#PAGE_PARAMS# #ADS_HEAD_SCRIPTS# #MICRODATA#

Function of multiple sclerosis-protective HLA class I alleles revealed by genome-wide protein-quantitative trait loci mapping of interferon signalling


Authors: Christian Lundtoft aff001;  Pascal Pucholt aff001;  Juliana Imgenberg-Kreuz aff001;  Jonas Carlsson-Almlöf aff002;  Maija-Leena Eloranta aff001;  Ann-Christine Syvänen aff002;  Gunnel Nordmark aff001;  Johanna K. Sandling aff001;  Ingrid Kockum aff003;  Tomas Olsson aff003;  Lars Rönnblom aff001;  Niklas Hagberg aff001
Authors place of work: Rheumatology and Science for Life Laboratories, Department of Medical Sciences, Uppsala University, Uppsala, Sweden aff001;  Molecular Medicine and Science for Life Laboratory, Department of Medical Sciences, Uppsala University, Uppsala, Sweden aff002;  Centre for Molecular Medicine, Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden aff003
Published in the journal: Function of multiple sclerosis-protective HLA class I alleles revealed by genome-wide protein-quantitative trait loci mapping of interferon signalling. PLoS Genet 16(10): e1009199. doi:10.1371/journal.pgen.1009199
Category: Research Article
doi: https://doi.org/10.1371/journal.pgen.1009199

Summary

Interferons (IFNs) are cytokines that are central to the host defence against viruses and other microorganisms. If not properly regulated, IFNs may contribute to the pathogenesis of inflammatory autoimmune, or infectious diseases. To identify genetic polymorphisms regulating the IFN system we performed an unbiased genome-wide protein-quantitative trait loci (pQTL) mapping of cell-type specific type I and type II IFN receptor levels and their responses in immune cells from 303 healthy individuals. Seven genome-wide significant (p < 5.0E-8) pQTLs were identified. Two independent SNPs that tagged the multiple sclerosis (MS)-protective HLA class I alleles A*02/A*68 and B*44, respectively, were associated with increased levels of IFNAR2 in B and T cells, with the most prominent effect in IgDCD27+ memory B cells. The increased IFNAR2 levels in B cells were replicated in cells from an independent set of healthy individuals and in MS patients. Despite increased IFNAR2 levels, B and T cells carrying the MS-protective alleles displayed a reduced response to type I IFN stimulation. Expression and methylation-QTL analysis demonstrated increased mRNA expression of the pseudogene HLA-J in B cells carrying the MS-protective class I alleles, possibly driven via methylation-dependent transcriptional regulation. Together these data suggest that the MS-protective effects of HLA class I alleles are unrelated to their antigen-presenting function, and propose a previously unappreciated function of type I IFN signalling in B and T cells in MS immune-pathogenesis.

Keywords:

B cells – Flow cytometry – Gene expression – Interferons – NK cells – Single nucleotide polymorphisms – T cells

Introduction

Interferons (IFNs) are cytokines central for the immune response against viruses, bacteria and certain malignancies [1, 2]. The production and response to IFNs are strictly controlled, since an inappropriate activation of the IFN system can lead to the break of immune tolerance and the development of autoimmune diseases [3]. Based on receptor usage, the IFNs can be grouped into three different subclasses. Type I IFNs (including IFN-α/β), signal through the ubiquitously expressed IFNα/β-receptor (IFNAR). Type II IFN (IFN-γ) signals through the IFN-γ receptor (IFNGR) mainly expressed by monocytes, B cells and T cells, whereas type III IFNs (IFN-λ) signal through the IFN-λ receptor primarily expressed by epithelial cells [4]. IFN receptor signaling is transmitted via Janus kinas-mediated phosphorylation of transcription factors belonging to the signal transducer and activator of transcription (STAT) family, which drive the transcription of a large number of genes, including chemokines and HLA molecules [5].

Genome-wide association studies (GWAS) have identified thousands of inflammatory disease-associated single nucleotide polymorphisms (SNPs) [6] and many of these are located in genes connected to the IFN system [7]. The majority of disease-associated SNPs are located in non-coding regions of the genome and the underlying molecular mechanisms are mostly unknown. Data from genome-wide expression quantitative trait loci (eQTL) studies demonstrate that the regulatory effects of genetic variants are often context-dependent and cell-type specific [812]. Because protein levels to a large extent are buffered against transcriptional variation through post-transcriptional regulatory mechanisms, the degree to which an eQTL translates to a corresponding protein QTL (pQTL) or conversely is highly variable [13]. Hence, it is important to complement eQTL studies with systematic studies of how genetic variation influences protein levels. Several large-scale genome-wide pQTL studies have analyzed the genetic regulation of frequencies of immune cell subsets, cell-surface protein expression and the plasma proteome during homeostatic conditions [1420], but genome-wide context-dependent and cell-type specific pQTL-studies are more limited [21, 22].

Despite the large number of disease associations in genes linked to the IFN system, and the fact that genetic variants contributing to a balanced IFN response probably have been positively selected for during evolution, there has been no comprehensive study to investigate the genetic regulation of the type I and type II IFN receptor responses on the protein level. In this study we performed a genome-wide cell type-specific pQTL mapping of 45 IFN related traits in six subsets of immune cells from 303 healthy individuals.

Results

Protein QTL mapping of IFN-related traits identifies seven novel pQTLs

To study the effects of genetic variation on the type I and type II IFN systems, we analysed cell surface levels of the ligand binding high-affinity subunit of the IFN-α/β receptor (IFNAR2) and the ligand-binding subunit of the IFN-γ receptor (IFNGR1) as well as down-stream effects of receptor ligation, such as phosphorylation of transcription factors (pSTAT1 and pSTAT4), chemokine production (CXCL9 and CXCL10), and HLA class I and HLA class II receptor levels in six immune cell subsets in PBMCs from 303 healthy individuals using flow cytometry (Fig 1A). Based on surface levels of IFN receptors and the IFN-responsiveness in different cell types, 45 traits were selected for further studies (Fig 1B and S1 Fig). Genome-wide association analysis of these 45 traits identified eight independent genome-wide significant pQTLs (p < 5.0E-8; Fig 1C and Table 1; S2 Data all variants with a p-value < 1.0E-4). Three of the pQTLs regulated IFN receptor levels, whereas five pQTLs controlled IFN receptor responses. There was no overlap between IFN-α and IFN-γ induced pQTLs and the majority of pQTLs were cell-type specific (Fig 1D). Furthermore, of the eight pQTLs, one was a cis-pQTL and seven were trans-pQTLs. All pQTLs were located in non-coding regions of the genome, but one SNP was in complete LD with a missense SNP in FCGR2A (rs1801274, His167Arg) previously associated with systemic lupus erythematosus [23], ulcerative colitis [24] and Kawasaki disease [25]. Follow-up studies revealed that this pQTL was a false positive due to differential Fc-receptor binding of the anti-IFNGR1 mAb (S2 Fig) [26]. This finding stresses the fact that genetic associations between SNPs in the Fc receptor locus and traits determined using antibodies [1416] should be interpreted cautiously and preferably validated using antibodies without Fc-binding properties.

pQTL analysis of IFN-α and IFN-γ receptor levels and responses.
Fig. 1. pQTL analysis of IFN-α and IFN-γ receptor levels and responses.
(A) Schematic overview of the study design. For detailed experimental procedures see the Methods section. (B) Depiction of the 45 traits selected for pQTL analysis marked in green. (C) A combined Manhattan plot of the 45 pQTL analyses with the lowest p-value for each SNP depicted in the plot. The red line denotes the genome-wide significance level of p < 5.0E-8. (D) A heat-map with p-values of the eight genome-wide significant SNPs for all 45 traits studied. Blue indicates a decrease in minor allele carriers, while red indicates an increase in minor allele carriers. (C and D) p-values from the full model as described in the Methods section. (n = 303).
Tab. 1. Summary of genome-wide significant pQTLs.
Summary of genome-wide significant pQTLs.

In summary, we identify seven independent pQTLs regulating IFN receptor expression levels or IFN receptor responses.

Multiple cis-pQTLs regulate IFNAR2 surface levels in a cell-type specific manner

The only genome-wide significant cis-pQTL identified was an association between rs2186355 located upstream of IFNAR2 and IFNAR2 surface protein levels in CD56dim NK cells (p = 1.3E-11). A closer analysis of this locus revealed a second independent pQTL at rs28488669 regulating IFNAR2 levels in CD56dim NK cells in an additive manner (Fig 2A and 2B, p-value for test of interaction = 0.41). While the effect of rs2186355 was specific for CD56dim and CD56bright NK cells, rs28488669 also regulated IFNAR2 surface levels in CD8+ and CD4+ T cells (p < 5.0E-6; Fig 2C and S3 Fig). Both genetic variants have previously been reported as cis-eQTLs for IFNAR2 mRNA expression in whole blood [27], but to the best of our knowledge, this is the first time that these SNPs have been associated with IFNAR2 at the protein level and in a cell-type specific manner.

Cell-type and trait specific effects of pQTLs.
Fig. 2. Cell-type and trait specific effects of pQTLs.
(A-C) Data on IFNAR2 cell surface levels. (A) Regional association plot for IFNAR2 surface levels in CD56dim NK cell and (B) boxplots stratified by rs28488669 and rs2186355. (C) Heat-map of the relation between p-values for the two cis-IFNAR2 pQTLs in different cell-types as specified. (D-E) Data on IFN-α induced phosphorylation of STAT4 (pSTAT4) and STAT1 (pSTAT1). (D) Regional association plot of the IFN-α induced pSTAT4 in CD56dim NK cells and (E) boxplots stratified by rs2298260. (F) Heat-map of the relation between pQTL p-values for IFN-α induced pSTAT1 and pSTAT4 in indicated cell-types. (A and D) The genome-wide significance level (p < 5.0E-8) is denoted by a red line. (A-F) p-values from the full model using single (A, D-F), or two additive SNPs (B-C), (n = 303). Boxplots show median, IQR and range. gMFI = geometric mean fluorescence intensity.

A pQTL in the FOCAD locus associates with an increased type I IFN response in NK cells

The strongest pQTL for an IFN-induced trait was the association between the intronic SNP rs7388989 in FOCAD and IFN-α-induced pSTAT4 in CD56dim NK cells (p = 1.1E-9). This SNP is in LD (r2 = 0.79) with the missense SNP rs2298260 (Thr36Ala) in HACD4 (previously named PTPLAD2), which in addition to IFN-α-induced pSTAT4 in CD56dim NK cells (p = 3.6E-9) was associated with IFN-α-induced pSTAT4 in CD56bright NK cells as well as IFN-α-induced pSTAT1 in monocytes and CD56bright NK cells (p < 5.0E-6; Fig 2D–2F, and S4A and S4B Fig). Conditional analysis showed that both SNPs represented the same association signal (S4C Fig). The increased IFN-α-response in monocytes from minor allele carriers of rs2298260 translated into an increased production of CXCL9 (p = 1.1E-3; S3D Fig), whereas no effect was observed for CXCL10 (p = 0.94; S3E Fig). A SNP in complete LD with rs2298260 (rs2275888, r2 = 1.0) was previously identified as a cis-eQTL for IFNB1 in LPS-stimulated monocytes, which drove a trans-eQTL network consisting of multiple IFN-response genes [8]. Importantly, in contrast to the increased IFN-α-induced pSTAT4/pSTAT1 observed in minor allele carriers of rs2298260 in our study, the LPS-induced expression of IFNB1 was reduced in minor allele carriers in that study. Together these data highlight the importance of context-specific effects of regulatory SNPs.

B cells and T cells carrying the MS-protective HLA-A*02, A*68 and B*44 alleles have increased IFNAR2 surface levels in B cells and T cells

The top pQTL was a trans-association between rs2735099 located 1.2 kb downstream of HLA-A on chromosome 6 and IFNAR2 surface levels in B cells (p = 3.0E-26; Fig 3A, top panel). Conditioning on rs2735099 revealed a second independent signal from the intronic SNP rs17199328 located in HLA-B (p = 1.6E-11; Fig 3A, middle panel). The effect of rs2735099 and rs17199328 was additive with minor allele carriers displaying increased B cell IFNAR2 surface levels (Fig 3B). Similarly, increased IFNAR2 surface levels were seen in CD8+ and CD4+ T cells from minor allele carriers of the two SNPs (Fig 3C and 3D), but not in NK cells or monocytes (pCD56dim = 0.05, pCD56bright = 0.15, pmonocyte = 0.13; S5A Fig). As the genes encoding IFNAR1 and IFNGR2 are located adjacent to IFNAR2, we investigated the possibility of co-regulatory effects, but neither IFNAR1 nor IFNGR2 surface levels in B cells were associated with rs2735099 or rs17199328 (pIFNAR1 = 0.13, pIFNGR2 = 0.32, n = 95; Fig 3E and 3F).

Two independent SNPs additively associate with IFNAR2 protein levels in B cells and T cells.
Fig. 3. Two independent SNPs additively associate with IFNAR2 protein levels in B cells and T cells.
(A) Regional association plots of the trans-IFNAR2 pQTL in B cells before (top) and after conditioning for rs2735099 (middle) and rs17199328 (bottom). (B-D) Boxplots of IFNAR2 levels in B cells (B), CD8+ T cells (C) and CD4+ T (D) cells stratified by rs2735099 and rs17199328 (n = 301). (E-F) IFNAR1 (E) and IFNGR2 (F) surface levels in B cells (n = 95). p-values from the full model using single (A, E and F), or two additive SNPs (B-D). Boxplots show median, IQR and range. gMFI = geometric mean fluorescence intensity.

Imputation of HLA-A and HLA-B alleles from genotype data showed that the minor allele of rs2735099 tagged the HLA-A*02 and HLA-A*68 alleles, whereas rs17199328 tagged the HLA-B*44 allele (Fig 4A). Notably, both A*02 and B*44 are protective for multiple sclerosis (MS) [2830] and a MS-protective effect of HLA-A*68 has also been suggested [28]. Consequently, these three alleles will hereafter be referred to as the as the MS-protective class I alleles.

Increased IFNAR2 protein levels in B cells and T cells from carriers of the MS-protective class I alleles.
Fig. 4. Increased IFNAR2 protein levels in B cells and T cells from carriers of the MS-protective class I alleles.
(A) Correlation between rs2735099 and rs17199328 genotypes and imputed HLA-A (n = 301) and HLA-B alleles (n = 303). (B) The proportion of B cells subsets in bulk B cells (n = 95). (C) IFNAR2 surface levels in subsets of B cells from healthy donors (n = 95). (D) IFNAR2 surface levels in subsets of B cells and T cells from healthy donors (n = 22) and PwMS (n = 26). (B-D) Data are stratified by the sum of the MS-protective class I alleles HLA-A*02, HLA-A*68 and HLA-B*44, as indicated. p-values from the full model (B-C) or a simple linear regression (D), using the sum of MS-protective alleles (0–3) as a continuous variable. Boxplots show median, IQR and range. gMFI = geometric mean fluorescence intensity.

Sub-phenotyping of B cells from 95 individuals revealed that the altered IFNAR2 levels were not due to differences in the proportion of B cell subsets (Fig 4B), but instead driven mainly by differential IFNAR2 surface levels in the IgDCD27+ memory and IgDCD27 double-negative (DN) B cell subsets (pmemory = 5E-14, pDN = 2E-10; Fig 4C). In T cells, the effect of the MS-protective class I alleles was evident in both naïve and memory subsets (S5B Fig). The increased IFNAR2 levels in B cells carrying MS-protective class I alleles was replicated in an independent set of 22 healthy individuals and 26 persons with MS (PwMS; Fig 4D).

In summary, we identify two additive pQTLs resulting in increased IFNAR2 surface levels in B and T cells carrying the MS-protective class I alleles A*02, A*68 and B*44.

Post-transcriptional mechanisms underlie the increased IFNAR2 protein levels in carriers of the MS-protective class I alleles

IFNAR2 is expressed as a soluble secreted isoform (IFNAR2a) and two membrane-bound isoforms consisting of a full-length signalling receptor (IFNAR2c) and a truncated non-signalling receptor (IFNAR2b), which only differ in their intra-cellular domains (Fig 5A) [31]. The anti-IFNAR2 mAb used in our initial pQTL analysis targets the extracellular part of IFNAR2 and thus detects both membrane-bound IFNAR2 isoforms. To study whether the association between the MS-protective class I alleles and IFNAR2 protein levels was due to altered transcription, IFNAR2 mRNA expression levels were determined in B cells (n = 47) using primers that either amplify both membrane-bound isoforms (IFNAR2bc), or primers specific for the IFNAR2b or IFNAR2c isoforms. There were no significant differences in total mRNA levels of IFNAR2 (IFNAR2bc), nor its isoforms, attributed to the MS-protective class I alleles (Fig 5B). The lack of association was confirmed in RNA expression microarray data on B cells [32] (n = 288; p > 0.10 for the three probes targeting IFNAR2) and RNA-sequencing (RNA-seq) data on lymphoblastoid cell lines (LCLs) [33] (n = 373 of European descent; p > 0.10 for 8 different IFNAR2 transcripts).

Decreased type I IFN response in B cells carrying the MS-protective class I alleles.
Fig. 5. Decreased type I IFN response in B cells carrying the MS-protective class I alleles.
(A) A schematic picture of the IFNAR2 isoforms. (B) IFNAR2bc, IFNAR2b and IFNAR2c mRNA levels in isolated B cells determined using qRT-PCR. mRNA levels are normalized to the expression of the reference gene RPL13A and expressed as arbitrary units (AU) (n = 47, simple linear regression using the sum of MS-protective class I alleles (HLAsum) as independent variable). (C) Correlation between IFNAR2bc mRNA levels and IFNAR2 surface receptor levels determined using flow cytometry (n = 47). Data are binned into 0, 1 and ≥2 HLAsum. p-values and Pearson’s correlation coefficient (r) from linear regressions of each group are shown in coloured text. p-values from multiple linear regressions of the combined data using HLAsum (continuous variable) and mRNA values as independent variables in the full model are denoted in black. (D) Heat-maps comparing the correlation between IFNAR2 levels and IFN-α induced pSTAT1 (top) or pSTAT4 (bottom) in indicated cell-types with and without including HLAsum as a covariate in the full model (n = 303). (E-G) Correlation between IFNAR2 protein levels and IFN-α (2000 IU/ml) induced pSTAT1 (E), pSTAT4 (F) and CXCL10 (G) in B cells. p-values as described in (C), (n = 303). Boxplots show median, IQR and range. gMFI = geometric mean fluorescence intensity.

Correlation analysis between mRNA and protein levels revealed that while IFNAR2bc mRNA levels correlated with IFNAR2 protein levels in carriers of ≥2 MS-protective class I alleles, such correlation was not present in individuals carrying 0 MS-protective class I alleles (Fig 5C).

In summary, these data demonstrate that post-transcriptional mechanisms underlie the association between IFNAR2 protein levels and the MS-associated class I alleles.

Decreased response to type I IFN stimulation in B cells and T cells carrying the MS-protective class I alleles

In order to study the downstream consequences of altered IFNAR2 protein levels, we evaluated the level of STAT phosphorylation after stimulation with IFN-α. Intriguingly, IFNAR2 levels did not correlate with IFN-α-induced pSTAT1 or pSTAT4 in B cells unless adjusting for the number of MS-protective class I alleles (Fig 5D), thereby demonstrating an effect of the MS-protective class I alleles on downstream IFNAR2 signalling. IFNAR2 levels correlated positively with IFN-α-induced pSTAT1 and pSTAT4 in B cells from individuals with 0 MS-protective class I alleles, but not in in B cells from carriers of ≥2 MS-protective class I alleles who had reduced STAT phosphorylation in response to IFN-α despite having higher IFNAR2 levels (Fig 5E and 5F). Downstream of the impaired type I IFN-induced STAT phosphorylaton, a decreased production of CXCL10 was observed in B cells carrying the MS-protective class I alleles (Fig 5G). Similar to B cells, both CD8+ T cells and CD4+ T cells carrying the MS-protective class I alleles displayed a decreased pSTAT1 and pSTAT4 in response to IFN-α (S6 Fig).

In an effort to investigate whether the reduced IFNAR response in carriers of the MS-protective class I alleles was due to increased levels of the dominant negative IFNAR2b isoform, we raised polyclonal anti-IFNAR2b and anti-IFNAR2c antibodies in rabbits. However, none of these antibodies, nor a commercial anti-IFNAR2c antibody, detected endogenous levels of IFNAR2 in primary cells.

In summary, these data for the first time link the MS-protective class I alleles to a decreased type I IFN response in B cells and T cells.

The MS-protective class I alleles are cis-eQTLs for HLA-J

As trans-pQTLs are commonly due to secondary effects from a cis-eQTL regulating protein expression of an intermediary protein [34], we next analyzed RNA-seq data from LCLs [33] for cis-eQTLs of the MS-protective class I alleles within a region extending 1 Mb from the two SNPs tagging these alleles. Assuming that the additive effect on IFNAR2 expression of HLA-A*02/A*68 and B*44 was driven by the same mechanisms, we searched for eQTLs with independent additive effects from the class I alleles. Using these criteria, we identified increased expression of the two pseudogenes HLA-J and HCG4P5 in LCLs carrying the MS-protective class I alleles (p < 3.2E-4, corresponding to a p-value of < 0.05 after Bonferroni correction for 155 genes analysed in cis; Fig 6A). The HLA-J RNA sequencing data mapped to three different transcripts. The eQTL was specific for the HLA-J-001 (ENST00000494367.1) transcript (Fig 6A inset) and the eQTL association signal peak overlapped with the IFNAR2 pQTL peak (Fig 6B). HCG4P5 has only one transcript, HCG4P5-001 (ENST00000429656.1) and the eQTL and pQTL association signals did not colocalize (Fig 6B). To search for co-regulated genes in trans, we performed a global eQTL analysis, but no trans-eQTLs were observed (p < 2.1E-6, corresponding to a p-value of 0.05 after Bonferroni correction for 23,722 analyzed genes).

The MS-protective class I alleles are <i>cis</i>-eQTLs of <i>HLA-J</i>.
Fig. 6. The MS-protective class I alleles are cis-eQTLs of HLA-J.
(A) Cis-eQTL analysis within ± 1 Mb from the HLA-A*02/A*68 and HLA-B*44 tag SNPs in LCLs (n = 373). The p-value threshold of p < 3.2E-4, corresponding to a Bonferroni adjusted p-value of < 0.05 (155 analysed genes) is denoted with dotted lines. Data for 28 of the 155 genes are shown. The inset shows the transcript specific cis-eQTL p-values of HLA-J. (B) Correlation between p-values from the IFNAR2 pQTL in B cells and the cis-eQTL in LCLs for the transcript HLA-J-001 (left) and HCG4P5-001 (right). The HLA-A*02/A*68 and the B*44 tagging SNPs are marked in red and blue respectively. (C) Cis-meQTL analysis of the HLA-A*02/A*68 tagging SNP rs2735099 in primary B cells (n = 43). The p-value threshold of p < 2.1E-6, corresponding to a Bonferroni adjusted p-value of < 0.05 (23, 722 analysed CpG-sites) is denoted with dotted red lines.

As differential DNA methylation in transcriptional regulatory region is one possible mechanism driving eQTLs, we next analysed DNA methylation data from primary B cells [35]. This meQTL analysis identified a strong correlation between HLA-A*02/A*68 carriership and decreased methylation levels of a CpG-site 1 kb upstream of HLA-J (cg22012577, p = 4.6E-11; Fig 6C).

Together these data suggest that meQTL-mediated transcriptional regulation of HLA-J is a possible candidate mechanism to further explore in order to delineate the molecular mechanisms underlying the increased IFNAR2 protein levels and/or the uncoupling of IFNAR2 levels from type I IFN activation in carriers of the MS-protective class I alleles.

Discussion

In this study we performed an extensive characterization of the genetic regulation of the type I and type II IFN receptor levels and their response in immune cells through genome-wide pQTL analysis. We identify several SNPs regulating IFN receptor levels and their responses in a cell-type specific manner. Our most important finding is the identification of two SNPs tagging the HLA-A*02/A*68 and the HLA-B*44 alleles that are independently and additively associated with increased surface levels of the IFNAR2 subunit in B cells and T cells. Notably, HLA-A*02 and HLA-B*44 are both protective for the development of MS [2830], a chronic inflammatory disease of the central nervous system driven by the interaction of genetic and environmental factors [36]. Through its close relatedness to HLA-A*02, HLA-A*68 has been suggested to carry the same protective effect on MS as A*02 [28]. Protective effects of the infrequent HLA-B*38 and HLA-B*55 alleles have also been demonstrated [2830]. Although we did not detect an effect of HLA-B*38 and HLA-B*55 on IFNAR2 levels in our study, firm conclusions cannot be drawn as the number of individuals carrying these alleles were low.

HLA-A*02 is beside HLA-DRB1*15:01 the strongest genetic association in MS, but the causal mechanisms underlying the protective function for MS is still unknown. Conceivably, such mechanisms may both include allelic differences in antigen presentation to CD8+ T cells, which may have both pathogenic and protective roles in MS or its experimental models [3740], or effects unrelated to antigen presentation through regulation of mRNA expression of genes in the HLA locus [41].

The association between IFNAR2 levels and the MS-protective class I alleles was most prominent in B cells, and in particularly in the memory and double-negative subsets of B cells. Increased IFNAR2 levels were also seen in these B cell subsets from PwMS carrying the MS-protective class I alleles, but not in T cells. Although the absence of an effect in T cells from PwMS might reflect a power issue, these data confirm that the effect on IFNAR2 protein levels of the MS class I alleles are stronger in B cells than T cells. While classically considered a T cell driven disease, the importance of B cells in MS pathology is now well established, not least by the fact that anti-CD20 mAb is an effective treatment [42]. Both memory and double-negative B cells provide a link between the periphery and CNS as peripheral blood memory and double-negative B cells are clonally related to B cells found in the CSF of PwMS [43]. Memory B cells are of particular interest in MS as Epstein-Barr virus, which is a strong environmental risk factor for MS, reside in these cells. Moreover, the autoproliferation of brain-homing potentially pathogenic effector T cells is dependent on memory B cells [44], and depletion of memory B cells has been suggested as a possible mechanism of action for several of the therapies used today including IFN-β [45], anti-CD20 mAb, cladribine and natalizumab [46].

While cis-pQTLs are mainly driven by transcriptional regulation of the corresponding gene, trans-pQTLs are rarely directly connected to alteration in mRNA expression of the corresponding gene, but are rather driven by transcriptional regulation of a causal intermediate protein in cis of the pQTL [34]. Consistent with this observation, the increased protein levels of IFNAR2 in B cells from carriers of the MS-protective class I alleles were not under transcriptional control. Instead, we found that IFNAR2 protein levels were uncoupled from mRNA abundance in carriers lacking the MS-protective class I alleles.

Data from functional studies demonstrate that, despite presenting with higher IFNAR2 levels, B cells carrying the MS-protective class I alleles have a reduced pSTAT1, pSTAT4 and production of CXCL10 in response to type I IFN stimulation. CXCL10 is a potential biomarker for MS as the levels are elevated in cerebrospinal fluid (CSF) from PwMS and correlate with clinical characteristics such as IgG index, CSF mononuclear counts and T2 lesions [47, 48]. A decreased type I IFN response was also evident in CD4+ T and CD8+ T cells carrying the MS-protective class I alleles, but these effects were not as strong as in B cells. Although speculative, these data would fit in a model where the increased IFNAR2 levels in carriers of the MS-protective class I alleles are due to increased levels of the dominant inhibitory isoform IFNAR2b. The lack of isoform-specific anti-IFNAR2 antibodies precluded studies to directly determine whether this is true, and it is thus possible that other mechanisms, such as alterations in receptor trafficking/sorting or negative feedback signalling, are affected [49, 50].

Given the therapeutic use of IFN-β in MS, the association of decreased type I IFN activation in B cells and T cells carrying the MS-protective class I alleles may at first sound counter-intuitive. However, our data are in agreement with the fact that primary human immune cells carrying the MS-protective TYK2 variant rs34536443 (Pro1104Ala) have a reduced response to IFN-β stimulation [51, 52] and that mice homozygous for the corresponding mutation (Pro1124Ala) are completely protected from developing the animal model of MS, experimental autoimmune encephalomyelitis (EAE) [52]. The protective effects of reduced type I IFN signalling in carriers of the MS protective TYK2 variant might be driven by a shift between Th1/Th17 and Th2 polarization of T cells [51, 52]. A pathogenic role of type I IFN in the initiation of the MS disease is supported by the fact that antibody-mediated neutralization of type I IFN in the EAE model leads to a less severe disease during the early, but not in the established, phase of the disease [53]. Type I IFNs also have protective effects in EAE as both IFN-β and IFNAR1 knock-out mice have an increased disease severity [54, 55], but this effect is confined to the myeloid cells [54]. Thus, the effects of type I IFNs are complex and the effects on MS pathology probably differ depending on the cell-type and phase of the disease and whether a complete or partial reduction of IFN signalling is studied.

In an effort to determine the underlying molecular mechanism of the altered IFNAR2 levels we identified the IFN-α-induced pseudogene HLA-J [56] as a cis-eQTL for both MS-protective class I alleles. The differential expression of HLA-J was possibly mediated via altered DNA methylation as a meQTL upstream of HLA-J was observed for the HLA-A*02/*68 alleles. Although the term pseudogene implicates the absence of a functional effect, there is emerging data demonstrating biological effects of pseudogenes, including the regulation of immune responses during inflammation and infection [5759]. Allelic variants of HLA-J have previously been shown to influence HLA-A surface protein levels [60] and it is possible that the same mechanisms operate for IFNAR2. However, further studies are needed in order to determine whether HLA-J mRNA levels are causatively linked to IFNAR2 protein levels, or whether these two intermediate phenotypes of the MS-protective class I alleles are epiphenomenally related.

Although a definite functional link between MS pathology and the observed differences in type I IFN response, in particular in B cells remains to be established, our data challenge the prevailing view that the protective influence of HLA-A*02 and B*44 are mechanistically related to preferences in peptide binding [61], either in the defence against a pathogen associated with MS such as EBV [62], or presentation of peptides related to CD8+ T cells involved in suppressing disease [39]. Instead, we put forward strong and unexpected evidence that their influence on MS is related to the type I interferon system. Further studies to unravel the detailed molecular mechanisms underlying the altered IFNAR2 protein levels and signalling in carriers of the MS-protective class I alleles are warranted as this may identify novel therapeutic targets to reduce the I IFN signalling specifically in B cells and T cells.

Materials and methods

Ethic statement

The study was approved by the Regional Ethical Review Board in Uppsala (2009/013) and Karolinska Institute (2009/2107-31/2). Written informed consent was obtained from all participants.

Study participants

304 healthy individuals were recruited from Uppsala Bioresource. One individual was excluded during the genetic data quality control (see below) leaving a total of 303 healthy individuals. The median age was 52 years (range 22–81 years), 78% were females and 70% were CMV-seropositive. PBMCs were isolated from buffy coats using Ficoll density gradient centrifugation, and aliquots of 20E6 PBMCs were cryopreserved in FCS supplemented with 10% DMSO (Sigma-Aldrich). For replication, 22 additional healthy individuals from Uppsala Bioresource and 26 persons diagnosed with MS according to the revised McDonald criteria [63] were recruited. The median age of the PwMS was 41 years (range 28–63 years) and 85% were females. Of the PwMS, 24 presented with a relapsing-remitting MS in remission and two with a secondary-progressive MS. One of the PwMS was treatment naïve and 25 PwMS were treated with natalizumab. PBMCs were isolated from CPT-tubes and cryopreserved in FCS supplemented with 10% DMSO.

Interferon stimulation of PBMCs

PBMCs were thawed in complete medium (RPMI-1640 (Sigma-Aldrich) supplemented with 2 mM HEPES, penicillin, streptomycin, L-glutamine and 10% FCS) and rested overnight in 6-well plates at 37°C in a cell incubator with 5% CO2. Cells were transferred to V-bottomed 96-well plates and fluorescently barcoded using CellTrace Violet (Invitrogen, 0 nM (unstimulated), 83 nM (IFN-α) and 667 nM (IFN-γ)), and subsequently cultured in the presence or absence of IFN-α2b (IntronA, MSD) or IFN-γ (PeproTech) at ~EC100 concentrations (2000 U/ml and 1 ng/ml, respectively) [64]. IFNAR2 and IFNGR1 levels were determined before stimulation. Phosphorylation of STAT1 and STAT4 was determined after 20 min, intra-cellular levels of CXCL9 and CXCL10 after 7 h with golgiplug (BD Biosciences) added during the last 6 h, and HLA class I and HLA class II expression after 28 h.

For the genome-wide association analysis, PBMCs were analysed in four consecutive experiments, and cells from 16 individuals were included as internal controls in all four experiments to allow for batch normalisation.

Flow cytometry

Cells that were analysed for pSTAT and CXCL were fixed in 2% paraformaldehyde (PFA, Electron Microscopy Sciences) before flow cytometric stainings, whereas cells analysed for IFNR and HLA were fixed in 2% PFA after the last staining. Barcoded cells from each stimulus were pooled before flow cytometric stainings. For analysis of pSTAT and CXCL, cells were permeabilized with Perm Buffer III (BD Biosciences) or 0.5% saponin (Sigma-Aldrich), respectively. Stainings were performed in PBS supplemented with 2 mM EDTA and 0.5% human serum albumin (Octapharma) and FcBlock (Miltenyi Biotec) using two different 8-color panels with fluorochrome-conjugated antibodies (see below) and gating schemes as specified in S7 Fig. The IFNAR2 mAb REA124 was validated by transfection of HEK293T cells with siRNA targeting IFNAR2 ((dTdT-)GCACCATAGTGACACTGAA-dTdT), and plasmids encoding IFNAR2b or IFNAR2c (#RC201212 and #RC238664, Origene) (S8 Fig).

Flow cytometry data were acquired on a BD FACSCanto II instrument (BD Biosciences) and analysed with FlowJo version 10.5.3. IFN-induced read-outs were determined by subtracting the geometric mean fluorescence intensity (gMFI) in unstimulated cells from stimulated cells, i.e. ΔgMFI = gMFIIFN−gMFIunstimulated.

Antibodies used in flow cytometry

Mouse monoclonal antibodies

CD3, REA613, APC-Vio770, Miltenyi Biotec, Cat# 130-113-136; RRID:AB_2725964

CD3, UCHT1, PerCP-Cy5.5, BD Biosciences, Cat# 560835; RRID:AB_2033956

CD4, SK3, BV510 BD Biosciences, Cat# 562970; RRID:AB_2744424

CD8, SK1, FITC, BD Biosciences, Cat# 345772

CD14, MφP9, FITC, BD Biosciences, Cat# 345784

CD16, 3G8, BV510, BD Biosciences, Cat# 563829; RRID:AB_2744296

CD19, HIB19, FITC, BD Biosciences, Cat# 555412; RRID:AB_395812

CD20, H1, PE-Cy7, BD Biosciences, Cat# 561175; RRID:AB_10562032

CD20, REA780, PE-Vio770, Miltenyi Biotec, Cat# 130-111-340; RRID:AB_2656074

CD27, M-T271, PE-Cy7, BD Biosciences, Cat# 560609; RRID:AB_1727456

CD45RA, HI100, APC-H7, BD Biosciences, Cat# 560674; RRID:AB_1727497

CD56, N901, PerCP-Cy5.5, Beckman Coulter, Cat# A79388

CD57, TBO1, PE-Cy7, eBiosciences, Cat# 25-0577-42; RRID:AB_2573354

IgD, IA6-2, APC-Cy7, Biolegend, Cat# 348218; RRID:AB_11203722

IFNAR2, REA124, APC, Miltenyi Biotec, Cat# 130-099-560; RRID:AB_2652223

IFNGR1, 92101, PE, R&D systems, Cat# FAB673P; RRID:AB_2264548

pSTAT1, 4a, AF647, BD Biosciences, Cat# 562070; RRID:AB_10896129

pSTAT4, 38-pSTAT4, BD Biosciences, Cat# 562073; RRID:AB_10895804

CXCL9, B8-11, PE, BD Biosciences, Cat# 566013; RRID:AB_2739458

CXCL10, REA334, APC, Miltenyi Biotec, Cat# 130-104-963; RRID:AB_2651478

HLA-ABC, REA230, PE, Miltenyi Biotec, Cat# 130-101-448; RRID:AB_2652082

HLA-DRPQ, REA332, APC, Miltenyi Biotec, Cat# 130-104-824; RRID:AB_2652180

Polyclonal antibodies

IFNAR1 rabbit polyclonal, Abgent, Cat# AP8550c; RRID:AB_10553443

IFNAR2c, rabbit polyclonal, Abcam, Cat# ab56070; RRID:AB_880736

IFNGR2, goat polyclonal, APC, R&D systems, Cat# FAB773A; RRID:AB_2121710

donkey polyclonal anti-Rabbit IgG, PE, BD Biosciences, Cat# 558416; RRID:AB_1645469

Genotyping, quality control and imputation of genetic variants

DNA from the healthy individuals were genotyped on the Infinium Global Screening Array v2.0 (Illumina). The genetic structure of the study population was analysed using EIGENSOFT [65] after excluding long-range linkage disequilibrium (LD) regions [66], SNVs with MAF < 0.05 and SNVs in LD r2 > 0.2 using the HapMap3 cohort as reference population. All individuals mapped to the CEU population (< 10 SD from the mean HapMAP CEU population for the first 5 principle components) and passed the exclusion criteria for heterozygosity (> 5 SD from population mean), call rate (< 98%) and discordance for reported sex. One individual was excluded due to relatedness (π-hat > 0.1875, PLINK software v1.90).

Variants with a call rate < 98%, a MAF < 1% or with deviation from Hardy-Weinberg equilibrium (p < 0.00001) were excluded. Imputation of additional variants was performed at the Sanger imputation service [67] using the Haplotype reference consortium r1.1 reference and the “pre-phase with EAGLE2 and impute” pipeline [68]. Imputed genotype calls with a genotype probability score below 0.9 were set to missing and variants with an info score below 0.8, a MAF < 0.01, deviation from Hardy-Weinberg equilibrium (p < 0.0001) or a call rate < 95% were excluded. The final dataset contained 303 individuals and 5,084,123 genetic variants.

Imputation of HLA alleles

HLA types were imputed using HIBAG version 1.18.1 [69] and the multi-ethnic prediction model specific for Illumina Infinium Global Screening Array v2.0.

CMV status

CMV status was determined serologically at the first visit at the blood donation centre. For 38 donors with missing information, CMV status was imputed from 270 flow cytometry-based traits (cell proportions and gMFI levels in the pQTL screening) using missForest version 1.4 [70]. The random forest-based method showed an average prediction accuracy of 86% in 100 trails in which 10% of individuals with known CMV status were randomly left out for imputation.

pQTL mapping

Flow cytometry data were normalised for inter-experimental variation by applying factor correction, i.e. the average of internal control samples for the individual experiment divided by the average of internal control samples for all four experiments. Each trait was assessed for normality by Shapiro–Wilk test and visual evaluation of Q-Q plots. Data that were not normally distributed were square root or log-transformed. The pQTL analysis was performed by linear regression using PLINK version 1.90b6.7 [71]. Age, sex, CMV status, experiment batch, year and season (summer/winter) of blood sampling were included as covariates and is referred to as the “full model”. For the genome-wide association analysis, the standard threshold of p < 5.0E-8 was applied, whereas for other analyses, a p-value < 0.05 was considered significant.

mRNA expression analysis

B cells and monocytes were isolated sequentially from 40E6 PBMCs from 48 healthy individuals through positive selection of CD19+ cells followed by positive selection for CD14+ cells (MACS Cell Separation, Miltenyi Biotec). RNA was isolated using NucleoSpin RNA kit (Macherey-Nagel), and cDNA generated using QuantiTect Reverse Transcription kit (Qiagen). B cell mRNA from one individual was excluded due to an A260/A280 ratio of 0.9. Fast TaqMan qPCR was performed using TaqMan Fast Advanced Master Mix (Thermo Fisher Scientific) and TaqMan gene expression assays: Hs01022064_m1 (IFNAR2c), Hs00174198_m1 (IFNAR2bc), Hs00988304_m1 (IFNGR1), Hs03043885_g1 (RPL13A) (all from ThermoFischer). mRNA levels of IFNAR2b were determined with custom designed primers and probes [72]: Forward primer: CTATTCACAGGTGCAGTCATAATGC, Reverse primer:: GCACGCTTGTAATCCCAGCTA, Taqman probe: FAM-CAGTCGTCCTGCCTAAGCTTCCCCA-TAMRA. Data was acquired using a 7500 Fast Real-Time PCR System (Applied Biosystems) and analysed by the 2-ΔCt method using RPL13A as the reference gene.

eQTL analysis of publically available datasets

Microarray data on IFNAR2 expression in B cells from 288 individuals [32] were analysed by linear regression adjusting for sex using the proxies rs2735099 > rs2735097 (R2GBR = 0.88) and rs17199328 > rs2596477 (R2GBR = 1.0). RNA-seq data on LCLs from 373 individuals of European descent [33] were analysed using a linear regression adjusting for sex and population using the proxy rs2735099 > rs2735097 (R2EUR = 0.92).

meQTL analysis

Previously published data on CpG DNA methylation levels (HM450K BeadChip, Illumina) in B cells from 43 healthy individuals [35], were analysed for meQTLs within a 1 MB window of rs2735099 using a linear regression model including age and sex as covariates.

Statistical analysis

Statistical analyses were performed in R version 3.5.1 [73] or GraphPad Prism version 8. All statistical tests were two-tailed.

Supporting information

S1 Fig [a]
Cell-type specific interferon receptor expression and response.

S2 Fig [a]
A false positive pQTL signal resulting from differential Fc-receptor binding.

S3 Fig [a]
Cell-type specificity of IFNAR2 cis-pQTLs.

S4 Fig [a]
Characterization of the IFN-α induced pSTAT/ pQTL.

S5 Fig [a]
Cell-type specificity for the IFNAR2/ pQTL.

S6 Fig [r]
Decreased type I IFN response in T cells carrying the MS-protective class I alleles.

S7 Fig [a]
Gating strategy.

S8 Fig [a]
Validation of the monoclonal IFNAR2 antibody REA124.

S1 Data [xlsx]
An excel file containing raw data on geometric mean fluorescence intensities, covariates and HLA-A and HLA-B imputed alleles.

S2 Data [xlsx]
An Excel file with p-values for all 45 traits for each SNP with p < 1E-4 for at least one trait in the pQTL mapping.

S3 Data [xlsx]
An Excel file with data underlying figures.


Zdroje

1. Trinchieri G. Type I interferon: friend or foe? J Exp Med. 2010;207(10):2053–63. doi: 10.1084/jem.20101664 20837696

2. Ivashkiv LB. IFNgamma: signalling, epigenetics and roles in immunity, metabolism, disease and cancer immunotherapy. Nat Rev Immunol. 2018;18(9):545–58. doi: 10.1038/s41577-018-0029-z 29921905

3. Crow MK, Rönnblom L. Type I interferons in host defence and inflammatory diseases. Lupus Sci Med. 2019;6(1):e000336. doi: 10.1136/lupus-2019-000336 31205729

4. Hagberg N, Rönnblom L. Systemic Lupus Erythematosus—A Disease with A Dysregulated Type I Interferon System. Scand J Immunol. 2015;82(3):199–207. doi: 10.1111/sji.12330 26099519

5. Rusinova I, Forster S, Yu S, Kannan A, Masse M, Cumming H, et al. Interferome v2.0: an updated database of annotated interferon-regulated genes. Nucleic Acids Res. 2013;41(Database issue):D1040–6. doi: 10.1093/nar/gks1215 23203888

6. Buniello A, MacArthur JAL, Cerezo M, Harris LW, Hayhurst J, Malangone C, et al. The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic Acids Res. 2019;47(D1):D1005–D12. doi: 10.1093/nar/gky1120 30445434

7. Bronson PG, Chaivorapol C, Ortmann W, Behrens TW, Graham RR. The genetics of type I interferon in systemic lupus erythematosus. Curr Opin Immunol. 2012;24(5):530–7. doi: 10.1016/j.coi.2012.07.008 22889593

8. Fairfax BP, Humburg P, Makino S, Naranbhai V, Wong D, Lau E, et al. Innate immune activity conditions the effect of regulatory variants upon monocyte gene expression. Science. 2014;343(6175):1246949. doi: 10.1126/science.1246949 24604202

9. Lee MN, Ye C, Villani AC, Raj T, Li W, Eisenhaure TM, et al. Common genetic variants modulate pathogen-sensing responses in human dendritic cells. Science. 2014;343(6175):1246980. doi: 10.1126/science.1246980 24604203

10. Dimas AS, Deutsch S, Stranger BE, Montgomery SB, Borel C, Attar-Cohen H, et al. Common regulatory variation impacts gene expression in a cell type-dependent manner. Science. 2009;325(5945):1246–50. doi: 10.1126/science.1174148 19644074

11. Naranbhai V, Fairfax BP, Makino S, Humburg P, Wong D, Ng E, et al. Genomic modulators of gene expression in human neutrophils. Nat Commun. 2015;6:7545. doi: 10.1038/ncomms8545 26151758

12. Kim-Hellmuth S, Bechheim M, Putz B, Mohammadi P, Nedelec Y, Giangreco N, et al. Genetic regulatory effects modified by immune activation contribute to autoimmune disease associations. Nat Commun. 2017;8(1):266. doi: 10.1038/s41467-017-00366-1 28814792

13. Liu Y, Beyer A, Aebersold R. On the Dependency of Cellular Protein Levels on mRNA Abundance. Cell. 2016;165(3):535–50. doi: 10.1016/j.cell.2016.03.014 27104977

14. Orru V, Steri M, Sole G, Sidore C, Virdis F, Dei M, et al. Genetic variants regulating immune cell levels in health and disease. Cell. 2013;155(1):242–56. doi: 10.1016/j.cell.2013.08.041 24074872

15. Roederer M, Quaye L, Mangino M, Beddall MH, Mahnke Y, Chattopadhyay P, et al. The genetic architecture of the human immune system: a bioresource for autoimmunity and disease pathogenesis. Cell. 2015;161(2):387–403. doi: 10.1016/j.cell.2015.02.046 25772697

16. Aguirre-Gamboa R, Joosten I, Urbano PCM, van der Molen RG, van Rijssen E, van Cranenbroek B, et al. Differential Effects of Environmental and Genetic Factors on T and B Cell Immune Traits. Cell Rep. 2016;17(9):2474–87. doi: 10.1016/j.celrep.2016.10.053 27818087

17. Patin E, Hasan M, Bergstedt J, Rouilly V, Libri V, Urrutia A, et al. Natural variation in the parameters of innate immune cells is preferentially driven by genetic factors. Nat Immunol. 2018;19(3):302–14. doi: 10.1038/s41590-018-0049-7 29476184

18. Sun BB, Maranville JC, Peters JE, Stacey D, Staley JR, Blackshaw J, et al. Genomic atlas of the human plasma proteome. Nature. 2018;558(7708):73–9. doi: 10.1038/s41586-018-0175-2 29875488

19. Yao C, Chen G, Song C, Keefe J, Mendelson M, Huan T, et al. Genome-wide mapping of plasma protein QTLs identifies putatively causal genes and pathways for cardiovascular disease. Nat Commun. 2018;9(1):3268. doi: 10.1038/s41467-018-05512-x 30111768

20. Lagou V, Garcia-Perez JE, Smets I, Van Horebeek L, Vandebergh M, Chen L, et al. Genetic Architecture of Adaptive Immune System Identifies Key Immune Regulators. Cell Rep. 2018;25(3):798–810 e6. doi: 10.1016/j.celrep.2018.09.048 30332657

21. Li Y, Oosting M, Deelen P, Ricano-Ponce I, Smeekens S, Jaeger M, et al. Inter-individual variability and genetic influences on cytokine responses to bacteria and fungi. Nat Med. 2016;22(8):952–60. doi: 10.1038/nm.4139 27376574

22. Li Y, Oosting M, Smeekens SP, Jaeger M, Aguirre-Gamboa R, Le KTT, et al. A Functional Genomics Approach to Understand Variation in Cytokine Production in Humans. Cell. 2016;167(4):1099–110 e14. doi: 10.1016/j.cell.2016.10.017 27814507

23. Bentham J, Morris DL, Graham DSC, Pinder CL, Tombleson P, Behrens TW, et al. Genetic association analyses implicate aberrant regulation of innate and adaptive immunity genes in the pathogenesis of systemic lupus erythematosus. Nat Genet. 2015;47(12):1457–64. doi: 10.1038/ng.3434 26502338

24. McGovern DP, Gardet A, Torkvist L, Goyette P, Essers J, Taylor KD, et al. Genome-wide association identifies multiple ulcerative colitis susceptibility loci. Nat Genet. 2010;42(4):332–7. doi: 10.1038/ng.549 20228799

25. Khor CC, Davila S, Breunis WB, Lee YC, Shimizu C, Wright VJ, et al. Genome-wide association study identifies FCGR2A as a susceptibility locus for Kawasaki disease. Nat Genet. 2011;43(12):1241–6. doi: 10.1038/ng.981 22081228

26. van der Pol W, van de Winkel JG. IgG receptor polymorphisms: risk factors for disease. Immunogenetics. 1998;48(3):222–32. doi: 10.1007/s002510050426 9683667

27. Westra HJ, Peters MJ, Esko T, Yaghootkar H, Schurmann C, Kettunen J, et al. Systematic identification of trans eQTLs as putative drivers of known disease associations. Nature Genetics. 2013;45(10):1238–U195. doi: 10.1038/ng.2756 24013639

28. International Multiple Sclerosis Genetics C, Wellcome Trust Case Control C, Sawcer S, Hellenthal G, Pirinen M, Spencer CC, et al. Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis. Nature. 2011;476(7359):214–9. doi: 10.1038/nature10251 21833088

29. De Jager PL, Jia X, Wang J, de Bakker PI, Ottoboni L, Aggarwal NT, et al. Meta-analysis of genome scans and replication identify CD6, IRF8 and TNFRSF1A as new multiple sclerosis susceptibility loci. Nat Genet. 2009;41(7):776–82. doi: 10.1038/ng.401 19525953

30. Moutsianas L, Jostins L, Beecham AH, Dilthey AT, Xifara DK, Ban M, et al. Class II HLA interactions modulate genetic risk for multiple sclerosis. Nat Genet. 2015;47(10):1107–13. doi: 10.1038/ng.3395 26343388

31. Lutfalla G, Holland SJ, Cinato E, Monneron D, Reboul J, Rogers NC, et al. Mutant U5A cells are complemented by an interferon-alpha beta receptor subunit generated by alternative processing of a new member of a cytokine receptor gene cluster. EMBO J. 1995;14(20):5100–8. 7588638

32. Fairfax BP, Makino S, Radhakrishnan J, Plant K, Leslie S, Dilthey A, et al. Genetics of gene expression in primary immune cells identifies cell type-specific master regulators and roles of HLA alleles. Nat Genet. 2012;44(5):502–10. doi: 10.1038/ng.2205 22446964

33. Lappalainen T, Sammeth M, Friedlander MR, t Hoen PA, Monlong J, Rivas MA, et al. Transcriptome and genome sequencing uncovers functional variation in humans. Nature. 2013;501(7468):506–11. doi: 10.1038/nature12531 24037378

34. Chick JM, Munger SC, Simecek P, Huttlin EL, Choi K, Gatti DM, et al. Defining the consequences of genetic variation on a proteome-wide scale. Nature. 2016;534(7608):500–5. doi: 10.1038/nature18270 27309819

35. Imgenberg-Kreuz J, Sandling JK, Almlöf JC, Nordlund J, Signer L, Norheim KB, et al. Genome-wide DNA methylation analysis in multiple tissues in primary Sjogren's syndrome reveals regulatory effects at interferon-induced genes. Ann Rheum Dis. 2016;75(11):2029–36. doi: 10.1136/annrheumdis-2015-208659 26857698

36. Olsson T, Barcellos LF, Alfredsson L. Interactions between genetic, lifestyle and environmental risk factors for multiple sclerosis. Nat Rev Neurol. 2017;13(1):25–36. doi: 10.1038/nrneurol.2016.187 27934854

37. Friese MA, Fugger L. Pathogenic CD8(+) T cells in multiple sclerosis. Ann Neurol. 2009;66(2):132–41. doi: 10.1002/ana.21744 19743458

38. Mustafa M, Vingsbo C, Olsson T, Ljungdahl A, Hojeberg B, Holmdahl R. The major histocompatibility complex influences myelin basic protein 63-88-induced T cell cytokine profile and experimental autoimmune encephalomyelitis. Eur J Immunol. 1993;23(12):3089–95. doi: 10.1002/eji.1830231207 7504988

39. Mustafa M, Vingsbo C, Olsson T, Issazadeh S, Ljungdahl A, Holmdahl R. Protective influences on experimental autoimmune encephalomyelitis by MHC class I and class II alleles. J Immunol. 1994;153(7):3337–44. 7522259

40. Issazadeh S, Kjellen P, Olsson T, Mustafa M, Holmdahl R. Major histocompatibility complex-controlled protective influences on experimental autoimmune encephalomyelitis are peptide specific. Eur J Immunol. 1997;27(6):1584–7. doi: 10.1002/eji.1830270640 9209515

41. Kular L, Liu Y, Ruhrmann S, Zheleznyakova G, Marabita F, Gomez-Cabrero D, et al. DNA methylation as a mediator of HLA-DRB1*15:01 and a protective variant in multiple sclerosis. Nat Commun. 2018;9(1):2397. doi: 10.1038/s41467-018-04732-5 29921915

42. Hauser SL, Bar-Or A, Comi G, Giovannoni G, Hartung HP, Hemmer B, et al. Ocrelizumab versus Interferon Beta-1a in Relapsing Multiple Sclerosis. N Engl J Med. 2017;376(3):221–34. doi: 10.1056/NEJMoa1601277 28002679

43. Palanichamy A, Apeltsin L, Kuo TC, Sirota M, Wang S, Pitts SJ, et al. Immunoglobulin class-switched B cells form an active immune axis between CNS and periphery in multiple sclerosis. Sci Transl Med. 2014;6(248):248ra106. doi: 10.1126/scitranslmed.3008930 25100740

44. Jelcic I, Al Nimer F, Wang J, Lentsch V, Planas R, Jelcic I, et al. Memory B Cells Activate Brain-Homing, Autoreactive CD4(+) T Cells in Multiple Sclerosis. Cell. 2018;175(1):85–100 e23. doi: 10.1016/j.cell.2018.08.011 30173916

45. Rizzo F, Giacomini E, Mechelli R, Buscarinu MC, Salvetti M, Severa M, et al. Interferon-beta therapy specifically reduces pathogenic memory B cells in multiple sclerosis patients by inducing a FAS-mediated apoptosis. Immunol Cell Biol. 2016;94(9):886–94. doi: 10.1038/icb.2016.55 27265253

46. Baker D, Marta M, Pryce G, Giovannoni G, Schmierer K. Memory B Cells are Major Targets for Effective Immunotherapy in Relapsing Multiple Sclerosis. EBioMedicine. 2017;16:41–50. doi: 10.1016/j.ebiom.2017.01.042 28161400

47. Sorensen TL, Tani M, Jensen J, Pierce V, Lucchinetti C, Folcik VA, et al. Expression of specific chemokines and chemokine receptors in the central nervous system of multiple sclerosis patients. J Clin Invest. 1999;103(6):807–15. doi: 10.1172/JCI5150 10079101

48. Huang J, Khademi M, Fugger L, Lindhe O, Novakova L, Axelsson M, et al. Inflammation-related plasma and CSF biomarkers for multiple sclerosis. Proc Natl Acad Sci U S A. 2020;117(23):12952–60. doi: 10.1073/pnas.1912839117 32457139

49. Claudinon J, Monier MN, Lamaze C. Interfering with interferon receptor sorting and trafficking: impact on signaling. Biochimie. 2007;89(6–7):735–43. doi: 10.1016/j.biochi.2007.03.014 17493737

50. Ivashkiv LB, Donlin LT. Regulation of type I interferon responses. Nat Rev Immunol. 2014;14(1):36–49. doi: 10.1038/nri3581 24362405

51. Couturier N, Bucciarelli F, Nurtdinov RN, Debouverie M, Lebrun-Frenay C, Defer G, et al. Tyrosine kinase 2 variant influences T lymphocyte polarization and multiple sclerosis susceptibility. Brain. 2011;134(Pt 3):693–703. doi: 10.1093/brain/awr010 21354972

52. Dendrou CA, Cortes A, Shipman L, Evans HG, Attfield KE, Jostins L, et al. Resolving TYK2 locus genotype-to-phenotype differences in autoimmunity. Sci Transl Med. 2016;8(363):363ra149. doi: 10.1126/scitranslmed.aag1974 27807284

53. Isaksson M, Ardesjö B, Rönnblom L, Kämpe O, Lassmann H, Eloranta ML, et al. Plasmacytoid DC promote priming of autoimmune Th17 cells and EAE. Eur J Immunol. 2009;39(10):2925–35. doi: 10.1002/eji.200839179 19637225

54. Prinz M, Schmidt H, Mildner A, Knobeloch KP, Hanisch UK, Raasch J, et al. Distinct and nonredundant in vivo functions of IFNAR on myeloid cells limit autoimmunity in the central nervous system. Immunity. 2008;28(5):675–86. doi: 10.1016/j.immuni.2008.03.011 18424188

55. Teige I, Treschow A, Teige A, Mattsson R, Navikas V, Leanderson T, et al. IFN-beta gene deletion leads to augmented and chronic demyelinating experimental autoimmune encephalomyelitis. J Immunol. 2003;170(9):4776–84. doi: 10.4049/jimmunol.170.9.4776 12707359

56. Messer G, Zemmour J, Orr HT, Parham P, Weiss EH, Girdlestone J. HLA-J, a second inactivated class I HLA gene related to HLA-G and HLA-A. Implications for the evolution of the HLA-A-related genes. J Immunol. 1992;148(12):4043–53. 1602142

57. Cheetham SW, Faulkner GJ, Dinger ME. Overcoming challenges and dogmas to understand the functions of pseudogenes. Nat Rev Genet. 2019. doi: 10.1038/s41576-019-0196-1 31848477

58. Rapicavoli NA, Qu K, Zhang JJ, Mikhail M, Laberge RM, Chang HY. A mammalian pseudogene lncRNA at the interface of inflammation and anti-inflammatory therapeutics. Elife. 2013;2. doi: 10.7554/eLife.00762 23898399

59. Chiang JJ, Sparrer KMJ, van Gent M, Lassig C, Huang T, Osterrieder N, et al. Viral unmasking of cellular 5S rRNA pseudogene transcripts induces RIG-I-mediated immunity. Nat Immunol. 2018;19(1):53–62. doi: 10.1038/s41590-017-0005-y 29180807

60. Gu S, Fellerhoff B, Muller N, Laumbacher B, Wank R. Paradoxical downregulation of HLA-A expression by IFNgamma associated with schizophrenia and noncoding genes. Immunobiology. 2013;218(5):738–44. doi: 10.1016/j.imbio.2012.08.275 23083632

61. Friese MA, Jakobsen KB, Friis L, Etzensperger R, Craner MJ, McMahon RM, et al. Opposing effects of HLA class I molecules in tuning autoreactive CD8+ T cells in multiple sclerosis. Nat Med. 2008;14(11):1227–35. doi: 10.1038/nm.1881 18953350

62. Sundqvist E, Sundström P, Linden M, Hedström AK, Aloisi F, Hillert J, et al. Epstein-Barr virus and multiple sclerosis: interaction with HLA. Genes Immun. 2012;13(1):14–20. doi: 10.1038/gene.2011.42 21776012

63. Polman CH, Reingold SC, Banwell B, Clanet M, Cohen JA, Filippi M, et al. Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Ann Neurol. 2011;69(2):292–302. doi: 10.1002/ana.22366 21387374

64. Hagberg N, Joelsson M, Leonard D, Reid S, Eloranta ML, Mo J, et al. The STAT4 SLE risk allele rs7574865[T] is associated with increased IL-12-induced IFN-gamma production in T cells from patients with SLE. Ann Rheum Dis. 2018;77(7):1070–7. doi: 10.1136/annrheumdis-2017-212794 29475858

65. Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006;38(8):904–9. doi: 10.1038/ng1847 16862161

66. Price AL, Weale ME, Patterson N, Myers SR, Need AC, Shianna KV, et al. Long-range LD can confound genome scans in admixed populations. Am J Hum Genet. 2008;83(1):132–5; author reply 5–9. doi: 10.1016/j.ajhg.2008.06.005 18606306

67. McCarthy S, Das S, Kretzschmar W, Delaneau O, Wood AR, Teumer A, et al. A reference panel of 64,976 haplotypes for genotype imputation. Nat Genet. 2016;48(10):1279–83. doi: 10.1038/ng.3643 27548312

68. Loh PR, Danecek P, Palamara PF, Fuchsberger C, Y AR, H KF, et al. Reference-based phasing using the Haplotype Reference Consortium panel. Nat Genet. 2016;48(11):1443–8. doi: 10.1038/ng.3679 27694958

69. Zheng X, Shen J, Cox C, Wakefield JC, Ehm MG, Nelson MR, et al. HIBAG—HLA genotype imputation with attribute bagging. Pharmacogenomics J. 2014;14(2):192–200. doi: 10.1038/tpj.2013.18 23712092

70. Stekhoven DJ, Buhlmann P. MissForest—non-parametric missing value imputation for mixed-type data. Bioinformatics. 2012;28(1):112–8. doi: 10.1093/bioinformatics/btr597 22039212

71. Chang CC, Chow CC, Tellier LC, Vattikuti S, Purcell SM, Lee JJ. Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience. 2015;4:7. doi: 10.1186/s13742-015-0047-8 25722852

72. Serana F, Sottini A, Ghidini C, Zanotti C, Capra R, Cordioli C, et al. Modulation of IFNAR1 mRNA expression in multiple sclerosis patients. J Neuroimmunol. 2008;197(1):54–62. doi: 10.1016/j.jneuroim.2008.03.019 18482773

73. R Core Team. R: A language and environment for statistical computing: R Foundation for Statistical Computing; 2018 [Available from: https://www.R-project.org/.


Článek vyšel v časopise

PLOS Genetics


2020 Číslo 10
Nejčtenější tento týden
Nejčtenější v tomto čísle
Kurzy Podcasty Doporučená témata Časopisy
Přihlášení
Zapomenuté heslo

Zadejte e-mailovou adresu, se kterou jste vytvářel(a) účet, budou Vám na ni zaslány informace k nastavení nového hesla.

Přihlášení

Nemáte účet?  Registrujte se

#ADS_BOTTOM_SCRIPTS#