Peroxisome Proliferator-Activated Receptor γ and PGC-1α in Cancer: Dual Actions as Tumor Promoter and Suppressor

2.1. PPARγ as a Tumor Suppressor in Cancer

Our previous review summarized that PPARγ inhibits cell proliferation and induces apoptosis through the upregulation of Phosphatase and Tensin Homolog (PTEN), downregulation of survivin, downregulation of X-linked inhibitor of apoptosis (XIAP), suppression of NF-κB and glycogen synthase kinase (GSK)-3β, upregulation of cyclin-dependent kinase (CDK) inhibitors, downregulation of CDK and cyclin D1, downregulation of COX-2, upregulation of Krüppel-Like Factor 4 (KLF4), upregulation of Bax, downregulation of Bcl-2, and inhibition of telomerase activity and hTERT expression through modulation of the Myc/Mad/Max network [8]. This review briefly describes and summarizes new molecular mechanisms of PPARγ-related tumor suppression since 2012 (Table 1, Figure 2).

Understanding the role of PPARγ in cancer was improved by developing new synthetic and natural ligands of PPARγ and performing overexpression and knockdown experiments. PPARγ agonist troglitazone inhibits colon cancer cell growth through the inactivation of NF-κB by suppressing GSK-3β activity [24]. Emerging data suggest that PPARγ acts as a tumor suppressor by inactivating NF-κB through different mechanisms. For example, Lee et al. demonstrated that 4-O-methylhonokiol (MH), a PPARγ agonist, has antitumor activity in prostate cancer through increased PPARγ activity and p21-mediated suppression of NF-κB activity as observed by the loss of MH-induced growth inhibition and NF-κB inhibition in a p21 siRNA knockdown experiment [25]. In addition, overexpression of PPARγ was shown to inhibit cell proliferation and tumor growth via degradation of NF-κB by acting as an E3 ligase [26]. Hou et al. demonstrated that PPARγ inhibits mucin 1- (MUC1-) C-mediated cell proliferation via MUC1-C ubiquitination and degradation [27]. MUC1-C is known as an oncoprotein and interacts with IκB kinase, NF-κB/p65, and signal transducer and activator of transcription factor 3 (Stat3), p53, or BAX in order to activate the downstream pathway associated with tumor growth [47–52]. Efatutazone, a third-generation PPARγ agonist, has been reported to inhibit esophageal squamous cell carcinoma (ESCC) cell proliferation in vitro and in vivo through increased p21Cip protein levels via inactivation of Akt [28].

Several recent studies have shown that PPARγ agonists inhibit the survival of cancer stem cells (CSCs) [29, 30, 53–55]. PPARγ and RXR agonists were demonstrated to inhibit interleukin-6 (IL-6) promoter activity and reduce MMP-2 and MMP-9 expression and activity in tumor-associated fibroblasts [29]. Prost et al. demonstrated that pioglitazone, a PPARγ agonist, eradicates CSCs via the decreased expression of STAT5 and HIF-2α in chronic myeloid leukemia [30].

The Wnt/β-catenin signaling pathway plays an important role in the occurrence and development of cancer [56, 57]. Guo et al. reported that PPARγ overexpression inhibits the proliferation and migration of gastric cancer cells through downregulation of telomerase reverse transcriptase (TERT) and enabled homolog (ENAH) via inhibition of β-catenin [31]. Mammalian enabled (Mena), encoded by ENAH, is an actin-regulatory protein involved in controlling cell motility and cell-cell adhesion, which are important for the development of metastatic potential [58]. TERT and ENAH are new targets of the Wnt/β-catenin signaling pathway [59, 60]. Recently, activation of canonical Wnt signaling was reported to directly act on aerobic glycolysis and increase vessel formation in colon cancer through the Wnt target gene pyruvate dehydrogenase kinase 1 (PDK1) [61]. Via PDK1 activation, pyruvate is converted into acetyl-CoA, which enters the TCA cycle and is converted into citrate, which stimulates protein synthesis. Accumulation of metabolic intermediates (such as aspartate, glycine, serine, and ribose) in cells promotes de novo nucleotide synthesis, contributing to growth and proliferation [62]. In addition, blocking the Wnt pathway decreases PDK1 expression via transcriptional regulation and inhibits in vivo tumor growth [61].

Pseftogas et al. reported that PPARγ activation has a tumor suppressive effect by upregulating the expression of tumor suppressor Cyld, as the Cyld promoter has PPARγ binding sites [32]. Cyld was identified as a tumor suppressor gene that is causally associated with the development of inherited cylindromas [63]. The gene encodes a protein (CYLD) possessing a carboxyl-terminal ubiquitin-specific protease domain that selectively hydrolyzes K63- and M1-linked polyubiquitin chains [64]. A number of studies have suggested a role for CYLD in the growth suppression of different types of cancer cells, such as colon, hepatocellular, lung, melanoma, and breast cancer (reviewed in [65]). CYLD can inhibit several growth and antiapoptotic signaling pathways, including the NF-κB, JNK, p38, Wnt, Akt, and Notch pathways [65].

Rovito et al. demonstrated that PPARγ activation downregulates CXCR4 gene expression through recruitment of the silencing mediator of retinoid and thyroid hormone receptor (SMRT) corepressor to PPRE within the CXCR4 promoter and then inhibits breast cancer cell migration and invasion [33]. CXCR4, a seven-transmembrane G-protein-coupled receptor for stromal-cell derived factor-1α (SDF-1α), has been shown to be expressed in human breast cancer cells, and activation of the SDF-1α/CXCR4 axis is important in breast cancer migration and metastasis [66, 67].

2.2. PPARγ as a Tumor Promoter in Cancer

Our previous review described that PPARγ has tumor-promoting activity through the upregulation of β-catenin and c-Myc expression, upregulation of COX-2, upregulation of the expression of vascular endothelial growth factor (VEGF) and VEGF receptors, and upregulation of MMP-1 [8]. This review briefly introduces the action mechanisms of PPARγ as a tumor promoter (Figure 3).

Recently, increasing evidence has indicated that PPARγ acts as a tumor promoter [68–74]. Downregulation of PPARγ by siRNA knockdown or treatment with PPARγ antagonist GW9662 has been shown to inhibit the growth of cancer cells, suggesting a tumor-promoting effect for PPARγ in these cells [68–70]. PPARγ was shown to protect ErbB2-positive breast cancer cells from palmitate-induced toxicity [75]. In addition, PPARγ was demonstrated to play a crucial role in the maintenance of stemness in ErbB2-positive breast cancer cells; PPARγ antagonist GW9662 induces apoptosis and inhibits tumorsphere formation and tumor formation through the inhibition of lipogenic genes (ACLY, MIG12, FASN, and NR1D1) and stem cell-related genes (KLF4 and ALDH) [71]. CSCs have been identified as subpopulations of cells within tumors that promote tumor growth and recurrence [76–78].

Kesanakurti et al. demonstrated that PPARγ is involved in radiation-induced epithelial-to-mesenchymal transition (EMT) in glioma by interacting with p21-activated kinase 4 (PAK4), resulting in increased Nox1 expression and reactive oxygen species (ROS) [72]. EMT is a developmental transdifferentiation program facilitating the formation of highly motile cells with stem cell characteristics [79, 80]. EMT is also involved in increased metastatic potential and treatment resistance in cancer [81, 82]. The PAKs are a family of serine/threonine kinases involved in embryonic development, cytoskeletal remodeling, cell motility, and cell proliferation [83, 84], and aberrant expression of PAK4 has been shown to promote cancer cell proliferation and invasion [85–87].

A recent study using PPARγ siRNA showed that PPARγ suppression inhibits cell proliferation, colony formation, and tumorigenicity in vivo [73]. In addition, PPARγ upregulated VEGF expression through the binding of PPARγ in the promoter region of VEGF in prostate cancer cells [73]. Patitucci et al. demonstrated that PPARγ activation is involved in steatosis-associated liver cancer and provided evidence supporting the pharmacological modulation of hepatic PPARγ activity as a therapeutically relevant strategy in hepatic malignancy associated with activated Akt2 and PPARγ signaling [74].

3.1. Tumor-Promoting Functions of PGC-1α

As described above, PGC-1α is a regulator of PPARγ activity. Thus, the abnormalities in PGC-1α expression may affect PPARγ function. However, there was little report supporting that PGC-1α expression directs PPARγ activity in cancer. Thus, this review focuses on the role of PGC-1α, independent of the role of PPARγ in cancer. Literature works supporting the tumor-promoting functions of PGC-1α have increased [17, 34–40, 42, 91–93]. Shiota et al. showed that PGC-1α promotes cell growth through the activation of androgen receptor in prostate cancer cells by observing cell growth inhibition with PGC-1α knockdown experiments [17]. In addition, PGC-1α was increased in tumor samples from arsenic-induced skin cancer patients and may be associated with increased cell proliferation and enhanced mitochondrial biogenesis [34]. Bhalla et al. showed that PGC-1α promotes carcinogenesis and tumor growth through the induction of lipogenic enzymes (acetyl-CoA carboxylase and fatty acid synthase) using genetically modified PGC-1α mice [35]. That study demonstrated that PGC-1α knockout mice had decreased chemically induced liver and colon carcinogenesis, suggesting that PGC-1α may stimulate carcinogenesis [35]. Similarly, Shin et al. first demonstrated that overexpression of PGC-1α enhances cell proliferation and tumorigenesis via the upregulation of Sp1 and acyl-CoA binding protein [36]. It was also reported that PGC-1α overexpression leads to increased antioxidant enzymes (catalase, superoxide dismutase) and decreased ROS-induced apoptosis [36]. Similarly, PGC-1α knockdown significantly decreased cell number and induced apoptosis in PGC-1α positive melanoma cell lines, suggesting that PGC-1α is crucial in the survival of PGC-1α positive melanoma cells [37]. In addition, superoxide dismutase 2 protein levels were decreased in PGC-1α depleted melanoma cells. Moreover, ectopic expression of PGC-1α in melanoma cells increased the expression of ROS detoxifying genes. These data support the hypothesis that PGC-1α plays an important role in activating the ROS detoxification gene program to maintain melanoma cell survival [37]. Vazquez et al. also demonstrated that there was a significant reduction in tumor size in PGC-1α depleted cells, implying PGC-1α may be important in tumor progression [37]. De novo lipogenesis is a distinctive anabolic feature of malignant cells [94]. Carbons from glucose and glutamine supply cytoplasmic citrate for fatty acid synthesis with the help of lipogenic enzymes [94]. Glutamine can serve as an anaplerotic mitochondrial fuel and seems to be important for tumor survival [95]. In ErbB2-positive breast cancer cells, the PGC-1α/ERRα complex directly regulates the expression of glutamine metabolism enzymes, leading to the provision of glutamine carbons to de novo fatty acid synthesis [38]. PGC-1α overexpression, or ERRα activation, confers growth advantages of breast cancer cells even under limited nutrients, supporting the correlative clinical data that high expression of PGC-1α is associated with poor prognosis, possibly related to the activation of its downstream glutamine pathway target genes [38]. It was reported that PGC-1α expression is affected by various transcriptional pathways. One example is that melanocyte-lineage transcription master regulator and oncogene MITF activated PGC-1α expression in melanoma [37, 91]. The decrease in mitochondrial membrane potential and increased ROS production with a decrease in glutathione, cystathionine, and 5-adenosylhomocysteine were observed in PGC-1α-depleted melanoma cell lines, suggesting that intrinsic apoptotic pathway is activated in PGC-1α-depleted melanoma cells [37]. Another example is that the androgen receptor-AMP-activated protein kinase (AMPK) signaling axis increased expression of PGC-1α to drive growth advantages in prostate cancers [39]. It was also shown that PGC-1α expression was significantly higher in lung adenocarcinomas with wild type p53 than in tumors with mutant p53 [40]. Cell proliferation was inhibited by PGC-1α siRNA knockdown experiments in H1944 lung adenocarcinoma cells [40]. In metabolic stress conditions, PGC-1α was shown, in complex with p53, to coactivate the transcription of cell cycle inhibitors, while it was also shown to promote the expression of genes related to mitochondrial biogenesis. These two functions are cooperative processes that promote cell survival. Moreover, oxidative stress in PGC-1α knockdown cells resulted in p53-induced apoptosis [96]. In turn, it was also shown that increased expression of PGC-1α might prevent p53-induced cell death by maintaining an adequate balance between oxidative phosphorylation and glycolysis [97].

Some studies have examined the effect of PGC-1α on angiogenesis. PGC-1α has been reported to activate the production of VEGF through the estrogen-related receptor α- (ERRα-) dependent pathway [98]. PGC-1α was shown to regulate HIF-1α activity. Increased PGC-1α expression enhances oxygen consumption, resulting in decreased local oxygen tension and increased HIF-1α stability [99]. In addition, HIF-2α is a transcriptional target of PGC-1α, even though the involved transcriptional mechanism is not clear [100]. ERRα is overexpressed in many cancers and its inhibition reduces cell proliferation. Recent studies suggest an important role for the interaction between PGC-1α and ERRα in cancer (reviewed in [15]). Kinase suppressor of ras 1 (KSR1), a molecular scaffold of the Raf/MEK/extracellular signal-regulated kinase (ERK) cascade, has been demonstrated to promote oncogenic Ras-dependent anchorage-independent growth through the activation of PGC-1α and ERRα [92]. Interestingly, recent study shows that PGC-1α plays an important role in the metastatic switch. LeBleu et al. demonstrated that circulating mammary epithelial cancer cells exhibit increased PGC-1α expression, enhanced mitochondrial biogenesis, and oxidative phosphorylation, which may contribute to distant metastasis and poor patient outcome [93]. In addition, PGC-1α knockdown decreased ATP production, reduced actin cytoskeleton remodeling, lowered anchorage-independent survival, and decreased intra-/extravasation, which are all checkpoints that prevent metastasis in MDA-MB-231 breast cancer and B16F10 melanoma cells [93]. LeBleu et al. also showed that PGC-1α expression in invasive cancer cells was significantly associated with the formation of distant metastases in a clinical analysis of human invasive breast cancers [93].

3.2. Anticancer Functions of PGC-1α

As opposed to the tumor-promoting role of PGC-1α described above, several studies have shown that PGC-1α has anticancer effects. As described above, PGC-1α is decreased in colon [88], breast [89], and ovarian cancer cells [41], and PGC-1α overexpression in human ovarian cancer cell line Ho-8910 has been shown to induce apoptosis via downregulation of Bcl-1 and upregulation of Bax, suggesting that PGC-1α may be a contributor to the inhibition of tumor growth [41]. Lee et al. found that PPARγ activation and PGC-1α overexpression by adenovirus infection in HepG2 human hepatoma cells induced E-cadherin upregulation and inhibited cell motility [42]. One report showed that PGC-1α overexpression induced apoptosis via ROS accumulation in HT29 and HCT116 colorectal cancer cells. In addition, PGC-1α overexpression reduced tumor growth in an HT29 xenograft model, suggesting a role of PGC-1α as a tumor suppressor [43]. Zhang et al. reported that von Hippel-Lindau- (VHL-) deficient clear cell renal carcinomas exhibited higher levels of HIF-1α and enhanced glycolysis [101]. HIF-1α is known to induce the expression of transcriptional repressor Dec1, which leads to the suppression of PGC-1α expression and the inhibition of mitochondrial respiration [102]. However, the enforced PGC-1α expression in VHL-deficient cells, despite the restoration of mitochondrial function, did not block the inhibition of cell growth and enhanced sensitivity to cytotoxic therapies in oxidative stress conditions [102]. This is in line with clinical clear cell carcinoma data that showed the correlation of higher mitochondrial mass with reduced tumor aggressiveness [103], and the association of lower PGC-1α levels with worse patient outcome [102]. It was shown that PGC-1α attenuates stress responses necessary for cancer cell survival, by interacting with heat-shock factor 1 [104]. Wang and Moraes revealed that increased PGC-1α expression due to treatment with PPAR panagonist (bezafibrate) increased mitochondrial biogenesis, resulting in an inhibition of cancer cell proliferation under glycolytic conditions and inhibition of invasion [44]. In addition, PGC-1α downregulation by miRNA-217 led to the promotion of cancer cell proliferation in breast cancer cells, suggesting a role of PGC-1α as a tumor suppressor [105]. Recently, Torrano et al. showed that PGC-1α suppresses metastasis of prostate carcinoma through an ERRα-dependent transcriptional program [45]. Highly metastatic melanoma cells expressed lower levels of PGC-1α [46, 106]. In turn, these PGC-1α-low cells expressed higher levels of integrin, TGFβ, and Wnt signaling components involved in metastasis. It was shown that genetic depletion of PGC-1α increased metastasis in poorly invasive melanoma cells [46]. In contrast, PGC-1α overexpression in melanoma cells by ectopic expression or exposure to BRAF^V600E inhibitor vemurafenib suppressed metastasis through the direct regulation of inhibitor of DNA binding protein 2 (ID2) and inhibition of TCF-mediated gene transcription [46].

As described above, there have been many studies of the role of PGC-1α in tumor progression. However, it is still not sure if PGC-1α acts as a tumor promoter or tumor suppressor, and to date it is thought that its effect on tumor varies depending on the tissue context and tumor type (reviewed in [107]).