Genetic analysis of RNF213 p.R4810K variant in non-moyamoya intracranial artery stenosis/occlusion disease in a Chinese population

The ring finger protein 213 (RNF213) was identified as a susceptibility gene for moyamoya disease (MMD) recently [1, 2]. Several studies revealed high frequencies of the same RNF213 variant- p.R4810K (rs112735431, GenBank accession number AB537889) in East Asian MMD patients compared to normal controls (ORs > 100), which was illustrated as a founder mutation in Japanese, Korean and Chinese patients [1‐9]. Indeed, in Japan and Korea, the majority (~80%) of MMD patients carried at least one allele of RNF213 p.R4810K, which was significantly higher than that in the general population [1, 2, 4‐6, 8‐16]. In China, the carrier rate of RNF213 p.R4810K in MMD patients was about 20% [1, 3, 17‐19]. MMD risk was highly increased by this variant in the Asian population.

RNF213 located in chromosome 17q25.3, encoding a 596 kDa protein which functions both as an AAA‑type ATPase and an E3 ligase [20]. AAA-ATPases mediate various cell functions, including membrane fusion/transport, proteolysis, protein disaggregation/refolding, DNA recombination/repair and mitosis/meiosis [21]. AAA-ATPase dysfunction can cause several diseases, such as PEX1/PEX6 mutations cause multiple organ degeneration [22, 23] and Cdc48 mutations cause amyotrophic lateral sclerosis [24, 25]. E3 ligase activity may play a role in protein degradation or signaling processes [1]. Even the complete physiological functions of RNF213 are still unknown, knockdown of RNF213 in zebrafish leads to abnormal sprouting and irregular diameter of intracranial vessels, suggesting some contribution to vascular formation [1]. Previous studies revealed that a wide spectrum of phenotypes could occur within a family unit despite the members having the identical p.R4810K variant, with some individuals showing the typical phenotype of MMD such as bilateral stenosis/occlusion of the terminal portion of the internal carotid arteries, some showing only unilateral or middle cerebral artery stenosis/occlusion, and others with no abnormalities [1]. Miyawaki et al. reported that a particular subset of Japanese intracranial major artery stenosis/occlusion (ICASO) mainly diagnosed as atherosclerosis not MMD, associated with RNF213 p.R4810K variant (odds ratio, 16.8; 95% confidence interval, 3.81–74.5; p < 0.0001), suggesting that RNF213 p.R4810K variant might cause various severities of ICASO [4, 6]. Recently, Bang et al. also reported that RNF213 p.R4810K is a susceptibility variant not only for MMD but also for ICASO in Korean (odds ratio, 22.3; 95% confidence interval, 3.0–164.1; p < 0.0001) [13]. ICASO is an important and the most frequent cause of cerebral ischemic stroke among patients of Asian ancestry [26]. Liu et al. explored the association of RNF213 p.R4810K with MMD in Chinese population [3, 5, 17‐19], however, no association data was available about this variant with ICASO in China.

This study totally included 715 patients with non-MMD ICASO and 507 normal individuals without known cerebrovascular diseases. Clinical characteristics are shown in Table 1. Six of 715 ICASO patients (0.84%) and 2 of the 507 normal controls (0.39%) had the RNF213 p.R4810K variant (all heterozygotes). Even no statistically significant association was observed, the carrier rate of RNF213 p.R4810K was higher in ICASO group than that in the normal individuals (Odds ratio, 2.14; 95% confidence interval, 0.43–10.63; p = 0.56).

Table 2 shows the clinical characteristics of 6 non-MMD ICASO patients with the RNF213 p.R4810K variant. The Fig. 1 shows the MRA images of the 4 patients with ICASO identified with the p.R4810K variant (the digital MRA images of the other 2 patients with the p.R4810K variant in the ICASO group was not available due to the fact that the patients were referrals from the other hospitals). These ICASO patients showed partial stenosis or occlusion of the intracranial major artery without abnormal vascular networks in the basal ganglia on MRA. All the patients had hypertension, 4 patients had diabetes and 2 elderly patients also had coronary heart disease for more than 20 years.

Table 3 shows the distribution of RNF213 p.R4810K in MMD, Non-MMD ICASO and normal individuals in Japan, Korea and China. The carrier rate of RNF213 p.R4810K in Chinese MMD and non-MMD ICASO patients was significantly lower than that in Korea and Japan.

RNF213 was a susceptibility gene for MMD [25]. Previous studies showed that in Japan and Korea, the founder variant RNF213 p.R4810K was much more frequent in MMD patients (~80%) than in the general population (~1.0%), significantly increased MMD risk (ORs > 100) [1, 2, 5, 8, 9, 25]. However, as a susceptibility gene, Chinese MMD patients have significantly different genetic architecture. The carrier rate of RNF213 p.R4810K in Chinese MMD and general population is about 20 ~ 30% and 0.3% respectively, accounting for less part of MMD risk [1, 3, 5, 17‐19]. The genetic result is consistent with the unique epidemiological and clinical characteristics of Chinese MMD. In China, no significant difference in sex distribution of MMD, a female predominance is not observed compared to Japan and South Korea. Moreover, the familial occurrence of MMD is lower and the symptoms at the onset are different from those in Japan and South Korea [28]. Genetic heterogeneity is proposed to be partially responsible for the different clinical features of MMD in different ethnicities. It has been proposed that other rare variants of RNF213 may be causative mutations for MMD patients. Zhang et al. systematically investigated RNF213 variants of 255 Chinese MMD, revealing that p.R4810K was identified in 31.4% MMD patients and additional 25 rare variants (absent in controls) were identified in 10.6% of patients without p.R4810K variant. Segregation analysis supported the association between MMD and 2 novel variants (p.H4014T and p.R4160Q) [18]. Similar with MMD, genetic heterogeneity of ICASO in different population was indicated. In previous Japanese and Korean studies [4, 6, 13], a particular subset of ICASO mainly diagnosed as atherosclerosis, associated with RNF213 p.R4810K variant (odds ratios, 16.8 and 22.3; 95% confidence intervals, 3.81–74.5 and 3.0–164.1; both p < 0.0001 for Japanese and Korean, respectively). These findings strongly indicate that some cases of ICASO ascribed to unknown etiology or atherosclerosis might be caused by RNF213 p.R4810K variant. They proposed that RNF213 p.R4810K variant could contribute to the high prevalence of intracranial atherosclerotic stroke in Asians. However, in our study, only 0.84% Chinese ICASO patients (6/715) carried RNF213 p.R4810K variant, no statistically significant association of RNF213 p.R4810K variant with ICASO was observed. RNF213 p.R4810K variant is unlikely playing major role in Chinese ICASO. Population-specific variants might contribute a lot to the pathogenesis. In addition, it has been widely accepted that MMD and ICASO is caused by both genetic and environmental factors. In this study, all 6 ICASO patients with RNF213 p.R4810K variant had hypertension and 4 patients had diabetes. Other unknown genetic and environmental factors may trigger ICASO together. Further systematically investigate Chinese-specific genetic variants of ICASO is needed.

The limitations of this study should be mentioned. First, the analysis of the RNF213 variant was exclusively focused on the RNF213 p.R4810K variant. The other RNF213 variants such as p.H4014T and p.R4160Q were not evaluated. Further comprehensive genetic analysis of RNF213 and other potential genes is necessary to determine whether patients with ICASO without the p.R4810K variant have other variants or not. Second, because the strictly matched cerebrovascular disease-free controls were not available in this study, we used the allele frequency of p.R4810K in previous published control population as control in the association analysis. This may perturb the association results. However, such an approach could be justifiable for rare variants and could not perturb the results significantly, especially in the study with a low carrier rate of target variant in the case group. Third, due to the allele frequency of RNF213 R4810K was low in both case and normal individual group in China, statistic power might be low when exploring the association between RNF213 p.R4810K variant and ICASO in this study. Relevant studies with larger sample sizes are needed to validate our findings.

Even no statistical significance, the carrier rate of RNF213 p.R4810K was higher in ICASO group than that in the normal individuals in China. However, the carrier rate of RNF213 p.R4810K in Chinese ICASO patients was significantly lower than that in Korea or Japan. Population genetic heterogeneity was highly indicated. Further systematic genetic epidemiology studies with larger sample sizes focusing on Chinese-specific genetic variants and environmental risk factors of ICASO are needed.