In this study, our aim was to investigate the association between 9p21.3-variants and the severity of CAD in the presence and absence of T2D. Results from our analysis showed that two 9p21.3-variants (rs4977574 and rs2383207) from the known CAD-associated haplotype, are also associated with severity of CAD. We also showed that these variants are associated with severity of CAD in subjects with and without T2D. Additionally, a third SNP rs10738610, located inside the same CAD-associated haplotype, was found to be significantly associated with severity of CAD in subjects with T2D. The significance of the association for this SNP in the pooled analysis was slightly above the threshold; however, this heightened significance was strengthened after stratification by T2D. Association of rs4977574 with severity of CAD was confirmed in an independent study in subjects without T2D. The results of the association for SNP rs1333049, which is a well-documented variant in the association with incidence of CAD and is in high LD with rs494457 (r
= 0.89), rs2383207 (r
= 0.89), and rs10738610 (r
= 0.93), showed no association with severity of CAD in the Italian and German studies. However, Dandona et al.  in the Canadian study demonstrated that this SNP predicted the severity of coronary atheromatous burden in patients without T2D.
From our results, we hypothesized that although the 9p21.3 locus does not contain any genes, there are two different pathological mechanisms harboring the same haplotype, one that shares a common genetic platform between molecular mechanisms of CAD and T2D and the other, an independent one that is linked to the pathogenesis of CAD alone. Additionally, disagreement on the associations with severity of CAD of well-documented 9p21.3 variants associated with CAD variants may be due to various factors, including phenotype heterogeneity and difference on risk-allele frequencies among participating studies. The mechanisms of severity of CAD and T2D in patients with acute coronary syndrome (ACS) are complex, and they are not so clear at present. While many genetic studies, including GWAS, for CAD and T2D, have identified the 9p21.3 locus as a common risk region (determined by two adjacent haplotypes), there are no functional variants (non-coding variants) within this region. Therefore, the link between the 9p21.3 locus and the genetic backbones of these two diseases still remains poorly understood. The 9p21.3 locus lies in proximity to CDKN2A and CDKN2B. These cyclin-dependent kinase inhibitors (CDKNs) generate several transcript variants that are involved in diverse biological processes, including cell proliferation, cellular senescence, cell cycle arrest, induction of apoptosis, and activation of caspase activity, to name a few [35, 36]. Some of these biological processes are presumably implicated in the pathogenesis of atherosclerosis as well as in the other complex phenotypes (i.e., myocardial infarction, aortic aneurysm, ischemic stroke, T2D, glioma, and malignant melanoma). Recent works have speculated that the pathogenic role of the 9p21.3 locus in atherosclerosis is influenced by gene expression [37, 38] and a transcriptional and translational control mechanism presumably caused by cis- and trans-acting elements located in the vicinity of this region. Yet, no cis- and/or trans-acting mechanisms have been identified. Other works have elaborated many discussions about the analysis of the 9p21.3 locus and the appropriate phenotypic characterization. Thus, genotypic variability among 9p21.3 variants may be due to wrong phenotyping rather than sequence variations influenced by gene expression. Moreover, it seems that the significance of variants found at this locus is changeable to the phenotype of interest, its severity, and the presence of combined risk and environmental factors that correlated with the phenotype.
In our study, we have a well-characterized phenotype for severity of CAD in subjects with and without T2D. We performed association analyses in three different white populations of European ancestry, totaling 5,886 subjects. This enabled us to study the relationship between the 9p21.3 locus, including CAD- and T2D-associated variants, and the severity of CAD in subjects with and without T2D. Moreover, to reduce phenotype-genotype bias, we used additional scoring methods for assessing extension of CAD. Then, we conducted association analyses following two approaches for assessing the relationship between severity of CAD and 9p21.3-variants. Our finding showed (1) that rs4977574 and rs2383207 at the 9p21.3 locus are associated with severity of CAD regardless of T2D status and (2) that rs10738610 is associated with severity of CAD comorbid with T2D.
This study has limitations: (1) in order to characterize the 9p21.3 locus more effectively in subjects with severity of CAD and T2D, more variants need to be included in the analysis and hence tested for association with this phenotype; (2) observed differences among genetic effects of 9p21.3 variants across populations may be due to the quantitative method in which angiographic scores were calculated in each participating study; (3) lack of availability of variants rs2383207 and rs10738610 in replication stage demands validation of their associations with severity of CAD in other white population studies with similar characteristics; (4) phenotype heterogeneity among participating studies is likely to be present since each study was previously designed to answer its own research question and had predefined inclusion and exclusion criteria; (5) In the Italian study, although we cannot discard the presence of atherosclerosis in the 1,212 controls, preliminary testing between 9p21.3 variants and incidence of CAD obtained from follow-up records showed no association with CAD; (6) lack of availability of 3-vessel disease subjects in the German study may had underestimated the genetic effect of SNP rs1333049 attained in the replication stage; and (7) lack of availability of genotype data of SNP rs497754 in late-onset CAD subjects with severity of CAD in the Canadian study requires a replication in another cohort with similar characteristics as to strengthen the significance of this SNP with severity of CAD regardless on the onset of the disease.