This case–control study evaluated blood samples obtained from femoral veins of unrelated Thai decedents within 24 h after death. All decedent subjects underwent postmortem examination under the jurisdiction of the Department of Forensic Medicine, Faculty of Medicine Siriraj Hospital, Mahidol University, which is located in Bangkok, Central Thailand. The study protocol was ethically approved by the Siriraj Institutional Review Board, Faculty of Medicine Siriraj Hospital, Mahidol University. The subjects were divided into SUNDS and control groups. SUNDS cases were unrelated subjects who died during sleep at night with previous healthy history and a negative comprehensive postmortem investigation including histological and toxicological examination that can explain the cause of death. Control cases were unrelated subjects who died of any cause of death excluding SUNDS. To minimize the effect of genetic backgrounds that could lead to false-positive or false-negative results (Hu and Ziv 2008), control subjects were divided into three subgroups: subjects who resided in regions other than Northeastern Thailand (DR group, 99 subjects), age-matched subjects who resided in Northeastern Thailand (NE group, 28 subjects), and subjects older than 40 years who resided in Northeastern Thailand (NE+ group, 24 subjects). All control subjects were healthy and aged 15 years or older with no history of cardiac arrhythmic diseases, radiotherapy, chemotherapy, bone marrow transplantation, blood transfusion within 2 years, other SCN5A-related conditions (e.g., certain cardiovascular diseases), and genetic diseases. Subjects who developed some degree of decomposition or who were embalmed were excluded from the study. Subject information, including age, race, medical history, family history, autopsy findings, and cause of death, was documented.
All blood samples were stored at 4 °C shortly after being drawn from the femoral veins of the decedents. Genomic DNA for each subject was extracted from blood samples within 72 h after collection and quantified using a Nanodrop® ND-1000 (Thermo Fisher Scientific, Inc., Waltham, MA, USA).
Genetic testing
Genetic screening of all coding regions of SCN5A was performed in all SUNDS cases using polymerase chain reaction (PCR) and direct Sanger sequencing. A total of 40 pairs of PCR primers spanning all 28 exons of SCN5A were designed and employed for genetic screening (primer sequences available on request). PCR reactions were performed in a PTC-100™ Programmable Thermal Controller (MJ Research, Inc., St. Bruno, Quebec, Canada). Sanger sequencing was performed in an ABI 3730XL DNA Analyzer (Applied Biosystems, Inc. Foster City, CA, USA), using BigDye® Terminator Cycle Sequencing Kit version 3.1 (Applied Biosystems). The same primers were used for direct sequencing and PCR. Sequence chromatograms were then analyzed using Sequencing Analysis 5.2 software (Applied Biosystems).
The obtained sequencing data were then compared with the corresponding referenced cDNA sequence of SCN5A retrieved from GenBank® (NG_008934.1) using Mutation Surveyor® (Demo) (SoftGenetics, PA, USA). To identify the association between SCN5A variants and SUNDS, variants were selected by focusing on SNP information for CDX (Chinese Dai in Xishuangbanna, China), CHS (Southern Han Chinese, China), KHV (Kinh in Ho Chi Minh City, Vietnam), and JPT (Japanese in Tokyo, Japan) populations, including allele frequency, linkage disequilibrium (LD), and single-nucleotide polymorphism (SNP) locations. SNPs with a minor allele frequency of approximately 5% from the Exome Aggregation Consortium (ExAC), 1000 Genomes Project (http://www.1000genomes.org/ensembl-browser), and NHLBI GO Exome Sequencing Project (GO-ESP, http://snp.gs.washington.edu/EVS) databases were preferable.
These identified variants were also screened in the controls using high-resolution melting (HRM) analysis. Genetic screening was performed using a CFX96™ Real-Time PCR Detection System (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The primers of each variant suitable for HRM were designed using Primer3 version 0.4.0 (http://bioinfo.ut.ee/primer3-0.4.0/) (primer sequences available on request). Melting files generated using Precision Melt Analysis™ Software (Bio-Rad Laboratories, Inc., Hercules, CA, USA) were then analyzed. To validate the variants obtained from additional HRM patterns, large amplicons containing additional patterns were subjected to further genetic screening by PCR and Sanger sequencing.
Genetic variant analysis
All identified variants were filtered for candidate mutations based on allele frequency in the 1000 Genomes Project database, functional mutation types, predictive software programs, and a review of the related published literature. To determine whether the identified genetic variants were novel mutations, the variants were also compared with published data, including Single Nucleotide Polymorphism Database (dbSNP) (http://www.ncbi.nlm.nih.gov/snp/), the 1000 Genomes Project database, the GO-ESP database, and the inherited arrhythmias database (http://www.fsm.it/cardmoc/). To predict the deleterious effects of identified variants, three web-based in silico prediction programs were used in this study, namely MutationTaster (http://www.mutationtaster.org/), PROVEAN (http://provean.jcvi.org/index.php), and Human Splicing Finder (http://www.umd.be/HSF3/HSF.html).
Any variant identified in both SUNDS cases and controls was designated as a polymorphism. Polymorphisms with a minor allele frequency (MAF) exceeding 0.01 among the controls were classified as common polymorphisms. If the MAF was less than 0.01 among the controls, the polymorphism was regarded as rare polymorphism. Any variants observed only in SUNDS cases were designated as mutations. All nonsense, frameshift, and splice site variants were considered to be candidate pathogenic mutations, unless identified as a polymorphism.
Statistical analysis
All statistical analyses (unless otherwise specified) were performed using PASW Statistics for Windows, Version 18 (SPSS, Inc., Chicago, IL, USA). Continuous variables are shown as the mean ± standard deviation. The allele frequencies for each SNP site and genotype frequency for each polymorphism were calculated using the web-based SNPStat program (http://bioinfo.iconcologia.net/snpstats/start.htm). The genotype frequencies of the detected coding region polymorphisms in the SUNDS and control groups were tested for deviation from Hardy–Weinberg equilibrium using Pearson’s chi-square goodness-of-fit test (de Finetti) and Fisher’s exact test (SNPStat), respectively. Differences in genotype and allele frequencies between the SUNDS and control groups were also analyzed using Pearson’s chi-square goodness-of-fit test (de Finetti) and Fisher’s exact test (SNPStat), respectively. The association between multiple detected variants and LD was calculated using online SNPStat. To identify the detected variants that were inherited together, SNPStat was used to construct and estimate the haplotype frequency. The association between haplotypes and SUNDS was analyzed via logistic regression. The association between the detected polymorphisms and SUNDS risk was estimated by calculating the odds ratio (OR) and 95% confidence interval (95% CI) using Pearson’s chi-square test analysis. p values < 0.05 were considered statistically significant.