We genotyped 17 SNPs located within the IL21R gene in two independent, ethnically divergent populations of lupus patients and controls. Our European-derived cohort included 2,573 independent lupus cases and 3,075 healthy controls while our Hispanic cohort included 657 independent lupus cases and 265 healthy controls. Principal component analyses were used to examine sample homogeneity in both cohorts. A total of 234 and 156 samples were identified as outliers in the European-derived and Hispanic cohorts, respectively, and were excluded from further analysis. Subsequently, GC analysis was performed and revealed no significant population substratification (λ= 1.13 and 1.17, in the European-derived and Hispanic cohorts, respectively). Fifteen of the 17 SNPs genotyped had a minor allele frequency of at least 1%, which we required for analysis. Genotyping success rate was ≥99.2%, and ≥97.8% in the European-derived and Hispanic cohorts, respectively.
We found a genetic association between rs3093301 (A/G) and lupus in both the European-derived and the Hispanic cohorts. The lupus-associated allele (A) had a frequency of 66.6% among European-derived lupus cases compared to 63.7% among controls (OR= 1.13, χ2 = 9.42, p=0.0022) (, ). This association was confirmed in the Hispanic cohort, which had an associated allele frequency of 64.4% in lupus cases compared to 55.9% in controls (OR= 1.43, χ2 = 9.16, p=0.0025) (, ). Using Cochran-Mantel-Haenszel test, the meta-analysis odds ratio for the associated allele in rs3093301 in lupus patients compared to controls was 1.16 (95%CI= 1.08-1.25, meta-analysis p value= 1.0×10-4) in the combined European-derived and Hispanic samples.
Table 1 Genetic association between SNPs within IL21R and lupus in a cohort of European-derived lupus patients and controls. Cases analyzed included 2,245 females and 239 males, and controls analyzed included 2,009 females and 921 males. Only SNPs with minor (more ...)
Linkage disequilibrium (LD) plots for the IL21R SNPs analyzed in the European-derived (A), and the Hispanic (B) lupus patients and controls. Values depicted represent pair-wise correlation coefficient (r2).
Table 2 Genetic association between SNPs within IL21R and lupus in a cohort of Hispanic lupus patients and controls. Cases analyzed included 494 females and 63 males, and controls analyzed included 167 females and 42 males. Only SNPs with minor allele frequencies (more ...)
IL21R is located in proximity to the IL4R gene, another important candidate in autoimmunity. However, IL21R and IL4R are located in separate haplotype blocks in the European-derived samples (CEU) genotyped in the International HapMap Project (). To further confirm that the genetic effect we describe in rs3093301 is not a surrogate for a genetic association in IL4R, we genotyped 5 SNPs within the IL4R gene in both our European-derived and Hispanic cohorts. None of the genotyped SNPs in IL4R (rs2107356, rs2234895, rs1805011, rs1805013, and rs2074570) showed an association with lupus in either cohort (p>0.05).
To determine whether the lupus-risk allele in rs3093301 is more frequently associated with any clinical or serological manifestation in lupus patients, we determined the frequency of the various lupus manifestations in female lupus patients homozygous for the risk allele (A/A) (n=197) compared to patients homozygous for the protective allele (G/G) (n=62) in the European-derived cohort. Variables tested included malar rash, discoid rash, photosensitivity, oral ulcers, arthritis, pericarditis, pleuritis, proteinuria, urine cellular casts, seizures, psychosis, hemolytic anemia, leucopenia, lymphopenia, thrombocytopenia, the presence of anti-dsDNA, anti-Sm, anti-nRNP, anti-Ro, anti-La, and anti-ribosomal P antibodies. We found that the A/A genotype in rs3093301 is associated with the presence of malar rash in European-derived female lupus patients compared to the G/G genotype (60.9% versus 35.5%, OR=2.83, CI = 1.56-5.13, χ2
= 12.31, p=0.00045). This association remains significant after correcting for multiple testing using Bonferroni correction (corrected p= 0.0095). It was previously reported that reduced expression of IL21R in peripheral blood B cells is associated with nephritis in lupus patients (8
). We found no difference in the frequency of proteinuria or urine cellular casts between lupus patients with the A/A genotype compared to the G/G genotype in rs3093301 (p= 0.51 and 0.14, respectively).
We have previously reported a genetic association between IL21
and lupus. The SNP rs907715 within the IL21
gene was the most significantly associated SNP in our European-derived cohort (20
). To test if there is gene-gene interaction between IL21
upon lupus susceptibility, we performed logistic regression analysis to determine if the simultaneous presence of the risk alleles in the IL21
lupus-associated SNP (rs907715) and the IL21R
lupus associated SNP (rs3093301) could explain the increased susceptibility to lupus. We find no evidence for a synergistic effect for the presence of the risk alleles in rs907715 and rs3093301 (interaction p=0.58), suggesting no genetic epistasis between the IL21
loci upon lupus susceptibility.