|Home | About | Journals | Submit | Contact Us | Français|
Previous work has provided strong evidence for a role of peroxisome proliferator-activated receptor β/δ (PPARβ/δ) and transforming growth factor-β (TGFβ) in inflammation and tumor stroma function, raising the possibility that both signaling pathways are interconnected. We have addressed this hypothesis by microarray analyses of human diploid fibroblasts induced to myofibroblastic differentiation, which revealed a substantial, mostly reverse crosstalk of both pathways and identified distinct classes of genes. A major class encompasses classical PPAR target genes, including ANGPTL4, CPT1A, ADRP and PDK4. These genes are repressed by TGFβ, which is counteracted by PPARβ/δ activation. This is mediated, at least in part, by the TGFβ-induced recruitment of the corepressor SMRT to PPAR response elements, and its release by PPARβ/δ ligands, indicating that TGFβ and PPARβ/δ signals are integrated by chromatin-associated complexes. A second class represents TGFβ-induced genes that are downregulated by PPARβ/δ agonists, exemplified by CD274 and IL6, which is consistent with the anti-inflammatory properties of PPARβ/δ ligands. Finally, cooperative regulation by both ligands was observed for a minor group of genes, including several regulators of cell proliferation. These observations indicate that PPARβ/δ is able to influence the expression of distinct sets of both TGFβ-repressed and TGFβ-activated genes in both directions.
Peroxisome proliferator-activated receptors (PPARs) are nuclear receptors that function as ligand-inducible transcription factors (1–3). The three PPAR subtypes (PPARα, PPARβ/δ and PPARγ) activate their target genes through binding to PPAR response elements (PPREs) as heterodimers with members of the retinoid X receptor (RXR) family. PPARs play a central role in lipid metabolism by serving as sensors for fatty acids and fatty acid metabolites with major function as modulators of metabolic and inflammatory processes. Consequently, the transcriptional activity of PPARs is modulated not only by natural fatty acids, but also by lipid-derived metabolites such as prostaglandins J2 and I2, leukotriene A4, 15-hydroxyeicosatetraenoic acid and 1-palmitoyl-2-oleoyl-sn-glycerol-3-phosphocholine (4–7). PPARs also play essential roles in developmental processes, wound healing, cell differentiation and proliferation and many associated diseases, including arteriosclerosis, diabetes, fibrosis, inflammatory disorders and cancer (8–12), which led to the development of numerous subtype-selective, high-affinity ligands (13).
We and others have shown that PPARβ/δ plays an essential role in regulating the differentiation, function and proliferation of tumor stroma cells (14–16). Ppard-null mice show gross alterations of tumor endothelial cells and fibroblasts, resulting in a high proportion of immature, dysfunctional microvessels and increased numbers of myofibroblastic cells (14). Consistent with these in vivo data, overexpression of PPARβ/δ inhibited the proliferation of cultured fibroblasts (14). Likewise, the prostacyclin mimetic Treprostinil inhibited the proliferation of lung fibroblasts concomitant with the transcriptional activation of PPARβ/δ (17). A regulatory role for PPARβ/δ in myofibroblasts has also been shown in a cell culture model of cardiac fibrosis, i.e. neonatal rat cardiac fibroblasts induced to myofibroblast transdifferentiation by culturing on a rigid substrate (18). Finally, different PPAR subtypes have been shown to play a role in experimentally induced lung fibrosis, and it has been suggested that PPARβ/δ agonists may attenuate disease progression by inhibiting myofibroblast proliferation and function (19).
A cytokine present in vast amounts in many tumors and playing a pivotal role in both tumor stroma function, inflammation and tissue fibrosis is the transforming growth factor-β (TGFβ) (20), suggesting that PPARβ/δ and TGFβ signaling pathways may functionally interact. To test this hypothesis, we performed microarray analyses of human lung fibroblasts induced to differentiate into myofibroblastic cells by TGFβ and analyzed the influence of PPARβ/δ agonists on the transcriptional profile. This study revealed an extensive, mainly reverse crosstalk of the transcriptional pathways regulated by PPARβ/δ and TGFβ, leading to the definition of distinct classes of genes. Class A genes are repressed by TGFβ, which is, at least in part, due to the induction of the corepressor SMRT and is counteracted by PPARβ/δ agonists. These include many known PPAR target genes with functions in lipid metabolism. A prominent example is the ANGPTL4 gene, which encodes an important regulator of lipid metabolism and presumptive modulator of metastasis (21,22). In contrast, class B genes are induced by TGFβ and downregulated by PPARβ/δ agonists. These genes include IL6, which may be relevant in view of the reported anti-inflammatory and anti-fibrotic properties of PPARβ/δ.
TGFβ1 was purchased from Sigma-Aldrich (Karlsruhe, Germany), GW501516, GW1929 and GW7647 from Axxora (Lörrach, Germany), and L165,041 from Calbiochem (Merck, Darmstadt, Germany).
WI-38 cells were obtained from the ATCC and maintained in DMEM/MCDB105 (1:1, PAA, Cölbe, Germany/Sigma, Steinheim, Germany) supplemented with 10% fetal bovine serum, 100U/ml penicillin and 100µg/ml streptomycin in a humidified incubator at 37°C and 5% CO2. Differentiation by TGFβ1 was carried out in serum-free medium as described (23,24).
Cells were fixed with ethanol (70%), stained by indirect immunofluorescence using a polyclonal α-SMA antibody (Sigma, Steinheim, Germany) visualized by a Cy5-labeled secondary antibody (Molecular Probes A11029, Invitrogen, Karlsruhe, Germany), and counterstained with Hoechst 33258 (Invitrogen). Slides were evaluated with a Leica RMB 3 microscope equipped with fluorescence optics. For quantitative evaluation of SMA stress fibers detected by immunofluorescence, cells showing strong, weak or no staining were counted separately. A total of ~750 cells in eight microscopic fields were counted per sample.
Cells were seeded at a density of 5×105 cells per 6cm dish in 4ml DMEM with 10% fetal calf serum (FCS) and cultured for 2h. 1280ng small-interfering RNA (siRNA) in 100µl OptiMEM (Invitrogen) and 20µl HiPerfect (Qiagen, Hilden, Germany) were mixed and incubated for 5–10min at room temperature prior to transfection. The cells were replated 24h post-transfection at a density of 5×105 cells per 6cm dish. Transfection was repeated 48h after start of the experiment, and cells were passaged after another 24h. Forty-eight hours following the last transfection, cells were incubated in serum-free medium for 24h before stimulation. The NCOR2 siRNA pool was composed of the following sequences:
control siRNA (#1022563, Qiagen, Hilden, Germany).
Complementary DNA (cDNA) was synthesized from 0.1–1µg of RNA using oligo(dT) primers and the Omniscript kit (Qiagen, Hilden, Germany). Quantitative polymerase chain reaction (qPCR) was performed in a Mx3000P Real-Time PCR system (Stratagene, La Jolla, CA, USA) for 40 cycles at an annealing temperature of 60°C. PCR reactions were carried out using the Absolute QPCR SYBR Green Mix (Abgene, Hamburg, Germany) and a primer concentration of 0.2µM following the manufacturer’s instructions. L27 was used as normalizer. Comparative expression analyses were statistically analyzed by Student’s t-test (two-tailed, equal variance) and Bonferroni correction. The sequences of the primers are as follows:
Chromatin immunoprecipitation (ChIP) was performed as described (6), except that nuclei were resuspended at 2.5×107/ml, and 60 pulses were applied during sonication. The following antibodies were used: IgG pool, I5006 (Sigma-Aldrich, Steinheim, Germany), α-PPARβ/δ, sc-7197 (Santa Cruz, Heidelberg, Germany); α-SMRT, ab24551 (Abcam, Cambridge, UK). Comparative binding analyses were statistically analyzed by Student’s t-test (two-tailed, equal variance) and corrected for multiple hypothesis testing by the Bonferroni method. Primer sequences were as follows:
RNA was isolated using the Nucleospin RNA II kit (Macherey-Nagel, Düren, Germany). RNA quality was assessed using the Experion automated electrophoresis station with RNA StdSens chips (Bio-Rad, Munich, Germany). For microarray studies, total RNA samples were amplified and labeled using the Agilent Quick Amp Labeling Kit (Agilent, Santa Clara, CA, USA) according to the manufacturer's instructions. The amplification procedure consists of reverse transcription of total RNA, including spike-in with an oligo(dT) primer bearing a T7 promoter, followed by in vitro transcription of the resulting cDNA with T7 RNA polymerase in the presence of dye labeled CTP to generate multiple fluorescence labeled copies of each messenger RNA (mRNA). After purification, the labeled aRNA was quantified and hybridization samples were prepared according to the manufacturer's instructions. Human Agilent 4-plex Array 44K were used for the analysis of the gene expression of the different samples in a reference-design assay. As a reference, a pool of all samples was used. This reference was labeled with Cy3, while the samples were labeled with Cy5 dye. The hybridization assembly was performed as described in the Agilent Microarray Hybridization Chamber User Guide (G2534-90001). After a 17-h hybridization at 65°C, slides were washed as described by the manufacturer and subsequently scanned using an Agilent DNA Microarray Scanner G2505C; scan software: Agilent Scan Control Version A.8.1.3; quantification software: Agilent Feature Extraction Version 10.5.1.1 (FE Protocol GE_105_Dec08). Raw microarray data were normalized using the ‘loess’ method implemented within the marray package of R/BioConductor (www.bioconductor.org). Regulated probes were selected on the basis that the logarithmic (base 2) average intensity value was ≥6, and that the fluctuation between replicates was ≤50%.
The purpose of the present study was to investigate whether PPARβ/δ and TGFβ signaling pathways functionally interact. As an experimental model, we used diploid human lung fibroblasts (WI38 cells) induced by TGFβ to differentiate into myofibroblast-like cells. In order to characterize this system, we first studied the expression of the myofibroblast marker genes ACTA2 (coding for smooth muscle α-actin; SMA), COL4A1 (encoding collagen type IV α1) and SM22A (coding for smooth muscle protein 22-α). As shown in Figure 1A and B, TGFβ induced the expression all three genes. Increased levels of ACTA2 and COL4A1 mRNA were detectable after 6h and reached maximum levels after 24–36h (Figure 1A). In the same experimental setup, no significant effect of the PPARβ/δ agonists GW501516 or L165,041 on the TGFβ-mediated induction of ACTA2, COL4A1 and SM22A was detectable (Figure 1B), suggesting that the ligand-mediated activation of PPARβ/δ does not affect the myofibroblastic differentiation of WI38 cells.
Concomitantly with the induction of these marker genes, SMA-containing stress fibers, a hallmark of differentiating myofibroblasts, were readily detectable after 24h exposure of WI38 cells to TGFβ (Figure 1C). Consistent with the marker gene expression data in Figure 1B, treatment with the PPARβ/δ agonist GW501516 had no detectable effect on stress fiber formation by TGFβ (Figure 1D).
As the deletion of Ppard in mice has been associated with myofibroblastic differentiation in the tumor stroma, we also investigated whether the inhibition of PPARβ/δ expression in WI38 cells might affect the differentiation status of these cells. Supplementary Figure S1 shows that ACTA2 expression indeed increased after the siRNA-mediated knockdown of PPARβ/δ. Taken together, these observations suggest that PPARβ/δ plays a role in preventing myofibroblastic transdifferentiation under basal conditions, but that its activation by ligands does not prevent TGFβ-induced differentiation.
The fact that PPARβ/δ ligands do not affect the TGFβ-induced differentiation of WI38 cells makes this experimental system suitable to study possible interactions of these signaling pathways in myofibroblasts without interference by an altered differentiation state. Such interactions could, for instance, affect the functional activation or metabolic activity of these cells. We therefore used this model to address two questions: (i) does TGFβ alter the regulation of PPARβ/δ target genes, and (ii) do PPARβ/δ ligands impinge on TGFβ-mediated transcriptional signaling events that are associated with, for instance, inflammatory or fibrotic responses.
To identify potential functional interactions between TGFβ and PPARβ/δ signaling pathways, we performed microarray analyses of WI38 cells, either untreated (solvent) or treated with GW501516 (0.3µM), TGFβ1 (2ng/ml) or both ligands for 24h (EMBL-EBI ArrayExpress, accession number E-MEXP-2750). As illustrated by the Venn diagram in Figure 2A, 5039 probes indicated regulation by TGFβ and 143 probes regulation by GW501516 (≥1.3-fold change) with an overlap of 117 probes. These correspond to 74 different annotated genes regulated by both ligands.
To determine cooperative or antagonistic effects exerted by TGFβ and GW501516, we compared for individual genes the transcriptional outcome of exposing WI38 cells to both ligands to that of treatment with either ligand alone, as described in the following sections.
The effect of GW501516 on TGFβ-mediated regulation was determined by plotting the relative expression levels measured after co-treatment with both ligands against the expression levels measured after treatment with TGFβ alone. The dot plot in Figure 2B identifies different set of probes showing distinct responses to TGFβ and GW501516.
‘Class A’ probes, which represent the major group defined in the present study, indicate repression by TGFβ that is counteracted by GW501516. This pattern was observed for a total of 136 probes, including 122 different annotated genes (cutoff ≥1.3-fold upregulation by GW501516; red data points in Figure 2B; Supplementary Table S1). The characteristic expression pattern of class A genes in response to TGFβ and GW501516 is shown in Figure 3A, and validated by RT-qPCR (Figure 4) for ANGPTL4 (angiopoietin-like 4), PDK4 (pyruvate dehydrogenase kinase 4), CPT1A (carnitine palmitoyltransferase 1A) and ADRP (adipose differentiation-related protein). Several representative genes of this class are listed in Table 1.
‘Class B’ probes indicate a counteractive effect of GW501516 on TGFβ-mediated activation. This class encompasses 22 probes, representing 21 annotated genes (cutoff ≥1.3-fold difference for TGFβ plus GW501516 relative to TGFβ alone; blue data points in Figure 2B; Supplementary Table S1). Their characteristic expression pattern in response to TGFβ and GW501516 is shown in Figure 3B. The RT-qPCR data in Figure 5 confirm that PPARβ/δ activation counteracts the TGFβ-mediated induction of the class B genes IL6 (interleukin-6), CD274 (B7-H1) and CLDN1 (claudin 1), which was clearly detectable 6h after application of GW501516, pointing to a direct effect of the PPARβ/δ ligands. No effect on the TGFβ-mediated induction of IL6 was seen with the PPARγ ligand GW1929 or the PPARα agonist GW7647 (Figure 5D), suggesting that the observed effect is PPARβ/δ-specific.
Cooperative regulation was also detectable for several probes (Figure 2B; not highlighted; class C and D in Supplementary Table S1), suggesting that GW501516 is able to influence the expression of distinct sets of both TGFβ-repressed and TGFβ-activated genes in both directions. Class C includes KIT, FOXQ1 and TOP2A, which code for the tyrosine kinase receptor KIT, the transcription factor forkhead box Q1 and topoisomerase II, respectively. All three genes have been associated with cell cycle progression and tumorigenesis and may thus be of particular interest with respect to the function of TGFβ and PPARβ/δ in tumor and tumor stroma cells.
We next determined for individual probes the effect of TGFβ on GW501516 inducibility. This was achieved by plotting the induction by GW501516 in the presence of TGFβ (fold GW501516 plus TGFβ/TGFβ alone) against the induction by GW501516 in the absence of TGFβ (Figure 2C). The predominant probe set identified by this analysis indicates increased induction (≥1.3-fold) by GW501516 in the presence of TGFβ (shown as triangles in Figure 2C). Surprisingly, a substantial number of these probes are identical to those showing repression by TGFβ and counter-regulation by GW501516 (red data points in Figure 2B and C). This overlap (Figure 2D) includes 37% of all class A probes (45/122) and 40% of all GW501516-induced sequences (45/112). The concomitant sensitization by TGFβ to activation by PPARβ/δ agonists and the reversal of the repressive effect of TGFβ by these ligands is also illustrated by the data in Figure 4 and Table 1. These findings suggest that the TGFβ-mediated repression of class A genes and its reversal by PPARβ/δ agonists are functionally linked.
Finally, we addressed the molecular mechanisms that contribute to the regulation of class A genes. The activating and repressive activities of PPARs have been linked to interactions with proteins that serve as coactivators or corepressors, which in turn have profound effects on the local chromatin structure (9,25). Analysis of our microarray revealed a higher expression of several genes encoding corepressors of nuclear receptors in TGFβ-treated cells relative to solvent controls. These include NCOR1 (coding for NCOR), NCOR2 (encoding SMRT), SHARP, LCOR, SIN3B, MTA1 and CALR (Figure 6A). Previous work by several laboratories has established a role for the corepressors NCOR and SMRT in transcriptional repression by unliganded PPARβ/δ in vivo (9,25–28). Upregulation of NCOR2 was observed in RT-qPCR experiments already 6h after treatment with TGFβ, whereas the induction of NCOR1 was statistically not significant at this time point (Figure 6B).
We therefore analyzed whether TGFβ might influence the recruitment of SMRT to the PPREs of the ANGPTL4 gene in vivo. Figure 6C shows that this is indeed the case. TGFβ treatment induced a 2.2-fold enhanced recruitment relative to solvent-treated cells, which was decreased to 1.3-fold in the presence of GW501516. This correlates well with the observed changes in ANGPTL4 expression, pointing to a causal relationship between the regulation of class A genes and the recruitment of SMRT in response to TGFβ and GW501516.
To test this hypothesis, we analyzed the impact of NCOR2 siRNA interference on TGFβ and GW501516-regulated ANGPTL4 and PDK4 gene expression. As shown in Figure 7A (left), the treatment of WI38 cells with NCOR2 siRNA reduced NCOR2 expression to 28–46% relative to cells exposed to control siRNA. The same treatment also attenuated the TGFβ-mediated repression of both PPAR target genes, whose relative expression levels increased in NCOR2 siRNA-treated cells from 0.23 to 0.50 for ANGPTL4, and from 0.16 to 0.40 for PDK4 (Figure 7A and B), respectively. This increased basal level expression was paralleled by a decreased inducibility by PPARβ/δ ligands in the presence of TGFβ, which dropped by ~50% for both genes (Figure 7A and C). The fact that similar patterns were seen with both ANGPTL4 and PDK4 indicates that the regulatory mechanism identified in this study is not gene-specific. Taken together, these observations clearly establish a functional connection between SMRT, TGFβ and the transcription of PPARβ/δ target genes.
Several lines of evidence strongly suggest that PPARβ/δ plays a role in regulating the differentiation and function of tumor stroma and inflammatory cells, pointing to a crosstalk of PPARβ/δ and cytokine signaling pathways. A cytokine with a pivotal function in inflammation and tumorigenesis is TGFβ. In the present study, we tested this hypothesis by asking whether PPARβ/δ and TGFβ signaling pathways functionally interact and modulate the transcriptional activity of common target genes in diploid human fibroblasts induced to differentiate into myofibroblast-like cells.
The potential interaction of transcriptional signaling pathways regulated by PPARβ/δ and TGFβ was analyzed by determining the genome-wide transcriptional profile of WI38 cells treated with TGFβ, a PPARβ/δ agonist or both ligands. The data obtained from this analysis point to an extensive crosstalk of the transcriptional signaling pathways regulated by PPARβ/δ and TGFβ (Figures 2 and and3).3). Bioinformatic analyses identified several classes of genes showing distinct responses to the combined action of TGFβ and PPARβ/δ agonists. Two of these classes that are of particular interest are characterized by the following distinct features (Figures 2B and and3):3): (i) repression by TGFβ, which is counteracted by PPARβ/δ agonists (class A genes; Table 1), and (ii) induction by TGFβ, which is counteracted by PPARβ/δ agonists (class B genes). In both cases, PPARβ/δ agonists significantly inhibited the effect of TGFβ, indicating that this mode of interaction is a major feature of the interaction of these pathways.
We also determined for individual probes the effect of TGFβ on ligand-mediated PPARβ/δ activation. This analysis identified a major set of genes, representing mostly classical PPAR target genes, such as ANGPTL4, PDK4, ADRP and CPT1A, which show increased induction by GW501516 in the presence of TGFβ (Figure 2C). These genes overlap to a large extent (37%) with class A genes (red data points in Figure 2B and C), indicating that the enhancement of GW501516 inducibility by TGFβ is functionally linked to their repression by TGFβ.
It has previously been shown that the ANGPTL4 gene is induced by TGFβ in human breast cancer cell lines (21), which is in apparent contrast to the findings reported in the present study. It is, however, well established that TGFβ frequently exerts opposite effects on target gene expression in mesenchymal and epithelial cells, and that neoplastic transformation can subvert TGFβ-mediated transcriptional regulation (29). It would thus be conceivable that the ANGPTL4 gene is also subject to a similarly complex regulatory network. Consistent with this hypothesis is our observation (30) that ANGPTL4 transcription is induced by TGFβ in the epithelial cell line HaCaT (31) and in WPMY-1 cells, which is a SV40-transformed cell line derived from human prostate carcinoma-associated fibroblasts (32). These findings suggest that the ANGPTL4 gene is a useful model to investigate the molecular mechanisms underlying the cell type-specific and transformation-dependent effects of TGFβ-triggered transcriptional signaling pathways.
In the absence of ligands, PPARβ/δ target genes can be repressed through the recruitment of corepressors to PPRE-bound PPARβ/δ-RXR heterodimers, such as NCOR and SMRT (9,25–28). In the present study, we tested the hypothesis that TGFβ may enhance the formation or function of these repressor complexes. In such a scenario, TGFβ would lead to a decreased transcriptional activity in the absence of ligands, and PPARβ/δ agonists induce the dissociation of corepressors and their replacement with coactivators, thereby counteracting the TGFβ effect. Our data are consistent with this model: (i) the NCOR2 gene (coding for SMRT) is a transcriptional target of TGFβ (Figure 6A and B); (ii) the TGFβ-induced NCOR2 expression leads to an increased recruitment of the SMRT corepressor to the ANGPTL4 PPREs in vivo (Figure 6C); (iii) this enhancement of SMRT recruitment is markedly diminished by the PPARβ/δ agonist GW501516 (Figure 6C); (iv) the siRNA-mediated inhibition of NCOR2 expression leads to a strong derepression of ANGPTL4 transcription and an inhibition of TGFβ-mediated repression (Figure 7A and B); and (v) the same treatment also reduced the inducibility by PPARβ/δ ligands in the presence of TGFβ (Figure 7A and C). These findings provide compelling evidence for a functional link between the TGFβ-induced expression of SMRT, the impact of TGFβ on PPARβ/δ target genes and the counteracting effects of PPARβ/δ ligands. Importantly, similar siRNA effects were also observed with another class A gene, the PPARβ/δ target gene PDK4 (Figures 4 and and7).7). This suggests that the regulatory mechanism identified here is not a peculiar feature of the ANGPTL4 gene, but appears to a have a broader relevance. Collectively, our findings establish a clear functional connection between the induction of corepressor expression by TGFβ and the transcription of PPARβ/δ target genes, as are illustrated by the model in Figure 8.
The data in Figure 7A and C indicate that after knockdown of NCOR2 expression, TGFβ still represses ANGPTL4 and PDK4 transcription, albeit to a reduced extent. This suggests that SMRT may not be the only corepressor relevant in this context, and that the PPARβ/δ repressor complex is probably subject to additional regulatory mechanisms triggered by TGFβ. This is supported by the observation that the overall expression level induced by PPARβ/δ ligands is higher than that observed after treatment with ligand plus TGFβ (Figure 4). Consistent with this hypothesis, TGFβ induces several other corepressor genes, such as CALR (calreticulin), LCOR, MTA1, SHARP and SIN3B (Figure 6A), which may play a role in the formation of PPARβ/δ repressor complexes, as previously published for SHARP (9,25–28). The clarification of these questions will be the subject of future studies aiming at a precise dissection of the molecular mechanism involved in the regulation of class A genes by PPARβ/δ and TGFβ.
The genes represented by the second group are induced by TGFβ, which is diminished by PPARβ/δ agonists (Figure 2B, blue data points). This group contains several genes that are potentially relevant in view of the known function of PPARβ/δ in modulating the immune responses. One of these is interleukin-6, a cytokine with both pro-inflammatory and anti-inflammatory properties and a vast range of biological and pathophysiological activities, including a role in tissue fibrosis (33). Time course experiments suggest that repression of TGFβ-mediated IL6 induction by PPARβ/δ ligands is a direct event, because it is detectable within 6h post-treatment (Figure 5C). The IL6 gene is regulated by multiple transcription factors, including NFκB and C/EBP, which have been suggested to interact with PPARs in different experimental systems (34,35). It is possible that the inhibitory effect of PPARβ/δ on TGFβ-induced IL6 transcription is also associated with these transcription factors. Another potentially interesting gene in this context is CD274 coding for B7-H1, a membrane-bound ligand that modulates the activation or inhibition of lymphocytes and myeloid cells (36). Taken together, these data suggest that in differentiating myofibroblasts PPARβ/δ agonists counteract the effects of TGFβ for a subset of target genes with functions in immune regulation, highlighting the relevance of these compounds as potential anti-fibrotic and anti-inflammatory drugs.
We also detected cooperation of the two signaling pathways for several genes (Supplementary Table S1; class C and D). The cooperatively repressed genes (class C) include the cell cycle and tumorigenesis promoting genes KIT, FOXQ1 and TOP2A. This is of potential interest, because we observed cooperative effects of TGFβ and GW5101516 also on cell-cycle regulation. Thus, GW501516 not only inhibited cell-cycle progression in untreated WI38 cells, but also enhanced the inhibitory effect of TGFβ (Figure S2). The cooperative regulation of genes that have been associated with the cell cycle may thus provide an explanation for the cooperation of GW501516 and TGFβ in the inhibition of cell-cycle progression. However, it cannot be ruled out at present that the cell-cycle effects mediated by the two ligands are functionally unrelated. Inhibition of cell proliferation by PPARβ/δ ligands has previously been reported for a number of other cell lines of different origins, but the underlying molecular mechanisms remain largely obscure (10).
The Deutsche Forschungsgemeinschaft (Mu601/12-1 and SFB/TR17); Genomics and Bioinformatics core facility of the LOEWE-Schwerpunkt ‘Tumor and Inflammation’. Funding for open access charge: Research grant (DFG).
Conflict of interest statement. None declared.
Supplementary Data are available at NAR Online.