WHAT IS ALREADY KNOWN ON THIS TOPIC

Androgen deprivation can work synergistically with androgen receptor targeted vaccination to prolong the time to progression and survival in murine prostate cancer models. However, the optimal sequence of this combination, and its effect on the tumor immune microenvironment, have not been previously evaluated.

WHAT THIS STUDY ADDS

In murine models of prostate cancer, we demonstrate that there is a sequence preference to the delivery of androgen deprivation and androgen receptor targeted vaccination, and this is mediated by differences in antigen-specific CD8+T cells and myeloid cells within the tumor immune microenvironment. The antitumor efficacy of this combination was improved by reducing the migration of Gr-1+myeloid derived suppressor cells using a CXCR2 antagonist or by intermittent use of androgen deprivation therapy (ADT).

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

The pTVG-AR vaccine is being evaluated in clinical trials. The current findings suggest that its optimal use, or potentially other anti-prostate tumor vaccines, will be prior to androgen deprivation, and this may be improved using approaches such as intermittent ADT or bipolar androgen therapy that may impede the migration or infiltration of immunosuppressive myeloid cells.

Background

Prostate cancer is the second leading cause of cancer-related death among men in the USA.1 Androgen deprivation therapy (ADT) plays a central role in the treatment of prostate cancer, particularly in cases of advanced or metastatic disease. The growth and progression of prostate cancer hinge on the androgen receptor (AR) pathway. ADT functions by suppressing androgens, such as testosterone, or inhibiting their binding to the AR, aiming to impede the growth of prostate cancer cells.2 However, despite its initial efficacy, a significant challenge emerges in the form of castration resistance, with a median life expectancy of less than 3 years for patients with metastatic castration-resistant prostate cancer.3

Early investigations demonstrated that ADT, achieved through surgical or chemical means, led to a reduction in prostate size and triggered tumor apoptosis.4 In addition to its direct antitumor effects, there is growing evidence demonstrating that ADT exhibits immunostimulatory properties, including thymic regrowth, increased production of naïve T cells, enhanced immune cell infiltration into the prostate (both lymphoid and myeloid cells), enhanced antigen presentation and elevated antibody responses to prostate antigens.5–8 Preclinical studies have illustrated that ADT can enhance the efficacy of various immunotherapeutic approaches such as anticancer vaccines, presenting a promising avenue for the strategic harnessing of both antitumor and immunostimulatory effects to improve immunotherapy outcomes in prostate cancer.9–13 For example, Ardiani et al showed that male mice, on vaccination with a yeast-based Twist-encoded vaccine, displayed initial Twist-specific CD4+T cell proliferation. Notably, the combination of Twist vaccination and enzalutamide further enhanced CD4+T cell proliferation compared with control or enzalutamide-treated mice. This effect was also observed in the TRAMP model, where mice receiving the combination of the Twist vaccine and enzalutamide exhibited increased overall survival compared with control mice or those treated with monotherapy.12 Kwilas et al also demonstrated that a Twist poxvirus vaccine, when administered with enzalutamide, led to a significant improvement in overall survival.9 Combining a poxviral vaccine targeting PSA, PROSTVAC, with ADT in a prostate-specific PSA transgenic mouse model demonstrated a significant increase in PSA-specific T cells and an increase in T-cell interferon (IFN)-γ production.14 In a phase 2 clinical study involving patients with PSA progression, the combination of PROSTVAC and nilutamide, with nilutamide preceding the vaccine, resulted in a median time to treatment failure of 5.2 months from the initiation of combination therapy. In contrast, for patients receiving the vaccine followed by nilutamide, the median time to treatment failure was 13.9 months from the start of combination therapy, suggesting a possible benefit related to the sequence of administration of these agents.15 16

Paradoxically, the use of ADT leads to increased expression of the AR as a compensatory mechanism.17 In fact, castration resistance is often accompanied by mutations or gene amplifications leading to increased AR expression.18–21 For this reason, we have previously evaluated the AR as a vaccine target, one that might be strategically targeted in combination with ADT. We previously demonstrated that ADT treatment of prostate cancer cells increased their recognition by AR-specific CD8+T cells.22 We also showed that a DNA vaccine encoding the ligand-binding domain (LBD) of the AR (pTVG-AR), given with or without ADT, could elicit antigen-specific CD8+T cells, slow prostate tumor development, delay tumor growth, and prolong overall survival in murine models of prostate cancer.22 23 In a multicenter phase I study conducted in patients who had recently initiated ADT, we demonstrated that vaccination with pTVG-AR led to increased AR-specific Th1-biased immunity and a delayed time to castration resistance in immunized patients.24

Given that others have demonstrated that ADT can be immunomodulatory,25 26 and there may be a sequence preference to its use with vaccination, in the current report we investigated the sequence of ADT given with the pTVG-AR DNA vaccine in preclinical murine models of prostate cancer. Our goal was to determine whether, any by what mechanism, there was a sequence preference to administration, and whether changes to the tumor immune microenvironment might inform optimal approaches that could be applied to future human clinical trials using this or other prostate cancer vaccines.

Materials and methodsCell lines

TRAMP-C1 and Myc-CaP cell lines were obtained from ATCC (Manassas, Virginia, USA) and maintained according to ATCC recommendations and tested for Mycoplasma contamination.

Mice

FVB/NJ mice (stock #001800) and C57BL/6 mice (stock #000664) were obtained from The Jackson Laboratory (Bar Harbor, Maine, USA) and housed in micro-insulator cages under aseptic conditions. A2/TRAMP mice were generated by crossing homozygous TRAMP mice and HHDII-DR1 mice.23 All animal studies were conducted under an institutional animal care and use committee (IACUC)-approved protocol.

Tumor implantation

FVB mice were implanted with 1×106 cells in phosphate-buffered saline (PBS) into the flank of 4–6 weeks old FVB mice, or 6–10 weeks old NRG mice, respectively. Similarly, wild type male C57BL/6J mice aged 4–6 weeks were injected subcutaneously into the flank with 1×106 TRAMP-C1 cells in 1:1 ratio in PBS: Matrigel (Corning, New York, USA. CB354248).

Tumor growth studies

Tumors were measured twice weekly via calipers until the tumors reached 2 cm3. Tumor volumes were calculated as (long axis×short axis2)/2. Mice were randomized to treatment groups and treatments were started when the tumor volume reached 0.1–0.2 cm3.

Androgen deprivation therapy

Mice were treated subcutaneously with either degarelix (25 mg/kg) (Ferring Pharmaceuticals, Parsippany, New Jersey, USA) or a vehicle sham (PBS) treatment every 28 days starting when the tumor volume reached 0.1–0.2 cm3.

DNA immunization

100 µg of pTVG4 or pTVG-AR (GeneScript, Piscataway, New Jersey, USA) per mouse was administered intradermally in the ear pinna of TRAMP-C1 or Myc-CaP tumor-bearing mice. The first dose of vaccine was given the day after tumor implantation, with additional doses given weekly afterwards in the pTVG4 → ADT and pTVG-AR → ADT groups. For ADT → pTVG4 and ADT → pTVG-AR groups, the immunizations were started 1 day after ADT administration. In the A2-TRAMP studies, immunizations were started when mice reached 16 weeks of age, for 4 weeks, followed by ADT administered every 28 days. In the pTVG4 →ADT and pTVG-AR →ADT groups, a boost was administered prior to ADT. Conversely, in the ADT →pTVG4 and ADT → pTVG-AR groups, immunizations were initiated after ADT, with subsequent boosts administered post ADT every 28 days.

Testosterone administration

Mice were administered 0.9 mg testosterone cypionate (Cipla USA, Warren, New Jersey, USA). This was resuspended in 100 µL of sesame oil and delivered subcutaneously twice a week for 2 weeks beginning 14 days post ADT treatment.27 28

CXCR2 antagonist

CXCR2 antagonist, reparixin (Selleckchem, Houston, Texas, USA), was reconstituted in Tween-80 and PBS in a 1:4 ratio and administered subcutaneously at 5 mg/kg on the left flank thrice a week for 3 weeks post ADT administration.

Clodronate liposomes

Mice were injected with clodronate liposomes intravenously, using 200 µL of 5 mg/mL suspension of clodronate liposomes (Liposoma clodronate liposomes, Amsterdam, the Netherlands), twice weekly for 3 weeks post ADT treatment. Control groups received empty liposomes at the same dose and schedule.29

Flow cytometry

Tumors were collected from mice at different time points following different treatments, then digested for 1–2 hours at 37°C in mouse cell culture medium: Roswell Park Memorial Institute 1640 medium (RPMI 1640) with L-glutamine, 10% fetal calf serum, 200 U/mL Pen/Strep, 5% sodium pyruvate, 5% HEPES, and 50 µM β-MeOH supplemented with 2 mg/mL collagenase, 0.2 mg/mL DNAse I, and one tablet protease inhibitor (Sigma-Aldrich, St. Louis, Missouri, USA, Cat# 11697498001) per 50 mL digest solution. Digests were then passed through 100 µm screens. 5×106 cells were plated and Fc blocked (BD, Franklin Lakes, New Jersey, USA, 553142) for 20 min at 4°C. Cells were then stained for 30 min at 4°C with the viability dye Ghost Dye Red 780 (Tonbo 13–0865 T100) and the following antibodies: CD11b-BB515 (BD 564454), CD25-BB700 (BD 566498), GR-1-PE-CF594 (BD 562710), CD3-PE-Cy7 (eBiosciences Thermo Fisher Scientific, Waltham, Massachusetts, USA 25-0031-82), MHCII-BV421 (BioLegend San Diego, California, USA 107632), CD45-BV510 (BD 563891), CD4-BV605 (BioLegend 100451), CD19-BV711 (BD 563157), CD11c-APC (BD 550261), CD8-AF700 (100730) and cells were then fixed and permeabilized with the eBioscience Foxp3/Transcription Factor Staining Buffer Set overnight at 4°C (Thermo Fisher 00-5523-00). Cells were then stained with intracellular antibodies for 30 min at 4°C: FoxP3-PE (Thermo Fisher 12-5773-82), Ki67-BV421 (BD 562899), TNF-α (BioLegend 506323), Granzyme B (Thermo Fisher 61-8898-82) Perforin-FITC (BioLegend 353309). Flow cytometry was performed on a Thermo Fisher Attune NxT cytometer and data were analyzed using FlowJo V.10 software. Gates were set according to fluorescence minus one control.

In vitro studiesELISpot

Measures of antigen-specific immune response were performed by IFN-γ ELISpot (Bio-Techne, R&D Systems, Minneapolis, Minnesota, USA). Briefly, CD8+T cells were harvested from tumors of treated mice using Easy EasySep Mouse CD8+T Cell Isolation Kit (STEMCELL Technologies, Cat# 19853) following ADT treatment and added together with splenocytes from naïve FVB mice as feeder cells in a 1:10 ratio. AR25 (SRMLYFAPDLVFNEY), a 15-mer AR-specific dominant CD8+peptide epitope in FVB mice, or a pool of 62 15-mer amino acids spanning the amino acid sequence of the AR ligand-binding domain, was used as a measure for antigen-specific response measured after 48 hours of antigen stimulation. Positive control included stimulating the CD8+T cells with αCD3/αCD28 coated beads (Thermo Fisher, Cat# 11456D) at a ratio of two beads per CD8+T cell. Experimental results are shown as the number of IFN-γ spot-forming units per 105 CD8+T cells normalized against media-alone control wells.

Statistical analysis

Tumor growth data were analyzed by fitting a linear mixed-effects model with Geisser-Greenhouse correction and used to compare group means among treatment groups. Flow cytometry data were analyzed via analysis of variance (ANOVA) followed by Tukey’s multiple-comparison test when comparing all conditions to baseline, or planned contrasts when comparing only two conditions at each time point. Survival analysis was conducted using a Mantel-Cox log-rank test. For all comparisons, p values≤0.05 were considered statistically significant with asterisks *p<0.05, **p<0.01, and ***p<0.001. For ELISpot analyses, comparisons between treatment groups and media only controls were made using a Student’s t-test, with p<0.05 defining a significant antigen-specific T-cell response. One-way ANOVA with Tukey’s multiple comparisons test was performed for time course experiments with asterisks indicating *p<0.05, **p<0.01, and ***p<0.001. All analysis was performed on GraphPad Prism V.10.0.0, GraphPad Software, Boston, Massachusetts, USA.

ResultsSequence of vaccination and ADT, with vaccination given prior to ADT, (pTVG-AR → ADT) significantly improved antitumor responses in murine prostate tumor models

We have previously reported that combining ADT with AR-targeted vaccination led to greater antitumor effects in murine models of prostate cancer.22 This led to a phase I trial, demonstrating the vaccine’s safety, its ability to elicit or augment T cells specific for the AR LBD, and its potential to delay castration resistance.24 Consequently, we wished to determine whether there was an optimal use or sequence of these treatments to efficiently delay or prevent the emergence of castration-resistant disease in murine prostate tumor models. Hence, as a first study, we examined AR-targeted vaccination delivered either before or after ADT. As depicted in figure 1A, Myc-CaP tumor cells were implanted subcutaneously in male FVB mice, and when tumors reached a volume of 0.2–0.3 cm3 they were treated with degarelix (25 mg/kg subcutaneously every 28 days). Mice were also immunized weekly with a DNA vaccine encoding the ligand binding domain of the androgen receptor (pTVG-AR), or control vector (pTVG4), beginning either 1 day after tumor implantation (before androgen deprivation), or 1 day after degarelix treatment. As shown in figure 1B (and online supplemental figure 1), immunization prior to ADT elicited a greater antitumor response and significantly improved survival (figure 1C). Similar results were found in a separate murine model of prostate cancer in which C57BL/6 mice implanted with TRAMP-C1 tumor cells (figure 1D); AR-targeted vaccination prior to ADT resulted in a significantly greater antitumor response and improved overall survival (figure 1E,F, and online supplemental figure 2). In this model, mice received an additional weekly vaccination prior to ADT, given the slower growth rate of these tumors requiring 4 weeks to become palpable. Similarly, transgenic HLA-A2-expressing TRAMP mice, mice that develop autochthonous prostate tumors with age driven by the SV40 large T antigen under a prostate-specific promoter, were immunized in sequences before or after ADT, as depicted in online supplemental figure 3A. Mice that received vaccination prior to ADT showed a non-significant trend toward improvement in overall survival (online supplemental figure 3B). Overall, these findings demonstrated that vaccination in combination with ADT had a stronger antitumor effect in combination and was dependent on the order of administration, with pTVG-AR → ADT leading to superior antitumor responses.

Figure 1

Sequence of vaccination and ADT, with vaccination given prior to ADT, (pTVG-AR → ADT) significantly improved antitumor responses in murine prostate tumor models. FVB mice were implanted with Myc-CaP tumor cells and treated with degarelix (ADT), with pTVG4 (vector control) or pTVG-AR delivered before or after ADT and followed for tumor growth. Shown is a schema (panel A), tumor growth curves (panel B), and Kaplan-Meier curves depicting survival (time to a tumor size of 2 cm3 or death, panel C). Similarly, male C57BL/6 were implanted with TRAMP-C1 tumor cells, treated with ADT, and with DNA immunization with pTVG4 or pTVG-AR initiated either before or after ADT. For tumor growth curves, asterisks demonstrate significant (*p<0.05, **p<0.01 **p<0.001) differences as assessed by linear mixed effects model with Geisser-Greenhouse correction and Tukey’s multiple comparisons test with individual variances. Kaplan-Meier curves were compared using the log-rank test with asterisks indicating *p<0.05, **p<0.01, and ***p<0.001. Results are each from one experiment, with n=7–8 animals per group, and are representative of two independent experiments (see online supplemental figure 1). ADT, androgen deprivation therapy; AR, androgen receptor.

ADT led to an increase in prostate tumor-infiltrating T cells and myeloid cells, and this was affected by vaccination

It has been previously demonstrated by us and others that ADT can lead to increases in tumor-infiltrating CD4+ and CD8+T cells.5 6 We next asked whether the sequence of immunization, with immunization given prior to ADT, affected the infiltration of tumors by T cells. A similar study was performed as in figure 1A, but tumors were collected at several time points following treatment for evaluation of immune cell compositions via flow cytometry, as shown in figure 2A. Representative flow plots for T cells on days 21 and 36 and myeloid derived suppressor cells (MDSCs) on days 36 and 60 are shown in figure 2B (and the gating strategy is shown in online supplemental figure 4). We found that CD4+T and CD8+T cells increased in the tumor microenvironment in all mice at day 36, 14 days after ADT administration, but with the pTVG-AR → ADT combination group, CD4+T and CD8+T cells were further increased (figure 2C,D) and significantly reduced by day 60. Gr-1+MDSCs were significantly reduced at day 36 while significantly increased in all groups following ADT regardless of vaccination (figure 2E); there were little to no changes in regulatory CD4+T cells (Tregs, figure 2F). Taken together, these data suggested that ADT initially resulted in increased numbers of infiltrating T cells, which were further augmented by immunization prior to ADT. However, the effects of vaccination were transient, and at a later time point after ADT the tumor microenvironment had many Gr1+MDSCs and few T cells.

Figure 2

ADT led to an increase in prostate tumor-infiltrating T cells and myeloid cells, and T-cell infiltration was increased with vaccination prior to ADT. Myc-CaP tumor cells were implanted in male FVB mice and treated with ADT, and DNA immunization was delivered before or after ADT. Additional groups received either vector (pTVG4) or vaccine (pTVG-AR) alone, without ADT. Tumors were sampled at different days for flow cytometry analysis. Shown are a schema (panel A) and representative dot plots (panel B) of CD4+CD3+ and CD8+CD3+ T cells (collected on days 21 and 36), or CD11b+Gr1+ MDSCs (panel H, collected on days 36 and 60). CD4+T cells (panel C), CD8+T cells (panel D), and CD11b+Gr-1+ MDSC (panel E) are shown as a percentage of CD45+cells. CD4+FoxP3+ Tregs (panel F) are shown as a percentage of CD4+T cells. Panels C–F were compared using one-way analysis of variance with Tukey’s multiple comparisons test; asterisks indicate *p<0.05, **p<0.01, and ***p<0.001. Results are from one experiment and are representative of two independent experiments. ADT, androgen deprivation therapy; AR, androgen receptor; MDSC, myeloid derived suppressor cell, Tregs, regulator T cells.

Immunization prior to ADT leads to antigen-specific immune responses

We next investigated whether immunization prior to ADT led to infiltration of antigen-specific T cells. A similar study was performed in which mice were immunized, treated with ADT, and tumors were harvested on day 37, 2 weeks following ADT (figure 3A). Tumor-infiltrating CD8+T cells were isolated and evaluated by IFN-γ ELISpot for AR-specific T cells (figure 3B,C). We found that CD8+T cells specific for AR, and notably for AR25, the dominant epitope for AR in FVB mice (online supplemental figure 5), were infiltrating tumors in greater numbers in mice immunized prior to receiving ADT (figure 3B). CD8 T cells recovered from tumors following immunization also expressed more granzyme B and perforin (online supplemental figure 6). ADT was required for this infiltration, because AR-specific T cells were not detected in greater abundance following vaccination without ADT (online supplemental figure 7). We next wished to determine whether CD8+T cells elicited with vaccination prior to ADT were specifically responsible for the improved antitumor efficacy. Mice were treated with IgG or αCD8 for 3 days in the week prior to receiving ADT (figure 3D). As shown in figure 3E (and online supplemental figure 8), CD8+T cell depletion after immunization and prior to ADT abrogated the delay in time to castration-resistant tumor growth and significantly worsened animal survival (figure 3F). Overall, these results indicate that CD8+T cells elicited and/or activated by vaccination prior to ADT are crucial in mediating enhanced antitumor responses.

Figure 3

Immunization prior to ADT led to infiltration of tumors by antigen-specific CD8+T cells. Myc-CaP tumor cells were implanted in male FVB mice, treated with ADT and DNA immunization as described before, and tumors were harvested 2 weeks after ADT (schema in panel A). CD8+CD3+T cells were isolated and interferon-γ ELISpot was performed with media alone, AR pool peptides, AR-25, or anti-CD3/anti-CD28 beads. Raw data are shown in panel B and quantified in panel C. Myc-CaP tumor cells were implanted in male FVB mice, treated with ADT and DNA immunization before or after ADT, and mice received IgG (control) or CD8 depleting antibody on days 16, 18 and 20, before ADT treatment (schema in panel D). Shown are the tumor growth curves (panel E) and Kaplan-Meier survival curves (panel F). Panel C was compared using unpaired t-tests. Tumor growth curve comparisons were made using one-way analysis of variance with Tukey’s multiple comparisons test. Kaplan-Meier curves were compared using the log-rank test. Asterisks indicate *p<0.05, **p<0.01, and ***p<0.001. Results are from one experiment and are representative of two independent experiments. ADT, androgen deprivation therapy; AR, androgen receptor; SFU, spot-forming units.

MDSCs increased following ADT led to impaired antitumor response to vaccination

The accumulation of MDSC following ADT has been previously demonstrated in Myc-CaP tumors,30 similar to what we demonstrated in figure 2E. MDSC are also known to accumulate in tumors and the peripheral blood of patients with advanced, castration-resistant prostate cancer, and have been implicated in the observation that advanced prostate tumors are relatively devoid of infiltrating T cells.31 We hypothesized that ADT-induced MDSC accumulation might impair the function or presence of tumor-infiltrating T cells, even following vaccination, ultimately contributing to the development of castration-resistant tumors. To evaluate the timing and effect of ADT on MDSC infiltration, Myc-CaP tumor-bearing mice were treated with or without ADT (figure 4A) and followed for tumor growth (figure 4B) and overall survival (figure 4C). Tumors were sampled from cohorts of mice with similar tumor volumes, as indicated in figure 4B, to evaluate for the presence of tumor-infiltrating immune populations. In animals not treated with ADT, there was an increase in tumor-infiltrating CD4+ (figure 4D) and CD8+ (figure 4E) T cells over time, but no significant increase in MDSCs (figure 4F). In contrast, in ADT-treated tumors, there was a marked increase in CD4+ and CD8+ T cells 12 days after ADT, but a marked decline thereafter (figure 4D,E). This was associated with a concurrent significant increase in CD11b+Gr-1+ MDSCs (figure 4F). We next sought to determine whether depletion of tumor-infiltrating MDSCs could enhance antitumor immune responses. As shown in figure 4G, tumor-bearing mice were treated with vaccine, ADT, and control-encapsulated or clodronate-encapsulated liposomes, a treatment effective at depleting precursors of MDSCs, twice a week following ADT. As shown in figure 4H (and online supplemental figure 9), treatment with clodronate liposomes significantly improved antitumor responses in mice receiving AR-targeted vaccination and ADT, in either treatment sequence, as well as overall survival (figure 4I). Tumors collected from mice treated with clodronate liposomes at day 62 demonstrated an increase in CD4+ (figure 4J) and CD8+ (figure 4K) T cells. These findings suggested that treatments aimed at interrupting the infiltration of MDSC following ADT might significantly improve the treatment effects of vaccination and ADT.

Figure 4

ADT leads to infiltration of MDSCs, and depletion of these populations improved antitumor responses in combination with DNA vaccines. Myc-CaP tumor cells were implanted in male FVB mice and treated with ADT or control as indicated in panel A. Shown are the tumor growth curves and time points at which groups of animals were euthanized for tumor assessments (panel B) and Kaplan-Meier survival curves (panel C). Tumors collected were evaluated by flow cytometry for CD4+T cells (panel D), and CD8+T cells (panel E) and CD11b+Gr-1+ MDSC (panel F). Myc-CaP tumor cells were implanted in male FVB mice, treated as before, and then received control-encapsulated or clodronate-encapsulated liposomes as indicated (schema in panel G). Shown are the tumor growth curves (panel H) and Kaplan-Meier survival curves (panel I). In a parallel study, tumors were collected at day 62 and evaluated by flow cytometry for CD4+T cells (panel J), and CD8+T cells (panel K) and CD11b+Gr-1+ MDSC (panel L). For tumor growth curves, asterisks demonstrate significant (p<0.05) differences as assessed by linear mixed effects model with Geisser-Greenhouse correction and Kaplan-Meier curves were compared using the log-rank test with asterisks indicating *p<0.05, **p<0.01, and ***p<0.001. For panels D–F and J–L, comparisons were made using one-way analysis of variance with Tukey’s multiple comparisons test; asterisks indicate *p<0.05, **p<0.01, and ***p<0.001. Results are from single experiments with n=5 (panels D–F) or 3 (panels J–L) mice per group. ADT, androgen deprivation therapy; AR, androgen receptor; MDSC, myeloid derived suppressor cell.

CXCR2 blockade improves antitumor efficacy in the pTVG-AR→ ADT combination by infiltrating antigen-specific CD8+T cells and inhibiting recruitment of Gr1+MDSCs

As clodronate liposomes can also deplete macrophages and do not selectively deplete MDSCs, we next wished to evaluate agents that more specifically impair MDSC recruitment or function. Blocking CXCR2 has been demonstrated to inhibit the recruitment of MDSC,32 and hence we specifically sought to determine whether the inclusion of a CXCR2 inhibitor could improve the antitumor responses of the combination treatment (figure 5A). As demonstrated in figure 5B (and online supplemental figure 10), mice treated with reparixin showed improved antitumor responses and overall survival (figure 5C). Tumors from mice treated in the pTVG-AR → ADT sequence exhibited a significant reduction in MDSCs (figure 5D) and a significant increase in CD8+T cells (figure 5F). Among tumor-infiltrating CD8+T cells present several weeks after ADT, AR-specific CD8+T cells were slightly but significantly increased in frequency (figure 5G–I).

Figure 5

CXCR2 blockade improved anti-tumor efficacy in the pTVG-AR →ADT combination. Myc-CaP tumor cells were implanted in male FVB mice, treated with vaccine, ADT and reparixin as indicated in panel A. Shown are the tumor growth curves (panel B) and Kaplan-Meier survival curves (panel C). A similar study was performed, and tumors were collected on day 56 and evaluated by flow cytometry for CD4+T cells (panel D), and CD8+T cells (panel E) and CD11b+Gr-1+ MDSC (panel F). In a parallel experiment, tumors were harvested on day 37 and CD8+CD3+T cells were isolated (schema in panel G). Interferon-γ ELISpot was performed with media alone, AR pool peptides, AR-25, or anti-CD3/anti-CD28 beads, with results as shown in panel H and quantified in panel I. For tumor growth curves, asterisks demonstrate significant (p<0.05) differences as assessed by linear mixed effects model with Geisser-Greenhouse correction and Kaplan-Meier curves were compared using the log-rank test with asterisks indicating *p<0.05, **p<0.01, and ***p<0.001. For panels D–F, comparisons were made using one-way analysis of variance with Tukey’s multiple comparisons test; asterisks indicate *p<0.05, **p<0.01, and ***p<0.001. Panel I was compared using unpaired t-tests. Results are from single experiments with n=5 mice per group. ADT, androgen deprivation therapy; AR, androgen receptor; MDSC, myeloid derived suppressor cell.

ADT followed by testosterone supplementation led to reduced accumulation of MDSCs although significantly worsened antitumor responses in the Myc-CaP tumor model

Our findings indicated that prolonged treatment with ADT caused early infiltration of T cells, but this was of limited duration and eventually tumors accumulated MDSCs. This suggested that it might be preferable to use a shorter course of ADT, or use it intermittently, in combination with vaccination. Because degarelix has a long half-life, we used supplemental testosterone to quickly reverse the effects of androgen deprivation. Hence, we sought to determine whether cycling ADT and testosterone might reduce the intratumoral accumulation of MDSCs. Male FVB mice were implanted with Myc-CaP tumor cells and administered ADT. Testosterone (or control) was administered twice a week, 2 weeks post ADT (figure 6A). As seen in the tumor growth curves, testosterone caused the tumors to grow significantly faster compared with mice treated with ADT alone (figure 6B and online supplemental figure 11) and shortened their overall survival (figure 6C). In a parallel experiment, tumors were harvested at post-treatment points to detect tumor infiltrating immune populations. Interestingly, we found that MDSCs were significantly reduced at day 66 following testosterone treatment (figure 6D), but there was not an associated increase in CD4+ or CD8+T cells (figure 6E,F).

Figure 6

ADT and testosterone reversal led to reduced accumulation of MDSCs although significantly worsened antitumor responses in the Myc-CaP tumor model. Myc-CaP tumor cells were implanted in male FVB mice, treated with ADT and testosterone as indicated in panel A. Shown are the tumor growth curves (panel B) and Kaplan-Meier survival curves (panel C). In a parallel study, tumors were collected on days 34, 46 and 66 and evaluated by flow cytometry for CD11b+Gr-1+ MDSC (panel D), CD4+T cells (panel E), and CD8+T cells (panel F). For tumor growth curves, asterisks demonstrate significant (p<0.05) differences as assessed by linear mixed effects model with Geisser-Greenhouse correction and Kaplan-Meier curves were compared using the log-rank test with asterisks indicating *p<0.05, **p<0.01, and ***p<0.001. For panels D–F, comparisons were made using one-way analysis of variance with Tukey’s multiple comparisons test; asterisks indicate *p<0.05, **p<0.01, and ***p<0.001. Results are from one experiment with n=5 mice per group. ADT, androgen deprivation therapy; AR, androgen receptor; MDSC, myeloid derived suppressor cell.

ADT followed by testosterone supplementation reduced accumulation of MDSCs and significantly improved antitumor responses when combined with vaccination in TRAMP-C1 tumor-bearing mice, but not Myc-CaP tumor-bearing mice

Given that an intermittent schedule of ADT could reduce MDSC accumulation, we next sought to determine if the addition of vaccination prior to ADT could improve the antitumor efficacy. Myc-CaP tumors were immunized intradermally weekly with pTVG-AR or vector control 1-day post-tumor implantation and weekly thereafter. Groups received ADT at day 22, and then received testosterone or control 14 days later, as shown in figure 7A. From the tumor growth curves, shown in figure 7B (and online supplemental figure 12), treatment with vaccination improved the antitumor response compared with vector control, however, treatment with testosterone led to faster tumor growth and shorter survival (figure 7C). As before, treatment with testosterone led to a significant reduction of MDSCs within tumors (figure 7D), but vaccination with ADT and testosterone did not lead to a greater infiltration of CD4+ and CD8+ T cells (figure 7E,F).

Figure 7

ADT and testosterone reversal led to reduced accumulation of MDSC and significantly improved antitumor responses when combined with vaccination in the TRAMP-C1 tumor model but not in the Myc-CaP model. Myc-CaP tumor cells were implanted in male FVB mice, and treated with ADT and DNA immunization with or without testosterone, as indicated in panel A. Shown are the tumor growth curves (panel B) and Kaplan-Meier survival curves (panel C). A similar study was performed, and tumors were collected on days 56 and evaluated by flow cytometry for CD11b+Gr