The Impact of PPARγ Genetic Variants on IBD Susceptibility and IBD Disease Course

PPARγ is a nuclear receptor that regulates numerous pathways including cytokine expression and immune responses and plays an important role in controlling colon inflammation. We aimed at determining the occurring PPARγ SNPs, at predicting the haplotypes, and at determining the frequency outcome in inflammatory bowel disease (IBD) patients in comparison with healthy controls. We determined genetic variants in the coding exons and flanking intronic sequences of the NR1C3 gene in 284 IBD patients and 194 controls and predicted NR1C3 haplotypes via bioinformatic analysis. We investigated whether certain NR1C3 variants are associated with susceptibility to IBD or its disease course. None of the detected 22 NR1C3 variants were associated with IBD. Two variants with allelic frequencies over 1% were included in haplotype/diplotype analyses. None of the NR3C1 haplotypes showed association with IBD development or disease course. We conclude that NR1C3 haplotypes are not related to IBD susceptibility or IBD disease activity.


Introduction
Crohn's disease (CD) and ulcerative colitis (UC) are chronic recurrent forms of inflammation of the gastrointestinal tract (inflammatory bowel disease, or IBD), which are characterized by an onset in young adulthood and by an unpredictable disease course that may lead to debilitating complications [1]. The combined prevalence of CD and UC is estimated to be 100 to 200 per 100 000 individuals in developed countries [2]. IBD is thought to be of multifactorial genesis including a complex interplay among genetic, environmental, microbial, and immune factors [3]. The exact molecular pathogenesis of IBD is not yet fully elucidated [4]. Although great advances have been made in the clinical management of IBD, curative therapeutic strategies do not exist.
It has been recognized that components of pro-and antiinflammatory signalling cascades seem to play an important role in the pathogenesis of IBD. Initially, pro-and antiinflammatory cytoplasmic receptors that are activated by bacterial lipopolysaccharides, such as nucleotide oligomerisation domain (NOD) 2/caspase recruitment domain (CARD) 15, and NOD1/CARD4, have been especially studied in the past and have been identified as IBD susceptibility genes. These findings emphasized the pivotal role of an interaction between enteric microbes and the intestinal immune system in the pathophysiology of IBD [5][6][7][8].
Current evidence suggests that an additional receptor, PPARγ, plays an important role in the regulation of colon inflammation. PPARγ belongs to the nuclear receptor family that consists of approximately 50 different transcription factors, which are known to be involved in the regulation of a wide range of different biological processes. PPARγ controls the expression of a large number of different genes and was initially identified as an important regulator of genes involved in lipid metabolism and insulin sensitisation [9]. PPARγ acts through heterodimerisation with another nuclear receptor, retinoid X receptor α (RXRα). The heterodimer binds to specific DNA response elements within the promoters of its target genes (peroxisome proliferator response elements, PPREs) [10]. PPARγ is mostly expressed in adipose tissue and the large intestine. Kidney, liver, and small intestine express intermediate levels, whereas PPARγ is barely found in muscle [11].
PPARγ is known to modulate the expression of key transcription factors and kinases involved in inflammatory signalling cascades such as NF-κB, c-Jun, c-Fos, and nuclear 2 PPAR Research factor of activated T cell (NFAT). Thereby, PPARγ is able to inhibit the mucosal production of proinflammatory cytokines, such as interleukin 1β (IL-1β) and tumor necrosis factor-α (TNFα), and to downregulate the expression of various adhesion molecules [12,13]. Based on these findings, it has been demonstrated in mouse models that activation of PPARγ leads to an efficient reduction of the severity of intestinal inflammation by suppressing excessive immunoinflammatory responses [14,15]. As a consequence, PPARγ is currently under investigation as a potential target for novel anti-inflammatory agents [9]. Because of its central role in the regulation of colon inflammation, we hypothesized that PPARγ could be a putative susceptibility gene for the development of IBD. The PPARγ gene NR1C3 is located on chromosome 3 and is composed of 9 exons. Alternative splicing yields three different protein isoforms, PPARγ1, PPARγ2, and PPARγ3, which differ in the amino acid composition at their 5 ends. The isoform PPARγ2 is the most abundant PPARγ protein found in a number of human tissues [11]. PPARγ is known to be polymorphically expressed. Several SNPs have been described, one of which has been shown to have consequences for both the conformational protein structure and protein function [16,17]. So far, only a few studies have assessed the role of a few discrete NR1C3 gene polymorphisms in IBD pathogenesis. A systematic study to comprehensively investigate the role of global polymorphic features of the NR1C3 gene with a focus on its role in IBD susceptibility, such as the one presented here, has not been previously performed.
In the present study, we aimed at determining all occurring mutations and SNPs in the exonic regions of the PPARγ gene NR1C3, at bioinformatically predicting the arising haplotypes, and at evaluating their association with the risk to develop IBD and with IBD activity in a well-sized cohort of IBD patients and non-IBD controls.

Control Subjects.
One hundred and ninety-four non-IBD controls were recruited from gastroenterological patients undergoing surveillance colonoscopy, who did not show any symptoms of IBD. History of colorectal cancer was used as an exclusion criterion for both IBD patients and non-IBD controls. All subjects provided their written informed consent to be included in the study. Ethical approvals were obtained from the local medical ethical committees of all study sites involved in the collection of non-IBD samples.

Statistical Analysis of Allele Frequencies and Genotype
Distributions. To detect differences in SNP distribution between case and control groups or between disease activity

Haplotype and Diplotype
Analysis. The FAMHAP software was used to calculate the haplotypes and diplotypes based on the detected SNPs and mutations in the NR1C3 gene and to detect differences in haplotypes and diplotype distributions in case and control groups. FAMHAP performs a permutation test on associations between estimated haplotypes and the affection state based on Monte Carlo simulations. A value of P < 0.05 was considered to be significant.
Haplotype and diplotype calculations were performed on 256 IBD patients and 148 non-IBD controls, from which all sequence data of adequate quality were obtained. To allow referral to specific haplotypes, a frequency-based priority criterion was used to name haplo-and diplotypes (e.g., H 1 or D 1 for the most often occurring haplotype or diplotype, Table 8).

Calculation of Linkage Disequilibria.
Linkage disequilibria (LD) were calculated using the r 2 statistics. Calculations were performed using the software package Haploview (http://www.broadinstitute.org/scientific-community/ science/programs/medical-and-population-genetics/haploview/haploview).  Table 1 shows the demographic data of the individuals included into the NR1C3 analysis. The sequence data were screened for genetic variation in the NR1C3 gene, using the Basic Local Alignment Search Tool (BLAST; http://www.ncbi.nih.gov) and the GenBank entry NT 02257.18 as the reference sequence. As shown in Tables 2 and 3, altogether 22 variants were detected, whichwith exception of one mutation (one individual was found to be homozygous for variant number 6)-were in Hardy-Weinberg equilibrium. The majority of variants were singlenucleotide substitutions. Only one variant was characterised by a base-pair insertion leading to a frame shift. The majority of variants have not yet been described in the NCBI SNP database. Nine variants were found in exonic regions, 11 variants were found in intronic regions, one variant was detected in the 5 -prime region, and 1 variant was found within the 3 end of NR1C3. Only two of the detected variants (no. 1 and no. 7, Table 2) lead to nonsynonymous amino acid exchanges within the PPARγ protein. Furthermore, only two variants occurred with an allelic frequency of more than one percent (rs1801282 and rs3856806), thus fulfilling a definition of a genetic polymorphism ( Table 3).

NR1C3 Sequence
The two most often occurring variants rs1801282 and rs3856806 were found to be in moderately strong linkage ( Figure 1, r 2 = 40%, SNP numbers 4 and 19) This finding is in good agreement with previous publications [19][20][21]. With the exception of two individuals, who were carriers of mutation numbers 18 and 20 or 7 and 10 in combination, no subject ever carried more than one rare NR1C3 variant. Table 3, only two variants (rs1801282 and rs3856806) occurred in an allele frequency higher than 1% in IBD and non-IBD control group. No significant differences in allele frequencies were observed. Table 4 shows the results when comparing the frequency of NR1C3 genotypes carrying distinct genetic variants in heterozygous or homozygous form in IBD patients and non-IBD controls. No significant differences in the distribution     Tables 5, 6 and 7, none of the mentioned factors were significantly associated with the occurrence of variants rs1801282 or rs3856806.  Furthermore a comparison of the mean values of leukocyte or CRP concentrations in plasma did not show any significant differences between IBD patients carrying the variant rs1801282 or rs3856806 and non-variant carriers within the patient group (data not shown).

Haplotype and Diplotype
Analysis. The two NR1C3 genetic variants rs1801282 and rs3856806, which occurred in an allele frequency of higher than 1%, were included in the bioinformatic haplotype prediction using the computer programme FAMHAP. For this analysis, all individuals were considered, for which the sequencing outcome of all 22 variant loci was complete. Thus, it was possible to include 256 IBD patients (126 UC and 130 CD patients) and 148 controls. As shown in Table 8, four haplotypes (H1 to H4) were predicted to be in best reconstruction for both cohorts leading to eight different putatively occurring diplotypes (D1 to D8). All haplotypes and five diplotypes were predicted to occur at a frequency higher than 1% in the non-IBD control group.
Neither the overall haplotype nor the overall diplotype pattern varied significantly between the IBD group (or the IBD subgroups) and the control group. The result remained non-significant when investigating a possible relationship between the occurring haplotype or diplotype distribution pattern and disease activity (abundance of EIMs, fistulas, or high activity indices; Tables 9-11).

Discussion
PPARγ is an important modulator of pro-and antiinflammatory signalling cascades involving NF-κB. The importance of PPARγ is underlined by the fact that efficient anti-inflammatory effects can be reached when targeting PPARγ therapeutically. An important example of an antiinflammatory acting drug, which exerts agonistic effects on PPAR-γ and which is widely used in the therapy of especially UC is 5-aminosalicylic acid (5-ASA) [22].
Former studies, which investigated the impact of a polymorphic expression of PPARγ on diseases characterized by proinflammatory processes, focused specifically only on the analysis of the two commonly occurring NR1C3 SNPs rs1801282 and/or rs3856806. Several investigations showed that these NR1C3 gene variants are putatively associated with a moderately higher risk for the development of lifestyleassociated diseases (e.g., metabolic syndrome, coronary  artery disease, and type 2 diabetes) or for colorectal cancer [23][24][25][26][27][28][29]. However, these findings were only partly supported in subsequent meta-analyses [30][31][32].
In the study presented here, we aimed at comprehensively investigating the occurring polymorphisms within the PPARγ gene by sequencing all exonic regions and neighbouring intronic sequences. We analysed the frequency of arising genotypes and haplotypes in a large cohort of IBD patients and non-IBD control subjects and investigated the impact of the observed NR1C3 variants on IBD susceptibility and disease course.
Interestingly, NR1C3 appears to be strongly conserved. Only the two genetic variants rs1801282 and rs3856806, which have already been described in the literature and which are characterized by a moderately strong linkage behaviour, were found to occur in an allelic frequency of >1%. This observation, together with the fact that all other detected NR1C3 gene mutations occurred alone in >99% of all cases and never in combination with other PPARγ gene variants in any individual included in our study, supports the important physiological function of PPARγ, which apparently does not allow a highly polymorphic expression of this protein.
We did not find any significant association of distinct NR1C3 haplotypes with higher IBD susceptibility or with a modified IBD course. A few other studies have hitherto investigated the impact of a polymorphic PPARγ expression  on IBD susceptibility. These studies focused mainly on the investigation of the Pro12Ala polymorphism (rs1801282) and its putative influence on UC disease risk. These studies showed heterogeneous results. While [33] observing a significantly higher frequency of homozygous Pro12Ala SNP carriers in UC patients compared to controls in a Danish cohort, Shrestha et al. only observed a putative relationship between a higher UC disease activity and the occurrence of the Pro12Ala variant in a Dutch population, which they could not confirm in a Chinese cohort [34]. A third study focused specifically on the functional impact of the Pro12Ala SNP and showed that this variant appears to be associated with lower PPARγ mRNA levels in diseased mucosa of UC patients. This finding was combined with a higher prevalence of the Ala-variant in UC patients, when compared to CD patients and healthy controls. The latter observation, however, derives from only a relatively small number of individuals (29 UC and 10 CD patients, 134 controls), which were included in the analysis [35]. Two additional small studies did not find any significant impact of the SNP Pro12Ala on disease susceptibility for CD [36] or UC [37]. In the context of the heterogeneous study outcomes published so far, our study rather supports the hypothesis that a polymorphic expression of the PPARγ gene NR1C3 does not significantly influence the IBD risk or the course of the IBD forms, CD and UC.

Conclusions
In conclusion, we have performed a comprehensive study analyzing the role of NR1C3 genetic variants in IBD susceptibility and IBD course in a Swiss cohort of IBD patients. We showed that the polymorphic expression of the PPARγ gene is not a general modulating risk factor for IBD.