Harry J. McArdle, Rowett Research Institute, Greenburn Road, Bucksburn, Aberdeen AB21 9SB, UK (e-mail: email@example.com).
Iron and copper are both essential micronutrients and are required for a wide variety of enzymatic and other processes within the developing foetus. Transfer of both nutrients across the placenta is tightly regulated. In this review, we consider their mechanisms of transport, how the transfer is modulated in response to nutritional requirements and how the two metals interact. Iron uptake is via the transferrin receptor, followed by endocytosis, acidification of the vesicle, and release of the iron into the cytosol, and transfer across the basolateral membrane. Many of the genes involved have been identified, and, to varying extents, their mechanisms of regulation clarified, but there are still unanswered questions and conundrums. For example, although the ion channel DMT1 (now formally known as slc11a2) is essential for iron uptake in the gut, knockout mice, which have no slc11a2 protein, have apparently normal transfer across the placenta. There must, therefore, be an alternative mechanism, which remains unclear, although nonspecific calcium channels have been proposed as one possibility. For copper, uptake is a carrier-mediated process, and intracellular transfer is mediated by proteins known as chaperones. Efflux is through ATPases, but their localisation and how they are regulated is only now being elucidated. Regulation of copper proteins appears to be different from that of iron, with localisation of the protein, rather than changing levels, being responsible for altering rates of transfer. This may not be true for all the proteins and genes involved in the delivery of copper, and, again, there is much that remains to be clarified. Finally, we consider the interactions that occur between the two metals, reviewing the data that show how alterations in levels of one of the nutrients changes that of the other, and we examine the hypotheses explaining the interactions.
Copper and iron are essential micronutrients, critical for normal health and development (1). These metals play a central role in oxidation and reduction reactions, and deficiencies, whether from dietary or genetic causes, have a wide range of deleterious consequences (2).
Iron deficiency during pregnancy is disturbingly common. Even in developed countries, anaemia during pregnancy, over and above the normal drop in haemoglobin concentration caused by blood volume expansion, can be found in a significant proportion of pregnancies (3). Babies born to iron deficient mothers have been shown to be smaller, to have developmental problems, especially neurological (4) and, in later life, to have a higher risk of hypertension and other diseases (5–8).
There are no reliable biomarkers for copper status, so whether deficiency is a significant public health problem is not clear (9). We do know, however, that dietary intake in women aged 19–24 years is generally below recommended levels (10, 11), and there are data suggesting that this is more of a problem during pregnancy, when requirements increase.
The metabolism of the two metals is tightly interlinked. This was first shown in 1927, when administration of iron salts to anaemic rats failed to reverse their condition, whereas giving copper in the form of ashed foods returned haemoglobin levels to normal (12). It is also clear that transport of the metals is regulated at the level of the placenta (13). In this review, we examine the mechanisms of uptake and transfer across the placenta, comparing and contrasting as appropriate, and we also discuss how they interact and how transfer is regulated. We have rather downplayed the species differences that may exist in the model systems we discuss. We have chosen to do this deliberately, since the data suggest that the mechanisms are very similar in all species with haemochorial placentas, such as mouse, rat and human.
Iron uptake and transfer across the placenta
Despite its central importance in foetal development, the mechanism of iron transfer across the placenta has received surprisingly little attention. Because of this, in many stages in the transfer, we have had to extrapolate from duodenal studies, speculating that the mechanisms are similar in the two epithelia. Such data as there are available support this conclusion, but the caveat must always be borne in mind (for a diagrammatic representation, see Fig. 1).
The first step in uptake, however, is different between the two epithelia. Iron in serum is bound to transferrin, rather than being present in a low-molecular weight form as it is in the gut (14). Transferrin binds to transferrin receptor 1 on the microvillar membrane (15). Following binding, the complex is taken up into the placenta via coated vesicles (16). The vesicle becomes acidified and the iron dissociates from the transferrin. The iron effluxes from the endosome into the cytosol, probably through the iron channel protein DMT1, and the transferrin receptor/transferrin complex recycles back to the surface of the cell (16, 17).
Whether and how DMT1 is involved in the placenta is not completely straightforward. In an elegant series of experiments, Andrews et al. (18) demonstrated that DMT1 (Slc11a2) is not essential for iron delivery to the foetus. They showed that knockout mice (scc11a2−/−) were born anaemic, but that the total iron content in the newborn animals was similar to wild types. Importantly, they also demonstrated that the liver in the slc11a2−/− actually had higher levels of iron than wild-type (18). This stongly suggests that alternative pathways of iron transfer exist in the placenta. What these pathways comprise has not yet been identified. By contrast, following birth, the knockout mice died of iron deficiency quite rapidly, demonstrating that the protein is essential in absorption across the gut.
Iron metabolism in other cell types is regulated through a complex series of protein–protein interactions (19). Uptake of iron into the duodenal cell, unlike the trophoblast, is via DMT1, following reduction of Fe3+ to Fe2+ by DcytB (20). Expression of DcytB and DMT1 is regulated by enterocyte levels of iron (21). Efflux of iron from the enterocyte to the circulation is through ferroportin (IREG1) (22, 23) as Fe2+. The iron is oxidised to Fe3+ by a protein called hephaestin, a copper oxidase that is very similar in structure to ceruloplasmin (24) and then binds to serum transferrin. Functional studies required to demonstrate this mechanism directly in the placenta have not been carried out, but there are immunological data to support the hypothesis (23).
Absorption of iron into the portal circulation is regulated by a protein called hepcidin. Hepcidin was first identified as a microbial protein, is produced by the liver, and negatively regulates iron absorption, probably by binding to, and accelerating proteolysis of, IREG1 (25). Expression of hepcidin drops in iron deficiency and in the maternal liver during pregnancy (26). This results in a decrease in ferroportin breakdown and, together with rises in expression of other genes in the maternal liver and gut, explains the increased absorption of iron that occurs during pregnancy (26). Hepcidin is thought to act by binding directly to IREG1, inducing internalisation of the protein (25).
The iron overload related gene for haemochromatosis (Hfe) is expressed in the placenta at both protein and mRNA levels. It is found in the syncytiotrophoblast at quite high levels, predominantly on the apical membrane (27). Interestingly, the data show strongly that the protein is closely associated with transferrin receptor. Why this should be so is not clear. In the intestine, ablation of the Hfe gene improves survival in slc11a2 knockout mice (18). Perinatal haemochromatosis is similar in terms of iron metabolism to the adult form, but whether the cause is a defect in the gene is not always apparent (27). Clearly, homozygosity for at least some forms of haemochromatosis cannot be lethal in the perinatal period, or we would not have an adult form of the disease.
Iron in the developing foetus is accumulated against a concentration gradient, which is not surprising given the mechanism of absorption and, in the case of maternal iron deficiency, the placenta can protect the foetus significantly. For example, in our own studies, we fed rats a diet with iron levels that induced a reduction of iron stores in the maternal liver of approximately 73% by day 21.5 of pregnancy. By contrast, the foetal liver levels had dropped only to 50% of controls (28). In the placenta itself, the iron levels also decrease, and this change is associated with alterations in cytokine expression (28), specifically tumour necrosis factor (TNF)-α, TNF-αR1 and leptin (28). How these factors are inter-regulated is not known. However, we do know that the protection mechanism involves increased expression of transferrin receptor in the placenta, together with a rise in DMT1, but no change in mRNA levels for ferroportin (29).
In pregnancy, we have shown that hepcidin is produced by the foetal liver, and levels are, as expected, related to foetal liver levels of iron (L. Gambling and H. J. McArdle, unpublished data). However, there are no clear published data demonstrating the links and the interactions between the liver, signalling the requirements of the developing foetus. It seems we have much yet to learn.
Copper uptake and transfer across the placenta
The molecular mechanism of placental copper uptake and transfer is much less well understood than that of iron (Fig. 2). As for iron, copper transfer across the placenta increases during gestation (30). Uptake is through a high affinity carrier, Ctr1. Ctr1 is expressed early in pregnancy, and homozygous mutant embryos die early in gestation (31). Ctr1(−/−) embryos can be recovered at E8.5 but are severely developmentally retarded and morphologically abnormal. Histological analysis reveals discontinuities and variable thickness in the basement membrane of the embryonic region and an imperfect Reichert’s membrane, features that are consistent with a lack of activity of the cupro-enzyme lysyl oxidase, an enzyme essential for collagen cross linking (31).
Once taken up by the placenta, we assume the mechanism of transfer is similar to that described for other cells (e.g. gut cells). Cu is bound to one of a series of chaperone proteins, which deliver the metal to its target molecule. There are two Cu-ATPases expressed in placenta, ATP7A, the Menkes protein, and ATP7B, the protein altered in Wilson disease (32).
How the two ATPases are linked to deliver copper to the foetal circulation is only now beginning to be elucidated. In other tissues, the location of the ATPase is altered by the copper status. In the liver, for example, ATP7B is found in intracellular vesicles when copper levels are low, but increased copper results in translocation to the bile canalicular membrane so that copper can be excreted into the bile (33). Similarly, ATP7A is located either intracellularly or on the basal membrane in enterocytes.
In placenta, ATP7A is located in several different cell types, whereas ATP7B is found only in syncytiotrophoblast (32). Intriguingly, protein levels do not appear to change during gestation, which implies that the increase in transfer seen as development progresses (30) is related to localisation of the protein.
Using a trophoblast cell model, Hardman et al. (34) have suggested that the ATPases act in a concerted fashion. ATP7A is located on the basolateral membrane and excretes copper into the foetal circulation. ATP7B, in contrast, is either in a perinuclear compartment or, in the presence of high copper, translocated to the microvillar membrane, thereby transporting copper back to the maternal circulation. Curiously, ATP7B mutant animals are born with low copper levels in the liver and high levels in the placenta, which rather contradicts this model (35). Even such a simple model, therefore still needs work.
Interactions between copper and iron during pregnancy
As mentioned in the Introduction, it has been known for some time that the metabolism of these two nutrients is inter-related, but it is only much more recently that the mechanisms are beginning to be unravelled. During pregnancy, both copper and iron deficiency have marked effects on the metabolism of the other metal (13).
Effect of altered iron status on copper metabolism
We have investigated whether alteration of iron status during pregnancy alters copper levels and whether the effects can be explained by changing levels of gene expression (36). In the mother, iron deficiency results in an increase in copper levels in the liver. This is associated with an increase in serum copper levels in the mother and in activity of ceruloplasmin in the maternal serum. Interestingly none of the genes associated with copper metabolism are changed at mRNA level (36).
In the placenta, as in the gut, iron efflux is associated with a copper oxidase (37). This protein is thought to oxidise Fe2+ to Fe3+ prior to incorporation into foetal transferrin (37). In iron deficiency, the activity increases in the placenta (38). This makes teleological sense because it will facilitate incorporation into foetal transferrin.
Within the placenta itself, in iron deficiency, copper levels rise to higher than any other tissue measured, including the maternal liver (36). This observation fits with the mechanisms for copper transfer suggested by Hardman et al. (39). Uptake is through a carrier mediated process (40), whereas efflux is through the ATPase, ATP7A. If this is the limiting step, one would expect a build up of copper within the placenta in order to avoid overload to the foetus. However, copper levels in the foetal liver are actually lower than controls. Quite why this should occur is not apparent.
Effect of altered copper status on iron metabolism
Present data on the effect of mild copper deficiency on maternal to foetal transfer of iron are contradictory. A study by Wapnir et al. (41) showed foetal liver iron levels increased in copper deficiency, whereas another study reported a decrease (42). Our own data supported the latter result (43). With the induction of deficiency, ceruloplasmin levels decrease in maternal serum (43). Ceruloplasmin is required for efflux of iron from the maternal liver, and patients with aceruloplasminemia (44), or ceruloplasmin knockout mice (45), have increased levels of iron in the liver. As might be expected, delivery of iron to the foetus is compromised. Serum iron levels drop and foetal liver iron levels also drop. Interestingly, the levels of iron within the placenta do not change, suggesting that the conceptus can protect the placenta at the expense of the developing embryo and foetus.
Expression data in the placenta suggest that it is responding to the indirectly induced iron deficiency in the same way as it does to a iron deficiency itself (43). mRNA levels of both transferrin receptor and the IRE-regulated form of DMT1 rise as copper decreases, whereas mRNA of the non-IRE regulated form and of ferroportin does not change (43). This has some intriguing implications, primarily that copper deficiency also induces iron deficiency.
Given the high incidence of maternal anaemia during pregnancy, and that many of those who are prescribed iron supplements show no significant improvement, it is tempting to conclude that at least some of the cases are a consequence of low copper. It would be interesting to perform a comparison between providing iron supplements only or iron together with copper.
Summary and conclusions
In this brief review, we have considered copper and iron uptake and transfer across the placenta, and how the two essential micronutrients interact. It is clear that there is a very tight regulation of transfer and, despite the research efforts that have been undertaken, it is also clear that we have much to learn. The demonstration that there is a significant interaction between the two makes it apparent that, when considering supplementing with, for example, iron, copper status must also be taken into account.
The authors acknowledge the support of the Scottish Government, the European Union and the International Copper Association. We are grateful to Dr Helen Hayes, Ms Lynne Beattie and the staff of the Animal Resources Unit for technical assistance in our work cited in this review. Grants: Scottish Executive Environment and Rural Affairs Department, EU FPVI (NuGO and EARNEST), International Copper Association.