PPAR-γ Modulators as Current and Potential Cancer Treatments

Worldwide, cancer has become one of the leading causes of mortality. Peroxisome Proliferator-Activated Receptors (PPARs) is a family of critical sensors of lipids as well as regulators of diverse metabolic pathways. They are also equipped with the capability to promote eNOS activation, regulate immunity and inflammation response. Aside from the established properties, emerging discoveries are also made in PPAR’s functions in the cancer field. All considerations are given, there exists great potential in PPAR modulators which may hold in the management of cancers. In particular, PPAR-γ, the most expressed subtype in adipose tissues with two isoforms of different tissue distribution, has been proven to be able to inhibit cell proliferation, induce cell cycle termination and apoptosis of multiple cancer cells, promote intercellular adhesion, and cripple the inflamed state of tumor microenvironment, both on transcriptional and protein level. However, despite the multi-functionalities, the safety of PPAR-γ modulators is still of clinical concern in terms of dosage, drug interactions, cancer types and stages, etc. This review aims to consolidate the functions of PPAR-γ, the current and potential applications of PPAR-γ modulators, and the challenges in applying PPAR-γ modulators to cancer treatment, in both laboratory and clinical settings. We sincerely hope to provide a comprehensive perspective on the prospect of PPAR-γ applicability in the field of cancer treatment.


FUNCTIONS OF PPAR-g
PPAR-g, a subtype of PPARs, is most expressed in adipose tissue, with two isoforms PPARg1 and PPARg2, the former being more biologically and universally expressed in various tissues, while the latter is more restricted to adipose tissue, but could be induced by high-fat diet (HFD) in other tissues. This tight linkage between PPAR-g and fat has significantly attributed to the multi-functions of PPAR-g in metabolism, vascularization, inflammation, cell cycle regulation, differentiation, and migration, rendering it a special niche in the treatment of metabolic disorders and cancers.

Regulation of Lipid and Glucose Metabolism
PPAR-g is highly implicated in metabolism of lipids, which is a potent facilitator of adipogenesis and fatty acid storage (10,22). The activation of PPAR-g decreases circulating lipids which are majorly triglycerides and free fatty acids. Meanwhile, it selectively promotes lipid uptake by inducing adipogenesis and increasing lipid storage (20). PPAR-g is pivotal throughout the entire process of adipogenesis. It is an essential regulator of adipocyte differentiation, a probable participant in adipocyte self-renewal, and a requisite for mature adipocyte function (11). Alongside the metabolic changes of lipids induced by PPAR-activation, a series of adipokines is up-regulated. They participate in multiple actions, including insulin sensitivity regulation, inflammation, tumorigenesis, and global metabolism adjustment. Adiponectin may act on hepatic glucose output with additional vascular benefits on eNOS facilitated by other intermediates, e.g. shock protein 90 and Src homology region 2-containing protein tyrosine phosphatase 2, etc., through the PI3K-Akt pathway (20); resistin, leptin, and tumor necrosis factor-a (TNF-a), majorly functions in inflammatory reactions.
All mentioned adipokines, working individually or collectively, are factors affecting system insulin sensitivity. This collaborative effect on glucose metabolism is realized through the activation of genes responsible for insulin-dependent glucose uptake (GLUT4, IRS-1, IRS-2, and c-Cbl associated protein), and key genes involved in glucose-stimulated insulin secretion (GSIS), via the activation of PPAR-g (10,11,20).

Inhibition of Inflammation and Tumorigenic Implications
Being a part of PPARs superfamily, PPAR-g also actively regulates inflammation and immunity responses. On the genetic level, PPAR-g may activate trans-repression on proinflammatory genes with SUMOylation. SUMOylation is a post-translational modification, namely the conjugation of PPAR-g with Small Ubiquitin-like Modifier (SUMO) upon ligand activation. After a nuclear corepressor complex is bound, nuclear factor kappa B (NF-kB) target genes are transrepressed, which stabilizes NF-kB in a repressed, promoterbound state (10,20). In addition, PPAR-g may regulate the immune system through dendric cells (DC) and macrophages. It affects DC functions by altering antigen uptake, maturation, activation, migration, cytokine production, and lipid antigen presentation. In macrophages, PPAR-g inhibits genes encoding pro-inflammatory molecules while activating the expression of anti-inflammatory mediators. It also manipulates cell differentiation to inhibit wild-type proinflammatory 'M1' macrophages, while facilitating the maturation of antiinflammatory 'M2' macrophages, rendering anti-inflammatory effect bilaterally (11,20).
Strong regulative properties of PPAR-g on inflammation and immunity indicate its potential in cancer immunotherapy. Multiple signaling pathways may be engaged in antitumorigenic activities by the activation of PPAR-g. In PPAR-gattenuated mice melanoma cells, the infiltration of the myeloidderived suppressor cells (MDSCs) exhibits overall non-specific inflammatory responses. A corrective effect is achieved upon the ligand-binding of PPAR-g, and tumor growth is inhibited. This effect is achieved via the mTOR pathway, the subsequent blocking of MDSCs ROS overproduction, and possibly the RAGE pathway (23,24). Macrophages are highly plastic and heterogeneous. In the context of cancer, tumor-associated macrophages (TAMs) are particularly abundant and proproliferative within tumors. They are engaged in immunosuppression and angiogenesis, supporting tumor growth and metastasis. When activated, according to Gionfriddo, G., et al., PPAR-g can reduce the secretion of M1 pro-inflammatory and pro-tumor M2-cytokines without affecting macrophage polarization, yielding an anti-cancer effect (25). However, other studies contended the changes of polarization. Li, T., et al. found that in macrophages derived from human monocytic leukemia cell lines, a tumor suppressor called docking protein-1 (DOK1) can be activated with the activation of PPAR-g. It was supposed to induce polarization of macrophages towards an inflammatory phenotype with increased release of pro-inflammatory cytokines and reduced PD-L1 expression (26).

Induction of Cell Differentiation
Differentiation grades are a critical criterium in assessing the malignancy of tumors, which holds great significance in prognosis and development of clinical treatment. PPAR-g possesses the differentiation-inducing effect, which is associated with multiple different mechanisms yet has not been entirely elucidated (27,28). The presence of PPAR-g expression is required in tissue development, the placenta and heart, without which serious damage could be done. The deficiency of PPAR-g, according to Barak et al., interferes with terminal differentiation of the trophoblast and placental vascularization, leading to severe myocardial thinning and death (29). In breast cancer cells, both in vitro and in vivo, the activation of PPAR-g induces terminal differentiation of cancer cells into adipocytes and causes lipoapoptosis with up-regulated expressions of C/EBPb (30). Cells which go through epithelial-mesenchymal transition (EMT) and trans-differentiated afterward are growth-arrested, losing their cellular plasticity (31). PPAR-g activation also causes degradation of b-catenin and inactivation of the downstream Wnt/b-catenin pathway, and thus increases ketogenesis by inducing mitochondrial 3-hydroxy-3-methylglutaryl-CoA synthase 2 (HMGCS2), a rate-limiting ketogenic enzyme in the synthesis of ketone body b-hydroxybutyrate (bHB), which contributes to the regulation of intestinal cell differentiation (32)(33)(34). As cancer cells are characterized by the symbolic energy metabolism reprogramming, namely the Warburg effect, ketogenesis contributes to intestinal cell differentiation via the inhibition of glycolysis, thus attenuating the progression of human colon cancer cells in vitro (35). On the protein level, the activation of PPAR-g promotes epithelial differentiation and stabilizes the differentiated phenotype by upregulating the expression of key proteins like keratins, E-cadherin, alkaline phosphatase, and developmentally-regulated-GTP-binding protein 1, carcinoembryonic antigen (CEA), differentiationrelated gene-1 (DRG1), possibly through the interaction with Hic-5 which may serve as a PPARg co-activator (10, 20).

Suppression of Cell Proliferation
Limitless cell proliferation with self-sufficient growth signals and insensitivity to contrary or anti-growth signals are the hallmarks of cancer tissue (36); thus, the inhibition of cell proliferation is a major constituent in the attenuation of tumor growth. The antiproliferative effect from the activation of PPAR-g is fulfilled by a variety of mechanisms, such as the induction of cell cycle arrest, the direct up-regulation of certain genes, and the activation of various signaling pathways involved in immunity and inflammatory responses.
Activated upon ligand-binding, according to Zurlo et al., PPAR-g can interact with Sp1, stimulate p21(waf1/cip1) gene transcription, and consequently induce a G0/G1-phase arrest in human colorectal cancer and gastric carcinoma cells (32)(33)(34)37). Multiple cyclins and cyclin-dependent kinases (CDKs) are positive regulators of cell cycle progression, and upon agonistbinding, PPAR-g can facilitate cyclin D1 ablation, inducing cell arrest (38)(39)(40). Another study conducted on rats with breast cancer in vivo has showed the direct up-regulation of PTPRF gene expression by the activation of PPAR-g, which in part inhibits tumor cell proliferation (41). Additionally, upon PPAR-g activation, decreased levels of migration and invasion of human gastric adenocarcinoma cells via downregulating the ERKsignaling pathway (40) have also been demonstrated in a study through long-term patient investigation (42).
Multiple signaling pathways are implicated in the immunoresponses towards cancer cells induced by the activation of PPAR-g: (1) the blockade of the RAGE signaling weakens the recruitment and accumulation of further myeloid cells, including MDSCs, and the inhibition on T and natural killer cells, ameliorating T cell tolerance and strengthening anti-tumor immunity (24,43); (2) the inhibition of Toll-like receptor 4 (TLR4)-dependent mitogen-activated protein kinase (MAPK) pathway leads to the decreased activity of downstream NF-kB pathway (44); (3) the blocking of the mTOR signaling pathway not only downregulates proliferation, migration and invasion of mice prostate cancer and melanoma, but also upregulates apoptosis and autophagy both in vitro and in vitro (23,45), which were consistent with other studies confirming the increased production of reactive oxygen species (ROS) additively (23,40,46); (4) the co-activation and high expression of C/EBPa and PPAR-g, with the former inhibiting Forkhead box C1 (FOXC1)'s promoter activity, an important cancer-associated gene in tumor, can significantly inhibit cell proliferation, migration, invasion, and colony information (47); (5) the activation of PPARg-LXRa-ABCA1 pathway by lycopene has been demonstrated to inhibit both androgen-dependent prostate LNCaP cancer cells and androgen-independent prostate cancer cells DU145 and PC-3 (48).

Regulation of Cell Apoptosis
Another hallmark of cancer is the avoidance and evasion of programmed cell death. Canonical pathways of apoptosis include the intrinsic mitochondrial pathway and the extrinsic pathway mediated via the activation of cell surface death receptors. The intrinsic apoptotic signaling pathway primarily involves the activation of the proapoptotic Bcl-2 family members Bax and Bak, which facilitates the release of cytochrome c from the mitochondria and subsequent caspase-9 cleavage or activation, followed by the final cleavage or activation of the downstream effector caspases such as caspase-3 and -7, resulting in apoptosis. This is a pathway negatively regulated by several anti-apoptotic Bcl-2 family members, e.g., Bcl-2 and Bcl-XL. Apoptotic signaling through the extrinsic pathway is initiated by ligandbinding of death receptors which belong to the tumor necrosis factor (TNF) receptor superfamily, or the induction of trimerization of the said receptors. After that, the intracellular death domain of the death receptors subsequently recruits adapter proteins, forming a death-inducing signaling complex (DISC) which helps to recruit procaspase-8 to the DISC. When caspase-8 is activated, downstream effector caspases such as caspase-3 and -7 will then take effect. The effector caspases can also be activated by death receptors indirectly through caspase-8-mediated cleavage of Bid, which facilitates Bax activation and the subsequent release of cytochrome c from the mitochondria. Thus, the Bid cleavage links both apoptotic pathways (40).

Promotion of Tumorigenesis
Although flooding evidence supports the tumoricidal effect of PPAR-g, its pro-tumorigenic potential cannot be ignored.
Indeed, the counterpart's testimony is also too grounded to be dismissed reasonably, given the consideration that most effects are achieved in a dose, concentration, cancer type and individual-specific manner. Current evidence on the procancer effect of PPAR-g revolves around the intercellular and extracellular adhesion, inflammation, metabolism, cell proliferation, rendering it a negative prognostic biomarker. The carcinoembryonic antigen (CEA) is an important component of the intercellular adhesive forces, and its overexpression can be pro-oncogenic, which may even induce the blockage of differentiation of rat myoblast cells. Upon ligand binding, PPAR-g can induce an increase in CEA-dependent intercellular adhesion (10). In concert, the up-regulated expression of vascular endothelial growth factor A (VEGF-A) and Vimentin, a major intermediate filament protein that plays an important role in cell adhesion, migration, angiogenesis and neurite extension, is relate to the activation of PPAR-g (52). Additionally, the inactivation of PPAR-g by antagonists interferes with cancer cells' adhesion to the extracellular matrix, disrupting survival signals, and thus inducing anoikis, a special form of apoptosis (40).
The activation of PPAR-g displays a contradictory effect in immunotherapies equally. According to Wu, B., et al., the inactivation of PPAR-g boosts the efficacy of aPD-L1 and aPD-1 antibodies in the treatment against murine mammary tumors (53). The activation of PPAR-g in turn suppresses the expression and secretion of inflammatory factors, proinflammatory chemokines, which may reprogram the immune microenvironment of cancer to be less "inflamed", consequently enabling resistance of cancer cells to immune-directed therapies (54).
Metabolic-wise, researchers have found that the activation of PPAR-g could improve glucose and lipid metabolism, proving for the high demand of energy in the proliferation, migration and invasion of human prostate cancer cells and mice model (52). Interestingly, regarding factors involved in cell cycle, another study has found that the pro-apoptotic genes c-Myc and Bax are inhibited while cell-cycle regulators PCNA, cyclin D1 and COX-2 are promoted along with PPAR-g activation (40,55). In several studies, with the cell culture of human renal cell carcinoma, human prostate cancer xenografts in nude mice, as well as brain cancer in vivo, the increased expression of PPAR-g are all prognostically negative and pro-metastatic (56)(57)(58). Figure 1 illustrates the main mechanisms of PPAR-g in the regulation of tumor.

MODULATORS OF PPAR-g
PPAR-g is activated either endogenously or exogenously, respectively by polyunsaturated fatty acids, fatty acid derivatives such as 15-deoxy-delta-12,14-prostaglandin J2 (15d-PGJ2), nitrated fatty acids (10), and thiazolidinediones (TZDs) majorly, as well as other novel synthetic compounds with non-thiazolidinedione scaffold (27). PPAR-g modulators have been used for multiple types of cancer, because it can inhibit tumor growth by directly interfering with related signaling pathways, or indirectly enhancing the tumor cell sensitivity to the radiotherapy, and get utilized in combination with other drugs to promote efficacy synergistically ( Table 1).

PPAR-g Agonist: Thiazolidinediones Family
Thiazolidinediones (TZDs) are synthesized PPAR-g agonists, also known as glitazones, which started drawing considerable attention since the discovery of the prototypical drug ciglitazone. Through PPAR-g transactivation and trans-repression, by altering the expression of related genes to promote fatty acids storage in adipocytes, and consequently prioritizing the oxidation of carbohydrates, more specifically glucose, for cellular energy supply, TZDs are primarily employed in the treatment of type 2 diabetes (T2D) as insulin sensitizers. The first generation of TZDs includes troglitazone, which was invented and approved by FDA in 1997, only to be withdrawn from the market after a mere three years due to serious hepatotoxicity. Alarming side effects had led to the appearance of the second generation of TZDs (122). Derived from the parent compound thiazolidinedione, members of the second generation of TZDs, including pioglitazone and rosiglitazone, were developed by altering the metabolic product of ciglitazone, though the original one was never introduced into clinical use.
Recently, TZDs also present anti-cancer effects dependently or independently of PPAR-g activation, both on the transcriptional and the protein level. Mechanisms are involved in cell cycle, apoptosis, hormonal reactions, and stromal regulation, including partial depletion of intracellular Ca2+ store, proteasomal degradation of related proteins to induce cell cycle arrest and apoptosis, transcriptional repression of sex hormone receptors, and decreased activation of macrophages (123,124).
Before the elaboration on specific drugs, it is argued that there exist several lines of findings in the anti-tumor properties of TZDs in question of the dependence of PPAR-g activation. First, the potency of effects exemplifies a manner irrespective of the activation levels of PPAR-g, but more reliant on drug chemical structures, evidenced by the disparity of apoptosis-inducing ability between troglitazone and pioglitazone. Secondly, the susceptibility of tumor cells towards TZDs is in no strict accordance with PPAR-g activation levels, where cancer with low expression of PPAR-g respond better than those with overexpression. Thirdly, there is a considerable increase in the antitumor potency of the PPAR-g-inactive analogues than their parent compounds, like D2TG and D2CG, which are developed from troglitazone and ciglitazone, respectively. Lastly, the apoptotic effect of TZDs and derivatives is unshaken by the siRNA-mediated knockdown of PPAR-g in certain cell lines (123).

Ciglitazone
Ciglitazone is the prototype of all TZDs developed in the early 1980s, which has never been approved for clinical application in diabetic treatments due to its weak therapeutic efficacy. However, this debutante led way of several potent insulin sensitizers, the  analogues of which, troglitazone, pioglitazone and rosiglitazone, into the therapeutic market in the following years. Since the finding on the varied expression of PPAR-g in cancer of different differentiation stages and its involvement in cell proliferation, however, ciglitazone has been striving to make a comeback into the medical field as a novel potential anti-tumor agent. The anti-tumor effects exerted by ciglitazone extend into various perspectives, including cell cycle, apoptosis, and sex hormones, targeting different categories of cancers. Distinguished from its peers and derivatives, the effects of ciglitazone and its analogue links to PPAR-g independent. During the crucial G1/S phase transition in cell cycle, ciglitazone can suppress cancer progression either by partially depleting intracellular calcium stores, consequently causing the deactivation of eukaryotic initiation factor 2 (eIF2) to inhibit translation initiation (125); or functioning as a cyclin D1-ablative agent via proteasomal degradation, which results in cell cycle arrest (39,(59)(60)(61). Furthermore, ciglitazone is also equipped with the capabilities of effecting apoptotic pathways both intrinsically and extrinsically. Through blocking the interactions between Bcl-2 and Bcl-xL with Bak and other pro-apoptotic proteins, ciglitazone can suppress the anti-apoptotic functions of Bcl-xL and Bcl-2, ameliorating the aberrant intrinsic apoptotic activities (62). TNF-related apoptosis inducing ligand (TRAIL) is a member of the tumor necrosis factor (TNF) family of cytokines, an extrinsic apoptosis inducer. By selectively inhibiting FLIP (FLICE inhibitory protein), an apoptosis-suppressing protein, ciglitazone can block early events in TRAIL/TNF family death receptor signaling and eliminate cancer cells with apoptosis via the extrinsic pathway (63).
Aside from the universal anti-proliferative effects concerning cell cycle and apoptosis, ciglitazone also possesses properties associated with sex hormone receptors. The impacts of hormone receptors on cancer status are distinctive of hormone-dependent cancers. In prostate cancer cells, ciglitazone and its analogue reduce prostate-specific antigen (PSA) secretion with and without a decrease in the expression of androgen receptor (AR), independent of PPAR-g activation. However, in androgen-independent C4-2 prostate cancer cells, the ciglitazone-induced regulation of AR is PPAR-g-dependent (64,65). In breast cancer cells, alongside the aforementioned repressive effect on cyclin D1, the expression of estrogen receptor (ER) alpha is simultaneously inhibited by ciglitazone, possibly yielding a synergistic anti-tumor effect (39,60).
Ciglitazone not only displays significant potential in monotherapy, but combines with other drugs to improve the efficacy. The combination of ciglitazone and RXR-a ligands promotes growth inhibition and apoptosis. Moreover, the co-administration of ciglitazone and drugs outside of conventional cancer medication, such as lovastatin and phenylbutyrate, triggers TNF-a-related apoptosis inducing ligand and gammaradiation and the decrease of cancer cell viability (124).
Ciglitazone can also be a potential chemo-preventative agent due to its pro-differentiation effect. Multiple PPAR-g agonists, including ciglitazone, are capable of reducing brain tumor stem cell expansion and promoting differentiation and selfrenewal (66).

Troglitazone
Troglitazone (TGZ), the first generation of TZDS, is a dualagonist of PPAR-a and PPAR-g, more strongly of the latter, which underlines its multi-functionalities in tackling inflammation on top of diabetes (126). It was the first-ever anti-diabetic drug with the mechanism of insulin sensitization, carrying the belief of reducing cardiovascular risks associated with diabetes upon addressing the primary metabolic defect (127,128). However, after the linkage to the estimated 430 liver failures and the reported 63 liver failure deaths during the meagre span of three years since the FDA approval in 1997, TGZ was withdrawn from the market in 2000. And since then, extensive studies have been carried out on its hepatotoxicity, the understanding from which in turn took a twist and shed a light on the treatment of cancer by TGZ's anti-tumor properties (129).
On the transcriptional level, TGZ exerts effects with transactivation and trans-repression of genes, with the former majorly mediating anti-proliferation and -migration, and prodifferentiation, and the latter involving in inflammatory responses (124). In MCF-7 breast cancer cell line, TGZ reverses the effects induced by 12-O-tetradecanoylphorbol-13acetate (TPA), by inhibiting nuclear factor kB (NF-kB) and activator protein-1 DNA binding, and inhibiting matrix metalloproteinases (MMPs) which play a significant role in tissue remodeling, through a PPAR-g-dependent mechanism (67). Additionally, in prostate cancer PC-3 cell line, cancer growth is attenuated via the up-regulation of E-cadherin and glutathione peroxidase 3 (GPx3) upon TGZ administration, in a dose-and PPAR-g-dependent manner, with the downregulation of which being characteristic of epithelial-mesenchymal transition (EMT) (68). Similar results are shown in another study conducted on xenograft mice with human colon cancer cells, where both the E-cadherin/beta-catenin system and the Drg-1 gene expression were modulated, yielding a differentiation-promoting effect (130). Among all cancer, Differentiated Thyroid Carcinoma (DTC) holds a special place for TZDs, because follicular thyroid cancer is the only known neoplasm associated with a PPAR-g fusion gene product with over 30% follicular thyroid carcinoma expressing PAX8/PPAR-g (124). Aside from the traditional anti-tumor properties of antiproliferative, proapoptotic, and differentiating effects, the increased expression of sodium-iodide symporter and restoration of radioiodine uptake renders TGZ a potential agent in treatments against DTC, especially those resistant to conventional therapies (69)(70)(71).
Whereas anti-tumor evidence has emerged outside the PPAR activating property of TGZ where no ligand-binding of PPAR is detected. By TGZ's direct targeting epidermal growth factor receptor (EGFR) to induce its internalization and degradation by the endo-lysosomal degradation machinery, growth arrest of tumor cells can be achieved by the subsequent inhibition of EGFinduced Akt phosphorylation (124,131). In another study on pancreatic cancer in vivo, TGZ promotes mitochondriamediated apoptosis and moderately inhibits cell migration via the JNK pathway without marked adverse effects, in a PPAR-gindependent manner (72).
Furthermore, there is a mixed result in phase 2 trials in demonstration of the clinical efficacy of TGZ in monotherapy. Little to no improvement is witnessed in studies on metastatic colon cancer and breast cancer, while a significant decrease in serum cancer-specific antigen is detected in prostate cancer patients administrated with TGZ (73,132,133).
Despite the currently discouraging data from TGZ monotherapy, the results from the co-administration of troglitazone with other compounds are preferable. Evidence has been emerging in diverse categories of medication: the joint use of RXR-a ligands and TGZ displays apoptotic and growth inhibitory effect in gastric carcinoma and breast cancer cell lines (74,75); chemotherapeutic drugs, including cisplatin and paclitaxel, can induce growth inhibition of lung cancer cell lines, in synergy with TGZ (76,77); in ovarian and breast cancer cell lines refractory to conventional therapies, combined administration of cell signaling molecules, including TRAIL (TNF related apoptosis inducing ligand), and TGZ achieved promising apoptotic results synergistically (78)(79)(80); among other studies, various medication, including statin (lovastatin), NSAIDs (aspirin), estrogen modulator (tamoxifen) and x-ray, have all been reported to work synergistically with TGZ against thyroid cancer, glioblastoma cancer, lung cancer, breast cancer and cervix cancer as anti-tumor agents (134)(135)(136)(137)(138).
Chemo-preventative effect of TGZ against cancer has also been contended. In chemically induced rat colon and oral squamous cancer model, TGZ administration significantly reduces aberrant formations and activities (81, 82).

Pioglitazone
Pioglitazone (PGZ), derived from ciglitazone as a secondgeneration TZD, entered the US market for therapeutic use for type 2 diabetes in 1999, and shows therapeutic value in monotherapy and in co-administration as an adjuvant agent to conventional medication against various categories of cancer.
Several studies have supported that when administered alone, PGZ shows certain anti-tumor therapeutic effects both dependent on its PPAR-g activation property and independent. In a PPAR-g-dependent manner, PGZ decreases the expression level of b-catenin protein, a crucial carcinogenic mediator, and suppresses glioma cells growth and invasion both in vitro and in vivo as in mice xenograft (83). In breast cancer, the increased level of aromatase has been associated with heightened inflammation and a worse prognosis. The activation of PPAR-g by PGZ can halt the progress of breast cancer via multiple mechanisms, targeting estrogen receptor (ER) and aromatase, either by inhibiting ER expression through the PTEN pathway or inducing proteasome-dependent degradation of ER, or inhibiting aromatase via the PGE2 and BRCA1 pathways (139). The elevated expression of receptor for advanced end glycation products (RAGE) has been reported in the pathological biopsy of human HCC tissues. After being treated with PGZ, there is the increased expression of PPAR-g and decreased expression of RAGE of HCC cells in a dose-dependent manner, which indicates an association between PGZ administration and the inhibition on the growth and invasion of HCC cells (84). The significant metabolic regulating property of PPAR-g can also contribute to the anti-tumor effect. According to Srivastava et al., the activation of PPAR-g by PGZ induces a metabolic switch, inhibits pyruvate oxidation, by suppressing pyruvate dehydrogenase kinase 4 (PDK4) or b-oxidation of fatty acids, causes a marked increase in reactive oxygen species (ROS) levels, and eventually cell cycle arrest in lung cancer cells (85). Independent of PPAR-g activation, there is evidence indicating the suppression of the expression of cyclooxygenase-2 (COX-2) and interleukin-8 (IL-8) after the administration of PGZ, exemplifying an anti-angiogenic effect on pancreatic cancer cells in vitro and colon cancer cells in vivo (86,87). Another in vitro study also revealed an interesting anti-tumor effect of PGZ independent of PPAR-g activation. According to Saiki et al., the inhibitory effect of PGZ on the growth of human leukemia cell lines shows a selective pattern in leukemic, rather than normal human hematopoietic progenitor cells (140). Another study also on leukemia has revealed further evidence on PGZ's anti-cancer effect in combined treatment with arsenic trioxide (ATO), which may be achieved via the suppression of PI3K/Akt pathway (141). Meanwhile, a group of Japanese scientists found that used alongside Cisplatinum (CDDP), PGZ could arrest human osteosarcoma cell's chemoresistance, prohibiting cell proliferation and controlling necrosis (142). As for studies conducted in vivo, PGZ has also yielded promising performance in suppressing the progress of various cancer. In addition to anti-proliferative and anti-metastatic effect of PGZ on the aforementioned mice xenograft of human pancreatic cancer and colon cancer, there is also evidence of its antiangiogenetic efficacy targeting non-small cell lung cancer (NSCLC) via PPAR-g activation and the subsequent inhibition of NF-kB transcriptional activity (88). Similar to its analogues, PGZ also shows a chemo-preventative asset: neoplasm occurrence is suppressed in chemical-induced early-stage HCC when dosed with PGZ (143). This protective effect is also corroborated by a phase I clinical trial, where researchers found PGZ is well tolerated in prevention against radiationinduced cognitive decline (RICD) among patients undergoing radiotherapy (89).
Several phase II clinical trials were set out to test the efficacy of PGZ as an adjuvant agent in treating patients of refractory or advanced cancer. A combinational scheme consisting of trofosfamide, rofecoxib (a COX-2 inhibitor currently not in market) and clinically relevant doses of PGZ has exerted encouraging results in patients with chemo-refractory melanoma and soft tissue sarcoma, as well as advanced vascular malignancies, bringing stabilization and remission (90,91).
Despite all the positive evidence, unfortunately, clinical data are depicting another not-so-inspiring picture. Dosage of PGZ in T2D patients has been suggested to be associated with risk of bladder cancer (144)(145)(146), and the longer of the term, the higher of the risks (147,148). Yet further studies shall be carried out on the specific histological types of neoplasms, as well as the differences and biases between dosage course, diabetic pathology, as evidenced by others contending the decreased incidence of breast cancer (149), the unaffected risks of bladder cancer with short-term dosage (150), and no incidence of malignancy in normal urothelial transitional epithelium (NUTE) cells treated with PGZ (151). Concerns on the protumorigenic effect PGZ has been witnessed to express offer another explanation: the dichotomized behavior, where the activation of PPAR-g inhibits the invasion of cancer cells yet fuels the tumorigenic transition of the myeloid lineage (152). The "permissive" cell condition theory brought forward by Annicotte et al., where PGZ only effect as an anti-tumorigenic agent in the presence of valproic acid, an HDAC inhibitor, can counteract the mentioned pro-tumorigenic effects (153).

Rosiglitazone
Rosiglitazone (ROSI) is another high-affinity PPAR-g agonist, a second-generation TZD, originally developed as an insulin sensitizer for the treatment of diabetes. ROSI was patented in 1987 and approved for medical use in 1999, however, following its annual sales peak in 2006 at approximately $2.5-billion, it diverged onto a slope of falling profits as sales plummeted to just $9.5-million in 2012 after a 2007 meta-analysis linking the use of ROSI as an anti-diabetic with increased risks of cardiovascular incidents and heart attacks (154). Some urged FDA to take ROSI off the market, and from 2011 to 2013, stricter restrictions were forced by FDA on its purchase; but later in 2013, those restrictions were lifted after FDA's review on a 2009 trial failing to show the alleged heart attack risks.
The anti-cancer effect upon PPAR-g activation made researchers interested in this property of ROSI, and recent data has been encouraging. Marked results are being observed on various human malignant tumors, including gastric, breast, colorectal, adrenocortical and pancreatic cancer types; in vivo, ROSI also exhibits a preventative effect against multiple cancers (155). Like other members of the TZDs family, ROSI's anticancer effect is yielded via both PPAR-g-dependent and PPAR-gindependent pathways, as well as both in monotherapy and in combined treatment with chemotherapeutics. In a PPAR-gdependent manner, the anti-tumor activities of ROSI include the induction of apoptosis and autophagy, proliferation and metastasis inhibition, amelioration of multidrug resistance (MDR), lifting of immune suppression, and inhibition of angiogenesis.

PPAR-g-Dependent Anti-Cancer Effects
The regulation of cell proliferation relies heavily on diverse signaling pathways, among which there are two important ones, the PI3K/AKT/mTOR and the MAPK/ERK pathways. There is a binding domain within the PTEN promoter, which is a natural inhibitor of the PI3K/AKT pathway, and ROSI increases the expression of PTEN through activating PPAR-g in HCC cell lines (92,93). As for the MAPK/ERK pathway, fulfilling the communication of signals from a membrane receptor to the cell nucleus, activated PPAR-g halts the phosphorylation of ERK, thus inhibiting the tumor growth (94,95). Apoptotic cell factors are vital participants in the process of cell apoptosis, hence the inhibition of proliferation, and evidence is accumulating on the effects of ROSI in this perspective. Among a plethora of factors, there is FasL, inducing apoptosis by binding to its receptor Fas, whose expression ROSI is observed to enhance (156); COX-2's overexpression may be protumorigenic, and its inhibition has been demonstrated to improve therapy outcome in colon cancer, upon the activation of PPAR-g by ROSI, and vice versa (93,98); NHE1, a osmotic homeostatic regulator whose decreased expression give rise to tumor cell death sensitization, whose promoter has a PPRE that PPAR-g can bind to and inhibit its expression, has been reported to be significantly decreased in expression by ROSI in T2DM patients with breast cancer (96); BRCA1, a tumor suppressor gene whose overexpression induces apoptosis, is heavily PPARg-responsive, whose protein expression can be promoted to induce apoptosis by ROSI administration in human MCF-7 cells (97); the Bcl-2 family is a potent modulator of cell apoptosis with both inhibiting and accelerating gene products, and ROSI induces apoptosis by upregulate the expression of one accelerating factor, Bax (98). Tumor cell apoptosis could be counteracted by autophagy, where apoptotic mediators get degraded; on the other hand, excessively activated autophagy causes cell death (157,158). In an in vitro study, ROSI inhibits the proliferation of adrenocortical cancer cells by promoting autophagy through increasing the expressions of beclin-1 and Lamp-1 (99).
Metastasis is a distinctive behavior of malignant tumors, and a major reason for failed treatments. According to a recent study, ROSI is capable of downregulating C-X-C motif chemokine 12 (CXCL12)-induced migration, invasion and PI3K/Akt activation in a prostate cancer cell line, through the inhibition of the CXCL12/C-X-C chemokine receptor type 4 (CXCR4) axis (100). Apart from CXCR, ROSI has been found to suppress the expression of some migration-associated genes, including MMP-7, COX-2 and TIMP-1, consequently reducing tumor metastasis (101).
Multidrug resistance (MDR) is another key issue in the way toward satisfying treatment results. Encouragingly, ROSI exerts reversing effects in chemo-resistant ovarian cancer and melanoma cells via suppression of Frizzled-1 (FZD1), which enhances the inhibition of Wnt/b-catenin pathway, and the subsequent decreased expression of MDR1/P-gp, which is a common cause of MDR (102).
PPAR-g has been known for its anti-inflammatory properties, majorly contributed by the reduction of inflammatory cytokines; in a tumor microenvironment, tumor-associated inflammation could result in immune suppression involving T regulatory cells and myeloid-derived suppressor cells (MDSC), whose accumulation is a great threat to a successful cancer treatment. ROSI can reverse the immune-suppressed state by limiting early MDSC accumulation and intra-tumoral T regulatory cells in pancreatic cancer mice model (103,104).
Angiogenesis is a vital process in tumor growth and metastasis. Basic fibroblast growth factor (bFGF) and VEGF, two mitogenic cytokines, are highly responsible for endothelial cell growth and differentiation, hence angiogenesis. In a nontumor model, ROSI has been reported to inhibit the activities of those two factors; in myeloma cell line, ROSI inhibits insulin-like growth factor 1 (IGF-1) or HIF-1a, which could boost VEGF's pro-angiogenesis effects, via PI3K/AKT and ERK signaling in a PPAR-g-dependent fashion (105, 106).

PPAR-g-Independent Anti-Cancer Effects
Despite its nature being a PPAR-g agonist, ROSI could still partially exert its anti-tumor effects even with PPAR-g inhibitor or siRNA knockdown.
Among all effects, induction of apoptosis may be the most preserved without activating PPAR-g pathway, which could be achieved via inhibition of nuclear factor kappa-lightchain-enhancer of activated B cells (NF-kB) pathway and prostaglandin E2 (PGE2), as well as regulation of other apoptosis-associated cell factors. HCaRG, a novel calciumregulated gene, is found to be in control of cell proliferation and differentiation, and has been reported to under-express in cancer cells and more expressed in normal tissue adjacent in patients with renal cell carcinomas (159). According to Solban et al., ROSI induces apoptosis by inhibiting NF-kB activity through upregulating HCaRG (160). 15-PGDH is a tumor suppressor gene which is responsible for inactivating PGE2, and limited degradation of the latter might lead to increased tumor growth not only by promoting apoptosis resistance, but also angiogenesis, tumor invasiveness, as well as inhibiting immune surveillance. Hazra et al. indicated that ROSI increases the expression of 15-PGDH to inhibit NSCLC growth (107). Other cell factors in the play of apoptosis include Jab1, a potential oncogene and critical in tumorigenesis, which could be down-regulated by ROSI in liver cancer cells (108); death receptor 5 (DR5), a member of the tumor necrosis factor (TNF)-receptor superfamily, which can be activated by TNFrelated apoptosis inducing ligand (TNFSF10/TRAIL/APO-2L), and transduces apoptosis signal, and to be found up-regulated by ROSI to enhance TRAIL-induced apoptosis in various cancer cells (109); tuberous sclerosis complex 2 (TSC2), a tumor suppressor protein and modulator of the mTOR pathway, could be activated by ROSI with the consequent suppression of mTOR signaling, cell cycle halting, and inhibits growth of NSCLC cells (110).
Epithelial-mesenchymal transition (EMT) is a process highly associated with cancer progression and metastasis, where epithelial cells gain migratory and invasive properties (161). MET is the opposite. According to Tai et al., ROSI inhibits tumor proliferation and metastasis via inducing MKP-1 through promoting MET, consequently suppressing both tumor proliferation and metastasis (162).

ROSI in Monotherapy and Combined Treatments
There are a plethora of studies focusing on utilizing ROSI in monotherapy as well as an adjuvant drug in combined administration with other treatments, which is showing a promising picture when put together. In cell cultures and xenografts, ROSI exhibits anti-tumor functions, involving apoptosis, cell growth, differentiation, migration, invasion, and tumor associated angiogenesis inhibition, as well as immune system regulation, in various malignancy, and some of them have entered phase II trials. When used in combination, ROSI exerts an enhancing effect upon chemo drugs or radiosensitivity, and/or a drug-resistance rescuing function. Several phase I cancer clinical trials are carrying out ROSI's effects associated with chemotherapeutics (155).

Other Modulators of the TZDs Family
Apart from the aforementioned well-known members of the TZDs family, other modulators in this category have been reported to own certain anti-tumor properties, including the amelioration of multidrug resistance, differentiation induction, etc.
Efatutazone is a novel PPAR-g agonist and of the third generation of TZDs. Two separate studies by Ni et al. both reported that Efatutazone facilities the treatment of EGFR-TKIresistant lung adenocarcinoma, by promoting the protein expression of PPAR-g and phosphatase and tensin homolog (PTEN), causing the inactivation of the Akt pathway without affecting the transcriptional levels, which exerts a synergistic effect with LXRa, a member of another class of nuclear hormonal receptor reported being potential targets for the prevention and treatment of multiple cancers, agonist T0901317 (111,112). Ductal carcinoma in situ (DCIS), a pre-invasive breast lesion, is the precursor of invasive ductal carcinoma. Efatutazone with short-term and low-dose administration in MCFDCIS xenograft and C3(1)/Tag transgenic mice induce a higher level of lactational and luminal cell differentiation and delay this process, reportedly to having activated PPAR-g and subsequently downregulating Akt phosphorylation but leaving the ERK pathway unaltered (113). Balaglitazone is another member of the TZDs family. Upon treatment with balaglitazone, the expression of PPAR-g and PTEN elevated, which leads to the P-glycoprotein-mediated multidrug resistance partly reverses in doxorubicin-resistant human myelogenous leukemia (K562/DOX) cells (114).

Non-TZD PPAR-g Modulators
Except for TZDs, there are other PPAR-g ligands uncovered with anti-cancer properties.
The overactivation of WNT/b-catenin pathway has been associated with oxidative stress and inflammation, consequently contributing to carcinogenesis. It has been suggested that in gliomas, canonical WNT/b-catenin pathway and PPAR-g act in an opposed fashion (163). Resveratrol, a polyphenolic phytoalexin extracted from grape, inhibits uterine sarcoma cells proliferation by blocking WNT/b-catenin pathway in a PPAR-g-dependent manner (115). Another PPAR-g agonist, Curcumin, can inhibit the WNT/b-catenin pathway, as well as control circadian clocks relevant to this particular pathway, to restrict the growth of tumor (164). In addition, in diffuse large B cell lymphoma, Curcumin could also induce cell apoptosis and G2 cell cycle arrest by inhibiting the Akt/mTOR pathway (116). Bavachinin (Bnn) is extracted from a type of Chinese herb, named Psoraleacorylifolia. Bnn promotes the activation of PPAR-g and generates ROS to suppress the growth of A549 lung cancer cell line (117). Cladosporols A is a natural metabolite of Cladosporium tenuissimum, which can fully activate PPAR-g, similar to Rosiglitazone, and Cladosporols B is the oxidated product of the former, which owns partial PPAR-g agonizing effect. Both of them exemplify the capacity to limit the proliferation of colon HT-29 cancer cells while the latter is more efficient (118). Hu et al. demonstrated that Alpha-Tocopherol, one component of Vitamin E, targets PPAR-g and increases the expression of its downstream PTEN whereas downregulates Akt in rat esophageal cancer model induced by N-nitrosomethylbenzylamine (NMBA), indicating the chemoprevention role of Alpha-Tocopherol in early stage of esophageal cancer (119). Hydroxysafflor-Yellow is another PPAR-g ligand originated from safflower. It is effective in limiting gastric cancer cells BGC-823 proliferation through apoptosis mechanism (120). Triple-negative breast cancer (TNBC) is an aggressive subtype with the absence of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor-2 (HER2). Despite the declining mortality of breast cancer globally, owning to earlier diagnosis, TNBC subtype still causes high fatality, with metastasis and relapse being major clinical features in the late stage of TNBC and causing death directly. A novel PPAR-g agonist VSP-17 targets EMT, indicated by elevated E-cadherin level, successfully inhibiting the metastasis of MDA-MB-231 cell line and xenograft model (121).

DISCUSSION AND CONCLUSION
Cancer has long been regarded as the most detrimental disease for its high morbidity and mortality. However, it is gratifying that many novel therapeutic targets, such as PPAR-g, have been proved to show great potential in the inhibition of tumor proliferation and metastasis. By analyzing recent basic research, we have elucidated the functions of PPAR-g to interpret the underlying mechanisms of PPAR-g modulators in exerting anti-tumor effect. PPAR-g regulates metabolism of lipid and glucose, inhibits inflammation and immunity responses, suppresses cell proliferation and induces cell differentiation and apoptosis. It is reported that upon PPAR-g activation, the fatty acid oxidation in T cells are increased, and the subsequent increase in the proliferation of Tumor-Reactive CD8 + T Cells facilitates the efficacy of anti-PD-1 therapy (165). On the contrary, PPAR-g appears to promote tumorigenesis via enhancing intercellular adhesion and restraining apoptosis.
With the expression and secretion of inflammatory factors and pro-inflammatory chemokines suppressed, resistance to immunotherapy is likely to form in a PPAR-g dependent manner, which immensely diminishes anti-tumor potential of wild type T cells compared with PPAR-g-/-T cells (166). Moreover, PPAR-g modulators can be divided into TZDs family which comprises three generations and non-TZD natural compounds mainly derived from herbs and plants.
Based on the analysis of laboratory results and clinical study, we can conclude that PPAR-g modulators play an auxiliary role rather than a leading role in tumor treatment. Compared with promising but paradoxical performance in weakening cancer proliferation and metastasis, the ability of PPAR-g modulators to reinforce the tumor cell sensitivity to the therapies including radiotherapy, chemotherapy and TKIs is more pragmatic and closer to clinical application. In addition, whether the anti-tumor mechanism of TZDs is realized via PPAR-g and which tumors are prone to PPAR-g modulators remain to be further discovered. By establishing this thorough review on PPAR-g modulators, we hope to provide valuable insight into how we can better tackle cancer in the future.