Regulatory T cells in colorectal cancer patients suppress anti-tumor immune activity in a COX-2 dependent manner
- First Online:
- Cite this article as:
- Yaqub, S., Henjum, K., Mahic, M. et al. Cancer Immunol Immunother (2008) 57: 813. doi:10.1007/s00262-007-0417-x
- 265 Views
Naturally occurring regulatory T (TR) cells suppress autoreactive T cells whereas adaptive TR cells, induced in the periphery, play an important role in chronic viral diseases and cancer. Several studies indicate that cyclooxygenase (COX) inhibitors prevent cancer development of colon adenomas and delay disease progression in patients with colorectal cancer (CRC). We have shown that adaptive TR cells express COX-2 and produce PGE2 that suppress effector T cells in a manner that is reversed by COX-inhibitors.
Methods and results
Here we demonstrate that CRC patients have elevated levels of PGE2 in peripheral blood, and CRC tissue samples and draining lymph nodes display increased numbers of FOXP3+ TR cells. Depletion of TR cells from PBMC enhanced anti-tumor T-cell responses to peptides from carcinoembryonic antigen. Furthermore, the COX inhibitor indomethacin and the PKA type I antagonist Rp-8-Br-cAMPS significantly improved the anti-tumor immune activity.
We suggest that adaptive TR cells contribute to an immunosuppressive microenvironment in CRC and inhibit effector T cells by a COX-2–PGE2-dependent mechanism and thereby facilitate tumor growth. Therapeutic strategies targeting TR cells and the PGE2–cAMP pathway may be interesting to pursue to enhance anti-tumor immune activity in CRC patients.
KeywordsHumanColorectal cancerRegulatory T cellsCOX-2PGE2
Cyclooxygenase type 2
- TR cells
Regulatory T cells
Colorectal cancer (CRC) ranks second in incidence and mortality among malignant disease in both men and women in developed countries, although the recent advances in surgical technique and use of neo-adjuvant chemotherapy have improved clinical outcomes [4, 23, 46]. Most cases of CRC are sporadic, and both genetic background and environmental factors are important. Increasing evidence suggests that chronic inflammation plays a role in the pathogenesis of cancer, including CRC [1, 10]. Epidemiologic studies indicate that regular use of aspirin or nonsteroidal anti-inflammatory drugs (NSAIDs) can reduce the incidence of CRC in humans by 40–50% [17, 20, 39]. The chemopreventive role of NSAIDs and aspirin is thought to be exerted through the inhibition of cyclooxygenases (COXs), specifically COX-2, and prostaglandin (PG) production . Exogenous PGE2 treatment protects small intestinal adenomas from NSAID-induced regression in APCmin mice, a murine model for familial adenomatous polyposis (FAP), directly demonstrating that the anti-tumor effect of NSAIDs is related to reduction of PGE2 levels . However, use of COX-2 inhibitors to prevent CRC in humans has been connected with an increased frequency of adverse cardiovascular events [3, 34].
COX exists in two isoforms, named COX-1 and COX-2. COX-1 is constitutively expressed in a broad range of cells and tissues, and its enzymatic activity is important for renal function, gastric mucosal integrity and platelet aggregation. In contrast, COX-2 expression is normally absent in most cells and tissues, but is induced in response to proinflammatory cytokines . The COX-2 derived PGE2 plays a predominant role in inflammatory diseases such as arthritis. PGE2 exerts its cellular effects by binding to its cognate receptors (EP1–EP4) that belong to the family of seven transmembrane G protein coupled receptors.
Several reports have demonstrated that PGE2 promotes tumor growth through stimulation of EP receptor signaling with subsequent enhancement of cellular proliferation, promotion of angiogenesis, inhibition of apoptosis, stimulation of tumor invasion, and suppression of immune responses . Possible mechanisms involved are activation of peroxisome proliferator-activated receptors δ (PPARδ), epidermal growth factor receptor (EGFR), Ras-MAPK pathway, nuclear factor kappa B (NFκB), and Wnt signaling . The central role of PGE2 in tumorigenesis has been further confirmed through homozygous deletion of its receptors. Mice with deletions in EP1 and EP4 receptors, but not in EP3, were partially resistant to colon carcinogen mediated induction of abnormal crypt foci [29, 45]. Furthermore, EP2 disruption in APCΔ716 KO mice leads to decreased size and number of polyps . Both PGE2 production and level of COX-2 expression is increased in CRC compared with normal colonic mucosa [14, 21]. A recent study has shown association between high levels of a urinary PGE2 metabolite and CRC risk .
Carcinoembryonic antigen (CEA) is a 180 kD glycoprotein that is highly expressed in colon epithelial cells during embryonic development. The level of expression is low in adult tissues, but is overexpressed in almost all CRC. Several reports have shown that CEA is immunogenic, making it an attractive target for immunotherapeutic purposes and cancer vaccine development . CEA-epitopes that induce anti-tumor immune responses in both CD4+ and CD8+ T cells have been identified [6, 19].
The role of CD4+CD25+ regulatory T (TR) cells in suppressing anti-tumor immune activity is subject to considerable interest [30, 36]. TR cells express the forkhead transcription factor FOXP3 and maintain immunological tolerance by suppression of autoreactive T cells [12, 15, 35]. Naturally occurring TR cells develop in the thymus, however CD4+CD25- T cells in the periphery convert to CD4+CD25+FOXP3+ adaptive TR cells and acquire regulatory function under certain circumstances [28, 43]. Accumulating evidence indicates that adaptive TR cells are enriched in tumor tissue, draining lymph nodes, malignant effusions and peripheral blood from patients with various cancers, suggesting that the tumor microenvironment may foster immune tolerance by attracting and/or inducing dominant inhibitory cells .
We have shown that continuous antigenic stimulation of human peripheral CD4+CD25- T cells gives rise to CD4+CD25+FOXP3+ adaptive TR cells which suppress autologous CD4+CD25- responder T cells . Furthermore, we have reported a novel mechanism of immune suppression whereby adaptive TR cells express COX-2 and suppress effector T cells in a PGE2-dependent fashion that can be reversed with COX-2 inhibitors or PGE2-receptor antagonists . In addition, we and others have implicated PGE2 in induction of FOXP3 in CD4+CD25- T cells that contributes to modulation of T cell function [2, 28]. In a murine lung cancer model, inhibition of COX-2 enhanced anti-tumor immune responses [37, 38]. In line with these observations, we hypothesized that chronic immune stimulation in a cancer microenvironment induces TR cells that inhibit anti-tumor immune responses in a COX-2–PGE2–cAMP-dependent manner. In this study we demonstrate that CRC patients have high levels of PGE2 in plasma, and that TR cells suppress anti-CEA immune responses in a manner that can be reversed by COX-inhibitor or PKA-antagonist to the same extent as depletion of TR cells.
Materials and methods
Patients and clinical study control
Clinical characteristics of patients included in the study of tumor tissue analysis and anti-tumor immune analysis
Tumor size (mm)
Isolation of cells
30–50 ml of peripheral blood was drawn from patients or controls in EDTA tubes. Peripheral blood mononuclear cells (PBMC) were isolated by Isopaque-Ficoll (Lymphoprep, Nycomed Pharma AS, Oslo) gradient centrifugation. CD4+CD25+ T cells were isolated using a CD4+CD25+ regulatory T cell isolation kit according to the manufacturer’s manual (Miltenyi Biotec, Auburn, CA). The cells were routinely analyzed by flow cytometry and the purity of CD4+CD25+ and CD4+CD25- T cell populations were consistently >98%. In anti-tumor immune assays, CD25+ cells were depleted directly from PBMC after incubation with CD25 microbeads (Miltenyi Biotec, Auburn). Cells were cultured in RPMI 1640 (Gibco, Paisley, UK) supplemented with 10% heat-treated FCS, 100 U/ml penicillin/streptomycin, 1 mM sodium pyruvate and 1:100 non-essential amino acids (further described as complete medium).
We selected six sequences of the CEA protein for peptide synthesis, based on previously published data . The peptides were synthesized using an in-house multipep automated peptide synthesizer (INTAVIS Bioanalytical Instruments AG) following standard Fmoc-chemistry protocol with following sequences: CEA 1–6D [YLSGADLNL] , CEA 61–69 [HLFGYSWYK] , CEA 268–278 [QYSWFVNGTF], CEA 318–327 [TYACFVSNL] , CEA 570–579 [SYLSGANLNL] , and CEA 571–579 [YLSGANLNL] . Scrambled control peptides were generated the same way with the sequences: [SNYLLALGN], [SGYLLALDN], [LGSYHFWKY], [NYSLSLALGN], [YWVGFQFSNT], and [NVLFSTCYA]. The peptides were subjected to high performance capillary electrophoresis (HPCE), isolated with >90% purity and quality assessed by mass spectrometry. They were dissolved in DMSO at a concentration of 5 mg/ml and further diluted in complete medium and used at a final concentration of 5 μg/ml.
Flow cytometric analysis of lymphocytes
Cells were fixed in 4% paraformaldehyde (PFA) and permeabilized in FACS permeabilizing solution (BD BioSciences, San Jose, CA) prior to staining with CD4 APC, CD8 APC, CD3 PerCP, CD25 PE, anti-human IFN-γ PE, anti-human TNF-α FITC, (BD BioSciences Pharmingen, Cayman Chemicals). The CD25+ T cells were tested for FOXP3 expression by staining with APC anti-human FOXP3 staining kit, according to manufacturer’s instructions (clone PCH-101, eBioscience). The cells were washed once in PBS containing 1% BSA before acquiring data on a flow cytometer (FACSCalibur, BD Biosciences, San Jose, CA) and analyzed using FlowJo software (Tree Star, San Carlos, CA).
Peptide stimulation and cytokine assays
The PBMC, PBMC depleted of CD25+ cells, or PBMC co-cultured with autologous CD25+ T cells were stimulated with 5 μg/ml staphylococcal enterotoxin B (SEB) (Sigma-Aldrich), CEA-peptides (5 μg/ml), or complete medium for 18–20 h. Brefeldin A (Sigma-Aldrich) was added to a final concentration of 5 μM for the last 6 h of incubation. When used, COX inhibitor indomethacin (25 μM) (Sigma-Aldrich) or PKA type I antagonist Rp-8-Br-cAMPS (1 mM) (Lauras AS, Oslo, Norway) were added 90 min prior to activation with SEB or CEA-peptides. In all co-culture experiments, CD25+ T cells were prestained with 2 μM CFSE in order to separate TR cell population from responding effector T cells. The cells were fixed, permeabilized, stained for CD3, CD4, IFN-γ and TNF-α and analyzed by flow cytometry as described above. The SEB was used as a positive control for activation in the assay. The cytokine levels in un-stimulated cells were subtracted from the levels of activated cells to look at the induction in cytokine secretion.
Prostagandin E2 measurement by ELISA
Peripheral blood from patients and controls was centrifuged and plasma isolated and stored at −80°C, thawed, and subsequently analyzed for PGE2 concentration according to the manufacturer’s instructions (R&D, London, UK).
Preparation of tissue specimens and multicolor immunostaining
Large intestine resection specimens were fixed in 10% buffered formalin and processed by standard methods for pathologic evaluation. Relevant specimens were obtained from the tumor and unaffected mucosa (n = 11) as well as from regional draining lymph nodes (n = 12). Formalin-fixed and paraffin-embedded tissue was cut at 4 μm, deparaffinized, rehydrated and then subjected to antigen retrieval by cooking for 20 min in 0.01 M citrate buffer (pH 6.0) in a water bath at 98°C. Two-color immunofluorescence staining was performed as described elsewhere . To determine the density and phenotype of FOXP3+ T cells a rat mAb specific for FOXP3 (clone PCH101, IgG2a, 1/20; eBiosciences) was combined with mouse mAbs to either CD25 (clone 4C9, IgG2b, 1/100; Novocastra, UK) or COX-2 (#160112, IgG1, 1 μg/ml; Cayman Chemicals, Ann Arbor, MI). Primary antibody combinations were incubated overnight at +4°C followed by incubation with a mixture of Cy3-labeled donkey anti-rat IgG (2.5 μg/ml; Jackson ImmunoResearch Laboratories, West Grove, PA) and biotinylated subclass-specific goat anti-mouse IgG2b (10 μg/ml) or biotinylated goat anti-mouse IgG1 (2.5 μg/ml) (both from Jackson ImmunoResearch Laboratories) for 3 h, and finally Cy2-streptavidin (1 μg/ml; Amersham, UK) for 1 h. In addition, a rabbit antiserum to CD3 (1/50; DAKO, Denmark) was incubated overnight either with anti-FOXP3 (clone PCH101) or anti-COX2 (# 160112) followed by Alexa 488-labeled goat anti-rabbit IgG (2 μg/ml; Molecular Probes Invitrogen) combined with either Cy3-labeled donkey anti-rat IgG or Cy3-labeled goat anti-mouse IgG1 (Jackson ImmunoResearch Laboratories) for 3 h. In all staining experiments, negative controls were obtained both by omission of primary mAbs and by parallel incubation with irrelevant isotype- and concentration-matched primary mAbs. The specimens were examined at ×40 magnification by confocal microscopy (Leica, TCS SP, Heidelberg, Germany) and images were recorded with the confocal acquisition program (Leica TCS-NT). The images were further processed using Adobe Photoshop and Adobe Illustrator.
The data were analyzed using SPSS for Windows. The plasma levels of PGE2 in patients and controls were compared using Mann–Whitney U test. Anti-CEA responses in patients and controls were compared by Wilcoxon signed ranks test. Differences with two-sided P < 0.05 were considered significant.
CRC patients have elevated levels of PGE2 in plasma
FOXP3+ TR cells accumulate in CRCs and regional lymph nodes
TR cells inhibit anti-tumor immune responses in patients with CRC
TR cells inhibit anti-CEA responses in CRC patients in a COX-2–PGE2-dependent manner
Tumor cells are immunogenic and tumor-infiltrating lymphocytes (TILs) in melanoma, CRC, and ovarian cancer inhibit tumor growth and are associated with improved prognosis [9, 32, 49]. Interestingly, the type, density, and location of immune cells in CRC may have a higher predictive power than the prognosis estimated by the UICC-TNM classification . However, tumor-specific TR cells at the tumor site play a significant role in the suppression of anti-tumor immune activity which may adversely affect the prognosis . Several studies have shown that depletion of TR cells promote anti-tumor immune responses and tumor rejection [8, 41]. Several human cancers, including gastrointestinal, lung, and ovarian tumors, are also associated with a higher proportion of TR cells in the peripheral blood [8, 11, 36, 47]. In ovarian cancer the presence of TR cells in TILs has been shown to be an independent risk factor associated with poor prognosis . Numerous tumor-derived soluble factors such as vascular endothelial growth factor, IL-10, TGF-β and PGE2 may also act in concert to establish an immunosuppressive microenvironment which promotes T cell tolerance and tumor immune evasion . Consequently, modulation of the regulatory function of TR cells may be an important avenue to improve anti-tumor immune activity.
Continuous antigen stimulation of naïve human peripheral CD4+ T cells leads to induction of CD4+CD25+FOXP3+ TR cells that express COX-2 and suppress effector T cells by a PGE2–cAMP dependent mechanism . Here we show that patients with CRC have high plasma levels of PGE2, that TR cells from these patients inhibit anti-CEA immune responses and that this immunosuppression can be reversed ex vivo in an equally effective manner by treatment with the COX-inhibitor indomethacin or the PKA type I antagonist Rp-8-Br-cAMPS as by depletion of TR cells. However, in some patients COX-inhibitor improved the anti-CEA immune responses more than depletion of TR cells. Thus, the role of PGE2 from other immune cells cannot be ruled out. In the present study we have assessed T cell immune responses in PBMC cultures, and PGE2 from activated monocytes may have contributed to the additional immunosuppression unmasked by COX-inhibitor.
Several studies have demonstrated a significant infiltration of FOXP3+ TR cells into CRCs [26, 27]. The presence of TR cells in the tumor microenvironment and in draining lymph nodes with metastasis diminishes anti-tumor immune responses; however, previous studies have not elucidated the mechanism of suppression. In our study, the majority of TR cells at the tumor site were COX-2 negative or their presence were overshadowed by the strong staining in the CRC cells. However, COX-2+ TR cells were present in draining lymph nodes with and without metastasis, suggesting that COX-2+ TR cells contribute to the immune suppression by secretion of PGE2 in the lymph nodes where tumor antigens are presented by APC. The high expression of COX-2 in tumor epithelial cells is probably the main source of the elevated PGE2 level in plasma observed in these patients, which may drive both tumorigenesis, induction of TR cells, and local immune suppression. However, as indomethacin reverses the TR cells-mediated suppression of anti-tumor immune responses in PBMC from CRC subjects ex vivo where plasma PGE2 is washed out, the production of PGE2 in COX-2+ TR cells appears to be sufficient to inhibit immune responses. This argues that TR cells mediated PGE2 production in situ in draining lymph nodes is sufficient to inhibit the immune responsiveness to metastatic cells from the primary tumor. COX-inhibitors prevent development of CRC by several mechanisms as PGE2 affects cell proliferation, migration, apoptosis and angiogenesis. Improved anti-tumor immune function by inhibition of TR cells may be another mechanism that contributes to the beneficial effects of COX-inhbitors.
In conclusion, the present report supports the notion that TR cells suppress anti-tumor immune responses in a COX-2–PGE2–cAMP-dependent manner and that this mechanism may represent a target for future cancer immunotherapy.
We are grateful for the routine pathology data made available from Department of Pathology, Ullevaal University Hospital and technical assistance with preparation of tissue specimens and multicolor immunostaining by Linda Kristiansen, Vigdis Wendel, and Aaste Aursjo (The Pathology Clinic, Rikshospitalet-Radiumhospitalet Medical Center). This work was supported by grants from the Norwegian Functional Genomics Programme (FUGE), The Research Council of Norway, The Norwegian Cancer Society, Novo Nordic Foundation Committee, and the European Union grant no. 037189. S. Yaqub and M. Mahic are fellows of the Norwegian Cancer Society.