Fenofibrate in cancer : mechanisms involved in anticancer activity

To review the mechanisms of anti-cancer activity of fenofibrate (FF) Objective: and other Peroxisome Proliferator Activator Receptor α (PPARα) agonists based on evidences reported in the published literature. We extensively reviewed the literature concerning FF as an off target Methods: anti-cancer drug. Controversies regarding conflicting findings were also addressed. The main mechanism involved in anti-cancer activity is Results: anti-angiogenesis through down-regulation of Vascular Endothelial Growth Factor (VEGF), Vascular Endothelial Growth Factor Receptor (VEGFR) and Hypoxia Inducible factor-1 α (HIF-1α), inhibition of endothelial cell migration, up-regulation of endostatin and thrombospondin-1, but there are many other contributing mechanisms like apoptosis and cell cycle arrest, down-regulation of Nuclear Factor Kappa B (NF-kB) and Protein kinase B (Akt) and decrease of cellular energy by impairing mitochondrial function. Growth impairment is related to down-regulation of Phospho-Inositol 3 Kinase (PI3K)/Akt axis and down-regulation of the p38 map kinase (MAPK) cascade. A possible role should be assigned to FF stimulated over-expression of Tribbles Homolog-3 (TRIB3) which inhibits Akt phosphorylation. Important anti-cancer and anti-metastatic activities are due to down-regulation of MCP-1 (monocyte chemotactic protein-1), decreased Metalloprotease-9 (MMP-9) production, weak down-regulation of adhesion molecules like E selectin, intercellular adhesion molecules (ICAM) and Vascular Endothelial Adhesion Molecules (VCAM), and decreased secretion of chemokines like Interleukin-6 (IL-6), and down-regulation of cyclin D-1. There is no direct link between FF activity in lipid metabolism and anticancer activity, except for the fact that many anticancer actions are dependent from PPARα agonism. FF exhibits also PPARα independent anti-cancer activities. There are strong evidences indicating that FF can disrupt Conclusions: growth-related activities in many different cancers, due to anti-angiogenesis and anti-inflammatory effects. Therefore FF may be useful as a complementary adjunct treatment of cancer, particularly included in anti-angiogenic protocols like those currently increasingly used in glioblastoma. There are sound reasons to initiate well planned phase II clinical trials for FF as a complementary adjunct treatment of cancer. Referee Status:


Introduction
Fenofibrate (FF) is a drug of the fibrate class (a fibric acid derivative) that has been used since 1975 to reduce cholesterol (LDL and VLDL) and triglyceride levels and increase HDL in patients at risk of cardiovascular disease and for treatment of atherosclerosis (1 and 47). FF is one of the most commonly prescribed fibrates, and has a well known efficacy and tolerability profile 1 .
FF seems to lower lipid levels by activating peroxisome proliferatoractivated receptor alpha (PPARα), a nuclear receptor which acts as a ligand activated transcriptional factor and activates lipoprotein lipase and reducing apolipoprotein CIII expression. These activities increase lipolysis and eliminate triglyceride-rich particles 2 .
PPARs are a widely distributed family of nuclear receptors. Three isoforms have been identified: PPARα, PPARβ/δ, and PPARγ. Ligand binding activates these receptors that play key roles in cellular energy homeostasis, modulating glucose and lipid metabolism. PPARα as illustrated in Figure 1, is the molecular target of the fibrate class of drugs, which act as agonistic ligands of PPARα. Other fibrates like clofibrate and bezafibrate are also ligands for this receptor. Poly-unsaturated fatty acids are the natural ligands.
We shall not go any deeper into lipid metabolism activities of FF because our goal is to consider the effects of this pharmaceutical in cancer prevention and treatment rather than in cardiovascular risk.
Before the year 2000, all publications on FF considered anti-lipogenic properties of this drug with no mention of possible anticancer activity.
We only found two publications of FF activities before than that may be related with cancer: 1) Marx et al. (1999) describe that FF reduces the expression of vascular cell adhesion molecule 1 (VCAM-1) in human endothelial cells 3 .
In 2002 two findings on FF anti-cancer therapy were important: 1) It was demonstrated that PPARα and PPARδ downregulate NF-kB induced translocation to the nucleus in endothelial cells 5 and 2) PPARα activators inhibited endothelial cell migration by targeting Akt phosphorylation in endothelial cells 6 . Upon ligand binding, PPARα dimerizes with RXR (retinoic X receptor) and both interact with peroxisome proliferators responsive elements of the target gene. Coactivator proteins and RNA polymerase are recruited and the transcription machinery is set to work (trans-activation). When co-repressor molecules are recruited, trans-repression is unleashed and no transcription is produced. FF has the capacity to induce hepatocarcinoma in rodents, but this effect seems specific for this species, as in humans it has been shown to have cytotoxic effect on HepG2 hepatoma cell line at high concentrations and in a dose dependent manner 7 .
Varet et al. (2003) 8 demonstrated that FF inhibits angiogenesis in vitro and in vivo.
MCP-1 (monocyte chemotactic protein-1) is a protein that recruits and activates monocytes during inflammatory processes but also plays a role in cancer: it increases proliferation and invasion of CaP cells (prostate cancer) 9 . FF inhibits expression of MCP-1 on activated endothelial cells 10 .
PPARα agonists like FF were found to inhibit endothelial VEGFR2 11 expression.  In diabetic II hyperlipidemic patients, FF decreased E-selectin by 10% and ICAM-1 by 4% and no change of VCAM-1 was detected 12 .
But the first real hint towards a possible anticancer activity of FF was provided by Holland et al. in 2004 13 who showed by transcriptome analysis of endometrial cancer cells an overexpression of PPARα. This finding led them to the investigation of FF activity on tumor cells. FF reduced proliferation and increased apoptosis of cancer cells. At the same time, a second publication 14 showed the FF potential to reduce metastasis of melanoma cells in an experimental setting.
In ooforectomized rats, treated with estradiol and FF for 30 days, the uterine mass decreased, uterine glands had normal structure and there were no cases of atypical hyperplasia 15 .
Kubota et al. 16 found that apoptosis induced by FF in cultured human hepatocytes was due to caspase-dependent apoptosis by inhibiting phosphorylation of Akt, in a PPARα independent manner.
The role of chemokines produced by different stromal cells stimulating proliferation and angiogenesis in cancer tissues is well known. FF exerts a monocyte suppressing activity and reduces secretion of IL-6 and MCP-1 17 .
Studying the possible anti-rheumatic activity of FF it was observed that this compound inhibits NF-kB 18 .
After these preliminary hints indicating the FF possible anti-cancer activities of FF, great amount of research and publications were dedicated to this issue. We summarized these findings in table 1 to  table 15 according to anti-cancer activity disclosed.

FF anticancer mechanisms
On artificial grounds, but for better understanding, we have presented the anti-cancer activity of FF according to the main protumor factor/pathway affected by the drug. Angiogenesis inhibition as described in Table 1  FF was included in many multi-agent anti-angiogenic regimens. One consisted of FF, celecoxib, thalidomide with metronomic low dose cyclophosphamide and etoposide. Patients were less than 21 year old with recurrent or progressive tumors. Half of the patients obtained benefits (CR + PR + SD) 30 .  Treatment with FF significantly reduced proliferation and increased cell death in endometrial cancer cells, suggesting that altered expression of nuclear hormone receptors involved with fatty acid metabolism leads to deregulated cellular proliferation and apoptosis. PPARα was increased more than 4-folds in endometrial cancer cells. Other metronomic anti-angiogenic multidrug protocols included FF as one of the pharmaceuticals, particularly for children with embryonal brain tumors and other malignancies 31-33 .
The COMBAT Protocol 34 included low-dose daily temozolomide, etoposide, celecoxib, vitamin D, FF and retinoic acid and was used in 74 children with advanced refractory/relapsed solid tumors with two years overall survival of 43%.
The use of FF as part of anti-angiogenic multidrug protocols especially in pediatric cancer is constantly increasing.
Using a PPARα agonist like Wy-14643 in mice injected with tumor cells showed that treated animals had a marked reduction in tumor size and vascularization 35 .
Apoptosis induced by FF is caspase-dependent. In the case of clofibrate, apoptosis occurs through caspase 2 and 3 activation and ER stress in Jurkat cells 44 . Similar results were observed in Yoshida AH130 hepatoma cells 45 .
PPARα is increased in high grade renal cell carcinoma (RCC), but this does not provide any information about the functional status of this receptor, because in RCC the inhibition of PPARα induces apoptosis and agonists produce little or no effect 46 .
In 1983 Pascal et al. 47 investigated the cardiovascular and antiarteriosclerotic activities of FF and demonstrated that FF inhibited platelet derived growth factor (PDGF) stimulating activity on growth of cultured smooth muscle. Ten years later Munro et al. 48 showed that FF is not a specific inhibitor of PDGF because smooth muscle cells growth was equally growth-inhibited by FF when the culture was stimulated with fetal calf serum, PDGF or basic fibroblast growth factor (bFGF). Our conclusion based on these two publications is that FF is a growth inhibitor in general (as least regarding vascular smooth muscle).
Anti-proliferation activity of FF has been found in many non-tumor tissues besides vascular smooth muscle, e.g. mesangial cells 49 through inhibition of PI3K/AKT and ERK1/2 signaling pathways or by overexpression of TRIB3 (tribbles homolog 3) which inhibits Akt phosphorylation and slows cell cycle or causes arrest in G1/S 50 . In lymphocytes, FF also up-regulates TRIB3 causing cell cycle arrest 51,52 .
Endothelin-1 is a protein that increases cardiac fibroblast proliferation. PPARα agonists inhibit cardiac fibroblast proliferation downregulating endothelin-1 53 . FF also reduced c-jun expression in cardiac fibroblasts 54 . Endothelin-1 is an activator of the p38 mitogen activated kinase cascade. FF down-regulation of endothelin-1 also down-regulates the MAPK cascade in cardiomyocites 55 .
FF reduced the IFNγ and IL-1β-induced cell proliferation of astrocytes in culture 56 . Table 3 depicts the anti-proliferation activities of FF in cancer.
The work by Saidi et al. 59 needs further discussion. The authors noticed that in Ishikawa endometrial cancer cells FF enhanced growth inhibition when ATRA was simultaneously used. ATRA by itself had no effect on growth. This is a logical finding because PPARα forms a heterodimer with RXR before binding DNA at the peroxisome proliferators responsive element. So this synergy between FF and ATRA regarding growth inhibition seemed a PPARα-dependent activity. Apoptosis also was increased with the combination of these drugs.
Paradoxically, RNAi inhibition of PPARα showed only a minor reduction in FF effect and ATRA combined with FF showed minor differences in growth inhibition with or without PPARα RNAi.
After 48 hours of treatment the difference was approximately 40% less viability in cells treated with FF plus ATRA and no RNAi against those with RNAi. We hypothesize that RNAi inhibition of PPARα needs at least 48 hours to make the viability difference. So that growth inhibition seems, at least partially, as PPARα dependent.
Unfortunately FF and FF plus ATRA showed no differences in tumor size and growing in vivo compared with control group receiving no drugs. The combination of FF and retinoic acid is a potent inhibitor of Ishikawa endometrial cancer cell growth in vitro. Cell cycle arrest is produced at G1/S. Cyclin D1 was down-regulated. These results could not be reproduced in vivo.

Yokoyama Y, 2007 60
Clofibric acid (a PPARα agonist) significantly suppressed the growth of OVCAR-3 tumors xenotransplanted s.c. and significantly prolongs the survival of mice with malignant ascites derived from DISS cells as compared with control. Microvessel density was diminished and apoptosis was also found. VEGF and PGE2 were also diminished.
Yamasaki D, 2011 61 FF suppresses growth of the human hepatocellular carcinoma cell (Huh7) via PPARα-independent mechanisms and produces G1 arrest caused by the reduction of cyclin A and E2F1 and accumulation of the cyclin-dependent kinase inhibitor p27. This activity was not modified by PPARα antagonists. Inhibition of Akt phosphorylation by increased CTMP was also observed.
Chang NW, 2011 62 FF highly suppresses the formation of squamous cell carcinoma in an oral-specific 4-nitroquinoline 1-oxide/arecoline mouse model, decreases the tumor size, and increases the immunoreactivity of EGFR and COX2 in oral dysplasia, but decreases EGFR and COX2 expressions in SCC. These molecular events might be linked to the EGFR and COX2 regulatory pathways.
Huang J, 2013 63 FF is capable to suppress B-cell lymphoma growth. This growth suppression is independent of angiogenesis inhibition.
Binello E, 2014 42 FF shows anti-proliferative pro-apoptotic effects on high-grade gliomas and anti-invasive effects on glioma stem cells.
Li T, 2014 41 FF has anti-proliferation effects on breast cancer cell lines.
Wang H, 2014 64 FF can inhibit the growth and migration of human ovarian cancer cell SKOV3 in vitro, and to some extent induce apoptosis.

Liang H, 2014 65
Synergistic inhibitory effects on cancer cell proliferation by simultaneous application of FF and budesonide. FF inhibited cell proliferation in both TP53 wild type and deficient lung cancer cells.
The anti-proliferation effect of budesonide in TP53 wild type A549 cells was abolished in SK-MES-1 cells that do not have wild type TP53 protein.
The work by Chang et al. 62 suggests that FF may be useful for prevention of oral SCC because in an experimental setting FF was capable of reducing the incidence of tumors and also the progression from pre-neoplastic stage to SCC. FF at low doses lacked antitumor activity.
In spite of the known fact that glucocorticoids induce chemotherapy resistance in most of the solid tumors 66 , Liang et al. 65 found that FF and budesonide had synergistic anti-proliferative effect on lung cancer cells with intact TP53.
Inflammation plays a very important role in carcinogenesis and tumor progression. NF-kB pathway is an essential actor of the proinflammatory and anti-apoptotic activity 67-69 .
FF has the capacity to down-regulate NF-kB activity according to evidences gathered in Table 4. Through this PPARα-dependent mechanism, FF exerts anti-inflammatory activity. Besides, it also has non PPARα-dependent anti-inflammatory activity through upregulation of SHP (small heterodimer partner). Table 4 strongly support the FF anti-inflammatory activity mediated through NF-kB down-regulation and also PPARα independent mechanisms.

Evidences reported in
One of the proposed mechanisms of FF inhibiting NF-kB activity is depicted in Figure 5.
The research studies reported in Table 5 are evidence of downregulation of Akt phosphorylation by FF, but Piwowarczyk et al. 83 working with prostate cancer cells (DU-145) and endothelial cells (HUVEC) co-cultures found that FF increased levels of phosphorylated Akt in both HUVEC and DU-145 cells. They found that Akt phosphorylation was essential for FF increase of endothelial barrier ( Figure 6).
Mitochondrial uncoupling proteins (UCP) are mitochondrial anion carrier proteins that separate oxidative phosphorylation from ATP synthesis with energy lost as heat and reduction of mitochondrial membrane potential 90  where UCP2 expression was down-regulated grew faster than cells expressing UCP2 92 . They also found that loss of UCP2 produced a metabolic change toward glucose metabolism, decreased fatty acid oxidation and increased proliferation.
Evidence supports that FF decreases intracellular energy through inhibition of mitochondrial enzymes in a similar way as metformin.
On a theoretical basis, we may assume that there might be synergism with metformin on this ground. FF improved systemic inflammatory responses through the nuclear orphan receptor SHP (small heterodimer partner) and UCP2 (uncoupling protein 2). This is PPARα independent anti-inflammatory mechanism.
Schen W, 2014 78 FF down-regulated NF-kB in endotoxin induced uveitis. All inflammatory factors like cytokine production, vessel density, vascular leukostasis and inflammatory cell infiltration was also down-regulated.
Binello E, Germano IM, 2014 79 FF treatment decreased glioblastoma stem cell invasion in vitro. Treatment decreased the expression of NF-kB and cyclin D1 in a dose dependent and p53 independent manner.   PPARα involvement with cancer metabolism has been extensively reviewed by Grabacka and Reiss 93 .
Another enzyme down-regulated by FF is FAS (fatty acid synthase) 94 which is highly expressed in many cancer tissues. Fatty acid synthase (FAS) is a multicomplex enzyme that intervenes in endogenous synthesis of fatty acids and particularly palmitate. Abnormal fatty acid (FA) synthesis is one of the common features of many cancer cells and FAS has been identified as part of cancer controlling networks. Human cancers that over-express FAS, are usually associated with poor prognosis 95-99 .
The expression of adhesion molecules on the endothelial cell surface is critical for cells rolling in the vascular lumen to achieve tethering and adhesion to the vascular wall and eventually achieving diapedesis and colonization in the case of potentially metastatic cells or leukocyte recruitment to atherosclerotic lesions.
PPARα regulates gene expression of certain adhesion molecules in response to unsaturated fatty acids and fibric acid derivatives like FF. This control is achieved probably through inhibition of TNFα induced NF-kB activation 100 .  In glioblastoma FF is accumulated in the mitochondrial fraction, followed by an immediate impairment of mitochondrial respiration at the level of complex I of the electron transport chain. This mitochondrial action sensitizes tested glioblastoma cells to the PPARα-dependent metabolic switch from glycolysis to fatty acid β-oxidation.

The research by Marchesi et al. 100 that demonstrated reduction in adhesion molecules with FF treatment is important for two reasons:
1) It was performed in humans (10 hypertriglyceridemic patients).
2) The amount of reduction in fasting conditions (near 45% reduction for ICAM and around 33% reduction for VCAM levels).
Empen et al. (2003) described 10% reduction of E-selectin after six weeks treatment with FF, but found no major changes with VCAM-1 and ICAM-1 levels 12 .
Piwowarczyc et al. 83 demonstrated a new FF effect: increased endothelial cell adhesion to the susbstratum and increased adhesion between endothelial cells by activation of focal adhesion kinase (FAK). These impedes cell diapedesis through the vessel wall, which is an important objective to decrease metastatic risk. Figure 6 depicts this activity.
The production of the metastatic cascade is a complex process in which there are many successive steps that we shall not analyze in depth in this review.
But for a better understanding lets remember the main steps 107 : 1) Primary tumor growth and angiogenesis.
2) Future metastatic cells free themselves from the primary tumor.
10) Start growth in the colonized site including angiogenesis.
Cell motility, invasion, angiogenesis and the function of connexins and adhesion molecules are an essential part of this cascade and have already been considered. In table 8 we describe only specific research work relative to metastasis and FF.
COX-2 is the rate-limiting enzyme in prostaglandin synthesis that catalyzes the production of prostaglandins and thromboxanes from arachidonic acid, and has been associated with growth regulation and carcinogenesis in many tumors. The COX2/PGE2 pathway may be considered a pro-tumor pathway at least in certain cancers where elevated levels of COX2 have been identified. Most colorectal carcinomas and many adenomas exhibit this elevation 109-110 . One of the postulated mechanisms by which COX2/PGE2 signaling stimulates cell growth is through the activation of β-catenin 111 . COX2 is implicated in breast cancer progression and invasiveness 112,114 . In stage III breast cancer, COX2 over-expression is an unfavorable prognostic sign, and according to Kim et al. gives ground for using COX2 inhibitor combinations 113 . Simeone et al. identified the pathway leading to increased invasiveness in breast cancer: COX2 /protein kinase C/interleukin-8/urokinase-type plasminogen activator pathway 115 . COX2 is also associated with angiogenesis and metastasis 116 .
Many cancers harbor increased COX2 activity including lung, colorectal, breast and squamous cell carcinoma of the upper digestive system [117][118][119] . COX2 down-regulation is an important issue in many cancers. We have been describing the action of FF on different pro-tumor proteins in an artificially separate manner, but many of these proteins share their activity in the pro-tumor evolution or are part of the same pathway. This is the case of NF-kB and COX2 in the progression towards cancer of Barret's esophagus in which Table 7. Anti-adhesion and anti-chemokine activities.

Reference FF activities
Marx N, 1999 3 FF inhibits VCAM-1 transcription probably by inhibiting TNFα induced NF-kB activation.   Irradiating a murine microglial cell line with γ-rays led to increased expression of IL-1β and TNFα, Cox2. FF significantly decreased over-expression probably through NF-kB down-regulation.    Treatments with FF and troglitazone (a PPARγ ligand) strongly decreased the Sema6B mRNA. The drop in Sema6B mRNA level and in protein content was more important when the treatment combined the action of FF or troglitazone and 9-cis-retinoic acid. Combined arsenic trioxide and FF exert a significant effect on epithelial-mesenchimal transformation of A549 cells, which may be related with the expression of E-cadherin and Snail.  increased NF-kB activity is linked to increased IL-8 and COX2 expression 120 . As described above, FF is active against both: NF-kB and COX2.
The anticancer activities of FF are pleiotropic. Besides proliferation and angiogenesis down-regulation representing the main antitumoral effects, there are many others that will be described in the next three tables like ovarian aromatase inhibition, AMPK activation, IGF-I down-regulation, etc.
The IGF-1 receptor signaling system is a contributing factor in invasion, migration and proliferation of glioblastoma and became a legitimate target in the treatment of this pathology 128 . FF has experimentally shown to inhibit this system and decrease growth and invasion 82, 126,127 . A sort of IGF-1 trap was designed by D'Ambrosio et al. 129 that inhibited tumor growth in vivo and induced apoptosis.
These publications reported contrasting results: two of the showed that FF increased radiation sensitivity 133,134 and one showed decreased radiation sensitivity 132 . As we shall see latter, PPARα agonists are tissue-specific and species-specific. This may explain the difference. In the first case the experiments were performed on HeLa cells and in the second and third studies, experiments were performed on squamous cell carcinoma cells. In all three cases human cells were used, so the difference may lay in tissue-specific behavior or may be due to the fact that in the second experiment the environment was particularly hypoxic.
Semaphorins are a large family of axon guidance molecules. They interact with their receptors, plexins and neuropilins, and play important roles in a growing list of diverse biological systems, including cancer progression and tumor angiogenesis. Some semaphorins can activate tumor progression and angiogenesis, while others may have the opposite effect.
There is abundant literature on semaphorins 3,4 and 7. Little is known about semaphorin-6B. It is known that there is an association between gastric cancer 137 , gliobastoma 135 and certain breast cancers (MCF-7 breast adenocarcinoma cell line 136 ) and semaphoring-6B but the exact nature of this association is still poorly understood.
According to Ge 137 , inhibition of semaphoring-6B expression via RNA interference inhibited the migration, adhesion and invasion abilities of SGC-7901 gastric cancer cells in vitro.
SEMA6B transcript was down-regulated in two human glioblastoma cell lines (T98G and A172) when a prolonged treatment with ATRA was performed 138 .
The SEMA6B gen has a PPARα binding site in the promoter region so that the interrelation of PPARα and the gene effectively is present, and this interrelation is a negative one and probably exerts anti-tumoral effects.
In a large population study in Finland to assess the overall risk of prostate cancer in people taking cholesterol lowering medication, no decrease in risk was seen either with statins nor fibrates 142 . The statistics is significantly biased regarding fibrates: the population is too small (around 220 patients with prostate cancer receiving fibrates out of a total number of prostate cancer cases of 24.723).
HIV treatment with protease inhibitors has a frequent side effect: lipodystrophy syndrome. FF has been very effective in the treatment of this syndrome 143 . Nelfinavir and ritonavir are protease inhibitors which have shown many anti-cancer activities. We have postulated nelfinavir as a complementary off target treatment for cancer 144 .
The association of FF with nelfinavir may prevent lipodystrophy, hypertriglyceridemia and elevation of lipoparticules [145][146][147] . There is no experimental proof of synergy of this interaction regarding cancer treatment, but both drugs share certain characteristics that make synergy a very plausible feature.
Nelfinavir is a proteasome inhibitor, and proteasome is the site where PPARα is dissembled, so nelfinavir should prolong PPARα's life. Moreover nelfinavir and FF share the following activities: Akt inhibition, anti-angiogenesis, decreased proliferation, increased apoptosis, reduction of MMP2 and MMP9, and increased p21 144 . Table 17 shows similarities and differences between nelfinavir and FF.
The main purpose of this table is to illustrate the similarities and differences between this two drugs regarding cancer: nelfinavir down-regulates CDK-2 and FF down-regulates cyclin D-1 reinforcing cell cycle slowdown. Both are strong anti-angiogenic agents. FF also down-regulates COX2 which is an important step in decreasing angiogenesis.

Discussion
PPARα is highly expressed in certain cancer cells like endometrium 27 , prostate 148 , bladder 149 , certain breast cancer cell lines 150 , NSCLC 152 and others.
In these cases, the administration of a PPARα agonist like FF increased apoptosis and decreased proliferation.
In human breast cancer cell lines MCF-7 and MDA-MB-231, PPARα was overexpressed but use of agonists of this receptor increased proliferation 150 .
In stark contrast, Li et al. 152 tested FF in 12 breast cancer cell lines (including MDA-MB-231) and in all of them FF was effective as proliferation inhibitor. This effectiveness was independent of PPARα expression but was linked to triple negative condition. Paradoxically MDA-MB-231 was the more sensitive to inhibition of proliferation by FF treatment. In human colon cancer tissues PPARα is underexpressed when compared with normal tissues 151 . Using PPARα ligands in APC Min /+ mice to evaluate polyp formation, those treated showed decreased number of polyps and decreased size. We may conclude that overexpression or under-expression of PPARα in cancer tissues is not an indicator of future response to FF or other PPARα agonists.
Another issue to consider is the species-specific response to PPARα agonists: for instance, FF induces hepatocarcinogenesis in rodents but not in humans, insulin resistance in mouse but not in humans, oxidative stress in mouse heart but not in human heart. Human liver has a lower expression of PPARα than rodents 153 (The differences of PPARα in human liver has been extensively described in the review by Roberts 154 ).
There are important differences between rat and human hepatic cells. Thus this species-and tissue-specificity suggests that we should be cautious when interpreting the results of many of the published investigations. PPARα agonists research results obtained in rodents should not be taken for granted in humans.
The mechanism of FF anticancer activity may differ in different tumors: in mantle cell lymphoma it seems to induce apoptosis by inhibiting the TNFα/NF-kB axis 39 while in triple negative breast cancer it requires activation of NF-kB in order to induce apoptosis 41 .
PPARα agonists like FF are actively investigated as anti-cancer drugs, but paradoxically, PPARα inhibitors may also work against cancer. This is the case of renal cell carcinoma 46 where an inhibitor of this receptor produced cell cycle arrest and apoptosis.
After these necessary clarifying precautions, we must consider in detail the hard evidence collected in the medical literature that gives ground for FF as a complementary adjunct pharmaceutical in cancer therapy.
Phanigraphy et al. 21  FF also uses the AMPK pathway to produce its anti-tumoral effects 160 . This is a PPARα independent action that has been demonstrated in human oral squamous cell carcinoma (OSCC) where FF inhibited cell migration and invasion and reduced expression of MMP 1, 2, 7 and 9. LKB1 and AMPK were up-regulated after FF treatment. When AMPK was inhibited (with protein C) the antiinvasive effect was significantly reduced.
Metformin is another activator of the AMPK pathway. It has not been tested with FF but it is quite possible that there might be synergy in this activity. Another coincidence between metformin and FF is the decrease in cellular energy production.
The molecular mechanism of many of the FF anti-cancer activities is known since 2006 161 and can be summarized as inhibition of COX2 and VEGF at the transcription level by interfering with AP-1 binding to DNA and decreased expression of NF-kB. In susceptible cells there is a negative cross talk between PPARα and AP-1 161 .
Another mechanism postulated in anti-cancer activity is the disruption of the tumor-host stroma symbiosis due to anti-angiogenesis and anti-inflammatory activity 162 .
Finally we have to mention that FF effectively reduces nuclear SREBP-2 but not SREBP-1 163-166 . Nelfinavir and PUFA inhibit mature SREB-1 which is a transcription factor that promotes FAS synthesis, so these products may complement FF anti-cancer activity.

Conclusion
As observed in the tables, FF exerts polyvalent anti-cancer activities that deserve further research in the clinical setting.
FF is a PPARα agonist drug developed for treatment of elevated triglycerides and LDL cholesterol reducing cardiovascular risk that may be repurposed to be used in cancer due to its anti-angiogenic, anti-inflammatory, anti-proliferative, anti-metastatic, anti-adhesive, anti-invasive and pro-apoptotic activities in certain cancers.

Future perspectives
FF has already been incorporated in anti-angiogenic protocols for the treatment of glioblastoma. Probably in the future it will form part of protocols based on repurposed drugs directed to inhibit angiogenenesis like celecoxib, nelfinavir, and metformin.
Colorectal and prostate cancer seem good candidates for these therapies.

Competing interests
No competing interests were disclosed.

Grant information
The author(s) declared that no grants were involved in supporting this work. This in an excellent comprehensive review of the literature related to Fenofibrate anticancer effects, which definitely deserves to be indexed after revision.
I have three comments which should be implemented: It should be mentioned that fenofibrate is a prodrug, which is processed by blood esterases to fenofibric acid, and that fenofibric acid is potent agonist of PPARa. Therefore, there are two very different substances, fenofibrate and fenofibric acid, which could have very different anticancer properties. As far as I understand the system, mitochondrial and growth factor-mediated effects are rather associated with unprocessed fenofibrate, which in opposite to fenofibric acid is capably of interacting with biological membranes.
Reference 89 in table 6 should be updated (see below); and the issue of Fenofibrate and mitochondrial respiration deserves a separate paragraph since it explains how unprocessed fenofibrate can target tumor cells, being at the same time much less harmful to normal cells. Other described mechanisms could be involved as a consequence of this primary mitochondrial action, which actually happens in minutes. Thank you for your kind contribution to this review. A new version is under way, which will include all the issues you remarked: A figure will be introduced which will show the molecular structure of FF and fibric acid. The difference between the pharmacologic actions of both will be discussed.
Membrane rigidity due to FF will be explained under a new heading and the importance of this effect will be further discussed. The difference in cell distribution of FA and FF will be also considered.
Reference in Table 6 will be updated and reference 158 will be corrected.

Remains sincerely yours Tomas Koltai
There are no conflicts of interests Competing Interests: