The Ethical Committee of Soochow University has approved this study (approval number: ME81772029). Informed consent was obtained from all healthy blood donors and the victim’s family before the investigation.

A total of 524 controls and 158 SCD cases were included in the study, all of which were ethnic Han Chinese without any genetic relationship. The samples of SCD victims were collected from the following institutions during autopsy cases for medicolegal purpose: Soochow University and Institute of Forensic Sciences (China Ministry of Justice) representing Eastern China (20 cases), Sun Yat-sen University representing Southern China (103 cases) and Guizhou Medical University representing Southwestern China (35 cases). The blood samples were collected for the current study following the inclusive and exclusive criteria described previously (Wang et al., 2017; Zou et al., 2019). In brief, detailed forensic pathological examination was rigorously performed with each SCD case to confirm the coronary heart diseases as the main cause of SCD. No other fatal changes were detected, in addition to varying degrees of coronary atherosclerosis. Age-matched (±5 years) healthy controls were recruited from the same region during the same period as the SCD case, but those with any positive family history of SCD or cardiovascular disorders were excluded. During autopsies on traffic accident victims at the Forensic Medicine Expert Center of Soochow University, a total of 15 human heart tissues were obtained. The myocardium tissues were collected during forensic autopsies and were saved at −80°C until RNA and DNA extraction.

The GSE41571 and GSE23561 mRNAs profiles from Gene Expression Omnibus (GEO) database were used to analyze the correlation between NPC1 and metabolic syndrome and coronary artery disease (CAD) (Lee et al., 2013; Grayson et al., 2011). Information on all variants of NPC1 was retrieved from the variation module of the dbSNP database (https://ncbi.nlm.nih.gov/snp/, build 155, access date: June 2021). In sight of uncommon risk alleles with an evolutionary disadvantage, minor allele frequencies >0.15 was established for variant selection. Based on the Genotype-Tissue Expression (GTEx) data, expression quantitative trait loci (eQTL) analysis and Gene Expression Profiling Interactive Analysis (GEPIA) were used to assess the impact of the variants on target gene expression and to measure the expression correlations between genes, respectively (GTEx Consortium 2013). The 1000 Genomes Project data on 445 lymphoblastoid cell lines were used to validate the validity of the results of the analysis (Sudmant et al., 2015; Ferreira et al., 2016). Linkage disequilibrium (LD) analysis was analyzed using R package (Version 4.0.3) (Chan 2018). In addition, genotype data were obtained through the Ensembl Release 105 (December 2021) (http://asia.ensembl.org/Homo_sapiens/Info/Index).

Genomic DNA from blood samples and tissues was extracted using Blood DNA Kit or Blood Spots DNA Kit (TIANGEN, Beijing, China). All primers for the amplification were synthesized from Genewiz Company (Suzhou, China). Following the PCR reaction, the genotyping PCR products were detected by 7% non-denaturing PA-gel electrophoresis (PAGE) and visualized through the silver staining method. Genotyping process was performed in a double-blinded way (Zhu et al., 2012). After sample randomization, 20 DNA samples (10 from case and 10 from control) were gel purified and sequenced to confirm the results. In addition, two independent investigators verified approximately 10% (20 case samples and 190 control samples) of total DNA samples to confirm the consistency of genotyping. The following primers were used for genotyping: Forward: 5′-GTG​ATG​TGA​TCT​CTG​CTC​ATT-3′, Reverse: 5′-GGG​CAA​CAA​GTG​GAA​CA-3´.

Total RNAs of each group were extracted from myocardium tissue samples using TRIzol reagent (Invitrogen) and quantified by NanoDrop 2000c Spectrophotometer. The amplification was performed using the SYBR Green dye on the LightCycler real-time PCR system (Roche). The results were calculated using the 2^−ΔΔCT algorithm. To exclude potential bias across different conditions, GAPDH was used as the internal reference. The primer sequences were as follows: NPC1- Forward: 5′-GCA​CCT​TTT​ACC​ATC​ACT​CCT​G-3′, NPC1- Reverse: 5′-GGC​CAC​AGA​CAA​TAG​AGC​AGT-3´; GAPDH- Forward: 5′-CTC​TCT​GCT​CCT​CCT​GTT​CGA​C-3′, GAPDH- Reverse: 5′-TGA​GCG​ATG​TGG​CTC​GGC​T-3´.

Total proteins in human myocardial tissues were extracted with RIPA (Radio Immunoprecipitation Assay) lysates. Protein samples were subjected to 8% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), then transferred onto PVDF membranes. After transfer, the membranes were blocked with 5% BSA for 1 h and incubated overnight with the following primary antibodies: Anti-NPC1 (Cat# AB134113, 1:1,000 dilution, Abcam Antibodies); Anti-GAPDH (Cat#AB-M-M001, 1:1,000 dilution, Goodhere Biotechnology). After three TBST rinses, the PVDF membranes were incubated again for 1 h with secondary antibodies (1:1,000 dilution, Biyotime Biotechnology). Finally, the bands were detected with the ECL reagent (ChemiScope, Shanghai).

The 293T cell lines were cultured in a 37°C humidified chamber that was supplemented with 5% CO₂ and Dulbecco’s Dodified Eagle Medium (DMEM) containing 1% penicillin-streptomycin and 10% fetal bovine serum (FBS). The cell lines were characterized by Genetic Testing Biotechnology Corporation (Suzhou, China) using short tandem repeat (STR) markers.

Approximately 350-bp fragments containing rs150703258 located in intergenic regions close to NPC1 gene were synthesized and subcloned into KpnI and SacI sites of pGL3-control vector by Genewiz Company (Suzhou, China). Wild type (pGL3-WT) or mutant type (pGL3-MT) vectors contain the delete allele or insert allele, respectively. The direction and sequence of vectors were verified by direct sequencing.

In one well of a 24-well plate, 1×10⁵ cells were seeded for 24 h and then co-transfected with approximately 400 ng reconstructed vector and 40 ng pRL-SV40 vector (Promega) using jetPRIME^® transfection reagent (Polyplus-transfection^®). As controls, cells transfected with empty pGL3-control vector were used. Luciferase assays were performed at 24 h using a dual luciferase assay system and each group had four duplicate wells and the experiments were repeated at least three times.

JASPAR (http://jaspar.genereg.net/) was used for transcription factors binding site prediction (Castro-Mondragon et al., 2022). The UCSC genome browser is used for visualization the results. For each TF binding profile in the JASPAR CORE collection, genomes were scanned for matches. PWMScan was used to compute the relative scores and p-values (Ambrosini et al., 2018). Shaded boxes represent predicted binding sites for each of the TF profiles in the JASPAR CORE collection. The shading of the boxes indicates the p-value of the profile’s match to that position (scaled between 0 and 1,000 scores, where 0 corresponds to a p-value of 1 and 1,000 to a p-value ≤ 10⁻¹⁰).

The SPSS (IBM SPSS Statistics Subscription) and GraphPad Prism 8.4.2 were employed for statistical analysis. Depending on the normality of data, Student’s t-test was used to compare continuous data. The odds ratio (OR) and 95% confidence interval (95% CI) were estimated using binary logistic regression to evaluate the association between rs150703258 and SCD risk. The χ² test was used in Hardy-Weinberg equilibrium testing. All tests were two-sided and p-values < 0.05 were considered statistically significant.

The workflow of this study was presented in Figure 1. First, the expression level of NPC1 in patients with metabolic syndrome and CAD seems to be higher than that of healthy subjects and a similar trend can be seen from the comparison between the two groups of ruptured plaque and stable plaque (Figure 2A), indicating up-regulation of the NPC1 may lead to an increased risk of cardiac disease. As shown in Supplementary Table S1, we identified 27 variants with a MAF >0.15 within 3′ UTR of NPC1 and the range of 10-kb downstream region with available frequency. Then, we selected five mutation sites based on eQTL and 1000G data analysis. The four variants were SNPs (rs12966114, rs4800487, rs303761 and rs1367085) and the other one was indel polymorphism (rs150703258). To follow the well-established analysis protocols based on capillary gel electrophoresis combined with PCR-based assays in forensic genetics, indels and other length polymorphisms were given priority in marker selection. Thus, the indel variant rs150703258 was selected as the candidate variant for association investigation. As shown in Figure 2B, rs150703258 polymorphism showed a significant genotype--phenotype association with NPC1. Additional LD analysis indicated that rs150703258 and rs167336, which was associated with waist circumference in GWAS analysis, were in the same LD block (Figure 2C).

Sample demographic characteristics were given in Table 1. The characteristics of the SCD cases were provided in Supplementary Table S2. Their average age of current SCD cases was around 51-years old and there was a male predilection. Coronary atherosclerosis with various degrees was found in all the SCD cases. Three alleles were found by sequencing in the current cohort. The mutation genotyping results for rs150703258 were shown in Figure 3. A total of five rare 12-bp allele (4-bp longer than the insertion allele) of rs150703258 was discovered in the current cohort. The rare allele ([GTAT]₃) frequency was calculated to be 0.36%. The observed genotype distributions in the control group have no deviation to the Hardy-Weinberg equilibrium (p > 0.05).

Genotypic and allelic frequencies of rs150703258 and association analysis were shown in Table 2. Compared with the ins/ins genotype, an increased SCD risk was observed for individuals carrying ins/del genotype (adjusted OR: 1.663, 95% CI: 1.11–2.50, p = 1.35 × 10 − 2 ). Similar trend was also found in the dominant model (adjusted OR: 1.659, 95% CI: 1.13–2.45, p = 1.01 × 10 − 2 ). The 8-bp deletion allele could contribute to a 32.9% increased risk of SCD (OR: 1.329, 95% CI: 1.03–1.72, p = 2.89 × 10 − 2 ). Together, our results indicated that deletion allele of rs150703258 was associated with SCD susceptibility in Chinese populations. In addition, as shown in Supplementary Figure S1, body mass index (BMI) analysis revealed that the genotype of rs150703258 was also associated with BMI (mean BMI = 23.34, 24.55 and 25.07 for ins/ins, ins/del and del/del genotype, respectively).

The GTEx database was used for eQTL analysis to identify target genes potentially regulated by rs150703258. Our eQTL analysis observed that the rs150703258 variant is highly related to NPC1 and C18orf8 expression levels. In human arteries and heart tissues, both genes have been shown to be critical eQTL genes (Figure 4A).

To further investigate the genotype-phenotype correlations between rs150703258 and NPC1 gene, NPC1 expression at mRNA level and protein level was validated by quantitative real-time PCR and western blot in human myocardium tissue samples, respectively. The relative mRNA expression level of NPC1 in samples with ins/del and del/del genotype was significantly higher than those with ins/ins genotype (Figure 4B) with a fold change = 2.73, p < 0.0001. At protein level, the same trend was also observed in western blot analysis (Figure 4C). The results showed that the deletion allele genotype of rs150703258 was significantly associated with increased gene expression, indicating its potential regulatory role.

Based on Gene Expression Profiling Interactive Analysis, the expression of NPC1 and C18orf8 presents a significant correlation in human blood, arteries and heart tissues (Figure 5). The similar expression patterns of NPC1 and C18orf8 indicate that the two genes may be co-regulated (R = 0.82 in whole blood).

Using pGL3-control vector containing insertion or deletion allele of rs150703258 variant, we performed luciferase reporter assays to confirm whether the indel variant affects transcription activity. As shown in Figure 6, there was a significant difference in luciferase activities between the groups transfected with vectors carrying insertion allele (pGL3-MT) and those carrying deletion allele (pGL3-WT). Luciferase activity in the pGL3-WT transfection group was higher than that in the pGL3-MT transfection group, indicating that rs150703258 showed regulatory properties on gene transcription in this study. Additionally, we surmised that the deletion allele of rs150703258 may regulate gene expression by recruiting transcription factors. In silico analysis of this region showed that the rs150703258 deletion variant may create transcription factor binding sites (Figure 7).