Treatment of patients bearing human papillomavirus (HPV)-related cancers with synthetic long-peptide (SLP) therapeutic vaccines has shown promising results in clinical trials against premalignant lesions, whereas responses against later stage carcinomas have remained elusive. We show that conjugation of a well-documented HPV-E7 SLP to ultra-small polymeric nanoparticles (NP) enhances the antitumor efficacy of therapeutic vaccination in different mouse models of HPV+ cancers. Immunization of TC-1 tumor-bearing mice with a single dose of NP-conjugated E7LP (NP-E7LP) generated a larger pool of E7-specific CD8+ T cells with increased effector functions than unconjugated free E7LP. At the tumor site, NP-E7LP prompted a robust infiltration of CD8+ T cells that was not accompanied by concomitant accumulation of regulatory T cells (Tregs), resulting in a higher CD8+ T-cell to Treg ratio. Consequently, the amplified immune response elicited by the NP-E7LP formulation led to increased regression of large, well-established tumors, resulting in a significant percentage of complete responses that were not achievable by immunizing with the non-NP–conjugated long-peptide. The partial responses were characterized by distinct phases of regression, stable disease, and relapse to progressive growth, establishing a platform to investigate adaptive resistance mechanisms. The efficacy of NP-E7LP could be further improved by therapeutic activation of the costimulatory receptor 4-1BB. This NP-E7LP formulation illustrates a “solid-phase” antigen delivery strategy that is more effective than a conventional free-peptide (“liquid”) vaccine, further highlighting the potential of using such formulations for therapeutic vaccination against solid tumors. Cancer Immunol Res; 6(11); 1301–13. ©2018 AACR.
Human papillomaviruses (HPV) are the most common sexually transmitted agents worldwide, and it is estimated that most of the world population will come in contact with these viruses at least once in their lifetime (1). Chronic infection with oncogenic HPVs can lead to the development of asymptomatic neoplasias in the cervix and vulva and in the head and neck region (1–3). In some patients, these lesions eventually progress to symptomatic squamous cell carcinomas (1).
In contrast to many types of cancer, for which finding a suitable antigen for therapeutic vaccines might require complex analyses of the tumor and prediction of neoantigen binding to MHC (4, 5), HPV+ malignancies can be targeted by using the viral oncoproteins (6). Clinical trials using HPV oncoprotein–derived synthetic long-peptide (SLP) vaccines to treat patients with early-stage vulvar and cervical intraepithelial neoplastic lesions have shown promising results, in rare cases leading to complete responses (7–9). However, SLP vaccines have poor efficacy in most such patients, as well as in those with malignant disease (7, 10). Clinical responses to SLP vaccines are dependent on dendritic cells (DC) that process SLPs into short peptides for presentation on MHC molecules (11–13) to elicit CD8+ T-cell responses (7, 9). Although SLPs can elicit a strong CD4+ type 1 T helper (Th1) response in patients treated with HPV vaccines, the induced CD8+ T cell counterpart is usually weaker (7–9, 14). In addition to the magnitude of the CD8+ T-cell response, the CD8+ T-cell/regulatory T-cell (Treg) ratio is a prognostic marker for HPV+ cancers (15). Therapeutic vaccination of HPV+ cancer patients typically leads to an increased abundance of Tregs (16), and larger premalignant lesions show a greater increase in intratumoral Tregs compared with smaller ones (7). The stimulation of Treg infiltration consequent to therapeutic vaccination is suspected to contribute to the poor efficacy of SLP vaccines (17). As such, identifying vaccine formulations, which both enhance recruitment of activated CTLs while avoiding the concomitant induction of Tregs, may be necessary for efficacy against cervical and other HPV+ malignancies.
Bioengineering technologies are proving applicable to immunotherapy, for example, the development of lymph node–targeting vaccines (18–21), including an ultra-small polymeric nanoparticle (NP) formulation can boost immune responses to model tumor antigens (18). The chemical structure of this 30-nm diameter NP, with disulfide bonds on the surface, enables loading with thiol-bearing antigenic payload molecules (22). Their small size enables these NPs to rapidly reach draining lymph nodes after subcutaneous and intradermal administration, where they can be efficiently taken up by DCs (22). Once internalized by DCs, the disulfide bonds that bind the antigen payload to the surface of the NPs are broken, and the released antigen is preferentially processed and loaded onto MHC-I molecules to activate the CD8+ T-cell response (18, 22, 23).
We hypothesized that NP conjugation of SLPs derived from the HPV protein E7 (E7LP) would boost immune responses against E7-expressing neoplasias, in particular, well-established solid tumors in different anatomical locations. To test this hypothesis, the E743-77 SLP (E7LP; ref. 24) was conjugated to NPs, and compared with the non–NP-conjugated free (liquid) E7LP in several mouse models of HPV+ cancers. We comparatively evaluated efficacy of the two vaccine formulations in mice bearing tumors derived from inoculating HPV16 E6/7–transformed TC-1 cells, which have been widely used to test therapeutic strategies against HPV+ cancers (24–28), and a new squamous cell carcinoma cell line (termed SC1) derived from a skin tumor in an HPV16 transgenic mouse (29, 30). We assessed the two vaccines in mice bearing subcutaneous TC-1 and SC1 tumors, orthotopic TC-1 tumors at the cervico-vaginal tract (25), and experimental TC-1 lung metastases (31). The results presented below demonstrate that conjugating an E7 SLP, which is currently being evaluated in clinical trials (7–10, 32), to a solid-phase platform has clear benefits compared with a conventional free/liquid vaccine.
Materials and Methods
Mice were immunized with a total amount of 15 μg of E7LP as free peptide or in the NP-bound formulation, and 20 μg (subcutaneous TC-1 tumors experiment) or 40 μg (intravaginal tumors, lung metastases and SC1 experiments) of CpG was used as adjuvant. For the free E7LP vaccine, the long peptide was first dissolved in DMSO and then diluted in PBS prior to immunization. Control mice were treated with PBS. Groups were assigned by randomizing the mice according to the tumor volume or to the bioluminescence signal to ensure a similar size distribution. All the mice from one experiment were immunized together on the same day. All tumor-bearing mice were immunized once, unless otherwise stated. Subcutaneous immunizations of mice were performed in the four limbs using the Hock method (33). Subcutaneous TC-1 tumor–bearing mice were immunized when mean tumor size was around 100 mm3 (usually at day 10–12 after implantation) unless otherwise stated, or when tumor size was around 170 mm3 for the anti-4-1BB combination experiment. Cervico-vaginal tumor-bearing mice were immunized subcutaneously in the back near the tail. Subcutaneous SC1 tumor–bearing FVB/N H2b mice were immunized once, 47 days after tumor implantation when the mean tumor size reached approximately 100 mm3.
Mice, tumor cells, and antibody treatments
For all TC-1 implantation experiments, C57BL/6NCrl mice were used. C57BL/6NCrl mice (aged 6–8 weeks) were purchased from Charles River and kept under pathogen-free conditions at the animal facility of Ecole Polytechnique Fédérale de Lausanne (or at the animal facility of the Centre Hospitalier Universitaire Vaudois for the cervico-vaginal experiment). All experiments were performed in accordance with Swiss law and with approval of the Cantonal Veterinary Office of Canton de Vaud, Switzerland.
For the subcutaneous model, TC-1 cells (kindly provided by Prof TC Wu, Johns Hopkins University, Baltimore, MD) were cultured in DMEM (Gibco), 10% FBS (Gibco), penicillin/streptomycin (100 U/mL penicillin, 100 μg/mL streptomycin; Gibco), and were implanted subcutaneously on the flank of C57BL/6NCrl mice with inoculations of 5 × 105 cells in 100 μL HBBS. The TC1 cell line was authenticated by the expression of E6 and E7, cultured for 2 weeks (4–5 passages) prior implantation and tested for Mycoplasma. Tumor growth was measured with a caliper using the formula V = W2Lπ/6 and mice were sacrificed when tumor size reached 1,000 mm3.
For the cervico-vaginal tumor model, tumors were induced as described in ref. 25. Briefly, anesthetized diestrus-synchronized female mice were pretreated with 4% nonoxynol-9 (Abcam) for 6 hours, and were then intravaginally implanted with 5 × 104 TC-1 cells engineered to express a luciferase reporter gene (TC-1-Luc). Tumor growth was monitored by bioluminescence index (BLI) using the Xenogen imaging system after intraperitoneal injection of 50 mg/kg d-luciferin (PerkinElmer) in PBS. Mice were sacrificed when interruption criteria defined in the animal license were reached.
For the lung metastasis model, 5 × 105 TC-1-Luc cells were resuspended in 100 μL of PBS and injected into the tail vein of female mice. Lung metastases were monitored by BLI after intraperitoneal injection of 50 mg/kg d-luciferin (PerkinElmer) in PBS. Images were acquired using a Photon Imager (IVIS Spectrum) system and data analyzed with the provided software (IVIS). Mice were imaged twice a week, and were sacrificed when interruption criteria defined in the animal license were reached.
SC1 cells were generated from a squamous cell carcinoma of the skin of a K14HPV16/H-2b mouse (29). The K14HPV16/H-2b line was generated by crossing K14HPV16/FVB/N (H-2q) mice (34–36) with C57BL/6 (H-2b) mice to introduce the H-2b locus. F1 mice were backcrossed for 11 generations to FVB/N, selecting for the H-2b locus in every generation. This genetic configuration allows K14HPV16/H-2b mice to present E7-derived peptides on MHC-I molecules while maintaining the FVB/N background that is permissive for squamous carcinogenesis. SC1 cells were cultured in DMEM (Gibco), 10% FBS (Gibco), penicillin/streptomycin (100 U/mL penicillin, 100 μg/mL streptomycin; Gibco), and 1 × 106 SC1 cells resuspended in 100 μL HBBS/Matrigel (Corning; 1:1 solution) were implanted subcutaneously on the flank of FVB/N H-2b mice. Tumor growth was measured with a caliper using the formula V = W2Lπ/6.
For CD4+ and CD8+ T-cell depletion, TC-1 tumor–bearing mice received intraperitoneal injections every 4 days of 10 mg/kg CD4 (clone GK1.5, BioXcell) and/or 10 mg/kg CD8 antibody (clone 53-6.7, BioXcell). Antibody treatment started 3 days before vaccination and was continued for 3 weeks. CD4+ and CD8+ T-cell depletion was monitored weekly by flow cytometry analyses on the blood. Mice with less than 95% depletion were excluded from the analyses.
For the anti-4-1BB treatment, mice received 3 doses of 350 μg of anti-4-1BB (clone LOB12.3, BioXcell) intraperitoneally every 3 days starting at day 12 together with the vaccine.
CpG-B 1826 oligonucleotide (5′-TCCATGAGCTTCCTGACGTT-3′ as phosphorothioated DNA bases) was purchased from Microsynth and used as an adjuvant in both vaccine formulations. HPV16 E7 long peptide GQAEPDRAHYNIVTFCCKCDSTLRLCVQSTHVDIR (aa 43–77, purity > 90%) was purchased from Think Peptides and the Protein and Peptide Chemistry Facility, UNIL and used for NP conjugation and immunization. HPV16 E7 CD8+ T-cell peptide RAHYNIVTF was purchased from Think Peptides and used for restimulation.
NP synthesis and conjugation
NPs were synthesized, functionalized, and characterized as described previously (22, 23, 37). For antigen conjugation, HPV16 E7 long peptide was dissolved in DMSO and incubated for 12 hours in endotoxin-free water in the presence of NPs and guanidine hydrochloride (AppliChem) at room temperature. NP-E7LP was purified by size-exclusion chromatography using CL-6B matrix (Sigma-Aldrich), eluted, and stored in PBS at room temperature. The size of NPs before and after conjugation was determined by dynamic light scattering and remained around 30 nm. E7LP loading on the nanoparticles was measured by BCA assay (Thermo Fisher Scientific). NPs alone have been shown to have no adjuvant activity (38, 39).
Cell preparation for flow cytometry and antigen-specific in vitro restimulation
Spleens were harvested and gently disrupted through a 40-μm filter (Thermo Fisher Scientific). Red blood cells were lysed using ACK lysis buffer, and cells were filtered again through a 40-μm filter before use. Tumors were harvested and minced using a scalpel and digested for 45 minutes using collagenase A (0.33 U/mL, Roche), dispase (0.85 U/mL, Roche), DNAse I (144 U/mL, Roche) in RPMI medium with intermittent shaking at 37°C. For CD8+ T-cell antigen-specific restimulation, cell suspensions from either spleen or tumor were cultured at 37°C for 6 hours in a 96 well plate in the presence of 1 μg/mL of the HPV16 CD8+ T-cell peptide RAHYNIVTF. After the first 3 hours of culture, Brefeldin A (Sigma-Aldrich) was added to a final concentration of 5 μg/mL. All the cells were cultured in Iscove's modified Dulbecco's medium (Gibco) supplemented with 10% FBS (Gibco) and penicillin/streptomycin (100 U/mL penicillin, 100 μg/mL streptomycin; Gibco).
Flow cytometry analyses were performed 9 days after immunization of tumor-bearing mice. For surface staining and blocking, cells were incubated for 15 minutes on ice with the antibodies diluted in PBS 2% FBS. Before tetramer and antibody staining, all cell suspensions from tumor or spleen were blocked with anti CD16/32 (BioLegend). Cells were then labeled with fixable live/dead cell viability reagent (Invitrogen) in PBS for 15 minutes on ice. E7-specific CD8 T cells were stained with a tetramer-recognizing HPV16 E7 peptide RAHYNIVTF (49-57) presented by H-2Db (University of Lausanne, UNIL, Lausanne, Switzerland) for 30 minutes at room temperature before surface antibody staining. If no intracellular staining was performed after surface staining, cells were fixed with 2% PFA in PBS for 15 minutes on ice. For intracellular staining, cells were permeabilized and fixed with the Foxp3/Transcription Factor Staining Buffer Set Kit (eBioscience) following the manufacturer's instructions and then incubated overnight with the antibodies diluted in 1× Permeabilization buffer provided with the kit. After staining, cells were washed and resuspended in PBS 2% FBS for analyses. Samples were acquired on a Gallios analyzer (Beckman Coulter) and data were analyzed using FlowJo software (Tree Star Inc.). For flow cytometry analyses, all lymphocytes were gated on single, live cells. Antibodies used for flow cytometry: CD3 (clone 145-2C11, Thermo Fisher Scientific), CD4 (clone RM4-5, BioLegend), CD8a (clone 5H10, Thermo Fisher Scientific), B220 (clone RA3-6B2, Thermo Fisher Scientific), IFNγ (clone XMG1.2, BioLegend), TNFα (clone MP6-XT22, Thermo Fisher Scientific), GZB (clone NGZB, Thermo Fisher Scientific), Foxp3 (clone FJK-16s, Thermo Fisher Scientific), CD25 (clone PC61.5, Thermo Fisher Scientific), CD45 (30-F11, Thermo Fisher Scientific), CD11b (clone M1/70, Thermo Fisher Scientific), CD11c (clone N418, BioLegend), Ly6G (clone 1A8, BioLegend), Ly6C (clone HK1.4, Thermo Fisher Scientific), CD206 (clone C068C2, BioLegend), MHCII (clone M5/114.15.2, BioLegend), 4-1BB (clone 17B5, BioLegend), GITR (clone DTA-1, BD Horizon), ICOS (clone 7E.17G9, BD Pharmingen), and OX-40 (clone OX-86, BioLegend).
Tumors were harvested, embedded in optimal cutting temperature medium (OCT; Sakura), and immediately frozen on dry ice. Ten-micron–thick sections were cut from OCT-embedded tumors using a cryostat and collected on Superfrost Plus glass slides (Thermo Fisher Scientific). Tissue sections and OCT-embedded tumors were stored at −80°C. For immunofluorescence staining, sections were fixed in ice-cold methanol (Thermo Fisher Scientific) for 10 minutes before proceeding. Slides were washed with PBS to remove the remaining OCT and then blocked for 45 minutes at room temperature with PBS + 5% BSA + 2.5% FBS and then stained with primary antibodies diluted in PBS + 1% BSA overnight at 4°C in a humidified chamber. On the following day, slides were washed with PBS and stained with secondary antibodies diluted in PBS + 1% BSA for 1 hour at room temperature. Before mounting, slides were washed again in PBS and then covered with mounting media (Dako) containing DAPI (Roche, 5 μg/mL). Coverslips (Menzel Glaser) were applied to the slides and sealed using nail polish. Images were acquired using a Leica DM5500B and processed using Fiji (ImageJ). The following antibodies were used for immunofluorescence staining: CD8 (clone 53-6.7, eBioscience), keratin 14 (clone poly19053, BioLegend), ICOS (clone 7E.17G9, BD Biosciences), F4/80-PE (clone BM8, eBioscience), CD11c-FITC (clone N418, BioLegend), MRC1-Alexa Fluor 647 (clone C068C2, BioLegend), and PD-L1-PE (MIH5, eBioscience).
RNA extraction and real-time PCR
RNA from tumor tissue was isolated with the miRNeasy kit (Qiagen) and 1 μg of total RNA was subjected to cDNA synthesis using the iScript cDNA synthesis kit (Bio-Rad). Real-time PCR was conducted using the Rotor Gene SYBR green PCR kit (Qiagen) with 25 ng of cDNA on a Rotor Gene Q instrument (Qiagen) and the following primers (Microsynth):
The relative amount of cDNA was calculated with the ΔΔCt method using RPL13a as the reference gene and the TC-1 tumor sample as calibrator.
cDNA from TC-1 tumors was amplified with the E7-specific primers ATGCATGGAGATACACCTAC and ATTATGGTTTCTGAGAACAGA by Platinum Taq polymerase (Invitrogen) and cloned into pSC-A employing the StrataClone PCR cloning kit (Agilent Technologies). The insert containing the HPV16 E7 gene sequence was evaluated by Sanger sequencing using the T3 primer: TTAACCCTCACTAAAGG (Microsynth).
Statistical analyses were performed in GraphPad Prism 7. Flow cytometry data and CD8+ T cells quantification by histology were compared using Student t test. Survival curves were compared using Log-rank (Mantel–Cox) test. Normalized growth curves of SC1 tumors and mean TC-1 tumor growth curves were compared using two-way ANOVA. Statistic significance is indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; n.s., not significant.
Immunization with NP-E7LP enhances systemic immune responses and prolongs survival
Classical (free-peptide/liquid) therapeutic vaccine formulations against HPV16-driven cervical cancers have had equivocal therapeutic responses against established carcinomas, motivating efforts to improve immunization strategies. Thus, we sought to determine whether therapeutic vaccination with the NP-conjugated E7LP (NP-E7LP) could be more effective than the free E7LP formulation at eliciting systemic E7-specific CD8+ T-cell responses in mice with solid tumors, and to comparatively assess effects of the two vaccines on overall survival. C57Bl/6 mice were subcutaneously implanted with TC-1 cells. When the mean size of the palpable solid tumors was approximately 100 mm3, we administered either NP-E7LP+CpG, free E7LP+CpG, or PBS as control (Fig. 1A).
We began by comparatively evaluating the systemic adaptive immune response. Tumor-bearing mice were sacrificed 9 days after immunization and E7-specific CD8+ T cells in the spleen were quantitated by flow cytometry analyses (Fig. 1B and C). Mice that received the NP-E7LP showed an increase in the frequency of E7-specific CD8+ T cells that was 3- to 16-fold higher compared with mice that were immunized with the free E7LP. E7-specific CD8+ T cells were undetectable in PBS-treated mice, suggesting the absence of a spontaneous CTL immune response (Fig. 1B and C).
To comparatively evaluate the functionality of the E7-specific CD8+ T cells elicited by the two vaccine formulations, splenocytes harvested 9 days after immunization were restimulated in vitro with the class I MHC–restricted E7 peptide RAHYNIVTF (26). NP-E7LP–treated mice showed a significant increase in cells expressing IFNγ (Fig. 1D; Supplementary Fig. S1), TNFα (Fig. 1E; Supplementary Fig. S1), and Granzyme B (GZB; Fig. 1F; Supplementary Fig. S1) compared with mice that were immunized with free E7LP. A 3-fold increase in TNFα and IFNγ double-positive cells was also detected in the NP-E7LP immunized mice, indicating that polyfunctional CD8+ T cells were being generated in higher proportions (Fig. 1G; Supplementary Fig. S1). Consistent with the observed absence of E7-specific CD8+ T cells in PBS-treated animals (Fig. 1B and C), the cytokines and GZB were not detected in this group (Fig. 1D–F; Supplementary Fig. S1). These results demonstrate that the NP-E7LP vaccine formulation is superior to the free E7LP vaccine at inducing a systemic CD8+ T-cell response, which is characterized by a larger pool of E7-specific cells and increased production of the proinflammatory cytokines and GZB involved in killing target cells.
Congruent with the enhanced immune response elicited by the NP vaccine, mice that received the NP-E7LP formulation survived significantly longer than those who received the free form of E7LP, while all the PBS-treated control mice reached the defined endpoint (tumor volume > 1 cm3) starting at day 20 postimplantation (Fig. 1H). Although overall survival was improved in both vaccinated groups, 5 of 9 tumors in mice treated with free E7LP did not respond to vaccination and kept growing, and these mice had to be euthanized around day 20, similarly to PBS-treated mice (Fig. 1H and I). In contrast, 9 of 9 NP-E7LP–treated mice showed tumor regression after immunization; moreover, in the NP-E7LP–treated group, tumor shrinkage was faster and tumor size remained stable (below 250 mm3) for a longer period than in the 4 of 9 transitory responders to the free E7LP (Fig. 1I). Analysis of the mean tumor volumes highlights the superior ability of NP-E7LP at causing tumor regression (Supplementary Fig. S2). Among the NP-E7LP–treated mice, 3 of 9 mice showed a complete response with the disappearance of a palpable mass, and one was still tumor-free at day 137 postimplantation (Fig. 1H and I). In contrast, none of the mice in the free E7LP–treated group showed a complete response (Fig. 1I). Mice that received a second inoculation of NP-E7LP 7 days after the first one did not show improvement in survival nor in tumor growth rate, compared with the group that received only a single dose of the NP-E7LP vaccine (Fig 1H and I; Supplementary Fig. S2).
To determine whether CD8+ and/or CD4+ T cells contribute to the effects of NP vaccination, the NP-E7LP immunization of TC-1 tumor–bearing mice was combined with CD4 or CD8 antibody treatment (or a combination of the two). CD4+ T-cell depletion did not affect the efficacy of the vaccine, revealing that CD4+ T-cell help is not required for the enhanced effect of the NP-E7LP formulation (Supplementary Fig. S3A–S3C). In contrast, CD8+ T-cell depletion completely abolished the effect of the vaccine, demonstrating that the antitumor immune response elicited with NP-E7LP is dependent on CD8+ T cells (Supplementary Fig. S3A–S3C). Combined depletion of CD8+ and CD4+ T cells was similar to CD8+ T-cell depletion alone, supporting the conclusion that CD4+ T cells do not play a role in the antitumor immune response in this context (Supplementary Fig. S3A–S3C). Although the E7LP we used contains both a CD4+ and a CD8+ T-cell epitope, in light of the dispensability of the CD4+ T-cell response for the antitumor effects, we decided to focus subsequent experiments on CD8+ T cells.
Our initial experiments were designed to treat well-established solid tumors. Next we vaccinated TC-1 tumor–bearing mice at an earlier timepoint, incipient neoplasia, similarly to what has been reported in a previous study (40). We vaccinated at day 7 after implantation, when the mean tumor volume was approximately 22 mm3, and 32% of the mice lacked a palpable mass. In this setting, both formulations led to prolonged responses, although the NP-E7LP vaccine appears superior (Supplementary Fig. S4A and S4B). As such, immunization of mice with incipient neoplasia may not be fully informative about the effects of therapies aimed at well-established solid tumors, where the relative benefits of the NP-E7LP formulation are evident.
Tumors from NP-E7LP–treated mice have increased tumor-specific CD8+ T-cell infiltrates
We next assessed the abundance of E7-specific CD8+ T cells inside the TC-1–derived tumors, and characterized associated effects on the tumor microenvironment (TME), comparing the NP-E7LP with the free E7LP vaccine. Using the previously described protocol (Fig 1A), TC-1 tumor–bearing mice were sacrificed 9 days after vaccination. Flow cytometry analyses of the tumors revealed a CD8+ T-cell response that reflected the systemic responses described above in the spleen. Thus, tumors harvested from mice treated with the NP-E7LP vaccine had significantly higher CD8+ T-cell infiltrates compared with tumors from free E7LP–treated mice, while control mice treated with PBS did not have appreciable CD8+ T-cell infiltration (Fig. 2A). The fraction of E7-specific CD8+ T cells determined by tetramer staining followed the same trend, with NP-E7LP–treated mice having significantly more tumor-infiltrating antigen-specific cells than free E7LP–treated mice (Fig. 2B). Consistent with these results, immunostaining of tumor tissue sections from NP-E7LP–treated mice with anti-CD8 showed substantial infiltrates of CD8+ T cells distributed throughout the lesion. In contrast, tumors from the free E7LP–treated group contained only scattered CD8+ T-cell infiltrates, while CD8+ T cells were virtually absent in the PBS-treated control groups (Fig. 2C; Supplementary Fig. S5). To assess the functionality of the tumor-infiltrating E7-specific cells, we performed a restimulation assay using disaggregated cells from whole tumors. Intracellular staining for IFNγ, TNFα, and GZB revealed that cytokine production by CD8+ T cells was significantly higher in mice that received NP-E7LP compared with free E7LP, whereas it was virtually absent in tumors derived from PBS-treated mice (Fig. 2D–F), underscoring the increased immune activity of the NP-E7LP vaccine over the free E7LP formulation.
We next analyzed innate immune cell infiltrates in treated TC-1 tumors 9 days postvaccination. Irrespective of the formulation, vaccination caused a significant increase in intratumoral CD11b+F4/80+ macrophages compared with control PBS treatment (Supplementary Fig. S6A). Despite the general increase, the subpopulation of M2-like CD206+ (MRC1+) macrophages was significantly lower in both immunized groups (Supplementary Fig. S6B). Vaccination was also associated with an increase in CD11bHiLy6C+Ly6G− monocytes (Supplementary Fig S6C); there was, however, no significant change in CD11bHiLy6C−Ly6G+ neutrophils, which were rare in infiltrating TC-1 tumors (Supplementary Fig. S6D). Although more abundant in treated mice, no significant differences in F4/80−CD11b+CD11c+ and F4/80−CD11b−CD11c+ DCs were detected between cohorts that received the two vaccines, although the F4/80−CD11b−CD11c+ DCs were significantly increased in the NP-E7LP–treated group compared with PBS-treated controls (Supplementary Fig. S6E and S6F). Thus, the NP-E7LP formulation behaved similarly to the free E7LP form in regard to eliciting innate immune infiltration, suggesting that the enhanced efficacy of the NP-E7LP vaccine is predominantly due to a larger influx of tumor-specific and cytokine-producing CD8+ T cells rather than to specific changes in innate immune cell populations.
NP-E7LP limits intratumoral Treg accumulation and improves CD8+ T-cell/Treg ratio
It has been shown that NPs can direct antigens toward the intracellular pathway that leads to peptide cross-presentation and loading on MHC-I molecules (22, 23). Thus, we reasoned that the NP E7LP-conjugated vaccine might be increasing peptide loading onto MHC-I molecules and thereby impacting the intratumoral CD8+ T-cell/Treg ratio. We analyzed Treg cells infiltrating TC-1 tumors by flow cytometry 9 days after immunization of tumor-bearing mice. In contrast to a previous study performed with TC-1 tumors, in which Tregs were reportedly abundant in untreated tumors (41), we detected relatively few Tregs in this context. Vaccination with free E7LP led to a 4.5-fold higher proportion of intratumoral Tregs compared with the NP-E7LP–treated group (Fig. 3A). Tregs were present in NP-E7LP–vaccinated mice at a similarly low frequency to the PBS-treated control mice (Fig. 3A). Consequently, the CD8+ T-cell to Treg ratio was considerably improved in NP-E7LP–treated tumors compared with the free-E7LP group (Fig. 3B). Total CD4+ T-cell proportions were unchanged upon vaccination with both formulations and similar to those in PBS-treated tumors (Fig. 3C). Thus, vaccination using NP-bound antigen limited the generation and consequent intratumoral accumulation of Tregs following therapeutic vaccination.
NP conjugation improves long-peptide vaccination in different tumor models
We next compared the NP-E7LP formulation with the conventional free long peptide vaccine against TC-1 tumors arising in the cervico-vaginal tract and in lung metastases. To compare the vaccines in an advanced disease setting, orthotopic cervico-vaginal TC-1 tumor–bearing mice were immunized at day 14 after implantation, when tumors were well established. NP-E7LP vaccination significantly improved survival of tumor-bearing mice with 25% surviving more than 150 days; in the free E7LP group, all the mice reached the endpoint shortly after 50 days (Fig. 4A and B). When mice bearing well-established TC-1 lung metastases (at day 9 post intravenous inoculation) were treated, (Fig. 4C and D), efficacy was also improved, with 28.5% of the mice surviving for more than 100 days postimplantation, whereas free E7LP vaccine treated mice all reached endstage before day 50. Tumors and lung metastases collected from NP-E7LP–immunized mice 9 days after vaccination showed an increase in CD8+ T-cell infiltrates compared with the free E7LP group, while PBS-treated control had minimal CD8+ T-cell infiltrates in either setting (Fig. 4E and F; Supplementary Fig. S7A and S7B).
To test the efficacy of NP-E7LP immunization in another setting, we employed squamous carcinoma cells (SC1) that express the HPV16 early region genes, representing the bona fide cell type transformed by HPVs. SC1 cells were implanted subcutaneously, and mice were immunized when mean tumor volume was 100 mm3. Vaccination with NP-bound E7LP led to tumor regression/stabilization, whereas tumors from mice that received either free E7LP or PBS kept growing (Fig. 4G). Treatment with NP-E7LP led to a significant increase in intratumoral CD8+ T cells compared with the other groups (Fig. 4H and I). These assays underscore the superior efficacy of NP-E7LP over the free E7LP vaccine in physiologically distinct disease settings.
Efficacy-associated changes in the TME are progressively lost upon relapse
Despite increased systemic and local immune responses with the NP-E7LP, most of the subcutaneous TC-1 tumors eventually relapsed to progressive tumor growth after an initial phase of tumor shrinkage (response) that was followed by an intermediate phase where the tumor size remained relatively stable for a period ranging from 10 to 50 days (Fig. 1I). We sought to characterize the changes in the TME that are associated with these three phases (Supplementary Fig. S8A) in mice that received the NP-E7LP vaccine.
We analyzed expression of the E7 oncogene, potential mutation of the E7-encoded CD8+ T-cell peptide epitope contained within the SLP vaccine, and expression of several components of the antigen processing and presentation machinery, comparing responding tumors with relapsed ones. We found no significant change that could explain the loss of responsiveness toward the cancer cells (Supplementary Fig. S8B).
We then examined the T cells in the TME and found that CD8+ T-cell infiltration progressively decreased after the response phase (Fig. 5; Supplementary Fig. S9A). CD8+ T-cell activation, as indicated by ICOS staining, was high in the response phase but was already lost in the stable disease phase (Fig. 5; Supplementary Fig. S9B).
Macrophages were present in the TME at all stages, although the expression of F4/80 appeared to be upregulated in responding and stable tumors (Fig. 5, Supplementary Figs. S9C and S10). The DC- and M1-macrophage marker CD11c (42) was absent in PBS-treated tumor sections but present throughout the sections of response-phase tumors, and then progressively lost during transition to stable disease and relapse phase tumors (Fig. 5; Supplementary Figs. S10 and S11A). Staining for the M2 macrophage marker MRC1 (CD206; ref. 42) was conversely increased in untreated and relapsed tumors (Fig. 5; Supplementary Figs. S10 and S11B). These data indicate that the abundance of DCs and the frequency and phenotype of macrophages changes concordant with CD8+ T-cell activation. It has previously been reported that macrophages are modulated by the activity of CD8+ T cells and required for the antitumor response against TC-1 tumors elicited by vaccination (24). As such, it seems likely that the DC and macrophage dynamics in the stable disease and relapse phases are consequent to reduced recruitment of activated CD8+ T cells.
Analysis of PD-L1 expression revealed that it was upregulated in the response phase when activated CTLs are abundant, likely induced by CD8+ T-cell–secreted IFNγ, whereas it was virtually absent in untreated tumors and in the stable and relapse phases (Fig. 5; Supplementary Fig. S11C), suggesting that PD-L1 expression is not the primary mechanism responsible for the progressive loss of CD8+ T cells.
NP-E7LP synergizes with agonistic anti-4-1BB/CD137 to further boost tumor rejection
Early immunization with a traditional free SLP vaccine of TC-1 tumor–bearing mice led to discernibly enhanced immune responses and survival when given in combination with an agonistic anti-4-1BB antibody (40). In light of the evident loss of CD8+ T-cell activity in the stable disease and relapse phases (Fig. 5; Supplementary Fig. S9B), we comparatively evaluated the effects of anti-4-1BB in the context of both vaccines. To extend the previous study and more rigorously evaluate the potential benefit of this agonistic antibody, we treated larger, later-stage tumors that are poorly responsive to immunization with free E7LP alone. TC-1 tumor–bearing mice were immunized when mean tumor volume was around 170 mm3. Mice treated with free E7LP reached endstage within 3 weeks after implantation of TC-1 cells, whereas, as in previous experiments, all the mice treated with the NP-E7LP formulation survived significantly longer (Fig. 6A and B). Anti-4-1BB significantly improved survival in combination with both vaccine formulations (Fig. 6A and B). The improved efficacy of free E7LP + anti-4-1BB was comparable with that of NP-E7LP alone, and the inclusion of anti-4-1BB further improved its therapeutic efficacy. Comparison of mean tumor volumes in the treatment cohorts (Supplementary Fig. S12) was time-limited and thus not fully informative due to the statistical necessity of prematurely terminating the comparison when the first free E7LP–treated mice reached the veterinary-defined endpoints.
To investigate the mechanism by which anti-4-1BB increased the efficacy of the two vaccines, we analyzed intratumoral immune infiltrates 9 days postimmunization with and without anti-4-1BB treatment. The anti-4-1BB treated groups showed only a modest increase in total E7-specific CD8+ T-cell infiltrates (Supplementary Fig. S13A). However, we observed an increase in CD44+KLRG1+ terminal effector E7-specific CD8+ T cells (Supplementary Fig. S13B) in the context of anti-4-1BB treatment that correlated with a significantly higher infiltration of these cells into the tumors, particularly in the cohort treated with NP-E7LP vaccine plus anti-4-1BB (Supplementary Fig. S13C). The CD8+ T-cell/Treg cell ratio showed a trend toward an increase upon anti-4-1BB treatment in the NP-E7LP group (Supplementary Fig. S13D). We also observed a trend toward increased frequency of E7-specific CD8+ T cells expressing activation markers (4-1BB, GITR, ICOS, and OX-40; Supplementary Fig. S13E–S13H) and elevated percentages of IFNγ-, TNFα-, and GZB-producing E7-specific CD8+ T cells upon anti-4-1BB treatment (Supplementary Fig. S13I–S13K). These data underscore the enhanced therapeutic efficacy of the NP vaccine in combination with anti-4-1BB.
Therapeutic vaccination is an attractive strategy to direct the immune system against specific cancer-associated antigens in solid tumors. CD8+ T cells play a fundamental role in tumor rejection, and thus induction of a potent CD8+ T-cell response is a major goal of therapeutic vaccines. Cervical cancer and other HPV-induced malignancies may be attractive prototypes for immunotherapy in general and therapeutic vaccines in particular, given that the viral oncogenes encode neoantigenic oncoproteins that drive the disease. In patients with HPV+ tumors, generating therapeutically efficacious CTL responses has proven difficult to achieve with classical vaccination strategies involving peptide/protein vaccines (43), including studies utilizing the HPV E7-derived SLP that produced, at best, modest CD8+ T-cell responses (7–9, 14). In terms of efficacy, E7 SLPs have proven to be effective in a minority of patients bearing early premalignant HPV+ lesions, whereas bona fide malignancies have been largely refractory (7, 10), underscoring the need for better therapeutic modalities.
Herein we show that by conjugating an E7-derived SLP to synthetic ultra-small (30 nm) NPs, it is possible to significantly boost the CD8+ T cell response following immunization (by 3–16 fold), consistent with the previously reported (18, 22, 23) ability of NPs to enhance MHC-I antigen presentation by efficient delivery of their antigenic payload to cross-presenting DCs. The NP-E7LP vaccine formulation was capable, in comparison with the non-NP–conjugated free E7LP, of eliciting a systemic immune response characterized by a larger pool of E7-specific CD8+ T cells producing activation-associated cytokines and GZB.
The NP-E7LP vaccine induced greater infiltration of tumors by CD8+ T cells in both subcutaneous and cervico-vaginal TC-1 tumors as well as in TC-1–derived lung metastases, in comparison with the free E7LP. The results highlight the ability of the NP-E7LP–generated CD8+ T cells to efficiently reach different anatomic locations and perform their effector function. Therapeutic vaccination with NP-E7LP of mice bearing TC-1–derived solid tumors led to tumor shrinkage in 100% of the treated animals, significantly increasing their lifespan. In contrast, in this disease setting, free E7LP only produced a response in about 50% of the mice. The NP-E7LP vaccine conveyed long-term survival in 25% and 28.5% of the mice with orthotopic TC-1 tumors or experimental lung metastases, respectively, whereas all of mice treated with the free E7 long peptide rapidly succumbed to their disease. Similar results were observed in mice bearing subcutaneous HPV16+ squamous cell carcinomas, extending the comparative benefits to the histologic tumor type associated with HPV transformation. Thus, in SC1 tumor–bearing mice, NP-E7LP immunization led to higher intratumoral CD8+ T-cell infiltrates compared with the classical free E7LP, resulting in tumor regression/stabilization in marked contrast to the free E7LP vaccine, which was ineffective this setting.
These data clearly demonstrate that NP conjugation can improve upon the capability of SLPs to induce tumor antigen–specific CD8+ T cells in sufficient quantity and activity to produce substantive regressions of well-established solid tumors, leading to appreciable survival benefit.
Promising results have also been reported for other distinct vaccine delivery platforms based on larger particles (44, 45) and lymph node–targeting vaccines, either alone (19) or in complex combinatorial strategies (46). Compared with an impressive but elaborate combinatorial therapeutic regimen (46), we show that a relatively simple approach, solely based on conjugation of an antigenic SLP to a NP-based delivery system and its administration in a single dose is sufficient to elicit a significant antitumor response not achievable with a traditional unconjugated “liquid” formulation.
Collectively, the results from the current study and other reports clearly demonstrate that a variety of solid-phase and lymph-node targeting systems have considerable potential to improve upon the generation of immune responses following therapeutic vaccination. It will be auspicious in future studies to compare these distinctive antigen delivery systems head-to-head in the same experimental setting, to rigorously assess their relative benefits.
We have, in addition, confirmed the results of a previous study describing the potential of combining therapeutic vaccination with anti-4-1BB (40), strengthening the proof of concept by demonstrating herein that this agonistic antibody can be beneficially combined with a lymph node targeting solid-phase antigen delivery system. Collectively, the data establish that anti-4-1BB treatment promotes the generation of terminal effector CD8+ T cells that infiltrate the tumors in higher numbers, are more activated, and produce higher amounts of cytokines. We suggest that combinations of novel antigen-delivery platforms, such as the NPs described herein, with agonistic antibodies, including anti-4-1BB/CD137, could present a new foundation upon which to build future treatment strategies for HPV+ cancer patients.
Another recognized limitation of therapeutic vaccines based on SLPs is the concomitant generation of Tregs in response to the immunization (16, 40), an undesirable response that has also been observed in HPV+ cancer patients, particularly those with larger/advanced lesions (7). Consistent with this result, we observed that immunization of tumor-bearing mice with free E7LP was associated with an increase in tumor-infiltrating Tregs that resulted in a low CD8+ T-cell/Treg ratio. In contrast, immunization with the NP-E7LP did not increase intratumoral Treg infiltrates, resulting in a much higher CD8+ T-cell/Treg ratio. These data suggest that, by means of NP conjugation it may be possible to restrict the generation of Tregs and thereby help circumvent one of the limitations of SLP vaccines. This effect is likely due to the ability of NPs to direct the delivery of antigens to the intracellular processing pathway responsible for peptide cross-presentation on MHC class I molecules (22, 23) ultimately favoring a higher intratumoral CD8+ T-cell/Treg ratio.
Despite the substantial increase in both systemic and local antitumor immune responses achievable with the NP-E7LP vaccine, we observed that most tumors eventually relapsed. Tumor shrinkage was followed by a relatively long stable-disease phase that lasted up to 50 days in subcutaneously transplanted mice, followed by relapse to progressive tumor growth. Our data indicate that the relapse is not due to a loss of antigen presentation by tumor cells, but it is associated with a reduction of intratumoral CD8+ T cells and possibly a loss of their activity. These changes are evidently not caused by the presence of immunosuppressive M2 macrophages or by expression of the inhibitory molecule PD-L1 (47, 48). We envisage that the treatment regimen with NP-bound SLP described herein could be used as a model in which to study changes associated with immunoediting (49) as well as strategies aimed to reactivate the immune response during the stable disease/equilibrium phase, which could be relevant for patients with cancer treated with immunotherapies. This escape mechanism could explain why, in subcutaneous TC-1 tumor–bearing mice, a second inoculation of the NP-E7LP vaccine, performed 7 days after the first one, failed both to promote greater tumor shrinkage and to improve survival, compared with mice that received only a single administration.
In conclusion, our data illustrate the potential for this particular and for other NP-based lymph node–targeting platforms (19, 44–46) to serve as effective delivery vehicles that boost the efficacy of neoantigen-based therapeutic vaccines in patients, most obviously to target HPV+ cancers with the well-validated E7LPs, but potentially also other cancers for which stimulatory neoantigens have been identified (50). It is possible that the use of antigenic peptides coupled with such systems will enhance many immunotherapeutic vaccine strategies for solid tumors, whether single or combinatorial, involving SLPs. This proposition warrants consideration for FDA investigational new drug, enabling vaccine development and clinical evaluation.
Disclosure of Potential Conflicts of Interest
M.A. Swartz has ownership interest (including stock, patents, etc.) in a patent on the nanoparticle technology, but it is not licensed. No potential conflicts of interest were disclosed by the other authors.
Conception and design: G. Galliverti, M. Tichet, D. Nardelli-Haefliger, M.A. Swartz, D. Hanahan, S. Wullschleger
Development of methodology: G. Galliverti, S. Hauert, M.A. Swartz
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): G. Galliverti, M. Tichet, S. Domingos-Pereira, S. Wullschleger
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): G. Galliverti, M. Tichet, S. Domingos-Pereira, M.A. Swartz, D. Hanahan, S. Wullschleger
Writing, review, and/or revision of the manuscript: G. Galliverti, M. Tichet, S. Domingos-Pereira, S. Hauert, D. Nardelli-Haefliger, M.A. Swartz, D. Hanahan, S. Wullschleger
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): G. Galliverti, S. Wullschleger
Study supervision: D. Nardelli-Haefliger, M.A. Swartz, D. Hanahan, S. Wullschleger
We thank B. Torchia, M.-W. Peng, and M.A. Gaveta for technical support; Dr. R. Guiet and the Bioimaging and Optics Platform (BIOP) Core Facility at EPFL for developing image analyses tools; Prof. T.C. Wu from Johns Hopkins University for providing TC-1 cells; and Prof. Daniel E. Speiser and Prof. Pedro Romero for comments on the manuscript. This work was supported by a Sinergia grant 160742 from the SNSF.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.