Structural and Functional Consequences of Mitochondrial Biogenesis in Human Adipocytes in Vitro

Introduction: Mitochondrial biogenesis is a complex process, and several factors and signaling pathways regulate this process in muscle or brown adipocytes. The aim of the study was to explore pathways affecting mitochondrial biogenesis and fatty acid oxidation (FAO) in human white adipocytes.

Methods: Human preadipocytes obtained from liposuction samples were differentiated in vitro. On the 10th day of differentiation, 4 μm forskolin and 1 μm peroxisome proliferator-activated receptor-γ (PPARγ) agonist (pioglitazone, rosiglitazone, or GW 929) or 10 μm PPARα agonist (WY-14,643) were added to the media for 96 h. Quantitative real-time PCR was used to determine gene expression/mitochondrial copy number and ¹⁴C-labeled palmitate to measure direct energy dissipation.

Results: The treatment of adipocytes with forskolin increased mitochondrial copy number and the expression of genes involved in mitochondrial biogenesis (PPARγ coactivator 1α and transcriptional factor A) and fatty acid oxidation (PPARα and medium-chain acyl-coenzyme A dehydrogenase). The end (CO₂) and intermediate products (¹⁴C-labeled acid-soluble products) of FAO were also increased after forskolin treatment. PPARγ and PPARα agonists increased mitochondrial copy number, uncoupling protein 1, medium-chain acyl-coenzyme A dehydrogenase, and carnitine palmitoyltransferase 1, but did not change PPARα, PPARγ coactivator 1α, or transcriptional factor A mRNA levels. FAO was higher after rosiglitazone, GW 929, and WY-14,643 but not after pioglitazone treatment.

Conclusions: Pharmacological activation of the cAMP or PPARγ pathway pushes the white adipocyte down the oxidative continuum. The direct energy-dissipating effects could be significant tools to treat obesity and to improve insulin resistance in type 2 diabetic patients by reduction of fat accumulation in adipocytes or by reprogramming fatty acid metabolism.

MITOCHONDRIA ARE THE organelles where several important cellular functions occur including amino acid biosynthesis and fatty acid oxidation (FAO). The mitochondrial DNA (mtDNA) is a small DNA molecule of about 16.5 kb, present in multiple copies in an individual mitochondrion, encoding several proteins involved in oxidative phosphorylation (for review see Refs.1 and 2). Mitochondrial number and function are changed in response to external stimuli and diseases. In patients with insulin resistance and type 2 diabetes and even in those with positive family history for diabetes, mitochondrial metabolism and ATP synthesis are reduced in concert with a reduction of key factors regulating mitochondrial biogenesis (2–6).

Initially, attention was focused on tissues with high oxidative capacity such as skeletal muscle. For example, Petersen et al. (7) showed a 40% reduction in muscle mitochondrial oxidative phosphorylation activity in the elderly that was associated with insulin resistance. Our recent results showed a reduction of mitochondrial copy number (mtDNA) in adipose tissue from diabetic volunteers when compared with healthy controls (3). In the same study, we observed that in vivo treatment with pioglitazone (PIO), a peroxisome proliferator-activated receptor-γ (PPARγ) agonist and commonly used insulin-sensitizing drug, restored diminished mtDNA content to levels observed in nondiabetic controls (3).

A number of reports indicate that PPARγ coactivator 1α (PGC-1α), highly expressed in brown adipocytes, heart, and skeletal muscle, is a key factor driving mitochondrial biogenesis (8–10). PGC-1α expression is up-regulated after cold exposure or β-adrenergic stimulation in brown adipose tissue (BAT) and muscle (11, 12) and increased during exercise in muscle (13). More recent experiments revealed a marked induction of PGC-1α mRNA in the liver during fasting, implying a stimulatory role of PGC-1α in gluconeogenesis and fatty acid oxidation (14, 15).

Although PGC-1α expression is low in white adipose tissue (WAT) compared with muscle and brown fat, the literature suggests a role for PGC-1α in mitochondrial biogenesis and oxidative metabolism in WAT. Semple et al. (16) reported 3-fold lower mRNA level of PGC-1α in sc fat of morbidly obese people. We recently reported that PGC-1α mRNA expression is significantly reduced in WAT obtained from type 2 diabetic subjects compared with nondiabetic controls, and activation of the PPARγ by PIO reversed the reduced level of PGC-1α mRNA (3). Ectopic expression of PGC-1α in rodent white adipose cells increased the cellular content of mtDNA and altered the adipocyte phenotype toward an oxidative phenotype including an increase in uncoupling protein 1 (UCP-1) (17). In another study, hyperleptinemia decreased the amount of WAT and increased β-oxidation, PPARα, and PGC-1α mRNA in WAT (18, 19).

PGC-1α interacts with and coactivates many nuclear (20–22) and nonnuclear receptors such as nuclear respiratory factors (NRF)-1 and -2 (10) and mitochondrial transcription factor A (mtTFA), which is directly responsible for mtDNA replication and transcription.

Based on our in vivo results (3) we hypothesized that PPARγ agonists, some of them commonly used in treatment of type 2 diabetes, might up-regulate mitochondrial biogenesis and FAO in vitro. Additionally, in our current study, we describe the effect of a stimulator of cAMP activity (forskolin) and a PPARα agonist (WY-14,643) on the expression of genes regulating mitochondrial biogenesis and FAO. For this purpose we measured mtDNA copy number and the expression of genes regulating mitochondrial biogenesis (PGC-1α and mtTFA) and FAO [PPARα, medium-chain acyl-coenzyme A dehydrogenase (MCAD), and carnitine palmitoylotransferase 1 (CPT-1)] in human adipocytes cultured in vitro. Furthermore, we assessed the impact of activation of these pathways on the oxidation of fatty acids. The direct energy-dissipating effects of these pathways could be significant tools to treat obesity and to improve insulin resistance in type 2 diabetic patients by reduction of fat accumulation in adipocytes or by reprogramming fatty acid metabolism.

Materials and Methods

Cell culture

Human WAT was obtained from liposuction samples (mixture of abdomen, hips, flanks, thighs, and gynecomastia) from five subjects (three women and two men; mean age, 41.8 ± 11.9 yr; body mass index, 27.9 ± 8.2 kg/m²; range, 22.5–39.1 kg/m²). The participants were healthy according to clinical examination and laboratory tests and did not take any medications. All adipose tissue specimens were obtained under a protocol approved by the Institutional Review Board of the Pennington Biomedical Research Center with informed consent of the patients.

The method for preadipocyte isolation was based on a modification of previously published procedures (23, 24). Briefly, 50–100 ml of tissue was subjected to 0.1% collagenase type I digestion with 1% BSA in PBS. Preadipocytes were grown to 80–90% confluence in stromal medium (DMEM/F-10) supplemented with fetal bovine serum and antibiotics. After 48–72 h, the cells were resuspended at a concentration of 3 × 10⁴ cells/cm². On the second day, the stromal medium was changed to the differentiation medium (DMEM/F-10, 3% FBS, 33 μm biotin, 17 μm pentothenate, 1 μm PPARγ agonist rosiglitazone (ROSI), 100 nm insulin, 1 μm dexamethasone, 0.2 mm 3-isobutyl-1-methylxanthine, 1× penicillin/streptomycin/fungizone) for 3 d. Cells were than changed to adipocyte medium (the same as differentiation medium but without ROSI and 3-isobutyl-1-methylxanthine), and the medium was changed every 3rd day. On the 10th day of differentiation, forskolin (4 μm) (EMD Biosciences, Calbiochem, San Diego, CA), PIO (1 μm) (PPARγ agonist; Takeda Pharmaceuticals, Lincolnshire, IL) alone or in combination with T0070907 (1 μm) (PPARγ antagonist; Calbiochem), ROSI (1 μm) (PPARγ agonist; Cayman Chemical, Ann Arbor, MI), GW 929 (100 nm) (PPARγ agonist; Calbiochem), or WY-14,643 (10 μm) (PPARα agonist; Alexis, San Diego, CA) were added to the media for 96 h (media and treatments were changed every 24 h). PPARγ agonists and antagonist were prepared in dimethylsulfoxide, and the remaining compounds were diluted in medium. Dimethylsulfoxide or medium was added to control cells (without treatments) to obtain the same concentration as in compound-treated conditions.

RNA and DNA extraction

Total RNA from cultured adipocytes was isolated with Trizol reagent (Invitrogen, Carlsbad, CA), purified with RNeasy columns (QIAGEN, Valencia, CA) and quantified spectrophotometrically. DNA was extracted from the same cells, after degradation of protein and RNA with Trizol, by phenol-chloroform extraction and ethanol precipitation according to the manufacturer’s procedure. The total amount of DNA recovered was determined by spectrophotometry.

Real-time RT-PCR for RNA

Real-time RT-PCR for tested genes was performed using a TaqMan 100Rxn PCR core reagent kit (Applied Biosystems, Roche, Branchburg, NJ) as described previously (3, 25). The sequences of primers, probes, and accession numbers for each gene are listed in Table 1. All expression data were normalized by dividing the amount of target by the amount of cyclophilin B present as an internal control.

TABLE 1.

Oligonucleotide sequences for primer/probe sets used in TaqMan analysis

Gene	Accession no.	Forward primer	Probe	Reverse primer
ND 1		CCCTAAAACCCGCCACATCT	CCATCACCCTCTACATCACCGCCC	GAGCGATGGTGAGAGCTAAGGT
PGC-1α	NM_013261	CACCAAACCCACAGAGAACA	CGCAGTCACAACACTTACAAGCCAAAC	GGGTCATTTGGTGACTCTGG
mtTFA	NM_003201	CCCAGATGCAAAAACTACAGAACTAA	TCCAACGCTGGGCAATTCTTCTA	TCCGCCCTATAAGCATCTTGA
Cyt c	NM_001916	TGGCCCCTCCCATCTACAC	AGAGTTTGACGATGGCACCCCAGCTA	ATCCTTGGCTATCTGGGACATG
NRF-1	NM_005011	CGTTGCCCAAGTGAATTATTCTG	TTGTTCCACCTCTCCATCAGCCA	CCCTGTAACGTGGCCCAAT
PPARα	NM_005036	GCTTTGGCTTTACGGAATACCA	AGCCATCTGAGCCAGGACAGCTTCCTAA	TGAAAGCGTGTCCGTGATGA
CD36	NM_001001548	AGTCACTGCGACATGATTAATGGT	CAGATGCAGCCTCATTTCCACCTTTTG	CTGCAATACCTGGCTTTTCTCA
MCAD	AF251043	TGCCCTGGAAAGGAAAACTTT	TGTAGAGCACCAAGCAATATCATTTATG	GTTCAACTTTCATTGCCATTTCAG
CPT-1	D87812	CCAGAGCAGCACCCCAAT	CATCTGCTACAGGGCCAAAGCCACCT	CTGCAATCATGTAGGAAACTCCATAG
UCP-1	NM_021833	ACGACACGGTCCAGGAGTTC	TCACCGCAGGGAAAGAAACAGCACC	ACCAGCTAAAATCTTGCTTCCTAAAC
CyclophilinB	XM_042251	GGAGATGGCACAGGAGGAAA	CATCTACGGTGAGCGCTTCCCCG	CGTAGTGCTTCAGTTTGAAGTTCTCA

ND 1, NADH dehydrogenase subunit 1.

Real-time PCR for mtDNA

Relative amounts of nuclear DNA (nDNA) and mtDNA were determined by quantitative real-time PCR as described previously (3). The ratio of mtDNA to nDNA reflects the tissue concentration of mitochondria per cell.

FAO assay

Fat oxidation was measured in differentiated adipocytes after 48 h (three subjects) or 96 h (two subjects) of culture with specific treatments (described above) performed in duplicate. Cells were preincubated with a glucose- and serum-free medium for 90 min, followed by a 4-h incubation with [1-¹⁴C]palmitate (NEN Life Science Products, Boston, MA) (2 μCi/ml) and 10 μm nonlabeled (cold) palmitate. Palmitate was coupled to fatty acid-free BSA in a molar ratio 5:1. After incubation, ¹⁴CO₂ (end product of FAO) and ¹⁴C-labeled acid-soluble products (ASPs; water-soluble intermediate metabolites of FAO) were measured according to Muoio et al. (26), with some modifications (27). Briefly, assayed medium was transferred into a 48-well trapping plate. The plate was clamped and sealed, and perchloric acid was injected through a rubber diaphragm into the medium, driving CO₂ through a tunnel into an adjacent well where it was trapped in 1 n NaOH. After trapping, medium was spun twice for the measurement of [¹⁴C]ASPs. Aliquots of NaOH and medium were transferred into scintillation vials, and radioactivity was measured on a multipurpose scintillation counter (model LS 6500; Beckman Instruments, Fullerton, CA). After assay, cells were placed on ice, washed twice with ice-cold Dulbecco’s PBS, and harvested into 0.05% SDS lysis buffer for subsequent protein measurement. Assays were performed in duplicate, and data were normalized to protein content. Absolute rates of FAO were adjusted for specific activity to account for the dilution of [¹⁴C]palmitate with unlabeled palmitate.

Glycerol concentration assay

The concentration of glycerol in media after 48 h of culture of differentiated adipocytes with tested treatments was performed in duplicate using a microanalysis analyzer CMA 600 (CMA Microdialysis AB, Solna, Sweden). Linearity for glycerol assay was 2–500 μmol/liter and adjusted for cellular protein content determined by bicinchoninic acid protein assay (Pierce, Rockford, IL).

Statistical analyses

Results from in vitro experiments are expressed as means ± se. Given differences between donors, data are presented as a percentage ± se of appropriate control value (set as 100%). Statistical analyses were performed using Statistica (version 6; StatSoft Inc., Tulsa, OK). Significant differences were established by one-way ANOVA with least significant differences post hoc and assumed as significant if P ≤ 0.05.

Results

Expression of genes related to mitochondrial biogenesis and FAO

To examine the pathways regulating mitochondrial biogenesis and FAO in WAT, we examined an in vitro model of differentiated human adipocytes treated 96 h with a cAMP/protein kinase A (PKA) activator, forskolin or agonists of PPARγ and PPARα.

Effect of forskolin

To determine whether forskolin affected mitochondrial biogenesis we measured mtDNA copy number and the expression of genes regulating mitochondrial biogenesis. Four days of treatment of differentiated adipocytes with forskolin significantly increased the mitochondrial copy number (∼44% increase; P < 0.01) (Fig 1D) and the expression of genes involved in mitochondrial biogenesis: PGC-1α (2-fold increase; P < 0.0001) (Fig. 1A), mtTFA (50% increase; P < 0.0001) (Fig. 1B), and cytochrome c (Cyt c) (2.5-fold increase; P < 0.0001) (Fig. 1C). Forskolin did not change expression of NRF-1, believed to be involved in mitochondrial biogenesis (data not shown). Additionally, forskolin increased a marker of brown adipocytes, UCP-1 mRNA expression, 6-fold (Fig. 1E). The incubation of adipocytes with forskolin caused a an up-regulation of mRNA expression for PPARα (P < 0.0001) (Fig. 2A) and MCAD (P ≤ 0.05) (Fig. 2B), key factors driving FAO in several tissues, but not CPT-1, the enzyme catalyzing entry of fatty acids into the mitochondrial matrix (Fig. 1C).

Fig. 1.

Effects of forskolin (F) (4 μm); PIO (1 μm), ROSI (1 μm), and GW 929 (GW) (100 nm), PPARγ agonists; WY-14,643 (WY) (10 μm), PPARα agonist; T0070907 (T) (1 μm), PPARγ antagonist; and T0070907 plus PIO (T+PIO) on mRNA expression of PGC-1α (A), mtTFA (B), Cyt c (C), mtDNA copy number (D), and UCP-1 (E). Differentiated in vitro human adipocytes were cultured for 96 h without (control) or with specific treatments. Media and treatments were changed every day. RNA and DNA were extracted, and quantitative real-time PCR was applied to determine the levels of various mRNAs or mtDNA copy number. Data represent the mean percentage expression levels of each treatment relative to controls. Results were collected from 10–12 replicate experiments in five to six donors. Conditions with similar letters are not significantly different.

Fig. 2.

Effects of forskolin (F) (4 μm); PIO (1 μm), ROSI (1 μm), and GW 929 (GW) (100 nm), PPARγ agonists; WY-14,643 (WY) (10 μm), PPARα agonist; T0070907 (T) (1 μm), PPARγ antagonist; and T0070907 plus PIO (T+PIO) on mRNA expression of genes involved in fatty acid transport and oxidation: PPARα (A), MCAD (B), CPT-1 (C), and CD36 (D). Differentiated in vitro human adipocytes were cultured for 96 h without (control) or with specific treatments. Media and treatments were changed every day. RNA and DNA were extracted, and quantitative real-time PCR was applied to determine the levels of various mRNAs or mtDNA copy number. Data represent the mean percentage expression levels of each treatment relative to controls. Results were collected from 10–12 replicate experiments in five to six donors (for CD36 from five replicates in two donors). Conditions with similar letters are not significantly different.

Effect of PPAR agonists

PPARγ as well as PPARα ligands significantly increased mtDNA copy number (Fig. 1D) and mRNA for UCP-1 (7- to 40-fold increase) (Fig. 1E). The addition of the PPARγ antagonist T0070907 (28) significantly reduced the stimulatory effect of PIO on mtDNA copy number (Fig. 1D). This antagonist given alone did not change mtDNA compared with control (Fig. 2D). Unexpectedly, PPARγ agonists did not change mRNA expression for PGC-1α, a key factor involved in mitochondrial biogenesis (Fig. 1A), or mitochondrial transcriptional factor A (Fig. 1B). Cyt c mRNA expression was higher after ROSI treatment (P < 0.05) (Fig. 1C) and tended to increase after PIO, GW 929, and WY-14,643 (P = 0.081) (Fig. 1C). The presence of the PPARγ antagonist T0070907 (28) in the culture, alone or in combination with PIO, significantly reduced mRNA levels of TFA (P < 0.001) (Fig. 1B) and Cyt c (P < 0.05) (Fig. 1C) but not PGC-1α (Fig. 1A).

The treatment of adipocytes with PPARγ (PIO, ROSI, and GW 929) or PPARα (WY-14,643) agonists did not significantly change mRNA expression for PPARα (Fig. 2A). The culture of adipocytes with the specific PPARγ antagonist T0070907 alone or in combination with PIO also had no effect on PPARα mRNA level. However, all of the factors (PPARγ as well PPARα agonists) markedly and significantly increased mRNA expression for MCAD (Fig. 2B) and CPT-1 (Fig. 2C). The classic PPARα target gene, CPT-1 mRNA, increased about 4-fold with all tested agonists of PPARγ (P < 0.0001) (Fig. 2C) and about 2.5-fold after PPARα activation (P < 0.05) (Fig. 2C). The stimulatory effect of the PPARα agonist on CPT-1 and MCAD mRNA levels was significantly smaller than PPARγ agonists. Furthermore, the addition of the specific PPARγ antagonist T0070907 in combination with PIO abolished the stimulatory effect of PIO on MCAD and CPT-1 mRNA expression (Fig. 2, B and C, respectively). Interestingly, culture of the cells with PPARγ antagonist alone also diminished the expression of the MCAD and CPT-1 compared with control. We have chosen MCAD as a marker of mitochondrial FAO capacity to maintain consistency with previously published data (3).

FAO

The end (CO₂) and intermediate products (ASPs) of FAO were measured after 48 and 96 h of adipocyte culture with forskolin, PPARγ (PIO, ROSI, and GW 929) or PPARα (WY-14,643) agonists. Forskolin markedly increased the level of ASPs after 96 h (3-fold increase) (Fig. 3, A and B) as well as CO₂ at the two tested time points (2- to 3-fold increase) (Fig. 3, C and D). PPARγ as well PPARα agonists generally did not change ASPs levels after 48 h of incubation but slightly (45–100%) increased after 96 h. The CO₂ levels were clearly higher (64–152%) after ROSI, GW 929, and WY-14,643 but not after PIO treatment. This was consistent across the four replicates, each from a different donor. We observed similar patterns in CO₂ release after 48 and 96 h of treatment (Fig. 3, A and B).

Fig. 3.

Effect of forskolin (F) (4 μm); PIO (1 μm), ROSI (1 μm), and GW 929 (GW) (100 nm), PPARγ agonists; and WY-14,643 (WY) (10 μm), PPARα agonist, on ¹⁴C-labeled ASPs (water-soluble intermediate metabolites) and ¹⁴CO₂ end products of FAO determined in human differentiated adipocytes after 48 and 96 h of culture with specific treatments. Data represent the mean percent fat oxidation of each treatment relative to controls. Results were collected from four to six replicate experiments in two to three donors.

Glycerol concentration

Glycerol levels were measured after 48 h of treatment of differentiated adipocytes. Addition of forskolin caused a statistically significant, approximately 4-fold increase in glycerol released to the media (Fig. 4). In contrast, culture of adipocytes with PPAR agonists (γ or α) did not change glycerol levels relative to control (Fig. 4).

Fig. 4.

Glycerol concentration determined in media after 48-h culture of human differentiated adipocytes with forskolin (F) (4 μm); PPARγ agonists PIO (1 μm), ROSI (1 μm), and GW 929 (GW) (100 nm); and PPARα agonist WY-14,643 (WY) (10 μm). Results were collected from four replicates. Conditions with similar letters are not significantly different.

Discussion

Adipocytes are classically divided into two distinct categories: white lipid-storing adipocytes and brown lipid-oxidizing adipocytes. Earlier in vivo studies from our laboratory suggest that the distinction between white and brown adipocytes is less clear than originally presumed (3), because activation of the PPARγ pathway in vivo increased mitochondrial number and genes involved in fat oxidation. These changes were concomitant with a clinically significant reduction in blood triglyceride levels. In the absence of histological localization, those earlier studies could not distinguish the emergence/differentiation of quiescent brown adipocytes in the white fat depot from the true remodeling/conversion of a white adipocyte into a more oxidative tan phenotype. Our present studies were undertaken to better understand pathways regulating mitochondrial biogenesis and FAO in human differentiated adipocytes.

We found that forskolin, a well-known activator of the cAMP/PKA pathway, increased mitochondrial copy number measured by real-time PCR. Forskolin also increased the transcription of mitochondrial biogenesis genes (PGC-1α and mtTFA) and nuclear encoded mitochondrial gene Cyt c. PGC-1α is a positive regulator of mitochondrial biogenesis, and its expression is increased after cold exposure, hyperphagia, or adrenergic stimulators affecting the cAMP/PKA pathway in brown adipocytes and muscle (29, 30). PGC-1α can be activated by at least two ways, first by cAMP/PKA cascade activating PGC-1α gene expression and second by activation of the p38 MAPK and/or cAMP/p38 MAPK cascade leading to direct phosphorylation of the PGC-1α protein itself (8). In the present study, we found that forskolin treatment significantly increased expression of the majority of tested genes involved in mitochondrial biogenesis including PGC-1α, mtTFA, and Cyt c. UCP-1, a specific marker and key regulator of thermogenesis in BAT, was also increased after forskolin treatment. Many reports suggest that the expression and activity of UCP-1 in BAT is predominantly regulated by the sympathetic system via stimulation of β-adrenergic receptors (31). The increase in UCP-1 expression in our experiments was accompanied by an increase in FAO. Our results showed that forskolin induces the expression of several genes regulating mitochondrial FAO in WAT as observed in other previously described tissues (31). We observed higher expression of PPARα and MCAD but not CPT-1 after forskolin treatment. Additionally, we found markedly higher FAO measured after 48 and 96 h.

We observed a 3.5-fold higher level of glycerol released from adipocytes cultured in the presence of forskolin but no effect of forskolin on expression of the fatty acid transporter FAT/CD36. This indicates that forskolin affects triglyceride stores rather than the import of fatty acids from the extracellular space. Summarizing, the effect of forskolin on mitochondrial biogenesis and FAO in human differentiated adipocytes is strong and similar to that previously reported in rodent brown adipocytes and muscle (9, 32).

In contrast to forskolin, which exhibited a direct and robust stimulatory effect on mitochondrial biogenesis/FAO pathways, the effect of PPAR agonist on mitochondrial biogenesis and FAO is subtly different in the present study. Our results showed that all three tested PPARγ agonists significantly increased mtDNA but had a smaller effect on Cyt c and no effect on PGC-1α and mtTFA, factors implicated in mitochondrial replication. In contrast to forskolin, the treatment of cells with a PPARγ antagonist alone (28) or in combination with PIO did not change PGC-1α mRNA expression but significantly diminished TFA and Cyt c mRNA levels. These in vitro results are only partially consistent with our recently published data in vivo showing that long-term activation of PPARγ by PIO in type 2 diabetic patients significantly increased mtDNA copy number and the expression of genes involved in mitochondrial biogenesis (PGC-1α and mtTFA but not NRF-1) (3). PIO also had a stimulatory effect on the expression of genes involved in fatty acid transport and β-oxidation; those are classically thought of as PPARα targets (3). PIO did not change energy expenditure or substrate oxidations as measured by indirect calorimetry (33). One interpretation of our in vitro results is the possibility of reprogramming of white fat cells into an intermediate fat-oxidizing phenotype, i.e. tan or brownish adipocyte. Furthermore, adipose tissue contributes to whole-body energy at rest, albeit only minimally compared with the lean tissue. This may explain the differences between our results and whole-body energy expenditure and fat metabolism. It has been reported that PPARγ activators [thiazolidinediones (TZDs) and non-TZD agonists] increase mitochondrial biogenesis, FAO, and white adipocyte remodeling in rodents (31, 34–36). Recently, Wilson-Fritch et al. (35) demonstrated a stimulatory effect of ROSI on mitochondrial biogenesis and remodeling of WAT with concomitant 30% increase of FAO and oxygen consumption in adipocytes isolated from mice. They observed a 2-fold increase in mitochondrial mass and a higher expression of mitochondrial proteins involved in fatty acid transport and oxidation (35); the mechanisms responsible for these described processes remain to be explored. Although Wilson-Fritch et al. (35) predicted only that PGC-1α could play a role in mitochondrial biogenesis, they did not report the gene expression or protein level for this key regulator. In another study performed in rodents, PPARγ activation up-regulated UCP-1 and other thermogenic genes (e.g. PGC-1α) mRNA expression but without an increase in whole-body energy expenditure (31). Higher energy expenditure after non-TZD PPARγ agonist treatment occurred only after β-adrenergic stimulation (31).

In the present study, the higher mtDNA copy number and increased expression of genes involved in fatty acid transport/β-oxidation had a similar pattern for three tested PPARγ agonists. However, these PPARγ agonists did not affect FAO in the same way as determined by the levels of end and intermediate products. We found a clear increase (about 100%) in ASPs and CO₂ production determined after 48 or 96 h of incubation for ROSI and GW 929 but not for PIO (all are PPARγ agonists). The addition of PIO did not change CO₂ levels after 48 or 96 h, although ASP concentration was about 50% higher after 96 h. This observation suggests that although activation of PPARγ receptors by PIO and ROSI cause similar changes in body weight and fat remodeling, they have divergent actions on energy dissipation; this may depend upon coactivator recruitment (37, 38). Interestingly, specific activation of PPARγ by ROSI and GW increased levels of intermediate and end products of FAO but did not change PPARα mRNA expression, as we found in vivo. We can only predict that activity of PPARα, a key factor driving FAO in several tissues, could be changed in our study not by TZDs but by higher expression of lipolysis products. It has been recently shown that lipoprotein lipase generates PPARα ligands in endothelial cells (39). Several reports indicate a stimulatory effect of PPARγ inducers on lipoprotein lipase activity in other tissues (25). The question of why PIO increases expression of genes driving the transport of fatty acids into mitochondria (CPT-1) and the first step of FAO (MCAD), but has no effect on FAO, remains unresolved. Another important question is why PIO and ROSI, both TZDs/PPARγ compounds, differentially regulate FAO.

PPARγ activation increased the expression of mRNA for UCP-1 in agreement with our previous in vivo results and many others (31, 35, 40, 41). A functional PPAR response element has been found within the UCP-1 promoter (42), and PPARγ agonists promote differentiation in BAT as well as in WAT. UCP-1 is induced in both tissues by TZDs in rodents. The cells expressing UCP-1 were described as newly created brown adipocytes. This is not possible in our fully differentiated adipocytes, suggesting plasticity of white adipocytes into tan adipocytes. The enhanced expression of this brown-specific gene UCP-1 was also observed after exogenous insertion of PGC-1α into human adipocytes (41). In our study we demonstrated higher mtDNA copy number and UCP-1 mRNA level, but PGC-1α mRNA did not change. The possible explanation for higher mtDNA copy number and no effect on PGC-1α mRNA expression could be involvement of additional PPARγ ligand-dependent cascades regulating nuclear-encoded mitochondrial genes or directly phosphorylating PGC-1α protein, not gene expression (for review see Ref.8). The limitation of the study was measurement of only gene expression and FAO products. We did not determine protein content.

The studies presented herein demonstrate, in fully differentiated adipocytes, that pharmacological activation of the cAMP or PPARγ pathway pushes the white adipocyte down the oxidative continuum. This is consonant with earlier studies that genetically modify adipocytes to alter the oxidative phenotype in similar cells (41, 43). The importance of this observation lies in the clinic, where the pharmacological activation of these pathways is known to improve the features of the metabolic syndrome (33, 44). This suggests that the progress of an adipocyte along an oxidative pathway might play a key role in the genesis of the metabolic syndrome.

Acknowledgements

We acknowledge the helpful technical assistance of Jana Smith and Gail Kilroy.

This work was supported by Takeda Pharmaceuticals North America as an unrestricted educational grant.

First Published Online October 4, 2005

Abbreviations:

 
• ASP,

  Acid-soluble products;

 
• BAT,

  brown adipose tissue;

 
• CPT-1,

  carnitine palmitoyltransferase 1;

 
• Cyt c,

  cytochrome c;

 
• FAO,

  fatty acid oxidation;

 
• MCAD,

  medium-chain acyl-coenzyme A dehydrogenase;

 
• mtDNA,

  mitochondrial DNA;

 
• mtTFA,

  mitochondrial transcription factor A;

 
• NRF,

  nuclear respiratory factor;

 
• PGC,

  PPARγ coactivator 1α;

 
• PIO,

  pioglitazone;

 
• PKA,

  protein kinase A;

 
• PPAR,

  peroxisome proliferator-activated receptor;

 
• ROSI,

  rosiglitazone;

 
• TZD,

  thiazolidinedione;

 
• UCP-1,

  uncoupling protein 1;

 
• WAT,

  white adipose tissue.