Skip to main content
NCBI home page
Brain logoLink to Brain
. 2015 Jan 23;138(3):632–643. doi: 10.1093/brain/awu405

Genetic variants are major determinants of CSF antibody levels in multiple sclerosis

An Goris 1,*,✉, Ine Pauwels 1,*, Marte W Gustavsen 2,3, Brechtje van Son 4, Kelly Hilven 1, Steffan D Bos 2,3, Elisabeth Gulowsen Celius 2, Pål Berg-Hansen 2,3, Jan Aarseth 5, Kjell-Morten Myhr 5,6, Sandra D’Alfonso 7,8, Nadia Barizzone 7, Maurizio A Leone 8,9, Filippo Martinelli Boneschi 10,11, Melissa Sorosina 11, Giuseppe Liberatore 10,11, Ingrid Kockum 12, Tomas Olsson 12, Jan Hillert 13, Lars Alfredsson 14, Sahl Khalid Bedri 13, Bernhard Hemmer 15,16, Dorothea Buck 15, Achim Berthele 15, Benjamin Knier 15, Viola Biberacher 15, Vincent van Pesch 17, Christian Sindic 17, Annette Bang Oturai 18, Helle Bach Søndergaard 18, Finn Sellebjerg 18, Poul Erik H Jensen 18, Manuel Comabella 19, Xavier Montalban 19, Jennifer Pérez-Boza 19, Sunny Malhotra 19, Jeannette Lechner-Scott 20, Simon Broadley 21, Mark Slee 22, Bruce Taylor 23, Allan G Kermode 24,25, Pierre-Antoine Gourraud 26; International Multiple Sclerosis Genetics Consortium22,26,#, Stephen J Sawcer 27, Bettina Kullle Andreassen 28, Bénédicte Dubois 1,4, Hanne F Harbo 2,3
PMCID: PMC4408440  PMID: 25616667

The presence of antibodies in the CSF is a hallmark of multiple sclerosis, but reasons for patient-to-patient differences in antibody levels remain largely unknown. In a genome-wide association screen of almost 7000 patients from 9 countries, Goris et al. identify six genetic variants that influence antibody levels and disease presentation.

Keywords: multiple sclerosis, CSF, oligoclonal bands, immunoglobulin, genetics

Abstract

Immunological hallmarks of multiple sclerosis include the production of antibodies in the central nervous system, expressed as presence of oligoclonal bands and/or an increased immunoglobulin G index—the level of immunoglobulin G in the cerebrospinal fluid compared to serum. However, the underlying differences between oligoclonal band-positive and -negative patients with multiple sclerosis and reasons for variability in immunoglobulin G index are not known. To identify genetic factors influencing the variation in the antibody levels in the cerebrospinal fluid in multiple sclerosis, we have performed a genome-wide association screen in patients collected from nine countries for two traits, presence or absence of oligoclonal bands (n = 3026) and immunoglobulin G index levels (n = 938), followed by a replication in 3891 additional patients. We replicate previously suggested association signals for oligoclonal band status in the major histocompatibility complex region for the rs9271640*A-rs6457617*G haplotype, correlated with HLA-DRB1*1501, and rs34083746*G, correlated with HLA-DQA1*0301 (P comparing two haplotypes = 8.88 × 10−16). Furthermore, we identify a novel association signal of rs9807334, near the ELAC1/SMAD4 genes, for oligoclonal band status (P = 8.45 × 10−7). The previously reported association of the immunoglobulin heavy chain locus with immunoglobulin G index reaches strong evidence for association in this data set (P = 3.79 × 10−37). We identify two novel associations in the major histocompatibility complex region with immunoglobulin G index: the rs9271640*A-rs6457617*G haplotype (P = 1.59 × 10−22), shared with oligoclonal band status, and an additional independent effect of rs6457617*G (P = 3.68 × 10−6). Variants identified in this study account for up to 2-fold differences in the odds of being oligoclonal band positive and 7.75% of the variation in immunoglobulin G index. Both traits are associated with clinical features of disease such as female gender, age at onset and severity. This is the largest study population so far investigated for the genetic influence on antibody levels in the cerebrospinal fluid in multiple sclerosis, including 6950 patients. We confirm that genetic factors underlie these antibody levels and identify both the major histocompatibility complex and immunoglobulin heavy chain region as major determinants.

Introduction

Multiple sclerosis is a neurological disease characterized by inflammation, demyelination and axonal degeneration, and is an important cause of disability in young adults (Compston and Coles, 2008). The aetiology is not known, but both genetic and environmental factors influence disease susceptibility (Compston and Coles, 2008). An association between multiple sclerosis and the human leukocyte antigen (HLA) genes in the major histocompatibility complex (MHC) region was identified early on (Jersild et al., 1972), and was later refined to four independent association signals of which HLA-DRB1*15:01 is the strongest (The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011). Recently, knowledge of non-HLA associations in multiple sclerosis has increased extensively and immunologically relevant genes are overrepresented among those mapping close to the identified risk variants (The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011; The International Multiple Sclerosis Genetics Consortium, 2013).

An immunological hallmark in multiple sclerosis is the finding of oligoclonal bands (OCBs) and/or increased immunoglobulin G (IgG) index in the CSF (Stangel et al., 2013). OCBs are reported to be observed in 90–95% of patients in Northern Europe, and are composed predominantly of IgG. The IgG index contrasts the CSF/serum IgG ratio to the CSF/serum albumin ratio (Link and Huang, 2006). There are strong indications that antibody levels in the CSF are influenced by genetic factors, as CSF abnormalities are seen in 19% of unaffected siblings of patients with multiple sclerosis as opposed to 4% of healthy unrelated individuals (Haghighi et al., 2000). Also, the OCB positive rates, as well as the IgG index are reported to correlate with ethnicity (Fukazawa et al., 1998; Kikuchi et al., 2003; Rinker et al., 2007; Lechner-Scott et al., 2011; Yoshimura et al., 2014), although not observed in all studies (Berg-Hansen et al., 2013).

The role of the HLA loci in determining OCB status has been highlighted previously (Fukazawa et al., 1998; Kikuchi et al., 2003; Imrell et al., 2006; Idiman et al., 2009; Wu et al., 2009; Romero-Pinel et al., 2011; Leone et al., 2013; Mero et al., 2013; Yoshimura et al., 2014). Most recently, large study populations from Scandinavia and Italy have shown HLA-DRB1*15∶01 to be associated with OCB-positive and HLA-DRB1*04∶04 with OCB-negative multiple sclerosis (Leone et al., 2013; Mero et al., 2013). Furthermore, the immunoglobulin heavy chain (IGHC) region was recently reported to be highly correlated with IgG index in German and Belgian patients with multiple sclerosis (Buck et al., 2012).

In the present study we aim to further determine the genetic impact on IgG levels in the CSF in multiple sclerosis. We combine genome-wide single nucleotide polymorphism (SNP) data (The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011) with available CSF data in a meta-analysis of 3059 patients with multiple sclerosis and replicate top-hits from the screening phase in 3891 additional samples. The combined data set of 6950 patients from nine countries is the largest study population so far to investigate both clinical and genetic factors associated with OCB status and IgG index and we replicate three previously described associations and identify three new genetic differences underlying CSF phenotypes.

Materials and methods

Study populations

The 3059 multiple sclerosis samples included in the screening phase of this study are a subset of the samples that were previously used in a genome-wide association study on multiple sclerosis susceptibility (The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011). Eight countries provided OCB status (positive or negative) for 3026 patients in the screening phase, while five countries provided IgG index for 938 patients (Table 1). In the replication phase, 3891 additional multiple sclerosis samples (3842 with OCB status and 2188 with IgG index) were provided by eight countries. The combined screening and replication cohort included OCB status on 6868 patients and IgG index on 3126 patients from the following countries: Australia, Belgium, Denmark, Germany, Italy, Norway, Spain, Sweden and USA (Table 1). For 3044 patients, both OCB status and IgG index were available. For the genetic analyses, 3478 patients with OCB status and 2072 patients with IgG index survived genetic quality control in the replication phase, leading to a final sample size of 6504 for OCB status and 3010 for IgG index in the combined genetic analyses.

Table 1.

Demographic characterization of the multiple sclerosis study populations

Country Female:male ratio PP (%) Age at onset, years, mean (range) MSSS mean (range) OCB (n) OCB+ (%) IgG (n) IgG index mean (range) IgG+ (%) Disease duration at lumbar puncture, years, median (range)
SCREEN (n = 3059) 2.28 13 33 (9–60) 4.43 (0.03–9.98) 3026 88 938 1.05 (0.29–7.27) 64 1 (−6–49)
Belgium (n = 358) 1.93 13 34 (9–60) 5.31 (0.03–9.98) 356 89 165 1.12 (0.43–5.36) 68 2 (−2–46)
Denmark (n = 197) 2.28 2 32 (12–59) 4.30 (0.11–9.76) 189 90 63 1.24 (0.42–7.27) 87 0 (0–39)
Italy (n = 565) 1.80 30 33 (10–60) 4.47 (0.04–9.96) 564 90 133 0.90 (0.34–3.19) 46 1 (0–41)
Norway (n = 827) 2.56 12 34 (10–59) 5.09 (0.26–9.86) 827 89 NA NA NA 3 (0–41)
Spain (n = 62) 1.42 41 33 (10–60) 5.57 (0.21–9.83) 62 79 NA NA NA 4 (0–20)
Sweden (n = 443) 2.52 4 33 (12–57) 3.01 (0.04–9.96) 439 88 355 1.00 (0.3–5) 61 1 (–6–49)
Germany (n = 444) 2.67 6 35 (15–58) 4.70 (0.11–9.84) 426 90 222 1.06 (0.29–3.94) 73 4 (0–19)
USA (n = 163) 2.70 10 34 (13–60) 5.54 (0.74–9.88) 163 75 NA NA NA 3 (0–36)
REP (n = 3891) 2.27 10 34 (2–72) 3.95 (0.03–9.96) 3842 87 2188 1.03 (0.06–6.20) 64 1 (−6–48)
Australia (n = 343) 3.05 10 35 (2–64) 4.08 (0.05–9.88) 343 71 127 0.98 (0.06–3.1) 61 1 (0–48)
Belgium (n = 359) 1.92 12 35 (11–72) 4.16 (0.03–9.94) 356 91 243 1.09 (0.37–4.45) 69 1 (0–44)
Denmark (n = 274) 1.98 6 32 (14–61) 4.15 (0.04–9.63) 274 93 274 1.27 (0.35–6.20) 77 2 (−1–32)
Italy (n = 759) 2.15 8 31 (7–66) 3.25 (0.09–9.86) 759 88 512 0.89 (0.24–2.97) 55 1 (0–35)
Norway (n = 1127) 2.49 13 34 (12–68) 4.47 (0.13–9.95) 1127 89 85 1.26 (0.4–5.5) 82 3 (0–38)
Spain (n = 242) 1.81 18 32 (13–65) 4.39 (0.15–9.96) 242 81 161 1.10 (0.4–5.53) 65 1 (0–40)
Sweden (n = 380) 2.65 12 38 (12–68) 4.97 (0.10–9.87) 380 90 379 1.03 (0.3–3.8) 68 0 (−6–44)
Germany (n = 407) 1.99 3 33 (12–64) 3.40 (0.10–9.77) 361 91 407 0.94 (0.37–5.2) 55 0 (−1–25)
Open in a new tab

PP = primary progressive multiple sclerosis; MSSS = Multiple Sclerosis Severity Score; REP = replication; OCB+ = oligoclonal band positive status, IgG+ = IgG index>0.7.

CSF analyses

All patients included in this study fulfilled Poser or McDonald multiple sclerosis criteria (Poser et al., 1983; McDonald et al., 2001) or were diagnosed as clinically isolated syndrome [n = 7 (0.2%) in screening phase (Germany); n = 138 (3.5%) in replication phase (Germany)], based on the combination of (i) a clinical symptom being typical and suggestive for multiple sclerosis; (ii) MRI of the brain and in most cases also of the spinal cord demonstrating typical multiple sclerosis lesions fulfilling the criteria of dissemination in space (Barkhof criteria or Swanton criteria); and (iii) CSF parameters. In most countries recruiting samples to this study, lumbar puncture is done routinely as part of the diagnostic process and is seldom repeated unless the diagnosis is unclear. OCB status is considered positive when more than one OCB was seen in the CSF that was not present in the serum. IgG index is considered positive for values >0.7. In this study the IgG index value was used in quantitative trait statistical analyses after transforming the IgG index ratio by taking the logarithm (base 2) (log2 IgG).

Clinical analyses

The clinical parameters gender, age at onset, Multiple Sclerosis Severity Score (Roxburgh et al., 2005), disease course and disease duration at lumbar puncture were included in the clinical analyses. Statistical analyses of CSF phenotypes related to demographic and clinical parameters were done in R 2.14.1 (www.r-project.org) using linear regression for IgG index and logistic regression for the binary OCB status. We included covariates in the analyses as indicated.

Screening phase

Quality control and analysis of genetic data provided in the multiple sclerosis genome-wide association study (The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011) were performed with Plink v1.07 (Purcell et al., 2007). For OCB status, we used a qualitative trait analysis in a logistic regression with binary outcome (OCB pos/neg). For IgG index, we applied a quantitative trait analysis in a linear regression with allele dosage as independent variable and IgG index as dependent outcome. In the screening phase, gender and the seven main principal components previously determined in this data set (The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011) were associated with IgG index and hence added as covariates for IgG index analyses. Evidence of association was tested in samples from each country individually and combined in a fixed-effects meta-analysis (assuming a qualitative trait for OCB and quantitative trait for IgG) over all countries. P-values < 5 × 10−8 and <10−4.5 in the screening phase were considered as genome-wide significant and suggestive evidence, respectively, as applied previously (The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011). For each lead SNP a conditional analysis of the surrounding 3 Mb region was performed, and SNPs reaching P < 10−4 were considered suggestive in this analysis.

Power for OCB analysis was determined mainly by the typical frequency of OCB-negative status in a multiple sclerosis study population. In the screening phase, we had 80% power to detect suggestive evidence (P < 10−4.5) for variants with a minor allele frequency of 0.20 and an odds ratio (OR) of 1.6. Power was 80% for variants explaining 2.5% of the variation in the distribution of IgG index seen in patients with multiple sclerosis.

Replication phase

Forty-two SNPs were selected for replication. Of these, 38 SNPs were brought forward to replication based on the results from the screening phase; 32 lead SNPs reaching P < 10−4.5, an additional proxy marker for rs6457617 (rs9275224 with r2 = 1), and five SNPs with conditional association signals of P < 10−4. Additionally, we added two SNPs that were previously suggested to be associated with OCB status (Leone et al., 2013; Mero et al., 2013) and two SNPs reported to be associated to autoantibody presence in other tissues (Ferreira et al., 2010; Plagnol et al., 2011) and showing association or a trend for association with IgG index in our study. The SNPs were genotyped using Sequenom MassARRAY® iPLEX® technology (Sequenom Inc.). Sequenom MassARRAY® Designer software v3.1 was used to design primers and extension probes. As part of validating the design, a test set of genotyping data from 86 individuals generated with Illumina platforms was included. Seven SNPs were replaced by proxies with moderate to high linkage disequilibrium (r2 > 0.63). For two SNPs, no assay could be designed and for two others genotyping assays failed, therefore a total of 38 SNPs were successfully included in the replication analysis (i.e. 28 lead SNPs and one synonym, five conditional SNPs and four SNPs selected from previous studies) (Supplementary material). Seventy-five nanograms of genomic DNA was used in 5 μl reactions in 384-well plates. The amplified resin-treated DNA was spotted with a Sequenom MassARRAY® Nanodispenser (Sequenom Inc.) on a SpectroCHIP Array. The SpectroCHIP Arrays were analysed using a Sequenom MALDI-TOF mass spectrometer (Sequenom Inc.). Genotyping calls were automatically generated using the MassARRAY® TyperAnalyzer software v4.0 (Sequenom Inc.) and were validated by manual review of the raw mass spectra scatter plots.

Genotyping quality control was performed in samples grouped per country. Samples with >4/38 missing genotypes (<89.5% sample success rate) and SNPs with <95% genotyping success rate were excluded from further analysis. No SNPs deviated from Hardy-Weinberg equilibrium (P < 10−4), except for a known multiple sclerosis susceptibility SNP in the MHC region in the Norwegian population. Analysis was performed per country with a linear model including gender as covariate for IgG and a logistic model for OCB, followed by a fixed-effects meta-analysis over all countries. An effect was considered replicated when reaching P < 0.05 in the replication phase.

Combined analyses

Analysis was performed by a fixed-effects meta-analysis over all cohorts (screening and replication cohorts per country as described previously). The percentage of the variance in IgG index explained by variants was calculated by subtracting adjusted r2 from a full model with that from the baseline linear model in R. Evidence for interaction between variants, defined as deviation from a multiplicative model, was investigated in a linear (IgG index) or logistic (OCB status) regression in R.

Major histocompatibility complex analyses

In the screening phase HLA-A, -B, -C, -DRB1, -DQA1 and -DQB1 genotypes were imputed from SNP data as described previously (Dilthey et al., 2011; The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011). The most likely imputed alleles were used as input for testing of association of common HLA alleles with CSF phenotypes using the same regression approach as described for SNPs above. Independent effects of SNPs/haplotypes in the MHC region were investigated with a likelihood ratio test for OCB status or F-test for IgG index comparing a null model and alternative model for phased haplotypes with a minor haplotype frequency of 0.05 including country, seven principal components (screening phase only) and gender as covariates (–chap option in Plink v1.07). The most likely model was obtained by starting with a null model with equal effects for all haplotypes consisting of known and novel replicated MHC signals and allowing separate effects for individual haplotypes, in order of significance, as long as this improved the model with nominal significance.

Multiple sclerosis genetic burden

In the screening phase, a genetic risk score, multiple sclerosis genetic burden, was calculated on the basis of the number of risk alleles weighted by their effect on multiple sclerosis risk according to the method described previously (Gourraud et al., 2011; Harbo et al., 2014). Two scores were calculated, both including the 57 risk SNPs and either including or excluding alleles imputed as above for the four HLA effects (HLA-DRB1*15:01, DRB1*03:01, DRB1*13:03 and HLA–A*02:01) established in our previous genome-wide association study on susceptibility including this data set (The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011). A logistic regression analysis with country as covariate was performed for OCB status, whereas a linear regression analysis including country, gender and seven principal components was performed for IgG index in R.

Results

Clinical and CSF analyses

The main demographic and clinical characteristics of patients included in the screening and replication phase, stratified by country of origin, are described in Table 1. Among 6868 patients with multiple sclerosis with known OCB status, 6033 (88%) were OCB-positive. The IgG index level in patients recruited from different countries is shown in Supplementary Fig. 1. An elevated IgG index was found in 1996 of 3126 patients with known IgG index (64%). IgG index was on average lower in Italian and higher in Danish and Norwegian patients. For 3044 patients, both OCB status and IgG index were available. Increased IgG index was highly correlated with OCB status (P < 2 × 10−16) in the combined data set (Table 2 and Supplementary Fig. 2). Overall, 62% of the patients with multiple sclerosis were positive and 10% were negative for both OCB and IgG index (Supplementary Table 1). On average, 26% of the patients were OCB-positive but did not have an increased IgG index. An increased IgG index in OCB-negative patients with multiple sclerosis was rare (2%).

Table 2.

Correlation of CSF measures with demographic and clinical variables of the included multiple sclerosis patients

Screening
 Replication
 Combined
Log IgG versus n Fold changee P n Fold changee P n Fold change IgG index (95% CI)e P
Gender (female)a 938 1.136 1.4 × 10−4 2188 1.115 7.6 × 10−7 3126 1.124 (1.085–1.165) 1.9 × 10−10
OCB-status (positive)b 905 1.637 <2 × 10−16 2139 1.640 <2 × 10−16 3044 1.639 (1.559–1.724) <2 × 10−16
Age at onsetb 728 0.996 0.024 2110 0.998 0.065 2838 0.998 (0.996–0.999) 0.013
Disease course (PP)c 712 0.928 0.27 2106 1.031 0.42 2818 1.008 (0.944–1.076) 0.81
MSSSc 679 1.021 0.0032 1761 1.009 0.064 2440 1.014 (1.006–1.021) 4.0 × 10−4
Disease duration lumbar punctured 707 1.002 0.55 1899 0.998 0.35 2606 0.999 (0.996–1.002) 0.65
OCB-status versus n ORf P n ORf P n OR OCB pos (95% CI)f P
Gender (female)a 3026 1.172 0.19 3842 1.406 0.0011 6868 1.298 (1.111–1.514) 9.7 × 10−4
Age at onsetb 2607 0.982 0.0033 3753 0.992 0.092 6360 0.989 (0.982–0.996) 0.0028
Disease course (PP)c 2549 1.066 0.74 3715 0.894 0.49 6264 0.989 (0.782–1.261) 0.93
MSSSc 1694 1.040 0.18 2401 1.005 0.85 4095 1.023 (0.987–1.060) 0.21
Disease duration lumbar punctured 1343 1.012 0.39 2154 0.997 0.81 3497 1.003 (0.987–1.020) 0.71
Open in a new tab

PP = primary progressive; MSSS = Multiple Sclerosis Severity Score. Covariates included: acountry; bcountry and gender; ccountry, gender and age at onset; dcountry, gender and age at lumbar puncture (LP). eFold change in IgG index; fodds ratio for OCB positivity. Nominally significant P-values are indicated in bold.

Gender was highly correlated with both IgG index and OCB status. Females had on average a 1.12-fold higher IgG index (P = 1.9 × 10−10) and were 1.3-fold more likely to be OCB-positive than males (P = 9.7 × 10−4) (Table 2 and Supplementary Fig. 3). There was no correlation either between IgG index or OCB status and the interval between onset and lumbar puncture (median of 1 year). Younger patients had higher IgG indices (P = 0.013), and were more likely to be OCB-positive (P = 0.0028). We found no correlation between multiple sclerosis subtype (bout onset or progressive onset) and IgG index or OCB status after correction for gender and age at onset. Patients with a higher IgG index were more severely affected, as assessed by the Multiple Sclerosis Severity Score (P = 4.0 × 10−4).

Genome-wide association screen

After quality control of the screening data, 485 522 SNPs in 3026 samples and 485 236 SNPs in 938 samples were available for the analysis of OCB status and IgG index, respectively (Supplementary Table 2). Genomic inflation factor was 0.987 for OCB and 1.015 for IgG index (Supplementary Fig. 4).

We first performed a genome-wide analysis of all SNPs and CSF data available (Fig. 1). Markers in the MHC region on chromosome 6 were associated with OCB status with genome-wide significance (P < 5 × 10−8). We identified two regions showing genome-wide significance for association based on the IgG index, the MHC region on chromosome 6 and the IGHC region on chromosome 14. In addition, 14 lead SNPs and two conditional SNPs, and 15 lead SNPs and three conditional SNPs reached suggestive evidence for OCB status and IgG index, respectively (Tables 3 and 4), and were thus taken forward to replication.

Figure 1.

Figure 1

Open in a new tab

Manhattan plot of genome-wide association screen for (A) OCB status and (B) IgG index. Red line indicates genome-wide significance, blue line suggestive evidence.

Table 3.

Genetic associations for OCB -positive versus -negative patients with multiple sclerosis

Screen (n = 3026)
 Replication (n = 3478)
 Combined (n = 6504)
CHR POS SNP Minor Allele P OR SNP P OR P OR (95% CI)
Primary signals
2 179005067 rs3997876 A 1.22 × 10−5 0.65 NA NA NA NA NA
3 12763938 rs1467025 A 2.20 × 10−5 0.71 rs1467025 0.88 0.99 0.0030 0.85 (0.76–0.95)
3 53531263 rs979220 G 3.12 × 10−5 0.69 rs979220 0.85 1.02 0.0057 0.84 (0.74–0.95)
3 109630418 rs3900940 G 1.74 × 10−5 1.51 NA NA NA NA NA
4 169864539 rs9312352 G 3.35 × 10−7 0.6 rs9312352 0.22 1.14 0.0040 0.81 (0.7–0.93)
6 19718096 rs9465463 A 5.54 × 10−6 0.57 rs2743602a 0.67 1.06 NA NA
6 32700178 rs9271640 A 6.21 × 10−10 1.81 rs9271640 2.15 × 10−4 1.38 4.91 × 10−12 1.56 (1.38–1.77)
7 30113824 rs12701006 A 7.00 × 10−6 0.55 rs12701006 0.89 0.98 6.37 × 10−4 0.71 (0.58–0.86)
7 40906212 rs11767956 G 1.74 × 10−5 1.44 rs11767956 0.94 0.99 0.0050 1.17 (1.05–1.31)
9 18243014 rs11790235 A 2.36 × 10−6 0.48 rs11790235 0.22 0.83 2.44 × 10−5 0.63 (0.51–0.78)
10 34249465 rs743107 G 1.78 × 10−5 0.65 rs743107 0.5 0.94 4.80 × 10−4 0.78 (0.68–0.9)
10 71371517 rs3793821 G 1.57 × 10−6 0.61 rs16927103b 0.38 0.91 NA NA
18 46778159 rs9807334 A 3.58 × 10−6 0.58 rs9807334 0.02 0.76 8.45 × 10−7 0.66 (0.56–0.78)
23 72267781 rs723923 A 1.78 × 10−5 0.51 NA NA NA NA NA
23 142351548 rs6636962 A 1.59 × 10−6 0.61 rs6636962 1.00 1.00 7.01 × 10−4 0.78 (0.68–0.9)
Conditional signals
3 108688432 rs1299704 G 7.14 × 10−5 0.73 NA NA NA NA NA
6 32601567 rs34083746 G 4.14 × 10−5 0.61 rs3957148c 0.059 0.80 2.38 × 10−4 0.73 (0.62–0.86)
Previously reported associations
3 141582383 rs17411949 A 0.0002 0.56 rs17411949 0.65 1.08 0.014 0.75 (0.6–0.94)
6 98125801 rs9320598 G 0.00298 0.63 rs9489141d 0.23 0.84 NA NA
Open in a new tab

Correlation between replication and screen SNPs: ar2 = 1; br2 = 0.79; cr2 = 0.63; dr2 = 1. CHR = chromosome; POS = position (hg18). Signals meeting criteria for replication are indicated in bold.

Table 4.

Genetic associations for IgG index in patients with multiple sclerosis.

Screen (n = 938)
 Replication (n = 2072)
 Combined (n = 3010)
CHR POS SNP Minor allele P Beta log IgG index SNP P Beta log IgG index P Fold change IgG index (95% CI)
Primary signals
3 65383112 rs12493497 C 1.16 × 10−5 0.138 rs12493497 0.40 0.018 0.0015 1.04 (1.02–1.07)
3 115661169 rs10934275 G 2.68 × 10−5 −0.179 rs10934275 0.35 −0.028 0.0014 0.95 (0.92–0.98)
4 26927863 rs4692195 G 2.52 × 10−5 −0.151 rs4692195 0.73 −0.008 0.0080 0.96 (0.94–0.99)
6 32771829 rs6457617 A 1.07 × 10−10 −0.206 rs6457617 1.29 × 10−8 −0.127 6.72 × 10−17 0.90 (0.88–0.92)
6 125378498 rs1413598 A 1.15 × 10−5 0.151 NA NA NA NA NA
7 56016231 rs3816089 A 2.54 × 10−5 0.264 rs3816089 0.91 0.005 0.012 1.06 (1.01–1.12)
9 77456395 rs1328565 G 1.22 × 10−5 0.198 rs1328565 0.93 −0.003 0.018 1.04 (1.01–1.08)
10 95961456 rs12769135 C 2.72 × 10−5 −0.148 rs12769135 0.32 0.024 0.13 0.98 (0.95–1.01)
10 132158923 rs448669 C 2.00 × 10−5 0.186 rs448669 0.25 −0.033 0.17 1.02 (0.99–1.06)
12 120146349 rs3887080 A 2.24 × 10−5 0.2 rs3887080 0.10 0.054 1.60 × 10−4 1.07 (1.03–1.11)
13 26015959 rs10507368 G 8.63 × 10−6 −0.163 rs10507368 0.93 0.002 0.020 0.97 (0.94–1.00)
13 86546361 rs9516451 A 5.55 × 10−6 0.284 rs9516451 0.34 −0.040 0.086 1.04 (0.99–1.09)
14 99354933 rs4905899 A 1.92 × 10−5 −0.134 rs4905899 0.35 −0.020 0.0015 0.96 (0.94–0.99)
14 105243925 rs11621145 G 1.48 × 10−14 0.263 rs11621145 1.97 × 10−24 0.23 3.79 × 10−37 1.18 (1.13–1.22)
16 75867997 rs2343047 C 2.72 × 10−6 −0.151 rs1594049a 0.23 −0.026 NA NA
17 9982874 rs12162156 A 1.69 × 10−5 0.15 rs12162156 0.54 0.014 0.0040 1.04 (1.01–1.07)
22 24768495 rs6004913 G 3.48 × 10−6 0.183 rs6004913 0.93 0.002 0.0066 1.04 (1.01–1.08)
Conditional signals
3 117229505 rs9289038 C 0.0001 −0.145 rs9289038 0.28 0.0337 0.10 0.97 (0.94–1.01)
6 30183822 rs3132679 A 6.61 × 10−5 0.151 rs3132679 0.10 0.0444 2.60 × 10−4 1.06 (1.03–1.09)
7 56540030 rs10249528 C 5.02 × 10−6 0.158 rs10256440b 0.21 −0.0303 NA NA
Open in a new tab

Correlation between replication and screen SNPs: ar2 = 0.97; br2 = 1. Covariates: gender and seven principal components in screening phase, gender in replication phase. CHR = chromosome; POS = position (hg18). Signals meeting criteria for replication are indicated in bold.

When analysing the 57 multiple sclerosis risk SNPs previously identified in the genome-wide association study (The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011) (Supplementary Table 3), one SNP (rs12368653) showed nominally significant association with OCB positivity. For IgG index, 7 of these 57 SNPs (12%) reached nominal significance. The multiple sclerosis associated risk allele increased the IgG index for five of these SNPs. As the contribution to the overall multiple sclerosis risk of each of these SNPs is small, we also estimated the multiple sclerosis genetic burden. This genetic risk score was composed of 57 established non-HLA multiple sclerosis risk SNPs and four classical HLA alleles (Gourraud et al., 2011; The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011). A higher multiple sclerosis genetic burden including the HLA alleles was associated with increased IgG index {fold change = 1.11 [95% confidence interval (CI): 1.07–1.15], P = 1.33 × 10−9} and OCB status [OR = 1.51 (95% CI: 1.32–1.74), P = 3.44 × 10−9], whereas these associations were substantially reduced when considering only non-HLA risk SNPs [IgG index: fold change = 1.06 (95% CI: 1.00–1.13), P = 0.038; OCB: OR = 1.19 (95% CI: 0.96–1.48), P = 0.12], indicating the MHC region as the main contributor.

Subsequently, we investigated SNPs known from the literature to be associated with the presence of antibodies in diverse other tissues (Pubmed and Catalogue of Genome-wide Association Studies 4/12/2012) (Supplementary Table 4). A SNP tagging HLA-DRB1*15:01 (rs3129934), which is reported to be protective against IgA deficiency, increased IgG index (P = 1.37 × 10−7) as well as the risk of OCB positivity in multiple sclerosis (P = 3.7 × 10−4).

Replication

A total of 42 SNPs were taken forward to the replication phase and 38 SNPs were successfully genotyped (Supplementary Table 5). After quality control (Supplementary Table 6), 3478 patients with OCB status and 2072 with IgG index data available were analysed. The same two regions that were associated with genome-wide significance with IgG index in the screening phase, i.e. the MHC and IGHC region, reached genome-wide significance in the replication phase as well (Tables 3 and 4). One novel non-MHC region (rs9807334 near ELAC1/SMAD4) was replicated with nominal significance for association with OCB status.

Combined analyses

Given the highly significant correlation between OCB status and IgG index, we compared results from a detailed analysis of the MHC class II region for both characteristics (Supplementary Table 7 and Fig. 2). The most likely model in the combined analysis included two independent effects in the MHC class II region on OCB status; a major effect of the rs9271640*A-rs6457617*G haplotype and an additional effect of rs3957148 (tagging SNP with r2 = 0.63 for rs34083746 identified in the screening phase). The first signal is highly correlated with the HLA-DRB1*1501 haplotype (r2 = 0.94), whereas the second is strongly correlated with HLA-DQA1*0301 (rs34083746: r2 = 0.93, rs3957148: r2 = 0.64), and more modestly with both HLA-DRB1*0401 (rs34083746: r2 = 0.39) and *0404 (rs34083746: r2 = 0.26). Carrying the MHC SNP allele associated with OCB-positive status at these loci, was associated with a 2-fold higher likelihood of being OCB-positive than carrying the allele associated with OCB-negative status at all three loci [OR = 2.25 (95% CI: 1.84–2.75), P = 8.88 × 10−16] (Fig. 2A). The rs9271640*A-rs6457617*G haplotype, correlated with DRB1*1501, was associated not only with OCB status, but also IgG index (P = 1.59 × 10−22) (Fig. 2B). Independent of this haplotype, rs6457617*G had an additional effect increasing IgG index (P = 3.68 × 10−6). Notably, the association of rs6457617 with IgG index was observed in both OCB-positive (P = 4.22 × 10−10) and OCB-negative multiple sclerosis patients (P = 6.12 × 10−4). SNP rs6457617 is only modestly associated with any classical HLA allele (r2 < 0.30). SNP rs34083746/rs3957148, correlated with DQA1*0301 and associated with OCB negative status, did not seem to influence IgG index levels. No interaction amongst the associated SNPs in the MHC region or between these and the SNP in the IGHC region was observed (data not shown).

Figure 2.

Figure 2

Open in a new tab

Most likely model for MHC class II haplotypes (composed of SNPs rs9271640-rs6457617-rs3957148) and CSF phenotype in the combined analysis. (A) OCB status and (B) IgG index. GAG haplotype is correlated with HLA-DQA1*0301 (r2 = 0.64), AGA haplotype with HLA-DRB1*1501 (r2 = 0.94). P-values are given compared to reference haplotypes.

The previously reported association of IgG index with SNPs in the IGHC region was replicated as highly significant in the combined data set (P = 3.79 × 10−37) and explains 4.7% of the variance in IgG index in the entire data set. The MHC and IGHC region, the major determinants of variation in IgG index, together explain 7.75% of variance and combined with gender and country, account for 12.65% of variance in IgG index throughout Europe. Forest plots for all association signals per country are given in Supplementary Figs 5 and 6.

We also investigated the correlation between the main four established genetic effects and age at onset, disease course and Multiple Sclerosis Severity Score with covariates gender and, for the latter two, age at onset. After correcting for multiple testing, the A-allele of rs9271640 (that was associated with increased IgG levels and likelihood of OCB-positive status), was associated with lower age at onset (Pscreen = 4.44 × 10−4, Preplication = 0.082, Pcombined = 5.68 × 10−4). In contrast, at the IGHC locus, the allele that increased IgG index showed association with a later age at onset (Pscreen = 0.011, Preplication = 0.0066, Pcombined = 8.23 × 10−5) (Supplementary Fig. 7).

Discussion

This study is the largest to date that investigates genetic differences, clinical and demographic characteristics in relation to OCB status (n = 6868) and IgG index (n = 3126) in multiple sclerosis. Our findings strongly support that multiple sclerosis patients with and without OCBs and/or increased IgG index are genetically distinct. This has previously been suggested in other, smaller studies (Kikuchi et al., 2003; Imrell et al., 2006; Idiman et al., 2009; Wu et al., 2009; Romero-Pinel et al., 2011; Buck et al., 2012; Leone et al., 2013; Mero et al., 2013; Yoshimura et al., 2014) and is now confirmed. In accordance with our previous observations and those of others (Lechner-Scott et al., 2011; Mero et al., 2013; Stangel et al., 2013), patients with multiple sclerosis and high CSF antibody levels, as characterized by OCB-positive status and/or high IgG index, more often are female and seem to have a lower age at onset and higher Multiple Sclerosis Severity Score. We find that the two CSF parameters included in this study, OCB status and IgG index, are highly correlated (Mayringer et al., 2005; Link and Huang, 2006; Rinker et al., 2007), supporting that these measurements reflect the same immunological process.

Recently, the IGHC locus on chromosome 14 was reported to be associated with IgG index in German and Belgian patients with multiple sclerosis (Buck et al., 2012). This is convincingly replicated in the present extended study with the finding of a strong association of rs11621145 to IgG index (Pcombined = 3.79 × 10−37). In addition, we identify two novel association signals in the MHC region as major genetic determinants for IgG index; rs9271640, which is correlated with the HLA-DRB1*15:01 allele, and rs6457617, which is located near the HLA-DQA1 gene but not known to tag any conventional HLA-allele. An interaction between HLA and IGHC has been suggested (Pandey, 2013); however, we do not find any evidence to support this hypothesis. Together, these three SNPs from the MHC and IGHC regions explain ∼7.75% of the variance in IgG index.

Several studies have suggested that OCB-positive and -negative patients with multiple sclerosis are associated with different HLA-DRB1 alleles (Kikuchi et al., 2003; Imrell et al., 2006; Idiman et al., 2009; Wu et al., 2009; Romero-Pinel et al., 2011; Leone et al., 2013; Mero et al., 2013; Yoshimura et al., 2014). We observe a significant association of SNP rs9271640 tagging HLA-DRB1*15:01 with OCB status and a high correlation of the multiple sclerosis genetic burden including HLA risk alleles with OCB status. Our findings hence support previous observations of the major established multiple sclerosis risk allele, HLA-DRB1*15:01, being more strongly associated with OCB-positive multiple sclerosis than OCB-negative multiple sclerosis. The present study also detects a significant difference between OCB-positive and -negative patients with regard to rs34083746. This SNP is highly correlated with the DQA1*0301 allele shared between DRB1*04 haplotypes. Hence, our findings are consistent with several HLA-DRB1*04 alleles having been shown to increase risk for OCB-negative multiple sclerosis, but not OCB-positive multiple sclerosis, in previous studies from Europe and Japan (Kikuchi et al., 2003; Imrell et al., 2006; Mero et al., 2013; Yoshimura et al., 2014). The combination of variation in the MHC region has a major impact on the OCB status of patients with multiple sclerosis, resulting in a >2-fold difference in the odds of being OCB-positive. In addition to replicating previously suggested associations to OCB status within the MHC region, our analyses provide evidence for a novel locus to be associated with OCB status, rs9807334 near the ELAC1/SMAD4 genes. The SMAD4 gene, a signal transduction protein in the tumor growth factor beta pathway, has previously been implicated in class switch recombination and in experimental autoimmune encephalomyelitis and multiple sclerosis (Park et al., 2005; Meoli et al., 2010; Huss et al., 2011) and the same allele has previously been suggested as associated with vaccine response (Ovsyannikova et al., 2012).

The CSF phenotype association signals in the MHC region we observe have been associated with susceptibility and antibody levels in other diseases. The HLA-DRB1*1501-DQB1*0601 haplotype has been associated with either the presence or increased quantity of immunoglobulins of the IgG, IgA and IgM families both in healthy controls and in disease, including total immunoglobulins (Ferreira et al., 2010), antibodies induced by viruses such as Epstein-Barr virus (Rubicz et al., 2013), and autoantibodies in type 1 diabetes (Ishii et al., 2005) and Sjögren syndrome (Gottenberg et al., 2003) though an opposite correlation is seen for few other antibody responses (Sundqvist et al., 2014). HLA-DQA1*0301, highly correlated with SNP rs34083746, is associated with autoantibody negative disease in ketosis-prone diabetes (Oak et al., 2014). SNP rs6457617 is a major susceptibility factor in rheumatoid arthritis and systemic sclerosis, independent of classical HLA alleles (The International MHC and Autoimmunity Genetics Network, 2009; Radstake et al., 2010; Allanore et al., 2011; Orozco et al., 2014). Of note, the risk allele for these autoimmune diseases decreases the IgG levels in multiple sclerosis. The IGHC locus, on the other hand, has not been associated with antibody levels in other diseases as yet. Overall, the MHC region may enhance antibody production and OCB-positive status, whereas the IGHC region has been proposed to influence clearance of antibodies from the CSF and hence immunoglobulin levels (Buck et al., 2012).

Other studies have shown that IgG levels correlate with the number of B cells, more particularly the number of plasmablasts, in the CSF (Rudick et al., 1999; Cepok et al., 2001, 2005). B cell follicles in the meninges have been found in a subset of patients with multiple sclerosis and are suggested to play a pathogenic role (Magliozzi et al., 2007). Moreover, a prognostic potential of findings in the CSF has been proposed, as positivity for OCB is reported to double the risk of developing clinically definite multiple sclerosis in patients with clinically isolated syndrome (Tintore et al., 2008). Also, some studies suggest that the inflammation in the brain reflected by CSF alterations may correlate with neurodegeneration and disease progression (Stangel et al., 2013). Our findings of a lower age at onset and a higher Multiple Sclerosis Severity Score in the patients with multiple sclerosis with marked CSF antibody levels support this hypothesis. In our study, HLA-DRB1*15:01 increases CSF antibody levels and lowers age at onset, as reported previously (Hensiek et al., 2002; The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2, 2011). However, at the IGHC locus, the allele increasing IgG index seems correlated with higher age at onset, indicating that the mechanisms underlying the correlation between genetic determinants, CSF antibody levels and clinical outcome are not fully understood as yet.

The present study is well-powered, and the clinical characteristics of the patients included support that they are representative of patients with multiple sclerosis in general. On average, OCB positivity was found in 88% and an elevated IgG index in 64% of patients, in accordance with earlier studies (Dobson et al., 2013; Stangel et al., 2013). Lumbar puncture is routinely performed as a part of the diagnostic process in most of the countries included, but a possible selection bias must be kept in mind when interpreting the clinical findings. OCBs and/or increased IgG index in the CSF are clinically important hallmarks in multiple sclerosis, and in this large study we show that these disease phenotypes are associated with both genetic variants and clinical and demographic characteristics. The presence of intrathecal immunoglobulin M has been reported as an additional CSF characteristic in multiple sclerosis patients that may be associated with unfavourable outcome or aggressive disease (Stangel et al., 2013). The study of genetic factors underlying immunoglobulin M levels is currently hampered by the availability of data but will be of future interest.

In summary, in this large study including 6950 patients with multiple sclerosis, we confirm that genetic variants in the immunologically important regions of MHC and IGHC influence OCB status and IgG index in patients with multiple sclerosis.

Acknowledgements

We thank the patients who participated in this study. We acknowledge the International Multiple Sclerosis Genetics Consortium (IMSGC) and Wellcome Trust Case Control Consortium 2 (WTCCC2) for the collaboration with the establishment of the genotypes that were included in the screening phase of this project. Please see Supplementary material for the IMSGC membership list. We thank Leentje Cosemans and the Genomics Core Facility at the KU Leuven (Belgium) for technical assistance with the replication phase and Prof. Koen Poesen for advice. From Norway we thank collaborators from the Multiple Sclerosis Research group in Oslo, especially Anja Bjølgerud and Inger-Lise Mero, and the Norwegian Multiple Sclerosis Register and Biobank for contribution with samples and data. The Italian group thanks the PROGEMUS (PROgnostic GEnetic factors in MUltiple Sclerosis) Consortium, the PROGRESSO Consortium, Dr Mariaemma Rodegher, Dr Lucia Moiola, Dr Bruno Colombo, Dr Vittorio Martinelli, Dr ssa Gabriella Passerini and Prof. Giancarlo Comi. The German group gratefully acknowledges Verena Grummel for her continuous contributions to the DNA biobank. From Australia, Prof. William Carroll, Dr Marzena Fabis-Pedrini, Mr Sean Burnard and Mrs Malgorzata Krupa are thanked for their contributions.

Glossary

Abbreviations

HLA

human leukocyte antigen

IgG

immunoglobulin G

IGHC

immunoglobulin heavy chain

MHC

major histocompatibility complex

OCB

oligoclonal band

SNP

single nucleotide polymorphism

Funding

The Belgian contribution to the project was supported by the Research Fund KU Leuven (OT/11/087 to A.G. and B.D.), the Belgian Charcot Foundation (Research Grant to A.G.), and the Research Foundation Flanders (FWO-Vlaanderen; Clinical Investigator Fellowship to B.D.). The Norwegian part of the study was supported by The Research Council of Norway (project 143153 and 196776 by H.F.H.), University of Oslo (project 211567 by H.F.H.), NevroNor, Research Council of Norway (project 176958 by K.M.M.) and The Western Norway Regional Health Authority, Stavanger, Norway. The Italian contribution was supported by the Italian Foundation for Multiple Sclerosis [FISM grants ‘Progetto Speciale Immunochip’, 2011/R/14, Fondazione cariplo (n° 2010‐0728)], and the CRT Foundation, Turin, Italy. The German contribution was supported by the German Ministry for Education and Research (BMBF), the German Competence Network Multiple Sclerosis (KKNMS, Control-MS, 01GI0917), the Deutsche Forschungsgemeinschaft (SyNergy Cluster, JE 530/1‐1), and the Munich Biomed Cluster M4. The Danish contribution was supported by grants from the Danish Multiple Sclerosis Society, the Danish Council for Strategic Research [grant number 2142‐08‐0039] and the Warwara Larsen Foundation. The Spanish contribution was supported by the ‘Red Española de Esclerosis Múltiple (REEM)’ sponsored by the FEDER-FIS and the ‘Ajuts per donar Suport als Grups de Recerca de Catalunya’, sponsored by the ‘Agència de Gestió d’Ajuts Universitaris i de Recerca’ (AGAUR), Generalitat de Catalunya, Spain. P.A.G. is a recipient of the Race to erase MS Award and the European Federation for Immunogenetics Julia Bodmer Award.

Supplementary material

Supplementary material is available at Brain online.

References

  1. Allanore Y, Saad M, Dieude P, Avouac J, Distler JH, Amouyel P, et al. Genome-wide scan identifies TNIP1, PSORS1C1, and RHOB as novel risk loci for systemic sclerosis. PLoS Genet. 2011;7:e1002091. doi: 10.1371/journal.pgen.1002091. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Berg-Hansen P, Smestad C, Sandvik L, Harbo HF, Celius EG. Increased disease severity in non-Western immigrants with multiple sclerosis in Oslo, Norway. Eur J Neurol. 2013;20:1546–52. doi: 10.1111/ene.12227. [DOI] [PubMed] [Google Scholar]
  3. Buck D, Albrecht E, Aslam M, Goris A, Hauenstein N, Jochim A, et al. Genetic variants in the immunoglobulin heavy chain locus are associated with the IgG index in multiple sclerosis. Ann Neurol. 2012;73:86–94. doi: 10.1002/ana.23749. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Cepok S, Jacobsen M, Schock S, Omer B, Jaekel S, Boddeker I, et al. Patterns of cerebrospinal fluid pathology correlate with disease progression in multiple sclerosis. Brain. 2001;124(Pt 11):2169–76. doi: 10.1093/brain/124.11.2169. [DOI] [PubMed] [Google Scholar]
  5. Cepok S, Rosche B, Grummel V, Vogel F, Zhou D, Sayn J, et al. Short-lived plasma blasts are the main B cell effector subset during the course of multiple sclerosis. Brain. 2005;128(Pt 7):1667–76. doi: 10.1093/brain/awh486. [DOI] [PubMed] [Google Scholar]
  6. Compston A, Coles A. Multiple sclerosis. Lancet. 2008;372:1502–17. doi: 10.1016/S0140-6736(08)61620-7. [DOI] [PubMed] [Google Scholar]
  7. Dilthey AT, Moutsianas L, Leslie S, McVean G. HLA*IMP—an integrated framework for imputing classical HLA alleles from SNP genotypes. Bioinformatics. 2011;27:968–72. doi: 10.1093/bioinformatics/btr061. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Dobson R, Ramagopalan S, Davis A, Giovannoni G. Cerebrospinal fluid oligoclonal bands in multiple sclerosis and clinically isolated syndromes: a meta-analysis of prevalence, prognosis and effect of latitude. J Neurol Neurosurg Psychiatry. 2013;84:909–14. doi: 10.1136/jnnp-2012-304695. [DOI] [PubMed] [Google Scholar]
  9. Ferreira RC, Pan-Hammarstrom Q, Graham RR, Gateva V, Fontan G, Lee AT, et al. Association of IFIH1 and other autoimmunity risk alleles with selective IgA deficiency. Nat Genet. 2010;42:777–80. doi: 10.1038/ng.644. [DOI] [PubMed] [Google Scholar]
  10. Fukazawa T, Kikuchi S, Sasaki H, Hamada K, Hamada T, Miyasaka K, et al. The significance of oligoclonal bands in multiple sclerosis in Japan: relevance of immunogenetic backgrounds. J Neurol Sci. 1998;158:209–14. doi: 10.1016/s0022-510x(98)00127-0. [DOI] [PubMed] [Google Scholar]
  11. Gottenberg JE, Busson M, Loiseau P, Cohen-Solal J, Lepage V, Charron D, et al. In primary Sjogren’s syndrome, HLA class II is associated exclusively with autoantibody production and spreading of the autoimmune response. Arthritis Rheum. 2003;48:2240–5. doi: 10.1002/art.11103. [DOI] [PubMed] [Google Scholar]
  12. Gourraud PA, McElroy JP, Caillier SJ, Johnson BA, Santaniello A, Hauser SL, et al. Aggregation of multiple sclerosis genetic risk variants in multiple and single case families. Ann Neurol. 2011;69:65–74. doi: 10.1002/ana.22323. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Haghighi S, Andersen O, Rosengren L, Bergstrom T, Wahlstrom J, Nilsson S. Incidence of CSF abnormalities in siblings of multiple sclerosis patients and unrelated controls. J Neurol. 2000;247:616–22. doi: 10.1007/s004150070130. [DOI] [PubMed] [Google Scholar]
  14. Harbo HF, Isobe N, Berg-Hansen P, Bos SD, Caillier SJ, Gustavsen MW, et al. Oligoclonal bands and age at onset correlate with genetic risk score in multiple sclerosis. Mult Scler. 2014;20:660–8. doi: 10.1177/1352458513506503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Hensiek AE, Sawcer SJ, Feakes R, Deans J, Mander A, Akesson E, et al. HLA-DR 15 is associated with female sex and younger age at diagnosis in multiple sclerosis. J Neurol Neurosurg Psychiatry. 2002;72:184–7. doi: 10.1136/jnnp.72.2.184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Huss DJ, Winger RC, Cox GM, Guerau-de-Arellano M, Yang Y, Racke MK, et al. TGF-beta signaling via Smad4 drives IL-10 production in effector Th1 cells and reduces T-cell trafficking in EAE. Eur J Immunol. 2011;41:2987–96. doi: 10.1002/eji.201141666. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Idiman E, Ozakbas S, Dogan Y, Kosehasanogullari G. The significance of oligoclonal bands in multiple sclerosis: relevance of demographic and clinical features, and immunogenetic backgrounds. J Neuroimmunol. 2009;212:121–4. doi: 10.1016/j.jneuroim.2009.04.014. [DOI] [PubMed] [Google Scholar]
  18. Imrell K, Landtblom AM, Hillert J, Masterman T. Multiple sclerosis with and without CSF bands: clinically indistinguishable but immunogenetically distinct. Neurology. 2006;67:1062–4. doi: 10.1212/01.wnl.0000237343.93389.35. [DOI] [PubMed] [Google Scholar]
  19. Ishii M, Hasegawa G, Fukui M, Obayashi H, Ohta M, Ogata M, et al. Clinical and genetic characteristics of diabetic patients with high-titer (>10,000 U/ml) of antibodies to glutamic acid decarboxylase. Immunol Lett. 2005;99:180–5. doi: 10.1016/j.imlet.2005.03.001. [DOI] [PubMed] [Google Scholar]
  20. Jersild C, Svejgaard A, Fog T. HL-A antigens and multiple sclerosis. Lancet. 1972;1:1240–1. doi: 10.1016/s0140-6736(72)90962-2. [DOI] [PubMed] [Google Scholar]
  21. Kikuchi S, Fukazawa T, Niino M, Yabe I, Miyagishi R, Hamada T, et al. HLA-related subpopulations of MS in Japanese with and without oligoclonal IgG bands. Human leukocyte antigen. Neurology. 2003;60:647–51. doi: 10.1212/01.wnl.0000048202.09147.9e. [DOI] [PubMed] [Google Scholar]
  22. Lechner-Scott J, Spencer B, de Malmanche T, Attia J, Fitzgerald M, Trojano M, et al. The frequency of CSF oligoclonal banding in multiple sclerosis increases with latitude. Mult Scler. 2011;18:974–82. doi: 10.1177/1352458511431729. [DOI] [PubMed] [Google Scholar]
  23. Leone MA, Barizzone N, Esposito F, Lucenti A, Harbo HF, Goris A, et al. Association of genetic markers with CSF oligoclonal bands in multiple sclerosis patients. PLoS One. 2013;8:e64408. doi: 10.1371/journal.pone.0064408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Link H, Huang YM. Oligoclonal bands in multiple sclerosis cerebrospinal fluid: an update on methodology and clinical usefulness. J Neuroimmunol. 2006;180:17–28. doi: 10.1016/j.jneuroim.2006.07.006. [DOI] [PubMed] [Google Scholar]
  25. Magliozzi R, Howell O, Vora A, Serafini B, Nicholas R, Puopolo M, et al. Meningeal B-cell follicles in secondary progressive multiple sclerosis associate with early onset of disease and severe cortical pathology. Brain. 2007;130(Pt 4):1089–104. doi: 10.1093/brain/awm038. [DOI] [PubMed] [Google Scholar]
  26. Mayringer I, Timeltaler B, Deisenhammer F. Correlation between the IgG index, oligoclonal bands in CSF, and the diagnosis of demyelinating diseases. Eur J Neurol. 2005;12:527–30. doi: 10.1111/j.1468-1331.2005.00997.x. [DOI] [PubMed] [Google Scholar]
  27. McDonald WI, Compston A, Edan G, Goodkin D, Hartung HP, Lublin FD, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol. 2001;50:121–7. doi: 10.1002/ana.1032. [DOI] [PubMed] [Google Scholar]
  28. Meoli EM, Oh U, Grant CW, Jacobson S. TGF-beta signaling is altered in the peripheral blood of subjects with multiple sclerosis. J Neuroimmunol. 2010;230:164–8. doi: 10.1016/j.jneuroim.2010.10.028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Mero IL, Gustavsen MW, Saether HS, Flam ST, Berg-Hansen P, Sondergaard HB, et al. Oligoclonal band status in Scandinavian multiple sclerosis patients is associated with specific genetic risk alleles. PLoS One. 2013;8:e58352. doi: 10.1371/journal.pone.0058352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Oak S, Gaur LK, Radtke J, Patel R, Iyer D, Ram N, et al. “Masked” and overt autoantibodies specific to the DPD epitope of 65 kDa glutamate decarboxylase (GAD65-DPD) are associated with preserved beta cell functional reserve in Ketosis-Prone Diabetes. J Clin Endocrinol Metab. 2014;99:E1040–4. doi: 10.1210/jc.2013-4189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Orozco G, Viatte S, Bowes J, Martin P, Wilson AG, Morgan AW, et al. Novel rheumatoid arthritis susceptibility locus at 22q12 identified in an extended UK genome-wide association study. Arthritis Rheumatol. 2014;66:24–30. doi: 10.1002/art.38196. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Ovsyannikova IG, Kennedy RB, O’Byrne M, Jacobson RM, Pankratz VS, Poland GA. Genome-wide association study of antibody response to smallpox vaccine. Vaccine. 2012;30:4182–9. doi: 10.1016/j.vaccine.2012.04.055. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Pandey JP. Immunoglobulin GM allotypes in multiple sclerosis. Ann Neurol. 2013;73:148. doi: 10.1002/ana.23804. [DOI] [PubMed] [Google Scholar]
  34. Park SR, Seo GY, Choi AJ, Stavnezer J, Kim PH. Analysis of transforming growth factor-beta1-induced Ig germ-line gamma2b transcription and its implication for IgA isotype switching. Eur J Immunol. 2005;35:946–56. doi: 10.1002/eji.200425848. [DOI] [PubMed] [Google Scholar]
  35. Plagnol V, Howson JM, Smyth DJ, Walker N, Hafler JP, Wallace C, et al. Genome-wide association analysis of autoantibody positivity in type 1 diabetes cases. PLoS Genet. 2011;7:e1002216. doi: 10.1371/journal.pgen.1002216. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Ebers GC, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol. 1983;13:227–31. doi: 10.1002/ana.410130302. [DOI] [PubMed] [Google Scholar]
  37. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007;81:559–75. doi: 10.1086/519795. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Radstake TR, Gorlova O, Rueda B, Martin JE, Alizadeh BZ, Palomino-Morales R, et al. Genome-wide association study of systemic sclerosis identifies CD247 as a new susceptibility locus. Nat Genet. 2010;42:426–9. doi: 10.1038/ng.565. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rinker JR, Trinkaus K, II, Naismith RT, Cross AH. Higher IgG index found in African Americans versus Caucasians with multiple sclerosis. Neurology. 2007;69:68–72. doi: 10.1212/01.wnl.0000265057.79843.d9. [DOI] [PubMed] [Google Scholar]
  40. Romero-Pinel L, Martinez-Yelamos S, Bau L, Matas E, Gubieras L, Maria Pujal J, et al. Association of HLA-DRB1*15 allele and CSF oligoclonal bands in a Spanish multiple sclerosis cohort. Eur J Neurol. 2011;18:1258–62. doi: 10.1111/j.1468-1331.2011.03379.x. [DOI] [PubMed] [Google Scholar]
  41. Roxburgh RH, Seaman SR, Masterman T, Hensiek AE, Sawcer SJ, Vukusic S, et al. Multiple sclerosis severity score: using disability and disease duration to rate disease severity. Neurology. 2005;64:1144–51. doi: 10.1212/01.WNL.0000156155.19270.F8. [DOI] [PubMed] [Google Scholar]
  42. Rubicz R, Yolken R, Drigalenko E, Carless MA, Dyer TD, Bauman L, et al. A Genome-wide integrative genomic study localizes genetic factors influencing antibodies against Epstein-Barr Virus Nuclear Antigen 1 (EBNA-1) PLoS Genet. 2013;9:e1003147. doi: 10.1371/journal.pgen.1003147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Rudick RA, Cookfair DL, Simonian NA, Ransohoff RM, Richert JR, Jacobs LD, et al. Cerebrospinal fluid abnormalities in a phase III trial of Avonex (IFNbeta-1a) for relapsing multiple sclerosis. The multiple sclerosis collaborative research group. J Neuroimmunol. 1999;93:8–14. doi: 10.1016/s0165-5728(98)00174-x. [DOI] [PubMed] [Google Scholar]
  44. Stangel M, Fredrikson S, Meinl E, Petzold A, Stuve O, Tumani H. The utility of cerebrospinal fluid analysis in patients with multiple sclerosis. Nat Rev Neurol. 2013;9:267–76. doi: 10.1038/nrneurol.2013.41. [DOI] [PubMed] [Google Scholar]
  45. Sundqvist E, Buck D, Warnke C, Albrecht E, Gieger C, Khademi M, et al. JC polyomavirus infection is strongly controlled by human leucocyte antigen class II variants. PLoS Pathog. 2014;10:e1004084. doi: 10.1371/journal.ppat.1004084. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. The International MHC and Autoimmunity Genetics Network. Mapping of multiple susceptibility variants within the MHC region for 7 immune-mediated diseases. Proc Natl Acad Sci USA. 2009;106:18680–5. doi: 10.1073/pnas.0909307106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. The International Multiple Sclerosis Genetics Consortium. Analysis of immune-related loci identifies 48 new susceptibility variants for multiple sclerosis. Nat Genet. 2013;45:1353–60. doi: 10.1038/ng.2770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. The International Multiple Sclerosis Genetics Consortium and The Wellcome Trust Case Control Consortium 2. Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis. Nature. 2011;476:214–9. doi: 10.1038/nature10251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Tintore M, Rovira A, Rio J, Tur C, Pelayo R, Nos C, et al. Do oligoclonal bands add information to MRI in first attacks of multiple sclerosis? Neurology. 2008;70(13 Pt 2):1079–83. doi: 10.1212/01.wnl.0000280576.73609.c6. [DOI] [PubMed] [Google Scholar]
  50. Wu JS, Qiu W, Castley A, James I, Joseph J, Christiansen FT, et al. Presence of CSF oligoclonal bands (OCB) is associated with the HLA-DRB1 genotype in a West Australian multiple sclerosis cohort. J Neurol Sci. 2009;288:63–7. doi: 10.1016/j.jns.2009.10.005. [DOI] [PubMed] [Google Scholar]
  51. Yoshimura S, Isobe N, Matsushita T, Masaki K, Sato S, Kawano Y, et al. Genetic and infectious profiles influence cerebrospinal fluid IgG abnormality in Japanese multiple sclerosis patients. PLoS One. 2014;9:e95367. doi: 10.1371/journal.pone.0095367. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Brain are provided here courtesy of Oxford University Press

RESOURCES