Introduction
Effective cancer immunotherapies should eradicate cancer cells and block the immunosuppression that occurs in cancer microenvironments [
1,
2]. Although CTLs play a major role in anti-tumor responses, increasing evidence indicates that the induction of cytotoxic effects is necessary, but not sufficient, to control tumor progression [
3]. The function of CTLs is affected by systemic and local immunosuppressive environments associated with tumor growth. The lytic activity of CTLs in the tumor microenvironment can be suppressed by myeloid-derived suppressive cells (MDSCs), tumor-associated macrophages (TAMs) and regulatory T cells (Tregs) that surround the tumor [
4‐
7]. Increasing the number of MDSCs generates natural suppressive activity in cancer patients [
8] and tumor-bearing mice [
9], and systemic accumulation of MDSCs is induced by various factors associated with cancers and several pathological conditions. In addition, CD4
+CD25
+ Tregs increase at tumor sites in mice and humans during lung [
10], head and neck [
11], breast [
12] and ovarian cancers [
13]. CD4
+CD25
+ Treg-depleting approaches have revealed that reduced Treg numbers improve anti-tumor responses and the inhibition of tumor growth [
14,
15]. Accordingly, successful cancer immunotherapy requires modulation of the immunosuppressive effects of tumor-associated MDSCs, M2 macrophages and Tregs.
Bacterial lipoproteins can be modified at the N-terminus with di- or triacyl glyceryl-cysteine units, which are recognized by TLR2 [
16,
17]. In addition to their TLR2 activity and their ability to induce dendritic cell maturation, recombinant lipoproteins stimulate a cytokine expression profile that is different from that of synthetic lipopeptides [
18]. We recently applied this platform technology to produce a recombinant mutant form of E7 (rlipo-E7m) to treat HPV-associated diseases. We observed that the administration of rlipo-E7m completely inhibited tumor growth [
19]. In the present study, we used the TLR9 agonist CpG ODN in combination with rlipo-E7m to treat large tumors. Our data indicate that the combination of rlipo-E7m and CpG ODN dramatically eliminates large tumors. Moreover, CpG ODN synergized with a TLR2 agonist–conjugated antigen to induce the systemic and local production of cytotoxic CD8
+ T cells and decrease the number of immunosuppressive cells both locally and systemically. These findings suggest that CpG ODN combines with a recombinant lipoprotein exhibiting TLR2 agonist activity to enhance anti-tumor immunity and block local immunosuppressive cells. These results demonstrate that the combination of CpG ODN and recombinant lipoprotein represents a feasible approach for the development of cancer vaccines.
Discussion
The limited success of cancer immunotherapy reflects the induction of CTL responses that cannot eliminate cancer cells in the presence of tumor-infiltrated immunosuppressive cells (i.e., MDSCs, Tregs and M2 macrophages). To develop a new generation of immunotherapeutic approaches, the enhancement of CTL responses and reduction of immunosuppressive cell numbers must be induced simultaneously. We recently developed rlipo-E7m, which possesses TLR2 agonist activity and robust CD8-mediated anti-tumor activity against palpable tumors in the absence of exogenous adjuvants [
19]. In clinical situations, immunotherapeutic reagents must eliminate large tumors and overcome the immunosuppressive barriers of the tumor environment. In this report, we demonstrate that rlipo-E7m alone could not inhibit the growth of large tumors following a single-dose treatment. Although multiple doses did eliminate tumors in some mice, rlipo-E7m in the presence of CpG ODN dramatically eliminated large tumors. In addition, the numbers of tumor-associated immunosuppressive cells, including MDSCs, Tregs and TAMs (especially M2-like cells), were decreased following treatment with rlipo-E7m/CpG. These results demonstrate that rlipo-E7m/CpG not only induces strong CTL responses but also modulates tumor-associated immunosuppressive cells. The synergistic TLR-mediated stimulation of DCs increases the production of inflammatory cytokines and promotes Th1-polarized immunity, which drives CTL responses [
22,
23]. The combination of TLR3 and TLR9 stimulation enhances the priming efficiency of a DNA vaccine [
24]. TLR2/TLR3 agonists or TLR3/TLR9 agonists synergistically activate DCs and subsequently increase the number of activated T cells [
25]. These results indicate that high-quality CTL responses are induced through a combination of multiple TLR agonists. TLRs also reduce the immunosuppressive activity of MDSCs [
26], TAMs [
27,
28] and Tregs [
29,
30].
Although the synergistic activation of DC and T cell responses through multiple TLR agonists has been reported [
22,
25,
31,
32], these synergistic effects are not observed when combinations of TLR2 and TLR9 agonists are used [
25,
32]. Similar results were obtained when rlipo-E7m (a TLR2 agonist) was combined with a TLR9 agonist to stimulate bone marrow-derived DCs (Additional file
5: Figure S5). However, synergistic activation was detected when plasmacytoid DCs were used (Figure
2a). We further measured TLR2 and TLR9 transcripts in different DC subsets and found that the expression of TLR9 was higher in pDCs than in BMDCs or splenic DCs. However, the expression of TLR2 was higher in BMDCs than in pDCs or splenic DCs (Additional file
6: Figure S6). We suggest that the synergistic effects of rlipo-E7m and CpG may be due to the balance of TLR2 and TLR9 expression in pDCs. Moreover, antigen-specific CD8
+ T cell responses were not improved when the HIV Env-derived peptide (KQIINMWQEVGKAMYAPPISGQIRRIQRGPGRAFVTIGK) was mixed with both TLR2 and TLR9 agonists [
32]. In contrast, immunization with the TLR2 agonist-fused antigen (rlipo-E7m) substantially increased CTL responses in the presence of the TLR9 agonist (Figure
3). These conflicting results may result from the lipid moiety of rlipo-E7m, which contains an unsaturated fatty acid, or the fact that rlipo-E7m efficiently co-delivers both antigen and TLR2 agonists to antigen-presenting cells, and the TLR9 agonist might enhance the cross-presentation of antigen to CD8
+ T cells. We also found that rlipo-OVA/CpG-pulsed pDCs induced higher T cell proliferation than rlipo-OVA alone (Additional file
7: Figure S7a). The rlipo-OVA/CpG-pulsed BMDCs did not induce higher T cell proliferation than rlipo-OVA alone (Additional file
7: Figure S7a). However, it would be nice if the combination could be reproduced in a different antigen model. We produced a model antigen recombinant lipo-ovalbumin (rlipo-OVA) and measured its therapeutic effects with CpG (Additional file
7: Figure S7b). We found that rlipo-OVA/CpG had stronger anti-tumor effects in EG7 (ovalbumin-expressing cell line)-bearing mice than rlipo-OVA alone. These results indicate that TLR9 agonists enhance antigen presentation of TLR2 agonist-fused antigens to CD8
+ T cells.
In addition to the dramatic anti-tumor effects of rlipo-E7m/CpG, the inhibition of local immunosuppressive cells was also observed following this combined treatment. CpG administration induces differentiation and blocks the immunosuppressive function of MDSCs [
26,
33] and Tregs [
34]. Our results also indicate that the administration of CpG alone mildly reduced the numbers of tumor-infiltrating MDSCs and TAMs. Interestingly, the administration of rlipo-E7m/CpG significantly improved the reduction of immunosuppressive cell numbers compared to treatment with CpG alone (Figure
6). In tumor-bearing mice, a high percentage of M2-like TAMs was detected among tumor-infiltrating leukocytes. This observation is consistent with a previous report indicating that the depletion of TAMs promotes the infiltration of lymphocytes [
35]. We observed that treatment with rlipo-E7/CpG enhanced the number of M1-like TAMs and reduced the number of M2-like TAMs (Figure
6 c, d). Interestingly, CD80
+ TAM numbers were dramatically increased in the rlipo-E7m/CpG group. CD80
+ TAMs have an M1-like phenotype that produces Th1-biased cytokines for M1 polarization [
36,
37]. Furthermore, CD80 is a co-stimulatory molecule in antigen-presenting cells that provides the second signal for priming T cells. Switching of tumor-infiltrating macrophages from M2 to M1 has been reported upon use of a combination of CpG ODN and anti-IL-10 antibodies to suppress tumor growth [
38]. Here, our data further indicate that rlipo-E7m/CpG significantly increases the number of antigen-specific CD8
+ T cells during tumor infiltration (Additional file
8: Figure S8). Reduction of the number of TAMs or immunosuppressive cells may not completely eradicate tumor growth; thus, local cytotoxic T lymphocytes are critical for killing tumor cells. The numbers of local CD8
+ T cells and antigen-specific CD8
+ T cells were significantly increased in the tumor microenvironment (Additional file
8: Figure S8). We speculate that these infiltrating CD8
+ T cells might secrete IFN-γ or other cytokines to shift from the M2 to the M1 phenotype. Although we observed that systemic administration of CpG ODN alone did not induce anti-tumor effects or reduce immunosuppressive cell numbers, CpG ODN may amplify the effects of rlipo-E7m and efficiently eliminate large tumors. Therefore, our findings suggest that the induction of CTL responses and the reduction of immunosuppressive cell numbers are critical for eliminating large tumors.
In conclusion, a single administration of recombinant lipoprotein induced strong anti-tumor immunity in the presence of a TLR9 agonist. Anti-tumor immunity resulted from the induction of antigen-specific CD8+ T cells and the reduction of immunosuppressive cells in the tumor microenvironment. Currently, we are investigating the potential critical cells and cytokines involved in the local inhibition of immunosuppressive cells.
Materials and methods
Animals, cell line and reagents
C57BL/6 mice were purchased from the National Laboratory Animal Center, Taiwan. TLR9-KO mice were purchased from Oriental Bioservice, Inc. (Tokyo, Japan). All experimental mice were maintained in a pathogen-free environment at the Laboratory Animal Center of the National Health Research Institutes (NHRI). The animals were used in compliance with institutional animal health care regulations, and all animal experimental protocols were approved by the NHRI Institutional Animal Care and Use Committee. For experimentally induced neoplasia in mice, the allowable tumor burden and criteria for euthanasia complied with the NCI Frederick ACUC Guidelines (Involving Experimental Neoplasia Proposals in Mice and Rats, 2006). Tumor survival was determined based on 20% weight loss, unexpected moribundity or an inability to obtain food or water.
The TC-1 cell line expressing the HPV-16 E6 and E7 oncoproteins was a kind gift from Dr. T-C. Wu (Johns Hopkins University, USA) [
39]. The cells were grown in RPMI-1640 medium supplemented with 10% (v/v) fetal bovine serum, 50 units/mL penicillin/streptomycin, 0.5 mM sodium pyruvate, 20 mM HEPES (Biological industries, Beit Haemek, Israel) and 0.5 μM β-mercaptoethanol at 37°C under 5% CO
2.
Oligodeoxynucleotide 1668 (CpG-ODN) was purchased from Invitrogen (Grand Island, NY). Its sequence was 5′-TCC ATG ACG TTC CTG ACG TT-3′ with a phosphorothioate backbone. Recombinant mouse GM-CSF and FLT-3 ligand were purchased from PeproTech, and lipopolysaccharide (LPS; Escherichia coli endotoxin serotype 055:B5) was purchased from Sigma-Aldrich. Carboxyfluorescein diacetate succinimidyl ester (CFSE) and propidium iodide (PI) were purchased from Invitrogen™. The PE-conjugated HPV16E749-57/MHC I tetramer was purchased from Beckman Coulter, Inc. The antibodies used in this study, with their respective clones in parentheses, were anti-CD16/32 (2.4G2), anti-CD4 (GK1.5), anti-CD8 (53-6.7), anti-F4/80 (BM8), anti-Gr-1 (RB6-8C5), anti-CD11b (M1/70), anti-IFN-γ (XMG1.1), anti-TNF-α (MP6-XT22), anti-IL-10 (JESS-16E3), anti-Foxp3 (FJK-16s) (all purchased from eBioscience®) and anti-CD45 (EM-05) (GeneTex, Inc). The chemotherapy drug cisplatin was purchased from Sigma Aldrich®.
Generation of dendritic cell subsets
The pDCs were derived from C57BL/6 mouse bone marrow [
40]. Briefly, the tibias were removed from 6-12-week-old mice and rinsed in 75% ethanol. The bone marrow cells were then flushed out and passed through a 70-μm nylon cell strainer (BD Falcon) with lymphocyte culture medium (LCM, RPMI-1640 medium supplemented with 10% (v/v) fetal bovine serum, 50 units/mL penicillin/streptomycin, 20 mM HEPES and 0.5 μM β-mercaptoethanol). After centrifugation at 1,200 rpm for 10 minutes, the bone marrow cells were lysed in 3 mL of RBC lysis buffer (BioLegend®) for 3 minutes, and 10 mL of LCM was added to halt the lysis. The cells were again centrifuged at 1,200 rpm for 10 minutes, and the cell supernatant was discarded. The cells were subsequently resuspended in LCM, and 2 × 10
6 cells were seeded into a 90 × 15 mm Petri dish (α-Plus) with 10 mL of LCM as well as 100 ng/mL of FLT-3 ligand (PeproTech) or 20 ng/ml of GM-CSF (PeproTech). The cells were incubated at 37°C under 5% CO
2 for 3 days, and another 10 mL of LCM containing 100 ng/mL of FLT-3 ligand or 20 ng/ml of GM-CSF was added to the cell culture plates (day 7, CD11c
+ cells ~75%). The floating BMDCs or pDCs were harvested on day 6 or day 7, respectively, and 2 × 10
5 DCs were seeded into a 96-micro-well plate with 0.1 mL of LCM. The stimulating ligand was dissolved in LCM and subsequently added to the DC culture for an additional 24 hours of incubation. For the DC activation analysis, several secretory cytokines in the culture supernatants were detected by ELISA. All assays were performed in duplicate in three independent experiments.
Immunization and tumor challenge
To evaluate therapeutic anti-tumor effects, TC-1 cells (2 × 10
5 per mouse) were implanted subcutaneously into the left flanks of naïve C57BL/6 mice 7, 14 or 25 days prior to immunization. The mice were arbitrarily assigned to groups (6 per group) and were immunized subcutaneously in the dorsum with the indicated doses of rlipo-E7m [
19], either alone or as an admixture with 10 μg of CpG ODN, in a total volume of 100 μL in PBS for each mouse. To monitor tumor growth, the tumors were measured with electronic calipers three times weekly. The tumor volume was calculated using the formula length x width
2 × 1/2.
TC-1 cancer cells (2 × 10
5) were inoculated into C57BL/6 mice by intravenous injection to establish an experimental animal model of metastatic lung cancer [
41]. After 14 days, a single dose of PBS, rlipo-E7m, CpG or rlipo-E7m/CpG was subcutaneously injected into the mice to evaluate the therapeutic effects of these compounds.
ELISPOT assay
The IFN-γ ELISPOT assay was performed according to the manufacturer’s instructions (eBioscience). Briefly, the ELISPOT plate (MSIP, Millipore) was pre-coated with anti-mouse IFN-γ capture antibody (AN18) overnight and subsequently blocked with LCM at room temperature for 2 hours. Splenocytes (1 × 106 per well) were plated in duplicate along with an H2-Db-restricted CTL epitope (HPV16E749-57) at a concentration of 10 μg/mL or with control peptides at 37°C for 48 hours. Following incubation, the plates were washed and incubated with biotinylated anti-mouse IFN-γ detection antibody (R46A2) for 2 hours, followed by incubation with avidin-HRP for 30 minutes and color development with AEC substrate reagents according to the manufacturer’s instructions (Sigma).
Flow cytometry
To characterize the populations of myeloid-derived suppressor cells (MDSCs), regulatory T cells (Tregs) and tumor-associated macrophages in immunized mice, RBC-lysed splenocytes or tumor cells derived from tumor-free and/or tumor-bearing mice were incubated with an anti-CD16/CD32 antibody to block non-specific binding and were subsequently stained with fluorochrome-conjugated monoclonal antibodies for surface marker and intracellular staining according to the manufacturer’s instructions (eBioscience®). Briefly, the cell suspension was stained with FITC-conjugated anti-CD4, PE-conjugated anti-CD25 and PE-Cy5-conjugated anti-Foxp3 to quantify Tregs; PE-conjugated anti-CD11b and PE-Cy7-conjugated anti-Gr-1 to quantify MDSCs; FITC-conjugated anti-F4/80 and PE-conjugated anti-CD11b to quantify TAMs and APC-conjugated anti-CD45 to quantify tumor-infiltrating leukocytes. For the intracellular detection of IFN-γ-secreting CD8+ T cells, RBC-lysed splenocytes derived from immunized C57BL/6 mice were re-stimulated with 10 μg/mL of H2-Db-restricted CTL epitopes (HPV16 E749-57) for 48 hours. The cells were subsequently harvested and stained with FITC-conjugated anti-CD8 and PE-conjugated anti-IFN-γ using a standard intracellular protocol. The E7-specific CD8+ T cells were stained with a PE-conjugated H-2Db/RAH tetramer and FITC-conjugated anti-CD8. The percentage of H-2Db/RAH+ CD8+ T cells was determined by flow cytometry (FACSCalibur, BD Bioscience, San Jose, CA). All of the data were analyzed using a FACSCalibur flow cytometer and the CellQuest software. All analyses were conducted on a gated lymphocyte population.
In vivo cytolytic assay
To detect antigen-specific cytolytic activity in the immunized mice in vivo, specific or irrelevant peptide-pulsed syngeneic splenocytes were used as target cells in a killing assay. The RBC-lysed splenocytes were counted and divided into two equal portions. These two portions were incubated at a density of 2 × 107 cells/ml with a specific peptide (HPV16 E749-57) and a non-specific peptide (OVA257-264) at 37°C for 30 minutes. To differentiate between the peptide-pulsed target cells, the two subsets of cells were labeled with CFSEhi (5 μM) and CFSElo (0.5 μM) at 37°C for 10 minutes. Both samples of cells were resuspended at 2 × 107 cells/ml and subsequently mixed at a 1:1 ratio (1 × 107:1 × 107) in PBS prior to adoptive transfer into immunized mice via tail vein injection 7 days after the previous immunization. The experimental cells were harvested 18 hours after adoptive transfer and analyzed using a FACSCalibur flow cytometer (BD Bioscience). The percentage of specific lysis was calculated using the following equation: % Specific lysis = [1-(%CFSE hi/%CFSE lo)] × 100.
Depletion of leukocyte subpopulations in vivo
CD4+, CD8+ or NK cells were depleted in vivo using 0.5 mg of anti-CD4 (GK1.5, eBioscience®), anti-CD8 (53-6.7, eBioscience®) or anti-NK1.1 (PK136, BioLegend®) injected intraperitoneally into mice one day prior to immunization. Rat IgG or mouse IgG (0.5 mg) (Invitrogen™) was used as the control antibody. The depletion efficiency was ~90% as determined by flow cytometry. Mice were implanted with TC-1 cells (2 × 105 per mouse) subcutaneously. Seven days later, 10 μg of rlipo-E7m/CpG was injected s.c. into the dorsum in a total volume of 100 μL. To monitor tumor growth, the tumors were measured with electronic calipers three times weekly. The tumor volume was calculated using the formula length × width2 × 1/2.
Statistical analysis
Statistical analyses were performed using Prism version 5.02 (GraphPad, CA, USA). Kaplan-Meier analysis was performed on the survival rates of the mice. The statistical significance of the differences between the groups was assessed using a two-tailed Student’s t test. For all results, P < 0.05 was considered statistically significant.
Competing interests
The authors declare no financial or commercial conflicts of interest.
Authors’ contributions
LSC, YCY and CCW performed the experiments with contributions from CHL, and HWC analyzed the data. HMH and SJL designed the experiments and wrote the manuscript. All authors read and approved the final manuscript.