Hereditary hemochromatosis (HHC) is one of the most frequent genetic disorders in humans. In healthy individuals, absorption of iron in the intestine is tightly regulated by cells with the highest iron demand, in particular erythroid precursors. Cloning of intestinal iron transporter proteins provided new insight into mechanisms and regulation of intestinal iron absorption. The aim of this study was to assess whether, in humans, the two transporters are regulated in an iron-dependent manner and whether this regulation is disturbed in HHC. Using quantitative PCR, we measured mRNA expression of divalent cation transporter 1 (DCT1), iron-regulated gene 1 (IREG1), and hephaestin in duodenal biopsy samples of individuals with normal iron levels, iron-deficiency anemia, or iron overload. In controls, we found inverse relationships between the DCT1 splice form containing an iron-responsive element (IRE) and blood hemoglobin, serum transferrin saturation, or ferritin. Subjects with iron-deficiency anemia showed a significant increase in expression of the spliced form, DCT1(IRE) mRNA. Similarly, in subjects homozygous for the C282Y HFE mutation, DCT1(IRE) expression levels remained high despite high serum iron saturation. Furthermore, a significantly increased IREG1 expression was observed. Hephaestin did not exhibit a similar iron-dependent regulation. Our data show that expression levels of human DCT1 mRNA, and to a lesser extent IREG1 mRNA, are regulated in an iron-dependent manner, whereas mRNA of hephaestin is not affected. The lack of appropriate downregulation of apical and basolateral iron transporters in duodenum likely leads to excessive iron absorption in persons with HHC.
- iron transporter
- iron metabolism
- divalent cation transporter
- divalent metal transporter
- iron regulated gene 1
hereditary hemochromatosis (HHC) is one of the most prevalent inherited disorders in humans. It is associated with excessive intestinal iron absorption that may lead to damage of various organs, especially liver, pancreas, heart, and pituitary gland. As reported by Feder et al. (13), it is primarily the result of a missense mutation within theHFE gene (mostly cysteine 282 to tyrosine). In HHC, this mutation leads to iron accumulation in the body and may lead to liver cirrhosis or heart disease, and eventually death, if not treated by phlebotomy (4, 12). The HFE protein neither binds nor transports iron but, rather, physically interacts with the transferrin receptor (TfR) (14, 21, 30, 45). Some investigators have proposed (22) that the HFE protein competes with diferric transferrin for the TfR and thereby reduces the overall uptake of iron into cells, maintaining low but sufficient intracellular levels of iron (35, 38). However, it has also been reported that HFE protein can facilitate the uptake of transferrin-bound iron into certain cells (26).
In the mammalian body, iron absorption is tightly regulated by the iron demand for erythropoiesis and occurs only in the small intestine, in particular in the duodenum (18). Iron uptake is a two-step process, as revealed by studies of naturally occurring mutations in animals (3) and the characterization of recently cloned transporter genes (19, 24). At the apices of intestinal villi, enterocytes express divalent cation transporter 1 [DCT1, also known as divalent metal transporter 1 (DMT1) or Nramp2], which is essential for cellular uptake of iron from the intestinal lumen (15, 19). The transport from enterocytes into the serum, at the basolateral side of enterocytes, is accomplished by the coordinated action of the transporter iron-regulated gene 1 [IREG1, also known as metal transporter protein 1 (MTP1) or ferroportin1 (1, 11, 24)] and the multicopper ferroxidase hephaestin (44). The mRNAs of DCT1 and IREG1 contain sequences resembling iron-responsive elements (IREs), previously found in the mRNAs of ferritin and TfR, that presumably regulate their expression by intracellular iron in a cell-specific manner (for review, see Refs. 34 and 42). DCT1, IREG1 mRNA, and IREG1 protein have been reported to be upregulated in response to chronic iron deficiency in the intestine (1, 19, 24,43). Abboud and Haile (1) isolated IREG1 (which they call MTP1) by the ability of its RNA to bind the IRE binding protein-1 (IRP1) immobilized on a column. Likewise, binding of IRP1 to the mRNA of DCT1 has been reported recently (46).
Interestingly, intestinal iron absorption, which is mediated by DCT1 and IREG1, occurs in duodenal villi (1, 8, 24), whereas the “modulator of iron homeostasis,” HFE, is normally expressed in crypt cells (7, 31). How HFE, which is presumed to act as a modulator of the affinity of TfR for transferrin, can regulate iron absorption in villus cells is still not completely understood. Recently, it was shown (28, 33) that the lack of expression of hepcidin, a peptide synthesized in liver, leads to iron overload in mice, similar to HFE or β2-microglobulin knockout mice. Whether this peptide acts as a regulator of genes involved in iron absorption, and how this regulation occurs, remains to be determined.
A model proposed by Conrad et al. in 1964 (10) describes a possible mechanism for regulation of iron absorption mediated by the amount of iron absorbed by crypt cells in the intestine. This body iron “sensing mechanism” in crypt cells might be disturbed in HHC, leading to imbalance in intestinal iron absorption in villus cells (37). The observed iron-dependent regulation of the two iron transporters, DCT1 and IREG/MTP1, raises the question of a possible dysregulation of these transporters in HHC. In two previous studies, carefully selected HHC patients homozygous for C282Y were compared with control subjects, subjects with secondary iron overload, and iron-deficient subjects for their expression level of total DCT1 and IREG1. DCT1 and IREG1 were found to be increased in HHC. However, this study did not differentiate between the two DCT1 splice forms (with and without an IRE) (47, 48), and the expression of hephaestin was not examined.
Recently, two rare forms of HHC not related to mutations in theHFE gene have been described. One of them is caused by a missense mutation in the IREG1 gene (29) and the other one by a defect in the recently identified second TfR isoform, called TfR2 (36). These findings highlight the importance of iron transport in controlling body iron homeostasis.
In the present study, we investigated the expression of the mRNAs of both the apical and basolateral iron transporters in human upper intestinal biopsies in subjects with iron-deficiency anemia, normal iron, and iron overload. We compared the data from patients with homozygous HHC, subjects with and without the major (C282Y) and minor (H63D) mutations of HFE associated with HHC at different levels of serum ferritin, and transferrin saturation (TfS). We found that, in HHC patients without liver disease, DCT1(IRE) and IREG1 are significantly upregulated compared with control individuals, whereas no changes were detected for hephaestin. We also assessed the portion of DCT1 upregulation that is attributable to HHC by studying the regulation of DCT1 in normal individuals as a function of blood hemoglobin, serum iron, and ferritin levels. The “normal” regulation of DCT1 in human intestine by differentiating the two splice forms has not been reported heretofore.
MATERIALS AND METHODS
Proximal duodenal biopsy samples (from the second portion within 5 cm of the duodenal bulb) were obtained from nine patients with classic HHC (C282Y +/+, H63D −/−) and 27 patients with wild-type genotype (C282Y −/−, H63D −/−). Patients were recruited from the clinic or the endoscopy suite of University of Massachusetts Memorial Healthcare. One patient with phenotypic hemochromatosis volunteered to undergo endoscopy for study purposes. All other patients volunteered for duodenal biopsies during medically indicated upper gastrointestinal endoscopies. In all cases, upper gastrointestinal endoscopies were performed in subjects who were ingesting diets that were not limited or supplemented with iron and after an overnight fast of at least 12 h. The study protocol and consent form were approved by the Human Subjects Committee of the University of Massachusetts. All patients were older than 18 years and gave their written, informed consents.
At the time of endoscopy, blood was obtained for hematocrit, hemoglobin, serum ferritin, serum iron binding capacity, transferrin, and HFE mutational analysis. The duodenal biopsies were obtained with standard mucosal biopsy forceps inserted through standard gastrofiberscopes (GIF140; Olympus, Tokyo, Japan). They were immediately placed into 300 μl of Ultraspec Total RNA isolation reagent (Biotecx, Houston, TX). Typically, a total of six biopsies, each ∼5 mg wet weight, were placed into three 1.5-ml microcentrifuge tubes, two specimens per tube. A seventh biopsy was also collected, immediately placed into a cryovial, and snap frozen in liquid N2 for histochemical studies. Tissue in Ultraspec was treated following the manufacturer's instructions, RNAs were stored at −20°C under isopropanol, and before first strand synthesis they were centrifuged, washed in ethanol, and dissolved in 20 μl water.
Random primed first strand cDNA was synthesized from each biopsy sample with an AMV-RT kit, following the manufacturer's instructions (Roche, Indianapolis, IN) and using 1 μg total RNA as a template (OD260). Genomic sequences for IREG1 and hephaestin were obtained from the high throughput genome sequence library of GenBank (IREG1 accession no. AC012488; hephaestin accession no. AC020739). For β-actin and DCT1, the earlier published genomic sequences (23,27) were used to design oligonucleotide primers.
The PCR primers synthesized were as follows: β-actin (GenBank accession no. M10277)2282GGGAAATCGTGCGTGACATT2301 and2583CACGAAACTACCTTCAACTCC2603, exons 3–4; IREG1 (GenBank accession no. AC012488)224TGCTATCTCCAGTTCCTTGC205 and160203TGTCTTCTCCTGCAACAACA160184, exons 1–5; hephaestin (GenBank accession no. AC020739)79303TATACCATCCACCCTCATGG79284 and63319GCATCCTATTGCTCTCCTGA63338, exons 2–4; DCT1(non-IRE) (GenBank accession nos. AF064481 and AF064483)1535CCCATCCTCACATTTACGAG1554 and953ATCCCAGAGTCCAAGACACA934, exons14–17; and DCT1(IRE) (GenBank accession nos. AF064481 and AF064482)1535CCCATCCTCACATTTACGAG1554 and921CCCTAATCCAGTTCTAAG904 exons 14–16a (IRE 3′ end), respectively.
Quantitative PCR amplification was performed with the Lightcycler device/software and the SYBRgreen DNA master kit (Roche) following the manufacturer's instructions. After initial denaturation at 95°C for 30 s, each gene was independently amplified and detected using the following cycles (usually 40 cycles): β-actin: 95°C, 0 s; 62°C, 4 s; 72°C, 5 s; 82°C, 2 s (fluorescent detection; gain: 20); IREG1: 95°C, 0 s; 58°C, 8 s; 72°C, 18 s; 82°C, 3 s (fluorescent detection; gain: 15); hephaestin: 95°C, 0 s; 53°C, 10 s; 72°C, 15 s; 82°C, 2 s (fluorescent detection; gain: 30); DCT1(IRE): 95°C, 0 s; 57°C, 5 s; 72°C, 12 s; 82°C, 1 s (fluorescent detection; gain: 20); and DCT1(non-IRE): 95°C, 0 s; 53°C, 10 s; 72°C, 15 s; 82°C, 2 s (fluorescent detection; gain: 30). Each PCR amplification was followed by a melting curve analysis (97°C, 5 s; 65°C, 15 s; followed by 0.1°C ramping by continuous fluorescent measurement to 99°C) to control for specific amplification products. Each sample was analyzed at least twice; in some cases, additional analysis of a second set of specimens was performed to examine intraindividual variations. Data obtained for DCT1, IREG1, and hephaestin were normalized by comparison with β-actin expression; therefore, the results might be called semiquantitative.
Nonradioactive in situ hybridization was performed as described previously (19), using a digoxigenin-labeled cRNA probe that contained 800 bases of DCT1(IRE) 3′ untranslated sequence (nucleotides 1719–2423, GenBank accession no. AB004857). Frozen sections (8 μm) were cut from the intestinal biopsies in a cryostat and captured onto Superfrost Plus microscope slides (Fisher Scientific, Pittsburgh, PA). The sections were fixed and acetylated and then hybridized at 68°C for 36 h to the DCT1(IRE) probe (approximate concentration, 100 ng/ml). Hybridized probe was visualized using alkaline phosphatase-conjugated antidigoxigenin Fab fragments (Roche) and 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium substrate (Kierkegard and Perry Laboratories, Gaithersburg, MD). Sections were rinsed several times in 100 mM Tris, 150 mM NaCl, and 20 mM EDTA, pH 9.5, and coverslipped with glycerol gelatin (Sigma, St. Louis, MO). Control sections were incubated in an identical concentration of the sense probe transcript.
Statistical and regression analyses were performed using the computer software Prism (GraphPad Software, San Diego, CA). Linear regression was calculated by the least squares method, andr 2 and P values are shown in the graphs to indicate the significance of the linear regression analyses. Significance was determined using the unpaired t-test method.
We studied a total of 36 subjects (see Table1 for details), of whom 27 wereHFE normal [having neither the C282Y nor the H63D mutation; individuals with prefix WT (wild-type) and in the following called “controls”] and 9 of whom were homozygous for C282Y (i.e., have classic HHC; individuals with prefix HHC). We divided the two groups further into five different categories: controls without liver disease (WT − LD), controls with iron-deficiency anemia without liver disease (WT − Fe), controls with liver disease (WT + LD), and HHC patients with or without liver disease (HHC ± LD). The reason for subdividing individuals with liver disease is the observation that liver disease can result in extensive iron accumulation. This division avoids mixing of two different effects influencing body iron homeostasis.
The genomic organizations of the human hephaestin and IREG1 genes were assembled from sequences published by the human genome project. The primers for quantitative PCR of DCT1, IREG1, and hephaestin were designed to amplify mRNA sequences spanning over at least one intron, to rule out amplification of genomic DNA (Fig.1), which was excluded by melting curve analysis using the Lightcycler software and by agarose gel electrophoresis (not shown).
As expected, WT − Fe subjects exhibited significantly lower mean serum TfS values (P < 0.0001) than controls (Fig. 2 A). Furthermore, they had significantly higher mean values for DCT1(IRE) mRNA expression (P = 0.0078) than WT − LD controls, with a mean value ∼10 times higher for anemic subjects than for subjects without anemia (Fig. 2 B). This increase was also detectable by in situ hybridization analysis on sections of frozen intestinal biopsies (Fig. 3). Between these two groups, no significant difference for DCT1(non-IRE) expression (Fig.2 C) or hephaestin (not shown) was observed. With respect to IREG1 mRNA expression, a modest, though statistically not significant, increase of expression in WT − Fe subjects was observed, in contrast with an earlier publication (47). We failed to detect an effect of age or gender on DCT1 mRNA expression in this study (data not shown). Control subjects with normal iron status exhibited a scattering for both DCT1(IRE) and DCT1(non-IRE) expression within roughly one order of magnitude [after normalization to β-actin (Fig.2 A)], in agreement with an earlier publication that measured the total DCT1 content (46). Normalized mean values for DCT1(IRE) were ∼10-fold higher than DCT1(non-IRE), a difference also observed for intestinal DCT1 expression in other mammals (A. Rolfs and M. A. Hediger, unpublished observations).
When we compared the WT − LD control group with the HHC − LD group (n = 5, subjects HHC35–HHC39 in Table 1), we not only observed significant differences in TfS (P = 0.0001, Fig. 2 A), with HHC − LD having about three times higher values, but we also saw a significantly higher DCT1(IRE) expression in the HHC − LD group (P = 0.013, Fig. 2 B). In addition, expression of IREG1 in the HHC − LD group was increased compared with controls (P = 0.04, Fig. 2 D). No differences in DCT1(non-IRE) (Fig. 2 C) or hephaestin (not shown) mRNA expression could be detected between these groups. The expression level of DCT1(IRE) (Fig. 2 B) in the HHC − LD group was the same as for the WT − Fe group, despite markedly different serum iron data (Table 1, Fig. 2 A). This suggests that, in contrast to iron-deficiency anemia, in HFE C282Y+/+ patients, expression of IREG1 and DCT1(IRE) are not correlated with or dependent on serum iron values. This lack of correlation may be due to the disturbance of iron homeostatic signaling in homozygous C282YHFE subjects, analogous to the recently described alterations of hepcidin in mice (28, 33).
We failed to observe statistically significant differences for DCT1(with or without IRE), IREG1, or hephaestin mRNA expression between WT − LD and WT + LD subjects. The number of liver disease patients, or their diversity, might make it difficult to derive any conclusive data, since the etiology and stage of liver disease may influence expression of these genes. Indeed, in controls, expression levels of hephaestin mRNA were within one order of magnitude, regardless of serum TfS or ferritin. In contrast, patient WT09, who has chronic hepatitis C (Table 1), had modestly elevated hephaestin mRNA levels (not shown). The liver disease in this patient may have influenced the expression of the hephaestin gene in an iron-independent manner.
Among controls, we found inverse relationships of DCT1(IRE) mRNA expression and serum TfS (Fig.4 A;r 2=0.49, P = 0.016), serum ferritin (Fig. 4 B; r 2=0.65,P = 0.005), and blood hemoglobin (Fig. 4 C;r 2=0.46, P = 0.015). Our data indicate that, in humans, the IRE-containing form of DCT1 mRNA is regulated in the same way in response to iron deprivation as in rat and mouse (8, 19). We did not observe significant correlations among duodenal IREG1 (see Fig. 6 A), DCT1(non-IRE), or hephaestin mRNA expression and any of the blood data studied in the WT − LD group.
With respect to HHC − LD patients who, as already described, exhibited high DCT1(IRE) mRNA expression despite elevated TfS values (Fig. 2, A and B), no inverse relationship between DCT1(IRE) and serum ferritin or TfS was detectable (Fig. 5 A,r 2=0.26, P = 0.4; and Fig.5 B, r 2=0.67, P = 0.08, respectively). For comparison, linear regression lines of controls from Fig. 4 are shown in Fig. 5. When correlating with blood hemoglobin, we found an inverse linear relationship that did not differ significantly from that of the WT − LD group (Fig.5 C), presumably because hemoglobin synthesis is not affected in HHC. In contrast to this, IREG1 mRNA expression and serum TfS exhibited a linear relationship for the HHC − LD group (Fig. 6 B;r 2=0.878, P = 0.02,n = 5), which was not paralleled by a similar relationship in WT − LD subjects (Fig. 6 A). These findings demonstrate that the correlation between serum iron levels and intestinal iron absorption is disturbed in HFE C282Y +/+ patients.
Numerous advances in iron metabolism have been made in the last 5 years, ushered in by cloning and characterization of the HFEgene and its mutations in HHC. In addition, several proteins responsible for intestinal iron absorption have been identified, including DCT1, IREG1, and hephaestin. Iron-dependent regulation of these transporters has not yet been the subject of in depth analysis in humans. The data described here, revealing an inverse correlation between DCT1(IRE) mRNA expression and serum TfS, serum ferritin, and blood hemoglobin (Figs. 4,A–C), emphasize the relationship of iron absorption and body iron homeostasis. The upregulation of DCT1 mRNA in iron deficiency is consistent with studies in rats (19) and mice (17) kept on diets low in iron, although in our patients, blood loss rather than iron-deficient diets was the cause of iron deficiency. The increase in DCT1 mRNA levels is most probably due to increased stability of the splice form of DCT1, which contains an IRE in its 3′ untranslated region, since the other splice form without the IRE did not exhibit a similar regulation.
In contrast, in HFE C282Y +/+ patients, an inverse relationship for DCT1(IRE) and TfS was not observed and expression of DCT1(IRE) mRNA was higher than anticipated by the serum TfS. HHC − LD patients, despite high serum TfS, exhibit elevations of DCT1(IRE) mRNA expression to the same degree as controls with iron-deficient anemia. Moreover, duodenal DCT1(IRE) mRNA expression in these individuals seems less closely correlated to serum ferritin levels, since untreated patients having similar serum ferritin levels (HHC32 and HHC33) had different serum TfS (Table 1) and DCT1(IRE) expressions.
Our observations on HFE C282Y +/+ patients are in agreement with studies of DCT1(IRE) regulation in homozygous mice with a gene disruption of HFE (17). More recent studies (9, 16, 40) suggest that HFE-regulated iron absorption involves additional, yet-to-be-determined factors. In one study comparing normal and β2-microglobulin knockout mice (9), no significant difference in duodenal DCT1 protein amount was observed. In another recent report of HFEknockout mice (16), a strong influence of the genetic background on the level of iron accumulation was observed in three different inbred mouse strains, suggesting that additional heritable factors influence the phenotype and might account for the differences in expression and regulation of DCT1 reported from different laboratories (9, 17, 40). It has also been reported that, in the duodena of HHC patients and in HFE knockout mice, IRP binding affinity is increased, leading to increased TfR expression (2, 32). These findings support the model of a defect in iron sensing in crypt cells of the small intestine of HHC subjects, leading to inappropriately high levels of expression of genes involved in iron metabolism and transport in differentiating villus cells (37), as if these iron-replete subjects had iron-deficiency anemia. However, how mutation or disruption of theHFE gene results in altered signaling that makes maturing villus cells behave as if the body had iron-deficiency anemia remains unknown. Recent results in mice suggest that the oligopeptide hepcidin (also referred to as LEAP-1) may be a molecule that reports the status of tissue (especially liver) iron stores to duodenal crypt cells (28, 33). In normal subjects, such signaling aids in appropriate regulation of mucosal uptake and transport of iron by modulation of expression of DCT1(IRE) and IREG1. It is possible that patients with HHC or mice with HFE gene disruption lack the ability to transduce the hepcidin signal to their duodenal enterocytes. However, it is currently unknown whether the HFE-TfR complex interacts with hepcidin and whether such an interaction triggers a specific response in enterocytes and macrophages.
With respect to the expression of IREG1 mRNA, our HHC − LD patients exhibited a significant upregulation compared with controls, in agreement with recent findings (47, 48). IREG1 mRNA, which contains in its 5′-untranslated region an IRE, might be regulated at the transcriptional level (Figs. 2 D and 6 A), as described for ferritin or the erythroid form of 5-aminolevulinate synthase (20, 25, 34, 39, 41, 42). It should be noted that the observed linear correlation for IREG1 with serum TfS in HHC − LD occurs at a level of iron loading not normally seen in healthy individuals. Further studies with a larger number of patients are required to investigate the observed upregulation in hemochromatosis. In contrast, no iron-dependent regulation of the hephaestin gene expression was observed.
It has been known for many years that iron overload occurs in liver cirrhosis, and recent results indicate that the great majority of patients with end-stage chronic liver disease and iron overload do not have HHC (i.e., are not C282Y +/+) (5, 6). Although we found no evidence for upregulation of intestinal iron transport in cirrhosis, this possibility cannot be ruled out, since none of our liver disease patients had iron overload (Table 1), as shown by liver biopsies (data not shown). Therefore, how and why patients with end-stage liver disease develop iron overload remains to be elucidated.
We thank the nursing and clinical staff of the University of Massachusetts Memorial Gastrointestinal Endoscopy Unit for their help in performing duodenal mucosal biopsies and care of patients studied.
Current address for J. G. Kohlroser: Division of Gastroenterology, Albany Medical College, Albany, NY
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants RO1-DK-57782 to M. A. Hediger and RO1-DK-38825 and DK-92326 to H. L. Bonkovsky and by a grant from the American College of Gastroenterology to J. G. Kohlroser and H. L. Bonkovsky.
Addresses for reprint requests and other correspondence: M. A. Hediger, Harvard Institutes of Medicine, Room 570, 77 Avenue Louis Pasteur, Boston, MA 02115 (E-mail:) or H. L. Bonkovsky, University of Massachusetts Medical School, Rm. S6–737, 55 Lake Ave. North, Worcester, MA 01651 (E-mail: ).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2002 the American Physiological Society