INK4/ARF Transcript Expression Is Associated with Chromosome 9p21 Variants Linked to Atherosclerosis

Background Genome-wide association studies (GWAS) have linked common single nucleotide polymorphisms (SNPs) on chromosome 9p21 near the INK4/ARF (CDKN2A/B) tumor suppressor locus with risk of atherosclerotic diseases and type 2 diabetes mellitus. To explore the mechanism of this association, we investigated whether expression of proximate transcripts (p16INK4a, p15INK4b, ARF, ANRIL and MTAP) correlate with genotype of representative 9p21 SNPs. Methodology/Principal Findings We analyzed expression of 9p21 transcripts in purified peripheral blood T-cells (PBTL) from 170 healthy donors. Samples were genotyped for six selected disease-related SNPs spanning the INK4/ARF locus. Correlations among these variables were determined by univariate and multivariate analysis. Significantly reduced expression of all INK4/ARF transcripts (p15INK4b, p16INK4a, ARF and ANRIL) was found in PBTL of individuals harboring a common SNP (rs10757278) associated with increased risk of coronary artery disease, stroke and aortic aneurysm. Expression of MTAP was not influenced by rs10757278 genotype. No association of any these transcripts was noted with five other tested 9p21 SNPs. Conclusions/Significance Genotypes of rs10757278 linked to increased risk of atherosclerotic diseases are also associated with decreased expression in PBTL of the INK4/ARF locus, which encodes three related anti-proliferative transcripts of known importance in tumor suppression and aging.


Introduction
Recent human genome-wide association studies (GWAS) have shown that SNPs with high minor allele frequencies near the INK4/ARF (CDKN2A/B) tumor suppressor locus on chromosome 9p21 correlate with risk of atherosclerotic diseases (coronary artery disease (CAD) [1,2,3,4,5], ischemic stroke [6,7], and abdominal aortic aneurysm [6]) and type 2 diabetes mellitus (T2DM) [8,9,10]. Evidence from our lab has suggested that expression of p16 INK4a , one of the proteins encoded by the INK4/ARF locus, controls pancreatic islet proliferation and mass with aging [11], and therefore could plausibly explain the link to T2DM. The mechanism underlying the association of these SNPs with atherosclerotic diseases, however, is less clear.
In particular, it is not known if these genetic variants are associated with altered expression of cis-encoded transcripts. The most replicated SNPs associated with strongest effect sizes for atherosclerotic disease and T2DM are .100 kb centromeric to the INK4/ARF locus (Fig. 1), which encodes three cell cycle inhibitors (p15 INK4b , p16 INK4a , ARF, reviewed in [12]) and a recently discovered putative non-coding RNA designated ANRIL [13]. There are no known cis-regulatory elements of the INK4/ ARF locus located at such a distance from the locus. It has also been suggested that a more distant gene on 9p21, MTAP (Fig. 1), could mediate disease associations [3,4].
In this work, we sought to determine whether there is a link between these disease-associated SNPs and expression of 9p21 transcripts. We found that individuals harboring a common SNP genotype associated with increased risk for atherosclerotic disease demonstrated markedly reduced expression of INK4/ARF-encoded transcripts in peripheral blood T-cells (PBTL), indicating a potentially important role for cell cycle inhibitors in atherosclerotic disease.

Ethics Statement
This study was approved by the University of North Carolina Institutional Review Board with all participants providing informed written consent.

Subjects, mRNA analysis and SNP genotyping
The subject characteristics will be described elsewhere in details (YL and NES, submitted). In brief, PBTLs were obtained from 170 healthy subjects and isolated to .90% purity by Magnetic Activated Cell Sorting (MACS). MACS, RNA preparation and transcript expression analysis were performed in PBTL as described ( [14], and YL and NES, submitted). Transcript expression levels are presented as log 2 -transformed expression (DDC T , ref. [15]). Genomic DNA was prepared from buffy coats using Puregene Blood Core Kit (Qiagen) according to manufacture's instructions. All SNP genotyping was performed at the UNC Mammalian Genotyping Core using TaqmanH SNP Genotyping assays by automatic genotype calling. Allele frequencies and linkage disequilibrium (LD) in studied subjects were comparable to those from HapMap (Table S1). All SNP genotypes were in Hardy-Weinberg Equilibrium (HWE) determined by Haploview software.

Statistical analysis
Linear models with different slopes and intercepts for each genotype were used to compare the effect of genotype on INK4/ ARF transcripts adjusting for age. ANOVA and Kruskal-Wallis tests were used to compare the mean gene expression across the genotypes (Table S2). A Student's t-test was used to compute pvalues for the pair-wise comparisons between the genotypes, and p-values were adjusted using Bonferroni method for multiple comparisons (Table S2). All analysis was performed using SAS (v. 9.1.3, SAS Institute Inc).

Results
We selected six SNPs for genotyping based on strength of association in previous studies: two associated with atherosclerosis (rs10757278 and rs518394) [1,2,3,4], two with T2DM (rs10811661 and rs564398) [8,9,10], one with frailty (rs2811712) [16], and one with melanoma (rs11515, in the 39 UTR of p16 INK4a and ARF) [17] ( Fig. 1). These SNPs were also chosen to have limited LD across the locus ( Fig. 1 and Table S1). While no significant associations were noted between expression of any 9p21 transcript and five of the six SNPs analyzed (not shown), a strong association was noted between INK4/ARF transcript expression (log 2 -transformed or DDC T ) and rs10757278 genotype ( Fig. 2 and Table S2). This SNP and nearby SNPs in high LD have been reproducibly associated with atherosclerotic disease in independent studies [1,2,3,4,6,7].
We observed significantly reduced expression of ARF, p15 INK4b , p16 INK4a , and ANRIL in GG vs. AA individuals ( Fig. 2A and 2B). The increase in expression in AA vs. GG donors was 2.1 to 2.8fold for all four INK4/ARF transcripts, with an intermediate effect seen in AG individuals for each transcript. These effects were highly significant after adjustment for multiple comparisons (p,0.0001, Table S2). While expression of p16 INK4a is highly dynamic in PBTL with age, differences between genotypes were significant at all ages (Fig. 2B). Moreover, the p16 INK4a and ARF have two shared exons (Fig. 1, reviewed in [12]), and aggregate p16 INK4a /ARF expression detected with primers/probe pairs that span the shared exons (Exon2/3) were also significantly increased in AA individuals compared to GG individuals ( Fig. 2A). In contrast, expression of a more distant 9p21 transcript, MTAP, as The LD plot of D9/LOD was generated using Haploview software [42]. The strength of LD increases from white to blue to red: white (disequilibrium coefficient (D9) ,1 and LOD score ,2); blue (D9 = 1 and LOD score ,2); pink (D9,1 and LOD score $2); and bright red (D9 = 1 and LOD score $2). well as genes located on other chromosomes (p21 CIP and IRF8) was not associated with rs10757278 genotype (Fig. 2A). As the G-allele correlates with increased risk of atherosclerotic disease [6,18], these data indicate that decreased expression of all members of the INK4/ARF locus in PBTL is associated with a higher risk of atherosclerotic disease.

Discussion
Here, we show that the expression of INK4/ARF locus transcripts is significantly reduced in individuals harboring a common allele of a 9p21 SNP ,120 kb from the INK4a/ARF locus. This observation suggests a surprising human heterogeneity in the expression of genes that have been associated with various aspects of mammalian cancer and aging (reviewed in [19]), including a particular importance in lymphocyte aging [20]. We believe this observation can be most parsimoniously explained by the existence of a polymorphic cis-regulatory element that influences expression of the INK4/ARF locus and that is closely linked to rs10757278. Cis-regulatory element located more than 100 kb apart from affected open reading frames (ORFs) are welldescribed at other loci [21].
It is important to note that rs10757278 and linked SNPs in high LD (e.g. rs1333049) are associated with family history of atherosclerotic disease, but not other traditional risk factors including body mass index, diabetes, tobacco use, blood pressure, C-reactive protein or lipid levels [2,22,23]. The risk genotypes, however, are strongly associated with coronary artery calcification, premature atherosclerosis and angiographically characterized disease [2,24,25]. Therefore, we believe the present finding supports a direct link between INK4/ARF expression and atherogenesis.
We believe a reasonable argument can be made for how each of the INK4/ARF transcripts might modulate the earliest stages of atherosclerosis. The INK4-class cyclin dependent kinase (CDK) inhibitors, of which p15 INK4b and p16 INK4a are founding members, block cell cycle progression by inhibiting the activity of the proliferative kinases CDK4 and CDK6. A role for cell cycle inhibitors has been postulated in promoting favorable vascular remodeling and preventing pathologic intimal hyperplasia by regulating G1 to S phase progression in vascular smooth muscle cells (reviewed in [26]). In support of this notion, mice lacking p16 INK4a have been shown to be more sensitive to vascular intimal hyperplasia in a carotid artery injury model [27], suggesting that p16 INK4a protects against atherogenesis by limiting a pathologic vascular reaction to injury in some circumstances. Likewise, TGFb signaling is thought to attenuate atherogenesis [28,29], and the anti-proliferative effects of TGF-b signaling are in part mediated by p15 INK4b (ref. [30]). ARF, on the other hand, inhibits the cell cycle by activating the p53 tumor suppressor. In murine systems, Arf expression and has been shown to exert a general anti-aging effect [31,32], presumably through an enhancement of DNA repair. Likewise, Arf expression has been shown to play a specific role in vascular biology in the development of the murine eye [33]. Together, expression of p15 INK4b , p16 INK4a or ARF, either alone or in combination could plausibly retard atherogenesis, most likely by limiting pathogenic vascular proliferation.
As the putative non-coding RNA ANRIL spans the riskassociated SNPs, ANRIL might be directly affected by the unknown causative genetic variant in strong LD with rs10757278. Importantly, Polycomb group (PcG) complexes potently repress INK4/ARF locus expression [34,35,36], and cisacting non-coding RNAs have been shown to play an important role in PcG-mediated repression at other loci (e.g. Xist at the inactive X-chromosome, reviewed in [37,38]). Therefore, it is possible that altered ANRIL expression could potentially regulate INK4/ARF repression through PcG complexes, explaining why all four INK4/ARF transcripts are coordinately associated with rs10757278 genotype. On the other hand, ANRIL could also influence the expression of other genes in trans-as has been recently reported for HOTAIR at Hox loci [39].
Importantly, however, it is likely the effects of INK4/ARF expression on the development of atherosclerosis do not result from their expression in PBTL, but rather, observations in PBTL may serve as a surrogate for INK4/ARF expression in other tissues. As cis-regulatory elements may act as enhancers in one tissue and repressors in another, it is also possible that genetic variants in LD with rs10757278 genotype might be associated with increased INK4/ARF transcript expression in other tissues of greater relevance to atherogenesis (e.g. endothelial progenitors or vascular smooth muscle cells). Several investigators have argued that cellular senescence, which is induced by p16 INK4a , promotes atherosclerosis (e.g. [40,41]), and we do not believe the present results are necessarily inconsistent with that view. Carefully designed murine studies will serve to elucidate the effects of INK4/ARF expression on atherogenesis.
In summary, our data suggests that a common genetic variant in strong LD with rs10757278 at chromosome 9p21 influences the expression of the INK4/ARF-associated transcripts, but not MTAP, in PBTL. These observations provocatively suggest a crucial role for cell cycle inhibition in the development of atherosclerotic disease in humans.