Transient increase of the labile iron pool in HepG2 cells
by intravenous iron preparations
Brigitte Sturm, Hans Goldenberg and Barbara Scheiber-Mojdehkar
Department of Medical Chemistry, University of Vienna, Austria
Intravenous iron, used for the treatment of anemia in
chronic renal failure and other diseases, represents a possible
source of free iron in tissue cells, particularly in the liver. In
this study we examined the effect of different sources of
intravenous iron (IVI) on the labile iron pool (LIP) which
represents the nonferritin-bound, redox-active iron that is
implicated in oxidative stress and cell injury. Furthermore,
we examined the role of the LIP for the synthesis of ferritin.
We used HepG2 cells as a well known model for hepatoma
cells and monitored the LIP with the metal-sensitive fluor-
escent probe, calcein-AM, the fluorescence of which is
quenched on binding to iron. We showed that steady state
LIP levels in HepG2 cells were increased transiently, up to
three-fold compared to control cells, as an adaptive response
to long-term IVI exposure. In relation to the amount of iron
in the LIP, the ferritin levels increased and the iron content of
ferritin decreased. As any fluctuation in the LIP, even when it
is only transient (e.g. after exposure to intravenous iron in
this study), may result either in impairment of synthesis of
iron containing proteins or in cell injury by pro-oxidants.
Such findings in nonreticuloendothelial cells may have
important implications in the generation of the adverse
effects of chronic iron exposure reported in dialysis patients.
Keywords: intravenous iron; labile iron pool; ferritin; liver;
protein synthesis.
Parenteral iron preparations are used widely for the
treatment of iron deficiency anemia in patients under

As the half-life of intravenous iron is several hours,
depending on the molecular properties of the individual
preparations [6,9], the tissues of the body are confronted
with this form of iron at relatively high concentrations (in
the range between 10 and 500 l
M
) depending on the dose
used and the rate of its infusion.
Further, a recent study suggesting that the life expectancy
of dialysis patients may be dependent on the dosage regimen
of intravenous iron (IVI) underscores the need of investi-
gation of the biochemical and pathobiochemical conse-
quences of its accumulation [10]. The administration of
large doses of parenteral iron may therefore be associated
with morbidity and mortality, in particular from infections.
These concerns arise, in part, from the known role of iron as
a growth factor for bacteria [11,12], its suspected inhibition
of neutrophil and endothelial function [13–18], the induc-
tion of protein oxidation [19], the ability to initiate oxidative
reactions [5] and clinical studies relating iron overload to
infectious morbidity [20–23].
The primary source of danger stems from the potential
release of iron into the plasma as Ôlabile plasma ironÕ [24],
as well as from the so-called cellular labile iron pool (LIP),
Correspondence to B. Scheiber-Mojdehkar, Department of Medical
Chemistry, Waehringerstr. 10, A-1090 Vienna, Austria.
Fax: + 43 1 4277 60881, Tel.: + 43 1 4277 60827,
E-mail:
Abbreviations: LIP, labile iron pool; IVI, intravenous iron; EPO,
erythropoietin; Tf, transferrin; Fe, ferrum; Fe-PP, ferric-pyrophos-

should not be neglected when judging the applied dosage
of intravenous iron.
Materials and methods
Materials
Calcein and its acetoxymethylester (calcein-AM) were
obtained from Molecular Probes. The iron chelator,
salicylaldehyde isonicotinoyl hydrazone (SIH), was a gen-
erous gift from P. Ponka (Lady Davis Institute for Medical
Research, Montreal, Canada) and was prepared as 5 m
M
stock solution in dimethylsulfoxide. Diethylene triamine
pentaacetate (DTPA), Fe-PP, cycloheximide and Hepes
were from Sigma.
Iron preparations (intravenous iron, IVI)
The preparations for testing were Venofer (ferric saccharate)
from Vifor (St. Gallen, Switzerland); Ferrlecit (ferric
gluconate) from Rhone-Poulenc Rorer (A. Nattermann
and Cie) and INFeD (ferric dextran) from Schein Pharma-
ceuticals. The preparations were dissolved in phosphate
buffered saline [NaCl/P
i
(m
M
):137,NaCl;2.7,KCl;1.45,
Na
2
HPO
4
;8.45,Na
2

measuring leakage of lactate dehydrogenase (LDH) into
the culture medium [28]. LDH activity was determined
spectrophotometrically with a test kit (Boehringer) by
means of Cobra Integra 700 autoanalyzer (Roche, Swit-
zerland). Enzyme activity in the medium was calculated as
percentage of the total intracellular and extracellular LDH
activity.
Toxicity of the iron preparations to HepG2 cells was
tested by a neutral red cytotoxicity assay [29]. After
preincubation of the cells with parenteral iron, cells were
washed and incubated with neutral red for 3 h. Then the
cells were washed with NaCl/P
i
and incubated with 200 lL
of 50% ethanol, 1% acetic acid (v/v) in distilled water for
20 min and absorbance at 540 nm was measured in a
fluorescence plate reader (Victor II) from Perkin Elmer.
Iron uptake into the LIP
In order to show that parenteral iron preparations
increase the cellular LIP, HepG2-cells were first incubated
with the fluorescent metal sensor, calcein-AM (0.25 l
M
)
at 37 °C in DMEM, buffered with 20 m
M
Hepes for
15 min. After calcein-loading, the cells were washed three
times and reincubated in DMEM, containing 20 m
M
Hepes and anti-calcein Igs [made by M. Hermann,

of the fast permeating chelator isonicotinoyl
salicylaldehyde hydrazone (SIH).
Inhibition of protein synthesis
Cells were preincubated with IVI (75 l
M
) and cyclohexi-
mide (15 lgÆmL
)1
) for the indicated times. The cell mono-
layer was then washed free of any surface-bound iron with
DMEM containing 50 l
M
DTPA and two more washings
3732 B. Sturm et al. (Eur. J. Biochem. 270) Ó FEBS 2003
with DMEM alone. Finally, the cellular LIP was measured
as described above.
Ferritin quantification by ELISA
Cells were incubated with 75 l
M
of IVI for the indicated
times. The cell monolayer was then washed free of any
surface-bound iron with DMEM containing 50 l
M
DTPA
and two more washings with DMEM alone. The cells were
lysed on ice in NP-40 lysis buffer containing 1% NP-40 and
1m
M
phenylmethanesulfonyl fluoride in 150 m
M

Results
Effect of IVI on the LIP
IVI taken up by HepG2 cells entered the labile iron pool
(Fig. 1). The LIP was assessed by the calcein-based method.
Cells were incubated with calcein-AM and baseline fluor-
escence was registered. Then various concentrations of IVI
were added and changes in calcein-fluorescence were
measured. Within the first 15 min of incubation with IVI,
the LIP increased (i.e. baseline fluorescence decreased)
between 8 and 25% depending on the iron source and the
concentration of iron calculated from the stoichiometric
composition. (Table 1). Exact concentrations could not be
obtained reliably because the cell-free calibration and the
assessment in the cellular system were apparently not
exactly equal. Ferric pyrophospate nominally represents
ÔfreeÕ iron and was most effective, followed by Ferrlecit,
Venofer and INFeD. This order corresponds to the known
physico-chemical stability of the iron complexes [6].
Adaptive response of the LIP to extracellular IVI
Exposure to extracellular IVI resulted in concentration
dependent quenching of the intracellular calcein fluores-
cence (Fig. 1, Table 1). This indicates that iron from
extracellular IVI was taken up into the cultured hepatocytes
and transiently incorporated into the LIP. To further
substantiate the adaptive response of the cells to the iron
challenge by the intravenous iron preparations, LIP meas-
urements at different time points after iron addition to the
culture medium were performed. Within the time frame
between 0 and 24 h of incubation with IVI, the LIP changed
in different ways depending on the source of iron (Fig. 2).

from different IVI preparations (Venofer, Ferrlecit, INFeD, Fe-PP)
were added. Control cells were incubated with cell culture medium
alone. Calcein fluorescence was determined within 15 min at 37 °C.
Quenching of fluorescence was referred to percentage of control.
Shown are the mean ± SEM from triplicates of three independent
experiments.
Preparation
IVI concentration (l
M
iron)
25 75 150
Venofer 8.3 ± 2.5 16.4 ± 2.2 17.8 ± 4.5
Ferrlecit 10.8 ± 4.0 16.9 ± 2.2 18.6 ± 0.8
INFeD 7.8 ± 2.7 10.1 ± 0.7 13.4 ± 1.3
Fe-PP 19.3 ± 0.3 21.4 ± 0.9 25.2 ± 1.5
Ó FEBS 2003 Intravenous iron and the labile iron pool (Eur. J. Biochem. 270) 3733
preparations, the maxima and the time course were
quantitatively different, i.e. the maxima were reached later
(after 4 h with Ferrlecit, and after 6 h with Venofer and
INFeD), were smaller and the decrease to the baseline was
slower, but in principle all IVI sources showed a similar
behaviour.
The transient increase in LIP after exposure to extracel-
lular IVI was not caused by cell damage as assessed by
means of lactate dehydrogenase release (LDH-release to the
medium was less than 5% of total LDH with 150 l
M
IVI)
and neutral red cytotoxicity test (neutral red incorporation
was not changed compared to untreated cells after exposure

M
iron) for up to 24 h (A) Venofer;
(B) Ferrlecit; (C) INFeD; (D) Fe-PP. Control cells were incubated with the cell culture medium alone. Then cells were loaded with calcein-AM
(0.25 l
M
), washed and incubated with DMEM, containing 20 m
M
Hepes and anti-calcein Ig. After registration of the baseline fluorescence, the
amount of intracellular metal bound to calcein (Ca-Fe) was assessed by addition of 100 l
M
of the fast permeating chelator SIH. Calcein
fluorescence was measured when the signal reached full fluorescence and remained stable (after 2 min). Shown are the mean ± SEM from
triplicates of three independent experiments.
3734 B. Sturm et al. (Eur. J. Biochem. 270) Ó FEBS 2003
cellular ferritin content increased with time and the rate of
the increase paralleled the increase in the LIP in the first few
hours of incubation, but was steeper in the time between 4
and 24 h for the iron sources with apparently slower iron
release, namely Venofer and INFeD (Fig. 4). Whereas with
Fe-PP and Ferrlecit, a cellular ferritin, content of 15 ng
ferritin per mg protein was already reached after 8 h of
incubation, it needed 24 h of incubation with Venofer and
INFeDtoreachthesameferritincontent.Thetimecourse
of ferritin increase corresponded to the decrease in LIP back
to the steady-state level: whereas with Fe-PP and Ferrlecit
the LIP was back to control level after 8 h, this took more
time with the two other iron preparations (Fig. 2A–D).
Apparently, the higher the initial increase in the LIP, the
faster ferritin synthesis is turned on, leading to quicker
disappearance of labile iron.

IVI. HepG2 cells were incubated for 0–8 h with IVI (75 l
M
iron) and
cycloheximide (15 lgÆmL
)1
). The control was incubated with cyclo-
heximide without IVI. Then cells were loaded with calcein-AM,
washed and incubated with DMEM, containing 20 m
M
Hepes and
anti-calcein Ig. After registration of the baseline fluorescence, the
amount of intracellular metal, bound to calcein (Ca-Fe), was assessed
by addition of 100 l
M
of the fast permeating chelator SIH. Calcein
fluorescence was measured when the signal reached full fluorescence
and remained stable (after 2 min). Shown are the mean ± SEM from
triplicates of three independent experiments.
Fig. 4. Synthesis of ferritin during long-time exposure to 75 l
M
iron
from IVI. HepG2cellswereexposedtoIVIbetween0and24h,
washed to remove surface bound iron, lysed, sonicated and stored at
)80 °C until used. The ferritin content of the lysate was determined by
ELISA as described in the Materials and methods section and corre-
lated to a standard curve. Shown are the mean ± SEM from dupli-
cates of three independent experiments.
Fig. 5. Molar ratio of iron and ferritin. Cells were incubated with
75 l
M

pool, formally also called Ôchelatable iron poolÕ because of
its accessibility to iron chelators [30,37]. This LIP is a
normal part of the total cellular iron, but it is kept small and
tightly regulated by the control mechanisms of cellular iron
homeostasis. When this balance gets out of control, free iron
can accumulate and cause oxidative damage, mainly by
reaction with ever-present reactive oxygen species (ROS)
like superoxide, hydrogen peroxide or organic peroxides
[38–40].
When the cellular LIP rises, the iron regulatory proteins
(IRPs) lose their ability to bind to iron responsive elements
(IRE) in several mRNAs. This, among other effects, leads to
an increase in the synthesis of ferritin, the major iron storage
protein. Iron bound to ferritin is harmless; thus ferritin is the
major defense against iron toxicity. Oxidative stress appar-
ently inactivates binding of IRP to IRE too and this initiates
cellular protection [41].
In hepatocytes, incubation with 100 l
M
low molecular
weight iron for 18 h doubled the LIP [42] and significantly
increased their ferritin content. We also show that iron from
the parenteral preparations enter the LIP in a time- and
concentration dependent manner. We chose the concentra-
tions between 25 and 75 l
M
iron because the fluorescence-
based method is limited with respect to the amount of iron
in the LIP. Higher concentrations of IVI lead to statistically
invalid and rather erratic results. Moreover, this concentra-

pattern of behavior similar to the increase of the LIP. The
more iron appeared in the LIP the faster the synthesis of
ferritin took place. But in general, at the endpoint (24 h) of
our IVI uptake experiments, the ferritin content was almost
the same in all cases.
HepG2 cells cultivated under normal tissue culture
conditions (DMEM-medium supplemented with 10% fetal
calf serum) are relatively iron poor. Accordingly, they have
a very low ferritin content. In this study, we show that the
ferritin of these cells is almost iron-saturated (4000 iron
atoms per molecule ferritin) and after uptake of iron from
the iron complexes into the LIP, the cells change their
metabolism according to the amount of incorporated iron
into the LIP. Control cells have highly iron loaded ferritins:
under these conditions iron from the preparations taken up
by the cells is not immediately scavenged by existing ferritin
and therefore can increase the labile iron pool. As the LIP is
suspected to regulate cellular iron metabolism (and possibly
also other known/or yet unknown enzymes or proteins
with/or without iron responsive elements) according to its
size, it is necessary that the size of the LIP is really sensitive
to incoming iron.
With iron-poor ferritin, this sensitivity to incoming iron
would be much weaker: it could immediately scavenge all
new iron from the LIP and almost no increase in the LIP
could result. The consequence of this scenario would be that
the size of the LIP is less dependent on nontransferrin-
bound iron uptake and therefore the cells need much more
time and higher amounts of incoming (and possible toxic)
iron to accommodate their metabolism according to the

loading. LIP levels return to the constitutive level of normal
tissue culture due to incorporation of labile iron into ferritin.
As any fluctuation in the LIP, even when it is only transient
(such as that following exposure to intravenous iron) may
result either in impairment of synthesis of iron containing
proteins or in cell injury by pro-oxidants [43], such findings
in nonreticuloendothelial cells may have important impli-
cations in the generation of the adverse effects of chronic
iron exposure reported in dialysis patients.
Acknowledgements
This work was supported by the Austrian Research Found (# FWF
P147842-PAT) and Hochschuljubilaeumsstiftung der Stadt Wien
(# H-83/2000).
References
1. Sunder-Plassmann, G. & Ho
¨
rl, W.H. (1995) Importance of iron
supply for erythropoietin therapy. Nephrol. Dial. Transplant. 10,
2070–2076.
2. Besarab, A., Kaiser, J.W. & Frinak, S. (1999) A study of par-
enteral iron regimens in hemodialysis patients. Am.J.KidneyDis.
34, 21–28.
3. MacDougall, I.C. (2001) Present and future strategies in the
treatment of renal anemia. Nephrol. Dial. Trans. 16, 50–55.
4. Kooistra, M.P., Niemantsverdreit, E.C., van Es, A., Mol-Beer-
man,N.M.,Strycenberg,A.&Marx,J.J.M.(1998)Ironabsorp-
tion in erythropoietin-treated hemodialysis patients: Effects of iron
availability, inflammation and aluminium. Nephrol. Dial. Trans.
13, 82–88.
5. Zager, R.A., Johnson, A.C., Hanson, S.Y. & Wasse, H. (2002)

14. Cantinieaux, B., Boelaert, J., Hariga, C. & Fondu, P. (1988)
Impaired neutrophil defense against Cersinis enterocolitica
in patients with iron overload who are undergoing dialysis.
J. Laboratory Clin. Med. 111, 524–528.
15. Boelaert, J.R., Cantinieaux, B., Hariga, C. & Fondu, P. (1990)
Recombinant erythropoietin reverses polymorphonuclear granu-
locyte dysfunction in iron-overloaded dialysis patients. Nephrol.
Dial. Transplant. 5, 504–417.
16. Veys, N., Vanholder, R. & Ringoir, S. (1992) Correction of defi-
cient phagocytosis during erythropoietin treatment in main-
tenance hemodialysis patients. Am. J. Kidney Dis. 19, 58–62.
17. Patruta, S.I., Edlinger, R., Sunder-Plassmann, G. & Hoerl, W.H.
(1998) Neutrophil impairment associated with iron therapy in
hemodialysis patients with functional iron deficiency. J. Am. Soc.
Nephrol. 9, 655–663.
18. Rooyakkers, T.M., Stroes, E.S.G., Kooistra, M.P., van Faassen,
E.E., Hider, R.C., Rabelink, T.J. & Marx, J.J.M. (2002) Ferric
saccharate induces oxygen radical stress and endothelial dys-
function in vivo. Eur. J. Clin. Invest. 32, 9–16.
19. Tovbin, D., Mazor, D., Vorobiov, M., Chaimovitz, C. &
Meyerstein, N. (2002) Induction of protein oxidation by intra-
venous iron in hemodialysis patients: Role of inflammation.
Am. J. Kidney Dis. 40, 1005–1012.
20. Mossey, R.T. & Sondheimer, J. (1985) Listeriosis in patients
with long-term hemodialysis and transfusional iron overload.
Am. J. Med. 7, 397–400.
21. Seifert,A.,vonHerrath,D.&Schaefer,K.(1987)Ironoverload,
but not treatment with desferrioxamine favours the development
of septemia in patients on maintenance hemodialysis. Q. J. Med.
65, 1025–1024.

lactic dehydrogenase: an analytical assessment of current assays.
Clin. Chem. 9, 391–395.
29. Babich, H., Sardana, M.K. & Borenfreund, E. (1988) Acute
cytotoxicities of polynuclear aromatic hydrocarbons determined
in vitro with the human liver tumour line, HepG2. Cell Biol.
Toxicol. 4, 295–309.
30. Breuer, W., Epsztejn, S. & Cabantchik, Z.I. (1995) Iron acquired
from transferrin by K562 cells is delivered into a cytoplasmic pool
of chelatable iron. J. Biol. Chem. 270, 24209–24215.
31. Collins, A., Ebben, J. & Ma, J. (1997) Frequent IV iron dosing is
associated with higher infectious deaths. J. Am. Soc. Nephrol. 8,
190 A.
32. MacDougall, I.C. (2000) Intravenous administration of iron in
epoetin-treated hemodialysis patients-which drugs, which regi-
men? Nephrol. Dial. Trans. 15, 1743–1745.
33. Besarab, A. (1999) Parenteral Iron Therapy: Safety and Efficacy.
Seminars Dialysis. 12, 237–242.
34. Besarab, A., Frinak, S. & Yee, J. (1999) An indistinct balance: The
safety and efficacy of parenteral iron therapy. J. Am. Soc. Nephrol.
10, 2029–2043.
35. Kuehn, L.C. & Hentze, M.W. (1992) Coordination of cellular iron
metabolism by post-transcriptional gene regulation. J. Inorg.
Biochem. 47, 183–195.
36. Klausner, R.D., Rouault, T.A. & Harford, J.B. (1993) Regulating
the fate of messenger RNA:the control of cellular iron metabo-
lism. Cell 72, 19–28.
37. Jacobs, A. (1977) Low molecular weight intracellular iron trans-
port compounds. Blood 5, 4331–4336.
38. Halliwell, B. & Gutteridge, J.M.C. (1990) Role of free radicals and
catalytic metal ions in human disease: An overview. Methods