A systematic review and meta-analysis of HLA class II associations in patients with IgG4 autoimmunity

Autoimmune diseases caused by pathogenic IgG4 subclass autoantibodies (IgG4-AID) include diseases like MuSK myasthenia gravis, pemphigus vulgaris or thrombotic thrombocytopenic purpura. Their etiology is still unknown. Polymorphisms in the human leukocyte antigen (HLA) gene locus, particularly in HLA-DRB1, are known genetic susceptibility factors for autoimmune diseases. We hypothesized a similar role for HLA polymorphisms in IgG4-AID and conducted a systematic review and meta-analysis with case–control studies on IgG4-AID based on MOOSE/ HuGENet guidelines. Genotype (G) and allele (A) frequencies of HLA-DQB1*05 (G: OR 3.8; 95% CI 2.44–5.9; p < 0.00001; A: OR 2.54; 95% CI 1.82–3.55; p < 0.00001) and HLA-DRB1*14 (G: OR 4.31; 95% CI 2.82–6.59; p < 0.00001; A: OR 4.78; 95% CI 3.52–6.49; p < 0.00001) and the HLA-DRB1*14-DQB1*05 haplotype (OR 6.3; 95% CI 3.28–12.09; p < 0.00001/OR 4.98; 95% CI 3.8–6.53; p < 0.00001) were increased while HLA-DRB1*13 (G: OR 0.48; 95% CI 0.34–0.68; p < 0.0001; A: OR 0.46; 95% CI 0.34–0.62; p < 0.00001) was decreased in IgG4-AID patients. In conclusion, the HLA-DQB1*05, HLA-DRB1*14 alleles and the HLA-DQB1*05-DRB1*14 haplotype could be genetic risk factors that predispose for the production of pathogenic IgG4 autoantibodies and the HLA-DRB1*13 allele may protect from IgG4 autoimmunity.

Legend for supplementary figures S1-S104: † Study was included in the qualitative synthesis, but excluded from the meta-analysis as it did not fit all selection criteria. ‡ Study did not differentiate between disease subgroups of pemphigus or CIDP.
* Study was included after discussion with W.B.. ** Study in which the same control group was used for pemphigus foliaceus and pemphigus vulgaris, here data were pooled for analysis.

Protocol for a HuGE association review and meta-analysis
Cover sheet: Are distinct alleles of the HLA-DRB1/DQB1 gene loci associated with class I IgG4 autoimmune diseases: a systematic association review and meta-analysis Background: The human leucocyte antigen (HLA) genes are expressed on antigen-presenting cells and epithelial cells in the thymus, and they code for the MHC class I and II cell surface receptors that are essential for presenting antigen to T-cells. MHC class II molecules (HLA-DR, HLA-DQ, and HLA-DP) present peptides to T helper cells. They are comprised of two different protein chains, α and β, that are non-covalently associated. HLA-DRB1 is the gene that codes for the β1 -chain of the DR isotype and HLA-DQB1 codes for the β1 -chain of the DQ isotype of the MHC class II cell surface receptor. HLA genes, especially the HLA-DRB1 gene, are highly polymorphic. In 2015, 1883 different alleles for HLA-DRB1 have been reported (Arango et al., 2017). The highest polymorphism is in the β1 domain of HLA-DRB1 which is in contact with the peptides that are presented to T-cells. The gene products of distinct HLA alleles will present distinct subsets of peptides (Pritchard et al., 2015). HLA genes are considered as a major contributor to genetic susceptibility to develop autoimmune diseases, and specifically HLA-DR has been linked to presentation of autoantigens to developing T-cells in the thymus and is associated with a range of autoimmune diseases, e.g. rheumatoid arthritis, diabetes mellitus type I, multiple sclerosis or systemic lupus erythematosus (Alvarez et al., 2015). It is thought that the presence of distinct HLA alleles may contribute to an aberrant presentation of self-peptide to autoreactive T helper cells in the thymus, thereby providing one of the first steps in the pathogenesis of autoimmune diseases (Arango et al., 2017). Therefore, observing an association of specific alleles with distinct autoimmune diseases are important as they may foster our understanding of disease aetiology and pathogenesis and may present a diagnostic soft marker.

Rationale
IgG4 autoimmune diseases (IgG4-AID are a new group of autoimmune diseases first collectively described in 2015 (Huijbers et al., 2015) that are caused by pathogenic IgG4 autoantibodies targeting different antigens and sharing important mechanistic and therapeutic commonalities (Koneczny, 2018) (table 1). They affect one of four different organs (1. The peripheral and central nervous system, 2. skin and mucosa, 3. the kidneys and 4. the vascular system) and were previously not recognized to be related, but they share (1) IgG4 as predominant antibody subclass, and (2) the same pathogenic mechanism, which is an IgG4-mediated blocking of protein-protein interaction instead of complement mediated injury of target-organs (Koneczny, 2018, Koneczny et al., 2013), (3) they also share low disease prevalences of less than 5 per 10,000 (Koneczny, 2020, Autoimmun. Rev. in press). In MuSK MG and pemphigus vulgaris, there is a geographical decline of cases with degrees of latitude with high prevalence in the south and lower prevalence in the north, at least in Europe IgG4-AID are classified by the degree of evidence for IgG4 pathogenicity (Koneczny, 2018), with IgG4 being proven as directly pathogenic in class I IgG4-AID. Class II and class III IgG4-AID are considered as potential IgG4-AID, but the evidence to date is insufficient and does not allow to include or exclude class II/III in the group of validated class I IgG4-AID. The six known class I IgG4-AID are 1.MuSK myasthenia gravis (antibodies to MuSK), 2. pemphigus vulgaris (antibodies to desmoglein 3), 3.pemphigus foliaceus (antibodies to desmoglein 1 and/or 3), 4. chronic inflammatory demyelinating polyneuropathy (CIDP) with antibodies against Contactin-1, 5. CIDP with antibodies against Neurofascin 155 and 6. thrombotic thrombocytopenic purpura with antibodies against ADAMTS13 (Koneczny, 2020, Autoimm. Rev. in press). A narrative review indicated a frequent genetic association of the individual IgG4-AID with a recurrent list of HLA supertypes (HLA-DRB1*04,11,14,15 and DQB1*05, Table 1), which suggests a shared immunopathogenesis (Koneczny, manuscript in preparation). However, it was a narrative review and not a systematic review with meta-analysis, which would provide stronger evidence of an association. To date, no systematic review or association review has been conducted to analyse whether IgG4-AID are overall genetically associated with the same HLA alleles. In this association review with meta-analysis, we aim to investigate if there is an overall association between class I IgG4-AID and HLA gene loci following the HuGENet™ HuGE Review Handbook, version 1.0.
Primary objective: Are individual class I IgG4-AID associated with distinct variants of the HLA gene loci?
Secondary objective: Are class I IgG4-AID collectively associated with distinct variants of the HLA gene loci?  2) Studies reporting the association of gene variants of the HLA-DR and/or HLA-DQ gene locus, including allele, genotype or haplotype frequency.
We will extract and collect data from tables and running text in the included manuscripts. If data is incomplete, unpublished or unavailable, we will attempt to retrieve the data by contacting the corresponding authors of the study by email.
We will extract the following information from each included study: Primary author, year of publication, full bibliographic information, standardized (e.g. Vancouver), demographic information of patients and controls (sex, age), country, type of IgG4-AID and/or type of autoantibody, affected organ, type of control, HLA typing method, sample size, genotype frequency, allele frequency, haplotype frequency, HLA supertype frequency, OR and 95%CI. If the data is combinable, we aim to conduct a meta-analysis for association between HLA variants (HLA allele, genotype, haplotype and/or supertype) and class I IgG4-AID individually and, if feasible, collectively.
Null hypothesis: HLA-DRB1*04,11,14,15 and/or HLA-DQB1*05 are not associated with class I IgG4-AID individually or collectively. We will calculate odds ratios (ORs) for allele, haplotype and genotype frequencies with 95% confidence intervals (95% CIs) to assess strength of association. The heterogeneity of the included studies will be measured using Cochran Q's test and I 2 value. The combined effect of the included studies (pooled OR) will be calculated with the Mantel-Haenszel test.
The publication bias will be tested by using funnel plots and the Egger's test. P values will be adjusted for multiple comparisons by the false discovery rate (FDR) method and will be corrected separately for single locus analysis and haplotype analysis. Corrected P values <.05 will be considered statistically significant.

Potential conflicts of interest
The authors declare no conflict of interest.