Peroxisome Proliferator-Activated Receptors-Alpha and Gamma Are Targets to Treat Offspring from Maternal Diet-Induced Obesity in Mice

Aim The aim of the present study was to evaluate whether activation of peroxisome proliferator-activated receptor (PPAR)alpha and PPARgamma by Bezafibrate (BZ) could attenuate hepatic and white adipose tissue (WAT) abnormalities in male offspring from diet-induced obese dams. Materials and Methods C57BL/6 female mice were fed a standard chow (SC; 10% lipids) diet or a high-fat (HF; 49% lipids) diet for 8 weeks before mating and during gestation and lactation periods. Male offspring received SC diet at weaning and were subdivided into four groups: SC, SC/BZ, HF and HF/BZ. Treatment with BZ (100 mg/Kg diet) started at 12 weeks of age and was maintained for three weeks. Results The HF diet resulted in an overweight phenotype and an increase in oral glucose intolerance and fasting glucose of dams. The HF offspring showed increased body mass, higher levels of plasmatic and hepatic triglycerides, higher levels of pro-inflammatory and lower levels of anti-inflammatory adipokines, impairment of glucose metabolism, abnormal fat pad mass distribution, higher number of larger adipocytes, hepatic steatosis, higher expression of lipogenic proteins concomitant to decreased expression of PPARalpha and carnitine palmitoyltransferase I (CPT-1) in liver, and diminished expression of PPARgamma and adiponectin in WAT. Treatment with BZ ameliorated the hepatic and WAT abnormalities generated by diet-induced maternal obesity, with improvements observed in the structural, biochemical and molecular characteristics of the animals' livers and epididymal fat. Conclusion Diet-induced maternal obesity lead to alterations in metabolism, hepatic lipotoxicity and adverse liver and WAT remodeling in the offspring. Targeting PPAR with Bezafibrate has beneficial effects reducing the alterations, mainly through reduction of WAT inflammatory state through PPARgamma activation and enhanced hepatic beta-oxidation due to increased PPARalpha/PPARgamma ratio in liver.


Introduction
Obesity and comorbidities (metabolic syndrome, MS, leading to type 2 diabetes mellitus, DM2, cardiovascular disease, CVD, and non-alcoholic fatty liver disease, NAFLD) is due not only to environmental factors but also to maternal nutrition [1]. According to the Developmental Overnutrition Hypothesis, maternal overnutrition leads to permanent alterations in appetite control, neuroendocrine behavior and/or energetic metabolism in offspring during development, leading to obesity in adulthood even in the absence of excessive energy intake [2,3].
WAT is able to express and secrete several cytokines called adipokines, such as leptin, adiponectin and resistin, and has a role in the secretion of the proinflammatory cytokines tumor necrosis factor (TNF)-alpha, interleukin (IL)-6, monocyte chemoattractant protein (MCP)-1 and plasminogen activator inhibitor (PAI)-1 [4]. Obesity and its outcomes are closely linked to an increase in both visceral WAT and adipocyte size, highly affecting the function of this organ [5]. Visceral fat accumulation is a cornerstone for the development of obesity-related pathologies, such as NAFLD, in contrast with subcutaneous WAT [6,7].
NAFLD is the hepatic component of MS and is associated with obesity, insulin resistance (IR) and DM2 [8,9]. As the condition progresses, NAFLD evolves to non alcoholic steatohepatitis (NASH) because of continuous inflammation and the peroxidation of lipids. NASH can progress to cirrhosis through hepatic fibrosis and advance to hepatocarcinoma [10,11]. In addition, maternal diet during gestation and lactation can induce IR and NAFLD in offspring [3,12].
Peroxisome proliferator-activated receptors (PPARs) are a family of transcription factors (TFs) existing in three isoforms: PPARalpha, PPARbeta and PPARgamma. They are closely linked to carbohydrate, protein, and lipid metabolism as well as to cellular proliferation [13,14]. BZ is often used in clinical practice as a hypolipidemic drug. It was reported as a pan-PPAR agonist activating all three isoforms, but it preferentially activates PPARalpha and PPARbeta [15,16]. BZ has been demonstrated to prevent DM2 both in patients with CVD and in patients having undergone heart attack [17]. Furthermore, it has been shown to result in improvement of NASH in an animal model [18]. Therefore, the present study was undertaken to evaluate whether Bezafibrate could attenuate alterations in the livers and WAT of male mice offspring from diet-induced obese dams.

Animals and diet
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (NIH Publication number 85- 23, revised 1996). The protocol was approved by Ethics Committee for Animal Experimentation of the State University of Rio de Janeiro (Protocol Number CEUA/012/ 2011). All efforts were made to minimize suffering.
Animals were maintained under controlled conditions (2062uC, humidity 60610% and 12 h/12 h dark/light cycle) and had free access to food and water. Four-week-old females were randomly divided into two nutritional groups (n = 10), standard-chow (SC; 10% of energy from lipids) or high-fat diet (HF; 49% of energy from lipids), and were fed these diets for 8 weeks. Food intake of dams was measured by Compulse v 2.7.13 (Harvard/Panlab, Barcelona, Spain) two weeks before mating, and body mass (BM) was measured weekly. At 12 weeks old, females were mated with breeding males, and during gestation and lactation, the females continued receiving the same diets. Dams received experimental diets during 8 weeks pre-mating in order to assure that they were overweighed at the time of conception. At weaning, male offspring were subdivided into four groups (n = 10), to include BZ treatment: SC, SC/BZ, HF, and HF/BZ (according to Figure 1). Treatment with BZ started at 12 weeks of age and was maintained for three weeks. Food intake of offspring was measured by Compulse two weeks before sacrifice, and body mass (BM) was measured weekly. Offspring from HF and SC dams were weaned onto the same diet, SC diet. Therefore, metabolic abnormalities observed in HF offspring could be attributed to diet-induced maternal obesity as animals received balanced diet at postweaning period. All diets were supplemented with purified nutrients by PragSolucoes (Jau, Sao Paulo, Brazil) and were in accordance with the American Institute of Nutrition's recommendations (Table 1). BZ (Sigma-Aldrich Co., St. Louis, Mo., USA) was added to the diets at the dose of 100 mg/kg, which is a murine dose that can be compared to the clinical human dose of 10 mg/ kg [19].

Oral glucose tolerance test (OGTT)
OGTT was performed at one week before mating in dams, and at the end of treatment in offspring. Mice were fasted for 6 h before glucose administration. Glucose (1 g/kg) was given orally at time 0; tail blood was collected at fasting (i.e., baseline) and then at 15, 30, 60 and 120 min after glucose loading for glucose determination (Glucometer Accu-Chek, Roche, SP, Brazil). The area under the curve (AUC; millimolar per minute) was calculated.
The values of glucose concentration at time 0 were used to determine the fasting glycemia.  Sacrifice and tissue extraction Before sacrifice, animals were deprived of food for 6 h and then deeply anaesthetized (150 mg/kg of sodium pentobarbital, intraperitoneally). Blood samples were rapidly obtained by cardiac puncture, and plasma was separated by centrifugation (1206 g for 20 min) at room temperature and stored individually at 220uC until assay. Both the liver and the fat pads (inguinal, retroperitoneal and epididymal fat pads) were carefully dissected, weighed and prepared for analysis. Portions of the liver and the epididymal fat pad were kept in a freshly prepared fixative solution (formaldehyde 4% w/v, 0.1 M phosphate buffer pH 7.2) during 48 h and prepared to microscopy. Additional portions of the liver and epididymal fat pad were rapidly frozen for molecular analysis by immunoblotting and RT-qPCR. Remaining liver was prepared for biochemical analysis, and the adiposity index calculated and determined as the ratio between the sum of the masses of all fat pads divided by the total body mass, represented as a percentage [20].

Homeostasis model assessment of insulin resistance index (HOMA-IR)
Serum insulin levels were determined using the Milliplex mouse serum adipokine panel kit MMHMAG-44K-08 and the fasting serum glucose were obtained at the sacrifice. The HOMA-IR index was calculated as: (fasting serum glucose X fasting serum insulin/22.5) [21] Adipocyte microscopy Epididymal adipose tissue samples that were fixed, were embedded in Paraplast Plus (Sigma-Aldrich, St. Louis, MO, USA), sectioned at 5 mm and stained with hematoxylin and eosin (HE). Ten non-consecutive random microscopic fields were analyzed per animal on a light microscope (Leica Microsystems GmbH, Wetzlar, Germany) and an Infinity 1-5c camera (Lumenera Co., Otawa, ON, Canada). The mean diameter of at least 50 adipocytes per animal was measured using Image Pro Plus software v7.01 (Media Cybernetics, Silver Spring, MD, USA). The numerical density per area of adipocyte was evaluated in consideration of the number of adipocytes in a frame of known area when they did not hit two consecutive lines of the system (forbidden lines). The system was produced with the STEPanizer web-based system (www.stepanizer.com) [22].

Liver microscopy and biochemistry
Random fragments of the liver were prepared for light microscopy. Small pieces of liver were embedded in Paraplast Plus (Sigma-Aldrich Co., St Louis, Mo., USA), sectioned at 5 mm and then stained with HE. Several slices were cut per fragment, and five microscopic fields per animal were analyzed at random. Digital images (same system described) and a test system made up of 36 test points (P T ) produced with the STEPanizer web-based system were used for the analysis. The volume density (Vv) was estimated by point counting for hepatocytes and steatosis [23]: Vv[structure] = Pp[structure]/P T , where Pp is the number of points that hit the structure. Then, the numerical density per area of hepatocyte binucleation (binucleation/mm 2 ) was estimated. A total of 50 digital images were evaluated per group, and all binucleation within the frame were counted except for those that hit the ''forbidden lines'' [24].
Several fragments of the liver of each animal were frozen at 280uC for further biochemical analysis. The hepatic triglyceride levels were measured. Briefly, 50 mg of frozen liver tissue was placed in an ultrasonic processor with 1 ml of isopropanol. The homogenate was centrifuged at 2000 g, and 5 ml of the supernatant was used to measure TAG with a kit and an automated spectrophotometer (Bioclin System II, Quibasa Ltda., Belo Horizonte, Brazil) [23].

Immunoblotting
For liver and WAT analyses, total proteins were extracted in homogenizing buffer with protease inhibitors. Next, the homogenates were centrifuged at 4uC, and the supernatants were collected. Equal quantities of total protein were resuspended in SDS-containing sample buffer, heated for 5 min at 100uC and separated by SDS/PAGE. After electrophoresis, the proteins were electroblotted onto a polyvinyldifluoride transfer membrane (Amersham Biosciences, Piscataway, N.J., USA). The efficiency of the transfer was visualized by Ponceau solution staining. The membrane was blocked by incubation with non-fat dry milk. Antibodies against sterol regulatory element binding protein-1c (SREBP-1c), PPARalpha, PPARgamma, adiponectin, glucose-6phosphatase (G6Pase), Phosphoenolpyruvate carboxykinase (PEPCK) and beta-actin were purchased from Santa Cruz. Hepatic and WAT homogenates were incubated with various polyclonal antibodies (as specified in the results section). Beta-actin served as a loading control for proteins.
All protein expression was detected using an ECL (enhanced chemiluminescence) detection system (Amersham Bioscience). Signals were visualized by autoradiography and determined by quantitative analysis of digital images of gels using ImageJ software, version 1.44 (NIH, imagej.nih.gov/ij, USA). The integral absorbance values were measured.

Liver RT-qPCR
Total RNA was extracted from approximately 50 mg of liver tissue using Trizol reagent (Invitrogen, CA, USA). RNA amount was determined using Nanovue (GE Life Sciences) spectroscopy, and 1 mg of RNA was treated with DNAse I (Invitrogen). Synthesis of the first strand cDNA was performed using Oligo (dT) primers for mRNA and Superscript III reverse-transcriptase (both Invitrogen). Quantitative real time PCR (RT-qPCR) was performed using a Biorad CFX96 cycler and the SYBR Green mix (Invitrogen). Primers for qPCR were designed using the Primer3 online software and are indicated in Table 2. Endogenous control TBP (TATA Box binding-protein) was used to normalize the expression of the selected genes. Efficiencies of RT-qPCR for the target gene and the endogenous control were approximately equal, and were calculated through dilution series of cDNA. Real Time PCR reactions were conducted as follows: after a pre-denaturation and polymerase-activation program (4 min at 95uC), forty-four cycles, each one consisting of 95uC for 10 s and 60uC for 15 s were followed by a melting curve program (60 to 95uC with heating rate of 0.1uC/s). Negative controls consisted of wells in which cDNA was substituted for deionized water. The relative expression ratio (RQ) of mRNA was calculated by the equation 2 2DDCt , in which 2DCT expresses the difference between number of cycles (CT) of the target genes and the endogenous control.

Data analysis
Values are shown as the mean and the standard error of mean (SEM). In the cases where we could confirm homoscedasticity of variances, comparisons among groups were made by a t-test or ANOVA followed by Holm-Sidak post-hoc test. A P-value#0.05 was considered statistically significant (GraphPad Prism version 6.01 for Windows).

Dams (before mating period)
Dams had no difference in BM (17.360.2 g) at 1 month of age. The HF dams had a higher energetic intake (+6%, P,0.03) and a higher BM gain than the SC dams (+29%, P = 0.043) during the 8 weeks before mating. Moreover, the AUC of the OGTT (+17%, P,0.02) and the fasting glucose levels (+13%, P,0.02) were higher in the HF dams than in the SC dams. These parameters are shown in Table 3.

Offspring
Food and energy intake and BM evolution. The HF offspring had a higher food intake than the SC offspring (+21%, P = 0.0015). The SC/BZ and HF/BZ offspring ingested 17% and 14% less food than the SC and HF offspring, respectively (P = 0.0064). In regards to BM, there were no differences in the BM of the offspring up to 12 weeks of age, but after 13 weeks of age, the HF offspring were heavier than the SC offspring, and at the end of experiment, the HF offspring were 10% heavier than the SC offspring (P = 0.0024). Both groups of drug-treated offspring showed a reduced BM, and at the end of experiment, the SC/BZ and HF/BZ groups were 10% and 24% lighter, respectively, than their counterparts (P = 0.0061 and P,0.0001). Table 4 and Figure 2 show these findings.
Carbohydrate metabolism. The HF offspring had higher fasting glucose values than the SC offspring (+23%, P = 0.0009), while the HF/BZ offspring showed a decrease in fasting glucose levels in comparison with the HF (236%, P,0.0001), the SC (223%, P = 0.002) and the SC/BZ (221%, P = 0.0017) offspring. Additionally, as shown in table 4, impairment in glucose tolerance in the HF offspring in comparison with the SC offspring (+26% for the AUC, P = 0.0041) was observed, which was minimized by the BZ treatment in the HF/BZ offspring (227% for the AUC, P = 0.0003). Moreover, also presented in table 4, the HOMA-IR Table 2. RT-qPCR primers and respective sequences.  indicates that the HF offspring presented insulin resistance when compared to the SC offspring (+162%, P,0.05) and the treatment with BZ improved this parameter in the HF/BZ offspring (265%, P,0.05). Plasma analysis. Plasma TG levels were affected by maternal obesity as the HF offspring had a higher concentration of TG than the SC offspring (+23%, P,0.0001). However, the HF/BZ offspring showed an improvement in this parameter in comparison to the HF offspring, with a consistent decrease of 26% (P,0.0001). Furthermore, the plasma TG levels in the HF/BZ offspring were lower than those in the SC (210%, P = 0.0008) and the SC/BZ (26%, P = 0.011) offspring (Table 4).
Adiponectin, IL-6, insulin, leptin and resistin levels in offspring were also affected by diet-induced maternal obesity, but BZ improved all these parameters, especially in the HF/BZ group, as shown in table 4.
Liver mass and hepatic TG levels. The HF offspring had greater liver mass than the SC offspring (+19%, P = 0.008), and both groups of treated offspring had increased liver masses in relation to their controls (+103% for SC/BZ; +53% for HF/BZ, P,0.0001). Despite presenting larger liver masses, both treated offspring groups presented reduced hepatic triglycerides: the SC/ BZ group showed lower concentrations than the SC group (215%, P = 0.003), and the HF/BZ group had a reduction of 16% when compared to the untreated HF group (P = 0.0006). Conversely, the HF offspring had a higher hepatic triglyceride concentration than the SC offspring (+21%, P = 0.0003). All of these findings are described in table 5.
Liver steatosis and binucleation. The presence of liver steatosis was confirmed in the HF group, and its improvement was shown in the HF/BZ group using stereological quantification. Fewer lipid droplets were observed in the livers of the SC group (11.6961.3%), which was considered normal, while the SC/BZ group exhibited a sharp decrease (249%, P = 0.014) in liver lipid droplet levels. Although steatosis was substantial in the HF group (+143%, P,0.0001), the BZ treatment promoted a significant decrease in the HF/BZ group (285%, P,0.0001). These findings are depicted in table 5.
Numerical density of binucleation showed that the HF group had higher profiles of binucleation than the SC group (+57%, P = 0.004, table 4), while the HF/BZ group showed a decrease in the binucleation profile in relation to the untreated HF group (231%, P = 0.004, table 5).
Fat pad mass and adipose index. The HF mice showed an increase in epididymal (+24%, P = 0.0008) and retroperitoneal (+26%, P = 0.007) fat pad mass and a decrease in inguinal fat pad mass (220%, P = 0.0006) compared with the SC mice. The SC/ BZ group presented a decrease (216%, P = 0.014) in epididymal fat pad mass compared with the SC group, whereas, the HF/BZ group presented a decrease in epididymal and retroperitoneal fat pad mass compared with untreated HF group (243%, 244%, respectively, P,0.0001). The epididymal/inguinal fat pad mass ratio was higher for the HF group than for the SC group (+31%, P = 0.016), and treatment with BZ reversed this ratio in the HF/ BZ group (237%, P,0.0001). Similarly, the adipose index was elevated for the HF group when compared to the SC group (+8%, P = 0.004), and the HF/BZ group showed a decrease in the adipose index (213%, P,0.0001) when compared with the untreated HF group (Table 5).
Adipocyte diameter. The adipocytes of the HF mice were considerably enlarged in comparison to those of the SC mice (+24%, P = 0.0004). The SC/BZ and HF/BZ mice had adipocytes of a decreased size (216%, P = 0.009; 234%, P,0.0001, respectively) in comparison with their controls. The numerical density of adipocytes per unit area in the HF group was smaller than in the SC group (231%, P = 0.0005); however, the SC/BZ group showed an elevated numerical density of adipocytes per unit area when compared to the SC group (+31%, P = 0.0005). The HF/BZ group showed similar results, with the numerical density of adipocytes per unit area of adipocytes per unit area 77% higher than in the untreated HF group (P,0.0001) and 21% higher than in the SC group (P = 0.02) (table 5). Adipocyte diameter distribution (ADD). When ADD was performed, the HF group had higher numbers of large adipocytes than the SC group (P,0.0001), most likely due to maternal obesity. However, the HF/BZ group had higher amounts of small adipocytes than the HF group (P,0.0001), with similar values to the SC and SC/BZ groups. The ADD can be observed in figure 3.
Liver immunoblotting. In order to verify the hepatic IR, G6Pase and PEPCK enzymes were assessed by immunoblotting. The HF offspring presented an elevated expression of G6Pase (P = 0.001) while the HF/BZ offspring showed a decreased expression of the same enzyme (P,0.05) ( Figure 4A). Similar results were found in relation to PEPCK in that the HF offspring presented an elevated expression of PEPCK (P,0.05) and the HF/BZ offspring, a decrease expression of the same enzyme (P,0.05) ( Figure 4B). In consideration of beta-oxidation and lipogenesis, PPARalpha expression was decreased in the HF group when compared to the SC group (P,0.0001) and increased in both BZ-treated groups in comparison with their counterparts (P,0.0001); however, the SC/BZ group showed elevated expression of this TF in comparison with the HF/BZ group (P,0.0001) ( Figure 4C). In contrast, we observed that SREBP-1c expression was higher in the HF group than in the SC group (P = 0.0007); nevertheless, treatment with BZ did not affect the expression of this TF ( Figure 4D). Although PPARgamma expression in the HF group was also elevated in comparison to the SC group (P = 0.01), it was decreased in the HF/BZ group in relation to the untreated HF group (P,0.0001) ( Figure 4E).
Considering the results above, the PPARalpha/SREBP-1c ratio was lower in the HF group compared to the SC group (P = 0.0176), higher in the SC/BZ group compared to the SC group (P,0.0001) and higher in the HF/BZ group compared to the untreated HF group (P = 0.008) ( Figure 5A). Likewise, the PPARalpha/PPARgamma ratio was higher in the SC/BZ group compared to the SC group (P,0.0001) and in the HF/BZ group compared to the HF group (P,0.0001). Moreover, the PPARalpha/PPARgamma ratio was lower in the HF group compared to the SC group (P = 0.01) ( Figure 5B).
WAT immunoblotting. There is no significant difference between the SC offspring and the HF offspring in relation to PPARalpha expression, but both treated group presented an elevation in the expression of this TF (P = 0.001 for the SC/BZ group in comparison with the SC group and P,0.05 for the HF/ BZ group in comparison with the HF group, figure 6A). PPARgamma expression in adipose tissue was lower in the HF group than in the SC group (P = 0.003), and both drug-treated groups presented an elevation in the expression of this TF (P = 0.0004 for the SC/BZ group in comparison with the SC group and P = 0.0147 for the HF/BZ group in comparison with the HF group, figure 6B). Similarly, adiponectin expression was decreased in the HF group (P = 0.03), and treatment with BZ restored the expression in both treated groups (P,0.0001) ( Figure 6C).    Liver RT-qPCR. Considering the relative mRNA expression of PPARalpha and its target gene, CPT-1, it was found that both treated groups presented an elevation of both mRNAs (P,0.001) and, inversely, the HF group presented a decreased expression of these mRNAs (P,0.05 for PPARalpha mRNA and P,0.01 for CPT-1 mRNA). In addition to these findings, relative mRNA expression of PPARgamma was increased in the HF group (P,0.001) while the HF/BZ presented a decreased expression (P,0.05). Interestingly, CD36, a target gene of PPARgamma, presented an elevated expression in both treated groups (P,0.001) and the HF group also presented a slight elevation of this protein (P,0.05). Relative mRNA expression of SREBP-1c was elevated in both treated groups and also in the HF group (P,0.001) (Figures 7 and 8).
Hence, the PPARalpha/SREBP-1c and the PPARalpha/PPARgamma ratios were higher in both treated groups when compared to the SC group and to the HF group (P,0.001) and the HF group also presented a decreased ratio in comparison to the SC group (P,0.05) (Figure 9).

Discussion
Diet-induced obesity in dams during pre-mating, gestation and lactation periods produced offspring with noticeable BM gain, high levels of plasma and hepatic TGs, high levels of proinflammatory and low levels of anti-inflammatory adipokines, impairment of glucose metabolism, abnormal fat pad mass distribution, higher numbers of larger adipocytes, hepatic steatosis, high mRNA expression of lipogenic proteins and its target genes concomitant to decreased expression of PPARalpha and CPT-1 in liver, and diminished expression of PPARgamma and adiponectin in WAT. All these findings could be accounted for by diet-induced maternal obesity once all offspring received a balanced diet at weaning instead of the same diet as their mothers.
Treatment with BZ ameliorated the hepatic defects generated by maternal obesity as well as WAT remodeling, with improvement in the structural, biochemical and molecular characteristics of the animals' livers and epididymal WAT. These benefits can be attributed to drug administration and the negative energy balance observed in treated animals since such improvements were not exhibited by animals exposed to diet-induced maternal obesity without being treated.
HF offspring had hyperphagia, an expected behavior accounted for by hypotalamic modifications owing to excessive maternal energy intake, leading to obesity in adulthood [25,26]. BZ yielded negative energetic balance with decreased food intake in both treated offspring groups and this observation agrees with the reduced body mass observed in treated offspring compared to their counterparts [27].
Diet-induced maternal obesity triggers increased non-esterified fatty acids (NEFA), triglycerides, glucose and inflammatory cytokine influx to the fetus, which provoke dyslipidemia and glucose intolerance in the HF offspring [3,26]. Moreover, the HF offspring also showed higher levels of plasma insulin and leptin than the SC offspring. Increased adiposity index correlates with elevated circulating levels of leptin, which in the pancreas act as a regulator of insulin secretion in a so called ''adipose-insular axis''. Under normal physiological conditions, insulin stimulates adipogenesis and leptin production, whereas leptin inhibits the production of insulin by pancreatic beta cells. However, as the individual put on weight, adipose tissue stores increase and leptin reaches supraphysiological levels and loses its capacity to suppress insulin production. As a result, hyperleptinemia is followed by hyperinsulinemia, which leads to hyperglycemia in the long term [28]. Furthermore, it is widely known that normal levels of insulin inhibit G6Pase and PEPCK gene transcription and overexpression of these enzymes is a consequence of IR [29]. In this study, it was demonstrated that the HF offspring presented an elevated expression of both enzymes that characterizes IR. On the other hand, treatment with BZ reduced the expression of theses enzymes in both treated groups, pointing the effects of BZ on improvement of insulin sensibility.
Activation of PPARalpha by BZ is widely known as an important step to produce its hypolipidemic effects [30]. Reducing adiposity index and modifying WAT's secretion profile, BZ yields normalization of leptin levels and restoration of adequate insulin levels. This explanation agrees with glucose metabolism improvement observed in the drug-treated offspring from diet-induced obese dams.
The HF offspring's WAT showed elevated numbers of larger adipocytes in comparison with the SC offspring, complying with expansion of visceral WAT and hypertrophy of adipocytes. Thus, these animals also showed altered metabolism of WAT, showing increased release of pro-inflammatory adipokines such as IL-6, resistin and TNF-alpha and decreased adiponectin plasma levels as observed in previous studies [5,31]. Treatment with BZ has beneficial effects attenuating the pro-inflammatory adipokine profile in the HF/BZ offspring. Several works have demonstrated that activation of the PPAR isoforms could change the obesityinduced inflammatory state [18,32]. PPARalpha activation serves as a negative regulator of the inflammatory process by antagoniz-  ing the activity of tissue factor pathways linked to inflammation such as the nuclear factor (NF)-kappaB [33].
PPARgamma activation in WAT is responsible for decreased production of adipokines that cause IR, such as TNFalpha and resistin [34]. Additionally, PPARgamma activation causes apoptosis of mature and large adipocytes and increases the population of small adipocytes, which are more insulin sensitive [35]. These findings are in accordance with results found here in drug-treated offspring such as increased expression of PPARgamma in WAT, the increase in the number of adipocytes per area and in the population of small adipocytes, as well as decreased adipocyte diameter, indicating an improvement of morphology and consequently metabolism of this tissue. These data indicate that BZ is a potent drug that can be used to reverse the WAT inflammatory state through PPARgamma activation. Besides, PPARalpha also has an important role in mitochondrial fatty acid oxidation in adipocytes even with lower PPARalpha mRNA expression level in adipocytes than in the liver. Furthermore, activation of this TF also leads to adipocytes differentiation, suggesting that PPARalphaspecific adipogenic pathways exist, although the effect of PPARalpha seems to be partially shared with that of PPARgamma [36]. Our results show that both drug-treated groups presented an elevation in the expression of PPARalpha in WAT. Interestingly, there is no difference between the SC offspring and the HF offspring, but it is important to consider and highlight that the HF offspring were weaned onto a balanced diet.
Adiponectin down-regulation in obesity is a key factor in the development of NAFLD because it is closely linked to IR and diabetes [18,20]. Likewise plasmatic levels, WAT expression of adiponectin was lower in the HF offspring than in the SC offspring. Both conditions were reversed by treatment with BZ. Furthermore, adiponectin is a direct target of regulation by PPARgamma, whose expression was also increased in WAT from our treated animals, corroborating our results [37]. When treated with Pioglitazone, a PPARgamma agonist, 3T3-L1 adipocytes presented enhanced insulin sensitivity through an upregulation of adiponectin receptor 2 (AdipoR2) and adiponectin, a novel pathway that might be addressed [38].
The high hepatic TG levels found in the HF offspring was alleviated by treatment in HF/BZ offspring. These data correlate with hepatic steatosis, confirming the effects of diet-induced maternal obesity in the development of NAFLD [3]. Moreover, higher level of hepatocyte binucleation in the HF offspring than in the SC offspring indicates an unsuccessful attempt to recover from liver damage [12]. Targeting PPARs with BZ reduced hepatic steatosis as well as binucleation rate, leading to improvement in hepatic structure.
PPARalpha activation by fibrates in rodents induces hepatic peroxisomal proliferation and hepatomegaly [39]. In fact, the dose administered caused an increase in liver mass in both treated groups; however, this effect is lost in humans, even when fibrates maintain PPARalpha activation. It has been proposed that PPARalpha is more highly expressed in the hepatic tissue in rodents than in humans [40]. However, the role of PPARalpha as the master regulator of hepatic lipid metabolism is well conserved between these species [41].
TG accumulation in the liver is a result of increased hepatic fatty acid synthesis and/or decreased beta-oxidation [42]. PPARs target differently hepatic beta-oxidation and lipogenesis [42] and in order to gain insight into molecular mechanism involved into these processes, evaluation of the expression of some proteins as well as mRNA expression are relevant. PPARalpha target genes encode enzymes involved with mitochondrial beta-oxidation [43], by which PPARalpha has a role in energy homeostasis, contributing greatly to energy production via oxidative phosphorylation generating ATP [42]. Our findings showed that both drug-treated groups exhibited high levels of PPARalpha expression and that conversely, the HF offspring showed lower levels of PPARalpha expression after exposure to diet-induced maternal obesity. The expression of PPARalpha correlates inversely with hepatic steatosis degree. CPT-1, a target gene of PPARalpha, is the mitochondrial gateway for fatty acid entry into the mitochondrial matrix, being the master regulator of the hepatic mitochondrial beta-oxidation [42,44], and it was reported that patients with NAFLD presents a decreased expression of CPT-1 gene [45]. Our findings showed that BZ could restore the expression of CPT-1 in both treated groups, ameliorating the hepatic steatosis caused by maternal dietinduced obesity.
Although PPARgamma is expressed at low concentrations in hepatic tissue, patients with NAFLD/NASH exhibit significantly high levels [46]. Concomitantly, SREBP-1c is also found in high levels in patients with NAFLD, acting with PPARgamma to favor lipogenesis and accumulation of triglycerides in hepatic tissue [3,47]. According to our data, diet-induced maternal obesity led to an increase in both PPARgamma and SREBP-1c in HF offspring, which justifies the steatosis found in this group. However, treatment with BZ decreased PPARgamma expression in the liver of the HF/BZ offspring in comparison to the untreated HF offspring, reducing hepatic lipogenesis.
Fatty acid translocase (FAT)/CD36, a PPARgamma target gene, is involved with long chain fatty acid (LCFAs) transport into mitochondria, correlating with oxidative capacity of the liver as long as CPT-1 is also present [44,48]. Although (FAT)/CD36 does not play a significant role in fatty acid uptake in liver, it makes possible the initial metabolization of LCFA, consisting of shortening its chain to enable the fatty acid to enter the mitochondria through CPT-1 action (maximum of 20 carbons in the molecule) [48]. Although HF group presented enhanced expression of (FAT)/CD36 mRNA, those animals also had lower levels of CPT-1 mRNA. Hence, the acetyl-coA resulting from the shortening of LCFA that cannot enter beta-oxidation pathway and might be deviated to lipogenesis and/or ketone bodies formation [42]. Besides, excessive fatty acids could be stored as lipid droplets. Conversely, the increased mRNA levels of CD36 in treated groups reinforce the activation of the three PPARs isoforms by BZ as PPARbeta activation increases the capacity of PPARgamma to induce CD36 transcription [49]. A previous work showed that Bezafibrate (Pan-PPAR agonist) overcome Troglitazone (total PPARgamma agonist) potential to induce CD36 [50]. In the present study, interplay between (FAT)/CD36 and CPT-1, both higher expressed in treated groups, guarantees the maximum oxidative activity within hepatic mitochondria, enhancing fatty acid metabolization and, thus, reducing hepatic steatosis.
The reduced PPARalpha/SREBP-1c ratio observed in the untreated HF offspring indicates higher SREBP-1c activation with lower PPARalpha expression, agreeing with a previous report [51] that showed higher SREBP-1c/PPARalpha ratio in obese patients. Although BZ did not modify SREBP-1c expression levels in the HF group, the activation of PPARalpha in the treated groups is more significant than SREBP-1c activation, favoring betaoxidation. Similarly, the PPARalpha/PPARgamma ratio was significantly decreased in the offspring from diet-induced obese dams and increased in the treated offspring.
The literature emphasizes that both PPARgamma and SREBP-1c are essential to trigger hepatic lipogenesis [51]. However, SREBP-1c seems to be an auxiliary of PPARgamma [52], which emerges as the master regulator of hepatic lipogenesis. In this way, SREBP-1c deletion in mice results in 50% less fatty acid synthesis [53], whereas mice deficient in liver-specific PPARgamma are completely protected against steatosis development [54]. In the present study, treatment with BZ did not modify SREBP-1c expression in the HF/BZ offspring when compared to the HF offspring. Despite unchanged expression of SREBP-1c mRNA after treatment with BZ, levels of PPARgamma mRNA were considerably reduced, being crucial to the lower hepatic triglycerides and liver steatosis found in this treated group.
Additionally, the relative abundance of PPARalpha in normal liver depicted by and increased PPARalpha/PPARgamma ratio in treated groups, causes this receptor to act as a regulator of fatty acid catabolism, minimizing the need to store these lipids in hepatocytes. Therefore, PPARalpha could protect the liver against PPARgamma lipogenic activity [55].
All of these findings support the conclusion that BZ can ameliorate hepatic metabolic abnormalities because it causes an increase in beta-oxidation via PPARalpha activation and rise in the target gene CPT-1 in association with a decrease in PPARgamma expression and no change in SREBP-1c expression.
The observation that increased PPARalpha/PPARgamma ratio counters hepatic lipogenesis, even with a lack of effect upon SREBP-1c expression, creates new perspectives aiming to unravel pathways by which PPARalpha can protect the liver against lipogenesis induced by PPARgamma overexpression as well as the precise mechanisms that underlie SREBP-1c and PPARgamma interaction in the development of NAFLD.
The present study has certain limitations. The major aim of this study was to verify the efficiency of BZ to reverse the alterations caused by maternal diet-induced obesity in offspring's liver and WAT. Therefore, we examined the activation of PPARalpha and PPARgamma through BZ as well as proteins related to betaoxidation and lipogenesis and metabolic parameters. Further studies are necessary to clarify separate actions of PPARalpha and PPARgamma in both organs as a pair-feeding group would enrich the discussion about BM loss and the role of PPARs in this context.
In conclusion, diet-induced maternal obesity lead to alterations in metabolism, hepatic lipotoxicity and adverse liver and adipose tissue remodeling caused in the offspring. Targeting PPAR by BZ has beneficial effects reducing the alterations, mainly through reduction of WAT inflammatory state through PPARgamma activation and enhanced hepatic beta-oxidation due to increased PPARalpha/PPARgamma ratio in liver, resulting in higher expression of target genes involved with this pathway. The relevance of these observations stem from the fact that the effects of maternal obesity are not restricted to the first generation and might compromise future generations, leading to a continuation in increasing rates of obesity and DM2 worldwide.