Gastrin-Deficient Mice Have Disturbed Hematopoiesis in Response to Iron Deficiency

Abstract

Gastrins are peptide hormones important for gastric acid secretion and growth of the gastrointestinal mucosa. We have previously demonstrated that ferric ions bind to gastrins, that the gastrin-ferric ion complex interacts with the iron transport protein transferrin in vitro, and that circulating gastrin concentrations positively correlate with transferrin saturation in vivo. Here we report the effect of long-term dietary iron modification on gastrin-deficient (Gas^−/−) and hypergastrinemic cholecystokinin receptor 2-deficient (Cck2r^−/−) mice, both of which have reduced basal gastric acid secretion. Iron homeostasis in both strains appeared normal unless the animals were challenged by iron deficiency. When fed an iron-deficient diet, Gas^−/− mice, but not Cck2r^−/−mice, developed severe anemia. In iron-deficient Gas^−/−mice, massive splenomegaly was also apparent with an increased number of splenic megakaryocytes accompanied by thrombocytosis. The expression of the mRNA encoding the iron-regulatory peptide hepcidin, Hamp, was down-regulated in both Cck2r^−/− and Gas^−/−mice on a low-iron diet, but, interestingly, the reduction was greater in Cck2r^−/− mice and smaller in Gas^−/− mice than in the corresponding wild-type strains. These data suggest that gastrins play an important direct role, unrelated to their ability to stimulate acid secretion, in hematopoiesis under conditions of iron deficiency.

Under normal conditions, 0.5–2 mg of iron is absorbed from food each day via the duodenum, and about the same amount is lost through bleeding and through sloughing of skin and mucosal cells (1). Approximately 0.1% of body iron is bound to transferrin in the plasma and forms a dynamic iron transit pool. The largest consumers of iron in the body are developing erythrocytes within the bone marrow (2). The requirement of erythrocytes for iron exceeds the amount taken up from food each day and therefore, to keep up with the daily erythrocyte iron demand, iron from senescent erythrocytes is recycled by reticuloendothelial macrophages and returned to plasma transferrin. Iron is also used for the production of myoglobin and a wide range of metalloproteins in other tissues. Excess iron is stored in most body cells as a macromolecular complex with the protein ferritin (3). The hepatocytes of the liver and the macrophages of the bone marrow and spleen are particularly well adapted for iron storage and constitute the major physiological iron reservoir in the body.

The primary regulator of body iron concentrations is the circulating peptide hepcidin, which is produced by the liver (4–6). Hepcidin blocks intestinal iron absorption by binding to the iron exporter, ferroportin, on the basolateral membrane of enterocytes and facilitating its internalization and degradation (7). Hepcidin also suppresses iron release from hepatocytes and reticuloendothelial macrophages by the same mechanism. Stimuli that decrease the expression of HAMP, the gene encoding hepcidin, include iron deficiency, anemia, administration of erythropoietin, and enhanced erythropoiesis (8–11). Such stimuli allow increased iron absorption and increased iron flux to the erythroid marrow. Hypoxia has also been shown to decrease hepcidin both in vivo and in vitro (8), but when erythropoietic activity is impaired, hepcidin is no longer suppressed even during severe anemia (10, 11).

It has been known for many years that gastric acid improves the absorption of several important nutrients, particularly iron, calcium, and vitamin B₁₂ (12–15). Gastric acid is essential for the absorption of dietary nonheme iron (15, 16) because it keeps the pH of the stomach low, thus allowing iron to remain soluble and available for reduction to the ferrous state, which is the form in which the majority of iron crosses the intestinal epithelium (17–19). The gastrointestinal hormone gastrin (Gamide) is responsible for stimulating the secretion of gastric acid (20). Gamide is produced from preprogastrin through a number of proteolytic cleavages and a final amidation step. Processing intermediates of progastrin such as glycine-extended gastrin (Ggly) are also biologically active in stimulating growth of the colon and accelerating the development of colorectal cancer (21, 22) but are not directly involved in acid secretion (23). Ggly circulates at similar concentrations to Gamide.

We have shown previously that both Gamide and Gly bind two ferric ions in vitro (24). Iron binding is essential for the biological activity of Ggly in vitro (25) and in vivo (26) but is not required for Gamide activity (27). Both Gamide and Ggly interact with apotransferrin in vitro (28, 29), and a positive correlation between circulating gastrin concentrations and transferrin saturation has also been identified in vivo in young mice and in hypergastrinemic patients (30). These observations formed the basis for our proposal that circulating gastrins may act as chaperones for the uptake of ferric ions by apotransferrin (30). The requirement for gastrin-stimulated acid secretion for efficient iron absorption also points to an important relationship between gastrin and iron homeostasis. In this paper we further explore the role of gastrins in iron homeostasis and, in particular, investigate the effects of long-term dietary iron deficiency in mice with altered circulating gastrin concentrations. We conclude that gastrins appear to play an important role in hematopoiesis under conditions of iron deficiency.

Materials and Methods

Dietary iron modification in mice

Balb/c mice carrying a deletion in the gastrin gene [Gas^−/− (31)] were compared with Balb/c wild-type mice. These mice have reduced gastric acidity because of the absence of gastrin (31). C57BL/6 mice carrying a deletion in the CCK2 receptor gene (Cck2r^−/−) have reduced gastric acidity, and the reduction in gastric acid secretion in Cck2r^−/− mice results in a compensatory increase in circulating gastrins (32, 33). Cck2r^−/− mice were compared with C57BL/6 wild-type mice. The Balb/c Gas^−/− mice were also back-crossed for 12 generations into the C57BL/6 strain to permit direct comparison with Cck2r^−/− mice for some of the critical experiments. Mice (3 wk of age) were weaned onto iron-modified diets. All mice were maintained under standard laboratory conditions on a 12-h light, 12-h dark cycle with free access to deionized water and food containing normal, low, or high iron (160, 3, or 6000 mg of iron per kg wet weight, respectively) (34, 35) for 6 wk. As part of the phenotypic assessment, mice were monitored daily, and their health was recorded using an animal health scoring system (36) approved by the Austin Health Animal Ethics Committee.

Hematological and histological analysis

Mice were anesthetized by ketamine/xylazine injection [100 mg/kg ketamine (Pfizer, Auckland, New Zealand) and 10 mg/kg xylazine (Troy Laboratories, Smithfield, Australia)], and blood was extracted by cardiac puncture. Fresh blood (250 μl) was placed in tubes containing EDTA as an anticoagulant. Erythrocyte parameters (red blood cell count, hemoglobin concentration, hematocrit, and mean cell volume), platelet count, and white cell differential count were measured using an Advia 120 automated hematological analyzer (Bayer, Tarrytown, NY). Hematoxylin and eosin (H&E) staining and Masson's trichrome staining on fixed and paraffin-embedded spleens and decalcified femurs was performed according to standard techniques. Megakaryocyte counts of bone marrow and spleen were performed by manual counting after staining with antibody against Von Willebrand factor (Dako, Glostrup, Denmark) at a 1:200 dilution. Twenty fields from femur (×20) and spleen (×20) were scored.

Iron status analysis

The rest of the collected fresh blood (∼700 μl) was used to extract serum. Serum ferritin and total iron-binding capacity were quantitated by the Division of Laboratory Medicine at Austin Health (Melbourne, Australia), using the Tina-quant immunoturbidimetric assay on a Hitachi 917 automatic analyzer (Roche Diagnostics, Castle Hill, Australia). Transferrin saturation was calculated as [(serum iron)/ total iron-binding capacity] × 100. Liver nonheme iron was quantitated using a previously described colorimetric method (37).

Hamp, ferroportin-1, and Dmt-1 mRNA quantitation

Total RNA was isolated from approximately 100 mg of frozen liver or duodenum of all mice using TRIzol reagent (Invitrogen, Melbourne, Australia) according to the manufacturer's instructions. Total RNA (5 μg) was used for cDNA synthesis with the Superscript II First Strand Synthesis system (Invitrogen). The resulting cDNA transcripts of hepatic or duodenal mRNA were used for real-time PCR amplification using the ABI PRISM 7700 Sequence Detector (Applied Biosystems, Melbourne, Australia) and Taqman chemistry. The following primers were used: Hamp forward, 5′-AGCACCACCTATCTCCATCAACA-3′; Hamp reverse, 5′-GCTTTCTTCCCCGTGCAA-3′; Hamp MGB probe, 5′-CTGCAGCTCTGTAGTCT-3′; Dmt-1 forward, 5′-GCGGCCAGTGATGAGTGAGT-3′; Dmt-1 reverse, 5′-ATGCCACCGGCAATCCT-3′; Dmt-1 MGB probe, 5′-CAGCCTATTCCATTGGAA-3′. Total Hamp mRNA was measured in these studies. Fpn-1 mRNA expression was quantitated using the mouse-specific ferroportin gene expression assay purchased from Applied Biosystems. Gene expression was quantitated relative to 18S RNA expression.

Ferroportin-1 and Dmt-1 protein expression

Protein was isolated from duodenal epithelial cells of all mice using the triple-detergent buffer (50 mm Tris HCl, pH 8; 150 mm NaCl; 0.1% sodium dodecyl sulfate; 1% Nonidet P40; 0.5% sodium deoxycholate) on ice for 30 min and then centrifuged for 5 min at the maximum speed at 4 C. Supernatants were collected and protein concentrations determined using the Bradford reagent (Sigma, Melbourne, Australia). Protein (20 μg) was subjected to 10% SDS-PAGE. The proteins were transferred onto Hybond-C-Extra nitrocellulose (Amersham Biosciences, Piscataway, NJ), blocked for 1 h in Tris-buffered saline containing 0.05% Tween 20 and 5% fat-free milk. Next, the nitrocellulose was incubated overnight at 4 C with rabbit polyclonal antibodies directed against either Dmt-1 or Ferroportin-1 both at 1:5000 dilution. After removal of the primary antibody, a secondary anti-IgG antibody (horseradish peroxidase labeled) was used, and cross reactivity was visualized using ECL reagent (Amersham Pharmacia Biotech, Piscataway, NJ) according to the manufacturer's instructions. Band densities were quantified by densitometric analysis using Multigage software (Fujifilm Medical Systems, Stamford, CT). Glyceraldehyde-3-phosphate dehydrogenase concentrations were used to demonstrate equal protein loading onto gels.

Serum erythropoietin quantitation

Erythropoietin (EPO) and thrombopoietin (TPO) in the serum were quantitated using an ELISA assay (R&D Systems, Minneapolis, MN) according to the manufacturer's recommendations.

Statistical analysis

Statistical significance for single comparisons of normally distributed data was determined by Student's t test or for data that were not normally distributed by Mann-Whitney rank sum test. For multiple comparisons Student's t test was used followed by the Bonferroni correction. For all experiments the mouse group size was six to 10. All statistics were analyzed with the statistical program SigmaStat and graphed using SigmaPlot (Jandel Scientific, San Rafael, CA).

Results

Gastrin-deficient mice on an iron-deficient diet develop gross splenomegaly

Although phenotypically normal when fed a normal-iron diet, the health of the Gas^−/− mice deteriorated during 6 wk on a low-iron diet. The animal sickness score, which included assessment of body weight, appearance, and behavior, was 11 ± 1 for the Gas^−/− mice on the low-iron diet, compared with a score of 6 ± 1 for the wild-type mice on the same diet. Furthermore, the observation that the Gas^−/− animals on the low-iron diet developed splenomegaly (Fig. 1) suggested that extramedullar hematopoiesis was likely increased. The spleen weights of the Gas^−/− mice (on a Balb/c background) fed the low-iron diet were significantly greater (average 367 mg) compared with the wild-type mice (114 mg). No significant difference in spleen size was observed between the Gas^−/− (125 mg) and wild-type mice (100 mg) on the normal-iron diet (Fig. 1A). To ensure that the development of splenomegaly in response to iron deficiency was not strain specific, Gas^−/− mice on a C57BL/6 background were also weaned onto the low-iron diet with a similar result (Fig. 1B). In this case, the spleen to body weight ratio was 6-fold greater (P < 0.01) in the Gas^−/− mice on the low-iron diet compared with wild-type C57BL/6 mice on the same diet (Fig. 1C), but the magnitude of the difference was enhanced by the relative splenic atrophy observed in the wild-type mice (Fig. 1B). Intriguingly, Cck2r^−/− mice, which have the same hypochlorhydric phenotype as Gas^−/− mice but have elevated concentrations of circulating gastrins, did not sicken or develop splenic abnormalities when fed a low-iron diet (Fig. 1, B and C).

Fig. 1.

On a low-iron diet, Gas^−/− mice, but not Cck2r^−/− mice, present with splenomegaly. A, On the Balb/c background, spleens from Gas^−/− mice on a low-iron diet were 2.8-fold larger and significantly paler than spleens from wild-type mice. No difference in spleen size was observed between Gas^−/− and wild-type Balb/c mice on the normal-iron diet. B, On the C57BL/6 background Gas^−/− mice on the low-iron diet also had enlarged spleens compared with the wild-type mice. The spleen size in Cck2r^−/− mice did not change with iron deficiency compared with wild-type mice. Spleens were significantly smaller in wild-type C57BL/6 mice on a low-iron diet than on a normal-iron diet. C, The spleen to body weight ratio of all mice is shown on both normal-iron diet (gray coarse hatched bars) and low-iron diet (black coarse hatched bars). The spleen to body weight ratio of Gas^−/− mice on either the Balb/c or C57BL/6 background on a low-iron diet was increased compared with the corresponding wild-type mice on the same diet. Data are means ± sem, where n = 6 (three males and three females); *, P < 0.05; **, P < 0.01. WT, Wild type.

To investigate whether any gross morphological defects were responsible for the significant enlargement of the spleens in the Gas^−/− mice on the low-iron diet, spleen histology was analyzed by H&E (Fig. 2) and Masson's trichrome stains (Supplemental Fig. 1 published on The Endocrine Society's Journals Online web site at http://endo.endojournals.org). On the normal-iron diet, H&E-stained sections were normal in Gas^−/− mice on both backgrounds but, in the same mice on a low-iron diet, the subcapsular region of the spleen was grossly enlarged with expanded red pulp and fewer lymphocyte-containing follicles. In contrast, the histology of the spleen was normal in wild-type mice on the low-iron diet with a clear separation of white pulp from red pulp. The Masson's trichrome stain highlighted the presence of increased amounts of collagen within the splenic capsule and throughout the spleen in the Gas^−/− mice (on both backgrounds) on the low-iron diet (see Supplemental Fig. 1 for representative Balb/c Gas^−/− data). Although the spleens of wild-type C57BL/6 mice atrophied with iron deficiency, the white pulp was clearly separated from red pulp (Fig. 2). Furthermore, no difference in spleen morphology was found in the Cck2r^−/− mice on either normal or iron-deficient diets compared with the wild-type C57BL/6 mice.

Fig. 2.

Gas^−/− mice on a low-iron diet have abnormal splenic architecture. On a low-iron diet Gas^−/− mice on both backgrounds have abnormal splenic architecture as shown by H&E stain. The spleen appeared normal in Balb/c Gas^−/− mice and C57BL/6 Gas^−/− mice on a normal-iron diet, or in wild-type Balb/c and C57BL/6 mice on either diet. The Cck2r^−/− mice on both the normal- and low-iron diets have splenic architecture similar to the wild-type C57BL/6 mice on the normal-iron diet. A picture from a representative section from the spleen from a mouse from each group is shown at ×4 magnification. WT, Wild type.

Hematopoiesis in mice with altered circulating gastrin concentrations

Gas^−/− mice on both strain backgrounds developed severe anemia after 6 wk on the low-iron diet. Hematocrit, mean corpuscular hemoglobin, and mean corpuscular volume decreased to a much greater degree in the Gas^−/− mice on either the Balb/c (Table 1) or C57BL/6 (Table 2) background on the low-iron diet compared with the wild-type mice on the same diet. The apparent difference in red blood cell count between the Gas^−/− and wild-type Balb/c mice on the low-iron diet was not statistically significant but was significant in the Gas^−/− mice on the C57BL/6 background (Table 2). Interestingly, the Cck2r^−/− mice on the low-iron diet had milder anemia and a hematopoietic profile more similar to the wild-type C57BL/6 mice than the Gas^−/− mice (Table 2).

Table 1.

Hematological profile of Gas^−/− mice on the Balb/c background and wild-type Balb/c mice on normal- or low-iron diets

Peripheral blood	Normal-iron diet		Low-iron diet
	Wild-type	Gas−/−	Wild-type	Gas−/−
RBC (106/μl)	10.4 ± 0.3	10.3 ± 0.2	10.5 ± 0.4	8.0 ± 1.1
HCT (%)	49.7 ± 1.7	48.5 ± 1.2	45.2 ± 2.5	33.0 ± 4.9#
MCH (pg)	16.0 ± 0.1	15.5 ± 0.1	13.0 ± 0.6**	10.6 ± 0.5***,##
MCV (fl)	47.9 ± 0.4	47.0 ± 0.3	43.0 ± 0.5***	40.0 ± 0.8***,#
WBC (103/μl)	3.4 ± 0.4	1.7 ± 0.6	3.7 ± 0.7	2.1 ± 0.4
Neutrophils	0.48 ± 0.09	0.37 ± 0.07	0.70 ± 0.20	0.28 ± 0.07
Lymphocytes	2.67 ± 0.45	1.18 ± 0.47	2.60 ± 0.62	1.97 ± 0.75
Monocytes	0.09 ± 0.04	0.04 ± 0.01	0.09 ± 0.01	0.02 ± 0.00###
Eosinophils	0.06 ± 0.01	0.06 ± 0.03	0.08 ± 0.01	0.01 ± 0.01##

Means ± sem where n = 6; WBC, white blood cells; RBC, red blood cells; HCT, hematocrit; MCH, mean corpuscular hemoglobin; MCV, mean corpuscular volume.

Comparisons of the effect of a low iron diet on Gas^−/− and wild-type mice: **, P < 0.01;

^***, P < 0.001.

Comparisons of the effect of gastrin deficiency in mice on the low-iron diet: ^#, P < 0.05;

^##, P < 0.01;

^###, P < 0.001.

Table 2.

Hematological profile of Gas^−/−, Cck2r^−/− on the C57BL/6 background and wild-type C57BL/6 mice on normal- or low-iron diets

Peripheral blood	Normal iron diet			Low iron diet
	C57BL/6 WT	Gas−/−	Cck2r−/−	C57BL/6 WT	Gas−/−	Cck2r−/−
RBC (106/μl)	9.2 ± 0.7	10.2 ± 0.2	10.1 ± 0.1	10.2 ± 0.3	6.3 ± 1.1**,#	9.2 ± 0.4
HCT (%)	46.9 ± 3.2	49.7 ± 1.2	48.5 ± 0.8	44.2 ± 0.2	23.6 ± 5.1**,#	38.3 ± 2.1***
MCH (pg)	15.9 ± 0.8	14.9 ± 0.1	15.4 ± 0.1	12.8 ± 0.2	10.1 ± 0.6***,#	11.5 ± 0.3***
MCV (fl)	49.2 ± 1.0	48.5 ± 0.4	48.1 ± 0.4	42.6 ± 1.1***	36.5 ± 1.6**,#	41.4 ± 1.3*
WBC (103/μl)	5.8 ± 1.1	6.1 ± 0.6	5.9 ± 0.9	1.3 ± 0.1**	3.3 ± 0.3**,##	5.3 ± 0.5*,###
Neutrophils	0.54 ± 0.09	0.63 ± 0.08	0.63 ± 0.1	0.43 ± 0.02	0.49 ± 0.07	0.45 ± 0.05
Lymphocytes	4.5 ± 0.9	4.7 ± 0.5	4.5 ± 0.7	0.56 ± 0.04***	2.4 ± 0.3**	4.3 ± 0.5###
Monocytes	0.17 ± 0.02	0.19 ± 0.03	0.16 ± 0.03	0.06 ± 0.02*	0.09 ± 0.01*	0.13 ± 0.02
Eosinophils	0.23 ± 0.07	0.22 ± 0.06	0.14 ± 0.04	0.09 ± 0.06	0.11 ± 0.06	0.21 ± 0.12

Means ± sem where n = 6; WBC, white blood cells; RBC, red blood cells; HCT, hematocrit; MCH, mean corpuscular hemoglobin; MCV, mean corpuscular volume; WT, wild type.

Comparisons of the effect of a low-iron diet on Gas^−/−, Cck2r^−/− and wild-type mice: *, P < 0.05;

^**, P < 0.01;

^***, P < 0.001.

Comparisons of the effect of gastrin deficiency or hypergastrinemia in mice on the low-iron diet: ^#, P < 0.05;

^##, P < 0.01;

^###, P < 0.001.

Because erythropoietin (EPO), the main cytokine involved in erythropoiesis, is up-regulated in iron deficiency to stimulate erythropoiesis, the concentrations of EPO in the serum were measured (Fig. 3A). EPO concentrations were significantly increased in both Gas^−/− (267-fold) and wild-type (17-fold) Balb/c mice on the low-iron diet compared with the same strains on the normal-iron diet. A significant increase in EPO concentrations (16-fold) was observed in the Gas^−/− mice compared with the wild-type Balb/c mice on the low-iron diet (Fig. 3A). Thus both the hematological parameters and the EPO concentration demonstrated that the Gas^−/− mice on a low-iron diet develop anemia that was much more severe than in wild-type mice on the same diet. No differences were observed in the erythropoietic parameters between Gas^−/− and wild-type Balb/c mice on the normal-iron diet (Table 1), so it was not surprising that EPO concentrations were also similar (Fig. 3A). Because TPO and EPO belong to the same hematopoietic growth factor subfamily, the concentrations of TPO in the circulation were also determined (Fig. 3B). However, although a difference was noted in serum TPO concentration between the Gas^−/− mice and wild-type Balb/c mice fed the normal-iron diet, the TPO concentrations in Gas^−/− and wild-type mice on both the normal or the low-iron diets were not significantly different (Fig. 3B).

Fig. 3.

Gas^−/− mice on a low-iron diet have significantly increased erythropoietin but not thrombopoietin concentrations. A, In Gas^−/− mice (gray bars) on a low-iron diet, circulating EPO concentrations were significantly increased compared with wild-type Balb/c mice (white bars) on the same diet. No difference in EPO concentrations was observed between the Gas^−/− mice on a normal-iron diet and the wild-type Balb/c mice on the same diet. B, The concentration of TPO was significantly increased in the Gas^−/− mice compared with the wild-type Balb/c mice on the normal-iron diet. No difference in TPO concentration was observed between the Gas^−/− and wild-type Balb/c mice on the low-iron diet, or between either strain on the low-iron diet and the corresponding strain on the normal-iron diet. Data are means ± sem, where n = 6 (three males and three females); **, P < 0.01.

To determine whether the high EPO concentration had an effect on other hematopoietic cell lineages, white blood cells (WBC) were counted. Although the total WBC numbers were not significantly different between the Balb/c Gas^−/− mice and their wild-type controls on either iron diet, the differential counts revealed a significant decrease in monocytes and in eosinophils on the low-iron diet in the Balb/c Gas^−/− mice compared with wild-type mice (Table 1). The WBC counts did not reveal any significant differences between the C57BL/6 Gas^−/− and C57BL/6 Cck2r^−/− mice and their wild-type controls on the normal-iron diet (Table 2). However, the WBC count including lymphocytes and monocytes (Table 2) in the wild-type C57BL/6 mice on the low-iron diet was significantly lower than on the normal-iron diet, presumably because of the splenic atrophy (Figs. 1 and 2). Although the WBC count was also significantly decreased in both Gas^−/− and Cck2r^−/− mice on the low-iron diet compared with the normal-iron diet, the WBC numbers in both strains were significantly higher than in the wild-type mice on the same diet (Table 2). The lymphocyte and monocyte numbers did not change in the Cck2r^−/− mice on the low-iron diet although both were decreased in the Gas^−/− mice on the low-iron diet, compared with the same strains on the normal-iron diet (Table 2).

The number of megakaryocytes in the bone marrow and spleen of the Gas^−/− and wild-type Balb/c mice was also measured by immunohistochemistry with an antibody targeted against Von Willebrand factor (Supplemental Fig. 2). No difference was observed in the numbers of bone marrow megakaryocytes in the Gas^−/− and wild-type Balb/c mice on either normal or low-iron diets (Fig. 4A). Although numbers of splenic megakaryocytes were also similar in the Gas^−/− and wild-type Balb/c mice on the normal-iron diet, on the low-iron diet the number of megakaryocytes was significantly greater in the Gas^−/− mice than in the wild-type Balb/c mice (Supplemental Fig. 2 and Fig. 4B). Because megakaryocytes are platelet precursors, the numbers of circulating platelets were also determined in the Gas^−/− and wild-type Balb/c mice. As expected, the high number of megakaryocytes in the Gas^−/− mice on the low-iron diet led to thrombocytosis (Fig. 4C). Taken together these data indicate significant abnormalities in hematopoiesis in iron-deficient Gas^−/− mice.

Fig. 4.

Gas^−/− mice on a low-iron diet show megakaryopoiesis in the spleen and thrombocytosis. A, Megakaryocyte numbers were similar in the bone marrow of the Gas^−/− mice (gray bars) and the wild-type Balb/c mice (white bars) on both normal- and low-iron diets. B, Megakaryocyte numbers in the spleen were significantly increased only in the Gas^−/− mice on the low-iron diet compared with the same mice on the normal-iron diet. C, Platelet number in the circulation was increased in the Gas^−/− mice on the low-iron diet compared with the wild-type Balb/c mice. No difference in circulating platelets was observed between the Gas^−/− and the wild-type Balb/c mice on the normal-iron diet. Data are means ± sem, where n = 6 (three males and three females); *, P < 0.05.

Iron status in mice with altered circulating gastrin concentrations

To determine the effects of the iron-modified diet on iron homeostasis in mice with altered circulating gastrin concentrations, iron status was assessed at the end of the 6-wk study by measurement of serum ferritin and hepatic iron concentrations. In all mice, the low-iron diet led to significant iron deficiency, and the high-iron diet led to iron overload as evident from the observed changes in serum ferritin, hepatic iron, and transferrin saturation (Fig. 5). Although the concentration of serum ferritin in the Cck2r^−/− mice on the low-iron diet was significantly lower than in wild-type mice undergoing the same treatment, no other significantly different responses to changes in dietary iron were found between the values for serum ferritin, hepatic iron, and transferrin saturation in the mice with altered circulating gastrin concentrations and their respective wild-type controls. These observations suggest that Gas^−/− and Cck2r^−/− mice can still obtain sufficient iron from the diet despite the reduction in gastric acid secretion in both strains.

Fig. 5.

Iron status of Gas^−/− mice and Cck2r ^−/− mice on diets containing normal, low, or high iron. A, Serum ferritin levels were significantly decreased in the Cck2r^−/− mice (hatched bars) on the low-iron diet compared with C57BL/6 wild-type mice (black bars) on the low-iron diet. No difference in serum ferritin levels was observed between the Cck2r^−/− mice and the C57BL/6 wild-type mice on the normal- and high-iron diets. In the Gas^−/− mice (gray bars) and the Balb/c wild-type mice (white bars) on the low- and high-iron diet, serum ferritin levels were significantly decreased and increased, respectively. However, there was no difference between the Gas^−/− and Balb/c wild-type mice. B, Hepatic iron was significantly decreased in all mice on the low-iron diet and increased in all mice on the high-iron diet. C, Transferrin saturation was significantly decreased in all mice on the low-iron diet and increased in all mice on the high-iron diet. Data are means ± sem, where n = 10 (five males and five females); *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Hamp mRNA expression in mice with altered circulating gastrin concentrations

Because we have previously identified a link between gastrin and iron status (30), the regulation of iron homeostasis was examined by measuring the expression of Hamp mRNA in mice with altered circulating gastrin concentrations on normal-, low-, and high-iron diets. A significant decrease in Hamp mRNA was observed in both Cck2r^−/− (169-fold) and wild-type mice (12-fold) on the low-iron diet compared with the same strains on the normal-iron diet (Fig. 6A). Hamp mRNA expression was also significantly decreased in both Gas^−/− and wild-type mice on the low-iron diet. Conversely, Hamp mRNA expression was significantly increased in Cck2r^−/−, Gas^−/−, and wild-type mice on the high-iron diet (Fig. 6A). Concentrations of hepcidin were not measured because of the lack of commercially available antibodies suitable for Western blotting.

Fig. 6.

Hamp mRNA expression is decreased in Cck2r ^−/− mice and increased in Gas^−/− mice on a low-iron diet, and Fpn-1 and Dmt-1 mRNA expression adjusts accordingly. A, Hamp mRNA expression is decreased in all mice on a low-iron diet and increased in all mice on a high-iron diet (Cck2r^−/− mice (hatched bars), C57BL/6 wild-type mice (black bars), Gas^−/− mice (gray bars), Balb/c wild-type mice (white bars). Hamp mRNA expression was decreased in Cck2r^−/− mice on a low-iron diet compared with wild-type mice on the same diet, and significantly increased in the Gas^−/− mice on a low-iron diet compared with wild-type mice on the same diet. B, Fpn-1 mRNA expression was increased in all mice on a low-iron diet and decreased in all mice on a high-iron diet. C, Dmt-1 mRNA expression was increased in all mice on a low-iron diet and decreased in all mice on a high-iron diet. The differences in Hamp mRNA expression were reflected in Dmt-1 and Fpn-1 expression. Data are means ± sem, where n = 10 (five males and five females); *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Significant differences in Hamp mRNA expression were observed between Cck2r^−/− and Gas^−/− mice and their respective wild-type strains on the low-iron diet. Thus Hamp mRNA expression in mice on the low-iron diet was 18-fold lower in the hypergastrinemic Cck2r^−/− strain compared with the wild-type strain. Conversely, Hamp mRNA expression was 43-fold higher in the Gas^−/− mice on the low-iron diet compared with the wild-type mice on the same diet. This observation suggests that gastrins modulate the response of hepcidin to iron deficiency.

Duodenal Dmt-1 and Fpn-1 expression in mice with altered circulating gastrin concentrations

To determine whether the observed changes in Hamp mRNA were associated with alterations in the expression of other iron-related genes, Fpn-1 expression was measured. The expression of Fpn-1 mRNA showed an inverse trend to the expression of Hamp mRNA in all mice, because it was increased with iron deficiency and decreased with iron overload (Fig. 6B). Furthermore, in the iron-deficient Cck2r^−/− mice the concentration of Fpn-1 mRNA was 1.9-fold lower than in the wild-type mice on the same diet.

The concentrations of Fpn-1 protein determined by Western blotting showed similar trends to the mRNA expression, although the magnitude of the differences was not as great. On the normal-iron diet no difference in Fpn-1 concentration [expressed as a percentage of glyceraldehyde-3-phosphate dehydrogenase (GAPDH)] was observed between the wild-type and Cck2r^−/− mice (78 ± 11% and 69 ± 5%, respectively), whereas on the low-iron diet Cck2r^−/− mice had a significantly decreased Fpn-1 concentration (93 ± 2%) compared with wild-type mice (112 ± 2%). In the Gas^−/− mice on the normal-iron diet no significant difference in Fpn-1 concentration was observed between the wild-type and Gas^−/− mice (80 ± 5% and 93 ± 5%, respectively), whereas on the low-iron diet Fpn-1 protein concentrations were increased in both Gas^−/− mice (147 ± 11%) and wild-type mice (126 ± 10%), although no significant difference was detected between them. Insufficient tissue was available from the mice on the high-iron diet to assess Fpn-1 concentrations, but based on mRNA quantitation the values should be similar to mice on normal diets.

Changes in the expression of Hamp mRNA led to inverse changes in the expression of Dmt-1 mRNA secondary to alterations in the expression of the iron exporter ferroportin by hepcidin. Dmt-1 mRNA expression was significantly increased in all mice fed the low-iron diet but significantly decreased only in the Cck2r^−/− mice on the high-iron diet (Fig. 6C). Furthermore, a significant increase in Dmt-1 mRNA expression was observed in Cck2r^−/− mice on the low-iron diet compared with wild-type animals on the same diet (Fig. 6C). Conversely only a 10-fold increase in Dmt-1 mRNA was observed in the Gas^−/− mice fed the low-iron diet compared with a 25-fold increase in the corresponding wild-type mice fed the same diet, although the difference was not statistically significant.

Dmt-1 protein concentrations showed similar trends to Dmt-1 mRNA expression, although the magnitude of the differences was not as great. On the normal-iron diet no significant difference in Dmt-1 protein concentration was observed between the wild-type and Cck2r^−/− mice (95 ± 1% and 112 ± 5%, respectively), whereas on the low-iron diet Cck2r^−/− mice had significantly increased Dmt-1 protein concentrations (130 ± 2%) compared with wild-type mice (111 ± 2%). In the Gas^−/− mice on the normal-iron diet no significant difference in Dmt-1 protein concentration was observed between the wild-type and Gas^−/− mice (62 ± 3% and 73 ± 8%, respectively) whereas on the low-iron diet Gas^−/− mice had decreased Dmt-1 protein concentrations (95 ± 8%) compared with wild-type mice (120 ± 3%).

The differences in the magnitude of the changes in Dmt-1 and Fpn-1 mRNA mirror the observed differences in Hamp mRNA concentrations not only between mice of all strains on the different diets but also, and more importantly, between either Cck2r^−/− or Gas^−/− mice and their respective wild-type strains on the low-iron diet. This observation strengthens the conclusion that gastrins modulate the response to iron deficiency.

Discussion

We have previously reported a positive correlation between circulating gastrin concentrations and transferrin saturation in young mice (30). This observation suggests that gastrins can modulate transferrin saturation in the first few weeks of life and thus potentially contribute to an increased availability of iron to cells within the body. In this study we have explored the gastrin-iron link further by examining how mice with altered gastrin concentrations respond to diets of differing iron content.

When mice with altered circulating gastrin concentrations were placed on a low-iron diet for 6 wk, the health of the Gas^−/− mice deteriorated significantly when compared with the corresponding wild-type mice. In wild-type animals, erythropoiesis is impaired as iron supply diminishes, resulting in anemia. On further examination, the Gas^−/− mice (on both Balb/c and C57BL/6 backgrounds) on the low-iron diet were also found to develop severe anemia (although their iron stores were the same as in the wild-type mice) as shown by reduction in hemoglobin, hematocrit, and mean corpuscular volume. Interestingly, the absence of any overt health problems and the lack of severe anemia in the Cck2r^−/− mice on the low-iron diet suggest that the contribution of gastrins to erythropoiesis in wild-type mice is not mediated via the CCK2 receptor.

On the low-iron diet the Gas^−/− mice, but not the Cck2r^−/− mice, also developed splenomegaly with altered splenic architecture containing hypertrophied red pulp (normally responsible for degradation of senescent erythrocytes) and atrophied white pulp (lymphatic tissue), likely a result of extramedullary hematopoiesis. Similar changes in splenic architecture have been described previously in mice transgenically modified to express high amounts of EPO (38). Furthermore, the enlarged spleens of the Gas^−/− mice contained high numbers of megakaryocytes, and this was accompanied by thrombocytosis. Iron depletion itself does not appear to be the cause, because wild-type mice on the low-iron diet exhibited similar reductions in serum ferritin, hepatic iron and transferrin saturation to their Gas^−/− counterparts without an altered platelet count. Although concentrations of circulating TPO, the major megakaryocytopoietic cytokine, were not significantly altered, previous studies in both mouse and humans have established that EPO can stimulate megakaryocyte formation in vitro and elevate megakaryocyte and platelet counts in vivo (39–42). Thus, the dramatic elevation in circulating EPO concentration observed specifically in Gas^−/− mice is a likely explanation for some of the changes in splenic architecture and for the increased platelet count.

The data presented here also reveal significant changes in the numbers of WBC in the Gas^−/− mice on an iron-deficient diet. Thus, there was a significant decrease in monocytes and eosinophils (Balb/c background), or in lymphocytes and monocytes (C57BL/6 background), between the Gas^−/− and wild-type mice on the iron-deficient diet. The evidence available in the literature for effects of gastrins on leukocytes is conflicting. Gastrin negatively regulates the mobility of neutrophils, macrophages, and lymphocytes in vitro (43, 44) and induces leukocyte-endothelial interactions and contributes to the inflammation caused by H. pylori in rats (45). CCK receptors have been detected on rat splenic monocytes and on a subset of human peripheral monocytes and shown to mediate CCK-dependent chemotaxis (46). The presence of CCK2R mRNA has been variously reported to be restricted to polymorphonuclear, but not mononuclear, leukocytes from normal human peripheral blood (47), or to peripheral lymphocytes (48). Finally many human leukemia cell lines express the CCK2R mRNA and respond to exogenous gastrin (47). Clearly more detailed study is warranted of the involvement in leukocyte development of nonamidated gastrins acting via receptors distinct from the CCK2 receptor (49, 50).

Dramatic differences in the ability of the Gas^−/− and Cck2r^−/− mice to regulate iron homeostasis were revealed in response to iron deficiency. Although in general, Hamp mRNA expression was decreased in all mice on the low-iron diet and increased in all mice on the high-iron diet (Fig. 6A), the Gas^−/− and Cck2r^−/− mice on the low-iron diet did not behave in the same way as their wild-type counterparts. Thus, Hamp mRNA expression in the hypergastrinemic Cck2r^−/− mice on the low-iron diet was significantly reduced compared with the wild-type animals on the same diet, to a point where Hamp mRNA was almost undetectable (Fig. 6A). Hamp mRNA also decreased significantly in Gas^−/− mice on the low-iron diet, but the observation that this decrease was significantly less than that seen in wild-type mice (Fig. 6A) suggests a possible misregulation of hepcidin expression under iron-deficient conditions.

In iron-deficient wild-type animals, Hamp expression is down-regulated to enhance iron release from cells to provide erythrocytes with sufficient iron. Erythropoetic activity is required for the down-regulation of Hamp-1 expression by anemia (10, 11), and this down-regulation can occur independently of the effects of hypoxia and EPO. The impaired erythropoiesis observed in the Gas^−/− mice thus may result in relatively high Hamp mRNA even in the presence of very high circulating EPO concentrations as there is reduced demand for iron for erythropoiesis. Subsequently, the inappropriately high Hamp expression in the Gas^−/− mice on the low-iron diet can lead to the sequestering of iron within various cell types via the down-regulation of ferroportin. The reduced flow of iron from storage cells and the reduced iron absorption could further impair erythropoiesis. In contrast, in the wild-type mice on the low-iron diet, Hamp is completely down-regulated, duodenal expression of Dmt-1 mRNA is increased, and thus dietary iron uptake and iron flow from storage cells should be maximal and sufficient to provide developing erythrocytes with enough iron to avoid severe anemia.

Despite the differences in hepcidin expression, our data suggest that, provided dietary iron supply is adequate, the regulation of iron homeostasis in mature mice with altered circulating gastrin concentrations is normal, in agreement with our previous observations of these mice on iron-replete diets (30). The similarities in liver iron content suggested that both Gas^−/− and Cck2r^−/− mice were of comparable iron status to their respective wild-type controls, even though both strains have reduced acid output (31, 33). Presumably the mice with altered circulating gastrin concentrations can still obtain sufficient iron from the diet despite the increased luminal pH. A similar situation has been reported previously in human patients with suppressed acid secretion after long-term treatment with the proton pump inhibitor omeprazole. Such patients did not develop iron deficiency unless they had underlying anemia (19). Unlike iron homeostasis, calcium homeostasis is significantly affected by a reduction in acid secretion as Cck2r^−/− mice develop mild hypocalcemia, and the observation of secondary hyperparathyroidism and increased bone resorption suggests that only mobilization of calcium from bone prevents a more severe hypocalcemia in these animals (51). Furthermore, the osteoporotic phenotype in Cck2r^−/− mice is completely rescued by dietary supplementation with calcium gluconate, but only partially reversed with calcium carbonate, which is less soluble than calcium gluconate at neutral pH (51).

A further possible explanation for the anemia and for the changes in splenic architecture in Gas^−/− mice on the low-iron diet involves the interaction between gastrins and serum transferrin. We have previously demonstrated with surface plasmon resonance and cross-linking assays that both Gamide and Ggly interact with iron-free transferrin but not with iron-loaded transferrin and that the interaction is dependent on binding of iron to the gastrin peptides (28, 29). We have proposed, on the basis of differences in transferrin saturation between 4-wk-old Gas^−/− mice, Cck2r^−/− mice, and their respective wild-type strains, that circulating gastrins may act as chaperones for the uptake of ferric ions by apotransferrin (30). The effect appears to be transitory, because no differences in transferrin saturation were observed between 10-wk-old animals of the above strains (30). The absence of significant differences in transferrin saturation between Gas^−/− mice, Cck2r^−/− mice, and their respective wild-type strains after feeding low- or high-iron diets for 6 wk after weaning (i.e. 9 wk after birth) (Fig. 5) is consistent with our previous data. Further experiments with dietary modification of shorter duration will be required to test the hypothesis that changes in the gastrin-transferrin interaction contribute to the observed phenotype.

In summary, impaired erythropoiesis in Gas^−/− mice, but not Cck2r^−/− mice, on a low-iron diet results in severe anemia that is accompanied by splenomegaly. The different responses to a low-iron diet between the Gas^−/− mice and the Cck2r^−/− mice was not the result of strain differences because when the Gas^−/− mice were back-crossed to C57BL/6 mice the inability to regulate iron homeostasis, reduced health, and splenomegaly were still evident. Ultimately, the decrease in erythropoiesis leads to relative blunting of the Hamp response and reduced iron supply to the plasma. The intriguing observations of differences in the responses of Gas^−/− mice and Cck2r^−/− mice to dietary iron deprivation provide evidence for the involvement of gastrin peptides in the regulation of hematopoiesis and suggest an important direct role for gastrins, distinct from their stimulation of acid secretion, in the response to iron deficiency.