Olanzapine, but not aripiprazole, weight-independently elevates serum triglycerides and activates lipogenic gene expression in female rats Free

Abstract

Metabolic adverse effects such as weight gain and dyslipidaemia represent a major concern in treatment with several antipsychotic drugs, including olanzapine. It remains unclear whether such metabolic side-effects fully depend on appetite-stimulating actions, or whether some dysmetabolic features induced by antipsychotics may arise through direct perturbation of metabolic pathways in relevant peripheral tissues. Recent clinical and preclinical studies indicate that dyslipidaemia could occur independently of weight gain. Using a rat model, we showed that subchronic treatment with olanzapine induces weight gain and increases adipose tissue mass in rats with free access to food. This effect was also observed for aripiprazole, considered metabolically neutral in the clinical setting. In pair-fed rats with limited food access, neither olanzapine nor aripiprazole induced weight gain. Interestingly, olanzapine, but not aripiprazole, induced weight-independent elevation of serum triglycerides, accompanied by up-regulation of several genes involved in lipid biosynthesis, both in liver and in adipose tissues. Our findings support the existence of tissue-specific, weight-independent direct effects of olanzapine on lipid metabolism.

Introduction

Life expectancy for schizophrenia patients is greatly reduced compared to that of the general population (Tiihonen et al.2009), with increased risk of somatic conditions such as cardiovascular disease (CVD) largely accounting for the mortality gap (Colton & Manderscheid, 2006). Treatment with certain atypical antipsychotic drugs, most notably clozapine and olanzapine, has consistently been shown to increase the risk of obesity, dyslipidaemia, and type 2 diabetes (Allison et al.1999; American Diabetes Association, 2004; Henderson, 2001), all established CVD risk factors (Despres et al.2008). Some newly introduced atypical antipsychotics, such as aripiprazole, are associated with much lower risk of metabolic adverse effects than olanzapine and clozapine (Leucht et al.2009; Newcomer et al.2008; Stip & Tourjman, 2010). In fact, add-on treatment of aripiprazole in patients treated with olanzapine reduced body weight and improved serum lipid profiles (Henderson et al.2009). The widespread use of metabolically unfavourable antipsychotics in clinical practice may be explained by superior symptom-relieving effects (Essali et al.2009; Komossa et al.2010; Lieberman et al.2005). Interestingly, the occurrence of weight gain and hypertriglyceridaemia has been suggested to predict superior clinical improvement of schizophrenia symptoms in some studies (Ascher-Svanum et al.2005; Bai et al.2006; Procyshyn et al.2007). Insight into the molecular basis of the complex balance between therapeutic efficacy and adverse effects may pave the way for improved drug therapy in schizophrenia.

Antipsychotic-induced weight gain probably represents a major risk factor for dyslipidaemia in patients. However, increased serum triglycerides and other metabolic derangements have been demonstrated to occur independently of weight gain for both clozapine and olanzapine (Birkenaes et al.2008; Meyer, 2002; Procyshyn et al.2007), suggesting that molecular mechanisms distinct from those causing hyperphagia and weight gain may underlie some of the metabolic disturbances associated with atypical antipsychotic agents. The nature of these mechanisms remains to be established. We have previously demonstrated that in cultured cells, several antipsychotic drugs, among them clozapine and olanzapine, activate lipogenic gene expression controlled by sterol regulatory element-binding protein (SREBP) transcription factors (Ferno et al.2005, 2006). These lipid-stimulating effects, suggested as being relevant for some of the metabolic disturbances associated with antipsychotic drug treatment (Ferno et al.2005; Raeder et al.2006), have later been confirmed by others (Lauressergues et al.2010; Yang et al.2007). Two SREBP isoforms exist, with SREBP1 mainly regulating fatty-acid synthesis, and SREBP2 controlling genes involved in cholesterol biosynthesis and transport (Shimano, 2001).

Molecular studies in psychotic patients spur ethical and practical challenges, and reliable preclinical models are therefore essential in psychopharmacological research (Boyda et al.2010). In rats, olanzapine frequently mimics the metabolic disturbances observed in humans, with increased adiposity consistently reported in both female and male rats (Albaugh et al.2006; Minet-Ringuet et al.2006a). However, olanzapine-induced hyperphagia and weight gain are extensively documented only in female rodents, with less robust effects observed in male rats (Cooper et al.2005, 2007; Minet-Ringuet et al.2006a, b). Olanzapine-induced effects on serum lipids in rats are equivocal (Albaugh et al.2006; Boyda et al.2010; Kalinichev et al.2005). With respect to aripiprazole, data on metabolic disturbances in rats are so far ambiguous (Han et al.2008; Kalinichev et al.2005).

In the present study, two subsequent, partially overlapping experiments were performed. In expt I, we investigated the effect of olanzapine and aripiprazole on body-weight gain and serum lipids in rats. In order to investigate potential weight-independent dyslipidaemic effects of the obesogenic drug olanzapine, a group of rats with food restricted to the level of vehicle-treated rats (pair-fed) was included. Expt II was a follow-up study, focusing on weight-independent effects of both olanzapine and aripiprazole. We compared the effects of olanzapine and aripiprazole on a number of lipid parameters and on lipogenic gene transcription across several metabolically active tissues.

Materials and methods

Drugs

Olanzapine and aripiprazole were suspended in 4% carboxymethyl cellulose (CMC). Care was taken to protect the drugs from light exposure. Plasma levels of antipsychotic drugs were determined by means of liquid chromatography/mass spectrometry [see Supplementary materials and methods (available online) for a detailed description].

Animals and treatment schemes

All experiments were performed in accordance with the guidelines of the Norwegian Committee for Experiments on Animals. Rats were kept under standard conditions with an artificial 12-h light/dark cycle (lights on 07:00 hours) under constant 48% humidity. Ad libitum-fed rats were allowed free access to standard laboratory chow during the experimental period, while pair-fed rats received an amount of food corresponding to that consumed by the control group during the previous 24 h. Rats were allowed free access to tap water. This study consists of two subsequent independent biological experiments. In the inital experiment (expt I), we investigated the effect of olanzapine and aripiprazole on body-weight gain and serum lipids in female Sprague–Dawley rats weighing (mean±s.d.) 261±12 g on the first day of treatment and housed n=3 per cage. We chose to house rats together in order to reduce stress, which could potentially influence food intake. The following treatment groups were included (n=9): vehicle, olanzapine ad libitum-fed, olanzapine pair-fed, aripiprazole ad libitum-fed. Each animal, and chow in each cage, was weighed daily, and average daily food intake for each animal was calculated. Vehicle or antipsychotic drug was administered by gavage twice daily (08:00–09:00 and 14:00–15:00 hours), with total daily dosage of 6 mg/kg olanzapine or aripiprazole split into two separate administrations (administration volume 2 ml/kg). A second experiment (expt II) was performed in order to follow-up on the weight-independent effects in expt I, focusing both on olanzapine and aripiprazole. In expt II we also aimed to investigate the effect of antipsychotic drug treatment on white adipose tissue (WAT) mass, using both dissection of WAT depots and MRI. Expt II was identical to expt I, with a few important exceptions: rats (n=10) were housed individually, and an aripiprazole pair-fed treatment group was included.

The duration of drug exposure was 13 d, and animals were sacrificed on day 14. The last drug dose prior to sacrifice was administered 18–23 h prior to decapitation. All animals were fasted from 21:00 hours on the day prior to sacrifice, with dissection starting at 09:00 hours the following day. Animals were anaesthetized using isoflurane prior to decapitation. Trunk blood was collected in pre-cooled tubes, left at 4°C for 25–30 min and centrifuged at 4°C for 10 min (3000 g) to extract serum. Samples from other tissues were frozen in liquid nitrogen immediately after dissection. Tissue samples were stored at −80°C until analysis, whereas serum samples were kept at −20°C.

Ethics statement

The animal experiments described in this article were approved by Norwegian Committee for Experiments on Animals (Forsøksdyrutvalget, FDU), and were performed in accordance with the Norwegian Animal Welfare Act and international guidelines. Care was taken to minimize the suffering of animals at all stages of the experiments.

Behavioural analysis

Possible sedative effects of olanzapine and aripiprazole were assessed in expt I. The number of seconds spent on grooming, rearing, sitting/standing or smaller movements was manually quantified during 180 s of recorded film by an experienced animal technician.

Serum glucose and lipid measurements and fatty-acid composition analysis

The levels of glucose, triglycerides, phospholipids, and cholesterol in serum and liver were measured enzymatically on the Hitachi 917 system (Roche Diagnostics, Germany) using glucose (Gluco-quant), triglyceride (GPO-PAP) and cholesterol (CHOD-PAP) kits (Roche Diagnostics) and a phospholipid kit (DiaSys Diagnostic Systems, Germany). Serum and liver lipids were extracted according to the method of Bligh and Dyer (Bligh & Dyer, 1959), evaporated under N₂ and re-dissolved in isopropanol before analysis. Serum lipid composition was analysed by gas–liquid chromatography (see Supplementary materials and methods for detailed description).

Serum insulin, leptin and adiponectin measurements

Truncal vein blood was collected in precooled tubes, left on ice for 30 min and centrifuged at 3000 g for 10 min. Serum was transferred to pre-cooled Eppendorf tubes immediately after centrifugation and stored at −20°C. Serum insulin, leptin and adiponectin levels were assessed as previously described (Caminos et al.2005).

Sample preparation, RNA extraction cDNA synthesis and real-time PCR

Tissue samples (∼20 mg of liver tissue or ∼100 mg of adipose tissue) were homogenized using a TissueLyser (Qiagen, USA). RNA extraction and cDNA synthesis were performed as previously described (Ferno et al.2009). Primers (Table 1) were designed using PrimerExpress (Applied Biosystems, USA) or Primer3 (Rozen & Skaletsky, 2000). The relative gene expression levels were determined by means of the comparative ΔC_t method. In liver samples, the expression of target genes was normalized relative to the endogenous control ribosomal protein, large, P0 (Rplp0), while genes run in WAT were normalized against both Rplp0 and β-actin (Actb); results are presented as fold changes normalized against P0.

Immunoblotting

Tissue samples (100 mg liver tissue, 250 mg adipose tissue) were homogenized in lysis buffer and centrifuged at 12 000 g for 10 min at 4°C (Lopez et al.2010). Thirty µg total protein from each sample was separated on NuPAGE 4–12% Bis-Tris Gels using SDS buffer and blotted onto Invitrolon™ PVDF membranes. PVDF membranes were blocked with 5% BSA in TBST prior to incubation with primary antibody at 4°C overnight or at room temperature (RT) for 1 h, followed by incubation with secondary antibody at RT for 1 h. The primary antibodies used were: ACC1: 04-322 (Millipore, USA); α-tubulin: T5168 (Sigma Aldrich, USA); FAS: SC20140 (Santa Cruz, USA); SREBP1 (recognizing both SREBP1a and −1c isoforms): SC8984 (Santa Cruz). Signal intensity measurements were performed using the ImageJ software (National Institutes of Health, USA).

Quantification of adipose tissue

In order to investigate whether weight gain correlates with increased WAT mass, we ran a separate experiment (expt II) using two alternative methods to compare the weights of dissected adipose tissue depots by the end of the treatment period with semi-quantitative MRI measurements of the changes in pararenal adipose tissue volumes during treatment. Following 13 d of drug treatment, animals were sacrificed on day 14 after overnight fasting, with immediate dissection of mesenteric, retroperitoneal, pararenal, and parametrial (periovarian) adipose tissue depots, which were weighed separately. Animals were subject to MRI scans on treatment days −1 (the day prior to initiation of treatment) and day 13 (the day prior to sacrifice). MRI images were acquired using a 7-T horizontal bore magnet (Pharmascan 70/16, Bruker BioSpin, Germany) operating at 300 MHz, using protocols described in detail in the Supplementary materials and methods section. An estimate of the relative change in pararenal fat volume was made using a robust thresholding scheme on the MR images, as described in Supplementary materials and methods. In short, an estimated change in adipose tissue volume during the treatment period was calculated by summing up voxels containing fat signals in selected slices with visible kidney tissue, as the kidneys represent reliable anatomical landmarks, and substracting the number of fat-containing voxels on the day prior to treatment initiation from the number of fat-containing voxels on the day prior to sacrifice.

Statistical analysis

Data are expressed as mean±s.e.m. Food intake was analysed by repeated-measures two-way ANOVAs, with treatment as between-subject variable and time as within-subject variable in both experiments. Body weight changes were analysed using the same method. In expt I, four treatment groups were included (control, olanzapine ad libitum-fed, olanzapine pair-fed, aripiprazole ad libitum-fed), whereas expt II contained five treatment groups, as an aripiprazole pair-fed group was added. One-way ANOVA followed by Tukey's post-hoc test or Student's t test was used to analyse statistical significance for each time-point. Pearson's bivariate correlation analysis was used in order to examine correlations between various metabolic parameters as described below. For all other comparisons, we used two-sided Student's t test. p values <0.05 were considered statistically significant. All tests were conducted with PASW Statistics version 18 (PASW Statistics; SPSS Inc., USA) software.

Results

Serum levels of antipsychotics correspond with a short half-life

The average serum concentration of olanzapine (mean±s.d.) at the time of sacrifice (18–23 h after last dose) was very low both in the olanzapine ad libitum-fed group (average+s.d.) (0.27±0.10 nm) and in the olanzapine pair-fed group (0.80±0.57 nm). Serum aripiprazole levels were also negligible (0.23±0.17 nm).

Olanzapine and aripiprazole induce hyperphagia and body-weight gain

Initially (expt I), we investigated how 13 d treatment with olanzapine or aripirazole affected daily food intake (Fig. 1a), cumulative food intake (Fig. 1b) and weight gain (Fig. 1c) in female rats. In order to investigate any potential hyperphagia-independent effects of the well established obesogenic drug olanzapine, we included an olanzapine pair-fed group, with food intake identical to the level consumed by vehicle-treated rats. Repeated-measures two-way ANOVA was performed for average daily food intake (from n=3 rats per cage) with treatment and time as factors. A statistically significant main effect on food intake for treatment was evident [F(2, 6)=21.91, p<0.01]. A Tukey's HSD post-hoc test revealed that both olanzapine and aripiprazole induced significantly different food intake from vehicle-treated controls. Similarly, cumulative body-weight gain was analysed using a repeated-measures two-way ANOVA with time and treatment as factors. Significant main effects [F(3, 32)=12.01, p<0.001] and significant treatment×time interaction effects [F(48, 133)=2.37, p<0.001] were observed. Statistically significant differences between the drugs for each time-point was determined by one-way ANOVA analysis followed by Tukey's post-hoc test, as indicated in Fig. 1c. In expt II, we focused on weight-independent effects of both olanzapine and aripiprazole. With respect to food intake (measured for each rat individually), a significant main effect was observed for treatment [F(2, 27)=16.31, p<0.001] and a trend towards significance was observed for the time×treatment interaction effect [F(24, 32)=1.645, p=0.1] (Supplementary Fig. S1 a). In line with elevated food intake, cumulative body-weight gain was increased, with a significant main effect of treatment [F(4, 45)=20.03, p<0.001] and significant treatment×time interaction effect [F(36, 62)=3.59, p<0.001]. In expt II, weight gain in the aripiprazole ad libitum-fed group was more pronounced than in expt I (Supplementary Fig. S1 b), whereas no weight gain was observed in the aripiprazole pair-fed group. Of note, one-way ANOVA followed by Tukey's post-hoc test revealed significantly reduced body weight in the olanzapine pair-fed group relative to the control (p⩽0.05) and aripiprazole pair-fed (p⩽0.01) groups.

Olanzapine, but not aripiprazole, induces weight-independent elevation in serum triglyceride levels

Interestingly, olanzapine significantly elevated fasting serum triglyceride levels both in ad libitum-fed rats that gained weight (mean±s.e.m., relative to vehicle, 129±10%, p⩽0.05) and in pair-fed rats that did not (126±7%, p⩽0.05), while no such changes were induced by aripiprazole, despite the increased weight gain (Fig. 2a). Other serum lipids were unaffected by olanzapine exposure, whereas aripiprazole slightly reduced serum cholesterol and phospholipid levels relative to controls (Fig. 2a and Table 2). Similarly, in expt II olanzapine induced significantly elevated serum triglyceride levels in the ad libitum-fed group (166±9%, p⩽0.001) and a trend towards increased serum triglyceride levels (126±12%, p=0.09) in pair-fed rats, despite marked weight loss in this treatment group. In contrast, no significant change in serum triglycerides was evident following aripiprazole treatment, neither in ad libitum-fed (114±13%, p=0.37) nor in pair-fed (101±8%, p=0.91) rats (Supplementary Fig. S2). The relationship between body-weight gain and serum lipids was explored using Pearson's correlation coefficient. In expt I, when analysing the whole group (n=36), serum triglycerides displayed a significant, but weak correlation [r(36)=0.42, p=0.012] with body-weight gain. However, since weight gain and lipids were differentially affected by the different treatments, correlation analyses were also performed separately for each treatment group. No significant correlation was observed between body-weight gain and any of the lipid parameters in the separate treatment groups, suggesting that the effect of treatment was more pronounced than the effect of weight gain. Similarly, no significant correlation between body-weight gain and serum lipids was found in any of the treatment groups in expt II (data not shown). Next, we measured fatty-acid composition in total serum lipids in expt I. Indeed, in line with the observed transcriptional up-regulation of the Δ9 fatty-acid desaturase stearoyl-CoA desaturase 1 (SCD1), as described below, levels of C18:1Δ9 (oleic acid) were significantly increased by olanzapine both in ad libitum-fed rats (121±6%, p⩽0.01) and in pair-fed rats (125±6%, p⩽0.01), whereas aripiprazole had no effect (Fig. 2b). Based on the olanzapine-induced changes in serum lipids, we examined the effect of drug treatment on hepatic lipids. Except for a slight reduction in hepatic cholesterol in olanzapine-exposed pair-fed animals, hepatic lipids were not significantly changed in any of the treatment groups compared to control (Fig. 1d). No significant correlation was found between body-weight gain and hepatic lipids, neither for all animals analysed together nor for the separate treatment groups.

Olanzapine and aripiprazole induce few alterations in serum glucose and metabolically relevant hormones

We observed no effect on fasting serum glucose levels, neither by olanzapine nor by aripiprazole (Table 2). Serum insulin levels were unaltered in olanzapine-treated rats, while a statistically significant reduction (81±6%, p⩽0.05) was induced by aripiprazole (Table 2). Serum levels of leptin were unaltered, whereas serum adiponectin was moderately increased in the olanzapine pair-fed group (Table 1).

Olanzapine increases lipogenic and adipogenic gene expression in parametrial WAT

Based on our previous findings of olanzapine-induced SREBP activation in cultured cells, we examined the expression of numerous lipid homeostasis genes in metabolically active peripheral tissues in expt I, including parametrial (visceral) WAT, subcutaneous WAT, liver, and brown adipose tissue (BAT). In parametrial WAT of ad libitum-fed rats, we found that olanzapine induced up-regulation of SREBP-controlled fatty-acid biosynthetic genes, including acetyl-CoA carboxylase 1 (Acc1), fatty-acid synthase (Fasn) and stearoyl-CoA desaturase (Scd1) (Table 3). A moderate, non-significant trend towards transcriptional up-regulation was also observed for aripiprazole (Table 3). Of note, the average fold change of the three genes was positively correlated with cumulative weight gain, both for aripiprazole [r(9)=0.90, p=0.001] and olanzapine [r(9)=0.87, p=0.003]. Nevertheless, a marked, significant up-regulation of Acc1, Fasn and Scd1 was also present in the olanzapine pair-fed treatment group (Table 3). As could be expected from the lack of body-weight gain in this group, no significant correlation was found between transcriptional activation and body-weight gain [r(9)=0.50, p=0.17]. The olanzapine-induced transcriptional activation of Acc1 was confirmed at the protein level, with statistically significant elevation of ACC1 in both ad libitum-fed and pair-fed rats (Fig. 3a). Notably, despite the marked effect on SREBP target genes in parametrial WAT, olanzapine affected neither expression levels of the Srebp1c gene itself, nor SREBP1 protein levels in this adipose tissue (Table 3, Fig. 3a). Several genes essential in triglyceride synthesis, e.g. Dgat1 (diacylglycerol O-acyltransferase 1) and Dgat2, not considered SREBP targets, followed the pattern above, with up-regulation in one or both olanzapine treatment groups, but not in the aripiprazole group (Table 3). In expt II, olanzapine also induced significant transcriptional induction of key lipogenic enzymes (Acc1, Fas, Scd1) in parametrial WAT, evident both in ad libitum-fed and in pair-fed rats. Several of these genes were also up-regulated by aripiprazole in ad libitum-fed rats, possibly due to the excessive weight gain, as no changes were observed in the aripiprazole pair-fed group (Supplementary Table S1).

Furthermore, we examined several genes involved in cholesterol metabolism, mainly controlled by the SREBP2 transcription factor. Cholesterol biosynthesis genes such as sterol regulatory element-binding protein 2 (Srebp2), 3-hydroxy-3-methylglutaryl-CoA reductase (Hmgcr) and 3-hydroxy-3-methylglutaryl-CoA synthase 1 (Hmgcs1) were up-regulated in parametrial WAT by olanzapine, with most pronounced effects in ad libitum-fed rats (Table 3). Interestingly, these effects were also seen for aripiprazole, with significant up-regulation of Srebp2 and Hmgcs1, as well as the cholesterol esterification gene sterol O-acyltransferase 1 (Table 3). Due to the close links between dyslipidaemia and insulin resistance, we examined the transcription of the insulin-responsive glucose transporter 4 (Glut4), which was markedly up-regulated both by olanzapine and aripiprazole (Table 3).

Subcutaneous adipose tissue: WAT depot-specific effects of antipsychotics on lipid-related gene expression

Olanzapine also up-regulated lipogenic genes in subcutaneous WAT, although the increase was less pronounced than in parametrial WAT, both at the transcriptional and protein levels (Table 3, Fig. 3b). Notably, in contrast to the effects observed in parametrial WAT, olanzapine markedly up-regulated Srebp1c transcription in subcutaneous WAT both in ad libitum-fed and pair-fed rats (Table 3). This was confirmed at the protein level, although reaching statistical significance only in the pair-fed group (Fig. 3b). For aripiprazole, the effect on fatty-acid biosynthesis genes in subcutaneous WAT was marginal, with down-regulation of Scd1 as the only significant finding (Table 3). Directly opposite to what we observed in the parametrial WAT, olanzapine induced down-regulation of Pparg, encoding the nuclear receptor transcription factor PPARγ, considered the master regulator of adipogenesis (Rosen et al.2002). Of note, transcription of the Pparg gene and several of its target genes was down-regulated by both olanzapine and aripiprazole. Diverging effects in parametrial and subcutaneous adipose depots were also observed for Glut4, as well as for genes encoding lipases and adipokines, which were down-regulated in subcutaneous WAT both by olanzapine and aripiprazole (Table 3).

Antipsychotic drugs induce minor effects on hepatic lipid-related gene expression

In the liver, normally representing the primary site of lipogenesis, the effects of olanzapine on the expression of fatty-acid and cholesterol metabolism genes were subtle (Table 3). Still, Western blots demonstrated that ACC1 and FASN protein levels were significantly elevated by olanzapine both in ad libitum-fed and in pair-fed rats, with a similar trend for the SREBP1 protein (Fig. 3c). Aripiprazole, on the other hand, significantly reduced FASN and SREBP1 protein levels (Fig. 3c, Table 3).

The effect of antipsychotic drugs on thermogenic markers in BAT

In our study we found that antipsychotic-induced weight gain was related to increased food intake. However, reduced energy expenditure in BAT has also been suggested to be involved in antipsychotic-induced weight gain, and we therefore examined the expression levels of thermogenic markers in BAT. We found that olanzapine markedly decreased the expression of genes encoding the thermogenic markers uncoupling protein 1 (Ucp1) and peroxisome proliferator-activated receptor gamma, co-activator 1 alpha (Ppargc1a) in BAT, evident both in ad libitum-fed and in pair-fed rats (Fig. 4a). Aripiprazole did not significantly alter the expression of Ucp1, whereas reduced expression levels of Pparg1a were observed (0.41±0.09, p⩽0.05) (Fig. 4a). Sedation, a recognized adverse effect of antipsychotic agents, could also decrease energy output. However, evaluating four different measures of locomotor activity, no statistically significant changes were found relative to vehicle-exposed rats (Fig. 4b).

Weight gain is related to increased adipose tissue mass

In order to investigate whether antipsychotic-induced weight gain is constituted by increased adipose tissue mass and whether the lipogenic effects of olanzapine could increase adipose mass independently of weight gain, a separate experiment (expt II) was set up, in which both a non-invasive MRI technique and weighing of fat after dissection were employed (Fig. 5, Table 4). Correlation analysis yielded a positive correlation [r(50)=0.51, p=0.0001] between cumulative weight gain and total dissected WAT mass (four anatomical depots, Table 4) when data for all animals were analysed together. The same was observed in the aripiprazole ad libitum-fed group [r(n=10)=0.77, p⩽0.01], but not in the olanzapine ad libitum-fed group. In contrast, volume estimates based on the MRI images (Fig. 5) demonstrated a clear trend towards significant increase in adipose tissue mass in the olanzapine ad libitum-fed group relative to control animals (324±84%), p=0.09), while no significant alteration was observed in the corresponding aripiprazole treatment group, despite the observed weight gain and increased mesenteric WAT mass (Table 4). Pearson's correlation analysis yielded no correlation between dissected WAT mass and MRI estimates of WAT volume (data not shown).

Discussion

In the present study, we demonstrated that both olanzapine and aripiprazole induced hyperphagia and weight gain in female rats. In contrast, the effects on blood lipids differed between the drugs. Serum triglyceride levels were elevated by olanzapine in both ad libitum-fed and pair-fed rats, whereas no such increase was observed for aripiprazole. Similarly, olanzapine, but not aripiprazole, weight-independently induced lipogenic gene expression in peripheral tissues, with the most pronounced effects in perimetrial WAT. The different propensity of these drugs to induce lipogenic effects may be relevant for their different metabolic profiles reported in humans (Rummel-Kluge et al.2010).

Drug-induced weight gain is caused by hyperphagia

Drug-induced weight gain was highly correlated with increased food intake, which, in agreement with previous findings, indicated that the obesogenic effects of the drugs are mainly caused by hyperphagia (Albaugh et al.2006; Arjona et al.2004; Coccurello et al.2008; Ferno et al.2011; Goudie et al.2002). However, different mechanisms may coincide, and our finding of reduced thermogenic markers in BAT supported a recent study suggesting decreased thermogenesis as a relevant mechanism for olanzapine-induced weight gain in rats (Stefanidis et al.2009). Nevertheless, the lack of weight gain in pair-fed rats in our study demonstrated that drug-induced reduction in BAT thermogenesis was not solely sufficient to induce weight gain.

Weight gain was positively correlated with dissected adipose tissue mass when all animals were analysed as one group. We investigated MRI as a potential non-invasive method of adipose tissue quantification, which would facilitate regular monitoring of WAT mass during treatment with obesogenic drugs. However, despite the acquisition of high-quality images, we failed to demonstrate correlation between body-weight gain and MRI-based WAT volume estimates. Furthermore, we found no correlation between dissected WAT mass and MRI volume estimates. Clearly, the MRI protocol used needs further development. Ideally, the number of MRI slices should be increased in order to acquire images of all intra-abdomnial WAT depots, circumventing the challenge of limited acquisition time for each rat due to issues such as time spent anaesthetized, body temperature increase, and general time consumption.

Serum triglycerides are elevated by olanzapine treatment, independent of weight gain

Serum triglycerides were elevated by olanzapine treatment, both in ad libitum-fed rats that gained weight and in pair-fed rats that did not, suggesting that this increase occurred independently of weight gain. The fact that aripiprazole did not increase serum triglycerides, despite inducing marked body-weight gain, supports this notion, which was confirmed by correlation analysis of triglyceride levels and body-weight gain for each treatment group. It should be noted that when analysing all treatment groups as one, a moderate but significant positive correlation between serum triglycerides and body-weight gain was observed, demonstrating that although weight gain alone did not account for the triglyceride increase in olanzapine-treated rats, this factor probably contributes to dyslipidaemic effects. The observation that elevated triglyceride levels were not entirely attributable to weight gain is of major importance, as it indicates that mechanisms other than those involved in increased body weight are relevant for antipsychotic-induced dyslipidaemia. This notion is in agreement with evidence from several clinical studies demonstrating significant elevation in serum triglycerides, independent of weight gain, in patients treated with olanzapine (Birkenaes et al.2008; Meyer, 2002; Procyshyn et al.2007).

Olanzapine, but not aripiprazole, exerts direct lipogenic transcriptional activation

The transcriptional up-regulation of SREBP-controlled lipogenic genes in WAT observed in both olanzapine-treated and aripiprazole-treated ad libitum-fed rats agrees with the established fact that hyperphagia-induced weight gain is associated with increased WAT mass and up-regulation of lipogenic gene expression (Shimano, 2001). Increased lipogenic gene expression was also observed in olanzapine pair-fed rats that did not gain weight, an effect not observed in pair-fed, aripiprazole-treated rats. This suggests that similar to the effect on serum triglyceride levels, the major impact on lipogenic transcriptional activation by olanzapine was related to pharmacological effects in WAT and not simply a secondary effect of weight gain. Similarly, marked lipogenic activation by olanzapine, but not aripiprazole, was observed at the protein level in the liver. Our demonstration of weight-independent lipogenic up-regulation by olanzapine is in agreement with previous findings of elevated levels of the SREBP1 target gene Fasn in intra-abdominal adipose tissue from male rats chronically treated with olanzapine without gaining weight (Minet-Ringuet et al.2007), and with a clinical study showing BMI-independent up-regulation of the fatty-acid biosynthesis genes Fasn and Scd1 in peripheral blood cells from a group of patients receiving olanzapine (Vik-Mo et al.2008). Scd1 is an SREBP1 target gene placed at the branch point between lipid synthesis/storage and fatty-acid oxidation, and may prove a critical regulatory step in the development of metabolic disorders by acting as a ‘metabolic switch’ between fatty-acid synthesis and catabolism (Ntambi et al.2002). SCD1 catalyses delta 9 monodesaturation of fatty acids, and the olanzapine-specific elevation of the monodesaturated fatty-acid oleic acid in serum observed in our study suggests that increased Scd1 expression by olanzapine has functional consequences, possibly relevant for olanzapine-induced metabolic disturbances. Interestingly, our findings confirm a recent study of elevated plasma 18:1/18:0 ratio after treatment with olanzapine as well as other antipsychotic drugs (McNamara et al.2011). Indeed, antipsychotic agents have also been shown to increase the proportion of monounsaturated fatty acids in plasma from patients (Kaddurah-Daouk et al.2007).

Lipogenic transcriptional activation in visceral WAT possibly occurs independently of SREBP activation

Although olanzapine-induced up-regulation of SREBP-controlled lipogenic genes was evident in parametrial WAT, it was not accompanied by the expected SREBP1 activation, neither at the transcriptional nor at the protein level. However, in subcutaneous WAT, with only moderate elevation of lipogenic SREBP1-controlled gene expression, we observed a striking elevation of Srebp1c expression and SREBP1 protein levels. These seemingly paradoxical findings may be explained by the fact that fatty-acid biosynthesis in adipose tissue is not exclusively controlled by SREBP1 (Sekiya et al.2007). In fact, it has been demonstrated that WAT-specific transgenic overexpression of Srebp1c reduced fat accumulation in mice, with concomitant down-regulation of PPARγ-controlled genes involved in adipocyte differentiation (Shimomura et al.1998), a state resembling the antipsychotic-induced effects observed in subcutaneous WAT in our study. PPARγ agonists, such as thioglitazones, have been used to treat metabolic disturbances in patients (Festuccia et al.2009). Thus it is tempting to speculate that PPARγ antagonistic effect of antipsychotic drugs could represent a molecular mechanism involved in their associated metabolic dysfunction, including glucose dysregulation. We found few alterations in fasting glucose and insulin levels, but studies employing clamping techniques have demonstrated antipsychotic-induced glucose dysregulation (Albaugh et al.2010; Chintoh et al.2009). It should be noted that marked down-regulation of PPARγ target genes was observed both in rats treated with olanzapine and in rats treated with aripiprazole, and thus the apparent PPARγ antagonistic effect does not offer a straightforward explanation for the antipsychotic-induced metabolic disturbances observed in the clinical setting.

In summary, we have demonstrated that both olanzapine and aripiprazole induce significant weight gain in female rats, attributable mainly to hyperphagia. Interestingly, olanzapine elevated serum triglycerides independently of weight gain, an effect that was not observed for aripiprazole. The concurrent olanzapine-induced up-regulation of lipogenic gene expression in adipose tissues implied a potential mechanism of antipsychotic-induced dyslipidaemia, possibly involving the SREBP transcription system. We also showed that olanzapine appears to exert PPARγ antagonistic effects in subcutaneous WAT, a property shared by aripiprazole. Further studies are required in order to elucidate the role of these effects in the development of metabolic dysfunction in patients.

Note

Supplementary material accompanies this paper on the Journal's website.

Acknowledgements

We greatly appreciate the excellent technical assistance from Marianne S. Nævdal, Liv Kristine Øysæd, Kari Williams, Pavol Bohov, Kari H. Mortensen, and Cecilie Brekke Rygh. We acknowledge the research infrastructure provided by the Norwegian Microarray Consortium (NMC; www.microarray.no), a national FUGE technology platform (Functional Genomics in Norway; www.fuge.no), by the Molecular Imaging Center (MIC), Department of Biomedicine, University of Bergen, and by the laboratory animal facility (Vivarium), University of Bergen. This study has been supported by grants from the Research Council of Norway (including the FUGE programme and ‘PSYKISK HELSE’ program), Norwegian Council for Mental Health, ExtraStiftelsen Helse og Rehabilitering, Helse Vest RHF, and Dr Einar Martens Fund (J. F.), the Medical Research Council, UK (A. V.-P.: G0802051), Wellcome Trust (A. V.-P.: 065326/Z/01/Z), Xunta de Galicia (M. L.: 10PXIB208164PR), and Fondo Investigationes Sanitarias (M. L.: PS09/01880), Ministerio de Ciencia e Innovación (M. L.: RyC-2007-00211), the European Community's Seventh Framework Programme (FP7/2007-2013) under grant agreement no. 245009 and by Ministerio de Educacion y Ciencia (CD: BFU2008-02001). CIBER de Fisiopatología de la Obesidad y Nutrición is an initiative of ISCIII.

Statement of Interest

None.

References