A Systematic Review of Antipsychotic Drug Effects on Human Gene Expression Related to Risk Factors for Cardiovascular Disease

Home , INSIG2, Mesoridazine, Triflupromazine

ORIGINAL ARTICLE A systematic review of antipsychotic drug effects on human gene expression related to risk factors for cardiovascular disease

DL Foley and A Mackinnon

Psychosis is associated with an elevated risk for cardiovascular disease. We reviewed evidence for a causal association between experimentally controlled antipsychotic drug exposure and a change in the expression of genes relevant to cardiovascular disease in human cell lines. Reports from SCOPUS - V.4 (Elsevier) and MEDLINE (ISI) were assessed for global or candidate gene expression analysis, tissue and cell type, tissue source or cell line, antipsychotic drug and dosage, length of drug exposure and statistically signiﬁcant fold change in gene expression after drug exposure; 29 eligible studies analysed gene expression in the brain, eye (as a model of neuronal cells), heart, kidney (as a model of any cell), liver, pancreas or skin. Antipsychotic drugs alter the expression of numerous genes related to cardiovascular health, including genes under the control of the sterol regulatory element binding protein transcription factors that control lipid and fatty acid biosynthesis.

The Pharmacogenomics Journal (2014) 14, 446–451; doi:10.1038/tpj.2014.8; published online 11 March 2014

INTRODUCTION Other genetic mechanisms may explain a larger proportion of Cardiovascular disease is the leading cause of death worldwide1 and cardiovascular risk associated with antipsychotics. Drugs bind to individuals with schizophrenia are twice as likely to die from transcription factors and thereby alter the expression of the genes cardiovascular disease as the wider community.2 Risk factors begin under their control. Antipsychotics could alter cardiovascular risk to change within several weeks of first exposure to antipsychotics profile by targeting the transcription of genes associated with and include a visible change in body weight and clinically signi- cardiovascular health. The cumulative effect of a treatment- ficant changes in cholesterol, triglyceride and glucose levels.3,4 The induced gene expression cascade could generate a large change impact on quality of life is profound.5,6 Randomized clinical trials in risk profile and is also consistent with the time course of 13 have established that the degree of change varies by drug and antipsychotic drug effects. Treatment-induced genetic effects duration of exposure.4 Cigarette smoking, a poor diet and may be an important aspect of antipsychotic drug action that sedentariness are prevalent among individuals with psychosis but influences the development of cardiovascular side effects but the the cumulative effect of lifestyle risk factors cannot explain the rapid relevant literature has not previously been reviewed. changes observed after the first exposure to antipsychotic drugs; The aim of this study is to systematically review evidence for there appears to be a direct effect of the antipsychotics themselves. antipsychotic drug effects on human gene expression associated One possible explanation is a direct effect of antipsychotics on with risk factors for cardiovascular disease. neuronal receptor binding profiles. Antipsychotics modulate neuronal receptors,7 altering appetite, satiety and peripheral metabolism as well as neurotransmission.8 All antipsychotics modulate dopamine (especially dopamine 2) receptors in the brain. Clozapine MATERIALS AND METHODS and olanzapine also modulate serotonin, glutamate, histamine, Original studies of antipsychotic drug effects on human gene expression alpha-adrenergic and muscarinic receptors7 and are associated with related to cardiovascular risk were selected for review. Only studies using the greatest weight gain.9 Weight gain has been associated with an experimentally controlled dose of a single antipsychotic drug for a known period of time were included to ensure accurate attribution of the antipsychotic blockade (antagonism) of serotonin receptor 1A and cause of any observed change in gene expression. Drug dose was selected 2C (HTR1A, HTR2C), histamine 1 (HRH1, previously known as H1)and 7 based on cell viability and estimated clinical relevance by study authors. dopamine receptor 2. Diabetes has been associated with Any study using human cell lines or tissue was eligible. Autopsy and living antipsychotic antagonism of muscarinic receptor 3 (CHRM1, whole-person studies were excluded because lifetime exposure to drugs, previously known as M1).7 Histamine and muscarinic receptor multiple drug exposures and the moderating effects of other variables antagonism has been hypothesised to explain glucose could not be controlled. abnormalities, and histamine antagonism has been hypothesised Eligible reports were retrieved from SCOPUS-V.4 (Elsevier) and to explain increased food intake and peripheral effects on MEDLINE (ISI) (from inception through August, 2013). Factors related to metabolism.8 Common DNA sequence variants could moderate cardiovascular risk profile in patients with their first treated episode of psychosis defined the search string (see Supplementary e-Table 24): effects of antipsychotics and have been reviewed in relation to 10 ‘antipsychotic gene expression’ AND (weight OR BMI OR waist OR ‘central antipsychotic-induced weight gain. Promising candidates included obesity’ OR adipose OR fat OR glucose OR insulin OR triglycerides OR variants in HTR2C and the dopamine transporter (SLC6A3, previously cholesterol OR ‘blood pressure’ OR ‘heart rate’ OR cortisol OR exercise OR known as DAT1) but these explained only a small proportion of leptin OR adiponectin OR interleukin OR VCAM OR selectin OR c-peptide observed weight gain,10 reminiscent of the small effects of common or QT interval OR creatinephosphokinase OR EKG OR grehlin OR resistan DNA sequence variants on risk for psychosis.11,12 OR visfatin OR cardiovascular). Eligible studies were divided into two sets.

Orygen Youth Health Research Centre, Centre for Youth Mental Health, University of Melbourne, Parkville, Victoria, Australia. Correspondence: Dr DL Foley, Orygen Youth Health Research Centre, 35 Poplar Road, Parkville, Victoria 3052, Australia. E-mail: [email protected] Received 15 December 2013; revised 22 January 2014; accepted 31 January 2014; published online 11 March 2014 Antipsychotic effects on human gene expression DL Foley and A Mackinnon 447

Figure 1. Global gene expression studies.

Global expression studies studies that could not control drug exposure experimentally, Global expression studies identified the genes with the largest change in including autopsy studies and those that combined autopsy and expression after exposure to antipsychotics, genes that were preferentially surgically removed tissue from living patients, (4) studies where an targeted by antipsychotics (Figure 1). Targeted replication studies tested if antipsychotic was considered only as a moderator of the effect of expression of those genes changed after exposure to other antipsychotics, another substance on gene expression and (5) studies where in different cell lines or tissues or over variable time periods. Global human experimental studies were described but antipsychotic- expression studies also assessed the change in total gene expression to induced gene expression was conducted only in animal cell lines. estimate the impact of antipsychotic drugs on cellular DNA repair mechanisms. This identified the degree of compromised DNA repair across all actively transcribed genes. Global expression studies in the brain Fourteen genes related to lipid and fatty acid biosynthesis were Candidate gene expression studies preferentially upregulated in a global expression analysis in glia Candidate gene studies tested apriorihypotheses about the effects of after exposure to clozapine/haloperidol.15 This was due to antipsychotics on expression changes in prespecified genes (Figure 2). These increased proteolytic cleavage of the sterol regulatory element were often associated with cell signalling pathways involving Akt that is binding protein (SREBP) transcription factors SREBP1 and SREBP2 strongly implicated in cardiovascular health and disease. These pathways are therefore briefly noted where it illustrates the connections among studies. (encoded by the SREBF1 and SREBF2 genes) that control Data extracted were global versus replication versus candidate expression of these genes, especially SREBP2. SREBP2 expression analysis, tissue and cell type, tissue source or cell line, preferentially activates the genes that control the synthesis and antipsychotic drug and dosage, length of drug exposure, gene for which uptake of cholesterol (Figure 1, Supplementary Table 1). data were reported, expression-fold change in that gene after drug Genes under the control of SREBP were also preferentially exposure and the associated standard error and exact P-value when upregulated in a second global expression analysis in glia and in 14 reported (Supplementary Tables 1 and 2). DataThief III version 1.6 was (a model for) neurons, across 18 different antipsychotics: used to extract data points from published bar charts and graphs if the aripiprazole, chlorpromazine, clozapine, fluphenazine, haloperidol, numeric fold change was not reported. The HUGO Gene Nomenclature iloperidone, loxapine, mesoridazine, molindone, olanzapine, Committee database was used to define gene symbols and names in perphenazine, promazine, quetiapine, risperidone, thioridazine, reviewed studies; previous aliases can be found on the gene cards website 16 (http://www.genecards.org/) and noted here only when that symbol or trifluoperazine, triflupromazine and ziprasidone. The most name was reported by cited studies. Gene function and disease downregulated genes had no common known function and associations were extracted from gene cards, tumour genes from some had no known function.16 www.tumor-gene.org/ and genes that were the subject of commercial Upregulation of genes related to cholesterol metabolism or genetic tests from www.genetests.org/ (Supplementary Table 3). transport by antipsychotic drugs was independently replicated by two targeted expression studies. These focussed on cholesterol- related genes because of the greater effect of antipsychotics on RESULTS SREBP2 than SREBP1 controlled gene expression in the first global Our search string identified 241 articles of which 29 met eligibility study.15 Genes that control cholesterol metabolism were criteria.15–37 We excluded (1) hypothesis and review papers, (2) preferentially upregulated in the brain in multiple glial and studies of mood stabilisers (lithium or valproate) or dopamine neuronal cell lines after exposure to chlorpromazine, clozapine, antagonists that were not typical or atypical antipsychotics, (3) haloperidol, olanzapine or risperidone.18,19

& 2014 Macmillan Publishers Limited The Pharmacogenomics Journal (2014), 446 – 451 Antipsychotic effects on human gene expression DL Foley and A Mackinnon 448

Figure 2. Candidate gene expression studies.

Another global study looked at the total product (RNA Candidate gene expression studies of the hypothesised recovery) of all actively transcribed genes in three skin neuroprotective effects of antipsychotic drugs in the brain (fibroblast) cell lines after exposure to antipsychotics.17 One Exposure to olanzapine and clozapine, but not haloperidol, cell line was especially sensitive to DNA damage by ultraviolet increased the expression of the AKT1 gene in neurons, which irradiation and was therefore used to model the effect of may be a neuroprotective effect.21 AKT1 is one of the three antipsychotics in the presence of compromised DNA repair isoforms of Akt. AKT1 controls cell cycle progression, inhibits mechanisms. In the normal (control) cell lines, recovery of RNA apoptotic processes (cell death), induces protein synthesis synthesis occurred within 1 h of ultraviolet irradiation after pathways and promotes cell survival and general tissue growth. exposure to chlorpromazine and haloperidol. The ultraviolet- AKT1 is therefore implicated in neuroprotection (adaptive cell sensitive cell line showed markedly depressed RNA recovery growth in the brain) and cancer (uncontrolled cell growth) after exposure to chlorpromazine and haloperidol. Complete (Figure 2, Supplementary Table 2). recovery was not reached even after 24 h. Antipsychotics may Exposure to clozapine but not haloperidol increased expression therefore adversely affect DNA repair in cell lines already of NRG1 and VAMP1 but not SNAP25 in a human brain aggregate sensitive to DNA damaging agents. containing both glia and neurons.22 This also was interpreted as a neuroprotective effect.22 NRG1 exerts its effects via the Ras–mitogen-activated protein kinase and phosphoinositide Replication in peripheral tissue 3-kinase–Akt signalling pathways through ErbB receptors. Expression of LDLR, a key cholesterol regulating gene under the Exposure to haloperidol increased expression of the IDE gene in control of SREBP2, was significantly increased in liver cells by neurons at low doses but decreased expression at higher doses.23 olanzapine.20 Expression of APOE was significantly increased in Decreased IDE expression would increase insulin levels. This was liver by clozapine, olanzapine and haloperidol.19 Expression of interpreted as neuroprotective, insulin is known to improve ABCA1 was significantly increased in liver by clozapine but not by learning and memory and as possibly related to adverse olanzapine or haloperidol.19 Expression of NPC1L1 was metabolic side effects because insulin affects hunger and satiety significantly increased in liver by haloperidol but not by signals.23 Activated Akt/protein kinase B mediates several clozapine or olanzapine.19 APOE, ABCA1 and NPC1L1 are all metabolic effects of insulin including glucose uptake. cholesterol transport genes under the control of the liver X Exposure to aripiprizole but not haloperidol increased BDNF, receptors (NR1H2, previously known as LXRa, and NR1H3, GSK3b and BCL2 expression in neurons.24 This was associated with previously known as LXRb; Figure 1, Supplementary Table 1). activation of the mitogen-activated protein kinase and Variation in the length of drug exposure produced different phosphoinositide 3-kinase/Akt pathways and subsequent patterns of upregulated genes.19 The expression of cholesterol activation of the transcription factor CREB and interpreted as transport genes controlled by the liver X receptor were upregulated possibly neuroprotective. Exposure to olanzapine also increased after 24–48 h of drug exposure, a significant delay relative to the BDNF expression in neurons.25 This also involved CREB-mediated activation of genes under the control of SREBP which occurred after gene transcription through several kinases including protein 6 h. Liver X receptors modulate the expression of SREBP1. kinase A, phosphoinositide 3-kinase, protein kinase C and Ca2 þ / The expression of SREBF2, HMGR, LDLR, SREBF1 and FASN in calmodulin dependent protein kinase II.25 liver was upregulated by haloperidol, clozapine, risperidone and Sertindole and pimozide blocked the neuronal potassium ziprasidone, variably at various doses, and expression of ABCA1 in channel gene KCNH7 (previously known as hERG-3).28 This liver was upregulated by haloperidol, risperidone and ziprasidone was interpreted as possibly inhibiting neurons with higher but not clozapine.33 A potential mechanism for the hepatic firing frequencies.28 Akt regulates voltage-dependent potassium antipsychotic drug side effects is the induction of endoplasmic channels. reticulum stress that activates the protein kinase R-like Exposure to haloperidol and quetiapine decreased expression endoplasmic reticulum kinase (PERK) dependent branch of the of several key enzymes of antioxidant metabolism in neurons unfolded protein response, which may then induce the SREBP1 including SOD1, SOD2, GPX, CAT, GST, GGT and GSR.26 Effects were and 2 pathways resulting in accumulation of lipids in not in the hypothesised direction. Reduced antioxidant enzyme hepatocytes.36 activity is associated with increased oxidation (increased oxidative

The Pharmacogenomics Journal (2014), 446 – 451 & 2014 Macmillan Publishers Limited Antipsychotic effects on human gene expression DL Foley and A Mackinnon 449 stress), which may cause neural damage. Increased oxidative heart disease, diabetes, multiple cancers and Alzheimer’s disease. stress inhibits the Akt pathway. Eleven genes were the subject of commercial diagnostic tests Exposure to clozapine decreased expression of CCND1 in glia, in unrelated to psychosis (Supplementary Table 3). a cell line that lacks PTEN (a tumour suppressor gene) activity, and arrested cell cycle progression at G0/G1 phase.27 This was interpreted as evidence clozapine may modulate the PI3K/Akt/ DISCUSSION GSK-3b pathway because it controls CCND1 expression and as Antipsychotic drugs alter the expression of numerous genes possibly relevant for tumour suppression.27 related to cardiovascular health. Many of these genes are Clozapine and risperidone inhibited T-bet expression, critical for associated with cell signalling pathways involving Akt, which is a Th1 differentiation, clozapine enhanced the expression of STAT6 molecular sensor implicated in both cardiovascular disease and and GATA3, critical for Th2 differentiation, haloperidol did not, cancer.38 Antipsychotic drugs are therefore a potentially important suggesting variation in the immunomodulatory (and therefore source of gene expression-mediated risk for cardiovascular disease, neuroprotective) effects of different antipsychotic drugs.35 and they identify gene targets that affect multiple phenotypes. Increased lipogenesis is a risk factor for cardiovascular disease and Candidate gene expression studies of hypothesised adverse side a major hallmark of cancer cells. This systematic review has effects of antipsychotics in the brain therefore uncovered a potentially important nexus of pleiotropic Exposure to haloperidol but not amisulpride increased expression of genes that may help explain why individuals with psychosis have the dopamine 2 receptor gene in neurons.30 Exposure to lipoic acid an elevated risk for cardiovascular disease, in part induced by for 24 h before exposure to haloperidol reversed the effect. Lipoic antipsychotics, and an altered risk for cancer in the form of a lower risk for lung, colorectal and prostate cancer but a higher risk for acid reduces oxidation of lipids and therefore reduces oxidative 39 stress. This was interpreted as evidence of a mechanism (oxidative breast cancer. Pleiotropy may therefore link important aspects of stress plus dopamine receptor 2 upregulation) that may underlie the antipsychotic drug efficacy, the development of adverse side development of tardive dyskinesia, a neurological side effect of effects and the altered cancer risk associated with psychosis. antipsychotics.30 Akt is a downstream target of dopamine receptor signalling. Dopamine receptor 2 receptors are essential for the The global perspective: lipids and fatty acids, and DNA repair inhibition of Akt by dopamine (Figure 2, Supplementary Table 2). Global expression studies showed that antipsychotic drugs Exposure to haloperidol did not alter the expression of UBL5 preferentially upregulate genes that control the synthesis and (ubiquitin-like 5 also known as the beacon gene) in neurons.31 This regulation of lipids and fatty acids in the brain. They do this by was interpreted as evidence that a difference in UBL5 expression activating the SREBP transcription factors, the master regulators of in brains of normal weight controls versus obese controls and lipid and fatty acid biosynthesis.40,41 These findings were haloperidol-treated individuals with schizophrenia was not due to replicated in four different cancer cell lines derived from glia treatment with antipsychotics. UBL5 is involved in the regulation and neurons, and in a retinal model of neurons, which was not of energy metabolism, food intake and obesity. derived from cancerous tissue. This is a strong evidence the global findings are not cell line or cell type specific and that they are not Candidate gene expression studies of hypothesised adverse side simply a byproduct of the unusual characteristics of cell lines effects of antipsychotics in peripheral tissue derived from cancerous tissues. Other studies showed that Gene expression altered by antipsychotic drugs in the brain may antipsychotics also activate SREBP in the liver and, when drug also be altered in peripheral tissues (Figure 1), but the expression of exposure is extended, the liver X receptor and consequently the only one candidate gene KCNH2 (previously known as hERG)has genes under their control. Liver X receptors are cholesterol- been analysed in brain and peripheral tissue after exposure to sensing nuclear receptors that also regulate lipid metabolism and antipsychotics. Mesoridazine blocked KCNH2 in the kidney cell line transport. HEK293, consistent with the development of long QT syndrome.29 Exposure to antipsychotics also decreased the total expression Blockage of KCNH2 by antipsychotics prolongs the QT interval and of all actively transcribed genes in cells already sensitive to DNA can cause sudden cardiac death. Although HEK293 is derived from damage. This may represent an important multiple hit model for kidney tissue it is typically used as a model of any cell, and it has individuals taking antipsychotics and exposed to the polycyclic been stably transfected with KCNH2 cDNA (Figure 2, Supplementary aromatic hydrocarbons in cigarette smoke or other environmental Table 2). toxins that compromise DNA repair.

Candidate gene expression studies of antipsychotic drug targets The candidate gene perspective: Akt and cell signal transduction in peripheral tissue pathways A screen of 1040 drugs showed that a subset of phenothiazine Antipsychotic drugs also modulate signal transduction pathways, antipsychotics modulated the expression of the insulin gene in in part by altering the expression of transcription factors and 37 genes associated with those pathways and in part by direct pancreas. Other antipsychotic drugs also increased insulin gene 21 expression in pancreas and this effect could be entirely explained activation of the signalling kinase Akt. Akt affects glucose metabolism, insulin resistance, risk for type 2 diabetes42 and many by SMAD3, which is a downstream effector of the transforming 38,43 growth factor beta pathway.32 SMAD3 has been implicated in the aspects of cardiovascular health and disease. The three highly related isoforms of Akt, AKT1, AKT2 and AKT3, are all expressed in regulation of adiposity (obesity), glucose tolerance and insulin 38 sensitivity. Akt (AKT2) is an important signalling molecule in the the heart. insulin signalling pathway (Figure 2, Supplementary Table 2). Altered gene expression in the brain after exposure to antipsychotics was associated with cell growth and cell death. Clozapine but not haloperidol upregulated the expression of 44 FOXA1, a hepatic transcription factor, and its regulated (lipogenic) The latter is consistent with reduced brain volume. Antipsychotic gene CIC in liver.34 drugs associated with the greatest weight gain and the worst metabolic side effects, clozapine and olanzapine, signiﬁcantly upregulated AKT1, whereas haloperidol did not. An increased An overview of gene function and gene-disease associations expression of AKT1 could be neuroprotective in the brain but, if Genes whose expression was altered by exposure to antipsychotic sustained, could also damage the heart.43 Antipsychotic drugs drugs included many genes known to be associated with coronary upregulate the insulin gene in the pancreas and the insulin-

& 2014 Macmillan Publishers Limited The Pharmacogenomics Journal (2014), 446 – 451 Antipsychotic effects on human gene expression DL Foley and A Mackinnon 450 degrading enzyme gene in the brain. If antipsychotic drugs also Implications for the development of cardiovascular side effects alter insulin levels in the heart this would have a direct negative Activation by antipsychotic drugs of genes associated with hepatic 45 effect on heart muscle mediated primarily through Akt. Some lipid homoeostasis may partly explain the altered risk for antipsychotic drugs regulate KCNH (hERG) genes and block cardiovascular disease after first exposure to antipsychotics. The voltage-dependent potassium channels that can lead to an genes associated with lipid homoeostasis whose expression elongation of the QT interval and sudden cardiac death. profiles were most strongly altered by antipsychotics may be useful peripheral markers of emerging cardiovascular side effects. Change in their expression may provide an evidence-based trigger Gaps in our perspective for drug switching or dose adjustment including trials of non- 54 Global expression analyses should be conducted in tissues relevant continuous exposure. Genes regulated by antipsychotics also to the development of cardiovascular disease to determine if they provide a potential peripheral target against which proposed mirror the pattern and degree of altered regulation of genes under interventions to improve cardiovascular health can be tested. the control of SREBP in the brain. No study has characterised gene expression in the heart after exposure to antipsychotics. Phos- Caveats and sources of possible bias phoinositide 3-kinase/Akt may activate SREBPs46,47 but no study has examined the effect of exposure to antipsychotics on AKT2 or (1) The cell lines used as model systems are not models of AKT3 expression in any tissue or AKT1 expression anywhere but schizophrenia. This is important if disease-specific effects glia. The effects of antipsychotic drugs on the expression of all Akt moderate drug related effects. CHRM3 (previously known as isoforms need to be characterised and the circumstances under M3) antagonism associated with risk for diabetes in schizo- which antipsychotics activate or inhibit phosphoinositide 3-kinase/ phrenia, for example, is not associated with risk for diabetes in Akt. As insulin regulates glucose and lipid metabolism, and risk for other patient populations.7 diabetes is increased among those with treated psychosis, the (2) The entire gene expression cascade induced by antipsychotics regulation of insulin by antipsychotics, and forkhead transcription may not be captured by the studies reviewed here (for factors, requires further study. example, immediate early genes).55 Testing if common environmental toxins make DNA more (3) Long term effect of antipsychotic drugs, at varying doses with sensitive to damage from antipsychotics is an important area for or without breaks and often in combination with other drugs, further study because compromised DNA repair is associated with 48 cannot be inferred from short-term experimental models. a shortened life span, rapid ageing and multiple disease risks. (4) The therapeutic dose for antipsychotic drugs is not defined. Antipsychotic drug levels vary across tissues and determining what constitutes a therapeutic dose (equivalent) in cell lines is Implications for the aetiology and treatment of schizophrenia a pragmatic decision. The activation by antipsychotic drugs of genes associated with (5) The pattern and severity of adverse antipsychotic drug effects fatty acid and lipid homoeostasis may be the common central vary but moderators of those effects on gene expression, mechanism by which they achieve their antipsychotic activity.16,49 including baseline (pretreatment) levels of gene expression, SREBP in glia control the synthesis of lipids and fatty acids that are are unknown. involved in a variety of glia–neuron interactions and neuronal (6) The proportion of change in cardiovascular risk profile that functions.50 The myelination theory of schizophrenia ascribes could be explained by gene expression changes induced by perturbed neurodevelopment and/or neurotransmission to subtle antipsychotic drugs is unknown. disruptions in myelination. Cholesterol is an indispensable component of myelin membranes. Cholesterol availability in One report has extended the studies reviewed here into clinical oligodendrocytes, a type of glial cell, is a rate-limiting factor for medicine. Vik-Mo et al.56 analysed peripheral blood from patients brain maturation51 and SREBP2 controls the cholesterol synthesis with psychosis. Expression of two lipid biosynthesis genes, FAS pathway during myelination.50 Upregulating genes that control and SCD, identified by global expression studies were upregulated cholesterol and fatty acids in the brain may correct an existing in peripheral blood from olanzapine-treated patients compared abnormality among individuals with psychosis. Akt also with unmedicated patient controls. This is an important proof of contributes to the acute regulation of cholesterol metabolism principle. through activating SREBP2.47 The cell growth and proliferation activated by Akt are supported by SREBP cholesterol production, which is required for the creation of new cell membranes.47 Akt is CONFLICT OF INTEREST a positive effector that can amplify this process whereas sterols The authors declare no conflict of interest. like cholesterol are the over-riding negative regulator.47 Akt, the liver X receptor and SREBF, and the LDLR and ABCA1 genes all interact in glia.52 Factors that regulate SREBP, the liver X receptor ACKNOWLEDGMENTS and Akt in vivo may therefore provide novel insights into the We wish to acknowledge the support of the Colonial Foundation (Australia) to the aetiology of psychosis and new targets for therapy. Common gene Orygen Youth Health Research Centre. variants individually explain very little of the risk for schizophrenia but characterising the effects of rare variants in the genes that encode SREBP and the three Akt isoforms may prove informative. REFERENCES AKT1, SREBF1, and SREBF2 are candidate genes for schizophrenia. 1 World Health Organization. 2011; http://www.who.int/mediacentre/factsheets/ SREBF2, and genes under the control of SREBP including SCAP and fs310/ en/index.html. INSIG2, are candidate genes for antipsychotic-induced weight 2 Saha S, Chant D, McGrath J. A systematic review of mortality in schizophrenia: is gain.10 AKT1 is a known cancer gene and SREBF2 and INSIG2 when the differential mortality gap worsening over time? Arch Gen Psychiatry 2007; 64: mutated are implicated in cancer.39 1123–1131. 3 The Lancet. No mental health without physical health. Lancet 2011; 377: 611. Change in biomarkers thought to be relevant to the aetiology of 4 Foley DL, Morley KI. Systematic review of early cardiometabolic outcomes of the psychosis, such as S100B, may also index effects of antipsychotic first treated episode of psychosis. Arch Gen Psychiatry 2011; 68: 609–616. drug effects on gene expression, both directly and indirectly via 5 Heald A. Physical health in schizophrenia: a challenge for antipsychotic therapy. 53 obesity and insulin resistance. Eur Psychiatry 2010; 25(Suppl 2): S6–11.

The Pharmacogenomics Journal (2014), 446 – 451 & 2014 Macmillan Publishers Limited Antipsychotic effects on human gene expression DL Foley and A Mackinnon 451 6 Gentile S. A systematic review of quality of life and weight gain-related issues in 30 Deslauriers J, Lefrançois M, Larouche A, Sarret P, Grignon S. Antipsychotic- patients treated for severe and persistent mental disorders: focus on aripiprazole. induced DRD2 upregulation and its prevention by a-lipoic acid in SH-SY5Y neu- Neuropsychiatr Dis Treat 2009; 5: 117–125. roblastoma cells. Synapse 2011; 65: 321–331. 7 Nasrallah HA. Atypical antipsychotic-induced metabolic side effects: insights from 31 Bernstein HG, Lendeckel U, Dobrowolny H, Stauch R, Steiner J, Grecksch G et al. receptor-binding profiles. Mol Psychiatry 2008; 13: 27–35. Beacon-like/ubiquitin-5-like immunoreactivity is highly expressed in human 8 Teff KL, Kim SF. Atypical antipsychotics and the neural regulation of food intake hypothalamus and increased in haloperidol-treated schizophrenics and a rat and peripheral metabolism. Physiol Behav 2011; 104: 590–598. model of schizophrenia. Psychoneuroendocrinology 2008; 33: 340–351. 9 Roerig JL, Steffen KJ, Mitchell JE. Atypical antipsychotic-induced weight gain: 32 Cohen T, Sundaresh S, Levine F. Antipsychotics activate the TGFb pathway insights into mechanisms of action. CNS Drugs 2011; 25: 1035–1059. effector SMAD3. Mol Psychiatry 2013; 18: 347–357. 10 Lett TAP, Wallace TJM, Chowdhury NI, Tiwari AK, Kennedy JL, Mu¨ller DJ. Phar- 33 Canfrań-Duque A, Casado ME, Pastor O, Sańchez-Wandelmer J, de la PenãG, macogenetics of antipsychotic-induced weight gain: review and clinical implica- Lerma M et al. Atypical antipsychotics alter cholesterol and fatty acid metabolism tions. Mol Psychiatry 2012; 17: 242–266. in vitro. J Lipid Res 2013; 54: 310–324. 11 Lee SH, DeCandia TR, Ripke S, Yang J. Schizophrenia Psychiatric Genome-Wide 34 Menga A, Infantino V, Iacobazzi F, Convertini P, Palmieri F, Iacobazzi V. Insight into Association Study Consortium (PGC-SCZ); International Schizophrenia Consortium mechanism of in vitro insulin secretion increase induced by antipsychotic clo- (ISC); Molecular Genetics of Schizophrenia Collaboration (MGS), Sullivan PF, zapine: role of FOXA1 and mitochondrial citrate carrier. Eur Neuropsycho- Goddard ME et al. Estimating the proportion of variation in susceptibility to pharmacol 2013; 23: 978–987. schizophrenia captured by common SNPs. Nat Genet 2012; 44: 247–250. 35 Chen ML, Tsai TC, Wang LK, Lin YY, Tsai YM, Lee MC et al. Clozapine inhibits Th1 12 Girard SL, Xiong L, Dion PA, Rouleau GA. Where are the missing pieces of the cell differentiation and causes the suppression of IFN-g production in peripheral schizophrenia genetics puzzle? Curr Opin Genet Dev 2011; 21: 310–316. blood mononuclear cells. Immunopharmacol Immunotoxicol 2012; 34: 686–694. 13 Thomas EA. Molecular profiling of antipsychotic drug function: convergent 36 Lauressergues E, Bert E, Duriez P, Hum D, Majd Z, Staels B et al. Does endoplasmic mechanisms in the pathology and treatment of psychiatric disorders. Mol Neu- reticulum stress participate in APD-induced hepatic metabolic dysregulation? robiol 2006; 34: 109–128. Neuropharmacology 2012; 62: 784–796. 14 Tummers B. Data Thief 111. 2006; http://datathief.org/. 37 Kiselyuk A, Farber-Katz S, Cohen T, Lee SH, Geron I, Azimi B et al. Phenothiazine 15 Fernø J, Raeder MB, Vik-Mo AO, Skrede S, Glambek M, Tronstad KJ et al. Anti- neuroleptics signal to the human insulin promoter as revealed by a novel psychotic drugs activate SREBP-regulated expression of lipid biosynthetic genes high-throughput screen. J Biomol Screen 2010; 15: 663–670. in cultured human glioma cells: a novel mechanism of action? Pharmacogenomics 38 Sussman MA, Vo¨lkers M, Fischer K, Bailey B, Cottage CT, Din S et al. Myocardial J 2005; 5: 298–304. AKT: the omnipresent nexus. Physiol Rev 2011; 91: 1023–1070. 16 Polymeropoulos MH, Licamele L, Volpi S, Mack K, Mitkus SN, Carstea ED et al. 39 Tabare´s-Seisdedos R, Dumont N, Baudot A, Valderas JM, Climent J, Valencia A Common effect of antipsychotics on the biosynthesis and regulation of fatty acids et al. No paradox, no progress: inverse cancer comorbidity in people with other and cholesterol supports a key role of lipid homeostasis in schizophrenia. Schi- complex diseases. Lancet Oncol 2011; 12: 604–608. zophr Res 2009; 108: 134–142. 40 Shimano H. Sterol regulatory element-binding proteins (SREBPs): transcriptional 17 Barak Y, Ron IG, Sirota P, Sobe T, Slor H. The effect of chlorpromazine and regulators of lipid synthetic genes. Prog Lipid Res 2001; 40: 439–452. haloperidol on DNA transcription. Int Clin Psychopharmacol 1996; 11: 193–197. 41 Shimano H. SREBPs: physiology and pathophysiology of the SREBP family. FEBS J 18 Fernø J, Skrede S, Vik-Mo AO, Håvik B, Steen VM. Drug-induced activation of 2009; 276: 616–621. SREBP-controlled lipogenic gene expression in CNS-related cell lines: marked 42 Zdychova´ J, Komers R. Emerging role of Akt kinase/protein kinase B signaling in differences between various antipsychotic drugs. BMC Neurosci 2006; 7:69–79. pathophysiology of diabetes and its complications. Physiol Res 2005; 54: 1–16. 19 Vik-Mo AO, Fernø J, Skrede S, Steen VM. Psychotropic drugs up-regulate the 43 O’Neill BT, Abel ED. Akt1 in the cardiovascular system: friend or foe? J Clin Invest expression of cholesterol transport proteins including ApoE in cultured human 2005; 115: 2059–2064. CNS- and liver cells. BMC Pharmacol 2009; 9: 10–16. 44 Ho BC, Andreasen NC, Ziebell S, Pierson R, Magnotta V. Long-term antipsychotic 20 Yang LH, Chen TM, Yu ST, Chen YH. Olanzapine induces SREBP-1-related adipo- treatment and brain volumes: a longitudinal study of first-episode schizophrenia. genesis in 3T3-L1 cells. Pharmacol Res 2007; 56: 202–208. Arch Gen Psychiatry 2011; 68: 128–137. 21 Aubry JM, Schwald M, Ballmann E, Karege F. Early effects of mood stabilizers on 45 Iliadis F, Kadoglou N, Didangelos T. Insulin and the heart. Diabetes Res Clin Pract the Akt/GSK-3beta signaling pathway and on cell survival and proliferation. Psy- 2011; 93(Suppl 1): S86–S91. chopharmacology 2009; 205: 419–429. 46 Krycer JR, Sharpe LJ, Luu W, Brown AJ. The akt-SREBP nexus: cell signaling meets 22 Chana G, Lucero G, Salaria S, Lozach J, Du P, Woelk C et al. Upregulation of NRG-1 lipid metabolism. Trends Endocrin Met 2010; 21: 268–276. and VAMP-1 in human brain aggregates exposed to clozapine. Schizophr Res 47 Luu W, Sharpe LJ, Stevenson J, Brown AJ. Akt acutely activates the cholestero- 2009; 113: 273–276. genic transcription factor SREBP-2. Biochim Biophys Acta 2012; 1823: 458–464. 23 Bernstein HG, Ernst T, Lendeckel U, Bukowska A, Ansorge S, Stauch R et al. 48 Hakem R. DNA-damage repair; the good, the bad, and the ugly. EMBO J 2008; 27: Reduced neuronal expression of insulin-degrading enzyme in the dorsolateral 589–605. prefrontal cortex of patients with haloperidol-treated, chronic schizophrenia. J 49 Patent application 20110118313. http://www.faqs.org/patents/app/20110118313. Psychiatr Res 2009; 43: 1095–1105. 50 Camargo N, Smit AB, Verheijen MHG. SREBPs: SREBP function in glia-neuron 24 Park SW, Lee JG, Ha EK, Choi SM, Cho HY, Seo MK et al. Differential effects of interactions. FEBS J 2009; 276: 628–636. aripiprazole and haloperidol on BDNF-mediated signal changes in SH-SY5Y cells. 51 Saher G, Bru¨ggerB,Lappe-SiefkeC,Mo¨bius W, Tozawa R, Wehr MC et al. High cho- Eur Neuropsychopharmacol 2009; 19: 356–362. lesterol level is essential for myelin membrane growth. Nat Neurosci 2005; 8: 468–475. 25 Lee JG, Cho HY, Park SW, Seo MK, Kim YH. Effects of olanzapine on brain-derived 52 Guo D, Reinitz F, Youssef M, Hong C, Nathanson D, Akhavan D et al. An LXR neurotrophic factor gene promoter activity in SH-SY5Y neuroblastoma cells. Prog agonist promotes GBM cell death through inhibition of an EGFR/AKT/SREBP-1/ Neuropsychopharmacol Biol Psychiatry 2010; 34: 1001–1006. LDLR-dependent pathway. Cancer Discov 2011; 1: 442–456. 26 Schmidt AJ, Hemmeter UM, Krieg JC, Vedder H, Heiser P. Impact of haloperidol 53 Steiner J, Myint AM, Schiltz K, Westphal S, Bernstein HG, Walter M et al. S100B and quetiapine on the expression of genes encoding antioxidant enzymes in serum levels in schizophrenia are presumably related to visceral obesity and human neuroblastoma SH-SY5Y cells. J Psychiatr Res 2009; 43: 818–823. insulin resistance. Cardiovasc Psychiatry Neurol 2010; 2010: 480707. 27 Shin SY, Choi BH, Ko J, Kim SH, Kim YS, Lee YH. Clozapine, a neuroleptic agent, 54 Remington G, Kapur S. Antipsychotic dosing: how much but also how often? inhibits Akt by counteracting Ca2 þ /calmodulin in PTEN-negative U-87MG Schizophr Bull 2010; 36: 900–903. human glioblastoma cells. Cell Signal 2006; 18: 1876–1886. 55 Sakai S, Ikematsu K, Matsuo A, Tsai CT, Nakasono I. Expression of C-fos, Fos-B, 28 Kang J, Chen X-L, Rampe D. The antipsychotic drugs sertindole and pimozide Fosl-1, Fosl-2, Dusp-1 and C-jun in the mouse heart after single and repeated block erg3, a human brain K þ channel. Biochem Biophys Res Commun 2001; 286: chlorpromazine administrations. Leg Med 2010; 12: 284–288. 499–504. 56 Vik-Mo AO, Birkenaes AB, Fernø J, Jonsdottir H, Andreassen OA, Steen VM. 29 Su Z, Martin R, Cox BF, Gintant G. Mesoridazine: an open-channel blocker of human Increased expression of lipid biosynthesis genes in peripheral blood cells of ether-a-go-go-related gene K þ channel. J Mol Cell Cardiol 2004; 36: 151–160. olanzapine-treated patients. Int J Neuropsychop 2008; 11: 679–684.

Supplementary Information accompanies the paper on the The Pharmacogenomics Journal website (http://www.nature.com/tpj)

& 2014 Macmillan Publishers Limited The Pharmacogenomics Journal (2014), 446 – 451