The naturally occurring CD4+ regulatory T cells (Treg) are continuously produced by the thymus and gradually settled in secondary lymphoid organs. They constitute 5–15% of the overall CD4+ T cell population and exert a strong suppressive activity on multiple components of the immune system. Although their repertoire in antigen recognition is quite diverse, they preferentially recognize tissue specific self-antigens. As many tumors express self-antigens, CD4+ Treg do not only dominantly suppress the activation and expansion of different effector cells capable of mediating autoimmunity, but also effective antitumor responses.1
So far the most specific marker for naturally occurring CD4+ Treg, at least in mice, is FoxP3, a member of the forkhead family of DNA-binding transcription factors. FoxP3 is highly expressed in naturally occurring CD4+ Treg and clearly linked to their suppressive function.2, 3, 4 Other molecules that are constitutively expressed on CD4+ Treg are the IL-2 receptor α-chain (CD25), cytotoxic lymphocyte-associated antigen-4 (CTLA-4) and different members of the tumor necrosis factor (TNF) receptor superfamily like the glucocorticoid-induced TNF related protein (GITR). These molecules, however, do not uniquely distinguish the CD4+ Treg from conventional CD4+ T cells that can temporally upregulate these molecules following activation.5, 6, 7, 8, 9 Recent data also suggest an important role of Toll-like receptors (TLRs) in controlling Treg.10 TLR2-triggering on murine Treg in combination with T cell receptor ligation resulted, both in vitro and in vivo, in proliferation of the otherwise anergic Treg. Moreover, the suppressive phenotype of Treg was temporarily abrogated in the presence of a TLR2-ligand, but was fully regained after removal of the TLR2-trigger.11, 12
CD4+ Treg are known to recirculate through lymphoid tissues and are able to enter various sites under pathological conditions.13 Recent studies emphasized that CD4+ Treg cells can infiltrate tumor masses and locally hinder the function of effector cells in different types of murine and human tumors.14, 15, 16, 17, 18, 19, 20 Their potency is most clearly demonstrated by in vivo studies. In mice, depletion of CD4+ Treg for a limited time provoked tumor regression without inducing apparent autoimmunity.21, 22, 23 However, still little is known about the in vivo accumulation and activation of Treg during tumor growth.
Brain tumors differ from other tumors in that they arise at an immune specialized compartment characterized by the lack of conventional lymphatics and the presence of a blood-brain-barrier, which selectively regulates the uptake of leucocytes.24 Besides their special location, many other mechanisms have been identified, which enable brain tumors to regularly escape from immune surveillance, such as secretion of immunosuppressive cytokines like TGF-β and IL-10 or immunosuppression by cell-mediated interactions via CD70, CD95-ligand or HLA-G.25, 26
To analyze the contribution of CD4+ Treg in the immune escape of brain tumors, we examined the presence and function of CD4+ Treg in an orthotopic, syngeneic murine glioma model. Our data presented here provide evidence that CD4+FoxP3+ Treg infiltrate and accumulate in gliomas during tumor growth, are activated inside the tumor and strongly suppress antiglioma responses. Furthermore, we demonstrate that the depletion of Treg prior to the tumor challenge is sufficient to break immune tolerance and to induce an effective antiglioma immune response, which can be further enhanced by additional anti-CTLA-4 blockade.
Material and methods
Mice and tumor cell lines
Eight- to 10-week-old female C57BL/6 mice were obtained from Charles River Wiga (Sulzfeld, Germany). They were kept under pathogen-free conditions in the animal facility of our institute. All animal experiments were approved by the Animal Experimental Committee of the Radboud University Nijmegen Medical Centre and were performed in accordance with institutional and national guidelines.
The GL261 mouse glioma cell line was kindly provided by U. Herrlinger (Bonn, Germany). GL261 cells were cultured in vitro in Iscove's IMDM (Gibco, Gaithersburg, MD) supplemented with 9% fetal calf serum (Sigma-Aldrich), 100 U/ml penicillin and 100 μg/ml streptomycin (1% P/S, Gibco) and 20 μM β-mercaptoethanol (complete medium). GL261 cells express MHC class I, but are MHC class II-negative. After INF-γ pretreatment GL261 cells upregulate MHC class I-molecules and become MHC class II-positive. CD25, CTLA-4, CD4 or CD8 was not detectable on GL261 cells (data not shown). The melanoma cell line B16-F10 was obtained from the American Type Culture Collection and cultured in complete medium. Cells were regularly monitored for mycoplasma contamination by PCR.
Brain tumor model
Prior to injection, GL261 cells were harvested, washed twice in phosphate-buffered saline (PBS), counted and adjusted to 2 × 104 cells in 5 μl of PBS in a 26-gauge Hamilton syringe. Mice were anaesthetized with isoflurane. After shaving the scalp and making an incision, a burr hole was made in the skull 2 mm lateral to the midline and 2 mm anterior to the bregma using a dental drill. Then, GL261 cells were injected over 1 min at a depth of 2.5 mm below the dura mater into the right cerebral hemisphere. Animals were observed daily and killed by cervical dislocation when characteristic symptoms such as hunched posture, reduced mobility and significant weight loss (>20%) occurred. In this model, the appearance of these signs reliably predicted the time of death within 2–3 days of anticipation. Animals without such symptoms were regarded as long-term survivors after 90 days. Representative animals in each treatment cohort were euthanized at selected time points to obtain tissues (brain tumors, lymphoid organs) for immunological analyses.
Isolation of tumor-infiltrating lymphocytes and tumor-derived Treg
For isolation of tumor-infiltrating lymphocytes (TIL) mice were anaesthetized with isoflurane and perfused through the left cardiac ventricle with isotonic Ringer's solution + 1 U/ml heparine. Brains were harvested, the tumor bearing part of the right cerebral hemisphere was dissected and weighted. At the final stage, when severe clinical symptoms became apparent, the tumor mass itself was collected. TIL were isolated by enzymatic digestion and modified Ficoll-Hypaque centrifugation as previously described.27 Briefly, brain tissue was incubated with Collagenase Type IA (50 mg/ml), DNAse Type I (10 μg/ml) and a trypsine inhibitor (Sigma-Aldrich) in Hanks Balanced Salt Solution (HBSS) with Ca2+, Mg2+ at 37°C. After 45 min EDTA was added (12.5 mM final), a cell suspension was prepared, filtered and transferred to a 15 ml tube. Thereafter, a 1g sedimentation was performed, the supernatant was collected 30 min later and put on a modified Ficoll gradient (75% Ficoll/25% RMPI supplement with 10% FCS). The interphase mainly containing myelin debris and dead cells was removed and the pellet was used for further analysis after washing. Brain-infiltrating lymphocytes (BIL) from the left, nontumor bearing cerebral hemisphere were prepared similarly. For Treg isolation, cells were subsequently put on a second density gradient using different concentrations of Percoll (70, 37, 30%). After centrifugation, cells from the 37%/30% interphase containing a high proportion of activated Treg cells were harvested, incubated with anti-CD25-FITC mAb and sorted with anti-FITC-MACS beads (Miltenyi Biotec, Germany). CD8 T cells and B cells were then depleted by anti-CD8 or anti-CD19 DynaBeads (Invitrogen) according to the manufacturer's instructions. After purification, the remaining cells were collected, washed, counted and used for the suppression assay.
Antibodies and flow cytometry
Directly labeled mAbs used for staining were anti-CD4-APC (RM4.4), CD8b (53–6.72)-APC, INF-γ-PE, anti-CD25-FITC (clone 7D4), CXCR4-FITC (kindly provided by R. van der Voort). All antibodies and isotype controls were obtained from BD Pharmingen, CA. A purified hamster antimouse mAb (BD Pharmingen) was used for staining of CTLA-4. FITC conjugated goat antihamster IgG polyclonal antibodies were used for detection. Rat antimouse GITR antibodies were purified from culture supernatant of DTA-1 hybridoma cells (generously provided by S. Sakaguchi) and used for staining. GITR expression was detected by FITC-conjugated goat antirat IgG polyclonal antibodies. Free binding sites were blocked by incubation with total rat IgG. Rat antimouse FoxP3-PE (clone FJK, staining according to instructions by manufacturer) was obtained from eBioscience, CA. Analysis of cell surface markers on lymphocytes was performed using a FACScalibur™ (Becton Dickinson) and CELLQuest™ software.
T cell purification and suppression assay
Spleens were mashed, filtered and CD4+ T cells were purified using antimouse-CD4 Microbeads (MACS, Miltenyi Biotec) resulting in a 95% pure CD4+ T cell population as measured by flow cytometry. Freshly sorted CD4+ naïve T cells (5 × 104) were labeled with CFSE (5 μM), seeded in a 96-well plate cultured with tumor-derived Treg cells (5 × 104) for 4 days in the presence of 1 μg/ml soluble anti-CD3 (clone 145-2C11, BD Pharmingen), 105 irradiated APC (CD4-MACS bead depleted splenocytes) in complete medium containing 5 cU/ml hIL-2. Resting cultured Treg cells, derived from a previously established Pam3Cys-SKKKK (Pam) stimulated Treg cell line, were used as control. Suppression/proliferation was monitored by FACS-analysis of the CFSE fluorescence after staining with anti-CD4-APC.
Treatment of GL261-bearing mice
Depletion of Treg was achieved by injecting mice i.p. with 250 μg of antimurine CD25 mAbs (PC61, rat IgG1 isotype) on day −3 before tumor inoculation. Treatment with antimurine CTLA-4 mAbs (9H10, hamster IgG isotype) started 3 days later. Antibodies were delivered intraperitoneally at 100 μg in PBS, followed by 2 injections of 50 μg, 3 and 6 days later. Control rat IgG and hamster IgG were obtained from Sigma-Aldrich. Depletion of CD8 or CD4 T cell subsets in naive C57BL/6 mice was achieved by i.p. injections of depleting mAb GK1.5 (anti-CD4) or 2.43 (anti-CD8) on days −1, 0 or day +5, 6 after tumor inoculation, respectively. A total of 200 μg of mAb was injected per mouse.
Immunomonitoring for GL261-specific effector cells
Splenocytes (1–2 × 106) were restimulated with irradiated INF-γ pre-treated GL261 in a 24-well plate (5 × 104 per well) and cultured in 1.5 ml complete medium supplemented with 10 cU human IL-2. After 3 days a Ficoll gradient was performed to remove dead cells and remaining cells were replated with IL-2. After 7 days, cells were harvested and used for an intracellular INF-y staining. Cells (5 × 104) were plated in a 96-well plate and stimulated with different target cells (GL261, B16F10) in the presence of Brefeldin A (10 μg/ml). Cells were harvested after 5 hr coculture. Surface staining was performed with CD4-APC or CD8-APC. Intracellular INF-γ was stained after fixation and permeabilization with PFA4% and 0.1% Saponine.
Determination of anti-GL261 humoral response
For the detection of glioma-specific antibodies against GL261 glioma cells, sera from treated animals were pooled (3–4 mice per group) at day 48 after tumor inoculation. Sera from nontumor bearing naïve mice and untreated tumor bearing mice were used as controls. GL261 cells were incubated with different serum dilutions (1/160–1/5120) for 30 min at 4°C, washed, then incubated either with a rat antimouse IgG1-FITC-conjugated antibody or a rat antimouse IgG2a-FITC conjugated antibody (Becton-Dickinson) and analyzed on a FACS flow cytometer. GL261 cells without serum incubation were used as negative control.
Histology and immunohistochemistry
For histological analysis, animals were anaesthetized and perfused as described earlier at day 17 or day 90 after intracerebral tumor inoculation. Brains were either snap frozen in 2-methylbutane (Merck, Wertheim/Main, Germany) and stored at −80°C (day 17) or fixed in 4% PFA and embedded in paraffin (day 90) until used for sectioning. Coronal sections (5 μm) were made. Immunohistochemistry was carried out on cryosections using hamster antimouse CD3 antibodies (BD Pharmingen). Binding of the primary antibodies was detected using a biotin-labeled goat antihamster IgG secondary antibody (BioSource International, CA), horseradish peroxidase-conjugated avidin and a freshly prepared solution of AEC substrate (3-amino-9-ethyl-carbazole; Sigma, Buchs, Switzerland) and H2O2. Sections were counterstained with hematoxylin and eosin (H&E). Myelin staining was performed on paraffin sections using Luxol Fast Blue and H&E as a counterstain.
A two-tailed Student t-test was used to analyze for significant differences between 2 treatment groups. One-way ANOVA was used to analyze differences between 3 and more groups. A Post test was performed with p values <0.05. Differences in survival between 2 treatment groups were analyzed for significance using the log-rank test. Data were analyzed by GraphPad Software 4.0.
Predominant infiltration of CD4+FoxP3+ Treg into a murine glioma
To dissect the role of regulatory T cells in brain tumors, we implanted syngeneic GL261 glioma cells intracerebrally into the right hemisphere of C57Bl6/N mice. Mice were perfused at 4, 10, 17, 24 days after tumor inoculation and at the final stage when severe clinical symptoms became apparent. Subsequently, TIL from the tumor bearing part of the right cerebral hemisphere, BIL from the left, nontumor bearing cerebral hemisphere, cervical lymph node cells and splenocytes were isolated and analyzed for the presence of CD4+FoxP3+ Treg. Strikingly, a time dependent increase in the frequency of CD4+FoxP3+ Treg relative to the total number of CD4+ T cells at the tumor bearing site was observed. A significant increase in the frequency of CD4+FoxP3+ Treg was already detectable 10 days after tumor inoculation (20.7 ± 1.2)%, reaching a maximum of (38.5 ± 6.1)% Treg of total CD4+ cells at the final tumor stage when mice were symptomatic. In contrast, the Treg/total CD4+ T cell ratio in the spleen (13.1% ± 0.7%–16.2% ± 1.6%) and cervical lymph nodes (13.6% ± 1.1%–18.9% ± 4.4%) remained essentially the same during tumor growth apart from a slight increase at the final tumor stage (Fig. 1a). Moreover, the Treg/total CD4+ T cell ratio at the tumor bearing site also exceeded that from the nontumor bearing site of the brain. Comparison of the absolute numbers of CD4+FoxP3+ Treg revealed an up to 300-fold increase in Treg numbers at the tumor-bearing site relative to the Treg numbers at the nontumor bearing site of the brain (Fig. 1b). Finally, FoxP3-immunhistochemistry revealed that Treg are preferentially located at the tumor site (not shown). Altogether, these data demonstrate that CD4+FoxP3+ Treg accumulate in intracerebrally growing glial brain tumors.
Intratumoral CD4+FoxP3+ Treg display an activated phenotype
Next, we investigated the expression of the Treg activation marker CD25 on CD4+FoxP3+ Treg isolated from the tumor bearing and nontumor bearing part of the brain during tumor growth. As shown in Figure 2a, from day 17 onwards an increase in the percentage of CD25+ Treg as well as of the mean fluorescence intensity of CD25 was detected on Treg present in the tumor bearing part relative to the nontumor bearing part of the brain. Previously, we and others have demonstrated that CD25 is upregulated following activation.6, 11, 28 To further proof the accumulation of activated Treg in the tumor, we stained freshly isolated CD4+FoxP3+ Treg isolated from different organs of a glioma bearing symptomatic mouse for their expression of CD25, GITR, CTLA-4 and the chemokine receptor CXCR4. Remarkably, all 4 Treg activation markers were by far most abundant on the tumor-infiltrating CD4+FoxP3+ Treg. Especially, GITR and CTLA-4 expression levels were strongly upregulated on tumor-derived Treg as compared with the Treg isolated from cervical lymph node or spleen (Fig. 2b). According to these data, we concluded that glioma-derived CD4+FoxP3+ Treg not only accumulate at the tumor site, but also exhibit an activated phenotype.
Freshly isolated intratumoral Treg are highly suppressive
To determine the suppressor activity of intratumoral Treg, CD25+CD4+FoxP3+ Treg were purified from glioma bearing symptomatic mice as described in Material and methods and tested directly in vitro for their ability to inhibit the proliferation of naïve CD4+ T cells in a suppression assay. Their potency was compared with pure CD25+CD4+FoxP3+ Treg generated using our recently developed Pam3Cys-SKKKK (Pam)-Treg expansion protocol.11 As shown in Figure 3, glioma-derived Treg were as effective as Pam-Treg in suppressing the proliferation of CFSE-labeledCD4+ T cells. We note that both glioma-derived Treg and Pam-Treg expressed similar levels of CD25. Titration experiments indicated that the Pam-Treg used were still able to mediate suppression at a Treg: responder T cell ratio of 1:16, further emphasizing the potency of glioma-derived Treg (data not shown).
Anti-CD25 treatment inhibits intratumoral accumulation of Treg and results in effective antiglioma immune responses
The accumulation of CD25+CD4+FoxP3+ Treg in the glioma tumors, let us to investigate the impact of anti-CD25 mAb mediated Treg depletion on tumor development in the absence of vaccination. Hereto, mice were injected i.p. with anti-CD25 mAbs (PC61) 3 days before tumor inoculation. Figure 4a shows that over 90% of the CD4+CD25+ cells in the spleen are depleted at the time of tumor inoculation (day 0) after injection of anti-CD25 mAbs. Moreover, the percentage of CD4+CD25+ cells in the spleen was still significantly diminished at day 17 after tumor inoculation. Analysis of TIL isolated from treated and nontreated animals at day 17 after tumor inoculation demonstrated that the anti-CD25 treatment strongly impedes with the intratumoral accumulation of Treg within a growing glioma. The administration of anti-CD25 mAbs resulted in both a significant reduction of CD4+FoxP3+ Treg and a significant decrease in CD25+ CD4+FoxP3+ cells within the TIL relative to control-treated animals. Moreover, CD4+FoxP3+ Treg reduction of anti-CD25 treated animals was more prominent within the TIL compared with the splenocytes (Fig. 4b). Further analysis of the TIL revealed that the CD25 expression levels on CD4+FoxP3+ Treg of anti-CD25 treated animals treatment were significantly lower than those on CD4+FoxP3+ Treg of control-treated animals (Fig. 4c). Remarkably, about 50% (54.5%, p < 0.0001) of the anti-CD25 mAb-treated mice survived without developing clinical symptoms, while all isotype control mAb treated mice developed severe clinical symptoms (Fig. 4d). These data thus demonstrate that anti-CD25 mAb treatment provokes regression of glioma tumors by reducing the number of CD4+FoxP3+ Treg, including CD25+CD4+FoxP3+ glioma-infiltrating Treg, in the absence of antitumor vaccination.
The antiglioma immune response induced after Treg depletion is CD4 and CD8 T cell dependent
Analysis of the immune response revealed that anti-CD25 treatment strongly enhanced the number of INF-γ producing T cells as compared with control tumor bearing mice, while INF-γ producing T cells were not detected in naïve mice (Fig. 5). INF-γ secretion appeared to be glioma-specific, as B16 melanoma cells failed to enhance INF-γ secretion following anti-CD25 treatment. Moreover, anti-CD25 mAb-treatment enhanced both GL261-specific CD4+ and CD8+ INF-γ producing T cells after restimulation with INF-γ pretreated GL261 cells (which express MHC class II). To further analyze the importance of CD4+ and CD8+ T cells in tumor eradication following Treg depletion, we codepleted CD4+ or CD8+ T cells in combination with Treg before or after tumor challenge. Survival data shown in Figure 6 demonstrated that both CD4+ and CD8+ T cells play a crucial role in controlling glioma outgrowth. CD4 depletion prior to tumor inoculation completely abrogated the antitumor-response (Fig. 6a). Similarly, depletion of CD8+ T cells prior to tumor inoculation also strongly affected the immune response, although CD8 depletion was somewhat less effective as CD4 depletion (Fig. 6b). Severely reduced antitumor immunity was also observed following CD4 depletion after tumor inoculation (Fig. 6c) and was even completely abrogated when CD8+ T cells were depleted (Fig. 6d). These data indicate a crucial role for CD4+ T cell help during both the priming phase and effector phase of the response as well as of CD8+ T cells. Thus both CD4+ and CD8+ T cells are important to eradicate glioma tumors after CD25-mediated depletion of suppressive Treg.
Anti-CTLA-4 blockade synergizes with Treg depletion in tumor eradication
To evaluate whether the depletion of regulatory T cells in combination with CTLA-4 blockade results in a further increase of therapy efficacy, mice were either treated with anti-CD25 monoclonal antibodies (PC61, 250 μg i.p.) 3 days before tumor challenge, anti-CTLA-4 monoclonal antibodies (9H10, 100 μg i.p.) at day 0, +3 and day +6 after tumor inoculation or a combination of both and monitored for clinical symptoms. Whereas anti-CTLA-4 treatment alone was almost as effective as Treg depletion with rejection of 50% of the gliomas, all mice were capable of rejecting the tumor when we combined Treg-depletion with CTLA-4 blockade (Fig. 7a). Analysis of the antitumor immune response in these mice revealed that the increase in therapy efficacy was accompanied by a significant increase in the number of specific CD8+ INF-γ producing T cells and a moderate increase in the number of glioma-specific CD4+ INF-γ producing T cells as compared with anti-CTLA-4 or anti-CD25 treatment alone. Direct ex vivo analysis of total TIL isolated from CD25/CTLA-4 mAb treated mice demonstrated their potent capacity to produce large amounts of INF-γ, whereas INF-γ producing TIL were essentially absent in untreated control mouse (Fig. 7b). Further analysis showed that besides cellular immunity also tumor-specific humoral immunity was boosted by anti-CD25 plus anti-CTLA-4 treatment (Fig. 7c). Finally, anti-CD25 plus anti-CTLA4 treated mice that had survived the first tumor challenge, were rechallenged with a 5-fold higher intracranial tumor dose. As shown in Figure 7d all mice survived without any further treatment, whereas all of the control mice died. These data thus indicate that a strong protective memory response was induced after the combination therapy with anti-CD25 and anti-CTLA-4 mAbs.
The immune response after anti-CD25 and anti-CTLA-4 treatment is directed against the tumor, but not towards normal brain tissue
To determine whether the immune response after anti-CD25 and anti-CTLA-4 treatment is associated with a risk of induction of deleterious immune response against normal brain tissue, we clinically and histologically evaluated mice for neurological sequelae. All animals that survived the tumor challenge appeared healthy and did not show neurological symptoms. Serial brain sections of anti-CD25 and anti-CTLA-4 treated mice were stained for myelin with Luxol Fast Blue and analyzed for the occurrence of subclinical signs of autoimmunity like myelin damage. Despite the presence of large numbers of T cells that surrounded and infiltrated a growing glioma at day 17 after tumor inoculation (Fig. 8a), neuropathological analyses at day 90 after tumor challenge failed to demonstrate any signs of demyelination. Perivascular lymphocytic infiltrates, which would be consistent with the induction of autoimmunity, were not observed. Residual tumor cells could not be detected, although some siderophages were present at the site of the previous inoculation of tumor cells (Fig. 8b). These data suggest that the effector cells after Treg depletion and anti-CTLA-4 treatment are directed against tumor-specific antigens, but not against normal brain tissue antigens.
CD4+FoxP3+ Treg are crucially involved in controlling the magnitude and type of immune responses against self- and non-self antigens.29 Several studies indicate that Treg play a significant role in the suppression of antitumor immunity, but still little is known regarding their presence and function within tumors.30
Here we demonstrate that CD4+ Treg are able to infiltrate tumors growing at the immune privileged site of the brain. Using a syngeneic mouse glioma model, we observed a time-dependent increase in the frequency and number of CD4+FoxP3+ Treg within the brain tumor during glioma growth. Our data further show that intratumoral CD4+FoxP3+ Treg upregulate several Treg activation markers during tumor growth. Direct ex vivo analysis of CD25-expression on CD4+FoxP3+ Treg revealed that both CD4+FoxP3+ Treg with high expression levels of CD25 and CD4+FoxP3 with no or low CD25-expression can be found within the brain tumor at an early tumor stage. In contrast, at late tumor-stages, especially when mice are symptomatic, almost all glioma-derived CD4+FoxP3+ Treg expressed CD25. Moreover, these Treg also displayed increased expression levels of GITR and CTLA-4. Furthermore, we could proof that these intratumoral CD25+CD4+FoxP3+ Treg are highly efficient in suppressing the activation of naïve T cells in a suppression assay.
We note that CD4+FoxP3+ Treg predominantely accumulated within the tumor and not in healthy brain tissue. Therefore, it is tempting to speculate that the tumor cells are themselves involved in Treg recruitment and activation, especially because GL261 glioma cells can express MHC-class II molecules under inflammatory conditions. This explanation is further supported by our recent observation that MHC-class II transduced B16 melanoma cells are more heavily infiltrated by CD4+FoxP3+ Treg after intracerebral inoculation than wildtype B16 cells (unpublished observation). Other possible candidates are local antigen-presenting cells like resident microglia cells or tumor-infiltrating dendritic cells that have scavenged tumor debris.31 Particularly, at late tumor-stages massive tissue damage with destruction of blood-brain barrier and tumor cell necrosis will occur, thereby facilitating the mobilization and activation of Treg.32
Previous studies also recognized the presence of CD25 high and CD25 low or negative CD4+FoxP3+ Treg subsets in mice. The latter are regarded as a reservoir of resting inactive Treg, which can be rapidly mobilized and become CD25+ upon activation.33 Furthermore, there is substantial evidence that activated Treg are more efficient regulatory cells than untreated Treg.6, 11, 28, 33 On the basis of our data that CD4+FoxP3+CD25− are abundant at the early tumor stages, we favor the idea that they are recruited along with CD4+FoxP3+CD25+ Treg cells to the brain and become activated within the tumor microenvironment thereby contributing to the high numbers of CD4+FoxP3+CD25+ Treg present in the late-tumor stages. However, we cannot exclude that part of the CD4+FoxP3+CD25+ Treg directly differentiate from naïve CD4+FoxP3−CD25− T cells or originate from CD4+FoxP3− CD25+ Thelper cells that are converted into CD4+FoxP3+ Treg inside the tumor.
There is ample evidence that T cell receptor stimulation and IL-2 are required for Treg activation. IL-2 is a key factor promoting CD25 expression. In the absence of IL-2 from activated T cells, CD25 expression is lost on Treg.6, 34, 35 Besides TCR triggering and IL-2 signalling, the tumor microenvironment of gliomas could also directly support Treg activation and development further explaining the large amount of Treg present in these tumors. The glioma microenvironment has been reported to contain high local concentrations of immunosuppressive cytokines such as TGF-β. TGF-β is known to induce FoxP3 expression in CD4+CD25− T cells and to promote the acquisition of regulatory properties in these cells.36, 37 Moreover, TGF-β seems to be necessary for the maintenance or function of Treg by regulating continuous FoxP3 expression, as TGFβ−/− mice develop autoimmune response within several weeks after birth.38, 39
The prominent accumulation of Treg within a glioma prompted us to investigate the effects of Treg depletion on tumor outgrowth. Treg depletion using CD25-specific mAbs has been previously shown to promote tumor rejection in mice.14, 18, 19, 23 However, the observed accumulation of Treg in the intracranial glioma model now allowed us to directly assess the effect of anti-CD25 treatment on the intratumoral accumulation of Treg. Direct ex vivo analysis of glioma-infiltrating Treg demonstrated that the administration of anti-CD25 mAbs (PC61) significantly reduces the number of CD4+FoxP3+ Treg, including CD25+CD4+FoxP3+ glioma-infiltrating Treg. These data are in keeping with recent publications that show that PC61 treatment leads to a rapid disappearance of CD25+CD4+FoxP3+ cells within blood and spleen.40, 41 The fact that the Treg depletion was more efficient within the TIL than the spleen might be explained by a delay in the recruitment of Treg from the periphery to the tumor, as Treg might first settle in the spleen and then migrate to the tumor after recovering from anti-CD25 depletion.
Further experiments showed that the depletion of Treg prior to the intracerebral inoculation of glioma cells efficiently uncovered an otherwise suppressed immune response against glioma-specific tumor antigens. Both glioma-specific INF-y secreting CD4+ and CD8+ T cells were detectable after depletion of Treg. Codepletion experiments confirmed that CD4+ T cells are required for both the initiation and effector phase of the immune responses and hence for effective tumor rejection in vivo. The requirement for CD8+ T cells was most prominent in the effector phase of the antitumor response.
Previous studies have indicated that treatment of mice with anti-CD25 mAbs may not represent the optimal means of inducing tumor immunity through Treg depletion and is only beneficial within a limited time window as antibody administration at the time of tumor inoculation or later is less effective.22 At these later time points anti-CD25 mAbs will not only deplete Treg, but also affect the effector cells that are involved in tumor rejection.
Our further studies revealed that antiglioma treatment is strongly improved when Treg depletion is followed by CTLA-4 blockade. The combination of CTLA-4 blockade and depletion of CD25+ Treg cells resulted in maximal therapeutic efficacy. The exact mechanism of CTLA-4 blockade is still a matter of debate. Some reports argue that anti-CTLA-4 treatment is likely to influence the action of Treg, but more recent data imply that the antitumor effects of CTLA-4 blockade are due to increased T cell activation rather than inhibition of Treg.42, 43 In our model, detailed analysis showed that CTLA-4 blockade and CD25 depletion markedly enhanced the frequency of glioma-specific effector T cells with a predominant increase in the CD8+ T cell numbers. This is in line with previous observations that the frequencies of antigen-specific CD8+ T cells are increased most prominently if CTLA-4 blockade is performed in CD25-depleted animals.44 The combination treatment also induced glioma-specific humoral response with high titers of glioma-specific IgG2a antibodies indicating a strong TH1 immune response.
As autoimmune toxicity has been associated with Treg depletion and CTLA-4 blockade in several preclinical and clinical studies,44, 45, 46 we analyzed mice for signs of autoimmunity. Although C57Bl/6 mice are susceptible to experimental allergic encephalomyelitis (EAE),47, 48 we did not observe any clinical or histological evidence of the induction of encephalomyelitis in any of our mice during an observation period of more than 90 days. The absence of visible CNS autoimmunity after Treg depletion plus CTLA-4 blockade implies that the local immune resonse is mainly directed against glioma-associated antigens in our model system.
In summary, this study demonstrates that Treg play a significant role of in glioma immune escape. We show here that CD4+FoxP3+ Treg gradually accumulate in growing gliomas and display an activated phenotype. Treg depletion by anti-CD25 treatment resulted in a significant decrease of Treg within the tumor and enhanced antitumor immunity against glioma which is further boosted by CTLA-4 blockade without inducing signs of autoimmunity. Further studies will be necessary to establish even more effective treatment protocols to cure preestablished gliomas by either regulating the suppressive function of intratumoral Treg or changing the intratumoral balance of Treg and effector T cells (Teff) towards an increasing Teff/Treg ratio. As impaired T cell responses due to an relative increase of Treg in the CD4 compartment in patients with malignant glioma have been reported49 and Treg can be detected within human glioma tissue (based on available data50 and our own observations) glioma-infiltrating Treg represent an attractive target for glioma immunotherapy.
This work was supported by a grant from the German Research Foundation (Dr. Oliver M. Grauer, GR2089/1-1) and funds from the Dutch Cancer Society (Prof. Gosse J. Adema, KWF2004-2893).