Variants of genes encoding TNF receptors and ligands and proteins regulating TNF activation in familial multiple sclerosis

Abstract Introduction Numerous genetic variants have been associated with susceptibility to multiple sclerosis (MS). Variants located in genes involved in specific pathways, such as those affecting TNF‐α, can contribute to the risk of MS. The purpose of this study was to determine whether variants of these genes are associated with greater risk of MS. Methods We used whole‐exome sequencing to study genes coding for TNF‐α receptors and ligands, and proteins promoting TNF‐α expression in 116 individuals from 19 families including at least two MS patients. We compared patients with MS, patients with other autoimmune diseases, and healthy individuals. Results Greater polymorphism was observed in several genes in families with familial MS compared to the general population; this may reflect greater susceptibility to autoimmune diseases. Pedigree analysis also revealed that LT‐α variants rs1041981 and rs2229094 and LT‐β variant rs4647197 were associated with MS and that LT‐β variant rs4647183 was associated with other autoimmune diseases. The association between autoimmune disease and TNFAIP2 variant rs1132339 is particularly noteworthy, as is the fact that TNFAIP6 variant rs1046668 appears to follow a recessive inheritance pattern. Conclusions Our findings support the idea that the risk of familial MS is associated with variants of signaling pathways, including those involving TNF‐α.


| INTRODUC TI ON
TNF-α receptors are transmembrane proteins that were discovered in 1984 to be associated with apoptosis, due to the death domain located on the intracellular part of the protein. To date, a total of 29 receptors and 19 ligands have been described in humans. As well as in apoptosis, this protein superfamily is involved in cell differentiation and proliferation through the activation of a range of signaling pathways, largely associated with mechanisms involved in immunity. These processes also involve proteins that promote | 1179

TORRE-FUENTES ET al.
TNF-α expression, 10 of which have been described. On account of their involvement in immune mechanisms, many of these genes are thought to play a role in multiple sclerosis (MS). We analyze the presence of genomic variants of these genes in the families of patients with familial MS.

| Study sample
Our sample comprised 116 members of 19 families including at least two members diagnosed with MS according to the 2010 McDonald criteria. 1 We studied the exomes of all individuals, establishing three groups: patients with MS, patients with other autoimmune dis-

eases (AID) included in the American Autoimmune Related Diseases
Association's list of autoimmune diseases, and healthy individuals.
Families were classified as type A, in which all patients with MS belonged to the same generation, or type B, in which more than one generation was affected by MS, as described in a previous study. 2 For all participants, we gathered data on demographic variables, personal history, clinical variables, age of onset, progression time (defined as the time from the presentation of the first neurological symptom associated with MS to the date of inclusion in the study), and clinical form of MS. This study was approved by our center's Ethics Committee. All participants included in the study gave written informed consent. cording to specific quality criteria and analyzed using the Variant Caller tool to identify nucleotide variations as compared to the reference genome. All identified variants were annotated using the latest available version of the Ion Reporter™ software (Thermo Fisher Scientific, Inc). We aimed to identify single-nucleotide variants and indels located within exons and splice sites of the genes analyzed, which would cause protein alterations with a minimum variant allele frequency (VAF) of 40%.

| Analysis of variants
Variants were described using the dbSNP record, when available.
The results of the descriptive analysis of variants are expressed as absolute frequencies (%), means (standard deviation), or medians (interquartile range). The More Powerful Quasi-Likelihood Score test (MQLS-XM) 5,6 was used to test associations between patients and controls; this test is used with samples including related individuals. Allele frequencies were analyzed to test for deviations from Hardy-Weinberg equilibrium. The Bonferroni correction for multiple comparisons was applied. We analyzed the variants detected in each group, accounting for family type. We also calculated odds ratios (OR) with 95% confidence intervals (CI). We used the chi-square test to compare frequencies between groups where the MQLS-XM test could not be applied. These comparisons were adjusted with the Fisher test or Yates correction when necessary (Table S2) Bioinformatics filtering also detected possible variants not previously registered on dbSNP.

| Description of the sample
Our sample of 116 participants from 19 families included 43 patients with MS (37.1%), 16 patients with AIDs other than MS (13.8%), and 57 healthy individuals (49.1%). Table 1 shows the characteristics of the cohort.

| Analysis of variants of TNF genes
All 164 SNPs followed Hardy-Weinberg equilibrium, where P-values below the Bonferroni-adjusted threshold (P < .0005) would indicate significant deviations. MAFs were consulted in the gnomAD browser.
We used the More Powerful Quasi-Likelihood Score test (MQLS; MQLS-XM), which accounts for relatedness between subjects using corrected variance. We observed no significant differences in allele frequencies between patients with MS and unaffected individuals.
We used the chi-square test to analyze differences between type A and type B families, applying Fisher test or Yates correction when necessary.
Information on all SNPs found in our cohort is listed in Table S2.
Although we did not observe significant differences in allele frequencies between patients with MS and unaffected individuals, some One missense variant of TNFRSF10B, rs1047266, was found in a higher percentage of patients with MS (14.0%) than unaffected individuals (8.2%; P = .013; MQLS). We also found trends to significance for variants rs2234167 of TNFRSF14, rs35041805 of TNFRSF19, and rs1046668 of TNFAIP6, with higher frequencies for patients with MS than for individuals without the disease.

| Frequency of variants identified in our cohort compared to their minor allele frequency
We used the chi-square test to analyze differences in allele frequencies between our cohort and the general population. The following variants of genes encoding TNF receptors showed greater frequencies in our cohort than the MAF: TNFRSF10B variant rs13265018, TNFRSF8 variant rs1763642 was much less frequent in families with MS (Table S3).
In our cohort, two variants of genes encoding TNF ligands were more frequent than would be expected from MAF data: LT-α variant rs2229094 and TNFSF10 variant rs16845759 (Table S3).

| Pedigree analysis of nonsynonymous exonic variants
Some families met criteria for establishing an association between MS and a variant of a gene encoding a TNF receptor. The genes meeting these criteria were TNFRSF10A and TNFRSF19 ( Figure 1).
Furthermore, the LT-α and LT-β TNF ligand-coding genes ( Figure 2) and the LITAF and TNFAIP2 genes, regulating TNF expression, met these criteria in the pedigree analysis ( Figure 3).

| Homozygosity of nonsynonymous exonic variants identified in the cohort
We used the chi-square test to compare differences in the percent-

| D ISCUSS I ON
Our study focused on one of the immune pathways related to MS.
Some of these genes have previously been related to the risk of the disease. 8 The search for genetic variations in any of these genes could lead to abnormal behavior of the pathway, potentially affecting susceptibility to MS.
Although none of the variants showed statistically significant differences in the comparison of frequencies between individuals with and without MS after correction for multiple comparisons, some variants showed a trend toward statistical significance.
We detected a rare TNFRSF19 variant, rs35041805, which has not previously been associated with MS nor other diseases. This variant and another polymorphism of this gene, rs9550987, were segregated with the disease in two families and in one family, respectively. Previous studies have shown that this gene is downregulated in B cells from siblings with MS. 9 TROY, the protein encoded by this gene, has also been linked to the inhibitory effects of myelin inhibitors, 10 and its expression has been shown to be upregulated in brain tissues of patients with MS. 11 The LT-β gene also presented a rare variant, rs4647187, which may be associated with the disease. LT-β, located in the 6p21.33 region, is part of the major histocompatibility complex (MHC) and has been associated with rheumatoid arthritis. 12 We may therefore hypothesize that rs4647187 contributes to the risk of autoimmune disease in these families.
Moreover, the LT-α gene has been linked to MS. 8,13 We identified two families displaying an association between two variants of this gene (rs1041981 and rs2229094) and MS; this association merits deeper analysis. Variant rs2229094 is particularly interesting because its frequency in the cohort was greater than we would expect given its MAF; we may therefore hypothesize that it increases the risk of autoimmune disease in these families. In addition, an increased LT-α protein response has been reported in patients with MS, which suggests that LT-α plays a role in the pathogenesis of the disease. 14 Mutations in LT-α could impact this role and the signaling pathway.
TNFSF10 encodes the TRAIL protein, which may be involved in MS pathogenesis, according to some studies. 15,16 There are four different types of TRAIL surface receptor. The genes coding for the TRAIL receptors are TNFRSF10A, TNFRSF10B, TNFRSF10C, and TNFRSF10D. The rs1047266 polymorphism of TNFRSF10B was found in a higher percentage of patients with MS than individuals without the disease. This polymorphism has not previously been associated with MS. However, TRAIL has been proposed as a marker of response to treatment with interferon. 17,18 In a study of 509 patients with MS receiving this treatment, no association with interferon beta was found. 19 While we did not observe significant differences, we detected two other variants of one of the TRAIL receptors, TNFRSF10A, rs20557 and rs17620, which segregated with the disease. This association should be further studied in other samples to evaluate its potential relationship with MS.
Interestingly, in one family presenting the rs1132339 variant of The above-mentioned polymorphism should be studied in the future.

| CON CLUS IONS
Genome-wide association studies and genome studies targeting specific mutations have identified numerous genetic variants associated with susceptibility to MS, but each of these explains only a very small percentage of the risk of developing MS. However, these variants are located in genes involved in specific pathways, which suggests that risk may reside in alterations to these pathways, rather than in specific genes. 23 It is evident that the relationship between TNF-α, its receptors, the ligands that bind to them, and proteins that promote its activation, as well as the TNF-α receptor-associated factors analyzed elsewhere, is associated with autoimmunity, comprising several signaling pathways involved in AIDs. This gives rise to the need to study potential associations between MS and variants of the genes encoding these proteins.
We analyzed these genes in the context of familial MS, identifying several associations which merit consideration in future studies. It should be noted that we address only familial cases of MS; these genes may not represent the same risk for sporadic cases.
We have not detected many variants previously associated with MS. 8,24 This may be explained by the size of our sample and a lack of statistical power. However, we found several variants with greater frequency in MS cases, suggesting a potential role in the disease. We should emphasize that greater polymorphism was observed in many genes in the families of patients with familial MS than in the general population; this may reflect greater susceptibility to autoimmune disease. Pedigree analysis also showed that the rs1041981 and rs2229094 variants of the LT-α and rs4647197 of LT-β genes were associated with MS and that LT-β variant rs4647183 was associated with other autoimmune diseases. The association between TNFAIP2 variant rs1132339 and AIDs, and the recessive inheritance pattern of TNFAIP6 variant rs1046668 are particularly noteworthy. While these variants require further study, including functional studies analyzing their pathogenicity, 25 these findings support the idea that the risk of familial MS may be associated with variants affecting signaling pathways, including those involving TNF-α.

ACK N OWLED G M ENTS
We thank our patients and their families for the information and DNA samples provided for our study. We are also grateful to the Spanish Society of Neurology's translation department for their assistance with the manuscript.

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to declare.

DATA AVA I L A B I L I T Y S TAT E M E N T
The datasets that support the findings of this study are available in the European Genome-Phenome Archive repository at https:// ega-archi ve.org/datas ets/EGAD00001005952, reference number EGAD00001005952.