Purpose: The use of costimulatory molecules targeting distinct T-cell signaling pathways has provided a means for triggering and enhancing antitumor immunity; however, it is still not fully understood what types of costimulatory molecules are suitable for the combination in tumor therapy. Our purpose in this study is to establish an effective antitumor immune approach by using costimulatory molecule 4-1BBL in combination with soluble PD-1.
Experimental Design: The murine H22 hepatocarcinoma served as an ectopic tumor model. Local gene transfer was done by injection with naked plasmid p4-1BBL and/or psPD-1. The synergistic mechanism of dual-gene therapy was elucidated by detecting the change of gene expression of immunoregulatory factors in tumor microenvironment. The effects of immunotherapy were evaluated by testing the function of tumor-specific T cells, measuring tumor weight or volume, survival of mice, and H&E staining of tissues.
Results: 4-1BBL expressed by normal nonimmune cells effectively enhanced antitumor immune response but up-regulated PD-L1 and did not reduce IL-10 and transforming growth factor-β (TGF-β). sPD-1 synergized with 4-1BBL to establish efficient antitumor immune environment, including down-regulation of IL-10 and TGF-β, further up-regulation of interleukin (IL)-2 and IFN-γ, and higher CD8+ T-cell infiltration. The combined treatment by 4-1BBL/sPD-1 eradicated tumors from mice with small amounts of preexistent tumor cells or tumors from ∼60% of individuals with larger amounts of preexistent tumor cells.
Conclusions: Our findings in this report imply a great potential of 4-1BBL in combination with sPD-1 in tumor therapeutics with the in vivo existent tumor cells as antigens.
- Tumor Immunotherapy
- Gene therapy
- soluble PD-1
The application of defined tumor antigens as therapeutic cancer vaccines has not yet resulted in a significant efficacy in cancer patients (1–4), which is partly explained by the existence of other undefined epitopes of tumor antigens in many cases (5) so that the therapeutic tumor vaccines may not cover all of residual tumor cells after different treatments. Therefore, it might be more valuable to use the in vivo existent tumor cells as the repertoire of tumor antigens for the generation of the pool of tumor-specific T cells to eradicate the small amount of residual tumor cells in vivo. The obstacle for this strategy is that the immunogenicities of self/tumor antigens on tumor cells in vivo are too weak to effectively induce the activation of tumor-specific T cells (5). However, the utilization of costimulatory molecules targeting distinct T-cell signaling pathways provides a means to trigger an effective antitumor immune response with tumor cells as antigens.
4-1BB, an attractive costimulatory molecule for tumor immunotherapy, is expressed on activated T cells and antigen-presenting cells such as dendritic cells (6–9). Therapies by utilizing the 4-1BBL/4-1BB signaling pathway have been shown to generate antitumor effects in various model systems (10–13). Recently, 4-1BBL and anti–4-1BB antibody have been used in tumor therapy in combination with other therapeutic elements such as cytokine interleukin (IL)-12, tumor vaccines, radiation therapy, and anti–CTLA-4 antibody (6, 13–16). The activation of 4-1BB signaling usually results in the production of a large amount of IFN-γ, which is capable of inducing or up-regulating the expression of PD-L1 (also B7-H1) on both immune cells and tumor cells (17–19). As coinhibitory molecules, PD-L1 and its receptor PD-1 mediate the inhibition of T-cell response (20, 21) and thus play an important role in down-regulating the antitumor immunity at effector phase. These findings imply a possibility of further improvement of antitumor effect of 4-1BBL by reducing the related immunosuppressive elements.
Blockade of the PD-L1/PD-1 pathway with antibody or soluble PD-1 (sPD-1) has been proved to enhance antitumor immunity (22–24) and inhibit tumor growth (18, 25). Our previous study has also confirmed that sPD-1 can reduce IFN-γ–mediated immunosuppressive effect (26). In this study, we tried to use sPD-1 to counteract the negative immunoregulatory elements existent in the process of tumor immunotherapy with 4-1BBL. 4-1BBL alone effectively enhanced antitumor immune response but up-regulated PD-L1 and did not reduce IL-10 and TGF-β in tumor microenvironment. Such disadvantages were compensated by sPD-1. In addition, a more ideal curative effect on tumor-bearing mice was realized by using 4-1BBL in combination with sPD-1. These findings verify that the modulation of 4-1BBL treatment with sPD-1 can trigger robust antitumor immunity with in vivo existent tumor cells as antigens.
Materials and Methods
Animals and cell lines. Six- to 8-week-old female BALB/c mice were purchased from Center of Medical Experimental Animals of Hubei Province (Wuhan, China). The animals were maintained in Tongji Medical Animal Facilities under pathogen-free conditions. All studies involving mice (including tumor cell inoculation and tumor weight in experiments) were approved by the institute's Animal Care and Use Committee. Baby hamster kidney (BHK) cell line, mouse H22 hepatocarcinoma cell line (BALB/c background), and mouse melanoma B16F1 cell line (C57BL/6 background) were purchased from China Center for Type Culture Collection (Wuhan, China) and cultured in DMEM with 10% FCS.
Plasmids. Eukaryotic expression vector p4-1BBL carrying full-length cDNA of murine 4-1BBL, psPD-1 carrying the cDNA encoding extracellular domain of murine PD-1 (sPD-1; ref. 25), and pGRA carrying the cDNA encoding extracellular domain of mouse IFN-γ receptor α-chain (26) were constructed by insertion of cDNA into plasmid pcDNA3.1 (Invitrogen, Carlsbad, CA) and were preserved in our laboratory.
Gene transfection in vitro and in vivo. Plasmids were prepared and analyzed as previously described (23, 25). The in vitro transfection of BHK cells was done with FuGENE 6 reagent (Roche Diagnostics, Indianapolis, IN) following the manufacturer's instructions. G418 (500 μg/mL) was used to select the stable transfection clones.
The in vivo transfection was done by direct local injection. Briefly, naked plasmid (100 μg in 100-μL saline) was injected directly into muscle (i.m. injection) of inoculation site. The plasmids were injected every 3 days because the mRNA expression by the vector decreased 3 days after i.m. injection of plasmids as observed in the preliminary experiments and our previous studies.
Conventional reverse transcription-PCR and real-time PCR. Total RNAs were isolated, using TRIzol reagent (Invitrogen) according to the manufacturer's instructions, from cells or muscle tissues of normal mice or tumor marginal tissues of tumor-bearing mice. A reverse transcription-PCR (RT-PCR) procedure was used to determine relative quantities of mRNA (OneStep RT-PCR kit, Qiagen, Valencia, CA). Twenty-eight PCR cycles were used for all of the analyses and the mRNA of β-actin was used as the internal control. The sequences of primers for detection of 4-1BBL mRNA were as follows: sense, 5′-CTGTGTTCCTATCTTCACCC-3′; antisense, 5′-TGTCTTCTTCGTACCTCAG-3′. The primers for other genes were previously described (25).
Quantitative real-time PCR for PD-L1, IL-2, IFN-γ, IL-10, and TGF-β1 genes was done using primers and procedures previously described (23, 25). The results were expressed as the expression level of each gene relative to that of housekeeping gene β-actin.
Western blot analysis. Seventy-two hours after transfection, cells and tissue samples were lysed or homogenized for Western blot as described (27). Rat anti-mouse 4-1BBL monoclonal antibody (TKS-1) was kindly provided by Prof. Hideo Yagita (Juntendo University, Tokyo, Japan). Antimouse PD-1 polyclonal antibody was purchased from R&D Systems (Minneapolis, MN).
Proliferation assay. BHK cells transfected with p4-1BBL, pcDNA3.1, or without transfection were cultured and fixed in 96-well culture plates with glutaric dialdehyde. Mouse splenocytes were cultured in wells and stimulated with HSP70-H22 peptide complexes (23, 25). To determine the proliferation of T cells, 1.0-μCi [3H]thymidine was added during the last 10 h of 72-h culture and then the incorporation of [3H]thymidine was measured in a MicroBeta TriLux liquid scintillation counter (Wallac, Turku, Finland).
ELISA. Mouse spleen cells were stimulated in the same way as mentioned earlier. The cells were passaged 72 h later and cultured for another 48 h. The levels of IFN-γ and IL-2 in the supernatants were assessed by ELISA using murine IFN-γ or IL-2 ELISA Kit (eBioscience, San Diego, CA) according to the manufacturer's protocol.
Cytotoxicity assay. Splenocytes were stimulated with HSP70-H22 peptide complexes in the presence or absence of BHK-expressed 4-1BBL, as described above, and cultured for 7 days in the presence of 20 units/mL IL-2. Then, the cells were used as effector cells for cytotoxicity assay. In other experiments, splenocytes from tumor-bearing mice were individually stimulated with HSP70-H22 peptide complexes in vitro for 3 days and used as effector cells. The flow cytometry–based method (27) was used for cytotoxicity assay. Briefly, carboxyfluorescein diacetate succinamidyl ester (CFSE)–labeled H22 and B16F1 target cells were incubated with effectors at different effector-to-target ratios at 37°C for 4 h. After staining with dye 7-amino-actinomycin D (7-AAD) to mark dead or dying cells, cytotoxicity was evaluated by flow cytometry. Cytolysis was determined by 7-AAD+CFSE+ cells / total CFSE+ cells.
Animal experiments and treatment protocol. BALB/c mice were inoculated with H22 cells by injection of 1 × 105 cells into right hind thigh muscle. On day 2 after inoculation, mice from treatment groups received 100 μg of plasmid DNA by i.m. injection (local naked DNA transfection). Fifty micrograms of p4-1BBL and 50 μg of psPD-1 were used in combination treatment. Gene delivery was done every 3 days for the indicated times in different experiments. The mice of control groups received an equal volume of saline or equal amount of pcDNA3.1 plasmid. In some experiments, HSP70-H22 peptide complexes (25) were used as tumor vaccine and injected together with plasmids as indicated. Mice were sacrificed and tumors were dissected and weighed on the indicated day after inoculation. In some experiments, the incidence of tumor in mice and the survival of mice were recorded.
Histology and immunohistochemistry. Mouse muscle tissues of inoculation sites of the treated mice were surgically excised, fixed for 12 to 24 h in 4% formalin, embedded in paraffin, and sectioned for H&E staining. For indirect immunostaining, fresh tissues were embedded in optimum cutting temperature solution and cut into 10-μm sections. Sections were fixed by acetone and incubated overnight at 4°C with rat anti-mouse CD8 monoclonal antibody diluted at 1/100. Biotinylated rat anti-mouse immunoglobulin G was used as secondary antibody, followed by streptavidin-conjugated horseradish peroxidase in the third step. The antibodies were purchased from Abcam, Inc. (Cambridge, MA).
Statistics. Results were expressed as mean ± SD and interpreted by ANOVA repeated measures test and Kaplan-Meier analysis. Differences were considered to be statistically significant when P < 0.05.
Normal nonimmune cell–expressed 4-1BBL possesses antitumor immune function. Agonistic 4-1BB antibody is generally used to trigger 4-1BB signal in tumor immunity. We constructed a 4-1BBL–expressing vector (p4-1BBL) as an alternative strategy to activate 4-1BB signal. BHK cells were transfected with p4-1BBL and expression of p4-1BBL was identified by RT-PCR and Western blot (Fig. 1A ). Thereby, mouse splenocytes were stimulated with HSP70-H22 peptide complexes in the presence or absence of BHK-expressed 4-1BBL. The results showed that both the cell proliferation (Fig. 1B) and the production of IL-2 and IFN-γ (Fig. 1C) were significantly increased in the presence of BHK-expressed 4-1BBL. When the splenocytes were stimulated with HSP70-H22 peptide complexes for 7 days, BHK-expressed 4-1BBL also significantly enhanced the specific cytotoxicity of the activated lymphocytes to H22 tumor cells but not to melanoma B16 cells (Fig. 1D). To validate the in vitro results in vivo, we then, on the basis of identifying the expression of p4-1BBL in muscle (Fig. 1A), treated H22 tumor-bearing mice by i.m. injection with p4-1BBL. The significant inhibition of tumor growth was observed at different time points in p4-1BBL group (Fig. 1E). Taken together, these results indicate that 4-1BBL expressed by normal nonimmune cells can trigger antitumor immune response and inhibit tumor growth; however, the effect of 4-1BBL/4-1BB signal alone is limited and not enough to eradicate tumor in vivo.
The immune response profile is improved by using 4-1BBL in combination with sPD-1. One of the possibilities for the limitation of 4-1BBL–mediated antitumor effect is the up-regulation of PD-L1 by 4-1BBL–promoted production of IFN-γ. To verify this, we first checked the effect of 4-1BBL on PD-L1 by analyzing the transcription activity of PD-L1 gene in tumor tissues after treatment by i.m. transfection with p4-1BBL or in combination with plasmid pGRA, the expression vector of a soluble IFN-γ receptor. The treatment with p4-1BBL significantly increased the transcription activities of both IFN-γ and PD-L1 genes, whereas the transcription of PD-L1 gene was significantly reduced by soluble IFN-γ receptor (Fig. 2A ). The results confirmed that the up-regulation of PD-L1 by 4-1BBL is mediated by IFN-γ.
Because the blockade of IFN-γ activity in the process of tumor immunotherapy is unreasonable, and we have found that the effective blockade of PD-L1/PD-1 pathway can be realized by the utilization of sPD-1 (23–26), here we hypothesized that sPD-1 could be used to improve 4-1BBL–mediated antitumor immune response. To test this, we first identified the expression of 4-1BBL and sPD-1 after dual-gene delivery (Fig. 2B) and then analyzed the transcription activity of PD-L1 gene in tumor tissue after different treatments. The results showed that the transcription activity of PD-L1 gene in tumor tissue was increased (Fig. 2C) even in the presence of sPD-1. The transcription activities of IFN-γ and IL-2 genes were further increased, whereas the transcription activities of IL-10 and TGF-β genes were significantly decreased, which were not decreased by 4-1BBL alone (Fig. 2D). The modulation of IL-10 and TGF-β in the tumor microenvironment is obviously due to the presence of sPD-1, not 4-1BBL. These data indicate that the combination of 4-1BBL and sPD-1 can remodulate tumor microenvironment and make it bias to antitumor immunity.
sPD-1 enhances 4-1BBL–mediated antitumor immunity. The above data revealed that the immune response profile can be improved if 4-1BBL is used in combination with sPD-1. We then analyzed whether this biased antitumor immunity could generate stronger cytotoxicity to tumor cells and result in the accumulation of more lymphocytes in tumor microenvironment. The cytotoxicity of splenocytes to tumor cells was analyzed on day 14 after tumor inoculation. The results showed that the cytotoxicity of splenocytes to H22 tumor cells (not B16 melanoma cells) was further increased in 4-1BBL/sPD-1 combination group, compared with that of single-gene (4-1BBL or sPD-1) treated group (Fig. 3A ). Immunohistochemical studies on day 28 after tumor inoculation revealed a significant difference in the amount of tumor-infiltrating T cells between control and treated mice. The amount of CD8+ cells in the tumor marginal tissue of 4-1BBL/sPD-1 combination group was significantly increased as compared with that of single-gene treated groups (Fig. 3B). The curative effect was also observed at different time points. Tumor growth in each treatment group was significantly suppressed compared with that in pcDNA3.1 control group (Fig. 3C), and the suppression of tumor growth by 4-1BBL/sPD-1 treatment was much stronger than that by single-gene treatment, indicating a synergistic antitumor effect of 4-1BBL and sPD-1.
Tumor vaccine cannot fortify the antitumor effect of 4-1BBL/sPD-1. Either 4-1BBL or sPD-1 has been reported to enhance the antitumor efficacy of tumor vaccines (tumor-specific antigen or antigen-pulsed dendritic cells) by augmenting the response of immune cells to antigen (17, 23, 28, 29). We then asked whether tumor vaccine was still needed to trigger more efficient antitumor immune response in vivo if 4-1BBL is used in combination with sPD-1. In our animal tumor model system, both 4-1BBL and sPD-1 did have a synergistic effect with tumor vaccine (HSP70-H22 peptide complexes; Fig. 4A ), indicating that the presentation of more tumor antigens is necessary if 4-1BBL or sPD-1 is used alone. However, the cocktail by joining tumor vaccine, 4-1BBL, and sPD-1 did not further enhance the antitumor effect, which was evaluated by tumor suppression after the treatment (Fig. 4B), the cytotoxicity of splenocytes to H22 cells (Fig. 4C), and the transcription activities of different cytokine genes (Fig. 4D). Moreover, after 4-week treatment, the incidence of tumor was significantly decreased in both 4-1BBL/sPD-1 and 4-1BBL/sPD-1/vaccine groups, and tumor formation was also delayed; however, there was no significant difference between these two groups (Fig. 4E). These data indicate that the presentation of more tumor antigens is not necessary any more if 4-1BBL is used in combination with sPD-1 in tumor-bearing host.
Tumor cell amount and treatment time influence immunotherapeutic efficacy. The above results showed that 4-1BBL/sPD-1 effectively established antitumor environment in tumor-bearing mice, but the tumors were not completely suppressed in some mice. Therefore, we further analyzed the influence of tumor cell amount and the time of treatment on the therapeutic effect of 4-1BBL/sPD-1. When the mice were inoculated with 1 × 104 H22 tumor cells, the tumors were completely inhibited by a 4-week treatment, evaluated by the incidence of tumor and the survival of mice (Fig. 5A ). When the mice were inoculated with a 10-fold dosage of tumor cells (1 × 105), all of the mice in control group died within 40 days, but tumors, after 4-week treatment, were completely inhibited in ∼40% of mice, which survived for more than 100 days. Compared with 4-week treatment, the prolonged treatment (8 weeks) significantly reduced the incidence of tumor and increased the survival rate to ∼60% (P < 0.05, Kaplan-Meier analysis; Fig. 5A).
We then repeated the above experiment by inoculating mice with 1 × 105 H22 cells and further analyzed the efficacy of 4-1BBL/sPD-1. In the treatment groups, some mice formed tumor, and further treatment could delay the development of tumor, but the tumors eventually reached the size similar to that of control group (Fig. 5B). After a 4-week treatment, there was no palpable tumor in some mice, and H&E staining of the tissues from the inoculation sites of these mice showed the existence of a few tumor cells in some individuals and abundant immune cells, but no tumor cells, in other mice without palpable tumor (Fig. 5C). The existence of residual tumor cells revealed the reason why some mice developed tumor after 4-week treatment. After 8-week treatment, some “cured” mice were sacrificed for H&E staining. Tumor cells were not observed; however, the immune cells still existed (Fig. 5C). After 8-week treatment, some cured mice were fed for 4 more weeks. H&E staining of the tissue from the inoculation site showed that the immune cells were gone and the tissues recovered (Fig. 5C), indicating that tumor was really eradicated by the treatment of p4-1BBL/psPD-1 if tumor was not formed after 8-week treatment.
Many tumor immunotherapy efforts are focused on the generation of strong T-cell response against tumor by combination approaches including the utilization of costimulatory molecules. The proper combination of different therapeutic molecules can only be realized on the basis of the choice of suitable molecules and the elucidated mechanism through which they synergize each other. In this study, we chose 4-1BBL and sPD-1 to enhance the activation and function of T cells at both early and late phases of antitumor immune response. Our data show that a proper combination of immunoregulatory factors can establish an antitumor environment with in vivo existent tumor as the source of antigens. In the presence of 4-1BBL and sPD-1, the endogenous tumor antigens can be very efficiently utilized to induce antitumor immune responses that the responses cannot be further amplified by the addition of exogenous antigens such as tumor vaccine. Therefore, in our animal tumor model, 4-1BBL and sPD-1 synergistically established efficient antitumor immune environment to eradicate tumors from mice with small amounts of preexistent tumor cells or to eradicate tumors from more than half of individuals with larger amounts of preexistent tumor cells.
In this study, 4-1BBL is chosen for the purpose of augmentation of T-cell activation through costimulation. We show that the 4-1BBL–expressing vector does not have to enter into specific cells such as tumor cells or immune cells, and the expression of 4-1BBL in nontumor/nonimmune cells can efficiently exert its function in the experiments both in vitro and in vivo. 4-1BBL is capable of enhancing the expression of B7 molecules on dendritic cells to augment the dendritic cell–mediated second signal (B7/CD28 signaling) for T-cell activation (7, 8). In particular, 4-1BBL/4-1BB signaling, unlike B7/CD28, is very important for the effector phase of T-cell response (30–32). The stimulation of 4-1BBL can prevent the apoptosis of CD8+ cells after several cell division cycles (28). More importantly, many CD8+ T cells lose CD28 on their surface (32–34) in the later phase of immune response, and, without further activation, these CD8+CD28− T cells can be converted into suppressor T cells (35) to inhibit immune response at the effector phase. The CD8+CD28− T cells produced in immune response can only be further activated by the 4-1BBL/4-1BB pathway but not by the B7/CD28 pathway (32).
Although 4-1BBL has been shown to produce antitumor effects in various model systems (10–13), our present study revealed that 4-1BBL–induced IFN-γ production resulted in the up-regulation of PD-L1 in our tumor model, and 4-1BBL did not effectively decrease the expression of immunosuppressive cytokines IL-10 and TGF-β in tumor microenvironment. These data indicate that there is still plenty of room for further improvement of immune response when 4-1BBL is used alone. Thus, we further chose sPD-1 to cooperate with 4-1BBL for a better therapeutic efficacy.
Our data show that the blockade of PD-L1 by sPD-1 not only increased the transcription activities of IL-2 and IFN-γ genes in priming phase (first 2 weeks after tumor inoculation) but also improved antitumor immune response at later phase (second 2 weeks). The combination of 4-1BBL with sPD-1 significantly decreased the expression of IL-10 and TGF-β in the treated mice, resulting in the further increased expression of IL-2 and IFN-γ and the accumulation of CD8+ T cells in tumor microenvironment 4 weeks after tumor inoculation. The reduced expression of both IL-10 and TGF-β by sPD-1 favors the prevention of the formation of suppressor T cells, as we previously reported that blockade of the PD-L1/PD-1 pathway with sPD-1 resulted in the reduced amount of regulatory T cells in tumor microenvironment, which was evaluated on the basis of the down-regulation of foxp3 gene expression in tumor environment (23). In addition, it has been reported that blockade of the B7/CTLA-4 pathway enhances the antitumor immune response concomitant with the formation of autoimmune response (36), but that is not the case when the PD-L1/PD-1 pathway is blocked (25). In this study, we further showed that maintaining a relative long treatment period with 4-1BBL/sPD-1 was more beneficial without the induction of autoimmunity because we did not observe any abnormal changes in treated mice, including the functional test of liver and kidney (data not shown).
In summary, the therapeutic strategy with 4-1BBL in combination with sPD-1 boosts not only the chance of the development of a successful means for tumor immunotherapy but also the understanding of the undisclosed roles played by immune system in the face of malignancies. It implicates a potent benefit for the clearance of the residual tumor cells after surgical removal of tumor.
Grant support: National Development Program (973) for Key Basic Research of China, no. 2002CB513100, and the National Natural Science Foundation of China, no. 30471587.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: H. Xiao and B. Huang contributed equally to this work.
- Received August 29, 2006.
- Revision received December 8, 2006.
- Accepted December 27, 2006.