Introduction

Feeding behavior is a complex process regulated by a vast neural circuitry and associated neuropeptides, as well as by circulating nutrients, hormones and cytokines. These peripheral inputs are integrated within neuronal populations located in the arcuate nucleus (ARC) of the hypothalamus (Plum et al. 2006). The ARC contains neurons that produce the orexigenic peptides, neuropeptide Y (NPY) and agouti-related peptide (AgRP), and anorexigenic neurons that synthesize pro-opiomelanocortin (Pomc) and the cocaine- and amphetamine-regulated transcript (CART) (Munzberg and Myers 2005; Niswender et al. 2004). Signals from the ARC arrive to second-order neurons located mainly in the lateral (LH) and paraventricular (PVN) nuclei, which lead to orexigenic and anorexigenic responses, respectively, through regulation of neuropeptides and sympathetic activity (Sakurai et al. 1998; Elmquist et al. 1999; Schwartz et al. 2000; Flier 2004; Myers and Olson 2012).

The adipocytokine, leptin, crosses the blood–brain barrier (BBB) and acts on the hypothalamus by increasing the synthesis of anorexigenic neuropeptides and suppressing the orexigenic neuropeptides, which results in the inhibition of food intake (Jequier 2002; Dhillon and Belsham 2011). Obesity is a worldwide epidemic that contributes to the development of insulin resistance and type 2 diabetes (T2D) (Kahn et al. 2006). The disruption of the actions of the central nervous system (CNS) may contribute to obesity onset (Purkayastha and Cai 2013). In rodents submitted to high-fat diet (HFD), the hypothalamic anorexigenic effect of insulin and leptin is diminished (Münzberg et al. 2004; De Souza et al. 2005; Prada et al. 2005; Moraes et al. 2009). A HFD has shown to increase inflammatory markers in the hypothalamus that impair food intake control, contributing to adiposity (Münzberg et al. 2004; De Souza et al. 2005; Zabolotny et al. 2008).

The intrauterine period and early life are critical time-windows in which neuroendocrine axes may be programmed by environmental stressors, such as nutrient shortage (Remacle et al. 2007). As such, pups born to dams fed on a low-protein diet during gestation and lactation present long-lasting perturbations in the expression profile of hypothalamic neuropeptides and these alterations are associated with increased food intake (Orozco-Solis et al. 2011). However, whether the leptin signaling disruption that occurs in obesity (Münzberg et al. 2004) underlies the effects of protein restriction upon hypothalamic neuropeptide expression profile remains uncertain.

Taurine (Tau) is a semi-essential amino acid present in high concentrations in the brain and is responsible for the maintenance of intracellular osmotic balance (L’Amoreaux et al. 2010). Tau suppresses the activation of the endoplasmic reticulum (ER) stress response in the CNS (Pan et al. 2010; Gharibani et al. 2013). In the hypothalamus, Tau decreases NPY expression and enhances the anorexigenic effects of insulin (Solon et al. 2012). In the periphery, Tau also improves insulin secretion (Ribeiro et al. 2009; Batista et al. 2012), increases insulin receptor phosphorylation (Camargo et al. 2013), and prevents obesity (Tsuboyama-Kasaoka et al. 2006; Batista et al. 2013b). Here, using a microarray analysis, we found that several genes involved in the hypothalamic control of food intake are altered in protein-restricted mice. The association of protein restriction and HFD also changed the profile of hypothalamic mRNA expression. These alterations were associated with hypothalamic leptin resistance, which impairs food intake control in protein-restricted mice that are fed on a HFD diet. Tau supplementation prevents the disruption of leptin signaling induced by HFD and modulates the expression of several hypothalamic genes.

Materials and methods

Animals and diets

All experiments were approved by the ethics committee at UNICAMP (certificate no.: 2826-1). Weaned 30-day-old male C57Bl/6 mice were obtained from the breeding colony at UNICAMP and maintained at 21 ± 2 °C on a 12 h light–dark cycle with free access to food and water. The mice were randomly assigned to the following groups: (a) Control mice fed on a normo-protein diet (14 % protein) for 14 weeks (C group, n = 20); (b) protein-restricted mice fed on a diet containing 6 % protein for 14 weeks (R group, n = 18); (c) C mice submitted to HFD during the last 8 weeks of experimental period (CH group, n = 20); (d) R mice submitted to HFD during the last 8 weeks of the treatment (RH group, n = 20); (e) CH mice supplemented with 5 % Tau (CHT group, n = 19); or (f) RH-supplemented mice (RHT group, n = 19). Tau was administered in the drinking water from weaning until the end of the experiment. Diets were prepared according to AIN-93 guidelines (Reeves 1997), using casein as a protein source and were purchased from PRAGSOLUÇÕES Biociências© (Jaú, SP, BRA). All diet compositions have been previously described (Batista et al. 2013b).

Biometric and nutritional parameters

During the 8 weeks of HFD treatment, mice from each group were placed in metabolic cages and had their food intake and body weight (BW) measured (Batista et al. 2013b). Food intake was registered three times per week and then the mean daily food intake was determined and expressed either as g/BW/day or Kcal/day (Bachmanov et al. 2002; Bol et al. 2009; Batista et al. 2013b). At the end of the experimental period (15th week), after 12 h of fasting, the final BW and nasoanal length were measured to obtain the Lee Index [from the ratio of BW (g)1/3/nasoanal length (cm) × 1000] (Bernardis and Patterson 1968) in all groups of mice. Subsequently, all mice were euthanized by decapitation and the retroperitoneal and perigonadal fat pads and brains were removed and weighed. Blood was collected in heparinized tubes (1:1000) and centrifuged at 2655g for 15 min, and the plasma was stored at −20 °C. Commercial kits were used according to the manufacturer’s instructions for measurements of plasma triglycerides (TG), cholesterol (CHOL) (Roche/Hitachi; Indianapolis, USA), and total protein concentrations (Laborlab, Guarulhos, SP, BRA). Plasma glucose was measured using a glucose analyzer (Accu-Chek Performa, Roche Diagnostic, Switzerland), and insulin was measured by radioimmunoassay using human insulin radiolabelled with 125I as tracer, rat insulin as standard (Crystal Chem. Inc., Downers Grove, IL, USA), and rat insulin antibody (donated by Dr Leclerq-Meyer, Free University of Brussels, Brussels, Belgium). The charcoal–dextran method was used to separate free insulin from antibody bound to 125I-insulin (Ribeiro et al. 2010). The limits for detection ranged from 0.02 to 5 ng/mL and the coefficient of variation was lower than 10 %.

TAU content in the hypothalamus

The analysis of TAU content in the hypothalamus was performed with reverse-phase HPLC method. The hypothalamus was disrupted by sonication in 150–200 µL of a solution containing 0.2 M perchloric acid, 0.01 M EDTA, and 100 mM alfa-aminobutyric acid as an internal standard. The homogenates were centrifuged at 20,000g at 4 °C for 15 min. An aliquot of the supernatant (10 µL) was dried in SpeedVac, and derivatized with phenylisothiocyanate (Bidlingmeyer et al. 1987). The phenylthiocarbamoyl–amino acid derivatives in each sample were resuspended in 0.14 M sodium acetate buffer containing 0.05 % triethylamine, pH 7.5, and then separated using a LDC MP3000 HPLC binary system (Milton Roy Co., Riviera Beach, FL, USA) and Picotag C18 5 μm column, 3.9 × 150 mm (Waters Corporation, Milford, MA, USA) and detected by UV (254 nm). Elution was performed at 1 ml/min at 38 °C using a gradient of 0.14 M sodium acetate, 0.05 % triethylamine and pH 5.7 as solution A, and acetonitrile/H2O (60:40, v/v) as solution B. Amino acid standards (Standard H from Pierce Protein Reagent Products, Pittsburgh, PA, USA) were derivatized and analyzed together with the samples. The amino acid peak identification was carried out by the amino acid retention time, and area integration by LCTalk software (Thermo Separation Products). Amino acid content for hypothalamus was expressed as nmol/mg hypothalamus weight.

Microarray analysis

Global hypothalamic gene expression was measured using the microarray, GeneChip® Mouse Genome 430 2.0 Array, from Affymetrix® Company (Santa Clara, CA, USA). The array contains 45,000 probe sets that analyze the expression level of over 39,000 transcripts and variants from over 34,000 well-characterized mouse genes. After decapitation, mice brains were removed from the skull and placed in a ventral position. The entire hypothalamus was dissected away from the optic chiasm (anterior) and mammillary bodies (posterior) and total RNA was extracted using Trizol® (Invitrogen, São Paulo, SP, BRA). A total of 200 ng RNA was used for cDNA synthesis, which was processed with the 3′ IVT Express Kit Assay. The hybridization, washing, and revelation steps with streptavidin-PE were performed using the Hybridization, Wash, and Stain kit, according to the manufacturer’s instructions. The arrays were analyzed using an Affymetrix® GeneChip® Scanner 3000 (Affymetrix, Santa Clara, CA, USA) at the DNA microarray facility (LMA), Brazilian Synchrotron Light Laboratory (Campinas, SP, BRA). Scanned images were analyzed in the R environment using the Affy and RankProd packages from Bioconductor using the algorithm MAS 5. We normalized the arrays and then calculated the signal intensity and the detection (the presence or absence of expression). For statistical analysis, we used the Rank Product test, which is biologically motivated and designed to test and detect differentially expressed genes in replicated microarray experiments. This is a simple non-parametric statistical method based on ranks of fold changes. We used a P value <0.04 and a percentage of false positives (pfp) < 0.1, in order to consider a given transcript as differentially expressed. The number of probe sets in each comparison is shown in Table 3. Pathway enrichment analyses were conducted using the MetaCoreTM platform (http://www.genego.com/metacore.php, GeneGo, St. Joseph, MI, USA) (Gentleman et al. 2004; Gautier et al. 2004; Team 2011; Hong et al. 2013).

Real time RT-PCR

The microarray analysis of the hypothalamic neuropeptides was confirmed by real time quantitative RT-PCR. The hypothalamic mRNA extraction was performed using Trizol®, according to the manufacturer’s instructions (Invitrogen, São Paulo, SP, BRA). A sample of 2 µg of each RNA was used for each reverse transcription reaction (RT) with random primers using DTT (100 mM), a dNTP mix (10 mM), and the enzyme SuperScript II (200U; Invitrogen, São Paulo, SP, BRA). The samples were then incubated for 50 min at 42 °C to obtain the cDNAs. The PCR reactions were completed with a final volume of 20 µL and contained 4 µL cDNA (0.1 µg/µL), 10 pM of NPY (Forward 5′TACTCCGCTCTGCGACACTA3′, Reverse 5′TCTTCAAGCCTTGTTCTGGG3′), or AgRP (Forward 5′GAGTTCCCAGGTCTAAGTCTGAATG3′, Reverse 5′ATCTAGCACCTCCGCCAAAG3′), or Pomc (Forward 5′GGCTTGCAAACTCGACCTC3′, Reverse 5′TGACCCATGACGTACTTCCG3′) or CART (Forward 5′ACCTTTGCTGGGTGCCCGTG3′, Reverse 5′ TGCAACGCTTCGATCAGCTCC3′) primers and SYBR Green PCR Master Mix (Applied Biosystems, São Paulo, SP, BRA). The transcripts were detected using the 7000 real time PCR system (Applied Biosystems, São Paulo, SP, BRA), and the threshold cycle values were used to calculate the number of cycles required to first detect a statistically significant increase in fluorescence intensity due to the binding of SYBR Green to a molecule of double-stranded DNA. The values were normalized by the values of control gene, rlpl0 (ribosomal protein large P0, or 36B4) (Forward 5′GAGGAATCAGATGAGGATATGGGA3′, Reverse 5′AAGCAGGCTGACTTGGTTGC3′) and were shown as relative expression over the control group level (2-ΔΔCT).

Evaluation of leptin action on food intake

During the last week of treatment (14th week), mice were acclimatized in metabolic cages for 24 h and fasted for 6 h, from 1:00 to 7:00 p.m., when feeding activity is low (Orozco-Solis et al. 2011). At the end of this period, an intraperitoneal (i.p.) injection of saline (0.9 % NaCl) was administered and food was made available. The same protocol was repeated for two consecutive days. Food intake was measured 12 h after saline administration on each day and the average of two consecutive days was determined. On the fourth and fifth days, the same mice were fasted for 6 h (1:00 to 7 p.m.) before receiving a single i.p. injection of leptin (500 ng/g BW; recombinant mouse leptin from ©Merck—Calbiochem cat. #429705, Darmstadt, Germany) according to (Vieira et al. 2012) with minor modifications. Food intake was measured 12 h after leptin administration and the average of two consecutive days was determined.

Western blot

The hypothalamic leptin action was evaluated in mice from all groups. After evaluation of food intake on the second day of leptin administration, mice were fasted for 1 h before another leptin load for hypothalamic leptin signaling assessment. Conscious mice received an i.p. injection of leptin (500 ng/g BW) or saline (0.9 % NaCl). After 45 min (without food access), the mice were euthanized by decapitation, and the hypothalamus was removed and immediately homogenized in antiprotease buffer containing 100 mM Tris pH 7.5, 10 mM sodium pyrophosphate, 100 mM sodium fluoride, 10 mM EDTA, 10 mM sodium vanadate, 2 mM PMSF, and 1 % Triton X-100. The extracts were then centrifuged at 15,294g at 4 °C for 40 min to remove insoluble material. The protein concentration was determined by the Bradford dye method (Bradford 1976), using bovine serum albumin as a standard curve. For PAGE and Western blot analysis, the samples were treated with a Laemmli sample buffer containing dithiothreitol. After heating to 95 °C for 5 min, the proteins were separated by electrophoresis (70 μg protein/lane, 10 % gels). Following electrophoresis, the proteins were transferred to nitrocellulose membranes. The membranes were treated with a blocking buffer (5 % non-fat dried milk, 10 mM Tris, 150 mM NaCl, and 0.02 % Tween 20) and were subsequently incubated overnight with a polyclonal antibody against the phospho-signal transducer and activator of transcription 3 (pSTAT3; 1:1000, cat. #9131S) or total STAT3 (1:1000, cat. #4904S). Both primary antibodies were from Cell Signaling Technology® (Danvers, MA, USA). Detection was performed after a 2 h incubation period using a horseradish peroxidase-conjugated secondary antibody (1:10,000, Invitrogen, São Paulo, SP, BRA). The band intensities were quantified by optical densitometry using the free software Image J Tool (http://ddsdx.uthscsa.edu/dig/itdesc.html). The densitometry values obtained from pSTAT3 were normalized by STAT3 protein expression. After assaying the target proteins, Western blotting was repeated using a rabbit polyclonal antibody to the glyceraldehyde 3-phosphate dehydrogenase protein (GAPDH; 1:1000; cat. no sc-25778, Santa Cruz Biotechnologies, CA, USA) as an internal control.

Statistical analysis

The results are presented as mean ± SEM for the number of samples (n) indicated. Statistical analyses were carried out using a Student’s t test (to compare results from saline and leptin-stimulated conditions, Fig. 3) or one-way analysis of variance (ANOVA) followed by Duncan’s post-test with the Statistica 5.0 software (Stantsoft, Tulsa, OK, USA), and the level of significance was set at P < 0.05. Prior to ANOVA analysis, a Gaussian distribution of the samples was assumed based on the Kolmogorov–Smirnov normality test using Graph-Pad Prism version 5.00 for Windows (GraphPad Software, San Diego, CA, USA).

Results

Metabolic characterization

At the end of the experimental period, R mice showed a lower body weight and nasoanal length (BW) compared with C mice (P < 0.05; Table 1). The HFD treatment increased BW and body length in the CH and RH groups (P < 0.05). Tau supplementation reduced BW in the CHT group, when compared with the CH mice (P < 0.05). HFD efficiently induced obesity, as the CH and RH groups showed a higher Lee index and retroperitoneal and perigonadal fat pads in comparison with the C and R groups, respectively (P < 0.05; Table 1). Tau supplementation decreased fat deposition only in the CHT mice (P < 0.05), but the treatment efficiently normalized the Lee index in both the RHT and CHT groups (Table 1). The R mice exhibited an enhanced ratio between brain weight and BW, when compared with the C mice (P < 0.05). All HFD-fed groups, regardless of Tau supplementation, showed a decrease in the brain/BW ratio in comparison with their respective controls (P < 0.05; Table 1). Hypothalamus weight was not modified either by malnutrition (R 7.1 ± 0.5 and C 7.1 ± 0.6 mg), HFD (CH 8.9 ± 0.2 and RH 6.6 ± 0.9 mg) or Tau supplementation (CHT 6.3 ± 0.5 and RHT 5.9 ± 0.4 mg). Additionally, hypothalamic Tau concentration was elevated by nearly 1.8-fold in RHT and CHT mice compared with their respective controls (P < 0.0001; Table 1). Of note, Tau content in the hypothalamus of R mice was also significantly higher than of that observed in C hypothalamus (P < 0.005; Table 1).

Table 1 Final body weight (BW), nasoanal length, Lee index, retroperitoneal (RP), and perigonadal (PG) fat pads, brain weight, and hypothalamic Tau concentration in C, CH, CHT, R, RH, RHT mice
Full size table

All HFD mice showed higher fasting glycaemia, when compared with their respective controls (P < 0.05; Table 2). Tau treatment normalized glycaemia only in the CHT group (Table 2), whereas, under fed conditions no alterations in plasma glucose levels were observed between the groups (Table 2). In addition, only the CH mice developed hyperinsulinemia, in both the fasted and fed states, when compared with the C mice (P < 0.05; Table 2). Tau supplementation normalized plasma insulin in the CHT mice (Table 2). The plasma lipid profile was also modulated by the nutritional interventions. The TG levels were unaltered during fasting, but were higher in the RH and RHT fed mice when compared with R (P < 0.05; Table 2). The HFD treatment increased plasma CHOL in the normal and protein-restricted mice in both nutritional states (P < 0.05; Table 2). Tau supplementation resulted in a reduction in CHOL only in the fasted CHT mice, when compared with the CH mice (P < 0.05; Table 2). In addition, the low-protein diet treatment lowered total plasma proteins, in comparison with the C group (P < 0.05; Table 2). The HFD treatment recovered plasma proteins in all groups (Table 2).

Table 2 Biochemical nutritional parameters analyzed in fasted and fed C, CH, CHT, R, RH, and RHT mice
Full size table

Food intake

Figure 1 shows the weekly ratio between food consumption and BW. Protein-restricted mice showed enhanced food intake, as judged by mean food intake during the 8 weeks of HFD treatment, when compared with the C group (P < 0.05; Fig. 1c). All the HFD mice had a lower food intake than the C mice (P < 0.05; Fig. 1c).

Fig. 1
figure 1

Food consumption in mice undergoing protein-restriction with or without HFD and Tau supplementation. The food intake of (a) the C, CH, CHT and (b) the R, RH, RHT mice was recorded for eight weeks. Mean food intake during the HFD treatment (c). Food (d), calorie (e) and protein intake (f) recorded during the last week of the experimental period. Data are mean ± SEM (n = 4–8). Asterisk different from C; ampersand different from CH; hash different from R; dollar different from RH. One-way ANOVA followed by Duncan’s post-test (P < 0.05)

Full size image

During the last week of treatment, the R mice were persistently hyperphagic (P < 0.05; Fig. 1d), but their calorie intake was matched to that of the C mice (Fig. 1e). In addition, although the RH and CH groups showed a lower amount of food intake, when compared with their respective controls (P < 0.05; Fig. 1d), the calorie intake was higher due to the hypercaloric nature of the diet (P < 0.05; Fig. 1e). Tau supplementation modified the total food intake only in the CHT mice (P < 0.05; Fig. 1d) and decreased the total calorie intake, when compared with the CH mice (Fig. 1e). However, the RHT mice persistently showed a hypercaloric intake at the end of the experiment (Fig. 1e). Protein intake in the R mice remained low despite an increase in food intake (Fig. 1d, f). The low food intake in the HFD mice also favored a lower protein intake; however, the CHT mice presented an increase in protein intake, when compared with the CH mice (Fig. 1f).

Hypothalamic gene expression profile

The altered food consumption observed in the protein-restricted mice and the hypercaloric intake observed in the RH and RHT groups suggest a deleterious action of the low-protein diet upon the hypothalamic control of food intake. Hence, we performed a microarray analysis of the hypothalamus in order to gain insight into the possible mechanisms responsible for these alterations.

The groups of mice exhibited changes in global hypothalamic gene expression. Hypothalamic genes were grouped into (1) control of food intake (Table 4); (2) tricarboxylic acid cycle (TCA), oxidative phosphorylation and glycolysis/gluconeogenesis (Table 5); (3) unfolded protein response, apoptosis, and endoplasmic reticulum (ER) stress (Table 6); and (4) signal transduction leptin/insulin signaling (Table 7).

The low-protein diet significantly changed the expression of 787 probe sets, 54 % of which were up-regulated and 46 % were down-regulated in the R mice. Furthermore, in the CH hypothalamus, the expression of 769 probe sets was modified, 56 % of which were up-regulated and 44 % down-regulated in comparison with the C group. Protein-restricted mice fed on a HFD modified the expression of 739 probe sets, with an up-regulation of 55 % and a down-regulation of 45 %, when compared with the R group. Finally, Tau supplementation altered the expression of 429 probe sets, with an up-regulation of 32 % and down-regulation of 68 % in the CHT mice, when compared with the CH group. Moreover, in the RHT mice, there were alterations in 540 probe sets, where 26 % were up-regulated and 74 % down-regulated, when compared with the RH group (Table 3).

Table 3 General data from the Affymetrix microarray analyses performed using the GeneChip® Mouse Genome 430 2.0
Full size table

Microarray analysis showed that hypothalamic neuropeptides involved in the control of food intake are regulated by diet and/or Tau treatment (Table 4). In the R group, NPY and AgRP gene expressions were higher, when compared with the C group (P < 0.0004 and P < 0.04; Table 4). Both the CH and RH mice showed a significant decrease in these orexigenic neuropeptides, when compared with the C mice (P < 0.0002 and P < 0.0001) and R mice (P < 0.0001), respectively (Table 4). Tau supplementation decreased the NPY mRNA content in the RHT group, when compared with the RH mice (P < 0.0001; Table 4), whereas it did not modify AgRP mRNA expression between the other groups. In addition, in the R mice, there was a decrease in CART prepropeptide (CARTpt) expression, when compared with the C mice (P < 0.0001), without any alteration in Pomc mRNA. Protein-restricted mice fed on a HFD diet showed a higher expression of Pomc and CARTpt genes, in comparison with the R group (P < 0.0001; Table 4). In the CH mice, only the expression of CARTpt was increased (P < 0.0007). Tau supplementation increased the gene expression of these anorexigenic neuropeptides only in the RHT group (Table 4).

Table 4 Genes involved in the control of food intake expressed by fold change between two groups
Full size table

In addition to genes related to food intake, we also found that protein restriction, HFD and/or Tau treatment modified the expression of several genes involved in different metabolic pathways in the hypothalamus (Tables 5, 6, 7).

Table 5 Hypothalamic genes involved in tricarboxylic acid cycle, glycolysis/gluconeogenesis, and oxidative phosphorylation expressed by fold change between two groups
Full size table
Table 6 Hypothalamic genes involved in unfolded protein response, apoptosis, and ER stress pathway expressed by fold change between two groups
Full size table
Table 7 Hypothalamic genes involved in leptin and insulin signaling, and the leptin signaling via PI3 K-dependent pathway/inflammation JAK-STAT pathway, expressed as fold change difference between two groups
Full size table

The genes involved in the tricarboxylic acid cycle (TCA), glycolysis/gluconeogenesis, and oxidative phosphorylation are shown in Table 5. Microarray analysis showed that 6 genes that play a role in glycolysis/gluconeogenesis were down-regulated in the R group, when compared with the C group (Table 5). In the hypothalamus from CH mice, 4 genes were up-regulated and 4 down-regulated. In the RH group, the mRNA content for 2 genes was up-regulated and for 4 genes was reduced (Table 5). Tau supplementation decreased Aldoc, but enhanced Mdh1, Aldoa, and Tpi1 in CHT, when compared with CH (Table 5). For oxidative phosphorylation, 15 genes were down-regulated by protein restriction (Table 5). The HFD treatment reduced the expressions of 8 genes and increased the expressions of 12 mRNAs in the hypothalamus of CH mice. In RH, 8 genes were down-regulated and 10 genes were up-regulated (Table 5). Tau prevented the HFD-induced alterations in genes involved in oxidative phosphorylation, as 9 genes showed an inverted pattern of expression, compared with the CH group (Table 5). Supplementation in the RHT group only reversed the expression of 6 genes (Table 5).

Genes involved in the unfolded protein response, apoptosis, and ER stress pathway were also modified between the groups and are shown in Table 6. Among the 23 genes involved in these processes that were modified by diet and/or Tau treatment, in the hypothalamus of R mice, 10 were down-regulated and 7 were up-regulated, when compared with the C mice (Table 6). Twelve genes were up-regulated and a decrease in 1 gene occurred in CH mice. In RH mice, 6 genes were up-regulated and 4 genes were reduced (Table 6). Tau treatment reversed the HFD-induced gene expression alteration in 7 mRNAs in the CH and 6 mRNAs in the RH mice that are involved in unfolded protein response, apoptosis, and ER stress pathway (Table 6).

The leptin signaling via PI3 K-dependent pathway and inflammation JAK-STAT pathway were altered in hypothalamus by diet and/or Tau supplementation and are shown in Table 7. In the R group, 15 genes involved in these pathways were up-regulated and 4 reduced in comparison with the C group (Table 7). HFD consumption lowered the content of 10 and 13 hypothalamic genes in the CH and in the RH group, respectively (Table 7). Tau supplementation reversed the expressions of 6 genes in the CHT mice and 2 genes in RHT mice (Table 7).

Gene expression of neuropeptides involved in feeding behavior

Figure 2 shows the gene expressions of the orexigenic and anorexigenic neuropeptides that were detected at significant levels in microarray analyses and were modified by diet and/or Tau supplementation. The hypothalamus from the R mice presented increases of 37 and 60 % in NPY and AgRP gene expressions, respectively, when compared with the C group (P < 0.05; Fig. 2a, b. Conversely, in the CH mice, reductions in NPY and AgRP of 46 and 42 %, respectively, were observed compared with the C mice (P < 0.05; Fig. 2a, b). In the RH group, there was a 31 % decrease in the mRNA content of NPY and AgRP compared with the R mice (P < 0.05; Fig. 2a, b). Tau did not alter the HFD-induced alterations in the expression of orexigenic neuropeptides (Fig. 2a, b).

Fig. 2
figure 2

Orexigenic and anorexigenic neuropeptide gene expression in the hypothalamus of mice that underwent protein-restriction with or without HFD and Tau supplementation. Hypothalamic mRNA extracts were processed for real time RT-PCR measurement of a NPY, b AgRP, c Pomc and d CART gene expression in the C, CH, CHT, R, RH, and RHT groups. Data are mean ± SEM (n = 5–8). Asterisk different from C; hash different from R. One-way ANOVA followed by Duncan’s post-test (P < 0.05)

Full size image

The mRNA content of Pomc and CART did not differ between the R and C mice (Fig. 2c, d). HFD treatment increased the expressions of Pomc and CART in the CH mice (P < 0.05; Fig. 2c, d). In the RH mice, no modifications in anorexigenic neuropeptides were observed in the hypothalamus. Tau supplementation did not modify these parameters in the CHT group (Fig. 2c, d); however, Pomc and CART mRNA were significantly increased in the RHT group (P < 0.05; Fig. 2c, d). It is important to highlight that the control gene rlpl0 was not altered by either dietary or Tau interventions.

Leptin-induced anorexigenic response and hypothalamic signaling

Finally, the hypothalamic control of food consumption was evaluated using leptin in all groups (Fig. 3). Leptin administration significantly reduced the 12 h food intake in the R and C mice (P < 0.03 and P < 0.004; Fig. 3a). This response was abolished in normal and protein-restricted mice fed a HFD. The anorexigenic response to leptin was restored in both Tau-supplemented groups (P < 0.04 in CHT and P < 0.03 in RHT; Fig. 3a). In the same mice, the effect of leptin on the hypothalamus was confirmed by the ratio of phosphorylated STAT3 (pSTAT3) to STAT3 total protein after an acute ip leptin administration (Fig. 3b). Leptin induced a 79 and 56 % increase in the pSTAT3/STAT3 protein expression in the C and R groups, respectively, (P < 0.008 and P < 0.04; Fig. 3b). No modification in pSTAT3/STAT3 content was observed for the RH and CH mice after leptin administration. Tau supplementation increased pSTAT3/STAT3 hypothalamic protein content in the CHT and RHT mice by 69 and 81 %, respectively (P < 0.02 and P < 0.002; Fig. 3b). No alteration in hypothalamic STAT3 protein expression was observed in the groups (Fig. 3c).

Fig. 3
figure 3

Food intake profile and hypothalamic STAT3 activation in response to leptin in mice that underwent protein-restriction with or without HFD and Tau supplementation. a Mean ± SEM (n = 4–6) of caloric intake after saline or leptin administration. After a 6 h period of fasting, mice of all groups received an i.p. injection of saline (0.9 % NaCl), and food intake was measured after 12 h. On another day, the same mice received an i.p. injection of leptin (500 ng/g BW). b pSTAT3/STAT3 protein expression in the hypothalamus of the C, CH, CHT, R, RH and RHT mice. Mice of both experimental groups received an i.p. injection of 0.9 % saline (−) or leptin (500 ng/g BW) (+). After 45 min, the hypothalamus was collected and used for immunoblotting experiments. c STAT3 and GAPDH (internal control) hypothalamic protein content. Bars represent the mean ± SEM of the values, determined by optical densitometry (n = 4–6 rats). Asterisk leptin is different from saline-treated mice (Student’s t test; P < 0.05)

Full size image

Discussion

Protein-restricted mice showed lower BW, body length, and total protein levels and a higher brain to BW ratio (Tables 1 and 2), which are typical features of a low-protein diet treatment (Amaral et al. 2010; Batista et al. 2012, 2013b). The HFD treatment effectively induced obesity in normal and restricted-protein mice by increasing BW, fat stores, and the plasma levels of CHOL and TG. In accordance with a previous report (Batista et al. 2013b), we also verified that Tau supplementation only prevented fat deposition in normal-protein HFD mice (Table 1).

Protein deficiency during early life may lead to obesity and glucose intolerance in adulthood (Hales and Barker 1992). Recently, our laboratory demonstrated that protein-restricted mice that are fed on a HFD developed obesity, hyperleptinemia, glucose intolerance, and insulin resistance. These features were associated with increased calorie intake (Batista et al. 2013b). Here, using the same rodent model, we aimed to extend these findings through a comprehensive evaluation of food intake, gene expression profile, and central leptin signaling in the hypothalamus. In addition, we investigated whether Tau supplementation modulated these processes, since central administration of this amino acid exerts anorexigenic effects (Solon et al. 2012).

Mice that were fed on a low-protein isocaloric diet showed increased food consumption, yet calorie intake was matched to normal-protein mice (Fig. 1e). This effect was associated with an increase in the expression of NPY and AgRP (Fig. 2; Table 4). In fact, a low-protein diet was reported to result in reduced activation of satiety centers, leading to increased food intake (Du et al. 2000). In the offspring of dams submitted to a protein-restricted diet, during gestation and lactation, hyperphagia associated with an increased expression of NPY and AgRP in the hypothalamus was described (De Souza et al. 2008; Orozco-Solis et al. 2011). In peripheral tissues, offspring of dams fed on a low-protein diet during gestation showed increased insulin sensitivity at 3 months of age (Latorraca et al. 1998); however, at 17 months, insulin resistance and diabetes were established (Petry et al. 2001). In addition, rats that underwent protein-restriction during lactation showed a lower hypothalamic STAT3 protein expression (Lisboa et al. 2012) and reduced leptin-induced anorexigenic effects (Passos et al. 2004). In contrast to this study, the adaptive gene expression of hypothalamic neuropeptides during protein restriction, observed in our study, was not due to impairment of hypothalamic leptin action, as leptin administration efficiently suppressed food consumption and increased pSTAT3/STAT3 protein expression in the R mice (Fig. 3a, b). Circulating leptin and its gene expression have been consistently reported to be unaltered in mice exposed to protein restriction during gestation or post-weaning (Bol et al. 2009; Batista et al. 2013b). However, we previously demonstrated that R mice had lower pIRS (phosphorylated insulin receptor substrate)-1 protein content in the hypothalamus (Camargo et al. 2013), which may contribute to central insulin resistance and increased orexigenic neuropeptide expression.

In a previous report (Batista et al. 2013b) and in this study, hypophagia was observed during the experimental period in both RH and CH mice, whereas increased caloric intake was verified during the last week of HFD treatment. It is known that HFD modifies hypothalamic neuropeptide expressions during the experimental period. Increased NPY mRNA content in the ARC was observed during the first 2 weeks in diet-induced obesity in rats, whereas at 12 weeks, these rats showed a decrease in NPY gene expression (Levin and Dunn-Meynell 2002). Here, we also observed a reduction in the expression of NPY and AgRP in the hypothalamus of RH and CH mice by microarray and PCR analysis (Fig. 2a, b). Levin and Dunn-Meynell (2002) suggested that an increase in orexigenic neuropeptides may contribute to obesity onset; however, the adaptive reduction of this neuropeptide with the prolongation of high energy intake may act to combat weight gain. Here, we observed increased mRNA content of CART and Pomc in the hypothalamus of normal-protein obese mice, whereas this response was blunted in protein-restricted mice fed on a HFD (Fig. 2c, d). A similar neuropeptide expression profile was previously observed in normal-protein rodents fed on a HFD (Wang et al. 1999; Ziotopoulou et al. 2000). Thus, the protein-restriction background may lead to differential adaptive responses to the HFD.

We also evidenced that the impairment in hypothalamic leptin signaling may be partially involved in caloric food intake and hypothalamic gene expression alterations in the RH and CH mice, since these groups of mice did not present any increase in pSTAT3/STAT3 protein expression (Fig. 3b) or inhibition of food intake (Fig. 3a) after leptin administration. Hyperleptinemia and impaired response to exogenous leptin are hallmarks of leptin resistance that are frequently described in HFD rodents (Münzberg et al. 2004; Moraes et al. 2009; Koch et al. 2014). Accordingly, leptin was elevated by thirteen- and eightfold in the plasma of CH and RH mice, respectively (Batista et al. 2013b). Taken together our data indicate the development of a full leptin resistance phenotype in both CH and RH mice.

A HFD was shown to increase the expression of inflammatory markers in the hypothalamus (De Souza et al. 2005). This pathway was involved in the desensitization of the hypothalamic leptin action (Romanatto et al. 2007; Amaral et al. 2006). In addition, it is important to mention that, in the hypothalamus of RH and CH mice, a decrease in pIRS-1/IRS-1 protein expression was shown (Camargo et al. 2013), which probably contributed to alterations in hypothalamic gene expressions and feeding behavior in these groups.

In a previous study, Tau showed anorexigenic actions in the hypothalamus by decreasing NPY gene expression and enhancing insulin action (Solon et al. 2012). Here, Tau supplementation had no effect upon the HFD-induced impairment of orexigenic neuropeptides (Table 4; Fig. 2). However, in RHT mice, the mRNA levels of the anorexigenic neuropeptides Pomc and AgRP were elevated (Fig. 2c, d). Furthermore, Tau reduced calorie consumption only in the CHT group (Fig. 1e), while it efficiently preserved the leptin hypothalamic action in both RHT and CHT mice, since both groups showed decreased food intake (Fig. 3a) and increased pSTAT3/STAT3 (Fig. 3b) ratio after leptin injection. Consistent with improved leptin signaling, the HFD-induced hyperleptinemia was previously shown to be completely prevented in CHT but, only partially in RHT mice (Batista et al. 2013b). Solon et al. (2012) reported that intracerebroventricular administration of Tau increased JAK2/STAT3 pathway activation. This amino acid also shows anti-inflammatory properties by decreasing cytokine levels in the brain and peripheral tissues (Lin et al. 2013; Li et al. 2013; Su et al. 2014). In addition, hypothalamic pAkt/Akt protein content was higher in CHT mice (Camargo et al. 2013). This evidence suggests that the preservation of leptin signaling in the hypothalamus partially controls food intake in RHT. As such, since Tau did not improve hypothalamic pAkt expression during basal conditions in this group (Camargo et al. 2013), this mechanism probably contributes to the persistent alteration in calories consumed by RHT mice.

The microarray analyses also demonstrated that the diet and/or Tau supplementation also alters the expression of mRNA related to TAC, glycolysis/gluconeogenesis, and oxidative phosphorylation (Table 5); unfolded protein response, apoptosis, and ER stress pathway (Table 6); and leptin and insulin signaling and inflammatory pathways (Table 7).

Our results showed that, in the R hypothalamus, there is a reduction in several genes involved in glycolysis/gluconeogenesis, TAC, and oxidative phosphorylation. Previously, modifications in TCA and mitochondrial oxidative phosphorylation genes were reported by microarray analysis in islets isolated from fetal protein-restricted rats (Reusens et al. 2008). In Zucker-fatty rats, the enhanced protein expression of several mitochondrial subunits of the respiratory chain in the hypothalamus leads to a neural glucose hypersensitivity, which impairs the redox signaling, promoting oxidative stress, and more hypothalamic damage (Colombani et al. 2009). In accordance with our results (Table 6), the HFD groups also had an increase in the expression of genes related to the unfolded protein response and ER stress. This pathway is involved in leptin desensitization in the hypothalamus, which may enhance NPY and AgRP mRNA levels, promoting hyperphagia (Ozcan et al. 2009).

Tau shows antioxidant properties and regulates mitochondrial function. Tau normalized the expression of genes involved in TCA, glycolysis, cellular defense, cellular growth and proliferation in fetal pancreatic islets from protein-restricted dams (Reusens et al. 2008). In PC12 cells, Tau decreased the ER stress activation induced by peroxide (Pan et al. 2010). Tau also reduced superoxide production in the mitochondria of cardiomyocytes by enhancing electron transport chain activity (Jong et al. 2012). Protein-restricted mice supplemented with Tau showed better redox balance and a reduction in ER stress markers such as p-PERK and BIP (Batista et al. 2013a, b; Cappelli et al. 2013). In accordance with all these studies, we herein report that Tau supplementation decreased the fold change in genes related to the unfolded protein response, apoptosis, and the ER stress pathway in the hypothalamus of the RHT and CHT mice (Tables 5, 6), which may contribute to better leptin signaling in these groups. Importantly, gene-profiling data was obtained in whole hypothalamus. Thus, further nuclei-specific analyses should provide more insight into Tau’s regulatory actions upon gene expression and feeding behavior. Because BBB cells and neurons express Tau transporter (Borg et al.1976; Huxtable and Lippincott 1982; Lee and Kang 2004), it is plausible that at least part of the effects of Tau supplementation described here occur due to a direct action upon the hypothalamic nuclei. In support of this notion, hypothalamic Tau concentration was dramatically elevated in supplemented mice.

In summary, food intake and expression of orexigenic neuropeptides in the hypothalamus were increased by protein restriction and decreased by the HFD, despite increased calorie intake. Tau supplementation exerted profound effects upon hypothalamic gene expression profile that were associated with the prevention of HFD-induced central leptin resistance. The effects of Tau supplementation upon the CNS should be further investigated and considered for the treatment of obesity.