The CD6 Multiple Sclerosis Susceptibility Allele Is Associated with Alterations in CD4⁺ T Cell Proliferation

Multiple sclerosis (MS) is an inflammatory disease of the CNS that shares elements of its genetic architecture with other autoimmune diseases (1). Pathologically, there are perivenular infiltrates of CD4⁺ and CD8⁺ T cells in the CNS white matter and meninges with demyelinating lesions and loss of axons in both white and gray matter (2, 3). There are also marked changes in systemic immune function with loss of regulatory T cell (Treg) function and increases in myelin-reactive CD4⁺ T cells (4–6). Interestingly, unbiased genome-wide association scans have identified susceptibility loci almost exclusively in regions containing genes with immune function, including CD6, CD25, CD40, CD58, HLA A, HLA B, HLA-DRB1, IL2RA, IL7R, IL12A, IRF8, STAT3, and TNFRSF1A (7–11). The identification of susceptibility loci is a major step forward in MS, but we do not yet understand how allelic variation influences immune function. This transition remains a major challenge in the understanding of MS and of other human autoimmune diseases.

CD6 is a 130-kDa type 1 transmembrane glycoprotein expressed on the surface of CD4⁺ and CD8⁺ T cells and, to a lesser extent, on B cells. CD6 stimulation plays an important role in the maintenance of T cell activation (12–19) because blocking interactions between CD6 and its ligand activated leukocyte cell adhesion molecule (ALCAM) result in diminished proliferation of T cells (12). The ALCAM-binding site is located at the membrane-proximal scavenger receptor cysteine-rich (SRCR) domain 3 of CD6 and is present in most of the five intracellular and three extracellular isoforms of CD6 that have been described (20–22). The CD6Δd3 splice variant, created by excluding exon 5, is lacking the membrane-proximal SRCR domain 3 with the ALCAM-binding site (23). The MS risk allele rs17824933^G is located in intron 1 of the CD6 gene (7) and is found at a frequency of 44% of subjects of European ancestry from the PhenoGenetic Project, a cohort of >1200 healthy subjects recruited from the Greater Boston Metropolitan area that supports the functional characterization of genetic variation; subjects homozygous for the risk allele (with an rs17824933^GG genotype) are found at 7.6% frequency in this population.

The Brigham and Women’s Hospital PhenoGenetic Cohort with a collection of 1200 healthy control subjects served as basis for subject recruitment and sample collection. These subjects were genotyped and had the following characteristics: female/male sex ratio was 60:40%; race distribution was 14% African American, 12% Asian American, 68% white, and 6% Hispanic; smoking history was 82% never smokers, 9% former smokers, and 9% smokers; mean age was 24.3 y (range, 18–50 y); mean body mass index was 22.5 (range, 13–50); and MHC was 51% HLA A2, 20% HLA A3, 23% HLA-DR2, and 20% HLA-DR1 and -DR4. All subjects have been comprehensively genotyped for allelic variants related to susceptibility to autoimmune diseases. The study was conducted in compliance with the Declaration of Helsinki. Approval was obtained from the local ethics committee before study initiation, and written informed consent was obtained from all patients before performing any study procedures.

Primary human lymphocytes were isolated from peripheral blood of healthy donors by Ficoll gradient, and CD4⁺ and CD8⁺ T cells were purified by MACS separation. CD4⁺ and CD8⁺ T cells were cultured in RPMI 1640 medium, 10% (v/v) FCS (Life Technologies, Paisley, U.K.).

Healthy control subjects of the Brigham and Women’s Hospital PhenoGenetic cohort were genotyped on the Affymetrix Genome-wide Human SNP Array 6.0 (Genechip 6.0) at the Broad Institute’s Center for Genotyping and processed for quality control using the PLINK software suite. We applied its standard quality-control pipeline for subjects (genotype success rate >95%, sex concordant, excess interheterozygosity/intraheterozygosity) and for single nucleotide polymorphisms (SNPs) (Hardy–Weinberg equilibrium p > 1 × 10⁻⁶; minor allele frequency > 0.01, genotype call rate > 0.95; mishap test > 1 × 10⁻⁹) to these data.

The complete results of the genome-wide association analysis are available to search by SNP id, chromosome, or base pair location on the International Multiple Sclerosis Genetics Consortium Web site (http://www.imsgc.org).

Full-length CD6 and the CD6Δd3 isoform were amplified by PCR as previously described (13). The following primers were used for amplification of full-length CD6: 5′-TAGTAGTCTAGAAGCCAGAGACAGCTCCAGAC-3′ (forward) and 5′- TAGTAGGTCGACCGAGATTCCTTGTTC-3′ (reverse). The CD6Δd3 isoform was amplified using the following primers: 5′-TAGTAGTCTAGATTGTCGACATCAGAGCACCAGTCC-3′ (forward) and 5′-TAGTAGCTCGAGCGAGATTCCTTGTTCTC-TATT-3′ (reverse). Total CD6 was analyzed using commercial probes (Hs00198752_m1; Applied Biosystems) specific for the exon 1/2 boundary (5′-AGGTCATCCATCTCCAGCCCCACCTGACCAGCTCAACACCAGCAGTGCAGAGAGTGAGCTGGCTTCCGGTCC-3′), and CD6 isoforms containing exons 8 and 9 were analyzed using commercial probes (Hs00945716_m1; Applied Biosystems) specific for the exon 8/9 boundary (5′-AAAGAAGTTTTCATGCTGCCCATCCAGGTCCAGGCCCCGCCCCCTGAG GACTCAGACT-3′). Relative fold changes of mRNA expression were calculated by the ΔΔcycle threshold method, and the amount of target gene was normalized to GAPDH (Applied Biosystems).

Full-length CD6 expression in CD4⁺ T cells was selectively knocked down using a transient direct small interfering RNA (siRNA) delivered by electroporation nucleofection (Amaxa), according to the manufacturer’s instructions. Nonspecific siRNA was used as control. Expression levels of full-length CD6 and the CD6Δd3 isoform were analyzed by real-time PCR.

CD4⁺ T cells were cultured in 96-well, round-bottom tissue culture plates at a density of 2 × 10⁴ cells/well in RPMI 1640 medium, 10% (v/v) FCS (Life Technologies), and stimulated with 4.5-mm polystyrene beads (Dynal/Invitrogen) covalently linked to anti-CD3 Abs (0.5 μg/ml), anti-CD28 Abs (0.5 μg/ml), and recombinant human ALCAM-Fc fusion protein (rhALCAM-Fc; 1 μg/ml). Cell proliferation was assessed at day 4 by [³H]thymidine incorporation.

The PE-conjugated and FITC-conjugated anti-CD3, anti-CD4, anti-CD8, and anti-CD6 mAbs were purchased from BD Pharmingen (San Diego, CA). The allophycocyanin-conjugated anti-CD4 mAb was purchased from Dako North America (Carpinteria, CA). The goat anti-human IgG Ab and its biotin-, FITC-, and PE-conjugated F(ab′)₂ fragment derivatives were purchased from Dako North America. Living cells were gated using Live/Dead fixable dead cell stain kit (Molecular Probes, Eugene, OR). Labeled cells were analyzed on a FACSCanto cytofluorometer equipped with the FACS Diva research software (BD Bioscience, San Diego, CA).

T cells were cultured in the presence of the anti-CD3 mAb OKT3 (0.5 μg/ml) and anti-CD28 mAb 15E8 (0.5 μg/ml). Where indicated, cells were incubated with 1 μg/ml anti-CD6 mAb M-T605 or 1 μg/ml isotype control. After 3 or 6 d, supernatants were collected and cytokine production was determined by ELISA as previously described (24). Matched pair capture and biotinylated detection Abs for determination of IFN-γ (NIB42, B133), IL-10 (JES3-19F1, JES3-12G8), and IL-13 (JES10-5A2, B69-2) were purchased from BD Bioscience. Recombinant human IFN-γ, IL-10, and IL-13 were purchased from Chiron (Ratingen, Germany).

Freshly isolated CD4⁺ T cells were labeled with 5 μM CFSE (Invitrogen, Karlsruhe, Germany) for 10 min and washed five times with PBS/FBS. CSFE-labeled CD4⁺ T cells (5 × 10⁴ cells/well) were cultured in the presence of the anti-CD3 mAb OKT3 (0.5 μg/ml) and anti-CD28 mAb 15E8 (0.5 μg/ml). Where indicated, cells were incubated with 1 μg/ml anti-CD6 mAb M-T605 or 1 μg/ml isotype control. After 3 and 6 d, CSFE dilution was analyzed by flow cytometry, and the number of cycling and noncycling CFSE-labeled T cells was determined.

Unpaired Student t test was used to compare RNA expression, cell proliferation, and cytokine production between T cells from donors with CC, CG, or GG genotype. Statistical analysis was performed using GraphPad Prism (GraphPad Software, La Jolla, CA).

In this study, we examined the correlation of the CD6 rs17824933^G allele with changes in the immune function of healthy subjects. We use healthy subjects to investigate correlations between MS susceptibility alleles and immunophenotype because the phenotype and function of immune cells in patients with autoimmune diseases are likely to be confounded by immunotherapies or by effects of the inflamed state of autoimmune disease. PhenoGenetic Project subjects are self-reported to be free of autoimmune diseases or chronic infectious diseases; they were genotyped for rs17824933, and subjects were recalled for blood donation based on their genotype. The characteristics of the donors who were recalled for the functional assays are summarized in Supplemental Table I.

Using real-time PCR, we analyzed the expression pattern of CD6 isoforms in primary CD4⁺ and CD8⁺ T cells obtained from the donors. The relative expression of the CD6Δd3 isoform is significantly higher on CD4⁺ and CD8⁺ T cells isolated ex vivo from subjects homozygous for the risk allele (rs17824933^GG) when compared with the same cells extracted from subjects homozygous for the protective allele (rs17824933^CC). Although total CD6 expression does not significantly differ between the different groups of subjects, the relative expression of full-length CD6 is diminished in CD4⁺ and CD8⁺ T cells from rs17824933^GG subjects (Fig. 1). A decrease in full-length CD6 mRNA correlates with increased levels of CD6Δd3 mRNA (Supplemental Fig. 1). Furthermore, the overall expression of CD6 on the cell surface was analyzed by flow cytometry and does not differ between the two groups of subjects (Supplemental Fig. 2). The expression patterns of other known CD6 isoforms are not affected by the risk allele (Supplemental Fig. 3).

It has been previously shown that the CD6 ligand, ALCAM, is induced on CD4⁺ T cells after activation, and blocking this interaction with inhibitory anti-CD6 mAb inhibits sustained entry into cell cycle (12, 25–30). Thus, we activated highly purified CD4⁺ T cells from donors with the rs17824933^CC, rs17824933^CG, or rs17824933^GG genotypes for 3 and 6 d with anti-CD3 and anti-CD28 mAb. Proliferation of the activated CD4⁺ T cells was assessed by CSFE dilution. Although no differences in proliferation of CD4⁺ T cells from donors of different genotypes was observed on day 3 of activation, a significant decrease in proliferation was observed in CD4⁺ T cells from rs17824933^GG subjects (58.4% proliferation) as compared with rs17824933^CC subjects (78.8% proliferation; p = 0.01) at day 6 (Fig. 2). Thus, a simple measure of entry into cell cycle by CD4⁺ T cells is influenced by the CD6 genotype.

We hypothesized that the decreased proliferative response in rs17824933^GG subjects was due to the relative increase of the CD6Δd3 splice variant lacking the membrane-proximal SRCR domain 3 with the ALCAM-binding site (23). To directly explore this mechanism, we established a model system with CD4⁺ T cells exclusively expressing the risk CD6Δd3 isoform without expression of full-length CD6. Expression of isoforms containing exon 5 in CD4⁺ T cells of subjects with rs17824933^CC genotype was specifically inhibited using exon 5-specific siRNA (Fig. 3A), thus increasing the relative proportion of CD6Δd3 to CD6 isoforms; as a control, CD4⁺ T cells from the same subjects were transfected with randomly generated control siRNA, allowing the comparison of identical CD4⁺ T cells differing only in the expression level of CD6 isoforms containing exon 5. Using this model system, we found that stimulation with a rhALCAM-Fc induced significant lower entry into cell cycle of CD4⁺ T cells in the context of the exon 5-specific siRNA construct versus CD4⁺ T cells derived from the same donor and transfected with control siRNA (p = 0.0111; Fig. 3B). These data demonstrate that increasing the relative proportion of the CD6Δd3 isoform in vitro recapitulates the diminished proliferative capacity of T cells seen in rs17824933^GG subjects who also have an increased proportion of the CD6Δd3 isoform.

We then examined the proliferative response of CD4⁺ T cells to anti-CD3 and anti-CD28 in the context of inhibition of the CD6-ALCAM costimulatory pathway, postulating that inhibitory effects would be observed only in subjects expressing high relative amounts of CD6 isoforms with the ALCAM-binding domain. Specifically, we activated highly purified CD4⁺ T cells from donors with each of the three rs17824933 genotypes for 3 and 6 d with anti-CD3 and anti-CD28 mAbs. The interaction between CD6 on CD4⁺ T cells and its ligand ALCAM on activated PBMCs was blocked with the CD6-specific Ab M-T605 as previously described (12). Although the M-T605 binding domain is located in the extracellular SRCR domain 1 and is not close to the ALCAM-binding domain in the extracellular SRCR domain 3, the assembly of the CD6 molecule provides an explanation for the prevention of ALCAM binding by M-T605: binding of M-T605 to domain 1 results in obstruction of the access to the ALCAM-binding site (Fig. 4A) (31). Inhibitory effects were observed only in subjects expressing high relative amounts of the CD6 isoforms with the ALCAM-binding domain. Blocking CD6 reduced the proliferation of CD4⁺ T cells from rs17824933^CC subjects by 27.3%, whereas no effect was seen in CD4⁺ T cells from subjects with one (0.6% inhibition; p = 0.0001) or two (0.4% inhibition; p = 0.0001) rs17824933^G alleles, indicating that proliferation of CD4⁺ T cells with CG or GG genotype is not influenced by CD6 stimulation (Fig. 4B, 4C). Similar differences were seen in IFN-γ production after blocking CD6 on CD4⁺ T cells from subjects of the different groups of genotype (Fig. 4D).

MS is an inflammatory neurodegenerative disease with complex causative factors. The mechanisms underlying the pathogenesis of MS are not yet completely known. Infiltrations of CD4⁺ and CD8⁺ T cells in the CNS white matter are involved in the autoinflammatory process (2). Recent data suggest a possible role for autoreactive Th1 and Th17 cells in mediating tissue inflammation and demyelination in the CNS (32). Although there is a potential paradox in that the risk allele results in less activation of CD4⁺ cells, this may be explained by the presence of CD4⁺ T cell populations that both induce and suppress immune responses. Specifically, there are two major regulatory CD4⁺ T cell populations: Tregs (FOXP3⁺CD127^lowCD25^hi) and the type 1 Treg cells population expressing IL-10. We have described defects in both Tregs and type 1 Treg cells in patients with MS (33, 34). Depletion of the Tregs in either experimental animal models or in humans with FOXP3 gene deletion results in spontaneous autoimmune disease. Tregs bind to stromal tissue via CD6 ligation to CD166 that is overexpressed in the endothelium of patients with MS (35); thus, loss of the ALCAM-binding domain could prevent Tregs from entering the CNS, thereby providing a possible mechanism by which the risk allele may lead to loss of suppressor immune function. However, direct in vivo experiments will be required to understand which effect ultimately dominates.

We currently do not know what the causal variant may be; rs17824933 is simply the best marker for the causal variant in our MS susceptibility analyses. A review of SNPs in linkage disequilibrium with rs17824933 (r² > 0.5 in HapMap Caucasians of European origin in Utah) did not reveal a variant with a clear potential to affect splicing of exon 5. Dedicated fine-mapping efforts will be needed to identify the set of genetic variation whose effect could be captured by rs17824933. Nevertheless, it is clear that this rs17824933 tags a haplotype of CD6 that is associated not only with susceptibility to MS but also with evidence of variation in the expression of CD6 and downstream function.

The genetic component in MS is indicated by a 20–40% increased relative risk for monozygotic twins of affected individuals (36). Moreover, siblings of affected individuals exhibit a 15–20% increased relative risk for development of MS (37). Although autoimmune diseases are more likely to be triggered by complex genetics than by a single genetic variation, not all genes associated with autoimmune diseases might be causal genes contributing to the onset of disease. Deeper insights into the functional consequences of genetic variations are crucial for the elucidation of how genes contribute to the development of autoimmune diseases. To our knowledge, our results are the first report of a direct relationship between disease-associated CD6 SNP and T cell function. Extension of this approach to other susceptibility alleles is required to enable the elucidation of pathways contributing to the onset of inflammatory diseases.

We thank I. Wood for donor coordination and blood sample collection.

The authors have no financial conflicts of interest.