The CD58 protective allele increases CD58 RNA expression. (A) To illustrate the association of the rs2300747^G marker to higher CD58 expression, we have plotted the distribution of RNA expression values by genotype and show the results of pairwise comparisons between genotype categories. The location of the mean value for each category is denoted by a black line. We present the results of the genotypic comparisons among the 89 LCLs of East Asian ancestry that offer a much more robust estimate of the allele's effect given the higher frequency (0.66) of rs2300747^G in this population. The allele frequency of rs2300747^G is much lower (0.13) in the smaller collection of 60 LCLs of European ancestry; this precludes a robust pairwise assessment of genotype categories. The cell surface expression of CD58 in these cell lines shows the same pattern of association to genotype—higher surface expression in rs2300747^GG subjects—but does not reach statistical significance given the limited number of cell lines that are available (data not shown). (B) To validate the initial observation obtained from the LCL data, we present evidence that the correlation of higher CD58 RNA expression with rs2300747^G is also seen ex vivo in RNA data obtained from mononuclear cells of subjects with remitting-relapsing MS or CIS. Since a single MS subject had the rare rs2300747^GG genotype, this subject was pooled with the 29 rs2300747^AG heterozygotes (green) and compared to the 209 rs2300747^AA homozygotes (red) with MS or CIS. The location of the mean value for each category is denoted by a black line.

CD2 costimulation strongly induces FoxP3 expression in regulatory T cells (Treg). (A) Cultures of CFSE-labeled responder T cells with or without CD4⁺CD25^high regulatory T cells, stimulated with αCD3/αCD2 (Top) or αCD3/αCD28 (Bottom) mAbs demonstrated 39% and 31% suppression at day 4, respectively. The percentage of cells not diluting CFSE is shown in parentheses. (B) Shown here is ex vivo staining for intracellular FoxP3of CD4⁺CD25⁻ responder T cells and the CD4⁺CD25^high regulatory T cells to demonstrate the purity of the initial pool of cells; 92% of cells are FoxP3⁺. (C) Cultures of responder T cells only (only CFSE^high/low) or regulatory T cell cocultures (CFSE^neg CD4⁺CD25^high regulatory T cells and CFSE^high/low target responder T cells) were stained for intracellular FoxP3 after 4 days of culture. Expression of FoxP3 is shown on the gated responder T cells (CFSE^high/low) or regulatory T cells (CFSE^neg). Percent positive cells as compared to isotype control antibodies and mean fluorescence activity (MFI) are shown. These data are representative of 8 independent experiments performed. (D) On average, the expression of FoxP3 in CFSE^neg CD4⁺CD25^high regulatory T cells after 4 days of culture is 2.1-fold greater following anti-CD2 costimulation when compared to anti-CD28 costimulation. The mean fluorescence intensity of FoxP3 expression is shown for each tested individual under each stimulation condition. Additional details are provided in Table S6 which demonstrates that the expression of FoxP3 increases in a dose-dependent manner following anti-CD2 costimulation.

The minor alleles of the 2 SNPs with the lowest association P-values are found on the same haplotype that spans most of the CD58 gene. The segment of chromosomal DNA under study is shown in black, with its physical boundaries noted based on human genome assembly 18; the location of the CD58 gene is shown in light blue. The flanking genes are IGSF3 (centromeric, navy blue) and ATP1A1 (telomeric, outside of the segment shown here). Below, we show the 31 SNPs and 1 insertion/deletion polymorphism (indel: −/T) found within this chromosomal segment that we have genotyped. These polymorphisms are divided into 3 distinct blocks of linkage disequilibrium based on genotyping data from both cases and healthy control individuals from the Brigham and Women's Hospital collection (see Fig. S1 for the detailed linkage disequilibrium structure and list of haplotypes found in this region). All haplotypes in the second block of linkage disequilibrium with a frequency >0.05 in the Brigham and Women's Hospital samples are shown with their unique sequences of alleles. The ancestral alleles of each polymorphisms, as recorded in the HapMap (27) or dbSNP resources (http://www.ncbi.nlm.nih.gov/projects/SNP/), are used to determine the ancestral haplotype, which serves as the reference haplotype, at the top of the figure. The derived (non-ancestral) allele of a SNP is highlighted by a navy blue circle on each haplotype. The minor alleles of the 2 SNPs (rs12044852 and rs2300747; Table S2) with a P value < 0.0005 in our primary association analysis to MS are highlighted in yellow; both minor alleles are on the same haplotype that is found at a frequency of 8% in the Brigham and Women's Hospital samples. Their location is highlighted in red. It is interesting to note that this 8% haplotype, which is underrepresented in subjects with MS, is the most divergent of the haplotypes with a frequency >5% in subjects of European ancestry studied here. The frequency of each haplotype is noted to the right, and the dendritic tree illustrates a possible evolutionary relationship of the different haplotypes.

Expression of CD58 RNA is enhanced during the remission phase of MS in untreated subjects. We compare the expression of CD58 RNA found in whole blood during the clinically defined remission and relapse phases of MS. For each phase, 10 unique untreated subjects with MS were sampled, and individual values of CD58 RNA expression were normalized to the mean CD58 RNA expression of 20 healthy control subjects. On average, the expression of CD58 increases 1.7-fold in whole blood during a relapse relative to the expression found in 20 healthy control subjects, and the distribution of the relative change observed in the remission phase is significantly different from that seen during the relapse phase (P = 0.011). This difference remains significant after removing the extreme outlier of each category (relapse and remission) (P = 0.040).