Relevance of Peroxisome Proliferator Activated Receptors in Multitarget Paradigm Associated with the Endocannabinoid System

Cannabinoids have shown to exert their therapeutic actions through a variety of targets. These include not only the canonical cannabinoid receptors CB1R and CB2R but also related orphan G protein-coupled receptors (GPCRs), ligand-gated ion channels, transient receptor potential (TRP) channels, metabolic enzymes, and nuclear receptors. In this review, we aim to summarize reported compounds exhibiting their therapeutic effects upon the modulation of CB1R and/or CB2R and the nuclear peroxisome proliferator-activated receptors (PPARs). Concomitant actions at CBRs and PPARα or PPARγ subtypes have shown to mediate antiobesity, analgesic, antitumoral, or neuroprotective properties of a variety of phytogenic, endogenous, and synthetic cannabinoids. The relevance of this multitargeting mechanism of action has been analyzed in the context of diverse pathologies. Synergistic effects triggered by combinatorial treatment with ligands that modulate the aforementioned targets have also been considered. This literature overview provides structural and pharmacological insights for the further development of dual cannabinoids for specific disorders.


Introduction
The concept of designed multiple-targeting ligands seems to appear with modern drug discovery approaches and especially in 2004 when Morphy et al. [1] published "From magic bullets to designed multiple ligands". Over the past years, this emerging polypharmacological-based therapeutic approach has been explored for multifactorial diseases, for example Alzheimer's disease, multiple sclerosis, diabetes, coronary heart disease, cancer, and rheumatoid arthritis [2][3][4]. Synergistic effects while reducing side effects are the pursued key goals compared to combination therapies. Nowadays, better understanding of protein/ligand complexes allows rationalizing the design of new multitarget ligands [3,5]. In this context, the endocannabinoid system (ECS) by itself represents a goldmine in multitarget therapeutic strategies.
Three decades ago, the first elements of the endocannabinoid signaling system were discovered thanks to ∆ 9 -tetrahydrocannabinol (THC: Figure 1), one of the two major components of the plant Cannabis sativa, the other one being cannabidiol (CBD; Figure 1) [6]. The identification of the principal biological target of THC, the CB 1 cannabinoid receptor (CB 1 R) [7], led to the discovery of two endocannabinoids, N-arachidonoyl-ethanolamine (AEA; anandamide, Figure 1) and 2-arachinoylglycerol (2-AG, Figure 1) a few years later [8][9][10]. Then, a second cannabinoid receptor, CB 2 R, was identified [11]. Thus, in the early part of the last decade, ECS was considered to comprise two cannabinoid receptors, CB 1 R and CB 2 R, the endocannabinoids AEA and 2-AG, and enzymes responsible for their degradation, the fatty acid amide hydrolase (FAAH) and the monoacylglycerol lipase (MAGL) (see [12] for a review). The ECS has been shown to regulate a variety of physiological and pathological activities, such as appetite, pain, memory, and inflammation [12]. transcription [37]. The activities of PPARs are mainly associated with fatty acid oxidation and metabolism. PPARα and PPARβ/δ are highly expressed in liver (PPARα), brown adipose tissue (PPARα), intestine (PPARβ/δ) heart, skeletal muscle, kidney, and skin. The two major isoforms of PPARγ, PPARγ1 and PPARγ2, are differently expressed being PPARγ1 expressed in numerous cells included immune and brain cells, whereas PPARγ2 is especially present in white and brown adipose tissues. Several fatty acid derivatives bind and/or activate PPARs. Fibrates and thiazolidinediones (TZDs) are two conventional classes of PPAR ligands (for recent developments of synthetic PPAR ligands, see [38]). Troglitazone, rosiglitazone, and pioglitazone are PPARγ agonists used for treatment of type 2 diabetes mellitus (T2DM). However, the first two have been withdrawn from the market by most countries due to hepatotoxicity and cardiovascular side effects respectively, pioglitazone not being exempt of serious side effects. Several PPARα agonists, including ciprofibrate, fenofibrate, and bezafibrate, are approved drugs in some countries with a clear risk of hepatotoxicity. So far, no synthetic cannabinoids apart from synthetic THC are in the market due to adverse outcomes, mainly psychotropic effects (CB 1 R agonists and antagonists), lack of efficacy (CB 2 R agonists), or possible immunosuppression (CB 2 R agonists). Thus, emerging strategies are being explored in academic and pharmaceutical laboratories [14]. One of these strategies focuses on ECS-PPARs for developing multiple target therapeutic agents. Comprehensive reviews from O'Sullivan et al. [31,39,40] describe the activation of PPARs by some cannabinoids including endocannabinoids, phytocannabinoids, and synthetic cannabinoids, and cannabinoids' therapeutic effects are mediated by PPARs. The mechanisms of action involved in the cannabinoids/PPARs interactions are not yet being resolved, even though there has been increasing evidence supporting the ability of PPARs activation to mediate some of the therapeutic effects of cannabinoids over the past 15 years.
CBRs and PPARs have shown to exhibit biological relevance in common pathophysiological contexts. For instance, both CB 1 R and PPARα have shown a therapeutic role in the regulation of lipid metabolism. Moreover, because of their immunomodulatory profile, CB 2 R and PPARγ have been widely studied for the management of diverse inflammatory diseases. These shared properties evidence the potential of multitargeting strategies in the context of specific diseases.
In this review, we will explore the connection between PPARs and elements of the ECS through the action of different cannabinoids reported so far to be engaged in this relationship.

CB 1 R-PPAR Modulation
Certain therapeutic outcomes triggered by cannabinoids have shown to be mediated through the modulation of CB 1 R along with the PPAR subtypes PPARα or PPARγ. Dual targeting CB 1 R/PPAR as well as co-administration have been reported as beneficial pharmacological strategies in the course of numerous pathologies including obesity, arthritis, cancer, epilepsy, and alcohol use disorder [31,41,42].

CB 1 R-PPARα
Concomitant actions at CB 1 R and PPARα receptors have shown to be involved in the therapeutic effects exhibited by cannabinoids in metabolic syndrome. CB 1 R blockage is widely known for the reduction of body weight gain in obese subjects [43][44][45]. Moreover, PPARα activation has shown to play a therapeutic role in the regulation of lipid metabolism and obesity [46,47]. Taking advantage of these modulatory profiles, diverse studies have pursued CB 1 R/PPARα dual targeting to tackle obesity. For instance, fatty acid amide derivatives conjugated with amphetamines have been claimed as CB 1 R/PPARα dual ligands with antiobesity properties [48,49]. The most potent compound of this series, the oleic acid-dihydroxyamphetamine (OLHHA) (Figure 1), a weak CB 1 R antagonist and moderate PPARα agonist, proved to induce satiety and control food intake, reduce body fat, and regulate fat metabolism in rats [48,49]. As demonstrated by these authors, molecular mechanisms mediating OLHHA antiobesity effects involve CB 1 R and PPARα, while FAAH and TRPV1 receptors do not contribute to food intake modulation.
Further investigations using OLHHA revealed its ability to improve non-alcoholic fatty liver disease (NAFLD) in an genetic animal model of obesity [50]. In this study, the immunohistochemical and histological analysis of liver and plasma samples of lean and obese Zucker rats upon OLHHA chronic treatment confirmed its anti-steatotic and hepatoprotective profile. A significant decrease in hepatic lipid accumulation, reduction of plasma levels of triglycerides, and cholesterol along with anti-apoptotic activity was also observed in obese rats. These hepatoprotective properties were related to an increase of CB 1 R expression and a downregulation of lipogenesis-related enzymes, while changes in PPARα mRNA expression between treated and vehicle rats were not significant. Even though these studies evidence the therapeutic potential of OLHHA, additional research is needed to fully determine its mechanism of action in obesity-associated fatty liver.
It is worth noting that the CB 1 R/PPARα dual profile of OLHHA was also proved to be effective at reducing alcohol intake [51]. In animal models of alcohol consumption, treatment with OLHHA was able to significantly reduce alcohol self-administration and voluntary alcohol consumption without triggering tolerance or toxicity. Therefore, this compound is a promising lead not only for the management of eating disorders and associated pathologies but also to treat alcohol use.
In an effort to obtain dual ligands targeting CB 1 R and PPARα, the diarylpyrazole core of the CB 1 R antagonist/inverse agonist SR141716A (Rimonabant, Figure 1) was fused to the phenoxypropanoate pharmacophore of the fibrates (fenofibrate, PPARα agonist, Figure 1), obtaining the so-called rimonabant fibrates [52]. The most potent compound of this series, 2 (Figure 1), exhibited nanomolar activity as a PPARα agonist in luciferase reporter gene assays and a CB 1 R antagonist in mouse vas deferens contractile response assays. Even though its therapeutic potential has not been proved yet, this dual ligand may impair metabolism through different molecular mechanisms.
Another therapeutic strategy involving both targets was developed upon the peripheral blockade of CB 1 R in a diet-induced obese (DIO) mouse model [53]. The peripherally restricted antagonist AM6545 was able to reduce hepatic steatosis and improved liver injury through PPARα as shown by its lack of activity in PPARα knockout DIO mice. However, AM6545 failed to directly bind or modulate PPARα. Thus, the antisteatotic effects of AM6545 are not the result of direct dual targeting but rather due to the ability of CB 1 R to regulate hepatic PPARα. The authors suggested that antagonizing CB 1 R by AM6545 may increase the levels of hepatic endocannabinoid-like compounds, oleoylethanolamide (OEA, Figure 1) and palmitoylethanolamide (PEA, Figure 1), which may then directly activate PPARα [53].
Evidence from animal models shows that the combinatorial treatment of OEA and SR141716A is a successful approach for the control of obesity [54]. OEA is a shorter monosaturated analogue of AEA that has shown to activate PPARα while being devoid of CBR activity. This endogenous PPARα agonist had already been reported to reduce body weight and regulate satiety via this nuclear receptor [55,56]. The combinational therapy of the aforementioned CB 1 R antagonist with OEA in obese Zucker rats resulted in an improved reduction on feeding, body weight gain, and cholesterol levels along with an inhibition of enzymes involved in lipid biosynthesis, evidencing the synergistic effects of both drugs [54]. These results support combining both mechanisms of action to provide a more efficient treatment for the management of obesity.
In addition to obesity, other pathologies have shown to be impacted by mechanisms that involve CB 1 R and PPARα modulation. The activation of these targets has been respectively studied for their neuroprotective effects [41,[57][58][59][60]. An example of dual CB 1 R and PPARα activation in this context is the antiepileptic profile of the endocannabinoid PEA [61]. PEA, a non-saturated analogue of OEA, exhibited anti-absence effects in a rat model of absence epilepsy (WAG/Rij rats), which were reversed by SR141716A and by the PPARα antagonist, GW6471. This endogenous compound is a PPARα agonist but lacks affinity toward the cannabinoid receptors CB 1 R and CB 2 R [31,62]. Therefore, these authors postulate that its CB 1 R-mediated antiepileptic effects are due to an enhancement of AEA activity by an entourage effect [61], which was also observed in previous PEA studies in other biological systems [63,64].
Likewise, the antinociceptive properties of certain cannabinoids have shown to be mediated through CB 1 R and PPARα. For instance, the endocannabinoid PEA exhibits analgesic effects via PPARα direct and CB 1 R indirect activation in an osteoarthritic chronic pain rat model [65]. Its behavioral effects were antagonized by SR141716A and GW6471; however, the implication of other targets, including the TRPV1 channel and the orphan GPCR GPR55, cannot be ruled out. In the same study, behavioral tests demonstrated that the antinociceptive properties of the synthetic cannabinoid agonists HU210 and WIN55,212-2 ( Figure 1) are not due to a dual CB 1 R/PPARα mechanism but mainly mediated by the cannabinoid receptor.
Moreover, the co-activation of both targets using the endocannabinoid AEA and the PPARα agonist GW7647 also demonstrated effective pain reduction [66]. Their synergistic effects significantly decreased pain behavior in a mouse model of acute chemical-induced pain.

CB 1 R-PPARγ
Pharmacological cannabinoid effects can also be mediated through CB 1 R/PPARγ dual mechanisms. Both receptors have shown to be involved in pathological processes including pain, tumor growth, or obesity [67]. PPARγ and CB 1 R have been associated to diverse types of cancer. Extensive research has proved the therapeutic utility of CB 1 R activation in the progress of a wide variety of tumors [68,69]. Moreover, the proapoptotic and antiproliferative properties of diverse cannabinoids have shown to be at least partially mediated by PPARγ activation [70]. Therefore, it is not surprising that the anticancer actions of specific cannabinoids are mediated through a dual mechanism. This antitumor CB 1 R/PPARγ profile can be exemplified by chromenopyrazoledione 4 (Figure 1), which is a cannabinoid quinone that exerts antiproliferative effects in hormone-sensitive prostate cancer in vitro and in a murine xenograft model [71]. Experiments in the androgen-sensitive LNCaP cell line demonstrated that this compound induces cancer cell death through a mechanism that involves PPARγ and CB 1 R activation as well as oxidative stress. Moderate CB 1 R binding affinity was reported, but direct PPARγ activation remains to be examined. Even though its ability to inhibit tumor growth was confirmed in prostate cancer xenograft mice, the suggested dual CB 1 R/PPAR mechanism needs to be confirmed in vivo.
The non-intoxicating phytocannabinoid CBD has also been reported to exert antitumor actions through CB 1 R and PPARγ in specific types of cancer [70,72,73]. Studies in colorectal carcinoma cell lines showed that CBD was able to significantly reduce cell proliferation via CB 1 R, PPARγ, and TRPV1 activation. Its ability to protect DNA from oxidative damage and enhance endocannabinoid levels were also observed upon CBD treatment [73]. CBD had been previously shown to bind and activate PPARγ [74]; nevertheless, weak CB 1 R activity was reported for this phytocannabinoid [75]. In light of that, the authors attribute CB 1 R-mediated antiproliferative effects in colon cancer cells to indirect activation due to endocannabinoids enhancement [73]. It is important to note that CBD antitumor properties in other types of cancer have not been related to dual CB 1 R and PPARγ [72]. For instance, in lung cancer cell lines, CBD mediates proapoptotic effects via PPARγ and COX-2 but not through CB 1 R, CB 2 R, or TRPV1, as demonstrated upon treatment with their corresponding antagonists [76].
Although strategies targeting CB 1 R/PPARα have been explored in further depth for the treatment of obesity, the concomitant modulation of CB 1 R and PPARγ may also be useful for the control of metabolic syndrome and related disorders. A recent study demonstrated that these two targets are involved in the antiobesity and antiadipogenic effects of leaves extracts from Mangifera indica (EMI) [77]. Treatment with EMI led to a reduction of food intake and adipose tissue in a rat model of cafeteria diet-induced obesity. These effects were accompanied by increased PPARγ and decreased CB 1 R mRNA expression. Curiously, the major bioactive component of EMI, the xanthone glycoside mangiferin (Figure 1), produced proadipogenic effects in the same rat model. The presence of other phenolic compounds identified in the extracts may enable the beneficial effects produced by EMI when compared to isolated mangiferin [77]. Studies to explore the direct modulation of CB 1 R and PPARγ with the extracts and with the isolated compounds remain to be done.
The role of CB 1 R and PPARγ has also been extensively demonstrated in inflammatory processes [78][79][80]. Indeed, ∆ 9tetrahydrocannabinolic acid (THCA), a phytogenic precursor of THC, exerts anti-arthritis activity through CB 1 R and PPARγ pathways [81]. In a murine model of collagen-induced arthritis, THCA was able to significantly decrease inflammatory biomarkers, synovial hyperplasia, and cartilage damage. These effects were abolished upon treatment with either SR141716 or T0070907 (PPARγ antagonist). THCA direct activation of both CB 1 R and PPARγ was confirmed by competitive binding assays and functional studies [81,82]. At CB 1 R, it was shown to act as an orthosteric agonist or as a positive allosteric modulator in the presence of the synthetic agonist CP-55,940 [81]. It also behaves as a CB 2 R inverse agonist; however, this effect was not involved in ∆ 9 -THCA anti-arthritis properties.
Collectively, these data demonstrate the therapeutic potential of CB 1 R-PPARα and CB 1 R-PPARγ pharmacological strategies in a wide range of pathologies. Approaches targeting these ECS receptors can effectively trigger synergistic actions that may offer better hopes in the course of diverse diseases. However, so far, clinical trials remain to be performed in order to confirm the results observed in animal models. A summary of CB 1 R-PPAR compounds associated by target and disease is displayed in Table 1.

CB 2 R-PPAR Modulation
Dual targeting CB 2 R/PPAR has also been explored as a promising therapeutic approach in the course of diverse diseases.

CB 2 R-PPARγ
Since CB 2 R and PPARγ share immunomodulatory properties, multitarget strategies for these receptors are primarily focused on pathologies with a pro-inflammatory component. In fact, several natural and synthetic dual ligands have shown promising results as therapeutic options in preclinical and clinical studies for arthritis, fibrotic diseases, multiple sclerosis, neurodegenerative disorders, substance abuse and mood disorders, cancer, and metabolic dysfunction (see Table 1). On the one hand, cannabinoids targeting CB 2 R show beneficial effects on rheumatoid arthritis by inhibiting inflammation and osteoclastogenesis in preclinical assays [126][127][128]. On the other hand, PPARγ is expressed in synoviocytes, where it regulates the expression of inflammatory mediators and osteoclast differentiation [129,130]. Ajulemic acid (AJA) (Figure 1) is a synthetic derivative of ∆ 9 -THC-11-oic acid with beneficial results in rheumatoid arthritis, as shown in experimental in vitro and animal models. AJA reduces inflammation through both CB 2 R and PPARγ activation [83,131,132], leading to a significantly reduction in pro-inflammatory mediators [131][132][133][134], edema [133,135,136], synovitis [133], bone and cartilage damage [133,137], and osteoclastogenesis [138]. Similar effects were observed with β-caryophyllene (BCP, also trans-caryophyllene) (Figure 1), which is a selective CB 2 R phytochemical compound that also binds to PPARα and upregulates PPARγ in a CB 2 R-dependent manner [84,139,140].
BCP prevented inflammation and cartilage damage in an experimental animal model of arthritis by reducing pro-inflammatory cytokines and recovering PPARγ expression [84]. These effects were abrogated by AM630, a CB 2 R antagonist, suggesting CB 2 R-involvement.
A combined multitarget approach can be a valuable strategy in other chronic autoimmune diseases, such as systemic sclerosis or scleroderma. CB 2 R agonists have shown anti-fibrotic and anti-inflammatory effects in animal models of fibrosis [141][142][143], whereas PPARγ agonists can block the pro-fibrotic transforming growth factor beta (TGF-β)/SMAD (small mothers against decapentaplegic) signaling pathway at the nuclear level [144]. Accordingly, AJA and the cannabidiol quinones VCE-004.3 and VCE-004.8 (also known as EHP-101 Figure 1), reduced the fibrotic and inflammatory response in animal models of systemic sclerosis by interfering with SMAD transcriptional activity and decreasing TGF-β-induced collagen synthesis and release [85][86][87]145]. Treatment with these dual compounds also modified the transcriptome signature, reducing gene expression of proinflammatory and pro-fibrotic signals [86][87][88]146]. AM630 was able to attenuate the effects of both quinones, while the PPARγ antagonist T0070907 only prevented VCE-004.8 effects but not the ones of VCE-004.3 [86,87]. However, that does not necessarily exclude PPARγ involvement, since both quinones and AJA bind at different site of the ligand-binding domain (LBD) than other canonical PPARγ ligands, including T0070907 [86,100]. These favorable effects have advanced these CB 2 R-PPARγ ligands into clinical trials with promising results regarding systemic sclerosis [146,147].
VCE-004.8 and BCP have shown very promising results over multiple sclerosis, which is an autoimmune demyelinating disease characterized by severe neuroinflammation [89,90,[148][149][150]. Both ligands were able to significantly improve the clinical features in experimental animal models by reducing the exacerbated inflammation in the central nervous system (CNS) [89,90,148,149] and, in the case of BCP, by also decreasing oxidative stress. While the effects of VCE-004.8 over multiple sclerosis pathogenesis are CB 2 Rand PPARγ-independent [90], BCP beneficial effects were reported to depend on CB 2 R activation at lower concentrations and PPARγ activation at higher concentrations [150].
Multiple reports showed that BCP and the natural product 4 -O-methylhonokiol (MHK) (Figure 1) prevented neuronal cell death with subsequent cognitive improvement in several Alzheimer's disease animal models [91,92,[151][152][153][154][155][156]. These natural products exerted their therapeutic effects by decreasing oxidative stress and inflammation [91,92,153,[155][156][157]. Moreover, oral treatment with MHK increased amyloid β clearance by modulating the expression of β-secretase, the amyloid precursor protein), and several amyloid β-degrading peptidases [91,[152][153][154]. Although the molecular target through which MHK exerts these effects has not yet been determined, this neolignan shows affinity for cannabinoid receptors with a 50-fold selectivity over CB 2 R [158] while also binding to the LBD of both PPARα and PPARγ [159]. On the other hand, the therapeutic effects of BCP over Alzheimer's disease rely mainly on CB 2 R activation, and to a lesser extent, on PPARγ activation [92,156]. In fact, BCP has shown neuroprotective effects mainly in rodent models in other neurodegenerative diseases, such as Parkinson's disease, thanks to its antioxidative and anti-inflammatory properties [93,160,161]. In these cases, the effects were prevented by AM630, implying CB 2 R involvement [160][161][162].
Remarkably, research has shown that BCP might also be useful in substance abuse disorders. Treatment with the sesquiterpene decreased alcohol consumption and the conditioned-rewarding effect of ethanol in a murine model of voluntary alcohol intake [94]. The response was thought to be mediated via CB 2 R activation, since co-administration with AM630 abrogated BCP effects. On the other hand, high doses of BCP also prevented cocaine self-administration and relapse in rats in a CBR-independent manner [95]. However, in this study, PPARα and PPARγ agonists had a similar effect to BCP, and co-administration with PPARα or PPARγ antagonists reversed BCP-induced improvement. It is worth noting that substance use disorders and anxiety frequently co-occur [163], and BCP has also shown anxiolytic and antidepressant effects via CB 2 R activation [164].
Diverse studies have shown the antitumorigenic effects of cannabinoids and the ECS as a target, which have been reviewed recently [165]. This also includes the antiproliferative effects of CB 2 R [166][167][168] and PPARγ [35]. The dual CB 2 R/PPARγ ligands AJA, MHK, and BCP have proven to possess beneficial effects as antineoplastic agents in different experimental models. In several tumor cells, AJA inhibited tumor growth in a CB 2 Rdependent, CB 1 R-independent manner, and when tested in vivo, oral administration of AJA delayed the appearance and size of tumors in nude mice [96]. Moreover, AJA also had a modest improvement in survival [169]. It is worth noting that pro-inflammatory cytokines can promote microenvironments favoring tumor growth and invasiveness [170], and AJA can decrease the expression and secretion of these cytokines [132,134]. Meanwhile, MHK and BCP promote cell cycle arrest and subsequent apoptosis in several tumor cell lines by stimulating PPARγ activity and decreasing NF-κB DNA binding activity [97,98,[171][172][173][174], which is a pattern already observed with other PPARγ ligands [175,176]. GW9662, a PPARγ antagonist, reversed the anti-cell growth and apoptotic effects of MHK and decreased phosphatase and tensin homolog (PTEN) expression [97,171], whereas AM630 prevented BCP antiproliferative response [98]. Other well-established CB 2 R/CB 1 R and CB 2 R ligands, such as WIN55,212-2 and JHW-015 respectively, have shown that their antineoplastic effects are also mediated through PPARγ activation [55,99].
Obesity and metabolic syndrome are associated to chronic low-grade inflammation. BCP has shown promising properties as a therapeutic agent in metabolic disorders. In preclinical studies, this sesquiterpene stimulates insulin secretion and sensitization [177,178], reduces glucose plasma levels and gastrointestinal absorption [179,180], decreases hyperlipidemia [179], promotes an anti-inflammatory and antioxidant environment, and protects against diabetic complications [139,181]. These effects are mediated through a combination of CB 2 R, PPARα, and PPARγ activation, and they have been recently reviewed elsewhere [103,182]. Other dual ligands, such as VCE-004.8, have also shown beneficial properties in animal models of obesity. Although this quinone shows a lower adipogenic profile than PPARγ full agonists, treatment with VCE-004.8 was able to significantly reduce weight gain, fat mass, and adiposity [100]. Moreover, VCE-004.8 decreased crown-like structures, which is a feature of macrophage infiltration in the adipose tissue of obese mice. The beneficial effects of VCE-004.8 on obesity might not be exclusively related to PPARγ activation. Research shows that beyond its effects on inflammation, CB 2 R is also involved in energy homeostasis and food intake [183]. Meanwhile, MHK displayed a low-to-moderate effect over body weight gain, fat mass, and adipocyte hypertrophy in mice [101]. However, MHK has protective effects in diabetes-related pathologies. In highfat diet-fed mice, MHK prevents fibrosis, organ injury, and lipid accumulation in liver, heart, and kidney [101,184,185]. Remarkably, MHK significantly downregulated CB 1 R overexpression and increased the hepatic concentrations of several N-acetylethanolamines, including AEA, OEA, and PEA [186].

CB 2 R-PPARα
The co-activation of CB 2 R and PPARα has been much less explored, although some evidence suggests that simultaneously targeting both receptors may also be an interesting pharmacological option in some pathology.
Ulcerative colitis, one of the main types of inflammatory bowel disease, is characterized by chronic persistent inflammation in the gastrointestinal tract. Oral administration of PEA has been shown to protect against edema and mucosa erosion in a dinitrobenzene sulfonic acid (DNBS)-induced ulcerative colitis murine model [102]. Moreover, this acylethanolamide displayed anti-inflammatory effects that were counteracted by CB 2 R, GPR55, or PPARα antagonists. Since PEA does not bind directly to CB 2 R [62], authors suggest that PEA might modulate CB 2 R through the entourage effect. In fact, it is well established that PEA can inhibit AEA inactivation and thus activate CBRs indirectly [63].
As summarized herein, several CB 2 R-PPARγ dual ligands have shown therapeutic relevance in preclinical and/or clinical studies for the management of inflammation, cancer, and neurodegeneration. CB 2 R and PPARα concomitant actions remain to be further investigated in disease models.

FAAH-PPAR Modulation
The degradation of enzymes, uptake, and intracellular transport upregulate the level of endocannabinoids [13]. Some of the pathophysiological effects of the inhibitors of these proteins have been described to be mediated through PPARs (for an understanding review, see Pistis and O'Sullivan [40]). Thus, we will report here the most recent advances that have been especially focused on FAAH inhibitors and PPARα agonists from these last years. FAAH is a serine hydrolase responsible for the degradation of AEA [187,188], 2-AG [189], PEA, and OEA [39] among other fatty acid compounds. Dual FAAH/PPARs inhibitors associated with diseases presented below are reported in Table 1.
The contribution of PPARα in the mechanism of action of FAAH inhibitors to reduce diverse nausea and abused-drug reward conditions has been evidenced in several preclinical studies [106,109]. Curiously, two FAAH inhibitors, URB597 and PF-3845 (Figure 1), showed different mechanisms of action to reduce acute nausea and anticipatory nausea in rodent models of conditioned [106]. PF-3845 suppressed acute nausea via PPARα, but not CB 1 R, whereas URB597 reduced anticipatory nausea via CB 1 R, but not PPARα. Despite the fact that these two nausea conditions have different molecular basis, the divergence between the two FAAH inhibitors may be due to a difference of selectivity and potency on FAAH as suggested by the authors, PF-3845 being more potent and selective than URB597 [106].
Dual FAAH/PPARα mechanisms of action were also investigated in opioid withdrawal. For instance, the aversive effects of acute naloxone-precipitated morphine withdrawal by oleoyl glycine (OlGly; Figure 1) was shown to be mediated by both CB 1 R and PPARα in a rat model [109]. OlGly has been described as an FAAH inhibitor and PPARα agonist in vitro but not as CBRs ligand CBRs [110]. These data suggest the indirect activation of CB 1 R via FAAH inhibition and subsequent elevation of AEA levels. Following with studies on OlGly in naloxone-precipitated morphine withdrawal in rodent models, Ayoub et al. [108] designed and synthesized a OlGly derivative, oleoyl alanine (HU595, Figure 1). In these experiments, the effect of HU595 was prevented by pre-treatment with either the PPARα antagonist MK886 or the CB 1 antagonist AM251, suggesting a PPARα and CB 1 R mechanism of action. Interestingly, at equivalent doses, its effects lasted longer than for OlGly, suggesting that it could be due to the enhanced stability of HU595 to hydrolysis [108]. It is worth mentioning that the mechanism of action of OlGly in reward and withdrawal conditions cannot be generalized as dual. For instance, OlGly was effective in nicotine reward and withdrawal [110], and this effect was proposed to be mediated by PPARα rather than FAAH. This was based on the fact that in a nicotine-dependent mice-conditioned place preference paradigm, the PPAR-α antagonist GW6471 prevented the effect of OlGly, whereas the lack of activity at FAAH was based on in vitro luciferase assays in which OlGly showed low agonism [110].
While designing new N-acylethanolamine acid amidase (NAAA) inhibitors based on azetidine-nitrile pharmacophore, Malamas et al. [105] identified potent and selective dual NAAA-FAAH inhibitors in vitro. For instance, the inhibitor 52 ( Figure 1) was found to be highly potent at NAAA and FAAH, showing negligible activity at MAGL, ABHD6, and cathepsin K. Such dual inhibitors represent an interesting strategy that combines two distinct anti-inflammatory molecular pathways. NAAA inhibition stimulates the PEA/PPAR-α anti-inflammatory signaling pathway, whereas FAAH inhibition activates the CBRs by increasing AEA levels. Using the same FAAH/NAAA dual strategy, Wu et al. [104] explored carmofur (Figure 1), a 5-fluorouracil used clinically as an antineoplastic drug, based on the fact that this molecule contains a urea group, which is a common group in inhibitors of FAAH and NAAH. In an acute lung injury model, the anti-inflammatory effect of carmofur was blocked by the PPARα antagonist MK886 and the CB 2 R antagonist SR144528. From a mechanistic point of view, carmofur did not bind either PPARα or CB 2 R in competitive binding assays, but it has been confirmed as FAAH and NAAA inhibitor in enzymatic assays. Thus, its mechanism of action has been suggested to be the through indirect activation of PPARα and CB 2 R [104].
Recently, Fotio et al. [107] studied the FAAH inhibition on tolerance to the antinociceptive effects of morphine in mice by monitoring nociceptive thresholds by the tail immersion assay. The FAAH inhibitor URB597 prevented the development of morphine tolerance without altering nociceptive thresholds. These effects were mediated by a mechanism involving CB 2 R and CB 1 R and PPARα [107]. The peripherally restricted FAAH inhibitor URB937 did not modified morphine tolerance in these assays, revealing a central mechanism. Therefore, combining a central acting FAAH inhibitor and an opioid could be an interesting strategy to multimodal analgesia, as suggested by these authors [107].
Similar to PPARγ agonists, FAAH inhibitors have been implicated in antiproliferative, pro-apoptotic, and cytotoxic effects on several cancerous cell (recently reviewed by Brunetti and colleagues [190]). In this context, a dual PPAR/FAAH strategy could offer synergistic effects for tumor management.

Other CBR-PPAR Modulatory Profiles
In addition to cannabinoid/PPAR compounds acting preferentially at one of the two CBRs, some of them exhibit potential therapeutic actions through either receptors or none of them, these last ones being included here from a structural point of view. An association of disease and cannabinoids presented below is reported in Table 1.
The phytocannabinoid cannabimovone (CBM; Figure 1) is a natural terpenoid structurally related to CBD that is devoid of significant affinity for either CB 1 R or CB 2 R [192], but it has been recently described as a PPARγ agonist that is able to stimulate insulin signaling in vitro [193].
In 2005, O'Sullivan and co-workers [120] provided strong evidence that THC is a PPARγ ligand that can produce PPARγ-mediated effects independent of direct CBRs activation. In their assays realized in rat aorta, THC induced time-dependent vasorelaxation through the increased bioavailability of nitric oxide and hydrogen peroxide production. This effect was unaffected by CB 1 R antagonism but inhibited by the PPARγ antagonist GW9662. THCA, the acidic precursor of THC, is a less efficacious agonist for both CB 1 R and CB 2 R than THC but displays higher potency at PPARγ than THC, as reviewed by Moreno-Sanz [194]. This activity at PPARγ has been confirmed in a model of Huntington's disease where THCA showed neuroprotective activity through a PPARγ-dependent pathway [82].
The controlled oxidation of CBG, a minor component of Cannabis sativa but precursor of CBD and THC, led to the identification of the cannabigerol quinone VCE-003 ( Figure 1) [118]. VCE-003 has been revealed to be a PPARγ partial agonist with no significant activity at CB 1 R and only moderate activity at CB 2 R. It has been explored successfully for its neuroinflammatory and neuroprotective properties in a murine model of multiple sclerosis [118]. A derivative of VCE-003, VCE-003.2 (also known as EHP-102; Figure 1), which does not exhibit activity at CBRs, binds and activates PPARγ transcriptional ac-tivity in competitive assays, as does its parent compound VCE-003. VCE-003.2 showed neuroprotective effects in a murine model of Huntington's disease that was prevented by co-administration of the PPARγ antagonist T0070907 and in a murine model of Parkinson's disease [114][115][116][117].
PPARγ plays a role in some of the potential therapeutic effects of the synthetic highaffinity CB 1 R and CB 2 R agonist WIN55,212-2. For instance, in a amyloid β-induced neuroinflammation model, WIN55,212-2 increases PPAR-γ signaling through CB 1 R, in addition to its direct action on both CB 1 R and CB 2 R [113]. It has also been described that the effects of WIN55,212-2 on behavioral sensitization induced by abused drugs were mediated through PPAR-γ [121].
As already mentioned, diverse endocannabinoids have been implicated in PPARs activation, and some of their mechanisms of action remain elusive. Studies such as the ones performed by Kaminski and co-workers [195] suggest that endocannabinoid metabolism may play an important role in the action of endocannabinoids at PPARs. In the mentioned studies, a COX-2 downstream metabolite of 2-AG, 15-desoxy-∆ 12,14 -prostaglandin-J 2 -glycerol ester (15d-PGJ 2 -G; Figure 1) which is a PPARγ activator, has been shown to be in part responsible for interleukin (IL)-2 suppression by 2-AG through effectively a PPARγ mechanism of action.
The direct action of cannabinoids on PPARs or their signaling has been widely reported, whereas PPAR agonists for activity at the CBRs have been much less studied. The phytochemical product magnolol (Figure 1), which is structurally related to MHK, is a PPARγ partial agonist that binds in a dimeric mode to the nuclear receptor [196]. Magnolol has been found to possess a wide variety of pharmacological actions that have been recently reviewed elsewhere [197]. Through PPARγ modulation, this neolignan exerts antiinflammatory actions in animal models of acute lung injury [198], ulcerative colitis [199], and Alzheimer's disease [200], as well as cardioprotective effects [201,202], and beneficial properties in metabolic disorders [203]. Interestingly, magnolol has also been found to act as a partial agonist of CB 1 R and CB 2 R receptors [204], and its derivatization has led to the design of novel synthetic cannabinoid ligands with improved affinity [205].

CBR-PPARα
In a computational approach, CBD was considered a poor ligand for the PPARα receptor type [206]. In the same study, the CBD derivatives cannabidiolic acid (CBDA; Figure 1), cannabigerolic acid (CBGA; Figure 1), and cannabigerol (CBG; Figure 1) were predicted dual PPARα/γ by performing MD simulations and docking using the crystallographic structures of PPARα and PPARγ, which was subsequently confirmed by luciferase reporter gene assays [206]. Recently, Rock et al. [125] showed that the anti-nausea effect of another phytocannabinoid, THCA, was blocked by a PPARα antagonist, whereas the effect of CBDA was mediated by the 5-HT 1a . Thus, they proposed the co-administration of CBDA and THCA as a combined therapy for treating chemotherapy-induced nausea through different mechanisms.
At PPARα, THC has been mentioned by Sun et al. [41] not binding to the purified PPARα ligand binding domain and not having effects on PPARα-driven transcriptional activity, whereas in studies by Takeda et al. [122], THC stimulated the activity of PPARα. This later showed that THC not only increases the expression of fatty acid 2-hydroxylase (FA2H) via the upregulation of PPARα expression in human breast cancer MDA-MB-231 cells but that THC also interferes with the PPARβ/δ-mediated inhibition of PPARα.
WIN55,212-2 was shown to increase PPAR-γ signaling, and it also displays PPARαbinding affinity with biphasic effects on PPARα gene-transcription activity [41]. This PPARα activation could explain the higher lipolytic effect of WIN55,212-2 compared to other cannabinoids. There are also evidence that the anticonvulsant effects of WIN55,212-2 assessed in mice could be mediated through CB 1 R and PPARα, since these effects are blocked by a combination of CB 1 R and PPARα antagonists [123].
The PPARα agonist fenofibrate binds both CB 1 R and CB 2 R with submicromolar affinity [207]. In [ 35 S]-GTPγS binding assays, fenofibrate was characterized as a CB 2 R agonist, whereas at CB 1 R, it acts as partial agonist at low concentrations and negative allosteric modulator at high concentrations. In cell-based CB 2 R/β-arrestin assays, fenofibrate was also identified as a CB 2 R agonist [208]. Thus, this PPARα activator can be considered a multitarget agent due to its possible activity at CBRs in addition to its primary PPAR target.
The acylethanolamides OEA and PEA are structurally related to the endogenous cannabinoid AEA with which they compete for the endocannabinoid-metabolizing enzyme, FAAH. In luciferase reporter gene assays and LBD of human PPARs, OEA binds PPAR-α with high affinity and PPARβ/δ but not PPARγ, whereas PEA only binds PPAR-α [56,209]. They also act at TRPV1 and GPR55 [13]. As already commented in this review, OEA and PEA do not bind CBRs; however, some of their activities have been suggested to be mediated through CBRs.
In the context of a dual PPAR/CBR strategy, it is worth mentioning the randomized, placebo-controlled, double-blind controlled trial realized on healthy human subjects in which a state of increased gut permeability was induced by aspirin [124]. In this clinical study, CBD and PEA could reduce human inflamed gastrointestinal permeability in vivo, suggesting the potential of these compounds for clinical use in inflammation bowel diseases.
As detailed in this section, a range of cannabinoids exert their therapeutic effects, acting preferentially via PPAR. On the other hand, specific compounds mediate their beneficial actions through both CB 1 R and CB 2 R along with a PPAR subtype. These pharmacological profiles are also worth exploring due to their anti-inflammatory, neuroprotective, or antitumoral properties.

Conclusions and Future Perspectives
The ECS is involved in a variety of multifactorial pathologies including neurodegenerative diseases, cancer, and metabolic syndrome. In this context, multitargeting strategies offer better hopes for an effective treatment due to the synergistic therapeutic effects triggered upon the modulation of diverse signaling pathways.
Molecular targets of the cannabinoids include the cannabinoid receptors CB 1 R and CB 2 R, orphan GPCRs, such as GPR18 and GPR55, ionotropic receptors such as GlyRs or GABA receptors, TRP channels including TRPV1-4, TRPA1, and TRPM8, and the PPAR family of nuclear receptors. The intrinsic complexity of the ECS along with variations associated to each disease phenotype led to diverse therapeutic options targeting this signaling system. This review provides a summary of reported molecules mainly exerting therapeutic actions via PPARs and the cannabinoid receptors and/or FAAH. In this context, promising dual strategies are synergistic CB 1 R blockage-PPARα activation, which offers therapeutic opportunities for the treatment of obesity or the coactivation of CB 2 R and PPARγ, which has shown great efficacy in the management of cancer or fibrosis.
Molecules exhibiting therapeutic actions mediated by CBR/PPAR targeting include endogenous, phytogenic, and synthetic compounds (the functional profile of these compounds has been summarized in Table 2). For instance, the endogenous N-acetylethanolamine PEA directly modulates PPARα and indirectly modulates the CBRs exerting antiepileptic, analgesic, or anti-inflammatory effects in different in vivo models. Phytogenic molecules including CBD, MHK, or BCP have also shown therapeutic properties mediated through CBR and PPAR responses. It is also worth highlighting the promising potential of cannabinoid acids. While the natural THC analogue THCA displays anti-arthritis activity through CB 1 R and PPARγ pathways, the synthetic THC acid derivative AJA exhibits antitumoral and anti-inflammatory effects via CB 2 R and PPARγ activation. The promising pharmacological potential of the latter has led it into clinical trials for chronic inflammatory diseases including systemic sclerosis, cystic fibrosis, dermatomyositis, and systemic lupus erythematosus [146,169,215]. The most remarkable synthetic compounds with this dual profile are the cannabinoid quinones. In fact, the cannabidiol quinone derivative developed by Emerald Health Pharmaceuticals, VCE-004.8, has entered clinical trials for its anti-inflammatory and neuroprotective properties [147]. Dual FAAH-PPAR activity has also been highlighted for its multitarget therapeutic value. Indeed, molecules acting as FAAH inhibitors and PPAR activators have shown beneficial actions for the treatment of pain, cancer, or inflammation.
Tridimensional structures of the cannabinoid and nuclear receptors and the fatty acid enzyme FAAH have been recently solved. While numerous crystallographic studies have been reported for all PPAR types, CB 1 R and CB 2 R have not been elucidated until the past few years. PPARα and γ have been solved in complex with a wide range of ligands [216][217][218][219]. However, thus far, the only reported PPAR structure bound to a cannabi-noid is the AJA-PPARγ complex [210]. In what concerns the cannabinoid receptors, crystal structures of the active and inactive states [15][16][17] and more recently cryo-EM CB 1 R-G i and CB 2 R-G i structures in complex with agonists [19,21,22] have helped the understanding of their activation and signaling mechanism. Moreover, a high-resolution crystal structure of FAAH bound to its inhibitor URB597 has also been reported [220]. All these structures may help understanding the molecular basis of cannabinoid dual targeting and guide future drug design.
It is important to highlight that compounds exhibiting CBR/PPAR dual profile might be able to reach cell membrane and nuclear receptors. Cannabinoids are lipophilic molecules that reach their CBR binding site upon diffusion through the lipid bilayer from the extracellular milieu. The lipid GPCR structures reported in the last years have shed light into a better understanding of how cannabinoids reach their targets [15][16][17]19,21,22]. Portals in the transmembrane helices may facilitate their access to the orthosteric binding site while allosteric sites have been found at the receptor interface. However, it is not fully determined how cannabinoids access the PPAR binding crevice. On the basis of reviewed studies, O'Sullivan [31,39,40] suggested different mechanisms of action: (i) Activation of the CBRs by cannabinoids induces signaling cascades that indirectly activate PPAR; (ii) Cannabinoids can bind directly to PPARs; (iii) Cannabinoids can be transported to PPARs by fatty acid-binding proteins (FABPs) s; (iv) Cannabinoid metabolites can be the ones that activate PPARs ( Figure 2). These mechanisms may vary depending on each molecule; however, further mechanistic studies remain to be done in order to determine cannabinoid-PPAR interactions. In addition to well-known unbiased modulatory profiles, functionally selective or biased agonists have recently emerged as a promising pharmacological approach in GPCRs [221]. Increasing evidence has shown that β-arrestin 1 and β-arrestin 2 can also recruit signaling mediators to the activated receptors, thus eliciting signaling cascades In addition to well-known unbiased modulatory profiles, functionally selective or biased agonists have recently emerged as a promising pharmacological approach in GPCRs [221]. Increasing evidence has shown that β-arrestin 1 and β-arrestin 2 can also recruit signaling mediators to the activated receptors, thus eliciting signaling cascades beyond G protein-mediated signals [221]. Functional selectivity at CBRs has been intensively studied in the past few years [222][223][224]. However, how biased modulators of CBRs toward one or the other signaling pathway affect these CBR/PPAR dual strategies remains to be characterized.
Interestingly, it has been reported that after the GPCR activation of specific receptors (such as the κ and δ opioid receptors [225]), β-arrestin 1, but not β-arrestin 2, can be actively transported to the nucleus [226]. Nuclear translocation correlates to an increased gene transcription of cell proliferation regulating genes, and presumably, histone modification and chromatin remodeling [225]. Moreover, β-arrestin 1 has been shown to directly interact with the LBD of both PPARα and PPARγ and modulate their transcriptional activity [227,228]. Although there has been no evidence yet in this regard, research is needed to determine if cannabinoid receptor activation could modulate PPAR transcriptional activity via β-arrestin 1 nuclear translocation and whether biased cannabinoid ligands elicit different responses through this mechanism. In addition to CBR functional selectivity, the therapeutic relevance of CBR allosteric modulation/PPAR activation should be further explored under diverse physiopathological conditions.
Extensive research supports the relevance of CBR-PPAR dual targeting in the context of a broad range of pathologies. In fact, multitarget drugs can offer an improved therapeutic profile in complex diseases, as already occurs in multifactorial disorders such as cancer, neurodegenerative diseases, or psychiatric disorders [229]. In this sense, CB 2 R and PPARγ modulators are at the forefront of multitarget cannabinoid-related drugs. Compounds such as AJA or EHP-101 have successfully reached clinical stages where they are being evaluated in autoimmune disorders, including systemic sclerosis. As a result of their polypharmacological profile, these drugs can offer therapeutic advances over the current combinatorial strategies, not only because of the synergistic effect that their molecular targets exert but also by improving the pharmacokinetic profile and reducing the risk of toxicity and drug-drug interactions when multiple drugs are administered.
While the studies reviewed herein have greatly enhanced our understanding of the cannabinoids/PPARs interactions, there are important challenges that remain to be investigated. With the exception of the aforementioned CB 2 R/PPARγ compounds, all the data summarized herein comes from animal models; therefore, the translational potential of these multitarget drugs remains to be determined in clinical trials. The design of selective and potent dual ligands may help further understanding the relevance of multitargeting approaches within the ECS.