Year: 1998
Type: review
Abstract: Until August 1996, the molecular basis for the autosomal recessive disease of iron metabolism, hereditary haemochromatosis (HH), was completely unknown. Although its association with HLA-A3 ( Simon et al, 1976 ) on chromosome 6p had been known for over 20 years, little progress had been made in identifying the underlying molecular mechanisms of the disease. Positional cloning approaches were initiated by several groups worldwide but were generally hampered by the lack of a specific chromosomal rearrangement associated with the disease and the low recombination frequency in this region of chromosome 6. Nevertheless, linkage disequilibrium analyses of several different patient populations provided evidence that the gene would ultimately reside well distal to the class I region of the MHC ( Jazwinska et al, 1995 ; Raha-Chowdhury et al, 1995 ; Seese et al, 1996 ). Finally in August 1996, Feder et al (1996 ) published the identity of the possible HH disease gene. Through an extensive linkage disequilibrium analysis over several megabases on chromosome 6p, a 250 kb stretch of genomic DNA from a unique yeast artificial chromosome clone spanning the linkage disequilibrium peak was chosen for large-scale sequencing. Fifteen genes were identified within this region; several were analysed for mutations in HH patients. A single missense mutation (C282Y) was found in the homozygous state in approximately 80% of patients. The gene, termed HLA-H by Feder et al (1996 ) but subsequently redesignated to its originally proposed name of HFE by the Nomenclature Committee of the Genome Database ( Bodmer et al, 1997 ), is an MHC class Ib gene. Class Ib genes are distinct from MHC class Ia genes due to their limited expression and lack of polymorphism ( Stroynowski & Fischer-Lindahl, 1994). Considerable scepticism greeted the discovery of HFE due to the limited data supporting a plausible mechanistic scenario for HH based upon an HLA-like molecule. In view of this, the headlines in a Nature Genetics editorial describing the finding summed up the scientific field's initial reaction to the work: 'Haemochromatosis ... definite maybe!' ( Anonymous, 1996). Interestingly, 2 years earlier, Dorak et al (1994 ) had fortuitously proposed that a putative human counterpart of a mouse class Ib molecule, the thymus leukaemia antigen (TL), may represent the HH disease gene. With the identification of the HFE gene, the goal of unravelling its biological role in iron metabolism has begun. A mutation in an HLA class Ib gene suggests one of several extant possibilities, namely that haemochromatosis may have an immunological basis, which is supported by several animal models and clinical observations. We review these findings and the possible role of HFE in HH, although it must be acknowledged that the HFE protein may have evolved a unique function well separated from any immunological underpinning. Hereditary haemochromatosis (HH) is a common inherited autosomal recessive disorder of iron metabolism with an estimated prevalence of 3–8 per 1000 in the Caucasian population ( Edwards & Kushner, 1993). Heterozygote carriers are found at a prevalence of approximately 1 in 10 but are rarely affected. Homozygous individuals absorb excessive amounts of dietary iron from birth which progresses over time; the gradual accumulation of iron eventually results in massive iron overload in several organ systems. The classically described presentation of 'bronze diabetes' is the result of skin hyperpigmentation and diabetes mellitus. However, many other organs are also affected which may result in arthropathy, hepatic cirrhosis and subsequent hepatocellular carcinoma, cardiomyopathies and hypogonadotropic hypogonadism by the fourth or fifth decade of life ( Bothwell et al, 1995 ). Early diagnosis followed by phlebotomy treatments to remove iron-rich red blood cells and deplete total body iron stores can prevent these manifestations and normalize life expectancy. Some patients require the removal of hundreds of units of blood and must maintain phlebotomy treatments for life. Diagnosis is currently accomplished by screening with serum transferrin saturation and serum ferritin levels, whereas liver biopsy with quantitative iron determination is often used to irrefutably document iron overload and to assess liver damage. The discovery of the HFE gene now enables genetic diagnosis which may modify current clinical practice. A recent United States Centers for Disease Control/National Human Genome Research Institute conference (1997) addressed this and other related issues. The underlying physiological defect in HH is not yet known; nevertheless, it is clear that iron preferentially accumulates in parenchymal cells of affected organs rather than the usual iron storage cells of the reticuloendothelial cell (RE)/macrophage system. In normal individuals iron stores are regulated at the level of intestinal absorption; the only means of excreting iron is through the loss of cells normally sloughed from the epithelium of the gut, urinary tract and skin, as well as through blood loss (e.g. menses) and pregnancy ( Lash & Saleem, 1995). In HH the intestinal transfer rate of iron to the serum is abnormally high ( Conrad et al, 1994 ). In addition, neither the intestinal epithelial cells nor the tissue RE/macrophage cells accumulate iron in HH until late in the course of the iron overload process. Prior to the discovery of HFE, several tissue and cell types from HH patients, including liver ( Pietrangelo et al, 1991 ), peripheral blood monocytes ( Adams et al, 1991 ) and duodenum ( Lombard et al, 1990 ), had been analysed for abnormalities in the major molecules involved in iron metabolism. These proteins include transferrin, transferrin receptor, ferritin, iron regulatory protein and tumour necrosis factor. Although various perturbations in the expression of these molecules have been demonstrated, prompting much speculation regarding the HH biochemical defect, no consistent physiological hypothesis for its cause had emerged ( Conrad et al, 1994 ) nor had a suitable in vitro model system been developed ( Zheng et al, 1994 ). In view of the relative gap in knowledge regarding the relationship of HFE to the well-characterized iron regulatory proteins, it appears appropriate to begin to explore alternative possibilities with respect to HFE's biological function. HFE belongs to the human major histocompatibility complex (MHC) class I gene family which is composed of approximately 20 class I genes, pseudogenes and gene fragments ( Geraghty et al, 1992a ). The common structure of all MHC class I genes consists of a leader sequence (approximately 25 amino acids in length), three external domains (termed α1, α2 and α3 of 90 amino acids each), a transmembrane domain (∼20 amino acids) and a short cytoplasmic segment (variable in length). The α1 and α2 domains fold to form a pocket or groove composed of a β pleated-sheet base flanked by helical walls; the groove is also termed the antigen recognition site which is the region that binds processed peptides ( Young et al, 1995 ). The α3 domain is intimately associated with beta 2-microglobulin (β2m) and is largely conserved among class I genes. β2m is a 12 kD protein which noncovalently binds to the MHC class I 45 kD heavy chain and is obligatory for normal class Ia cell surface expression. Three loci constitute the highly polymorphic class Ia (or classic) genes termed HLA-A, HLA-B and HLA-C. These three highly homologous genes are expressed in all somatic cells and function in the presentation of both self and foreign peptides, a process which regulates the CD8+ arm of T-cell recognition and cytotoxicity ( York & Rock, 1996). Each of these genes codes for an HLA heavy chain which binds both β2m and peptides during assembly of the protein complex. A multisubunit transport protein termed TAP (for transporter associated with antigen processing) transfers proteolysed peptide fragments by an energy-dependent mechanism into the lumen of the rough endoplasmic reticulum where the peptides (generally of 8 or 9 amino acids in length) are available to complex with nascent class I heavy chains should the peptide possess the appropriate motif or anchor residues complementary to amino acids within the heavy chain antigen recognition site ( Ljunggren et al, 1995 ; Young et al, 1995 ; Rammensee et al, 1993 ). Cell surface class Ia expression therefore requires a functional TAP transport mechanism. In contrast to the class Ia gene assemblage, which has been very thoroughly investigated due to its direct involvement in T-cell education and functionality, the MHC class Ib genes constitute a group of poorly characterized molecules. Collectively, the class Ib genes possess several general characteristics which contrast sharply with those of the class Ia group. First, they are expressed with relative tissue specificity, with many restricted to the placenta and/or gut. Second, they are not polymorphic ( Klein & Figueroa, 1986). Indeed, nucleotide differences observed within allelic class Ib gene families are usually silent, third-position codon changes, whereas the class Ia genes are highly polymorphic in the α1 and α2 domains to allow for structural diversity in the binding of various peptides. In addition, most class Ib molecules do not require TAP for their expression and their functions have been difficult to determine ( Shawar et al, 1994 ). Although low-level HFE expression appears to be fairly ubiquitous, its expression, which may be TAP independent since TAP knockout mice do not accumulate iron ( Santos et al, 1996 ), is high in the gut ( Parkkila et al, 1997 ). In humans, five class Ib genes are found within the MHC region on chromosome 6p and are termed HLA-E, HLA-F, HLA-G, MICA and MICB ( Geraghty et al, 1992b ). HLA-F ( Geraghty et al, 1990 ) has a broad pattern of expression within lymphoid tissues, whereas HLA-G ( Geraghty et al, 1987 ) is detected on placental cells where it has been predicted to be involved in both quelling maternal immunological reactivity toward paternal HLA antigens ( Kovats et al, 1990 ) and protection from natural killer (NK) cell-mediated lysis ( Pazmany et al, 1996 ). The MICA and MICB genes (for MHC class I chain related genes), whose expression is relegated to the gut, are highly divergent class Ib members which reside between the TNF and HLA-B loci ( Bahram et al, 1994 ). MIC genes do not appear to manifest a strict requirement for β2m binding and are not dependent upon TAP for expression. These characteristics suggest that either they do not bind peptide or they derive peptide from a TAP-independent pathway ( Groh et al, 1996 ). Several other class Ib genes reside outside of the MHC region and, like the MIC genes, are expressed within the intestine. These genes are highly divergent from class Ia sequences and do not share any significant sequence identity. Included in this group are the CD1 genes, MR1, Zn-α2-glycoprotein and the FcRn gene which possesses nucleotide similarity to an Fc receptor ( Martin et al, 1987 ; Hashimoto et al, 1995 ; Simister & Mostov, 1989; Araki et al, 1988 ). Although CD1 is involved in the selection of natural T cells which secrete IL-4 ( Bendelac et al, 1995 ), no biological information is yet available for MR1 and the Zn-α2-glycoprotein. HFE is the newest member of the class Ib gene family to be described. A phylogenetic analysis of the HFE gene, based on a comparison of α1 and α2 domains from several class Ia and Ib genes, confirms that the HFE gene, like the CD1, FcRn, MIC and Zn-α2-glycoprotein genes, is a highly divergent member of the class Ib family which probably shared a common progenitor several hundreds of millions of years ago (Fig 1). HFE is closely related to its mouse orthologue ( Hashimoto et al, 1997 ), suggesting that HFE may subserve an important and conserved function. and α2 comparisons). The evolutionary conservation of HFE is also apparent, based on a comparison of the conserved antigen recognition site with those of the class Ia proteins. Of the amino acids comprising the functionally important B and F pockets ( Young et al, 1995 ), HFE is quite similar to HLA-A2. In the B pocket, HFE and HLA-A2 have only one nonconservative amino acid difference. Within the F pocket, HFE and HLA-A2 have three nonconservative amino acid differences. Importantly, only one amino acid difference exists between the human and mouse HFE proteins in these same regions. HFE demonstrates certain similarity to human FcRn, a class Ib molecule that does not bind a ligand within its antigen recognition site ( Story et al, 1994 ). Like FcRn, HFE contains a proline at position 188 in the floor of the antigen recognition site; the FcRn residue has been suggested to clip or close its antigen recognition site ( Burmeister et al, 1994 ). However, human CD1d also possesses a proline at this site whilst retaining the ability to bind peptide ligands despite its unusually narrow antigen recognition site ( Zeng et al, 1997 ). Thus, the overall evidence from structural data is inconclusive regarding the HFE protein's predicted ability to bind ligands. Based on Northern blot analysis, low-level HFE expression is detected in many tissues with the exception of brain. Distinctly higher levels are observed in the ovary, heart, kidney and pancreas, whereas the highest levels of expression are seen in small intestine and liver. RNase protection analysis, offering slightly greater sensitivity, demonstrated only minimally detectable levels in fetal brain and lymphoblasts ( Feder et al, 1996 ). At the protein level, immunohistochemical analysis using an antibody to a C-terminal peptide revealed HFE protein staining in epithelial cells and subendothelial leucocytes throughout the entire GI tract ( Parkkila et al, 1997 ). Staining of the entire plasma membrane was seen in non-polarized epithelial cells such as the stratified squamous epithelium of the oesophagus. Immunoreactivity was restricted to the basolateral surfaces of most polarized epithelial cells. Interestingly, an intracellular perinuclear staining pattern was observed in the crypts of the duodenum, jejunum and ileum, with little or no staining of the villi or Brunner's glands. Little immunoreactivity was seen in the apical or basolateral membranes of the small intestine epithelium. In the liver, HFE expression was prominent in the basolateral surface of the plasma membranes of the bile duct epithelium sinusoidal lining cells. No immunoreactivity could be demonstrated in the pancreas or brain. A single cysteine to tyrosine substitution at position 282 (C282Y) in the HFE protein (a G to A transition at nucleotide 845) was observed within the majority of HH patients initially reported (83% of 178 patients). A second sequence difference, which produces a histidine to aspartic acid substitution at position 63 (H63D) (Fig 2) due to a C to G transversion at nucleotide 187, bears an interesting but direct relationship to HH in that compound heterozygotes (C282Y/H63D) succumb to the disease, although with reduced penetrance ( Beutler, 1997a). m and achieve cell surface expression. The common C282Y mutation was predicted to abrogate the formation of an intra-α3 domain disulphide bond required for the binding of β2m and subsequent cell surface expression ( Feder et al, 1996 ). Transfection experiments have recently been performed using both the wild-type and mutant cDNA constructs ( Feder et al, 1997 ). It was observed that wild-type HFE bound β2m whereas the C282Y mutation, but not the H63D mutation, abolished this interaction. Immunofluorescence studies demonstrated that the wild-type and H63D HFE proteins were expressed on the cell surface whereas the C282Y HFE mutant remained completely intracellular, suggesting that an abnormality in protein trafficking and/or cell-surface expression of HFE is the fundamental biochemical defect involved in HH (Fig 2). The occurrence of the C282Y mutation in the homozygous state has been analysed in HH patients in a number of geographical areas. Variable frequencies have been observed: 83% ( Feder et al, 1996 ) and 82% ( Beutler et al, 1996 ) in the United States, 91% ( Jouanolle et al, 1996 ) and 72% ( Borot et al, 1997 ) in France, 64% in Italy ( Carella et al, 1997 ), 100% in Australia ( Jazwinska et al, 1996 ) and 91% in the United Kingdom ( The UK Haemochromatosis Consortium, 1997). Essentially no normal have been to of the mutant HFE frequency studies that the C282Y mutation is populations that are not of ( Beutler & et al, 1997 ; et al, 1997 ; et al, 1997 ). In all patient the H63D than the C282Y mutation whereas the H63D mutation demonstrated a similar frequency among HH patients and These data are consistent with HH a disease, whereas the role of the H63D although associated with HH ( Beutler, Thus, the that the H63D may represent a polymorphism in linkage with an as yet mutation can not be In the of any mutations a significant number of patients with normal HFE genes may possess mutations in other genes either involved in the expression of HFE or involved in iron metabolism. in HFE have been predicted to to iron associated with other ( Beutler, such as ( et al, al, 1997 ) and ( et al, 1997 ). The association of HH in individuals of has to the that the mutation in HFE in a population ( Simon et al, ). The of the HFE mutation has been estimated at between and based upon chromosomal analysis ( et al, 1997 ). Thus, the disease most probably in the of the The high frequency of the mutant gene suggests that the mutation may a such as protection from iron in of ( 1996). The intestine has been the organ underlying the defect in iron metabolism in HH ( Bothwell et al, ; & HH patients appear to both normal iron into the intestinal and normal of iron into the intracellular storage of the However, the rate of iron transfer from the to the plasma is ( et al, 1991 ). the normal and into the are by the of iron to the results in the of iron in the intestinal cells. The normal physiological transport of iron from the the basolateral membrane and into the ( Bothwell et al, ) is poorly Several proteins involved in iron and have been although their direct physiological role in iron is not known ( et al, 1987 ; Conrad et al, 1993 ). where the HH within this of transport The role of a class Ib protein in this process is also not but suggests of immunological involvement which is supported by the number of that are also in the small intestine and in association with ( significant of iron in HH with the system. and of the system are major of iron through the of red blood cells and their and tissue are also the same cells which and foreign peptide antigens in the of MHC molecules to T a process known as antigen presentation ( York & Rock, 1996). A in HH is the accumulation of iron in parenchymal such as the normal storage in cells seen in other iron overload and of iron from cells has been reported ( et al, ) which the transfer In ( et al, ) an in ferritin by blood to the defect which may to the lack of iron accumulation in cells. The abnormalities in both the and the of an underlying common biochemical defect by the loss of normal HFE protein function. The of the iron regulatory protein in iron metabolism through of ferritin and transferrin receptor gene expression was reported to be in HH monocytes to normal and iron overload ( et al, 1997 ). However, in HH patients with normal iron was not Thus, the of appears to be to the HH in monocytes would prevent ferritin and may represent a molecular mechanism underlying the storage of iron in cells. perturbations in the of HH patients have been et al (1994 ) found abnormally high in peripheral blood T in HH patients. were with the number of required to deplete iron and were not normal iron The high were due to a in CD8+ T cells ( et al, 1994 ). The of both and cells the chain of the T-cell receptor as well as CD8+ cells have been reported to be ( et al, 1994 ). These clinical data may suggest as to general in T cells from the the and in CD8+ cells from an abnormality in class I related to iron Within the CD8+ an of CD8+ T with a in the of CD8+ T was also was in HH patients the HLA-A3 expression was normal in cells. A higher of but not cells was also found within the peripheral CD8+ T-cell in HH patients relative to ( et al, 1997 ). CD8+ T cells from HH patients were of to but CD8+ T cells not the normal CD8+ T also these immunological are to the of cell-surface HFE molecules or are the result of and were the to that β2m knockout mice succumb to iron overload in a human HH in many parenchymal iron by high serum transferrin saturation and an iron transfer rate ( et al, 1994 ). finding was by & suggested that the iron was by a gene in the region termed The of the β2m knockout mice significant for class I involvement in iron The iron overload in the mice could be by with cells from liver from normal β2m mice of the same A in iron stores from the hepatic to cells within 1 followed by a in total iron stores at However, during this transferrin saturation and transfer rate remained These results suggest that cells class I molecules could not the intestinal defect, but could the defect in the liver. It is possible that cells lymphoid and/or within the fetal liver, ultimately a role in the for the role of T cells in iron metabolism is supported by the of iron overload in mice for the recombination ( Santos et al, 1994 ). In contrast to mice not succumb to parenchymal iron suggests that the HFE protein is distinct from the majority of class I heavy chains whose expression is TAP dependent ( Young et al, 1995 ). the homozygous knockout of the gene, whose is an molecule involved in the selection of T cells its with class I also to iron overload ( Santos et al, 1996 ). However, in CD8+ levels in mice for class I genes ( et al, 1993 ) and in mice for the TAP peptide transporter which class Ia expression ( Ljunggren et al, 1995 ). These with the human that although a defect in HFE can cause abnormalities in the the loss of cells has no on iron suggests that HFE may function of the immunological and in HH patients. The identification of HFE as an HLA molecule several as to its role in iron metabolism. a class Ib molecule, does the HFE protein bind a peptide is a in unravelling the of cells such as intestinal although not as have been to as cells in vitro (e.g. with ( et al, ; et al, ). parenchymal cells that HFE in the gut, liver and antigen to iron a bound peptide ligand is processed for a number of are the ligand TAP it at the as are many class Ib it from a to iron evidence a role for T cells in iron metabolism. the HFE protein with a or and, does it T cells mechanism by which HFE may function is of iron such as and IL-4 have upon iron metabolism. A for a T cell which could with HFE is the natural T which appears to with a of class Ib proteins including CD1 and ( Bendelac et al, 1994 et al, 1995 ; et al, 1996 ). cell not which is consistent with normal in mice ( Santos et al, 1996 ). natural T cells are the most T-cell within the liver ( & other in cause in iron storage through other important such as T cells in cause iron overload ( Santos et al, 1994 ). The (C282Y) mutation for only of the of HH in many the that mutations in other of the HFE iron pathway may for the other of iron is an which a role in a of A number of genes may be involved in the transport and storage of iron including a of molecules such as protein is homologous to an molecule which has recently been to be in the mouse model ( et al, 1997 ). The discovery of HFE now the for to the HFE protein's important role in iron metabolism and the mechanisms underlying the of genetic iron The current evidence is with the that at one of the major of iron HFE, may be an immunological molecule may for a significant between iron and ( et al, 1995 ). The that HFE has a divergent role from the yet be and would in with the of the class Ib family ( was supported by from the and the 1 to
Source: British Journal of Haematology
Institution: Pennsylvania State University
Language: English
Subfield: Hematology
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W1548610513