Atherosclerosis is currently thought to be an inflammatory disease with deposits of cholesterol and inflammatory cells in the arterial wall [1‐3]. In the later stages, thrombosis forms, and this leads to fatal events, like myocardial and cerebral infarctions [4, 5]. The etiology of atherosclerosis is multi-factorial, in that it includes various genetic and environmental factors [6, 7]. It is known that genetic factors play a pivotal role in atherosclerosis, but the influence of specific genes on susceptibility is not fully understood [8‐10]. The discovery of novel risk factors for atherosclerosis may lead to efficient prevention and new therapeutic targets for this fatal disease.

Recently, two independent genome-wide association studies (GWAS) of atherosclerotic diseases in Caucasians pointed to two SNPs, closely resided in the 31.1 Kb (intron 14 ~ exon 19) region of the RYR3 gene [11, 12]. RYR3 gene is located on chromosome 15q14-q15, and encode a large intracellular homotetrameric protein (> 2MDa) that comprises 4780 amino acids [13‐15]; RYR3 protein resides on the sarcoplasmic reticulum membrane and releases Ca2+ from intracellular stores to regulates intracellular calcium concentration [16, 17]. The RyR3-deficient mice lack calcium regulation properties in arterial smooth muscle cells, which lead to dysregulation of arterial tone [18]. In human arterial endothelial cells, Ca2+ release mediated by RYR3, but not by the RYR1 and RYR2, plays a role in endothelial vasodilation [19]. Thus, it is conceivable that RYR3 is a good candidate for atherosclerosis susceptible gene. The rs877087 polymorphism, which is an intron SNP, was associated with stroke in a Caucasian population from 4 large cohorts (n = 19,602) [11]. The rs2229116, which is an Ile to Val non-synonymous SNP, was associated with common carotid intima-media thickness (cIMT), a clinical parameter for subclinical atherosclerosis [20]. This was found in a specific group of Caucasian males infected with HIV and were treated with highly active antiretroviral therapy (n = 177) [12]. Since rs877087 and rs2229116 on RYR3 gene is separated by 31.1 Kb, we hypothesize that this region might confer susceptibility to atherosclerosis, and thus studied this region in detail using tag SNPs and haplotypes.

The previous GWAS showed that the RYR3 gene polymorphisms rs877087 and rs2229116 was associated with stroke and cIMT, respectively [11, 12]. In this study, we investigated these two SNPs on RYR3 gene and the 31.1 Kb region in between employing elderly Japanese. We observed that both rs877087 and rs2229116 was associated with PAI. Our study suggested that the atherosclerosis risk allele is T for rs877087, as the same risk allele from the GWAS of stroke. It should be noted that the minor allele frequency of rs877087 is 0.035 for C allele in Japanese, and 0.42 for T allele in Caucasian according to the HapMap data. rs2229116 was associated with PAI and atherosclerosis severity in common carotid after adjusted for conventional risk factors (Table 2). However, our risk allele was opposite from the original paper [12], in that, the GWAS for cIMT the risk allele was G, but in our study it was A (Table 3). The LD patterns between rs877087 and rs2229116 from HapMap data showed great differences between CEU and JPT (Figure 1 and Additional file 3: Figure S1 and Additional file 4: Figure S2). As reference to rs2229116, the LD values of two populations showed large difference (ΔD’ > 0.5) in several SNP sites, including the 2 tag SNPs (rs2132207 and rs658750), suggesting that the haplotypes within this region may be different between CEU and JPT (Figure 1). Therefore, different haplotypes might account for the different observation between Japanese and European. In order to further investigate the region between the two reported SNPs, we further selected 2 tag SNPs (rs2132207 and rs658750) and build haplotypes (Table 1). Haplotype linear regression analyses revealed that the carriers of TAGG haplotype had a decrease trend of atherosclerosis severity in both common carotid arteries and PAI. Also, an increase trend was observed in TGGA haplotype in common carotid arteries and TAAA haplotypes in PAI, respectively (Table 4). In haplotype logistic regression model for atherosclerosis severity trait, we observed TAGG haplotype had a preventive effect in both common carotid arteries and PAI (Table 4). Although we tried to explain the opposite effect in Japanese and Caucasian, using trans-ethnic fine mapping, and selected these two tag SNPs which gave a significant difference in LD value between two ethnics, we could not explain the association with either of the SNP. Nevertheless, taken together the results of haplotype analysis, it is likely that the non-synonymous polymorphism rs2229116 (Val/Ile) itself may be the primary determinant for the observed associations (Table 4).

Along with the function of RYR3 in vascular cells, recent studies have found that RYR3 have function in T cells, which is involved in the inflammation process of atherosclerotic lesions [27, 28]. RYR3 appear to regulate Ca2+ signaling in T lymphocytes [29], in which, calcium signaling is essential for its activation and differentiation [30]. The association of T cell activation markers and carotid arterial stiffness was also reported in a group of HIV-infected women [31]. These lines of evidence suggest that, intracellular calcium mobilization mediated by RYR3 may be involved in atherosclerosis process through its common participants, such as arterial smooth muscle cells, endothelial cells and T lymphocytes.

Our study has some limitations. The subjects of the present study were consecutive autopsy cases of patients in a geriatric hospital. Therefore, the selection bias and whether the subjects can represent the demographics of Japan should be evaluated. The autopsy rate of the hospital was kept around 40% and those who died suddenly outside the hospital and medicolegal cases were not autopsied for this study. The leading mortality probability at the age of 75 years according the census data in ‘Abridged Life Tables For Japan 2010’ by the Ministry of Health, Labor and Welfare of Japan (http://​www.​mhlw.​go.​jp/​english/​database/​db-hw/​lifetb10/​index.​html) were as follows; 25.9% men vs. 16.5% women by malignant neoplasms, 15.2% men vs. 20.3% women by heart diseases, 10.4% men vs. 12.1% women by cerebrovascular diseases, and 15.3% men vs. 12.4% women by pneumonia. Almost all death rates for major diseases at 75 years of age in Tokyo were consistent with our autopsy data, except for the frequency of malignant neoplasms is a little higher and that of cerebrovascular diseases is lower in our autopsy subjects. Since the national health insurance covers the whole population of Japan and in Japan most death occurred in hospitals, the selection bias from socioeconomic difference of the subjects seems minimal. Altogether, the subjects in our study did not differ greatly from the standard elderly residents in Tokyo, and may represent the general population, including young residents, in Japan rather than a selected hospitalized population.

RYR3 gene rs877087 and rs2229116 polymorphisms are associated with atherosclerosis in elderly Japanese. Functional studies are still needed to identify the causal variation.