Objective Interferon-gamma (IFNG) is a type II interferon playing diverse roles in innate and adaptive immune systems. Elevated expression of IFNG has been associated with systemic lupus erythematosus (SLE). This study examined the association of IFNG polymorphisms with SLE susceptibility.
Methods Five tag single-nucleotide polymorphisms (SNP) and eight variations in all known regulatory sequences affecting IFNG expression within and around IFNG were genotyped in 1759 unrelated Korean subjects. SLE susceptibility association was assessed by comparing 742 SLE patients and 1017 unaffected controls using multivariate logistic regression analysis with adjustment for age and gender.
Results SLE susceptibility association was significant with rs2069705 in the promoter (adjusted OR 2.27, p=0.0024) and marginal with rs3181032 in the promoter (p=0.037), rs2430561 in intron 1 (p=0.022) and rs2069718 in intron 3 (p=0.026) in a recessive genetic model. Five other SNP showed no association and four other variations were not polymorphic.
Conclusion Several SNP in IFNG are associated with SLE susceptibility, and the risk allele of an associated SNP (rs2430561) located in an NF-κB binding site has elevated IFNG expression versus the non-risk allele, supporting that elevated IFNG expression is associated with increased SLE susceptibility.
Statistics from Altmetric.com
Systemic lupus erythematosus (SLE; OMIM 152700) is a chronic autoimmune disease caused by diverse genetic and environmental factors, and several loci and genes have been associated with susceptibility to SLE using genome-wide scan and candidate gene approaches. In this case–control study, we investigated several polymorphisms in the interferon-gamma gene (IFNG) for association with susceptibility to SLE. IFNG is located on chromosome 12q14.1, an SLE susceptibility locus previously identified in a genome-wide linkage study using white families, among many other genes.1 Furthermore, IFNG is excessively expressed in peripheral blood T cells of SLE patients.2 In murine models of lupus, deletion of Ifng or inhibition of its receptor leads to an improvement of lupus-like symptoms.3 4 The elevated IFNG expression in CD4+ cells of human SLE patients and a murine lupus model could be explained by defective DNA methylation.5 6 In this study, however, a genetic contribution of IFNG to SLE susceptibility was discovered other than the epigenetic contribution.
Subjects and genotyping
A total of 1759 unrelated Korean subjects including 742 patients with SLE and 1017 healthy controls was recruited at Hanyang University Hospital for Rheumatic Diseases, and their demographic characteristics are shown in supplementary table 1 (available online only). Patients fulfilled the American College of Rheumatology criteria. Genomic DNA were obtained from peripheral blood cells of the subjects and genotyped using single-base extension reactions of MassARRAY (Sequenom, San Diego, California, USA), according to the manufacturer's instructions.
Selection of variations for association tests
A total of 13 variations located within and around IFNG (table 1) was selected for SLE association tests. Five single nucleotide polymorphisms (SNP) had been genotyped in the Korean HapMap Project (http://www.khapmap.org) for the IFNG gene (5 kb gene plus 5 kb upstream) in a Korean population, and their pair-wise linkage disequilibrium (LD) values were not high (0.00 ≤ r2≤ 0.68). Only two of them (rs2069718 and rs1861493) and one additional SNP (rs1861494) had been genotyped for IFNG in the International HapMap Chinese and Japanese populations (http://www.hapmap.org), but the additional SNP rs1861494 was fully correlated (r2= 1.00) with rs1861493, which was genotyped in both Korean and International HapMap Project is an official name for an international project. Accordingly, all HapMap SNP for IFNG could be tagged by the five Korean HapMap SNP.
In addition, all the variations in the known IFNG regulatory regions7 were retrieved from the dbSNP database (http://www.ncbi.nlm.nih.gov/projects/SNP), including unvalidated ones (supplementary table 2, available online only). Ten variations were located in six conserved non-coding sequences (CNS) with enhancer activity for IFNG expression,8 but three of them were not polymorphic according to the Korean or International HapMap Project data for Asians. A SNP (rs2430561) in a nuclear factor kappa B (NF-κB) transcription factor binding site was additionally selected,9 but the non-polymorphic rs2069707 on a heat shock factor (HSF) binding site was not.10 Seven additional SNP were located in the promoter regions where DNA methylation epigenetically regulates IFNG expression in T helper cell lineage,5 but none of them changed a CpG sequence or were polymorphic. No SNP had been found in the 14-bp 5'-untranslated region known to affect translation efficiency.7 A total of eight variations was thus chosen from all the known cis-acting elements for IFNG expression to be examined in this study in addition to the five tag SNP.
Associations with SLE susceptibility and manifestation were assessed by logistic regression analysis with adjustment for the slight differences in age distribution and gender ratio between the cases and controls using SPSS version 11.5. For multiple-testing correction, an effective degree of freedom was estimated at 7.53 for the set of nine tested SNP using the EDF program11 and the statistical significance level was set at α = 0.0066 (0.050/7.53). Association was considered significant for p < 0.0066 but marginal when 0.0066 ≤ p ≤ 0.050. The genotype frequencies in the control subjects were under Hardy–Weinberg equilibrium in the χ2 test. LD patterns and haplotypes were analysed using Haploview version 4.1, and haplotype association was examined using the χ2 test. Statistical power was calculated using the PGA1 program.11
In this study, SLE susceptibility association was assessed for 13 variations located within and around IFNG (table 1). Five SNP were selected to tag all the IFNG SNP previously genotyped in the Korean and International HapMap Projects, and eight variations (seven SNP and one insertion) were retrieved from the dbSNP database (including unvalidated ones) to examine all the known IFNG regulatory regions,7 including CNS regions with enhancer activity for IFNG expression.8
When 1759 unrelated Korean subjects were genotyped for the 13 variations, four CNS variations (rs4913413, rs35332482, rs61923089 and rs60694874) were not polymorphic (table 1) and the remaining nine SNP were not highly correlated with each other (0.01 ≤r2≤ 0.66 in the control samples), except that a non-tag, regulatory SNP rs2430561 was almost fully correlated (r2= 0.98) with a tag SNP rs2069718 (figure 1).
The nine common SNP were tested for association with SLE susceptibility in a codominant genetic model using multivariate logistic regression, with adjustment for age and gender by comparing 742 SLE patients and 1017 healthy controls (table 1). The IFNG promoter SNP rs2069705 was significantly associated with SLE susceptibility (p = 0.0047). Marginal association was observed (p = 0.029) with two correlated SNP, rs2430561 and rs2069718, but was accounted for by their moderate correlations (r2= 0.55 in the controls) with the significantly associated rs2069705, because association disappeared when adjusted for rs2069705 genotypes in conditional logistic regression analysis.
The three SNP showed SLE susceptibility association additionally in a recessive genetic model. The association was significant with rs2069705 (adjusted odds ratio (OR) 2.27, p = 0.0024) and marginal with rs2430561 (adjusted OR 2.48, p = 0.022) and rs2069718 (adjusted OR 2.42, p = 0.026). In addition, a tag SNP rs3181032 was marginally associated with SLE in a recessive model (adjusted OR 9.81, p = 0.037), although the number of minor allele homozygotes was too small (one control and seven patients).
According to the Korean HapMap data, the marginally associated SNP rs2069718 in IFNG was fully correlated (r2= 1.00) with rs2216164 and rs7304531 located at 32 kb and 19 kb downstream from IFNG, respectively. Only two haplotypes were common when constructed with two IFNG intron SNP (rs2069718 and rs2430561) and two downstream SNP (rs2216164 and rs7304531), and the risk haplotype was marginally associated with SLE susceptibility (p = 0.031, supplementary table 3, available online only).
All 13 variations were additionally tested for association with all American College of Rheumatology diagnosis criteria for SLE listed in supplementary table 1 (available online only). None of them was associated with any manifestation, except that the significantly SLE-associated SNP rs2069705 was marginally associated with malar rash in a codominant (p = 0.016) or recessive (p = 0.019) genetic model (table 2).
This study reports the genetic association of IFNG SNP with susceptibility to SLE, which is consistent with previous functional implications of high expression levels of IFNG in the pathogenesis of lupus in both human disease2 and murine models.3 4 Although a significantly associated SNP (rs2069705) in the promoter has not been examined for its effects on IFNG expression, a marginally associated SNP (rs2430561) in intron 1 has been demonstrated to affect the IFNG expression level, as the minor A allele had higher affinity to nuclear transcription factor NF-κB in an electrophoretic mobility shift assay9 and exhibited a higher enhancer activity in an expression reporter assay12 than the major T allele. Furthermore, the minor A allele was highly correlated with the (TG)12 allele of a TG-repeat microsatellite polymorphism (located adjacent to rs2430561)9 13 which has been associated with increased IFNG secretion from cultured peripheral blood mononuclear cells.14
Although polymorphisms in IFNG have been associated with susceptibility to various diseases including immune diseases, with a manifestation (arthritis) within SLE patients15 16 and with a certain histological phenotype within lupus nephritis patients,17 none of them has been associated with SLE susceptibility. The expression-associated IFNG SNP rs2430561 was not associated with SLE susceptibility in a white population of 24 cases and 36 controls18 and a Thai population of 154 cases and 154 controls,16 and the adjacent TG-repeat microsatellite polymorphism was not associated with SLE susceptibility in a mixed white and African-American population of 136 cases and 99 controls.15 These three previous cohorts were too small to detect the OR of 2.48 observed in this study (power 8%, 21% and 16%, respectively, vs 81%).
In summary, SLE susceptibility is associated with polymorphisms in IFNG, supporting previous associations of IFNG expression with SLE susceptibility. This association needs to be examined in other populations by well-designed replication studies, because the LD patterns of the SNP in and around IFNG are very diverse among the Korean and International HapMap populations.
KK and SKC contributed equally to this work and are joint first authors.
Funding This study was supported by grants from the Research Program for New Drug Target Discovery (20090083335 to CK), Korean HapMap Project (M10504000005 to CK) and Korea Healthcare Technology R&D Project (A010252 and A080588 to S-CB; A091327 to S-KC).
Competing interests None.
Ethics approval The study was approved by the Institutional Review Board of Hanyang University Medical Center, Seoul, Korea.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.