Effects of the second-generation antipsychotic drugs aripiprazole and olanzapine on human adipocyte differentiation

Second-generation antipsychotics (SGAs), used as the cornerstone treatment for schizophrenia and other mental disorders, can cause adverse metabolic effects (e.g. obesity and type 2 diabetes). We investigated the effects of SGAs on adipocyte differentiation and metabolism. The presence of therapeutic concentrations of aripiprazole (ARI) or its active metabolite dehydroaripiprazole (DARI) during human adipocyte differentiation impaired adipocyte glucose uptake while the expression of gene markers of fatty acid oxidation were increased. Addi- tionally, the use of a supra-therapeutic concentration of ARI inhibited adipocyte differentiation. Furthermore, olanzapine (OLA), a highly obesogenic SGA, directly increased leptin gene expression but did not affect adipo- cyte differentiation and metabolism. These molecular insights are novel, and suggest that ARI, but not OLA, may directly act via alterations in adipocyte differentiation and potentially by causing a switch from glucose to lipid utilization in human adipocytes. Additionally, SGAs may effect crosstalk with other organs, such as the brain, to exert their adverse metabolic effects.


Background
Schizophrenia is a severe and chronic psychiatric disorder characterized by positive (e.g. delusions and hallucinations), negative (e.g. social withdrawal), and cognitive (e.g. cognitive inflexibility and impaired problem solving) symptoms (McCutcheon et al., 2020). Atypical or second-generation antipsychotics (SGAs) are medications commonly prescribed to treat schizophrenia and other psychiatric disorders (e.g. bipolar disorder), being highly efficient in suppressing both positive and negative symptoms. However, they are associated with weight gain and, as a consequence, metabolic alterations, such as insulin resistance, dyslipidemia, cardiovascular disease, and type 2 diabetes (T2D) (Nasrallah, 2008a, Rojo et al., 2015. Different SGAs have different propensities to induce adverse metabolic effects, e.g. olanzapine (OLA) is a high-risk drug associated with worse metabolic outcomes, while aripiprazole (ARI) is considered more metabolically neutral (Hirsch et al., 2017).
Adipose tissue is an important endocrine organ involved in the regulation of whole-body energy balance. Adipose tissue expansion occurs due to an increase in adipocyte number (hyperplasia) and/or an increase in cell size (hypertrophy) (Fuster et al., 2016). New adipocytes are generated by recruitment and differentiation from precursor cells (preadipocytes) located in the adipose tissue, in a process so-called adipocyte differentiation or adipogenesis (Tang and Lane 2012). An increase in cell number by the differentiation of new adipocytes, as opposed to increased cell size, is considered protective against obesity-associated metabolic complications (Gustafson et al., 2015). It has been shown that the number of preadipocytes in subcutaneous adipose tissue (SAT) that can undergo differentiation is reduced in obesity, resulting in the accumulation of triglycerides in the existing adipocytes, leading to adipocyte hypertrophy, which is associated with hypertension, dyslipidemia, ectopic lipid accumulation, and T2D (Bays et al., 2008). Previous studies have shown that supra-therapeutic concentrations of SGAs, such as OLA, facilitate lipid storage and stimulate adipogenesis in mouse and human adipose stem cell models (Sertié et al., 2011;Nimura et al., 2015;Bába et al., 2019), an effect that could play a role in SGA-induced weight gain. However, the knowledge of the impact of therapeutic concentrations of SGAs in human primary preadipocytes differentiation and function is limited.
Our recent study showed that ARI and OLA at therapeutic concentrations have a mild direct effect on lipid and glucose metabolism of mature adipocytes after short-term adipose tissue exposure (up to 72h) (Sarsenbayeva et al., 2019;Sarsenbayeva et al., 2021). This data suggests that SGAs might affect adipocytes before they are mature (during differentiation), longer/chronic treatments might be needed, or the adverse effects could be mediated through other insulin-sensitive tissues and/or the central nervous system.
In this study, we aimed to investigate the direct effects of ARI and OLA on human primary adipocyte differentiation and glucose metabolism, and the expression of genes regulating adipocyte function, after a long-term exposure during differentiation.

Subjects
SAT samples were obtained from 6 healthy subjects (4 women/2 men, BMI: 26.3-30.3 kg/m 2 , age: 20-63 years) by needle aspiration of the lower part of the abdomen after local dermal anesthesia with lidocaine (Xylocaine, AstraZeneca, Södertälje, Sweden), and used for the isolation of preadipocytes. Subjects' clinical characteristics are shown in Table 1. Blood samples were collected after overnight fasting (>10 h) for biochemical analysis at the Department of Clinical Chemistry (Uppsala University Hospital). We excluded subjects with type 1 diabetes, endocrine disorders, cancer, or other major diseases, as well as subjects having ongoing medication with antipsychotics, antidepressants, or neuroleptic drugs. The study was approved by the Regional Ethics Review Board in Uppsala (Dnr, 2018/385), and the reported investigations have been carried out following principles endorsed by the Declaration of Helsinki. All participants gave their written informed consent.

Fluorescent lipid staining for differentiation rate
Fluorescent staining of neutral lipids was performed on days 10-14 of differentiation (n = 5) as previously described . Cells were fixed with 4% formaldehyde (Histolab, Gothenburg, Sweden), and a combination of two fluorescent dyes were used: BODIPY 493/503 for neutral lipids (green; Molecular Probes, OR, USA) and Hoechst 33342 for nuclei (blue; Invitrogen, MA, USA). Cells were imaged using the ImageXpress Pico (Molecular Devices, CA, USA) and the InCellis Fluorescent microscope (Bertin Instruments, Montigny-le-Bretonneux, France). Image acquisition on the ImageXpress Pico was performed in a 6x6 square image scan, using a 10× magnification, covering ~20% of the central area of each well. Hoechst 33342 staining was imaged using the DAPI channel and BODIPY staining using the FITC channel, and image analyses were performed on the images obtained from the ImageXpress Pico instrument using the CellRe-porterXpress software. The percentage of lipid-positive cells was calculated using the automated cell scoring software function "Cell differentiation" which identifies cells that either contain lipids (positive cells) or that do not contain any lipids (negative cells). Lipid-positive area was measured using the software function "Cell count", with marker set for the FITC channel and calculated as the total stain area of BODIPY per total number of cells.

Gene expression
Total RNA was extracted from differentiating preadipocytes (days 0, 7, and 14 post-induction) (n = 3-6) using the RNeasy lipid tissue mini kit (Qiagen, Hilden, Germany), as previoulsy reported . cDNA was synthesized using a High Capacity cDNA Reverse Transcriptase kit (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's guidelines. The concentration and purity of total RNA were measured with the Nanodrop (Thermo Scientific). mRNA expression was determined using TaqMan gene expression assays (Thermo Fisher) shown in Table 2. Gene expression was detected using the QuantStudio 3 sequence detection system (Applied Biosystems) and calculated using a 2 − deltaCt . The gene expression levels were normalized to the housekeeping gene GUSB for human samples and Gapdh for mice samples, and all samples were run in duplicates.

Glucose uptake in differentiated adipocytes
Glucose uptake was performed on days 13-14 of the differentiation (n = 3-5) in control and SGA-incubated cells using luminescence Glucose Uptake-GLO™ kit (Promega, Madison, WI, USA) according to the manufacturer's instructions and as previously reported . Briefly, cells were washed twice with warmed PBS and then incubated for 2 h in Krebs Ringer HEPES (KRH) buffer containing 0.01% BSA (Sigma), with 5 mM glucose (Sigma), 200 nM adenosine (Sigma), and pH 7.4 followed by washing twice with KRH without glucose. Cells were then incubated at 37 • C for 30 min in KRH medium without glucose, and without or with physiological and supra-physiological concentrations of insulin (25 and 1000 μU/mL, respectively). 1 mM (final concentration) of 2-deoxyglucose was added during the last 10 min of the incubation. The glucose uptake quantification was done as per the manufacturer's instructions and corrected for total protein concentrations.

Cell viability assay
Cell viability was measured on day 14 of differentiation using CyQUANT™ LDH Cytotoxicity Assay Kit (C20300; Invitrogen, Waltham, MA) per manufacturer's instructions (n = 3). Briefly, 10X Lysis Buffer was added onto the plates with adipocytes differentiated without or with SGAs, and the plates were incubated for 45 min at 37 • C. After incubation, 1X Positive Control and Reaction Mixture were added, and plates were incubated at room temperature for 30 min protected from light. Thereafter, Stop Solution was added, and absorbance was measured at 490 and 680 nm on a plate reader (SpectraMax® iD3, Molecular Devices). Cell viability was calculated as relative to control.

Animals
Animal studies developed at the Instituto de Investigaciones Biomedicas Alberto Sols (Madrid, Spain) were approved by the Ethics Committee of the Consejo Superior de Investigaciones Científicas (CSIC, Spain) and conducted per the guidelines for animal care of Comunidad de Madrid (Spain) and the Directive 2010/63/EU of the European Parliament and the Council.
To test the effects of OLA and ARI on adipose tissue gene expression in vivo, 12-week old male mice were used, and the study was performed as previously described (Ferreira, Folgueira et al. 2022). Briefly, the animals were kept in a controlled environment (temperature: 22-24 • C, humidity: 55%), with 12 h of light-dark cycles and fed with regular rodents' chow diet (A04, Panlab, Barcelona, Spain) (n = 9) or the same diet supplemented with OLA (n = 9) or ARI (n = 8) for 7 months. The calculations were made so that, with a regular food intake of 4 g/day, the mice (with a weight between 30 and 35 g) received 5 mg/kg/day of OLA or ARI. After the treatment, during the light phase of the diurnal cycle, mice were euthanized by decapitation. White adipose tissue, both visceral-epididymal and subcutaneous-inguinal, was dissected, snap-frozen, and stored at − 80 • C until used for gene expression analyses.

Statistical analysis
Data are shown as mean ± SEM unless stated otherwise. All data were first checked for normality using the Shapiro-Wilk test and analyzing histograms. A comparison of the mean of more than two groups was made using repeated measures ANOVA and corrected for the false discovery rate using the original Benjamini and Hochberg method. All statistical analyses were performed using GraphPad Prism 9 software, and p < 0.05 was considered statistically significant.

SGAs effects on human adipocyte differentiation
Human primary preadipocytes were differentiated in the absence or presence of ARI, DARI, and OLA for 14 days. The adipocyte differentiation rate was determined by measuring the number of differentiating adipocytes, adipocyte lipid accumulation, and the expression of key adipogenic mRNA markers. We have previously shown that these measures were inhibited after the knockout of the master regulator of adipogenesis, PPARG, in human preadipocytes (Kamble et al., 2020).
When present during differentiation, therapeutic levels of ARI and supra-therapeutic concentration of DARI caused a minor but significant reduction in the number of differentiating preadipocytes (up to 4%; p < 0.05), and in the amount of lipids accumulated per adipocyte (up to 10%; p < 0.05) (Fig. 1A-C). Similarly, the supra-therapeutic concentration of ARI caused only a minor decrease in the number of differentiating preadipocytes (~6%; p < 0.001); however, it markedly decreased the amount of lipids accumulated per adipocyte (~70%; p < 0.01), compared to the control (Fig. 1A-C).
Regarding the effects of SGAs on the expression of adipogenic gene markers, overall, we observed that ARI on days 7 and 14, and DARI on day 7, at supra-therapeutic concentrations, reduced the expression of all measured differentiation markers (PPARG, CEBPA, CD36, and ADIPOQ) (Fig. 1D-G). In addition, therapeutic concentrations of ARI reduced CEBPA and ADIPOQ expression, and DARI reduced CEBPA expression on day 7, but these effects were lost by day 14 of adipogenesis (p < 0.05) ( Fig. 1E and G). Interestingly, therapeutic concentrations of ARI and supra-therapeutic concentration of DARI increased PPARG expression on day 14 of adipogenesis (p < 0.05) (Fig. 1D). OLA did not affect differentiation rate at therapeutic (data not shown) or supra-therapeutic concentrations (Fig. 1A-D), except for an almost 2-fold up-regulation of LEP expression at the end of differentiation compared to the control (p < 0.01) (Fig. 1H). Furthermore, leptin mRNA levels were significantly increased by treatment with supra-and therapeutic concentrations of ARI and DARI on day 14 of differentiation (p < 0.05) (Fig. 1H). Adipocyte cell viability was not affected after differentiation in the presence of either SGA (Fig. 1I).

Analysis of the SGA effects on the expression of lipid storage markers
Lipid storage gene markers FASN, DGAT2, and LPL expression were significantly reduced during and after differentiation (days 7 and 14) in Comparisons were made between treatments at either time point and not between time points. Data are shown as mean ± SEM, n = 4-6. FASN data has been log-transformed. ARI, aripiprazole, OLA, olanzapine; DARI, dehydroaripiprazole. *p < 0.05, **p < 0.01, ***p < 0.001. the presence of a supra-therapeutic concentration of ARI (p < 0.05) ( Fig. 2A-C). LPL mRNA expression was also reduced with therapeutic concentrations of ARI and supra-therapeutic concentration of DARI on day 7 of differentiation (p < 0.05), but these effects were not observed on day 14 (Fig. 2C). The expression of the lipolytic genes HSL and ATGL was significantly reduced by the supra-therapeutic concentration of ARI (days 7 and 14) (p < 0.05) (Fig. 2D and E). In addition, ATGL gene expression on day 7 was dose-dependently reduced by ARI treatment (Fig. 2E), and DARI reduced mRNA expression of HSL on day 14 and ATGL on day 7 of differentiation (p < 0.05) (Fig. 2D and E).

Analysis of glucose uptake in differentiated adipocytes after SGAs treatment
Glucose uptake was measured in adipocytes differentiated in the presence or absence of SGAs. ARI and DARI dose-dependently decreased basal and insulin-stimulated glucose uptake up to ~70% (Fig. 3A). However, the relative response to insulin was similar between SGAs and control conditions, except for a significant reduction with a supratherapeutic concentration of ARI (p < 0.001 for maximal insulin stimulation). GLUT1 mRNA expression was dose-dependently decreased up to 80% by ARI and DARI on day 14 of differentiation (p < 0.05) (Fig. 3B). The expression of GLUT4 transporter was dose-dependently decreased up to 90% by both ARI and DARI on day 7 and by supratherapeutic concentrations of ARI and DARI on day 14 of differentiation (p < 0.05) (Fig. 3C). In addition, the protein levels of GLUT4 followed the gene expression pattern and were decreased up to 60% by the supra-therapeutic concentration of ARI and both supra-and therapeutic concentrations of DARI on day 14 of differentiation, when compared to the control condition (p < 0.05) (Fig. 3D and E). OLA had no impact on adipocyte glucose uptake or the expression of the glucose transporters at therapeutic (data not shown) or supra-therapeutic concentrations (Fig. 3A-E).

Evaluation of the effects of SGAs on gene expression of mitochondrial function and lipid oxidation markers
The expression of mitochondrial biogenesis marker TFAM and the lipid oxidation markers PDK4, CPT1B, and ACO1 was decreased by the supra-therapeutic concentration of ARI on days 7 and 14 compared to control on the respective days (p < 0.05, Fig. 4A-D). In addition, the supra-therapeutic concentration of DARI reduced the expression of TFAM and PDK4, but only on day 7 (p < 0.05, Fig. 4A-B). In contrast, therapeutic concentrations of ARI and supra-and therapeutic concentrations of DARI increased the expression of TFAM (p < 0.05, Fig. 4A) and lipid oxidation markers PDK4, CPT1B, and ACO1 (p < 0.06, Fig. 4B-D) on day 14 of differentiation.

Gene expression in adipose tissue from ARI-and OLA-treated mice
The in vitro outcomes of the SGAs-mediated effects on the expression of differentiation, lipid storage, and mitochondrial function markers in human preadipocytes were validated in inguinal and epididymal adipose tissue depots from mice treated with ARI and OLA for 7 months.
Treatment with OLA increased the weight of the animals, compared to baseline (Ferreira, Folgueira et al. 2022), but not with ARI. However, there was an increase in epididymal adipose tissue weight after OLA and ARI treatment (by 89% and 132%, respectively; p < 0.05), and inguinal adipose tissue weight by ARI (by 140%; p < 0.05). In epididymal adipose tissue, ARI and OLA treatments decreased the expression of Pparg (p < 0.001) (Fig. 5A). Similar to the in vitro results in human preadipocytes, OLA and ARI increased the expression of Lep in both inguinal (OLA p = 0.07 and ARI p < 0.05) and epididymal adipose tissue (OLA p = 0.05 and ARI p < 0.01) (Fig. 5B). The lipid storage marker Fasn was decreased in both inguinal and epididymal adipose tissues (p < 0.05) (Fig. 5C). Furthermore, the lipid oxidation marker Pdk4 was increased by both ARI and OLA treatments in epididymal adipose tissue (p < 0.05) (Fig. 5D).

Discussion
This study investigated the effects of ARI, its active metabolite DARI, and OLA on human adipocyte differentiation. The results show that while ARI and DARI can directly impair adipocyte differentiation and glucose uptake, OLA only acts to increase leptin mRNA expression in differentiated adipocytes.
Therapeutic concentrations of ARI and DARI had only minor or no effects on the differentiation of human preadipocytes and expression of adipogenic and lipogenic gene markers. On the contrary, the supra- therapeutic concentration of ARI markedly decreased lipid accumulation by ~70%, which was in accordance with the reduction in the expression of genes regulating adipocyte differentiation (e.g. PPARG, CD36, CEBPA) and lipid storage (e.g. FASN, DGAT2) and adipocyte differentiation. However, the translational relevance of this finding is debatable since the supra-therapeutic concentration used is ~10 times higher than commonly found in the plasma of patients taking ARI, which varies roughly from 0.5 to 2 μM with doses of 7.5-60 mg/day (Kirschbaum et al., 2008). OLA treatment had no effects on adipocyte differentiation and adipocyte lipid and glucose metabolism in vitro, suggesting that commonly reported OLA-induced weight gain (Citrome et al., 2011;Huang et al., 2020) is not due to direct effects on adipocyte development. Conversely, OLA has previously been described to enhance adipocyte differentiation and increase triglyceride accumulation during in vitro differentiation in human and rodent precursor cells (Bába et al., 2019;Nimura et al., 2015;Sertié et al., 2011;Yang et al., 2007). However, it is important to highlight that previous studies used different cell models, incubation periods, and much higher concentrations of OLA (up to 100 μM) than found in the serum of the patients under OLA treatment.
The divergent effects could be explained by the different cell models and/or supra-therapeutic concentrations of SGAs used in previous in vitro studies (Grajales et al., 2019;Ferreira, Grajales et al. 2020). In accordance with the findings in human preadipocytes, OLA and ARI treatments in mice had no effect on the expression of differentiation markers in inguinal adipose tissue. Nevertheless, in epididymal adipose tissue ARI and OLA treatment reduced Pparg expression. These results also highlight the importance of future studies to explore SGAs-induced effects in cells from human visceral adipose tissue (VAT) to elucidate potential depot-specific effects since the results presented in this work are exclusively in preadipocytes isolated from human SAT.
The treatment with ARI and its metabolite dose-dependently reduced (up to 70%) the basal and insulin-stimulated glucose uptake in differentiated adipocytes. Insulin-stimulated glucose uptake reduction caused by the supra-therapeutic concentration of ARI could be attributed to the lack of adipocyte differentiation in those cultures, as insulin responsiveness and GLUT4 expression are characteristics of mature adipocytes (Hauner et al., 1998;Fernyhough et al., 2007). However, the therapeutic concentrations of ARI, as well as DARI, inhibited adipocyte glucose uptake without major effects on adipocyte differentiation, so this effect can be considered differentiation-independent. Moreover, except for the highest ARI concentration, cultures treated with ARI and DARI did not show an altered insulin response, suggesting that the inhibitory effects are independent of insulin signaling. Accordingly, decreased glucose uptake was most likely due to the decreased expression of glucose transporters GLUT1 and GLUT4. Reduction in adipocyte glucose uptake after ARI and DARI treatment may contribute to impaired glucose handling in adipose tissue and hyperglycemia, as observed in patients with T2D (Pereira et al., 2016;Boersma et al., 2018). Even though ARI is considered metabolically safer, compared to OLA, previous studies have reported aripiprazole-induced hyperglycemia, and even cases of diabetic ketoacidosis in humans (Dhamija andVerma 2008, Makhzoumi et al., 2008).
We also observed an increase in the expression of genes important for mitochondrial biogenesis and lipid oxidation (e.g. TFAM, PDK4, CPT1B) by ARI and DARI, which could indicate increased lipid oxidation rates in adipocytes (Vernochet et al., 2012;Warfel et al., 2016;Pettersen et al., 2019;De Carvalho et al., 2021). However, bearing that measuring gene expression of mitochondrial function markers is an indirect measure of mitochondrial activity, further functional studies are required. Enhanced lipid oxidation in combination with the decreased glucose uptake lead us to hypothesize that there could be a potential metabolic switch from glucose to lipid utilization in adipocytes. In animal studies, it has previously been shown that SGAs such as OLA and clozapine can cause a rapid switch from glucose utilization to whole-body lipid oxidation in mice, while ARI has little or no effects (Albaugh et al., 2012;Klingerman et al., 2013). Decreased glucose uptake and increased lipid oxidation in adipocytes, as observed in our results, could be protective against weight gain (Vijayakumar et al., 2012) and potentially explain Fig. 5. Gene expression in inguinal and epididymal adipose tissue depots from SGAs-treated mice for 7 months. mRNA expression of (A-C) adipocyte differentiation and function markers, (D) lipid oxidation gene marker in inguinal and epididymal adipose tissue of control, and ARI-and OLA-treated mice. Data are shown as mean ± SEM, n = 9 for controls, n = 9 for OLA, and n = 8 for ARI. ARI, aripiprazole, OLA, olanzapine. *p < 0.05, **p < 0.01, ***p < 0.001. why patients taking ARI have fewer reported weight-related side effects (Hirsch et al., 2017). However, further in vitro and in vivo functional assessments of potential energy substrate switch following ARI treatment are needed.
Leptin is an energy balance regulating adipokine secreted mainly by white adipose tissue, hence circulating leptin levels are strongly correlated to body fat percentage (Considine et al., 1996a). It is well-known that leptin blood levels are increased in obesity (Considine et al., 1996a), reflecting leptin resistance that may account for disturbances in body weight regulation (Sáinz et al., 2015). ARI-and OLA-treated mice in vivo showed increased levels of Lep expression in both epididymal and inguinal adipose tissue depots. An increase in leptin upon treatment with SGAs has also previously been shown in patients taking the drugs (Pérez-Iglesias, Ortiz-Garcia de la Foz et al., 2014, Potvin et al., 2015. Leptin increase in vivo could reflect the increase in adiposity after the SGAs treatment, as also previously suggested in patients during antipsychotic treatment (Haupt et al., 2005). However, our in vitro data argue that ARI, DARI, and OLA treatment can directly up-regulate LEP expression in human adipocytes. Leptin is considered a major CNS satiety signal, but it also plays an important role in the periphery. Leptin can acutely inhibit the translocation of GLUT4 in murine muscle L6 cells (Sweeney et al., 2001) and regulate lipolysis in adipocytes (Zeng et al., 2015). However, whether increased leptin mRNA expression is associated with increased leptin protein levels and impairment in glucose and lipid metabolism in adipocytes requires further investigation.
ARI and OLA target a spectrum of different neurotransmitter receptors e.g. dopaminergic, serotonergic, histaminergic etc. (Roth et al., 2004). However, they differ in drug pharmacology, from the receptors they target with varying relative affinities, to their agonism/antagonism properties, e.g. OLA antagonizes dopamine receptor D 2 , serotonin receptor 5-HT 2C , and muscarinic M3 receptor, while ARI acts as a partial agonist of D 2 and 5-HT 1 (Nasrallah 2008, Siafis et al., 2018. Specific receptor-binding profiles of SGAs have previously been associated with the specific metabolic side effects; for example, the antagonism of 5-HT 2C , D 2 , and H 1 has been linked to the antipsychotic induced weight gain and diabetes, and partial agonism of D 2/3 or 5-HT 1 may disrupt insulin secretion by beta cells (Allison et al., 1999;Matsui-Sakata et al., 2005;Correll et al., 2011, Guenette et al., 2013Ustione et al., 2013;Bennet et al., 2015). Varying propensities to cause metabolic adverse-effects, as well as different adverse-effect profiles, may be due to the differences in receptor-binding profiles of the SGAs (Nasrallah, 2008a). Future studies are warranted to investigate the link between receptor-binding profiles and the divergent metabolic side-effects induced by ARI and OLA.
This study has some limitations. The differentiation in the presence of SGAs was performed in a limited number of experiments and only in preadipocytes from SAT, thus, future studies should also focus on the effects on VAT preadipocytes. Additionally, the in vitro study in human adipocytes doesn't necessarily reflect the in vivo setting in humans. Furthermore, the therapeutic and supra-therapeutic concentrations of the drugs used to treat human adipocytes were based on the measures in the plasma of the patients receiving treatment (Kirschbaum et al., 2008;Citrome et al., 2009). However, since SGAs are lipophilic drugs, the concentration in adipose tissue could be higher than the concentration in plasma. Additionally, this study our lacks functional assessment of lipid oxidation.
These data suggest that ARI, but not OLA, may directly act via inhibition of subcutaneous adipocyte differentiation and glucose uptake, and increase the expression of lipid oxidation genes. These molecular insights suggest that ARI may cause a switch from glucose to lipid utilization in adipocytes, while OLA may require longer chronic treatments and/or the crosstalk with other organs, such as the brain, to exert their side-metabolic effects.

Declaration of competing interest
All authors have read the journal's authorship agreement and policy on disclosure of potential conflicts of interest. The authors declare no conflict of interest.

Data availability
Data will be made available on request.