Behçet s disease (BD) is a refractory systemic inflammatory

Transcription

1 Localization of the Pathogenic Gene of Behçet s Disease by Microsatellite Analysis of Three Different Populations Nobuhisa Mizuki, 1,2 Masao Ota, 3 Kazuro Yabuki, 1,2 Yoshihiko Katsuyama, 3 Hitoshi Ando, 2 Gerassimos D. Palimeris, 4 Evangelia Kaklamani, 4 Massimo Accorinti, 5 Paola Pivetti-Pezzi, 5 Shigeaki Ohno, 1 and Hidetoshi Inoko 2 PURPOSE. Behçet s disease (BD) is known to be associated with HLA-B51 in many ethnic groups. However, the pathogenic gene responsible for BD is as yet unknown. To localize the critical region of the pathogenic gene, microsatellite markers distributed around the HLA-B gene were investigated. The BD patients studied were of three ethnic origins: Japanese, Greek, or Italian. METHODS. The total group consisted of 172 BD patients, of whom were 95 Japanese, 55 Greek, and 22 Italian. Eight polymorphic microsatellite markers distributed within 1100 kb of the HLA-B gene were analyzed using PCR and subsequent automated fragment detection by fluorescent-based technology. RESULTS. Among the eight markers, allele 348 of the MIB microsatellite was remarkably common in all three BD populations ( Japanese, Pc ; Greek, Pc ; Italian, Pc 0.11). However, HLA-B51 was found to be the marker most strongly associated with BD in each population ( Japanese, Pc ; Greek, Pc ; Italian, Pc ). In genotypic differentiation between the patients and controls, only HLA-B51 was found to be significantly associated with BD in all three populations. Stratification analysis suggested that significant associations of BD with MICA and other microsatellites resulted from a linkage disequilibrium with HLA-B51. CONCLUSIONS. These results suggest that the pathogenic gene of BD is HLA-B51 itself and not other genes located in the vicinity of HLA-B. (Invest Ophthalmol Vis Sci. 2000;41: ) Behçet s disease (BD) is a refractory systemic inflammatory disorder characterized by four major symptoms consisting of oral aphthous ulcers, ocular lesions, skin lesions, and genital ulcerations, and occasionally by inflammation in tissues and organs throughout the body, including the vascular system, central nervous system, gastrointestinal tract, lungs, kidneys, and joints. This disease is distributed worldwide, but a higher prevalence was found among the Asian and Eurasian populations along the Silk Route stretching to the countries of the Mediterranean region. 1 The incidence of this From the 1 Department of Ophthalmology, Yokohama City University School of Medicine, Yokohama, Kanagawa, Japan; 2 Department of Genetic Information, Division of Molecular Life Science, Tokai University School of Medicine, Kanagawa, Japan; 3 Department of Legal Medicine, Shinshu University School of Medicine, Nagano, Japan; 4 Department of Hygiene and Epidemiology, University of Athens Medical School, Athens, Greece; and 5 Institute of Ophthalmology, University La Sapienza, Rome, Italy. Supported by Grants-in-Aid from the Ministry of Education, Science, Sports, and Culture of Japan ( and ); a grant from the Ministry of Health and Welfare of Japan; and a research grant from the Kanagawa Academy of Science and Technology. Submitted for publication January 3, 2000; revised May 5, 2000, and June 26, 2000; accepted July 13, Commercial relationships policy: N. Corresponding author: Hidetoshi Inoko, Department of Genetic Information, Division of Molecular Life Science, Tokai University School of Medicine, Bohseidai, Isehara, Kanagawa , Japan. hinoko@is.icc.u-tokai.ac.jp disease in Japan is approximately 15:100,000, whereas that in North America and Europe is very low at 1:500,000. 2,3 There have been no reports of clinical cases of BD among Black populations. Our group as well as others have presented evidence for an association between BD and HLA-B51, one of the split antigens of HLA-B5, and this genetic marker was found to be most strongly associated with the disease. 1,4 HLA genes are scattered throughout the HLA gene region, which occupies more than 3500 kb on chromosome 6 (6p21.3). This strong association between BD and HLA-B51 has been confirmed in patients of many ethnic groups, particularly those from the Middle to Far East, including Turks, Greeks, Italians, French, English, Tunisians, Israelis, Jordanians, Iranians, Saudi Arabians, Kuwaitis, Han Chinese, Koreans, Taiwanese, and Japanese. One attractive hypothesis is that BD spread through Asian and Eurasian populations from Japan to the Middle East, together with its associated HLA allele, HLA-B51, as a result of the movement of nomadic or Turkish tribes traveling the Silk Route. 1 The etiology and pathogenesis of BD have remained unclear, but it has been assumed that an infectious agent, immune mechanism, and genetic factor are involved in the onset of this disease. In severe cases, BD predominantly affects men, and so it is not clear whether BD should be classified as an autoimmune disorder. The mean age at onset is within the third decade, children are rarely affected, and few neonatal cases have been reported. The main microscopic finding in most sites of active BD is an immune-mediated occlusive vasculitis. Investigative Ophthalmology & Visual Science, November 2000, Vol. 41, No Copyright Association for Research in Vision and Ophthalmology

2 IOVS, November 2000, Vol. 41, No. 12 Microsatellite Analysis in Behçet s Disease 3703 At the cellular level, CD4 T cells have been found in perivascular inflammatory exudates, and Th1 cells were shown to respond to various stimuli by producing interleukin (IL) 2, interferon (IFN), and tumor necrosis factor (TNF). 5 In several recent studies of BD patients, an increased number of T cells have been found in peripheral blood and in involved tissues, and a phenotypically distinct subset of these cells has been especially observed at sites of inflammation. 6 8 Furthermore, significant T-cell proliferative responses to mycobacterial 65-kDa heat shock protein (HSP) peptides and to their human 60-kDa HSP homologues have also been noted in BD patients. 9,10 Therefore, BD is probably not a simple hereditary disease, and its onset might be triggered by exogenous antigen(s) found on bacteria, viruses, or other microorganisms. In our previous study, the MICA (MHC class I chain-related gene A) gene located 46 kb centromeric of the HLA-B gene appeared to be a strong candidate as the gene responsible for susceptibility to BD, based on its chromosomal localization and its restricted and heat shock induced expression in epithelial cells, 11 its predicted immunologic function as a ligand of V 1 T cells, 12 and a strong association between a certain transmembrane (TM) microsatellite allele, MICA-A6, and BD. 13 However, the possibility of a primary association of MICA-A6 with BD was lower in the Greek population 14 and lack of an association between BD and the MICA-TM microsatellite polymorphism was reported in the Spanish population. 15 As a result of our recent extensive analyses of genetic polymorphism in the extracellular ( 1, 2, and 3) domains of MICA, a strong association between MIC-A009 and BD was found in the Japanese, Palestinian, and Jordanian populations that could be explained by a strong linkage disequilibrium with HLA- B51. 16,17 The MICA gene was therefore suggested not to be directly involved in the pathogenesis of BD, and the pathogenic gene responsible for BD has not yet been clearly localized. Namely, it is still uncertain as to whether HLA-B or an unknown gene in a tight linkage disequilibrium with it is the actual pathogenic gene of this disease. We have recently completed genomic sequencing of the entire HLA class I region spanning approximately 1.8 Mb (1800 kb) from the MICB gene to the HLA-F gene. This region contains the MICA, MICB, HLA-B, and HLA-C genes, and we have identified 758 microsatellite repeats (short tandem repeats [STR]) included within it These microsatellites are densely spread over the BD candidate region and should serve as valuable polymorphic markers for precise mapping of the pathogenic gene. In this study, we performed association analysis using HLA-B and eight microsatellites distributed within a 1100-kb portion around this gene as genetic markers to precisely pinpoint the location of the pathogenic gene responsible for BD. Because genetic recombination, interchromosomal crossover, and genetic exchange may vary depending on ethnic origin, three different populations consisting of Japanese, Greeks, and Italians were enrolled in this study. METHODS Subjects The total group consisted of 95 Japanese patients with BD and 132 healthy Japanese controls, 55 Greek patients with BD and 52 healthy Greek controls, and 22 Italian BD patients and 28 healthy Italian controls, and all were investigated for genetic associations with polymorphic markers in the HLA class I region. Patients were diagnosed according to ISG (International Study Group) diagnostic criteria for Behçet s disease 21 at the uveitis clinic of Yokohama City University (Japanese patients), at the rheumatology and ophthalmology clinics of Athens University (Greek patients), or at the ophthalmology clinic of University La Sapienza (Italian patients). All the BD patients were seen as outpatients for a period of more than 1 year. As for the Greek and Italian patients with BD, these were again diagnosed by some of the authors of the present study (S.O., N.M., and K.Y.) according to standard criteria proposed by the Japan Behçet s Disease Research Committee. All the control subjects were healthy volunteers unrelated to each other or to the patients and were matched to the patients by ethnic origin and age within a range of 5 years. All the patients and controls agreed to a blood examination according to the guidelines of the Declaration of Helsinki. Serologic HLA Class I Typing Serologic HLA class I typing was performed using peripheral blood lymphocytes and the standard microlymphocytotoxicity technique. Analysis of Eight Microsatellite Repeat Polymorphisms Among 38 microsatellites established to serve as informative polymorphic genetic markers of the HLA class I region (PIC: polymorphism content value 0.66, average number of alleles 8.9), 22,23 8 were selected for use in this study. These were C1-2-A, MICA-TM, MIB, C1-4-1, C1-2-5, C1-3-1, C2-4-4, and C3-2-11, and all are located around the MICA and HLA-B genes (Fig. 1). These markers were densely distributed at the following distances from the HLA-B gene: C1-2-A, 147 kb centromeric; MICA-TM, 46 kb centromeric; MIB, 24 kb centromeric; C1-4-1, 6 kb centromeric; C1-2-5, 62 kb telomeric; C1-3-1, 111 kb telomeric; C2-4-4, 282 kb telomeric; and C3-2-11, 912 kb telomeric. PCR primers and conditions were the same as previously described. 22,23 The forward primer was labeled at the 5 end with 6-FAM, HEX, or TET (PE Biosystems, Foster City, CA). To determine the number of microsatellite repeats, PCRamplified products were denatured for 5 minutes at 100 C, mixed with formamide-containing stop buffer, and electrophoresed on 4% polyacrylamide gels with 8 M urea in an automated DNA sequencer ( 373A sequencer; PE Biosystems). The number of microsatellite repeats was estimated with Genescan 672 software (PE Biosystems) by means of the local Southern method using GS500 TAMRA (PE Biosystems) as a size marker. Statistical Analysis Gene and phenotype frequencies at the eight microsatellite and HLA-B loci were estimated by direct counting. The distribution of alleles in patients with BD was compared with that of normal controls, and significance was tested by the 2 method using continuity correction and Fisher s exact probability test with Yate s correction. Furthermore, the P value was corrected by multiplying by the number of microsatellites or HLA-B alleles (corrected P [Pc] value). A Pc value 0.05 was considered statistically significant. To control for the effect of linkage disequilibrium between loci, the Mantel-Haenszel weighted odds ratio (OR) was used. 24 A P value 0.1 was accepted as statistically significant. Genotypic differentiation testing was

3 3704 Mizuki et al. IOVS, November 2000, Vol. 41, No. 12 TABLE 1. Statistical Significance of Alleles Associated with Behçet s Disease in the Japanese Group Marker No. of Alleles Allele Control (n 132)* Patient (n 95)* RR 2 P Pc C1-2-A (12.9) 27 (28.4) MICA-TM 5 A6 62 (47.0) 67 (70.5) A (27.3) 7 (7.4) A5 71 (53.8) 40 (42.1) MIB (12.1) 38 (40.0) (7.6) 27 (28.4) (30.3) 9 (9.5) C (65.9) 82 (86.3) (25.9) 11 (11.6) (36.4) 22 (23.2) HLA-B 25 B51 18 (13.6) 56 (58.9) C (15.9) 38 (40.0) (0.8) 15 (15.8) C (47.0) 57 (60.0) (23.5) 7 (7.4) (80.3) 66 (69.5) C (16.7) 35 (36.8) (81.1) 60 (63.2) C (20.5) 33 (34.7) Alleles showing Pc values of 0.5 are listed. RR, relative risk. * Values in parentheses are percentages. performed with the Markov chain method included within the GENEPOP software package. 25,26 The dememorization period was 1000 steps. A P value 0.01 was considered statistically significant. RESULTS Table 1 shows the statistically significant alleles associated with BD in the Japanese population at the eight microsatellite and HLA-B loci, namely C1-2-A, MICA-TM, MIB, C1-4-1, HLA-B, C1-2-5, C1-3-1, C2-4-4, and C3-2-11, all of which are distributed in the 1100-kb HLA class I region from centromere to telomere 22,23 (Fig. 1). All the alleles in each microsatellite marker were named on the basis of the amplified fragment size length. Among the eight microsatellite markers, allele 348 of MIB was most strongly associated with BD (Pc ). The phenotype frequency (PF) of allele 348 of MIB was particularly high (40.0%) in the Japanese BD patients compared with that (12.1%) of the healthy Japanese controls. Alleles 178 and 202 of C1-2-5 (178, Pc ; 202, Pc ), allele 344 of MIB (Pc ), allele A6 of MICA-TM (Pc ), and allele 217 of C1-4-1 (Pc ) were also represented to a remarkably significant degree in the patient group, whereas allele A5.1 of MICA-TM (Pc ) and allele 336 of MIB (Pc ) were extremely reduced in frequency in this group, with Pc values There were no BD-associated alleles in the C1-2-A, C1-3-1, C2-4-4, or C markers showing Pc values No allele at the C locus gave rise to the statistical significance (Pc 0.05). It was noteworthy that 56 of 95 patients (PF 58.9%) were HLA-B51 positive compared with 18 of 132 controls (PF 13.6%), indicating that this locus had the strongest association with BD (Pc ). Table 2 shows the statistically significant alleles in terms of BD association among the Greek population. Results were similar to those seen in the Japanese population, although the numbers of samples from the Greek patients and controls were relatively small. Therefore, the statistical significance (Pc value) seen in this group was not particularly high compared with that obtained with the Japanese group. In the Greek popula- TABLE 2. Statistical Significance of Alleles Associated with Behçet s Disease in the Greek Group Marker No. of Alleles Allele Control (n 52)* Patient (n 55)* RR 2 P Pc C1-2-A (11.5) 18 (32.7) (25.0) 5 (9.1) MICA-TM 5 A6 30 (57.7) 44 (80.0) MIB (28.8) 38 (69.1) C (63.5) 50 (90.9) (42.3) 11 (20.0) (23.1) 5 (9.1) HLA-B 27 B51 12 (23.1) 43 (78.2) C (7.7) 17 (30.9) C (19.2) 27 (49.1) (26.9) 6 (10.9) Alleles showing Pc values of 0.5 are listed. RR, relative risk. * Values in parentheses are percentages.

4 IOVS, November 2000, Vol. 41, No. 12 Microsatellite Analysis in Behçet s Disease 3705 TABLE 3. Statistical Significance of Alleles Associated with Behçet s Disease in the Italian Group Marker No. of Alleles Allele Control (n 28)* Patient (n 22)* RR 2 P Pc C1-2-A (17.9) 0 (0) NS (0.56) MICA-TM 5 A6 16 (57.1) 17 (77.3) 2.55 NS NS NS (0.68) MIB (17.6) 12 (54.5) (46.4) 4 (18.1) C (64.3) 19 (90.9) (42.9) 3 (13.6) HLA-B 24 B51 5 (17.9) 15 (68.2) C (0) 5 (22.7) C (10.7) 7 (31.8) C (10.7) 10 (45.5) Alleles showing Pc values of 0.5, and C1-2-A-256 and MICA-A6 alleles are listed. RR, relative risk; NS, not significant. * Values in parentheses are percentages. tion, allele 348 of MIB was again most strongly associated with BD (Pc ) among the eight microsatellite markers. Allele 217 of C1-4-1 (Pc ), allele 291 of C1-3-1 (Pc 0.011), and allele 188 of C1-2-5 (Pc 0.040) were significantly increased in frequency in the patient group, with Pc values Among the C1-2-A, MICA-TM, C2-4-4, and C microsatellite markers, no allele showed this degree of significance. Again, it should be noted that 43 of 55 patients (PF 78.2%) were HLA-B51 positive compared with 12 of 52 controls (PF 23.1%), indicating that this marker had the strongest association with BD at any locus investigated (Pc ). Table 3 shows statistically significant alleles in terms of BD association among the Italian group. Because the numbers of samples from the Italians were much smaller (22 patients and 28 controls) than those of the other groups, no microsatellite locus reached statistical significance with Pc value Only allele 348 of MIB (Pc 0.11), allele 213 of C (Pc 0.12), and allele 217 of C1-4-1 (Pc 0.15) were relatively higher in frequency in the patient group than in the healthy controls. It is worth noting that HLA-B51 was again the most significant marker in terms of BD association, with PF of 68.2% for the patients (15 of 22), compared with 21.4% for the controls (6 of 28; Pc ). Age- and sex-stratified analyses were also performed for each population; however, these factors were found not to influence the results. Taken together, allele A6 of MICA, allele 348 of MIB, allele 217 of C1-4-1 and HLA-B51, all within the 46-kb segment between the MICA and HLA-B genes (Fig. 1), were commonly higher in frequency, with considerably low Pc values in the patients of the three ethnic groups. Therefore, to elucidate which of the two loci, MICA or HLA-B, is the actual pathogenic gene for BD, the degree of association of MICA-A6, MIB-348, C or 188 of C1-2-5 with the BD patients stratified for the possible confounding effect of HLA-B51 was estimated for each population by calculation of the Mantel-Haenszel weighted odds ratio. As a result, no primary association with the disease was observed with respect to the MICA-A6, MIB- 348, or C allele in any of the ethnic groups studied (data not shown). In contrast, when an association between HLA-B51 and BD patients stratified for the possible confounding effect of MICA-A6, MIB-348, or C was estimated in each population, a distinctively significant association between B51 and the disease was still observed in all these three different ethnic groups (Table 4). Furthermore, the patient and control groups were compared with respect to genotypic differences in allelic distribution (genotypic differentiation test; Table 5 and Fig. 1). Generally speaking, genotypic distribution should be identical within the same ethnic group. If a variation between patient and control groups is observed at a locus, allelic distribution at this site would be presumed to be under the influence of some TABLE 4. Association of HLA-B51 with BD Stratified for the Effect of MICA-A6, MIB-348, or C in the Japanese, Greek, and Italian Groups Weighted OR 2 P 95% CI Japanese MICA-A OR MIB OR C OR Greek MICA-A OR MIB OR C OR Italian MICA-A OR MIB OR C OR OR, odds ratio; CI, confidence interval.

5 3706 Mizuki et al. IOVS, November 2000, Vol. 41, No. 12 TABLE 5. Genotypic Differentiation between the Normal and Patient Groups of Japanese, Greek, or Italian Ethnicity Marker Japanese (Cont. 132, BD 95) Greek (Cont. 52, BD 55) Italian (Cont. 28, BD 22) C1-2-A MICA-TM MIB C HLA-B C C C C Values are P SE. Cont., controls; BD, Behçet s disease. genetic bias. In the Japanese population, P values were significantly low (P 0.01) at seven loci (MICA-TM, MIB, C1-4-1, HLA-B, C1-2-5, C1-3-1, C2-4-4), and especially low P values (P ) were observed at loci MICA-TM, MIB, HLA-B, and C Furthermore, in the Greek population, P values were significantly low with respect to the HLA-B locus (P ) and relatively low for C1-4-1 (P ), which is located only 6 kb centromeric of HLA-B. In the Italian population, only the HLA-B locus gave rise to a significantly low P value (P ). DISCUSSION It has been well known that BD has a strong association with HLA-B51 in many ethnic groups including the Asian, Eurasian, and Mediterranean populations from Japan through to the Mediterranean basin. In our previous study the MICA gene located 46 kb centromeric of the HLA-B gene has been strongly implicated in conferring a susceptibility to BD. 13 However, the possibility of a primary association of MICA-A6 with BD was lower in the Greek population, 14 and lack of an association between BD and the MICA-TM microsatellite polymorphism was reported in the Spanish population. 15 Furthermore, higher frequencies of specific microsatellite allele could be explained by a strong linkage disequilibrium with HLA-B51. 16,17 Therefore, the MICA gene was suggested not to be directly involved in the pathogenesis of BD, and the pathogenic gene responsible for BD has not yet been clearly localized. In this study, we systematically performed an association analysis using microsatellite markers densely spread over the region of the BD candidate gene. Because BD is one of the few diseases in which the same HLA association is found among many ethnic groups, we not only studied the Japanese BD patients but the Greek and Italian BD patients as well in an effort to pinpoint the critical gene for the development of this disease. Statistically significant differences between the patient and control groups were found at one or more alleles of several microsatellite markers. The MICA-A6, MIB-348, C , and HLA-B51 alleles in particular were commonly increased in frequency with exceedingly low Pc values in the patient groups of the three ethnic populations studied. There were no common alleles that had increased or decreased in frequency at any microsatellite locus centromeric of MICA or telomeric of HLA-B. Accordingly, the MICA-A6 MIB-348 C HLA- B51 haplotype appears to have worldwide predominance among groups of BD patients. Of these alleles, it must be FIGURE 1. Genotypic differences in the Japanese, Greek, and Italian groups with respect to the location of HLA-B and microsatellite loci used for association analysis of Behçet s disease. The locations of all markers used in this study are displayed under the map along with the distance (kb) to neighboring markers. The C1-2-A microsatellite is the most centromeric of the markers. Reciprocal logarithmic P values obtained by genotypic differentiation testing are plotted on the vertical axis. The locations of polymorphic markers are plotted on the horizontal axis. (, Japanese; E, Greek;, Italian).

6 IOVS, November 2000, Vol. 41, No. 12 Microsatellite Analysis in Behçet s Disease 3707 emphasized that HLA-B51 is by far the one most strongly associated allele with BD (Japanese, Pc ; Greeks, Pc ; Italians, Pc ). Furthermore, in stratification analyses of the confounding effect of MICA-A6, MIB-348, and C on HLA-B51 association and vice versa, only the association of B51 was remarkably significant in all the three ethnic groups, indicating a primary role for HLA- B51 in the development of BD. Thus, the significant increase in the frequency of MICA-A6, MIB-348, and C in the patient groups can be explained by a linkage disequilibrium with HLA-B51. Genotypic differentiation testing between the Japanese patients and controls, as calculated by the Markov chain method, clearly showed highly significant P values for four loci: MICA-TM, MIB, HLA-B, and C Two other loci, C1-4-1 and C1-3-1, also had significant P values The allelic distribution at each of these loci was thought to be under the strong influence of some genetic bias. On the other hand, none of the P values for locus C1-2-A or C indicated statistical significance. These results suggest that the search for the pathogenic gene responsible for the development of BD can be narrowed to within the 157-kb interval between the MICA and C1-3-1 loci (Fig. 1). Furthermore, genotypic differentiation testing of the Greek patients and controls, significantly low P values were obtained for two loci only, HLA-B and C1-4-1, located 6 kb centromeric of HLA-B. In the Italian population, a significant P value was obtained only for the HLA-B locus, whereas none of the eight microsatellite markers showed significant P values It was confirmed that the BD patients of the three groups enrolled in this study shared quite similar clinical features according to ISG diagnostic criteria for Behçet s disease and standard criteria proposed by the Japan Behçet s Disease Research Committee. Collectively, HLA-B51 was the allele with the strongest association with BD in each of the groups studied. A highly significant association between HLA-B51 and BD was observed in all three ethnic groups, even after stratification for the possible confounding effect of the nearby genetic markers closely linked to HLA-B51, such as MICA-A6, MIB-348, and C Furthermore, on genotypic differentiation testing between the patients and controls, significant P values were obtained only for the HLA-B locus in any of the three ethnic groups studied. Recently, using Japanese BD patients, we have narrowed the location of the critical segment for the BDcausative gene to 46 kb between the MICA and HLA-B genes by means of the exact test of Hardy-Weinberg proportion for multiple alleles at microsatellite loci. 27 There is still a possibility that a specific allele or mutation in an unknown gene within this segment in a strong linkage disequilibrium with HLA-B51 is directly involved in BD pathogenesis. In fact, our genomic sequence analysis of the HLA class I region suggested the presence of three new genes: NOB1, NOB2, and NOB3 (new organization associated with HLA-B [NOB]), located between the MICA and HLA-B genes. However, all three have recently been established to be pseudogenes with expressed homologues on other chromosomes (Yabuki et al., unpublished observations). Furthermore, it should be emphasized that in this study three microsatellites, MICA-TM (46 kb distant from HLA-B), MIB (24 kb distant from HLA-B), and C1-4-1 (6 kb distant from HLA-B), were located in this critical segment very near to the HLA-B gene, but all three showed a much weaker association with BD than did HLA-B in any statistical analysis used, including the Fisher s exact P value test, the Mantel- Haenszel weighted odds ratio test with 95% confidence intervals, and the genotypic differentiation test. These results clearly indicated that the actual pathogenic gene involved in the development of BD is HLA-B51 (HLA-B*51 at the DNA allele level) itself. However, because HLA-B51 is tightly linked with MICA-A6 and because all the HLA-B51 positive individuals in these populations also possess MICA-A6, the possibility remains that the MICA-A6 allele represents an additional risk factor or further amplifies the risk of developing BD. The existence of HLA-B51 negative BD patients may be due to the influence of other genetic factor(s) and/or various external environmental or infectious agent(s). Use of genetic markers for association analysis of disease, especially complex and multifactorial diseases, has been suggested to have several limitations, one of which is that this method tends to give rise to type I and type II errors. 28,29 In this respect, it appears necessary to increase the number of BD patients from a widened selection of ethnic backgrounds. In addition, various independent statistical methods for association analysis, such as linkage disequilibrium mapping, haplotype analysis, the exact test of the Hardy-Weinberg equilibrium, the haplotype relative risk test, and the transmission disequilibrium test would also be useful in increasing our understanding of the pathogenesis of this disease. References 1. Ohno S, Ohguchi M, Hirose S, Matsuda H, Wakisaka A, Aizawa M. Close association of HLA-Bw51 with Behçet s disease. Arch Ophthalmol. 1982;100: Ohno S. Behçet disease in the world. In: Lehner T, Barnes CG, eds. Recent Advances in Behçet s Disease. London: Royal Society of Medicine Services; Moutsopoulos HM. Behçet s syndrome. In: Isselbacher J, Braunwald E, Wilson D, Martin B, Fauci A, Kasper D, eds. Harrison s Principles of Internal Medicine. New York: McGraw-Hill; 1994: Mizuki N, Ohno S. Molecular genetics (HLA) of Behçet s disease. Yonsei Med J. 1998;38: Valente RM, Hall S, O Duffy JD, Conn DL. Vasculitis and related disorders: Behçet s disease. In: Kelly H, Ruddy S, eds. Textbook of Rheumatology, Vol. II, 5th ed. Philadelphia: Saunders; 1997: Suzuki Y, Hoshi K, Matsuda T, Mizushima Y. Increased peripheral blood gamma delta T cells and natural killer cells in Behçet s disease. J Rheumatol. 1992;19: Hamzaoui K, Hamzaoui A, Hentati F, et al. Phenotype and functional profile of T cells expressing gamma delta receptor from patients with active Behçet s disease. J Rheumatol. 1994;21: Esin S, Gul A, Hodara V, et al. Peripheral blood T cell expansions in patients with Behçet s disease. Clin Exp Immunol. 1997;107: Hasan A, Fortune F, Wilson A, et al. Role of gamma delta T cells in pathogenesis and diagnosis of Behçet s disease. Lancet. 1996;347: Kaneko S, Suzuki N, Yamashita N, et al. Characterization of T cells specific for an epitope of human 60-kDa heat shock protein (hsp) in patients with Behçet s disease (BD) in Japan. Clin Exp Immunol. 1997;108: Groh V, Bahram S, Bauer S, Herman A, Beauchamp M, Spies T. Cell stress-regulated human major histocompatibility complex class I gene expressed in gastrointestinal epithelium. Proc Natl Acad Sci USA. 1996;93: Groh V, Steinle A, Bauer S, Spies T. Recognition of stress-induced MHC molecules by intestinal epithelial gammadelta T cells. Science. 1998;279:

7 3708 Mizuki et al. IOVS, November 2000, Vol. 41, No Mizuki N, Ota M, Kimura M, et al. Triplet repeat polymorphism in the transmembrane region of the MICA gene: a strong association of six GCT repetitions with Behçet s disease. Proc Natl Acad Sci USA. 1997;94: Yabuki K, Mizuki N, Ota M, et al. Association of MICA gene and HLA-B*5101 with Behçet s disease in Greece. Invest Ophthalmol Vis Sci. 1999;40: Gonzalez-Escribano MF, Rodriguez MR, Aguilar F, Alvarez A, Sanchez-Roman J, Nunez-Roldan A. Lack of association of MICA transmembrane region polymorphism and Behçet s disease in Spain. Tissue Antigens. 1999;54: Mizuki N, Ota M, Katsuyama Y, et al. Association analysis between the MIC-A and HLA-B alleles in Japanese patients with Behçet s disease. Arthritis Rheum. 1999;42: Wallace GR, Verity DH, Delamaine LJ, et al. MIC-A allele profiles and HLA class I associations in Behçet s disease. Immunogenetics. 1999;49: Mizuki N, Ando H, Kimura M, et al. Nucleotide sequence analysis of the HLA class I region spanning the 237- kb segment around the HLA-B and -C genes. Genomics. 1997;42: Shiina T, Tamiya G, Oka A, et al. Nucleotide sequencing analysis of the 146-kilobase segment around the IkBL and MICA genes at the centromeric end of the HLA class I region. Genomics. 1998;47: Shiina T, Tamiya G, Oka A, Takishima N, Inoko H. Genome sequencing analysis of the 1.8 Mb entire human MHC class I region. Immunol Rev. 1999;167: International Study Group for Behçet s Disease. Criteria for diagnosis of Behçet s disease. Lancet. 1990;335: Grimaldi MC, Clayton J, Pontarotti P, Cambon-Thomsen A, Crouau- Roy B. New highly polymorphic microsatellite marker in linkage disequilibrium with HLA-B. Hum Immunol. 1996;51: Tamiya G, Ota M, Katsuyama Y, et al. Twenty-six new polymorphic microsatellite markers around the HLA-B, -C and -E loci in the human MHC class I region. Tissue Antigens. 1998;51: Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst. 1950;22: Raymond M, Rouseset F. An exact test for population differentiation. Evolution. 1995;49: Raymond M, Rouseset F. Genepop (version 1.2): population genetics software for exact tests and ecumenicism. J Hered. 1995; 86: Ota M, Mizuki N, Katsuyama Y, et al. The critical region for Behçet s disease in the human major histocompatibility complex is reduced to a 46-kb segment centromeric of HLA-B, by association analysis using refined microsatellite mapping. Am J Hum Genet. 1999;64: Kidd KK. Associations of disease with genetic markers: Dèjà vu all over again. Am J Med Genet. 1993;48: Hodge SE. What association analysis can and cannot tell us about the genetics of complex disease. Am J Med Genet. 1994;54: