A'.v/) Derinaiot t')V7: h: l.tj /.?V Ciiprrigt}!

P Mnnt<sguard 1Q97
Prinfeil in Deniiunk • .-III rii;tils reseryett
Experimental Dennatologg
ISSN 0906-6705

Molecular analysis of different phases in

human wound healing
Petri JB, Konig S, Haupt B, Hatistcin LJ-F, Herrmann K. Molecular J. B. Petri, S. Konig,
analysis of different phases in human wound healing. B. Haupt, U.-F. Haustein
Exp Dermalol 1997: 6: 133-139. © Munksgaard, 1997 and K. Herrmann
Department of Dermatology. Leipzig University,
Abstract: Culttn'ed granulation libroblasls grown from punch biopsies of Germany
the same lower arm area, obtained 3, 6, 9 and 14 days after wounding,
were used as a human wound healing model in comparison to quiescent
fibroblasls. We investigated the expression of key extracellular matrix com-
ponents at the protein level by flow cytometry and mRNA steady state
levels by Northern blotting of the different fibroblasts and compared these
data lo the ability lo migrate towards a chemolaetic signal. Proeollagen
al (I), fibronectin and matrix metalloprotease-1 synthesis was strongly tip-
regulated at the mRNA steady state level on days 3 and 14. Tissue inhibi-
tor of metalloprotease-1 mRNA is only 20" n down-regulated between day
3 and 14. Chemota.xis towards conditioned medium reflects a net effect Key words: gene expression - mRNA analysis -
of several factors and is distinctly different from chemotaxis towards plate- chemotaxis - human wound healing model
lel-derived growth factor, which peaks al day 3. Compared to the protein J. Bernliard Petri, Hautklinik. Universitat Leipzig.
level, the enhaneed expression of the corresponding PDGF receptor (3 Liehigstr. 21. D-04103 Leipzig. Germany
cliain mRNA is delayed by 3 lo 6 days. PDGF receptor a shows no regula- TeL: xx49 341 9718651
tory ehanges during the observation period. This data further supports Fax: xx49 341 9718659
the idea thai functionally divergent subpopulations of fibroblasts exist e-mail: petb(» server3.medizin.uni-leipzig.de
during wound healing. Accepted tor publication 4 February 1997

As outlined in more detail in a previous paper (1),
we obtained results at the cellular protein level that
show regulation oi' key adhesion molecules such as
the (3|, a, and ay-chain of the VLA integrins on
granulation fibroblasts during the time course of
cutaneous wound repair. In particular during the
first 6 days an up-regulation has been observed.
Wound healing is regulated by communication and
cooperation between resident cells (e.g., fibro-
blasts) in the surrounding area of the wound bed
and circulating immune cells (maerophages, neu-
trophils, monocytes, and lymphocytes).
In the early phase of wound healing activated
platelets liberate cytokines such as PDGF and
other mediators (2-4), which initiate specialized
functions of fibroblasts, endothelial cells and itn-
mune cells. A provisional matri.x composed of fi-
brin, fibronectin and mainly platelets is formed.
Gebauer et al. (5) discuss a stimulus by cytokines
as primary activating signal lo express or re-
arrange different integrin subunits on the fibro-
blast surface membrane. After the initial trigger
the specific granulation fibroblast functions are
maintained by autocrine stimulalion.
Krieg et al. detected fibroblasts with clearly diver-
ging collagen synthesis in different topographic
layers of the skin (6) and located collagen "high-
producers'" to the vicinity of vessels and mono-
nuclear cells. The latter stimulate fibroblast func-
tions vta soluble mediators (TGF-p, IL-l; IL-6 (7)).
Several groups (5, 8, 9) developed a "fibroblast stem
cell" systetn from the idea that subpopulations of
fibroblasts with different metabolic functions exist.
Cultivation of fibroblasts //; vitro after out-
growth from fresh biopsies in monolayer cultures
(10) is frequently used in eell biology. Although the
//; vivo situation probably is not fully modeled, it
can be assumed that cells that actively grow out
from tissue and migrate are most likely also active
in the initial phase of wound healing. Therefore,
we choose this system for investigations on human
granulation fibroblasts.
With the present study we addressed the follow-
ing objectives:- 1. Do cultured granulation fibro-
blasts tnaintain their specific activated status [ad-
hesion molecules, proeollagen a 1(1), fibronectin
and MMP I expression] at the level of gene ex-
pression and thus can be used as a wound healing

133
Petri et al.
model? 2. Does the cliemolaetic potency change
during wound healing and is it reflected in a rel-
evant receptor?
Methods
Human wound liealiiii^ model and cell culture
Punch biopsies from the lower arm dermis were
obtained from three age-matched healthy volun-
teers (age 22-24) according to established ethical
guidelines. Granulation tissue was taken on days
3. 6, 9 and 14 also by repeated punching in the
same area. In this paper we show the results of one
donor as Northern blots, densitometric values are
given for all three donors and the chemotaxis data
are combined data.
Monolayer cultures of fibroblasts were estab-
lished by outgrowth from the punch biopsies and
maintained under standard conditions (DMEM,
lO'Mi fetal calf serum, glutamine, penicillin, strepto-
mycin; reference 5).
Chemotaxis
The experiments were carried out using a Boyden
chamber system with 8 |.un pore size polycarbonate
filters from Nucleopore (Costar, Bodenheim, Ger-
many). The filters were soaked in ethanol for 15
min, rinsed extensively in distilled water, soaked
for 10 min in 0.5'Mi acetic acid at 50°C, washed and
boiled in a solution of 5 mg/1 gelatin in water and
finally dried at 80°C.
Fibroblasts were harvested after trypsin/ethyl-
enediamine tetraacetic acid (EDTA)-treatment, ad-
justed to 2X10'' cell.s/ml and used for the assay.
The chambers were incubated at 37°C in 5"/) CO2
in humidified air for 4 h. Cells that were attached
to the upper side but did not migrate through the
filter were removed mechanically. The filters were
fixed and stained with "Diff-Quiek" (Merz and
Dade, Unterschleissheim, Germany) and counted
at 160X magnification in 10 microscope fields per
filter. Assessment of random migration was carried
out using the same concentration of attractants in
the bottom and in the top compartment of the
chambers. Fibroblast-conditioned medium was
obtained from a confluent fibroblast culture with
FCS-free medium (10 ml) after incubation for 24
h. PDGF-AB (Dianova, Hamburg, Germany) was
diluted to a concentration of 25 ng/ml in serum-
free Dulbeceo's Modified Eagle Medium (DMEM;
Bioehrom, Berlin, Germany).
Flow eytometry
The cells were detached with 0.05'^ trypsin, 0.02'Mi
EDTA in phosphate buffered saline. 4x10'^ cells/
134
ml were incubated for 45 min at 4°C with the pri-
mary monoclonal antibody directed to the p-sub-
unit of the PDGF-receptor (57 |,ig/ml; Dianova)
and for 30 min at 4°C with a FlTC-labelled sec-
ondary antibody (Dianova). The measurements
were carried out on an EPICS PROFILE II (Coul-
ter, Hialeah, FL, USA).
niRNA analysis
mRNA was isolated from 70 to 80% confluent hu-
man cell eultures, quiescent dermal fibroblasts as
control and fibroblasts from granulation tissue of
day 3, 6, 9 and 14 after wounding using the "Quick-
Prep Micro mRNA purification kit" (Pharmacia-
Biotech, Sweden). Cells gave comparable results be-
tween passages four to eight. Approximately 500 ng
ofmRNA was eleetrophoresed in an \.2"A, agarose
3 chemotactic response to PDGF-AB and conditioned medium
'•PDGF
200 • CM
150
100
1
1
50
0 Bllt
1
1
days after wounding
DF 3 6 9 14 DF 3 6 9 14
«- 5.6 kB *• 28S
•*».»*•». 18S
Fiifure L (a) Clicinolaxis of gniniilalion libroblasls towards
fibroblast conditioned medium and PDGI- A13 of human der-
mal and granulation fibroblasts. The data are shown as
averages±standard deviation of Ihree independent e.\periments
in triplicate. Untreated cells were trypsini/ed and used for the
chemotaxis assay. Fibroblast-conditioned medium or medium
containing PDGF AB. respectively, was filled into the lower
conipartnient of the Boyden chamber. CM: conditioned me-
dium, (b) Left-liand panel; Northern blot with 7 ng niRNA
from granulation libroblasts IVom different days after wounding
(3-14) probed with an antisense probe to the PDGF (3 chain
gene. DF: dermal fibroblasts. Right-hand panel: ethidium bro-
mide stained gel showing equal aniotints of 18S and 28S rRNA.

gel (containing \"/a fortnaldehyde) and transferred
to nylon membranes (Qiagen, Hilden, Germany).
Filters were baked for 2 h at 80°C and hybridized
with 100 ng/ml digoxigenin-UTP-labeled RNA ("ri-
boprobe") for 16 h. The hybrids were made visible
by an enzyme-linked immunoassay and a sub-
sequetit enzyme-catalyzed chcmilutniniscent reac-
tion, with CSPD (Tropix, Heidelbetg, Germany).
The riboprobe for GAPDH was transcribed //; vitro
from a cloned 389 bp-PCR-product into the Smal-
site of cloning vector pBluescript SK+ (Stratagene,
La Jolla, CA, USA). The probes for MMP Tand
TIMP were obtained accordingly ftotn cloned RT-
PCR products. The proeollagen al(I) (Hf677) and
the integrin Pi clones were a gift from Prof. Krieg,
Cologne.
Results
We investigated quiescent skin fibroblasts and
granulation fibroblasts obtained frotn lower artn
punch biopsies from experimentally wounded, age-
tnatched healthy donors by chemotaxis. Northern
and flow cytometer analysis.
Cfiemota.xis of granulation fihrohlasts and PDGF-
receptor mRNA
F'wsi, we tested the chetnotactic response towards
the PDGF-AB isoforni and fibroblast-conditioncd
cell cell
number number
skin fibroblasts
(16%)
a. fluorescence intensity b. fiuorescenee intensity
Fii^urc 2. (a) F'low eytometrie analysis of PDF'G receptor (i
(panel a) and PDFG reeeptor a (panel b) expressed on granu-
lation libroblasls during diiTerent stages of wotind healing com-
pared to quieseent dermal ftbroblasts. Fluorescence intensity ol
PDFG receptors is shown as solid black lines, isotype-inatched
IgG controls are given as gray lines.
Molecular analysis in hnman wound healing
mRNA steady state level
1 2 3 4 5
collagen a1(l)
MMP1
fjbronectin
TIMP-1
f« integrin (CD 29)
GAPDH
Figure .•?. Steady state mRNA Northern blots of quieseent der-
mal and graiuiiation libroblasts from days 3 to 14 hybridized
with ptobes lor piocollagen type al (I), MMP-l (matrix metal-
loprotease-1), Hbroneetin and P, ehain of VLA integrins.
GAPDH was used to standardize. I: eontrol libroblasts; 2: day
3; 3: day 6: 4: day 9; 5: day 14.
medium as chemoattractants. As demonstrated in
Fig. la, the chemotactic response to PDGF
reaches a tnaximum at day 3, followed by a de-
crease at day 6 and 9, whereas the chetnotactic po-
tency of granulation fibroblasts from day 14 re-
turns to the control level. A reverse answer of
granulation fibroblasts from day 3 to 14 towards
fibroblast-conditioned tnediutn as chetnoattractant
was obvious and probably reflects a net effect of
diffetent factors. Ttanscripts of PDGF receptors a
and P genes were investigated from the same time
points. The course o'[ PDGF-receptor (3 followed a
similar time course as the chemotaxis experiment,
but was delayed atid reached a tnaxitnutn on day 6
(Fig. lb). hi contrast, PDGF-receptor a (data not
shown) did not show tnarked ehanges throughout
the observation period, which coincides with the
flow cytotnetric data (tiext paragraph).
Flow cvtontetric analysis of PDGF a and (i
receptor
By flow cytometry, we analyzed the density of the
PDGF feceptors a and P on different granulation
135
Petri et al.
collagen a1 (I) mRNA MMPImRNA

conlrol day 3 day 6 day 9 control day 3 days day 9 day 14

ribroblasl!! granulallon nbrobl»U b ribroblasts granulation fibroblasts

flbronectin mRNA TIMP I mRNA

cantrol dBy 3 day 6 days dayH conln)l day 3 day 6 day 6 day 14
fibroblasts granulation fibrobliBtl d Hbroblasls granulation fibrobiatts

integrin |31 (CD29) mRNA

Ftgme 4. Calculated densitoiiietric values (with SD) of mRNA

level data of thiee donors, normalized to the respective
control day 3 day 8 day 9 day 14 GAPDH amounts, (a) Proeollagen al(I): (b) MMP I: (c)
ribroblasts granulation fibroblasts libronectin: (d) TIMP I; (e) Pi chain of VLA integrins.

fibroblasts. Granulation fibroblasts from day 3
after wounding show a significant increase in re-
ceptor density compared to quiescent fibroblasts
(Fig. 2a). The density was still elevated on days 6
through 14, although slightly reduced compared to
day 3. Since PDGF-AB is able to bind to both the
dimerized PDGF-receptor subtypes a a and aP, we
also measured the density of the a receptor on
granulation fibroblasts. We found no increase in
receptor density throughout the observed period
of wound healing for the receptor a (Fig. 2b).
Analysis of mRNA
We present Northern data (Fig. 3) and the respect-
ive diagrams with the calculated densitometric
values (Fig. 4a-e), which depict the course of ex-
pression. The steady state levels of proeollagen
type al(I) tnRNA are tnarkedly increased in the
fibroblast outgrowth of days 3, 6, and 14 (Fig. 4a).
Proeollagen al(I) and fibronectin (Fig. 4c) show a
very similar eourse of expression. The expression
of the PI chain of the VLA integrins, which is also
part of the fibroneetin receptor (integrin a5 p|) was
up-regulated with a maxitnum between 140 and
160% at day 6 and did not return to the basic level
during the experiment (Fig. 4e).
In agreement with the results for proeollagen
al(I), we observed a minitnal expression of nietal-
loprotease-1 (MMP-1) mRNA on day 6 and 9
(Fig. 4b). We interpret the tnarkedly increased ex-
pression of MMP-l tnRNA on day 14 (less pro-
nounced on day 3) wilh an enhanced turnover rate
of proeollagen al(I) on days 3 and 14. Particularly
around day 14 the fibroblasts not only produce
new collagen, but also degrade already synthesized
collagen to restrueture the newly fortned tissue.
Furthermore, we analyzed the regulation of
TIMP-1, the inhibitor of MMP-1, and found only a
20% reduction on the transeriptional level (Fig. 4d).
Between dilTerent outgrowth days of granulation
tissue no significant alterations were detected.

136

The tnRNA levels were normalized to the
GAPDH mRNA amount in the same tnRNA
preparation. Thus, small differences in niRNA
loading could be corrected as well.
Discussion
Wound healing is morphologically subdivided into
different stages by regulated interactions of stimu-
lating and inhibiting signals. Locally produced and
secreted cytokines play an important role in this
process (5, 12, 13, 14). During wound repair Ihe
expression of proteins involved with the extracellu-
lar matrix changes at certain stages of this process.
With our present experitnents we could show the
cellular responses on three differenl levels: chemo-
taxis, surface expression and mRNA levels.
Chemotaetic abilities vary dttrittg wottttd Itealittg
Imtnigraling granulation libroblasts arise from
resting fibroblasts in the surtounding tissue. These
activated fibroblasts undergo phenotypic changes
after starting tnigration. Directed tnigration to-
wards chemoattraclants is closely connected wilh
cyclic adhesion and de-adhesion of cell surface
integrins to components of the extracellular tnal-
rix. The expression of the participating compon-
ents is regulated by adhesion, the action of cyto-
kines and chemoattractants.
It has been shown that various substances (e.g.,
PDGF, epidermal growth factor, transfortning
growth factor p or basic fibroblast growth factor)
promote migration of fibroblasts iit vitro (14-16).
The observed different responses of granulation
fibroblasts towards conditioned mediutn refiect an
increasing net effect of different factors (e.g.,
PDGF, basic fibroblasl growth factor, fibronectin,
collagenous peplides) frotn day 3 to 14. The ob-
served opposite course of PDGF action alone on
granulation libroblasts with a tnaxitnum on day 3
supports the idea that PDGF exerts its effect early
during wound healing by ptotnoting fibroblast mi-
gration into Ihe wound area. Hayashi el al. (13)
found thai all three isofortns of PDGF are aetive
chemoattractants, although with different signal
strength.
PDGF reeeptor sitbtypes a attd />'
The protein data correlate very well with the
mRNA data. Only on day 14, the mRNA steady
state level for the PDGF reeeplor P teturned to the
original level of quiescent fibroblasts. In contrast,
the level in flow cytotnetry was reduced on day 14
compared to day 3, but not as low as the level of
quiescent libroblasts. This differenee ean be undet-
Molecular analysis in human wound healing
stood as a consequence of the shut down or reduc-
tion of receptor transcription, whereas the turn-
over of the translation product takes more time.
Therefore, the reduction is effective earlier on the
tnRNA level than on the protein level. However,
the cotnplementation of tnRNA-data by FACS
analysis confirtns the causal relattonship between
the chemotaclic activity of PDGF-AB and PDGF
receptor P tnRNA expression. In contrast to the P
receplor chain, the a receptor is not tnodulaled
during the observed days.
The data for PDGF receptor p, obtained as
mRNA level, show a slightly different titne course.
The maximal induction of expression occurs at day
6, but also returns to the basic level at day 14. So
far, we cannot decide whether (in analogy to other
lytosine kinases such as TGF-P receptors) a cyto-
solic "pool" of receptors for PDGF is translocated
to the surface in response to the stimulus or
whether sufficient receptors are on the surface of
fibroblasts that allow an immediate response to the
stitnulus. After the signal the PDGF teceptor gene
expression is first turned on and then again down-
regulated to the physiological level.
E.xpressiott oJ grctntthitiott Jibroblasts mottitored at
the mRNA level
The regulation of fibroblast metabolistn in depend-
enee of the above synergistie or antagonistic acting
cytokines has been shown. Apparently, the process
of wound healing is maintained by a complex net-
work of autocrine loops. In experimental systems
this network is diffieult to sitnulate. However, the
pattern of gene activation, tneasured as gene tran-
scripts, can be observed over a period of up to
eight cell passages. This allowed us to use tnono-
layer cultures of granulation fibroblasts and study
the effeel on the gene expression.
Ptocollagen al(I) and in parallel matrix melallo-
protease-1 (MMP-1) tnRNA was strongly upregu-
lated in the early phase, which can be explained
with the requirement for a rapid pielitninary clo-
sute of the wound area. At day 6 proeollagen al(l)
synthesis is still enhanced. The observed down-
regulation of MMP 1 at days 6 and 9 supports
the idea of a "quiet phase" of angiogenesis and
restructuring. The tnarked up-regulation of both
ptocollagen al (1) and MMP 1 atound day 14
marks a second phase, the beginning of temodel-
ling of the wound atea. Fibronectin, which gener-
ally shows a similar course of up- and down-regu-
lation as proeollagen al(l), possibly serves as a
guide and slide rail for newly formed collagen
fibers.
The data on mRNA levels for proeollagen «1(I)
and MMP 1 support the assutnption that both
137
Petri et al.
genes arc coordinalely expressed. During the initial
phase of collagen synthesis there is relatively little
degradation of the newly synthesized collagen.
Only after wound closure and the onset of remod-
elling processes, the level of MMP-1 is dramati-
cally increased. MMP-I could also be required for
degradation of ECM debris in the wound area. In
keratinocytes MMP-I induction is maximal be-
tween 12-24 h after wounding and does not require
infiammation (17). Apparently MMP-1 is dif-
ferently regulated in different dermal cell types.
Frotii a physiological standpoint one would ex-
pect that TIMP as the inhibitor of MMP-1 would
somehow be regulated at the niRNA level, particu-
larly around day 3 and 14. However, from our data
we cannot conclude whether TIMP regulates
MMP-1 during wound healing or not.
With respect to the regulated phases of wound
healing, the cootdinated up- and down-regulation
can be explained physiologically. Fibroblasts show
their maximal activity during the early wound
healing stage, the invasion of the granulation
tissue. They divide very rapidly and reconstruct the
affected tissue by maximal protein synthesis. After
the first solid wound closure the granulation
fibroblasts start to remodel the original structure.
The observations with granulation fibroblasts in
cell culture support our hypothesis (5) of subpopu-
lations of fibroblasts, which, after the initial trig-
gering, remain in their differentiated status for a
long period of time in cell culture. In normal
tissue, these subpopulations usually behave like
"normal" fibroblasts. Only after activation fibro-
blasts differentiate into granulation fibroblasts and
respond to the particular requiretnents of wound
closure and healing. The fate of the activated
fibroblasts after the event has not been elucidated
yet.
Furthermore, experimental data indicate that
fibroblasts from donors of various ages differ at
least phenotypically (8, 9). Funaro et al. (18) ob-
tained monoclonal antibodies that distinguish be-
tween fetal and adult fibroblasts. However, they
could not prove a fibroblast specificity of these
antibodies. The study of specific subpopulations of
fibroblasts will make it possible to address some of
these questions in more detail.
The results presented here not only show that
indeed a particular set of fibroblasts does remain
in an activated stage during wound healing, but
allow conclusions for disturbed wound healing as
well. The pathophysiology of the regulation of
fibroblasts can further experimentally support this
model. If this holds true, it should be possible to
mimic such an activated subset by providing genet-
ically engineered fibroblasts on wounds, where the
connective tissue is unable to produce the particu-
138
lar subset. In this way our improved understanding
of wound healing eventually could lead to im-
proved therapeutic strategies.
Acknowledgements
We thank the following sponsors: Deutsche torschungsgcmein-
schaft (DFG Hc-2146/l-l) and Bundesministerium fur For-
schung und Technologic (FKZ .'?22-4001-OIZZ9!().Vl. 12 and
322 4()()I-O1ZZ9I()3/1.21). We also thank Mrs H. Gcdicke and
E. Schnabel for skilll'ul technical, assistance.
References
1. Petri J B. Saalbach A, Haupt B, Piercr M. Haustcin U-F,
Herrmann K. /;; vitro analysis of adhesion molecule ex-
pression and gel contraction with human granulation
libroblasts. Wound Rep Reg 1997: 5: 1 S.
2. Kulozik M. Heckmann M. Maucli C, ScharlTctter K, Krieg
T Cylokinc regulation of collagen metabolism during
wound healing /;; vitro and /// vivo. In: .lansseii H, Rooman
R, Robertson J I S, eds. Wound healing. Peterslield, UK:
Wrightson Biomedical Publishing Ltd, 1991: 33-39.
3. C lark R A F. Cutaneous wound repair: a review with em-
phasis on integrin receptor expression. In: .lanssen 11. Roo-
man R, Robertson .1 1 S, eds. Wound healing. Peterslield,
t)K: Wrightson Biomedical Publishing Ltd, iV^I: 7 17.
4. Meycr-lngold W. Eichner W. Platelcl-derived growth factor.
Cell Biol Int 1995: 19: 3X9 39S.
5. Gebauer S, Saalbach A. Mittag M et al. Impression of ad-
hesion molecules during wound healing characterization
ol granulation libroblasts. .1. I'tir Anasthesie und
Intensivbehandlung Cm rent Aspects in Wound Healina J
A 1 1996: 3: 32 42.
6. Krieg T. Meurcr M. Systemic scleroderma clinical and
pathophysiological aspects. .1 Am Acad Dermatol 19XS- IX-
457 484.
7. Kulo/ik M. .Scharffetter K. Lankat-Buttgereit B. Heck-
mann M. Krieg T Co-localization ol transforming growth
faclor (32 with rel(I) proeollagen niRNA in tissue sections
of patients with systemic sclerosis. .1 Clin Invest I99()- Sfv
917 922.
H. Bayreuther K, Rodemann H P Kommel R. Dittman K,
Albiez M. Fraiicz P I. Human skin libroblasts in vitro dif-
ferentiate along a terminal cell lineage. Proc Natl Acad Sci
USA, I9SX: X5: 5112-51 16.
9. Bayreuther K. Rodemann H P Francz P I. Maier K. Dif-
ferentiation of librobiast stem cells. .1 Cell Sci Suppl 1988-
10: 115-130.
10. (Jranstein R D, Deak M R, .lacques SL ct al. The systemic
administration of interferon gamma inhibits collagen syn-
thesis and acute inflammation in a murine skin wounding
model. .1 Invest Dermatol 1989: 93: 18-27.
11. Katz A M, Rosenthal D, Sauder D N. Cell Adhesion mol-
ecules structLU'c, function, and implication in a variety of
cutaneous and other pathologic conditions. Int .1 Ocrmatol
1991: 30: 153-160.
12. Dimcan M R. Bcrman B. Van de Water .1. Boyd R L. Wick
Ci, Cjershwin M E. Mononuclear cells isolated IVom librotic
skin lesions in avian scleroderma conslilulivcly produce
libroblast-activating cytokines and immunoglobulin M. Int
Arch Allergy Immunol 1995: 107: 519 526."
13. llayashi N. Takehara K. Soma Y. Differential cheinotactic
responses mediated by platelet-derived growth lactor a-
aitd (5-receptors. Arc Bioehem Biophys 1995: 322: 423 428.
14. Ashcroft G S, Horan M A. Ferguson M W .1. The elTects
of ageing on cutaneous wound healing in mammals, .1 Anat
1995: 187: I 26.

15. Adelmann-Grill B C. Wach 1-, Cully Z, Hein R. Krieg T
Chemotactic migration of normal dermal libroblasts to-
wards epidermal growth factor and its modulation by plate-
let-derived growth factor and transforminn erowth factor-
beta. Eur .1 Cell Biol 1990: 51: 322-326.
16. Postlethwaite A E. Dermal fibroblast function. In: Nickol-
off B J. ed. Dermal immune system. Boca Raton: CRC
Press, Inc., 1993: 163-184.
17. Inoue M. Kratz G. Haegerstrand A. Stahle-Backdahl M.
Molecular analysis in human wound healing
Collagenase expression is rapidly induced in wound-edge
keratinocytes after acute injur>' in human skin, persists dur-
ing healing, and stops at re-epithelization. .1 Invest Derma-
tol 1995: 104: 479^83.
18. Funaro A. Bellone G. DeMontc L et al. Generation and
selection of monoclonal antibodies identifying surfaee mol-
ecules discriminating between fetal and adult fibroblasts. J
Clin Lab Anal 1989: 3: 50-56.
139