Original Article

Genes and Immunity (2011) 12, 191–198; doi:10.1038/gene.2010.59; published online 23 December 2010

Exploring the CLEC16A gene reveals a MS-associated variant with correlation to the relative expression of CLEC16A isoforms in thymus

I-L Mero1,2, M Ban3, Å R Lorentzen2,4, C Smestad1, E G Celius1, H Sæther2, H Saeedi2, M K Viken2, B Skinningsrud5, D E Undlien5, J Aarseth6, K-M Myhr6, S Granum7, A Spurkland7, S Sawcer3, A Compston3, B A Lie2,8 and H F Harbo1,4,8

  1. 1Department of Neurology, Oslo University Hospital, Oslo, Norway
  2. 2Institute of Immunology, Oslo University Hospital, Oslo, Norway
  3. 3Department of Clinical Neurosciences, University of Cambridge, Addenbrooke's, Cambridge, UK
  4. 4Institute of Clinical Medicine, University of Oslo, Oslo, Norway
  5. 5Department and Institute of Medical Genetics, Oslo University Hospital and University of Oslo, Oslo, Norway
  6. 6The Norwegian multiple sclerosis registry and biobank, Department of Neurology, Haukeland University Hospital, Bergen, Norway
  7. 7Institute of Basal Medical Science, University of Oslo, Oslo, Norway

Correspondence: Dr I-L Mero, Department of Neurology, Oslo University Hospital, Ullevål, N-0407 Oslo, Norway. E-mail: ilmero@rr-research.no

8These authors contributed equally to this work.

Received 3 May 2010; Revised 29 July 2010; Accepted 3 August 2010; Published online 23 December 2010.



Genomewide association studies have implicated the CLEC16A gene in several autoimmune diseases, including multiple sclerosis (MS) and type 1 diabetes. However, the most associated single-nucleotide polymorphism (SNP) varies, and causal variants are still to be defined. In MS, two SNPs in partial linkage disequilibrium with each other, rs6498169 and rs12708716, have been validated at genomewide significance level. To explore the CLEC16A association in MS in more detail, we genotyped 57 SNPs in 807 Norwegian MS patients and 1027 Norwegian controls. Six highly associated SNPs emerged and were then replicated in two large independent sample sets (Norwegian and British), together including 1153 MS trios, 2308 MS patients and 4044 healthy controls. In combined analyses, SNP rs12708716 gave the strongest association signal in MS (P=5.3 × 10−8, odds ratio 1.18, 95% confidence interval=1.11–1.25), and was found to be superior to the other SNP associations in conditional logistic regression analyses. Expression analysis revealed that rs12708716 genotype was significantly associated with the relative expression levels of two different CLEC16A transcripts in thymus (P=0.004), but not in blood, possibly implying a thymus- or cell-specific splice regulation.


CLEC16A; MS; finemapping; gene expression



Genomewide association studies (GWAS) have demonstrated that single-nucleotide polymorphisms (SNPs) in the C-type lectin domain family 16, member A (CLEC16A) gene at chromosome 16p13, influence susceptibility in both multiple sclerosis (MS) and type 1 diabetes (T1D).1, 2, 3 Multiple replication and candidate gene studies have verified this region as being disease associated in T1D and MS, and association with CLEC16A has also been found in anti-cyclic citrullinated peptide (CCP) negative rheumatoid arthritis, juvenile idiopathic arthritis, Addison's disease and NOD2-negative Crohn's disease.4, 5, 6, 7, 8, 9 The association of CLEC16A SNPs across multiple immune-mediated diseases and the observation that the gene is almost uniquely expressed on immune cells, make a common effect on autoimmunity likely. The CLEC16A protein belongs to the C-type lectin family whose immune functions include differentiation of self versus non-self glycoproteins and the induction of the appropriate immune response.10 CLEC16A is an atypical C-type lectin carrying only a short C-type lectin domain unable to bind carbohydrates, and the function of the protein is still unknown.9 Limited data exists on CLEC16A isoforms, and despite its considerable length spanning 200000bp, only three transcripts are annotated at the University of California Santa Cruz (UCSC) Genome Browser on Human Feb. 2009 (GRCh37/hg19) Assembly (http://genome.ucsc.edu). All the strongest SNP associations found so far are situated in seemingly noncoding regions, and their biological consequences are largely unknown.

The original CLEC16A SNP association in MS, found by the International MS Genetics Consortium (IMSGC) GWAS in 2007, was rs6498169 in intron 22.2 This SNP association has been replicated and confirmed at genomewide significance level in MS (P<5 × 10−7).6, 11 In T1D, the strongest association so far has been found for four SNPs situated in intron 19 and 22; rs12708716,9 rs725613, rs2903692 and rs17673553,1 which are all in fairly strong linkage disequilibrium (LD) (r2=0.66–1.00; haploview v 4.1, hapmap release 21, CEU sample set (Utah residents with Northern and Western European ancestry from the Centre de'Etude du Polymorphism Human (CEPH) collection).12, 13 The LD between the MS-associated rs6498169 and the T1D SNPs is modest (r2=0.18–0.26),12, 13 and association with rs6498169 has not been found in T1D.8, 14 In contrast, the T1D-associated SNPs rs12708716 and rs725613 have indeed also shown strong association with MS in an IMSGC study including more than 2000 trio families, 5000 cases and 10000 unrelated controls (P=1.6 × 10−15), and in a Sardinian study (P=6.4 × 10−5), respectively.4, 15

The reported CLEC16A SNP associations also vary for other autoimmune diseases. The originally described MS-associated SNP rs6498169 has been shown to be associated with anti-CCP-negative rheumatoid arthritis and juvenile idiopathic arthritis, but not anti-CCP-positive rheumatoid arthritis, Addison's disease and NOD2-negative Crohn's disease.5, 8, 16 Vice versa the T1D SNPs rs2903692 and rs12708716 have shown association with NOD2-negative Crohn's and Addison's disease, respectively.5, 8

Thus to date, MS is the only disease described to be associated with both rs6498169 and the LD block containing the four T1D associated SNPs. The only fine-mapping study of CLEC16A performed in MS so far is an Australian study (n=1146 cases and n=1309 controls) of several MS risk genes, where 44 tagging SNPs in the CLEC16A gene pointed towards rs6498169, the most significantly associated CLEC16A SNP in the original IMSGC GWAS.2, 17 It does not appear from the paper whether rs12708716 was tested in this study.

In order to further explore the CLEC16A SNP association in MS, we conducted fine mapping in two phases, screening a 57 SNP panel in a Norwegian MS patient-control sample set in the first phase and replicating the top hit findings in two large independent Norwegian and British sample sets in the second phase. Furthermore, we studied the expression levels of different CLEC16A transcripts in thymus and blood in order to correlate these with the genotyping results.



Screening phase

Our screening phase comprised a 57 SNP panel specifically designed to (i) explore SNPs in LD with the MS-associated rs12708716, (ii) test SNPs previously reported in autoimmune disease and (iii) tag the entire CLEC16A gene. The panel was genotyped in a Norwegian sample set, which after removal of three individuals with more than 50% missing genotypes, included 804 MS cases and 1027 controls. Five of 57 markers failed in the screening phase: rs794423 and rs12927773 did not genotype because of assay failure, whereas rs45537935, rs7204935 and rs3177383 were genotyped as monomorphic. Genotype success rates were above 95% for all remaining markers, and all SNPs were in Hardy–Weinberg Equilibrium (P>0.001) in patients and controls. Association results and inclusion criteria for all SNPs in the screening phase can be found in Supplementary Table 1.

Five markers were associated with a P<0.001 in the screening phase (Table 1). All were included in the study as part of the tagging SNP set, except rs6498169, previously reported to be associated with MS and anti-CCP-negative rheumatoid arthritis.2, 6, 8, 11 The most significantly associated SNP, rs12923849 has previously showed some degree of association with T1D.1 The other three top hits in our screening phase; rs9934231, rs6498168 and rs7206912 have not previously been reported to be associated in MS or other autoimmune diseases. Because of failure of genotyping rs6498168 in the replication phase, it was replaced by a SNP in complete LD (r2=1), rs8060411, and this SNP was for completeness, also tested in the screening panel.

Neither the MS-associated rs12708716 (P=0.006) nor any of the SNPs selected on the basis of LD with this variant (that is, r2>0.6) were among these top hits. However, several of them showed significant association in the screening (P<0.01) (Supplementary Table 1). As the magnitude of their association was negligibly different from rs12708716, we chose to only include the already confirmed associated SNP rs12708716 in our replication.

To evaluate the correlation of the six SNPs chosen for replication, we looked at their LD pattern in the screening sample set (Supplementary Figure 1). Rs12923849, rs9934231 and the previously reported rs12708716 were relatively uncorrelated with the other top hit SNPs from the screening (r2less than or equal to0.29), whereas rs6498168, rs7206912 and rs6498169 showed correlation with each other (r2>0.70). All six SNPs were in strong LD as measured by D′ (D′>0.65). In general, strong D′-LD, but limited r2-LD was observed across the entire region screened (Supplementary Figure 1).

Replication and combined phase

Six SNPs were genotyped in the Norwegian and the British replication sample sets. All SNPs had a genotype success rates>95% and were in Hardy–Weinberg Equilibrium (P>0.001) in patients and controls. No Mendelian errors were found within the genotyped trios. Removal of individuals (15 Norwegian and 19 British) and trios (n=40) with less than 50% genotype success rates, left 1201 Norwegian cases, 2054 Norwegian controls, 1113 British trios, 1097 UK cases and 1966 UK controls for further analyses. The results from the replication and combined analyses of the screening and replication phases are shown in Table 2. In the Norwegian replication sample set, all SNPs except rs12923849 were replicated to be associated at the 5% significance level, with rs7296012 being the most strongly associated (P=2.3 × 10−4). In the UK replication sample set, only rs12708716 (P=4.5 × 10−4) and rs12923849 (P=0.015) were significantly associated in the trios and case controls combined. The association with rs12708716 had previously been found for the entire UK sample set in an IMSGC study.4 However, for all the other SNPs included in the replication, the trends of association in the British samples were in the same direction as for the two Norwegian sample sets.

Combining the screening and replication data including genotypes of 1113 trios, 3102 MS patients and 5047 controls (Table 2) collected in Norway and the UK, the strongest association was seen for SNP rs12708716 (P=5.3 × 10−8, P corrected=3.2 × 10−7, odds ratio=1.18, 95% confidence intervals=1.11–1.25), which correlates with earlier reports in MS and T1D.4, 9 The other SNPs also reached highly significant P-values on their own, including SNP rs7206912, not previously demonstrated to be disease associated (P=3.4 × 10−6, P corrected=2.1 × 10−5, odds ratio=1.14, 95% confidence intervals=1.08–1.21). Significant population heterogeneity (P<0.05) was seen for all SNPs except the most associated rs12708716, and to adjust for this all combined analyses were calculated with nationality as confounder.

LD, haplotype and logistic regression calculations

The correlation between the most significantly associated SNP rs12708716 and the other five replicated top hit SNPs was modest (r2<0.4; Figure 1a), but LD at the D′ level was high (D′>0.65; Figure 1b). The relationship between the six top hit SNPs was therefore further explored by haplotype analyses (Table 3), which interestingly showed that only six haplotypes were commonly seen (>5%) in the tested populations. Among these six haplotypes there was only one significant risk haplotype carrying all the risk alleles of the six replicated SNPs, and one highly associated protective haplotype harbouring all the protective alleles. One other haplotype identical for all SNPs except carrying the risk allele of SNP rs12923849, showed a protective effect (P=0.02), which would speak against this SNP being a causal variant. Among the not significantly associated haplotypes, were haplotypes that carry either the risk or protective allele for rs12708716 and rs9934231. They were also seen to always carry all nonrisk alleles of the three SNPs mapping to one LD block; rs7260912, rs806411 and rs6498169. Lastly, we performed logistic regression analyses on all top hit SNPs to further explore the dependency of these SNP associations. The other five top hit SNPs lost their significant association when conditioning on rs12708716, whereas rs12708716 remained highly significant when conditioning on the other top hit SNPs (Table 4). These data suggest that rs12708716 is the best candidate for being the causal variant in MS.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

LD plots of the SNPs in the replication phase on the basis of genotypes from all sample sets included in the combined analysis. (a) D′-LD. (b) R2-LD.

Full figure and legend (112K)

Expression analyses

We analysed CLEC16A RNA expression in 42 thymic tissue samples, and in 24 peripheral whole-blood samples from healthy control subjects. Not knowing which isoforms to expect in our thymus samples, we first used an assay covering all known transcripts, and found that CLEC16A transcripts were abundantly expressed both in thymic tissue and in peripheral blood. We went on to test the expression of two CLEC16A transcripts; comprising a short (21 exons) and a full-length protein variant (24 exons) in the thymus tissue and peripheral blood samples in view of their rs12708716 genotype. A third transcript is also described in Genome Browser (http://genome.ucsc.edu), however, this comprises only four exons, and was not further investigated. Four thymus samples and two peripheral blood samples were excluded because of cycle threshold (CT) s.d.>0.1. The remaining samples were divided into two groups: homozygotes for the risk allele (peripheral blood n=14, thymus n=17) and carriers of the protective allele (peripheral blood n=8, thymus n=21). The rationale for this division was to increase statistical power as few thymus samples (n=6) and no blood samples were homozygotes for the protective allele. No association with rs12708716 genotype was found looking at the expression levels of the assay comprising all isoforms nor for the short and long isoform separately. However, we found that the relative expression of the short and long isoform was significantly associated with rs12708716 genotype in the thymus samples (P=0.004), where a higher relative proportion of the shorter isoform versus the long isoform was found for the individuals harbouring the risk genotype AA. As we compared the relative CTs of the two isoform assays directly and not through a standard curve, we cannot tell whether this relationship is because of an increase in the short isoform or a decrease in the long isoform. The same relationship was not found for the peripheral blood samples (P=0.76) (Figure 2).

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Correlation between rs12708716 genotype and the relative distribution of the short- and long isoform. (a) peripheral blood and (b) Thymus. The y-axis represents (CT mean short isoform/CT mean long isoform). CT is inversely related to quantity, thus, there is a higher relative proportion of the short isoform in thymus samples with the risk genotype AA compared with AG and GG. No peripheral blood samples harboured the rarest genotype GG. P-values were calculated by Mann–Whitney U-test.

Full figure and legend (23K)



The results of our fine-mapping analysis of CLEC16A in a large sample set including 1113 MS trios, 3102 MS patients and 5047 healthy controls from Norway and the UK, indicated that the strongest CLEC16A SNP association in MS was the previously reported MS-associated SNP rs12708716. Association at the other SNPs tested seemed to be related to LD with rs12708716, with the risk allele from each of the other SNPs occurring on the same haplotype as the risk allele of rs12708716. Logistic regression showed that rs12708716 withstood conditioning on the five other top hit SNPs. Likewise the other SNPs lost significance conditioning on rs12708716. Altogether, among the SNPs tested the MS-associated SNP rs12708716 appeared to exert the superior association.

Haplotype analysis and LD-D′ showed that there had been little historical recombination on the main haplotype carrying the rs12708716 risk allele. However, the r2 was relatively low because of frequency differences between the associated alleles, providing sufficient power to demonstrate primary association to rs12708716 by logistic regression analyses. The rs12708716 SNP and the four other SNPs comprising the risk haplotype identified in our study are non-coding SNPs located in intron 10, 19, 20 and 22. Among these, rs12708716 is the best candidate for being the causal variant even though we cannot rule out variants in strong LD with this SNP. Interestingly, Todd and colleagues have performed extensive sequencing on the CLEC16A gene region, and found neither non-synonymous or rare variants within the gene, nor any variant in the neighbouring genes that overcome the association of the CLEC16A SNP rs12708716 in T1D.9, 18

Both up- and downstream of the CLEC16A gene are other interesting candidate genes for autoimmune disease, including the MHC2 transcription factor CIITA and the cytokine signalling suppressor SOCS1 gene (http://genome.ucsc.edu/). However, CLEC16A lies within its own LD block, and little D′ LD is seen with these two candidate genes on either side (www.hapmap.org). Thus, the disease association signal picked up by GWAS is likely to come from the CLEC16A gene itself. In our screening panel, we genotyped one SNP downstream of SOCS1, rs243329 (P=0.03), which has been found associated with T1D, and one SNP in CIITA, rs8048002 (P=0.71), which has shown association with Addison's disease.7, 9 Both SNPs should based on low LD (D′<0.35, r2<0.1), be independent of the CLEC16A-associated SNPs found in MS (see Supplementary Figure 1). Still, evidence of a role for SOCS1 and CIITA in MS is emerging. Recently, Zuvich et al.19 published a study showing association of a polymorphism in the SOCS1 gene associated in MS. Another study by Bronson et al.20 showed association between MS susceptibility and variation in CIITA in patients with the MS predisposing HLA-DRB1*1501 allele. Even if these associations are apparently independent of CLEC16A because of very limited LD between the associated polymorphisms in either gene,19, 20 polymorphisms within the CLEC16A gene could potentially affect expression levels of SOCS1 or CIITA, or vice versa. Thus, a role of all three genes in MS should be further investigated.

To date, only limited expression data on CLEC16A has been published. CLEC16A encodes a suggested immune receptor, and shows association with multiple autoimmune diseases. This could mean that mutations in the gene cause a general autoimmune dysfunction. As the thymus is relevant for the establishment of central tolerance, it is relevant to study the expression of CLEC16A in thymic tissue. We found association between rs12708716 genotype and the relative expression of a short (21 exons) and long (24 exons) CLEC16A transcript in thymus. In addition to fewer exons (lacks exon 22–24), the short transcript contains a 48bp deletion not found in the long transcript comprising the exon 10–11 boundary (http://genome.ucsc.edu). rs12708716 is situated in intron 19, whereas rs9934231, rs8060411, rs7206912 and rs6498169, which are also on the described MS-associated haplotype are in intron 10, 20 or 22. Thus, they cluster in some proximity to regions where differences are seen in the transcripts. Hakonarson et al.1 have previously studied CLEC16A expression in NK cells in view of the T1D-associated SNP rs2903692 and found that homozygotes for the A allele showed a trend towards increased CLEC16A mRNA expression (KIAA0350/CLEC16A Affymetrix probe set 231221 (Affymetrix, Santa Clara, CA, USA) ). Interestingly, this SNP rs2903692 is in strong LD (r2=0.88) with rs12708716. We saw no such direct relationship with genotype and expression level of CLEC16A transcript neither in thymus nor whole blood, nor did we find any effect on relative isoform expression in whole blood as seen in the thymus samples. Of note, thymus samples consist mainly of thymocytes, whereas peripheral whole-blood samples are very heterogenous in cellular composition. It is possible that with more detailed analysis of CLEC16A expression in subsets of white blood cells, the same relationship between CLEC16A transcripts and genotypes would be observed. Our results could therefore speak for thymus-specific and or cell-specific regulatory effect on splicing mechanisms being picked up by rs12708716.

In conclusion, in the most extensive fine-mapping study of the CLEC16A gene in MS performed to date, we find a significant association between MS susceptibility and rs12708716. Furthermore, the rs12708716 genotype is associated with the relative expression of CLEC16A transcripts in thymus, but not in whole blood, suggesting thymus-specific splicing regulation being involved in the expression of this plausible immune receptor gene.


Materials and methods


The Norwegian patient sample sets were gathered in the Oslo MS registry and biobank (n=619) and the Norwegian MS registry and biobank (n=1394), and were sampled in Oslo with surrounding areas and nationwide, respectively. In the screening phase, we genotyped 807 Norwegian MS patients from the Oslo- and the Norwegian MS registry and biobank, and 1027 healthy controls from the Norwegian bone marrow registry. Of these, 644 patients and 1023 controls had previously been genotyped for rs12708716 in an IMSGC study.4 Norwegian replication samples consisted of 1206 MS patients from the Norwegian MS registry and biobank, and 2064 Norwegian blood donors as healthy controls. Additionally, 22 duplicates were genotyped within the screening phase and 245 duplicates were typed in both the screening phase and replication phase for quality control.

The British replication sample set consisted of 1153 trios, 1102 UK MS cases, 1980 controls from the 1958 British birth cohort (www.b58cgene.sgul.ac.uk) and 277 duplicates. The whole UK sample set had previously been genotyped for rs12708716 as part of an IMSGC study,4 whereas 1350 of the patients and all UK controls had been included in the IMSGC GWAS in 2007, thereby being genotyped for rs6498169.2All patients included in the study were diagnosed according to Poser and/or McDonald criteria.21, 22 Demographic information of the patient sample sets is given in Table 5. In the expression analyses, we used peripheral whole blood samples from healthy bone marrow donors, and thymic tissue samples obtained from 42 Norwegian children under the age of 13 undergoing corrective cardiac surgery at Rikshospitalet. Whole blood was collected on Tempus tubes (Applied Biosystems, Foster city, CA, USA), and thymic tissue samples were collected in RNAlater solution (Ambion, Austin, TX, USA) immediately during surgery.

Informed consent was achieved from all participants or their parents (for the thymus samples), and the study was approved by the appropriate local ethical authorities.

SNP panel design and genotyping

The SNP selection for the initial screening phase was done in three stages. The majority of the screening samples (644 cases and 1023 controls) had previously showed nominally significant association with rs12708716 (P=0.005) in an IMSGC study.4 We therefore chose to further explore putative associations represented by this known associated SNP rs12708716, and selected a panel of SNPs in moderate-to-strong LD with rs12708716 (r2>0.6, MAF>0.01). We then used a tagging approach among these SNPs to remove redundant SNPs in almost complete LD (r2>0.95). This left us with 14 SNPs representing 59 SNPs in moderate LD with the reference SNP rs12708716. Second, we incorporated these 14 SNPs by force inclusion into a general mapping panel of the gene (r2>=0.8 and MAF>0.05, capturing 241 SNPs). In this general map, we also forced to include SNPs of particular interest, like possible splice site SNPs or SNPs reported to be associated with other autoimmune disease, including rs6498169. As the CLEC16A gene is fairly contained within its own LD block, making the previously reported CLEC16A association signals likely to originate from genetic variant(s) within the gene itself (www.hapmap.org), we confined our SNP selection area to either end of the longest flanking distance of the CLEC16A gene chr16p13:10945–11184kb. (dbSNP build 129 SNP track for the Human March 2006 assembly (http://genome.ucsc.edu/). Finally, we added SNPs of special interest from the neighbourhood genes SOCS1 and CIITA or the intergenic region separating these genes. The inclusion criterion for each SNP included in the screening panel is given in Supplementary Table 1.

For the SNP selection, we used the Tagger programme23 incorporated in Haploview v 4.1,12 and SNP genotype data from HapMap release 21, CEU analysis panel (samples collected from people living in Utah with ancestry from northern and western Europe).13 The sequences for primer design were obtained from dbSNP build 129, SNP track for the Human March 2006 (hg18) assembly at the UCSC website (http://genome.ucsc.edu/). Our final panel of 57 SNPs was genotyped by Sequenom technology on the Centre for interactive genetics; Cigene, Norwegian University of Life Sciences (UMB) AAs.

Replication genotyping of our five top hits from the screening phase, along with the already described associated SNP rs12708716 was performed by TaqMan technology. (Applied Biosystems). As the TaqMan assay did not work for one of the top hit SNP (rs6498168), we replaced this SNP with one in complete LD (r2=1; rs8060411), which was then genotyped in both the screening cohort and the replication cohorts. A total of 1029 Norwegian blood donors were previously genotyped for rs12708716 by Skinningsrud et al.7, 8 For the 1958 birth controls, SNPs rs9934231, rs12708716, rs6498169 and rs12923849, had already been genotyped by the Wellcome Trust Case-Control Consortium and, hence, their genotypes were obtained through the Wellcome Trust Case-Control Consortium website (www.wtccc.org.uk). Further, SNP rs12708716 has previously been genotyped in the complete UK cohort .and does not represent new data.4

Expression analyses

Whole-blood RNA from healthy Norwegian bone marrow donors was isolated using the Tempus spin RNA kit and included a DNAse digestion step by AbsoluteRNA Wash Solution (Applied Biosystems). Complementary DNA was produced by the reverse transcriptase kit provided by Eurogentec (Eurogentec, Liege Science Park, Seraing, Belgium). Thymus RNA isolation was performed with TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocols from whole tissue biopsies from human thymi. The complementary DNA synthesis was performed on total RNA by employing SuperScrip III reverse transcriptase (Invitrogen, Cat. No: 18080-051) and random hexamers with ~1000ng input of total RNA. DNAse digestion was performed with DNaseI (RNase-free) M0303S (BioLabs New England, Ipswich, MA, USA).

An assay comprising known CLEC16A RNA transcripts, and differential expression of short isoform BC112897 (21 exons) and full-length peptide NM_015226.1/AB002348.3 (24 exons) were measured by real-time PCR on an ABI7900HT system using predesigned gene expression assays available from Applied Biosystems. All assays had primers covering exon–exon transition to avoid DNA contamination. B-Rajii RNA was isolated using the RNeasy Mini Kit by Quiagen Inc. (Valencia, CA, USA) and used as positive control for the two isoforms. Expression of the two transcripts in B-Rajii cells was before the real-time experiment confirmed by PCR and gel electrophoresis. RNase P was used as endogenous control/house keeping gene and corrected for when looking at either transcripts alone. Triplicates were used for all samples included in the experiment. Where one replicate was clearly an outlier, this was omitted to leave a comparison of the two closest replicates. CT standard deviation cut off was set to 0.1. As the sample input for the measurement of long- and short isoform were the same, CT mean was measured directly and not through RNase P.

Statistical analyses

The data quality control including PED CHECK of the trio samples, genotype success rates and Hardy–Weinberg Equilibrium was performed in Plink v 1.06.24 Haploview v 4.1 was used for LD plot analyses.12 Association testing, conditional logistic regression analyses and haplotype analyses were carried out in Unphased v Haplotype frequency threshold was set to 5%. In the combined analyses, we found significant heterogeneity for several SNPs between the UK and Norwegian population. This was corrected for in all the combined analyses, by setting nationality as confounder. Two of the SNPs (rs12708716 and rs6498169) on the identified MS susceptibility haplotype have previously been confirmed associated with MS at genomewide significance level, and our primary goal for combined data analyses was to single out a primary association. Hence, the P-values are presented noncorrected in the combined material. The comparison of CLEC16A expression and genotype was done by comparing the CT mean distribution among the homozygotes for the risk allele versus heterozygotes and noncarriers for the risk allele by Mann–Whitney U-test performed in GraphPad. Prism v. 5.01 (GraphPad Software Inc., San Diego, CA, USA).


Conflict of interest

The authors declare no conflict of interest.



  1. Hakonarson H, Grant SF, Bradfield JP, Marchand L, Kim CE, Glessner JT et al. A genome-wide association study identifies KIAA0350 as a type 1 diabetes gene. Nature 2007; 448: 591–594.
  2. International Multiple Sclerosis Genetics Consortium (IMSGC). Risk alleles for multiple sclerosis identified by a genomewide study. N Engl J Med 2007; 357: 851–862.
  3. Wellcome Trust Case Control Consortium. Genome-wide association study of 14000 cases of seven common diseases and 3000 shared controls. Nature 2007; 447: 661–678.
  4. International Multiple Sclerosis Genetics Consortium (IMSGC). The expanding genetic overlap between multiple sclerosis and type I diabetes. Genes Immun 2009; 10: 11–14.
  5. Marquez A, Varade J, Robledo G, Martinez A, Mendoza JL, Taxonera C et al. Specific association of a CLEC16A/KIAA0350 polymorphism with NOD2/CARD15(-) Crohn's disease patients. Eur J Hum Genet 2009; 17: 1304–1308.
  6. Rubio JP, Stankovich J, Field J, Tubridy N, Marriott M, Chapman C et al. Replication of KIAA0350, IL2RA, RPL5 and CD58 as multiple sclerosis susceptibility genes in Australians. Genes Immun 2008; 9: 624–630.
  7. Skinningsrud B, Husebye ES, Pearce SH, McDonald DO, Brandal K, Boe WA et al. Polymorphisms in CLEC16A and CIITA at 16p13 are associated with primary adrenal insufficiency. J Clin Endocrinol Metab 2008; 93: 3310–3317.
  8. Skinningsrud B, Lie BA, Husebye ES, Kvien TK, Forre O, Flato B et al. A CLEC16A variant confers risk for juvenile idiopathic arthritis and anti-CCP negative rheumatoid arthritis. Ann Rheum Dis 2009; 69: 1471–1474.
  9. Todd JA, Walker NM, Cooper JD, Smyth DJ, Downes K, Plagnol V et al. Robust associations of four new chromosome regions from genome-wide analyses of type 1 diabetes. Nat Genet 2007; 39: 857–864.
  10. Geijtenbeek TB, Gringhuis SI. Signalling through C-type lectin receptors: shaping immune responses. Nat Rev Immunol 2009; 9: 465–479.
  11. Hoppenbrouwers IA, Aulchenko YS, Janssens AC, Ramagopalan SV, Broer L, Kayser M et al. Replication of CD58 and CLEC16A as genome-wide significant risk genes for multiple sclerosis. J Hum Genet 2009; 54: 676–680.
  12. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005; 21: 263–265.
  13. International HapMap Consortium 2007. A second generation human haplotype map of over 3.1 million SNPs. Nature 2007; 449: 851–861.
  14. Martinez A, Perdigones N, Cenit M, Espino L, Varade J, Lamas JR et al. Chromosomal region 16p13: further evidence of increased predisposition to immune diseases. Ann Rheum Dis 2010; 69: 309–311.
  15. Zoledziewska M, Costa G, Pitzalis M, Cocco E, Melis C, Moi L et al. Variation within the CLEC16A gene shows consistent disease association with both multiple sclerosis and type 1 diabetes in Sardinia. Genes Immun 2009; 10: 15–22.
  16. Bronson PG, Ramsay PP, Seldin MF, Gregersen PK, Criswell LA, Barcellos LF. A candidate gene study of CLEC16A does not provide evidence of association with risk for anti-CCP-positive rheumatoid arthritis. Genes Immun 2010; 11: 504–508.
  17. Perera D, Stankovich J, Butzkueven H, Taylor BV, Foote SJ, Kilpatrick TJ et al. Fine mapping of multiple sclerosis susceptibility genes provides evidence of allelic heterogeneity at the IL2RA locus. J Neuroimmunol 2009; 211: 105–109.
  18. Nejentsev S, Walker N, Riches D, Egholm M, Todd JA. Rare variants of IFIH1, a gene implicated in antiviral responses, protect against type 1 diabetes. Science 2009; 324: 387–389.
  19. Zuvich RL, McCauley JL, Oksenberg JR, Sawcer SJ, De Jager PL, Aubin C et al. Genetic variation in the IL7RA/IL7 pathway increases multiple sclerosis susceptibility. Hum Genet 2010; 127: 525–535.
  20. Bronson PG, Caillier S, Ramsay PP, McCauley JL, Zuvich RL, De Jager PL et al. CIITA variation in the presence of HLA-DRB1*1501 increases risk for multiple sclerosis. Hum Mol Genet 2010; 19: 2331–2340.
  21. 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: 121–127.
  22. 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: 227–231.
  23. de Bakker PI, Yelensky R, Pe’er I, Gabriel SB, Daly MJ, Altshuler D. Efficiency and power in genetic association studies. Nat Genet 2005; 37: 1217–1223.
  24. 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: 559–575.
  25. Dudbridge F. Likelihood-based association analysis for nuclear families and unrelated subjects with missing genotype data. Hum Hered 2008; 66: 87–98.


We would like to thank all MS patients and healthy controls for their participation in the study. The study is funded by grants from The South-Eastern Norway Regional Health Authority, the Research Council of Norway; Scientific Advisory Council Ullevål, Oslo University Hospital; Norwegian Foundation for Health and Rehabilitation, the Norwegian Diabetes Association, the Oslo MS Association and the Odd Fellow MS society. We thank all contributors to the collection of samples and clinical data in the Norwegian MS Registry and Biobank. The Norwegian MS Registry and Biobank is supported by the Research Council of Norway, Haukeland University Hospital and Western Norway Regional Health Authority. The Norwegian Bone Marrow Donor Registry, Rikshospitalet, Oslo University Hospital are acknowledged for providing Norwegian controls. Harald Lindberg is thanked for collection of thymus samples. The Centre for interactive genetics; Cigene, Norwegian University of Life Sciences (UMB) AAs is thanked for performing Sequenom analyses. The University of Cambridge and colleagues at the Department of Clinical Neuroscience Addenbrooke's hospital are acknowledged for laboratory assistance and much appreciated collaboration in connection with the research stay of IL Mero in this department. This study makes use of data generated by the Wellcome Trust Case-Control Consortium. A full list of the investigators who contributed to the generation of the data is available from www.wtccc.org.uk. Funding for the project was provided by the Wellcome Trust under award 076113. We acknowledge use of DNA from the British 1958 Birth Cohort collection, funded by the Medical Research Council Grant G0000934 and the Wellcome Trust Grant 068545/Z/02. This work was also supported by the Medical Research Council (G0700061), the Wellcome Trust (084702/Z/08/Z) and the Cambridge NIHR Biomedical Research Centre.

Supplementary Information accompanies the paper on Genes and Immunity website