Gut Supplement, 1991, S11l-S115 Sil Genetic haemochromatosis Gut: first published as 10.1136/gut.32.Suppl.S111 on 1 September 1991. Downloaded from

A B Bomford, I W Dymock, E B D Hamilton

Early years pathology was undertaken and a method devel- By the time of his arrival at King's College oped with Peter Scheuer and AR Muir to assess Hospital in 1966 Roger Williams had already hepatic iron overload in relatives with early iron published extensively in the field of haemo- loading.' This was based on the characteristic chromatosis. During the years 1959-1965 while electron microscopic appearances ofthe iron core working at the Royal Free Hospital with Sheila of ferritin which is distributed freely within the Sherlock he focused on two areas. The first was cytoplasm of hepatocytes. The counting of dif- the study of families of patients with haemo- fusely scattered electron dense particles provided chromatosis to determine whether there was a a means ofquantifying tissue iron concentrations genetic component to the disease and, if so, as compared with liver sections from control identify the mode of inheritance. The second subjects. If iron stores are only minimally area was the use of radioactive tracer techniques increased the excess hepatic iron, being in a to investigate iron absorption, in particular the soluble form, does not produce a marked reac- changes in response to treatment and its meas- tion when cytochemical methods such as Pearls urement in the relatives ofindex patients. stain are used. Application of this method to a A detailed family study of patients with group of 46 relatives, mainly siblings and chil- haemochromatosis culminated in two papers dren, showed that 74% had increased concentra- published in 1962.'2 This type of study was tions ofhepatic iron.2 A strong case was made on timely because the view that haemochromatosis the basis of these findings for an autosomal was an inborn error ofiron metabolism, based on recessive mode ofinheritance with heterozygotes the early work of Sheldon,3 had come under showing a minimal increase in hepatic iron stores attack by an American physician, R A Mac- and partial biochemical abnormalities such as donald. Sheldon's hypothesis which was formu- raised concentrations of serum iron. lated before the classical theories of iron balance The second abnormality that was found to had been published,4 was that haemochromatosis unequivocally characterise haemochromatosis was an iron storage disease in which tissue was the increased absorption of dietary iron. damage was the result of excess iron deposition; Even this simple assertion, however, was contro- the iron accumulated because of an inherited, versial in the 1960s because patients with greatly

albeit unknown metabolic defect. He considered increased iron stores had been observed to absorb http://gut.bmj.com/ that the excess iron would eventually lead to normal amounts of iron at the time of presenta- tissue damage manifesting as the clinical features tion.5 Working with Chris Pitcher it was soon of hepatic cirrhosis, diabetes mellitus and hypo- established that iron absorption was indeed gonadism.3 Macdonald's opposing view that iron within the normal range in most patients at overload and cirrhosis were not pathogenetically presentation.7 With reduction of iron stores by linked was based on clinical, pathological and venesection, absorption rose to high levels and

experimental evidence.5 It is worthwhile briefly only became suppressed as iron stores reaccumu- on September 24, 2021 by guest. Protected copyright. considering how this evidence was marshalled lated. The group of index patients and relatives because it is here that the controversy over was investigated by the use of double isotope whether hepatic iron overload occurs in alcoholic techniques and whole body counting, good cirrhosis has its origins. Macdonald suggested agreement being found for the results obtained that iron overload in cirrhosis was always the using the two methods in patients with a wide result of exposure to excess dietary iron, the range of iron loading and serum iron concentra- major source of which he considered to be tions.' Over half of the relatives examined had alcoholic drinks, iron in food from cooking increased absorption of dietary iron, a mean of utensils and pharmacological iron.56 It is difficult 26% of the dose as compared with 5% in the to provide quantitative data for dietary iron controls, when this was taken as a radiolabelled levels, however, and it is also necessary to iron salt with a standard meal.9 It is clear in postulate a separate cause for the portal cirrhosis, retrospect that many of the relatives examined The Institute ofLiver the major aetiological groups in the 1960's being were heterozygotes who, it is now known, do not Studies and Department alcoholic, malnutrition, postviral and cryp- accumulate significant amounts of iron in the of Rheumatology, King's togenic. The experimental evidence against iron liver or elsewhere and it is difficult to sustain the College Hospital and School of Medicine and deposition causing tissue injury was stronger; all argument that iron absorption continues at this Dentistry, London the attempts to produce liver damage in animals rate in carriers of a single gene. A B Bomford by dosing with enteral or parenteral iron5 had E B D Hamilton failed. Tissue damage could only be produced by Stepping Hill Hospital, the concomitant administration ofa diet deficient TREATMENT AND NATURAL HISTORY Southport, Cheshire in choline. Although venesection therapy is now common- I W Dymock Because of these difficulties the approach place and the accepted form of treatment of iron Correspondence to: with the Dr A B Bomford, Institute of taken was to search for evidence that there was a loaded patients haemochromatosis, per- Liver Studies, King's College heritable component to iron overload with the vasive controversy over whether the disease was a School of Medicine and to claim in Dentistry, Denmark Hill, inference that haemochromatosis was a distinct distinct pathological entity led the London SE5 9PJ. pathological entity. A detailed analysis ofhepatic 1966 that this form of treatment did not improve S112 Genetichaemochromatosis

prognosis or clinical features.6 At the time, right hand. Flexion of these joints was limited

reports of the effectiveness of venesection in and radiographs showed cystic changes in the Gut: first published as 10.1136/gut.32.Suppl.S111 on 1 September 1991. Downloaded from removing excess tissue iron and producing clini- metacarpal heads, sometimes with infracturing cal improvement were limited to isolated case ofthe cysts on to the joint surface which could be studies or small series.'°0' In 1969 an analysis of irregular. There was narrowing ofthe joint space patients who had been venesected to iron defi- but not erosions or ulnar deviation of the pha- ciency by the removal of 15-20 g iron, was langes. We agree with Resnick that this pattern published jointly with the Department of and radiological features ofarthritis in the hands Medicine, Royal Free Hospital.'0 This compared frequently differs from that seen in other cases of survival in 40 treated patients with a group of 18 pyrophosphate arthropathy.2" untreated patients whose records were available The chondrocalcinosis was most frequently at King's College Hospital and who had pres- seen in the knees both in the fibro and hyaline ented before the idea had taken hold that deple- cartilage, but it was also found in the wrists, tion of excess tissue iron could be beneficial. symphasis pubis and other large joints. Together Although treatment was associated with a with Eric Bywaters we have studied the spines at decrease in the calculated five year mortality autopsy from six patients and observed calcium from 67% to 11% the groups were not strictly pyrophosphate in the outer part of the interver- comparable with respect to age of presentation tebral discs and in the ligamentum flavum.2' This and clinical complications of iron overload. A was not associated with back pain but the subsequent analysis of a larger patient series in radiographs gave superficial resemblance to which the age of presentation of treated and ankylosing spondylitis. untreated groups was the same, survival was Arthritis starting in the hands has been again found to be significantly improved by reported as one of the presenting features of treatment." In this latter study a correction was genetic haemochromatosis in some patients2223 made for clinical differences between the two but as the disease progresses, arthritis ofthe hips groups, such as the presence ofdiabetes mellitus can be a major problem causing severe disability, and cirrhosis, which could have weighted sur- which in many cases can be relieved by arthro- vival in favour ofthe treated patients. Both these plasty. In our 20 year study of 112 patients, 10 studies differentiated clearly between those clini- had hip surgery; also noted were characteristic cal features that could be expected to improve radiological and histological features in which with treatment - that is, pigmentation, hepato- there was unusual lifting of the hyaline cartilage megaly, cardiac failure, and diabetes in about a offthe underlying bone in the hip joint.24 third ofpatients, from those which were irrevers- In addition to this destructive metabolic ible - arthritis, hypogonadism and portal hyper- arthritis which can lead to damaged joints, it is tension. In spite ofreports ofreversal ofcirrhosis important to stress that sometimes an acute after iron depletion'2 we consider it a rare event crystal synovitis or pseudogout caused by and in the majority of patients cirrhosis remains deposits of calcium pyrophosphate in the car- http://gut.bmj.com/ many years after the complete removal of iron tilage can occur. These inflammatory episodes with the attendant increased risk ofhepatocellu- have been known to lead to the mistaken diag- lar carcinoma.'3 nosis ofrheumatoid arthritis. Nearly 80% of the patients in our series had The presence ofiron in the synovium was first developed cirrhosis by the time of presentation. recorded by Sheldon.3 We have examined syn- More recently, increased numbers of patients ovium obtained by needle biopsy in 22 patients have been diagnosed at the precirrhotic stage and and at necropsy in five and related the findings to on September 24, 2021 by guest. Protected copyright. analysis of a large group of German patients has the clinical and radiological evidence ofarthropa- shown that these patients have a normal life thy and to the measurement of liver and total expectancy whereas cirrhotic, diabetic, and the body iron stores.2" At the time of the biopsy the more heavily iron loaded patients have a reduced iron did not relate directly with the joint changes, life expectancy. 14 The high risk ofhepatocellular but correlated well with liver iron and the iron cancer, over 200-fold increase, was noted in this burden removed by venesection therapy. study. The incidence, severity and radiological pro- gression of the arthritis and chondrocalcinosis was recorded over 10 years in 18 patients. ARTHRITIS AND HAEMOCHROMATOSIS Chondrocalcinosis did not get less or disappear, The recognition that arthritis frequently occurs despite adequate venesection therapy but spread in patients with genetic haemochromatosis is one to involve other joints and became more severe, of the important clinical additions to our knowl- being found in seven patients initially and in 13 at edge in recent years and this arthritis can be the the end of the study. The knees and wrists were cause ofmuch pain and disability. Since the first the joints most frequently involved.26 case reports by Schumacher in 1964," we have, Iron salts act as nucleating agents for crystal over the past 20 years, made a detailed study of formation and it has been shown that low concen- the specific features, the pattern of joint involve- trations of ferric salts promote crystal growth. ment and how it develops.""19 The amount ofpyrophosphate needed for crystal Arthritis was found in more than half the formation was reduced to 25% ofthat required in patients, and the presence of widespread the absence ofiron.27 McCarty has also attempted chondrocalcinosis as a result of the deposition of to show that the removal of such crystals may be calcium pyrophosphate was a distinguishing fea- inhibited by iron.2' He induced haemosiderosis ture. One of the earliest recognisable signs was in rabbit knees by injecting autologous blood and often bony swelling of the second and third measured the rate of clearance of "Sr and 4'Ca metacarpophalangeal joints more marked in the from the joints. This was significantly reduced in Bomford, Dymock, Hamilton S113

haemosiderotic joints suggesting inhibition of transferrin saturation and serum ferritin concen- intracellular pyrophosphatase and confirming tration.' The demonstration of tight linkage earlier work that the erythrocyte form of the between the HLA locus and the haemo- Gut: first published as 10.1136/gut.32.Suppl.S111 on 1 September 1991. Downloaded from enzyme was inhibited by ferrous ions.29 chromatosis gene has also been used to show that Additional evidence supporting the relation- iron loaded alcoholic patients are homozygous ship between the arthritis of haemochromatosis for the gene and do not differ significantly from with iron overload is that a similar type of joint non-drinking homozygotes.? damage has been reported in patients with sec- ondary haemochromatosis as a result of heredi- tary spherocytosis.' Age is also an important FERRITIN AND TRANSFERRIN RECEPTOR: factor; we found that the arthritis was signifi- EXPRESSION AND REGULATION cantly more frequent in those patients whose first Transferrin, its receptor and ferritin provide an symptom of haemochromatosis appeared after essential sequence for making iron available to the age of 50 years. This age factor is also seen in the cell.4' Disturbances in the structure, function the pyrophosphatase arthropathy that occurs in and regulation ofthis sequence have been repeat- patients with primary hyperparathyroidism. edly sought in genetic haemochromatosis. An early report claiming that the ferritin phenotype was abnormal in genetic haemochromatosis42 was HAEMOCHROMATOSIS GENE not confirmed in studies from this laboratory.43 The most important advance in understanding Serum ferritin, however, has been found to be a the pathogenesis of genetic haemochromatosis reliable means ofassessing tissue iron that can be came in 1975 with the report by Simon and mobilised by venesection4 and is widely used in colleagues that the frequency ofthe HLA antigen detecting affected relatives in family screening A3 was increased in a group of unrelated studies.45 Functional studies have used periph- patients.3 32 This association was confirmed by an eral blood monocytes because of their acces- analysis ofour own patients33 and by reports from sibility and because of the observation that the several other centres.3' Pedigree analysis substan- cells of the reticuloendothelial system are spared tiated the conclusion ofearlier studies ofthe iron in the iron loading process in genetic haemo- loading phenotype that the disease was chromatosis and may be unable to retain stored recessively transmitted.2 The discovery of link- iron.' Ferritin synthesis and transferrin receptor age of the haemochromatosis allele to the HLA function have been shown to be normal.54 The locus has had an enormous impact on the man- only site at which an abnormality has been agement of the condition and in clarifying the consistently observed is the epithelial cells ofthe problem ofiron loaded alcoholic patients, as well duodenum. This was first noted 25 years ago by as localising the disease gene to the short arm of Crosby who observed paucity of collections of chromosome 6. The haemochromatosis gene ferritin iron cores in duodenal biopsies studied

being tightly linked to the HLA locus is inherited by electron microscopy.47 http://gut.bmj.com/ with the particular HLA haplotype on which it is During the last 10 years the genes encoding the carried. In family studies, subjects who have the ferritin H and L chains, transferrin and its same two HLA haplotypes as the index patient receptor have been cloned and sequenced and will have homozygous genetic haemo- their chromosomal locations determined. This chromatosis, those with one shared haplotype are has important implications for our understand- heterozygotes, while subjects carrying neither of ing of genetic haemochromatosis. First, apart

the affected haplotypes are normal homo- from a ferritin pseudogene, none of the genes on September 24, 2021 by guest. Protected copyright. zygotes."'43 The only exceptions are offspring of localise to the short arm of chromosome 6 heterozygous-homozygous matings and patients indicating that genetic haemochromatosis is not in whom recombination of genetic material has caused by structural abnormalities of these taken place. While there is strong evidence that genes. Second, examination of the ferritin sub- the disease gene is closely linked to the region unit mRNAs revealed a highly conserved encoding the HLA class I genes, the location of sequence of 28 nucleotides in the 5'untranslated the disease allele relative to the A and B loci is not region in amphibia, birds, and man.4' Structural known. One approach to identifying the gene is analysis of this region predicted that it would to examine restriction fragment length polymor- assume a stem loop configuration in the form ofa phism of DNA using probes to HLA class 1 paired stem supporting a single stranded loop genes, other genes close to this region and to comprised of the bases CAGUGU.4' Analysis of overlapping sequences ofadjacent DNA.3637 the transferrin receptor mRNA revealed a sim- The prevalence of genetic haemochromatosis ilarly conserved region, although in this case has recently been estimated from the results of a there were important differences in that five stem study in the United States of over 11 000 blood loop structures were found in the 3'untranslated donors.3' A transferrin saturation of 62% and region.' The importance of these structures above was used to recognise the homozygous which have been termed iron responsive genotype, because this has previously been elements,5' is that they provide the cisacting shown to recognise 92% of such cases.39 Serum component of a translational control mechanism ferritin was found to indicate only those with that allows for the coordinate regulation by cells heavy iron loading. The frequency of homo- of the transferrin receptor (iron uptake) and zygosity was calculated to be 0 0045, corres- ferritin (iron storage). The transacting element of ponding to a gene frequency of 0-067. These this homoeostatic mechanism is a cytosolic pro- figures are similar to those derived from pedigree tein which in the absence of iron binds to iron analyses from several centres using a combina- responsive elements to suppress ferritin mRNA tion of HLA typing and biochemical assays for translation5' and increase translation of transfer- S1 14 Genetichaemochromatosis

rin receptor mRNA by increasing its stability.52 3 Sheldon JH. Haemochromatosis. London: Oxford University Press, 1935. Conversely, when iron is abundant, the iron 4 McCance RA, Widdowson EM. Absorption and excretion of Gut: first published as 10.1136/gut.32.Suppl.S111 on 1 September 1991. Downloaded from responsive element binding protein dissociates iron. Lancet 1937; ii: 680-4. 5 Macdonald RA. Hemochromatosis and hemosiderosis. Spring- from iron responsive elements so that ferritin field, Illinois: Thomas, 1964. mRNA translation proceeds and iron is stored in 6 Macdonald RA. Primary haemochromatosis: inherited or acquired? Prog Haematol 1965; 5: 324-57. newly synthec-ised ferritin; transferrin receptor 7 Williams R, Manenti F, Williams HS, Pitcher CS. Iron mRNA rapidly decays, the protein becomes less absorption in idiopathic haemochromatosis before, during and after venesection therapy. BMJ 1965; ii: 78-81. abundant and iron uptake by the cell is 8 Pitcher CS, Williams HS, Parsonson A, Williams R. The decreased. How the iron responsive element measurement of iron absorption by the double isotope technique. BrJ Haematol 1965; 2: 633-41. binding protein senses iron concentrations or 9 Williams R, Pitcher CS, Parsonson A, Williams HS. Iron what form the iron takes is not yet known; in vitro absorption in the relatives of patients with idiopathic haemochromatosis. Lancet 1965; ii: 1243-6. experiments suggest that the redox state of the 10 Williams R, Smith PM, Spicer EJF, Barry M, Sherlock S. protein is important in determining its binding to Venesection therapy in idiopathic haemochromatosis: An analysis of 40 treated and 18 untreated patients. Q J Med mRNA." 1969; 38: 1-16. The expression of ferritin and the transferrin 11 Bomford A, Williams R. Long term results of venesection therapy in idiopathic haemochromatosis. QJ7Med 1976; 45: receptor can be used as an indicator of the iron 611-23. status ofthe cell and their study provides a means 12 Fellows IW, Stewart M, Jeffcoate WJ, Smith PG, Toghill PS. Hepatocellular carcinoma in primary haemochromatosis in ofinvestigating homoeostatic mechanisms. In an the absence ofcirrhosis. Gut 1988; 29: 1603-6. investigation of cellular iron homoeostasis in 13 Bomford A, Walker RJ, Williams R. Treatment of iron overload including results in a personal series of 85 patients genetic haemochromatosis we found that both with idiopathic haemochromatosis. In: Kief H, ed. Iron proteins were appropriately regulated in the metabolism and its disorders. Amsterdam: Excerpta Medica, 1975: 324-30. liver.37 This is illustrated by the finding that in 14 Niederau C, Fischer R, Sonnenberg A, Stremmel W, Tram- untreated patients with increased concentrations pisch HJ, Stromeyer G. Survival and cause of death in cirrhotic and noncirrhotic patients with primary oftissue iron, receptor protein was undetectable, hemochromatosis. NEnglJMed 1985; 313: 1256-62. using an antireceptor monoclonal antibody. In 15 Schumacher HR. Hemochromatosis and arthritis. Arthritis Rhewnatism 1964; 7: 41-50. venesected patients with normal iron stores, the 16 Hamilton E, Williams R, Barlow KA, Smith PM. The receptor was re-expressed, being found in par- arthropathy ofidiopathic haemochromatosis. QJ7Med 1968; 37: 171-82. enchymal and non-parenchymal cells in a cyto- 17 Dymock IW, Hamilton EBD, Laws JW, Williams R. Clinical plasmic and membraneous distribution.5" In and radiological analysis of 63 patients with iron overload. Ann Rheum Dis 1970; 31: 469-76. contrast, the transferrin receptor was inap- 18 Atkins CJ, Mclvor J, Smith PM, Hamilton E, Williams R. in the villus of Studies in haemochromatosis and in idiopathic propriately regulated enterocytes chondrocalcinosis. QJMed 1970; 34: 71-82. the duodenal mucosa." At this site, in spite of 19 Hamilton EBD. Arthritis in ochronosis, haemochromatosis increased iron stores the was and Wilsons Disease. In: Scott JT, ed. Copemans textbook of greatly receptor the rheumatic diseases. Edinburgh, Churchill Livingstone, expressed with the same intensity as in controls. 1986. In patients with B thalassaemia and transfusional 20 Adamson TC, Resnick CS, Guera J, Wint VC, Weisman MH, iron overload, however, the enterocyte receptor Resnick D. Hand and wrist arthropathy in hemochromatosis

and calcium pyrophosphate deposition disease; distinct http://gut.bmj.com/ was down This observation is radiographic features. Radiology 1983; 147: 377-81. clearly regulated.54 21 Bywaters EGL, Hamilton EBD, Williams R. The spine in significant because iron absorption is increased at idiopathic haemochromatosis. Ann Rheum Dis 1971; 30: this site in both conditions as a result ofa 453-65. primary 22 Gordon DA, Little HA. The arthropathy ofhaemochromatosis genetic disorder of iron metabolism in genetic without haemochromatosis. Arthritis Rheumatism 1973; 16: haemochromatosis and because of increased 305-12. 23 Jonge-Bok JM, Macfarlane JD. The articular diversity ofearly erythroid drive in B thalassaemia. Both groups haemochromatosis. JBonejointSurg 1987; 69-B: 41-4. showed down regulation of the receptor in the 24 Axford JS, Bomford A, Revell P, Watt I, Williams R, Hamilton EBD. Hip arthropathy in genetic hemo- on September 24, 2021 by guest. Protected copyright. gastric mucosa which is an appropriate response chromatosis: Radiographic and histologic features. Arthritis the increased iron stores. Rheumatism 1991; 34: 357-61. considering greatly 25 Walker RJ, Dymock IW, Ansell ID, Hamilton EBD, Williams These findings are supported by recent reports in R. Synovial biopsy in haemochromatosis arthropathy. Ann ferritin determined RheumDis 1972; 31: 98-102. which expression, by radio- 26 Hamilton EBD, Bomford AB, Laws JW, Williams R. The immunoassay" and immunocytochemical"6 natural history of arthritis in idiopathic haemochromatosis. in duodenal QJrMed 1981; 50: 321-9. methods, was reduced enterocytes in 27 Hearn PR, Russell RGG, Elliott JC. Formation product of genetic haemochromatosis. It appears that calcium pyrophosphate in vitro and the effect of iron salts. of two Clin SciMol Med 1978; 54: 29. although coordinate regulation the pro- 28 McCarty DJ, Palmer DW, Garanci JC. Clearance of calcium teins is maintained in the enterocyte in genetic pyrophosphate dihydrate crystals in vivo. Effect of synovial their of hemosiderosis. Arthritis Rheumatism 1981; 24: 706-10. haemochromatosis pattern expression 29 McCarty DJ, Pepe PF, Solomon SD, Cobb J. Inhibition of indicates that this cell type is iron deficient. This human erythrocyte pyrophosphatase activity by calcium, returns us to cupric and ferrous ions. Arthritis Rheumatism 1970; 13: 336. paradox Crosby's original observa- 30 Berry EM, Miller JP. Hereditary spherocytosis, haemo- tion of nearly 30 years ago when he found a chromatosis, diabetes mellitus and chondrocalcinosis. Proc R Soc Med 1973; 66: 9-10. paucity offerritin iron cores in enterocytes.47 It is 31 Simon M, Pawlotsky Y, Bourel M, Fauchet R, Genetet BG. not clear at present at which point(s) in the Hemochromatose idiopathic. Maladie associee a l'antigene tissulaire HLA-A3? Nouv Press Med 1975; 3: 1432. regulatory pathway an abnormality lies nor how 32 Simon M, Bourel M, Fauchet R, Genetet B. Association this abnormality relates to the haemochromatosis of HLA-A3 and HLA-B14 antigens with idiopathic haemochromatosis. Gut 1976; 17: 332-4. gene on chromosome 6; it is possible to rule out a 33 Bomford A, Eddiston ALWF, Kennedy L, Batchelor R, structural defect of the iron responsive element Williams R. Histocompatibility antigens as markers of abnormal iron metabolism in patients with idiopathic binding protein as the gene encoding this protein haemochromatosis and their relatives. Lancet 1977; i: 327-9. localises to chromosome 9.50 34 Simon M, Brissot P. The genetics of haemochromatosis. J Hepatol 1988; 6: 116-24. 35 Edwards CQ, Dadone MM, Skolnick MH, Kushner JP. Hereditary haemochromatosis. Clin Haematol 1982; 11: 1 Scheuer PJ, Williams R, Muir AR. Hepatic pathology in 411-35. relatives of patients with haemochromatosis. Jf Pathol Bact 36 David V, Paul P, Simon M, et al. DNA polymorphism related 1962; 84: 53-64. to the idiopathic haemochromatosis gene: evidence in a 2 Williams R, Scheuer PJ, Sherlock 5. The inheritance of recombinant family. Hum Genet 1986; 74: 113-20. idiopathic haemochromatosis: A clinical and liver biopsy 37 Lord DK, Dunham I, Campbell RD, Bomford A, Strachan T, study of 16 families. QJMed 1962; 31: 249-65. Cox TM. Molecular analysis of the human MCH class I Bomford, Dymock, Hamilton S115

region in hereditary haemochromatosis. A study by pulsed- 47 Crosby WH. The control of iron balance by the intestinal field gel electrophoresis. Hum Genet 1990; 85: 976-84. mucosa. Blood 1963; 22: 441-9.

38 Edwards CQ, Griffen LM, Goidgar D, Drummond C, 48 Aziz N, Munro HN. Iron regulates ferritin mRNA translation Gut: first published as 10.1136/gut.32.Suppl.S111 on 1 September 1991. Downloaded from Skolnick MH, Kushner JP. Prevalence ofhemochromatosis through a segment of its 5' untranslated region. Proc Natl among 11,065 presumably healthy blood donors. N EnglJ7 Acad Sci USA 1987; 84: 8478-82. Med 1988; 318: 1355-62. 49 Casey JL, Hentze MW, Koeller DM, et al. Iron-responsive 39 Dadone MM, Kushner JP, Edwards CQ, Bishop DT, elements; regulatory sequences that control mRNA levels Skolnick MH. Hereditary hemochromatosis; analysis of and translation. Science 1988; 240: 924-8. laboratory expression of the disease by genotype in 18 50 Klausner RD, Harford JB. Cis-trans models for post- pedigrees. AmJ Clin Pathol 1982; 78: 196-207. transcriptional gene regulation. Science 1989; 246: 870-2. 40 LeSage GD, Baldus WP, Fairbanks VF, et al. Hemo- 51 Leibold EA, Munro HN. Cytoplasmic protein binds in vivo to chromatosis; genetic or alcohol-induced? Gastroenterology a highly conserved sequence in the 5' untranslated region of 1984; 84: 1471-7. ferritin heavy- and light-subunit mRNAs. Proc Natl Acad 41 Bomford AB, Munro HN. Transferrin and its receptor; their Sci USA 1988; 85: 2171-5. roles in cell function. Hepatology 1985; 5: 870-5. 52 Muilner EW, Neupert B, Kuhn LC. A specific mRNA binding 42-Powell LW, Alpert E, Isselbacher KJ, Drysdale JW. Abnor- factor regulates the iron-dependent stability of cytoplasmic mality in tissue isoferritin distribution in idiopathic transferrin receptor mRNA. Cell 1989; 58: 373-82. haemochromatosis. Nature 1974; 250: 333-5. 53 Lombard M, Bomford A, Hynes M, et al. Regulation of the 43 Bomford A, Lis Y, McFarlane IG, Williams R. Variation in the hepatic transferrin receptor in hereditary hemochromatosis. distribution of two human heart ferritin species. Isoferritin Hepatology 1989; 9: 1-5. profile and subunit composition in normal and iron overload 54 Lombard M, Bomford AB, Polson RJ, Bellingham AJ, subjects. BiochemJ7 1977; 167: 309-12. Williams R. Differential expression oftransferrin receptor in 44 Beamish MR, Walker R, Miller F, et al. Transferrin iron, duodenal mucosa in iron overload; evidence for a site- chelatable iron and ferritin in idiopathic haemochromatosis. specific defect in genetic hemochromatosis. Gastroenterology BrJHaematol 1974; 27: 219-28. 1990; 98: 976-84. 45 Basset ML, HallidayJW, Ferris RA, Powell LW. Diagnosis of 55 Whittaker P, Skikne BS, Covell AM, et al. Duodenal iron hemochromatosis in young subjects: predictive accuracy of proteins in idiopathic hemochromatosis. J Clin Invest 1989; biochemical screening tests. Gastroenterology 1984; 87: 628- 83: 261-8. 33. 56 Fracanzani AL, Fargion S, Romeno R, et al. Immuno- 46 Brink B, Disler P, Lynch S, et al. Patterns of iron storage in histochemical evidence for a lack of ferritin in duodenal dietary iron overload and idiopathic haemochromatosis. absorptive epithelioid cells in idiopathic hemochromatosis. J Lab Clin Med 1976; 88: 725-31. Gastroenterology 1989; 96: 1071-8. http://gut.bmj.com/ on September 24, 2021 by guest. Protected copyright.