ORIGINAL ARTICLE

Olanzapine promotes fat accumulation in male rats bydecreasing physical activity, repartitioning energy andincreasing adipose tissue lipogenesis while impairinglipolysisVL Albaugh1, JG Judson1, P She1, CH Lang1, KP Maresca2, JL Joyal2 and CJ Lynch1

1Department of Cellular and Molecular Physiology, The Pennsylvania State University College of Medicine, Hershey,PA, USA and 2Molecular Insight Pharmaceuticals, Cambridge, MA, USA

Olanzapine and other atypical antipsychotics cause metabolic side effects leading to obesityand diabetes; although these continue to be an important public health concern, theirunderlying mechanisms remain elusive. Therefore, an animal model of these side effects wasdeveloped in male Sprague–Dawley rats. Chronic administration of olanzapine elevated fastingglucose, impaired glucose and insulin tolerance, increased fat mass but, in contrast to femalerats, did not increase body weight or food intake. Acute studies were conducted to delineatethe mechanisms responsible for these effects. Olanzapine markedly decreased physicalactivity without a compensatory decline in food intake. It also acutely elevated fasting glucoseand worsened oral glucose and insulin tolerance, suggesting that these effects are adiposityindependent. Hyperinsulinemic-euglycemic clamp studies measuring 14C-2-deoxyglucoseuptake revealed tissue-specific insulin resistance. Insulin sensitivity was impaired in skeletalmuscle, but either unchanged or increased in adipose tissue depots. Consistent with theolanzapine-induced hyperglycemia, there was a tendency for increased 14C-2-deoxyglucoseuptake into fat depots of fed rats and, surprisingly, free fatty acid (FFA) uptake into fat depotswas elevated approximately twofold. The increased glucose and FFA uptake into adiposetissue was coupled with increased adipose tissue lipogenesis. Finally, olanzapine loweredfasting plasma FFA, and as it had no effect on isoproterenol-stimulated rises in plasmaglucose, it blunted isoproterenol-stimulated in vivo lipolysis in fed rats. Collectively, theseresults suggest that olanzapine exerts several metabolic effects that together favor increasedaccumulation of fuel into adipose tissue, thereby increasing adiposity.Molecular Psychiatry advance online publication, 23 March 2010; doi:10.1038/mp.2010.33

Keywords: atypical antipsychotics; insulin resistance; nutrient partitioning; hyperinsulinemic-euglycemic clamp; free fatty acid uptake; lipogenesis

Introduction

Atypical antipsychotic drugs are a mainstay ofpsychiatric pharmacotherapy and are used to treat avariety of psychiatric illnesses, most notably schizo-phrenia. These second generation drugs lack theextrapyramidal and other debilitating movement sideeffects that complicated treatment with first genera-tion compounds. The popularity of the atypicalantipsychotics, especially olanzapine, led to theobservation that these drugs have distinct metabolicside effects that were not detected during preclinicaltesting, including insulin resistance, diabetes, and

obesity.1–6 The mechanisms underlying these meta-bolic side effects in human beings are not known.However, case reports and other observational studieshave suggested that effects of atypical antipsychoticson glucose homeostasis and insulin sensitivity pre-cede weight gain in human beings.7–9 Regardless,clinically significant weight gain is observed inalmost one-third of patients treated with olanzapine,predisposing these patients to the metabolic syn-drome and additional comorbid conditions (reviewedin Newcomer6,10). Drug therapy is typically continuedin patients that are resistant to the weight-gain pro-moting effects of antipsychotics, though the weight-gain-independent effects of these drugs are less wellunderstood in human beings. Indeed, several studieshave suggested that atypical antipsychotics mayperturb lipid as well as glucose metabolism11–13.Olanzapine, one of the most popular antipsychotics,confers the greatest potential for weight gain and is thefocus of our current effort.

Received 29 July 2009; revised 28 January 2010; accepted 16February 2010

Correspondence: Professor CJ Lynch, Department of Cellular andMolecular Physiology, Pennsylvania State University College ofMedicine, 500 University Drive, MC-H166, Hershey, PA 17033,USA.E-mail: clynch@psu.edu

Molecular Psychiatry (2010), 1–13& 2010 Nature Publishing Group All rights reserved 1359-4184/10 $32.00

www.nature.com/mp

Numerous animal studies have examined thepotential metabolic side effects of olanzapine, thoughmany questions remain unanswered. A commonfinding is that, unlike the clinical situation, olanza-pine-associated weight gain in rats is sex dependentwith female, but not male rats, displaying a robustincrease in body weight gain.14–18 Body weight gain infemale rats is strongly associated with hyperphagia,whereas effects of olanzapine in male rats seem to belimited to increased meal size without altering 24 hfood intake.19–22 However, recent reports23,24 haveshown that olanzapine does indeed have metaboliceffects in male rats. When treated chronically ( > 3weeks), olanzapine increased adiposity without alter-ing food intake or body weight gain, similar to chronic(6 weeks) administration in dogs.25 The mechanism ofthe sexual dimorphism in rats and the increasedadiposity in male rats and dogs is currently unknown,although these weight-gain-independent effects inanimals suggest that weight-gain-resistant patientscontinued on olanzapine therapy may also be atincreased metabolic risk.

In the following report, we confirm that chronicolanzapine administration leads to increased deposi-tion of fat (adiposity) without a change in total bodyweight in male rats and show that these increases infat mass can be detected as early as the seventhtreatment day. After acute olanzapine treatment,animals eat more food for their amount of physicalactivity, which declines. Flux of glucose and fattyacid fuels toward adipose tissue is increased, whichis associated with increased oxygen consumption thatseems to be secondary to increased lipogenesis. Theincreased fuel uptake and storage is also associatedwith impaired lipolysis. Thus, olanzapine seems toincrease adiposity by repartitioning excess nutrientstoward adipose tissue and, by impairing mobilizationof that stored energy, effectively traps it there.Although many patients experience weight gain inresponse to olanzapine, our findings suggest amechanism whereby olanzapine may silently increasemetabolic risk by increasing adiposity without affect-ing body weight per se.

Materials and methods

AnimalsThe animal facilities and protocol were reviewed andapproved by the Institutional Animal Care and UseCommittee of the Penn State Hershey College ofMedicine. Male, Sprague–Dawley rats (B200–225 g)were purchased from Charles River Laboratories(Cambridge, MA, USA) and maintained on a 12:12-hlight/dark cycle with food (Harlan-Teklad RodentChow, no. 2018; Madison, WI, USA) and waterprovided ad libitum. Animals were singly housedduring chronic experiments with experimentalgroups matched for body weight. In acute studies,experimental groups were matched for body weightand housed in groups of five. Before the beginningof individual experiments, animals were routinely

acclimated ( > 1 week) to laboratory conditions bydaily gavage and handling to reduce potential stresseffects during experiments.

Drug preparation/administrationOlanzapine was a gift from Neuland Laboratories(Hyderabad, India) and Dr Reddy’s Laboratories(Hyderabad, India). Olanzapine was administeredeither by oral gavage (acute studies) or self-adminis-tered in cookie dough (chronic studies), similar toearlier studies.14 For gavage, olanzapine solution wasprepared by dissolving pure compound in 0.1 N HClin distilled water, adjusting to pH 5.5 with 0.01 N

NaOH, and adding distilled water to reach the desiredconcentration. Vehicle was similarly pH-adjusted,distilled water. For cookie dough self-administration,separate batches of dough were freshly prepared froma pre-prepared dry mix every few days as needed andrefrigerated at 4 1C.

Chronic drug studiesAnimals were exposed chronically to olanzapine byself-administration of drug-containing cookie dough.Animals were provided 20 g dough per kg bodyweight in 1-oz American portion-control cups (waxpaper) from Hoffmaster (Oshkosh, WI, USA; stock no.60–100). Cookie dough was administered dailybetween 1600 and 1800 hours. At the time of drugadministration, body weight and 24-h food intakewere measured. For chronic treatment, the dose ofolanzapine was 4 mg kg–1 per day during days zerothrough six. On the seventh treatment day, the dosewas increased to 8 mg kg–1 per day. The dose wasagain increased on day 14 to 12 mg kg–1 per daythrough the end of the study. This dosing scheme wasadapted from studies in female rats, in which thisramped-dosing prolonged drug-induced increases inbody weight gain.14

Acute drug studiesRats were randomly allocated to experimental groups.For acute oral glucose tolerance tests (OGTT) andinsulin tolerance tests (ITT), dosing was identical toour earlier studies in female rats.14 Briefly, one-thirdof the daily dose was administered on the first daybetween 0800 and 0900 hours, whereas the remainingtwo-thirds dose was administered before the darkcycle between 1600 and 1700 hours (total dailydose = 4 or 10 mg kg–1). Food and water were providedad libitum. The following morning, one-third of thedaily dose was administered between 0700 and 0800hours. At this point, animals were food restricted,whereas water remained available ad libitum. Fourhours later, the remaining two-thirds dose wasadministered (total dose = 4 or 10 mg kg–1). One hourafter the fourth and final drug or vehicle gavage, theglucose or ITT were performed.

In the calorimetry time-course studies, olanzapine(10 mg kg–1) was provided at the time indicated in thefigure in ad libitum fed animals, as indicated. For allother acute experiments, end points were measured

Olanzapine-induced adiposityVL Albaugh et al

2

Molecular Psychiatry

after two doses of vehicle or drug (olanzapine = 4 or10 mg kg–1). The first dose of olanzapine or vehiclewas administered before the start of the dark cycle atB1800 h. Food and water were provided ad libitum oranimals were food deprived. The following morning,a second dose of vehicle or olanzapine was adminis-tered typically between 0700 and 0800 hours. Thefinal drug or vehicle gavage was always given 2 hbefore the beginning of any experiment or blood/tissue sampling.

Locomotor activity and energy expenditureLocomotor activity was measured using infraredtechnology (OPT-M3, Columbus Instruments). Thecounts of three-dimensional beam breaking (X total,X ambulatory or ‘y’, and Z) were measured for 24 h.Energy expenditure was measured concurrently usingindirect calorimetry (Oxymax, Columbus Instru-ments). Constant airflow (B8 ml g–1 per min) wasdrawn through the chamber and monitored by a mass-sensitive flow meter. O2 and CO2 concentrations weremeasured at the inlet and outlet of sealed chambers tocalculate O2 consumption and CO2 expiration. Eachchamber was measured for 1 min at 15-min intervals.Animals were acclimated to the specialized calori-metry cages and had ad libitum access to food andwater throughout the 24 h period.

Body composition analysisLongitudinal changes in body composition weretracked non-invasively in conscious animals usingan 1H-NMR analyzer (Bruker LF90 proton-NMRMinispec; Bruker Optics, The Woodlands, TX, USA)for rapid determination of total body fluid, lean andadipose tissue masses, and then returned to theirrespective cages. Initially, groups were matched fortotal body adiposity and lean body mass (week 0) andthen subsequent measurements were made weekly ondays 7, 14, 21, 28, and 35.

Glucose and ITTOGTT or ITT were performed after either chronic oracute drug treatment. Chronically exposed rats werefood restricted for 14 h, whereas acutely exposedanimals were food restricted for B5 h before theOGTT or ITT. Acutely exposed animals received drugaccording to the acute-dosing regimen (see above),whereas chronically exposed animals received halfthe daily dose of drug (6 mg kg–1) 1 h before the start ofthe OGTT or ITT. During the OGTT, basal bloodsamples and glucose measurements were obtainedand then glucose was given through oral gavage (acuteOGTT: 2.5 g glucose per kg; chronic OGTT: 1.5 gglucose per kg); more glucose was used in acutestudies as, at the time of the OGTT, animals were B1month younger than at the time of the OGTT afterchronic treatment. Blood samples were collectedthrough a tail snip at 30-min intervals for 120 min.Samples were then centrifuged at 1800 g for 10 min at4 1C; plasma was separated and frozen at �80 1C forfurther analysis. During the ITT, after basal blood

samples were collected, animals received an intraper-itoneal (i.p.) injection of insulin (0.75 U kg–1, Humu-lin-R, Lilly, Indianapolis, IN, USA) and blood glucosewas measured every 30 min for 2 h.

Hyperinsulinemic-euglycemic clamp experimentsClamp experiments were performed in chronicallycatheterized animals. Before catheter implantation,rats were housed in groups, but singly housed aftersurgery. Animals were anesthetized using volatileisofluorane (4% induction, 1.5% maintenance). Usingaseptic surgical technique, a single arterial polyethy-lene catheter (PE-50) for blood sampling was im-planted in the left carotid artery such that the tip layin the aortic arch. A venous catheter (PE-50) with asilicon tip (silicon tubing; ID: 0.51 mm, OD: 0.94 mm;Braintree Scientific) was inserted in the right jugularvein for infusion of non-radioactive glucose andinsulin. A second ‘hybrid’ venous catheter (PE-50/PE-10) was inserted for uninterrupted infusion ofradiolabeled glucose (3-[3H]-glucose). At the conclu-sion of the operation, catheters were filled with aheparinized saline solution (200 U ml–1) to preventlumen thrombus formation and sealed with metalplugs. During post-operative recovery, animals weregiven subcutaneous, 37 1C saline and intramuscularceftriaxone (100 mg kg–1). At least 4 days recovery wasallowed before conducting clamp experiments. Atthis time, body weight was greater than or within 5%of the pre-operative value. Animals not meeting thiscriterion were not used in clamp experiments.

All hyperinsulinemic-euglycemic clamp experi-ments were performed at the same time in themorning on conscious, unrestrained animals after14 h of food restriction to ensure that the animalswere in the post-absorptive state. On the eveningbefore clamp experiments, animals received eithervehicle or olanzapine (10 mg kg–1) through oral gavagebetween 1700 and 1800 hours and then foodrestricted. The following morning animals receivedanother gavage of vehicle or olanzapine (10 mg kg–1)and then were placed in small, dark cages for theclamp experiment. During the basal period (t =�120to 0 min), saline was infused through the venouscatheter and a tracer amount of tritiated glucose(3-[3H]-glucose; Perkin-Elmer, Waltham, MA, USA)was infused as a primed-continuous infusion (10 mCibolus, 0.2 mCi min–1) through the hybrid venouscatheter for measurement of basal hepatic glucoseoutput. At time zero (t = 0), a primed-continuousinsulin infusion (75 mU kg–1 bolus, 1 mU kg–1 permin, 0.3% BSA) was started and glucose (20%dextrose) was co-infused to maintain euglycemia(B100 mg per 100 ml). Blood glucose concentrationwas monitored every 10–15 min. In addition, the rateof tritiated glucose infusion was increased to 0.4mCi -min–1 to maintain a constant specific activity for themeasurement of insulin-stimulated glucose kinetics.Blood samples ( < 250ml) were collected at �20, 0, 60,120, 160, and 180 min for measurement of insulin andspecific activity of plasma glucose.

Olanzapine-induced adiposityVL Albaugh et al

3

Molecular Psychiatry

Tissue glucose uptakeA tracer amount of 1-[14C]-2-deoxyglucose (8 mCi; MPBiochemicals, Irvine, CA, USA) was given as anintravenous bolus during the steady-state phase(t = 140 min) of the hyperinsulinemic-euglycemicclamp for the determination of the in vivo rate ofglucose uptake, or tracer was alternatively given as abolus in 14 h food restricted or ad libitum fed animals.Serial blood samples were collected (t = 142, 145, 150,160, 170, 180) to determine the area under the curve(AUC) of plasma 2-deoxyglucose activity during thefinal 40 min of the experiment, or similar in vivolabeling periods in food restricted or fed animals. Atthe end of the clamp (t = 180 min), animals wereeuthanized and tissues quickly excised and immedi-ately freeze clamped in liquid nitrogen. Frozentissues were powdered and then homogenized inice-cold, 0.5 N perchloric acid (0.4 ml per 100 mgtissue). Homogenized samples were centrifuged at3000 g for 15 min. The supernatant was collected andneutralized with an equal molar amount of potassiumhydroxide and then assayed for total 14C radioactivityusing a dual-channel liquid scintillation counter(Beckman Coulter, Fullerton, CA, USA). The 14Cradioactivity in these samples represents the totalcounts from both the 2-deoxyglucose and the phos-phorylated 2-deoxyglucose present in the tissuesample. An aliquot of the neutralized extract wassubjected to Somogyi extraction, which removes thenon-phosphorylated 2-deoxyglucose and thencounted. Thus, total counts of phosphorylated2-deoxyglucose were calculated as the total countsminus the counts remaining after Somogyi extraction.

Adipose tissue fatty acid uptakeFree fatty acid (FFA) uptake by adipose tissue depotswas measured using the non-metabolizable fatty acidanalog, radioisotopic b-methyl-iodo-phenyl-pentade-canoic acid ([125I]-BMIPP), as described earlier andvalidated.26 At B0700 hours, a tracer amount of [125I]-BMIPP (10 mCi) was given as an intravenous bolus(t = 0 min). Serial blood samples ( < 250ml) werecollected (t = 2, 5, 10, 20, 30, 40) to determine theAUC for plasma [125I]-BMIPP during the 40 minin vivo labeling period. After final blood samples werecollected (t = 40 min), animals were euthanizedand adipose depots, that is epididymal, pararenal,and subcutaneous (over the abdomen) depots, wereexcised and freeze clamped with aluminum tongscooled in liquid nitrogen. Total [125I]-BMIPP activitywas measured in plasma (25 ml) and whole tissue(B100 mg) for calculation of FFA metabolic rate.Tissue and plasma 125I radioactivity was countedusing a gamma counter (Beckman Coulter).

Tissue lipogenesisMeasurement of in vivo lipogenesis using 3H2Owas measured and calculated as described earlier,27

with the exception that animals were not anesthe-tized during the isotope-labeling period. MaleSprague–Dawley rats were housed two per cage and

randomized to control or olanzapine groups with adlibitum access to water and chow. The mean bodyweights between the two groups did not significantlydiffer (Con: 340±3 g vs Olz: 340±6 g). Olanzapine(10 mg kg–1) or vehicle was administered by oralgavage at 0800 and 2000 hours on the first day oftreatment. On the morning of the second day, animalsreceived a third dose of drug or vehicle and thenreceived 4.5 mCi of 3H2O dissolved in saline 15 minafter oral gavage. Two hours after the radioisotopeinjection, tissues were quickly excised under iso-fluorane anesthesia and immediately freeze clampedin liquid nitrogen and stored at �84 1C.

Tissues were subsequently powdered and 150–300 mg was transferred to liquid nitrogen cooledscrew cap glass centrifuge tubes. Total lipid wasextracted using a two-step Bligh and Dyer extractionusing the yield increasing modification: mixture of1 M saline and 0.2 M phosphoric acid instead ofwater.28,29 The results are reported as mol 3H incorpo-rated into total lipid per hour based on individualplasma-specific activities corrected for estimatedplasma water (92%).

Isoproterenol challenge testTo gauge adipose tissue lipolytic capacity, an iso-proterenol challenge test was performed. Isoproter-enol (Isuprel) was purchased from AbbottLaboratories (Abbott Park, IL, USA) and dilutedin sterile normal saline immediately before use.Animals were given ad libitum access to food andwater and received the final vehicle or olanzapinegavage 2 h before the start of the challenge test. Aftercollection of basal blood samples, an i.p. injection(2 ml kg–1) of isoproterenol (0.05 mg kg–1) was given tovehicle and olanzapine-treated animals. Serial bloodsamples were collected at 30-min intervals for 2 h tomeasure blood glucose and to prepare plasma bycentrifugation for FFA and glycerol analysis.

Analytical procedures and metabolite measurementsAscensia Contour blood glucometers were chosen tomeasure whole blood glucose (Bayer Healthcare LLC,Mishawaka, IN, USA) because of the small samplesize required (B0.6 ml) and correction for differencesin hematocrit. Plasma insulin concentrations weredetermined by ELISA kit for rat insulin (Mercodia,Uppsala, Sweden). Glucose-specific activity wasdetermined on plasma deproteinized with equalmolar volumes of barium hydroxide and zinc sulfate.Samples were incubated at 4 1C overnight and thendried to remove 3H2O. Counts remaining represent theradioactivity from plasma 3-[3H]-glucose. Colori-metric assays were used to measure FFAs (WakoDiagnostics, Richmond, VA, USA) and glycerol(Cayman Chemical, Ann Arbor, MI, USA).

CalculationsIn chronic feeding studies ‘food-conversion effi-ciency’ was calculated as the ‘change in body weight’divided by the ‘cumulative food intake’.30 During

Olanzapine-induced adiposityVL Albaugh et al

4

Molecular Psychiatry

euglycemic clamp studies, the glucose rates ofappearance (Ra) and disappearance (Rd) were calcu-lated using the isotope dilution method.31 Briefly,glucose turnover rate was calculated as the ratio of therate of [3H]-glucose infusion (dpm per min) to theglucose-specific activity (dpm per mg). Hepaticglucose output in the basal state was defined asthe tracer-determined rate of glucose turnover. Duringthe steady-state portion of the clamp experiment(t = 120–180 min), the rates of glucose appearanceand disappearance are equal and thus the residualhepatic glucose output during the clamp equals thetotal glucose turnover minus the exogenous glucoseinfusion rate. The glucose metabolic rate, Rg (nmol g–1

tissue per min) in each tissue was determined bytaking the total counts of phosphorylated 2-deoxyglucoseand dividing by the AUC of plasma 14C-2-deoxyglucoseduring the 40 min in vivo labeling period and multi-plying by the plasma glucose concentration.32,33

Under non-euglycemic clamp conditions, tissuemetabolic clearance rate for glucose was calculatedby dividing the calculated Rg by the plasma glucoseconcentration. The adipose tissue FFA metabolic rate,Rf, was calculated by taking the total tissue counts of[125I]-BMIPP and dividing by the AUC of [125I]-BMIPPduring the 40 min in vivo labeling period and multi-plying by the mean plasma FFA concentration.26 ForITT, the inverse area under the curve (ITT-AUC) wascalculated by integrating the decline in the plasmaglucose relative to time 0, as blood glucose normallydecreases after insulin injection.

Statistical analysesFor all results, data are expressed as the mean±s.e. Tocalculate the statistical significance (P < 0.05), Stu-dent’s t-test or one-way analysis of variation withBonferroni’s multiple comparison post-test was usedwhen appropriate. A Welch correction was applied ifthe variances between groups were significantlydifferent. Asterisks indicate a particular level ofsignificance (***P < 0.001, **P < 0.01, *P < 0.05). Sam-ple sizes for each experiment are included in therespective figure legends. All statistical analyses anddata manipulations were made using GraphPad Prismor InStat computer software (GraphPad Software,San Diego, CA, USA).

Results

Effects of chronic olanzapine on male ratsEarlier work has shown that olanzapine increasesadiposity over an extended time course (B3 weeks),though the mechanism of this effect is not clear.23,24

Similar to earlier reports, neither body weight gain(Figure 1a) nor 24 h food intake (Figure 1b)differed between vehicle- and olanzapine-treated ratsthroughout the study. Moreover, food-conversionefficiency was not different during the first weekof treatment (23.1±1.4 vs 21.8±1.8%, for control-and olanzapine-treated groups, respectively) or after3 weeks of treatment (17.0±0.8 vs 15.1±1.4%).

Figure 1 Effects of chronic olanzapine on body weight, foodintake, adiposity, oral glucose tolerance, and insulin tolerance.Animals were trained to eat drug-free cookie dough andthen allocated to experimental groups for chronic olanzapineor vehicle treatment as indicated (see ‘Materials and methods’for detailed-dosing protocol). Data represent the mean±s.e. (n = 8–10). Asterisks indicate a significant difference(*P < 0.05, **P < 0.01, ***P < 0.001) in the olanzapine groupcompared with the control group for each measurement.(a, b) Body weight and 24 h food intake were measured dailyand data represent the mean±s.e. (c) Percent total bodyadiposity was measured weekly by 1H-NMR for 5 weeks.(d, e) On day 28 of olanzapine or vehicle treatment, an oralglucose tolerance test was performed. Animals were foodrestricted for 14 h and administered a half-dose of olanzapine(6 mg kg–1) or vehicle 1 h before beginning the OGTT. Baselineblood samples were collected 1 h after the gavage and then aglucose solution (1.5 g kg–1) administered orally. Serial bloodsamples were collected at 30-min intervals for 2 h after glucosegavage for measurement of (d) blood glucose and (e) plasmainsulin. Data represent the mean±s.e. (n = 5–10). (f) On day 42of olanzapine treatment, an insulin tolerance test wasperformed. Animals were food restricted for 14 h before thetest and given a half-dose of olanzapine (6 mg kg–1) or vehicle1 h before the challenge. The response to injected insulin wasmeasured for 120 min as the change in baseline blood glucose.

Olanzapine-induced adiposityVL Albaugh et al

5

Molecular Psychiatry

In contrast to weight gain and food intake, olanzapineincreased percent of total body fat by the seventhday of treatment (Figure 1c). The increase intotal body fat relative to control animals was main-tained for the duration of the study, but reached aplateau at 3 weeks. Percentages of adipose, lean,and fluid masses over the course of the study areshown in Table 1. Accordingly, as olanzapinedid not change body weight gain during chronicadministration, the observed increase in fat masswas balanced by a decrease in total body lean mass(Table 1).

Physical activity was measured during the third weekof olanzapine or vehicle treatment, as decreased activityhas been implicated in olanzapine-induced obesity infemale rats.15,34 However, locomotor activity did notdiffer between experimental groups during the thirdtreatment week [‘X dimension’ 25 820±1121 vs 22 445±1616; ‘Y dimension’ 14 589±820 vs 12 201±1095;‘Z dimension’ 2431±285 vs 1899±278 (n = 10), allvalues ‘control vs olanzapine’, respectively].

Increased adiposity is strongly associated withinsulin resistance, and thus we challenged animalswith an oral glucose tolerance test after 4 weeks ofolanzapine treatment. Consistent with their increasedadiposity, the olanzapine-treated animals displayedhyperglycemia (Figure 1d) and hyperinsulinemia(Figure 1e) after 14 h of food restriction. Glucose and

insulin concentrations remained significantly ele-vated compared with vehicle-treated controls through-out the OGTT. Moreover, the AUCGlucose was signifi-cantly elevated by olanzapine [194±78 vs 401±134(mM �min), P < 0.01], suggesting decreased whole-body insulin sensitivity. To confirm such changes,an insulin tolerance test was performed on day 42of drug treatment. Once again, drug-treated animalsshowed elevated plasma glucose after 14 h foodrestriction (5.0±0.2 vs 6.1±0.2 mM, P < 0.001).Olanzapine blunted the fall in blood glucoseafter insulin injection (Figure 1f), with glucoseconcentrations remaining elevated compared withcontrol levels at 60, 90, and 120 min post-injection.The ITT-AUC was lower in the olanzapine groupcompared with control animals [6226±564 vs3797±569 (min �% baseline), P < 0.01], which furthersuggests decreased insulin sensitivity.

Effect of acute olanzapine on locomotor activity andenergy expenditureAs primary disturbances in energy expenditure mightincrease adiposity and weight gain over a similar timeperiod, we examined potential drug effects on theseparameters. Locomotor activity and energy expendi-ture were measured after acute oral olanzapineadministration (Figures 2b and c, indicated byarrows) at the beginning of the dark cycle (shadedarea) and then again at the beginning of the light cycleon the following morning. As seen in Figure 2a,olanzapine significantly decreased locomotor activityin all dimensions (x, y, z) during the dark and lightcycles, but had no effect on 24 h food intake (data notshown); thus relative food intake was more thanexpected for the amount of physical activity. Leptin,an adipocyte-derived hormone and known satietyfactor, has been posited as contributing to thehyperphagia observed in female rats after olanzapinetreatment.14 Consistent with those observations, plas-ma leptin concentration was also decreased afteracute olanzapine-treated male rats [5 h fast: con-trol = 5.29±0.55 (n = 9), Olz = 3.16±0.44 (n = 10),P < 0.01]. Thus, decreases in plasma leptin may alsocontribute to the relative increase in food intake.

O2 consumption (Figure 2b) and CO2 production(Figure 2c) were measured throughout a completedark and light cycle. Unexpectedly, the olanzapine-treated group had a similar VO2 (Figure 2d) and VCO2

(Figure 2e) relative to vehicle-treated controls duringthe dark cycle, whereas these parameters weresignificantly elevated during the light cycle. Rectaltemperature was also measured in separate groupsof animals before and 2 h after olanzapine gavage,though no change from baseline body temperaturewas detectable (data not shown). Thus, energyexpenditure is not decreased after olanzapine ad-ministration, even though drug-treated animalshave less locomotor activity. These results show thatolanzapine-treated rats use some of the excess caloricintake to support a process other than locomotoractivity.

Table 1 Minispec proton-NMR body composition dataduring chronic olanzapine treatment (4–12 mg kg–1) inmale rats

Time Weightcomponent

Control(%)

Olanzapine(%)

Week 0 Adipose 20.7±0.5 20.9±0.6Lean 65.1±0.7 65.8±0.5Fluid 6.3±0.2 6.2±0.2

Week 1 Adipose 20.7±0.8 22.6±0.4*Lean 59.2±1.4 58.3±1.3Fluid 5.3±0.2 5.0±0.2

Week 2 Adipose 23.3±0.7 25.3±0.5*Lean 63.5±0.8 62.1±0.6Fluid 5.7±0.3 5.3±0.2

Week 3 Adipose 25.2±0.5 27.5±0.4**Lean 63.4±0.3 62.1±0.4*Fluid 5.7±0.1 5.6±0.1

Week 4 Adipose 24.7±0.6 27.4±0.4**Lean 61.3±0.9 60.5±0.7Fluid 5.5±0.3 5.2±0.2

Week 5 Adipose 25.7±0.7 27.6±0.5*Lean 62.8±0.6 61.3±0.3*Fluid 6.2±0.1 6.2±0.1

Abbreviation: NMR, nuclear magnetic resonance.Values are mean±s.e.; n = 10 rats per group; data representthe average component mass in percent of total bodyweight; asterisks indicate significantly different (*P < 0.05,**P < 0.01) compared with measurements from time-matched control animals.

Olanzapine-induced adiposityVL Albaugh et al

6

Molecular Psychiatry

Effects of acute olanzapine on glucose tolerance andinsulin action

The worsening glucose and insulin tolerance ob-served after chronic treatment with olanzapine mightbe secondary to increased adiposity. To test thishypothesis, OGTTs and ITTs were conducted inseparate cohorts of rats on the second treatment day

before significant accumulation of adipose tissue. Asseen in Figure 3, acute olanzapine-treated ratsdisplayed mild hyperglycemia (Figure 3a) and hyper-insulinemia (Figure 3b) after 5 h of food restrictionafter a low dose of olanzapine (4 mg kg–1 per day),suggesting insulin resistance. During the OGTT,glucose concentration tended to be higher relative tocontrol values, with baseline and 120 min bloodglucose measurements being significantly elevatedin the olanzapine group. Plasma insulin was similarbetween experimental groups throughout the OGTT,even though increased circulating insulin was ex-pected given the mild elevations in blood glucose. Tobetter assess insulin sensitivity, an insulin tolerancetest (Figure 3c) was conducted in vehicle-treated ratsor in rats injected with one of two doses of olanzapine(4 or 10 mg kg–1 per day). Olanzapine administrationled to a dose-dependent decrease in inverse ITT-AUC(Figure 3c). The 4 mg kg–1 per day dose decreased theinverse ITT-AUC relative to the control group[6054±344 vs 4436±415 (min �% change), P < 0.01],whereas the 10 mg kg–1 per day dose further decreasedthe inverse ITT-AUC [4436±415 vs 1514±296(min �% change), P < 0.001]. Together, these datashow that insulin resistance after olanzapine admin-istration precedes and is, therefore, independent ofthe change in body composition observed in thechronically treated animals.

As drug-induced insulin resistance precedes thechanges in body composition, it is tempting tospeculate that it might have a function in thedevelopment of the olanzapine-induced adiposity.For example, Caro et al.35 proposed a mechanism ofobesity wherein tissue-specific insulin resistance, inwhich muscle but not fat was affected, would lead toshunting of substrates away from muscle and towardfat. To test this hypothesis, hyperinsulinemic-eugly-cemic clamp studies using tracer methodology wereconducted on the second treatment day. Plasmainsulin concentration was maintained at a mild,hyperinsulinemic level to examine drug effects onperipheral tissues as well as hepatic insulin sensitiv-ity (plasma insulin: 200±28 vs 205±25 pM in controlvs olanzapine-treated animals, respectively). Duringthe basal state (14 h food restricted), the tracer-determined rate of glucose turnover, which is equiva-lent to hepatic glucose output (Figure 3d), did notdiffer between experimental groups. Moreover, in-sulin-mediated suppression of hepatic glucose out-put, a measure of hepatic insulin sensitivity, wassimilar in the control and olanzapine groups.Consistent with two earlier studies,36,37 the tracer-calculated rate of whole-body glucose disposal duringthe clamp was significantly decreased in the olanza-pine group (Figure 3e). Insulin-mediated glucosedisposal was increased by B85% in control animals,whereas olanzapine-treated animals showed no in-crease in whole-body glucose disposal, confirming theinsulin resistance suggested by the glucose and ITT.

To determine the tissue(s) responsible for theinsulin resistance, tissue glucose metabolic rates (Rg)

Figure 2 Energy expenditure and locomotor activity afteracute olanzapine administration. Animals were placed inspecialized chambers to measure locomotor activity andenergy expenditure using indirect calorimetry. After accli-mation, olanzapine (10 mg kg–1) or vehicle solution wasadministered by oral gavage (indicated by arrows). Animalsretained ad libitum access to food and water for the durationof the experiment. (a) VO2 and (b) VCO2 were measured at15-min intervals for 24 h. (c) 24 h locomotor activity wasmeasured as an indicator of physical activity. Backgroundshading or lack thereof indicates the dark and light cycles,respectively. Locomotor activity was significantly differentfrom control in each dimension during the dark and lightcycles (P < 0.001). (d) Average VO2 and (e) average VCO2 forthe dark and light cycles were calculated. All data representthe mean±s.e. (n = 12), asterisks indicate significant differ-ences (***P < 0.001, **P < 0.01).

Olanzapine-induced adiposityVL Albaugh et al

7

Molecular Psychiatry

were calculated for muscle, heart, liver, and adiposetissues during the steady-state portion of the eugly-cemic clamp (Table 2). Fed insulin concentrationswere not significantly different (data not shown) andthe plasma insulin was clamped to approximate theinsulin in the fed state. Consistent with the decreasein insulin-stimulated whole-body glucose disposal,Rg was decreased by 39% in the heart as well as by31% in gastrocnemius and 40% in soleus, whichrepresent both fast- and slow-twitch skeletal musclefibers, respectively. Unlike skeletal muscle Rg, a 53%increase in subcutaneous adipose tissue glucoseuptake was observed relative to control animals,whereas epididymal and pararenal adipose tissue Rg

remained unchanged. Thus, olanzapine acutely per-turbs insulin sensitivity in a tissue-specific manner,decreasing insulin sensitivity in muscle tissue, whileeither increasing or not affecting sensitivity inadipose depots.

Tissue-specific glucose uptake was also measuredin 14 h food-restricted animals, under similar condi-tions as the basal period of the clamp studies. Tissuemetabolic clearance rate instead of tissue Rg is shownin Table 2, as olanzapine elevated glucose and insulinconcentrations under these conditions. Despite theinsulin resistance, tissue clearance of glucose byskeletal muscle and two adipose depots was notaffected, presumably because the plasma glucose andinsulin were elevated. However, clearance by sub-cutaneous adipose tissue was increased. Basal uptakewas also increased in the heart.

Effects of olanzapine on adipose tissue FFA fluxIn 14 h food-restricted animals, acute olanzapinetreatment decreased plasma FFA (Figure 4a), as wellas glycerol [17.5±1.3 vs 9.6±1.1, P < 0.001]. This wassurprising as insulin resistance is normally associatedwith elevated FFA. To examine the mechanism

Figure 3 Effects of acute olanzapine on oral glucose tolerance, insulin tolerance, and whole-body insulin sensitivity.Asterisks indicate significant differences compared with control (*P < 0.05, ***P < 0.001). (a, b) An oral glucose tolerance testwas conducted on the second treatment day after acute olanzapine (4 mg kg–1) gavage (see ‘Materials and methods’ for dosingprotocol). Animals were food restricted for 5 h before the OGTT and received a final olanzapine or vehicle gavage 1 h beforethe start of the OGTT. Baseline blood samples were collected at 1 h after gavage and then glucose solution (2.5 g kg–1)administered orally. Serial blood samples were collected at 30-min intervals for 2 h after glucose gavage for measurement of(a) blood glucose and (b) plasma insulin. Data represent the mean±s.e. (n = 12). (c) An insulin tolerance test was performedin another group of animals on the second treatment day. Animals received high-dose olanzapine (10 mg kg–1), low-doseolanzapine (4 mg kg–1), or vehicle solution by oral gavage 1 h before beginning the tolerance test. Baseline blood samples werecollected and then insulin (0.75 U kg–1, i.p.) injected. Blood glucose was measured for 120 min as an indicator of insulinsensitivity. Data represent the mean±s.e. (n = 18–20). (d, e) Hyperinsulinemic-euglycemic clamp studies were conductedafter acute olanzapine administration (10 mg kg–1). (d) Hepatic glucose output was measured during basal (14 h foodrestricted) and clamp conditions. Basal hepatic glucose output (HGO) was equal to the total glucose turnover during the basalperiod, whereas the clamp was the residual HGO during the final 40 min of the clamp. (e) Whole-body glucose disposal, ameasure of whole-body insulin action, was measured during hyperinsulinemic-euglycemic clamp conditions and comparedwith basal glucose turnover. Data represent the mean±s.e. of glucose turnover during the last 40 min of the clampexperiment (n = 10–14). (f) Adipose tissue glucose uptake in ad libitum fed animals was measured after acute olanzapineadministration (10 mg kg–1). Data represent the mean±s.e. (n = 8).

Olanzapine-induced adiposityVL Albaugh et al

8

Molecular Psychiatry

Table 2 Tissue metabolic clearance rates and insulin-stimulated glucose metabolic rates for individual tissues duringeither basal or euglycemic-hyperinsulinemic clamp conditions in male rats on the second day of acute olanzapine treatment(10 mg kg–1)

Tissue Basal conditions metabolic clearancerate (ml g–1 per tissue per min)

% Changefrom

control

Clamp conditions glucose metabolicrate (nmol g–1 per tissue per min)

% Changefrom

control

Control Olanzapine Control Olanzapine

Gastrocnemius 1.28±0.10 1.62±0.24 ns 198±20 136±11 �31%*Soleus 2.36±0.28 3.20±0.46 ns 381±25 229±36 �40%**Heart 36.5±5.6 124.9±24.5 þ 242%*** 375±37.0 221±26.7 �39**Epididymal adipose 0.67±0.09 0.59±0.10 ns 51±6 51±8 nsPararenal adipose 0.67±0.12 0.74±0.09 ns 65±13 75±8 nsSubcutaneous adipose 1.62±0.25 2.68±0.37 þ 65%* 89±8 136±16 þ 53%*Liver 10.71±1.46 11.04±2.05 ns 195±23 202±16 ns

Values are mean±s.e.; n = 6–14 per tissue; asterisks indicate significant differences (*P < 0.05, **P < 0.01) compared withcontrol; ns, not significantly different from control.

Figure 4 Acute effects of olanzapine on circulating and adipose tissue FFA uptake, lipogenesis, and mobilization of FFAand glucose in response to an isoproterenol challenge. (a) Plasma FFAs were measured in 14 h food-restricted animals afterolanzapine (10 mg kg–1) gavage. (b) In a separate cohort, FFA uptake into adipose tissue was measured on the second day ofolanzapine treatment using a non-metabolizable FFA analog (125I-BMIPP) in animals that had ad libitum access to food andwater. Two hours after the final olanzapine dose (10 mg kg–1), an intravenous bolus of the FFA tracer was given and bloodsamples were collected during a 40-min in vivo labeling period. Epididymal, pararenal, and subcutaneous fat pads wereharvested and measured for tracer uptake, an index of adipocyte FFA uptake. (c) In another cohort of animals, adipose tissuelipogenesis was measured in the well-fed state after acute olanzapine (10 mg kg–1). After final administration of olanzapine,3H2O was administered through i.p. injection and tissue samples were collected after a 120-min in vivo labeling period.Epididymal, retroperitoneal, and subcutaneous adipose tissues were excised, extracted for total lipid, and then counted for3H content as an index of lipogenesis. (d–f) An isoproterenol challenge test was conducted to assess the effects of olanzapine(10 mg kg–1) on isoproterenol-stimulated lipolysis and hepatic glucose output. Animals had ad libitum access to food andwater. Baseline blood samples were collected and then isoproterenol (0.01 mg kg–1) administered. Serial blood samples at30-min intervals for 2 h were collected to measure the lipolytic response, as measured by the change in (c) FFA and (d)glycerol from isoproterenol-stimulated lipolysis, as well as the (e) hepatic glycolytic response. Data represent the mean±s.e.(n = 8–10). Asterisks indicate significant differences compared with time-matched control values (*P < 0.05, ***P < 0.001,****P < 0.0001).

Olanzapine-induced adiposityVL Albaugh et al

9

Molecular Psychiatry

underlying this effect and to further elucidate themechanisms of elevated adiposity development inthis model, we measured the uptake of FFA intoadipose tissue, rates of in vivo lipogenesis, andFFA mobilization after a mild-dose isoproterenolchallenge under ad libitum fed conditions in whichFFA levels were not different (data not shown).

FFA metabolic rates were measured using the non-metabolizable FFA analog 125I-BMIPP in ad libitumfed animals (Figure 4b). Olanzapine increasedFFA metabolic rate twofold in epididymal, pararenal,and subcutaneous adipose depots.

Given that adipose tissue uptake of FFA and/orglucose tended to be increased by olanzapine, wehypothesized that lipogenesis might be also elevated,which could potentially explain the elevated oxygenconsumption and chronic increases in fat mass indrug-treated animals. To test this hypothesis, wemeasured in vivo lipogenesis in fed rats by measuringadipose tissue incorporation of 3H2O into triglyceride.As seen in Figure 4c, a trend for increased epididymaladipose tissue lipogenesis was detectable (P = 0.1),though significant increases in lipogenesis (P <0.0001) were detectable in both the retroperitonealand subcutaneous adipose tissue depots (75 and266% higher, respectively).

Adipocytes and hepatocytes mobilize, respectively,stored fatty acids and glucose fuels in the post-absorptive state, during longer periods of fooddeprivation, or in times of stress when catechola-mines are elevated. Lipolytic impairment could serveas an additional pathway leading to the chronicchanges in body composition by trapping stored fuels,and thus we tested lipolytic capacity in vivo with anisoproterenol challenge test in fed animals in whichthe FFA are not different. After injection of a low doseof isoproterenol, rises of plasma glycerol (Figure 4d)and FFAs (Figure 4e) were blunted by olanzapine.Even though lipolysis was attenuated, peak isoproter-enol-stimulated hepatic glucose output remainedintact (Figure 4f). Thus, the effects of olanzapine onlipolysis probably do not involve the early steps inthe b-adrenergic signaling pathway shared by thelipolytic and hepatic glucose output response toisoproterenol. Overall, through the above actions,olanzapine seems to concurrently promote energyrepartitioning to and trapping within triglyceridestores of adipose tissue.

Discussion

In this paper, we confirm that chronic olanzapineadministration in male rats increases total bodyadiposity independent of a change in either bodyweight gain or food intake. The increased fat masswas detectable by 1H-NMR within the first week ofdrug administration and is associated with impairedoral glucose tolerance and insulin resistance. Thisstudy extends these original findings by showing thatolanzapine has at least four acute metabolic effectsthat may act in concert to favor adipose tissue

deposition and development of insulin resistance.First, olanzapine decreases physical activity withouta compensatory reduction in food intake, resulting ina relative over-nutrition in drug-treated animals. Thiswas associated with a decline in the satiety hormoneleptin. Second, olanzapine causes tissue-specificchanges in insulin sensitivity that support a tendencyfor increased glucose uptake into adipose tissue.Third, FFA uptake into adipose tissue was alsoincreased, along with drug-stimulated increases inadipose tissue lipogenesis in some depots. Last,olanzapine impaired in vivo lipolysis during anisoproterenol challenge, which suggests that adiposetissue is functionally trapping newly synthesized andstored metabolic fuels. These acute effects of olanza-pine are present on the second treatment day and areassociated with detectable increases in adipose tissueby the seventh treatment day. In this model, theaccumulation of adipose tissue is self-limiting, reach-ing a plateau after B3 weeks of dose ramping, butmost likely further worsens the drug-induced insulinresistance and glucose intolerance observed beforebody composition is affected.

Decreased physical activity without a compensatorydecrease in food intakeA prolonged decrease in physical activity withoutreduction of food intake favors the accumulation ofadipose tissue by excess nutrient storage. In contrastto human beings, olanzapine-induced weight gainand hyperphagia in rats is sexually dimorphic.Studies have shown that female rats, but not males,are susceptible to drug-induced weight gain14,34, andour results are consistent with those findings.23,24

However, although female rats show decreased activ-ity and increased food intake,15,34 male rats showdecreased activity without changes in food intake.Decreased locomotor activity was not observed at 3weeks, suggesting that this effect lessens with time.Interestingly, percent body fat also reached plateauover a similar time course. Thus, in the short term,male rats have a relative over-nutrition for their givenlevel of activity. It is tempting to speculate that thismismatch is due to the drug-induced reduction inplasma leptin concentrations we observed here andearlier in females. Although obesity drives higherplasma leptin concentrations in human beings, it hasbeen noted that plasma leptin concentrations tend tobe lower in patients receiving olanzapine thanexpected based on BMI.38 The reason for lack ofweight gain despite eating more food than expectedfor physical activity seems to be due to maintainedenergy expenditure. The energy-consuming processmight represent activation of an adaptive thermogen-esis. However, that should decrease body weight,which is not consistent with deposition of adiposetissue and loss of lean mass that occurs with chronicolanzapine treatment. A more likely alternative is thatenergy expenditure is increased, at least in part, bymale rats expending energy for de novo lipogenesis,which is supported by the observed increase in body

Olanzapine-induced adiposityVL Albaugh et al

10

Molecular Psychiatry

fat and elevated lipogenesis found in several adiposetissue depots. In individuals consuming a mixedmacronutrient diet (carbohydrates, fat, protein)de novo lipogenesis usually accounts for a small fractionof total energy expenditure ( < 5%), as most storedlipid is recycled from dietary sources.39 However,de novo lipogenesis from carbohydrate precursors canaccount for upward of B20–30% of energy expendi-ture.40,41 The standard rat chow used in this study isB6% crude fat and, therefore, increased adipositywould be thought to come mostly from carbohydrateprecursors. It is noteworthy that adipose tissue didnot exhibit the insulin resistance observed in skeletalmuscle and, therefore, was under different conditionsof either increased glucose uptake or a trend forincreased glucose uptake. Moreover, the percentincrease of lipogenesis observed in subcutaneousadipose tissue was much great than that of epididy-mal or retroperitoneal fat, consistent with the appar-ent increase in insulin sensitivity observed in thesubcutaneous depot as measured by the increasedinsulin-stimulated glucose uptake during a euglyce-mic clamp.

Increased energy uptake into fatTissue-specific insulin resistance has been posited tounderlie the pathogenesis of obesity in human beingsand monogenetic obesities in rodents.35 We notedsuch tissue-specific insulin resistance after acuteolanzapine. Under some conditions, a tendency orsignificant increase in glucose uptake into fat wasobserved. In addition, there was a large increase inFFA uptake into adipose tissue in fed animals. In arecent study we showed that other peripheral tissueswere similarly affected, and this may explain thesparing of glucose/hyperglycemia and apparent mus-cle insulin resistance after acute olanzapine.14 In-creased glucose and FFA uptake seems to be anadditional mechanism through which olanzapinecontributes to increased adiposity during chronictreatment.

It should be emphasized that these effects are acuteand, therefore, not secondary to increased adiposity,but may instead promote the accumulation of fat seenwith chronic treatment. Earlier studies have docu-mented weight-gain-independent insulin resistanceduring hyperinsulinemic-euglycemic clamp stu-dies.36,37 Consistent with those reports, we show asimilar decrease in insulin-stimulated whole-bodyglucose disposal (Figure 3e). Furthermore, this de-creased glucose disposal is due to skeletal muscleinsulin resistance (Table 2), a conclusion suggestedfrom earlier studies, but most likely undetectedbecause of small sample size.37 In contrast, glucoseuptake in epididymal and pararenal adipose tissuewas unchanged, whereas subcutaneous adipose tissueshowed increased glucose flux. The mechanism bywhich skeletal muscle becomes insulin resistant andnot adipose tissue is currently unknown. Moreover,the differential effects on glucose flux in visceral andsubcutaneous adipose depots is not known, but may

be due to intrinsic differences between these distincttissue subtypes.42–44 As olanzapine elevated plasmaglucose and insulin in the fed and food-restrictedstates, excess glucose should be repartition to adiposetissue with the prevailing differences in tissue insulinsensitivity.

Increased adipose tissue glucose uptake relativeto skeletal muscle would be expected to increaseadipocyte triglyceride storage, which is further po-tentiated by the nearly twofold increase in FFA fluxpresent in visceral and subcutaneous fat depots. Theincreased FFA flux likely contributes to triglyceridestorage by providing FFAs for esterification tophosphoglycerol moieties from glycolysis. Thesechanges in glucose and FFA flux are consistent withincreased adiposity and adipocyte hypertrophyreported earlier.45 As tissue FFA uptake is both passiveand regulated by insulin-sensitive fatty acid trans-porters,46 the mechanism of the increased tissue FFAuptake may involve either increased fatty acidoxidation, stimulation of fatty acid transporters by adirect drug effect, or by drug-induced increases inplasma insulin affecting lipogenic pathways. In-creased VO2 during the light cycle supports thetheory of increased lipogenesis.

Impaired fuel mobilization from adipose tissue

Another factor contributing to increased adiposityduring olanzapine administration is impaired mobi-lization of stored fuels. Normally, stored adiposetriglyceride is mobilized through lipolysis during thepost-absorptive state to conserve glucose fuel for thebrain. Alternatively, FFA are mobilized in times ofstress by elevated plasma catecholamines. Olanza-pine blunted increases in plasma FFA and glycerol in14 h food-restrict rats, suggesting lipolytic impair-ment that was confirmed by decreased in vivolipolysis after injected isoproterenol in fed animalswhen FFA were not significantly different betweenthe experimental groups. Consistent with these find-ings, decreased plasma FFA have also been reportedin patients and healthy volunteers after acute(2–3 weeks) olanzapine treatment.12,13 Moreover,catecholamine-stimulated lipolysis is attenuated inadipose tissue harvested from chronically treatedrats,45 suggesting a primary defect in the adipocyte.The mechanism of the decreased lipolysis is currentlyunknown, though several hypotheses may be con-sidered. Decreased sympathetic tone will reducelipolysis;47 however, this is likely not responsiblebecause the response to exogenous catecholaminesremains intact. Earlier radioligand binding studieshave not detected binding of olanzapine tob2-adrenergic receptors48 and hepatic glucose output,which is regulated by b2-adrenergic receptors inresponse to isoproterenol, was not affected. It ispossible that olanzapine inhibits b3-adrenergic recep-tors that are found only on adipose tissue. As many ofthe early steps in b-adrenergic signaling are sharedbetween b2- and b3-adrenergic receptors in fat and

Olanzapine-induced adiposityVL Albaugh et al

11

Molecular Psychiatry

b2-adrenergic receptors in liver, it seems unlikely thatthese are the target of the impairment.

In summary, olanzapine increases the accumulationof fat mass in male rats independent of a change inbody weight gain and food intake. Current findingsshow that olanzapine has at least five effects thatcollectively predispose male rats to increased adip-osity, namely decreased physical activity, tissue-specific insulin resistance leading to repartitioningof glucose fuel, increased uptake of FFA into fat,increased adipose tissue lipogenesis, and impairedlipolysis, which traps excess fuel taken up as well asnewly formed triglyceride. Further studies are neededto identify the molecular mechanisms leading to theseearly drug effects and determine their relevance toantipsychotic-induced weight gain and insulin resis-tance seen clinically.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgments

The expert technical assistance of Beth Halle Gern,Stephanie Goshorn, Jamie G Spicer, and Ali Nairizi isgreatly appreciated. VLA is the recipient of a SeedGrant from the American Medical Association, Foun-dation. JGJ was supported by an institutional NIDDKSTEP-UP grant (R25 DK078381). This project wassupported by NIH (DK062880) and the PennsylvaniaDepartment of Health using Tobacco SettlementFunds. The Department of Health specifically dis-claims responsibility for any analyses, interpreta-tions, or conclusions.

References

1 Ananth J, Parameswaran S, Gunatilake S. Side effects of atypicalantipsychotic drugs. Curr Pharm Des 2004; 10: 2219–2229.

2 Ebenbichler CF, Laimer M, Eder U, Mangweth B, Weiss E, Hofer Aet al. Olanzapine induces insulin resistance: results from aprospective study. J Clin Psychiatry 2003; 64: 1436–1439.

3 Haupt DW, Newcomer JW. Hyperglycemia and antipsychoticmedications. J Clin Psychiatry 2001; 62(Suppl 27): 15–26; discus-sion 40–41.

4 Newcomer JW. Abnormalities of glucose metabolism associatedwith atypical antipsychotic drugs. J Clin Psychiatry 2004;65(Suppl 18): 36–46.

5 Newcomer JW. Metabolic risk during antipsychotic treatment. ClinTher 2004; 26: 1936–1946.

6 Newcomer JW. Second-generation (atypical) antipsychotics andmetabolic effects: a comprehensive literature review. CNS Drugs2005; 19(Suppl 1): 1–93.

7 Fertig MK, Brooks VG, Shelton PS, English CW. Hyperglycemiaassociated with olanzapine. J Clin Psychiatry 1998; 59: 687–689.

8 Avella J, Wetli CV, Wilson JC, Katz M, Hahn T. Fatal olanzapine-induced hyperglycemic ketoacidosis. Am J Forensic Med Pathol2004; 25: 172–175.

9 Torrey EF, Swalwell CI. Fatal olanzapine-induced ketoacidosis.Am J Psychiatry 2003; 160: 2241.

10 Newcomer JW. Metabolic considerations in the use of antipsycho-tic medications: a review of recent evidence. J Clin Psychiatry2007; 68(Suppl 1): 20–27.

11 Sacher J, Mossaheb N, Spindelegger C, Klein N, Geiss-Granadia T,Sauermann R et al. Effects of olanzapine and ziprasidone onglucose tolerance in healthy volunteers. Neuropsychopharmaco-logy 2008; 33: 1633–1641.

12 Kaddurah-Daouk R, McEvoy J, Baillie RA, Lee D, Yao JK,Doraiswamy PM et al. Metabolomic mapping of atypical anti-psychotic effects in schizophrenia. Mol Psychiatry 2007; 12: 934–945.

13 Vidarsdottir S, de Leeuw van Weenen JE, Frolich M, Roelfsema F,Romijn JA, Pijl H. Effects of olanzapine and haloperidol on themetabolic status of healthy men. J Clin Endocrinol Metab 2010; 95:118–125.

14 Albaugh VL, Henry CR, Bello NT, Hajnal A, Lynch SL, Halle Bet al. Hormonal and metabolic effects of olanzapine and clozapinerelated to body weight in rodents. Obesity 2006; 14: 36–51.

15 Arjona AA, Zhang SX, Adamson B, Wurtman RJ. An animal modelof antipsychotic-induced weight gain. Behav Brain Res 2004; 152:121–127.

16 Cooper GD, Pickavance LC, Wilding JP, Halford JC, Goudie AJ. Aparametric analysis of olanzapine-induced weight gain in femalerats. Psychopharmacology (Berl) 2005; 181: 80–89.

17 Goudie AJ, Smith JA, Halford JC. Characterization of olanzapine-induced weight gain in rats. J Psychopharmacol 2002; 16: 291–296.

18 Pouzet B, Mow T, Kreilgaard M, Velschow S. Chronic treatmentwith antipsychotics in rats as a model for antipsychotic-inducedweight gain in human. Pharmacol Biochem Behav 2003; 75: 133–140.

19 Lee MD, Clifton PG. Meal patterns of free feeding rats treated withclozapine, olanzapine, or haloperidol. Pharmacol Biochem Behav2002; 71: 147–154.

20 Thornton-Jones Z, Neill JC, Reynolds GP. The atypical antipsy-chotic olanzapine enhances ingestive behaviour in the rat:a preliminary study. J Psychopharmacol 2002; 16: 35–37.

21 Hartfield AW, Moore NA, Clifton PG. Effects of clozapine,olanzapine and haloperidol on the microstructure of ingestivebehaviour in the rat. Psychopharmacology (Berl) 2003; 167:115–122.

22 Benvenga MJ, Leander JD. Increased food consumption byclozapine, but not by olanzapine, in satiated rats. Drug Dev Res1997; 41: 48–50.

23 Minet-Ringuet J, Even PC, Goubern M, Tome D, Beaurepaire RD.Long term treatment with olanzapine mixed with the food in malerats induces body fat deposition with no increase in body weightand no thermogenic alteration. Appetite 2006; 46: 254–262.

24 Cooper GD, Pickavance LC, Wilding JP, Harrold JA, Halford JC,Goudie AJ. Effects of olanzapine in male rats: enhanced adiposityin the absence of hyperphagia, weight gain or metabolicabnormalities. J Psychopharmacol 2007; 21: 405–413.

25 Ader M, Kim SP, Catalano KJ, Ionut V, Hucking K, Richey JM et al.Metabolic dysregulation with atypical antipsychotics occurs in theabsence of underlying disease: a placebo-controlled study ofolanzapine and risperidone in dogs. Diabetes 2005; 54: 862–871.

26 Shearer J, Coenen KR, Pencek RR, Swift LL, Wasserman DH,Rottman JN. Long chain fatty acid uptake in vivo: comparison of[125I]-BMIPP and [3H]-bromopalmitate. Lipids 2008; 43: 703–711.

27 Robinson AM, Girard JR, Williamson DH. Evidence for a role ofinsulin in the regulation of lipogenesis in lactating rat mammarygland. Measurements of lipogenesis in vivo and plasma hormoneconcentrations in response to starvation and refeeding. Biochem J1978; 176: 343–346.

28 Hajra AK. On extraction of acyl and alkyl dihydroxyacetonephosphate from incubation mixtures. Lipids 1974; 9: 502–505.

29 Bligh EG, Dyer WJ. A rapid method of total lipid extraction andpurification. Can J Biochem Physiol 1959; 37: 911–917.

30 Ivy JL, Cortez MY, Chandler RM, Byrne HK, Miller RH. Effects ofpyruvate on the metabolism and insulin resistance of obese Zuckerrats. Am J Clin Nutr 1994; 59: 331–337.

31 Bergman RN, Hope ID, Yang YJ, Watanabe RM, Meador MA, YounJH et al. Assessment of insulin sensitivity in vivo: a critical review.Diabetes Metab Rev 1989; 5: 411–429.

32 Crist GH, Xu B, Lanoue KF, Lang CH. Tissue-specific effects of invivo adenosine receptor blockade on glucose uptake in Zuckerrats. FASEB J 1998; 12: 1301–1308.

Olanzapine-induced adiposityVL Albaugh et al

12

Molecular Psychiatry

33 Lang CH, Dobrescu C, Meszaros K. Insulin-mediated glucoseuptake by individual tissues during sepsis. Metab: Clin Exp 1990;39: 1096–1107.

34 Fell MJ, Anjum N, Dickinson K, Marshall KM, Peltola LM, VickersS et al. The distinct effects of subchronic antipsychotic drugtreatment on macronutrient selection, body weight, adiposity, andmetabolism in female rats. Psychopharmacology (Berl) 2007; 194:221–231.

35 Caro J, Sinha M, Dohm G, California Uo. Insulin resistance inobesity. In: Bray G, Ricquier D, Spiegelman B (eds). UCLASymposia on Molecular and Cellular Biology. Wiley-Liss: Key-stone, Co, 1989, pp 203–218.

36 Chintoh AF, Mann SW, Lam L, Giacca A, Fletcher P, Nobrega J et al.Insulin resistance and secretion in vivo: effects of different anti-psychotics in an animal model. Schizophr Res 2009; 108: 127–133.

37 Houseknecht KL, Robertson AS, Zavadoski W, Gibbs EM, JohnsonDE, Rollema H. Acute effects of atypical antipsychotics on whole-body insulin resistance in rats: implications for adverse metaboliceffects. Neuropsychopharmacology 2007; 32: 289–297.

38 Haupt DW, Luber A, Maeda J, Melson AK, Schweiger JA,Newcomer JW. Plasma leptin and adiposity during antipsychotictreatment of schizophrenia. Neuropsychopharmacology 2005; 30:184–191.

39 Acheson KJ, Schutz Y, Bessard T, Ravussin E, Jequier E, Flatt JP.Nutritional influences on lipogenesis and thermogenesis after acarbohydrate meal. Am J Physiol 1984; 246(1 Pt 1): E62–E70.

40 Chwalibog A, Thorbek G. Energy expenditure by de novolipogenesis. Br J Nutr 2001; 86: 309.

41 Acheson KJ, Schutz Y, Bessard T, Anantharaman K, Flatt JP,Jequier E. Glycogen storage capacity and de novo lipogenesisduring massive carbohydrate overfeeding in man. Am J Clin Nutr1988; 48: 240–247.

42 Jensen MD, Johnson CM. Contribution of leg and splanchnic freefatty acid (FFA) kinetics to postabsorptive FFA flux in men andwomen. Metabolism 1996; 45: 662–666.

43 Nelson RH, Basu R, Johnson CM, Rizza RA, Miles JM. Splanchnicspillover of extracellular lipase-generated fatty acids in overweightand obese humans. Diabetes 2007; 56: 2878–2884.

44 Nielsen S, Guo Z, Johnson CM, Hensrud DD, Jensen MD.Splanchnic lipolysis in human obesity. J Clin Invest 2004; 113:1582–1588.

45 Minet-Ringuet J, Even PC, Valet P, Carpene C, Visentin V, Prevot Det al. Alterations of lipid metabolism and gene expression in ratadipocytes during chronic olanzapine treatment. Mol Psychiatry2007; 12: 562–571.

46 Luiken JJ, Dyck DJ, Han XX, Tandon NN, Arumugam Y, Glatz JFet al. Insulin induces the translocation of the fatty acid transporterFAT/CD36 to the plasma membrane. Am J Physiol EndocrinolMetab 2002; 282: E491–E495.

47 Hucking K, Hamilton-Wessler M, Ellmerer M, Bergman RN. Burst-like control of lipolysis by the sympathetic nervous system in vivo.J Clin Invest 2003; 111: 257–264.

48 Bymaster FP, Calligaro DO, Falcone JF, Marsh RD, Moore NA,Tye NC et al. Radioreceptor binding profile of the atypicalantipsychotic olanzapine. Neuropsychopharmacology 1996; 14:87–96.

Olanzapine-induced adiposityVL Albaugh et al

13

Molecular Psychiatry