|Home | About | Journals | Submit | Contact Us | Français|
Pulmonary fibrosis is a progressive scarring disease with no effective treatment. Transforming growth factor (TGF)-β is up-regulated in fibrotic diseases, where it stimulates differentiation of fibroblasts to myofibroblasts and production of excess extracellular matrix. Peroxisome proliferator–activated receptor (PPAR) γ is a transcription factor that regulates adipogenesis, insulin sensitization, and inflammation. We report here that a novel PPARγ ligand, 2-cyano-3,12-dioxoolean-1,9-dien-28-oic acid (CDDO), is a potent inhibitor of TGF-β–stimulated differentiation of human lung fibroblasts to myofibroblasts, and suppresses up-regulation of α–smooth muscle actin, fibronectin, collagen, and the novel myofibroblast marker, calponin. The inhibitory concentration causing a 50% decrease in aSMA for CDDO was 20-fold lower than the endogenous PPARγ ligand, 15-deoxy-Δ12,14-prostaglandin J2 (15 d-PGJ2), and 400-fold lower than the synthetic ligand, rosiglitazone. Pharmacologic and genetic approaches were used to demonstrate that CDDO mediates its activity via a PPARγ-independent pathway. CDDO and 15 d-PGJ2 contain an α/β unsaturated ketone, which acts as an electrophilic center that can form covalent bonds with cellular proteins. Prostaglandin A1 and diphenyl diselenide, both strong electrophiles, also inhibit myofibroblast differentiation, but a structural analog of 15 d-PGJ2 lacking the electrophilic center is much less potent. CDDO does not alter TGF-β–induced Smad or AP-1 signaling, but does inhibit acetylation of CREB binding protein/p300, a critical coactivator in the transcriptional regulation of TGF-β–responsive genes. Overall, these data indicate that certain PPARγ ligands, and other small molecules with electrophilic centers, are potent inhibitors of critical TGF-β–mediated profibrogenic activities through pathways independent of PPARγ. As the inhibitory concentration causing a 50% decrease in aSMA for CDDO is 400-fold lower than that in rosiglitazone, the translational potential of CDDO for treatment of fibrotic diseases is high.
This article describes a novel mechanism by which a peroxisome proliferator–activated receptor γ agonist blocks myofibroblast differentiation in lung fibroblasts. This has applications for the design of novel treatments for lung fibrosis.
Idiopathic pulmonary fibrosis affects approximately 5 million people worldwide. The mechanism of disease initiation and progression is poorly understood, and the disease is refractory to current therapies (1). Fibrotic remodeling of lung tissue is also an important facet to the pathology of other lung diseases, including sarcoidosis and asthma (2). Fibrosis is hypothesized to be a defect in regulation of wound healing, resulting in the production of excess extracellular matrix components (3, 4). Fibroblasts are major effector cells in the development of pulmonary fibrosis; they differentiate to myofibroblasts and produce excess extracellular matrix components, including collagen and fibronectin (4–6). The increased extracellular matrix leads to alteration of normal lung architecture and impaired gas exchange.
The cytokine, transforming growth factor (TGF)-β1, induces the differentiation of fibroblasts to myofibroblasts in vitro (1, 6), and is implicated as a key cellular inducer of fibrotic diseases (reviewed in References 2, 7), including idiopathic pulmonary fibrosis (8–10). Overexpression of TGF-β in rodent lung leads to severe, irreversible pulmonary fibrosis (11, 12), whereas inhibiting TGF-β in vivo decreases the development and progression of pulmonary fibrosis (13–15). TGF-β gene transcription is initiated by binding of TGF-β to its receptor, which leads to phosphorylation of the signaling proteins, Smad2/3, recruitment of Smad4, and translocation of the Smad2/3/4 complex to the nucleus, where it interacts with coactivator proteins, such as CREB (cAMP response element binding protein) binding protein (CBP)/p300. CBP/p300 is a histone acetyl transferase (HAT) implicated in transcriptional regulatory pathways, including TGF-β (16).
Peroxisome proliferator–activated receptor (PPAR) γ is a transcription factor. After ligand binding, PPARγ heterodimerizes with the retinoid X receptor (RXR), and is translocated to the nucleus, where it binds to PPARγ response elements (PPREs) (14, 17). PPARγ agonists regulate fat metabolism and adipogenesis (18). Endogenous PPARγ ligands include prostaglandin 15-deoxy-Δ12,14-prostaglandin J2 (15 d-PGJ2), as well as lysophosphatidic acid and 15S-hydroxyeicosatetraenoic acid. The most well studied synthetic agonists are the thiazolidinedione class of drugs, such as rosiglitazone and pioglitazone, which are currently prescribed as insulin sensitizers in type II diabetes (19). PPARγ agonists have potentially important roles as anti-inflammatory and antifibrotic agents (20, 21), and are also involved in the regulation of cellular differentiation (22, 23). We and others have reported that PPARγ agonists, including thiazolidinediones, have antifibrotic effects on human lung fibroblasts in vitro (24, 25). Interestingly, rosiglitazone is reported to be 100-fold more effective at promoting fibroblast-to-adipocyte differentiation than 15 d-PGJ2 (26), but is much less effective than 15 d-PGJ2 at inhibiting differentiation of lung fibroblasts to myofibroblasts (24). This suggests that the proadipogenesis and antifibrotic properties of PPARγ ligands may act via different mechanisms.
2-Cyano-3,12-dioxoolean-1,9-dien-28-oic acid (CDDO) is a novel PPARγ ligand. Although both 15 d-PGJ2 and CDDO bind PPARγ with high affinities (Ki [inhibition constant] 1.2 × 10−9 and 10−8, respectively) (26, 27), and regulate adipogenesis in a PPARγ-dependent manner, there are emerging data that they mediate some of their biological activities via a PPARγ-independent pathway (22, 24, 28, 29). 15 d-PGJ2 is the final cyclopentenone metabolite of PGD2 in the cyclooxygenase pathway. The cyclopentenone prostaglandins have an α/β-unsaturated ketone ring with electrophilic carbons that are highly susceptible to Michael addition reactions, forming covalent bonds with free reduced thiols, such as glutathione and accessible cysteine residues of intracellular proteins (see Figure E3 in the online supplement) (30). For example, 15 d-PGJ2 mediates its anti-inflammatory effects in part via covalent modification of cysteine residues in the NF-κB components IkB and p65 (31). CDDO also has an α/β-unsaturated ketone, and is potentially reactive in modifying cellular proteins (22, 29, 32). Recent data from our laboratory and others suggest that the electrophilic carbon on 15 d-PGJ2 is important for meditating its PPARγ-independent effects (22, 29, 32).
Here, we report that CDDO has profound antifibrotic effects on human lung fibroblasts in vitro at 20-fold lower concentrations than 15 d-PGJ2 and 400-fold lower concentration than rosiglitazone. Moreover, the effects are independent of PPARγ-mediated transcription and dependent on the presence of a strong electrophilic center. This independent mechanism does not involve the classic TGF-β/Smad signaling pathway or the alternate activator protein (AP)-1 and mitogen-activated protein kinase pathways, but does involve alterations of TGF-β–mediated acetylation of CBP/p300.
Normal human lung fibroblast cell strains were derived from tissue explants and maintained in Eagle's minimum essential medium (Life Technologies, Gaithersburg, MD) supplemented with 10% FBS (Sigma, St. Louis, MO), 2 mM L-glutamine, penicillin (100 U/ml), streptomycin (100 μg/ml), and amphotericin (0.25 μg/ml; Invitrogen, Carlsbad, CA) at 37°C in 7% CO2. These explants were derived from anatomically normal lung tissue from patients undergoing surgical resection for benign hamartoma or for small peripheral nodules. Patient samples were obtained with informed consent under the approval of the Institutional Review Board of the University of Rochester. These cells are morphologically consistent with fibroblasts, they express collagen and vimentin, and they do not express CD45, factor VIII, or cytokeratin (33). Cells were used at low passages, typically 4–8. The PPARγ agonists, 15 d-PGJ2 (Biomol, Plymouth Meeting, PA), 9,10-dihydro-15-deoxy-Δ12,14-PGJ2 (CAY10410; Cayman Chemical Co., Ann Arbor, MI), and CDDO (NIH-RAID Program and Reata Pharmaceuticals, Dallas, TX), were prepared as 10-mM stocks in DMSO and added to cell cultures to the final concentrations indicated. GW9662 and prostaglandin A1 (PGA1; Cayman Chemical), and diphenyl diselenide (DSPS; Sigma) were prepared in the same manner. Recombinant human TGF-β1 was purchased from R&D Systems (Minneapolis, MN).
Primary human lung fibroblasts (100,000 cells/well) were plated in triplicate in six-well plates (Falcon/Becton Dickson, Franklin Lakes, NJ) and treated with TGF-β and/or PPARγ agonists, as described subsequently here. In some experiments, the fibroblasts were infected with a lentiviral vector expressing a dominant negative PPARγ or the empty vector (see below) for 24 hours before treatment with TGF-β and PPARγ agonists. Cell lysates were prepared using either an NP-40–based lysis buffer or the commercial ActiveMotif Nuclear Extract kit (Active Motif, Carlsbad, CA). Lysates containing 2 or 5 μg of protein were separated by 10% SDS-PAGE under reducing conditions, transferred to nitrocellulose membranes, and proteins of interest were detected by Western blotting and chemiluminescence (Western Lightning, Perkin-Elmer, Wellesley, MA). To analyze collagen and fibronectin levels, precast 4–15% SDS-PAGE gradient gels (Bio-Rad, Hercules, CA) were used. Lysates were examined for expression of α-smooth muscle actin (SMA; Sigma), calponin (DAKO, Carpinteria, CA), fibronectin (Sigma), Smad2/3, pSmad2/3, Smad4, or acetylated CBP/p300 (Cell Signaling, Beverly, MA) or collagen I (Santa Cruz, Carlsbad, CA). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH; Abcam, Cambridge, MA), β-actin (Abcam), and/or vinculin (Sigma) were used as loading controls. Densitometry of the resulting bands was performed using Kodak Molecular Imaging Software (Kodak, Rochester, NY), and normalized to the loading control.
RNA was isolated from homogenized primary human lung fibroblast cultures using RNeasy, according to the manufacturer's protocol (Qiagen, Valencia, CA). RNA (1.0 μg) was incubated with PCR buffer, 0.5 μg of oligo (dT)12–18 primer (Invitrogen), 10 mM deoxy-nucleotide-triphosphate (dNTP) for 10 minutes at 70°C and 5 minutes in ice water, followed by addition of 40 U of recombinant RNasin RNase inhibitor (Promega, Madison, WI), 0.1 mM DTT, and 200 U of Superscript III reverse transcriptase (RT; Invitrogen). The mixture was further incubated for 5 minutes at room temperature, 60 minutes at 50°C, and 15 minutes at 70°C. The reaction contents were diluted to 80-μl volume and stored at −20°C. Negative controls contained no RT enzyme.
Quantitative real-time RT-PCR reactions were performed using a Bio-Rad iCycler with SYBR Green Supermix (Bio-Rad) according to the supplier's recommended protocol, with the following modifications. For amplification of the collagen I (COL1A1) message, the reaction contained 4 mM MgCl2 and 0.1 μM of each primer, whereas, for collagen III (COL3A1), α-SMA (ACTA2), and GAPDH, the reactions contained 3 mM MgCl2 and 0.2 μM of each primer. Oligomer primers were ordered from Real Time Primers (Elkins Park, PA) (COL1A1) and Integrated DNA Technologies (Coralville, IA) (COL3A1, ACTA2, and GAPDH). The primer sequences were as follows: COL1A1 sense, 5′- GGA CAC AGA GGT TTC AGT GG-3′; COL1A1 anti-sense, 5′- CCA GTA GCA CCA TCA TTT CC-3′; GAPDH sense, 5′- AAG GTC GGA GTC AAC GGA TTT GGT-3′; GAPDH anti-sense, 5′- AGC CTT GAC GGT GCC ATG GAA TTT-3′; ACTA2 sense, 5′- TCT GGA GAT GGT GTC ACC CAC AAT-3′; ACTA2 anti-sense, 5′-AAT AGC CAC GCT CAG TCA GG-3′; COL3A1 sense, 5′-AGC AA AGT TTC CTC CGA GGC CAG-3′; COL3A1 antisense, 5′-GGA GAA TGT TGT GCA GTT TGC CCA-3′.
Primary human lung fibroblast strains were cultured and treated as described above. At harvest, cells were trypsinized, washed twice with PBS, and fixed with 100% methanol for 30 minutes at 4°C. The cells were then washed twice with PBS/0.1% Tween-20 (PBS-T) and blocked with 5% normal mouse serum for 20 minutes at room temperature. Blocked cells were washed once with PBS-T and incubated with fluorescently labeled (Cy3) mouse anti–α-SMA antibody (Sigma) at a 1:400 dilution in PBS-T plus 1% BSA, overnight at 4°C. The next day, cells were washed twice with PBS-T, once with PBS, and stained with Draq5 (Exira, San Diego, CA) at a 1:200 dilution in PBS. Samples were analyzed directly on an Imagestream System 100 (Amnis Corporation, Seattle, WA). A gating strategy was employed to discriminate live single cells from debris and clumps of cells based on the intensity and aspect ratio of nuclear staining (Draq5). Single cells were identified as fibroblasts or myofibroblasts based on the intensity and fractional area of α-SMA staining. See the online supplemental Materials and Methods for more details.
Plasmids encoding a Flag-tagged wild-type PPARγ (pcDNA-Flag-PPARγ1) and a Flag-tagged dominant-negative PPARγ (pcDNA-Flag-PPARγ1–L466A/E469A) were a kind gift from V.K.K. Chatterjee (University of Cambridge, Cambridge, UK). The cDNA sequences were amplified by PCR using primers that added a restriction site (NheI and NotI) at each end. The PCR product was then digested with NheI and NotI and subcloned into a NheI-NotI–digested pCDH1-MCS1-EF1-copGFP vector (System Bioscience, Mountain View, CA). These vectors were named LV-PPARγ and LV-PPARγ-DN, and the vector without PPARγ cDNA was termed LV-empty.
Human embryonic kidney-293FT cells (Invitrogen) were grown to 50–70% confluency in Dulbecco's modified Eagle's medium (GIBCO, Carlsbad, CA) supplemented with 10% FBS in T-175 flasks. Subsequently, the vesicular stomatitis virus G (VSV-G)-pseudotyped HIV vector was generated by cotransfecting with 5 μg of envelope vector (pomp–VSV-G), 14 μg of transfer vector, and 14 μg of packaging vector (pomp-Δ89.2), using Lipofectamine LTX (Invitrogen) following the manufacturer's recommendations. Cells were split into two new T-175 flasks 6 hours after transfection. Supernatants were collected 48 and 72 hours after transfection. Virus was harvested by ultracentrifugation at 50,000 × g for 2 hours at 4°C using a Beckman SW28 rotor (Beckman Coulter, Inc., Fullerton, CA). The concentrated virus stocks were titered on human embryonic kidney-293FT cells based on green fluorescent protein (GFP) expression.
Primary lung fibroblasts cultured in six-well plates were cotransfected using Fugene6 (Roche Applied Science, Indianapolis, IN) with an AP-1 luciferase reporter (a gift from Dr. Sanjay Maggirwar, University of Rochester) (34) and a CMV–β-galactosidase control construct (a gift from Dr. T. Gasiewicz, University of Rochester) (35). After 24 hours, the cells were washed and then treated with TGF-β (5 ng/ml) with or without 15 d-PGJ2 (5 μM), or CDDO (1 μM) in medium (10% FBS), and harvested after a further 24-hour incubation. Luciferase activity was measured using a luciferase assay system (Promega) in a luminometer (Packard Instruments, Meriden, CT) and normalized to β-galactosidase activity, determined by a colorimetric assay (Promega). The experiments were performed in triplicate wells.
Cell viability was assessed by 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide assay (36). Fibroblasts were plated in triplicate at a density of 5,000 cells per well in 96-well plates, and treated with TGF-β and PPARγ agonists for 72 hours at the same concentrations previously used here. Production of the colored reaction product was measured at 560 nm, and the results were normalized to the negative control wells.
All data are expressed as means (±SD). A Student's unpaired t-test and ANOVA were used to establish statistical significance. Results were considered significant if P was less than 0.05.
To determine the efficacy of CDDO, 15 d-PGJ2, and rosiglitazone at inhibiting TGF-β–driven differentiation of fibroblasts to myofibroblasts, human lung fibroblasts were treated with 5 ng/ml of TGF-β and cotreated with either CDDO, 15 d-PGJ2, or rosiglitazone. The maximum concentrations used (5 μM 15 d-PGJ2, 1μM CDDO, and 20 μM rosiglitazone) were chosen for having the strongest inhibitory effect on TGF-β-stimulated α-SMA up-regulation without affecting cell viability (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide assay; data not shown). Myofibroblast differentiation was assessed using the common myofibroblast marker, α-SMA, by immunocytochemistry, Western blot, Imagestream flow cytometry, and RT-PCR. In addition, the expression of a novel myofibroblast marker, calponin, was examined. Calponin is a calmodulin-, F-actin–, and tropomyosin-binding protein that is involved in smooth muscle contraction (37). CDDO, 15 d-PGJ2, and rosiglitazone all inhibited myofibroblast differentiation, as determined by Western blot for α-SMA (Figure 1A) and by immunocytochemistry (Figure E1). CDDO and 15 d-PGJ2 also efficiently inhibited the novel pulmonary myofibroblast marker, calponin (Figure 1B). The inhibitory concentration causing a 50% decrease in aSMA CDDO was 20-fold lower than that for 15 d-PGJ2 with respect to α-SMA expression (0.04 and 0.7 μM, respectively) (Figure 1A), and sixfold lower with respect to calponin (0.2 and 1.3 μM, respectively) (Figure 1B and data not shown). We also measured α-SMA mRNA levels. Similar to the protein data, α-SMA mRNA was induced by TGF-β, and was significantly decreased by CDDO and 15 d-PGJ2 (Figure 1C). By all three measures of myofibroblast differentiation, rosiglitazone was much less effective at inhibiting TGF-β–driven myofibroblast differentiation than either CDDO or 15 d-PGJ2.
We also examined lung fibroblasts treated with TGF-β and CDDO or 15 d-PGJ2 by Imagestream analysis, a new flow cytometry technique that images cells in addition to determining their fluorescence emissions. We have refined this technique and used intracellular staining to identify single-cell expression and changes in α-SMA staining with TGF-β treatment. This technique is valuable, because it is an objective measure of α-SMA on an individual cell basis. Myofibroblasts were distinguished from fibroblasts based on the area and intensity of α-SMA staining, as described in Materials and Methods. TGF-β induced myofibroblast differentiation; after 72 hours, 35% of lung fibroblasts were defined as myofibroblasts by our gating strategy. Both 15 d-PGJ2 and CDDO inhibited this differentiation by approximately 50%. (Figure 2 and Figure E2).
We examined the effects of 15 d-PGJ2 and CDDO on extracellular matrix production, an important profibrotic effector function of lung fibroblasts and myofibroblasts, by determining the expression of fibronectin and collagen. Both CDDO and 15 d-PGJ2 suppressed TGF-β–induced fibronectin expression, as determined by Western blotting (Figure 3A). In addition, TGF-β stimulated up-regulation of collagen I and III protein (data not shown), and mRNA (Figures 3B and 3C) was significantly decreased by CDDO as well as 15 d-PGJ2 and rosiglitazone.
We used complimentary pharmacological and genetic approaches to analyze the mechanism of inhibition of TGF-β–stimulated α-SMA expression by CDDO and 15 d-PGJ2. GW9662 is a highly specific PPARγ antagonist that covalently binds to a cysteine residue within the binding pocket of PPARγ, permanently altering its ability to bind PPARγ ligands. PPARγ agonist–driven adipogenesis is PPARγ dependent, and is completely inhibited by GW9662 (18). Primary lung fibroblasts were pretreated with GW9662, followed by the addition of TGF-β and either CDDO or 15 d-PGJ2. After 72 hours, myofibroblast differentiation was determined by Western blot for α-SMA. GW9662 did not prevent the inhibition of TGF-β–stimulated myofibroblast differentiation by CDDO or 15 d-PGJ2 (Figure 4A).
Using a genetic approach, we overexpressed a dominant-negative PPARγ (LV-PPARγ-DN) in primary human lung fibroblasts using a lentiviral vector that results in stable coexpression of the DN PPARγ and GFP. Greater than 90% of cells infected with the construct expressed GFP and the Flag-tagged DN PPARγ (data not shown). Overexpression of LV-PPARγ-DN did not significantly reduce the ability of either CDDO or 15 d-PGJ2 to inhibit TGF-β–induced myofibroblast differentiation. α-SMA expression was 45% inhibited by 15 d-PGJ2 in fibroblasts infected with LV-PPARγ-DN compared with 39% inhibition in fibroblasts infected with the control vector (LV-empty). Similarly, CDDO reduced α-SMA expression by 75% in LV-PPARγ-DN–infected fibroblasts compared with 77% in the control cells (Figure 4B). Interestingly, LV-PPARγ-DN plus TGF-β resulted in a strong up-regulation of α-SMA compared with fibroblasts treated with LV-empty and TGF-β. This suggests that, in the absence of exogenous PPARγ ligands, α-SMA expression and myofibroblast differentiation are regulated by the PPARγ transcription factor, perhaps involving endogenous or media-derived PPARγ ligands. However, as the LV-PPARγ-DN did not alter the ability of 15 d-PGJ2 and CDDO to inhibit α-SMA expression, their effects must be due largely to PPARγ-independent effects.
CDDO and 15 d-PGJ2 contain an α/β-unsaturated ketone that can form covalent bonds with free sulfhydryls, such as occur in glutathione and accessible cysteine residues in cellular proteins (Figure E3) (38, 39). There are reports that the PPARγ-independent effects of 15 d-PGJ2 may be mediated via this active site (31). CAY10410 (9,10-dihydro-15-deoxy-delta12,14-PGJ2) is an analog of 15 d-PGJ2 that lacks the unsaturated ketone, and is resistant to metabolism by glutathione addition (39). CAY10410 can activate PPRE-dependent transcription (40), but has significantly diminished PPARγ-independent activity compared with 15 d-PGJ2 (22). CAY10410 is a much less effective inhibitor of TGF-β–stimulated myofibroblast differentiation than 15 d-PGJ2 at the same concentration (Figure 5A), demonstrating that the electrophilic center of 15 d-PGJ2 is important in mediating this bioactivity. To further investigate the importance of the electrophilic center, we tested two potent electrophiles, DSPS and PGA1. Both of these compounds significantly reduced TGF-β–induced α-SMA expression (Figure 5B).
Our data suggest that CDDO and 15 d-PGJ2 mediate their antifibrotic effect by a PPARγ-independent pathway. One mechanism by which these ligands may mediate their activity is by interference with TGF-β–mediated Smad signaling. Primary human lung fibroblasts were pretreated with CDDO and 15 d-PGJ2 for 1 hour, then treated with TGF-β and harvested after 15 or 30 minutes. Total and phospho-Smad2/3 were determined in whole-cell lysates and nuclear extracts, and Smad4 was determined in nuclear extracts by Western blot. TGF-β treatment of human primary lung fibroblasts resulted in rapid phosphorylation of Smad2/3 within 15 minutes, with a slight further increase by 30 minutes. There was no significant difference in Smad2/3 phosphorylation between TGF-β alone and TGF-β plus 15 d-PGJ2 or CDDO (Figure 6A). TGF-β treatment also induced rapid nuclear translocation of Smad4 that was unaffected by cotreatment with CDDO or 15 d-PGJ2 (Figure 6B). Another important TGF-β–mediated signaling pathway involves the transcription factor, AP-1 (41). Primary lung fibroblasts were transfected with an AP-1 reporter construct and treated with TGF-β and 15 d-PGJ2 or CDDO. TGF-β alone induced an eightfold increase in AP-1 activity. Interestingly, both 15 d-PGJ2 and CDDO alone also significantly induced AP-1 activity, and the combination of TGF-β and CDDO was twice as potent as TGF-β alone (Figure 6C), demonstrating that 15 d-PGJ2 and CDDO do not inhibit myofibroblast differentiation by inhibiting AP-1 signaling.
Recruitment of CBP/p300 to the Smad2/3/4 complex is required for maximal TGF-β–driven up-regulation of α-SMA (42). Primary human lung fibroblasts were treated with TGF-β alone or with TGF-β plus 1 μM CDDO for 72 hours, then harvested for Western blotting. TGF-β treatment resulted in significant acetylation of CBP/p300 that was completely inhibited by CDDO (Figure 7).
PPARγ ligands have both anti-inflammatory and antiscarring effects. In primary human lung fibroblasts, rosiglitazone and 15 d-PGJ2 inhibit expression of α-SMA and collagen (24, 25). Others have demonstrated in animal models of lung fibrosis, as well as in other organs, that PPARγ ligands have antifibrotic effects (20, 25, 43). Here, we report that CDDO is a novel antifibrotic agent that potently inhibits TGF-β–stimulated differentiation of human lung fibroblasts to myofibroblasts via a largely PPARγ-independent pathway, with an inhibitory concentration causing a 50% drop (40 nM) that is 20-fold lower than 15 d-PGJ2 and 400-fold lower than rosiglitazone.
Because of the exciting potential of PPARγ ligands as antifibrotic agents, it is important to understand their mechanism of action. The anti-inflammatory and antifibrotic effects of the both synthetic and natural PPARγ ligands have been reported to involve both PPARγ-dependent and -independent mechanisms (22, 28, 44). The PPARγ-dependent mechanism requires ligand binding to PPARγ, heterodimerization with RXR and its ligand, nuclear translocation, binding to the PPRE, and induction of gene transcription, whereas the PPARγ-independent mechanism or mechanisms have not yet been fully determined.
To investigate whether the antifibrotic effects of CDDO and 15 d-PGJ2 are PPARγ dependent or independent, we used a small molecule inhibitor of PPARγ and overexpression of a dominant-negative PPARγ construct. GW9662, an irreversible PPARγ antagonist, did not restore myofibroblast differentiation in fibroblasts treated with TGF-β and CDDO or 15 d-PGJ2 (Figure 4A). Similarly, when lung fibroblasts were transfected with LV-PPARγ-DN, CDDO and 15 d-PGJ2 were still able to significantly inhibit TGF-β–driven myofibroblast differentiation (Figure 4B). Interestingly, when PPARγ activity was blocked with either GW9662 or the DN construct, treatment with TGF-β resulted in significantly greater induction of α-SMA than in cells with intact PPARγ signaling cascades treated with TGF-β. This suggests that, in the absence of potent exogenous agonists, PPARγ opposes TGF-β–stimulated myofibroblast differentiation via a PPARγ-dependent mechanism involving endogenous or media-derived PPARγ ligands, such as eicosanoids and other fatty acids (21, 45). Blocking PPARγ function removes this opposing regulatory pathway, increasing the profibrotic effects of TGF-β. Note that, although rosiglitazone is a strong PPARγ-dependent promoter of adipogenesis, it exhibits little PPARγ-independent activity, and is a relatively poor inhibitor of myofibroblast differentiation compared with other agents (24, 26, 46, 47). We hypothesize that myofibroblast differentiation is regulated by both PPARγ-dependent and PPARγ-independent pathways, but that the PPARγ-independent mechanisms are much stronger, and compounds that access the independent pathways are more effective at inhibiting TGF-β–stimulated myofibroblast differentiation at lower concentrations. Therefore, compounds that access PPARγ-independent pathways are likely to have greater clinical potential as antifibrotic agents.
Inspection of the structures of the PPARγ ligands suggests that the strong PPARγ-independent effects of 15 d-PGJ2 and CDDO might be mediated by their electrophilic α/β-unsaturated ketone. These compounds are susceptible to Michael addition reactions by nucleophiles of free sulfhydryls, such as occur in glutathione and accessible cysteine residues in cellular proteins (Figure E3B) (30, 31), and it has been reported that 15 d-PGJ2 modifies the activity of the estrogen receptor by covalently binding to its DNA binding domain (48). We tested this hypothesis using CAY10410, an analog of 15 d-PGJ2 that lacks the α/β-unsaturated ketone and is unable to undergo this reaction (Figure E3A). CAY10410 was a much less effective inhibitor of α-SMA expression than 15 d-PGJ2 at the same concentration (Figure 5A). We further demonstrated the importance of the electrophilic carbon by using two other strong electrophilic compounds (DSPS and PGA1), which potently decreased TGF-β–induced expression of α-SMA (Figure 5B). These new findings support the concept that the electrophilic carbon has a central role in mediating inhibition of TGF-β–stimulated myofibroblast differentiation by a PPARγ-independent pathway.
PPARγ-independent effects of PPARγ ligands have been reported in several papers; however, no common pathway applicable to all cell types has been identified. For example, 15 d-PGJ2 decreases Smad2/3 phosphorylation in human hepatic stellate cells (49), whereas it inhibits nuclear localization without affecting phosphorylation in human glomerular mesangial cells (50). Our data demonstrate that, in primary human lung fibroblasts, neither 15 d-PGJ2 nor CDDO decrease TGF-β–induced phosphorylation of Smad2/3, or alter Smad localization. It has been reported that Smad3, but not Smad2, is required for the fibrotic response (51, 52); however, the antibodies used is this study detect both Smad2 and Smad3. Therefore, a role for Smad3 alone is not completely ruled out.
Alternately, pioglitazone inhibits profibrotic changes in mouse glomerular mesangial cells by suppressing AP-1 activity (41). However, neither 15 d-PGJ2 nor CDDO inhibits AP-1 activation in human lung fibroblasts (Figure 6C). In fact, both ligands alone activated an AP-1 reporter, and cotreatment with TGF-β and either ligand resulted in more AP-1 activity than TGF-β alone (Figure 6C). We have found that 15 d-PGJ2 and CDDO induce heme oxygenase-1 (HO-1), an enzyme involved in protecting cells from oxidative stress (H. E. Ferguson and colleagues, unpublished data), and it has been reported that HO-1 induces AP-1 activity (53). We hypothesize that, because 15 d-PGJ2 and CDDO are strong electrophiles and can react with and consume glutathione (Figure E3), they induce oxidative stress and HO-1 in lung fibroblasts, leading to enhanced AP-1 activity.
We report here the novel finding that CDDO inhibits acetylation of CBP/p300. CBP/p300 is a critical transcriptional cofactor, the activity of which is regulated in part by acetylation (54). CBP/p300 contains a HAT domain that can acetylate the core histones, H3 and H4, which relaxes the chromatin, making it accessible to other transcription factors. The HAT domain of CBP/p300 can acetylate other transcription factors as well, and it has been shown that acetylation of Smad3 is required for the expression of TGF-β–responsive genes, including α-SMA (16, 42, 55). Upon binding of TGF-β to its receptor, Smad3 is phosphoryated, recruits Smad4, and the Smad3/4 complex is translocated to the nucleus, where it binds to the promoters of TGF-β–responsive genes (Figure E4). Smad3 recruits binding of the CBP/p300 cofactor, which becomes acetylated itself by an unknown mechanism. The HAT domain of CBP/p300 acetylates Smad3, which is required for transcriptional activation, and the core histones, H3 and H4, which converts the chromatin from a closed, transcriptionally inactive conformation to an open, transcriptionally active conformation. Acetylation of CBP/p300 results in enhanced HAT activity (54). We have shown that CDDO dramatically inhibits acetylation of CBP/p300 (Figure 7). Therefore, we hypothesize that CDDO inhibits myofibroblast differentiation by inhibiting acetylation of CBP/p300, which, in turn, inhibits acetylation of Smad3 and histones, thereby blocking transcription from TGF-β–responsive genes.
Overall, we have demonstrated for the first time that PPARγ ligands, such as CDDO and other small molecules with electrophilic carbons, are potent inhibitors of TGF-β–induced myofibroblast differentiation and extracellular matrix production in human lung fibroblasts that act via a PPARγ-independent pathway. This pathway may involve a novel mechanism of action for CDDO, inhibiting α-SMA expression by dysregulating acetylation of CBP/p300. CDDO is an orally active agent, with a long half-life, and is in clinical trials for the treatment of several forms of cancer, including leukemia, solid tumors, and myelomas. The translational potential of this compound is thus high. Our results also suggest that developing other electrophilic compounds that target PPARγ-independent pathways may be of great clinical utility for scarring diseases of the lungs and other tissues.
This work was supported in part by National Institutes of Health grants HL75432, DE011390, ES01247, HL095402, EY017123, EY015836, T32-HL066988, T32-HL007152, ES07026, HL75432-0452, and HL75432-0451.
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
Originally Published in Press as DOI: 10.1165/rcmb.2009-0006OC on March 13, 2009
Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript.