Association study between variants of AMP-activated protein kinase catalytic and regulatory subunit genes with antipsychotic-induced weight gain

Journal of Psychiatric Research 46 (2012) 462e468

Contents lists available at SciVerse ScienceDirect

Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires

Association study between variants of AMP-activated protein kinase catalytic and regulatory subunit genes with antipsychotic-induced weight gain Renan P. Souza a, b, c,1, Arun K. Tiwari a, b,1, Nabilah I. Chowdhury a, b, Rolando B. Ceddia d, Jeffrey A. Lieberman e, f, Herbert Y. Meltzer g, h, James L. Kennedy a, b, Daniel J. Müller a, b, * a

Neurogenetics Section, Neuroscience Department, Centre for Addiction and Mental Health, Toronto, ON, Canada Department of Psychiatry, University of Toronto, Toronto, ON, Canada Laboratory of Neurosciences, University of the Far South Catarinense, Criciuma, SC, Brazil d School of Kinesiology & Health Science, Faculty of Health, York University, Toronto, Canada e Department of Psychiatry, College of Physicians and Surgeons, Columbia University, New York City, NY, USA f New York State Psychiatric Institute, New York City, NY, USA g Psychiatric Hospital, Vanderbilt University, Nashville, TN, USA h Department of Psychiatry and Behavioral Sciences, Northwestern University, Feinberg School of Medicine, Chicago, IL, USA b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 October 2011 Received in revised form 15 December 2011 Accepted 5 January 2012

Weight gain and metabolic syndrome are the most common deleterious side effects following treatment with second generation antipsychotic drugs such as clozapine and olanzapine. However, the mechanisms underlying these negative effects of second generation antipsychotic drugs are not fully understood. In this study we investigate whether variants in the genes coding for the a-catalytic (PRKAA1, PRAKAA2) and the b regulatory subunits (PRKAB1 and PRKAB2) of the cellular energy sensor AMP-activated protein kinase (AMPK) are associated with antipsychotic-induced weight gain. To accomplish this, ten polymorphisms in 208 schizophrenia or schizoaffective disorder patients treated with clozapine, haloperidol, risperidone or olanzapine for up to 14 weeks were analyzed. Significant association was observed between rs3766522 in PRKAB2 (AA vs. AT þ TT; p ¼ 0.022) and rs10789038 in PRKAA2 (GG þ GA vs. AA, p ¼ 0.023) with weight change (%) in patients of European ancestry following treatment with clozapine or olanzapine. Allelic association of the T-allele of rs3766522 (p ¼ 0.019) and the G-allele of rs10789038 (p ¼ 0.041) with weight change (%) was also observed. Analysis of raw weight gain revealed that carriers of the T-allele of rs3766522 (AT þ TT, 4.3 kg  3.7) gained more weight than the AA-genotype carriers (2.5 kg  4.5, p ¼ 0.042). Similarly, carriers of the G-allele of rs10789038 (GG þ GA, 4.2 kg  4.5) gained more weight than AA-homozygotes (1.5 kg  2.9, p ¼ 0.014) under antipsychotic treatment. In conclusion, we observed significant associations between polymorphisms in AMPK subunit genes and weight gain induced by clozapine and olanzapine. Ó 2012 Elsevier Ltd. All rights reserved.

Keywords: AMPK Weight gain Association Clozapine PRKAB2

1. Introduction Antipsychotic medications are an important component in the pharmacological management of many psychotic conditions. Although second generation antipsychotics have many notable benefits compared to their older counterparts, the use of some of these drugs has been associated with changes in neuroendocrine signaling pathways that cause weight gain, diabetes (including

* Corresponding author. Pharmacogenetics Research Clinic, Neurogenetics Section, Centre for Addiction and Mental Health, 250 College Street, R30, Toronto, Ontario, Canada M5T 1R8. Tel.: þ1 416 535 8501x6851; fax: þ1 416 979 4666. E-mail address: [email protected] (D.J. Müller). 1 These authors made equal contribution to this study. 0022-3956/$ e see front matter Ó 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.jpsychires.2012.01.010

acute diabetic ketoacidosis), and an atherogenic lipid profile (increased LDL-cholesterol and triglyceride levels and decreased HDL-cholesterol) (Muller and Kennedy, 2006). Despite their serious impact on overall physical and psychological health (Crisp et al., 2000), mechanisms and predictors for antipsychotic-related weight gain are poorly understood. Pharmacogenetic approaches may provide a number of distinct advantages in the search for informative correlates of psychotropic drug response, aiming at a more “individualized” therapy (Basile et al., 2002b; Muller and Kennedy, 2006). Weight gain may be considered a robust phenotype for pharmacogenetic studies as heritability of weight regulation is observed in twin, adoption and family studies (Comuzzie and Allison, 1998). In addition, concordance of weight gain is observed in monozygotic twins and sibling pairs exposed to

R.P. Souza et al. / Journal of Psychiatric Research 46 (2012) 462e468

antipsychotics (Gebhardt et al., 2010; Theisen et al., 2005; Wehmeier et al., 2005). These findings support the role of genetic factors in the development of obesity, as well as in weight gain induced by antipsychotic drugs. Association of variants in candidate genes such as the serotonin receptor 2C (5-HT2C) and leptin have been consistently reported (De Luca et al., 2007a; Ellingrod et al., 2007). In this context, AMPK is a central molecule integrating nutrient and hormonal signals to maintain cellular and whole-body energy homeostasis (Hardie, 2007; Lim et al., 2010; Viollet et al., 2010). AMPK is a well-conserved heterotrimeric serine/threonine kinase consisting of a catalytic a and two (b and g) regulatory subunits. The a and b subunits have two isoforms (a1 and a2; b1 and b2), whereas three isoforms of the g subunit (g1, g2, and g3) exist. These subunits are coded by the PRKAA1/PRKAA2, PRKAB1/ PRKAB2, and PRKAG1/PRKAG2/PRKAG3 genes, respectively. The combinations of these isoforms can theoretically form 12 different heterotrimers (Hardie, 2007; Viollet et al., 2010). Along with splice variants and the use of alternative promoters, these combinations lead to a diverse array of AMPK complexes with distinct tissuespecific distributions and metabolic effects (Viollet et al., 2010). At the cellular level, AMPK functions as a fuel gauge by sensing alterations in AMP/ATP ratio. This kinase is generally activated in conditions that deplete energy (hypoxia, ischemia and low nutrient availability) or accelerate ATP consumption (e.g. exercise) (Hardie, 2007). AMPK is expressed throughout the brain and its activation in the hypothalamus is associated with increased food intake. In fact, the injection of adenovirus coding for either a constitutively active or a dominant negative form of AMPK into the basal medial hypothalamus led to increased and decreased food intake, respectively (Minokoshi et al., 2004). Hypothalamic AMPK activity has been reported to be inhibited by the anorexigenic hormones leptin and insulin, as well as by high glucose and refeeding (Minokoshi et al., 2004). Conversely, ghrelin and cannabionoids up-regulate hypoyhalamic AMPK activity and increase food intake (Andersson et al., 2004; Kola et al., 2005). Additionally, AMPK has been demonstrated to mediate some of the metabolic effects of the widely-used antidiabetic drugs biguanides and thiazolidinediones (Hardie, 2007). Pharmacological studies have also shown that orexigenic antipsychotics such as clozapine, olanzapine, and quetiapine induce AMPK phosphorylation in mice hypothalami, while antipsychotic drugs such as haloperidol, risperidone, ziprasidone, and aripirazole, which are not associated with weight gain, do not elicit this effect (Kim et al., 2007). Furthermore, administration of clozapine at low doses selectively activated AMPK activity in the hypothalamus, but not in cerebral cortex and cerebellum. Clozapine also reversed the reduction in hypothalamic AMPK activity induced by leptin and insulin in mice (Kim et al., 2007). These effects were not observed in mice lacking the Histamine receptor H1 (HRH1) and triprolidine an antagonists H1 receptor stimulated phosphorAMPK to the same extent as clozapine (Kim et al., 2007). Activation of AMPK via HRH1 is interesting as orexigenic antipsychotics (e.g. clozapine, olanzapine) show the highest affinity for H1 receptor (Kim et al., 2007). Furthermore, an earlier study showed that affinity of antipsychotics for H1 receptor was positively correlated with weight gain and the affinity of H1 receptor alone could correctly classify most of the antipsychotic drugs as causing weight gain or not (Kroeze et al., 2003). Altogether, these observations suggest that alterations in the ability of AMPK to exert its metabolic functions in central and peripheral tissues may lead to dysfunctional alterations in food intake and body weight regulation. In fact, association studies have suggested that specific genetic variants in AMPK subunits may alter AMPK function and confer susceptibility to the development of type II diabetes, increase cholesterol levels, and impair glucose and lipid metabolism (Horikoshi et al., 2006; Spencer-Jones et al., 2006). Recently, an

463

intronic polymorphism rs10074991 in the PRKAA1 gene has been associated with change in body mass index (BMI) following w12 weeks of clozapine treatment (Jassim et al., 2011). This SNP along with rs6588640 in PRKAA2 was also associated with BMI at the start of first antipsychotic treatment (Jassim et al., 2011). Therefore, considering that AMPK gene variants have been associated with susceptibility to metabolic changes and that antipsychotic drugs are able to modulate this kinase, we investigated if polymorphisms in AMPK subunits were related to antipsychoticinduced weight gain. The primary hypothesis is that SNPs in AMPK subunit genes will be associated with antipsychotic-induced weight gain in a relatively more homogeneous sample of European ancestry patients treated with medication with the highest risk of weight gain (clozapine and olanzapine). Since other antipsychotics can also lead to weight gain (e.g. risperidone), we also tested a secondary hypothesis that these SNPs may be associated with weight gain irrespective of the weight gain liability of the drug in the European ancestry patients. Here, we provide evidence that weight gain in schizophrenia or schizoaffective disorder patients undergoing treatment with clozapine or olanzapine are associated with polymorphisms in PRKAA2 and PRKAB2. 2. Methods 2.1. Subjects A total of 208 schizophrenia or schizoaffective disorder patients diagnosed according to DSM-III-R or DSM-IV criteria were included in the study. The diagnoses were made or reviewed by experienced psychiatrists (Sample A: DJM, Sample B&C: JAL and HYM). After complete description of the study to the participants, written informed consent was obtained in line with each institution’s research ethics review board guidelines. Detailed demographic and clinical characteristics are provided in Table 1 and have been published previously (Tiwari et al., 2010a,b). Patients for this study were recruited at three different sites: Sample A (Charite University Medicine, Berlin, Germany). Patients 18e60 years old with DSM-IV diagnosis of schizophrenia or schizoaffective disorder and minimum positive and negative symptom scale score of 60 were included. Patients were given antipsychotic medication and assessed up to 6 weeks (Table 1). Exclusion criteria included pregnancy or breast feeding, organic brain disorder, severe head injuries, severe or unstable comorbidities (Hepatitis C, HIV, Thyroid disorder or diabetes mellitus), substance dependence (polytoxicomania), clinically relevant mental retardation and severe personality disorder. Sample B (Case Western Reserve University, Cleveland, Ohio, USA) included schizophrenia patients diagnosed according to DSMIIR criteria (18e60 years) that were either treatment refractory or intolerant to treatment with typical antipsychotics. Before starting treatment with clozapine, a washout period of 7e14 days was allowed. Patients did not receive any co-medication unless it was required for clinical reasons. Clozapine dosage was not fixed and serum levels were monitored to ascertain compliance. Patients were treated with clozapine for 6 weeks and weight change (%) from baseline weight was calculated [for further details see Masellis et al. (1998)]. Exclusion criteria included pregnancy, organic brain disorder, severe head injuries, previous medical conditions which required treatment and were not stable (Hepatitis C, HIV, Thyroid disorder or diabetes mellitus), substance dependence, clinically relevant mental retardation and severe personality disorder. Sample C (Hillside Hospital in Glen Oaks, New York, USA) included 18e60 year old patients with chronic schizophrenia or schizoaffective disorder according to DSM-IV criteria. These patients had earlier shown suboptimal response to previous treatment with typical

464

R.P. Souza et al. / Journal of Psychiatric Research 46 (2012) 462e468

Table 1 Demographic and clinical characteristics of the patients under study. Characteristics Gender Female Male Age Initial body weight (kg) Weight change (kg) Weight change (%) Weight gain <7% 7% Baseline BPRSc Study duration (wk) Ethnicity European-American African-American Others Drugs prescribed Clozapine Haloperidol Olanzapine Risperidone Othersd

Sample Ab n ¼ 79 n (%)

Sample Bb n ¼ 72 n (%)

Sample Cb n ¼ 57 n (%)

Total n ¼ 208 n (%)

31 (39.2) 48 (60.8) 35.10  11.6 80.17  15.7 3.22  3.8 3.90  4.6

27 (37.5) 45 (62.5) 33.28  8.5 74.57  13.6 3.72  4.5 5.32  6.5

9 (15.8) 48 (84.2) 40.47  9.1 84.95  17.4 4.43  6.4 5.73  8.4

67 (32.2) 141 (67.8) 35.94  10.3 79.46  15.9 3.72  4.9 4.89  6.5

59 (74.7) 20 (25.3) 51.10  14.9 5.18  1.5

46 (63.9) 26 (36.1) 51.42  14.4 6.00

37 (66.1) 19 (33.9) 54.01  7.7 11.57  3.9

142 (68.6) 65 (31.4) 52.28  12.5 7.19  3.5

77 (97.5) 2 (2.5) 0

51 (70.8) 21 (29.2) 0

13 (22.8) 33 (57.9) 11 (19.3)

141 (67.8) 56 (26.9) 11 (5.3)

10 6 15 23 25

72 (100) 0 0 0 0

12 11 22 12 0

(12.7) (7.6) (19.0) (29.1) (31.6)

(21.1) (19.3) (38.6) (21.1)

94 17 37 35 25

(45.2) (8.2) (17.8) (16.8) (12.0)

p-value 0.008 <0.001 0.004a 0.636a 0.503a 0.322 0.039a <0.001a <0.001

<0.001

a

KruskaleWallis test. The samples were included from three different clinical studies. c BPRS: Brief Psychiatric Rating Scale. Baseline BPRS scores were not available for a subset of the patients (n ¼ 150). To achieve comparability with sample B, BPRS total scores were extracted from PANSS ratings for samples A and C. d Includes antipsychotic drugs: fluphenazine; aripiprazole, quetiapine, ziprasidone and amisulpride. b

antipsychotic drugs. This was defined by persistent positive symptoms and a poor level of functioning over past two years. Patients were randomly assigned to receive either clozapine (500 mg/day), olanzapine (20 mg/day), risperidone (8 mg/day) or haloperidol (20 mg/day) in a 14 week double blind study. Patients with a history of non-response to clozapine, risperidone, or olanzapine, defined as an unambiguous lack of improvement despite a contiguous adequate trial of risperidone or olanzapine for at least 6 weeks, or clozapine for at least 14 weeks were excluded. Similarly, patients with a history of clozapine, olanzapine, risperidone, or haloperidol intolerance, as well as those who received an intramuscular formulation of antipsychotic medication within 30 days before randomization were also excluded [for further details see Volavka et al. (2002)]. Sample B and C have been described in more detail in previous studies (Basile et al., 2002a; De Luca et al., 2007b; Muller et al., 2005a,b). Ethnicity was determined using ancestry information for the last two generations provided by the participants. The subject was classified as being of European (or African) ancestry if they had all the parents and grand-parents of European (or African) descent. 2.2. Genotyping Blood samples were transported to the Centre for Addiction and Mental Health (CAMH) in Toronto, ON, Canada and genomic DNA was extracted using the high-salt method (Lahiri and Nurnberger, 1991). Single nucleotide polymorphisms (SNPs) were genotyped using TaqManÒ SNP Genotyping Assays (Applied Biosystems Inc, Foster City, CA). Laboratory staff was blind to most of the clinical and demographical information, including weight changes. Ten percent of the total sample was randomly regenotyped to check genotyping accuracy. 2.3. Selection of the polymorphisms Haploview 4.1 (Barrett et al., 2005) was used to select tagSNPs in the PRKAA1, PRKAA2, PRKAB1 and PRKAB2. SNPs in the CEU

population (Utah residents with ancestry from northern and western Europe) of the HapMap database with a minimum allele frequency higher (MAF) than 0.05 and HardyeWeinberg equilibrium (HWE) p > 0.001 were used for tagSNP selection. Three SNPs in each catalytic PRKAA1 and PRKAA2 subunit genes were selected covering 77% and 64%, respectively, of the total variants that have met the MAF criteria mentioned above. Two variants in each regulatory subunit were needed to cover 100% of the variants in both PRKAB1 and PRKAB2 subunit genes. 2.4. Statistical analysis Categorical variables were compared using Pearson’s c2 test and continuous variables were analyzed using Student’s t-test or analysis of variance (Statistical Package for the Social Sciences (SPSS), version 13.0). Association tests between genotype and weight gain were performed using analysis of covariance (ANCOVA) with weight change (%) from baseline as the dependent variable. Genotypes were entered as fixed factor and baseline weight and duration of treatment were used as covariates. UNPHASED version 3.0.10 (Dudbridge, 2003, 2008) was used for haplotype comparisons and Haploview 4.1 (Barrett et al., 2005) was used for calculating linkage disequilibrium (LD). Power calculations were carried out using Quanto 1.2.4 (Gauderman and Morrison, 2006). Assuming a minor allele frequency of 0.15 a sample size of n ¼ 208, we had more than 80% power to detect a mean difference of 2.5% between carriers and non-carriers of the risk genotype in an additive model. 3. Results All the SNPs tested in this study were in HardyeWeinberg equilibrium (p > 0.05, Table 2). High LD but low r2 was observed for all the SNPs except between rs2746355 and rs2796528 (D0 ¼ 0.909 and r2 ¼ 0.791) in the PRKAA2 gene (Supplementary Table 1). The demographic characteristics of the study sample are presented in Table 1. The three clinical sites differed significantly in several clinical variables but showed similar rates of change in

Table 2 Association analysis of SNPs with antipsychotic-induced weight gain at genotypic and/or allelic level. SNP (location)

Genotype

Weight change (%)* all patients (n ¼ 208)

PRKAA1 (Chr5, 38.8 kbp)

rs3805494 (Intron 1)

GG GA AA GG GA AA CC CA AA GG GA AA TT TA AA GG GA AA CC CA AA TT TG GG AA AT TT CC CA AA

5.59 3.85 2.54 5.73 4.80 4.75 4.81 5.30 0.33 5.45 4.96 3.99 4.90 4.80 4.74 4.87 4.86 4.79 4.15 5.19 4.52 0.93 3.42 5.33 4.74 4.99 4.16 4.89 2.77 6.73

rs1002424 (Intron 6)

rs6882903 (Intron 6)

PRKAA2 (Chr1, 63.10 kbp)

rs10789038 (Intron 1)

rs2746355 (Intron 1)

rs2796528 (Intron 1)

PRKAB1 (Chr12, 13.67 kbp)

rs278152 (Intron 5)

rs17408578 (Intron 5)

PRKAB2 (Chr1, 17.44 kbp)

rs3766522 (Intron 2)

rs6689934 (Intron 4)

                             

6.9 (123) 5.2 (64) 8.3 (13) 4.89 (21) 6.5 (89) 6.9 (91) 6.6 (114) 6.2 (80) 6.6 (9) 5.9 (40) 7.0 (104) 5.5 (52) 6.5 (133) 6.6 (66) 5.6 (6) 6.3 (132) 7.2 (63) 5.1 (7) 8.0 (21) 6.0 (109) 6.7 (74) 4.3 (3) 5.2 (40) 6.7 (161) 7.2 (115) 5.4 (77) 5.2 (12) 6.6 (190) 3.3 (9) 3.7 (4)

p-value

Weight change (%)* European Americans, all drugs (n ¼ 141)

p-value

Weight change (%)* European Americans (Ola or Cloz; n ¼ 84)

p-value

Allele

Count (freq)

Mean (Variance)

p-value

HWE p-value

0.215

4.34  5.6 (74) 3.48  5.1 (51) 1.69  8.5 (10) 6.26  3.6 (13) 2.91  4.5 (55) 4.23  6.7 (68) 3.64  5.5 (72) 4.89  5.6 (61) 2.07  4.20 (7) 4.78  5.5 (36) 3.97  6.3 (64) 3.01  4.5 (33) 4.00  5.6 (98) 3.74  5.8 (41) 0.73 (1) 4.05  5.5 (98) 3.62  6.1 (40) 0.73 (1) 3.43  6.8 (10) 4.39  4.8 (78) 3.26  6.6 (52) 0.93  4.3 (3) 3.42  5.2 (40) 4.24  5.9 (139) 2.66  5.6 (68) 5.18  5.5 (61) 4.54  5.3 (11) 3.85  5.8 (127) 2.06  3.2 (7) 6.72  3.7 (4)

0.386

5.44  5.9 3.49  5.7 3.79  3.2 6.53  3.5 3.48  4.4 4.94  6.9 3.71  4.9 6.09  6.4 1.89  3.3 5.41  6.1 5.62  5.9 2.23  4.6 4.83  6.2 3.94  4.6 0.73 (1) 4.99  6.0 3.61  5.2 0.73 (1) 3.59  7.4 4.86  5.4 4.26  5.9 3.69 (1) 3.54  5.4 4.82  5.9 2.94  5.2 6.27  5.9 5.91  4.9 4.66  5.8 0.00 (3)

0.378

G A

117 (0.70) 49 (0.30)

4.79 (31.9) 3.56 (31.9)

0.215

0.239

0.596

G A

49 (0.30) 117 (0.70)

4.35 (32.5) 4.50 (32.5)

0.756

0.903

0.012 0.093a

C A

127 (0.76) 41 (0.24)

4.37 (32.3) 4.92 (32.3)

0.519

0.236

0.074b 0.023a,b

G A

77 (0.48) 83 (0.52)

5.51 (32.1) 3.74 (32.1)

0.041

0.509

0.480a

T A

140 (0.83) 28 (0.17)

4.66 (32.2) 3.71 (32.2)

0.429

0.295

(56) (26)

0.308a

G A

138 (0.83) 28 (0.17)

4.73 (32.5) 3.41 (32.5)

0.293

0.286

(7) (42) (35)

0.830

C A

56 (0.33) 112 (0.67)

4.54 (32.3) 4.48 (32.3)

0.985

0.252

0.407a

T G

22 (0.13) 146 (0.87)

3.55 (32.2) 4.65 (32.2)

0.390

0.673

0.073 0.022a

A T

122 (0.73) 46 (0.27)

3.87 (31.3) 6.18 (31.3)

0.019

0.869

0.219

C A

163 (0.98) 3 (0.02)

4.58 (32.3) 0.00

0.133

0.867

0.968

0.052

0.388

0.952

0.897

0.373

0.169

0.923

0.458

0.146

0.007

0.482

0.750a

0.621a

0.538

0.307a

0.048

0.394

(39) (39) (5) (7) (35) (41) (46) (35) (3) (20) (37) (23) (57) (26)

(20) (63) (44) (34) (6) (80)

R.P. Souza et al. / Journal of Psychiatric Research 46 (2012) 462e468

Gene

*Mean  standard deviation (number of individuals). Significant values are in bold font. Ola or Cloz ¼ Olanzapine or Clozapine treated patients. a SNPs with low number of individuals in a genotypic group were merged with the heterozygous genotypic group. b KruskaleWallis test; HWE: HardyeWeinberg Equilibrium, calculated for the patients with European ancestry on RiskMed (clozapine and olanzapine).

465

466

R.P. Souza et al. / Journal of Psychiatric Research 46 (2012) 462e468

weight (Table 1). Differences in baseline weight can influence the individual weight gain and duration of treatment was different among the clinical sites. Therefore, these variables were entered as a covariate in the association analysis. In the total sample, no differences in distribution of weight change (%) across genotypic groups were observed (n ¼ 208; p > 0.05, Table 2). None of the SNPs were associated with antipsychotic-induced weight gain in the overall sample set (p > 0.05, Table 2). Since the sample consisted of diverse ancestries, we conducted an association analysis in the patients of European ancestry only (n ¼ 141). Significant association was observed between rs6882903 in PRKAA1 (p ¼ 0.007) and rs3766522 in PRKAB2 (p ¼ 0.048) subunits and antipsychoticinduced weight gain in patients of European ancestry. The drugs prescribed at the different clinical sites have different propensities to cause weight gain. Therefore, we further narrowed down on patients treated with only clozapine or olanzapine (n ¼ 84), the drugs with the highest propensities to cause weight gain and other metabolic disturbances (Nasrallah, 2008). Significant association was observed with rs3766522 in PRKAB2 (AA vs. AT þ TT; p ¼ 0.022, Table 2) and rs10789038 in PRKAA2 (GG þ GA vs. AA, p ¼ 0.023) with antipsychotic-induced weight gain. This was also reflected in the allelic analysis where the T-allele of rs3766522 (p ¼ 0.019) and G-allele of rs10789038 (p ¼ 0.041) were associated with weight change (Table 2). In terms of raw weight gain (in kilograms), carriers of the T-allele of rs3766522 (AT þ TT, 4.3 kg  3.7) gained more weight than the AA-genotype carriers (2.5 kg  4.5, p ¼ 0.042). Similarly, carriers of the G-allele of rs10789038 (GG þ GA, 4.2 kg  4.5) gained more weight than AAhomozygotes (1.5 kg  2.9, p ¼ 0.014). A trend of association with weight change (%) was observed for rs6882903 in PRKAA1 (CC vs. CA þ AA, p ¼ 0.093) subunit gene but this was not reflected in the allelic analysis (p ¼ 0.519). We further explored two SNP haplotypes across the genes tested and did not observe significant association of any of the haplotypes (p > 0.05, data not shown). On applying gene-wide correction for multiple independent tests our observations of genotypic association of rs3766522 (pcorrected[0.019  2 tests] ¼ 0.038) and rs10789038 (pcorrected[0.023  2 tests] ¼ 0.046) in the samples of European ancestry on clozapine or olanzapine remained significant. However, none of our observations will be significant if we perform a study-wide correction for multiple testing. If we use the commonly used cut-off criteria of 7% weight gain from baseline, we observed that carriers of the T-allele (AT þ TT) are 3-fold more likely to gain more than 7% weight than individuals with AA-genotype (OR ¼ 3.0 CI: 1.1e8.1; p ¼ 0.027). Similarly, carriers of the G-allele (GG þ GA) of rs10789038 are also over 3-fold more likely to gain more than 7% weight than individuals with the AA-genotype (OR ¼ 3.6, CI: 0.95e13.6, p ¼ 0.049). However, none of the above observations will be significant if a study-wide correction for multiple tests was carried out. 4. Discussion In this study we report the novel findings that intronic SNPs in the genes coding for the catalytic a2 (PRKAA2) and the regulatory b2 (PRKAB2) subunits of AMPK are associated with antipsychoticinduced weight gain in schizophrenia or schizoaffective disorder patients. AMPK plays an important role in the maintenance of whole-body energy homeostasis by regulating central and peripheral mechanisms that control food intake, as well as glucose and lipid metabolism (Hardie, 2007; Viollet et al., 2010). The Nterminus portion of the a-catalytic subunit has a serine/threonine protein kinase domain, while the C-terminal region contains the b1-binding domain required for the formation of the heterotrimer complex (Hardie, 2007). The b subunit acts as a tethering domain for a and g subunits and serves as a scaffold that allows the

assembly of the abg complex (Viollet et al., 2010). The interaction of AMP with the cystathionine b-synthase sequence repeats (CBS domains) of the g regulatory subunit exerts a direct allosteric effect that increases AMPK activity up to 10-fold (Suter et al., 2006). This also promotes the phosphorylation by the upstream kinase LKB1 of a critical threonine residue (Thr-172) within the a-catalytic subunit, which further activates AMPK by up to 100-fold, accounting for the majority of its activity (Suter et al., 2006). Structural alterations in different subunits or defective assembling of the heterotrimer may lead to inability of AMPK to exert its role as an energy sensor and major regulator of substrate metabolism with important implications for whole-body glucose and lipid homeostasis. Therefore, analysis of polymorphisms in the catalytic and regulatory subunits of AMPK could reveal clinically relevant genetic polymorphism(s). In fact, polymorphism and haplotypes in the PRKAA2 have been previously associated with type II diabetes (Horikoshi et al., 2006; Keshavarz et al., 2008), although reports to the contrary are also available (Sun et al., 2006). Furthermore, polymorphisms in the gene coding for the a2 subunit (PRKAA2) of AMPK have been associated with serum concentrations of lipoproteins such as LDLcholesterol and HDL-cholesterol (Spencer-Jones et al., 2006). Mice lacking the a2 subunit of AMPK on a high-fat diet exhibit increased body weight and fat mass, hyperglycemia and insulin resistance compared to wild type mice (Viollet et al., 2003a,b). Thus, the association of the SNP rs10789038 with antipsychotic-induced weight gain found in our study may indeed be of significance. The SNP rs10789038 is not in high correlation with the SNP rs6588640 (r2 ¼ 0.038, HapMap) reported to be associated with BMI before start of first antipsychotic treatment (Jassim et al., 2011). In addition, we did not observe association of rs1002424 in PRKAA1 which is highly correlated with rs10074991 (r2 ¼ 0.951) associated with change in body mass index (BMI) in the study by Jassim et al. (2011). Thus our observation of association is distinct from the previous report (Jassim et al., 2011). However, it is possible that rs10789038 and rs6588640 might be a surrogate for functional polymorphism(s) not yet identified at this locus. The gene PRKAB2 like the PRKAA2 is expressed predominantly in the skeletal and cardiac muscles with a2b2g1complexes exerting most of the metabolic regulatory effects of AMPK in these tissues (Viollet et al., 2010). In addition, the a2 and b2 but not the a2eb1 or a1 subunit, translocate to the nucleus and induce transcription of downstream genes (Suzuki et al., 2007). In fact, AMPK is wellknown for its ability to phosphorylate and activate major transcriptional regulators such as the peroxisome proliferator-activated receptor (PPAR)-g co-activator-a1 (PGC-a1), which promotes mitochondrial biogenesis (Jager et al., 2007; Zong et al., 2002). The SNP rs376652 associated with antipsychotic-induced weight gain is intronic with no known functional significance. Further studies are required to test whether or not the polymorphisms described by us in schizophrenia patients are accompanied by alterations in AMPK activity. A limitation of our study resides in the inclusion of three samples (A, B, and C) that differed in some clinical and demographic characteristics (Table 1). However, these samples included clinically comparable population of patients. Moreover, patients of sample B and sample C were recruited in the early 1990’s when second generation antipsychotics were largely unavailable. Therefore, it is unlikely that the patients received significant amounts of second generation antipsychotics. Other limitations included the reduced gene coverage for the subunits PRKAA1 and PRKAA2 and our relatively small sample size. Although we had enough power to detect clinically relevant weight change in the larger sample set (n ¼ 208), we did not have enough power in the reduced sample of European American on clozapine or olanzapine (n ¼ 84). Another limitation is that none of the above observations will be significant if a study-

R.P. Souza et al. / Journal of Psychiatric Research 46 (2012) 462e468

wide correction for multiple tests was employed. Therefore, it is essential to replicate these observations in larger sample sets. In summary, this study presents novel findings regarding the potential role of variants in the a-catalytic and b-regulatory subunits of AMPK in the development of dysfunctional metabolic alterations in schizophrenia patients. Studies in larger sample sets with SNPs covering the regulatory region of these genes coupled with functional analysis may shed light into their role in antipsychotic-induced weight gain and other related metabolic disorders.

Role of funding source The study was funded by the Canadian Institute of Health Research operating grant (Genetics of antipsychotics induced metabolic syndrome, MOP 89853) to DJM. The funding source had no role in design of the study, analysis of data or writing of the manuscript.

Author contributions The contribution of the authors is as follows: Renan P. Souza carried out the genetic analysis and wrote the first draft of the manuscript; Arun K. Tiwari carried out the statistical analysis and revised the manuscript; Nabilah I. Chowdhury was involved in the genotyping of the SNPs and editing of the manuscript; Rolando B. Ceddia was involved in the editing and gave critical inputs for the study; Jeffrey A. Lieberman and Herbert Y. Meltzer were involved in the collection of the patient sample and revision of the manuscript. James L. Kennedy and Daniel J. Müller were involved in the conception, design of the study, supervised all the genetic and statistical analysis, and edited the manuscript. Conflict of interest RPS/AKT/NC/DJM/RBC: report no competing interests. HYM has received grants or is a consultant to: Abbott Labs, ACADIA, BristolMyers Squibb, Eli Lilly, Jansse, Pfizer, Astra Zeneca, Glaxo Smith Kline, Memory, Cephalon, Minster, Aryx, and BiolineRx. HYM is a shareholder of ACADIA. JAL reports having received research funding or consulting or educational fees from Astra Zeneca, Bristol-Myers Squibb, Eli Lilly, Forest, GlaxoSmithKline, Janssen, Novartis, Pfizer, and Solvay. JLK has been a consultant to GSK, Sanofi-Aventis, Dainippon-Sumitomo.

Acknowledgments RPS has received Canadian Institute of Health Research (CIHR/ Rx&DeWyeth, XWY93967) fellowship. AKT is a recipient of the Center for Addiction and Mental Health (CAMH) postdoctoral fellowship and a NARSAD 2010 young investigator award. RBC is a recipient of the CIHR New Investigator Award. DJM has received a CIHR operating grant (Genetics of antipsychotics induced metabolic syndrome, MOP 89853); a NARSAD Young Investigator Award, a CIHR Michael Smith New Investigator Salary Prize for Research in Schizophrenia, and an OMHF New Investigator Fellowship. JLK received a CIHR operating grant.

Appendix. Supplementary material Supplementary material related to this article can be found online at doi:10.1016/j.jpsychires.2012.01.010.

467

References Andersson U, Filipsson K, Abbott CR, Woods A, Smith K, Bloom SR, et al. AMPactivated protein kinase plays a role in the control of food intake. Journal of Biological Chemistry 2004;279:12005e8. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005;21:263e5. Basile VS, Masellis M, De Luca V, Meltzer HY, Kennedy JL. 759C/T genetic variation of 5HT(2C) receptor and clozapine-induced weight gain. Lancet 2002a;360: 1790e1. Basile VS, Masellis M, Potkin SG, Kennedy JL. Pharmacogenomics in schizophrenia: the quest for individualized therapy. Human Molecular Genetics 2002b;11: 2517e30. Comuzzie AG, Allison DB. The search for human obesity genes. Science 1998;280: 1374e7. Crisp AH, Gelder MG, Rix S, Meltzer HI, Rowlands OJ. Stigmatisation of people with mental illnesses. The British Journal of Psychiatry 2000;177:4e7. De Luca V, Mueller DJ, de Bartolomeis A, Kennedy JL. Association of the HTR2C gene and antipsychotic induced weight gain: a meta-analysis. International Journal of Neuropsychopharmacology 2007a;10:697e704. De Luca V, Muller DJ, Hwang R, Lieberman JA, Volavka J, Meltzer HY, et al. HTR2C haplotypes and antipsychotics-induced weight gain: X-linked multimarker analysis. Human Psychopharmacology 2007b;22:463e7. Dudbridge F. Pedigree disequilibrium tests for multilocus haplotypes. Genetic Epidemiology 2003;25:115e21. Dudbridge F. Likelihood-based association analysis for nuclear families and unrelated subjects with missing genotype data. Human Heredity 2008;66:87e98. Ellingrod VL, Bishop JR, Moline J, Lin YC, Miller del D. Leptin and leptin receptor gene polymorphisms and increases in body mass index (BMI) from olanzapine treatment in persons with schizophrenia. Psychopharmacology Bulletin 2007; 40:57e62. Gauderman WJ, Morrison JM. A computer program for power and sample size calculations for genetic-epidemiology studies, http://hydra.usc.edu/gxe; 2006. Gebhardt S, Theisen FM, Haberhausen M, Heinzel-Gutenbrunner M, Wehmeier PM, Krieg JC, et al. Body weight gain induced by atypical antipsychotics: an extension of the monozygotic twin and sib pair study. Journal of Clinical Pharmacy and Therapeutics 2010;35:207e11. Hardie DG. AMP-activated/SNF1 protein kinases: conserved guardians of cellular energy. Nature Review Molecular Cell Biology 2007;8:774e85. Horikoshi M, Hara K, Ohashi J, Miyake K, Tokunaga K, Ito C, et al. A polymorphism in the AMPKalpha2 subunit gene is associated with insulin resistance and type 2 diabetes in the Japanese population. Diabetes 2006;55:919e23. Jager S, Handschin C, St-Pierre J, Spiegelman BM. AMP-activated protein kinase (AMPK) action in skeletal muscle via direct phosphorylation of PGC-1alpha. Proceedings of the National Academy of Sciences of the United States of America 2007;104:12017e22. Jassim G, Ferno J, Theisen FM, Haberhausen M, Christoforou A, Havik B, et al. Association study of energy homeostasis genes and antipsychotic-induced weight gain in patients with schizophrenia. Pharmacopsychiatry 2011;44: 15e20. Keshavarz P, Inoue H, Nakamura N, Yoshikawa T, Tanahashi T, Itakura M. Single nucleotide polymorphisms in genes encoding LKB1 (STK11), TORC2 (CRTC2) and AMPK alpha2-subunit (PRKAA2) and risk of type 2 diabetes. Molecular Genetics and Metabolism 2008;93:200e9. Kim SF, Huang AS, Snowman AM, Teuscher C, Snyder SH. Antipsychotic druginduced weight gain mediated by histamine H1 receptor-linked activation of hypothalamic AMP-kinase. Proceedings of the National Academy of Sciences of the United States of America 2007;104:3456e9. Kola B, Hubina E, Tucci SA, Kirkham TC, Garcia EA, Mitchell SE, et al. Cannabinoids and ghrelin have both central and peripheral metabolic and cardiac effects via AMP-activated protein kinase. Journal of Biological Chemistry 2005;280: 25196e201. Kroeze WK, Hufeisen SJ, Popadak BA, Renock SM, Steinberg S, Ernsberger P, et al. H1-histamine receptor affinity predicts short-term weight gain for typical and atypical antipsychotic drugs. Neuropsychopharmacology 2003;28:519e26. Lahiri DK, Nurnberger Jr JI. A rapid non-enzymatic method for the preparation of HMW DNA from blood for RFLP studies. Nucleic Acids Research 1991;19:5444. Lim CT, Kola B, Korbonits M. AMPK as a mediator of hormonal signalling. Journal of Molecular Endocrinology 2010;44:87e97. Masellis M, Basile V, Meltzer HY, Lieberman JA, Sevy S, Macciardi FM, et al. Serotonin subtype 2 receptor genes and clinical response to clozapine in schizophrenia patients. Neuropsychopharmacology 1998;19:123e32. Minokoshi Y, Alquier T, Furukawa N, Kim YB, Lee A, Xue B, et al. AMP-kinase regulates food intake by responding to hormonal and nutrient signals in the hypothalamus. Nature 2004;428:569e74. Muller DJ, De Luca V, Sicard T, King N, Hwang R, Volavka J, et al. Suggestive association between the C825T polymorphism of the G-protein beta3 subunit gene (GNB3) and clinical improvement with antipsychotics in schizophrenia. European Neuropsychopharmacology 2005a;15:525e31. Muller DJ, Kennedy JL. Genetics of antipsychotic treatment emergent weight gain in schizophrenia. Pharmacogenomics 2006;7:863e87. Muller DJ, Klempan TA, De Luca V, Sicard T, Volavka J, Czobor P, et al. The SNAP-25 gene may be associated with clinical response and weight gain in antipsychotic treatment of schizophrenia. Neuroscience Letters 2005b;379:81e9.

468

R.P. Souza et al. / Journal of Psychiatric Research 46 (2012) 462e468

Nasrallah HA. Atypical antipsychotic-induced metabolic side effects: insights from receptor-binding profiles. Molecular Psychiatry 2008;13:27e35. Spencer-Jones NJ, Ge D, Snieder H, Perks U, Swaminathan R, Spector TD, et al. AMPkinase alpha2 subunit gene PRKAA2 variants are associated with total cholesterol, low-density lipoprotein-cholesterol and high-density lipoproteincholesterol in normal women. Journal of Medical Genetics 2006;43:936e42. Sun MW, Lee JY, de Bakker PI, Burtt NP, Almgren P, Rastam L, et al. Haplotype structures and large-scale association testing of the 50 AMP-activated protein kinase genes PRKAA2, PRKAB1, and PRKAB2 [corrected] with type 2 diabetes. Diabetes 2006;55:849e55. Suter M, Riek U, Tuerk R, Schlattner U, Wallimann T, Neumann D. Dissecting the role of 50 -AMP for allosteric stimulation, activation, and deactivation of AMPactivated protein kinase. Journal of Biological Chemistry 2006;281:32207e16. Suzuki A, Okamoto S, Lee S, Saito K, Shiuchi T, Minokoshi Y. Leptin stimulates fatty acid oxidation and peroxisome proliferator-activated receptor alpha gene expression in mouse C2C12 myoblasts by changing the subcellular localization of the alpha2 form of AMP-activated protein kinase. Molecular and Cellular Biology 2007;27:4317e27. Theisen FM, Gebhardt S, Haberhausen M, Heinzel-Gutenbrunner M, Wehmeier PM, Krieg JC, et al. Clozapine-induced weight gain: a study in monozygotic twins and same-sex sib pairs. Psychiatric Genetics 2005;15:285e9. Tiwari AK, Rodgers JB, Sicard M, Zai CC, Likhodi O, Freeman N, et al. Association study of polymorphisms in cholecystokinin gene and its receptors with antipsychotic induced weight gain in schizophrenia patients. Progress in NeuroPsychopharmacology and Biological Psychiatry 2010a;34:1484e90.

Tiwari AK, Zai CC, Likhodi O, Lisker A, Singh D, Souza RP, et al. A common polymorphism in the cannabinoid receptor 1 (CNR1) gene is associated with antipsychotic-induced weight gain in Schizophrenia. Neuropsychopharmacology 2010b;35:1315e24. Viollet B, Andreelli F, Jorgensen SB, Perrin C, Flamez D, Mu J, et al. Physiological role of AMP-activated protein kinase (AMPK): insights from knockout mouse models. Biochemical Society Transactions 2003a;31:216e9. Viollet B, Andreelli F, Jorgensen SB, Perrin C, Geloen A, Flamez D, et al. The AMPactivated protein kinase alpha2 catalytic subunit controls whole-body insulin sensitivity. Journal of Clinical Investigation 2003b;111:91e8. Viollet B, Horman S, Leclerc J, Lantier L, Foretz M, Billaud M, et al. AMPK inhibition in health and disease. Critical Reviews in Biochemical and Molecular Biology 2010;45:276e95. Volavka J, Czobor P, Sheitman B, Lindenmayer JP, Citrome L, McEvoy JP, et al. Clozapine, olanzapine, risperidone, and haloperidol in the treatment of patients with chronic schizophrenia and schizoaffective disorder. American Journal of Psychiatry 2002;159:255e62. Wehmeier PM, Gebhardt S, Schmidtke J, Remschmidt H, Hebebrand J, Theisen FM. Clozapine: weight gain in a pair of monozygotic twins concordant for schizophrenia and mild mental retardation. Psychiatry Research 2005;133:273e6. Zong H, Ren JM, Young LH, Pypaert M, Mu J, Birnbaum MJ, et al. AMP kinase is required for mitochondrial biogenesis in skeletal muscle in response to chronic energy deprivation. Proceedings of the National Academy of Sciences of the United States of America 2002;99:15983e7.