Interrelationships between tissue iron status and erythropoiesis during postweaning development following neonatal iron deficiency in rats

Abstract

Dietary iron is particularly critical during periods of rapid growth such as in neonatal development. Human and rodent studies have indicated that iron deficiency or excess during this critical stage of development can have significant long- and short-term consequences. Since the requirement for iron changes during development, the availability of adequate iron is critical for the differentiation and maturation of individual organs participating in iron homeostasis. We have examined in rats the effects of dietary iron supplement following neonatal iron deficiency on tissue iron status in relation to erythropoietic ability during 16 wk of postweaning development. This physiological model indicates that postweaning iron-adequate diet following neonatal iron deficiency adversely affects erythroid differentiation in the bone marrow and promotes splenic erythropoiesis leading to splenomegaly and erythrocytosis. This altered physiology of iron homeostasis during postweaning development is also reflected in the inability to maintain liver and spleen iron concentrations and the altered expression of iron regulatory proteins in the liver. These studies provide critical insights into the consequences of neonatal iron deficiency and the dietary iron-induced cellular signals affecting iron homeostasis during early development.

adequate dietary iron is critical during the periods of rapid growth that include the fetal and neonatal stages. Iron deficiency during these critical periods of development can induce general metabolic dysfunction by causing tissue hypoxia and affecting the development of organs, and these outcomes may not be reversed by iron supplementation (22, 30). On the other hand, excessive iron can generate free radicals via the Fenton and Heber-Weiss reactions to mediate detrimental effects (28). Studies from humans and experimental animals indicate that iron absorption is extremely high in neonatal mammals and decreases to adult levels after weaning, suggesting a developmental regulation of iron absorption (1, 10). These age-dependent changes in absorption have been attributed to the intrinsic characteristics of the intestine rather than bioavailability (12, 14). Iron deficiency and hypoxia in neonatal rats is known to cause a delay in the maturation of intestinal enzymes (6, 18). The precise regulation of physiologically active iron in the tissues combined with the developmental changes in neonates will be critical for later development and for preventing a risk of chronic diseases (8).

Iron availability for erythropoiesis and other vital cellular functions is regulated by intestinal uptake, release/storage from hepatocytes, and recycling from red blood cells. An antimicrobial peptide hepcidin, a key regulator of iron metabolism, regulates intestinal iron absorption and iron recycling by macrophages (13, 19). Variation in hepcidin expression by transcriptional regulators primarily acts through IL-6/STAT3 and BMP/Smad signaling pathways (4). However, dietary iron-dependent hepcidin regulation and the proteins that mediate this pathway are still being investigated. Since hepcidin regulation can also be influenced by inflammation, hypoxia, and augmented erythropoiesis, several of these factors may be present concomitantly during pathological situations to regulate hepcidin expression (5, 9). Studies have indicated that hepcidin expression during such situations is determined by the strength of the individual stimuli, and erythropoietic drive can inhibit both inflammatory and iron-sensing pathways (17).

Although age-dependent differences in the individual tissue responses to iron deficiency and iron overload have been studied, information on the effects of dietary iron supplementation following neonatal iron deficiency on erythropoietic and tissue iron status is lacking. This information is vital to understand how respective organs responsible for iron homeostasis respond to dietary iron following iron deficiency during the critical stages of development. Therefore, we investigated the responses of liver, spleen, and bone marrow to postweaning iron-adequate diet following a neonatal iron deficiency. Using a rat model, we examined 1) iron concentrations in the liver, spleen, and bone marrow; 2) expression of iron regulatory proteins in the liver; and 3) the ability to produce erythroid progenitors in the bone marrow and the spleen, during 16 wk of postweaning development.

EXPERIMENTAL PROCEDURES

Dietary treatments.

Sprague-Dawley rats for breeding stocks were purchased from Harlan Sprague Dawley (Indianapolis, IN). Iron-adequate (80 μg iron/g diet) and iron-deficient (4 μg iron/g diet) diets were purchased from Teklad (Harlan Laboratories, Madison, WI) with ferric citrate as the iron source. The iron-deficient diet contained all components of the iron-adequate diet with the exception of ferric citrate, and the iron concentration in the diet was verified by atomic absorption spectrophotometry after wet digestion with nitric acid. All rats received food and deionized distilled water ad libitum and were housed in a temperature (22 ± 1°C)- and humidity (40%)-controlled room maintained on a 12:12-h light-dark cycle.

Female breeder rats were fed an iron-adequate diet (80 μg iron/g diet) for 3 wk prior to mating. Pregnant dams continued to be fed an iron-adequate diet until gestational day 15 (G15), at which time they were divided into two groups: 1) a control group fed an iron-adequate diet (80 μg iron/g diet) and 2) an iron-deficient group fed an iron-deficient diet (4 μg iron/g diet), through postnatal day 23 (P23) (weaning). On P2, litters were culled to 10 rats per litter. After weaning on P23, pups from both dietary groups were fed an iron-adequate diet (80 μg iron/g diet) for up to 16 wk. This feeding schedule produced two groups of experimental rats: 1) iron adequate throughout (IA), and 2) iron deficient from G15 to P23 and then iron adequate beginning at P23 (IDIA). All experimental protocols were in accordance with the National Institutes of Health animal care guidelines and were approved by the Pennsylvania State University Institutional Animal Care and Use Committee.

Hemoglobin, hematocrit, and transferrin saturation.

Whole blood was collected from IA and IDIA rats at P23 and at 2, 4, 8, 12, and 16 wk postweaning. Hemoglobin was analyzed with the cyanomethemoglobin method according to the manufacturer's instructions (Sigma-Aldrich, St. Louis, MO). Briefly, whole blood samples (20 μl) were collected in microcapillary tubes and mixed with 4 ml Drabkin's solution (0.1% sodium bicarbonate, 0.005% potassium cyanide, and 0.02% potassium ferricyanide). Values were determined photometrically at 540 nm and calculated relative to a standard curve (Sigma-Aldrich). Hematocrit was calculated after centrifugation of blood samples in heparinized microcapillary tubes (13,700 g, 5 min, room temperature). Transferrin saturation was calculated as serum iron/TIBC × 100.

Serum Epo.

Serum erythropoietin (Epo) concentrations of IA and IDIA rats were determined by immunoassay (ELISA) using a Quantikine Rat Erythropoietin assay kit and according to the manufacturer's instructions (R&D Systems, Minneapolis, MN).

Liver, spleen, and bone marrow iron concentration.

IA and IDIA rats were decapitated on P23, or 2, 4, 8, 12, or 16 wk postweaning after brief exposure to CO₂. One lobe from each liver was quickly dissected and stored in RNAlater (Sigma, St. Louis, MO) for total RNA isolation. A second lobe of liver was stored in RIPA buffer (Pierce, Rockford, IL) containing complete mini protease inhibitor cocktail (Roche, Mannheim, Germany) for protein analysis. The remaining liver and the spleen tissue samples were used to estimate tissue iron content. All samples were stored at −80°C until further processing. Each treatment group included eight rats per time point that represented four litters.

Liver and spleen tissue iron levels were quantified colorimetrically by a modification of the method of Cook (34). Liver (200 mg) and spleen (100 mg) tissues were digested overnight in 3 M HCl/10% trichloroacetic acid at 65°C, followed by the addition of chromogen (0.01% bathophenanthroline sulfonate, 0.1% thioglycolic acid, and 4 M sodium acetate). For a standard curve, serial dilutions of a ferric iron standard (Sigma-Aldrich) were used. Color was allowed to develop for 15 min, and absorbance was measured at 535 nm.

Bone marrow (femurs) was collected at 4 and 8 wk postweaning from IA and IDIA rats. Bones were immediately placed into fixative (4% paraformaldehyde) in phosphate buffer. After 48 h fixation, bones were decalcified and processed for embedding into paraffin by standard methods. Sections were cut at 5 μm, deparaffinized in xylene, and hydrated in distilled water. Sections were then placed in staining solution, consisting of equal volumes of 4% potassium ferrocyanide and 4% hydrochloric acid (Polysciences, Warrington, PA), for 30 min at room temperature. Following a rinse in distilled water, sections were counterstained in 1% nuclear fast red (Polysciences) for 5 min and washed in distilled water. Sections were then dehydrated through ascending alcohols to xylene and mounted in Poly-Mount for microscopic observation.

Serum GDF15 and TWSG1 measurement.

For GDF15 and TWSG1 measurements, serum collected from IA and IDIA rats were pooled by treatment group. The total protein content of the pooled serum was assessed by the BCA protein assay (Thermo Scientific, Rockford, IL). An equal quantity of total protein (20 μg) was then loaded on 10% SDS-PAGE, and expression of GDF15 and TWSG1 was measured by Western blotting technique using anti-GDF15 and anti-TWSG1 antibodies (Abcam, Cambridge, MA), respectively.

Total RNA isolation, reverse transcription and real-time PCR.

For isolating total RNA from liver and spleen tissues, samples were pooled by treatment group for each time point. Total RNA from liver and spleen tissue was isolated using an RNeasy mini kit (Qiagen, Valencia, CA) following the protocol provided by the manufacturer. To eliminate genomic DNA contamination in the RNA preparation, on column DNAse treatment was performed during RNA isolation. RNA quality was examined with standard 260/280 nm spectrophotometer readings, and this ratio was maintained at >2.1. One microgram of total RNA was used to prepare cDNA by using an RT² First Strand Kit (SABiosciences, Frederick, MD). The PCR reaction mix (25 μl reaction volume) consisted of 12.5 μl of iQ SYBR Green Supermix (Bio-Rad), 200 nM of primers (hepcidin forward taggacaggaataaataatggg, hepcidin reverse cagaaagcaagactgatgacag, BMP6 forward gggagagggactggagcccg, BMP6 reverse gtcgctggcgtggggagaac), and 1 μl of cDNA. PCR reactions were performed with the following cycling conditions: 1 cycle at 95°C for 10 min, 40 cycles of 30 s at 60°C and 45 s at 72°C. cDNA prepared from spleen tissue was analyzed by using a real-time PCR array (Rat TGF-β BMP signaling pathway, SABiosciences, Frederick, MD) following the procedure provided by the manufacturer. Data were analyzed by the standard ΔΔCt method with β-actin as an internal control.

Protein isolation and Western blot analysis.

Liver tissues collected were pooled by treatment group for each time point and homogenized in cold RIPA buffer (Pierce) containing protease inhibitor cocktail (Roche, Indianapolis, IN). Cell debris were centrifuged (20,000 g, 20 min, 4°C), and the total protein content of the extracts was assessed by using the BCA protein assay (Thermo Scientific). A total of 10 μg of protein was loaded onto a 10% sodium dodecyl sulfate polyacrylamide gel and separated by electrophoresis (SDS-PAGE). Separated proteins were transferred to nitrocellulose membrane (Bio-Rad, Hercules, CA) using a Mini Trans-Blot Cell (Bio-Rad). Membranes were blocked with phosphate buffer (137 mmol/l NaCl, 2.7 mmol/l KCl, 10 mmol/l Na₂HPO₄, 1.76 mmol/l KH₂PO₄, pH 7.4) containing 0.05% Tween 20 and 5% nonfat dried milk. Rabbit polyclonal antibodies to ferritin and TfR2 (Santa Cruz Biotechnology, Santa Cruz, CA) were used to detect expression levels of respective proteins. Respective horseradish peroxidase-conjugated secondary antibodies (Sigma-Aldrich) were used to identify the protein band of interest with use of Super Signal West Pico Chemiluminescent substrate (Pierce). The membranes were stripped and reprobed with anti-β-actin antibody (Sigma-Aldrich) to confirm equal loading and transfer. Protein expression levels were quantitated by densitometric scanning using a Carestream Molecular Imaging System (New Haven, CT) and expressed as percentage of control.

Erythroid progenitor assays.

Erythroid colonies were assayed as previously described (20, 25). Briefly, cells isolated from bone marrow and spleen (1 × 10⁶ cells) were plated in methylcellulose media containing 3 U/ml EPO (Stem Cell Technologies, Vancouver, BC), 15 ng/ml bone morphogenic protein 4 (BMP4; R&D Systems), and 2.5 ng/ml interleukin 3 (IL-3; Sigma) in triplicate. CFU-E (colony-forming unit-erythroid) and BFU-E (burst-forming unit-erythroid) were scored by acid benzidine staining (11) after incubation for 2 and 5 days, respectively.

Statistical method.

All experimental groups contained six to eight rats per time point that represented four litters. Data were compiled from three independent trials. Data are presented as means ± SE. Differences between means were tested for statistical significance by using t-test probabilities. Differences were considered significant at P < 0.05.

RESULTS

Postweaning iron-adequate diet restores hemoglobin and serum Epo concentrations but leads to erythrocytosis in rats that experienced iron deficiency during the neonatal period (IDIA).

Hemoglobin concentrations were significantly lower in IDIA rats at P23 and at 2 wk postweaning (P < 0.05). But hemoglobin concentrations in IDIA rats were restored to the level of IA rats by 4 wk postweaning (Fig. 1A). Serum Epo concentrations were significantly elevated in IDIA rats at P23 and at 2 and 4 wk postweaning. However, by 8, 12, and 16 wk serum Epo levels were similar to IA rats (P < 0.001 for all; Fig. 1B). Serum transferrin saturation was significantly reduced in IDIA rats at 2, 4, 8, and 12 wk postweaning (Fig. 1C; P < 0.05). Hematocrit levels in IDIA rats were significantly reduced at P23 because of the iron deficiency but were restored to the level of IA rats at 2 wk postweaning. From 4 to 16 wk, IDIA rats exhibited significantly elevated hematocrit (P < 0.05; Fig. 1D). Paradoxically, erythrocytosis was associated with normal serum Epo levels.

Fig. 1.

Effect of postweaning dietary iron on hemoglobin, hematocrit, transferrin saturation, and serum erythropoietin concentration. A: changes in hemoglobin concentrations of rats that were iron adequate throughout (IA) and iron deficient from gestational day 15 to postnatal day 23 (IDIA; *P < 0.05 relative to respective IA rats). B: serum erythropoietin (Epo) concentrations (*P < 0.001 relative to respective IA rats) n = 6 to 8 per diet group per time point. C: transferrin saturation (*P < 0.01 relative to respective IA rats). D: hematocrit (*P < 0.05 relative to respective IA rats).

IDIA rats are unable to maintain liver and spleen iron concentrations during postweaning development.

Liver and spleen iron concentrations were significantly reduced in IDIA rats at P23 (P < 0.05); however, 2 wk of iron-adequate diet restored liver and spleen iron levels in IDIA rats to that of IA rats. In contrast to IA rats, after 4 wk of iron-adequate diet, IDIA rats showed a significantly reduced iron concentration in both liver (P < 0.05) and spleen tissues (P < 0.05) that was maintained through 16 wk of supplementation (Fig. 2, A and B). In contrast to the liver and spleen, histochemical analysis of bone marrow sections at 4–16 wk postweaning showed an increased Prussian blue reaction in bone marrows of IDIA rats, suggesting the presence of more iron relative to IA rats (Fig. 2C).

Fig. 2.

Effect of postweaning dietary iron on liver, spleen, and bone marrow iron concentration. A: liver iron concentration (*P < 0.05 relative to respective IA rats). B: spleen iron concentration (*P < 0.05 relative to respective IA rats), n = 6 to 8 per diet group per time point. C: representative bone marrow section. Prussian blue reaction for the demonstration of iron (n = 6 per diet group).

Expression of iron-regulatory proteins in the liver.

To further investigate the regulation of iron absorption during the postweaning period, we examined the expression of the mRNA encoding the iron-regulatory peptide hepcidin and BMP6, an endogenous regulator of hepcidin, in the livers of rats at different time points during postweaning development. Expression of hepcidin in IDIA rats was significantly reduced at P23 and at 2 and 4 wk postweaning (P < 0.01 for all; Fig. 3A). Expression of BMP6 in IDIA rats was significantly elevated only at 2 wk following weaning (P < 0.001; Fig. 3B). Expression of ferritin, an indicator of iron storage in the liver, was reduced in IDIA rats at P23 (P < 0.05); however, 2 wk of iron-adequate diet following weaning was able to restore ferritin levels similar to IA rats (Fig. 3C). Liver ferritin was again significantly reduced by 8 wk postweaning (P < 0.05; Fig. 3C), indicating that stored iron was being utilized to meet erythropoietic demand. Expression of TfR2 in the liver of IDIA rats was significantly elevated at P23 and at 2 wk postweaning (P < 0.05 for both; Fig. 3C), whereas TfR2 expression was not changed significantly in these rats at 8 wk postweaning despite low ferritin levels.

Fig. 3.

Effect of postweaning dietary iron on the expression of iron regulatory genes. RT-PCR analysis. A: hepcidin. B: BMP6 in liver tissue. cDNA were prepared from pooled liver tissues of IA (n = 8) and IDIA (n = 8) rats. Data were analyzed by the ΔΔCt method; expression was normalized to β-actin expression and presented as fold change. Hepcidin, *P < 0.01 relative to respective IA group; BMP6, *P < 0.001 relative to respective IA group. C: 10 μg of total protein isolated from pooled liver tissues of IA (n = 8) and IDIA (n = 8) rats, subjected to SDS-PAGE and then analyzed by Western blot with anti-ferritin (left) and anti-TfR2 (right) antibodies. Proteins detected were quantitated by densitometry and the ratio was determined in arbitrary units. *P < 0.05 relative to respective IA group.

Serum GDF15 and TWSG1 levels during postweaning development.

A significant elevation in serum GDF15 levels in IDIA rats was observed at 4 and 8 wk postweaning relative to IA rats whereas serum TWSG1 levels were found to be significantly elevated in IDIA rats at 4 wk postweaning (Fig. 4, P < 0.05).

Fig. 4.

Expression of GDF15 and TWSG1 in the serum of IA and IDIA rats. Total protein content of the pooled serum was assessed by BCA protein assay. An equal quantity of total protein (20 μg) was loaded on 10% SDS-PAGE, and expression of GDF15 and TWSG1 was measured by Western blotting technique with anti-GDF15 and anti-TWSG1 antibodies. The ratio was determined in arbitrary units. *P < 0.05 relative to respective IA group (n = 8 per diet group).

Bone marrow erythropoiesis was suppressed and splenic erythropoiesis was activated in IDIA rats during postweaning development.

Erythropoiesis occurs in the fetal liver during early development, and after birth, the bone marrow becomes the predominant site of definitive steady-state erythropoiesis. To analyze erythrocyte producing ability, an erythroid progenitor assay was performed during postweaning development. The numbers of early BFU-E and late CFU-E erythroid progenitors were significantly reduced (P < 0.01) in the bone marrow of IDIA rats compared with IA rats at all time points (Fig. 5, A and B). Erythroid colony numbers as well as size were found to be reduced in IDIA rats compared with age-matched IA rats during the postweaning period (Fig. 5C). Erythropoiesis shifts to the spleen at times of anemic stress. We measured splenic CFU-E and BFU-E scores in IDIA rats and both CFU-E and BFU-E were significantly increased (P < 0.001) compared with IA rats during the postweaning period (Fig. 6, A and B). In addition, erythroid colony number as well as size was found to be increased in the spleens of IDIA rats compared with age-matched IA rats during postweaning development (Fig. 6C). These data suggest that during postweaning development, bone marrow erythropoiesis is suppressed in IDIA rats and splenic erythropoiesis is elevated (stress erythropoiesis). The stress erythropoiesis may be the cause of the erythrocytosis observed in IDIA rats. Splenic cDNA analysis using an RT-PCR array (Rat TGF-β BMP signaling pathway) demonstrates an increased expression of bone morphogenic protein 4 (BMP4) and bone morphogenic protein 2 (BMP2), which are responsible for the activation of stress erythropoiesis (Fig. 6D; P < 0.0001) (20, 25). In addition, splenic weights of IDIA rats were found to be increased during the postweaning period (4 to 16 wk) compared with age-matched IA rats (Fig. 6E; P < 0.005).

Fig. 5.

Expansion of erythroid progenitor cells in the bone marrows of IA and IDIA rats during postweaning development. A: relative colony-forming units-erythroids (CFU-E) per 1 × 10⁶ bone marrow cells. B: relative burst-forming units-erythroids (BFU-E) per 1 × 10⁶ bone marrow cells. *P < 0.01 relative to respective IA group; n = 6 to 8 per diet group per time point. C: representative erythroid colonies.

Fig. 6.

Expansion of erythroid progenitor cells in the spleen of IA and IDIA rats during postweaning development. A: relative CFU-E per 1 × 10⁶ splenic cells. B: relative BFU-E per 1 × 10⁶ splenic cells. *P < 0.001 relative to respective IA group. n = 6 to 8 per diet group per time point. C: representative erythroid colonies. D: RT-PCR analysis of cDNA prepared from pooled spleen tissue from IA and IDIA rats (n = 8 per each diet group) using TGF-β BMP signaling pathway PCR array. Data was analyzed by the standard ΔΔCt method with β-actin as an internal control. Changes in the expression of BMP2 and BMP4 genes represented in the array are presented as fold change. E: increase in splenic weight of IDIA rats compared with IA rats during postweaning development (n = 8 per diet group).

DISCUSSION

Our rat model of neonatal iron deficiency has enabled us to examine how peripheral organ systems respond to dietary iron treatment during postweaning development. The reduction in dietary iron from G15 through P23 resulted in the development of an anemia in mothers. Pups nursed by these mothers utilize their reserve iron more rapidly and experience iron deficiency during this active growth (neonatal) period. Providing iron-adequate diet for 4 wk during the postweaning period restored hemoglobin concentrations in IDIA rats. Hematocrit values were elevated in these rats beginning at 4 wk postweaning and reaching a maximum at 8 wk postweaning (20% more compared with age-matched IA rats). Since elevated hematocrit has been reported to occur during many hematological abnormalities such as erythrocytosis and Chuvash polycythemia (15, 16), this observation prompted us to investigate further how iron availability in IDIA rats relates to erythropoietic ability and tissue iron concentrations during postweaning development.

At P23, IDIA rats had depleted liver iron and splenic iron stores compared with IA rats, which is reflective of receiving an iron-deficient diet in neonatal life. Two weeks of iron-adequate diet postweaning helped to restore liver and spleen iron concentrations; however, following this initial 2 wk of iron-adequate diet, IDIA rats were not able to maintain their liver and spleen iron concentrations. This observation was supported by the liver ferritin data showing that 2 wk of iron-adequate diet restored ferritin concentrations in IDIA rats, but by 8 wk, ferritin concentrations were depleted to 30% that of age-matched IA rats. These data indicate that, during postweaning development, there is a dysregulation of iron homeostasis in IDIA rats despite sufficient iron availability.

In the liver, hepcidin transcription responds to the body's iron requirement (21, 36), and BMP6 has been identified as an upstream regulator of hepcidin (3, 23). Expression of hepcidin in the liver of IDIA rats was greatly reduced at P23 as we would expect during iron deficiency; however, this suppressed hepcidin expression continued through 4 wk of iron-adequate diet. Suppression in hepcidin expression despite sufficient iron availability during the first 4 wk of postweaning development resembles the conditions of iron deficiency, hypoxia, and erythropoietic expansion (2). In this study, 2 wk of postweaning iron-adequate diet upregulated BMP6 expression in the liver of IDIA rats compared with age-matched IA rats. BMP6 binds to type 1 and type 2 BMP receptors and coreceptor hemojuvelin on hepatocytes, leading to activation of SMAD proteins. Phosphorylated SMAD 1/5/8 forms complexes with SMAD4 and translocates to the nucleus to activate the target hepcidin gene promoter (7). Our observation on the upregulation of BMP6 and suppressed hepcidin expressions, during the time of sufficient iron availability in IDIA rats during postweaning development, recapitulates conditions observed during ineffective erythropoiesis (7, 11, 34) and hepcidin regulation is subjected to additional regulatory effects.

Several studies have attempted to identify signals from the bone marrow that communicate the demand for iron, and these investigations have resulted in identifying growth differentiation factor 15 (GDF15) and twisted gastrulation (TWSG1) as a candidate molecules. Using a transcriptional profiling approach during erythropoiesis, GDF15 was previously found to be upregulated in thalassemic serum and is known to suppress hepcidin expression in vitro (33). In this study, we have investigated whether GDF15 or TWSG1 as an erythroid signal might have played a role in the suppression of hepcidin expression during postweaning development in IDIA rats. An increase in GDF15 protein levels in the serum suggests that the erythroblast signal GDF15 might have contributed to hepcidin suppression at least during the initial period of postweaning development in IDIA rats. Elevated TWSG1 observed in the serum of IDIA rats at 4 wk postweaning might have also contributed to suppress hepcidin expression by interfering with BMP signaling as a BMP antagonist. However, how dietary iron alters BMP/SMAD signals through these two erythroblast signals during postweaning development needs further investigation.

Several studies have indicated that apoptotic erythroblasts produce GDF15, which contributes to extraerythroid tissue iron overloading due to ineffective erythropoiesis (24, 27). Although GDF15 is not considered a major erythroid regulator of iron under normal conditions, it has been proposed to contribute to hepcidin suppression and iron overloading during ineffective erythropoiesis (29). Studies have also indicated that TWSG1 secreted from erythroblasts may contribute to iron loading by inhibiting BMP-mediated hepcidin expression (32). In this study we did not observe iron overload, at least in liver and spleen tissues of IDIA rats, while feeding an iron-adequate diet during the 16 wk postweaning period. Prussian blue reaction in the bone marrow sections of IA and IDIA rats indicates that there is an increased presence of iron in the bone marrows of IDIA rats during the postweaning period. This observation may indicate the inability to utilize iron to maintain steady-state erythropoiesis in the bone marrow of IDIA rats during postweaning development. Mechanisms that affect normal erythropoietic ability in these rats need to be further investigated.

Bone marrow remains the primary site of steady-state erythropoiesis after birth; however, splenic erythropoiesis has been reported to occur in adult mammals during iron deficiency, severe blood loss, and certain pathological conditions (15, 26). In the present study, IDIA rats showed a 50% reduction in erythroid progenitor cells in the bone marrow of IDIA rats compared with the age-matched IA rats. These data suggest that during postweaning development, there is an inability to produce sufficient erythroid progenitors in the bone marrow of rats that experienced iron deficiency during the neonatal period. In addition, IDIA rats showed a twofold higher level of erythroid progenitor cells in the spleens, suggesting that, to compensate for the defect in bone marrow erythropoiesis, splenic erythropoiesis is activated in these rats. We also observed an increase in splenic weights of IDIA rats during postweaning development. It has been reported that activation of stress erythropoiesis is BMP4 dependent (20, 25), and our RT-PCR array analysis shows a twofold increase in the expression of BMP4 expression in the spleens of IDIA rats, suggesting that postweaning dietary iron while reducing bone marrow erythropoiesis induced BMP4 dependent stress erythropoiesis in these rats. There are several factors that promote splenic erythropoiesis such as severe blood loss and mutations in von Hippel-Lindau (VHL) gene and hypoxia-inducible factors (15). This is, to our knowledge, the first report suggesting that providing adequate dietary iron following neonatal iron deficiency may promote splenic erythropoiesis. Interestingly, stress erythropoiesis in IDIA rats is active despite normal concentrations of Epo, at 8 to 16 wk postweaning in this study.

In conclusion, we demonstrated that a postweaning iron-adequate diet following neonatal iron deficiency in rats adversely affects the ability to maintain liver and spleen iron status and suppresses erythroid differentiation in the bone marrow. To compensate for the decreased bone marrow erythropoiesis, a rapid expansion of erythroid progenitors in the spleen results in splenomegaly and erythrocytosis. This physiological model helps us to probe systemic iron regulation following neonatal iron deficiency: specifically, dietary iron-induced molecular signals. In addition, these data present information on how individual organs responsible for iron homeostasis respond to postweaning dietary iron (a time at which human tissue samples are unavailable for analysis), which is essential for the development of dietary recommendations and therapies for anemia during postnatal development.

GRANTS

This work was supported by Agriculture and Food Research Initiative Grant 35200-18231 from the USDA National Institute for Food and Agriculture.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).