Abstract
Purpose: Cancer immunotherapy, such as vaccination, is an increasingly successful treatment modality, but its interaction with chemotherapy remains largely undefined. Therefore, we explored the mechanism of synergy between vaccination with synthetic long peptides (SLP) of human papillomavirus type 16 (HPV16) and cisplatin in a preclinical tumor model for HPV16.
Experimental Design: SLP vaccination in this preclinical tumor model allowed the elucidation of novel mechanisms of synergy between chemo- and immunotherapy. By analyzing the tumor immune infiltrate, we focused on the local intratumoral effects of chemotherapy, vaccination, or the combination.
Results: Of several chemotherapeutic agents, cisplatin synergized best with SLP vaccination in tumor eradication, without requirement for the maximum-tolerated dose (MTD). Upon SLP vaccination, tumors were highly infiltrated with HPV-specific, tumor necrosis factor-α (TNFα)- and interferon-γ (IFNγ)–producing T cells. Upon combined treatment, tumor cell proliferation was significantly decreased compared with single treated and untreated tumors. Furthermore, we showed that TNFα strongly enhanced cisplatin-induced apoptotic tumor cell death in a JNK-dependent manner. This is consistent with upregulation of proapoptotic molecules and with enhanced cell death in vivo upon combined SLP vaccination and cisplatin treatment. In vivo neutralization of TNFα significantly reduced the antitumor responses induced by the combined treatment.
Conclusion: Taken together, our data show that peptide vaccination with cisplatin treatment leads to decreased tumor cell proliferation and TNFα-induced enhanced cisplatin-mediated killing of tumor cells, together resulting in superior tumor eradication. Clin Cancer Res; 21(4); 781–94. ©2014 AACR.
Previously, we have shown that synthetic long peptides (SLP) vaccination against human papillomavirus type 16 (HPV16) oncogenic proteins is safe and induces functional T-cell responses in mice and humans. Although vaccination induces potent clinical responses in mice and in patients with premalignant lesions, no clinical responses were observed in patients with cervical cancer. Here, we show that SLP vaccination can be safely combined with seven relevant chemotherapeutics without hampering the vaccine-induced antitumor response. In fact, vaccination synergizes strongly with cisplatin, without being dependent on the maximum-tolerated dose (MTD), thereby reducing toxicity. SLP vaccination induces tumor infiltration of tumor necrosis factor-α (TNFα) and interferon-γ (IFNγ)–producing T cells. Of these cytokines, TNFα was able to sensitize several mouse and human tumor cell lines for cisplatin-mediated killing. In vivo neutralization of TNFα strongly decreases the synergistic antitumor effect of cisplatin with peptide vaccination, indicating the crucial role of TNFα in tumor eradication. Our study indicates that vaccine-induced entry of TNFα-producing T cells into the tumor significantly enhances the tumoricidal effects of chemotherapy without requirement for the MTD.
Introduction
Cancer immunotherapy based on the activation of the patient's own immune system, in particular T cells, has recently shown remarkable successes. Characteristically, clinical responses to active immunotherapy are often delayed compared with conventional debulking reagents (1–3), begging the question whether fast-acting chemotherapeutics can be combined with immunotherapy. In several clinical trials, synergy between chemo- and immunotherapy was indeed observed (4, 5). Some chemotherapeutic agents can induce immunogenic cell death, featuring release of proinflammatory factors from dying cells with the potential to mature dendritic cells (DC), thereby activating CD8+ T cells (6–8). In addition, chemotherapeutics can convert the tumor microenvironment into a site permissive for vaccination by enhancement of antigen density (9). In about 50% of the patients with cervical cancer, the causative virus type is high-risk human papillomavirus type 16 (HPV16; ref. 10). The two viral oncoproteins E6 and E7 are required for maintenance of the transformed cell state (11), creating ideal targets for therapeutic vaccination (12). Previous results from our group show that vaccination with synthetic long peptides (SLP) vaccines, directed against HPV16 E6 and E7, can effectively cause tumor regression in preclinical mouse tumor models and eradication of lesions caused by HPV16 in patients with premalignant disease (13–15). Complete remission of grade 3 HPV16+ vulvar intraepithelial neoplasia correlated with strong and broad SLP vaccine-induced T-cell immunity. However, in end-stage cervical cancer, this SLP vaccine causes less robust HPV-specific T-cell responses without an apparent effect on survival (16, 17). We now show in preclinical models that therapeutic HPV16 E7 SLP vaccination has remarkable synergy with defined chemotherapeutic agents through mechanisms that are clearly distinct from those previously reported. The DNA cross-linking agent cisplatin displayed the strongest synergy with vaccination. The mechanisms of action are abundant infiltration with polyfunctional vaccine-induced CD8 T cells that synergize with cisplatin by tumor necrosis factor-α (TNFα)–mediated increased sensitivity of tumor cells to cisplatin-mediated killing, associated with reduced tumor cell proliferation after combination therapy. Notably, combination of therapeutic SLP vaccination with cisplatin leads to superior tumor eradication without added toxicity.
Materials and Methods
Mice
Female C57BL/6 mice were purchased from Charles River Laboratories and housed in the animal facility of the Leiden University Medical Center (LUMC; Leiden, the Netherlands). The congenic strains Thy1.1 (CD90.1) and Ly5.1 (CD45.1) and ovalbumin-specific T-cell receptor (TCR) transgenic OT-I mice, all on a C57BL/6 background, were bred in house. All mice were housed in individually ventilated cage (IVC) systems under specific pathogen-free conditions and used at 8 to 10 weeks of age. Experiments were approved by the Animal Experiments Committee of the LUMC, in line with the guidelines of the European Committee.
Tumor cell lines and culture conditions
Tumor cell line TC-1 (a kind gift from T.C. Wu, John Hopkins University, Baltimore, MD) was generated by retroviral transduction of lung fibroblasts of C57BL/6 origin with HPV16 E6 and E7 and c-H-ras oncogenes (18) and maintained as previously described (14). Tumor cell line C3 was generated by transfection of B6 mouse embryonic cells (MEC) with the complete HPV16 genome and maintained as previously described (19). The methylcholanthrene-induced fibrosarcoma cell line MCA205 was kindly provided by Prof. T. Tüting (University of Bonn, Germany). All cell lines were tested by PCR for rodent viruses with negative results. The DC line D1 was provided by P. Ricciardi-Castagnoli (University of Milano-Bicocca, Milan, Italy; ref. 20). Human cervical adenocarcinoma cell line HeLa, human cervical carcinoma SiHa and MCA205 cells were cultured in RPMI-1640 (Gibco), supplemented with 8% FCS (Greiner), 100 IU/mL penicillin/streptomycin (Gibco), and 2 mmol/L glutamine (Gibco). Mycoplasma tests routinely performed for all cell lines by PCR were negative.
Tumor (-treatment) experiments
On day 0, C57BL/6 mice were subcutaneously inoculated with 1 × 105 TC-1 tumor cells or 5 × 105 C3 tumor cells in 200-μL PBS and 0.2% BSA. When a palpable tumor was present (day 8 for TC-1 experiments, day 14 for C3 experiments), mice were split into groups with comparable tumor size and were treated (intraperitoneal, i.p.) with chemotherapy. On the basis of the results with cisplatin, doses of chemotherapy used were similar to the reported maximal-tolerated dose (MTD), but provided in a split dose with an interval of 1 week, resulting in the following doses: cisplatin (10 and 4 mg/kg), doxorubicin (4 mg/kg), oxaliplatin (5 mg/kg), gemcitabine (100 mg/kg), carboplatin (40 mg/kg), paclitaxel (20 mg/kg) at day 8 and 9, or with topotecan (2 mg/kg) at days 8, 9, 10, and 11. At the same time of the first chemotherapy treatment, mice with and without chemotherapy were vaccinated in the contralateral flank with synthetic long HPV16 E743–77 peptide (GQAEPDRAHYNIVTFCCKCDSTLRLCVQSTHVDIR), dissolved in PBS and emulsified in Montanide ISA 51 on day 9 upon tumor challenge or with 4 mg/kg cisplatin in NaCl i.p. on day 15 upon tumor challenge or with the combination (Fig. 1A).
In vivo antibody treatment
For in vivo CD4+ and CD8+ T-cell depletion, mice were injected i.p. on day 7 and then every 6 days onward with 100 μg of the monoclonal antibodies (mAbs) GK1.5 and YTS-169, respectively. All mice used had a >99% depletion of the targeted T-cell population as measured by flow cytometry. For in vivo TNFα neutralization, mice were injected i.p. on day 7 and then twice weekly until day 23 with 450 μg of mAb MP6-XT22. mAbs were prepared and purified as described previously (21).
Adoptive transfer experiments
For adoptive transfers, 0.5 × 106 OT-I Thy1.1 or 2 × 106 OT-II Ly5.1 T cells were negatively enriched via a CD8+ and CD4+ enrichment kit (BD), respectively, and together intravenous (i.v.) injected in Thy1.2/Ly5.2-recipient mice. Subsequently, the mice were vaccinated with two SLPs, one containing the SIINFEKL epitope (OVA241-270) to stimulate OT-I cells and the other containing the (OVA317-347) epitope to stimulate OT-II cells (22). One group also received cisplatin (4 mg/kg) via i.p. injection. On day 7, cisplatin treatment was repeated and on day 10 the mice were booster vaccinated with SLP in PBS with 20 μg adjuvant CpG. The presence of congenic marked T cells in the spleens and lymph nodes (LN) on day 14 was analyzed by flow cytometry. In addition, to determine the cytokine production, cells were incubated for 5 hours in the presence or absence of SIINFEKL peptide and Brefeldin A after which an intracellular cytokine staining was performed as previously described (14).
For analysis of the capacity of T cells to populate the tumor, both donor (Thy1.1/Ly5.2 and Thy1.2/Ly5.1) and recipient mice (Ly5.2/Thy1.2) mice were injected with 1 × 105 TC-1 tumor cells on day −8. On day −5, CD8-depleting antibodies were injected i.p. in the recipient mice. All CD8+ T cells were depleted but CD8+ T cells reappeared on day 5 as measured by flow cytometry. Both donor and recipient mice were injected on day 0 with SLP in Montanide. Donors with the congenic markers Thy1.1/Ly5.2 also received cisplatin (i.p.) on day 0. Two days after vaccination, spleens and vaccine draining LNs were isolated from both donor mice, mixed at 1:1, and transferred to recipient mice. After 8 days, the tumor, spleen, and tumor draining LN were isolated from each mouse and analyzed for the presence of donor cells by flow cytometry. Relative infiltration was calculated by setting the percentage of CD8+ T cells originated from the peptide treated mice on 100, the percentage of T cells originated from the cisplatin + peptide–treated mice as a percentage of that.
Analysis of tumor-infiltrating immune populations
Isolated tumors from mice transcardially perfused with PBS-EDTA were mechanically disrupted in small pieces and incubated for 15 minutes at 37°C in Iscove's modified Dulbecco's medium (IMDM)–containing Liberase (Roche). Single-cell suspensions were next prepared by mincing the tumors through a 70-μm cell strainer (BD Biosciences). For cell-surface staining, cells were resuspended in staining buffer (PBS + 2% FSC + 0.05% sodium azide) and incubated with H-2Db tetramers containing HPV16 E749–57 peptide (RAHYNIVTF), 7-aminoactinomycin D (for dead exclusion; Invitrogen), CD19 (clone 1D3), CD8a (clone 53–6.7), CD3 (clone 145-2C11), CD45.2 (clone 104), class II (clone M5/114.15.2), CD44 (clone IM7) for 30 minutes at 4°C. Fluorescent conjugated antibodies were purchased from BD Biosciences and eBioscience. For intracellular Ki-67 staining, surface-stained cells were fixed and permeabilized with the FoxP3 staining buffer, and subsequently incubated with FITC-labeled Ki-67–specific antibodies or isotype control (eBioscience).
Furthermore, single-cell suspensions of tumor(-infiltrate) were incubated for 5 hours with 40,000 D1 cells preloaded with the SLP HPV16 E743–77 (10 μg/mL) in the presence of Brefeldin A (2 μg/mL) after which an intracellular cytokine staining was performed.
Microscopy
Isolated tumors from transcardially perfused mice were embedded in Tissue-Tek and immediately frozen in 2-methylbutane cooled on dry ice. Cryostat sections (4–5 μm) were cut at −20°C, dried overnight at 60°C, and then frozen at −20°C.
After thawing, samples were fixed with acetone, dried, washed in PBS, incubated with biotinylated CD4 mAb (clone RM4-5; Pharmingen), washed and stained with CD3 mAb (Alexa Fluor 647, clone 17A2; BioLegend) and streptavidin Alexa Fluor 546 (molecular probes) in PBS with mouse serum. Images (×200) were captured with a confocal laser scanning microscope (Zeiss LSM 510). For analysis of mitosis activity, slides were stained with hematoxylin and eosin (H&E). Cell- and mitosis-counts were scored by two independent investigators. In vivo apoptosis was analyzed on cryostat sections by the use of terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) technology (in situ cell death detection kit, POD; Roche) according to the manufacturer's instructions. TUNEL staining was visualized with the substrate DAB (3,3′-diaminobenzidine tetrahydrochloride) after which the samples were counterstained with hematoxylin (Dako). Tissue sections treated without TdT were used as negative control.
CTL assay
Tumor-free C57BL/6 mice were immunized with synthetic HPV16 E743–77 peptide and 20μg CpG. Eight days after boosting, splenocytes were isolated, restimulated with the specific peptide for 7 days, and used as effectors in a CTL assay against TC-1 target cells. Interferon-γ (IFNγ; Prospec) exposed or nonexposed TC-1 tumor cells were incubated with various concentrations of cisplatin and incubated overnight. The next day, cells were washed, incubated for 1 hour with 1 μg/mL of E7 peptide (RAHYNIVTF), and then used in a 4-hour 51Cr release assay. Target cells were then labeled with 100 μL 51Cr for 1 hour, washed, and plated into a 96-well round-bottomed plate at a density of 2,000 tumor cells per well together with T cells (of which 4% was E7-specific as measured by tetramer staining) at the indicated E:T ratios. After 4 hours, cells were harvested, and the percentage 51Cr release was measured using a gamma counter.
Cell viability assays
For the various assays, TC-1 was seeded at 8,000 cells per well with 20 ng (250 international units) mouse TNFα (Prospec) and 2 μg/mL cisplatin. Concentrations of topotecan, carboplatin, and paclitaxel are indicated in Supplementary Figure S4F. C3 was seeded at 10,000 cells per well with 4 ng/mL TNFα and 6 μg/mL cisplatin, MCA205 was seeded at 7,000 cells per well with TNFα and cisplatin (concentrations are indicated in Supplementary Figure S4D). HeLa cells were plated at 18,000 cells per well with 2 μg/mL cisplatin and 25 ng/mL hTNFα, 40 μg/mL etanercept (Enbrel), or supernatant of phytohemagglutinin (PHA; 0.5 μg/mL)-stimulated peripheral blood mononuclear cells (PBMCs) of a healthy donor. SiHa cells were seeded at 20,000 cells per well with 100 ng/mL TNFα and 12 μg/mL cisplatin. Real-time apoptosis induction was quantified using a live cell apoptosis assay as previously described (23). Briefly, tumor cells were seeded into a GreinerBioOne 96-wells uClear plate and incubated for 24 hours. Then, cisplatin and/or TNFα were added to the cells. Binding of Annexin V–Cy3 conjugate to phosphatidyl serine present on the membranes of apoptotic cells was quantified in time by imaging with a BD Pathway 855 imager (Becton Dickinson). The total area of Annexin V–Cy3 fluorescence per image was quantified using Image Pro (Media Cybernetics). Cell death induced by cisplatin with supernatant of peptide-treated tumors was analyzed by Annexin V staining using flow cytometry.
Cell viability was determined using a standard colorimetric MTT (3-4,5-dimethylthiazol-2-yl-2, 5-diphenyl-tetrazolium bromide) reduction assay (Trevigen) and performed via the setup used for the live cell apoptosis assay. Analysis was performed 22 hours after cisplatin and/or TNFα exposure. Control cells without agents were cultured in parallel using the same conditions with comparable media changes. Absorbance was measured with a microplate reader (Bio-Rad) at a test wavelength of 570 nm, and a reference wavelength of 655 nm. For each treatment, cell viability was evaluated as a percentage using the following equation: (OD of treated sample/mean OD of untreated sample) × 100.
Quantitative real-time PCR
TC-1 tumor cells were harvested and then seeded at 1.3 × 105 cells per well (6 wells plate). After 24 hours of incubation, TNFα and/or cisplatin were added. Twenty-four hours later, cells were harvested and suspended in TRizol. Total RNA was isolated from cells using an RNeasy Mini Kit (Qiagen) according to the manufacturer's protocol. cDNA was synthesized using SuperScript III Reverse Transcriptase (Invitrogen). Oligonucleotide primers are compiled in Table 1. Reactions were performed using iQ SybrGreen supermix (Bio-Rad) on the my-iQ Real-Time PCR (Bio-Rad) with an annealing temperature of 60°C. Cycle threshold (Ct) values were normalized to the expression levels of the GAPDH and β-actin genes.
Gene . | Forward primer . | Reverse primer . |
---|---|---|
BCL-XL (Bcl2l1)a | TGGAGTAAACTGGGGGTCGCATCG | AGCCACCGTCATGCCCGTCAGG |
Bfk (Bcl2l15)a | CCCACTATCCCACGCTGCTA | TCCTCTCCTGAGCATGGTTCA |
Brucea | ACATCCGACAAGCAACGGTTA | AGCACGGTGAAGGATTTCTGA |
c-IAP2 (Birc3)a | TATTTGTGCAACAGGACATTAGGAGT | TCTTTCCTCCTGGAGTTTCCA |
TNFαb | ACGGCATGGATCTCAAAGAC | GGTCACTGTCCCAGCATCTT |
Ki-67b | CCTTTGCTGTCCCCGAAGA | GGCTTCTCATCTGTTGCTTCCT |
Xaf1a | CTGTGGCTCTGCCTTTGAAGT | CGTCTCCTGTGGGAAGATCAG |
GAPDHb | TGTGTCCGTCGTGGATCTGA | TTGCTGTTGAAGTCGCAGGAG |
β-actinb | GATGTCACGCACGATTTCC | GGGCTATGCTCTCCCTCAC |
PCNAb | GAACCTCACCAGCATGTCCA | ATTCACCCGACGGCATCTTT |
Gene . | Forward primer . | Reverse primer . |
---|---|---|
BCL-XL (Bcl2l1)a | TGGAGTAAACTGGGGGTCGCATCG | AGCCACCGTCATGCCCGTCAGG |
Bfk (Bcl2l15)a | CCCACTATCCCACGCTGCTA | TCCTCTCCTGAGCATGGTTCA |
Brucea | ACATCCGACAAGCAACGGTTA | AGCACGGTGAAGGATTTCTGA |
c-IAP2 (Birc3)a | TATTTGTGCAACAGGACATTAGGAGT | TCTTTCCTCCTGGAGTTTCCA |
TNFαb | ACGGCATGGATCTCAAAGAC | GGTCACTGTCCCAGCATCTT |
Ki-67b | CCTTTGCTGTCCCCGAAGA | GGCTTCTCATCTGTTGCTTCCT |
Xaf1a | CTGTGGCTCTGCCTTTGAAGT | CGTCTCCTGTGGGAAGATCAG |
GAPDHb | TGTGTCCGTCGTGGATCTGA | TTGCTGTTGAAGTCGCAGGAG |
β-actinb | GATGTCACGCACGATTTCC | GGGCTATGCTCTCCCTCAC |
PCNAb | GAACCTCACCAGCATGTCCA | ATTCACCCGACGGCATCTTT |
aPrimers were obtained from Eurogentec.
bPrimers were obtained from Sigma-Aldrich.
Statistical analysis
Survival for differentially treated mice was compared using the Kaplan–Meier method and the log-rank (Mantel–Cox) test. Additional statistical methods are stated in the legends. All P < 0.05 were considered significant.
Results
Cisplatin displays the strongest synergy with SLP vaccination
To test which chemotherapeutics have favorable antitumor effects in combination with therapeutic SLP vaccination, the HPV16 SLP vaccine was combined with seven clinically relevant chemotherapeutic agents. Mice were inoculated with TC-1 tumor cells, expressing the HPV16 oncoproteins E6 and E7, and tumor size was longitudinally measured. When a palpable tumor was present on day 8 after tumor challenge, treatment was started. Systemic treatment with chemotherapeutic agents alone, given twice with a 1-week interval (Fig. 1A), displayed a modest transient effect on tumor outgrowth (Fig. 1B–D and Supplementary Fig. S1A). SLP vaccines as a single therapy, given in Montanide to mimic the currently used clinical vaccine strategy in patients with HPV16+ tumors (13, 16, 17), induced only a temporal decrease in tumor size in most mice and complete tumor eradication in 16% of the mice. Strikingly, none of the chemotherapeutic treatments, at the doses used, significantly impaired the impact of peptide vaccination on tumor growth. Combined treatment consisting of oxaliplatin, doxorubicin, or paclitaxel with SLP vaccination did not enhance overall survival when compared with SLP vaccination alone (Fig. 1B and Supplementary Fig. S1A). In contrast, the chemotherapeutics topotecan, gemcitabine, carboplatin, or cisplatin clearly synergized with vaccination, showing significantly better survival in groups of mice treated with the chemo-immunotherapy combination, compared with either treatment alone (Fig. 1C and D and Supplementary Fig. S1A). In particular, cisplatin synergized well with SLP vaccination, resulting in a survival of 75% to 80% of the mice. To avoid significant weight loss (Supplementary Fig. S1B), a lower, less toxic, dose of cisplatin was used. Similar to the higher dose, the lower cisplatin dose combined with peptide vaccination completely eradicated tumors in approximately 75% of the mice (Fig. 1D and Supplementary Fig. S1B). There was indeed no significant difference between these two groups of chemo-immunotherapy-treated mice (P = 0.47), emphasizing the fact that in combination with immunotherapy, dosing of chemotherapy can be lower, avoiding unnecessary toxicity such as weight loss. Mice successfully treated with this combination remained tumor free for at least 2 months. When such tumor-free mice were rechallenged with TC-1 tumor cells, all were protected (data not shown). In a different murine tumor model for cervical cancer, a transformed B6 MEC line with the complete HPV16 genome (C3), the combined SLP, and cisplatin treatment again resulted in a stronger decrease in tumor size than single treatment (Supplementary Fig. S1C). Of note, after day 32, many of the shrunken C3 tumors grew out again, possibly via immune-escape.
In TC-1 tumor-bearing mice treated with both cisplatin and SLP vaccination, we observed that while elimination of CD4+ T cells had no effect on treatment efficacy, CD8+ T cell depletion completely abolished tumor control by the combined treatment (Supplementary Fig. S1D). Together, these data show that clinically relevant chemotherapy can be safely combined with SLP vaccination. We selected the effective and nontoxic combination with lower dose cisplatin to further investigate the mechanisms of synergy.
Effect of cisplatin chemotherapy on T-cell numbers, function, and homing capacity
To understand the mechanisms underlying the synergy observed with cisplatin plus SLP vaccination, we first asked whether chemotherapy can affect the antigen-specific expansion and differentiation of T cells. To this end, mice were injected i.v. with OT-I and OT-II TCR transgenic T cells. On this same day, these mice were vaccinated with two SLPs comprising the T-cell epitopes recognized by OT-I and OT-II T cells, emulsified with Montanide (Fig. 2A). Treatment with gemcitabine, carboplatin, or cisplatin did not affect either the numbers of OT-I/OT-II T cells or their cytokine production at the doses tested (Fig. 2B–E, and not shown).
Similar results were found in experiments in which the endogenous vaccine-induced HPV16-specific T-cell response in TC-1 tumor-bearing mice was measured. The percentages of circulating HPV16-specific tetramer-positive CD8+ T cells in mice vaccinated with SLP alone were comparable with those in mice treated with SLP vaccination and cisplatin (Fig. 2F and G). In addition, expression of the activation molecules CD25, CD69, CD137, and KLRG1 appeared unaffected on both CD4+ and CD8+ T cells, at least for cisplatin treatment (Supplementary Fig. S2).
To explore whether cisplatin-exposed T cells have differential homing capacity to the tumor microenvironment compared with unexposed T cells, adoptive transfers of congenically marked endogenous T cells from tumor-bearing mice, treated with SLP vaccine (Ly5.1) or with SLP vaccine and cisplatin (Thy1.1), were transferred into the same, CD8+ T cell–depleted host mice (Thy1.2/Ly5.2). Eight days later, spleens, LNs, and tumors were harvested and the presence of adoptively transferred cells was analyzed, revealing that cisplatin-exposed T cells were equally efficient in repopulating the spleen, LNs, and tumor as nonexposed T cells (Fig. 2H). Taken together, we conclude that the synergy observed between some chemotherapeutic agents and vaccination is not related to overt changes in systemic T-cell immunity.
Vaccination strongly increases the number of intratumoral polyfunctional CTLs
Next, we focused on effects of combined chemo-immunotherapy on the tumor microenvironment. Tumor size significantly affects various immune parameters (24, 25). Therefore, we used a treatment protocol (Fig. 3A) in which cotreatment still effectively synergized (Supplementary Fig. S3), but analyzed the tumors at the start of the regression phase (Fig. 3B). Analysis of the immune infiltrate in the tumor microenvironment at the start of regression showed that the percentage of leukocytes significantly increased upon treatment with cisplatin, peptide, and the combined treatment. The combination treatment resulted in a total of approximately 60% leukocytes in the tumor environment on day 19 (Fig. 3C and D). Given the significant tumor regression observed upon combined treatment, enhanced leukocyte infiltration at later time points is expected but has not been tested.
Although cisplatin was capable of inducing CD45+ leukocyte infiltration, very few of these cells were CD8+ T cells (Fig. 4A–F). Interestingly, substantial numbers of CD8 T cells were entering the tumor upon peptide vaccination or the combination treatment (7.8 times more CD8+ T cells in the combined treatment group compared with untreated; Fig. 4A). Confocal microscopy using antibodies to CD4 (red) and CD3 (blue) confirmed our observation that vaccination strongly increases the density of CD8+ T-cell infiltration (characterized as CD3+/CD4− cells) in the tumor beds (Fig. 4B and C).
Further characterization of the tumor-infiltrating CD8+ T-cell population revealed a high expression of Ki-67 (Fig. 4D) as a marker of proliferation, and showed that a very high percentage (∼40%) specifically recognized the tumor antigen E7 (Fig. 4E). When isolated, many of the tumor-infiltrating T cells produced the proinflammatory cytokines IFNγ and TNFα after in vitro exposure to E7-presenting antigen-presenting cells (D1 cell line; Fig. 4F).
Together, these data show that although chemotherapy and vaccination can both enhance leukocyte infiltration of the tumors, IFNγ- and TNFα-producing antigen-specific CD8 T cells enter this tumor only upon specific vaccination. Because cisplatin addition to vaccination had no effect on these CD8 T cells, we hypothesized that the enhanced antitumor responses observed upon combined treatment were not due to direct effects on the tumoricidal CD8 T cells but rather due to differences in the tumor cells.
Combined treatment of vaccination with cisplatin decreases tumor cell proliferation
While the intratumoral T cells expressed high levels of Ki-67, we noticed that the Ki-67 levels in tumor cells of the combined treatment group were considerably lower than in tumor cells from untreated, cisplatin-, or peptide-treated mice, suggesting a decrease in tumor cell proliferation caused by the combined interaction of cisplatin and peptide (Fig. 5A). To corroborate this, the numbers of mitotic figures per high-power field in H&E-stained sections, a well-reproducible, strong prognosticator of disease in LN-negative patients (26), were compared between the untreated group and the various treatment groups. This analysis confirmed that upon either cisplatin or peptide vaccination alone, the division rate of tumor cells declined but the lowest number of mitoses was observed upon combined cisplatin and vaccination treatment (Fig. 5B). Following in vitro treatment of tumor cells with TNFα, IFNγ, and/or cisplatin, both Ki-67 and proliferating cell nuclear antigen (PCNA) levels were not decreased (data not shown), suggesting that the decreased proliferation of tumor cells is not a result of direct exposure to either one of these T cell–produced cytokines. Together, these data show that combined cisplatin and SLP vaccination decrease the proliferation of tumor cells more extensively in vivo than either treatment alone.
Combined treatment of vaccination with cisplatin enhances tumor cell death and is caused by the combined action of TNFα and cisplatin
Tumor regression and clearance ultimately depend on tumor cell death. We therefore analyzed in vivo cell death by the TUNEL assay, a technique that detects fragmented DNA resulting from apoptosis. Strikingly, we observed that the combination treatment significantly enhanced tumor cell death compared with the untreated and cisplatin-treated tumors on day 17 already (Supplementary Fig. S4A) and also on day 19 (Fig. 6A and B).
First, we hypothesized that cisplatin might enhance the susceptibility of tumor cells to T-cell killing. To test this, TC-1 tumor cells were pretreated with various doses of cisplatin. The highest dose induced approximately 40% cell death and the lowest dose almost no cell death as measured by the MTT assays (data not shown). After washing, the tumor cells were incubated with splenocytes containing 4% HPV16-specific CD8+ T cells (data not shown). Although tumor cells were efficiently killed by these splenocytes in these short cytotoxicity assays, cisplatin pretreatment did not enhance sensitivity toward CTL-mediated killing (Fig. 6C). Thus, the observed synergy between SLP vaccination and cisplatin cannot be explained by an increased sensitivity of cisplatin-treated TC-1 tumor cells toward CTL-mediated killing as measured by short chromium release assays. Therefore, we hypothesized that the enhanced killing of tumor cells observed in vivo rather depends on prolonged and sustained cooperation between T-cell effector function and cisplatin activities on tumor cells.
Because we have seen that tumors from vaccinated mice are highly infiltrated with IFNγ- and TNFα-producing T cells, we hypothesized that the synergistic effect on tumor cell death was due to an enhancement of cisplatin-induced cell death by these cytokines. We therefore examined the effect of IFNγ and TNFα on cisplatin-induced tumor cell death. By live apoptosis imaging, it became apparent that coincubation of cisplatin with various dosages of IFNγ had no effect on cell death (Fig. 6D and data not shown). Strikingly, TNFα exposure strongly sensitized tumor cells to cisplatin-mediated killing (Fig. 6D; Supplementary Fig. S4B; Supplementary Movies S1–S4). This observation was confirmed by performing an MTT assay with recombinant TNFα (Fig. 6E), and supernatant of overnight incubated single-cell suspensions of peptide-treated tumors (Supplementary Fig. S4C).
To test whether TNFα sensitization of cisplatin-treated tumor cells is a general phenomenon, similar experiments were done with mouse tumor cell lines C3 and MCA 205 and human cervical cancer cell lines SiHa and HeLa, yielding comparable results (Fig. 6F and Supplementary Fig. S4D and S4E). Similarly, the supernatant of activated PBMCs could sensitize HeLa tumor cells for cisplatin-induced killing (Supplementary Fig. S4E and Fig. 6G). Because the addition of the TNFα inhibitor etanercept fully abolished this synergy (Fig. 6G), we concluded that specifically the TNFα in the supernatant of these human T cells sensitized for cisplatin cell death. In addition, we observed that TNFα could sensitize tumor cells for topotecan- and carboplatin-induced tumor cell death but not for paclitaxel-induced apoptosis (Supplementary Fig. S4G).
Enhanced cell death of TC-1 cells was accompanied by a TNFα-induced upregulation of the antiapoptotic molecule c-IAP2, and a cisplatin-induced upregulation of the proapoptotic gene Xaf1. Remarkably, treatment with both cisplatin and TNFα strongly enhanced expression of the proapoptotic molecule Bfk, suggesting a crucial role for this molecule in the observed synergy (Fig. 6H). Expression of the antiapoptotic molecules Bruce and Bcl-xL was not altered by TNFα, cisplatin, or combined exposure (data not shown).
TNFα binding to TNFR1 and TNFR2 can induce signaling leading to either cell survival or cell death. The activation status of the MAPK family member JNK determines cellular outcome (27). In addition, it has been found that combined cisplatin/TNFα treatment of renal proximal tubular cells induces JNK activation, leading to a shift in apoptotic programming and, causing tubular cell apoptosis at much lower cisplatin levels than in the absence of TNFα (28). We hypothesized that a similar mechanism may play a role in the synergy between vaccine-induced T-cell responses and cisplatin treatment. When TC-1 tumor cells were pretreated with the JNK inhibitor SP600125, cisplatin–TNFα-induced cell death was largely inhibited (Fig. 6I), indicating that JNK activation is important in the TNFα-mediated sensitization of tumor cells to cisplatin-induced killing. Next, we used TNFα-neutralizing antibodies to examine the role for TNFα in the synergy between cisplatin and peptide vaccination observed in vivo. Strikingly, the use of TNFα-neutralizing antibodies strongly decreased the synergy between cisplatin and peptide vaccination (Fig. 6J).
Together, these data show that TNFα sensitizes for cisplatin in a JNK-dependent manner causing synergistic cell death. This is confirmed by the observation that tumors from the combined treatment group contain significantly more apoptotic tumor cells than tumors from single-treated mice. Moreover, we show that vaccination induces TNFα-producing T cells in the tumor. When the mice, treated with peptide vaccination and cisplatin, receive TNFα-neutralizing antibodies, the synergy between the two treatments is strongly decreased, emphasizing the important role for TNFα in the cisplatin + peptide–induced antitumor response.
Discussion
In this study, we show that SLP vaccination combined with clinically relevant chemotherapeutic agents acts synergistically in tumor eradication. Combined treatment of low-dose cisplatin with peptide significantly enhanced survival compared with single treatment with high-dose cisplatin. Effective antitumor responses induced by the combined treatment were accompanied by enhanced leukocyte infiltration into the tumor, in particular by vaccine-specific CD8+ T cells. Because our vaccination protocol includes subcutaneous antigen delivery in Montanide in the opposing flank, T cells must have been actively attracted to the contra-lateral tumor site. In the local tumor environment, these T cells produce TNFα and IFNγ, indicating that they are fully functional.
Reportedly, TNFα and IFNγ can drive cancer cells into senescence, corresponding to a decreased Ki-67 expression (29). Although the cisplatin-treated TC-1 tumors infiltrated with vaccine-induced TNFα- and IFNγ-producing T cells show a decreased proliferation in vivo, in vitro treatment of tumor cells with TNFα, IFNγ, and/or cisplatin was not associated with decreased Ki-67 and PCNA levels (measured by qPCR, data not shown). This suggests that, at the doses used, TC-1 tumor cells were not driven into senescence by TNFα and IFNγ. The two HPV16 oncoproteins E6 and E7 are expressed by TC-1 tumors (18). A recent study on mouse keratinocytes reported that the oncoprotein E7 can interfere with IFNγ-mediated JAK1–JAK2–STAT1–IRF-1 signaling pathways (30). Because STAT1-deficient β-cancer cells resisted TNFα- and IFNγ-induced senescence (29), we hypothesize that E7 expression by TC-1 tumor cells protects them from complete senescence.
Over one third of patients treated with cisplatin develop renal injury within 10 days after a single dose of cisplatin. A molecule strongly associated with cisplatin-induced (nephro-) toxicity is TNFα (28, 31–33). Although this cytokine has been proposed as a strong trigger of cancer cell death (34), its life-threatening toxicity when administered systemically prohibits systemic clinical use (31, 35). Intratumoral administration of TNFα remains challenging for some primary tumors and especially for metastasis. In contrast to renal parenchymal cells (36), TC-1 tumor cells did not express TNFα after cisplatin treatment (data not shown). Strikingly, we show here that SLP vaccination induces large numbers of TNFα-producing T cells that are likely further stimulated to produce TNFα by antigen encounter within the tumor. In turn, this allows synergistic cell death of tumor cells when combined with cisplatin (Fig. 6), and other (carboplatin and topotecan) but not all chemotherapeutics. Previously, it was observed that HPV16 E7-specific immune responses combined with cisplatin treatment can enhance apoptotic tumor cell death (37, 38). Here, we confirm these observations and additionally show that TNFα has a key function in this enhanced cell death, indicated by the decreased synergy between cisplatin and vaccination in antitumor responses when TNFα was neutralized by systemic antibody treatment. Also, incubation of tumor cells with TNFα combined with low doses of cisplatin can enhance proapoptotic molecule expression in tumor cells and thus increases the level of cell death. Because cisplatin with TNFα-mediated apoptosis is induced after a minimum of 10 hours, this mechanism is likely not involved in short in vitro cytotoxicity assays as performed us (Fig. 6C) and others (37, 39). However, the synergy is relevant when TNFα-producing T cells and cisplatin are together with tumor cells for a longer period as occurs in the in vivo situation.
In addition to the previously described chemotherapy-enhanced sensitivity for granzyme B–mediated tumor cell death by intratumoral T cells (39, 40), the synergy between TNFα and cisplatin is a novel mechanism involved in successful chemo-immunotherapy. We postulate that both mechanisms can be involved in previous studies in which synergy was observed between cisplatin and efficient T cell–based immunotherapy (9, 37, 38, 41–44).
In addition to cisplatin-induced cell death, TNFα can enhance doxorubicin-induced apoptosis (45). However, it was recently shown that although doxorubicin enhances the RNA levels of TNFα by tumor cells, this cytokine is not involved in doxorubicin-induced antitumor responses in various tumor models (46). The levels of TNFα induced by the vaccine-induced CD8+ T-cell infiltrated in the tumor beds are likely to be much higher than the levels produced by chemotherapy-treated tumors, explaining why these T cells can potentiate cisplatin-mediated apoptosis, resulting in the remarkable synergistic antitumor activity of cisplatin and SLP vaccination seen in the current work. These data show that the introduction of different forms of active T cell–based immunotherapies into clinical practice calls for detailed analyses of the different forms of chemotherapy that best synergize with and support the immunotherapy. What is a scourge in the toxicity of cisplatin toward kidney tubular epithelial cells (28, 36) turns out to be a blessing in the enhanced demise of tumor cells helped along by the locally high levels of TNFα produced by vaccine-induced tumor-specific T cells.
Other investigators have shown marked differences in the induction of immunogenic cell death by different chemotherapeutic compounds (47). This is obviously of major importance in the absence of additional robust stimulation of tumor-specific T-cell responses, such as possible by SLP vaccination. In the case of this study, induction of a therapeutic T-cell immune response by chemotherapy itself is of no major consequence, because of the robust SLP vaccine that is capable of inducing a powerful therapeutic immune response by itself. Upon vaccination combined with cisplatin, the tumor microenvironment is highly infiltrated with leukocytes, including the HPV-specific, cytokine-producing, tumoricidal T cells. This work shows that together with TNFα produced by the abundant T cells in the tumor, cisplatin enhances cell death and causes decreased proliferation of tumor cells. These data elucidate mechanisms by which therapeutic vaccination synergizes with cisplatin for efficient systemic antitumor T-cell responses, providing a strong rationale for implementation of this type of chemo-immunotherapy.
Disclosure of Potential Conflicts of Interest
C.J.M. Melief is an employee of and has ownership interest in ISA Pharmaceuticals. S.H. van der Burg is a consultant/advisory board member for ISA Pharmaceuticals. C.J.M. Melief and S.H. van der Burg are named as inventors on a patent on the use of synthetic long peptides as vaccines (US 7.202.034), which is held by Leiden University Medical Center and licensed to ISA pharmaceuticals. No potential conflicts of interest were disclosed by the other authors.
Authors' Contributions
Conception and design: T.C. van der Sluis, F. Ossendorp, B. van de Water, R. Arens, S.H. van der Burg, C.J.M. Melief
Development of methodology: T.C. van der Sluis, E.S. Jordanova, A. Sloots, F. Ossendorp, R. Arens, S.H. van der Burg, C.J.M. Melief
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T.C. van der Sluis, S. van Duikeren, S. Huppelschoten, E.S. Jordanova, E.Beyranvand Nejad, A. Sloots, L. Boon, V.T.H.B.M. Smit
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T.C. van der Sluis, S. van Duikeren, S. Huppelschoten, E.S. Jordanova, V.T.H.B.M. Smit, M.J.P. Welters, B. van de Water, R. Arens, S.H. van der Burg, C.J.M. Melief
Writing, review, and/or revision of the manuscript: T.C. van der Sluis, E.S. Jordanova, L. Boon, V.T.H.B.M. Smit, M.J.P. Welters, F. Ossendorp, B. van de Water, R. Arens, S.H. van der Burg, C.J.M. Melief
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. van Duikeren
Study supervision: B. van de Water, R. Arens, S.H. van der Burg, C.J.M. Melief
Acknowledgments
The authors would like to thank W. Benckhuijsen and J.W. Drijfhout for providing peptides and K. Franken for providing tetramers. The authors also thank L. Fredriksson for assistance with the live apoptosis imaging and H. Brouwers and D.T.S.L Jansen for technical assistance.
Grant Support
This study has been conducted by Leiden University Medical Center, which holds a patent on the use of synthetic long peptides as vaccine (US 7.202.034).
This work was supported by a grant from Dutch Cancer Society, KWF 2009-4400 to C.J.M. Melief and S.H. van der Burg and a Gisela Thier grant from LUMC to R. Arens.
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.