{"paper_id":"c9bb59b4-48f0-4771-b49b-50bf455acadb","body_text":"Angiogenesis, the development of new capillaries from preexisting\nmicrovessels, plays a crucial role in several normal physiological processes,\nsuch as embryonic development, ovulation, wound healing, as well as tissue and\norgan regeneration. Angiogenesis also constitutes a crucial step in the aetiology\nof diverse pathological states, including cancer, diabetic retinopathy,\nage-related macular degeneration, psoriasis, and rheumatoid arthritis [ 1 ,  2 ]. In the last few years, the complicated\nbiochemical mechanisms governing neovessel formation have been well\nestablished. These include the proliferation of endothelial cells (ECs) from\npreexisting capillaries, the breakdown and reassembly of the extracellular\nmatrix (ECM) and the morphogenic process of endothelial tube formation [ 2 ,  3 ]. Numerous growth factors,\nincluding vascular endothelial growth factor (VEGF) family, basic fibroblast\ngrowth factors (bFGFs), platelet-derived growth factor (PDGF), hepatocyte\ngrowth factor (HGF), placenta growth factor (PGF), matrix metalloproteinases\n(MMPs), ephrin family, angiopoetin-1 (Ang-1), interleukins (IL-2, -6, -8), as\nwell as various endothelial surface molecules such CD31, CD34, CD36, CD144, and\na v b 3  integrins, have been found to control essential\nsteps within angiogenesis process [ 1 – 3 ].\nThe generation and release of antiangiogenic factors, such as interferon (INF)\n- α , - β , - γ , platelet factor 4 (PF4), and tissue\ninhibitors of MMPs (TIMPs) contribute to the coordinated downregulation of the\nangiogenic process within physiologic angiogenesis [ 4 ].\nPeroxisome proliferator-activated receptors\n(PPARs) are members of the nuclear hormone receptor superfamily of ligand-activated\ntranscription factors and include three different isotypes: PPAR- α , PPAR- β / δ , and PPAR- γ  [ 5 ,  6 ]. PPAR- γ , the most extensively studied amongst them,\nfunctions as ligand-activated transcription factor by binding to specific DNA\nsequences, termed to as peroxisome proliferator response elements (PPREs), in\nthe promoter of the target genes only as a heterodimer with the retinoid X receptor\n(RXR) [ 7 – 9 ]. PPRE has been\nmainly identified in the upstream regulatory sequences of genes related to\nmetabolic pathways [ 7 – 9 ]. In addition, recent studies have\nrevealed that PPAR- γ  can regulate gene expression independently of PPRE, either\nby suppressing growth hormone protein-1 (GHP-1), a transcription factor\ninvolved in pituitary specific gene expression, or by interfering with the\nfunction of activator protein-1 (AP-1), signal transducer and activator of\ntranscription-1 (STAT-1) and nuclear factor- κ B (NF- κ B) [ 7 ,  10 – 12 ]. In this context, the\nidentification of a sumoylation-dependent pathway by which PPAR- γ  represses transcriptional activation of\ninflammatory response genes has recently been reported [ 13 ]. This mechanism\nprovides a possible explanation for how ligand-bound PPAR- γ  activation can be converted from an activator\nof transcription to a promoter-specific repressor of NF- κ B target genes [ 13 ].\nA wide range of natural and synthetic structurally diverse compounds has\nbeen reported as potent PPAR- γ  ligands. The long chain polyunsaturated fatty acids and their\nderivatives, such as 15-deoxy-Δ 12,14 -prostaglandin J 2  (15d-PGJ2), as\nwell as nitrolinoleic\nacids are known natural occurring PPAR- γ  ligands [ 14 ,  15 ]. Recently, curcumin, a well-documented anticancer\nphytochemical component of turmeric, has been shown to exert anti-inflammatory\nfunctions via upregulation of PPAR- γ  activation [ 16 ]. Thiazolidinediones\n(TZDs) and tyrosine-based derivatives, such as glitazars (tesaglitazar,\nfarglitazar), constitute the most well-known synthetic ligands [ 17 ,  18 ], while relatively lower binding affinity for PPAR- γ  has also been reported for some nonsteroidal anti-inflammatory drugs (NSAIDs) [ 19 ]. TZDs represent a promising class of oral antidiabetic agents,\nsome of which are already marketed drugs (pioglitazone-PGZ and rosiglitazone-RGZ)\nfor the treatment of type II diabetes mellitus [ 20 ].\nInterestingly, a wide spectrum of action for TZDs beyond the treatment of\ndiabetes, including anti-inflammatory and antineoplastic properties, as well as targeting signaling pathways\nimplicated in atherosclerosis and osteoporosis has been reported [ 21 – 23 ]. In the last decade, more than 1000 PPAR- γ  ligands belonged to several distinct chemical classes have\nbeen synthesized and evaluated for their binding and transactivation to their\nreceptor. In this aspect, screening drug-like characteristics in the chemical space of PPAR- γ \nligands have currently been considered as an emerging demand in the aim to\ndiscover more potent compounds with improved absorption, distribution,\nmetabolism, excretion/toxicity (ADME/Tox) properties, avoiding potential toxic\nside effects, as well as pharmacokinetic and pharmacodynamic problems [ 24 ,  25 ].\nTo date, there has been a substantial accumulation of evidence that PPAR- γ  ligands exert\nregulatory effects on angiogenesis process related to diverse disease states,\nincluding cancer and diabetes [ 26 – 28 ]. It is also well documented that they directly affect tumor cells by inhibiting cell growth and inducing\ncell differentiation and apoptosis in various cancer types [ 21 ,  29 ,  30 ]. In view of the fact that\nangiogenesis is implicated in tumor development and metastasis and its\ninhibition could serve as potent antitumor side-therapeutic approach, the current\nreview summarizes the latest knowledge of the role of PPAR- γ  ligands in angiogenesis related to cancer, highlighting in\nthe underlying mechanisms.\n\nTumor angiogenesis constitutes an\nessential component of tumor growth, invasion, and metastasis that depends on a\nnet balance of angiogenic and antiangiogenic mediators, which are secreted by\nboth tumor and host infiltrating cells [ 31 ]. Currently, it is well established that\nthis dynamic balance between angiogenic stimulators and inhibitors, controls\nthe angiogenic signaling cascade governing the transformation of a tumor from a\nnonangiogenic to an angiogenic phenotype [ 32 ]. The acquisition of angiogenic phenotype has been considered\nas a rate-limiting step in tumor progression, which allows the tumor to\ntransform from a small lesion to a rapidly expanding mass with metastatic\npotency [ 33 ]. On the other hand, human tumors arise in the absence of\nangiogenic activity and may exist in a microscopic dormant state for months to\nyears without neovascularization [ 34 ]. In this context, hypoxia, developed\nwithin rapidly proliferating tissues or as a result of the occlusion of blood\nvessels, has been considered as a primary physiological regulator of the\nangiogenic switch [ 35 ]. The key mediators of this response are members of the\nhypoxia-inducible factor (HIF) family of proteins that function as\ntranscriptional regulators, stimulating the expression of a multitude of genes\nimportant for oxygen homeostasis [ 36 ,  37 ]. In addition, HIF has been found to\nenhance the expression of several angiogenic mediators, including VEGF-R1,\nVEGF-R2, Ang-1, Ang-2, MMP-2, and MMP-9 in malignant tumors [ 36 ,  38 ].\nIn response to hypoxia, tumor cells\nturn on the angiogenic signaling cascade by secreting various potent angiogenic\nmediators, such as VEGF, PDGF, bFGFs, angiopoetins, HGF, fibronectin, and\nheparanase that in turn activate endothelial cells of preexisting capillaries\nto produce MMPs for the collapse of ECM [ 39 ]. Degradation of ECM by MMPs allows\nendothelial cells to migrate in response to chemotactic growth factors, including\nVEGF, PDGF, and bFGFs [ 33 ,  39 ]. Members of CXC chemokine family, such as IL-2,\n-6, -8, and integrins  α \n v \n β \n 3 , are also involved in the angiogenic cascade. It should be noted that\nin the case of high progressive tumors, the release of endogenous\nantiangiogenic factors are insufficient to counteract the net effect of\nangiogenic ones. Thus, the formation of new blood vessel is formed after\nattracting accessory cells, mainly pericytes and smooth muscle cells, producing\na new basement membrane and a firm ECM [ 39 ,  40 ]. The above-mentioned angiogenic\nmediators have been joined by others including Notch/Delta, semaphorin, ephrin,\nand roundabout/slit families of proteins [ 40 ]. Besides this, blockage of NF- κ B activity has been shown to reduce VEGF gene\nexpression in highly malignant tumor cells, since a binding site for this\ntranscription factor has been identified within the VEGF promoter [ 41 ]. Each\nof the sequential steps within angiogenic cascade could be considered as a\npotential single target for the development of new drug candidates against\ntumor vasculogenesis.\nCurrently, numerous therapeutic approaches have been\ndesigned in the aim to control tumor angiogenic cascade by targeting the above-mentioned\nangiogenic mediators [ 40 ,  42 ]. In this context, more than a few angiogenesis inhibitors have\nalready been approved for the treatment of cancer, while several compounds are\nin the late stage of clinical trials. The main category of the antiangiogenic\ncompounds exerts its action indirectly either by neutralization of\ntumor-derived angiogenic factors or preventing the receptors/signaling pathways\nof these growth factors. In this regard, VEGF isoforms and their tyrosine\nkinase receptors VEGFRs, as well as epidermal growth factor (EGF) and its\nreceptor (EGFR) are currently explored in clinical trials as drug candidates\nagainst cancer [ 43 – 45 ].\nWith\nrespect to angiogenesis inhibitors, several angiostatic compounds, such as\nendostatin, thrombospodin-1 (TSP-1), tumstatin, angiostatin, and 16-kDa\nN-terminal fragment of human prolactin (16K hPRL) have been reported to directly\nand selectively suppress endothelial cell migration inducing EC apoptosis and\ncell cycle arrest within tumor neovascularization [ 46 ,  47 ]. It should be\nmentioned that most of these angiostatic compounds are also naturally occurring\nmolecules that compensate with angiogenic factors in order to control\nangiogenic cascade in normal physiologic conditions. In addition, targeting MMPs\nby such agents has been reported, underlining the importance of ECM remodeling\nduring angiogenesis process. Activation of NF- κ B may also be a possible mechanism of such\nangiostatic agents to induce EC apoptosis and to improve immune response within\nangiogenesis process [ 46 ,  47 ].\n\nPPAR- γ  ligands can regulate\ntumor angiogenesis via direct effects on ECs proliferation and migration and/or\nthrough indirect mode of action by affecting the counterbalance between\nangiogenic and antiangiogenic mediators ( Figure 1 ,  Table 1 ).\nPPAR- γ  has been reported to be expressed in endothelial cells and PPAR- γ  ligands are well established to exert direct effects on them [ 48 ,  49 ]. PPAR- γ  activation by either naturally occurring or\nsynthetic ligands resulted in potent inhibition of growth factor-induced differentiation\nand proliferation in human\numbilical vein endothelial cells (HUVECs) and choroidal\nendothelial cells (CECs) [ 48 ,  49 ]. In this regard, PPAR- γ  dependent mode of action has been shown to stimulate caspase-mediated\nECs apoptosis [ 50 ]. Importantly,\nRGZ levels able to inhibit ECs proliferation are readily achieved in patients\nundergoing standard antidiabetic RGZ treatment [ 51 ]. Moreover, both RGZ and PGZ, at relative pharmacological\nconcentrations, resulted in a strong prevention of VEGF-induced tube formation\nand ECs migration [ 52 ,  53 ]. Mechanistically, it has been supported that angiogenesis inhibition by RGZ in HUVECs\ninvolves a proapoptotic mechanism which includes the implication of the PPAR- γ -mediated NO production and the maxi-K channel activation [ 54 ]. Maxi-K\nchannels, essential mediators of vascular remodeling and angiogenesis, are\nsynergically regulated by various intracellular second messengers including NO\n[ 54 ]. Hence, a possible proapoptotic mechanism for the PPAR- γ -mediated NO production has been\nsuggested [ 55 ]. Recently, pigment epithelium-derived factor\n(PEDF), a potent antiangiogenic glycoprotein, has been shown to stimulate\nHUVECs apoptosis through sequential induction in the expression and\ntranscriptional activity of PPAR- γ . PEDF upregulated\np53 expression via PPAR- γ , supporting evidence that p53 may be a major\ntarget in PPAR- γ  mediated ECs apoptosis [ 56 ].\nPPAR- γ  has\nalso been reported to be expressed in tumor ECs, presenting a relative\noverexpression in tumor-induced endothelial sprouts compared to normal endothelium.\nIn this case, endothelial and tumoral cells have been shown to display\ninhibition even at low TZDs doses [ 57 ]. Importantly, TZDs inhibited tumor cell\ninvasion across blood vessel endothelium. In fact, RGZ at concentrations close\nto the range of its binding affinity for PPAR- γ \n[ 8 ] exerted inhibitory effects on tumor angiogenesis in malignant cell lines\nand in immunodeficient mice with transplanted tumors [ 57 ]. In this regard, it\nshould be mentioned that RGZ concentrations of 5  μ M and higher led to the phosphorylation of\neIF-2 α \nin HUVECs, supporting\nevidence that the inhibition of ECs proliferation could also be mediated\nthrough a PPAR- γ \nindependent pathway. However, at even lower concentration range (0.1–1  μ M), at which PPAR- γ  is activated, RGZ was capable of exerting even stronger\nantiproliferative effects on ECs in vitro [ 57 ]. In this\ncontext, the concentration range of PPAR- γ  ligands should be taken into\ncareful consideration, because over a concentration limit, which may be varied\namongst the different types of cells, in vitro, as well as amongst different\nspecies, in vivo, receptor-independent actions could be elicited. Such PPAR- γ  mode of action has recently been reviewed by\nFeinstein et al., who suggested an alternative mitochondrial target for TZDs,\ntermed as mitoNEET [ 12 ]. To this point, it should be noted that higher doses of RGZ were less\neffective in inhibiting angiogenesis and hence lung metastasis than lower doses\nthat are actually comparable to the serum levels of RGZ in diabetic patients [ 27 ,  51 ]. Overall, although PPAR- γ  ligands can also induce EC apoptosis as mentioned in the previous\nparagraph, it is unlikely that they do this under physiological conditions as\nthis may result in a severe thrombosis. Thus, it should be emphasized the fact PPAR- γ  ligands may target better EC proliferation as shown by Panigrahy et\nal. [ 27 ] and Freed et al. [ 51 ].\nOrthotopic\nimplantation of H2122 nonsmall cell lung adenocarcinoma cells overexpressing\nPPAR- γ  into the lungs of nude mice attenuated tumor\ngrowth and metastasis by selective inhibition of invasive metastasis, and\nactivation of pathways that promote a more differentiated epithelial phenotype\n[ 73 ]. This evidence deserves special attention since both\nangiogenesis and invasion are crucial for the formation of metastasis and the recurrence\nof tumors. Moreover, reintroduction of exogenous TSP1 or its peptide\nderivative ABT510 can reverse the angiogenic switch, and thus blocking tumor\nexpansion. TSP-1 is a well-known potent angiogenesis inhibitor that targets ECs\nfor apoptosis through signaling cascade at its receptor CD36. In tumor\nxenografts, TGZ, RGZ, and 15d-PGJ2 coupled to ABT510 suppressed angiogenesis\nand induced ECs apoptosis in a CD36 dependent manner [ 62 ]. In this context, 15d-PGJ2\ntreatment upregulated CD36 surface expression in human monocytic cell line THP-1\nby enhancement of CD36 gene transcription [ 74 ]. Thus, PPAR- γ \ncould be considered as a critical regulator of CD36 expression, as both natural\nand synthetic PPAR- γ  ligands are capable of increasing CD36\nexpression [ 75 ].\nReceptor-mediated effects for PPAR- γ  ligands\nin inhibiting angiogenesis through direct mode of action on endothelium seem to\nbe dominated [ 28 ,  57 ]. In this regard, PPAR- γ  knockout mice embryos died on day 10 of life because of interference\nwith the terminal differentiation pattern of trophoblasts, as well as the loss\nof vascular development in the placenta [ 76 ,  77 ]. It has also been suggested\nthat PPAR-binding protein (PBP), a coactivator of PPAR- γ , may\nfunction as a negative modulator of ECs proliferation [ 77 ]. Such genetic data\nprovides additional evidence that PPAR- γ  functions as modulator of angiogenesis; however, receptor-independent action should\nnot be excluded. In this aspect, Artwohl et al. showed PPAR- γ -independent\nantiproliferative effects on HUVECs associated with lactate release, possibly\ndue to inhibition of mitochondrial function [ 78 ].\nBeyond the direct mode of action on\nthe endothelium, PPAR- γ  ligands have been reported to downregulate\nangiogenesis process via indirect mechanisms by modulating the levels of the endogenous angiogenesis mediators ( Figure 1 ,  Table 1 ). In this context, VEGF/VEGFR signaling pathway seems to be a key target for PPAR- γ \nligands in inhibiting angiogenesis. Xin at al. provided the first evidence that 15d-PGJ2 reduced VEGFRs\nm-RNA levels in HUVECs [ 48 ]. It\nhas also been supported that PPAR- γ  ligands may have bifunctional properties in\nKDR gene expression that involve the enhancement of Sp1-DNA binding in absence\nof ligand by PPAR- γ  itself and the suppression of Sp1-DNA-binding in presence\nof PPAR- γ  ligands [ 79 ]. Moreover, PPAR- γ  activation has been shown\nto downregulate leptin and tumor necrosis factor (TNF- α ), two well-known angiogenesis-inducing factors\n[ 80 ,  81 ]. In fact, PPAR- γ  activation by TZDs attenuated leptin gene\nexpression both in vivo and in vitro [ 82 ,  83 ] and blocked\nleptin-induced ECs migration through inhibition of Akt and eNOS signaling [ 84 ].\nThis evidence suggests that endothelial phosphatase and tensin homologue\nmutated on chromosome ten (PTEN), a negative regulator of PI3K → Akt signaling,\nmay play a crucial role in the ECs antimigratory actions of TZDs [ 84 ].\nTumor-associated angiogenesis has\nbeen reported to be indirectly suppressed by blocking the expression of\nangiogenic stimulators in response to PPAR- γ \nligand activation. In this regard, PPAR- γ \nactivation by TGZ or PGZ diminished the production of the angiogenic ELR + CXC\nchemokines IL-8 (CXCL8), ENA-78 (CXCL5), and Gro- α  (CXCL1) in human non-small-cell lung cancer\ncell line A459 [ 63 ]. This effect was ascribed to the negative modulation of NF- κ B activation [ 63 ]. In\naddition, CGZ was found to decrease PGE 2  production through\ndownregulation of cyclooxygenase-2 (COX-2) expression in human non-small-cell\nlung carcinoma A427 and A549 cell lines [ 64 ]. Interestingly, utilization of a\ndominant negative PPAR- γ  construct\nrevealed that the effect of CGZ on both COX-2 and PGE2 was mediated through\nPPAR- γ \nindependent pathways [ 64 ]. Another study demonstrated that 15d-PGJ2 attenuated\nthe expression of Ang-1 and hence the angiogenic process through the\nangiopoietin-Tie2 system in the gastric cancer cell line MKN45 [ 67 ]. Ang-1 is\ninvolved in the regulation of maturation and stabilization of the vascular wall,\nand thus it might be a potential target for inhibiting tumor angiogenesis. Moreover,\nin a model of human anaplastic thyroid carcinoma, RS5444, a novel high-affinity\nPPAR- γ  agonist exerted\npotent antiangiogenic action, in vivo, by decreasing CD31, a specific molecular\nmarker of blood vessels [ 71 ]. In this regard, PPAR- γ  ligand treatment (TZDs,\n15d-PGJ2, and RS1303) dose-dependently suppressed cell proliferation by\ninducing apoptosis instead of differentiation in five human anaplastic\ncarcinoma cell lines (MSA, IAA, ROA, K119, and KOA-2) [ 61 ]. Recently, CGZ has also\nbeen shown to produce antitumor effects against ovarian cancer, in vitro and in\nvivo, in conjunction with reduced angiogenesis and induction of apoptosis [ 65 ].\nIn this case, CGZ induced antitumor effects were comparable to that of\ncisplatin and were ascribed to inhibition of VEGF production in relation to PGE 2  reduction, an endogenous stimulator of angiogenesis and invasiveness [ 65 ]. PPAR- γ  ligands have also been shown to repress VEGF\ngene expression via a PPAR- γ -responsive\nelement (PPRE) in the VEGF gene promoter in both primary and transformed human\nendometrial cell cultures [ 60 ]. This study provided substantial evidence that\nPPAR- γ  ligands may be exploited pharmacologically to\ninhibit pathological vascularization in complications of pregnancy,\nendometriosis, and endometrial adenocarcinoma [ 60 ].\nAs mentioned in  Section 3.1 , RGZ\nsuppressed tumor angiogenesis by direct mode of action in endothelium; however,\nindirect antiangiogenic effects have also been reported [ 57 ]. More to the\npoint, RGZ, at low doses, in vitro,\ninhibited bovine capillary ECs and reduced VEGF production by tumor cells [ 57 ].\nRGZ also suppressed angiogenesis in the chick chorioallantoic membrane, in the\navascular cornea, in vivo, as\nwell as in a variety of primary tumors, such as glioblastoma U87 and Lewis lung\ncarcinoma cells, in vitro [ 57 ].\nLikewise, both PGZ and 15d-PGJ2 have been shown to inhibit, dose- and \ntime-dependently, VEGF and bFGF secretion in human renal cell carcinoma cells [ 66 ].\nImportantly, antiangiogenic effects were observed at the dose of 5  μ M PGZ, a level that is\nalso obtained in diabetic patients after standard PGZ treatment [ 66 ]. On the\nother hand, there is nonavailable data so far concerning the effect of PPAR- γ  ligand treatment on the expression\nand/or secretion of antiangiogenic mediators. In this regard, future studies focused\non the impact of PPAR- γ  ligands in mediators, such as endostatin,\nTSP-1, tumstatin, angiostatin, and 16K hPRL are strongly recommended.\nAngiogenesis constitutes a crucial\nstep for tumor invasion and formation of metastasis. In this aspect, PPAR- γ  ligand treatment attenuated\nthe invasiveness of pancreatic tumor cells, reducing MMP-2 and -9 protein\nlevels and activity [ 70 ]. Moreover, the secretion of the invasive factor tissue\nplasminogen activator (tPA) was decreased by RGZ treatment in pancreatic tumor\nAsPC-1 cells through receptor mediated mechanisms [ 58 ]. Treatment of the highly\naggressive human breast cancer cell line MDA-MB-231 with synthetic and natural\nPPAR- γ \nligands, at \nnoncytotoxic concentrations, also resulted in a\nsignificant inhibition of the invasive capacity [ 59 ]. In fact, TIMP-1 was\nupregulated by PPAR- γ  ligand\ntreatment, while the gelatinolytic activities of gelatinases in the conditioned\nmedia were decreased [ 59 ]. Moreover, PPAR- γ \nligands downregulated the invasive potential of anaplastic thyroid carcinoma\ncells, and this effect was prominent in 3 cell lines, which exhibited higher\nexpression level of the PPAR- γ  gene or protein [ 61 ].\nClinical evidence from a pilot study enrolled 6 patients\nwith angiosarcoma and hemangioendothelioma, revealed that the angiostatic\ntriple combination of PGZ, rofecoxib, and metronomic trofosfamide exhibited\nhigh efficacy in the palliative care of patients [ 85 ]. Until this study, antiangiogenic\ndrugs such PGZ and rofecoxib had not been considered for the treatment of human\nangiosarcoma. In support of this view, a case report study has demonstrated\nthat this novel antiangiogenic therapy was effective in a patient with endemic Kaposi\nsarcoma and led to partial remission that was stable for 18 months without\nsignificant side effects [ 86 ]. Hence, targeting PPAR- γ  may prove to be a potential therapeutic strategy in combined treatments\nwith conventional chemotherapy for patients with vascular disorders [ 87 ].\n\nThe most comprehensive data so far render\nPPAR- γ  ligands as potent inhibitors of angiogenesis;\nhowever, there are several lines of evidence to support that PPAR- γ  ligand activation can also trigger\nangiogenic cascade ( Table 1 ). In fact, increased VEGF mRNA levels and induction of angiogenesis in response to\nPPAR- γ  ligands treatment have been reported both in vitro and in vivo [ 88 – 91 ]. Interestingly, TZDs have been considered as potential\npharmacological agents for angiogenesis induction in the treatment of ischemic\nartery disease [ 89 ]. Recent clinical evidence has also demonstrated\nthat RGZ treatment improved endothelial progenitor cell (EPC) number and\nmigratory activity in diabetic patients [ 92 ,  93 ].\nIn addition, PGZ treatment was found to improve endothelial function by\nincreasing the number and the migratory capacity of EPCs in animal and human studies\n[ 94 ,  95 ]. Another study has revealed that eNOS upregulation induced by RGZ may\nbe the dominant mechanism through which RGZ enhanced angiogenesis [ 91 ].\nHowever, Gensh et al. did not observe upregulation of vascular eNOS mRNA\nexpression or setback of the PGZ-induced increase of EPCs in the presence of\n1-NAME, a NOS inhibitor [ 94 ]. These authors suggested that TZDs may regulate\nEPCs by a mechanism independent of eNOS [ 94 ]; however, further studies based on\npharmacologic blocking or knockout modeling of eNOS are strongly recommended in\norder for precise conclusion to be drawn. Importantly, taking into account the\ndiscrepancy in literature, Gensh et al. assumed that TZDs may play a\ndouble-edged role in angiogenesis signaling by promoting the number and\nmigration of EPCs at lower tissue concentrations obtained by systematic\ntreatment, whereas the antiangiogenic effects are elicited at higher local\nconcentrations [ 94 ]. This major remark has also been reported in the case of\nbreast cancer cells where low concentration of PPAR- γ  ligands increase cell proliferation in\ncontrast to the higher concentrations that suppress cell growth [ 96 ]. The urgent demand to define and monitor the dosage\nof PPAR- γ  ligands in clinical trials for cancer therapy\nis thoroughly discussed by Panigrahy et al. [ 27 ]. In this aspect, special\nattention deserves the fact that atorvastatin, a 3-hydroxy-3-methylglutaryl\ncoenzyme A (HMG-CoA) reductase inhibitor, and PGZ increased myocardial 15d-PGJ2 levels in the rat myocardium and HUVECs [ 97 ]. 15d-PGJ2 was produced mainly via COX-2 and activated PPAR- γ . Interestingly, it was\nsupported that PPAR- γ  activation was exclusively mediated by 15d-PGJ2 in the case of atorvastatin, whereas PGZ activated directly PPAR- γ  or indirectly via 15d-PGJ2 [ 97 ]. Thus, these recent findings raise the question\nwhether the final effect of PPAR- γ  ligands is completely ascribed to the dose of\nPPAR- γ  ligand\ntreatment or in addition to the induction of endogenous PPAR- γ  activators, such as 15d-PGJ2. It should also be taken into account the fact\nthat endogenous nitrated fatty acids that comprise a class of nitric\noxide-derived, PPAR- γ  dependent and cell signaling mediators can modulate systematic\ninflammatory responses within physiological concentration ranges [ 98 ].\nThere\nis also substantial evidence, which suggests that PPAR- γ  ligands stimulate tumor angiogenesis. In this\ncontext, 15d-PGJ2 treatment was\nfound to dose-dependently increase the VEGF mRNA expression in both human\nandrogen-independent PC-3 prostate and 5637 urinary bladder carcinoma cells [ 68 ].\nIn addition, 15d-PGJ2 resulted in upregulation of VEGF expression through the\ninduction of heme oxygenase (OH)-1 ERK1/2 phosphorylation in human breast\ncancer MCF-7 cells, thus contributing to increased angiogenesis in this type of\ntumor cells [ 69 ]. Nimesulide, a selective COX-2 inhibitor, although at\nrelatively high concentrations, enhanced VEGF secretion from pancreatic cancer\ncells in vitro, as well as from\nboth COX-2-positive and COX-2-negative pancreatic tumors through PPAR- γ  activation [ 72 ]. Importantly, in the case of COX-2-negative pancreatic tumors,\nnimesulide-stimulated VEGF production was considerably associated with enhanced\nangiogenesis and tumor growth [ 72 ]. Besides this, VEGF was differentially\nincreased, according to the differentiation state of the cells, by the three\nPPAR isotypes, - α , - β / δ , and - γ , in two different human\nurinary bladder cancer cell lines, RT4 and T24, derived from grade-I and\ngrade-III tumors, respectively [ 99 ]. The PPAR ligand-induced VEGF expression\nseemed to be PPAR-specific and involved an indirect mechanism requiring an\nintermediary regulatory protein through the MAP (ERK1/2) kinase pathway,\nprobably by a modulation of the phosphorylation state of PPARs [ 99 ]. Immunohistochemical\nanalysis in human bladder tumor specimens also revealed statistically\nsignificant associations between PPAR- γ  and several angiogenic factors, such as VEGF,\nbFGF, platelet-derived endothelial cell growth factor (PDECGF), and EGFR in\nrespect to the incidence of tumor recurrence or progression [ 100 ]. On the other\nhand, no statistically significant differences were observed between PPAR- γ  immunoreactivity and angiogenesis parameters in skin cancer, whereas\nthe microvessel density was significantly higher in actin keratosis and\nsquamous cell carcinoma that expressed PPAR- β / δ  [ 101 ]. These clinical data on PPAR- γ -induced signaling implicated in the expression of crucial angiogenic\nfactors in human neoplasia may unfold the development of new therapeutic\napproaches in those types of cancer in which excessive angiogenesis represents\na negative prognostic factor.\n\nAs hypoxia is a key regulator of the\nangiogenic switch, hypoxia-induced angiogenesis is gaining gradually increasing\ninterest as a potential target for cancer therapy. In human bladder tumors and\ncell lines, several components of the hypoxia response pathway, including HIF-1 α  and HIF-2 α  have been considered as\nimportant cofactors of the regulation of VEGF [ 102 ]. Recent findings have\nrevealed that PPAR- γ  can modulate arterial remodeling associated with hypoxic hypertension [ 103 ].\nIn fact, RGZ was found to attenuate and reverse pulmonary arterial remodeling\nand neomuscularization in rats subjected to chronic hypoxia [ 104 ]. Decreased\npulmonary arterial (PA) remodeling in RGZ-treated animals was associated with\ndecreased smooth muscle cell proliferation, decreased collagen and elastin\ndeposition, and increased matrix MMP-2 activity in the PA wall [ 104 ]. In this\naspect, PPAR- γ \nmRNA levels were found significantly lower in human adhesion fibroblasts\ncompared to normal ones in response to hypoxia [ 105 ]. Moreover, hypoxia has\ndemonstrated to reduce the mRNA levels of PPAR- γ  protein in human proximal renal tubular\nepithelial cells (HPTECs). However, knockout of HIF-1 α  with its dominant negative form did not block\nthe hypoxia-induced reduction in PPAR- γ  expression [ 106 ]. In this regard, substantial\nevidence has revealed that 15d-PGJ2 can modulate the activities of several\ntranscriptional factors, such as NF- κ B and AP-1, including also HIF-1 [ 107 ]. The\nregulation of the aforementioned redox-sensitive transcription factors by\n15d-PGJ2 was not necessarily mediated via PPAR- γ  activation, but rather involves covalent\nmodification or oxidation of their critical cysteine residues acting as a redox\nsensor [ 107 ]. Overall, targeting hypoxia-induced angiogenesis by PPAR- γ  ligands may prove to be a promising therapy\nfor the treatment of cancer; however, the precise mechanisms involved in\nhypoxia-induced angiogenesis process remain to be clarified.\n\nAt the present, there is quite a\nlot of evidence to support that PPAR- γ  may be considered as therapeutic\ntarget for diverse disease states in which excessive angiogenesis is\nimplicated, including cancer. The most comprehensive data so far have revealed that PPAR- γ  ligands are capable of inhibiting\nangiogenesis implicated in tumor malignant transformation and expansion. Targeting\nECs proliferation and migration seems to be a dominant effect of PPAR- γ  ligands on tumor angiogenesis.\nIndirect mechanisms that involve the counterbalance between a multitude of endogenous\nangiogenic and antiangiogenic factors further account for the inhibitory\neffects of PPAR- γ  ligands on\ntumor angiogenesis. According\nto these data, PPAR- γ  ligands may unfold new perspectives in clinical use against primary\ntumor growth and metastasis, since tumors that exhibit multidrug resistance are\neffectively targeted by antiangiogenic chemotherapy. Such perspectives could be\nclinically relevant, as PGZ and RGZ are orally administered FDA-approved drugs,\nalready been used by million patients undergoing standard antidiabetic\ntreatment.\nOn the other hand, there are\nseveral lines of evidence that PPAR- γ  ligands can also enhance tumor angiogenesis progression under certain\nconditions. This controversy could\nbe attributed to the pleiotropic action of PPAR- γ \nligands, possibly via cofactors, either coactivators or corepressors. Such\ndiscrepancies may also be ascribed either to differences in time and dose of\nPPAR- γ  ligand\ntreatment, or to differences among the various organisms and types of cells\nthat have been studied. It should be taken into account that angiogenesis is a\nmultifaceted process that involves a wide range of mediators capable of\ninducing or suppressing angiogenesis in addition to the degree of tissue\nhypoxia. Consequently, the final outcome is difficult to be assessed accurately\nand depends significantly on experimental models and/or treatment conditions. Moreover,\neach type of cancer in humans presents individual and distinct vascular pattern\non the microenvironment in which it is located. Thus, it should be taken into careful\nconsideration the type of cancer being treated when deciding an appropriate\ntherapeutic strategy.\nIn this aspect, the use of\ndifferent cancer models, in vitro\nand in vivo, are strongly recommended\nto further define the molecular interactions amongst PPAR- γ , angiogenic/antiangiogenic factors, and tumor progression markers\nwithin the distinct cancer types. Future research effort should also be\norientated to the clinical evaluation of PPAR- γ  expression in aggressive tumor\ncancers in which various angiogenic/antiangiogenic factors exhibit high\nprognostic value. Such studies could delineate the potential of PPAR- γ  ligands in future anticancer therapeutic strategies, either alone or combined\nwith conventional chemotherapy.","source_license":"CC-BY-4.0","license_restricted":false}