3A). In conclusion, the humoral anti-peptide response of RA patients appeared more complex and less specific than the cellular anti-peptide response. In the present study, we found that a proportion of RA patients (21%) developed autoimmune T-cell responses specific for a major determinant contained in the sequence 117–133, which

is located in the second RNA-binding domain of hnRNP-A2. This proportion appears low at first sight but it should be considered that these patients had established disease and were treated with various immunomodulatory agents. Although some patients (11%) with osteoarthritis also reacted to the major T-cell epitope LBH589 nmr 120–133, it should be noted that these patients did not receive immunosuppressive medication. Therefore, the proportion of positive RA patients may be underestimated. The difficulty of identifying autoimmune T-cell epitopes is highlighted by a study on celiac disease, in which the patients had to be challenged Nutlin-3a order with gliadin to detect the dominant epitope 17.

The two peptides 117–133 and 120–133 preferentially recognized by PBMC from RA patients both contain the 9-mer core sequence 123–131 binding to various RA-associated HLA molecules, as determined by TEPITOPE analysis (Table 1 and Fig. 4). Nevertheless, only two patients reacted to both peptides (Table 2 and Supporting Information Table 2). This result may be linked to a differential presentation by various HLA molecules and recognition by various T-cell repertoires. Indeed, DR10 may present, and/or the selected T cell may recognize, 117–133 but not 120–133, and it would be the opposite for DR7,

whereas DR1-restricted T cells would recognize both peptides (Table 2). Since the sequence 117/120–133 binds to various RA-associated HLA alleles, it might be linked to pathogenicity in different ethnic populations. Indeed, DR*0101 and *0401 are present in Caucasians 1, whereas DR*1001 is often found in populations originating from the Mediterranean area, such as Spain, Greece, and Saudi Arabia 18–20. Moreover, HAS1 peptide 117/120–133 binds well to DR*0405 (Fig. 4), the major HLA-allele associated with severe and erosive RA in Japan 14. Although patients with SLE may also be reactive to the hnRNP-A2 antigen 21, it is unlikely that they recognize the RA dominant epitope 117/120–133, since susceptibility to lupus is associated to HLA-DR2 (DR15) and DR3 22 and peptides 117/120–133 are predicted extremely bad binders to these alleles (Fig. 4). The TEPITOPE program is best designed to accurately identify promiscuous epitopes, i.e. epitopes binding to many HLA class II molecules.