Epidermis-Type Lipoxygenase 3 Regulates Adipocyte Differentiation and Peroxisome Proliferator-Activated Receptor γ Activity

Plasmids.Mouse eLOX3 was cloned into pcDNA3.1 (Invitrogen) and pBABE (generously provided by Ormond MacDougald) by PCR. Deletion of amino acids (aa) N155 to P195 and generation of myc-tagged eLOX3 were performed by PCR. All clones were verified by sequencing. pBABE-ADD1R was a generous gift from Bruce M. Spiegelman. UASGal-TK-Luc (where UAS is upstream activator sequence, TK is thymidine kinase, and Luc is luciferase) was kindly provided by Ronald M. Evans. The cytomegalovirus (CMV)-β-galactosidase was from Clontech. Gal4-PPARα, -PPARδ, and -PPARγ ligand binding domains (LBD) were kindly donated by Jan Fleckner (Novo Nordisk A/S). Gal4-retinoid X receptor α (RXRα) LBD contains the rat RXRα LBD (aa 202 to 467) cloned in frame to the Gal4 DNA-binding domain (DBD) in pM (Clontech). pDR1 Luc, MSCV-PPARγ (where MSCV is murine stem cell virus), and MSCV-PPARγ Q286P were generous gifts from Bruce M. Spiegelman. pcDNA3.1mychisA XOR was a generous gift from Toren Finkel. Enhanced green fluorescent protein (EGFP)-PPARγ was constructed by cloning full-length PPARγ in frame to EGFP in CMV-EGFP (Stratagene). The PPARγ-DE domains were inserted into pGEX-5x-1 (GE Healthcare) placing a glutathione S-transferase (GST) tag in the N terminus of the PPARγ fragment. pCMX-T7-TRAP220 was generously provided by Eckhardt Treuter. pSUPER-RETRO was kindly supplied by Reuven Agami. Target sequences for eLOX3 were identified using a small interfering RNA (siRNA) target identifier from Ambion (Austin, TX). Oligonucleotides were annealed and cloned into pSUPER-RETRO. The following target sequences were used: GTATCTCACCGCAATCATC, AAGACCTTCACTTCAGAGTAC, and AATCATCTTTAATTGCTCCGC.

Cell culture, differentiation, transduction, siRNA transfection, and oil red O staining.Wild-type mouse embryonic fibroblasts (MEFs) and MEFs lacking a functional retinoblastoma tumor suppressor gene (Rb^−/−) were generous gifts from Jiri Bartek and Jiri Lukas. MEFs and 3T3-L1 preadipocytes were grown, differentiated, and transduced as described previously (21). For the hepoxilin differentiation experiments, vehicle or hepoxilins at a 4 or 10 μM final concentration were added to differentiation medium at days 0, 1, 2, and 3. N-Acetyl-cysteine (NAC) was dissolved in water and vehicle, or NAC (5 mM) was added to the differentiation medium for the periods indicated in the Fig. 7 legend. eLOX3 SMART pool RNA interference (RNAi) (Dharmacon) was transfected into 3T3-L1 cells at confluence, using DeliverX plus (Panomics) according to the manufacturer's instructions. For oil red O staining, plates were washed in phosphate-buffered saline (PBS), and cells were fixed in 3.7% paraformaldehyde for 1 h and stained as described previously (21).

Immunofluorescence analysis.EGFP-PPARγ and myc-eLOX3 were transfected into 50% confluent 3T3-L1 cells using Metafectene (Biontex). After 24 h, cells were washed twice in PBS, pH 7.2, and fixed with 90% methanol at −20°C. Following fixation, cells were washed in PBS, pH 7.2, and blocked for 1 h at room temperature in 3% horse serum (HS)-PBS. Incubation with monoclonal anti-myc antibody (Roche 9E10) was performed at a 1:50 dilution in 3% HS-PBS for 1 h at room temperature. Cells were subsequently washed in PBS, pH 7.2, and incubated with a Cy3-conjugated anti-mouse antibody (Jackson ImmunoLabs) at 1:250 for 1 h at room temperature. Finally, cells were washed and mounted using fluorescent mounting medium (Dako). Visualization was performed on a Zeiss Axiovert 200 M laser scanning confocal microscope with a pinhole of 1, and scanning was performed within selected ranges of 495 to 548 nm for EGFP and from 580 nm for Cy3. Image processing was performed with Zeiss software.

Real-time PCR and protein analysis.Total RNA was purified as described previously (21). Prior to reverse transcription, RNA was treated with RQ1 DNase (Promega), and reverse transcription was performed essentially as described previously (35). Expression levels of genes of interest were normalized to expression the of TATA box-binding protein (TBP). Primer sequences are available on request.

For protein analysis, whole-cell extracts, electrophoresis, blotting, visualization, and stripping of membranes were performed as described previously (21). Primary antibodies were rabbit anti-eLOX3 (35), mouse anti-PPARγ (Santa Cruz sc-7273), rabbit anti-GLUT1 (Abcam ab652), rabbit anti-GLUT4 (Abcam ab654), rabbit anti-TFIIB (Santa Cruz sc-274), and rabbit anti-aP2 (kindly provided by David A. Bernlohr). Secondary antibodies were horseradish peroxidase-conjugated anti-rabbit and anti-mouse antibodies from Dako.

LOX activity assay and detection of HETEs by enzyme-linked immunosorbent assay (ELISA).Detection of LOX activity was performed essentially as described previously (9, 52) with increasing amounts of baicalein added to the enzymatic reaction as indicated in Fig. 2D.

ELISAs for the specific detection of 12(S)-HETE and 15(S)-HETE were from R&D Systems and were used according to the manufacturer's instructions. Briefly, organic extracts of medium from 3T3-L1 preadipocytes were loaded on C₁₈ columns. Fractions from ethyl acetate elutions were evaporated to dryness, resuspended in ethanol, and assayed.

Transfections.Sixty percent confluent Rb^−/− MEFs were transfected using Metafectene (Biontex) according to the manufacturer's instructions. Six hours after transfection, medium was changed to AmnioMax basal medium supplemented with antibiotics and hepoxilin A3 (HXA3), hepoxilin B3 (HXB3) (Biomol), L165041 (kindly provided by Merck), rosiglitazone (kindly supplied by Novo Nordisk A/S), or LG1069 (kindly supplied by Novo Nordisk A/S). All compounds were dissolved in dimethyl sulfoxide (DMSO) or ethanol. Where no compounds were added, an equal volume of vehicle was added. Cells were harvested after 12 h, and luciferase and β-galactosidase activities were measured according to standard protocols.

Insulin-stimulated glucose uptake.Glucose uptake was measured in 3T3-L1 adipocytes. On day 6 of differentiation, the cells were treated with vehicle, rosiglitazone, or hepoxilins. Forty-eight hours later the cells were incubated with serum-free Dulbecco's modified Eagle's medium (DMEM) supplemented with antibiotics at 37°C for 1 to 2 h. Afterwards, the cells were washed once with Krebs-Ringer-HEPES (KRP) buffer and incubated with KRP buffer for 30 min at 37°C. Insulin was added to a final concentration of 0, 10, or 100 nM, and cells were incubated at 37°C for 15 min. Glucose uptake was initiated by the addition of 100 μl of KRP buffer supplemented with 10 mM glucose (5 mCi/liter 2-deoxy-d-[¹⁴C]glucose), and cells were incubated for 10 min at 37°C. Glucose uptake was terminated by three washes with ice-cold PBS. The cells were lysed with 1% Triton X-100, scintillation fluid was added, and radioactivity was determined.

Liquid chromatography mass spectrometry (LC/MS).Sample preparation and instrument setup were performed as described previously (1). Screening was carried out by fragmenting the ion [M-H]⁻ signal m/z 335.1 (M is hepoxilin A3 or B3) and monitoring the transitions at [M-H₂O-H]⁻ signal m/z 317, [M-2H₂O-H]⁻ signal m/z 299, [M-COO-H]⁻ signal m/z 291, [M-H₂O-COO-H]⁻ signal m/z 273, and [M-2H₂O-COO-H]⁻ signal m/z 255. The intensities of the signals were recorded in ion counts per second (icps).

Ligand binding assay.A LanthaScreen time-resolved fluorescence resonance energy transfer (TR-FRET) PPARγ competitive binding assay (Invitrogen) was used according to the manufacturer's instructions with increasing concentrations of hepoxilin A3, hepoxilin B3, or rosiglitazone, as indicated in Fig. 4B.

GST pulldown.GST-PPARγ DE was expressed in Escherichia coli by induction with 0.1 mM isopropyl-β-d-thiogalactosidase at 30°C. Cells were disrupted by sonication, and the cell lysate was subsequently incubated with glutathione-Sepharose (GE Healthcare). The TRAP220 fragment was produced by in vitro translation (TnT; Promega) in the presence of [³⁵S]methionine (Amersham). Beads and in vitro translated proteins were incubated in pulldown buffer (20 mM Tris-HCl, pH 8.0, 100 mM NaCl, 10 mM EDTA, 0.5% NP-40, 1% skim milk, and protease inhibitors [Complete; Roche]) for 15 min at 4°C in the presence of vehicle, rosiglitazone, HXA3, or HXB3. Beads were washed once with pulldown buffer containing skim milk and twice with pulldown buffer without skim milk. All buffers contained vehicle or ligands. Beads were then boiled in SDS lysis buffer and resolved using SDS-PAGE. [³⁵S]methionine-labeled proteins were visualized by autoradiography.

Docking of hepoxilins.To model docking of hepoxilin A3 and B3 to the PPARγ ligand-binding pocket, the Gold Suite software package was applied (CCDC Gold Suite, Cambridge, United Kingdom). The two hepoxilins were prepared using Corina, version 3.00 (Molecular Networks, Erlangen, Germany), for generating three-dimensional structures, Omega, version 2.2.1 (Openeye Scientific Software Inc., Santa Fe, NM), for energy minimization, and Open Babel, version 2.1.0 (http://openbabel.sourceforge.net/ ), for calculating protonation states at physiological pH. Sybyl, version 8.0 (Tripos, St. Louis, MO), was applied for protein preparation. The Brookhaven Protein Data Bank (PDB) was searched for crystallographic data of agonist-PPARγ complexes (7). Among the agonist-PPARγ complexes that included two fatty acids bound simultaneously in the ligand-binding domain, the PDB entry 2vsr had the highest crystal structure resolution, so its PPARγ ligand-binding pocket was chosen to set up a docking protocol. The resulting docking poses were visualized with LigandScout, version 2.0 (56). The best docking poses were determined with respect to their GOLDScores and plausibility. For example, a docking pose that represented known agonist-PPARγ interactions and was ranked second in terms of GOLDScore was preferred over the highest-ranked docking pose that formed mostly implausible ligand-protein interactions.

Data presentation.All data shown are representatives of at least two independent experiments performed at least in duplicate. All error bars represent standard deviations.

eLOX3 promotes adipocyte differentiation of Rb^−/− mouse embryonic fibroblasts (MEFs).We previously showed that LOX inhibitors prevented adipogenesis if added during the early stages of 3T3-L1 adipocyte differentiation and that this inhibition was circumvented by the addition of a PPARγ agonist (35). These findings suggested a LOX-dependent activation of PPARγ during the initiation of adipocyte differentiation, implicating the unconventional LOX, eLOX3, in the process since this is the only LOX detectable in differentiating 3T3-L1 preadipocytes (35).

To investigate the possible role of eLOX3 in activation of PPARγ, we used Rb^−/− MEFs as a model system. Whereas more than 90% of 3T3-L1 preadipocytes differentiate when induced by the standard methylisobutylxanthine, dexamethasone, and insulin (MDI) protocol, Rb^−/− MEFs require a PPARγ agonist in the differentiation cocktail in order to undergo adipose conversion (21). Rb^−/− MEFs are furthermore a particularly sensitive system for detecting PPARγ ligands as the retinoblastoma protein (pRB) is a PPARγ corepressor (15). Therefore, we initially used Rb^−/− MEFs as a model for identification of enzymes and pathways involved in activation of PPARγ and, hence, possibly PPARγ ligand production.

Rb ^{− / −} MEFs were transduced with a retroviral vector encoding eLOX3 or empty vector, selected, and differentiated according to the MDI protocol. Vector control cells barely differentiated, but cells expressing eLOX3 underwent adipocyte differentiation as seen by increased oil red O staining (Fig. 1A) and expression of the adipocyte marker genes aP2 and PPARγ2 (Fig. 1B). As observed for 3T3-L1 preadipocytes (35), addition of the LOX inhibitor baicalein from the time of induction and throughout the differentiation period prevented differentiation of Rb^−/− MEFs with forced expression of eLOX3 (Fig. 1A and B). If baicalein abolished adipocyte differentiation by inhibiting eLOX3 activity and eLOX3 activity preceded PPARγ activation, we predicted that inclusion of the PPARγ-specific agonist rosiglitazone would overcome the inhibitory effect of baicalein. Indeed, addition of rosiglitazone rescued adipocyte differentiation of baicalein-treated Rb^−/− MEFs overexpressing eLOX3 (Fig. 1A and B). These results show that eLOX3 induced differentiation in a model system that would otherwise require addition of an exogenous PPARγ activator. To investigate if eLOX3 expression was perturbed in Rb^−/− MEFs and to test if reintroduction of eLOX3 at least in part augmented adipogenesis by restoring eLOX3 expression levels, we examined eLOX3 protein levels by immunoblotting. Intriguingly, eLOX3 levels were decreased in Rb^−/− MEFs compared to levels in wild-type MEFs (Fig. 1C), suggesting that low expression of eLOX3 in Rb^−/− MEFs contributed to their impaired ability to undergo adipocyte differentiation.

• Open in new tab
• Download powerpoint

FIG. 1.

Forced expression of eLOX3 increases differentiation of Rb^−/− MEFs. Cells were transduced with either an empty or eLOX3-expressing vector and selected using puromycin. Cells were induced to differentiate according to the MDI protocol with the inclusion of either vehicle, 20 μM baicalein, or 20 μM baicalein and 0.5 μM rosiglitazone from time of induction. Eight days after induction, cells were scored for differentiation. (A) Cells stained for lipid accumulation using oil red O. (B) Expression of adipocyte marker genes, PPARγ2 and aP2, was analyzed by Western blotting. (C) Protein levels of eLOX3 in wild-type and Rb^−/− MEFs at the onset of differentiation as analyzed by Western blotting. TFIIB was included as a loading control. One representative experiment out of three independent experiments performed in duplicate is shown.

ADD1 has been implicated in the production of PPARγ ligands (26), but the mechanism by which ADD1 leads to increased PPARγ ligand production has not been elucidated. Using transient transfections, we observed that ADD1 induced the expression of a luciferase reporter gene controlled by a 2.5-kb fragment of the eLOX3 promoter (data not shown), indicating that augmentation of eLOX3 expression might underlie ligand production in response to forced expression of ADD1. To further examine this possibility, we transduced Rb^−/− MEFs with a retroviral vector expressing the E-box-selective ADD1 mutant (ADD1R), previously shown to be as efficient as wild-type ADD1 in inducing PPARγ ligand production (26). Forced expression of ADD1R, indeed, caused adipogenesis in Rb^−/− MEFs. This effect was abolished by the addition of baicalein, but not indomethacin, a cyclooxygenase inhibitor. Addition of rosiglitazone to the baicalein-treated cells rescued differentiation (see Fig. S1A to D in the supplemental material). Collectively, these results indicated that forced expression of ADD1R may promote PPARγ ligand production and adipogenesis in a LOX-dependent manner, underscoring the ability of LOXs to promote adipogenesis in vitro. However, two points strongly argue against the idea that ligand production induced by forced expression of ADD1R reflects the mechanism for ligand generation and activation of PPARγ during normal adipocyte differentiation. eLOX3 mRNA decreases upon induction of adipocyte differentiation, which is opposite to the increase in ADD1 expression. Whereas eLOX3 mRNA drops at the onset of adipogenesis, the eLOX3 protein level remains stable, suggesting that eLOX3 activity is regulated by other means than increased mRNA expression. Besides augmenting eLOX3 mRNA levels, forced expression of ADD1R also induced expression of platelet-type 12(S)-LOX (p12SLOX) (see Fig. S1E in the supplemental material), an enzyme with the ability to produce substrates for eLOX3. Expression of p12SLOX is, however, undetectable in 3T3-L1 cells (35), which disagrees with the assumption that this enzyme is the one responsible for generating substrates for eLOX3 during adipogenesis.

eLOX3 enzymatic activity is required to promote adipocyte differentiation.To test if eLOX3 activity is required to induce adipocyte differentiation of Rb^−/− MEFs, we constructed an enzymatically inactive eLOX3 mutant. Deletion of amino acids N155 to P195 (eLOX3Δ) abolished the enzymatic activity, as measured by the ability to convert 15(S)-HPETE into 13(R)-hydroxy-14(S),15(S)-HepEtrE (Fig. 2C). Whereas expression of full-length eLOX3 increased adipogenic conversion in Rb^−/− MEFs, expression of eLOX3Δ in Rb^−/− MEFs did not, as determined by oil red O staining and expression of PPARγ2 and aP2 (Fig. 2A and B). Furthermore, to elucidate whether eLOX3 activity was directly affected by baicalein, we tested the effect of increasing amounts of baicalein on the in vitro activity of eLOX3, as well as epidermal-type 12(S)-LOX (e12SLOX) and leukocyte-type 12(S)-LOX (l12SLOX). Addition of baicalein inhibited eLOX3 to a degree comparable to the inhibition of the two 12(S)-LOXs (Fig. 2D), suggesting that baicalein inhibits adipocyte differentiation by directly inhibiting eLOX3.

• Open in new tab
• Download powerpoint

FIG. 2.

The enzymatic activity of eLOX3 is required for its adipogenic effect. Rb^−/− MEFs were transduced with vectors expressing eLOX3, the eLOX3Δ inactive mutant, or vector control and selected using puromycin. Eight days after induction with MDI, cells were scored for adipogenic conversion. (A) Cells were stained for lipid accumulation using oil red O. (B) Expression of adipocyte marker genes, PPARγ and aP2, was analyzed by Western blotting. Data are representative of three independent experiments performed in duplicate. (C) Enzymatic properties of the eLOX3Δ mutant. Cells were transiently transfected with wild-type eLOX3, eLOX3Δ, or LacZ control expression vectors, lysed in LOX assay buffer, and incubated with 100 μM 15(S)-HPETE for 15 min at 37°C. Products were extracted and run on a high-performance liquid chromatograph, and eluates were monitored at 205 nM. Depicted is conversion of 15(S)-HPETE into 13(R)-hydroxy-14(S),15(S)-HepEtrE with the activity of wild-type eLOX3 set to 100%. Data are representative of two independent experiments. (D) Baicalein inhibits enzymatic activity of known LOXs. Expression vectors harboring cDNAs encoding different LOXs were transiently transfected into 293 cells. Cells were lysed in LOX assay buffer and incubated with appropriate substrate and increasing amounts of baicalein. Enzymatic activation is depicted as percent activity with activation in the absence of baicalein set to 100%. (E and F) Identified products of eLOX3, HXA3 and HXB3, increase adipose conversion of Rb^−/− MEFs. Cells were induced to differentiate according to the MDI protocol with addition of 10 μM HXA3 or HXB3 or vehicle control at the time of induction and on days 1, 2, and 3. Adipogenic conversion was investigated 8 days after induction. (E) Morphological differentiation and staining of lipids by oil red O. (F) Expression of adipogenic marker genes by real-time PCR. PEPCK, phosphoenolpyruvate carboxykinase. One representative experiment out of three independent experiments performed in duplicate is shown.

Hepoxilins promote adipocyte differentiation.eLOX3 converts HPETE intermediates into epoxyalcohol fatty acids in a hydroperoxide isomerase reaction. The 12-HPETE-derived epoxyalcohol fatty acids are known as hepoxilins (42, 63, 65). Hepoxilins are a group of epoxyalcohols derived from hydroperoxy fatty acids. Their function is best studied in relation to upholding the function of the epidermis. The group is normally segregated into two types of hepoxilins, A3 and B3, based on the distance between the epoxy and hydroxyl group. Hepoxilins of the A type are normally much more unstable than the B type hepoxilins. Type A hepoxilins are easily hydrolyzed either by acids or in cells by epoxide hydrolases and are therefore difficult to detect in biological extracts (8). To examine whether hepoxilin A3 or hepoxilin B3 could recapitulate the effects observed upon forced eLOX3 expression, we tested their ability to restore adipogenesis in the Rb^−/− MEFs. Rb^−/− MEFs were induced to differentiate according to the MDI protocol with the inclusion of 10 μM hepoxilin A3, hepoxilin B3, or vehicle control during the first 4 days of differentiation. Eight days after induction, cells were stained with oil red O (Fig. 2E), and RNA was harvested for analysis of expression of differentiation markers (Fig. 2F). Vehicle-treated cells did not differentiate, whereas inclusion of either hepoxilin A3 or B3 promoted adipocyte differentiation of the Rb^−/− MEFs.

Forced expression eLOX3 induces adipose conversion of 3T3-L1 preadipocytes, and hepoxilins are produced during the initial stages of differentiation.Having established that eLOX3 can stimulate adipose conversion in Rb^−/− MEFs, we sought to determine if eLOX3 also was able to induce adipocyte differentiation in 3T3-L1 preadipocytes. 3T3-L1 preadipocytes readily undergo adipocyte differentiation when induced according to the MDI protocol, and an elevation of the cAMP level at the initiation of differentiation has been linked to the transient induction of PPARγ ligand production (54). The decrease in adipose conversion of 3T3-L1 preadipocytes by the exclusion of the cAMP-elevating compound can be circumvented by the inclusion of an exogenous PPARγ ligand. We therefore hypothesized that forced expression of eLOX3 could increase adipogenesis in 3T3-L1 preadipocytes induced with only dexamethasone and insulin. Indeed, ectopic expression of eLOX3 augmented adipocyte differentiation of 3T3-L1 preadipocytes (Fig. 3A). Furthermore, under the same conditions both hepoxilin A3 and B3 simulated adipose conversion of 3T3-L1 preadipocytes to a degree similar to rosiglitazone as assessed by oil red O and marker gene expression (Fig. 3B and C).

• Open in new tab
• Download powerpoint

FIG. 3.

eLOX3 and its products induce adipogenesis in 3T3-L1 preadipocytes. Cells were transduced with either empty vector or eLOX3-expressing vector and selected using puromycin. Cells were induced to differentiate with dexamethasone and insulin. Eight days after induction, cells were scored for differentiation as measured by oil red O staining (A). (B and C) 3T3-L1 cells were induced to differentiate with dexamethasone and insulin and with addition of HXA3, HXB3, rosiglitazone, or vehicle control at the time of induction and on days 1, 2, and 3. Adipogenic conversion was investigated 8 days after induction. (B) Morphological differentiation and staining of lipids by oil red O. (C) Expression of adipogenic marker genes by real-time PCR. Data are represented as induction relative to the induction by rosiglitazone over vehicle control. Induction by rosiglitazone is set to 100%. PEPCK, phosphoenolpyruvate carboxykinase. (D) LC/MS analyses of conditioned medium samples from 3T3-L1 cells treated with either DI or MDI for 24 h. Only peaks from signals corresponding to species with a molecular mass similar to that of hepoxilin (335.1 Da) are shown. The peak from the eluted hepoxilin is marked by a red arrow. (E) MS/MS analyses of hepoxilin standards (left and middle panels) and the coeluting peak from day 1 medium from 3T3-L1 cells induced to differentiate with MDI. One representative experiment out of two independent experiments performed in duplicate is shown.

Tzameli and colleagues have shown that PPARγ ligand(s) is produced early and transiently during adipocyte differentiation in a cAMP-dependent manner, and the ligand(s) is detectable in the medium (54). We therefore examined conditioned medium from 3T3-L1 cells for the presence of hepoxilins during the early stages of the adipogenic conversion process using liquid chromatography tandem mass spectrometry (LC/MS/MS). By LC, hepoxilin A and hepoxilin B coeluted whether spiked with blank medium or not (data not shown). As ligand production was shown to be cAMP dependent, we analyzed conditioned medium from cells treated with DI or MDI for 24 h. Only lipids extracted from 3T3-L1 cells treated with MDI yielded substances with a chromatographic retention time compatible with the presence of hepoxilin A3 or B3 in the extract (Fig. 3D). The MS/MS fragmentation patterns of blank medium samples spiked with hepoxilin A3 or B3 are characterized by diagnostic fragment ions with m/z of 317, 299, 273, and 275, but it is possible at least partly to distinguish between the two by comparing the m/z 317 and 273 intensities. The intensity of fragment m/z 317 was higher than that of the m/z 273 fragment for hepoxilin B3 (ratio 317/273, ∼ 2.61), while the opposite was observed for hepoxilin A3 (ratio 317/273, ∼ 0.62). Furthermore, only fragmentation of hepoxilin B3 resulted in an ion with an m/z of 291. A similar pattern was observed when the pure compounds were analyzed. By MS/MS analysis of the sample derived from the MDI-treated sample, a fragmentation pattern clearly compatible with the presence of hepoxilins was observed. The presence of an ion with an m/z of 291 and the m/z 317/273 ratio of 2.63 suggests that at least hepoxilin B3 was generated during this early stage of the differentiation program, but the unstable nature of hepoxilin A3 makes it possible also that this hepoxilin is produced together with hepoxilin B3 (Fig. 3E). Moreover, consistent with the results of Tzameli and coworkers, the production of hepoxilin(s) was dependent on elevated cAMP levels.

Hepoxilins activate PPARγ.To test if hepoxilin A3 and B3 directly affected the activity of PPARγ, we transiently transfected Rb^−/− MEFs with Gal4-DBD-PPARγ-LBD, a UASGal-TK-luciferase reporter, and a CMV-β-galactosidase normalization vector and treated the cells with increasing amounts of hepoxilin A3 or B3. Figure 4A demonstrates that both hepoxilins activated PPARγ in a dose-dependent manner. At a concentration of 10 μM, the activation reached 40% of that observed with 0.5 μM rosiglitazone (Fig. 4A).

• Open in new tab
• Download powerpoint

FIG. 4.

Hepoxilin A3 and B3 bind and activate PPARγ. (A) HXA3 and HXB3 activate PPARγLBD in a dose-dependent manner. Rb^−/− MEFs transfected with a Gal4-DBD-PPARγ-LBD expression vector and a UASGal-TK-luciferase reporter were treated with vehicle, 0.5 μM rosiglitazone, or increasing concentrations of HXA3 or HXB3. Luciferase activity was normalized to ß-galactosidase activity and is presented as fold change over vehicle only control. One representative experiment out of four independent experiments performed in triplicate is shown. (B) HXA3, HXB3, and rosiglitazone displace a fluorescent ligand from the PPARγ-LBD. RFU, relative fluorescence units. One representative experiment out of three independent experiments performed in triplicate is shown. (C) Rb^−/− MEFs transfected with pDR1 Luc and either PPARγ, PPARγ Q286P, or empty vector treated with 1 μM rosiglitazone, 10 μM HXA3, or 10 μM HXB3. Luciferase activity was normalized to β-galactosidase activity. Data represent three independent experiments performed in triplicate. (D) Potentiation of insulin-stimulated glucose uptake by vehicle, rosiglitazone, HXA3, or HXB3. Differentiated 3T3-L1s were treated with the indicated compounds, and basal or insulin-stimulated uptake of radioactively labeled glucose was measured. Data are presented as fold uptake over the vehicle only control. Data represent two independent experiments performed in octuplicate. (E) GST-PPARγ DE pulldown of TRAP220 in the presence of vehicle, 1 μM rosiglitazone, 10 μM HXA3, or 10 μM HXB3. One representative experiment out of two independent experiments is shown. (F) 3T3-L1 cells were transfected with GFP-PPARγ2 and myc-eLOX3, and localization of the two fusion proteins was determined by confocal microscopy.

PPARδ has been shown to stimulate adipocyte differentiation (4, 5, 22, 36). Furthermore, retinoid X receptor α (RXRα) has been shown to bind fatty acids (14), and liganded RXR enhances the transactivating potential of the PPARγ/RXR heterodimer (51). Therefore, we examined whether the hepoxilins activated PPARδ or RXRα using transient transfection of Rb^−/− MEFs with Gal4-DBD-PPARδ-LBD or Gal4-DBD-RXRα-LBD as described above. Neither hepoxilin A3 nor hepoxilin B3 activated RXRα, and they only very weakly activated PPARδ, suggesting that the adipogenic effect of hepoxilins is not mediated via activation of PPARδ or RXRα (data not shown). In addition, Gal4-DBD-PPARα-LBD was not activated by either hepoxilin A3 or B3 in Rb^−/− MEFs (data not shown). To further test whether hepoxilin A3 or B3 bound directly to PPARγ through its ligand binding pocket, we determined their relative binding affinities in a ligand displacement assay (Fig. 4B). Both hepoxilins displaced the binding of fluorescent ligand to PPARγ with 50% K_d (K_d50) values for rosiglitazone, hepoxilin A3, and hepoxilin B of 27 nM, 11 μM, and 4 μM, respectively. The apparently high K_d50 value for the hepoxilins may reflect the inherent instability of the epoxide ring in these compounds in aqueous solution. However, collectively our results indicate that the effects observed on both activation of PPARγ and adipocyte differentiation could be mediated through direct binding to the PPARγ LBD.

Hypothetically, hepoxilins might augment PPARγ-mediated transactivation in a manner independent of binding to the ligand binding domain. To examine this possibility, we took advantage of a PPARγ mutant (PPARγ Q286P) in which a glutamine-to-proline exchange of residue 286 abolishes the ability of PPARγ to bind and become activated by virtually all known ligands (48, 55). Figure 4C shows that hepoxilin A3 and B3 increased the activity of wild-type PPARγ, whereas both hepoxilins failed to augment the activity of PPARγ Q286P, supporting the notion that hepoxilins activate PPARγ by binding to the ligand binding domain and thereby activating the AF2 function.

Synthetic PPARγ agonists have been widely used clinically because of their ability to increase insulin sensitivity. We therefore examined if the hepoxilins could increase insulin signaling in differentiated 3T3-L1 adipocytes by measuring insulin-dependent glucose uptake. Both hepoxilin A3 and B3 potentiated insulin-stimulated glucose uptake to levels similar to rosiglitazone (Fig. 4D). Neither the two hepoxilins nor rosiglitazone affected the protein levels of the two glucose-transporters, GLUT1 and GLUT4 (see Fig. S2 in the supplemental material).

PPARγ agonists increase transactivation by facilitating binding of coactivators. TRAP220 (Mediator Subunit 1, Med1) can be directly recruited to PPARγ in a ligand-dependent manner and is part of the mediator complex shown to be required for adipogenesis (18, 19). Furthermore, full PPARγ agonists have been shown to recruit TRAP220 more efficiently than partial PPARγ agonists (37). We therefore examined the ability of hepoxilins to enhance binding of TRAP220 to PPARγ. Hepoxilin B3 enhanced binding of TRAP220 to the PPARγ LBD to a level comparable to that of rosiglitazone, whereas recruitment mediated by hepoxilin A3 was lower (Fig. 4E). The low level of recruitment of TRAP220 in response to hepoxilin A3 may at least in part be due to the highly unstable nature of hepoxilin A3. To examine in more detail the ability of hepoxilins to recruit cofactors to PPARγ, we attempted to use a microarray approach to determine interactions between PPARγ and peptides derived from 18 known coregulators (31). However, incompatibility between hepoxilins and dithiothreitol (DTT) needed in the buffers for this assay prevented acquisition of consistent data (results not shown).

As an alternative method to examine the possible binding of hepoxilin A3 and B3 to PPARγ, the putative binding modes of both compounds were determined by docking into the ligand-binding pocket (see Fig. S3 in the supplemental material). Schwabe and colleagues recently showed that two molecules of 9-HODE, which is structurally similar to hepoxilin A3 and B3, bind simultaneously to the PPARγ ligand-binding pocket (24). When two molecules of either hepoxilin A3 and B3 were docked to the PPARγ ligand-binding pocket, ligand-protein interactions similar to the binding mode of 9-HODE were observed. Therefore, we hypothesized that two molecules of the hepoxilins could bind simultaneously to PPARγ. Based on this assumption, the putative binding mode of hepoxilin A3 included three hydrogen bonds formed between the oxygen atoms of the carboxylic acid group of the first molecule of hepoxilin A3 and Ser289, His449, and Tyr473 as well as one hydrogen bond located between the hydroxyl moiety of the second hepoxilin A3 and Glu295. In addition, the hydroxyl moiety of the first hepoxilin A3 molecule is involved in hydrogen bonds with one of the carboxyl oxygens of the second hepoxilin A3 as well as with the epoxy oxygen of the same molecule. Moreover, the hepoxilin A3 molecule that formed hydrogen bonds to the active site residues Ser289, His449, and Tyr473 established hydrophobic contacts to residues Ala278, Ile281, Leu356, and Phe360, whereas the second molecule exhibited hydrophobic interactions with residue Ile341.

The best docking pose for hepoxilin B3 had two hydrogen bonds between the oxygen atoms of the carboxylic acid moiety of the first hepoxilin B3 and residues His449 and Tyr473 and another hydrogen bond between the hydroxyl moiety of this ligand and residue Ser289. The second hepoxilin B3 molecule forms one hydrogen bond between its hydroxyl moiety and Glu295. Furthermore, one hydrogen bond was established between the hydroxyl moiety of the first hepoxilin B3 and one carboxyl oxygen of the second hepoxilin B3 molecule. Finally, the hepoxilin B3 molecule predicted to bind to the active site residues Ser289, His449, and Tyr473 exhibited hydrophobic interactions with residues Leu330, Val339, Ile341, and Met364, whereas the second hepoxilin B3 molecule was involved in a hydrophobic interaction with residue Ile262.

Spatial uncoupling of the site of production of a signaling molecule and the mediator could pose a problem for the conduction of a signal, especially for compounds containing structures as unstable as the epoxide ring in hepoxilins. We therefore examined the localization of eLOX3 and PPARγ in 3T3-L1 preadipocytes. As can be seen in Fig. 4F, PPARγ localized solely to the nucleus, whereas eLOX3 was present both in the cytoplasm and the nucleus. Its presence in the nucleus near PPARγ indicates that hepoxilin could be generated close to PPARγ, leading to local concentrations sufficient to activate the nuclear receptor.

Downregulation of eLOX3 in 3T3-L1 blocks adipocyte differentiation.As eLOX3 has a stimulatory effect on adipogenesis, we speculated if eLOX3 was required for adipocyte differentiation of 3T3-L1 preadipocytes. We therefore knocked down expression of eLOX3 in 3T3-L1 cells and measured their ability to undergo adipose conversion. Cells were transduced with five different short hairpin RNA (shRNA)-expressing vectors specifically designed to silence eLOX3 or an empty vector control. Pools of selected cells were seeded and either harvested for analysis of eLOX3 expression (day 0) or induced to differentiate according to the standard MDI protocol. Six days after induction, cells were Oil red O stained, and protein was harvested for analysis of differentiation markers (Fig. 5A and B). Three shRNA constructs, sheLOX3-1, sheLOX3-2, and sheLOX3-3, reduced eLOX3 expression below the limit of detection on day 0 (Fig. 5B). On day 6, expression of eLOX3 had been partially restored in cells transduced with sheLOX3-1, whereas eLOX3 expression was still significantly suppressed in cells transduced with sheLOX3-2 and sheLOX3-3. Of note is the striking inverse relationship between eLOX3 expression on day 6 and adipocyte differentiation. eLOX3 expression was most efficiently repressed in cells transduced with sheLOX3-2 and sheLOX3-3, and these cells did not differentiate, as measured by oil red O staining and expression of the adipocyte markers, aP2 and PPARγ. Slight expression of eLOX3 in sheLOX3-1-transfected cells correlated with expression of PPARγ and aP2.

• Open in new tab
• Download powerpoint

FIG. 5.

Knockdown of eLOX3 expression prevents adipocyte differentiation of 3T3-L1 cells. Cells were transduced with shRNA constructs against eLOX3, selected with puromycin, and induced to differentiate according to the MDI protocol. (A) Cells stained with oil red O. (B) Detection of eLOX3 expression and adipocyte markers using Western blotting. TFIIB was included as a loading control. One representative experiment out of three independent experiments performed in duplicate is shown. (C and D) 3T3-L1 preadipocytes were transfected with either scrambled or eLOX3 RNAi using DeliverX Plus and induced to differentiate in the presence or absence of 0.5 μM rosiglitazone. (C) Cells were stained for lipid accumulation using oil red O. (D) Expression of PEPCK mRNA as a marker for differentiation was examined by real-time PCR. One representative experiment out of two independent experiments performed in triplicate is shown.

As an alternative for downregulation of eLOX3 expression, we next employed RNAi as a tool to knock down eLOX3 expression. As seen in Fig. 5C and D, eLOX3 silencing by RNAi partly recapitulated the phenotype observed by shRNA silencing of eLOX3. We assume that the observed differences in degree of suppression of adipocyte differentiation relate to differences in the efficiency of RNAi transfection versus retroviral transduction and selection. Importantly, however, the impaired adipogenic conversion observed by eLOX3 silencing could be bypassed by adding rosiglitazone during differentiation, as visualized by oil red O staining (Fig. 5C), and expression of phosphoenolpyruvate carboxykinase (PEPCK), an adipocyte marker (Fig. 5D).

XOR and eLOX3 synergistically enhance PPARγ-mediated transactivation.As described above, eLOX3 cannot by itself use arachidonic acid as a substrate but can catalyze only the conversion of HPETEs into epoxyalcohols. HPETEs are normally considered to be generated solely by other LOXs. This makes eLOX3 a counterintuitive candidate for PPARγ agonist production as no other LOXs are present at detectable levels in 3T3-L1 cells (35). Still, we observed increased levels of HETEs, the degradation products of HPETEs, upon induction of differentiation (Fig. 6A). The dioxygenation catalyzed by LOXs resembles that of lipid auto-oxidation, suggesting that the presence of ROS during adipocyte differentiation could result in generation of HPETEs. Recently, XOR was shown to be implicated in the production of PPARγ ligands during adipogenesis (13), and blocking the ROS-generating site of XOR abolished ligand production. We therefore hypothesized that XOR could generate HPETEs through ROS formation, thereby cooperating with eLOX3 in the production of PPARγ-activating hepoxilins. Auto-oxidation of arachidonic acid by ROS results in the generation of isoprostanes, malondialdehydes, 4-hydroxy-nonenals, and HPETEs (40), of which the latter can be converted into hepoxilins by eLOX3.

• Open in new tab
• Download powerpoint

FIG. 6.

eLOX3 cooperates with XOR in the activation of PPARγ. (A) 15(S)- and 12(S)-HETE in the medium from wild-type MEFs before (day 0) and 24 h after (day 1) induction with MDI. Data represent two independent experiments performed in duplicate. (B) Rb^−/− MEFs were transiently transfected using UASGal-TK-luciferase reporter, Gal4-DBD-PPARγ-LBD, and either XOR, eLOX3, or both. Transfected cells were stimulated with 5 μM arachidonic acid. Luciferase activity was normalized to ß-galactosidase activity and is presented as fold change over vector only control. One representative experiment out of four independent experiments performed in triplicates is shown. (C) Expression of XOR in wild-type and Rb^−/− MEFs was analyzed by real-time PCR and normalized to TBP expression. Data represent two independent experiments performed in triplicate.

To test the effect of XOR and eLOX3 on PPARγ activation, we examined the transactivating activity of Gal4-DBD-PPARγ-LBD in the presence of either XOR, eLOX3, or both (Fig. 6B). Interestingly, in the presence of low concentrations of arachidonic acid (5 μM), XOR or eLOX3 alone minimally increased the activity of the PPARγ-LBD. However, when both were expressed, PPARγ-mediated transactivation was synergistically enhanced. Even though hepoxilins efficiently restored adipogenesis in the Rb^−/− MEFs (Fig. 2), forced expression of eLOX3 resulted in only a modest rescue of adipocyte differentiation (Fig. 1 and 2), suggesting that the generation of eLOX3 substrates is diminished in the Rb^−/− MEFs. Indeed, these MEFs fail to induce expression of XOR upon administration of the differentiation cocktail (Fig. 6C). It was, however, impossible to examine a synergistic effect of eLOX3 and XOR on adipocyte differentiation as forced expression of XOR blocks adipogenesis (13). Collectively, these data suggest that eLOX3 and XOR cooperate in the activation of PPARγ during the early stages of adipocyte differentiation.

The antioxidant N-acetyl cysteine (NAC) inhibits adipose conversion.If the generation of PPARγ ligand(s) involved a ROS-mediated conversion of arachidonic acid into eLOX3 substrates, we hypothesized that scavenging ROS using an antioxidant compound would have an antiadipogenic effect. NAC is commonly used as a tool to remove ROS. As predicted, NAC abolished adipose conversion of 3T3-L1 preadipocytes as measured by oil red O and marker gene expression (Fig. 7A and B). Furthermore, the NAC-mediated inhibition was circumvented by the inclusion of an exogenous PPARγ ligand during differentiation. The decline in PPARγ mRNA levels by rosiglitazone was expected based on previous reports (11, 47). As PPARγ ligand(s) is produced early and transiently during adipocyte differentiation, NAC should be required early in order to abolish adipogenesis. Indeed, inclusion of NAC from day 0 to day 4 was sufficient to inhibit adipocyte differentiation (Fig. 7C and D). Furthermore, addition of NAC from day 2 onwards resulted in a modest inhibition, whereas inclusion from day 4 or day 6 to the end of differentiation had no effect on adipose conversion (data not shown). Collectively, these data indicate a requirement for early production of ROS for induction of adipocyte differentiation. Interestingly, the level of ROS was recently shown to increase during the early stages of adipocyte differentiation. It was suggested that ROS was important for activation of C/EBPβ and that antioxidants inhibited adipogenesis by blocking activation of C/EBPβ (32). However, the inclusion of the antioxidant NAC in the differentiation cocktail reduced C/EBPβ mRNA expression only marginally and augmented rather than inhibited induction of the C/EBPβ-responsive gene XOR (Fig. 7E), suggesting that ROS-dependent effects on C/EBPβ activity do not affect XOR expression.

• Open in new tab
• Download powerpoint

FIG. 7.

The antioxidant NAC abolishes adipose conversion. (A and B) 3T3-L1 preadipocytes were induced to differentiate according to the MDI protocol with the inclusion of rosiglitazone (0.5 μM) and/or NAC (5 mM) or vehicle alone. Differentiation was determined by oil red O staining (A) or expression of adipocyte marker genes (B). ACC, acetyl coenzyme A carboxylase; HSL, hormone-sensitive lipase; GLUT4, facilitated glucose transporter, member 4. Data represent two independent experiments performed in duplicate. (C and D) 3T3-L1 preadipocytes were induced to differentiate according to the MDI protocol with the inclusion of rosiglitazone (0.5 μM) and/or NAC (5 mM) or vehicle alone from day 0 to day 4. Differentiation was determined by oil red O staining (C) and expression of adipocyte marker genes (D). Data represent two independent experiments performed in duplicate. (E) 3T3-L1 cells were induced to differentiate in the presence and absence of 5 mM NAC. Expression of C/EBPβ and XOR was analyzed at indicated time points. Data represent two independent experiments performed in duplicate.