Journal of Gastroenterology and Hepatology (1999) 14, 745–749

drugs; for example, the toxicity of antituberculous drugsis exacerbated by other hepatotoxins.5

The progressive iron loading that occurs in hered-itary haemochromatosis leads to hepatic fibrosis,cirrhosis and hepatocellular carcinoma.6 In haemochro-matosis, alcohol has been shown to be a cofactor in thedevelopment of fibrosis and cirrhosis7 as has chronicviral hepatitis.8 Could the reverse be true? Could adegree of iron loading, short of that of haemochro-matosis, act as a damaging agent in liver diseases ofother aetiology?

CHRONIC VIRAL HEPATITIS

It is emerging that iron is a cofactor in the liver diseaseof chronic viral hepatitis.9 For example, patients withchronic hepatitis C who have higher hepatic iron storesmay be less likely to respond to interferon than patientswith lower hepatic iron stores10 and exhibit increasedhepatic inflammation.11 In some patients with chronichepatitis C, phlebotomy alone can improve or even nor-malize liver function tests.12 Furthermore, it has longbeen recognized that chronic liver disease, of whateveraetiology, may exhibit excess iron in the liver, but

INTRODUCTION

It is a basic tenet in the medical care of patients withliver disease that agents which damage the liver act inconcert. Thus, for example, even before clear evidenceemerged, it was universally recommended that patientswith any form of liver disease should severely moderatetheir alcoholic intake. We have now reached the situa-tion in which evidence has accumulated to provide ascientific basis for this practice.

Patients with hepatitis B who take alcohol have liverfunction tests that are more severely abnormal thanthose with a similar intake but without the virus.1

Patients with liver disease secondary to hepatitis B showearlier development of cirrhosis and hepatocellular car-cinoma if they take alcohol.2 For patients with hepati-tis C, there is evidence that the virus and alcoholtogether are more damaging to the liver than eitheralone. For example, liver function tests in such patientsimprove less with abstinence than in the absence ofhepatitis C.3 Recent, careful study of the progress ofliver disease in patients with hepatitis C has shown thatalcohol, along with male gender and age are factors thatcan be clearly related to the development of fibrosis.4

The same principle applies to liver damage caused by

REVIEW

The haemochromatosis gene: A co-factor for chronic liverdiseases?

DK GEORGE, LAWRIE W POWELL AND MS LOSOWSKY

Clinical Sciences Unit, Queensland Institute of Medical Research and The University of Queensland,Brisbane, Australia

Abstract There is increasing evidence that hepatotoxins, such as alcohol and the hepatitis viruses, actas co-factors in causing hepatic fibrosis and cirrhosis. For example, alcohol aggravates the hepatic damageproduced by iron in hereditary haemochromatosis.We present evidence that the reverse is also true, thatis, that iron loading of mild to moderate degree due to heterozygosity or homozygosity for the haemochro-matosis genetic mutations acts as a significant hepatotoxin aggravating hepatic damage from other causesof liver disease.These include non-alcoholic steatohepatitis, chronic hepatitis C, porphyria cutanea tardaand possibly primary liver cell cancer. However, any additional hepatotoxic effect is due to the hepatic ironconcentration and not the mutations in the haemochromatosis genes.© 1999 Blackwell Science Asia Pty Ltd

Key words: cirrhosis, fibrosis, haemochromatosis, heterozygote, HFE, iron.

Correspondence: Lawrie W Powell, Queensland Institute of Medical Research, The Bancroft Centre, PO Royal Brisbane Hospital, Brisbane, QLD 4029, Australia. Email: <lawrieP@qimr.edu.au>

Accepted for publication 2 March 1999.

usually not to the extent seen in homozygoushaemochromatosis.13 The cause of such increases inhepatic iron stores has previously not been established.It has, in one view, been attributed to the effects ofalcohol,14 but the subject is now regarded as more openand continues to attract attention.15 There is evidenceof increased hepatic iron deposition associated with cir-rhotic liver disease,15 perhaps in some way consequentfrom spontaneous or surgical portocaval shunting.16

However, the reason why this occurs in some patientsbut not others has, so far, escaped elucidation andwhether this increased hepatic iron results from or con-tributes to progressive hepatic fibrosis has not previ-ously been established.

In considering a problem in human disease it is alwayshelpful to have an animal model or some confirmationof a hypothesis in other species. Fortunately, such is thecase here. For example, the concerted action of differenthepatotoxins and iron is seen in the rat in the combinedeffects of carbon tetrachloride and iron,17 carbon tetra-chloride, alcohol and iron,18 alcohol, high-fat diet andiron19 and in the conjunction of copper and iron.20 In allthese rat models iron exacerbates hepatic fibrosis andeven, in the latter model, hepatocellular carcinoma.There are sound reasons to believe that such interac-tions might also occur in human liver diseases mediatedthrough increased lipid peroxidation and subsequentincreased collagen production. Increased lipid peroxida-tion is apparent in chronic hepatitis C,9 alcoholic liverdisease21 and iron overload21,22 with iron also catalysingas well as causing lipid peroxidation reactions.23

HAEMOCHROMATOSIS GENE

The recent cloning of the haemochromatosis gene24

HFE, and the identification of the common C282Ymutation responsible for most cases24–26 has provided an opportunity to re-examine the situation. It has long been known that putative heterozygotes for haemo-chromatosis (based on HLA typing within an affectedfamily) can have a degree of excess iron accumulationshort of that which produces the full disease. Up to 25%of healthy heterozygotes, based on HLA typing, haveabnormal iron indices (i.e. increased serum ferritinlevels and/or transferrin saturation27,28) and liver biop-sies in these subjects demonstrate hepatic iron concen-trations significantly greater than controls but less thanhomozygotes.28,29 On its own, this mild increase inhepatic iron stores is currently thought to be benign,but this may not be the case in the presence of otherliver diseases. Heterozygosity for haemochromatosis hasbeen suggested as a factor in other iron-loading condi-tions such as sideroblastic anaemia30 and PCT, but theheterozygote contribution could not previously befirmly established. It is now possible to investigate therole of the haemochromatosis mutation in the develop-ment of these diseases. Heterozygosity or homozygosityfor the C282Y haemochromatosis mutation has nowbeen shown to be present in almost half the cases ofPCT in the UK,31 Australia32 and the USA,33 a signifi-cantly higher prevalence than in control populations.

746 DK George et al.

Presumably the C282Y mutation is a precipitatingfactor for PCT consequent on the increased hepaticiron stores that are caused by this mutation, althoughthis has not yet been established. Furthermore, theeffect of the C282Y mutation and hepatic iron storeson liver damage in these patients has not been assessed.A second mutation, H63D, is also found on haemo-chromatosis chromosomes that do not contain theC282Y mutation.24 The H63D mutation appears tohave a minor role in iron overload although it may besignificant in some cases when associated with a C282Ymutation (i.e. compound heterozygotes).24,34This muta-tion may also play a role in the development of PCT incertain populations as a significantly greater proportionof Italian patients with PCT possess the H63D muta-tion than controls.35

Does a usually benign condition, such as heterozy-gosity for haemochromatosis, actually worsen liverdamage in coexisting liver disease and, if such an effectoccurs, is it an indirect effect through increased hepaticiron stores or due to a direct, perhaps immunomodula-tory effect of this major histocompatibility complex class1-like gene? The demonstration of increased prevalenceof the HFE mutations in a fibrotic liver disease com-pared with a control population would support a director indirect association between these mutations and theprogression of fibrosis in these diseases. However, theabsence of a statistically significant difference does notexclude an association if any effect of the HFE muta-tions is mediated through increased iron stores. As onlyup to 25% of heterozygotes may develop increasedhepatic iron stores and other causes of increased ironstores clearly occur,24,26,36 relatively small study popula-tions may not reveal an indirect effect of HFE mutationsas a significantly increased gene prevalence.

NON-ALCOHOLICSTEATOHEPATITIS

We have shown that the C282Y haemochromatosismutation is a cofactor in liver damage in patients withnon-alcoholic steatohepatitis (NASH).37 We found theprevalence of the C282Y mutation to be significantlyincreased in patients with NASH compared with a largecontrol population and NASH patients with this muta-tion also had significantly higher iron stores thanpatients without C282Y. More importantly, we demon-strated that increased iron stores, even iron stores at theupper end of the currently accepted normal range,usually but not always caused by the C282Y mutation,were associated with increased hepatic fibrosis. TheC282Y mutation itself did not have an independentassociation with fibrosis, only an indirect effect throughincreasing hepatic iron stores. These findings have nowbeen confirmed by a similar study in the USA.38

CHRONIC HEPATITIS C

Similar studies are now being reported for chronichepatitis C and in alcoholic liver disease although mostof these studies have yet to be published in full and the

results are, so far, conflicting. Smith et al. found no dif-ference in the prevalence of the C282Y mutation inpatients with chronic hepatitis C compared with a smallcontrol population.39 A greater proportion of patientswith the C282Y mutation had cirrhosis and detectableiron stores in the liver (as assessed by Perls’ stain) thanpatients without the C282Y mutation.The mean gradeof fibrosis was also significantly greater in chronichepatitis C patients with the C282Y mutation. How-ever, some chronic hepatitis C patients without theC282Y mutation also had positive Perls’ stains.Whetherthe effect of the C282Y mutation on fibrosis was medi-ated through increased iron stores or by a direct effectof the mutation was not addressed in this study. Thisconclusion is further supported by the recent study ofPiperno et al. who showed by multiple regression analy-sis that the HFE mutations and alcohol were the twofactors independently associated with increased hepaticiron and increased fibrosis in patients with chronic viralhepatitis.40 These studies support the role of the C282Ymutation41 and/or H63D mutation40 in causing somebut not all increased hepatic iron stores in chronichepatitis C. However, they do not necessarily support adirect relationship between the C282Y mutation andfibrosis.42

ALCOHOLIC LIVER DISEASE

Studies of the role of HFE in alcoholic liver disease arepreliminary and at present limited to one study pub-lished in full and others in abstract or letter form. Onebrief report has suggested that the C282Y mutation waspresent with a greater frequency in patients with alco-holic liver disease than controls and was associated withgreater liver damage,43 but this has not been substan-tiated by other studies.44,45 Grove et al. found thatneither the C282Y nor the H63D mutation was signif-icantly increased in alcoholic liver disease patients withfibrosis and cirrhosis and also concluded that thesemutations had no effect on hepatic iron stores infibrotic/cirrhotic alcoholic liver disease.44 However,most of the patients in their study group already hadestablished cirrhosis, a factor which may itself cause sig-nificant iron loading as discussed above. In addition,hepatic iron stores were assessed by the grade of hepa-tocellular Perls’ stain only and not by hepatic iron con-centration measured biochemically. In homozygoushaemochromatosis, alcohol consumption does notincrease total hepatic iron stores but seems to cause ironto redistribute from hepatocytes to the reticuloen-dothelial system.46 If a similar process occurs in het-erozygotes with increased hepatic iron stores, the degreeof iron loading would be underestimated by scoringhepatocellular iron only. In spite of these factors,patients with fibrotic alcoholic liver disease with one ormore copy of the C282Y mutation were more likely tohave a degree of visible hepatocellular iron stores thanage and sex-matched patients without this mutation.The relationship between hepatic iron stores and fibro-sis was not specifically addressed in this study and therelationship between the HFE mutations, hepatic iron

Iron, HFE and liver disease 747

stores and fibrosis was not assessed in precirrhoticpatients to eliminate the possible iron loading effect ofcirrhosis per se.

There has been much recent interest in progressiveiron overload (of the degree seen in advancedhaemochromatosis) in patients with end-stage liverdisease, often requiring liver transplantation. Severalgroups of investigators have reported on the frequencyof HFE mutations in such patients.47–49 Among a totalof 88 patients transplanted for iron overload at threedifferent USA institutions, only nine (10%) werehomozygous for the C282Y mutation. Similar observa-tions were made by Pascoe et al. from Australia.50 Thus,most of these patients appear to have gross iron over-load secondary to end-stage liver disease.14 Prospectivemore detailed studies on such patients with cirrhosisand iron overload are needed.

SUMMARY

From the available data, a clear picture of the role ofthe haemochromatosis mutations in chronic liver dis-eases has yet to emerge and further studies are requiredto clarify a number of issues. It appears, however, thatthe HFE C282Y mutation is not consistently presentwith greater frequency in chronic liver diseases whichsuggests that a direct effect of the gene on liver diseasesis unlikely.The C282Y mutation is generally associatedwith higher hepatic iron stores although various factorsmay make this association less apparent and non-HFEassociated increased iron stores clearly occur. Increasedhepatic iron stores are associated with increased fibro-sis in NASH but similar associations have not beenaddressed in other liver diseases. The association ofmildly increased hepatic iron stores with the C282Ymutation in NASH implicates these iron stores,although mild, as a contributory cause of fibrosis andnot vice versa.

If future studies are to provide definitive results, largedisease populations are required in view of the likelyhigh prevalence of HFE mutations in control popula-tions. Precirrhotic liver disease groups should bestudied because of the possibility that cirrhosis per seincreases hepatic iron stores. The relationship betweenhepatic iron stores and fibrosis should be specificallyaddressed in such studies.

REFERENCES

1 Villa E, Rubbinani L, Barchi T et al. Susceptibility ofchronic symptomless HbsAg carriers to ethanol-inducedhepatic damage. Lancet 1982; ii: 1243–4.

2 Ohnishi K, Iida S, Iwama S et al. The effect of chronichabitual alcohol intake on the development of liver cir-rhosis and hepatocellular carcinoma: Relation to hepatitisB surface antigen carriage. Cancer 1982; 49: 672–7.

3 Yokoyama H, Ishii H, Moriya S et al. Relationshipbetween hepatitis C virus subtypes and clinical features ofliver disease seen in alcoholics. J. Hepatol. 1995; 22:130–4.

4 Poynard T, Bedossa P, Opolon P et al. Natural history ofliver fibrosis progression in patients with chronic hepati-tis C. Lancet 1997; 349: 825–32.

5 Farrell GC. Drug-induced acute hepatitis. In: Farrell GC,ed. Drug-induced Liver Disease. Edinburgh: Churchill Liv-ingstone 1994; 247–99.

6 Powell LW, Jazwinska E, Halliday JW. Primary iron over-load. In: Brock JH, Halliday JW, Pippard MJ, Powell LW,eds. Iron Metabolism in Health and Disease. London: WBSaunders Co Ltd 1994; 227–70.

7 Loreal O, Deugnier Y, Moirand R et al. Liver fibrosis ingenetic haemochromatosis. J. Hepatol. 1992; 16: 122–7.

8 Piperno A, Fargion S, D’Alba R et al. Liver damage inItalian patients with hereditary haemochromatosis ishighly influenced by hepatitis B and C virus infection. J.Hepatol. 1992; 16: 364–8.

9 Bonkovsky HL, Banner BF, Rothman AL. Iron andchronic viral hepatitis. Hepatology 1997; 25: 759–68.

10 Olynyk JK, Reddy KJ, Di Bisceglie AM et al. Hepatic ironconcentration as a predictor of response to interferonalpha therapy in chronic hepatitis C. Gastroenterology1995; 108: 1104–9.

11 Boucher E, Bourienne A, Adams P, Turlin B, Brissot P,Deugnier Y. Liver iron concentration and distribution inchronic hepatitis C before and after interferon treatment.Gut 1997; 41: 115–20.

12 Hayashi H, Takikawa T, Nishimura N, Yano M, IsomuraT, Sakamoto N. Improvement of serum aminotransferaselevels after phlebotomy in patients with chronic activehepatitis C and excess iron. Am. J. Gastroenterol. 1994; 89:986–8.

13 Searle J, Kerr JFR, Halliday JW, Powell LW. Iron storagedisease. In: MacSween R, Anthony PP, Scheuer PJ,Portman BC, Burt, AD, eds. Pathology of the Liver.London: Churchill Livingstone, 1994; 219–41.

14 MacDonald RA. Hemochromatosis and cirrhosis in dif-ferent geographical areas. Am. J. Med. Sci. 1965; 249:36–46.

15 Ludwig J, Hashimoto E, Porayko MK et al. Hemosidero-sis in cirrhosis: A study of 447 native livers. Gastroenterol-ogy 1972; 62: 61–72.

16 Conn HO. Portacaval anastomosis and hepatic hemo-siderin deposition: A prospective, controlled investigation.Gastroenterology 1972; 112: 882–8.

17 Kent G,Volini FI, Minick OT et al. Effect of iron loadingupon the formation of collagen in the hepatic injuryinduced by carbon tetrachloride. Am. J. Path. 1964; 45:129–55.

18 MacKinnon MA, Clayton C, Plummer J et al. Iron over-load facilitates hepatic fibrosis in the rat/low dose carbontetrachloride model. Hepatology 1995; 21: 1083–8.

19 Tsukamoto J, Home W, Kamimura S et al. Experimentalliver cirrhosis induced by alcohol and iron. J. Clin. Invest.1995; 96: 620–30.

20 Kats J, Kobune M, Kohgo Y et al. Hepatic iron depletionprevents spontaneous development of fulminant hepatitisand liver cancer in Long-Evans Cinnamon rats. J. Clin.Invest. 1996; 98: 923–9.

21 Macdonald G, Bridle KR, Ward PJ et al. Hepatic ironstores and steatosis are associated with lipid peroxidationand fibrosis. J. Gastroenterol. Hepatol. 1998; 13: A152.

22 Houglum K, Ramm GA, Crawford DH, Witzum JL,Powell LW, Chojkier M. Excess iron induces hepatic

748 DK George et al.

oxidative stress and transforming growth factor beta 1 in genetic hemochromatosis. Hepatology 1997; 26: 605–10.

23 Britton RS. Metal-induced hepatotoxicity. Semin. LiverDis. 1996; 16: 3–12.

24 Feder JN, Gnirke A, Thomas W et al. A novel MHC class1-like gene is mutated in patients with hereditaryhaemochromatosis. Nat. Genet. 1996; 13: 399–408.

25 Jazwinska EC, Cullen LM, Busfield F et al. Haemochro-matosis and HLA-H. Nat. Genet. 1996; 14: 249–51.

26 The UK Haemochromatosis Consortium. A simplegenetic test identified 90% of UK patients withhaemochromatosis. Gut 1997; 41: 841–4.

27 Bassett ML, Halliday JW, Powell LW. HLA typing in idio-pathic hemochromatosis: Distinction between homozy-gotes and heterozygotes with biochemical expression.Hepatology 1981; 1: 2120–6.

28 Bulaj ZJ, Giffen LM, Jorde LB, Edwards CQ, Kushner JP.Clinical and biochemical abnormalities in people het-erozygous for hemochromatosis. N. Engl. J. Med. 1996;335: 1799–805.

29 Bassett ML, Halliday JW, Powell LW.Value of hepatic ironmeasurements in early haemochromatosis and determina-tion of the critical iron level associated with fibrosis. Hepa-tology 1986; 6: 24–9.

30 Barron R, Grace ND, Sherwood G et al. Iron overloadcomplicating sideroblastic anaemia: Is the gene forhaemochromatosis responsible? Gastroenterology 1989; 96:204–6.

31 Roberts AG, Whatley SD, Morgan RR et al. Increased frequency of the haemochromatosis Cys282Tyr mutationin sporadic porphyria cutanea tarda. Lancet 1997; 349:321–3.

32 Stewart KA, Busfield F, Jazwinska EC et al. The C282Ymutation in the haemochromatosis gene (HFE) andhepatitis C virus infection are independent cofactors forporphyria cutanea tarda in Australian patients. J. Hepatol.1998; 28: 404–9.

33 Bonkovsky HL, Poh-Fitzpatrick M, Pimstone N et al. Por-phyria cutanea tarda, hepatitis C and HFE gene mutationsin North America. Hepatology 1998; 27: 1661–9.

34 Bacon BR, Powell LW, Adams PC, Kresina TR, HoofnagleJH. Molecular medicine and hemochromatosis: At thecrossroads. Gastroenterology 1999; 116: 193–207.

35 Samprietro M, Piperno A, Lupinica L et al. High preva-lence of His63Asp HFE mutation in Italian patients withporphyria cutanea tarda. Hepatology 1998; 27: 181–4.

36 Moirand R, Mortaji AM, Loreal O, Paillard F, Brissot P,Deugnier Y. A new syndrome of liver iron overload withnormal transferrin saturation. Lancet 1997; 349: 95–7.

37 George DK, Goldwurm S, Macdonald GA et al. Increasedhepatic iron in non alcoholic steatohepatitis is associatedwith increased fibrosis. Gastroenterology 1998; 114:311–19.

38 Jawaid Q, Leclair P, Tortorelli K, Cobb J, Lambrecht R,Bonkovsky H. Non-alcoholic steatohepatitis (NASH) andhereditary hemochromatosis (HHC). Gastroenterology1998; 114: A1264.

39 Smith BC, Grove J, Guzail MA et al. Heterozygosity forhereditary hemochromatosis is associated with more fibrosis in chronic hepatitis C. Hepatology 1998; 27:1695–9.

40 Piperno A, Gergani A, Malosio I et al. Hepatic iron over-

load in patients with chronic viral hepatitis: Role of HFEgene mutations. Hepatology 1998; 28: 1105–9.

41 Hezode C, Cazeneuve C, Coue O et al. Liver iron over-load in patients with chronic active hepatitis C.Prevalence, role of the hemochromatosis gene and relationwith hepatic histological lesions. Hepatology 1997; 26:A1120.

42 Gehrke SG, Riedel HD, Fitscher BA, Goeser WS, Kalli-nowski B. Prevalence of mutations in the hemochromato-sis gene HLA-H in patients with chronic hepatitis C.Influence on histological activity, iron metabolism, andresponse to interferon therapy. Hepatology 1997; 26:A1128.

43 Teare J, Mantafounis D,Yallow R,Thursz M,Thomas HC.Increased frequency of the haemochromatosis allele inpatients with alcoholic liver disease. J. Hepatol. 1997; 26(Suppl. 1): 57.

44 Grove J, Daly AK, Burt AD et al. Heterozygotes for HFEhave no increased risk of advanced alcoholic liver disease.Gut 1998; 43: 262–6.

45 Aldersley MA, Howdle PD, Wyatt JI, Robinson PA,

Iron, HFE and liver disease 749

Markham AF. Haemochromatosis gene mutation in liverdisease patients. Lancet 1997; 349: 1025–6.

46 Deugnier YM, Loreal O,Turlin B et al. Liver pathology ingenetic haemochromatosis: A review of 135 homozygouscases and their bioclinical correlations. Gastroenterology1992; 102: 2050–9.

47 Fiel MI, Schiano TD, Bodenheimer HC et al. Genotypic,biochemical and histological analysis for hereditaryhemochromatosis in liver transplant patients. Hepatology1997; 26: 159A.

48 Brandhagen DJ, Porayko MK, Moyer RP et al. Genetictesting for hemochromatosis in liver transplant recipientswith hepatic explant hemosiderosis. Hepatology 1997; 26:158A.

49 Minguillan JT, Lee RG, Britton R et al. Genetic markersfor hemochromatosis in patients with cirrhosis and ironoverload. Hepatology 1997; 26: 158A.

50 Pascoe A, Kerlin P, Powell LW et al. Spur cell anaemiaand hepatic iron stores in patients with alcoholic liverdisease undergoing orthotopic liver transplantation. Gut1998; (in press).