Genotyping and Frequency of PCSK9 Variations Among Hypercholesterolemic and Diabetic Subjects

Abstract

Non-synonymous single-nucleotide polymorphism (SNPs) in the gene for proprotein convertase subtilisin/kexin type 9 (PCSK9) can influence cholesterol and glucose metabolism, leading to increased risk of cardiovascular disease and diabetes. To determine the frequency of four common PCSK9 SNPs, L10Ins, A56V, I474V, and E670G, in a population sample (n = 98) of the Hail region of Kingdom of Saudi Arabia. Blood was collected from participants; serum cholesterol, blood glucose and glycated hemoglobin were determined; genomic DNA was extracted and PCR amplicons from SNP-containing PCSK9 exons were subjected to Sanger sequencing. Out of 98 participants. 10 (10.20%) carried none of the SNPs, 2 (2.04%) the L10ins/A56V linked SNPs, 35 (35.71%) the I474V SNP, 22 (22.45%) both the I474V and E670G SNPs, and 29 (29.59%) the E670G SNP. Of the 30 eucholesterolemic diabetics patients, 11 (36.66%) carried the I474V SNP, 10 (33.33%) the E679G SNP and 6 (20%) the I474V/E679G. SNPs. Of 63 diabetic patients, 26 (41.26%) carry I474V SNP and 22 (34.92%) carry E670G SNP. Our data demonstrated that the I474V and E670G PCSK9 variants are very frequent in the Hail region of Saudi Arabia and are found at even higher frequency among diabetics. Further investigations are needed to determine whether these variations or another variant segregating with them can explain its apparent association with diabetes in this population.

Introduction

Proprotein convertase subtilisin/kexin type 9 (PCSK9) down-regulates the level of cell-surface low-density lipoprotein receptor (LDLR) by inducing its lysosomal degradation, not only in the liver but also in insulin-producing pancreatic cells [1]. LDLR mediates the cellular uptake of circulating LDL-cholesterol (LDL-C). Liver LDLR is responsible for the clearance of plasma LDL-C [2]. Liver-secreted plasma PCSK9 antagonizes this action. Increased PCSK9 expression or activity would reduce LDL-C clearance, causing hypercholesterolemia, a risk factor for cardiovascular disease (CVD). Decreased PCSK9 expression or activity would increase LDL-C clearance, causing hypocholesterolemia, and protecting from CVD [3]. In pancreatic beta cells, LDLR-mediated entry of excess extracellular LDL-C impairs insulin production and secretion as well as cell survival [4–6]. The increase of PCSK9 activity in the bloodstream may down-modulate this entry and thus protect the cells. PCSK9-deficient mice are hypoinsulinemic and prediabetic [7].

Over the years, more than 160 genetic variations have been identified in the exons of human PCSK9. Many of them affect plasma cholesterol. Loss-of-function (LOF) variations lower PCSK9 expression (production or secretion) or activity (LDLR binding and degradation), reducing plasma cholesterol. Inversely, gain-of-function (GOF) variations do the reverse and increase plasma cholesterol, Potent LOF variations (e.g. R46L, Y142X, C769X) have been associated with significant protection against coronary artery disease (CAD) [8], and potent GOF mutations (e.g. S127R, D374Y) have been associated with potentially lethal autosomal dominant hypercholesterolemia (ADH) [9]. Correlations between PCSK9 and cholesterol levels have been established in many populations [10–13]. Interestingly recent large meta-analyses have shown that carriers of PCSK9 LOF variations to be at higher risk of diabetes mellitus [14, 15].

Hypercholesterolemia and diabetes are highly prevalent in Kingdom of Saudi Arabia [16–19]. In 2013, in the kingdom city of Hail, we randomly collected blood samples from subjects affected or not by these conditions. We singled out one particular subject who depicts a phenotype of obesity, diabetes, and hypercholesterolemia. Whole Exome Sequencing on the genomic DNA of this patient uncovered two variations in exons 9 (I474V) and 12 (E670G) of the PCSK9 gene [20]. In this current study, we used Sanger sequencing to investigate these PCSK9 exons on the DNA from the whole blood cohort to determine the frequency of these variations in this population sample.

Methods

Subjects

The present study was included 98 participants, out of which four groups were formed: 10 healthy participants (HS), 25 hypercholesterolemic participants (HC), 30 eucholesterolemic diabetic participants (DP) and 33 hypercholesterolemic/diabetic participants (HC/DP), all of the participants provided informed written permission before ingoing the investigational practice. The diagnosis of hypercholesterolemia was made according to the following criteria total cholesterol above 4 mmol/L. Also, the diagnosis of diabetes was according to American Diabetes Association patient have fasting glucose ≥ 7 mmol/L will be diabetic. Our investigation protocol was performed according to the Helsinki Declaration and institutional ethics committee approval was obtained from College of Medicine, University of Hail, Saudi Arabia.

Blood samples were taken for biochemical analysis following overnight fasting. Serum total cholesterol and fasting blood glucose levels were assayed using the routine clinical technique. Blood glycosylated hemoglobin (HbA1c) was assayed according to Fluckiger and Winterhalter [21].

Whole Blood Genomic DNA Extraction, PCR and Sanger Sequencing

Genomic whole blood DNA was extracted and purified using QIAamp DNA Blood Mini Kit from Qiagen (Hilden, Germany). DNA was quality checked using Qubit fluorimeter (Invitrogen, Carlsbad, CA). We proceeded to Sanger sequencing as followed. Using primers (Table 1) flanking PCSK9 exons 1, 9 and 12 which contain the L10ins/A56V, A474V and E670 variations, respectively, PCR amplicons were generated, purified and sequenced in both directions on a Genetic Analyzer 3500XL automated sequencer with BigDye terminator cycle sequencing reagents (Applied Biosystems, Riyadh, SA).

Table 1.

Primers name used in PCR amplification

Primers names	Exons	Sequences
PCSK9-474F	9	GCC CTC CTC TCT CCT ACC AT
PCSK9-474R		GTC ACC TCC ATG CGC TCG CCC CG
PCSK9-1F	1	CCGTTCAGTTCAGGGTCTGAG
PCSK9-1B		TCAAGATCGTGCCAAGCG
PCSK9-12F	12	TGGTAGGCATCTGTCTATCTCC
PCSK9-12B		GAAGCATCCCCATCCCC

Statistical Analysis

The data analysis was carried out using statistical software SPSS 18.0 (SPSS Inc., Chicago, IL, USA). Data are presented as a mean ± standard error. Statistical analyses were performed using one-way ANOVA. P < 0.05 was considered statistically significant.

Results and Discussion

The 98 participants were divided into four groups: 10 healthy participants (HS), 25 hypercholesterolemic participants (HC), 30 eucholesterolemic diabetic participants (DP) and 33 hypercholesterolemic/diabetic participants (HC/DP) (Table 2). Most were male (62/98).

Table 2.

Distribution of Common PCSK9 Variations among healthy, hypercholesterolemic, eucholesterolemic diabetic and hypercholesterolemic/diabetic participants

	N (F/M)	PCSK9 Variations
		None	L10/A53V	I474V	I474V/E679G	E670G
None (NO)	10 (3/7)	5	1	2	2	0
High cholesterol (HC)	25 (10/15)	2	1	7	8	7
Diabetes mellitus (DM)	30 (10/20)	3	0	11	6	10
HC and DM (HC/DM)	33 (13/20)	0	0	15	6	12
Total	98 (36/62)	10	2	35	22	29

Alterations in the levels of fasting blood sugar (FBS), glycated hemoglobin (HbA1c) and total cholesterol among healthy, hypercholesterolemic, eucholesterolemic and hypercholesterolemic/diabetic participants were depicted in Figs. 1, 2 and 3 respectively. Hypercholesterolemic, eucholesterolemic and hypercholesterolemic/diabetic participants showed a significant elevation in fasting blood sugar, glycated hemoglobin and total cholesterol levels as compared to that of healthy participants.

Fig. 1.

Level of fasting blood glucose in healthy, hypercholesterolemic, eucholesterolemic diabetic and hypercholesterolemic/diabetic participants (Mean ± SE)

Fig. 2.

Level of glycated hemoglobin in healthy, hypercholesterolemic, eucholesterolemic diabetic and hypercholesterolemic/diabetic participants (Mean ± SE)

Fig. 3.

Level of serum cholesterol in healthy, hypercholesterolemic, eucholesterolemic diabetic and hypercholesterolemic/diabetic participants (Mean ± SE)

The 3 PCSK9 exons of interest were sequenced from the genomic DNA of all participants; common and variant alleles were identified (Figs. 4, 5, 6, 7). Eighty-eight participants i.e. (89.79%) carried at least 1 SNP: 2 participants (2.04%) the exon-1 L10ins and A53V SNPs, 35 participants (35.71%) the exon-9 I474V SNP, 22 participants (22.44%) the exon-9 I474V and exon-12 E670G SNPs, and 29 participants (29.59%) the exon-12 E670G SNP (Table 2). This frequency distribution is quite different from that observed in the United States and Canadian sample populations in which the minor allele frequency (MAF) of the L10ins/A56V, I474V, and E670G SNPs are around 16, 13, and 4%, respectively [22, 23]. It should be mentioned that the L10ins SNPs is invariably linked to A56V in Caucasians, but not sub-Saharan Africans and that it is a LOF phenotype in the former population, but not in the latter [23].

Fig. 4.

Exon 1 Normal subjects and variant allele. PCSK9 gene sequencing electropherogram around SNP A53V (exon 1). a Alanine codon at the position 53 (common allele GCC). b Nonsense mutation. c.158 T > C; GCC → GTC (Ala → Val) (variant allele)

Fig. 5.

Exon 1 CTG repeats: a normal subject b variant allele. PCSK9 gene sequencing electropherogram around SNP L10Ins (exon 1). a Common allele normal subjects. b Variant alleles (eucholesterolemic subjects) characterized by Mutation in leucine repeat, insertion of CTG into CTG repeat

Fig. 6.

Exon 9 a Heterozygous allele; b variant allele. 6 PCSK9 gene sequencing electropherogram around SNP I474V (exon 9). a Heterozygous alleles from hyperglycemic and hypercholesterolemic subjects around Ile474Val SNP. b Homozygous alleles, hyperglycemic and hypercholesterolemic subjects characterized by c.1420 A > G, CGT → CAT, (Ile → Val) (variant allele)

Fig. 7.

PCSK9 gene sequencing electropherogram around SNP E670G (exon 12). a Common allele at position 2009 normal subjects. b Variant alleles (hypercholesterolemic subjects)

Most striking on this table is the fact that 26 out of 63 diabetic participants, DM and HC/DM combined (41.27%) carried the I474V SNPs and 22 out of 63 patients (34.92%) carry E670G SNP, altogether representing a situation more than double the combined frequency of all the other alleles.

The high frequency of the I474V and E670G variants among diabetic participants cannot be explained on the basis of current knowledge on the cellular biology of PCSK9 variants. These variants cannot be ascribed to a LOF or GOF phenotype since it has been associated with a wide range of plasma cholesterol levels [22, 24]. However, these variants were observed in two Omani ADH patients [25]; they have also been found to segregate with CAD in a cohort of Chinese subjects [26]. Whether these PCSK9 variants directly contributes to these pathologic conditions remain to be elucidated by in-depth studies. The influence of other genetic mutations tightly linked to this variation cannot be excluded as yet.

In our study, carriers of the I474V and E670G variants have a mean plasma cholesterol of 4.31 ± 0.23 mmol/L compared to 2.14 ± 0.04 in carriers of L10ins/A56V variants (P = 0.001 by 1-way ANOVA) and 5.39 ± 0.51 in carriers of I474V/E670G variants (not significantly different) (Fig. 1). The small number of participants and the anti-cholesterol treatment received by nearly half (49/98) of them do not permit any strong inference as to the LOF or GOF nature of these variations in this population sample. Nonetheless, the lower plasma cholesterol of the six carriers of L10ins/A56V variations is well in line with the LOF phenotypic effect in other populations [23, 26, 27].

Agarose gel electrophoresis images showing PCSK9 PCR products amplified from blood genomic DNA a exon 1, b exon 9, and c exon 12. Lane (1) and lane (2) respectively represent amplicons harboring common alleles and variant alleles

Conclusion

In this pilot study, we have determined the frequency of four common PCSK9 SNPs in a small sample of the population of the Hail region of Saudi Arabia. A frequency of 35.71 and 29.59% for the I474V and E670G variations were the most common; both were roughly 60% of diabetic patients. We also confirmed the association of L10ins/A56V variations with the lower plasma cholesterol carriers. More studies are needed to determine why this reportedly neutral variation appears to pathogenic in this population.

Abbreviations

ADH

Autosomal dominant hypercholesterolemia

CAD

Coronary artery disease

GOF

Gain of function

LDL-C

Low-density lipoprotein-cholesterol

LDLR

Low-density lipoprotein receptor

LOF

Loss of function

PCSK9

Proprotein convertase subtilisin/kexin type 9

SNP

Single-nucleotide polymorphism

FBG

Fasting blood glucose

HbA1c

Glycated hemoglobin