CD4+CD25+ regulatory T cells suppress tumor immunity but are sensitive to cyclophosphamide which allows immunotherapy of established tumors to be curative

Authors


Abstract

We investigated the mechanisms of immune tolerance raised by tumors by comparing immunogenic and tolerogenic tumor cell clones isolated from a rat colon carcinoma. When injected into syngeneichosts, the immunogenic REGb cells yield tumors that are rejected, while the tolerogenic PROb cells yield progressive tumors and inhibit the regression of REGb tumors. We show here that PROb tumor volume is correlated with an expansion of CD4+CD25+ regulatory T lymphocytes in lymphoid tissues. These cells delay in vivo the rejection of REGb tumors and inhibit in vitro T cell-mediated immune responses against REGb cells through a mechanism that requires cell contact between effector and regulatory T cells and involves TGF-β. While total T cells fromPROb tumor-bearing rats yield no apparent anti-tumor immune response, depletion of CD25+ T cells restores this reactivity. A single administration of cyclophosphamide depletes CD4+CD25+ T cells in PROb tumor-bearing animals, delays the growth of PROb tumors, and cures rats bearing established PROb tumors when followed by an immunotherapy which has no curative effect when administered alone. These results demonstrate the role of CD4+CD25+ regulatory T cells in tumor-induced immune tolerance and the interest of regulatory T cell depletion to sensitize established tumors to immunotherapy.

Abbreviation:
CTX:

Cyclophosphamide

1 Introduction

Progressive tumors can escape immune recognition and destruction by actively establishing an immune tolerance involving immunosuppressive T lymphocytes 13.Several subsets of immunosuppressive CD4+ and CD8+ T cells have been identified 4. Recently, naturally occurring cells exhibiting immunosuppressive (regulatory)functions were also identified within the CD4+ T cell subset which expresses CD25, the α-chain of the IL-2 receptor, in both rodents 5, 6 and humans7. These regulatory CD4+CD25+ T cells contribute to the prevention of autoimmune disorders by controlling the activity of autoreactive T lymphocytes. They inhibit the proliferation of autoimmune CD25 T cells through still undetermined mechanisms 8. It has been suggested that CD4+CD25+ regulatory T cells could also contribute to the immune tolerance of cancer 9, 10, and proposed that the therapeutic depletion of these cells could result in an improved response to cancer immunotherapy 1114. However, none of these studies demonstrates a therapeutic benefit of CD4+CD25+ T cell depletion on fully established tumors.

In the present study, we investigated the occurrence and functions of CD4+CD25+ regulatory T cells in a rat tumor model. Two clones established from the same tumor cell line were used 15. When injected into syngeneic rats, the REGb cell line yields tumors that completely regress through a T cell-dependent immune response that leads to tumor cell killing by tumor-infiltrating MΦ 16. The PROb cell line yields progressive and lethal tumors, while maintaining tumor rejection antigen expression, since PROb cells are rejected bysyngeneic rats immunized against REGb cells 15. When injected into rats bearing an established PROb tumor, REGb cells are not rejected and yield growing and lethal tumors.

In the present work, we demonstrate that CD4+CD25+ T cells undergo a progressive, tumor volume-related expansion in lymphoid tissues of PROb tumor-bearing rats. CD4+CD25+ T cells from these animals are able to prevent the induction of T cell-mediated immune response elicited by immunogenic REGb cells. A single administration of the alkylating agent cyclophosphamide (CTX) efficiently depletes CD4+CD25+ regulatory T cells and permits to cure established PROb tumors with an otherwise poorly efficient immunotherapy associating tumor cells and BCG.

2 Results

2.1 CD4+CD25+ cell ratio in the spleen and tumor-draining lymph nodes increases with tumor progression

The prevalence of CD4+ splenic T cells that co-expressed CD25 was determined using double-staining FCM (Fig. 1a). This ratio, which was found to be 10.8±2.0% of CD4+ T cells in tumor-free rats, increased in correlation with PROb tumor volume (r=0.970, n=20, p<0.001), reaching about 40% in rats bearing 1 cm3 tumors (Fig. 1b). The prevalence of CD4 T cells co-expressing CD25 was low, with more than 90% of the CD25+ T cells co-expressing CD4. The mean actual number of CD4+CD25+ T cells in the spleen increased from 2.1×106 (range: 1.4×106–2.6×106) in tumor-free rats to 3.6×106 (2×106–4.8×106) and 5.8×106 (4.2×106–9.4×106) in rats bearing 21- and 42-day-old PROb tumors, respectively (n=4). Similar results were obtained in the axillary lymph nodes, in which the ratio of CD4+ T cells co-expressing CD25 increased from 9±2% in tumor-free rats to 21±3% and 38±4%, 21 and 42 days after tumor challenge, respectively. FCM analysis showed that CD4+CD25+ splenic T cells expressed TCR α/β and the RT6–2 alloantigen, whereas only low levels of CD45RC and no Thy-1 (CD90) could be detected on these cells (data not shown). Altogether, these results demonstrate that tumor progression is associated with CD4+CD25+ T cell expansion in the lymphoid organs.

Figure 1.

Increase in the ratio of CD4+ T cells co-expressing CD25 during PROb tumor progression. (a) FCM analysis of spleen cells isolated from tumor-free rats and rats bearing PROb tumors. Cells were double-labeled with anti-CD4 and anti-CD25 mAb. In the individual determinations represented here, the percentage of CD4+ cells that co-expressed CD25 was 9% (tumor-free rat), 22% (21-day-old tumor) and 40% (42-day-old tumor), respectively. These plots are representative of the results obtained with five rats in each group. (b). Correlation between the ratio of CD4+ T cells co-expressing CD25 in the spleen of 20 rats and the PROb tumor volume.

2.2 CD25+ splenic T cells from PROb tumor-bearing rats delay the rejection of the immunogenic REGb tumors

Injection of the REGb immunogenic tumor cell variant into tumor-free rats yields small tumors and triggers a T cell-dependent immune response, which induces complete rejection of these tumors 3 weeks after tumor cell injection 15. Co-injection of REGb cells with the purified CD25+ spleen cells from rats bearing established PROb tumors yielded tumors which reached a maximal volume tenfold larger than that of tumors obtained after injection of REGb cells alone (Fig. 2a). This effect was not related to a contamination of spleen cells with metastatic PROb cells since injection of spleen cells from rats bearing PROb tumors, in the absence of REGb cells, did not yield any tumor. Tumors induced by co-injection of REGb cells and CD25+ spleen cells from PROb tumor-bearing animals were ultimately rejected, but the rejection was significantly delayed. In contrast, the CD25-depleted fraction from PROb tumor-bearing rat spleen cells completely inhibited tumor appearance when co-injected with REGb cells (Fig. 2a).

CD25+, but not CD25 T cells, isolated from the spleen of PROb tumor-bearing rats inhibited IFN-γ production in mixed cultures of immune T cells from rats which rejected REGb tumors, dendritic cells and killed REGb cells (Fig. 2b). This inhibition was not observed when CD25+ T cells were separated from the other cells populations by a permeable membrane (Fig. 2c). In addition, this inhibition was prevented by a blocking anti-TGF-β mAb (Fig. 2d), whereas it was not influenced by an anti-rat IL-10 antibody (not shown). Taken together, these results indicate that CD25+ T cells contained in the spleen of PROb tumor-bearing rats negatively regulate the function of tumor immune lymphocytes, leading to tumor rejection through a process involving TGF-β and a direct or close contact between lymphocyte subpopulations.

Then, we analyzed whether an injection of the immunogenic REGb cell variant which generates always regressing tumors also induced an accumulation of CD4+CD25+ regulatory T cells in the spleen. We found that the ratio of the splenic CD4+ T cells that co-expressed CD25 increased from 9±2 to 15±2 and 21±3%, 14 and 21 days after REGb cell injection, respectively. However, in difference to the CD4+CD25+ cells generated in rats bearing PROb tumors, these cells were unable to block IFN-γ production by splenic T lymphocytes of rats immunized after several REGb cell injections. Indeed, CD4+CD25+ T cells from rats bearing a 21-day-old regressive REGb tumor secreted large amounts of IFN-γ when cultured with dendritic cells isolated from the spleen of a rat immunized after three injections of REGb cells (Fig. 3a) and induced MΦ-mediated PROb cell killing (Fig. 3b).

We also explored CD4+CD25+ splenic T cells in another model of regressive tumor obtained after injection of a PROb cell variant in which cytochrome c expression was down-regulated by stable transfection of an antisense construct. These cells, due to an imbalance in the respiratory chain, had a high susceptibility to die. When injected into syngeneic rats, this cell variant induced tumors which regressed in all the animals after a few weeks of progression (E. Schmitt et al., manuscript in revision). The ratio of CD4+ T cells co-expressing CD25 increased from 10±2% to 16±2% in the spleens of three rats bearing 14-day-old cytochrome c antisense PROb tumors. In difference to CD4+CD25+ splenic T cells from three PROb tumor-bearing rats, CD4+CD25+ T cells from three rats bearing a regressing cytochrome c antisense-transfected tumor produced large amounts of IFN-γ (Fig. 3c) and induced MΦ to kill the PROb cells (Fig. 3d). Thus, these CD4+CD25+ T cells, like that of rats bearing REGb tumors, exerted clearly an effector rather than a regulatory immune function. These observations indicate that CD4+ T cells with a CD25+ phenotype exhibit an immunosuppressive regulatory function when isolated from animals bearing a progressing tumor, whereas they function as activated effector cells when isolated from animals bearing regressing tumors.

Figure 2.

Effects of CD25+ T cells from PROb tumor-bearing rats on the immune response to REGb cells. (a) Tumor volume kinetics in rats receiving an s.c. injection of 1×106 REGb cells alone (open diamonds) or mixed with 5×106 T cells from the spleen of rats bearing 42-day-old PROb tumors, these T cells being either previously enriched with (solid triangles) or depleted of (open squares) CD25+ T cells (n=5 rats per group). (b) IFN-γ concentrations were measured in the supernatants of 3-day co-cultures of 1×104 mitomycin C-treated REGb cells with 104 dendritic cells and 2.5×105 T cells isolated from the spleen of rats immunized against REGb cells alone (control) or mixed with 2.5×105 CD25+ or CD25 T cells isolated from the spleen of rats bearing 28-day-old PROb tumors. One representative of four independent experiments is shown (mean ± SD of triplicates). (c) The same experiment was performed by using a Transwell system that separates CD25+ or CD25 T cells from the other cells by a porous membrane. Results shown are from one representative of two independent experiments (mean ± SD of triplicates). (d) The same experiment was performed in the absence or presence of a blocking mAb against TGF-β1, TGF-β2 and TGF-β3 (50 μg/ml). Shown are means ± SD of triplicates. *p<0.001; NS: not significant. Similar results were obtained with anti-TGF-β mAb at a concentration of 100 μg/ml.

Figure 3.

Effector functions of CD4+CD25+ splenic T cells isolated from rats bearing a regressive tumor. Regressive tumors were produced by injecting s.c. either REGb cells, or a PROb cell variant in which cytochrome c expression was down-regulated by transfection of an antisense construct (PRO cyto-c as). CD4+CD25+ T cells were isolated from the spleen 21 days after PROb or REGb tumor cell injection (a, b), or 14 days after PROb or PRO cyto-c as tumor cell injection (c, d), and studied for IFN-γ production (a, c) and the capacity of inducing M Φ -mediated cytotoxicity to PROb cells (b, d). Shown are means ± SD of triplicates; n=3 rats per group.

2.3 CD25+ T cell depletion restores the anti-tumor immune function of splenic T lymphocytes from rats bearing established PROb tumors

T lymphocytes isolated from the spleen of rats bearing PROb tumors produced low IFN-γ levels when cultured with PROb cells and splenic dendritic cells. Depletion of CD25+ cells from this T cell population resulted in a dramatic increase in IFN-γ secretion (Fig. 4a). We previously reported that T lymphocytes from tumor-immune rats had no direct cytotoxic effect on PROb cells, but strongly stimulated the tumoricidal activity of peritoneal and tumor-infiltrating MΦ through IFN-γ secretion 16. In contrast to unfractioned splenic T cells from PROb tumor-bearing rats, which did not activate MΦ-mediated cytotoxic activity, CD25-depleted T cells from the same animals induced MΦ to kill the PROb cells (Fig. 4b). This effect was mediated by IFN-γ since it was inhibited by addition of neutralizing antibodies to IFN-γ in the same co-culture system.

No cytotoxic effect was observed when total T lymphocytes or the CD25+ cell fraction were substituted for CD25 T cells (Fig. 4c). Stimulation of this MΦ-mediated cytotoxicity by mixtures of CD25+and CD25 T cells from tumor-bearing rats was directly dependent on the CD25+/CD25 ratio (Fig. 4d). Splenic T cells from naive rats were not able to produce IFN-γ and stimulate the cytotoxic activity of MΦ towards PROb cells even after CD25+ T cell depletion (Fig. 4b). Taken together, these results demonstrate that tumor-specific effector T cells, able to kill tumor cells through IFN-γ secretion and MΦ activation, are present in the spleen of tumor-bearing rats. These cells are functionally blocked by the abundant CD25+ regulatory T cell subpopulation raised by the tumor since CD25+ T cell depletion can restore anti-tumor activity of these effector cells.

Figure 4.

CD25+ T cell depletion restores an immune response in T cells isolated from the spleen of tumor-bearing rats. (a) IFN-γ levels in the supernatants of 3-day mixed cultures of 1×104 mitomycin C-treated PROb cells and 1×104 dendritic cells with 1×105 total (SLc) or CD25-depleted (CD25) T cells from the spleen of rats bearing 28-day-old PROb tumors (TB) or tumor-free rats (TF). (b) MΦ-mediated cytotoxicity was determined with a crystal violet assay after 48 h of co-culture of 1×104 PROb cells with 1×105 MΦ in the presence of 1×105 total (SLc) or CD25-depleted (CD25) T cells isolated from the spleen of rats bearing 42-day-old PROb tumors (TB) or tumor-free rats (TF). (c) Cytotoxicity of peritoneal MΦ on PROb cells after addition of supernatants from co-cultures of 1×105 total (SLc) or CD25-depleted (CD25) T cells isolated from the spleen of rats bearing 42-day-old PROb tumors (TB) or tumor-free rats (TF). A blocking anti-rat IFN-γ mAb was added to the supernatant of CD25-depleted T cells from tumor-bearing rats. (d) Cytotoxic effects of MΦ against PROb cells was determined with a crystal violet assay after 48 h of co-culture of 1×104 PROb cells with 1×105 MΦ and 5×105 T lymphocytes isolated from the spleen of tumor-bearing rats at different ratios of CD25+ to CD25 T cells. In (a–d) results from one of four independent experiments are shown (mean ± SD of triplicates).

2.4 Cyclophosphamide depletes regulatory CD25+ T lymphocytes in the spleens of PROb tumor-bearing rats

In an attempt to deplete CD4+CD25+ regulatory T cells in vivo and to restore the function of tumor-specific effector cells, we injected PROb tumor-bearing rats with various cytotoxic or immunomodulatory agents: CTX, methotrexate, vinblastin, 5-fluorouracil, doxorubicin, cisplatin, 6α-methylprednisolone and a murine anti-rat CD25 mAb. Only CTX, methotrexate and the anti-CD25 mAb induced a decrease in the CD4+CD25+/CD4+ splenic T cell ratio in the spleen resected 7 days after a single drug injection (Fig. 5a). This ratio dropped from 21±2% in untreated tumor-bearing rats to 6±3% (p<0.05), 11%± 3% (p<0.05) and 14%± 2% (p<0.05), in rats which received CTX, methotrexate and anti-CD25 mAb, respectively. Only splenic lymphocytes from CTX- and methotrexate-treated tumor-bearing rats induced MΦ-mediated PROb cell killing (Fig. 5b), demonstrating that these two drugs were able to restore an anti-tumor immune response in spleen cells from tumor-bearing animals.

We further investigated the kinetics of splenic CD4+ T cells at the days following CTX injection. The whole population of CD4+ T cells in the spleen began to decrease 1 day after the injection and then returned to its pretreatment level after 5 days. The CD4+CD25+ subpopulation decreased more slowly, reaching its lowest level 7 days after injection. This subpopulation returned to its pretreatment level after 28 days. Thus, the CD4+CD25+/CD4+ ratio increased during the first 2 days after CTX injection, then decreased to reach its nadir on day 7 after CTX injection. Afterwards, the ratio increased and returned to its pretreatment level after 28 days (Fig. 5c). The capacity of splenic lymphocytes from PROb tumor-bearing rats to induce MΦ-mediated cytotoxicity 7 days after CTX treatment was negatively correlated with the ratio of CD4+CD25+/CD4+ T cells (Fig. 5c, d). This capacity was reduced by the addition of CD25+, but not CD25 splenic T cells from untreated PROb tumor-bearing rats (Fig. 5e). Splenic T cells from CTX-treated tumor-free, instead of PROb tumor-bearing rats, did not activate the cytotoxic effect of MΦ against PROb cells (data not shown). These results indicate that a single injection of CTX strongly depletes CD25+ T cells in rats bearing PROb tumors and restores the anti-tumor activity of effector T cells obtained from these animals.

Figure 5.

 CTX-induced depletion of CD4+CD25+ regulatory T cells from PROb tumor-bearing rats and restoration of the T cell-dependent MΦ cytotoxicity. (a) Groups of four rats bearing 21-day-old PROb tumors received one injection of either CTX (30 mg/kg i.p.), methotrexate (MTX; 2 mg/kg i.v.), vinblastin (VLB; 0.1 mg/kg i.v.), adriamycin (DXR; 1.5 mg/kg i.v.), 5-fluorouracil (5FU; 20 mg/kg i.v.), cisplatin (CDDP; 3 mg/kg i.v.), methylprenisolone (MP; 15 mg/kg i.v.), or the anti-CD25 mAb (1 mg i.p.). FCM analysis of the ratio of CD4+ cells co-expressing CD25 to total CD4+ T cells in the spleen was performed 7 days after the treatment. (b) MΦ-mediated cytotoxicity was evaluated by a crystal violet assay 48 h after co-culture of 1×104 PROb cells with 1×105 MΦ and 5×105 splenic lymphocytes from groups of four rats bearing 21-day-old-PROb tumor collected 7 days after indicated treatments. (c) Evolution of the rate of splenic CD4+ T cells (filled squares), CD4+CD25+ T cells (filled diamonds) and CD4+CD25+/CD4+ ratio (open diamonds) after one injection of CTX (30 mg/kg i.p.) in 21-day-old PROb tumor-bearing rats (n=4). (d) Evolution of MΦ-mediated cytotoxicity evaluated as in (b). (e) Effects of a single injection of CTX on the capacity of splenic lymphocytes from rats bearing 21-day-old PROb tumors to induce MΦ-mediated cytotoxicity against PROb in a co-culture of 1×104 PROb cells with 1×105 MΦ and 1×105 splenic lymphocytes. Column 1: untreated rats; column 2: rats treated with CTX 7 days before spleen resection. Effects of the addition of 2.5×105 CD25+ (column 3) or CD25 (column 4) lymphocytes from untreated tumor-bearing rats. Results are representative of four independent experiments (mean ± SD of triplicates).

2.5 A single injection of cyclophosphamide delays tumor growth and increases the efficacy of an active immunotherapy

A single CTX injection (25 mg/kg) significantly delayed tumor growth in rats bearing established 21-day-old PROb tumors, compared with non-treated rats (p<0.001), but the tumors still progressed (Fig. 6a). This delay did not result from a direct cytotoxic effect of CTX on the tumor cells, since the CTX treatment did not demonstrate any significant effect on PROb tumors growing in T cell-deficient nude rats (Fig. 6b) or nude mice (data not shown).

Since CTX-induced regulatory CD4+CD25+ T cell depletion was not sufficient for obtaining a lasting complete response of established PROb tumors, we sought to enforce its effect by raising an active immune response against the tumor cells. We found that an intradermal (i.d.) injection of PROb cells (1×106) mixed with Mycobacterium bovis Bacillus Calmette-Guérin (BCG; 80,000 CFU) in tumor-free rats induced complete protection against tumors resulting from a subsequent subcutaneous (s.c.) injection of 1×106 PROb cells (n=10). In rats bearing established PROb tumors, this treatment significantly delayed tumor progression (p<0.001), but did not cure the animals since tumor progression was still observed (unpublished results and Fig. 6a).

To determine whether CTX-induced CD25+ T cell depletion could sensitize animals with established tumors to this immunization procedure, rats bearing established 21-day-old PROb tumors (mean tumor volume: 184±30 mm3) received a single CTX injection (25 mg/kg), followed 7 days later by an i.d. injection of PROb cells mixed with BCG. This combined treatment induced a complete regression of established tumors, and all the treated rats remained tumor-free 90 days after tumor cell injection (Fig. 6a). These animals were durably immunized, as no tumor appeared in six rats which received a second PROb cell challenge 6 weeks after the combined treatment. In contrast, s.c. injection of a syngeneic glioma cell line induced tumors which grew similarly in rats immunized against PRO cells after the combined treatment and in naive animals (data not shown).

The effect of CTX was confirmed in another experiment in which three groups of four rats bearing established PROb tumors (100±33 mm3) were either left untreated, received an unique injection of CTX (25 mg/kg), or received a CTX injection followed 7 days later by an i.d. injection of PROb cells mixed with BCG. The tumors progressed in all animals of the first two groups, with a mean tumor volume of 611±47 mm3 and 329±109 mm3, respectively, 4 weeks after CTX injection. Again, the combined treatment induced a complete regression of PROb tumors and the four animals remained tumor-free when sacrificed 44 days after treatment.

Figure 6.

Sensitization to immunotherapy of established PROb tumors after CTX-induced regulatory T cell depletion. (a) Comparison of PROb tumor growth (mean ± SD) in four groups of six rats, which received, 21 days after s.c. injection of 1×106 PROb cells, either CTX followed 7 days later by an i.d. injection of PROb cells mixed with BCG (open squares), or CTX alone (filled diamonds), or an i.d. injection of a mixture of 1×106 PROb cell with BCG alone on day 28 (filled triangles). A control group received no treatment (filled squares). All rats presented progressively growing tumors and were sacrificed at day 77, except for rats receiving combined treatment, which were all tumor-free at day 90. (b) Comparison of PROb tumor growth (mean ± SD) in two groups of five nude rats bearing 21-day-old PROb tumors which received (filled squares) a single injection of CTX (arrow) or were not treated (open diamonds).

3 Discussion

It was recently reported that CD4+CD25+ regulatory T cells accumulated in the peripheral blood and tumor microenvironment of patients with various carcinomas 1719. The present study enforces these observations by showing that CD4+CD25+ T cell expansion in the lymphoid organs of tumor-bearing animals is correlated with the tumor volume. These cells express TCR α / β, RT6 and low levels of CD45RC, but do not express Thy-1, a phenotype similar to that of regulatory T lymphocytes involved in self tolerance in the rat 6, 20.

We compared two clones of tumor cells obtained from the same carcinoma and sharing tumor rejection antigens but differing in their capacity for inducing a tumor-specific immune response. Theseclones made possible to explore the effects of regulatory T cells from animals bearing a progressive tumor on the immune response induced by an immunogenic tumor clone. CD4+CD25+ T cells isolated from rats bearing the progressive PROb tumor delayed the regression of tumors induced by co-injected REGb cells and prevented T lymphocytes from rats immunized with the immunogenic REGb cells from secreting IFN-γ, a cytokine that mediates tumor regression through MΦ activation 16. This interaction required a close contact between regulatory T cells from tumor-bearing animals and immune T cells from rats which had rejected the regressive cell variant. The role of TGF-β in the immune suppression mediated by regulatory CD4+CD25+ T cells gives rise to controversy (reviewed in 21). In our experimental model, the suppressive effect of CD4+CD25+ T cells was abrogated by addition of a blocking antibody against TGF-β.

In difference to CD4+CD25+ T cells isolated from the spleen of PROb tumor-bearing rats, those isolated from the spleen of REGb tumor-bearing rats exerted tumor-suppressing effector functions, notably through IFN-γ secretion. CD25 is known to be expressed by both activated and regulatory CD4+ T cells 5. We previously demonstrated that immunogenicity of the REGb cell variant could be related to its capacity to die and release antigenic proteins when injected in vivo19, 22. In contrast, injected PROb cells did not die and gave rise to tolerogenic progressive tumors. CD4+CD25+ T lymphocytes with an activated rather than regulatory function could also be isolated from the spleen of rats bearing regressing tumors obtained after injection of a PROb cell variant in which cytochrome c expression was down-regulated by transfection of an antisense construct. Cytochrome c depletion enhances PROb cell death and immunogenicity in vivo (E. Schmitt et al., manuscript in revision), further indicating the relationship between cell sensitivity to death and modulation of the immune response.

Interestingly, whereas total splenic T lymphocytes from rats bearing established PROb tumors seemed to be devoid of any reactivity against tumor cells, CD25+ T cell depletion revealedthat the spleens from these animals contained CD25 T cells exhibiting anti-tumor activity. This observation indicated that the immune dysfunction characterizing advanced human and experimental cancers might be reversible by depleting CD25+ T cells.

Antibody-mediated depletion of CD25+ T cells was shown to facilitate the induction of tumor immunity in rodent models 1014. In all these reports, anti-CD25 antibody injection preceded or was performed in the 2 days following tumor cell injection. We searched for a more efficient method to break down the immune tolerance due to tumor-induced CD4+CD25+ regulatory T cells, and to obtain curative effects on established tumors. Among a panel of cytotoxic or immunosuppressive agents, CTX was the most efficient in decreasing the ratio of CD4+CD25+/CD4+ cells in the spleen and in restoring an anti-tumor immune response in tumor-bearing animals. Low doses of CTX were previously shown to reduce suppressor T cell function in mice 23 and in cancer patients 24, and to restore the efficacy of adoptive immunotherapy in experimental tumors 25]. However, the nature of the depleted suppressor cells was not characterized.

A single CTX injection has no major direct effect on the PROb tumor cells, as it does not significantly modify the growth of PROb tumors in athymic nude rats and mice. CTX alone delays tumor progression but does not cure PROb tumors in immunocompetent rats. The most important finding of the present study is that CTX treatment followed 7 days later, at the time of maximal CD4+CD25+ regulatory T cell depletion, by an active, tumor-specific immunotherapy, cured all the rats bearing well-established PROb tumors. Immunotherapy consisted in an i.d. injection of PROb cells mixed with BCG, a treatment able to prevent PROb tumor growth when administered before tumor cell injection. An immunotherapy associating BCG with autologous tumor cells has been reported to givesignificant, even if limited, clinical benefits after colon cancer resection 26. The cure of PROb tumors induces a lasting immune protection against a second challenge with the sme tumor line, but not against an unrelated glioma.

Altogether, these results demonstrate that expansion of CD4+CD25+ regulatory T cells during tumor progression raises a potent barrier against tumor rejection induced by immune CD25 T cells and constitutes an additional obstacle for achieving a successful cancer immunotherapy. Depletion of CD4+CD25+ regulatory T cells can be achieved by a single injection of CTX and potentiates the efficacy of an active tumor-specific immunotherapy. Breaking down CD4+CD25+ regulatory T cell efficiency thus appears to be an attractive strategy for circumventing immune tolerance induced by tumors and to tip the balance towards an effective immunotherapy against established tumors.

4 Materials and methods

4.1 Animals and cell lines

Syngeneic BD-IX strain rats, 10–12 weeks of age, were bred in our laboratory. The PROb and REGb cell lines were established from a chemically-induced colon carcinoma in a BD-IX strain rat 15. PROb cells stably transfected with cDNA encoding cytochrome c introduced in an antisense orientation in pTARGET were also used. GV1A1, a cell line obtained from a BD-IX strain ratglioma, was obtained from Dr. Eva Klein (Karolinska Institute, Stockholm, Sweden). Cells were cultured in RPMI 1640 medium supplemented with 10% FBS and were monitored routinely and found to be free of Mycoplasma infection. PROb and REGb tumors were obtained by injecting 1×106 tumor cells s.c. in the thoracic wall of 10–12-week-old BD-IX rats. Animal use and handling were performed according to the French laws for animal experimentation.

4.2 Cytotoxic and immunomodulatory agents

Rats were treated with a single injection of CTX (30 mg/kg, i.p.), methotrexate (2 mg/kg, i.p.), vinblastin (0.1 mg/kg, i.v.), 5-fluorouracil (20 mg/kg, i.v.), doxorubicin (1.5 mg/kg, i.v.), cisplatin (3 mg/kg, i.v.), or 6α-methylprednisolone (15 mg/kg, i.p.), all obtained from Sigma Aldrich (St-Quentin-Fallavier, France) or an anti-rat CD25 (OX-39) mAb (1 mg per rat, i.p.) from Serotec (Oxford, GB). Rabbit polyclonal antibodies against rat IFN-γ and IL-10 (Serotec) and mAb against TGF-β1, TGF-β2 and TGF-β3 (R & D Systems, Oxon, GB) were used as immunomodulatory components in vitro.

4.3 Antibodies and FCM analysis

We obtained murine anti–rat mAb against CD4 (W3/25-FITC), CD25 (OX-39), TCR α / β (R7/3), CD90 (Thy-1) (OX-7), and CD45RC (OX-22); and isotype control mAb (Serotec) and donkey anti-mouse Ig-conjugated R-phycoerythrin (DAM-PE; Jackson, West Grove, PE) were used for FCM. Two-color FCM analyses were performed on 1×106 cells incubated with OX-39, then with DAM-PE. The cells were incubated with W3/25-FITC and then analyzed by FACScan (Becton Dickinson, Mountain View, CA) using CellQuest Software.

4.4 Lymphocyte and dendritic cell isolation

Cells from spleen and axilliary lymph nodes were mechanically dissociated before passage through a nylon wool column to obtain a suspension containing at least 95% T cells after FACScan analysis. The separation of CD25+ T cells was performed using an anti-CD25 mAb and anti-mouse IgG1-coated magnetic beads (Miltenyi Biotec, Auburn, CA) according to the manufacturer's instructions. About 90% of the positively selected cells were CD4+CD25+ based on FCM analysis. Enriched populations of splenic dendritic cells were isolated as previously reported 27. These populations were comprised of 50–60% dendritic cells that were identified by their distinctive morphology and labeling with anti-MHC class II mAb.

4.5 Inhibition of immune T cell response to tumor antigens

T lymphocytes (1×105) and dendritic cells (1×104), isolated from the spleen of a BD-IX rat immunized by three monthly injections of REGb cells, were mixed with 1×104 REGb cells. Tumor cells were pretreated with mitomycin C (50 μ g/ml for 2 h) to prevent proliferation. After a 3-day culture period in the presence of splenic lymphocytes from immunized rats, IFN-γ levels were determined in the culture supernatant using a colorimetric ELISA kit for rat IFN-γ (R & D). To determine whether the effect of regulatory T cells on tumor immune T cells required a direct contact, a Transwell system (Corning Costar, Brumath, France) was used, CD4+CD25+ cells from tumor bearing rats being seeded in the upper chamber, whereas T lymphocytes isolated from the spleen of immunized rat were seeded in the bottom compartment with dendritic cells and tumor cells.

4.6 Restoration of an immune response against PROb cells in vitro

T lymphocytes (1×105) and dendritic cells (1×104) isolated from the spleen of BD-IX rats bearing 28-day-old PROb tumors were mixed with 1×104 mitomycin C-treated PROb cells. In some wells, lymphocytes were depleted from CD25+ cells by incubation with anti-CD25 mAb and rabbit complement (Low Tox; Cedarlane, Hornby, Canada). After a 3-day mixed culture, the IFN-γ levels were measured in the supernatant using an ELISA kit.

4.7 Induction of cytotoxicity against PROb cells in vitro

Splenic T lymphocytes from naive or 42-day-old PROb tumor-bearing rats were mixed with resident peritoneal MΦ from naive BD-IX rats and PROb cells, at a ratio of 1/10/50. The lymphocytes were depleted or not depleted in CD25+ cells. After a 48 h mixed culture, the density of the residual attached PROb cells was evaluated after crystal violet staining as previously reported 16.

4.8 Cyclophosphamide treatment and active immunotherapy

Tumor-free or tumor-bearing rats (n=5–6 per group) received a single i.p. injection of CTX (25 mg/kg; Sigma) or vehicle. Seven days later, an i.d. injection of 1×106 living PROb cells mixed with BCG (80,000 CFU; Pasteur-Merieux, Paris, France) was performed and the tumor volume was evaluated weekly using a caliper.

4.9 Statistics

Statistical significance was determined by the two-tailed Student's t-test or analysis of variance, and p values less than 0.05 were considered as significant.

Acknowledgements

This work was sponsored in part by the French National League against Cancer (National, Burgundy and Saône-&-Loire committees). We thank J. H. Landor, S. Gurbuxani and L. Zitvogel for helpful comments and advice. We thank M. Moutet and A. Fromentin for technical assistance.

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

  6. 6

    WILEY-VCH

Ancillary