PubMed A service of the U.S.

National Library of Medicine

and the National Institutes of Health

My NCBI
[Sign In] [Register]
PubMed
SearchDatabase name forSearch term
Go Clear

1: Ann Clin Biochem. 1983 Mar;20 Pt 2:80-5

Influence of liver isoferritin profile on quantitative measurement of

total ferritin protein in human liver, with emphasis on iron overload.

Zuyderhoudt FM, Linthorst C, Jörning GG, Ladiges NC, Brugman AM.

Ferritin was isolated from normal human liver and from iron-loaded human liver by
gel chromatography and by ultracentrifugation. From each of these ferritin batches
several isoferritin fractions were isolated by preparative isoelectric focusing. It was
our aim to have at our disposal isoferritin fractions with relatively large pI ranges but
distinctly different isoferritin profiles on analytical isoelectric focusing. These
fractions were compared for their immunoreactivity. The total protein in the isoferritin
fractions was measured by the nitrogen content. The immunoreactivity of the
isoferritin fractions was measured as the ratio of the reaction in immunoassays to the
nitrogen content of these fractions. We used radial immunodiffusion and enzyme-
linked immunoassay to measure immunoreactivity. The immunoreactivity did not
change obviously with the isoferritin composition of the isolated fractions. It is
concluded that pathological changes in the isoferritin composition that might occur in
liver ferritin during iron overload does not significantly influence the quantitative
measurement of liver ferritin protein by immunological methods.

 
 Ann Clin Biochem 1983; 20: 80-85

Influence of liver is 0 ferritin profile on quantitative

measurement of total ferritin protein in human
liver, with emphasis on iron overload
F M J ZUYDERHOUDT, C LINTHORST, G G A JÖRNING, N C J J LADlGES,
AND A M BRUGMAN
From the Laboratory of Experimenta/ Medicine, University of Amsterdam, Wilhe/mina Gasthuis,
Eerste He/mersstraat 104, 1054 EG Amsterdam, Tlre Netherlands

SUMMARY Ferritin was isolated from normal human liver and from iron-Ioaded human liver by
gel chromatography and by ultracentrifugation. From each of these ferritin batches several isorerritin
fractions were isolated by preparative isoelectric focusing. Jt was our aim to have at our disposal
isoferritin fractions with relatively large pI ranges but distinctly different isoferritin profiles on
analytical isoelectric focusing. These fractions were compared for their immunoreactivity. The
total protein in the isoferritin fractions was measured by the nitrogen content. The immunoreactivity
of the isoferritin fractions was measured as the ratio of the reaction in immunoassays to the nitrogen
content of these fractions . We used radial immllnodiffusion and enzyme-linked immunoassay to
measure immllnoreactivity. The immunoreactivity did not change obviously with the isoferritin
composition of the isolated fractions. It is concluded that pathological changes in the isoferritin
composition that might occur in liver ferritin during iron overload does not significantly influence
the quantitative measurement of liver ferritin protein by immunological methods.

It has been established that the measurement of
serum ferritin provides an important contriblltion
to the estimation of the body's iron stores. l2
However, from several stLldies it became apparent
that the amount of ferritin protein could nol always
be measllred accurately. It seems to depend on the
kind of ferritins in the circlllation and the antiserum
llsed. 23 Comparable problems arose with tissue
. ferritins. Use of antisera raised against spleen,
heart. and Iiver ferritins revealed different immuno­
reactivlty in immunoassays.1 Heart as weil as liver
ferritins are composed' of protein species with
different isoelectric focllsing profiles. 4 Obviously
variation of the isoferritin profile could influence
the amount of ferritin protein estimated with
immunoassays.
During diseases accompanied by massive deposi­
tion of iron in the Ijver the ferritin protein content
of the Iiver increases in order to incorporate the
increased amount of Jiver iron. Tn these circum­
stances the'· isoferritin spectrum seems to shift to the
more basic ferritin species. 5 It has been sllggested
that during mobilisation of iron from the iron­
loaded Iiver the original isoferritin profile is restored. 5
80
However, such shifts have not been confinned by
Wagstaff et a/. 4
We are interested in the amount of ferritin protein
in liver biopsies from patients with disorders of iron
metabolism. Because of the variability of the iso­
ferritin profiles we had to investigale if the amollnt
of liver ferritin protein measured with iml11uno­
assays is infJuenced by the kind of iso-ferritin profile
of the liver. If so, then Iiver ferritin in 'normal' liver
and in iron-loaded Iiver cannot be measured and
compared with an immunoassay llsing one kind of
antibody and one kind of ferrilin sdindard. In order
to study these phenomena we isolated Jiver ferritin
from norm al Iiver and from the liver of a patient
with iron overload callsed by numerOllS blood
transfusions. The isolated ferritin batches, HF I
(normal liver) and HF II (iron-Ioaded Iiver), were
separated into fractions with different ranges of
iso-ferritins. These fractions were compared with
[egard to their immllnoreactivity using radial
immunodiffusion (RID) and enzyme-linked immllno­
assay (ELISA). The nitrogen content was llsed as
reference value for the amount of protein in each
fraction.
lnfluence ojliver isojerritin profile on quantitative measurement ojlotaljerritin protein 81

Materials and methods polyacrylamide-gel electrophoresis was performed

ISOLATION PROCEDURE OF L1VER FERRJTIN
PROTEIN
Ferritin /rom norma//iver (HF 1)
In essence, the procedure was the same as that of
Wagstaff et af. using Sephadex G-200 and Sepharose
6B gel filtration. 4 This method, which we found to
be effective in the isolation of rat liver ferritin, was
less satisfactory for human liver ferritin. Densito­
metric scanning of flat bed SDS polyacrylamide gel
electrophoresis plates revealed impurities up to 10"/".
as described by the manufacturer (LKB, multiphore)
in gels of 7·5% acrylamide v,:ith 0 ·2% SDS afler
heating the samples (l00°C) .with 1 % SDS for 5
minules. ResuJts are expressed as a percentage of
total protein measured after densitometric scanning
of the plates.
The ferritin standard used in immunoassays was
prepared from normal liver by ultracentrifugation.o
This preparation was composed of a broad spectrum
of liver isoferritins as was revealed by analytical
isoelectric focusing. '~ .

Ferri/injrom iron-/oaded fiver (HF 11) PREPARATION OF ANTISERUM Tà.uVER

Because of the problems encountered with the gel FERRJTlN .. ' " ,
chromatographic method these ferritins were isol­ A rabbit received 1 mg of purified 'nórmall human
ated by ultracentrifugation , as described by Penders liver ferritin in 50 % Freund's complete adjtwant in
e/ af. G After the first centrifugation step at 150000 g eight intracutaneous injections. After four weeks
(middle of the tube) a glass-like top layer was the rabbit received two intramuscular booster
scraped off from the red-brown peil et. The remaining injections of 0·5 mg ferritin in 0·5 mi saline, whiCh
pellet was dissolved and centrifuged again. Densito­ was repeated after one week. The animal was
metric scanning of flat bed SDS polyacrylamide thereafter bied twice a week, and further booster
gel electrophoresis ' plates revealed virtually no
impurities.
Radial immunodiffusion, enzyme-linked immuno­
assay, and the measurement of the mean iron
content of ferritin were performed as described
previously.7 8
The nitrogen content of the fractions was
measured, after wet ashing of samples, by reaction
with ninhydrine using (NH 4hS04 as standard .
Preparative isoelectric focusing on a sucrose
gradient was performed with a type 8101 column
(LKB, procedure as advised by the manufacturer).
Fractions were collected between approximately
pH 4·9 and pH 6·0, as will be detailed later. All
fractions were extensively dialysed against phosphate
buffer, pH 7·4 (10 mmo!/!), to remove sucrose and
ampholines.
Analytical isoelectric focusing and staining for
protein was performed with the LKB multiphore
nat bed apparatus, as described by the manufacturer.
The electrophoresis started with 200 V and 2·5 W
for approximately 15 hours and was followed by
some 3 homs with 400 V and 2·5 W.

S/aining jor iron

- -­
The gel was immersed in JO% (w/v) K 4Fe(CN)R for
10 minutes. Thereafter colour development was
stimulated by immersing the gel for 5 minutes in Q-----
Hel, 2 mol/I, followed by short rinsing with HF1 HF2
deminera!ised water. The gel has to remain acid, Fig. 1 SDS po fyacryfamide gel e fec/ropltoresis 0)
otherwise it wil! become turbid . ferri/ins isola/ed from normaf fiver (HF I) and from iron­
The isoferritin fractions were visualised by foaded fiver (HF ll): (a) undissocia/ed ferri/ill; (b) im­
dcn,itometric scanning of the isoelectric focusing purilies; (c) sl/huni/s of frrri/ill muf some degrcrdntiol/
plates after staining. Sodium dodecylsulphate (SDS) produc/s of these subuni/s.
82 Zuyderhoudt, Linthorst, Jörning, Ladiges and Brugmm/

InJections we re given as necessary. The specificity ISOELECTRIC FOCUSING PROFILES OF

of the antibodies was established by immuno­ FERRITIN PREPARATIONS

clcctrophorcsis. HP I AND HF 11 AFTER STAINING FOR IRON

pH measurements on the analytical isoelectric Because of impurities in HF J, the focused pre para­
focusing plates were performed with an Jngold tions Were stained for (ferritin) iron, so that both
surface electrode at room temperature. preparations can be compared, In Fig, 2 the iso­
electric focusing profiles showed that HF 11 (iron­
Results loaded liver) was composcd slightly more of basic
isoferritins whereas HF I (nonnal liver) showed a
PURITY OF THE ISOFERRITIN PREPARATIONS more even distribution of iron-containing molecules
Ferritin from normal liver (HF I) over the complete pH range. This sm all difference
SDS polyacrylamide gel electrophoresis revealed might be due partly to different isolation procedures
impurities of approximately 10% (Fig. 1). With of HF I and HF Il.
preparative isoelectric focusing some impurities
could be eliminated because they were more acidic lSOELECTRIC FOCUSING PROFILES OF
or more basic than the liver isoferritins. The isolated ISOFERRITIN FRACTIONS ISOLATED FROM
isoferritin fractions contained protein impurities of HF I AND HF II AFTER STAINING FOR
approximately 5 %. PROTEIN
Figure 3 shows the densi tometric scans of the isolated
Ferritin from iron-loaded liver (HF II) isoferritin fractions from HF J, indicating that
With SDS polyacrylamide gel electrophoresis several fractions were distinctly different with regard
virtually no impurities were detectable in HF IJ to isoferritin composition. Figure 4 shows the
(Fig. 1) so the isolated isoferritin fractions from
HF II did not show any impurities either. Analytica!
isoelectric focusing profiles of the heat-supernatant
pH 20'C
from liver homogenate before ultracentrifugation . Normal' human liver
we re compared with those of purified ferritin after 5·60
staining of the gels for (ferritin) iron. Scarcely any
differences could be detected, indicating that no
obvious selection of iron-containing ferritins had
occurred during ultracentrifuga tion.
\ .,.,
~.
5'00
f\ '
_ ._ ._ Iron- Ioade<! liver HF-ll c \ t~
~'50
____'Normal · liver HF-I I.
I·
~
c
o I'
/, . , pH 20'C
.D
Cs
I'
I ' I 560 ~
VI
i'.
/
u
a­
c0 I
.D I I 5'/,0
Cs iI
«'"
'1 .D
./ \
I/ \ \
\ I 5·00 .\
i'
.1 . \ r .\ ••••• 111
\ ' \ .\ __ ._121
.,/' \
1,-60
. \
.. \ , _131
I' \ , . \ ----I~)
1 \.
'-
' .. \ .. .. . 15)
$ . ..... \

\ .\
"
. :-..').
1 1. . ,
...... ::
Distance Distance
Fig. 2 Analy/ical isoelec/ric focusing profile of ferri/ins Fig. 3 Analy/ical isoelec/ric focusing pro/iles of iso­
isola/ed from normalliver and from jron-loaded liver. ferri/in fractions isolated from normalliver.
Jnftllence of liver isoferrilill profile on quamilalive measurement of lolal ferrilin prolein 83

lron-Ioaded human liver

densitometric scans of isolated isoferritin fractions
from HF II; thispicture is quite complicated
pH 20·C because of extensive ' overlap of the fractions. Only
580 the three most abundant fractions of HF IJ (2 to 4)
could be reconcentrated and refocused . These
fractions show profound differences in isoelectric
focusing pattern, as is illustrated on the original
isoelectric focusing plate as weil as on the corres­
ponding densitometric scan (Fig. 5).
5·00 It is important to note th at the different isolated
fractions from HF I as weil as from HF. Ir we re
concentrated independently, so summatio~ of the
460
isoferritin profiles will not exactly reflect the profiles
of theoriginal preparations HFJ at1dHI;H (Fig. 2).

IMMUNOREACTIVITY OF ISOFER'RITJt-1
••••• 111 FRACTIONS ISOLATED FROM HF I AND HF IJ
__ ._ (21
____ (3) The amount of protein measured as nitrogen is
. ..... .. (41
cornpared with the amount of ferritin protein
_IS) measured by radial immunodiffusion (RID) and
enzyme-linked immunoassay (ELISA). .
Table 1 shows the isoferritin fractions 1 to 5 from
Oistance HF J. Duplo variation of the measurements was
less than 6 %. The relation between the amount of
Fig. 4 Analytical isoelectric focusing profiles of iso· protein nitrogen and the immunologically measured
ferritin fractions isolated from iron·loaded liver. ferritin protein was varia bie, but no systematic

pH 20·C

580

5·40

(lJ
u 500
§
..0
o'­
Vl
..0
<! HO

_. __ (21
- ­ __ (3)
....... (4)

(a) (b)

Distance
Fig. 5 (a) Analylical isoelectric focusing profiles of recoflcentrated isoferritin fractions 2 to 4 isolated from iron-loaded
liver. (b) Con'esponding isoelectric focusing gel showing the original preparation HF IJ (extreme left) followed by tlle
isolated isoferritin fractions 2 to 4 (from leftto right) .
 84 Zuyderhoudt, LinlhorSI, Jörning, Ladiges alld Brugl1lG/1

Table 1 lmmunoreaclivity of iso/aled isoferritinfraclions tissue from both 'normal' patients and patients with
from norma/ liver iron overload.
mg of protein measured by RID or ELISA From Tables 1 and 2 one can conclude that no
HFI: . x 100%
mg of protein measured as N large differences in immunoreactivity can be
Fractfon RID EL/SA expected between liver ferritin preparations wilh
isoferritin compositions such as lhose used in this
I 83 82
study. Whatever method we used to measure the
2
3
103
92
-'
83 ferritin protein content of oLir purified isol'erritin
4 101 93
5 84 92
fractions, the results did not differ significanlly
(TabIe 3).
x- 93% )(=88% Similar results were obtained by Wagstaff el a/ D
SD~ 9% SD= 6% with ferritins isolated from human leukaemie cells.
"FraClion 2 is omitled because no ELISA resull was available. In another study, however, Wagstaff el a/. 4 showed
that the immunoreactivity of isolated liver iso­
ferritins with a very restricted pI range varicd
Table 2 lmmunoreaclivily of üolaled isoferrilin fraclions widely. These results seem to be contradictory to
from iron-Ioaded /iver our findings but difTerences between the experi­
mg ofprotein measured by RID or ELISA ments of Wagstaff et a/. 4 and our work might
HF II: x 100% -
mg of protein measured as N " explain tbis.
Fracllon RID EL/SA An explanation of our results might be that each
of our isoferrilin fractions could be composed of
I 96 112
2 107 102 isoferritins with both low and high immunoreactivity.
3 102 102 After all our isoferritin fractions exhibit a much
4 95 104 broader pI range than the fractions isolated by
5 103 106
Wagstaff el a/. 4 This mixed composition of our iso­
X=IOI% )(=105% ferrilin fractions might lead to a mean immuno­
SD= 5% SD= 4%
reactivity which did not vary considerably in our
isoferritin preparations.
Indications for the mixed isoferritin composition
change of immunoreactivity from fractions 1 to 5 of the fractions in our study are the multiple bands
occurred. Impurities in these fractions have prob­ in the analytical isoelectric focusing profiles
ably contributed to this variabiJity. (Figs 3 to 5) and the iron content of thesefractions.
Table 2 shows the isoferritin fractions 1 to 5 from lsoferritins with significant differences in immuno­
,HF 11. Duplo variation of the measurements was reactivity have previously showndifferences with
less than 6%. The immunoreactivity of the fractions respect to the iron content. 4 In our study, the
'., did not differ significantly.

Discussion Table 3 Ferrilin prolein in isoferrilin fracliolls as

\
measured by differenl melhods
, Gel chromatographic isolation of ferritins from the The ferritin content of each fraction is expressed as the
, livers Qf patients with no known disturbances of iron percentage of the total ferritin content of the sum of these
fractions.
metabolism' ,appeared to be difficult. Repeatedly
some JO % of non-ferritin protein was detectable Prepararion FraClion PrOleilJ measured
after SDS polyacrylamide gel electrophoresis. We AsN ByRID By EL/SA
did' not try to get rid of these impurities by ion­
HF Y" I 36 34 35
exchange chromatography because we feared that 3 24 25 23
this method would select specific isoferritins and 4 21 23 22
disturb the purpose of this study. Part of the 5 19 18 20
impurities could be excluded from the isoferritin Total (±2 mg) 100 100 100
fraclions during preparative isoelectric focusing,
HF IJ I 4 4 4
Figures 3 to 5 show that, in our hands, preparative 2 30 J2 2~
isoelectric focusin~ of ferritins in a sucrose gradient 3 19 19 19
4 40 38 41
resltlts in fractions with extensive overlap of the 5 7 7 7
isoelectric focusing profiles and quite large pI ranges
in each fraction. But we assume that these fractions Tot.1 (±7 mg) 100 100 100
'c over most of lhe profiles that cao be found in liver "Fraction 2 is omitled bec.use no ELISA result was av.ilable
/nflllence ol liver isolerrifin profile on qllanfitative meaSlfremenfs ol fofal lerrifin profein 85
Table 4 Percenlage of iron in ferrilin in Ihe isoferrilill References
froclions
1 Munro HN, Linder Me. Fenitin: Structure, bio­
Pr~pa,allon Frae/Jon % ofiron in
ferritin pro/ein
synthesis, and role in iron meiabolism. Physiol Rel'.
1978; 58: 317-96.
HF! 1 18 · 4 2 Worwood M. Serum ferritin. In: Jacobs A, Worwood
2 17·3 M. eds. Iron in biochemistry alld medicille IJ. London
3 16 · 4
4 18 ·3 and New York: Academie Press 1980: 203-44.
5 20 · 2 3 Hazard JT, Drysdale JW. Ferritinaemia in cancer.
Nature 1977; 265: 755-6.
HF 11 I 25 ·0
4 Wagstaff M, Worwood M, Jacobs A. Properties -of
2 23 ·0
3 24 · 4 hu man tissue isoferritins. Biochem J 1978;: 173 :
4 22 · 7 969-77.
5 22 · 8 • Powell LW, McKeering LV, Halliday JW. Allera ­
tions in tissue rerritins in iron storage disorders.
Gul 1975; 16: 909-12. - .
fractions did not show much difference in immuno­
• Penders TJ, De Rooij-Dijk HH, Leynse B. Rapid:
reactivity and, accordingly, the iron content of the isolation of ferritin by means of ultracentrirugation.
isoferritin fractions did not reveal considerable Biochim Biophys Acla 1968; 168: 588-90.
variation (Tabie 4). 7 Zuyderhoudt FMJ, Hengeveld P, Van Gooi J,
Thc difference bet ween the mean immuno­ Jörning GGA . A method for measurement of liver
reactivity of 'normal' liver ferritin HF land the iron fractions in needie biopsy specimens and some
ferritin from _iron-Ioaded liver (approx 91 % and results in acute liver disease. Clill Chim Acla 1978~
103 %, respectively; Tables 1 and 2) can be almost 86: 313-21.
8 Zuyderhoudt FMJ, Boers W, Linthorst C, el al. An
fully explained by the non-ferritin protein impurities
enzyme-linked immunoassay for ferritin in human
in HF I (Fig. 1). Immunoreactivity of ferritin
serum and rat plasma and the influence of the iron in
preparations from normal and iron-Ioaded livers serum ferritin on serum iron measurement during
has been reported to be similar. 4 acute hepatitis. Clin Chim Acta 1978; 88: 37-44.
Our study shows that in cases of iron overload the • Wagstaff M, Worwood M, Jacobs A. Biochemical
tolal amount of ferritin protein in a liver biopsy and immunological characterization of ferritin from
specimen 7 can be measured accurately with Ijver leukaemie cells. Brit J Haemalol1980; 45: 263-74.
ferritin as a standard and our rabbit anliserum to
liver fcrritin. Accepled for publicalion 8 July 1982