A majority of Brazilian patients with rheumatoid arthritis HLA-DRB1 alleles carry both the HLA-DRB1 shared epitope and anti-citrunillated peptide antibodies.

The objective of the present study was to evaluate the contribution of the shared epitope (SE), the rheumatoid arthritis (RA) protection model, and the occurrence of anti-cyclic citrullinated peptide (anti-CCP) antibodies in RA patients from a genetically diverse population. One hundred and forty Brazilian RA patients and 161 matched controls were typed for HLA-DRB1 alleles using amplified DNA hybridized with sequence-specific oligonucleotide probes or primers. Patients were stratified according to the presence or absence of SE (DRB1*0401, *0404, *0405, *0101, *1001, and *1402), of the DERAA alleles (DRB1*0103, *0402, *1102, *1103, *1301, *1302, and *1304), and X (all other alleles). Anti-CCP antibodies were measured by ELISA. The combined frequency of SE-positive alleles was significantly greater (76.4 vs 23.6%, P < 0.0001) than the controls. The SE/SE and SE/X genotypes were over-represented (P < 0.0001, OR = 6.02) and DERAA/X was under-represented in RA patients (P < 0.001, OR = 0.49), whereas the frequencies of the SE/DERAA, X/X and X/DERAA genotypes were not significantly different from controls. The frequency of anti-CCP antibodies was higher in SE-positive patients than in SE-negative patients (64.6 vs 44.7%, P = 0.03; OR = 2.25). Although the Brazilian population is highly miscegenated, the results of this study support the findings observed in most genetically homogeneous populations with RA; however, they are not mutually exclusive but rather complementary. The participation of DRB1-DERAA alleles in protection against RA was also observed (OR = 0.4; 95%CI = 0.23-0.68).

Autoantibodies to cyclic citrullinated peptides (anti-CCP antibodies) are highly specific for RA, can be detected years before the first clinical manifestation of RA (17), and are reported to be a good predictor of the development of RA (18).Because the contribution of the SEcontaining HLA alleles to the pathogenesis of RA is not well understood, the novel information on the association of SE alleles with anti-CCP-positive disease (17) led us to evaluate the hypothesis that the SE alleles are a risk factor for anti-CCP antibodies.
Since the Brazilian population is genetically diverse, presenting a distinctive pattern of DR alleles and haplotypes, the evaluation of RA susceptibility models in this highly miscegenated population can contribute to the understanding of such associations.In the present study, we evaluated the contribution of DRB1 alleles stratified according to the clustering of SE + and SE -alleles, the RAP model alleles and the presence of anti-CCP antibodies among Brazilian patients with RA.

Patients and controls
A total of 140 (109 women) patients aged 24-79 years (median = 51), including 98 Whites, self-described to be of Western and Southern European ancestry, 24 Mulattoes (Caucasian and Black admixtures), and 18 Blacks, historically of mostly Bantu, Benin and Senegal African ancestry (19) presenting RA were evaluated at the University Hospital of the Faculty of Medicine of Ribeirão Preto, Ribeirão Preto, SP, Brazil, and at the University Hospital of the Faculty of Medical Sciences of Campinas, Campinas, SP, Brazil.All patients were typed for HLA-DRB1 alleles and fulfilled the 1987 revised criteria of the American College of Rheumatology for the diagnosis of RA (20).As controls, we evaluated 161 randomly selected normal blood donors of similar ethnic background from the same geographical region (21), including 86 Whites, self-described to be of Western and Southern European ancestry, 47 Blacks (Black ancestry up to the third generation), historically of mostly Bantu, Benin and Senegal African ancestry, as reviewed by Figueiredo et al. (19), and 28 Mulatto (Caucasian and Black admixtures) individuals.The population assignments were made by the same investigator who asked the donors about their ancestry.All individuals were submitted to a medical examination that showed no evidence of previous underlying diseases, and presented negative serology for Chagas' disease, B-and C-hepatitis, and HIV infection.The study protocol was approved by the Ethics Committee of the Faculty of Medicine of Ribeirão Preto, University of São Paulo, and all subjects gave written informed consent to participate in the study.

Autoantibody detection
The detection of anti-CCP2 IgG antibodies was performed using commercially available ELISA kits containing synthetic peptides (Quanta Lite anti-CCP2 Inova, San Diego, USA).ELISA was performed according to manufacturer instructions.Serum samples presenting results >20 U/mL were considered to be positive for the anti-CCP antibody.Rheumatoid factor was evaluated in all patients by nephelometry.

Allele clustering according to the SE and the RAP models
The frequency of DRB1 alleles considered to pertain to  (12,13).For clarity, this study uses the term DERAA-encoding alleles but does not differentiate between the direct effect of these alleles and the effect of other alleles in linkage with the DERAA-encoding alleles.For analysis, 6 groups were formed according to the presence of DRB1 alleles, as follows: group A, homozygosity for the SE (SE/SE); group B, 1 SE allele and 1 no SE and no DERAA (SE/X); group C, 1 SE and 1 DERAA allele (SE/DERAA); group D, no SE and no DERAA (X/X); group E, 1 DERAA allele (X/DERAA), and group F, homozygosity for DERAA alleles (DERAA/DERAA) (see Table 2).

Statistical analysis
HLA frequencies observed in patients and controls were compared using the two-tailed Fisher exact test with the approximation of Woolf.Differences were considered to be significant at P < 0.05.The odds ratio (OR) was used to estimate the strength of the associations.The detection of the strongest association between alleles was performed by the method of Svejgaard and Ryder (25).The variables (gender, race, age) observed in patients and controls were compared using unpaired t-test.Differences were considered to be significant at P < 0.05.

Frequencies of DRB1 and DRB1 alleles clustered as SE and DERAA alleles
The individual comparison of the frequency of the HLA alleles associated with RA, i.e., the HLA-DRB1*0401, *0404 (*0408), *0405, *0101, and *1001 alleles, observed in this series was significantly different from the respective frequency observed in controls (Table 1).Nevertheless, when the frequency of the SE alleles was considered, it was observed in 39.6% of RA patients and in only 12.5% of the controls Table 1.Table 1.Table 1.Table 1 (Pc < 0.0001; Table 1), with an OR of 6.02.The frequency of DERAA alleles in RA patients was decreased when compared with controls (11.42 vs 20.8%, P = 0.003, OR = 0.49), conferring protection against the development of RA (Table 1). www.bjournal.com.br

DERAA-positive alleles (RAP model), and the influence of the SE
To study the effect of the presence of the DERAA motif on susceptibility to RA, patients and controls were divided into 6 groups according to their HLA-DRB1 genotypes (Table 2).The overall effect of DERAA on the development of RA was assessed by comparing the presence of the DERAA motif (groups C, E, and F) versus its absence (groups A, B, and D).The presence of DERAA was underrepresented in the RA patients (19.3%) versus controls (37.2%),conferring an OR of 0.40 (P = 0.0006; 95%CI = 0.23-0.68).The effect of the DERAA on the absence of the SE alleles was assessed by comparing group E (X/DERAA) plus group F (DERAA/DERAA) with group D (X/X).DERAA subjects had a reduced risk of developing RA (OR = 0.7), although the observed effect was not statistically significant (P = 0.6; 95%CI = 0.39-1.59).

Association between SE individuals and presence of anti-CCP antibodies
Anti-CCP antibody was present in 58% of the patients and rheumatoid factor was present in 60.7%.All patients with positive anti-CCP antibody also presented positive rheumatoid factor.Ninety-three patients carried at least one dose of shared epitope (Table 3).The frequency of anti-CCP antibodies was higher in SE-positive patients (64.5%) when compared with SE-negative patients (44.7%;P = 0.03, OR = 2.63; see Table 4).The distribution of anti-CCP antibodies was analyzed according to specific SErelated alleles; we did not observe a higher frequency of anti-CCP antibodies in RA patients carrying DRB1*0401, *0404, and *0101 compared with patients who do not carry these alleles.Table 3. Table 3. Table 3. Table 3. Table 3 Data are reported as number of individuals with percent in parentheses unless otherwise stated.RA = rheumatoid arthritis; Anti-CCP = anti-cyclic citrullinated peptide antibodies; RF = rheumatoid factor; SE = shared epitope; NA = not applicable; ND = not determined.The variables (gender, race, age) observed in patients and controls were compared using unpaired t-test.Differences were considered to be significant at P < 0.05.SE-positive observed in patients and controls were compared using the twotailed Fisher exact test with the approximation of Woolf.Differences were considered to be significant at P < 0.05 (*P < 0.0001).
Table 4. Table 4. Table 4. Table 4. Data are reported as number of individuals with percent in parentheses unless otherwise stated.Presence or not of anti-CCP antibodies in SE-positive or SE-negative RA patients was compared using the two-tailed Fisher exact test with the approximation of Woolf.Differences were considered to be significant at P < 0.05.

Discussion
The association between RA and HLA-DR4 antigens was first reported in the 1970s using cellular techniques (1), and has recently been reexamined at the genomic DNA level (2,3).The susceptibility to RA has been reported to vary according to the ethnic background of the patients, supporting the "shared epitope" hypothesis (10,11).The extraordinary genetic diversity of HLA alleles and haplotypes seen in highly miscegenated populations may result from the aggregation of distinctive alleles from separate racial and ethnic groups.The gene bank of the modern Brazilian population represents the contribution of individuals of varied racial and ethnic groups.Historic immigrations from Europe have distributed various Caucasian populations, mainly Portuguese, Spanish, Italian and German, across Brazil.Brazilian Amerindian populations still exist as semi-isolated tribal groups, but their genes are also represented in modern urban populations.In a previous study, we observed that the frequencies of DRB1*15/ 16 (*02), *03, *12, *13, and *09 were similar among Whites, Blacks and Mulattoes.In contrast, the frequencies of DRB1*01, *04, *07, *08, *11, and *14 varied between the different groups.In particular, DRB1*01 genes were detected at approximately the same frequency in Whites and Mulattoes, but were greater in Blacks.DRB1*04 was similar in Whites and Mulattoes, but less in Blacks.Regarding the frequency of individual alleles, HLA-DRB1*0101 and *1001 were significantly greater in Blacks compared with Whites (Pc = 0.01 and Pc = 0.01, respectively).The polymorphism of DRB1*04 alleles seen in Brazilian Whites was greater than that observed in the Black population (21).Since the DRB1*04, *01 and *10 allelic frequencies are different for White and Black individuals, Pina et al. performed a study on RA patients using exclusively Afro-Brazilian descendants (Pina FP, Conde RC, Louzada-Júnior P, Donadi EA, Bertolo MB, unpublished data).They concluded that Afro-Brazilians with RA inherited a few HLA genes, which are characteristic of RA not only in African people, but also in Europeans and Asians who originated the Brazilian population.In Chile, where HLA-DR4 encoding DRB1 alleles do not encode susceptibility to RA (26), HLA-DR9 was associated with RA (27).In Japan, DRB1*0901 was associated with RA after exclusion of alleles *0101 and *0405 (28).In Kuwait, the association of HLA-DR3 with RA is explained by a high frequency of HLA-DR3 together with a relatively low frequency of HLA-DR4 alleles (29).In the present study, we did not observe an association with either DRB1-09 or DRB1-03 alleles.In fact, all DRB1 SE alleles were observed in Brazilian RA patients, including the rare DRB1*1402 allele that is com-monly observed in individuals of Amerindian heritage, reflecting the high diversity of the population of the present study (21).Thus, in highly miscegenated populations the evaluation of SE frequency instead of individual alleles pertaining to the SE may be more important in terms of immunogenetic susceptibility to RA.
The events involved in the pathogenesis of RA are still unclear, but certainly the etiology is multifactorial.The shared epitope of HLA-DRB1 is the most important genetic risk factor.Environmental risk factors are less understood.Smoking is a candidate, associated with increased citrullinated cyclic peptide antibodies (30)(31)(32).In this study, we investigated the associations between HLA class II alleles and RA and the protective effects of the DERAA-encoding HLA-DRB1 alleles on RA susceptibility.The question of whether the effect of DERAA is truly protective or is merely the result of the absence of predisposing SE encoding HLA-DRB1 alleles is the subject of some controversy.In the present study, the comparison of subgroups (Table 2) allowed differentiation of the effects of protection and nonprotection.In addition, in our population, we showed that DERAA-encoding HLA-DRB1 alleles independently reduce the risk of developing RA.These findings agree with similar results observed in more homogeneous groups such as Dutch (13) and Swiss (33) populations.
It has been demonstrated that peptides carrying the DERAA motif are naturally processed by human antigenpresenting cells, and it has been suggested that the protective effect of DERAA is mediated by a specific protective T cell response (34).Although our results show that the presence of a predisposing haplotype is not required to observe the protective effect associated with DERAA, it is possible that the DERAA sequence, particularly the DRB1*13 alleles, not only protects against RA but is also associated with a milder outcome in other diseases, such as reduced progression to active chronic hepatitis C and B (35).Theses findings are intriguing and indicate the importance of elucidating the biologic pathways underlying these associations, because they might unveil new insight into immune regulation in relation to the HLA system.
What does this tell us about the disease mechanism?Assuming differential risks for no susceptibility alleles, this hypothesis is able to explain data in the literature by referring to mutations in amino acid sequence at positions 67-74 in the HLA-DRB1 molecule.Thus, this version supports the notion that this sequence is central in conferring susceptibility to RA and does not indicate a role for another gene within the HLA complex.The fact that local amino acid substitutions induce differential risks suggests that the interaction of this region with presented peptides, superantigens or invariant chain may be important in the www.bjournal.com.brpathogenesis of this disease (13,14).
We observed a higher frequency of anti-CCP antibodies in SE-positive patients (64.5%) compared with SEnegative patients (44.7%).In contrast, we did not observe a higher frequency of anti-CCP antibodies in RA patients that carried specific SE-related alleles, such as DRB1*0401, *0404, and *0101.Huizinga et al. (17) reported that the SE alleles were only associated with anti-CCP-positive RA and not with anti-CCP-negative disease, indicating that the SE alleles are not associated with RA as such, but rather with a distinct phenotype of the disease.Although no formal conclusions on causality can be drawn from this association study, these findings suggest that anti-CCP antibodies mediate the association between SE alleles and RA (36).It would be of interest to repeat the present study by following the development of anti-CCP antibodies and RA in healthy asymptomatic persons with and without SE alleles.Nevertheless, the present findings constitute an important refinement of the long-known association between HLA and RA by indicating that the SE alleles are not primarily associated with RA, but rather with anti-CCP antibody positivity.
Although SE represents the major contribution to RA susceptibility, other genes acting individually or in concert with SE may also be involved (37,38).In addition, environmental factors may further induce epigenetic alterations contributing to RA pathogenesis (39,40).
We have presented several lines of evidence indicating the involvement of DRB1 alleles in the susceptibility to RA.The RAP, SE and P4 polymorphism models are all associated with the HV3 region of the DRB1 gene and are useful to understand the immunogenetic susceptibility to RA.The findings of the present study support current RA models, and also support the idea that these models are not mutually exclusive but rather complementary as proposed by Zanelli et al. (16).Even though the Brazilian population is highly miscegenated, overall the present findings agree with others observed in genetically more homogeneous populations.

Table 1 .
. DRB1 and shared epitope (SE) allele frequencies among patients with rheumatoid arthritis (RA) and controls.