WHAT IS ALREADY KNOWN ON THIS TOPIC

Therapeutic cancer vaccines are an active area of investigation towards improving outcomes for patients with human papillomavirus (HPV)-associated malignancies, being frequently used in combination with other immunotherapeutic agents, particularly checkpoint inhibitors.

WHAT THIS STUDY ADDS

In this study, we employ a novel combination comprising an HPV16-specific vaccine with the immunocytokine NHS-IL12 and the class I histone deacetylase (HDAC) inhibitor Entinostat and demonstrate strong tumor control through heightened CD8+ T-cell functionality, reduced regulatory T cell infiltration, and altered polarization of macrophages in the tumor. The immune effects of each single agent and doublet were interrogated in comparison to triple therapy to understand how each agent contributes to the mechanism of action.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

This study provides rationale for the use of HPV vaccine, NHS-IL12, and a class I HDAC inhibitor such as Entinostat in the clinical setting for patients with HPV16-associated malignancies, including those with programmed cell death protein-1/programmed death ligand-1 refractory tumors. This study also demonstrates the complexity in dissecting the contribution of individual agents when used in triple therapy.

Introduction

An estimated 4.5% of all cancers worldwide (~630,000/year) are attributable to human papillomavirus (HPV).1 In the USA, the Centers for Disease Control and Prevention estimates >47,000 HPV-associated cancers are diagnosed each year.2 Though prophylactic HPV vaccinations have become more widespread in recent years, they are not expected to have a major impact on projected oropharynx cancer incidence in the USA through 2045.3 Additionally, as vaccination rates are higher in the Americas compared with other parts of the world, it will likely be even longer before prophylactic HPV vaccination demonstrates an impact on the incidence of HPV-associated malignancies worldwide.4

Currently, the standard of care (SOC) for HPV-associated malignancies includes surgical resection, chemotherapy, and radiation5; however, advanced stage HPV-associated cancers have poor prognosis and response to SOC.6 7 Despite improved outcomes with programmed cell death protein-1 (PD-1) targeted therapies, treatment resistance is common and presents a significant challenge. Modest overall response rates (ORR) from 17–26% in cervical cancer,8 9 to 24–32% in head and neck squamous cell carcinoma (HNSCC),10 11 and responses of less than 10% in patients with checkpoint refractory disease12 highlight a significant unmet need to develop novel therapies for these patients.

While >200 HPV strains have been identified, >70% of all HPV-related malignancies are attributed to the high-risk strains HPV16 or HPV18.1 13 HPV16 is associated with ~60% and 70% of cervical and HPV+ HNSCC, respectively.14 15 Due to high tumor expression of viral proteins E6 and E7, HPV-associated cancers are a prime target for therapeutic vaccines.16 Multiple HPV16-targeted vaccines are currently in Phase II/III clinical development.17 18 The liposomal-based vaccine PDS0101 (designated HPV vaccine), comprising R-DOTAP and HLA-unrestricted HPV16 E6 and E7 peptides,19 20 was shown to induce strong HPV16-specific responses in preclinical and clinical studies.21 22 In a Phase I study (NCT02065973), this HPV vaccine was shown to be safe and to induce multifunctional CD8+ and CD4+ T-cell responses specific to HPV16.22 The vaccine is currently involved in multiple Phase II clinical trials in combination with SOC (NCT04580771), checkpoint inhibition (NCT04260126, NCT05232851), or other immunotherapeutics (NCT04287868).

Recently, encouraging preclinical and clinical data have been observed with the HPV vaccine in combination studies with immunotherapeutic agents such as PDS01ADC (hereafter referred to as NHS-IL12). NHS-IL12 is a tumor-targeting immunocytokine consisting of two molecules of interleukin (IL)-12 fused to the histone binding antibody NHS76, which localizes to necrotic regions of the tumor.23 24 Produced primarily by antigen-presenting cells, IL-12 is an important cytokine in cancer immunotherapy by bridging innate and adaptive immunity. IL-12 plays a role both in T-cell priming and activation of effector responses, polarizing T cells towards a pro-inflammatory type 1 (Th1) response, driving interferon (IFN)-γ secretion, and promoting the establishment of effector memory T cells.25 26 NHS-IL12 demonstrated a good safety profile in patients with solid malignancies and reported encouraging clinical monotherapy activity despite a lack of objective responses.27 28 Preclinically, NHS-IL12 demonstrated increased antitumor efficacy when combined with SOC and immunotherapies.21 23 25 29 Triple combination of HPV vaccine, NHS-IL12, and the α-programmed death ligand-1/transforming growth factor, beta receptor II fusion protein bintrafusp alfa showed superior antitumor activity in preclinical models, with improved tumor T-cell infiltration and clonality.21 Clinical studies conducted with the trio yielded promising results with an ORR of 35% in all HPV16+ cancers, including 88% ORR in checkpoint naïve patients.30 However, the development of bintrafusp alfa has been discontinued; therefore, an alternative combination partner is needed to maximize the potential of HPV vaccine plus NHS-IL12.

Overexpression of histone deacetylases (HDACs) has been implicated in cancer formation and progression.31 Combining HDAC inhibitors with immunotherapy has shown promising antitumor activity in preclinical models.32–34 In particular, the class I HDAC inhibitor Entinostat created a favorable tumor microenvironment (TME) by upregulating antigen presentation machinery, suppressing myeloid-derived suppressor cells, as well as CD4+ regulatory T cells (Tregs) via FoxP3 downmodulation.33 35–37 Entinostat was shown to be safe and well tolerated in multiple clinical studies.38 Previous preclinical studies with Entinostat have shown strong synergy with sequential NHS-IL12 administration.32 34 Introductory treatment with Entinostat promoted tumor necrosis allowing increased NHS-IL12 to localize to the tumor.32 Combination therapy augmented CD8+ T-cell effector function and antigen specificity and induced favorable macrophage differentiation towards antitumor phenotypes.32

For the reasons outlined above, we hypothesized that triple therapy encompassing HPV vaccine, NHS-IL12, and Entinostat may result in strong antitumor effects. Since αPD-1 is currently part of SOC therapy for several advanced HPV-associated cancers, we also examined combining the double therapy of HPV vaccine and NHS-IL12 with αPD-1. This study is the first to examine a therapeutic cancer vaccine in combination with an immunocytokine and an HDAC inhibitor in the HPV setting and demonstrate a mechanism of action. Here, we report strong synergism between HPV vaccine, NHS-IL12, and Entinostat, resulting in prolonged survival in αPD-1 refractory tumor models of HPV-associated cancers. Using flow cytometry, single-cell RNA sequencing (scRNA-seq), and multiplex immunofluorescence, we show triple combination therapy increased CD8+ T-cell frequency and functionality and induced favorable macrophage differentiation resulting in a highly immune active pro-inflammatory TME.

Materials and methodsReagents

Under a Collaborative Research and Development Agreement (CRADA) with the National Cancer Institute (NCI), PDS0101 (HPV vaccine) and NHS-(murine) IL-12 (murine PDS01ADC) were obtained from PDS Biotechnology. Entinostat was provided by Syndax Pharmaceuticals, also under a CRADA with the NCI. Low-fat diet of 35% sucrose enriched with Entinostat for a target dose of 6 mg/kg/day was obtained from Research Diets. Anti-mouse PD-1 (clone RMP1-14-CP153), CD8α (clone 2.43) and CD4 (clone GK1.5) antibodies were purchased from Bio X Cell.

Cell lines

The HPV16 E6/E7+ lung cell line TC-1 was a kind gift from Dr T C Wu (Johns Hopkins University) and cultured as previously described.21 The HPV16 E6/E7+ murine oropharyngeal cell line mEER39 40 was a generous gift from Dr Clint Allen (NCI) and was cultured as recommended in media containing Iscove’s Modified Dulbecco’s Medium, Ham’s F12, 10% fetal bovine serum, 1% pen/strep, and 5 µg/L epidermal growth factor. All cell lines were used at low passage (<8) and tested negative for Mycoplasma (MycoAlert Kit, Lonza).

Mice

C57Bl/6 mice aged 6–8 weeks were obtained from the NCI Frederick National Laboratory for Cancer Research and co-housed under specific pathogen-free conditions with a 12 hours:12 hours light/dark cycle in rooms at 72°F±2°F and 30–70% relative humidity in an Association for Assessment and Accreditation of Laboratory Animal Care-accredited National Institutes of Health (NIH) animal facility. All studies were approved by the NIH Institutional Animal Care and Use Committee.

Murine tumor studies

TC-1 cells (5×104) or mEER cells (5×105; mixed 1:1 with Matrigel (Corning Life Sciences)) were injected subcutaneously (s.c.) on the right flank of female C57Bl/6 mice. When tumors reached an average of 50 mm3 (TC-1) or 120 mm3 (mEER), mice were randomized to low fat control or 6 mg/kg/day Entinostat diet for 10 days, with PDS0101 (300 µg R-DOTAP plus 40 µg HPV peptide mixture, s.c.) administered weekly for a total of three doses. Once control tumors reached an average of 200 mm3 (TC-1) or 240 mm3 (mEER), mice received the first of three NHS-IL12 doses (2 µg s.c.) given every other day. At least 70 days after tumor resolution, naïve and cured mice were implanted or re-challenged, respectively, with 5×105 mEER cells as described above. In select studies, αPD-1 (200 µg) was administered intraperitoneally (i.p.) on the same schedule as the PDS0101 vaccine. Tumors were measured twice weekly with digital calipers and volume was determined as (short diameter2×long diameter)/2. Tumor growth inhibition was calculated as (1 – (average tumor volume of treatment group / average tumor volume of control group)) × 100. Mouse body weights were recorded weekly to monitor health. In select studies, CD8 or CD4 depleting antibodies (100 µg/100 µL, i.p.) were administered on days 11, 12, and 13 after tumor implantation, followed by weekly injections for the duration of the experiment. Depletions were confirmed in peripheral blood via flow cytometry.

Flow cytometry

Staining for flow cytometry was conducted on ice using fluorescently labeled antibodies diluted in phosphate-buffered saline (PBS) supplemented with 1% bovine serum albumin, 10 mM HEPES, and 1 mM EDTA (online supplemental table 7). Intracellular staining was performed using the FoxP3/Transcription Factor Kit (eBioscience), according to manufacturer’s instructions. Live cells were identified using Live/Dead Fixable Dye (Thermo Fisher). HPV16 E7-specific CD8+ T cells were identified using a tetramer specific to the immunodominant murine epitope RF9 (RAHYNIVTF, catalog MHC-LC291, Creative Biolabs), whose human reactivity is unconfirmed.41 Cell counts were established using 123 count eBeads (Invitrogen). Data were acquired on an Attune Flow Cytometer (Thermo Fisher) or an LSRIIFortessa flow cytometer with FACSDiva V.9.0 software (BD Biosciences). Quality control was performed using peak extraction and cleaning-oriented quality control (PeacoQC)42 and data were analyzed using FlowJo V.10 (BD Biosciences) using the gating strategy shown in online supplemental table 8 and figure 6.

Statistical analysis

Statistical analyses were performed using GraphPad Prism V.10.2.3 (GraphPad Software) and are listed in figure legends; only significant differences are shown. Significant threshold was set at *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Data availability

Single-cell RNA and T-cell receptor (TCR) sequencing data were deposited in the Gene Expression Omnibus (GEO) database (GSE295238). All other data are available on reasonable request.

Code availability

All scRNA-seq data analysis and visualization were performed using the NIH Integrated Data Analysis Platform using R programs developed on the Foundry platform. All scripts used will be available on GitHub at the time of publication.

ResultsAddition of Entinostat versus addition of αPD-1 to PDS0101 vaccine plus NHS-IL12 yields superior tumor control against HPV16+ TC-1 tumors

To examine the antitumor effects of triple therapy encompassing HPV vaccine plus NHS-IL12, we compared the addition of Entinostat or αPD-1 to the dual-agent combination in αPD-1 refractory43 44 TC-1 tumor-bearing mice (figure 1A–C). An initial dose of HPV vaccine was given at treatment onset, followed by weekly boosts, with αPD-1 administered on the same schedule. Consistent with prior studies, a lead-in of Entinostat was given prior to administration of NHS-IL12 to promote increased deposition of NHS-IL12 in the TME.32 34 As NHS-IL12 contains the human NHS76 antibody, only a single cycle of therapy can be examined in syngeneic mouse models. Thus, NHS-IL12 was administered around the second vaccine dose to promote increased immune activation. Combining HPV vaccine and NHS-IL12 with either Entinostat or αPD-1 induced strong tumor control and increased survival compared with PBS control, although tumor escape was observed on therapy cessation on day 23 (figure 1B–C). Entinostat-containing combination therapy resulted in complete tumor resolution in 4/19 mice and improved survival compared with αPD-1 combination therapy. To better understand the differences between these two combination therapies, immune analysis was conducted 23 days after tumor implant. In the TME, both therapies increased CD8+ T-cell infiltration and HPV16-specific (RF9+) CD8+ T cells. However, only Entinostat combination therapy reduced tumor infiltration of Treg, yielding greatly improved CD8+:Treg and RF9+CD8+:Treg ratios (online supplemental figure 1A,B). Both combination therapies elicited strong increases in CCL5, CXCL9 and other T-cell recruiting chemokines in the tumor, with protein levels noted higher with Entinostat combination (online supplemental figure 1C). Analysis of splenic CD8+ T cells indicated a larger increase in the frequency of CD8+ and proliferating (Ki67+) CD8+ T cells with Entinostat combination compared with αPD-1-based therapy (online supplemental figure 1D). Similar to tumor chemokines, both combination therapies induced heightened IFN-γ, tumor necrosis factor (TNF)-α, and IL-2 serum levels, most elevated with Entinostat-containing therapy versus αPD-1-containing therapy (online supplemental figure 1E). Moreover, αCD3/αCD28 TCR stimulation ex vivo of splenic CD8+ T cells significantly increased IFN-γ and/or TNF-α production with the Entinostat combination, while the αPD-1 combination yielded higher levels of PD-1+ Tim-3+ CD8+ T cells, suggestive of increased exhaustion (online supplemental figure 1F). Collectively, these data demonstrate that both combination therapies elicit strong antitumor effects over a single cycle of therapy, although a more favorable antitumor immunity and survival benefit was noted with Entinostat-based therapy. Thus, we proceeded to further examine the combination of HPV vaccine, NHS-IL12, and Entinostat, hereafter referred to as triple therapy.

Figure 1

Combination therapy of PDS0101 (V), NHS-IL12 (N), and Entinostat (ENT (E)) induces potent tumor suppression. (A) Treatment timeline: 5×104 TC-1 cells were implanted s.c. in the right flank of female C57Bl/6 mice. When tumors averaged ~50 mm3, mice were randomized to Entinostat (6 mg/kg/day) or control chow for 10 days and the first of three weekly doses of PDS0101 (100 µL s.c.) and αPD-1 (200 µg i.p.) were administered. NHS-IL12 (2 µg s.c.) was administered every other day for a total of three doses when tumors averaged 200 mm3. PBS (100 µL s.c.) was administered to control mice on the same schedule as the PDS0101 vaccine and NHS-IL12. (B) Individual growth curves and (C) survival from two independent experiments (n=14–19 mice/group); inset denotes median overall survival (mOS, days). (D) Treatment schematic for subsequent studies of PDS0101, NHS-IL12, and Entinostat combination therapy. (E) Aggregate (mean±SEM) and percentage of tumor growth inhibition (% TGI), plus (F) individual tumor TC-1 growth curves over one cycle of therapy from two independent studies (n=16–17 mice/group). (G) Number of mice in each treatment cohort with decreasing tumor volumes or tumors <400 mm3 at 30 days post TC-1 tumor implantation, and (H) Kaplan-Meier plot with mOS (days) from two independent studies (n=14 mice/group). (I) Survival data for triple combination therapy in the mEER tumor model (n=8–9 mice/group). Tumor volumes: two-way analysis of variance with Tukey’s multiple comparisons test. Survival: Mantel-Cox test; only significant differences shown. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. i.p., intraperitoneally; PBS, phosphate-buffered saline; s.c., subcutaneously.

Triple combination therapy with HPV vaccine, NHS-IL12, and Entinostat yields potent antitumor activity in murine HPV16+ tumors refractory to αPD-1 therapy

To dissect the tumor-suppressive ability of triple combination therapy encompassing HPV vaccine, NHS-IL12, and Entinostat against HPV16+ tumors refractory to αPD-1 therapy,43 44 TC-1 tumor-bearing mice were treated as described (figure 1D). As the human nature of the NHS antibody in the NHS-IL12 molecule restricts these studies to a single cycle of therapy, tumor control was used as the primary endpoint. Significant tumor growth inhibition of 69.8% was observed after one cycle of therapy (day 24) in mice treated with triple combination, superior to that observed with HPV vaccine plus Entinostat (46.1%) or NHS-IL12 (54.6%), or Entinostat combined with NHS-IL12 (67.3%) (figure 1E). Analysis of tumor control in individual mice at the end of a single cycle of therapy demonstrated 88.2% of mice treated with triple therapy with decreasing tumor volumes, significantly higher than that observed in mice treated with Entinostat+NHS-IL12 (52.9%, p=0.0239), HPV vaccine+NHS-IL12 (5.9%, p<0.0001), and HPV vaccine+Entinostat (0%) (figure 1F–G, online supplemental table 1). At the end of a single treatment cycle, triple therapy yielded the highest frequency of tumors <400 mm3 (88.2%), approaching significance versus Entinostat+NHS-IL12 (58.8%, p=0.0519), and significantly higher versus all other treatment cohorts (figure 1G, online supplemental table 2). Despite a single cycle of therapy, we sought to evaluate its effects on survival to test the potential of triple therapy for clinical translation, where on a planned clinical study, patients will receive multiple cycles of therapy. A single cycle of triple therapy led to a significant increase in overall survival compared with all treatment groups except Entinostat+NHS-IL12 (figure 1H). A trend towards increased survival with triple therapy versus Entinostat+NHS-IL12 doublet was observed in a meta-analysis of three independent studies after a single cycle of therapy (p=0.0888, online supplemental figure 1G). Similar results were observed in mEER tumor-bearing mice, where triple therapy elicited full tumor clearance and full protective memory on tumor rechallenge, not observed with any doublet or monotherapies (figure 1I, online supplemental figure 2). Collectively, these data demonstrate that the combination of all three agents is required for the highest tumor control and survival benefit in αPD-1 refractory HPV16+ tumor models.

Analysis of the frequency and functionality of CD8+ T cells in the tumor

Next, we sought to identify the mode of action driving the antitumor effects observed with Entinostat triple therapy in TC-1 tumor-bearing mice. Immune depletion studies demonstrated CD8+ but not CD4+ T cells to be determinant to the antitumor efficacy of triple therapy (online supplemental figure 3). To gain further insight into the role of CD8+ T cells and the immune-mediated mechanisms driving tumor response during one cycle of triple therapy, tumors treated as described (figure 1D) were harvested 21–24 days after implant (10–13 days after treatment onset) and examined via flow cytometry and multiplex immunofluorescence (days 22 and 24), and scRNA-seq (day 21). All combination therapies significantly increased the frequency of CD8+ T cells to similar levels as detected by flow cytometry (figure 2A, online supplemental table 3), while HPV vaccine+NHS-IL12 double and triple combination therapies demonstrated the largest increases in CD8+ T-cell frequency, associated with significant transcriptomic and proteomic enrichment in chemokines associated with T-cell recruitment (online supplemental figure 4A,B). Triple therapy induced the strongest reduction of tumor Treg by scRNA-seq (figure 2B, online supplemental table 4). Transcriptomic analysis of the tumor immunome demonstrated each single agent to modestly increase the frequency of CD8+ T cells, with corresponding decreases in Treg and natural killer (NK) cell frequencies. NK cell frequency was further reduced with Entinostat+NHS-IL12 and HPV vaccine+Entinostat doublets (figure 2B, online supplemental table 4). Reduction of CD4+ T cells and Treg in the tumor with triple combination therapy was confirmed with multiplex immunofluorescence and flow cytometry (online supplemental figure 4C-G). Transcriptomic analysis of CD8+ T cells present in triple therapy tumors revealed enrichment in activated pathways associated with T-cell activation, migration, cytotoxicity, and effector function compared with PBS control (figure 2C). CD8+ T cells from triple therapy tumors also displayed increased gene expression of granzymes and Ifng, and reduced expression of the exhaustion-associated transcriptional factor Tox (figure 2D). Granzyme expression was driven primarily by Entinostat and augmented with NHS-IL12. While Tox reduction was seen in all combination cohorts receiving NHS-IL12, the immunocytokine alone had minimal effect on Tox expression. Importantly, all three agents were necessary for maximal Ifng expression with single agent and double therapies showing only minor increases in expression. Significant elevation in soluble tumor IFN-γ protein levels was confirmed by LEGENDplex cytometric bead array (figure 2E). Of note, NK cells were also found to express Ifng by scRNA-seq (online supplemental figure 5). HPV vaccine+NHS-IL12, Entinostat+NHS-IL12, and triple therapy treatment cohorts all displayed significantly elevated levels of IFN-γ protein in the tumor, with the highest levels seen in Entinostat+ NHS-IL12 and triple therapy. Granzyme B+CD8+ T cells were also confirmed to be significantly elevated with triple combination therapy by flow cytometry and multiplex immunofluorescence (figure 2F–G). Of note, while all treatment cohorts receiving more than one agent showed an overall increase in CD8+ T-cell infiltration (figure 2A), only the triple therapy resulted in a significant increase in granzyme B+CD8+ and granzyme B+ effector memory CD8+ T cells in the tumor (figure 2F), suggesting all three agents are required to achieve maximal T-cell functionality. Similar effects were observed in peripheral granzyme B+ CD8+ T cells in mEER tumor-bearing mice (online supplemental figure 2E). These data indicate the cooperative effects of Entinostat, NHS-IL12, and HPV vaccine to potentiate CD8+ T-cell function in the TME conducive to tumor control and clearance.

Figure 2

Improved CD8+ T-cell function with triple combination therapy. (A) Frequency of tumor CD8+ T cells per treatment cohort determined by flow cytometry. (B) Frequency of tumor lymphocyte populations per treatment cohort determined by scRNA-seq. (C) Select activated GO, Hallmark, C7, KEGG, and Reactome pathways in CD8+ T cells from tumors treated with triple therapy compared with PBS control by scRNA-seq. (D) Heatmap of gene expression in tumor CD8+ T cells determined by scRNA-seq. (E) Tumor IFN-γ protein concentration per treatment cohort. (F) Frequency of GzmB+ CD8+ T cells and GzmB+ CD8+ TEM in the tumor determined by flow cytometry. (G) Representative immunofluorescence staining (40×) for tumor CD8 and GzmB from eight PBS controls and six triple therapy-treated tumors. Graph shows quantification of GzmB+ CD8+ cell density in PBS controls and triple therapy tumors; Mann-Whitney test. Flow cytometry data and tumor IFN-γ concentration show median of eight mice/group from two independent studies; statistical differences determined by Kruskal-Wallis test with Dunn’s multiple comparisons test unless otherwise noted; only significant differences shown. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ENT (E), Entinostat; GzmB, granzyme B; IFN, interferon; KEGG, Kyoto Encyclopedia of Genes and Genomes; N, NHS-IL12; NK, natural killer; PBS, phosphate buffered saline; TEM, effector memory T cells; TIL, tumor-infiltrating lymphocytes; Treg, regulatory T cells; V, PDS0101 vaccine.

Cytolytic T-cell transcriptome analyses in the TME

To further interrogate the mechanism of action associated with CD8+ T-cell responses elicited by each agent and combination, sequencing of TCR gene clonotypes was performed at the single-cell level. Clones were defined as CD8+ T cells with matching TCR alpha and beta sequences. Compared with PBS control, all treatment cohorts displayed increased clonality (fewer clones/100 cells) and clonal expansion (larger number of clones present in >1 cell, online supplemental table 5).

To examine phenotypic and functional differences between clones, CD8+ T cells from all treatment cohorts with a paired TCR alpha/beta sequence were filtered and re-clustered. Expression of key phenotypic markers allowed for each cluster to be assigned a T-cell transcriptional profile (figure 3A–B).45 Most clusters predominantly expressed genes associated with exhaustion (Pdcd1, Ctla4, Havcr1, Lag3, and Tigit; Tex_1-6. One cluster had very high expression of Prf1 and multiple granzymes and was thus classified as cytotoxic T lymphocyte-like (TCTL), while another expressing high levels of Mki67 and Gins2 was composed of proliferating CD8+ T cells. Three clusters displayed gene expression profiles consistent with differential developmental states. Stem-like or naïve CD8+ T cells (Tstem/naïve) were characterized by heightened expression of Il7r, Tcf7, and Sell along with Ccr7 and Ly6a expression. CD8+ T cells with high levels of Ccr7 and Il7r and moderate expression of Tcf7 were classified as naïve/central memory-like (Tnaïve/CM), while effector memory-like T cells (TEM) displayed elevated Il7r and Gzmk expression in the absence of Ccr7 and Tcf7.

Figure 3

Triple therapy increased cytotoxic transcriptional profile in top tumor CD8+ T-cell receptor (TCR) clones. (A) UMAP shows all tumor CD8+ T-cell clusters with paired TCR α and β sequences. (B) Dot plot shows expression levels of key transcriptional phenotypic markers by cluster. (C) Frequency of specific CD8+ T-cell transcriptional profiles in the top five clones (designated A–E) for each treatment group. ENT (E), Entinostat; N, NHS-IL12; PBS, phosphate buffered saline; Tctl, cytotoxic T lymphocyte-like; Tem, effector memory-like T cells; Tex, exhuasted T cells; Tnaive/cm, naïve/central memory-like; Tstem/naive, stem-like or naïve CD8+ T cells; Tprol, proliferating CD8+ T cells; UMAP, Uniform Manifold Approximation and Projection; V, PDS0101 vaccine.

An examination of the most prominent clones in each treatment cohort revealed the highest enrichment of TCTL with triple therapy (figure 3C). Only three clones were present in more than one cell with PBS control treatment, with the predominant transcriptional profile being Tnaïve/CM. While NHS-IL12 treatment predominantly resulted in Tex2 (high in Tox), HPV vaccine and Entinostat monotherapies induced a notable but modest presence of cells with the TCTL transcriptional profile. Combination therapy with HPV vaccine plus NHS-IL12 or Entinostat also yielded minor expression of TCTL, though exhaustion profiles were most common. HPV vaccine+Entinostat also resulted in a high frequency of TEM in the most prevalent clone. An expansion of TCTL was observed in the most prominent clone in the Entinostat+NHS-IL12 doublet; however, only triple therapy demonstrated >50% of cells in each of its top five clones exhibiting the TCTL transcriptional profile. This highlights a marked increase in functionality achieved only with the combination of all three agents.

Triple therapy increased HPV16 E7+-specific CD8+ T-cell responses, and expansion of a cytotoxic phenotype clone exclusive to vaccinated cohorts

To better assess the role of vaccination in the antitumor response, we examined the specificity of CD8+ T cells in the TME. The HPV vaccine contains the murine HPV16 E7 RF9 epitope. As expected, all vaccinated cohorts displayed a significant increase in the frequency of RF9+CD8+ T cells (figure 4A). A trending increase in RF9 antigen-specific CD8+ T cells was also observed with Entinostat+NHS-IL12 and HPV vaccine+Entinostat doublet therapies.

Figure 4

Triple combination therapy increased tumor antigen-specific CD8+ T cells with increased functionality. (A) Frequency of HPV16 E7 (RF9+) antigen-specific tumor CD8+ T cells per treatment cohort determined by flow cytometry. Graph shows median of 8 mice/group from two independent studies, Kruskal-Wallis test with Dunn’s multiple comparisons test; only significant differences shown. (B) Frequency of tumor CD8+ T-cell transcriptional profiles in the top 10 T-cell receptor clones in all treatment groups. (C) Frequency of tumor CD8+ T-cell transcriptional profiles in vaccine-induced clone (clone 2) by treatment group. (D) Dot plot of expression levels of key gene markers in the vaccine-induced clone (clone 2) by treatment group. *p<0.05, **p<0.01, ***p<0.001. ENT (E), Entinostat; N, NHS-IL12; PBS, phosphate buffered saline; Tctl, cytotoxic T lymphocyte-like; Tem, effector memory-like T cells; Tex, exhausted T cells; Tnaive/cm, naïve/central memory-like; Tstem/naive, stem-like or naïve CD8+ T cells; Tprol, proliferating CD8+ T cells; V, PDS0101 vaccine.

To further interrogate antigen specificity in the tumor, CD8+ TCR clones across all treatment cohorts were examined. Whereas most clones were found to be unique to each therapy, the second most prominent clone across all samples (clone 2, figure 4B) was exclusive to vaccine-treated cohorts, suggesting clone 2 to be likely induced by the vaccine. Frequency analysis by treatment cohort revealed clone 2 to be enriched with triple therapy and strongly favored the TCTL expression profile (figure 4C). A closer look at the gene expression of clone 2 expressing cells across all vaccinated cohorts confirmed heightened expression of granzymes and Prf1 along with increased Ifng, and reduced levels of exhaustion markers (Ctla4, Lag3, Tigit, Tox, figure 4D). Notably, clone 2 was the most prominent TCR clone elicited by triple therapy (clone A in figure 3C, V+N+E graph). Together, this suggests triple therapy increases the frequency and cytotoxic capability of T-cell clones induced by the vaccine.

Role of macrophages in antitumor response

We next examined the impact of each agent on tumor-associated macrophages (TAMs). Overall macrophage infiltration was significantly reduced with triple combination therapy (figure 5A). However, macrophage polarization was found to shift from a pro-tumor, immunosuppressive, M2-like phenotype towards a favorable antitumor, pro-inflammatory, M1-like phenotype (figure 5B–C). Single agents demonstrated little to no effect on macrophage polarization. Combination of HPV vaccine + Entinostat or Entinostat + NHS-IL12 yielded modest increases in M1-like macrophages, with a large population of M2-like macrophages remaining. In contrast, HPV vaccine + NHS-IL12 led to a dramatic reduction in the frequency of M2-like macrophages; however, the increase in M1-like macrophages remained modest. Triple therapy resulted in a similar reduction in M2-like macrophages, but with a heightened level of M1-like macrophages (figure 5C). While there was no significant difference in M1:M2 ratio between the doublets and triple therapy when assessed via flow cytometry (M1: CD38+, M2: CD206+), the expanded number of genes used for phenotypic determination via scRNA-seq (online supplemental table 6) allowed for a more precise view of the changes to macrophage polarization, demonstrating the strongest shift with triple therapy. Closer examination of transcriptional states in TAMs by scRNA-seq revealed M1-like macrophages (figure 5D), M2-like macrophages (figure 5E), and other (non M1-like or M2-like) macrophages (figure 5F) to be enriched in activated pathways associated with inflammatory response and immune activation with triple therapy. M2-like macrophages and other macrophages also displayed activation of antigen presentation pathways (figure 5E–F). Interestingly, M2-like macrophages were found to have activated glycolytic pathways (figure 5E). As glycolysis in macrophages is strongly associated with a pro-inflammatory M1-like phenotype,46 47 this suggests these macrophages expressing key M2-like markers (Mrc1, Cd163) may be transitioning towards a more antitumor M1-like phenotype. Gene expression analysis of key transcriptional phenotypic markers across all macrophages and monocytes showed the strongest expression of pro-inflammatory M1-like markers (Cd38, Nos2, Hif1a, Il1b, Sell) with triple therapy (figure 5G). Triple therapy macrophages and monocytes also had reduced proliferation (Mki67) and increased chemokine expression (Ccl5, Cxcl9, Cxcl10, Cxcl1, Cxcl2, Cxcl3), indicating a more proinflammatory, immune active TME supportive of T-cell recruitment.

Figure 5

Triple therapy reduced tumor-associated macrophages (TAM) and modulated their biology towards a more tumor-suppressive gene expression profile. (A) TAM infiltration and (B) M1:M2-like TAM ratio quantified by flow cytometry. Graphs show the median of 8 mice/group from two independent studies. Kruskal-Wallis test with Dunn’s multiple comparisons test, only significant differences shown. (C) TAM polarization and M1:M2-like ratio determined by single-cell RNA sequencing. (D–F) Select GO, Hallmark, C7, KEGG, and Reactome pathways activated with triple therapy compared with PBS control in (D) M1-like macrophages, (E) M2-like macrophages, and (F) other macrophages. (G) Expression of key gene signatures in all tumor macrophages and monocytes by treatment group. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ENT (E), Entinostat; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; N, NHS-IL12; PBS, phosphate buffered saline; V, PDS0101 vaccine.

To gain a deeper insight into myeloid cells in the tumors and how each agent contributes to the overall phenotype, a trajectory analysis was performed on all macrophages and monocytes (figure 6). The red highlighted pathway in figure 6A shows the evolution of macrophage cell states from the lowest to highest pseudotime, denoting transcriptional transition from a pro-tumor M2-like state (cluster 1) to an antitumor M1-like phenotype (cluster 5) via clusters 4-0-2 (figure 6A–B). Examining the expression of a broad range of gene markers across all clusters confirmed cluster 5 to have the highest expression of genes associated with an antitumor, pro-inflammatory M1-like transcriptional state, including Cd38, Nos2, Il1b, and Tnf (figure 6C). We next compared the frequency of each cluster between treatment cohorts to determine how therapy affected the development and maturation of tumor macrophages (figure 6D). Entinostat monotherapy yielded no noticeable changes in cluster frequency. While NHS-IL12 and HPV vaccine as single agents led to reductions in cluster 1, there were no notable increases in antitumor pro-inflammatory macrophage clusters, indicative of a small shift in macrophage maturation within the M2-like and intermediate clusters. HPV vaccine+Entinostat also showed no notable changes in cluster distribution, while Entinostat+NHS-IL12 yielded a modest increase in cluster 5 while retaining the prominent cluster 1 population. Both HPV vaccine+NHS-IL12 and triple therapy yielded dramatic reductions in cluster 1 and other low pseudotime M2-like clusters. However, despite the limited effect of any of the monotherapies, triple therapy resulted in a prominent and exclusive enrichment of the strongly pro-inflammatory cluster 5. Notably, trajectory analysis demonstrated that in the absence of Entinostat, the HPV vaccine+NHS-IL12 doublet drove preferential enrichment in TAM cluster 6 (figure 6D), representing an intermediate transcriptional state in the M2-li