HLA-B27 and Ankylosing Spondylitis: functional aspects of HLA-B27 molecules and other factors involved in the disease pathogenesis

Ankylosing Spondylitis (AS) is a chronic rheumatic disease affecting primarily the spine and the sacroiliac joints whose pathogenic mechanism is still poorly understood (Tsui FW et al., 2014). AS has been found strongly associated with HLA-B27 and, recently, with ERAP1, an aminopeptidase relevant in determining the HLA-B27 peptide repertoire (Reveille JD, Maganti RM, 2009; Evans DM et al., 2011). Not all HLA-B27 alleles predispose to AS. HLA-B*2709 differing from the worldwide and more frequent AS-associated HLA-B*2705 by only the His116Asp substitution, is not associated with AS (Paladini F et al., 2005; Chen H et al., 2012). The functional/structural comparison of these two alleles offers a unique opportunity to gain insight into the role of the HLA-B27 molecules in spondyloarthritis. In addition to be associated with AS, HLA-B27 has been found to protect very efficiently against a number of common viral infections. This could be explained by the remarkable performance of the HLA-B27-restricted and virus-specific CD8+ T cells (Neumann-Haefelin C et al., 2006; Sorrentino R et al., 2014). Our hypothesis, based on preliminary data, is that the HLA-B*2705 molecules can present a broader spectrum of peptides than B*2709, among which peptides possessing non-canonical binding motifs and that this is the basis of its dualistic role in both virus protection and autoimmunity. A recent study carried out in my laboratory reported that HLA-B*2705 molecules could exploit an unconventional route of antigen presentation that is proteasome- and TAP-independent. To this purpose, chimeric proteins carrying viral epitopes were used, able to cross the cell membranes by virtue of HIV-TAT transducing domain (Bettosini F et al., 2005; Magnacca A et al., 2012). The initial aim of my PhD thesis was to investigate whether other HLA locus B molecules, such as the HLA-B*0702, could exploit this uncanonical pathway of antigen processing employed by the HLA-B27. Therefore, we used the chimeric protein system as carrier for the viral immunodominant epitope, pEBNA3A (RPPIFIRRL), derived from EBV and restricted for the HLA-B*0702. To generate pEBNA3A-specific CD8+ T cells, we stimulated the PBMCs derived from a B*0702 positive subject who was at the same time B*2705 positive and AS affected. Surprisingly, the CD8+ T cells reactivated by stimulation with the HLA-B*0702-restricted EBNA3A peptide were recognizing this peptide both in the proper restriction context, that is the B*0702, as well as in that of B*2705 but barely in that of B*2709 allele. The same results were achieved with other two B*0702/B*2705 positive AS patients. Subsequently, extending our analysis to individuals who were B*2705 or B*2709 carriers but B*0702 negative, we found that the B*2705, but not the B*2709, could accommodate unexpected peptides on its binding groove only in B*2705 positive subjects. In light of these data, the second aim of my PhD project was to elucidate the meaning and the relevance of such uncanonical presentation mediated by the B*2705 molecule, but not by the B*2709, using a combination of functional studies, TCR characterization and theoretical approaches (molecular modeling and dynamic simulations). In conclusion, this study shows an unexpected flexibility of B27 molecules, particularly of the B*2705, in the presentation of peptides considered unfit for its binding groove, giving another clue for its implication in autoimmunity.