Tumor Regulatory T Cells Potently Abrogate Antitumor Immunity¹

The CD4⁺CD25⁺ regulatory T cells (Treg)³ constitute 5–10% of peripheral CD4⁺ T cells in naive mice (1). Tumors may differentiate, expand, recruit, and activate Treg (tumor Treg) via multiple mechanisms (2, 3, 4, 5). The frequency of tumor Treg is commonly elevated in tumors, peripheral blood, or lymphoid organs of many tumor-bearing hosts (2). When compared with naive Treg, suppressor functions of tumor Treg have been reported to be ineffective, comparable, or enhanced in various tumor models (5, 6, 7, 8, 9, 10, 11). Tumor Treg may promote local tumor growth at tumor sites, and may be relevant to the progression of systemic disease in the peripheral blood or lymphoid organs (12). Thus, tumor Treg may be a major obstacle for immunotherapy of cancer (2, 4, 5). It has become clear that understanding the mechanisms of action by Treg in tumor immunity is critical in developing an effective tumor vaccine or immunotherapy (2, 5).

Treg suppress multiple cells such as CD4⁺ and CD8⁺ T cells, B cells, NK cells, NKT cells, dendritic cells (DC) and macrophages (13, 14, 15, 16, 17). Cell targets and stages of the immune response that are critical for Treg-mediated suppression are still unclear (13, 14, 15, 16, 17). Several biological events of Treg-mediated suppression including inhibition of proliferation, cytokine production, differentiation, and migration by effectors have been reported (13, 14, 15, 16, 17). Also, down-regulation of expression of costimulatory molecules on DC, disruption of stable naive T-DC contact, induction of tolerogenic APCs, reduction of DC ability to present Ag and stimulate T cell proliferation, and control of cytotoxicity of tumor-infiltrating DC have been suggested (13, 14, 15, 16, 17, 18). Several mechanisms of Treg-mediated suppression have been proposed: 1) a yet unknown direct cell-cell contact; 2) killing B, NK, or CD8⁺ T cells via molecules such as perforin or granzyme B; 3) metabolic disruption via molecules such as CD39 or CD73; 4) modulating DC via molecules such as CTLA-4, LFA-1, LAG-3 or neuropilin-1; 5) indirect suppression via cytokines such as IL-10, TGF-β, or IL-35; and 6) in vivo consumption of survival and growth-promoting cytokines (13, 14, 15, 16, 17, 19, 20, 21, 22, 23). Recent reports show that Treg suppress mast cells via cell-cell contact involving OX40-OX40L signaling in mice and kill autologous CD8⁺ T cells by Fas-mediated apoptosis in humans (24, 25).

Multiple mechanisms underlying Treg-mediated immune suppression may be not mutually exclusive, and may vary depending on the nature of the immune response being regulated in various diseases and/or models (13, 14, 15, 16, 17). One complication, in particular, is how tumor Treg, which are usually elevated and likely activated during tumor progression, execute suppressor function (2, 3, 4, 5).

Adoptive tumor-primed CD4⁺ T cell transfer induced an effective host CD8⁺ T cell-dependent tumor protection in an aggressive spontaneous metastatic murine breast tumor model (26). Using this tumor model, we tested the suppressor function of both tumor and naive Treg in vitro and in vivo.

BALB/c mice were purchased from Taconic Farms and housed in specific pathogen-free conditions in the University of Pittsburgh animal facility. All animal procedures were performed according to approved protocols and in accordance with recommendations for the proper use and care of laboratory animals. Murine breast tumor cell 4T1.2-Neu (27) was maintained in DMEM (Irvine Scientific) supplemented with 10% FBS (HyClone), 2 mM glutamine (Invitrogen), antibiotic antimycotic solution (Sigma-Aldrich), and G-418 (500 μg/ml) (Invitrogen). Murine colon carcinoma cell CT26 (American Type Culture Collection) was cultured in RPMI 1640 (Irvine Scientific) supplemented with 10% FBS, 2 mM glutamine, and antibiotic antimycotic solution.

Six to 8-wk-old female BALB/c mice were inoculated s.c. with 4T1.2-Neu (1 × 10⁵) in 20 μl of endotoxin-free 1×PBS (Sigma) at the fourth mammary fat pad. Three to 4 wk post tumor inoculation, CD4⁺CD25⁺ T cells were purified from splenocytes of tumor-bearing mice or age-matched naive BALB/c mice using mouse CD4⁺CD25⁺ regulatory T cell isolation kit (Miltenyi Biotec) (Purity was confirmed by flow cytometry and consistently resulted in > 95%.). For surface staining, splenocytes or purified CD4⁺CD25⁺ T cells were stained by anti-CD4-PETXRED (or -FITC) (GK1.5) and anti-CD25-PE (PC61 or 7D4) or anti-CTLA-4-biotin mAb (UC10-4B9) (streptavidin-FITC as a second Ab) (isotype control of each Ab was used in control staining) (BD Biosciences or eBioscience), and analyzed by flow cytometry. For intracellular staining of CTLA-4, purified tumor CD4⁺CD25⁺ or naive CD4⁺CD25⁺ T cells were cultured in RPMI 1640, 10% FBS in 12-well plate at 37°C, 5% CO₂ for 4 d. Monensin (BD Biosciences) (4 μl/ml) was added for additional 4 h of culture. Cells were washed and resuspended in Cytoperm/Cytofix solution (BD Biosciences) for 20 min on ice, and subsequently stained using anti-CTLA-4-biotin mAb as described above. The expression of Foxp3 protein in purified tumor CD4⁺CD25⁺, tumor CD4⁺CD25⁻, or naive CD4⁺CD25⁺ T cells was detected using the specific Foxp3 (FJK-16s) intracellular staining kit (eBioscience).

Preparation of tumor-primed CD4⁺ T cells was described previously (26). In brief, BALB/c mice were injected i.p. with 600 μg of anti-CD25 mAb (purified from hybridoma cell PC61 culture supernatants using mAb purification kit). Three days later, these mice were inoculated s.c. with 4T1.2-Neu (1 × 10⁵) as described above. Tumor-primed CD4⁺ T cells were purified from splenocytes of anti-CD25 mAb-pretreated mice that rejected the tumor using anti-mouse CD4 microbeads (Miltenyi Biotec). Purified tumor-primed CD4⁺ T cells (2 × 10⁶/ml) were cultured in the presence or absence of tumor Treg, tumor CD4⁺CD25⁻ T cells or naive Treg (2–8 × 10⁶/ml) in RPMI 1640/10% FBS at 37°C, 5% CO₂ for 2 days. The concentration of IL-2 in the culture supernatants was determined by ELISA (BD Biosciences).

Activation of naive CD8⁺ T cells in vitro was described previously (26). In brief, tumor-primed CD4⁺ T cells were obtained as described above. CD11c⁺ DC or CD8⁺ T cells were purified from splenocytes of naive BALB/c mice using anti-mouse CD11c or CD8 microbeads (Miltenyi Biotec). Purified CD11c⁺ DC were loaded with 4T1.2-Neu tumor Ag. Tumor-primed CD4⁺ T cells (2 × 10⁵), tumor Ag-loaded DC (2 × 10⁵), and naive CD8⁺ T cells (2 × 10⁵) were cultured in the presence or absence of purified tumor CD4⁺CD25⁻ T cells, tumor Treg, or naive Treg (2–8 × 10⁵) in 200 μl of RPMI 1640, 10% FBS at 37°C, 5% CO₂ for 2 days. The concentration of IFN-γ in the culture supernatants was determined by ELISA (BD Biosciences).

Naive splenic CD11c⁺ DC and tumor-primed CD4⁺ T cells were prepared as described above. CD11c⁺ DC (2 × 10⁵) and tumor-primed CD4⁺ T cells (2 × 10⁵) were cultured alone, with tumor Treg or naive Treg (2–8 × 10⁵) in the presence of LPS (1 μg/ml; Sigma-Aldrich) in 200 μl RPMI 1640 10% FBS at 37°C, 5% CO₂ for 18 or 48 h. The concentration of IL-12 (p40) in the culture supernatant was measured by ELISA (BD Biosciences). These cells were harvested and treated with 5 mM EDTA. Fc receptor binding of Abs was minimized by incubation with rat anti-mouse CD16/CD32 mAb (BD Biosciences) before staining with anti-CD11c-allophycocyanin (HL3) and anti-CD80-PE (16-10A1) or anti-CD86-PE (GL1) (isotype control of each Ab was used in control staining) (BD Biosciences or eBioscience), and analyzed by flow cytometry. Propidium iodide (BD Biosciences) was used to check cell viability. Forward and side scatter were used to exclude cell debris. The flow cytometric data were analyzed using FlowJo software (Tree Star).

BALB/c mice were adoptively transferred i.v. with tumor-primed CD4⁺ T cells (1 × 10⁷). One day later, these mice were inoculated s.c. with 4T1.2-Neu (1 × 10⁵). Primary tumor was observed in all tumor-inoculated mice. Tumor rejection usually started on day 9 and completed on day 21 post tumor inoculation. To examine Treg-mediated suppression of effectors in vitro, day 21 post tumor inoculation, mice that rejected the tumor were sacrificed and splenic cells (1 × 10⁵) were restimulated in vitro with mitomycin C-treated 4T1.2-Neu (1 × 10⁴) in the presence of tumor Treg or naive Treg (1 × 10⁵) in 200 μl of RPMI 1640/10% FBS at 37°C, 5% CO₂ for 2–3 days. To examine Treg-mediated suppression of effectors in vivo, day 21 post tumor inoculation, effectors in mice that rejected the tumor were restimulated in vivo by immunization s.c. with mitomycin C-treated 4T1.2-Neu (2 × 10⁶) and these mice were adoptively transferred i.v. with tumor Treg or naive Treg (1 × 10⁷) at the same time. Six days later, these mice were sacrificed and splenic cells (4 × 10⁵) were cultured without in vitro stimulation in 200 μl of RPMI 1640 10% FBS at 37°C, 5%CO₂ for 3 days. The concentration of IFN-γ in the culture supernatant was measured by ELISA.

Tumor-primed CD4⁺ T cells (1 × 10⁷) and tumor Treg or naive Treg (1 × 10⁷) were adoptively cotransferred i.v. into naive BALB/c mice on day −1. These mice were inoculated s.c. with 4T1.2-Neu (1 × 10⁵) on day 0. On day 5, CD8⁺ T cells were purified from pooled tumor-draining lymph nodes (TDLN) using anti-mouse CD8 microbeads. CD11c⁺ DC were purified from splenocytes of naive BALB/c mice as descried above, and loaded with 4T1.2-Neu or CT26 tumor Ag (26). Tumor Ag-loaded DC (4 × 10⁴) and purified CD8⁺ T cells (1 × 10⁵) were cultured in 200 μl of RPMI 1640 10% FBS at 37°C, 5% CO₂ for 5 days. The concentration of IFN-γ in the culture supernatants was determined by ELISA.

To examine the influence of Treg on antitumor immunity at the early stage of the immune response induced by adoptively transferred tumor-primed CD4⁺ T cells, tumor Treg or naive Treg (1 × 10⁷) and tumor-primed CD4⁺ T cells (1 × 10⁷) were adoptively cotransferred i.v. into naive BALB/c mice on day −1. These mice were inoculated s.c. with 4T1.2-Neu (1 × 10⁵) on day 0. To examine the influence of Treg on antitumor immunity at the effector phase of the immune response induced by adoptively transferred tumor-primed CD4⁺ T cells, tumor-primed CD4⁺ T cells (1 × 10⁷) were adoptively transferred i.v. into naive BALB/c mice on day −1. These mice were inoculated s.c. with 4T1.2-Neu (1 × 10⁵) on day 0. Tumor-rejection usually started on day 9 post tumor inoculation, adoptive tumor Treg or naive Treg (1 × 10⁷) transfer thus was performed on day 9. Primary tumor was observed in all tumor-inoculated mice and measured by electric caliper in two perpendicular diameters every other day. Mice were sacrificed for humane reasons when a primary tumor reached 10 mm in mean diameter, when ulceration, bleeding, or both developed, or when mice became ill due to metastatic diseases.

We analyzed significance between groups with Student’s t test. For evaluation of data from animal survival studies, we used the log rank test. All analyses were performed using Prism software (GraphPad Software). A p value <0.05 is considered to be statistically significant.

The increased frequency of CD4⁺CD25⁺ T cells in TDLN or tumors has been reported in a 4T1 tumor model (11, 28). This population was also elevated in splenocytes of 4T1.2-Neu tumor (a 4T1 related model)-bearing mice when compared with naive mice (Fig. 1, A and B). Tumor progression induced splenomegaly with a marked increase in absolute numbers of CD4⁺CD25⁺ T cells after 2-wk tumor inoculation (data not shown). When compared with CD4⁺ T cells from splenocytes of naive mice, CD4⁺ T cells from splenocytes of tumor-bearing mice expressed higher levels of CD25 and CTLA-4 (Fig. 1, A, C, and D). In line with naive Treg (data not shown), tumor Treg but not tumor CD4⁺CD25⁻ T cells expressed Foxp3 (Fig. 1,E). Importantly, tumor Treg but not tumor CD4⁺CD25⁻ T cells suppressed the production of IL-2 by tumor-primed CD4⁺ T cells and naive CD8⁺ T cell activation in vitro (Fig. 1, F and G). These data show that, in the breast tumor model, CD4⁺CD25⁺ T cells from splenocytes of tumor-bearing mice are actually functional Treg.

Tumor-primed CD4⁺ T cells spontaneously produced IL-2 in culture (26). To test whether there was any special property of tumor Treg, tumor-primed CD4⁺ T cells were cultured alone, with tumor Treg or naive Treg at ratios of 1:1, 1:2, or 1:4 (CD4:Treg). The production of IL-2 by tumor-primed CD4⁺ T cells was monitored. As shown in Fig. 2 A, both tumor Treg and naive Treg inhibited the production of IL-2 by tumor-primed CD4⁺ T cells in a dose-dependent manner. These data suggest that tumor Treg ability to inhibit tumor-primed CD4⁺ T cell activity is comparable to naive Treg in vitro.

Naive CD8⁺ T cells were activated by tumor Ag-loaded DC in the presence of tumor-primed CD4⁺ T cells in vitro (26). To test the special property of tumor Treg, we examined tumor Treg or naive Treg ability to suppress naive CD8⁺ T cell activation in vitro. Tumor-primed CD4⁺ T cells, tumor Ag-loaded DC, and naive CD8⁺ T cells were cultured alone, with tumor Treg, or naive Treg at ratios of 1:1:1:1, 1:1:1:2, or 1:1:1:4 (CD4:DC:CD8:Treg). IFN-γ in the culture supernatants was monitored. As shown in Fig. 2 B, tumor Treg inhibited naive CD8⁺ T cell activation in vitro in a dose-dependent manner. However, naive Treg did not do so. These data show that tumor Treg ability to suppress naive CD8⁺ T cell activation is distinguishable from naive Treg in vitro.

It has been shown that both naive Treg and tumor Treg inhibited DC function in vitro and in vivo (22, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39). To examine the influence of tumor Treg or naive Treg on DC, we compared the expression of CD80 and CD86 and the production of IL-12 by splenic DC in vitro. Splenic DC and tumor-primed CD4⁺ T cells were cultured alone, with tumor Treg, or naive Treg in the presence of LPS at ratios of 1:1:1, 1:1:2, or 1:1:4 (DC:CD4:Treg). When compared with naive Treg, tumor Treg down-regulated CD80 and CD86 expression, and inhibited the production of IL-12 by DC in a dose-independent manner in vitro (Fig. 3, A and B, data not shown). These data indicate that tumor Treg potently suppress DC function in vitro.

Adoptive tumor-primed CD4⁺ T cell transfer ignited potent protective antitumor immunity (26). To determine whether tumor Treg or naive Treg inhibit antitumor immunity at the effector phase of the immune response, mice were adoptively transferred with tumor-primed CD4⁺ T cells (1 × 10⁷) on day −1, and inoculated with tumor cells on day 0. Tumor Treg or naive Treg (1 × 10⁷) were adoptively transferred on day 9. As shown in Fig. 4,A, neither tumor Treg nor naive Treg suppressed antitumor immunity at the effector phase of the immune response. To understand the mechanism behind this observation, we examined tumor Treg or naive Treg ability to inhibit effectors generated by adoptively transferred tumor-primed CD4⁺ T cells. Effectors from tumor-rejection mice were cultured with tumor Treg or naive Treg in vitro. As shown in Fig. 4,B, neither tumor Treg nor naive Treg suppressed effectors in vitro. To confirm this in vivo, tumor Treg or naive Treg were adoptively transferred into tumor-rejection mice. As shown in Fig. 4 C, neither tumor Treg nor naive Treg suppressed effectors in vivo. These data suggest that, in this model, effectors are resistant to suppression by tumor Treg or naive Treg. This may result in the failure of Treg-mediated suppression of antitumor immunity at the effector phase of the immune response.

Next, we explored the influence of tumor Treg or naive Treg on the induction of antitumor immunity. We first tested whether tumor Treg abrogate tumor-specific CD8⁺ T cell responses induced by adoptively transferred tumor-primed CD4⁺ T cells in TDLN. Mice were adoptively cotransferred with tumor-primed CD4⁺ T cells (1 × 10⁷) and tumor Treg or naive Treg (1 × 10⁷) on day −1, and inoculated with tumor cells on day 0. On day 5, CD8⁺ T cells were purified from pooled TDLN. The in vivo tumor-specific CD8⁺ T cell responses were evaluated ex vivo by stimulation of purified CD8⁺ T cells with tumor Ag-loaded DC. As shown in Fig. 5,A, CD8⁺ T cells purified from TDLN of mice that were adoptively transferred by tumor-primed CD4⁺ T cells alone or cotransferred by tumor-primed CD4⁺ T cells and naive Treg responded to a tumor-specific restimulation. However, CD8⁺ T cells purified from TDLN of mice that were adoptively cotransferred by tumor-primed CD4⁺ T cells and tumor Treg did not do so (Fig. 5,A). We next examined whether tumor Treg suppress antitumor immunity at the early stage of the immune response. Mice were adoptively cotransferred with tumor-primed CD4⁺ T cells (1 × 10⁷) and tumor Treg or naive Treg (1 × 10⁷) on day −1, and inoculated with tumor cells on day 0. As shown in Fig. 5 B, only tumor Treg inhibited antitumor immunity at the early stage of the immune response. These data indicate that, in this model, tumor Treg potently abrogate tumor-specific CD8⁺ T cell responses in TDLN and antitumor immunity at the early stage of the immune response induced by adoptively transferred tumor-primed CD4⁺ T cells.

Both Treg and effectors could be primed during tumor progression (40, 41). Thus, Foxp3 expression together with suppressor function (a cardinal feature of Treg) may be the best way to judge whether tumor CD4⁺CD25⁺ T cells are actually Treg (14). In this report, CD4⁺CD25⁺ but not CD4⁺CD25⁻ T cells from splenocytes of tumor-bearing mice expressed Foxp3, inhibited tumor-primed CD4⁺ T cell activity and CD8⁺ T cell activation (Fig. 1), suggesting they are functional Treg.

During tumor progression, tumors may trigger Treg activation in vivo (3, 5). The function of Treg from tumor-bearing mice was different from naive Treg activated by anti-CD3/IL-2 in vitro (42). Suppressive activity of Treg from cancer patients was enhanced when compared with Treg from healthy donors (9, 10). In a related 4T1 model, suppressive activity of tumor-infiltrating Treg was increased (11). It is therefore interesting to know whether tumor Treg are functionally distinguishable from naive Treg (5). Treg are usually expanded and activated in vitro, and adoptively transferred into mice to examine their function in vivo. In this breast tumor model, sufficient Treg were obtained from splenocytes (but not lymph nodes or tumors) of tumor-bearing mice at a later stage of tumor progression (3–4 wk post tumor inoculation), which made it possible to evaluate suppressor function of tumor Treg expanded and activated under a physiological (tumor progression) condition.

With regard to the inhibition of the production of IL-2 by tumor-primed CD4⁺ T cells in vitro, tumor Treg was comparable to naive Treg (Fig. 2,A). However, only tumor Treg suppressed naive CD8⁺ T cell activation in vitro (Fig. 2,B). Naive Treg must be activated to exhibit suppressor function in vitro (43). In this model, naive Treg were cultured with tumor-primed CD4⁺ T cells or tumor-primed CD4⁺ T cells and tumor Ag-loaded DC, and exhibited suppressor function (Fig. 2,A), indicating that naive Treg were activated during ex vivo culture because tumor-primed CD4⁺ T cells spontaneously produced IL-2 (Figs. 1,D and 2 A; Ref. 26), which is required to activate Treg (43, 44).

DC have been suggested to be the most relevant targets of Treg in vivo (33, 34). Consistently with others (35), we found that tumor Treg hindered DC ability to stimulate T cell proliferation in vitro and DC incubated with tumor Treg in vitro induced less robust Ag-specific T cell responses in vivo after vaccination (data not shown). A recent report indicates that naive Treg in vitro down-regulate CD80 or CD86 expression on splenic DC at a ratio of 1:10 (DC:Treg) (39). In this study, tumor Treg in vitro down-regulated CD80 or CD86 expression and reduced the production of IL-12 by splenic DC at a ratio of 1:1 (DC:Treg) (Fig. 3), indicating a potent suppressor function of tumor Treg. These data suggest that tumor Treg might be important in disarming DC function, although whether their effects on DC account for immunopathology is still an open question to be determined.

It has been speculated that the mechanisms of Treg-mediated suppression observed in vitro may be different from the mechanisms operative in vivo (23, 45, 46). Therefore, it is important to examine the function of tumor Treg and naive Treg in vivo. Treg depletion in tumors resulted in an advanced tumor regression and adoptive transfer of purified Treg inhibited adoptively transferred tumor-specific CD8⁺ T cell- or NK cell-mediated antitumor immunity in vivo (19, 47, 48, 49). These data suggest that tumor Treg or naive Treg suppress antitumor effectors. In this model, neither tumor Treg nor naive Treg impaired antitumor immunity at the effector phase of the immune response induced by adoptively transferred tumor-primed CD4⁺ T cells (Fig. 4,A). This correlated with the observation that effectors from tumor-rejection mice were resistant to suppression by tumor Treg or naive Treg in vitro or in vivo (Fig. 4, B and C). Thus, strategies aimed at generating effectors that are resistant to Treg suppression may be promising for an effective tumor vaccine or immunotherapy.

Treg have been indicated to mediate suppression of T cell immunity at the early stage of the immune response (2, 50). Tumor Treg suppressed naive CD8⁺ T cell activation in vitro (Fig. 2,B). Furthermore, tumor Treg in vivo suppressed antitumor immunity at the early stage of the immune response induced by adoptively transferred tumor-primed CD4⁺ T cells (Fig. 5,B). Tumor-primed CD4⁺ T cells, tumor Treg, and naive Treg were detected in TDLN after adoptive transfer (data not shown). Importantly, tumor Treg (but not naive Treg) potently abrogated tumor-specific CD8⁺ T cell responses in TDLN (Fig. 5 A). This observation is in line with a report that activated CD8⁺ T cells were not detectable in TDLN in 4T1 tumor-bearing mice (11). These data indicate that in vivo tumor Treg suppress tumor-specific CD8⁺ T cell responses at the early stage of the immune response induced by adoptively transferred tumor-primed CD4⁺ T cells. This can be interpreted that 1) tumor Treg abrogate naive CD8⁺ T cell priming, 2) CD8⁺ T cells primed in the presence of tumor Treg are anergic, or 3) tumor Treg suppress function and/or differentiation of primed CD8⁺ T cells. The precise mechanisms of action by tumor Treg in TDLN will be investigated in the future studies. It is now recognized that Treg function may be compartmentalized (15, 51, 52). How the function of tumor Treg from other compartments compares to that of the spleen needs further investigation.

In summary, in this study, in vitro data show the comparable ability between tumor Treg and naive Treg to inhibit tumor-primed CD4⁺ T cell activity, and the potent ability of tumor Treg to suppress CD8⁺ T cell activation and hinder DC function. In vivo results demonstrate that tumor Treg are potent to abrogate tumor-specific CD8⁺ T cell responses in TDLN and suppress antitumor immunity at the early stage of the immune response induced by adoptively transferred tumor-primed CD4⁺ T cells. Collectively, these data suggest that tumor Treg effectively abrogate antitumor immunity.

We are indebted to H. Shen, D. Falkner, J. Chen, H. Noh, and C. Donahue (University of Pittsburgh) for their help, and the University of Pittsburgh Experimental Animal Facility technicians for animal care.

The authors have no financial conflict of interest.