In complex trait genetics in general and, notably the autoimmune diseases, there is a fast approaching challenge of how to move from gene-wide association and fine-mapping studies to functional pathways and therapeutics. MS is a case in point: genetic studies identify HLA-DRB1*1501 as a primary risk factor in MS, although the close genetic proximity of HLA-DQB1*0602 and strong linkage disequilibrium between the two make it almost impossible to distinguish the contribution of these genes except through functional studies. Most functional studies relating to HLA products in MS have focused on DRB1*1501, with little consideration for the potential contribution of DQB1*0602-associated autoimmunity to pathogenesis. Our recent studies in HLA-DR15-Tg mice showing that susceptibility to MOBP-induced EAE is determined by HLA-DQB1*0602, and not by DRB1*1501 , were the first to implicate DQB1*0602-associated autoimmunity in the pathogenesis of MS and to suggest DQ6 as an important disease predisposing, rather than just a disease-modifying allele, as previously suggested . We now show that the DQB1*0602-associated pathogenic autoimmunity against MOBP is not a limited case, and that DQB1*0602-autoimmunity against other CNS antigens may also play a role in pathogenesis of MS. We show that disease-susceptibility to PLP, a highly encephalitogenic protein and one of the most prominent and 'MS-implicated' antigens in the CNS, is also determined by DQB1*0602, and not by DRB1*1501 gene products of the HLA-DR15 haplotype. Thus, HLA-DRB1*1501 transgenics were found to be refractory to PLP disease induction, whereas the HLA-DQB1*0602 transgenics were susceptible, via T-cells reactive against PLP139-151 and PLP175-194 encephalitogenic epitopes. These findings have important bearing on the candidacy of the DQB1*0602 allele as genetic risk factor for MS.
Analysis of the T-cell autoimmunity against PLP, which showed that PLP is immunogenic for HLA-DRB1*1501- and HLA-DQB1*0602-Tg mice also revealed that the T-cell autoimmunity against PLP in both transgenics is predominantly directed against PLP139-151 and PLP175-194 epitopes, which are the immunodominant and encephalitogenic epitopes for SJL/J mice . However, although both transgenics could mount CD4+ T-cells reactive against these co-dominant PLP epitopes, the PLP139-151 and PLP175-194-reactive T-cells that developed in DQB1*0602-Tg mice were of pathogenic Th1/Th17-type cells that caused the development of clinical EAE in DQB1*0602-Tg mice, while the PLP139-151 and PLP175-194-reactive T-cells were of a more Th2-type that conferred resistance to PLP-EAE in DRB1*1501-Tg mice. Further investigation is required to understand how recognition of the same epitopes in the context of DQB1*0602 class-II molecule drives Th1/Th17 pathogenic T-cell autoimmunity, while the recognition of same epitopes in the context of DRB1*1501 class-II molecule drives more Th2 type autoimmunity. These findings in the reductionist transgenic models are quite challenging since both DRB1*1501 and DQB1*0602 molecules are co-expressed in HLA-DR15+ MS patients, raising the question on whether the net T-cell autoimmunity against PLP, and particularly against the PLP139-151 and PLP175-194 epitopes, in HLA-DR15+ MS would be of Th1/Th17 pathogenic autoimmunity or more of anti-inflammatory Th2 type. This question can be resolved by analyzing the Th1/Th17/Th2-type of T-cell reactivity against PLP, particularly against PLP139-151 and PLP175-194 epitopes by PBLs of HLA-DR15+ MS patients, compared to the Th1/Th17/Th2-type response by their DRB1*1501- and DQB1*0602-restricted T-cells reactive against these epitopes. In this context, however, it is worth mentioning that (DRB1*1501 × DQB1*0602)F1 double Tg mice, expressing both DRB1*1501 and DQB1*0602 molecules were recently found to be resistant to EAE induction by either PLP139-151 or PLP175-194 epitope (Kaushansky et al., unpublished data). Why and by what mechanisms the susceptibility to EAE induced by these immunodominant epitopes was not inherited in the F1 double-Tg mice as a dominant trait, as is usually the case for EAE susceptibility to other encephalitogenic antigens/epitopes in laboratory animals, is now under investigation in our laboratory.
The central question of whether HLA-DRB1*1501, HLA-DQB1*0602, or their co-expression is the primary risk factor in MS has not been fully resolved by genetic studies due to their linkage disequilibrium. Although most MS genetics studies in Caucasian populations suggest HLA-DRB1*1501 as the genetic risk factor in MS [13–15], several other genetic studies in unique populations suggest that HLA-DQ alleles may also be a risk factor. In a small cohort of Norwegian MS patients, some patients were identified who carried DQB1*0602 or DQB1*0603 without DRB1*1501, but none who were DRB1*1501 without DQB1*0602 . Related observations were made in a relatively small sample of Hong Kong Chinese patients with MS. In that population, DR15 is expressed without DQB1*0602 and DQB1*0602 without DR15, so that one can ask whether either or both of the dissociated genes appear to confer an enhanced risk. It was found that the enhanced risk was associated with DQB1*0602 . A similar case has been made for susceptibility in Afro-Brazilians, where the frequency of DQB1*0602 among patients is higher than that for the main DR15 allele, DRB1*1503, in that unique population . A caveat here is the analysis of HLA-DR and -DQ associations conducted in a large cohort of African American MS patients, which showed a selective association with HLA-DRB1*1501 and not with HLA-DQB1*0602 was identified . Thus, while primacy of HLA-DRB1*15 or HLA-DQB1*0602 has not been conclusive in the African populations that show greater haplotypic diversity than Europeans and distinct patterns of linkage disequilibrium, a potential contribution of HLA-DQB1*0602 to MS susceptibility could still be inferred. The possibility that both HLA-DRB1*1501 and HLA-DQB1*0602 loci influence susceptibility to MS through epistatic interactions has been demonstrated in Canadian MS cohorts, where the HLA-DQA1*0102, which showed no independent association, was found to interact strongly with HLA-DRB1*1501 in trans, increasing MS risk in the presence of HLA-DRB1*1501 and playing a protective role in its absence . Nonetheless, our data showing pathogenic DQB1*0602-associated autoimmunity against MOBP in HLA-Tg mice  and against PLP (this study) offer a rationale and potential mechanisms for the HLA-DQB1*0602 association with MS.
The reductionist experiment of transgenic lines, separating HLA-DR and HLA-DQ, offers a chance to dissociate these presentation experiments in a genetic setup rarely seen in humans because of the rarity of recombination events between these loci. Studies with HLA class-II Tg mice have previously demonstrated HLA-DR-dependent disease induced by MBP, PLP, and MOG [24–27]. However, while susceptibility to MBP- and MOG-induced EAE in HLA-Tg mice was determined by the DRB1*1501 allele of the HLA-DR15 haplotype, the previously reported susceptibility to PLP91-110 epitope was associated with the DRB1*0301, a non-Caucasian MS-associated allelic gene of the HLA-DR3 haplotype. Our results show that, unlike the previously reported HLA-DR-dependent susceptibility to MBP-, PLP-, or MOG-induced EAE, the pathogenic autoimmunity against PLP, as well as against MOBP , was dependent on HLA-DQB1*0602 rather than HLA-DRB1*1501. This DQB1*0602-associated susceptibility is in striking contrast with human and transgenic mouse studies suggesting a protective role for HLA-DQ6. The closely related allele DQB1*0601, most typically found in South Asian populations, has been reported in two human studies to be associated with protection from MS [31, 46]. In accordance with those findings, studies in HLA-Tg-mice [29, 30] argued that the presence of HLA-DQB1*0601 could exert an epistatic protective effect on HLA-DR-dependent, anti-myelin autoimmunity. Thus, while HLA-DRB1*1502 Tg mice were susceptible to MOG-induced EAE, the HLA(DRB1*1502 × DQB1*0601) double-Tg mice were resistant . The protective effect of HLA-DQ6(DQA1*0103/DQB1*0601) was reported in more detail in PLP-induced EAE, in which HLA-DR3(DRB1*0301)-Tg mice were susceptible to PLP91-110-induced disease, whereas the HLA-DR3(DRB1*0301) × DQ6(DQB1*0601) double-Tg mice were protected . In our studies, using the Caucasian MS-associated allelic genes (of the HLA-DR15 haplotype) we show the opposite scenario for the role of the HLA-DQ6 in anti-myelin autoimmunity in HLA-Tg mice, as the HLA-DQB1*0602 gene product determined susceptibility to PLP- as well as to MOBP-induced EAE.
The PLP-induced EAE in HLA-DQB1*0602 transgenic mice showed a typical caudo-rostral clinical progression that was associated with CNS demyelination, axonal damage and with optic neuritis. However, unlike usually observed in PLP-induced disease or other EAE models in the wild type mice, the CNS pathology in phPLP175-194-induced EAE in HLA-DQB1*0602 transgenics was more pronounced in the brain rather than in the spinal cord. Such a strong involvement of cerebellum and brainstem exceeding that of spinal cord, which has been suggested by Stromnes et al.  to be a feature of Th17-driven disease, corresponded to the high Th17 secretion by PLP175-194-primed LNC derived from DQB1*0602-Tg mice. It should be noted that the DRB1*1501-Tg line (originated from C. David and maintained in our animal facility as homozygotic line for several years) which is refractory to PLP139-151- and to PLP175-194-induced EAE, or to MOBP15-36- induced EAE , is susceptible in our hands to EAE induced by MOG35-55 and to MBP89-104 (data not shown), indicating that the disease-resistance or -susceptibility is the result of a selective DRB1*1501-associated Th2 autoimmunity, and DQB1*0602-associated Th1/Th17 autoimmunity, against PLP and MOBP in the setting of the these HLA-humanized mice. Whether such a selective preference for Th2 autoimmunity against PLP or MOBP in the context of DRB1*1501, and Th1/Th17 autoimmunity in the context of DQB1*0602, occurs also in HLA-DR15+ MS should be examined by ex-vivo analysis the patients' responses to these myelin antigen/epitopes in the context of DRB1*1501- and DQB1*0602-associated antigen presentation.
Over the last three decades, autoimmunity against MBP, PLP, and more recently also against other CNS myelin proteins, such as MOG, MOBP and OSP, has been extensively investigated in MS patients, HLA class-II Tg mice, and in wild-type mice, as major target antigens in MS. Clearly we do not yet know the full extent of CNS proteins that may become autoimmune targets in MS, but the credentials of PLP as a target, both in human MS and mouse models are very compelling. When comparing PLP and MBP, PLP is stronger and dominant encephalitogen at least in some EAE models, particularly in SJL/J mice where PLP139-151 and PLP175-194 are the co-dominant encephalitogenic PLP epitopes . TCRPLP transgenic mice on the SJL/J background develop spontaneous EAE with a relatively high frequency  compared with TCRMBP transgenic mice on B10.PL background . Further, upon EAE induction with whole spinal cord homogenate, the dominant T cell response is directed against PLP139-151 . These PLP encephalitogenic peptides are also seen in MS patients' responses. A study on high avidity myelin specific T cells in MS documented that both PLP139-151 and the PLP178-191 epitopes are key targets of high avidity T-cell allels and clearly elevated in MS versus healthy controls . Perhaps most tellingly, a clinical/MRI relapse triggered in a patient as consequence of an altered peptide ligand (APL) clinical trial was temporally correlated with spread of their response to PLP 190-209 . Hence, the potential contribution of the autoimmunity against PLP to the pathogenesis of MS, which is likely to be DQB1*0602-associated in HLA-DR15+ MS, should be quite significant.