Targeting the COX-2/PGE₂ axis to enhance NK and T cell immunotherapy in brain tumors

Abstract

Aggressive brain tumors such as glioblastoma (GBM) remain among the most lethal human cancers, with a median survival of only 15 months despite multimodal treatment. Their resistance arises from a triad of barriers—the blood–brain barrier (BBB), marked intratumoral heterogeneity, and a profoundly immunosuppressive tumor microenvironment (TME). Immunotherapeutic strategies based on natural killer (NK) and T cells, leveraging antigen-independent cytotoxicity and antigen-specific precision, respectively, offer potential breakthroughs but are often limited by chronic neuroinflammation. A key driver of TME suppression is prostaglandin E2 (PGE2), produced via the cyclooxygenase-2 (COX-2) pathway. PGE2 exerts a dual role: Intracellularly, it can promote apoptosis, whereas extracellularly, it fosters tumor progression, immune evasion, and therapeutic resistance. Through activation of EP2 and EP4 receptors, PGE2 signals via Gαs proteins to elevate cyclic adenosine monophosphate (cAMP), leading to impaired cytotoxic immunity. This signaling downregulates NK cell activating receptors (e.g., NKG2D, NKp30), induces CD8⁺ T cell exhaustion, and promotes regulatory T cell expansion. The COX-2/PGE₂ axis further mediates resistance to checkpoint inhibitors, CAR-T therapy, and chemotherapy by enhancing neuronal excitation through EP1 receptor activation in GBM. Targeting this pathway has therefore emerged as a compelling therapeutic strategy, which can restore NK and T cell function and sensitize tumors to immunotherapy. Combining PGE₂ modulation with next-generation NK/T cell approaches—including CAR-NK and CAR-T platforms—holds promise to overcome immune resistance and redefine therapeutic paradigms for GBM and other central nervous system malignancies.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s00262-026-04372-5.

Introduction

Brain tumors, particularly glioblastoma (GBM), remain among the most lethal cancer. Despite surgery, radiotherapy, and chemotherapy, median survival is around 15 months. Treatment failure reflects three major barriers: restricted drug delivery across the blood–brain barrier (BBB), marked tumor heterogeneity driving resistance [1], and a profoundly immunosuppressive tumor microenvironment (TME) [2]. Natural killer (NK) and T cell-based therapies have emerged as promising strategies to overcome these limitations, representing a conceptual advance beyond conventional treatments [3].

NK cells rapidly eliminate stressed or MHC-deficient tumor cells through perforin and granzyme release [4], whereas engineered T cells—such as chimeric antigen receptor (CAR)-modified cells targeting EGFRvIII or IL13Rα2—provide antigen-specific cytotoxicity [5]. Early clinical studies established feasibility and safety [6], while later trials demonstrated tumor regression but highlighted challenges including antigen escape [7]. Recent approaches aim to improve persistence and trafficking while overcoming TME- and BBB-associated constraints [3].

A key mediator of immune suppression in GBM is prostaglandin E2 (PGE2). Produced via the cyclooxygenase-2 (COX-2) pathway, PGE2 dampens NK and T cell activity through prostaglandin E2 receptor 2 and 4 (EP2/EP4) signaling, while promoting macrophage polarization, angiogenesis, and stem-like features [8]. Although mechanistic insights have expanded, critical gaps remain regarding receptor-specific effects, subset-selective immune regulation, and therapy resistance [9].

Given that 5-year survival in GBM is below 5% [10], NK and T cell therapies offer significant clinical potential. While NK cells provide broad cytotoxicity and safety, T cells offer antigen specificity and memory [11]. Integrative strategies that counter prostaglandin-mediated suppression may enhance efficacy [12, 13]. This review summarizes recent advances and outlines future directions for combining immune-based therapies with anti-prostaglandin approaches in brain tumors.

The role of the PGE₂ axis in tumor immunotherapy

Neuroinflammation and the PGE₂ axis in brain tumors

Neuroinflammation is characterized by immune activation within the central nervous system (CNS). In particular, acute activation of microglia and astrocytes via nuclear factor kappa-B (NF-κB) and janus kinase and signal transducer and activator of transcription (JAK/STAT) enhances cytokine and chemokine release, promotes NK cytotoxicity, and facilitates CD8⁺ T cell priming and BBB trafficking. This phase temporarily strengthens antitumor immunity across GBM, medulloblastoma, meningioma, diffuse midline glioma (DMG), and diffuse intrinsic pontine glioma (DIPG) [14–22].

Chronic neuroinflammation: immunosuppressive transition

Persistent inflammation shifts the TME toward immune evasion. Tumor-derived interleukin-6 (IL-6), transforming growth factor-beta (TGF-β), colony-stimulating factor-1 (CSF-1), and PGE2-driven M2 microglia polarization and the subsequent transition of astrocytes toward an immunosuppressive phenotype collectively promote angiogenesis and tumor growth by secreting interleukin-10 (IL-10), vascular endothelial growth factor (VEGF), and TGF-β [23–25]. In GBM, regulatory T cells (Tregs) and myeloid-derived suppressor cells (MDSCs) expansion, angiogenesis, and effector exhaustion limit the efficacy of NK-, T-, and CAR-based therapies [26, 27]. This chronic state highlights the need to strategically modulate neuroinflammation to restore antitumor immunity.

PGE2 as a central immune checkpoint

Neuroinflammation profoundly influences NK and T cell functions within the TME of brain tumors. Extracellular PGE2 suppresses NK cells by reducing NKG2D and NKp30 and impairs T cell function through EP2/EP4–cAMP signaling, decreasing IFN-γ and promoting exhaustion (Fig. 1) [28–30]. It also enhances PD-L1 expression and reinforces myeloid-mediated suppression [31]. Although intracellular PGE2 may induce apoptosis under hypoxia, rapid export shifts its net effect toward immune evasion [8, 32–35].

Fig. 1.

Schematic representation of the PGE2-mediated inhibition of NK cells. PGE2, released by tumor cells, binds to the EP4 receptor on NK cells, increasing cyclic adenosine monophosphate (cAMP) levels and suppressing migration, cytotoxicity, and cytokine production. EPs antagonists block this suppression, potentially restoring NK cell functions and boosting antitumor immunity

Therapeutic implications: targeting the PGE2 axis

As highlighted in Table 1, major obstacles include BBB restriction, an immunosuppressive microenvironment enriched in TGF-β/IL-10/PGE₂, regulatory immune dominance, checkpoint signaling, and antigenic heterogeneity. Most NK and CAR-T trials therefore remain early phase with limited durability [36–41]. Chronic PGE2-driven inflammation intersects with these barriers by suppressing effector receptors, sustaining myeloid polarization, and reinforcing checkpoint pathways. Targeting COX-2, mPGES-1, EP2/EP4, or PGE2 transport directly addresses this upstream inflammatory node. Integrating PGE2 axis inhibitors with NK, CAR-T, or CAR-NK platforms offers a rational combinatorial approach. By restoring effector competence and reprogramming the suppressive niche, PGE2 blockade may enhance persistence, cytotoxicity, and therapeutic durability in refractory CNS tumors.

Table 1.

Clinical trials involving NK and T cell therapies for brain tumors

Trial ID	Trial name	Phase	Type	Condition(s)	Reference	Main findings
NCT02271711	Expanded natural killer cell infusion in treating younger patients with recurrent/refractory brain tumors	I	NK cells	Recurrent/refractory medulloblastoma, AT/RT, ependymoma	[42]	Proved that repeated intraventricular NK cell infusions are safe and feasible in children with recurrent brain tumors, with one patient achieving stable disease
NCT02100891	Phase II STIR trial: haploidentical transplant and donor natural Killer cells for solid tumors (STIR)	II	NK cells	Solid tumors, including high-risk malignant brain tumors (recurrent/refractory)	[43]	Demonstrated a 72% disease control rate and 64% one-year survival using donor NK cells after haploidentical transplants for solid tumors
NCT02130869	A pilot study of immunotherapy including haploidentical NK cell infusion following CD133 + positively selected autologous hematopoietic stem cells in children with high-risk solid tumors and lymphomas	I	NK cells	High-risk brain tumors (among others)	[44]	Confirmed the safety and feasibility of combining autologous stem cell transplants with haploidentical NK cell infusions for high-risk pediatric tumors
NCT01588769	Tolerability and efficacy of ALECSAT administered to glioblastoma multiforme patients	I	NK cells (cytotoxic lymphocytes)	Recurrent glioblastoma	[45]	Established that the ALECSAT protocol is well tolerated in glioblastoma patients, showing potential for slowing disease progression
NCT03383978	Intracranial injection of NK-92/5.28.z cells in patients with recurrent HER2-positive glioblastoma	I	NK cells	Recurrent HER2 + glioblastoma	[46]	Provided first-in-human proof that intracranial CAR-NK cell injections are safe and do not cause severe neurotoxicity in HER2 + glioblastoma patients
NCT02839954	CAR-pNK cell immunotherapy in MUC1 positive relapsed or refractory solid tumor	I/II	NK cells (CAR-pNK)	Recurrent glioma	[47]	Demonstrated that combined MUC1 and PD-L1-targeted CAR-NK cell therapy is safe and well tolerated in patients with relapsed or refractory solid tumors, showing no cytokine release syndrome
NCT04185038	CAR T cell therapy targeting B7-H3 for malignant CNS tumors	I	T cells (CAR-T)	Malignant CNS tumors, DIPG, DMG	[48]	Preliminary results showed that repetitive brain injections (intracerebroventricular) of B7-H3-targeting CAR-T cells were well tolerated and provided a promising overall survival benefit for pediatric patients with DIPG
NCT05768880	Study Of B7-H3, EGFR806, HER2, And IL13-Zetakine (Quad) car T cell locoregional immunotherapy for malignant CNS tumors	I	T cells (CAR-T)	Malignant CNS tumors, DIPG, DMG	[49]	This ongoing phase 1 trial is evaluating the safety of a “Quad” CAR-T cell therapy targeting four antigens (B7-H3, EGFR806, HER2, and IL13-zetakine) to overcome tumor heterogeneity in pediatric brain tumors
NCT01875601	NK white blood cells and interleukin in children and young adults with advanced solid tumors	I	NK cells	Refractory pediatric malignant solid tumors	[42]	Established the feasibility and safety of harvesting and expanding activated autologous NK cells for dose-escalated infusion following lymphodepleting chemotherapy in children with advanced solid tumors
NCT04214730	Study of NK combined with chemotherapy for advanced solid tumor	-	NK cells	Advanced solid cancer (non-operable)	[50]	Investigated the safety and efficacy of combining NK cell infusions with chemotherapy to enhance antitumor responses in patients with non-operable advanced solid cancers
NCT02208362	Genetically modified T-cells in treating patients with recurrent or refractory malignant glioma	I	Autologous NK/T	Recurrent glioblastoma	[36]	Proved that local delivery of IL-13Rα2-targeted CAR-T cells is safe and feasible, with 50% of recurrent glioblastoma patients achieving stable disease or better
NCT01454596	CAR T cell receptor immunotherapy targeting EGFRvIII	II	CAR-T	GBM	[37]	Found that intravenous infusion of EGFRvIII-directed CAR-T cells followed by IL-2 is safe and leads to successful cell trafficking into the brain, despite a median progression-free survival of 1.3 months
NCT02209376	Autologous T cells redirected to EGFRVIII-with a chimeric antigen receptor	I	CAR-T	EGFRvIII + GBM	[51]	Proved that a single peripheral dose of EGFRvIII-directed CAR-T cells can mediate antigen loss and induce adaptive resistance in recurrent glioblastoma, with one patient maintaining stable disease for over 18 months
NCT00730613	Cellular adoptive immunotherapy using genetically modified T-lymphocytes	I	CAR-T	Recurrent or refractory high-grade malignant glioma	[38]	Validated the safety and bioactivity of intracranial IL13-zetakine + CD8 + T cell clones, observing transient antitumor activity in recurrent glioblastoma patients
NCT04270461	NKG2D-based CAR T-cells immunotherapy	I	CAR-T	GBM, other solid tumors	[52]	Highlighted the potential of NKG2D CAR-T cells to target broadly expressed stress-induced ligands and overcome the immunosuppressive barriers of the solid tumor microenvironment
NCT04717999	Pilot study of NKG2D CAR-T	–	CAR-T	Recurrent glioblastoma	[53]	Assessing the therapeutic potential and technological innovations of CAR-T cells to target specific antigens like EGFRvIII, HER2, and IL-13Rα2 in glioblastoma
NCT05131763	NKG2D-based CAR T-cells immunotherapy	I	CAR-T	GBM, other solid tumors	[54]	Evaluating the safety and clinical activity of NKG2D-based CAR-T cells administered intravenously or via hepatic portal artery for relapsed/refractory solid tumors
NCT06186401	Anti-EGFRvIII synNotch receptor induced Anti-EphA2/IL-13Ralpha2 CAR (E-SYNC) T cells	I	CAR-T	EGFRvIII + GBM	[41]	Evaluating the safety and preliminary efficacy of a “next-generation” synNotch-gated CAR-T therapy (E-SYNC) designed to precisely target EGFRvIII while overcoming glioblastoma heterogeneity
NCT03415100	Pilot study of NKG2D-ligand targeted CAR-NK cells	I	CAR-NK	Colorectal cancer (HCT116), potential GBM application	[55]	Demonstrated that local infusion of mRNA-engineered NKG2D CAR-NK cells is safe and can successfully reduce tumor burden and malignant ascites in patients with metastatic colorectal cancer
NCT03579927	CAR.CD19-CD28-zeta-2A-iCasp9-IL-15-transduced cord blood NK cells, high-dose chemotherapy, and stem cell transplant	II	CAR-T	Diffuse midline glioma (DMG), DIPG	[56]	Investigating the safety and clinical activity of combining CD19-targeted cord blood NK cells with high-dose chemotherapy and stem cell transplants to treat B cell lymphomas
NCT04550663	Study to evaluate the safety and effectiveness of NKG2D-based CAR-T cells infusion	I	CAR-T	GBM	[57]	Assessing the safety and antitumor effectiveness of NKG2D-based CAR-T cell infusions as a treatment for patients with glioblastoma
NCT04077866	Evaluate the safety and efficacy of B7-H3 CAR-T	II	CAR-T	Recurrent or refractory glioblastoma	[58]	Evaluating the safety, feasibility, and maximum tolerated dose of B7-H3-targeted CAR-T cells delivered via intratumoral or intracerebroventricular injection for recurrent glioblastoma
NCT03056339	Umbilical & cord blood (CB) derived car-engineered NK cells	II	CAR-NK	Relapsed/refractory CD19 + B lymphoid malignancies	[59]	Reported a 73% objective response rate in 11 patients with relapsed/refractory B cell malignancies, with cord blood-derived CAR-NK cells persisting for up to one year without major toxicities
NCT05627323	Study of the dual delivery of CHM-1101, an autologous chlorotoxin-chimeric antigen receptor (CLTX-CAR) cell product	I	CAR-T	Recurrent or progressive glioblastoma	[60]	Ongoing phase 1b study evaluating the safety and feasibility of dual-delivered chlorotoxin-directed CAR-T cells (CHM-1101) in patients with MMP2+ recurrent or progressive glioblastoma

PGE2 axis-driven resistance to NK- and CAR-based immunotherapies in GBM and the CNS

Clinical status of adoptive cell therapies in brain cancer

Recent clinical investigations in malignant brain tumors have evolved from conventional adoptive lymphocyte transfer toward increasingly refined genetic engineering platforms designed to enhance tumor specificity and effector potency. Although overall safety profiles have been acceptable across early-phase studies, durable clinical responses remain elusive. In the context of T and chimeric antigen receptor T cells, precision targeting of tumor-associated antigens such as EGFRvIII (NCT01454596) and IL13Rα2 (NCT00730613) initially demonstrated measurable antitumor activity; however, these responses were frequently transient, with rapid disease recurrence driven by antigen loss variants and progressive T cell dysfunction, including exhaustion phenotypes that undermine sustained cytotoxicity [7, 39, 51]. Parallel efforts employing natural killer and CAR-modified NK cells have capitalized on their intrinsic, major histocompatibility complex-independent cytolytic capacity. Clinical approaches using HER2-directed NK-92 cells or MUC1-targeted CAR-NK platforms have provided a comparatively safer, allogeneic “off-the-shelf” strategy with reduced incidence of immune effector cell-associated neurotoxicity. Nonetheless, limited in vivo persistence and suboptimal expansion within the immunosuppressive brain microenvironment continue to constrain their long-term therapeutic efficacy [6].

The mechanism of failure: the PGE2-mediated “immune barrier”

The COX-2/PGE₂ axis operates as a pivotal driver of immunotherapy resistance in malignant brain tumors, functioning as an upstream regulator that coordinates immune suppression at multiple levels. Rather than acting through a single inhibitory pathway, it establishes a layered defensive network that simultaneously undermines both innate and adaptive immune effectors. Within this framework, extracellular PGE₂ imposes a profound functional paralysis on adoptively transferred lymphocytes, effectively driving what clinicians often describe as NK and T cell “fatigue”. By engaging EP2 and EP4 receptors on therapeutic cells, PGE₂ activates intracellular cAMP/PKA signaling cascades that blunt effector competence at multiple levels. In T and CAR-T cells, this pathway suppresses IL-2 production and proliferative expansion, while reprogramming cellular metabolism toward an exhausted phenotype that favors the establishment of immunologically “cold” tumor niches [61]. In parallel, NK and CAR-NK cells experience transcriptional and surface downregulation of critical activating receptors, including NKG2D and DNAM-1 [31]. Because these molecules constitute the principal activating machinery for NK-mediated cytotoxicity, their loss functionally disables tumor recognition, rendering even highly engineered CAR-NK constructs incapable of efficiently engaging malignant targets.

Beyond intrinsic effector dysfunction, structural and microenvironmental constraints further compromise therapeutic efficacy. The BBB restricts the trafficking and infiltration of intravenously administered immune cells into the tumor bed [62]. Even when this obstacle is bypassed through intracranial delivery, the PGE₂-enriched milieu within the brain TME operates as a biochemical barricade, limiting intratumoral penetration and sustained activity. Concurrently, PGE₂ reshapes the myeloid compartment by recruiting myeloid-derived suppressor cells and polarizing macrophages toward an M2-like suppressive state [35, 63]. These reprogrammed cells secrete inhibitory mediators such as TGF-β and IL-10, establishing a self-amplifying immunosuppressive circuit that persists beyond the limited life span and functional window of infused effector cells.

Compounding these challenges, antigenic escape and cellular heterogeneity remain formidable barriers. Malignant brain tumors display extensive intercellular diversity, and PGE₂-driven inflammatory signaling enhances transcriptional plasticity within tumor populations. This adaptive reprogramming enables tumor cells to downregulate or lose target antigens—such as EGFRvIII—thereby evading recognition by CAR-engineered T or NK cells that are designed to detect these specific molecular signatures [64]. Collectively, these layered mechanisms underscore how the PGE₂ axis not only dampens effector cell fitness but also remodels the tumor ecosystem to actively resist precision immunotherapy.

Clinical dilemmas: PGE2-mediated immune exclusion and effector paralyzation

The failure of NK, T, CAR-T, and CAR-NK therapies in brain cancer clinical trials stems from a synergistic “immune exclusion” effect, where the PGE2 axis acts as the primary orchestrator of the immunosuppressive landscape. The core dilemma lies in the functional paralyzation of effector cells: While T and CAR-T cells are hindered by EP2/EP4-mediated signaling that induces rapid metabolic exhaustion [61], NK and CAR-NK cells suffer from a loss of “blind” cytotoxicity due to the PGE2-driven downregulation of NKG2D receptors [31]. This is compounded by the structural limitations of the BBB, which restricts cell trafficking [62], and the extreme antigenic heterogeneity that allows for rapid escape—a phenomenon particularly prevalent in EGFRvIII-targeted CAR-T trials [51, 65]. Furthermore, unlike the safer but shorter-lived NK-based approaches [6], T cell therapies often struggle with the narrow therapeutic window between efficacy and high-grade neurotoxicity [51]. Consequently, current clinical failures suggest that achieving sustained tumor regression requires more than antigen recognition; it necessitates a “combination armoring” strategy, such as pairing synNotch systems [41] with COX-2 inhibitors [66], to physically dismantle the PGE2-rich metabolic barrier.

Receptor-specific mechanisms of PGE2-mediated immune suppression in brain tumors

PGE2 exerts multifaceted effects on brain tumor biology through four G-protein-coupled receptors (EP1, EP2, EP3, and EP4), each with distinct signaling pathways and roles in immune modulation. These receptors differentially regulate the TME and immune effector functions, particularly in NK and T cells. Despite their significance, the precise contributions of each receptor to immune suppression and tumor progression remain underexplored and are often treated as a monolithic entity in therapeutic strategies. This section provides a comprehensive analysis of the EP receptor-specific effects on immune suppression, their implications for brain tumor immunotherapy, and the critical gaps that require further investigation.

EP2 and EP4 receptors: predominant mediators of immune suppression

The EP2 and EP4 receptors, both coupled to Gαs proteins, elevate intracellular cyclic adenosine monophosphate (cAMP) levels via adenylyl cyclase activation, profoundly suppressing immune effector functions in the brain tumor TME [49]. These receptors are the primary drivers of PGE2-mediated immune evasion, affecting NK and T cell activities, which are critical for antitumor responses.

In NK cells, signaling through EP2 and EP4 receptors reduces the expression of crucial activating receptors, such as NKG2D and NKp30, which hinders the release of cytotoxic granules containing perforin and granzymes, as well as the production of cytokines like IFN-γ. Recent investigations have revealed that PGE2, acting via EP2/EP4, modifies the receptor expression patterns in NK cells, diminishing their ability to release granules and migrate toward tumor sites in GBM models [50]. Similarly, in CD8 + T cells, EP2/EP4-mediated cAMP elevation disrupts glycolysis and ribosome biogenesis, which are critical for effector function and proliferation [35]. Single-cell RNA sequencing (scRNA-seq) analyses of GBM have revealed that EP2/EP4 signaling induces metabolic reprogramming, leading to T cell exhaustion and diminished antitumor responses.

Beyond direct immune cell suppression, EP2 and EP4 receptors promote an immunosuppressive TME by recruiting myeloid-derived suppressor cells (MDSCs) and polarizing glioma-associated macrophages (GAMs) toward an M2 phenotype. Preclinical studies on brain tumor types, including medulloblastoma and meningioma, have indicated that PGE2-driven MDSC accumulation inhibits NK cell cytotoxicity and T cell activation, creating a permissive environment for tumor growth [67]. In GBM, EP2 inhibition attenuates COX-2-associated tumor progression, reduces MDSC infiltration, and enhances immune surveillance [68]. Comparable effects have been observed in DMG, where EP4 signaling exacerbates TME immunosuppression, limiting the efficacy of CAR-T cell therapies.

The preclinical success of selective antagonists underscores the therapeutic potential of targeting EP2/EP4. For instance, EP2 inhibitors reduced the tumor burden in GBM mouse models by restoring NK and T cell functions, while EP4 antagonists [69] such as HTL0039732 are under investigation in clinical trials for solid tumors, including brain malignancies (NCT05944237) [70]. These findings highlight that EP2 and EP4 are critical targets for enhancing the efficacy of immunotherapy across brain tumor subtypes.

EP1 and EP3 receptors: emerging roles in tumor biology and immune modulation

In contrast to the well-characterized immunosuppressive roles of EP2 and EP4, the EP1 and EP3 receptors—coupled to Gαq and Gαi proteins, respectively—are less studied but increasingly recognized for their contributions to brain tumor progression and immune regulation [71]. Their distinct signaling pathways suggest nuanced roles that warrant further investigation.

The EP1 receptor, which activates protein kinase C (PKC) and increases intracellular calcium via phospholipase C, has been implicated in proinflammatory signaling and tumor cell behavior [66]. In GBM, EP1 mediates neuronal excitation through calcium influx, upregulates synaptic proteins, and elevates glutamine/asparagine levels, all of which confer chemoresistance [9]. A study highlighted EP1’s role in neuron–tumor crosstalk, where PGE2-driven calcium signaling enhanced tumor cell survival and resistance to therapies, a mechanism potentially relevant to DMG and DIPG, owing to their midline locations. Additionally, EP1 may amplify inflammation via nuclear factor kappa-B (NF-κB) activation, potentially increasing MDSC recruitment and TME immunosuppression in meningioma and medulloblastoma models. However, direct evidence is limited.

The EP3 receptor, which inhibits cAMP via Gαi signaling, is primarily associated with tumor cell migration and invasion. In GBM cell lines (e.g., A172), EP3 activation via PGE2 enhances migration through TRPM7 channel modulation and PKA-dependent pathways, thereby promoting tumor invasiveness [72]. Emerging evidence suggests that EP3 may also indirectly influence immune cell function by altering TME cytokine profiles, potentially skewing T cell responses toward a Th2 phenotype in medulloblastoma and reducing NK cell cytotoxicity in DMG [73]. However, its immune-specific effects remain underexplored compared to EP2/EP4.

The distinct signaling of EP1 and EP3 suggests their potential synergistic roles with EP2/EP4 in shaping the TME. For example, in meningiomas, EP1-driven inflammation may enhance EP2/EP4-mediated MDSC recruitment, thereby creating a compounded immunosuppressive effect [35]. Similarly, in pediatric tumors, such as DIPG, EP3’s role in tumor cell survival may complement EP4’s immune suppression, complicating immunotherapy outcomes [74].

Unresolved mechanisms and critical gaps

Despite these advances, significant gaps persist in our understanding of EP receptor-specific contributions to immune suppression and tumor progression across brain tumor types. The differential expression and activity of EP1–EP4 receptors vary by tumor subtype, with mesenchymal GBMs exhibiting higher EP2/EP4 expression, correlating with aggressive behavior and poor prognosis. In contrast, proneural GBMs and certain medulloblastomas may rely more on EP1/EP3 for tumor survival, necessitating subtype-specific investigations.

The compensatory mechanisms of EP receptors remain poorly defined. For example, inhibition of EP2 may upregulate EP3 signaling, potentially sustaining tumor migration or immune evasion, as suggested in preclinical GBM models [69]. Spatial transcriptomics and scRNA-seq are required to map receptor expression and interactions within the heterogeneous TMEs of GBM, meningioma, and pediatric gliomas. Such approaches could elucidate how EP1-driven neuronal excitation in the DMG/DIPG interfaces with EP4-mediated immune suppression, thereby providing information for combination therapies.

PGE2-mediated suppression of NK and T cell subsets in brain tumors

PGE2 does not exert uniform immunosuppressive effects across the immune landscape; rather, it selectively targets specific subsets of NK and T cells and modulates their functions in a context-dependent manner. This selective vulnerability is particularly evident in the TME. By downregulating key receptors, impairing cytokine production, and promoting exhaustion or regulatory phenotypes, PGE2 contributes to immune evasion and therapy resistance.

NK cells: subset-specific vulnerabilities and functional impairment

NK cells, comprising the CD56dim (primarily cytotoxic) and CD56bright (cytokine-producing) subsets, are critical innate effectors in brain tumor surveillance. PGE2 preferentially disrupts these subsets through EP2/EP4 receptor signaling, leading to the downregulation of activating receptors (e.g., NKG2D and NKp30) and alterations in chemokine receptor expression (e.g., shifts in CXCR3/CXCR4), which impairs migration and tumor infiltration [75, 76]. In GBM, PGE2 induces NK cell anergy, characterized by increased secretion of immunosuppressive cytokines such as IL-6 and IL-8, alongside reduced IFN-γ production, thereby diminishing antitumor cytotoxicity [77].

Subset-specific effects are notable: PGE2 appears to have a profound impact on CD56dim NK cells, compromising their perforin/granzyme-mediated killing, whereas CD56bright subsets may exhibit altered cytokine profiles that favor TME immunosuppression. Preclinical studies in GBM models have demonstrated that PGE2 signaling via EP2/EP4 reprograms NK cell gene expression, leading to dysfunction and reduced degranulation, particularly in disseminated tumor cells. A 2024 study highlighted PGE2-induced NK cell dysfunction in disseminated GBM cells, emphasizing impaired recruitment and activation as the key mechanisms of evasion [78].

This suppression also extends to other brain tumors. In medulloblastoma, the COX-2/PGE2 axis promotes immune evasion by inhibiting NK cell cytotoxicity, potentially through MDSC recruitment and TME remodeling, as observed in preclinical models where PGE2 correlates with reduced NK infiltration [79] (Table 2). Herein, we summarized ongoing clinical trials targeting the COX-2/PGE₂ axis and related prostaglandin receptors (EP2, EP4, and DP) in advanced solid tumors, highlighting their translational potential for brain tumors. Most agents, such as E7046, ONO-4578, AAT-007, TPST-1495, and vorbipiprant, are EP4 or dual EP2/EP4 antagonists used during early-phase (I or Ib/II) investigations. These compounds aim to reverse PGE2-mediated immunosuppression, restore NK and T cell function, and enhance checkpoint inhibitor efficacy. Although not yet brain tumor-specific, these trials provide a pharmacological foundation for integrating EP receptor blockade with NK/T cell therapy, particularly for GBM or diffuse midline glioma (DMG), where COX-2 and PGE₂ are the key immune-modulating factors (Table 2).

Table 2.

Clinical trials targeting PTGS2 (COX-2), EP, and DP pathways in cancer

Agent	Target	Cancer indication	Phase	NCT number(s)	Main findings
E7046	EP4 antagonist	Advanced/metastatic solid tumors	Phase I/Ib	NCT02540291; NCT03152370; NCT04432857[80–82]	Demonstrated acceptable safety and induced objective responses in combination with radiotherapy or pembrolizumab
ONO-4578	EP4 antagonist	Advanced malignancies	Phase I	NCT03661632; NCT03155061[83, 84]	Showed manageable tolerability and promoted CD8+ T cell infiltration when combined with PD-1 inhibitors
AAT-007	EP4 antagonist	Advanced/metastatic tumors	Phase I	NCT02538432; NCT03658772 [85, 86]	Confirmed a favorable safety profile and biological activity in reversing prostaglandin-mediated immunosuppression
TPST-1495	Dual EP2/EP4 antagonist	Advanced/metastatic tumors	Phase I	NCT04344795[87]	Exhibited greater potency in blocking PGE2 signaling compared to single antagonists, with early signs of tumor control
Ifetroban	TXA2 receptor antagonist (relevant to PG pathways)	High-risk malignant solid tumors	Phase I/II	NCT03694249 [88]	Investigated for its potential to inhibit platelet-mediated tumor metastasis with a well-tolerated safety profile
Vorbipiprant	EP4 antagonist	Metastatic colorectal cancer (combined with PD-1 inhibitor)	Phase Ib/II	NCT05205330[89]	Showed promising clinical activity in refractory MSS-CRC patients by enhancing the efficacy of PD-1 blockade

In mouse and human models, PGE2 consistently suppressed NK cell cytotoxicity and T cell proliferation through EP2/EP4-dependent cAMP signaling. Early studies [90, 91] demonstrated that PGE₂ reduces NK activation markers (e.g., NKp30, TRAIL) and increases inhibitory KIR expression. More recent single-cell analyses [92, 93] confirmed that tumor-derived PGE₂ disrupts IL-2/STAT5 signaling and effector differentiation in CD8⁺ T cells, while EP2/EP4 blockade restores cytokine production and proliferation. In contrast, PGD₂ signaling appears immunostimulatory, enhancing cytotoxic and inflammatory gene signatures in NK and CD8⁺ MAIT cells [94], suggesting a context-dependent divergence among prostaglandins (Supplementary Table). However, there are currently no meta-analyses (or even disease‑specific systematic reviews) focused on E7046, ONO‑4578, AAT‑007, TPST‑1495, or vorbipiprant in primary brain tumors or gliomas. Most studies of these agents are small, early phase, and heterogeneous, or conducted in mixed solid tumor populations rather than glioma-specific cohorts, limiting the availability of comparable data for formal evidence synthesis. Moreover, brain tumor-specific challenges—including small patient populations, BBB constraints, evolving endpoints, and rapidly shifting treatment paradigms—have slowed the accumulation of robust clinical evidence, leaving the field dominated by preclinical and exploratory studies.

Together, these tables emphasize that PGE₂-driven EP2/EP4 signaling is a central suppressive axis in NK and T cell dysfunction, whereas pharmacological inhibition or receptor-specific modulation represents a rational strategy to improve immunotherapy efficacy in brain tumors. Integrating EP antagonist therapy with CAR-NK or CAR-T cell approaches could thus restore effector activity and counteract the immunosuppressive TME characteristics of GBM and DMG.

T Cells: differential impacts on effector, memory, and regulatory subsets

T cells, encompassing CD8 + cytotoxic, CD4 + helper, and Treg subsets, are vital for adaptive antitumor immunity but are selectively suppressed by PGE2 in brain tumor TMEs. PGE2 primarily targets CD8 + T cells via EP2/EP4-mediated cAMP elevation, reducing IFN-γ production, promoting exhaustion (e.g., via PD-1 upregulation), and disrupting metabolic pathways like glycolysis. While PGE2 suppresses T-cell proliferation and IL-2 production by preventing cells from reaching a functional activation state, it also does so without necessarily decreasing the expression of the IL-2 receptor itself [95]. In GBM, this leads to diminished effector function and impaired tumor infiltration, with studies showing PGE2’s role in metabolic reprogramming, which fosters T cell senescence. γδ T cells, a subset with innate-like properties, are also inhibited through cAMP/PKA signaling, limiting their cytotoxic potential against glioma cells [15, 92].

PGE2 further skews T cell dynamics by promoting Treg expansion and suppressing memory T cell formation. In GBM, the COX-2/PGE2 axis enhances Treg recruitment and function, downregulates major histocompatibility complex class II (MHC II) on microglia, and impairs CD4 + T cell activation, thereby tipping the balance toward immunosuppression. A study linked COX-2/PGE2 signaling to Treg promotion, correlating poor prognosis and resistance to checkpoint inhibitors [96]. Memory T cells (e.g., central and effector memory subsets) remain underexplored; however, emerging data suggest that PGE2 hinders their persistence and reduces long-term antitumor surveillance [92, 97, 98].

In addition to GBM, the effects of PGE2 are also manifested in other brain tumors. In medulloblastoma, PGE2 suppresses CD8 + T cell responses through TME modulation, promoting Treg dominance and limiting immunotherapy efficacy. Meningiomas exhibit PGE2-driven T cell exhaustion, with elevated COX-2 expression correlating with increased Tregs and reduced effector T cell infiltration. In DMG and DIPG, PGE2 exacerbates T cell dysfunction, impairing CAR-T cell therapies by fostering an immunosuppressive milieu rich in Tregs and exhausted CD8 + subsets, as evidenced in clinical trials in which PGE2 levels predict poor outcomes. Recent studies on T cell metabolism in malignant brain tumors highlight PGE2’s role in altering circulating metabolites that favor Treg expansion over effector responses [12, 15, 99].

Impact of PGE2 on dendritic cell function and DC–NK/T cell crosstalk

The bidirectional crosstalk between dendritic cells (DCs) and NK cells is fundamental for orchestrating early innate activation and for licensing effective downstream T cell immunity. Within the TME, however, elevated PGE₂ disrupts this regulatory circuitry and reprograms DCs toward a tolerogenic state, thereby attenuating both NK and T cell responses [100, 101]. Through EP2 and EP4 signaling, PGE₂ interferes with DC maturation, diverting them from fully immunogenic antigen-presenting cells into tolerogenic DCs characterized by suppressed IL-12 production, reduced expression of co-stimulatory molecules (CD80/CD86), and enhanced expression of inhibitory mediators [100, 102]. The reduction of IL-12 critically impairs NK cell activation and Th1 polarization [100, 103], while upregulation of PD-L1 and indoleamine 2,3-dioxygenase establishes a suppressive interface that dampens lymphocyte priming and expansion [104]. Collectively, these molecular alterations convert DCs from immune initiators into enforcers of tolerance.

Functionally, this reprogramming dismantles DC-dependent NK cell priming and recruitment. Under physiological conditions, mature DCs sustain NK proliferation and cytotoxic competence through IL-12 and IL-15 delivery. In a PGE₂-rich milieu, diminished IL-12 leads to inadequate NK activation and reduced IFN-γ secretion [103], while altered production of chemokines such as CXCL9 and CXCL10 limits NK cell trafficking into the tumor core [103, 105]. As a result, NK cells fail to accumulate and execute effective cytotoxic responses. The tolerogenic shift extends to adaptive immunity: PGE₂-conditioned DCs provide insufficient co-stimulatory signals, promote T cell exhaustion or anergy, and favor the expansion of Foxp3⁺ regulatory T cells, further consolidating local immune suppression [101, 104].

In GBM, where PGE₂ levels are frequently elevated, this axis entrenches a multilayered blockade that compromises NK-, CAR-T-, and checkpoint-based therapies. Pharmacologic inhibition of COX-2 or selective blockade of EP2/EP4 signaling represents a rational strategy to restore DC immunogenicity, recover IL-12 production, and re-establish productive DC–NK/T cell interactions. By converting a tolerogenic, “cold” microenvironment into a more inflamed and responsive state, targeting the COX-2/PGE₂ axis may substantially enhance the efficacy and durability of immunotherapy in GBM [101, 103, 105].

Future directions and therapeutic strategies targeting PGE₂-driven immunosuppression in brain tumors

Despite these insights, significant gaps remain in understanding PGE2’s interactions with resident brain cells (e.g., microglia and astrocytes) to amplify subset-specific suppression across tumor types. For instance, how PGE2 modulates microglia-derived factors to differentially affect CD56dim versus CD56bright NK cells or CD8 + effectors versus Tregs has not been fully elucidated, particularly in non-GBM tumors such as medulloblastoma or meningioma. Multi-omics approaches, including scRNA-seq and spatial transcriptomics, can dissect these dynamics and reveal subset-specific vulnerabilities and PGE2-driven metabolic reprogramming [15, 106, 107].

Therapeutic proposals include combining PGE2 inhibitors (e.g., COX-2 antagonists or EP2/EP4 blockers) with CAR-NK therapies targeting antigens such as EGFRvIII in GBM models or GD2 in DMG/DIPG to restore NK cytotoxicity and T cell effector functions. In medulloblastomas and meningiomas, integrating PGE2 blockade with Treg-depleting agents can enhance antitumor immunity [12, 66, 108].

Future studies should prioritize patient-derived organoids and humanized mouse models to validate these strategies and addressing translational barriers and subtype-specific variations in PGE2’s immunosuppressive effects. Such efforts could significantly improve the efficacy of NK and T cell-based immunotherapies for diverse brain tumors [109, 110]. PGE2 contributes significantly to therapeutic resistance across a spectrum of brain tumors, including GBM, medulloblastoma, meningioma, DMG, and DIPG. PGE2 undermines both conventional and novel therapeutic modalities by modulating TME and immune effector functions, acting as a pivotal mediator of resistance. This section examines PGE2’s role in conferring resistance to traditional therapies such as radiotherapy and chemotherapy, as well as emerging immunotherapies, including checkpoint inhibitors, oncolytic virotherapy, CAR-T cell therapies, and BiTEs. It integrates preclinical and clinical evidence to elucidate resistance mechanisms and proposes strategies to overcome them with a focus on diverse brain tumor types [9, 15, 111].

Resistance to conventional therapies

PGE2 promotes resistance to standard treatments, notably radiotherapy and chemotherapy, via multiple molecular pathways. In GBM, PGE2 enhances radioresistance by upregulating inhibitor of DNA binding 1 (Id1), a transcription factor that sustains tumor cell survival under radiation stress. Studies have demonstrated that PGE2, via EP2/EP4 signaling, activates Id1 to maintain stemness and inhibits apoptosis in irradiated GBM cells, thereby reducing treatment efficacy. Similarly, PGE2 contributes to TMZ resistance, a cornerstone chemotherapy for GBM, by inducing DNA repair mechanisms and promoting tumor cell survival through β-catenin and PI3K/AKT pathways [112, 113].

This resistance also extends to other brain tumors. In medulloblastoma, PGE2-driven COX-2 signaling enhances tumor cell survival after radiation by upregulating anti-apoptotic proteins, such as Bcl-2, correlating with increased recurrence rates. In meningiomas, elevated PGE2 levels in the TME are associated with resistance to adjuvant chemotherapy, such as hydroxyurea, through enhanced proliferation and angiogenesis mediated by EP2/EP4 receptors. For DMG and DIPG, PGE2’s role in chemotherapy resistance is emerging, with preclinical evidence suggesting that it sustains tumor cell stemness and inhibits apoptosis in hypoxic niches, thus complicating treatment outcomes in these pediatric malignancies [64].

Resistance to emerging immunotherapies

PGE2 significantly impairs the efficacy of novel immunotherapies, including checkpoint inhibitors and oncolytic virotherapy, by promoting immunosuppressive TME and weakening immune effector functions. Specifically, in GBM, PGE2 induces T cell exhaustion through the EP2/EP4 receptor-mediated elevation of cAMP, which upregulates exhaustion markers such as PD-1 and LAG-3 on CD8 + T cells. This mechanism contributes to diminished responses to checkpoint blockade therapies (e.g., anti-PD-1/PD-L1), as preclinical models have demonstrated that elevated PGE2 levels correlate with reduced objective response rates and worse outcomes. Notably, LAG-3 + CD4 + T cell infiltration, promoted by PGE2-mediated pathways, is associated with poor prognosis and can be targeted to enhance the efficacy of checkpoint inhibitors [114].

In oncolytic virotherapy, PGE2 impairs therapeutic efficacy in GBM and DMG by suppressing NK cell cytotoxicity and antiviral immune responses essential for viral tumor lysis and immune activation. PGE2 reduces production of proinflammatory cytokines like IFN-γ and TNF-α, which are crucial for the immune-mediated clearance of tumors during virotherapy. Targeting PGE2 restored immune activation and enhanced virotherapeutic responses in preclinical models [115]. Additionally, NK cell therapies combined with antibody strategies have demonstrated the potential to counteract PGE2’s immunosuppressive effects in glioma models [116].

Resistance to CAR-T and BiTE therapies

CAR-T cell therapies and bispecific T cell engager (BiTEs) targeting tumor-associated antigens such as EGFRvIII, IL13Rα2, or GD2 face considerable resistance mediated by PGE2 in brain tumors. In GBM, PGE2 promotes T cell exhaustion and reduces CAR-T cell persistence by downregulating critical effector cytokines like IFN-γ and IL-2, while expanding immunosuppressive Tregs within the TME. This immunosuppressive network involving MDSCs and M2 macrophages impedes effective CAR-T cell function, contributing to the transient responses and relapse observed in clinical trials targeting EGFRvIII (e.g., NCT01454596) [37, 117].

In DMG and DIPG, PGE2 exacerbates CAR-T resistance by impairing T cell infiltration across the BBB and suppressing NK cell support, which is critical for synergistic antitumor effects. Preclinical data suggest that PGE2’s interaction with EP2/EP4 receptors on CAR-T cells reduces their proliferative capacity and effector function, particularly in pediatric gliomas with a high hypoxic burden. Similarly, BiTE therapies, which redirect T cells to tumor cells via dual antigen binding, are hindered by PGE2’s induction of T cell anergy and Treg dominance, as observed in meningioma models, where PGE2 correlates with reduced BiTE efficacy.

In DMG and DIPG, PGE2-driven mechanisms further hinder CAR-T cell efficacy by restricting T cell infiltration across the BBB and suppressing NK cell support, which is vital for synergistic antitumor activity. Preclinical evidence demonstrates that PGE2 engagement with EP2/EP4 receptors on CAR-T cells diminishes their proliferation and effector functions in pediatric gliomas marked by high hypoxia [118, 119]. Similarly, BiTE therapies are compromised by PGE2-induced T cell anergy and Treg enrichment, as seen in meningioma models, where PGE2 is correlated with reduced therapeutic efficacy [120].

The emerging data further highlight PGE2’s role in neuron–tumor crosstalk, where it promotes chemoresistance through EP1-mediated neuronal excitation, increasing synaptic activity and glutamine/asparagine levels in GBM and potentially DMG. This resistance mechanism, which is counteracted by BBB-permeable celecoxib derivatives, underscores PGE2’s broad impact on both conventional and advanced therapies.

Additionally, emerging data have highlighted a novel resistance mechanism via neuron–tumor crosstalk in the TME, wherein PGE2 activates neurons through the EP1 receptor. This neuronal excitation enhances synaptic activity and elevates glutamine and asparagine levels, which are metabolites that foster chemoresistance in GBM and potentially in DMG. A BBB-permeable celecoxib derivative (compound 11) was developed to effectively inhibit PGE2 signaling and reduce neuronal excitation and tumor growth in preclinical models, thus presenting a promising approach to overcome PGE2-mediated resistance [9].

Integrative gaps and future directions

The full spectrum of PGE2-mediated resistance mechanisms across brain tumor types remains incompletely understood. The interplay between PGE2 and other TME components such as microglia, astrocytes, and neuronal elements in driving resistance is underexplored. For example, neuron–microglia contacts govern PGE2 tolerance mechanisms via TLR4-mediated pathways in microglia, highlighting the critical role of neural cell interactions in PGE2 regulation within the brain environment [121]. Evidence also shows that PGE2 activates neurons via the EP1 receptor, promoting neuronal excitation that supports tumor resistance via enhanced synaptic plasticity and metabolic changes [9]. This neuronal involvement remains insufficiently investigated, especially beyond GBM, in non-GBM tumors such as meningioma and medulloblastoma, where PGE2’s subtype-specific impact and interactions with astrocytes have yet to be fully defined [122].

Therapeutic strategies to overcome PGE2-mediated resistance include COX-2 inhibitors, such as celecoxib, and selective EP receptor antagonists to prime the TME for enhanced immunotherapy efficacy. COX-2 inhibition has been shown to potentiate immunotherapy responses in brain tumors by reducing immunosuppressive prostaglandins and improving effector T cell responses [123, 124]. In meningiomas, COX-2 is ubiquitously expressed, and celecoxib inhibits meningioma cell growth, suggesting that COX-2 inhibitors may also reduce tumor-associated immunosuppression and edema, potentially enhancing immunotherapy, such as BiTEs [125]. Preclinical work in GBM models has demonstrated that EP4 antagonists restore NK and T cell activity, improving outcomes in combination with CAR-T therapies targeting antigens such as EGFRvIII [126]. In DMG, CAR-T therapies targeting GD2 combined with EP receptor blockade show promise for overcoming PGE2-driven immune suppression.

Future research should leverage multi-omics tools, including spatial transcriptomics and metabolomics, to comprehensively map PGE2 resistance pathways across tumor subtypes and developmental stages [127]. Model systems such as patient-derived organoids and humanized mouse models are critical for validating combination strategies including the integration of PGE2 inhibitors with oncolytic virotherapy or checkpoint inhibitors [128]. Such approaches have the potential to overcome the immunosuppressive milieu and improve therapeutic outcomes for diverse brain tumors by reshaping their treatment landscape.

Future perspectives

To advance brain tumor immunotherapy, several research and therapeutic directions have been proposed to address the challenges posed by PGE2 and the TME.

Receptor specificity governs the dual effects of PGE2

The molecular mechanisms that determine whether PGE₂ exerts intracellular pro-apoptotic or extracellular pro-tumorigenic effects require further investigation. Multi-omics approaches should map PGE2 synthase (e.g., mPGES-1) regulation and EP receptor expression across tumor subtypes (GBM, medulloblastoma, meningioma, and DMG/DIPG). Clustered regularly interspaced short palindromic repeats (CRISPR)-based gene editing in patient-derived organoids can force intracellular PGE2 accumulation, potentially flipping its role in enhancing tumor cell apoptosis while inhibiting secretion.

Targeting subset-specific immune suppression

PGE2 exerts distinct effects on NK (CD56dim vs. CD56bright), and T cell (CD8 + , γδ, Treg, memory) subsets underscore the need for subset-specific immunomodulatory strategies. Multi-omics profiling of immune cells in the TME of diverse brain tumors can identify biomarkers of PGE2-driven suppression, thereby guiding the development of targeted therapies. For instance, combining EP2/EP4 antagonists with CAR-NK therapies targeting EGFRvIII in GBM or GD2 in DMG could restore cytotoxic function, whereas Treg-depleting agents may enhance T cell responses in meningiomas.

Overcoming therapy resistance

By modulating tumor-intrinsic and immune pathways, PGE2 promotes resistance to radiotherapy, chemotherapy, checkpoint inhibitors, virotherapy, CAR-T cells, and BiTEs, necessitating a coordinated therapeutic framework. Pre-treatment with BBB-permeable COX-2 inhibitors (e.g., celecoxib derivatives) or EP receptor antagonists can prime the TME and enhance the efficacy of immunotherapy. Clinical trials should evaluate these combinations, particularly in pediatric tumors, such as DMG/DIPG, where PGE2-driven neuronal crosstalk complicates treatment. Humanized mouse models and organoids can be used validate these strategies across different tumor types.

Enhancing delivery across the BBB

The BBB is a significant barrier to NK and T cell infiltration. Innovations such as convection-enhanced delivery (CED), focused ultrasound, and chemokine receptor engineering (e.g., CXCR2) should be optimized to improve cell delivery in GBM, medulloblastoma, and meningiomas. Combining these drugs with PGE2 inhibitors could enhances immune cell homing and persistence.

Developing precision therapeutics

The development of selective EP receptor antagonists or allosteric modulators is critical to minimize off-target effects while targeting PGE2-mediated suppression. Phase I/II trials, such as those exploring EP4 antagonists (NCT05944237), should be expanded to include brain tumor cohorts with a focus on subtype-specific responses (e.g., mesenchymal GBM vs. proneural). Biomarker-driven approaches, such as identifying high-PGE2 tumors via liquid biopsies or imaging, can personalize treatment.

Exploring non-GBM tumors

While GBM has dominated the research, medulloblastoma, meningioma, and DMG/DIPG require dedicated studies to understand the role of PGE2 in their unique TMEs. For instance, meningioma’s peritumoral edema and medulloblastoma development may amplify effects of PGE2, necessitating tailored therapeutic strategies. These approaches aim to bridge translational gaps by leveraging advanced models and precision therapies to integrate PGE2 modulation with NK/T cell strategies to improve outcomes across brain tumor types.

Discussion

The therapeutic management of GBM continues to be constrained by three interlocking barriers: the restrictive BBB, pronounced intratumoral heterogeneity, and a deeply immunosuppressive TME. Within this network, the COX-2/PGE2 cascade operates as a pivotal inflammatory hub that bridges chronic neuroinflammation with functional immune paralysis. Although intracellular PGE2 can exert pro-apoptotic effects under hypoxic stress through cAMP-dependent mitochondrial pathways, its rapid extracellular release fundamentally redefines its biological role. Secreted PGE2 engages EP1–EP4 receptors and shifts toward a tumor-promoting mediator that enhances proliferation, immune escape, and therapeutic resistance.

Mechanistically, signaling through EP2 and EP4 on immune effector populations elevates intracellular cAMP, triggering a broad inhibitory program. In natural killer cells, this results in the suppression of activating receptors such as NKG2D and NKp30, thereby attenuating cytotoxic recognition. In CD8⁺ T lymphocytes, PGE2 signaling perturbs metabolic fitness and ribosomal activity, accelerating functional exhaustion and diminishing clonal expansion. These coordinated effects establish a biochemical shield that protects GBM from immune surveillance.

Beyond direct suppression of cytotoxic cells, PGE2 reshapes the immune network by reprogramming dendritic cells into a tolerogenic state through EP2/EP4 engagement. Such dendritic cells exhibit impaired IL-12 production, compromising NK cell priming and Th1 polarization, while concurrently upregulating inhibitory ligands including PD-L1. This altered antigen-presenting landscape fosters regulatory T cell expansion and facilitates recruitment of myeloid-derived suppressor populations, reinforcing a self-perpetuating suppressive circuit. In GBM specifically, PGE2 signaling further promotes alternative macrophage polarization and enhances neuronal excitability through EP1 activation. The latter contributes to chemoresistance by coupling synaptic hyperactivity with metabolic reprogramming, linking neurobiology directly to tumor resilience.

Translational efforts targeting this axis are gaining momentum. Small-molecule antagonists such as E7046, ONO-4578, AAT-007, and the dual EP2/EP4 inhibitor TPST-1495 have demonstrated acceptable safety profiles and preliminary biological activity in early-phase clinical studies across solid tumors. Combination regimens—for example, ONO-4578 with PD-1 blockade—have shown enhanced CD8⁺ T cell infiltration, supporting the rationale for prostaglandin pathway inhibition as an immunologic sensitizer. Additionally, brain-penetrant derivatives of celecoxib are being explored to mitigate PGE2-driven neuronal excitation and tumor progression. Nevertheless, most trials remain phase I or Ib/II and are frequently conducted in heterogeneous solid tumor cohorts rather than glioma-specific populations, underscoring the gap between mechanistic insight and disease-focused validation.

Looking forward, improving GBM outcomes will likely require a deliberate strategy of combinatorial immune reinforcement. Integrating EP receptor antagonism with advanced adoptive platforms—including CAR-engineered NK or T cells and programmable systems such as synNotch circuits—may restore effector persistence by dismantling the prostaglandin-enriched metabolic barrier. Equally critical are innovative delivery approaches, such as focused ultrasound or convection-enhanced infusion, to overcome central nervous system pharmacologic constraints. When coupled with subset-resolved multi-omics profiling to guide patient stratification, these approaches offer a coherent framework for translating pathway biology into clinical benefit. In this context, dual targeting of the PGE2 axis alongside cellular immunotherapy represents a mechanistically grounded strategy with the potential to recalibrate the therapeutic trajectory of GBM.

In conclusion, PGE₂ represents a central immunometabolic regulator in malignant brain tumors, integrating chronic neuroinflammation with profound immune suppression and therapeutic resistance. By attenuating NK and CD8⁺ T cell cytotoxicity, reprogramming dendritic cells toward tolerogenic phenotypes, promoting suppressive myeloid polarization, and coupling neuronal hyperactivity to metabolic adaptation, the COX-2/PGE₂ axis establishes a multilayered tumor-protective network. Its context-dependent duality—exerting intracellular pro-apoptotic effects under defined stress conditions while functioning extracellularly as a potent pro-tumorigenic mediator—constitutes both a biological complexity and a therapeutic opportunity. Targeting this pathway, particularly through inhibition of EP2 and EP4 signaling, in rational combination with advanced NK and T cell-based immunotherapies, offers a mechanistically grounded strategy to restore effector function and overcome resistance across GBM and other aggressive brain tumors, including medulloblastoma, meningioma, diffuse midline glioma, and diffuse intrinsic pontine glioma. Moving forward, translating these insights into meaningful clinical benefit will require integrative multi-omics profiling, tumor- and immune subtype-resolved analyses, and innovative central nervous system delivery platforms. Collectively, synergistic modulation of the PGE₂ axis alongside adoptive cellular immunotherapy has the potential to fundamentally reshape the therapeutic landscape of malignant brain tumors.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

All authors contributed to the study conception and design. Manuscript preparation, literature review and analysis were performed by CJS and JFK. The first draft of the manuscript was written by CJS, JFK, YHC, and TIH. All authors read and approved the final manuscript.

Funding

This research was supported by the National Science and Technology Council of Taiwan (grant no. 114-2320-B-038-012), Taipei Medical University (TMU113-AE1-B16), TMU–Ji Yan Biomedical Co., Ltd. Industry–Academia Collaboration Project (grant no. A-113-086, A-113-100 and A-114-041), and Sunny Brain Tumor and Brain Disease Research and Development Fund (106-5310-001-400).

Data availability

No data were generated during preparation of the manuscript.

Declarations

Conflict of interest

The authors have no relevant financial or non-financial interests to disclose.

Ethical approval

Not applicable.