Schizophrenia is a severe mental illness. It is reported that approximately 5% of schizophrenic patients commit suicide during their lifetimes (1). Atypical antipsychotic drugs such as olanzapine (OLZ) and clozapine (CLO) have been developed for patients suffering from schizophrenia, bipolar disorder, and autism spectrum disorders to reduce suicide rates and improve longevity (2). These drugs antagonize serotonin 5-HT2C-receptor and dopamine D4-receptor, thereby reducing the symptoms of schizophrenia and causing relatively few extrapyramidal symptoms (EPS) compared to typical antipsychotic drugs (3). However, atypical antipsychotic drugs have been reported to be frequently associated with hyperphagia-induced obesity and metabolic disorders such as dyslipidemia, hyperglycemia, hyperinsulinemia, and hyperlipidemia (4, 5). More than two-thirds of the patients who receive atypical antipsychotic drugs have experienced weight gain of 1-5 kg within the first 4 weeks. The weight gain was more evident in females (6). Despite these clinical side effects, the precise mechanism of action of atypical antipsychotic drugs on the development of hyperphagia and obesity has not been well established.

The hypothalamus plays a pivotal role in the regulation of appetite and energy metabolism. The hypothalamic arcuate nucleus (ARH) is a central area that comprises anorexigenic proopiomelanocortin (POMC) neurons and orexigenic agouti-related protein (AgRP) neurons which receive signals from circulating hormones and nutrients to maintain the body weight and control energy homeostasis (7). In particular, POMC neurons project to neighboring hypothalamic areas such as the paraventricular hypothalamus (PVH) and dorsomedial hypothalamus (DMH). POMC neurons secrete α-melanocortin-stimulating hormone (αMSH) and β-endorphin (β-END) which are responsible for reduced food intake and enhanced energy expenditure (8, 9). Numerous studies have shown that these neural circuits have a crucial role in sending signals from circulating hormones and nutrients to the secondary hypothalamic areas (10, 11).

Several studies have shown that OLZ up-regulates the mRNA expression of orexigenic neuropeptides such as Npy and Agrp and down-regulates the anorexigenic Pomc mRNA expression in the hypothalamus (12-14). One study showed that OLZ administration directly alters the expression of neuropeptides regardless of feeding conditions such as ad libitum or pair-fed conditions. This suggests that the expression changes are not the secondary effects of olanzapine-induced hyperphagia (12). Despite the OLZ-induced changes in neuropeptide expression in the hypothalamus, little is known about how OLZ directly affects POMC neurons.

Metformin (MET) has been widely used since 1957 for the treatment of type 2 diabetes (15, 16); the drug reduces food intake and enhances insulin sensitivity (17, 18). For years, MET has been introduced to patients who had experienced OLZ-induced weight gain and metabolic disorders (19-21). Despite plenty of evidence that supports MET treatment for OLZ-induced metabolic disturbances and obesity, the mechanism through which MET ameliorates olanzapine-induced metabolic disturbances has not been well established. Moreover, since the weight gain itself can be a factor that affects metabolism and POMC neurons, the “pure” olanzapine effects must be elucidated. In the present study, we investigated the POMC neurons of mice that were treated with OLZ but were not yet obese and observed a decrease in numbers, projections, and leptin sensitivity. This result shows how olanzapine itself, but not obesity induces metabolic disturbances.

Short-term administration of olanzapine and metformin modulated hypothalamic neuropeptides before altering body weight and food intake in female mice

Chronic administration of olanzapine (OLZ) has been broadly reported to induce hyperphagia-induced obesity although the exact timing and pattern of obesity development differ according to the experimental animal and administration method used (12-14, 22, 23). To investigate the causative effect of OLZ-induced weight gain, we focused on the effect of short-term administration of olanzapine before the onset of obesity. Our data showed that short-term (5 days) administration of OLZ (5 mg/kg) did not induce changes in body weight and food intake (Fig. 1A-C). Since weight gain is a secondary phenotype of hyperphagia (7, 24), we first examined the expression of the appetite-associated hypothalamic neuropeptides before the weight gain. Previous reports have shown that OLZ administration for 1 week decreased the anorexigenic Pomc expression and increased orexigenic Npy and Agrp expression, which were accompanied by significant weight gain (12-14). Consistently, we found that 5 mg/kg OLZ administration for 5 days decreased Pomc expression and slightly increased Npy and Agrp expression in the hypothalamus before the onset of obesity (Fig. 1D). Administration of lower dosage OLZ (2 mg/kg) for 5 days failed to decrease hypothalamic Pomc expression (Supplementary Fig. 1). Metformin (MET) has been commonly prescribed to ameliorate OLZ-induced weight gain and metabolic disorders (25, 26). To investigate the beneficial effect of MET, MET was coadministered with OLZ in female mice. The results showed that MET recovered hypothalamic Pomc expression without causing any changes in the body weight, food intake, and Npy/Agrp expression compared to only OLZ administration in female mice (Fig. 1B-D).

MET restored OLZ-induced reduction in POMC neuronal distribution

To investigate the causes of OLZ- or OLZ + MET-induced changes in Pomc expression, we next examined the distribution of POMC neurons in the female hypothalamus. Hypothalamic POMC neurons are distributed in the arcuate nucleus (ARH) between bregma −1.22 mm and bregma −2.30 mm and are mostly located rostrally. Our data showed that the number of β-endorphin⁺ (POMC neuron marker) POMC neurons was decreased by OLZ administration and returned to normal by MET coadministration (Fig. 2). These changes were observed across the entire rostro-caudal area of the hypothalamus (Fig. 2), which suggests that the decreased Pomc mRNA expression was due to the decrease in the number of POMC neurons (Fig. 1D and Fig. 2). We also analyzed the soma sizes of POMC neurons; however, there were no changes in the neuronal size among the three groups (Fig. 2).

MET restored OLZ-induced reduction in POMC axonal projection

POMC neurons innervate the paraventricular hypothalamus (PVH) and dorsomedial hypothalamus (DMH), which are indispensable for the regulation of food intake and energy expenditure (27). Our data showed that POMC axonal projections to PVH and DMH were decreased by OLZ administration, and this alteration was restored by MET coadministration (Fig. 3A, B). In addition, we normalized the projection intensity by taking the number of ARH POMC neurons into account since the changes in projection intensity of β-endorphin⁺ POMC neurons could be due to the decrease in the POMC neuron numbers. The calculation results showed that OLZ demonstrated a tendency to decrease POMC projection but MET coadministration significantly recovered the OLZ-induced decreased POMC projection (Fig. 3C, D). Of note, MET administration alone failed to significant changes in the distribution or number of POMC neurons (Supplementary Fig. 2).

MET ameliorated OLZ-induced leptin resistance in POMC neurons

Leptin is produced by adipose tissue and is essential for the homeostasis of food intake and systemic metabolism (28). Leptin activates POMC neurons through the leptin receptor to suppress appetite and stimulate energy expenditure (28, 29). We next assessed whether OLZ-induced POMC neuronal changes are associated with leptin resistance. Consequently, we injected 1 μg leptin into the hypothalamic 3V and observed POMC activation following a 5-days administration of OLZ or OLZ + MET in 7-week-old female mice (Fig. 4A, B). The result showed that OLZ significantly impeded leptin-induced POMC activation, which was recovered by MET coadministration (Fig. 4C). Taken together, our data demonstrate that OLZ decreases Pomc mRNA expression, POMC neuronal number, POMC projection, and leptin sensitivity of POMC neurons. Moreover, pharmacological administration of MET led to the recovery of OLZ-induced abnormalities in POMC neurons.

Our study demonstrates the effect of OLZ and OLZ + MET administration on the hypothalamic POMC neurons and leptin sensitivity. Since OLZ is widely known to cause hyperphagia-induced weight gain in females, we measured the hypothalamic Pomc mRNA expression, POMC neuron number, POMC innervation, and leptin sensitivity in POMC neurons in the female mouse. The measurements were done before the onset of obesity so that it was possible to discriminate between the changes caused by OLZ administration alone and the changes caused by obesity. Our results suggested that OLZ directly causes hypothalamic POMC neuronal changes, which might lead to obesity in long term.

Leptin acts on its receptors on POMC neurons, which leads to the activation of POMC neurons (7, 30-32). The activation of hypothalamic POMC neurons results in increased energy expenditure and reduced food intake, which is essential in metabolic homeostasis (28). c-Fos is a well-known indicator of neuronal activity. Also, it is known from a previous study that leptin administration increases c-Fos expression in the hypothalamic POMC neurons (33). Of note, leptin-resistant (db/db) genetically obese mice failed to demonstrate leptin-induced c-Fos expression in POMC neurons (34). Another study showed that endocannabinoid/CB1 receptor (CB1R) antagonism significantly recovered c-Fos immunoreactivity blunted by HFD in ARH neurons (35). In the present study, we injected leptin into the 3V of OLZ- or OLZ + MET-treated mice. OLZ-injected mice showed fewer c-Fos⁺ POMC neurons than the OLZ + MET group. This implies that the POMC neurons of OLZ-treated mice were less sensitive to leptin than that of the OLZ + MET group. Hence, it would be legit to claim that MET restores OLZ-induced leptin resistance.

OLZ-induced leptin resistance in POMC neurons could be explained based on hypothalamic inflammation, which is a well-known cause of leptin resistance (36, 37). Accumulating data have shown that OLZ induces inflammation in the hypothalamus as well as peripheral tissues (37-39). A recent study showed that OLZ stimulates astrocytes via toll-like receptor-4 signaling (39). In this study, dose-dependent OLZ treatment induced hypothalamic endoplasmic reticulum (ER) stress in the cultured human astrocytes, which led to uncontrolled hyperphagia. Plenty of evidence has suggested that inflammatory cytokines such as IL-1β, IL-6, and TNFα trigger leptin resistance by increasing the expression of the suppressor of cytokine signaling 3 (SOCS3) and protein tyrosine phosphatase 1B (PTP1B). They are known to be negative regulators of hypothalamic leptin signaling as they interrupt leptin receptor (ObRb)-Janus kinase (Jak)2 signaling (40, 41). Although there is still a possibility that OLZ directly induces gene expression related to leptin resistance, it is conceivable that OLZ causes hypothalamic inflammation before the onset of obesity, which in turn leads to hypothalamic leptin resistance.

MET is used as an AMPK activator and AMPK is known as an anti-inflammatory molecule (42). AICAR, a pharmacological AMPK activator, inhibits lipopolysaccharide (LPS)-induced expression of inflammatory cytokines such as IL-1β, IL-6, and TNFα in astrocyte, microglia, and peritoneal macrophage primary cells (42). In the experimental autoimmune encephalomyelitis (EAE) model, AMPK activation demonstrated anti-inflammatory and immunomodulatory effects by preventing the infiltration of inflammatory cells across the blood-brain barrier (BBB) (43). Based on this evidence, it is hypothesized that MET improves OLZ-induced leptin resistance in hypothalamic POMC neurons partly through its anti-inflammatory action.

In this study, we demonstrated that OLZ induces a decrease in the number of POMC neurons, POMC projections, and leptin sensitivity in POMC neurons, which were reversed by MET coadministration. Further studies are needed to investigate the role of hypothalamic inflammation in OLZ-induced leptin resistance.

Animals

All animal procedures were approved by the Institutional Animal Care and Use Committee of the Hallym University (Chuncheon, Korea). Seven-week-old C57BL/6 female mice were purchased from DBL (Chungbuk, Korea). Animals were group-housed in a temperature-controlled room (22 ± 1°C) with a 12 h light-dark cycle (lights on at 8 AM). Mice had free access to a standard chow diet and water ad libitum unless otherwise indicated.

Drug administration

Olanzapine (Sigma, #O1141) was first dissolved in dimethyl sulfoxide (DMSO) and subsequently prepared in normal saline (0.9% NaCl) to reach the final concentration. Metformin (Sigma, #PHR1084) was dissolved in normal saline. Olanzapine (5 mg/kg/d) and metformin (300 mg/kg) were administered daily by oral gavage (between 8:00 AM and 9:00 AM) for the indicated period. Control mice were administered with a corresponding vehicle at the body weight balanced volume.

Immunofluorescence staining

Mice were deeply anesthetized with isoflurane inhalation and transcardially perfused with a 50 ml ice-cold normal saline followed by 50 ml 4% paraformaldehyde (PFA) via the left ventricle of the heart. Whole brains were collected and post-fixed with 4% PFA for 16 h at 4°C and dehydrated in PBS-based 30% sucrose solution for 48 h. Coronal brains including the hypothalamic area were sectioned 30 μm thick using a cryostat (Leica, Wetzlar, Germany). Brain slices were stored at −70°C in a deep freezer. For POMC neuron staining, hypothalamic slices were permeabilized in 0.5% PBST (Triton X-100 in PBS) for 5 min and blocked with 3% bovine serum albumin (BSA) in 0.5% PBST solution at room temperature (RT) for 1 h. The slices were incubated with anti-mouse β-endorphin antibody (1:1,000, Phoenix Pharmaceuticals, #H-022-33) in a blocking solution at 4°C for 48 h and then at RT for 1 h. For c-Fos and β-endorphin double staining, hypothalamic slices were blocked with 3% BSA in 0.5% PBST solution at RT for 1 h and incubated with anti-c-Fos antibody (1:1000, Synaptic System, #226 003) in blocking solution at 4°C for 48 h. Next, the slices were stained following β-endorphin staining procedure. After washing, hypothalamic slices were incubated with the appropriate Alexa-Flour 488-, 546-, or 555-conjugated secondary antibodies (1: 1000, Invitrogen) at RT for 1 h. Fluorescence was taken using confocal microscopy (Carl Zeiss 710, Germany).

Hypothalamic gene expression

Mice from Control, OLZ, and OLZ+MET groups were kept under freely-fed conditions and sacrificed by decapitation at indicated timepoints. Mediobasal hypothalamic tissue blocks were obtained, quickly frozen in liquid nitrogen, and stored at −70°C in a deep freezer. Total RNAs were extracted using TRIzol (Life Technologies, #15596018) according to the manufacturer’s protocol. RNAs were reverse transcribed to generate cDNA. The expression levels were determined using real-time PCR analysis using the Pomc, Agrp, Npy, and Gapdh primers. The mRNA expression levels were normalized to that of Gapdh.

Intracerebroventricular injection

A stainless-steel cannula (26 gauge) was implanted into the 3rd ventricle (3V) of C57BL/6 mice (Stereotaxic coordinates: 1.5 mm caudal to bregma and 5.5 mm ventral to the sagittal sinus). Following a 7-day recovery period, the correct positioning of each cannula was confirmed by observing a vigorous water intake following administration of 50 ng angiotensin-2. Animals with a negative drinking response to angiotensin-2 were excluded from this study.

POMC neuron leptin sensitivity test

Leptin (1 μg, R&D Systems, #498-OB) was dissolved in 2 μl of normal saline before ICV administration. Leptin was injected 24 h after the 5th day of Veh, OLZ, OLZ + MET administration. Mice were cardiac perfused at 45 min after leptin injection for immunofluorescence staining.

Statistics

All data values are presented as mean ± standard error of the mean (SEM). Statistical analyses were performed using Prism version 9.3.0 (GraphPad). Statistical significance among the groups was tested using one-way or repeated-measures analysis of variance (ANOVA) followed by a post hoc least significant difference test. Statistical significance was defined by a *P < 0.05, **P < 0.01, and ***P < 0.001.

This research was supported by Hallym University Research Fund, 202110-005 (HRF-202110-005).

The authors have no conflicting interests.