Introduction

Proprotein convertase subtilisin/kexin type 9 (PCSK9) is a serine protease that plays a key role in the regulation of low-density lipoprotein cholesterol (LDL-C) by promoting the degradation of low-density lipoprotein receptors (LDLR) through a post-transcriptional mechanism [1–4]. The PCSK9 binds directly to the EGF-A domain of LDLR and targets the receptor to the lysosome for degradation, leading to reduced clearance of LDL-C from circulation [5, 6]. Gain-of-function (GOF) mutations in the PCSK9 gene cause hypercholesterolemia and cardiovascular disease (CVD) [7–10], whereas loss-of-function (LOF) mutations are associated with reduced plasma levels of LDL-C and protection from CVD [11, 12].

PCSK9 is a highly polymorphic gene, with over 40 single nucleotide polymorphisms (SNPs) reported [13]. The relationship of common PCSK9 polymorphisms, e.g., R46L, I474V, and E670G, with serum lipid levels has been investigated. Nevertheless, the association between PCSK9 polymorphisms and lipid levels varies across populations. Several studies in European, African American, and white populations, hypobetalipoproteinemia subjects, as well as in the Dallas Heart Study (DHS) and the PROSPER study group have shown that PCSK9 46L carriers had decreased LDL-C levels [14–22]. Additionally, Caucasian, African Canadian [23], and Japanese [24] individuals carrying the V allele of PCSK9 I474V had lower total cholesterol (TC) and/or LDL-C levels than non-carriers. However, PCSK9 I474V was not associated with lipid levels in healthy UK men [15], Brazilian subjects [26], and the PLIC study [25]. Moreover, in German [27], Brazilian [26], and Tunisian populations [28], as well as in the Belgium Stroke Study [29], the PLIC study [25], and the LCAS [30], PCSK9 E670G was associated with increased TC and/or LDL-C levels or an increased risk of atherosclerosis. In healthy UK men [15], Taiwanese [31], Guangxi Bai Ku Yao, and Han Chinese populations [32], and in the PROSPER study group [14], PCSK9 E670G was not associated with increased TC and/or LDL-C levels or increased risk of atherosclerosis.

Apolipoprotein E (ApoE) is a component of plasma chylomicrons, chylomicron remnants, VLDL, IDL, and HDL. ApoE acts as a ligand for LDLR and LDL-related protein (LRP) [33]. There are three common alleles (E2, E3, and E4) in the APOE gene, which code for six genotypes of E2E2, E3E3, E4E4, E2E3, E2E4, and E3E4 [33]. The E4 allele is associated with higher levels of LDL-C and total cholesterol, and E2 with lower levels compared with the E3 allele [34]. In addition, E4 allele is associated with the lower levels of HDL-C [35]. It has been reported that the E4 allele is associated with increased risk of CVD [34, 35]. Moreover, the relationship between PCSK9 and APOE polymorphisms on lipid levels and intima media thickness had been reported in a previous study [25].

The effect of PCSK9 and APOE polymorphisms on lipid levels in a Southern Thai population has not been examined. Thus, the present study aimed to investigate the effects of these gene polymorphisms on lipid levels in Southern Thai subjects.

Materials and Methods

Study Population and Data Collection

The study was conducted on 495 unrelated healthy volunteers of a Southern Thai population from urban and rural areas in Pattalung and Nakhon Si Thammarat provinces. Pregnant women, individuals with diabetes mellitus and secondary dyslipidemia due to renal, liver or thyroid disease, and those individuals taking medication to lower lipid levels were not included in the study. Waist circumference (WC) was measured midway between the lower rib margin and the iliac crest. Body mass index (BMI) was calculated as a ratio of weight in kg and square height in meters (kg/m2). Blood pressure was measured, and data on lifestyle and sociodemographic factors were collected using a questionnaire. The questionnaire included education, occupation, previous diseases, current medication use, cigarette smoking habits, alcohol consumption, and physical activity. Written informed consent was obtained from every subjects. The study protocol was approved by the ethics committee of Walailak University.

Laboratory Examinations

Blood samples were collected from subjects after a 12 h fast. Serum and plasma were separated by centrifugation at 3000 rpm for 10 min and were later used for the analysis of biochemical parameters. Blood was collected in EDTA for DNA extraction. Fasting blood sugar (FBS) and lipid profiles were measured using enzymatic kits. HDL-C was determined after the precipitation of the apolipoprotein (apo) B containing lipoproteins by phosphotungstate/magnesium chloride. LDL-C was calculated using the Friedewald formula.

DNA Extraction

Genomic DNA was extracted from the buffy coat using a GeneAid kit (GeneAid Biotech Ltd., Taiwan). The quality and quantity of DNA were estimated using a NanoDrop 2000 UV–Vis spectrophotometer (Thermo Fisher Scientific Inc., USA). DNA samples were stored at −80 °C until analysis.

Polymorphism Detection

Genotyping for PCSK9 and APOE polymorphisms was conducted using the polymerase chain reaction and restriction fragment length polymorphism (PCR–RFLP) method with slight modifications of the previously described protocol [15, 27, 36]. Each PCR reaction contained 100 ng of genomic DNA, 0.2 µmol/l of each primer (eurofins MWG GmbH, Germany), 200 µmol/l dNTPs, 0.625 U Taq polymerase, and PCR buffer [50 mmol/l KCl, 1.5 mmol/l MgCl2, 10 mmol/l Tris–HCl (pH 8.3)] [New England BioLabs (NEB), USA], for all polymorphisms, except PCSK9 R46L and APOE, where 10 % (v/v) DMSO was added. The final reaction volume was 25 μL. Enzymes, primers, and PCR conditions used in this study are shown in Table S1. The PCR reactions were performed using GeneAmp PCR system 9700 thermal cycler (Applied Biosystems, USA). PCR products from PCSK9 R46L, I474V, E670G, and APOE assays were digested with RsaI (NEB, USA), BccI (NEB, USA), Sau96I (NEB, USA), and AflIII and HaeII (NEB, USA), respectively, at 37 °C for overnight except I474V, which was incubated at 37 °C for 1 h. Restriction fragments were identified by 2 and 4 % (w/v) agarose gel electrophoresis for R46L, E670G, and for I474V, APOE, respectively. The PCR products and restriction fragments were stained with ethidium bromide and visualized under UV light.

Statistical Analyses

Continuous variables were expressed as the mean ± standard deviation. Categorical variables were presented as percentages. The observed genotype frequencies were compared with the expected frequencies and checked for Hardy–Weinberg equilibrium. Differences in genotypic and allelic distributions between the groups were estimated using the Chi square (χ 2) test. Data normality was determined by the Kolmogorov–Smirnov test. Mean differences between genders and genotypes were assessed by the two-tailed unpaired Student’s t test for normally distributed parameters and the Mann–Whitney U test for nonparametric. For multiple comparisons of means between genotypes, one-way ANOVA followed by Tukey’s multiple comparison test was performed. A p value <0.05 was considered statistically significant. The analyses were performed using SPSS version 17 (SPSS, Evanston, IL, USA).

Results

Subjects’ Characteristics

Most of rural subjects were farmers, whereas the majority of the urban subjects were government officers. Urban subjects had a higher education level than rural subjects, and the current alcohol consumption was significantly higher in urban subjects than in rural subjects. However, there were no significant differences in the number of current smokers, former smokers, and exercise habits between rural and urban subjects (Table S2). The anthropometric and biochemical variables of the study subjects (307 urban, 188 rural) are shown in Table 1. The urban subjects (total group) showed significantly higher TC and LDL-C levels and lower BMI, HDL-C, and FBS levels compared with the rural subjects (all p < 0.05).

Table 1 Anthropometric and biochemical variables of study subjects
Full size table

Genotypic and Allelic Distributions

The genotypic and allelic frequencies of PCSK9 R46L, I474V, and E670G and APOE polymorphisms in urban and rural subjects are shown in Table 2. All of the observed genotypic frequencies were in Hardy–Weinberg equilibrium. The genotypic and allelic frequencies of PCSK9 I474V, E670G, and APOE polymorphisms, but not PCSK9 R46L polymorphism were not significantly different between urban and rural subjects (Table 2).

Table 2 Genotypic and allelic frequencies of PCSK9 (R46L, I474V, and E670G) and APOE polymorphisms in urban and rural subjects
Full size table

Relationship Between PCSK9 Genotypes and Anthropometric and Biochemical Variables

The anthropometric and biochemical variables according to PCSK9 genotypes in males, females, and the total group in urban and rural subjects are shown in Tables 3, 4, and Table S3, respectively.

Table 3 Anthropometric and biochemical variables according to PCSK9 genotypes (R46L, I474V, and E670G) in males in urban and rural subjects
Full size table
Table 4 Anthropometric and biochemical variables according to PCSK9 genotypes (R46L, I474V, and E670G) in females in urban and rural subjects
Full size table

The 46L urban male carriers showed significantly higher diastolic blood pressure (DBP) and FBS levels (p < 0.05), whereas the 46L urban female carriers had significantly lower TC and LDL-C levels (p < 0.05). In the rural subjects, the statistical analysis was not determined between 46L carriers and non-carriers because of the lack of the L allele. In males, compared with the urban 46RR genotype, rural 46RR genotype was associated with significantly lower BMI, WC, TC, LDL-C levels, and higher HDL-C levels (p < 0.05). In contrast, compared with the urban female 46RR genotype, the rural female 46RR genotype was associated with significantly higher BMI, WC, DBP, triglyceride (TG), HDL-C, FBS levels, and lower TC and LDL-C levels (p < 0.05).

The 474V rural female carriers had significantly lower HDL-C levels compared with non-carriers (p < 0.05). In comparison with the urban male 474II genotype, the rural male 474II genotype had significantly lower BMI, WC, DBP, TC, LDL-C levels, and higher HDL-C levels (p < 0.05). In contrast, the rural female 474II genotype had significantly higher BMI, WC, DBP, TG, HDL-C, FBS levels and lower LDL-C levels compared with the urban female 474II genotype (p < 0.05). In addition, the rural female 474IV genotype had significantly higher DBP, and lower HDL-C levels compared with the urban female 474IV genotype (p < 0.05).

Moreover, the 670G urban female carriers had significantly higher TC and LDL-C levels than non-carriers (p < 0.05). In contrast, no significant differences in lipid levels and other parameters between 670G carriers and non-carriers among the rural subjects were observed. In males, compared with the urban 670EE genotype, the rural 670EE genotype had significantly lower BMI, WC, DBP, TC, LDL-C levels, and higher HDL-C levels (p < 0.05). In females, rural 670EE genotype showed significantly higher BMI, WC, DBP, TG, HDL-C, FBS levels, and lower LDL-C levels compared with urban 670EE genotype (p < 0.05). Furthermore, significantly higher TC and LDL-C levels were observed in urban females with the EG genotype compared with rural females with the EG genotype (p < 0.05).

Relationship Between APOE Genotypes and Anthropometric and Biochemical Variables

The anthropometric and biochemical variables according to APOE genotypes in males and females, as well as in the total group, in the urban and rural subjects are shown in Table 5 and Table S4, respectively. Subjects with the E2E4 genotype were excluded from this analysis. The APOE genotype had a significant effect on TC and LDL levels in both urban males and females (p < 0.05). No significant differences were observed in other parameters. In comparison with APOE3 carriers, APOE2 carriers showed significantly lower TC and/or LDL-C levels in both urban males and females, whereas APOE4 carriers showed significantly higher TC and LDL-C levels than APOE3 carriers in urban males (p < 0.05).

Table 5 Anthropometric and biochemical variables according to APOE genotypes in males and females in urban and rural subjects
Full size table

PCSK9 E670G Polymorphism and Relation to APOE Polymorphism

The analysis of a combination of PCSK9 E670G and APOE polymorphisms in the urban subjects showed an increase in TC and LDL-C levels, with an increase from the lowest level in non-APOE4-PCSK9 670EE carriers to the highest level in APOE4-PCSK9 670EG carriers (p value for trend = 0.000 for both TC, and LDL-C; Table 6). In addition, non-APOE4-PCSK9 670EG carriers had significantly higher TC and LDL-C levels than non-APOE4-PCSK9 670EE carriers (p < 0.05).

Table 6 Anthropometric and biochemical variables of carriers of the PCSK9 E670G according to APOE alleles in urban subjects
Full size table

Discussion

In the present study, we investigated the association of PCSK9 and APOE polymorphisms with lipid levels in a Southern Thai population. Urban subjects had significantly higher TC and LDL-C, and lower BMI, HDL-C, and FBS levels than did rural subjects, results consistent with a previous study [37]. We observed a significantly higher frequency of current alcohol consumption and higher education in urban subjects relative to rural subjects. Most urban subjects were government officers, whereas most rural subjects were farmers, suggesting that the differences in lipid profiles, BMI, and FBS levels between urban and rural subjects could be explained by differences in lifestyle and sociodemographic factors.

In the present study, the frequency of L allele of PCSK9 R46L polymorphism was 2.28 and 0 %; in the urban and rural subjects, respectively. This allele frequency in urban subjects was similar to the Caucasian individuals (2.40 %) [23], whereas lower frequency of the L allele of R46L was observed in an Italian (1.04 %) population [18], in the CARDIA study (1.64 % in whites) [16], in the Dallas Heart Study (DHS) (1.6 % in whites, 0.75 % in Hispanics, and 0.28 % in black) [22], in the Copenhagen City Heart Study (1.22 %) [19], Copenhagen General Population Study (1.41 %) [19], and Copenhagen Ischemic Heart Disease Study (1.18 %) [19]. In addition, a higher frequency of L allele of R46L was observed in the French Canadian population (4.80 %) [23].

The frequency of the V allele of PCSK9 I474V polymorphism was 2.77 and 1.86 % in the urban and rural subjects, respectively. This finding in urban subjects was similar to that of a Japanese population (2.48 %) [24]. However, the higher frequency of the V allele of I474V has been observed in many populations: Brazilian subjects (17.59 %) [26], Caucasian Canadian (12.60 %) [23], African Canadians (19.40 %) [23], African Americans (22 %) [23], and healthy UK men (17.68 %) [15], as well as in the Dallas heart study (DHS) (18 % in whites, 9.7 % in Hispanics, and 22 % in blacks) [22] and the PLIC study (17.59 %) [25].

The frequency of the G allele of PCSK9 E670G polymorphism was 1.95 and 1.06 %, in urban and rural subjects, respectively. This finding in urban subjects was similar to that in reported in a Guangxi Bai Ku Yao (2 %) population [32], whereas a higher frequency of the G allele was observed in Germany (5 %) [27], the UK (3.4 %) [15] ), Chinese Taiwanese (11.9 %) [31], and Han Chinese (4.8 %) [32] populations, as well as in the Dallas Heart Study (DHS) (3.6 % in whites, 4.2 % in Hispanics, and 26.0 % in blacks) [22], the TexGen population (4.4 %) [30], the Lipoprotein Coronary Atherosclerosis Study (LCAS) (7.4 %) [30], and the PROSPER study group (6 %) [14]. Overall, we indicate that the frequencies of the PCSK9 46L, 474V, and 670G alleles vary across populations.

In the present study, we found that the 46L urban male carriers had significantly higher DBP and FBS compared with non-carriers. In contrast, the 46L urban female carriers had significantly lower TC and LDL-C levels compared with non-carriers. Our results in urban females were consistent with previous studies. Several populations in the UK [15], Sweden [17], Italy [18], Denmark [19], white populations [21], and hypobetalipoproteinemia subjects, as well as in the PROSPER study group [14], the CARDIA Study [16], and the Bogalusa Heart Study [20] have shown that 46L carriers are associated with lower LDL-C levels.

In contrast to 46L urban carriers, the 474V rural female carriers had significantly lower HDL-C levels than non-carriers (p < 0.05). It has been shown that 474V carriers are associated with lower LDL-C and PCSK9 levels in Caucasian, African Canadian [23], and Japanese [24] populations, but the association of 474V carriers with lipid levels was not observed in healthy UK men [15], Brazilian subjects [26], or in the PLIC study [25]. This suggests that differences in study population characteristics, study design, and the frequency of the 474V allele may explain the inconsistent results.

Furthermore, the 670G urban female carriers had significantly higher TC and LDL-C levels than did non-carriers. Our findings were inconsistent with the previous studies in Germany [27], which showed that the association of E670G with increased LDL-C in men, but not in women. However, our results were consistent with previous studies conducted in Brazilian [26] and Tunisian populations [28], as well as in the Belgium stroke study [29], the PLIC study [25], and the LCAS [30], all which showed the association of E670G with an increased risk of CVD. In the present study, the association of PCSK9 E670G with lipid parameters was not observed in the rural subjects. This finding was compatible with the healthy UK men [15] and the PROSPER study group [14], which showed no association of E670G with any significant effects on lipid levels or CVD risk. Moreover, some studies demonstrated the preventive effect on CVD risk by E670G. In Han populations, 670G carriers had significantly higher serum HDL-C and APOAI levels in males, and lower ApoB levels and ApoAI/ApoB ratio in females than the non-carriers [32]. In a Chinese population in Taiwan, E670G was found to be associated with low LDL-C levels are not a risk variant for CVD [31]. Altogether, the discrepancies of the association of PCSK9 R46L, I474V, and E670G polymorphisms with lipid levels among various populations may be due to ethnicity and/or environmental factors, as well as gene–environment interactions.

In the present study, the various effects on anthropometric data and lipid levels between urban and rural carriers of the 46RR, 474II, 474IV, 670EE and 670EG genotypes in both males and females were observed. Our results suggest that the differences in lifestyle and sociodemographic factors between urban and rural subjects may be an influential factor on the anthropometric data and lipid profiles according to each genotype. In a previous study, Aung et al. [38] reported that alcohol consumption can modify the effects of the PCSK9 E670G polymorphism on serum TC and LDL-C levels. Subjects with the AA genotype of PCSK9 E670G benefit more from alcohol consumption than do the subjects with the AG genotype.

Additionally, we found that APOE4 carriers had increased TC and LDL-C levels relative to APOE3 carriers in the urban males. APOE2 carriers had decreased TC and/or LDL-C levels compared with APOE3 carriers in both urban males and females. Our results are similar to other studies [39, 40]. Finally, we also observed the combined effect of APOE4 and PCSK9 E670G polymorphisms on TC and LDL-C levels in the urban subjects. A significant trend of increased TC and LDL-C was observed in non-APOE4-PCSK9 670EE carriers to APOE4–PCSK9 670EG carriers in urban subjects. Our finding was similar to the PLIC study, which showed that APOE2-PCSK9-670EE carriers had a more favorable lipid profile and decreased intima media thickness compared with APOE4–PCSK9-670G carriers [25].

In conclusion, we demonstrated the association of PCSK9 and APOE polymorphisms with serum lipid levels in a Southern Thai population. PCSK9 R46L, I474V, and E670G polymorphisms may represent genetic risk factors for CVD in urban males, rural females and urban females, respectively. However, females with PCSK9 R46L showed a favorable lipid profile that may protect against CVD in urban subjects. The combination of APOE and PCSK9 E670G polymorphisms may be suggestive of an independent factor for the determination of lipid levels. Our study had several limitations, including a small sample size, the low frequencies of the minor alleles of PCSK9 polymorphisms, and a lack of data on nutritional status. Further study using a larger sample is needed to confirm our findings. Additionally, the interaction between PCSK9 polymorphisms and environmental factors on lipid profiles is required.