Background. Despite the significant efforts made by the world’s medical community in the field of prevention and treatment of stroke, this disease remains the second leading cause of death worldwide, and a leading cause of adult disability [1], cognitive decline and dementia [2]. Ischemic stroke (IS) comprises up to 85% of all cases [3]. The risk of IS is determined by a complex interplay of genetic and environmental factors partly acting through modifiable risk factors such as high blood pressure [4, 5].

The previous studies have already revealed genetic markers associated with the development of IS [6, 7, 8], as well as related phenotypes/risk factors, such as hypertension [9, 10, 11], atherosclerosis [12, 13], and thrombosis [14]. However, despite the large number of studies conducted in the world, the molecular mechanisms of IS development remain poorly understood.

One of the most promising fields in this area is the study of chaperone proteins. A critical decrease in perfusion in IS leads to hypoxic damage to brain tissues [15]. Cascades of ischemic and reperfusion damage to neurons and glia include hundreds of individual links that lead to cell death. It is the critical accumulation of damaged enzymes and structural proteins that causes irreversible changes that lead to necrosis or programmed cell death [16]. The main function of chaperones is to restore the native tertiary and quaternary structure of damaged proteins. There is already enough information that confirms the neuroprotective role of chaperones in IS, for example, the HSP class of chaperones ("heat shock" proteins) [17].

Recently, a new class of chaperones, "Hero proteins", has been discovered, which main mechanism of action is the physical shielding of other proteins. Such a function is provided by a flexible structure and a high negative or positive charge of molecules [18]. Moreover, Hero proteins can suppress some forms of pathogenic protein aggregates in cells and in animal models of neurodegenerative diseases in vivo. The described properties of Hero let us suggest their wide involvement in the physiological and pathological processes occurring in cells with ischemic damage.

The aim of this study was to analyze association of the tagger SNP rs2900262 in gene C9orf16 encoding the member of Hero proteins with the risk and clinical manifestations of ischemic stroke.

Material and methods. A total of 1929 unrelated Russians (861 patients with IS and 1068 healthy individuals) from Central Russia were recruited into the study. The Ethical Review Committee of Kursk State Medical University approved the study protocol. All the participants gave written informed consent before the enrollment in this study, subject to the following inclusion criteria: self-declared Russian descent, a birthplace inside of Central Russia.

Baseline and clinical characteristics of the study population are listed in Table 1. The patients were enrolled in to the study in two periods: from the Regional Vascular Center of Kursk Regional Clinical Hospital between 2015 and 2017 [19, 20] and Neurology Clinics of Kursk Emergency Medicine Hospital Kursk between 2010 and 2012 [21, 22]. All the patients were examined by qualified neurologists. The diagnosis of IS was made in the acute phase of stroke, according to the results of the neurological examination and computed tomography and/or magnetic resonance imaging of the brain. The patients were recruited consecutively. The IS patients were enrolled under the following exclusion criteria: hepatic or renal failure or endocrine, autoimmune, oncological, or other diseases that can cause an acute cerebrovascular event; intracerebral hemorrhage, hemodynamic or dissection-related stroke, traumatic brain injury. All the patients with IS had a history of hypertension and received antihypertensive therapy.

The control group was compiled from healthy volunteers with no clinical evidence of cerebrovascular, cardiovascular, or other chronic diseases and with normal blood pressure without antihypertensive therapy. Healthy individuals were included in control group if they had a systolic blood pressure less than 130 mm Hg and/or a diastolic blood pressure less than 85 mm Hg on at least 3 separate measurements. Control subjects were enrolled from Kursk hospitals during periodic medical examinations at public institutions and industrial enterprises of Kursk region. This group was recruited from the same population and during the same period [23, 24].

The following criteria were used in the selection of SNPs: the SNP must be tagging, have a minor allele frequency of at least 0.05 in the European population and be characterized by a high regulatory potential. According to the bioinformatic tool GenePipe (https://snpinfo.niehs.nih.gov/snpinfo/selegene.html), which was used to select SNPs based on the reference haplotypic structure of the Caucasian population (CEU) of the project HapMap, the only tag SNP in the gene C9orf16 (also known as BBLN, ID:79095) is rs2900262. This genetic variant is localized in the intron. Several bioinformatic resources were used to assess the regulatory potential of this SNP. Bioinformatic resource HaploReg (v4.1) established a connection of rs2900262 C9orf16 with modifications of histones marking promoters in 15 different tissues; it established the localization of this SNP in the regions of hypersensitivity to deoxyribonuclease in 7 issues, regions of 4 regulatory motifs (https://pubs.broadinstitute.org/mammals/haploreg/haploreg.php). According to rSNPBase (http://rsnp.psych.ac.cn/index.do), it is characterized by distal transcriptional regulation, the regulation mediated by RNA-binding proteins and regulation mediated by microRNA. The instrument RegulomeDB allowed to reveal that rs2900262 C9orf16 is characterized by a regulatory coefficient of 1b (eQTL + TF binding + any motif + DNase Footprint + DNase peak) (http://regulome.stanford.edu/). According to the date presented in Ensembl genome browser, this genetic variant is characterized by an average frequency of the minor T allele in European populations of 0.057 (https://www.ensembl.org/). Thus, rs2900262 C9orf16 was selected for this study, which meets the necessary criteria for inclusion in the study.

Genetic Analysis. DNA analysis was carried out at the Laboratory of Genomic Research of Research Institute for Genetic and Molecular Epidemiology of Kursk State Medical University (Kursk, Russia). Approximately 5 mL of venous blood was collected from the cubital vein of each participant into EDTA-coated tubes and maintained at −20◦C until processed. Genomic DNA was extracted from thawed blood samples by the standard procedure of phenol/chloroform extraction and ethanol precipitation. Genotyping of the SNP was done using Taq‐Man–based real‐time polymerase chain reaction assays according to the protocol designed in the Laboratory of Genomic Research, Research Institute for Genetic and Molecular Epidemiology. Primers and probes were designed (http://primer3.ut.ee/) using the Primer3 online program [25], and then synthesized by the Syntol company (Moscow, Russia). The 2 flanking primers such as forward 5′- ATATACCATGAGGGGGTTGGA -3′ and reverse 5′- GAGGAGCTGGAGGAGCGTAT-3′ and also allele-specific probes T-5′-FAM-CCTTTCCCAGAACACA-RTQ1-3′ and 5′-ROX-CCTTTCCTAGAACACA-BHQ2-3′ were used for genotyping of the polymorphism. A real-time PCR was done in a 25-mL reaction mixture containing 1.5 U of Hot Start Taq DNA polymerase (Biolabmix, Novosibirsk, Russia), about 1 μg of DNA, 0.25 μM each primer, 250 μM dNTPs, 3.0 mM MgCl₂, 1xPCR buffer [67 mM Tris-HCl, pH 8.8, 16.6 mM (NH4)₂SO₄, 0.01% Tween-20]. The amplification reaction consisted of an initial denaturation for 10 minutes at 95°C, followed by 39 cycles of 92°C for 30 seconds and 58°C for 1 minute. Figure 1 shows allelic discrimination plot for SNP rs2900262 C9orf16 assay designed for this study.

The plot shows clear separation between the signals derived from allele 1 (rs2900262-C, FAM fluorescent dye) or allele 2 rs2900262-T, ROX fluorescent dye). Genotypes CC, CT, and TT are shown as circles, triangles and squares, respectively. Primer sequences used for genotyping are available upon request. To ensure quality control, 10% of DNA samples were genotyped in duplicates blinded to the case-control status. The concordance rate was >99%.

Allele and genotype frequencies in the case and control groups were counted and compared by the Fisher exact test to identify significant departures from Hardy–Weinberg equilibrium. Genotype frequencies in the study groups and their associations with disease risk were analyzed using the SNPStats software (https://www.snpstats.net/start.htm) [26]. For the analysis of associations of genotypes, dominant, recessive, and additive models were considered; taking into account the number of models of associations, a correction for multiple testing was applied using Benjamini-Hogberg false-discovery rate approach.

Multiple logistic regression analysis was performed to evaluate the associations of rs2900262 C9orf16 genotypes with the risk of ischemic events. Continuous variables were analyzed for normal distribution by the Kolmogorov–Smirnov test. Linear regression analysis was used to evaluate the association between rs4644832 and age of ischemic events, which was previously transformed to a normal variable through the procedure of inverse transformation of ranks. All associations were adjusted for age, gender and smoking status.

The following bioinformatics recourses were used for analysis of regulatory potential of rs2900262 C9orf16:

- The rSNPBase tool (http://rsnp.psych.ac.cn/index.do) was used to evaluate the effect of SNPs on proximal transcriptional regulation, distal transcriptional regulation, RNA-binding protein-mediated regulation, and miRNA-mediated regulation (accessed on 7 May 2022) [27].

- RegulomeDB (Version 1.1) (http://regulome.stanford.edu/) was used to estimate the regulatory coefficient of SNP (accessed on 7 May 2022) [28].

- GTExportal (http://www.gtexportal.org/) was used to analyze the expression levels of the studied genes in brain, whole blood, and blood vessels, as well as to analyze the binding of SNPs to quantitative expression trait loci (eQTLs) (accessed on 7 May 2022) [29].

- Additionally, to analyze the binding of rs2900262 C9orf16 to quantitative expression trait loci (eQTL) in peripheral blood, the eQTLGen resource was used (https://www.eqtlgen.org/) (accessed on 7 May 2022) [30].

- Bioinformatics resource HaploReg (v4.1) (http://archive.broadinstitute.org/mammals/haploreg/haploreg.php) was used to assess the association of rs2900262 C9orf16 with histone modifications marking promoters and enhancers: mono-methylation at the 4th lysine residue of the histone H3 protein (H3K4me1), tri-methylation at the 4th lysine residue of the histone H3 protein (H3K4me3), the acetylation at the 9th lysine residues of the histone H3 protein (H3K9ac), acetylation of the lysine residues at N-terminal position 27 of the histone H3 protein (H3K27ac). This resource has also been used to analyze the localization of SNPs in regions of DNase hypersensitivity, regions of regulatory motifs, and sites binding with regulatory proteins (accessed on 7 May 2022) [31].

- To assess the effect of rs2900262 C9orf16 on the binding of transcription factors (TFs) to DNA, depending on the carriage of the reference/alternative alleles, the atSNP Function Prediction online tool was used (http://atsnp.biostat.wisc.edu/search; accessed on 14 May 2022) [32]. TFs were included in the analysis only under the condition of a high and very high degree of influence of SNP on the interaction of TFs with DNA, calculated on the basis of a positional weight matrix. The subsequent analysis of the possible joint participation of TFs associated with the reference/SNP alleles in common biological processes directly related to the pathogenesis of IS was carried out using the Enrichr online tool (https://maayanlab.cloud/Enrichr/) [33]. Biological functions controlled by transcription factors associated with SNP rs2900262 C9orf16 were used as functional groups (accessed on 14 May 2022). The level of significance of gene ontologies, taking into account multiple tests, was assessed by the Benjamini–Hochberg method (False Discovery Rate) [34].

- For bioinformatics analysis of interactions between the C9orf16 gene and cigarette smoke components, the Comparative Toxicogenomics Database resource (CTD) available online, was used (http://ctdbase.org) (accessed on 16 May 2022) [35]. CTD provides the ability to analyze specific interactions between genes and chemicals in vertebrates and invertebrates based on data obtained from published scientific studies worldwide. This tool was used to analyze binary interactions involving one chemical and one gene/protein.

The Cerebrovascular Disease Knowledge Portal (CDKP) (https://cd.hugeamp.org/) was used for  bioinformatic analyze of the associations of rs2900262 C9orf16 with atherosclerosis-associated diseases, intermediate phenotypes, and risk factors for ischemic stroke (such as blood pressure, the level of triglycerides, the level of  low density lipoproteins, etc.) (accessed on 18 May 2022) [36]. 

Results and discussion. The allele frequencies in the study patients according to the data published in the 1000 Genomes Project, Phase 3 (http://www.ensembl.org (accessed on 30 May 2022)) were comparable with those reported in following European populations: British in England and Scotland (GBR), Iberian populations in Spain (IBS), Toscani in Italy (TSI) (Table 2).

The genotype frequencies of rs2900262 C9orf16 in study groups are presented in Table 3. The distribution of genotypes frequencies corresponded to the Hardy-Weinberg equilibrium both in the group of patients and in the control group (P>0.05).

Analysis of the general group of patients/controls did not show an association of rs2900262 C9orf16 with the risk of IS (Table 3).

Due to a possible sex dimorphism in the associations of genetic variants with the risk of IS, an analysis of the relationship of rs2900262 C9orf16 with the development of IS separately in males and females was carried out. It turned out that this genetic variant does not affect the risk of the disease in males. As shown in Table 4, SNP rs2900262 C9orf16 was found to be associated with a increased risk of IS in females. In particular, according to the dominant model, genotypes C/T-T/T was associated with an increased risk of IS (corOR=1.74, 95%CI 1.07–2.82, Padj=0.025, P^FDR=0.042). According to the additive model, the rs2900262 genotypes were also associated with an increased risk of IS (corOR=1.69, 95%CI 1.06–2.71, Padj=0.028, P^FDR=0.042).

The analysis of associations of rs2900262 C9orf16 with of IS risk depending on the smoking status revealed that this genetic variant is associated with an increased risk of IS exclusively in smoking individuals according to the dominant model (OR=1.92, 95%CI 1.09–3.37, Padj=0.019, P^FDR=0.042). Analysis of the log-additive model also revealed the association of rs2900262 C9ORF16 with the increased risk of IS in smokers (OR=1.79, 95%CI 1.04–3.08, Padj=0.028, P^FDR=0.042) (Table 5).

Analysis of the effects of rs2900262 C9orf16 on clinical manifestations of IS revealed that in carriers of genotypes C/T-T/T, ischemic stroke manifests on average 2 years earlier than in those of genotype C/C (Recessive model: Padj=0.017; Difference= -1.98, 95% CI=-3.61 - -0.36, Padj= 0.017, P^FDR=0.026). According to the additive model, rs2900262 C9ORF16 is also associated with a decrease in the age of manifestation of IS (Difference= -1.82, 95% CI=-3.38 - -0.26, Padj= 0.015, P^FDR=0.026) (Table 6).

The C9orf16 gene is expressed in brain tissues, blood vessels, and whole blood (Figure 2). In brain tissues, MeTPM varies from 50.20 to 172.2; in blood vessels - from 64.62 to 79.12; in whole blood MeTPM=50.32 (Fig. 2).

C9orf16 (also known as BBLN), a gene with unknown functions, has not been investigated for the risk of IS. However, in a previous study, a change in the expression of this gene in the blood of patients with Parkinson's Disease was found, which indicates a possible significant role of this gene in neuroprotection [37].

Due to the fact that the use of PubMed and Google Scholar search resources did not allow us to detect studies aimed at studying the associations of the studied Tag SNP rs2900262 C9orf16 with the development of IS, we used the bioinformatic resource Cardiovascular Disease Knowledge Portal to interpret the phenotypic effects of this genetic variant (CVDKP), (https://cvd.hugeamp.org), which combine and analyze the results of genetic associations of the largest consortia for the study of cardio- and cerebrovascular diseases, as well as their intermediate phenotypes. It was found that the protective C allele is associated with a decrease in BMI, Body fat percentage, Systolic blood pressure, Diastolic blood pressure, Mean arterial pressure, Pulse pressure, Creatine kinase. At the same time, this allele is associated with an increase in levels of HDL cholesterol и AF-SNP hypertension interaction (Table 7). Thus, this genetic variant may play an important role in the regulation of blood pressure, lipid metabolism, and may also influence the association of SNPs with the development of atrial fibrillation. Taking into account that atrial fibrillation acts as a significant risk factor for ischemic stroke of cardioembolic type, the influence of rs2900262 C9orf16 on the risk of IS may include a range of molecular mechanisms involved in various links in the pathogenesis of the disease.

Bioinformatic analysis was additionally performed to interpret the functional effects of rs2900262 C9orf16. The use of the GTEx Portal resource and the eQTL Calculator search query showed that the C allele (with a protective effect on the risk of IS development established in our study) affects an increase in gene expression in blood vessels. At the same time, the C allele is characterized by multidirectional effects relative to the effect on the level of expression in brain tissues: carriers of the C allele showed an increase in the expression of the C9orf16 gene in Brain - Caudate (basal ganglia) and Brain – Cerebellum and, conversely, a decrease in gene expression in Brain – Cortex и Brain - Frontal Cortex (BA9) (Table 8).

Resource eQTLGen Consortium (https://www.eqtlgen.org/cis-eqtls.html) showed that the risk allele T affects the decrease in expression of gene URM1 (Z-score = -10.6884, P =1.15×10^-26) and DNM1 (Z-score = -4.67, P = 2.9×10^-6), as well as an increase in gene expression SH3GLB2 (Z-score = 4.44, P = 8.8×10^-6) и RP11-395P17.3 (Z-score = 5.78, P = 7.41×10^-9) through cis-eQTL effects in blood cells. Gene URM1 acts as a ubiquitin-like protein (UBL) that is covalently conjugated via an isopeptide bond to lysine residues of target proteins such as MOCS3, ATPBD3, CTU2, USP15 and CAS. The thiocarboxylated form serves as substrate for conjugation and oxidative stress specifically induces the formation of UBL-protein conjugates (https://www.genecards.org/cgi-bin/carddisp.pl?gene=URM1&keywords=URM1). DNM1 gene encodes a member of the dynamin subfamily of GTP-binding proteins. The encoded protein possesses unique mechanochemical properties used to tubulate and sever membranes, and is involved in clathrin-mediated endocytosis and other vesicular trafficking processes. Actin and other cytoskeletal proteins act as binding partners for the encoded protein, which can also self-assemble leading to stimulation of GTPase activity (https://www.genecards.org/cgi-bin/carddisp.pl?gene=DNM1&keywords=DNM1). SH3GLB2 is a Protein Coding gene.  Enables identical protein binding activity (https://www.genecards.org/cgi-bin/carddisp.pl?gene=SH3GLB2&keywords=SH3GLB2).

            Subsequent analysis revealed a significant effect of rs2900262 C9orf16 on histone modifications. According to the HaploReg database (v4.1), this genetic variant is located in the region of DNA binding to histone H3, characterized by mono-methylation at the 4th lysine residue of the histone H3 protein (H3K4me1) and marking enhancers in all brain tissues represented by HaploReg (v4.1), peripheral blood cells, as well as tri-methylation at the 4th lysine residue of the histone H3 protein (H3K4me3) and marking promoters in the brain tissues and blood. The effect of this histone mark is enhanced by acetylation of the lysine residues at N-terminal position 27 of the histone H3 protein (H3K27ac), marking enhancers in peripheral blood cells and all brain tissues, as well as the acetylation at the 9th lysine residues of the histone H3 protein (H3K9ac), marking promoters in blood cells and all brain tissues, except Brain Hippocampus Middle (Table 9).

The analysis of  TFs using atSNP source revealed that TFs binding to the protective allele C, are involved in joint regulation of leukocyte adhesion to vascular endothelial cell (GO:1904994; Padj=0.02), regulation of vasculature development (GO:1901342; Padj=0.02), regulation of blood vessel endothelial cell migration (GO:0043535; Padj=0.02), positive regulation of defense response (GO:0031349; Padj=0.03), regulation of angiogenesis (GO:0045765; Padj=0.045), regulation of inflammatory response (GO:0050727; Padj=0.045) (Table 10).

It is noteworthy that the association of rs2900262 C9orf16 with the risk of IS was found in women and was not found in men. Previously, Smith et al. showed that C9orf16 expression may be negatively regulated by estradiol [38]. Thereby decreased expression of C9orf16 in T allele carriers may manifest female-specifically being more impeded in women.

            Similarly, in contrast to non-smokers, smoking patients were characterized by more strong association between mutant allele and the risk of ischemic stroke. This finding may have the same ground since some tobacco components apparently may decrease C9orf16 expression. Using bioinformatic resourse Comparative Toxicogenomics Database we revealed that tobacco smoke and its components are able to down-regulate expression of C9orf16 as well as up-regulate methylation of C9orf16 promoter (Table 11).

Considering that an increase in methylation leads to a decrease in gene expression, studies conducted in this direction provide additional evidence of the effect of smoking on a decrease in the expression of the C9orf16 gene and an increase in "risk" effects regarding the development of IS in carriers of variant genotypes C/C-C/T.            

Conclusion. To the best of our knowledge, this is the first study that showed the association of rs2900262 C9orf16 with the risk of IS, sex-specific effect in this association, and the trigger effect of smoking on the association of this genetic variant with the development of the disease. Bioinformatic analysis revealed a high regulatory potential of rs2900262 C9orf16, particular cis-eQTL-mediated effects on the regulation of gene expression in blood, blood vessels and brain, binding to transcription factors involved in biological processes with significant role in the pathogenesis of IS, as well as epigenetic regulation of this SNP. Future studies have yet to accumulate the data disclosing detailed networks of its probable implication in ischemic stroke risk and course.

Financial support

This research was funded by Russian Science Foundation (No. 22-15-00288, https://rscf.ru/en/project/22-15-00288/).