EP2 Antagonists (2011-2021): A Decade's Journey from Discovery to Therapeutics.

Nonsteroidal anti-inflammatory drugs (NSAIDs) such as aspirin, ibuprofen, and naproxen are nonspecific inhibitors of both COX isoenzymes ( Figure 1A ). These drugs are widely used to reduce pain, fever, and inflammation, but their mechanisms of action were not fully elucidated until the early 1970s. 28 NSAIDs act by inhibiting the production of prostanoids, which may initiate signaling pathways that are essential to some important physiological functions in many organs and tissues under normal conditions; thus, their adverse effects can be widespread. Depending on the specific drug and dose used, NSAIDs can damage the gastrointestinal tract and lead to kidney, cardiovascular, and liver dysfunctions, which are believed to be caused by inhibition of COX-1. 29 In the hopes of mitigating the side effects associated with a nonselective COX inhibition, focus has been shifted to COX-2 for more specificity. The COX-2 enzyme was originally discovered and sequenced in 1988 and then confirmed in 1991, 30 fostering the development of selective small-molecule inhibitors for COX-2 (coxibs) ( Figure 1A ). Less than eight years later, the first coxibs were developed and then quickly introduced to the market, with Vioxx (rofecoxib) and Celebrex (celecoxib) launched in 1999. These selective COX-2 inhibitors are very effective to alleviate pain and other inflammation-related complications without causing major gastrointestinal ulceration and bleeding or other nonselective COX inhibitors-associated side effects and, thus, were thought safer than the conventional NSAIDs. 31 Unfortunately, it soon became evident that coxibs were associated with severe cardiac side effects in patients taking them as long-term medication, 32 – 34 suggesting that a certain COX downstream prostanoid signaling pathway might function to protect the microvascular systems. 35 Targeting the inducible mPGES-1 to inhibit the PGE 2 synthesis from COX-2-derived PGH 2 without affecting other types of prostanoids was thought more specific than inhibiting COX-2 itself. However, designing and testing compounds to inhibit this enzyme have proved more convoluted than previously thought due to the interspecies differences in the sequence and structure of the mPGES-1 enzyme. 36 , 37 Thus, to circumvent these complications derived from inhibiting biosynthetic enzymes for PGE 2 , the downstream PGE 2 receptor EP2 has been proposed as an alternative and hopefully more specific target owing to its leading role in the COX-2-mediated inflammatory signaling cascade. 16

Genetic ablation of EP2 (EP2 −/− ) was an inhibitory strategy available in early investigations to elucidate the physiological and pathological roles of the EP2 receptor. Though very useful, it was also complicated by the developmental and other homeostatic adjustments in mice that resulted in hypertension and reduced litter size. 58 – 60 In the meantime, small-molecule agonists of the EP2 receptor that have been widely used include butaprost, CP-533536, CAY10399 , ONO-AE1–259, and C-9 ( Figure 2 ). 16 , 61 , 62 However, butaprost is only ~18-fold selective for EP2 over EP3; 63 CP-533536 (evatanepag) is ~64-fold selective over the EP4 receptor; 64 CAY10399 and ONO-AE1–259 are highly selective on EP2 but have prostanoid-like structures; C-9 is quite selective against other PGE 2 receptors but less than fourfold selective over the TP receptor. 65 Positive allosteric modulators for EP2 with nonprostanoid structures have been reported and provide alternative chemical probes to study the receptor in the presence of PGE 2 ; 66 , 67 however, the uncertainty of their pharmacokinetic properties impedes their uses in vivo. A critical reason to explain the frustration in the early investigations of the EP2 receptor in health and disease was the lack of a suitable EP2 receptor antagonist. 16 Among all nine currently known prostanoid receptors, EP2 was the last one that did not have a truly selective small-molecule antagonist until 2011. Although the underlying reason is unclear, a recent study on the cryo-electron microscopy structure of the EP2-G s complex revealed its unconventional conformation in the active state when compared to those of all currently known active type 1 GPCRs. 68 Prior to 2011, a commonly used antagonist to study the EP2 receptor was AH-6809 ( Figure 2 ). Unfortunately, AH-6809 is neither selective nor potent—it weakly inhibits both EP1 and EP2 without palpable preference—and is unsuitable for in vivo studies. 63 To overcome these limitations, researchers from Pfizer, Emory, and Amgen independently developed the first truly selective EP2 antagonists with well-characterized pharmacokinetic and pharmacodynamic profiles that are suitable for both in vitro and in vivo testing. 69 – 71 The rest of this review focuses on the discovery and optimization of these EP2 antagonists, followed by their evaluations in various preclinical models of peripheral and central inflammation-associated conditions. In 2011, Pfizer reported its first EP2-selective antagonist, compound 1 ( Figure 3 ), which showed promise in both in vitro and in vivo assays. 69 Compound 1 did not show any agonistic activity, nor did it exert any effects on EP2 without the presence of PGE 2 or another agonist. These results indicate that it specifically counteracts the effects of PGE 2 or other agonists acting on the EP2 receptor. 69 , 72 Compound 1 also displayed promising pharmacokinetic profiles and good tissue-penetrating properties and, thus, allowed studies on EP2 antagonism in an in vivo setting. Moreover, it was able to induce a relaxation of mouse trachea, indicating its therapeutic potential in asthmatic diseases. Two years later, Pfizer reported its second EP2-selective antagonist, compound 2 ( Figure 3 ), which was found ~10-fold more potent than the previously reported compound 1 for the antagonism of PGE 2 (functional EP2 K i : 0.63 vs 7.6 nM). 73 Just like compound 1 , the newly reported compound showed good selectivity on EP2 over other G s -coupled prostanoid receptors. Both Pfizer compounds, but not EP4-selective antagonist CJ-042794, powerfully antagonized the PGE 2 -provoked cAMP signaling and the consequent inhibition of IgE-dependent histamine release in human lung mast cells in a competitive mechanism, although both EP2 and EP4 receptors were expressed in these cells. 73 These interesting findings suggest that the principal mechanism whereby PGE 2 inhibits a mediator release in lung mast cells is mainly through activation of the EP2 receptor, and thus, these EP2 antagonists likely have therapeutic potential to treat asthma and other respiratory diseases. The brain penetration of these two compounds has not been reported by Pfizer; however, a recent study revealed that compound 1 was unable to cross the blood-brain barrier, 26 suggesting that its action is only limited in the periphery. Currently, there is no study on the structure–function relationships (SAR) of these Pfizer compounds, nor are there any new EP2 antagonists reported by Pfizer. Their current status for further development is unknown. In 2008, utilizing a set of cell-based time-resolved fluorescence resonance energy transfer (TR-FRET) assays of cAMP formation, researchers at Emory University performed a high-throughput screening (HTS) of 262 371 compounds ( https://pubchem.ncbi.nlm.nih.gov/bioassay/1422 ). Because of the lack of a truly selective EP2 antagonist back then, their goal was to develop compounds that selectively inhibit the EP2 receptor and thus can be used to study a prolonged seizure-induced brain inflammation. Consequently, a series of small molecules were identified as competitive antagonists of the human EP2 receptor. Among these, compound 3 ( Figure 4A ) is the most potent compound with a functional Schild K B of 2.4 nM for the antagonism of PGE 2 and showed a plasma half-life ( t 1/2 ) of 0.6 h and a brain-to-plasma ratio of 0.3. 70 They then performed the first SAR study based on the 3-aryl-acrylamide scaffold in compound 3 and disclosed that an indole moiety, ethylene linker, acrylamide, and methoxyphenyl or halogenphenyl promoted activities in the nanomolar level. Meanwhile, the introduction of fluorine atoms into the compound 3 scaffold improved the metabolic stability but slightly decreased the potency of the compounds, leading to the discovery of compound 4 ( Figure 4A ). 14 , 70 , 74 – 77 Compared to compound 3 , compound 4 is less potent with a functional Schild K B of 17.8 nM on human EP2 but has an acceptable extended plasma half-life ( t 1/2 ) of 1.6 h and a favorable brain-to-plasma ratio of 1.6. In addition, both compounds 3 and 4 demonstrated a similar potency on human, mouse, and rat EP2 receptors, 14 , 70 justifying their use in both animal models and human conditions. 78 Off-target profiling of compounds 3 and 4 displayed negligible effects on a panel of more than 40 critical ion channels, enzymes, receptors, and neurotransmitter transporters, 14 , 15 attesting to their safety for in vivo uses. Additional optimization and SAR studies on the core structure of compounds 3 and 4 led to the development of second-generation EP2 antagonists compound 5 and compound 6 , in which the indol-1-yl is replaced with an indol-3-yl and the acrylamide linker is replaced with an amide ( Figure 4B ). 79 , 80 These modifications enable compounds 5 and 6 to maintain the nanomolar efficacy and enhance the selectivity to EP2 receptor over DP1. In addition, an introduction of nitrogen-containing heterocycles greatly improves their water solubility, yielding the water-soluble EP2 antagonists. Further SAR study reveals that the tetrazole ring is good for EP2 potency, solubility, and metabolic stability. 75 On the one hand, indeed, tetrazole ring-containing compound 5 has a functional EP2 K B of 48.5 nM and is highly soluble (500 μ M in water), metabolically stable in vivo (plasma half-life in mice: 10.5 h), and peripherally restricted (brain-to-plasma ratio: 0.02). 79 On the other hand, another SAR study on the middle phenyl ring reinforces the notion that nitrogen in the ring can enhance the aqueous solubility, leading to the discovery of compound 6 , which has a Schild K B of 9.7 nM on EP2, a water solubility of 2.5 mM, a brain-to-plasma ratio of 0.4, and a plasma half-life of 2.4 h in mice. 80 Overall, these two second-generation EP2 antagonists show much improved selectivity against the other three G α s -coupled prostanoid receptors (DP1, EP4, and IP) when compared to their precursors. Compound 7 was among the original hits from the Emory HTS and possesses a distinct chemical scaffold, a carbamothioylacrylamide backbone ( Figure 4C ). This compound has a functional EP2 K B of 8.8 nM, and with a systemic administration in mice it showed a plasma half-life of 2.7 h and a brain-to-plasma ratio of 0.02. 81 The low brain penetration of compounds 5 and 7 enables these two compounds to be ideal tools to study the functions of the EP2 receptor in peripheral diseases associated with chronic inflammation such as rheumatoid arthritis and chronic obstructive pulmonary disease, in which EP2 appears to play essential pathogenic roles. 82 – 84 Molecular docking using the recently solved cryo-electron microscopy (cryo-EM) structure of the human EP2 receptor and G s protein complex (PDB code: 7CX3) revealed the simulated interactions between the EP2 receptor and Emory compounds using Schrödinger software, 68 , 85 exemplified by compound 3 ( Figure 5 ). Understanding the dynamic three-dimensional (3-D) interactions between these competitive antagonists and the EP2-G s complex might help with better rational designs to develop the next-generation EP2 antagonists with more balanced potency and selectivity. In 2015, an HTS of a small-molecule library by researchers at Amgen led to the identification of compound 8 ( Figure 6 ) as a novel EP2 antagonist with a moderate to mild potency across the human, mouse, and rat receptors. In selectivity tests, it was more than 400-fold selective against the EP1 receptor, 300-fold selective against EP3, but only 50-fold selective over EP4, the other G α s -coupled receptor for PGE 2 . Compound 8 had a low metabolic stability in liver microsomes and significantly inhibited both CYP3A4 and CYP2D6 and 71 raised concerns of drug–drug interactions because these two CYP enzymes metabolize more than 50% of clinically important drugs. Compound 8 then underwent several lead-optimization steps to improve its metabolic stability, reduce its inhibition on CYP3A4 and CYP2D6, and increase the selectivity against the EP3 and EP4 receptors. Subsequent SAR studies revealed that the replacements for the benzothiophene ring, pyridine ring, and methyl ether of compound 8 could address all these liabilities ( Figure 6 ) and, eventually, resulted in the discovery of compound 9 , which instead has a fluoroindole ring, pyridone ring, and chlorobenzoxazepine ( Figure 6 ). These changes afforded compound 9 a single-digit nanomolar binding affinity for the EP2 receptor and significantly improved microsomal stability. Importantly, compound 9 demonstrated a comparable potency across the human, mouse, and rat EP2 receptors, was greater than 660-fold selective in the EP2 cAMP assay against DP1, EP4, and IP receptors, and was devoid of CYP inhibition. 71 With an oral dose, compound 9 had a plasma half-life of 3.4 h in CD-1 mice and brain-to-plasma ratios of 0.7 and 0.9 in C57BL/6 mice and Sprague–Dawley rats, respectively. 71 The favorable in vivo half-life and high brain penetration as well as a capability of increasing the macrophage-mediated clearance of amyloid-beta plaques ex vivo raise the therapeutic potential of compound 9 for Alzheimer’s disease (AD) and other inflammation-associated CNS conditions. 26 , 71

The prostaglandin receptor EP2 is coupled to a heterotrimeric G s protein complex consisting of α, β , and γ subunits. The activation of EP2 receptor by PGE 2 quickly separates the G s complex into G α and G βγ , which in turn regulate diverse downstream signaling molecules that also can crosstalk with many other pathways ( Figure 1B ). 11 First and foremost, G α s activates adenyl cyclase (AC) to increase cellular levels of cAMP, thereby activating protein kinase A (PKA) and exchange factor directly activated by cAMP (EPAC). On the one hand, the cAMP-stimulated EPAC then, in turn, can activate its downstream effectors Rap GTP-binding proteins 1/2 (RAP1/2) to regulate many pathogenic events such as neuroinflammation and neurotoxicity ( Figure 1A ). 10 , 38 , 39 On the other hand, the phosphorylated PKA acts on the transcription factor cAMP response element-binding protein (CREB), which then translocates to the nucleus, where it can regulate the transcription of responding genes. 16 Additionally, EP2 receptor, through the dissociated G βγ subunits upon PGE 2 binding, activates the phosphoinositide 3-kinase (PI3K)/protein kinase B (PKB or Akt) pathway, which causes the phosphorylation and inactivation of glycogen synthase kinase 3 β (GSK-3 β ), eventually leading to the stabilization and nuclear translocation of β -catenin and the expression of growth-promoting genes and inflammation. 40 – 42 Activated PI3K/Akt may also phosphorylate the transcription factor nuclear factor κ B (NF- κ B), which then translocates to the nucleus to regulate the transcription of a wide variety of genes involved in inflammation. 43 , 44 Interestingly, PGE 2 /EP2 signaling can transactivate the epidermal growth factor receptor (EGFR) likely via PKA and Src, leading to the activation of the Ras/Raf/mitogen-activated protein kinase kinase (MEK)/extracellular signal-regulated kinase (ERK) pathway and cell proliferation and invasion. 44 – 46 The EP2 receptor also mediates a G protein-independent pathway via association with β -arrestin, 47 resulting in the activation of a c-Jun N-terminal kinase (JNK)/profilin 1 (Pfn-1)/filamentous actin (F-actin) pathway for cell migration and proliferation. 48 In addition, the transactivation of EGFR by PGE 2 /EP2 is likely mediated by β -arrestin, which can phosphorylate and activate Src ( Figure 1B ). 49 Unlike other prostanoid receptors (e.g., EP4), the EP2 receptor usually does not undergo internalization and thus resists desensitization even upon a repeated stimulation by PGE 2 , 2 , 50 allowing for a continuous receptor signaling and drug effects. The EP2 receptor is widely expressed in the human body and has been primarily detected in the central nervous system (CNS), bones, reproductive system, leukocytes, and smooth muscle, 51 , 52 where it regulates a variety of physiological functions, such as immunoregulation, ovulation and fertilization, vasodilation, osteoclasts, regulation of intraocular pressure, neuronal plasticity, learning, and memory. However, the deregulated PGE 2 signaling via the EP2 receptor may contribute to the molecular and cellular mechanisms of a number of inflammation-associated pathogenic processes, such as tumorigenesis, oxidative stress, amyloid β (A β ) accumulation, neurotoxicity, neurodegeneration, and neuroinflammation. 16 , 17 , 19 , 38 , 53 – 57 The broad pathophysiological roles of the excessive PGE 2 /EP2 signaling taken together lend support to a translational strategy suppressing the EP2 receptor to harness the uncontrolled inflammation in these disease states.

COX-2 is commonly expressed in a variety of types of cancers and has been well-known to promote inflammatory microenvironments, angiogenesis, immune evasion, and treatment resistance. 9 , 159 – 161 Most of these pro-tumor effects of COX-2 are believed to be largely attributed to the PGE 2 signaling mediated by the EP2 receptor, because the genetic deletion of EP2 led to the suppression of tumor development and the progression of various backgrounds in early studies. 15 , 44 , 162 – 168 However, considering other significant complications in EP2 −/− mice, the results from the EP2 ablation were unable to be validated by a reliable pharmacological approach due to the lack of truly selective EP2 antagonists with adequate pharmacokinetic and pharmacodynamic profiles. The discovery of brain-permeable and brain-impermeable EP2 antagonists enable pharmacological strategies to study multifaceted roles of EP2 in tumor development and progression, chronic inflammation, metastasis, angiogenesis, and multidrug resistance in animal models of cancers of both the periphery and brain. As the third most prevalent cancer worldwide, colon cancer constitutes the fourth most common cause of cancer-related death. 169 It has long been observed that a regular consumption of aspirin or other NSAIDs is highly correlated with a reduction in the mortality rates of sporadic colorectal cancer patients. In addition, NSAIDs have been shown to facilitate the self-regression of familial colon polyposis, a precancerous condition. 170 – 172 Results from these epidemiological analyses suggest that the pathogenesis of colon cancer is highly associated with inflammatory responses in the colon, and thus suppressing tumor inflammation might prevent the development and progression of colon cancer. Although the mechanism whereby COX cascade-mediated inflammation promotes colon cancer largely remains elusive, the downstream PGE 2 /EP2 signaling pathway has been thought to contribute to tumor formation and progression. 168 , 173 , 174 Interestingly, a deficiency of the EP2 receptor (EP2 −/− ), but not EP1 or EP3, selectively decreased tumor formation in azoxymethane (AOM)/dextran sodium sulfate (DSS)-induced colon tumorigenesis, a colitis-associated cancer model, accompanied by downregulated pro-inflammatory genes such as TNF- α , IL-6, chemokine (C-X-C motif) ligand 1 (CXCL1), and COX-2. 175 These findings are consistent with results from early studies on mouse models of intestinal polyp formation, 162 mammary hyperplasia, 164 skin tumor, 165 and lung tumorigenesis, in which the genetic ablation of the EP2 receptor in mice relieved tumor-associated inflammation and suppressed tumor development. Importantly, in the colitis-associated colorectal cancer model, systemic administration of EP2 antagonist compound 1 daily for 80 d potently suppressed colon tumorigenesis induced by AOM/DSS in a dose-dependent manner ( Table 2 ). 175 Results from this study also suggest that EP2 in tumor-associated fibroblasts and neutrophils may facilitate colon tumorigenesis by amplifying the inflammation and shaping the tumor microenvironment. 168 , 175 Given that the PGE 2 /EP2 signaling likely serves as a key node of chronic inflammation in the colon tumor microenvironment, EP2 antagonists might represent promising candidates of NSAIDs-alternative for chemoprevention of colon cancer. Gliomas constitute ~80% of all primary malignant brain tumors in humans, and more than 80% of these cases are classified by the World Health Organization (WHO) as grade IV tumor–glioblastoma. With the current standard treatment, namely, surgical resection followed by concurrent radiotherapy and chemotherapy with Temozolomide, the prognosis of glioblastoma remains poor with a median overall survival below 15 months, and less than 10% of patients can survive more than five years. 176 – 178 Among several comprehensive factors that render malignant glioma particularly difficult to treat is that most antitumor agents including many immunotherapeutic drugs are unable to reach the tumor sites due to their poor brain penetration. 179 The development of new therapeutics with an adequate efficacy for this most devastating and deadly type of brain cancer is certainly an urgent unmet need. 180 , 181 Inflammation within the brain emerged as a key contributor to many forms of brain cancer. 182 COX-2, as a chief pro-inflammatory mediator, is often induced in intracranial tumors, 183 , 184 and it has been shown to promote the growth, migration, angiogenesis, and immune evasion of malignant gliomas. 185 – 187 It appears that COX-2/mPGES-1/PGE 2 signaling axis is highly associated with the aggressiveness of human gliomas, and the EP2 receptor is a key G α s -coupled receptor that mediates COX-2/PGE 2 -initiated cAMP signal pathways in human malignant glioma cells. 188 Interestingly, inhibition of the EP2 receptor by the antagonist compound 4 reduced the COX-2 activity-driven glioblastoma cell proliferation, invasion, and migration, and it caused cell-cycle arrest at G0-G1 and apoptosis of the brain tumor cells. 188 Moreover, oral administration of compound 4 twice daily for four consecutive weeks suppressed malignant glioma growth and increased survival rates of mice that harbored intracranial tumors formed by human glioblastoma cells ( Table 1 ). 188 These results together suggest that PGE 2 signaling via EP2 receptor contributes to the malignancy of human glioma cells and that the selective EP2 antagonists with adequate brain penetration such as compounds 3 and 4 might represent promising candidates as new therapeutic agents for malignant glioma. 189 Neuroblastoma is the third-most common type of pediatric cancers and accounts for nearly 15% of cancer-related deaths in young children. 160 Despite marked advances in tumor diagnosis and management during the past decades, the five-year survival rates for patients with high-risk neuroblastoma remain below 50%. 190 COX, a conventional inflammatory executor, has recently been demonstrated to provide the essential driving force for neuroblastoma pathogenesis. 191 PGE 2 highly presents in various tumor tissues including those of neuroblastoma, where the COX expression is elevated. 192 , 193 A daily oral intake of COX inhibitors decreased the burden of tumors with an amplification of the MYCN oncogene (encoding a transcription factor), 160 reduced the presence of tumor-associated innate immune cells, 194 and delayed the progression of 11q-deleted neuroblastoma. 193 Likewise, the inhibition of PGE 2 direct synthase mPGES-1 by Compound III (CIII) or 934 was found to suppress neuroblastoma with the MYCN amplification or 11q deletion, 195 , 196 attesting to an essential role for PGE 2 in the high-risk neuroblastoma. Interestingly, correlation analyses on neuroblastoma patient data sets revealed a positive relationship between the COX/PGE 2 /EP2 signaling axis and the aggressiveness of a human neuroblastoma. 197 A cell-based TR-FRET assay method identified EP2 as the primary G α s -coupled receptor that mediates PGE 2 -initiated cAMP signaling in neuroblastoma cells with various risk factors. 197 A preliminary study using compounds 4 and 7 demonstrated that pharmacological inhibition of the EP2 receptor in mice substantially impaired the growth of human neuroblastoma xenografts and the associated angiogenesis and downregulated pro-inflammatory cytokines in tumor tissues ( Table 1 ). 197 Collectively, these preliminary results suggest that PGE 2 via EP2 receptor increases the growth and malignant potential of human neuroblastoma cells; selective EP2 antagonists might provide novel alternative therapeutic strategies to COX or mPGES-1 inhibition for this devastating type of pediatric cancer.

Ischemic stroke accounts for ~87% of all stroke cases, but due to its complex pathophysiology and relatively narrow intervention window the current treatment remains mainly limited to intravenous thrombolysis. 126 COX-2 is highly regulated by neuronal activities and is often rapidly and robustly induced within the brain by cerebral ischemia. 127 , 128 A large number of early studies using various rodent stroke models have shown that the induced COX-2 by cerebral ischemia might contribute to the neuronal injury, whereas the administration of selective COX-2 inhibitors or a genetic ablation of COX-2 reduced infarct volumes. 127 – 131 However, the therapeutical targeting of COX-2 has been increasingly dampened due to extensive complications of cerebrovascular and cardiovascular systems caused by chronic uses of COX-2 inhibitors as medication. Strategies therapeutically targeting the downstream prostanoid receptors that are responsible for the deleterious effects of the COX cascade have been proposed. 23 , 25 , 53 , 132 Several early studies utilizing global congenital EP2 knockout mice found that EP2 receptor activation might play some neuroprotective role following an ischemic stroke, because EP2 deficiency exacerbated the infarction after a middle cerebral artery occlusion (MCAO) with or without reperfusion. 133 – 135 However, in a recent animal study using more sophisticated and more specific knockout strategies, it was found that the postnatal deletion of EP2 or the conditional ablation of its neuronal form, but not the endothelial or myeloid form, decreased the cortical infarction and neurological deficit in mice after a 45 min episode of MCAO followed by reperfusion. 19 The contradiction between these early and recent studies can be explained by the broad complications and neurological dysfunctions that are caused by the global congenital deletion of EP2, as the receptor plays some essential functions in the brain, such as synaptic transmission, sensorimotor gating, synaptic plasticity, and cognitive functions. 8 , 124 , 125 The postnatal and conditional deletion of EP2 would preserve its normal neurophysiological functions during the early stages of brain development, and it thus would avoid the cognitive and behavioral impairments observed in the global congenital EP2 knockout mice. 19 As such, the conditional and postnatally induced deletion of EP2 provided more specific and reliable strategies to investigate the EP2 receptor-mediated neuroinflammation following ischemic stroke in the latest study. 19 Nonetheless, because of the significant contribution of PGE 2 /EP2 signaling to COX-2 cascade-mediated neuronal inflammation and injury in excitotoxic and ischemic injuries, 18 , 19 several EP2-selective small-molecule antagonists were evaluated for therapeutic potential in animal models of an ischemic stroke with the Amgen compound 9 being the first one ( Table 2 ). Systemic treatment with compound 9 at 4.5 h and again at 24 h after ischemia start improved the neurological scores as well as decreased the weight loss and infarct volumes in mice after an episode of 45 min of MCAO followed by 72 h of reperfusion. 19 Likewise, intraperitoneal injection of compound 4 with three doses that were administered at 4.5, 12, and 24 h following an MCAO for 45 min significantly improved the overall neurological score and histological lesion size in male mice ( Table 1 ). 21 Moreover, treatment with compound 4 was shown to downregulate the expression of prototypical inflammatory cytokines IL-1 β , IL-6, and TNF- α in injured sites after MCAO. Similar results derived from studies using two different EP2 antagonists suggest that the beneficial effects of these compounds are unlikely related to a specific chemical structure; they rather are caused by their inhibition on EP2 receptor, that is, the prospective mechanisms of drug action. These proof-of-concept preclinical studies consistently support EP2 antagonism by brain-permeable EP2 antagonists such as compounds 4 and 9 as an emerging strategy to treat an ischemic stroke.

Sepsis arises from the dysfunction of the regulatory mechanisms that govern the immune system during the response to various infections. Notably, almost 70% of patients with sepsis will develop sepsis-associated encephalopathy, which induces neuroinflammation, a key feature associated with long-term cognitive deficits. The COX-2/PGE 2 /EP2 signaling pathway has been implicated in playing a pivotal role in neuroinflammation through the elevation of pro-inflammatory cytokines, chemokines, and COX-2 expression within the brain. The systemic administration of the selective EP2 antagonist compound 4 after LPS-triggered systemic inflammation facilitated the recovery of body weight, mitigated brain inflammation as evaluated by pro-inflammatory mediators such as IL-1 β , IL-6, TNF- α , chemokine (C–C motif) ligand 2 (CCL2), and COX-2 as well as microgliosis, and prevented the loss of synaptic proteins such as the postsynaptic density protein 95 (PSD-95) and synaptophysin. In a panel of behavioral tests that were performed approximately one month after an LPS injection in mice, compound 4 treatment ameliorated the depression-like behavior in a sucrose preference test and cognitive deficits in a novel object recognition test ( Table 1 ). 158 These results recognize EP2 receptor inhibition as a new strategy to alleviate neuroinflammation and long-term affective and cognitive issues of sepsis survivors.

A prolonged convulsive seizure or status epilepticus (SE) is a common life-threatening condition that can cause significant morbidity and mortality. Resulting from the failure of the mechanisms in charge of seizure termination, SE is traditionally defined as a continuous seizure lasting longer than 30 min or two or more separate seizures without a complete recovery of consciousness between any of them. 86 Despite the introduction of a newer generation of antiseizure drugs (ASDs) during the past few decades, SE remains the second most common neurological emergency only after acute stroke. With more than 20% mortality across various populations, SE accounts for hundreds to thousands of annual deaths in the United States (U.S.) alone. 87 SE in humans triggers a cascade of molecular, cellular, and systemic events in the brain that can eventually aggravate the occurrence of unprovoked seizures—that is, development of epilepsy—in survivors. 88 Prolonged seizures induced by proconvulsive agents such as kainic acid, pilocarpine, and organophosphate in experimental rodents trigger similar reactions within brain tissue, such as pro-inflammatory processes engaging cytokines, reactive gliosis, blood-brain barrier breakdown, neuronal death, and changes in the synaptic efficacy. 89 – 91 Given their recapitulation of many important features of human SE conditions, such as mortality, behavioral abnormality, and the development of life-long epilepsy, 91 these animal models have been commonly used in preclinical studies to evaluate COX inhibitors, EP2-selective antagonists, and other anti-inflammatory compounds for therapeutic benefits, 92 as both COX-2 and EP2 receptor in the brain are rapidly induced by SE. 18 , 93 , 94 As a commonly used proconvulsive agent, pilocarpine with systemic administration can quickly trigger a continuous seizure-like behavior in mice and rats, which closely simulates the SE observed in human. Without intervention, a pilocarpine-induced SE usually lasts for hours, making it a suitable model to study SE-related neuronal injury, oxidative stress, and inflammation in the brain. 95 – 97 As the first EP2-selective antagonist that was ever tested in an animal seizure model ( Table 1 ), compound 3 dramatically reduced a hippocampal injury when administered in mice beginning 1 h after the termination of a pilocarpine-induced SE. 70 The neuroprotective effect of compound 3 was extended by broader beneficial effects of its analogue compound 4 in the same mouse model of SE ( Table 1 ), where compound 4 reduced the mortality rate of pilocarpine-treated mice by 30% at the end of the first week and by 35% after two months. 14 , 93 Moreover, the neuroprotective effect of compound 4 was accompanied by a substantial downregulation of pro-inflammatory mediators, such as oxidative stress enzymes, cytokines, and chemokines, in addition to blocking the feedback amplification of the COX-2 cascade upon EP2 activation. 14 , 22 , 93 , 94 Taken together, these findings demonstrated that the EP2 inhibition completely recapitulated the multiple benefits of a conditional deletion of COX-2 from a restricted population of forebrain neurons in a mouse pilocarpine model 98 and suggested that COX-2-mediated pathogenic effects after the SE should be mainly attributed to its downstream PGE 2 /EP2 signaling. 14 , 99 Continual efforts in medicinal chemistry and lead optimization to improve the drug-like properties of small-molecule EP2 antagonists 3 and 4 led to the discovery of a second-generation EP2 antagonist, TG8–260 ( 10 ). Unlike the first-generation EP2 antagonists ( Figure 4A ), compound 10 does not possess an acrylamide moiety; however, its structure was not fully disclosed until recently. 100 Nonetheless, compound 10 was tested in a TR-FRET cAMP functional assay with a Schild K B of 13.2 nM and a water solubility of 238 μ M. Strikingly, compound 10 showed a more than 500-fold selectivity to EP2 over other G α s -coupled prostanoid receptors DP, EP4, and IP, 100 which is a significant improvement compared to compounds 3 and 4 . With an intraperitoneal administration in C57BL/6 mice, compound 10 showed a terminal plasma half-life of 2.8 h and an extremely low brain penetration with a brain-to-plasma ratio of 0.02, which is ~15-and 80-fold less than that of compounds 3 and 4 , respectively. Interestingly, a 60 min episode of pilocarpine-induced SE in Sprague–Dawley rats was able to significantly increase the brain-to-plasma ratio of compound 10 from 0.03 to 0.05, suggesting that the SE-induced blood-brain barrier (BBB) opening facilitated its entry to the brain parenchyma. 100 EP2 receptor inhibition by compound 10 administered in adult male Sprague–Dawley rats beginning 2 h after a pilocarpine-induced SE considerably reduced hippocampal neuroinflammation and gliosis ( Table 1 ). However, SE-triggered neuronal injury and BBB breakdown were not mitigated by the treatment with compound 10 , 100 suggesting that the lack of brain penetration prevented the compound from providing favorable beneficial effects like what the first-generation EP2 antagonists did in the pilocarpine model of seizures. Nevertheless, both first-generation (compound 4 ) and second-generation (compound 10 ) EP2 antagonists exhibited powerful anti-inflammatory activities following pilocarpine-induced SE. As the two most commonly used chemoconvulsants in animal models, pilocarpine and kainic acid trigger prolonged seizures that share most fundamental commonalities that simulate human SE conditions, such as mortality, behavioral abnormality, and the development of life-long epilepsy. 14 , 89 , 91 , 96 , 101 However, the proconvulsant effect of pilocarpine is caused by its direct activation of the muscarinic acetylcholine receptor subtype M1, 102 while kainic acid induces experimental seizures through selectively acting on the glutamate receptor subtypes GluK1 in interneurons and GluK2 in principal neurons. 103 , 104 Therefore, these two convulsive agents are likely to have different peripheral and central effects. As such, it is critical to use both models to test therapeutic agents to identify and exclude any model-specific outcomes. The beneficial effects of EP2 inhibition in pilocarpine-treated mice and rats encouraged studies to test EP2 antagonists in a kainic acid model of SE, and compound 4 was first selected owing to its adequate in vivo half-life and favorable brain penetration in rodents. 14 , 74 It has been found that EP2 inhibition by systemic administration of compound 4 after a 1 h episode of kainic acid-induced SE in mice reduced seizure-triggered functional deficits, cytokine induction, reactive gliosis, BBB impairment, and hippocampal damage. 18 Importantly, most of the beneficial effects by compound 4 were independently reproduced in a similar mouse kainic acid model of SE ( Table 1 ). 27 The common, broad, reproducible benefits from a post-SE treatment with compound 4 after pilocarpine- and kainic acid-induced seizures eliminate the possibility that the neuroprotective and anti-inflammatory effects of EP2 antagonism following SE is model-specific and reinforce the feasibility of blocking PGE 2 /EP2 signaling as an adjunctive strategy to treat prolonged seizures. In a recent study, BI1029539, a selective mPGES-1 inhibitor, prevented an upregulation of P-glycoprotein expression and transport activity in capillaries from humanized mPGES-1 mice after kainic acid-induced SE. 105 Whether PGE 2 /EP2 is also involved in SE-promoted P-glycoprotein expression and transport activity in capillaries remains to be determined. Evidence from mounting preclinical studies over the past two decades suggests that the irregular brain-derived neurotrophic factor (BDNF) signaling via its tropomyosin-related kinase receptor B (TrkB) is essential to acquired epilepsy of etiologies such as traumatic brain injury (TBI) and de novo SE. 106 – 109 It thus has been proposed that blocking BDNF/TrkB signaling or the downstream effector phospholipase C γ 1 (PLC- γ 1) might provide promising strategies to interrupt acquired epileptogenesis. 109 – 111 Interestingly, both COX-2 and BDNF in the hippocampus were rapidly elevated by kainic acid or pilocarpine-induced SE with the induction of COX-2 temporally leading that of BDNF. 20 COX-2 inhibition by SC-58125, a potent selective COX-2 inhibitor, 112 or EP2 antagonism by compound 4 , prevented a BDNF elevation in the hippocampus following a pilocarpine-induced SE in rats and mice. Likewise, treatment with compound 4 after a kainic acid-induced SE in mice decreased the SE-triggered phosphorylation of the cAMP response element-binding protein (CREB) and the activation of the BDNF/TrkB signaling in the hippocampus. 20 It is well-known that the CREB activation can upregulate the BDNF expression and TrkB activation, 113 , 114 which in turn can maintain or increase the CREB phosphorylation through a Ca 2+ /calmodulin-dependent protein kinase (CaMK)-dependent mechanism. 115 These interesting findings together suggest that the COX-2 via PGE 2 /EP2 signaling regulates the hippocampal BDNF/TrkB pathway following prolonged seizures ( Figure 1 ). Therefore, EP2 inhibition by selective antagonists might also provide a novel strategy to suppress the aberrant CREB/BDNF/TrkB activity during an acquired epileptogenesis. Diisopropyl fluorophosphate (DFP) is an organophosphorus-based agent with a cholinergic toxicity that can provoke the onset of SE in rodents within minutes and thus is commonly used to model an exposure to nerve agents in humans. A DFP exposure in rats leads to early consequences of cholinesterase inhibition such as whole-body motor convulsions, muscle weakness, and SE, followed days later by neuronal death, neuroinflammation, gliosis, BBB breakdown, weight loss, muscle weakness, and gastrointestinal dysfunction. The long-term consequences of DFP exposure include anxiety behaviors, cognitive deficits, and unprovoked seizures. 116 , 117 Treatment with compound 4 after a DFP-induced SE in rats with six doses beginning 80–150 min after the SE onset substantially decreased the weight loss, acute neuronal death in the hippocampus, inflammatory cytokine burst, reactive microgliosis, and the BBB breakdown in the days after the SE ( Table 1 ). 116 , 118 A selective inhibition of the EP2 receptor by compound 4 in rats did not prevent the anxiety-like behavior assessed by open-field and light-dark box tests four weeks following the DFP exposure; however, it did alleviate the DFP SE-induced memory impairment measured by a novel object recognition test 6–12 weeks after the SE. 116 , 119 These substantial benefits on intermediate and long-term consequences of SE by treatment with EP2 antagonist compound 4 after a DFP exposure reinforce the therapeutic potential of EP2 inhibition in organophosphate-induced pathologies. Note that a single systemic administration of compound 4 approximately 1 h prior to the seizure induction by pilocarpine, kainic acid, or DFP did not change the behavioral seizure progression, the latency to SE, or the duration of SE in these rodent SE models ( Table 1 ). 14 , 18 , 116 , 119 The absence of an effect on behavioral seizures by the EP2 antagonist compound 4 was further validated by an electroencephalography (EEG) recording in pilocarpine-treated mice 14 as well as in DFP-treated rats. 116 , 119 Thus, the neuroprotection, cognitive improvement, and other benefits observed in EP2 antagonist-treated mice and rats were not direct outcomes of an anticonvulsant effect; they instead likely resulted from the anti-inflammatory effects of EP2 receptor inhibition. The lack of an antiseizure effect of EP2 antagonists also rules out the likelihood of the EP2 inhibition as a monotherapy for SE. Rather, these preclinical studies together raise the feasibility that inhibition of the EP2 receptor might provide an adjuvant strategy for the current first-line antiseizure drugs to treat SE. 22 Febrile seizures are caused by high body temperature and are the most common type of seizure in infants and young children. Retrospective studies reveal that persistent febrile seizures or febrile status epilepticus (FSE) in childhood is highly associated with an increased risk of temporal lobe epilepsy (TLE) in adulthood. 120 , 121 Similar to the pilocarpine and kainic acid-induced seizures, the experimental febrile status epilepticus (eFSE) induced in immature rats caused strong inflammatory reactions with a rapid and long-lasting elevation of pro-inflammatory cytokines including interleukin 1 β (IL-1 β ), tumor necrosis factor α (TNF- α ), and IL-6, astrocytic, and microglial activation as well as the activation and upregulation of the COX-2/mPGES-1/PGE 2 /EP2 signaling axis. 122 , 123 However, administration of the EP2 antagonist compound 4 with only three doses beginning 4 h after the eFSE had minimal effects on the total number of aberrant EEG spikes recorded in the first 7–60 d after the eFSE, and it did not significantly alter the duration of each spike series provoked by the eFSE ( Table 1 ). 123 However, whether a longer-term treatment with compound 4 can decrease the development of an abnormal hyperexcitability following eFSE remains to be determined. The discrepancy of EP2 inhibition after pilocarpine/kainic acid-induced SE in adult rodents and eFSE in newborn rats might be at least partially caused by the difference between mature and infantile brains ( Table 1 ). It is possible that the targeting of a specific inflammatory signaling in juvenile epilepsy might alternatively lead to the activation of other inflammatory pathways in response or may simply be inadequate to afford meaningful impacts on the epilepsy development. It is important to know that the EP2 is highly expressed in the developing mouse brain but declines to a very low level in an adult mouse brain, suggesting that EP2 receptor may play some fundamental roles in the brain during the early stages of its development when the EP2 expression remains high. 19 , 25 Indeed, the genetic ablation of the EP2 receptor led to impaired cognition, emotional behaviors, and hippocampal long-term potentiation (LTP) and long-term depression (LTD). 8 , 124 , 125 Thus, EP2 inhibition during the early development stages may impose more profound adversity in the immature brain than it does to the adult brain. All these potential unwanted effects of EP2 inhibition in newborn animals together may profoundly compromise its therapeutic benefits observed in adult SE models.

Retinitis pigmentosa is a group of diseases that involve the breakdown and loss of cells in the retina and is characterized by a progressive degeneration of photoreceptors in the retina. Consisting of a class of heterogenous inherited eye disorders, retinitis pigmentosa can be caused by a host of gene mutations and develop into partial or complete blindness, as currently there is no effective treatment available for patients. Neuroinflammation mediated by activated microglia within the retina is thought to facilitate the progression of this disorder. 156 In retinal degeneration 10 (rd10) mice, an experimental model for retinitis pigmentosa, COX-1 expression, was found to be upregulated within retinal microglia, compared to the wild-type peers. 157 The genetic ablation or pharmacological inhibition of COX-1 alleviated microglial activation as well as preserved the retinal photoreceptor, retinal function, and visual performance in these rd10 mice, suggesting the involvement of COX/PGE 2 signaling in the pathophysiology of the disease. Interestingly, pharmacological inhibition of the EP2 receptor by compound 4 twice daily largely prevented the activation of retinal microglia and suppressed the excessive generation of prototypic pro-inflammatory cytokines including IL-1 β and TNF- α . Furthermore, a reduction in the degeneration of retinal photoreceptors in rd10 mice that were treated by compound 4 was also observed. Consequently, the retinal function and visual performance were largely preserved in rd10 mice by a compound 4 treatment. 157 Taken together, the EP2 inhibition precisely recapitulated the beneficial effects of COX-1 deletion or inhibition in the rd10 retina ( Table 1 ), highlighting the potential and feasibility of EP2 antagonists in treating retinitis pigmentosa.

Over the past decade, significant progress has been made in developing small-molecule antagonists selectively targeting the EP2 receptor for the treatment of peripheral and central inflammation-associated conditions. Compared to the first-generation EP2 antagonists, the second-generation analogues in general are more selective and possess drug-like properties with diverse pharmacokinetic profiles. Some of these compounds such as compounds 4 and 9 have a favorable brain penetration and thus showed excellent therapeutic potential in preclinical models of epilepsy, cerebral ischemia, and brain tumor; others like TG6–129 and compound 1 cannot cross the BBB at all and demonstrated tremendous efficacy in managing peripheral cancers including colon cancer and neuroblastoma. Despite the challenges in early years, the development of novel small molecules that precisely target the PGE 2 /EP2 signaling pathway has promising therapeutic potential to afford profound anti-inflammatory effects in a broad range of inflammation-associated conditions ( Tables 1 and 2 ). Currently, there is no human clinical trial on EP2-selective antagonists; however, a dual antagonist TPST-1495 targeting both EP2 and EP4 is in a current phase 1a/1b study for solid tumors ( https://clinicaltrials.gov/ct2/show/NCT04344795 ). The recently solved cryo-EM structures of the human EP2 receptor and G s protein complex are expected to play some fundamental roles in rational designs to develop the next-generation EP2 antagonists with further improved pharmacokinetic and pharmacodynamic properties for clinical applications ( Figure 5 ). 68 In addition, the use of nanoparticles encapsulated by macrocycle-based compounds such as calixarenes for the controlled release of these potential therapeutic agents might also help in targeted therapy. 198 Many CNS disorders are now considered as inflammation-associated neurological conditions, highlighting the commonality and essentiality of inflammatory signaling pathways in the pathophysiological mechanisms of these diseases. The brain disorders associated with profound neuroinflammatory causes and pathologies can be acute conditions including strokes 199 – 201 and epileptic seizures, 92 , 202 , 203 or chronic diseases, for example, inflammatory and neuropathic pain 204 , 205 and neurodegenerative diseases such as AD, Parkinson’s disease (PD), and amyotrophic lateral sclerosis (ALS). 206 – 209 Accumulating evidence from recent studies suggests that the COX-2/mPGES-1/PGE 2 /EP2 signaling axis may act as a key culprit of the maladaptive immune and inflammatory responses in nearly all these CNS conditions ( Tables 1 and 2 ). 17 , 19 , 21 – 23 , 26 , 57 , 143 , 210 , 211 It is worth noting that the activation of the EP2 receptor leads to the elevation of pro-inflammatory cytokines and NF- κ B, which can further induce COX-2 and thus sustain and even amplify the inflammatory chain reactions. Specifically blocking the PGE 2 /EP2 signaling, a key node of the inflammatory networks, by brain-permeable EP2 selective antagonists is designated to break this vicious self-reinforcing feedback loop of neuroinflammation but does not affect pathways mediated by other prostanoids that may play physiological and beneficial functions under normal conditions. Cancer biology research has been shifting from the focus on cancer cells alone to a more inclusive concept of the cancer microenvironment comprised of cancer cells, cancer-associated fibroblasts, vascular cells, and infiltrating immune cells. 212 Tumor-elicited inflammation has been increasingly recognized for its essential roles in shaping the plasticity of the cancer microenvironment by independently amplifying immune responses and the release of cytokines and many other inflammatory mediators and growth factors that act together to facilitate tumor initiation, growth, progression, and metastasis. 213 The COX-2/PGE 2 /EP2 signaling pathway has been shown to be commonly upregulated in multiple cancer types and to be involved in cancer-related chronic inflammation, immune suppression, angiogenesis, tumor invasion and metastasis, and multidrug resistance. 44 More importantly, elevated PGE 2 /EP2 signaling was identified both within the tumor and its microenvironment. In addition, PGE 2 signaling via the EP2 receptor can also crosstalk with the epidermal growth factor (EGF)/EGF receptor (EGFR) transduction axis, which in turn can activate several other signal pathways, such as mitogen-activated protein kinase (MAPK), PI3K/Akt, signal transducer and activator of transcription (STAT), and PLC, thereby facilitating the proliferation, migration, differentiation, and survival of tumor cells ( Figure 1B ). 44 Therefore, targeting the EP2 receptor by these EP2 antagonists has the potential to serve as a comprehensive treatment strategy to directly impede the proliferation of cancer cells ( Tables 1 and 2 ). They may also suppress the cancer-nourishing molecular and cellular components within the cancer microenvironment and, thus, can potentially be developed as adjunct therapies for current standard anticancer drugs. To determine this possibility in the future, it is important to evaluate combined treatment engaging both EP2 antagonists and front-line chemotherapy or immunotherapy drugs in animal models. The ongoing pandemic of the coronavirus disease 2019 (COVID-19) or severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has once again demonstrated the importance of effective immune responses to a host and the devastating consequences when they are dysregulated to become out of control. Cytokine storm as a term was first used to describe the engraftment complications of acute graft versus host disease (GVHD) after allogeneic bone marrow transplantation, where inflammatory cytokines, such as IL-1, IL-2, interferon γ (IFN- γ ), and TNF- α , may play critical and unique roles. 214 , 215 Despite the term’s long history, there is still no single or widely accepted definition likely due to its wide array of causes and outcomes. 216 Generally, it is thought of as an overactivation of the immune system characterized by an extreme elevation of systemic cytokine circulation and hyperactive immune cells, both of which can have detrimental effects on multiorgan systems. 216 COX-2/PGE 2 /EP2 signaling has been shown to play a considerable role in regulating the immune response, specifically through an induction of a wide array of pro-inflammatory cytokines and chemokines including IL-1 β , IL-6, TNF- α , and CCL2 (MCP-1) in various disease conditions. 18 , 21 , 26 , 38 , 55 , 157 , 197 Interestingly, spike proteins of the COVID-19 virus have been shown to induce a COX-2 expression, which could implicate its involvement in facilitating the cytokine storm in COVID-19. 217 , 218 NSAIDs are primary medicines that are widely used to alleviate pain, fever, and inflammation, which are the common symptoms of COVID-19 patients. 219 Indeed, treatment with NSAIDs has been demonstrated to stifle the inflammatory response to COVID-19, but it also altered the host’s ability to produce protective antibodies. 220 Given that there is no clear evidence of risk or benefit for the use of NSAIDs in COVID-19 patients, 221 further research is needed to better define the role of NSAIDs and particularly COX-2 inhibitors in patients with COVID-19 infection. 222 Both EP2 and EP4 receptors were detected in human lung mast cells, but it is the EP2 that mediates the cAMP signaling and inhibitory effects of PGE 2 on IgE-dependent histamine release in these cells, 73 suggesting a potential use of EP2 antagonists to alleviate the inflammation in asthma and other respiratory diseases. Whether the EP2 receptor is also involved in excessive inflammatory reactions and cytokine storm that are caused by SARS-CoV-2 would be a very interesting question to investigate. Therefore, future studies should also be directed to determine whether the selective EP2 antagonists are able to provide more therapeutic efficacy and specificity than COX-targeting drugs as a novel strategy to combat the uncontrollable cytokine storm following COVID-19 infection.

Synaptic plasticity in nociceptive spinal pathways has been proposed as an important mechanism that might contribute to the amplification of nociceptive signaling under conditions such as acute postoperative pain and chronic pain associated with a peripheral inflammation. 141 COX-2 has been found constitutively expressed in both spinal neurons and radial glia under naive conditions; 142 , 143 however, it also can be substantially upregulated in the spinal dorsal horn after injury, 144 , 145 suggesting that a COX-2/PGE 2 cascade might be involved in the regulation of synaptic LTP within the superficial dorsal horn of the spinal cord. Indeed, both phospholipase A2 (PLA2) selective inhibitor arachidonyl trifluoromethyl ketone (AACOCF3) and COX-2 selective inhibitor nimesulide were able to prevent spike timing-dependent LTP (or tLTP) at sensory synapses onto spinoparabrachial neurons. 143 Interestingly, selective EP2 antagonist compound 1 fully recapitulated the inhibitory effects of AACOCF3 and nimesulide on tLTP, suggesting that the amplification of ascending nociceptive transmission by the spinal superficial dorsal horn network is likely governed by PLA2/COX-2/PGE 2 /EP2 signaling axis. Endometriosis is a chronic gynecological disease that currently does not have a cure and is often characterized by devastating chronic pain and infertility. Changes in pain perception in women are believed to engage a number of pro-inflammatory mediators in efferent peripheral nerve endings. As a well-known mediator of inflammation and nociception in inflammatory and neuropathic pain, 146 – 148 PGE 2 has been found within endometriosis lesions where COX-2 expression is substantially induced. 149 – 151 Interestingly, in a recent study on a preclinical mouse model of endometriosis, COX-2, EP2, and EP4 were found upregulated in endometriosis lesions, dorsal root ganglia, spinal cord, thalamus, and forebrain, suggesting an amplification process along the pain neuroaxis caused by endometriosis. 152 Intraperitoneal injection of the EP2-selective antagonists compound 1 and compound 4 , but not EP4 antagonist L-161982, significantly reversed abdominal or paw hyperalgesia in this model of endometriosis ( Tables 1 and 2 ). Further study showed that the oral administration of compound 1 in mice with endometriosis also resulted in dramatically reduced allodynia in both abdomen and hind-paw tests in a time-dependent manner ( Table 2 ). 152 Taken together, these interesting findings using two different EP2-selective antagonists demonstrated the feasibility of targeting the EP2 receptor as an emerging strategy to treat inflammatory pain and hyperalgesia.

Intracranial aneurysm or brain aneurysm is a weak or thin spot on a cerebral artery or vein that balloons out and fills with blood. More than 50% of intracranial aneurysm patients are unaware of their conditions and thus are left untreated, increasing the chances of aneurysm rupture and fatal subarachnoid hemorrhage. 153 Inflammatory and immunological reactions have long been known in unruptured cerebral aneurysms and may be related to the formation and rupture of aneurysms. 154 Examining the expression of COXs, PGESs, and EP receptors revealed that COX-2, mPGES-1, and EP2 were induced in endothelial cells in the walls of a cerebral aneurysm in both human patients and rats with an induced cerebral aneurysm. 155 Interestingly, the incidence of cerebral aneurysm in mice was significantly prevented by the COX-2 inhibitor celecoxib and the genetic ablation of the EP2 receptor but not by that of EP1, EP3, or EP4. The deficiency of the EP2 receptor also suppressed NF- κ B-mediated chronic inflammation in cerebral aneurysm lesions. 155 These findings revealed an essential role of COX-2/mPGES-1/PGE 2 /EP2/NF- κ B signaling axis in the development and enlargement of a cerebral aneurysm. Oral administration of the EP2 antagonist compound 1 suppressed intracranial aneurysm that was induced in rats in a dose-dependent manner as well as prevented NF- κ B activation, COX-2 induction, and the expression of monocyte chemoattractant protein 1 (MCP-1), leading to the reduction in macrophage infiltration into intracranial aneurysm lesions ( Table 2 ). 43 Thus, targeting the EP2 receptor by small-molecule antagonists could be a novel pharmacological strategy to treat developing intracranial aneurysms and reduce the potential risk of a fatal subarachnoid hemorrhage.

Cyclooxygenase (COX) is the key enzyme in the two-step biosynthesis of prostanoids from arachidonic acid (AA), a 20-carbon fatty acid, after it is liberated from the cell membrane by cytoplasmic phospholipase A2 (cPLA2) ( Figure 1A ). 1 , 2 COX has two currently known isoforms, namely, COX-1 and COX-2, which are encoded by ptgs1 and ptgs2 , respectively. On the one hand, the COX-1 isozyme is constitutively expressed to play some essential roles, such as maintaining the integrity of the gastrointestinal tract lining and preserving normal renal and cardiovascular functions. 1 COX-2, on the other hand, is usually expressed at low basal levels in most normal tissues and organs. However, it can be rapidly and robustly upregulated in response to inflammatory and mitogenic stimuli, thereby commonly being considered as an early response gene. 1 , 2 Both COX-1 and COX-2 catalyze AA to prostaglandin H2 (PGH 2 ), which is an intermediate molecule and can be further converted by tissue-specific prostanoid synthases to five forms of prostanoids ( Figure 1A ): prostaglandin D2 (PGD 2 ), prostaglandin E2 (PGE 2 ), prostaglandin F2 α (PGF 2 α ), prostacyclin or prostaglandin I2 (PGI 2 ), and thromboxane A2 (TXA 2 ). 1 These short-lived lipid-like bioactive molecules in turn dynamically activate several membrane-bound G protein-coupled receptors (GPCRs): four receptors (EP1, EP2, EP3, and EP4) activated by PGE 2 ; two (DP1 and DP2) activated by PGD 2 ; each of the other three prostanoids (PGF 2 α , PGI 2 , and TXA 2 ) acts on only one receptor, that is, FP, IP, and TP, respectively ( Figure 1A ). 3 , 4 Interestingly, despite their sequence and structural similarities, COX-1 and COX-2 have different major enzymatic products. 5 Particularly, COX-2 is found to be functionally coupled to microsomal prostaglandin E synthase 1 (mPGES-1), which is also inducible and catalyzes COX-2-derived PGH 2 into PGE 2 in response to inflammatory or injurious stimuli. 2 The mPGES-2 and cytosolic PGES (cPGES) are the other two PGES isoenzymes that synthase PGE 2 from COX-1-produced PGH 2 at basal physiological conditions ( Figure 1A ). PGE 2 is known to produce pleiotropic effects throughout the body including immunogenicity, inflammation, vasodilation, platelet aggregation, and memory formation, and it is thought to perpetuate the COX-2 expression through a positive feedback loop. 6 – 9 PGE 2 regulates these physiological and pathological events via interactions with four GPCRs, EP1–EP4 ( Figure 1A ). The EP1 receptor is G α q -coupled to regulate the mobilization of cytosolic Ca 2+ and the activation of protein kinase C (PKC); EP2 and EP4 are linked to G α s for cAMP-dependent pathways; the EP3 receptor is mainly coupled to G α i to downregulate the cAMP signaling. 10 , 11 PGE 2 signaling via the EP2 subtype has been demonstrated to be involved in many of the immunoregulatory, inflammatory, neurotoxic, and other pathophysiological effects that are mediated by the COX-2/PGE 2 signaling cascade under various disease conditions 10 , 12 – 27

The pathology of AD is characterized by the development of amyloid β (A β ) plaques and neurofibrillary tangles. 136 At the onset of AD, activated microglia are designated to contribute to the clearance of the neurotoxic A β plaques through phagocytosis. 137 As the plaques aggregate, however, microglia activated by a long-term exposure to A β can be skewed from the phagocytic phenotype to a neuroinflammatory phenotype, and this transition during the disease progression has been proposed to be mediated by PGE 2 signaling. It has been shown that PGE 2 induces an inhibitory effect on a microglia/macrophages-mediated phagocytosis through the EP2 receptor. 138 An upregulation of the EP2 receptor is associated with an increased pro-inflammatory gene expression and a decreased clearance of A β plaques in which microglia are involved. 57 It has been noted that the EP2 signaling appears to suppress pretty much all beneficial aspects of activated microglia including phagocytosis, termination of pro-inflammatory responses, release of neurotrophic factors, and lysosomal function. 139 When the microglial EP2 receptor is ablated, microglial phagocytotic activity is enhanced, and the aggregation of A β is mitigated. 140 Thus, pharmacologically targeting the EP2 receptor might provide a novel therapeutic strategy to restore the beneficial functions of healthy microglial and prevent or even reverse the progression to AD and cognitive aging. Among the newly discovered selective EP2 antagonists, compound 9 was evaluated in an ex vivo assay using brain slices that were harvested from 18-month-old Tg2576 mice, a common experimental model for AD. 71 It was observed that treatment with compound 9 increased the macrophage-mediated A β phagocytosis in a concentration-dependent manner. 71 In a more recent study, two EP2 antagonists were evaluated in the context of cognitive aging: brain permeable compound 9 and brain impermeable compound 1 ( Table 2 ). Interestingly, both Amgen and Pfizer compounds restored cellular bioenergetics and microglia/macrophage phagocytosis and improved the hippocampal synaptic plasticity as well as the spatial memory in aged mice. 26 Together, the improved bioenergetics and immune shifting induced by the EP2 antagonism contributed to the recovery of hippocampal synaptic protein levels and the long-term potentiation (LTP) in the CA1 region. Interestingly, compound 1 —though it was unable to penetrate the brain parenchyma—attenuated pro-inflammatory responses not only in the blood but also in the CNS. 26 It is postulated that the peripheral EP2 inhibition by compound 1 may cause as-yet-unidentified alterations in the blood that beneficially influence the aging cerebrovascular endothelium or cross the BBB to directly improve the neuronal function. 26 It is also possible that the compromised BBB integrity in an aged brain enables compound 1 to cross the BBB as compound 10 did in rats undergoing a pilocarpine-induced SE. 100