Effects of atorvastatin on CYP3A4 and CYP3A5 mRNA expression in mononuclear cells and CYP3A activity in hypercholeresterolemic patients

Clinica Chimica Acta 421 (2013) 157–163

Contents lists available at SciVerse ScienceDirect

Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim

Effects of atorvastatin on CYP3A4 and CYP3A5 mRNA expression in mononuclear cells and CYP3A activity in hypercholeresterolemic patients Maria Alice V. Willrich a,⁎, 1, Alice C. Rodrigues a, Alvaro Cerda a, Fabiana D.V. Genvigir a, Simone S. Arazi a, Egidio L. Dorea b, Marcia M.S. Bernik b, Marcelo C. Bertolami c, Andre Faludi c, Alvaro Largura d, Linnea M. Baudhuin e, Sandra C. Bryant f, Mario Hiroyuki Hirata a, Rosario Dominguez Crespo Hirata a a

Department of Clinical and Toxicological Analyses School of Pharmaceutical Sciences University of Sao Paulo Sao Paulo SP Brazil University Hospital University of Sao Paulo Sao Paulo SP Brazil c Dante Pazzanezze Institute of Cardiology Sao Paulo SP Brazil d Paranalise Laboratory Curitiba PR Brazil e Department of Laboratory Medicine and Pathology Mayo Clinic Rochester MN USA f Division of Biomedical Statistics and Informatics Mayo Clinic Rochester MN USA b

a r t i c l e

i n f o

Article history: Received 10 December 2012 Received in revised form 31 January 2013 Accepted 6 March 2013 Available online 15 March 2013 Keywords: Pharmacogenomics Atorvastatin CYP3A4 CYP3A5 mRNA expression CYP3A activity

a b s t r a c t Background: Variability of response to statins has been related to polymorphisms in genes involved in cholesterol homeostasis and statin metabolism, such as CYP3A4 and CYP3A5. We investigated the effects of atorvastatin on CYP3A4 and CYP3A5 mRNA expression in mononuclear cells and on CYP3A activity and their interactions with common variants. Methods: Unrelated individuals (n = 121) with hypercholesterolemia (HC) were treated with atorvastatin (10 mg/day/4 weeks). Ninety-two normolipidemic (NL) subjects were selected as a control group. Genotype analysis of CYP3A4*1B (rs2740574), CYP3A4*22 (rs35599367), CYP3A5*3C (rs776746), and CYP3A5*1D (rs15524) and mRNA levels in peripheral blood mononuclear cells (PBMCs) were estimated. CYP3A activity was phenotyped by the urinary cortisol to 6-beta-hydroxy-cortisol ratio. Results: LDL cholesterol reduction in response to atorvastatin was positively correlated with change in CYP3A4 (R2 = 0.039, p = 0.037) and CYP3A5 (R2 = 0.047, p = 0.019) mRNA levels and negatively correlated with CYP3A activity (R2 = 0.071, p = 0.022). CYP3A5*3C (AGT haplotype) was associated to lower basal CYP3A5 mRNA expression in HC (p b 0.045), however none of the haplotype groups impacted treatment. Conclusion: It is likely that cholesterolemia status changes promoted by atorvastatin play a role in regulating CYP3A4 and CYP3A5 mRNA expression in PBMCs, as well as CYP3A activity. CYP3A5*3C (AGT haplotype) also contributes for the variability of CYP3A5 mRNA levels in PBMCs. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Statins are reversible inhibitors of 3-hydroxy-3-methyl-glutaryl-CoA reductase, which catalyzes the rate-limiting step in the de novo biosynthesis of cholesterol by acetyl-CoA [1]. Reduction of intracellular cholesterol subsequently induces the expression of the low-density lipoprotein (LDL) receptor, resulting in increased LDL uptake by hepatic and peripheral cells, thereby dramatically reducing circulating cholesterol levels [2]. Pleiotropic effects include improving endothelial function, decreasing

⁎ Corresponding author at: Department of Clinical and Toxicological Analyses, School of Pharmaceutical Sciences, University of Sao Paulo, Av. Prof. Lineu Prestes, 580, B17, 05508-900 Sao Paulo, SP, Brazil. Tel.: + 55 11 3091 3660. E-mail address: [email protected] (M.A.V. Willrich). 1 Present address: Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, USA, 200 First Street SW 55905. 0009-8981/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cca.2013.03.007

vascular inflammation and enhancing plaque stability [3]. Statin use contributes to an approximate 30% reduction in cardiovascular disease (CVD) [4], and statins are one of the most commonly prescribed drugs worldwide. Atorvastatin is administered orally as an active acid form and is metabolized primarily by cytochrome P450 3A4 (CYP3A4) and CYP3A5 in the gut and liver, and it is further conjugated by uridine diphosphate glucuronosyl transferases 1A1 and 1A3 [5]. The uptake and efflux membrane drug transporters expressed in liver and intestinal cells contribute to the overall bioavailability of the atorvastatin [5–7]. CYP3A4 and CYP3A5 are important drug metabolizing enzymes due to their abundance in the intestinal epithelium and liver, and their ability to metabolize a large number of chemically unrelated drugs [8]. Inter-individual variability in CYP3A4 and CYP3A5 liver expression is very high (10- to 100-fold). This variability has been suggested to be associated with variants in the genes encoding CYP3A4 and CYP3A5 and

158

M.A.V. Willrich et al. / Clinica Chimica Acta 421 (2013) 157–163

may contribute to the differences found in response to several drugs, including atorvastatin and other statins [9,10]. There is also great interindividual response to statin treatment. It is thought that genetics may play an important role in the biodisposition and safety of statins [11]. Understanding and addressing the pharmacogenomic variability of statins has been challenging and numerous efforts have been undertaken to discover a surrogate marker for statin response to treatment. CYP3A4 and CYP3A5 expression and activity in the liver are major determinants of drug efficacy, toxicity, and hence, therapeutic outcome [10]. Hepatocytes are the main site of action for statins. Increased CYP3A mRNA expression and CYP3A protein levels in hepatocytes have been demonstrated to be increased by statins [12], ranging from 6 to 20 fold the basal level [13,14]. Moreover, CYP3A4 mRNA expression was increased when HepG2 cells were incubated with atorvastatin, as was CYP3A5, though to a lesser extent [15]. Analysis of mRNA expression in peripheral blood mononuclear cells (PBMCs) has been used to evaluate the effects of lipid-lowering drugs on peripheral tissues in vivo [16,17]. Considering their involvement in pathophysiological mechanisms of cardiovascular-related diseases, mainly through the inflammatory pathways, expression in mononuclear cells could be used to evaluate modification of expression observed with this system [18]. In this study, we investigated the effects of atorvastatin on PBMCs CYP3A4 and CYP3A5 mRNA expression and CYP3A activity in vivo, measured by 6-beta-hydroxy-cortisol/cortisol ratio, and evaluated the interaction with common genetic variants of CYP3A4 and CYP3A5 in hypercholesterolemic individuals. 2. Subjects and methods 2.1. Subjects and study protocol The characteristics of this study design and protocol have been previously reported [16,19,20]. Briefly, 458 individuals were selected to participate in the study at two clinical research centers in Sao Paulo, Brazil. Participants were screened and determined to be hypercholesterolemic (HC) with LDL cholesterol levels higher than 160 mg/dL (n = 210) or normolipidemic (NL; n = 248). Participants with hypertriglyceridemia (triglycerides ≥ 400 mg/dL), hypothyroidism, diabetes mellitus, liver or kidney diseases, and other forms of secondary dyslipidemia were excluded from this study. Overall 121 HC and 92 NL individuals consented and completed the study protocol. HC patients went through a low fat diet (according to American Heart Association Step One Diet) for four weeks [21]. Individuals who still had LDL cholesterol greater than 160 mg/dL were treated with atorvastatin 10 mg orally once daily for four weeks. HC patients were evaluated for serum concentrations of lipids to assess response to atorvastatin treatment, and alanine aminotransferase (ALT) and creatine kinase (CK) to monitor liver and muscle toxicity side effects. Patients continued taking other medications throughout the study, including diuretics, angiotensin converting enzyme (ACE) inhibitors, beta-adrenergic blockers, and calcium channel blockers. Information on age, body mass index (BMI), blood pressure, menopause status, hypertension, obesity, family history of cardiovascular disease (CVD), tobacco smoking, alcohol consumption, exercise and current medications were recorded. The study protocol was approved by the Ethics Committees of the Dante Pazzanese Institute of Cardiology, University Hospital and School of Pharmaceutical Sciences at the University of Sao Paulo, Sao Paulo, Brazil. 2.2. Samples Blood samples were collected after a 12 hour fast. Patients undergoing atorvastatin treatment had blood drawn prior to and after the

4 week treatment. Serum total cholesterol, high-density lipoprotein (HDL) cholesterol and triglycerides were measured using routine enzymatic methods. CK and ALT were measured by kinetic methods. LDL and very low-density lipoprotein (VLDL) cholesterol were calculated [22]. Plasma apolipoprotein AI (apoAI) and apolipoprotein B (apoB) were determined by nephelometry. Morning urine samples were collected. Urinary creatinine was measured by the Jaffe method. An aliquot was frozen at − 80 °C for 6-beta-hydroxy-cortisol (6BOHC) and cortisol detection by high performance liquid chromatography (HPLC). 2.3. CYP3A4 and CYP3A5 genotyping Genomic DNA was isolated from 1 mL EDTA-anticoagulated whole blood samples using a salting-out method [23]. CYP3A4*1B (rs2740574) was detected by polymerase chain reaction (PCR) followed by restriction fragment analysis (RFLP) as previously published [24]. CYP3A5*3C (rs776746) and CYP3A4*22 (rs35599367) were detected using TaqMan Genotyping Assays [Life Technologies (Foster City, CA), catalog numbers C__26201809_30 and C_593013445_10]. Each PCR reaction contained 4 μL of Universal Master Mix (2×) (Life Technologies), 0.4 μL of the specific Genotyping Assay (20×), and 20 ng of DNA. The real-time PCR assays were carried out in 96-well plates using a 7500 Real-Time PCR system (Applied Biosystems, Foster City, CA, USA). The thermal cycling protocol consisted of an initial cycle of 10 min at 95 °C followed by 40 cycles at 92 °C for 15 s, 60 °C for 1 min using standard 7500 conditions. CYP3A5*1D (rs15524) was detected by DNA sequencing. Primer sequences and PCR thermal cycling conditions were previously published [20]. For quality control purposes, 5% of all samples were repeated and genotype accuracy was confirmed by DNA sequencing using the capillary electrophoresis system Mega BACE 1000 (Amersham Pharmacia Biotech, Upsala, Sweden) after PCR amplification using sets of primers designed with Primer Premier® v. 5.0 (Premier Biosoft International, USA). 2.4. RNA extraction and CYP3A4/CYP3A5 mRNA expression by real-time PCR EDTA-anticoagulated blood was diluted in phosphate-buffered saline (1:1) and this suspension was layered in Histopaque-1077 (Sigma-Aldrich, St. Louis, MO, USA) and centrifuged for 30 min at 400 g at room temperature. PBMCs were collected from the interphase and immediately underwent RNA extraction using TRIzol® reagent (Invitrogen Corporation, CA, USA) according to the manufacturer's instructions. Complementary DNA (cDNA) was produced from 1 μg of total RNA by SuperScriptTM II Reverse Transcriptase RNase H- with 10 μM random primers (Invitrogen Corporation, CA, USA) and CYP3A4 and CYP3A5 mRNA were measured by a TaqMan quantitative PCR assay. The reference gene was chosen after analysis using the software geNorm [25]. Five reference genes were tested [hypoxanthine phosphoribosyl transferase I (HRPT1), hydroxymethyl-bilane synthase (HMBS), beta-2-microglobulin (B2M), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and ubiquitin C (UBC)]. The most stable gene for PBMCs under experimental conditions was GAPDH, so it was used as a reference gene. The real-time PCR assays were carried out in 96-well plates using a 7500 Real-Time PCR system (Applied Biosystems, Foster City, CA, USA). CYP3A4 primer sequences were: F: TCAATAACAGTC TTTCCATTCCTCAT and R: TTCGAGGCGACTTTCTTTCATC, and the probe sequence was FAMTM TGTTTCCAAGAGAAGTTACMGBNFQ. For CYP3A5, primers were F: CTATCGTCAGGGTCTCTGGAAATT and R: ACGTTCCC CACATTTTTCCATA and the probe sequence was FAMTMACACAGAGT GCTATAAAAMGBNFQ. For GAPDH, primers were F: GGAAGGTGAAG GTCGGAGTCA, R: CTGGAAGATGGTGATGGGATTTC and the probe was VIC®TCAGCCTTGACGGTGC MGBNFQ. The thermal cycler protocol

M.A.V. Willrich et al. / Clinica Chimica Acta 421 (2013) 157–163

consisted of 40 cycles at 95 °C for 15 s and 60 °C for 1 min. Efficiency of all real-time PCR assays was considered adequate if between 90% and 110%, established by a standard curve of different cDNA dilutions and their slopes (Efficiency = 10 (−1/slope) − 1). Samples were tested in duplicate. All reaction plates were tested with no template controls in order to evaluate possible contamination of the reagents with PCR products. The relative expression of the target gene was analyzed using a comparative ΔCT method [26]. 2.5. CYP3A activity assay: 6-beta-hydroxy-cortisol/cortisol ratio Cortisol to 6BOHC ratio was used as a marker to measure CYP3A activity. Samples were collected any time from 6 am to 9 am (in order to minimize the effect of the circadian rhythm on urinary excretion of the analytes). Samples were evaluated for quality using a urinary creatinine assay (NL, 101 ± 60; HC at baseline 78 ± 55; HC after treatment 66 ± 47 mg/dL). Cortisol and 6BOHC were measured by HPLC after liquid–liquid extraction of the urine samples based on a previously reported protocol [27]. Briefly, all reagents used for testing were HPLC-grade and were purchased from Merck (Merck, Sao Paulo, Brazil). The standards for cortisol, 6BOHC, and the internal standard (dexamethasone) were purchased from Sigma Aldrich (Sao Paulo, Brazil) and were of at least 98% purity. The standard solutions were prepared in HPLC-grade methanol at 1 mg/mL. The solutions were stable for up to 3 months when stored at − 20 °C. 200 μL of dexamethasone [1 mg/mL] was added to 500 μL of urine samples and homogenized by vortexing. Subsequently, 5 mL of dichloromethane were added. After mixing vigorously for 1 min, the solution was centrifuged for 5 min at 3200 g. The organic phase was removed and the solution evaporated at 40 °C under a nitrogen stream. Then, the sample was reconstituted with 200 μL of methanol and water (60:40, v/v). Assays were carried out using a LC-2010 HT gas chromatographer (Shimadzu Brazil, Sao Paulo, Brazil) with Luna 5 mm C18 100A 150 × 4.6 mm columns (Phenomenex Inc, Torrance, CA, USA). The chromatography was performed at room temperature with 60 μL of sample, an injection flow rate of 1.0 mL/min, and a mobile phase containing methanol and water (60:40, v/v). Detection was set at UV absorbance (245 nm) and results were expressed as 6BOHC/cortisol ratio. Analytical recovery was carried out by adding 100 ng of cortisol and 1000 ng 6BOHC in 500 μL of 1:50 diluted urine before the extraction procedure (n = 6). Creatinine on the diluted urine was 1.6 mg/dL and before the sample was spiked with the analytes of interest, it did not yield any detectable peaks on the chromatogram. The recovery was 92% ± 8% for dexamethasone (mean ± SD), 85% ± 7% for cortisol and 81% ± 12% for 6BOHC. The limit of quantitation (LOQ) was 1 μg/L for cortisol and 1 mg/L for 6BOHC. The limit of detection (LOD) was 1 μg/L for cortisol and 100 μg/L for 6BOHC. The estimated accuracy was 105% for cortisol and 78% for 6BOHC. The linearity range was 1–400 μg/L (r2 = 0.9932) for cortisol and 1–500 mg/L (r2 = 0.9931) for 6BOHC. The internal standard was used as a pre-analytical control and its peak area was used to assess the variability of the process. Runs were accepted when recovery of the internal standard was greater than 85%. Peaks were quantitated using 6BOHC/cortisol peak area ratio. 2.6. Statistical analysis Categorical variables were compared by the Fisher's exact test. The quality of the CYP3A4*1B, CYP3A5*3C, CYP3A5*1D genotypes were assessed using the Pearson's χ 2 test of deviation from Hardy– Weinberg equilibrium of SNPs of control (or case or both) subjects. Minor allele frequencies were compared between NL and HC subjects using the Fisher's exact test. Pair-wise linkage disequilibrium (LD) among SNPs was measured using the D′ index and r 2 values. Haplotype frequencies, overall haplotype associations and pair-wise haplotype associations compared to the wild-type haplotype (defined

159

as AGT) were assessed using the HaploStats analysis package in R [28,29]. Linear regression analysis was performed to examine associations between haplotypes and mRNA expression and activity both adjusted for possible covariates and unadjusted. Continuous variables were presented as mean ± standard deviation (SD) or median (range), as appropriate. NL and HC comparisons were analyzed by t-test or Wilcoxon rank sum test, as appropriate. Comparisons of continuous variables before and after atorvastatin treatment were analyzed by paired Wilcoxon signed rank test. Reduction in serum lipids in response to atorvastatin treatment were expressed as the value after treatment − the value before treatment. Associations among mRNA expression, activity, and individual SNPs were assessed using univariate linear regression on the log-transformed variables. Univariate general linear regression was used to determine possible covariates for and associations with LDL-c reduction. CYP3A4 and CYP3A5 expression can be considered a complex trait under multifactorial control [10], so multiple linear regression analysis was used to determine which variables could affect or predict mRNA expression. From those variables with univariate associations with mRNA expression, CYP3A activity and response to atorvastatin treatment, multiple linear regression models were used for assessing additional genotype effects and covariates using SAS® software (version 9.2, SAS Institute Inc.). Significance was defined as p-value b 0.05. 3. Results The characteristics of the HC and NL groups are described in Table 1. Both groups had the same proportion of female and male subjects enrolled (p = 0.297). NL subjects (n = 91) were younger than the HC subjects (n = 119) (about 10 years in average (p b 0.0001)). In both groups, Caucasians represented the majority (p = 0.552). BMI and obesity were higher in HC individuals (both p b 0.017). NL and HC subjects had approximately the same rate of hypertension (p = 0.091), but family history of CVD was higher in HC (p = 0.036). Exercise, tobacco and alcohol consumption were similar between both groups (p > 0.47). Since the HC group is older than the NL group, the menopause status in women from the HC group was also higher (p b 0.0001). NL and HC subjects had similar profiles of concomitant medication usage, such as CYP3A inhibitors and substrates (p > 0.16, Table 1). HC subjects had higher levels of total and LDL cholesterol, triglycerides, and apoB, and lower levels of apoAI than NL (p b 0.0001). HDLcholesterol levels were similar between the two groups (p = 0.56). Atorvastatin significantly reduced total, LDL cholesterol, apoB, and triglycerides levels (p b 0.0001) (Table 1), and lowered high-density lipoprotein (HDL) cholesterol levels. In addition, atorvastatin treatment did not cause a significant increase in CK or ALT levels (data not shown). Univariate general linear regression analysis showed that LDL-cholesterol reduction in response to atorvastatin was not influenced by age, gender, ethnicity, BMI, hypertension, CVD history, menopause status, and concomitant medications (CYP3A substrates or inhibitors) (p > 0.14, Table 2). On the other hand, exercise explained 4.2% of the LDL cholesterol reduction (Table 2). CYP3A4 (Fig. 1A) and CYP3A5 (Fig. 1B) mRNA baseline levels in PBMCs were lower in HC than in NL groups (p b 0.0001). For HC subjects, CYP3A4 and CYP3A5 mRNA expression was significantly correlated before (R2 = 0.22, p b 0.0001) and after atorvastatin treatment (R2 = 0.58, p b 0.0001), as determined by linear regression. Baseline levels of CYP3A4 and CYP3A5 mRNA expression levels were not influenced by BMI, ethnicity, gender, hypertension, CYP3A concomitant substrates or inhibitors (p > 0.089) (data not shown). However, both CYP3A4 and CYP3A5 mRNA expression levels at baseline were inversely associated with exercise (p b 0.017), while only CYP3A5 mRNA expression was inversely associated with family history of CVD (p = 0.025, data not shown). Average mRNA levels of CYP3A4 and CYP3A5 were not influenced after atorvastatin in HC individuals (p > 0.169) (Fig. 1), nevertheless

160

M.A.V. Willrich et al. / Clinica Chimica Acta 421 (2013) 157–163

Table 1 Data of hypercholesterolemic and normolipidemic individuals. Variable

NL (91)

HC basal (119)

P value1

HC Atorva (119)

Age, years Men Caucasian Body mass index, kg/m2 Obesity Hypertension Family history of CVD Exercise Tobacco smoking Alcohol comsumption Menopause status

47 ± 8 28% (25) 64% (55) 26 [18,39] 11% (10) 47% (42) 43% (38) 51% (45) 17% (15) 30% (27) 32% (17)

57 ± 11 35% (42) 68% (81) 27 [17,38] 27% (32) 59% (69) 58% (68) 46% (48) 22% (23) 31% (33) 85% (53)

b0.0001a 0.297c 0.552c 0.017 b 0.008c 0.091c 0.036 c 0.566 c 0.469 c 1.000 c b0.0001c

– – – – – – – – – – –

Concomitant medication CYP3A inhibitors CYP3A substrates CYP3A inducers None

14% (12) 16% (13) 0% (0) 78% (66)

8% (8) 17% (18) 0% (0) 80% (82)

0.160c 0.844c 1.000 c 1.000 c

– – – –

Lipid profile, mg/dL Total cholesterol Triglycerides HDL cholesterol LDL cholesterol Apolipoprotein AI Apolipoprotein B

174 [130,200] 82 [39, 151] 54 [36, 91] 101 [45, 131] 141.5 [68, 207] 76 [32, 129]

273 [218, 414] 143 [50, 379] 54 [30, 99] 181 [144, 319] 126.5 [40, 218] 136 [67, 198]

Minor allele frequency CYP3A4*1B rs2740574 [G] CYP3A4*22 rs35599367 [T] CYP3A5*3C rs776746 [A] CYP3A5*1D rs15524 [C]

33% 3% 33% 30%

20% 1% 29% 26%

b0.0001b b0.0001b 0.560b b0.0001b 0.0001b b0.0001b

P value2

b0.0001d 0.0001d 0.0002d b0.0001d 0.067d b0.0001d

196 [120, 268] 126 [34, 381] 52 [27, 94] 114 [67, 188] 128 [81, 230] 95.5 [59, 151]

0.049c 0.621c 0.538c 0.367c

– – –

Number of individuals is in parenthesis. Obesity was considered as BMI ≥ 30 kg/m2. Hypertension was blood pressure higher than 140/90 mm Hg or treatment with anti-hypertensive drugs [41]. Exercise was 30 min of aerobic activity at least three times a week. Current tobacco smoking was a daily intake of one or more cigarettes. Alcohol consumption was defined as one dose a day or less. A standard alcohol dose contains approximately 10 g of alcohol and its volume varies according to the beverage (350 mL of beer, 150 mL of wine or 50 mL of distilled spirits [42]. CVD: cardiovascular disease; HDL: high-density lipoprotein; LDL: low-density lipoprotein. Continuous variables are shown as median and interquartile range, compared by t-test (a) or Wilcoxon rank sum test (b). Categorical variables were compared by Fisher Exact test (c). Wilcoxon Signed Rank Sum test (d). 1 p-value refers to comparison between NL and HC basal groups. 2 p-value refers to comparison between HC basal and treatment with atorvastatin.

the linear regression analysis demonstrated that an increase in CYP3A4 and CYP3A5 mRNA expression after treatment is related to a greater LDL cholesterol reduction in response to atorvastatin (Table 2, p b 0.037). These variables explain, respectively, 3.9% and 4.7% of the LDL cholesterol reduction. Variation in CYP3A5 mRNA expression remained significantly associated with LDL reduction, even after adjusting for exercise or change in 6BOHC/cortisol (p b 0.043) (data not shown). None of the variables tested (BMI, hypertension, CVD history, gender, age, ethnicity, concomitant medication or menopause status) affected urinary 6BOHC/cortisol ratio before (p > 0.081) and after treatment (p > 0.10) (data not shown). 6BOHC/cortisol ratio showed lower values in HC subjects than in the NL group (p b 0.0001) (Fig. 2) and was influenced by age (p = 0.049), but not by other variables (p > 0.081, data not shown). There was no correlation between 6BOHC/cortisol ratio and PBMC CYP3A4 or CYP3A5 mRNA levels in NL or HC subjects in this study (p > 0.331, data not shown). Even though atorvastatin treatment did not modify 6BOHC/cortisol ratio in HC subjects (Fig. 2), univariate linear regression showed that LDL cholesterol reduction in response to atorvastatin was negatively correlated and partly explained (7.1%) by 6BOHC/cortisol ratio change (p = 0.022, Table 2) even after adjustment for exercise (data not shown). Genotype distribution of the studied polymorphisms were in HWE (p > 0.158), except for CYP3A4*1B (p = 0.006). The minor allele frequency (MAF) of the CYP3A4*1B allele was slightly higher in NL (33%) than in HC (20%, p = 0.049), while CYP3A4*22, CYP3A5*3C, and

Table 2 Univariate linear regression analysis of variables associated with LDL cholesterol reduction in response to atorvastatin. Variables

N

Age, years Men Caucasian Hypertension (mm Hg) Obesity (BMI > 30 kg/m2) Body mass index Family history of CVD Exercise Tobacco smoking Alcohol consumption Menopause status CYP3A inhibitors CYP3A substrates CYP3A4 mRNA change [2−DDCt] CYP3A5 mRNA change [2−DDCt] 6BOHC/Cortisol ratio change (*) CYP3A4*1B[ref genotype AA] GG AG CYP3A5*3C[ref genotype GG] AA AG CYP3A5*1D[ref genotype TT] CC CT

119 119 119 116 117 117 118 104 106 106 62 103 103 113 115 74 119

Slope (SE) −0.149 (0.251) 8.145 (5.822) 2.606 (6.013) −4.137 (5.687) −0.272 (6.322) 0.266 (0.656) −8.455 (5.672) 12.679 (5.967) 10.346 (7.233) 5.040 (6.483) 3.457 (11.457) 5.728 (11.366) −6.776 (7.992) 0.541 (0.256) 1.269 (0.535) −0.077 (0.033) 2.720 (10.273) −4.169 (6.805)

119 9.150 (9.461) −11.321 (5.850) 119 −10.332 (5.997) 9.777 (9.781)

p-value

R2

0.554 0.164 0.666 0.468 0.966 0.686 0.139 0.036 0.156 0.439 0.764 0.615 0.398 0.037 0.019 0.022 0.774 0.792 0.541 0.051 0.336 0.055 0.086 0.088 0.320

0.003 0.016 0.002 0.005 b0.0001 0.001 0.019 0.042 0.019 0.006 0.002 0.003 0.007 0.039 0.047 0.071 0.004

0.05

0.04

Notes: CVD: cardiovascular disease; 6BOHC: 6-beta-hydroxy-cortisol; SE: standard error of slope. (*) Calculated as ratio after atorvastatin minus ratio at baseline.

M.A.V. Willrich et al. / Clinica Chimica Acta 421 (2013) 157–163

161

Fig. 1. CYP3A4 (A) and CYP3A5 (B) mRNA levels for NL subjects and HC before and after atorvastatin treatment (10 mg/day/4 weeks). NL, normolipidemic subjects. HC, hypercholesterolemic subjects. Mean is represented as a cross (+), median as a hyphen (−) in the jitterplot. mRNA expression of each gene was measured by qPCR and normalized with GAPDH. Y axis is on log scale. The 2−DCt formula was used to calculate the relative amount of transcript in the sample.

CYP3A5*1D had similar frequencies between these cohorts (p > 0.37) (Table 1). Differences in CYP3A4 and CYP3A5 allele frequencies were observed between Caucasians (whites) (MAF: CYP3A4*1B [G], 0.08; CYP3A4*22 [T], 0.03; CYP3A5*3C [A], 0.15 and CYP3A5*1D [C], 0.16) and Afrodescents (MAF: CYP3A4*1B [G], 0.55; CYP3A4*22 [T], 0.00; CYP3A5*3C [A], 0.60 and CYP3A5*1D [C], 0.50) (p b 0.0001). However, this variable is based in skin color, which does not correlate well with ancestry in populations with a large genetic diversity such as ours [30–32]. Linkage disequilibrium was tested for CYP3A4 and CYP3A5 variants. Association was found for CYP3A4*1B and CYP3A5*3C, D′ = 0.7202 (p b 0.0001, r2 = 0.37). An even stronger association was observed for CYP3A5*3C and CYP3A5*1D, D′ = 0.9136 (p b 0.0001, r2 = 0.73). No linkage disequilibrium was found between CYP3A4*22 and the other variants (D′ = − 1.000 (p > 0.0611). CYP3A4*1B and CYP3A5*1D genotypes were not associated with LDL cholesterol reduction in response to atorvastatin (Table 2). The CYP3A5*3C reference genotype GG was very close to being significant (p = 0.051) and alone explained 5% of the LDL cholesterol reduction. Haplotype groups were formed based on the number of copies of each variant: CYP3A4*1B (A > G), CYP3A5*3C (A > G) and CYP3A5*1D

(C > T), respectively, with AGT as the reference, since it was the most frequent haplotype in our sample. CYP3A4*22 was not included for this analysis due to its low frequency. The frequencies of the haplotypes AGT, AAC, GAC, GAT, GGT and rare (AAT and AGC: less than 2%) were similar between NL and HC groups (p = 0.1460, data not shown). Univariate association analysis was assessed between CYP3A4 and CYP3A5 haplotypes and other variables in HC subjects. Ethnicity was the most strongly associated (p b 0.0001), as shown by allelic analyses. The relationship between CYP3A4/3A5 haplotypes and mRNA expression was carried out after adjusting for covariates (including but not limited to gender, ethnicity, hypertension, CVD history, and BMI) using the HaploStats package in R [28]. All HC patients had posterior probabilities of the assigned haplotype > 96%. CYP3A4 mRNA expression at baseline or post-treatment was not influenced by CYP3A4 and CYP3A5 haplotype in HC individuals (p = 0.887, data not shown). Interestingly, GAC and GAT haplotypes were associated with higher CYP3A5 mRNA basal levels in PBMCs compared with AGT haplotype in the HC group (pair-wise p = 0.0003 and p = 0.045, respectively, Fig. 3A). This relationship did not remain after treatment with atorvastatin (p = 0.919, Fig. 3B). CYP3A4 and CYP3A5 haplotypes were not related to baseline or post-treatment levels of urinary 6BOHC/cortisol ratio, even when adjusted for covariates (data not shown). 4. Discussion

Fig. 2. Urinary 6BOHC/cortisol ratio in NL and HC individuals before and after atorvastatin treatment (10 mg/day/4 weeks). Notes: NL, normolipidemic subjects. HC, hypercholesterolemic subjects. 6BOHC, 6-beta-hydroxy-cortisol. Mean is represented as a cross (+), median as a hyphen (−) in the jitterplot. Cortisol and 6BOHC were measured in morning urine samples by HPLC-UV. Y axis is on log scale.

Our observations demonstrated that LDL reduction in response to atorvastatin is explained by exercise as well as variation in CYP3A4 and CYP3A5 mRNA expression and CYP3A activity. CYP3A4 SNPs were not significantly associated to atorvastatin response. It is well established that an active lifestyle reduces cardiac risk and decreases lipid levels. Exercising regularly led to an additional reduction of 13% in LDL-c levels in our HC subject cohort. Previous reports regarding the impact of CYP3A4 and CYP3A5 mRNA expression and activity on atorvastatin treatment have been variable [15,33,34]. Although median values of CYP3A4 and CYP3A5 expression were apparently not influenced by atorvastatin due to their well known great variability (ten- to 40-fold variation in CYP3A4 liver mRNA expression has been reported [10,35]), the linear regression analysis showed that an increase in CYP3A4 and CYP3A5 mRNA expression levels was positively correlated to the reduction of LDL cholesterol. Therefore, a relationship between cholesterolemia and CYP3A4 and CYP3A5 mRNA expression in response to atorvastatin treatment was demonstrated. On the other hand, LDL cholesterol levels of HC subjects after treatment were still higher than in NL subjects, which might

162

M.A.V. Willrich et al. / Clinica Chimica Acta 421 (2013) 157–163

Fig. 3. Relationship of CYP3A5 haplotypes with mRNA levels at baseline (A) and after atorvastatin treatment (B) in hypercholesterolemics. Haplotype analysis was performed using HaploStats software. mRNA expression of each gene was measured by qPCR and normalized with GAPDH. The 2−DCt formula was used to calculate the relative amount of transcripts in the sample. Y axis is on log scale. Median values for each haplotype are represented as a hyphen (−) in the jitterplot. Haplotypes were formed in the following order: CYP3A4*1B, CYP3A5*3C and CYP3A5*1D. Reference haplotype was considered as the most frequent one, AGT.

explain why CYP3A4 and CYP3A5 mRNA expression in HC subjects did not increase up to the NL levels. A relevant aspect of this finding is that a greater decrease in LDL cholesterol levels (i.e. better response to treatment) was associated with a greater increase in CYP3A4 and CYP3A5 mRNA expression (positive slope). An increase in CYP3A4 and CYP3A5 mRNA levels would lead to enhanced atorvastatin activation and consequently, a greater LDL-c reduction would take place. The urinary 6BOHC/cortisol ratio reflects CYP3A activity in the liver after renal clearance for cortisol and its metabolite. The 6BOHC/cortisol ratio is advantageous in that it requires only an isolated urine sample and it is the only currently available non-invasive assay for CYP3A activity. Moreover, it has been suggested that this method is accurate enough for detecting drug induction or inhibition when the subjects are their own controls [36]. Given that CYP3A4 and CYP3A5 have overlapping substrates, an assay has not yet been developed that can tease out the activity of CYP3A4 and CYP3A5 individually. Overall, HC presented with lower 6BOHC/cortisol ratios when compared to NL subjects (Fig. 2) suggesting that CYP3A activity is lower in HC subjects. As reported by Honda and colleagues (2011) in hepatocytes [37], cholesterol is a substrate for 4-β-hydroxylation by CYP3A4. The product of that reaction, 25-hydroxy-cholesterol, is a potent HMG-CoA reductase inhibitor in vitro, and therefore, CYP3A may participate in the regulation of lipid metabolism [37]. A consequence of the fact that cholesterol is a substrate for CYP3A4 is that it may function as an inhibitor of the enzyme's activity [38]. Shinkyo and colleagues [39] have reported that cholesterol acts as a substrate and inhibitor of CYP3A4 in hepatocytes by binding to CYP3A4 and, at 10 μM, inhibiting CYP3A4 prototypic reactions. Therefore, the lower CYP3A activity levels found in HC subjects may be explained by the higher levels of serum cholesterol in those patients. Furthermore, using linear regression, we demonstrated that CYP3A activity was inversely correlated to the reduction of LDL cholesterol in response to atorvastatin. While the 6BOHC/cortisol ratio reflects the endogenous activity of CYP3A4 and CYP3A5, interactions between CYP enzymes and their ligands are often atypical. There is more than one mode of binding and conformational changes upon binding of some substrates [38]. One could hypothesize that, after the patients were given atorvastatin— that has a high affinity to CYP3A4/5, it would become the enzymes' first priority to metabolize the drug, with a lower priority given to their endogenous role of cortisol hydroxylation. This could explain

why we observed an association of lower levels of CYP3A activity in HC subjects after LDL-c reduction. More than a decade ago, it was stated that 90% of the variation in CYP3A4 activity was attributable to genetic factors [40]. It is intriguing that despite multiple groups' efforts in pharmacogenomic studies, the keystone for explaining the interindividual variability observed in the expression of CYP3A4 has yet to be identified. It has been suggested that the remaining heritable variability hitherto unexplained might be due to epigenetic regulation, not evaluated in this report. Or yet, as suggested by Wang et al. [34], the impact of the variants could be tissue-specific and influence mRNA and protein expression just in the liver, for instance. We used a candidate gene strategy to evaluate the relationship between individual polymorphism and haplotypes of CY3A4 and CYP3A5 with mRNA expression in PBMCs. To the best of our knowledge, the haplotype analyses as performed in this study were not previously reported. Interestingly the haplotypes GAT and GAC, but not the individual variants, were associated with increased CYP3A5 expression in HC patients. Baseline CYP3A5 mRNA expression was lower in AGT than in GAC or GAT haplotype carriers in HC subjects, confirming the CYP3A5*3C (middle G in the AGT haplotype) deleterious effect on CYP3A5 expression. CYP3A5*3C abolishes a splice site and creates a pseudoexon, thereby generating several non-coding mRNAs and a non-functional truncated protein. It is unclear as to why there is a difference in CYP3A5 mRNA expression in the HC group at baseline, but not after atorvastatin treatment. One hypothesis is that treatment with atorvastatin, and its accompanying pleiotropic effects may activate transcription factors and/or other biological pathways that compensate for the deleterious effect of CYP3A5*3C. Further investigations would be required to explore this hypothesis. It is likely that the high variability in CYP3A4 and CYP3A5 mRNA expression, as well as CYP3A activity, complicated the observation of atorvastatin effects. The wide range of response to atorvastatin treatment in a genetically diverse population such as ours created a heterogeneous HC group. Furthermore, a lack of liver biopsies from these patients hindered the ability to correlate the PBMC findings to the liver. Future studies in larger cohorts of patients will be necessary to reproduce and correlate PBMC findings with liver specimens. Future studies may further benefit from examining the effects of haplotypes in conjunction with other variations in other genes, such as the ones that code for

M.A.V. Willrich et al. / Clinica Chimica Acta 421 (2013) 157–163

phase II enzymes and efflux transporters such as SLCO1B1 and ABCB1, in order to better understand the impact of the combined variants on response to atorvastatin treatment. In conclusion, this study reports that changes in cholesterolemia status promoted by statins influence expression of CYP3A4 and CYP3A5 mRNA in PBMCs and affect CYP3A activity in different ways. Acknowledgments This study was supported by grants from CNPq (#502171/2004-9), Brasilia, and FAPESP (# 2008/06667-9), Sao Paulo, Brazil. M. A. V. Willrich, A. C. Rodrigues and F. D. V. Genvigir were recipients of scholarships from FAPESP. A. Cerda is the recipient of a fellowship from CONYCIT, Chile. M.H. Hirata and R.D.C. Hirata are recipients of fellowships from CNPq. This publication statistical analysis was supported by NIH/NCRR CTSA Grant Number UL1 RR024150. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH. References [1] Shitara Y, Sugiyama Y. Pharmacokinetic and pharmacodynamic alterations of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors: drug– drug interactions and interindividual differences in transporter and metabolic enzyme functions. Pharmacol Ther 2006;112:71–105. [2] Brown MS, Goldstein JL. Lowering plasma cholesterol by raising LDL receptors. N Engl J Med 1981;305:515–7. [3] Wang CY, Liu PY, Liao JK. Pleiotropic effects of statin therapy: molecular mechanisms and clinical results. Trends Mol Med 2008;14:37–44. [4] Vladutiu GD. Genetic predisposition to statin myopathy. Curr Opin Rheumatol 2008;20:648–55. [5] Neuvonen PJ, Backman JT, Niemi M. Pharmacokinetic comparison of the potential over-the-counter statins simvastatin, lovastatin, fluvastatin and pravastatin. Clin Pharmacokinet 2008;47:463–74. [6] Neuvonen PJ. Drug interactions with HMG-CoA reductase inhibitors (statins): the importance of CYP enzymes, transporters and pharmacogenetics. Curr Opin Investig Drugs 2010;11:323–32. [7] Grube M, Kock K, Oswald S, et al. Organic anion transporting polypeptide 2B1 is a high-affinity transporter for atorvastatin and is expressed in the human heart. Clin Pharmacol Ther 2006;80:607–20. [8] Zanger UM, Turpeinen M, Klein K, Schwab M. Functional pharmacogenetics/genomics of human cytochromes P450 involved in drug biotransformation. Anal Bioanal Chem 2008;392:1093–108. [9] Wojnowski L, Hustert E, Klein K, et al. Re: modification of clinical presentation of prostate tumors by a novel genetic variant in CYP3A4.J Natl Cancer Inst 2002;94: 630–1 [author reply 631–632]. [10] Lamba V, Panetta JC, Strom S, Schuetz EG. Genetic predictors of interindividual variability in hepatic CYP3A4 expression. J Pharmacol Exp Ther 2010;332:1088–99. [11] Rodrigues AC, Hirata MH, Hirata RD. The genetic determinants of atorvastatin response. Curr Opin Mol Ther 2007;9:545–53. [12] Kocarek TA, Dahn MS, Cai H, Strom SC, Mercer-Haines NA. Regulation of CYP2B6 and CYP3A expression by hydroxymethylglutaryl coenzyme A inhibitors in primary cultured human hepatocytes. Drug Metab Dispos 2002;30:1400–5. [13] Watanabe M, Kumai T, Matsumoto N, et al. Expression of CYP3A4 mRNA is correlated with CYP3A4 protein level and metabolic activity in human liver. J Pharmacol Sci 2004;94:459–62. [14] Takagi S, Nakajima M, Mohri T, Yokoi T. Post-transcriptional regulation of human pregnane X receptor by micro-RNA affects the expression of cytochrome P450 3A4. J Biol Chem 2008;283:9674–80. [15] Willrich MA, Hirata MH, Hirata RD. Statin regulation of CYP3A4 and CYP3A5 expression. Pharmacogenomics 2009;10:1017–24. [16] Cerda A, Genvigir FD, Arazi SS, et al. Influence of SCARB1 polymorphisms on serum lipids of hypercholesterolemic individuals treated with atorvastatin. Clin Chim Acta 2010;411:631–7. [17] Haas CE, Brazeau D, Cloen D, et al. Cytochrome P450 mRNA expression in peripheral blood lymphocytes as a predictor of enzyme induction. Eur J Clin Pharmacol 2005;61:583–93.

163

[18] Siest G, Jeannesson E, Marteau JB, et al. Transcription factor and drug-metabolizing enzyme gene expression in lymphocytes from healthy human subjects. Drug Metab Dispos 2008;36:182–9. [19] Genvigir FD, Rodrigues AC, Cerda A, et al. Effects of lipid-lowering drugs on reverse cholesterol transport gene expressions in peripheral blood mononuclear and HepG2 cells. Pharmacogenomics 2010;11:1235–46. [20] Willrich MA, Hirata MH, Genvigir FD, et al. CYP3A53A allele is associated with reduced lowering-lipid response to atorvastatin in individuals with hypercholesterolemia. Clin Chim Acta 2008;398:15–20. [21] Chahoud G, Aude YW, Mehta JL. Dietary recommendations in the prevention and treatment of coronary heart disease: do we have the ideal diet yet? Am J Cardiol 2004;94:1260–7. [22] Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 1972;18:499–502. [23] Salazar LA, Hirata MH, Cavalli SA, Machado MO, Hirata RD. Optimized procedure for DNA isolation from fresh and cryopreserved clotted human blood useful in clinical molecular testing. Clin Chem 1998;44:1748–50. [24] Cavalli SA, Hirata MH, Hirata RD. Detection of MboII polymorphism at the 5′ promoter region of CYP3A4. Clin Chem 2001;47:348–51. [25] Vandesompele J, De Preter K, Pattyn F, et al. Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes.Genome Biol 2002;3 [RESEARCH0034]. [26] Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(−Delta Delta C(T)) method. Methods 2001;25:402–8. [27] Rouits E, Boisdron-Celle M, Morel A, Gamelin E. Simple and sensitive high-performance liquid chromatography method for simultaneous determination of urinary free cortisol and 6beta-hydroxycortisol in routine practice. For CYP 3A4 activity evaluation in basal conditions and after grapefruit juice intake. J Chromatogr B Analyt Technol Biomed Life Sci 2003;793:357–66. [28] Lake SL, Lyon H, Tantisira K, et al. Estimation and tests of haplotype–environment interaction when linkage phase is ambiguous. Hum Hered 2003;55:56–65. [29] Schaid DJ, Rowland CM, Tines DE, Jacobson RM, Poland GA. Score tests for association between traits and haplotypes when linkage phase is ambiguous. Am J Hum Genet 2002;70:425–34. [30] 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;16:524–30. [31] Suarez-Kurtz G. Pharmacogenomics and the genetic diversity of the Brazilian population. Cad Saude Publica 2009;25:1650–1. [32] Suarez-Kurtz G, Genro JP, de Moraes MO, et al. Global pharmacogenomics: impact of population diversity on the distribution of polymorphisms in the CYP2C cluster among Brazilians. Pharmacogenomics J 2012;12:267–76. [33] Rosales A, Alvear M, Cuevas A, Saavedra N, Zambrano T, Salazar LA. Identification of pharmacogenetic predictors of lipid-lowering response to atorvastatin in Chilean subjects with hypercholesterolemia. Clin Chim Acta 2012;413:495–501. [34] Wang D, Guo Y, Wrighton SA, Cooke GE, Sadee W. Intronic polymorphism in CYP3A4 affects hepatic expression and response to statin drugs. Pharmacogenomics J 2011;11: 274–86. [35] Kuehl P, Zhang J, Lin Y, et al. Sequence diversity in CYP3A promoters and characterization of the genetic basis of polymorphic CYP3A5 expression. Nat Genet 2001;27: 383–91. [36] Galteau MM, Shamsa F. Urinary 6beta-hydroxycortisol: a validated test for evaluating drug induction or drug inhibition mediated through CYP3A in humans and in animals. Eur J Clin Pharmacol 2003;59:713–33. [37] Honda A, Miyazaki T, Ikegami T, et al. Cholesterol 25-hydroxylation activity of CYP3A. J Lipid Res 2011;52:1509–16. [38] Diczfalusy U, Bjorkhem I. Still another activity by the highly promiscuous enzyme CYP3A4: 25-hydroxylation of cholesterol. J Lipid Res 2011;52:1447–9. [39] Shinkyo R, Guengerich FP. Inhibition of human cytochrome P450 3A4 by cholesterol. J Biol Chem 2011;286:18426–33. [40] Ozdemir V, Kalow W, Tang BK, et al. Evaluation of the genetic component of variability in CYP3A4 activity: a repeated drug administration method. Pharmacogenetics 2000;10:373–88. [41] Mancia G, De Backer G, Dominiczak A, et al. 2007 guidelines for the management of arterial hypertension: The Task Force for the Management of Arterial Hypertension of the European Society of Hypertension (ESH) and of the European Society of Cardiology (ESC). J Hypertens 2007;25:1105–87. [42] WHO. World Health Organization Global information system on alcohol and death. Definition of an alcoholic dose. Available at http://www.who.int/gho/alcohol/en/ index.html.