Association of DIAPH1 gene polymorphisms with ischemic stroke

Demographic and clinical characteristics of the study population

Clinic-demographic characteristics of participants in the hypertension case-control study are summarized in Supplementary Table 1. Although study subjects were matched for age (5 year-group), hypertensive cases were on average 3.42 years older than controls (P < 0.001). Participants with hypertension had higher BMI, total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), glucose (GLU), and a higher rate of type 2 diabetes mellitus (T2DM) than controls (P < 0.001). No significant differences in gender, high-density lipoprotein cholesterol (HDL-C), or smoking and drinking statuses were observed (P > 0.05).

The characteristics of individuals in the stroke case-control study are summarized in Supplementary Table 2. Significant differences were observed among IS, HS, and controls for age, gender, smoking and drinking habits, hypertension, lipid profiles, and T2DM (P < 0.05). Post-hoc multiple comparisons showed that total TC, HDL-C, and LDL-C levels were significantly higher in IS cases than in controls. Compared to controls, HS cases were older, had higher levels of TC and HDL-C, and lower TG levels. These characteristics were adjusted as confounding factors when evaluating the association of DIAPH1 with stroke.

Association analysis of DIAPH1 variants in the case-control study of hypertension

In the case-control study of hypertension, the allele frequencies of the DIAPH1 rs3805691, rs251019, and rs11954998 SNPs in controls were consistent with Hardy-Weinberg equilibrium (HWE), whereas the allele frequencies of rs251018 and rs7703688 were not (HWEP = 0.004 and P < 0.001, respectively). Compared with CC/CT carriers, the rs251019 TT genotype was significantly associated with decreased risk of hypertension, after adjusting for covariates including age, gender, BMI, TC, TG, HDL-C, LDL-C, GLU, smoking status, and drinking status [adjusted odds ratio (OR), 0.791; 95% confidence interval (CI), 0.636-0.984; P=0.035] (Table 1).

Association analysis of DIAPH1 variants in the case-control study of IS

In the case-control study of stroke, the frequencies of all DIAPH1 SNPs in controls were consistent with HWE. Table 2 shows the results of the association analyses after adjusting for age, gender, smoking status, drinking status, TC, TG, LDL-C, HDL-C, T2DM, and hypertension. The additive model (CC vs CT vs TT) suggested that the rs3805691 variant was associated with decreased risk of IS (adjusted OR=0.782, 95% CI=0.700–0.874, P=1.3×10^-5). Compared with CC carriers, CT/TT carriers of the rs3805691 SNP were also at lower risk of IS (adjusted OR=0.712, 95% CI=0.622–0.815, P=7.986×10^-7). Comparable protective effects against stroke were observed for the rs251019 variant in both additive and dominant models. In contrast, rs7703688 and rs11954998 C allele carriers were associated with increased risk of IS [ORs (95% CIs) for the additive model were 1.721 (1.486-1.993) and 1.537 (1.356-1.743), P=4.139×10^-12 and P=2.058×10^-11, respectively]. These associations were still significant after Bonferroni correction. No significant association with IS was found for rs251018, nor between the five DIAPH1 SNPs and HS.

Furthermore, we conducted SNP association analysis stratified by TOAST subtypes (Supplementary Table 8). After covariates adjustment, both rs3805691 and rs251019 were negatively associated with small artery occlusion (SAO) and large-artery atherosclerosis (LAA) under the additive and dominant models. Meanwhile, the rs11954998 and rs7703688 variants were instead associated with increased risks of SAO and LAA under all three genetic models, while no association was detected between rs251018 and either SAO or LAA. Analysis for HS subtypes showed that compared to the TT/TC genotypes, the CC genotype of rs7703688 conferred higher risk for subarachnoid hemorrhage (SAH), whereas none of the DIAPH1 SNPs studied showed association with intracerebral hemorrhage (ICH) (Supplementary Table 9).

Association analysis of DIAPH1 variants in the cohort study of hypertension and stroke

The clinic-demographic characteristics of participants in the cohort study of hypertension and stroke are shown in Supplementary Table 3. No significant associations between selected DIAPH1 gene variants and hypertension were observed (Table 3). Regarding stroke, rs251018 GG genotype carriers showed significantly higher incidence rate than TT/TG carriers after adjusting for age, gender, TC, TG, HDL-C, LDL-C, smoking, drinking, BMI, T2DM, and hypertension. Increased risk for stroke was also found for rs7703688 genotypes in the additive and recessive models (P=0.049 and P=0.042, respectively; Table 3).

Comparison of DIAPH1 mRNA expression between IS and controls

Comparative analysis of mRNA expression for the selected SNPs was further conducted in 58 controls and 66 IS cases (43 SAO and 23 LAA). Compared with hypertensive controls, the expression of DIAPH1 mRNA was significantly downregulated in IS [0.773 (0.575, 1.088) vs 0.933 (0.775, 1.117), P = 0.003]. Results are depicted in Figure 1. The expression of DIAPH1 mRNA among the genotypes of rs3805691, rs251018, rs251019, rs11954998, and rs7703688 did not differ significantly, neither in IS cases nor in controls (Supplementary Figure 2). However, the DIAPH1 mRNA level in SAO was upregulated with rs251019 genotypes, especially in homozygous CC carriers (mean expression levels in TT, TC, and CC carriers were 0.742, 0.889, and 1.765, respectively, P_trend = 0.048; Supplementary Figure 3).

Study population

Case-control and cohort studies were conducted to investigate the association of polymorphisms in the DIAPH1 gene with hypertension and stroke. A total of 4,128 participants from the community hypertension survey were recruited between 2009 and 2010 from Guanlin and Xushe towns in Yixing city (Jiangsu, China). In the case-control study of hypertension, 2,012 patients with hypertension and 2,116 controls were recruited. Hypertension was determined as systolic blood pressure (SBP) ≥140 mmHg and/or diastolic blood pressure (DBP) ≥90 mmHg, or currently receiving anti-hypertensive medication. Subjects who had a clinical history of secondary hypertension were excluded. As the average age was higher in hypertension cases than in controls, 94 elder controls were selected from local communities for age-matching with hypertension cases. The demographic characteristics of the sample population are listed in Supplementary Table 1, and have been previously reported [25].

In the case-control study of stroke, 2,212 IS and 754 HS cases were recruited between 2013 and 2017 in the Peopleʼs Hospital of Yixing City. All stroke cases were admitted with first-time acute stroke within 72 hours. Individuals older than 85 years were excluded from the study. 2,590 age (± 5 years)- and gender-matched controls were selected from community members. The demographic and clinical characteristics of the studied population are summarized in Supplementary Table 2.

IS and HS sub-types were confirmed by a neurologist according to medical records of computed tomography (CT) and/or magnetic resonance imaging (MRI). As per TOAST criteria [26], 2,212 IS cases were classified into 1,199 SAO, 882 LAA, 108 cardiogenic cerebral embolisms (CE), 12 strokes of undetermined etiology (SUE), and 11 strokes of other determined etiology (SOE). In turn, 754 HS cases were classified into 103 SAH and 651 intracerebral hemorrhages (ICH).

For cohort studies, 2,116 participants with normal blood pressure were enrolled for the hypertension study, while 4,098 subjects were enrolled for the study assessing stroke. The clinic-demographic characteristics of these populations are listed in Supplementary Table 3. During a median follow-up time of 5.01 years, 613 cases of hypertension and 183 strokes (171 IS and 12 HS) were recorded. The flow chart of the study design is outlined in Supplementary Figure 1.

Interviews, physical examinations, and laboratory tests were conducted for all participants. Demographic characteristics including age, gender, smoking status and drinking status were obtained by trained research staff though a standard questionnaire. Weight, height, and blood pressure measurements were obtained by trained assistants according to standard protocols.

Drinking habit was defined as self-reported drinking frequency (current or past consumption of an alcoholic beverage at least 2 times per week for at least 6 months per year). Smoking habit was defined as current or past consumption of at least 20 cigarettes per week for at least 3 months per year.

The research protocol was approved by the ethics committee of Nanjing Medical University. All participants were informed in detail about the investigation and voluntarily signed the informed consent form.

SNP selection

We selected SNPs covering the DIAPH1 gene within 5 kb upstream and 2 kb downstream of the 5' and 3’ ends of the transcript, respectively, according to the International Hap MAP Project database (HapMap Data Rel 24/phase II Nov08, on NCBI B36 assembly, dbSNPb126). All SNPs were selected with the criteria of minor allele frequency (MAF) ≥0.05 and linkage disequilibrium (LD) r² ≥0.8. SNPs with predictive biological effects and functions were obtained from HaploReg v4.1 [27] and rerun as tagSNPs. Finally, five tagSNPs, rs3805691, rs251018, rs251019, rs7703688, and rs11954998 were selected. Corresponding biological information, primers, and probes are summarized in Supplementary Tables 4, 5.

Blood sampling and SNP genotyping

Blood samples were collected in EDTA-containing receptacles after overnight fasting (>10 h), and TC, TG, HDL-C, LDL-C, and GLU were measured. Stroke cases consented to donate 5 ml venous blood to the Department of Neurology after admission. Anticoagulated samples were stored at −20°C within 24 h after separating serum and plasma.

Genomic DNA was isolated applying a standard phenol-chloroform method. Genotyping was performed using the TaqMan allelic discrimination assay on a 7900HT Real-time PCR System (Applied Biosystems, Foster City, CA). Each plate included blank samples as negative controls to verify genotyping quality. Call rates for each SNP were > 99.9%.

DIAPH1 mRNA measurement

To explore whether DIAPH1 was differentially expressed between IS cases and controls, 66 inpatients with newly diagnosed IS were recruited from Yixing People's Hospital (from Jan to Nov. 2017). Considering that hypertension is a dominant characteristic of IS, we selected 58 age- and gender- matched hypertensive controls to compare the expression of DIAPH1 at the mRNA level. To this end, PBMCs were isolated within 4 h after blood extraction and DIAPH1 mRNA was isolated and quantified using a standard protocol (Supplementary Methods). Primer sequences for DIAPH1 and control GAPDH mRNAs are listed in Supplementary Table 6. In addition, DNA genotyping was conducted in both groups to investigate potential association of DIAPH1 polymorphisms with mRNA level variation. Clinic-demographic characteristics of IS cases and hypertensive controls are summarized in Supplementary Table 7.

Statistical analysis

Unpaired Student’s t-tests were used to assess inter-group differences for quantitative variables, presented as means ± SD. HWE for genotype frequencies was estimated with a Fisher's exact test in controls. For comparisons between not normally distributed independent samples, the Mann–Whitney U test was applied. Unconditional logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals (CIs) as well as to adjust for covariates. Cox regression was applied to estimate hazard ratios (HRs) and 95% CIs in the cohort study. A two-tailed P value of 0.05 was defined as the cutoff for statistical significance. All statistical analyses were performed with SPSS version 18.0 (SPSS, Inc., Chicago, IL).