This is the first study on the frequency distribution of all common sites of RNF213, including all exons and regulatory regions, in a Chinese cohort. The two SNPs, namely, rs8082521 and rs55996424, were identified to be associated with sICAS.
Protein RNF213 has a hexameric structure, which includes two functional domains: AAA + ATPase and E3 ligase; the protein dynamically changes its formation through adenosine-triphosphate (ATP)/adenosine diphosphate (ADP) binding and hydrolysis cycles.(3, 13) Gene RNF213 was initially identified as a susceptibility gene for MMD.(1, 3) Miyawaki et al. investigated the prevalence of p.R4810K (c.14576G > A) variant of RNF213 in a Japanese non-MMD ICAS patient population, and they found an association with anterior circulation ICAS.(14) Recently, Kamimura et al. found the RNF213 p.R4810K variant is correlated with early-onset ischemic stroke with anterior circulation stenosis in Japan.(15) According to statistics in 2012, approximately 16 million people in East Asian countries carry this genetic polymorphism.(2) In the process of study on MMD disease, it was found that RNF213 genetic variants may result in arterial fragility and susceptibility to hemodynamic stress,(16) and this mechanism can also lead to the pathogenesis of sICAS. In addition, a previous functional study suggested that Protein RNF213 may cause low angiogenic activity and high mitotic abnormalities.(1, 17) This finding has also been confirmed in animal experiments.(18, 19) Choi et al. used high-resolution magnetic resonance to analyze consecutive stroke patients, and draw the conclusion that RNF213 variant may be associated with vasculogenesis and smaller intracranial arteries which predisposing hemodynamic compromise in the presence of intracranial atherosclerosis.(20) On the basis of the above-stated studies, it may be stated that gene RNF213 and its transcriptional product may affect the morphogenesis and the morphological change in the intracranial artery,(21) thereby affecting the pathogenesis of sICAS.
The RNF213 common mutation, namely, rs8082521, is the conversion of cytosine (C) to adenine (A). The corresponding amino acid conversion is of glutamine (Q) to lysine (K). Limited studies have been conducted previously on the RNF213 rs8082521 site, and this site has not been reported in Caucasians. A study conducted with Japanese MMD patients in 2014 demonstrated a significant negative correlation between the RNF213 rs8082521 allele A and MMD, although no significant correlation was observed after performing adjustment for the p.R4810K genotype.(22) In the present study, it was observed that the rs8082521 locus genotype frequency and the allele frequency were significantly different between the sICAS group and the healthy control group. It was indicated that allele A decreases the risk of sICAS. Furthermore, the results of the analysis of the rs8082521 locus genotype frequencies revealed that rs8082521 correlated significantly with sICAS under the additive (AA vs. CC), dominant (AA + AC vs. CC), and recessive (AA vs. CC + AC) models (P = 0.0006; P = 0.0022; P = 0.0218). The results obtained were similar to the findings of previous research conducted in Japan on a related topic.
The RNF213 common mutation, namely, rs55996424, in which the base thymine (T) becomes adenine (A), is located in exon 17. Rs55996424 is involved in the coding of E3 ubiquitin-protein ligase RNF213 isoform 2, and the corresponding amino acid conversion is of lysine (K) to methionine (M). Only one previous study mentioned this mutation. A study conducted on Japanese hypertensive patients in 2015 demonstrated that there was no significant association between the rs55996424 and hypertension.(23) In our study, the rs55996424 locus genotype frequency and allele frequency were significantly different between the sICAS group and the healthy control group. It was indicated that the T allele increased the risk of sICAS. Furthermore, the results of the analysis of the rs55996424 locus genotype frequencies demonstrated that rs55996424 correlated significantly with sICAS under the additive (AA vs. TT) and recessive (AA vs. TT + AT) models (P = 0.0239; P = 0.0087). Although no statistically significant difference was observed under both additive and recessive models between the control and sICAS groups (P > 0.05/39 or P = 0.00128) after the Bonferroni correction. However, there remained a possibility of overcorrection, which could have resulted in false-negative results. The next step for future research could be increasing the sample size.
According to previous research, the reported pathogenic mutations, such as p.R4810K and p.R4859k, are located in the region between the exons 42 and 68, (1, 24) which corresponds to the region from the RING finger domain to the C-terminus of the RNF213 protein. The mutations in the C-terminus region of RNF213 could, therefore, be predicted to cause functional alterations in the protein.(25) We observed significant LD between rs55996424 and three other SNPs (rs8082521, rs77506504, and rs148731719) in both case and control groups; and significant LD between rs55996424 and nine other SNPs (rs72849841, rs61359568, rs61741791, rs35332090, rs61740658, rs116948489, rs113236556, rs112535386, and rs3185057) only in the control group. Among the aforementioned nine SNPs, rs116948489, rs113236556, rs112535386, and rs318505 are located in the 42–68 exon region. Since no significant LD was observed between rs55996424 and the four SNPs located in the 42–68 exon region, it may be stated with a certain degree of confidence that the pathogenicity of rs55996424 is not associated with the pathogenic genes that have been reported in previous studies.
A previous study conducted in Japan considered that nine SNPs (rs17857135, rs17853989, rs8082521, rs10782008, rs9913636, rs8074015, rs9674961, rs35332090, and rs61740658) located on the same haplotypes could be regarded as tag markers for p.R4810K susceptibility.(22) In our study’s haplotype analysis part, these nine SNPs were set as block 1 and block 3, and rs55996424 was set as block 2. The haplotype analysis was conducted in these three blocks, the result of which demonstrated no significant LD among these three blocks in both case and control groups. Therefore, the common mutation rs55996424 might be a novel independent risk factor for sICAS in the Chinese population.
The present study had certain limitations. Firstly, data of healthy control subjects was obtained from the GeneSky in-house database, and their clinical characteristic information was not available. Therefore, the impact of environmental risk factors could not be completely eliminated. Secondly, common mutations exhibit only minimal pathogenic effects and can explain only a small proportion (approximately 10%) of genetic risk, and sICAS is a polygenic disease. The third limitation was that this study was a single-center study, larger population-based multicenter studies are required in order to verify the findings of the present study.