T-Cell Receptor Repertoire In Hereditary Hemochromatosis: a Study of 32 Hemochromatosis Patients and 274 Healthy Subjects

T-Cell Receptor Repertoire in HereditaryHemochromatosis: A Study of 32 HemochromatosisPatients and 274 Healthy Subjects

Carla Cardoso, Graca Porto, Rosa Lacerda, Dolores Resende,Pedro Rodrigues, Fernanda Bravo, Jose Carlos Oliveira,Benvindo Justica, and Maria de Sousa

ABSTRACT: Low CD81 T lymphocyte numbers havecontributed to deciphering the genotype/phenotype dis-crepancies found in hereditary hemochromatosis (HH)patients genotyped for the Hfe mutations, C282Y andH63D. In this study, we extend the analysis of T lym-phocytes in HH to the T cell receptor (TcR) repertoire.Thirty-two HH patients (C282Y homozygous) and 274Hfe genotyped healthy subjects were studied. The follow-ing TcR chains were analyzed: Va2.3, Vb5.1, Vb5.2,Vb5.3, Vb6.7, Vb8, and Vb12 among the CD41 andCD81 populations. Lymphopenias and absence of expan-sions of the Vb5.2 and Vb12 chains in the CD81 poolwere seen in controls heterozygous for the C282Y muta-tion. Expansions in the control group were seen withinthe CD81 pool and were rare/absent within the CD41

pool. TcR expansions were found more frequent in pa-

tients with iron overload related pathology than in pa-tients without pathology. 9/16 of the patients with pa-thology have at least one expansion among the CD81 poola number significantly higher compared with patientswithout pathology (1/16). These findings suggest thatHfe has an effect in the shaping of T-cell populationseither directly, as indicated by the lymphopenia seen inthe two chains in C282Y heterozygous without iron over-load, or indirectly by contributing to iron overload pa-thology. Human Immunology 62, 488–499 (2001).© American Society for Histocompatibility and Immu-nogenetics, 2001. Published by Elsevier Science Inc.

KEYWORDS: hereditary hemochromatosis; Hfe; TcR;Va/b TcR expansions; lymphocytes

ABBREVIATIONSb2-m b2-microglobulinCID X-linked combined immunodeficiency diseaseHH hereditary hemochromatosisHBV hepatitis B virusHCV hepatitis C virusHIV human immunodefficiency virusHLA human leukocyte antigen

HTLV human T-lymphotropic virusMHC major histocompatibility complexRA rheumatoid arthritisRag1 recombinant activating gene 1TcR T-cell receptorTf transferrinTfSat transferrin saturation

INTRODUCTIONHereditary hemochromatosis (HH) is an autossomic re-cessive disorder of iron metabolism characterized by in-

creased iron absorption that leads to iron overload ofparenchymal cells in several organs [1–3]. Clinical con-sequences of iron accumulation include cirrhosis of theliver, hepatocellular carcinoma, diabetes, heart failure,arthritis, and hypogonadism.

The recently identified hemochromatosis gene, Hfe, isa new major histocompatibility complex (MHC) classI-like gene [4,5]. Two common mutations have beenidentified. The C282Y mutation is found in the vastmajority of Caucasian patients screened (for review seePorto and de Sousa [6]). It affects the a3 domain result-ing in a failure of the mutant protein to bind to b2-

From the Molecular Immunology and Pathology (C.C., G.P., R.L.,D.R., P.R., M.S.), ICBAS and Molecular Immunology, IBMC, Porto; andClinical Hematology (G.P., B.J.) and Clinical Chemistry (F.B., J.C.O.),Santo Antonio General Hospital, Porto, Portugal.

Address reprint requests to: Dr. Maria de Sousa, Molecular ImmunologyLaboratory, Institute for Molecular and Cell Biology (IBMC), Rua doCampo Alegre, 823, 4150-180 Porto, Portugal; Tel: 1351 (22)6074956; Fax: 1351 (22) 6098480; E-mail: [email protected].

Received November 21, 2000; revised February 2, 2001; accepted March1, 2001.

Human Immunology 62, 488–499 (2001)0198-8859/01/$–see front matter© American Society for Histocompatibility and Immunogenetics, 2001

Published by Elsevier Science Inc. PII S0198-8859(01)00233-6

microglobulin (b2-m) and, consequently, abolishes cellsurface expression [5]. The H63D mutation is found inpolymorphic frequencies in the normal population. Thismutation affects the a1 domain of the molecule but doesnot interfere with cell surface expression [5]. The rele-vance of this mutation/polymorphism (H63D) in thepathology of iron overload is still unclear.

In spite of the discovery of the Hfe gene and a mu-tation, which is present in the majority of HH casesamong Caucasians (C282Y), the clinical heterogeneityamong patients with the same genotype [7, 8] and theexistence of iron overload without Hfe mutations [9, 10]confirm that other factors must be critical in the regu-lation of iron metabolism.

Abnormalities of T lymphocyte numbers, particularlyCD81 T cells have been reported in patients with HH,which are related to the severity of the iron overload[11–14]. Evidence for a causal relationship between lowlymphocyte numbers and iron overload has gained supportfrom the gene knockout technology in mice. Iron overloadhas been described both in b2-m and in gd knockout mice[15–17]. The most severe experimental iron overload hasbeen observed, however, in mice double knockout forb2-m and Rag1 [18]. Bahram and coworkers [19], in astudy of lymphocyte populations in Hfe-deficient mice,found no differences in CD4 and CD8 numbers betweenthe knockout and control mice. More recently, Levy et al.[20] observed a significantly higher liver iron depositionin mice lacking both Hfe and b2-m molecules than inmice lacking Hfe alone. Taken together these observa-tions in animal models are in agreement with the previ-ous results with HH patients in which the CD4:CD8ratios were not related to the presence of Hfe mutationsbut with the expression of the disease (for review see deSousa and Porto [21]). The reverse possibility, i.e., that inHH patients iron overload may influence the T-cell pool,cannot be excluded. In the present study, we extend theanalysis of T lymphocytes in HH to the T cell receptor(TcR) repertoire. Thirty-two HH patients (homozygousfor the C282Y mutation) characterized by the acceptedbiochemical and pathologic markers of the disease [12]and 274 healthy subjects genotyped for the Hfe muta-tions were studied. The following TcR chains were an-alyzed: Va2.3, Vb5.1, Vb5.2, Vb5.3, Vb6.7, Vb8, andVb12 among the CD41 and CD81 T-cell pool.

MATERIAL AND METHODSPopulation StudiedPeripheral blood samples were obtained from 306 sub-jects genotyped for Hfe. The study population consistedof: (a) 32 HH patients characterized according to previ-ously described criteria [12] and “a posteriori” confirmedto be homozygous for the C282Y Hfe mutation; (b) 274

subjects without iron overload that included 180 familymembers of patients and 94 control subjects randomlyselected from the same geographical area of the patients;of these, 85 were heterozygous for the C282Y mutation,13 were compound heterozygous for the two mutations,75 were heterozygous and 18 were homozygous for theH63D mutation and 83 subjects had no Hfe mutations.The HH population consisted of 11 females and 21 malesaged between 13 and 70 years of age (mean 6 SD 546 6 13 years old) and the population without ironoverload consisted of 155 females and 119 males between8 and 85 years of age (mean 6 SD 5 41 6 17 years old)(Table 1). All subjects gave informed consent to partic-ipate in this study.

Exclusion CriteriaExclusion criteria for the purpose of the present studyincluded the finding of positive viral markers (HBV,HCV, HTLV, and HIV1 and 2) or a history of heavyalcohol consumption. Patients with acquired forms ofiron overload (namely due to hematologic disorders,chronic alcoholic or viral liver disease, and disturbance oflipid metabolism) or hemochromatosis patients nonho-mozygous for the C282Y mutation were excluded.

Definition of Iron Overload Related PathologyAmong HH PatientsIn sixteen patients iron accumulation in target organshad pathologic expression, as follows: liver fibrosis (n 55), cirrhosis (n 5 9), hepatocellular carcinoma (n 5 1),diabetes (n 5 5), hypogonadotrophic hypogonadism(n 5 6), pigmentation (n 5 8), arthropathy (n 5 9), andcardiac abnormalities (n 5 5) (see Table 5 for detailedcharacterization of the patients). The clinical diagnosis ofthe referred iron overload related pathologies was done inpatients by standard routine methods, namely, liver bi-opsy, magnetic resonance, endocrinologic functionaltests, radiologic imaging techniques, and echocardio-graphic examination, according to established criteria[22, 23]. The same routine clinical examinations were

TABLE 1 Population studied according to Hfegenotype, sex, and age

Hfe genotypeFemales

nMales

nGlobal

nAge (years)X 1 SD

C282Y/C282Y* 11 21 32 47 6 12C282Y/wt 52 33 85 36 6 17C282Y/H63D 6 7 13 35 6 16H63D/wt 42 33 75 45 6 17H63D/H63D 9 9 18 48 6 16wt/wt 46 37 83 43 6 17Total 166 140 306

* All subjects are HH patients; n 5 number of subjects.

489TcR Repertoire in Hereditary Hemochromatosis

done systematically in asymptomatic C282Y homozy-gous patients; the screening of iron related pathology,except liver biopsy, is not routinely performed in pa-tients with normal liver function tests.

Blood TestsPeripheral blood samples were obtained during morningfasting by venous punction. EDTA containing tubeswere used to collect the blood samples for determinationsof hematologic parameters, FACScan analysis and Hfegenotyping. Anticoagulant-free tubes were used for se-rum collection for determination of biochemical param-eters of iron metabolism. Hematologic and biochemicalparameters were determined by standard techniques inthe routine Hematology and Clinical Chemistry Labora-tories of Santo Antonio General Hospital (Porto, Portu-gal). Whole blood counts were done in a Coulter JSautomatic cell counter (Hialeah, FL, USA). Serum ironwas analyzed in a Cobas integra analyzer (Roche, Swit-zerland). Serum ferritin and serum transferrin were de-termined by the turbidimetric test in the same analyzer.

Detection of Hfe MutationsFor genotyping the Hfe gene the haemochromatosis genemutation assay (ViennaLab, Vienna, Austria) was usedaccording to the manufacturers’ specifications. Briefly,genomic DNA was isolated from whole blood, followedby polymerase chain reaction (PCR) amplification of theHfe gene, and subsequent detection of the two mutationsdone by allele-specific hybridization in microwells.Bound sequences are detected using a horseradish perox-idase-labeled antibody to the reporter molecule and colorsubstrates.

Peripheral Blood T-Cell PhenotypingMononuclear cells from fresh collected peripheral bloodwere stained from whole blood. The red blood cells werefirst lysed in 50 ml of lysis solution (10 mM Tris, 0.15M NH4Cl, pH 7.4) for 10 min at 37°C. Cells werewashed twice in PBS supplemented with 0.1% sodiumazide and 0.2% BSA (PBS-BSA). Staining was performedat 4°C for 30 min in the dark in staining solution(PBS-BSA) in round-bottomed microtiter plates(Greiner-Nurtingen, Germany) with ;0.5 3 106 cells/well. The following moAbs were used: CD3-FITC, CD8-PE, CD4-PE (DakoPats, Copenhagen, Denmark); F1(Va2.3), LC4 (Vb5.1), 1C1 (Vb5.215.3), W112(Vb5.3), OT145 (Vb6.7), 16G8 (Vb8), and S511(Vb12) all FITC-conjugated (T-Cell Sciences, Boston,MA, USA). A mouse IgG1 FITC/PE (Dakopats) was usedas a negative control. After the staining, cells werewashed twice with PBS-BSA and fixed in a final volumeof 500 ml of PBS containing 0.1% paraformaldehyde and0.1% sodium azide. At least 2 3 104 lymphocytes were

acquired in a FACScan (Becton Dickinson, MountainView, CA, USA) using forward and side-scatter charac-teristics, and subsequently analyzed using the Lysis IIprogram (Becton Dickinson, CA, USA). The percentagesof CD81 T cells positive for each TcR Va/b chain weredetermined after setting quadrants on CD8bright cells asdescribed in Arosa et al. [14].

The total number of CD41 and CD81 lymphocyteswere then estimated from the total lymphocyte counts bythe formula: CD81 (or CD41) cells 5 % of CD31CD81

(or CD31CD41) cells 3 total lymphocyte counts(3106/ml)/100.

Definition of T-Cell ExpansionsAn expansion was considered positive when the percent-age of cells positive for each TcR Va/b within the CD81

or CD41 pool were three times higher than the corre-sponding median [24, 25] found in the control popula-tion. The control population was defined as the popula-tion without iron overload and without Hfe mutations.

Statistical MethodsFor the purpose of statistical analysis, subjects were dividedaccording to the Hfe genotype. In addition, patients weredivided according to the presence/absence of iron overloadrelated organ pathology. Differences in the distribution ofvalues between groups were tested by the nonparametricKolmogorov-Smirnov two-sample test. The significance ofthe differences found in frequencies was tested using theChi-square test with Yates correction for small samples.Correction for multiple testing was done using the Dun-nett’s test based on the angular transformation of eachproportion. All testing was performed at the 0.05 level ofsignificance and all p values are two-sided.

RESULTSIron Related Parameters: Influence ofHfe MutationsThe biochemical characterization of the HH patients, allhomozygous for the C282Y mutation, and the normalsubjects grouped according to the Hfe genotypes, ispresented in Table 2. As expected, the group of the HHpatients had statistically significant higher mean valuesof serum iron, transferrin saturation (TfSat), and serumferritin when compared with subjects without Hfe mu-tations. When the biochemical parameters of the normalsubjects carriers of Hfe mutations were compared withthe subjects without Hfe mutations slight, but statisti-cally significant, differences were observed in iron pa-rameters (Table 2), although always within the normalrange. Subjects heterozygous for the C282Y mutation,compound heterozygous for both Hfe mutations, andhomozygous for H63D mutation had statistically signif-icant higher mean values of TfSat and lower mean values

490 C. Cardoso et al.

of plasma transferrin (Tf). Statistically significant highervalues of serum iron were seen in subjects compoundheterozygous for the two Hfe mutations and in subjectshomozygous for the H63D mutation when comparedwith subjects without Hfe mutations. No significantdifferences were observed in serum ferritin levels amongthe normal subjects genotyped for the Hfe mutations(Table 2) [26, 27].

T-Cell Receptor RepertoireTcR Va/b gene usage. The median values for TcR Va/bgene usage in CD41 and in CD81 of peripheral bloodsubsets, expressed in percentage, among the different Hfegenotyped subjects are presented in Table 3. It is note-

worthy the constancy of the median values of each Va/bwithin the CD41 and CD81 subsets among the differentgroups. However, the wide range of values seen in somecases reflects the existence of expansions as detailed in thenext section. The skewing of Vb5.11, Vb6.71, andVb81 chains toward the CD41 pool was observed in allHfe genotyped groups (Table 4). More than 80% of thesubjects have those chains preferentially expressed in theCD41 subset. The Vb121 chain appears to be skewedtoward the CD41 pool in more than 80% of the subjectsonly in the group of subjects carriers of the C282Ymutation, but not in the other Hfe genotyped groups(Table 4). In general, TcR expansions were seen in thepatient as well as in the control group without Hfe

TABLE 2 Biochemical characterization of the population analyzed grouped according to the Hfe genotype andsignificance differences between the groupsa

Hfe genotypeNumber of

subjectsFe

(mg/dl)Tf

(mg/dl)TfSat(%)

Ferritin(ng/ml)

C282Y/C282Yb

32 142 6 65† 195 6 43† 67 6 36† 210 6 390C282Y/wt 85 99 6 33 254 6 41* 33 6 11* 69 6 12C282Y/H63D 13 130 6 38† 236 6 30† 46 6 14† 74 6 24H63D/wt 75 96 6 38 258 6 25 32 6 12 61 6 16H63D/H63D 18 126 6 28† 240 6 41† 44 6 13† 101 6 49wt/wt 83 94 6 30 269 6 43 30 6 10 68 6 15

Fe 5 serum iron; Tf 5 serum transferrin; TfSat 5 transferrin saturation. Ferritin is summarized by geometric mean 6 SEM. The other biochemical parametersare presented by the arithmetic mean 6 S.D. according to the usual reference values for those parameters [26, 27].a The significance level of the comparison between the subjects carriers of Hfe mutations and the control group without Hfe mutations are indicated as * p ,0.05 and † p , 0.01.b All subjects are HH patients.

TABLE 3 TcR Va/b expression on CD41 and CD81 lymphocytes in subjects grouped according to the Hfegenotype

TcR

Hfe genotype

Cut-offd

(3x median)C282Y/C282Ya

(n 5 23–31)bC282Y/wt

(n 5 49–81)C282Y/H63D

(n 5 6–13)H63D/wt

(n 5 56–68)H63D/H63D(n 5 11–13)

wt/wt(n 5 57–70)

CD41 (%)Va 2.31 3.23 (1.69–42.95)c 3.15 (1.65–6.97) 3.34 (2.01–4.44) 3.28 (1.85–10.04) 3.22 (1.18–5.71) 3.55 (1.41–10.14) 10.65vb 5.11 2.38 (0.99–8.33) 2.60 (0.71–9.06) 2.95 (0.47–3.75) 2.66 (0.03–9.41) 2.73 (0.69–4.23) 2.69 (0.75–10.70) 8.07Vb 5.2

1

2.84 (1.33–17.74) 2.75 (0.45–5.78) 2.96 (0.94–4.04) 2.87 (0.84–5.71) 2.65 (1.65–5.04) 2.77 (1.44–7.68) 8.31Vb 5.31 1.29 (0.72–7.80) 1.23 (0.44–10.34) 1.28 (0.63–2.73) 1.22 (0.43–5.30) 1.32 (0.22–2.91) 1.22 (0.65–19.62) 3.66Vb 6.71 3.94 (0.34–9.59) 3.99 (0.02–11.23) 3.45 (3.07–6.06) 5.20 (0.81–8.31) 3.25 (0.46–6.71) 4.12 (0.39–15.60) 12.36Vb 81 4.01 (1.74–18.02) 4.46 (1.62–11.82) 3.86 (2.60–6.10) 4.11 (1.61–6.57) 3.70 (2.30–5.98) 4.29 (0.86–13.91) 12.87Vb 121 1.62 (1.22–6.99) 1.72 (0.81–3.16) 1.75 (1.61–2.00) 1.84 (0.89–13.11) 1.71 (0.90–2.24) 1.83 (0.53–5.82) 5.49

(n 5 24–31) (n 5 49–78) (n 5 7–13) (n 5 57–68) (n 5 11–13) (n 5 56–68)Va 2.31 2.68 (0.62–27.25) 2.78 (0.59–7.25) 3.21 (1.81–4.54) 2.73 (0.51–12.29) 3.05 (0.98–10.47) 2.80 (0.62–16.48) 8.4Vb 5.11 0.71 (0.20–6.18) 0.92 (0.14–10.63) 1.02 (0.09–6.24) 0.98 (0.06–23.05) 0.73 (0.10–2.97) 0.87 (0.26–11.67) 2.61Vb 5.21 2.44 (0.68–14.79) 2.51 (0.13–5.93) 2.26 (0.49–10.97) 2.64 (0.23–9.16) 1.66 (0.12–10.48) 2.75 (0.55–10.72) 8.25Vb 5.31 1.14 (0.44–6.41) 1.12 (0.21–22.80) 1.03 (0.22–3.88) 0.96 (0.22–5.55) 0.91 (0.15–1.61) 1.01 (0.36–6.14) 3.03Vb 6.71 1.06 (0.08–9.65) 1.32 (0.29–7.94) 1.40 (0.30–4.26) 1.29 (0.14–9.69) 0.87 (0.08–2.09) 1.35 (0.07–8.72) 4.05Vb 81 2.47 (0.41–29.10) 3.48 (0.55–33.07) 2.61 (0.80–9.37) 2.97 (0.28–14.45) 2.03 (0.77–10.25) 2.99 (0.83–10.78) 8.94Vb 121 0.98 (0.17–3.98) 0.97 (0.35–3.91) 1.35 (1.05–1.61) 1.18 (0.37–4.95) 1.79 (0.60–4.38) 1.37 (0.30–11.07) 4.11

a All subjects are HH patients; b n 5 number of subjects; c median (range); d for definition of an expansion see Material and Methods.


mutations (Figure 1). We observed that 15.5% of theHH patients (5/32) have at least one expansion amongthe CD41 T cells, and 31.25% of the patients (10/32)have at least one expansion in the CD81 T cell subset(Table 5). Expansions in control subjects were also seenboth in the CD41 (6/69, 8.7%) and in the CD81 subset(20/69, 29%) (Figure 1). None of the patients had ex-pansions in the Vb6.71 chain among the CD41 cells orin the Vb121 chain among the CD81 T-cell subset(Figure 1 and Table 5).

T-cell expansions relate to organ pathology in HH patientshomozygous for the C282Y mutation. When the number ofexpansions were analyzed separately in the group ofpatients with and without iron overload related organpathology (for characterization of patients see Materialand Methods and Table 5), the number of subjects withexpansions was much higher in the group with pathologythan in the one without pathology, both in the CD41

(4/16 or 25% vs 1/16 or 6.25%) and in the CD81 pool

(9/16 or 56% vs 1/16 or 6.25%) (Figure 1 and Table 5).This difference being significantly higher for the CD81

population (p 5 0.008, x2 with Yates correction). Themean ages of the HH patients with expansions andwithout expansions were not statistically significant(51 6 12 years old and 43 6 13 years old, respectively).In spite of the observation of more subjects with expan-sions within the CD81 and CD41 pool among the groupwith iron overload related organ pathology, no statisti-cally significant differences were found in total lympho-cyte counts, total numbers of CD41 and CD81 T cellsbetween the two patient groups (Table 6). In severalpatients, undergoing phlebotomy treatment, sequentialblood samples were available. Three representative ex-amples of the persistence of an expansion over periodsranging from 1 to 4 years old are shown in Figure 2. Theserial data in the case of the patient 076/001 weredetermined during the intensive phlebotomy treatment(Figure 2A). In the case of patient 001/001 the serialdeterminations were obtained 10 years after intensivetreatment, during the maintenance treatment (Figure2B). Patient 056/001 had serial determinations duringthe intensive treatment and in the maintenance treat-ment (Figure 2C). In addition, the serial data beforeand after intensive treatment of a patient (060/001) inwhom no expansions were found is illustrated (Figure2D).

FIGURE 1 Frequency of HH patients with Va/b expan-sions in the CD41 and CD81 T-cell subsets in comparisonwith controls without Hfe mutations (white bars). The groupof patients, all homozygous for the C282Y mutation, wasdivided according to the presence (black bars) or absence (graybars) of iron overload organ related pathology. Note that somesubjects have missing values for some chains.

TABLE 4 Mean ratio of CD41 to CD81 T cells that simultaneously express each Va/b chain among the Hfegenotyped subjects, and the proportion of subjects with those ratios .1

CD41/CD81 ratio

Hfe genotype

C282Y/C282Ya C282Y/wt C282Y/H63D H63D/wt H63D/H63D wt/wt

Va2.31 1.29 (72%)* 1.19 (67%) 0.85 (43%) 1.30 (76%) 1.04 (58%) 1.20 (64%)Vb5.11 2.67 (93%) 2.68 (92%) 3.66 (86%) 2.69 (91%) 3.2 (100%) 2.77 (93%)Vb5.21 1.13 (62%) 1.04 (56%) 1.18 (67%) 1.09 (57%) 1.77 (75%) 1.04 (55%)Vb5.31 1.30 (78%) 1.26 (67%) 1.32 (70%) 1.27 (72%) 1.39 (100%) 1.36 (73%)Vb6.71 3.23 (88%) 2.98 (91%) 2.96 (100%) 3.56 (97%) 3.89 (100%) 2.90 (96%)Vb81 1.79 (81%) 1.31 (80%) 1.46 (82%) 1.48 (79%) 2.07 (77%) 1.50 (81%)Vb121 1.89 (85%) 1.76 (83%) 1.46 (86%) 1.58 (77%) 1.06 (55%) 1.36 (65%)

* Arithmetic mean (percentage of subjects with the CD41/CD81 ratio of each Va/b chain .1.a All subjects are HH patients (shaded area).


CD41 and CD81 T-cell expansions among the other Hfegenotyped groups without iron overload. The observation ofan increased frequency of expansions in the patient groupraised the necessary question of the expansions occurringin response to iron overload itself. Therefore, the sameanalysis was done in the other Hfe genotyped groups

without iron overload in order to see whether Hfe mu-tations could influence the shaping of the T-cell reper-toire independently of iron overload. For that purposethe subjects without iron overload were grouped accord-ing to Hfe genotype. The frequency of individuals withexpansions in the two subsets among carriers of Hfe

TABLE 5 Characterization of HH patients at the time of first TcR V analysis

Clinical groupa Patient Age/sexExpansionsin CD41

Expansionsin CD81

TfSat(%)

Ferritin(ng/ml) Pathology

Phlebotomytreatment

With iron relatedpathology(n 5 16)

001/001 58/M None Vb5.1; Vb8 43 24,6 F1ECG1A M003/001 70/M Va2.3 Va2.3 65 23,4 C1P1D1A M019/001 43/M None None 69 30,6 C1H M019/011 43/M None None 138 2156 C1P I031/001 70/F None None 100 105 C1P1D1A M035/001 48/M None Vb5.3; Vb6.7; Vb8 32 9,6 C1P1H1Hep I039/010 61/F None Vb5.1 40 207,8 A I049/001 51/M None None 11 26 C1P1H1A I054/001 60/M Va2.3; Vb5.1; Vb5.2;

Vb5.3; Vb12Va2.3; Vb5.1;Vb5.2; Vb5.3

93 3502 A

056/001 48/F None Vb5.2 100 211 C1P1D1H1A I060/001 46/M None None 24 13 A M074/002 41/M None Vb6.7 145 4243 F I076/001 57/F None Vb6.7 80 1770 ECG1F1A I079/001 62/M Vb5.3 None 17 15,7 C1P1ECG1H M084/001 38/M Va2.3 None 104 5782 C1F1D1ECG1H I

AMA/MGC 31/M None Vb8 101 4405 F1P1D1ECG IWithout iron related

pathology(n 5 16)

003/003 61/F None None 60 60,5 None M005/003 32/M None None 70 857,2 None I010/004 46/F None None 22 7,1 None M010/023 45/M None None 6 15 None M015/006 46/M None None 78 78,3 None M015/007 50/F None None 33 58,6 None M034/001 41/F None None 56 64,9 None M035/008 26/F None None 143 285 None I035/009 23/M None None 66 139 None I047/001 39/M Vb8 None 117 50 None M047/004 13/M None None 40 24,3 None —067/001 42/M None None 62 212,3 None I082/001 44/F None Va2.3 62 103,5 None I085/001 43/M None None 93 1612 None I085/002 39/M None None 96 2400 None IANM 61/F None None 92 142 None I

a For clinical characterization see Patients and Methods. All patients were homozygous for C282Y mutation.

A 5 arthropathy; C 5 cirrhosis; D 5 diabetes; ECG 5 electrocardiograph abnormalities; F 5 fibrosis; H 5 hypogonadism; Hep 5 hepatocarcinoma; P 5pigmentation; I 5 intensive treatment (weekly phlebotomy); M 5 maintenance (periodical phlebotomy).

TABLE 6 Total lymphocyte counts, CD41 and CD81 T cells among the HH patients (C282Y/C282Y) groupedaccording to the presence or absence of iron overload related pathology

HH patients NumberTotal lymphocytes

(3106 cells/ml)CD81 T cells

(3106 cells/ml)CD41 T cells

(3106 cells/ml)

With pathology 16 1.88 6 0.69a 0.37 6 0.27 0.84 6 0.38Without pathology 16 2.04 6 0.79 0.47 6 0.22 0.94 6 0.42Significance level NS NS NS

a Arithmetic mean 6 standard deviation.

NS 5 not significant.


mutations was not statistically significant differentfrom that found in the control group (subjects onlycarrying the wt allele). In all groups, the frequency ofsubjects with expansions among the Va/b chains stud-ied were higher within the CD81 subset (26% inC282Y/wt; 38% in C282Y/H63D; 26% in H63D/wt;38% in H63D/H63D; and 29% in wt/wt subjects) thanwithin the CD41 T subset (6% in C282Y/wt; 0% inC282Y/H63D; 11% in H63D/wt; 0% in H63D/H63D; and 9% in wt/wt subjects) (Figure 3). In thegroup of subjects C282Y/wt the frequency of individ-uals with Vb5.11 (8/50, 16%) and Vb81 (6/70, 6.7%)

cell expansions within the CD81 pool was twofoldhigher than in controls (5/57, 8.8% and 2/64, 3.1%,respectively). This difference, however, did not reachstatistical significance. Furthermore, none of theC282Y/wt subjects had expansions in Vb5.21 andVb121 cells within the CD81 pool in contrast withthe controls (p 5 0.03 and p 5 0.02, respectively;Figure 3). In the light of this observation we comparedthe distribution of positive cells in those two chainsbetween the group of the C282Y/wt subjects and thewt/wt subjects (excluding the subjects with expan-sions). The former had statistically significant lowervalues of Vb121 cells in the CD81 subset (Kolmog-orov-Smirnov two sample test, DN 5 0.28, p 5 0.025;Figure 4b). The difference observed for the Vb5.21

cells among the CD81 subset was not statisticallysignificant (Kolmogorov-Smirnov two sample test,DN 5 0.25, p 5 0.08; Figure 4a). No differences wereobserved in the distribution of those two chains in theCD41 pool between the two Hfe genotyped groups.

In the group of subjects homozygous for the H63Dmutation, we observed that 25% (3/12) of subjects hadexpansions in the Va2.31CD81 T cells.

The significance of the differences observed amongHfe genotyped subjects persisted after correction formultiple testing with Dunnett’s test.

FIGURE 2 Illustration of serial determinations in four rep-resentative examples: (A) Patient 076/001 (expansion inVb6.71CD41 cells) during 4 months in the course of theintensive treatment; (B) Patient 001/001 (expansion inVb81CD41 cells) during 4 years of maintenance therapy (theintensive treatment was finished 10 years before); (C) Patient056/001 (expansion in Vb5.21CD41) during the intensivetreatment and 2 months of maintenance; and (D) Patient060/001 (no expansions were observed), serial data ofVb6.71CD41 cells during 3 months of the intensive treat-ment and 7 months after the intensive treatment. The discon-tinuous line shows the value that defined the criteria for anexpansion (see Patients and Methods). The arrow indicates theend of intensive treatment. Abbreviation: y.o. 5 years old.


DISCUSSIONThe start of the present study of the TcR repertoire andof the T-cell subpopulations in hereditary hemochroma-tosis patients preceded the discovery of Hfe, the candi-date gene for hemochromatosis, first designated asHLA-H [4]. The consistent finding of defective numbersof CD81 cells in some HH patients led to the questionwhether the quantitative abnormalities seen would affectrandomly or selectively the TcR Va/b chain repertoire.The availability of Hfe genotyping halfway through thestudy gave us the opportunity of extending it to Hfegenotyped family members of patients and controls. Thepopulation examined consisted of 306 subjects, making

it, to our knowledge, the largest study of the TcRrepertoire in a human population. The present study of32 HH patients, 180 family members without ironoverload and 94 additional controls, provided the oppor-tunity to determine whether the presence of Hfe muta-tions and/or iron overload could influence the shaping ofT-cell pools and the TcR repertoire.

TcR Va/b Expansions and Iron Overload RelatedOrgan PathologyIron overload per se did not appear to influence eithertotal or selected lymphocyte population numbers. How-ever, the frequency of expansions in the C282Y homozy-

FIGURE 3 Frequency of Hfe geno-typed subjects without iron overloadwith Va/b expansions in the CD41 andCD41 T-cell subsets in comparison withcontrols without Hfe mutations (whitebars). *Indicates the statistical signifi-cance (p , 0.05) between the frequencyof expansions in a particular Va/b chainin each group when compared with con-trols (wt/wt).


gous patients was significantly higher in those with ironoverload related organ pathology than in those withoutpathology. These expansions involved both CD41 andCD81 cells. It must be stressed that expansions observedare not related to the intensive phlebotomy treatment per seas clearly illustrated in serial data of representative patientspresented in Figure 2. No differences were found in totallymphocyte counts, CD41 and CD81 T cells betweenthese two patient groups, suggesting that the expansionsobserved do not alter the size of these populations.

Because one of the exclusion criteria was the positivefinding of the viral markers of HBV, HCV, HTLV,HIV1, and HIV2, it seems unlikely that viruses played arole in the differences found between the two groups ofpatients. CD8 expansions expressing different Vb afterprimary infection with HIV were reported in a serialanalysis of six HIV patients, suggesting that expansionscould be involved in an antigen specific immune re-sponse [28]. In the case of iron-related pathology, it isunclear whether changes can occur in the target loadedcells that could be “recognized” by circulating T lym-phocytes, and is a point that perhaps deserves furtheranalysis. TcR expansions have been reported in a group

of 71 heavy alcohol drinkers, particularly of Vb5.2among CD81 T-cell pool [25] and in other forms ofselective organ pathologies unrelated to infection,namely in rheumatoid arthritis (RA) [29–31] and inBehcets disease [24]. Mizushima et al. [29], in a study of15 RA subjects, described that some alterations in theTcR Va/b repertoire were attributable to CD41 T cells.A different study described clonal expansions of CD81

cells expressing Va12.1 in 15% of 46 RA patientsstudied [30] when compared with 68 healthy controls.Fitzgerald and collaborators [31], in a study published in1996, suggested that expansions in the 31 RA patientsanalyzed may be skewed compared with 65 healthy sub-jects, and found that Vb3 and Va12.1 were more fre-quently expanded in RA patients than in controls. InBehcets disease, T-cell expansions were at significantlyhigher incidence in CD41 subset in the 23 patientsstudied when compared with 15 control subjects [24].

TcR Va/b Expansions and Hfe MutationsTo analyze the influence of Hfe mutations, indepen-dently of iron overload, in the shaping of the TcRrepertoire, the frequency of expansions in Hfe genotyped

FIGURE 4 Cumulative distribution function of theVb5.21CD81 T cells (a) and the Vb121CD81 T cells(b) among C282Y/wt subjects and wt/wt subjects. Thestatistical significance of the difference between thedistributions of the values in the two Hfe genotypedgroups are indicated in the graph and were tested bythe nonparametric Kolmogorov-Smirnov two-sampletest.


subjects without iron overload was studied (detected byboth serum ferritin and TfSat levels; Table 2).

In the whole group of normal subjects (n 5 274), wefound frequent expansions in the CD81 T-cell popula-tion (ranging from 26% to 38% in subjects among thedifferent Hfe genotypes) and a lower or absent frequencyof expansions in the CD41 T-cell population (rangingfrom 0% to 11% among the different Hfe genotypes).This observation is in agreement with previous studiesreporting expansions in normal subjects within theCD81 T-cell pool and the absence of expansions withinCD41 T-cell pool [25, 31, 32]. However, in earlierstudies the subjects were not genotyped for Hfe.

A study of this dimension confirmed the observationmade by others of the existence of skewing of some Va/bchains toward CD41 or CD81 population [30, 33–38].In this study this was clearly the case for the Vb5.11

chain, previously known to be skewed toward the CD41

pool [33, 34]. We observed that more than 80% of thesubjects have the Vb6.71 chain and Vb81 chain skewedto the CD41 subset. In the case of Vb6.71 chain thisobservation was reported earlier in a study done by others[33, 35] and, in the case of Vb81, this finding wasobserved by Imberti and coworkers [38]. In the case ofthe Vb121chain we only observed a majority of individ-uals with this chain skewed toward the CD41 subset insubjects carrying the C282Y mutation. Variable patternsof expression of Vb121 chain toward the CD41 orCD81 subset or equal expression in the two pools werereported by Davey et al. [33], in a population where noother genetic markers were tested for.

In the present study an additional finding deservesattention: the absence of expansions in the Vb5.21 andVb121 cells among the CD81 subset in C282Y het-erozygous with statistically significant lower percentagesof Vb121CD81 T cells. It appears that the C282Ymutation (both in homozygous and in heterozygousform) affects, in a negative manner, the selection ofVb121CD81 cells. None of the carriers of the C282Ywere found to have expansions in this particular chain.Curiously, this chain did not appear to be affected inpeople carrying the H63D mutation. Previous work byGoldman and collaborators in 1992 [39] also describedstatistically significant low numbers of Vb5 and Vb12among the CD81 T cells in a novel X-linked combinedimmunodeficiency disease (CID). In the two cases, HHand CID, the defective numbers of the Vb chains are notrelated to infection. Moreover, the iron load status ofCID patients has not been reported.

The question of a putative role for Hfe in peptidepresentation is still under study (F. Lemonnier and H.Firat, personal communication). At present, it is thoughtthat Hfe is unlikely to contribute to antigen presentation[40], but its two mutations are linked to particular

MHC-class I molecules namely HLA-A3 (in the case theC282Y) and HLA-A29 (in the case of the H63D) [41–43]. This raises the possibility that classical MHC class Imay itself contribute to the observed results. In addition,cellular iron overload may induce yet not identifiedchanges in the target cells that, in turn, may be recog-nized by circulating T cells.

The recent observation by Hentze’s group (Muck-enthaler and Hentze, personal communication), a higherexpression of Hsp70 by the microarray technology inHeLa cells occurring when incubated with iron, supportsthe idea that stress proteins induced by iron could beinvolved in the immune response. Hsp70 is a proteinthat has been described as a chaperone with capacity toinduce cellular immunity [44] and with potential capac-ity to activate T cells in vivo and in vitro [45].

ACKNOWLEDGMENTS

The authors acknowledge Carla Simoes, Mafalda Fonseca,Graca Melo, and Eugenia Cruz for all the technical support.We thank Pedro Oliveira for the help in the statistical analysisof the data. We also gratefully thank Fernando Arosa, ElsaCardoso, and Margarida Correia Neves for helpful discussion.This work was funded by the Portuguese Foundation forScience and Technology (FCT) Grant #1323/95. Carla Cardosois the recipient of a PRAXIS XXI PhD fellowship grant(BD/13383/97).

REFERENCES

1. Cartwright GE, Edwards CQ, Kravitz K, Skolnick M,Amos DB, Johnson A, Buskjaer L: Genetics of hemochro-matosis. N Engl J Med 301:1291, 1979.

2. Cox TM, Lord DK: Hereditary hemochromatosis. Eur JHaematol 42:112, 1989.

3. Edwards CQ, Griffen LM, Goldgar DE, Drummond C,Skolnick MH, Kushner JP: Prevalence of hemochromato-sis among 11,065 presumably healthy blood donors.N Engl J Med 318:1355, 1988.

4. Feder J, Gnirke A, Thomas W, Tsuchihashi Z, RuddyDA, Basava A, Dormishian F, Domingo R, Ellis MC,Fullan A, Hinton LM, Jones NL, Kimmel BE, Kronmal,GS, Lauer P, Lee VK, Loeb DB, Mapa FA, McClelland E,Meyer NC, Mintier GA, Moeller N, Moore T, MorikangE, Prass CE, Quintana L, Starnes SM, Schatzman RC,Brunke KJ, Drayna DT, Risch NJ, Bacon BR, Wolff RK:A novel MHC class I like gene is mutated in patients withhereditary haemochromatosis. Nature Genet 13:399, 1996.

5. Feder JN, Tsuchihashi Z, Irrinki A, Lee VK, Mapa FA,Morikang E, Prass CE, Starnes SM, Wolff RK, Parkkila S,Sly WS, Schatzman RC: The hemochromatosis foundermutation in HLA-H disrupts beta2-microglobulin inter-action and cell surface expression. J Biol Chem 272:14025, 1997.

6. Porto G, de Sousa M: Variation of hemochromatosis prev-


alence and genotype in national groups. In Barton JC,Edwards CQ (eds): Hemochromatosis: genetics, patho-physiology, diagnosis and treatment. Cambridge, UK:Cambridge University Press, 2000.

7. Rhodes DA, Raha-Chowdhury R, Cox TM, Trwosdale J:Homozygosity for the predominant Cys282Tyr mutationand absence of disease expression in hereditary hemochro-matosis. J Med Genet 34:761, 1997.

8. Crawford DH, Jazwinska EC, Cullen LM, Powell LW:Expression of HLA-linked hemochromatosis in subjectshomozygous or heterozygous for the C282Y mutation.Gastroenterology 114:1003, 1998.

9. Shaheen NJ, Bacon BR, Grimm IS: Clinical characteristicsof hereditary hemochromatosis patients who lack theC282Y mutation. Hepatology 28:526, 1998.

10. Pietrangelo A, Montosi G, Totaro A, Garuti C, Conte D,Cassanelli S, Fraquelli M, Sardini C, Vasta F, Gasparini P:Hereditary hemochromatosis in adults without patho-genic mutations in the hemochromatosis gene. N Engl JMed 341:725, 1999.

11. Porto G, Reimao R, Goncalves C, Vicente C, Justica B, deSousa M: Haemochromatosis as window into the study ofthe immunological system: a novel correlation betweenCD81 lymphocytes and iron overload. Eur J Haematol52:283, 1994.

12. Porto G, Vicente C, Teixeira MA, Martins O, Cabeda JM,Lacerda R, Goncalves C, Fraga J, da Silva BM, Alves H,Justica B, de Sousa M: Relative impact of HLA phenotypeand CD4/CD8 ratios on the clinical expression of hemo-chromatosis. Hepatology 25:397, 1997.

13. Arosa FA, da Silva AJ, Godinho IM, ter Steege JC, Porto G,Rudd CE, de Sousa M: Decreased CD8-p56lck activity inperipheral blood T-lymphocytes from patients with hered-itary haemochromatosis. Scand J Immunol 39:426, 1994.

14. Arosa FA, Oliveira L, Porto G, da Silva BM, Kruijer W,Veltman J, de Sousa M: Anomalies of the CD81 T cellpool in haemochromatosis: HLA-A3-linked expansions ofCD81CD28-T cells. Clin Exp Immunol 107:548, 1997.

15. de Sousa M, Reimao R, Lacerda R, Hugo P, Kaufmann S,Porto G: Iron overload in b2-microglobulin deficientmice. Immunol Lett 39:105, 1994.

16. Santos M, Schilham M, Rademakers LHPM, Marx JJM, deSousa M, Clevers H: Defective iron b2-microglobulinknock-out mice recapitulates hereditary hemochromatosisin man. J Exp Med 184:1975, 1996.

17. Ten Elshof AE, Brittenham GM, Chorney KA, Page MJ,Gerhard G, Cable EE, Chorney MJ: Gamma delta intra-epithelial lymphocytes drive tumor necrosis factor-alpharesponsiveness to intestinal iron challenge: relevance tohemochromatosis. Immunol Rev 167:223, 1999.

18. Santos M, Clevers H, Marx JJM, de Sousa M: Iron over-load and Heart fibrosis in b2-microglobulin and Rag1-deficient mice. Am J Pathol 157:1883, 2000.

19. Bahram S, Gilfillan S, Kuhn LC, Moret R, Schulze JB,Lebeau A, Schumann K: Experimental hemochromatosis

due to MHC class I HFE deficiency: immune status andiron metabolism. Proc Natl Acad Sci USA 96:13312, 1999.

20. Levy JE, Montross LK, Andrews NC: Genes that modifythe hemochromatosis phenotype in mice. J Clin Invest105:1209, 2000.

21. de Sousa M, Porto G: The immunology of hemochroma-tosis. J Hepatol 28:1, 1998.

22. EASL International Consensus Conference in Haemochro-matosis. J Hepatol 33:485, 2000.

23. Barton JC, McDonnell SM, Adams PC, Brissot P, PowellLW, Edwards CQ, Cook JD, Kowdley KV, and the Hemo-chromatosis Management Working Group: Managementof hemochromatosis. Ann Intern Med 129:932, 1998.

24. Esin S, Gul A, Hodara V, Jeddi-Tehrani M, Dilsen N,Konice M, Andersson R, Wigzell: Peripheral blood T cellexpansions in patients with Behcet’s disease. Clin ExpImmunol 107: 520, 1997.

25. Arosa FA, Porto G, Cabeda JM, Lacerda R, Resende D, CruzE, Cardoso C, Fonseca M, Simoes C, Rodrigues P, Bravo F,Oliveira JC, Alves H, Fraga J, Justica B, de Sousa M:Expansions of CD81CD282 and CD81TcRVbeta5.21 Tcells in peripheral blood of heavy alcohol drinkers. Alco-hol Clin Exp Res 24:519, 2000.

26. Vicente C, Porto G, de Sousa M: Method for establishingserum ferritin reference values depending on sex and age.J Lab Clin Med 116:779, 1990.

27. Porto G, Vicente C, Fraga J, da Silva BM, de Sousa M:Importance of establishing appropriate local reference val-ues for the screening of hemochromatosis: a study of threedifferent control populations and 136 hemochromatosisfamily members. Hemochromatosis Clinical and ResearchGroup. J Lab Clin Med 119:295, 1992.

28. Pantaleo G, Demarest JF, Soudeyns C, Graziosi C, DenisJW, Borrow P, Saag MS, Shaw GM, Sekaly RP, Fauci AS:Major expansion of CD81 T cells with predominant Vbusage during primary immune response to HIV. Nature370:463, 1994.

29. Mizushima N, Kohsaka H, Nanki T, Ollier WER, CarsonDA, Miyasaka N: HLA-dependent peripheral T cell re-ceptor (TCR) repertoire formation and its modification byrheumatoid arthritis (RA). Clin Exp Immunol 110:428,1997.

30. DerSimonian H, Sugita M, Glass DN, Maier AL, Wein-blatt ME, Reme T, Brenner MB: Clonal Va12.11 T cellexpansions in the peripheral blood of rheumatoid arthritispatients. J Exp Med 177:1623, 1993.

31. Fitzgerald JE, Ricalton NS, Meyer A-C, West SG, KaplanH, Behrendt C, Kotzin BL: Analysis of clonal CD81 Tcell expansions in normal individuals and patients withrheumatoid arthritis. J Immunol 154:3538, 1995.

32. Posnett DN, Sinha R, Kabak S, Russo C: Clonal popula-tions of T cell equivalent in normal elderly humans: the Tcell equivalent to “benign monoclonal gammopathy.” JExp Med 179:609, 1994.

33. Davey MP, Meyer MM, Munkirs DD, Babcock D, Braun


MP, Hayden JB, Bakke AC: T-cell receptor variable betagenes show differential expression in CD4 and CD8 Tcells. Hum Immunol 32:194, 1991.

34. Cossarizza A, Kahan M, Ortolani C, Franceschi C, LondeiM: Preferential expression of V beta 6.7 domain on humanperipheral CD41 T cells. Implication for positive selec-tion of T cells in man. Eur J Immunol 21:1571, 1991.

35. Akolkar PN, Gulwani-Akolkar B, Pergolizzi R, BiglerRD, Silver J: Influence of HLA genes on T cell receptor Vsegment frequencies and expression levels in peripheralblood lymphocytes. J Immunol 150:2761, 1993.

36. Gulwani-Akolkar B, Shi B, Akolkar PN, Ito K, Bias WB,Silver J: Do HLA genes play a prominent role in deter-mining T cell receptor V alpha segment usage in humans?J Immunol 154:3843, 1995.

37. Silver J, Gulwani-Akolkar B, Akolkar PN: The influenceof genetics, environment, and disease state on the humanT-cell receptor repertoire. Ann N Y Acad Sci 756:28, 1995.

38. Imberti L, Sottini A, Spagnoli G, Primi D: Expression ofthe human V beta 8 gene product preferentially correlateswith class II major histocmpatibility complex restrictionspecificity. Eur J Immunol 20:2817, 1990.

39. Goldman AS, Palkowetz KH, Rudloff HE, Brooks EG,Schmalstieg FC: Repertoire of V alpha and V beta regionsof T cell antigen receptors on CD41 and CD81 periph-eral blood T cells in a novel X-linked combined immu-nodeficiency disease. Eur J Immunol 22:1103, 1992.

40. Lebron JA, Bennett MJ, Vaughn DE, Chirino AJ, SnowPM, Mintier GA, Feder JN, Bjorkman PJ: Crystal struc-ture of the hemochromatosis protein HFE and character-ization of its interaction with transferrin receptor. Cell93:111, 1998.

41. Mullighan CG, Bunce M, Fanning GC, Marshall SE,Welsh KI. A rapid method of haplotyping HFE muta-tions and linkage disequilibrium in a Caucasoid popula-tion. Gut 42:566, 1998.

42. Murphy S, Curran MD, McDougall N, Callender ME,O’Brien CJ, Middleton D: High incidence of the Cys 282Tyr mutation in the HFE gene in the Irish population—implications for haemochromatosis. Tissue Antigens 52:484, 1998.

43. Porto G, Alves H, Rodrigues P, Cabeda JM, Portal C,Ruivo A, Justica B, Wolff R, de Sousa M: Major histo-compatibility complex class I associations in iron over-load: evidence for a new link between the HFE H63Dmutation, HLA-A29, and non-classical forms of hemo-chromatosis. Immunogenetics 47:404, 1998.

44. Moroi Y, Mayhew M, Trcka J, Hoe MH, Takechi Y, HartlFU, Rothman JE, Houghton AN: Induction of cellularimmunity by immunization with novel hybrid peptidescomplexed to heat shock protein 70. Proc Natl Acad SciUSA 97:3485, 2000.

45. Breloer M, Fleischer B, von Bonin A: In vivo and in vitroactivation of T cells after administration of Ag-negativeheat shock proteins. J Immunol 162:3141, 1999.


T-Cell Receptor Repertoire In Hereditary Hemochromatosis: a Study of 32 Hemochromatosis Patients and 274 Healthy Subjects

Documents