Nervous System Expression of PPARγ and Mutant PPARγ Has Profound Effects on Metabolic Regulation and Brain Development

Abstract

Peroxisome proliferator activated receptor (PPARγ) is a nuclear receptor transcription factor that regulates adipogenesis and energy homeostasis. Recent studies suggest PPARγ may mediate some of its metabolic effects through actions in the brain. We used a Cre-recombinase-dependent (using Nestin^Cre) conditionally activatable transgene expressing either wild-type (WT) or dominant-negative (P467L) PPARγ to examine mechanisms by which PPARγ in the nervous system controls energy balance. Inducible expression of PPARγ was evident throughout the brain. Expression of 2 PPARγ target genes, aP2 and CD36, was induced by WT but not P467L PPARγ in the brain. Surprisingly, Nes^Cre/PPARγ-WT mice exhibited severe microcephaly and brain malformation, suggesting that PPARγ can modulate brain development. On the contrary, Nes^Cre/PPARγ-P467L mice exhibited blunted weight gain to high-fat diet, which correlated with a decrease in lean mass and tissue masses, accompanied by elevated plasma GH, and depressed plasma IGF-1, indicative of GH resistance. There was no expression of the transgene in the pancreas but both fasting plasma glucose, and fed and fasted plasma insulin levels were markedly decreased. Nes^Cre/PPARγ-P467L mice fed either control diet or high-fat diet displayed impaired glucose tolerance yet exhibited increased sensitivity to exogenous insulin and increased insulin receptor signaling in white adipose tissue, liver, and skeletal muscle. These observations support the concept that alterations in PPARγ-driven mechanisms in the nervous system play a role in the regulation of growth and glucose metabolic homeostasis.

Peroxisome proliferator activated receptor (PPARγ) is a nuclear receptor transcription factor best recognized for its role in adipogenesis and lipid metabolism. Agonists of PPARγ, thiazolidinediones (TZDs), offer protection against type 2 diabetes and have been shown to reduce cardiovascular events such as stroke and to reduce cardiovascular risk in renal failure (1–5). Unfortunately, TZDs are no longer routinely prescribed because of adverse effects, including weight gain, fluid retention, bone fracture, and possibly neoplasms (6).

PPARγ is expressed in the central nervous system (CNS) (7). Recent studies suggest that PPARγ in the brain regulates feeding behavior; and based on this, PPARγ activation in the brain is thought to mediate at least some of the adverse effects of TZDs (8, 9). Administration of TZD into the brain or expression of a constitutively active PPARγ in the hypothalamus promotes weight gain whereas inhibiting PPARγ prevents weight gain (9). Similarly, ablating PPARγ in neurons caused a reduction in food intake and increased energy expenditure (EE) in mice fed a high-fat diet (HFD) (8). Consistent with the antioxidant properties of PPARγ, TZD administration reduced oxidative stress in the arcuate nucleus, which was associated with increased activity of orexigenic agouti-related protein/neuropeptide Y neurons and decreased activity of anorexigenic proopiomelanocortin (POMC) neurons (10). Cell-specific ablation of PPARγ in POMC neurons protected mice from diet-induced obesity (11). We recently reported that inducible expression of dominant-negative (DN) PPARγ specifically in POMC neurons can affect metabolic homeostasis under certain dietary conditions, although the effects of PPARγ interference were more subtle that complete ablation of PPARγ in POMC neurons (12).

Because of these subtle effects caused by DN PPARγ interference in POMC neurons, we considered the possibility that interfering with PPARγ throughout the CNS may have more profound effects on energy homeostasis. Thus, we generated 2 transgenic mouse models in which either wild-type (WT) human PPARγ (PPARγ-WT) or the DN mutant (P467L) form of human PPARγ is conditionally expressed in the nervous system via the neuronal Nestin promoter. Neuron-specific expression of PPARγ-WT caused widespread abnormalities in brain development and premature death. On the contrary, global interference with PPARγ function in the CNS leads to impaired growth, resistance to diet-induced obesity associated with decreased lean mass, and a redistribution of adipose tissue, GH resistance, and abnormalities in glucose and insulin. Interestingly, the mice are exquisitely sensitive to exogenous insulin.

Materials and Methods

Generation of conditional human PPARγ transgenic mice

Transgenic mice conditionally expressing a human DN mutant (P467L) form of PPARγ were generated at the University of Iowa Gene Editing Facility. CAG-PPARγ-P467L mice carry a transgene designed to express both PPARγ (P467L) and the tdTomato reporter gene after its selective activation by Cre-recombinase (12, 13). Neural overexpression of the mutant (P467L) form of PPARγ was achieved by crossing CAG-PPARγ-P467L-T mice with Nestin-Cre mice (B6.Cg-Tg(Nes-cre)1Kln/J; stock 003771), purchased from The Jackson Laboratory) (14). Nestin^Cre/PPARγ-P467L and Nestin^Cre/PPARγ-WT along with their littermate controls were used for the experiments described here. All procedures have been approved by the University of Iowa Institutional Animal Care and Use Committee. Mice were kept under standard laboratory conditions, with free access to food and water.

Western blotting for transgenic PPARγ protein

About 100 mg of each tissue (brain, liver, spleen, pancreas, adipose, or muscle) were homogenized in a lysis buffer containing 50 mmol/L Tris-HCl buffer, 0.1 mmol/L EDTA, 0.1 mmol/L EGTA (pH 7.5), and 0.1% vol/vol sodium dodecyl sulfate, with protease inhibitors (Thermo Scientific). Protein concentration was determined using the BCA assay (Pierce Protein Research Products). Protein lysates were separated by SDS-PAGE gel and transferred to polyvinylidene difluoride membranes (Immobilon-P; Millipore). Membranes were blocked for 30 minutes at room temperature with 5% BSA in Tris-buffered saline with 0.1% Tween 20 (TBS-T). PPARγ expression was determined using a monoclonal anti-PPARγ antibody (C26H12; Cell Signaling Technology) at a dilution of 1:1000 (in 5% BSA/TBS-T). Secondary antibodies for PPARγ was a donkey antirabbit IgG (NA934; GE Healthcare Life Sciences) used at 1:10 000 dilution. Specificity was assess by comigration with recombinant PPARγ expressed in HEK293 cells and by the absence of signal in tissues from nontransgenic mice. AKT and phospho-AKT (Ser473) expression was detected using anti-AKT (9272; Cell Signaling Technology) and anti-phospho-AKT (9271; Cell Signaling Technology) antibodies at a dilution of 1:2000 for anti-phospho-AKT (in 5% BSA/TBS-T) and 1:2000 (in 5% skim milk) for total AKT. The secondary antibody for AKT was a goat antirabbit IgG horseradish peroxidase-linked antibody (7074; Cell Signaling) diluted 1:10 000 in 5% milk/TBS. Rabbit anti-β-actin was used as a loading control at a dilution of 1:10 000 (ab8229; AbCam). Primary antibodies were diluted and incubated overnight at 4°C. Secondary antibodies were applied for 1 hour at room temperature. Blots were treated with Pierce SuperSignal Western Pico chemiluminescence reagent for visualization.

Quantitative real-time PCR analysis

Brain tissue (100 μg) was suspended in ice-cold TRIzol (Invitrogen). Total RNA was isolated using RNeasy spin columns (RNeasy Mini Plus kit; QIAGEN) following the manufacturer's instructions. The RNA concentration was determined using a NanoDrop ND-1000. cDNA was generated using SuperScript (Invitrogen). Quantitative real-time PCR was performed using the TaqMan (applied Biosystems) gene expression assay from 10.0-ng cDNA in a total volume of 10.0 μL following the manufacturer's recommendations. The assay numbers for TaqMan (Applied Biosystems) probes were: Mm99999915_g1 (mouse glyceraldehyde 3 phosphate dehydrogenase [GAPDH]) and human PPARγ (Hs01115513_m1). The expression of tdTomato (forward, 5′-CAC CAT CGT GGA ACA GTA CG-3′ and reverse, 5′-GCG CAT GAA CTC TTT GAT GA-3′) was determined using SYBR green (Bio-Rad) according to the manufacturer's protocol. Mouse GAPDH was used as an internal control.

Diets

All mice were fed either 60% HFD (D12492; Research Diets) or isocaloric-match control diet (D12450J; Research Diets) starting at 5 weeks of age for 15 weeks. Body weights from group-housed mice were measured weekly at 8–9 am until 20 weeks of age. All experiments were performed at 20 weeks of HFD, unless otherwise reported. Daily food intake was assessed in individually housed mice.

Measurement of body composition

Body composition was measured at the end of diet treatment period (25 wk) in vivo by Nuclear Magnetic Resonance (Bruker Minispec LF-90). Both male and female Nestin^Cre/PPARγ-P467L and control mice (on either HFD or control diet) were evaluated.

Measurement of EE

Male Nestin^Cre/PPARγ-P467L and control mice, 20 weeks of age, on HFD were individually housed for 1 week before measurements. Mice were placed into temperature-controlled (30°C), insulated chambers for estimation of heat production by respirometry (15). First, the CO₂ (model CD-3A; AEI) and O₂ analyzers (model S-3A/II; AEI) were calibrated to standardized air containing 5000 ppm CO₂ and 20.50% O₂ (Praxair), respectively. Mice were then placed into water-jacketed, temperature-controlled, and air-sealed chambers, and the change in effluent O₂ and CO₂ concentrations were recorded using PowerLab and Chart software (AD Instruments). Flow was determined by mass flow meters (EM1; Sensiron) to standard temperature and pressure-corrected flow values. Heat production was estimated using the equation based on Lusk (1928): heat = O₂ consumed (VO₂) [1.232 respiratory exchange ratio (RER) + 3.815], where VO₂ = [ΔO₂%] · [STP-corrected flow] and RER = ΔCO₂%/ΔO₂%.

Male Nestin^Cre/PPARγ-P467L mice also underwent metabolic parameters measurements over a 3-day period using the Oxymax Comprehensive Lab Animal Monitoring System (Columbus Instruments). Briefly, mice were acclimatized to the cages for 12 hours before measurements began. The volumes of VO₂ and CO₂ produced by each mouse were measured through indirect calorimetry. RER was calculated as the ratio of VO₂ into CO₂ produced. EE was calculated using the Lusk equation: 3.815 + 1.232 × RER. Activity levels were counted as infrared beam breaks along the x-axis of the cage.

Fasting glucose and glucose tolerance test (GTT) and insulin tolerance test (ITT)

Mice were individually housed and fasted overnight (16–18 h). Fasting blood glucose levels were measured via Accu-Chek Aviva Blood Glucose Meter (Roche) using tail blood. GTT and ITT were performed as described (16). Four- to five-month-old mice were fasted overnight (GTT) or for 5 hours (ITT) in a clean cage. Body weight and basal glucose levels were obtained before ip glucose (2.0 g/kg) or insulin (1 U/kg) administration. Blood glucose levels were determined at the following time points: 15, 30, 60, and 120 minutes using a glucometer (Accu-Chek Aviva Glucose Meter; Roche) from tail-blood.

Insulin receptor signaling

Male mice were fasted overnight in a clean cage. On the test morning, the mice were anesthetized with ketamine-xylazine, and the jugular vein was cannulated for either vehicle (PBS) or insulin (5 U/kg) injections. Fifteen minutes after injection, the mice were killed, and the liver, perigenital white adipose tissue (WAT), and soleus muscle were collected. The tissues were then processed for immunoblotting, and the levels of phospho-AKT and total AKT were determined as described above.

Plasma measurements

Blood was extracted from the left ventricle of Nestin^Cre/PPARγ-P467L and control mice at the time of killing (typically at 8 am). Plasma from each blood sample was isolated using 0.5M EDTA and by centrifugation at 7000 rpm for 5 minutes at 4°C. Plasma levels of leptin, insulin, thyroid hormone (free T₃), and GH were measured by commercially available ELISA kits (Bio-Rad for leptin, Crystal Chem for insulin, Cusabio for free T₃, and Cloudclone for GH).

Statistics

All data are expressed as mean ± SEM. The mean between 2 groups were analyzed by 2-tailed Student's t test. The means of 2 or more groups and 2 genotypes were analyzed by a two-way ANOVA with Bonferroni post hoc tests unless otherwise stated. Analysis of covariance (ANCOVA) was used to analyze the effect of both genotype and body weight on EE and RER. Significance was recorded at P < .05.

Results

We generated a transgene designed to inducibly express either WT or DN PPARγ in a Cre-recombinase-dependent manner (12). To evaluate the role of PPARγ in the nervous system, CAG-PPARγ-WT transgenic mice, designed to inducibly express WT PPARγ (PPARγ-WT) and the tdTomato reporter gene, and CAG-PPARγ-P467L transgenic mice, designed to inducibly express PPARγ-P467L and tdTomato were crossed with mice transgenic for Cre-recombinase under the control of the pan-neuronal Nestin promoter (14). Abundant expression of PPARγ-P467L and PPARγ-WT mRNA was detected in the brain of Nestin^Cre/PPARγ-P467L and Nestin^Cre/PPARγ-WT mice, respectively (Figure 1A). Similarly, there was abundant expression of tdTomato mRNA, the reporter gene integrated into both transgene constructs in the brain of transgenic mice (Figure 1B). Nestin^Cre/PPARγ-P467L mice exhibited increased expression of human PPARγ protein in the hypothalamus, cortex, hippocampus, and brain stem, whereas no human PPARγ was detected in littermates that expressed only Cre-recombinase (Figure 1C). There was no evidence of human PPARγ protein expression in the liver, pancreas, or sc adipose tissue (Figure 1D). Similarly, there was no detectable PPARγ-P467L mRNA expression in perigenital adipose tissue (cycle threshold value, 33.2 ± 0.2 in Nestin^Cre/PPARγ-P467L vs 33.1 ± 0.2 in control) and brown adipose tissue (BAT) (cycle threshold value, 30.3 ± 0.3 in Nestin^Cre/PPARγ-P467L vs 32.3 ± 0.2 in control). Expression of 2 canonical PPARγ target genes, aP2 and CD36, was significantly increased in the brain of Nestin^Cre/PPARγ-WT mice (Figure 1, E and F). Consistent with the transcriptional impairment of PPARγ-P467L, there was no induction of aP2 or CD36 expression in brain from Nestin^Cre/PPARγ-P467L mice. These data confirm the nervous system activation of the transgene.

Figure 1.

Transgene expression in Nestin^Cre/PPARγ-P467L mice. A and B, Expression of PPARγ (A) and tdTomato (B) mRNA in brain of Nestin^Cre/PPARγ-P467L and Nestin^Cre/PPARγ-WT mice expressed as fold induction vs nontransgenic littermate controls as detected by quantitative real-time RT-PCR (Q-PCR). Nestin^Cre/PPARγ-P467L (n = 3), Nestin^Cre/PPARγ-P467L control (n = 3), Nestin^Cre/PPARγ-WT (n = 5), Nestin^Cre/PPARγ-WT control (n = 4); *, P < .01 by ANOVA. C, Western blotting of PPARγ protein expression in brain regions of Nestin^Cre/PPARγ-P467L and Control mice (30-μg protein per lane). Note that only 1.5 μg of protein was loaded in the control lane (labeled C) which is from HEK293 cells overexpressing PPARγ. HY, hypothalamus; CX, cortex; HP, hippocampus; BS, brainstem. D. Western blotting of PPARγ protein expression in peripheral tissues of Nestin^Cre/PPARγ-P467L and Control mice (30-μg protein per lane). L, liver; P, pancreas; A, sc adipose tissue. E and F, Expression of aP2 (E) and CD36 (F) mRNA in brain of Nestin^Cre/PPARγ-P467L and Nestin^Cre/PPARγ-WT mice expressed as fold induction vs nontransgenic littermate controls as detected by Q-PCR. Nestin^Cre/PPARγ-P467L (n = 3), Nestin^Cre/PPARγ-P467L control (n = 3), Nestin^Cre/PPARγ-WT (n = 5), and Nestin^Cre/PPARγ-WT control (n = 4); *, P < .02, **, P < .01 by ANOVA. Q-PCR was TaqMan assays normalized to GAPDH. Data are mean ± SEM.

At birth, Nestin^Cre/PPARγ-WT mice are indistinguishable from the rest of the litter. However, between P7 and P10, they weigh 1.0–1.5 g less than their littermate control mice. By P21, the Nestin^Cre/PPARγ-WT mice are severely growth restricted (body weight, 4.5 ± 0.12 vs 8.4 ± 0.2 g in controls; P < .0001) and generally struggle to survive. This accounts for the lower numbers of Nestin^Cre/PPARγ-WT mice genotyped after weaning (15 of 108 weanlings, 13%). Nestin^Cre/PPARγ-WT mice exhibit severe microcephaly and abnormal brain structural development. Histological analysis of their brains reveals significant distortion of the brain architecture (Figure 2). The striatum and the hippocampi are severely hypoplastic and are replaced by prominent ventricular spaces with accentuated posterior horns. Thinning of the cortical layers accentuates the ventricular enlargement. Agenesis of corpus callosum is present uniformly in all samples examined by histology with cingulate cortex also absent (n = 10). The presence of white matter was confirmed with OsO₄ staining suggest the absence of a myelination defect (data not shown). The striatal and hippocampal hypoplasia seem to vary in severity, such that 2 distinct phenotypes can be distinguished. Of the sections examined from 10 mice, 6 had a severe phenotype with significant hippocampal and striatal hypoplasia and prominent colpocephaly. Four samples had the milder phenotype characterized with development of the striatum and absence of ventricular enlargement. These brain defects are likely to account for the early demise, which precluded us from studying the metabolic phenotype of these mice.

Figure 2.

Brain histology in Nestin^Cre/PPARγ-WT mice. Representative sections from a control brain (A, C, and E) and comparable sections from a Nestin^Cre/PPARγ-WT brain (B, D, and F) showing severe hypoplasia of major brain structures. B, Hypoplasia of the caudate-putamen (cp) and consequent ex vacuo enlargement of the lateral ventricles (lv) can be clearly seen. Note in B and D, agenesis of the corpus callosum (cc). Also in B and D, the abnormal position and hypoplasia of the hippocampus (hc) can be clearly seen. F, Abnormal development of cerebellar white matter can be seen on the left side of the section. On the right side, severe hypoplasia and dysplasia of the cerebellar cortex is clearly evident. Scale bar, 5 mm.

Ablation of neuronal PPARγ leads to resistance to HFD-induced weight gain (8). We therefore hypothesized that overexpression of the DN PPARγ-P467L will recapitulate the phenotype. At 5 weeks of age, male Nestin^Cre/PPARγ-P467L mice weighed less than littermate controls (15.1 ± 0.8 vs 17.6 ± 0.6 g, P = .024). The reduction in body weight of male Nestin^Cre/PPARγ-P467L mice was sustained over time and more evident in mice fed 60% HFD (Figure 3, A and B). At 20 weeks of age (after 15 wk on HFD), male Nestin^Cre/PPARγ-P467L mice exhibited significantly lower body weight compared with their littermate controls (Figure 4A). Interestingly, the reduced body weight in Nestin^Cre/PPARγ-P467L mice was not due to a decrease in fat mass (Figure 4B) or fluid mass (Figure 4C) but to a reduction in lean mass (Figure 4D). Consistent with this, there was a significant decrease in the weight of the liver, heart, kidney, and spleen (Figure 4, E–H). Although there was no difference in total fat mass, there were redistribution of adipose tissue to regional depots, including a decrease in BAT (Figure 4I), and an increase in perigenital fat pad (Figure 4J). There was no change in inguinal sc adipose tissue (Figure 4K). The redistribution of adipose tissue did not result in any change in the circulating levels of leptin (Figure 4L).

Figure 3.

Body weight and weight gain in Nestin^Cre/PPARγ-P467L mice. Body weight (A and C) and change in body weight (B and D) are plotted in male (A and B) and female (C and D) Nestin^Cre/PPARγ-P467L mice in response to 10% isocaloric control diet or 60% HFD for 15 weeks starting at 5 weeks of age. The number of mice per group is indicated. Data in the graphs are expressed as the mean ± SEM; *, P = .0005, interaction of weeks × genotype, two-way ANOVA.

Figure 4.

Body composition in Nestin^Cre/PPARγ-P467L mice. Body weight (A), body composition (B–D), organ weights (E–K), and plasma leptin (L) are shown for male Nestin^Cre/PPARγ-P467L mice fed 60% HFD for 15 weeks starting at 5 weeks of age. The number of mice per group are as indicated in Figure 3. Data in the graphs are expressed as the mean ± SEM; *, P < .0001; **, P = .003; ***, P = .01 by t test.

In contrast to male mice, there was no difference in body weight in female Nestin^Cre/PPARγ-P467L mice compared with controls in response to 60% HFD (Figure 3, C and D). Similarly, there was no difference in body composition in female Nestin^Cre/PPARγ-P467L mice (fat mass: 23.2 ± 2.3 vs 20.0 ± 2.4 g in controls, P = .37; lean mass, 14.8 ± 0.8 vs 15.8 ± 0.6 g in controls, P = .31; fluid mass, 6.1 ± 0.4 vs 5.5 ± 0.4 g in controls, P = .26, n ≥ 7/group).

The decreased lean mass of the male Nestin^Cre/PPARγ-P467L mice was not due to muscle atrophy. Accordingly, there was no change in the level of mRNA expression of molecular markers of muscle atrophy such as Gadd45, Fbxo32, Cdkn1a, and Smox (data not shown). Moreover, histological examination of tibialis anterior muscle revealed no difference between genotypes in muscle fiber morphology (data not shown).

Analysis of 24-hour food intake, averaged over 5 days, showed no difference between male Nestin^Cre/PPARγ-P467L and littermate control mice (Figure 5A). Heat production trended to be lower in the Nestin^Cre/PPARγ-P467L males, but there was no statistical difference in EE when analyzed with ANCOVA (Figure 5B). There was also no difference in EE when heat production was normalized by lean body mass (Figure 5C). Neuronal ablation of PPARγ has been shown to increase EE through up-regulation of thermogenic genes such as uncoupling protein 1 (UCP1) in BAT (8). Nestin^Cre/PPARγ-P467L male mice exhibited similar levels of expression of UCP1 mRNA in BAT as control mice (Figure 5D). This is consistent with similar levels of EE between the genotypes. Interestingly, there was a trend toward decreased activity in male Nestin^Cre/PPARγ-P467L compared with controls but this did not reach significance (beam breaks: light phase, 68 ± 18 in Nestin^Cre/PPARγ-P467L; n = 5 vs 100 ± 24 in controls; n = 6, P = .33; and dark phase, 130 ± 14 vs 257 ± 57 in controls; P = .13).

Figure 5.

Energy intake and output in Nestin^Cre/PPARγ-P467L mice. Food intake (n = 8 for control, n = 6 for Nestin^Cre/PPARγ-P467L) (A), heat production analyzed by ANCOVA (n = 17 for control, n = 10 for Nestin^Cre/PPARγ-P467L) (B), heat production normalized by lean body mass (n = 17 for control, n = 10 for Nestin^Cre/PPARγ-P467L) (C), and UCP1 mRNA measured by quantitative real-time RT-PCR (Q-PCR) (n = 6 per group) (D) is shown. Data in the graphs are expressed as the mean ± SEM.

We next measured several metabolic hormones. There was no difference in the level of free thyroid hormone (T₃) in Nestin^Cre/PPARγ-P467L mice compared with control littermates (Figure 6A). Interestingly, plasma GH level trended to be elevated in Nestin^Cre/PPARγ-P467L mice, but this failed to reach significance (Figure 6B). Nevertheless, the trend toward an elevation suggests the mice may have GH resistance. Consistent with this, there was a significant decrease in plasma levels of IGF-1 (Figure 6C).

Figure 6.

Endocrine measures in Nestin^Cre/PPARγ-P467L mice. Plasma T₃ (n = 6) (A), GH (n = 5) (B), and IGF-1 (n = 7) (C). Data in the graphs are expressed as the mean ± SEM; *, P = .03.

Brain PPARγ has been shown to play a critical role in hepatic insulin sensitivity, and reduced IGF-1 may have implications for glucose homeostasis (8, 17). We therefore measured the levels of fasting glucose in Nestin^Cre/PPARγ-P467L mice after 15 weeks of HFD. Nestin^Cre/PPARγ-P467L mice exhibited significantly lower fasting glucose levels compared with controls (Figure 7A). Nestin^Cre/PPARγ-P467L mice were markedly hypoinsulinemic compared with controls in both the fed and fasting states (Figure 7B). This is unlikely due to expression of the transgene in insulin producing β-cells, because there was no evidence for human PPARγ or tdTomato expression in pancreas at the mRNA (Figure 7C) or protein levels (Figure 2D). We next performed GTT and ITT to better understand whether there was altered insulin sensitivity in these mice. We first measured glucose tolerance and insulin sensitivity in 10% fat control diet-fed male Nestin^Cre/PPARγ-P467L mice. Control diet-fed male Nestin^Cre/PPARγ-P467L mice were glucose intolerant (Figure 7D). Interestingly, the experiment assessing insulin sensitivity of control diet-fed Nestin^Cre/PPARγ-P467L mice had to be terminated as the mice became distressed from hypoglycemia. 60% HFD-fed Nestin^Cre/PPARγ-P467L male mice were glucose intolerant compared with their littermate controls at 60 minutes after infusion but ultimately cleared glucose by 120 minutes (Figure 7E). HFD-fed Nestin^Cre/PPARγ-P467L male mice demonstrated significantly enhanced response to exogenous insulin compared with controls (Figure 7F). Consistent with this, insulin receptor signaling was augmented in the WAT, liver, and muscle of Nestin^Cre/PPARγ-P467L mice fed HFD (Figure 8).

Figure 7.

Glucose homeostasis in Nestin^Cre/PPARγ-P467L mice. Fasting plasma glucose (*, P < .002; n = 11 control; n = 8 Nestin^Cre/PPARγ-P467L) (A), fasting and fed plasma insulin (*, P < .0001; **, P < .05, n = 6–7 per group) (B), and pancreatic expression of human PPARγ and tdTomato mRNA (n = 3 per group) (C) in Nestin^Cre/PPARγ-P467L and control mice. Intraperitoneal GTT in mice fed 10% fat control (D) and 60% HFD (E) is shown; *, P < .05 vs control mice; n = 5 per group. Intraperitoneal ITT in mice fed HFD (F) is shown; **, P = .0001; n = 6 per group. Data in the graphs are expressed as the mean ± SEM.

Figure 8.

Insulin signaling in Nestin^Cre/PPARγ-P467L mice. A, Representative Western blotting measuring phosphorylated AKT and total AKT in perigenital WAT, liver, and muscle is shown in response to a single injection of insulin (5 U/kg iv) or saline as vehicle. Tissues were collected 15 minutes after injection. B, Quantification of Western blottings (n = 4 samples per group). *, P < .05 insulin vs vehicle; **, P < .05 insulin-treated Nestin^Cre/PPARγ-P467L vs insulin-treated and control mice.

Discussion

A role for neuronal PPARγ in the regulation of energy homeostasis has recently been suggested. Whole-brain ablation of PPARγ results in protection against diet-induced obesity through mechanisms involving increased EE and decreased energy intake (8). POMC neuron-specific deletion of PPARγ recapitulates the whole-brain phenotype in inducing resistance to HFD and elucidates additional mechanisms underlying these protective effects, ie, increase in the levels of reactive oxygen species, which are required for the proper functioning of the anorexigenic POMC neurons (11). Expression of DN mutation of PPARγ specifically in POMC neurons has subtle effects under some dietary conditions (12).

Null mutations in PPARγ lead to embryonic lethality due to impaired placental development (18). In contrast, even though rare, DN mutations in PPARγ have been reported to cause severe metabolic disturbances and early onset hypertension in humans (17, 19). One such mutation (P467L) occurs in the ligand-binding domain of PPARγ and has been shown to interfere with the WT form of the receptor by competing for target gene promoter binding sites and to be transcriptionally defective (17, 20). We previously demonstrated the utility of using the PPARγ-P467L mutant as a potent repressor of experimentally validated PPARγ target genes, which otherwise show increased expression levels in response to PPARγ agonists (21, 22). We show here that just as in the blood vessel, the DN PPARγ-P467L mutation is defective in mediating transactivation of PPARγ target genes in the brain. These same PPARγ targets are induced when PPARγ-WT is expressed in neurons. Administration of rosiglitazone induces weight gain when PPARγ-WT is expressed selectively in POMC neurons (12).

Deletion of PPARγ in brain has been reported to result in significantly lower body weight compared with control mice at a young age and when fed regular chow diet (8). We observed the same in the Nestin^Cre/PPARγ-P467L at baseline, before treatment with HFD, and also during feeding with low-fat isocaloric control diet. The differences in body mass reported in the brain PPARγ knockouts disappear by 13 weeks (8), whereas in our model, it persisted through adulthood. The divergence in body weight was most pronounced under conditions of HFD feeding but was evident even with low-fat diet, suggesting growth impairment. Consistent with this, the weight of several organs was decreased in Nestin^Cre/PPARγ-P467L mice. Moreover, there was no difference in EE or energy intake. It is unlikely that the reduced growth in these mice is due to pituitary impairment, which was reported in some Nestin-CRE driver mice (23), because our model did not have reduced GH but rather a trend for increased GH. Instead of a state of GH deficiency, the marked decrease in IGF-1 is an indicator of GH resistance (24). Another contributor to poor growth in these mice may be related to low insulin levels. Abnormalities in the growth of children with insulin deficiency due to type1 diabetes has been recognized for many decades (25). Moreover, the role of insulin as a growth factor has been well established (26), because impaired growth is a significant complication of uncontrolled diabetes (27, 28). Infusion of insulin in the streptozotocin model of type 1 diabetes in rats, particularly when administered in the hepatic circulation, can stimulate (or reverse) growth restriction (29). Alternatively, the poor growth in these mice may be related to GH resistance. For example, Laron syndrome in humans is an abnormality of growth caused by resistance to GH (30). Hypoglycemia, low IGF-1 and elevated GH levels are hallmarks of the syndrome. The Nestin^Cre/PPARγ-P467L mice phenocopy at least some aspects of Laron syndrome and phenocopy models of Laron syndrome in mice and pigs (31, 32).

Neuronal ablation of PPARγ leads to impaired glucose tolerance and reduced insulin sensitivity (8). Like those mice, Nestin^Cre/PPARγ-P467L mice exhibited impaired glucose tolerance under control and HFD-fed states. However, unexpectedly, fasting glucose was significantly lower in the Nestin^Cre/PPARγ-P467L after 15 weeks of HFD than in controls. Furthermore, Nestin^Cre/PPARγ-P467L mice exhibited a marked sensitivity to exogenous insulin in an ITT, and a marked increase in insulin receptor signaling in adipose, liver, and muscle. This enhanced insulin sensitivity may explain the low circulating glucose and hypoinsulinemia. It should be noted, however, that our model stands in contrast with the insulin resistance phenotype of humans carrying DN PPARγ mutations. These patients are characterized by severe fasting hyperinsulinemia and hyperglycemia consistent with severe insulin resistance (17, 19). Of course, the mutations in humans affect every tissue in the body, whereas in our model, the interference with PPARγ is neuron specific.

The hypothalamus regulates energy and glucose balance by maintaining reciprocal neural communications with the preganglionic motor neurons of the autonomic nervous system (33). It has been recently shown that abnormal insulin signaling in POMC neurons, in response to HFD exposure, leads to altered POMC projections to the preautonomic paraventricular nucleus of the hypothalamus and thus impaired pancreatic parasympathetic innervation, demonstrated by decreased glucose-stimulated insulin secretion (34). Activation of PPARγ in POMC neurons has been shown to lead to decreased activity pattern of these neurons, which is abolished by treatment with PPARγ antagonist (10). Thus, we cannot formally rule out that the hypoinsulinemia observed in our model is due to impaired parasympathetic output to the pancreas as a result of altered input to parasympathetic preganglionic neurons.

Appendix

See Table 1.

Table 1.

Antibody Table

Peptide/ Protein Target	Antigen Sequence (if Known)	Name of Antibody	Manufacturer, Catalog Number, and/or Name of Individual Providing the Antibody	Species Raised in; Monoclonal or Polyclonal	Dilution Used	RRID
PPARG	Synthetic peptide corresponding to residues surrounding Asp69 of human PPARγ	PPARG	Cell Signaling, C26H12	Rabbit; monoclonal, 2435	1:1000	AB_2166051
Actin		Actin	AbCam, ab8229	Goat; polyclonal	1:10 000	AB_306374
Phospho-AKT	Synthetic phosphopeptide corresponding to residues surrounding Ser473 of mouse akt	P-AKT	Cell Signaling, 9271	Rabbit; polyclonal	1:2000	AB_329825
Total-AKT	Synthetic peptide corresponding to the C-terminal sequence of mouse Akt	AKT	Cell Signaling, 9272	Rabbit; polyclonal	1:2000	AB_329827
Antirabbit IgG		Antirabbit IgG	Cell Signaling, 7074V	Goat, antirabbit, IgG	1:10 000	AB_11178390
Antirabbit IgG		Antirabbit IgG	GE Healthcare Life Sciences, NA934	Donkey; antirabbit, IgG	1:10 000	AB_772206