Interactions between iron and fat in non-alcoholic fatty liver disease

Abstract

Non-alcoholic fatty disease (NAFLD) is estimated to affect approximately one billion people worldwide. However, the pathogenesis of this common disease is not well understood. This has hampered the development of effective treatments that might prevent complications such as liver failure, liver cancer and cardiac disease. Iron and mutations of the haemochromatosis gene (HFE) have long been considered to have pathogenic roles in the development of NAFLD although their significance and mechanisms of effect remain controversial. Recent data indicate a role for modest increases in iron in the generation of insulin resistance and adipose tissue dysfunction, both hallmark features of NAFLD.Modest increases in iron have been shown to occur with heterozygous HFE gene mutations. Due to their high prevalence, heterozygous HFE gene mutations are frequently observed in individuals with co-existent NAFLD and such mutations have been hypothesised to have a co-toxic role in NAFLD pathogenesis. Homozygous deletion of Hfe has been shown to lead to dysregulated lipid metabolism and liver injury in a mouse model of NAFLD. Chapter 3 describes a model of heterozygous deletion of the Hfe gene in mice fed a high calorie diet. This led to impaired homeostasis of both iron and glucose, although there were no alterations in hepatic lipid metabolism or liver injury.Chapter 4 seeks to further examine the links between iron and insulin resistance. Hepatic iron concentration (HIC) was measured using MR Ferriscan and serum concentrations of six adipokines were determined using a multiplex ELISA array from a cohort of 60 adults with NAFLD. After patients were randomised to six months of venesection and dietary advice (n=23) versus dietary advice alone (n=28), no differences were seen between groups in change in serum concentrations of adiponectin, leptin, resistin, retinol binding protein-4 (RBP-4), tumour necrosis factor alpha (TNFa) and interleukin-6 (IL-6). However, unexpectedly, a significant positive correlation between baseline HIC and serum adiponectin was found. This strengthened further after correction for age, gender and body mass index (rho=0.36, p=0.007). Furthermore, significant inverse correlations were identified between HIC and six surrogate measures of insulin resistance; Adipo-IR, serum insulin, serum glucose, HOMA-IR, HbA1C and hepatic steatosis. A positive correlation was noted between HIC and insulin sensitivity index. These data indicate that HIC positively correlates with serum adiponectin concentration and insulin sensitivity in patients with NAFLD, although at present, causality cannot be established.Adipose tissue dysfunction secondary to iron has been increasingly implicated in NAFLD pathogenesis. The remainder of the thesis focuses on adipose tissue iron homeostasis and its effects on adipose tissue biology. A model of adipocyte-specific ferroportin (Fpn1) knockout (FKO) was developed in order to selectively load in into adipocytes. This involved an Adipoq-Cre recombinase driven floxed deletion of Fpn1 in adipocytes (Chapter 5). This model demonstrated successful selective deletion of Fpn1, however this did not result in increased adipocyte iron stores. Similarly, FKO mice did not demonstrate evidence of adipokine dysregulation, or altered insulin sensitivity in hyperinsulinaemic-euglycaemic clamp studies. Likewise, there was no effect of adipocyte-specific ferroportin deletion on liver injury. This study has demonstrated that ferroportin is not a key determinant of adipocyte iron homeostasis in this model. Further studies are required in order to establish the key factors that regulate adipocyte iron homeostasis in animal models and humans.Finally, to test the hypothesis that iron modulates adipokine release, an in vitro model of adipocyte iron loading using differentiated Simpson-Golabi-Behmel Syndrome (SGBS) pre-adipocytes was developed (Chapter 6). Using a proteomic approach of stable isotope labelled amino acids in cell culture (SILAC), the effects of iron on the human adipocyte secretome were examined. This identified 60 proteins with a greater than two-fold change in secretome concentration and pl0.05 with iron treatment. Of these, it was found that secreted apolipoprotein E (ApoE) was reduced by 58% (p=0.001) and 76% (p=0.007) by SILAC and western blot respectively. Intracellular ApoE levels were increased more than 11-fold (p=0.0005) with iron treatment. These findings suggest that secretion of ApoE from adipocytes is inhibited by iron. This may be highly relevant with regard to NAFLD pathogenesis given the observation that ApoE knockout leads to steatohepatitis in mice fed a Western diet. Moreover, the broad range of proteins with differential secretion secondary to iron provides a platform for multiple future studies.In conclusion, this thesis describes a number of novel links between iron, insulin sensitivity and adipokine regulation relevant to NAFLD pathogenesis. In particular, impaired glucose homeostasis in a mouse model of heterozygous Hfe deletion was observed. Notably consistent links between HIC and multiple measures of insulin sensitivity were seen in a human cohort of patients with NAFLD. Finally the effect of iron on the human adipocyte secretome was examined, identifying profound inhibition of ApoE secretion by iron. These findings may offer opportunities for the development of effective new therapies for the treatment of NAFLD.