Weekly Digests
‹ Back to May

PGE(2) sabotages TIL expansion and effector differentiation

May 8, 2024

The bioactive lipid prostaglandin E2 (PGE2) is often detected in the tumor microenvironment (TME), and has been shown to play a role in immune escape and cancer progression. Two recent studies published in Nature assessed the effects of PGE2 on tumor-infiltrating lymphocytes (TIL). Lacher, Dörr, et al. investigated its impact on TCF1+CD8+ T cells in mice, while Morotti, Grimm, et al. assessed how PGE2 affects ex vivo TIL expansion for adoptive cell transfer (ACT) in humans.

PGE2 signaling is mediated by four receptors expressed on immune cells, of which two, EP2 and EP4 (encoded by Ptger2 and Ptger4 in mice), are known to suppress immune function. Therefore, Lacher, Dörr, et al. created a mouse model (Cd4crePtger2-/-Ptger4fl/fl) in which mice were germline EP2-deficient, and EP4 was deleted in CD4+ and CD8+ T cells; control mice were either WT mice or lacking only in EP2. The Cd4crePtger2-/-Ptger4fl/fl mice rejected PGE2-producing BRAFV600E melanoma tumors, while tumors grew in WT mice and in mice lacking only EP2. The tumor rejection in the Cd4crePtger2-/-Ptger4fl/fl mice was associated with an increase in CD8+ but not CD4+ TIL, suggesting the differences in rejection were due to CD8+ T cell-related effects.

scRNAseq and scTCRseq on CD8+ TIL from BRAFV600E tumors revealed 8 TIL clusters, which all expressed Pdcd1 (PD-1), Cd44, and Tox. Clusters 1 and 2 highly expressed stem-like TCF1+ T cell markers (Tcf7, Slamf6, and Il7r). Cluster 1 TCF1+ TILs expressed Sell (CD62L), Ccr7, and Bach2, while cluster 2 TCF1+ TILs expressed effector function and migration markers. The other clusters lacked Tcf7, and expressed genes associated with T cell differentiation and effector function, including Havcr2 (Tim3) and Cxcr6, suggestive of early and/or terminally differentiated TIL phenotypes. Unsupervised slingshot analysis showed a developmental trajectory from TCF1+ cells that progressed to CX3CR1hiTim3+ effector cells, then into CD7hiTim3+ effector cells, after which cells progressed to terminally differentiated subsets. In Cd4crePtger2-/-Ptger4fl/fl mice, there were highly expanded TIL clones, higher proportions of clones shared between TCF1+ cells and Tim3+CXCR6+ effector cells, and a shift towards early effector and terminally differentiated TIL populations. Mice only lacking EP2 signaling had no expansion of any differentiating effector TIL populations. The data suggested that PGE2 prevented differentiation and expansion of effector T cells that arise from TCF1+ TILs, and EP2/EP4 deficiency reversed this blockade.

Treatment with FTY720, which prevents newly-primed T cells from moving out of the lymph node, beginning on day 6 did not impact the intratumoral formation and expansion of Tim3+ TILs. However, these effects were blocked when FTY720 was used from day 1, suggesting that TCF1+ TIL locally generated effector responses when PGE2 signaling was blocked.

EP2/EP4 deficiency in TCF1+ TIL increased the expression of transcription factors related to effector differentiation, stimulatory cytokine signaling, and survival. This included increases in transcription factor activity related to IL-2 signaling, which led the researchers to hypothesize that PGE2 might impact the effects IL-2 has on TILs. In vitro experiments with TCF1+ TILs sorted from PGE2-deficient tumors showed that their expansion and differentiation into effector cells in response to IL-2 and CD3/CD28 stimulation was limited by PGE2. Further assessing these effects in antigen-experienced, repeatedly-activated (RA) TCF1+CD8+ T cells generated in vitro with RNAseq showed PGE2 treatment induced transcriptional changes, with increases in EP2/EP4-mediated cAMP signaling and T cell quiescence. Cells also downregulated T cell differentiation-associated mTORC1 and the IL-2 signaling pathway. Further, IL-2 stimulation of these PGE2 cells did not induce STAT5 phosphorylation and the cells had reduced expression of the IL-2 receptor γ-chain, indicating a dominant inhibitory effect of PGE2.

Morotti, Grimm, et al. were also interested in factors affecting the T cell response to IL-2 and investigated how the TME affects the response to IL-2 in human TILs during the rapid expansion protocol for ACT. scRNAseq and scTCRseq analyses of baseline and expanded TILs from patients with melanoma allowed for assessing the baseline TIL state and the subsequent response to IL-2 rapid expansion. There was an association between the baseline IL-2 signaling gene signature and TIL expansion. Assessing TME factors that may impact IL-2 responses in the in vitro cultures, the researchers found an inverse correlation between IL-2 signaling and PGE2-linked pathways.

To determine which TIL clonotypes could sense PGE2 in tumors, a gene signature for recent PGE2 exposure in CD8+ effector T cells was developed using an in vitro culture system with RA T cells. This signature was detected in tumor-reactive CD8+ TILs in baseline tumors, and its expression correlated with poor ex vivo expansion in response to IL-2, in particular in those CD8+ TIL that had a precursor-exhausted, terminally exhausted, or effector memory-like phenotype. Additionally, high levels of PGE2 in tumor fragments in TIL cultures correlated negatively with in vitro TIL expansion. EP2/EP4 blockade with small molecule antagonists reversed the inhibitory effect of PGE2 on TIL expansion.

Assessing how PGE2 negatively impacts the action of IL-2 on TIL, the researchers, similar to Lacher, Dörr, et al., found reduced surface expression of IL-2Rγc on TIL in response to PGE2, which was reversed by EP2/EP4 blockade. Additionally, exposure to PGE2 reduced the surface colocalization of IL-2Rβ and IL-2Rγc in TILs. Stimulation of PGE2-exposed RA T cells with exogenous IL-2 did not induce downstream signaling of IL-2Rγc (AKT, mTOR, S6), and this was also true when an IL-2 variant (IL-2v) that selectively binds to IL-2Rγc was used.

Bulk RNAseq of unstimulated and RA CD8+ T cells after PGE2 exposure revealed an increase in gene expression related to protein kinase A signaling, cAMP-dependent signaling, and metabolic processes. In RA CD8+ T cells, PGE2 inhibited several IL-2-mediated pathways associated with T cell proliferation and regulation of metabolism, leukocyte proliferation, and the mTOR pathway. Also, PGE2 induced transcriptional changes in RA CD8+ T cells related to mitochondrial and lipid metabolism, as it upregulated genes associated with dysfunctional exhaustion, cell cycle arrest, mTOR inhibition, lipid metabolism, and ferroptosis. There was also increased oxidative stress, mitochondrial dysfunction, and a lack of nucleotide synthesis.

Morotti, Grimm, et al. then assessed whether PGE2 blockade during TIL expansion protocols could restore the TIL response to IL-2. Blocking EP2/EP4 receptors in TIL increased the number of expanded TILs. Similarly, blocking COX1 and COX2 reduced PGE2 in the culture medium, increased TIL expansion, and increased IL-2Rγc expression on TIL. Further, it increased the TILs’ mitochondrial fitness and the expression of TCF7 and MYB, which are related to stem-like T cell longevity. Overall, COX inhibition resulted in more proliferation-competent precursor-like TILs in the culture, as there was an increase in stem-like CD39-TCF1+ cells and a decrease in TOX+TCF1- cells. Further, these effects seemed preferentially induced in tumor-reactive TILs, as they upregulated CD137 and IFNγ/TNF production in autologous in vitro tumor coculture assays.

To test whether this translated into better tumor control by the TIL product, TILs were transferred into NSG mice together with patient-derived tumor cells. Control cells could not control tumor growth, but the COX inhibitor-treated TILs induced tumor rejection. In NSG mice with patient-derived xenograft melanomas, these TIL were more abundant intratumorally and had higher CD103 and lower PD-1 expression.

Finally, Lacher, Dorr, et al. also tested the effects of blocking PGE2-EP2/EP4 signaling on tumor growth. WT or EP2/EP4-deficient OT-I T cells were transferred into mice challenged with MC38-OVA tumors. Both WT and EP2/EP4-deficient OT-I cells induced similar levels of TCF-1+ cells, indicating that lymphoid tissue priming was not impacted, but WT T cells halted expansion after tumor infiltration, while the EP2/EP4-deficient OT-I T cells persistently expanded in the TME. These effects were inhibited when IL-2R signaling was blocked, suggesting IL-2R signaling drove the expansion and differentiation of antigen-specific CD8+ TILs when PGE2-EP2/EP4 signaling was absent. When transferred into WT mice, these EP2/EP4-deficient OT-I cells induced complete rejection of MC38-OVA tumors.

Together, these studies show that tumoral PGE2 negatively impacted TIL expansion and effector differentiation. Blockade of the PGE2-PE2/PE4 pathway in the tumor may help induce better T cell antitumor responses. Given its effects on IL-2 response by the TIL, PGE2 might be an important therapeutic target for both IL-2-directed therapies (such as IL-2v) and TIL manufacturing.

Write-up by Maartje Wouters, image by Lauren Hitchings

Meet the researcher

This week, first author Sebastian Lacher answered our questions.

The Böttcher lab: From the left, the 4th person is Jan Böttcher and the 5th is Sebastian Lacher. Photo credit: Astrid Eckert/TUM

What was the most surprising finding of this study for you?
One of the most astonishing findings we made was how PGE2 affected CD8+ T cells in tumors. Contrary to our expectations, PGE2 did not seem to induce a dysfunctional gene expression program or promote T cell exhaustion, nor did it show a significant impact on the ability of cytotoxic CD8+ T cells to execute effector functions. Instead, PGE2 mediated its deleterious effect by throttling T cell expansion and effector differentiation from tumor-infiltrating stem-like CD8+ T cells – a mechanism that seems to be rather unique compared to other previously known inhibitory mechanisms limiting T cell immunity in cancer. A second astonishing finding was how effective tumor-infiltrating CD8+ T cells could mount anti-cancer responses and eliminate tumors when the T cells lacked the PGE2 receptors EP2 and EP4.

What is the outlook?
We hope that our findings may help to improve immunotherapies that are based on targeting T cells or using patient-derived T cells. A companion paper by Morotti & Grimm et al. (2024) already provides evidence that this is indeed the case by showing that PGE2 undermines the activity of human TILs and their use for adoptive T cell therapy (ACT). One of our key findings was that PGE2 impaired IL-2 responsiveness in TILs; therefore, it will be intriguing to see whether blocking the PGE2-EP2/EP4 axis can promote better responses to current immunotherapies using IL-2. Moreover, it will be exciting to see whether blocking PGE2 signaling will play a substantial role in therapies such as chimeric antigen receptor (CAR) T cell or antibody–drug conjugate therapy.

What was the coolest thing you’ve learned (about) recently outside of work?
During my vacation in Panama, my American friend and I went on a snorkeling trip to see the stunning marine life. While exploring Punta Laurel, I swam to a recommended spot with my underwater camera in tow. On the way, I came by some seaweed fields and discovered a sandbank in the middle, where I saw a 4-meter-long shark. I immediately panicked, turned around, and tried to capture the last moments of my life with my camera, wildly taking pictures around me, not aiming for anything. I just shouted, "Shark! Shark!" to my friend, and we both quickly got out of the water. Later, it turned out to be a nurse shark, described as sluggish, sociable, bottom-dwelling sharks that feed on invertebrates, thus posing no threat to humans. Well, now that I know what a nurse shark looks like, next time, I won't panic; instead, I'll aim for the shark and calmly capture the moment.


Lacher SB, Dörr J, de Almeida GP, Hönninger J, Bayerl F, Hirschberger A, Pedde AM, Meiser P, Ramsauer L, Rudolph TJ, Spranger N, Morotti M, Grimm AJ, Jarosch S, Oner A, Gregor L, Lesch S, Michaelides S, Fertig L, Briukhovetska D, Majed L, Stock S, Busch DH, Buchholz VR, Knolle PA, Zehn D, Dangaj Laniti D, Kobold S, Böttcher JP. PGE2 limits effector expansion of tumour-infiltrating stem-like CD8+ T cells. Nature. 2024 Apr 24. 

Morotti M, Grimm AJ, Hope HC, Arnaud M, Desbuisson M, Rayroux N, Barras D, Masid M, Murgues B, Chap BS, Ongaro M, Rota IA, Ronet C, Minasyan A, Chiffelle J, Lacher SB, Bobisse S, Murgues C, Ghisoni E, Ouchen K, Bou Mjahed R, Benedetti F, Abdellaoui N, Turrini R, Gannon PO, Zaman K, Mathevet P, Lelievre L, Crespo I, Conrad M, Verdeil G, Kandalaft LE, Dagher J, Corria-Osorio J, Doucey MA, Ho PC, Harari A, Vannini N, Böttcher JP, Dangaj Laniti D, Coukos G. PGE2 inhibits TIL expansion by disrupting IL-2 signalling and mitochondrial function. Nature. 2024 Apr 24. 

In the Spotlight...

A dietary commensal microbe enhances antitumor immunity by activating tumor macrophages to sequester iron

Sharma et al. identified the commensal LPIMB19 (LP) as a stimulator of CD8+ T cell responses, and verified its efficacy in multiple solid tumor models, including a human BLCA organoid model. The primary immunogenic effector was the capsular polysaccharide RHP, which endowed TILs with an effector-like cytotoxic phenotype and induced macrophage activation dependent on TLR2. Iron transport was upregulated in i.t. macrophages, promoting iron uptake (via LCN2 secretion), with concomitant depletion of iron in tumor cells and enhanced apoptosis. Both TLR2 and LCN2 were required for antitumor efficacy, suggesting a role for macrophage-mediated “nutritional immunity”.

Contributed by Morgan Janes

Facile generation of biepitopic antibodies with intrinsic agonism for activating receptors in the tumor necrosis factor superfamily

Jhajj, Schardt, and Khalasawi et al. report a systematic method for discovering single-chain antibodies with unique epitopes (relative to clinical-stage antibodies) targeting TNF receptors, which were then fused to the light chains of the clinical-stage IgGs to generate human biepitopic agonist antibodies, eliminating the need for animal immunization, humanization, and molecular reformatting. Biepitopic OX40 antibodies showed potent, FcγR-independent primary human CD4+ T cell activation. The competition-based antibody screening platform was generalizable to other members of the TNF receptor superfamily (CD137; 41-BB).

Contributed by Shishir Pant

Selective refueling of CAR T cells using ADA1 and CD26 boosts antitumor immunity

To support CAR T cell metabolism and functionality in the tumor microenvironment, Hu and Sarkar et al. developed “metabolically refueled” CAR T cells (MR-CAR-T) expressing CD26, a T cell costimulatory molecule, and ADA1, an enzyme that generates inosine as an alternative T cell energy source. ADA1 was fused to an anti-CD3 scFv to both activate and direct ADA1 to T cells. Compared to standard CAR T cells, MR-CAR-T improved in vitro migration, cytotoxicity, and proliferation; in vivo, MR-CAR-T increased tumor inosine concentration and T cell infiltration, leading to superior tumor control in HER2- or GPC3-expressing tumor models.

Contributed by Alex Najibi

Single-cell multiomic dissection of response and resistance to chimeric antigen receptor T cells against BCMA in relapsed multiple myeloma

To find markers of response and non-response to BCMA-targeted CAR T cell therapy, Rade and Grieb et al. applied longitudinal, single-cell multiomics on peripheral blood and bone marrow before and after CAR T cell infusion in heavily pretreated patients with relapsed/refractory MM. Differences between responders (5 CR) and non-responders (5 with poor outcomes) were identified at apheresis, prior to CAR T cell infusion. Non-responders had a suppressive microenvironment with increased CD39+ monocytes and fewer and functionally impaired CD8+ T cells and NK cells. After therapy, hyperexpanded CAR T cell clones had an exhausted PD-1+ phenotype compared to low/intermediate clones.

Contributed by Katherine Turner

RNA aggregates harness the danger response for potent cancer immunotherapy

Mendez-Gomez et al. showed that multi-lamellar tumor- and model-antigen RNA lipid particle aggregates (RNA-LPAs), delivered i.v., transfected fibroblastic stroma cells to activate cytosolic pattern recognition receptors to induce cytokines/chemokines, DC/lymphocyte migration, peripheral/intratumoral specific primary and recall responses, and tumor rejection in murine models. Tumor-derived RNA-LPA infusions in client-owned canines with terminal spontaneous gliomas were well tolerated, rapidly induced an immunologically “hot” TMEs, and improved survival outcomes. In a first-in-human glioblastoma trial, RNA-LPA infusions rapidly induced immune activation in the periphery and TME, and glioma antigen-specific responses.

Contributed by Paula Hochman

Up-regulated PLA2G10 in cancer impairs T cell infiltration to dampen immunity

Using a genome-wide in vitro screening platform, Zhang et al. identified phospholipase A2 group 10 (PLA2G10) protein as a candidate for T cell exclusion. PLA2G10 is frequently upregulated in human cancers and is associated with T cell exclusion and poor tumor immunity. PLA2G10 overexpression in immunogenic mouse tumors excluded T cells and dampened tumor immunity, resulting in resistance to anti-PD-1 immunotherapy. Mechanistically, PLA2G10 hydrolyzed phospholipids into small lipid metabolites, which inhibited chemokine-mediated T cell mobility. PLA2G10 blockade enhanced T cell infiltration and sensitized PLA2G10-overexpressing tumors to immunotherapy.

Contributed by Shishir Pant

Everything New this Week In...

Close Modal

Small change for you. Big change for us!

This Thanksgiving season, show your support for cancer research by donating your change.

In less than a minute, link your credit card with our partner RoundUp App.

Every purchase you make with that card will be rounded up and the change will be donated to ACIR.

All transactions are securely made through Stripe.