Polymorphisms of peroxisome proliferator-activated receptor γ (PPARγ) and cluster of differentiation 36 (CD36) associated with valproate-induced obesity in epileptic patients.

Rationale: Valproate (VPA) is a choice for the treatment of primary generalized epilepsies and partial epilepsies. Unfortunately, weight gain or obesity is one of the most frequent adverse effects of VPA treatment. Genetic factors were shown to be involved in the effect.Objective: The aim of this study was to investigate the association of selected single nucleotide polymorphisms (SNPs) of cluster of differentiation 36 (CD36) and peroxisome proliferator-activated receptor γ (PPARγ) with VPA-induced weight gain and obesity in epileptic patients.Methods: A total of 225 Chinese Han epilepsy patients receiving VPA treatment were recruited in the study. Height and weight for the calculation of body mass index (BMI) were measured at the initiation of VPA therapy and in the follow-up examination. A BMI of 25 kg/m² or higher was defined as obesity on the basis of the World Health Organization (WHO) criteria for Asian populations. Four SNPs in CD36 (rs1194197, rs7807607) and PPARγ (rs10865710, rs2920502) were genotyped using the Sequenom^® MassArray iPlex platform.Results: About 19.6% of epileptic patients receiving VPA therapy were found to become obese. After covariate analysis of age, gender, sex, height, initial BMI, and VPA dosage, the CD36 rs1194197 C allele and rs7807607 T allele (OR, 0.31; 95%CI, 0.13-0.72; P = 0.009 and OR, 0.38; 95%CI; 0.18-0.83; P = 0.02, respectively) were identified as protective factors for VPA-induced obesity. The PPARγ rs10865710 C allele carriers were found to be less likely to suffer from VPA-induced obesity compared with GG genotype carriers (OR, 0.04; 95%CI, 0.01-0.12; P < 0.001). After a Bonferroni correction for multiple comparisons, the genotypic associations of CD36 rs1194197 and PPARγ rs10865710 and the allelic association of CD36 rs7807607 with obesity remained statistically significant.Conclusions: Our data first indicated that CD36 and PPARγ polymorphisms may be associated with VPA-induced obesity and weight gain, suggesting that CD36 and PPARγ may have potential value in predicting VPA-induced obesity in Chinese Han epileptic patients.

VPA; Epileptic; Obesity; PPARγ; CD36; Genetic polymorphisms

Valproate (VPA) is a choice for primary generalized epilepsies and partial epilepsies (Williams et al. [44] ). In some countries, it is also used in the prophylactic treatment of migraines and in the treatment of neuropathic pain and psychiatric disorders (schizophrenia and bipolar disorder). The VPA-based antiepileptic therapy usually lasts years and sometimes even lasts for the whole lifetime. Although it is considered to be safe in most patients, drug-associated weight gain or obesity is one of the most common reasons for patients to discontinue VPA therapy (Kanemura et al. [19] ). A recent study from America indicated approximately 40.4% of VPA-treated patients gained more than 10% of their total body weight and developed obesity, whereas the percentage was only 8% in patients treated with other anticonvulsant drugs (Biton et al. [2] ). About 17% of Spanish epileptic patients treated with monotherapy of VPA developed obesity, and about 6% of patients went on to become overweight (Carmona-Vazquez et al. [6] ). However, the mechanism underlying VPA-induced weight gain and obesity remains unknown.

VPA-induced weight gain may be ascribed to multifactorial effects. This is because the progress of obesity induced by VPA in epileptic patients was always accompanied by metabolic disturbances, including hyperinsulinemia, hyperleptinemia, dyslipidemia, insulin resistance, and leptin (LEP) resistance (Verrotti et al. [43] ). Several studies considered VPA-related weight gain as a result of insulin elevation which promoted glycogenesis and lipogenesis in skeletal muscle (Pylvänen et al. [34] ; Verrotti et al. [42] ), while other studies indicated the obesity might result from VPA-induced increase of non-high-density lipoprotein cholesterol (Fang et al. [11] ; Yamamoto et al. [45] ). However, how VPA induces the metabolic dysfunction that facilitates weight gain or obesity is not well defined. Interestingly, Klein et al. ([21] ) found similar patterns of weight gain in monozygotic twins treated with VPA. A gender-matched twin and sibling pair study suggested it was VPA duration but not dosage that was positively associated with weight gain (Petty et al. [32] ). This evidence shed light on the effect of genetic variation in the development of VPA-induced obesity. In fact, several studies have uncovered the importance of polymorphisms in genes, such as cytochrome P450 2C19 (CYP2C19) (Noai et al. [30] ) and the LEP receptor (LEPR) (Li et al. [24] ), for predicting the risk of VPA-induced obesity.

Peroxisome proliferator-activated receptor γ (PPARγ) is a member of the ligand-activated nuclear hormone receptor superfamily and regulates expressions of battery genes involved in adipocyte differentiation, lipid metabolism and storage, and insulin sensitivity (Jeninga et al. [18] ). The activation of PPARγ can enhance the sensitivity of adipocytes to insulin via upregulating the expression of adiponectin and promote the formation of mature adipocytes (Scherer [36] ). VPA is a potential activator for PPARγ (Zuckermann et al. [47] ). Expressions of lipid metabolism-related genes induced by VPA were observed to be accompanied by activated PPARγ (Bai et al. [1] ), indicative of the important role of PPARγ in VPA-induced metabolic disorders. In the past decade, the association of common variations in PPARγ with metabolic syndrome induced by antipsychotic drugs, such as olanzapine or clozapine, was well studied in Americans (Tiwari et al. [40] ; Tiwari et al. [39] ), Germans (Brandl et al. [4] ; Staeker et al. [37] ), Polish (Grygiel-Gorniak et al. [14] ), Turks (Herken et al. [17] ), and Japanese (Ujike et al. [41] ). However, there has been no investigation into the correlation between PPARγ polymorphisms and VPA-induced obesity in Chinese epileptic patients.

A cluster of differentiation 36 (CD36) is a glycoprotein expressed at the surface of multiple cell types, including adipocytes, skeletal muscle cells, and hepatocytes (Su and Abumrad [38] ). It belongs to the B scavenger receptor family and shows high affinity to oxidized low-density lipoprotein (LDL) and long-chain fatty acid (FA). Results from transgenic mice revealed CD36 played an important role in fat metabolism (Hajri et al. [15] ). Correlations of CD36 polymorphisms with lipometabolic disorder-based diseases, such as cardiovascular diseases, atherosclerosis, metabolic syndrome, and obesity, have been reported in Japan (Miyaoka et al. [29] ), Spain (Bokor et al. [3] ), America (Ma et al. [27] ), and Germany (Heni et al. [16] ). Additionally, CD36 is also involved in oral fat perception. It was shown that CD36-null mice failed to display the preference for linoleic acid compared with wild-type mice (Laugerette et al. [22] ). Common variants in CD36 were found to be associated with oral fat preference and obesity in African Americans (Keller et al. [20] ). More importantly, the hepatic regulation of CD36 by VPA treatment was uncovered by our previous study (Bai et al. [1] ), implying the potential significance of CD36 in VPA-induced weight gain and obesity. However, the association of CD36 polymorphisms with VPA-induced obesity in epileptic patients has not yet been elucidated.

Hypothesizing that genetic polymorphisms of PPARγ and CD36 may play important roles in the development of VPA-induced weight gain, we investigated the associations of CD36 (rs1194197 and rs7807607) or PPARγ (rs10865710 and rs2920502) with VPA-induced obesity in Chinese Han epileptic patients. Our results first indicated that the major alleles of CD36 rs1194197, rs7807607, and PPARγ rs10865710 were associated with a lower risk of VPA-induced weight gain and obesity in Chinese Han epileptic patients. These data may provide clues for a better understanding of the potential mechanisms underlying VPA-induced weight gain and offer novel markers to predictively evaluate the risk of obesity in Chinese Han epileptic patients treated with VPA-based therapy.

This study was approved by the Human Investigation Ethics Committee of the School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou, China (http://clinicaltrials.gov Identifier No. NCT01172626). It was conducted in accordance with the Declaration of Helsinki and was consistent with applicable guidelines of good clinical practice. A total of 225 Chinese Han epilepsy patients who ranged from 18 to 60 years old and received VPA treatment were recruited in this study. The written informed consent had been obtained from all patients prior to enrolment. The exclusion criteria included failing to follow the prescription or to achieve complete seizure control, pregnancy, infancy, severe head injuries, major surgery during the study, substance dependence and additional medicines that induce weight gain, family history of obesity, a history of obesity before VPA therapy, allergy to VPA, mental deficiency and personality disorder, hepatitis C, HIV, thyroid disorder, diabetes mellitus, and other unaccepted diseases that may influence the present study. All patients in the study had been diagnosed with epilepsy according to the revised guideline of the International League Against Epilepsy (https://www.epilepsydiagnosis.org/) and had normal hepatorenal functions on the basis of the data from laboratory examinations and clinical records. According to the situation, the dosing regimen for each patient was maintained stably. All the patients were told to take proper exercise and have regular sleep (7-8 h) during the therapy. Smoking and drinking were absolutely forbidden. Demographic characteristics, clinical information, and prescriptions were collected from patients’ real-time electronic medical records at the Department of Neurology at the First Affiliated Hospital of Sun Yat-sen University. Height and weight for the calculation of body mass index (BMI) were measured at the initiation of VPA therapy and in the follow-up examination. The World Health Organization (WHO) criterion for the Asian obese population (BMI ≥ 25 kg/m2) was used as the definition of obesity (http://www.wpro.who.int/nutrition/documents/Redefining%5fobesity/en).

Whole blood samples (2 mL) from patients were collected in EDTA tubes and stored at − 80 °C. The extraction of DNA from whole blood was performed using the Generay DNA extraction Kit (Tiangen, Beijing, China) according to the manufacturer’s protocol. A DNA absorbance ratio (A260/A280) of 1.8 or higher was considered high purity, while the concentrations of extracted DNA were detected by a NanoDrop 2000 Spectrophotometer (Thermo Fisher, Rockford, IL, USA).

Four single nucleotide polymorphisms (SNPs) selected in this study were shown in Table 1 and were based on (1) minor allele frequency (MAF) greater than 0.1, (2) previously reported associations with metabolic disorders, and (3) previously reported functional effects. Genotyping of all polymorphisms was performed using the MassArray iPLEX technology platform (Sequenom, San Diego, CA, USA) according to the instructions for multiplexed genotyping analysis. Briefly, DNA (10 ng) samples were amplified through multiplex PCR with a Mastercycler nexus PCR amplifier (Eppendorf, Hamburg, Germany) and then were used for iPLEX extension reactions in a 384-well plate (Thermo Scientific, Rockford, IL, USA). The final products were desalted and dispensed into a 384-pad SpectroChip (Sequenom, San Diego, CA, USA). Allele detection was carried out using MALDI-TOF mass spectrometry (Sequenom, San Diego, CA, USA). MassArray Typer 4.0 software (Sequenom, San Diego, CA, USA) was used to analyze the data. Inspection of the clusters was performed to ensure a clear cluster separation with satisfactory signal-to-noise cut-off. SpectroChip data with a more than 10% call rate in the blank check or with a more than 25% call rate in the blank control were considered invalid and were repeated.

Data were expressed as mean ± SD for continuous variables and frequencies (%) for categorical variables. Student’s t test and Mann-Whitney U test were used for analyzing normally distributed and non-normally distributed continuous variables, respectively, while the Pearson χ2 test was used for analyzing categorical variables. The logistic regression model and covariance analysis were utilized for multivariate analysis to avoid confounding factors. Bonferroni corrections were used for multiple comparisons. The Hardy-Weinberg equilibrium was assessed using Haploview 4.2. A value of P < 0.05 was defined as statistically significant. All statistical analyses were performed using SPSS 21 software (IBM Corp., Armonk, NY, USA). The graphs were prepared using GraphPad Prism 5 (GraphPad Software Incorporated, San Diego, CA, USA).

A total of 225 Chinese Han epileptic patients that included 142 males (63.1%) and 83 females (36.9%) were recruited in the current study. The average age of the study population was approximately 25.80 ± 7.01 years old. All the patients were divided into the non-obese or obese group according to the aforementioned definition after treatment with VPA for a year. The prevalence of obesity was about 19.6% in our study. Among 225 patients, 134 subjects received VPA monotherapy, while 91 subjects were treated with VPA in combination with either lamotrigine (LTG) (n = 72), carbamazepine (CBZ) (n = 12), or oxcarbazepine (OXC) (n = 7). These other drugs were known to have little influence on weight; therefore, the patients treated with these drugs were also included in this study (Pickrell et al. [33] ). Consistently, no statistical discrepancy was found in the obesity rate among these four categories of treatment regimen. Other parameters of the study population were described in Table 2. There was no statistical discrepancy in the distribution of age, sex, initial weight, initial BMI, or VPA dosage between the non-obese and obese groups.

Table 3 showed the genotype distributions and allele frequencies for selected SNPs. The allelic distributions of these SNPs were in accordance with the Hardy-Weinberg equilibrium (P > 0.05). The allelic frequencies were almost consistent with the data from the NCBI database (http://www.ncbi.nlm.nih.gov/snp). The associations of selected SNPs with VPA-induced obesity can be observed in Table 3. Specifically, patients who were carriers of the C allele of CD36rs1194197 were significantly associated with a lower risk of VPA-induced obesity (CC + CT vs TT, P = 0.009), and the frequency of the C allele of rs1194197 in non-obese patients was higher than that in obese counterparts (64.6 vs 50.0%, P = 0.009) (Table 3). CD36 rs7807607 was also significantly associated with VPA-induced obesity. Patients with the T allele of CD36 rs7807607 had a lower obesity risk than those carrying the CC genotype (TT + CT vs CC, P = 0.02) (Table 3). The frequency of the T allele of rs7807607 was higher in non-obese patients (65.7 vs 48.9% in obese patients, P = 0.005) (Table 3).For PPARγ rs10865710, the frequency of the C allele was higher in non-obese patients (71.8 vs 36.4% in obese patients, P < 0.001) (Table 3). Patients who were carriers of the GG genotype had a higher risk of obesity during VPA therapy compared to those who carried CC + CG genotypes (GG vs CC + CG, P < 0.001) (Table 3). PPARγ rs2920502 was found to have no relationship with VPA-induced obesity in our study (Table 3). These results suggest that the CD36 and PPARγ gene polymorphisms may be associated with VPA-induced obesity in Chinese Han epileptic patients.

To gain a better understanding of the correlations between the CD36 and PPARγ gene polymorphisms and VPA-induced obesity, the association of genotypes with BMI change in the study patients was also examined. As shown in Table 4, no relationship was found between genotypes and baseline BMI. After VPA treatment for a year, patients who were carriers of the TT genotype of CD36 rs1194197, CC genotype of CD36 rs7807607, or GG genotype of PPARγ rs10865710 had higher BMI compared with those who carried the major alleles (P = 0.002, P = 0.01, and P < 0.001, respectively) (Table 4). Next, we examined the difference of the linear regression between mean BMI and therapeutic time between different genotypes (Fig. 1). The results showed that the regression line for CD36 rs1194197 TT or rs7807607 CC genotype exhibited significantly higher slope than that for the corresponding major allele (0.22 ± 0.02 vs 0.05 ± 0.002, P < 0.0001 and 0.20 ± 0.02 vs 0.06 ± 0.006, P < 0.0001, respectively) (Fig. 1a, b). Furthermore, the regression line for PPARγ rs10865710 CC + CG genotypes showed significantly lower slope as compared to that of the minor genotype (0.09 ± 0.008 vs 0.49 ± 0.04, P < 0.0001) (Fig. 1c). Table 5 showed the association of genotypes with BMI increase in patients after VPA treatment initiation. It was shown that, after 12 months, patients who were carriers of the CD36 rs1194197 TT genotype or rs7807607 CC genotype gained significantly more weight than those with the major alleles (P = 0.004 and P = 0.001, respectively) (Table 5). Patients with the major allele of PPARγ rs10865710 gained much less weight than those with the GG genotype at a period of 12 months after therapy initiation (P = 0.006) (Table 5). These data, thus, indicate that the CD36 and PPARγ gene polymorphisms may be closely associated with VPA-induced weight gain in Chinese Han epileptic patients.a-c Linear regression between mean BMI and therapeutic time in different genotypes. The difference of slopes between two regression lines was evaluated. P < 0.05 was considered statistically significant

VPA-induced obesity is closely associated with many chronic diseases, such as cardiovascular diseases, atherosclerosis, metabolic syndrome, and impaired social interaction, and therefore, it has become the most common reason for epileptic patients to discontinue VPA therapy (Breum et al. [5] ). Unfortunately, the mechanism underlying VPA-triggered weight gain has not yet been fully understood. In this study, we found that the major alleles of CD36 rs1194197 and rs7807607 and PPARγ rs10865710 may be protective factors for VPA-induced weight gain and obesity in Chinese Han epilepsy patients. The genotypic associations of CD36 rs1194197 and PPARγ rs10865710 and the allelic association of CD36 rs7807607 with obesity remained statistically significant after the Bonferroni correction (P < 0.0125). These data first indicated the predictive values of CD36 and PPARγ polymorphisms in VPA-induced obesity in Chinese Han epilepsy patients.

The high prevalence of obesity in VPA-treated epileptic patients has been reported in many countries. For instance, approximately 47% of patients gained more than 10% of their total body weight and developed obesity after VPA treatment, while the percentage was only 14% in patients treated with other anticonvulsant drugs (Corman et al. [8] ). A similar incidence (40.4%) was also found in America (Biton et al. [2] ). The obesity rate was 17% in Spanish epileptic patients treated with monotherapy of VPA with around 6% of patients becoming overweight (Carmona-Vazquez et al. [6] ). It was reported that about 27.5% of men and 56.4% of women gained more than 5 kg during the short-term VPA therapy in Austria and, therefore, suffered from individual sociopsychological problems (El-Khatib et al. [10] ). Compared with the Western countries, the incidence of VPA-induced obesity in Chinese Han epileptic patients was lower but cannot be ignored. In this study, the prevalence of obesity was approximately 19.6%, which is in accordance with our previous report (Li et al. [24] ). The discrepancy may be caused by different races, dietary patterns, or external influences.

PPARγ determines adiposity and energy balance by regulating adipocyte differentiation and fat metabolism (Zhou et al. [46] ). A recent study showed that PPARγ-null mice failed to develop adipose tissue, suggestive of an essential role of PPARγ in the formation of mature adipocytes (Doerks et al. [9] ). In humans, PPARγ polymorphisms were of high interest in a pharmacogenetic study of antipsychotic-induced weight gain (Lett et al. [23] ). Hypothesizing that PPARγ polymorphisms may play a significant role in VPA-induced weight gain and obesity, two SNPs in the PPARγ gene were investigated in this cohort. PPARγ rs10865170 resides in the promoter region, which is characterized by a C to G substitution in human beings. The allelic frequency of G in this study was about 0.34 and is similar to the reported data in the Chinese population (Luo et al. [26] ). PPARγ rs10865170 has been considered as an important genetic link to the regulation of lipid metabolism and total body weight (Mansoori et al. [28] ); however, the association of the PPARγ polymorphism with VPA-induced weight gain and obesity in Chinese Han epileptic patients is unclear. In this study, we first uncovered the association of the major allele of PPARγ rs10865170 with a lower risk of VPA-induced weight gain and obesity (P < 0.001). VPA was known to be a potential activator for PPARγ, whereas PPARγ proteins in carriers with the major allele of rs10865170 had lower transcriptional activity compared to homozygotes for the minor allele (Jeninga et al. [18] ). This may disrupt the inducible effect of VPA on PPARγ and inhibit PPARγ-mediated adipose differentiation and weight gain. Although the C allele in this study seemed to be a protective factor for VPA-induced weight gain or obesity, some other publications indicated that the G allele was associated with a lower prevalence of obesity (Luo et al. [26] ), implying the potential interaction between VPA and PPARγ rs10865170, which requires further exploration.

CD36 is one of the most important downstream genes for PPARγ. It is a high-affinity receptor for long-chain FAs and has been frequently investigated in lipid homeostasis regulation, including oral fat perception, FA metabolism, and lipid accumulation (Su and Abumrad [38] ). CD36 expression in the platelet membrane was highly correlated with hyperlipidemia that could lead to obesity (Bokor et al. [3] ), atherosclerosis (Love-Gregory et al. [25] ), coronary heart disease (Ma et al. [27] ), and fatty liver diseases (Farinelli et al. [12] ) in human beings. This has made CD36 an attractive biomarker for personalized medicine for many years. Results from our laboratory revealed the ability of VPA to induce the expression of CD36 (Bai et al. [1] ); therefore, we hypothesized that, to some extent, genetic variants in CD36 may be associated with VPA-induced weight gain and investigated the possibility. CD36 rs1194197is located in the upstream of 5’-UTR in the CD36 gene sequence and is characterized by a C to T substitution in humans. The allelic frequency of T was about 0.34 in this study, which is lower than the data from other countries (Farook et al. [13] ; Keller et al. [20] ). Previous reports indicated that the T allele of rs1194197 increased the risk of metabolic syndrome induced by antipsychotic therapy in Americans (Keller et al. [20] ; Love-Gregory et al. [25] ). However, the association of CD36 rs1194197 with VPA-induced obesity in Chinese Han epileptic patients remains unknown. In the present study, we first found that the CD36 rs1194197 T allele was associated with a higher risk of VPA-induced obesity (P = 0.009). The variant has been reported to affect the alternative splicing of exons and upstream promoters in the CD36 gene, which may promote the VPA-induced platelet expression of CD36 (Chang et al. [7] ; Rac and Safranow [35] ). As a result, the serum level of HDL was decreased, while serum levels of TG and FFA were elevated (Keller et al. [20] ). The excessive lipid in plasma stimulated adipose differentiation, leading to the developing weight gain in the period of VPA treatment. This evidence further confirmed our findings that the CD36 rs1194197 T allele may increase the risk of VPA-induced obesity.

CD36 rs7807607 is anchored to the upstream of 5’-UTR in the CD36 gene sequence, which is characterized by a T to C substitution in human beings. The allelic frequency of C in our study was about 0.40. It is similar to the data on the Chinese population in the NCBI database. A previous paper has demonstrated the correlation of the CD36 rs7807607 C allele with the increased risk of metabolic syndrome in Americans (Noel et al. [31] ). However, its association with VPA-induced obesity in Chinese Han epileptic patients has not yet been elucidated. In our investigation, we found that the CD36 rs7807607 minor allele was associated with a higher risk of VPA-induced obesity in Chinese Han epileptic patients (P = 0.005). Although the influence of the mutant allele on the function of CD36 is not clear, carriers of the minor allele of rs7807607 were found to be characterized by higher plasma levels of FFA and TG compared to those with homozygotes for the major allele (Farook et al. [13] ). The abnormity of the blood lipid metabolic triggered by this variant may aggravate VPA-induced metabolic disorders and increase the odds of drug-induced weight gain and obesity, which makes the minor allele of CD36 rs7807607 a risk factor for VPA-induced weight gain and obesity.

In the study, we found there was a small population of patients with higher baseline BMI in risk genotype groups. Although the difference of mean initial BMI between the two groups was insignificant, the higher mean BMI of baseline in risk genotypes is visible (Table 4 and Fig. 1). This might be a collective consequence of lipid metabolism-related genetic mutations that facilitated antiepileptic drug-induced weight gain as described in “Introduction.” Thus far, the exact mechanism underlying the associations of these genetic polymorphisms with VPA-induced weight gain is unknown and needs further investigation.

In summary, our findings first indicated that the major genotypes of CD36 rs1194197, rs7807607, and PPARγ rs10865710 may be protective factors for VPA-induced weight gain and obesity in Chinese Han epileptic patients, suggesting the predictive values of CD36 and PPARγ polymorphisms in individualized VPA-based antiepileptic therapy. Further study is needed as an extension of this retrospective study to confirm the present findings through diet control and selecting patients at a more similar baseline BMI. Moreover, further investigations related to the SNP interactions of reported genes and functional analysis in a larger sample size with the individualized alternation of lipid profile need to be performed. Such studies may be helpful for better risk evaluation so that patients with a high risk of drug-induced weight gain and obesity can be treated with more caution.

Xupeng Bai, Chuncao Xu, and Dingsheng Wen are co-first authors.

This work was supported by the National Key Research and Development Program (2017YFC0909303), National Natural Science Foundations of China [81730103 and 81573658], Guangdong Provincial Key Laboratory of Construction Foundation [2017B030314030], and the Science and Technology Planning Project of Guangdong Province (2017A020215147).

Prof. Min Huang and Prof. Jing Jin designed the research; Xupeng Bai and Dingsheng Wen performed the research; Chuncao Xu, Hongliang Li, and Yibei Chen assisted with the research; Prof. Min Huang, Prof. Prof. Xueding Wang, and Lie-min Zhou provided appreciated support for the acquisition of subjects. Xupeng Bai and Dingsheng Wen analyzed the data; Xupeng Bai and Prof. Jing Jin wrote the paper.

This study was approved by the Human Investigation Ethics Committee of the School of Pharmaceutical Sciences, Sun Yat-sen University, Guangzhou, China (Clinicaltrials.gov Identifier No. NCT01172626). It was conducted in accordance with the Declaration of Helsinki and was consistent with applicable guidelines of good clinical practice.

1 Bai X, Hong W, Cai P, Chen Y, Xu C, Cao D, Yu W, Zhao Z, Huang M, Jin J, Valproate induced hepatic steatosis by enhanced fatty acid uptake and triglyceride synthesis, Toxicol Appl Pharmacol, 2017, 324, 12, 25, 10.1016/j.taap.2017.03.022

• 2 Biton V, Mirza W, Montouris G, Vuong A, Hammer AE, Barrett PS, Weight change associated with valproate and lamotrigine monotherapy in patients with epilepsy, Neurology, 2001, 56, 172, 177, 10.1212/WNL.56.2.172
• 3 Bokor S, Legry V, Meirhaeghe A, Ruiz JR, Mauro B, Widhalm K, Manios Y, Amouyel P, Moreno LA, Molnàr D, Dallongeville J, HELENA Study group, Single-nucleotide polymorphism of CD36 locus and obesity in European adolescents, Obesity (Silver Spring), 2010, 18, 1398, 1403, 10.1038/oby.2009.412
• 4 Brandl EJ, Tiwari AK, Zai CC, Chowdhury NI, Lieberman JA, Meltzer HY, Kennedy JL, Müller DJ, No evidence for a role of the peroxisome proliferator-activated receptor gamma (PPARG) and adiponectin (ADIPOQ) genes in antipsychotic-induced weight gain, Psychiatry Res, 2014, 219, 255, 260, 10.1016/j.psychres.2014.05.031
• 5 Breum L, Astrup A, Gram L, Andersen T, Stokholm KH, Christensen NJ, Werdelin L, Madsen J, Metabolic changes during treatment with valproate in humans: implication for untoward weight gain, Metabolism, 1992, 41, 666, 670, 10.1016/0026-0495(92)90061-E
• 6 Carmona-Vazquez CR, Ruiz-Garcia M, Pena-Landin DM, Diaz-Garcia L, Greenawalt SR, The prevalence of obesity and metabolic syndrome in paediatric patients with epilepsy treated in monotherapy with valproic acid, Rev Neurol, 2015, 61, 193, 201
• 7 Chang R, Chou MC, Hung LY, Wang ME, Hsu MC, Chiu CH, Study of valproic acid-enhanced hepatocyte steatosis, Biomed Res Int, 2016, 2016, 95765034789392
• 8 Corman CL, Leung NM, Guberman AH, Weight gain in epileptic patients during treatment with valproic acid: a retrospective study, Can J Neurol Sci, 2015, 24, 240, 244, 10.1017/S0317167100021879
• 9 Doerks T, Copley RR, Schultz J, Ponting CP, Bork P, Systematic identification of novel protein domain families associated with nuclear functions, Genome Res, 2002, 12, 47, 56, 10.1101/gr.203201155265
• 10 El-Khatib F, Valproate, weight gain and carbohydrate craving: a gender study, Seizure, 2007, 16, 226, 232, 10.1016/j.seizure.2006.12.009
• 11 Fang J, Chen S, Tong N, Chen L, An D, Mu J, Zhou D, Metabolic syndrome among Chinese obese patients with epilepsy on sodium valproate, Seizure, 2012, 21, 578, 582, 10.1016/j.seizure.2012.06.001
• 12 Farinelli E, Giampaoli D, Cenciarini A, Cercado E, Verrotti A, Valproic acid and nonalcoholic fatty liver disease: a possible association?, World J Hepatol, 2015, 7, 1251, 1257, 10.4254/wjh.v7.i9.12514438499
• 13 Farook VS, Puppala S, Schneider J, Fowler SP, Chittoor G, Dyer TD, Allayee H, Cole SA, Arya R, Black MH, Curran JE, Almasy L, Buchanan TA, Jenkinson CP, Lehman DM, Watanabe RM, Blangero J, Duggirala R, Metabolic syndrome is linked to chromosome 7q21 and associated with genetic variants in CD36 and GNAT3 in Mexican Americans, Obesity (Silver Spring), 2012, 20, 2083, 2092, 10.1038/oby.2012.744287372
• 14 Grygiel-Gorniak B, Kaczmarek E, Mosor M, Przyslawski J, Bogacz A, Genetic background, adipocytokines, and metabolic disorders in postmenopausal overweight and obese women, Biochem Genet, 2016, 54, 636, 652, 10.1007/s10528-016-9743-z5018036
• 15 Hajri T, Hall AM, Jensen DR, Pietka TA, Drover VA, Tao H, Eckel R, Abumrad NA, CD36-facilitated fatty acid uptake inhibits leptin production and signaling in adipose tissue, Diabetes, 2007, 56, 1872, 1880, 10.2337/db06-1699
• 16 Heni M, Müssig K, Machicao F, Machann J, Schick F, Claussen CD, Stefan N, Fritsche A, Häring HU, Staiger H, Variants in the CD36 gene locus determine whole-body adiposity, but have no independent effect on insulin sensitivity, Obesity (Silver Spring), 2011, 19, 1004, 1009, 10.1038/oby.2010.251
• 17 Herken H, Erdal M, Aydin N, Sengul C, Karadag F, Barlas O, Akin F, The association of olanzapine-induced weight gain with peroxisome proliferator-activated receptor-gamma2 Pro12Ala polymorphism in patients with schizophrenia, DNA Cell Biol, 2009, 28, 515, 519, 10.1089/dna.2009.0893
• 18 Jeninga EH, Gurnell M, Kalkhoven E, Functional implications of genetic variation in human PPARγ, Trends Endocrinol Metab, 2009, 20, 380, 387, 10.1016/j.tem.2009.04.005
• 19 Kanemura H, Sano F, Maeda Y, Sugita K, Aihara M, Valproate sodium enhances body weight gain in patients with childhood epilepsy: a pathogenic mechanisms and open-label clinical trial of behavior therapy, Seizure, 2012, 21, 496, 500, 10.1016/j.seizure.2012.05.001
• 20 Keller KL, Liang LCH, Sakimura J, May D, van Belle C, Breen C, Driggin E, Tepper BJ, Lanzano PC, Deng L, Chung WK, Common variants in the CD36 gene are associated with oral fat perception, fat preferences, and obesity in African Americans, Obesity (Silver Spring), 2012, 20, 1066, 1073, 10.1038/oby.2011.374
• 21 Klein KM, Hamer HM, Reis J, Schmidtke J, Oertel WH, Theisen FM, Hebebrand J, Rosenow F, Weight change in monozygotic twins treated with valproate, Obes Res, 2005, 13, 1330, 1334, 10.1038/oby.2005.161
• 22 Laugerette F, Passilly-Degrace P, Patris B, Niot I, Febbraio M, Montmayeur JP, Besnard P, CD36 involvement in orosensory detection of dietary lipids, spontaneous fat preference, and digestive secretions, J Clin Invest, 2005, 115, 3177, 3184, 10.1172/JCI252991265871
• 23 Lett TA, Wallace TJ, Chowdhury NI, Tiwari AK, Kennedy JL, Muller DJ, Pharmacogenetics of antipsychotic-induced weight gain: review and clinical implications, Mol Psychiatry, 2012, 17, 242, 266, 10.1038/mp.2011.109
• 24 Li H, Wang X, Zhou Y, Ni G, Su Q, Chen Z, Chen Z, Li J, Chen X, Hou X, Xie W, Xin S, Zhou L, Huang M, Association of LEPR and ANKK1 gene polymorphisms with weight gain in epilepsy patients receiving valproic acid, Int J Neuropsychopharmacol, 2015, 18, pyv021, 10.1093/ijnp/pyv0214540110
• 25 Love-Gregory L, Sherva R, Schappe T, Qi JS, McCrea J, Klein S, Connelly MA, Abumrad NA, Common CD36 SNPs reduce protein expression and may contribute to a protective atherogenic profile, Hum Mol Genet, 2011, 20, 193, 201, 10.1093/hmg/ddq449
• 26 Luo W, Guo Z, Wu M, Hao C, Hu X, Zhou Z, Zhou Z, Yao X, Zhang L, Liu J, Association of peroxisome proliferator-activated receptor α/δ/γ with obesity, and gene-gene interaction, in the Chinese Han population, J Epidemiol, 2013, 23, 187, 194, 10.2188/jea.JE201201103700259
• 27 Ma X, Bacci S, Mlynarski W, Gottardo L, Soccio T, Menzaghi C, Iori E, Lager RA, Shroff AR, Gervino EV, Nesto RW, Johnstone MT, Abumrad NA, Avogaro A, Trischitta V, Doria A, A common haplotype at the CD36 locus is associated with high free fatty acid levels and increased cardiovascular risk in Caucasians, Hum Mol Genet, 2004, 13, 2197, 2205, 10.1093/hmg/ddh233
• 28 Mansoori A, Amini M, Kolahdooz F, Seyedrezazadeh E, Obesity and Pro12Ala polymorphism of peroxisome proliferator-activated receptor-gamma gene in healthy adults: a systematic review and meta-analysis, Ann Nutr Metab, 2015, 67, 104, 118, 10.1159/000439285
• 29 Miyaoka K, Kuwasako T, Hirano K-i, Nozaki S, Yamashita S, Matsuzawa Y, CD36 deficiency associated with insulin resistance, Lancet, 2001, 357, 686, 687, 10.1016/S0140-6736(00)04138-6
• 30 Noai M, Soraoka H, Kajiwara A, Tanamachi Y, Oniki K, Nakagawa K, Ishitsu T, Saruwatari J, Cytochrome P450 2C19 polymorphisms and valproic acid-induced weight gain, Acta Neurol Scand, 2016, 133, 216, 223, 10.1111/ane.12473
• 31 Noel SE, Lai CQ, Mattei J, Parnell LD, Ordovas JM, Tucker KL, Variants of the CD36 gene and metabolic syndrome in Boston Puerto Rican adults, Atherosclerosis, 2010, 211, 210, 215, 10.1016/j.atherosclerosis.2010.02.0092923842
• 32 Petty SJ, Kantor S, Lawrence KM, Berkovic SF, Collins M, Hill KD, Makovey J, Sambrook PN, O'Brien TJ, Wark JD, Weight and fat distribution in patients taking valproate: a valproate-discordant gender-matched twin and sibling pair study, Epilepsia, 2014, 55, 1551, 1557, 10.1111/epi.12745
• 33 Pickrell WO, Lacey AS, Thomas RH, Smith PE, Rees MI, Weight change associated with antiepileptic drugs, J Neurol Neurosurg Psychiatry, 2013, 84, 796, 799, 10.1136/jnnp-2012-303688
• 34 Pylvänen V, Knip M, Pakarinen A, Kotila M, Turkka J, Isojärvi JI, Serum insulin and leptin levels in valproate-associated obesity, Epilepsia, 2002, 43, 514, 517, 10.1046/j.1528-1157.2002.31501.x
• 35 Rac M, Safranow K, Molecular basis of human CD36 gene mutations, Mol Med, 2007, 13, 1, 10.2119/2006-00088.Rac
• 36 Scherer PE, Adipose tissue: from lipid storage compartment to endocrine organ, Diabetes, 2006, 55, 1537, 1545, 10.2337/db06-0263
• 37 Staeker J, Leucht S, Steimer W, Peroxisome proliferator-activated receptor gamma (PPARG) Pro12Ala: lack of association with weight gain in psychiatric inpatients treated with olanzapine or clozapine, Mol Diagn Ther, 2012, 16, 93, 98, 10.1007/BF03256433
• 38 Su X, Abumrad NA, Cellular fatty acid uptake: a pathway under construction, Trends Endocrinol Metab, 2009, 20, 72, 77, 10.1016/j.tem.2008.11.0012845711
• 39 Tiwari HK, Patki A, Lieberman J, Stroup TS, Allison DB, Leibel RL, Chung WK, Association of allelic variation in genes mediating aspects of energy homeostasis with weight gain during administration of antipsychotic drugs (CATIE study), Front Genet, 2011, 23202977
• 40 Tiwari AK, Brandl EJ, Weber C, Likhodi O, Zai CC, Hahn MK, Lieberman JA, Meltzer HY, Kennedy JL, Müller DJ, Association of a functional polymorphism in neuropeptide Y with antipsychotic-induced weight gain in schizophrenia patients, J Clin Psychopharmacol, 2013, 33, 11, 17, 10.1097/JCP.0b013e31827d145a
• 41 Ujike H, Nomura A, Morita Y, Morio A, Okahisa Y, Kotaka T, Kodama M, Ishihara T, Kuroda S, Multiple genetic factors in olanzapine-induced weight gain in schizophrenia patients: a cohort study, J Clin Psychiatry, 2008, 69, 1416, 1422, 10.4088/JCP.v69n0909
• 42 Verrotti A, Basciani F, Simone M, Trotta D, Morgese G, Chiarelli F, Insulin resistance in epileptic girls who gain weight after therapy with valproic acid, J Child Neurol, 2002, 17, 265, 268, 10.1177/088307380201700405
• 43 Verrotti A, Torre R, Trotta D, Mohn A, Chiarelli F, Valproate-induced insulin resistance and obesity in children, Horm Res, 2009, 71, 125, 131, 10.1159/000197868
• 44 Williams RS, Cheng L, Mudge AW, Harwood AJ, A common mechanism of action for three mood-stabilizing drugs, Nature, 2002, 417, 292, 295, 10.1038/417292a
• 45 Yamamoto Y, Terada K, Takahashi Y, Imai K, Kagawa Y, Inoue Y, Influence of antiepileptic drugs on serum lipid levels in adult epilepsy patients, Epilepsy Res, 2016, 127, 101, 106, 10.1016/j.eplepsyres.2016.08.027
• 46 Zhou J, Febbraio M, Wada T, Zhai Y, Kuruba R, He J, Lee JH, Khadem S, Ren S, Li S, Silverstein RL, Xie W, Hepatic fatty acid transporter Cd36 is a common target of LXR, PXR, and PPARγ in promoting steatosis, Gastroenterology, 2008, 134, 556, 567, 10.1053/j.gastro.2007.11.037
• 47 Zuckermann AM, Ragione RM, Baines DL, Williams RS, Valproic acid protects against haemorrhagic shock-induced signalling changes via PPARγ activation in an in vitro model, Br J Pharmacol, 2015, 172, 5306, 5317, 10.1111/bph.133205123713

PHOTO (COLOR)