Enhancement of anti-tumor immunity through local modulation of CTLA-4 and GITR by dendritic cells

Authors


Abstract

Cancer vaccines have now demonstrated clinical efficacy, but immune modulatory mechanisms that prevent autoimmunity limit their effectiveness. Systemic administration of mAbs targeting the immune modulatory receptors CTLA-4 and glucocorticoid-induced TNFR-related protein (GITR) on Treg and effector T cells augments anti-tumor immunity both experimentally and clinically, but can induce life-threatening autoimmunity. We hypothesized that local delivery of anti-CTLA-4 and anti-GITR mAbs to the sites where T cells and tumor antigen-loaded DC vaccines interact would enhance the induction of anti-tumor immunity while avoiding autoimmunity. To achieve this goal, DCs transfected with mRNA encoding the H and L chains of anti-mouse CTLA-4 and GITR mAbs were co-administered with tumor antigen mRNA-transfected DCs. We observed enhanced induction of anti-tumor immunity and significantly improved survival in melanoma-bearing mice, without signs of autoimmunity. Using in vitro assays with human DCs, we demonstrated that DCs transfected with mRNA encoding a humanized anti-CTLA-4 mAb and mRNA encoding a soluble human GITR-L fusion protein enhance the induction of anti-tumor CTLs in response to DCs transfected with mRNAs encoding either melanoma or breast cancer antigens. Based on these results, this approach of using local delivery of immune modulators to enhance vaccine-induced immunity is currently being evaluated in a phase I clinical cancer immunotherapy trial.

Introduction

APC-based cancer immunotherapy has demonstrated clinical activity 1; however, immune and tumor responses remain modest, emphasizing the need for new strategies to enhance anti-tumor immunity. When APCs such as DCs are exposed to Ag, they process and present Ag-derived peptides to T cells and provide co-stimulatory signals through CD80 and CD86, which engage the T-cell activating receptor CD28. CTLA-4 (CD152), upregulated on T cells for 2–3 days after activation, also binds CD80 and CD86 and mediates immune homeostasis and induction of tolerance to self-antigens 2. The combination of antagonistic anti-CTLA-4 mAbs and vaccination with cytokine-modified melanoma cells has synergistic effects on anti-tumor immunity in mice and humans 3, 4. Several clinical trials have shown that anti-CTLA-4 mAb administration augments clinical anti-tumor responses and, in a recent trial, improves survival in patients with metastatic melanoma 5.

Glucocorticoid-induced TNFR-related protein (GITR), expressed on Treg and activated T cells, also appears to regulate the induction of immune responses 6. Murine in vivo studies using an agonistic mAb to bind GITR have demonstrated enhanced anti-tumor immunity against poorly immunogenic tumors 7. In addition, co-administration of anti-GITR and anti-CTLA-4 mAbs in mice leads to eradication of tumors 7, 8. However, systemic administration of anti-CTLA-4 mAbs to cancer patients has been associated with severe, life-threatening autoimmunity 5, 9. Autoimmunity has also been induced in mice by systemic administration of anti-GITR mAbs.

To overcome these limitations, we have developed a novel approach to target delivery of such immune modulators, using DCs transfected with immune modulator-encoding mRNA, to sites where anti-tumor T cells are induced. We have previously reported that co-administration of DCs transfected with mRNA encoding the H and L chains of an agonistic anti-GITR mAb or soluble GITR-L enhances the induction of anti-tumor immunity in response to vaccination with tumor-associated Ag (TAA) mRNA-transfected DCs in a murine melanoma model and improves survival in tumor-bearing mice, while avoiding the induction of autoimmunity observed with systemic administration of anti-GITR mAb 10.

In our present study, we first evaluated the effect of combined modulation of both CTLA-4 and GITR using DCs transfected with mRNA encoding the H and L chains of anti-CTLA-4 and anti-GITR mAbs in a murine melanoma immunotherapy model. We then generated mRNAs encoding humanized H and L chains of an anti-human CTLA-4 mAb as well as a soluble human GITR-L fusion protein and demonstrated that DCs transfected with these mRNAs secrete functional immune modulating proteins that bind CTLA-4 and GITR, respectively. Finally, we demonstrated the immune-enhancing effect of DCs transfected with these mRNAs in two in vitro human immunotherapy models in which CTLs were induced in response to DCs loaded with either melanoma or breast TAAs using mRNA-transfected DCs.

Results

DCs transfected with immune modulator mRNAs enhance anti-melanoma immunity in vivo

We first evaluated the effect of co-administration of DCs transfected with mRNA encoding an anti-CTLA-4 mAb alone and of co-administration of DCs transfected with mRNA encoding both anti-GITR and anti-CTLA-4 mAbs on the induction of anti-tumor immunity in response to vaccination with DCs transfected with mRNA encoding the TAA TRP-2. As shown in Fig. 1A, in control mice vaccinated with actin mRNA-transfected DCs, the co-administration of anti-GITR and anti-CTLA-4 mAb-encoding mRNA-transfected DCs minimally prolonged survival, suggesting that these immune modulator mRNA-transfected DCs may, to some degree, augment anti-tumor immunity even in the absence of antigen-specific vaccination. In tumor-bearing mice vaccinated with DCs transfected with mRNA encoding TRP-2, survival was significantly prolonged when DCs transfected with either anti-GITR mAb-encoding mRNA or anti-CTLA-4 mAb-encoding mRNA were co-administered. Survival was further, although not significantly, prolonged when a combination of both anti-GITR mAb mRNA-transfected and anti-CTLA-4 mAb mRNA-transfected DCs were co-administered, with 80% of these mice being tumor free at 120 days. No signs of autoimmunity were observed in any vaccinated mice. Importantly, when the tumor diameter was measured as a function of time after tumor implantation, the co-administration of anti-CTLA-4 mAb mRNA-transfected and anti-GITR mAb mRNA-transfected DCs led to a significant reduction in tumor growth compared with co-administration of either immune modulator mRNA-transfected DC population alone (Fig. 1B). These data support the hypothesis that local delivery of immune modulators enhances cancer vaccine efficacy.

Figure 1.

Effect of local immune receptor modulation on anti-tumor immunity in response to tumor-specific immunization in a B16/F10.9 melanoma immunotherapy model. (A) 7–8-wk-old C57BL/6 mice (10/group) were implanted with 2.5×104 B16/F10.9 tumor cells s.c. in the flank region, then vaccinated 2 days later. The figure key depicts the antigen used and the immune receptor targeted in each group. For experiments testing the combination of TRP-2 tumor antigen mRNA-transfected DCs and mAb-encoding mRNA-transfected DCs, mice were immunized s.c. at the base of each ear pinna with 1×105 DCs for each DC group in 50 μL PBS/ear pinna. Total number of DCs per mouse was 4×105 for groups testing two combinations and 6×105 DCs per mouse for groups testing three combinations. Starting on day 10, tumor growth was monitored every other day by measuring the appearance of palpable (5–6 mm diameter) tumors and mice were sacrificed once the tumor diameter reached 20 mm. All surviving mice were tumor-free at day 120 with no evidence of autoimmunity, at which time they were sacrificed. Comparisons between groups at day 70 were performed using the log-rank test (Mantel–Haenszel test), *p<0.05, ***p≤0.001, ns=non-significant. Median survival day: TRP-2+IgG: 37; TRP-2+anti-CTLA-4: 45; actin+anti-GITR+anti-CTLA-4: 25.5 and actin+IgG: 18. All other groups, median: undefined. (B) In these same groups of mice, average tumor size+SEM versus days post-tumor implant is displayed. Tumor growth curves over time were compared using one-way ANOVA for repeated measures (p<0.0001) with Bonferroni multiple comparison post-test to compare the four groups indicated, *p≤0.05, **p≤0.01, ***p≤0.001. (C) 7–8-wk-old C57BL/6 mice (2/group) were implanted with 2.5×104 B16/F10.9 tumor cells and immunized as indicated in (A). Cells were harvested from the draining auricular lymph nodes 10-days post-immunization and the non-adherent cells were stimulated with DCs transfected with TRP-2 mRNA. Induction of TRP-2-specific CTLs was determined 5-days post-restimulation using a standard cytotoxicity assay as described in Materials and methods section. F10.9 and EL4 cells were used as targets. The standard deviation reflects variation between triplicate wells. **p<0.01 as compared with TRP-2+anti-GITR+anti-CTLA-4 group (filled circles) using paired two-tailed Student's t test. (D) 7–8-wk-old C57BL/6 mice (10/group) were implanted with 3.0×104 B16/F10.9 tumor cells s.c. in the flank region. Mice were vaccinated once 7-days after tumor implantation, at which time 3 mm diameter s.c. tumor deposits were palpable in all mice. DCs were transfected with the mRNA combinations as indicated in the figure key. Mice were immunized s.c. at the base of each ear pinna with 1.5×105 DCs in 50 μL PBS/ear pinna. Total number of DCs per mouse was 3×105 for all groups tested. Starting on day 8, tumor growth was monitored and recorded when tumor size was 5 mm and above. Mice were sacrificed once the tumor diameter reached 20 mm (used to determine percent survival). All surviving mice were tumor-free at day 60 with no evidence of autoimmunity, at which time they were sacrificed. Comparisons between groups on day 50 were performed using the log-rank test (Mantel–Haenszel test), ***p≤0.0001. Median survival day: TRP-2+IgG: 21.5; TRP-2+GITR-L: 29; TRP-2+anti-CTLA-4: 29; TRP-2+GITR-L+anti-CTLA-4: 42; actin+GITR-L+anti-CTLA-4: 16 and actin+IgG: 13.

To study the mechanism of the antitumor activity of locally delivered immune modulators, we depleted selected T-cell subsets before immunization. We observed that the protective anti-tumor immune response induced by vaccination with TRP-2 mRNA-transfected DCs was completely abrogated by depletion of either CD4+ or CD8+ T cells (Supporting Information Fig. 1), indicating that both CD4+ and CD8+ T cells are responsible for the antitumor activity of this vaccination strategy.

In mice vaccinated with TRP-2 mRNA-transfected DCs co-administered with DCs transfected with control IgG mRNA, depletion of Treg cells prior to vaccination minimally improved survival (Supporting Information Fig. 1A). However, the co-administration of DCs transfected with either anti-GITR mAb mRNA (Supporting Information Fig. 1B) or anti-CTLA-4 mAb mRNA (Supporting Information Fig. 1C) prolonged survival to a much greater extent than did Treg cell depletion. Furthermore, Treg cell depletion did not add to the effect achieved with anti-GITR or anti-CTLA-4 mAb-secreting DCs.

The CTL activity detected in lymphocytes harvested from lymph nodes draining the sites of DC vaccination correlated with the observed improvement in survival. As shown in Fig. 1C, the highest CTL activity against B16/F10.9 melanoma cells was induced when a combination of anti-GITR mRNA-transfected and anti-CTLA-4 mRNA-transfected DCs was co-administered with DCs transfected with mRNA encoding TRP-2. Importantly, non-specific lytic activity, as assessed using EL4 cells, was not significantly elevated in mice receiving co-administration of DCs transfected with mRNA encoding either or both of these immune modulatory mAbs.

We next evaluated our approach using a more stringent model in which mice were vaccinated 7 days after melanoma implantation. Additionally, in this experiment, DCs were co-transfected with a combination of TAA and immune modulatory mRNAs. To avoid heterologous associations between H and L chains of anti-CTLA-4 and anti-GITR mAbs in the mRNA-transfected DCs, instead of using anti-GITR mAb mRNA, DCs were transfected with mRNA encoding a soluble GITR-L fusion protein. In our prior study, we found that DCs transfected with either anti-GITR mAb mRNA or mRNA encoding soluble GITR-L had equivalent effects 10. As shown in Fig. 1D, a single vaccination with DCs co-transfected with TRP-2 mRNA, anti-CTLA-4 mAb mRNA, and mRNA encoding soluble GITR-L significantly prolonged survival in this stringent model. This prolongation was statistically significant when compared with vaccination with DCs transfected with either TRP-2 plus anti-CTLA-4 mAb mRNAs or TRP-2 plus soluble GITR-L mRNAs.

Human DCs transfected with immune modulator mRNAs secrete immune modulatory proteins

Our next goal was to generate mRNAs encoding immune modulators that could be used for human studies. We therefore cloned the H and L chains of an anti-human CTLA-4 mAb and replaced the murine C regions with those of human IgG. Based on our prior study 10, we constructed a gene encoding a soluble human GITR-L fusion protein.

We then evaluated the time course of immune modulator secretion by human DCs after mRNA transfection. As shown in Fig. 2, human monocyte-derived DCs transfected with mRNA encoding either the H and L chains of humanized anti-CTLA-4 or soluble GITR-L fusion protein secrete proteins that specifically bind to cells expressing their respective targets. In these assays, a sample of culture medium was collected at 6 h, but for all remaining time points, the entire supernatant was collected and replaced with fresh medium. Thus, for both anti-CTLA-4 mAb and soluble GITR-L, most of the secretion of immune modulator occurred within the first 6 h after mRNA transfection, with additional secretion continuing during the period from 24 to 48 h after mRNA transfection.

Figure 2.

Secretion of immune modulators by DCs transfected with mRNA encoding humanized anti-CTLA-4 mAbs or human GITR-L fusion proteins. DCs were transfected with mRNA encoding humanized anti-CTLA-4 mAb or human GITR-L fusion protein. 106 cells were electroporated with 10 μg of anti-CTLA-4-encoding H chain mRNA and 5 μg of L chain mRNA or 10 μg of GITR-L/Fc fusion protein mRNA. The supernatant was harvested at the time points indicated and concentrated by centrifugation through a 50 kDa cut-off filter (65–70-fold concentrated). A supernatant sample was collected at 6 h, then at each subsequent time point, all of the supernatant was harvested and fresh medium containing GMCSF and IL-4 was added to the mRNA-transfected DCs. The concentrate was incubated with CHO cells expressing CTLA-4 (CHO-CTLA-4) or GITR (CHO-GITR). Binding of anti-CTLA-4 mAb and GITR-L fusion proteins in the supernatant to the CHO cells was determined by staining with biotinylated anti-human IgG followed by streptavidin–allophycocyanin. The cells were also stained with a commercially available antibody as a positive control. Cells were analyzed on an FACScalibur flow cytometer. (A) Supernatant from DCs transfected with anti-CTLA-4 mAb-encoding mRNA incubated with CHO-CTLA-4 cells. (B) Supernatant from DCs transfected with anti-CTLA-4 mAb-encoding mRNA incubated with cells that do not express CTLA-4. (C) Supernatant from DCs transfected with GITR-L fusion protein-encoding mRNA incubated with CHO-GITR cells. (D) Supernatant from DCs transfected with GITR-L fusion protein-encoding mRNA incubated with cells that do not express GITR. The results shown are representative of two independent experiments.

Melanoma immunity is enhanced by DCs transfected with mRNA encoding soluble immune modulators

Using cells from HLA-A2+ human donors, we next evaluated the ability of DCs transfected with mRNA encoding soluble GITR-L and anti-CTLA-4 mAb to augment the induction of anti-melanoma CTL in response to DCs transfected with mRNA encoding four defined melanoma TAAs (MART, tyrosinase, MAGE-3, and gp100). As shown in Fig. 3A, DCs transfected with either TAA mRNA or GFP mRNA stimulated the induction of antigen-specific CTL, as assessed by using mRNA-transfected autologous DCs as targets. Co-incubation of autologous T cells with DCs transfected with mRNA encoding either soluble GITR-L fusion protein or anti-CTLA-4 mAb alone enhanced the induction of melanoma TAA-specific CTL activity, while co-incubation with DCs transfected with mRNA encoding both of these immune modulators demonstrated a further enhancement of antigen-specific CTL induction, without an increase in non-specific background CTL activity.

Figure 3.

Effect of DCs transfected with soluble GITR-L fusion protein mRNA and anti-CTLA-4 mAb mRNA on melanoma antigen-specific CTL responses. DCs transfected with TAA mRNA (combination of MART-1, gp100, MAGE-3 and tyrosinase mRNAs) or GFP mRNA, and either control (actin) mRNA, GITR-L mRNA, humanized anti-CTLA-4 mAb (α-CTLA-4) mRNA, or GITR-L mRNA+α-CTLA-4 mRNA were used to stimulate autologous T cells weekly twice followed by a CTL assay. Induction of TAA-specific CTLs was measured using mRNA-transfected DCs and tumor targets. This experiment was performed three times, using cells from different donors, each with similar results. Data show mean±SD. (A) DC targets: Targets were autologous DCs transfected with mRNA encoding TAAs MART-1, gp100, MAGE-3 and tyrosinase (DC-TAA) or transfected with GFP (DC-GFP). (B) Tumor cell targets: Because the donor for these experiments was HLA-A2+, CTL activity was assessed against the HLA-A2+ melanoma cell lines DM6 and A375, and the HLA-A2+ tumor cell line 293 that does not express TAAs MART-1, gp100, MAGE-3 or tyrosinase.

CTL activity against melanoma cells was also assessed. As shown in Fig. 3B, co-incubation with DCs transfected with mRNA encoding both GITR-L and anti-CTLA-4 mAb markedly enhanced the induction of melanoma TAA-specific CTL activity against DM6, an HLA-A2+ melanoma cell line which expresses all four melanoma TAAs used in these experiments, and against A375, an HLA-A2+ melanoma cell line which expresses only MAGE-3. No lytic activity was observed against the HLA-A2+ 293 cell line which does not express any of the four melanoma TAAs, suggesting that non-specific autoimmune responses were not induced in these in vitro experiments.

Breast cancer immunity is enhanced by DCs secreting soluble immune modulators

Again using cells from HLA-A2+ human donors in vitro, we evaluated the ability of DCs transfected with immune modulator mRNAs to augment the induction of antigen-specific CTL in response to DCs transfected with mRNA encoding four defined breast cancer TAAs (CEA, MUC1, MAGE-3, and HER2/neu). As shown in Fig. 4A, DCs transfected with a combination of mRNAs encoding all four breast cancer TAAs stimulated the induction of antigen-specific CTL activity against each of the four TAAs, as assessed using autologous DCs transfected with individual breast cancer TAA-encoding mRNAs as targets. Co-incubation with DCs transfected with mRNA encoding either GITR-L fusion protein or anti-CTLA-4 mAb alone enhanced the induction of antigen-specific CTL activity, while co-incubation with DCs transfected with mRNA encoding both of these immune modulators demonstrated an additive enhancement of tumor antigen-specific CTL induction, without an increase in non-specific background CTL activity.

Figure 4.

Effect of DCs transfected with soluble GITR-L protein mRNA and anti-CTLA-4 mAb mRNA on breast tumor antigen-specific CTL responses. DCs transfected with TAA mRNA (combination of CEA, MUC-1, MAGE-3 and HER2/neu mRNAs) or GFP mRNA, and either control (actin) mRNA, GITR-L mRNA, humanized anti-CTLA-4 mAb (α-CTLA-4) mRNA, or GITR-L mRNA+α-CTLA-4 mRNA were used to stimulate autologous T cells weekly twice followed by a CTL assay. Induction of TAA-specific CTLs was measured using DCs and tumor targets. This experiment was performed twice, using cells from different donors, both with similar results. Data show mean±SD. (A) Targets are DCs transfected with mRNA encoding the breast TAAs, HER2/neu, MUC-1, MAGE-3 or CEA. (B) Because the donors for these experiments were HLA-A2+, CTL activity was assessed against the HLA-A2+ breast cancer cell line MCF7 and MDA-231, the HLA-A2 breast cancer cell line T47D, and the HLA-A2+ tumor cell line 293 that does not express TAAs CEA, MUC-1, MAGE-3 or HER2/neu.

CTL activity against cultured breast cancer cells was also assessed. As shown in Fig. 4B, co-incubation with DCs transfected with mRNA encoding both GITR-L and anti-CTLA-4 mAb further enhanced the induction of CTL activity against HLA-A2+ MCF-7 cells (which express CEA, HER2/neu and MUC1), while stimulating no lytic activity against the HLA-A2+ 293 cell line (that does not express any of the four breast cancer TAAs used in this study), suggesting that non-specific autoimmune responses were not induced. In this experiment, activity against HLA-A2 T47D breast cancer cells (which express breast TAAs MAGE-3 and MUC1), may be due to TAA-derived peptides expressed in the context of non-A2 class I molecules possibly shared by the cell donor and the T47D cells, as expression of HLA alleles other than A2 was not assessed in these experiments.

Engineered DCs inhibit Treg induction and Treg-mediated suppression

We next assessed the ability of human monocyte-derived DCs transfected with mRNA encoding these two immune modulators to inhibit the induction of Treg cells in vitro. As shown in Fig. 5A, the number of CD25+CD127low Treg cells induced when naïve CD4+ CD25 T cells were stimulated with cytokine and PGE2-matured autologous DCs was markedly reduced when the DCs were transfected with mRNA encoding both anti-CTLA-4 mAb and soluble GITR-L. DCs transfected with mRNA encoding either immune modulator alone did not significantly reduce the induction of CD25+CD127low Treg cells (data not shown). Similar results were obtained using cells from two donors when flow cytometric staining for CD25 and Foxp3 was used to assess Treg cell induction (data not shown).

Figure 5.

Effect of DCs transfected with mRNA encoding anti-CTLA-4 mAb and soluble GITR-L protein on the induction of Treg cells in DC-T-cell co-cultures and the suppressive function of induced CD4+CD25+ cells. (A) CD4+CD25 T cells were isolated as described in Materials and methods section and cultured with mRNA-transfected matured DCs. After 10 days of culture, the induction of CD25+CD127low Treg cells was assessed by flow cytometry. This experiment was performed using cells from two donors, as indicated. (B) CD4+CD25+ T cells isolated from co-cultures of CD4+CD25 T cells with autologous DCs transfected with the indicated mRNAs were co-cultured with autologous CD4+CD25 T cells in the presence of anti-CD3 and anti-CD28 mAbs. After 4 days, cell proliferation was assessed using the WST-1 reagent, with absorbance at 450 nm reflecting the proliferative response. This experiment was performed using cells from two different donors, as indicated. Data show mean±SD. Comparisons were made using Student's t-test, *p<0.05, **p<0.01.

We then isolated these induced CD4+CD25+ T cells and determined their suppressive function in proliferation assays. These isolated CD4+CD25+ T cells were cultured with autologous CD4+CD25 cells and proliferation was determined after 4 days. As can be seen in Fig. 5B, CD4+CD25+ T cells induced by DCs transfected with mRNA encoding both anti-CTLA-4 mAb and GITR-L exhibited a significantly reduced capacity to inhibit CD4+CD25 T-cell proliferation compared with CD4+CD25+ T cells that had been induced in the presence of control mRNA-transfected DCs.

Discussion

Cancer immunotherapy is now demonstrating promising clinical activity 1; however, major improvements in clinical outcome will likely require strategies to target tumor-induced and normal host immune-modulatory mechanisms that tend to dampen the induction of T- and B-cell responses capable of eradicating established malignancies. Among the candidate target molecules are GITR and CTLA-4.

GITR is constitutively expressed on Treg cells, is expressed at low levels on multiple immune cells including naïve T cells, and is upregulated upon activation 6, 11. Treatment of mice with an agonistic anti-GITR mAb alone enhances anti-tumor immune responses 7, 12, 13. When combined with immunization against melanoma-related antigens, we and others have found that agonistic anti-GITR mAb-treated mice demonstrated increased tumor antigen-specific T-cell responses and protective tumor immunity 10, 14. These data suggest that systemic treatment with agonistic anti-GITR mAb, alone or as an adjunct to active immunization, increases the ability to overcome immune tolerance.

CTLA-4, like GITR, is constitutively expressed at high levels on Treg cells, and is also up-regulated on activated effector T cells. CTLA-4 is a negative regulator of T-cell activation which suppresses T-cell activity by binding the B7 molecules CD80 and CD86, and CTLA-4 blockade results in improved tumor immunity and tumor regression in several mouse models 3, 15. In clinical studies, treatment with antagonistic CTLA-4-specific mAb induced objective cancer regression in some patients vaccinated with a gp100-derived peptide 16. In a more recent study in patients with metastatic melanoma, regardless of immunization, treatment with anti-CTLA-4 mAb was found to significantly prolong survival 5.

A major concern with systemically administered immunomodulatory molecules is the induction of autoimmunity. Treatment with anti-GITR mAb stimulated autoimmunity in mice 12, 13, an effect we also observed in our previous study 10. Clinically, administration of the anti-CTLA-4 mAb ipilimumab has induced immune-related adverse events, including severe colitis that in some cases has resulted in death 5, 16, 17.

Importantly, in our current study using mRNA-transfected DCs to locally deliver these immune modulators, anti-tumor immunity was enhanced, but autoimmunity was not observed in mice in vivo. Furthermore, no increase in non-specific background immune responses against control target cells using human cells in vitro was detected (Figs. 3 and 4). Based on our results, we anticipate that local delivery of anti-CTLA-4 mAb and soluble GITR-L using mRNA-transfected DCs will circumvent the autoimmune complications associated with systemically administered mAbs, while still augmenting vaccine-induced anti-tumor immune responses in human patients.

Another advantage of the local release of the anti-CTLA-4 mAbs and GITR-stimulating molecules provided through mRNA transfection of DCs is the relatively short duration of their secretion at the key location where they are needed, at the site of T-cell activation. Previous studies have demonstrated that mRNA has a half-life of <4 h after transfection into DCs, suggesting that translation of the mRNA into protein is a transient event in the DCs. In our current study, we confirmed that for both soluble GITR-L and anti-CTLA-4 mAb, most of the secreted proteins were released within 6 h of mRNA transfection, with additional release occurring for up to 24–48 h (Fig. 2). Since human CD4+CD25+ Treg cells, as well as CD4+CD25 T cells, have been shown to up-regulate expression of CTLA-4 within 4 h of activation, this time course of DC secretion of anti-CTLA-4 mAb that we observed should be ideal for blocking CTLA-4 expressed by Treg and CD4+ T cells 18.

An important observation of our study is that local GITR stimulation and CTLA-4 inhibition combined to enhance the induction of immune responses; however, the mechanism for this additive effect remains to be fully elucidated. Cohen et al. observed that anti-GITR mAb treatment led to a significant reduction of tumor infiltrating Foxp3+ Treg cells, without globally suppressing Treg cell activity 14. Recent studies indicate that the enhanced immune response is due to ligation of GITR on the effector cells, making them resistant to Treg cell-mediated suppression, rather than to a direct negative regulatory effect on Treg cell function 19, 20. Anti-GITR mAb treatment may also be augmenting anti-tumor immunity in response to vaccination by directly co-stimulating effector T cells 19.

Tuyaerts et al., transfected human DCs with mRNA encoding GITR-L, generating DCs that expressed high levels of GITR-L on the cell surface. Such GITR-L-expressing DCs did not down-modulate Treg cell-mediated suppression, but did demonstrate enhanced T-cell co-stimulation and enhanced the induction of melanoma antigen-specific CD8+ T cells 20. In our in vitro human experiments, DCs transfected with mRNA encoding soluble GITR-L alone did not suppress the induction of Treg cells (data not shown). Only when DCs were transfected with mRNAs encoding both GITR-L and anti-CTLA-4 mAb did they suppress the induction of Treg cells (Fig. 5A), and such DCs also reduced the suppressive effects of Treg cells on CD4+ T-cell proliferation (Fig. 5B).

Regarding the effects of anti-CTLA-4, using a murine B16 melanoma model, Sutmullter et al. found that the immune response to vaccination as well as survival was further improved when anti-CTLA-4 mAb was administered along with depletion of Treg cells, suggesting that anti-CTLA-4 mAb does not primarily function by modulating CTLA-4 on Treg cells 21. Ribas and colleagues evaluated melanoma biopsies from patients treated systemically with anti-CTLA-4 mAb and noted large numbers of CD8+ CTL in regressing lesions, but did not observe a significant effect on infiltration by Treg cells 22. Similarly, in patients treated at the NCI with Ipilimumab, no effect on peripheral blood Treg cell numbers or activity either in vitro or in vivo was detectable 23, suggesting that the anti-tumor immune effects of anti-CTLA-4 mAb are not mediated directly by inhibition or depletion of Treg cells.

In our murine studies, the improved survival seen in TRP-2 mRNA-transfected DC-vaccinated mice when these mice were co-vaccinated with either anti-GITR mAb or anti-CTLA-4 mAb mRNA-transfected DC was superior to that seen with Treg cell depletion and was not further enhanced by Treg cell depletion (Supporting Information Fig. 1). This suggests that local delivery of these immune modulators not only overcomes Treg cell-mediated negative immune regulatory effects but also has additional effects on the effector T cells responsible for enhanced tumor immunity.

Mitsui et al. attributed the synergistic effect of anti-GITR and anti-CTLA-4 mAb co-administration to: i) anti-CTLA-4 mediated increased CD8+ T-cell infiltration into tumors and ii) anti-GITR mediated increase in cytokine secretion and Treg cell resistance 8. We observed additive effects in both the B16 mouse melanoma model in vivo (Fig. 1) and in human cells in vitro (Figs. 3 and 4) using DCs transfected with mRNA encoding immune modulators targeting both GITR and CTLA-4.

Based on the results presented in this report, we have initiated a phase I clinical trial in subjects with metastatic melanoma (ClinicalTrials.gov ID NCT01216436). In this study, subjects are vaccinated with an intranodal injection of DCs transfected with mRNA encoding melanoma TAAs MART, MAGE-3, gp100 and tyrosinase, and co-transfected with mRNA encoding soluble human GITR-L and/or anti-CTLA-4 mAb. Since human studies have shown that only a small percentage (1–2%) of DCs migrate to the draining lymph nodes after intradermal injection, and that this migration occurs over a period of several days, our DCs transfected with mRNA encoding soluble GITR-L and/or anti-CTLA-4 mAb would be unlikely to secrete these proteins by the time they migrated to the draining lymph nodes (see Fig. 2). To overcome this limitation, in our phase I clinical trial we are injecting the DCs directly into inguinal and axillary nodes under ultrasound guidance. We anticipate that this novel clinical approach will be safe and that the co-administration of DCs transfected with mRNA encoding soluble immune modulators targeting GITR and CTLA-4 will enhance anti-tumor immunity in response to vaccination, while avoiding the induction of autoimmunity seen with systemic administration of such immune modulators.

Materials and methods

mRNAs

The genes encoding murine melanoma antigen TRP-2, enhanced GFP and murine actin have been inserted into the pSP73-Sph/A64 plasmid, as described 10, 24, 25. The genes encoding human melanoma TAAs MART, tyrosinase, MAGE-3 and gp100 have been cloned into pcDNA3.1-64A, as described 26. Human HER2/neu was cloned using PCR and template cDNA synthesized from RNA isolated from the MCF-7 breast cancer cell line and ligated into pcDNA3.1-64A. The gene for MUC1 (kindly provided by Olivera Finn, University of Pittsburgh) was subcloned into pSP73-Sph/A64. CEACAM 6 was cloned using PCR and template cDNA synthesized from SW403 (ATCC) RNA, then inserted into plasmid pSP73-Sph/A64.

Cloning of the H and L chains of control rat IgG and the anti-mouse GITR mAb DTA-1, as well as a soluble mouse GITR-L fusion protein have been previously described 10. Similar methods were used to clone the genes encoding the H and L chains of anti-CTLA-4 mAb 9H10. Humanized H and L chains of the anti-CTLA-4 hybridoma A3.6B10 (ATCC, Manassas, VA, USA) were cloned into pSP73-Sph/A64 using standard techniques. A human GITR-L human Fc fusion protein was cloned and inserted into pSP73-Sph/A64 using techniques similar to those previously described for cloning of a murine fusion protein 10. Human CTLA-4 and GITR were cloned into pSP73-Sph/A64 using standard techniques. Refer to Supporting Information Material and Methods for cloning details.

For mRNA transcription, plasmids were digested with SpeI for use as a template for in vitro transcription reactions using the mMESSAGE mMACHINE T7 kit (Ambion, Austin, TX, USA) according to the manufacturer's protocol. mRNA was purified with the RNeasy mini kit.

Validation of human immune modulator secretion after mRNA transfection

Mature DCs were electroporated with anti-CTLA-4 H and L chain mRNA or GITR-L mRNA, then supernatants were harvested at various time points and concentrated. CHO cells were electroporated with mRNA encoding either human chimeric CTLA-4/CD28 or human GITR, then incubated at 37°C for 24 h. Cell-surface expression of CTLA-4 or GITR by these CHO cells was confirmed using commercially available allophycocyanin-labeled anti-CTLA-4 mAb (BD Pharmingen, San Jose, CA, USA) or anti-GITR mAb (R&D, Minneapolis, MN, USA). These CHO-CTLA-4 cells or CHO-GITR cells were incubated with the supernatants and bound anti-CTLA-4 mAb or GITR-L, respectively, was detected using biotinylated anti-human Ig (eBioscience, San Diego, CA, USA), streptavidin-allophycocyanin (eBioscience), and a FACScalibur flow cytometer (BD Biosciences, San Jose, CA, USA). Refer to Supporting Information Material and Methods for details.

Mice

Four to six wk-old C57BL/6 mice (H-2b) were obtained from the Jackson Laboratory (Bar Harbor, ME, USA). The investigators adhered to the “Guide for the Care and Use of Laboratory Animals” as proposed by the committee on care of Laboratory Animal Resources Commission on Life Sciences, National Research Council. The facilities at the Duke Vivarium are fully accredited by the American Association for Accreditation of Laboratory Animal Care, and all studies were approved by the Duke University Institutional Animal Care and Use Committee.

Cell lines and reagents

Cell lines used included the F10.9 clone of B16 melanoma (B16/F10.9), DTA-1 cells (anti-murine GITR mAb secreting hybridoma cells, kindly provided by Dr. Shimon Sakaguchi), CHO cells and EL4 thymoma cells. Cells were maintained in DMEM supplemented with 10% FCS, 25 mM HEPES, 2 mM L-glutamine and 1 mM sodium pyruvate.

Generation and electroporation of murine DCs

DCs were generated from bone marrow precursors and electroporated with mRNA as described previously 10. Transfected DCs were matured for 5 h in GM-CSF+IL-4 medium supplemented with LPS (Sigma L265L, E. coli 026:B6) at 100 ng/mL. Cells were harvested and washed twice prior to use.

B16/F10.9 melanoma tumor immunotherapy model

C57BL/6 mice were implanted with 2.5–3×104 B16/F10.9 tumor cells s.c. in the flank region. Mice were vaccinated once either 2 days or 7 days after tumor implantation. Tumor-bearing mice were immunized s.c. at the base of the ear pinna with 1–1.5×105 mRNA-transfected DCs/ear pinna in 50 μL PBS for a total of 100 μL/mouse. For experiments testing the combination of tumor antigen mRNA-transfected DCs and mAb-encoding mRNA-transfected DCs, mice were immunized with 1–1.5×105 DCs for each group for a combined 2–3×105 DCs/ear pinna in 50 μL PBS, for a total of 100 μL/mouse. Tumor growth was evaluated every other day starting on day 10. Mice were sacrificed once the tumor size reached 20 mm. Tumor-free mice were sacrificed after 65–70 days or as indicated in the figure legends.

CTL induction in vivo

Cells were harvested from the draining auricular LN of mice vaccinated as above (day 2) 7–10 days after immunization followed by adherence for 1 h. Ten million non-adherent lymphocytes were cultured with 2×105 stimulator cells (DCs electroporated with TRP-2-encoding mRNA) in 5 mL of RPMI +10% FCS per well in a 6-well tissue culture plate. Effector T cells were harvested after 5 days followed by a standard europium-release in vitro cytotoxicity assay, as previously described 10. See Supporting Information Materials and methods for more details.

Statistical analysis

For tumor studies, comparison between two groups was performed using the log-rank test (Mantel–Haenszel test). Tumor growth curves over time were compared using one-way ANOVA for repeated measures with Bonferroni multiple comparison post-test. Statistical significance in immune assays comparing two groups was done using paired two-tailed Student's t test. A probability of <0.05 (p<0.05) was considered statistically significant.

Human DC generation

Cellular material used in these experiments was obtained from human subjects following informed consent using protocols approved by the Duke University Investigational Review Board. DCs were generated from monocytes isolated by plastic adherence from PBMC collected by leukapheresis, as described previously 26.

DC maturation

RNA-electroporated DCs were matured for 6-8 h in AIM-V media containing GM-CSF (800 U/mL), IL-4 (500 U/mL), TNF-α (10 ng/mL), IL-1β (10 ng/mL), IL-6 (1000 U/mL), and PGE2 (1 μg/mL). All cytokines were obtained from Peprotech. PGE2 was purchased from Sigma.

In vitro stimulation of T cells with mRNA-transfected DCs

T cells, isolated from thawed PBMC, were stimulated with mRNA-transfected autologous DC, as previously described 26. Totally, 6–7 days after a single re-stimulation, T cells were harvested, counted and used as effector T cells in a europium-release CTL assay 10. Autologous DCs transfected with TAA-encoding mRNA and tumor cells were used as targets.

Assessment of Treg cell induction

The induction of Treg cells by mature DCs was assessed using the methods described by Banerjee, et al 27. Refer to Supporting Information Material and Methods for details.

Assessment of Treg cell-mediated suppression

Autologous CD4+CD25+ cells induced during autologous DC/CD4+CD25 T-cell co-cultures (see above) were isolated. The inhibitory effect of these induced CD4+CD25+ cells on the proliferation of CD4+CD25 cells in response to anti-CD3 (OKT-3) and anti-CD28 (CD28.2) mAb stimulation was then assessed. Refer to Supporting Information Material and methods for details.

Acknowledgements

We thank David Snyder for his expert technical assistance with the murine immunotherapy studies. This project was partially supported by grant UL1RR024128 from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH), and NIH Roadmap for Medical Research (SN and DB), and by VA Merit Review grants (SKP, DST).

Conflict of interest: The authors have no financial or commercial conflict of interest.

Ancillary