Haemochromatosis is a common, inherited disorder of iron metabolism. The gene (HFE), which is mutated in the majority of patients, has been cloned. The HFE protein, however, is not an iron transporter, but rather is thought to function as a regulator of iron absorption. The cloning of theHFE gene was followed rapidly by the identification of further proteins involved in the iron absorption pathway in the duodenum and transfer to the main iron storage site in the liver. The identification of a new metal ion transporter in the rat provided the first molecular information on the active absorption of metal ions by mammalian cells.1 This divalent metal transporter 1 (DMT1) has a broad substrate specificity that includes Fe²⁺ and a range of other divalent metal cations.1 (DMT1 is also referred to as divalent cation transporter 1, DCT1, and natural resistance-associated macrophage protein 2, Nramp2.) DMT1 mRNA is widely expressed, with high levels in the proximal duodenum, the site of absorption of iron and most other divalent metal ions.1 The biological importance of this transporter is shown by its involvement in two naturally occurring animal mutants of iron metabolism. Homozygous mutation ofDMT1 in the mouse is responsible for microcytic anaemia, and in the rat, the Belgrade phenotype of microcytic, hypochromic anaemia, with severe defects in intestinal and reticulocyte iron absorption.2 3 The Belgrade reticulocyte defect seems to be a failure to transport iron out of the transferrin cycle endosomes.

Alternative splicing of exons in the primary DMT1 RNA transcript results in two classes of mature DMT1 mRNAs: DMT1(IRE) includes an iron responsive element in the 3′ untranslated region; DMT1(non-IRE) does not.4 5 Duodenal DMT1 mRNA increases in response to dietary iron deficiency.1 This iron sensitive regulation may thus be mediated by the 3′ iron responsive element of the DMT1(IRE) form, in a similar way to regulation of the transferrin receptor.6 In HFErelated haemo- chromatosis, duodenal enterocytes have increased duodenal iron responsive protein activity, and decreased ferritin mRNAs, with respect to controls. Thus despite body iron overload, the evidence suggests that haemo- chromatotic duodenal enterocytes are paradoxically iron deficient.7 Consistent with this, DMT1 mRNA was increased in duodenal mucosa from patients with haemochromatosis, suggesting a possible mechanism for the increased duodenal iron uptake and iron over- load in haemochromatosis.8 Similarly, in theHFE knockout mouse, northern blot analysis indicated an increase in the DMT1(IRE) mRNA relative to controls in duodenum. However, in the iron overloaded liver of these mice, there was no increase in either total or DMT1(IRE) mRNA, as measured by northern blot analysis.4

In this issue (see page 270), Trinder and colleagues report RNA in situ hybridisation and immunohistochemical studies of DMT1 in liver and duodenum of rats, in response to altered iron stores. Convincing evidence is given for the specificity of the anti-DMT1 peptide antibody. Both the riboprobe used to detect mRNA and the antibody used for protein staining are common to the IRE and non-IRE forms of DMT1, and so total DMT1 staining is described. In duodenal villi from iron deficient animals, RNA in situ hybridisation indicated an increase in DMT1 mRNA with respect to controls. Immunohistochemistry did not detect any DMT1 protein staining in the crypts in any iron status, under the experimental conditions used, but DMT1 staining was detected, beginning at the crypt–villus junction and reaching highest levels in the upper half of the villus. Protein staining of the villus was highest in iron deficiency and least in iron overload. The mechanism underlying the change in expression of DMT1 as the duodenal enterocyte migrates from crypt to villus is not yet clear, although these observations are consistent with regulation of the villus DMT1(IRE) via its 3′ iron responsive element.

Trinder et al also describe some of the first observations on the pattern of expression ofDMT1 in the liver under different dietary iron conditions. Interestingly, a relative increase in the DMT1 protein staining and polarisation to the hepatocyte sinusoidal membranes was seen in iron excess. In liver, this different regulatory pattern (apparent upregulation and polarisation in iron excess) may indicate a difference in the regulatory response mechanism of DMT1 compared with that of the duodenal villi, probably involving expression of DMT1(non-IRE).

Polarisation of DMT1 to the sinusoidal plasma membranes of hepatocytes may explain the remarkable avidity of the liver for the highly reactive, non-transferrin bound (NTB) form of iron. This pool in normal individuals is minute or undetectable in the serum, where iron is normally bound to transferrin. However, under conditions where transferrin is fully saturated with iron, such as in haemochromatosis, substantial serum NTB iron may be detected.9 NTB iron stimulates both the formation of the highly reactive and damaging hydroxyl radical, and the peroxidation of membrane lipids. However, most serum NTB iron is extracted by the liver in the first pass.9 10 Trinder et alpropose that DMT1 on the microvillus membrane of hepatocytes clears the portal blood of NTB iron, and so reduces the risk of unregulated iron uptake and oxidative damage to cells elsewhere in the body. Conversely, in iron deficiency, relative downregulation of DMT1 in the liver, and upregulation in the duodenal enterocyte may permit iron to bypass the liver, bind transferrin, and be transported to body cells according to transferrin receptor expression.

This tissue specific regulation of DMT1 expression pattern reflects both the bodily requirements for iron, and its potentially damaging effects. Iron is essential because it can be readily and reversibly oxidised and reduced between its two common valencies, ferrous (2+) and ferric (3+). This property is exploited by iron dependent proteins and enzymes, to permit oxygen transport and the redox reactions of respiration. However, in excess, iron catalyses formation of reactive free radicals, leading to tissue damage. Given that our regulation of body iron content is at the level of absorption, the differing DMT1 expression pattern in duodenum and liver is clearly important to ensure regulation of this essential but toxic metal ion.