International Union of Basic and Clinical Pharmacology. CIX. Differences and Similarities between Human and Rodent Prostaglandin E₂ Receptors (EP1–4) and Prostacyclin Receptor (IP): Specific Roles in Pathophysiologic Conditions

1. T Lymphocytes

For a long time, PGE₂ through EP2 and EP4 activation and downstream cAMP-PKA signaling was believed to suppress both mouse and human T-cell activation and primary cytokine production [e.g., IL-2 and interferon (IFN)-γ] in response to antigens or mitogens (Brudvik and Tasken, 2012). However, this concept has been challenged recently with evidence that these receptors may actually act as an immune activator (Yao and Narumiya, 2019). In this respect, activation of EP2 and/or EP4 receptor by PGE₂ or their selective agonists facilitated mouse Th1 cell differentiation, and this was prevented by their respective antagonists (Yao et al., 2009, 2013). Interestingly, these effects involve both cAMP-PKA– and PI3K-Akt–signaling pathways (Yao et al., 2009, 2013). Findings linking PGE₂ and EP2/4 receptors to human autoimmune inflammatory diseases have been revealed by genetic association studies, which positively link enhanced PGE₂ signaling to IL-23/ type 17 helper T cell (Th17) signature genes and disease severity (Yao and Narumiya, 2019).

In other studies, an EP1-receptor agonist, ONO-DI-004, increased Th1 cell differentiation in wild-type mouse T cells but had little effect on differentiation of EP1-deficient Th1 cells (Nagamachi et al., 2007). On the one hand, PGE₂ and EP2/EP4-receptor agonists inhibit human Th1 cells by reducing the expression of the transcription factor T-box expressed in T cells and the production of the cytokine IFN-γ (Boniface et al., 2009; Napolitani et al., 2009). On the other hand, PGE₂ favors the production of type 2 cytokines IL-4 and IL-5 from human T cells but does not affect mouse type 2 helper T cell (Th2) cytokine production in vitro (Hilkens et al., 1995). Moreover, PGE₂ and EP2/EP4-receptor agonists significantly promote IL-17 and IL-22 production from both mouse and human Th17 cells because of induction of IL-23 and IL-1β receptors, an effect counteracted by their respective antagonists (Hilkens et al., 1995; Chizzolini et al., 2008; Boniface et al., 2009; Napolitani et al., 2009; Yao et al., 2009; Chen et al., 2010; Lee et al., 2019).

Although Th1, Th2, and Th17 cells modulate various proinflammatory and antimicrobial responses, regulatory T cells (Treg) usually act in an anti-inflammatory manner. It was previously reported that PGE₂ promotes both mouse and human Treg cell differentiation, especially in the tumor microenvironment (Baratelli et al., 2005; Sharma et al., 2005). In contrast, it was recently found that PGE₂ can also inhibit both mouse and human Treg development in vitro, which is mimicked by EP2- and EP4-receptor agonists and mediated by the cAMP-PKA pathway (Hooper et al., 2017; Li et al., 2017b; Maseda et al., 2018).

Similar to PGE₂, PGI₂ can also activate the cAMP pathway and regulate T-cell function, but depending on the context of disease model being investigated, it can have proimmunomodulatory or anti-immunomodulatory effects, including on T cells (Dorris and Peebles, 2012). In one of the early studies, IP-receptor agonists were reported to inhibit Th1 cell differentiation in vitro via a cAMP-dependent suppression of NF-κβ (Zhou et al., 2007a). In subsequent experiments by the same group, antigen- or IL-33–dependent Th2 cell function and allergic lung inflammation could be elicited by prostacyclin analogs both in vitro and in vivo, (Zhou et al., 2007b, 2018a). A critical role of the IP receptor in response to a fungal challenge in mouse lungs was also confirmed in PTGIR null mice, in which increased IL-5 and IL-13 responsiveness of CD4+ T cells to Alternaria sensitization, typically a response requiring IL-33, was observed (Zhou et al., 2018a). The situation appears to differ for other types of antigen reactions, in which activation of IP receptor by the agonist iloprost enhances Th1 differentiation (suppresses Th2 differentiation) in vivo and promotes Th1 cell–mediated inflammatory responses in a mouse model of contact hypersensitivity. This is also consistent with PTGIR null mice displaying much less contact hypersensitivity (Nakajima et al., 2010). Moreover, the IP-receptor agonists iloprost and cicaprost facilitated mouse Th17 cell differentiation and function and increased IL-17 and IL-22 production from human Th17 cells (Truchetet et al., 2012; Zhou et al., 2012). Collectively, the new findings indicate that PGE₂ and PGI₂ can use the cAMP pathway to promote inflammatory effector T-cell (e.g., Th1, Th17, and Th22) responses in vivo, although they primarily downregulate T cell–receptor activation. Thus, the role of G_s-coupled prostanoid receptors on T-cell function, although compelling, is complex and remains somewhat conflicting. Immunomodulation appears to be dependent on the animal species, the type of inflammatory disease studied, and, to an extent, whether data has been collected in vitro or in vivo.

2. Innate Lymphoid Cells

Innate lymphoid cells (ILCs) are a group of cells that secrete large amounts of prototypic T-cell cytokines, such as IFN-γ, IL-17, and IL-4, in response to appropriate stimuli, but, unlike T cells, they do not express T-cell receptors (Vivier et al., 2018). ILCs exert their functions by producing different cytokines (e.g., ILC1 cells produce IFN-γ, and ILC2 cells secrete IL-4, IL-5, and IL-13, whereas ILC3 cells mainly produce IL-17 and IL-22). In mice, PGE₂ and an EP4-receptor agonist, L-902,688, increased the production of IL-22 from ILC3 cells in vitro. Inhibition of PGE₂ production using a nonselective COX inhibitor indomethacin or blockade of EP4 signaling using either a selective EP4-receptor antagonist L-161,982 or deletion of EP4 receptor on T cells and ILCs reduced IL-22 production in vivo (Duffin et al., 2016). PGE₂ also promotes IL-22 production from human ILC3 (Duffin et al., 2016). In contrast to experiments in ILC3 cells, an EP4-receptor agonist, PGE₂-alcohol, mimicked PGE₂ inhibition of ILC2 cytokine production in mice, whereas the EP2-receptor agonist, butaprost, did not reduce ILC2 cytokine production significantly in the same setting (Zhou et al., 2018b). Similarly, although the EP4-receptor agonist L-902,688 mimicked PGE₂ suppression of cytokine production from human ILC2, an EP2-receptor agonist, butaprost, had little effect (Maric et al., 2018).

3. Dendritic Cells

Dendritic cells (DCs) are important cells for regulating innate immunity and for presenting antigens to T cells, with PGE₂ playing a critical role in the maturation of DCs (Kalinski, 2012; Jia et al., 2019). PGE₂ is required for the migration of DCs, allowing their homing to draining lymph nodes, and this is mimicked by EP4-receptor agonist in mouse DCs and by both EP2- and EP4-receptor agonists in human monocyte-derived DCs (Kabashima et al., 2003; Legler et al., 2006). Engagement of EP2 and EP4 receptors by their agonists promotes the production of the proinflammatory cytokine IL-23, which is critical for development and maturation of Th17 cells by both mouse and human DCs. These effects were found to be mediated through activation of the cAMP-PKA pathway and transcription factors CREB, NF-κB, and C/AATT enhancer-binding protein β (Kocieda et al., 2012;Shi et al., 2015; Ma et al., 2016). Moreover, EP2- and EP3-receptor agonists induced the generation of human tolerogenic DCs characterized by the induction of high levels of the immunosuppressant, IL-10, whereas an EP4-receptor agonist favored the development of inflammatory DC by promoting the production of IL-23 and Th17 polarization (Flórez-Grau et al., 2017). In other studies, the EP3-receptor agonist ONO-AE-248 was observed to inhibit the chemotaxis and costimulatory molecule expressions of mouse DCs in vitro and restricted DC cell function to fine-tune excessive skin inflammation in vivo (Shiraishi et al., 2013). A number of studies have reported a suppressive effect of the IP receptor in DCs. The IP-receptor agonists cicaprost, iloprost, and treprostinil inhibited the production of proinflammatory chemokines and cytokines from human monocyte-induced DCs stimulated by lipopolysaccharides (LPSs) or TNF-α (Hung et al., 2009; Yeh et al., 2011; Wang et al., 2017a). Furthermore, iloprost also suppressed mouse-airway DC function to inhibit Th2 differentiation and thereby reduced allergic lung inflammation in a mouse model of asthma (Idzko et al., 2007).

4. Macrophages

In human macrophages, PGE₂-EP2/EP4-cAMP signaling inhibits inflammatory cytokine production (e.g., TNF-α, IL-1β) and phagocytosis, promoting an M2-phenotype associated with an increase in IL-10 production. For example, the EP4-receptor agonist L-902,688 was found to inhibit human lung macrophage production of TNF-α, whereas the EP2 agonist butaprost was 400× less potent, suggesting a major role for the EP4 over the EP2 receptor in this study (Gill et al., 2016). In other studies, EP4 receptors may play a role in the resolution phase of inflammation (Sokolowska et al., 2015). It was found that PGE₂ and the EP4 agonist CAY10598 inhibited the activation of nucleotide-binding oligomerization domain-like receptors family pyrin domain-containing 3 inflammasome and IL-1β production in human primary monocyte-derived macrophages, whereas EP4-receptor antagonist GW627368X (Wilson et al., 2006) or EP4 knockdown reversed the PGE₂-mediated nucleotide-binding oligomerization domain-like receptors family pyrin domain-containing 3 inhibition (Sokolowska et al., 2015). Binding of advanced glycation end-products on human monocytes/macrophages activated T cells and reduced allograft survival, a process that was inhibited by PGE₂, the EP2-receptor agonist ONO-AE1-259, and the EP4-receptor agonist ONO-AE1-329. The inhibitory effects of PGE₂ were prevented by either by AH6809 (EP1/2 and DP1 antagonist) or AH23848 (EP4/TP antagonist) (Takahashi et al., 2010). In other situations, however, PGE₂, butaprost, and the EP4 agonist CAY10598 could inhibit 1,25-dihydroxy vitamin D3–induced production of human cationic antimicrobial protein-18 from human macrophages during Mycobacterium tuberculosis infection (Wan et al., 2018). Given that responses could partially be reversed by AH6809 (EP1/2 and DP1 antagonist) or L-161,982 (EP4 antagonist) but not L-798106 (EP3 antagonist) suggests a dual role for EP2 and EP4 receptors in restraining the innate immune response and prolonging microbial survival. Likewise, the killing of Klebsiella pneumoniae by rat alveolar macrophages was prevented by PGE₂ and treprostinil, with both agents acting in part through EP2 receptors (Aronoff et al., 2007).

As already documented with EP2/4 receptor agonists, PGI₂ analogs (iloprost, beraprost, treprostinil, and ONO-1301) are also able to suppress LPS-induced proinflammatory monocyte chemoattractant protein-1 (MCP-1) production from human monocytes and macrophages (Tsai et al., 2015). More recently, Aoki and colleagues recently reported that inactivation of the PGE₂-EP2- NF-κB–signaling pathway in mouse macrophages reduced macrophage infiltration and proinflammatory cytokine (e.g., MCP-1) production, leading to the prevention of intracranial aneurysms. They found that administration of EP2-receptor antagonist PF-04418948 (af Forselles et al., 2011) in rats reduced macrophage infiltration and intracranial aneurysm formation and progression (Aoki et al., 2017a,b). The IP-receptor agonist cicaprost stimulated vascular endothelial growth factor secretion but inhibited MCP-1 production from TNF-α–treated human monocyte-derived macrophages, whereas administration of an IP-receptor antagonist, RO3244794 (Bley et al., 2006), significantly reduced neovascularized lesion area in mouse choroidal neovascularization model (Woodward et al., 2019). In the same experimental setting, such effects could be mimicked by PGE₂ acting in part on EP4 receptors (as determined by the EP4 antagonist, GW627368). Taken together, this suggests that both IP and EP4 receptors may play a role macrophage-driven neovascularization.

5. Neutrophils

To kill invading microbes, neutrophils release their nuclear contents in an NADPH oxidase and reactive oxygen species –dependent manner, with neutrophil extracellular traps (NETs) playing a critical role in killing bacteria, fungi, or viruses by physically trapping them. PGE₂ plays a key role in this process and has been shown to inhibit human neutrophil function (such as superoxide production, migration, and antimicrobial peptide release), an effect that was prevented by AH6809 (EP1/2 and DP1 antagonist) but not by the EP4 receptor–selective antagonist ONO-AE2-227 (Turcotte et al., 2017). Other studies have confirmed that PGE₂ inhibits human NET formation through EP2 and EP4 receptors in vitro and via the EP2 receptor in vivo, in which the EP2 agonist butaprost suppressed NET formation in mice (Shishikura et al., 2016). Both mouse and human neutrophils overexpress COX-2 and PGE₂ post–bone marrow transplantation (Ballinger et al., 2006) and exhibit defective bacterial killing due to reduced NET formation (Domingo-Gonzalez et al., 2016). Reduced NET formation after bone-marrow transplantation in mice and humans could be restored by COX inhibitors or an EP2-receptor antagonist (PF-04418948) plus an EP4-receptor antagonist (ONO-AE3-208) (Domingo-Gonzalez et al., 2016). Activation of EP4 receptor prevented endotoxin-induced mouse neutrophil infiltration into airways (Konya et al., 2015). Similarly, EP4 receptor mediates PGE₂-induced enhancement of human pulmonary microvascular barrier function against neutrophil infiltration (Konya et al., 2013). Taken together, these relatively recent studies provide good evidence that PGE₂, through activation of both EP2 and EP4 receptors, suppresses the immune response of neutrophils. Thus, manipulation of these receptors therapeutically may prove useful in blocking pathologic NETosis in autoimmune diseases and/or aid the host response to infection.

6. Eosinophils and Mast Cells

Human and mouse eosinophils express EP2 and EP4 receptors, which mediate the effects of PGE₂ by blocking eosinophil responses, such as degranulation, chemotaxis, and production of reactive oxygen species (Mita et al., 2002; Sturm et al., 2008; Luschnig-Schratl et al., 2011). The underlying signaling pathways appear to involve PI3K, phosphoinositide-dependent kinase 1, and PKC but not the cAMP/PKA pathway (Luschnig-Schratl et al., 2011; Sturm et al., 2015). PGE₂ via the EP4 receptor inhibited the interaction of eosinophils with human pulmonary endothelial cells in vitro, including adhesion and transmigration (Konya et al., 2011). In other experimental settings, the EP2 receptor appears to inhibit the mobilization of eosinophils from guinea-pig bone marrow and allergen-induced eosinophil recruitment to mouse lung (Sturm et al., 2008) and is involved in IgE-dependent human-airway constriction in vitro by inhibiting mast-cell activation (Safholm et al., 2015).

In human mast-cell lines and primary cord blood-derived mast cells, EP2-, EP3-, and EP4-receptor proteins are expressed (Feng et al., 2006; Torres-Atencio et al., 2014). PGE₂ counteracted the hyperosmolar-induced degranulation of these mast cells via EP2 and EP4 receptors (Torres-Atencio et al., 2014). Human lung mast cells likewise express both EP2 and EP4-receptor mRNA, but it is the EP2 receptor that predominantly mediates the inhibitory effect of PGE₂ on histamine release in vitro (Kay et al., 2013). In contrast, EP3-receptor activation causes migration, adhesion, antigen-dependent degranulation, and IL-6 release of mouse mast cells (Nguyen et al., 2002; Weller et al., 2007; Sakanaka et al., 2008) and potentiates histamine release in human peripheral blood-derived mast cells (Wang and Lau, 2006).

Similar to PGE₂, PGI₂ attenuates the locomotion of human peripheral blood eosinophils and guinea-pig bone-marrow eosinophils via IP-receptor activation (Konya et al., 2010; Sturm et al., 2011). Unlike EP2/EP4 signaling in eosinophils, the inhibitory-effect PGI₂ is mediated by intracellular cAMP. Accordingly, endothelium-derived PGI₂ controls eosinophil-endothelial interaction and promotes the barrier function of lung endothelial cells to limit eosinophil adhesion and transendothelial migration (Konya et al., 2010). Thus, these data explain previous findings that deletion of the IP receptor in mice augmented allergen-induced eosinophilia in the lung and skin and enhanced airway remodeling (Takahashi et al., 2002; Nagao et al., 2003).

7. Hematopoietic Stem/Progenitor Cell and Leukemia

PGE₂ treatment of hematopoietic stem cell (HSC)/hematopoietic progenitor cell (HPC) from mice and humans promotes survival, proliferation, and engraftment in vitro (Hoggatt et al., 2009) These effects are recapitulated by the EP2-receptor agonist ONO-AE1-259 or the EP4-receptor agonist ONO-AE1-329, which increased mouse and human HSC/HPC colony formation and long-term bone-marrow reconstitution capacity of Lineage⁻Sca-1⁺c-Kit⁺ cells (Ikushima et al., 2013; Wang et al., 2017b). Moreover, treatment of bone-marrow mesenchymal progenitor cells with PGE₂ or the EP4-receptor agonist significantly increased their ability to support HSPC colony formation (Ikushima et al., 2013). Conversely, treatment with COX inhibitors increased HPCs in peripheral blood of both mice and humans. Similarly, administration of selective EP4-receptor antagonists L-161,982 or AH23848 expanded bone-marrow HPCs and enhanced HPC mobilization in mice induced by GM-CSF, whereas administration of selective EP4-receptor agonist (L-902,688), rather than EP1-, EP2-, or EP3-receptor agonists, reduced HPC mobilization (Hoggatt et al., 2013).

PGE₁ treatment impaired the persistence and activity of leukemic stem cells in a preclinical mouse chronic myelogenous leukemia (CML) model and a xenograft model of transplanted CML patient CD34⁺ HSCs/HPCs, and a nonselective EP2/EP3/EP4 agonist, misoprostol, conferred similar protection against CML, suggesting potential therapeutic strategy of CML by using PGE₁ or misoprostol (Li et al., 2017a). Along the same lines, the breakpoint cluster region-Abelson inhibitors imatinib and nilotinib were recently described to enhance PGE₂ biosynthesis in monocytes of healthy volunteers and CML patients, an effect that might contribute to their clinical efficacy in the treatment of CML (Bärnthaler et al., 2019).

1. Multiple Sclerosis

Numerous genome-wide associated studies suggest that polymorphisms in the 5p13.1 regulatory region near PTGER4 (encoding human EP4) are significantly associated with PTGER4 gene expression and the susceptibility to MS (De Jager et al., 2009; Matesanz et al., 2012). The level of PGE₂ was increased in cerebrospinal fluid of MS patients (Bolton et al., 1984). Compared with healthy individuals, Th17 cells from MS patients have higher levels of EP2 receptor, resulting in increased expression of proinflammatory cytokines like IFN-γ and GM-CSF and pathogenicity of Th17 cells (Kofler et al., 2014). Administration of EP2-receptor agonist did not affect proinflammatory cytokine production from Th17 cells of healthy individuals but increased IFN-γ and cerebrospinal fluid (CSF) 2 production from Th17 cells isolated from MS patients (Kofler et al., 2014). Studies using an animal model of MS [experimental autoimmune encephalomyelitis (EAE)] demonstrated that EP4-receptor gene deletion or pharmacological blockade of EP4 receptor during the immunization stage prevented EAE development in mice and downregulated Th1/Th17 cells (Yao et al., 2009; Esaki et al., 2010). Administration of an EP4-receptor agonist after peak disease response still reduces the peak EAE severity. These results thus suggest distinct roles of the EP4 receptor at different stages of EAE disease.

2. Rheumatoid Arthritis

In the animal model of carrageenan-induced paw inflammation, neutralization of PGE₂ by a monoclonal antibody prevented the development of tissue edema and hyperalgesia in affected paws, an effect associated with reduced IL-6 production (Portanova et al., 1996). Blockade of PGE₂-EP2/EP4 signaling using receptor antagonists similarly suppressed joint inflammation in the mouse model of allergen-induced arthritis, which again was related to a reduction in IL-6 (McCoy et al., 2002; Honda et al., 2006). IL-6 is the key cytokine that mediates Th17 cell development, and PGE₂ facilitates Th17 immune responses. Misoprostol, a PGE₂ analog binding to EP2, EP3, and EP4 receptors, exacerbated collagen-induced arthritis in mice through activating the inflammatory IL-23/IL-17 axis, whereas an EP4-receptor antagonist reduced arthritis in a mouse model (Sheibanie et al., 2007a; Chen et al., 2010). This supports previous observations in genetically modifed mice, in which deletion of inducible mPGES-1 and EP4, but not EP3 or EP2, reduced arthritic incidence and severity in simlar experimental models (McCoy et al., 2002; Trebino et al., 2003). The therapeutic perspectives associated with EP4-receptor blockade will be developed in Section X. D. Arthritis.

3. Inflammatory Bowel Disease and Colon Cancer

COX-2 activity in the colonic epithelial cells of IBD patients (Singer et al., 1998) and PGE₂ levels in the lesions of IBD patients are elevated (Schmidt et al., 1996). Similarly, gene polymorphisms in PTGER4 loci are associated with increased PTGER4 gene expression and susceptibility to Crohn disease, suggesting a critical role of EP4 receptors (Libioulle et al., 2007; Glas et al., 2012). PGE₂ is known to act in different ways in the gastrointestinal tract. For example, PGE₂ plays fundamental roles in maintaining the gastrointestinal epithelial barrier, and therefore, blockade of PGE₂ synthesis or the EP4 receptor was found to promote acute gastrointestinal injury in mice (Kabashima et al., 2002; Duffin et al., 2016) and induce gut damage in humans. However, misoprostol, an EP2/EP3/EP4 agonist, aggravated intestinal inflammation induced by 2,4,6-trinitrobenzene sulfonic acid in mice through promoting the inflammatory IL-23/IL-17 pathway (Sheibanie et al., 2007b). In contrast, genetic deletion of the PGE₂ synthase mPGES-1 or the EP4 receptor in T cells ameliorated T cell–mediated chronic intestinal inflammation in mice, which was associated with reduction of the development of inflammatory Th1 and Th17 cells in the intestine (Maseda et al., 2018). Therefore, PGE₂ may also facilitate T cell–mediated chronic intestinal inflammation in both mice and humans.

PGE₂ has long been known to be associated with the development and progression of colorectal cancer (CRC), and use of COX inhibitors has been suggested to prevent CRC (Chan et al., 2005). Genome-wide association studies have indicated that polymorphisms in the NAD+-dependent 15-hydroxyprostaglandin dehydrogenase (15-PGDH), an enzyme that breaks down PGE₂ into biologically inactive 15-keto-PGE₂, are associated with higher risk for CRC, whereas polymorphisms in PTGER2 were associated with lower CRC risk (Hoeft et al., 2010). PGE₂ promoted human LoVo colon cancer cell proliferation and migration through activation of PI3K/Akt and glycogen synthase kinase 3β/β-catenin pathways; this could be prevented by the nonselective EP2- or EP4-receptor antagonist, AH6809 (EP1/2 and DP1 antagonist) or AH23848 (EP4/TP antagonist), respectively (Hsu et al., 2017). PGE₂ and the EP2-receptor agonist butaprost promoted the survival of human colorectal carcinoma-15 cells, another human colon cancer cell line, and this was also prevented by AH6809 (Shehzad et al., 2014). In mice, deficiency of the EP2 receptor or treatment with the EP2-receptor antagonist, PF-04418948, reduced azoxymethane and dextran sodium sulfate–induced colon tumorigenesis associated with downregulation of proinflammatory genes (e.g., TNF-α, IL-6, CXCL1, and COX-2) in tumor-associated fibroblasts and neutrophils (Ma et al., 2015). Thus, further studies are needed to understand the differential roles of PGE₂ in gastrointestinal homeostasis, such as, for example, the underlying mechanisms of how PGE₂ protects against acute gastrointestinal injury and promotes mucosal regeneration but also promotes chronic (especially T cell–mediated) intestinal inflammation and colorectal cancer.

4. Lung Inflammation

PGE₂ and PGI₂ are generally bronchoprotective, although these two prostanoids can increase inflammatory Th1 and Th17 cell numbers and function under autoimmune inflammatory conditions in other organs. In patients with aspirin-exacerbated respiratory disease, the PGE₂-EP2 pathway is downregulated, but is associated with upregulation of PGD₂ and leukotrienes as well as overactivation of type 2 innate lymphoid cells (Rusznak and Peebles, 2019). Using immunohistochemistry on human bronchial biopsy, it was reported that patients with aspirin-sensitive asthma had increased bronchial mucosal neutrophil and eosinophil numbers but reduced percentages of T cells, macrophages, mast cells, and neutrophils expressing EP2 (Corrigan et al., 2012). Given that EP2-receptor agonists were the only prostanoid-EP agonists to inhibit cytokine production in peripheral blood mononuclear cells, the authors concluded that EP2 agonists might be beneficial in patients with asthma. Likewise, in mice, the PGE₂-EP2 signaling suppressed allergen sensitization and thus attenuated the development of Th2-polarized immunity and airway inflammatory responses (Zasłona et al., 2014). On the other hand, EP2-deficiency had the opposite effect and enhanced type 2 eosinophilic responses and IgE production in ovalbumin (OVA)-sensitized mice, whereas administration of misoprostol (EP2/EP3/EP4 agonist) to WT rather than EP2-deficient mice suppressed this inflammatory response and attenuated the IgE production (Zasłona et al., 2014). Results of this study provide evidence for a critical role of EP2 receptors in inhibiting airway inflammation through the dampening down of the Th2-cell cytokine surge. Such a view is not supported by Church and colleagues (2012), who reported that mPGES1-mediated PGE₂ production in the lung contributed to the enhancement of allergic responses at the effector phase after allergen challenge. Based on parallel studies in congenic COX-1/2–deficient mice, they concluded that the primary prostanoid that was protecting against allergic inflammation was PGI₂ and not PGE₂. This is consistent with much earlier studies reporting that IP receptor–deficient mice have more severe allergic reactions in the lung compared with WT mice (Takahashi et al., 2002) and with more recent studies, in which deletion of mPGES-1 increased vascular production of PGI₂ presumed to be from the redistribution of precursor PG substrate (Avendaño et al., 2018). Evidence supporting this notion that PGE₂ might actually drive allergic inflammation comes from Gao and colleagues (2016), who found that PGE₂-EP2 signaling in B cells actually promoted IgE production in OVA-induced asthma models. In trying to reconcile these conflicting results, it is important to note that genetic drift of mouse colonies almost certainly exists in different laboratories. This may in turn affect the nature of the allergic and inflammatory response action, including the amount of IgE and the cytokine profile generated with different immunization (e.g., the length of challenge) and experimental protocols (see Church et al., 2012; Gao et al., 2016 for further discussion).

Recently, PGE₂-EP4/EP2 signaling has been reported to inhibit mouse as well as human ILC2 cell activation, which may contribute to control of allergic lung inflammation (Maric et al., 2018; Zhou et al., 2018b). By using agonists and antagonists in mouse models, EP2/EP4 receptors were reported to abrogate acute lung injury and inflammation through actions on various immune cells, including T cells, macrophages, B cells, eosinophils, innate lymphoid cells, and endothelial cells (Sheller et al., 2000; Sturm et al., 2008; Birrell et al., 2015; Konya et al., 2015; Draijer et al., 2016; Felton et al., 2018). Cicaprost, a potent and relatively selective IP agonist, was shown to inhibit IL-33–induced allergic lung inflammation through suppression of Th2 and ILC2 responses (Zhou et al., 2016, 2018a; Jian et al., 2017). Similarly, administration of ONO-1301, a novel prostacyclin analog with TxA₂ synthase inhibitory activity, protected against OVA- and house dust mite–induced airway inflammation and remodeling in mice (Yamabayashi et al., 2012; Kimura et al., 2013).

These effects of PGE₂ on lymphocytes are relevant for lung diseases, such as hypersensitivity pneumonitis, sarcoidosis, pulmonary fibrosis, and bronchial asthma. In mouse models of asthma, the activation of the EP3 receptor on bronchial epithelial cells inhibited the allergen-induced expression of chemokines (Hirata and Narumiya, 2012). EP2-receptor agonists inhibited GM-CSF release. EP4 receptors in human macrophages inhibited proinflammatory cytokines (IL-8) release (Gill et al., 2016), whereas EP4-receptor agonists inhibited neutrophils infiltration in the mice lung (Konya et al., 2015). In knockout mice, the EP4 receptor was involved in eosinophil and neutrophil infiltration in in vivo animal models of asthma, chronic obstructive pulmonary disease (COPD), and inflammation (Birrell et al., 2015). In human lung-transplant recipients, genetic variation in PGES and EP4 coding genes have been found to be associated with primary graft dysfunction and decreased Treg suppressor cell function (Diamond et al., 2014). It is imperative to further investigate why PGE₂ acts mainly as a suppressant in lung inflammation, whereas it can exert both protective and proinflammatory activities in most other organ systems. With respect to manipulating PGE₂ levels experimentally, it is important to note that EP4 (and EP3) will be activated at 10-fold lower concentrations (of PGE₂) than EP2, whereas at high concentrations (≥100 nM), the PGF_2α (FP) receptor will be significantly activated as well as, in humans, the DP1 receptor (Clapp and Gurung, 2015). Taken together, this is likely to make the interpretation of the role of PGE₂ in regulating inflammation complex potentially hard to extrapolate between different studies.

5. Skin Inflammation and Cancer

PGE₂ synthases and EP receptors are expressed in both human and mouse skin. Multiple types of cells within the skin, such as mast cells, macrophages, dendritic cells, and keratinocytes, can all produce PGE₂. In a mouse model of delayed-type hypersensitivity, PGE₂ and its receptor agonists suppressed skin inflammation by increasing the production of immunosuppressive type 2 cytokines (e.g., IL-4 or IL-10) (Shreedhar et al., 1998; Miyauchi-Hashimoto et al., 2001). Blockade of endogenous PGE₂ production by a COX inhibitor or EP2 signaling by a relevant antagonist enhanced the production of the cytokine thymic stromal lymphopoietin from keratinocytes, whereas type 2 immune responses in the skin were attenuated by administering an EP2-receptor agonist (Sawada et al., 2019). Later on, PGE₂ was shown to promote contact hypersensitivity, whereas EP4-receptor antagonist or EP4 deficiency reduced hapten-induced skin inflammation, probably through the action of EP4 receptors on skin dendritic-cell migration and Th1/Th17 cell expansion (Kabashima et al., 2003; Yao et al., 2009, 2013). Administration of a selective EP1-receptor antagonist ONO-8713 during the sensitization stage also suppresses hapten-induced skin inflammation (Nagamachi et al., 2007). Furthermore, Lee and colleagues (2019) suggested that blockade of EP2 and EP4 signaling (using genetically modified animals and receptor antagonists) reduced the generation of pathogenic Th17 cells and psoriatic skin inflammation. Blockade of the PGE₂-EP4 pathway restricted allergic contact dermatitis in mice associated with reduced IL-22 production from T cells (Robb et al., 2018). In human inflamed skin (both atopic and psoriatic), the levels of PGE₂ and gene expression of its synthases and receptors were found to be increased, and effective therapies downregulated PGE₂ pathway–related gene expression (Fogh et al., 1989; Robb et al., 2018; Lee et al., 2019). In human keratinocytes, PGE₂ suppressed CCL7 expression through EP2 and EP3 receptors, leading to a reduction of inflammatory T-cell homing within the skin (Kanda et al., 2004). Thus, PGE₂ seems to have differential and partially opposing effects on different types of cells in human skin.

Development of skin tumor has long been known to be associated with enhanced COX-2–PGE₂-EP signaling (Rundhaug et al., 2011). Deficiency of mPGES-1, the key enzyme mediating PGE₂ synthesis, prevented B16 melanoma cell growth in vivo, and treatment with an EP4-receptor antagonist similarly inhibited not only the growth of B16 tumor cells but metastasis to bone marrow in mice (Inada et al., 2015). Oral, uveal, and cutaneous melanoma cells isolated from patients showed reduced IL-8 production in vitro, which was mimicked by the (EP3 > EP1) receptor agonist sulprostone and prevented by the EP3-receptor antagonist L-798106 (Venza et al., 2018).

Moving forward, a comprehensive understanding of the actions for PGE₂ and its receptors on distinct cell types during skin inflammation will be particularly important.

1. Vascular Tone Regulation

The characterization of prostanoid receptors in human blood vessels is particularly important, and these results could provide therapeutic approaches for different diseases. Because of the fact that obtaining fresh human blood vessels on a regular basis is tough, many studies have instead focused on vessels obtained from different experimental animal models. However, several differences occur between vessels derived from human and rodents in terms of the responsible EP-receptor subtype for vasoconstriction/vasorelaxation induced by PGE₂. For example, the control of vascular tone by PGE₂ in human renal artery is regulated somewhat differently than rodent renal artery. For example, in rat renal artery, 11-deoxy-PGE1 (EP2 and EP4-receptor agonist) induced relaxation, whereas butaprost (EP2-receptor agonist) is relatively ineffective (Tang et al., 2000). The contraction induced by higher concentrations of PGE₂ is mimicked by sulprostone, an (EP3 > EP1) receptor agonist (Tang et al., 2000). Similarly, in vivo renal blood flow studies in rats indicated that sulprostone caused transient renal vasoconstriction, whereas prolonged relaxation was obtained with EP4-receptor activation (Purdy and Arendshorst, 2000). In mice, the deletion of individual EP receptors demonstrated that EP2 receptor is partly involved in renal vasodilatation, whereas EP1 and EP3 receptors are involved in renal vasoconstriction (Imig et al., 2002). On the other hand, in human renal artery, no role for the EP2, EP1, and EP3 receptors was detected. Instead, the PGE₂-induced relaxation was mimicked by CAY10598 (EP4 agonist), and the PGE₂-induced contraction was blocked by the TP-receptor antagonist, S18886 (Eskildsen et al., 2014). Moreover, some studies have demonstrated that the pharmacology and mechanism underlying the effect of IP-receptor activation on vascular tone are different between human and rodent blood vessels (Clapp and Gurung, 2015).

Another difference between human and experimental animals has been observed in the regulation of pulmonary artery vascular tone by PGE₂. Studies performed in human pulmonary artery demonstrated that sulprostone (EP3 > EP1 agonist) contraction is insensitive to TP-receptor antagonists (EP169 and GR32191) (Qian et al., 1994). This finding suggests the involvement of EP3 receptors in PGE₂-induced contraction in human pulmonary artery. However, PGE₂-induced relaxation occurs via EP2 or EP4 receptors in rabbit pulmonary artery contracted by norepinephrine (Kitamura et al., 1976), unlike in proximal human pulmonary artery, in which PGE₂ fails to cause relaxation of arteries preconstricted with norepinephrine (Walch et al., 1999).

On the other hand, in some vessels similarities do exist between humans and experimental animals in terms of the role of PGE₂ in regulating vascular tone. Studies report that PGE₂-induced vasodilatation occurs via the EP4 receptor in saphenous vein derived from many species, including from human, rabbit, piglet, or guinea-pig tissue (Coleman et al., 1994; Lydford et al., 1996; Jones and Chan, 2005; Wilson and Giles, 2005; Foudi et al., 2011). It is important to note that pharmacological tools for the characterization of EP-receptor subtypes should be chosen carefully since they can have different effects on human and rodents. For example, GW627368X is frequently used as a selective EP4-receptor antagonist, but it can also bind to TP receptors in humans, whereas this is not the case in other species (Wilson et al., 2006).

Several studies demonstrated that PGE₂-induced contraction is mostly mediated by EP3 receptor in human “healthy” vessels, such as coronary, internal mammary, intercostal, and pulmonary arteries (Foudi et al., 2011; Kozłowska et al., 2012; Longrois et al., 2012; Ozen et al., 2015). By contrast, the EP1 receptor is involved in the contraction evoked by PGE₂ in several rodent vessels, including in rat renal artery, aorta, or mesenteric artery (Michel et al., 2007; Xavier et al., 2009; Silveira et al., 2014). On the other hand, the relaxant effect of PGE₂ in the majority of human vessels is mediated by the EP4 receptor, whereas in rodents, vasodilatation induced by EP2-receptor activation is also observed (Imig et al., 2002; Davis et al., 2004; Foudi et al., 2008, 2011; Maubach et al., 2009). In accordance with this finding, PGE₂ produces substantial hypertension in EP2 null mice (Kennedy et al., 1999). Somewhat variable effects of gene deletion on systolic blood pressure were reported, although mice in both studies developed profound salt-sensitive hypertension but not in controls (Kennedy et al., 1999; Tilley et al., 1999). Taken together, this suggests a major role for EP2 receptors in regulating vascular tone and sodium handling in the kidney, at least in mice.

PGI₂ induces vasodilatation by the activation of IP receptor. The effects of PGI₂ analogs on vascular tone have mostly been determined in in vitro studies using pulmonary arteries. PGI₂ analogs, such as treprostinil, iloprost, and beraprost induced relaxation in both human and rat pulmonary arteries (Benyahia et al., 2013, 2015; Shen et al., 2019). However, when the precontractile agent is endothelin-1, iloprost and treprostinil are able to relax human pulmonary artery but not rat pulmonary artery (Benyahia et al., 2015). On the other hand, TP-receptor activation by high doses of PGI₂ elicits contraction in some rodent vessels, although such an effect is not exhibited in human vessels (Xavier et al., 2009; Baretella and Vanhoutte, 2016). Relaxation by prostacyclin analogs in normal vascular preparations is consistently enhanced over a wide agonist concentration range through blocking EP3-receptor function or G_i coupling and suggests tonic activation of these receptors opposes the action of PGI₂ and its analogs in the cardiovascular system (Clapp et al., 2020). Overall, characterization of EP-receptor subtypes and effects of IP/EP-receptor agonist/antagonist on vascular tone are still not determined in several human vessels, such as coronary artery, carotid artery, or aorta. Further studies are warranted and will provide therapeutic approaches for different diseases, such as atherosclerosis or aneurysm.

1. Hypertension

The substantial roles of prostanoids in the regulation of blood pressure are highlighted by the prohypertensive effects of nonsteroidal anti-inflammatory drugs (NSAIDs) and COX-2 inhibitors (Snowden and Nelson, 2011; Ruschitzka et al., 2017). Likewise, when the PGIS gene is deleted, mice become hypertensive and show fibrosis and vascular remodeling in the kidney (Yokoyama et al., 2002) but not when the IP receptor is deleted (Hoshikawa et al., 2001), suggesting additional targets for PGI₂ in the regulation of vascular function, most likely through PPARs (Clapp and Gurung, 2015). Elevated expressions of mPGES-1 and COX-2 are found in hypertensive patients, mice, and rats versus their normotensive controls (Boshra et al., 2011; Avendaño et al., 2016). Furthermore, deletion of the mPGES-1 gene in hypertensive mice prevented the increased vasoconstrictor response to angiotensin 2 (Ang-II) (Avendaño et al., 2018). In accordance with this finding, an mPGES-1 inhibitor decreased the contractile response to noradrenaline in human arteries and veins (Ozen et al., 2017). Overall, both in human and rodents, the enzymes responsible for the synthesis of PGI₂ and PGE₂ are involved in the pathogenesis of arterial hypertension.

Under physiologic conditions, there is a balance between the effects of EP1/EP3 receptor (vasoconstriction) and those of EP2/EP4 receptor (vasodilation), whereas in hypertensive models and patients with cardiovascular disease, this balance is likely to be disrupted (Clapp and Gurung, 2015). Consistent with this notion, EP1 receptor participated in impaired vascular function observed in hypertensive animal models (Avendaño et al., 2016). Moreover, in mice, a reduced systolic blood pressure was observed upon treatment with various EP1-receptor antagonists (ONO-8713, AH6809, SC51322) and with deletion of the EP1 gene (Stock et al., 2001; Guan et al., 2007; Rutkai et al., 2009). EP3-receptor agonists, such as sulprostone (EP3 > EP1 agonist), MB28767, or SC46275 increased mean arterial pressure in wild-type mice (Zhang et al., 2000), and increased renal blood flow was observed in EP3-deficient mice (Audoly et al., 2001). Furthermore, EP2- and EP4-receptor knockout mice have elevated systolic blood pressure (Kennedy et al., 1999; Tilley et al., 1999; Xu et al., 2019). The role of EP2 in the regulation of blood pressure in mice is supported by the study, which revealed an association between a polymorphism of the EP2 gene and essential hypertension in men (Sato et al., 2007). In contrast, mostly EP3 agonists, such as sulprostone and misoprostol, which are used as clinical practice, are associated with an increased incidence of ischemic stroke or myocardial infarction in patients probably due to coronary constriction. Globally, these results suggest a strong role for the EP3 receptor in the control of arterial blood pressure (Guerci et al., 2013; Vital et al., 2013; Masclee et al., 2018; Mazhar et al., 2018; Schink et al., 2018). These in vivo observations are supported by the numerous in vitro data on the vasoconstrictor role of the EP3 receptor in human vasculature as mentioned previously (Qian et al., 1994; Norel et al., 2004a; Foudi et al., 2011; Longrois et al., 2012; Ozen et al., 2015).

The expression of PGIS and the concentrations of the stable PGI₂ metabolite (6-keto-PGF_1α or 2,3-dinor-6-keto-PGF_1α) measured in hypertensive patients were found to be similar or lower than those obtained for normotensive patients (Lemne et al., 1992; Klockenbusch et al., 2000; Vainio et al., 2004; Hellsten et al., 2012). A number of polymorphisms in the PGIS gene have been described that are not associated with essential hypertension in humans (Nakayama et al., 2002a, 2003). However, several studies performed in hypertensive animal models, including spontaneously hypertensive rats, Dahl salt-sensitive rats, deoxycorticosterone-salt hypertensive rats, and renovascular hypertensive rats, demonstrate increased levels of 6-keto-PGF_1α, probably via the induction of COX-2 expression (Ishimitsu et al., 1991; Matsumoto et al., 2016). Infusion of PGI₂ or its mimetics exhibited similar results in humans and rats and caused reductions of blood pressure (Pickles and O’Grady, 1982; Frölich, 1990; Kato et al., 1992; Zlatnik et al., 1999; Picken et al., 2019). Finally, polymorphisms of the PGIS promoter have been discovered and are associated with increased synthesis of PGIS (Stearman et al., 2014).

The potent vasodilatory effects of PGI₂ in the systemic and pulmonary circulation are well documented (Clapp and Gurung, 2015). The evidence that IP receptors per se regulate blood pressure under physiologic conditions is not supported by gene-deletion studies, in which mice are normotensive (Hoshikawa et al., 2001). However, loss of the IP receptors leads to the development of renovascular hypertension, and mice have an exaggerated hypertensive and remodeling effect of hypoxia in the lung. Thus in disease, the IP receptor appears to become dysfunctional in the vascular system and, in doing so, may unmask the contractile effects of PGI₂ as observed in spontaneously hypertensive rats (Félétou et al., 2009). Not only in hypertensive rodent models but also in diabetic and aged rats, IP-receptor signaling appears to be impaired. PGI₂ no longer caused vasodilatation but became a prominent endothelium-derived vasoconstrictor by activating TP receptors (Vanhoutte, 2011). However, in most human vascular preparations, PGI₂ or its analogs induced only relaxation (Benyahia et al., 2015; Foudi et al., 2017).

Data from clinical trials in pulmonary hypertensive (PH) patients showed that chronic treatment with PGI₂ analogs leads to a significant fall in pulmonary arterial pressure and pulmonary vascular resistance but does not lead to a drop in systemic blood pressure with standard doses (Picken et al., 2019). This may point to a “relative” pulmonary-selective effect of these agents, although common side effects of all IP agonists are headache and flushing, suggesting a potent vasodilatory effect on cerebral and skin blood vessels, respectively. The situation may be different in patients with systemic hypertension, although the effects of IP agonists on blood systemic arterial pressure in this condition have not been routinely studied.

2. Diabetes

The plasma concentrations of PGE₂ in diabetic patients were found to be similar or higher than those obtained for nondiabetic patients (Arisaka et al., 1986; Axelrod et al., 1986; Mourits-Andersen et al., 1986). In diabetic rats, the plasma levels of PGE₂ were also increased (Axelrod and Cornelius, 1984; Craven et al., 1987).

Studies performed in diabetic mice and rats demonstrated that increased vascular tone and diabetic nephropathy were reversed by either AH6809 or ONO-8713 (EP1-receptor antagonists). EP1-receptor overexpression was detected in diabetic rat glomeruli by using Northern blot analysis, and confirmation was obtained by using in situ hybridization. This observation suggests that EP1 receptors could contribute to the development of hypertension and nephropathy in diabetic rodents (Makino et al., 2002; Rutkai et al., 2009); however, the contribution of EP1 receptor in diabetic patients needs to be evaluated. The EP2 receptor could also be involved in diabetic retinopathy in both humans and rats, possibly because of accelerated retinal vascular leakage, leukostasis, and endothelial cell apoptosis (Wang et al., 2019). On the other hand, other studies performed on streptozotocin-induced diabetic rats showed that EP2 and EP4 receptors were involved in the protective roles of PGE₂ (Vennemann et al., 2012; Yasui et al., 2015), highlighting the complex role of these prostanoid receptors in different experimental models. Moreover, the suppression of EP4 receptor–associated protein has been suggested as a novel strategy for the treatment of diabetes in mice (Vallerie et al., 2016; Higuchi et al., 2019).

In both diabetic patients and mice models, upregulation of the EP3 receptor in pancreatic islet of Langerhans has been reported by using real-time PCR (Kimple et al., 2013; Amior et al., 2019). Blockade of the EP3 receptor by DG-041 (Su et al., 2008) in combination with activation of EP4 receptor increased β-cell proliferation in humans but not in mice (Carboneau et al., 2017). Moreover, DG-041 had no significant effects on the diabetic phenotype of mice (Ceddia et al., 2019), whereas L-798106 (EP3-receptor antagonist) has been shown to decrease insulin resistance in db/db mice (Chan et al., 2016). In contrast with this finding, genetic deletion of all three EP3 isoforms or deletion of only EP3α and EP3γ isoforms resulted in increased insulin resistance in mice when they were fed a high-fat diet (Ceddia et al., 2016; Xu et al., 2016).

The release of 6-keto-PGF_1α was significantly decreased in both diabetic patients and rats (Lubawy and Valentovic, 1982; Jeremy et al., 1987a; Brunkwall and Bergqvist, 1992; Kalogeropoulou et al., 2002; Bolego et al., 2006; Peredo et al., 2006), suggesting loss of circulating PGI₂. The expression of PGIS was lower in subcutaneous arteries from diabetic patients, whereas IP-receptor expression remained unchanged (Mokhtar et al., 2016b; Safiah Mokhtar et al., 2013). In contrast, another study performed on human platelet showed that IP-receptor expression was inversely correlated with HbA1c levels (Knebel et al., 2015), which may contribute to platelet hyperactivity in humans with type 2 diabetes. This is certainly consistent with the heightened thrombotic state observed in mice not expressing the IP receptor (Hoshikawa et al., 2001). Furthermore, PGIS- and IP-receptor expressions were decreased in Zucker diabetic fatty rats and streptozotocin-induced diabetic rats, respectively (Nasrallah and Hebert, 2004; Lu et al., 2005). Even though the expression of the IP receptor was not changed in diabetic patients (Safiah Mokhtar et al., 2013), PGI₂-induced relaxation was increased in coronary arterioles of patients with diabetes, presumably to compensate for decreased nitric oxide bioavailability (Szerafin et al., 2006). In contrast, there was no compensatory role of PGI₂ in streptozotocin-induced diabetic rats (Mokhtar et al., 2016a), whereas PGI₂-induced contraction was increased, and the relaxation induced by IP-receptor agonist was decreased in diabetic mice (Kimura et al., 1989; Przygodzki et al., 2015).

3. Abdominal Aortic Aneurysm

Dilatation and weakening of the aorta in abdominal aortic aneurysms (AAAs) are accompanied by an alteration in the blood vessel, such as an increase of local inflammation, smooth muscle cell apoptosis, elevated oxidation stress, and especially extracellular matrix degradation via matrix metalloproteinase (MMP) activity (Han et al., 2018). Several research groups (including our own) have demonstrated that PGE₂ release and mPGES-1 expression were increased in vascular preparations derived from AAA patients (Camacho et al., 2013; Solà-Villà et al., 2015; Gomez et al., 2016). These in vitro results are supported by clinical studies, which demonstrated a lower aneurysm growth rate in patients receiving NSAIDs (Walton et al., 1999). In line with human studies, Ang-II–induced AAA formation in mice resulted in increased PGE₂ levels and deletion of mPGES-1 protected against AAA formation (King et al., 2006; Wang et al., 2008).

The role of EP-receptor subtypes has also been investigated in the pathogenesis of AAA. In human AAA samples, higher EP4-receptor expression versus normal samples was demonstrated by real-time PCR studies (Camacho et al., 2013). Exposure of aortic smooth muscle cells (SMCs) and macrophages derived from human AAA preparations to EP4-receptor antagonists (CJ-42794 or ONO-AE3-208) decreased MMP activation and proinflammatory cytokines secretion (Cao et al., 2012; Yokoyama et al., 2012; Mamun et al., 2018). These findings suggested that EP4-receptor antagonists could be a therapeutic target for the treatment of AAA. Similarly, administration of EP4-receptor antagonist or deletion of the EP4 receptor in ApoE knockout mice reduced AAA formation via diminution of cytokine/chemokine levels and MMP activities (Cao et al., 2012; Yokoyama et al., 2012; Mamun et al., 2018). However, there is one contradictory study performed in hyperlipidemic mice. In this study, deficiency of EP4 receptor increased AAA formation induced by Ang-II (Tang et al., 2011b). As described above (in section II. A. 4. Macrophages), it has recently been shown that deletion or antagonism (PF-04418948) of the EP2 receptor is responsible for reduced macrophage infiltration and intracranial aneurysm formation in rodents (Aoki et al., 2017a,b; Shimizu et al., 2019). Taken together, these results suggest that aneurysm development involves SMC and macrophage crosstalk with EP2 and/or EP4-receptor activation, in human and rodent models. As a hypothesis, the EP receptors involved could be dependent on the type of aneurysm (aortic or intracranial artery). Overall, most studies have addressed the role of the EP4 receptor in human AAA development, and the roles of EP2 and EP4 receptors were investigated in rodent models. EP3-receptor subtypes were also detected in aortic SMCs derived from patients with or without AAA (Bayston et al., 2003), and some EP3 mRNA splice variants were differentially expressed between human aortic SMC derived from control versus AAA patients. Further investigations may shed light on a potential novel role of EP3 receptors in arterial aneurysms.

In contrast, only a few studies investigated the role of PGI₂ and its receptor in the pathogenesis of AAA. Solà-Villà and colleagues (2015) demonstrated that there were increased PGE₂ levels in human AAA samples, which were significantly correlated with enhanced levels of PGI₂ released from the tissue samples, whereas in other studies there was no change in PGI₂ levels from human AAA samples obtained from patients with Marfan syndrome (Soto et al., 2018). However, Wang et al. (2008) demonstrated that deletion of mPGES-1 gene increased 2,3-dinor-6-keto-PGF_1α concentrations in urine of mice with Ang-II–induced AAA. Although this suggested that PGI₂ has a protective effect against AAA formation in mice, further studies in both human and rodents need to be conducted to substantiate the role and mechanism of PGI₂ in AAA development.

4. Obesity

Obesity is defined as a low-grade inflammatory disease and associated with many cardiovascular diseases, including diabetes, hypertension, and metabolic syndrome (Ceddia et al., 2016). Since PGE₂ and PGI₂ are induced under inflammatory conditions, several studies have focused on the role of these mediators in the development of obesity.

Recently, we and other groups have demonstrated that PGE₂ levels in plasma and omental adipose tissue are greater in obese patients (García-Alonso et al., 2016; Ozen et al., 2019), whereas there are conflicting results in terms of the PGE₂ levels measured in obese rodent models (Pham Huu Chanh et al., 1987; Cunha et al., 2010; Rocha-Rodrigues et al., 2017). Another difference has been in the detected level of mPGES-1 expression between obese patients and obese rodents: mPGES-1 expression remained unchanged in the adipose tissue of obese patients, whereas it is decreased in those obtained for obese mice (Hétu and Riendeau, 2007; García-Alonso et al., 2016). EP3 mRNA expression is significantly and consistently upregulated in primary adipocytes isolated from high-fat diet–induced obese rats and human subjects (Chan et al., 2016). Moreover, EP3 mRNA levels are positively correlated with the body mass index in humans and TNF-α and MCP-1 levels in adipose tissue (Chan et al., 2016). On the other hand, downregulation of EP3 isoforms in high-fat diet–induced obese mice was reported by using both Western blot and real-time PCR techniques (Xu et al., 2016). In accordance with these findings, other studies indicated that EP3 knockout mice had an obese phenotype with abnormal lipid distribution and accumulation versus wild-type mice (Sanchez-Alavez et al., 2007; Ceddia et al., 2016). In contrast, short-term treatment DG-041, an EP3-receptor antagonist, had no effect on body composition and glycemic control in obese diabetic mice (Ceddia et al., 2019). Overall, the roles of mPGES-1 enzyme and EP3 receptor in obesity have been evaluated in many studies; however, there are several differences documented between human and rodents.

The roles of other EP-receptor subtypes have so far only been determined in obese animal models but not in obese patients, and this should be evaluated in future studies. The EP4-receptor agonist TCS 251 induced a greater relaxation in coronary arterioles derived from obese rats (Santiago et al., 2016), and this could be due to increased EP4-receptor expression in obesity. Enhanced EP4-receptor expression could have protective roles since the activation of EP4 receptor is involved in the reduction of adipose tissue inflammation (Tang et al., 2015; Yasui et al., 2015). On the other hand, only one study demonstrated that EP2 levels are significantly decreased in macrophages of obese diabetic mice (Hellmann et al., 2013); however, the role of EP2 receptor in the development of human obesity remains to be elucidated. Furthermore, it should be noted that in the Hellman study, only Western blot techniques were used for quantification of the EP receptor. Since antibodies for EP receptors are generally not very specific unless verified by small interfering RNAs, gene deletion, or expression systems, other techniques, such as in situ hybridization or PCR, are necessary for verification.

In humans, weight reduction induces a significant decrease of 6-keto-PGF_1α production in adipose tissue (Katz and Knittle, 1991). In contrast, in high-fat diet–induced obese rats, spontaneously hypertensive obese rats, or obese Zucker rats, there is no change, or there is a decrease in 6-keto-PGF_1α levels (Goodwill et al., 2008; Hodnett et al., 2009; Mendizábal et al., 2013; Vendrame et al., 2014; Lee et al., 2017; Lemaster et al., 2017). PGI₂-mediated vasodilatation is impaired in patients with obesity and metabolic syndrome (Limberg et al., 2013). In accordance with this finding, in obese Zucker rats, a decrease of PGI₂-induced vasodilatation is observed, and the contraction response induced by PGI₂ via TP-receptor activation is increased (Xiang et al., 2006; Baretella and Vanhoutte, 2016). On the other hand, another study performed on these rats demonstrated that neither IP nor TP-receptor expression was changed by immunofluorescence studies (Hodnett et al., 2009). Overall, even though 6-keto-PGF_1α production is different between human and rodents in obesity, vasodilator effects of PGI₂ were decreased in both obese patients and rodents. The mechanism underlying this decrease needs to be elucidated in obese patients; future investigations could thus provide novel therapeutic aspects especially for obesity-related cardiovascular diseases.

5. Atherosclerosis

The levels of PGE₂ are increased in preparations with atherosclerotic lesions and plasma from atherosclerotic patients and rodents (Rolland et al., 1984; Gómez-Hernández et al., 2006a; Pang et al., 2019). Elevated levels of mPGES-1 expression were observed in human atherosclerotic preparations (Gómez-Hernández et al., 2006a), and the deletion of mPGES-1 gene in mice retarded atherosclerosis and neointimal hyperplasia (Wang et al., 2006, 2011). This protection may in part be conferred by a compensatory increase in PGI₂ levels, which is associated with deletion of mPGES-1 in mice. Using double-knockout mice (mPGES-1 and IP), protection from injury was lost, and neointima formation was more severe compared with IP-deficient mice, confirming a role for PGE₂ as well as PGI₂ in restraining the neointimal response to injury (Hao et al., 2018).

There are contradictory results for the involvement of EP1 receptor in the pathogenesis of atherosclerosis. One study demonstrated that EP1 receptor was not detected in human carotid plaques (Cipollone et al., 2005), whereas another study demonstrated the expression of EP1 receptors in the shoulder of plaques by Western blot, PCR, and immunohistochemical techiques (Gómez-Hernández et al., 2006a). In addition, patients treated with statins had decreased EP1-receptor expression in atherosclerotic plaques, which may be associated with the beneficial effects of statins (Gómez-Hernández et al., 2006b). EP2-receptor expression was detected in human carotid and femoral plaques, but there was no difference in EP2-receptor expression levels between atherosclerotic and nonatherosclerotic preparations, as determined by Western blot, PCR, and immunohistochemical studies (Gómez-Hernández et al., 2006a,b; Muto et al., 2010). Interestingly, some oxidized phospholipids that accumulate in atherosclerotic lesions have been found to activate EP2 receptors and might hence play a role in atherogenesis in humans (Li et al., 2006). EP2 receptors have also been implicated in a mouse atherosclerosis model, although with a different angle: vascular SMC from EP2 knockout mice showed increased migration and proliferation, suggesting that EP2-receptor activation could be beneficial for the treatment of vascular remodeling, as observed in atherosclerosis (Zhu et al., 2011).

There is a significant increase of EP3-receptor expression in human carotid plaques, which was detected by Western blot, PCR, and immunohistochemical studies (Gómez-Hernández et al., 2006a). On the other hand, another study revealed that oxidation of low-density lipoprotein decreased EP3-receptor expression in human macrophages. The downregulation of EP3 expression by oxidized low-density lipoprotein resulted in impairment of EP3-mediated anti-inflammatory effects (Sui et al., 2014). Hypercholesterolemia and increased diet-induced atherosclerosis were observed in mice with genetic deletion of hepatocyte-specific EP3 receptor (Yan et al., 2017). These studies suggest that EP3-receptor activation could have beneficial roles in the treatment of atherosclerosis and hypercholesterolemia. In contrast, the study performed in mice using different prostaglandin receptor antagonists and small interfering RNA revealed that EP3α and β splice variants are involved in neointimal formation in response to injury (Zhang et al., 2013a).

The overexpression of EP4 receptor was detected in human carotid atherosclerosis by Western blot, PCR, and immunohistochemical studies, and the EP4 receptor was shown to be involved in destabilization of the plaques by the regulation of MMPs (Cipollone et al., 2005; Gómez-Hernández et al., 2006a). Similarly, PGE₁-OH, an EP4-receptor agonist, stimulated MMP-9 expression in macrophages of mice (Pavlovic et al., 2006). The deletion of the EP4 receptor in mice macrophages reduced aortic atherosclerosis (Babaev et al., 2008), whereas deletion of the EP4 receptor in bone marrow–derived cells in mice did not change atherosclerotic lesion size but increased inflammation (Tang et al., 2011a). A recent study demonstrated that hypercholesterolemia was observed in EP4 knockout mice, and treatment with EP4-receptor agonist (CAY10580) in mice fed with a high-fat diet prevented diet-induced hypercholesterolemia (Ying et al., 2018). Consistent with a role for EP4 receptors in re-endothelialization after angioplasty-wire injury, deletion of EP4 in endothelial cells was enhanced while EP4 agonists protected against neointimal formation (Hao et al., 2018). In humans, PGE₂ decreased chemokine levels in human macrophage through EP4-receptor activation, and this could prevent atherosclerotic plaque development (Takayama et al., 2002, 2006). The role of EP4 receptor in atherosclerosis was studied in more detail in both humans and mice by using genetic deletion or selective receptor agonist/antagonists; however, results regarding the role of other EP-receptor subtypes are contradictory, and mechanisms of their beneficial or harmful effects are not fully understood yet.

The link between PGI₂ signaling and atherosclerosis is highlighted by the side effects of COX-2 inhibitors via their inhibitory effect on PGI₂ levels. Generally, PGI₂ has been found to play athero-protective roles. At the time of the development of PGI₂ for the treatment of pulmonary arterial hypertension (PAH), much of the focus of the clinical use of PGI₂ was on the treatment of peripheral vascular diseases, such as critical limb ischemia associated with atherosclerosis and Buerger disease reviewed in Clapp and Gurung (2015). Moreover, PGI₂ has an inhibitory effect on platelet-derived growth factor (PDGF) production, which plays an important role on SMC proliferation and neointimal formation in atherosclerosis. Both antiproliferative and also lipid-lowering effects of PGI₂ were observed in aortic cells derived from both human and rodents and lead to antiatherosclerotic effects (Clapp and Gurung, 2015). The urinary levels of 2,3-dinor-6-keto-PGF_1α are greater in patients and mice with atherosclerosis and suggest that greater PGI₂ production could act as a compensatory mechanism. In addition, it should be noted that the measurement of the urinary metabolite 2,3-dinor-6-keto-PGF_1α reflects kidney synthesis PGI₂ and is not representative of total production in whole body. However, another study performed on patients with atherosclerotic diseases demonstrated that PGI₂ levels are not associated with major adverse cardiovascular events and vascular inflammation (Wang et al., 2018).

IP-receptor expression was decreased in human atherosclerotic plaques (Di Taranto et al., 2012), and an IP-receptor mutation was associated with atherothrombosis in a patient cohort with a high risk of cardiovascular disease (Arehart et al., 2008). Similarly, the genetic deletion of IP receptor in mice or pharmacological inhibition of PGI₂ by COX-2 inhibitors resulted into a significant acceleration in atherogenesis (Kobayashi et al., 2004; Gitlin and Loftin, 2009). In accordance with this finding, PGI₂ analogs (octimibate and BMY 42393) reduced early atherosclerosis in hyperlipidemic hamster (Kowala et al., 1993).

6. Cerebral Stroke

Two types of strokes are reported in literature: ischemic and hemorrhagic stroke (Mehndiratta et al., 2015). In a rat model of ischemia stroke, an upregulation of COX-2 mRNA is observed after the occlusion of the middle cerebral artery, and this leads to increased production of PGE₂ by 292% ± 57% after 24 hours (Nogawa et al., 1997). Similarly, PGE₂ production is increased after 24 hours of brain ischemia in mice, and it is associated to a significant increased level of COX-2 mRNA and proteins (Yokota et al., 2004). Similar results are found in humans, with increased mRNA and protein levels of COX-2 found in post-mortem infarcted human brain (Sairanen et al., 1998). The COX-2 inhibitor, NS-398, attenuated the elevation of PGE₂ in the postischemic brain and reduced the volume of infarction (Nogawa et al., 1997). Another study demonstrated that deletion of mPGES-1–coding gene in mice abolished postischemic PGE₂ production in the cortex, and that is associated with a reduction of myocardial infarction, edema, and cell death in comparison with WT mice (Ikeda-Matsuo et al., 2006). In rats, inhibition of autophagy after an ischemic stroke shows a significant decreased level of proinflammatory molecules, such as PGE₂ (He et al., 2019). The important role of PGE₂ is observed in adult human, wherein COX-2/PGE₂ pathway is associated with the middle cerebral artery occlusion and the hemorrhagic stroke in patients with Moyamoya disease caused by blocked arteries at the base of the brain. Furthermore, COX-2 and mPGES-1 were found abundant in the vascular walls of middle cerebral artery and superficial temporal artery in patients with Moyamoya disease (Zhang et al., 2016a). In line with these results, it was found that the polymorphism of COX-2 (-765 G>C) in humans is linked to a decreased risk of stroke (Cipollone et al., 2004), highlighting a strong link between COX-2 and enhanced cardiovascular risk.

The genetic deletion of EP1 receptor is related to a neurotoxic effect in mouse model of brain transient ischemia (Zhen et al., 2012). Treatment with the EP1-receptor antagonist, SC51089 or EP1 gene deletion demonstrated an improvement of middle cerebral artery occlusion in mice (Kawano et al., 2006) and reduced neuronal death after an episode of transient forebrain ischemia (Shimamura et al., 2013). Moreover, EP1 knockout mice were associated with a decreased ischemic lesion after stroke and increased cerebral blood flow (Ahmad et al., 2006; Saleem et al., 2007). Another study showed that in an ischemic stroke model, pretreatment with a specific EP1-receptor antagonist, ONO-8713, reduced the size of the infarct (Ahmad et al., 2008). Finally, it was shown that inhibition of EP1 receptor improved the survival of hippocampal slices (in culture) from mice with ischemic stroke induced by oxygen-glucose deprivation (Zhou et al., 2008). These observations suggested that EP1-receptor antagonists could be therapeutic targets in ischemic stroke.

The EP2 receptor appears to have an important beneficial role in reducing cerebral ischemia in an experimental model of stroke (Andreasson, 2010). Consistent with this, genetic deletion of EP2 receptor resulted in increased infarct volumes in mice (McCullough et al., 2004; Liu et al., 2005). Moreover, pharmacological activation of EP2 by ONO-AE1-259-01 significantly reduced the infarct volume in mice (Ahmad et al., 2010). However, recent studies demonstrated that neuronal EP2-receptor expression is induced after cerebral ischemia (Liu et al., 2019). Validation of the anti-EP2 antibody used in that study was performed by using cerebellar lysates and HEK cells overexpressing the EP2 receptor. The blockade of EP2 in mice contributed to cerebro-protection by reducing neuroinflammation (Liu et al., 2019), and EP2 knockout mice were shown to be more protected from intracerebral hemorrhage strokes (Leclerc et al., 2015b).

EP3 receptor is the most abundant receptor in brain (Leclerc et al., 2016) and appears to play an important role in acute ischemic stroke. A study showed that activation of EP3 receptor with ONO-AE-248 increased infarct size in experimental stroke (Ahmad et al., 2007). In the same context, the genetic deletion of EP3 receptor in a model of cerebral ischemia resulted in a reduction in cell death and infarction volumes induced by oxygen and glucose deprivation (Saleem et al., 2009). Deletion of EP3 receptor in mice suppressed damage to the blood-brain barrier, activation of microglia, and neutrophil infiltration into the ischemic cortex (Ikeda-Matsuo et al., 2011). In intracerebral hemorrhage, the genetic suppression of EP3 resulted in a decrease in intracerebral hemorrhage–induced brain damage and improved functional recovery. In addition, EP3 knockout mice showed a significant reduction in astrogliosis, microglial activation, blood-brain barrier degradation, and neutrophil infiltration. Overall, these studies suggested a detrimental role of the PGE₂-EP3–signaling axis in modulating brain injury, inflammation, and neurologic functional recovery (Leclerc et al., 2015a).

In a mouse model of cerebral ischemia, EP4 activation by ONO-AE1-329 reduced infarct volume, and deletion of EP4 exacerbated stroke injury (Liang et al., 2011). Similarly, treatment by an EP4 agonist, L-902,688, reduced infarct volume after ischemic stroke in mice and rats (Akram et al., 2013; DeMars et al., 2018). In humans, no difference of EP4-receptor expression was detected in blood of ischemic stroke patients versus asymptomatic patients by real-time PCR studies (Ferronato et al., 2011). Further studies are necessary to examine the roles of EP receptors and the potentially beneficial effects of agonists/antagonists in cerebral stroke.

The possible beneficial effect of PGI₂ in stroke was explored many years ago in humans. Accordingly, treatment with PGI₂ in patients diagnosed with ischemic stroke had a positive impact on stroke recovery with no neurologic deficit or minor residual hemiparesis (Gryglewski et al., 1983). Other clinical trials and studies in humans demonstrated that PGI₂ infusion showed beneficial effects after a stroke (Hsu et al., 1987). Moreover, PGI₂ agonists climprost (TTC-900) or TEI-7165 were described to have protective effects on postischemic neuronal damage in a gerbil model (Matsuda et al., 1997; Cui et al., 1999).

IP-receptor activation can attenuate anatomic and functional damage after ischemic stroke. The infarct volumes and neurologic deficit scores are significantly greater in IP knockout mice after both transient and permanent middle cerebral artery occlusion. Treatment with the IP-receptor agonists beraprost or MRE-269 before and after transient middle cerebral artery occlusion reduced the neurologic deficit score and infarct volume in WT mice (Saleem et al., 2010; Yang et al., 2017). Moreover, several studies performed on animals demonstrated beneficial roles of PGI₂ in cerebral blood flow (Bentzer et al., 2003; Lundblad et al., 2008). However, administration of PGI₂ did not change cerebral blood flow in human after subarachnoid hemorrhage (Rasmussen et al., 2015). On the other hand, in another study performed in patients with cerebral infarction, beraprost plus aspirin was found to be more effective than aspirin alone to reduce the recurrence of cerebral infarction or death (Chen et al., 2017). Moreover, an association between polymorphism of the IP-receptor gene and platelet activation was found in patients with cerebral infarction (Shimizu et al., 2013).

7. Arrhythmia

The study performed on rabbits demonstrated that PGE₂ prevented drug-induced torsade de pointes, which is life-threatening arrhythmia (Farkas and Coker, 2003). Antiarrhythmic effect of PGE₂ was also demonstrated in humans (Mest and Rausch, 1983). In contrast, microinjection of PGE₂ in rats or superfusion of the rat cardiac myocytes with PGE₂ resulted in tachycardia (Li et al., 1997; Zaretskaia et al., 2003), whereas 41% of women receiving misoprostol, EP2/EP3/EP4 agonist, had late decelerations or bradycardias (Kolderup et al., 1999). EP3 receptors located presynaptically on sympathetic nerve fibers supplying the heart of pithed rats strongly inhibit the neurogenic tachycardia. Sulprostone (EP3 > EP1), but not the IP/EP1-receptor agonist iloprost, inhibited the increase in electrically provoked heart rate dose-dependently. L-826266 (EP3-receptor antagonist) has no effect on basal heart rate or diastolic blood pressure but reduces the inhibitory effect of sulprostone (Kozłowska et al., 2012).

PGI₂ induced a marked reduction in the contraction rate of the rat cardiac myocytes and had a protective effect against the arrhythmias (Li et al., 1997). Similar effect was also observed in in vivo studies performed on rats and showed that low doses of PGI₂ reduced arrhythmias induced by coronary artery ligation or aconitine (Mest and Förster, 1978; Johnston et al., 1983). However, in humans, PGI₂ does not appear to have a cardiac antiarrhythmic effect and may increase the atrial and ventricular recurrent response. This effect could be related to an increase in adrenergic tone (Brembilla-Perrot et al., 1985). Increased excretion of 6-keto-PGF_1α has been observed in patients with ventricular arrhythmia (Chlewicka and Ignatowska-Switalska, 1992). There are very few studies regarding the role of IP and EP receptors in arrhythmia so far.

8. Pulmonary Circulation and Hypertension

In non-PH large pulmonary vessels (>2-mm diameter), activation of PGE₂ and PGI₂ receptors function in multiple ways to control vascular tone. Constriction of human pulmonary arteries induced by PGE₂ is mediated by activation of EP3 receptor (Qian et al., 1994; Norel et al., 2004a), whereas EP1-receptor activation mediates constriction of human pulmonary veins (Walch et al., 2001). Indeed, EP1 antagonists enhance iloprost and PGE₂-induced relaxation in human pulmonary veins and also underlie the contractions evoked by these two agents in the same tissue (Walch et al., 1999; Foudi et al., 2008). In large human pulmonary veins, relaxation to PGE₂ is mediated by the EP4 receptor (Foudi et al., 2008), whereas the EP2 agonist ONO-AE1-259 produces relaxation at relatively high concentrations, although a nonselective effect on the EP4 receptor is not excluded (Foudi et al., 2008). It is possible that the different effects of PGE₂ or EP-receptor agonists on pulmonary arteries as compared with veins are related either to differential expression of receptor subtypes or to differential coupling of receptors, such as the EP4 receptor (that can activate cAMP synthesis through G_s and also activate PI3K) (Hirata and Narumiya, 2012). On the other hand, IP-receptor agonists are well known to induce potent relaxation of pulmonary arteries (and veins) derived from human and rodent lungs (Haye-Legrand et al., 1987; Walch et al., 1999; Norel et al., 2004a,b; Benyahia et al., 2013, 2015). Apart from treprostinil, PGI₂ and other stable analogs (iloprost and beraprost) appear to be weaker venous rather than arterial dilators (Benyahia et al., 2013), likely to reflect the differential expression of prostanoid receptors involved in regulating tone in the lung. It should be noted that the functional roles of EP-receptor subtypes on pulmonary vascular tone are not well studied in resistance vessels, particularly in relation to human lung microvessels.

Contraction and thickening of arterial vascular wall in lung are characteristics of PH. The urinary excretion of the stable metabolite of PGI₂ (2,3-dinor-6-keto-PGF_1α) is decreased in patients with PH compared with control patients (Christman et al., 1992). In parallel with this observation, the density of PGIS detected by immunohistochemistry was lower in the pulmonary arterial endothelium of patients with severe PH compared with controls (Tuder et al., 1999). Recently, diminution of PGIS density and 6-keto-PGF_1α levels in pulmonary artery, pulmonary artery smooth muscle cells, and distal lung tissue derived from patients in PH group-III have been described (Ozen et al., 2020a). Furthermore, PGIS polymorphisms appeared to protect against the development of PAH in families known to harbor mutations that are strongly linked to the disease, suggesting that PGIS might act as a modifier gene influencing the penetrance in hereditary PAH (Stearman et al., 2014). Very recently, three rare loss-of-function PGIS variants were found in patients with idiopathic PAH, providing evidence that PGIS might be a susceptibility gene for PAH, possibly by causing endothelial apoptosis (Wang et al., 2020). Interestingly, patients with variants of the gene coding for PGIS (PTGIS) were more sensitive to the vasodilatory effects of iloprost, although the nature of such potentiation remains unknown.

Ex vivo studies performed on preparations derived from lung or pulmonary arteries have consistently shown not only a reduction of PGI₂ synthesis but also of the IP-receptor expression in both patients with PH and rats treated with monocrotaline and hypoxia that develop PH (Lai et al., 2008; Falcetti et al., 2010; Jiang et al., 2013; Li et al., 2018; Fan et al., 2019a; Clapp et al., 2020; Ozen et al., 2020a,b). The impairment of the PGI₂ pathway in PH lungs underlies the rationale for why the administration of vasodilators, such as PGI₂ or mimetics (IP agonists), are beneficial in the treatment of PH patients awaiting lung transplantation. This contrasts with EP4 and EP2 receptors, whose vascular expression was preserved or enhanced in human and experimental PAH, as confirmed by Western blot, real-time PCR, and immunohistochemical studies (Lai et al., 2008; Patel et al., 2018; Clapp et al., 2020). However, some of these IP agonists (epoprostenol, iloprost) have affinity for the EP1 receptor (Abramovitz et al., 2000; Whittle et al., 2012; Clapp and Gurung, 2015) with the potential of constriction of human pulmonary veins (Walch et al., 2001; Norel et al., 2004a). So far, there is no evidence for such a role of EP1 receptors in rodent pulmonary veins. Furthermore, some other IP agonists like treprostinil are also potent EP2 agonists (Whittle et al., 2012), which may combine with its potent activation of other prostanoid receptors (IP and DP1) to promote venodilation in pulmonary resistance vessels (Orie et al., 2013).

For these reasons, PGE₂ and activation of the EP receptors are also of interest in PH. PGE₂ concentrations in plasma were reduced in chronically hypoxic PH rats (Fan et al., 2019a), whereas in human pulmonary arteries exposed to hypoxia, increased levels of PGE₂ were detected (Yang et al., 2002). Recently, it has been shown that EP2-receptor expression levels were increased in human pulmonary artery SMCs and in the lungs derived from PAH patients (Patel et al., 2018; Clapp et al., 2020). This may not be surprising, given that EP2-receptor expression can be enhanced in response to PDGF and transforming growth factor-β (TGF-β), which are key drivers of SMC proliferation in PAH (Clapp et al., 2020). Based upon the high relative abundance of EP2 over IP (84-fold) and the fact that treprostinil has a 10-fold greater affinity at the EP2 receptor compared with the IP receptor (Whittle et al., 2012), it can be predicted that treprostinil will be more than two orders of magnitude (>800-fold) more active at EP2 versus IP receptors in PAH cells. This might help explain why activation of the EP2 receptor becomes a more prominent mechanism than the IP receptor to drive inhibition of pulmonary SMC proliferation by treprostinil. It should be noted that EP2 receptors are even more heavily expressed in adventitial fibroblasts in PAH with little evidence of IP-receptor expression (Clapp et al., 2020). EP2 receptors are known to have a range of inhibitory effects on fibroblast function, inhibiting migration, proliferation, and the transition of fibroblasts to myofibroblasts. These findings offer a new therapeutic perspective of how vascular wall thickening and fibrosis could be targeted via a signaling pathway that is robustly expressed in PAH.

It is important to understand the role of contractile prostanoid receptors because these could limit the doses of PGI₂ mimetics given therapeutically or potentially give rise to unwanted side effects. From studies conducted so far, it would appear that the EP3-receptor pathway is upregulated in PH, and this could occur for a number of reasons (Clapp et al., 2020). Increased expression of the EP3 receptor is found in human and mice pulmonary arteries after exposure to hypoxia by real-time PCR studies (Lu et al., 2015) in a monocrotaline model of PAH (Morrison et al., 2012) or in pulmonary arteries derived from PAH patients (Clapp et al., 2020). Furthermore, deletion of the EP3 receptor strongly inhibited the progression of PH induced by chronic hypoxia in rats. Similar results were obtained with the treatment of EP3-receptor antagonist, L-798106 (Lu et al., 2015). Likewise, increased sensitivity to the (EP3 > EP1) agonist sulprostone occurred in PAH arteries obtained from monocrotaline-treated rats, whereas beraprost (IP > EP3/TP agonist) caused contraction in distal human pulmonary arteries obtained from PAH patients with end-stage disease (Shen et al., 2019). Such studies suggest a gain of function of the EP3 receptor in PAH. This could be driven by EP3-TP–induced vasoconstrictive synergism, which has been described in numerous blood vessels, wherein priming with a TxA₂ mimetic or α1 receptor agonist markedly increases both the potency and size of contraction to EP3 agonists (Benyahia et al., 2015; Clapp and Gurung, 2015). The consequence of enhanced EP3-receptor activation would be to lower cAMP levels via its coupling to G_i, which would thus have the potential to counteract the effects of any IP-receptor agonist irrespective of whether they can directly activate these contractile receptors (Orie and Clapp, 2011). This may help to explain why vasorelaxation induced by PGI₂ analogs is consistently enhanced over the entire concentration range when EP3 receptors are inhibited (Orie and Clapp, 2011; Morrison et al., 2012). Overall, these findings suggest that EP3-receptor antagonists could be a therapeutic target for PH. Thus enhanced EP3-receptor expression together with IP-receptor downregulation may curtail the action of prostacyclin in PAH patients with severe disease or limit their therapeutic efficacy (Clapp et al., 2020).

EP4-receptor expression was not modified in pulmonary arteries or lungs derived from chronically hypoxic PH rats or monocrotaline-treated rats and patients with PAH; this was demonstrated by real-time PCR, Western blot, and immunohistochemical studies (Lai et al., 2008; Li et al., 2018; Fan et al., 2019a). However, our recent study demonstrated that there is a decrease of EP4-receptor expression in bronchi derived from PH group-III patients by Western blot and real-time PCR techniques (Ozen et al., 2020b). Another recent study performed on hypoxia-induced PH rats revealed that the beneficial effect of beraprost (PGI₂ analog) is mediated via the EP4 receptor–related pathway (Tian et al., 2019). In accordance with this finding, the EP4-receptor agonist, L-902,688, decreased PH right ventricular hypertrophy in hypoxic PH mice and monocrotaline-induced PH rats (Lai et al., 2018). In contrast, one study showed that during hypoxia, the vasoconstrictor effect of PGE₂ is mediated through the activation of the EP4 receptor on the rat intrapulmonary artery (Yan et al., 2013).

Finally, PH as well as parturition, abortion, or gastrointestinal ulcers are the only domains in which EP and IP receptors are therapeutic targets in clinical practice. Better knowledge of the prostanoid receptors involved and the selectivity and the potency of the compounds used in these clinical conditions is therefore of utmost importance. Most of our knowledge about PH and the development of pharmacological/therapeutic strategies has focused on PH group-I (PAH). More studies in PH patients from other groups are necessary; the recently published work with human bronchi (Ozen et al., 2020b) suggests that inhaled PGI₂ analogs may also have a promising therapeutic effect in PH group-III, which is one of the most common and lethal forms of PH. In PH group-III, PH is secondary to respiratory diseases, such as COPD, so in this case, inhaled PGI₂ could have a dual effect by decreasing airway resistance, thus supplying more of this vasorelaxant drug and oxygen to pulmonary arteries.

1. Expression of Prostaglandin E₂ Receptors in Platelets

A large number of studies demonstrated that human and mouse platelets express EP2, EP3, and EP4 receptors (Paul et al., 1998; Kuriyama et al., 2010; Hubertus et al., 2014; Pasterk et al., 2015). On the other hand, although Petrucci et al. (2011) showed the presence of the EP1 receptor in human platelets, other studies were not able to detect EP1 receptor in human and mouse platelets (Ma et al., 2001; Hubertus et al., 2014). Supporting these latter studies, the EP1-receptor agonist, ONO-DI-004, and the EP1-receptor antagonist, ONO-8713, did not alter human platelet aggregation (Iyú et al., 2010).

2. Prostaglandin E₂ and Prostaglandin E₂ Receptor 2

The action of PGE₂ to inhibit platelet aggregation at high concentrations is mediated by two receptors, namely EP2 and EP4. Both receptors are coupled to the G_s protein and increase the intracellular cAMP concentration. Real-time PCR (RT-PCR) and Southern blot analysis demonstrated that there is relatively low expression of EP2 receptor in human and mice platelets (Paul et al., 1998; Ma et al., 2001). Despite the low levels of EP2 receptors, the EP2-receptor agonists butaprost and ONO-AE1-259 inhibited platelet aggregation induced by the TP-receptor agonist U-46619 in human and mouse platelets to a similar extent (Kuriyama et al., 2010; Smith et al., 2010). This effect of the EP2-receptor agonist was absent in platelets of EP2-receptor knockout mice (Kuriyama et al., 2010).

3. Prostaglandin E₂ and Prostaglandin E₂ Receptor 3

Multiple isoforms of the EP3 receptor are present in human platelets, whose structures differ only by their carboxy-terminal tails responsible for the specificity for G-proteins (Kotani et al., 1995; Paul et al., 1998). EP3 has the potential to couple to G_s, G_i, or G_q proteins. In human platelets, four isoforms of the EP3 receptor (termed EP3-1b, EP3-II, EP3-III, and EP3-IV) were detected (Paul et al., 1998). EP3 splice variant distribution and function remain to be determined, but the overall effect of the EP3 receptor in human and mouse platelets is inhibition of cAMP production via G_i protein (Gray and Heptinstall, 1991; Ma et al., 2001). The activation of EP3 receptor potentiated platelet aggregation induced by different agents in both humans and mice (Philipose et al., 2010; Petrucci et al., 2011; Hubertus et al., 2014). These in vitro results were also confirmed in vivo using EP3 knockout mice, in which thrombotic responses to AA were decreased (Ma et al., 2001; Gross et al., 2007). Similar results were obtained in ferric chloride–induced thrombosis in mice (Gross et al., 2007).

To investigate the role of EP3 in thrombosis, EP3-receptor antagonist (DG-041) and agonists [sulprostone (EP3 > EP1 agonist), 17-phenyl trinor PGE₂ (EP1 agonist)] were used. EP3-receptor agonists increased platelet aggregation in humans and mice induced by different platelet agonists (Heptinstall et al., 2008; Pasterk et al., 2015; Theiler et al., 2016). Moreover, the (EP3 > EP1) receptor agonist, sulprostone, augmented the adhesion of human platelets to fibrinogen and collagen under low shear stress. This effect was prevented by the EP3-receptor antagonist L-798106 (Pasterk et al., 2015). Potentiation of PGE₂-induced platelet aggregation was inhibited by DG-041 in human, rat, and mouse platelets (Heptinstall et al., 2008; Singh et al., 2009; Smith et al., 2010; Tilly et al., 2014). In vivo studies performed on mice also demonstrated that DG-041 reduced thrombosis but had no effect on bleeding time (Tilly et al., 2014). In line with these findings, one clinical study performed on healthy volunteers demonstrated that DG-041 inhibited platelet function without increasing bleeding time (Fox et al., 2013). EP3-receptor antagonists could therefore be a novel therapeutic approach for the treatment of atherothrombosis without increasing the risk of hemorrhage, such as stroke or GI bleeding.

4. Prostaglandin E₂ and Prostaglandin E₂ Receptor 4

The inhibitory effect of PGE₂ on human platelet aggregation was abolished in the presence of MF-191, an EP4-receptor antagonist. Moreover, EP4-receptor agonist, ONO-AE1-329, inhibited platelet aggregation induced by the TP-receptor agonists U-46619, adenosin diphosphate, or collagen in human and mouse platelets. These data suggested that the inhibition of platelet aggregation by PGE₂ is mediated by EP4 receptor in humans (Philipose et al., 2010; Smith et al., 2010) and mice (Kuriyama et al., 2010). However, there is a greater inhibitory potency of ONO-AE1–329 in human platelets than in mouse platelets (Kuriyama et al., 2010; Mawhin et al., 2015). The PGE₂-induced inhibition of platelet aggregation was dramatically increased in EP3 and IP double-knockout mice, suggesting that EP2- and EP4-mediated inhibitory effect is augmented when the EP3 receptor is absent (Kuriyama et al., 2010). In both human and mouse platelets, the potentiating effect of PGE₂ on platelet aggregation via EP3 receptor is predominant over any inhibitory effects of EP2 and EP4 receptors (Gross et al., 2007; Iyú et al., 2010; Kuriyama et al., 2010; Hubertus et al., 2014). The inhibitory role of the EP4 receptor had probably been concealed for some time by the fact that it is the IP receptor rather than the EP2 or EP4 receptors that mediate the inhibitory effect of PGE₂ on mouse platelets, as revealed by studies using IP knockout mice (Fabre et al., 2001; Kuriyama et al., 2010).

1. Alzheimer Disease

Alzheimer disease (AD) is an age-related dementia that is not only characterized by β amyloid (Aβ) protein aggregation and accumulation but also by τ protein hyperphosphorylation in neurons (Guan and Wang, 2019). Neuroinflammation seems to play a critical role in the physiopathology of AD (Akiyama et al., 2000). Indeed, PGE₂ levels in CSF were increased in AD patients with mild memory impairment, whereas PGE₂ levels were decreased in end-stage AD patients (Combrinck et al., 2006). Interestingly, neuronal COX-2 expression was also reported to be downregulated in end-stage AD patients (Yermakova and O’Banion, 2001). mPGES-1 enzyme was present in neurons, microglia, and endothelium in human healthy brains, but its expression levels were upregulated in both mRNA and protein levels in AD patients (Chaudhry et al., 2008). The mPGES-1 gene disruption protected neurons from cytotoxic effect of Aβ 31–35 fragment in mice (Kuroki et al., 2012), indicating that mPGES-1-PGE₂ axis takes part in Aβ-eliciting harmful effects on neurons.

The role of EP1 receptor is still not clear in AD initiation and progression in humans. The genetic deletion of EP1 receptor decreased basal levels of Aβ in Swedish amyloid precursor protein (APPS)/presenilin-1 (PS1) mice model (APPS- and PS1-mutated AD model), and therefore, neurons in the absence of EP1 receptor are presumably more resistant to Aβ-induced toxicity (Zhen et al., 2012). The exacerbating role of EP1-receptor signaling was also suggested by an in vitro study: the blockade of EP1-receptor signaling with its antagonist (SC51089) in human neuroblastoma cell line (MC65) resulted in approximately 50% of reduction of Aβ-induced neurotoxicity (Li et al., 2013).

The genetic deletion of EP2 receptor in APPS mice decreased the protein levels of Aβ40 and Aβ42 as well as oxidative stress. Based on this study, an EP2-receptor signaling is considered to elicit proinflammatory and proamyloidogenic actions, at least in this AD model (Liang et al., 2005).

Another likely possibility is that EP2-receptor pathway may potentiate phagocytosis of Aβ42 by microglia in mouse AD model. Indeed, EP2 receptor has been shown to activate microglial function in other neurologic disorders, such as Parkinson Disease (PD) (see next paragraph). It was also reported in mouse primary neurons that pharmacological activation of EP2 or EP4 receptor rescued Aβ42-induced cell death in a cAMP-dependent manner (Echeverria et al., 2005). However, there is no strong evidence showing a protective role of EP2 pathway against AD in humans, and to date there is no ongoing clinical trial by targeting EP2 receptor for the treatments of AD patients (Cudaback et al., 2014).

Quantitative Western blot and immunohistochemistry analysis with human temporal cortex (postmortem tissues) demonstrated increased expression of EP3 receptor in mild cognitive impairment and a further increase in AD patients (Shi et al., 2012). These results were in accordance with the elevated expression of the EP3-receptor mRNA in hippocampus of mild-age APPS mice (AD model). The deletion of the EP3 receptor in these APPS mice decreased both Aβ40 and Aβ42 protein levels, with an attenuation of neuroinflammation and the reduction of proinflammatory gene expression, cytokine production, and oxidative stress (Shi et al., 2012). A recent study described a particular aspect of the EP3 pathway in APPS/PS1 mice; the treatment with EP3 > EP1 agonist, sulprostone, impaired synaptic plasticity of specific neurons in the hippocampus. Such a harmful effect of sulprostone was reversed by ONO-AE3-240, an EP3-receptor antagonist (Maingret et al., 2017). The studies on AD patients still remain to be evaluated for the use of EP3-receptor antagonists for clinical purposes (Cudaback et al., 2014).

EP4-receptor pathway involves the regulation of immune and proinflammatory responses in mouse AD models (Woodling and Andreasson, 2016). Indeed, the pharmacological suppression of EP4 receptor by using ONO-AE3-208 and also the genetic ablation of EP4 signaling lead to an improvement of cognitive function in mouse AD model (Hoshino et al., 2012). Nevertheless, EP4-targeted strategies have not been carried out clinically (Cudaback et al., 2014).

On the other hand, it was reported that PGI₂ counteracts PGE₂-induced IFN-γ production levels in mouse brain in an Aβ-dependent mechanism (Wang et al., 2016b). Interestingly, another group also demonstrated that PGI₂ ameliorated the exacerbating effect of PGE₂ on cognitive disorder in APPS/PS1 mice (Zheng et al., 2017). Overall, the studies investigating the role of IP and EP receptors in AD are mostly restricted to animal models, thus further studies are warranted to examine whether and how PGI₂ and PGE₂ are involved in the development and/or treatment of AD in humans.

2. Parkinson Disease

PD is characterized by a progressive loss of dopaminergic neurons in the substantia nigra. This complex and multifactorial disease could be explained by several molecular and cellular mechanisms, including neuroinflammation and microglial dysfunctions. The levels of PGE₂ were reported to be significantly elevated in the substantia nigra and other areas of PD patients’ brains (Mattammal et al., 1995).

EP1 receptor has been suggested to mediate PGE₂-induced neurotoxicity in rat dopaminergic neurons isolated from substantia nigra: EP1-receptor agonist 17-phenyl trinor PGE₂ induced toxic effects on dopaminergic neurons along with PGE₂, and 16-phenyl tetranor PGE₂ (stable analog of PGE₂) induced toxic effects on dopaminergic neurons, whereas EP2-receptor agonist butaprost and the (EP3 > EP1) receptor agonist sulprostone failed to exert any toxic actions (Carrasco et al., 2007). In rat microglia, an EP2-receptor signaling was reported to activate microglial functions in a cAMP-dependent manner, and thus, EP2 pathway appears to aggravate neuroinflammation. Indeed, EP2-receptor agonist, butaprost, as well as PGE₂ stimulated cAMP production in microglial cells, whereas EP1-receptor agonist, 17-phenyl trinor PGE₂, and EP4-receptor agonist, CAY10598, failed to elicit such actions. It is interesting that EP2-receptor activation exacerbated the rapid upregulation of mRNAs of proinflammatory genes, such as COX-2, IL-6, and iNOS (Quan et al., 2013). Kang et al. (2017) also found that COX-2–PGE₂-EP2-cAMP axis is involved in oxidopamine-induced neurotoxic effects in Neuro-2a (mouse neuroblastoma) and SH-SY5Y (originated from human bone marrow of a 4-year-old patient with neuroblastoma) cell lines, both of which mainly consist of dopaminergic neurons. Moreover, Johansson et al. (2013) demonstrated in an 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine–induced PD model that microglia-specific EP2 deficiency attenuates disease-induced microglial activation and prevents a loss in dopaminergic neurons in substantia nigra. In contrast to such an in vivo study, in rat primary in vitro midbrain cultured cells, an EP2-receptor pathway appears to exhibit a neuroprotective effect on dopaminergic neurons (Carrasco et al., 2008).

Although EP4 receptor is coupled to cAMP stimulation like EP2, the EP4-receptor pathway substantially works in a neuroprotective direction: in mouse in vivo model of PD, microglia-specific EP4 deficiency exacerbated microglial activation and T-cell infiltration in substantia nigra, and systemic administration of an EP4 agonist prevented a loss of dopaminergic neurons (Pradhan et al., 2017).

The exact role of PGI₂ in PD remains to be determined. However, it was reported that enforced PGI₂ synthesis by adenoviral gene transfer into substantia nigra prevents loss of dopaminergic neurons in oxidopamine-induced PD model. Based on this report, PGI₂ has a potential to exert a neuroprotective action in PD (Tsai et al., 2013). Further studies on the role of EP and IP receptor in patients with PD are necessary.

3. Huntington Disease

Huntington Disease (HD) is a neuropathology in which a genetic mutation can cause a large spectrum of symptoms, such as chorea and other motor disorders, but also cognitive disorders caused by an atrophy in basal ganglia (Anglada-Huguet et al., 2014).

In addition to the genetic etiology, the role of inflammation has been also described in the progression of HD. In fact, EP1-receptor antagonist, SC51089, has been shown to slow down the motor disorders and to ameliorate long-term memory decline in a mouse model of HD. Thus, the antagonism of EP1 receptor appears to improve many disorders in HD (Anglada-Huguet et al., 2014).

Another study from the same group demonstrated that an EP2/EP3/EP4 agonist, misoprostol, can also reduce memory decline in mouse model of HD. Authors concluded that among EP receptors, EP2 receptor promotes synaptic plasticity and delays neurodegeneration by stimulating the brain-derived neurotrophic factor expression (Anglada-Huguet et al., 2016). Despite these recent works, the impact and roles of IP/EP receptors in the pathophysiology of HD remain to be elucidated in humans also.

4. Multiple Sclerosis in Central Nervous System

MS is a chronic demyelinating disease of the CNS that leads to permanent cognitive and motor disabilities. MS is characterized by inflammation, oligodendrocyte loss, and axonal pathology. It was reported that both PGE₂ levels in CSF and COX-2 expression levels in demyelinating plaques are increased in MS patients, suggesting that COX-2–PGE₂ axis is involved in neuroinflammation (Mattsson et al., 2009; Palumbo et al., 2012). In addition, Kihara et al. (2009) demonstrated by using lipidomics that the main products of eicosanoid synthesis pathway in spinal cord are shifted from PGD₂ into PGE₂ in association with the onset of mouse MS model. In addition, they found that mPGES-1 deficiency can attenuate the symptoms of the disease, and there is a positive correlation between mPGES-1 immunoreactivity in microglia/macrophages and the severity of MS disease, indicating that mPGES-1-PGE₂ pathway plays a role in the progression of MS (Kihara et al., 2009). Kihara et al. (2009) proposed that at least EP1/EP2/EP4 receptors may participate in the PGE₂-elicited MS pathogenesis in various cell types, and the suppression of mPGES-1 activity by specific inhibitor may be a potential treatment against MS in humans.

In another mouse model of MS (induced by cuprizone), COX-2–PGE₂-EP2 axis has been proposed to play an important role by aggravating oligodendrocyte apoptosis during the onset of MS: AH6809 (EP1/2 and DP1 antagonist), reduced cuprizone-induced oligodendrocyte apoptosis, demyelination, neuroinflammation, and motor deficits. Indeed, the gene expression levels of EP receptors, including EP2, in brain were upregulated in the progressive stage of MS (after 5 weeks of cuprizone treatment) (Palumbo et al., 2012). In this MS model, PGI₂ level in the CSF is likely to be unchanged upon cuprizone administration (Palumbo et al., 2012). Interestingly, it was reported in mouse model of MS that a stable IP agonist, iloprost, suppresses demyelination and motor dysfunction, indicating that an IP-receptor agonist has a potential to prevent (or alleviate) symptoms of MS (Muramatsu et al., 2015). However, there are no data available on the role of IP receptor in humans with MS.

5. Amyotrophic Lateral Sclerosis

Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disease primarily involving motor neurons and characterized by several molecular and cellular dysfunctions, including oxidative stress, apoptosis, neuroinflammation, and glutamate toxicity. This disease has been shown to be closely associated with upregulation of proinflammatory pathway; PGE₂ levels in CSF, brain, and plasma were all increased in ALS patients (Almer et al., 2002; Iłzecka, 2003).

EP2 receptor has been shown to participate in neuroinflammation and progression of ALS in mouse model of familial ALS containing G93A-mutation in superoxide dismutase gene. The EP2-receptor deficiency improves motor strength and extends survival in association with systemic reductions in the levels of proinflammatory effectors, such as iNOS and NADPH oxidase, suggesting that suppression of EP2-receptor signaling may be a novel strategy for the treatment of ALS (Liang et al., 2008). It should be noted that the EP2-receptor signaling can induce the motor-neuron–like cell death in an in vitro system (Miyagishi et al., 2013).

The mRNA expression levels of alternatively spliced isoforms of EP3 receptor were characterized during the pathogenesis of ALS; EP3α and EP3γ mRNAs were detected in the WT lumbar spinal cord, but EP3β mRNA was undetectable. When the authors analyzed motor neurons dissected out of spinal cord, EP3γ mRNA was predominantly detected in motor neurons, whereas EP3α and EP3β mRNAs were undetectable. Such an EP3γ-biased expression in motor neurons was unchanged in ALS-model (G93A-mutated in superoxide dismutase gene) mice (Kosuge et al., 2015).

Bilak et al. (2004) reported that the EP2-receptor agonist butaprost as well as the EP3 > EP1 agonist sulprostone exerted a neuroprotective effect on motor neurons in slice culture model of ALS. It remains to be elucidated why G_s-coupled EP2 and G_i-coupled EP3 pathways share similar protective actions in chronic glutamate–induced ALS model.

Interestingly, it was reported that administration of the IP-receptor agonist, ONO-1301-MS, to ALS-model mice attenuates the expression of hypoxia-inducible factor 1α, which is a hypoxia marker known to be elevated in the spinal cord of mouse model of ALS and ALS patients (Tada et al., 2019). There has been no study yet on the role of PGI₂ in humans with ALS. Moreover, although PGE₂ levels have been shown to increase in patients with ALS, further studies are necessary to confirm the contribution of each EP-receptor subtype in these patients.

1. Prostaglandin E₂ and Prostacyclin Receptors in Peripheral Nervous System/Dorsal Root Ganglion Neurons

Aspirin and other NSAIDs are used for several pharmaceutical actions, especially as analgesics. Inflammation, caused by injury or other reasons, could be then perceived by the brain and all other systems as a signal of danger and/or attention through pain. PGE₂, through its EP1–4 receptors, and PGI₂, through its IP receptor, are considered to be pain mediators.

Dorsal root ganglion (DRG) neurons play an important role in the pain perception (i.e., nociception) and pain transmission. IP, EP1, EP3, and EP4 receptors were detected in DRG neurons of mice with in situ hybridization studies (Sugimoto et al., 1994; Oida et al., 1995). Another study also demonstrated that EP1, EP2, and EP4 receptors and only EP3γ isoform (but not EP3α or EP3β) were expressed in rat dissociated sensory DRG neurons with PCR detection (Southall and Vasko, 2001).

Several pharmacological studies focused on the roles of EP and IP receptors in DRG neurons, indicating that only IP and EP4 receptors are involved in cAMP production. An IP-receptor agonist, cicaprost, and EP4-receptor agonist, ONO-AE1-329, increased cAMP levels in rat DRG neurons in a concentration-dependent manner. On the other hand, ONO-DI-004 (EP1-receptor agonist), ONO-AE1-259 (EP2 agonist), sulprostone (EP3 > EP1 agonist), and ONO-AE-248 (EP3 agonist) were unable to alter cAMP levels, suggesting that PGE₂-induced cAMP production appears to be mediated by EP4 receptor without being affected by EP3 receptors (Wise, 2006).

2. Roles of Prostaglandin E₂ Receptor 3 in Pain Perception

Several studies have been performed to clarify the role of EP3 receptor in pain perception. Pharmacological stimulation with the EP3-receptor agonist, ONO-AE-248, resulted in antinociceptive effect in rat models of joint inflammation (Natura et al., 2013). Another study demonstrated in a mechanical nerve ligation-induced neuropathic pain model that among the four EP receptors, only EP3 deficiency attenuated nociceptive behavior and mechanical allodynia. Recently, EP3-induced CCL2 chemokine release has been proposed as a possible mechanism underlying EP3-induced neuropathic pain (Treutlein et al., 2018). Thus, EP3-receptor antagonist could be a therapeutic target to reduce chronic neuropathic pain, and studies performed in humans with neuropathic pain are necessary.

3. Roles of Prostaglandin E₂ Receptor 1, Prostaglandin E₂ Receptor 2, Prostaglandin E₂ Receptor 4, and Prostacyclin Receptor in Pain Perception

Apparently, EP1 receptor is involved in pain perception through its major role in peripheral nervous system (PNS) but not in CNS. The studies using EP1 knockout mice revealed that EP1 receptor is implicated in the perception of inflammation induced-heat/pain sensitization in PNS (Johansson et al., 2011). Moriyama et al. (2005) demonstrated that EP1 deficiency attenuates the activation efficiency of TRPV1, a nonselective cation channel, which is expressed in sensory neurons and activated by various noxious stimuli, such as heat, proton, and pepper constituent. Moreover, EP1-receptor antagonist, ONO-8713, mimicked the effect of EP1 deficiency, whereas an EP1-receptor agonist, ONO-DI-004, exacerbated capsaicin-induced TRPV1 activation in mice DRG neurons. The mechanisms underlying PGE₂-induced TRPV1 potentiation are likely different between human and mouse DRG neurons. PGE₂-induced TRPV1 potentiation was suppressed by the activation of metabotropic glutamate receptors 2 and 3 in mouse but not human DRG neurons (Sheahan et al., 2018). The inhibition of pain perception by selective EP1 antagonist remains to be confirmed in human models.

Inflammatory pain occurs in endometriosis. In a preclinical mouse model of endometriosis, EP2 receptor has been shown to mediate peripheral and central hyperalgesia. Real-time PCR studies demonstrated that the expression level of EP2/EP4 receptors and also COX-2 were significantly increased in endometriosis lesions compared with controls. An EP2-receptor antagonist, PF-04418948, was the most efficiently analgesic rather than EP4- or TRPV1- antagonist, since this drug showed suppressive effects on both peripheral and central hyperalgesia (Greaves et al., 2017). Using an chimeric endometriosis model in which human endometriotic cells were xenografted into nude mice, Arosh et al. (2015) demonstrated that selective inhibition of both EP2 and EP4 receptors suppresses proinflammatory state of DRG neurons and attenuates pelvic pain in endometriosis. Lin et al. (2006) reported that EP4 receptor plays pivotal roles in nociception and promotes inflammatory pain hypersensitivity, at least in mouse model. Indeed, they found that EP4-receptor expression is increased both in mRNA and protein levels in DRG neurons upon peripheral inflammation (and EP1-3 expression levels are unchanged) (Lin et al., 2006). The use of EP4-receptor antagonists, such as MF498 or AH23848, was associated with pain reduction, and hence, these drugs had analgesic effects in murine inflammation models (Lin et al., 2006; Clark et al., 2008). These analgesic actions of EP4-receptor antagonists have already been verified in humans (Jin et al., 2018), and these drugs may serve for the treatment of pain-associated inflammatory diseases, such as rheumatic disease or osteoarthritis. Indeed, Grapiprant, an EP4 antagonist, is already approved for the treatment of pain and inflammation in osteoarthritis in dogs (Rausch-Derra et al., 2016).

The IP receptor is highly involved in nociception, hyperalgesia, and inflammation. The effect of IP-receptor activation on the sensitization of rat sensory neurons is mediated by stimulation of adenylyl cyclase and phospholipase C in sensory neurons (Smith et al., 1998). Furthermore, the involvement of IP receptors in inflammatory pain was addressed by using IP-receptor knockout mice and acetic acid–induced writhing test (Murata et al., 1997; Bley et al., 1998). The observed effect was also confirmed by using two selective IP antagonists, RO1138452 (CAY10441) and RO3244794 (Bley et al., 2006). These antagonists were tested in a rat model of nociception; both antagonists inhibited carrageenan-induced mechanical hyperalgesia and edema formation (Bley et al., 2006). Furthermore, in a rat model of neuropathic pain, a stable PGI₂ analog (carbaprostacyclin) increased the neuronal activities in DRG and dorsal horn in a dose-dependent manner (Omana-Zapata and Bley, 2001).

The inhibition of PGI₂-IP pathway in rodent models of hyperalgesia and chronic arthritis showed a significant reduction of pain and associated inflammation. In this study, the injection of a prostacyclin analog, beraprost (IP > EP3/TP agonist), dose-dependently induced hyperalgesia, and such an effect was abolished by the simultaneous administration of IP-receptor antagonist (N-[4-(imidazo- lidin-2-ylideneamino)-benzyl]-4-methoxy-benzamide) (Pulichino et al., 2006). In contrast, a 7-day clinical use of another prostacyclin analog, iloprost (IP = EP1 > EP3 agonist), in rheumatoid arthritis patients demonstrated that iloprost has anti-inflammatory and analgesic properties (Gao et al., 2002). Another study reported on the analgesic effect of iloprost being similar to the use of tramadol, a powerful analgesic, in a clinical trial for patients with arthritis (Mayerhoefer et al., 2007). Such different efficacies of beraprost and iloprost could be explained by species difference and/or the different specificity profiles between iloprost (IP = EP1 > EP3) and beraprost (IP > EP3/TP) (Whittle et al., 2012; Alexander et al., 2019).

In patients with pulmonary hypertension, so far all IP-receptor agonists produce adverse events related to pain, including site pain, jaw pain, flushing, headache, and extremity pain, suggesting a key role of the IP receptor in mediating pain (Picken et al., 2019). Given that the highest odds ratio for jaw pain from meta-analysis was seen with beraprost (Picken et al., 2019), this might indicate involvement of EP3 in the joint pain. Transitioning from epoprostenol (synthetic PGI₂) to subcutaneous treprostinil increases site pain (Rubenfire et al., 2007), again suggesting additional receptors contributing—probably EP2 and DP1 because they are both expressed in the skin, and EP2 is expressed in the dorsal horn—which receive sensory information from primary afferents. One of the mechanisms underlying PGI₂-elicited pain modulation is explained by the potentiation of the TRPV1 receptor in mouse DRG neurons (Moriyama et al., 2005). Alternatively, as suggested for the role of PGE₂, PGI₂-IP signaling is also likely to sensitize glutamate pathway by inducing phosphorylation and translocation of GluR1 receptor in mouse DRG neurons in zymosan-induced mechanical hyperalgesia model (Schuh et al., 2014).

1. Water and Salt Regulation

It is well known that PGE₂ is the most important prostanoid in regulating water and solutes balance. Although COX-1 is constitutively expressed in the kidney, mice deficient in COX-1 appear to be healthy with no obvious renal defects. In contrast, COX-2 seems to play a more important role in regulating renal water transport (Li et al., 2017d). In rats, the EP3 receptor is reported to regulate water excretion in response to high salt intake; it decreases collecting duct-water permeability and increases water excretion. High-salt treatment increased COX-2–dependent PGE₂ production when the EP3 receptor was blocked by L-798106 in the thick ascending limb, whereas urine output was decreased when the EP3 receptor was activated by sulprostone (EP3 > EP1 agonist) (Hao et al., 2016). EP2 and EP4 receptors are reported to bypass vasopressin signaling and increase water reabsorption (Olesen and Fenton, 2013). In rodents, it was also shown that PGE₂ inhibits AC and NaCl reabsorption in thick ascending loop of Henle (Schlondorff and Ardaillou, 1986). By radioligand membrane binding and autoradiography, the localization of [³H]PGE₂ was demonstrated in proximal tubule as well as the glomeruli of human kidney, a distribution that is in accordance with the assumed site of action for the salt and water regulatory function of PGE₂ (Eriksson et al., 1990). However, mechanistic studies have been confined to animal experiments, and these findings need to be examined in humans.

2. Renin Release

PGE₂ and PGI₂ stimulate renin secretion and renin gene expression by activating cAMP formation in human juxtaglomerular cells (Wagner et al., 1998) probably through activation of the EP4 receptor (Kaminska et al., 2014). The same mechanism was demonstrated in mice (Jensen et al., 1996; Wang et al., 2016a).

According to Hao and Breyer (2007), in nephrotic syndrome, maintenance of normal renal function in human becomes dependent on COX-2–derived prostanoids, particularly PGE₂ and PGI₂, with an active participation of EP4, EP2, and IP receptors that mediate their vasodilator effect. In an equivalent rat model, a selective EP4 antagonist (L-161982) exacerbated proteinuria and glomerular cell apoptosis (Aoudjit et al., 2006). More specifically, and in diabetic nephropathy, the involvement of EP1 (Kennedy et al., 2007) and EP3 receptors (Hassouneh et al., 2016) in mediating disease progression has demonstrated that arginine vasopressin–mediated water reabsorption was reduced in sulprostone-treated or EP₁^−/− rats. These pathophysiologic effects were not proven in human. On the other hand, elevation of COX-2 and PGE₂ expression is the main feature of renal cell carcinoma in both human (Kaminska et al., 2014) and rat (Rehman et al., 2013).

1. Mucosal Protection

In rodents, endogenous PGI₂ and PGE₂ are constitutively produced in the stomach through constitutive COX-1. These two PGs reduce stomach acid secretion, activate mucosal blood flow, and facilitate the release of viscous mucus (Amagase et al., 2014; Takeuchi and Amagase, 2018). In human, the protective effect of PGE₂ analogs, such as 16,16-dimethyl PGE₂ (EP3 > EP2/EP4 agonist) against gastric-acid secretion and gastric-ulcer formation, was shown in the late 1960s (Robert et al., 1968, 1974). Further pharmacological results obtained in human and dogs proved the involvement of the EP3 receptor in inhibition of gastric-acid secretion (Robert et al., 1974; Tsai et al., 1995). In mice, the inhibitory action of prostaglandins (i.e., PGE₂ and PGI₂) on gastric-acid production in the damaged stomach by taurocholate sodium is mediated by activation of EP3 and IP receptors (Nishio et al., 2007; Takeuchi and Amagase, 2018). Similarly, EP3-receptor activation in rat is responsible for a reduced production of gastric acid induced by pentagastrin/histamine stimulations (Kato et al., 2005). PGI₂ and PGE₂ play crucial roles in gastric mucosal defense during induced cold-restraint stress through IP and EP4, respectively. Gastric lesions induced by 18 hours of cold-restraint stress are significantly increased in IP knockout mice compared with WT mice. In WT mice, pretreatment with iloprost and indomethacin in combination prevented gastric lesions caused by cold-restraint stress (Amagase et al., 2014).

NSAIDs are well known to be responsible for gastric lesions and stomach injuries, and PGE₂ can prevent and reverse these effects (Johansson et al., 1980). Gastric damage induced by indomethacin or by HCl/ethanol in rat could be prevented by PGE₂ or EP1-receptor agonists [17-phenyl trinor PGE₂ (EP1 agonist) and sulprostone (EP3 > EP1 agonist)], and this protection disappears in EP1 knockout mice (Araki et al., 2000; Suzuki et al., 2001; Takeuchi, 2012, 2014). This EP1-mediated effect is attributed to the inhibition of gastric hypermotility induced by NSAIDs (Takeuchi et al., 1986); most of these experiments have been done with rodents, so these receptor subtype roles in the gastrointestinal tract need to be confirmed in human. On the other hand, the clinical perspective to use EP1-receptor agonists should be limited to treat NSAID-induced damage, to stimulate bicarbonate production in the stomach, and to reduce acid reflux in esophagus (Takeuchi et al., 1997; Suzuki et al., 2001; Takeuchi and Amagase, 2018). Surprisingly, in a recent clinical study (20 patients), the EP1 antagonist ONO-8539 had a positive effect on acid-induced heartburn in healthy male subjects with gastroesophageal reflux disease (Kondo et al., 2017). Because this antagonist would probably act on symptoms or sensorial responses and not on the endogenous cause of acid reflux in esophagus, further investigations are necessary.

Mucus and bicarbonate secretions by epithelial cells are further physiologic mechanisms involved in preventing/healing gastric lesions. Indomethacin-induced gastric-mucosa lesions could be prevented by the administration of misoprostol (EP2/EP3/EP4 agonist), resulting in a lower edema average in gastric mucosa (Cavallini et al., 2006). Furthermore, administration of an EP4-selective agonist also significantly reduced indomethacin-induced apoptosis of human gastric-mucous epithelial cells (Jiang et al., 2009). These results suggest that EP4-activating reagents may be used to prevent NSAID-induced ulcers by maintaining mucous epithelial-cell survival.

Globally, gastrointestinal cytoprotection induced by selective EP3/EP4-receptor agonists could be very promising. On the one hand, these agonists may reduce gastric-acid secretion and mucosal inflammation (e.g., in colitis), and on the other hand, they increase mucus and bicarbonate productions (Robert et al., 1974; Takeuchi et al., 1997; Kato et al., 2005; Larsen et al., 2005; Takeuchi and Amagase, 2018). For this reason, misoprostol (Cytotec, the EP2/EP3/EP4 agonist), with its mucoprotective and antiacid properties, is already an effective treatment of gastrointestinal injury in clinic, which has been shown to be more effective than omeprazole (a proton pump inhibitor) (Taha et al., 2018; Kim et al., 2019). The effect of misoprostol could be probably through activation of the EP4 receptor. PGE₂-EP4 signaling has also a protective role in colon mucosal barrier in human and murine models. In EP4 knockout mice (EP4⁻/⁻) treated with dextran sodium sulfate (DSS), a loss of the colon epithelial barrier and the accumulation of neutrophils and CD4+ T cells in the colon were observed (Kabashima et al., 2002). In WT mice, the use of selective EP4-receptor antagonist (ONO-AE3-208) led to the development of severe DSS-induced colitis (Kabashima et al., 2002). However, the administration of selective EP4-receptor agonist (ONO-AE-734) to wild-type mice suppressed DSS-induced colitis (Kabashima et al., 2002). In human, the use of selective EP4-receptor agonist rivenprost (ONO-4819) in patients with mild-moderate ulcerative colitis significantly improved histologic scores and reduced the disease activity index after 2 weeks of therapy (Nakase et al., 2010).

2. Gastrointestinal Motility and Muscular Tone

Gastric cytoprotection and lesions are also associated with gastrointestinal motility (contraction/frequency of contraction) (Suzuki et al., 2001). In vivo, inhibition of PGE₂ and PGI₂ synthesis by indomethacin and many other NSAIDs in rat stomach or intestine cause increased motility (Takeuchi et al., 1986; Takeuchi, 2012). This increase is reversed by exogenous PGE₂ and by selective EP1 agonist and EP4 agonists when motility is measured in rat stomach and intestine, respectively (Suzuki et al., 2001; Kunikata et al., 2002; Takeuchi and Amagase, 2018). The mechanism associated with this NSAID increased motility in vivo is still unknown, despite that it was suggested to a central vagal stimulation (Yokotani et al., 1996). In contrast, many other in vitro studies using isolated gastrointestinal preparations (longitudinal muscle) have shown a contractile role for exogenous PGE₂ in human stomach and colon (Bassil et al., 2008; Fairbrother et al., 2011), rat gastric fundus and colon (Abraham et al., 1980; Sametz et al., 2000; Bassil et al., 2008; Iizuka et al., 2014), guinea-pig ileum and fundus (Coleman and Kennedy, 1985; Sametz et al., 2000), and mice ileum (Fairbrother et al., 2011). Numerous in vitro pharmacological studies using PGE₂ or selective agonists (e.g., ONO-DI-004 for EP1, ONO-AE-248 for EP3, sulprostone for EP3 > EP1) and antagonists (e.g., EP1A, ONO-8713, SC51089, and SC19220 for EP1, ONO-AE3-240 and L-798106 for EP3) have determined that EP1-receptor activation is responsible for contractions of human longitudinal muscle, whereas in rodent gastrointestinal muscles it is EP1- and EP3-receptor activations that are responsible (Coleman and Kennedy, 1985; Sametz et al., 2000; Fairbrother et al., 2011; Iizuka et al., 2014). This difference of effects induced by PGE₂ (or mimetics) on motility versus contraction of gastrointestinal tract in vitro, dependent or independent of neuronal (central) component, and the difference between species need more experiments.

One possible clinical application of this EP1 receptor on motility in humans is illustrated by lubiprostone (a PGE₁ derivate and chloride channel type 2 channel opener) as a treatment of constipation. On rat and human stomach longitudinal muscle and also on mouse intestinal circular muscle, lubiprostone has a contractile effect mediated by activation of EP1 receptor. However, EP4-receptor antagonists reduced the contractile action of lubiprostone on circular colon muscle in rat and human (Bassil et al., 2008; Chan and Mashimo, 2013).

In human, the pronounced contractile effect of PGI₂ analogs on isolated gastric smooth muscle may contribute to explaining abdominal pain and cramping associated with the use of these compounds in clinic. However, selexipag and its active metabolite MRE-269 (highly selective IP agonist) had few gastric side effects (Morrison et al., 2010). In contrast, other PGI₂ analogs, like iloprost and beraprost that have some affinity for EP3 and EP1 receptors, induced dose-dependent contraction of rat gastric fundus. Furthermore, the contraction to iloprost and beraprost was inhibited by an EP3-receptor antagonist ((2E)-3-(3Ј,4Ј-dichlorobiphenyl-2-yl)-N-(2-thienylsulfonyl)acrylamide) and EP1-receptor antagonists (SC19220 and SC51322), suggesting that EP1 receptors contributed to contraction of gastric fundus to iloprost and beraprost (Morrison et al., 2010).

1. Effect of Nonsteroidal Anti-Inflammatory Drugs

In the in vivo situation, inhibition/deletion of COX-2 reduces pain; however, this enzyme has been suggested to be responsible for the loss of bone mass or mineral density both in humans and rodents (Robertson et al., 2006; Blackwell et al., 2010; Nakata et al., 2018). Many studies have shown that in humans and rodents treated with NSAIDs as well as in mice lacking COX-2 gene the bone-healing process after fracture is impaired (Pountos et al., 2012; Vuolteenaho et al., 2008). In a similar way, other anti-inflammatory drugs, such as glucocorticosteroids (e.g., dexamethasone), are responsible for impaired osteogenesis and could account for osteoporosis induction (Weinstein, 2012; Whittier and Saag, 2016). Treatment of human bone-marrow stromal cells (BMSCs) in vitro with dexamethasone induces the expression of the EP2 and EP4 receptors. In this case, osteoblast differentiation from human BMSCs is impaired, and PGE₂ promotes adipogenesis instead of osteogenesis (Noack et al., 2015; Mirsaidi et al., 2017). Such a differentiation-controlling effect of PGE₂ on BMSC differentiation was not observed in mouse cells; the adipocyte differentiation is inhibited by PGE₂ (Inazumi et al., 2011; Fujimori et al., 2012). Interestingly, the beneficit of using NSAIDs to suppress the pathologic bone growth and heterotopic ossification has been shown in human and rat model (Zhang et al., 2013b; Lisowska et al., 2018b).