Tissue Anligens (1988) 31,254-258

HLA-DR4 genotype frequency and gender effect in familial rheumatoid arthritis 'M. A. KHAN, 'T. S. YAMASHITA, 'T. L. REYNOLDS, *F. WOLFE and 3M. K. KHAN

'Case Western Reserve University, Cleveland Metropolitan General Hospital, Cleveland, Ohio, 'Arthritis Center, Wichita, Kansas and 3Kent State University, Kent, Ohio, U.S.A.

We have studied HLA-A, B, C, and DR antigens in 75 unrelated white families, each with multiple cases of adult-onset definite or classical rheumatoid arthritis (RA) (by ARA criteria). There was no difference in age of onset or serological features of RA between males and females. HLA-DR4 phenotype frequency among female (68%) and male (71%) patients also did not differ significantly. The observed frequencies of HLA-DR4 genotypes (homozygous, heterozygous, and those lacking it) differed significantly (p<0.005) between affected and unaffected individuals. However, the observed genotype distribution did not differ from what is expected given the Hardy-Weinburg equilibrium. There is a direct correlation between the number of DR4 allele(s) carried (two, one, or zero) and the percent affected (p<0.026 for females and p<0.052 for males). These findings highlight the importance of discerning the additional genetic determinants, including those that are gender-associated, which influence susceptibility to RA.

Received for publication 15 December 1987, revised, accepted 1 February 1988

The association of rheumatoid arthritis (RA) with the HLA allele DR4 is well established (Stastny 1978, Khan et al. 1979, Khan & Khan 1986a, 1986b). This relationship indicates that genetic factors play a role in disease predispo- sition. However, precise mechanisms and mode of the DR4-associated susceptibility de- terminants are not understood. We have stud- ied 75 unrelated white families, each with two or more cases of adult-onset definite or classi- cal RA (by American Rheumatism Associ- ation criteria), typed for HLA-A, B, C, and DR antigens. This study is an extension of our earlier report which was based on the first 37 of these 75 families (Khan & Khan 1986a,

1986b). We report our analyses of HLA-DR4 genotypes and gender effect on the occur- rence of RA in order to understand better the immunogenetics of this disease.

Material and Methods

We studied 75 Caucasian families in which a diagnosis of definite or classical RA of adult onset (after 16th birthday) had been made in two or more blood relatives. All of these fam- ilies were unrelated with no known parental consanguinity. Available adult family mem- bers, both affected and unaffected, were ex-

DR 4 GENOTYPE FREQUENCY AND GENDER EFFECT IN FAMILIAL RA 255

amined for symptoms and signs of RA. All members considered as suffering from RA for this analysis fulfilled the requirements for classical or definite RA, as defined by the American Rheumatism Association (Ropes et al. 1958). Those family members who had possible or probable RA, or who appeared to have had chronic arthritis that went into com- plete remission without any residual clinical or radiological findings, were not considered to have had RA for our analyses. Among the 804 family members, several were deceased, were unavailable, or were young children. When an affected or unknown status relative was unavailable or deceased, attempts were made to obtain their medical records from their physicians and their families. Four hundred and fifty-five individuals were typed for histocompatibility antigens, 427 of whom were successfully HLA-DR typed.

HLA typing was performed using a stan- dard microlymphocytotoxicity method (Mic- key et al. 1980) testing for HLA-A, B, C, and DR specificities in the labs of Drs William Braun in Cleveland, Ohio and Peter Stastny in Dallas, Texas. Specificities were assigned according to the 9th International Histocom- patibility Workshop definitions. Rheumatoid factor was tested by the latex agglutination method using the RA-TEST Reagents Kit by Hyland DiagnosticsKooper Biomedical. An- tinuclear antibodies (ANA) were detected by the standard immunofluorescent techniques using rat kidney as substrate (Fritzler 1986). Antibody to double-stranded DNA (anti- dsDNA) was determined by indirect immu- nofluorescence using Crithidia luciliae as sub- strate (Ballou & Kushner 1986). Antibodies to extractable nuclear antigens (anti-ENA) were detected by counter-immunoelectropho- resis against rabbit thymus nuclear extract (Scopelitis et al. 1980). Circulating immune complexes (CIC) were tested by the Clq bind- ing assay (Zubler & Lambert 1977).

Analytical methods used were Chi-square

tests [Zar 19841 of goodness of fit and associ- ation, correlation and affected sib-pair analy- sis of linkage (Day & Simons 1976).

Results and Discussion

We studied 385 males (58 with RA, 24 of them probands) and 419 females (127 with RA, 51 of them probands). The proportion affected among females was higher than among males (31% versus 15%, p<0.0005). Rheumatoid factor was present in 80.4% of the patients with RA; 60.3% had ANA; 75.0% had CIC; 4.9% had anti-dsDNA; and 4.2% had anti- ENA. There was no difference in the serolog- ical features (frequency of RF, ANA, CIC, anti-dsDNA, and anti-ENA) between af- fected males and females. RF and ANA were observed significantly more commonly among DRCpositive family members than those members lacking DR4 (p<0.03). HLA-A, B, C, and DR typing was performed on 427 indi- viduals [138 with RA (96 females and 42 males) and 289 unaffected persons (155 fe- males and 134 males]. An additional 28 indi- viduals were typed for only HLA-A, B, and C (DR typing was unsuccessful).

Among the 144 parents of the probands, 26 (18.1%) were affected (17 of whom were available and DR typed) and 118 (81.9%) were unaffected (48 of whom were available and DR typed); 12 (70.6%) of the affected parents were DR4-positive compared with 29 (60.4%) unaffected parents. There were 254 siblings of the probands, 55 (21.7%) affected (36 of whom were available and DR typed) and 199 (78.3%) unaffected (91 of whom were available and DR typed); 25 (69.4%) of the affected siblings were DR4-positive compared with 49 (53.9%) unaffected siblings. Of the 123 offspring of the probands, 13 (10.6%) were affected (11 of whom were available and DR typed) and 110 (89.4%) were unaffected (64 of whom were available and DR typed). Seven of the 11 (63.6%) affected offspring

256 KHAN ET AL.

Table I. Genotype frequencies of affected and unaffected subjects.

HLA-DR4 Affected Unaffected Total

Obs (Exp) Obs (Exp) Obs (Exp)

+I+ 30 (28.3) 36 (32.2) 66 (59.2) +I- 65 (68.4) 121 (128.6) 186 (199.6) -I- 43 (41.3) 132 (128.2) 175 (168.2)

x: = 0.339 x: = 0.998 x: = 1.971

Total 138 289 427

No significant deviation from Hardy-Weinberg equilibrium was found.

were DRCpositive compared with 41 of the 64 (64.1%) unaffected offspring. Many of the unaffected offspring were much younger than age 40. The mean age of onset of R A in these families is 40.2 for males and 41.6 for females; the difference is not significant.

The availability of HLA-DR4 as a marker for RA has enabled us to further elucidate the genetic factors which lead to the development of the multi-factorial disease. By analysis of affected sib-pairs, we confirm our earlier re- sult that linkage exists between HLA and RA (Khan & Khan 1986b). Of the 75 families we have now studied, there are 37 with affected sib-pairs. Analysis of affected sib-pairs in these 37 families establishes linkage between HLA and RA, since 18 affected sib-pairs share both HLA haplotypes, 15 share one,

and four share none, significantly different (p<0.005) from 9.25, 18.5, and 9.25 expected if the disease and HLA were independent.

The observed HLA genotype frequencies (DR4 homozygous, DR4 heterozygous, and those lacking DR4) among affected and un- affected family members is shown in Table 1. The observed genotype frequency of affected and unaffected individuals differed signifi- cantly x: = 10.69, p<0.005). However, as il- lustrated in Table.1, the observed genotype distribution did not differ from what is ex- pected given the Hardy-Weinberg equilib- rium. DR4 homozygous (+/+) and hetero- zygous (+/-) patients, and those lacking DR4 (-/-), did not differ in mean age of onset or gender.

Table 2 shows the frequency of individuals

Table 2. Genotype frequencies of affected and unaffected subjects by gender.

HLA-DR4 Female Male Female vs male genotype

unaffected affected (YO) unaffected affected (YO) X: P

+I+ +/- -I-

20 23 (55%) 16 7 (30?'0) 3.2 NS 56 42 (43%) 65 23 (26%) 5.7 pX0.025 79 31 (28%) 53 12 (18%) 2.1 NS

NS = not significant.

DR 4 GENOTYPE FREQUENCY AND GENDER EFFECT IN FAMILIAL RA 257

Female (n = 251) r = 0.998 (p < 0.026)

Male (n = 176) r = 0.982 (p < 0.052)

-I- +I- +I+ DR4 Genotype

Figure 1 . There is a linear association between the number of DR4-allele(s) carried (2, 1, or 0) and the percent affected (p<0.026 for females and p<0.052 for males).

who were DR4 homozygous, DR4 heterozy- gous, and who lacked DR4, grouped by gen- der and affection status. DR4 frequency (+/+ and +/-) among male (30/42 (71%)) and fe- male (65196 (68%)) patients also did not differ significantly. A relative excess of RA was ob- served in females when compared with males for the three DR4 genotypes (+/+, +/-, -/-), but the difference is statistically signif- icant only among DR4 heterozygotes (Table 2). There was a linear association between the number of DR4 allele(s) carried (two, one, or none) and the percent affected (Figure l ) , indicating that the likelihood of developing RA was directly proportional to the number of DR4 alleles carried. It is of interest that Nepom et al. (1986) had noted that there was an excess of DR4(Dw4) and DR4(Dw14) ha- plotypes among their DR4 homozygous sero- positive childhood-onset RA patients. They suggested that this Dw heterozygous predis- position may indxate the possibility of gene complementation (two gene effect) that re- sults in increased susceptibility to RA, similar to what has been observed in insulin-depend- ent diabetes mellitus.

Acknowledgments

This work was supported by grants from the National Institute of Arthritis, Diabetes, and Digestive Kidney Diseases through grant number AM 20618 to the Northeast Ohio Multipurpose Arthritis Center; and the Irma Bender Arthritis Research Fund

References Ballou, S. P. & Kushner, I. (1986) Crilhidia luciliae

immunofluorescence test for antibodies to DNA. In Manual of Clinical Immunology, 3rd edn., eds. Rose, N. R. et al., pp. 740-743, American So- ciety of Microbiology, Washington, D.C.

Day, N. E. & Simons, M. J. (1976) Disease suscep- tibility genes - their identification by multiple case family studies. Tissue Antigens 8 , 109-119.

Fritzler, M. (1986) Immunofluorescent antinuclear antibody. In Manual of Clinical Immunology, 3rd edn., eds. Rose, N. R. et al., pp. 733-739, Amer- ican Society of Microbiology, Washington, D.C.

Khan, M. A., Kushner, I., Ballou, S. P. & Braun, W. E. (1979) Familial rheumatoid arthritis and HLA-DRw4 (letter). Lancel i, 921-922.

Khan, M. A. & Khan, M. K. (1986a) HLA studies in familial and sporadic rheumatoid arthritis. Dis- ease Markers 4, 67-76.

Khan, M. A. & Khan, M. K. (1986b) Analysis of genetic susceptibility of familial rheumatoid ar- thritis. Disease Markers 4, 77-84.

Mickey, M. R. & Terasaki, P. I. (1980) The serolog- ical data of the 8th Workshop and summary analysis. In Histocompatibility Testing 1980, ed. Terasaki, P. I., pp. 21-136, UCLA Tissue Typing Laboratory, Los Angeles.

Neporn. R . S.. Palmer, J., Kim, S. J.. Hansen, J. A, , Holbeck, S. L. & Nepom, G. T. (1986) Spe- cific genomic markers for the HLA-DQ subre- gion discriminate between DR4+ insulin-de- pendent diabetes mellitus and DR4+ seropos- itive juvenile rheumatoid arthritis. J Exp Med 164, 345-350.

Ropes, M. W., Bennett, G. A, , Cobbs, S., Jacox, R. & Jessar, R. A. (1958) Revision of diagnostic criteria for rheumatoid arthritis. Bull Rheum Dis 9, 175-176.

Scopelitis, E., Biundo, Jr. J. J . & Alspaugh, M. A. (1980) Anti-SS-A antibody and other antinuclear antibodies in systemic lupus erythematosus. Ar- thritis Rheum 1, 287-293.

Stastny, P. (1978) Association of the B-Cell alloan-

258 KHAN ET AL.

tigen DRw4 with rheumatoid arthritis. N Engl J Med 298,869-871,

Zar, J. H. (1984) Biostatistical Analysis, 2nd edn., pp. 40-70, Prentice-Hall, N. J .

Zubler, R. H. & Lambert, P. H. (1977) 1231-Clq- binding test for soluble immune complexes. Ann Rheum Dis 36, 27-30. U.S.A.

Address: Muhammad A . Khan, M.D. Division of Rheumatology Cleveland Metropolitian General Hospital 3395 Scranton Road Cleveland, Ohio 44109