Association Between CACNA1C Polymorphisms and Insomnia in a Chinese Han Population: A Case-Control Population-Based Analysis

A limited number of genome-wide association studies (GWAS) have shown that CACNA1C is a potential candidate gene for insomnia. The present study sought to investigate the association of CACNA1C gene polymorphisms and insomnia in a Chinese Han population. Twenty-one single nucleotide polymorphisms (SNPs) in the CACNA1C were genotyped by using matrix-assisted laser desorption ionization time-of-ight mass spectrometry (MALDI-TOF MS) with a Sequenom MassARRAY system in 205 patients with insomnia and 154 healthy controls. Data from all participants were systematically collected. Association of the polymorphisms with insomnia was evaluated by statistical analysis. Linkage disequilibrium (LD) and haplotype analysis were performed with Haploview v4.2 software. After adjustment for multiple comparisons and gender, age, body mass index (BMI), hypertension and glycated hemoglobin (HbA1c), only the rs2302729 in the recessive model (adjusted odds ratio [AOR]= 0.414, 95%CI= 0.220-0.771, P= 0.004) related to insomnia achieved signicance by the false discovery rate (FDR) Benjamini and Hochberg (BH) criterion. The haplotypes rs2302729-rs1051375, CA and TG were signicantly associated with insomnia (P<0.05). Our ndings contributed important evidence for the conrming of CACNA1C as a susceptibility gene for insomnia in the samples of Chinese Han population.


Introduction
Insomnia is one of the most common sleep disorders encountered in the population, frequently characterized by the subjective complaint of di culty falling or maintaining sleep, or non-restorative sleep, producing signi cant daytime symptoms, including di culty concentrating and mood disturbances. Studies from Western populations show that the incidence of insomnia was 6.8% to 19.2% (Jansson-Fröjmark and Linton 2008; Pallesen et al. 2014;Ford et al. 2015). While a Chinese crosssectional study recruited 2982 aged participants indicated that the prevalence of occasional insomnia and frequent insomnia was up to 17.07% and 12.17% respectively (Zhuang et al. 2016). These studies indicated that insomnia is a widespread public health problem globally, especially in China. Insomnia is associated with several negative sequela, including fatigue, impaired memory, inattention, irritability, and depression. Epidemiological and longitudinal studies have also shown that insomnia is a risk factor for the development of a variety of health problems, such as new-onset mood, anxiety and depression, impaired immune function, diabetes, obesity and cognitive impairment, and death related to cardiovascular disease (Bollu and Kaur 2019; Fernandez-Mendoza and Vgontzas 2013).
Despite insomnia being widespread and severely affected, the exact cause of insomnia is still largely unknown, and a better understanding of its pathogenic mechanism is essential to improve the quality of life and clinical outcomes of insomnia. Previous studies also have shown that non-genetic factors such as older age, female, depression, poor perceived health, pain, obesity, and socioeconomic status may increase the risk of insomnia (Hartz et al. 2007). Most notably, several aspects of sleep behavior such as timing, duration and quality have been demonstrated to be heritable (Byrne et al. 2013). The identi cation of the potential susceptibility genes and gene loci, and the exploration of the molecular genetics of insomnia are required to facilitate the early diagnosis and prevention of it, and personalized treatment.
However, the contribution of genetics to its etiology and pathophysiology remains largely unknown, while initial studies point in a few directions. Several genome wide association studies (GWAS) have been conducted on sleep phenotypes in humans (Byrne et (Semenova et al. 2017). Additionally, a twin modeling study (Lind et al. 2015) also indicated insomnia symptoms were found to be modestly heritable.
Nevertheless, many of these analyses had small sample sizes, or based on certain ethnic, and even many associations were not replicated in other independent samples (Byrne et al. 2013;Parsons et al. 2013).
The voltage-dependent calcium channel, L-type, alpha 1C subunit (CACNA1C) gene encodes CACNA1c, whose function is to mediate the in ux of calcium ions into the cell upon membrane depolarization, and is mainly involved in neurotransmitter release, neuronal excitability (Xu et al. 2015), and gene regulation (Bhat et al. 2012 Hammerschlag et al. 2017). In addition, GWAS also has many limitations (Emily 2018): First, tag-SNPs, only a selection of SNPs, were genotyped by SNPs array but not all SNPs. Second, GWAS suffer from a multiple testing issue which might be different from one statistical test to another due to the large number of SNPs tested when considering signi cance at the genome level. Finally, covariates which can play an important role in the analysis of disease susceptibility, while GWAS didn't taken it into consideration.
Given the susceptibilities of CACNA1C polymorphisms in insomnia is still unknown in Chinese population who had a high incidence of insomnia, we undertook an evaluation of 21 candidate SNPs in CACNA1C within an independent sample of Chinese Han population in this report.

Study Population and DNA Extraction
A case-control study included 205 patients with chronic insomnia and 154 age-matched healthy volunteers from the Department of Neurology at the Second A liated Hospital, University of South China (Hengyang, China) between Jul 2016 and Dec 2018. All subjects were unrelated Han Chinese from Hunan province. The diagnosis of chronic insomnia is based on the patient's subjective complaints about poor sleep quality, di culty starting or maintaining sleep, or early to wake, along with reports of severe distress or daytime consequences, and duration of at least 3 or more nights per week for three months or longer, according to the diagnostic criteria from the International Classi cation of Sleep Disorder, 3rd (ICSD-3) (American Academy of Sleep Medicine 2014). It is well-established that chronic insomnia disorder in ICSD-3 can be combined with many other medical conditions, such as the dependence of these comorbidities on pain, mental disorders, life events, and pregnancy (Bollu and Kaur 2019; Fernandez-Mendoza and Vgontzas 2013; Nacar and Taşhan 2019). As these conditions were not appropriate for insomnia genetic studies. Therefore, subjects with these conditions should exclude from the study. Speci cally, the exclusion criteria are: 1) patients with a standard total Self-rating Depression Scale (SDS) score of >62 points, or a standard total Self-rating Anxiety Scale (SAS) score of >59 points; 2) subjects with pregnancy, chronic pain, malignant tumor, dementia, stroke, moderate to severe obstructive sleep apnea hypopnea syndrome, chronic obstructive pulmonary disease, congestive heart failure, asthma, cirrhosis, renal impairment, and moderate to severe mental disorders; 3) subjects with a history of major emotional trauma six months before sampling, or a history of taking corticosteroids or psychotropic drugs.
All patients, whose chief complaint was poor sleeping quality, received the nal diagnosis by the patient's clinical manifestations, clinical history, Pittsburgh Sleep Quality Index (PSQI), SAS, SDS and Polysomnography (PSG) results. The scoring used for PSQI, SDS, and SAS was conducted according to reference (Zung 1965; Zung 1971; Tsai et al. 2005). As described in the study (Tsai et al. 2005), a total PSQI score >7 points indicates sleep disturbance, and the higher scores indicate more serious the condition.
The study protocol was approved by the Human Ethics Committees Review Board at the University of South China (no. 201602001), and signed informed consent was obtained from all participants. Data from all participants were systematically collected, including gender, age, body weight index (BMI), glycated hemoglobin (HbA1c), and blood pressure. History of insomnia were collected from the patients.
Venous blood samples were collected from all participants. Every genomic DNA was extracted from the blood samples using TIANamp Genomic DNA Kit (TIANGEN BIOTECH, Beijing, China). The DNA concentration was determined with the NanoDrop 2000 (Thermo Fisher Scienti c, MA, USA), and this was followed by dilution to a nal concentration of 500ng/uL. Diluted extracts and extracted serum samples were stored at -20°C until further analysis.

Candidate SNPs Selection and Primer Designing
Candidate SNPs in the CACNA1C were selected from several large genome-wide association studies (GWAS) (Byrne et al. 2013;Parsons et al. 2013) and the NCBI database (http://www.ncbi.nlm.nih.gov/projects/SNP/), and each SNP must have a minimum allele frequency (MAF) of greater than 0.05 in the Chinese Han population. A total of 25 tagSNPs were initially selected, such as rs2051990, rs7301906, rs7304986, rs7316184, rs16929275, rs16929276 and rs16929278 in intron 3; rs2302729 in intron 9; rs215976, rs216008, rs1051375 and rs1544514 in exon; rs2041135, rs2470433, rs4765975, rs7302540, rs7316246, rs7957163, rs10466907, rs10713809, rs11062319, rs11353034, rs11429670, rs12809807 and rs35440646 located in 3'UTR. Because the regions of homologous or near-homologous sequence in the genome can compete with the target sequence of the genetic locus for primer hybridization that leads to PCR failure, rs16929278 in intron 3, and rs10713809, rs11429670 and rs35440646 in 3'UTR were not tested. Finally, twenty-one SNPs were chosen as candidates for the following experiments. Primers for detection potential polymorphisms were designed with MassARRAY Assay Designer v3.1 (Sequenom, CA, USA). Pairs of PCR primers and single base extension (SBE) primers were provided for each SNP, and listed in Tables 1 and 2. Genotype Genotyping of each SNP locus was performed using a Sequenom MassARRAY system using matrixassisted laser desorption ionization time of ight mass spectrometry (MALDI-TOF MS) (Sequenom, CA, USA) at BGI (Beijing, China). All experimental operations were carried out in accordance with the manufacturers' instructions. Each total 5µL PCR reaction contained 0.625µL 1.25×PCR Buffer, 1.000 µL (500nM each) PCR primer mix (synthesized by Thermo Fisher Scienti c, MA, USA), 0.325µL 25mM MgCl 2 , 0.1 µL 25 mM dNTP mix, 0.1000 µL HotStar Taq (5U/µL), and 1 µL of DNA sample (20ng), and 1.850 µL HPLC water. Primary PCR was performed on the GeneAMP PCR system 9700 (Thermo Fisher Scienti c, MA, USA) with the following conditions: at 94°C for 5 min, 45 cycles of 94°C for 20 s, 56°C for 30 s, and 72°C for 60 s, and a nal extension at 72°C for 3 min. Then, non-incorporated dNTPs in the PCR products were dephosphorylated by adding a cocktail of 1.53 µL HPLC water, 0.17 µL Shrimp Alkaline Phosphatase (SAP) Buffer and 0.3µL SAP enzyme (1U/µL) (Sequenom, CA, USA) incubated on the GeneAMP PCR system 9700 at 37°C for 20 min, and then followed by 85°C for 5 min. SBE reaction contained 7 µL of SAP-treated PCR products and 2µL iPLEX mix (Sequenom, CA, USA). The iPLEX mix contained 0.2000µL iPLEX buffer, 0.2000µL iPLEX Termination mix, 0.940µL extend primer mix(0.625µM:1.25µM), 0.041µL iPLEX enzyme and 0.619µL HPLC water. The SBE reaction was performed with the following conditions: 94 °C for 30 s, followed by 40 cycles of one step at 94°C for 5 s with ve subcycles of 52 °C for 5 s and 80 °C for 5s, and followed by 72°C for 3 min. A total of 16 µL molecular HPLC water and each extension product were added to CLEAN Resin 384/6 MG Dimple plate (Sequenom, CA, USA) for further puri cation. Samples were rotated for 30 min on a tube rotator. After treatment, approximately 10 nL of products were spotted onto a 384-well SpectroCHIP bioarray (Sequenom, CA, USA), with the MassARRAY Nanodispenser RS1000 (Sequenom, CA, USA). The time of ight of these ionized products depends on the mass of each allele that will be measured by the mass spectrometer. The results were visualized on the MassArray Typer Analyzer v4.0 (Sequenom, CA, USA).

Statistical analysis
Clinical data and gene frequencies were analyzed using the t-test, Mann-Whitney U test and the χ 2 test with SPSS 22.0 (IBM, NY, USA). Values were presented as mean ± SD, median (M) and range, or the number and percentage (%) based on the type of data. We calculated original genotype frequencies for each of the 21 polymorphisms in insomnia and control groups, and assessed Hardy-Weinberg equilibrium for each SNP locus among controls. Considering the imbalance of the possible clinical insomnia risk factors between the case group and the control group, and the errors in multiple comparisons, the SNPs which were signi cant differences in the distributions (P-value <0.05 in the initial results) of genotype frequencies between insomnia and control participants were selected for further analysis. To identify the effect of genetic models on insomnia, analyses were performed using dominant (wild-type homozygote versus heterozygote and mutant-homozygote), recessive (wild-type homozygote and heterozygote versus mutant-homozygote), and additive (wild-type homozygote versus mutant-homozygote) modes for each polymorphism (Zhu et al. 2019). We estimated the genetic odds ratio (OR) using logistic regression analyses to compute relative risks and 95% con dence intervals (CI). To address the use of a pooled control group, genetic analyses were adjusted for the matching factors of gender, age, BMI, hypertension, and HbA1c. We then perform false discovery rate (FDR) Benjamini and Hochberg procedure (BH) (Benjamini et al. 2001), a type of signi cance-level adjustment for controlling the error in multiple comparisons, to account for the increased probability of reporting false positive results through multiple tests. The procedure for controlling the FDR at level 0.05, and FDR (BH) thresholds for each of the 21 individual SNPs was calculated as described (Benjamini et al. 2001). Linkage disequilibrium (LD) and haplotype analysis were performed with Haploview v4.2 software, and a detailed description of the haplotype analysis is given elsewhere (Gabriel et al. 2002). A difference of P <0.05 (two-sided) indicates signi cant.

Clinical data
The clinical characteristics of the subjects in the study are shown in Table 3. There were signi cant differences between the two groups in the systolic blood pressure (SBP) and HbA1c levels between the two groups (P <0.05). However, there was no signi cant difference in the age, gender, BMI, and diastolic blood pressure (DBP) between the case and control groups (P >0.05). The results of PSQI evaluation showed that all insomnia patients with a total score >7 points, and the overall PSQI average score was (14.23 ± 2.85), which suggests the cases had severe insomnia disorder. The measurement results of the SAS and SDS scale showed that none of the 205 patients had moderate to severe depression or anxiety. Furthermore, the PSG records revealed no evidence of moderate to severe obstructive sleep apnea hypopnea syndrome, or other sleep disorder. Table 4 lists the 21 polymorphisms examined and the observed genotype frequencies at each gene locus among the subjects. All the SNP markers were accorded with the Hardy-Weinberg equilibrium (HWE) (P>0.05). As shown, there were signi cant differences in the distributions of genotype frequencies only in three loci, rs1051375 and rs216008 in exon, and rs2302729 in intron 9, between the insomnia group and the control group (P<0.05). These three loci were selected for further analysis.

Genetic Analysis
Genetic contributions of the SNP with alleles to insomnia were analyzed by using dominant, recessive, and additive models' assessment. Table 5 presents three selected SNPs which, in any model of inheritance, were found in the initial analysis. As shown, following our multivariate logistic regressions and risk factor-adjusted analyses, only the recessive model and additive model of the rs216008, rs2302729 and rs1051375 polymorphism were found to have a risk-adjusted P-value less than 0.05 for association with insomnia. However, in analyses using the multiple tests with the FDR (BH) criterion, only the rs2302729 signi cant in the FDR analysis remained in the recessive model (AOR =0.416, 95%CI =0.220-0.781, P =0.004). In comparison with the CC+CT genotype at rs2302729, the TT genotype decreased the risk of the disease by 0.416 times, suggesting that the major C allele may be a risk factor for insomnia.

Haplotype association and Linkage disequilibrium
Haplotype structures of the CACNA1C gene were analyzed. Haplotype association analysis suggested twenty-one SNPs were identi ed that detected two LD blocks, among which four haplotypes were identi ed, as shown in Figure 1 and Table 6. The results showed that block 1 consisted of two closely linked SNPs (rs2302729 and rs1051375) with |D'| value is 0.83, and block 2 included ve linked SNPs (rs7316264, rs12809807, rs10466907, rs4765975, and rs7957163) with |D'| value is 0.81. In particular, the block 1, rs2302729-rs1051375, CA and TG haplotypes were associated with the risk of insomnia in the Chinese Han population (P<0.05), while the rs7316264-rs12809807-rs10466907-rs4765975-rs7957163 haplotypes were not associated with insomnia in this population (P>0.05, Table 6).

Discussion
In this case-control study, of the 21 SNPs analyzed here, we found only the recessive model of the rs2302729 in CACNA1C gene to be strongly associated with insomnia in our Chinese Hunan Han population, after adjustment for risk factor-adjusted analyses and multiple comparisons. We also observed a signi cant association between rs2302729-rs1051375, CA and TG haplotypes with insomnia.
The CACNA1C gene, located on chromosome 12p13.3, encodes an α-1 subunit of the L-type voltagedependent gated calcium channel (CACNA1c), which consists of 24 transmembrane segments, forming the pore mediate calcium ions (Ca 2+ ), a second messenger in the phosphatidylinositol system, pass into the cell. The channel transiently increases the membrane permeability for calcium-mediated cell membrane depolarization, playing an essential role in neurotransmitter release, neuronal excitability, and gene regulation (Mortensen 2013;Nakao et al. 2015). Sleep-wake cycle is regulated by several neurotransmitters, including calcium ions, and neuronal excitability (Xu et al. 2015). It is worth noting that the CACNA1c may be involved in mental illnesses, such as bipolar disorder and major depressive disorder (Kabir et al. 2017), and insomnia is the most common manifestation of those mental illness. The effect of CACNA1c on insomnia may overlap with bipolar disorder. Moreover, treating sleep regulation center with voltage-gated L-type Ca 2+ channel (LTCC) blocker diltiazem can signi cantly increase pentobarbital-induced total sleep (TS), non-rapid eye movement (NREM) sleep and slow wave sleep (SWS) ratio in NREM sleep (Cui et al. 2011). Therefore, CACNA1c may participate in insomnia by regulating calcium in ux, altering nerve excitability and neurotransmitter release. In recent years, as a candidate gene contributing to essential insomnia, CACNA1C has been highlighted in corresponding studies, despite the lack of replication (Byrne et al. 2013;Parsons et al. 2013;Kantojärvi et al. 2017). From a clinical and statistical research perspective, our data provide preliminary evidence for the establishment of CACNA1C as a susceptibility gene for insomnia.
In this study, only the recessive model of rs2302729 was found to have a risk-adjusted P-value less than FDR (BH) thresholds (Table 5). The rs2302729 located in intron 9 of CACNA1C. Current studies suggest that introns can alter not only transcript levels by affecting the rate of transcription, transcript stability, and nuclear export, but also the e ciency of precursor messenger RNA (pre-mRNA) splicing and mRNA translation (Shaul 2017). Sequence variants, such as SNPs, may lead to disorders in gene transcription, splicing and translation, resulting in abnormalities in gene expression and function (Caminsky et al. 2014). As reported, different splicing products of voltage-controlled calcium channels could lead to different functional effects, which can affect the in ux of extracellular calcium ions and the activation of calcium signaling (Kim et al. 2005), and the subsequent activation of mitogen-activated protein kinase (MAPK) signal in the suprachiasmatic nuclei, and then stimulate potential downstream transcription factors to regulate the expression of biological clock genes (Obrietan et al. 1998) leading to insomnia. Several variations in CACNA1C introns have been reported to predispose carriers to the development of insomnia (Byrne et al. 2013;Parsons et al. 2013;Kantojärvi et al. 2017). Though current research has failed to replicate the same SNP results, we successfully found an association of rs2302729 in intron 9 in recessive model with insomnia, and found that the rs2302729 T allele had a decreased risk of insomnia. It is unclear whether and how this mutation alters the expression of CACNA1C. It's worth noting that, by analyzing the nucleotide sequence, we found that the C allele at the rs2302729 locus could create a CpG site, which may serve as methylation site and regulate the gene expression, just like the rs1051375 locus (Kim et al. 2016). Though the recessive model and additive model of the rs216008 and rs1051375 polymorphism were found to have a risk-adjusted P-value less than 0.05 for association with insomnia in the present study, the associations were no longer signi cant by using the multiple tests with the FDR (BH) criterion (Table  5). This suggested that the rs216008 and rs1051375 polymorphisms may still be potential SNPs for insomnia. In fact, the alterations in the rs1051375 and rs216008 SNPs don't directly affect the primary structure of the CACNA1c. Nevertheless, previous study showed that genetic associations at rs1051375 has been observed in the Korean population of patients with bipolar disorder, and the G allele, the minor allele in the rs1051375 SNP, creates a CpG site containing a cytosine nucleotide adjacent to a guanine nucleotide (Kim et al. 2016). Most importantly, 70-80% of cytosines in the mammalian genome CpG sites are methylated (Jabbari and Bernardi, 2004), and DNA methylation pattern of CACNA1C is speci c in neuronal cells (Nishioka et al. 2013). It is worth mentioning that the bipolar disorder is characterized by circadian rhythm disturbances. In the current study, we excluded patients with moderate to severe mental disorders, though it is unclear whether insomnia patients with rs1051375 are predisposed to anxiety and depression. Interestingly, we nd that the rs2302729 and rs1051375 constituted a closely linkage disequilibrium block (Figure 1 and Table 6), and the rs2302729-rs1051375, CA and TG haplotypes were also signi cantly associated with insomnia in our population (P<0.05). Given that the recessive model of rs2302729 polymorphism is highly associated with insomnia, it was shown here that the polymorphism in rs1051375 was related to the insomnia before adjustment using FDR correction, probably due to the two SNPs were in signi cantly linkage disequilibrium. While another study of interindividual variability in mRNA expression and 12 splicing loci of CACNA1C in 65 heart tissue samples indicate that the mRNA ratio of CACNA1C for the rs216008 C and T alleles did not differ signi cantly from that of DNA ratio in all samples tested (Wang et al. 2006). However, the effect of rs216008 and rs1051375 on the expression of CACNA1C is still unclear. Therefore, additional research is needed to determine whether and how the mutations alter the expression and function of the CACNA1C gene in insomnia.
Our study also has some limitations. First, compared to the GWAS study, the sample size in our report was relatively small. Second, For the analysis of the genetic factors of insomnia, cases with many comorbidities on pain, mental disorders, life events, and many other confounding conditions were excluded from the study, and the selected polymorphisms were then entered into the risk factor-adjusted analysis which further controlled for gender, age, BMI, hypertension, and diabetes. While there were relatively few risk factors included in the logistic regression analysis. At last, since our controls, who were healthy individuals, had no complaints of sleep disorders and emotion disorders, they had not undergone SAS, SDS, and PSG assessments, and the comparison of these indexes between the insomnia group and the controls was not performed.

Conclusions
In summary, this study analyzed the relationship between twenty-one candidate polymorphisms in the CACNA1C gene and insomnia. Our results indicate that rs2302729 polymorphism in CACNA1C gene is signi cantly associated with insomnia, and CACNA1C may be a susceptibility gene for insomnia in the samples of Chinese Han population. Further well-designed studies with a larger population sample are still warranted to validate our ndings, and additional investigations also are necessary to de ne the mechanism underlying this relationship.