Influence of PCSK9 polymorphisms on plasma lipids and response to atorvastatin treatment in Brazilian subjects

Journal of Clinical Lipidology (2014) 8, 256–264

Influence of PCSK9 polymorphisms on plasma lipids and response to atorvastatin treatment in Brazilian subjects Jacqueline M. Anderson, PharmD, Alvaro Cerda, PhD, Mario H. Hirata, PhD, Alice C. Rodrigues, PhD, Egidio L. Dorea, PhD, Marcia M. S. Bernik, PhD, Marcelo C. Bertolami, PhD, Andre A. Faludi, PhD, Rosario D. C. Hirata, PhD* School of Pharmaceutical Sciences, University of Sao Paulo, Av. Prof. Lineu Prestes, 580, B17, Sao Paulo 05508–000, SP, Brazil (Drs. Anderson, Cerda, M.H. Hirata, and R.D.C. Hirata); Center of Molecular Biology and Pharmacogenetics, Scientific and Technological Bioresource Nucleus, Universidad de la Frontera, Temuco, Chile (Dr. Cerda); Institute of Biomedical Sciences, University of Sao Paulo, Sao Paulo, SP, Brazil (Dr. Rodrigues); University Hospital, University of Sao Paulo, Sao Paulo, SP, Brazil (Drs. Dorea and Bernik); and Dante Pazzanese Institute of Cardiology, Sao Paulo, SP, Brazil (Drs. Bertolami and Faludi) KEYWORDS: Proprotein convertase subtilisin/kexin type 9 (PCSK9); Single nucleotide polymorphism; Cholesterol; Atorvastatin; Pharmacogenomics

BACKGROUND: The proprotein convertase subtilisin/kexin type 9 (PCSK9) has a key role in the regulation of plasma low-density lipoprotein (LDL) cholesterol by enhancing the degradation of LDL receptor. Functional variants in PCSK9 have been associated with differences in plasma lipids and may contribute to the variability of the response to cholesterol-lowering drugs. OBJECTIVE: To investigate the influence of PCSK9 variants on plasma lipid profile and response to atorvastatin in Brazilian subjects. METHODS: PCSK9 E670G, I474V, and R46L single nucleotide polymorphisms (SNPs) and plasma lipids were evaluated in 163 hypercholesterolemics (HC) and 171 normolipidemics (NL). HC patients with indication for cholesterol-lowering drug therapy (n 5 128) were treated with atorvastatin (10 mg/ d/4 wk). PCSK9 SNPs were analyzed by real time polymerase chain reaction. RESULTS: Frequencies of the PCSK9 SNPs were similar between the HC and NL groups. Logistic regression analysis showed a trend of association between PCSK9 E670G and hypercholesterolemia after adjustment for covariates (P 5 .059). The 670G allele was associated with high basal levels of LDL cholesterol (P 5 .03) in HC patients using the extreme discordant phenotype method. No association tests were performed for R46L variant because of its very low frequency, whereas the I474V polymorphism and PCSK9 haplotypes were not related to hypercholesterolemia or variability on plasma lipids in both NL and HC groups (P . .05). LDL cholesterol reduction in response to atorvastatin was not influenced by PCSK9 genotypes or haplotypes. CONCLUSIONS: PCSK9 E670G polymorphism but not I474V contributes to the variability on plasma LDL cholesterol levels in hypercholesterolemic subjects. Both PCSK9 variants have no influence on cholesterol-lowering response to atorvastatin. Ó 2014 National Lipid Association. All rights reserved.

* Corresponding author. E-mail address: [email protected] Submitted October 30, 2013. Accepted for publication February 26, 2014.

Hypercholesterolemia has become an important public health problem because it is a major risk factor for diseases such as coronary artery disease (CAD), hypertension, and

1933-2874/$ - see front matter Ó 2014 National Lipid Association. All rights reserved. http://dx.doi.org/10.1016/j.jacl.2014.02.008

PCSK9 variants and atorvastatin response

257

stroke. It is caused by the interactions between environmental and genetic factors, such as variants in genes involved in the cholesterol biosynthesis, transport, and metabolic pathways.1–3 The autosomal dominant hypercholesterolemia (ADH) is a monogenic disorder that results in increase of the circulating low-density lipoprotein (LDL). Three genes have been related to ADH: LDL receptor (LDLR) that accounts for most of the identified mutations; apolipoprotein B (APOB); and, the most recently identified, the proprotein convertase subtilisin/kexin type 9 (PCSK9).4 The PCSK9 encodes a serine protease of the mammalian serine proprotein convertase family, which is responsible for the proteolytic maturation of secretory proteins such as growth factors, cytokines, pro-hormones, and receptors.5 The PCSK9 has a key role in the regulation of plasma LDL by mediating LDLR breakdown through a posttranscriptional mechanism, in which PCSK9 binds to the LDLR and interferes with the recycling of the receptor protein from the endosome to the cell surface, resulting in the receptors being rerouted to the lysosome for degradation. The region within the LDLR that is bound directly by PCSK9 corresponds to the first epidermal growth factorlike repeat located within the receptor EGF precursor homology domain.6 PCSK9 can also modify the rate of secretion of apo-B 100 from hepatocytes in an LDLR-independent manner. The interaction of the PCSK9 with apoB-100 results in increased production of apoB-100, possibly through the inhibition of intracellular degradation.7 PCSK9 is a highly polymorphic gene. Gain-of-function mutations in PCSK9 cause increased enzyme activity, which results in high rate of LDLR degradation and ADH-associated increase in LDL cholesterol. Loss-offunction mutations that inactivate PCSK9 result in low plasma levels of LDL cholesterol and confer protection for CAD.8 Previous studies have investigated the relationship of PCSK9 single nucleotide polymorphisms (SNPs), including the 2 common variants E670G and I474V, and changes in serum lipids showing inconsistent results. The PCSK9 E670G SNP was associated with polygenic hypercholesterolemia in men but not in women in a sample from a European population,9 whereas in a Chinese healthy population the 670G allele was associated with high serum high-density lipoprotein (HDL) cholesterol and ApoAI in men, and low serum ApoB levels and high the ApoAI/ApoB ratio in women.10 Canadian Caucasian individuals carrying the 474V variant had lower LDL cholesterol levels than noncarriers and the 670G allele was related to lower total cholesterol and triglycerides.11 In a Japanese healthy population, the 474V allele was associated with low levels of total and LDL cholesterol.12 On the other hand, both E670G and I474V SNPs were not associated with LDL cholesterol levels in black and white Americans from the Dallas Heart study8 and the Coronary Artery Risk Development in Young Adults Study.13

Statins are inhibitors of cholesterol biosynthesis that have high efficacy in lowering LDL cholesterol levels14 and result in clinically significant reductions in cardiovascular morbidity and mortality.15 Although statin treatment has proven beneficial effects, there is large variability in clinical response to statin therapy as well as in the incidence of adverse effects. Several studies have shown that this interindividual variability is associated with SNPs in genes related to lipid homeostasis and statins metabolism.16–19 Previous studies have demonstrated that carriers of lossof-function mutations in the PCSK9 can have a better response to statins suggesting that lipid-lowering by PCSK9 inhibitors may be synergistic to that achieved by statins treatment.20 Gain-of-function mutations in PCSK9 were associated with attenuated statin-mediated reduction of LDL cholesterol.21,22 The present study investigated the influence of the PCSK9 variants on plasma lipids and cholesterol-lowering response to atorvastatin in Brazilian subjects.

Anderson et al

Subjects and methods Subjects and study protocol A total of 334 individuals were randomly selected at the University Hospital of the University of Sao Paulo and Institute Dante Pazzanese of Cardiology (Sao Paulo, SP, Brazil). Individuals with LDL cholesterol higher than 160 mg/dL or on cholesterol-lowering therapy were classified as hypercholesterolemic (HC, n 5 163) and those with LDL cholesterol lower than 130 mg/dL and triglycerides lower than 150 mg/dL were considered normolipidemic (NL, n 5 171). Individuals with diabetes mellitus, thyroid, liver and kidney diseases, and triglycerides higher than 400 mg/dL and pregnant women or those on oral contraceptives were not included in this study. Information on age, body mass index (BMI), gender, hypertension, obesity, menopause status, cigarette smoking, physical activity, alcohol consumption, family history of CAD, and medication use was recorded. The study protocol was approved by the Ethics Committees of the University Hospital of University of Sao Paulo (Protocol # 811/08), School of Pharmaceutical Sciences of University of Sao Paulo (Protocol # 472), and Institute Dante Pazzanese of Cardiology (Protocol # 2077/2000). Each individual agreed to participate in the study by signing an informed consent form. All HC subjects were instructed to consume a lowcholesterol diet during 4 weeks without lipid-lowering therapy; after this period, those HC individuals with indications of cholesterol-lowering therapy (n 5 128), to reach the LDL cholesterol goal according to the IV Brazilian Guidelines for Dyslipidemia and Atherosclerosis Prevention,23 were started on atorvastatin treatment (10 mg/ d/4 wk), as previously described.24 The other 35 individuals without indication of atorvastatin treatment were

258 maintained in the HC group because all of them had LDL cholesterol values higher than 130 mg/dL after a low-cholesterol diet. At the end of the protocol, plasma lipid response to atorvastatin as well as creatine kinase (CK) and alanine aminotransferase (ALT) enzymes were measured to detect possible statin-related adverse drug reactions on muscle and liver tissues.

Blood samples and laboratory data Blood samples for biochemical profile measurements and genomic DNA extraction were drawn after an overnight (12 hour) fast. Patients receiving atorvastatin had blood drawn before and after the 4-week treatment. Total cholesterol, HDL cholesterol, and triglycerides were measured by routine enzymatic colorimetric methods. Very low-density lipoprotein (VLDL) and LDL cholesterol were calculated by Friedewald’s formula. ApoAI and apoB were measured by nephelometry. ALT and CK activities were determined by kinetic methods.

PCSK9 polymorphism analysis Genomic DNA was extracted from EDTA-anticoagulated blood samples using a salting-out procedure.25 SNPs of the PCSK9 rs505151 (E670G, c.2009A.G), rs562556 (I474V, c.1420A.G), and rs11591147 (R46L, c.137 G.T) were detected by TaqMan real-time polymerase chain reaction (PCR). The primers and probes were obtained in the format SNP Genotyping Assays (20!) from Life Technologies (Foster City, CA) under catalog numbers C_998744_10 (SNP E670G), C_998751_10 (SNP I474V), and C_2018188_10 (SNP R46L). The thermal cycling protocol consisted of initial cycle at 10 minutes at 95
C followed by 40 cycles at 92
C for 15 seconds and 60
C for 1 minute. The amplification was carried out in a 7500 fast real-time PCR system (Life Technologies). PCR assays were retested for 10% of the samples.

Statistical analysis Statistical analysis was performed using the Sigma Stat software v. 3.5 (SPSS Inc., Chicago IL) and Minitab v.15 Statistical Software (Minitab Inc., State College, PA). Continuous variables were compared by t-test or MannWhitney rank-sum test and categorical variables were compared by chi-square test. Differences in mean values before and after atorvastatin treatment were compared by paired t-test. Skewed variables were logarithmically transformed to improve normality for statistical analysis. Hardy-Weinberg equilibrium was evaluated by chi-square test and haplotype frequencies were estimated by the expectation–maximization algorithm using the Haploview software. The linkage disequilibrium (LD) was measured accessing the LD coefficient (Lewonstin’s D0 ) between each pair of SNPs using the Haploview software and we

Journal of Clinical Lipidology, Vol 8, No 3, June 2014 considered substantial LD with D0 $ 0.50. The extreme discordant phenotype (EDP) method was applied to examine the relationship of the PCSK9 variants with extreme basal levels of LDL cholesterol in HC subjects. The EDP approach contrasted the ratio of the variant allele frequencies at opposite ends of the baseline LDL cholesterol (which correspond to the lower and upper ends of the LDL cholesterol distribution histogram) measured at various cutoff points of basal levels of LDL cholesterol, corresponding to the percentiles of the number of HC subjects (10, 20, and 30 percentiles). Logistic and linear regression analysis was performed to evaluate the association of genotypes with hypercholesterolemia and LDL variation in response to atorvastatin. Additionally, age, gender, ethnics, obesity, hypertension, alcohol consumption, cigarette smoking, concomitant medication, and physical activity were included as covariates. Significance was considered at P , .05.

Results Clinical and laboratory data of the studied groups Clinical, laboratory, and anthropometric data of the HC and NL subjects before and after atorvastatin therapy are shown in Table 1. The mean values of age, BMI, and proportion of obese and hypertensive subjects and postmenopausal women were higher in the HC group than in NL subjects (P , .05). Both groups had more women than men, although the gender distribution was similar between the NL and HC groups. Ethnics, history of CAD, cigarette smoking, alcohol consumption, and physical activity frequencies were similar between the HC and NL groups. The HC group showed higher basal levels of total, LDL and VLDL cholesterol, triglycerides and ApoB than NL group (P , .05); however, the values of HDL cholesterol were similar in both groups (P 5 .389). After atorvastatin treatment (10 mg/d/4 wk), the serum lipids were significantly reduced with the exception of apoAI concentration, which did not change. According to CK and ALT monitoring, no adverse drug reaction cases were observed after atorvastatin treatment.

PCSK9 polymorphisms Genotype frequencies of PCSK9 E670G, I474V, and R46L polymorphisms were in the Hardy-Weinberg equilibrium confirming the random selection of the individuals. Frequencies of the PCSK9 E670G genotypes were similar between HC (AA: 73.0%; AG: 24.6%; and GG: 2.4%) and NL group (AA: 76.6%; AG: 21.6%; and GG: 1.8%, P . .05). For SNP I474V, genotypes had also similar frequencies between HC (AA: 68.7%; AG: 24.6%; and GG: 6.7%) and NL group (AA: 60.8%; AG: 35.1%; and GG: 4.1%, P . .05) (data not shown). The R46L polymorphism

Anderson et al Table 1

PCSK9 variants and atorvastatin response

259

Clinical, lipids and genetic data of hypercholesterolemic and normolipidemic subjects

Variable Clinical Number of individuals Age, y BMI, kg/m2 Men/women, % Ethnics (European/African), % Hypertension, % Obesity, % Menopause, % History of CAD, % Physical activity, % Cigarette smoking, % Alcohol consumption, % Plasma lipids Total cholesterol, mg/dL Baseline Treatment LDL cholesterol, mg/dL* Baseline Treatment HDL cholesterol, mg/dL* Baseline Treatment VLDL cholesterol, mg/dL* Baseline Treatment Triglycerides, mg/dL* Baseline Treatment ApoAI, mg/dL Baseline Treatment ApoB, mg/dL* Baseline Treatment PCSK9 polymorphism MAF, % E670G (c.2009A.G) I474V (c.1420A.G) R46L (c.137G.T) Haplotypes, % A1A2 G1A2 A1G2 G1G2

NL

HC

P value

171 47 6 7 26 6 4 74/26 61/39 35 15 29 44 47 22 29

163 55 6 10 28 6 4 66/34 64/36 55 28 75 55 48 22 28

,.001† ,.001† .138† .722† ,.001† .005† ,.001† .062† .931† .957† 0.835†

173 6 18 —

271 6 39 198 6 30

,.001‡ ,.001x

98 6 18 —

183 6 36 117 6 27

,.001† ,.001x

58 6 13 —

57 6 14 55 6 13

.389† ,.001x

16 6 5 —

30 6 13 26 6 11

,.001† ,.001x

82 6 28 —

151 6 67 132 6 54

,.001† ,.001x

141 6 27 —

136 6 27 139 6 28

,.001† .078x

85 6 22 —

144 6 27 99 6 20

,.001† ,.001x

12.6 21.6 1.8

14.7 19.1 0.3

.486 .457 .145

68.8 18.6 9.5 3.1

68.1 17.1 12.8 1.9

.845 .625 .172 .331

ApoAI, apolipoprotein AI; ApoB, apolipoprotein B; BMI, body mass index; CAD, coronary artery disease; HDL, high-density lipoprotein; LDL, lowdensity lipoprotein; MAF, minor allele frequency; VLDL, very low-density lipoprotein. 1,2: allele for I474V and E670G, respectively. Continuous variables are shown as mean 6 standard deviation. *Variables that were logarithmically transformed. Categorical variables were compared by chi-square test. †P values for NL vs HC basal compared by t-test. ‡Mann-Whitney rank-sum test. xP values for HC basal vs HC posttreatment compared by paired t-test.

also presents similar genotypes frequencies between NL (GG: 96.5%; GT: 3.5%; and TT: 0%) and HC group (GG: 99.4%; GT: 0.6%; and TT: 0%). The minor allele frequencies (MAF) of PCSK9 polymorphisms E670G (HC:

14.7%; NL: 12.6%; P 5 .486), I474V (HC: 19.1%; NL: 21.6%; P 5 .457) and R46L (HC: 0.3%; NL: 1.8%; P 5 .145) were also similar between HC and NL subjects (Table 1). Because of its very low frequency in both HC

260

Journal of Clinical Lipidology, Vol 8, No 3, June 2014

and NL groups, R46L SNP was not used for further association analysis in this study. Univariate logistic regression analysis showed that E670G and I474V polymorphisms are not related with hypercholesterolemia. However, after adjustment for covariates, there was a trend of association between E670G and hypercholesterolemia (P 5 .059; data not shown). No linkage disequilibrium was found between E670 G and I474V polymorphisms (D0 5 0.1). When these 2 variants were combined into haplotypes, 4 haplotypes were identified that had similar frequencies between HC and NL groups (P . .05) (Table 1).

PCSK9 polymorphisms, plasma lipids, and atorvastatin response The G allele carriers of E670G polymorphism shows a trend to have high basal LDL cholesterol levels in HC group (P 5 .07; Table 2). We applied the EDP methodology to examine the relationship of the PCSK9 variants with extreme basal levels of LDL cholesterol in HC subjects.

Table 2

This analysis revealed that the G allele frequency was higher in the group with the highest levels of LDL cholesterol (percentile 90, LDL.237 mg/dL) than in the group with the lowest levels (percentile 10, LDL ,161 mg/dL) at the 10 percentile cutoff point (P 5 .03), even after adjustment by relevant covariates. G allele carriers are 6 times more likely to present increased basal levels of LDL cholesterol than the noncarriers at the 10th extreme percentile; this association decreases as increasing the cutoff to 20 and 30 percentiles (Fig. 1). There was no significant association of the PCSK9 I474V polymorphism and plasma lipids in HC, even when EDP was performed (Table 2). Moreover, the studied haplotypes were not significantly associated with extreme basal levels of LDL cholesterol at any cut off percentile, but interestingly the rare haplotype GG showed to be related to higher LDL cholesterol since it was not present in any of the groups with extreme low LDL cholesterol values (up to percentile 30), which can interfere with the power of the statistical tests (data not shown).

Relationship of PCSK9 polymorphisms with serum lipids in hypercholesterolemics, before and after atorvastatin treatment E670G

Variable (mg/dL) Total cholesterol* Baseline Treatment Change, % HDL cholesterol* Baseline Treatment Change, % LDL cholesterol* Baseline Treatment Change, % VLDL cholesterol* Baseline Treatment Change, % Triglycerides* Baseline Treatment Change, % ApoAI Baseline Treatment Change, % ApoB* Baseline Treatment Change, %

I474V

EE (91)

EG 1 GG (37)

P value

II (87)

IV 1 VV (41)

P value

279 6 37 197 6 31 229.1 6 9.4

286 6 37 201 6 29 229.2 6 9.4

.311 .440 .962

280 6 34 198 6 30 228.7 6 9.8

283 6 43 198 6 32 229.8 6 8.5

.600 .909 .534

56 6 14 54 6 13 22.8 6 1.5

58 6 12 56 6 11 23.06 6 9.2

.168 .264 .901

56 6 13 54 6 12 22.7 6 10.1

57 6 15 55 6 14 23.1 6 10.4

.761 .859 .819

189 6 34 115 6 28 238.6 6 12.5

198 6 35 121 6 25 238.5 6 12.1

.073 .335 .946

191 6 31 117 6 28 238.1 6 13.4

195 6 39 117 6 26 239.5 6 10.1

.980 .964 .548

33 6 14 24 6 11 218.5 6 35.7

29 6 11 24 6 10 220.1 6 32.8

.127 .452 .809

32 6 13 26 6 10 210 6 34

31 6 13 25 6 10 216 6 25

.747 .403 .679

166 6 71 135 6 56 213.4 6 29.7

144 6 54 122 6 47 212.7 6 27.5

.127 .179 .809

160 6 67 133 6 55 220 6 30

156 6 67 126 6 53 218 6 35

.438 .366 .466

137 6 29 135 6 41 22.2 6 10.1

135 6 22 134 6 32 22.3 6 14.3

.687 .902 .956

137 6 27 141 6 30 2.9 6 11.5

134 6 28 134 6 25 20.89 6 11.1

.575 .237 .351

138 6 33 99 6 20 229.2 6 12.2

140 6 39 98 6 20 232.9 6 11.6

.460 .624 .130

144 6 27 100 6 19 229.6 6 12.1

143 6 26 98 6 21 231.0 6 12.1

.820 .651 .555

ApoAI, apolipoprotein AI; ApoB, apolipoprotein B; HDL, high-density lipoprotein; LDL, low-density lipoprotein; VLDL, very low-density lipoprotein. 2 indicates decrease. Continuous variables are shown as mean 6 standard deviation and compared by t-test or Mann-Whitney rank-sum test. *Variables that were logarithmically transformed.

Anderson et al

PCSK9 variants and atorvastatin response

Figure 1 Extreme discordant phenotype analysis of the relationship between baseline low-density lipoprotein (LDL) cholesterol and PCSK9 E670G and I474V polymorphisms. Data are expressed as the ratio of the minor variant allele frequencies at opposite ends of the baseline LDL cholesterol, measured at various cutoff points of LDL cholesterol levels, corresponding to the percentiles of the number of hypercholesterolemic subjects. The plotted ratios refer to high-to-low baseline LDL cholesterol.

PCSK9 variants did not influence the plasma lipid profile in response to atorvastatin in HC group (Table 2) and in normolipidemic subjects (Table 3). Goals for LDL cholesterol reduction analysis also showed lack of relationship with these variants (Table 4). Moreover we did not find any association between PCSK9 genotypes and LDL reduction after atorvastatin treatment in HC individuals in the multiple linear regression model even after adjustment for covariates gender, age, ethnics, BMI, and physical activity (data not shown).

Discussion Diverse studies have shown the influence of PCSK9 SNPs on serum lipids levels; however, the results are inconsistent and there are few studies that have investigated the role of these polymorphisms in statin response. We have investigated the relationship of 3 PCSK9 polymorphisms Table 3

261 with plasma lipids and atorvastatin response in Brazilian subjects. In this study, the MAF for 670G allele (total group: 13.6%) was similar to that previously reported by other studies in black populations (21%–26%) but was higher than that reported for Caucasian populations (3.0%–5.0%).8,9,11,13,26,27 The MAF for 46L allele (total group: 1.0%) was also similar to those reported in black and white populations (0.28%–3.2%).8,11,13,26 The frequency of 474V allele (total group: 20.3%) was similar to that found in North Americans (15%–22%), African Canadians (19.4%), and European (17.9%) populations.8,11,26,27 These differences probably result from ethnic differences among these studies. Brazilians are a highly complex population because of their high admixture, with Amerindian, European, and African ancestral roots.28 Nevertheless, it is important to highlight that the polymorphisms studied in this work were evaluated in an urban Southeast Brazilian cohort, which is predominantly composed by a population of European descendant, followed by an African ancestry population.29 It is noteworthy that this study classified the ancestral origin of individuals by self-description of skin color so we do not have precise estimates of the ancestral composition of the individuals. A recent study proposed the use of the African Component of Ancestry to assess the ancestry in case-control association studies. This approach could be an important step to eliminate a potential confounding factor, particularly when subjects are taken from a complex urban population.30 However, there was a good correlation between African Component of Ancestry values and color self-categorization in that population study indicating that the self-description of skin color may not be a confounding factor.30 In this study, the E670G and I474V SNPs and their haplotypes were not related with hypercholesterolemia, although there was a trend of association of E670 G polymorphism after adjustment for covariates. Similarly, Scartezini et al did not find association of these haplotypes and hypercholesterolemia in the Northwick Park Heart Study II.26 Moreover, in a previous case-control study that enrolled both hypercholesterolemic and normolipidemic

Relationship of PCSK9 polymorphisms with serum lipids in normolipidemics E670G

I474V

Variable (mg/dL)

EE (131)

Total cholesterol HDL cholesterol LDL cholesterol VLDL cholesterol* Triglycerides* ApoAI ApoB*

173 59 98 16 81 142 80

6 6 6 6 6 6 6

18 14 18 6 28 27 27

EG 1 GG (40) 174 57 100 17 84 140 90

6 6 6 6 6 6 6

19 10 18 5 27 24 28

P

II (104)

.720 .446 .510 .442 .442 .796 .054

174 59 99 16 80 141 86

6 6 6 6 6 6 6

18 13 18 5 27 26 21

IV 1 VV (67) 171 57 98 17 84 142 82

6 6 6 6 6 6 6

18 12 19 6 29 28 22

P .267* .203 .724* .522 .521 .839 .309

ApoAI, apolipoprotein AI; ApoB, apolipoprotein B; HDL, high-density lipoprotein; LDL, low-density lipoprotein; VLDL, very low-density lipoprotein. Number of individuals is in parentheses. Continuous variables are shown as mean 6 standard deviation and compared by t-test or Mann-Whitney rank-sum test. *Variables that were logarithmically transformed.

262

Journal of Clinical Lipidology, Vol 8, No 3, June 2014

Table 4

Relationship between PCSK9 polymorphisms and atorvastatin response according to LDL goal in hypercholestrolemics LDL goal ,100 mg/dL

LDL goal ,130 mg/dL

Polymorphism

, LDL goal

. LDL goal

OR (95% CI)

P value

, LDL goal

. LDL goal

OR (95% CI)

P value

E670G

30.8% 27.0% 31.0% 26.8%

69.2% 73.0% 69.0% 73.2%

Reference 0.83 (0.35–1.95) Reference 0.81 (0.35–1.86)

.836

71.4% 67.6% 67.8% 75.6%

28.6% 32.4% 32.2% 24.4%

Reference 0.83 (0.36–1.90) Reference 1.47 (0.63–3.41)

.826

I474V

EE EG 1 GG II IV 1 VV

(28) (10) (27) (11)

(63) (27) (60) (30)

.781

(65) (25) (59) (31)

(26) (12) (28) (10)

.488

CI, confidence interval; LDL, low-density lipoprotein; OR, odds ratio. Number of individuals is in parentheses.

Brazilian subjects, we also did not find relationship between other PCSK9 variants (R46L, A53V, A443T, H553R, and C679X) and hypercholesterolemia.31 One important finding of our study was that the association of the 670G allele with high basal levels of LDL cholesterol in the HC group, which confirms the results of previous investigations in which the E670G polymorphism was associated with hypercholesterolemia and high levels of LDL cholesterol. Chen et al found that a common PCSK9 haplotype encompassing the E670G variant was an independent determinant of plasma LDL cholesterol levels and of the severity of coronary atherosclerosis in the Lipoprotein Coronary Atherosclerosis Study (LCAS) and the TexGen populations from Houston, Texas.27 The E670G polymorphism was associated with plasma total cholesterol, LDL cholesterol, and apoB levels in LCAS population and accounted for 3.5% of plasma LDL cholesterol variability.27 In a study of a European population (506 patients attending the lipid clinic of the University Hospital in Hamburg, Germany), the PCSK9 E670G polymorphism was associated with polygenic hypercholesterolemia in men but not in women.9 This finding indicates that the gender can influence the effect of this SNP on lipid levels, but the explanation for this gender difference is still unknown. Conversely, in the Dallas Heart Study no association was found between PCSK9 E670G polymorphism and plasma LDL cholesterol in 3.543 Hispanic and non-Hispanic black and whites.8 This variant was also not associated with LDL cholesterol levels in black and white American subjects from the Coronary Artery Risk Development in Young Adults Study.13 Similarly Scartezini et al reported that this variant did not influence the lipid levels or coronary heart disease risk in healthy UK men from the Northwick Park Heart Study II.26 This discrepancy can be partially explained by the characteristics of the study population such as gender and ethnicity and by differences in sample size among the studies. The PCSK9 E670G polymorphism is located within the cysteine-rich C-terminal domain, which is required for PCSK9 to remain bound to the LDLR during endosomal acidification and consequently to LDLR degradation. The positively charged C-terminal domain of PCSK9 appears to be involved in electrostatic interactions with negatively charged structures of the LDLR.32,33 It is possible that

the amino acid change causes a modification in the charge of the protein after the glutamic acid has a negative charge and the glycine has no charge. This change could increase the affinity of PCSK9 for the LDLR, promoting a reduction in the number of the receptors mediated by PCSK9, which can explain the high levels of LDL in plasma. However, the conformational and functional alteration induced by the E670G SNP in PCSK9 is still unknown. The PCSK9 I474V polymorphism was not associated with variability in the serum lipid profile in both the NL and HC groups. These findings are similar to that found by Chen et al in the LCAS population comprising 372 Caucasian subjects who had plasma LDL cholesterol levels of 115 to 190 mg/dL despite diet and 1 or more coronary lesions27 and by Kotowski et al in the Dallas Heart Study, in which SNP was not associated with plasma LDL cholesterol.8 In an association study using a large cohort representing the general population in Japan, the 474V allele was associated with low levels of total and LDL cholesterol both in male and female subjects. However, in the same study, the I474V polymorphism was not associated with the incidence of myocardial infarction.12 This population showed a significantly lower frequency of the 474V allele, which may explain the discrepancy in the results, besides differences in study population characteristics and study design.12 The function of the I474V polymorphism is not known, but it is possible that this variant is influenced by another polymorphism in tight linkage once isoleucine at amino acid number 474 was not conserved in either rats or mice.12 Recently, the effect of PCSK9 I474V variant on the protein structure was investigated using an array of in silico tools.34 A detailed bioinformatics analysis was carried out on the mutant protein, and the authors reported that this variant does not alter the overall tertiary structure of the protein with respect to wild type. Further local folding of PCSK9 in and around the site of mutation remains unperturbed and the site of the mutation is localized away from the binding site to the LDLR and probably does not alter the binding affinity of PCSK9 with LDLR.34 In this study, the PCSK9 polymorphisms E670G and I474V did not influence the response to atorvastatin of HC subjects. To our knowledge, this is the first study that evaluates the relationship of I474V polymorphism and response to statin treatment.

PCSK9 variants and atorvastatin response

263

Previous investigations reported that E670G variant had no impact on statin response. The Prospective Study of Pravastatin in the Elderly at Risk study could not reveal a significant difference in LDL cholesterol response to pravastatin treatment or coronary heart disease risk reduction between carriers and noncarriers of the E670G variant in an elderly population.35 Lack of association between this SNP and LDL cholesterol reduction in response to atorvastatin (80 mg/d) or pravastatin (40 mg/d) was also reported by the Pravastatin or Atorvastatin Evaluation and Infection Therapy–Thrombolysis in Myocardial Infarction 22 trial that enrolled 22 hypercholesterolemic patients with acute coronary syndrome within the preceding 10 days of hospitalization.36 Similarly, Chen et al showed that this polymorphism did not influence the response to fluvastatin treatment.27 In contrast, hypercholesterolemic subjects from the Utrecht Cardiovascular Pharmacogenetics study that carried the 670G allele had a better response to statin treatment; however, this study has enrolled patients taking different doses, different time of exposure to the statin therapy (participants were considered exposed when the cumulative defined daily doses of statin use was more than 180 days) and different types of statins what can explain the divergence in results.37 A possible limitation of this study is the number of individuals enrolled. Because the frequency of R46L polymorphism was very low, it was not possible to evaluate its influence on serum lipids and atorvastatin response. Therefore, a higher number of individuals would be necessary obtain more conclusive information about the influence of PCSK9 genotypes/haplotypes on serum lipids and atorvastatin response in the Brazilian population. Besides, in this study, we were not able to investigate the functional activity of these variants that could provide a better understand of the relationship of these polymorphisms and plasma lipids and statin response. Although relevant clinical variables were used to adjust the models of statistical analysis in this work, another limitation was that dietary intake was not considered among these covariates, which is an important variable regarding dyslipidemia and atorvastatin response.

from the National Council for Research and Development (CNPq), Brasilia, Brazil; M.H.H. and R.D.C.H. are recipients of fellowships from CNPq; A.C. was a recipient of a fellowship from CONICYT, Chile. We thank the volunteers for their participation in this study. Competing interests: The authors declare that they have no competing interests.

Anderson et al

Conclusion We report that PCSK9 E670G polymorphism is associated with high basal levels of LDL cholesterol in HC subjects, whereas the I474V variant has no influence on plasma lipid profile. Both PCSK9 variants have no impact on atorvastatin response in Brazilian hypercholesterolemic subjects.

Acknowledgments This work was supported by grants from Sao Paulo Research Foundation (FAPESP) (grant #2008-06667-9), Sao Paulo, Brazil. J.M.A. was the recipient of a fellowship

References 1. Ramasamy I. Recent advances in physiological lipoprotein metabolism. Clin Chem Lab Med. 2013;12:1–33. 2. Third report of the National Cholesterol Education Program (NCEP) Expert Panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III) final report. Circulation. 2002;25:3143–3421. 3. Teslovich TM, Musunuru K, Smith AV, et al. Biological, clinical and population relevance of 95 loci for blood lipids. Nature. 2010;7307: 707–713. 4. Soutar AK, Naoumova RP. Mechanisms of disease: genetic causes of familial hypercholesterolemia. Nat Clin Pract Cardiovasc Med. 2007; 4:214–225. 5. Lopez D. PCSK9: an enigmatic protease. Biochim Biophys Acta. 2008; 1781:184–191. 6. Zhang DW, Lagace TA, Garuti R, et al. Binding of proprotein convertase subtilisin/kexin type 9 to epidermal growth factor-like repeat A of low density lipoprotein receptor decreases receptor recycling and increases degradation. J Biol Chem. 2007;282:18602–18612. 7. Sun H, Samarghandi A, Zhang N, Yao Z, Xiong M, Teng BB. Proprotein convertase subtilisin/kexin type 9 interacts with apolipoprotein B and prevents its intracellular degradation, irrespective of the low- density lipoprotein receptor. Arterioscler Thromb Vasc Biol. 2012;32: 1585–1595. 8. Kotowski IK, Pertsemlidis A, Luke A, et al. A spectrum of PCSK9 alleles contributes to plasma levels of low-density lipoprotein cholesterol. Am J Hum Genet. 2006;78:410–422. 9. Evans D, Beil FU. The E670G SNP in the PCSK9 gene is associated with polygenic hypercholesterolemia in men but not in women. BMC Med Genet. 2006;7:66. 10. Aung LHH, Yin RX, Miao L, et al. The proprotein convertase subtilisin/kexin type 9 gene E670G polymorphism and serum lipid levels in the Guangxi Bai Ku Yao and Han populations. Lipids Health Dis. 2011;10:5. 11. Mayne J, Ooi TC, Raymond A, et al. Differential effects of PCSK9 loss of function variants on serum lipid and PCSK9 levels in Caucasian and African Canadian populations. Lipids Health Dis. 2013;12:70. 12. Shioji K, Mannami T, Kokubo Y, et al. Genetic variants in PCSK9 affect the cholesterol level in Japanese. J Hum Genet. 2004;49: 109–114. 13. Huang CC, Fornage M, Lloyd-Jones DM, Wei GS, Boerwinkle E, Liu K. Longitudinal association of PCSK9 sequence variations with low-density lipoprotein cholesterol levels: the Coronay Artery Risk Development in Young Adults Study. Circ Cardiovasc Genet. 2009; 2:354–361. 14. Zhou Q, Liao J. Statins and cardiovascular diseases: from cholesterol lowering to pleiotropy. Curr Pharm Des. 2009;15:467–478. 15. Minder CM, Blumenthal RS, Blaha MJ. Statins for primary prevention of cardiovascular disease: the benefits outweigh the risks. Curr Opin Cardiol. 2013;28:554–560. 16. Rodrigues AC, Hirata MH, Hirata RD. The genetic determinants of atorvastatin response. Curr Opin Mol Ther. 2007;9:545–553. 17. Postmus I, Verschuren JJ, de Craen AJ, et al. Pharmacogenetics of statins: achievements, whole-genome analyses and future perspectives. Pharmacogenomics. 2012;13:831–840.

264 18. Cerda A, Genvigir FD, Arazi SS, et al. Influence of SCARB1 polymorphism on serum lipids of hypercholesterolemic individuals treated with atorvastatin. Clin Chim Acta. 2010;411:631–637. 19. Rodrigues AC, Perin PM, Purim SG, et al. Pharmacogenetics of OATP transporters reveals that SLCO1B1 c.388A.G variant is determinant of increased atorvastatin response. Int J Mol Sci. 2011;12:5815–5827. 20. Berge KE, Ose L, Leren TP. Missense mutations in the PCSK9 gene are associated with hypocholesterolemia and possibly increased response to statin therapy. Arterioscler Thromb Vasc Biol. 2006;26: 1094–1100. 21. Ding K, Kullo IJ. Molecular population genetics of PCSK9: a signature of recent positive selection. Pharmacogenet Genomics. 2008;18: 169–179. 22. Naoumova RP, Tosi I, Patel D, et al. Severe hypercholesterolemia in four British families with the D374Y mutation in the PCSK9 gene: long-term follow-up and treatment response. Arterioscler Thromb Vasc Biol. 2005;25:2654–2660. 23. Sposito AC, Caramelli B, Fonseca FA, et al. IV Brazilian guideline for dyslipidemia and atherosclerosis prevention: Department of Atherosclerosis of Brazilian Society of Cardiology. Arq Bras Cardiol. 2007;88(Suppl 1):2–19. 24. Cerda A, Genvigir FD, Willrich MA, et al. Apolipoprotein E mRNA expression in mononuclear cells from normolipidemic and hypercholesterolemic individuals treated with atorvastatin. Lipids Health Dis. 2011;10:206. 25. Salazar LA, Hirata MH, Cavalli SA, Machado MO, Hirata RDC. Optimized procedure for DNA isolation from fresh and cryopreserved clotted human blood useful in clinical molecular testing. Clin Chem. 1998; 44:1748–1750. 26. Scartezini M, Hubbart C, Whittall RA, Cooper JA, Neil AH, Humphries SE. The PCSK9 gene R46L variant is associated with lower plasma lipid levels and cardiovascular risk in healthy U.K. men. Clin Sci (Lond). 2007;113:435–441. 27. Chen SN, Ballantyne CM, Gotto AM Jr., Tan Y, Willerson JT, Marian AJ. A common PCSK9 haplotype, encompassing the E670G coding single nucleotide polymorphism, is a novel genetic marker

Journal of Clinical Lipidology, Vol 8, No 3, June 2014

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

for plasma low-density lipoprotein cholesterol levels and severity of coronary atherosclerosis. J Am Coll Cardiol. 2005;45:1611–1619. Suarez-Kurtz G, Pena SD, Struchiner CJ, Hutz MH. Pharmacogenomic diversity among Brazilians: influence of ancestry, selfreported color, and geographical origin. Front Pharmacol. 2012;3:191. Pena SD, Di Pietro G, Fuchshuber-Moraes M, et al. The genomic ancestry of individuals from different geographical regions of Brazil is more uniform than expected. PLoS One. 2011;6:e17063. Silbiger VN, Hirata MH, Luchessi AD, et al. Differentiation of African components of ancestry to stratify groups in a case-control study of a Brazilian urban population. Genet Test Mol Biomarkers. 2012;16: 524–530. Rodrigues AC, Sobrino B, Genvigir FD, et al. Genetic variants in genes related to lipid metabolism and atherosclerosis, dyslipidemia and atorvastatin response. Clin Chim Acta. 2013;417:8–11. Tveten K, Holla ØL, Cameron J, et al. Interaction between the ligandbinding domain of the LDL receptor and the C-terminal domain of PCSK9 is required for PCSK9 to remain bound to the LDL receptor during endosomal acidification. Hum Mol Genet. 2012;21:1402–1409. Holla ØL, Cameron J, Tveten K, et al. Role of the C-terminal domain of PCSK9 in degradation of the LDL receptors. J Lipid Res. 2011;52: 1787–1794. Al-Waili K, Al-Zidi WA, Al-Abri AR, et al. Mutation in the PCSK9 gene in Omani Arab subjects with autosomal dominant hypercholesterolemia and its effect on PCSK9 protein structure. Oman Med J. 2013;28:48–52. Polisecki E, Peter I, Robertson M, et al. PROSPER Study Group. Genetic variation at the PCSK9 locus moderately lowers low-density lipoprotein cholesterol levels, but does not significantly lower vascular disease risk in an elderly population. Atherosclerosis. 2008;200:95–101. Mega JL, Morrow DA, Brown A, Cannon CP, Sabatine MS. Identification of genetic variants associated with response to statin therapy. Arterioscler Thromb Vasc Biol. 2009;29:1310–1315. Peters BJ, Pett H, Klungel OH, et al. Genetic variability within the cholesterol lowering pathway and the effectiveness of statins in reducing the risk of MI. Atherosclerosis. 2011;217:458–464.