Development of insulin resistance in mice lacking PGC-1α in adipose tissues

Generation of FKO Mice.

Mice with a specific deletion in adipose tissue PGC-1α (FKO mice) were generated by breeding animals harboring a floxed PGC-1α allele (33) with mice transgenically expressing a cre recombinase under the control of the adiponectin promoter (Fig. 1A) (34). The excision efficiency of the adiponectin-driven cre recombinase was monitored by measuring protein and mRNA levels of PGC-1α. Because PGC-1α protein levels are very low in unstimulated adipose tissue, we induced PGC-1α levels with cold exposure (27). PGC-1α protein was clearly induced and detectable in both BAT and subcutaneous inguinal WAT (IWAT) of floxed control mice on cold exposure, but PGC-1α protein was not detected in these adipose tissues of FKO mice (Fig. 1B). Similarly, the transcript levels of PGC-1α were strongly reduced in BAT, IWAT, and epididymal WAT (EWAT) of FKO mice housed at 21 °C (Fig. 1C). The ablation of PGC-1α was restricted to adipose tissue; no alterations of PGC-1α transcript levels were detected in other known PGC-1α–expressing tissues, such as liver and muscle (Fig. 1C). Importantly, the FKO exhibited no compensatory increase in PGC-1β expression. The FKO mice were born in the expected Mendelian ratios and showed no growth abnormalities.

Characterization of Adipose Tissue Lacking PGC-1α Expression.

Given that PGC-1α has been implicated in the regulation of thermogenic, mitochondrial, and metabolic gene programs in various tissues, we studied gene expression changes in FKO mice by real-time PCR. Differentiation-dependent adipose expression of such genes as PPARγ, AP2, and Adiponectin was not altered in any of the adipose depots analyzed (Fig. 2 A–C). However, transcript levels of the BAT-enriched genes Cidea and UCP1 were reduced in WAT depots, but not in BAT depots (Fig. 2 A–C). Mitochondrial gene expression was normal in BAT depots (Fig. 2A), but altered in WAT depots (Fig. 2 B and C). The greatest defects were observed in the subcutaneous IWAT, with 20–50% reductions in transcripts of OXPHOS genes, genes of the tricarboxylic acid (TCA) cycle, and FAO genes (Fig. 2B). Reduced mRNA expression of CPT1β, PDK4, and FABP3 was also observed in EWAT (Fig. 2C). Despite these changes in gene expression, the adipose tissues of FKO mice appeared morphologically normal compared with control mice (Fig. 2D).

To determine whether the observed gene expression changes are cell-autonomous, the stromal-vascular fraction (SVF) of inguinal adipose tissue was isolated and differentiated in vitro. Inguinal SVF from both FKO and control mice differentiated equally well and accumulated ample amounts of lipids, as indicated by oil red O staining (Fig. 3A). In agreement with the in vivo data, adipocytes differentiated from FKO mice in culture expressed lower levels of Cidea and UCP1 (Fig. 3B). Several transcripts of mitochondrial genes were also reduced by 20–40% in adipocytes derived from the SVF of the FKO mice compared with controls (Fig. 3B). To investigate whether the observed reduction in gene expression would impact cellular physiology, oxygen consumption was measured in inguinal adipocytes with an oxygen-sensitive Clark electrode. Loss of PGC-1α resulted in a significant reduction of uncoupled oxygen consumption rate and trend toward lower basal respiration (Fig. 3C). These data show that loss of PGC-1α impairs basal mitochondrial and thermogenic gene expression in the subcutaneous WAT in ambient temperature.

As expected, the response to a thermogenic challenge was blunted in the FKO mice. Exposure to cold caused core body temperature to drop slightly but statistically significantly faster in FKO mice than in Flox control mice (Fig. S1A). However, FKO mice could survive for longer periods in a cold environment. Thermogenic gene expression was reduced on cold exposure in all analyzed fat pads of FKO mice compared with control mice (Fig. S1 B–D). In addition, the acquisition of a BAT-like phenotype on prolonged cold exposure was strongly blunted in IWAT of FKO mice compared with control mice (Fig. S1E).

Glucose Intolerance and Insulin Resistance in FKO Mice Fed an HFD.

To study the effect of adipose PGC-1α deregulation on glucose homeostasis, FKO and control mice were fed either regular chow or an HFD, and metabolic parameters were analyzed. FKO mice fed regular chow did not differ significantly from one another in body weight, body composition, or blood glucose and insulin levels (Fig. S2 A–D). No changes in free fatty acid (FFA) or triglyceride (TG) levels were detected, and all animals were similarly glucose-tolerant (Fig. S2 E–G).

The control and FKO mice fed an HFD gained similar amounts of weight (Fig. 4A) and exhibited no changes in body composition as measured by MRI (Fig. 4B). However, changes in glucose tolerance were apparent when the ability of control and FKO mice to deal with an acute glucose load was tested via a glucose tolerance test after 8 wk on an HFD. The FKO mice were clearly less able to clear a glucose bolus, as indicated by the shift in the blood glucose curve (Fig. 4C). Insulin tolerance was subsequently tested to determine whether this glucose intolerance was associated with systemic insulin resistance. Insulin sensitivity did not differ between control and FKO mice after 9 wk on the HFD; however, the FKO mice had higher starting glucose levels (Fig. S2H). After 16 wk on the HFD, the capacity to reduce blood glucose levels after a bolus of insulin was significantly reduced in FKO mice compared with control mice (Fig. 4D). HFD feeding also resulted in higher circulating FFA (after 12 wk on the HFD) and TG (after 17 wk on the HFD) levels in FKO mice compared with control mice (Fig. 4 E and F). Serum insulin levels were similar in the FKO and control mice (Fig. 4G).

Increased Liver Insulin Resistance in HFD-Fed FKO Mice.

To further understand how whole-body glucose metabolism was disturbed in FKO mice, we applied the hyperinsulinemic-euglycemic clamp technique on mice fed the HFD for 8 wk. Clamp studies allow for very accurate determination of insulin-dependent peripheral glucose uptake and liver glucose output in vivo. Under hyperinsulinemic conditions, compared with control mice, the FKO mice required less glucose infusion to maintain constant blood glucose levels, indicating decreased insulin sensitivity (Fig. 5A and Table S1). We investigated whether this was caused by a decrease in peripheral glucose uptake or more hepatic glucose output. Although peripheral glucose uptake was decreased only marginally (Fig. 5B), there was a significant decrease in glucose uptake in BAT of FKO mice (Fig. 5C), indicating insulin resistance in this tissue. Interestingly, the clamp study revealed a significant decrease in the suppression of hepatic glucose output by insulin (Fig. 5D). These data suggest that the observed decrease in whole-body glucose tolerance and insulin sensitivity in FKO mice is driven mainly by a liver insulin resistance.

Impaired Lipid Metabolism in HFD-Fed FKO Mice.

The consumption of an HFD can increase nonshivering thermogenesis (35, 36). Recent data have shown that activation of thermogenesis reduces the amount of circulating lipids in mice and humans (37, 38). Given the elevated levels of circulating FFA and TG in the HFD-fed FKO mice, we analyzed UCP1 expression in WAT and BAT depots of these mice. UCP1 expression was reduced by 50% in IWAT of FKO mice compared with control mice, but did not differ in EWAT (Fig. 6A). However, BAT UCP1 mRNA levels were 10% lower in FKO mice (Fig. 6A), suggesting a decrease in the thermogenic program in FKO mice compared with controls. The reduced expression of OXPHOS and TCA genes observed in the IWAT of FKO mice fed a chow diet was not apparent under conditions of an HFD (Fig. S3). EWAT mitochondrial gene expression also did not differ between FKO and control mice under this condition (Fig. S3). Interestingly, with the challenge of an HFD, decreases in some mitochondrial genes, including Cox5b, Idh3α, and SOD2, were observed in BAT of FKO mice compared with controls (Fig. S3). Transcript levels of genes involved in lipid turnover were also strongly reduced in BAT of HFD-fed FKO mice compared with controls, with a 50% reduction in FABP3 expression (Fig. 6B) and a 30–40% reduction in the expression of genes involved in FFA and TG uptake and breakdown, including CD36, LPL, and LipA (Fig. 6B). Changes of FAO transcript levels in IWAT and EWAT were also observed; the mRNA level of PDK4 was decreased in IWAT and EWAT, and a trend toward lower CD36 and LPL mRNA levels was observed in IWAT of HFD-fed FKO mice compared with control mice (Fig. 6 C and D). FKO mice also exhibited decreased expression of genes involved in adipose lipid synthesis; mRNA levels of DGAT, LXR, and SCD1 were decreased in BAT (Fig. 6B), and mRNA levels of LXR, SCD1, and SREBP1 were down-regulated in EWAT (Fig. 6D). These data suggest that decreases transcript levels of UCP1 and genes involved in lipid uptake and synthesis lead to elevated circulating blood lipid levels (Fig. 4 E and F). We next asked whether the higher levels of circulating lipids resulted in increased hepatic steatosis in FKO mice. Liver TG and FFA levels trended to be higher in FKO mice compared with controls, but this difference did not reach statistical significance (Fig. 6 E and F). Thus, FKO mice fed an HFD have reduced expression of UCP1 and genes regulating lipid metabolism as well as increased circulating lipid levels, all of which might contribute to the liver insulin resistance observed in these animals.

Animals.

All mice were maintained at 24 °C on a 12-h:12-h light-dark cycle. FKO mice were generated by crossing female mice homozygous for the floxed allele with heterozygous floxed male mice carrying a transgene expressing cre-recombinase under control of the adiponectin promoter. Mice carrying the floxed allele as well as mice carrying the cre transgene were backcrossed 10 times to the C57BL/6CR background. Animals were fed a regular chow diet or an HFD (60% kcal fat; Research Diets). All experiments were performed in accordance with the animal facility’s institutional Animal Care and Use Committee regulations.

Hyperinsulinemic-Euglycemic Clamping.

Mice were kept on an HFD for 7 wk. At 6 d before the experiment, an indwelling catheter was implanted into the left jugular vein, and clamping was performed as described previously (15).

Glucose and Lipid Measurements.

Glucose was measured in tail blood using a standard glucometer. Liver TGs were extracted using chloroform/methanol (2:1) mix, dried in a fume hood overnight, and dissolved in a solution containing 60% butanol and 40% of the Triton-X114/methanol mix (2:1; vol/vol). TGs and nonsterified fatty acids were measured by colorimetric assays (from Sigma-Aldrich and Wako, respectively).

RNA Isolation and Real-Time PCR Analysis.

Total RNA from cultured cells or mouse tissues was isolated by TRIzol extraction and purification using Qiagen RNeasy mini columns according to the manufacturer's instructions. For quantitative real-time PCR analysis, 1–1.5 μg of total RNA was reverse-transcribed using the Applied Biosystems High-Capacity cDNA Reverse-Transcription Kit. SYBR Green reactions using SYBR Green PCR Master Mix were assembled along with 250 nM primers according to the manufacturer’s instructions and performed with an ABI Prism 7900HT sequence detection system (Applied Biosystems). Relative expression of mRNAs was determined after normalization to Rps18. The Student t test was used to evaluate statistical significance. All primer sequences are available from the authors on request.

Western Blot Analysis.

Whole-cell extracts were prepared from adipose tissues by homogenization in lysis buffer containing 50 mM Tris (pH 7.4), 500 mM NaCl, 1% Nonidet P-40, 20% glycerol, and 1 mM DTT, supplemented with protease inhibitor mixture (Roche Diagnostics). Lysates were resolved by SDS/PAGE, transferred to PVDF membrane (Millipore), and probed with indicated antibodies. Proteins were visualized using the Supersignal West Dura HRP Detection Kit (Pierce). The PGC-1α and UCP1 antibodies were obtained from Calbiochem (mAb 4C1.3) and Abcam (ab10983), respectively.

Immunohistochemistry.

For histological analysis, adipose tissue was fixed in 4% formaldehyde, embedded in paraffin, cut into 6-μm sections, and mounted on slides. Immunohistochemistry was performed using the Vectastain Elite ABC Kit (Vector Laboratories) according to the manufacturer’s instructions using rabbit UCP1 antibody (ab10983; 1:1,000 dilution; Abcam).

SV Culture and Primary Adipocyte Differentiation.

Isolation of adipose SV fraction and differentiation of primary adipocytes was done as described previously (50). Oil red O staining was performed as described previously (51).

Oxygen Consumption Assay.

Primary SV cultures isolated from IWAT of FKO and Flox mice were induced to undergo adipogenesis. At day 7 of differentiation, oxygen consumption was measured in fat cells using a Strathkelvin Clark-type electrode. Oligomycin (1 μM; Sigma-Aldrich) was added to block coupled respiration.

Statistical Analysis.

Statistical significance (set at P < 0.05) was assessed by the Student t test or ANOVA using GraphPad Prism, as indicated. Values are reported as mean ± SEM.