A study was carried out to determine the functional attributes of CD4+ CD25+ regulatory T cells in cancer progression by suppressing antitumor immunity.
A study was carried out to determine the functional attributes of CD4+ CD25+ regulatory T cells in cancer progression by suppressing antitumor immunity.
Triple-color flow cytometry was used to study the phenotype expression of CD4+ CD25+ regulatory T cells and CD8+ T cells in the peripheral blood lymphocytes (PBLs) and tumor-infiltrating lymphocytes (TILs) of 57 cases of stage I to IV endometrial carcinoma. The expression of T cell subsets was correlated with clinical prognostic parameters.
The prevalence of CD4+ CD25+ T cells was significantly higher in the TILs than PBLs. The expression of CD4+ CD25+ regulatory T cells in cancer milieu correlated with the tumor grade, stage, and myometrium invasion. The expression of FOXP3 and GITR in CD4+ CD25+ regulatory T cells was lower in PBLs than TILs. Most tumor-infiltrating CD8+ T cells were CD28− CD45RA− CD45RO+ CCR7−, suggesting good terminal differentiation. Most of them had an activated role with CD69+ CD103+ CD152+. Functionally, both granzyme B and perforin were scarcely expressed in peripheral regulatory T cells but were highly expressed in peripheral regulatory T cells in the tumor microenvironment. In contrast, CD8+ cytotoxic T cells derived from PBLs expressed both granzyme B and perforin, and at significantly higher levels than in TILs. Further functional assays demonstrated that Th1 cytokines and cytotoxic molecules can be synchronously up-regulated in CD8+ cytotoxic T cells.
Regulatory T cells in the tumor microenvironment may abrogate CD8+ T cell cytotoxicity in a granzyme B- and perforin-dependent conduit. Decreases in both Th1 cytokines and cytotoxic enzymes are relevant for regulatory T cell-mediated restraint of tumor clearance in vivo. Of clinical significance, the expression of regulatory T cells in TILs may mediate T cell immune repression within cancer milieu and thus greatly correlate with cancer progression. Cancer 2010. © 2010 American Cancer Society.
CD4+ CD25+ T cells have been shown to mediate the process of infection, tolerance of transplantation, autoimmunity, and tumor immunity.1-4 It has been proposed that regulatory T cells can attenuate antitumor immune responses by suppressing the proliferation and cytokine production of effector T cells in cancer milieu.5, 6 Three types of CD4+ regulatory T cells have been partly characterized in humans. The first is CD4+ CD25-IL-10+ Foxp3lowtype 1 T regulatory (Tr1) cells, which are induced in the periphery on encountering antigens and secrete interleukin (IL)-10 and transforming growth factor (TGF)-β.7, 8 The second type is naturally occurring CD4+ CD25highFoxp3+ T cells, which arise directly in the thymus and have the ability to suppress the responses of both CD4+ CD25− and CD8+ CD25− T cells.9 Naturally occurring CD4+ CD25highFoxp3+ T cells express the transcription factor forkhead box p3 (Foxp3), glucocorticoid-induced tumor necrosis factor receptor family-related gene (GITR), and intracellular cytotoxic T lymphocyte-associated antigen-4. The third type is Th3 cells, which are dependent on IL-4 for functional differentiation.10
Recently, human studies have suggested that both IL-10 and TGF-β1 production by regulatory T cells may mediate immune suppression in the tumor microenvironment.6, 8 Importantly, in clinical human studies, an increased proportion of regulatory T cells in cancer milieu has been detected in patients with cancers of a diverse nature, which implicates a poor clinical outcome.11-16 In the case of tumors, manipulation of regulatory T cells holds great promise in immunotherapy for cancer. Several studies indicate that regulatory T cells can inhibit effective antitumor immune responses, and that their removal conversely endorses tumor rejection.17 The increasing number of studies of regulatory T cells in patients with cancer also points to the key role of these cells in related disease progression.17 However, the mechanisms by which regulatory T cells suppress antitumor immunity remain largely undefined.
In the present study, we tried to verify the essential immune-regulatory roles of regulatory T cells from tumor-infiltrating lymphocytes (TILs) of cancer patients. We compared the subsets of regulatory T cells in TILs and peripheral blood lymphocytes (PBLs) from patients with endometrial cancer (EC). Furthermore, we measured the expressions of cytokines and cytotoxic molecules of CD8+ T cells in functionally associated regulatory T cells from cancer milieu.
A total of 57 patients with stage I to IV EC were enrolled prospectively in this study between March 2006 and December 2008. All patients who underwent surgery signed the institutional review board-approved informed consent for release of specimens for research purposes. After the staging operation, the surgical specimens were examined carefully by experienced pathologists to exclude the possibility of coexisting malignancy. Each case of EC was evaluated for clinical and pathological parameters, including surgical stage, lymphatic or vascular permeation, lymph node metastatic status, and histologic grade. The histologic grades of EC included grade I, grade II, and grade III. Surgical staging of each patient was defined according to the 1988 modification of the International Federation of Gynecologists and Obstetricians staging for EC.
Tissue specimens were aseptically excised immediately after surgery from at least 4 different tumor sites. Tissue fragments were carefully washed and reperfused with phosphate-buffered saline to remove contaminated blood. TILs and PBLs were isolated by methods described in our previous study.18-22
PBLs and TILs were thawed and cultured in RPMI-1640 containing 10% human AB sera for 4 hours at 37°C in 5% CO2 in multiwell plates or flasks depending on the cell number at 1 × 106/mL. Cells stimulated by anti-CD3/anti-CD28–coated beads were first depleted of monocytes by a 2-hour adherence in tissue culture flasks. Nonadherent cells were removed and cultured at a 3:1 bead:cell ratio at a final concentration of 1 × 106/mL in RPMI-1640 containing 10% human AB sera for 2 days at 37°C in 5% CO2. Brefeldin A (10 mg/mL) was added 4 hours before analysis. Cells were harvested, and beads were removed by magnetic separation.
CD4+ T cells were enriched by magnetic cell sorting using a MiniMACS separator (Miltenyi Biotec, Bergisch Gladbach, Germany). Monoclonal antibodies labeled with fluorescein isothiocyanate (FITC), phycoerythin (PE), and peridinin chlorophyll protein (PerCP) (Immunocytometry System; Beckton-Dickinson, San Jose, CA) were used for 3-color flow cytometry. The following matchings were arranged: a mixture of PE-coupled specific monoclonal antibodies (MAbs): anti-CD25, anti-FoxP3, anti-45RO, anti-CD28, anti-CD69, anti-CD152, anti-GITR, anti-IL-2, anti-IL4, anti-IL10, anti-interferon (IFN)-γ, anti–tumor necrosis factor (TNF)-α; a mixture of FITC-coupled specific MAbs: anti-CD4, anti-CD103, anti-CD45RA, anti-CCR7, anti–granzyme B, antiperforin; and a mixture of PerCP-coupled specific MAbs: anti-CD8, anti-CD4, anti-CD25, and anti-CD3. For intracellular staining, cells were fixed and permeabilized with a Beckton-Dickinson Cytofix/Cytoperm Plus Fixation/Permeabilization Kit according to the manufacturer's instructions. A Simultest control (mouse immunoglobulin [Ig]G1-FITC+ IgG2a-PE) was used as a background control. Three-color flow cytometry was performed on a FACSCaliber flow cytometer (Beckton-Dickinson). The regional gate was set on FL1 (anti-CD3-PerCP). Flow cytometry data was analyzed using WinMDI software.
The percentage expression of a given marker was obtained for a subset of T lymphocytes. All data were expressed as mean ± standard error unless otherwise indicated. The t test for paired data was used to compare paired samples from the same patients. The association between clinical prognostic parameters and CD4+ CD25+ regulatory T cells expression was assessed using Pearson correlation test. Statistical significance was defined by a P value <.05.
Samples of peripheral blood and malignant endometrial tissue from EC patients were screened by flow cytometry for CD4+ CD25+ T cells (Fig. 1A). The mean ratio of CD4+ CD25+ regulatory T cells in the PBLs was significantly lower than that in TILs (23% ± 11% vs 31% ± 12%, respectively, P = .001; Fig. 1B).
The regulatory T cells from both PBLs and TILs expressed membranous CD4, CD25, and CD45RO, but rarely expressed CCR7+ CD45RO− (naive) (13% and 2%, respectively) (Fig. 2A). The regulatory T cells from TILs rarely expressed CCR7. The circulating regulatory T cells expressed significantly higher CCR7+ CD45RO+ (48% and 24%, respectively, P = .029) (central memory) and lower CCR7-CD45RO+ (35% and 65%, respectively, P = .009) (effector memory) than the tumor-infiltrating regulatory T cells. By gating on the CD4+ CD25+ regulatory T cells, 4 additional cell markers could be studied, FOXP3, CD152, CD103, and GITR (Fig. 2B). High levels of FOXP3, GITR, and CD152 expression were detected in CD4+ CD25+ regulatory T cells in all samples. The circulating regulatory T cells expressed almost no CD103. The expressions of FOXP3, CD103, and GITR on CD4+ CD25+ regulatory T cells were lower in those from PBLs than those from TILs (FOXP3, 38% and 60%, respectively, P = .025; CD103, 4% and 31%, respectively, P = .001; GITR, 20% and 65%, respectively, P = .040). Similar levels of CD152 expression were detected in CD4+ CD25+ regulatory T cells from blood and tissue (33% and 36%, respectively, P = .776).
The majority of TILs were CD4+ and CD8+ T lymphocytes, although the CD4+/CD8+ cell ratio varied substantially between samples. The expression of intracellular granzyme B and perforin, both markers of cytotoxic potential, was determined on CD4+ CD25+ regulatory T cells and CD8+ T lymphocytes from both peripheral blood and EC tissue. Analysis of the CD4+ CD25+ regulatory T cells derived from TILs and PBLs showed that the expression of granzyme B and perforin could be detected at diverse levels on CD4+ CD25+ regulatory T cells from TILs (granzyme B was detected in 12% and perforin in 19%), and was significantly more prominent than expression in PBLs. In contrast, only minimal expression levels of granzyme B and perforin could be detected on gated CD4+ CD25+ regulatory T cells from PBLs (granzyme B was detected in 3% and perforin in 5%) (Fig. 3A). Analysis of the CD8+ T lymphocyte fraction derived from TILs and PBLs showed that only minimal expression levels of granzyme B and perforin could be detected in gated CD8+ T cells from TILs (granzyme B was detected in 5% and perforin in 7%) (Fig. 3B). In contrast, the expression of granzyme B and perforin could be detected at diverse levels on CD8+ T lymphocytes from PBLs (granzyme B was detected in 13% and perforin in 34%), and was significantly more prominent than in those from TILs.
The percentage of CD4+ CD25+ regulatory T cells inversely correlated with the percentage of granzyme B- and perforin-expressing CD8+ T lymphocytes (Fig. 3C). There was a significant negative correlation between CD4+ CD25+ regulatory T cells and granzyme B-expressing (r = −0.651, P < .001) and perforin-expressing CD8+ T lymphocytes derived from PBLs (r = −0.708, P < .001). The expression of CD4+ CD25+ regulatory T cells was also strongly negatively correlated to granzyme B-expressing (r = −0.586, P = .004) and perforin-expressing CD8+ lymphocytes derived from TILs (r = −0.549, P = .042). Therefore, the existence of CD4+ CD25+ regulatory T cells may contribute, at least in part, to the functional compromise of CD8+ TILs.
The CD8+ lymphocytes from PBLs and TILs both expressed membranous CD28, CD45RA, and CD45RO (Fig. 4A, B). The CD8+ T lymphocytes from EC did not express CCR7 (Fig. 4B). The circulating CD8+ lymphocytes had a significantly more prominent fraction of CCR7+ CD45RO-naive T cells (24% and 1%, respectively, P < .001) and a smaller fraction of CCR7− CD45RO+ effector memory T cells (35% and 75%, respectively, P < .001) than infiltrating regulatory T cells. Similar fractions of CCR7+ CD45RO+ central memory T cells were detected in the CD8+ T lymphocytes derived from autologous blood and tissue samples (16% vs 11%, P = .133). The infiltrating CD8+ T lymphocytes expressed significantly higher CD69 (82% vs 2%, P < .001), CD103 (34% vs 0.4%, P < .001), and CD152 (38% vs 13%, P < .001) (Fig. 4C), but lower CD28 (34% vs 66%, P < .001) and CD45RA (35% vs 75%, P < .001) than circulating regulatory T cells (Fig. 4A). Most tumor-infiltrating CD8+ T lymphocytes were CD28− CD45RA− CD45RO+ CCR7−, suggesting good terminal differentiation. In addition, most of them were CD69+ CD103+ CD152+, signifying their activated role in the cancer milieu.
To assess the functional polarity of TILs, we measured the fractions of T cells secreting cytokines after anti-CD3/anti-CD28 costimulation. After stimulation, the percentage of TILs producing IL-2, IFN-γ, and TNF-α was similar to the PBLs (Fig. 5). The anti-CD3/anti-CD28–coated beads were able to induce production of IL-2, IFN-γ, and TNF-α in TILs derived from EC. We also measured the levels of the Th2 cytokines IL-4 and IL-10 in the TILs and found them to be at low levels similar to those in the PBLs. There was no significant production of Th2 cytokines in the TILs after stimulation. Our data demonstrated that the functional polarity of TILs and the responding capacity were intact, and that these T cells were able to mount a response against a tumor after appropriate activation stimuli.
To further study the correlated observation of Th1 cytokine production in TILs and PBLs from EC patients, we assessed the status of various cytotoxic molecules (perforin and granzyme B) on CD8+ T cells (Fig. 5). The mean percentage of T cells expressing perforin and granzyme B after stimulation was determined by intracellular staining and flow cytometry. After adequate in vitro stimulation, CD8+ T cells derived from TILs produced these cytotoxic molecules at levels similar to those in PBLs.
The presence of CD4+ CD25+ regulatory T cells in PBLs correlated with the tumor grade, stage, and myometrium invasion (Fig. 6). There were significant positive correlations between CD4+ CD25+ T cells and histological tumor grade (r = 0.696, P < .001) as well as cancer stage (r = 0.513, P < .001). The existence of CD4+ CD25+ regulatory T cells was also strongly correlated with cancer myometrium invasion (r = 0.456, P = .001). Similar trends for CD4+ CD25+ regulatory T cells could be shown in TILs, with positive correlations with the clinical tumor grade, stage, and cancer myometrium invasion (Fig. 6). The positive correlations among CD4+ CD25+ regulatory T cells and tumor grade (r = 0.371, P < .017), stage (r = 0.358, P < .022), and myometrium invasion (r = 0.484, P = .001) were strongly significant.
The prognostic significance of TILs has been a longstanding topic of debate. However, recent studies have revealed that a subset of CD4+ T cells, referred to as CD4+ CD25+ regulatory T cells, may accumulate in the tumor environment and suppress tumor-specific T cell responses, thereby hindering tumor rejection. Hence, predicting tumor behavior on the basis of an indiscriminate evaluation of tumor-infiltrating T cells may result in inconsistent prognostic accuracy.23 In the current study, we confirmed that the prevalence of CD25+ cells among CD4+ T cells was significantly higher in TILs than PBLs. We demonstrated that there is an increased percentage of CD4+ CD25+ regulatory T cells in EC patients, and linked this to T cell immune suppression. The presence of regulatory T cells in the tumor was also allied with a poor prognosis in patients with EC.
The prevalence of regulatory T cells was significantly correlated with clinicopathologic factors. Cell surface markers and FoxP3 expression analysis identified these cells phenotypically as regulatory T cells. Generally, these cells exist in markedly higher proportions within TILs, PBLs, and/or regional lymph node lymphocytes of cancer patients, and their frequencies are suggested to be strongly related to tumor progression and inversely correlated with the efficacy of treatment.24 Our study showed that the high prevalence of regulatory T cells in the cancer stroma as well as the peripheral blood of EC patients was closely correlated with clinically malignant features, including tumor grade, stage, lymph node metastasis, and myometrium invasion. The prevalence of regulatory T cells in TILs within the cancer milieu can be a positive indicator of tumor aggressiveness. It is plausible that an increase in regulatory T cells at the tumor site may endorse local tumor growth and invasion, whereas an increase in regulatory T cells in the peripheral circulation may be relevant to the progression of systemic metastasis. The correlation between tumor stage or grade and regulatory T cell levels appears to be stronger for PBL regulatory T cell than for TIL regulatory T cell levels (Fig. 6). It may be possible that the inhibition of immune response in the tumor microenvironment consists of several inhibitory mechanisms, including increasing regulatory T cells in TILs, which is revealed by our results. In contrast, the inhibitory mechanism to the systematic immune system relies on regulatory T cells more significantly than in the tumor microenvironment.
The mechanisms of regulatory T cell-mediated immunosuppression are poorly understood. Tumor-specific regulatory T cells require ligand-specific activation and cell-to-cell contact to exert their suppressive activity on tumor-specific effector cells (CD8+ cytotoxic T lymphocytes and CD4+ Th cells), which includes decreased cytotoxicity, proliferation, and Th1 cytokine secretion.24 In the present study of TILs, these regulatory T cells may have averted appropriate antitumor immune responses through certain cytolytic enzyme-dependent essential pathways. In our study, granzyme B was scarcely expressed in peripheral regulatory T cells, but was highly expressed in 5% to 30% of CD4+ CD25+ regulatory T cells in the tumor microenvironment. A similar finding was noted for perforin. Regulatory T cells derived from the tumor environment could induce natural killer (NK) and CD8+ T cell death in a granzyme B- and perforin-dependent fashion.25 Granzyme B and perforin are therefore relevant for regulatory T cell-mediated suppression of tumor clearance in vivo. Grossman et al demonstrated that activated human regulatory T cells can use the perforin-granzyme pathway to kill a variety of autologous immune cells in vitro.26 Granzyme B and perforin are both important for the ability of NK cells and CD8+ T cells to kill their targets. Regulatory T cells may use the perforin-granzyme pathway as a mechanism to suppress the function of immune cells through suppression of NK and CD8+ T cells that are responsible for clearing these tumors.25 Moreover, Zhao et al reported that activated murine regulatory T cells can suppress B cell proliferation also in a granzyme B- and perforin-dependent fashion.27 Gondek et al reported that activated murine regulatory T cells suppressed CD4+ CD25+ T effector cells via a granzyme B-dependent mechanism.28 Conspicuously, all these studies used in vitro models with activated regulatory T cells.
It is not yet clear whether activated regulatory T cells in vivo express granzymes, and also, whether these molecules are important for regulatory T cell-mediated suppression of antitumor immune responses. In the present study, we directly used flow cytometry to evaluate the granzyme B and perforin expression in CD4+ CD25+ regulatory T cells isolated from tumor tissues. We found that regulatory T cells in the tumor microenvironment can express significantly high levels of both granzyme B and perforin. On the contrary, granzyme B and perforin were rarely expressed by CD8+ T cell-derived TILs. Mempel et al29 reported that the regulatory T cell-induced failure of cytotoxic T lymphocytes (CTLs) to kill target cells in vivo was correlated with impaired release of lytic granules. These results suggested that regulatory T cells might suppress antitumor immune responses by abrogating the expressions of granzyme B and perforin in CD8+ T cells. Ex vivo evaluation of in vivo activated NK and T cells also showed that tumor-activated regulatory T cells can cause the death of NK and CD8+ T cells in a granzyme B- and/or perforin-dependent pattern.25 The studies described here suggest that regulatory T cells can suppress the ability of CD8+ T cells to clear tumors. Our data novelly showed that not only is the granzyme-perforin pathway important for the function of NK and CD8+ T cells, but it may reciprocally be used by regulatory T cells to restrain the activity of these cells.
Manipulations of these regulatory T cells are proposed as possible new strategies for the treatment of cancer by facilitating the loss of tolerance to self-antigens.30-34 The CD28 costimulatory pathway plays an important role in antitumor responses. Immunodeficient mice have been cured of established human and rodent tumors when treated with anti-CD28 antibodies.35 To determine the functional capacity of patient T cells and their potential utility for immunotherapy, we measured intracellular cytokine production after stimulation with anti-CD3 and anti-CD28 antibodies. We found that TILs can express significant levels of IL-2, IFN-γ, and TNF-α after in vitro activation with anti-CD3 and anti-CD28 antibodies. Essentially, all of these cytokines are necessary for the potentiation of the cytotoxic activity of T cells and NK cells. Both IFN-γ and TNF-α possess direct cytotoxic and cytostatic activity toward tumor cells. Increased production of these cytokines may be helpful in inducing cytolytic cell function, recruiting and expanding antigen-presenting cells, and mediating direct antitumor effects. Collectively, the present results suggest that the dysfunction of T cells may be a reversible phenomenon, dependent on the tumor-bearing environment of the patient. Ample evidence demonstrates that there are multiple immunosuppressive mechanisms that considerably dampen antitumor responses and weaken the activity of current immunotherapeutic regimens.36 Therefore, it is necessary to reverse tumor-mediated immunosuppression before immunotherapies can successfully be applied.
We found that most tumor-infiltrating CD8+ lymphocytes were CD28− CD45RA− CD45RO+ CCR7−, suggesting good terminal differentiation. In addition, most of them were CD69+ CD103+ CD152+, suggesting an activated role. Regulatory T cells can reversibly blunt T cell responses by selectively modulating the terminal effector function of primed CD8+ T cells. Regulatory T cells are proposed to reversibly suppress cytotoxic T cell function independent of effector differentiation.29 Regulated CTLs exhibited no defect in proliferation, induction of cytotoxic effector molecules and secretory granules, in situ motility, or ability to form antigen-dependent conjugates with target cells. Only granule exocytosis by CTLs was markedly impaired in the presence of regulatory T cells. Regulatory T cells can also selectively interfere with the release of cytolytic granules by CTLs in a reversible and TGF-β–dependent manner.29 By this means, regulatory T cells can attenuate CTL cytotoxicity without affecting priming or differentiation.
In summary, TILs derived from EC patients contain remarkable proportions of CD4+ CD25+ T cells. Our data suggest that regulatory T cells play an essential role, at least in part, in controlling the anticancer immune response to EC. Host immune surveillance is prominent in the early stages but decreases during tumor progression, inversely correlating with the increased prevalence of regulatory T cells. The present study signifies that a high prevalence of regulatory T cells can be an indicator of poor clinical prognosis. Secretion of granzyme B and perforin is largely constricted in TILs with this immunophenotype, which may contribute to immune dysfunction. Furthermore, despite the activated status of CD8+ cytotoxic T cells, minimal expressions of granzyme B and perforin still render them without cytotoxicity. Our study also shows that adequate in vitro activation of TILs is able to restore the appropriate antitumor immune response for subsequent immunotherapy. Consequently, CD4+ CD25+ Foxp3+ regulatory T cells are not only involved in functional down-regulation of T, NK, and NK-T cells, but also seriously interfere with the antigen presentation function of various cells, including dendritic cells.37 Therefore, enhanced recruitment of antitumor effector T lymphocytes to tumor tissue, in addition to inhibition of local regulatory T cells, may be an ideal target for improving cancer immunotherapy.30-34, 38-40 Strategies aimed at depletion/functional inhibition of these cells by molecular targeting must maintain a critical balance between tumor immunity and self-tolerance.34, 41-44 These immunomodulation pathways may therefore have potential applications in forthcoming treatments of cancer.
In the present study, we directly measured and analyzed the ex vivo correlations between phenotypically defined regulatory T and CD8 cellular populations derived from human cancer milieu. We have further extended our study to explore the functional cytokines and cytotoxic enzymes of regulatory T and CD8+ T cells.
This work was supported by grants from the National Science Council (NSC 95-2314-B-002-278-MY3, NSC 95-2314-B-002-262-MY3) and research grants from the National Taiwan University Hospital–Taipei Veterans General Hospital (NTUH-TVGH) Joint Research Program (VN97-12, 96VN-008).