Genetic models

Its genetic similarity to humans and its possibilities for genetic manipulation makes the rodent an attractive model. Two iron-deficient rodent models, microcytic anaemia (mk) mice and b rats, carry a glycine-to-arginine substitution at position 185 in the iron transporter gene Slc11a2 (encoding DMT1)^{(Reference Thompson, Molina and Donaghey38,Reference Canonne-Hergaux, Fleming and Levy58,Reference Jiang, Ranganathan and Lu59)} . The G185R mutation in DMT1 affects targeting of the protein to the brush border of the intestinal epithelium, the normal physiological site for iron uptake. All mk mice and b rats exhibit systemic, severe microcytic, hypochromic anaemia, with impaired intestinal iron absorption and defective iron metabolism in peripheral tissues^{(Reference Thompson, Molina and Donaghey38,Reference Canonne-Hergaux, Fleming and Levy58,Reference Jiang, Ranganathan and Lu59)} .

Donovan et al.^{(Reference Donovan, Lima and Pinkus60)} created two conditional knockout mouse models: Meox2-cre;Fpn ^flox/flox and vil-Cre-ER ^T2; Fpn ^flox/flox mice. They found that Fpn-deficient animals suffer from IDA, but accumulate iron in enterocytes, macrophages and hepatocytes^{(Reference Donovan, Lima and Pinkus60)}, consistent with the key role that Fpn plays in these cell types. The gut-specific inactivation of Fpn confirmed that it is critical for intestinal iron absorption. These studies have identified the main sites of Fpn activity and provide insight into haemochromatosis.

Effective iron export by Fpn requires the participation of two copper-dependent ferroxidases hephaestin (Heph) and ceruloplasmin (Cp). The mislocation of Heph in the sex-linked anaemia (sla) mouse leads to a defect in the release of iron from intestinal enterocytes into the circulation; in other words, higher cellular iron levels in intestinal enterocytes but lower systemic iron levels are present in these mice^{(Reference Chen, Su and Attieh61,Reference Kuo, Su and Chen62)} . Despite a deletion removing 192 amino acids, sla mice still produce a truncated Heph protein, but its mislocation from the basolateral membrane makes it unable to interact with laterally located Fpn, which contributes to the observed iron transport defects in sla mice^{(Reference Kuo, Su and Chen62)}. Newborn sla mice show IDA with low haemoglobin levels^{(Reference Kingston, Bannerman and Bannerman63)}. Recent studies have found that mice lacking Heph in the whole body (Heph ^−/−) and in the intestine alone (Heph ^int/int), mice with double knockout of Heph and Cp (Heph ^−/−Cp ^−/−) and mice with intestine-specific Heph deletion (Heph ^int/intCp ^−/−) exhibit dietary iron absorption defects and disrupted iron homeostasis, with serious systemic iron deficiency^{(Reference Fuqua, Lu and Frazer64,Reference Xu, Chen and Zheng65)} .

The study by Nicolas et al.^{(Reference Nicolas, Bennoun and Porteu66)} was a milestone in research on the molecular basis of anaemia. They found that transgenic mice with overexpressed hepcidin develop severe microcytic hypochromic anaemia. These transgenic animals are born with decreased body iron levels, reduced body size, pallor, and hairless and crumpled skin, which are related to erythrocyte abnormalities^{(Reference Nicolas, Bennoun and Porteu66)}. Model animals that overexpress hepcidin^{(Reference Nicolas, Bennoun and Porteu66)} or that lack the peptide^{(Reference Nicolas, Bennoun and Devaux67)} (upstream stimulatory factor 2 (Usf2)-knockout mice accompanied by iron overload, as mentioned in the section ‘Animal models of iron overload and their features’) are valuable tools for studying iron homeostasis in vivo and for deciphering the molecular mechanisms of hepcidin’s action as a major regulator of systemic iron homeostasis.

The lack of hepcidin repressor matriptase-2 (encoded by Tmprss6) results in increased hepcidin levels, triggering iron deficiency and anaemia^{(Reference Folgueras, de Lara and Pendás68,Reference Willemetz, Lenoir and Deschemin69)} . Tmprss6 ^−/− mice model the human disorder iron-refractory IDA. The anaemic phenotype of these mutant mice results from the blockage of intestinal iron export into plasma^{(Reference Folgueras, de Lara and Pendás68)}. Based on these findings, the authors conclude that matriptase-2 activity represents the first significant step in hepcidin regulation and iron homeostasis.

The hypotransferrinaemic mouse is a model of inherited transferrin deficiency^{(Reference Trenor, Campagna and Sellers70,Reference Bu and Bartnikas71)} . Two of its best-known phenotypes are severe IDA and massive tissue iron overload^{(Reference Trenor, Campagna and Sellers70,Reference Bu and Bartnikas71)} .

These rodents offer a unique model for studying both iron deficiency and excess. Although haematopoietic cells are severely iron deficient, most, if not all, other tissues show marked iron overload. These animals complement other model animals, including mk mice, b rats and sla mice, for the study of iron homeostasis. Although these models are generally studied on adult animals, in some cases the consequences of genetic mutations have been studied in early postnatal period, as we show in Table 1. We believe that these models have the potential to study iron metabolism and regulation in the neonatal period. However, future research is needed to focus on consequences in neonatal periods.

Dietary iron insufficiency

A low-iron diet for the animals is typically required in this nutritional iron deficiency model. Suckling chinchilla rabbits receiving only their mothers’ milk (without extra iron supplementation) from PND 15 to PND 40 showed a continuous decrease in plasma iron concentration, blood haemoglobin concentration, erythrocyte count and mean corpuscular volume (MCV) and subsequently developed pure microcytaemia^{(Reference Holter and Refsum72)}. An iron-deficient diet (4 ppm Fe) was given to Wistar rats for 32 d beginning at PND 0 to simulate anaemia^{(Reference DeMaman, Melo and Homem73)}. The rat data indicate that iron is essential for maintaining optic nerve cell structure and that structural damage from iron deficiency is not easily reverted by iron repletion^{(Reference DeMaman, Melo and Homem73)}. A large animal model found that 33% of housed lambs had haemoglobin values below 80 g/l (regarded as anaemia) at 3 weeks postpartum if no iron was supplemented^{(Reference Vatn and Framstad74)}.

In addition, pigs are promising animal models for iron disorder research. Owing to their lowest hepatic iron stores among all mammalian species, high iron requirements for fast growth, limited external supply caused by low iron content in the sow’s milk, and immature mechanisms of iron absorption, suckling piglets easily develop systemic IDA without iron supplementation^{(Reference Szabo and Bilkei75,Reference Svoboda and Drabek76)} . Research on neonatal piglets fed a low-iron diet demonstrated that iron deficiency had a negative impact on the development of brain white matter^{(Reference Leyshon, Radlowski and Mudd77)}, peripheral immunity^{(Reference Leyshon, Ji and Caputo78)}, the serum metabolome^{(Reference Dong, Wan and Yang79)}, hepatic histology^{(Reference Dong, Wan and Yang79)}, the development of colonic microbiota^{(Reference Knight, Wang and Donovan80)}, and volatile fatty acids concentrations^{(Reference Knight, Wang and Donovan80)}. A 28-d study of anaemic piglets showed a significant decrease in their haemoglobin concentration, haematocrit level and MCV in blood and hepatic non-haem and haem iron content, along with higher mRNA expression levels of renal erythropoietin and erythroferrone in the bone marrow^{(Reference Jończy, Mazgaj and Starzyński81)}. An interesting aspect of the study was that iron was found to be retained in the duodenum of anaemic piglets, with down-regulation of copper-related genes and decreased FPN and HEPH expression, which indicated that the regulation of intra-enterocytic iron absorption was probably related to the interaction between copper and iron metabolism proteins^{(Reference Jończy, Mazgaj and Starzyński81)}. Similarly, results from studies of iron-deficient rats given an iron-reduced diet at various stages of postnatal development (at 8 and 21 d; 6, 9, 12 and 36 weeks; and 1 year of age) have also shown that copper plays important role in overall body iron homeostasis^{(Reference Ravia, Stephen and Ghishan82)}. Chen et al.^{(Reference Chen, Huang and Su83)} also showed lower Heph and Cp ferroxidase activity and hepcidin expression but higher enterocyte mRNA and protein levels of Fpn1 in young (6-week-old) copper-deficient mice with systemic iron deficiency. This study indicates that copper is required for the proper processing and/or stability of Heph.

Neonatal iron imbalance due to maternal iron deficiency may also serve as a model. Pregnancy frequently leads to maternal iron deficiency in humans due to the increased demand for iron, which may have serious consequences for the offspring. Some studies have assumed that poor maternal iron status during pregnancy does not affect foetal or neonatal iron endowment, but severe iron deficiency does^{(Reference Raffaeli, Manzoni and Cortesi46,Reference Lönnerdal, Georgieff and Hernell84)} . Infants from moderately and severely anaemic mothers have lower birth weights and iron stores, placing them at higher risk of iron deficiency at an early age^{(Reference Lönnerdal, Georgieff and Hernell84)}. In an experimental rat model, female rats were given an iron-deficient diet (5.2 ppm Fe) for 6 weeks to produce iron-deficient pups^{(Reference Moos, Skjørringe and Thomsen85)}. They found that these pups had a significant (40%) reduction of iron level in the brain compared with pups from control mothers^{(Reference Moos, Skjørringe and Thomsen85)}. Similarly, in pregnant Sprague-Dawley rats fed a 3–6 ppm Fe diet, the brain iron concentration in the offspring was 60% lower in PND 7 rats and 19% lower in PND 28 rats, and the iron concentrations were lower in the liver and heart of PND 7 rats^{(Reference Rao, Tkac and Townsend86)}. This animal model also showed that iron deficiency decreased brain-derived neurotrophic factor function, which may be related to learning deficits^{(Reference Tran, Fretham and Carlson87)}. Earlier studies used guinea pigs because their brain growth spurt and myelogenesis, like in humans, occur prenatally, that is, earlier than in rats and mice^{(Reference Leblanc, Surette and Fiset88,Reference Shero, Fiset and Plamondon89)} . Leblanc et al.^{(Reference Leblanc, Surette and Fiset88)} showed that maternal iron deficiency led to lower levels of haemoglobin, haematocrit and MCV in guinea pig pups and that it was associated with a modified fatty acid profile in the offspring’s brain. Rhesus monkeys are also an excellent model in that they mimic the developmental risk factors for IDA in humans, given their similarities to humans in gestational physiology and brain development. Compromised iron status was seen in newborn rhesus monkeys when their mothers suffered from iron deficiency, and the newborns easily developed iron-deficient anaemia by 6 months postpartum if iron supplementation did not meet their growth-related needs^{(Reference Lubach and Coe90,Reference Sandri, Lubach and Lock91)} .

Foetuses dissected from the uterus can reflect the iron condition of infants in the prenatal period, and as such, they can be a model for premature infants. In a study from our laboratory, an experiment was performed using pregnant mice fed an iron-deficient diet containing 4 ppm Fe, and foetuses were dissected at embryonic day 18.5^{(Reference Mazgaj, Lipiński and Edison92)}. The foetuses from the anaemic mothers were found to have lower hepatic non-haem iron levels, drastic reductions in ferritin light chain levels, barely detectable hepcidin mRNA expression and a greater decrease in hepatic Fpn levels compared with controls^{(Reference Mazgaj, Lipiński and Edison92)}. Previous work has also reported different strains of rats, such as Rowett Hooded Lister^{(Reference Gambling, Charania and Hannah93,Reference Cornock, Gambling and Langley-Evans94)} , Wistar^{(Reference Cornock, Gambling and Langley-Evans94)} and Sprague-Dawley^{(Reference Crowe, Dandekar and Fox95)}, for studying foetal growth. Cornock et al.^{(Reference Cornock, Gambling and Langley-Evans94)} compared differences in foetal and maternal iron homeostasis between Wistar and Rowett Hooded Lister rats fed a low-iron diet prior to and during gestation. It is worth noting that, although lower iron content was found in both rat strains, there were considerable differences between them^{(Reference Cornock, Gambling and Langley-Evans94)}. Furthermore, understanding the mechanisms that govern iron homeostasis in the mother–placenta–foetus unit could aid in managing iron disorders during pregnancy. Sangkhae et al.^{(Reference Sangkhae, Fisher and Wong52)} used female and male heterozygous iron-regulatory protein (IRP) 1/hepcidin mice to generate placentas and embryos lacking IRP1/hepcidin or not, and they investigated the effects of maternal iron status on placental and foetal iron homeostasis. They confirmed, from the iron deficiency induced by a low-iron diet, that placental iron homeostasis is mediated by the iron-responsive elements/IRP system. The homeostatic response of the metabolically active placenta is to retain iron, which protects the placenta from severe iron deficiency and consequent reduced oxidative phosphorylation. Although this placental iron retention comes at the cost of foetal iron deficiency, it may ultimately protect the foetus from the more serious adverse effects of wider placental dysfunction.

In general, scientific studies using low iron intake in the diet both during pregnancy and postpartum indicate the key role of iron ions in physiological functions. The results underscore the importance of a careful approach when designing the route and regimen of iron supplementation for treating IDA in infants.

Phlebotomy

Taking blood samples can also cause a loss of total body iron and can induce anaemia (that is, phlebotomy-induced anaemia (PIA)). Animal models of PIA can be used to investigate the health status of full-term and preterm neonates who experience significant iron loss from frequent phlebotomy procedures. Serial-timed phlebotomy in neonatal mice led to abnormalities in the content and morphology of erythrocytes, as well as their haemoglobin, reticulocyte properties and immune cell responses – findings that are similar to those in humans^{(Reference Chung, Desiraju and Namachivayam96)}. Another study used a mouse model of PIA with adequate neurological development and haematocrit levels comparable to those found in premature infants in the neonatal intensive care unit to investigate the effect of anaemia on hippocampal neurochemistry and gene expression^{(Reference Wallin, Tkac and Stucker97)}. Once again, it was shown that, as haematocrit levels decrease in newborn mice with PIA, the level of phosphoethanolamine, the precursor of myelin, also decreases^{(Reference Wallin, Tkac and Stucker97)}. A recent study found the side effect of PIA in hippocampal function and showed that consequences of PIA may be based in sex-specific transcriptomic alterations^{(Reference Singh, Wallin and Abrahante Lloréns98)}.

In addition, negative iron balance was found in lambs with PIA^{(Reference Zamora, Guiang and Widness7)}. The risk of neonatal brain iron deficiency due to repeated phlebotomy is real, as iron is prioritised for erythrocyte production over brain accretion^{(Reference Zamora, Guiang and Widness7)}. Rydal et al.^{(Reference Rydal, Bhattarai and Nielsen99)} developed an experimental model of IDA in sows and offspring induced by blood removal combined with an iron-reduced diet during gestation. Piglets from the blood-removal group had significantly lower erythrocyte counts and haematocrit, as well as a tendency to have lower haemoglobin levels^{(Reference Rydal, Bhattarai and Nielsen99)}. Moreover, recent epidemiologic research indicates that severe anaemia may be an essential risk factor for the development of necrotising enterocolitis^{(Reference Arthur, Nalbant and Feldman100)}. Using a preclinical neonatal murine model of PIA, Arthur et al.^{(Reference Arthur, Nalbant and Feldman100)} showed that anaemia can enhance intestinal inflammation and barrier disruption, possibly by altering macrophage function, resulting in a type of predisposing intestinal injury that may increase the risk for necrotising enterocolitis. These observations may be the basis for further preclinical studies on the role of anaemia in common comorbidities.

Caesarean section

Caesarean section is a common surgical procedure for delivering full-term and preterm neonates. A systematic review and meta-analysis of full-term neonates showed that, compared with neonates born vaginally, those born by caesarean section had lower levels of haematocrit, haemoglobin and erythrocytes^{(Reference Zhou, Li and Zhu51)}. Similar results were observed in premature neonates^{(Reference Wu, Chou and Chen50)}, indicating that neonates delivered by caesarean section might be more likely affected by IDA in infancy. In rodent models, caesarean section is performed either at term (on the day of birth) or preterm (1 d before birth)^{(Reference Boksa and Zhang101–Reference Chiesa, Rabiei and Riffault103)}. It is important to note that premature mice could not survive if they were delivered 1 or 2 d earlier than spontaneous parturition^{(Reference Stelloh, Allen and Mattson102,Reference Loctin and Delost104)} .

Larger animals enable approaches that are not feasible in rodents and, thus, serve as important translational models to bridge the gap between rodent model studies and human clinical studies. Sheep have been a relevant and valuable species for modelling preterm birth. Not only is the sheep’s gestation period closer to that of humans, but their relatively large size allows specific foetal or pregnant sheep surgery. However, the difference in the stomach between ruminant and monogastric species such as pigs and humans limits the relevance of findings from sheep studies with regard to digestive tract development. Piglets are becoming popular animals for research on preterm offspring. Unlike rodents and sheep, the main advantage of the preterm pig is its similarity to the human baby in terms of size and the impaired respiratory, nutritional, immunological and metabolic responses after preterm birth^{(Reference Sangild, Thymann and Schmidt105)}. Preterm piglets present severe symptoms of prematurity, such as hepatic, digestive, locomotive and thermoregulatory immaturity^{(Reference Sangild, Thymann and Schmidt105,Reference Eiby, Wright and Kalanjati106)} . Eiby et al.^{(Reference Eiby, Wright and Kalanjati106)} established a pig model of the preterm neonate, in which piglets were delivered by caesarean section at six gestational ages, that is, at days 91, 94, 97, 100, 104 and 113 (term day 115). This experiment showed that the degree of organ development and maturation is proportionate to age and that only the late premature piglets had high survival rates^{(Reference Eiby, Wright and Kalanjati106)}. An earlier study found that there was a markedly low level of haemoglobin, erythrocytes, MCV, mean cell haemoglobin and especially packed cell volume (also known as haematocrit) in piglets delivered by caesarean section on the 111th or 112th day of pregnancy until treatment with iron^{(Reference Holter, Framstad and Aulie107)}. In addition, in our ongoing project, we are comparing the differences between premature piglets delivered at the gestational age of 109 d and full-term piglets; we found that premature piglets have lower body weights, erythrocyte parameters, plasma iron levels and transferrin saturation (TSAT)^{(Reference Wang, Lenartowicz and Mazgaj108)}. However, the most important aspect of producing a caesarean section piglet model is ensuring that it best represents human newborns. The pig model, thoroughly validated in terms of iron metabolism, can be used for diagnostic purposes, as well as to test new drugs or supplements. So far, however, few studies have focused on iron deficiency using larger animal models because working with these animals is time-consuming and expensive; feeding is a major limiting issue; and most importantly, the premature animals die easily.

Transfusion

Due to extravascular intramedullary haemolysis, people with TM require regular erythrocyte transfusions to survive, leading to systemic iron overload^{(Reference Origa56)}. For neonatal studies, a preclinical transfusion-dependent humanised mouse model of TM was generated by targeted gene replacement of the mouse adult globin genes in embryonic stem cells with a human γβ⁰ globin gene-switching cassette^{(Reference Huo, McConnell and Ryan109)}. The fully humanised TM mice were born alive, surviving solely upon human HbF and minor amounts of HbA2, and became progressively anaemic upon completion of the foetal-to-adult Hb switch; the majority died before weaning age. Weekly hypertransfusion therapy starting 2 d after birth inhibited erythropoiesis in the bone marrow, but elevated hepatic iron stores^{(Reference Huo, McConnell and Ryan109)}. In addition, different adult animal models have been applied for transfusion studies, including mouse^{(Reference Youssef, Rebbaa and Pampou110)}, rat^{(Reference d’Almeida, Jagger and Duggan111)}, dog^{(Reference Berndt, Buttenberg and Graw112)}, sheep^{(Reference Simonova, Tung and Fraser113)}, pig^{(Reference Jensen, Nielsen and Simonsen114)} and marmoset^{(Reference Yurimoto, Mineshige and Shinohara115)}. However, to date, few studies have been conducted on early postnatal iron metabolism using neonatal models. Given the differences between adults and newborns, it is imperative that relevant neonatal animal models of transfusion be developed to investigate adverse transfusion reactions.

Genetic model

GRACILE syndrome is a neonatal, lethal metabolic disorder with iron overload^{(Reference Visapää, Fellman and Vesa116)}. Affected infants present severe growth retardation and iron overload signs, including hepatic haemosiderosis, increased SF concentrations, hypotransferrinaemia with increased TSAT, and free plasma iron^{(Reference Visapää, Fellman and Vesa116)}. A transgenic mouse model with a c.232A>G mutation in Bcs1l has been developed to mimic GRACILE syndrome^{(Reference Levéen, Kotarsky and Mörgelin117)}. However, only small amounts of iron are detected in the liver of Bcs1l ^G/G mice, while liver iron accumulation is observed both at gestation and at birth stages in individuals with GRACILE syndrome^{(Reference Levéen, Kotarsky and Mörgelin117)}. Levéen et al.^{(Reference Levéen, Kotarsky and Mörgelin117)} speculated that this disparity may be related to differences in placental iron transport between mice and humans.

DGUOK deficiency is the most common genetic factor of human mitochondrial DNA depletion syndromes, a group of severe, phenotypically heterogeneous, recessively inherited disorders. Patients with DGUOK mutations show hepatic iron overload, which may progress to liver failure^{(Reference Pronicka, Węglewska-Jurkiewicz and Taybert118)}. Thus far, several animal models of these syndromes have been described, including DGUOK-deficient strains of mouse^{(Reference Zhou, Curbo and Zhao119)}, rat^{(Reference Bennett, Helbling and Meng120)} and zebrafish^{(Reference Munro, Horvath and Müller121)}, in addition to a DGUOK mutant mouse line with a phenylalanine-to-serine substitution of residue 180^{(Reference Borreguero, Wueest and Hantel122)}, but no iron disorder studies have been conducted with these models. Guo et al.^{(Reference Guo, Duan and He123)} established an in vitro liver disease model of liver organoids and hepatocytes developed from the pluripotent stem cells of patients with DGUOK mutation who had died at 2 months of age, along with associated controls and isogenic cell lines corrected by the CRISPR/Cas9 system. They used ferric ammonium citrate to induce iron overload and found the DGUOK mutant model to be more sensitive to iron overload-induced ferroptosis – which is mediated by nuclear receptor co-activator 4-dependent degradation of ferritin in lysosome and cellular labile iron release – than control^{(Reference Guo, Duan and He123)}.

Patients affected by TI do not require chronic blood transfusions for survival. However, non-transfused patients show hyperabsorption of iron in the gut due to dysregulation of hepcidin transcription^{(Reference Origa56)}. Two mouse models of TI are available: Hbb^{th1/th1(Reference Skow, Burkhart and Johnson124)} and Hbb^{th3/+(Reference Gardenghi, Marongiu and Ramos125)}. Adult Hbb^th1/th1 and Hbb^th3/+ mice exhibit signs of severe disease, such as hepatosplenomegaly, anaemia and aberrant erythrocyte morphology, which are comparable to those of patients affected by TI^{(Reference Gardenghi, Marongiu and Ramos125)}. Using the Hbb^th3/+ model, researchers have found that ablating the hepcidin regulator erythroferrone^{(Reference Kautz, Jung and Du126)} or the oral Fpn inhibitor vamifeport^{(Reference Kalleda, Flace and Altermatt127)} and injecting the mice with the haem oxygenase inhibitor tin protoporphyrin IX^{(Reference Garcia-Santos, Hamdi and Saxova128)} can ameliorate anaemia, ineffective erythropoiesis and dysregulated iron homeostasis. Unfortunately, mice completely lacking adult β-globin genes (Hbb^th3/th3) die late in gestation. A mouse model of TM was established by transplanting haematopoietic foetal liver cells harvested from Hbb^th3/th3 embryos at 14.5 d of gestation into lethally irradiated syngeneic adult recipients^{(Reference Gardenghi, Marongiu and Ramos125,Reference Rivella, May and Chadburn129)} . In addition, as noted above, the humanised murine model of TM is also useful for the development of new transfusion and iron chelation regimens^{(Reference Huo, McConnell and Ryan109)}, as well as preclinical trials of new drugs that could improve the quality of life of patients. Recently, Yang et al.^{(Reference Yang, Kang and Hu130)} generated a β-thalassaemia rabbit model using CRISPR/Cas9-mediated genome editing and found marked iron deposits in the rabbit spleen, liver and kidney.

Hereditary haemochromatosis (HH) is a common autosomal recessive disorder and the most common cause of iron overload in adults^{(Reference Murray and Kowdley131)}. Most people do not experience symptoms until old age, because patients need many years to build up the excess iron that causes the disease^{(Reference Murray and Kowdley131,Reference Whitington132)} . Although the classic haemochromatosis phenotype has been rarely reported early in life, neonatal screening for HH offers an opportunity for early clinical intervention, preventing the morbidity and mortality associated with HH. Four types of HH have been identified: type I haemochromatosis due to homeostatic iron regulator (HFE) mutation; type II haemochromatosis due to haemojuvelin (HJV) and hepcidin mutations (juvenile haemochromatosis); type III haemochromatosis related to transferrin receptor 2 (TfR2) mutation; and type IV haemochromatosis (Fpn disease). The most common HH subtype, highly prevalent in the white population, presents with an adult-onset and chronic phenotype and is caused by the HFE p.C282Y mutation^{(Reference Muckenthaler, Marques and Colucci133)}. The non-HFE HH subtypes are relatively rare and typically manifest during the juvenile period^{(Reference Muckenthaler, Marques and Colucci133)}. In juvenile HH, symptoms may appear in the second decade of life and are associated with mutations in the genes encoding hepcidin, HJV and phosphatidylinositol glycan anchor biosynthesis class A (PIGA), an enzyme involved in glycosylphosphatidylinositol (GPI)-anchor biosynthesis^{(Reference Muckenthaler, Marques and Colucci133)}. The lack of GPI synthesis results in improper processing of HJV and neuronal Cp^{(Reference Muckenthaler, Marques and Colucci133)}. Lack of a GPI anchor causes diminished HJV on the cell surface and reduced hepcidin synthesis^{(Reference Muckenthaler, Marques and Colucci133)}. Most importantly, signs of PIGA-mediated juvenile haemochromatosis have been observed as early as 2 years of age^{(Reference Muckenthaler, Marques and Colucci133)}. There are currently many mouse models for HH, including HFE ^{−/−(Reference Zhou, Tomatsu and Fleming134)}, HJV ^{−/−(Reference Huang, Pinkus and Pinkus135)}, TfR2 ^{Y245X(Reference Fleming, Ahmann and Migas136)}, TfR2 ^{−/−(Reference Wallace, Summerville and Lusby137)}, Fpn ^{ffe(Reference Zohn, De Domenico and Pollock138)} and USF2 ^{−/−(Reference Nicolas, Bennoun and Devaux67)}.

Dietary iron overload

Iron loading also potentially results from excessive intake of dietary iron, although this is uncommon in human infants. However, elevated iron supplementation has been widely used in animal models to imitate human iron overload. Excessive iron supplementation in early life causes hepatic and hippocampal iron overload, a decrease in duodenal DMT1 mRNA expression and drastic increases in hepatic hepcidin mRNA and ferritin heavy chain protein expression, as well as deficits in social novelty recognition in nursing term piglets^{(Reference Ji, Lönnerdal and Kim139)}. Despite high hepcidin expression, researchers have found that both mRNA and protein expression of Fpn1 in the duodenal mucosa and liver were unaffected by iron provision or birth weight^{(Reference Ji, Lönnerdal and Kim139,Reference Ji, Doan and Wang140)} . Authors speculated that hepcidin-induced Fpn1 degradation is hyporesponsive to iron excess in a nursing pig model^{(Reference Ji, Lönnerdal and Kim139,Reference Ji, Doan and Wang140)} . In a Wistar rat model used for investigating the effect of iron on the brain, the results showed that postnatal iron administration induces brain oxidative damage and memory deficits^{(Reference de Lima, Polydoro and Laranja17,Reference Schröder, Fredriksson and Vianna141)} . Guo et al.^{(Reference Guo, Zhang and Liu142)} designed dietary high-iron, medium-iron and low-iron groups of pregnant female mice and found that maternal high-iron supplementation affected reproductive performance and increased foetal iron levels at day 18.5. In another experiment, pregnant mice starting at 2 weeks of gestation and their offspring up to 30 weeks were supplied with excessive dietary iron^{(Reference Kim, Kim and Min143)}. Their data suggested that prolonged exposure to high iron levels from the foetal stage to adulthood alters lipid metabolism^{(Reference Kim, Kim and Min143)}.