An engineered Accum‐E7 protein‐based vaccine with dual anti‐cervical cancer activity

Abstract

Worldwide prevalence of cervical cancer decreased significantly with the use of human papilloma virus (HPV)‐targeted prophylactic vaccines. However, these multivalent antiviral vaccines are inert against established tumors, which leave patients with surgical ablative options possibly resulting in long‐term reproductive complications and morbidity. In an attempt to bypass this unmet medical need, we designed a new E7 protein‐based vaccine formulation using Accum™, a technology platform designed to promote endosome‐to‐cytosol escape as a means to enhance protein accumulation in target cells. Prophylactic vaccination of immunocompetent mice using the Accum‐E7 vaccine (aE7) leads to complete protection from cervical cancer despite multiple challenges conducted with ascending C3.43 cellular doses (0.5‐, 1.0‐, and 2.0 × 10⁶ cells). Moreover, the humoral response induced by aE7 was higher in magnitude compared with naked E7 protein vaccination and displayed potent inhibitory effects on C3.43 proliferation in vitro. When administered therapeutically to animals with pre‐established C3.43 or Tal3 tumors, the vaccine‐induced response synergized with multiple immune checkpoint blockers (anti‐PD‐1, anti‐CTLA4, and anti‐CD47) to effectively control tumor growth. Mechanistically, the observed therapeutic effect requires cross‐presenting dendritic cells as well as CD8 T cells predominantly, with a non‐negligible role played by both CD4⁺ and CD19⁺ lymphocytes. good laboratory practice (GLP) studies revealed that aE7 is immunogenic and well tolerated by immunocompetent mice with no observed adverse effects despite the use of a fourfold exceeding dose. In a nutshell, aE7 represents an ideal vaccine candidate for further clinical development as it uses a single engineered protein capable of exhibiting both prophylactic and therapeutic activity.

By linking Accum to the E7 oncoprotein, an enhanced accumulation and processing of E7‐derived immunogenic peptides take place on MHCI molecules leading to potent CD8 T‐cell activation.

Abbreviations

Accum

Accumulator

aE7

Accum‐E7

CD

cluster of differentiation

CTL

cytotoxic T‐lymphocyte

CTLA4

cytotoxic T lymphocytes‐associated protein 4

DC

dendritic cell

GLP

good laboratory practice

GMP

good manufacturing practice

HPV

human papilloma virus

IP

intraperitoneal

KO

knock‐out

NK

natural killer

OVA

ovalbumin

PD‐1

programmed death 1

PD‐L1

programmed death‐ligand 1

SC

subcutaneous

VLP

virus‐like particle

WT

wild‐type

1. INTRODUCTION

Human papillomavirus (HPV) can cause genital warts subsequently leading to cervical, anal, or head and neck cancers. ¹ , ² , ³ Prophylactic HPV vaccines were therefore developed to prevent these infections and their associated complications. The most widely used of all four commercially available HPV vaccines is Gardasil®‐9. ⁴ Made of L1 capsid proteins resembling those found on the surface of HPV types 6, 11, 16, 18, 31, 33, 45, 52, and 58, this noninfectious vaccine is considered safe and effective, with clinical trials reporting an efficacy of more than 90% at preventing HPV‐related diseases. ⁵ However, this vaccine is not perfect for multiple reasons. Besides the requirement for three doses, the vaccine manufacturing remains costly, which may be prohibitive to low‐income countries. ⁶ In addition, Gardasil®‐9 cannot provide complete protection against all 14 high‐risk HPV subtypes, as it only targets 9 of them. ⁶ Furthermore, the vaccine should be administered before HPV exposure implying that it cannot treat existing infections or established cervical cancer. ⁶ It is therefore recommended for children and young adults before becoming sexually active. Despite HPV vaccination being considered as an important weapon in the fight against cervical cancer, it remains far from being a “silver bullet.”

The HPV genome is divided into three main compartments: an early (E) region containing the E1, E2, E4, E5, E6, and E7 genes; a late (L) region encoding for the L1 and L2 genes; and an upstream regulatory region. ⁷ While the late L1 and L2 genes encode for the major and minor capsid proteins required for the progeny virions assembly, respectively, all E1 to E7 encoding genes play central roles in supporting viral DNA transcription and replication. ⁷ Interestingly, deletion of the L1 and L2 genes occurs following HPV viral DNA genome integration, which explains why prophylactic vaccines (targeting the L1 capsid proteins) are useless against HPV‐associated cancers. ⁸ The situation gets even more complex with the subsequent elimination of the E2 gene, a negative regulator of E6 and E7 gene expression, which results in exacerbated coexpression of these two oncogenes in HPV‐related lesions. ⁹ Interestingly, however, the sustained overexpression of these oncogenes in HPV‐associated malignancies highlight their importance as ideal targets for the potential development of therapeutic cervical cancer vaccines. ¹⁰ , ¹¹

The development of an effective therapeutic vaccine must involve minimal hurdles not only to simplify the production processes while avoiding losses but also to allow for a decrease in manufacturing cost. In that regard, protein‐based vaccination represents an interesting alternative to virus‐like particles (VLPs), as proteins are simpler to purify, which would ease production, lower cost, and promote scalability for mass production. ⁶ One example of HPV protein‐based vaccination tested the use of the minor capsid protein L2. ¹² , ¹³ Despite eliciting a humoral response, the immunogenicity of the L2 protein was considered weak compared with L1. ⁶ , ¹² , ¹³ Another reason blocking the use of the L2 as a potential therapeutic vaccine is its deletion following viral genome integration. ⁶ Although unfit as a vaccine candidate, these studies using the L2 protein provide a strong impetus for the possible use of proteins as an alternative approach to common VLP‐based vaccines in the design of future therapeutic vaccines.

Using the Accum™ technology as a mean to deliver proteins directly to the cytosol of antigen‐presenting cells by rupturing endosomal membranes, we report in this study an engineered Accum‐E7 (aE7)‐based protein vaccine capable of providing a durable memory response when used prophylactically. However, the selection of E7 as a target “non‐self” antigen widened the vaccine scope of application, as it can also control the growth of pre‐established cervical tumors through the generation of potent cytotoxic T lymphocytes (CTLs). Overall, this engineered aE7 vaccine is safe, simple to engineer and manufacture making it an elixir candidate capable of providing both prophylactic and therapeutic activity against cervical cancer.

2. MATERIALS AND METHODS

2.1. Animals and ethics

All female wild‐type (WT) C57BL/6 and Batf3 knock‐out (KO) mice (aged 6‐ to 8‐week‐old) used in this study were purchased from Jackson Laboratories. Animals were housed in a pathogen‐free environment at the animal facility of the Institute for Research in Immunology and Cancer (IRIC). Animal protocols were approved by the Animal Care Committee of Université de Montréal.

2.2. Cell line, reagents, and antibodies

The C3.43 cervical cancer cell line was licensed to Defence Therapeutics by the University of South California. The Tal3 cervical cancer cell line was licensed to Université de Montréal by the Office of Technology Transfer of the Johns Hopkins University . All cell culture reagents were purchased from Wisent Bioproducts. Recombinant E7 protein was purchased from Creative‐BioMart. All flow cytometry antibodies (anti‐PD‐L1, anti‐CTLA4, and anti‐CD47) were purchased from BD Biosciences and used according to the manufacturer's instructions. The GMP‐grade Montanide™ ISA720 adjuvant was purchased from Seppic Inc. The FTY770 chemical used to block T‐cell migration was purchased from Bio‐Techne. The 96‐well Nunc plates were purchased from ThermoFisher. All in vivo used antibodies targeting immune checkpoints (PD‐1, CTLA4, and CD47), 4‐1BB, CD4, CD8, CD19, and anti‐NK1.1 were purchased from BioXCell.

2.3. Engineering the aE7 vaccine

The Accum molecule was synthesized as previously described ¹⁴ and conjugated to recombinant E7. For the conjugation, recombinant E7 was solubilized at 1–10 mg/mL in sterile PBS. The SM(PEG)4 cross‐linker was added to the reaction for 1 h using a 50× molar excess ratio. The free SM(PEG)4 cross‐linker was discarded by centricon filtration and Sephadex column. Accum was added in the same molar excess ratio and incubated for 1 h. Free unlinked Accum was removed by centricon filtration and Sephadex column. Accum‐modified E7 was diluted in sterile PBS at a final concentration of 5–10 mg/mL as confirmed by ultraviolet absorbance.

2.4. Prophylactic vaccination studies

For prophylactic vaccination, female C57BL/6 mice (n = 5/group) were subcutaneously (SC) injected at days 0 and 14 with 50 μL recombinant E7 or aE7 (1 μg/dose) admixed at a 1:3 volume ratio with the Montanide™ ISA720 adjuvant. Control animals received 50 μL of saline solution admixed with the adjuvant (1:3 ratio). Two weeks following the second vaccination, the mice were SC challenged with 5.0 × 10⁵ C3.43 cells, and tumor growth was assessed over time. Two additional challenges (at days 30 and 60 post second vaccination) using 1.0‐ and 2.0 × 10⁶ C3.43 tumor cells, respectively, were given to the aE7‐vaccinated animals as depicted in the cognate panel. For the dosing study, a similar approach was used except that the vaccine was given at 0.10, 0.25, 0.50, 1.0, and 5.0 μg/dose.

2.5. Therapeutic vaccination studies

For therapeutic vaccination, female C57BL/6 mice (n = 5/group) received a SC injection of 5.0 × 10⁵ C3.43 or 1.0 × 10⁶ Tal3 cells at day 0. Three to five days later (or after appearance of palpable tumors with an average size of 35–50 mm³), the mice were SC injected with a 50 μL recombinant E7 or aE7 (1 μg/dose) mixed with the adjuvant (two injections 1 week apart) at a distant site for the first injection and on the opposite flank for the second dosing. Control animals received 50 μL of saline solution admixed with the adjuvant (1:3 ratio). For the dosing study, a similar approach was used except that the vaccine was given at 1.0, 2.5, 5.0, and 10.0 μg/dose. Treated animals were followed thereafter for tumor growth. For therapeutic vaccination in combination with immune checkpoint inhibitors (anti‐PD‐1, anti‐CTLA4, and anti‐CD47) or the agonist anti‐4‐1BB antibody, all mice received intraperitoneal (IP) injections of the antibodies (200 μg/per dose) every 2 days for a total of six doses over 2 weeks. For in vivo depletion studies, anti‐CD4, anti‐CD8, anti‐CD19, and anti‐NK1.1 were IP delivered once a week for three consecutive weeks (total of three IP injections at 200 μg/per dose). A similar approach was used for Batf3KO mice. As for the FTY770‐related study, the compound was delivered IP (5 mg/kg/dose) for a total of five IP injections as depicted on the cognate panel. ¹⁵

2.6. Antibody titer quantification by ELISA

Nunc MaxiSorp™ 96‐well plates were coated overnight with 1.0 μg of recombinant E7 protein diluted in coating buffer at 4°C. The following day, coated plates were washed and then blocked with 3% skim milk for 1 h at room temperature. Following that step, the plates were washed prior to adding the diluted sera (twofold dilutions were prepared). Following a 2‐h incubation period, the plates were washed prior to adding the secondary HRP‐linked anti‐mouse IgG antibody at a dilution of 1:1000. Two hours later, the plates were washed and then incubated at room temperature with HRP for 10–20 min. Following HRP quenching, the signal was detected using a SynergyH1 microplate reader (Biotek).

2.7. In vitro neutralization of C3.43 proliferation

To assess the neutralizing abilities of the generated antibodies, C3.43 cells plated at a density of 5.0 × 10⁵ cells were left for 24 h before adding control, E7‐, or aE7‐specific sera collected from vaccinated mice. All collected sera were used at a 1:100 dilution. Total cell count was performed 48 h later.

2.8. Murine GLP study to assess the immunogenicity and safety profiles of aE7

A GLP study was conducted on male and female C57BL/6 mice by an independent Clinical Research Organization (ITR Laboratories, ITR study number 701158). For control conditions, C57BL/6 mice aged 6–8 weeks (10 animals/group/sex) were either injected with PBS (control animals), or PBS admixed with the adjuvant (1:3 ratio). The test groups were separated as follows: low dose (2.5 μg/dose), medium dose (10.0 μg/dose), and high dose (40.0 μg/dose). A volume of 50 μL was used for all SC immunizations performed at days 0 (left flank) and 21 (right flank). Animals were observed at days −1, 7, 14, 20, 28, 35, and 42 for (i) any unusual physical sign, (ii) dosing site complications (Draize scoring), and (iii) body weight. Necropsy was also conducted on more than 30 different organs to assess the presence of adverse events. Sera collected from these mice were shipped to another independent Clinical Research Organization (CRO) (Immuni T) to quantify their antibody titers by ELISA (Immuni T study number CU0125‐S0730).

2.9. Statistical analysis

p‐Values were calculated using Student's t‐test or one‐way analysis of variance (ANOVA) where appropriate. Results are represented as average mean with SD error bars, and statistical significance is represented with asterisks: *p < 0.05, **p < 0.01, ***p < 0.001.

3. RESULTS

3.1. Prophylactic vaccination using the aE7 vaccine confers complete protection despite multiple challenges

We have recently shown how the prophylactic use of Accum‐linked OVA as a protein‐based vaccine can effectively control the growth of the EG.7 OVA‐expressing T‐cell lymphoma line. ¹⁴ However, the OVA system remains purely “artificial,” as the used xenoantigen is experimental in nature. We thus decided to explore the same approach using a more relevant tumor‐specific antigen. Since the development of cervical cancer relies on transformative events driven by the HPV16 E6 and E7 oncogenes (Figure 1A), we elected E7 in our vaccine design for two main reasons. Besides the fact that its function overlaps with the E6 oncoprotein, the CR1 and CR2 homology domains of E7 are highly conserved among HPV subtypes, and the E7 oncoprotein inhibits multiple cellular checkpoints (p53, p21, and RB‐E2F) compared with E6 (blocks p53 only). ¹⁶ , ¹⁷ Following its Accum conjugation, aE7 was admixed with the commercial GMP‐grade Montanide™ ISA720 adjuvant prior to vaccinating immune‐competent C57BL/6 mice (Figure 1B). In contrast to the naked E7 vaccination group, all aE7‐vaccinated mice remained tumor‐free following the initial C3.43 challenge (Figure 1C). To further test the potency and durability of the induced memory response, the same group was challenged with ascending cellular doses of C3.43 at days 30 and 60. Despite these subsequent challenges, no tumors appeared up to 3 months following the second vaccine dose (Figure 1C,D). We next conducted a dose‐finding vaccination study and observed complete protection using a 0.5–5.0‐μg dosing range, while tumors started appearing at lower doses (0.10 and 0.25 μg; Figure 1E,F). Altogether, these data demonstrate that the E7‐derived HPV16 oncoprotein conjugated to Accum™ can be effectively used to mount a durable protective immune response capable of halting cervical cancer establishment.

FIGURE 1.

The prophylactic potency of Accum‐E7 (aE7). (A) A general overview of the human papilloma virus (HPV) genome encoding for all open reading frames. (B) Schematic overview of the schedule used for prophylactic vaccination. (C) C3.43 tumor growth curves in animals vaccinated using recombinant E7 (blue) versus aE7 (purple). Control mice received saline admixed with the adjuvant (black). The dashed lines represent the two subsequent challenges that were given at days 30 and 60, respectively. (D) The Kaplan–Meier survival curve of the experiment shown in panel (C). (E) Prophylactic vaccination conducted using various aE7 doses. The used doses were 5 μg (red), 1 μg (blue), 0.5 (green), 0.25 μg (gray), and 0.1 μg (purple). The control group is represented by the black line. (F) The Kaplan–Meier survival curve of the experiment shown in panel (E). For experiments displayed in panels C–F, n = 5/group with **p < 0.01 and ***p < 0.001.

3.2. aE7 elicits antibodies capable of interfering with C3.43 proliferation in vitro

Since the aE7 vaccine is protein based, we next investigated its capacity to stimulate a humoral response, which may contribute to inhibiting cancer cell growth. To validate this hypothesis, antibody titers were quantified by ELISA using sera collected from vaccinated mice 3 weeks following the second vaccine dosing in the prophylactic setting (Figure 1B). The antibody titer in aE7‐vaccinated animals was almost 55 times higher compared with the naked E7 group (Figure 2A). To assess the potential relevance of these antibodies in this cervical cancer model, we next treated in vitro growing C3.43 cells with diluted sera (1:100) for 48 h prior to conducting a total cell count (Figure 2B). As expected, sera derived from aE7‐vacinated animals impaired cell proliferation by almost 30% compared with the remaining groups (Figure 2C,D). These data bolster the idea that aE7 could trigger a beneficial humoral response capable of impairing C3.43 proliferation.

FIGURE 2.

Accum‐E7 (aE7)‐induced antibodies capable of blocking C3.43 proliferation in vitro. (A) Antibody titer analysis using sera collected from E7‐ or aE7‐vaccinated animals. (B) A schematic diagram showing the experimental design used to assess antibody‐mediated cell inhibition. (C) C3.43 counts 48 h following addition of the diluted sera to cells in vitro. (D) Percentage of C3.43 cell proliferation inhibition derived from the counts obtained in panel (C). For this panel, n = 5/group with **p < 0.01, ***p < 0.001.

3.3. Therapeutic vaccination using aE7 synergizes with immune checkpoint inhibitors

Gardasil®‐9 is a highly effective prophylactic vaccine capable of inducing a protective response by targeting the L1 capsid proteins on the surface of HPV viral particles. However, this vaccine is inert once cervical cancer is established. As aE7 targets the E7 oncoprotein known to be continuously expressed by cervical cancer cells, we next asked whether the vaccine could be administered therapeutically. Since therapeutic vaccination is most likely to be used as a combo therapy with common immune checkpoint inhibitors, we first assessed the expression of PD‐L1, CTLA4, and CD47 on the surface of C3.43 cells and detected a positive signal for CD47 expression (Figure 3A). When combined with antibodies targeting these three immune checkpoints or the agonist 4‐1BB receptor to enhance the T‐cell response (Figure 3B), aE7 delayed tumor growth in all experimental groups with a pronounced effect observed when combined with anti‐CD47 (Figure 3C,D). These encouraging results prompted us to test whether the obtained immune response could be further enhanced through aE7 dose optimization. Indeed, a dose‐dependent antitumoral response was observed when anti‐CD47 was coadministered with ascending aE7 concentrations (Figure 3E,F), an observation further validated using both anti‐CTLA4 and anti‐PD‐1 inhibitors (Figure 3G,H). To further support our findings, we next tested our therapeutic vaccine against Tal3, a cell line capable of SC tumors in immunocompetent mice isolated from a novel mouse model of spontaneous high‐risk HPVE6/E7‐expressing carcinoma in the cervicovaginal tract by using Sleeping Beauty transposase. ¹⁸ Based on their immune checkpoint expression profile (Figure 4A), a therapeutic vaccination study was conducted using both anti‐CTLA4 and anti‐CD47 (Figure 4B). Akin to the previously obtained results, aE7 vaccination resulted again in efficient Tal3 growth control (Figure 4C) and 20% survival rate by day 40 (Figure 4D). Combining the same vaccine with both immune checkpoint inhibitors improved the effect leading to a complete survival rate by the end of the experiment (Figure 4C,D). Altogether, our studies demonstrate a synergy between multiple immune checkpoint inhibitors and aE7 therapeutic vaccination, with the anti‐CD47 combo being the most effective.

FIGURE 3.

Accum‐E7 (aE7) can be used to treat pre‐established C3.43 cervical cancer. (A) Representative flow cytometry assessing the expression of PD‐L1, CTLA4, and CD47 on the surface of C3.43 cells. (B) A schematic overview of the schedule used for therapeutic vaccination. (C) C3.43 tumor growth curves in vaccinated animals. The displayed groups are: saline/adjuvant (black), aE7 1 μg (gray), aE7 + anti‐PD‐1 (green), aE7 + anti‐CTLA4 (blue), aE7 + anti‐41BB (red), aE7 + anti‐CD47 (purple), anti‐CTLA4 (open blue), anti‐PD‐1 (open green), anti‐41BB (open red), and anti‐CD47 (open purple). (D) The Kaplan–Meier survival curve of the experiment shown in panel (C). (E) Therapeutic vaccination conducted using various aE7 doses. The used doses were: 1 μg (green), 2.5 μg (blue), 5 μg (purple), and 10 μg (red). The saline/adjuvant group is shown in black. (F) The Kaplan–Meier survival curve of the experiment shown in panel (E). (G) Therapeutic vaccination in combination with anti‐PD‐1 and anti‐CTLA4 using the 10‐μg vaccine dose. (H) The Kaplan–Meier survival curve of the experiment shown in panel (G). For panels (C)–(H), n = 5/group with *p < 0.05, **p < 0.01, and ***p < 0.001.

FIGURE 4.

Vaccination using Accum‐E7 (aE7) is also effective against pre‐established Tal3 cervical cancer. (A) Representative flow cytometry assessing the expression of CTLA4, CD47, PD‐1, and PD‐L1 on the surface of Tal3 cells. (B) A schematic overview of the schedule used for therapeutic vaccination. (C) Tal3 tumor growth curves in vaccinated animals. The displayed groups are: saline/adjuvant (black), anti‐CD47 + anti‐CTLA4 (blue), 10 μg aE7 (green), and 10 μg aE7 + anti‐CD47 + anti‐CTLA4 (red). (D) The Kaplan–Meier survival curve of the experiment shown in panel (C). For panels C and D, n = 5/group with *p < 0.05, **p < 0.01, and ***p < 0.001.

3.4. The therapeutic effect of aE7 relies on cross‐presenting dendritic cells (DCs) and CD8 T cells

To trigger an antitumoral response, a given DC must capture, process, and present antigens to responding CD8 T cells. This antigen cross‐presentation process is central to initiate potent cancer‐specific CTLs. We thus questioned whether endogenous cross‐presenting DCs are required for the aE7‐mediated therapeutic effect by comparing the antitumoral response in WT versus Batf3KO mice (deficient in cross‐presenting DCs; Figure 5A). For all subsequent studies, we chose to combine the vaccine with the anti‐PD‐1 immune checkpoint inhibitor, as it is widely used in clinical settings. Indeed, the absence of cross‐presenting DCs results in a 20% survival rate compared with 70% in WT immunocompetent mice (Figure 5B). Based on these observations, we next assessed the role played by other immune cells in mice undergoing various depletion targeting CD4, CD8, CD19, or NK cells (Figure 5C). Although the absence of NK cells had no impact on the elicited immune response, CD19⁺ B‐cell or CD4⁺ T‐cell depletion weakened the animals' ability to control tumor growth (Figure 5D). On the other hand, a full loss in therapeutic potency was observed in mice depleted from CD8 T cells (Figure 5D). If CD8⁺ T‐cell priming is central to controlling cervical cancer growth, then blocking T‐cell migration should lead to similar outcomes. We thus repeated the therapeutic vaccination except that animals were cotreated with Fingolimod (FYT770), a compound sequestering T cells in secondary lymphoid organs (Figure 5E). Although, FTY770 coadministration could still delay animal survival, all mice ended up succumbing to tumor development (0% survival at day 36 in contrast to 80% in aE7 + anti PD‐1 animals; Figure 5F). In summary, these observations allude to a central role for cross‐presenting DCs and CD8 T‐cell activity in mounting potent antitumoral immunity following aE7 vaccination.

FIGURE 5.

The therapeutic potency of Accum‐E7 (aE7) relies mainly on cross‐presenting dendritic cells (DCs) and CD8 T‐cell activity. (A) Schematic overview of the used vaccination schedule in WT versus Batf3KO mice. (B) The Kaplan–Meier survival curve of the experiment shown in panel (A). (C) Schematic overview of the vaccination schedule used for the in vivo depletion studies. (D) The Kaplan–Meier survival curve of the experiment shown in panel (C). (E) Schematic overview of the vaccination schedule used for the FTY770‐related study. (F) The Kaplan–Meier survival curve of the experiment shown in panel (E). For these studies, n = 5/group with *p < 0.05, **p < 0.01, and ***p < 0.001.

3.5. Vaccination using aE7 is safe and immunogenic (GLP study)

To evaluate the safety and immunogenicity profile of the vaccine, we next conducted a GLP study in mice testing three different doses (conducted by a third party). Although our therapeutic vaccination studies used an aE7 range of 1.0–10.0 μg, the GLP study tested a different range depicted as: 2.5 μg (low dose), 10.0 μg (medium dose), and 40.0 μg (high dose; Figure 6A). Body weight analysis conducted on vaccinated male (Figure 6B upper panel) and female (Figure 6B lower panel) immunocompetent mice revealed no negative fluctuations over 42 days regardless of the used dose. As for gross pathology, we chose to simplify data interpretation by focusing on the group receiving the highest aE7 dose (40.0 μg) as a representative group for the safety profile of the vaccine. As shown in Figure 6C, no major abnormalities were observed in any of the processed organs (red squares represent minor/negligible observations). Furthermore, analysis of antibody titers derived from male (Figure 7A) or female (Figure 7A) mice revealed two important points. First, no significant differences could be perceived between both sexes (Figure 7). Second, a plateau was reached using the 10.0 μg dose, as detected titers were not significantly different when compared with the highest dose (Figure 7). In sum, this GLP study confirms that the vaccine is immunogenic, safe, and well tolerated by animals even when used at a high dose (40.0 vs. 10.0 μg).

FIGURE 6.

The Accum‐E7 (aE7) is well tolerated by mice (good laboratory practice [GLP] study). (A) Schematic overview of the GLP study design. (B) Body weight analysis in male (upper panel) versus female (lower panel) mice. (C) Histopathology scores conducted on various tissues isolated from mice receiving the highest vaccine dose. Red squares represent animals with minor but detectable adverse effects. For this panel, n = 10/group/sex.

FIGURE 7.

Analysis of antibody titer using sera derived from the good laboratory practice (GLP) study. (A) Antibody titer analysis using sera collected from male mice at days 0, 21, and 42. (B) Similar to (A) except using sera collected from female mice. The horizontal black line displayed in both panels represents the lower level of quantification. For these panels, 10 mice were used per group and per sex.

4. DISCUSSION

Four HPV‐targeting prophylactic vaccines were developed as a preventive measure to cervical cancer. ⁴ However, no therapeutic vaccines are currently available to cure persistent HPV infections or established cancer—hence the need for a vaccine with a dual prophylactic and therapeutic function. In that regard, therapeutic vaccination could be defined as the ability to mount potent and effective cellular responses against HPV‐specific peptides (e.g., of “non‐self” origin) presented by MHCI molecules, while generating, in parallel, a memory imprint of the ongoing immune response. Thus, an ideal therapeutic vaccine should be efficiently captured and processed by DCs to present immunogenic peptide sequences to responding CD8 T cells. Despite multiple lackluster preclinical and clinical attempts (including ours) to vaccinate using HPV16‐derived E6 or E7 proteins (Figure 1C,D), the elicited immune responses were either weak, ineffective, or failed in demonstrating the hoped‐for clinical outcome. ¹⁹ , ²⁰ This is where the Accum™ technology stands, as it bypasses one of the major molecular obstacles encountered with the use of almost any immune‐based therapeutics or biologics: endosomal entrapment. In a nutshell, any endocytosed molecule is entrapped within the endosomal lumen where it is subjected, during endosomal maturation and/or lysosomal fusion, to a harsh acidic environment, nonspecific degradation by activated proteases, and damages inflicted by reactive oxygen species produced by V‐ATPases and/or NADPH oxidases (Figure 8, upper panel). As we previously demonstrated, Accum™ conjugation to the OVA xenoantigen triggers endosomal membrane breakdown allowing the captured antigen to directly enter the cytosol for efficient processing by the proteasomal machinery. ¹⁴ We thus used a similar approach on the E7 oncoprotein as a mean to enhance its bioaccumulation in the cytosol of DCs where nondamaged immunogenic peptides could be effectively presented to mount hefty CTL responses (Figure 8, lower panel). This vaccination modality would not only avoid the risk of host genome integration akin to genetic‐based vaccines, but the SC administered protein vaccine could spread to several lymphoid organs eliciting therefore a T‐cell response with a broader homing capacity to cancer‐infiltrated tissues. ²¹ , ²² , ²³

FIGURE 8.

A schematic diagram comparing the immune response to the naked E7 (upper panel) versus the Accum‐E7 (aE7) vaccine (lower panel).

Among all the salient observations made in this study, the protective effect of aE7 is astonishing as it appears to be long lasting. For instance, aE7‐vaccinated animals underwent multiple challenges following the second vaccine dose yet remained consistently protected from tumor growth. Nevertheless, the protective effect of the vaccine is sustained even if used at a half dose (0.5 vs. 1.0 μg). This implies that the vaccine is highly immunogenic most likely due to the Accum's ability to efficiently accumulate E7 in target DCs. Besides inducing a CTL response, aE7 triggers an antibody titer that is far superior to the one generated by naked E7. Although antibodies are generally believed to play pivotal roles in the control of cancer cell growth, our data could argue that aE7‐induced antibodies may inhibit or promote cancer cell death possibly via antibody‐dependent cellular cytotoxicity‐related mechanisms following recognition of cell surface peptide–MHCI complexes. ²⁴ , ²⁵ Additional studies are therefore warranted to assess the specificity and function of these antibodies prior to tailoring a more detailed mode of action.

The development of a novel prophylactic vaccine for cervical cancer is not attractive to the field, as it may not provide substantial advantages over commercially available vaccines (Cervaris™, Cecolin®, Gardasil®‐4, and Gardasil®‐9). However, the continuous expression of HPV‐related oncogenes in cervical cancer cells represents an opportunity to direct CTLs against oncogene‐derived peptides presented by MHCI complexes. Due to the negative role played by immune checkpoints in inhibiting CTL function in the tumor microenvironment, we first screened the C3.43 and Tal3 cell lines for the expression of the most common immune checkpoints and identified CD47 as the predominant one. This observation correlates with our therapeutic vaccination studies, as aE7 combined to anti‐CD47 or anti‐CD47/anti‐CTLA4 triggers an antitumoral response superior to anti‐CTLA4 or anti‐PD‐1 inhibitors. Although these studies suggest that aE7 should be used in a combination therapy regimen for the treatment of established cervical cancer, they do not highlight the mechanism by which the vaccine works. Using a combination of KO and in vivo depletion strategies, we found that aE7 relies on cross‐presenting DCs to prime CD8 T cells. This does not preclude the importance of other immune cells such as CD4⁺ T cells, which are known to support the development of CTLs via cytokine secretion, nor CD19⁺ B cells, which differentiate into plasma cells secreting antibodies capable of impeding tumor growth as shown in our studies. Interestingly, however, NK cell depletion had no impact on the vaccine potency clearly indicating an exclusive role for adaptive immunity in response to aE7 therapeutic vaccination.

Using the right tools in vaccine design is central to develop meaningful antitumoral responses. Several groups focused on reverse vaccinology or immune informatics to identify epitopes against specific antigens. ²⁶ Although these approaches can be less time‐consuming, cost‐effective, more accurate in identifying relevant or immunogenic sequences, and are generally safe, the final identified product(s) will always face endosomal entrapment combined to nonspecific antigen degradation and/or damages inflicted to immunogenic sequences. Our proof‐of‐concept study demonstrates that it is indeed possible to recycle previously tested antigens (e.g., E7) using the Accum™ technology or apply it to other vaccine modalities such as long polypeptide fragments, DNA‐ or even RNA‐based vaccines in order to enhance their therapeutic potency. For instance, it would be interesting to explore the use of an Accum‐E5 protein vaccine, as this HPV gene plays a role in transforming normal cells, downregulating MHCI expression, and blocking protein trafficking through the endoplasmic reticulum. ²⁷ , ²⁸ In other words, the versatility of the Accum™ technology could easily widen the scope for future vaccine design targeting HPV or other emerging pathogens in addition to cancer.

Taken together, our study shows that aE7 (i) provides potent protection (prophylactic vaccination) and durable memory responses, (ii) triggers antibodies exhibiting a non‐negligible role in fighting cervical tumors, (iii) controls established tumors when delivered in combination with several immune checkpoint inhibitors (therapeutic vaccination), and (iv) is relatively safe, well tolerated, and immunogenic by animals even when used at higher doses (GLP study). Besides additional physicochemical characterization (e.g., protein thermostability and hydrophilicity) to better understand whether Accum conjugation affects the structural conformation, stability, or half‐life of the E7 protein, further validations will be required to shown how this vaccine could synergize with other cancer immunotherapy approaches targeting HPV‐induced malignancies.

AUTHOR CONTRIBUTIONS

Jean‐Pierre Bikorimana: Data curation; formal analysis; investigation; methodology; writing – review and editing. Jamilah Abusarah: Conceptualization; data curation; formal analysis; writing – review and editing. Marina Gonçalves: Conceptualization; data curation; formal analysis; writing – review and editing. Roudy Farah: Conceptualization; data curation; formal analysis; writing – review and editing. Wael Saad: Conceptualization; data curation; formal analysis; writing – review and editing. Sebastien Talbot: Conceptualization; data curation; formal analysis; writing – review and editing. Daniela Stanga: Conceptualization; data curation; formal analysis; writing – review and editing. Simon Beaudoin: Conceptualization; data curation; methodology; writing – review and editing. Sebastien Plouffe: Formal analysis; writing – review and editing. Moutih Rafei: Conceptualization; formal analysis; funding acquisition; investigation; methodology; project administration; resources; supervision; validation; writing – original draft; writing – review and editing.

FUNDING INFORMATION

This study was funded by a research contract granted by Defense Therapeutics to Dr. M. Rafei (RB002835).

CONFLICT OF INTEREST STATEMENT

Daniela Stanga, Simon Beaudoin, and Sebastien Plouffe are employees of Defense Therapeutics Inc. and declare competing financial interest. Other authors do not have any conflict of interest.

ETHICS STATEMENT

Approval of the research protocol by an Institutional Reviewer Board: The GLP study was performed in compliance with regulations and test guidelines from the EMA, FDA, and ICH.

Informed Consent: N/A.

Registry and the Registration No. of the study/trial: The ITR GLP study number is 701158.

Animal Studies: Animal protocols were approved by the Animal Care Committee of Université de Montréal (protocol # 22‐065).

Bikorimana J‐P, Abusarah J, Gonçalves M, et al. An engineered Accum‐E7 protein‐based vaccine with dual anti‐cervical cancer activity. Cancer Sci. 2024;115:1102‐1113. doi: 10.1111/cas.16096