The anti-inflammatory and vasoprotective properties of mPGES-1 inhibition offer promising therapeutic potential

ABSTRACT

Introduction

Prostaglandin E₂ (PGE₂) is produced by cyclooxygenases (COX-1/2) and the microsomal prostaglandin E synthase 1 (mPGES-1). PGE₂ is pro-inflammatory in diseases such as rheumatoid arthritis, cardiovascular disorders, and cancer. While Nonsteroidal anti-inflammatory drugs (NSAIDs) targeting COX can effectively reduce inflammation, their use is limited by gastrointestinal and cardiovascular side effects resulting from the blockade of all prostanoids. To overcome this limitation, selective inhibition of mPGES-1 is being explored as an alternative therapeutic strategy to inhibit PGE₂ production while sparing or even upregulating other prostaglandins. However, the exact timing and location of PGH₂ conversion to PGD₂, PGI₂, TXB₂ or PGF_2α, and whether it hinders or supports the therapeutic effect of mPGES-1 inhibition, is not fully understood.

Areas covered

The article briefly describes prostanoid history and metabolism with a strong focus on the vascular effects of prostanoids. Recent advances in mPGES-1 inhibitor development and results from pre-clinical and clinical studies are presented. Prostanoid shunting after mPGES-1 inhibition is highlighted and particularly discussed in the context of cardiovascular diseases.

Expert opinion

The newest research demonstrates that inhibition of mPGES-1 is a potent anti-inflammatory treatment strategy and beneficial and safer regarding cardiovascular side effects compared to NSAIDs. Inhibitors of mPGES-1 hold great potential to advance to the clinic and there are ongoing phase-II trials in endometriosis.

GRAPHICAL ABSTRACT

KEYWORDS:

• mPGES-1
• PGE₂
• prostacyclin
• pharmacological inhibition
• inflammation
• cardiovascular disease
• prostanoid shunting

1. Introduction

Since their discovery in the 1930s, prostanoids have been found to have various physiological and pathophysiological functions. These include regulating blood flow, vascular permeability, smooth muscle cell function, renal filtration, gastrointestinal integrity, sleep regulation, platelet function, inflammatory processes, and reproduction [1,2].

Prostanoids are derived from arachidonic acid, which is part of the phospholipid bilayer of cell membranes. When intracellular Ca²⁺ levels increase, the cytosolic enzyme phospholipase A₂ (PLA₂) is activated, leading to the hydrolysis of phospholipids. This process generates free arachidonic acid and a lysophospholipid.

Arachidonic acid is then converted into PGH₂ by the enzymes cyclooxygenases (COX-1 and COX-2). Subsequently, PGH₂ is further transformed into prostaglandins and thromboxanes, collectively known as prostanoids (PGE₂, PGD₂, PGI₂, PGF_2α, and TXA₂) (see Figure 1). Cyclooxygenases are enzymes bound to the cell membrane that possess cyclooxygenase and peroxidase activities. Initially, arachidonic acid is oxidized with two molecules of oxygen (O₂), resulting in the formation of the unstable endoperoxide intermediate PGG₂. PGG₂ is then reduced to PGH₂ [3]. COX-1 and COX-2 are isozymes that differ in their expression patterns and tissue localization. While COX-1 is typically constitutively expressed, COX-2 is also inducible and contributes to prostanoid production during inflammation. PGE₂, in particular, is targeted in anti-inflammatory and analgesic treatment strategies. Inflammatory PGE₂ is produced from PGH₂ by an enzyme called microsomal prostaglandin E synthase 1 (mPGES-1), which was discovered by Jakobsson and colleagues in 1999. MPGES-1 is a microsomal, GSH-dependent, and inducible enzyme [4]. Prostanoids exert their effects through activation of rhodopsin-like seven transmembrane spanning G protein-coupled receptors (GPCRs), acting as endocrine, autocrine, or paracrine signaling molecules. There are nine main prostanoid receptors identified, including DP1 and DP2 (CRTH2) for PGD₂, EP1–4 for PGE₂, TP for TXA₂, IP for PGI₂, and FP for PGF_2α [5]. The activation of downstream target gene activation results from specific coupling of the prostanoid receptors to different G-proteins (see Table 1). Coupling to the Gq protein, which is the case for EP1, TP and FP activate phospholipase C (PLC) which then cleaves phosphatidylinositol 4,5-bisphosphate (PIP2) into inositol triphosphate (IP3) and diacylglycerol (DAG). IP3 consequently activates intracellular Ca²⁺ levels and DAG activates the protein kinase C (PKC). Gq proteins can also activate Rho kinase. The EP3 as well as the DP2 receptor couple to Gi protein which inhibits adenyl cyclase (AC). This reduces cyclic adenosine monophosphate (cAMP) levels. Coupling to the Gs protein in turn activates AC and stimulates cAMP levels and protein kinase A (PKA). EP2, EP4, DP1 and IP couple to the Gs protein.

Figure 1. Biosynthesis of the primary prostanoids (PGE₂, PGD₂, PGI₂, PGF_2α and TXA₂) and some of their metabolites. Enzymes are highlighted in gray squares. COX-1 (cyclooxygenase 1), COX-2 (cyclooxygenase 2), mPGES-1 (microsomal prostaglandin E synthase 1), mPGES-1 (microsomal prostaglandin E synthase 2), cPGES (cytosolic prostaglandin E synthase), PGIS (prostacyclin synthase), L-PGDS (lipocalin prostaglandin D synthase), H-PGDS (hematopoietic prostaglandin D synthase), TXS (thromboxane synthase), PGFS (prostaglandin F synthase), MGST3 (microsomal glutathione transferase 3).

Table 1. VSMC activation by G-protein coupled receptors. All prostanoids can act through the other than their primary receptors, however with lower affinity.

Signaling pathway	Second messenger	Primary Receptor	Agonist	Vascular effect
Gq	IP3 → Ca^2+↑, Rho-kinase ↑	TP	TXA2	constriction
		FP	PGF2α
		EP1	PGE2
Gi	cAMP↓	EP3	PGE2	constriction
		DP2 (CRTH2)	PGD2
Gs	cAMP↑	IP	PGI2	relaxation
		EP2	PGE2
		EP4	PGE2
		DP1	PGD2

1.1. Prostanoids as regulators of vascular tone

PGI₂ and TXA₂ play a crucial role in regulating vascular tone, thereby controlling blood pressure and blood flow throughout the body. Other prostanoids, such as PGE₂, PGD₂, and PGF_2α, also have an impact on vascular effects (see Table 1). Prostanoids interact directly with intracellular Ca²⁺ and cAMP levels through their respective receptors, influencing the degree of constriction of a vessel’s smooth muscle cells. The interaction between actin and myosin filaments, which is driven by the influx of Ca²⁺ from channels at the plasma membrane or from the sarcoplasmic reticulum (SR), is primarily initiated by the activation of Gq-protein coupled receptors (as depicted in Figure 2). Gq binding results in the hydrolysis of IP3 and the opening of Ca²⁺ channels in the SR. Calmodulin in the cytosol binds to Ca²⁺, activating myosin light chain (MLC) kinase. Phosphorylation of MLC, in turn, facilitates interaction between actin and myosin filaments, leading to constriction [6]. The main Gq-coupled vasoconstrictors include TXA₂, endothelin-1, angiotensin-II, vasopressin, and norepinephrine. Changes in intracellular cAMP levels can also regulate the phosphorylation of MLC. Signaling through Gi-coupled receptors results in decreased intracellular cAMP levels and consequent constriction. Alternatively, Gq proteins can activate Rho-kinase, which inhibits MLC dephosphorylation, leading to VSMC constriction [7].

Figure 2. Simplified schematic illustration of VSMC activation by GPCR signaling and NO. Activation of signaling pathways is indicated by ► and inhibition by ┤. Endothelium dependent relaxation can be induced by prostanoids through the activation of Gs coupled receptors on the VSMC. The major endothelium dependent vasodilator is NO (nitric oxide) produced by eNOS (endothelial nitric oxide synthase) which is activated by NO agonists like bradykinin (BK) or acetylcholine (ACh). NO production can be inhibited by the eNOS inhibitor ADMA (asymmetrical dimethylarginine). EDHF (endothelium dependent hyperpolarization factor) includes CYP450 products of arachidonic acid EETs (epoxyeicosatrienoic acids), H₂O₂ (hydrogen peroxide), potassium ions, gap junctions, and C-type natriuretic peptide. Endothelium dependent contraction of VSMC can be induced through Gi and Gq coupled receptors, activated by for example endothelin-I or prostanoids. Reactive oxygen species such as O₂ (superoxide) can react with NO to form OONO⁻ (peroxynitrite), a cytotoxic molecule that reduces NO availability and causes VSMC constriction. SOD (superoxide dismutase) is an antioxidant enzyme and catalyzes the dismutation of superoxide to O₂ (oxygen) and H₂O₂. VSMC constriction can also be induced by other Gq protein coupled receptor agonists such as angiotensin-II, vasopressin and norepinephrine. Additional abbreviations: MLCK (myosin light chain kinase), MLCP (myosin light chain phosphatase), SR (sarcoplasmic reticulum), cGMP (cyclic guanosine monophosphate), cAMP (cyclic adenosine monophosphate), IP3 (inositol triphosphate), PKA (protein kinase A), NO (nitric oxide). Please see text for more detailed descriptions on the depicted pathways.

When intracellular Ca²⁺ levels decrease, MLC phosphorylation also decreases, leading to relaxation of vascular smooth muscle cells (VSMC). Relaxation can also occur through inhibition of MLC phosphorylation or activation of MLC dephosphorylation. Increased cAMP levels, which activate PKA and inhibit MLCK action, mediate the inhibition of MLC phosphorylation. On the other hand, an increase in intracellular cGMP levels leads to MLC dephosphorylation and VSMC relaxation. PGI₂, an important endothelium-derived vasodilator, activates IP, a Gs-coupled receptor, and increases intracellular cAMP levels. PGI₂ serves as a mediator of inflammation, maintains homeostasis by inhibiting platelet aggregation and vasodilating blood vessels, and is considered the most important cardioprotective prostanoid.

Aside from PGI₂, endothelium-dependent relaxation can also be induced by nitric oxide (NO) and endothelium-derived hyperpolarization factor (EDHF). NO increases cGMP levels and induces MLC phosphatase in VSMC. Endothelial NO synthase (eNOS), the enzyme that produces NO in endothelial cells, is constitutively expressed and regulated by intracellular Ca²⁺ levels. It responds to inflammatory mediators and NO agonists such as bradykinin or acetylcholine [8]. NO production is also regulated through PKA-induced eNOS phosphorylation, which is induced by shear stress. NO diffuses from the endothelial cell into the VSMC, where it binds to guanylyl cyclase which produces cGMP, causing dephosphorylation of MLC, subsequently leading to VSMC relaxation [9]. EDHF causes vascular relaxation independently of NO and PGI₂ and is mediated by several potential factors, including CYP450 products of arachidonic acid (epoxyeicosatrienoic acids), H₂O₂, potassium ions, gap junctions, and C-type natriuretic peptide [10].

2. Exploring alternative strategies to traditional NSAIDs: promising therapeutic approaches with fewer side effects

Traditional NSAIDs, including aspirin, indomethacin, diclofenac, ibuprofen, and naproxen, inhibit both COX enzymes, while COX-2 selective inhibitors, also known as coxibs, such as celecoxib or etoricoxib, primarily act on COX-2. NSAIDs are commonly used as a first-line treatment for various diseases due to their anti-inflammatory, anti-pyretic, and analgesic properties, and are widely available over-the-counter for minor aches and pains, osteoarthritis, rheumatoid arthritis, migraine, and musculoskeletal injuries [11]. However, the use of NSAIDs and coxibs is associated with gastrointestinal and cardiovascular complications when used for prolonged periods. Inhibition of constitutive prostanoid production can lead to an imbalance in prostanoid homeostasis, which alters cell integrity, causing gastrointestinal side effects or reduced renal performance. While selective COX-2 inhibition was initially thought to be a safer alternative to COX-1 inhibition, blocking inducible COX-2 in endothelial cells can lead to severe cardiovascular problems, such as myocardial infarction, pulmonary hypertension, and heart failure [12]. Consequently, rofecoxib and valdecoxib have been withdrawn, and FDA approval for lumiracoxib and etoricoxib has been rejected [13–16], while etoricoxib is approved in Europe and other countries. The underlying molecular mechanism for the cardiovascular side effects of COX-2 inhibition is an imbalance of pro-thrombotic TXA₂, mainly produced by platelets via COX-1 in the vasculature, and anti-thrombotic PGI₂, mainly produced by vascular cells via COX-2 [17]. Except for low-dose aspirin, NSAIDs are associated with a negative class effect on cardiovascular safety [18]. Therefore, the use of these drugs is concerning, and alternative treatments are needed.

There are several alternative treatment strategies that may have fewer side-effects than traditional NSAIDs. These include dual COX and LOX inhibitors, nitric-oxide (NO)-NSAIDs, and mPGES-1 inhibitors. The combination of COX-1/COX-2 and 5-lipoxygenase (LOX) inhibition is believed to prevent the formation of TXA₂ in platelets, downregulate inflammatory prostanoids, and counteract gastric damage by blocking leukotrienes. Licofelone is a COX/LOX inhibitor that has shown promise in clinical trials for treating osteoarthritis with fewer gastrointestinal side effects than traditional NSAIDs [19]. NO-releasing NSAIDs (NO-NSAIDs), have shown promising effects in pre-clinical studies for treating inflammation and hypertension, and naproxcinod (NO-naproxen) is being studied clinically for treating osteoarthritis [20,21]. Selective inhibition of mPGES-1, the synthase of PGE₂ downstream of COX, has been proposed as an alternative target to COX. This approach could selectively reduce inflammatory PGE₂ without causing undesirable side effects by sparing other prostaglandins, especially PGI₂. Thereby, inhibition of mPGES-1 compared to other alternatives holds most promise.

2.1. Anti-inflammatory aspects of mPGES-1 inhibition

Recent advances in mPGES-1 inhibitor characterization and pre-clinical studies demonstrating the anti-inflammatory effects of mPGES-1 inhibition are reviewed by Bergqvist et al. Detailed descriptions of the following examples can be found in [22]. Several mPGES-1 inhibitors (CII, CIII, compound 4b, compound A) have been shown to reduce pouch exudate PGE₂ and kidney PGE₂ in the air-pouch model of inflammation and to have an anti-pyretic effect in LPS-induced fever [23–26]. The mPGES-1 inhibitor CIII protected the liver and accelerated liver repair compared with vehicle controls in a mouse model of hepatic ischemia reperfusion [27]. The mPGES-1 inhibitor compound 4b was compared with celecoxib during a 14-day observation period regarding its effects on the stomach of mice. Celecoxib caused bleeding ulcers at 50 mg/kg. No adverse effects on the stomach of the mice was observed for the mPGES-1 inhibitor [25]. Inhibition of mPGES-1 has also been reported to reduce tumor growth and block PGE₂ formation by cancer-associated fibroblasts in a xenograft neuroblastoma mouse model [28]. The cross-species inhibitors 934, 117, 118, 323 and 323 demonstrated improved selectivity and potency in various pre-clinical assays [29]. The inhibitors blocked human and rat mPGES-1 reversibly and presented bioavailabilities between 10–46%. Besides weak to moderate inhibition of mPGES-2 by all compounds and weak inhibition of COX-2 by 323 no other off target activation was reported. The inhibitors reduced PGE₂ levels in A549 cells and whole blood as well as in an air-pouch mouse model. Moreover, the inhibitors reduced paw-swelling in rats and constriction in human resistance arteries ex-vivo [29,30]. The selective human mPGES-1 inhibitor Vipoglanstad (GS-248) reduced acute lung injury and sepsis in a transgenic mouse model [31]. Treatment with Vipoglanstat ameliorated leukocyte infiltration in LPS-induced lung injury and prolonged survival of mice with severe sepsis.

Few mPGES-1 inhibitors have been studied in clinical trials (2013, LY3023703 [32]; 2015, GRC-27864 [33]; 2019, Vipoglanstat [34]; 2022, NS-580). Taken together, the inhibitors decreased LPS-induced PGE₂ and increased PGI₂ production in blood and/or urine of healthy subjects. The trial for LY3023703 was terminated due to experimentally assessed off-target compound specific liver toxicity observed in two study participants [35]. GRC-27864 was well tolerated by healthy volunteers up to a dose of 1000 mg. Vipoglanstat was recently evaluated in a phase II study (ClinicalTrials.gov Identifier, NCT number: NCT04744207). Vipoglanstat was well tolerated and potently inhibited mPGES-1, but no significant effects were observed on Raynaud’s syndrome symptoms in patients with systemic sclerosis. A phase II investigation of Vipoglanstat in endometriosis is currently being prepared (https://www.gesynta.se/news-media). The mPGES-1 inhibitor NS-580 is currently tested in a phase-II trial for the treatment of endometriosis and chronic pelvic pain syndrome https://rctportal.niph.go.jp/en/detail?trial_id=jRCT2031210685.

2.2. Shifting prostanoid balance for potential therapeutic benefits

The feared cardiovascular side effects of NSAIDs are commonly attributed to an imbalance between pro-thrombotic and anti-thrombotic mediators that can lead to thrombotic events, elevated arterial blood pressure, and heart failure. Inhibition of mPGES-1 has been hypothesized to cause similar adverse effects. However, recent studies have indicated a potentially anti-inflammatory and protective cardiovascular profile of mPGES-1 inhibition through a reduction in PGE₂ and a shift toward PGI₂. Blocking a terminal prostanoid synthase could redirect the substrate PGH₂ toward other prostanoid synthases. The upregulation or maintenance of prostanoid levels (PGD₂, PGF_2α, PGI₂, TXA₂) upon mPGES-1 inhibition largely depends on the tissue-specific expression of prostanoid synthases in pathological conditions [36]. This shunting phenomenon following mPGES-1 inhibition thereby introduces some uncertainty, raising concerns about potential side effects. However, it equally holds expectations of additional beneficial effects. Tissue- and disease-specific prostanoid shunting upon blockage of mPGES-1 has been observed in various in-vitro and in-vivo assays.

Our research group has recently identified differences in the prostanoid profile of several immune cells, with profound variations in PGH₂ redirection after mPGES-1 inhibition [37,38]. Inhibition of mPGES-1 in RAW264.7 cells and mouse bone marrow-derived macrophages (BMDMs) led to shunting toward the PGD₂ pathway including pro-resolving 15dPGJ₂ and its metabolites [37,38]. In BMDMs, TXB₂ and PGF_2α were also increased [37,38]. No significant shunting was observed in A549 cells treated with an mPGES-1 inhibitor [37,38]. Enhanced PGI₂ production was observed in rheumatoid arthritis synovial fibroblasts (RASFs) treated with an mPGES-1 inhibitor [37,38]. Furthermore, in human arteries stimulated by inflammatory signals ex-vivo, we observed that inhibiting mPGES-1 significantly reduced PGE₂ levels while leaving PGI₂ levels unchanged or even increased [30]. Conversely, COX-2 inhibition blocked the production of both prostaglandins [30]. Other studies have also reported a notable increase in PGI₂ following mPGES-1 inhibition in internal mammary arteries and saphenous veins [39]. Moreover, investigations involving mPGES-1 knockout mice and mPGES-1 inhibitors have revealed effects on lipid mediators beyond prostanoids [40,41].

In an angiotensin-II induced aortic aneurism mouse model, the redirection of PGH₂ to PGD₂ and PGI₂ in urine was associated with a reduction in oxidative stress and a protection from aortic aneurism formation [42]. PGD₂ was also increased in bone marrow derived dendritic cells from mPGES-1 KO mice [43], and the PGD₂ metabolite 15dPGJ₂ was increased in LPS treated mouse peritoneal macrophages lacking mPGES-1 [40]. Mice deficient in mPGES-1 exhibit decreased urinary levels of PGE₂, unaltered TXA₂ levels, and increased urinary production of PGI₂ [44]. In contrast to COX-2 inhibition or deletion, mPGES-1 deletion was found to protect against thrombogenesis and did not cause an increase in blood pressure in mice [44]. Treatment of mPGES-1 KO mice with LPS resulted in reduced platelet activation, aggregation, and release of platelet-derived microparticles compared to WT controls [45]. Depletion and inhibition of mPGES-1 has been found to be protective in models of inflammatory pain, despite enhanced production of the pain mediator PGI₂ [46,47]. In rats with post operative pain the mPGES-1 inhibitor UK4b showed equal efficacy compared with opioid treatment [48]. The pain attenuation was associated with reduced PGE₂ formation without changing the levels of other prostanoids. This suggests that inhibition of mPGES-1 could be a therapeutic strategy to reduce inflammatory pain.

Ex-vivo inhibition of mPGES-1 with CIII in larger human vessels led to decreased vascular tone via shunting to PGI₂ [39]. These results, indicate important vasoactive properties of mPGES-1 inhibitors. We have recently shown that blocking of mPGES-1 in resistance arteries from end-stage kidney disease (ESKD) patients and Non-ESKD controls significantly reduced adrenergic vasoconstriction ex-vivo. In opposite the COX-1/COX-2 inhibitor indomethacin as well as COX-2 selective inhibitors NS-398 and etoricoxib had no effect on NE-induced vasoconstriction [29,30]. In experiments with larger human arteries (saphenous vein and internal mammary artery) and intrarenal arcuate arteries from mice, inhibition of COX-2 caused enhanced constriction [39,49]. Furthermore, mPGES-1 inhibition in resistance arteries led to a significant increase in acetylcholine-induced dilation, suggesting that it may improve blood flow and reduce resistance in-vivo. Blocking PGI₂ signaling with an IP receptor antagonist did not restore reduced adrenergic constriction in human resistance arteries, nor did blocking PGE₂-EP4 or signaling through PPARγ. Exogenous PGE₂ demonstrated a biphasic effect, inducing dilation at nanomolar concentrations and constriction at micromolar concentrations [30]. Whether COX-1 or COX-2 is the primary source of PGI₂ in the vasculature is controversial and studies show conflicting results. Deletion of COX-2 in endothelial cells and VSMC has been shown to reduce systemic PGI₂ [50,51]. On the other hand, a dominant role of COX-1 in PGI₂ production in mouse aorta has been demonstrated [52]. Our laboratory’s immunohistochemistry results showed expression of mPGES-1, COX-1, PGIS, weak expression for COX-2, and receptor expression for PGE₂ (EP1–4), thromboxane (TP), and PGI₂ (IP) in resistance arteries, supporting the preferential role of COX-1 for PGI₂ production [53]. The overall results of ex-vivo vascular studies suggest that inhibition of mPGES-1 has beneficial vasoactive effects, including the promotion of PGI₂ signaling and possibly other pathways, such as local reduction of PGE₂ and creation of an antioxidant milieu that maintains formation of NO. In human increased PGI₂ production in blood and/or urine was observed in phase I or II clinical trials for the mPGES-1 inhibitors LY3023703 [32], GRC-27864 [33] and Vipoglanstat [34].

2.3. Inhibition of mPGES-1 in models of cardiovascular disease

Several models of cardiovascular disease have demonstrated the beneficial effects of blocking mPGES-1. Besides for example reduction of aneurism formation in hyperlipidemic mice [42], mPGES-1 deletion reduced VSMC proliferation and hyperplasia after vascular injury in mice with femoral artery injury [54]. Endothelial cells lacking the low-density lipoprotein receptor showed increased PGIS and TXS expression after mPGES-1 deletion, and LPS stimulation of mPGES-1-deficient VSMCs and macrophages resulted in a shift toward PGI₂ and TXA₂ production [55]. Inhibiting mPGES-1 could also protect the heart by sparing the endothelial nitric oxide synthase inhibitor ADMA pathway, as PGI₂ and PGE₂ can regulate ADMA levels via IP, PPARβ/δ, and EP4 receptors. Compared to COX-2 inhibition, which increased the formation of ADMA in plasma, mPGES-1 KO mice showed increased plasma levels of PGI₂ and no effect on ADMA pathway genes [56].

In mice with experimental MI following left coronary artery ligation, global deletion of mPGES-1 did not alter the post-MI cardiac dysfunction observed in wild type animals [57]. However, myeloid-specific deletion of mPGES-1 improved survival after MI without adverse cardiac remodeling [57]. Deletion of mPGES-1 did not increase ischemic myocardial injury 24 hours after coronary artery ligation in mice, unlike wild type mice treated with celecoxib [58]. On the other hand, global deletion of mPGES-1 or its deletion in bone marrow-derived myeloid cells impaired cardiac remodeling after MI in other studies [59,60]. Results from studies using mPGES-1 or COX KO mice and COX inhibitors are conflicting, indicating a need for a better understanding of the link between the prostanoid system and cardiac remodeling after MI.

Several studies suggest a protective role of PGE₂ signaling in the adverse effects seen with COX-2 and mPGES-1 depletion in various experimental models, including MI. The PGE₂-EP3 pathway has been shown to be important for cardiac recovery after MI in mice [61]. Another study described the protective effect of mPGES-1 through PGE₂-EP4 signaling in mice with acute myocardial ischemia-reperfusion injury [62]. A summary of the role and potential of PGE₂ in MI/IR injury can be found here [63].

Recent research from our group has shown that chronic administration of the mPGES-1 inhibitor CIII and 118 improved cardiac function in mice after MI compared to vehicle or celecoxib [64,65]. CIII reduced infarct area, increased scar thickness, decreased collagen I/III ratio, decreased fibrosis-related gene expression, and increased capillary density in the ischemic area. Shunting to urinary metabolites of PGI₂ after inhibition of mPGES-1 was positively correlated with cardiac function after MI. The PGI₂/PGE₂ urinary metabolite ratio correlated positively with ejection fraction, fractional shortening, and scar thickness. These findings highlight fundamental differences between experimental enzyme deletion and inhibition and suggest that pharmacological inhibition of mPGES-1 could improve the outcome after MI. Future studies are warranted to validate and expand our understanding of the underlying mechanisms.

3. Expert opinion

NSAIDs are among the most used and recognized medications worldwide. However serious gastrointestinal and especially cardiovascular side-effects associated with their use have led to a number of regulatory actions including the withdrawal of several NSAIDs (e.g. Vioxx in 2004), the introduction of ‘black box’ warnings and the reclassification of the over-the-counter medication diclofenac as prescription only in 2015 (UK). Fears of cardiovascular events caused by NSAIDs have become a public health issue and have generated a ‘guilt by association’ for other anti-inflammatory strategies that target prostanoid signaling pathways, such as inhibition of mPGES-1, which selectively targets PGE₂ synthesis downstream of the COX enzymes. However, inhibition of mPGES-1 appears to rebalance the prostanoid profile and reduce the cardiovascular side effects seen with NSAIDs, and is therefore likely superior to COX inhibition, and EP receptor targeting. Another obstacle to the development of mPGES-1 inhibitors has been the fact that inhibitors targeting the human mPGES-1 show little or no effects in rodent models. Only recently cross-species inhibitors have been developed allowing in-depth pre-clinical characterizations. In opposite to the ‘guilt by association,’ recent studies suggest a cardiovascular protective profile of mPGES-1 inhibition through a shift toward PGI₂. As has been shown, the redirection of PGH₂ to other prostanoids, such as cardioprotective PGI₂ or anti-inflammatory PGD₂ depends on the cellular composition of the tissue and disease-causing mechanisms, so studies of the larger lipid mediator network are needed for consequence estimation and optimal treatment. Preclinical studies show favorable results with mPGES-1 inhibitors in relieving inflammation and pain, improving vascular function by reducing adrenergic vasoconstriction, protecting against thrombogenesis and protecting against consequences of myocardial infarction. Selective reduction of PGE₂, redirection of PGH₂ to PGD₂ and PGI₂ and maintenance of the NO pathway are thought to contribute to the anti-inflammatory and cardioprotective effects observed after inhibition of mPGES-1.

The encouraging results of the use of mPGES-1 inhibitors in a rodent model of myocardial infarction still require further investigation of the underlying mechanisms. These include ablation of prostanoid receptors to identify activated signaling pathways, measurements of plasma ADMA levels at different timepoints as well as the optimal timing of intervention (prophylactic, during acute MI, or post MI). In most of the studies presented, treatment and disease induction collided and a better understanding of whether the inhibitor has similar or improved/worsened effects when used at different phases of the disease model studied could increase clinical relevance.

In conclusion, there is a great need for clinical research of mPGES-1 inhibitors in chronic inflammatory diseases such as rheumatoid arthritis, spondyloarthritis, and osteoarthritis, as well as diseases with a strong component of vasculopathy and in cancer. The cardioprotective effects observed in experimental models warrant phase II trials in MI, possibly also as prophylactic treatment in ‘at risk’ individuals for cardiovascular disease such as those with hypertension, diabetes mellitus or metabolic syndrome. It is encouraging that phase II studies are ongoing and planned in endometriosis, a complex female disease with components of hormonal regulation, angiogenesis, cell proliferation, inflammation and pain. The authors believe that additional mechanistic studies of mPGES-1 inhibitors are unlikely to further predict outcome in humans. Therefore, multiple phase II studies should be conducted, possibly in an academic setting led by leading scientists in the respective fields, which usually reduce study costs. There are several examples where the indication of a developed drug was planned or predicted for one disease but ultimately proved to be better in another disease. Secukinumab is one such interesting example, which went from an indication for rheumatoid arthritis to psoriasis arthritis and spondyloarthritis.

Overall, the results of recent pre-clinical and clinical investigations suggest that inhibition of mPGES-1 is a promising anti-inflammatory therapy with the potential to prevent and ameliorate inflammation and pain, cancer, cardiovascular diseases and other inflammatory conditions with cardiovascular comorbidities.

Abbreviations

ADMA	=	asymmetric dimethyl arginine
COX	=	cyclooxygenase
EC	=	endothelial cell
ESKD	=	end-stage kidney disease
LPS	=	lipopolysaccharide
MI	=	myocardial infraction
MI/IR	=	myocardial infarction ischemia reperfusion
mPGES-1	=	microsomal prostaglandin E synthase 1
NO	=	nitric oxide
Non-ESKD	=	non end-stage kidney disease
NSAIDs	=	non-steroidal anti-inflammatory drugs
PPAR	=	peroxisome proliferator activated receptor
PGI2	=	prostaglandin I 2 (prostacyclin)
PGE2	=	prostaglandin E 2
TXB2	=	thromboxane B 2
VSMC	=	vascular smooth muscle cells

Article highlights

• Inhibition of mPGES-1 affects PGE₂ and other prostanoids specific for the cell type and disease condition.

• Inhibition of mPGES-1 is anti-inflammatory and vasoactive.

• Inhibition of mPGES-1 may activate anti-inflammatory and cardioprotective pathways encouraging implications for the treatment of inflammation without cardiovascular side effects.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Acknowledgments

This review is partly based on the PhD thesis of Julia Steinmetz-Späh.

Additional information

Funding

This manuscript was funded by the Swedish Rheumatism Association (grant no: R-755861, R-931761), King Gustaf V’s 80 Years Foundation, The European Union’s Horizon 2020 research project ArthritisHeal under the Marie Skłodowska-Curie grant agreement (No 812890), The Swedish Cancer Society (grant no: 19-0183), The Cancer Society in Stockholm (grant no: 171073), The Swedish research council (2021-03420).