Among psychiatric diseases, psychosis represents one of the most impacting cause of disability (Vigo et al., 2016), with a consequent increased international trend in antipsychotic prescription (Verdoux et al., 2010). Both typical neuroleptics (such as chlorpromazine and haloperidol) and second-generation anti-psychotics (including clozapine, olanzapine, risperidone, and aripiprazole) have been widely shown to be associated with weight gain (Medici et al., 2016), adipose tissue dysfunctions and body mass index (BMI) alterations (Konarzewska et al., 2014), most likely via impaired adipokine and pro-inflammatory cytokine pathways (Kucerova et al., 2015), as well as increased food intake, delayed satiety signaling, insulin resistance and glucose dysregulation (Deng, 2013; Manu et al., 2015). However, some studies have described weight elevation, adipose tissue dysfunctions, as well as other metabolic alterations, also in drug-free and drug-naïve psychotic patients (Ryan et al., 2003, 2004; Spelman et al., 2007), suggesting that antipsychotic medication might not be the main responsible of these disturbances. Indeed, increased visceral fat distribution has been described as one of the most important pathogenetic process, accounting for weight impairment and enhanced risk of metabolic syndrome development in untreated psychotic patients (Thakore et al., 2002).

On the other hand, several studies have focused on first psychotic episode subjects to investigate whether drug-free patients were more likely prone to develop a metabolic disorder than general population (Britvic et al., 2013; Misiak et al., 2016). Unfortunately, only few studies from the pre-medication era are actually available and some biases in the methodological aspects, as well as the use of different diagnostic criteria, make their interpretation arduous. In particular, some studies have included, as “drug-free,” patients medicated for a short period of time (Mane et al., 2009). Hence, as the status of “drug-free” or “drug-naïve” is difficult to be found in a psychotic patient, studies on non-pharmacologic animal models of psychosis are particularly useful. One of the most used drug-free rodent model of psychosis is the social isolation rearing of young adult rats (Weiss and Feldon, 2001; Leng et al., 2004), which provides a non-pharmacologic method of inducing long-term alterations reminiscent of several symptoms seen in psychotic patients (King et al., 2009), including hyperreactivity to novel environments and cognitive impairment (Geyer and Ellenbroek, 2003; Lapiz et al., 2003). Indeed, social isolation after weaning represents a stressful condition for rats, normally living in a gregarious environment. When reared under isolation, they show a pattern of behavioral and neurobiological alterations, which are thought to be a product of a prolonged stress (Fone and Porkess, 2008). Although the underlying pathological mechanisms are poorly understood, social deprivation of rat pups during a critical period, such as from weaning through adulthood, appears as a strong psychosocial stress, which is thought to contribute to the etiology of mental disorders, such as psychiatric conditions in humans (Schiavone et al., 2013). Together with these symptoms, accumulating evidence suggests that deprivation of social interactions is significantly associated with several metabolic disturbances, including hypertension, cardiovascular dysfunctions, and adipose tissue dysfunctions (Bartolomucci et al., 2009).

Recent data highlighted a crucial role of redox imbalance, defined as a disequilibrium between free radical production and degradation, in adipose tissue dysfunctions (Jankovic et al., 2015b). In particular, both mitochondria and other reactive oxygen species (ROS) producers, such as the NADPH oxidase NOX enzymes or the nitric oxide synthase system have been described as key redox modulators in the adipose tissue (Daiber, 2010; Koh et al., 2010). On the other hand, decrease in antioxidant defense has been shown to be a crucial contributor to the onset and progression of adipocyte dysfunctions (Curtis et al., 2010; Jankovic et al., 2016; Yadav et al., 2016). We have previously demonstrated an increased expression of the NADPH oxidase NOX2 enzyme in the central nervous system (CNS) of socially isolated rats (Schiavone et al., 2009, 2012, 2017).

Thus, the main aim of the present study was to investigate the existence of a possible pathogenetic link between oxidative stress-related biomolecular alterations and white adipose tissue dysfunctions in a “drug-free” rat model of psychosis induced by 7 weeks of social isolation. To this purpose, we compared the amount of total and visceral fat in animals reared either in social isolation (7 weeks) or in social groups, by using dual-energy X-ray (DEXA) absorptiometry. mRNA levels of specific genes encoding for ROS producer and antioxidant enzymes, as well as for oxidative stress-induced damage in the adipose tissue, were also assessed. As secondary objective, we aimed to assess if alterations of the redox state might impact the expression of Adrb3 in visceral fat, given their crucial role in the pathophysiology of adipose tissue.

Adult male and female Wistar rats (Envigo, San Pietro al Natisone, Italy) weighting 250–280 g were housed at constant room temperature (22 ± 1°C) and relative humidity (55 ± 5%) under a 12 h light/dark cycle (lights on from 7:00 AM to 7:00 PM) for at least 7 days before the experiments. Food and water were available ad libitum. Procedures involving animals and their care were conducted in conformity with the institutional guidelines of the Italian Ministry of Health (D.L. 26/2014; protocol authorization number: 171/2017 PR) the Guide for the Care and Use of Laboratory Animals: Eight Edition, the Guide for the Care and Use of Mammals in Neuroscience and Behavioral Research (National Research Council, 2004), the Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes. All procedures involving animals were conducted in accordance to ARRIVE guidelines. Animal welfare was daily monitored through the entire period of experimental procedures. No signs of distress were evidenced, anyway all efforts were made to minimize the number of animals used and their suffering.

The social isolation protocol was performed as previously described (Schiavone et al., 2016). Briefly, one male and two females were housed together for mating (Leng et al., 2004). At weaning (postnatal day 21), pups were separated from their mothers and reared either as isolated rats (ISO; one rat per cage) or reared in group as control rats (GRP; three to four rats per cage). To avoid a litter effect, each litter contributed only one male subject to the GRP and one male subject to the ISO. All animals were reared in Plexiglas cages (ISO: 40.0 cm × 27.0 cm × 20.0 cm; GRP: 59.0 cm × 38.5 cm × 20.0 cm). ISO and GRP rats were housed in the same room, so that isolated rats maintained a visual, auditory, and olfactory contact with the other animals. Rats were disturbed only for cleaning purposes (changing the cage once a week for ISO and GRP) and for the weekly body weight measurement, throughout the whole experiment. At the end of the social isolation protocol, the assessment of behavior was performed by evaluating the locomotor activity of control and isolated animals by the Open Field test and the discrimination ability by using the Novel Object Recognition test, as previously described (Schiavone et al., 2009, 2012). Animals exposed to the social isolation procedure described in the present work showed an increased locomotor activity and a reduction in the discrimination index with respect to rats reared in group (data not shown), further confirming what previously observed from our group (Schiavone et al., 2009, 2012). All the other experimental procedures were conducted immediately after the behavioral assessment.

Grouped and isolated rats were weekly weighted. Data were expressed as body weight gain, thus as the difference between initial body weight (week 1 of the social isolation period) and final body weight (week 7 of the social isolation period). Individual food intake was evaluated weekly for ISO rats, while, for GRP animals, the mean of food intake was considered. Data were expressed as feeding efficiency, indicated as g of body weight gained for g of food intake.

Researchers performing analysis were blind with respect to the rat rearing conditions, as it was not possible to deduce from the labeling whether an animal was isolated or not. The social isolation procedure was performed in a dedicated part of the animal facility, not accessible to the investigators during the entire period of the social isolation protocol. The blinding of the data was maintained until the analysis was terminated.

At the end of the social isolation period, in vivo dual energy X-ray absorptiometry analysis (DEXA) was performed as previously described (Schiavone et al., 2016) with a body scan densitometer (Hologic Dexa Bone Densitometer, Hologic Italia S.R.L., Rome, Italy). Briefly, before measurements, body calibration scan was performed with the Hologic phantom for small animals. Animals were positioned ventrally with the forelimbs away from the trunk to scan the whole body. The appropriate software program for small animals (DEXA; L & R Hip Software Ver. 11.1 for Windows) was used. After the scan, three regions of interest (ROI) were marked; right femur (R1), T9-L5 vertebrae (R2) and L1-L6 vertebrae (R3) (Figure 1). All animal images were scanned and analyzed by the same operator. Total fat mass was represented by R1+R2+R3 fat mass, while visceral fat referred to R3 fat mass, as previously described (Gerbaix et al., 2010).

Visceral fat was collected from the posterior wall of the abdominal cavity, by an area delimited in the center by the spinal structure, from one side by the abdominal wall, on the upper part by the diaphragm and in the lower part by the pelvic floor.

Immediately after collection, visceral adipose tissue was frozen in liquid nitrogen and stored at -80°C until use. The qPCR experiments were conducted as previously described (Camerino et al., 2014). Briefly, 400 ng of total RNA was added to 1 μl dNTP mix 10 mM (Roche N.C. 11277049001, Switzerland) and 1 μl RandomHexamers 50 mM (Life Technologies C.N. n808-0127, United States) and incubated at 65°C for 5 min. Afterward, 4 μl 5X First Standard Buffer (Life Technologies C.N. Y02321), 2 μl 0.1 M DTT (Life Technologies C.N. Y00147) and 1 μl Recombinant RNA Ribonuclease Inhibitor 40 U/ml (Promega, C.N. N2511, United States) were added and incubated at 42°C for 2 min. One μl of Super Script II Reverse Trascriptase 200 U/ml (Life Technologies C.N. 18064-014) was added to each solution and incubated at 25°C for 10 min, at 42°C for 50 min and at 70°C for 15 min. Real-time PCR was performed in triplicate using the Applied Biosystems Real-time PCR 7500 Fast system (United States), MicroAmp Fast Optical 96-Well Reaction Plate 0.1 ml (Life Technologies C.N. 4346906) and MicroAmp Optical Adhesive Film (Life Technologies C.N. 4311971). The setup of reactions consisted of 8 ng cDNA, 0.5 μl of TaqManGeneExpression Assays (Life Technologies), 5 μl of TaqMan Universal PCR master mix No AmpErase UNG (2X) (Life Technologies C.N. 4324018) and Nuclease-Free Water not Diethylpyrocarbonate (DEPC-Treated) (Life Technologies C.N. AM9930) for a final volume of 10 μl. The following RT-TaqMan-PCR conditions were as follows: step 1: 95°C for 20 s, step 2: 95°C for 3 s and step 3: 60°C for 30 s; steps 2 and 3 were repeated 40 times. The results were compared with a relative standard curve obtained by five points of 1:4 serial dilutions. The mRNA expression of the genes was normalized to the best housekeeping gene phosphoribosyltransferase 1 (Hprt1) selected from glyceraldehyde-3-phosphate dehydrogenase (Gapdh) beta-actin (Actinb) and Hprt1 by BestKeeper and NorFinder software. TaqMan Hydrolysis primer and probe gene expression assays were obtained by Life Technologies with the following assay IDs: Uncoupling protein 1 (mitochondrial, proton carrier) (Ucp1) IDs: Rn00562126_m1; Cell death activator (Cidea) IDs: Rn04181355_m1; Solute carrier family 2 (facilitated glucose transporter), member 4 (Slc2a4) IDs: Rn01752377_m1; Acetyl-CoA carboxylase beta (Acacb) IDs: Rn00588290_m1; NADPH oxidase (Nox1) IDs: Rn00586652_m1; NADPH oxidase (Nox4) IDs: Rn00585380_m1; heme oxygenase (decycling) 1 (Hmox1) IDs: Rn00561387_m1; adrenergic beta 3 receptor (Adrb3) IDs: Rn00565393_m1 Hprt1 IDs: Rn01527840_m1; Actb IDs: Rn00667869_m1 and; Gapdh IDs: Rn_01775763_g1. All gene expression experiments were conducted following the MIQE guideline (Bustin et al., 2009).

PRDX1, NOX1 and Adrb3 protein levels in visceral fat were determined by enzyme-linked immunosorbent assay (ELISA), using commercial ELISA kits (Cloude-Clone Corporation, Houston, TX, United States). Assays were performed according to the manufacturer’s instructions. To normalize data and negate differences due to sample collection, protein concentration was determined by using the BCA (Thermo Fisher) assay kit. Each sample analysis was carried out in duplicate to avoid intra-assay variations.

8-OHdG quantification was performed by using the Highly Sensitive 8- OHdG Check Elisa Kit; (JaiCa, Japan), as previously described (Colaianna et al., 2013).

Data were analyzed using Graph Pad^® 6.0 software for Windows. Sample size for behavioral and DEXA assessment was calculated with GPower 3.1. Data were analyzed by unpaired Student’s t-test or by two-way ANOVA for repeated measures for feeding efficiency data. Data were normally distributed, however in case of non-normal distribution, as indicated by F-test, Welch’s correction was applied. For all tests, a p < 0.05 was considered statistically significant. Results are expressed as means ± mean standard error (SEM) or as percentage of increase compared to controls.

In order to evaluate if social isolation rearing might affect adipose tissue amount, we evaluated total and visceral fat mass by DEXA in rats reared in group and in animals exposed to a 7-week period of social isolation. We detected an increase of both total fat mass (29%) (Figure 2A and related inset, n = 10 controls and n = 8 isolated animals, Student’s t-test Welch’s correction, p < 0.01) and R3 fat mass (41%) (Figure 2B and related inset, n = 10 controls and n = 8 isolated animals, Student’s t-test with Welch’s correction, p < 0.05) in isolated rats with respect to control animals.

In order to assess if social isolation could induce any significant differences in body weight gain and food intake between control and isolated animals, we weekly measured these two parameters. At the beginning of the study (week 1 of the social isolation protocol) no difference was present between grouped and isolated rats (data not shown). As regard body weight gain, we found that the increase of body weight in isolated animals was comparable to control animals (Figure 2C, n = 10 controls and n = 8 isolated animals, p = 0.009). Furthermore, we evaluated feeding efficiency between groups. A statistical difference in this parameter was evident over time (Figure 2D, two-way ANOVA F_1,18 = 1.511, p < 0.0001), as physiologically expected, while no difference was evident after between group analyses (Figure 2D, two-way ANOVA F_1,18 = 0.0272, p > 0.05).

In order to analyze the possible association between increased visceral fat amount and redox disequilibrium in isolated animals, we performed qPCR for specific genes encoding for antioxidant enzymes and ROS producers in visceral fat. We detected a decreased mRNA expression of Prdx1, known to be a crucial component of the antioxidant defense (Aeby et al., 2016) reducing hydrogen peroxide and alkyling hydroperoxides, in isolated animals with respect to rats which were reared in group (Figure 3A, n = 4 controls and 4 isolated animals, Student’s t-test, p < 0.05), which was also translated in a decreased expression of the corresponding protein level (Figure 3B, n = 5 controls and 6 isolated animals, Student’s t-test, p < 0.05). This was associated with decreased mRNA levels of Ucp-1, described as a fine regulator of the antioxidant capacity in the white adipose tissue (Jankovic et al., 2015b) (Table 1, n = 4 controls and 6 isolated animals, Student’s t-test, p < 0.05).

An enhancement in the expression of free radical producers was also observed after 7 weeks of social isolation, in terms of increased Nox1, a NADPH oxidase isoform implicated in ROS production in adipose tissue (Matsuzawa-Nagata et al., 2008), at both mRNA and protein levels (Figure 4A, n = 4 controls and 4 isolated animals, Student’s t-test, p < 0.05; Figure 4B, n = 5 controls and 6 isolated animals, Student’s t-test, p < 0.01), as well as mRNA of Hmox-1 (Figure 4C, n = 5 controls and 5 isolated animals, Student’s t-test, p < 0.05), considered one of the most sensitive and reliable indicators of cellular oxidative stress and redox regulator (Poss and Tonegawa, 1997; Ryter and Choi, 2005). The observed disequilibrium between antioxidant defense and expression of ROS producers in visceral fat was further confirmed by a significant increase, in the same tissue, of 8OHdG, one of the most abundant marker of oxidative stress to DNA (Breen and Murphy, 1995), after 7 weeks of social isolation (Figure 4D, n = 4 controls and 5 isolated animals, Student’s t-test, p < 0.05). Seven weeks of social isolation did not increase Nox4 mRNA expression in socially isolated rats compared to controls (Figure 4E, n = 5 controls and 6 isolated animals, Student’s t-test, p = 0.9101).

In order to assess if social isolation-induced increased oxidative stress in visceral fat resulted in a molecular damage of this tissue, we quantified mRNA expression of genes whose increase has been related to adipose tissue dysfunctions, i.e., Cidea, Slc2a4, and Acacb (Ma et al., 2011; Wu et al., 2014; Feldo et al., 2015; Malodobra-Mazur et al., 2016). Seven weeks of social isolation caused a significant increase in Cidea mRNA levels (Table 1, n = 4 controls and 5 isolated animals, Student’s t-test, p < 0.001), as well as Slc2a4 (Table 1, n = 4 controls and 6 isolated animals, Student’s t-test, p < 0.01) and Acacb expression (Table 1, n = 4 controls and 5 isolated animals, Student’s t-test, p < 0.05) with respect to the rearing in non-isolation conditions.

In order to determine the impact of social isolation-induced increased oxidative stress in visceral fat and adrenergic beta receptors, we performed Real-Time PCR for the subtype 3. We found a significant elevation in the mRNA levels of Adrb3 in visceral fat derived from 7 week isolated animals compared to controls (Figure 5A, n = 4 controls and 4 isolated animals, Student’s t-test, p < 0.05). This was accompanied by a significant elevation of Adrb3 protein levels (Figure 5B, n = 5 controls and 6 isolated animals, p < 0.05).