Targeting the EP1 receptor reduces Fas ligand expression and increases the antitumor immune response in an in vivo model of colon cancer

Authors


Correspondence to: Aileen Houston, Department of Medicine, University College Cork, Clinical Sciences Building, Cork University Hospital, Cork, Ireland, Fax: +353-21-4345300, E-mail: a.houston@ucc.ie

Abstract

Despite studies demonstrating that inhibition of cyclooxygenase-2 (COX-2)-derived prostaglandin E2 (PGE2) has significant chemotherapeutic benefits in vitro and in vivo, inhibition of COX enzymes is associated with serious gastrointestinal and cardiovascular side effects, limiting the clinical utility of these drugs. PGE2 signals through four different receptors (EP1–EP4) and targeting individual receptor(s) may avoid these side effects, while retaining significant anticancer benefits. Here, we show that targeted inhibition of the EP1 receptor in the tumor cells and the tumor microenvironment resulted in the significant inhibition of tumor growth in vivo. Both dietary administration and direct injection of the EP1 receptor-specific antagonist, ONO-8713, effectively reduced the growth of established CT26 tumors in BALB/c mice, with suppression of the EP1 receptor in the tumor cells alone less effective in reducing tumor growth. This antitumor effect was associated with reduced Fas ligand expression and attenuated tumor-induced immune suppression. In particular, tumor infiltration by CD4+CD25+Foxp3+ regulatory T cells was decreased, whereas the cytotoxic activity of isolated splenocytes against CT26 cells was increased. F4/80+ macrophage infiltration was also decreased; however, there was no change in macrophage phenotype. These findings suggest that the EP1 receptor represents a potential target for the treatment of colon cancer.

Abbreviations
COX

cyclooxygenase

PGE2

prostaglandin E2

FasL

Fas ligand

CTL

cytotoxic T cell

Treg

regulatory T cell

TAM

tumor-associated macrophage

Numerous studies have demonstrated a link between chronic inflammation and cancer. One such inflammatory mediator is prostaglandin E2 (PGE2).[1] PGE2 is derived from arachidonic acid as a result of the activity of cyclooxygenases (COXs). Numerous mouse models of cancer have demonstrated that COX-2-derived PGE2 promotes tumor growth,[2, 3] with increased expression of COX-2 and PGE2 being found in various human malignancies. Moreover, inhibition of COX-2-derived PGE2 has significant chemotherapeutic benefits in vitro and in vivo.[1, 4-7] However, despite these anticancer benefits, inhibition of COX enzymes has been found to be associated with serious gastrointestinal and cardiovascular side effects,[4, 7] limiting the clinical utility of these drugs.

PGE2 activates four different G-protein-coupled receptors, EP1, EP2, EP3 and EP4, with targeting of the receptors offering the potential of antineoplastic activity with fewer side effects. Although most studies to date have identified the EP2 and EP4 receptors as being responsible for the tumor-promoting effects of PGE2,[8] the EP1 receptor may also be an effective target against colon cancer. Human colon cancer cells express the EP1 receptor in vivo,[9, 10] whereas EP1 receptor knockout mice have significantly fewer azoxymethane (AOM)-induced aberrant crypt foci[11] (ACF) and colon cancer development.[12] Furthermore, ONO-8711, a selective EP1 antagonist, significantly reduced AOM-induced ACF and intestinal polyp formation in APCMin mice.[11, 13] Moreover, Kitamura et al.[14] showed that the EP1 and EP4 receptor subtypes may have separate intrinsic roles and, to some extent, contribute to different aspects of colon tumorigenesis.[14] The EP4 antagonist was found to be more effective at reducing polyp size, whereas the EP1 antagonist was more effective at reducing polyp number.[14] Targeting the EP1 receptor was also shown not to affect prostacyclin production in human endothelial cells,[15] important given that inhibition of prostacyclin production by COX-2 selective inhibitors was shown to be one of the major contributors to the cardiovascular side effects of these drugs.[16]

Despite these promising findings, the mechanisms by which the EP1 receptor promotes tumorigenesis are unclear. Signaling through the EP1 receptor on colon tumor cells was recently shown by us to upregulate the expression of Fas ligand (FasL/CD95L) in vitro and may represent one potential mechanism.[9] We and others have shown that expression of FasL or its receptor Fas on tumor cells promotes tumor growth in vivo.[17, 18] However, whether induction of tumor-expressed FasL in response to signaling through the EP1 receptor occurs in vivo is unclear. Moreover, the EP1 receptor is expressed by both tumor cells and multiple immune cell types. Thus, the protumorigenic effects of the EP1 receptor could be due to PGE2 signaling through EP1 not on the tumor cells directly but rather on immune cells in the tumor microenvironment. For instance, PGE2 suppresses the effector functions of helper T cells, cytotoxic T cells (CTLs) and natural killer cells and enhances the accumulation of regulatory T (Treg) cells.[19] PGE2 also plays a role in the differentiation of monocytes toward an immunosuppressive or “M2-like” phenotype.[20] Such tumor-associated macrophages (TAMs) can play an important role in tumor progression. Whether signaling through the EP1 receptor suppresses the antitumor immune response in vivo is unknown. This has been explored in our study.

The findings of our study suggest that the EP1 receptor is a potential therapeutic target for the treatment of colon cancer. Blocking EP1 receptor signaling in established tumors was found to inhibit tumor growth in vivo. Suppression of tumor growth required inhibition of EP1 receptor signaling in both the tumor cells and nontumor cells in the tumor microenvironment and was associated with a reduction in FasL expression, reduced Treg cell infiltration and an improved antitumor immune response.

Material and Methods

Chemicals

Unless stated otherwise, all chemicals were purchased from Sigma-Aldrich (St. Louis, MO). The EP1 receptor antagonist ONO-8713 was a generous gift from Ono Pharmaceuticals, Osaka, Japan.

Mice and tumor model

Female BALB/c mice (4–6 weeks) were obtained from Harlan UK (Oxon, UK) and maintained in a pathogen-free environment in the animal facility of University College Cork. Animal experiments were performed in accordance with institutional guidelines using an Animal Research Committee-approved protocol. CT26 cells, a murine colon cancer cell line of BALB/c origin, was kindly provided by Dr. Stephen Todryk (Northumbria University, UK). Cells were maintained in vitro at 37°C in a 5% CO2 humidified atmosphere in DMEM supplemented with 100 µg/ml streptomycin, 100 U/ml penicillin and 10% fetal bovine serum. To establish subcutaneous tumors, mice were injected into the right flank with 2.0 × 105 tumor cells resuspended in 100 µl phosphate buffered saline (PBS). Tumor growth was monitored by regular measurement of tumor length (a) and width (b) using a Vernier calliper, and the volume was calculated as follows: ½ (a × b2). Animals were sacrificed after 48 days.

Generation of FasLlow/negative and EP1low/negative colon cancer cells

Cells were transfected with lentiviral particles containing three target-specific shRNAs against FasL (sc-35358-V), EP1 (sc-40170-V) or control lentiviral particles containing scrambled shRNA (sc-108080; Santa Cruz Biotechnology, Santa Cruz, CA), according to the manufacturers' instructions. Briefly, cells were seeded in 24-well plates at a concentration of 2 × 105 cells per milliliter. Cells were infected 24 hr later with lentiviral particles in the presence of 5 µg/ml polybrene. Cells were cultured in selection medium containing puromycin until resistant clones could be identified. Resistant clones were selected by limiting dilution. Knockdown of FasL and EP1 expression was determined by real-time reverse transcription polymerase chain reaction (RT-PCR) and Western blotting. The clones with the lowest level of FasL and EP1 expression were designated as CT26FasL shRNA and CT26EP1 shRNA, respectively, whereas the clone transfected with scrambled RNA was designated as CT26scr shRNA.

Proliferation assay

Cell proliferation was measured by resazurin reduction.[21] Cells were seeded at 2 × 105 cells per milliliter in 96-well plates. After incubation for 24 hr, media supplemented with 44 µM resazurin was added, and resazurin reduction to resorufin was measured fluorometrically using a GENios plate reader (TECAN, Grodig, Austria) and Xfluor spreadsheet software. Results obtained were expressed in fluorescence units (FUs), and the percentage viability was calculated as follows: (FU treated/FU control) × 100.

Tumor cell and TAM isolation

Tumors were sliced into 1–3 mm[3] pieces and incubated for 1 hr at 37°C with collagenase/dispase solution (Roche Diagnostics, Mannheim, Germany). After washing in PBS, a single-cell suspension was obtained by passing the cells through a cell strainer (Benton Dickonson, Franklin Lakes, NJ). TAMs were isolated by seeding the cell suspension on 24-well plates at a concentration of 0.5 × 105 cells per milliliter. Cells were washed 3 hr later with PBS to remove nonadherent cells. The adherent population was characterized by immunofluorescence and morphological criteria.

Splenocyte isolation

Single-cell suspensions were obtained by mechanical disruption with a syringe plunger in RPMI 1640 medium supplemented with 10% fetal calf serum. The suspensions were then passed through a cell strainer, and red cells were lysed using red cell lysis buffer. Cells were then washed and resuspended in complete media.

Autologous mixed lymphocyte-tumor reaction

Splenocytes were stimulated with IL-2 and irradiated CT26 tumor cells (6.2 Gy at 3 Gy/min) at a ratio of 12:1. After 5 days in culture, the in vitro stimulated splenocytes were collected and tested for their cytotoxic activity against CT26 cells using the Ziva Tox Ultrasensitive Cytotoxicity Assay (Jaden BioScience, San Diego, CA) according to the manufacturer's instructions. Briefly, BrdU was added for the last 4 hr of incubation, cells were fixed, washed and incubated with stringency solution. Cells were then incubated with anti-BrdU antibody conjugate solution, washed and incubated with preparation solution prior to addition of the CDP*Star®Chemiluminescent substrate. Chemiluminescence was detected using a Glomax multidetection system luminometer (Promega, Madison, WI).

Flow cytometry

Single-cell suspensions from tumor tissue were prepared. Monoclonal antibodies to CD8, F4/80 (BD Biosciences, NJ), CD4 and CD25 (eBioscience, San Diego, CA) were used to label the cells for phenotypic analysis. Antibodies to the transcription factor Foxp3 (eBioscience) were used to label permeabilized cells. Debris and dead cells were excluded from flow cytometric analysis using a selection gate on forward scatter and side scatter cellular properties. The frequency of Treg cells was assessed by gating CD4+ cells only and subsequently plotting CD25+ cells against Foxp3+ cells. Cell populations were assessed using the Accuri C6 Flow Cytometer System and CFlow commercial software.

Western blotting

Tumor cells were lysed for 1 hr on ice in ice-cold lysis buffer containing 50 mM Tris HCl (pH 8.0), 150 mM NaCl and 1% Triton-X 100, supplemented with complete protease inhibitors (Roche Diagnostics). Equal amounts of protein were separated on a 10% SDS-polyacrylamide gel and transferred to nitrocellulose membranes. Membranes were blocked for 1 hr at room temperature with 5% nonfat dry milk in PBS containing 0.1% Tween-20. Membranes were probed overnight at 4°C with anti-FasL-specific antibody (Abcam, Cambridge, UK) or anti-EP1-, anti-EP2-, anti-EP3- or anti-EP4-specific antibodies (Cayman Chemical, Ann Arbor, MI). Membranes were washed and incubated with the appropriate secondary antibody conjugated with HRP (Dako, Carpinteria, CA). The results were visualized by chemiluminescence detection (Millipore, Billerica, MA). As an internal control, all membranes were subsequently stripped of the first antibody and reprobed with anti-β-actin-specific antibody (Sigma-Aldrich).

RT-PCR

Total cellular RNA was isolated using the GenElute Mammalian Total RNA Mini Kit according to the manufacturer's instructions. cDNA was synthesized using the SuperScript Vilvo Kit (Invitrogen, Carlsbad, CA). RT-PCR was performed using an Applied Biosystems PRISM 7500 PCR system (Applied Biosystems, Foster City, CA) and Syber Green Jumpstart Taq ReadyMix. All samples were run in triplicate, and relative quantitation was calculated using the 2−ΔΔCt method.

PGE2 ELISA

Tumor cells were isolated and seeded at 0.5 × 105 cells per milliliter. After 24 hr, supernatants were harvested, and PGE2 levels were determined in triplicate by ELISA (Arbor Assays, Ann Arbor, MI).

Statistical analysis

Means with SEM are represented in each graph. Statistical analysis was performed using GRAPHPAD PRISM version 5.0 for Windows (GraphPad Software, San Diego, CA). p-values were calculated using the unpaired Student's t-test with p < 0.05 considered statistically significant.

Results

Blocking EP1 receptor suppresses tumor growth in vivo

To date, in vivo studies on the role of the EP1 receptor in colon tumorigenesis have investigated whether suppressing EP1 receptor signaling prior to tumor development, either using an EP1 receptor antagonist or EP1 receptor knockout mice, is effective in preventing tumorigenesis.[11, 12] However, whether targeting the EP1 receptor has therapeutic potential for the treatment of established tumors is unknown. To investigate this, BALB/c mice were injected subcutaneously with CT26 cells. Beginning either 20 days (n = 8) after tumor cell inoculation, at which time all mice had palpable tumors, or after 27 days (n = 8), the specific EP1 receptor antagonist ONO-8713 was administered orally as a powder mixed into their daily feed at 1,000 ppm; 1,000 ppm was selected based on the findings of a preliminary study using 500 and 1,000 ppm (Supporting Information Fig. S1). ONO-8713 is a potent second-generation EP1 receptor antagonist.[22] Consistent with having a Ki binding value for the EP1 receptor of 0.3 nM and a Ki value of greater than 1,000 nM for the other EP and IP receptor subtypes, it has been shown not to have agonistic or antagonistic action on the other prostanoid receptors.[22] Oral administration was used as this was successful in reducing ACF formation and intestinal polyp formation in APCMin mice.[11] Moreover, therapeutically, this is an ideal means of drug administration. Food consumption was monitored every second day, with no difference between the groups seen. As there was the possibility that the antagonist would not reach the site of tumor inoculation, ONO-8713 was administered by direct injection into the tumors of one group of mice (n = 5) three times a week (30 mg/kg per injection) beginning at Day 27. Although twice weekly injection of established tumors was used in the preliminary study (Supporting Information Fig. S1), this was increased to thrice weekly to determine if this would result in a greater reduction in tumor growth.

As shown in Figure 1a, blocking the EP1 receptor suppressed tumor growth in vivo. Beginning ONO-8713 administration at Day 20 (when the average tumor size was 0.26 cm3) significantly suppressed tumor growth by 65% (p = 0.023). Delaying the start of administration until Day 27 when the tumors were larger (average size = 0.5 cm3) also reduced tumor growth (42% reduction); however, this was not significant. Direct injection of established tumors (average size = 0.5 cm3) with ONO-8713 suppressed tumor growth by 78% (p = 0.023), with three of the five established tumors actually regressing in size. However, there was extensive necrosis present in these injected tumors.

Figure 1.

Inhibiting EP1 receptor signaling reduces the growth of pre-established tumors in vivo. (a) CT26 colon tumor cells (2 × 105) were subcutaneously inoculated into BALB/c mice. Inoculated tumor cells and treatments are indicated. Unless otherwise indicated, ONO-8713 was administered in the basal diet. Tumor growth was monitored as described in the Material and Methods section. Data points represent the mean value ± SEM and show the findings of one tumor challenge experiment. (b) EP1 receptor expression was suppressed in CT26EP1 shRNA cells. The efficiency of suppression of EP1 receptor was assessed by RT-PCR and Western blotting. For the RT-PCR, data were normalized to β-actin and analyzed using the 2−ΔΔCt method. Immunoblotting for β-actin was used as the loading control. (c) Proliferation of tumor cells ex vivo was unaffected by EP1 receptor expression or EP1 receptor signaling. Cell proliferation was determined by resazurin reduction. Proliferation was normalized to tumor cells isolated from CT26scr shRNA tumors. Data shown are the findings from three independent experiments. (d) Blocking EP1 receptor signaling does not affect PGE2 secretion by tumor cells. Tumor cells were isolated from all tumor-bearing mice and cultured ex vivo for 24 hr. Cell culture supernatants were harvested, and the level of PGE2 was measured in triplicate by ELISA. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

This reduction in tumor growth in vivo could be due to suppression of the EP1 receptor either on the tumor cells or nontumor cells in the tumor microenvironment. To evaluate this and to control for the effect of EP1 receptor expression by tumor cells, the EP1 receptor was suppressed in CT26 tumor cells prior to subcutaneous inoculation. Numerous clones were generated, and the clone with the greatest reduction was selected for in vivo analysis (Fig. 1b). Suppressing tumor expression of the EP1 receptor had no effect on tumor development but did result in a decrease in tumor growth (20%) in vivo (Fig. 1a), suggesting that inhibition of tumor cell growth in vivo is predominantly due to effects of the EP1 receptor antagonist on the host, rather than on the tumor cells directly. Indeed, the ex vivo growth of tumor cells isolated from the tumor-bearing mice was unaffected by EP1 receptor antagonism (Fig. 1c). However, CT26EP1 shRNA cells still expressed the EP1 receptor (Fig. 1b), albeit at a much lower level. This lower level of expression may still be capable of transducing a protumorigenic signal.

CT26 cells secrete PGE2 and express all four EP receptors (Supporting Information Fig. S2). Changes in the level of PGE2 in the tumors could thus affect signaling through the other receptors. To determine whether the reduced tumor growth correlated with changes in the level of PGE2, endogenous PGE2 levels in the tumors were determined by ELISA (Fig. 1d). Although PGE2 levels were slightly lower in treated tumors, there was no significant difference between the groups, suggesting that the reduced tumor growth was not due to alterations in PGE2-mediated signaling through the other EP receptors.

Inhibition of EP1 receptor suppresses expression of FasL by colon tumors

We have previously shown that PGE2 signaling through the EP1 receptor increases FasL expression in human colon tumor cells[9] and that suppressing FasL expression by tumor cells reduces tumorigenesis in vivo.[17] Consistent with these findings, both knockdown of the EP1 receptor by shRNA and treatment with ONO-8713 suppressed FasL expression in CT26 cells in vitro (Fig. 2a) and in tumors in vivo (Fig. 2e). Moreover, both tumor development (Fig. 2c) and growth (Fig. 2d) were significantly suppressed when the inoculated tumors had FasL expression suppressed in advance by shRNA (Fig. 2b). One-third of the mice inoculated with CT26FasL shRNA did not develop tumors (Fig. 2c), and in the mice that did develop tumors (7/11), the growth of the tumors was significantly reduced (p < 0.007; Fig. 2d). Interestingly, although the initial appearance of the tumors was delayed, by Day 48, however, CT26FasL shRNA tumors that did develop were similar in size to those that were treated with ONO-8713 beginning at Day 20, with no statistical difference between them.

Figure 2.

Signaling through the EP1 receptor upregulates FasL expression in vivo. (a) CT26 cells were treated with increasing concentrations of ONO-8713 for 24 hr, or FasL and EP1 receptor expression was suppressed in CT26FasL shRNA and CT26EP1 shRNA cells, respectively. FasL expression was determined by Western blotting. Expression of the housekeeping gene β-actin was used as an internal control. Results shown are representative of three independent experiments. (b) FasL expression was suppressed in CT26FasL shRNA cells. The efficiency of suppression of FasL was assessed by RT-PCR and Western blotting. RT-PCR data were normalized to β-actin and analyzed using the 2−ΔΔCt method. (c) Suppression of FasL expression reduces tumor development in vivo. Cells were subcutaneously inoculated into BALB/c mice. Mice were palpated thrice weekly to detect tumors. Data shown are the percentage of mice remaining tumor free on the indicated day. (d) Suppression of FasL expression reduces tumor growth in vivo. CT26scr shRNA or CT26FasL shRNA cells (2 × 105) were subcutaneously inoculated into BALB/c mice. ONO-8713 was administered in the basal diet with the day of initiation of treatment indicated. Tumor growth was monitored as described in the Material and Methods section. Data points represent the mean value ± SEM and show the findings of one tumor challenge experiment. (e) Blocking EP1 signaling suppresses FasL expression in vivo. Tumors were excised after 48 days. FasL expression in the excised tumors was analyzed by RT-PCR and Western blotting. For Western blotting, 30 µg of protein from all tumors per treatment group was pooled. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Incorporation of ONO-8713 into the diet of CT26FasL shRNA tumor-bearing mice beginning at Day 27 did not result in a further significant reduction in tumor volume (Fig. 2d). Given that the growth of CT26FasL shRNA tumors was already greatly reduced, it was perhaps unsurprising that ONO-8713 administration did not have any additive effects. However, from Day 36 onward, tumor growth was halted in these mice, in contrast to the slow increase in tumor growth seen in nontreated CT26FasL shRNA tumors (Fig. 2d). Moreover, although FasL was not reduced to the same extent in the CT26EP1 shRNA clones in vitro compared to the ONO-8713-treated cells and the CT26FasL shRNA clones, FasL mRNA and protein levels were reduced in the CT26EP1 shRNA-derived tumors in vivo (Fig. 2e). Despite this, growth of these tumors was significantly greater than that of CT26FasL shRNA tumors (p < 0.0179). Altogether, these findings suggest that PGE2 may have additional effects on cells in the tumor microenvironment that are affected by ONO-8713 treatment.

Blocking EP1 receptor signaling enhances intratumoral CD8+ T cells, suppresses intratumoral CD4+ T cells and Treg cells and increases CTL activity

To assess whether blocking PGE2-EP1 signaling affects T cell recruitment, tumors excised after 48 days were dissociated with collagenase/dispase, and single-cell suspensions were analyzed by flow cytometry. Infiltration of the tumors by CD4+ T cells was significantly decreased in all groups on blocking EP1 receptor signaling (Fig. 3a). In contrast, CD8+ T-cell infiltration was increased, although this was not significant (Fig. 3a). Antagonizing the EP1 receptor also reduced the number of CD4+CD25+Foxp3+ Treg cells present in the tumors, with a significant reduction occurring following oral administration of ONO-8713 beginning at Day 20 and direct injection of ONO-8713 into the tumors (Fig. 3b). In contrast, the level of Treg cells in CT26EP1 shRNA tumors was increased, suggesting that EP1 receptor signaling in cells in the tumor microenvironment, rather than the tumor cells themselves, plays an important role in Treg cell recruitment and/or expansion. Given that Treg cells are potent suppressors of the antitumor immune response,[23] failure to CT26EP1 shRNA tumors to block the expansion of Treg cells may account for the enhanced growth of these tumors in vivo relative to ONO-8713-treated mice.

Figure 3.

Blocking EP1 receptor signaling alters the intratumoral immune cell population. (a) Flow cytometric analysis revealed that suppressing EP1 receptor signaling significantly reduces the percentage of CD4+ T cells and increases the percentage of CD8+ T cells within the tumors. Data points represent the mean ± SEM; n ≥ 5 mice per group. (b) Representative dot plots show the gating strategy for the identification of Treg cells within the tumor tissues. Treg cells were characterized as the ratio of CD25+ Foxp3+ cells in the CD4 gate. Treg cell infiltration was suppressed by EP1 receptor antagonism. Data points represent the mean ± SEM; n ≥ 5 mice per group. (c) CTL activity is recovered in ONO-8713-treated tumor-bearing mice and in mice inoculated with CT26EP1 shRNA tumor cells. Splenocytes were isolated and after 5 days in vitro stimulation with irradiated CT26 cells and IL-2, cytotoxic activity was assessed as described in the Material and Methods section. Data represent splenocytes from one or two mice per treatment group and is representative of three to five independent experiments (****p < 0.0001; ***p < 0.001; **p < 0.01). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

CTLs are a critical component of the immune response to tumors, with CTL activity declining with progressive tumor growth. To determine if blocking the EP1 receptor affects PGE2-mediated suppression of CTL activity, CTL were generated from tumor-bearing mice and examined for cytotoxic activity against CT26 cells (Fig. 3c). CTL from CT26scr shRNA tumor-bearing mice had an average CTL activity of 33%, whereas CTL from nontumor-bearing mice exhibited a cytotoxicity of 85%. Cytotoxicity was significantly increased in effector cells generated from tumor-bearing mice orally administered with ONO-8713 beginning at Day 20 (83%; p = 0.0031) and directly injected with ONO-8713 (92%; p < 0.0001), relative to those derived from untreated CT26scr shRNA tumor-bearing mice. Although CTL generated from CT26EP1 shRNA tumors (70%; p < 0.0001) and tumor-bearing mice orally administered ONO-8713 beginning at Day 27 (77%; p < 0.0001) also exhibited increased CTL activity relative to those generated from nontreated tumor-bearing mice (33%), interestingly, this CTL activity was significantly less than that of the PBS control mice (85%), p = 0.0029 and p < 0.0001, respectively.

Blocking EP1 alters TAM infiltration but does not play a role in the polarization of TAMs

Cytokines implicated in the differentiation of macrophages toward an immunosuppressive M2-like phenotype include PGE2 and IL-6.[20] Given the important role played by TAM in tumor development, the effect of blocking the EP1 receptor on TAM infiltration and polarization was determined. Analysis revealed significantly reduced levels of F4/80+ cells within the tumors when administration of the antagonist began at Day 20 or following direct injection of the antagonist (Fig. 4a). In contrast, TAM levels were comparable to that of control tumors when the initiation of antagonist administration was delayed until Day 27 when the tumors were larger and when EP1 was directly suppressed in the tumor cells.

Figure 4.

Antagonizing the EP1 receptor reduces macrophage infiltration but does not alter macrophage polarization. (a) Flow cytometric analysis revealed that suppressing EP1 receptor signaling reduces the percentage of F4/80+ cells within the tumors. Data points represent the mean ± SEM; n ≥ 5 mice per group. (b) The EP1 receptor does not play a role in TAM polarization. Macrophages were isolated from a minimum of five excised tumors per treatment group as described in the Material and Methods section. RNA was isolated and pooled, and the changes in transcription of IL-12, IL-6 and NOS2 were determined by RT-PCR. Data were normalized to β2-microglobulin and analyzed using the 2−ΔΔCt method. Results represent the mean ± SEM. (c) Blocking signaling through the EP1 receptor does not alter secretion of PGE2 by TAM. Isolated macrophages (n = 3 per group) were cultured ex vivo for 24 hr. Cell culture supernatants were harvested, and the level of PGE2 was measured in triplicate by ELISA.

These macrophages retained characteristics of the M2-like phenotype. They were IL-12low, IL-6low and nitric oxide synthase (NOSlow; Fig. 4b), with no change in PGE2 secretion (Fig. 4c). Although macrophages from ONO-8713-injected tumors were IL-12high, IL-6high, NOShigh and secreted significantly less PGE2, suggestive of a more antitumorigenic M1-like phenotype, this may be due to the high level of necrosis present in these injected tumors. Indeed, ingestion of necrotic cells has been shown to increase the transcription of several cytokines by macrophages.[24]

Discussion

PGE2 is the most abundant prostaglandin found in a variety of human malignancies, facilitating tumor progression by stimulating cell proliferation and angiogenesis and suppressing the antitumor immune response.[1, 19] PGE2 released from apoptotic tumor cells may also help to repopulate tumors following chemotherapeutic and radiotherapeutic regimens, with PGE2 being shown to stimulate the growth of therapy-resistant tumor cells.[25] Despite being one of the PGE2 receptors implicated in tumorigenesis,[11, 12] the protumorigenic effects of the EP1 receptor are very poorly understood. Importantly from a therapeutic standpoint, in our study, we demonstrate that targeting the EP1 receptor significantly retards the growth of established tumors in vivo. This inhibition of tumor growth was unlikely due to direct inhibition of PGE2-induced tumor cell growth, as suppressing EP1 receptor expression in tumor cells alone was far less effective in reducing tumor growth in vivo than when multiple cell types within the tumor microenvironment were affected. Moreover, growth of the tumor cells in vitro was unaffected by suppression of the receptor.

One of the potential mechanisms by which the EP1 receptor mediates its protumorigenic effects is upregulation of FasL expression in tumor cells.[9] Fas signaling in response to binding of FasL has been shown to have numerous protumorigenic effects, with expression of Fas and FasL by malignant cells being associated with enhanced tumor growth, inflammation, metastases and apoptotic depletion of tumor-infiltrating lymphocytes in vivo.[26, 27] Indeed, suppressing EP1 receptor signaling was found to effectively suppress FasL expression by tumor cells in vitro and in vivo. However, despite exhibiting reduced FasL expression, the reduction in growth of CT26EP1 shRNA tumors was far less than that of CT26FasL shRNA tumors or ONO-8713-treated tumors. Although CT26EP1 shRNA clones exhibited reduced FasL expression, the level of expression was greater than that seen in the CT26FasL shRNA clones and following treatment with ONO-8713 in vitro. Oligomerization of FasL is required for triggering of Fas signaling,[28] and thus, the threshold of FasL expression may play a role in determining whether FasL promotes tumorigenesis, with the CT26EP1 shRNA tumor cells exhibiting a level of FasL expression sufficient to mediate the protumorigenic effects of FasL. Alternatively, the greater reduction in tumor growth seen in ONO-8713-treated tumors may be due to the effect of the antagonist on PGE2/EP1 receptor signaling in cells in the tumor microenvironment.

Immune cells are a prominent component of solid tumors. Such cells, if appropriately activated, can mediate tumor rejection.[20] The EP receptors are expressed by multiple immune cell types,[29] and thus, the EP1 antagonist could potentially suppress EP1 receptor signaling in these cells. In our study, we found that the protumorigenic effects of the EP1 receptor are also immunological in nature. Blocking EP1 receptor signaling reduced the level of intratumoral CD4+ T cells and increased the level of CD8+ T cells. CD4+ and CD8+ T cell recruitment was also altered in CT26EP1 shRNA tumors, suggesting that this alteration in T-cell recruitment may be due to changes in secretion of T-cell chemotactic factors by the tumor cells. For instance, both IL-16 and CCL5 (RANTES) have preferential effects on CD4+ T-cell chemotaxis.[30] Alternatively, this difference in the level of the intratumoral CD4+ and CD8+ T cells could be due to differences in the sensitivity of the T-cell subsets to the antiproliferative effects of PGE2, with CD8+ T cells being shown to be more susceptible to PGE2-mediated inhibition of proliferation than CD4+ T cells.[31] Although EP2 and EP4 receptors have been shown to predominantly mediate the antiproliferative activity of PGE2 on lymphocyte proliferation,[29] this study did not subdivide the lymphocytes into CD4+ and CD8+ cells.

CD8+ T cells, if appropriately activated, can mediate tumor rejection, with CD8+ CTL among the major antitumor effector mechanisms.[32, 33] Such antitumor activity is strongly suppressed by Treg cells.[23] Together with altered T-cell infiltration, blocking EP1 receptor signaling was also associated with reduced levels of Treg cells within the tumors and enhanced cytotoxic activity of T cells. Indeed, mice with the greatest reduction in tumor growth in vivo also had the greatest reduction in Treg cells and the highest CTL activity. Changes in the level of Treg cells and CTL activity are likely due to effects of the antagonist on nontumor cells in the tumor microenvironment, as the level of Treg cells remained high in CT26EP1 shRNA tumors. CTL generated from splenocytes from these mice also exhibited the least CTL activity of all the treatment groups. Indeed, this change in the level of Treg cell infiltration represents one of the major differences between ONO-8713-treated tumor cells and EP1-suppressed CT26EP1 shRNA tumor cells, suggesting that PGE2 acts in a paracrine fashion to recruit Treg cells to tumors.

How blocking the EP1 receptor alters Treg cell infiltration and/or expansion is unclear. Induction of Treg cells and suppression of CTL activity in response to PGE2 have previously been ascribed to activation of cAMP/protein kinase A (PKA) by EP2 and EP4.[34, 35] The EP1 receptor has traditionally been associated with the activation of Ca2+ signaling through coupling to Gq and the activation of phospholipase C. A recent study, however, has shown that the EP1 receptor can also activate PKA, independently of cAMP,[36] which may potentially account for this previously unknown role for the EP1 receptor in the induction of Treg cells in the tumor microenvironment.

The level of TAM present within the tumors was also suppressed on blocking of the EP1 receptor in the tumor microenvironment. TAMs are a major constituent of the leukocyte infiltrate in solid tumors and are recruited to tumors by tumor-derived chemotactic factors.[37] Such TAM have been shown to be skewed in tumors towards an immunosuppressive or M2-like phenotype by environmental cues such as PGE2, IL-10 and IL-6.[20, 38, 39] Skewing of TAM toward this M2 phenotype favors tumor progression by suppressing T-cell proliferation, stimulating tumor cell proliferation, angiogenesis, tumor cell migration and increasing stroma reaction. Consistent with the ability of TAM to promote tumor progression, those tumors that showed the greatest suppression of tumor growth also showed the greatest reduction in TAM. This reduction in TAM was not due to alterations in CCL2 (data not shown), suggesting other macrophage–chemotactic factors may play a more important role in this model. Moreover, although the level of PGE2 present in the tumor microenvironment following blocking of the EP1 receptor was unaltered, the macrophages retained an M2-like phenotype, suggesting that the EP1 receptor is involved in TAM recruitment but not polarization.

In conclusion, we have shown that the EP1 receptor mediates several of the protumorigenic effects of PGE2 and that EP1 receptor antagonism is effective in reducing the growth of established tumors. Given that we have previously shown that human colon tumors in vivo express the EP1 receptor,[9] and that EP1 receptor antagonists inhibit chemically induced breast cancer development in rats[40] and reduce the number of skin tumors per mouse following UVB exposure,[41] EP1 receptor antagonists may be good candidates as chemotherapeutic agents for not only colon but also other cancers.

Acknowledgements

The authors thank Jacquie Kelly, Kinga Gebolys and Aine Dorgan for their excellent technical assistance and the Molecular Virology Diagnostic and Research Laboratory (MDVRL) for the use of equipment.

Ancillary