High-resolution HLA class II sequencing of Swedish multiple sclerosis patients

1 INTRODUCTION

A recent family-study used high-resolution sequencing to report a novel predisposing association to MS for HLA-DPB1*104:01, independent of HLA-DRB1*15:01 haplotype. High-resolution sequencing also conferred negative MS association to HLA-DRB1*01:01:01, HLA-DQB1*03:01 and HLA-DQB1*03:03 (Osoegawa et al., 2021) and HLA-DRB1*14:04:01 (Vinoy et al., 2021).

Secondary HLA-DRB loci, including HLA-DRB3, DRB4 and DRB5, have been less defined for MS. HLA-DRB5*01:01 is found in linkage disequilibrium with HLA-DRB1*15:01 and therefore found in increased frequency among MS patients. Both HLA-DRB5*01:01 and HLA-DRB1*15:01 present peptides in heterodimer pocket-formation with non-polymorphic HLA-DRα. HLA-DRB5*01:01 has been indicated to modulate the presumed disease promoting HLA-DRB1*15:01 antigen-presentation since rare HLA-DRB5*null haplotypes among HLA-DRB1*15:03 African-American MS patients correlated with an increased risk of developing severe secondary progressive MS (Caillier et al., 2008).

Sweden has one of the highest prevalence of MS worldwide (Ahlgren et al., 2014; Ahlgren et al., 2012), partly explained by the high frequency of the HLA-DRB1*15:01-DQB1*06:02 haplotype in the population.

The aim of the studies was to define HLA genetics among MS patients not associated with HLA-DRB5*01:01:01-DRB1*15:01:01-DQA1*01:02:01-DQB1*06:02:01. Here, we used high-resolution sequencing to determine the associations between MS and HLA-DRB3, DRB4, DRB5, DRB1, DQA1, DQB1, DPA1 and DPB1. This kind of typing offers substantial advantages over methods of imputation and, therefore, more accurately captures the polymorphic complexity of the HLA region.

2 METHODS AND MATERIALS

3 RESULTS

4 DISCUSSION

From the present study, we first report that previously found HLA-associations to MS, most prominently to DR15-DQ6, can be confirmed in 100 MS patients from Southern Sweden. The three genotypes with increased frequency in MS included the HLA-DR15-DQ6 haplotype together with previously reported HLA-DRB1*08:01 (HLA-DR8-DQ4), HLA-DRB3*01-DR3-DQ2, while the association with HLA-DR7-DQ2 appear novel.

Here, we could not confirm recent allelic MS associations using high-resolution sequencing to HLA-DPB1*104:01 (predisposing), HLA-DQB1*03:01 and HLA-DQB1*03:03 (protective), and HLA-DRB1*14:04:01 (protective) (Osoegawa et al., 2021; Vinoy et al., 2021). These alleles were not found to be commonly present in the Swedish population, with exception of HLA-DQB1*03:01. In our study, we found HLA-DQB1*03:01 on the extended HLA DRB3*02:02:01-DQA1*05:05:01 haplotype, with various combinations of DRB1-alleles (11:01:01, 11:04:01, 12:01:01, 13:05:01). Due to high degree of polymorphism presented, we suggest that mechanistic studies would be necessary to fully determine the role of HLA-DQB1*03:01 to risk and progression of MS.

While HLA-DRB3*01:01:02 was considered protective against MS at the allelic level, the extended genotype, abbreviated HLA-DR15-DQ6-DRB3*01-DR3-DQ2, was considered a predisposing risk factor. This highlights the limitations of studies of single alleles relative to studies of extended haplotypes and genotypes. We speculate that the shifting association for HLA-DRB3*01:01:02 from protective to predisposing, could be related to mechanisms of epistasis. Previously, HLA-DQA1*01:02 was demonstrated to increase risk of MS if found in trans position to the haplotype of HLA-DR15-DQ6 (Lincoln et al., 2009).

Lastly, we report heterogeneity in HLA associations to MS given that, among 100 patients, 69 different extended HLA-DR-DQ genotypes were identified. Approximately half of all genotypes (36 of 69) and two-thirds of all patients (68 of 100) encompassed haplotypes of either HLA-DR15-DQ6, HLA-DRB3*01-DR3-DQ2 or HLA-DRB3*02-DR3-DQ2.

A substantial number of studies over the last 50 years have associated HLA genetics to MS (Hollenbach & Oksenberg, 2015). The extended HLA-DR15-DQ6 haplotype is consistently reported, both here and elsewhere, as the major genetic contributor to MS. Some suggestions of the presence of this extended haplotype at the age at onset (Hensiek et al., 2002; Masterman et al., 2000; The International Multiple Sclerosis Genetics Consortium & The Wellcome Trust Case Control Consortium 2, 2011), but not during the clinical course of MS, have been suggested (Barcellos et al., 2006). Although a dose effect may explain the data (Barcellos et al., 2006; Barcellos et al., 2003), our study failed to corroborate this finding, as the prevalence of HLA-DR15-DQ6 homozygosity in Swedish population controls was found to be fairly high [2.5% (16/636 controls) as compared to 6% (6/100 MS patients]. There are conflicting reports as to whether the predisposition to MS stems from the HLA-DR or DQ region. For example, strong linkage disequilibrium within the HLA-II region among European populations has made it difficult to establish separate influences of DRB1*1501 and DQB1*0602 in the extended HLA-DR15-DQ6 haplotype. Studies with inclusion of African American populations with diverse patterns of linkage disequilibrium found a selective association to DRB1*1501, independent of DQB1*0602 (McElroy et al., 2010; Oksenberg et al., 2004).

The group of non-DR15-DQ6 associated MS patients (44 out of 100 patients) is of interest, complementary triggers and pathways of autoimmunity could be present. It cannot be excluded that environmental factors are different between MS patients negative for HLA-DR15-DQ6 but positive for HLA-DR7-DQ2. Studies in larger and diversified cohorts would be needed to define MS in this group of patients.

Historically, the HLA-DRB1*03:01 allele has been associated with MS. An increased risk was reported when the HLA-DRB1*03:01 allele is homozygous (Moutsianas et al., 2015). In contrast our study found a similar frequency at the allelic level (patients 11.5%–12.0% controls), and there were no MS homozygous patients present [MS patients; 0/100 (0%) and controls; 10/636 (1.6%)]. Combining HLA-DRB1 and HLA-DRB345 alleles into haplotypes revealed a risk associated with HLA-DRB3*02-DR3-DQ2, while HLA-DRB3*01-DR3-DQ2 was associated with MS at a genotypic level. This exemplifies that risk associations and protective associations do not necessarily agree at the allelic, haplotypic and genotypic level.

The total of six MS-associated extended HLA-DR-DQ haplotypes found in the studies means that other haplotypes than DR15-DQ6 may contribute to MS. However, combined into genotypes, three MS-associated variants were found in the HLA-DR15-DQ6 haplotype combined with HLA-DR8-DQ4, HLA-DRB3*01-DR3-DQ2, HLA-DR7-DQ2. A plausible explanation for this outcome is the greater depth afforded by high-resolution sequencing, causing a more accurate reflection of the wide polymorphism found in the HLA region. Another plausible explanation is a less dominant effect of a few specific HLA genes. The notion of smaller influences by many genes has been supported by low large-family linkage studies (Sawcer et al., 2014). The high-resolution HLA typing makes it possible to further dissect the possible contribution of other alleles, perhaps haplotypes, to the risk of MS.

A weakness of our study was the fact that the HLA-DPA1 and HLA-DPB1 alleles could not be assigned to extended haplotypes and genotypes. This was largely attributed to the very high frequency of HLA-DPA1*01:03:01 (>80%) among both patients and controls, which precluded computation of the other rare allele variants among the families that were used as templates (Lind et al., 2019). Another limitation was the missing information of age at first symptom and family history of MS, these parameters would have been of importance especially among non-HLA-DR15-DQ6 patients.

It appears overly simplistic to characterize MS as having a single aetiology and pathogenesis. The heterogeneous clinical presentation, variable clinical course, inconsistency in genetic markers, unpredictable therapeutic response and diverse histopathological findings may be indicative of divergence in the demyelination pathways (Lucchinetti et al., 2000; McFarland & Martin, 2007). The high-resolution HLA as presented here would represent an important approach to better delineate MS aetiology and pathogenesis in the future.

In summary, we have confirmed previous HLA associations to MS, while also offering plausible explanations to help guide future research. Using high-resolution sequencing, we have highlighted the widespread polymorphisms as 69 distinct genotypes present among 100 Swedish MS patients and further studies in a larger number of patients will be needed to further delineate MS not associated with HLA-DR15-DQ6.2.

CONFLICT OF INTEREST

There are no conflicts of interest.