The Effect of β-Adrenergic and Peroxisome Proliferator–Activated Receptor-γ Stimulation on Target Genes Related to Lipid Metabolism in Human Subcutaneous Adipose Tissue

Nondiabetic patients were enrolled in a clinical trial performed in Baton Rouge, Louisiana, at the Pennington Biomedical Research Center (14 men and 43 women between 18 and 50 years of age with BMI 30–37 kg/m²). The characteristics of the study population were previously reported (20). Subjects were healthy and not taking thiazolidinediones, β-blockers, orlistat, sibutramine, ephedrine, phenylpropanolamine (Dexatrim), corticosteroids, statins, fibrates, cholesterol-binding drugs, or herbal supplements containing ephedrine and/or caffeine; abusing alcohol; or using other illicit drugs. Subjects were randomized into four groups: 1) placebo/placebo (PP), 2) ephedrine HCl plus caffeine/placebo (ECP), 3) placebo/pioglitazone (PPio), and 4) ephedrine HCl plus caffeine/pioglitazone (ECPio). The placebo, the pharmaceutical ephedrine HCl (25 mg) (Breathe Easy; Contract Pharmacal Corporation, Hauppauge, NY) and caffeine (200 mg) (Contract Pharmacal Corporation) were dosed as follows: one of each pill per day for 7 days at breakfast, increased to one of each pill at breakfast and lunch for the next 7 days, and then increased to one of each pill at breakfast, lunch, and dinner for the remainder of the 16-week protocol. Subjects who could not tolerate three doses of EC (ephedrine HCl plus caffeine) per day were allowed to continue on two doses per day. If subjects could not tolerate two doses per day they were dropped from the protocol. Medications were taken at least 4 h apart to reduce side effects (tremor and tachycardia). Pioglitazone was initiated at 15 mg/day and increased by 15 mg each week until the maximum dose of 45 mg was achieved. All subjects signed a consent form approved by the Pennington Biomedical Research Center Ethical Review Board after potential risks and procedures had been explained. At treatment initiation, all subjects were given a short standardized instruction on diet and behavior modification as well as instructions on healthy activity levels (walking). The primary end point was the percentage of fat measured with dual-energy X-ray absorptiometry after 16 weeks of treatment using a Hologic QDR 4500 DEXA.

Adipose tissue was obtained by Bergstrom needle biopsies from subcutaneous depots 6–10 cm lateral to the umbilicus at baseline (0 week) and after 12 weeks of treatment following an overnight fast. Fat samples were cleaned of visible connective tissue and blood vessels, immediately frozen, and stored at −70°C until assays.

Body fat mass and lean mass were measured on a Hologic Dual Energy X-ray Absorptiometer (QDR 4500A; Hologic, Waltham, MA). Visceral adiposity was determined by multislice computed tomography (CT) scanning using a General Electric High-Speed Computed Tomography scanner under an established protocol (21).

Total RNA from 50 to 100 mg adipose tissue was isolated with Trizol reagent (Invitrogen, Carlsbad, CA) and purified with RNeasy columns (QIAGEN, Valencia, CA) according to manufacturers’ procedure. The quantity and quality of the RNA were confirmed by Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA). TNA (total nucleic acid) was extracted from a separate sample of adipose tissue (20 mg) by isopropanol precipitation using MasterPure Complete DNA and RNA Pruification kit (EPICENTRE, Madison, WI). The total DNA recovered was determined by spectrophotometry.

Real-time RT-PCR was performed using a Taqman 100Rxn PCR Core Reagent Kit (Applied Biosystems, Roche, Branchburg, NJ) as previously described (15,18,22). Real-time RT-PCR was carried out in an ABI PRISM 7900 sequence detector (Applied Biosystems, Branchburg, NJ) using the following parameters: 1 cycle of 48° C for 30 min and 95° C for 10 min followed by 40 cycles at 95° C for 15 s and 60° C for 1 min. A standard curve was prepared by serial dilution of pooled total RNA and each gene/sample was compared with this standard curve. All expression data were normalized by dividing the amount of target gene by the amount of cyclophilin B applied as an internal control. To determine the tissue content of mitochondria per cell, real-time PCR for mitochondrial DNA was applied as described previously (22).

All values are presented in figures and tables as raw means ± SE unless otherwise noted as SD. Significant differences were assumed for P < 0.05. Gene expression data were analyzed by paired t test within treatment group using GraphPad Prism version 4.0 (GraphPad Software, San Diego, CA). Serial lipid data was analyzed using a mixed model (PROC MIXED) in SAS (Cary, NC). Pearson's correlations were performed in JMP version 3.0 (SAS, Cary, NC).

The characteristics of the overweight/obese study population are presented in supplemental Table 1 (found in an online-only appendix at http://dx.doi.org/10.2337/dc06-1962). Generally, the treatment groups were well balanced except for the baseline triglycerides, which were by chance higher in the PPio patients. One subject in the PPio and ECP groups and three subjects in the ECPio group were able to tolerate only two EC doses per day. The clinical parameters of the study population expressed as percent of change after 16 weeks of treatment with ephedrine, caffeine, and/or pioglitazone are presented in Fig. 1. We observed a 6.6% weight loss in the ECP group and 5.2% weight decrease after combined intervention with ECPio (Fig. 1A). The body weight of volunteers treated with pioglitazone alone or placebo did not change (Fig. 1A). Thus, the major treatment effects on body weight and adiposity were due to the EC treatment; the addition of pioglitazone did not increase body weight loss as seen in the preclinical rodent models. The pattern of change in percent fat paralleled the observed changes in body weight (Fig. 1B). Body fat decreased by 6.0% in the ECP group and 4.6% in the ECPio group, while body fat remained unchanged in the PPio and PP groups (Fig. 1B). Visceral and subcutaneous fat measured by CT decreased in proportion to the reduction in weight and body fat (Figs. 1C and D).

After combination therapy, triglycerides decreased by 41 mg/dl in the ECPio group, whereas treatment with EC alone or pioglitazone alone decreased the levels by 15.2 and 24 mg/dl, respectively (Table 1). HDL cholesterol increased by 7.8 mg/dl in the ECPio group, 3.7 mg/dl in the ECP group, and 5.1 mg/dl in the PPio group. Thus, the reduction in triglyceride levels can be largely attributed to the pioglitazone treatment, with no statistically significant impact from adding EC.

Pioglitazone treatment markedly increased mRNA expression for PEPCK1 (P = 0.01) (Fig. 2) but did not affect LPL (Fig. 2A), fatty acid transporter (FATP/CD36) (Fig. 2B), fatty acid synthase (data not shown), or stearoyl-CoA desaturase-1 (SCD-1) (Fig. 2D) mRNA levels when compared with the placebo-treated subjects. Pioglitazone treatment also increased carnitine palmitoyltransferase-1 (CPT-1, P < 0.01) (Fig. 3A), medium-chain acyl CoAdehydrogenase (MCAD, P < 0.05) (Fig. 3B), and malonyl-CoA decarboxylase (MLYCD) (P < 0.01) (Fig. 3C) but did not change PPAR-α (Fig. 3D) or cytochrome C (data not shown) mRNA levels. Treatment with pioglitazone alone did not change the expression of genes involved in mitochondrial biogenesis including PGC-1α, mtTFA, NRF-1, ERRα, AMPKα, or SIRTs 1–3 (data not shown). The upregulation of these three lipid oxidation genes (CPT01, MCAD, and MYLCD) suggests that the adipocytes might be reprogrammed for greater fat oxidation when stimulated by β-agonists.

The treatment of healthy subjects with the β-agonist ECP decreased only SCD-1 mRNA expression (P < 0.001) (Fig. 2D) but did not change the expression of other genes involved in lipid metabolism (Fig. 2).

The combined administration of these three drugs (ECPio) resulted in a higher mRNA expression for FATP/CD36 (P = 0.03) (Fig. 2B). Interestingly, only the combined treatments (ECPio) increased expression of SIRT1 and CD36 (both P = 0.03) (Figs. 2B and E). SIRT1 has been implicated in energy homeostasis and was recently described as a regulator of fat oxidation in adipocytes (23). The administration of pioglitazone or EC alone did not change SIRT1 mRNA expression (Fig. 2E). Importantly, the effect of pioglitazone to increase CPT-1, MCAD, and MYLCD was maintained in the combination (CPT-1 [P < 0.05, Fig. 3A], MCAD [P < 0.05, Fig. 3B], and MLYCD [P < 0.05, Fig. 3C]). The decrease of SCD-1 produced by EC was slightly blunted by the combination (P = 0.0004 for EC, P = 0.07 for ECP), suggesting one potential opposing action of the combination. Additionally, we did not observe any changes in mitochondrial copy number (mtDNA) between groups (Fig. 3D). There was no significant relationship between baseline mRNA for SIRT and fasting triglycerides. Similarly, the correlation between the change in SIRT1 and the change in triglycerides was not significant.

Given that the SIRT gene family has been implicated in the regulation of genes involved in fat oxidation, we explored the relationships between the expression of genes in the fat oxidation pathway and SIRT1 expression. Baseline mRNA for SIRT1 was positively correlated with PPAR-α (r = 0.65, P < 0.001) (Fig. 4A), MCAD (r = 0.61, P < 0.001) (Fig. 4B), LPL (r = 0.69, P < 0.001) (Fig. 4C), CD36 (r = 0.58, P < 0.001) (Fig. 4D), TFAM (r = 0.77, P < 0.001) (Fig. 4E), PGC-1α (r = 0.36, P < 0.001) (Fig. 4F), and ERRα (r = 0.35, P < 0.01) (data not shown) but not with cytochrome C (r = 0.22, data not shown), CPT-1 (r = 0.21, data not shown), or CAP (r = 0.25, data not shown). The mRNA expression for the mitochondrial SIRTs (SIRT3 and SIRT5) showed much lower correlations with TFAM (r = 0.32 and 0.54, respectively, P < 0.01; data not shown), LPL (r = 0.31 and 0.48, P < 0.01; data not shown), PPAR-α (r = 0.51 and 0.58, P < 0.01; data not shown), and MCAD (r = 0.43 and 0.44, P < 0.01; data not shown) (supplemental Table 2 of the online-only appendix).

Preclinical data suggest that combined PPAR-γ and β-adrenergic therapy synergize to increase oxidative capacity in adipose tissue. This study was performed to determine whether the combined pharmacological intervention with pioglitazone (PPAR-γ) and EC (β-adrenergic) would have a beneficial effect on body weight, lipids, and the expression of genes involved in lipolysis/lipogenesis, mitochondrial biogenesis, and oxidative metabolism in subcutaneous fat. It has been previously suggested that the existence of cross-talk between these two distinct pathways in the WAT in rodents produced weight loss beyond that seen with either agent alone (16). In the present study, the addition of pioglitazone to EC (ECPio group) had no additional effect on the loss of body weight. It should be noted that treatment pioglitazone alone did not change body weight. This is slightly surprising and contrasts with the commonly held view that activation of the PPAR-γ system will cause obligatory fat gain (21). In obese hypertensive subjects without diabetes, pioglitazone did not increase body weight (24).

Several previous studies have observed a decrease in visceral adipose tissue (VAT) with no changes in subcutaneous adipose tissue (SAT) in pioglitazone-treated patients, consistent with our results in the PP-treated patients. The reason total fat did not decrease could be due to one of two explanations. First, VAT could change without a change in whole-body SAT. A change in VAT of 13% for a depot that is only ∼4 kg represents only a 500-g change. One possibility is that dual-energy X-ray absorptiometry is not sensitive enough to detect this change. Second, VAT could decrease and SAT could increase by an equivalent amount. Note that the SAT measure by CT represents only abdominal SAT and VAT. An increase in gluteal femoral SAT or arm/leg SAT that matched the decrease in VAT would result in a “no change” result for SAT.

It has been proposed that the beneficial effect of thiazolidinediones on plasma lipids and glucose concentration in type 2 diabetic subjects is due to several mechanisms within white adipocytes and involves higher LPL leading to increased fatty acid transport into the adipocyte, increased fat catabolism in the cell, increased lipogenesis, and mitochondrial biogenesis (18,22,25). The unexpected finding of unchanged body weight after PPAR-γ activation in the present study could be partly explained by the fact that our subjects were healthy and did not have diabetes. Activation of PPAR-γ receptors may exert different effects on lipid metabolism in subcutaneous fat depots, depending on the diabetes status. Our subjects did not change body weight after pioglitazone treatment but significantly reduced their blood triglyceride levels without an increase in mRNA for LPL or FATP/CD36. The increased expression of several genes involved in β-oxidation (CPT-1, MCAD, and MLYCD) is consistent with our previous experiments in vivo and in vitro (18,22) and suggests that increased lipid oxidation and not sequestration/storage might be important for the improved triglycerides. Taken together, these results suggest that the benefit of ECP combination treatment is that weight loss effects of EC treatment are retained along with the hypotriglyceridemic actions of pioglitazone and a suggestion of a somewhat greater effect of the combination on triglycerides and HDL cholesterol. Pioglitazone increased CPT-I, MCAD, and MYLCD, whereas EC, a β-agonist, is a known activator of lipolysis in adipocytes. This suggests a potential for synergy between EC and pioglitazone for lipid oxidation: reprogramming of adipocytes toward oxidation by pioglitazone and activation of lipolysis by EC providing fatty acids substrate toward oxidation. One unexpected finding was the increase in SIRT1. Furthermore, since this was unique to the ECPio group, there may be signaling or transcriptional synergy underlying this result. We speculate that the ability of pioglitazone to activate the transcription of lipid-oxidizing genes, which we have previously shown are increased in diabetic subjects in vivo and confirmed in vitro, combined with the known effects of the β-adrenergic system to activate fat oxidation might work together to lower triglycerides. Larger studies of this combination in subjects with elevated triglycerides and low HDL are clearly warranted.

One unique feature of the combination treatment is the increase in CD36 and SIRT1 mRNA. This is in contrast to ECP alone or pioglitazone alone, which had no effect on the expression of these two genes. It is possible that these genes represent a larger set of genes that are specifically upregulated by the combination of PPAR-γ activation and cAMP activation. In contrast to the preclinical rodent models, this synergistic effect on gene expression did not translate into differences in weight loss. No other candidate genes measured in this study were uniquely regulated by the combination treatment.

In the present study, the activation of β-adrenergic receptors reduced body weight and blood triglycerides, but surprisingly did not change mRNA expression of several tested genes involved in lipid catabolism or mitochondrial biogenesis, as previously demonstrated in numerous rodent or in vitro studies (11,12,15). Activation of cAMP-dependent protein kinase pathway increases LPL expression during the differentiation of adipocytes but may decrease LPL activity through posttranscriptional modifications. Therefore, the lack of changes in LPL mRNA levels in our study may suggest that LPL activity was not changed; however, LPL activity was not measured. Furthermore, in the present study we did not note alterations in CPT-1 and CD36 mRNA expression. Regulation of lipid oxidation is certainly more complicated than simply changes in gene expression, as posttranslational mechanisms, such as allosteric regulation and translocation to the cell surface, play a critical role in determining final activities (26,27); thus, this result must be interpreted with caution. EC (ECP) reduced mRNA expression of SCD-1, whereas pioglitazone did not. Combined therapy with pioglitazone and β-adrenergic stimulation (ECPio group) tended to decrease SCD-1 mRNA expression. This effect was predominant with EC treatment but was not seen with pioglitazone treatment. SCD-1 is the rate-limiting enzyme in the desaturation of saturated fatty acids to monounsaturated fatty acids and targets lipids to TAG synthesis. We found a significant correlation between mRNAs for SCD-1 and fatty acid synthease at baseline (r = 0.51; P < 0.001; data not shown). Reduced SCD-1 activates pathways promoting fatty acid oxidation and decreases triglyceride synthesis in fat and muscle (28).

Our study showed that combination therapy with pioglitazone and EC significantly increased the expression of SIRT1 transcript level, whereas the administration of these treatments separately did not affect gene expression. SIRT1 is the closest mammalian ortholog of SIR 2 (silencing information regulator 2) that augments lifespan in lower organisms and mammalian cells (29–31) in response to caloric restriction. SIRT1 regulates fat mobilization in white adipocytes (23). We noted a strong relationship between baseline mRNAs for SIRT1 and genes involved in lipid uptake/oxidation (LPL, FATP/CD36, PPAR-α, and MCAD) and mitochondrial biogenesis (TFA and PGC-1α).

To summarize, these results indicate that combined therapy with PPAR-γ and β-adrenergic stimulators, representing two distinct intracellular signaling pathways, has a beneficial effect on body weight, plasma triglycerides, and lipid metabolism in subcutaneous fat depots through the downregulation of genes triggering fat accumulation (SCD-1) and upregulating genes required for fatty acid catabolism (SIRT1, CPT-1, MCAD, and MLYCD).

This work was supported by an unrestricted grant from Takeda Pharmaceuticals (to S.R.S.) and National Institutes of Health Grant DKO74772 (to T.W.G.).

The authors acknowledge Matt Hulver, Michele McNeil, David Hymel, Heather Loggins, and Leisel Hurder for advice and technical assistance and the volunteers who participated in this study.