Evaluation of the IL1 Gene Cluster Single Nucleotide Polymorphisms in Primary Open-Angle Glaucoma Pathogenesis

Introduction

Glaucoma is a heterogeneous group of optic neuropathies characterized by typical visual field loss and often associated with elevated intraocular pressure (IOP). It is the largest cause of irreversible blindness affecting ∼60 million people globally. Among various subtypes, primary open-angle glaucoma (POAG) is the most prevalent form (Tham et al., 2014).

Glaucoma is a complex genetic disease, of which 72% of cases exhibit a familial component. However, it often lacks a clear Mendelian pattern of inheritance (Gong et al., 2007). To date, at least 33 loci have been linked with POAG; however, so far only 5 genes have been identified namely Myocilin (MYOC) (Stone et al., 1997), Optineurin (OPTN) (Rezaie et al., 2002), WD repeat protein 36 (WDR36) (Monemi et al., 2005), NTF4 (Pasutto et al., 2009), and ASB10 (Pasutto et al., 2012). In addition, recently, MPP7 has been proposed as another novel candidate gene for POAG possibly influencing the aqueous humor dynamics in the eye (Vishal et al., 2016). Mutations in these genes account for a small percentage of POAG cases. However, in most cases, the pathogenesis still remains elusive. Apart from mutations in the candidate genes, common variants of many other genes are known to influence the pathogenesis of POAG (Ray and Mookherjee, 2009).

Aberrant immune modulation has long been hypothesized to be responsible for sporadic POAG, especially in cases having low to normal IOP (Tezel and Wax, 2004) referred as normotensive glaucoma (NTG). More than 30% of NTG cases have some form of immune dysregulation (Cartwright et al., 1992). Autoantibody against many self-proteins, including HSP-27, HSP-60, rhodopsin, α-crystallins, α-fodrin, γ-enolase, glutathione-S-transferase, tumor necrosis factor-α, and glycosaminoglycans, was identified in glaucoma patients (Tezel et al., 1998, 2004; Wax et al., 1998; Yang et al., 2001). It is not yet clear whether these autoantibodies are an epiphenomenon of glaucoma or act as a causative factor. However, experimental evidence suggests that these autoantibodies in the serum of glaucoma patients could significantly upregulate apoptosis-related proteins in cultured retinal ganglion cells (RGCs), which might cause higher RGC loss in glaucoma patients (Tezel and Wax, 2000). Recent work by Tezel et al. (2012) suggested that autoantibodies of some proteins (namely AIF, CREB-binding protein, ephrin type-A receptor, and huntington) could also act as potential POAG biomarkers.

Interleukins are a group of cytokines involved in different effector mechanisms in the immune system. Previous studies have reported increased level of interleukins in the aqueous humor of glaucoma patients, including IL8, IL6, IL1B, IFN-α, and GM-CSF (Kuchtey et al., 2010; Borkenstein et al., 2013). Previous reports also suggested that single nucleotide polymorphisms (SNPs) in the IL1 gene might be associated with POAG (Wang et al., 2006). A recent study from our laboratory demonstrated an association of a haplotype made by IL1A (−889C/T), IL1B (−511 C/T), and IL1B (+3953C/T) with non-high tension glaucoma (non-HTG) cases (POAG patient with presenting IOP <20 mmHg) in an east Indian POAG patient cohort (Mookherjee et al., 2010). In addition, the IL1 gene cluster is located in close proximity to an established POAG locus (GLC1B at 2cen-2q13) (Stoilova et al., 1996), which is linked to low tension glaucoma. In this study, we evaluated 68 SNPs spanning a region of ∼350 kb in the IL1 gene cluster for their possible association with east Indian POAG patient cohort, consisting of nine genes (IL1A, IL1B, IL1F7, IL1F9, IL1F6, IL1F8, IL1F5, IL1F10, and ILRN) in chromosome 2q13 (Nicklin et al., 2002).

Materials and Methods

In this biphasic study, a total of 392 patients and 354 controls were recruited. The clinical parameters and distribution of patients and controls in two phases are furnished in Table 1. Genomic DNA from the peripheral blood samples of POAG patients and controls was isolated as described previously (Mookherjee et al., 2010) with written consent from both the groups. The study protocol adhered to the tenets of the Declaration of Helsinki and was approved by the Human Ethics Committee of the Indian Institute of Chemical Biology, Kolkata, India.

A total of 68 SNPs from the IL1 gene cluster were chosen from HapMap Tag SNP data from the European Cohort (CEU) as an earlier analysis found similarity between our study cohort (Indo-European) and the European population (Indian Genome Variation Consortium, 2005, 2008); (MAF ≥0.1, R² ≥ 0.8). A few coding nonsynonymous SNPs were also included for their possible functionality (Supplementary Table S1; Supplementary Data are available online at www.liebertpub.com/gtmb).

In the first phase of the study, in addition to 68 SNPs, 3 previously reported (rs1800587, rs16944, and rs1143634) (Mookherjee et al., 2010) SNPs were genotyped by a MALDI-TOF-based Sequenom platform at The Centre for Genomic Application (TCGA, New Delhi). The in-house genotyping data of the three SNPs (rs1800587, rs16944, and rs1143634) obtained by restriction fragment length polymorphism (RFLP) from our previous study were used for quality control and ∼99% concordance was observed between RFLP and Sequenom genotypes. In the second phase, selected SNPs were genotyped by Sanger sequencing, using an automated DNA sequencer (ABI 3130XL; Applied Biosystems, Foster City, CA).

The allele frequencies of the SNPs were compared between patients and controls using a chi-square test. The p-values were corrected for multiple testing by the Single Nucleotide Polymorphism Spectral Decomposition (SNPSpD) software (Nyholt, 2004). Linkage disequilibrium (LD) was determined by the Haploview 4.2 software (www.broad.mit.edu/mpg/haploview). The bootstrap analysis was performed by resampling with replacement in the first phase of the study (see Supplementary Materials and Methods section for further details).

Results and Discussion

In the first phase of the study, a total of 68 SNPs were genotyped in 198 patients and 112 control subjects. Among these variants, seven were excluded from further analysis as six did not follow Hardy-Weinberg equilibrium and one was monomorphic in our dataset. The minor allele frequencies of the remaining 61 SNPs in the study cohort and population enlisted in HapMap database are furnished in Supplementary Table S1.

Three SNPs (rs3811046, rs3811047, and rs1013477) were found to be significantly associated with non-HTG cases, among which the first two are nonsynonymous, located in exon 2 of IL1F7, and are in complete LD, while the third one (rs1013477) is an intergenic SNP (Table 2). For rs3811046 (c134 G>T, Gly31Val), the c.134T allele and for rs3811047 (c.166 A>G, Thr42Ala), the c.166G allele were found to be overrepresented in the patient population (Table 2). For the intergenic SNP rs1013477 (A>G), the G allele was found to be overrepresented in the patient population (Table 2). However, these associations were found to be weak after p-values were corrected for multiple tests.

Interestingly, in silico analysis by SIFT (http://sift.bii.a-star.edu.sg/) revealed that the substitution of glycine (GGA) with valine (GTA) at the 31st codon and threonine (ACC) with alanine (GCC) at the 42nd codon in IL1F7 might alter the protein function (SIFT score 0.00 and 0.02, respectively).

A thorough literature survey revealed that IL1F7 might be involved in diseases resulting from aberrant immunomodulation. It is a member of the IL1 gene family (Nicklin et al., 2002) and acts as an antagonist for IL1 gene expression. It functions by controlling the inflammatory response and can suppress or reduce the function of many pro-inflammatory cytokines, including IL1A, IL1B, IL6, and IL1RN (Bufler et al., 2002; Sharma et al., 2008; Nold et al., 2010). IL1F7 was shown to be associated with ankylosing spondylitis (AS) (Pan et al., 2010; Pei et al., 2013) and inflammatory bowel disease (IBD) (Imaeda et al., 2013). High expression of IL1F7 was found in the inflamed colonic mucosa of IBD patients (Imaeda et al., 2013), as well as in the inflamed colonic mucosa of patients with ulcerative colitis and Crohn's disease (Imaeda et al., 2013). In the Chinese population, IL1F7 was found to be highly expressed in the affected synovia of AS patients and rs3811047 was found to be associated with AS (Pan et al., 2010). Elevated level of IL1F7 was also found in patients with atopic dermatitis (Fujita et al., 2013). In general, it is highly expressed in tissues with inflammation and presumed to inhibit both the innate and the adaptive immune responses.

Aberrant immunomodulation has also been implicated in glaucoma especially in NTG cases. A preliminary bootstrap analysis from phase 1 of the study, taking 100 patients and controls at random from the pool of 198 patients and 112 controls, indicated a weak association of rs3811047 and rs3811046 with non-HTG cases. Next, we evaluated these two SNPs in an additional 194 non-HTG cases and 242 controls from the same population. However, the preliminary observation could not be replicated in the validation cohort (Table 3).

The results from both the discovery and replication cohorts were combined (392 patients and 354 controls) to analyze the distribution of these two SNPs in the non-HTG patient and control cohorts. However, the combined dataset did not yield any significant association with non-HTG cases (Table 3).

In conclusion, our study did not reveal any potential association of the SNPs in the IL1 gene cluster with non-HTG patients. Further studies on additional large cohorts will help to validate our observation.