Rheumatic heart disease (RHD) is an autoimmune sequela

Transcription

1 HLA Class II Associations With Rheumatic Heart Disease Are More Evident and Consistent Among Clinically Homogeneous Patients Yajaira Guédez, MD; Alyaa Kotby, MD; Maha El-Demellawy, PhD; Ashraf Galal, MD; Glenys Thomson, PhD; Salah Zaher, MD; Samir Kassem, MD; Malak Kotb, PhD Background Discrepancies in reported HLA class II associations with rheumatic heart disease (RHD) may have been due to inaccuracies of serological typing reagents and/or lack of defined clinical classification of patients analyzed. The molecular association between HLA and RHD was investigated in patients with defined clinical outcome. Methods and Results Class II allele/haplotype distribution was determined in 2 groups of RHD patients (n 88) and a control group (n 59). Patients were divided into the mitral valve disease (MVD) category (ie, those with mitral regurgitation with or without mitral stenosis) and the multivalvular lesions (MVL) category, with impairment of aortic and/or tricuspid valves in addition to mitral valve damage. The MVD category (n 65) accounted for 74% of patients and included significantly fewer recurrent RF episodes compared with MVL patients (P 0.002). Conclusions Significant increases in DRB1*0701 and DQA1*0201 alleles and DRB1*0701-DQA1*0201 haplotypes were found in patients. Removal of the MVL patients from analysis increased the strength of HLA associations among the MVD sample. The frequency of DQA1*0103 allele was decreased and the DQB1*0603 allele was absent from the patient group, suggesting that these alleles may confer protective effects against RHD. DQ alleles in linkage disequilibrium with DR alleles appear to influence risk/protection effect: whereas the DRB1*13 DQA1* DQB1*0301 haplotype showed a trend toward risk, the DRB1*13-DQA1*0103-DQB1*0603 haplotype was absent in the RHD sample. Our data indicate that certain class II alleles/haplotypes are associated with risk or protection from RHD and that these associations appear to be stronger and more consistent when analyzed in patients with relatively more homogeneous clinical manifestations. (Circulation. 1999;99: ) Key Words: rheumatic heart disease antigens genetics Rheumatic heart disease (RHD) is an autoimmune sequela of group A streptococcal infections and one of the leading causes of morbidity and mortality in many parts of the world. 1 9 The disease is often preceded by rheumatic fever (RF) episodes that may, in susceptible individuals, progress to a chronic valvular disease. The relatively low attack rate of RF after untreated streptococcal pharyngitis (up to 2% to 3%), 10 the relatively high concordance rate for RF in monozygotic twins (19%) in comparison to dizygotic twins (2.5%), 10 and high familial incidence of RF 11 suggest the involvement of host genetic factors in susceptibility to RF with consequential progression to RHD. Several studies have suggested that genetic susceptibility to RF and RHD is linked to HLA class II alleles However, there has been an apparent discrepancy as to the nature of susceptibility and/or protective alleles. This may have been partly because the majority of investigations used less accurate serological HLA typing methods that can generate false results and fail to discriminate between allelic subgroups. Ethnic differences in the distribution of HLA alleles and the contribution of other genes that could display, in different populations, distinct linkage disequilibrium patterns with HLA DR or DQ alleles may also have contributed to these apparent discrepancies. Another confounding problem was the failure of some studies to separate RF patients with and without carditis or to differentiate between subcategories of RHD for analysis of genetic susceptibility to RHD. Few studies have attempted to analyze the HLA class II associations with specific clinical forms of RHD. 14,20 The question remains as to whether an association between class II allotype and RHD actually exists, and if so, whether it varies for different ethnic groups or is consistent among patients who share the same pattern of valve lesions regardless of their ethnic origin. Received September 22, 1998; revision received February 19, 1999; accepted March 16, From the Veterans Affairs Medical Center (Y.G., M.E., M.K.) and The University of Tennessee, Departments of Surgery, Microbiology, and Immunology (Y.G., M.E., M.K.), Memphis; the Pediatrics Department, Faculty of Medicine, Ain Shams University, Cairo (A.K.) and Alexandria University (A.G., S.Z., S.K.), Egypt; and the Department of Integrative Biology, University of California, Berkeley (G.T.). Correspondence to Malak Kotb, PhD, University of Tennessee, Memphis, 956 Court Ave, Suite A202, Memphis, TN mkotb@utmem1.utmem.edu 1999 American Heart Association, Inc. Circulation is available at

2 Guédez et al June 1, RF is an inflammatory condition that can have different manifestations, including polyarthritis, chorea, and carditis. 24 In most countries, the incidence of carditis in RF patients ranges from 30% to 90%, and the majority of RF patients with carditis develop RHD. 25 In Egypt, for example, the incidence of carditis among RF patients is 60%, and 90% of carditis patients develop RHD. 25 Genetic associations are more likely to be detected in clinically homogeneous groups of patients, and thus it is important to separate carditis patients from patients with other RF sequelae but without carditis. Failure to do so may mask important genetic associations by inclusion of clinically heterogeneous groups. We noted that in studies in which RF patients with carditis were analyzed separately from other RF sequelae or in which only RHD patients, the majority of whom had mitral valve disease (MVD), were studied, the reported HLA associations were rather similar ,18,23 Based on the above observations, we hypothesized that HLA class II associations with RHD may be more consistent if analyzed in patients with relatively homogeneous clinical outcome. We reasoned that by focusing on the RHD patients with documented patterns of mitral valve damage and a history of RF, we would be excluding RF patients without carditis and misdiagnosed patients with RF, which can skew the genetic analyses. Results of the present study support our hypothesis and indicate that certain HLA class II alleles and haplotypes are associated with risk/protection from RHD and that these associations are more evident in patients with MVD. Methods Subjects Two groups of unrelated patients with established RHD (n 88; 59% female, 41% male) from northern Egypt (Cairo and Alexandria) were analyzed both separately by clinic locale and then in combination for the purpose of cross-validation analysis. RHD-1 patients (n 53; mean age, years; 58.5% female and 41.5% male) attended the cardiology clinic at Ain Shams University in Cairo. RHD-2 patients (n 35; mean age, years; 60% female and 40% male) were from El-Shatby Clinic in Alexandria. Patients presented initially with an illness that fulfilled the updated Jones criteria for RF. 24 Diagnosis of valve lesions in follow-up visits was made by echocardiography and/or catheterization. RHD patients were further categorized into MVD (n 65; mean age, years) or multivalvular lesion (MVL; n 20; mean age, years) categories. The MVD category consisted of mitral regurgitation (n 49; 55.7% of all RHD; mean age, years) plus mixed mitral regurgitation mitral stenosis (n 16; 18% of all RHD; mean age, years). The MVD category in the RHD-1 and RHD-2 groups represented 62% (MVD-1) and 91% (MVD-2), respectively. Control subjects were unrelated healthy individuals (n 59; mean age, ; 32% female and 68% male) selected at random from the same geographic area who appeared to be free of autoimmune diseases and had no family history of RF. Internal review board approval (IRB 5938) was obtained for this study. DNA Typing of HLA Class II Genes DNA from whole blood of patients and control subjects was extracted by the Chelex method. 26 Low-resolution HLA-DR typing for DRB1*01 through 18 as well as for DRB3 (DR52), DRB4 (DR53), and DRB5 (DR51) was performed by polymerase chain reaction (PCR) with amplification with sequence-specific primers, 27,28 and 13 DNA control subjects provided by the 12th International Histocompatibility Workshop (IHW) were used as test references. 29 DNA was amplified in a final 50 L of1 PCR buffer (50 mmol/l KCl; 10 mmol/l Tris-HCl, ph 8.3; 0.001% wt/vol gelatin); each reaction mixture contained 200 mol/l of each dntp, 0.4 mol/l DR-specific primers, 0.2 mol/l DR internal control primers, with 1.25 U Taq polymerase (Promega) and 10 to 20 L genomic DNA. Amplification occurred for a total of 33 cycles with 1 minute each of denaturation at 95 C, annealing at 55 C, and extension at 72 C, with the last cycle having an additional 6 minutes of extension at 72 C. PCR products were separated on 2% agarose, the amplified bands were visualized, and the DR type was deduced. High-resolution subtyping of certain DRB1 and all DQB1 alleles was performed with the Innolipa reverse hybridization typing system (Innogenetics) according to the manufacturer s instructions. 30 A nonradioactive oligotyping method was used to examine HLA- DQA1 polymorphism. Primers, probes, and procedures were those indicated in the 12th IHW protocols. 29 Amplified DNA was immobilized on nylon membranes and hybridized with a panel of 19 digoxigenin-labeled probes. An anti digoxigenin alkaline phosphatase Fab antibody fragment (Boehringer Mannheim) was added and bound to any hybrid previously formed. Membranes were incubated with a chemiluminescent substrate (AMPPD, Boehringer Mannheim) and exposed briefly to x-ray films, and reactivity patterns were recorded from the autoradiograph and analyzed. Statistical Analysis HLA-DRB1, DQA1, and DQB1 allele frequencies in patients and healthy unrelated control subjects were compared. Typing of all 3 loci was performed on all patients and control subjects, and haplotype assignments were made with known patterns of linkage disequilibrium in Caucasians 31,32 and were confirmed by family-based genotyping of 12 Egyptian families with 1 to 7 siblings, which generated 11 haplotypes, confirming that the linkage disequilibrium patterns observed were the same as those reported for Caucasians. 31,32 Allele and haplotype frequencies were determined by the method of gene counting. Tests for differences in predisposing and protective effects of HLA class II alleles, haplotypes, and genotypes were performed by use of the odds ratio (OR) method 33 and the relative predispositional effects (RPE) method. 34 Statistical significance was examined by Fisher s exact test. The strength of associations was verified by the cross-validation method 35 (ie, significant associations found in RHD-1 or MVD-1 patients were crossvalidated against RHD-2 or MVD-2 patients, respectively). Crossvalidation against previous studies reporting the same association was also performed. Results Frequency of DRB Alleles in RHD Patients and Control Subjects The frequency of DRB1*0701 was significantly increased in patients over control subjects in the RHD-1 sample (OR 2.3, P 0.03), the RHD-2 sample (OR 2.9, P 0.010), and the combined sample (OR 2.5, P 0.007), where the frequency in patients was 21% compared with 9.3% in control subjects (Table 1). The strength of the DRB1*0701 association increased when the MVD group was analyzed separately (Table 1); the frequency of DRB1*0701 was 24%, compared with 9% in control subjects (OR 3.0, P 0.002), and the significant association of this allele was cross-validated in the MVD-1 and MVD-2 groups (P 0.003, and 0.018, respectively). The frequency of DRB1*0701 in patients with MVL (10%) was lower than in the MVD patients but not significantly different and also was not significantly different from that in control subjects. The DRB1*13 allele showed a trend toward a higher frequency in RHD patients compared with control subjects,

3 2786 HLA Association in RHD TABLE 1. DRB1 Allele Distribution in RHD Patients Compared With Healthy Control Subjects DRB1* Alleles *05 *06 Group *01 *02 *03 *04 *11 *12 *13 *14 *0701 *08 *0901 *1001 RHD 1 gf(p) OR/ (0.030) 2.3/ (n 106) RHD 2 (n 70) gf(p) OR/ (0.012) 0.19/ (0.046) 0.31/ (0.010) 2.9/ All RHD gf(p) OR/ (0.007) 2.5/ (n 176) MVD-1 (n 66) gf(p) OR/ (0.003) 3.4/ MVD-2 (n 64) gf(p) OR/ (0.018) 0.21/ (0.018) 2.72/ All MVD gf(p) OR/ (0.002) 3.0/ (n 130) MVL (n 40) gf(p) OR/ Control subjects gf (n 118) *The asterisk preceding the allele number indicates the allele molecular designation (typing at DNA level). Group: RHD-1/MVD-1 and RHD-2/MVD-2 are RHD or MVD patients (mitral regurgitation [MR] and MR mitral stenosis cases) from Cairo and Alexandria, respectively. MVL indicates multivalvular lesion cases (MR plus aortic and/or tricuspid regurgitation). Bold-face type highlights statistically significant associations for patients vs control subjects. gf (gene frequency), P (probability), OR (odds ratio), and 2 values are reported only for significant associations (P 0.05). n number of haplotypes (eg, 106 haplotypes from 53 individuals). Nature of valve lesion was not reported on 3 patients. whereas the DRB1*10, DRB1*02, and DRB1*11 alleles showed decreased frequencies, particularly in RHD-2 patients. However, neither of these effects was significant or cross-validated, even with RPE analysis after removal of DRB1*0701. Frequency of DQA1 and DQB1 Alleles in RHD Patients and Control Subjects A possible protective effect of DQA1*0103 was seen (Table 2); the frequency of this allele was lower in all RHD patients than in control subjects (P 0.048), with significantly negative association found in the RHD-2 (P 0.001) but not in RHD-1 patients. The strength of the negative association between the DQA1*0103 allele and RHD was increased when the MVD group was analyzed separately (P 0.016) TABLE 2. Distribution of DQA1 Alleles in RHD Patients and Healthy Control Subjects and was strongest in the MVD-2 group (P 0.002), but it was not cross-validated with MVD-1 patients. An increased frequency of DQA1*0201 was seen among patients. The association of DQA1*0201 with risk was stronger in the MVD category (P 0.013), and its significance was cross-validated in the MVD-1 (P 0.03) and MVD-2 (P 0.04) groups (Table 2). RPE analysis failed to reveal further associations after removal of either DQA1*0103 or DQA1*0201 alleles from the analysis. No significant association with either risk or protection was found in the MVL patients; however, the DQA1*0401 allele showed a trend toward increased frequency in this group. Similarly, although the DQA1*0601 allele was completely absent in MVD-1 and MVD-2 patients, it was found with a 3-fold higher frequency in MVL patients than in control DQA1* Alleles Group *0101/4 *0102 *0103 *0201 * *0401 * *0601 RHD 1 (n 106) gf(p) OR/ RHD 2 (n 70) gf(p) OR/ (0.001) 0.05/ (0.026) 2.39/ All RHD (n 176) gf(p) OR/ (0.048) 0.44/ (0.039) 1.93/ MVD-1 (n 66) gf(p) OR/ (0.030) 2.4/ MVD-2 (n 64) gf(p) OR/ (0.002) (0.05/ (0.042) 2.3/ MVD (n 130) gf(p) OR/ (0.016) 0.30/ (0.013) 2.3/ MVL (n 40) gf(p) OR/ Control subjects (n 118) gf *Typing methods did not discriminate between DQA1*0301/0302, DQA1 *0101/0104, or the DQA1 *0501/*0502/*0503. Group: see Table 1 footnote. Bold-face type highlights statistically significant associations for patients vs controls. gf (gene frequency), P (probability), OR (odds ratio), and 2 values are reported only for significant associations (P 0.05). n number of haplotypes (eg, 106 haplotypes from 53 individuals). Nature of valve lesion was not reported on 3 patients.

4 Guédez et al June 1, TABLE 3. Distribution of DQB1 Alleles in RHD Patients and Healthy Control Subjects DQB1* Alleles Group * *0301 *0302 *0303 *0402 *0501 *0601 *0602 *0603 *0604 *0607 Others RHD 1 gf(p) OR/ (0.020) 0.08/ (n 106) RHD 2 (n 70) gf(p) OR/ (0.022) 0.09/ # All RHD gf(p) OR/ (0.004) 0.05/ (n 176) MVD-1 (n 66) gf(p) OR/ # MVD-2 (n 64) gf(p) OR/ (0.029) 0.10/ # MVD (n 130) gf(p) OR/ (0.037) 0.21/ (0.011) 0.07/ MVL (n 40) gf(p) OR/ Control subjects gf (n 118) *Alleles DQB1 *0201/*0202 could not be distinguished and were referred to as DQB1 * Group: see Table 1 footnote. Bold-face type highlights statistically significant associations for patients vs controls. Alleles having a gf 3% in all the groups under study: DQB1 *0502, *0503, *0605, *0608, and *0609. gf (gene frequency) (ie, allele frequency), P, OR, and 2 values are reported only for significant associations (P 0.05). n number of haplotypes (eg, 106 haplotypes from 53 individuals). Nature of valve lesion was not reported on 3 patients. #Despite total absence of this allele, the P values approached but did not reach significance, probably because of sample size (RHD-2, P 0.058; MVD-1, P 0.066; MVD-2, P 0.071). subjects (P NS). The DQB1*0601 allele was absent from only the RHD-2 group (P 0.02) and thus the MVD-2 (P 0.03). Interestingly, the DQB1*0603 allele was completely absent in all patient groups, including patients with MVL, whereas its frequency was 5% in control subjects (P for all RHD and 0.01 for MVD, Table 3). This suggests a protective effect for this allele in RHD. DRB1-DQA1-DQB1 Haplotypes Associated With RHD The DRB1*0701-DQA1*0201 haplotype in this population was in linkage disequilibrium with either the DQB1*0201 or DQB1*0303 alleles and was significantly increased in the combined RHD (P 0.018) and the RHD-2 group (P 0.01), Table 4. This association was stronger in the MVD group (P 0.004) and was cross-validated between the MVD-1 (P 0.012) and the MVD-2 (P 0.018) groups. Predisposing TABLE 4. Significant HLA Haplotype Associations With Predisposition/Protection in RHD Patients effects of DRB1*0701 DQA1*0201 DQB1*0201-2/*0303 haplotypes may have been conferred by DRB1*0701 and/or DQA1*0201 alleles, which were significantly associated with risk in MVD patients. The frequency of DRB1*13 DQA1* DQB1*0301 was higher in all patient categories and was significantly increased in the combined RHD sample (OR 3, P 0.04), Table 4. Removal of the MVL patients increased the strength of this association in the MVD group (P 0.03), with a higher association found in the MVD-2 group (P 0.02). Failure to cross-validate the trend for a predisposing effect of this haplotype may have been related to the sample size. Highresolution subtyping revealed that DRB1*13 suballeles found on this susceptibility haplotype were DRB1*1303, *1304, *1307, *1310, and *1314. Interestingly, the DRB1*13-DQA1*0103-DQB1*0603 haplotype was completely absent in all patient groups, DRB1-DQA1-DQB1 Group *15-*0103-*0601 *0701-*0201-*0201-2/0303 *13-* *0301 *13-*0103-*0603 RHD-1 (n 106) HF(P) OR/ (0.039) 0.09/4.6 RHD-2 (n 70) HF(P) OR/ 2 0 (0.022) 0.06/ (0.010) 2.9/ All RHD (n 176) HF(P) OR/ (0.018) 2.2/ (0.044) 2.9/ (0.009) 0.06/7.6 MVD-1 (n 66) HF(P) OR/ (0.012) 2.9/ MVD-2 (n 64) HF(P) OR/ (0.029) 0.10/ (0.018) 2.72/ (0.022) 4.1/ MVD (n 130) HF(P) OR/ (0.037) 0.21/ (0.004) 2.8/ (0.033) 3.2/ (0.023) 0.08/5.6 MVL (n 40) HF(P) OR/ Control subjects (n 118) HF(P) OR/ Significant increase in the frequency of the DRB1 *0701-DQA1 *0201-DQB1 * haplotype alone was found in all RHD cases (P 0.045) and the MVD group (P 0.014). HF indicates haplotype frequency. The P, OR, and 2 values are reported only for significant associations (P 0.05). n number of haplotypes (eg, 106 haplotypes from 53 individuals). Nature of valve lesion was not reported on 3 patients.

5 2788 HLA Association in RHD TABLE 5. Previously Reported HLA Class II Associations in Patients With RF and RHD HLA Class II Associations Population Risk Protection Method* Clinical Form Patient Age, y (mean SD) No. of Patients No. of Control Subjects North India 21 DR3 Serology RHD Unknown North India 17 DR3 DR2 Serology RHD Unknown DQw2 Black (America) 20 DR2 Serology RHD (MR) Unknown Caucasian (America) 20 DR4 Serology RHD (MR) Unknown Black (South Africa) 23 DR1 Serology RHD (60% MVD) DR6 Caucasian (America) 22 DR4 DR6 Serology RHD Saudi 19 DR4 Serology RHD (MVL) Turkish 16 DR3 DR5 Serology RHD (51% MVD) DR7 Brazilian 18 DR7 Serology RHD/RF (carditis) DR53 Brazilian 15 DR53 RFLP RHD/RF Unknown DR16(2) Japanese 14 DRB1*1405 DQA1*0104 DQB1*0503 PCR-SSO RHD (86% MVD) Mexican 13 DRB1*0301 DRB1*0701 DQB1*0201 whereas its frequency was 4% in control subjects (P 0.009), Table 4. The DRB1*13 suballeles found on this protective haplotype were DRB1*1301, *1308, and *1316. The trend for a protective effect of this haplotype may have been conferred by the DQA1*0103 and/or DQB1*0603 alleles, which were individually associated with protection (Tables 2 and 3). The trend toward protective association with the DRB1*15-DQA1*0103-DQB1*0601 haplotype found in patients from Alexandria (RHD-2/MVD-2) but not Cairo was unexpected because of the ethnic homogeneity between these geographically proximal cities. It is also possible that certain genetic elements are associated with RHD regardless of ethnicity or strain of streptococcus, whereas other elements may be sensitive to differences in the serotypes of streptococcal strains circulating in different communities. Discussion The HLA association in RHD has been investigated in several studies, and although the data seem to support the hypothesis that such a link exists, a consistent association with specific allele(s)/haplotypes was lacking (Table 5). This might be attributable to difficulties associated with serological HLA typing methods used by the majority of these investigations, which can generate false results. For example, recently Ayoub et al 36 found that the previously described DR4 association in Caucasian American RHD patients did not hold when HLA was reassessed by DNA typing methods. Old serological typing reagents did not differentiate between DR allelic splits such as DR13 and DR14, which are splits of PCR-SSO RHD (52% MVD) 75% 40 y *Method: RFLP indicates restriction fragment-length polymorphism; PCR-SSO, PCR and sequence-specific oligonucleotide. MR indicates mitral regurgitation. Bold-face type highlights associations observed in this study. Recently, Ayoub et al 36 found that the previously described association with DR4 did not hold when analysis was performed by molecular typing. DR6, and provided no data on subdivisions of DQA and DQB alleles that are on the same haplotype and that may influence the effect of DR alleles on disease. Therefore, associations apparent only with molecular subdivisions of suballeles were missed. The predisposing/protective effects of particular class II suballeles may have been masked by other suballeles that are not associated with disease, and the differential effect on disease of a particular allele present on distinct haplotypes would not have been detected. In addition to these technical difficulties, different outcomes of RF may be associated with distinct genetic elements, and it is important and useful to analyze clinically homogeneous RF categories separately. The findings of our study (summarized in Table 6) underscore the importance of analyzing genetic association with disease in clinically homogeneous patients by molecular methods. If we focus on RHD patients with MVD, there is a high likelihood that these patients had carditis during their RF attack, and thus misdiagnosed patients or those with other categories of RF without carditis, who may mask genetic associations with RHD, would be excluded from the analysis. This distinction was particularly important in our study because as many as 40% of RF patients in Egypt are without carditis, 25 and this group would have masked the associations detected here. Furthermore, we believe that analysis of MVD and MVL patients separately increases the clinical homogeneity of patients. Indeed, our study shows that HLA associations were stronger when the MVL group was excluded from the analysis. Interestingly, the MVD patients experi-

6 Guédez et al June 1, TABLE 6. Summary of HLA Alleles/Haplotypes Associated With RHD in This Study Risk Protection DRB1*0701 DQA1*0201 DQA1*0103 DQB1*0603 DRB1*0701-DQA1*0201 DRB1*13-DQA1* DQB1*0301 DRB1*13-DQA1*0103-DQB1*0603 *The asterisk preceding the allele number indicates the allele molecular designation (typing at DNA level). Bold-face type highlights associations cross-validated in this study. The associations with DRB1*0701 allele and DRB1*0701-related haplotypes were cross-validated in this study as well as with previous studies in Turkish, Brazilian, and Mexican populations (Table 5). The DR 6 (13) haplotypes showed strong trends toward risk/protection but were not cross-validated in this study. enced a significant 5-fold fewer acute recurrent RF episodes compared with the MVL patients, with mean rates of RF recurrence of 0.7 and 3.3, respectively (P 0.002). In accordance with our hypothesis, studies in which RHD patients were analyzed separately from RF without carditis 15,18 and in which the MVD category accounted for the majority of cases 13,14,16,23 seem to have found similar class II associations with RHD, and these are consistent with the findings reported here (Tables 5 and 6). For example, increased frequencies of DRB1*0701 (DR7), DRB1*0301 (DR3), DR6, and DQB1*0201 alleles were found in RHD patients from different ethnic groups. 13,14,16,18,23 Consistent with these studies, we found a significant increase in the frequency of the DRB1*0701 allele in RHD patients (P 0.007), and this association with risk was even stronger when analysis was focused on the MVD group (P 0.002). Similarly, the entire DRB1*0701-DQA1*0201 haplotype was significantly associated with risk in our entire RHD sample (particularly in combination with DQB1*0201-2), and the strongest association with this haplotype was found in the MVD category (P 0.004). This association was crossvalidated in our MVD-1 and MVD-2 samples and also against previous studies in which the majority of patients had MVD. 13,16,18 Interesting trends and associations were also clustered around the DR6-related haplotypes. The DR6 antigen has 2 phenotypic splits encoded by the DRB1*13 or DRB1*14 alleles, each with several subtypes. In our study, the DRB1*13 DQA1* DQB1*0301 haplotype showed a trend toward association with risk and was significantly increased in patients with MVD. By contrast, the DRB1*13- DQA1*0103-DQB1*0603 haplotype was absent in all patient groups, which suggested a protective influence of this haplotype. An association between DR6 haplotypes and RHD was also reported in other studies. 14,23 Koyanagi et al 14 reported a high frequency of the DRB1*1405-DQA1*0104- DQB1*0503 haplotype in Japanese RHD patients with predominant MVD. A high frequency of DR6 was found in blacks from South Africa 23 (60% with MVD). Interestingly, a negative association between RHD and DR6 was reported in the Utah study 22 ; however, because serological typing was performed and thus no information on DR6 splits, suballeles, or haplotypes was provided, it is possible that the negative association with DR6 was conferred by the DRB1*13- DQA1*0103-DQB1*0603 haplotype, which in our study is showing a protective trend. Therefore, depending on the DR6 splits (DR13 or DR14) or the DRB1*13/DRB1*14 suballeles and the nature of additional elements present on the same haplotype (ie, DQA and DQB alleles), the DR6 haplotypes may either exert risk or confer protection from RHD. More than 1 predisposing allele and haplotype were detected in our study. Although individual alleles on risk haplotypes were significantly associated with RHD, we should not rule out the possibility that other genetic elements on these haplotypes may be more directly involved in disease susceptibility. In the MVD group, the frequencies of DRB1*0701-DQA1*0201 DQB1*0201-2/0303 and DRB1*13 DQA1* DQB1*0301 haplotypes were 22% and 10%, compared with 9% and 3% in control subjects, respectively. These levels of frequencies reported here as well as in Japanese 14 and Mexican 13 patients are to be expected for autoimmune diseases of infectious origin in which variations in the microorganisms in this case differences in serotype of group A streptococcal strains may influence outcome. Inasmuch as different microbe-derived autoimmunogenic peptides may be differentially presented by distinct alleles, HLA associations may become more evident if sorted by the nature of peptide binding motifs. 37,38 In analyzing the role of class II alleles/haplotypes in various diseases, it is important to consider that protective associations are equally, if not more, relevant than predisposing associations. For example, DQB1*0302 and DQB1*0201, which pose a strong risk in insulin-dependent diabetes, lack the protective Asp57 found in DQB1*0301, DQB1*0303, and DQB1*0401 alleles and especially in DQB1*0602, which confers protection even among relatives of insulindependent diabetes. 39 HLA alleles regulate immune responses to infections, 35 bind and present autoantigens with different affinities, play a role in T-cell repertoire selection, 40 and may themselves be target autoantigens. 39,41,42 Differential presentation of autoimmune peptides by protective and nonprotective or susceptibility alleles can have major effects on the development of pathogenic autoimmunity. Future structure function studies may reveal mechanisms by which certain alleles (eg, DQB1*0603) and the DRB1*13- DQA1*0103-DQB1*0603 haplotype may confer protection in RHD. In conclusion, our data support the hypothesis that apparent discrepancies among some studies investigating HLA class II associations with RHD might have been due, in part, to an inappropriate grouping of RF with and without carditis and/or heterogeneous clinical subgroups of RHD patients. It should be noted that the results from our study showing that the DRB1*0701-, DR6-, and DQB1*0201-related haplotypes confer susceptibility to MVD are in agreement with those reported for populations of Turkish, 16 Mexican, 13 South African, 23 and Japanese 14 RHD patients, in whom the MVD category constituted the majority of cases ( 50%). The data presented here suggest that the identification of class II allele haplotypes may provide an insight into the molecular mechanism of the disease and may be a useful tool in predicting the clinical outcome in RF patients, thereby affording new means of intervention or vaccine design.

7 2790 HLA Association in RHD Acknowledgments This work was supported by grants from the US Veterans Administration (Dr Kotb) and the National Institutes of Health (NIH AI to Dr Kotb and GM to Dr Thomson). Dr El-Demellawy was supported, in part, by the W.S. Fulbright Commission. References 1. Markowitz M, Gerber MA. The Jones criteria for guidance in the diagnosis of rheumatic fever: another perspective. Arch Pediatr Adolesc Med. 1995;149: Stollerman GH. Rheumatic fever. Lancet. 1997;349: Shulman ST. Complications of streptococcal pharyngitis. Pediatr Infect Dis J. 1994;13(suppl 1):S70 S Kaplan EL. Global assessment of rheumatic fever and rheumatic heart disease at the close of the century: influences and dynamics of populations and pathogens: a failure to realize prevention? Circulation. 1993; 88: Bisno AL. Group A streptococcal infections and acute rheumatic fever. N Engl J Med. 1991;325: Zabriskie JB, Gibofsky A. Genetic control of the susceptibility to infection with pathogenic bacteria. Curr Top Microbiol Immunol. 1986; 124: Ayoub EM. Acute rheumatic fever. In: Emmanouilides GC, Riemenschneider TA, Allen HD, Gutgesell HP, eds. Heart Disease in Infants, Children, and Adolescents. 5th ed. Baltimore, Md: Williams & Wilkins; 1995: Veasy LG, Hill HR. Immunologic and clinical correlations in rheumatic fever and rheumatic heart disease. Pediatr Infect Dis J. 1997;16: Sanyal SK. Long-term sequelae of the first attack of acute rheumatic fever during childhood. In: Emmanouilides GC, Riemenschneider TA, Allen HD, Gutgesell HP, eds. Heart Disease in Infants, Children and Adolescents. 5th ed. Baltimore, Md: Williams & Wilkins; 1995: DiSciascio G, Taranta A. Rheumatic fever in children. Am Heart J. 1980;99: Pickles WN. Rheumatic family. Lancet. 1943;2: Monplaisir N, Valette I, Bach J-F. HLA antigens in 88 cases of rheumatic fever observed in Martinique. Tissue Antigens. 1986;28: Debaz H, Olivo A, Perez-Luque E, Vasquez-Garcia MN, Burguete A, Chavez-Negrete A, Velasco C, Arguero R, Gorodeszky C. DNA Analysis of class II alleles in rheumatic heart disease in Mexicans. 22nd Annual ASHI Meeting Abstracts. Human Immunology. 1996;49(suppl): Koyanagi T, Koga Y, Nishi H, Toshima H, Sasazuki T, Imaizumi T, Kimura A. DNA typing of HLA class II genes in Japanese patients with rheumatic heart disease. J Mol Cell Cardiol. 1996;28: Weidebach W, Goldberg AC, Chiarella JM, Guilherme L, Snitcowsky R, Pileggi F, Khalil J. HLA class II antigens in rheumatic fever: analysis of the DR locus by restriction fragment-length polymorphism and oligotyping. Hum Immunol. 1994;40: Ozkan M, Carin M, Sonmez G, Senocak M, Ozdemir M, Yakut C. HLA antigens in Turkish race with rheumatic heart disease. Circulation. 1993; 87: Taneja V, Mehra NK, Reddy S, Narula J, Tandon R, Vaidya MC, Bhatia ML. HLA-DR/DQ antigens and reactivity to B cell alloantigen D8/17 in Indian patients with rheumatic heart disease. Circulation. 1989;80: Guilherme L, Weidebach W, Kiss MH, Snitcowsky R, Khalil J. Association of human leukocyte class II antigens with rheumatic heart disease in Brazilian population. Circulation. 1991;83: Rajapakse CNA, Halim K, Al-Orainey I, Al-Nozha M, Al-Aska AK. A genetic marker for rheumatic heart disease. Br Heart J. 1987;58: Ayoub EM, Barret DJ, Maclaren NK, Krischer JP. Association of class II human histocompatibility leukocyte antigens with rheumatic fever. J Clin Invest. 1986;77: Jhinghan B, Mehra NK, Reddy KS, Taneja V, Vaidya MC, Bhatia ML. HLA, blood groups and secretor status in patients with established rheumatic heart disease. Tissue Antigens. 1986;27: Anastasiou-Nana MI, Anderson JL, Carquist JF, Nanas JN. HLA-DR typing and lymphocyte subset evaluation in rheumatic heart disease: a search for immune response factors. Am Heart J. 1986;112: Maharaj B, Hammond MG, Appadoo B, Leary WP, Phil D, Pudifin DJ. HLA-A, B, DR, and DQ antigens in black patients with severe chronic rheumatic heart disease. Circulation. 1987;76: Dajani AS, Bisno AL, Chung KJ, Durack DT, Gerber MA, Kaplan EL, Millard HD, Randolph MF, Shulman ST, Watanakunakorn C. Guidelines for the diagnosis of rheumatic fever. Jones Criteria, 1992 update. Special Writing Group of the Committee on Rheumatic Fever, Endocarditis, and Kawasaki Disease of the Council on Cardiovascular Disease in the Young of the American Heart Association. JAMA. 1992;268: Kotby AA, El-Monim MTA, Hassan AS. Assessment of rheumatic fever and rheumatic heart disease in the Children s Hospital of Ain Shams University from Egypt J Pediatr. 1996;13: Walsh PS, Metzger DA, Higuchi R. Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. Biotechniques. 1991;10: Olerup O, Zetterquist H. HLA-DR typing by PCR amplification with sequence-specific primers (PCR-SSP) in 2 hours: an alternative to serological DR typing in clinical practice including donor-recipient matching in cadaveric transplantation. Tissue Antigens. 1992;39: Olerup O, Zetterquist H. DR low-resolution PCR-SSP typing: a correction and an up-date. Tissue Antigens. 1993;41: Bignon JD, Fernandez-Vina MA, Cheneau ML, Faucet R, Schreuder GMT, Clayton J, Marsh SGE, Charron D. HLA DNA class II typing by PCR-SSOP: 12th International Histocompatibility Workshop experience. In: Charron D, ed. HLA: Genetic Diversity of HLA: Functional and Medical Implications. Paris, France: EDK; 1997: Buyse I, Decorte R, Baens M, Cuppens H, Semana G, Emonds MP, Marynen P, Cassiman JJ. Rapid DNA typing of class II HLA antigens using the polymerase chain reaction and reverse dot blot hybridization. Tissue Antigens. 1993;41: Hors J, El Chenawi S, Djoulah S, Hafez M, Abbas F, El Borai MH, Kamel M, Abbal M, Cambon-Thomsen A, Mercier P, Reviron D, Magzoub MA, Rosner G, Delgado JC, Yunis E, Raffoux C, Tamouza R, Izaabel H, Hmida S, Benhamamouch S, Bessaoud K, Langaney A, Sanchez-Mazas A. HLA in North African populations: 12th International Histocompatibility Workshop NAFR report. In: Charron D, ed. HLA: Genetic Diversity of HLA: Functional and Medical Implications. Paris, France: EDK; 1997: Fernandez-Viña MA, Gao X, Moraes ME, Moraes JR, Salatiel I, Miller S, Tsai J, Sun Y, An J, Layrisse Z, Gazit E, Brautbar C, Stastny P. Alleles at four HLA class II loci determined by oligonucleotide hybridization and their associations in five ethnic groups. Immunogenetics. 1991;34: Woolf B. On estimating the relation between blood groups and disease. Ann Hum Genet. 1955;19: Payami H, Khan MH, Grennan DM, Sanders PA, Dyer PA, Thomson G. Analysis of genetic interrelationship among HLA-associated diseases. Am J Hum Genet. 1987;41: Hill AVS. HLA and infection. J R Coll Physicians Lond. 1992;26: Ayoub EM, Kotb M, Cunningham MW. Rheumatic fever pathogenesis. In: Stevens DL, Kaplan EL, eds. Streptococcal Infections. New York, NY: Oxford University Press: In press. 37. Thorsby E. HLA associated diseases. A summary of the 12th International Histocompatibility workshop component. In: Charron D, ed. HLA: Genetic Diversity of HLA: Functional and Medical Implications. Paris, France: EDK; 1996: Kotb M. Infection and autoimmunity: a story of the host, the pathogen, and the copathogen. Clin Immunol Immunopathol. 1995;74: Nepom GT. Class II antigens and disease susceptibility. Annu Rev Med. 1995;46: Schmidt D, Verdager J, Nuzhat A, Santamaria P. A mechanism for the major histocompatibility complex-linked resistance to autoimmunity. J Exp Med. 1997;186: Sinha AA, Lopez MT, McDevit HO. Autoimmune diseases: the failure of self tolerance. Science. 1990;248: McDevitt HO, Chinitz A. Genetic control of the antibody response: relationship between immune responses and histocompatibility (H-2) type. Science. 1969;163: