Granulocyte Kinetics

in Health and Disease

Tamar Tak
Granulocyte Kinetics in
Health and Disease

Tamar Tak
ISBN: 978-94-6299-331-0

Cover artwork: Stella de Hoog

Printing: Ridderprint B.V., Ridderkerk

Printed on FSC certified paper

Studies in this thesis were supported by the Dutch Lung Foundation (grant 3.2.10.052),
Mucoviszidose e.V (grant 1209), Else Kröner-Fresenius-Stiftung (grant EKFS-A17), Stichting
Sanquin Bloedvoorziening and the Netherlands Cystic Fibrosis Foundation (Christina
onderzoekssubsidie 2012).

Publication of this thesis was supported by Infection & Immunity Utrecht, Buchem B.V.,
Raaijmakers en Tak Enterprises B.V. and Stichting Astmabestrijding.
 Granulocyte Kinetics in
Health and Disease
De kinetiek van granulocyten in gezondheid en in ziekte
(met een samenvatting in het Nederlands)

Proefschrift

ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de rector
magnificus, prof.dr. G.J. van der Zwaan, ingevolge het besluit van het college voor
promoties in het openbaar te verdedigen op dinsdag 31 mei 2016 des middags te 12.45
uur

door

Tamar Tak

geboren op 12 januari 1986 te Oosterhout

Promotor:
Prof. dr. L. Koenderman

Copromotoren:
Dr. J.A.M. Borghans
Dr. K. Tesselaar
Promotiecommissie:
Prof. dr. B.J. Prakken

Prof. dr. R.J. de Boer

Prof. dr. D. Roos

Prof. dr. D. Hartl

Prof. dr. P. Pickkers

Paranimfen:

Rémy Laferte

Bart Hilvering
Voor oma
Table of Contents

Chapter 1 General Introduction 11

Part I. Kinetics under homeostatic conditions

Chapter 2 What's your age again? Determination of human neutrophil 31

half-lives revisited
Chapter 3 Estimating human neutrophil circulatory lifespan by short- 49
term deuterated-glucose labelling in vivo
Chapter 4 Kinetics of eosinophils and basophils in the circulation of 71
healthy humans
Chapter 5 Circulatory and maturation kinetics of human monocyte 79
subsets in vivo

Part II. Kinetics in acute inflammation

Chapter 6 Immune suppression by neutrophils and granulocytic 97

myeloid-derived suppressor cells: similarities and differences
Chapter 7 Parallel differentiation of normal and immune suppressive 125
neutrophils in man
Chapter 8 Neutrophil Mac-1 (CD11b) mediates influenza-induced 149
pathology in mice

Part III. Kinetics in chronic pulmonary inflammation

Chapter 9 Similar activation state of neutrophils in sputum of asthma 171

patients irrespective of sputum eosinophilia
Chapter 10 Granulocyte kinetics in the blood and sputum of patients 191
with eosinophilic and neutrophilic inflammation
Chapter 11 General discussion 215

Addendum I Nederlandse samenvatting 245

II Dankwoord 253
III Curriculum Vitae 255
IV List of publications 256
 Chapter 1

General Introduction
 General introduction

The myeloid compartment of the immune system

Immune cells originating from the bone marrow are divided into several classes. The first
division is into lymphoid and myeloid cells; the latter being the focus of this thesis. The 1
myeloid compartment can be further subdivided into the granulocytic compartment
(neutrophils, eosinophils and basophils) and the mononuclear compartment, comprised of
monocytes, macrophages and myeloid dendritic cells. Mononuclear myeloid cells are
relatively large cells with a round nucleus, whereas granulocytes are characterised by a multi-
lobed nucleus and a cytoplasm filled with granules. Myeloid cells can have different roles in
the immune system. Neutrophils are the most abundant circulating white blood cells and are
essential for host defence particularly against bacteria (1). With between 2*1010 and 5*1011
neutrophils circulating in the bloodstream, they are the first cells to arrive on a site of
inflammation where they eliminate pathogens by engulfing and digesting them in a process
called phagocytosis. Dendritic cells are professional antigen presenting cells. After antigen
uptake, they degrade the protein and present it as small peptides on their membrane to cells
of the adaptive immune system. Macrophages can also phagocytose pathogens. In contrast
to neutrophils they reside mostly in tissue instead of in the bloodstream. They express an
array of scavenger receptors that are used for clearing debris and dead or dying cells.
Monocytes can phagocytose pathogens, but are considered to be more important as potent
sources of cytokines and chemokines, thereby shaping the following immune response. Both
basophils and eosinophils play a role in the allergic immune response and are thought to
protect from parasitic infections (2, 3). Eosinophils have granules filled by cytotoxic proteins,
which can be released to eliminate large pathogens, such as parasites. Recent literature has
shown however, that eosinophil deficient mice do not have a higher parasite load in some
infection models (4), whereas in others eosinophils actually increase survival of the parasite
(5, 6). This has led to a re-appraisal of the actual eosinophil function towards a more immune-
modulatory function, for example by mediating plasma cell survival (7) or release of cytokines
and chemokines (8). In addition, eosinophils have been implicated in tissue remodelling and
repair (9).
Basophils are also known for their release of cytokines and chemokines, in particular
histamine. Basophils release granule-stored histamine upon contact with allergens via
allergen-specific IgE (10) and initiate an allergic immune reaction. In addition, basophils are a
source of Il-4 (11), and can thereby steer the immune system towards a Th2 response (12).

Phagocytes – an evolutionary perspective

To underline the importance of innate immune cells, one can study their emergence during
evolution. The evolution theory states that life is a struggle for survival in which organisms
will only survive if they are adapted to their environment (13). This environment contains
large numbers of microorganisms, which can be pathogenic. Therefore, each organism
requires some form of protection from micro-organisms. Cells capable of protecting against
microorganisms by phagocytosis arose in evolution together with the first multi-cellular
organisms, around a billion years ago, and have been found in organisms as primitive as the

13
Chapter 1

slime mould (14). Professional phagocytic cells with a similar morphology as granulocytes
arose somewhat later in evolution in primitive invertebrates, such as sponges and jelly fish
(15, 16). Thus, phagocytosing immune cells are present in almost all species of the animal
kingdom (17). Traits of an organism that are not essential for survival are often lost during
evolution, so the highly conserved nature of the phagocytic immune cell highlights its
importance for host survival.

1. The lifecycle of a granulocyte

Cell production in the bone marrow – trees and controversies
The granulocyte lifecycle (figure 1) begins in the bone marrow. Granulocytes are terminally
differentiated cells not capable of division. So in order to replenish their number, immature
cells are produced in the bone marrow from progenitor cells. Several stages of maturation
are recognised in the bone marrow (18). Traditionally, haematopoiesis is pictured as a
decision tree, in which multipotent haematopoietic stem cells lose (pluri)potency with each
step of maturation (19). After differentiation into a committed myeloid progenitor, they were
originally thought to lose their capability of differentiating into lymphoid cells. After further
maturation into the myeloblast stage only granulocytes, monocytes and macrophages can be
produced and total commitment to the granulocytic lineage was thought to start at the
promyelocyte stage, with a different promyelocyte for each type of granulocyte (19). More
recently, however, these findings have been disputed in a study suggesting that neutrophils
develop separately from the other myeloid cells in its own lineage (20) and by several other
studies that question whether the myeloid and lymphoid progenitors are truly committed (21,
22). During this whole differentiation programme, progenitor cells have been shown to
undergo asymmetric divisions, in which one daughter cell is different from the other (23),
although other reports have shown only symmetrical divisions (24). In addition, cytokines and
growth factors can skew the production of a certain type of cell. (25) However, the underlying
mechanisms of the decision of progenitor cells to differentiate into one cell or another in vivo
remain to be elucidated. Taken together, these findings indicate that surprisingly little is
known about the process leading to production of myeloid cells.

Granulocyte production in the bone marrow – lazy cells and the post-mitotic pool
A lot further down the haematopoietic tree, additional interesting processes take place. For
granulocytes, more maturation stages are recognised: Each granulocyte type has its own line
of differentiation from promyelocyte into myelocyte, metamyelocyte, followed by immature
cells with a banded nuclear morphology and finally a mature cell (18). During these
maturation steps, the progenitor cells lose their capability for proliferation, which occurs
during differentiation from myelocyte to metamyelocyte, making the myelocytes the last
dividing progenitor. In contrast to the other granulocytic progenitors, the myelocytic
compartment has a unique feature: the lazy pool (26). The lazy pool has been described for
neutrophils and is thought to comprise approximately half of the myelocytic compartment
(27, 28). It consists of cells that divide very rarely compared to the other half of the myelocytic

14
 General introduction

pool. The speed at which they double has been estimated to be 71 days, whereas active
myelocytes divide approximately once every day (29, 30).
After differentiation from myelocytes into metamyelocytes, progenitor cells can no 1
longer go into mitosis and thus, enter the post-mitotic pool (PMP). The time cells spent in the
PMP differs between the different granulocytes, with eosinophils having a post-mitotic pool
transit time (PMPtt) of only 4 days, where basophils and neutrophils have a PMPtt of
approximately 6 days (31, 32).

Figure 1. Different stages during neutrophil development in homeostasis and inflammation. Neutrophils are
continuously produced in the bone marrow. The first committed neutrophil progenitor is the myeloblast.
Together with the promyelocytes and the myelocytes, they comprise the mitotic pool of neutrophil progenitors,
as these three cell types can still divide. A special feature of the myelocytes is that not all of them are actively
dividing. An estimated 50% is thought to be in a resting state, called the lazy pool. From the stage after the
myelocytes onwards, neutrophil progenitor lose their capacity to divide and produce new cells. These cells are
called metamyelocytes, banded neutrophils and, finally, mature neutrophils. As long as these cells remain in the
bone marrow, they are termed the post-mitotic pool. Mature segmented neutrophils can finally enter the
circulation, where they can either be freely flowing in the blood or be in the marginated pool. Lastly, the
hypersegmented neutrophil has been described, the origin of which is unknown. During inflammation,
marginated, hypersegmented and bone marrow neutrophils are recruited to the bloodstream and the
production of new neutrophils is increased. Although it seems likely that the lazy pool becomes active to
contribute to this increased neutrophil output, this has not yet been shown.

The mature granulocyte

After release into the bloodstream, granulocytes exist under normal circumstances as either
freely circulating cells or as part of the marginated pool (33). The marginated pool was
identified when isolated neutrophils were labelled and reinfused, upon which fewer labelled
neutrophils were recovered than expected (34). Upon injection of high doses of
corticosteroids or adrenalin, these cells were rapidly mobilised back into the circulation,
leading to the prevailing idea of a marginated pool from which neutrophils can be rapidly
mobilized. The fraction of marginated cells differs somewhat between experiments, but
seems to be approximately 50% for neutrophils, and in a single experiment with eosinophils
as much as 80% of labelled cells were lost from the circulation (35). Infusion of adrenalin or
high doses of corticosteroids induced a rapid doubling of neutrophil counts caused by

15
Chapter 1

recruitment of marginated neutrophils back into the bloodstream (36). As the percentage of
labelled cells was similar before and after recruitment of the marginated pool, the two pools
are assumed to be constantly exchanging cells and thus in equilibrium (33, 36). Marginated
granulocytes are assumed to be adhering to or rolling on the endothelium. This margination
was originally postulated to occur mainly in the spleen, liver and lungs (35, 37, 38), as
reinfusion of radiolabelled neutrophils led to accumulation of label in these organs. However,
the presence of marginated neutrophils in the lungs was later shown to be an activation
artefact (39). It is unknown to what extent the experiments suggesting the existence of a
marginated pool were affected by this artefact. For basophils, none of these experiments
have been performed, so the existence of a marginated pool for these cells is unknown.

2. The lifecycle of the monocyte

Monocytes in blood and bone marrow
In contrast to the many different granulocytic progenitors in the bone marrow, the production
of new monocytes is relatively simple (figure 2). After the myeloblast stage, committed
monocyte progenitors are recognised in the bone marrow as promonocytes and immature
monocytes (18, 40). The maturation from the last division into a circulating mature monocyte
takes only 0.5 to 1 days (31, 41), which is very short compared to the PMPtt of granulocytes.
The situation of monocytes in blood is more complicated, with three different
monocyte subsets being recognised in humans by flow cytometric evaluation (42):
CD14++/CD16- classical monocytes, CD14++/CD16+ double positive intermediate monocytes
and CD14low/CD16+ non-classical monocytes. Each of these subsets has been proposed to
have a different function: classical monocytes are best at uptake of antigens, intermediate
monocytes activate the adaptive immune system by antigen presentation (43) and the non-
classical monocytes show a patrolling behaviour in the vasculature (44). Measurements on all
monocytes combined have indicated a lifespan of only a few days in circulation (31, 41, 45).
Also for monocytes the existence of a marginated pool has been suggested, although this
notion was based solely on calculations with production numbers and lifespans (46).

Differentiation in the monocytic compartment

The differentiation pattern of the different monocyte subsets and macrophages remains a
matter of debate. Traditionally, monocytes were thought to represent an intermediate stage
of cells that became macrophages after homing into the tissue (47). More recently, however,
it has become clear that at least in mice many types of macrophages do not originate in the
bone marrow (48). The other types that do arise from haematopoietic progenitors in the bone
marrow migrate to their niche before birth (49, 50) and are not replenished by the bone
marrow. Macrophages that are constantly being renewed from the bone marrow include
CD14+ skin macrophages (51), some cardiac macrophages (52), and intestinal macrophages
(53). For at least the skin and gut macrophages, it was shown in mice that they specifically
originate from classical monocytes (51, 53) and this is also most likely the case for skin CD14+
macrophages in humans (51).

16
 General introduction

Figure 2. Different stages in mononuclear cell development. Blood monocytes, skin CD14+ macrophages and
intestinal DC’s are constantly renewed from cells produced in the bone marrow. The monoblast is the last
mononuclear progenitor still capable of division. When they mature into immature monocytes and
promonocytes, they no longer undergo mitosis and are therefore recognised as part of the PMP. Monocytes
remain in the PMP only for a short time until they are released into the blood. In the blood, three monocyte
subsets exist: classical, intermediate and non-classical monocytes. These cells are usually considered to be
different maturation stages of the same cell, but this remains to be proven conclusively. CD14+ macrophages
and intestinal DC’s are continuously replenished from circulating classical monocytes, whereas most other
macrophages and DS’s are replenished by local proliferation. There is evidence for the existence of a
mononuclear marginated pool, but it is unknown of which monocyte subset this pool is comprised.

How the three monocyte subsets in blood are related is also unsure. In humans, IM
are placed between CM and NCM on the basis of: 1. surface marker expression measured by
flow cytometry (54), 2. gene expression patterns (43, 55) and 3. enhancer profiles in promotor
regions in the DNA (56). This has led to the prevailing idea that monocytes follow a linear
differentiation path from classical to intermediate to non-classical monocyte. Pulse-chase
labelling in macaques (57) and repopulation kinetics after HSCT strengthen this notion as
classical monocytes are the first to be labelled/repopulated, followed by the intermediate
and non-classical monocytes (51). This indicates that if the subsets are developmentally
related, the differentiation would have to occur from classical to intermediate to non-classical
monocyte. However, this developmental relation remains to be proven and current evidence
in humans is circumstantial at best.
Also in mice two monocyte subsets have been identified that are homologous to the
human classical and non-classical monocytes (58), although clear differences have also been
reported (59). Using microscopy (60) and adoptive transfer (53) of labelled cells, it has been
suggested that these murine subsets differentiate from classical into non-classical monocytes.

17
Chapter 1

Of note, however, the former study was performed during inflammation, and the latter
results might have been confounded by (undesired) co-transfer of monocytic progenitor cells
(61, 62).

3. Granulocyte kinetics in disease

Granulocytes in acute inflammation
Acute inflammation is found in patients with severe trauma (63), burn trauma (64), or sepsis
and in healthy individuals in response to systemic LPS challenge (65). Under these conditions
the number of freely flowing neutrophils in the bloodstream increases rapidly (36, 65). The
neutrophils responsible for this increase in the human endotoxemia model have been shown
to be different from those in the marginated pool (36). In fact these cells were shown to
consist of three phenotypically different neutrophils (65): CD16dim/CD62Lhigh neutrophils with
a banded nuclear phenotype, CD16high/CD62Ldim neutrophils which have on average more
nuclear segments and are therefore named hypersegmented neutrophils and lastly,
CD16high/CD62Lhigh neutrophils, which have a normal nuclear morphology. In the rest of the
thesis these cells are referred to as banded cells, hypersegmented cells and mature cells,
respectively. The three subsets have differences in functionality: banded neutrophils have
superior bactericidal capabilities (unpublished results by P. Leliefeld), whereas
hypersegmented neutrophils are deficient in bacterial killing. In contrast, the latter cells are
shown to inhibit T-cell proliferation in vitro (65). Although banded neutrophils observed in
the circulation during acute inflammation are phenotypically similar to banded bone marrow
neutrophils, and even though hypersegmented neutrophils have characteristics of aged
neutrophils (66, 67), the origin and kinetics of these two subsets remain to be elucidated
(figure 1).

Granulocytes in the lungs during inflammation

In healthy individuals only small numbers of granulocytes are found in the lung lumen (68).
However, when the lung is inflamed, such as in patients with cystic fibrosis (CF) or asthma,
granulocytes leave the bloodstream and migrate into the lungs (69, 70). In CF patients, mostly
neutrophils are found in the lungs, whereas in asthma patients both neutrophils and
eosinophils are found in increased numbers. In fact, different subtypes of asthma are
identified based on which granulocyte is present in the sputum (70): patients with more than
56% neutrophils and few eosinophils (<3%) in their sputum are considered to have
neutrophilic asthma, patients with >3% eosinophils and with “low” percentages of
neutrophils (<56%) are considered to have eosinophilic asthma. If both granulocytes are
present in high numbers, the phenotype is referred to as mixed. Lastly, patients in which both
cell types are present in numbers below the cut-off values of 56% and 3% are considered to
have the paucigranulocytic type of asthma (71).
In both diseases, granulocytes are thought to contribute to the inflammation in the
lungs. Deposition of neutrophil and eosinophil granule proteins (neutrophil elastase and
eosinophil major basic protein) is observed in both diseases. This indicates that degranulation

18
 General introduction

and release of these toxic compounds occurs and can cause damage to the lungs. In fact, the
amount of granule protein depositions was shown to correlate with epithelial damage in
asthma patients (72). In addition, neutrophils have been shown to be able to release their 1
DNA in a process called NETosis (73). The lungs of CF patients contain large amounts of free
DNA, the amount of which correlates to airflow obstruction (74). This free DNA contributes
to the high viscosity of the lining fluid that is characteristic of the disease, and, inhaled DNAse
is prescribed in these patients to treat the high viscosity that causes so much trouble (75).
Therefore, a role for granulocytes in both asthma and CF is highly likely.

Why determine granulocyte kinetics?

As granulocytes are implicated to contribute to the pathology of many diseases, it is logical to
use or develop therapeutics targeting these cells. However, interpreting the effects of these
therapeutics can be difficult when the kinetics of the cells they target are unknown. A good
example for difficulties in interpreting results can be found in the studies describing the effect
of anti-IL5, mepolizumab, in the treatment of eosinophilic asthma (76). Eosinophils are
dependent on the presence of IL-5 during their development in the bone marrow (77).
Mepolizumab is a monoclonal antibody targeting IL-5, which is very effective in reducing
eosinophil numbers in blood in hypereosinophilic syndrome (78, 79). As eosinophils are
implicated in the pathology of at least eosinophilic asthma (80), there were high expectations
for the use of mepolizumab in this disease. However, even though mepolizumab potently
reduced eosinophil numbers in the blood within 4 weeks, only a reduction of approximately
50% was observed in bronchial mucosa and sputum up to 20 weeks after the start of therapy
(76, 81). In addition, one study showed a marked delay (12 weeks) between the reduction of
blood eosinophil counts and reduction of exacerbations (82), although another study
suggested immediate clinical efficacy (83). Taken together, these studies raise important
questions regarding eosinophil kinetics in the lungs. Do eosinophils survive for a long time in
the lungs? Or alternatively, could the few eosinophils that are produced in the presence of
mepolizumab immediately migrate into the lungs, thereby explaining the large numbers of
eosinophils in the lungs in parallel with low numbers in the blood?
Another example lies in the interpretation of the effects of corticosteroids on disease.
Many in vitro studies have shown that corticosteroids can diminish cytokine-induced
eosinophil survival (84), whereas the same concentration of steroids prevents neutrophil
apoptosis in vitro (85, 86). This suggests that whilst reducing eosinophil-mediated
inflammation, corticosteroids might aggravate neutrophil-mediated inflammation. However,
it is unknown whether neutrophil apoptosis in vitro has any relation to the situation in vivo,
highlighting again, the need for reliable in vivo kinetical data.

Controversies regarding determination of granulocyte lifespan

Since determining the kinetics of granulocytes in the lungs and circulation is so important,
why are so little data available regarding this issue? As described above, granulocytes are
found in several compartments in the body: the bone marrow, the blood, the marginated pool

19
Chapter 1

and the tissues. Unfortunately, only the blood is easily accessible for kinetic analyses in
humans. In addition, most of the labelling methods used to track cells in vivo are toxic,
including radioactive tracers like tritium thymidine (3H-TdR), chromium (51Cr), technetium
(99Tc) or indium (111In), the mutagen BrdU or the nerve gas DFP, which has been used in
radioactive forms with 3H and 32P. The toxicity of 99Tc (87), 111In (87) and DFP (88) has been
shown to activate neutrophils. In addition, many of these labelling techniques require
neutrophils to be taken out of the body, isolated and re-infused, a process which at least in
the case of neutrophils has been shown to lead to aberrant homing characteristics (39). Due
to their toxicity, in vivo labelling techniques such as 3H-TdR administration have only been
employed in terminal cancer patients (31, 32, 89, 90), whom are likely to have altered
granulocyte kinetics due to the cancer-induced inflammation and or treatment. On top of
that, 3H-TdR has been suggested to alter eosinophil kinetics in vivo (90). Most textbooks
report a neutrophil half-life in circulation of 7h (91), which corresponds to an average
circulatory lifespan of approximately 10h. In the study that is referred to by most of these
textbooks, prisoners were injected with DF32P, which binds very strongly to neutrophil
elastase, and thus mostly labels neutrophils. The amount of radioactivity in the neutrophil
fraction decreased rapidly in this study, leading to the conclusion of a short neutrophil
circulatory half-life of 7 hrs. However, this approach has been shown to lead to a rapid loss of
neutrophils from the circulation (92) and the 7h half-life might, therefore, reflect the kinetics
of DFP-induced toxicity instead of the true neutrophil circulatory kinetics. These issues will be
discussed in detail in chapter two of this thesis.
More recently, non-toxic labelling strategies have been developed using stable isotope
(non-radioactive) labelling with deuterium (2H) instead of radioactive tritium (3H) (93).
Normally, only 0.015% of the hydrogen in the human body is enriched (2H), with the
remainder being protium (1H). This isotope is non-toxic. Only very high doses of deuterium,
with at least 30% of body water being deuterated have been reported to reduce the fertility
of male mice by decreasing spermatogenesis (94, 95) and have been shown to enhance
basophil activation ex vivo (96). Below these levels, however, deuterium labelling is
considered a safe isotope to perform kinetic studies. Body water enrichments of around 2
percent are more than sufficient to determine cell kinetics when administered as 2H2O (97),
and when deuterium is administered as glucose (6,6-2H2-glucose) the contribution of
deuterium to the total body water hydrogen pool is negligible.
Using these techniques, some estimates have been made regarding granulocyte
kinetics in the circulation. Using 6,6-2H2-glucose, neutrophil kinetics have been studied (98),
but as cells were followed too short to determine the complete uplabelling curve and no
downlabelling was studied, only the PMPtt could be determined from this study. An attempt
by our own group to determine circulatory neutrophil kinetics with 2H2O (99) was hampered
by the relatively slow kinetics of the label in the human body in comparison with the rapid
neutrophil kinetics (see also chapter 2). In addition, Li et al correctly pointed out that the
estimated average lifespan of 5.4 days that we previously reported (which corresponds to a
3.75d half-life) may not reflect the lifespan of neutrophils in blood, but instead a 3.75 day

20
 General introduction

division time of neutrophil progenitors in the bone marrow (BM) (100). Thus, there is ample
uncertainty regarding the circulatory kinetics of neutrophils in the circulation, whereas for
the other two types of granulocytes no in vivo labelling data from healthy volunteers is 1
available at all. Therefore, employing this technique will reveal crucial insights into the biology
of these cells.

Scope and outline of this thesis

The first part of this thesis aims to determine the kinetics of myeloid cells in healthy humans.
In the 1950’s and 1960’s many studies were performed to elucidate these kinetics, the results
of which can nowadays be found in text books. Unfortunately, the methods available in those
days were not ideal and cannot be reproduced for obvious ethical reasons. Chapter 2 of this
thesis presents an overview of methods used for tracking granulocytes, their resulting
conclusions and limitations. Subsequently, studies are presented to elucidate the kinetics of
myeloid cells by deuterated glucose administration, a technique that overcomes most of
these limitations and thereby represents a more accurate estimation of the kinetics of
neutrophils in chapter 3, of eosinophils and basophils in chapter 4, and of the three types of
monocytes in chapter 5. In addition, using computational modelling, we aim to elucidate
whether monocyte can differentiate from CM into IM and subsequently into NCM or whether
these three subsets are each a different cell type, differentiating separately from progenitors
in the bone marrow.
The second part of this thesis studies the kinetics of cells during acute inflammation.
These studies start with chapter 6, which presents a detailed analysis of the immune
suppressive neutrophils that are found in different types of disease. Chapter 7 also focuses
on neutrophil subsets in disease: the kinetics of the three neutrophil subsets found only
during acute inflammation will be determined in a model for human endotoxemia using
deuterated glucose. These findings are combined with a proteomics approach to shed light
on developmental relations between these three subsets. Chapter 8 presents another study
performed in the human endotoxemia model, but focusses on the kinetics of the three
monocyte subsets by multidimensional FACS analysis. The last chapter of this part, chapter 9
aims to determine whether suppressive neutrophils play a role in disease. In a murine model
of influenza infection, disease characteristics are compared between mice capable of CD11b-
mediated immune suppression and mice that lack CD11b and thus, cannot employ this
surface molecule in immune suppression.
The third and last part of this thesis describes granulocytes in respiratory diseases. It
starts with chapter 10, a flow-cytometric comparison between the neutrophils and
eosinophils in blood and induced sputum samples of asthma patients and concludes that
there are no differences in expression patterns between patients with different asthma
phenotypes. Finally, in chapter 11 we determined granulocyte kinetics in the blood and lungs
of patients with eosinophilic asthma and cystic fibrosis, again by in vivo deuterium labelling.
Lastly, chapter 12 discusses the results of the previous chapters in a broader perspective.

21
Chapter 1

References
1. Lekstrom-Himes JA, and Gallin JI. Immunodeficiency diseases caused by defects in
phagocytes. The New England journal of medicine. 2000;343(23):1703-14.
2. Klion AD, and Nutman TB. The role of eosinophils in host defense against helminth
parasites. The Journal of allergy and clinical immunology. 2004;113(1):30-7.
3. Voehringer D. The role of basophils in helminth infection. Trends in parasitology.
2009;25(12):551-6.
4. Swartz JM, Dyer KD, Cheever AW, Ramalingam T, Pesnicak L, Domachowske JB, Lee JJ,
Lee NA, Foster PS, Wynn TA, et al. Schistosoma mansoni infection in eosinophil
lineage-ablated mice. Blood. 2006;108(7):2420-7.
5. Gebreselassie NG, Moorhead AR, Fabre V, Gagliardo LF, Lee NA, Lee JJ, and Appleton
JA. Eosinophils preserve parasitic nematode larvae by regulating local immunity. J
Immunol. 2012;188(1):417-25.
6. Fabre V, Beiting DP, Bliss SK, Gebreselassie NG, Gagliardo LF, Lee NA, Lee JJ, and
Appleton JA. Eosinophil deficiency compromises parasite survival in chronic nematode
infection. J Immunol. 2009;182(3):1577-83.
7. Chu VT, Frohlich A, Steinhauser G, Scheel T, Roch T, Fillatreau S, Lee JJ, Lohning M, and
Berek C. Eosinophils are required for the maintenance of plasma cells in the bone
marrow. Nature immunology. 2011;12(2):151-9.
8. Jacobsen EA, Taranova AG, Lee NA, and Lee JJ. Eosinophils: singularly destructive
effector cells or purveyors of immunoregulation? The Journal of allergy and clinical
immunology. 2007;119(6):1313-20.
9. Kay AB, Phipps S, and Robinson DS. A role for eosinophils in airway remodelling in
asthma. Trends in immunology. 2004;25(9):477-82.
10. Ishizaka T, De Bernardo R, Tomioka H, Lichtenstein LM, and Ishizaka K. Identification
of basophil granulocytes as a site of allergic histamine release. J Immunol.
1972;108(4):1000-8.
11. Min B, Prout M, Hu-Li J, Zhu J, Jankovic D, Morgan ES, Urban JF, Jr., Dvorak AM,
Finkelman FD, LeGros G, et al. Basophils produce IL-4 and accumulate in tissues after
infection with a Th2-inducing parasite. J Exp Med. 2004;200(4):507-17.
12. Zhu J, Yamane H, and Paul WE. Differentiation of effector CD4 T cell populations (*).
Annual review of immunology. 2010;28(445-89.
13. Darwin C. On the origin of species. New York: D. Appleton and Co.; 1871.
14. Chen G, Zhuchenko O, and Kuspa A. Immune-like phagocyte activity in the social
amoeba. Science. 2007;317(5838):678-81.
15. Amulic B, Cazalet C, Hayes GL, Metzler KD, and Zychlinsky A. Neutrophil function: from
mechanisms to disease. Annual review of immunology. 2012;30(459-89.
16. Patterson MJ, and Landolt ML. Cellular Reaction to Injury in the Anthozoan
Anthopleura-Elegantissima. J Invertebr Pathol. 1979;33(2):189-96.
17. Ribeiro C, and Brehelin M. Insect haemocytes: what type of cell is that? J Insect Physiol.
2006;52(5):417-29.
18. Schudel L. Leitfaden der Blutmorphologie: Manual of blood morphology. Zürich:
Thieme; 1965.
19. Kawamoto H, Wada H, and Katsura Y. A revised scheme for developmental pathways
of hematopoietic cells: the myeloid-based model. Int Immunol. 2010;22(2):65-70.

22
 General introduction

20. Gorgens A, Radtke S, Mollmann M, Cross M, Durig J, Horn PA, and Giebel B. Revision
of the human hematopoietic tree: granulocyte subtypes derive from distinct
hematopoietic lineages. Cell reports. 2013;3(5):1539-52. 1
21. Adolfsson J, Mansson R, Buza-Vidas N, Hultquist A, Liuba K, Jensen CT, Bryder D, Yang
L, Borge OJ, Thoren LA, et al. Identification of Flt3+ lympho-myeloid stem cells lacking
erythro-megakaryocytic potential a revised road map for adult blood lineage
commitment. Cell. 2005;121(2):295-306.
22. Wada H, Masuda K, Satoh R, Kakugawa K, Ikawa T, Katsura Y, and Kawamoto H. Adult
T-cell progenitors retain myeloid potential. Nature. 2008;452(7188):768-72.
23. Beckmann J, Scheitza S, Wernet P, Fischer JC, and Giebel B. Asymmetric cell division
within the human hematopoietic stem and progenitor cell compartment:
identification of asymmetrically segregating proteins. Blood. 2007;109(12):5494-501.
24. Boll I, and Kuhn A. Granulocytopoiesis in human bone marrow cultures studied by
means of kinematography. Blood. 1965;26(4):449-70.
25. Lodish H. Molecular Cell Biology. W. H. Freeman; 2004.
26. Lajtha LG, Gilbert CW, Porteous DD, and Alexanian R. Kinetics of a Bone-Marrow Stem-
Cell Population. Annals of the New York Academy of Sciences. 1964;113(742-52.
27. Dresch C, Troccoli G, and Mary JY. Growth fraction of myelocytes in normal human
granulopoiesis. Cell and tissue kinetics. 1986;19(1):11-22.
28. Mary JY. Normal human granulopoiesis revisited. II. Bone marrow data. Biomedicine
& pharmacotherapy = Biomedecine & pharmacotherapie. 1985;39(2):66-77.
29. Boll IT, and Fuchs G. A kinetic model of granulocytopoiesis. Experimental cell research.
1970;61(1):147-52.
30. Stryckmans P, Cronkite EP, Fache J, Fliedner TM, and Ramos J. Deoxyribonucleic acid
synthesis time of erythropoietic and granulopoietic cells in human beings. Nature.
1966;211(5050):717-20.
31. Cronkite EP, Fliedner TM, Bond VP, and Rubini JR. Dynamics of hemopoietic
proliferation in man and mice studied by H3-thymidine incorporation into DNA. Annals
of the New York Academy of Sciences. 1959;77(803-20.
32. Steinbach KH, Schick P, Trepel F, Raffler H, Dohrmann J, Heilgeist G, Heltzel W, Li K,
Past W, van der Woerd-de Lange JA, et al. Estimation of kinetic parameters of
neutrophilic, eosinophilic, and basophilic granulocytes in human blood. Blut.
1979;39(1):27-38.
33. Summers C, Rankin SM, Condliffe AM, Singh N, Peters AM, and Chilvers ER. Neutrophil
kinetics in health and disease. Trends in immunology. 2010;31(8):318-24.
34. Athens JW, Raab SO, Haab OP, Mauer AM, Ashenbrucker H, Cartwright GE, and
Wintrobe MM. Leukokinetic studies. III. The distribution of granulocytes in the blood
of normal subjects. The Journal of clinical investigation. 1961;40(159-64.
35. Farahi N, Singh NR, Heard S, Loutsios C, Summers C, Solanki CK, Solanki K, Balan KK,
Ruparelia P, Peters AM, et al. Use of 111-Indium-labeled autologous eosinophils to
establish the in vivo kinetics of human eosinophils in healthy subjects. Blood.
2012;120(19):4068-71.
36. Athens JW, Haab OP, Raab SO, Mauer AM, Ashenbrucker H, Cartwright GE, and
Wintrobe MM. Leukokinetic studies. IV. The total blood, circulating and marginal
granulocyte pools and the granulocyte turnover rate in normal subjects. The Journal
of clinical investigation. 1961;40(989-95.

23
Chapter 1

37. Saverymuttu SH, Peters AM, Keshavarzian A, Reavy HJ, and Lavender JP. The kinetics
of 111indium distribution following injection of 111indium labelled autologous
granulocytes in man. British journal of haematology. 1985;61(4):675-85.
38. Szczepura KR, Ruparelia P, Solanki CK, Balan K, Newbold P, Summers C, Chilvers ER,
and Peters AM. Measuring whole-body neutrophil redistribution using a dedicated
whole-body counter and ultra-low doses of 111Indium. European journal of clinical
investigation. 2011;41(1):77-83.
39. Summers C, Singh NR, White JF, Mackenzie IM, Johnston A, Solanki C, Balan KK, Peters
AM, and Chilvers ER. Pulmonary retention of primed neutrophils: a novel protective
host response, which is impaired in the acute respiratory distress syndrome. Thorax.
2014;69(7):623-9.
40. Goasguen JE, Bennett JM, Bain BJ, Vallespi T, Brunning R, Mufti GJ, and International
Working Group on Morphology of Myelodysplastic S. Morphological evaluation of
monocytes and their precursors. Haematologica. 2009;94(7):994-7.
41. Whitelaw DM. Observations on human monocyte kinetics after pulse labeling. Cell and
tissue kinetics. 1972;5(4):311-7.
42. Ziegler-Heitbrock L, Ancuta P, Crowe S, Dalod M, Grau V, Hart DN, Leenen PJ, Liu YJ,
MacPherson G, Randolph GJ, et al. Nomenclature of monocytes and dendritic cells in
blood. Blood. 2010;116(16):e74-80.
43. Zawada AM, Rogacev KS, Rotter B, Winter P, Marell RR, Fliser D, and Heine GH.
SuperSAGE evidence for CD14++CD16+ monocytes as a third monocyte subset. Blood.
2011;118(12):e50-61.
44. Cros J, Cagnard N, Woollard K, Patey N, Zhang SY, Senechal B, Puel A, Biswas SK,
Moshous D, Picard C, et al. Human CD14dim monocytes patrol and sense nucleic acids
and viruses via TLR7 and TLR8 receptors. Immunity. 2010;33(3):375-86.
45. Mohri H, Perelson AS, Tung K, Ribeiro RM, Ramratnam B, Markowitz M, Kost R, Hurley
A, Weinberger L, Cesar D, et al. Increased turnover of T lymphocytes in HIV-1 infection
and its reduction by antiretroviral therapy. J Exp Med. 2001;194(9):1277-87.
46. van Furth R, and Sluiter W. Distribution of blood monocytes between a marginating
and a circulating pool. J Exp Med. 1986;163(2):474-9.
47. van Furth R, and Cohn ZA. The origin and kinetics of mononuclear phagocytes. J Exp
Med. 1968;128(3):415-35.
48. Epelman S, Lavine KJ, and Randolph GJ. Origin and functions of tissue macrophages.
Immunity. 2014;41(1):21-35.
49. Hashimoto D, Chow A, Noizat C, Teo P, Beasley MB, Leboeuf M, Becker CD, See P, Price
J, Lucas D, et al. Tissue-resident macrophages self-maintain locally throughout adult
life with minimal contribution from circulating monocytes. Immunity. 2013;38(4):792-
804.
50. Orkin SH, and Zon LI. Hematopoiesis: an evolving paradigm for stem cell biology. Cell.
2008;132(4):631-44.
51. McGovern N, Schlitzer A, Gunawan M, Jardine L, Shin A, Poyner E, Green K, Dickinson
R, Wang XN, Low D, et al. Human dermal CD14(+) cells are a transient population of
monocyte-derived macrophages. Immunity. 2014;41(3):465-77.
52. Epelman S, Lavine KJ, Beaudin AE, Sojka DK, Carrero JA, Calderon B, Brija T, Gautier EL,
Ivanov S, Satpathy AT, et al. Embryonic and adult-derived resident cardiac
macrophages are maintained through distinct mechanisms at steady state and during
inflammation. Immunity. 2014;40(1):91-104.

24
 General introduction

53. Varol C, Landsman L, Fogg DK, Greenshtein L, Gildor B, Margalit R, Kalchenko V,

Geissmann F, and Jung S. Monocytes give rise to mucosal, but not splenic,
conventional dendritic cells. J Exp Med. 2007;204(1):171-80. 1
54. Hijdra D, Vorselaars AD, Grutters JC, Claessen AM, and Rijkers GT. Phenotypic
characterization of human intermediate monocytes. Frontiers in immunology.
2013;4(339.
55. Wong KL, Tai JJ, Wong WC, Han H, Sem X, Yeap WH, Kourilsky P, and Wong SC. Gene
expression profiling reveals the defining features of the classical, intermediate, and
nonclassical human monocyte subsets. Blood. 2011;118(5):e16-31.
56. Schmidl C, Renner K, Peter K, Eder R, Lassmann T, Balwierz PJ, Itoh M, Nagao-Sato S,
Kawaji H, Carninci P, et al. Transcription and enhancer profiling in human monocyte
subsets. Blood. 2014;123(17):e90-9.
57. Sugimoto C, Hasegawa A, Saito Y, Fukuyo Y, Chiu KB, Cai Y, Breed MW, Mori K, Roy CJ,
Lackner AA, et al. Differentiation Kinetics of Blood Monocytes and Dendritic Cells in
Macaques: Insights to Understanding Human Myeloid Cell Development. J Immunol.
2015.
58. Sunderkotter C, Nikolic T, Dillon MJ, Van Rooijen N, Stehling M, Drevets DA, and
Leenen PJ. Subpopulations of mouse blood monocytes differ in maturation stage and
inflammatory response. J Immunol. 2004;172(7):4410-7.
59. Ingersoll MA, Spanbroek R, Lottaz C, Gautier EL, Frankenberger M, Hoffmann R, Lang
R, Haniffa M, Collin M, Tacke F, et al. Comparison of gene expression profiles between
human and mouse monocyte subsets. Blood. 2010;115(3):e10-9.
60. Dal-Secco D, Wang J, Zeng Z, Kolaczkowska E, Wong CH, Petri B, Ransohoff RM, Charo
IF, Jenne CN, and Kubes P. A dynamic spectrum of monocytes arising from the in situ
reprogramming of CCR2+ monocytes at a site of sterile injury. J Exp Med.
2015;212(4):447-56.
61. Geissmann F, Manz MG, Jung S, Sieweke MH, Merad M, and Ley K. Development of
monocytes, macrophages, and dendritic cells. Science. 2010;327(5966):656-61.
62. Shi C, and Pamer EG. Monocyte recruitment during infection and inflammation.
Nature reviews Immunology. 2011;11(11):762-74.
63. Visser T, Pillay J, Pickkers P, Leenen LP, and Koenderman L. Homology in systemic
neutrophil response induced by human experimental endotoxemia and by trauma.
Shock. 2012;37(2):145-51.
64. Peter FW, Schuschke DA, Barker JH, Fleishcher-Peter B, Pierangeli S, Vogt PM, and
Steinau HU. The effect of severe burn injury on proinflammatory cytokines and
leukocyte behavior: its modulation with granulocyte colony-stimulating factor. Burns.
1999;25(6):477-86.
65. Pillay J, Kamp VM, van Hoffen E, Visser T, Tak T, Lammers JW, Ulfman LH, Leenen LP,
Pickkers P, and Koenderman L. A subset of neutrophils in human systemic
inflammation inhibits T cell responses through Mac-1. The Journal of clinical
investigation. 2012;122(1):327-36.
66. Pillay J, Tak T, Kamp VM, and Koenderman L. Immune suppression by neutrophils and
granulocytic myeloid-derived suppressor cells: similarities and differences. Cellular
and molecular life sciences : CMLS. 2013.
67. Nauseef WM, and Borregaard N. Neutrophils at work. Nature immunology.
2014;15(7):602-11.

25
Chapter 1

68. Fortunati E, Kazemier KM, Grutters JC, Koenderman L, and Van den Bosch v J. Human
neutrophils switch to an activated phenotype after homing to the lung irrespective of
inflammatory disease. Clinical and experimental immunology. 2009;155(3):559-66.
69. Gibson RL, Burns JL, and Ramsey BW. Pathophysiology and management of pulmonary
infections in cystic fibrosis. Am J Respir Crit Care Med. 2003;168(8):918-51.
70. Simpson JL, Scott R, Boyle MJ, and Gibson PG. Inflammatory subtypes in asthma:
assessment and identification using induced sputum. Respirology. 2006;11(1):54-61.
71. Vijverberg SJ, Hilvering B, Raaijmakers JA, Lammers JW, Maitland-van der Zee AH, and
Koenderman L. Clinical utility of asthma biomarkers: from bench to bedside. Biologics.
2013;7(199-210.
72. Fujisawa T, Kephart GM, Gray BH, and Gleich GJ. The neutrophil and chronic allergic
inflammation. Immunochemical localization of neutrophil elastase. Am Rev Respir Dis.
1990;141(3):689-97.
73. Brinkmann V, Reichard U, Goosmann C, Fauler B, Uhlemann Y, Weiss DS, Weinrauch
Y, and Zychlinsky A. Neutrophil extracellular traps kill bacteria. Science.
2004;303(5663):1532-5.
74. Marcos V, Zhou-Suckow Z, Onder Yildirim A, Bohla A, Hector A, Vitkov L, Krautgartner
WD, Stoiber W, Griese M, Eickelberg O, et al. Free DNA in cystic fibrosis airway fluids
correlates with airflow obstruction. Mediators Inflamm. 2015;2015(408935.
75. Hodson ME. Aerosolized dornase alfa (rhDNase) for therapy of cystic fibrosis. Am J
Respir Crit Care Med. 1995;151(3 Pt 2):S70-4.
76. Flood-Page PT, Menzies-Gow AN, Kay AB, and Robinson DS. Eosinophil's role remains
uncertain as anti-interleukin-5 only partially depletes numbers in asthmatic airway.
Am J Respir Crit Care Med. 2003;167(2):199-204.
77. Menzies-Gow A, Flood-Page P, Sehmi R, Burman J, Hamid Q, Robinson DS, Kay AB, and
Denburg J. Anti-IL-5 (mepolizumab) therapy induces bone marrow eosinophil
maturational arrest and decreases eosinophil progenitors in the bronchial mucosa of
atopic asthmatics. The Journal of allergy and clinical immunology. 2003;111(4):714-9.
78. Rothenberg ME, Klion AD, Roufosse FE, Kahn JE, Weller PF, Simon HU, Schwartz LB,
Rosenwasser LJ, Ring J, Griffin EF, et al. Treatment of patients with the
hypereosinophilic syndrome with mepolizumab. The New England journal of medicine.
2008;358(12):1215-28.
79. Garrett JK, Jameson SC, Thomson B, Collins MH, Wagoner LE, Freese DK, Beck LA,
Boyce JA, Filipovich AH, Villanueva JM, et al. Anti-interleukin-5 (mepolizumab) therapy
for hypereosinophilic syndromes. The Journal of allergy and clinical immunology.
2004;113(1):115-9.
80. Busse WW, and Lemanske RF, Jr. Asthma. The New England journal of medicine.
2001;344(5):350-62.
81. Flood-Page P, Swenson C, Faiferman I, Matthews J, Williams M, Brannick L, Robinson
D, Wenzel S, Busse W, Hansel TT, et al. A study to evaluate safety and efficacy of
mepolizumab in patients with moderate persistent asthma. Am J Respir Crit Care Med.
2007;176(11):1062-71.
82. Nair P, Pizzichini MM, Kjarsgaard M, Inman MD, Efthimiadis A, Pizzichini E, Hargreave
FE, and O'Byrne PM. Mepolizumab for prednisone-dependent asthma with sputum
eosinophilia. The New England journal of medicine. 2009;360(10):985-93.
83. Haldar P, Brightling CE, Hargadon B, Gupta S, Monteiro W, Sousa A, Marshall RP,
Bradding P, Green RH, Wardlaw AJ, et al. Mepolizumab and exacerbations of

26
 General introduction

refractory eosinophilic asthma. The New England journal of medicine.

2009;360(10):973-84.
84. Lamas AM, Leon OG, and Schleimer RP. Glucocorticoids inhibit eosinophil responses 1
to granulocyte-macrophage colony-stimulating factor. J Immunol. 1991;147(1):254-9.
85. Meagher LC, Cousin JM, Seckl JR, and Haslett C. Opposing effects of glucocorticoids on
the rate of apoptosis in neutrophilic and eosinophilic granulocytes. J Immunol.
1996;156(11):4422-8.
86. McColl A, Michlewska S, Dransfield I, and Rossi AG. Effects of glucocorticoids on
apoptosis and clearance of apoptotic cells. ScientificWorldJournal. 2007;7(1165-81.
87. Bertrand-Caix J, Freyburger G, Bordenave L, Labrouche S, Bareille R, Baquey C,
Ducassou D, and Lefebvre F. Functional upregulation of granulocytes labeled with
technetium-99m-HMPAO and indium-111-oxinate. Journal of nuclear medicine :
official publication, Society of Nuclear Medicine. 1996;37(5):863-8.
88. Tsang JL, Parodo JC, and Marshall JC. Regulation of apoptosis and priming of
neutrophil oxidative burst by diisopropyl fluorophosphate. Journal of inflammation.
2010;7(32.
89. Fliedner TM, Cronkite EP, Killmann SA, and Bond VP. Granulocytopoiesis. Ii.
Emergence and Pattern of Labeling of Neutrophilic Granulocytes in Humans. Blood.
1964;24(683-700.
90. Parwaresch MR, Walle AJ, and Arndt D. The peripheral kinetics of human radiolabelled
eosinophils. Virchows Archiv B: Cell pathology. 19761):57-66.
91. Cartwright GE, Athens JW, and Wintrobe MM. The Kinetics of Granulopoiesis in
Normal Man. Blood. 1964;24(780-803.
92. Dresch C, Najean Y, and Beauchet J. In vitro 51Cr and 32P-DFP labeling of granulocytes
in man. Journal of nuclear medicine : official publication, Society of Nuclear Medicine.
1971;12(12):774-84.
93. Busch R, Neese RA, Awada M, Hayes GM, and Hellerstein MK. Measurement of cell
proliferation by heavy water labeling. Nature protocols. 2007;2(12):3045-57.
94. Hughes AM, Bennett EL, and Calvin M. Production of Sterility in Mice by Deuterium
Oxide. Proc Natl Acad Sci U S A. 1959;45(4):581-6.
95. Oakberg EF, and Hughes AM. Deuterium oxide effect on spermatogenesis in the
mouse. Experimental cell research. 1968;50(2):306-14.
96. Tung R, and Lichtenstein LM. In vitro histamine release from basophils of asthmatic
and atopic individuals in D2O. J Immunol. 1982;128(5):2067-72.
97. Vrisekoop N, den Braber I, de Boer AB, Ruiter AF, Ackermans MT, van der Crabben SN,
Schrijver EH, Spierenburg G, Sauerwein HP, Hazenberg MD, et al. Sparse production
but preferential incorporation of recently produced naive T cells in the human
peripheral pool. Proc Natl Acad Sci U S A. 2008;105(16):6115-20.
98. Macallan DC, Fullerton CA, Neese RA, Haddock K, Park SS, and Hellerstein MK.
Measurement of cell proliferation by labeling of DNA with stable isotope-labeled
glucose: studies in vitro, in animals, and in humans. Proc Natl Acad Sci U S A.
1998;95(2):708-13.
99. Pillay J, den Braber I, Vrisekoop N, Kwast LM, de Boer RJ, Borghans JA, Tesselaar K, and
Koenderman L. In vivo labeling with 2H2O reveals a human neutrophil lifespan of 5.4
days. Blood. 2010;116(4):625-7.
100. Li KW, Turner SM, Emson CL, Hellerstein MK, and Dale DC. Deuterium and neutrophil
kinetics. Blood. 2011;117(22):6052-3; author reply 3-4.

27
 Part I

Kinetics under homeostatic conditions

 Chapter 2
What’s your age again? – Determination of
human neutrophil half-lives revisited

Tamar Tak1, Kiki Tesselaar2 Janesh Pillay1,3, José A.M. Borghans2, Leo Koenderman1

1. Dept. of Respiratory Medicine, University Medical Center Utrecht,

the Netherlands
2. Dept. of Immunology, University Medical Center Utrecht,
the Netherlands
3. Dept. of Anesthesiology, University Medical Center Utrecht,
the Netherlands

J Leukoc Biol. 2013 Oct;94(4):595-601. doi: 10.1189/jlb.1112571.

Epub 2013 Apr 26.
Chapter 2

Abstract
Neutrophils are the most abundant white blood cells and are indispensable for host defense.
Recently, they have also been implicated in immune regulation and suppression. The latter
functions seem hard to reconcile with the widely held view that neutrophils are very short-
lived, with a circulatory half-life of less than 7 hours. To reopen the discussion on the average
neutrophil half-life, we review and discuss experiments performed in the 1950’s, 60’s and
70’s, as well as recent in vivo labeling experiments. We reappraise the current knowledge on
neutrophil half-lives, including their production in the bone marrow, their residency in the
circulation and marginated pool and their exit from the circulation.

32
 Neutrophil half-lives revisited

Introduction
Neutrophils are the most abundant immune cells in human blood under homeostatic
conditions (1). They comprise the first line of cellular defense against invading micro-
organisms and are indispensable for host defense (2). On the other hand, they can be
detrimental in acute and chronic inflammatory diseases as hyperactivation of these cells can
cause tissue damage. Recently, the concept is emerging that neutrophils have functions
2
beyond defense against pathogens. For example, they were shown to influence B-cell
activation and to inhibit T-cell activation and proliferation (3-5)).
These novel functions have received little attention, possibly due to the current
paradigm that neutrophils have a circulatory half-life of only 7 hours (6). In the search of the
typical neutrophil circulatory half-life, different methods have led to different conclusions. In
this review, we revisit the experiments that have been performed to estimate neutrophil half-
lives, discuss their strengths and limitations, and suggest improvements for further research.

The neutrophil life cycle

Many aspects of the neutrophil life cycle, from production in the bone marrow to apoptosis,
have been reviewed elsewhere (6-9). Here, we will discuss only those aspects that are
essential to understand the methods and results of experiments aimed at determining the
circulatory half-life of neutrophils. All experiments discussed throughout this review have
been performed in healthy humans, unless stated otherwise.

Neutrophil production
Neutrophils originate from neutrophil progenitors in the bone marrow (figure 1). The first
granulocyte-committed progenitor is the myeloblast, which can differentiate into three
different granulocyte lineages. Compared to other progenitors, few of these cells are found
in the bone marrow (10). Myeloblasts can further differentiate/mature into promyelocytes,
which are committed to the neutrophil lineage and can no longer produce basophils or
eosinophils. Subsequently, promyelocytes can mature into myelocytes. The cells in these
three stages of development are able to proliferate and hence are known as the mitotic pool
(7). Their division is believed to be symmetric, resulting in two cells with similar
characteristics, which may proliferate further, or go one step ahead in maturation (11).
When myelocytes become metamyelocytes, they can no longer divide and become
part of the post-mitotic pool (PMP). After the metamyelocyte stage, the cells mature into
banded and, finally, fully mature segmented neutrophils, which may enter the peripheral
blood (7). The time from the last myelocytic division to release into the blood, i.e. the PMP
transit time, is 6 to 7 days in healthy individuals, but can be shortened to 48 hours in diseased
conditions (12-14). The release of neutrophils from the bone marrow is regulated by the
chemokine receptors C-X-C motif receptor 2 (CXCR2) and CXCR4 and their ligands (CXCL1 and
2 for CXCR2 and CXCL12 for CXCR4) (9). CXCL1 and 2 are expressed in high levels in the
vasculature of the bone marrow, directing neutrophils towards the circulation. CXCL12 on the
other hand, is produced by endosteal osteoblasts and retains neutrophils in the bone marrow.
Under inflammatory conditions, cytokines such as G-CSF change the balance between these

33
Chapter 2

chemokines and induce neutrophil mobilization into the bloodstream (as extensively
reviewed previously (9)).

Figure 1. Production of the

neutrophil lineage in the bone
marrow. Myeloblasts,
promyelocytes and myelocytes are
capable of proliferation and, thus,
comprise the mitotic pool. The
mitotic pool is followed by the
subsequent maturation stages of
metamyelocytes, banded
neutrophils and segmented
neutrophils. These cells together are
called the post-mitotic pool (PMP).
After being fully matured,
segmented neutrophils can leave the
bone marrow and enter the
circulation. Numbers in brackets
represent the total amount of each
cell type per kilogram bodyweight,
as determined by Donohue et al (15).

The marginated pool

After entering the circulation, neutrophils are thought to occur in two populations: (1) a free
flowing pool present in blood and (2) a marginated pool of cells that are not retrieved by
blood sampling (6). It is largely unknown how the distribution of neutrophils between these
pools is regulated and whether these pools of cells are functionally different. The marginated
pool was identified in studies where healthy volunteers received infusions with ex vivo radio-
labeled autologous neutrophils. In these experiments, over 50% of the labeled neutrophils
disappeared from the circulation immediately after infusion (16). When adrenalin was
administered during the infusion of cells, the overall number of neutrophils in the
bloodstream increased and 64% of the labeled neutrophils could be retrieved (17). When
subjects performed exercise several hours after infusion of labeled cells, the percentage of
retrieved neutrophils even reached 87%. Since exercise induced cells to go from the
marginated to the circulating pool, but did not change the percentage of labeled cells in the
blood stream, it was concluded that the marginated pool is in equilibrium with the free
flowing pool of neutrophils (17).
The exact nature of this quickly mobilizable pool is unknown, but it is hypothesized
that it consists of cells that are too rigid to easily flow through narrow capillaries (18). The
most likely places for this to occur are in the liver, spleen and bone marrow, since radiolabeled

34
 Neutrophil half-lives revisited

neutrophils imaged with a gamma-camera (scintigraphy) were shown to quickly accumulate

in these organs (19). In addition, neutrophils were shown to take 2-10 minutes to transit
through these organs, much longer than expected if they were to flow freely through the
blood vessels (20, 21). The lungs have also been implicated as a site for neutrophil
margination (20), but recent studies have questioned this notion by showing that the transit
time through the lungs was dependent on the method of isolation of the neutrophils and their 2
activation state (19, 22). Neutrophils isolated by the method that induced the least activation
showed a mere 3.5 seconds longer pulmonary transit time than red blood cells (19, 22).
Besides influencing the site of margination, ex vivo manipulation of neutrophils was shown to
influence the size of the marginated pool, with an increase of margination proportional to the
duration of the labeling procedure (16). This indicates that there is a lot more uncertainty
about the size and characteristics of the marginated pool than is generally thought. Future
experiments should reveal the dynamic characteristics and size of the marginated pool under
physiological circumstances.

Neutrophil clearance
Under homeostasis, the constant production of new neutrophils in the bone marrow needs
to be counterbalanced by the clearance of the same amount of cells from the system. A
reduced clearance and concurrent neutrophilia occurs in many chronic inflammatory
diseases, and is involved in the pathogenesis of these diseases (23). The organs implicated in
neutrophil clearance are the same as those thought to house the marginated pool: the spleen,
liver and bone marrow (19, 24). Twenty-four and 48 hours after infusion of radio-labeled
neutrophils, these organs contained the highest concentrations of label (19, 24). Since very
few labeled neutrophils were found in the circulation at these time points, these findings were
interpreted to reflect the destruction of neutrophils in the spleen, liver and bone marrow (6).
However, these experiments did not give any direct evidence for neutrophil destruction in
these organs. Since the very same organs are thought to contain marginated neutrophils, it is
difficult to conclude whether the accumulation of label in these organs is due to neutrophil
destruction or migration.
Another possibility for neutrophil clearance is that neutrophils extravasate to
peripheral tissues, where they go into apoptosis. Migration of radiolabeled neutrophils to
peripheral tissues has indeed been shown to occur under non-homeostatic, inflammatory
conditions (25). However, during homeostasis most tissues do not contain neutrophils,
because whole-body imaging did not show any evidence for accumulation of ex vivo
radiolabeled neutrophils in peripheral tissues (26). Interestingly, in both mice and zebra fish,
neutrophils have been shown to migrate from peripheral tissues back into the bloodstream
by a process called reverse transmigration, demonstrating that extravasation does not
necessarily lead to neutrophil clearance in the tissue (27, 28).

In vitro experiments
In principal, the easiest approach to determine the average neutrophil half-life is by isolating
neutrophils from the blood of healthy volunteers and measuring their survival in vitro.
However, in this approach normal tissue environment is lacking, which most likely affects the

35
Chapter 2

half-life of the cells. For example, ex vivo manipulation of neutrophils is known to cause
priming, which has been shown to increase survival in vitro (23, 24, 29). On the other hand,
neutrophils in culture spontaneously produce reactive oxygen species, which is known to
induce apoptosis (30, 31). Based on these findings, in vitro experiments do not seem to reflect
the in vivo situation and, therefore, will not be discussed further in this review.

Labeling experiments
The most commonly used method of determining the neutrophil half-life in vivo is by pulse-
chase labeling experiments. Several different labeling procedures and different labels have
been used, each with their own benefits and limitations (table 1).

Ex vivo labeling
The first experimental approach to determine the half-life of neutrophils in humans that has
been performed was based on ex vivo labeling. In these experiments, blood was drawn from
a healthy volunteer, labeled in vitro with a radioactive tracer and reinfused into the volunteer.
The first tracer to be used for this purpose was diisopropyl fluorophosphate with a radioactive
phosphorous (DF32P) or hydrogen isotope (3H-DFP) (16). DFP is a cell permeable molecule that
binds to the hydroxyl group of serine proteases. Since neutrophils contain large amounts of
serine proteases (e.g. neutrophil elastase), they take up large amounts of label (32).
Subsequently, the amount of label was determined by measuring the radioactivity per
milligram of isolated neutrophils or by autoradiography on blood smears. Alternatively,
neutrophils were labeled with metastable technetium-99-hexametazime (99mTc-HMPAO),
indium-111 (111In) tropolonate, or chromium-51 (51Cr), which bind to cellular proteins (33), or
with tritiated thymidine (3HTdr), which is incorporated into the DNA of dividing cells during
the S-phase (12). Since neutrophils do not divide, 3HTdr was injected into a healthy volunteer
and after 7 days, when the percentage of labeled neutrophils in the blood was at its
maximum, blood was transfused into a second healthy volunteer. Over time, the radioactivity
was measured in the blood of the second volunteer.
Where 51Cr labeled cells revealed a neutrophil half-life of 17 hours, a number of ex vivo
DFP and the 3HTdr labeling studies all came to similar half-lives of 5-8 hours in peripheral
blood (table 1) (8, 16, 17, 32, 34). Despite the relatively consistent picture that emerges from
these studies, ex vivo labeling has a major limitation since it causes neutrophil activation,
which has been shown to alter neutrophil homing characteristics (19, 22, 29, 35) and thereby
affects the neutrophil circulatory half-life. Indeed, ex vivo manipulation of neutrophils has
been shown to have dramatic effects on their circulatory half-life in mice; in vivo labeled
neutrophils had a half-life of 8-10 hours, whereas ex vivo labeled neutrophils disappeared
from the blood with a half-life of only 1.5 hours (24, 36). Even though the in vivo labeled
neutrophil circulatory half-life in mice lies within the range of half-lives reported in human ex
vivo labeling studies, it cannot be taken as support for its correctness, as leukocyte half-lives
can differ tremendously between man and mouse (37).
In summary, since ex vivo labeling studies were performed with cells that likely had
altered homing characteristics, the results from these studies may not reflect the half-life of
normal, untouched neutrophils under homeostatic circumstances.

36
 Neutrophil half-lives revisited

In vivo DF32P labeling

Besides labeling neutrophils ex vivo with DF32P, Cartwright et al injected this tracer directly
into the bloodstream of healthy subjects (32). Since neutrophil elastase is present in all
maturation stages from myelocyte to mature neutrophil, not only circulating neutrophils, but
also bone marrow neutrophils and neutrophil progenitor cells were labeled. Immediately
after injection of DFP, labeled neutrophils disappeared from the circulation with a half-life of 2
7 hours. Nevertheless, labeled neutrophils were found up to 20 days later, which was
interpreted as release of newly produced neutrophils from the bone marrow.
DFP is a structural homologue of the nerve gas sarin and is extremely toxic. Dresch et
al (34) demonstrated that this toxicity affects neutrophil half-lives. The injection of DF32P
several hours after reinfusion of 51Cr labeled neutrophils into a healthy volunteer, resulted in
an immediate loss of one third of the leukocytes and of ~85% of 51Cr labeled cells and hence
in an underestimation of the neutrophil half-life. The results of this study show that DF32P is
not suitable for studying neutrophil half-life or margination.

In vivo 3HTdr labeling

3HTdr has also been used for in vivo labeling experiments. After injection, it is incorporated
into the DNA of neutrophil progenitors (12). After going through all subsequent maturation
stages, these progenitor cells give rise to fully mature, labeled neutrophils that start entering
the bloodstream. Several studies showed that the first 3HTdr labeled neutrophils enter the
circulation approximately 5 days after labeling (8, 12, 13, 38). Two to three days later, the
amount of label in the blood reaches its maximum and subsequently decreases in a similar
timespan. Ten to twelve days after injection of label, a second peak of labeled neutrophils
was seen in some volunteers, which was interpreted as a wave of new labeled cells appearing
from the bone marrow (13). This illustrates a drawback of determining the neutrophil half-
life by this method. In studies in which neutrophil progenitors are labeled, it is impossible to
distinguish between labeled cells surviving in the circulation and labeled cells that have
recently left the bone marrow. Steinbach et al. have calculated the half-life of in vivo 3HTdr
labeled neutrophils in terminal cancer patients and found it to be 17.3 hours (38). However,
the effect of cancer on neutrophil kinetics is unknown. In addition, 3HTdr can be toxic for bone
marrow cells, demonstrating the advantage of a truly non-toxic label (39).

In vivo deuterium labeling

Labeling with the stable hydrogen isotope deuterium (2H) is a relatively new technique that
does label cells without cytotoxic effects (14). 2H is administered to volunteers in the form of
heavy water (2H2O) or deuterated glucose (2H-glucose), which are used to produce
nucleotides via the de novo synthesis pathway and thus incorporated into the DNA of dividing
cells (40, 41). Incorporation of the isotope into the DNA of a certain cell population can be
quantified by a combination of gas chromatography and mass spectrometry (14). We have
used data from our long-term (63 days) 2H2O labeling study to estimate the average
neutrophil half-life in healthy volunteers (42, 43). The presence of 2H was measured in
neutrophil DNA and urine during and after administration of label. A mathematical model was
fitted to the data in order to estimate the average neutrophil half-life. The relatively slow gain

37
Chapter 2

and loss of label in blood neutrophils were interpreted to reflect a neutrophil circulatory half-
life of as much as 90 hours, far longer than described previously (43). However, as pointed
out by Li et al., the data could equally well be explained by a model in which BM neutrophils
have a half-life of 90 hours, while neutrophils have a much shorter circulatory half-life (44).
Future non-toxic in vivo labeling studies on bone marrow neutrophils should resolve this
issue.

Table 1. Overview of results from labeling experiments

Duration % cells in
Reported
Reference Tracer In vivo or ex vivo labelling of marginated
Half-life
labeling pool
Cartwright Ex vivo labeling of whole
DF32P 1h 6.7h 56%
(32) blood
Ex vivo labeling of whole
Mauer (16) DF32P 1h 5.3-9.0h 43-67%
blood
Ex vivo labeling of whole
Athens (17) DF32P 1h 6.6h 51%
blood
In vivo labeled blood Single
H-
3
Dancey (8) transfused into donor at dose 7.6h 41.5%
Thymidine
day 7 after labeling (<30min)
Single
Cartwright
DF32P In vivo dose 7.0h N.D.
(32)
(<10min)
Ex vivo labeling of isolated
Dresch (34) 51
Cr 45min 17.5h 48%
leukocytes
Steinbach 3
H- In vivo (terminal cancer
3-10 days 17.3h N.D
(38) Thymidine patients)
111
In-
6.3h 68%
tropolanate Ex vivo labeling of isolated
Peters (26) 5min
99m
Tc- leukocytes
4.2h 67%
HMPAO
Al-Janabi 99m
Tc-
Ex vivo 5min 4.3h 65%
(45) HMPAO
Pillay (43) 2
H2O In vivo 63 days 90h N.D

Neutrophil production in the bone marrow

An alternative way to estimate the neutrophil circulatory half-life is to reveal it from
neutrophil production numbers in the bone marrow compared to the amount of circulating
neutrophils. The number of neutrophils exported from the BM per day can be derived by
dividing the number of cells in the PMP by the average PMP transit time. The number of cells
in the PMP was estimated to be 8.1 * 109 cells per kilogram bodyweight, by extrapolation
from 59Fe labeling studies that were used to quantify the total number of cells in the erythroid
lineage (15, 46). The 59Fe study revealed that the total number of erythrocytes and
reticulocytes in the body 3.2 * 1011 per kilogram bodyweight, of which approximately 3%
resides in the bone marrow. Combining the known prevalence of each cell type in the bone
marrow with this number led to the conclusion that there are approximately 18 * 109
granulocytic bone marrow cells per kilogram body weight, of which 45% is in the PMP (15,

38
 Neutrophil half-lives revisited

46). However, It has been shown that neutrophils can migrate back to the bone marrow (47)
and one study showed that up to 50% of PMP cells undergo apoptosis before reaching the
bloodstream (48), demonstrating that the effective size of the PMP used in these equations
is probably an overestimation. In addition, the 59Fe labelling studies have never been
confirmed with different methods and thus, the accuracy of the estimated bone marrow cell
numbers remains unknown. 2
The second part of the equation, the transit time through the PMP, was estimated to
be 5 days, since in 3HTdr experiments the first labeled cells appear in the blood at this
timepoint (12, 13, 15). Although these experiments accurately measured the appearance of
the first labeled cells, they did not determine the average PMP transit time. Moreover,
labeling experiments utilizing a non-toxic label revealed the first labeled cells to appear after
6 rather than 5 days (14).
The amount of free flowing neutrophils in the bloodstream of an average healthy
human has been estimated to be 8-30 * 109 (32), to which the number of neutrophils in the
marginated pool should be added. As stated above, the consensus is that during homeostasis
the marginated pool comprises ~50% of all circulating neutrophils, although this number
should be used with caution, because it is based on ex vivo labeling experiments.
Combining the number of neutrophils in the blood, the PMP size and transit time, the
neutrophil lifespan in the blood was estimated to be 6.2 hours, which corresponds to a 4.3
hour circulatory half-life. Considering the probable overestimation of the effective PMP size
and an underestimation of the PMP transit time, this estimation is probably an
underestimation of the real neutrophil circulatory half-life.
A second method to estimate neutrophil production numbers in the bone marrow is
by determining the amount of divisions by neutrophil progenitors. In principle, one should be
able to deduce it from the total number of myelocytes and their doubling time. As described
above, early experiments estimated the total cellularity of the bone marrow to be 18 * 109
bone marrow cells per kilogram body weight (15, 46). Myelocytes are thought to comprise
24% of all bone marrow cells, yielding an estimated total number of 4.3 * 109 myelocytes per
kilogram bodyweight (7, 12, 15). In vitro experiments have revealed a myelocyte doubling
time of 60 hours. Since myelocytes in cell division were found to complete the full cell cycle
in 30 hours, the authors concluded that two thirds of the myelocytes were not actively
proliferating. These cells were dubbed the “lazy pool” and estimated to divide only once every
69 days (7). It remains unknown how representative these numbers are for the in vivo
situation. Alternatively, in vivo myelocyte doubling times can be derived from the fraction of
myelocytes undergoing cell division, and the time it takes for each cell to complete a cell-
division. Unfortunately, data are conflicting: flow cytometric evaluation of cellular DNA
contents has shown that in homeostasis, 6.3% of myelocytes are in the S, G2 or M-phase (49),
whilst another study revealed that less than 0.005% of myelocytes are in the M-phase (50)
and visual inspection of bone marrow smears estimated 0.79 – 5.2% of myelocytes to be in
the M-phase (7). As a consequence, neutrophil half-lives calculated from the different papers
yield results that differ up to 1000-fold.
Taken together, both the interpretation of in vivo deuterium labeling studies as well
as the above-mentioned alternative routes to quantify neutrophil circulatory half-lives
require better quantitative insights into neutrophil production rates in the BM. Future

39
Chapter 2

research should aim to quantify (1) how many neutrophils reside in the marginated pool (2)
myelocyte doubling times in vivo and (3) the effective size of the PMP.

The neutrophil after-life – after leaving the bloodstream

So far, we have only discussed the circulatory half-life of neutrophils, but – as already
mentioned – neutrophils do not exclusively reside in the bloodstream. There is a pool of
neutrophils that rapidly enters the bloodstream upon immune activation (e.g. in sepsis and
trauma) or during exercise (3, 17). Although some of these cells are probably coming from the
marginated pool, (6, 17), others are considered to come from outside the blood vessels. In
labeling studies, injection of LPS or corticosteroids not only increased the number of labeled
cells in the blood, but also decreased the average amount of label per cell. (17). This suggests
that not only marginated cells, but also other neutrophils are recruited to the bloodstream,
most probably involving immature cells from the bone marrow. Indeed, these cells were
shown to have a banded nuclear morphology and low CD16 expression (3). In addition to
these recent bone marrow emigrants, cells with a hypersegmented nucleus (> 4 lobes) were
found to enter the bloodstream (3). Considering that neutrophils increase their amount of
lobes during maturation, it is tempting to suggest that hypersegmented cells are older than
mature neutrophils.
Experiments by Vincent et al (51) have indeed shown that a pool of old cells exists in
the body and can be recruited to the bloodstream. In these experiments, one calve was
labeled with 3HTdr after which its bloodstream was surgically coupled to a genetically
identical sibling. After 2 hours, the connection of the bloodstreams was removed and the
number of labeled cells in the acceptor calve was followed. Forty-eight hours later, very few
labeled cells were detectable in peripheral blood. Upon receiving a dose of corticosteroids,
labeled cells were again detected in the bloodstream, at levels as high as 25% of the maximum
value, suggesting that at least a subpopulation of neutrophils survives for several days in the
body. This notion is strengthened by observations in mice, which have demonstrated that
neutrophils can survive up to 7 days in the tissues (52). Taken together, these data show that
neutrophils do not go into apoptosis immediately after leaving the bloodstream, but that a
marked pool of old neutrophils resides somewhere in the body. Whether these cells can
influence the measurement of neutrophil half-lives by labeling experiments depends on
whether these cells return to the circulation, as they do after injection of steroids (51).

Neutrophil half-lives under non-homeostatic conditions

Neutrophils are well known to migrate to sites of inflammation under the guidance of factors
such as C5a, CXCL8 (IL-8), platelet activating factor (PAF) and leukotriene B4 (LTB4) (reviewed
in (53)). Also, during acute inflammation, the total number of neutrophils in the bloodstream
increases dramatically (17). Based on these findings it is likely that inflammation influences
the neutrophil half-life. Ex vivo labeled neutrophils showed an increased circulatory half-life
compared to healthy volunteers in patients suffering from chronic myeloid leukemia (CML),
sarcoidosis or liver cirrhosis, in splenectomized individuals, and when healthy volunteers
received prednisolone (table 2) (54, 55). No difference was seen in patients suffering from
rheumatoid arthritis, active inflammation or in healthy volunteers receiving G-CSF (26, 45,

40
 Neutrophil half-lives revisited

56). One study suggested that G-CSF and GM-CSF administration lead to increased neutrophil
counts in peripheral blood by increasing the production rate, rather than the circulatory half-
life of neutrophils (57), while another study found evidence for an increased neutrophil half-
life after GM-CSF administration. Also in other diseases or non-homeostatic conditions, data
differ. Within one study, labeled neutrophils of different patients with active infections had
either longer or shorter half-lives than healthy controls (54). Similarly, the half-lives of 2
neutrophils from neutropenic patients differed, depending on whether neutropenia was drug
induced or not and whether the transfusion was autologous or from a healthy donor (table 2)
(54, 55).
Taken together, these data demonstrate that inflammation can influence neutrophil
half-lives, and that the influence varies between or even within diseases. Importantly, before
disturbances induced by disease can be understood, it is essential to fully understand the
situation under homeostatic circumstances.

Table 2. Neutrophil half-life under non-homeostatic conditions

Disease / Half-life Half-life
Reference pharmaceutic Labeling method (non-homeostatic) (control)
Transfusion of 3H-DFP labeled, isolated
Dale (55) Neutropenia neutrophils from volunteers receiving 20.3h
G-CSF and dexamethasone 9.6h
Transfusion of autologous 3H-DFP
Dale (2) (58) GM-CSF 13.1h
labeled blood
Prednisone 10.3h
Transfusion of autologous, ex vivo
Athens (17) Exercise 6.7h ns 6.8h
DF32P labeled blood
Adrenalin 5.8h ns
Transfusion of autologous, ex vivo 3H-
Price (56) G-CSF 13.5-15.9h ns 10.4h
DFP labeled blood
Ex vivo 51Cr labeling of isolated
Dresch (34) DFP injection 9.4h 17.5h
neutrophils
Rheumatoid Ex vivo 99mTc-HMPAO labeling of
Al-Janabi (45) 4.2h 4.3h
arthritis isolated white blood cells

Suspected Ex vivo 111In–tropolonate 5.8h 6.3h

Peters (26) inflammatory
disease* Ex vivo 99m-HMPAO 4.5h 4.2h
CML 82-140h
Sarcoidosis 17h
Liver Cirrhosis 15h
Drug induced
Transfusion of autologous, ex vivo 3H- 8.4, 13.2h
Meuret (54) neutropenia 7.6h
DFP labeled blood
Other
2.3-7h
Neutropenia
Infection 4-12h
Splenectomy 22.2h
ns = not significantly different from control
* Patients that are not diagnosed with inflammation are used for control half-life

41
Chapter 2

Concluding Remarks
Although the importance of the quantification of neutrophil half-lives and kinetics is evident,
there is still a lot of uncertainty about neutrophil turnover in BM and blood. Both old and new
labeling experiments suffer from technical limitations such as the toxicity of the tracers that
were used, ex vivo manipulation affecting neutrophil kinetics and migration, and the inability
to distinguish between labeled cells residing in the blood and those that are released by the
bone marrow. Alternative approaches that are based on the quantification of neutrophil
production in the bone marrow are hampered by the lack of consistent data. Therefore, we
strongly encourage initiatives that will address these issues with new experiments in which
neutrophils are not affected by label toxicity or ex vivo manipulation. Such studies should
reveal conclusive data on neutrophil production rates in the bone marrow and half-lives in
the circulation.

42
 Neutrophil half-lives revisited

References
1. Hsieh MM, Everhart JE, Byrd-Holt DD, Tisdale JF, and Rodgers GP. Prevalence of
neutropenia in the U.S. population: age, sex, smoking status, and ethnic differences.
Annals of internal medicine. 2007;146(7):486-92.
2. Lekstrom-Himes JA, and Gallin JI. Immunodeficiency diseases caused by defects in
phagocytes. The New England journal of medicine. 2000;343(23):1703-14. 2
3. Pillay J, Kamp VM, van Hoffen E, Visser T, Tak T, Lammers JW, Ulfman LH, Leenen LP,
Pickkers P, and Koenderman L. A subset of neutrophils in human systemic
inflammation inhibits T cell responses through Mac-1. The Journal of clinical
investigation. 2012;122(1):327-36.
4. Puga I, Cols M, Barra CM, He B, Cassis L, Gentile M, Comerma L, Chorny A, Shan M, Xu
W, et al. B cell-helper neutrophils stimulate the diversification and production of
immunoglobulin in the marginal zone of the spleen. Nature immunology.
2012;13(2):170-80.
5. Gabrilovich DI, and Nagaraj S. Myeloid-derived suppressor cells as regulators of the
immune system. Nature reviews Immunology. 2009;9(3):162-74.
6. Summers C, Rankin SM, Condliffe AM, Singh N, Peters AM, and Chilvers ER. Neutrophil
kinetics in health and disease. Trends in immunology. 2010;31(8):318-24.
7. Boll IT, and Fuchs G. A kinetic model of granulocytopoiesis. Experimental cell research.
1970;61(1):147-52.
8. Dancey JT, Deubelbeiss KA, Harker LA, and Finch CA. Neutrophil kinetics in man. The
Journal of clinical investigation. 1976;58(3):705-15.
9. Day RB, and Link DC. Regulation of neutrophil trafficking from the bone marrow.
Cellular and molecular life sciences : CMLS. 2012;69(9):1415-23.
10. Mary JY. Normal human granulopoiesis revisited. II. Bone marrow data. Biomedicine
& pharmacotherapy = Biomedecine & pharmacotherapie. 1985;39(2):66-77.
11. Mitsiadis TA, Barrandon O, Rochat A, Barrandon Y, and De Bari C. Stem cell niches in
mammals. Experimental cell research. 2007;313(16):3377-85.
12. Cronkite EP, Fliedner TM, Bond VP, and Rubini JR. Dynamics of hemopoietic
proliferation in man and mice studied by H3-thymidine incorporation into DNA. Annals
of the New York Academy of Sciences. 1959;77(803-20.
13. Fliedner TM, Cronkite EP, Killmann SA, and Bond VP. Granulocytopoiesis. Ii.
Emergence and Pattern of Labeling of Neutrophilic Granulocytes in Humans. Blood.
1964;24(683-700.
14. Macallan DC, Fullerton CA, Neese RA, Haddock K, Park SS, and Hellerstein MK.
Measurement of cell proliferation by labeling of DNA with stable isotope-labeled
glucose: studies in vitro, in animals, and in humans. Proc Natl Acad Sci U S A.
1998;95(2):708-13.
15. Donohue DM, Reiff RH, Hanson ML, Betson Y, and Finch CA. Quantitative
measurement of the erythrocytic and granulocytic cells of the marrow and blood. The
Journal of clinical investigation. 1958;37(11):1571-6.
16. Mauer AM, Athens JW, Ashenbrucker H, Cartwright GE, and Wintrobe MM.
Leukokinetic Studies. II. A Method for Labeling Granulocytes in Vitro with Radioactive
Diisopropylfluorophosphate (Dfp). The Journal of clinical investigation.
1960;39(9):1481-6.
17. Athens JW, Haab OP, Raab SO, Mauer AM, Ashenbrucker H, Cartwright GE, and
Wintrobe MM. Leukokinetic studies. IV. The total blood, circulating and marginal

43
Chapter 2

granulocyte pools and the granulocyte turnover rate in normal subjects. The Journal
of clinical investigation. 1961;40(989-95.
18. Wiggs BR, English D, Quinlan WM, Doyle NA, Hogg JC, and Doerschuk CM.
Contributions of capillary pathway size and neutrophil deformability to neutrophil
transit through rabbit lungs. J Appl Physiol. 1994;77(1):463-70.
19. Saverymuttu SH, Peters AM, Keshavarzian A, Reavy HJ, and Lavender JP. The kinetics
of 111indium distribution following injection of 111indium labelled autologous
granulocytes in man. British journal of haematology. 1985;61(4):675-85.
20. Peters AM. Just how big is the pulmonary granulocyte pool? Clin Sci (Lond).
1998;94(1):7-19.
21. Ussov WY, Aktolun C, Myers MJ, Jamar F, and Peters AM. Granulocyte margination in
bone marrow: comparison with margination in the spleen and liver. Scandinavian
journal of clinical and laboratory investigation. 1995;55(1):87-96.
22. Summers C, White JF, Mackenzie IM, Solanki C, Peters AM, and Chilvers ER.
Measurement of neutrophil pulmonary transit time in humans [abstract]. Am J Respir
Crit Care Med. 2009;179(A1333.
23. Simon HU. Regulation of eosinophil and neutrophil apoptosis--similarities and
differences. Immunological reviews. 2001;179(156-62.
24. Suratt BT, Young SK, Lieber J, Nick JA, Henson PM, and Worthen GS. Neutrophil
maturation and activation determine anatomic site of clearance from circulation.
American journal of physiology Lung cellular and molecular physiology.
2001;281(4):L913-21.
25. Saverymuttu SH, Peters AM, Reavy HJ, and Lavender JP. Measurement of granulocyte
migration and accumulation in inflammation in man. Clinical and experimental
immunology. 1983;52(3):607-12.
26. Peters AM, Roddie ME, Danpure HJ, Osman S, Zacharopoulos GP, George P, Stuttle
AW, and Lavender JP. 99Tcm-HMPAO labelled leucocytes: comparison with 111In-
tropolonate labelled granulocytes. Nuclear medicine communications. 1988;9(6):449-
63.
27. Buckley CD, Ross EA, McGettrick HM, Osborne CE, Haworth O, Schmutz C, Stone PC,
Salmon M, Matharu NM, Vohra RK, et al. Identification of a phenotypically and
functionally distinct population of long-lived neutrophils in a model of reverse
endothelial migration. Journal of leukocyte biology. 2006;79(2):303-11.
28. Yoo SK, and Huttenlocher A. Spatiotemporal photolabeling of neutrophil trafficking
during inflammation in live zebrafish. Journal of leukocyte biology. 2011;89(5):661-7.
29. Watson F, Robinson JJ, and Edwards SW. Neutrophil function in whole blood and after
purification: changes in receptor expression, oxidase activity and responsiveness to
cytokines. Bioscience reports. 1992;12(2):123-33.
30. Scheel-Toellner D, Wang K, Craddock R, Webb PR, McGettrick HM, Assi LK, Parkes N,
Clough LE, Gulbins E, Salmon M, et al. Reactive oxygen species limit neutrophil life
span by activating death receptor signaling. Blood. 2004;104(8):2557-64.
31. Aoshiba K, Nakajima Y, Yasui S, Tamaoki J, and Nagai A. Red blood cells inhibit
apoptosis of human neutrophils. Blood. 1999;93(11):4006-10.
32. Cartwright GE, Athens JW, and Wintrobe MM. The Kinetics of Granulopoiesis in
Normal Man. Blood. 1964;24(780-803.
33. Ronai PM. The elution of 51Cr from labelled leukocytes--a new theory. Blood.
1969;33(3):408-13.

44
 Neutrophil half-lives revisited

34. Dresch C, Najean Y, and Beauchet J. In vitro 51Cr and 32P-DFP labeling of granulocytes
in man. Journal of nuclear medicine : official publication, Society of Nuclear Medicine.
1971;12(12):774-84.
35. Bertrand-Caix J, Freyburger G, Bordenave L, Labrouche S, Bareille R, Baquey C,
Ducassou D, and Lefebvre F. Functional upregulation of granulocytes labeled with
technetium-99m-HMPAO and indium-111-oxinate. Journal of nuclear medicine : 2
official publication, Society of Nuclear Medicine. 1996;37(5):863-8.
36. Basu S, Hodgson G, Katz M, and Dunn AR. Evaluation of role of G-CSF in the production,
survival, and release of neutrophils from bone marrow into circulation. Blood.
2002;100(3):854-61.
37. den Braber I, Mugwagwa T, Vrisekoop N, Westera L, Mogling R, de Boer AB, Willems
N, Schrijver EH, Spierenburg G, Gaiser K, et al. Maintenance of peripheral naive T cells
is sustained by thymus output in mice but not humans. Immunity. 2012;36(2):288-97.
38. Steinbach KH, Schick P, Trepel F, Raffler H, Dohrmann J, Heilgeist G, Heltzel W, Li K,
Past W, van der Woerd-de Lange JA, et al. Estimation of kinetic parameters of
neutrophilic, eosinophilic, and basophilic granulocytes in human blood. Blut.
1979;39(1):27-38.
39. Smith WW, Brecher G, Stohlman F, Jr., and Cornfield J. Toxicity of tritiated thymidine
to bone marrow transplants. Radiation research. 1962;16(201-7.
40. Busch R, Neese RA, Awada M, Hayes GM, and Hellerstein MK. Measurement of cell
proliferation by heavy water labeling. Nature protocols. 2007;2(12):3045-57.
41. Macallan DC, Asquith B, Zhang Y, de Lara C, Ghattas H, Defoiche J, and Beverley PC.
Measurement of proliferation and disappearance of rapid turnover cell populations in
human studies using deuterium-labeled glucose. Nature protocols. 2009;4(9):1313-27.
42. Vrisekoop N, den Braber I, de Boer AB, Ruiter AF, Ackermans MT, van der Crabben SN,
Schrijver EH, Spierenburg G, Sauerwein HP, Hazenberg MD, et al. Sparse production
but preferential incorporation of recently produced naive T cells in the human
peripheral pool. Proc Natl Acad Sci U S A. 2008;105(16):6115-20.
43. Pillay J, den Braber I, Vrisekoop N, Kwast LM, de Boer RJ, Borghans JA, Tesselaar K, and
Koenderman L. In vivo labeling with 2H2O reveals a human neutrophil lifespan of 5.4
days. Blood. 2010;116(4):625-7.
44. Li KW, Turner SM, Emson CL, Hellerstein MK, and Dale DC. Deuterium and neutrophil
kinetics. Blood. 2011;117(22):6052-3; author reply 3-4.
45. al-Janabi MA, Jones AK, Solanki K, Critchley M, and Britton KE. A comparative study of
labelled neutrophil behaviour in patients with rheumatoid arthritis and normal
volunteers. Nuclear medicine communications. 1993;14(10):888-95.
46. Suit HD. A technique for estimating the bone marrow cellularity in vivo using 59Fe.
Journal of clinical pathology. 1957;10(3):267-9.
47. Rankin SM. The bone marrow: a site of neutrophil clearance. Journal of leukocyte
biology. 2010;88(2):241-51.
48. Colijn C, and Mackey MC. A mathematical model of hematopoiesis: II. Cyclical
neutropenia. Journal of theoretical biology. 2005;237(2):133-46.
49. Matarraz S, Fernandez C, Albors M, Teodosio C, Lopez A, Jara-Acevedo M, Cervero C,
Caballero G, Gutierrez O, and Orfao A. Cell-cycle distribution of different cell
compartments in normal versus reactive bone marrow: a frame of reference for the
study of dysplastic hematopoiesis. Cytometry Part B, Clinical cytometry.
2011;80(6):354-61.

45
Chapter 2

50. Behbehani GK, Bendall SC, Clutter MR, Fantl WJ, and Nolan GP. Single-cell mass
cytometry adapted to measurements of the cell cycle. Cytometry Part A : the journal
of the International Society for Analytical Cytology. 2012;81(7):552-66.
51. Vincent PC, Chanana AD, Cronkite EP, and Joel DD. The intravascular survival of
neutrophils labeled in vivo. Blood. 1974;43(3):371-7.
52. Cheretakis C, Leung R, Sun CX, Dror Y, and Glogauer M. Timing of neutrophil tissue
repopulation predicts restoration of innate immune protection in a murine bone
marrow transplantation model. Blood. 2006;108(8):2821-6.
53. Mukaida N, Harada A, and Matsushima K. Interleukin-8 (IL-8) and monocyte
chemotactic and activating factor (MCAF/MCP-1), chemokines essentially involved in
inflammatory and immune reactions. Cytokine & growth factor reviews. 1998;9(1):9-
23.
54. Meuret G, Hoffmann G, Fliedner TM, Rau M, Oehl S, Walz R, and von Klein-Wisenberg
A. Neutrophil kinetics in man. Studies using autotransfusion of 3 H-DFP labeled blood
cells and autoradiography. Blut. 1973;26(2):97-109.
55. Dale DC, Liles WC, Llewellyn C, Rodger E, and Price TH. Neutrophil transfusions:
kinetics and functions of neutrophils mobilized with granulocyte-colony-stimulating
factor and dexamethasone. Transfusion. 1998;38(8):713-21.
56. Price TH, Chatta GS, and Dale DC. Effect of recombinant granulocyte colony-
stimulating factor on neutrophil kinetics in normal young and elderly humans. Blood.
1996;88(1):335-40.
57. Aglietta M, Piacibello W, Sanavio F, Stacchini A, Apra F, Schena M, Mossetti C, Carnino
F, Caligaris-Cappio F, and Gavosto F. Kinetics of human hemopoietic cells after in vivo
administration of granulocyte-macrophage colony-stimulating factor. The Journal of
clinical investigation. 1989;83(2):551-7.
58. Dale DC, Liles WC, Llewellyn C, and Price TH. Effects of granulocyte-macrophage
colony-stimulating factor (GM-CSF) on neutrophil kinetics and function in normal
human volunteers. American journal of hematology. 1998;57(1):7-15.

46
Neutrophil half-lives revisited

47
 Chapter 3
Estimating human neutrophil circulatory
lifespan by short-term deuterated-glucose
labelling in vivo

Tak, T.1 Tesselaar, K.2 Drylewicz, J.2 Borghans, J.A.M.2 Koenderman, L.1

1. Department of Respiratory Medicine Laboratory of Translational Immunology,

University Medical Centre Utrecht, Utrecht, the Netherlands.
2. Department of Immunology, Laboratory of Translational Immunology,
University Medical Centre Utrecht, Utrecht, the Netherlands.

Submitted
Chapter 3

Abstract
Neutrophils are essential for host defence against invading pathogens and shape subsequent
immunity by acquiring different functional phenotypes. However, a basic characteristic, their
lifespan in blood under homeostatic conditions, remains uncertain. Previous attempts to
determine the neutrophil lifespan in human blood were hampered by sampling under non-
homeostatic conditions (e.g. cancer patients), reinfusion of ex vivo manipulated cells, or
toxicity of the label. A recent attempt by our group applying stable non-toxic 2H2O in vivo was
hampered by the relatively slow turnover of 2H2O in body water and by not taking bone
marrow kinetics into account.
Here, we employed short-term (6h) in vivo pulse labelling with 6,6-2H2-glucose in
healthy volunteers to estimate the circulatory lifespan of neutrophils by determining
deuterium enrichment in the DNA of blood neutrophils. Measured deuterium enrichments
were interpreted using both a model based on ordinary differential equations (ODE) and an
agent-based conveyor-belt model, both simulating label kinetics in both blood and bone
marrow. ODE models estimated implausible values for the combination of neutrophil
circulatory lifespan and division time of progenitor cells (t-div) in the BM (3.7h/4.8d and
4.5d/1h for lifespan and t-div, respectively). The agent-based model was not able to produce
exact estimates, but simulations with this model produced the best results with a neutrophil
circulatory lifespan between 2.5 and 4 days. Such relatively long circulatory lifespans allows
neutrophils to engage in relatively long-term processes, such as interactions with the adaptive
immune system.

50
 Neutrophil circulatory lifespan

Introduction
Neutrophils are the most common white blood cells and are indispensable for immune
defense (1). Traditionally, neutrophils are considered to only form a first line of defense
against bacteria and fungi, but more recently several other neutrophil functions have been
described (2, 3). These include tumoricidal activity (4), antigen presentation (5), immune
suppression (6) and trans-differentiation to neutrophil-dendritic hybrid cells (7). Besides
having a beneficial role in immunity against pathogens, neutrophils have been shown to be
detrimental in a wide range of diseases, making them a potential target for novel therapeutics 3
(8-10). However, some very basic neutrophil characteristics, including their average lifespan
in the bloodstream, remain a matter of debate (11, 12). Without knowledge of their lifespan,
it is difficult to fully understand the mechanisms by which these cells engage in immune
homeostasis and to predict the potential effects of therapeutics.
Much work has been done on elucidating the circulatory lifespan of neutrophils (13-
16). Unfortunately, as reviewed elsewhere (17), interpretation of these studies is hampered
by utilization of labelling methods which induce rapid depletion of neutrophils from the blood
(18, 19), by ex vivo manipulation of neutrophils leading to aberrant homing to the lungs when
reinfused in vivo (20) or by the use of toxic labels (15, 21). More recently, our group attempted
to estimate the circulatory lifespan of untouched neutrophils by utilization of non-toxic
deuterium labelling in vivo with heavy water (2H2O) (11). As human body water has a relatively
slow turnover rate, 2H2O labelling is not the ideal method to measure the lifespan of cells with
a relatively fast turnover rate, such as neutrophils. In addition, Li et al correctly pointed out
that the long estimated average lifespan of 5.4 days that we previously reported may in fact
not reflect the lifespan of neutrophils in blood, but may alternatively point at a long division
time of neutrophil progenitors in the bone marrow (BM) (12).
In order to gain a better insight into the in vivo neutrophil lifespan, we performed
short-term (6h) in vivo pulse-labelling in 32 healthy volunteers using 6,62H2–glucose, a source
of deuterium with a much higher turnover rate than water. Blood was drawn from these
individuals at different time points and incorporation of label in neutrophil DNA was
measured using GC/MS analysis. The resulting data were analysed using a model based on
ordinary differential equations (ODE) and an agent-based model that simulates maturation
kinetics in both blood and BM.

Methods
Study population
In total, 32 healthy volunteers (age 19-27, no allergic/immunological diseases, no medication
use, no excessive use of alcohol, no drug use and normal WBC differential counts) were
included in the study after giving written informed consent. The study was performed in
accordance to the Declaration of Helsinki, and was approved by the medical ethical
committee of the University Medical Centre Utrecht (#10-349/O). The labelling protocol was
adapted from Macallan et al, with small modifications (22): Volunteers received 12 oral doses

51
Chapter 3

of 6,6-2H2-glucose (Cambridge Isotopes Tewksbury, MA, USA) at a total of 1g/kg bodyweight

over a period of 6 hours. For twelve volunteers, blood was drawn before the start of the
labelling procedure and at five subsequent days. For the other twenty volunteers only one
time point was available after labelling. Sampling days differed between volunteers to
maximize the information regarding deuterium enrichment. All samples (92 in total) were
taken at the same time of day, to prevent effects caused by diurnal rhythms in cell
populations. One volunteer presented with gastroenteritis at day 10 during the study and was
excluded from the analysis.

Cell isolation
Whole blood white cells were obtained by erythrocyte lysis of heparinized blood using an ice-
cold NH4Cl solution (23). Subsequently, cells were isolated by FACS sorting (Aria IIb and III,
Becton Dickinson, Mountain View CA, USA). Neutrophils were gated based on FSC/SSC, and
as CD16high/CD192neg to exclude eosinophils (supplementary figure S1). Purity was 99.5% as
determined by re-analysis of the sorted samples and by microscopic evaluation of cytospin
preparations.

Measurement of DNA enrichment

DNA enrichment was measured as described in detail previously (22). In short, DNA was
isolated (NucleoSpin Blood, Macherey-Nagel, Düren, Germany), enzymatically hydrolyzed
into deoxyribonecleotides and derivatised into pentafluoro triacetate (PFTA) for analysis on a
GC/MS using selective ion monitoring (SIM). Only the adenosine derivate was analyzed, with
m/z ratios of 435 (M+0, unlabeled) and 437 (M+2, labelled) to limit effects of label re-
utilization, which is far more pronounced in pyrimidines. Equal amounts of PFTA were
analyzed for each sample to avoid abundance artefacts. DNA enrichment was corrected for
the availability of label in the blood as published previously (22). The cell-type specific factor
used to correct the fraction adenosine that is re-utilized instead of de novo produced was
determined in vitro as described in the next section.

In vitro cell assays

Several assays were performed in vitro with a neutrophil progenitor derived cell line (HL-60)
to determine important characteristics of the fate of the label inside the cell (time,
enrichment and de novo adenosine synthesis from 6,6-2H-glucose), To determine the fraction
of adenosine being produced by de novo synthesis instead of by the salvage pathway, cells
were cultured in IMDM medium containing 50% 6,6-2H-glucose and 8% FBS for 4 weeks as
published previously (22). Cells were isolated at several time points during the 4 weeks and
DNA 2H-enrichment was determined. When DNA enrichment reached a plateau, the ratio of
de novo synthesized/salvaged adenosine could be calculated as:
Ratio = (plateau enrichment) / (fraction label in medium).
HL-60 cells were pulse-labelled with 50% 6,6-2H-glucose containing medium for
6 hours to determine the effective pulse-length in neutrophil progenitors. After 6 hours the

52
 Neutrophil circulatory lifespan

cells were washed and cultured in medium with normal glucose (figure 1). The total glucose
concentration was kept constant during the whole experiment. Small samples were taken at
several time points to count the number of cells and to determine their DNA 2H-enrichment.
The time between the first uptake of label into the DNA and the first time point at which label
incorporation was at its maximal level (determined after 24h) was measured and used as the
time period of the pulse in all simulations (see below). The total amount of labelled adenosine
was calculated as DNA 2H enrichment multiplied by the number of cells. A linear regression
was performed on the labelling data from t=2h (when the first label was detected) up to t=6h 3
(when label was removed). The time during which 2H-incorporation takes place was estimated
as the time at which the linear regression line and the maximum amount of labelled
adenosine at t=24h intersect, minus the delay during the uplabelling phase.

Computational and mathematical modelling

Enrichment data were analyzed using mathematical modelling based on ordinary differential
equations (ODE’s) or using an agent-based model. Details on the models can be found in the
supplementary materials and methods. In short, all models assume progenitor cells in the
bone marrow to divide with an average division time t-div, during which label can be
incorporated into the DNA. The maximum level of enrichment of a cell reached after one
division is 50%, since only one DNA strand of each chromosome is newly synthesized during
cell division. Hereafter, one daughter cell remains a progenitor cell that will divide again (after
t-div hours), whereas the other cell will stop dividing and start maturing. After a delay during
which cells mature in the post-mitotic pool (PMP), the so-called PMP transit time (PMPtt),
cells enter the bloodstream. In the circulation, cells have an average lifespan after which they
leave the bloodstream. Cells are assumed not to return to the circulation, since the numbers
of returning cells have been shown to be very low in healthy volunteers (24). The marginated
pool is assumed to be constantly exchanging cells with the circulating pool, leading to similar
enrichments and kinetics in both pools as published previously (14, 25).

Software
FCS express 4 flow was used for analysis of FACS data, GraphPad Prism 6 for plotting and
correcting measurements with a standard curve. ODE models were fitted using Monolix
version 4.3.3s, whereas NetLogo 5.02 was used to develop the computer model with the add-
in “Behaviourspace” for high throughput analysis and R x64 3.1.2 for data transformations
and calculation of the sum of squared residuals (SSR).

Results
Determination of effective pulse length and percentage de novo adenosine synthesis by in
vitro 6,6-2H2-glucose labelling in the myeloid progenitor cell line HL60.
Long term culture of HL-60 cells with 50% 6,62H2–glucose revealed a maximum DNA 2H-
enrichment of 0.65 times the enrichment in the culture medium (figure 1A). This indicated
that 65% of the adenosine moiety in these cells was derived from de novo synthesis.

53
Chapter 3

Short-term (6h) labelling with 50% 6,62H2 –glucose was performed to determine the
intracellular kinetics of de novo adenosine synthesis from 6,62H2–glucose and its
incorporation into the DNA. The total amount of incorporated 2H adenosines was determined
by measuring 2H-DNA enrichment and cell numbers. During the first 1.0 to 1.5 hours of the
experiment little DNA 2H-enrichment was observed (figure 1B). After removal of 6,62H2–
glucose, DNA enrichment kept increasing for approximately 4h, at which point the maximum
amount of label was reached. Taken together, these data indicate that a 6 hour pulse of
6,62H2–glucose label leads to 2H being incorporated into the DNA for approximately 8h; the
effective pulse length used in the mathematical simulations (see below).

Figure 1. (A) Long term culture

of pro-myelocytic cell HL-60
cell line with 50% 6,62H2–
glucose, revealed a maximum
DNA enrichment of 0.65 times
of the enrichment in the
culture medium. Thus, 65% of
the adenosine moieties in
these cells are derived from de
novo synthesis and the
correction factor used in the
correction for plasma glucose
levels should be 0.65. On
average, cell numbers doubled
each 1.04 days. Short-term
labelling with 50% 6,62H2–
glucose was performed to
determine the kinetics of de
novo adenosine synthesis from
6,62H2–glucose and its
incorporation into the DNA.
Cells were grown in medium
containing 50% 6,62H2–glucose
for 6h, after which they were grown in normal medium. The glucose concentration was kept constant during the
experiment. The total amount of incorporated 2H adenosines was determined by measuring 2H-DNA enrichment
and multiplying by cell numbers. After removal of 6,62H2–glucose, DNA enrichment kept increasing for
approximately 4h, at which point a similar enrichment was reached as measured 24h after the start of the
experiment.

In vivo 6,6-2H2-glucose labelling in healthy volunteers.

In total, 32 healthy volunteers received twelve oral doses of 6,62H2–glucose (1 g/kg) over the
course of six hours. Availability of label in plasma was determined prior to and at 3 moments
during intake of labeled glucose, and 2H-enrichment in the DNA of blood neutrophils was
followed over time (figure 2). Availability of label in the plasma during the procedure was used
to correct DNA enrichments. The first labeled neutrophil DNA was detected in the circulation
at day 6 after intake of label, representing the time it takes for the first myelocytes to mature

54
 Neutrophil circulatory lifespan

into neutrophils and to be released into the blood (i.e. the post-mitotic pool transit time,
PMPtt). This is in agreement with previously published results obtained from healthy
volunteers (26), but slightly longer than observed under non-homeostatic conditions (15, 21,
27). Deuterium enrichment in peripheral blood neutrophils reached its peak at day 8 after
intake of label, and decreased almost to baseline at the end of the study (day 17/18).

Figure 2. Deuterium enrichment in plasma glucose and peripheral blood neutrophil DNA. (A) Deuterium
enrichment of plasma glucose during intake of 6,6-2H2-glucose, which determines the availability of label for
each volunteer. For every individual 4 time points were studied (mean ± SD, n=32). (B) Deuterium enrichment
of neutrophil DNA after label intake, with 6 time points for 12 volunteers and an additional single time point for
20 volunteers (92 data points in total), after correction for each individual's plasma enrichment and the factor
0.65??? The dashed curve joins the mean enrichment levels per time point, while error bars indicate SD.

Fitting data using ODE’s

In order to estimate the neutrophil lifespan in our experiments, we first fitted our data to an
ODE model adapted from the study performed by Vrisekoop et al. (28) (see supplementary
materials). As each volunteer had measurements on different moments after intake of label,
we used a mixed effects approach, which first fits all combined data to the model and then
estimates parameters for each individual. The model assumes 2H-incorporation into the DNA
of neutrophil progenitors in the bone marrow during the labelling phase and exponential loss
of label after cessation of label intake. Label enters the circulation with the same enrichment
kinetics as the bone marrow, with an additional delay in the PMP. Initially, the PMPtt was
assumed to be the same for each cell. Fits were performed using different starting values for
each parameter, assuming blood or BM to be the compartment with the slowest kinetics as
described by Pillay et al (11) and Li et al (12). Alternatively, starting values were chosen in
which both compartments had similar kinetics.
Fitting our model to the data resulted in two nearly identical fits with no significant
difference in quality (as determined by the Akaike Information Criterium) (figure 3A). The two
fits represented two extremes, with either progenitor cells dividing every 4.3 days and a
neutrophil circulatory lifespan of 3.7h or the bone marrow dividing a bit more than once an
hour and a 4.8 day circulatory lifespan. Both fits estimated the PMPtt to be on average 6.4
days.

55
Chapter 3

Figure 3. Fitting an ODE model to the

experimental data resulted in 2 fits of
equal quality and nearly identical curves
(A). The difference between the two fits is
whether the slowest kinetics take place in
the BM or in blood. For both solutions
(slow BM or slow blood), parameter
estimates for each individual are plotted
for the progenitor division time (B),
neutrophil circulatory lifespan (C) and
PMPtt (D).

The two extreme results of this model are difficult to reconcile with previous
publications on the behavior of circulatory neutrophils and/or neutrophil progenitors. An
average neutrophil circulatory lifespan of 3.7h is shorter than any report described to date in
healthy volunteers (17), and even the S-phase of the cell-cycle has been reported to be twice
as long (approximately 8h (29)). The model estimated the PMPtt of the first cells entering the
bloodstream to be 6.4 days on average. This long PMPtt is not in agreement with our labelling
data, since our results show the first appearance of labeled DNA in the blood between 5 and
6 days (figure 2B). In addition, the prerequisite in this model that all neutrophils have a similar
PMPtt has been challenged before (27). To mitigate this latter issue we added an extra
parameter to the model, which allows for variation in PMPtt between cells. In this variant of
the model, neutrophils were assumed to exit the PMP with a normally distributed variation
around an average PMPtt (see supplementary methods). Addition of this extra parameter
improved the fit to the early data points but not significantly improve the fit to the full dataset
and resulted in an even shorter circulatory lifespan (1.0h) or progenitor division time (2.4h).

Fitting the data applying agent-based modelling

As stated above, the ODE models we used to describe our labelling data resulted in estimates
of neutrophil circulatory lifespans or BM turnover that are improbable in the light of previous
experiments and cell biology. A caveat of the ODE model is that it assumes cells to behave
exponentially. This means that even if the average time between divisions in the BM (t-div) is
in the order of several days, some myelocytes can divide e.g. within 1 hour. It has been shown

56
 Neutrophil circulatory lifespan

that myelocyte divisions are not exponentially distributed in vivo. In fact, the time between
two myelocytes divisions appears to vary little, as shown by in vivo 3H-Tdr pulse-labelling (30,
31).
Therefore, an agent-based method was developed (see supplementary materials for
specifics) to analyze the pooled data from all volunteers. The agent-based model simulates
the kinetics of individual cells based on a conveyor-belt model as first described by Cartwright
et al (13) (Figure 4A). The model simulates both BM and blood kinetics, as well as a PMPtt. It
assumes that myelocytes that enter the cell cycle keep dividing until exit into the PMP. The 3
average time cells stay in the PMP until they enter the bloodstream as mature cells is the
PMPtt.
In the vasculature, neutrophils can be either circulating or marginated. Since the
marginated pool is thought to be in equilibrium with circulating neutrophils (25) we assume
that the level of deuterium enrichment is similar in both pools (14), and for simplicity did not
include the marginated pool in the model. We allowed for heterogeneity between cells by
adding a normally distributed variation around each parameter, i.e. each cell is assigned a
value for PMPtt from a normal distribution with mean parameter PMPtt and a SD of SD_PMP.
After each division, progenitor cells are assigned a new value drawn from the same
distribution for their next division.
The model was run ~375.000 times with different values for the division time in the
bone marrow (t-div), the PMPtt, the average circulatory lifespan, and different SDs for each
of these parameters. These values were based on the wide range of previous estimates on
neutrophil lifespan (5-192h) (17) and neutrophil progenitor division times (10-92h). We
allowed the average PMPtt to vary quite widely between cells, from 4-8 days. SD’s ranged
from 0 (no variation between cells) to a maximum of 40% of the mean parameter value. For
each circulatory lifespan that was tested, the exact same combination of other parameters
and parameter variations was run (>16.000). For each run, the quality of the fit to the
experimental data was determined by the sum of squared residuals (SSR), with a low SSR
indicating a better fit (Figure 3B). As agent-based models are stochastic, two different runs
with the same set of parameters can have a slightly different outcome. Therefore, several of
the best fits were re-run 10 times and plotted against the measured data (figure 3C-J).
Model runs with circulatory lifespans between 72 and 96 hours and t-div between 48 and 96h
mimicked the data best. Runs with shorter circulatory lifespans tended to show a too rapid
up and downlabelling while runs with even longer circulatory lifespans tended to show a too
slow down-labelling.
Due to the conveyor-belt nature of the model, runs with a circulatory lifespan shorter
than t-div and a low SD showed waves of labelled cells entering and rapidly leaving the
circulation (e.g. figure 3C). With a short circulatory lifespan, fits to the data improved when
there was more variation in the PMPtt. A large variation in PMPtt, however, leads to the
model prediction that labelled cells enter the circulation very early during the experiment (e.g
figure 3D). In contrast, model runs with a longer circulatory lifespan fitted the data better
when the variation in the PMPtt was low (e.g. figure 3I). To give more insight into which

57
Chapter 3

Figure 4. (A) Schematic representation of the neutrophil compartment as simulated by the agent-based model.
The model is a linear conveyor-belt model, in which cells subsequently go through each phase: division,
maturation in the post-mitotic pool and residence in circulation followed by clearance from circulation. To allow
for biological variation, cells go through each phase at a mean speed with a normally distributed variability. The
SD of the variation is an independent parameter for each phase. This model was run >375.000 times and the
quality of the fit to the experimental data was determined by the SSR method, in which a lower SSR indicates a
better fit. For each run, the SSR was plotted against the circulatory neutrophil lifespan (B), with a cutoff above
10 (indicating irrelevant fits) for clarity. The best fitting parameter combinations for a circulatory lifespan of 7,
15, 24, 42, 60, 72, 96 and 120 hours (C-J) were rerun 10 times and plotted against the measured enrichments.
Information on the values for the other parameters can be found in supplementary figure S2. Green lines
represent BM DNA enrichment, red lines represent blood neutrophil enrichment +/- 95% CI (light area’s) and
black dots represent the measured DNA enrichments +/- SD.

58
 Neutrophil circulatory lifespan

parameter combinations were used, and which showed the best fits to the experimental data,
all possible combinations were plotted, with the values of the best fits highlighted in
supplementary figure S2.

Discussion
Re-utilization of label
Label re-utilization can be a major concern in interpreting labelling studies (32). However, as
we are using a pulse-chase type of labelling and neutrophils have a PMPtt of approximately 6 3
days, re-utilization of labeled nucleotides from neutrophils removed from the circulation
cannot occur before day 6 after intake of label. After re-uptake of label, it will take an
additional 5-6 days before the re-utilized label can be found in the DNA of circulating
neutrophils due to the length of the PMPtt. Thus, reutilized label will not affect our 2H
enrichment data for at least the first 12 days. In addition, we confirm previous studies (26)
showing that the majority of the adenosine moiety is derived from de novo synthesis and not
by base salvage. Taken together, these data indicate that re-utilization of label is of less
concern than it might be during long-term labelling strategies.

ODE versus an agent-based model

The ODE models provided reasonable fits of the neutrophil DNA enrichment data (figure 2A).
However, the combination of estimates of these fits for neutrophil lifespan and t-div were
biologically not plausible. A caveat of ODE model is that they may not correctly describe the
kinetics of neutrophil progenitors in the BM. ODE models assume progenitor cells to behave
exponentially, while in fact they have been shown to have a very low variation in time
between divisions (30). In addition, the first ODE model we used assumed that every
neutrophil takes the same amount of time to mature and enter the bloodstream. This notion
has been challenged by other authors (27), but the addition of an extra parameter allowing
variation in the PMPtt did not produce significantly better fits. In addition, this extra
parameter led to even less biologically plausible results, with an estimated neutrophil
circulatory lifespan of only one hour.
We also developed and applied an agent-based model, which describes neutrophil
maturation as a conveyor belt, with bone marrow cells having little variation between
divisions. The model allowed for (but did not require) variation between PMPtt’s of different
cells. As shown in figure 4, this model was able to fit the measured data and produced
estimates for neutrophil lifespan and t-div in a biologically plausible range.
The agent based model was not able to produce exact parameter estimates and did
not take into account the information lying in repeated measurements per individual. It did,
however, better describe the behavior of cells in the BM as described in literature and
produced more biologically plausible estimates of a 72-96 h circulatory lifespan and a 72h t-
div. Therefore, we propose that the model may be a better option for the interpretation of
neutrophil 2H enrichment data in the context of cellular kinetics.

59
Chapter 3

Bone marrow kinetics

Even though we did not acquire BM samples from healthy volunteers, our computer model
made predictions regarding the turnover of neutrophil progenitors. Simulating the blood data
obtained after short term 2H-glucose labeling with the agent-based model yielded an estimate
of the average circulatory neutrophil lifespan of 72-96h, and a relatively long t-div (72h). This
relatively long t-div is in the same order of magnitude as the hypothetical one proposed by Li
et al in response to our 2H2O labeling study (12).
Little direct data is available on the turnover of neutrophil progenitors in the BM, the
myelocytes. The only published in vivo labelling experiments measuring myelocyte kinetics in
bone marrow employed tritium thymidine (3H-TdR) pulse-labelling in both humans (30) and
rats (31). This was followed by measurement of the fraction of labeled mitoses in myelocytes
obtained from bone marrow samples. In this experiment the percentage of 3H-TdR labelled
mitoses increased rapidly (>90% at 3h post administration of label) until all mitotic cells were
3H-TdR labelled. Fifteen hours after administration of label, the percentage of labelled

mitoses decreased rapidly again, to be followed by all mitoses being labelled again ~18h after
administration of label. These data demonstrated that all myelocytes that go into mitosis
when 3H-Tdr is present go into mitosis again after 18h. In addition, during the second
myelocyte division 100% of mitoses are labelled, indicating there is very little variation in the
time between myelocyte divisions and suggesting that myelocytes in a pulse-chase labeling
experiment do not behave exponentially.
Another study employed in vivo labelling with diisopropyl fluorophosphate-32 (DF32P)
and revealed two phases of label decrease in the blood, one of which was interpreted as the
kinetics in BM, with an average division time of 3 days (13). However, interpretation and
comparison of the 3 studies (13, 30), (31) are not straightforward, because the first
experiment was performed under non-homeostatic conditions in a patient with mild
polycythemia vera and the second was performed using DF32P. Treatment of individuals with
DFP in vivo has been shown to induce priming and rapid loss of neutrophils from the
circulation (18, 33). Therefore, the first phase of decrease of label in the latter study more
likely reflected DFP-induced loss of labelled cells rather than the normal half-life of blood
neutrophils. This also makes the second phase a combination of loss due to BM and blood
kinetics, without a possibility to distinguish between the two.
In vitro, myelocytes have been followed by video microscopy (34) and by 3H-TdR
labelling (35) revealing ex vivo cycle times of 30h and 22h, respectively. These estimates are
very near the estimated t-div of 18h described in vivo (18,19). In addition to direct
measurements, the turnover of myelocytes can be estimated from the percentage of
myelocytes in mitosis (mitotic index) and the duration of the mitosis (35). Interpretation of
such data should take into account that approximately 50% of myelocytes form a “lazy pool”
of cells, which do not divide within the timeframe of the labelling experiments and thus do
not contribute to the incorporation of label in neutrophils (35, 36). From these studies it was
estimated that myelocytes divide on average once every 21.5h (35), which is again very similar
to results obtained in vivo (30). Taken together, data on directly labelled BM 3H-TdR suggest

60
 Neutrophil circulatory lifespan

that the division time of neutrophil progenitor cells in BM is much faster than the 3.75 days
required to reconcile our 2H2O labelling data (11) with a short circulatory neutrophil lifespan.
Importantly, the BM turnover in vivo remains to be determined in healthy volunteers with a
non-toxic labelling strategy.
Our results do not seem to be in line with the in vivo and in vitro methods described
above (30, 34, 35), which indicated shorter neutrophil progenitor division times ranging from
18-30 hours. However, these data were obtained in vitro or from patients with hematological
disorders. It is unknown whether division times of myelocytes in vitro represent the situation 3
in vivo. In addition, BM has been shown to undergo increased turnover during non-
homeostatic conditions such as inflammation (37, 38). Therefore, it is possible that these data
reflected the BM kinetics under non-homeostatic conditions, whereas our predictions reflect
homeostatic conditions. Of importance is that restricting division times of myelocytes in the
BM to a maximum of 30h is associated with longer blood lifespans (supplementary figure S2).
Therefore, the estimates of our model will be conservative as our model predicts a relatively
long t-div.

Conclusion
Our study applying short-term labelling with 2H-glucose in vivo has circumvented problems
regarding long-term labelling with 2H2O and suggests that a relatively long division time of
progenitors in the BM is not necessarily associated with a short (<1 day) life span of
neutrophils in peripheral blood. Our study supports a relatively long (2-4 day) lifespan of
neutrophils in the peripheral blood under homeostatic conditions.
The best estimate of the circulatory neutrophil lifespan based on the agent-based
model is 3 – 9 times longer than that obtained using 3H-TdR or DF32P labelling, which can be
explained by the use of toxic labelling methods or non-homeostatic conditions (17). On the
other hand, the estimated circulatory lifespans are shorter than those previously presented
by our group using 2H2O labelling. As stated above, long term 2H2O labelling of volunteers is
not ideal for estimating the kinetics of circulating neutrophils. In addition, the model used in
the 2H2O study was only capable of determining the slowest turnover, either in blood or BM.
The results obtained by pulse-chase 6,62H2–glucose labelling were first analyzed by ODE
models taking label kinetics into account in both blood and BM resulted in implausible
estimates of either the circulatory neutrophil lifespan or the t-div of neutrophil progenitors
in the BM. Using an AB model to simulate the label kinetics, based on the known behavior of
neutrophil progenitor cells in the BM, we show that this model is not agreement with a short
neutrophil circulatory lifespan. Instead, simulations of out AB model suggest a neutrophil
circulatory lifespan of 3-4 days.

61
Chapter 3

References
1. Lekstrom-Himes JA, and Gallin JI. Immunodeficiency diseases caused by defects in
phagocytes. The New England journal of medicine. 2000;343(23):1703-14.
2. Muller I, Munder M, Kropf P, and Hansch GM. Polymorphonuclear neutrophils and T
lymphocytes: strange bedfellows or brothers in arms? Trends in immunology.
2009;30(11):522-30.
3. Nauseef WM, and Borregaard N. Neutrophils at work. Nature immunology.
2014;15(7):602-11.
4. Sagiv JY, Michaeli J, Assi S, Mishalian I, Kisos H, Levy L, Damti P, Lumbroso D, Polyansky L,
Sionov RV, et al. Phenotypic diversity and plasticity in circulating neutrophil
subpopulations in cancer. Cell reports. 2015;10(4):562-73.
5. Radsak M, Iking-Konert C, Stegmaier S, Andrassy K, and Hansch GM. Polymorphonuclear
neutrophils as accessory cells for T-cell activation: major histocompatibility complex class
II restricted antigen-dependent induction of T-cell proliferation. Immunology.
2000;101(4):521-30.
6. Gabrilovich DI, and Nagaraj S. Myeloid-derived suppressor cells as regulators of the
immune system. Nature reviews Immunology. 2009;9(3):162-74.
7. Matsushima H, Geng S, Lu R, Okamoto T, Yao Y, Mayuzumi N, Kotol PF, Chojnacki BJ,
Miyazaki T, Gallo RL, et al. Neutrophil differentiation into a unique hybrid population
exhibiting dual phenotype and functionality of neutrophils and dendritic cells. Blood.
2013;121(10):1677-89.
8. Gregory AD, and Houghton AM. Tumor-associated neutrophils: new targets for cancer
therapy. Cancer research. 2011;71(7):2411-6.
9. Bruijnzeel PL, Uddin M, and Koenderman L. Targeting neutrophilic inflammation in severe
neutrophilic asthma: can we target the disease-relevant neutrophil phenotype? Journal
of leukocyte biology. 2015;98(4):549-56.
10. Brown KA, and Treacher DF. Neutrophils as potential therapeutic targets in sepsis. Discov
Med. 2006;6(33):118-22.
11. Pillay J, den Braber I, Vrisekoop N, Kwast LM, de Boer RJ, Borghans JA, Tesselaar K, and
Koenderman L. In vivo labeling with 2H2O reveals a human neutrophil lifespan of 5.4
days. Blood. 2010;116(4):625-7.
12. Li KW, Turner SM, Emson CL, Hellerstein MK, and Dale DC. Deuterium and neutrophil
kinetics. Blood. 2011;117(22):6052-3; author reply 3-4.
13. Cartwright GE, Athens JW, and Wintrobe MM. The Kinetics of Granulopoiesis in Normal
Man. Blood. 1964;24(780-803.
14. Athens JW, Haab OP, Raab SO, Mauer AM, Ashenbrucker H, Cartwright GE, and Wintrobe
MM. Leukokinetic studies. IV. The total blood, circulating and marginal granulocyte pools
and the granulocyte turnover rate in normal subjects. The Journal of clinical investigation.
1961;40(989-95.
15. Cronkite EP, Fliedner TM, Bond VP, and Rubini JR. Dynamics of hemopoietic proliferation
in man and mice studied by H3-thymidine incorporation into DNA. Annals of the New
York Academy of Sciences. 1959;77(803-20.
16. Fliedner TM, Cronkite EP, Killmann SA, and Bond VP. Granulocytopoiesis. Ii. Emergence
and Pattern of Labeling of Neutrophilic Granulocytes in Humans. Blood. 1964;24(683-
700.

62
 Neutrophil circulatory lifespan

17. Tak T, Tesselaar K, Pillay J, Borghans JA, and Koenderman L. What's your age again?
Determination of human neutrophil half-lives revisited. Journal of leukocyte biology.
2013.
18. Dresch C, Najean Y, and Beauchet J. In vitro 51Cr and 32P-DFP labeling of granulocytes in
man. Journal of nuclear medicine : official publication, Society of Nuclear Medicine.
1971;12(12):774-84.
19. Vincent PC, Chanana AD, Cronkite EP, and Joel DD. The intravascular survival of
neutrophils labeled in vivo. Blood. 1974;43(3):371-7.
20. Summers C WJ, Mackenzie I, Solanki C, Peters A, Chilvers E. Measurement of neutrophil 3
pulmonary transit time in humans [abstract]. American Journal of Respiratory Critical
Care. 2009;197A(A1333.
21. Steinbach KH, Schick P, Trepel F, Raffler H, Dohrmann J, Heilgeist G, Heltzel W, Li K, Past
W, van der Woerd-de Lange JA, et al. Estimation of kinetic parameters of neutrophilic,
eosinophilic, and basophilic granulocytes in human blood. Blut. 1979;39(1):27-38.
22. Macallan DC, Asquith B, Zhang Y, de Lara C, Ghattas H, Defoiche J, and Beverley PC.
Measurement of proliferation and disappearance of rapid turnover cell populations in
human studies using deuterium-labeled glucose. Nature protocols. 2009;4(9):1313-27.
23. Mengelers HJ, Maikoe T, Hooibrink B, Kuypers TW, Kreukniet J, Lammers JW, and
Koenderman L. Down modulation of L-Selectin expression on eosinophils recovered from
bronchoalveolar lavage fluid after allergen provocation. Clinical and experimental allergy
: journal of the British Society for Allergy and Clinical Immunology. 1993;23(3):196-204.
24. Buckley CD, Ross EA, McGettrick HM, Osborne CE, Haworth O, Schmutz C, Stone PC,
Salmon M, Matharu NM, Vohra RK, et al. Identification of a phenotypically and
functionally distinct population of long-lived neutrophils in a model of reverse
endothelial migration. Journal of leukocyte biology. 2006;79(2):303-11.
25. Summers C, Rankin SM, Condliffe AM, Singh N, Peters AM, and Chilvers ER. Neutrophil
kinetics in health and disease. Trends in immunology. 2010;31(8):318-24.
26. Macallan DC, Fullerton CA, Neese RA, Haddock K, Park SS, and Hellerstein MK.
Measurement of cell proliferation by labeling of DNA with stable isotope-labeled glucose:
studies in vitro, in animals, and in humans. Proc Natl Acad Sci U S A. 1998;95(2):708-13.
27. Dancey JT, Deubelbeiss KA, Harker LA, and Finch CA. Neutrophil kinetics in man. The
Journal of clinical investigation. 1976;58(3):705-15.
28. Vrisekoop N, den Braber I, de Boer AB, Ruiter AF, Ackermans MT, van der Crabben SN,
Schrijver EH, Spierenburg G, Sauerwein HP, Hazenberg MD, et al. Sparse production but
preferential incorporation of recently produced naive T cells in the human peripheral
pool. Proc Natl Acad Sci U S A. 2008;105(16):6115-20.
29. Sahar E, Wage ML, and Latt SA. Maturation rates and transition probabilities of cycling
cells. Cytometry. 1983;4(3):202-10.
30. Stryckmans P, Cronkite EP, Fache J, Fliedner TM, and Ramos J. Deoxyribonucleic acid
synthesis time of erythropoietic and granulopoietic cells in human beings. Nature.
1966;211(5050):717-20.
31. Constable TB, and Blackett NM. The cell population kinetics of neutrophilic cells. Cell and
tissue kinetics. 1972;5(4):289-302.
32. Tofts PS, Chevassut T, Cutajar M, Dowell NG, and Peters AM. Doubts concerning the
recently reported human neutrophil lifespan of 5.4 days. Blood. 2011;117(22):6050-2;
author reply 3-4.

63
Chapter 3

33. Tsang JL, Parodo JC, and Marshall JC. Regulation of apoptosis and priming of neutrophil
oxidative burst by diisopropyl fluorophosphate. Journal of inflammation. 2010;7(32.
34. Boll I, and Kuhn A. Granulocytopoiesis in human bone marrow cultures studied by means
of kinematography. Blood. 1965;26(4):449-70.
35. Dresch C, Troccoli G, and Mary JY. Growth fraction of myelocytes in normal human
granulopoiesis. Cell and tissue kinetics. 1986;19(1):11-22.
36. Mary JY. Normal human granulopoiesis revisited. II. Bone marrow data. Biomedicine &
pharmacotherapy = Biomedecine & pharmacotherapie. 1985;39(2):66-77.
37. Aglietta M, Piacibello W, Sanavio F, Stacchini A, Apra F, Schena M, Mossetti C, Carnino F,
Caligaris-Cappio F, and Gavosto F. Kinetics of human hemopoietic cells after in vivo
administration of granulocyte-macrophage colony-stimulating factor. The Journal of
clinical investigation. 1989;83(2):551-7.
38. Rosinski M, Yarmush ML, and Berthiaume F. Quantitative dynamics of in vivo bone
marrow neutrophil production and egress in response to injury and infection. Annals of
biomedical engineering. 2004;32(8):1108-19.

64
 Neutrophil circulatory lifespan

Supplementary data

Supplementary figure 1. Sort strategy. Neutrophils were isolated by FACS-sorting based on (A) FSC/SSC, (B)
doublet exclusion and (C) CD16high/CD193neg. (D) Representative cytospin of sorted neutrophils stained with
May-Grünwald-Giemsa. Used magnification 25x, bar indicates 25 μm.

65
Chapter 3

66
Neutrophil circulatory lifespan

67
Chapter 3

Supplementary figure S2. Tested parameter values and the parameter values of the best fits. Tested values for
all parameters (PMPtt, t-div, and the three standard deviations) were plotted in black with the 10 best fitting
parameter sets in blue for each tested circulatory lifespan. Numbers indicate which set of parameters fitted best
within the tested lifespan (1 is best). Especially the short circulatory lifespans tend to show random patterns of
the 10 best fits, which may be due to overfitting as illustrated in figure 2.

68
Neutrophil circulatory lifespan

69
 Chapter 4
Kinetics of eosinophils and basophils in
the circulation of healthy humans

Tak, T.1 Tesselaar, K.2 Borghans, J.A.M.2 Koenderman, L.1

1. Department of Respiratory Medicine Laboratory of Translational Immunology,

University Medical Centre Utrecht, Utrecht, the Netherlands.
2. Department of Immunology, Laboratory of Translational Immunology, Univer-
sity Medical Centre Utrecht, Utrecht, the Netherlands.

Manuscript in preparation
Chapter 4

72
 Circulatory eosinophil and basophil kinetics

To the editor,
Eosinophils and basophils play an important role in the protection against invading
multicellular parasites, and exhibit a detrimental role in allergic diseases. Because of this,
therapeutics are in use to prevent the production of eosinophils (α-IL-5 and α-IL5R/CD125)
or the activation of basophils (α-IgE) in patients with a wide range of allergic diseases.
However, remarkably little is known about the kinetics of these cells in homeostasis, which
hampers the interpretation of the effects of these therapeutics. A clear example is the lack of
understanding of prolonged tissue eosinophilia in the asthmatic lung during treatment with 3
the anti-IL-5 antibody Mepolizumab (1).
Only one study has attempted to determine the circulatory lifespan of basophils in
humans, and estimated it to be 89h (2). However, this study was performed in terminal cancer
4
patients because of the toxicity of tritiated thymidine (3H-TdR) that was used in the study. It
is very likely that the pathogenesis of disease, its treatment and the toxicity of the label
influenced the cellular kinetics. The average circulatory lifespan of eosinophils was estimated
to be between 11 and 63h in patients with cancer (2, 3) or hypereosinophilia (3, 4) using
chromium-51 (4) and 3H-Tdr (2, 3). Another study in healthy volunteers reported a 25h
lifespan (5), but this was based on reinfusion of ex vivo manipulated cells into the volunteer;
a procedure that has been shown to activate both eosinophils and neutrophils (6). Ex vivo
manipulation of neutrophils has been shown to cause sequestration in the lung (7), which was
also observed for ex vivo labelled eosinophils (5).
We labeled 7 healthy volunteers for 6h with 12 oral half-hourly doses of 6,62H2–
glucose, adding up to 1g/kg bodyweight, as adapted from Macallan et al (8). Written informed
consent was obtained from all volunteers in agreement with the declaration of Helsinki.
During the labelling procedure, small blood samples were collected to determine the
availability of label in the plasma. One volunteer reported gastroenteritis at day 10, so for this
volunteer measurements from day 7 and onwards were excluded from analysis.
At various time points after intake of label, blood was obtained in sodium heparin
tubes. After lysis of erythrocytes in a 150 mM NH4Cl, 10 mM KHCO3 0.1 mM NA2EDTA buffer,
total leukocytes were stained with CD16-FITC (clone 3G8), CD193-Alexa647 (83103, both from
BD biosciences) and CD9-IgM (clone S32, a kind gift from Ir. D. Kanters). A secondary anti-IgM-
PE (Southern Biotech) was used for detection of CD9. Eosinophils and basophils were isolated
by FACS sorting as SSChigh/CD16-/CCR3+/CD9high and SSClow/CD16-/CCR3+/CD9high, respectively
(supplementary figure S1). Cellular DNA was isolated after sorting of typically >99% pure cell
populations. Subsequently, 2H-enrichment in the DNA was determined in the adenosine
moiety as previously described (8).
Computational modelling was performed as previously described in detail (chapter 2)
using an agent-based model that assumes an effective 6,6-2H2-glucose pulse-length of 8h and
takes into account incorporation of label in the bone marrow, followed by maturation in the
post mitotic pool (PMP), release from the bone marrow into the bloodstream and ultimately
clearance from the circulation.

73
Chapter 4

Computational models were run using NetLogo 5.0.2 and analyzed using R version
3.1.2. FACS data were analyzed with FCS Express Flow V5.01, graphs were made using
GraphPad Prism 6.05. Composite figures were made using Adobe Illustrator CC 2015.
Plasma glucose 2H-enrichment was determined to correct for availability of label,
(figure 1A). After correction, DNA 2H-enrichment data showed that the first labelled
eosinophils and basophils appeared 4/5 days after intake of label (figure 1B). This
demonstrates that the time between the last division of an eosinophil progenitor and release
of the first mature cells into the blood (PMP transit time; PMPtt) is approximately 4-5 days.
Thereafter, the differences between eosinophils and basophils became quite pronounced:
Eosinophils reached their maximum enrichment around day 7, while basophils reached their
maximum enrichment at around day 12 after intake of label. The previous study by Steinbach
et al using 3H-thymidine labelling in cancer patients found a shorter PMPtt for both
eosinophils (2-4 days) and basophils (3-4 days) (2). A possible explanation for these
discrepancies is that inflammation in cancer patients might reduce the PMPtt, such as
described for neutrophils in inflammation (9). We do show eosinophilic inflammation to lead
to a longer eosinophil PMPtt (see chapter 10), but this type of inflammation might alter the
PMPtt differently. Alternatively, the kinetics of 3H-Tdr incorporation might differ from those
of 6,6-2H2-glucose . Using the agent-based computational model, we performed 500,000 runs
with different parameter combinations with the following constraints: BM division times
between 18h-120h, PMPtt between 66-180h and circulatory lifespans of eosinophils and
basophils between 4-340h. For each of these runs, we determined how well the run described
the experimental data, using the sum of squared residuals (SSR). The enrichment curves of
eosinophils were best described by model runs with a circulatory lifespan between 66 and 88
hours, whereas for basophils model runs with circulatory lifespans between 192 and 312
hours mimicked the experimental data best (figure 1 C,D). As the model is stochastic, different
model runs may produce slightly different results, so the model was re-run 10 times with the
best estimates and plotted with the measured data (figure 1 E,F) These lifespans are much
longer than those previously reported for terminal cancer patients and based on reinfusion
of ex vivo labelled cells. These differences might, again, be explained by the presence of
inflammation in cancer patients and (pre)activation of ex vivo labelled eosinophils.
Although not based on deuterium measurements of bone marrow progenitors, our
computational model estimates that eosinophilic progenitors divide once every 96-120h and
basophilic progenitors every 48-84h. In rats, it was shown that approximately 1% of
eosinophilic progenitors are in mitosis (10). Assuming that the duration of the mitosis is about
1h (11), one can estimate that eosinophilic progenitors divide approximately once every 100h,
which would be in good agreement with our estimates. For basophils no data are available on
progenitor mitotic indices. Solid conclusions on the turnover of the different granulocyte
progenitor cells await direct measurement of deuterium incorporation in these cells from the
bone marrow.
In conclusion, we present the first kinetic data of untouched eosinophils and basophils
in the circulation of healthy individuals. Our analysis suggests that both eosinophils and

74
 Circulatory eosinophil and basophil kinetics

basophils have a PMPtt of 4-5 days, and that eosinophils and basophils have circulatory
lifespans of 66-88 h and 192-312 h, respectively.
Figure 1. 2H-enrichment
kinetics and computa-
tional modelling.
Plasma glucose 2H-
enrichment during the
labelling procedure (A).
DNA 2
H-enrichment 3
corrected for label
availability in plasma (B)
was detected at day 5 4
after intake of label in
both eosinophils and
basophils, after which
the two cell types show
marked differences in
kinetics. When fitting
the data with our
computational model,
the best fits for
eosinophils (C) were
reached with circulatory
lifespans of 66-88 h and
the best fits for
basophils for 192-312 h
(D) Arrows indicate the
best fit of each cell type,
which was rerun 10
times and plotted with
the measured data for
other parameters, see
supplementary table 1
for both eosinophils (E)
and basophils (F). Data
represent (A) means of 7
measurements at each
time point +/-95% CI, (B)
mean +/- SD of a total of
39 measurements from
7 volunteers (6 times 6
measurements, only 3
measurements for one volunteer), (C-D) best 200.000 individual model runs of ~500.000 and (E-F) the mean +/-
SD of 10 model re-runs with the estimated best parameters plotted with 39 individual measurements.

75
Chapter 4

References
1. Flood-Page PT, Menzies-Gow AN, Kay AB, and Robinson DS. Eosinophil's role remains
uncertain as anti-interleukin-5 only partially depletes numbers in asthmatic airway.
Am J Respir Crit Care Med. 2003;167(2):199-204.
2. Steinbach KH, Schick P, Trepel F, Raffler H, Dohrmann J, Heilgeist G, Heltzel W, Li K,
Past W, van der Woerd-de Lange JA, et al. Estimation of kinetic parameters of
neutrophilic, eosinophilic, and basophilic granulocytes in human blood. Blut.
1979;39(1):27-38.
3. Parwaresch MR, Walle AJ, and Arndt D. The peripheral kinetics of human radiolabelled
eosinophils. Virchows Archiv B: Cell pathology. 19761):57-66.
4. Dale DC, Hubert RT, and Fauci A. Eosinophil kinetics in the hypereosinophilic
syndrome. The Journal of laboratory and clinical medicine. 1976;87(3):487-95.
5. Farahi N, Singh NR, Heard S, Loutsios C, Summers C, Solanki CK, Solanki K, Balan KK,
Ruparelia P, Peters AM, et al. Use of 111-Indium-labeled autologous eosinophils to
establish the in vivo kinetics of human eosinophils in healthy subjects. Blood.
2012;120(19):4068-71.
6. Lukawska JJ, Livieratos L, Sawyer BM, Lee T, O'Doherty M, Blower PJ, Kofi M, Ballinger
JR, Corrigan CJ, Gnanasegaran G, et al. Real-time differential tracking of human
neutrophil and eosinophil migration in vivo. The Journal of allergy and clinical
immunology. 2014;133(1):233-9 e1.
7. Summers C, Singh NR, White JF, Mackenzie IM, Johnston A, Solanki C, Balan KK, Peters
AM, and Chilvers ER. Pulmonary retention of primed neutrophils: a novel protective
host response, which is impaired in the acute respiratory distress syndrome. Thorax.
2014;69(7):623-9.
8. Macallan DC, Asquith B, Zhang Y, de Lara C, Ghattas H, Defoiche J, and Beverley PC.
Measurement of proliferation and disappearance of rapid turnover cell populations in
human studies using deuterium-labeled glucose. Nature protocols. 2009;4(9):1313-27.
9. Terashima T, Wiggs B, English D, Hogg JC, and van Eeden SF. Polymorphonuclear
leukocyte transit times in bone marrow during streptococcal pneumonia. The
American journal of physiology. 1996;271(4 Pt 1):L587-92.
10. Bogatov LV. Mitotic indices of individual types of bone marrow cells of normal rats.
Bulletin of Experimental Biology and Medicine.69(2):198-200.
11. Killmann SA, Cronkite EP, Fliedner TM, and Bond VP. Mitotic Indices of Human Bone
Marrow Cells. 3. Duration of Some Phases of Erythrocytic and Granulocytic
Proliferation Computed from Mitotic Indices. Blood. 1964;24(267-80.

76
 Circulatory eosinophil and basophil kinetics

Supplementary Materials

Supplementary table 1. Values of parameter sets producing the best fits in figure 1 E and F.
Parameter Basophil Eosinophil
circulatory lifespan (h) 240 80
lifespan standard deviation (%) 20 30
tdiv precursors (h) 60 96
tdiv standard deviation (%) 0 20 3
average PMPtt (h) 136 124
PMPtt standard deviation (%) 30 20 4

Supplementary figure S1. Gating strategy for FACS sort. (A) Granulocyte and PBMC Singlets were gated based
on FSC/SSC. (B) Basophils were identified from SSClow PBMC’s as CD193+/CD9+. (C) Eosinophils were identified
from SSChigh granulocytes as CD193+/CD16-/CD9+. (D) Microscopic evaluation of May-Grünwald-Giemsa stained
cytospins show typical eosinophilic and basophilic granules. Object used: 100x (E) Re-analysis of sorted cells
show typical SSChigh/FSClow eosinophils (blue) and basophils (red) with a typical FSC/SSC in between that of
lymphocytes and monocytes.

77
 Chapter 5
Circulatory and maturation kinetics of human
monocyte subsets in vivo

Tak, T1. Drylewicz, J2. Conemans, L1. Borghans, JAM2. Tesselaar, K2. Koenderman, L1

1. Department of Respiratory Medicine Laboratory of Translational Immunology,

University Medical Centre Utrecht, Utrecht, the Netherlands.
2. Department of Immunology, Laboratory of Translational Immunology,
University Medical Centre Utrecht, Utrecht, the Netherlands.

Submitted
Chapter 5

Abstract
Monocytes are thought to be relatively short-lived cells that are essential for the immune
system. Their kinetics were previously studied assuming a single homogeneous monocyte
compartment. In fact, monocytes can be subdivided into at least three different subsets:
classical CD14++/CD16- (CM), intermediate CD14++/CD16+ (IM) and non-classical CD14+/CD16+
(NCM) monocytes, which are generally assumed to differentiate in a linear fashion from CM
via IM to NCM.
To investigate the circulatory and maturation kinetics of these monocyte subsets, we
employed in vivo deuterium labelling in humans. Fourteen volunteers were labelled with 6,6-
2H -glucose and subsequently 2H-DNA enrichment was determined over time in sorted
2
monocyte subsets. We used a mixed effect model to interpret the enrichment levels in the
different subsets, estimated their circulatory lifespans and assessed whether a linear
differentiation pattern is consistent with the labelling data.
We estimated the average circulatory lifespans of CM, IM and NCM to be 0.74, 1.1 and
3.1 days, respectively. Our data support the generally assumed linear differentiation from CM
via IM to NCM if the majority of CM does not differentiate into circulatory NCM and the last
step of this differentiation pathway takes place outside the circulation, suggesting that
monocytes leave the circulation as IM and re-enter as NCM.

80
 Monocyte circulatory and maturation kinetics

Introduction
Monocytes originate from the bone marrow (1), are distributed in the bloodstream, and can
differentiate into skin macrophages or intestinal DC’s (2). They play an essential role in the
defense against pathogens (3) and are implicated in a range of diseases (4).
Labelling experiments with 6,6-2H2-glucose (5) and 3H-thymidine (6) have suggested a
relatively short monocyte circulatory lifespan of 2-4 days. While these experiments
considered monocytes as a single population, the currently held view is that the blood
monocyte compartment consists of at least 3 distinct subsets: Classical CD14++/CD16-
monocytes (CM), intermediate CD14++/CD16+ monocytes (IM), and non-classical CD14-/CD16+
monocytes (NCM). The lifespans of these separate subsets remain to be determined. The
three subsets are characterized by gradually changing surface marker expressions (7),
differential gene profiling (8, 9), enhancer profiling (10), and different functionality (4, 11).
Their maturation kinetics also differ, as human CM have been shown to repopulate the 5
bloodstream more rapidly after haematopoietic stem cell transplantation than CD16+
monocytes (12). In rhesus macaques, which have comparable monocyte subsets, BrdU
labelling showed CM to be the first to become BrdU positive, followed by IM and later NCM
(13). The combination of gradually changing expression patterns, combined with their
consecutive repopulation/labelling kinetics has led to the prevailing idea that monocytes
follow a linear differentiation path from CM via IM to NCM (4, 10), although an alternative
model in which each subset develops in parallel from separate progenitors cannot be ruled
out.
In mice, two monocyte subsets have been described (14) for which a linear
differentiation pattern has been shown more directly, because 1) adoptively transferred CM
homologues were shown to differentiate into NCM (15, 16) and 2) in vivo imaged monocytes
were shown to lose the CM marker CCR2 whilst acquiring the NCM marker CXC3CR1 at sites
of sterile inflammation (17). However, due to doubts about the purity of adoptively
transferred CM preparations, the discussion on the developmental relation of the two
monocyte subsets under homeostatic conditions has not yet been settled (4, 15, 18).
Using in vivo 6,6-2H2-glucose pulse-chase labelling combined with mathematical
modelling, we here estimated the kinetics of the three human monocyte subsets under
homeostatic conditions, and assessed whether these labelling data are consistent with the
generally assumed linear differentiation pattern from CM via IM to NCM in the circulation.

81
Chapter 5

Results and discussion

Monocyte subset numbers and percentages (figure 1) were in agreement with previous
findings (4), with 88.5% CM, 4.4% IM and 7.1% NCM. Maximum DNA enrichment (figure 1)
was achieved first in CM (3-4 days), followed by IM (~4 days) and NCM (~8 days), in line with
recent results obtained in rhesus macaques (13), which seems to fit a putative linear
differentiation pattern. Our CM labelling data are in fact very similar to 3H-thymidine data on
the whole monocyte population (6, 19), which is not surprising since >80% of monocytes are
CM.

Figure 1. Analysis and isolation of circulating monocytes. Monocyte subsets were gated based on FSC/SSC (A),
singlets and CD14/CD16 expression (B). Resulting monocyte percentages (C) and numbers (D) showed CM to be
the most common circulating monocyte subset, followed by NCM and with IM being the smallest subset. FACS
analysis of sorted monocyte subsets (E) typically revealed a purity >99% and good separation between the
populations. Cytospin preparations stained with May-Grünwald-Giemsa (F) showed an increasingly mature
phenotype from CM to NCM as characterized by a more neutrophilic cytoplasm and increasingly dendritic
appearance. DNA 2H-enrichments after in vivo pulse-labeling with 6,6-2H2-glucose (G) were determined by GC-
MS and corrected for availability of label in the plasma during label administration and the fraction de novo
synthesized adenosine (see Methods). A,B,E,F: representative results from 1 experiment. C,D: Each circle
represents the mean of 6 measurements in one volunteer, lines indicate mean +/-SD for the 14 volunteers. F:
Objective used: 100x oil immersion lens, numerical aperture 1.30, scale bars indicate 10μm. G: 249
measurements from 14 individuals Mean +/- SD.

A mathematical model was fitted to the measured deuterium enrichment levels (figure
2A-B and supplementary methods) assuming all incorporation of label to occur at the
monocyte progenitor stage in the bone marrow (BM). After maturation in the post-mitotic
pool (PMP), the model assumes that cells enter the bloodstream as CM and subsequently

82
 Monocyte circulatory and maturation kinetics

mature into IM and NCM. Models assuming differentiation from IM into NCM directly in the
circulation failed to give good fits to the enrichment data (see supplementary figure S1). We
therefore extended the model with a delay during which IM differentiate into NCM outside
the circulation. This delay significantly increased the fit to the labeling data, and was
estimated to be 1.8 (1.4-2.1) days. Of note, adding a delay between CM and IM did not
significantly improve the fit to the data and was therefore not included in the model. Based
on the model fits, the average circulatory lifespans of cells in the three monocyte subsets
were estimated to be 0.74 (95% CI: 0.71-0.77), 1.1 (0.78-1.4) and 3.1 (2.7-3.6) days for CM,
IM and NCM, respectively (figure 2C, and table S1). In addition, the turnover rate of monocyte
precursors in the bone marrow was estimated to be 0.43 (0.42-0.45) per day while the
average maturation time in the PMP was 1.95 (1.74-2.1) days.
From the estimated circulatory lifespans and the percentages of monocytes in the
three subsets we could calculate the percentage of CM that differentiate into circulating IM 5
and the percentage of circulating IM that differentiate into circulating NCM (see
supplementary methods). We estimated that on average only 4.4% (2.4-6.4) of CM
differentiate into IM, while 53% (36.5-70.6) of IM mature into circulating NCM. This is in line
with previous experiments that suggested monocyte-derived macrophages in skin (12) and
DC’s in the gut (16) to be directly derived from CM, and not from IM or NCM, implying that
not all CM become IM. Alternatively, there might be pools of IM and NCM in the blood that
can nevertheless not be obtained by venipuncture, and therefore seem to be “lost” from the
circulation. The existence of non-circulating monocytes has been suggested in the form of
“patrolling” NCM (11) adhering to blood vessel walls or the marginated monocytes described
in mice (20, 21).
Even though our results can be fitted with a linear CM-IM-NCM differentiation, our
results do not exclude the possibility that any or all subsets develop separately. In the case of
IM and NCM, this would require longer retention in the PMP and can only truly be excluded
using lineage tracing experiments.
Our results have important implications for monocytes under homeostasis: we show
that all three monocyte subsets have a rapid turnover in circulation. Our data support the
hypothesis of a linear CM-IM-NCM differentiation if the majority of CM does not differentiate
into circulatory NCM and the last step of this differentiation pathway takes place outside the
circulation, suggesting that monocytes leave the circulation as IM and re-enter as NCM. From
these results important questions arise, such as what determines whether a monocyte
follows the CM/IM/NCM differentiation pathway or becomes a macrophage or DC, why IM
leave the circulation and how these processes are involved in immune homeostasis and
pathogenesis of immune-mediated diseases.

83
Chapter 5

Figure 2. Linear differentiation model used to describe the data (A). The model consists of a bone marrow
compartment in which label is incorporated during progenitor divisions. Subsequently, cells differentiate in a
linear fashion from progenitor cells into CM, IM and NCM. Two delays were added in the model. The first is a
delay between incorporation of label in progenitor cells and release as CM into the bloodstream: the PMPtt. The
second delay was added during maturation from IM to NCM, assuming that this differentiation step occurs
outside the circulation, after which monocytes return to the blood stream as NCM. From the resulting lifespans
and the known relative abundance of each subset, the "loss" of cells from the linear CM-IM-NCM differentiation
pattern was determined. Using this model, each parameter was estimated from all pooled data. The best fit of
the model (solid curves) was plotted together with the measured enrichment levels (dots) (B). Using a mixed
effect model, we estimated for each individual the cellular lifespans in the different monocyte subsets (C), the
delays in the PMP and between IM and NCM differentiation (D), and the percentage of cells that does not
differentiate from CM into IM or from IM into NCM (E). Data were obtained from 14 individuals. Best fits for
each individual can be found in supplementary figure S2.

Materials and Methods

Human volunteers
Five healthy volunteers and 9 asthma patients were included in the study after giving written
informed consent in accordance to the Declaration of Helsinki. Monocyte (subset) numbers
and kinetics did not differ between controls and asthmatics (see supplementary figure 3).
Therefore, data from both groups was combined for analysis. This study was approved by the
University Medical Centre Utrecht ethical review board (METC).
Volunteers received an oral dose of 1g 6,6-2H2-glucose (Cambridge Isotopes) per kg of
bodyweight over a period of 6h as published previously (22). Blood was drawn at six time
points within 17 days after labelling.

84
 Monocyte circulatory and maturation kinetics

Cell isolation
Monocyte populations were isolated from total leukocytes prepared by RBC lysis from
heparinized blood (23). Monocyte subpopulations were isolated by FACS sorting on a BD
AriaIII (see figure 1 for gating strategy). Resulting preparations were typically >99% pure.

DNA enrichment
After isolation of DNA from the different monocyte subsets, DNA 2H-enrichment was
determined as published previously (24) by GC-MS on adenosine derivatized to pentafluoro-
triacetate, with m/z ratios of 435 and 437 for unlabeled and labeled derivatives, respectively.
Resulting enrichments were corrected for each individual's availability of label in plasma and
the fraction de novo synthesized adenosine (0.65) as described previously (22).
5
Mathematical modelling
Monocyte lifespans were estimated using a mixed effects model with Monolix 4.3.3 taking
into account kinetics and delays in BM, blood and non-circulatory compartments as well as
relative cell numbers. For detailed description, see supplementary methods.

85
Chapter 5

References
1. van Furth R, and Cohn ZA. The origin and kinetics of mononuclear phagocytes. J Exp
Med. 1968;128(3):415-35.
2. Auffray C, Sieweke MH, and Geissmann F. Blood monocytes: development,
heterogeneity, and relationship with dendritic cells. Annual review of immunology.
2009;27(669-92.
3. Serbina NV, Jia T, Hohl TM, and Pamer EG. Monocyte-mediated defense against
microbial pathogens. Annual review of immunology. 2008;26(421-52.
4. Wong KL, Yeap WH, Tai JJ, Ong SM, Dang TM, and Wong SC. The three human
monocyte subsets: implications for health and disease. Immunologic research.
2012;53(1-3):41-57.
5. Mohri H, Perelson AS, Tung K, Ribeiro RM, Ramratnam B, Markowitz M, Kost R, Hurley
A, Weinberger L, Cesar D, et al. Increased turnover of T lymphocytes in HIV-1 infection
and its reduction by antiretroviral therapy. J Exp Med. 2001;194(9):1277-87.
6. Whitelaw DM. Observations on human monocyte kinetics after pulse labeling. Cell and
tissue kinetics. 1972;5(4):311-7.
7. Hijdra D, Vorselaars AD, Grutters JC, Claessen AM, and Rijkers GT. Phenotypic
characterization of human intermediate monocytes. Frontiers in immunology.
2013;4(339.
8. Wong KL, Tai JJ, Wong WC, Han H, Sem X, Yeap WH, Kourilsky P, and Wong SC. Gene
expression profiling reveals the defining features of the classical, intermediate, and
nonclassical human monocyte subsets. Blood. 2011;118(5):e16-31.
9. Zawada AM, Rogacev KS, Rotter B, Winter P, Marell RR, Fliser D, and Heine GH.
SuperSAGE evidence for CD14++CD16+ monocytes as a third monocyte subset. Blood.
2011;118(12):e50-61.
10. Schmidl C, Renner K, Peter K, Eder R, Lassmann T, Balwierz PJ, Itoh M, Nagao-Sato S,
Kawaji H, Carninci P, et al. Transcription and enhancer profiling in human monocyte
subsets. Blood. 2014;123(17):e90-9.
11. Cros J, Cagnard N, Woollard K, Patey N, Zhang SY, Senechal B, Puel A, Biswas SK,
Moshous D, Picard C, et al. Human CD14dim monocytes patrol and sense nucleic acids
and viruses via TLR7 and TLR8 receptors. Immunity. 2010;33(3):375-86.
12. McGovern N, Schlitzer A, Gunawan M, Jardine L, Shin A, Poyner E, Green K, Dickinson
R, Wang XN, Low D, et al. Human dermal CD14(+) cells are a transient population of
monocyte-derived macrophages. Immunity. 2014;41(3):465-77.
13. Sugimoto C, Hasegawa A, Saito Y, Fukuyo Y, Chiu KB, Cai Y, Breed MW, Mori K, Roy CJ,
Lackner AA, et al. Differentiation Kinetics of Blood Monocytes and Dendritic Cells in
Macaques: Insights to Understanding Human Myeloid Cell Development. J Immunol.
2015.
14. Sunderkotter C, Nikolic T, Dillon MJ, Van Rooijen N, Stehling M, Drevets DA, and
Leenen PJ. Subpopulations of mouse blood monocytes differ in maturation stage and
inflammatory response. J Immunol. 2004;172(7):4410-7.

86
 Monocyte circulatory and maturation kinetics

15. Geissmann F, Manz MG, Jung S, Sieweke MH, Merad M, and Ley K. Development of
monocytes, macrophages, and dendritic cells. Science. 2010;327(5966):656-61.
16. Varol C, Landsman L, Fogg DK, Greenshtein L, Gildor B, Margalit R, Kalchenko V,
Geissmann F, and Jung S. Monocytes give rise to mucosal, but not splenic,
conventional dendritic cells. J Exp Med. 2007;204(1):171-80.
17. Dal-Secco D, Wang J, Zeng Z, Kolaczkowska E, Wong CH, Petri B, Ransohoff RM, Charo
IF, Jenne CN, and Kubes P. A dynamic spectrum of monocytes arising from the in situ
reprogramming of CCR2+ monocytes at a site of sterile injury. J Exp Med.
2015;212(4):447-56.
18. Shi C, and Pamer EG. Monocyte recruitment during infection and inflammation.
Nature reviews Immunology. 2011;11(11):762-74.
19. Cronkite EP, Fliedner TM, Bond VP, and Rubini JR. Dynamics of hemopoietic
proliferation in man and mice studied by H3-thymidine incorporation into DNA. Annals 5
of the New York Academy of Sciences. 1959;77(803-20.
20. van Furth R, and Sluiter W. Distribution of blood monocytes between a marginating
and a circulating pool. J Exp Med. 1986;163(2):474-9.
21. Barletta KE, Cagnina RE, Wallace KL, Ramos SI, Mehrad B, and Linden J. Leukocyte
compartments in the mouse lung: distinguishing between marginated, interstitial, and
alveolar cells in response to injury. Journal of immunological methods. 2012;375(1-
2):100-10.
22. Macallan DC, Asquith B, Zhang Y, de Lara C, Ghattas H, Defoiche J, and Beverley PC.
Measurement of proliferation and disappearance of rapid turnover cell populations in
human studies using deuterium-labeled glucose. Nature protocols. 2009;4(9):1313-27.
23. Pillay J, Ramakers BP, Kamp VM, Loi AL, Lam SW, Hietbrink F, Leenen LP, Tool AT,
Pickkers P, and Koenderman L. Functional heterogeneity and differential priming of
circulating neutrophils in human experimental endotoxemia. Journal of leukocyte
biology. 2010;88(1):211-20.
24. Busch R, Neese RA, Awada M, Hayes GM, and Hellerstein MK. Measurement of cell
proliferation by heavy water labeling. Nature protocols. 2007;2(12):3045-57.

87
Chapter 5

Supplementary Information

Supplementary figure S1. Best fits of models with IM to NCM maturation inside the circulation. Models
assuming IM to differentiate into NCM in the circulation without delay (A) do not produce good fits for NCM
uplabelling. Models allowing a fraction of IM to leave the circulation, while the rest remains in the circulation
for a longer amount of time (B) do not produce good fits either. The model used in the rest of the paper assuming
IM to leave the circulation during maturation into NCM does produce good fits to the enrichment data. Lines
indicate fits, symbols represent means +/- range of measured enrichments.

Supplementary figure S2. Best fits of the mathematical model to the enrichment levels of the three monocyte
subsets for each volunteer. Some data points are missing because not enough cells were collected to ensure
reliable quantification of 2H-DNA enrichment.

88
 Monocyte circulatory and maturation kinetics

5
Supplementary figure S3. Comparison of blood monocyte subsets between healthy controls and asthma
patients. DNA 2H-enrichment kinetics for CM (A), IM (B) and NCM (C) show no obvious differences between
patients and healthy controls. No statistically significant differences were found between healthy volunteers
and asthma patients in subset percentages (D), absolute cell numbers (E) and estimated cellular lifespans (F). All
plots represent data from 5 healthy volunteers and 9 asthma patients, with subset percentages (D) and counts
(E) representing mean values of each individual during the study.

89
90
Chapter 5

Supplementary table 1. Parameter estimates for each individual

TURNOVER D (DAYS-1) LIFESPAN (DAYS)* DELAY Δ (DAYS) LOSS (FRACTION)2
ID BM CM IM NCM CM IM NCM BM->CM IM->NCM CM->IM IM->NCM
1 0.43 (0.04) 1.41 (0.18) 1.69 (0.62) 0.18 (0.04) 0.71 (0.09) 0.59 (0.22) 5.59 (1.26) 1.69 (0.19) 1.83 (0.54) 0.96 0.85
2 0.42 (0.03) 1.38 (0.2) 1.41 (0.4) 0.33 (0.08) 0.72 (0.1) 0.71 (0.2) 3.05 (0.76) 1.83 (0.19) 2.81 (0.41) 0.94 0.76
3 0.45 (0.03) 1.26 (0.17) 0.58 (0.11) 0.26 (0.06) 0.8 (0.11) 1.71 (0.32) 3.84 (0.83) 2.76 (0.15) 1.98 (0.84) 0.98 0.10
4 0.44 (0.04) 1.4 (0.17) 0.41 (0.08) 0.46 (0.13) 0.71 (0.08) 2.43 (0.47) 2.16 (0.6) 1.72 (0.15) 1.08 (0.3) 0.98 0.00
5 0.43 (0.03) 1.3 (0.17) 3.35 (0.99) 0.31 (0.07) 0.77 (0.1) 0.3 (0.09) 3.18 (0.69) 1.7 (0.16) 0.84 (0.21) 0.84 0.85
6 0.36 (0.03) 1.22 (0.18) 1.49 (0.73) 0.42 (0.1) 0.82 (0.12) 0.67 (0.33) 2.39 (0.57) 2.14 (0.17) 1.69 (0.4) 0.96 0.73
7 0.41 (0.03) 1.37 (0.19) 0.9 (0.27) 0.25 (0.06) 0.73 (0.1) 1.11 (0.34) 4.02 (0.98) 1.77 (0.17) 2.64 (0.56) 0.97 0.34
8 0.46 (0.03) 1.32 (0.19) 0.96 (0.21) 0.31 (0.09) 0.76 (0.11) 1.05 (0.23) 3.21 (0.96) 2.54 (0.16) 2.09 (0.42) 0.97 0.30
9 0.43 (0.03) 1.43 (0.23) 1.03 (0.34) 0.35 (0.07) 0.7 (0.11) 0.97 (0.32) 2.85 (0.6) 1.69 (0.15) 1.95 (0.42) 0.97 0.13
10 0.46 (0.03) 1.49 (0.16) 1.06 (0.37) 0.36 (0.08) 0.67 (0.07) 0.94 (0.33) 2.77 (0.62) 1.63 (0.08) 1.28 (0.41) 0.96 0.31
11 0.41 (0.03) 1.18 (0.13) 0.87 (0.19) 0.34 (0.08) 0.85 (0.09) 1.15 (0.26) 2.95 (0.68) 1.61 (0.16) 1.8 (0.38) 0.97 0.48
12 0.47 (0.03) 1.41 (0.17) 1.16 (0.25) 0.4 (0.09) 0.71 (0.09) 0.87 (0.19) 2.52 (0.59) 2.25 (0.11) 2.01 (0.41) 0.93 0.53
13 0.46 (0.03) 1.42 (0.16) 0.83 (0.18) 0.46 (0.11) 0.71 (0.08) 1.21 (0.26) 2.19 (0.55) 2.02 (0.15) 1.53 (0.36) 0.97 0.09
14 0.47 (0.03) 1.39 (0.22) 0.9 (0.16) 0.38 (0.07) 0.72 (0.11) 1.11 (0.2) 2.67 (0.51) 1.93 (0.12) 1.91 (0.44) 0.98 0.58
MEAN 0.43 1.35 1.19 0.34 0.74 1.06 3.10 1.95 1.82 0.95 0.43
MEDIAN 0.43 1.38 0.99 0.34 0.72 1.01 2.90 1.80 1.87 0.04 0.29
Values indicate mean (SE) in days, as estimated by Monolix
∗ calculated as 1/d, half-lives can be calculated as ln(2)/d
 Monocyte circulatory and maturation kinetics

Supplementary Methods

Monocyte dynamics model

We constructed a model of monocyte development in the bone marrow and differentiation
in the periphery. In this model, monocyte progenitors (M) proliferate in the bone marrow at
a rate pm and mature for a fixed time Δ before exiting into the blood as classical monocytes
M1 (CD14++CD16-) at a rate �. Monocytes M1 are lost at a rate dm1 and a fraction α1 of them
mature into intermediate (CD14++CD16+). Intermediate M2 are lost at a rate dm2 and a
fraction α2 eventually mature into non-classical monocytes M3 (CD14+CD16+) with a fixed
delay δ. Non-classical monocytes M3 are lost at a rate dm3.
The biological model is described by the following equations:
5
��
= �� � − ��
��

���
= � ��� − Δ� − ��� ��
��

���
= �� ��� �� − ��� ��
��

���
= �� ��� �� �� − �� − ��� ��
��

where ���� describes the dynamics of monocyte progenitors in the bone marrow.

By assuming that each cell population count is constant over time, the fraction of labeled DNA
in each population denoted LM for the precursor in the bone marrow and L1, L2 and L3 in the
periphery at any given time t is hence given by:

During intake of label:

���
= �� − �� ��
��

After intake of label:

���
= −�� ��
��

91
Chapter 5

���
= ��� �� �� − Δ� − ��� ��
��

���
= �� ��� �� − ��� ��
��

���
= �� ��� �� �� − �� − ��� ��
��

Each parameter (pm, dm1, dm2, dm3, Δ, δ, α1, α2) was modeled as the sum of a population
(fixed) parameter (a) and a random effect (bi) allowing each parameter to be different from
one individual to another: �� = � + �� . Each random effect was assumed to be normally
distributed with a variance to be estimated: �� ~��0, �� �. Hence, for each biological
parameter, two parameters were estimated: one for the average value and one for the
variance of the random effect. Parameters were estimated using the software Monolix®
(version 4.3.3) (1).

In some multi-compartment models, it can be difficult to distinguish between label retained

for a long period in the BM and slow kinetics of circulating cells (2). However, this alternative
with a BM turnover of 1.0 d-1 and CM lifespan of 2.0 d produced a statistically significant
poorer fit (Akaike Information Criteria of -1960 vs -1981).

Supplementary References
1. Lavielle M, Mentre F. Estimation of population pharmacokinetic parameters of saquinavir
in HIV patients with the MONOLIX software. J Pharmacokinet Pharmacodyn. 2007;34(2):229-
249.
2. Li KW, Turner SM, Emson CL, Hellerstein MK, Dale DC. Deuterium and neutrophil kinetics.
Blood. 2011;117(22):6052-6053; author reply 6053-6054.

92
Monocyte circulatory and maturation kinetics

93
 Part II

Kinetics in acute inflammation

 Chapter 6
Immune suppression by neutrophils and
granulocytic myeloid derived suppressor cells:
similarities and differences

Janesh Pillay1,2, Tamar Tak1, Vera M. Kamp1 and Leo Koenderman1

1. Department of Respiratory Medicine, University Medical Center Utrecht, Utrecht,

The Netherlands
2. Department of Anesthesiology, University Medical Center Utrecht,
Utrecht, The Netherlands

Cell Mol Life Sci. 2013 Oct;70(20):3813-27. doi: 10.1007/s00018-013-1286-4.

Epub 2013 Feb 20.
Chapter 6

Abstract
Neutrophils are essential effector cells in the host defense against invading pathogens.
Recently novel neutrophil functions have emerged besides their classical anti-microbial role.
One of these functions is the suppression of T-cell responses. In this respect, neutrophils share
similarities with granulocytic myeloid derived suppressor cells (G-MDSCs). In this review, we
will discuss the similarities and differences between neutrophils and G-MDSCs. Various types
of G-MDSCs have been described, ranging from immature to mature cells shaping the immune
response by different immune suppressive mechanisms. However, all types of G-MDSCs share
distinct features of neutrophils, such as surface markers and morphology. We propose that
G-MDSCs are heterogeneous and represent novel phenotypes of neutrophils, capable of
suppressing the immune response. In this review we will attempt to clarify the differences
and similarities between neutrophils and G-MDSCs and attempt to facilitate further research.

98
 Suppressive neutrophils and G-MDSCs

Introduction
Neutrophils are important effector cells in the innate immune response against invading
micro-organisms (1). The cells possess multiple powerful mechanisms enabling them to
migrate towards, engage with, in particular, small targets and kill them intracellularly (1). The
importance of these cells is illustrated by the fact that neutrophils and/or neutrophil like cells
already developed early in evolution (2). Cells with phagocytic function and neutrophil-
specific proteins are now found in species ranging from simple organisms such as sea fan
corrals (3) to complex organisms such as mammals (4).
The evolution from simple to complex organisms resulted in the origin of the adaptive
immune system. This review will focus on recent data showing the existence of multiple
functional phenotypes of neutrophils that, beyond their well-recognized anti-microbial
functions, are able to steer and shape the adaptive immune system. But before reviewing
these functional phenotypes in detail, it is important to first discuss recent data with respect
to: 1. definitions for priming and phenotypes and 2. the life cycle and compartmentalization
of neutrophils. 6
Switching phenotype and priming: two distinct mechanisms
In this review, we define granulocytic myeloid derived suppressor cells (G-MDSCs) as a
phenotype of neutrophils. A phenotype refers to a cell that either in the bone marrow or by
instruction in the periphery (fig. 1) develops towards a cell with a specialized function, which
distinguishes it from other cells. In the case of G-MDSCs, this would be their ability to suppress
the adaptive immune response. It is only recently that neutrophils are accepted to have
multiple phenotypes and surprisingly, little is known regarding the occurrence and induction
mechanisms of these neutrophil phenotypes. In contrast to induction of phenotypes, priming
can also modulate the functionality of neutrophils. Non-primed neutrophils are relatively
refractory to activation, limiting aspecific activation. This process functions as a safe lock
mechanism and has been extensively reviewed elsewhere (5, 6). Only after priming (typically
by a cytokine, chemokine or bioactive lipid), a neutrophil can optimally exert functions such
as the generation of a respiratory burst induced by fMLF (7), or chemotaxis (8). Priming is a
mechanism distinct from changing of phenotype, as it reversibly potentiates effector
functions of neutrophils but does not change their overall function.

99
Chapter 6

Figure 1. Priming versus functional

phenotypes of neutrophils. The figure
illustrates that phenotypes are defined as
cells that retain specialized functions for a
prolonged time. Priming refers to the
mechanism that is rapidly and reversibly
induced by soluble or cell associated
mediators such as platelet activating factor
(PAF) (6), which potentiate functions of
neutrophils but do not change their overall
function. Priming can potentiate all
different phenotypes and functions, such
as migration, production of ROS and
phagocytosis.

The life cycle of a neutrophil

Despite the consensus regarding the importance of neutrophils in host defense surprisingly
little is known about very basic characteristics of these cells in respect to their life cycle. As
stated above, it is only recently that neutrophils are accepted to have multiple phenotypes. A
possible reason that neutrophil subtypes were overlooked is the view that they are short-
lived cells, which perform their duty and subsequently rapidly go into apoptosis in the tissue.
This view is based on experiments labeling and tracing neutrophils with radioactive isotopes
(9-12). These experiments, which used ex-vivo and potentially toxic labeling techniques
showed a peripheral blood half-life of only 7-25 hours. Our recent paper using in-vivo labeling
with the stable isotope 2H suggests a half-life of 3.9 days (13). This result remains a matter of
debate as Li et al suggested that the observed results could also be explained by a 3.9 day
division time of neutrophil progenitors (14). Moreover, the view that neutrophils in tissue
cannot return to the peripheral blood has been challenged by several studies. Already in 1974,
Vincent et al (15) showed in calves that after disappearance of most labeled neutrophils from
blood, hydrocortisone can induce their return into the circulation, where they stay for at least
another 24 hours. More recently, several studies have provided additional evidence that
support the view that neutrophils do not simply die by apoptosis in the tissues, but move to
additional sites in the body. These studies show homing of neutrophils to secondary lymphoid
tissue (16) and reverse migration of cells over endothelium in vitro and in vivo (17, 18).
Reverse migration and remobilization of neutrophils has also very elegantly been shown in a

100
 Suppressive neutrophils and G-MDSCs

zebra fish larvae demonstrating migration of neutrophils from a site of inflammation toward
different organs throughout the organism (19).
Taken together, these data demonstrate that at least a subpopulation of neutrophils
can survive for much longer than previously appreciated, allowing more time for these cells
to switch phenotypes and exert functions beyond cytotoxicity against invading pathogens.

Myeloid derived suppressor cells

One such phenotype is the recently described myeloid derived suppressor cells (MDSC). These
cells were firstly identified at the beginning of this century and described as immature
myeloid cells that suppress immune responses in the spleens of tumor bearing mice (20-22).
Such immune suppression was earlier attributed to myeloid cells, but this activity was
confined to differentiated cells such as macrophages (23). As research progressed on these
immature myeloid cells, it became clear that they consisted of a heterogeneous group of cells,
consisting of (precursors of) granulocytes and monocytes, and that these cells were not
always immature (24). The term Myeloid derived suppressor cell was coined in 2007 by
Gabrilovic et al. (25) to encompass the heterogeneity of these cells.
6
Considering the granulocytic component of MDSCs, there is still discussion on their
differences and similarities with neutrophils. Recently, research on neutrophils described
various novel neutrophil functions, such as antigen presentation, inhibition of immune
responses and induction of B-cell class switching (26-28). In addition, it has been known for
decades that neutrophils reside in the spleen in health and disease (29), a location frequently
sampled for MDSCs (30-32). As the research fields concerning neutrophils and granulocytic
MDSCs seem to have evolved in separate ways, this review will attempt to clarify the
differences and similarities between these cells and attempt to unify and guide further
research.

Identification of neutrophils and G-MDSCs

G-MDSCs are MDSCs of granulocytic origin. According to this definition, these cells can belong
to one of three different types of granulocytes: neutrophils, eosinophils and basophils.
However, only neutrophils have been described as a component of MDSCs (33, 34). Multiple
surface markers and characteristics that identify G-MDSCs have been described. Before going
into detail about the different G-MDSC characteristics, we will first clearly define how to
identify a neutrophil in order to discuss the similarities and differences with G-MDSCs.

Neutrophil identification
The gold standard to identify a neutrophil is by visual inspection under a light microscope.
When stained with May-Grünwald-Giemsa or similar, neutrophils can be easily distinguished
by the shape of their nucleus and cytoplasmic color/granularity (fig. 2). The nucleus should
either have a band or (hyper)segmented shape and a light pink/purple cytoplasm filled with
similarly colored (“neutrophilic”) granules (fig. 2) (35).

101
Chapter 6

Identification of neutrophils by flow cytometry may be more convenient than visual

inspection, as the latter is a more laborious and subjective method. In mice, flow cytometric
identification of neutrophils can easily be performed by using the neutrophil specific marker
Ly6G (36). Traditionally, Ly6G is combined with CD11b, but this is not necessary when using
the specific Ly6G antibody 1A8 (36).
Human neutrophils lack a marker similar to Ly6G, but can be reliably identified
nonetheless (table 1). In studies on MDSCs, CD11b and CD33 are traditionally used as markers
for human MDSCs. However, these markers are expressed on all cells of the myelocytic
lineage and on NK cells, so they are not specific enough to identify human neutrophils (37-
39). Other markers used are CD14 and CD15. Neutrophils (or G-MDSCs) are found to be
CD14neg/low and CD15pos, whereas monocytes (or Mo-MDSCs) are CD14high and CD15neg/low (34).
Unfortunately, these two markers are not sufficient to identify neutrophils, as eosinophils
have a similar CD15 expression (40). We suggest CD16 as an additional marker, as mature
neutrophils are CD16high, eosinophils are CD16neg and monocytes either CD16neg or CD16int.
Therefore, CD16 allows for distinction between these two types of granulocytes. An additional
advantage of using CD16 is that its expression varies between the different stages of
neutrophil maturation: neutrophil progenitors capable of dividing are CD16neg, with
increasing expressions in metamyelocytes, banded and mature neutrophils, respectively (37).
CD16 alone is not enough to identify neutrophils, since NK cells and monocytes also express
this marker (41).
In short, we suggest the use of Ly6G for identification of murine neutrophils and the
combination of CD14, CD15 and CD16 for identification of human neutrophils. We do want to
emphasize the importance of visual inspection, which remains the gold standard to identify
neutrophils. Visual inspection should routinely be performed in order to eliminate the
possibility of other cell types expressing neutrophil markers under certain clinical conditions.

Table 1. Expression of the markers commonly used to identify human neutrophils or G-

MDSCs.
Neutrophil Eosinophil Monocyte NK cell
CD14 -/+ - ++ -
CD15 ++ ++ -/+ -
CD16 ++ - + ++
CD11b ++ ++ ++ ++
CD33 + + + +

G-MDSCs versus neutrophils

As mentioned above, G-MDSCs have been implicated to have similar surface expression of
CD14, CD15 and CD16 as neutrophils. However, there seems to be one prime feature that
distinguishes them from normal neutrophils: immune suppression. Several methods have
been proposed to distinguish between the suppressive G-MDSCs and circulating neutrophils
and will be discussed below.

102
 Suppressive neutrophils and G-MDSCs

Identification of G-MDSCs – Flow Cytometry

Several papers show differences between G-MDSCs and normal neutrophils in the expression
of cell surface markers visualized by flow cytometry (table 2). Greifenberg et al (42) identified
two subsets of neutrophils with a different CD11b expression in the spleens of healthy mice.
Of these two populations, only the relatively low (but still positive) CD11b expressing cells
were found to be immunosuppressive and, therefore, exhibited characteristics of G-MDSCs.
Youn et al found an increased proportion of neutrophils expressing SLAMF4 (CD244) in mice
bearing several different tumors. In some, but not all, of these tumor models, there was also
an increased population of neutrophils expressing CSF1-R (CD115). When they compared the
CD244 positive and negative populations, only the CD244pos cells were found to be
immunosuppressive (43). The consequences of these findings for the human situation remain
to be established (table 2).

Table 2. Reported differences in flow cytometric markers between suppressive and non-
suppressive neutrophils
6
Murine/
Cell identification Model/Disease Source Ref
human
CD11b↓ banded Murine Healthy mice Spleen Greifenberg (42)
Tumor (CD115: only EL4,
CD244↑ CD115↑ Murine CD244: EL4, LLC, MC38, Spleen Youn (43)
B16F10)
Tumor (non-small cell lung
IL-4Rα↑ Human Blood Mandruzzato (45)
cancer)
CD62L↓ CD11c↑ (CD32↑
CD35↑ CD45↑ CD66b↑) Human Trauma, LPS injection Blood Pillay (28)
hypersegmented
CD11b↑ BAFF↑ CD15↓
CD16↓ CD62L↓↓ CD62P↓
Human Healthy Spleen Puga (27)
CD86↑a CD95↑a CD27↓a
CD40L↓a
a. B-helper Neutrophil subtype 2 only

In humans, an enhanced expression of the IL-4Rα (CD124) was found on suppressive cells.
This marker was found on the G-MDSCs of patients with non-small cell lung carcinoma (44).
However another paper found CD124 expression to correlate only with immunosuppression
by monocyte derived MDSCs (45). Therefore, it remains uncertain whether CD124 can be used
to identify human G-MDSCs.
In severely injured patients and a human acute inflammation model, our group has
identified distinct neutrophil subsets of which the CD62Ldim/CD16bright subset was
immunosuppressive (28). In contrast to the findings by Greifenberg, who showed G-MDSCs
to be lower in CD11b expression, this CD62Ldim/CD16bright subset showed a trend of higher
CD11b expression (42). Other markers upregulated in these suppressive cells were CD11c,

103
Chapter 6

CD32, CD35, CD45 and CD66b. The suppressive cells could, however, not be clearly
distinguished on the basis of these latter markers.
Puga et al. show two different subtypes of neutrophils in the human spleen, named
NBH1 and NBH2 (B-cell helper neutrophils). These subtypes have a higher expression of B-cell
activating factor (BAFF) and CD11b, and lower expressions of CD15, CD16, CD62P and CD62L
compared to blood cells. Additionally, the NBH2 cells have a higher CD27, CD40L, CD86 and
HLA-II compared to both circulating and NBH1 neutrophils (27). Unfortunately, they only
assessed immune suppression by splenic neutrophils as a whole. Therefore, it is unclear
whether only one of these two subtypes or both are suppressive and which markers can
distinguish between suppressive and normal neutrophils.
In conclusion, many markers are shown to distinguish suppressive G-MDSCs from non-
suppressive neutrophils. However, so far none of these findings has been confirmed by other
papers and some findings are contradictory (e.g. CD11b, IL-4Rα). Thus, to date no single or
combined expression of surface markers can reliably identify suppressive neutrophils or G-
MDSCs in either humans or mice.

Identification of G-MDSCs – Density centrifugation

Centrifugation of blood over a layer with a density of 1.077g/ml is a common step in the
isolation of leukocytes from whole blood (46). Due to their relatively high density, neutrophils
end up below the layer, on top of the erythrocyte fraction, whereas the PBMC fraction is
found in the interphase between this layer and the plasma. Schmielau and Finn (47), and
Rodriquez et al (48) found immunosuppressive G-MDSCs in the PBMC fraction of cancer
patients. These cells show an activated phenotype, characterized by increased CD66b and
CD11b expression. Also, they show the immune suppression to be mediated by the CD66b
expressing cells (48). However, they did not show whether the neutrophils with normal
density in the same patients were also suppressive, therefore it remains uncertain whether
density centrifugation can distinguish between suppressive and non-suppressive cells. In vitro
activation of neutrophils from healthy donors resulted in neutrophils with a similar density
and suppressive capabilities, indicating that in this system, G-MDSCs might be activated
neutrophils (48). Density centrifugation remains a widely used method for the isolation of
human MDSCs in cancer patients, but there is still a lack of data on the differences between
these G-MDSC and neutrophils from these patients (49).

Identification of G-MDSCs – Gene profiling

Even though it is not possible to isolate cells based on gene expression patterns, it is likely
that cells with different functions will have different gene expression profiles. Fridlender et al
show differences in the transcriptome of naïve bone marrow neutrophils in healthy mice,
blood G-MDSCs from tumor bearing mice and tumor associated neutrophils (TANs) (50). The
cells from the blood of tumor bearing mice have a low expression of mRNA for cytokines and
chemokines compared to TANs. Compared to bone marrow cells, G-MDSCs show a low mRNA
expression of granule proteins, NADPH complex subunits and peroxidases. Unfortunately, the

104
 Suppressive neutrophils and G-MDSCs

location of neutrophils can influence their functionality (51), so it is unclear whether these
differences were specific for G-MDSCs or a result of different localization/maturation. For
instance, it is likely that neutrophils produce their granule and respiratory burst proteins
during maturation and store them for later use, explaining the high amounts of mRNA for
these proteins (52, 53).
Another transcriptome analysis by Youn et al. compared neutrophils from naïve and
tumor-bearing mice. It showed an upregulation of MPO and proteins involved in cell-cycle
pathways in G-MDSCs from tumor bearing mice. In contrast, neutrophils from naïve mice
show an upregulation in mRNA for cytokines, chemokines, proteases and other pro-
inflammatory proteins (43).
Other proteins found to be upregulated in G-MDSCs are Arginase I (48, 50, 54, 55),
iNOS (56) and IL-10 (54). As these three proteins are directly involved in mechanisms of
immune suppression by G-MDSCs, they will be described in more detail in the section below.

Identification of G-MDSCs– nuclear morphology

MDSCs are in general described as young or immature cells (57). The nuclear morphology of
6
neutrophils provides a simple tool to assess their age. Neutrophils possess a distinct nuclear
morphology in different stages of development (fig. 2). Early progenitors have a round
nucleus, which changes during maturation in the horseshoe, or “banded”, shape of a human
immature neutrophil (a ring-shape in mice). When these cells fully mature, the nucleus starts
showing indentations and is called segmented. When the nucleus has 5 or more segments in
humans, or a cloverleaf-shape in mice, it is called hypersegmented. Since neutrophils gain
more indentations and segments upon maturation, it is tempting to address hypersegmented
cells as “old”. However, there is evidence that segmented and hypersegmented neutrophils
in humans are of similar age (58).
In the paper from Greifenberg et al mentioned above, the G-MDSC population had a
clear ring-shaped morphology, whereas the cells with a segmented nucleus were not
suppressive (42). This supports the notion of G-MDSCs being young/immature cells. Also,
Fridlender et al show in a tumor model that immunosuppressive TANs are mostly immature,
whereas after TGF-β inhibition, the TANs were found to be hypersegmented and did not
suppress tumor growth, thus implying loss of immune suppression [37].
Other papers, however, show no difference in nuclear morphology for the suppressive
cells (43, 59). Similarly, Dumitru et al (34) have extensively reviewed the phenotype of
suppressive G-MDSCs in human cancers and found them to be segmented in 8 out of 9 papers
where the nuclear morphology was assessed (44, 47, 48, 55, 60-63). In addition, in our model
of acute inflammation and in severely injured patients, we have shown only the
hypersegmented cells to be immunosuppressive (28).
Taken together, nuclear morphology is not a good indication for immunosuppressive
functions and, therefore, of G-MDSCs. However, these differences do indicate the existence
of several distinct G-MDSC subtypes.

105
Chapter 6

Figure 2. Schematic representations and images of the nuclear morphology of human and murine neutrophils
during subsequent stages of development. Myelocytes mature into metamyelocytes, banded neutrophils and
finally into mature segmented neutrophils. Neutrophils may also become hypersegmented, with more than 4
nuclear lobes (human) or a cloverleaf shape (mouse). It is unknown whether hypersegmented neutrophils are
more mature than segmented neutrophils.

G-MDSC identification and subtypes – conclusion

When studying potential G-MDSCs (or suppressive neutrophils), one should first ascertain the
cells of interest to be neutrophils. This can be done by flow cytometric determination of CD14,
CD15 and CD16 expression and, ideally, assessing nuclear morphology after cell sorting.
Density centrifugation is not a suitable method for isolating suppressive neutrophils, as it
cannot distinguish suppressive cells from non-suppressive activated cells.
In various studies, different surface markers are shown to distinguish G-MDSCs or
suppressive neutrophils from their non-suppressive counterparts (table 2). However, there
are differences in expression of (activation) markers and nuclear morphology between these
suppressive subsets. This demonstrates that there are several G-MDSC phenotypes, possibly
reflecting differences in localization, clinical condition or origin.

106
 Suppressive neutrophils and G-MDSCs

Mechanisms of immune suppression by suppressive neutrophils and G-MDSCs and their

relevance to disease
Proliferation of T-cells is influenced by many environmental factors. These factors, such as
cytokines, growth factors, and amino acids are easily altered in an inflammatory environment
in the presence of other inflammatory cells such as neutrophils and G-MDSCs. Suppression of
T-cell responses can be achieved by depletion of essential amino acids from the
microenvironment such as L-arginine (64), (massive) generation of reactive oxygen species
(47), or through cell-cell contact (fig. 3) (28).
Production of anti-inflammatory cytokines such as IL-10 by neutrophils has been
proposed (59, 65). However this was only observed in murine neutrophils (66) and will,
therefore, not be discussed in this review.
Recently studies have shown that, in addition to limiting T-cell responses, G-MDSCs
limit NK-cell responses and activation to vaccinia virus (67). This was dependent on H2O2
production by G-MDSCs. Other studies have shown reduced NK-cell responses by G-MDSCs in
pregnancy, cancer and in the tumor environment, however no mechanism of suppression was
reported (68-70).
6

The role of arginase in T-cell suppression by MDSCs

Arginase-1 (ARG1) was shown to be important in the suppression of immune responses by
MDSCs in various murine models (71). ARG1 metabolizes L-arginine into L-ornithine and urea.
This depletes L-arginine from the micro-environment. The amino acid L-arginine has multiple
roles such as its importance in wound healing (72). In addition, it is the only endogenous
substrate for the production of nitric oxide (NO) by inducible nitric oxide synthase (iNOS) (73).
L-arginine is necessary for T-cell proliferation, as in the absence of L-arginine, the cell cycle of
proliferating T-cells arrests in the G0-G1 phase. (64).
Several mechanisms have been described to explain this L-arginine depletion
mediated inhibition of proliferation. L-arginine influences the expression of the T cell receptor
ζ chain (TCRζ, CD247) (74), and ARG-1 has been shown to downregulate TCRζ expression, and
T-cell activation at the level of TCR expression (75, 76). The TCR/CD3 expression is regulated
by continuous internalization and recycling of receptors. The level of surface expression of
the receptor regulates the ability of a T cell to become activated. The rate of synthesis of the
TCRζ-chain is rate limiting to that of the other TCR/CD3 chains. Therefore, this chain is
critically important in the regulation of TCR/CD3 internalization and recycling as it stabilizes
the TCR/CD3 complex on the cell membrane. (77). A second mechanism by which a depletion
of L-arginine results in T-cell suppression has recently been described. Feldemeyer et al show
that dephosphorylation of cofilin is decreased by depletion of L-arginine. Cofilin is a protein
necessary for the remodeling of F-actin (78), which is essential for the formation of an
immunological synapse and T-cell proliferation (79).
ARG1 is widely expressed in murine myeloid cells and macrophages. However, in
humans it has only convincingly been shown in neutrophils (33, 80). Neutrophil ARG1 is
synthesized in their myelocyte and metamyelocyte stage and located in the gelatinase

107
Chapter 6

containing granules of mature neutrophils (52, 80). It is implicated in the host defense against
fungi (80). Activated neutrophils exocytose a form of ARG1 that is catalytically active at pH
9.5-10.5 (52, 81). This ARG1 becomes active at a physiological pH of 7.5 only after cleavage
by a co-factor. The co-factor responsible for this cleavage has not been identified, but it has
been suggested to be located in azurophil neutrophil granules.
Release of ARG1 by neutrophils requires cellular activation and degranulation of both
tertiary (gelatinase) and azurophllic granules. As stated above, human MDSCs have been
shown to co-localize with PBMCs when isolated by density separation. Interestingly fMLF-
activated neutrophils from healthy volunteers show similar behavior and co-localize with
PBMCs (47). In patients suffering from severe traumatic injury, the increase of ARG1 activity
has also been attributed to activated neutrophils in the PBMC fraction (82). In addition,
increased serum ARG1 correlates with degranulated neutrophils in patients with glioblastoma
multiforme (62). These findings could implicate that G-MDSCs in humans that inhibit T-cell
proliferation via an ARG1 mediated mechanism are simply activated granulocytes (48).
As described above, ARG1 expression in myeloid cells of mice and humans is
essentially different (83). Human studies have only correlated the degree and occurrence of
ex vivo measured ARG1 mediated T-cell suppression to disease progression. Murine studies
mostly focused on the association of ex vivo T-cells suppression and occurrence of MDSCs in
the spleen. The direct contribution of MDSCs to in vivo T-cell suppression in T-cell mediated
diseases has remained largely uninvestigated, although recently, ARG1 has been shown to
limit graft vs host disease (GVHD) in mice. In this study ARG1 expressing monocytic MDSC
were generated by ex vivo incubation with G-CSF, GM-CSF and IL-13. Adoptive transfer of
these ARG1 expressing cells or administration of pegylated-ARG1 limited pathology in this
model (71).

Reactive Oxygen Species

A hallmark of neutrophils and G-MDSCs is the potential to produce large amounts of reactive
oxygen species (ROS). These are generated by the NADPH-oxidase complex in neutrophils. A
detailed and schematic description of the generation of ROS is presented by Nathan and Ding
(84). Generation of superoxide anion (O2-) is the first oxygen radical produced. O2- can be
converted to two substances that have been shown to mediate lymphocyte suppression.
Firstly, O2- can react with NO, producing reactive nitrogen species such as peroxynitrite. NO
is generated by inducible nitric oxide synthase (iNOS) using L-arginine as substrate, linking the
generation of reactive nitrogen species to L-arginine metabolism as described above. Reactive
nitrogen species are utilized in some models by monocytic MDSCs but not by G-MDSC and
neutrophils, and will, therefore, not be discussed in this review (57).
The second substance formed from O2- is H2O2 (hydrogenperoxide). H2O2 can be
converted by myeloperoxidase to hypochlorous acid (HOCl-). H2O2 can suppress lymphocyte
proliferation through various mechanisms by inducing apoptosis, decreasing Nf-κB activation,
downregulating TCRζ and oxidation of cofilin (85-87).

108
 Suppressive neutrophils and G-MDSCs

Cofilin remodeling of F-actin is essential for the T-cell effector function. Oxidation of cofilin
results in its loss of Ser3 phosphorylation (86). Dephosphorylated cofilin is unable to mediate
actin depolimerization, thus severely disturbing actin dynamics and impairing T-cell activation
(79). Similar to L-arginine depletion, oxidative stress correlates with TCRζ expression,
although the exact mechanism is not known. In addition, oxidative stress blocks Nf-κB
activation leading to impaired T-cell activation (87).
Of note is that regulatory T-cells have been shown to be resistant to oxidative stress
(88). This suggests that regulatory T-cells are less suppressed than other T-cells, thus
enhancing the overall suppressive effect of H2O2 in vivo.
Suppression of T-cell activation and proliferation requires high concentrations of H2O2
(47, 86), which can be provided by the presence of large numbers of neutrophils at the site of
T-cell activation. This might be due to the fact that hydrogen peroxide is unstable and is
rapidly converted to H2O and O2. Indeed, activated neutrophils or G-MDSCs in cancer patients
have been shown to inhibit T-cell responses in a H2O2 dependent manner (47).
The relevance of H2O2 in the context of G-MDSC or neutrophil mediated suppression
is difficult to study in animal models. This is mainly due to the diverse biological functions of
6
H2O2. Besides immune suppression, H2O2 and its metabolites are involved in bacterial killing
(89). In addition it functions as a signaling molecule necessary for diverse cellular functions
(89) including chemotaxis of immune cells. It has recently been shown that H2O2 is a potent
inducer of chemotaxis of neutrophil-like immune cells in a model of tissue injury in zebrafish
(90). Hydrogen peroxide might, therefore, also indirectly contribute to microbial clearance by
attracting immune cells and killing bacteria. These functions of H2O2 are indispensable in
immune processes and, therefore, complicate the interpretation of studies targeting H2O2 to
define its role in immune suppression by G-MDSCs.
Caution must be taken in interpreting ex vivo suppression of T-cell proliferation
mediated by H2O2. Manipulation and isolation of neutrophils and G-MDSCs might lead to cell
priming and aberrant activation. Also, adhesion to plastic culture dishes might result in
cellular activation, degranulation and reactive oxygen species production resulting in vitro
suppression of T-cell responses (91). Activation of large number neutrophils from healthy
volunteers has been shown to suppress T-cell responses ex vivo (86). Therefore, at least two
possibilities exist on how H2O2 results in immune suppression in vivo. Firstly, a general
oxidative environment described by Klemke et al. in which ‘normal’ activated neutrophils
mediate immune suppression. Secondly as described below, low concentrations of H2O2 can
be delivered via the formation of an immunological synapse providing specific and direct
suppression of T-cell responses. It would be useful to distinguish between these two
mechanisms in future studies concerning G-MDSCs and neutrophil suppression by H2O2.

109
Chapter 6

Figure 3. Mechanisms of
suppression by G-MDSC and
suppressive neutrophils.
Suppression can be mediated
by extracellular arginase,
extracellular ROS, or ROS in
an immunological synapse. Al
these mechanisms result in
reduced T-cell proliferation,
via decreases in extracellular
L-arginine, cofilin, TCRζ
expression, NF-κB activation
or unknown mechanisms.

Immunological synapse formation, the requirement of cell-to-cell contact

The potency of the above-described suppressive mechanisms would be greatly enhanced by
cell-to-cell contact and the formation of an immunological synapse. H2O2 has a short half-life
and can be degraded by many endogenous anti-oxidants. Therefore, release into a synapse
would potentiate and concentrate local concentrations of H2O2, H2O2 is produced in an
immunological synapse between T-cells and macrophages and dendritic cells during antigen
presentation, and results in decreased lymphocyte activation (92, 93). We have recently
shown that a subset of neutrophils in human inflammation is capable of directly delivering
H2O2 to the surface of lymphocytes and thereby limiting T-cell activation and proliferation
(28). This contact was dependent on CD11b/CD18, an integrin abundantly expressed by G-
MDSCs in mice. However in mice, no requirement of cell-to-cell contact suppression by G-
MDSCs was found. A very recent study showed that in patients with gastric cancer, G-MDSCs
isolated from the tumor site, suppressed T-cells in a contact dependent manner (94).
Regretfully, no experiments were performed in this latter study to further elucidate the
suppressive mechanism.

Distribution of neutrophils and G-MDSC’s in lymphoid organs.

In order to modulate the function and proliferation of T-cells, neutrophils or G-MDSCs need
to come in contact with or in close proximity to T-cells (95). T-cell proliferation is normally
considered to take place in secondary lymphoid organs such as lymph nodes and the spleen
(96). Recently, T-cell proliferation has also been shown at the site of inflammation (97, 98). In
order to suppress these T-cells, neutrophils will have to be present at these sites. Indeed,

110
 Suppressive neutrophils and G-MDSCs

many studies show neutrophil homing to sites of T-cell proliferation, which will be reviewed
in the following section.

Neutrophils in lymphoid organs.

Spleen: Neutrophils are known to migrate to the spleen both under homeostatic and
pathological conditions (29). Reinfusion of ex-vivo 111Indium labeled neutrophils in healthy
controls showed the majority of label in the bone marrow, spleen and liver (29, 99). These
studies imply that considerable amounts of neutrophils rapidly home to the spleen after
release from the bone marrow. In addition, in mice about 10% of reinfused radiolabeled
neutrophils migrated towards the spleen, which was not influenced by maturation status of
neutrophils or inflammation (100). It is important to emphasize that ex vivo manipulation of
the cells could induced subtle changes affecting their homing behavior in vivo (101).
In the spleen, neutrophils reside on the border of the red and white pulp (102, 103)
and the marginal zone under normal homeostatic conditions, whereas T-cells are found in the
white pulp during homeostasis (102). Consequently neutrophils should migrate to the white
pulp in order to contact the T-cells or vice versa. Neutrophil migration to the white pulp has
6
been shown after intraperitoneal injection of LPS in mice. This was shown to be CD14
dependent (103). Also, after surgical trauma neutrophils were found to co-localize with T-cells
in the spleen (76). These data demonstrate that neutrophils migrate towards the T-cell zones
of the spleen in acute systemic inflammation.
Lymph nodes: During inflammation, neutrophils are found to migrate to lymph nodes
(16, 104-110). Already in 1987, neutrophil trafficking from lung to draining lymph nodes was
described in dogs (1). In this study fluorescent microspheres were instilled in the lung of dogs
and phagocytosed by neutrophils and macrophages. After 40 hours, almost half of the cells in
the draining lymph node were neutrophils containing microspheres (104). Also, in a more
physiological model of antigen uptake (110) neutrophils can migrate to draining lymph nodes
(16, 105). Neutrophils were detected in lymph nodes during infections with mycobacterium
bovis (106), Salmonella (107), and different parasites (108-110). In some of these models,
neutrophils were shown to alter (16, 110) or even inhibit the inflammatory response (105,
109). The route of migration towards (106-110) the lymph nodes was via the lymphatic system
(16, 104-106, 108).

Suppressive neutrophils and G-MDSC in the spleen

Almost all studies regarding G-MDSCs in literature were performed with Ly6G positive cells
isolated from the spleen (30-32). However, not all of these Ly6G positive neutrophils in the
spleen can suppress T-cells (42). Influx of G-MDSCs into the spleen in mice has been seen both
in acute and chronic inflammation such as cancer models (30), parasite infection
(Trypanosoma cruzi) (31), and superantigen stimulation (staphylococcal enterotoxin) (32).
Numbers of G-MDSCs were increased up to 10-fold 14 days after Trypanosoma cruzi infection
(31). During superantigen stimulation, suppressive neutrophils with highly segmented nuclei

111
Chapter 6

were sorted from the spleen (32), these cells bear a resemblance to the hypersegmented
CD16bright/CD62Ldim neutrophils that are found in the blood after LPS challenge (28).
Some cancer models increase hematopoiesis, resulting in increased cycling of
hematopoietic stem cells and hematopoietic activity in the spleen (111). Younos et al show
by in vivo BrdU labeling that in tumor bearing mice granulocytic proliferation mainly takes
place in the spleen, whereas, in control mice granulocytic cells predominantly proliferate in
the bone marrow (112). The CD3+ cells in this model proliferate less in the tumor bearing
mice, but unfortunately, they do not show that this immune suppression is a direct effect of
the spleen granulocytes. There were also no microscopic pictures of these cells, to show their
maturation stage (112).

Suppressive neutrophils and G-MDSCs in the lymph nodes

Less data is available to show suppressive neutrophils or G-MDSCs in lymph nodes. Sepsis
induced an influx of immature myelocytes capable of T-cell suppression in lymph nodes. These
cells could be detected 10-14 days after sepsis and remained present in the lymph nodes for
at least 12 weeks after sepsis. Cytospins obtained during this study showed a heterogeneous
group of cells consisting of both monocytic and granulocytic origin (113).
Vascular Endothelial Growth Factor (VEGF) is able to induce MDSCs in cancer models
and is a factor important for immune evasion in several cancer models (114). Upon infusion
of VEGF, myeloid cells, including neutrophils, were massively increased in lymph nodes (115).
Unfortunately, the capacity of these granulocytes to suppress T-cells was not tested. Another
indication that MDSCs can migrate to lymph nodes came from a study of Watanabe et al (116).
They showed that proliferation of T-cells in the lymph nodes of leukocyte depleted mice was
low when injected with spleen cells (containing both T-cells and MDSCs) from tumor bearing
mice, compared to proliferation after injection with control mice spleen cells (116).
Proliferation was measured in vitro using cells isolated from lymph nodes. Unfortunately this
model did not discriminate between granulocytic and monocytic MDSCs, so further research
is necessary to draw definite conclusions about the presence and importance of suppressive
neutrophils in lymph nodes.

T-cell proliferation outside the lymphoid organs

T-cell proliferation is not restricted to lymphoid organs, because T cell proliferation was also
found e.g. at sites of viral infection (97, 98, 117-120). In influenza infection, proliferating T-
cells in the lungs contribute substantially to the total number of cytotoxic T-cells in the lung
(97, 117). Also, the persistence and reactivation of influenza-specific CD8+ memory T-cells can
take place in mice without secondary lymphoid organs (118). Similarly, CD8+ T-cells
proliferate outside the secondary lymphoid organs in a model of Herpes simplex virus (HSV)
reactivation. In this model, infected sensory dorsal root ganglia (DRGs) are transplanted into
naïve mice, inducing proliferation in the DRGs of both memory CD8+ T-cells from graft (98)
and newly recruited CD8+ T-cells from the host (119). Even further, in RSV infected mice, CD4+
memory T-cells proliferate and differentiate in the lung, but not in the lymph nodes (120).

112
 Suppressive neutrophils and G-MDSCs

Taken together, this shows that T-cells can proliferate at sites of viral infection, which is
exactly the place where vast amounts of neutrophils are found (121, 122). Therefore,
although it may contribute, neutrophil migration towards the secondary lymphoid organs is
not necessary to dampen the immune response.

Origin of G-MDSCs and suppressive neutrophils

Many papers show only a subset of neutrophils to be suppressive. Even further, these
suppressive subsets show differences in (flow cytometric) expression patterns and nuclear
morphology (27, 28, 43-45). The difference between normal neutrophils and the different
types of suppressive neutrophils may lie in the presence of cytokines or growth factors, (e.g.
G-CSF and VEGF) (114, 115, 123) in localization or in their origin (27). Few studies have
addressed the origin of suppressive phenotypes, therefore we will briefly discuss four
hypotheses regarding the origin(s) of these suppressive cells (fig. 4):

1. Suppressive neutrophils might originate from normal, fully maturated cells. These
cells acquire a suppressive phenotype under certain (inflammatory) conditions. They
6
can either retain their mature nuclear morphology (1m) or become hypersegmented
(1h).
2. Cells do not fully mature before exiting from the bone marrow. Progenitors have been
found in the peripheral blood under conditions of severe systemic inflammation
caused by infection or trauma (124, 125). These cells are neutrophil progenitors, which
migrate to the tissue and subsequently become suppressive.
3. An altered or a dedicated suppressive granulopoiesis, underlie the production of G-
MDSCs, as suggested by the role of G-CSF in several papers (71, 123). This results in
either immature (3i) or mature (3m) cells with a suppressive phenotype.
4. Instead of being produced in the bone marrow, suppressive cells might be produced
by extramedullary granulopoiesis. This would result in either immature (4i) or mature
(4m) cells with a suppressive phenotype. For example, Youn et al. (43) described G-
MDSCs from tumor-bearing mice were produced in the spleen, whereas neutrophils
from healthy mice originated from the bone marrow.

At this moment it is unclear which of these mechanisms underlie the induction of G-MDSC’s
and whether multiple mechanisms co-exist. Further research is required to elucidate the
origin of different suppressive phenotypes, and whether differences between suppressive
phenotypes are caused by differences in their origin or by alternative activation.

113
Chapter 6

Figure 4. The origin of G-MDSCs remains unknown. Hypothetically, these calls can arise from mature (1) or
immature (2) neutrophils receiving signals to become suppressive. Alternatively, there may be a dedicated
granulopoiesis, which only produces suppressive cells. This granulopoiesis can take place either in the bone
marrow (3) or extramedullary (4). Additionally, these cells can be immature (i), mature (m) or hypersegmented
(h).

Conclusion – A novel hypothesis: G-MDSCs are a phenotype of neutrophils

Neutrophils do not belong to a single homogenous population of cytotoxic cells with a sole
function to eliminate invading microorganisms. In fact, these cells can engage with and
modulate T-cells and, thereby, shape the adaptive immune system. The lack of consensus
regarding nomenclature of these suppressive cells, their heterogeneity, and the lack of
suppressive assays in many studies makes it difficult to draw overall conclusions. However,
these studies support the hypothesis that multiple types of suppressive neutrophils exist,
capable of mediating immune suppression by different mechanisms. Given the recent
advances in neutrophil biology, illustrating their plasticity, we hypothesize that G-MDSCs
might be a functional heterogenic subset of neutrophils. At this moment in time, it is
uncertain how many neutrophil phenotypes exist. It is, however, clear that targeting
neutrophils or G-MDSCs as clinical intervention is only effective with knowledge of the
different pro- and anti-inflammatory phenotypes, and when origin and kinetics of these cells
are adequately elucidated.

114
 Suppressive neutrophils and G-MDSCs

References
1. Nathan C. Neutrophils and immunity: challenges and opportunities. Nat Rev Immunol.
2006;6(3):173-82.
2. Patterson MJ, and Landolt ML. Cellular reaction to injury in the anthozoan
Anthopleura elegantissima. J Invertebr Pathol. 1979;33(2):189-96.
3. Mydlarz LD, Holthouse SF, Peters EC, and Harvell CD. Cellular responses in sea fan
corals: granular amoebocytes react to pathogen and climate stressors. PLoS One.
2008;3(3):e1811.
4. Cavaillon JM. The historical milestones in the understanding of leukocyte biology
initiated by Elie Metchnikoff. J Leukoc Biol. 2011;90(3):413-24.
5. Condliffe AM, Kitchen E, and Chilvers ER. Neutrophil priming: pathophysiological
consequences and underlying mechanisms. Clin Sci (Lond). 1998;94(5):461-71.
6. Coffer PJ, and Koenderman L. Granulocyte signal transduction and priming: cause
without effect? Immunol Lett. 1997;57(1-3):27-31.
7. Koenderman L, Yazdanbakhsh M, Roos D, and Verhoeven AJ. Dual mechanisms in
priming of the chemoattractant-induced respiratory burst in human granulocytes. A
Ca2+-dependent and a Ca2+-independent route. J Immunol. 1989;142(2):623-8. 6
8. Hakansson L, and Venge P. Priming of eosinophil and neutrophil migratory responses
by interleukin 3 and interleukin 5. Apmis. 1994;102(4):308-16.
9. Dancey JT, Deubelbeiss KA, Harker LA, and Finch CA. Neutrophil kinetics in man. J Clin
Invest. 1976;58(3):705-15.
10. Cartwright GE, Athens JW, and Wintrobe MM. The Kinetics of Granulopoiesis in
Normal Man. Blood. 1964;24(780-803.
11. Dresch C, Najean Y, and Bauchet J. Kinetic studies of 51Cr and DF32P labelled
granulocytes. Br J Haematol. 1975;29(1):67-80.
12. Steinbach KH, Schick P, Trepel F, Raffler H, Dohrmann J, Heilgeist G, Heltzel W, Li K,
Past W, van der Woerd-de Lange JA, et al. Estimation of kinetic parameters of
neutrophilic, eosinophilic, and basophilic granulocytes in human blood. Blut.
1979;39(1):27-38.
13. Pillay J, den Braber I, Vrisekoop N, Kwast LM, de Boer RJ, Borghans JA, Tesselaar K, and
Koenderman L. In vivo labeling with 2H2O reveals a human neutrophil lifespan of 5.4
days. Blood. 2010;116(4):625-7.
14. Li KW, Turner SM, Emson CL, Hellerstein MK, and Dale DC. Deuterium and neutrophil
kinetics. Blood. 2011;117(22):6052-3; author reply 3-4.
15. Vincent PC, Chanana AD, Cronkite EP, and Joel DD. The intravascular survival of
neutrophils labeled in vivo. Blood. 1974;43(3):371-7.
16. Maletto BA, Ropolo AS, Alignani DO, Liscovsky MV, Ranocchia RP, Moron VG, and
Pistoresi-Palencia MC. Presence of neutrophil-bearing antigen in lymphoid organs of
immune mice. Blood. 2006;108(9):3094-102.
17. Woodfin A, Voisin MB, Beyrau M, Colom B, Caille D, Diapouli FM, Nash GB, Chavakis
T, Albelda SM, Rainger GE, et al. The junctional adhesion molecule JAM-C regulates
polarized transendothelial migration of neutrophils in vivo. Nat Immunol.
2011;12(8):761-9.
18. Buckley CD, Ross EA, McGettrick HM, Osborne CE, Haworth O, Schmutz C, Stone PC,
Salmon M, Matharu NM, Vohra RK, et al. Identification of a phenotypically and
functionally distinct population of long-lived neutrophils in a model of reverse
endothelial migration. J Leukoc Biol. 2006;79(2):303-11.

115
Chapter 6

19. Yoo SK, and Huttenlocher A. Spatiotemporal photolabeling of neutrophil trafficking

during inflammation in live zebrafish. J Leukoc Biol. 2011;89(5):661-7.
20. Kusmartsev SA, Li Y, and Chen SH. Gr-1+ myeloid cells derived from tumor-bearing
mice inhibit primary T cell activation induced through CD3/CD28 costimulation. J
Immunol. 2000;165(2):779-85.
21. Almand B, Clark JI, Nikitina E, van Beynen J, English NR, Knight SC, Carbone DP, and
Gabrilovich DI. Increased production of immature myeloid cells in cancer patients: a
mechanism of immunosuppression in cancer. J Immunol. 2001;166(1):678-89.
22. Gabrilovich DI, Velders MP, Sotomayor EM, and Kast WM. Mechanism of immune
dysfunction in cancer mediated by immature Gr-1+ myeloid cells. J Immunol.
2001;166(9):5398-406.
23. Otsuji M, Kimura Y, Aoe T, Okamoto Y, and Saito T. Oxidative stress by tumor-derived
macrophages suppresses the expression of CD3 zeta chain of T-cell receptor complex
and antigen-specific T-cell responses. Proc Natl Acad Sci U S A. 1996;93(23):13119-24.
24. Peranzoni E, Zilio S, Marigo I, Dolcetti L, Zanovello P, Mandruzzato S, and Bronte V.
Myeloid-derived suppressor cell heterogeneity and subset definition. Curr Opin
Immunol. 2010;22(2):238-44.
25. Gabrilovich DI, Bronte V, Chen SH, Colombo MP, Ochoa A, Ostrand-Rosenberg S, and
Schreiber H. The terminology issue for myeloid-derived suppressor cells. Cancer Res.
2007;67(1):425; author reply 6.
26. Ostanin DV, Kurmaeva E, Furr K, Bao R, Hoffman J, Berney S, and Grisham MB.
Acquisition of antigen-presenting functions by neutrophils isolated from mice with
chronic colitis. J Immunol. 2012;188(3):1491-502.
27. Puga I, Cols M, Barra CM, He B, Cassis L, Gentile M, Comerma L, Chorny A, Shan M, Xu
W, et al. B cell-helper neutrophils stimulate the diversification and production of
immunoglobulin in the marginal zone of the spleen. Nat Immunol. 2012;13(2):170-80.
28. Pillay J, Kamp VM, van Hoffen E, Visser T, Tak T, Lammers JW, Ulfman LH, Leenen LP,
Pickkers P, and Koenderman L. A subset of neutrophils in human systemic
inflammation inhibits T cell responses through Mac-1. J Clin Invest. 2012;122(1):327-
36.
29. Saverymuttu SH, Peters AM, Keshavarzian A, Reavy HJ, and Lavender JP. The kinetics
of 111indium distribution following injection of 111indium labelled autologous
granulocytes in man. Br J Haematol. 1985;61(4):675-85.
30. Movahedi K, Guilliams M, Van den Bossche J, Van den Bergh R, Gysemans C, Beschin
A, De Baetselier P, and Van Ginderachter JA. Identification of discrete tumor-induced
myeloid-derived suppressor cell subpopulations with distinct T cell-suppressive
activity. Blood. 2008;111(8):4233-44.
31. Goni O, Alcaide P, and Fresno M. Immunosuppression during acute Trypanosoma cruzi
infection: involvement of Ly6G (Gr1(+))CD11b(+ )immature myeloid suppressor cells.
Int Immunol. 2002;14(10):1125-34.
32. Cauley LS, Miller EE, Yen M, and Swain SL. Superantigen-induced CD4 T cell tolerance
mediated by myeloid cells and IFN-gamma. J Immunol. 2000;165(11):6056-66.
33. Luckner-Minden C, Fischer I, Langhans CD, Schiller M, Kropf P, Muller I, Hohlfeld JM,
Ho AD, and Munder M. Human eosinophil granulocytes do not express the enzyme
arginase. J Leukoc Biol. 2010;87(6):1125-32.

116
 Suppressive neutrophils and G-MDSCs

34. Dumitru CA, Moses K, Trellakis S, Lang S, and Brandau S. Neutrophils and granulocytic
myeloid-derived suppressor cells: immunophenotyping, cell biology and clinical
relevance in human oncology. Cancer Immunol Immunother. 2012;61(8):1155-67.
35. L S. Leitfaden der Blutmorphologie. Thieme; 1965.
36. Daley JM, Thomay AA, Connolly MD, Reichner JS, and Albina JE. Use of Ly6G-specific
monoclonal antibody to deplete neutrophils in mice. J Leukoc Biol. 2008;83(1):64-70.
37. Elghetany MT. Surface antigen changes during normal neutrophilic development: a
critical review. Blood Cells Mol Dis. 2002;28(2):260-74.
38. Hernandez-Caselles T, Martinez-Esparza M, Perez-Oliva AB, Quintanilla-Cecconi AM,
Garcia-Alonso A, Alvarez-Lopez DM, and Garcia-Penarrubia P. A study of CD33 (SIGLEC-
3) antigen expression and function on activated human T and NK cells: two isoforms
of CD33 are generated by alternative splicing. J Leukoc Biol. 2006;79(1):46-58.
39. Griffin JD, Linch D, Sabbath K, Larcom P, and Schlossman SF. A monoclonal antibody
reactive with normal and leukemic human myeloid progenitor cells. Leuk Res.
1984;8(4):521-34.
40. Spitalnik PF, Spitalnik SL, Danley JM, Lopez AF, Vadas MA, Civin CI, and Ginsburg V.
Carbohydrate-specific monoclonal antibodies bind to human granulocytes and
stimulate antibody-dependent cellular cytotoxicity. Arch Biochem Biophys. 6
1989;271(1):168-76.
41. Ruiz-Cabello F, Lopez Nevot MA, Garrido A, and Garrido F. A study of GRM1
monoclonal antibody that reacts with natural killer cells and granulocytes. Nat Immun
Cell Growth Regul. 1987;6(2):99-108.
42. Greifenberg V, Ribechini E, Rossner S, and Lutz MB. Myeloid-derived suppressor cell
activation by combined LPS and IFN-gamma treatment impairs DC development. Eur J
Immunol. 2009;39(10):2865-76.
43. Youn JI, Collazo M, Shalova IN, Biswas SK, and Gabrilovich DI. Characterization of the
nature of granulocytic myeloid-derived suppressor cells in tumor-bearing mice. J
Leukoc Biol. 2012;91(1):167-81.
44. Liu CY, Wang YM, Wang CL, Feng PH, Ko HW, Liu YH, Wu YC, Chu Y, Chung FT, Kuo CH,
et al. Population alterations of L-arginase- and inducible nitric oxide synthase-
expressed CD11b+/CD14(-)/CD15+/CD33+ myeloid-derived suppressor cells and CD8+
T lymphocytes in patients with advanced-stage non-small cell lung cancer. J Cancer
Res Clin Oncol. 2010;136(1):35-45.
45. Mandruzzato S, Solito S, Falisi E, Francescato S, Chiarion-Sileni V, Mocellin S, Zanon A,
Rossi CR, Nitti D, Bronte V, et al. IL4Ralpha+ myeloid-derived suppressor cell expansion
in cancer patients. J Immunol. 2009;182(10):6562-8.
46. Boyum A. Isolation of mononuclear cells and granulocytes from human blood.
Isolation of monuclear cells by one centrifugation, and of granulocytes by combining
centrifugation and sedimentation at 1 g. Scand J Clin Lab Invest Suppl. 1968;97(77-89.
47. Schmielau J, and Finn OJ. Activated granulocytes and granulocyte-derived hydrogen
peroxide are the underlying mechanism of suppression of t-cell function in advanced
cancer patients. Cancer Res. 2001;61(12):4756-60.
48. Rodriguez PC, Ernstoff MS, Hernandez C, Atkins M, Zabaleta J, Sierra R, and Ochoa AC.
Arginase I-producing myeloid-derived suppressor cells in renal cell carcinoma are a
subpopulation of activated granulocytes. Cancer Res. 2009;69(4):1553-60.
49. Poschke I, and Kiessling R. On the armament and appearances of human myeloid-
derived suppressor cells. Clin Immunol. 2012;144(3):250-68.

117
Chapter 6

50. Fridlender ZG, Sun J, Mishalian I, Singhal S, Cheng G, Kapoor V, Horng W, Fridlender G,
Bayuh R, Worthen GS, et al. Transcriptomic analysis comparing tumor-associated
neutrophils with granulocytic myeloid-derived suppressor cells and normal
neutrophils. PLoS One. 2012;7(2):e31524.
51. Boxio R, Bossenmeyer-Pourie C, Steinckwich N, Dournon C, and Nusse O. Mouse bone
marrow contains large numbers of functionally competent neutrophils. J Leukoc Biol.
2004;75(4):604-11.
52. Jacobsen LC, Theilgaard-Monch K, Christensen EI, and Borregaard N. Arginase 1 is
expressed in myelocytes/metamyelocytes and localized in gelatinase granules of
human neutrophils. Blood. 2007;109(7):3084-7.
53. Borregaard N, Theilgaard-Monch K, Sorensen OE, and Cowland JB. Regulation of
human neutrophil granule protein expression. Curr Opin Hematol. 2001;8(1):23-7.
54. Goulart MR, Pluhar GE, and Ohlfest JR. Identification of myeloid derived suppressor
cells in dogs with naturally occurring cancer. PLoS One. 2012;7(3):e33274.
55. Zea AH, Rodriguez PC, Atkins MB, Hernandez C, Signoretti S, Zabaleta J, McDermott D,
Quiceno D, Youmans A, O'Neill A, et al. Arginase-producing myeloid suppressor cells
in renal cell carcinoma patients: a mechanism of tumor evasion. Cancer Res.
2005;65(8):3044-8.
56. Liu Y, Peng SY, Wang L, and Wang WJ. [Effects of extracts of cheezheng pain relieving
plaster on nitric oxide and iNOS expression in macrophages induced by
lipopolysaccharides]. Yao Xue Xue Bao. 2009;44(8):863-7.
57. Gabrilovich DI, and Nagaraj S. Myeloid-derived suppressor cells as regulators of the
immune system. Nat Rev Immunol. 2009;9(3):162-74.
58. Fliedner TM, Cronkite EP, Killmann SA, and Bond VP. Granulocytopoiesis. Ii.
Emergence and Pattern of Labeling of Neutrophilic Granulocytes in Humans. Blood.
1964;24(683-700.
59. Noel G, Wang Q, Schwemberger S, Hanson C, Giacalone N, Haar L, and Ogle CK.
Neutrophils, not monocyte/macrophages, are the major splenic source of postburn IL-
10. Shock. 2011;36(2):149-55.
60. Porembka MR, Mitchem JB, Belt BA, Hsieh CS, Lee HM, Herndon J, Gillanders WE,
Linehan DC, and Goedegebuure P. Pancreatic adenocarcinoma induces bone marrow
mobilization of myeloid-derived suppressor cells which promote primary tumor
growth. Cancer Immunol Immunother. 2012;61(9):1373-85.
61. Srivastava MK, Bosch JJ, Thompson JA, Ksander BR, Edelman MJ, and Ostrand-
Rosenberg S. Lung cancer patients' CD4(+) T cells are activated in vitro by MHC II cell-
based vaccines despite the presence of myeloid-derived suppressor cells. Cancer
Immunol Immunother. 2008;57(10):1493-504.
62. Sippel TR, White J, Nag K, Tsvankin V, Klaassen M, Kleinschmidt-DeMasters BK, and
Waziri A. Neutrophil degranulation and immunosuppression in patients with GBM:
restoration of cellular immune function by targeting arginase I. Clin Cancer Res.
2011;17(22):6992-7002.
63. Brandau S, Trellakis S, Bruderek K, Schmaltz D, Steller G, Elian M, Suttmann H, Schenck
M, Welling J, Zabel P, et al. Myeloid-derived suppressor cells in the peripheral blood
of cancer patients contain a subset of immature neutrophils with impaired migratory
properties. J Leukoc Biol. 2011;89(2):311-7.
64. Rodriguez PC, Quiceno DG, and Ochoa AC. L-arginine availability regulates T-
lymphocyte cell-cycle progression. Blood. 2007;109(4):1568-73.

118
 Suppressive neutrophils and G-MDSCs

65. De Santo C, Arscott R, Booth S, Karydis I, Jones M, Asher R, Salio M, Middleton M, and
Cerundolo V. Invariant NKT cells modulate the suppressive activity of IL-10-secreting
neutrophils differentiated with serum amyloid A. Nat Immunol. 2010;11(11):1039-46.
66. Davey MS, Tamassia N, Rossato M, Bazzoni F, Calzetti F, Bruderek K, Sironi M, Zimmer
L, Bottazzi B, Mantovani A, et al. Failure to detect production of IL-10 by activated
human neutrophils. Nat Immunol. 2011;12(11):1017-8; author reply 8-20.
67. Fortin C, Huang X, and Yang Y. NK cell response to vaccinia virus is regulated by
myeloid-derived suppressor cells. J Immunol. 2012;189(4):1843-9.
68. Sceneay J, Chow MT, Chen A, Halse HM, Wong CS, Andrews DM, Sloan EK, Parker BS,
Bowtell DD, Smyth MJ, et al. Primary Tumor Hypoxia Recruits
CD11b+/Ly6Cmed/Ly6G+ Immune Suppressor Cells and Compromises NK Cell
Cytotoxicity in the Premetastatic Niche. Cancer Res. 2012;72(16):3906-11.
69. Mauti LA, Le Bitoux MA, Baumer K, Stehle JC, Golshayan D, Provero P, and
Stamenkovic I. Myeloid-derived suppressor cells are implicated in regulating
permissiveness for tumor metastasis during mouse gestation. J Clin Invest.
2011;121(7):2794-807.
70. Elkabets M, Ribeiro VS, Dinarello CA, Ostrand-Rosenberg S, Di Santo JP, Apte RN, and
Vosshenrich CA. IL-1beta regulates a novel myeloid-derived suppressor cell subset 6
that impairs NK cell development and function. Eur J Immunol. 2010;40(12):3347-57.
71. Highfill SL, Rodriguez PC, Zhou Q, Goetz CA, Koehn BH, Veenstra R, Taylor PA,
Panoskaltsis-Mortari A, Serody JS, Munn DH, et al. Bone marrow myeloid-derived
suppressor cells (MDSCs) inhibit graft-versus-host disease (GVHD) via an arginase-1-
dependent mechanism that is up-regulated by interleukin-13. Blood.
2010;116(25):5738-47.
72. Wittmann F, Prix N, Mayr S, Angele P, Wichmann MW, van den Engel NK, Hernandez-
Richter T, Chaudry IH, Jauch KW, and Angele MK. L-arginine improves wound healing
after trauma-hemorrhage by increasing collagen synthesis. J Trauma. 2005;59(1):162-
8.
73. Morris SM, Jr. Arginine: master and commander in innate immune responses. Sci
Signal. 2010;3(135):pe27.
74. Rodriguez PC, Zea AH, Culotta KS, Zabaleta J, Ochoa JB, and Ochoa AC. Regulation of T
cell receptor CD3zeta chain expression by L-arginine. J Biol Chem.
2002;277(24):21123-9.
75. Yachimovich-Cohen N, Even-Ram S, Shufaro Y, Rachmilewitz J, and Reubinoff B.
Human embryonic stem cells suppress T cell responses via arginase I-dependent
mechanism. J Immunol. 2010;184(3):1300-8.
76. Makarenkova VP, Bansal V, Matta BM, Perez LA, and Ochoa JB. CD11b+/Gr-1+ myeloid
suppressor cells cause T cell dysfunction after traumatic stress. J Immunol.
2006;176(4):2085-94.
77. D'Oro U, Munitic I, Chacko G, Karpova T, McNally J, and Ashwell JD. Regulation of
constitutive TCR internalization by the zeta-chain. J Immunol. 2002;169(11):6269-78.
78. Feldmeyer N, Wabnitz G, Leicht S, Luckner-Minden C, Schiller M, Franz T, Conradi R,
Kropf P, Muller I, Ho AD, et al. Arginine deficiency leads to impaired cofilin
dephosphorylation in activated human T lymphocytes. Int Immunol. 2012;24(5):303-
13.
79. Burkhardt JK, Carrizosa E, and Shaffer MH. The actin cytoskeleton in T cell activation.
Annu Rev Immunol. 2008;26(233-59.

119
Chapter 6

80. Munder M, Mollinedo F, Calafat J, Canchado J, Gil-Lamaignere C, Fuentes JM, Luckner

C, Doschko G, Soler G, Eichmann K, et al. Arginase I is constitutively expressed in
human granulocytes and participates in fungicidal activity. Blood. 2005;105(6):2549-
56.
81. Rotondo R, Bertolotto M, Barisione G, Astigiano S, Mandruzzato S, Ottonello L, Dallegri
F, Bronte V, Ferrini S, and Barbieri O. Exocytosis of azurophil and arginase 1-containing
granules by activated polymorphonuclear neutrophils is required to inhibit T
lymphocyte proliferation. J Leukoc Biol. 2011;89(5):721-7.
82. Bryk JA, Popovic PJ, Zenati MS, Munera V, Pribis JP, and Ochoa JB. Nature of myeloid
cells expressing arginase 1 in peripheral blood after trauma. J Trauma. 2010;68(4):843-
52.
83. Ochoa AC, Zea AH, Hernandez C, and Rodriguez PC. Arginase, prostaglandins, and
myeloid-derived suppressor cells in renal cell carcinoma. Clin Cancer Res. 2007;13(2 Pt
2):721s-6s.
84. Nathan C, and Ding A. SnapShot: Reactive Oxygen Intermediates (ROI). Cell.
2010;140(6):951- e2.
85. Gelderman KA, Hultqvist M, Holmberg J, Olofsson P, and Holmdahl R. T cell surface
redox levels determine T cell reactivity and arthritis susceptibility. Proceedings of the
National Academy of Sciences of the United States of America. 2006;103(34):12831-6.
86. Klemke M, Wabnitz GH, Funke F, Funk B, Kirchgessner H, and Samstag Y. Oxidation of
cofilin mediates T cell hyporesponsiveness under oxidative stress conditions.
Immunity. 2008;29(3):404-13.
87. Malmberg KJ, Arulampalam V, Ichihara F, Petersson M, Seki K, Andersson T, Lenkei R,
Masucci G, Pettersson S, and Kiessling R. Inhibition of activated/memory (CD45RO(+))
T cells by oxidative stress associated with block of NF-kappaB activation. J Immunol.
2001;167(5):2595-601.
88. Mougiakakos D, Johansson CC, and Kiessling R. Naturally occurring regulatory T cells
show reduced sensitivity toward oxidative stress-induced cell death. Blood.
2009;113(15):3542-5.
89. Wittmann C, Chockley P, Singh SK, Pase L, Lieschke GJ, and Grabher C. Hydrogen
peroxide in inflammation: messenger, guide, and assassin. Adv Hematol.
2012;2012(541471.
90. Niethammer P, Grabher C, Look AT, and Mitchison TJ. A tissue-scale gradient of
hydrogen peroxide mediates rapid wound detection in zebrafish. Nature.
2009;459(7249):996-9.
91. Ginis I, and Tauber AI. Activation mechanisms of adherent human neutrophils. Blood.
1990;76(6):1233-9.
92. Gelderman KA, Hultqvist M, Pizzolla A, Zhao M, Nandakumar KS, Mattsson R, and
Holmdahl R. Macrophages suppress T cell responses and arthritis development in mice
by producing reactive oxygen species. The Journal of clinical investigation.
2007;117(10):3020-8.
93. Varga G, Balkow S, Wild MK, Stadtbaeumer A, Krummen M, Rothoeft T, Higuchi T,
Beissert S, Wethmar K, Scharffetter-Kochanek K, et al. Active MAC-1 (CD11b/CD18) on
DCs inhibits full T-cell activation. Blood. 2007;109(2):661-9.
94. Zhuang Y, Peng LS, Zhao YL, Shi Y, Mao XH, Chen W, Pang KC, Liu XF, Liu T, Zhang JY, et
al. CD8(+) T Cells That Produce Interleukin-17 Regulate Myeloid-Derived Suppressor

120
 Suppressive neutrophils and G-MDSCs

Cells and Are Associated With Survival Time of Patients With Gastric Cancer.
Gastroenterology. 2012.
95. Haverkamp JM, Crist SA, Elzey BD, Cimen C, and Ratliff TL. In vivo suppressive function
of myeloid-derived suppressor cells is limited to the inflammatory site. Eur J Immunol.
2011;41(3):749-59.
96. von Andrian UH, and Mackay CR. T-cell function and migration. Two sides of the same
coin. N Engl J Med. 2000;343(14):1020-34.
97. Lawrence CW, Ream RM, and Braciale TJ. Frequency, specificity, and sites of expansion
of CD8+ T cells during primary pulmonary influenza virus infection. J Immunol.
2005;174(9):5332-40.
98. Wakim LM, Waithman J, van Rooijen N, Heath WR, and Carbone FR. Dendritic cell-
induced memory T cell activation in nonlymphoid tissues. Science.
2008;319(5860):198-202.
99. Szczepura KR, Ruparelia P, Solanki CK, Balan K, Newbold P, Summers C, Chilvers ER,
and Peters AM. Measuring whole-body neutrophil redistribution using a dedicated
whole-body counter and ultra-low doses of 111Indium. Eur J Clin Invest.
2011;41(1):77-83.
100. Suratt BT, Young SK, Lieber J, Nick JA, Henson PM, and Worthen GS. Neutrophil 6
maturation and activation determine anatomic site of clearance from circulation. Am
J Physiol Lung Cell Mol Physiol. 2001;281(4):L913-21.
101. Summers C, White JF, Mackenzie IM, Solanki C, Peters AM, and Chilvers ER.
Measurement of neutrophil pulmonary transit time in humans [abstract]. Am J Respir
Crit Care Med. 2009;179:A1333(
102. Nolte MA, Hoen EN, van Stijn A, Kraal G, and Mebius RE. Isolation of the intact white
pulp. Quantitative and qualitative analysis of the cellular composition of the splenic
compartments. Eur J Immunol. 2000;30(2):626-34.
103. Kesteman N, Vansanten G, Pajak B, Goyert SM, and Moser M. Injection of
lipopolysaccharide induces the migration of splenic neutrophils to the T cell area of
the white pulp: role of CD14 and CXC chemokines. J Leukoc Biol. 2008;83(3):640-7.
104. Harmsen AG, Mason MJ, Muggenburg BA, Gillett NA, Jarpe MA, and Bice DE. Migration
of neutrophils from lung to tracheobronchial lymph node. J Leukoc Biol.
1987;41(2):95-103.
105. Yang CW, Strong BS, Miller MJ, and Unanue ER. Neutrophils influence the level of
antigen presentation during the immune response to protein antigens in adjuvants. J
Immunol. 2010;185(5):2927-34.
106. Abadie V, Badell E, Douillard P, Ensergueix D, Leenen PJ, Tanguy M, Fiette L, Saeland
S, Gicquel B, and Winter N. Neutrophils rapidly migrate via lymphatics after
Mycobacterium bovis BCG intradermal vaccination and shuttle live bacilli to the
draining lymph nodes. Blood. 2005;106(5):1843-50.
107. Bonneau M, Epardaud M, Payot F, Niborski V, Thoulouze MI, Bernex F, Charley B,
Riffault S, Guilloteau LA, and Schwartz-Cornil I. Migratory monocytes and granulocytes
are major lymphatic carriers of Salmonella from tissue to draining lymph node. J
Leukoc Biol. 2006;79(2):268-76.
108. Chtanova T, Schaeffer M, Han SJ, van Dooren GG, Nollmann M, Herzmark P, Chan SW,
Satija H, Camfield K, Aaron H, et al. Dynamics of neutrophil migration in lymph nodes
during infection. Immunity. 2008;29(3):487-96.

121
Chapter 6

109. Pesce JT, Liu Z, Hamed H, Alem F, Whitmire J, Lin H, Liu Q, Urban JF, Jr., and Gause WC.
Neutrophils clear bacteria associated with parasitic nematodes augmenting the
development of an effective Th2-type response. J Immunol. 2008;180(1):464-74.
110. Tacchini-Cottier F, Zweifel C, Belkaid Y, Mukankundiye C, Vasei M, Launois P, Milon G,
and Louis JA. An immunomodulatory function for neutrophils during the induction of
a CD4+ Th2 response in BALB/c mice infected with Leishmania major. J Immunol.
2000;165(5):2628-36.
111. Balducci L, and Hardy C. High proliferation of granulocyte-macrophage progenitors in
tumor-bearing mice. Cancer Res. 1983;43(10):4643-7.
112. Younos IH, Dafferner AJ, Gulen D, Britton HC, and Talmadge JE. Tumor regulation of
myeloid-derived suppressor cell proliferation and trafficking. Int Immunopharmacol.
2012;13(3):245-56.
113. Delano MJ, Scumpia PO, Weinstein JS, Coco D, Nagaraj S, Kelly-Scumpia KM, O'Malley
KA, Wynn JL, Antonenko S, Al-Quran SZ, et al. MyD88-dependent expansion of an
immature GR-1(+)CD11b(+) population induces T cell suppression and Th2 polarization
in sepsis. J Exp Med. 2007;204(6):1463-74.
114. Jayaraman P, Parikh F, Lopez-Rivera E, Hailemichael Y, Clark A, Ma G, Cannan D,
Ramacher M, Kato M, Overwijk WW, et al. Tumor-expressed inducible nitric oxide
synthase controls induction of functional myeloid-derived suppressor cells through
modulation of vascular endothelial growth factor release. J Immunol.
2012;188(11):5365-76.
115. Gabrilovich D, Ishida T, Oyama T, Ran S, Kravtsov V, Nadaf S, and Carbone DP. Vascular
endothelial growth factor inhibits the development of dendritic cells and dramatically
affects the differentiation of multiple hematopoietic lineages in vivo. Blood.
1998;92(11):4150-66.
116. Watanabe S, Deguchi K, Zheng R, Tamai H, Wang LX, Cohen PA, and Shu S. Tumor-
induced CD11b+Gr-1+ myeloid cells suppress T cell sensitization in tumor-draining
lymph nodes. J Immunol. 2008;181(5):3291-300.
117. McGill J, and Legge KL. Cutting edge: contribution of lung-resident T cell proliferation
to the overall magnitude of the antigen-specific CD8 T cell response in the lungs
following murine influenza virus infection. J Immunol. 2009;183(7):4177-81.
118. Moyron-Quiroz JE, Rangel-Moreno J, Hartson L, Kusser K, Tighe MP, Klonowski KD,
Lefrancois L, Cauley LS, Harmsen AG, Lund FE, et al. Persistence and responsiveness of
immunologic memory in the absence of secondary lymphoid organs. Immunity.
2006;25(4):643-54.
119. Wakim LM, Gebhardt T, Heath WR, and Carbone FR. Cutting edge: local recall
responses by memory T cells newly recruited to peripheral nonlymphoid tissues. J
Immunol. 2008;181(9):5837-41.
120. Wissinger EL, Stevens WW, Varga SM, and Braciale TJ. Proliferative expansion and
acquisition of effector activity by memory CD4+ T cells in the lungs following
pulmonary virus infection. J Immunol. 2008;180(5):2957-66.
121. Perrone LA, Plowden JK, Garcia-Sastre A, Katz JM, and Tumpey TM. H5N1 and 1918
pandemic influenza virus infection results in early and excessive infiltration of
macrophages and neutrophils in the lungs of mice. PLoS Pathog. 2008;4(8):e1000115.
122. Heidema J, Lukens MV, van Maren WW, van Dijk ME, Otten HG, van Vught AJ, van der
Werff DB, van Gestel SJ, Semple MG, Smyth RL, et al. CD8+ T cell responses in

122
 Suppressive neutrophils and G-MDSCs

bronchoalveolar lavage fluid and peripheral blood mononuclear cells of infants with
severe primary respiratory syncytial virus infections. J Immunol. 2007;179(12):8410-7.
123. Waight JD, Hu Q, Miller A, Liu S, and Abrams SI. Tumor-derived G-CSF facilitates
neoplastic growth through a granulocytic myeloid-derived suppressor cell-dependent
mechanism. PLoS One. 2011;6(11):e27690.
124. Kroepfl JM, Pekovits K, Stelzer I, Fuchs R, Zelzer S, Hofmann P, Sedlmayr P, Dohr G,
Wallner-Liebmann S, Domej W, et al. Exercise Increases the Frequency of Circulating
Hematopoietic Progenitor Cells, But Reduces Hematopoietic Colony-Forming
Capacity. Stem Cells Dev. 2012.
125. Gibson PG, Manning PJ, O'Byrne PM, Girgis-Gabardo A, Dolovich J, Denburg JA, and
Hargreave FE. Allergen-induced asthmatic responses. Relationship between increases
in airway responsiveness and increases in circulating eosinophils, basophils, and their
progenitors. Am Rev Respir Dis. 1991;143(2):331-5.

123
 Chapter 7
Parallel differentiation of normal and immune-
suppressive neutrophils in man

Tamar Tak1, Patrick Wijten2, Marjolein Heeres3, Peter Pickkers4, Arjan Scholten2,
Albert J.R. Heck2, Nienke Vrisekoop1 Luke Leenen3, José A.M. Borghans5, Kiki
Tesselaar5, Leo Koenderman1

1. Department of Respiratory Medicine, Laboratory of Translational Medicine,

University Medical Center Utrecht, Utrecht, the Netherlands
2. Biomolecular Mass Spectrometry and Proteomics, Utrecht Institute for
Pharmaceutical Sciences and Bijvoet Center for Biomolecular Research,
Utrecht University, Utrecht, The Netherlands
3. Department of Trauma Surgery, University Medical Center Utrecht, Utrecht,
the Netherlands
4. Laboratory of Pediatric Infectious Diseases, Radboud University Medical
Centre, Nijmegen, The Netherlands
5. Department of Immunology, Laboratory of Translational Medicine, University
Medical Center Utrecht, Utrecht, the Netherlands

Submitted
Chapter 7

Abstract
During acute inflammation three neutrophil subsets are found in the blood neutrophils
resembling conventional mature neutrophils, neutrophils with a banded nucleus and T-cell
suppressing neutrophils with a hypersegmented nucleus. Little is known about the origins,
properties and developmental relation of these subsets.
We determined the maturation times and proteomic profiles of these neutrophil
subsets during lipopolysaccharide induced acute inflammation. Cellular kinetics was
determined with DNA 2H-enrichment after in vivo pulse chase experiments with 6,6-2H2-
glucose. Labeled banded neutrophils were detected already after 5 days. Labeled
hypersegmented and classical neutrophils both appeared later at day 7, and had similar
kinetics suggesting a non-linear differentiation relationship. In line with this, comparison of
the proteomes by hierarchical clustering placed hypersegmented neutrophils separately from
the other two subsets.
These data imply a linear relation between banded and classical neutrophils.
Hypersegmented neutrophils are a separate subset which is recruited to the bloodstream in
response to inflammation.

126
 Parallel neutrophil subset differentiation

Introduction
Neutrophils engage in resolution of infection by phagocytosis and subsequent killing of
pathogens. They are also vital in maintaining the immunological balance during and after
infection. Decreased functionality of neutrophils can lead to infectious complications and
even sepsis (1). On the other hand, hyperactivation can lead to severe pro-inflammatory
conditions such as acute respiratory distress syndrome (ARDS). Neutrophils balance
hyper/hypo-activation of the immune system by release of pro-inflammatory and toxic
mediators (2, 3), and suppression of immune responses (4).
Acute inflammation induces recruitment of neutrophil phenotypes that are not
present in either circulation or marginated pool (5) during homeostasis. Several studies have
identified such inflammatory phenotypes using different characteristics (6-10), such as
differential 3-fucosyl-N-acetyl-lactosamine (CD15) and Fc-gamma receptor-III (CD16)
expression (6), IL-4Rα (CD124) expression (8), low buoyant density (9), arginase expression
(10) or immunomodulatory functionality (7). Based on the differential expression of (CD16)
and L-selectin (CD62L) (11), we have discriminated at least three different subsets of
inflammatory neutrophils: 1. Neutrophils resembling prototypical mature blood neutrophils
characterized by a CD16bright/CD62Lbright (mature); 2. CD16dim/CD62Lbright cells with a banded
nuclear morphology (banded) and 3. CD16bright/CD62Ldim cells with a hypersegmented nucleus 7
that are capable of suppressing T-cell proliferation (11) (hypersegmented). Nothing is known
regarding the origin of these different subsets. Based on their nuclear morphology, the
banded and hypersegmented have been hypothesized to be young and aged neutrophils
respectively, which are both residing in or returned to the bone marrow (4, 12) during
homeostasis. However, to date, this hypothesis remains to be tested.
The human endotoxemia model is a valuable model to study acute systemic
inflammation (13) and induces changes in blood immune cells reminiscent of those found in
septic (14, 15) or severely injured patients (16). These latter conditions also induce the
recruitment of banded and hypersegmented neutrophils. Transcriptome analyses of
neutrophils (subsets) in the human endotoxemia model (17, 18) revealed large differences
between banded and hypersegmented neutrophils, but this study did not include the mature
neutrophil subset. Moreover, subset characterization of neutrophils by transcriptome
analysis is hampered by the fact that these cells have a low transcriptional capacity (19) and
many changes are taking place at the protein level. Proteins are being produced and stored
during maturation in the bone marrow (3, 20) and can for example be lost by degranulation.
These changes will be missed by mRNA analysis.
Our study was designed to gain more insight into the properties and origins of the
three neutrophil subsets found in acute inflammation. We performed high-end proteomics
analysis on highly purified blood neutrophil subsets to assess differences between the three
subsets observed during acute inflammation. In vivo pulse chase deuterium (2H) labelling was
performed to determine the maturation times of the different subsets.

127
Chapter 7

Materials and Methods

Subjects
Samples were obtained from 20 healthy human volunteers who participated in a human
endotoxemia trial (Clinicaltrials.gov identifier NCT01766414), in which 10 volunteers were
treated with a C1-esterase inhibitor and 10 volunteers were treated with a placebo. All data
presented here are from the placebo group unless stated otherwise. The study was approved
by the ethics review board of the Radboud University Medical Center and in compliance with
the declaration of Helsinki (adapted by WMA general assembly, Fortaleza, Brazil 2013), ICH
Good Clinical Practice guidelines and in compliance with the rulings of the Dutch Medical
Research Involving Human Subjects Act (WMO). Written informed consent was obtained from
all study participants. Subjects were screened and found healthy by physical examination,
electrocardiography and hematological laboratory values. All volunteers were negative for
HIV and hepatitis B antibodies in serum. Subjects with febrile illness in the two weeks before
the study were excluded, as were subjects taking prescription drugs. Subjects were asked to
refrain from caffeine and alcohol use in the 24h prior to the experiment and refrain from food
intake 12h prior to the experiment.

In vivo 2H-labelling
The relative age of neutrophil subsets was determined using in vivo 2H-labelling. In vivo
labeling was performed as described previously (21, 22). In short, all twenty volunteers were
asked to drink 12 half-hourly doses of 6.67 grams of metabolic grade 6,6-2H2-glucose (total
80g, Cambridge Isotope Laboratories, Tewksbury, MA, USA) at three to eleven days before
LPS administration (see below). The 6,6-2H2-glucose label is metabolized and incorporated
into the DNA of all dividing cells via the de novo nucleotide synthesis pathway17,18. During the
labelling procedure, blood was collected by finger prick to determine the availability of label
in plasma, as described previously (22).

Experimental endotoxemia model

LPS challenge was performed according to a strict protocol as described previously (23). In
short, after admission to the research medium care unit of the Radboud Medical Center, heart
rate and blood pressure were monitored starting at t=-1h until discharge at t=8h. A cannula
was placed in an antecubital vein to permit infusion of hydration fluid (1.5L 2.5%
glucose/0.45% saline) from t=-1h until t=0h. Then a single dose of 2ng/kg bodyweight LPS (US
Standard Reference Endotoxin Escherichia Coli O:113, obtained from the Pharmaceutical
Development Section of the NIH, Bethesda, MD, USA) was injected (t=0h). LPS injection was
followed by continuous infusion of hydration fluid (2.5% glucose/0.45% saline, 150ml/h) from
t=0 to t=8. The course of LPS-induced symptoms (headache, shivering, nausea, vomiting,
muscle pain, and back pain) was scored every 30 minutes on a 6-point Likert scale (0=no
symptoms, 5=very severe symptoms; vomiting=3 points), adding up to a maximum of 28
points (24). WBC counts were analyzed on a routine heamatocytometer at t=0, 1, 2, 6 and 8h.

128
 Parallel neutrophil subset differentiation

Neutrophil isolation
Leukocytes were isolated from blood anticoagulated with sodium heparin. After erythrocytes
lysis in ice-cold isotonic erythrocyte lysis buffer (150 mM NH4Cl, 10 mM KHCO3 and 0.1 mM
NA2EDTA),leukocyte preparations were stained with antibodies in PBS supplemented with
0.32% trisodium citrate and 10% human pasteurized plasma solution (Sanquin, Amsterdam,
the Netherlands): CD14-ECD (clone RMO52, Beckman Coulter Pasadena, CA, USA), CD62L-PE
(clone SK11) and CD16-Alexa647 (clone 3G8) (both from BD Biosciences, San Jose, CA, USA).
Neutrophil subsets were sorted using an AriaIII FACS sorter (BD, Biosciences). Neutrophils
were identified based on a FSC/SSC granulocyte gate, doublet exclusion based on SSC
height/width, CD14neg and CD16pos. Subsequently, neutrophil subsets were sorted as
CD16dim/CD62Lbright (banded), CD16bright/CD62Lbright (normal) and CD16bright/CD62Ldim
(hypersegmented) (supplementary figure 1). Sorted populations were re-analysed (typically
>99% pure), washed 3 times in ice-cold PBS, snap-frozen and stored in liquid nitrogen until
further processing.

Determination of 2H enrichment
DNA was isolated from sorted cell populations using a NucleoSpin Blood kit (Macherey-Nagel,
Düren, Germany), hydrolysed and derivatized to pentafluoro triacetate (PFTA) as described in 7
detail previously (22). Relative quantities of unlabeled and 2H labeled PFTA were determined
with an Agilent 7980A/5975C GC-MS in negative chemical ionization mode scanning for m/z
435 (M+0, unlabeled) and m/z 437 (M+2, labelled). Resulting enrichments were corrected for
natural background enrichment and availability of 6,6-2H2-glucose in plasma as described in
chapter 3.

Protein identification and quantification

A detailed description of the proteomics approach can be found in the supplementary
methods section. In short, 3*106 cells of the different neutrophil subsets were isolated from
blood derived from three volunteers in the placebo group 3 h after LPS infusion. Cells of each
subset were digested with trypsin, labeled with a different label for each subset and pooled
in a 1:1:1 ratio for each individual subject before analysis by Mass Spectrometry.
Samples were fractionated using strong cation exchange chromatography (SCX).
Fractions were analyzed using a LTQ-Orbitrap LC-MS/MS (25). Peptides were identified by
searching the resulting peak lists against the Uniprot database (Homo sapiens), with exclusion
of common contaminants. Since glucose is not used for amino acid production and
glycosylated peptides were not measured, in vivo 2H-labelling did not affect protein
quantifications. All results have been deposited in the ProteomeXchange Consortium via the
PRIDE partner repository (26) (dataset identifier PXD001674; DOI: 10.6019/PXD001674).

129
Chapter 7

Hierarchical clustering
Hierarchical clustering was employed to quantify the dissimilarities between the neutrophil
subsets and the individual samples. Proteins with quantification in only one of three
individuals in any of the subsets were excluded from analysis. Remaining missing values were
mean imputed. Using the pvclust plugin (version 1.3-0) for R (version 3.1.1), a Euclidean
dissimilarity matrix was calculated from protein abundances and clustered hierarchically
using the averages method. Significance of the resulting clusters was tested with multiscale
bootstrapping (10,000 repeats) and reported as approximate unbiased p-values +/- 95% CI of
the p-value (27).

Clustering of protein expression data

Sums of protein quantifications from three individual samples were standardized to an
average expression of 0, with a standard deviation of 1. Proteins with missing values for two
or more individuals in any of the three subsets were excluded from analysis. Standardized
expression profiles were subjected to Fuzzy C-means clustering using R with the Mfuzz plugin
(2.3.1) (28). In addition to assigning each protein into a cluster, Fuzzy clustering indicates how
well each protein fits in each cluster. By choosing a minimal required membership value,
outliers are excluded from analysis, making Fuzzy clustering preferable over hard clustering
methods such as the k-means method.
The required number of clusters was determined by calculating the average centroid
distance for clustering runs with two up to forty clusters (29). After seven clusters, addition
of extra clusters did little to improve clustering results. Thus, the amount of clusters was set
to seven. To prevent clustering of random data while retaining an as low as possible false-
negative rate, the optimal value for fuzzifier parameter m was calculated as described
previously (29), resulting in a value of 3.69.

Gene ontology (GO) enrichment analysis

Proteins with cluster membership values of at least 0.33 were analyzed by gene ontology
enrichment analysis. Gene names of identified proteins were obtained from the Uniprot
database. Subsequently, GO-terms of biological processes, molecular functions and cellular
components were obtained for each protein and analyzed using the Gene Ontology
enrichment analysis and visualization tool (GOrilla) (accession date: January 20th 2015) (30,
31). Enrichment was determined for proteins in each cluster compared to the background set
of all identified proteins using an exact mHG p-value computation, corrected for multiplicity
using Benjamini and Hochberg’s FDR correction (30, 32). As this resulted in a list of up to 120
significantly enriched GO-terms in a cluster, results were summarized by removing highly
similar GO-terms with REViGO and visualized in R using the ggplot2 plugin (version 0.9.3.1)
(33).

130
 Parallel neutrophil subset differentiation

Statistics on clinical data

Clinical data were tested for significant deviation from t=0 using one-way repeated measures
ANOVAs with p-values corrected for multiplicity using Dunnett’s correction.

Results
Kinetics of clinical and inflammatory parameters after endotoxin challenge
As shown previously (34), infusion of LPS in healthy volunteers induced clinical symptoms
corresponding with a systemic inflammatory response syndrome (SIRS). Disease scores were
increased 1.5 hours after LPS injection, followed by an increase in heart rate and body
temperature at 2-5.5 hours post LPS (figure 1). The mean arterial pressure decreased from 2
hours after LPS injection until the end of the experiment.
A B

Heart Rate (bpm)

Disease score

C D
Temperature ( C)

Pressure (mmHg)

Figure 1. Clinical parameters after LPS injection. Disease severity scores (A) are increased at 1.5 hours after LPS
injection (t=0), followed by an increased heart rate (B), core temperature (C) and a decreased mean arterial
pressure (D). Data are represented as median +/- 95%CI with n=10. * indicates p<0.05, ** indicates p<0.01, ***
indicates p<0.001 and **** indicates p<0.0001 compared to t=0, as determined by repeated measures one-way
ANOVA, corrected for multiplicity by Dunnett’s multiple comparisons test.

Counts of all leukocytes were decreased one hour after injection (figure 2A-B).
Circulating monocyte numbers normalized after 6h, whereas lymphocyte, eosinophil and
basophil numbers remained low until 8h after LPS injection. After an initial decrease in
neutrophil numbers at 1h after LPS, numbers increased 2.5-fold of the pre-LPS count. FACS

131
Chapter 7

analysis revealed 3 neutrophil subsets based on CD16/CD62L expression (figure 2C), which
appeared at 3h and 6h post-LPS (figure 2D). Microscopic examination of sorted subsets
showed an increased average number of nuclear lobes (hypersegmentation) in the
CD16bright/CD62Ldim population and a banded nuclear phenotype for the CD16dim/CD62Lbright
population as published previously (figure 2E) (11).

Figure 2. Kinetics of WBCs and neutrophil subsets after LPS injection. Neutrophils counts (A) are decreased 1h
after LPS injection (t=0), followed by increased counts. Counts of other leukocytes (B) are also decreased at t=1h,
but recovered at t=6 or t=8h. Different neutrophil subsets (C) were identified based on CD16 and CD62L
expression. The appearance of the banded and hypersegmented subsets (D) was followed over time and started
at t=3. After sorting, microscopic analysis (E) of the different subsets reveals an increasing nuclear lobularity
from banded to mature to hypersegmented neutrophils. Symbols and bars (A,B,D) represent mean +/- 95%CI
(n=10, placebo group). * indicates p<0.05, ** indicates p<0.01, *** indicates p<0.001 and **** indicates
p<0.0001 compared to t=0, as determined by repeated measures one-way ANOVA with Dunnett’s correction for
multiplicity. (E ) individual measurements with lines representing medians +/- IQR (n=7-8). * indicates p<0.05,
** indicates p<0.01 as determined by one-way ANOVA with Holm-Sidak correction for multiplicity.

In vivo 2H-labelling
Since neutrophil progenitors do not proliferate after the myelocyte stage, in vivo 2H-labelling
can be used to determine the time required for the different subsets to mature Under
homeostasis, the first 2H was detected in peripheral blood neutrophil DNA at day 7 after
intake of label (21), indicating that conventional neutrophils mature for more than 5 days in
the bone marrow after the last division of their progenitors (myelocytes). Both the mature
and hypersegmented populations isolated at three hours post-LPS injection show enrichment
curves very similar to neutrophils isolated before LPS challenge (figure 3). Therefore, also
these subsets took approximately 5 day to mature. In contrast, the enrichment curve of the

132
 Parallel neutrophil subset differentiation

banded neutrophil population is shifted to the left, with maximum enrichment as early as day
five, indicating that the banded population is released two days earlier from the bone
marrow, and thus more immature than the other two subsets. No significant differences in
labelling kinetics were observed between placebo and C1-inhibitor treated groups
(supplementary figure S2). Therefore, labelling data were combined from both treatment
groups.
H-enrichment (fraction)

H-enrichment (fraction)
2

Figure 3. Maturation kinetics of neutrophil subsets. Healthy volunteers were labelled with 6,62H2-glucose over
the course of 6h. (A) Availability of label at the day of glucose intake, N=20 mean +/- SD. (B) DNA 2H2-enrichment
of the different inflammatory neutrophil subsets 3 hours after LPS-injection at 3-11 days after 2H2-glucose intake 7
revealed that banded exit the bone marrow two days before the other subsets. Fraction enrichment was
corrected for availability of label. Symbols indicate mean +/- SD, N=4 volunteers at each timepoint at a total of
20 volunteers. * indicates p<0.05 for banded vs mature, # indicates p<0.05 for banded vs pre-LPS and ***
indicates p<0.001 for banded vs all other subsets as determined by a repeated measures 2-way ANOVA with
Tukey’s correction for multiplicity.

Hierarchical clustering of protein expression data

Proteome analysis of the three neutrophil subsets obtained from 3 individuals infused with
LPS, identified and quantified 2161 unique proteins (see supplementary table 1). The relation
between the three neutrophil subsets and each replicate were assessed using hierarchical
clustering on proteins with no more than one missing value in each of the subsets (figure 4,
1755 out of 2161 proteins). For hierarchical clustering a Euclidean dissimilarity matrix was
created to determine the differences between all samples. Subsequently, a dendrogram was
created to visualize the relations between samples. Replicates of each subset clustered
together in the dendrogram. Using multi-scale bootstrapping, we determined that the
existence of these clusters is supported by our data with a probability of over 99.9% (27). The
replicates of banded and mature neutrophils clustered together, and were significantly
separate from the hypersegmented neutrophils. Use of normalized protein expressions or
different distance matrix calculations did not result in formation of different clusters (data
not shown).

133
Chapter 7

Figure 4. Hierarchical clustering of protein expression data. (A) Heatmap of normalized protein expression
profiles. (B) Hierarchical clustering of non-normalized protein expression using averages clustering on an
Euclidean distance matrix. Clusters were tested for statistical significance using multi-scale bootstrapping, which
revealed four statistically significant clusters (indicated by red boxes): one for each neutrophil subset and one
for the banded and mature neutrophils together. P-values are approximately unbiased (AU) p-values indicating
the chance that a cluster exists. I.e. an AU p-value of >.95 indicates a <.05 chance that a cluster does not exist.
Numbers between brackets indicate the standard error of the AU p-value estimation.

Clustering of protein expression profiles

To further examine the differences between the three neutrophil subsets, protein expression
profiles were examined using functional enrichment analysis. For this, proteins were first
clustered based on their relative expression in each subset. Prior to clustering, the optimal
number of clusters was determined using the average centroid distance method, as described
previously (29). The average centroid distance method determines how well all proteins are
clustered for an increasing number of clusters. When addition of extra clusters does not lead
to a better result (similar to a scree plot), the optimal number of clusters is reached. For our
dataset, this number was determined to be seven (figure 5).
The resulting seven protein clusters are clearly different in their relative expression
between the three subsets. For example, there is a cluster of proteins which are highly
upregulated in banded neutrophils (cluster 1), a cluster of neutrophils highly upregulated in
hypersegmented neutrophils (cluster 3) and two clusters with expression increasing or
decreasing from banded to mature to hypersegmented neutrophils (clusters 6 and 4). A total
of 618 proteins had an expression pattern that differed too much from the seven clusters to

134
 Parallel neutrophil subset differentiation

be assigned to any of the clusters. These unclustered proteins were treated as a separate
cluster.

Figure 5. Clustering of protein expression profiles. Clustering analysis divides normalized protein expression
patterns to into 7 distinct protein expression clusters. Three clusters have high expression in one subset 7
compared to the others (clusters 1, 3 and 5 for hypersegmented, banded and mature, respectively), two clusters
have decreased expression in one subset (clusters 2 and 7 for hypersegmented and banded) and two clusters
have gradually increasing protein expression (cluster 6) or decreasing (cluster 4) protein expression from banded
to mature and hypersegmented. Each line represents the expression pattern of one protein. Line colors indicate
the cluster membership value, i.e. the higher the membership value, the more similar to the cluster average. A
minimum cluster membership value of 0.33 was used for further analysis. Numbers between brackets indicate
the number of proteins in each cluster with cluster membership values above 0.33.

Gene ontology enrichment

Gene ontology enrichment analysis was performed on the seven clusters of proteins and the
cluster of non-classified proteins. GO-terms describe biological processes, molecular
functions or cellular component and are annotated to genes/proteins based on evidence from
literature. For gene ontology enrichment analysis protein clusters are tested for enrichment
of GOterms in a cluster of proteins in comparison to the GOterms in the whole dataset (all
2161 identified proteins). The unclustered proteins and cluster 5 did not show any
enrichment. In addition, none of the enrichments in clusters 3 and 7 were statistically
significant after correction for multi-testing. Clusters 1, 2, 4 and 6 on the other hand were
significantly enriched for GO-terms. (See table 1 for the 10 most significantly enriched
GOterm in each cluster and supplementary table 2 for the complete list of enrichments).

135
136
Table 1. Top 10 most significantly enriched GO-terms per cluster.
Cluster 1: FDR Cluster 4: FDR
Enrichment Type Enrichment Type
Hypersegmented ↑ corrected p Band>Mature>Hyper corrected p
Chapter 7

extracellular space 4.98E-09 2.43 Component ribonucleoprotein complex 1.07E-15 2.34 Component
extracellular region 1.98E-08 2.69 Component ribosomal subunit 2.57E-13 3.55 Component
plasma membrane 5.60E-06 1.58 Component RNA metabolic process 4.25E-13 1.89 Process
cotranslational protein targeting to
response to organic substance 1.08E-05 1.86 Function 4.97E-13 3.32 Process
membrane
SRP-dependent cotranslational protein
defense response 1.40E-05 1.86 Function 5.26E-13 3.36 Process
targeting to membrane
response to external stimulus 1.75E-05 2.07 Function protein targeting to ER 7.39E-13 3.28 Process
response to oxygen-
2.22E-05 2.04 Function nucleic acid metabolic process 7.50E-13 1.78 Process
containing compound
response to external biotic
2.41E-05 2.25 Function translational initiation 9.36E-13 3.03 Process
stimulus
protein localization to endoplasmic
response to chemical 2.52E-05 1.69 Function 1.36E-12 3.2 Process
reticulum
establishment of protein localization to
response to biotic stimulus 3.41E-05 2.22 Function 1.54E-12 3.2 Process
endoplasmic reticulum
Cluster 2: FDR Cluster 6: FDR
Enrichment Type Enrichment Type
Hypersegmented ↓ corrected p Hyper>Mature>Band corrected p
ribonucleoprotein complex 3.72E-07 2.25 Component extracellular vesicular exosome 2.32E-04 1.26 Component
nucleic acid binding 1.01E-05 1.61 Function extracellular vesicle 2.90E-04 1.26 Component
extracellular membrane-bounded
RNA binding 2.02E-05 1.66 Function 3.86E-04 1.26 Component
organelle
RNA metabolic process 4.30E-05 1.82 Process vesicle 5.18E-04 1.23 Component
poly(A) RNA binding 9.61E-05 1.7 Function extracellular region part 5.57E-04 1.24 Component
nucleic acid metabolic process 3.81E-04 1.67 Process extracellular organelle 5.80E-04 1.26 Component
structural constituent of
9.23E-04 2.99 Function membrane-bounded vesicle 8.88E-04 1.25 Component
ribosome
ribosomal subunit 1.55E-03 2.91 Component actin binding 7.31E-03 2.18 Function
negative regulation of Wnt signaling
mRNA metabolic process 1.82E-03 2.04 Process 2.00E-02 2.45 Process
pathway
nicotinamide nucleotide metabolic
translational elongation 2.04E-03 2.89 Process 2.02E-02 3.72 Process
process
 Parallel neutrophil subset differentiation

Using ReViGO (accession date 22-01-2015), highly similar GO-terms were removed. The
remaining GO-terms for biological processes (figure 6) were clustered based on semantic
similarity. Thereby, enrichments of proteins involved in biological processes can clearly be
recognized. Cluster 1, which consists of proteins upregulated in hypersegmented neutrophils,
contains genes involved in adhesion and activation, response to stimuli, regulation of these
responses and regulation of the immune system in general. Cluster 2, consisting of proteins
downregulated in hypersegmented neutrophils shows enrichment of proteins involved in
protein shuttling, ribosomal proteins and proteins involved in RNA metabolism, splicing,
translation etc. Similarly, cluster 4, which also contains proteins downregulated in
hypersegments consists mostly of proteins involved in mRNA processing and biosynthesis.
Lastly, cluster 6 shows a low expression in banded neutrophils, which increases in mature and
hypersegmented neutrophils. This clusters shows enrichment in proteins involved in
nucleotide metabolism, MHC class I antigen presentation, protein degradation and
suppression of the Wnt-signaling pathway. Taken together, the functional enrichment
analysis indicated banded cells as being younger cells still in the process of producing large
amounts of proteins, whereas hypersegmented neutrophils show enrichment of proteins
involved in immune regulation, which is in line with their T-cell suppressing capacity (11).
7
Discussion
In addition to their previously ascribed differences in morphology and function (11, 35) we
here show that the three neutrophil subsets present in blood during acute inflammation
display large differences in protein expression profiles. The immune suppressive capability of
hypersegmented neutrophils is represented by their increased expression of proteins
involved in immune regulation compared to the other two subsets.
Although neutrophils were shown in vitro to retain some capability for de novo protein
production (36, 37), they lose most of their ribosomes, their nucleolus and rough ER during
maturation from banded to mature neutrophils in the bone marrow (19). Therefore, cytokine
production by neutrophils is orders of magnitude lower compared to monocytes and it also
takes longer, with the earliest protein production detected at 3-6h after stimulation (38). This
is in agreement with our findings of reduced expression of ribosomal and RNA
processing/synthesizing proteins in mature cells compared to the younger banded cells
(figure 7). The hypersegmented neutrophil subset shows an even further decrease in
expression of proteins involved in mRNA and protein production, suggesting a very low
capacity for de novo protein synthesis.

137
 Chapter 7

Semantic space y

Semantic space x

Figure 6. Gene ontology enrichment analysis for protein involvement in cellular processes. The seven clusters
from figure 5 were subjected to gene ontology enrichment analysis. After correction for multi-testing, four
clusters revealed significant GO enrichments, which are plotted based on semantic similarities (i.e. similar terms
group together). Colors indicate FDR p-values for each term and the size indicates how general a term is, with a
smaller size for a more specific term. Highly similar terms were removed for clarity, but are available in
supplementary table 2.

138
 Parallel neutrophil subset differentiation

A 40S ribosomal subunit

1000 Banded
Mature
Hyper
Abundance (ppm)

100

10

1
x
S4a

S11
S12
S13
S4
S115
S1a
S16
S17
S18
S29
S20
S21
S23
S24
S25
S27
S38
0
S2
S33

S5
S6
S7
S8
S19
SA

5
S

B 1000
60S ribosomal subunit
Banded
Mature
Hyper
Abundance (ppm)

100

10

7
1
L1A
L3
L4
L5
L6
L77
a
L8
L19
L1 0
L1a
L12
L1 3

L14
L15
L17
L18
L29
L20
L22
L2 3
L2a
L24
L6
L227
L2a
L38
L30
L32
L4
L335
L3a
L3 6
7a
P0
P1
P2

5
3

C Eukaryotic Translation Initiation & Elongation Factors

1000 Banded
Mature
Hyper
Abundance (ppm)

100

10

1
EIIF3 I

F4G1
EEI IF2

E F1A1
EEEF B2

F E1

S33A1
S3A2
S3A3
S3B1
S3B2
S3B4
B5
EI EI1AY
EI2A1B

EI F3A
E F3 B

E IF K
EIF3 M

EI IF4B
E F3 J

E I F4 4 E

F A
E 5B

EE1G
EIF3 F
FG
EIF2K2
EIF2B1
EIF2S1
EI2S2
EIF3 3

EIIF3C

E 3H

EI IF33L
EIF4S6
EIF4A1
EI4A2
EF 3

EIF4 E3

EIF5 5

EE IF6

EEF11D

F2
E I F3

G
F S

F A

EEF1
F F

S
F
EI

Initiation Elongation Splicing

Figure 7. Expression pattern of ribosomal proteins. Gene ontology enrichment analysis in figure 6 and table 1
show enrichment of GO-terms for the ribosome, mRNA processing and translation in clusters 3 and 4. To further
illustrate this finding, we have plotted the expression patterns of all 65 detected ribosomal proteins of (A) the
40S ribosomal subunit, (B) the 60S ribosomal subunit and (C) an additional 37 eukaryotic initiation/elongation
factors (EIF/EEF) and 7 mRNA splicing factors. Each circle represents a single measurement, lines connect
measurements within the same individual. The abundance of some proteins lay below the detection limit
(approximately 2 ppm, depending on protein length and sequence).

139
Chapter 7

Protein expression profiles indicated many proteins to have a gradual increase or

decrease in expression from banded to mature and to hypersegmented (clusters 4 and 6),
which supports the hypothesis that hypersegmented neutrophils are a later neutrophil
maturation stage (4, 12). However, pulse-chase labelling with 6,6-2H2-glucose clearly
demonstrate that hypersegmented neutrophils have identical kinetics for maturation time
and/or disappearance rate from blood. This shows that hypersegmented neutrophils need
the same time to mature as conventional neutrophils and are, thus, not aged mature
neutrophils.
The highly similar maturation kinetics of mature and hypersegmented neutrophils was
not reflected by comparable protein expression profiles, as hierarchical clustering of the
samples revealed the mature subset to be more similar to the two days younger banded
neutrophils than to hypersegmented neutrophils. The difference in protein expression
profiles can be explained by either a rapid change in protein expression by neutrophils already
in the blood during the 3 hours after LPS injection, or by hypersegmented neutrophils being
a subset with a different protein expression recruited into the bloodstream. Over 300 proteins
were upregulated in hypersegmented neutrophils compared to the other two subsets (cluster
1). Considering the low ribosomal content of neutrophils (19) and the further decreased
expression of proteins involved in translation, transcription and protein transport (clusters 2
and 4) in hypersegmented neutrophils, it seems unlikely that hypersegmented neutrophils
rapidly synthesize proteins in response to LPS. In addition, cluster 1 contains transmembrane
proteins which are normally produced in the rough ER, an organelle that is lacking in mature
neutrophils (19). Furthermore, the hypothesis that hypersegmented cells originate from a
rapid differentiation from normally segmented neutrophils during lps challenge poses the
question why an even larger number of neutrophils with a similar age did not respond in a
similar manner and retained a normal nuclear phenotype (Figure 2D).
The hypothesis of hypersegmented neutrophils as a separate subset is supported by
the fact that the majority of neutrophils under homeostasis does not reside in the peripheral
blood. Large numbers of neutrophils have been found in lungs, spleen, liver, bone marrow
and the marginated pool (6, 39, 40). Important experiments performed in the 1960’s
demonstrated that the increase in neutrophil numbers after LPS was due to cells entering the
bloodstream that were different compared to homeostatic cells. This was different compared
to neutrophilia in response to exercise, adrenalin or prednisone, which caused recruitment of
neutrophils from the marginated pool that consisted of cells similar to those in the circulation
before challenge (5, 40, 41). Banded neutrophils most likely originate from the bone marrow,
considering the large number of banded neutrophils found in this organ (42) under
homeostatic conditions. Hypersegmented neutrophils can also be found in the BM, albeit at
lower numbers than banded and mature neutrophils (43). Therefore, the bone marrow is also
a potential source of hypersegmented neutrophils. Alternatively, neutrophil reserves have
been described in the liver and spleen and are therefore also potential organs where
hypersegmented neutrophils reside in homeostasis (6, 39, 40).

140
 Parallel neutrophil subset differentiation

Taken together, the proteome expression profiles of the three neutrophil subsets
indicate that these subsets with specialized functions evolve with different functional
proteomes. Our study provides evidence that hypersegmented neutrophils follow a separate
differentiation pathway compared to normally segmented neutrophils with a similar
maturation time. These immune suppressive cells are only recruited to the blood stream
during acute inflammatory conditions.

141
Chapter 7

References
1. Hotchkiss RS, Coopersmith CM, McDunn JE, and Ferguson TA. The sepsis seesaw:
tilting toward immunosuppression. Nature medicine. 2009;15(5):496-7.
2. Tecchio C, Micheletti A, and Cassatella MA. Neutrophil-derived cytokines: facts
beyond expression. Frontiers in immunology. 2014;5(508.
3. Rorvig S, Ostergaard O, Heegaard NH, and Borregaard N. Proteome profiling of human
neutrophil granule subsets, secretory vesicles, and cell membrane: correlation with
transcriptome profiling of neutrophil precursors. Journal of leukocyte biology.
2013;94(4):711-21.
4. Pillay J, Tak T, Kamp VM, and Koenderman L. Immune suppression by neutrophils and
granulocytic myeloid-derived suppressor cells: similarities and differences. Cellular
and molecular life sciences : CMLS. 2013.
5. Athens JW, Haab OP, Raab SO, Mauer AM, Ashenbrucker H, Cartwright GE, and
Wintrobe MM. Leukokinetic studies. IV. The total blood, circulating and marginal
granulocyte pools and the granulocyte turnover rate in normal subjects. The Journal
of clinical investigation. 1961;40(989-95.
6. Puga I, Cols M, Barra CM, He B, Cassis L, Gentile M, Comerma L, Chorny A, Shan M, Xu
W, et al. B cell-helper neutrophils stimulate the diversification and production of
immunoglobulin in the marginal zone of the spleen. Nature immunology.
2012;13(2):170-80.
7. Gabrilovich DI, and Nagaraj S. Myeloid-derived suppressor cells as regulators of the
immune system. Nature reviews Immunology. 2009;9(3):162-74.
8. Mandruzzato S, Solito S, Falisi E, Francescato S, Chiarion-Sileni V, Mocellin S, Zanon A,
Rossi CR, Nitti D, Bronte V, et al. IL4Ralpha+ myeloid-derived suppressor cell expansion
in cancer patients. J Immunol. 2009;182(10):6562-8.
9. Schmielau J, and Finn OJ. Activated granulocytes and granulocyte-derived hydrogen
peroxide are the underlying mechanism of suppression of t-cell function in advanced
cancer patients. Cancer research. 2001;61(12):4756-60.
10. Rodriguez PC, Ernstoff MS, Hernandez C, Atkins M, Zabaleta J, Sierra R, and Ochoa AC.
Arginase I-producing myeloid-derived suppressor cells in renal cell carcinoma are a
subpopulation of activated granulocytes. Cancer research. 2009;69(4):1553-60.
11. Pillay J, Kamp VM, van Hoffen E, Visser T, Tak T, Lammers JW, Ulfman LH, Leenen LP,
Pickkers P, and Koenderman L. A subset of neutrophils in human systemic
inflammation inhibits T cell responses through Mac-1. The Journal of clinical
investigation. 2012;122(1):327-36.
12. Nauseef WM, and Borregaard N. Neutrophils at work. Nature immunology.
2014;15(7):602-11.
13. Lowry SF. Human endotoxemia: a model for mechanistic insight and therapeutic
targeting. Shock. 2005;24 Suppl 1(94-100.
14. Calvano SE, Xiao W, Richards DR, Felciano RM, Baker HV, Cho RJ, Chen RO, Brownstein
BH, Cobb JP, Tschoeke SK, et al. A network-based analysis of systemic inflammation in
humans. Nature. 2005;437(7061):1032-7.
15. Wong HR, Freishtat RJ, Monaco M, Odoms K, and Shanley TP. Leukocyte subset-
derived genomewide expression profiles in pediatric septic shock. Pediatric critical
care medicine : a journal of the Society of Critical Care Medicine and the World
Federation of Pediatric Intensive and Critical Care Societies. 2010;11(3):349-55.

142
 Parallel neutrophil subset differentiation

16. Visser T, Pillay J, Pickkers P, Leenen LP, and Koenderman L. Homology in systemic
neutrophil response induced by human experimental endotoxemia and by trauma.
Shock. 2012;37(2):145-51.
17. de Kleijn S, Kox M, Sama IE, Pillay J, van Diepen A, Huijnen MA, van der Hoeven JG,
Ferwerda G, Hermans PW, and Pickkers P. Transcriptome kinetics of circulating
neutrophils during human experimental endotoxemia. PloS one. 2012;7(6):e38255.
18. de Kleijn S, Langereis JD, Leentjens J, Kox M, Netea MG, Koenderman L, Ferwerda G,
Pickkers P, and Hermans PW. IFN-gamma-stimulated neutrophils suppress
lymphocyte proliferation through expression of PD-L1. PloS one. 2013;8(8):e72249.
19. Zucker-Franklin D. Atlas of blood cells: function and pathology. Edi. Ermes; 1988.
20. Borregaard N, Sehested M, Nielsen BS, Sengelov H, and Kjeldsen L. Biosynthesis of
granule proteins in normal human bone marrow cells. Gelatinase is a marker of
terminal neutrophil differentiation. Blood. 1995;85(3):812-7.
21. Macallan DC, Fullerton CA, Neese RA, Haddock K, Park SS, and Hellerstein MK.
Measurement of cell proliferation by labeling of DNA with stable isotope-labeled
glucose: studies in vitro, in animals, and in humans. Proc Natl Acad Sci U S A.
1998;95(2):708-13.
22. Macallan DC, Asquith B, Zhang Y, de Lara C, Ghattas H, Defoiche J, and Beverley PC.
Measurement of proliferation and disappearance of rapid turnover cell populations in
human studies using deuterium-labeled glucose. Nature protocols. 2009;4(9):1313-
27. 7
23. Leentjens J, Kox M, Koch RM, Preijers F, Joosten LA, van der Hoeven JG, Netea MG,
and Pickkers P. Reversal of immunoparalysis in humans in vivo: a double-blind,
placebo-controlled, randomized pilot study. Am J Respir Crit Care Med.
2012;186(9):838-45.
24. Kox M, van Eijk LT, Zwaag J, van den Wildenberg J, Sweep FC, van der Hoeven JG, and
Pickkers P. Voluntary activation of the sympathetic nervous system and attenuation
of the innate immune response in humans. Proc Natl Acad Sci U S A.
2014;111(20):7379-84.
25. Wijten P, van Holten T, Woo LL, Bleijerveld OB, Roest M, Heck AJ, and Scholten A. High
precision platelet releasate definition by quantitative reversed protein profiling--brief
report. Arterioscler Thromb Vasc Biol. 2013;33(7):1635-8.
26. Vizcaino JA, Reisinger F, Cote R, and Martens L. PRIDE: Data submission and analysis.
Curr Protoc Protein Sci. 2010;Chapter 25(Unit 25 4.
27. Suzuki R, and Shimodaira H. Pvclust: an R package for assessing the uncertainty in
hierarchical clustering. Bioinformatics. 2006;22(12):1540-2.
28. Kumar L, and M EF. Mfuzz: a software package for soft clustering of microarray data.
Bioinformation. 2007;2(1):5-7.
29. Schwammle V, and Jensen ON. A simple and fast method to determine the parameters
for fuzzy c-means cluster analysis. Bioinformatics. 2010;26(22):2841-8.
30. Eden E, Navon R, Steinfeld I, Lipson D, and Yakhini Z. GOrilla: a tool for discovery and
visualization of enriched GO terms in ranked gene lists. BMC bioinformatics.
2009;10(48.
31. Consortium GO. 2015.
32. Benjamini Y, and Hochberg Y. Controlling the False Discovery Rate: A Practical and
Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society Series
B (Methodological). 1995;57(1):289-300.

143
Chapter 7

33. Supek F, Bosnjak M, Skunca N, and Smuc T. REVIGO summarizes and visualizes long
lists of gene ontology terms. PloS one. 2011;6(7):e21800.
34. van der Meer W, Pickkers P, Scott CS, van der Hoeven JG, and Gunnewiek JK.
Hematological indices, inflammatory markers and neutrophil CD64 expression:
comparative trends during experimental human endotoxemia. Journal of endotoxin
research. 2007;13(2):94-100.
35. Kamp VM, Pillay J, Lammers JW, Pickkers P, Ulfman LH, and Koenderman L. Human
suppressive neutrophils CD16bright/CD62Ldim exhibit decreased adhesion. Journal of
leukocyte biology. 2012;92(5):1011-20.
36. Lindemann SW, Yost CC, Denis MM, McIntyre TM, Weyrich AS, and Zimmerman GA.
Neutrophils alter the inflammatory milieu by signal-dependent translation of
constitutive messenger RNAs. Proc Natl Acad Sci U S A. 2004;101(18):7076-81.
37. Fessler MB, Malcolm KC, Duncan MW, and Worthen GS. A genomic and proteomic
analysis of activation of the human neutrophil by lipopolysaccharide and its mediation
by p38 mitogen-activated protein kinase. The Journal of biological chemistry.
2002;277(35):31291-302.
38. Zimmermann M, Aguilera FB, Castellucci M, Rossato M, Costa S, Lunardi C, Ostuni R,
Girolomoni G, Natoli G, Bazzoni F, et al. Chromatin remodelling and autocrine
TNFalpha are required for optimal interleukin-6 expression in activated human
neutrophils. Nat Commun. 2015;6(6061.
39. Saverymuttu SH, Peters AM, Keshavarzian A, Reavy HJ, and Lavender JP. The kinetics
of 111indium distribution following injection of 111indium labelled autologous
granulocytes in man. British journal of haematology. 1985;61(4):675-85.
40. Summers C, Rankin SM, Condliffe AM, Singh N, Peters AM, and Chilvers ER. Neutrophil
kinetics in health and disease. Trends in immunology. 2010;31(8):318-24.
41. Athens JW, Raab SO, Haab OP, Mauer AM, Ashenbrucker H, Cartwright GE, and
Wintrobe MM. Leukokinetic studies. III. The distribution of granulocytes in the blood
of normal subjects. The Journal of clinical investigation. 1961;40(159-64.
42. Donohue DM, Reiff RH, Hanson ML, Betson Y, and Finch CA. Quantitative
measurement of the erythrocytic and granulocytic cells of the marrow and blood. The
Journal of clinical investigation. 1958;37(11):1571-6.
43. Bohinjec J. Myelokathexis: chronic neutropenia with hyperplastic bone marrow and
hypersegmented neutrophils in two siblings. Blut. 1981;42(3):191-6.

144
 Parallel neutrophil subset differentiation

Supplementary Information
Supplementary Tables will be made available online with the publication of this paper.

Supplementary Methods

Protein isolation and labelling

Cell pellets were reconstituted in a pH 8.0 buffer containing 100mM Tris, , 10mM DTT, 2%
SDS, and protease inhibitor cocktails (Complete Mini and PhosSTOP, roche diagnostics).
Samples were sonicated for a total of 4 minutes and subsequently centrifuged at 14000g to
remove cellular debris. Using the Filter Aided Sample Preparation (FASP) method1, the SDS
buffer was replaced by a 8M urea buffer pH 8.0, alkylated with 50mM iodoacetamide and
digested by 4h incubation with Lys-c (Wako, Richmond, VA), followed by an overnight Trypsin
incubation (Promega, Madison, WI) and subsequent acidification to 5% formic acid. After
desalting with Sep-Pak C18 cartridges (Waters Corporation, Milford, MA), resulting peptides
were labelled by stable isotope dimethyl labelling as described previously2 (1H2-
formaldehyde, 2H2-formaldehyde and 13C-2H2-formaldehyde for light, medium and heavy
label, respectively). Peptides of the three neutrophil subsets were mixed in a 1:1:1 ratio, as
7
determined by LC-MS.

Fractionation
Samples were dried in vacuo and resuspended in 10% formic acid and fractionated using
Strong Cation Exchange (SCX) chromatography on an Agilent 1100 HPLC system equipped
with a 0.8x50mm (3.5µm resin) Zorbax BioSCX-series II analytical column (both from Agilent
Technologies, Waldbronn, Germany) and a 2.1x15mm (40μm i.d., 10Å Resin) Opti Lynx C18
trapping column (Optimize Technologies, Oregon, OR). Peptides were loaded on the trapping
column in a 0.05% formic acid, pH 2.9 buffer and eluted using 0.05% formic acid, 80%
acetonitrile, pH 2.9 supplemented with non-linearly increasing concentrations of up to
500mM NaCl for SCX fractionation. A total number of 50 SCX fractions were collected and
dried in a vacuum centrifuge. Fractions containing 2+, 3+ and 4+ peptides were reconstituted
in 10% formic acid for further analysis.

Liquid chromatography and tandem mass spectrometry (LC-MS/MS)

Samples were analyzed with Proxeon EASY-nLC 1000 coupled to a LTQ-Orbitrap Q-Exactive
(Thermo Fisher Scientific, Bremen, Germany) equipped with a 20mm*100µm i.d. ReproSil-Pur
C18-AQ (Dr. Maisch GmbH, Ammerbuch, Germany) trapping column and a 450mm*50µm i.d.
Poroshell 120 EC-C18 (Agilent Technologies, Little Fall, DE, USA) analytical column (packed in
house with 3 and 2.7μm resin, respectively). Fractions were submitted to 45-180 minute
elutions (based on peptide content measured by 220nm UV-intensity) with 0.1% formic acid
containing an increasing concentration of 7 to 30% acetonitrile. Between fractions Columns
were washed with 100% acetonitrile and equilibrated down to 7%. Nanospray was achieved
using a distally coated fused silica emitter (inner and outer diameters 375μm and 20μm,

145
Chapter 7

respectively; made in-house) biased to 1.7 kV. High resolution full survey scans (35,0000
FHMW) were acquired on the orbitrap from m/z 350 to 1500, whereas the MS2 scan is
acquired at a resolution of 17.500. The ten most intense precursors were isolated with an
isolation width of 1.5 m/z and fragmented using High Energy Collision Dissociation (HCD). The
target ion setting was 3*106 for MS and 5*104 for MS/MS, with a maximum fill-time of 250 ms
and 120 ms.

Data analysis: Identification and Quantitation

Peptide identification was performed by searching the peak lists against a concatenated
target-decoy database of human protein sequences in the Uniprot database (release
2013_12) using Proteome Discoverer version 1.4.0.288 (Thermo Scientific, Bremen,
Germany), supplemented with a common contaminants database using the Mascot search
engine version 2.3 (Matrix Science, London, United Kingdom). Search parameters allowed up
to two missed trypsin cleavage sites per peptide, carbamidomethylation was assumed for all
cysteine residues, allowed methionine oxidation as a result of sample handling and assumed
labelling on both lysine residues and N-termini. Mass tolerances were set at 50ppm for
precursor peptides, while fragment mass tolerance was set at 0.05Da using HCD
fragmentation. Use of the percolator algorithm3 resulted in a false-discovery rate of less than
one percent. Identified peptides were filtered for an ion score of 20, and a peptide rank of 1
so only the best matching peptide is used. Resulting proteins with evidence only on cDNA
level were manually re-analysed for matching proteins with evidence on protein levels.
All results have been deposited in the ProteomeXchange Consortium via the PRIDE
partner repository4 with dataset identifier PXD001674 and DOI: 10.6019/PXD001674).

Supplementary references

1 Wisniewski, J. R., Zougman, A. & Mann, M. Combination of FASP and StageTip-based

fractionation allows in-depth analysis of the hippocampal membrane proteome.
Journal of proteome research 8, 5674-5678, doi:10.1021/pr900748n (2009).
2 Boersema, P. J. et al. In-depth qualitative and quantitative profiling of tyrosine
phosphorylation using a combination of phosphopeptide immunoaffinity purification
and stable isotope dimethyl labeling. Molecular & cellular proteomics : MCP 9, 84-99,
doi:10.1074/mcp.M900291-MCP200 (2010).
3 Kall, L., Storey, J. D., MacCoss, M. J. & Noble, W. S. Assigning significance to peptides
identified by tandem mass spectrometry using decoy databases. Journal of proteome
research 7, 29-34, doi:10.1021/pr700600n (2008).
4 Vizcaino, J. A., Reisinger, F., Cote, R. & Martens, L. PRIDE: Data submission and
analysis. Curr Protoc Protein Sci Chapter 25, Unit 25 24,
doi:10.1002/0471140864.ps2504s60 (2010).

146
 Parallel neutrophil subset differentiation

Supplementary Figures

7
Supplementary figure S1. Sort logic and sort results of neutrophil subsets. Granulocytes were gated on FSC/SSC
(A), with exclusion of doublets (B) and CD14pos monocytes (C). Subsequently, neutrophil subsets (D) were gated
as CD16dim/CD62Lbright, CD16high/CD62Lbright and CD16high/CD62Ldim. (E) Re-analysis of the sorted cell populations
shows a clear separation between the subsets and microscopic analysis of MGG stained cytospin preparations
shows the expected banded (F), mature (G) and hypersegmented (H) nuclear morphology.

Mature t=3h Supplementary Figure S2.

0.10 Deuterium enrichment for
Placebo
placebo and C1-esterase
Enrichment (fraction)

Enrichment (fraction)

C1E inhibitor
inhibitor treated groups.
0.05 No differences were
observed between kinetics
of placebo and treated
0.00 volunteers in the
3 5 7 9 11 neutrophils before LPS
Time of LPS injection
2 injection and the three
(days post H-glucose intake)
subsets after LPS injection.
Each symbol represents a
single measurement in one
Enrichment (fraction)

Enrichment (fraction)

volunteer, with 2
volunteers per condition
per timepoint, to a total of
20 volunteers.

147
 Chapter 8
Neutrophil Mac-1 (CD11b) mediates influenza-
induced pathology in mice

Tamar Tak1,7, Janesh Pillay1,2,7, Guruswami Karnam3, Okan Bastian4,7, Louis Boon5,
Marco Viveen6, Tomasz P. Rygiel4, Frank E. Coenjaerts6 , Linde Meyaard4,7,
Leo Koenderman1,7

1. Department of Respiratory Medicine, University Medical Center Utrecht,

Utrecht, The Netherlands
2. Department of Anesthesiology, University Medical Center Utrecht,
Utrecht, The Netherlands
3. Department of immunology, University Medical Center Utrecht,
Utrecht, The Netherlands
4. Department of Traumatology, University Medical Center Utrecht,
Utrecht, The Netherlands
5. Bioceros B.V. Utrecht, the Netherlands
6. Department of bacteriology, University Medical Center Utrecht,
Utrecht, The Netherlands

Manuscript in preparation
Chapter 8

Abstract
Influenza virus infections can lead to life-threatening pathology through immune-mediated
tissue damage. In various experimental models, this damage is dependent on T-cells. In
humans, we have recently identified hypersegemented neutrophils capable of suppressing T-
cells, which requires neutrophil CD11b/CD18. Therefore, we tested the hypothesis that
immune suppression by neutrophils can ameliorate pathology after influenza infection.
WT and CD11b-/- mice were infected with A/HK/2/68 (H3N2) influenza virus.
Different conditions to rescue from pathology were studied: T-cell depletion and adaptive
transfer of wt-neutrophils in CD11b-/- mice. Disease was monitored by weight loss, leukocyte
infiltration and immunohistochemistry.
We demonstrated that CD11b-/- mice suffered increased weight-loss compared to
wild-type animals, accompanied with increased lung damage, leukocyte infiltration and
presence of hypersegmented neutrophils in the lungs. The increased pathology in CD11b-/-
mice was T-cell dependent as determined by T-cell depletion. In addition, disease in CD11b-/-
mice was accompanied by higher numbers of T-cell in the lungs early during infection
compared to WT mice. These differences in pathology were not associated with an increased
viral load, confirming that pathology is immune-mediated rather than caused by virus-
induced damage. A single adoptive transfer of wild-type neutrophils partly restored
protection against influenza-induced pathology, demonstrating the importance of neutrophil
CD11b.
Our data indicate that neutrophil CD11b is involved in preventing pathology in
influenza induced T-cell mediated disease.

150
 Neutrophil CD11b mediates influenza-induced pathology

Introduction
Influenza viruses cause seasonal epidemics each year. Infection leads to an immune response
aimed at viral clearance. However, activation of the immune system may lead to lung damage
(1, 2) and in rare cases, a potentially fatal cytokine-storm (3). Thus, the antiviral response
needs to be tightly regulated.
Large numbers of neutrophils are recruited to the lungs during influenza infection.
These cells are viewed as beneficial in bacterial and fungal infections due to their capacity to
phagocytose and kill these pathogens. Their role in anti-viral responses, however, is
controversial (4). In influenza infection, the role of neutrophils is largely considered to be
detrimental, as they can cause tissue damage by exaggerated production of reactive oxygen
species (ROS) and other cytotoxic compounds (5). Supporting this latter view is the finding
that low doses of neutrophil-depleting antibodies lead to a protective effect in experimental
influenza infection (2). This has led to the prevailing idea that neutrophils must be inhibited
in influenza infection in order to limit pathology. On the other hand, neutrophil depletion with
high doses of anti-neutrophil antibodies leads to increased pathology and death (2, 6). This
indicates that neutrophils are not only detrimental in influenza infection and suggests a
protective role for neutrophils in influenza infection.
A possible mechanism by which neutrophils protect from influenza-induced pathology
is by their recently described ability to suppress immune responses (7). One of the
mechanisms by which neutrophils can inhibit immune responses is by suppressing T-cell 8
proliferation in a CD11b/CD18 and H2O2-dependent manner (8). To our knowledge
neutrophils are not required for anti-viral responses (2) and influenza-induced pathology is T-
cell dependent (9). This allowed us to test our hypothesis that CD11b dependent neutrophil-
induced T-cell suppression prevents influenza-induced pathology in mice.
To investigate the hypothesis that CD11b dependent T-cell suppression by neutrophils
limits influenza-induced pathology, we infected wild-type (WT) and CD11b-/- mice with a sub-
lethal dose of influenza and followed disease progression over time. To assess the role of
CD11b on neutrophils specifically, CD11b-/- mice were reconstituted with WT neutrophils by
adoptive transfer.

Materials and Methods

Mice and infections
Wild type C57BL/6J mice were obtained from Charles River Laboratories (Erkrath, Germany).
CD11b-/- mice on C57BL/6J background were kindly provided by Dr. J. Leusen and bred at the
specific pathogen-free unit of the Utrecht University Central Animal Laboratory (GDL).
Experiments were approved by the local Ethical Committee for Animal Experiments (DEC) in
accordance with current Dutch and European legislation.
Influenza A/HK/2/68 (H3N2) virus was grown in fertilized hen’s eggs as published
previously (9). Infection was performed by intranasal administration of 0.03
heamagluttination units (HAU) in 50 μl PBS under mild ether anaesthesia unless stated

151
Chapter 8

otherwise. Mice were weighed once every 24h to monitor disease progression. A loss of 20%
weight from day 0 was considered the humane endpoint in accordance with Dutch
regulations. When any of the mice reached the humane endpoint (usually at day 8), all
animals were sacrificed by i.p. injection of 12mg Sodium Pentobarbital (Veterinary Medicine
Pharmacy, Utrecht, the Netherlands) for sample collection. For depletion of T-cells, 20μg anti-
CD4 (clone GK1.5) and 50μg anti-CD8 (clone YTS169, both from Bioceros, Utrecht, the
Netherlands) was injected at days -1 and 1 relative to infection.

Neutrophil isolations
Mature neutrophils were isolated from bone marrow using discontinuous Percoll gradients
as published previously (10). In short, bone marrow cavities of the femur and tibia of donor
mice were flushed with HBSS and lysed for RBCs using isotonic ammoniumchloride solution
(150 mM NH4Cl, 10 mM KHCO3 and 0.1 mM NA2EDTA). Subsequently, cells were centrifuged
over a discontinuous 45, 50, 55 and 81% (vol/vol) isotonic percoll gradient. Neutrophils
obtained from the 55/81% interface were typically >90% pure as determined by May-
Grünwald-Giemsa staining. 107 neutrophils were injected into the tail vein 2-3 hours before
influenza infection.

Sample collection
Cells were obtained by bronchoalveolar lavage (BAL), and from spleen and lungs. For BAL, the
thorax was opened and the trachea exposed. A small aperture was made in the trachea and
a 1.3mm cannula from a BD Insyte Cathether (BD, San Jose, CA, USA) was inserted into the
trachea. Subsequently, BAL was obtained by attaching a syringe to the cannula and washing
the lungs twice with 500 μl PBS.
Lungs were perfused to remove blood from by cutting the vena cava and injecting 4ml
of PBS into the right ventricle of the heart. Right lungs were cut into small pieces and digested
for 30 minutes at 37°C with 0.6 mg/ml DNAse I and 1.6 mg/ml collagenase (both from Roche,
Basel, Switzerland) in PBS. Single cell suspensions were obtained from lungs by passing
samples through a 100 μm cell strainer. Left lungs were placed in metal air-tight containers
and snap-frozen in liquid nitrogen for qPCR analysis (see below).
Spleen single cell suspensions were obtained by mechanically passing spleens through
a 100 μm cell-strainer, followed by RBC lysis with isotonic ammoniumchloride solution as
described above.

Analysis of BAL samples

Cell counts in BAL samples were determined using a Z1 Coulter counter (Beckman Coulter,
Brea, CA, USA). BAL protein quantification was performed on the supernatant of centrifuged
BAL fluid with a Pierce BCA kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the
manufacturer’s instruction. 570nm absorbance was measured in a FLUOstar OPTIMA plate
reader (BMG LABTECH, Offenburg, Germany). Protein concentrations were determined using
a standard curve of known concentrations BSA.

152
 Neutrophil CD11b mediates influenza-induced pathology

Cytospins of 1x105 cells were stained with May-Grünwald Giemsa and differential
counts were performed in a blinded manner by a trained pathologist. Neutrophils were
subdivided into banded, normally segmented and hypersegmented on their nuclear
morphology as described previously (11). Banded cells were recognised by their doughnut-
shaped nucleus without any visible segmentation, whereas hypersegmented neutrophils
were identified with a high degree of segmentation in which at least two connections
between segments were narrow wire-like structures (width <5% of the segment and >3μm
long) and segments arranged as the leaves of a clover instead of circular. Everything in
between banded and hypersegmented nuclear morphology was considered a normally
segmented neutrophil (see also figure 5b). Cytospins with <50 neutrophils were excluded
from analysis.

Histological preparations
Lung histology sections were prepared as published previously (12). In short, mice were
exsanguinated by cutting the caudal vena cava. The trachea was cannulated and the lungs
were fixed in situ by intratracheal instillation of 4% formaldehyde in PBS at a constant 25cm
fluid pressure for 5 min. After fixation, lungs were collected and stored in 4% formaldehyde
for at least 24h until further processing. Lungs were paraffin embedded by placing tissue in
increasing concentrations of ethanol, followed by xylene and lastly by paraffin. After
embedding, 3μm slides were cut in the dorsal ventral plane and transferred to glass slides.
Samples were deparaffinised by xylene, rehydrated using descending concentrations of
8
ethanol and stained with hematoxylin and eosin (H&E). After another dehydration in
ascending ethanol concentrations and xylene, cover slides were mounted with Cytoseal
(Thermo Fisher Scientific). Microscopic images were obtained using a Leica DMRXE
microscope equipped with a 5x objective (Leica Microsystems, Wetzlar, Germany). The
percentage open lung area was determined on the lower tips of each lung using ImageJ 1.47T
(see supplementary figure 1). A region of interest was made to include only lung tissue and
exclude blood vessels, broncheoli and cutting artefacts. Subsequently, images were
converted to 32-bit greyscale, a threshold was placed (identical for each image) and the
percentage of open lung area was calculated as percentage of pixels with values below the
threshold. Means of the two lungs were used for analysis.

Flow Cytometry
Single cell preparations were processed for flow cytometry in FACS buffer (PBS supplemented
with 0.1% BSA and 10μg/ml DNAse I (Roche). Cells were stained on ice with antibodies (see
supplementary table 1) for 30 minutes. Cells were washed twice, fixed for 2min with 1%
formaldehyde, washed and resuspended in FACS buffer. Samples were analysed on a Gallios
flow cytometer (Beckman Coulter). A known number of TrueCount beads (Life technologies)
was added to each sample, thereby allowing determination of absolute cell numbers. T-cells
were identified as SSClow, CD3+, and either CD4 or CD8 single positive. Neutrophils were
identified as SSCintermediate, Ly6Ghigh and autofluorescencenegative (see supplementary figure S2).

153
Chapter 8

Antibodies used were CD3ε-FITC clone eBio500A2, CD11b-PB clone M1/70 (both
eBiosciences, San Diego, CA, USA), CD8a-PO clone 5H10 (Thermo Fisher Scientific, ), CD4-PB
clone RM4-5, CD62L-PC7 clone MEL-14 and Ly6G clone 1A8 (BD Pharmingen, San Jose, CA,
USA).

Quantitative PCR
Relative viral loads were determined on snap-frozen left lungs as published previously (13).
In short, RNA was isolated using a Roche MagNALyser (Roche, Basel, Switzerland).
Subsequently, cDNA was created from isolated DNA and analysed in duplo on a Bio-Rad MyiQ
PCR machine (Bio-rad, Hercules, CA, USA). Viral RNA content was corrected with GAPDH RNA
using the delta-delta CT method (14). The following primers were used: Influenza forward
AAG ACC AAT CCT GTC ACC TCT GA, reverse CAA AGC GTC TAC GCT GCA GTC C, GAPDH
forward GCA CAG TCA AGG CCG AGA AT and reverse GCC TTC TCC ATG GTG GTG AA (15, 16).

Statistical analysis
Calculations of statistical analysis were performed using Prism 6.07 (GraphPad Software, La
Jolla, CA, USA). Timecourse experiments were compared by 2-way ANOVA, whereas
measurements were tested by Mann-Whitney U test or Kruskall-Wallis analysis of variance
where applicable. All p-values were corrected for multiplicity where applicable. Flow
cytometric data was analysed with FCS Express V4.07 (De Novo Software, Los Angeles, CA,
USA).

Results
CD11b-/- mice display increased pathology after influenza infection
To investigate the role of CD11b in influenza infection, WT and CD11b-/- mice were infected
intranasally with various doses of influenza A/HK/2/68 (H3N2) virus. Low doses of influenza
virus (0.015 and 0.022 HAU) (Figure 1A-B) did not induce pathology in either WT and CD11b-
/- as determined by monitoring the bodyweight of the animals. At a dose of 0.03 HAU (Figure

1C), CD11b-/- mice showed a marked loss of bodyweight starting at day 5-6 post-infection,
whereas in WT animals, only a small decrease in weight was observed. At the highest dose of
0.15 HAU (Figure 1D) a more severe weight loss was observed in both WT and CD11b-/- groups.
At day 8, the weight loss was so severe that two CD11b-/- mice lost more than 20% weight and
had to be euthanized according to local regulations. In all groups, mice that were not
euthanized recovered spontaneously from infection at day 9. These data indicate that CD11b-
/- mice have more severe disease after influenza infection.

154
 Neutrophil CD11b mediates influenza-induced pathology

8
Figure 1. Weight loss after infection with different doses of influenza virus. CD11b-/- mice display increased
weight loss after influenza infection at a dose of 0.15 HAU. At a dose of 0.03 HAU, only the CD11b-/- mice lose
weight and at lower doses neither the WT or CD11b-/- show signs of disease. N=8 per group, pooled data from
two separate experiments. Error bars indicate SEM, * indicates a difference between WT and CD11b-/- with
p<0.05 as determined by two-way ANOVA after correction with Sidak’s multiple comparisons test.

Increased weight loss in CD11b-/- mice is accompanied by lung damage, leukocyte

infiltration, but not by viral load
At day 8 post-infection, mice were sacrificed and bronchoalveolar lavage (BAL) or whole lungs
were harvested for analysis. Microscopic analysis of in situ fixed lungs revealed an increased
leukocyte infiltration in CD11b-/- mice compared to WT (Figure 2A) accompanied with a
decrease in the percentage lung lumen area on lung histology sections (Figure 2B). Increased
leukocyte infiltration was confirmed by FACS analysis as increased neutrophil infiltration into
the lungs (Figure 2C) and an increased cell count in the BAL fluid (Figure 2D). Differential
counts (Figure 2E) revealed this increase to be mainly due to neutrophil and lymphocyte
infiltration, without changes in the absolute number of alveolar macrophages. To quantify the
extent of lung damage, we determined the protein content in the BAL fluid as lung damage
leads to protein leakage from the bloodstream into the lungs. The extent of lung damage to
was greater in CD11b-/- mice than in WT mice (Figure 2F). Despite the large differences in
pathology, no differences were detected in viral load, indicating that CD11b does not seem
to play a role in the anti-viral response and that differences in pathology are not due to direct
damage evoked by the virus.

155
Chapter 8

Figure 2. Pathology after influenza infection. CD11b-/- mice show more pathology 8 days post-infection. (A)
Representative H&E slides of the lower portion of the lungs show more leukocyte infiltration and inflammation.
(B) quantitation of H&E slides shows a decreased lung lumen in the CD11b-/- mice. Median values of both lungs,
N=4 per group, objective used: 5x. (C) Neutrophil infiltration in the lungs is increased at both day 3 and day 8
post-infection N=4-8. (D) total leukocyte infiltration in the lungs was determined by counts of the BAL fluid, N=8.
(E) differential counts of BAL leukocytes show the increased infiltration in the CD11b-/- mice to be neutrophils
and lymphocytes, N=3 (WT) or N=8 (CD11b-/-). (F) Protein content was taken as a measure of lung damage
(N=14). (G) Viral loads did not differ between the two groups, N=2/4 (day 3) & N=9/12 (Day8). Data are pooled
from 1 (B), 2 (C-E) or 3 (F,G) separate experiments. *indicates p<0.05 and ** p<0.01 as determined by Mann-
Whitney U-test with Holm-Sidak correction for multiplicity where required (D).

156
 Neutrophil CD11b mediates influenza-induced pathology

A role for T-cells in influenza induced pathology

Previous publications have shown influenza-induced pathology to be T-cell dependent (9). To
determine whether this is also the case in CD11b-/- mice, both WT and CD11b-/- mice were
treated with CD4 and CD8 T-cells depleting antibodies or PBS 1 day prior to Infection and at 1
day post-infection. Differential counting of BAL samples (Figure 3A) revealed a decrease in
lymphocyte numbers in the CD4/CD8 depleted CD11b-/- mice, whereas lymphocyte numbers
were already low in untreated WT animals without a decrease after T-cell depletion. Flow
cytometric analysis of digested lungs (Figure 3B) and spleens (Figure 3C) showed significant
reductions of both CD4+ and CD8+ T-cell numbers in both genotypes treated with depleting
antibodies (75 to 1200-fold reduction). T-cell depletion had no further effect on the already
low weight loss in WT animals (Figure 3D), but CD11b-/- mice showed a partial recovery from
disease after T-cell depletion (Figure 3F), indicating that the influenza-induced pathology in
these mice is also T-cell dependent.

Figure 3. Pathology is mediated by T-cells in both WT and CD11b-/-. Injection of depleting antibodies against
CD4 and CD8 at 2 days before and one day after influenza infection reduces the number of lymphocytes in (A)
the BAL and T-cell numbers in both lung (B) and spleen (C). (D) T-cell depletion has no effect on weight loss in
WT animals, but limits weight loss in CD11b-/- animals (E). Cell numbers indicate means +/- SD with N=3-4 (A-C).
Time courses represent mean +/- SEM with N=5-8 from two separate experiments. *, **, *** and **** indicate
p<0.05, 0.01, 0.001 and 0.0001, respectively as determined by one-way ANOVA with Holm-Sidak’s multiple
comparisons test (A-C) or two-way ANOVA with Sidak’s multiple comparisons test (E).

157
Chapter 8

T-cell numbers during influenza infection

As we have shown that T-cells play a role in influenza-induced pathology , we measured T-cell
numbers during the course of influenza infection. At day 3 post-infection both CD4 and CD8
T-cell counts were higher in the lungs of CD11b-/- mice compared to WT animals (Figure 4A).
However, lung T-cell counts were similar later (day 8) during infection (Figure 4B), when the
largest difference in pathology was observed. This difference was specific for the lungs, as T-
cell counts were similar in the spleen at both timepoints (Figure 4C-D) and might indicate that
differences in T-cell numbers early during infection can lead to differences in pathology at
later timepoints.

Figure 4. T-cell numbers after infection. At day 3 post-infection, an in-creased amount of both CD4 and CD8 T-
cells was seen in the lungs (A), but not in the spleens (C) of CD11b-/- mice. At day 8 post-infection, no differences
were observed in either the lungs (B) or spleens (D). N=3-4 (day 3) or 7-8 (day 8). * indicates p<0.05 as
determined by Mann-Whitney U-test.

Neutrophil phenotype
Acute inflammation in humans is accompanied by the recruitment of immune
hypersegmented neutrophils, which have been shown to be capable of suppressing T-cell
proliferation in a CD11b dependent manner (8). So far, hypersegmented neutrophils have
been described in the blood during acute inflammation (8), and are found in synovial fluid of
inflamed joints in juvenile idiopathic arthritis (unpublished data). As T-cell numbers were

158
 Neutrophil CD11b mediates influenza-induced pathology

increased in CD11b-/- animals, we determined neutrophil number in the lungs and spleen and
determined the nuclear morphology of BAL neutrophils at the day at which the most weight
loss was observed (day 8 post-infection). Neutrophil numbers were increased in the lungs of
CD11b-/- mice both at day 3 and day 8 post-infection (Figure 2C) and at day 8 post-infection in
the BAL (Figure 2D). Besides an increase in number, differential count of BAL neutrophils
revealed a higher percentage of hypersegmented neutrophils in CD11b-/- mice (Figure 5A).

Figure 5. Neutrophil nuclear morphology.

(A) Cytospins of BAL cells stained with
May-Grünwald-Giemsa were assessed for
the presence of (left to right) banded,
normally segmented or hypersegmented
nuclear morphology. WT mice had a higher
percentage banded neutrophils whereas
CD11b-/- mice had more hypersegmented
neutrophils. (B) Representative examples
of banded, normally segmented and
hypersegmented neutrophils. Objective
used: 100x, error bars indicate 5μm. Data
represent means +/- 95% CI, N=4-8 from
two separate experiments. * and **
indicate p<0.05 and 0.01 as assessed by
Mann-Whitney U test corrected for 8
multiplicity with Holm’s sequential
Bonferroni method.

CD11b expressing neutrophils are responsible for protection against influenza-induced

pathology
As shown above, lack of CD11b expression lead to increased pathology after influenza
infection, which we hypothesized to be due to neutrophil CD11b-mediated responses.
However, CD11b is also expressed on other myeloid cells, such as eosinophils, monocytes,
dendritic cells and macrophages. Therefore, we set out experiments to rescue CD11b-/- mice
from disease by adoptive transfer of 107 WT neutrophils 2h prior to infection. To exclude the
effect of just additional neutrophils being present, the control group received a similar
amount of CD11b-/- neutrophils. Adoptive transfer of WT neutrophils decreased weight loss
at day 8 post-infection compared to CD11b-/- mice either with or without receiving CD11b-/-
neutrophils (p=0.03 and p=0.0002, respectively). The rescue from pathology does not seem
to be complete, as WT mice have a slightly higher weight at day 8 post-infection compared to
WT transferred CD11b-/- mice (97.8% compared to 94.5% of day 0), but this difference was
not significant (p=0.33).
The observed effect was present only in animals infected under ether anaesthesia.
Animals anaesthetized with isoflurane no rescue by adoptive transfer of WT neutrophils,
albeit differences between WT and CD11b-/- without transfer were also smaller.

159
Chapter 8

WT vs CD11b-/-
A Ether B Isoflurane
WT
CD11b-/-
Weight (% of day 0)

100
*

90

80
0 2 4 6 8
Time post-infection (d)

Adoptive transfer
C D Isoflurane
Ether Isoflurane
KO+WT -/-+WT
CD11b
Weight (% of day 0)

100 KO+KO -/-+CD11b-/-

CD11b

90

80
0 2 4 6 8
Time post infection (d)

Figure 6. Adoptive transfer of WT neutrophils rescues from disease, but is dependent on anaesthesia. WT and
CD11b-/- mice were infected with influenza under sedation with either ether or isoflurane. The difference in
weightless between WT and CD11b-/- is more pronounced in animals sedated with ether during infection (A)
than the isoflurane anesthetised group (B), despite a 1.5x higher dose for the isoflurane anesthetised animals.
IV injection of 106 WT neutrophils 3h before infection decreased the pathology in CD11b-/- mice (C), whereas
adoptive transfer of CD11b-/- neutrophils did not have any effect. However, this effect was not observed in mice
infected under isoflurane sedation (D). Data represent means +/- SEM, with pooled data of 5 experiments (N=17-
20) (A,C) or pooled data from 2 individual experiments (N=7-8) (B,D). *, **, *** and **** indicate p<0.05, 0.01,
0.001 and 0.0001, respectively, as determined by 2-way ANOVA corrected for multiplicity by Sidak’s multiple
comparison test.

Discussion
Infection by influenza virus causes pathology that is usually mild. However, in rare cases an
exaggerated immune response is found that can lead to ARDS (3). This is associated with high
morbidity and mortality (3). In addition, some influenza viruses are highly pathogenic, with
the 1918 Spanish Flu (H1N1) being the most extreme case (17). Indeed, many studies have
shown that a runaway immune response can cause the influenza-induced pathology (2).
We demonstrate that a lack of CD11b leads to more severe pathology after influenza
infection in mice and that adoptive transfer of WT neutrophils rescues CD11b-/- mice from

160
 Neutrophil CD11b mediates influenza-induced pathology

pathology. Previous studies examining the role of neutrophils in influenza infection have
relied on the depletion of neutrophils (3,5) or both neutrophils and monocytes (18, 19) by
application of depleting antibodies. These studies all show more influenza-induced pathology
after high doses of neutrophil depleting antibodies. However, neutrophil depletion can have
serious side-effects: neutrophil-targeting antibodies have been shown in humans to be
capable of causing immune mediated pathology that potentially leads to fatal transfusion-
related lung injury (TRALI) (20, 21). In mice, neutrophil depletion in the context of TNF-α-
induced inflammation leads to death within 1 hour (22), which raises the question whether
the observed effect of neutrophil depletion in influenza infection might be due to depletion-
induced inflammation Our study relied on the complete absence CD11b instead of neutrophil
depletion and, therefore, provides more robust evidence of the involvement of neutrophils
in suppression of influenza-induced pathology.
CD11b has been shown to be involved in neutrophil extravasation, but a lack of CD11b
does not influence neutrophil extravasation into the lungs (23, 24). This was confirmed in our
data, which showed increased neutrophil homing into the lungs and BAL of KO mice (figure
2C and E). The increased neutrophil infiltration in CD11b-/- mice might be caused by the
increased lung damage and inflammation. Alternatively, as a large number of infiltrating
neutrophils is hypersegmented, the increased infiltration might reflect infiltration of immune-
suppressive cells which are unable to limit inflammation due to their lack of CD11b.
Neutrophils have been shown to phagocytose influenza particles in vitro (25, 26), CD11b
is critical for phagocytosis of opsonized particles (27) and CD11b has a long list of other ligands
8
(28). Therefore, it is possible that CD11b is required for viral clearance. However, we do not
detect an increase in viral load for the CD11b-/- mice compared to controls, making a defect
in viral killing unlikely to be causative of differences in pathology. This finding is agreement
with a previous study in which neutrophil depletion did not affect viral loads (2). On the other
hand, Tate et al found that control animals had cleared all virus at day 8 post-infection,
whereas in neutrophil-depleted animals were deficient (6). Along similar lines, Wareing et al.
showed neutrophil recruitment to the lungs is not required for viral clearance (29), making a
direct interaction of neutrophil CD11b with viral particles an unlikely cause for the increased
pathology in CD11b-/- mice.
We propose that CD11b prevents influenza-induced pathology by suppressing T-cell
proliferation for the increased pathology is accompanied with increased T-cell numbers in the
lungs 3 days post-infection (figure 4). This likely concerns the bystander T-cell activation (30)
instead of influenza-specific T-cells, because viral loads are not different between WT and
CD11b-/- mice. Interestingly, the higher number and percentage of hypersegmented
neutrophils in the BAL of CD11b-/- mice suggests that severe pathology is followed by more
recruitment of hypersegmented neutrophils to control runaway immune activation.
Differences in tissue T-cell numbers are observed at day 3 post infection, whereas
differences in weight loss are seen only at day 8. In addition, adoptively transferred WT
neutrophils were not found in the lungs or spleen of CD11b-/- mice at day 8 post-infection
(data not shown), which further suggests that the difference in pathology originates from

161
Chapter 8

early on in the infection. Repeated infusions of WT neutrophils might be able to reduce T-cell
numbers at later time points and further limit pathology.
Differences in pathology were observed between animals anesthetised with ether or
isoflurane. To reach a similar extend of weight loss, a 1.5 times higher viral load was
administered in isoflurane anesthetised mice (figure 6). Similar findings have been described
previously in a different infection model (31) and might be accountable to either the immune-
suppressive effects of isoflurane (32, 33) or the lung damage induced by ether. In isoflurane
anaesthetized CD11b-/- mice, adoptive transfer of WT neutrophils did not reduce pathology.
Possibly, the damage induced by ether causes neutrophils to home to the lungs where they
can affect the immune response early in infection. On the other hand, the difference in
pathology between WT and CD11b-/- mice was also smaller in animals infected under
isoflurane anaesthesia. The lack of an observed effect of adoptive transfer may also have been
due to a smaller treatment window.
In conclusion, we demonstrated neutrophil CD11b to be essential in preventing
pathology during infection with non-lethal doses of influenza virus. This effect on pathology
was not mediated by increased viral clearance as viral loads were not affected. CD11b-/- mice
have increased T-cell numbers early in infection and hypersegmented neutrophils were found
in the BAL during infection. Therefore, we propose CD11b-mediated suppression of T-cell
responses as preventing pathology during influenza infection. In addition, activating CD11b
likely prevents pathology, thereby offering an interesting therapeutic target.

162
 Neutrophil CD11b mediates influenza-induced pathology

References
1. Scherle PA, Palladino G, and Gerhard W. Mice can recover from pulmonary influenza
virus infection in the absence of class I-restricted cytotoxic T cells. J Immunol.
1992;148(1):212-7.
2. Brandes M, Klauschen F, Kuchen S, and Germain RN. A systems analysis identifies a
feedforward inflammatory circuit leading to lethal influenza infection. Cell.
2013;154(1):197-212.
3. de Jong MD, Simmons CP, Thanh TT, Hien VM, Smith GJ, Chau TN, Hoang DM, Chau
NV, Khanh TH, Dong VC, et al. Fatal outcome of human influenza A (H5N1) is associated
with high viral load and hypercytokinemia. Nature medicine. 2006;12(10):1203-7.
4. Geerdink RJ, Pillay J, Meyaard L, and Bont L. Neutrophils in respiratory syncytial virus
infection: A target for asthma prevention. The Journal of allergy and clinical
immunology. 2015;136(4):838-47.
5. Snelgrove RJ, Edwards L, Rae AJ, and Hussell T. An absence of reactive oxygen species
improves the resolution of lung influenza infection. European journal of immunology.
2006;36(6):1364-73.
6. Tate MD, Deng YM, Jones JE, Anderson GP, Brooks AG, and Reading PC. Neutrophils
ameliorate lung injury and the development of severe disease during influenza
infection. J Immunol. 2009;183(11):7441-50.
7. Gabrilovich DI, and Nagaraj S. Myeloid-derived suppressor cells as regulators of the 8
immune system. Nature reviews Immunology. 2009;9(3):162-74.
8. Pillay J, Kamp VM, van Hoffen E, Visser T, Tak T, Lammers JW, Ulfman LH, Leenen LP,
Pickkers P, and Koenderman L. A subset of neutrophils in human systemic
inflammation inhibits T cell responses through Mac-1. The Journal of clinical
investigation. 2012;122(1):327-36.
9. Rygiel TP, Rijkers ES, de Ruiter T, Stolte EH, van der Valk M, Rimmelzwaan GF, Boon L,
van Loon AM, Coenjaerts FE, Hoek RM, et al. Lack of CD200 enhances pathological T
cell responses during influenza infection. J Immunol. 2009;183(3):1990-6.
10. Kim ND, Chou RC, Seung E, Tager AM, and Luster AD. A unique requirement for the
leukotriene B4 receptor BLT1 for neutrophil recruitment in inflammatory arthritis. J
Exp Med. 2006;203(4):829-35.
11. Pillay J, Tak T, Kamp VM, and Koenderman L. Immune suppression by neutrophils and
granulocytic myeloid-derived suppressor cells: similarities and differences. Cellular
and molecular life sciences : CMLS. 2013.
12. Braber S, Verheijden KA, Henricks PA, Kraneveld AD, and Folkerts G. A comparison of
fixation methods on lung morphology in a murine model of emphysema. American
journal of physiology Lung cellular and molecular physiology. 2010;299(6):L843-51.
13. Hendriks J, Xiao Y, Rossen JW, van der Sluijs KF, Sugamura K, Ishii N, and Borst J. During
viral infection of the respiratory tract, CD27, 4-1BB, and OX40 collectively determine
formation of CD8+ memory T cells and their capacity for secondary expansion. J
Immunol. 2005;175(3):1665-76.

163
Chapter 8

14. Rao X, Huang X, Zhou Z, and Lin X. An improvement of the 2^(-delta delta CT) method
for quantitative real-time polymerase chain reaction data analysis. Biostat Bioinforma
Biomath. 2013;3(3):71-85.
15. Ward CL, Dempsey MH, Ring CJ, Kempson RE, Zhang L, Gor D, Snowden BW, and
Tisdale M. Design and performance testing of quantitative real time PCR assays for
influenza A and B viral load measurement. J Clin Virol. 2004;29(3):179-88.
16. Pauklin S, and Petersen-Mahrt SK. Progesterone inhibits activation-induced
deaminase by binding to the promoter. J Immunol. 2009;183(2):1238-44.
17. Taubenberger JK, and Morens DM. The pathology of influenza virus infections. Annu
Rev Pathol. 2008;3(499-522.
18. Fujisawa H. Neutrophils play an essential role in cooperation with antibody in both
protection against and recovery from pulmonary infection with influenza virus in mice.
J Virol. 2008;82(6):2772-83.
19. Tate MD, Brooks AG, and Reading PC. The role of neutrophils in the upper and lower
respiratory tract during influenza virus infection of mice. Respiratory research.
2008;9(57.
20. Santamaria A, Moya F, Martinez C, Martino R, Martinez-Perez J, and Muniz-Diaz E.
Transfusion-related acute lung injury associated with an NA1-specific antigranulocyte
antibody. Haematologica. 1998;83(10):951-2.
21. Curtis BR, and McFarland JG. Mechanisms of transfusion-related acute lung injury
(TRALI): anti-leukocyte antibodies. Critical care medicine. 2006;34(5 Suppl):S118-23.
22. Abbitt KB, Cotter MJ, Ridger VC, Crossman DC, Hellewell PG, and Norman KE. Antibody
ligation of murine Ly-6G induces neutropenia, blood flow cessation, and death via
complement-dependent and independent mechanisms. Journal of leukocyte biology.
2009;85(1):55-63.
23. Mizgerd JP, Kubo H, Kutkoski GJ, Bhagwan SD, Scharffetter-Kochanek K, Beaudet AL,
and Doerschuk CM. Neutrophil emigration in the skin, lungs, and peritoneum:
different requirements for CD11/CD18 revealed by CD18-deficient mice. J Exp Med.
1997;186(8):1357-64.
24. Lu H, Smith CW, Perrard J, Bullard D, Tang L, Shappell SB, Entman ML, Beaudet AL, and
Ballantyne CM. LFA-1 is sufficient in mediating neutrophil emigration in Mac-1-
deficient mice. The Journal of clinical investigation. 1997;99(6):1340-50.
25. Ratcliffe D, Migliorisi G, and Cramer E. Translocation of influenza virus by migrating
neutrophils. Cell Mol Biol. 1992;38(1):63-70.
26. Yamamoto K, Suzuki K, Suzuki K, and Mizuno S. Phagocytosis and ingestion of influenza
virus by human polymorphonuclear leucocytes in vitro: electronmicroscopy studies. J
Med Microbiol. 1989;28(3):191-8.
27. van Spriel AB, Leusen JH, van Egmond M, Dijkman HB, Assmann KJ, Mayadas TN, and
van de Winkel JG. Mac-1 (CD11b/CD18) is essential for Fc receptor-mediated
neutrophil cytotoxicity and immunologic synapse formation. Blood. 2001;97(8):2478-
86.

164
 Neutrophil CD11b mediates influenza-induced pathology

28. Solovjov DA, Pluskota E, and Plow EF. Distinct roles for the alpha and beta subunits in
the functions of integrin alphaMbeta2. The Journal of biological chemistry.
2005;280(2):1336-45.
29. Wareing MD, Shea AL, Inglis CA, Dias PB, and Sarawar SR. CXCR2 is required for
neutrophil recruitment to the lung during influenza virus infection, but is not essential
for viral clearance. Viral Immunol. 2007;20(3):369-78.
30. Tough DF, Borrow P, and Sprent J. Induction of bystander T cell proliferation by viruses
and type I interferon in vivo. Science. 1996;272(5270):1947-50.
31. Eory ML, Zanuzzi CN, Fuentealba NA, Sguazza GH, Gimeno EJ, Galosi CM, and Barbeito
CG. Effects of different anesthetics in the murine model of EHV-1 infection. Vet Pathol.
2013;50(5):849-56.
32. Markovic SN, Knight PR, and Murasko DM. Inhibition of interferon stimulation of
natural killer cell activity in mice anesthetized with halothane or isoflurane.
Anesthesiology. 1993;78(4):700-6.
33. Anderson SL, Duke-Novakovski T, and Singh B. The immune response to anesthesia:
part 1. Vet Anaesth Analg. 2014;41(2):113-26.

165
Chapter 8

Supplementary figure S1. Image analysis. (A) The percentage of lung lumen area was determined on H&E
stained slides of in situ fixed lungs. (B) Colour images were converted to 8-bit greyscale. (C) An ROI was made to
exclude large blood vessels, bronchi and possible cutting artefacts. (D) A threshold was set (same value for each
image) and the percentage open area was determined using ImageJ’s “analyse particles” function without
limitations on size and circularity. For each mouse, the lower tips of both lungs were analysed, the means of
which were plotted in figure 2B.

166
 Neutrophil CD11b mediates influenza-induced pathology

Supplementary figure S2. Gating strategy. Lungs digested with collagenase and DNAse I were homogenised and
analysed by flow cytometry. (A) T-cells were identified as SSClow/CD3+ and further subdivided into CD4+ and CD8+
T-cells. (B) Neutrophils were identified as SSCintermediate/Ly6Ghigh/autofluorescencelow in both WT and CD11b-/-
mice. A similar gating strategy was used for detection of T-cells and neutrophils in the spleen.

167
 Part III

Kinetics in chronic pulmonary infection

 Chapter 9
Similar activation state of neutrophils in
sputum of asthma patients irrespective of
sputum eosinophilia

Tamar Tak1, Bart Hilvering1, Kiki Tesselaar2, Leo Koenderman1

1. Department of Respiratory Medicine and 2Immunology, Laboratory of

Translational Immunology, University Medical Centre Utrecht,
the Netherlands
2. Department of Immunology, Laboratory of Translational Immunology,
University Medical Centre Utrecht, the Netherlands

Clin Exp Immunol. 2015 Nov;182(2):204-12. doi: 10.1111/cei.12676.

Epub 2015 Sep 16
Chapter 9

Abstract
Background: Inflammatory phenotypes of asthma are associated with differences in
disease characteristics. It is unknown whether these inflammatory phenotypes are reflected
by the activation status of neutrophils in blood and sputum.
Methods: We obtained peripheral blood and induced sputum from 21 asthma
patients and stratified our samples based on sputum eosinophilia resulting in two groups
(>3% eosinophils: n=13, <3%: n=8). Eosinophils and neutrophils from blood and sputum were
analysed for expression of activation- and degranulation markers by flow cytometry. Data
were analysed by both classical, non-parametric statistics and a multi-dimensional approach,
using principal component analysis (PCA).
Results: Patients with sputum eosinophilia were characterised by increased ACQ
scores and blood eosinophil counts. Both sputum neutrophils and eosinophils displayed an
activated and degranulated phenotype compared to cells obtained from blood. Specifically,
degranulation of all granule types were detected in sputum cells, combined with an increased
expression of the activation markers (activated) Mac-1 (CD11b), Programmed Death-Ligand 1
(CD274) and a decreased expression of CD62L. CD69 expression was only increased on
sputum eosinophils. Surface marker expression of neutrophils was similar in the presence or
absence of eosinophilia, either by single or by multi-dimensional analysis.
Conclusion: Sputum neutrophils were highly activated and degranulated irrespective
of sputum eosinophilia. Therefore, we conclude that differences in granulocyte activation in
sputum and/or blood are not associated with clinical differences in the two groups of asthma
patients. The finding of PD-L1 expression on sputum granulocytes suggests an immuno-
modulatory role of these cells in the tissue.

172
 Granulocyte activation state in asthma phenotypes

Background
An estimated 300 million individuals worldwide are affected by asthma.(1) Asthma is
characterized by airway inflammation, bronchial hyperresponsiveness and reversible airway
obstruction. Several different inflammatory phenotypes have been identified in asthma,
which are accompanied by different clinical characteristics. (2) Sputum eosinophilia is
associated with bronchial hyperresponsiveness, high FeNO and (specific) IgE levels. In
addition, the presence of either neutrophils or eosinophils in sputum is associated with a
decreased FEV1 (3, 4). These inflammatory phenotypes are identified by microscopically
evaluating the percentages of neutrophils and eosinophils in sputum or by transcriptomic
profiling of whole sputum samples.(2, 5, 6) In addition to microscopic evaluation, cells
obtained by sputum induction can be analysed by flow cytometry.(7-11) Sputum granulocytes
analysed by this technique have been shown to display an activated phenotype, with
upregulation of markers for activation and degranulation on either neutrophils,(11)
eosinophils(12) or both.(13) However, up to date there are no reports on in depth analysis of
activation markers comparing blood and sputum granulocytes using multi-dimensional
analysis. In addition, no studies have been published that adequately compare sputum
granulocyte expression profiles between patients with different asthma phenotypes. A single
study did compare the expression levels of two classical activation markers between patients
with moderate and severe asthma, but did not find a correlation between expression and
disease severity.(13) Transcriptomic profiling of whole blood and sputum samples showed 8
upregulation of neutrophil defensins and proteases in the blood of neutrophilic asthma
patients, and significant differences between sputum samples of patients with different
asthma phenotypes. (6, 14) It is unknown, however, whether these differences reflect 9
differences in expression of the granulocytes themselves or differences in cell numbers.
Traditionally, granulocyte activation markers include adhesion receptors such as the
integrin Mac-1 (CD11b/CD18) and L-selectin (CD62L). Other activation markers described to
be upregulated on circulating or sputum granulocytes from asthma patients are the CD11b
activation epitope CBRM1/5 (15) and the Intercellular Adhesion Molecule-1 (ICAM-1, CD54);
the latter adhesion receptor was proposed as a therapeutic target for antigen-induced acute
airway inflammation.(16) The activation marker CD69, which is well known as a T-cell
activation marker(17), was only shown to be upregulated on eosinophils from
broncheoalveolar lavage (BAL) when compared to blood cells.(15)
Neutrophils and eosinophils possess different types of granules containing
antimicrobial and pro-inflammatory proteins. For neutrophils, release of these proteins to the
outside of the cell (degranulation) occurs sequentially in response to increasing strength of
activation signals, with secretory vesicles degranulating by the most mild stimulus, followed
by tertiary, specific and azurophilic granules.(18-20) The marker for neutrophil tertiary- and
eosinophil secretory granules (CD11b) was shown to be upregulated on cells from both BAL
and sputum, compared to blood granulocytes.(15) Markers for specific and
azurophilic/crystalloid granules were shown to be upregulated on BAL and on sputum
granulocytes in several diseases, but so far not on sputum cells from asthma patients.(15, 21,

173
Chapter 9

22) Degranulation of neutrophil secretory vesicles was shown to occur already in the blood of
asthma patients.(23, 24)
Another activation marker of interest, Programmed Death-Ligand 1 (PD-L1, CD274),
has been implicated in a murine model of airway hyperresponsiveness.(25) The
immunomodulatory protein CD274 is not as well characterized in granulocytes as in
lymphocytes, but was shown to responsible for suppression of T-cell responses by interferon-
γ stimulated neutrophils (26) and is highly expressed in granulomas in the lungs of sarcoidosis
patients, although its expression on granulocytes was not tested by flow cytometry.(27)
Up to now, all immunophenotyping of sputum cells has been performed by analysis of
single-dimensional marker expression, which ignores the interaction between multiple
markers. To overcome this issue, marker expressions and interactions can be studied using
principal component analysis (PCA). PCA is a multivariate analysis method that detects
systematic variability within multiple parameters and explores correlations between these
parameters.(28) It transforms datasets with a large number of measured parameters into a
smaller number of parameters, called principal components. As the resulting smaller number
of components is more easily interpreted, it is a preferred technique for analysis of large
datasets and forms the basis of cluster approaches used to identify asthma phenotypes.(29,
30)
In this study we investigated differences in granulocyte activation and degranulation
by flow cytometric evaluation of neutrophils and eosinophils isolated from blood and sputum
of asthma patients with and without sputum eosinophilia, both by single and
multidimensional analysis.

Methods
Study population and ethics
This study was approved by the local medical ethics committee. Patients were recruited at
the pulmonary outpatient clinic of the University Medical Centre Utrecht (table 1 patient
baseline characteristics) and gave written informed consent in accordance to the Declaration
of Helsinki (seventh revision, Fortaleza, 2013).
Patients were included according to the following criteria: age 18-75 years, having
adult asthma defined by the GINA guidelines on clinical features (episodic shortness of breath,
particularly at night and often accompanied by cough and wheezing) and reversibility of FEV1
(forced expiratory volume in 1s) upon inhalation of 400ug salbutamol (≥12% predicted or
≥200ml) and/or airway hyperresponsiveness to histamine (PC20 <8mg/ml).(1) Patient
numbers were calculated to be able to distinguish medium effect sizes with a power of 0.8
and an α of 0.05 using G*Power 3.1.3. (31)
Exclusion criteria for the study were smoking in the last twelve months, a smoking
history ≥10 pack years, treatment with antibiotics <4 weeks ago or confirmed allergic
bronchopulmonary aspergillosis.
Blood and sputum samples were stratified on presence or absence of sputum
eosinophilia based on a cut-off value of three percent sputum eosinophils, as determined by

174
 Granulocyte activation state in asthma phenotypes

differential count (described below). All samples were collected between 9AM and 1PM to
minimize effects of diurnal variation such as found in FeNO and haematological parameters.
(32, 33)

Lung function and clinical parameters

After inclusion of patients, the Asthma Control Questionnaire (ACQ) and Medication
Adherence Report Scale (MARS) were filled out. FEV1 measurements were performed by peak
flow measurement using the PiKo-1 (nSpire Health, Longmont, Co, USA) and FeNO was
determined using NIOX MINO® (Aerocrine, Solna, Sweden) with an expiration time of 10
seconds. Absolute blood eosinophil counts were calculated from percentages obtained by
FACS-analysis (see below) and WBC counts obtained froma Cell-Dyn haematology analyser
(Abbot Diagnostics, North Chicago, Il, USA).

Blood processing
Erythrocytes were lysed from sodium heparin blood using lysis buffer consisting of 150 mM
NH4Cl, 10 mM KHCO3 and 0.1 mM NA2EDTA dissolved in ddH2O. Resulting total leukocyte
preparations were washed and resuspended in staining buffer consisting of PBS
supplemented with trisodium citrate (0.32% w/v) (both prepared by the UMCU pharmacy)
and human pasteurized plasma solution (10% w/v, Sanquin, Amsterdam, the Netherlands).

Sputum induction and processing

8
Sputum was induced by inhalation of 0.9-5% saline aerosols and processed as published
previously using Sputolysin prepared from a 10x stock (Merck Millipore, Darmstadt, Germany)
and supplemented with NaCl to reach an osmolarity of 280-290mOsm.(34, 35) Cytospin slides 9
of sputum cells were stained with May-Grünwald-Giemsa for differential cell count. Samples
containing >80% squamous epithelial cells were excluded from analysis. Percentages of cells
in sputum were calculated after exclusion of squamous epithelial cells.(36) Remaining cells
were resuspended in staining buffer for FACS staining procedure.

Flow cytometry
Samples were stained with antibodies for 30 minutes on ice at a maximum concentration of
5 x 106 cells/ml and washed twice before analysis on a Gallios flow cytometer (Beckman
Coulter, Pasadena, CA, USA).
Granulocytes in blood were identified based on FSC/SSC and CD16 was used to
differentiate into neutrophils (CD16+) and eosinophils (CD16-).For sputum samples a different
sorting strategy was chosen due to the presence of epithelial cells and alveolar macrophages
with a high SSC (Figure 1). Epithelial cells do not express CD11b and alveolar macrophages are
highly autofluorescent at emission wavelengths of around 450nm when excited by a 405nm
laser. Therefore, after exclusion of debris on FSC/SSC, granulocytes were identified as CD11b+
and 405/450nm-autofluorescencelow. Subsequently, they were differentiated into neutrophils
and eosinophils based on CD16 expression and SSC. Analysis was performed on at least 250
neutrophils/eosinophils. Degranulation of neutrophils was measured using CR1 (CD35),
integrin αM (CD11b), CEACAM-8 (CD66b) and LAMP-3 (CD63) antibodies for secretory,

175
Chapter 9

tertiary, specific, and azurophilic, respectively.(19) For eosinophils, LAMP-3 was used as a
marker for crystalloid (specific/secondary) granule degranulation (37) and CD11b as a marker
for secretory (sombrero) vesicles (20). Since CD63 is also expressed on activated platelets, its
expression was only determined on cells negative for platelet marker CD41.(38)
Antibodies used were CD11b-APC-Alexa750 (clone Bear1), CD16-Krome Orange (3G8),
CD274-PeCy7 (PDL1.3.1) and CD62L-ECD (DREG56) from Beckman Coulter (Pasadena, CA,
USA), CD35-FITC (E11) from Biolegend (San Diego, CA, USA. CD63-PE (H5C6), CD69-PeCy7
(FN50), IgG1 isotype control-PE (X40), IgG2a isotype control-FITC (X39) and the Annexin-V PE
apoptosis kit I from BD (San Jose, CA, USA), active CD11b-Alexa700 (CBRM1/5) from
eBiosciences (San Diego, CA, USA), CD54-PE (MEM-111) from EXBIO Praha (Vestec, Czech
Republic) and CD41-FITC (VIPL3) from Life Technologies (Carlsbad, CA, USA).

10-3 10-2 103 104 105 106

CD16
Figure 1. Gating strategy to identify sputum granulocytes. (A) A FSC/SSC gate is used to exclude most debris
and lymphocytes. (B) Subsequently, cells positive for CD11b and negative for ~450nm autofluorescence were
gated to exclude residual lymphocytes and alveolar macrophages (which are highly autofluorescent). (C) The
resulting granulocytes were subdivided in neutrophils and eosinophils based on SSC and expression of CD16. (D)
The resulting neutrophil and eosinophil population show FSC/SSC patterns similar to those in blood, with a
higher SSC for eosinophils and higher FSC for neutrophils.

Data analysis and representation

FCS express 4.0 (De Novo Software, Los Angeles, CA, USA) was used for evaluation of flow
cytometric data and determination of median fluorescence intensities. Data were plotted as
boxes representing median +/-IQR with error bars plotted according to Tukey’s method in
Prism 6.04 (GraphPad Software, La Jolla, CA, USA). Statistical analysis was performed in SPSS
Statistics 22 (IBM, Armonk, NY, USA). Samples were compared using multiple Wilcoxon-
Mann-Whitney tests or Related-samples Wilcoxon signed-rank tests where applicable.

176
 Granulocyte activation state in asthma phenotypes

Correction for multiple testing was performed using the Bonferroni Method. Correlations
between parameters were determined using Spearman’s Correlation Coefficient.
Principal Component Analysis (PCA), was performed as described extensively
elsewhere.(28) In short, PCA was performed on the MFIs of all measured parameters (except
isotype controls) using SPSS Statistics 22. MFIs were classified as numerical data, missing
values were mode imputed and discretization was used. Scree plots (not shown) were used
to determine the required amount of components. Data points with object scores of >3.5 or
<-3.5 were considered as outliers and excluded from the analysis. Graphs of object scores and
loading plots were made in Prism 6.04.

Results
Patient characteristics
Sputum and blood samples were obtained from 21 patients (table 1 patient baseline
characteristics). Thirteen patients (68%) had more than 3% eosinophils in their sputum
samples. This sputum eosinophilia was accompanied with higher eosinophil counts in
peripheral blood (p=0.007), less well controlled asthma (p=0.035) and trends toward higher
FeNO, more medication use and lower absolute FEV1 (p=0.09, 0.09 and 0.053 respectively).

Table 1. Patient baseline characteristics at time of sputum induction grouped for sputum eosinophilia
>3% sputum eo <3% sputum eo p-value *
Number 13 8 NA 8
Gender M/F 9/4 4/4 0.88
Age (years) 45 (19-66) 47 (25-54) 0.45
BMI (kg/m2) 27 (22-35) 27 (22-32) 0.885 9
FeNO (ppb) 32 (16-215) 23 (12-42) 0.09
ACQ score 2.3 (0.0-4.4) 1.3 (0.14-2.0) 0.035
Blood Eosinophil Count (109/mL) 0.68 (0.1-1.2) 0.18 (0.0-0.5) 0.007
Medica�on † 4 (3-5) 3.5 (0-4) 0.09
FEV1 (L) 2.4 (1.5-3.9) 3.5 (2.3-4.0) 0.053
FEV1 (% predicted) 76 (61-105) 89 (61-113) 0.16
Reversibility FEV1 ‡ 0.47 (0.0-16.6) 0.43 (0.0-4.9) 0.46
Sputum eosinophils (%) § 29 (3-82) 0 (0-1) NA
Sputum neutrophils (%) § 39 (10-77) 49 (19-83) 0.41
Values are medians +/- range unless indicated otherwise.
* Based on Wilcoxon-Mann-Whitney test or Fisher’s exact test where appropriate
† 5-point ordinal scale based on guidelines of the British Thoracic Society , with 0) no medication, 1) inhaled
SABA when required 2) low dose ICS + SABA 3) low/medium dose ICS + LABA, or medium dose ICS + LABA 4) High
dose ICS +/- LABA and leukotrien receptor antagonist test and 5) High dose ICS + OCS +/- LABA
‡ Percentage reversibility in FEV1 (corrected for age, gender, length and bodyweight) a�er inhala�on of 2x
100μg salbutamol
§ As determined by microscopic evaluation
Abbreviations: BMI: body mass index, FeNO: forced exhaled nitric oxide, ACQ: Asthma Control Questionnaire,
FEV1: Forced Expiratory Volume in 1 sec, NA: Not applicable.

177
Chapter 9

Viability of sputum granulocytes

After processing of sputum samples, the viability of the cells in the sample was determined
by Annexin-V and 7-AAD staining.(39) Sputum neutrophils and eosinophils were typically 90-
95% alive (Figure 2) and 5-10% necrotic or late apoptotic. Less than 1% of sputum cells was
Annexin-V single positive and, thus, early apoptotic.

Figure 2. Viability of sputum

granulocytes. Sputum cells are typically
>90% viable, with <1% early apoptotic
Annexin-V single positive cells and the
remainder being necrotic or late
apoptotic. Bars depict means with SD
patients with (n=13) and without (n=8)
sputum eosinophilia.

Expression of activation markers in blood and sputum

Both neutrophils and eosinophils showed an activated phenotype in sputum compared to
blood (Figure 3 and Supplementary Figure 1). CD62L was shed from sputum neutrophils and
eosinophils. An increased CD69 expression was found only on sputum eosinophils, whereas
CD11b, CBRM1/5 and CD274 were upregulated on both cell types. Expression of CD54 was
not detected, irrespective of cell or sampling location. Increased expression of markers for
tertiary (CD11b), specific (CD66b) and azurophilic/crystalloid (CD63) granules indicated a
highly degranulated state for sputum granulocytes.
In contrast to the marked differences in expression patterns between blood and
sputum granulocytes, no significant differences were detected when comparing expression
levels of markers on blood and sputum cells isolated from individuals with high and low
eosinophil counts in sputum (Figure 4). In addition, no significant correlations were found
between ACQ score, FEV1 (percentage predicted and/or absolute value) and expression of
any of the granulocyte markers (data not shown). Due to the absence or small numbers of
eosinophils in patients with <3% sputum eosinophils surface marker expressions of these cells
could not be sufficiently compared between patient groups.

178
 Granulocyte activation state in asthma phenotypes

Figure 3. Expression profiles of blood granulocytes compared to sputum. Expression of activation markers on
blood and sputum neutrophils (A-C, N=21) and eosinophils (E-F, N=13). Since CD11b is both a classical activation
marker and a marker for degranulation of tertiary granules, it is displayed twice. Boxes represent medians +/-
IQR with whiskers of 1.5 IQR as according to Tukey’s method. Light grey fills represent data points below the
isotype control median measured on blood granulocytes.
* indicates p<0.05, ** p<0.001 significant differences between blood and sputum as determined by multiple
Wilcoxon signed-rank test corrected for multiplicity by Bonferroni correction.

179
Chapter 9

Figure 4. Expression profiles of sputum granulocytes in patients with presence of absence of sputum
eosinophilia. Expression profiles of (A) blood neutrophils, (B) blood eosinophils and (C) sputum neutrophils are
similar in patients with presence or absence of sputum eosinophilia. Boxes represent medians +/- IQR with
whiskers of 1.5 IQR, as according to Tukey’s method. Light grey fills indicate values below the median isotype
control. No statistically significant differences were found as determined by multiple Wilcoxon-Mann-Whitney
tests with Bonferroni correction (p<0.05 after multiplicity correction).

180
 Granulocyte activation state in asthma phenotypes

Principal component analysis (PCA)

PCA was performed on the neutrophil data and eosinophil data for the patients with adequate
numbers of sputum eosinophils (Figure 5). Scree plots (not shown) indicated that all PCAs
comparing blood and sputum samples required 2 components and the PCA comparing the
two asthma phenotypes required 4 components.
Principal component 1 accurately separates samples from blood and sputum for both
neutrophils (Figure 5A) and eosinophils (Figure 5B), demonstrating that receptor expression
profiles differ between sampling locations. The first principal component accounted for a
total of 55% and 52% of the variation in receptor expression for neutrophils and eosinophils,
respectively. Principal component 2 described a further 16% and 23% of the variation.
In contrast, none of the four principal components distinguished between patients
with or without eosinophils in their sputum after a PCA on the combined data of blood
eosinophils, blood neutrophils and sputum neutrophils (Figure 5C). Separate PCA of surface
marker data for blood eosinophils, blood neutrophils and sputum neutrophils did also not
distinguish between patient groups (Supplementary Figure 2).

Discussion
In our study, asthma patients (n=21) were characterized by presence or absence of sputum
eosinophilia (>3% eosinophils: n=13, <3%: n=8) at the time of inclusion. At inclusion, the two
patient groups showed clear differences in clinical parameters. Patients with sputum
8
eosinophilia had more uncontrolled asthma, blood eosinophilia and trends towards higher
FeNO, increased medication use and lower FEV1.
However, there were no differences in the expression of activation and degranulation 9
markers between the two patient groups. No correlations were found between surface
expression markers and markers of disease severity such as ACQ or FEV1. Post-hoc analysis of
the effect size for each marker revealed a median effect sizes r of 0.146 (range 0.032-0.345),
indicating that any effects missed due to the small sample size would have been small.
Therefore, we conclude that activation or degranulation of granulocytes in both blood and
sputum are not associated with clinical differences in eosinophilic and non-eosinophilic
asthma, defined by sputum analysis. The absence of differences in marker expressions on
sputum neutrophils is in line with the finding of an activated phenotype of neutrophils in the
BAL of both healthy volunteers and COPD patients (40) and supports the hypothesis that the
process of recruitment to the per se leads to granulocyte activation. Our results do confirm
that sputum granulocytes have increased expression of the classical activation marker CD11b
and decreased CD62L expression.(8, 10-13) Both neutrophils and eosinophils display a highly
degranulated phenotype, with upregulation of markers for tertiary, specific and azurophilic
granules on neutrophils and crystalloid and secretory granules on eosinophils. Interestingly,
the expression of secretory granule marker CR1 (CD35) was already high on blood cells and
did not increase further on sputum cells. This can be explained by the fact that secretory
vesicles are the first to fuse with the membrane and that in the peripheral blood of asthma

181
Chapter 9

182
 Granulocyte activation state in asthma phenotypes

Figure 5. Multi-dimensional analysis of flow cytometric data. Object scores of each patient (left) and loading
plots of each marker(right) of principal component analysis (PCA) of receptor expression levels on blood and
sputum neutrophils (A) and eosinophils (B). Dashed lines originate from the mean object scores of a group.
Component 1 showed a clear separation between blood and sputum cell types with some residual variation
explained by component 2. Loading plots indicate the contribution of each marker in a component, with a large
distance from 0 indicating a large influence. Markers which are not (differently) expressed remain close to zero
on the x-axis, so play little role in the component 1. Markers upregulated in sputum have positive values for
component 1 (e.g. CD11b), whereas downregulated markers (CD62L) have negative values. PCA analysis did not
discriminate between patient groups (C) in any of the four components. Group sizes were 21 (A),13 (B) and 21
(C).

patients this process has already taken place.(23, 24) Alternatively, CR1 may have been shed
from the cell surface, as described in other diseases.(41)
CD54 has been described as an eosinophil activation marker,(42, 43) which is
expressed on sputum eosinophils.(12) However, in line with another study,(8) we did not
detect CD54 expression on sputum granulocytes, whilst using the same antibody clone (MEM-
111). The reason for this discrepancy remains to be elucidated. We show CD69 to be an
activation marker expressed on sputum eosinophils, just as found on BAL eosinophils.(15)
Another important finding is the high upregulation of immune-regulatory protein
CD274 on sputum neutrophils, and to a lesser extent on eosinophils (see Figure 3). Immune
suppressive blood neutrophils have been shown to upregulate mRNA for this marker during
acute inflammation and employ it for suppression of T-cell proliferation.(26) In the lungs,
8
these cells cannot be as easily identified, as cells have shed CD62L after leaving the
bloodstream, but the expression of CD274 does support the view that sputum neutrophils
might have a suppressive phenotype. Interestingly, high expression has also been found in 9
sarcoid lung granulomas, even though the expression of CD274 on neutrophils was not
studied specifically. In addition, blockade of PD-1 (the receptor for CD274) pathway restores
T-cell functioning in vitro.(27) In conclusion, the expression of CD274 in the lungs of asthmatic
patients favours the hypothesis that sputum cells can modulate inflammation in asthma
rather than merely causing tissue damage and perpetuation of the inflammatory response.
In conclusion, granulocytes in sputum display a highly activated and degranulated
phenotype compared to granulocytes in peripheral blood. However, sputum granulocytes
receptor profiles do not differ in presence or absence of sputum eosinophilia in patients with
asthma. Furthermore, we found the immune-inhibitory protein CD274 to be specifically
expressed on sputum cells, supporting the hypothesis that sputum granulocytes can have an
immune-modulatory instead of a detrimental role in asthma.

183
Chapter 9

References
1. (GINA). Global Initiative for Asthma. Gina report, global strategy for asthma management
and prevention. Updated February 2014 Accessed February 2, 2014.
2. Simpson JL, Scott R, Boyle MJ, and Gibson PG. Inflammatory subtypes in asthma:
assessment and identification using induced sputum. Respirology. 2006;11(1):54-61.
3. Woodruff PG, Khashayar R, Lazarus SC, Janson S, Avila P, Boushey HA, Segal M, and Fahy
JV. Relationship between airway inflammation, hyperresponsiveness, and obstruction in
asthma. The Journal of allergy and clinical immunology. 2001;108(5):753-8.
4. Wagener AH, de Nijs SB, Lutter R, Sousa AR, Weersink EJ, Bel EH, and Sterk PJ. External
validation of blood eosinophils, FENO and serum periostin as surrogates for sputum
eosinophils in asthma. Thorax. 2014.
5. Pavord ID, Pizzichini MM, Pizzichini E, and Hargreave FE. The use of induced sputum to
investigate airway inflammation. Thorax. 1997;52(6):498-501.
6. Baines KJ, Simpson JL, Wood LG, Scott RJ, Fibbens NL, Powell H, Cowan DC, Taylor DR,
Cowan JO, and Gibson PG. Sputum gene expression signature of 6 biomarkers
discriminates asthma inflammatory phenotypes. The Journal of allergy and clinical
immunology. 2014;133(4):997-1007.
7. Brooks CR, van Dalen CJ, Hermans IF, and Douwes J. Identifying leukocyte populations in
fresh and cryopreserved sputum using flow cytometry. Cytometry Part B, Clinical
cytometry. 2013;84(2):104-13.
8. Dominguez Ortega J, Leon F, Martinez Alonso JC, Alonso Llamazares A, Roldan E, Robledo
T, Mesa M, Bootello A, and Martinez-Cocera C. Fluorocytometric analysis of induced
sputum cells in an asthmatic population. Journal of investigational allergology & clinical
immunology. 2004;14(2):108-13.
9. Lay JC, Peden DB, and Alexis NE. Flow cytometry of sputum: assessing inflammation and
immune response elements in the bronchial airways. Inhalation toxicology.
2011;23(7):392-406.
10. Vidal S, Bellido-Casado J, Granel C, Crespo A, Plaza V, and Juarez C. Flow cytometry
analysis of leukocytes in induced sputum from asthmatic patients. Immunobiology.
2012;217(7):692-7.
11. Alexis N, Soukup J, Ghio A, and Becker S. Sputum phagocytes from healthy individuals are
functional and activated: a flow cytometric comparison with cells in bronchoalveolar
lavage and peripheral blood. Clin Immunol. 2000;97(1):21-32.
12. Hansel TT, Braunstein JB, Walker C, Blaser K, Bruijnzeel PL, Virchow JC, Jr., and Virchow
C, Sr. Sputum eosinophils from asthmatics express ICAM-1 and HLA-DR. Clinical and
experimental immunology. 1991;86(2):271-7.
13. in 't Veen JC, Grootendorst DC, Bel EH, Smits HH, Van Der Keur M, Sterk PJ, and Hiemstra
PS. CD11b and L-selectin expression on eosinophils and neutrophils in blood and induced
sputum of patients with asthma compared with normal subjects. Clinical and
experimental allergy : journal of the British Society for Allergy and Clinical Immunology.
1998;28(5):606-15.

184
 Granulocyte activation state in asthma phenotypes

14. Baines KJ, Simpson JL, Wood LG, Scott RJ, and Gibson PG. Systemic upregulation of
neutrophil alpha-defensins and serine proteases in neutrophilic asthma. Thorax.
2011;66(11):942-7.
15. Johansson MW. Activation states of blood eosinophils in asthma. Clinical and
experimental allergy : journal of the British Society for Allergy and Clinical Immunology.
2014.
16. Gundel RH, Wegner CD, Torcellini CA, Clarke CC, Haynes N, Rothlein R, Smith CW, and
Letts LG. Endothelial leukocyte adhesion molecule-1 mediates antigen-induced acute
airway inflammation and late-phase airway obstruction in monkeys. The Journal of
clinical investigation. 1991;88(4):1407-11.
17. Risso A, Cosulich ME, Rubartelli A, Mazza MR, and Bargellesi A. MLR3 molecule is an
activation antigen shared by human B, T lymphocytes and T cell precursors. European
journal of immunology. 1989;19(2):323-8.
18. Gullberg U, Bengtsson N, Bulow E, Garwicz D, Lindmark A, and Olsson I. Processing and
targeting of granule proteins in human neutrophils. Journal of immunological methods.
1999;232(1-2):201-10.
19. Borregaard N, and Cowland JB. Granules of the human neutrophilic polymorphonuclear
leukocyte. Blood. 1997;89(10):3503-21.
20. Adkinson NF, Bochner BS, Burks W, Busse WW, and Holgate ST. Middleton's Allergy:
Principles and Practice. Elsevier - Health Sciences Division; 2013.
21. Mallia P, Message SD, Contoli M, Gray KK, Telcian A, Laza-Stanca V, Papi A, Stanciu LA,
8
Elkin S, Kon OM, et al. Neutrophil adhesion molecules in experimental rhinovirus
infection in COPD. Respiratory research. 2013;14(72. 9
22. Tirouvanziam R, Gernez Y, Conrad CK, Moss RB, Schrijver I, Dunn CE, Davies ZA,
Herzenberg LA, and Herzenberg LA. Profound functional and signaling changes in viable
inflammatory neutrophils homing to cystic fibrosis airways. Proc Natl Acad Sci U S A.
2008;105(11):4335-9.
23. Berends C, Hoekstra MO, Dijkhuizen B, de Monchy JG, Gerritsen J, and Kauffman HF.
Expression of CD35 (CR1) and CD11b (CR3) on circulating neutrophils and eosinophils
from allergic asthmatic children. Clinical and experimental allergy : journal of the British
Society for Allergy and Clinical Immunology. 1993;23(11):926-33.
24. Lundahl J, Skedinger M, Hed J, Johansson SG, and Zetterstrom O. Lability in complement
receptor mobilization of granulocytes in patients with bronchial hyperreactivity. Clinical
and experimental allergy : journal of the British Society for Allergy and Clinical
Immunology. 1992;22(9):834-8.
25. Akbari O, Stock P, Singh AK, Lombardi V, Lee WL, Freeman GJ, Sharpe AH, Umetsu DT,
and Dekruyff RH. PD-L1 and PD-L2 modulate airway inflammation and iNKT-cell-
dependent airway hyperreactivity in opposing directions. Mucosal immunology.
2010;3(1):81-91.

185
Chapter 9

26. de Kleijn S, Langereis JD, Leentjens J, Kox M, Netea MG, Koenderman L, Ferwerda G,
Pickkers P, and Hermans PW. IFN-gamma-stimulated neutrophils suppress lymphocyte
proliferation through expression of PD-L1. PloS one. 2013;8(8):e72249.
27. Braun NA, Celada LJ, Herazo-Maya JD, Abraham S, Shaginurova G, Sevin CM, Grutters J,
Culver DA, Dworski R, Sheller J, et al. Blockade of the Programmed Death-1 Pathway
Restores Sarcoidosis CD4+ T Cells Proliferative Capacity. Am J Respir Crit Care Med. 2014.
28. Linting M, and van der Kooij A. Nonlinear principal components analysis with CATPCA: a
tutorial. Journal of personality assessment. 2012;94(1):12-25.
29. Wu W, Bleecker E, Moore W, Busse WW, Castro M, Chung KF, Calhoun WJ, Erzurum S,
Gaston B, Israel E, et al. Unsupervised phenotyping of Severe Asthma Research Program
participants using expanded lung data. The Journal of allergy and clinical immunology.
2014;133(5):1280-8.
30. Haldar P, Pavord ID, Shaw DE, Berry MA, Thomas M, Brightling CE, Wardlaw AJ, and Green
RH. Cluster analysis and clinical asthma phenotypes. Am J Respir Crit Care Med.
2008;178(3):218-24.
31. Faul F, Erdfelder E, Lang AG, and Buchner A. G*Power 3: a flexible statistical power
analysis program for the social, behavioral, and biomedical sciences. Behavior research
methods. 2007;39(2):175-91.
32. Sennels HP, Jorgensen HL, Goetze JP, and Fahrenkrug J. Rhythmic 24-hour variations of
frequently used clinical biochemical parameters in healthy young males--the Bispebjerg
study of diurnal variations. Scandinavian journal of clinical and laboratory investigation.
2012;72(4):287-95.
33. Mattes J, Storm van's Gravesande K, Moeller C, Moseler M, Brandis M, and Kuehr J.
Circadian variation of exhaled nitric oxide and urinary eosinophil protein X in asthmatic
and healthy children. Pediatric research. 2002;51(2):190-4.
34. Rutgers SR, Timens W, Kaufmann HF, van der Mark TW, Koeter GH, and Postma DS.
Comparison of induced sputum with bronchial wash, bronchoalveolar lavage and
bronchial biopsies in COPD. The European respiratory journal : official journal of the
European Society for Clinical Respiratory Physiology. 2000;15(1):109-15.
35. Efthimiadis A, Spanevello A, Hamid Q, Kelly MM, Linden M, Louis R, Pizzichini MM,
Pizzichini E, Ronchi C, Van Overvel F, et al. Methods of sputum processing for cell counts,
immunocytochemistry and in situ hybridisation. The European respiratory journal
Supplement. 2002;37(19s-23s.
36. Simpson JL, McElduff P, and Gibson PG. Assessment and reproducibility of non-
eosinophilic asthma using induced sputum. Respiration; international review of thoracic
diseases. 2010;79(2):147-51.
37. Mahmudi-Azer S, Downey GP, and Moqbel R. Translocation of the tetraspanin CD63 in
association with human eosinophil mediator release. Blood. 2002;99(11):4039-47.
38. van Velzen JF, Laros-van Gorkom BA, Pop GA, and van Heerde WL. Multicolor flow
cytometry for evaluation of platelet surface antigens and activation markers. Thromb Res.
2012;130(1):92-8.

186
 Granulocyte activation state in asthma phenotypes

39. Vermes I, Haanen C, Steffens-Nakken H, and Reutelingsperger C. A novel assay for

apoptosis. Flow cytometric detection of phosphatidylserine expression on early apoptotic
cells using fluorescein labelled Annexin V. Journal of immunological methods.
1995;184(1):39-51.
40. Fortunati E, Kazemier KM, Grutters JC, Koenderman L, and Van den Bosch v J. Human
neutrophils switch to an activated phenotype after homing to the lung irrespective of
inflammatory disease. Clinical and experimental immunology. 2009;155(3):559-66.
41. Danielsson C, Pascual M, French L, Steiger G, and Schifferli JA. Soluble complement
receptor type 1 (CD35) is released from leukocytes by surface cleavage. European journal
of immunology. 1994;24(11):2725-31.
42. Walker C, Rihs S, Braun RK, Betz S, and Bruijnzeel PL. Increased expression of CD11b and
functional changes in eosinophils after migration across endothelial cell monolayers. J
Immunol. 1993;150(9):4061-71.
43. Yamamoto H, Sedgwick JB, Vrtis RF, and Busse WW. The effect of transendothelial
migration on eosinophil function. American journal of respiratory cell and molecular
biology. 2000;23(3):379-88.

187
Chapter 9

Supplementary Figures

Supplementary figure S1. Representative histograms of all measured markers in blood (red) and sputum (blue).
Filled, solid lines represent the relevant marker/antibody, dashed lines indicate isotype controls Histograms are
binned and normalised for peak values.

188
 Granulocyte activation state in asthma phenotypes

Supplementary figure S2. Object scores of PCA on marker expression for each cell type and location. PCA was
8
performed on A) blood neutrophils, B) sputum neutrophils and C) sputum eosinophils separately. None of these
analyses was able to discriminate between patient groups. Dashed lines originate from the median object score
of the two patient groups. N=13 for >3% sputum eosinophils and N=8 for <3% sputum eosinophils.
9

189
 Chapter 10

Granulocyte kinetics in the blood and sputum

of patients with eosinophilic and neutrophilic
inflammation.

T. Tak 1,3, L. Conemans1,3, B. Hilvering1,3, C. van Aalst1,3, D. Hartl4, J.A.M. Borghans2,3,

K. Tesselaar2,3 and L. Koenderman1,3.

1. Departments of Respiratory Medicine, University Medical Center Utrecht,

Utrecht, the Netherlands
2. Departments of Immunology, University Medical Center Utrecht, Utrecht,
the Netherlands
3. Laboratory of Translational Immunology, University Medical Center
Utrecht, Utrecht, the Netherlands
4. Department of Pediatrics, University of Tübingen, Tübingen, Germany

Manuscript in preparation
Chapter 10

Abstract
Little is known about the kinetics of inflammatory cells in blood and sputum of patients with
inflammatory lung disease. The kinetics of granulocytes in circulation and sputum were
studied by labelling 11 healthy volunteers, 6 cystic fibrosis (CF) patients and 9 eosinophilic
asthma (EA) patients with 6,6-2H2-glucose for 6 hours. Neutrophils, eosinophils and free DNA
(only in CF sputum) were isolated from both blood and sputum at different time points, and
DNA 2H-enrichment was determined by GC-MS.
The neutrophil compartment in CF patients exhibited a ~1 day shorter maturation time
in bone marrow compared to healthy controls, while the neutrophil compartment of EA
patients was unaffected. In contrast, the eosinophil compartment in EA patients was
characterized by a ~1 day longer maturation time in BM compared to healthy controls, while
the dynamics of eosinophils of CF patients were marginally affected. In addition, eosinophil
turnover in the circulation was significantly faster in EA patients than in healthy controls.
While the kinetics of neutrophils in blood and sputum were highly comparable, eosinophils
seemed to follow different labelling patterns in blood and sputum. These findings suggest
that neutrophils survive only briefly in sputum, while eosinophils in sputum can survive much
longer. The 2H-kinetics of free DNA in the sputum of CF patients was similar to that of
neutrophils in blood and sputum, yet with a delay of around 1 day.
These findings provide critical insight into the behaviour of both neutrophils and
eosinophils in blood and lungs during inflammatory conditions. Our data support longevity of
eosinophils in the sputum of EA and CF patients, but contrast the widely held view that
neutrophils survive for relatively long periods of time in the sputum of lung patients with
neutrophil inflammation.

192
 Granulocyte kinetics in chronic pulmonary diseases

Introduction
Accumulation of eosinophils in the bronchial airways and sputum is characteristic for
eosinophilic asthma (EA) (1, 2), whereas in cystic fibrosis (CF) neutrophils are the predominant
immune cell(3). In both diseases, both types of granulocytes are highly activated in the
sputum and BAL when they are detected in these samples (4). Granulocytes obtained from
the lung lumen express markers indicative for degranulation (5, 6). In addition, depositions of
cytotoxic granule proteins (Eosinophil cationic protein and neutrophil elastase) are found in
the sputum of both patient groups (7-9). Characteristic for CF is neutrophil-derived DNA in
the sputum and BAL, which has been reported to contribute to disease severity (10).
Eosinophils are a therapeutic target in EA patients as improper mobilization and
activation of these cells are considered to be involved in the pathogenesis of asthma(1). At
this moment only two therapeutic strategies interfere significantly with these processes in
vivo: treatment with glucocorticosteroids (GCS) (11) and injection of monoclonal antibodies
directed against interleukin (IL)-5 (12-15). Treatment with GCS is associated with lowering of
eosinophil numbers in the sputum (16). This is thought to be mediated by eosinophil
apoptosis, as treatment of eosinophils in vitro leads to apoptosis of these cells (17). However,
clear data supporting this hypothesis in vivo are still lacking. Treatment with Mepolizumab
reduces eosinophil production in the bone marrow by blocking the
proliferation/differentiation factor IL-5 (12). Surprisingly, this treatment was shown to reduce
the number of circulating eosinophils by more than 90% in 4 weeks, whereas only a 50%
reduction was observed in eosinophil numbers in bronchial biopsies and sputum samples,
even after 10 weeks of therapy (18, 19). This raises the question whether eosinophils in the
lungs are long-lived, or whether the few eosinophils that are produced in the presence of
Mepolizumab preferentially home towards the lungs of asthma patients.
Neutrophils are a therapeutic target in lung diseases associated with neutrophilic 10
inflammation such as COPD and CF. However, up to date, no medication is available that can
reduce neutrophil production in the bone marrow other than cytostatic drugs such as
azathioprine and methotrexate. These have been tested in severe asthma (20) and CF (21),
but with limited success.
GCS are commonly used for the treatment of asthma and CF (11) although their role
in the treatment of CF is controversial (17). Some studies argue that GCS are increasing
neutrophil survival and might actually prevent the resolution of inflammation. In asthma, GCS
have been proven particularly successful in reducing symptoms in patients with eosinophilic
asthma, but less in patients with non-eosinophilic asthma (22). The lack of effect of GCS in CF
and other neutrophil-associated diseases might be attributed to the in vitro finding that GCS
induce eosinophil apoptosis, but prevent neutrophil apoptosis(23, 24). Oral administration of
GCS (OCS) was shown to increase neutrophil numbers in the bronchial mucosa of asthma
patients, which might be in support of the hypothesis that GCS cause increased neutrophil
survival. In the same study, however, inhaled GCS (ICS) did not have an effect on neutrophil
numbers in sputum or mucosa, which might indicate that the increased neutrophil number
with OCS are not caused by increased survival in the lungs (25).

193
Chapter 10

Similar to the situation with eosinophils, little is known about neutrophil kinetics in
the lung in health and disease. Estimates for the circulatory lifespan of neutrophils range from
7h to 5 days (26, 27) and chapter 3). After leaving the bloodstream, neutrophils have been
postulated to survive for several more days in tissues (28), although these results were
obtained using toxic labelling strategies(29) in cats. In human lungs it has been suggested that
neutrophils survive for at least two days, as reinjection of 111In-radiolabelled neutrophils in
COPD patients resulted in detection of radioactivity 1 to 5 days after all radiolabelled
neutrophils had left the circulation (30). Unfortunately, however, in the latter study sputum
samples were not analysed for the presence of intact neutrophils and ex vivo manipulation
may have affected neutrophil homing (29, 31).
In vivo deuterium (2H) labelling allows labelling of untouched cells with a non-toxic
and non-radioactive label (32). Here we report the deuterium labelling results of 11 healthy
volunteers, 6 CF patients and 9 EA patients who have been labelled with 2H in the form of 6,6-
2H -glucose (1 g/kg bodyweight) for 6 hours. Neutrophils and eosinophils were isolated from
2
blood and sputum at several time points after intake of label. Based on the deuterium
enrichment in the DNA of these cells, we aimed to determine the kinetics of these cells in
both blood and lungs. In addition, we have estimated the turnover of free DNA in the sputum
of CF patients.

Materials and Methods

Approval
Patients were recruited from the outpatient clinic in the UMC Utrecht, healthy controls were
recruited by advertisements in the UMC Utrecht and Utrecht University in Utrecht, the
Netherlands. All participants gave written informed consent in accordance to the Declaration
of Helsinki (seventh revision, Fortaleza, 2013). This study was approved by the local ethical
review board (METC) in accordance to Dutch legislation.

Study population
Healthy volunteers, eosinophilic asthma patients and CF patients were all included with an
age between 18-75 years. For patient characteristics, see table 1. Adult asthma was defined
according to the GINA guidelines, based on clinical features (episodic shortness of breath,
particularly at night and often accompanied by cough/wheezing) and reversibility of FEV1
(forced expiratory volume in 1s) upon inhalation of 400µg salbutamol (≥12% predicted or
≥200ml) and/or airway hyperresponsiveness to histamine (PC20 <8mg/ml) (11).
Subsequently, eosinophilic asthma was defined by sputum eosinophilia (≥3%) on the most
recent sputum induction.
CF patients were identified by proven CF by sweat test (sweat Cl concentration >
60mmol/l) or 2 disease-causing mutations in the CFTR gene. Most exclusion criteria were
similar for all groups: smoking (previous 12 months or >10 pack years), chronic infections (e.g.
STD’s, HIV), and drug or alcohol abuse. For healthy volunteers the use of any medication
(except contraceptives and pain killers) was an additional exclusion criterion. Asthma patients

194
 Granulocyte kinetics in chronic pulmonary diseases

with proven allergic bronchopulmonary aspergillosis were also excluded. Since CF-related
diabetes might lead to aberrant incorporation of label (33), insulin-dependent CF patients
were excluded from the study.

Labelling and sampling procedure

In vivo 2H-labelling was performed as published previously (32). In short, volunteers were
asked to refrain from food intake for at least 4h prior to the start of the procedure. Volunteers
drank 12 half-hourly doses of metabolic grade 6,6-2H2-glucose amounting to a total of 1g/kg
bodyweight (Cambridge Isotope Laboratories, Tewksbury, MA, USA). 6,6-2H2-glucose was
metabolized and incorporated into the DNA of dividing cells via the de novo nucleotide
synthesis pathway17,18. To determine the availability of label (plasma glucose enrichment)
during the labelling procedure, small blood samples were collected on filter paper by finger
pricks (34). Blood glucose concentration was determined from frozen plasma samples using
a Precision XceedPro analyser (Abbott Diabetes Care, Alameda, CA, USA). Cells were isolated
from blood collected by venapuncture before start of the labelling procedure and at five later
time points. Induced sputum was obtained at two time points after collection of blood. Time
points varied between volunteers to maximize the information obtained about cellular
kinetics.

10

Figure 1. Flowchart of the study design

Blood processing
Blood was collected in sodium heparin tubes. Erythrocytes were lysed using lysis buffer
consisting of 150 mM NH4Cl, 10 mM KHCO3 and 0.1 mM NA2EDTA dissolved in ddH2O.
Resulting total leukocyte preparations were washed and resuspended in FACS staining buffer
consisting of PBS supplemented with trisodium citrate (0.32% w/v) (both from UMCU

195
Chapter 10

pharmacy) and human pasteurized plasma solution (10% w/v, Sanquin, Amsterdam, the
Netherlands).

Sputum induction and processing

Sputum was induced by inhalation of 0.9-5% saline aerosols unless patients had a predicted
FEV1 below 30% and patients were asked for spontaneous sputum. Single cell suspensions
were obtained by 15 min 37°C incubation with 4:1 (w/vol) Sputolysin (Merck Millipore,
Darmstadt, Germany) adjusted to 280-290mOsm with NaCl, whilst slowly dissociating using a
wide-tip Pasteur pipette (35, 36). For CF sputa, bovine pancreas DNase I (Roche, Basel,
Switzerland) was added during the last 10 min at a concentration of 1mg/ml. Resulting cell
suspensions were filtered through a 48μm nylon filter and washed with PBS. Absolute cell
counts in sputum were determined using a CELL-DYN Emerald haemocytometer (Abbott
Diagnostics, Abbott Park, IL, USA). Differential counts were performed on cytospin slides of
sputum cells stained with May-Grünwald-Giemsa. Percentages of cells in sputum for
determination of asthma subtype were calculated after exclusion of squamous epithelial cells
as published previously(37). For calculation of absolute neutrophil and eosinophil counts from
total sputum counts, squamous epithelial cells were not excluded from analysis. Remaining
cells were washed and resuspended in PBS for FACS staining procedure. Free DNA was
collected from CF sputum supernatants before addition of DNAse I by spinning down small
samples to remove cells. DNA was isolated using a NucleoSpin Blood kit according to
producers' instructions (MACHEREY-NÄGEL, Düren, Germany).
DNA concentrations from samples were determined using a NanoDrop 2000 (Thermo
scientific, Waltham, MA, USA). The total amount of free DNA was calculated from the DNA
concentration and the weight of the sputum sample.

Granulocyte isolation
Single cell suspensions from sputum were stained with LIVE/DEAD violet cell staining kit (Life
technologies (Thermo Fisher Scientific, Waltham, MA, USA) in PBS. After washing, cells were
stained in FACS staining buffer with the following antibodies: CD14-ECD (clone RMO52,
Beckman Coulter Pasadena, CA, USA), CD62L-PE (clone SK11), CD16-FITC (clone 3G8) (both
from BD Biosciences, San Jose, CA, USA), CD193-Alexa647, CD206-PerCP/Cy5.5 (both from
Biolegend, San Diego, CA, USA) and CD9 (unlabelled clone S32, a kind gift from ing. D.
Kanters). Anti-CD9 was counterstained after three washes with Goat-anti-mouse-IgM-PC7
(SouthernBiotech, Birmingham, AL, USA). Blood samples were stained similarly, but without
LIVE/DEAD violet and CD206.
Cells were isolated using an AriaIII FACS sorter (BD, Biosciences). Blood granulocytes
were identified based on a FSC/SSC granulocyte gate, doublet exclusion and CD14neg (see
supplementary figure 1 for gating strategy). Subsequently, neutrophils were identified as
CD16high and eosinophils as CD9high. Sputum cells were gated similarly, with addition of a
Live/DEAD stain to exclude necrotic cells and highly auto-fluorescent cells. Sorted populations
were re-analysed (typically >99% pure). Cellular DNA was isolated from fresh samples using a

196
 Granulocyte kinetics in chronic pulmonary diseases

NucleoSpin Blood kit according to producers' instructions (MACHEREY-NAGEL, Düren,

Germany).

Determination of DNA 2H enrichment

DNA was hydrolysed and derivatized to pentafluoro triacetate (PFTA) as described in detail
previously (34). Relative quantities of unlabelled and 2H-labeled PFTA were determined with
an Agilent 7980A/5975C GC-MS in negative chemical ionization mode scanning for m/z 435
(M+0, unlabeled) and m/z 437 (M+2, labelled). Resulting enrichments were corrected for
natural background enrichment, availability of 6,6-2H2-glucose in plasma and the ratio
adenosine that is de novo synthesised/obtained by base salvage (0.65, see chapter 3) as
published previously (34).

Estimation of PMPtt and kinetics by mathematical modelling

A mixed effect model (as described in chapter 3) was fitted to the deuterium enrichment data
of all individuals. This model has the advantage of being able to estimate parameters for each
individual, and aids the quantification of the PMPtt of neutrophils and eosinophils, but has
difficulties distinguishing whether the rate limiting step of the label kinetics takes place in the
blood or the bone marrow (BM). The alternative estimate with the rate limiting step taking
place in the BM produced similar results for PMPtt and turnover of the rate-limiting step.

Data analysis
Flow cytometry data was analysed using FCS express 5.01 (De Novo Software , Los Angeles,
CA, USA), statistical analysis was performed using GraphPad Prism 6.07 (GraphPad Software,
La Jolla, CA, USA), parameter estimations were performed using Monolix 4.3.3. (Lixoft, Orsay,
France).
10
Results
Patient characteristics
Eleven healthy controls, 6 CF patients and 9 EA patients were included in the study (see
flowchart figure 1 and table 1). From all study participants 6 blood samples were collected
and from all patients and 5 healthy controls two additional sputum samples were obtained.
Compared to controls, CF patients had a higher percentage of neutrophils in sputum. All 9
asthma patients had >3% eosinophils in their sputum during their most recent sputum
induction; one patient also had >56% sputum neutrophils and was therefore considered as
having a mixed asthma phenotype. Two patients with an eosinophilic asthma phenotype
switched to a mixed phenotype, between the last clinical sputum induction and the first
sputum induction in the study. One patient had changed form an eosinophilic to a
neutrophilic asthma phenotype. Thus, the long-term reproducibility of the asthma phenotype
in this study was 66% and the reproducibility regarding the presence of eosinophils was 89%.
No differences were observed between the first and the second sputum induction in any of

197
Chapter 10

the asthma patients, indicating that the short-term reproducibility of sputum inductions in
our small cohort was high.

Table 1. Patient Characteristics.

Patient characteristics Controls CF EA
N 11¶ 6 9
Male (%) 4 (36%) 2 (33%) 5 (56%)
Age (years) 24 (19-32) 26 (22-36) 56 (40-72)
BMI 22.7 (19.2-24.7) 24.2 (18.4-35.3) 27.2 (23.2-32.2)
FEV1 (L) N.D. 1.77 (1.23-2.72) 2.45 (1.47-3.1)
FEV1 (% predicted) † N.D. 51.6 (29-93) 73.7 (52.2-96.6)
Total WBC counts 6.6 (5.7-7.8) 8.9 (6.7-11.1) 7.8 (4.7-14.2)
Blood neutrophil count (106/ml) 4.1 (2.6-8.6) 4.68 (3.47-5.56) 4.6 (2.6-9.4)
Blood eosinophil count (106/ml) 0.21 (0.05-1.09) 0.23 (0.12-0.40) 0.54 (0.1-2.2)*
Sputum neutrophils (%)§ 1.4 (0-12.4) 76.7 (51.8-96.8)** 38.3 (0-83.8)*
Sputum eosinophils (%)§ 0 (0-0) 3.0 (0-12.9) 24.7 (0-72.6)*
FeNO (ppb) 18 (11-35) 14 (5-48) 60 (11-198)
ACQ N/A 2.31 (1.28-3.14) 1.53 (0.14-2.57)
Medication score‡ N/A 4.11 (3-5)
Data represent mean (range) unless indicated otherwise.
¶ N=5 for sputum measurements
† Percentage reversibility in FEV1 (corrected for age, gender, length and bodyweight) a�er inhala�on of 2x
100μg salbutamol.
§ As determined by microscopic evaluation, excluding epithelial cells.
‡ 5-point ordinal scale based on guidelines of the British Thoracic Society , with 0) no medication, 1) inhaled
SABA when required 2) low dose ICS + SABA 3) low/medium dose ICS + LABA, or medium dose ICS + LABA 4) High
dose ICS +/- LABA and 5) High dose ICS + OCS +/- LABA.
*,**: significantly different from control with p<0.05 and p<0.0001 as determined by Kruskall-Wallis test with
Dunn’s correction for multiplicity.
Abbreviations: BMI: body mass index, FeNO: forced exhaled nitric oxide, ACQ: Asthma Control Questionnaire,
FEV1: Forced Expiratory Volume in 1 sec, N/A: Not applicable.

Cell numbers in sputum

Differential counts on leukocytes of all sputum samples (figure 2A) revealed that healthy
control sputum samples contained mostly macrophages, CF sputum samples contained
mostly neutrophils and EA sputum samples contained approximately 40% of both neutrophils
and eosinophils.
Absolute cell counts (figure 2B) were highest in CF patients, followed by EA and were
lowest in healthy controls. For both patient groups all sputum samples contained enough
neutrophils (figure 2C) for deuterium enrichment analysis, except for two EA patient samples.
Eosinophils (figure 2D) were present in adequate numbers for detection of deuterium
enrichment in 12 out of 18 EA sputum samples and in 7 out of 12 CF sputum samples. In
addition to having the lowest absolute cell counts, sputum samples from healthy volunteers
consisted of >90% epithelial cells (26 and 56% for CF and EA, respectively). Therefore,

198
 Granulocyte kinetics in chronic pulmonary diseases

neutrophil numbers in sputum samples from controls were too low for reliable determination
of DNA 2H-enrichment. No eosinophils were detected in healthy control sputum.

B CNeutrophils in sputum D
10 9 **
****
10 8

10 7

Cells
10 6

10 5

10 4

N.D.
Controls CF EA

E F
10

Figure 2. Cell composition and numbers in sputum samples. Differential counts of sputum leukocytes (A) were
determined by microscopic evaluation of May-Grünwald-Giemsa stained cytospin slides of induced sputum
samples and show large differences in sputum cell composition. Absolute numbers of total sputum cells (B)
(including epithelial cells), neutrophils (C) and eosinophils (D). For both patient groups, free DNA was extracted
from sputum supernatants (E). Asthma patients were recruited as having EA based on their most recent clinical
sputum differential count. At the time of the study, however, the phenotype of some patients had changed into
a neutrophilic or mixed asthma phenotype (F). Each sputum sample was depicted separately for healthy controls
(N=10 samples), CF patients (N=10-12) and EA patients (12-18). Data represent medians (A) or geometric means
with 95%CI (B-E). * indicates p<0.05, ** p<0.01, *** p<0.001 and ****p<0.0001 as determined by Mann-
Whitney U test or Kruskall-Wallis test with Dunns multiple comparisons test.

199
Chapter 10

Plasma availability of 6,6-2H2-glucose label

To correct for differences between individuals in 6,6-2H2-glucose uptake, DNA enrichments
were corrected as if all newly synthesised DNA was 2H labelled. Theretofore, we obtained
small blood samples by finger prick at 0.5h, 3h and 5.5h after intake of label and determined
the relative plasma enrichment by GC-MS (figure 3A). At 30 min post-intake of label, plasma
glucose enrichment was significantly lower in the CF group compared to healthy controls
(p=0.0056). This difference does not appear to be caused by an altered glucose metabolism,
since insulin-dependent diabetic patients were not included in our study and fasting plasma
glucose concentrations were normal (figure 3B). The calculation used to determine the
correction factor (see (34)) assumes that a slow increase in plasma glucose enrichment results
in slow down labelling. Therefore, correction factors did not differ between our patient
groups.

Figure 3. Plasma glucose

enrichment. Plasma glucose 2H-
enrichment (A) was determined
before intake of label and 30 min, 3h
and 5.5h after intake of label. CF
patients show lower enrichment 30
min after the first intake of label,
whereas this difference was gone 3h
after the first intake of label. Plasma
glucose concentrations at t=0 (B) did
not show significant differences
between controls and CF and EA
patients, although EA patients
tended to have slightly higher
correction factor plasma glucose concentrations. The
5 effect on the calculated availability
of label (C) was negligible, however,
4
Correction factor

as the average correction factor

3 used to correct DNA enrichments
2
did not differ between groups. Data
represent means +/- 95% CI, with
1 Control N=11, CF N=6 and Asthma
0 N=9 at each time point. Plasma
Ctrl CF EA glucose enrichment data (A) were
nudged (-0.1h for controls, +0.1h for CF) to increase visibility of error bars. ** indicates p=0.006 for CF compared
to control, as determined by two-way ANOVA with Holm-Sidak correction for multiplicity.

Granulocyte kinetics in blood

6,6-2H-glucose is incorporated into the DNA during the S-phase of the cell cycle. Since neither
neutrophils nor eosinophils are capable of proliferation, DNA 2H-enrichment is only measured
after cells have differentiated from progenitors in the bone marrow (neutrophilic and
eosinophilic myelocytes) into mature granulocytes and are released into the bloodstream (32,

200
 Granulocyte kinetics in chronic pulmonary diseases

38). This delay is generally referred to as the post-mitotic pool transit time (PMPtt). Under
homeostatic conditions, the neutrophil PMPtt was described as 6-7 days (32). We obtained
similar results in our healthy controls, with low neutrophil DNA enrichment until day 6 post-
intake of label and a higher enrichment at day 7 (Figure 4A). As shown in chapter 4, for
eosinophils, no DNA enrichment was detected in healthy control samples up to day 4, with a
large increase at day 5.
During neutrophil inflammation such as found in CF patients, a significantly higher
neutrophil DNA enrichment was observed already at days 5 and 6 post intake of label (Figure
4A), suggesting a shortened PMPtt. Estimation of the PMPtt for each volunteer (Figure 4C)
with a mathematical model also showed a significantly shorter PMPtt for CF patients
compared to controls. For asthma patients, neutrophil enrichment was significantly higher
compared to controls at day 7, although estimates for each volunteer do not show a
significant decrease in PMPtt.

10

Figure 4. Blood DNA 2H enrichment kinetics. After intake of label, neutrophils (A) and eosinophils (B) were
isolated and their DNA 2H enrichment was determined for 11 healthy controls, 6 CF patients and 9 asthma
patients. Best estimates of PMPtt (C) and the turnover of the rate-limiting compartment (D) (either blood or
BM). N indicates neutrophils, E indicates eosinophils. Data represent (A,B) median enrichments +/- SEM from a
total of 66, 36 and 54 measurements for controls, CF and EA patients, respectively or estimates for each
individual with lines indicating means and error bars indicating 95%CI. * indicates p<0.05, ** p<0.01 and ***
p<0.001 as determined by (A,B) two-way ANOVA with Dunnett’s multiple comparisons test or (C,D) one-way
ANOVA with Tukey’s multiple comparisons test.

201
Chapter 10

Eosinophil DNA 2H-enrichment kinetics in blood (Figure 4B) was similar in CF patients
compared to controls. In EA patients, however, clear differences were observed. The first
detectable DNA enrichment in blood of EA patients was 1-2 days delayed and first seen at day
6 with a low enrichment. This suggests that in EA patients, the eosinophil PMPtt is longer than
in healthy controls and is confirmed by the estimated PMPtt for individuals (0.5d longer). The
maximal eosinophil 2H DNA enrichment was observed at day 9, whereas in healthy controls
the maximum enrichment was reached at day 7. In addition, estimates on the lifespan in the
slowest pool showed that despite their longer maturation time, eosinophils have more rapid
kinetics (figure 4D), whereas no differences were found in neutrophil lifespans.

Granulocyte kinetics in sputum

For five out of 11 healthy controls and all patients, two induced sputum samples were
obtained (except for one volunteer, from whom spontaneous sputum was used). From these
samples, granulocytes were isolated and their DNA 2H-enrichment was determined. As stated
above, cell numbers were too low in control sputum samples to allow reliable DNA 2H
quantification. Most sputum samples from CF and EA patients contained enough neutrophils
for reliable quantification of 2H enrichment, while for eosinophils, 17 out of 30 samples
contained enough cells. In addition to normal neutrophils, 11 sputum samples of CF patients
also contained large numbers of CD16low neutrophils, which were sufficient for deuterium
enrichment analysis.

Figure 5. DNA 2H-enrichment of paired sputum and blood granulocytes. DNA 2H enrichment was plotted over
time for sputum cells together with paired blood enrichments before sputum induction for neutrophils from CF
(A) and EA (C) patients as well as for eosinophils (B,D). Arrows indicate two individuals where enrichment in
eosinophils and neutrophils preceded the enrichment in blood taken at the same moment.

202
 Granulocyte kinetics in chronic pulmonary diseases

DNA 2H enrichment kinetics of sputum neutrophils were remarkably similar to those of blood
neutrophils in both CF and EA patients, while CD16low neutrophils from CF patients seemed
to reach their peak of labelling with a minor delay of about 1 day compared to classical
neutrophils in blood and sputum. Eosinophils, however, did show differences in sputum
compared to blood. When the highest enrichment was reached in blood at day 7, the
enrichment was lower in all four sputum eosinophil samples from the same time point. In
contrast, after enrichments decreased in blood, they increased or remained high in two out
of three EA sputum eosinophils samples and 3 out of 4 CF sputum samples. Interestingly, in
some cases enrichment in sputum cells appears to precede enrichment in blood, such as at
day 5 for eosinophils in one individual and at day 6 in the neutrophils of another individual
(indicated with arrows).

Effects of corticosteroid use on granulocyte kinetics

As stated above, GCS have been described to induce eosinophil apoptosis, but to decrease
neutrophil apoptosis in vitro (23, 24). To determine the effects of steroid use we grouped our
patients into three different groups: No steroid use (3 CF patients), ICS use (3 CF, 7 EA) or ICS
combined with OCS use (2 EA). No clear differences were observed in the 2H-enrichment
kinetics of blood neutrophils and eosinophils and sputum neutrophils between these
treatment groups (figure 6). For sputum eosinophils, however, there seemed to be a
difference between the three treatment groups. Despite a low number of measurements,
patients receiving OCS showed lower DNA enrichment in sputum eosinophils compared to
the two other treatment groups. Patients that did not receive any GCS showed the highest
enrichment in sputum eosinophil DNA. Since lower enrichments indicate a lower turnover,
GCS use may be associated with eosinophils surviving longer in the lungs.

10

203
Chapter 10

Figure 6. Effect of GCS on granulocyte kinetics. Enrichment data from both CF and asthma patients were
compared between patients taking ICS, ICS and OCS, or no GCS. Data were plotted for blood neutrophils (A)
blood eosinophils (B), sputum neutrophils (C), and sputum eosinophils (D). Data consist of a total of 18 blood
and 6 sputum measurements from 3 CF patients not receiving GCS, 60 blood and 20 sputum samples from 10
patients receiving ICS, and 12 blood and 4 sputum samples from 2 EA patients receiving ICS and OCS. Points
indicate means from duplicate measurements.

2H-enrichment kinetics of free DNA

The sputum of CF patients contains large amounts of free DNA. From 6 CF patients a total of
12 measurements could be obtained (figure 7). Free DNA 2H-enrichment kinetics showed
similar uplabelling kinetics compared to neutrophils in sputum, which were the most likely
source of free DNA in sputum. There appears to be, however, a 1 day delay in free DNA 2H-
enrichment compared to blood and sputum neutrophil enrichments and most sputum
samples show a relatively high enrichment at later measurements.
To remove free DNA from the sputum and improve lung function, some CF patients
inhale nebulized recombinant human deoxyribonuclease I (Dornase alpha) (39). In our study
group, three out of six patients used Dornase alpha on a daily basis. The other three patients
did not use Dornase alpha or any other DNA-degrading medication. We observed a trend
towards a higher sputum DNA content, but when comparing the 2H-enrichment kinetics of
free DNA between patients with and without the use of Dornase alpha, we did not observe
any clear differences. This suggests that Dornase alpha does not influence the turnover of
free DNA in the lungs of CF patients.

204
 Granulocyte kinetics in chronic pulmonary diseases

Enrichment (fraction)
Enrichment (fraction)

Figure 7. 2H-enrichment of free DNA in sputum of CF patients. Sputum samples of CF patients contain large
amounts of free DNA. This free DNA was isolated and its 2H-enrichment was determined and plotted together
with 2H enrichment kinetics of blood and sputum neutrophils (A) and separately for patients with or without the
use of Dornase alpha (B). Also the total amount of free DNA in sputum samples was compared between CF
patients with and without the use of Dornase alpha (C). Data was obtained from (A) 12 sputum samples with
matched blood samples or (B-C) 12 sputum samples from 6 volunteer, 3 of whom with and three without the
use of Dornase alpha.
10
Discussion
Glucose kinetics in plasma
The CF patients in our study showed delayed plasma 6,6-2H2-glucose enrichment, which might
point to CF-associated diabetes (33). However, none of our patients used insulin or displayed
elevated fasting plasma glucose levels (figure 2B). Therefore, a more likely cause might for
the delayed label uptake might be due to the previously described delayed gastric emptying
in CF patients (40) (REF). Delayed gastric emptying would lead to a slower uptake of
deuterated glucose in the intestine and thereby a delayed uptake in the blood. However,
gastric emptying was not assessed in our patient population. The delayed uptake of glucose
did not have an overall effect on 2H labelling, however, since at later time points there was
no difference between the three groups nor was there a difference between the correction
factor for availability of label.

205
Chapter 10

Neutrophil kinetics in blood

The blood of CF patients contained a small percentage of both CD16low/CD62Lhigh and
CD16high/CD62Llow neutrophils (figure 8), which is a clear indication of inflammation.
Pulmonary inflammation has been shown to reduce the PMPtt in mice (41) and, indeed, DNA
2H-enrichment kinetics revealed a shorter neutrophil PMPtt in CF patients compared to

healthy controls. In agreement with these data we have shown in acute inflammation
CD16low/CD62Lhigh banded neutrophils to appear 2 days earlier in the blood compared to
normal neutrophils under normal conditions (see chapter 7). On average, however, only 5%
of the neutrophils in the blood of CF patients were CD62Lhigh/CD16low. Five percent of the
maximal measured enrichment (~0.10) would result in a fraction measured enrichment at day
5 of only 0.005. As the measured enrichment at day 5 is far higher than 0.005, the reduced
PMPtt in CF patients is unlikely to be caused solely by immature cells entering the circulation.
Instead, it indicates that also the CD16high/CD62Lhigh neutrophils are released early from the
BM and, thereby, causing an increase in neutrophil maturation times.
Neutrophil progenitor turnover has been shown to increase in response to GM-CSF
(42), a cytokine that has been shown to be elevated in some CF patients (43). We did not
observe an increased turnover of the rate-limiting neutrophil compartment, however (figure
4D). This indicates that either the BM turnover is not increased in CF patients, or that the BM
is not the rate-limiting step in the kinetics of neutrophil 2H-DNA enrichment.

Eosinophil kinetics in blood

Eosinophils in EA patients have a delayed maturation rate, which is in contrast to the more
rapid maturation of neutrophils in CF patients. Eosinophils with labelled DNA appear in the
blood of EA patients one day later than in healthy controls and CF patients. In addition,
estimations of the turnover of the slowest eosinophil pool turnover were higher in EA patients
compared to controls and CF patients (figure 8D). Interestingly, eosinophil numbers were
increased in the blood of EA patients. This could be caused by a delayed eosinophil clearance
from the circulation, or an increased number of (dividing) eosinophilic progenitors in the BM.
The first option appears not plausible, as there is no longer circulatory lifespan. The second
option of an increased number of eosinophilic progenitors in the BM is more like as it has also
been shown in eosinophilic inflammation (44) and would allow for an increased eosinophil
count even despite a slower turnover of individual progenitors in the BM. To prove that the
increased eosinophil numbers are due to only an increase in the number of dividing
eosinophilic progenitors would require estimation of the kinetics of these cells with 2H in the
bone marrow. The required multiple bone marrow biopsies are unlikely to be performed due
to ethical constraints .
As stated above, corticosteroids have been shown to shorten eosinophil lifespans in
vitro (24) and are widely believed to reduce sputum eosinophilia, even though this latter
notion has been questioned (45). When comparing patients taking no corticosteroids, only
ICS or combined ICS and OCS, no clear differences were observed in blood eosinophil kinetics
(Figure 6B). This indicates that the observed differences in eosinophil kinetics in EA patients

206
 Granulocyte kinetics in chronic pulmonary diseases

compared to CF patients and healthy controls are unlikely to be caused by the use of
corticosteroids.

Figure 8. Percentages of neutrophil

subsets. During cell sorts, the
percentages of CD16low/CD62Lhigh (A) and
CD16high/CD62Llow (B) neutrophils were
determined. CF patients have a trend
towards a higher percentage of immature
CD16low/CD62Lhigh neutrophils and a
significant increase in the percentage of
CD16high/CD62Llow neutrophils, indicating
an ongoing inflammation in CF patients.
Data represent means of 4-6
measurements for each volunteer, with
N=5 for controls, N=6 for CF patients and
N=9 for EA patients.

Granulocyte kinetics in sputum

One of our most remarkable findings was that neutrophil 2H-enrichment kinetics in sputum
almost exactly followed blood neutrophil deuterium enrichments. This implies that
neutrophils are continuously replaced by neutrophils from the bloodstream, and thus remain
intact in the sputum only very shortly. It would also explain the lack of effect of treatment
with GCS, as the turnover of sputum neutrophils may be too high to be able to detect
differences in DNA-enrichment with 6,6-2H2-glucose. An alternative hypothesis is that
neutrophils are recruited to the sputum only in response to sputum induction (46) and are
not present under normal conditions in these patients. We deem this explanation unlikely
because high neutrophil numbers are present in spontaneous samples in CF patients (one of 10
which is presented in this study). Another argument is the finding that sputum neutrophils
are often found in tightly packed sputum plugs that are unlikely to be formed during short
term sputum induction.
CD62Llow/CD16low sputum neutrophils from CF patients seemed to have delayed 2H-
enrichment kinetics, although the number of samples with adequate cell numbers was low.
The delay in uplabelling is approximately one day, indicating that at least some neutrophils
survive in sputum for one day. In vitro, CD62Llow/CD16low have been shown to be pre-
apoptotic (47), which also fits with CD62Llow/CD16low sputum neutrophils in CF patients being
older cells.
Although the number of sputum eosinophil samples was relatively small, there is a
good indication for differences in kinetics between blood and sputum cells of EA patients. At
day 7, when most volunteers had a high DNA enrichment in their circulating eosinophils, the
enrichment in sputum eosinophils was only half of that in blood. In addition, two out of three
later time points show that enrichment remained high when blood eosinophil DNA
enrichments were approaching their baseline levels. Together, these data suggest that

207
Chapter 10

eosinophils in EA patients reside longer in sputum than in blood: when the first labelled
eosinophils enter the sputum, the presence of long-lived unlabelled sputum eosinophils
lowers the observed average 2H-DNA enrichment in sputum, and the longevity of the newly-
entered labelled eosinophils subsequently causes the average deuterium labelling in sputum
to stay high for a relatively long period. The suggested long lifespan of eosinophils in sputum
may explain why preventing eosinophil production with anti-IL-5 does not deplete lung or
sputum eosinophils very well (18, 19). Although for CF patients, the differences between
blood and sputum eosinophils were less pronounced, they also showed a trend towards a
longer eosinophil survival in sputum, as in most patients eosinophil DNA enrichment
remained high when enrichment in blood eosinophils was already declining.
Interestingly, one EA patient and one CF patient showed a higher eosinophil 2H
enrichment in sputum than in the circulation during the uplabelling phase of the experiment.
Similarly, in one EA patient the same was observed for neutrophils. Even though these are
only few data points, it suggests that labelled cells can appear in sputum before they appear
in the circulation. This might be caused either by preferential homing of some eosinophils to
the lungs or by production of eosinophils from progenitors in bronchial tissue, as previously
reported (12).

Lack of effect of dornase alpha on free DNA kinetics

Free DNA is a major component of the sputum of CF patients (figure 2) and causes a high
sputum viscosity (48), which can lead to pulmonary obstruction. We show that 2H-enrichment
remains high in free DNA of CF patients for a longer period of time than sputum neutrophil
DNA, indicating that free DNA is more slowly cleared from the lungs of CF patients. Dornase
alpha did not appear to have an effect on the clearance of free DNA albeit the numbers of
patient samples included were low. The methods used in our study allow for the isolation of
DNA fragments as small as 200bp, larger than often observed in patients with or without
Dornase alpha (49). Thus, Dornase alpha might not reduce the DNA length enough to affect
enrichment kinetics as determined by our method.

Conclusions
In summary, marked differences in granulocyte kinetics were found in patients with
neutrophil and eosinophil driven diseases: Neutrophilic inflammation in the airways of CF
patients leads to an accelerated maturation of neutrophils in the bone marrow. Eosinophilic
inflammation, on the other hand, results in a slower eosinophil maturation, but more rapid
turnover in the total eosinophil compartment. The enrichment of neutrophils in sputum
mimics that of blood neutrophils, indicating that these cells have a short residence time in
sputum. Our data suggest, however, that eosinophils survive much longer in sputum than in
blood. If anything, in patients taking OCS, eosinophils seem to survive even longer. Taken
together, neutrophilic and eosinophilic inflammation have very different effects on
granulocyte kinetics.

208
 Granulocyte kinetics in chronic pulmonary diseases

References
1. Bousquet J, Chanez P, Lacoste JY, Barneon G, Ghavanian N, Enander I, Venge P,
Ahlstedt S, Simony-Lafontaine J, Godard P, et al. Eosinophilic inflammation in asthma.
The New England journal of medicine. 1990;323(15):1033-9.
2. Simpson JL, Scott R, Boyle MJ, and Gibson PG. Inflammatory subtypes in asthma:
assessment and identification using induced sputum. Respirology. 2006;11(1):54-61.
3. Cantin AM, Hartl D, Konstan MW, and Chmiel JF. Inflammation in cystic fibrosis lung
disease: Pathogenesis and therapy. J Cyst Fibros. 2015;14(4):419-30.
4. Fortunati E, Kazemier KM, Grutters JC, Koenderman L, and Van den Bosch v J. Human
neutrophils switch to an activated phenotype after homing to the lung irrespective of
inflammatory disease. Clinical and experimental immunology. 2009;155(3):559-66.
5. Tirouvanziam R, Conrad CK, Bottiglieri T, Herzenberg LA, Moss RB, and Herzenberg LA.
High-dose oral N-acetylcysteine, a glutathione prodrug, modulates inflammation in
cystic fibrosis. Proc Natl Acad Sci U S A. 2006;103(12):4628-33.
6. Tak T, Hilvering B, Tesselaar K, and Koenderman L. Similar activation state of
neutrophils in sputum of asthma patients irrespective of sputum eosinophilia. Clinical
and experimental immunology. 2015;182(2):204-12.
7. Fujisawa T, Kephart GM, Gray BH, and Gleich GJ. The neutrophil and chronic allergic
inflammation. Immunochemical localization of neutrophil elastase. Am Rev Respir Dis.
1990;141(3):689-97.
8. Halmerbauer G, Arri S, Schierl M, Strauch E, and Koller DY. The relationship of
eosinophil granule proteins to ions in the sputum of patients with cystic fibrosis.
Clinical and experimental allergy : journal of the British Society for Allergy and Clinical
Immunology. 2000;30(12):1771-6.
9. Downey DG, Martin SL, Dempster M, Moore JE, Keogan MT, Starcher B, Edgar J, Bilton 10
D, and Elborn JS. The relationship of clinical and inflammatory markers to outcome in
stable patients with cystic fibrosis. Pediatr Pulmonol. 2007;42(3):216-20.
10. Marcos V, Zhou-Suckow Z, Onder Yildirim A, Bohla A, Hector A, Vitkov L, Krautgartner
WD, Stoiber W, Griese M, Eickelberg O, et al. Free DNA in cystic fibrosis airway fluids
correlates with airflow obstruction. Mediators Inflamm. 2015;2015(408935.
11. (GINA). Global Initiative for Asthma. Gina report, global strategy for asthma
management and prevention. Updated February 2014 Accessed February 2, 2014.
12. Menzies-Gow A, Flood-Page P, Sehmi R, Burman J, Hamid Q, Robinson DS, Kay AB, and
Denburg J. Anti-IL-5 (mepolizumab) therapy induces bone marrow eosinophil
maturational arrest and decreases eosinophil progenitors in the bronchial mucosa of
atopic asthmatics. The Journal of allergy and clinical immunology. 2003;111(4):714-9.
13. Nair P, Pizzichini MM, Kjarsgaard M, Inman MD, Efthimiadis A, Pizzichini E, Hargreave
FE, and O'Byrne PM. Mepolizumab for prednisone-dependent asthma with sputum
eosinophilia. The New England journal of medicine. 2009;360(10):985-93.
14. Haldar P, Brightling CE, Hargadon B, Gupta S, Monteiro W, Sousa A, Marshall RP,
Bradding P, Green RH, Wardlaw AJ, et al. Mepolizumab and exacerbations of

209
Chapter 10

refractory eosinophilic asthma. The New England journal of medicine.

2009;360(10):973-84.
15. Ortega HG, Liu MC, Pavord ID, Brusselle GG, FitzGerald JM, Chetta A, Humbert M, Katz
LE, Keene ON, Yancey SW, et al. Mepolizumab treatment in patients with severe
eosinophilic asthma. The New England journal of medicine. 2014;371(13):1198-207.
16. Green RH, Brightling CE, McKenna S, Hargadon B, Parker D, Bradding P, Wardlaw AJ,
and Pavord ID. Asthma exacerbations and sputum eosinophil counts: a randomised
controlled trial. Lancet. 2002;360(9347):1715-21.
17. Adachi T, Motojima S, Hirata A, Fukuda T, Kihara N, Kosaku A, Ohtake H, and Makino
S. Eosinophil apoptosis caused by theophylline, glucocorticoids, and macrolides after
stimulation with IL-5. The Journal of allergy and clinical immunology. 1996;98(6 Pt
2):S207-15.
18. Flood-Page PT, Menzies-Gow AN, Kay AB, and Robinson DS. Eosinophil's role remains
uncertain as anti-interleukin-5 only partially depletes numbers in asthmatic airway.
Am J Respir Crit Care Med. 2003;167(2):199-204.
19. Flood-Page P, Swenson C, Faiferman I, Matthews J, Williams M, Brannick L, Robinson
D, Wenzel S, Busse W, Hansel TT, et al. A study to evaluate safety and efficacy of
mepolizumab in patients with moderate persistent asthma. Am J Respir Crit Care Med.
2007;176(11):1062-71.
20. Chung KF, Wenzel SE, Brozek JL, Bush A, Castro M, Sterk PJ, Adcock IM, Bateman ED,
Bel EH, Bleecker ER, et al. International ERS/ATS guidelines on definition, evaluation
and treatment of severe asthma. The European respiratory journal : official journal of
the European Society for Clinical Respiratory Physiology. 2014;43(2):343-73.
21. Ballmann M, Junge S, and von der Hardt H. Low-dose methotrexate for advanced
pulmonary disease in patients with cystic fibrosis. Respir Med. 2003;97(5):498-500.
22. Wark PA, and Gibson PG. Clinical usefulness of inflammatory markers in asthma. Am J
Respir Med. 2003;2(1):11-9.
23. Cox G. Glucocorticoid treatment inhibits apoptosis in human neutrophils. Separation
of survival and activation outcomes. J Immunol. 1995;154(9):4719-25.
24. Meagher LC, Cousin JM, Seckl JR, and Haslett C. Opposing effects of glucocorticoids on
the rate of apoptosis in neutrophilic and eosinophilic granulocytes. J Immunol.
1996;156(11):4422-8.
25. Nguyen LT, Lim S, Oates T, and Chung KF. Increase in airway neutrophils after oral but
not inhaled corticosteroid therapy in mild asthma. Respir Med. 2005;99(2):200-7.
26. Cartwright GE, Athens JW, and Wintrobe MM. The Kinetics of Granulopoiesis in
Normal Man. Blood. 1964;24(780-803.
27. Pillay J, den Braber I, Vrisekoop N, Kwast LM, de Boer RJ, Borghans JA, Tesselaar K, and
Koenderman L. In vivo labeling with 2H2O reveals a human neutrophil lifespan of 5.4
days. Blood. 2010;116(4):625-7.
28. Arndt H, Kubes P, Grisham MB, Gonzalez E, and Granger DN. Granulocyte turnover in
the feline intestine. Inflammation. 1992;16(5):549-59.

210
 Granulocyte kinetics in chronic pulmonary diseases

29. Tak T, Tesselaar K, Pillay J, Borghans JA, and Koenderman L. What's your age again?
Determination of human neutrophil half-lives revisited. Journal of leukocyte biology.
2013.
30. Ruparelia P, Szczepura KR, Summers C, Solanki CK, Balan K, Newbold P, Bilton D, Peters
AM, and Chilvers ER. Quantification of neutrophil migration into the lungs of patients
with chronic obstructive pulmonary disease. Eur J Nucl Med Mol Imaging.
2011;38(5):911-9.
31. Summers C, Singh NR, White JF, Mackenzie IM, Johnston A, Solanki C, Balan KK, Peters
AM, and Chilvers ER. Pulmonary retention of primed neutrophils: a novel protective
host response, which is impaired in the acute respiratory distress syndrome. Thorax.
2014;69(7):623-9.
32. Macallan DC, Fullerton CA, Neese RA, Haddock K, Park SS, and Hellerstein MK.
Measurement of cell proliferation by labeling of DNA with stable isotope-labeled
glucose: studies in vitro, in animals, and in humans. Proc Natl Acad Sci U S A.
1998;95(2):708-13.
33. Kelly A, and Moran A. Update on cystic fibrosis-related diabetes. J Cyst Fibros.
2013;12(4):318-31.
34. Macallan DC, Asquith B, Zhang Y, de Lara C, Ghattas H, Defoiche J, and Beverley PC.
Measurement of proliferation and disappearance of rapid turnover cell populations in
human studies using deuterium-labeled glucose. Nature protocols. 2009;4(9):1313-27.
35. Rutgers SR, Timens W, Kaufmann HF, van der Mark TW, Koeter GH, and Postma DS.
Comparison of induced sputum with bronchial wash, bronchoalveolar lavage and
bronchial biopsies in COPD. The European respiratory journal : official journal of the
European Society for Clinical Respiratory Physiology. 2000;15(1):109-15.
36. Efthimiadis A, Spanevello A, Hamid Q, Kelly MM, Linden M, Louis R, Pizzichini MM,
Pizzichini E, Ronchi C, Van Overvel F, et al. Methods of sputum processing for cell
10
counts, immunocytochemistry and in situ hybridisation. The European respiratory
journal Supplement. 2002;37(19s-23s.
37. Simpson JL, McElduff P, and Gibson PG. Assessment and reproducibility of non-
eosinophilic asthma using induced sputum. Respiration; international review of
thoracic diseases. 2010;79(2):147-51.
38. Steinbach KH, Schick P, Trepel F, Raffler H, Dohrmann J, Heilgeist G, Heltzel W, Li K,
Past W, van der Woerd-de Lange JA, et al. Estimation of kinetic parameters of
neutrophilic, eosinophilic, and basophilic granulocytes in human blood. Blut.
1979;39(1):27-38.
39. Gutteridge C, and Kuhn RJ. Pulmozyme--Dornase alfa. Pediatr Nurs. 1994;20(3):278-9.
40. Couturier O, Bodet-Milin C, Querellou S, Carlier T, Turzo A, and Bizais Y. Gastric
scintigraphy with a liquid-solid radiolabelled meal: performances of solid and liquid
parameters. Nuclear medicine communications. 2004;25(11):1143-50.

211
Chapter 10

41. Terashima T, Wiggs B, English D, Hogg JC, and van Eeden SF. Polymorphonuclear
leukocyte transit times in bone marrow during streptococcal pneumonia. The
American journal of physiology. 1996;271(4 Pt 1):L587-92.
42. Aglietta M, Piacibello W, Sanavio F, Stacchini A, Apra F, Schena M, Mossetti C, Carnino
F, Caligaris-Cappio F, and Gavosto F. Kinetics of human hemopoietic cells after in vivo
administration of granulocyte-macrophage colony-stimulating factor. The Journal of
clinical investigation. 1989;83(2):551-7.
43. Moser C, Jensen PO, Pressler T, Frederiksen B, Lanng S, Kharazmi A, Koch C, and Hoiby
N. Serum concentrations of GM-CSF and G-CSF correlate with the Th1/Th2 cytokine
response in cystic fibrosis patients with chronic Pseudomonas aeruginosa lung
infection. APMIS. 2005;113(6):400-9.
44. Wood LJ, Inman MD, Watson RM, Foley R, Denburg JA, and O'Byrne PM. Changes in
bone marrow inflammatory cell progenitors after inhaled allergen in asthmatic
subjects. Am J Respir Crit Care Med. 1998;157(1):99-105.
45. Baigelman W, Chodosh S, Pizzuto D, and Cupples LA. Sputum and blood eosinophils
during corticosteroid treatment of acute exacerbations of asthma. Am J Med.
1983;75(6):929-36.
46. Nightingale JA, Rogers DF, and Barnes PJ. Effect of repeated sputum induction on cell
counts in normal volunteers. Thorax. 1998;53(2):87-90.
47. Homburg CH, de Haas M, von dem Borne AE, Verhoeven AJ, Reutelingsperger CP, and
Roos D. Human neutrophils lose their surface Fc gamma RIII and acquire Annexin V
binding sites during apoptosis in vitro. Blood. 1995;85(2):532-40.
48. Picot R, Das I, and Reid L. Pus, deoxyribonucleic acid, and sputum viscosity. Thorax.
1978;33(2):235-42.
49. Brandt T, Breitenstein S, von der Hardt H, and Tummler B. DNA concentration and
length in sputum of patients with cystic fibrosis during inhalation with recombinant
human DNase. Thorax. 1995;50(8):880-2.

212
 Granulocyte kinetics in chronic pulmonary diseases

Supplementary Figures

10

Supplementary figure S1. Sort logic. Blood granulocytes (A) were identified as SSChigh singlets and divided into
neutrophils and eosinophils based on CD9 and CD19. For cells from CF (B) and EA (C) sputums, an additional
CD14--/LiveDEAD- gate was used. Some CF sputums contained large numbers of CD16low neutrophils and were
therefore also isolated. Eosinophils (D) and neutrophils from both CF (E) and EA (F) sputum samples were largely
intact after sort as determined by microscopic evaluation of MGG stained cytospin slides. Objective used: 100x,
scale bar indicates 10μm.

213
Chapter 11

General discussion
Chapter 11

216
 General discussion

The research presented in this thesis is focused on the maturation and kinetics of granulocytes
and monocytes in healthy humans, during acute inflammation and during chronic pulmonary
inflammation. The main method to assess cellular kinetics in this thesis was in vivo pulse-
chase labelling with 6,6-2H2-glucose. This non-toxic label was orally administered to
volunteers and incorporated into the DNA of dividing cells during the S-phase of the cell cycle.
The label could later be detected in the DNA of cell populations by GC-MS analysis. As the
lifecycle of a cell depends on several factors, interpretation of these experiments was done
by mathematical/computational modelling. In the first part of the thesis this method was
applied to healthy humans to estimate the circulatory lifespans of neutrophils, eosinophils,
basophils and monocytes and was used to assess the developmental relationship between
the three different monocyte subsets. The second part of the thesis focused on the neutrophil
subsets that are found in the circulation during severe inflammatory conditions. The
developmental relation between these subsets was discussed as well as their role in a model
of influenza infection. The third and last part of the thesis discussed the role of neutrophils
and eosinophils in pulmonary inflammatory disease. In this chapter, the conclusions of these
three different parts are discussed in a broader perspective.

Pulse length of 6,6-2H2-glucose

During the course of a 6,6-2H2-glucose labelling experiment the amount of label is not
completely constant. In addition, the uptake of 6,6-2H2-glucose can differ between individuals
or patient groups and, thereby, influence the incorporation of label into the DNA. To remove
this source of intra-individual variation, measured DNA 2H-enrichments were corrected for
differences in 6,6-2H2-glucose availability in plasma. This was done by determining the plasma
glucose enrichment during the labelling procedure and estimating in each individual how
much label was available for incorporation into the DNA. Subsequently, using the estimated
fraction label available during the labelling procedure, measured DNA enrichments were
corrected to the values they would have had if all newly incorporated nucleotides were 2H
labelled (1). This correction for label availability was a vital part in the determination of 2H 11
enrichment kinetics and, ultimately, the estimation of cell kinetics. Several important issues
need to be addressed that are important for this estimation:

1. The fraction of 6,6-2H2-glucose needs to be monitored in the circulation. This can easily
be done by taking small amounts of blood and measuring the enrichment by GC-MS,
as performed in all the labelling studies presented in this thesis and previous
publications(1).
2. It is necessary to know the fraction of de novo synthesised nucleotides compared to
nucleotides originating from the salvage pathway; only de novo synthesised
nucleotides can incorporate 6,6-2H2-glucose. As shown in chapter 3, this can be
estimated by growing the cell of interest in medium containing a known amount of
label for several divisions and measuring the DNA enrichment.

217
Chapter 11

3. Some aspects of the behaviour of label (e.g. metabolism) inside the cell need to be
known and, specifically, the effective pulse-length for the cell under investigation. We
found that this effective pulse length may differ from the period the label was
administered (see chapter 3). Firstly, it takes more than one hour until 6,6-2H2-glucose
is metabolised into adenosine and incorporated into the DNA (see figure 1 of Chapter
3, as the first label is detected at 2h after the start of the experiment. Secondly, after
removal of label, we found that in vitro grown HL-60 cells keep incorporating 2H
labelled adenosine for several hours until they reach a plateau of 2H-enrichment (at
around 10h) that corresponds to a situation in which the maximal amount of 2H is
incorporated into the DNA (estimated at 24h). By combining these two findings, the
pulse-length of the labelling procedure can be estimated at approximately 8 h.

The importance of correct estimations of the plasma glucose correction is illustrated in figure
1. It shows an estimation of circulatory basophil kinetics by both the model based on ordinary
differential equations (ODE model) and the agent based (AB) model (see chapter 3) with
different assumed lengths of the effective pulse length. When assuming that 6,6-2H2-glucose
is only incorporated during the 6h that label is administered, neither model predicts a high
enough enrichment to produce a good fit (see figure 1A-B). The fit improves starting from a
pulse-length of 8h. The quality of the fit increases up to a 10h pulse length, after which the
quality decreases again. The resulting estimates of circulatory basophil lifespans and division
times (tdiv) of their BM precursors are also significantly affected by the effective pulse length,
especially when using the ODE model. Using this model, the estimated lifespan of basophils
increases from 4 days when assuming a 8h pulse length, to more than 20 days assuming a 20h
pulse-length. The best fits of the AB model also show longer basophil circulatory lifespans
with increasing effective pulse-lengths, but the effect is not as strong as for the ODE model.
Taken together, these results clearly show that correct estimates for pulse-length are
paramount for an adequate estimation of cell kinetics using 6,62H2-glucose. Considering that
none of the previously published studies using 6,6-2H2-glucose labelling have determined the
effective pulse length (1), such estimates may have to be re-evaluated taking into account the
importance of the pulse-length.
In chapter 3 we have estimated the effective pulse-length for the pro-myelocytic HL-
60 cell line to be 8h when 6,6-2H2-glucose is administered for 6h. When assuming an 8h
instead of a 6h effective pulse-length fits improve, both for the ODE and AB model.
Performing this assay in vitro is not ideal, however, as one should take into account that the
concentrations of nucleotides and glucose differ between human blood and cell culture
media. Also, even though HL-60 cells originate from promyelocytes, they are not likely to
behave exactly like neutrophil myelocytes in vivo. Unfortunately, it is very challenging to
perform bone marrow studies with multiple sampling in healthy volunteers in vivo.

218
 General discussion

Figure 1. Effect of 6,62H2-glucose effective pulse length on lifespan estimations. Basophil DNA 2H-enrichment
data from an experiment were 6,62H2-glucose was administered for 6 hours were fitted to an ODE model
assuming different effective 6,62H2-glucose pulse lengths. Best fits of the ODE (A) and the AB (B) model show
differences when increasing the effective pulse-length. The resulting estimate of these fits (C,D) show that
assuming longer effective pulse-lengths leads to longer lifespan estimates. Basophil 2H-enrichments represent 6
measurements from 6 individuals; AB parameter estimates represent the mean of the 10 best fits.

Re-evaluating granulocyte circulatory lifespans.

As stated in chapters 2 and 3 of this thesis, a neutrophil lifespan of 10h as found in most text-
books is based on experiments (2-4) under conditions that have been proven to severely
affect neutrophil kinetics (5, 6). The previous attempt using 63 day non-toxic 2H2O labelling
on the other hand, suffered from a difficulty in distinguishing between blood and BM kinetics
11
(7). In addition, the dynamics of body water may be too slow to reliably measure the kinetics
of a cell subset as fast as granulocytes. The latter problem can be circumvented by the use of
short-term pulse labelling with 6,6-2H2-glucose, since the turnover of glucose in the blood is
very high (half-life of approximately 0.5 h) (8).
A commonly used method to analyse experimental labelling data is by fitting a model
described by ordinary differential equations (ODE) to the experimental measurements.
Unfortunately, for neutrophils, such models turned out to have difficulties distinguishing
between blood and bone marrow (BM) kinetics. This is illustrated in chapter 3, where the two
ODE models used to estimate the circulatory neutrophil lifespan lead to either a very short
lifespan in blood or a very short time between divisions in the BM (tdiv). For neutrophils, this
would mean that on average they remain in the circulation for less than 4h, shorter than any
previous report studying the neutrophil lifespan. Alternatively, the model suggests a much
longer circulatory lifespan of several days, in line with Pillay et al (9), yet with an unlikely short

219
Chapter 11

time between progenitor divisions in the BM (less than 1h). A similar situation is seen when
estimating eosinophil circulatory lifespans from a deuterium labelling dataset (see chapter 4,
figure 1). For both basophils and pDCs, the ODE model does not estimate short lifespans or
tdiv. In contrast it estimates nearly equal values for both parameters (see table 1 for estimates
of several different cell types).
Estimation of monocyte kinetics by ODE models proved to be more feasible. We
studied whether monocyte labelling data were compatible with a linear differentiation
pattern from classical (CM) to intermediate (IM) to non-classical monocytes (NCM). Since
NCM have kinetics very different from IM, they must have a relatively long circulatory lifespan
compared to IM. One might expect the kinetics of IM and NCM and, specifically, the longer
NCM lifespan to place limits on what value the BM and CM can reach (e.g. their lifespan
cannot be very short) However, when analysing CM as a single subset (the parallel model, as
described below), results are only slightly different (see table 1).
The ODE models presented in chapters 3 estimated circulatory lifespans of
neutrophils, eosinophils and basophils that were either much shorter than those suggested
by the AB model, or much longer, yet with extremely rapidly dividing progenitor cells (see
table1). The estimates obtained in the early studies (3, 4, 10-12) are more likely to be
underestimating circulatory lifespans due to ex vivo manipulation or label toxicity. Therefore,
we sought an alternative strategy to analyse the results from our labelling studies. One of
these alternatives is the agent-based model (AB model). In AB models, a large number of
individual “agents“ can be modelled, each with their own behaviour. AB models are
particularly well-suited to model the behaviour of swarming animals (13). In the case of the
AB model in chapter 3 these “agents” are individual cells. This model might be better suited
to describe the behaviour of bone marrow cells, as ODE models assume exponential kinetics
of cells, e.g. they assume that there are always cells that divide within an hour after their last
division, even if on average cells divide every few days. As stated before (chapter 3),
neutrophil progenitors were shown to have very little variation in time between divisions (14),
which can be mimicked in an AB model. AB models, on the other hand, have practical
disadvantages. The behaviour of each agent needs to be tracked and updated separately and,
therefore, AB models require much more computer processing time than ODE models. For
example, the data in chapter 4 required more than one week of full-time calculations on a
normal desktop PC.
Another advantage of ODE models compared to AB models is that they can perform a
parameter estimation, in which parameter values are changed in small steps. During each
step, the quality of the fit is determined. If the quality of the fit has improved, the parameters
are changed again in the same direction. If not, the parameter values are changed into a
different direction, until changing the parameters does not significantly improve the fit
anymore. Since this is difficult to perform with an AB model, the models in chapters 3 and 4
were run with a large number of different parameter value combinations. The runs of the
model that described the data best seemed biologically plausible. Interestingly, the AB model

220
 General discussion

estimated a neutrophil circulatory lifespan somewhere between the two different estimates
of the ODE model.

Table 1. Estimated circulatory lifespans, time between divisions in the BM (tdiv) and PMPtt for different models
in days.

ODE fixed PMPtt ODE variable PMPtt Agent-based Literature*

Optimum** / Optimum 1 Optimum Optimum Optimum 2

Parameter (Blood) 2 (BM) 1 (Blood) (BM)
Lifespan 4.8 0.15 4.5 0.04 2.5-4 1
Neutrophils tDiv 0.04 4.5 0.1 4.6 1-2 0.75-1
PMPtt§ 6.5 6.3 6.2 6.2 5.7 2.8
Lifespan 8.1 0.6 8.3 0.04 2.8-3.7 1.1
Eosinophils tDiv 0.8 8.4 0.1 8.4 3-5 -
PMPtt§ 4.4 4.7 3.7 3.8 5.1 4.3
Lifespan 3.9 3.6 2.9 0.2 8-12 3.7
Basophils tDiv 3.9 4.0 2.7 4.2 2-3.5 -
PMPtt§ 6.6 6.7 2.2 2.8 5.7 4.3
Lifespan 4.1 4.2 4.8 1.9 6-10 -
pDC tDiv 3.9 4.0 0.6 5.7 1-3 -
PMPtt§ 4.3 4.2 0.005.6 0.4 4-4.3 -
ODE parallel ODE linear
differentiation differentiation
Optimum** /
Blood BM Blood BM
Parameter
Lifespan 2.3 0.7 2.0 0.7 1 4.3
Classical
tDiv 0.7 2.3 1.1 2.3 2.5 3
Monocyte†
PMPtt§ 1.2 2.1 1.6 1.9 2 0.5-1.1
* 3H-Tdr data from (10) and (15)
** By choosing different initial parameter values for tdiv and lifespan, the ODE model will find optimal fits with
a long circulatory lifespan and short tDiv (optimum 1), or a long tdiv and short cirulatory lifespan (optimum 2),
which describe the data equally well. For basophils, and pDC’s, these results were very similar in case of the ODE 11
model with a fixed PMPtt, indicating that there might be only one optimum.
† For CM, models with either CM developing separately or linearly into IM and NCM are depicted.
§ PMPtt’s represent the PMPtt of all cells, of the fastest cell or the average of all cells for the ODE model assuming
that all cells have exactly the same PMPtt (fixed PMPtt), the ODE model allowing for a normally distributed
variation in PMPtt (variable PMPtt) and AB models, respectively.

When adapting the AB model to represent the three monocyte subsets and running
this model with the estimates obtained from the ODE model, the model produced fairly good
fits. When slightly increasing tdiv, the fits improved even further (see figure 2). Interestingly,
fits with a good quality were only produced when assuming a relatively large variation around
all parameters (a SD of approximately 40% of values for tdiv, PMPtt and the circulatory
lifespans of each subset). This is in contrast to the best fits of the AB model to the neutrophil
data, where the best fits were always with little variation around the parameters (SD of <20%)
(see chapter 3, supplementary figure S2). This suggests that there is a difference between

221
Chapter 11

neutrophils and monocytes, which makes the former more difficult to model with ODE
models. The reason for this might lie in monocyte progenitors behaving differently, for
instance by having a more random time between divisions as compared to neutrophil
progenitors. Unfortunately, however, no kinetic data is available regarding monocyte or
neutrophil progenitors in the BM.

ODE model Agent-based model

0.08 8

Classical
C
0.06 6 I
Intermediate
Non-classical
0.04 4

0.02 2

0.00 0
0 5 10 15 20 0 5 10 15 20
Time (days) Time (days)

Figure 2. ODE and AB model fits of the monocyte DNA 2H–enrichment kinetics. The monocyte lifespans
estimated by the ODE model described in chapter 4 (A) were used for a run with the AB model (B), which showed
a similar fit to the data, although only when assuming a longer tdiv than predicted by the ODE model (60h instead
of 48h), which might be attributable to the differences behaviour of the BM in both models.

As the ODE and AB models produce different estimates for some cells (i.e.
granulocytes), one may wonder what is the underlying difference between the AB and the
ODE models? As stated above, one difference may lie in the timing of neutrophil progenitor
divisions, while another difference may be related to the way cells enter and exit the PMP. In
the AB model, progenitor cells can divide with very little variation between divisions, whereas
ODE models assume exponential behaviour of cells. As a result, in ODE models, some cells can
divide again very shortly after their previous division, regardless of the value of tdiv. This is
illustrated in chapter 3 by the rapid decrease of label in the BM after cessation of label
intake(figure 2A). In the AB model, on the other hand, progenitor cells cannot divide quickly
after their previous division (chapter 3, figure 3), which results in a slower initial decrease in
BM 2H enrichment.
After a progenitor division, a cell can either enter the PMP or stay a progenitor and
keep dividing. As the AB model assumes the PMP to be a conveyor-belt, a long tdiv combined
with a very short circulatory lifespan leads to a rapid increase in label in the circulation when
the first cells exit the PMP. If the circulatory lifespan is short, these cells will rapidly exit the
circulation again and be replaced by unlabelled cells. When the progenitor cells in the BM
divide again, new labelled cells will enter and leave the circulation again, albeit with only half
the amount of label in each cell. The resulting enrichment will look like a series of peaks in 2H-
enrichment with low enrichment in between. An example of this can be seen in figure 3C of
chapter 3. The ODE model makes a different assumption about the BM and PMP. It assumes

222
 General discussion

cells to divide and enter the PMP randomly. Thus, the fraction label entering the PMP is
exactly the same as the fraction label in the BM at that time. After a delay in the PMP, the
blood kinetics closely reflect the BM kinetics in the case of a short circulatory lifespan. Since
the uplabelling in the BM is very rapid (8h), this assumption leads to a very rapid increase in
label in the periphery (such as in chapter 3, figure 2).
Taken together, chapters 2-5 demonstrate that one should be careful when using
mathematical/computational models for estimation of circulatory kinetics for cells, which are
predominantly produced in pools at locations other than the circulation. Chapter 3 and 4 and
table 1 show this to be the case for the three granulocytes, as ODE models have difficulties in
separating BM and blood kinetics of these cells. Also other important cell types, such as NK
cells (16) are to a large extent derived from progenitor cells in the BM under homeostasis and
their circulatory lifespans might, therefore, not easily be adequately assessed. It also asks for
a re-appraisal of labelling studies performed on cells produced in the BM, which are modelled
without taking the BM into account (17, 18). Direct measurements of 2H enrichment in bone
marrow samples would help to strengthen the conclusions in chapters 3 and 4. However,
these data are not only difficult to obtain, but may also turn out to be difficult to interpret.
For example, not all neutrophil myelocytes are actively dividing in vitro, (19, 20) but the
fraction of these non-dividing (“lazy”) cells in vivo is unknown. Furthermore, some myelocytes
are derived from myelocyte divisions, but others are derived from differentiating pro-
myelocytes. If a large fraction of myelocytes were derived from promyelocytes (the myelocyte
precursor), it would still be difficult to interpret the myelocyte data, similar to the
interpretation of data from mature cells. Thus, detailed bone marrow 2H-enrichment data
may be needed to aid the estimation of granulocyte circulatory lifespans.

Linear or parallel monocyte differentiation

In chapter 5, our in vivo labelling data support the hypothesis that monocytes differentiate in
a linear fashion from classical monocytes (CM) via intermediate monocytes (IM) to non-
classical monocytes (NCM). On top of that, the calculations used in this chapter made 11
important predictions on the monocyte maturation process.
1. They predict that only 3% of the CM mature into IM, due to the large number of CM
compared to NCM and their relative circulatory lifespans.
2. For a good fit, the model required a delay between IM and NCM differentiation taking
place outside the circulation.
Data that support these predictions can be found in experiments showing CM to differentiate
into tissue dendritic cells (DC’s) (21) or skin macrophages (22), and the finding that NCM in
individuals with brain tumours contain tumour-specific peptides (unpublished data by Van
Dongen et al. presented at the NVVI 2014).
The monocyte 2H-labelling kinetics can also be explained with an alternative model,
assuming that each monocyte subset has its own progenitor in the bone marrow, PMPtt and
circulatory lifespan (figure 3). With this assumption, subset lifespans can be estimated for
each subset separately. This results in a somewhat longer estimated IM circulatory lifespan

223
Chapter 11

compared to the results from the linear differentiation model (1.6 vs 1.0 d) and a somewhat
shorter lifespan for both CM (0.5 vs 0.7 days) and NCM (2.8 vs 3.5d). Thus, although
monocytes might differentiate from CM into IM and NCM, the alternative is not ruled out on
the basis of the labelling data in chapter 5. In addition, several combinations of parallel and
linear differentiation are possible, such as CM maturing into IM, whereas NCM develop
separately. As long as these models do not violate the order of CM acquiring label first,
followed by IM and later NCM, they will most likely be able to explain our labelling data. A
developmental relation between CM, IM and NCM is supported by data obtained by gene
arrays, which showed that the transcriptome of IM lies in between that of CM and NCM (23).
Furthermore, surface marker expression data obtained by flow cytometry showed changes in
expression to be gradual from CM to NCM, without discrete differences in any of the
measured markers(24).

Figure 3. Linear versus parallel differentiation of the monocyte subsets. Monocytes have been hypothesized
to differentiate in a linear fashion from CM to IM to NCM (A). Alternatively, they could differentiate in parallel,
with each subset originating from a different progenitor. Both assumptions were tested using MONOLIX, as
described in chapter 5. Population fits of both models produce good fits (C-D), with small differences, such as in
the predicted moment of label appearing in the DNA of non-classical monocytes. The parallel differentiation
model predicts a somewhat shorter CM and NCM circulatory lifespan and a longer IM lifespan.

224
 General discussion

To definitively answer the question whether the three monocyte subsets differentiate
in a linear, parallel or mixed fashion, new experiments will have to be performed. Preferably,
these experiments include lineage tracing experiments utilizing genetic barcoding (such as
done in mice (25)) or by cell mutation analysis of somatic mutations in humans (26). These
experiments should be performed on a single-cell basis, to determine whether differences in
lineage are discrete or continuous.

Hypersegmented neutrophils
Under homeostatic conditions, the majority of circulating neutrophils appear to be one
homogeneous population. During acute inflammation, however, at least three neutrophil
subsets can be identified, each with a different nuclear morphology and functionality (27): 1.
Banded, CD16low/CD62Lhigh neutrophils, which are shown to be better at containing bacteria
than the other two subsets (unpublished results by P. Leliefeld). 2. CD16high/CD62Lhigh
neutrophils with a functionality and nuclear phenotype similar to that of neutrophils under
homeostasis and 3. Hypersegmented, CD16high/CD62Llow neutrophils which have been shown
to be capable of suppressing T-cell proliferation in vitro (27). In chapter 7, we induced
recruitment of these three neutrophil subsets to the circulation by injecting healthy
volunteers with a single dose of 2ng LPS per kg bodyweight. Three hours after injection, large
numbers of these three neutrophils subsets were present in the blood. In chapter 7, we
propose that the hypersegmented neutrophil is a separately developing neutrophil subset,
which is rapidly mobilized to the blood in response to inflammation. This is in contrast to our
initial idea that hypersegmented neutrophils are a further differentiation stage of normal
neutrophils induced by the injection LPS. This conclusion is based on several arguments:
1. Age. Mature and hypersegmented neutrophils have similar 2H-enrichment kinetics.
This suggests that if hypersegmented neutrophils have differentiated from mature
neutrophils in response to LPS, this differentiation would have to occur in less than 3
hours (the time between LPS injection and blood withdrawal).
2. Protein expression levels. These indicate that mature neutrophils are more similar to 11
the two day younger banded neutrophils than to the hypersegmented neutrophils,
even though the latter have had a similar maturation time. Thus, if hypersegmented
neutrophils were derived from normally segmented cells, the change in their
proteome would be larger than that induced by 2 days of maturation.
3. The translational capacity. This is very poor in hypersegmented neutrophils (see
chapter 7). In fact, when calculating the increase in abundance for each protein
upregulated in hypersegmented neutrophils compared to normally segmented
neutrophils and subsequently dividing this abundance by the total amount of detected
protein in hypersegmented cells, one can calculate newly synthesized protein (mass)
as follows:

225
Chapter 11

݂݅ �������������� � ��������������� :
����� ����������� �� �������� � ��������������� � ����������������
for each protein.

Now to determine the total percentage of newly synthesized proteins in hypersegmented

neutrophils we calculate:

∑ ����� ����������������
���������� ����� ����������� �
∑ ��������������

Based on our proteomics data, this calculation shows that 13% of the hypersegmented
neutrophil proteome is newly synthesized compared to normally segmented neutrophils.
When performing the same calculation for the amount of newly synthesized protein in the
differentiation from banded neutrophil into mature neutrophil, this percentage is “only”
7.6%. Taken together, this makes it unlikely that hypersegmented neutrophils differentiate
from normally segmented neutrophils during the 3 hours after LPS challenge.

Figure 4. CD62Llow neutrophils after administration

of Plerixafor. Blood was obtained 24h after
administration of Plerixafor and analysed for the
presence of CD62Llow neutrophils. Flow cytometric
data RBC lysed whole blood were gated as SSChigh
granulocytes, with exclusion of eosinophils,
monocytes and NK-cells by CD193, CD14 and CD56
expression, respectively.

Circumstantial evidence suggest that CD62Llow neutrophils are the result of inflammation as
they are described in many severe inflammatory conditions, including trauma (28), the human
endotoxemia model (27), burn injury (unpublished data by KM Groeneveld) and cancer (29).
In addition, this phenotype was described in more mild cases of immune activation in
otherwise healthy volunteers, such as after sleep deprivation (30) and after administration of
G-CSF (31). In all these studies it is unclear whether the cells are formed by activation during
the disease or whether they are mobilized from a pre-existing non-activated pool outside the
blood. An important finding in this regard is that CD62Llow neutrophils have been observed in
the blood after injection of plerixafor; a condition without inflammation. Plerixafor is a CXCR4

226
 General discussion

antagonist which is used to mobilize hematopoietic progenitor cells from the BM into the
blood (32). Signalling trough SDF1/CXCR4 axis has been described as a BM retention signal
(33). Thus, blockade of CXCR4 does not lead to inflammation, but it leads to recruitment of
CD62Llow neutrophils to the circulation (figure 4). This clearly indicates that CD62Llow cells can
enter the bloodstream most likely from the bone marrow in the absence of inflammation.
Many questions regarding the origin and function of hypersegmented neutrophils
remain, however, which will be further discussed in the remainder of the discussion.

The origin and function of neutrophil nuclear segmentation

In chapters 6 and 7 of this thesis, the function and origin of hypersegmented neutrophils has
been discussed. It is unknown why hypersegmented neutrophils show increased lobularity of
their nuclei or why neutrophils have a segmented nucleus in the first place. Segmentation of
the neutrophil (or neutrophil-homologues) nucleus is an evolutionary conserved trait, as it
has been reported in animals ranging from invertebrates (30) to fish, birds and mammals (27-
29) This clearly shows the evolutionary pressure to retain this nuclear shape. The shape of a
cell's nucleus is at least in part determined by the relative expression of Lamin-A/C versus
Lamin-B (34), with neutrophils losing Lamin-A/C expression during differentiation. The
importance of the Lamin-B receptor in nuclear segmentation comes from studies focussing
on mutations in this gene (32). Individuals carrying these mutations are hallmarked by
neutrophils with only two very round lobes when heterozygous and a completely round
nucleus when homozygous (35). This condition is called the Pelger-Huët anomaly.
Surprisingly, this condition is completely asymptomatic when heterozygous (36, 37). When
homozygous, skeletal defects occur, but immunological defects are rarely found (37, 38). The
only cases of Pelger-Huët anomaly in which immunological symptoms have been described
were characterized by an aberrant chromosome 22 (39). This indicates the importance of
additional genes for neutrophil nuclear segmentation, since none of the Lamins nor their
receptors are located on this chromosome and not all family members with Pelger-Huët
anomaly had this enlarged chromosome (39). 11
Wolf et al provided some evidence that nuclear segmentation allows neutrophils to
move through smaller gaps than cells with a round nucleus (40) and might, therefore,
facilitate extravasation. However, this study compared neutrophils (segmented nucleus) with
lymphocytes (round nucleus) and did not take into account that other differences between
these cell types might underlie differences in cellular movement. Data describing migration
and extravasation in Pelger-Huët are conflicting. One study showed impaired neutrophil
migration through skin windows (41). Another study showing impaired neutrophil migration
through a filter only in some patients (39), whereas in the same study no differences were
observed in other patients with a similar nuclear phenotype (39). In addition, normal
migration under an agar gel (42) or through filters (43) and the absence of clinical symptoms
in nearly all cases (36, 39) are indicative that migration of neutrophils with round nuclei is not
affected in Pelger-Huët anomaly.

227
Chapter 11

Based on the above, it is unlikely that neutrophil hypersegmentation leads to increased

migration. In fact, banded neutrophils migrate even better through a filter in a Boyden
chamber assay (44) and migration through a restrictive collagen matrix also showed no
differences in ability to squeeze through small pores between the three neutrophil subsets
(figure5A). Also, migration does not induce hypersegmentation, since neutrophils obtained
from the sputum of asthma patients show no increased lobularity (see fig 5C).

A B C

Band Mature Hyperseg

Figure 5. Effect of neutrophil segmentation on migration speed (A) and neutrophil nuclear segmentation in
samples obtained from (B) healthy volunteers 3h after injection of LPS or (C) from sputum samples from asthma
and CF patients and the synovium of patients with juvenile idiopathic arthritis (JIA) patients.

Alternatively, nuclear segmentation might reflect differences in chromosomal

localisation. Chromosomes have been shown to have specific localisations within the nucleus
for specific chromosomes, called chromosome territories (45). Chromosome territories have
been linked to transcriptional activity, allowing gene enhancers and silencers to affect
transcriptional activity on a different chromosome. In addition, the nuclear organisation
affects accessibility of DNA and thereby transcription (46). Chromosome rearrangements
have been shown to differ between cell types and during cell differentiation (46). For
neutrophils, the existence of chromosome territories is suggested by the fact that
chromosome localisation within the nucleus is not random (47). We did not asses
chromosome localisation in the neutrophil subsets, but alterations in chromosome territories
are likely due to the different shape of the nucleus and the large differences in expression of
proteins involved in transcription and translation as shown in chapter 7. In addition, porcine
neutrophils have been shown to change their chromosome arrangement after stimulation
with LPS (48), although subsequent research reported these differences to be very small (49).
Taken together, neutrophil hypersegmentation is unlikely to allow more rapid
migration and extravasation and neutrophil segmentation may not be of any consequence for
neutrophil function, as Pelger-Huët anomaly does not induce immunological defects.
Alternatively, differences in segmentation might reflect differences in transcriptional and
translational activity such as shown in chapter 7 and by de Kleijn et al(50).

228
 General discussion

Why do neutrophils get hypersegmented?

As described above, neutrophils lose Lamin-A/C expression during differentiation and Lamin-
B-receptor mutations lead to reduced nuclear segmentation. The proteomics analysis
performed on the three neutrophil subsets in chapter 7 does not show a difference in
expression of Lamin-B receptor, but expression of Lamin-B1 increased in hypersegmented
neutrophils (500 vs 430ppm, p<0.001) with a trend towards increased Lamin-B2 expression
(p=0.07 after multi-testing correction).
Determination of neutrophil hypersegmentation is hallmarked by the presence of
large numbers of neutrophils with ≥6 nuclear lobes (51). Hypersegmentation of neutrophil
nuclei has particularly been described in DNA-replication deficiencies, mostly caused by
vitamin B12 deficiency or DNA-replication poisons, although neutrophils with six lobes have
also been seen in about 10% of healthy children (52). CD62Llow neutrophils with an increased
nuclear lobularity, such as seen in the human endotoxemia model and in trauma patients (27),
usually have 4-5 nuclear lobes.
It is often suggested that neutrophils characterized by increased lobularity of their
nuclei represent older cells (53-55). However, in chapter 7 we provide evidence that these
cells do not have an increased cellular age, which is in agreement with previous experiments
with 3H-TdR in one individual (56). The latter study showed no difference in the PMPtt and
labelling kinetics between neutrophils with two and neutrophils with more than 4 lobes.
This does not mean, however, that neutrophils cannot become hypersegmented when
aging. For example, when culturing neutrophils ex vivo with the supernatant of UV irradiated
keratinocytes some neutrophils survived for at least 5 days (unpublished data by C. van Aalst).
After 2 days of culture, all of these neutrophils obtained a hypersegmented phenotype with
4-5 lobes arranged as a cloverleaf such as seen in murine neutrophils (chapter 8, figure 5).
These neutrophils with a cloverleaf shaped nucleus have also been observed in the synovial
fluid of juvenile idiopathic arthritis patients and in the sputum of some CF patients (figure 5C).
Cells with a cloverleaf type of nuclear segmentation are typically not found in the peripheral
blood under inflammatory conditions. Therefore, it is unknown whether this cloverleaf 11
hypersegmentation is a different form of hypersegmentation or a later stage of neutrophil
development.

Where do hypersegmented neutrophils come from?

In chapter 7, we propose that the hypersegmented neutrophil is a separately developing
neutrophil subset, which is rapidly mobilized to the blood. It remains unclear, however, where
these cells come from. This question is difficult to answer with the techniques we have been
using up to now. We detected the hypersegmented neutrophil subset by flow cytometry
based on CD16/CD62L expression (27) or, alternatively, by an increased expression of CD11c.
Unfortunately, neither low CD62L expression nor high CD11c expression is a conclusive
characteristic of a hypersegmented neutrophil. CD62L shedding is observed in neutrophil
endothelial transmigration (57), so neutrophils in tissues are CD62Llow regardless of their
nuclear phenotype. This is illustrated by the results shown in chapter 9, which show that

229
Chapter 11

sputum cells are low in CD62L , yet do not show increased nuclear segmentation (figure 5C).
CD11c surface expression might not be ideal for detection of hypersegmented neutrophils
either, because this is a protein stored in ficolin and gelatinase containing granules (58). The
protein remains on the surface after degranulation of these vesicles. Thus, also activated
neutrophils will show increased expression of this marker.
Even though the markers used to identify hypersegmented neutrophils are
ambiguous, it is tempting to speculate about the origin of hypersegmented neutrophils.
Several non-circulating neutrophil populations have been described in literature. For
example, reinfusion of radiolabelled neutrophils shows homing of these cells to the spleen,
lungs, liver and bone marrow of healthy volunteers (59, 60), although retention in the lung
was later described to be an artefact of ex vivo manipulation (6). In addition, in the spleen
two neutrophil subsets have been described that are phenotypically different from blood
neutrophils. One of these two subsets was described to be capable of suppressing T-cell
proliferation and have decreased CD62L expression (61). This would make the spleen a
potential site of origin for hypersegmented neutrophils. However, follow-up studies failed to
detect neutrophil subsets in fresh human spleen samples (62), questioning this hypothesis.
The detection of neutrophil homing to the liver under homeostasis is thought to be due to
the clearing of neutrophils from the circulation by the liver (63). Neutrophil numbers are low
in the liver under homeostasis (64), making the liver an unlikely location from which
hypersegmented neutrophils can be released during homeostasis. The last organ discussed
here to which large numbers of neutrophils have been shown to migrate under homeostasis
is the BM (59). The bone marrow contains large numbers of neutrophils in all stages of
differentiation and has also been implicated as a site of neutrophil clearance (65). In addition,
we have observed the presence of hypersegmented neutrophils in the bone marrow
(unpublished data) and blockade of the BM-homing receptor CXCR4 (66) recruits CD62Llow
neutrophils to the circulation as described above. Therefore, the BM seems a likely site of
origin for hypersegmented neutrophils.
Besides demonstrating the accumulation of neutrophils in several organs, reinfusion
of radiolabelled neutrophils into healthy humans has also indicated the existence of a
marginated pool of neutrophils (67). During these experiments, the number of radiolabelled
neutrophils that were recovered from the circulation after reinfusion was lower than
expected. Subsequent injection of high doses of corticosteroids or adrenalin led to recovery
of the expected number of radiolabelled neutrophils in the peripheral blood (68). Although
reinfusion of ex vivo manipulated neutrophils is far from ideal when investigating neutrophil
kinetics, exercise and injection of adrenalin or corticosteroids have been shown to rapidly
increase neutrophil counts in healthy volunteers (68). Experiments from our lab in which the
blood of rowers and cyclists was analysed before and after exercise indeed showed an
increased neutrophil count, but no evidence of additional CD62Llow neutrophils entering the
circulation (unpublished results from B. Hilvering and S. van Staveren). Therefore, it is unlikely
that hypersegmented neutrophils originate from the marginated pool.

230
 General discussion

A final possibility for the origin of hypersegmented neutrophils might be that cells
return from the tissue into the blood. These cells have been identified in the circulation in
healthy volunteers and in increased numbers during chronic inflammation (69). These reverse
transmigrated neutrophils were identified by their expression of ICAM-1 (CD54). This remains
a matter of debate as replicating studies have not been able to detect an appreciable
expression of ICAM-1 (CD54) on blood neutrophils (see chapter 9). Nevertheless, the process
of reverse transmigration has been shown extensively in vivo in mice and zebrafish (70, 71),
albeit only under inflammatory conditions. This raises the question whether hypersegmented
neutrophils are reverse transmigrated. Hypersegmented neutrophils do not express CD54
(27), and have a totally different receptor expression pattern compared to reverse
transmigrated neutrophils (27, 69). With respect to their function, reverse transmigrated
neutrophils have been shown to have normal phagosome acidification in zebrafish (72) as
opposed to hypersegmented human neutrophils, which do not have adequate phagosome
acidification (unpublished results by Pieter Leliefeld). Furthermore, reverse transmigrated
neutrophils are associated with causing lung damage in mice (70) instead of being
immunomodulatory like human hypersegmented neutrophils.
Taken together, we hypothesize that the BM may be the main source of CD62Llow
hypersegmented neutrophils, because: 1. Small numbers of hypersegmented neutrophils can
be observed in the BM and 2. Blockade of BM-homing receptor CXCR4 results in recruitment
of hypersegmented neutrophils to the circulation.

Sputum granulocytes
Sputum induction is a technique by which cells can be obtained from the central airways (73).
Both neutrophils and eosinophils can be observed in the sputum of asthma patients and are
thought to contribute to pathology in this disease (74, 75). Sputum eosinophilia is usually
defined as 2-3% eosinophils in sputum and the cut-off defining sputum neutrophilia ranges
from 40% to 76% neutrophils (76-78). Based on the percentage of these cells in induced-
sputum samples, asthma can be subdivided into four different immunological subtypes: 1. 11
eosinophilic asthma (EA) : patients with only sputum eosinophilia, 2. neutrophilic asthma
(NA): patients with only sputum neutrophilia 3. mixed eosinophilic/neutrophilic asthma:
patients with both cell types present in sputum and 4. paucigranulocytic asthma: patients
with neither sputum eosinophilia nor sputum neutrophilia (76). Determining the
immunological asthma subtype can be useful in the clinic as some types of medication are
only effective in patients with eosinophilic asthma (79) The results of sputum inductions have
also been used to adapt treatment regimens, leading to a lower number of exacerbations
(80).

Limitations of sputum inductions

Even though the classification of asthma into 4 immunological subtypes has proven to be
useful, it should be applied with caution. For example, as the cut-off for neutrophilic asthma
differs between studies (76, 77), patients can be classified as having neutrophilic asthma in

231
Chapter 11

one study, while being classified as paucigranulocytic asthma in the other. In addition, one
may wonder what is the relevance of discrete cut-off points in terms of percentages of cells
in sputum. It is unlikely that patients with 53% neutrophils in sputum show large differences
with patients having 55% neutrophils in sputum, as percentages do not say much about
absolute numbers of cells. As for patients with EA, the percentage of sputum eosinophils can
differ greatly. For example, in chapter 10, eosinophil percentages ranged from 3.4% to 81%.
In this study, we observed a correlation between the percentage of eosinophils in sputum and
the results of the asthma control questionnaire (ACQ) and absolute FEV1 (figure 6). Therefore,
we feel that classification of patients based on granulocyte percentages in sputum should be
done with caution. In addition, when used in the clinic, not only the asthma phenotype should
be taken into account, but also the percentage on which the phenotype classification has
been based. Furthermore, if successful treatment reduces the percentage of sputum
eosinophils (80), eosinophilic asthma patients on treatment may suddenly be classified as
non-eosinophilic, which might explain the reported variability in sputum eosinophil
percentages over time (78). However, even though it might be ideal for research purposes to
perform sputum inductions in treatment-naïve individuals, it is undesirable to have
moderate-to severe patients stop therapy to classify their asthma phenotype. Besides, it is
likely that patients with persistent sputum eosinophilia despite treatment belong to an
immunologically and clinically different asthma subtype than patients with sputum
eosinophilia who respond to treatment.

Sputum granulocytes in healthy volunteers

In the studies presented in chapter 9 and 10, sputum samples were compared between two
groups of patients without healthy controls. In previous reports comparisons were typically
made between these patient groups and healthy controls (81). Unfortunately, in our studies
we have been unable to obtain sufficient numbers of granulocytes from healthy controls (see
figure 2, chapter 10), which may be related to the relatively young age of our healthy controls
(19-32 years of age). Indeed, the percentage of neutrophils in the sputum has been shown to
increase with age (82), and aging has been associated with increased low-grade inflammation
in a process dubbed “inflammaging” (83). Such low-grade inflammation might lead to
neutrophils homing towards the lungs. Taken together, inI absence of inflammation in a truly
healthy individual, granulocytes are probably not present in the sputum.

The role of sputum granulocytes in asthma

As stated before, both neutrophils and eosinophils are implicated in causing pathology in
asthma. Both sputum neutrophilia and eosinophilia have been shown to correlate with
decreased a Forced expiratory volume in 1s (FEV1) and Asthma Control Questionnaire (ACQ)
symptom scores (84-86). Airway hyperresponsiveness correlates positively with eosinophil
numbers (86), but negatively with neutrophil numbers (87). In the patient group in chapter 9,
however, ACQ score were decreasing as the percentage of neutrophils increased (figure 6). It
should be noted, though, that the study population included only 3 patients with neutrophilic

232
 General discussion

asthma and 2 patients with a mixed phenotype. The involvement of eosinophils in asthma is
further strengthened by the effect of the IL-5 blocking antibody Mepolizumab, which reduces
eosinophil production in the bone marrow, slightly increases FEV1 and reduces the amount
of exacerbations in EA patients, compared to placebo treated patients (79, 88, 89). After
cessation of therapy, the number of exacerbations increased again to a similar level as the
placebo group (90). It should be noted that the latter trials were performed in specific patients
with sputum eosinophilia despite treatment, while previous trials including all types of
asthma failed to show clinical improvement (91). This indicated that the patient group with
persistent sputum eosinophilia despite treatment might differ from the general asthma
population (92). In addition, eosinophil numbers in the lung mucosa and sputum samples
were only reduced by 50% after 20 weeks of Mepolizumab treatment and eosinophil major
basic protein (MBP) deposition was not affected at all (93). The lack of effect on sputum
eosinophils by Mepolizumab might be explained by a long eosinophil residence time in the
lungs. Although we have not been able to estimate how long eosinophils survive in the lungs,
our labelling data presented in chapter 10 do point towards a much longer eosinophil survival
in the lungs than in blood.

FEV1 (absolute)
100 Neutro
Eo

ACQ Neutro Eo
% Cells

R² 0.268 0.508
50 p 0.016 <0.001

FEV1 Neutro Eo
R² 0.165 0.417
p 0.068 0.002
0
1.5 2.0 2.5 3.0 3.5 4.0
FEV1 (L)
Figure 6. Correlation between sputum granulocyte percentages and disease characteristics. The percentage of
sputum neutrophils (●) and eosinophils (o) were plotted against (A) ACQ scores and the absolute FEV1 (B).
Correlations were determined by linear regression. N=21 for each cell type, patient characteristics can be found 11
in chapter 9. Lines indicate best linear fits with filled areas representing the 95% CI of the linear regression.

The role of granulocytes in CF

As discussed in chapter 10, neutrophils are the most abundant cell type in the sputum of CF
patients. In addition, we show in chapter 10 that the circulatory neutrophil kinetics are
perturbed in CF patients compared to healthy controls. Neutrophils in CF patients were also
found to have a shorter PMPtt. In rabbits, similar results have been seen in experimental
streptococcal pneumonia(94). This indicates that there is constant systemic inflammation
present in CF patients. Interestingly, only low percentages of immature CD16low neutrophils
were present in the blood of these patients, suggesting that mature neutrophils of CF patients
also exit the BM earlier than in healthy controls.
In the sputum of CF patients, neutrophils survive for only a very short time (< 1 day),
as the 2H-enrichment curves for blood and sputum neutrophils are highly similar. This would

233
Chapter 11

indicate that preventing neutrophil infiltration into the lungs of CF patients may be a better
therapeutic option than targeting neutrophils that are present in the lungs.

Conclusions
In conclusion, this thesis re-evaluates the circulatory kinetics of neutrophils, eosinophils and
basophils. It shows that neutrophil kinetics are disturbed in patients with the neutrophil-
driven inflammatory disease CF, whereas EA patients with eosinophil driven inflammation
have disturbed eosinophil kinetics. The combination of in vivo 6,62H2-glucose and proteomic
analysis of the three neutrophil subsets present during acute inflammation suggests that
hypersegmented neutrophils form a separately developing neutrophil subset recruited to the
circulation when needed. In sputum, both neutrophils and eosinophils are in an activated
state. Neutrophils do not survive long in sputum, while eosinophils appear to live much longer
in sputum than in the circulation. Monocytes may indeed follow the linear differentiation path
from CM via IM to NCM, but our data suggest that IM cells leave the circulation while maturing
into NCM.

234
 General discussion

References
1. Macallan DC, Asquith B, Zhang Y, de Lara C, Ghattas H, Defoiche J, and Beverley PC.
Measurement of proliferation and disappearance of rapid turnover cell populations in
human studies using deuterium-labeled glucose. Nature protocols. 2009;4(9):1313-27.
2. Mauer AM, Athens JW, Ashenbrucker H, Cartwright GE, and Wintrobe MM.
Leukokinetic Studies. II. A Method for Labeling Granulocytes in Vitro with Radioactive
Diisopropylfluorophosphate (Dfp). The Journal of clinical investigation.
1960;39(9):1481-6.
3. Cartwright GE, Athens JW, and Wintrobe MM. The Kinetics of Granulopoiesis in
Normal Man. Blood. 1964;24(780-803.
4. Peters AM, Roddie ME, Danpure HJ, Osman S, Zacharopoulos GP, George P, Stuttle
AW, and Lavender JP. 99Tcm-HMPAO labelled leucocytes: comparison with 111In-
tropolonate labelled granulocytes. Nuclear medicine communications. 1988;9(6):449-
63.
5. Dresch C, Najean Y, and Beauchet J. In vitro 51Cr and 32P-DFP labeling of granulocytes
in man. Journal of nuclear medicine : official publication, Society of Nuclear Medicine.
1971;12(12):774-84.
6. Summers C, Singh NR, White JF, Mackenzie IM, Johnston A, Solanki C, Balan KK, Peters
AM, and Chilvers ER. Pulmonary retention of primed neutrophils: a novel protective
host response, which is impaired in the acute respiratory distress syndrome. Thorax.
2014;69(7):623-9.
7. Li KW, Turner SM, Emson CL, Hellerstein MK, and Dale DC. Deuterium and neutrophil
kinetics. Blood. 2011;117(22):6052-3; author reply 3-4.
8. Macallan DC, Fullerton CA, Neese RA, Haddock K, Park SS, and Hellerstein MK.
Measurement of cell proliferation by labeling of DNA with stable isotope-labeled
glucose: studies in vitro, in animals, and in humans. Proc Natl Acad Sci U S A.
1998;95(2):708-13.
9. Pillay J, den Braber I, Vrisekoop N, Kwast LM, de Boer RJ, Borghans JA, Tesselaar K, and 11
Koenderman L. In vivo labeling with 2H2O reveals a human neutrophil lifespan of 5.4
days. Blood. 2010;116(4):625-7.
10. Steinbach KH, Schick P, Trepel F, Raffler H, Dohrmann J, Heilgeist G, Heltzel W, Li K,
Past W, van der Woerd-de Lange JA, et al. Estimation of kinetic parameters of
neutrophilic, eosinophilic, and basophilic granulocytes in human blood. Blut.
1979;39(1):27-38.
11. Cronkite EP, Fliedner TM, Bond VP, and Rubini JR. Dynamics of hemopoietic
proliferation in man and mice studied by H3-thymidine incorporation into DNA. Annals
of the New York Academy of Sciences. 1959;77(803-20.
12. Parwaresch MR, Walle AJ, and Arndt D. The peripheral kinetics of human radiolabelled
eosinophils. Virchows Archiv B: Cell pathology. 19761):57-66.
13. Vodovotz Y, and An G. Translational Systems Biology. Boston: Academic Press;
2015:111-35.

235
Chapter 11

14. Stryckmans P, Cronkite EP, Fache J, Fliedner TM, and Ramos J. Deoxyribonucleic acid
synthesis time of erythropoietic and granulopoietic cells in human beings. Nature.
1966;211(5050):717-20.
15. Whitelaw DM. Observations on human monocyte kinetics after pulse labeling. Cell and
tissue kinetics. 1972;5(4):311-7.
16. Ali AK, Nandagopal N, and Lee SH. IL-15-PI3K-AKT-mTOR: A Critical Pathway in the Life
Journey of Natural Killer Cells. Frontiers in immunology. 2015;6(355.
17. Zhang Y, Wallace DL, de Lara CM, Ghattas H, Asquith B, Worth A, Griffin GE, Taylor GP,
Tough DF, Beverley PC, et al. In vivo kinetics of human natural killer cells: the effects
of ageing and acute and chronic viral infection. Immunology. 2007;121(2):258-65.
18. Mohri H, Perelson AS, Tung K, Ribeiro RM, Ramratnam B, Markowitz M, Kost R, Hurley
A, Weinberger L, Cesar D, et al. Increased turnover of T lymphocytes in HIV-1 infection
and its reduction by antiretroviral therapy. J Exp Med. 2001;194(9):1277-87.
19. Mary JY. Normal human granulopoiesis revisited. II. Bone marrow data. Biomedicine
& pharmacotherapy = Biomedecine & pharmacotherapie. 1985;39(2):66-77.
20. Dresch C, Troccoli G, and Mary JY. Growth fraction of myelocytes in normal human
granulopoiesis. Cell and tissue kinetics. 1986;19(1):11-22.
21. Varol C, Landsman L, Fogg DK, Greenshtein L, Gildor B, Margalit R, Kalchenko V,
Geissmann F, and Jung S. Monocytes give rise to mucosal, but not splenic,
conventional dendritic cells. J Exp Med. 2007;204(1):171-80.
22. McGovern N, Schlitzer A, Gunawan M, Jardine L, Shin A, Poyner E, Green K, Dickinson
R, Wang XN, Low D, et al. Human dermal CD14(+) cells are a transient population of
monocyte-derived macrophages. Immunity. 2014;41(3):465-77.
23. Zawada AM, Rogacev KS, Rotter B, Winter P, Marell RR, Fliser D, and Heine GH.
SuperSAGE evidence for CD14++CD16+ monocytes as a third monocyte subset. Blood.
2011;118(12):e50-61.
24. Hijdra D, Vorselaars AD, Grutters JC, Claessen AM, and Rijkers GT. Phenotypic
characterization of human intermediate monocytes. Frontiers in immunology.
2013;4(339.
25. Perie L, Hodgkin PD, Naik SH, Schumacher TN, de Boer RJ, and Duffy KR. Determining
lineage pathways from cellular barcoding experiments. Cell reports. 2014;6(4):617-24.
26. Evrony GD, Lee E, Mehta BK, Benjamini Y, Johnson RM, Cai X, Yang L, Haseley P,
Lehmann HS, Park PJ, et al. Cell lineage analysis in human brain using endogenous
retroelements. Neuron. 2015;85(1):49-59.
27. Pillay J, Kamp VM, van Hoffen E, Visser T, Tak T, Lammers JW, Ulfman LH, Leenen LP,
Pickkers P, and Koenderman L. A subset of neutrophils in human systemic
inflammation inhibits T cell responses through Mac-1. The Journal of clinical
investigation. 2012;122(1):327-36.
28. Visser T, Pillay J, Pickkers P, Leenen LP, and Koenderman L. Homology in systemic
neutrophil response induced by human experimental endotoxemia and by trauma.
Shock. 2012;37(2):145-51.

236
 General discussion

29. Hao S, Andersen M, and Yu H. Detection of immune suppressive neutrophils in

peripheral blood samples of cancer patients. Am J Blood Res. 2013;3(3):239-45.
30. Lasselin J, Rehman JU, Akerstedt T, Lekander M, and Axelsson J. Effect of long-term
sleep restriction and subsequent recovery sleep on the diurnal rhythms of white blood
cell subpopulations. Brain Behav Immun. 2015;47(93-9.
31. Visser T, Beijnen v, T.A., Kamp VM, Ulfman LH, Heeres M, Westinga K, Leenen LP, and
Koenderman L. In: University U ed. Monitoring & modulation of the neutrophil
response. Utrecht; 2012:50-64.
32. Bilgin YM, and de Greef G. Plerixafor for stem cell mobilization: the current status.
Curr Opin Hematol. 2015.
33. Martin C, Burdon PC, Bridger G, Gutierrez-Ramos JC, Williams TJ, and Rankin SM.
Chemokines acting via CXCR2 and CXCR4 control the release of neutrophils from the
bone marrow and their return following senescence. Immunity. 2003;19(4):583-93.
34. Yabuki M, Miyake T, Doi Y, Fujiwara T, Hamazaki K, Yoshioka T, Horton AA, and Utsumi
K. Role of nuclear lamins in nuclear segmentation of human neutrophils. Physiol Chem
Phys Med NMR. 1999;31(2):77-84.
35. Hoffmann K, Dreger CK, Olins AL, Olins DE, Shultz LD, Lucke B, Karl H, Kaps R, Muller
D, Vaya A, et al. Mutations in the gene encoding the lamin B receptor produce an
altered nuclear morphology in granulocytes (Pelger-Huet anomaly). Nat Genet.
2002;31(4):410-4.
36. Speeckaert MM, Verhelst C, Koch A, Speeckaert R, and Lacquet F. Pelger-Huet
anomaly: a critical review of the literature. Acta haematologica. 2009;121(4):202-6.
37. Hoffmann K, Sperling K, Olins AL, and Olins DE. The granulocyte nucleus and lamin B
receptor: avoiding the ovoid. Chromosoma. 2007;116(3):227-35.
38. Siegert E, Beier L, and Grabner H. [Homozygotic Pelger-Huet anomaly]. Kinderarztl
Prax. 1983;51(4):164-9.
39. Repo H, Vuopio P, Leirisalo M, Jansson SE, and Kosunen TU. Impaired neutrophil
chemotaxis in Pelger-Huet anomaly. Clinical and experimental immunology. 11
1979;36(2):326-33.
40. Wolf K, Te Lindert M, Krause M, Alexander S, Te Riet J, Willis AL, Hoffman RM, Figdor
CG, Weiss SJ, and Friedl P. Physical limits of cell migration: control by ECM space and
nuclear deformation and tuning by proteolysis and traction force. The Journal of cell
biology. 2013;201(7):1069-84.
41. Park BH, Dolen J, and Snyder B. Defective chemotactic migration of
polymorphonuclear leukocytes in Pelger-Huet anomaly. Proc Soc Exp Biol Med.
1977;155(1):51-4.
42. Johnson CA, Bass DA, Trillo AA, Snyder MS, and DeChatelet LR. Functional and
metabolic studies of polymorphonuclear leukocytes in the congenital Pelger-Huet
anomaly. Blood. 1980;55(3):466-9.

237
Chapter 11

43. Matsumoto T, Harada Y, Yamaguchi K, Matsuzaki H, Sanada I, Yoshimura T, Honda M,

and Tanaka R. Cytogenetic and functional studies of leukocytes with Pelger-Huet
anomaly. Acta haematologica. 1984;72(4):264-73.
44. Kamp VM, Pillay J, Lammers JW, Pickkers P, Ulfman LH, and Koenderman L. Human
suppressive neutrophils CD16bright/CD62Ldim exhibit decreased adhesion. Journal of
leukocyte biology. 2012;92(5):1011-20.
45. Cremer T, and Cremer M. Chromosome territories. Cold Spring Harb Perspect Biol.
2010;2(3):a003889.
46. Cremer T, and Cremer C. Chromosome territories, nuclear architecture and gene
regulation in mammalian cells. Nat Rev Genet. 2001;2(4):292-301.
47. Aquiles Sanchez J, Karni RJ, and Wangh LJ. Fluorescent in situ hybridization (FISH)
analysis of the relationship between chromosome location and nuclear morphology in
human neutrophils. Chromosoma. 1997;106(3):168-77.
48. Yerle-Bouissou M, Mompart F, Iannuccelli E, Robelin D, Jauneau A, Lahbib-Mansais Y,
Delcros C, Oswald IP, and Gellin J. Nuclear architecture of resting and LPS-stimulated
porcine neutrophils by 3D FISH. Chromosome Res. 2009;17(7):847-62.
49. Mompart F, Robelin D, Delcros C, and Yerle-Bouissou M. 3D organization of telomeres
in porcine neutrophils and analysis of LPS-activation effect. BMC Cell Biol. 2013;14(30.
50. de Kleijn S, Kox M, Sama IE, Pillay J, van Diepen A, Huijnen MA, van der Hoeven JG,
Ferwerda G, Hermans PW, and Pickkers P. Transcriptome kinetics of circulating
neutrophils during human experimental endotoxemia. PloS one. 2012;7(6):e38255.
51. McKenzie SB, and Williams JL. Clinical Laboratory Hematology. Pearson; 2010.
52. Sipahi T, Tavil B, and Unver Y. Neutrophil hypersegmentation in children with iron
deficiency anemia. Pediatr Hematol Oncol. 2002;19(4):235-8.
53. Pillay J, Tak T, Kamp VM, and Koenderman L. Immune suppression by neutrophils and
granulocytic myeloid-derived suppressor cells: similarities and differences. Cellular
and molecular life sciences : CMLS. 2013.
54. Nauseef WM, and Borregaard N. Neutrophils at work. Nature immunology.
2014;15(7):602-11.
55. Murphy P. The neutrophil. Springer Science & Business Media; 1976.
56. Fliedner TM, Cronkite EP, Killmann SA, and Bond VP. Granulocytopoiesis. Ii.
Emergence and Pattern of Labeling of Neutrophilic Granulocytes in Humans. Blood.
1964;24(683-700.
57. Venturi GM, Tu L, Kadono T, Khan AI, Fujimoto Y, Oshel P, Bock CB, Miller AS, Albrecht
RM, Kubes P, et al. Leukocyte migration is regulated by L-selectin endoproteolytic
release. Immunity. 2003;19(5):713-24.
58. Rorvig S, Ostergaard O, Heegaard NH, and Borregaard N. Proteome profiling of human
neutrophil granule subsets, secretory vesicles, and cell membrane: correlation with
transcriptome profiling of neutrophil precursors. Journal of leukocyte biology.
2013;94(4):711-21.

238
 General discussion

59. Saverymuttu SH, Peters AM, Keshavarzian A, Reavy HJ, and Lavender JP. The kinetics
of 111indium distribution following injection of 111indium labelled autologous
granulocytes in man. British journal of haematology. 1985;61(4):675-85.
60. Bathe M, Shirai A, Doerschuk CM, and Kamm RD. Neutrophil transit times through
pulmonary capillaries: the effects of capillary geometry and fMLP-stimulation. Biophys
J. 2002;83(4):1917-33.
61. Puga I, Cols M, Barra CM, He B, Cassis L, Gentile M, Comerma L, Chorny A, Shan M, Xu
W, et al. B cell-helper neutrophils stimulate the diversification and production of
immunoglobulin in the marginal zone of the spleen. Nature immunology.
2012;13(2):170-80.
62. Nagelkerke SQ, aan de Kerk DJ, Jansen MH, van den Berg TK, and Kuijpers TW. Failure
to detect functional neutrophil B helper cells in the human spleen. PloS one.
2014;9(2):e88377.
63. Hong C, Kidani Y, N AG, Phung T, Ito A, Rong X, Ericson K, Mikkola H, Beaven SW, Miller
LS, et al. Coordinate regulation of neutrophil homeostasis by liver X receptors in mice.
The Journal of clinical investigation. 2012;122(1):337-47.
64. Knolle PA, Gerken G, Loser E, Dienes HP, Gantner F, Tiegs G, Meyer zum Buschenfelde
KH, and Lohse AW. Role of sinusoidal endothelial cells of the liver in concanavalin A-
induced hepatic injury in mice. Hepatology. 1996;24(4):824-9.
65. Rankin SM. The bone marrow: a site of neutrophil clearance. Journal of leukocyte
biology. 2010;88(2):241-51.
66. Suratt BT, Petty JM, Young SK, Malcolm KC, Lieber JG, Nick JA, Gonzalo JA, Henson PM,
and Worthen GS. Role of the CXCR4/SDF-1 chemokine axis in circulating neutrophil
homeostasis. Blood. 2004;104(2):565-71.
67. Athens JW, Raab SO, Haab OP, Mauer AM, Ashenbrucker H, Cartwright GE, and
Wintrobe MM. Leukokinetic studies. III. The distribution of granulocytes in the blood
of normal subjects. The Journal of clinical investigation. 1961;40(159-64.
68. Athens JW, Haab OP, Raab SO, Mauer AM, Ashenbrucker H, Cartwright GE, and 11
Wintrobe MM. Leukokinetic studies. IV. The total blood, circulating and marginal
granulocyte pools and the granulocyte turnover rate in normal subjects. The Journal
of clinical investigation. 1961;40(989-95.
69. Buckley CD, Ross EA, McGettrick HM, Osborne CE, Haworth O, Schmutz C, Stone PC,
Salmon M, Matharu NM, Vohra RK, et al. Identification of a phenotypically and
functionally distinct population of long-lived neutrophils in a model of reverse
endothelial migration. Journal of leukocyte biology. 2006;79(2):303-11.
70. Woodfin A, Voisin MB, Beyrau M, Colom B, Caille D, Diapouli FM, Nash GB, Chavakis
T, Albelda SM, Rainger GE, et al. The junctional adhesion molecule JAM-C regulates
polarized transendothelial migration of neutrophils in vivo. Nature immunology.
2011;12(8):761-9.
71. Yoo SK, and Huttenlocher A. Spatiotemporal photolabeling of neutrophil trafficking
during inflammation in live zebrafish. Journal of leukocyte biology. 2011;89(5):661-7.

239
Chapter 11

72. Ellett F, Elks PM, Robertson AL, Ogryzko NV, and Renshaw SA. Defining the phenotype
of neutrophils following reverse migration in zebrafish. Journal of leukocyte biology.
2015.
73. Alexis NE, Hu SC, Zeman K, Alter T, and Bennett WD. Induced sputum derives from the
central airways: confirmation using a radiolabeled aerosol bolus delivery technique.
Am J Respir Crit Care Med. 2001;164(10 Pt 1):1964-70.
74. Djukanovic R, Roche WR, Wilson JW, Beasley CR, Twentyman OP, Howarth RH, and
Holgate ST. Mucosal inflammation in asthma. Am Rev Respir Dis. 1990;142(2):434-57.
75. Bruijnzeel PL, Uddin M, and Koenderman L. Targeting neutrophilic inflammation in
severe neutrophilic asthma: can we target the disease-relevant neutrophil
phenotype? Journal of leukocyte biology. 2015;98(4):549-56.
76. Schleich FN, Manise M, Sele J, Henket M, Seidel L, and Louis R. Distribution of sputum
cellular phenotype in a large asthma cohort: predicting factors for eosinophilic vs
neutrophilic inflammation. BMC Pulm Med. 2013;13(11.
77. Fleming L, Tsartsali L, Wilson N, Regamey N, and Bush A. Sputum inflammatory
phenotypes are not stable in children with asthma. Thorax. 2012;67(8):675-81.
78. Kupczyk M, Dahlen B, Sterk PJ, Nizankowska-Mogilnicka E, Papi A, Bel EH, Chanez P,
Howarth PH, Holgate ST, Brusselle G, et al. Stability of phenotypes defined by
physiological variables and biomarkers in adults with asthma. Allergy.
2014;69(9):1198-204.
79. Haldar P, Brightling CE, Hargadon B, Gupta S, Monteiro W, Sousa A, Marshall RP,
Bradding P, Green RH, Wardlaw AJ, et al. Mepolizumab and exacerbations of
refractory eosinophilic asthma. The New England journal of medicine.
2009;360(10):973-84.
80. Jayaram L, Pizzichini MM, Cook RJ, Boulet LP, Lemiere C, Pizzichini E, Cartier A, Hussack
P, Goldsmith CH, Laviolette M, et al. Determining asthma treatment by monitoring
sputum cell counts: effect on exacerbations. The European respiratory journal : official
journal of the European Society for Clinical Respiratory Physiology. 2006;27(3):483-94.
81. Belda J, Leigh R, Parameswaran K, O'Byrne PM, Sears MR, and Hargreave FE. Induced
sputum cell counts in healthy adults. Am J Respir Crit Care Med. 2000;161(2 Pt 1):475-
8.
82. Thomas RA, Green RH, Brightling CE, Birring SS, Parker D, Wardlaw AJ, and Pavord ID.
The influence of age on induced sputum differential cell counts in normal subjects.
Chest. 2004;126(6):1811-4.
83. Franceschi C, Capri M, Monti D, Giunta S, Olivieri F, Sevini F, Panourgia MP, Invidia L,
Celani L, Scurti M, et al. Inflammaging and anti-inflammaging: a systemic perspective
on aging and longevity emerged from studies in humans. Mech Ageing Dev.
2007;128(1):92-105.
84. Woodruff PG, Khashayar R, Lazarus SC, Janson S, Avila P, Boushey HA, Segal M, and
Fahy JV. Relationship between airway inflammation, hyperresponsiveness, and

240
 General discussion

obstruction in asthma. The Journal of allergy and clinical immunology.

2001;108(5):753-8.
85. Wagener AH, de Nijs SB, Lutter R, Sousa AR, Weersink EJ, Bel EH, and Sterk PJ. External
validation of blood eosinophils, FENO and serum periostin as surrogates for sputum
eosinophils in asthma. Thorax. 2014.
86. Louis R, Lau LC, Bron AO, Roldaan AC, Radermecker M, and Djukanovic R. The
relationship between airways inflammation and asthma severity. Am J Respir Crit Care
Med. 2000;161(1):9-16.
87. Shaw DE, Berry MA, Hargadon B, McKenna S, Shelley MJ, Green RH, Brightling CE,
Wardlaw AJ, and Pavord ID. Association between neutrophilic airway inflammation
and airflow limitation in adults with asthma. Chest. 2007;132(6):1871-5.
88. Nair P, Pizzichini MM, Kjarsgaard M, Inman MD, Efthimiadis A, Pizzichini E, Hargreave
FE, and O'Byrne PM. Mepolizumab for prednisone-dependent asthma with sputum
eosinophilia. The New England journal of medicine. 2009;360(10):985-93.
89. Ortega HG, Liu MC, Pavord ID, Brusselle GG, FitzGerald JM, Chetta A, Humbert M, Katz
LE, Keene ON, Yancey SW, et al. Mepolizumab treatment in patients with severe
eosinophilic asthma. The New England journal of medicine. 2014;371(13):1198-207.
90. Haldar P, Brightling CE, Singapuri A, Hargadon B, Gupta S, Monteiro W, Bradding P,
Green RH, Wardlaw AJ, Ortega H, et al. Outcomes after cessation of mepolizumab
therapy in severe eosinophilic asthma: a 12-month follow-up analysis. The Journal of
allergy and clinical immunology. 2014;133(3):921-3.
91. Flood-Page P, Swenson C, Faiferman I, Matthews J, Williams M, Brannick L, Robinson
D, Wenzel S, Busse W, Hansel TT, et al. A study to evaluate safety and efficacy of
mepolizumab in patients with moderate persistent asthma. Am J Respir Crit Care Med.
2007;176(11):1062-71.
92. Wenzel SE. Eosinophils in asthma--closing the loop or opening the door? The New
England journal of medicine. 2009;360(10):1026-8.
93. Flood-Page PT, Menzies-Gow AN, Kay AB, and Robinson DS. Eosinophil's role remains 11
uncertain as anti-interleukin-5 only partially depletes numbers in asthmatic airway.
Am J Respir Crit Care Med. 2003;167(2):199-204.
94. Terashima T, Wiggs B, English D, Hogg JC, and van Eeden SF. Polymorphonuclear
leukocyte transit times in bone marrow during streptococcal pneumonia. The
American journal of physiology. 1996;271(4 Pt 1):L587-92.

241
 Addendum

Nederlandse samenvatting I
Dankwoord II
Curriculum Vitae III
List of publications IV
Addendum

244
 Addendum

Appendix I. Nederlandse samenvatting

De wereld om ons heen zit vol met bacteriën en andere micro-organismen. Het
immuunsysteem beschermt ons lichaam tegen schadelijke micro-organismen. Witte
bloedcellen vormen een belangrijk onderdeel van het immuunsysteem. Er zijn tientallen
verschillende witte bloedcellen met verschillende functies. Sommige witte bloedcellen maken
bacteriën direct dood door ze op te eten, andere witte bloedcellen maken antistoffen aan die
bacteriën duidelijker zichtbaar maken voor de bacterie-etende cellen en weer andere witte
bloedcellen onthouden welke bacteriën ze eerder hebben gezien, zodat je er immuun tegen
wordt.

Er bestaan tientallen verschillende witte bloedcellen, maar dit proefschrift gaat voor
het grootste deel over de witte bloedcellen die direct bacteriën en andere micro-organismen
aanvallen (zie figuur 1), zoals de monocyten en de drie soorten granulocyten: De neutrofiele
granulocyt, ook wel de neutrofiel, de eosinofiele granulocyt ofwel eosinofiel en de basofiele
granulocyt, de basofiel. Deze cellen heten granulocyten omdat ze vol zitten met korrels
(granulen). De naamgeving basofiel, eosinofiel en neutrofiel slaat dan weer op de soort
kleurstoffen die deze granulen aankleuren (basisch, zuur/eosien of pH-neutraal). De andere
witte bloedcellen die in dit proefschrift worden behandeld zijn de monocyten. De monocyten
zijn dan ook weer onder te verdelen in drie verschillende types, de klassieke, de niet-klassieke
en de intermediaire monocyten. Van de monocyt wordt verondersteld dat deze de voorloper
is van twee andere soorten cellen: de dendritische cel en de macrofaag. Deze twee celtypen
zitten over het algemeen niet in het bloed maar in de weefsels.
Figuur 1. Witte bloedcellen.
Foto’s van de verschillende
witte bloedcellen die aan
bod komen in dit
proefschrift. De foto’s zijn
gemaakt onder de
microscoop met een
vergroting van 10.000x. De
lijn links onderin is 10μm op
Ad
de foto (1μm is 1/1000ste
mm). Normaal gesproken
zijn witte bloedcellen
doorzichtig of wit met een
lichte groene zweem, maar
door kleurstoffen toe te
voegen kunnen de celkern
(donkerpaars), de inhoud
van de cel en de korrels van
granulocyten (verschillende
kleuren) mooi worden
zichtbaar gemaakt.

245
Addendum

Witte bloedcellen en ontstekingsziekten

Zoals eerder gezegd hebben witte bloedcellen de taak om het lichaam te verdedigen tegen
bacteriën en andere micro-organismen. Helaas kunnen witte bloedcellen zich soms verkeerd
gedragen, wat tot ziekte kan leiden. De korrels van basofiele granulocyten bevatten
bijvoorbeeld histamine. Als deze histamine vrij komt, kan een allergische reactie ontstaan.
Voor neutrofielen en eosinofielen kan het resultaat nog erger zijn: de inhoud van hun korrels
is namelijk bedoeld om bacteriën dood te maken en is zeer giftig. Als deze giftige stoffen vrij
komen in het lichaam, kunnen ze grote schade aanrichten. Daarom moet het vrij laten komen
van deze stoffen goed worden gereguleerd.

Slechte regulatie van witte bloedcellen lijkt te gebeuren in veel auto-immuun- en

ontstekingsziekten. Astma is een voorbeeld van een ontstekingsziekte: In mensen met deze
ziekte zijn granulocyten in de luchtwegen te vinden, terwijl deze bij gezonde mensen alleen
dieper in de longen te vinden zijn, in de longblaasjes. Daarnaast lijken deze granulocyten ook
nog eens heel erg geactiveerd en komen de giftige stoffen uit de korrels van granulocyten vrij
in de long. Bij taaislijmziekte (Cystic Fibrosis, CF) zijn witte bloedcellen ook betrokken. Er zitten
namelijk zeer grote aantallen witte bloedcellen in de longen van mensen met deze ziekte, nog
veel meer dan bij astma. Bij CF gaan deze cellen naar de longen omdat mensen met CF
bacteriën niet goed uit hun longen kunnen krijgen. Bij astma komen witte bloedcellen naar
de longen omdat er een ontsteking gaande is, alhoewel dat niet hoeft te komen door een
infectie met een bacterie of een virus. Bij beide ziekten kunnen de witte bloedcellen grote
schade aanrichten aan de longen. De giftige stoffen die uit hun korrels komen maken
longblaasjes kapot, waardoor de longen minder goed werken. Ook kan er littekenweefsel
ontstaan waardoor de longen minder elastisch worden. Met minder elastische longen kan je
minder goed ademhalen, zoals bijvoorbeeld vaak gebeurd bij rokers die COPD krijgen.

Astma en CF zijn chronische ziekten, deze gaan niet zo maar over en de ontsteking is
eigenlijk altijd aanwezig. Acute ontstekingen zijn veroorzaakt door bijvoorbeeld een auto-
ongeluk of een bloedvergiftiging. Een ontstekingsreactie bestaat bijvoorbeeld uit koorts en
een verhoogde hartslag. Er gebeuren echter ook allerlei dingen met de witte bloedcellen. Er
komen bijvoorbeeld twee soorten neutrofielen het bloed in, die daar normaal niet zijn.
Normale neutrofielen hebben een celkern die bestaat uit 3 delen (lobben of segmenten, zie
figuur 1). De twee nieuwe soorten neutrofielen hebben een andere vorm van celkern (voor
een voorbeeld, zie pagina 147). De staafkernige neutrofiel heeft helemaal geen aparte
segmenten, maar een hoefijzervormige kern en is extra goed in het doden van bacteriën. De
andere nieuwe neutrofiel heeft juist meer segmenten en wordt daarom een
hypergesegementeerde neutrofiel genoemd (hypersegment). Deze cel is juist helemaal niet
goed in het doodmaken van bacteriën, maar kan andere witte bloedcellen afremmen en
daarmee mogelijk ontstekingen remmen.

246
 Addendum

Hoe lang leven witte bloedcellen?

Mijn onderzoek richt zich onder andere op de vraag hoe lang witte bloedcellen leven in het
bloed en in de longen. In de jaren ’60 van de 20e eeuw is een aantal studies uitgevoerd om
deze vraag te beantwoorden. Deze studies leidden tot de conclusie dat neutrofielen maar 10
uur leven in het bloed, eosinofielen 24 uur en basofielen 4 dagen. Deze studies zijn echter
uitgevoerd bij terminaal zieke kanker patiënten, met cellen die het lichaam uit zijn verwijderd,
met gebruik making van zeer giftige stoffen om de cellen te volgen of een combinatie van
deze factoren. Al deze factoren hebben grote invloed op de levensduur van cellen in het
bloed. De giftige stof DFP, bijvoorbeeld, heeft tot gevolg dat 90% van de witte bloedcellen uit
het bloed verdwijnt. Dit heeft uiteraard een grote invloed op de levensduur van deze cellen.
Ook is bekend dat sommige ziekten effect kunnen hebben op de levensduur van witte
bloedcellen alsook het uit het lichaam halen van deze cellen. Hoofdstuk 2 van dit proefschrift
gaat hier dieper op in en leidt tot de conclusie dat we eigenlijk helemaal niet weten hoe lang
deze cellen leven in het lichaam.

Waarom wil je weten hoe lang witte bloedcellen leven?

Maar waarom willen we dan weten hoe lang een witte bloedcel leeft? Waarom is dit
belangrijk? De reden om dit uit te zoeken is om beter te begrijpen hoe witte bloedcellen
werken. Dat kan helpen bepaalde ziekteprocessen en de effecten van medicijnen beter te
begrijpen. Een voorbeeld hiervan is het medicijn Mepolizumab en de klinische onderzoeken
die zijn uitgevoerd om te testen hoe goed het werkt. Mepolizumab is een medicijn dat de
aanmaak van eosinofielen remt. Aangezien van eosinofielen wordt gedacht dat ze mede de
oorzaak zijn van astma, leek Mepolizumab een geschikt medicijn om deze ziekte te
behandelen. Na het inspuiten van Mepolizumab werd het aantal eosinofielen in het bloed van
astma patiënten inderdaad veel lager, tot ongeveer 10-20% van de originele waarde 4 weken
na behandeling. Deze lage waarden bleven behouden tot zeker het einde van de 20 weken
durende studie. Echter, de patiënten in deze studie hadden niet minder last van hun astma:
Hun longfunctie werd niet beter, hun symptomen werden niet minder en de kwaliteit van
leven (bepaald met vragenlijsten) was ook niet hoger aan het einde van de studie. Naar
aanleiding van deze studie werd in twijfel getrokken of eosinofielen wel betrokken zijn bij
Ad
astma. Want als er geen eosinofielen zijn om ziekte te veroorzaken, waarom worden de
patiënten dan niet beter? Het antwoord op deze vraag werd gevonden toen men ook in de
longen van astma patiënten ging kijken naar de aantallen eosinofielen. Deze aantallen bleken
namelijk maar te zijn gehalveerd na 12 weken. Deze resultaten wezen er op dat eosinofielen
veel langer leven in de longen dan in het bloed van astma patiënten of dat in aanwezigheid
van Mepolizumab de weinige eosinofielen die worden aangemaakt voor het grootste deel
naar de longen toe gaan. Latere studies gaven nog een aanwijzing dat eosinofielen misschien
langer leven in de longen: als Mepolizumab namelijk voor langere tijd wordt gegeven, bleek
het aantal ernstige astma aanvallen wel flink te verminderen. Daarom wordt tegenwoordig,

247
Addendum

zo’n 15 jaar na de eerste studies Mepolizumab voorgeschreven voor patiënten met ernstig
astma.

Als men van tevoren had geweten dat eosinofielen misschien veel langer zouden leven
in de longen dan in het bloed, dan zouden de resultaten van de Mepolizumab studies niet zo
verassend zijn geweest en had Mepolizumab veel eerder beschikbaar kunnen komen voor
patiënten. Wat is gebeurd in de studies met Mepolizumab geldt voor alle studies met
medicijnen die invloed hebben op witte bloedcellen: als we niet begrijpen hoe deze cellen
zich gedragen, kunnen we de resultaten van deze studies maar moeilijk interpreteren of
voorspellen. Maar wat is dan de beste manier om te bepalen hoe lang witte bloedcellen
leven? Zoals beschreven in hoofdstuk 2 zijn de methodes die gebruikt zijn in het verleden niet
ideaal. Deze waren niet ideaal omdat ze gebruik maakten van giftige (en soms ook
radioactieve) stoffen of omdat de cellen uit het bloed moesten worden gehaald om ze te
labelen. Daarom is voor dit proefschrift gebruik gemaakt van een heel nieuwe methode: in
vivo deuterium labelen.

De levensduur van cellen bepalen met deuterium

Deuterium is een vorm van waterstof, wat weer bekend is als de H in H2O. Waterstof is het
lichtste element waarvan een atoom bestaat uit één proton in de kern en één elektron er
omheen. Deuterium is een variant van waterstof (een isotoop), maar dan met een extra
neutron in zijn kern. Daarom wordt het ook wel zwaar waterstof genoemd. Deuterium komt
in de natuur voor, maar slechts 0.015% van alle waterstof in de natuur is deuterium. De
belangrijkste eigenschap voor ons onderzoek is, echter, dat deuterium zich ongeveer
hetzelfde gedraagt als normaal waterstof en bovenal dat het niet giftig is. De levensduur van
cellen kan bepaald worden met deuterium door vrijwilligers suiker te laten drinken waarvan
twee waterstofatomen zijn vervangen voor deuterium atomen. Deze suiker wordt in het
lichaam opgenomen en gebruikt om onder andere nieuwe witte bloedcellen aan te maken.
Door na inname van deze glucose witte bloedcellen uit het lichaam te halen en te kijken of er
deuterium aanwezig is, kan je er achter komen of er deuterium aanwezig was in het lichaam
toen deze cel werd aangemaakt. En als je op genoeg momenten kijkt hoe veel deuterium er
in de witte bloedcellen aanwezig is, kan je bepalen hoe lang deze cellen leven.
Figuur 2. Deuterium. Waterstof is het
lichtste element dat er is. Een normaal
waterstof atoom bestaat uit een enkele
proton, waar zich een elektron omheen
beweegt. Deuterium is een zwaardere
vorm van waterstof omdat het naast
een proton ook een neutron in zijn
celkern heeft. Verder gedragen de twee
vormen van waterstof zich bijna
helemaal hetzelfde en daarom is
deuterium ook niet gevaarlijk.

248
 Addendum

Dit proefschrift
Dit proefschrift laat zien hoe witte bloedcellen zich gedragen in het lichaam van gezonde en
zieke mensen. Het proefschrift begint met kijken naar hoe lang witte bloedcellen nou eigenlijk
leven in het bloed van gezonde mensen.

Deel 1. Bepalen van de levensloop van witte bloedcellen in gezonde mensen

In hoofdstuk 2 wordt beschreven hoe men vroeger heeft geprobeerd te bepalen hoe lang
witte bloedcellen leven in het bloed en waarom die methodes niet optimaal zijn geweest. In
de jaren ’60 en ’70 van de vorige eeuw zijn namelijk veel van dit soort studies gedaan over
neutrofielen. Deze werden echter uitgevoerd met heel erg giftige (radioactieve) stoffen of
met andere methodes die het gedrag van witte bloedcellen beïnvloedden. De conclusie van
dit hoofdstuk is dan ook dat we eigenlijk nog helemaal niet weten hoe lang neutrofielen leven
in het bloed en hetzelfde geldt voor veel andere soorten witte bloedcellen. Korter en
zakelijker en in een nieuwe alinea: In hoofdstuk 3 wordt daarom geprobeerd een schatting te
maken van hoe lang neutrofielen leven in het bloed maar dan met gebruikmaking van het niet
giftige deuterium-glucose. Hierbij liepen we echter tegen een probleem aan: deuterium
wordt ingebouwd in het beenmerg tijdens de productie van witte bloedcellen. Er blijft echter
ook nog deuterium achter in het beenmerg, waardoor het erg moeilijk is om te achterhalen
of de levensduur die wij bepaald hebben nou de levensduur is van neutrofielen in het bloed,
óf dat het eigenlijk laat zien hoe lang deuterium achterblijft in het beenmerg.

Normaal gesproken worden wiskundige modellen gebruikt om de metingen van

deuterium experimenten te interpreteren, maar door het effect van deuterium in het
beenmerg komen deze modellen op antwoorden die biologisch gezien onmogelijk zijn. Deze
wiskundige modellen gaan er van uit dat witte bloedcellen zich gedragen op een manier die
te vatten is in een wiskundige vergelijking. Er zijn echter aanwijzingen dat dit helemaal niet
zo is en daarom is in hoofdstuk 3 een ander soort model beschreven dat de levensloop van
een witte bloedcel realistischer volgt. Met dit model komen we tot de conclusie dat
neutrofielen ongeveer 3 dagen in het bloed leven. Ditzelfde model is vervolgens in hoofdstuk Ad
4 gebruikt om de levensduur van eosinofielen en basofielen te schatten op ongeveer 3 dagen
voor eosinofielen en op meer dan 8 dagen voor basofielen.

In de studie in hoofdstuk 5 wordt opnieuw deuterium gebruikt, maar nu om de

levensduur van de drie soorten monocyten te bepalen en meteen ook meer inzicht te krijgen
op wat voor manier de monocyten aan elkaar verwant zijn. De drie soorten monocyten lijken
namelijk erg veel op elkaar. Ze kunnen verschillende stadia in het leven van één monocyt zijn.
Het is ook mogelijk dat de drie soorten monocyten allemaal aparte celtypes zijn die los van
elkaar in het beenmerg ontstaan. De eerste mogelijkheid is getest in hoofdstuk 5 en komt tot
de conclusie dat het mogelijk is dat klassieke monocyten intermediaire monocyten worden
en daarna niet-klassieke monocyten. Dit kan echter alleen onder 2 voorwaarden: 1. Minder

249
Addendum

dan 5% van de klassieke monocyt wordt een intermediaire en vervolgens niet-klassieke

monocyt. Het grootste deel van de klassieke monocyten moet daarom ofwel dood gaan of
naar het weefsel gaan en niet meer terug komen. En aanname 2 is dat in de periode dat een
cel verandert van een intermediaire naar een niet-klassieke monocyt bevindt deze cel zich
buiten de bloedbaan en komt pas weer in het bloed als een klassieke monocyt. Als we
aannemen dat aan deze voorwaarden wordt voldaan, leven de drie monocyten gemiddeld
0,74, 1,1 en 3,1 dagen in het bloed voor de klassieke, intermediaire en niet-klassieke monocyt.
Deze resultaten sluiten mooi aan bij oude studies die hebben laten zien dat klassieke
monocyten naar de weefsels kunnen gaan en daar een dendritische cel of een macrofaag
kunnen worden. Daarnaast zijn er aanwijzingen dat veel niet-klassieke monocyten uit de
bloedbaan zijn geweest.

Deel 2. Witte bloedcellen in acute ontsteking

Vanaf deel 2 gaat het proefschrift niet alleen meer over witte bloedcellen in gezonde mensen,
maar ook over hoe deze cellen zich gedragen als er een ontsteking is. Dit hangt af van het
soort ontsteking, acuut of chronisch. Hoofdstuk 6 is een review artikel over de
ontstekingsremmende functie die neutrofielen kunnen aannemen in een acute ontsteking. In
de studie die is beschreven in hoofdstuk 7 wordt weer deuterium gebruikt om te achterhalen
hoe de twee extra soorten neutrofielen die we zien bij acute ontsteking zich verhouden tot
de normale neutrofiel. De staafkernige neutrofiel lijkt twee dagen eerder uit het beenmerg te
komen dan normale neutrofielen en lijkt daarom een voorloper te zijn van normale
neutrofielen die eerder het beenmerg uit komen bij acute ontsteking. Van de hyper
gesegmenteerde neutrofiel lijkt het dan logisch dat deze ouder zou zijn dan een normale
neutrofiel: hoe ouder de cel, hoe meer segmenten. De hyper gesegmenteerde neutrofiel is
echter precies even oud als de normale neutrofiel. Daardoor lijkt het alsof hyper
gesegmenteerde neutrofielen heel snel ontstaan uit normale neutrofielen als er een
ontsteking aan de gang is. Als we kijken naar de eiwitten waar de drie soorten neutrofielen
van gemaakt zijn, met een techniek genaamd “proteomics”, dan lijkt er echter iets anders aan
de hand te zijn: Van elk van de 2200 meest voorkomende eiwitten in deze cellen is de
hoeveelheid bepaald en hieruit blijken de staafkernige neutrofielen en de normale
neutrofielen meer op elkaar te lijken dan de normale en hyper gesegmenteerde neutrofielen.
Het verschil tussen eiwitten in hyper gesegmenteerde en normale neutrofielen is dus groter
dan het verschil dat je krijgt in 2 dagen groei van staafkernige neutrofiel naar normale
neutrofiel. En dat terwijl de hypersegment even oud is. Het verschil tussen normale en
staafkernige neutrofielen is zelfs zo groot dat het onwaarschijnlijk is dat hyper
gesegmenteerde neutrofielen in korte tijd ontstaan uit normale neutrofielen. Dit betekent
dat hyper gesegmenteerde neutrofielen apart ontwikkelen van de andere twee soorten
neutrofielen en opgeslagen liggen buiten de bloedbaan om alleen gebruikt te worden
wanneer ze nodig zijn. Als we er dan nog achter komen hoe deze cellen het bloed in te krijgen
zijn zonder acute ontsteking, dan zouden we in de toekomst misschien ontstekingen kunnen

250
 Addendum

remmen door alleen ontstekingsremmende neutrofielen naar het bloed of naar een plek met
een ontsteking te leiden.

De werking van deze ontstekingsremmende neutrofielen is echter alleen nog maar

aangetoond buiten het lichaam. Het is heel moeilijk om te bewijzen dat deze cellen ook in het
lichaam ontstekingen remmen. Daarom is in hoofdstuk 8 geprobeerd om aan te tonen dat dit
ook zo gebeurt in het lichaam. Omdat dit soort proeven niet mag worden gedaan in mensen,
zijn in dit hoofdstuk twee soorten muizen gebruikt: normale muizen en muizen waarvan de
neutrofielen geen ontstekingen kunnen remmen omdat ze een cruciaal eiwit (genaamd MAC-
1) niet aan kunnen maken. Onder gewone omstandigheden zijn muizen zonder MAC-1
helemaal gezond, maar zodra er een ontsteking optreedt (in dit geval een infectie met het
influenza virus) dan worden deze muizen erger ziek dan normale muizen. Geef je de muizen
zonder MAC-1 neutrofielen van muizen mét MAC-1, dan worden ze minder ziek dan muizen
die neutrofielen krijgen van andere muizen zonder MAC-1. Daarnaast is het niet zo dat muizen
MAC-1 gebruiken om het virus op te ruimen, want er is geen verschil in de hoeveelheid virus
tussen de verschillende soorten muizen. Deze resultaten tonen aan dat
ontstekingsremmende neutrofielen ook ontstekingen remmen in het lichaam en dat het
toedienen van ontstekingsremmende neutrofielen ontsteking kan remmen. Daarom is het in
de toekomst misschien mogelijk om ontstekingsremmende neutrofielen toe te dienen aan
mensen met een ontsteking.

Deel 3. Witte bloedcellen in chronische longziekten.

Het laatste deel van dit proefschrift gaat over witte bloedcellen bij chronische longziekten,
zowel in de luchtwegen als in het bloed. Het plan bestond om deuterium te gebruiken om het
gedrag van witte bloedcellen in de luchtwegen te bepalen, maar eigenlijk was er nog niet heel
veel bekend over de cellen die zich daar bevinden. Daarom is in hoofdstuk 10 eerst een
uitgebreide analyse gedaan op deze cellen. Cellen zijn uit de luchtwegen gehaald met behulp
van een sputum-inductie. Hierbij ademen vrijwilligers (astma patiënten) een zoutoplossing in
en moeten daarna op een speciale manier hoesten om kleine klompjes slijm uit de luchtwegen
te krijgen. Dit slijm bevatte grote hoeveelheden witte bloedcellen, voornamelijk neutrofielen,
Ad
eosinofielen en macrofagen. Daarnaast bleek dat deze cellen in goede staat waren: meer dan
90% van de cellen was in leven. Daarnaast bleek dat deze cellen prima geschikt waren om te
gebruiken in een deuterium studie. De cellen waren echter wel allemaal erg geactiveerd en
er waren aanwijzingen dat ze de giftige inhoud van hun korrels vrij hadden laten komen.
Wanneer we keken naar de correlatie tussen ziekteverschijnselen en de percentages witte
bloedcellen in het sputum, dan blijkt dat hoe hoger het percentage eosinofielen in het
sputum, hoe meer last de patiënt had van zijn/haar astma. Echter, bij een hoger percentage
neutrofielen in astma hadden patiënten minder last van hun astma. Daardoor lijkt het er op
dat neutrofielen niet zo schadelijk zijn in de longen van astmapatiënten als gedacht.

251
Addendum

Daarnaast is ook nog gekeken naar verschillen tussen mensen met verschillende
soorten astma. Afhankelijk van het percentage sputum eosinofielen (meer of minder dan 3%),
kunnen astma patiënten worden onderverdeeld in eosinofiel of non-eosinofiel astma. De
activatiestatus van de neutrofielen en eosinofielen verschilde echter helemaal niet tussen de
twee soorten astma, in bloed noch in de luchtwegen. Dit wijst er dan weer op dat zodra een
neutrofiel of eosinofiel naar de luchtwegen toe gaat, hij geactiveerd is.

Aangezien gebleken was dat de cellen in het sputum van astma patiënten in goede
staat waren, was het mogelijk om ze te isoleren voor een deuterium-studie. Daarom werden
9 patiënten met eosinofiel astma gevraagd om deuterium-glucose te drinken. Om een
vergelijking te kunnen maken met een ziekte die niet eosinofiel gedreven is, maar juist meer
neutrofiel gedreven lijkt, werden ook 6 CF patiënten gevraagd om deuterium-glucose in te
nemen. Uit deze studie bleek dat neutrofielen in het bloed van CF patiënten minder tijd nodig
hadden om volwassen te worden in het beenmerg. De eosinofielen van patiënten met
eosinofiel astma, daarentegen, hadden meer tijd nodig om volwassen te worden. Daarnaast
bleven ze ook veel korter in het bloed dan eosinofielen van gezonde mensen en CF patiënten.
Dit geeft dus aan dat de levensloop van eosinofielen is verstoord in een eosinofiel gedreven
ziekte (eosinofiel astma), terwijl de levensloop van neutrofielen aangetast is in een neutrofiel
gedreven ziekte (CF). Over de levensduur van cellen in het sputum is het moeilijker om
conclusies te trekken vanwege het lage aantal geslaagde metingen. Het leek er echter op dat
neutrofielen snel dood gaan in het sputum, terwijl er aanwijzingen waren dat eosinofielen
juist lang overleven in de luchtwegen. Dit zou dan in overeenkomst zijn met onze hypothese
dat eosinofielen lang in de longen leven en dat het daardoor heel lang duurt voor het medicijn
Mepolizumab werkzaam wordt. Hiervoor moeten echter meer metingen worden verricht, het
liefst ook nog bij mensen die Mepolizumab gebruiken.

Conclusie

Dit proefschrift heeft veel waardevolle inzichten geleverd in het leven van neutrofielen,
eosinofielen, basofielen en monocyten. De eerste drie celtypen lijken langer in het bloed te
leven dan eerder was aangenomen. Daarnaast lijkt het er op dat er in ieder geval één extra
soort neutrofiel bestaat, welke alleen wordt ingezet bij acute ontstekingen. Voor de drie
soorten monocyt toont het proefschrift aan dat ze een verschillend ontwikkelingsstadium
kunnen zijn van dezelfde cel, maar alleen als niet alle cellen het hele programma doorlopen.
Ten slotte laat het proefschrift zien dat de levensloop van cellen verschilt bij mensen met
chronische longziekten; niet alleen in de longen, maar ook in het bloed.

252
 Addendum

Appendix II. Dankwoord

In tegenstelling tot wat veel mensen denken, werken wetenschappers niet eenzaam en alleen
in het lab. Integendeel, wetenschap doe je vooral door met veel mensen samen te werken, te
praten en kennis te delen. En dat gaat gepaard met een hele hoop koffie drinken. Heel veel
mensen hebben bijgedragen aan de totstandkoming van dit proefschrift en ik wil hen met dit
dankwoord de lof geven die hen toekomt.

Allereerst Leo: lang, lang geleden is mijn lab-leven begonnen bij jou met mijn bachelor-scriptie
en een stage van twee weken. Deze stage is uitgelopen op een master stage en uiteindelijk
dit proefschrift. Ik heb er geen seconde spijt van gehad dat ik zo veel tijd op je lab heb
doorgebracht. Ik heb verschrikkelijk veel geleerd van onze discussies. Dan bedoel ik niet alleen
onze discussies over de neutrofiel, maar ook over wetenschap en hoe je om moet gaan met
wetenschappers. Wat ik hierna ook bereik, ik zal het voor een groot deel te danken hebben
aan de jaren onder jouw leiding. En laten we ook niet vergeten dat we ook een hoop lol
hebben gehad, net zo belangrijk!

José, waar ik wat meer quick & dirty werk, ben jij heel erg precies en nauwkeurig, dus dat vult
elkaar lekker aan. Het proefschrift en de papers zijn er in ieder geval veel en veel beter door
geworden. Dus ik ga proberen daar wat van mee te nemen in de toekomst.

Kiki, je bent heerlijk recht door zee, dat is erg verfrissend in het wetenschappelijk wereldje.
En ik ben blij dat Leo en ik onze liefde voor de neutrofiel een beetje hebben weten over te
dragen.

Alle collega’s en oud-collega’s van het longenlab: Janesh, Vera, Sacha, Bart, Lennard, Deon,
Leo/Lei, Corneli, Suus, Jeroen, Pieter, Jan, Okan, Adéle, Michel, Marjolein, Tjaakje, Erinke, Na,
Laurien en Nienke. De gezelligheid in het lab en op congressen maakt dat ik het altijd erg naar
m’n zin heb gehad.

Sortoperators Gerrit, Pien, Koos en Jeroen, ontzettend bedankt voor alle hulp. Ik tel alleen al
in dit proefschrift meer dan 250 sorts, ieder met een stuk of 6 populaties om het lekker
moeilijk te maken. En dan waren er nog alle proefsorts en sorts, omdat ik weer eens iets Ad
nieuws had bedacht. Jullie hulp heeft er voor gezorgd dat deze studies überhaupt mogelijk
waren. Jullie mooie gedicht hield me blij gedurende donkere nachtjes sorteren. En Corneli,
zonder jouw hulp in het lab had ik het ook nooit gered.

Alle gesorteerde samples moesten daarna nog eens geanalyseerd worden op de GC-MS en
die had daar niet altijd even veel zin in. Daarom wil ik ook Liset, Sigrid en Vera bedanken voor
de hulp om de GC-MS elke keer toch weer aan de praat te krijgen. Patrick, jouw werk op de
nog hippere mass-spec hebben geresulteerd in de prachtige resultaten van hoofdstuk 7.

Julia, wat heb je toch veel geholpen om wijs te geraken uit de deuterium databrij. Wiskundige
vergelijkingen zijn nog steeds een mysterie voor me, dus voor je hulp daarmee verdien je
zeker een aantal zinnen in mijn dankwoord. Daar. Zijn. Ze. Dan. Ook mijn andere collega’s uit

253
Addendum

de kamer wil ik bedanken. Mariona, Lodewijk, Esther en soms Hilde. Ondanks (of dankzij) wat
jullie mijn Takkie hebben aangedaan hadden we erg veel lol. Zorg goed voor het beestje.

En dan wil ik ook alle mensen van het LTI bedanken. Sinds ons lab is opgenomen in het LTI is
het erg gezellig. En dat is mede dankzij de borrel-commissies, natuurlijk.

Guru and Tomek, thanks for helping me with the fluffy ladies in the infection unit. Manuela,
Sharmayne en Tamara, jullie bedankt dat jullie zo goed voor onze dames hebben gezorgd.

Buiten het lab heb ik ook heel veel hulp gehad en in het bijzonder van Simone. Het papierwerk
bijhouden van onze drie klinische trials, bloed prikken, rochels opvangen, alle vrijwilligers
opvangen, ondervragen en vaak ook urenlang glucose geven, je deed het allemaal. Zonder
jou had ik deze studies nooit uit kunnen voeren en je bent echt een fantastische aanwinst
voor de afdeling. De longfunctieassistenten Mitzy, Simon en Monique en
researchverpleegkundige Suzan, bedankt voor de hulp bij de sputuminducties. Verder ook
heel veel dank voor drs. Renée Schweizer, dr. Ed van de Graaf en drs. Ferdinand Teding van
Berkhout voor het helpen om de patiënten te rekruteren voor studie.

Bart en Lennard, de wetenschap heeft artsen zoals jullie hard nodig. Want het blijkt dat een
onderzoeker als ik wel veel af weet van het immuunsysteem, maar slechts heel weinig over
de geneeskunst en wat het nou is om zo’n ziekte te hebben. Marga, Saskia, Sonja, Arief en
Mathias van de apotheek, bedankt voor het regelen dat we de deuterium-glucose konden en
mochten gebruiken en voor het bereiden van de drankjes voor de proefpersonen.

Niet alleen in Utrecht heb ik veel mensen te danken voor hun hulp met patiënten, maar ook
in Nijmegen. Het waren heftige studies, maar dankzij de hulp van iedereen bij de Medium
Care unit ging het allemaal erg soepel. Zeker Marjolein en Roger bedankt voor alle uren die
jullie met de proefpersonen doorbrachten.

Uiteraard moet ik onze proefpersonen bedanken. Klinisch onderzoek is alleen mogelijk met
jullie inzet en ik heb er erg van genoten om zo veel met jullie te kunnen praten. Een
wetenschapper in het lab kan misschien vergeten waar het onderzoek eigenlijk om draait,
maar dankzij jullie weet ik weer precies waarom ik dit werk doe: namelijk om vreselijke
ziekten zoals astma en CF de wereld uit te helpen of anders de ziekte wat draaglijker te maken.

Stella, de voorkant is echt prachtig geworden. Nog mooier dan ik had gehoopt.

Alle Utrechtse en de Oosterhoutse vrienden, bedankt voor alle gezelligheid over de jaren. Als
je echte vrienden hebt, heb je eigenlijk niks meer nodig. En echte vrienden, dat zijn jullie.

Pap, mam en Shanti, als iemand de fout in gaat dan zeggen ze vaak dat hij waarschijnlijk een
slechte jeugd heeft gehad. Maar wanneer iemand iets bereikt zeggen ze nooit dat hij wel een
goede jeugd zal hebben gehad. Dus daarom wil ik dat deze keer wel doen: Ik heb een
fantastische jeugd gehad en zonder jullie had ik nooit zo ver kunnen komen.

Lieve Lisanne, voor mij is Parijs alleen de stad van de liefde als jij er bent.

254
 Addendum

Appendix III. Curriculum Vitae

Tamar Tak was born on the 12th of January in Oosterhout, Noord Brabant, the Netherlands.
After completing his secondary education at the Sint-Oelbert Gymnasium, he started his
bachelor’s degree Biomedische Wetenschappen at the Utrecht University in 2004.
In 2008 he obtained his bachelor’s degree while he had already enrolled in the
Biomedical Sciences master’s programme Immunity & Infection in Utrecht. During this
master’s he did his 9-month internship in the department of Respiratory Medicine under
supervision of Prof. Leo Koenderman and Dr. Henk Honing on the subject of Fc-Receptor
regulation. Afterwards, he moved to Australia for a 7-month internship in the lab of Prof.
Jacob George at the Storr Liver Unit of the Westmead Millenium Institute in Sydney. Here, he
studied how the Hepatitis C virus causes fat accumulation in the liver and how the virus
depends on the fat metabolism for its survival. In 2009 Tamar obtained his Master’s degree
which was conferred cum laude.
In 2010, Tamar started his PhD in the department of Respiratory Medicine, which
became part of the Laboroatory of Translational Immunology, under supervision of promotor
prof. dr. Leo Koenderman and copromotors dr. Kiki Tesselaar and dr. José Borghans. The
results of this PhD research are reported in this thesis.
Tamar is currently working as a post-doctoral researcher in the lab of dr. Leïla Perié at
Institut Curie Institute in Paris, in the Quantitative Immuno-Haematology department of the
Physical Chemistry unit. Here, he aims to unravel the family tree of dendritic and monocytic
cells using genetic barcoding and other lineage tracing experiments.

Ad

255
Addendum

Appendix IV. List of publications

Tak T, Hilvering B, Tesselaar K, and Koenderman L. Similar activation state of neutrophils in
sputum of asthma patients irrespective of sputum eosinophilia. Clinical and experimental
immunology. 2015;182(2):204-12.

Tak T, Tesselaar K, Pillay J, Borghans JA, and Koenderman L. What's your age again?
Determination of human neutrophil half-lives revisited. Journal of leukocyte biology. 2013.

Pillay J, Tak T, Kamp VM, and Koenderman L. Immune suppression by neutrophils and
granulocytic myeloid-derived suppressor cells: similarities and differences. Cellular and
molecular life sciences : CMLS. 2013.

Pillay J, Kamp VM, van Hoffen E, Visser T, Tak T, Lammers JW, Ulfman LH, Leenen LP, Pickkers
P, and Koenderman L. A subset of neutrophils in human systemic inflammation inhibits T cell
responses through Mac-1. The Journal of clinical investigation. 2012;122(1):327-36.

Tak T, Tesselaar K Drylewicz J Borghans JAM, Koenderman L. Estimating human neutrophil

circulatory lifespan by short-term deuterated-glucose labelling in vivo. Submitted

Tak T, Drylewicz J, Conemans L, Borghans JAM, Tesselaar K, Koenderman L. Circulatory and

maturation kinetics of human monocyte subsets in vivo. Submitted

Tak T, Wijten P, Heeres M, Pickkers P, Scholten, A. Heck AJR, Vrisekoop N, Leenen L, Borghans
JAM, Tesselaar K, Koenderman, L. Parallel differentiation of normal and immune suppressive
neutrophils in man. Submitted

256