Nephrol Dial Transplant (2004) 19: 2592–2597
doi:10.1093/ndt/gfh278
Advance Access publication 3 August 2004
Original Article
Oxidative stress induced by iron released from transferrin
in low pH peritoneal dialysis solution
Yasuyoshi Yamaji1, Yuichi Nakazato2, Naoki Oshima1, Matsuhiko Hayashi2 and Takao Saruta2
1
Kidney Disease Center and Internal Medicine, Saitama Social Insurance Hospital, Saitama-City, Saitama, Japan and
Department of Internal Medicine, School of Medicine, Keio University, Tokyo, Japan
2
Abstract
Background. Transferrin binds extracellular iron and
protects tissues from iron-induced oxidative stress.
The binding of iron and transferrin is pH dependent
and conventional peritoneal dialysis (PD) solutions
have unphysiologically low pH values. Herein, we
investigated whether conventional PD solution releases
iron from transferrin and if the released iron causes
oxidative stress.
Methods. Effects of PD solutions on iron binding to
transferrin were examined with purified human transferrin and transferrin in dialysates drained from PD
patients. Oxidative stress induced by iron released
from transferrin was evaluated in terms of the
formation of thiobarbituric acid reactive substance
(TBARS) and protein carbonylation in the human red
blood cell (RBC) membrane. The iron deposition in
peritoneal tissue from PD patients was evaluated by
Perls’ staining with diaminobenzidine intensification.
Results. Low pH PD solution released iron from
transferrin. This iron release occurred within 1 min. Iron
release was not observed in neutralized PD solution.
Iron released from transferrin in low pH PD solution
increased TBARS formation and protein carbonylation in the human RBC membrane. Iron deposition,
which is prominent in the fibrotic area facing the
peritoneal cavity, was observed in the peritoneum of
PD patients.
Conclusions. Iron released from transferrin in low pH
PD solution can produce oxidative stress in the
peritoneum of a PD patient. Neutralizing PD solution
can avoid this problem. Iron deposition in the peritoneum may participate in the pathogenesis of peritoneal fibrosis in PD patients.
Correspondence and offprint requests to: Yasuyoshi Yamaji, MD,
Kidney Disease Center and Internal Medicine, Saitama Social
Insurance Hospital, 4-9-3 Kitaurawa, Urawa-ku, Saitama-City,
Saitama, 330-0074, Japan. Email: [email protected]
Keywords: dialysate; fibrosis; iron; oxidative stress;
peritoneal dialysis; peritoneal function; pH
Introduction
Chronic peritoneal dialysis (PD) results in peritoneal
injury, which leads to a progressive reduction in dialytic
efficiency. While high glucose concentration, contaminated glucose degradation products, and low pH in PD
solutions have been implicated in the pathogenesis, the
precise mechanisms of peritoneal injury are not well
understood [1].
Iron is a metal essential to life and is necessary for a
wide variety of metabolic processes, such as oxygen
transport, electron transfer, nitrogen fixation and DNA
synthesis. Iron exists in two oxidative states, ferrous
(Fe[II]) and ferric (Fe[III]), which can donate or accept
electrons. Although these redox reactions are important for biological reactions, they can also be hazardous
to cells. Redox reactions involving iron play a key
role in the formation of harmful free radicals that
damage living cells via various pathways. To avoid the
hazardous effects of iron, extracellular iron is bound by
transferrin. In addition, two-thirds of serum transferrin
exists as apotransferrin and will quickly capture the free
iron which is released from the cell [2].
The affinity of iron for transferrin is a pH-dependent
process. In plasma (pH 7.4), binding between iron and
transferrin is very strong, whereas virtually no binding
occurs at pH<4.5. Iron is released from transferrin in
the low pH environment of endosomes after receptorbound transferrin is internalized into cells [2]. In the
extracellular space, iron is never released from transferrin under physiological conditions [3].
Conventional PD solutions have unphysiologically
low pH values, around pH 5.0. While several PD
solutions that have more physiological pH values are
available [4,5], low pH PD solutions are still widely
used.
Nephrol Dial Transplant Vol. 19 No. 10 ß ERA–EDTA 2004; all rights reserved
Iron release from transferrin in low pH PD solution and oxidative stress
We report here that iron is released from transferrin
in low pH PD solutions and iron released from
transferrin has oxidative effects in these solutions.
These effects are avoided by neutralizing the pH of the
PD solution. In addition, peritoneal iron deposition
was observed in PD patients.
Subjects and methods
Preparation of iron-saturated human transferrin
Iron was added to 0.2 mM human apotransferrin (Sigma,
St Louis, MO, USA) in 0.1 M N-(2-hydroxyethyl)piperazineN0 -(2-ethanesulfonic acid) (HEPES), 0.025 M NaHCO3, pH
7.4, using sufficient 0.01 M ferrous ammonium sulfate in
0.01 M HCl to bring the protein to 90% saturation [6]. After
a 24 h equilibration period in a 5% CO2 incubator, transferrin
was dialysed against phosphate-buffered saline (PBS: phosphate 1 mM, NaCl 150 mM, pH 7.4). The iron concentration
and unsaturated iron binding capacity (UIBC) were measured
by the clinical laboratory at Saitama Social Insurance
Hospital using an autoanalyser method. Measured saturation
levels were typically 90%. For some experiments, transferrin
was concentrated by ultrafiltration with Ultrafree-15
Centrifugal Filter Units (nominal molecular weight limit
30 000; Millipore, Bedford, MA, USA).
Determination of the effects of PD solution on
iron binding to transferrin
Iron-saturated transferrin (TIBC, 200 mg/dl) was applied
to cassette-type dialysis material (Slide-A-Lyzer, molecular
weight cut-off of 10 000; Pierce, Rockford, IL, USA) and
dialysed against a PD solution (1.36 g/dl glucose, 132 mEq/l
Naþ, 3.5 mEq/l Ca2þ, 0.5 mEq/l Mg2þ, 96 mEq/l Cl,
40 mEq/L lactate, pH 5.2) for 12 h at 37 C. The iron concentration and UIBC were then measured using an autoanalyser method.
Iron binding to transferrin increases the OD 465 of
transferrin (465 nm, 2500 M1 cm1/iron) [6] and iron release
from transferrin was also measured spectrophotometrically.
One millilitre of PD solution was added to 0.01 ml of concentrated iron-saturated transferrin (TIBC, 25 000 mg/dl) and
OD 465 was monitored sequentially. As a control, transferrin
was diluted with PBS (pH 7.4).
Determination of iron release from transferrin in
drained dialysates from PD patients
Drained dialysates from PD patients were concentrated
50-fold by ultrafiltration with Ultrafree-15 Centrifugal
Filter Units (Millipore) and dialysed against fresh PD solution. Since non-transferrin-bound iron is bound by negatively
charged proteins in biological fluid [7], MgCl2 (final
concentration, 20 mM) was added to the PD solution to
prevent non-specific iron binding to proteins. After the
addition of MgCl2, the pH of the PD solution was adjusted
with NaOH. Adding MgCl2 to the PD solution did not affect
iron release from purified human transferrin, as determined
by a spectrophotometric assay. After 12 h of dialysis, the iron
concentration and UIBC were determined as described above.
2593
Red blood cell ghost membrane preparation
Heparinized blood was obtained from healthy volunteers.
Red blood cells (RBC) were washed three times with PBS (pH
7.4) with removal of the buffy coat. After the addition of
lysing buffer [5 mM sodium phosphate, 0.5 mM ethylenediaminetetraacetic acid (EDTA), pH 8.0, 4 C], the ghost
membranes obtained were washed five times with lysing
buffer at 4 C. Between each wash the ghost pellet was passed
five times through a 22-gauge needle to promote further lysis
of the cells. After contaminating red material had been
removed by centrifugation (800 g, 1 min), the white ghost
membranes were washed with low pH buffer (0.5 mM EDTA,
10 mM sodium acetate buffer, pH 4.0) twice and then washed
with PBS (pH 7.4) three times.
Determination of oxidative stress on the RBC
membrane in PD solution
Incubation was initiated by adding PD solution to the RBC
membrane ghost pellet. After the incubation at 37 C for 12 h,
membrane ghosts were washed three times with PBS before
thiobarbituric acid reactive substance (TBARS) and protein
carbonylation were determined.
To initiate the thiobarbituric acid (TBA) reaction, 60 ml of
5.4% sodium dodecyl sulphate (SDS), 300 ml of 20% acetate
buffer (pH 3.5), 10 ml of 0.8% butylhydroxytoluene, 300 ml of
0.8% TBA and 35 ml of 1 mM FeCl3 were subsequently added
[8]. After incubation on ice for 1 h, the reaction mixture was
boiled for 1 h. TBARS was extracted with butanol and then
measured by spectrophotometric assay (OD 532 nm).
Determination of protein carbonyl groups by
immunoblotting
The protein carbonyl groups were determined by immunoblotting [9] using a kit (OxyBlotTM; Intergen, Purchase, NY,
USA). In brief, RBC membrane protein was solubilized with
SDS and the protein carbonyl group was labelled with
dinitrophenylhydrazone (DNP). The protein samples were
then subjected to SDS–polyacrylamide gel electrophoresis
and then transferred to a nitrocellulose membrane. After
blocking with 5% skim milk, the membrane was probed
with rabbit anti-DNP antibody. After extensive washing
with Tris-buffered saline (pH 7.4) with 0.005% Tween-20, the
membrane was incubated with horseradish peroxidase conjugated goat anti-rabbit IgG. Immunoreactive proteins were
visualized with a chemiluminescence detection system (Pierce),
followed by exposure to Kodak X-ray film. Densitometric
analysis was performed using the public domain NIH
ImagePC program.
Iron tissue staining
Iron depositions in peritoneal tissue samples from 12 PD
patients (three autopsy specimens, nine peritoneal biopsy
specimens) and four autopsy cases of haemodialysis patients
who had never received CAPD therapy were examined using
Perls’ reaction with diaminobenzidine (DAB) intensification
[10]. After Perls’ reaction, sections were incubated with 0.5%
DAB in 0.1 M phosphate buffer at pH 7.4 for 20 min,
followed by 15 min in the same medium containing 0.005%
H2O2. The reaction was stopped by rinsing in deionized water
2594
Y. Yamaji et al.
for 30 min and then counterstained with haematoxylin.
Negative control slides were prepared in the same way
through the DAB intensification without the Perls’ solution
preincubation. No positive staining was detected in the
negative control sections.
Statistical analysis
Data are presented as means±SEM. Statistical significance
was evaluated using the paired t-test or one-way analysis of
variance (ANOVA) followed by Fisher’s PLDS test.
Differences at P<0.05 were considered significant.
Results
solution (pH 5.2), the transferrin saturation rate
determined by spectrophotometric assay was also
decreased, as shown in Figure 1B.
The time course of iron release from transferrin in
conventional PD solution was determined by spectrophotometric assay. As shown in Figure 2, 40% of iron
was released within 1 min.
Iron release from transferrin in PD solution
is pH dependent
As shown in Figure 3, neutralizing the pH of the PD
solution prevented the effects of the PD solution on
iron release from transferrin. The pH of the PD
solution was adjusted with NaOH or HCl. At a pH
Conventional PD solution (pH 5.2) causes
iron release from transferrin
Transferrin saturation (%)
A
100
100
Transferrin saturation (%)
Dialysing purified iron-saturated human transferrin
against conventional PD solution (pH 5.2) decreased
the transferrin saturation rate, as shown in Figure 1A.
When concentrated iron-saturated purified transferrin
stock (pH 7.4) was diluted with conventional PD
80
PBS pH 7.4
80
*
60
*
*
*
40
*
PD solution
20
0
0
*
60
5
15
10
Time (minutes)
40
Fig. 2. Iron release from transferrin occurs within 1 min in
conventional PD solution (pH 5.2). The time course of iron release
from transferrin was determined by spectrophotometric assay.
*P < 0.05 vs control.
20
0
Before the dialysis
After the dialysis
against PD solution
100
100
80
60
*
40
20
0
Dilution with PBS
pH 7.4
Dilution with PD
solution
Fig. 1. Conventional PD solution (pH 5.2) causes iron release from
transferrin. (A) Effect of dialysing iron-saturated purified human
transferrin against conventional PD solution (pH 5.2). The transferrin saturation rate was determined using an autoanalyser
method. (B) Effect of diluting iron-saturated purified human
transferrin stock (pH 7.4) with conventional PD solution (pH 5.2).
The transferrin saturation rate was determined by spectrophotometric assay. *P<0.05.
Transferrin saturation (%)
Transferrin saturation (%)
B
80
*
60
*
40
*
20
0
4.0
5.0
6.0
7.0
8.0
pH of PD solution
Fig. 3. Iron release from transferrin in PD solution is pH
dependent. Neutralizing the pH of PD solution avoided the effects
of the PD solution on iron release from transferrin. *P<0.05 vs
before the dialysis.
pH of PD
solution
Transferrin saturation
before dialysis (%)
Transferrin saturation
after dialysis (%)
pH 5.2
pH 7.4
37.6±4.7
37.6±4.7
17.8±2.6a
35.1±5.2
Data are means±SEM.
a
P<0.05 vs before dialysis.
above 6.0, no iron release was observed. The release of
iron from transferrin in low pH PD solutions was
reversible, since neutralizing the pH with NaOH
increased the transferrin saturation rate, compared to
the control value.
Low pH PD solution caused iron release from
transferrin in drained dialysates
The iron concentration and TIBC of the drain dialysate
from PD patients were 1.1±0.3 and 3.2±0.8 mg/dl,
respectively. Table 1 shows the effects of dialysing
drained dialysates from the PD patients against PD
solution. As described in Subjects and methods, we
added 20 mM MgCl2 to the PD solution to prevent
non-specific iron binding to proteins. This addition of
MgCl2 did not affect the iron release from purified
human transferrin. Dialysing the drained dialysates
against a low pH PD solution decreased the transferrin
saturation. This effect was not observed with neutralized PD solution. The transferrin saturation rate in
serum from PD patients was also decreased by
dialysing against a low pH PD solution containing
20 mM MgCl2 (data not shown).
Iron released from transferrin in low pH PD
solution produces oxidative stress on membrane
lipids and proteins
To investigate the effects of iron released from
transferrin on oxidative stress on biological materials,
we incubated human RBC membranes in PD solution.
When iron-saturated transferrin (100 mg iron/dl) was
added to the PD solution at pH 5.2, the production of
TBARS was markedly enhanced, as shown in Figure 4.
This effect of transferrin was abolished by addition
of the iron chelator, deferroxamine (2 mM). When
neutralized PD solution was used, no effects of
iron-saturated transferrin were observed.
Western analysis of the protein carbonyl content of
the RBC membrane showed the protein carbonyl
content to be markedly increased by the addition of
iron-saturated transferrin to the PD solution (pH 5.2).
Figure 5 presents the densitometric analysis of protein
carbonylation (molecular weight between 100 and
30 kDa). This effect was also abolished by deferroxamine. Iron-saturated transferrin had no effect in PD
solution at pH 7.0.
2595
*
5
4
3
2
1
0
C
TF
TF
TF
C
+DEF
pH 5.2
pH 7.0
Fig. 4. Iron released from transferrin in low pH PD solution exerts
oxidative stress on membrane lipids; TBA reaction. Addition of
iron-saturated transferrin (TF) to the PD solution at pH 5.2
enhanced the TBARS production in human RBC membranes. This
effect of transferrin was abolished by the addition of iron chelator
deferroxamine (DEF). When neutralized PD solution (pH 7.0) was
used, the effect of iron-saturated transferrin was not observed.
*P<0.05 vs control.
*
6
Protein Carbonyl Content
(Relative increase)
Table 1. Effect of pH PD solutions on transferrin saturation rate in
drained dialysate from PD patients
TBARS (Relative increase)
Iron release from transferrin in low pH PD solution and oxidative stress
4
2
0
C
TF
TF
C
TF
+DEF
pH 5.2
pH 7.0
Fig. 5. Iron released from transferrin in low pH PD solution exerts
oxidative stress on membrane proteins; protein carbonylation. The
protein carbonyl content was markedly increased by the addition of
iron-saturated transferrin (TF) to the PD solution (pH 5.2). This
effect of transferrin was abolished by the addition of deferroxamine
(DEF). When neutralized PD solution (pH 7) was used, this effect
of iron-saturated transferrin was not observed. *P<0.05 vs control.
Iron deposition in the peritoneum of PD patients
We examined peritoneal iron deposition in PD patients
using Perls’ staining with DAB intensification. In all
peritoneal samples we examined, some form of iron
deposition was observed. In several cases, iron deposition was prominent around the fibrotic area facing the
peritoneal cavity (Figure 6A). Iron deposition was also
2596
Y. Yamaji et al.
Fig. 6. Peritoneal iron deposition in PD patients. Peritoneal iron deposition in PD patients was evaluated by Perls’ staining with DAB
intensification. (A) Iron deposition was prominent around the fibrotic area facing the peritoneal cavity. (B) Iron deposition was also
observed in some vascular walls. (C and D) In two cases of sclerosing encapsulating peritonitis, extensive iron deposition was observed
(black bar ¼ 100 mm).
observed in some vascular walls (B). In two cases of
sclerosing encapsulating peritonitis, iron deposition
was extensive (C and D). In the autopsy specimens
from four haemodialysis patients that lacked CAPDassociated morphological changes, iron deposition was
not observed.
Discussion
Our results clearly show that iron is released from
transferrin in conventional low pH PD solution and
that the released iron exerts oxidative stress. Drained
dialysates from PD patients contain transferrin, which
releases iron in low pH PD solution. Since the iron
release occurs within 1 min, it can be concluded that
iron is released from transferrin in some portion of
the residual PD solution after new PD solution is
introduced into the peritoneal cavity, before the
neutralization of PD solution occurs. Low pH PD
solution may decrease pH in the interstitial space of the
peritoneum and cause iron release. Iron deposition was
observed in all peritoneal samples from PD patients,
and iron deposition was prominent in the fibrotic area
facing the peritoneal cavity. While the origin of this
iron cannot be determined, the iron deposition may be
attributable to the iron released from transferrin.
Various degrees of diffuse peritoneal fibrosis has
been documented in the patients who have been on
long-term PD [1]. Contributions of oxidative stress to
tissue injury and fibrosis have been shown in a variety
of experiments [11]. We showed that iron released from
transferrin in low pH PD solution exerts strong
oxidative stress. Although the concentration of ironsaturated transferrin in peritoneal fluid is lower than
the concentration we used in these in vitro experiments, the peritoneum of PD patients is exposed to the
iron-induced oxidative stress whenever new low pH PD
solutions are introduced into the peritoneal cavity.
Since various oxidative reactions are irreversible,
the iron-induced oxidative damage accumulates in
the peritoneum of patients receiving long-term PD
with low pH PD solutions.
We found iron deposition in the fibrotic peritoneum of PD patients and extensive iron deposition
was observed in the severely fibrotic peritoneum of
patients with sclerosing encapsulating peritonitis.
Excess iron deposited in the liver has been shown
to cause hepatic fibrosis, through oxidative stress, in
several animal models [12]. In addition, intraperitoneal administration of a high dose of iron dextran
reportedly causes peritoneal fibrosis in rat PD
models [13]. These observations support the hypothesis that oxidative stress induced by iron released
from transferrin in low pH PD solution may play
an important role in peritoneal injury and fibrosis in
PD patients.
Several immunohistochemical studies have shown
that advanced glycation end products (AGEs) accumulate in the peritoneal tissue of PD patients [14]. This
suggests that a high concentration of glucose results in
local generation of AGEs in the PD patient peritoneum. It has been shown that oxidative stress plays a
key role in the formation of AGEs. Since metalcatalysed oxidation accelerates the formation of AGEs
[15], iron released from transferrin in low pH PD
solution may contribute to peritoneal AGEs formation
in PD patients.
For the treatment of anaemia in dialysis patients,
the use of intravenous iron is widely recommended
[16]. Recently, several studies in haemodialysis patients
have shown intravenous iron to induce oxidative
stress [17], which may accelerate vascular complications. Iron excess may also enhance the pre-existing
Iron release from transferrin in low pH PD solution and oxidative stress
risk for infectious complications or malignant disorders
in dialysis patients [18,19]. In addition to these proposed hazardous effects of iron treatment in dialysis
patients, our results raise the possibility that iron excess
may enhance the peritoneal injury in PD patients.
Neutralizing the pH of PD solution, i.e. raising the
pH above 6.0, prevents iron release from transferrin
and also avoids the oxidative stress produced by ironsaturated transferrin. Since a low pH environment
reportedly enhances the oxidative stress induced by
iron [20], a low pH PD solution may enhance the
oxidative stress induced by iron deposited in tissues as
we observed in peritoneal tissues. This possible
oxidative stress can also be avoided by neutralizing
the PD solution.
Recently, several PD solutions with a neutral pH
value have been introduced. Several markers of
peritoneal injury in drained dialysates are reportedly
improved by switching from conventional PD solution
to the neutral PD solutions [4,5]. While these neutral
PD solutions also have the advantage of lower
concentrations of cytotoxic glucose degradation products, the iron-related mechanisms we proposed herein
may also contribute to the improved biocompatibility
of these solutions. While the use of a neutral PD
solution for 2 years does not improve the peritoneal
transport characteristics [4], a longer study period may
reveal some benefits of neutral PD solution in
peritoneal transport parameters. It is also possible
that neutralized PD solution would decrease the
incidence of sclerosing encapsulating peritonitis,
which is a potentially fatal complication of PD.
Acknowledgements. We are grateful to Tomofusa Migita,
Kazuyuki Komuro and staff of the clinical chemistry laboratory
in Saitama Social Insurance Hospital for their technical assistance.
Conflict of interest statement. We have had no involvements that
might raise questions of bias in the work reported or in the
conclusions, implications, or opinions stated.
References
1. Mactier RA. The spectrum of peritoneal fibrosing syndromes in
peritoneal dialysis. Adv Perit Dial 2000; 16: 223–228
2. Ponka P. Cellular iron metabolism. Kidney Int Suppl 1999; 69:
S2–S11
3. Leake DS. Does an acidic pH explain why low density
lipoprotein is oxidised in atherosclerotic lesions? Atherosclerosis
1997; 129: 149–157
2597
4. Jones S, Holmes CJ, Krediet RT et al. Bicarbonate/lactatebased peritoneal dialysis solution increases cancer antigen 125
and decreases hyaluronic acid levels. Kidney Int 2001; 59:
1529–1538
5. Rippe B, Simonsen O, Heimburger O et al. Long-term clinical
effects of a peritoneal dialysis fluid with less glucose
degradation products. Kidney Int 2001; 59: 348–357
6. Thompson CP, Grady JK, Chasteen ND. The influence of
uncoordinated histidines on iron release from transferrin.
A chemical modification study. J Biol Chem 1986; 261:
13128–13134
7. Gosriwatana I, Loreal O, Lu S, Brissot P, Porter J, Hider RC.
Quantification of non-transferrin-bound iron in the presence of
unsaturated transferrin. Anal Biochem 1999; 273: 212–220
8. Kikugawa K, Kojima T, Yamaki S, Kosugi H. Interpretation
of the thiobarbituric acid reactivity of rat liver and brain
homogenates in the presence of ferric ion and ethylenediaminetetraacetic acid. Anal Biochem 1992; 202: 249–255
9. Robinson CE, Keshavarzian A, Pasco DS, Frommel TO,
Winship DH, Holmes EW. Determination of protein carbonyl
groups by immunoblotting. Anal Biochem 1999; 266: 48–57
10. Pang JH, Jiang MJ, Chen YL et al. Increased ferritin gene
expression in atherosclerotic lesions. J Clin Invest 1996; 97:
2204–2212
11. Poli G, Parola M. Oxidative damage and fibrogenesis. Free
Radic Biol Med 1997; 22: 287–305
12. Tsukamoto H, Horne W, Kamimura S et al. Experimental liver
cirrhosis induced by alcohol and iron. J Clin Invest 1995; 96:
620–630
13. Park SE, Twardowski ZJ, Moore HL. Chronic administration
of iron dextran into the peritoneal cavity of rats. Perit Dial Int
1997; 17: 179–185
14. Honda K, Nitta K, Horita S et al. Accumulation of advanced
glycation end products in the peritoneal vasculature of continuous ambulatory peritoneal dialysis patients with low ultrafiltration. Nephrol Dial Transplant 1999; 14: 1541–1549
15. Sajithlal GB, Chithra P, Chandrakasan G. The role of metalcatalyzed oxidation in the formation of advanced glycation end
products: an in vitro study on collagen. Free Radic Biol Med
1998; 25: 265–269
16. Eschbach JW. Current concepts of anemia management in
chronic renal failure: impact of NKF-DOQI. Semin Nephrol
2000; 20: 320–329
17. Roob JM, Khoschsorur G, Tiran A, Horina JH, Holzer H,
Winklhofer-Roob BM. Vitamin E attenuates oxidative stress
induced by intravenous iron in patients on hemodialysis. J Am
Soc Nephrol 2000; 11: 539–549
18. Weinberg ED. Iron therapy and cancer. Kidney Int Suppl 1999;
69: S131–S134
19. Patruta SI, Horl WH. Iron and infection. Kidney Int Suppl
1999; 69: S125–S130
20. Morgan J, Leake DS. Oxidation of low density lipoprotein
by iron or copper at acidic pH. J Lipid Res 1995; 36:
2504–2512
Received for publication: 11.10.03
Accepted in revised form: 6.2.04