Association of IFI200 Gene Polymorphisms with Susceptibility to Systemic Lupus Erythematosus

Detection of apoptosis

Within 60 minutes after blood draw, peripheral blood mononuclear cells (PBMC) from 15 healthy donors containing different IFI16 genotypes were isolated by Ficoll-Hypaque gradient (Robbins Scientific, Sunnyvale, CA, USA) and resuspended in RPMI-1640 medium (Sigma, New York, NY, USA) with 10% fetal bovine serum (Gibco, Karlsruhe, Germany) and 100 U/ml penicillin (Gibco) and 100 μg/ml streptomycin (Gibco) at concentration 5 × 10⁶ cells/ml. These cells were plated into 24-well plastic tissue culture plates. PBMC were treated with doxorubicin (Pharmacia and Upjohn GmbH, Freiburg, Germany) at a concentration of 10 μg/ml and then cultured at 37°C in 5% CO₂ for 24 h. This condition was determined to be effective to induce highest p53 expression in dose (1, 10, and 100 μg/ml) and time (8 and 24 h) response experiments. p53 mRNA expression was studied using real-time reverse transcription-polymerase chain reaction (RT-PCR) with the following primers: 5′TGG CCA TCT ACA AGC AGT CAC A3′ and 5′GCA AAT TTC CTT CCA CTC GGA T3′, as described¹⁰. Finally, the percentage of apoptotic cells was measured by FACScan flow cytometry (Becton Dickinson, Mountain View, CA, USA) for annexin V-FITC and propidium iodide binding according to the manufacturer’s protocol (BD Biosciences Pharmingen, San Diego, CA, USA). Cells positive only for annexin V were counted as apoptotic.

Quantification of IFIX and IFI16 isoforms

Expression of IFIX and IFI16 isoforms was semiquantified by conventional RT-PCR using the following primers: 5′GGA ACA GAG TCA GCA TCC and 5′GTT ATT TGA TAT CCT TGT CC for IFIX ¹¹; and 5′CAT CTT CGG ACT CCT CAG and 5′GTT CAG CAC CAT CAC TTC for IFI16. cDNA from leukocytes of 19 SLE patients containing different genotypes was amplified in a Perkin Elmer/GeneAmp PCR 2400 system. The PCR conditions consisted of initial denaturation at 94°C for 5 min, followed by 30 cycles of denaturation (94°C, 30 s), annealing (60°C for IFIX gene, 62°C for IFI16 gene, 30 s), extension (72°C, 30 s), and final extension (72°C, 7 min). PCR products were loaded in 1.5% Tris-acetate agarose gel and analyzed by electrophoresis in Tris-acetate buffer at 100 volts for 40 min followed by staining with 50 μg/ml ethidium bromide. The density of product bands was semiquantified using Gel Doc™ MZL software (Bio-Rad).

Statistical analysis

The association between genotypes and the expression of isoforms or apoptosis was analyzed by GraphPad Prism, version 4.0 (Graph-Pad Software, San Diego, CA, USA). The Mann-Whitney U test was used in comparisons between 2 groups, while Kruskal-Walls H was applied in case of analysis of more than 2 groups. A p value < 0.05 was considered significant.

Eleven putative functional single-nucleotide polymorphisms (SNP) were selected from MNDA, IFIX, IFI16, and AIM2 genes according to computational prediction and genotyped as described in Table 1. The percentage of genotyping for each SNP was 100%. There were no missing genotyping data. All SNP were in Hardy-Weinberg equilibrium when we compared the observed with the expected genotype frequencies of each SNP (p > 0.05). PLINK v1.05 program¹² was used to calculate p values, empirical p values (100,000 permutations), and odds ratios (95% confidence intervals) for genotype and haplotype associations. Linkage disequilibrium was determined by JLIN, a Java-based linkage disequilibrium plotter¹³ (Table 2).

A significant association of the minor allele frequency from 3 SNP of IFIX and IFI16 genes was found between SLE patients and healthy controls (Table 3). The G13792T (rs856084) in IFIX and the C6771G (rs866484) and A23201G (rs1772414) in IFI16 showed significant differences (OR 0.73, 95% CI 0.54–0.98, empirical p = 0.033; OR 1.33, 95% CI 1.00–1.75, empirical p = 0.05; and OR 1.37, 95% CI 1.03–1.80, empirical p = 0.030, respectively). The association of both SNP (rs866484 and rs1772414) in IFI16 gene seems to be a recessive effect model. The presence of 2 G alleles (GG) conferred a significant p = 0.009 (empirical p = 0.023) for rs866484 and p = 0.013 (empirical p = 0.032) for rs1772414. The marginal significance results are probably due to the limited sample size in this study. We performed haplotype analysis of significant SNP rs866484 and rs1772414 with the strongest linkage disequilibrium (D’ = 0.814 and r² = 0.643) by comparing each haplotype with another 3 haplotypes between SLE patients and controls. Our finding showed that the GG was a risk haplotype (OR 1.41 and p = 0.017, empirical p = 0.031), whereas CA was a protective haplotype (OR 0.73 and p = 0.032, empirical p = 0.041).

Regarding the selection as putative functional SNP, we tried to prove our hypothesis of 3 positive SNP as follows. SNP rs866484 (C6771G) of IFI16 is a nonsynonymous SNP whose amino acid is altered from threonine to serine. This SNP is located on an A-type repeat containing p53 binding site. Binding of IFI16 protein to the C-terminus of p53 has been reported to stimulate the transcription of p53-responsive reporter plasmids and leads to susceptibility to apoptosis of cells¹⁴. We hypothesized that the G risk allele may bind p53 with higher affinity and lead to more apoptosis. We conducted our experiment by inducing the highest p53 mRNA expression with doxorubicin at a concentration of 10 μg/ml for 24 h. The association of SNP rs866484 and apoptosis rate was studied by flow cytometry analysis. We were unable to determine the association between SNP and apoptosis (Figure 1A). This negative finding may be because threonine and serine are in the same amino acid group. Nevertheless, further study should be performed using direct protein-protein interaction methods to prove the direct role of this SNP in protein structure that affects p53 binding. In addition, we tested the association of the 2 SNP within the intron of IFIX (rs856084) and IFI16 (rs1772414) genes with the expression of the isoforms (α and ß for IFIX; A, B, and C for IFI16) by reverse transcription-polymerase chain reaction, as described³. However, we found no differences in isoforms among the genotypes (Figure 1B, 1C, 1D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

A. The effect of SNP rs866484 genotypes (CC, CG, and GG; n = 5, 6, and 5, respectively) on apoptotic response of healthy PBMC treated with 10 μg/ml doxorubicin for 24 h. B, C, D. The ratio of short product/full-length product grouped by SNP rs856084 genotypes of IFIX (GG, GT, and TT; n = 3, 9, and 3) and SNP rs1772414 genotypes of IFI16 (AA, AG, and GG; n = 7, 6, and 6). Each data symbol represents an individual sample; horizontal lines show median values.

Our results suggest that these SNP are not likely to be functional SNP. We could not exclude the possibility that these SNP are in linkage disequilibrium with nearby causative SNP. Interestingly, one report in SLE in Caucasian patients also suggested an association signal in the IFIX-IFI16 intergenic region¹⁵. More extensive research using dense SNP and increased sample sizes is required to clarify the role of IFIX and IFI16 gene in SLE.