3T3-L1 cell line was purchased from American Type Culture Collection (Manassas, VA). The HEK293A cell line was from Invitrogen (Long Island, NY). Cell culture supplies were from Fisher Scientific (Agawa, MA) or Gibco Life Technology (Long Island, NY). TransLucent reporter vector for PPARγ [PPRE(+)-Luc reporter gene containing the PPARγ responsive element (PPRE)] 21 was from Panomics, Inc (Redwood City, CA). Renilla luciferase control reporter vector pRL-null and a dual luciferase reporter assay kit system were from Promega (Madison, WI). Recombinant adenovirus encoding glutathione peroxidase and its parental adenovirus Ad-5 were from Genecore of Iowa University (Ames, IA). Troglitazone was from Biomol Inc (Plymouth meeting, PA). Other chemicals, reagents, and solvents were from Sigma (St. Louis, MO), unless noted elsewhere.

3T3-L1 preadipocytes were grown in Dulbecco's minimum essential medium (DMEM) with 10% calf serum, pencicillin (100 IU), and streptomycin (100 IU). Differentiation was induced on day 2 post confluence using DMEM with 10% fetal bovine serum (FBS), 0.5 mM methylisobutylxanthine, 1 μM dexamethasome, and 170 nM insulin. After 48 h, medium was changed to DMEM plus 10% FBS and 170 nM insulin. Cells were used for incubation with octanoate or other effectors 6–9 days thereafter, at which point >90% of the cells accumulated lipid droplets.

Triglycerides (TAG) synthesis in adipocytes uses both pre-made and de novo synthesized fatty acids, with the glycerol backbone comes primarily from glucose-derived glycerol-3-phosphate. In this work, we assess the effects of octanoate on each of these steps using [9,10-³H] triolein (1 μCi/ml, 0.5% lipid emulation, measures lipoprotein lipase activity), ³H₂O (25 μCi/ml, measures fatty acid synthase activity), and [U-³H] glucose (10 μCi/ml, measures net TAG synthesis) as the substrates. Cells were incubated with labeled substrates individually for 3 h (a linear range of 0 – 5 h was confirmed in preliminary experiments), washed 4 times with warm PBS containing 1% BSA. Cells were then lysed, extracted with organic solvent, and the lipid components were separated by TLC 22. The TAG fraction was scraped for scintillation counting directly or for methylation reaction as described before 22. After removal of the methyl acyl esters, the aqueous phase containing the glycerol moiety was used for scintillation counting 22. Non-specific binding was measured by exposing cells to the same medium but washed immediately (usually <5% of that incorporated into the cellular TAG pool after the 3 h incubation).

To measure β-oxidation of octanoate, cells were grown in T25 flask, incubated in serum-free DMEM containing 0.5% BSA overnight, and then treated with exogenous [1-¹⁴C] octanoate (0.5 mM, 1 μCi/ml) together with the desired effectors for 2 h. The release of ¹⁴CO₂was measured as described before 22. To measure the β-oxidation of oleate, cells grown in 6-well plates were pre-incubated in serum-free DMEM with [9,10-³H] oleate (1 μCi/ml, <0.1 μM) overnight. Exogenous oleate was removed by washing the cells with PBS containing 0.5% BSA. Cells were then incubated with or without insulin or L-carnitine for 2 h. The release of ³H₂O into the medium was measured as described 23.

HEK293A cells were transfected with PPARγ2 using a recombinant retrovirus encoding a full-length cDNA of mouse PPARγ2 (a gift from Dr. Spiegelman BM, Harvard University) followed by neomycin selection. The 293A-PPARγ2 cells thus generated were then grown in DMEM with 10% FBS to 60% confluence and co-transfected with PPRE(+)-Luc reporter and Rluc reporter vectors using Effectene Transfection Reagent from Qiagen (Valencia, CA). Octanoate or other effectors were added to cell culture 48 h after the transfection. Cells were harvested for a dual luciferase assay using a commercial kit (Promega).

Total RNA was isolated using Trizol method (Invitrogen, Carlsbad, CA) and reverse transcription from mRNA to cDNA was performed as described before 24. Intron-spanning PCR primers were designed using a web-based program provided by Roche. House keeping gene mouse HPRT was used as the endogenous reference. SYBG-based real-time PCR was conducted in 20 μl reaction mix containing 10 μl PCR enzyme mix (Qiagen), 2 μl cDNA, 3 μl primer mix (final 1.5 μM for each primer), and 5 μl nuclease-free water, using a Rotorgene 3000A system. Amplification parameters consisted of initial enzyme activation at 95°C for 10 min and 45 cycles of three-step PCR (denature 5 s at 95°C, annealing 10 s at 60°C, and extension 20 s at 72°C). The specificity of products generated for each set of primers was examined for each fragment using a melting curve and gel electrophoresis. Reactions were run in triplicates and data calculated as the change in cycle threshold (Ct) for the target gene relative to the Ct for HPRT. To confirm the relationship between Ct values and mRNA levels, primers were calibrated by using serial dilutions of cDNA.

Cellular ROS was measured with a protocol modified from the literature 25. Briefly, 3T3-L1 adipocytes were incubated with octanoate with or without the antioxidant, N-acetylcysteine (NAC), for 24 h. Dichlorofluorescein diacetate (2 μM, DCFH-DA, Molecular Probe, Eugene, OR), a cell permeable nonfluorescent precursor, was then added to the cells and the incubation was extended for another 30 min. Within the cells, DCFH-DA is hydrolyzed by nonspecific esterases to release DCF, which is readily oxidized by intracellular ROS. The oxidized product emits green fluorescence (ex 488 nm, em 525 nm). At the end of incubation, cells were washed with warm KRB buffer and immediately imaged under a polarizing/fluorescent microscope (Nikon Eclipse TE200). Caution was taken to ensure that cells from different samples were exposed to excitation for identical period of time (30 s) and photographed using the same exposure time (15 s) and receiver gain (1.0) using a Nikon digital camera (original magnification 10x).

In this study, TAG synthesis was measured by isotope-tracing method using substrates at different stages of the biosynthesis cascade. As shown in Figure 1, pre-incubation with octanoate for 3 days substantially inhibited the de novo fatty acid synthesis (Fig. 1A), the incorporation of exogenous fatty acids into TAG (Fig. 1B), and the net synthesis of intracellular TAG (Fig. 1C). These results indicated that octanoate induced a comprehensive inhibition of lipogenesis. Since measurement was done in the absence of octanoate, the inhibitory effect was likely sustained through modulation of gene expression. To test this possibility, we performed quantitative real-time PCR analysis for selected lipogenic genes. As shown in Figure 2, octanoate induced a large decrease in the expression of key enzymes involved in fatty acid uptake and triglyceride synthesis. These included lipoprotein lipase (LPL), fatty acid synthase (FAS), and diacylglycerol acyltransferase 2 (DGAT2). The reduction in LPL and FAS correlates with the specific reduction in fatty acid uptake from exogenous triolein (Fig. 1B) and de novo fatty acid synthesis from H₂O (Fig. 1A), respectively. A reduction in DGAT2 might contribute to reduction of overall triglyceride synthesis (Fig. 1A–C). Besides, octanoate also inhibited the expression of CD36, a protein that has been shown to be required for efficient lipid storage 26. Among the targets tested, CD36 and LPL have been demonstrated as PPARγ target genes 2728. As for FAS and DGAT2, although with no defined PPRE in their promoters, both are drastically induced during preadipocyte differentiation, a process that is tightly controlled by the activation of PPARγ. Hence, these two genes can be considered as indirect downstream targets of PPARγ. It is not surprising that we detected a large decrease in mRNA of PPARγ in association with the changes in the aforementioned lipogenic genes (Fig. 2). In parallel, we also detected a similar decrease in the expression of PPARδ (Fig. 2). Co-incubation with a synthetic PPARγ agonist, troglitazone, largely restored the expression of the lipogenic genes as well as PPARγ amd PPARδ (Fig. 2). Troglitazone also caused a 1.5 fold increase in TAG synthesis and diminished the inhibitory effect of octanoate (19 and data not shown).

Results above suggest that sustained anti-liogenic effect of octanoate is correlated with reduced expression of lipogenic genes, and the effect was reversible by co-treatment with a PPARγ synthetic ligand. This distinguishes octanoate from common fatty acids that have been shown to activate PPARγ 2930313233343536. One of the unique properties of octanoate is that it might be activated within the mitochondria 12 and enters the β-oxidation pathway independent of CPT-I 37. This allows octanoate to be β-oxidized in the presence of glucose and insulin, conditions under which long-chain fatty acids are primarily channeled to esterification. This prediction, however, has not been firmly established in adipocytes. Peculiarly, several isoforms of medium chain acyl CoA synthetase have been recently identified and cloned 383940, with none expressed to an appreciable level in adipocytes 41. Because β-oxidation is linked to ROS generation 25424344, the potential molecular signals for regulation of lipogenesis, it is important to test whether β-oxidation of octanoate is truly independent of CPT-I in adipocytes. For this purpose, we measured the generation of ¹⁴CO₂ from octanoate and tested its response to insulin, L-carnitine, oleate, and etomoxir, factors that modulate CPT-I via different mechanisms. As shown in Figure 3A, β-oxidation of octanoate was slightly inhibited (~18%) by insulin, a hormone that promotes the generation of the natural inhibitor of CPT-I 37, and Etomoxir, a pharmaceutical inhibitor of CPT-I. On the other hand, L-carnitine, an activator of CPT-I, caused a ~60% inhibition of octanoate oxidation. A combination of L-carnitine and exogenous oleate further enhanced the inhibition (> 85%). In contrast, β-oxidation of oleate was increased by L-carnitine more than 2 fold but inhibited by insulin by about 60% (Fig. 3B), consistent with the literature 37. These results indicate that in adipocytes, octanoate was mainly oxidized independent of CPT-I (> 80%). A small fraction (< 20%), that was sensitive to insulin and etomoxir, might be activated in the cytosol and hence depend on CPT-I to enter the mitochondria. The observation that L-carnitine inhibited, rather than promoted, β-oxidation of octanoate suggests that activation of CPT-I largely increased the transport of endogenous fatty acids into the β-oxidation pathway which compete with octanoate for the enzymes downstream from CPT-1. This competition was further enhanced in the presence of added oleate.

A common consequence of fatty acid oxidation, as compared to that of glucose, is the generation of NADH and FADH2. This increases electron flow through the mitochondrial electron transport chain and increases electron leak to produce ROS 4546. Because ROS has been shown to activate the stress-responsive protein kinases that lead to the inhibition of PPARγ 474849, we hypothesize that ROS might be an important mediator for the anti-lipogenic effects of octanoate. To test this possibility, we first measured the ROS intensity in adipocytes treated with octanoate. As shown in Figure 4A, incubation with octanoate for 24 h significantly increased the intracellular ROS intensity, which was blocked by co-treatment with N-acetylcysteine. Secondly, we measured the effects of octanoate on PPARγ transcriptional activity using a PPRE(+)-Luc reporter gene assay. As shown in Figure 4B, pre-incubation with octanoate for 24 h inhibited PPARγ transcription activity by about 50%, and this inhibition was blocked by a co-treatment with N-acetylcysteine. As a positive control, we show that exogenous ROS, generated by a mixture of xanthine and xanthine oxidase, also inhibited PPARγ transcriptional activity, an effect blocked by N-acetylcysteine. Finally, we transfected mature adipocytes with the adenovirus encoding glutathione peroxidase 1 (GPx-1), a cytosolic isoform of GPx which detoxifies ROS 47. As shown in Figure 4C, octanoate inhibited TAG synthesis by about 40% in adipocytes transfected with the parent adenovirus Ad-5. This inhibition was offset by over-expression of GPx-1.