Background

Toll-like receptors (TLRs) are evolutionarily conserved receptors, the largest constituents of the pattern recognition receptor family that serve as key players in innate immune responses.1 2 Multiple TLRs have been identified in humans and mice (TLR1-TLR10 for humans, TLR1-TLR9 and TLR11-TLR13 for mice).3 These receptors recognize conserved moieties associated with pathogens (pathogen-associated molecular patterns) or damaged cells (damage-associated molecular patterns), and initiate a variety of downstream signaling cascades that ultimately lead to the activation and maturation of immune cells involved in innate and adaptive immune responses.4 The capability of TLRs to initiate innate and adaptive immune responses has positioned them to serve as targets to engage during vaccination, to augment vaccine-mediated immune responses.5 Consequently, synthetic agonists for TLRs have been extensively studied as adjuvants for traditional vaccines, and multiple recent reports have established the applicability of TLR agonists as adjuvants for cancer vaccines.6–9

In our previous studies, we found that using TLR agonists as vaccine adjuvants for peptide or DNA antitumor vaccines resulted in greater antitumor T cell function.10 This was mediated in part by decreased expression of a T cell checkpoint receptor, PD-1,11 on vaccine-activated CD8+T cells. Specifically, we found that certain TLR agonists (notably TLR1/2, TLR7/8, and TLR9) stimulated the secretion of IL-12 from professional antigen presenting cells (APCs) and this downregulated the expression of PD-1 on CD8+T cells at the time of activation. We also found that the combination of multiple TLR agonists could further improve the antitumor function of CD8+T cells activated by vaccination.12 For example, certain TLR agonist combinations (notably TLR1/2+TLR3, TLR1/2+TLR9, and TLR3+TLR9) resulted in further downregulation of PD-1 as well as other checkpoint receptors, notably lymphocyte activation gene 3 (LAG-3),13 and CD160,14 in mouse CD8+T cells via upregulation of IL-12 and IFNβ secretion from professional APCs. These combinations also augmented the antitumor efficacy of peptide or DNA vaccines when used as adjuvants in murine tumor models.

Interestingly, we also found that TLR agonists (notably the combination of TLR3 and TLR9) led to increased expression of certain checkpoint receptors, such as cytotoxic T-lymphocyte-associated protein 4 (CTLA-4),15 T cell immunoglobulin and mucin-domain containing 3 (TIM-3),16 V-domain Ig suppressor of T cell activation (VISTA)17 and T cell immunoreceptor with Ig and ITIM domains (TIGIT),18 on CD8+T cells when used during activation of these cells.12 Many studies have shown that disruption of T cell checkpoint receptor signaling with immune checkpoint blockade can improve the antitumor efficacy of CD8+T cells. For example, we previously reported that OT-I T cells activated by a high-affinity epitope had increased PD-1 expression, and administration of αPD-1 or αPD-L1 along with the vaccine demonstrated greater antitumor efficacy in murine tumor models.19 20 These findings suggest that disruption of downstream signaling of specific T cell checkpoint receptors, which are upregulated following TLR agonist treatment, could have even more profound effects on the antitumor efficacy of vaccine-activated CD8+T cells.

In the current study, we hypothesized that the combination of TLR agonists (TLR3 and TLR9) and specific immune checkpoint blockade, when used at the time of T cell activation with vaccines, would elicit greater antitumor activity. Using ovalbumin-expressing E.G7 tumor (eg,7-OVA-PDL1high) and MyC-CaP prostate murine tumor models, we assessed the antitumor efficacy of CD8+T cells following stimulation with vaccines, including or excluding TLR3 and TLR9 agonists, and including or excluding immune checkpoint blockade (αPD-1, αCTLA-4, αLAG-3, αVISTA, αTIM-3, and/or αTIGIT). The effects of treatment on tumor-infiltrating lymphocytes were assessed.

Materials and methodsMice

OT-I (C57BL/6-Tg(TcraTcrb)1100Mjb/J, Stock No: 003831), OT-II (B6.Cg-Tg(TcraTcrb)425Cb n/J, Stock No: 004194)), C57BL/6J (Stock No: 000664), and FVB/NJ (FVB, Stock No: 001800) mice were purchased from The Jackson Laboratory (Jax, Bar Harbor, Massachusetts, USA). Mice were maintained in microisolator cages under aseptic conditions.

Cell lines

E.G7-OVA (derivative of EL4) and MyC-CaP cells were obtained from ATCC (Manassas, VA, Cat. # CRL-2113, #CRL-3255) and maintained via the ATCC-recommended culture methods. E.G7-OVA cells were lentivirally transduced to constitutively express PD-L1, as previously described.20

Peptides

Ovalbumin peptides for H-2b-restricted class I epitope (SIINFEKL) and class II epitope (ISQAVHAAHAEINEAGR, OVA-II) were synthesized, and the purity and identity were confirmed by mass spectrometry and gas chromatography (LifeTein, Hillsborough, New Jersey, USA). Peptides were reconstituted in DMSO (2 mg/mL) and stored at −80°C until use.

TLR agonists

TLR3 agonist (Poly(I:C) HMW, vac-pic) was purchased from InvivoGen (San Diego, CA). TLR9 agonist (ODN 1826) was obtained from Integrated DNA Technologies (San Diego, CA). Both agonists were dissolved in endotoxin-free physiological water from InvivoGen (San Diego, CA). Doses chosen for administration to animals were as previously reported.12

In vitro assaysOT-II splenocyte stimulation

Splenocytes were isolated from the spleens of OT-II mice, disaggregated using a mesh screen, and then treated to osmotically lyse red blood cells with an ammonium chloride/potassium chloride lysis buffer (0.15 M NH4Cl, 10 mM KHCO3, 0.1 mM EDTA). Splenocytes were then cultured in RPMI 1640 medium supplemented with L-glutamine, 10% fetal calf serum, 200 U/mL penicillin/streptomycin, 1% sodium pyruvate, 1% HEPES, and 50 µM β-MeOH with 2 µg/mL class-II OVA peptide (ISQAVHAAHAEINEAGR). TLR agonists were added 1 hour before stimulating with the peptide at the following concentrations: 10 µg/mL Poly(I:C) HMW, 5 µM ODN 1826. After 72 hours, cells were analyzed via flow cytometry with the following antibodies: CD3-FITC (BD 555274), CD4-BUV395 (BD 563790), CD8-BV786 (BD 563332), PD-1-PECF594 (BD 562523), and Live/Dead Ghost dye 780 (Tonbo, San Diego, CA 13-0865-T100).

Treg suppression assay

CD8+T cells were isolated from OT-I splenocytes via immunomagnetic negative selection (StemCell Vancouver, BC, Canada, 19853), then labeled with carboxyfluorescein succinimidyl ester (CFSE, Biolegend 423801) according to the manufacturer’s instructions. OT-II splenocytes were stimulated as above and Tregs were then isolated using the EasySep Mouse CD4+CD25+ regulatory T cell Isolation Kit II (StemCell 18783). 1×105 labeled CD8+T cells were cultured together with Tregs at a 1:1 ratio. CD8+T cells were stimulated with anti-CD3/anti-CD28 coated beads (Thermo Fisher 11 456D) at a ratio of 2 beads per CD8+T cell. 30 units/mL of murine IL-2 was then added, and cells were cultured for 72 hours in 96-well plates. CD8+T cells were analyzed for proliferation (loss of CFSE) via flow cytometry. For analysis of cytokines synthesized by CD3+CD8+T cells, Tregs from stimulated OT-II splenocytes were isolated as above and co-cultured at a 1:1 ratio with CD3+CD8+T cells from wild-type mice as described above. CD3+CD8+ T cells were stimulated with anti-CD3/anti-CD28 coated beads at a ratio of 2 beads per CD3+CD8+ T cell. Cells were cultured for 4 hours at 37℃ in the presence of GolgiStop (containing monensin, BD 554724). Cells were then washed, blocked with 50%PBS/FBS for 30 mins, and stained for cell surface markers with the following antibodies: CD3-PE-Cy5 (Biolegend 100310), CD4-BV711 (Biolegend 100550), CD8-Alexa700 (Biolegend 100730), and Live/Dead Ghost Dye 780 (Tonbo 13-0865 T100). Cells were then incubated for 30 min at 4℃, washed, and incubated at 4℃ in FoxP3 Fixation buffer overnight (eBioscience 00-5523-00). The next day, cells were washed with permeabilization buffer, then stained with the following intracellular antibodies: Perforin-FITC (Invitrogen 11-9392-82), IFNγ-PE (BD 554412), Granzyme B-PE-eFluor610 (Invitrogen 61-8898-82), and TNFα-PE-Cy7 (eBioscience 25-7423-41).

Tumor treatment studiesAntibody treatments

All antibodies were administered at 100 µg intraperitoneally. An αPD-1 hybridoma (clone G4, a gift by Dr. Lieping Chen) was produced by Envigo (Madison, WI). αCTLA-4 (BE0164), αLAG-3 (BE0174), αTIGIT (BE0274), αTIM-3 (BE0115), αVISTA (BE0310), and Armenian Hamster IgG (BE0091) were purchased from BioXCell. Treg-depleting αCTLA-4 (αCTLA-4 IgG2a, αTreg) was produced by Neoclone (Madison, WI) from a hybridoma provided by Bristol-Myers Squibb (New York City, NY).

Tumor studies with E.G7-OVA tumors

E.G7-OVA-PD-L1high cells were injected subcutaneously into 6-week-old C57BL/6 mice. Because these tumors have different growth rates in male and female mice, mice of only one sex (female) were used for each of these studies. When tumors were palpable and similarly sized (day 7, ~0.1 cm3), 2×106 naive OT-I splenocytes were adoptively transferred to each mouse intraperitoneally. The following day, mice were immunized subcutaneously with 100 µg SIINFEKL peptide and coadministered with TLR agonists at the following doses: TLR3 [Poly(I:C) HMW, 100 µg/mouse], TLR9 (ODN 1826, 50 µg/mouse). Mice were further treated with immune checkpoint blockade, αPD-1, αTIM-3, αTIGIT, αVISTA, αLAG-3, αCTLA-4, αTreg, IgG control, or antibody combinations, the day following immunization. Tumor volume was measured using calipers and calculated in cubic centimeters according to the formula: (π/6)×(long axis)×(short axis).2 In a parallel study, tumors were collected and digested in collagenase, DNAse I, and protease inhibitors for 2 hours at 37°C, passed through a 100 mm mesh screen, and analyzed by flow cytometry with the following antibodies: SIINFEKL H2Kb tetramer-BV421 (NIH Tetramer Core Facility), CD3-PE-Cy5 (Biolegend 100310), CD4-BV711 (Biolegend 100550), CD8-Alexa700 (Biolegend 100730), CD25-PE-CF594(BD 562694), CD44-Alexa488 (Biolegend 103016), CD45-PerCP-Cy5.5 (Biolegend 103132), CD62L-BV510 (Biolegend 104441), CD127-APC (Biolegend 121122), KLRG1-PE-Cy7 (Biolegend 138416), FOXP3-PE (Thermo Fisher 12-5773-82), 4-1BB-BV605 (BD 106110), and Live/Dead Ghost dye 780 (Tonbo, San Diego, CA 13-0865-T100).

Tumor studies with MyC-CaP tumors

Six-week-old FVB/NJ mice were implanted with 1×106 MyC-CaP cells administered subcutaneously. Only male mice were used for these studies, given that the tumor of origin is male specific. Each mouse was then immunized intradermally with 100 μg DNA vaccine (pTVG-AR), or vector control (pTVG-4) weekly, beginning 1 day after tumor implantation. TLR agonists were coadministered with the vaccine intradermally at the following doses: TLR3 (Poly(I:C) HMW, 100 µg/mouse), TLR9 (ODN 1826, 50 µg/mouse). Mice were further treated with immune checkpoint blockade, αPD-1, αLAG-3, αCTLA-4, or IgG control the day following each immunization. Tumor volumes were measured as described above. In a parallel study, tumors were collected after treatment and digested in collagenase, DNAse I, and protease inhibitors for 2 hours at 37°C, passed through a 100 mm mesh screen, and analyzed by flow cytometry with the following antibodies: CD3-FITC (BD 555274), CD4-BUV395 (BD 563790), CD8-BV785 (BD 563332), CD25-PE-CF594 (BD 562694), CD44-PE-Cy7 (BD 561283), CD45-BV510 (BD 563891), CD62L-BV605 (BD 563252), CD127-APC (Biolegend 121122), KLRG1-BV711 (BD 564014), FOXP3-PE (Thermo Fisher 12-5773-82), 4-1BB-PerCP-eF710 (Lifetech 46-1371-82), and Live/Dead Ghost dye 780 (Tonbo, San Diego, CA 13-0865-T100).

Statistical analysis

Group mean comparisons were performed using GraphPad Prism software, V.8.4.3. Analysis of variance (ANOVA) followed by Tukey’s multiple-comparison test was used to compare individual group means. In samples for which ANOVA was not applicable, the linear mixed effects model with Geisser-Greenhouse correction was used to compare group means among treatment groups. Survival analysis was conducted using a Mantel-Cox log-rank test. For all comparisons, p values ≤0.05 were considered statistically significant.

ResultsCombining TLR agonists with αCTLA-4, but not αPD-1, improved the efficacy of peptide-activated OT-I CD8+ T cells in suppressing E.G7 tumor growth

We previously demonstrated that a combination of TLR agonists, administered at the time of T cell activation by vaccination, could improve the antitumor activity of CD8+T cells.12 This treatment led to a decrease in PD-1 expression, and a concurrent increase in CTLA-4 expression, on vaccine-activated CD8+T cells. Consequently, we wished to evaluate whether combining TLR agonists with immune checkpoint blockade, αCTLA-4 or αPD-1, further affected T-cell mediated antitumor activity. E.G7 tumor cells (expressing ovalbumin and PD-L1) were implanted in C57BL/6 mice and permitted to grow until palpable (~7 days). OT-I splenocytes were then adoptively transferred, and the following day mice were immunized with 100 µg SIINFEKL (OVA) peptide alone, or peptide with TLR3 and TLR9 agonists. Mice were further treated with αPD-1 and/or αCTLA-4 the day following immunization, as indicated in figure 1A. The antitumor efficacy of vaccine-activated CD8+T cells was improved with αPD-1 and αCTLA-4 combined, but not either with either antibody alone (figure 1B,C, online supplemental figure 1A–C). The addition of TLR3 and TLR9 agonists to peptide immunization produced a greater suppression of tumor growth, and this was significantly increased with the further addition of αCTLA-4. However, the antitumor effect of TLR agonist and vaccine treatment was abrogated with the addition of αPD-1. As shown in figure 1D, the combination of TLR agonists and αCTLA-4 led to an increase in tumor-infiltrating antigen-specific CD8+T cells, and an increase in the CD8-to-CD4 Treg ratio. No differences were observed in the activation state (4-1BB expression) of tumor-infiltrating antigen-specific CD8+T cells following these different treatments (figure 1E), or the memory phenotype of these CD8+T cells (online supplemental figure 1E,F).

Figure 1

Combining TLR agonists with αCTLA-4, but not αPD-1, improved the efficacy of peptide-activated OT-I CD8+T cells in suppressing E.G7 tumor growth. (A) Ovalbumin-expressing E.G7 (eg,7-OVA-PD-L1high) cells were implanted in C57BL/6 mice and permitted to grow until tumors were palpable (7 days). OT-I splenocytes were then adoptively transferred and mice were immunized subcutaneously the following day with SIINFEKL (OVA) peptide alone, or with TLR agonists [TLR3 (Poly I:C) and TLR9 (ODN1826)]. Mice were further treated with αPD-1 and/or αCTLA-4 the day following immunization. (B) Shown are the tumor growth curves (mean+SE, n=7 animals per group). (C) Survival plots using the time to death or when tumors reached 2 cm3 in size, whichever occurred first. Data shown are from one of two independent studies (see online supplemental figure 2). (D) In a parallel study, animals were treated as in A, but tumors were collected at day 15 and evaluated for the frequency of infiltrating CD3+CD8+ tetramer+ T cells per gram of tumor and the number of Tregs (CD3+CD4+ CD25+ FoxP3+). (E) Tumor-infiltrating CD3+CD8+ tetramer+T cells were further evaluated for 4-1BB expression by flow cytometry. Asterisks indicate p<0.05 assessed by the mixed-effects model with Geisser-Greenhouse correction and Tukey’s multiple comparisons test with individual variances (B), by log-rank test (C), or by the one-way ANOVA with Tukey’s multiple comparisons test (D, E). Error bars represent SEM.

TLR agonists activated Tregs with lower PD-1 expression, and these Tregs were more functionally suppressive of CD8+ T cell proliferation and worsened the antitumor efficacy

We have previously shown that αPD-1 can increase the suppressive activity of Tregs.21 Given this, we tested whether TLR agonists affected the activation of Tregs and modulated PD-1 expression, as a possible explanation for why addition of αPD-1 might have abrogated the antitumor efficacy of vaccination with TLR agonists. Splenocytes from the spleens of OT-II mice were stimulated in vitro for 3 days with class II OVA peptide (OVA-II) or media alone in the presence or absence of TLR3 and TLR9 agonists. As shown in figure 2A, stimulation of splenocytes with TLR agonists significantly decreased the expression of PD-1 on Tregs. To determine whether Tregs conditioned with TLR agonists were more functionally suppressive of CD8+T cells, Tregs were purified from splenocytes and activated with OVA-II peptide and TLR agonists for 72 hours. They were then co-incubated with naïve OT-I CD3+CD8+T cells labeled with CFSE, in the presence of IL-2 and anti-CD3/anti-CD28 beads. Tregs stimulated with OVA-II peptide and TLR agonists significantly suppressed CD3+CD8+T cell proliferation compared with Tregs stimulated in the absence of TLR agonists, with OVA-II OVA peptide alone (figure 2B). In addition, these CD3+CD8+ T cells showed decreased effector function with reduced production of IFNγ, perforin, granzyme B, and TNFα (figure 2C). To determine whether the decreased antitumor efficacy of CD8+T cells observed following vaccination with TLR agonists and αPD-1 was dependent on Tregs, similar murine tumor studies were conducted using an αCTLA-4-targeting antibody that also depletes Tregs (αTreg).22 E.G7 tumor-bearing mice were treated as in figure 1A, but with immune checkpoint blockade including αPD-1, αCTLA-4, or αTreg. As shown in figure 2D and online supplemental figure 3, while treatment with the αTreg antibody alone with peptide vaccine had minimal effect, its use when combined with TLR agonists treatment significantly improved the antitumor efficacy and restored the antitumor efficacy when combined with αPD-1.

Figure 2

TLR agonists activated Tregs by lowering PD-1 expression, and these Tregs were more functionally suppressive of CD8+T cell proliferation and antitumor efficacy. Splenocytes were collected from the spleens of OT-II mice and stimulated in vitro for 3 days with class II OVA peptide (ISQAVHAAHAEINEAGR, OVA-II) or media alone (untreated) in the presence or absence of TLR agonists (TLR3+TLR9). (A) The median fluorescence intensity (MFI) of PD-1 was determined by flow cytometry in activated CD4+T cells (CD3+CD4+ CD25+ FoxP3-) and Tregs (CD3+CD4+ CD25+ FoxP3+). (B) Purified Tregs were coincubated with naïve OT-I CD3+CD8+ T cells labeled with CFSE, in a 1:1 ratio, in the presence of IL-2 and anti-CD3/anti-CD28 beads. Proliferation of CD3+CD8+ T cell was analyzed via flow cytometry by CFSE loss, and the % of cells with CFSE loss is shown. Upper panel shows representative flow cytometry data, and lower panel shows quantification (n=3 per condition). (C) Purified Tregs were coincubated with CD3+CD8+ cells from wild type mice in a 1:1 ratio for 48 hours, and analyzed for the expression of IFNγ, perforin, Granzyme B, and TNFα. Shown are the MFI of expression of these in CD3+CD8+ cells. (D) Treatment of mice with E.G7-OVA-PD-L1high tumors was conducted as described in (A). The day following peptide immunization (with or without TLR agonists), mice were treated with immune checkpoint blockade (or control IgG), using αPD-1, αCTLA-4, or a CTLA-4-targeting antibody that depletes Treg (αTreg). Animals receiving αTreg or αCTLA-4 received additional antibody treatment on days 11 and 13. Shown are the tumor growth curves (mean+SE, n=7 animals per group). Asterisks indicate p<0.05 assessed by the one-way ANOVA with Tukey’s multiple comparisons test (A and B) or by the mixed-effects model with Geisser-Greenhouse correction and Tukey’s multiple comparisons test with individual variances (C). Error bars represent SEM.

Combination of TLR agonists and αLAG-3 with peptide-activated CD8+T cells elicited prolonged suppression of E.G7-OVA-PDL1high tumor growth

In addition to PD-1 and CTLA-4, multiple other T cell checkpoint receptors are expressed on CD8+T cells following their activation, including LAG-3, TIM-3, TIGIT, and VISTA.12 We, therefore, next wished to determine whether combining TLR agonists with antibodies blocking LAG-3, TIM-3, TIGIT, or VISTA might similarly improve the antitumor efficacy of vaccine-activated CD8+T cells. Ovalbumin-expressing E.G7 tumor-bearing C57BL/6 mice were treated as shown in figure 1A. In this case, the immune checkpoint blockade used αTIM-3, αTIGIT, αVISTA, αLAG-3, or αCTLA-4, administered the day following immunization. As shown in figure 3A,B and online supplemental figure 4A–C, immune checkpoint blockade with peptide activation alone demonstrated little antitumor effect compared with controls. Combinations with αTIM-3 or αTIGIT showed no benefit over TLR agonist treatment alone. Combination with αVISTA was similar to treatment with αCTLA-4 in improving the antitumor effect relative to TLR agonist treatment alone. The greatest tumor suppression was observed when mice were treated with TLR agonists and αLAG-3. However, this treatment did not lead to significantly greater numbers of tumor-infiltrating antigen-specific CD8+T cells (figure 3C), CD8+T cells with greater activation (figure 3D), or differences in CD8+T cell memory phenotype (online supplemental figure 4D,E) compared with mice treated with vaccine, TLR agonists and CTLA-4 blockade.

Figure 3

Combination of TLR agonists and αLAG-3 with peptide-activated CD8+T cells elicited prolonged suppression of E.G7-OVA-PDL1high tumor growth. Ovalbumin-expressing E.G7 cells (eg,7-OVA-PD-L1high) were implanted in C57BL/6 mice and permitted to grow until tumors were palpable (7 days). As in figure 1A, OT-I splenocytes were then adoptively transferred and mice were immunized subcutaneously the following day with SIINFEKL (OVA) peptide, with or without TLR agonists. Mice were then treated with immune checkpoint blockade, using αTIM-3, αTIGIT, αVISTA, αLAG-3, αCTLA-4, or IgG control, the day following immunization. (A) Shown are the tumor growth curves (mean+SE, n=7 animals per group). (B) Survival plots using the time to death or when tumors reached 2 cm3 in size, whichever occurred first. (C) In a parallel study, animals were treated as in A but tumors were collected at day 16 and evaluated for the frequency of infiltrating CD3+CD8+ tetramer+ T cells per gram of tumor and the number of Tregs (CD3+CD4+ CD25+ FoxP3+). (D) Tumor-infiltrating CD3+CD8+ tetramer+ T cells were further evaluated for 4-1BB expression by flow cytometry. Asterisks indicate p<0.05 assessed by the mixed-effects model with Geisser-Greenhouse correction and Tukey’s multiple comparisons test with individual variances (A), by log-rank test (B), or by one-way ANOVA with Tukey’s multiple comparisons test (C, D). Error bars represent SEM. Results are from one experiment and are representative of data from three independent experiments (shown in online supplemental figure 5 and 6). ANOVA, analysis of variance; MFI, median fluorescence intensity; TLR, toll-like receptor.

Peptide-activated CD8+T cells had greater antitumor efficacy against E.G7 tumors when delivered with TLR agonists and αLAG-3 than with αPD-1 and αLAG-3

We have previously reported that the specific combination of αPD-1 and αLAG-3, when used together with vaccination, significantly increased the antitumor efficacy of cancer vaccines.23 We consequently wished to evaluate whether the combination of TLR agonists and αLAG-3 showed better vaccine-mediated antitumor efficacy than the αPD-1+αLAG-3 combination. Similar murine tumor studies with ovalbumin-expressing E.G7 were conducted as in figure 1A. As shown in figure 4A,B and online supplemental figure 7A–C, while the combination of vaccine with αPD-1 and αLAG-3 led to improved antitumor growth over controls, the combination of TLR agonists and αLAG-3 with vaccine demonstrated significantly greater antitumor growth control. This treatment was associated with slightly greater numbers of antigen-specific tumor-infiltrating CD8+T cells and improved CD8-to-Treg ratio (figure 4C), however, T cell activation (figure 4D) was not significantly different. Similarly, treatment was not associated with marked changes in CD4+T cells or the memory phenotype of CD8+T cells (online supplemental figure 7D,E). The treatment effect of TLR agonists and αLAG-3 required vaccine activation of antigen-specific CD8+T cells, since treatment with TLR agonists and αLAG-3 without peptide vaccine activation had no demonstrable antitumor activity (online supplemental figure 9).

Figure 4

Peptide-activated CD8+T cells had greater antitumor efficacy against E.G7 tumors when delivered with TLR agonists and αLAG-3 than with αPD-1 and αLAG-3. Ovalbumin-expressing E.G7 cells were implanted in C57BL/6 mice and permitted to grow until tumors were palpable (7 days). As in figures 1 and 3, OT-I splenocytes were then adoptively transferred and mice were immunized subcutaneously the following day with SIINFEKL (OVA) peptide, with or without TLR agonists. Mice were then treated with αPD-1 and/or αLAG-3 (or IgG control), the day following immunization. (A) Shown are the tumor growth curves (mean+SEM, n=7 animals per group). (B) Survival plots using the time to death or when tumors reached 2 cm3 in size, whichever occurred first. Data shown are representative of two independent studies (see online supplemental figure 8). (C) In a parallel study, animals were treated as in A but tumors were collected at day 16 and evaluated for the frequency of infiltrating CD3+CD8+ tetramer+T cells per gram of tumor and the number of Tregs. (D) Tumor-infiltrating CD3+CD8+ tetramer+T cells were further evaluated for 4-1BB and expression by flow cytometry. Asterisks indicate p<0.05 assessed by the mixed-effects model with Geisser-Greenhouse correction and Tukey’s multiple comparisons test with individual variances (A), by log-rank test (B), or by the one-way ANOVA with Tukey’s multiple comparisons test (C, D). Error bars represent SEM. ANOVA, analysis of variance; MFI, median fluorescence intensity; TLR, toll-like receptor.

Combination therapy of TLR agonists and immune checkpoint blockade improves the antitumor efficacy of a DNA vaccine in a murine prostate tumor model

The studies above using ovalbumin as a model tumor antigen indicated that combinations of TLR agonists and immune checkpoint blockade might further improve the function of CD8+T cells activated by vaccination. Consequently, we wished to evaluate this in a more relevant tumor model targeting a native antigen, effectively using TLR agonists and immune checkpoint blockade as vaccine adjuvants. We have previously reported that a DNA vaccine encoding the ligand-binding domain of the androgen receptor (pTVG-AR) can elicit antigen-specific CD8+T cells with cytolytic function.24 We have also demonstrated that coadministration of a DNA vaccine with TLR agonists, or with αPD-1 and/or αLAG-3, further improved the antitumor activity of vaccination.10 12 23 For these studies, MyC-CaP prostate tumor cells were implanted in male FVB mice. Mice were then immunized intradermally weekly, beginning the day after tumor implantation, with control vector (pTVG4) or pTVG-AR, and coadministered with TLR3 and TLR9 agonists. Mice were then treated with αPD-1, αCTLA-4, αLAG-3, IgG control, or combinations of antibodies, the day after each vaccination (figure 5A). As shown in figure 5B and online supplemental figure 10, the combination of vaccine with immune checkpoint blockade showed modest effects on tumor growth. In contrast, the combination of TLR agonists with αPD-1 showed inferior antitumor efficacy compared with vaccination with TLR agonists and IgG control, similar to our findings in EG.7 tumor-bearing mice. The combination of TLR agonists and αCTLA-4 or αLAG-3 further suppressed tumor growth and significantly prolonged survival (figure 5C). Compared with the group vaccinated with TLR agonists and IgG control, immunization with TLR agonists and αCTLA-4 led to an increase in the number of tumor-infiltrating CD8+T cells and an increased CD8-to-Treg ratio (figure 5D). This treatment also demonstrated an increased frequency of short-lived effector cells (SLEC) among CD8+T cells (online supplemental figure 10F). Immunization with TLR agonists and αLAG-3 demonstrated an increased expression of 4-1BB on tumor-infiltrating CD8+T cells (figure 5E).

Figure 5

Combination of TLR agonists and immune checkpoint blockade improves the antitumor efficacy of a DNA vaccine in a murine prostate tumor model. Male FVB mice were implanted subcutaneously with 106 MyC-CaP tumor cells. (A) Mice were immunized intradermally weekly beginning on day 1 with control vector (pTVG4) or DNA encoding AR ligand-binding domain (pTVG-AR), and delivered alone or with TLR agonists. Mice were then treated with αPD-1, αCTLA-4, αLAG-3, combinations of antibodies, or IgG control the day after each vaccination. (B) Shown are the growth curves for each group (mean+SEM, n=6–7 animals per group). (C) Survival plots using the time to death or when tumors reached 2 cm3 in size, whichever occurred first. Data shown are representative of two independent studies (see online supplemental figure 11). (D) In a parallel study, animals were treated as in A, but tumors were collected at day 33 and evaluated for the frequency of infiltrating CD3+CD8+ T cells per gram of tumor and the number of Tregs (CD3+CD4+ CD25+FoxP3+). (E) Tumor-infiltrating CD3+CD8+ T cells were further evaluated for 4-1BB expression by flow cytometry. Asterisks indicate p<0.05 assessed by the mixed-effects model with Geisser-Greenhouse correction and Tukey’s multiple comparisons test with individual variances (B), by log-rank test (C), or by one-way ANOVA with Tukey’s multiple comparisons test (D, E). Error bars represent SEM. ANOVA, analysis of variance; TLR, toll-like receptor.

Discussion

We have previously found that T cell activation in the presence of TLR agonists resulted in altered expression of multiple T cell checkpoint receptors and elicited greater vaccine-mediated antitumor efficacy.12 This indicated that further disrupting the signaling of T cell checkpoint receptors by combining TLR agonists with immune checkpoint blockade should elicit better antitumor function when used with cancer vaccines. In particular, we hypothesized that blockade of T cell checkpoint receptors whose expression was increased following TLR agonist stimulation (CTLA-4, TIM-3, VISTA and TIGIT) would improve the antitumor activity of CD8+T cells activated by vaccine and TLR agonist treatment. Consequently, we investigated whether combining TLR agonists (TLR3 and TLR9) and immune checkpoint blockade improved the antitumor function of CD8+T cells activated by vaccination. We found that combining TLR3 and TLR9 agonists with αCTLA-4, αLAG-3, or αVISTA led to further suppression of tumor growth, while combinations with αTIM-3 or αTIGIT showed no further improvement. Combining αPD-1 with TLR agonists unexpectedly resulted in greater tumor growth, and this was due to activation of CD4+Treg. Taken together, these results suggest (1) Depletion of CD4+Treg may be necessary when combining TLR agonists and αPD-1 to improve the antitumor efficacy of vaccines; (2) the change in expression of immune checkpoint receptors on CD8+T cells activated in the presence of TLR agonists may not predict the optimal immune checkpoint blockade to deliver with vaccine and TLR agonists; and (3) TLR agonists may be superior to αPD-1 when used in combination with other immune checkpoint blockade agents (such as αCTLA-4 or αLAG-3) and antitumor vaccines.

Our findings indicate that depletion of CD4+Treg may be necessary when combining TLR agonists and αPD-1 to improve the antitumor efficacy of vaccines. Tregs are known to infiltrate many types of cancers and serve as one of the major cell lineages that suppress antitumor immunity.25 26 Studies have reported that Tregs are responsible for ‘hyperprogression’ of cancer that has been observed in a subset of patients treated with αPD-1, suggesting that αPD-1 treatment can affect the function of Tregs.27–30 We have, in fact, previously demonstrated that αPD-1 can promote the activity of Tregs, resulting in suppression of CD8+T cell proliferation.21 Our in vitro studies used Treg from OT-II mice and identified that TLR agonists can also activate the suppressive function of Tregs, and this is associated with lower PD-1 expression on Tregs. Conceivably these Treg from OT-II mice could be different from endogenous Treg in EG.7 tumor-bearing mice in terms of PD-1 expression. While PD-1 expression was not evaluated in those studies, endogenous Tregs in MyC-CaP tumors showed a trend of lower numbers of Treg, but no differences in PD-1 expression on these Treg, following treatment with TLR agonists (online supplemental figure 10E). Separate from our studies, a recent study has shown that the balance of PD-1 expression on lymphocytes is important, and that patients with greater numbers of PD-1+Tregs relative to PD-1+CD8+ T cells had significantly decreased survival.31 Given our findings that TLR agonists elicited greater antitumor efficacy when delivered with vaccine without αPD-1, lowered PD-1 expression on both CD8+T cells and Tregs by TLR agonists might not have impacted the balance of PD-1 expression. At the same time, further addition of αPD-1 may have unfavorably modified the balance of PD-1 expression between Tregs and CD8+T cells. In any case, our findings suggest that depletion of CD4+Treg is necessary if these TLR agonists and αPD-1 are combined and delivered together with antitumor vaccines.

Our results also demonstrate that the change in expression of immune checkpoint receptors on CD8+T cells activated in the presence of TLR agonists was not associated with optimal immune checkpoint blockade when delivered with vaccine and TLR agonists. In a previous study, we found that T cells activated by vaccination along with TLR3 and TLR9 agonists resulted in upregulation of several T cell checkpoint receptors, notably CTLA-4, VISTA, TIGIT, and TIM-3, on CD8+T cells.12 Therefore, we hypothesized that blocking any of these receptors would further improve the antitumor efficacy of vaccination along with TLR agonists. The combination of TLR3 and TLR9 agonists with αCTLA-4 or αVISTA improved tumor growth suppression when used with vaccine, while combinations with αTIM-3 or αTIGIT did not. Interestingly, we found that blockade for LAG-3, a T cell checkpoint receptor which we had shown is downregulated in CD8+T cells (similar to PD-1) following activation in the presence of TLR3 and TLR9 agonists, elicited greater suppression of tumor growth when combined with TLR agonists and vaccine. The reason for these different outcomes following treatment with TLR agonists and different immune checkpoint blockade is not fully understood. However, this might be due to different biological functions of each T cell checkpoint receptor on T cells and its role during T cell activation. For example, CTLA-4 regulates proliferation of T cells32 while LAG-3 and TIGIT are more likely to engage in early steps in TCR signaling to attenuate T cell activation.33 34 TIM-3 signaling is known to coordinate the apoptosis of T cells following activation.35 More research will be needed to fully understand the function of immune checkpoint receptors and their blockade during vaccine-mediated activation. Nevertheless, our findings suggest that combinations of TLR3 and TLR9 agonists and specific immune checkpoint blockade, notably αCTLA-4, αLAG-3, and/or αVISTA, are of particular interest when combined with antitumor vaccines. Further studies are needed to determine if they can similarly modulate the function of human CD8+T cells.

Our findings suggest that TLR agonists (notably TLR3 and TLR9) may be superior to αPD-1 when used in combination with other immune checkpoint blockade (such as αCTLA-4 or αLAG-3), at least when these agents are delivered with a T cell activating treatment such as an antitumor vaccine. As demonstrated, this is likely due to the concurrent activation of CD4+Treg by αPD-1 treatment. Specifically, we observed that TLR3 and TLR9 agonists delivered with antigen-specific immunization and CTLA-4 or LAG-3 blockade led to better tumor suppression than vaccination delivered with a combined checkpoint blockade (either αPD-1+αCTLA-4 or αPD-1+αLAG-3). Notably, in these tumor models, there was little antitumor efficacy of combined immune checkpoint receptor blockade in the absence of vaccination, suggesting that similar approaches could be used for immunologically “cold” tumors, like prostate cancer, for which there is little therapeutic efficacy using immune checkpoint blockade alone. However, combined immune checkpoint blockade, notably with αPD-1 and αCTLA-4, has demonstrated remarkable antitumor efficacy for certain malignancies, with even greater efficacy than αPD-1 or αCTLA-4 treatment alone.36 Similarly, αPD-1 and αLAG3 combinations are being explored for multiple malignancies, and this combination has been demonstrated to be superior to αPD-1 monotherapy for patients with melanoma.37 However, combined immune checkpoint blockade combinations generally have more toxicity, and more than half of patients treated with αPD-1 and αCTLA-4 have experienced grade 3 or 4 adverse events.