Therapeutic HPV Cancer Vaccine Targeted to CD40 Elicits Effective CD8⁺ T-cell Immunity

Human papillomavirus (HPV) infection is the most common sexually transmitted infection in the United States, with most sexually active individuals acquiring HPV. Approximately 79 million people are currently infected with HPV, and 14 million are newly infected each year in the United States (1, 2).

Of the more than 150 different types of HPV (3), high-risk HPV strains (HPV16 and 18) are strongly associated with many cancers of the cervix, vagina, vulva, penis, and anus (4, 5). Up to 22% of adults are HPV16-seropositive, but most primary infections are cleared without sequelae (6–8). The clinical observations are similar for HPV18. However, in a small but significant proportion of individuals, their immune systems fail to eradicate the virus and it becomes latent. Such persistent infection can lead to cancers. A recent U.S. population-based study conducted by the Centers for Disease Control and Prevention shows that 66% of cervical cancers, 55% of vaginal cancers, 79% of anal cancers, and 62% of oropharyngeal cancers are attributable to HPV16 or 18. Each year in the United States, an estimated 26,000 new cancers are attributable to HPV, about 17,000 in women and 9,000 in men (9). Among these cancers, HPV-associated head-and-neck cancers, including oropharyngeal squamous cell carcinomas, have recently risen dramatically in men under 50 years old, whereas the incidence of HPV-negative oropharyngeal cancers has decreased (10). Although many of the HPV-positive tumors can be cured with modern multidisciplinary treatment approaches, development of new and effective therapeutic vaccines against HPV-associated malignancies is of importance to bring better benefit to the patients.

HPV oncogenes E6 and E7 drive malignant transformation by inactivating the p53 and retinoblastoma tumor suppressor genes (11). Such tumor-associated antigens (TAA, E6/7) of viral origin are constitutively expressed in HPV-infected nonmalignant and malignant cells. As a consequence, E6 and E7 are ideal immunotherapeutic target antigens to evoke cellular immune responses, particularly CD8⁺ cytotoxic T lymphocytes (CTL), which are the major effector cells that kill tumors. These unique characteristics of HPV-associated malignancies, which are distinct from many other cancers where TAAs are of host origin, provide an opportunity to develop effective immunotherapeutic vaccines. Several candidate vaccines are currently being investigated, including E6- and/or E7-derived peptides (12–15), proteins (16, 17), plasmids (18, 19), and live-vectored vaccines (20–23). In addition to potential safety concerns for some of these vaccine models, especially for immune-compromised individuals, eliciting robust CD8⁺ CTL-mediated immunity still remains a major challenge (24–26).

Among various types of antigen presenting cells (APC), dendritic cells (DC) are the most efficient at cross-presenting antigens to CD8⁺ T cells (27, 28). This ability of DCs makes them ideal cellular targets for the rational design of cancer vaccines. Indeed, targeting antigen to DCs via their surface receptors using conjugates of anti-DC surface receptor and antigen can result in greatly enhanced cellular immunity (29). We have found that targeting antigens to DCs via CD40 results in greater antigen-specific CD8⁺ CTLs than targeting to any of nine other DC surface receptors tested (30). It is therefore reasonable to test whether an immunotherapeutic vaccine designed to target CD40 has the potential to bring clinical benefit to HPV-associated cancer patients. In this study, we assessed the preclinical efficacy of a CD40-targeting vaccine for HPV16-associated malignancies in patients in vitro and human CD40 transgenic (hCD40Tg) mice in vivo. Data from this study support the development of a CD40-targeting immunotherapeutic vaccine for HPV-associated malignancies.

Antibodies and tetramers used are described in the Supplementary Materials and Methods. IL2, IL7, and IL15 were purchased from PeproTech. Overlapping 15-mer peptides (staggered by 11 amino acids) spanning the entire HPV16.E6 and HPV16.E7 proteins and human prostate-specific antigen (PSA) were purchased from Mimotopes. Poly(I:C) was purchased from Invivogen. Montanide ISA 51 was acquired from SEPPIC. Mouse CD8⁺ T-cell (negative) and CD4⁺ T-cell (positive) enrichment kits were purchased from StemCell.

A mAb to CD40 (clone 12E12), specific for the ectodomains of human CD40, was generated (30, 31). The specificity of αCD40 (clone 12E12) mAb was verified by its specific binding to human CD40 expressed on CHO cells transfected with the full-length human CD40 (Supplementary Fig. S1A and S1B) and by ELISA by measuring its binding to the recombinant human CD40–Fc fusion protein (Supplementary Fig. S1C). Recombinant fusion protein of αCD40 and HPV16.E6/7 has been described: αCD40VH2-HPV16.E6/7 (GenBank KP684039) paired with the corresponding light chain αCD40VK2-hIgGK sequence within KM660792. Human IgG4 heavy chains with two site mutations (S228P and L235E; ref. 32) were made to further abolish non-specific binding to Fc receptors and to stabilize the labile C–C bond typical of IgG4. Recombinant αCD40VH2-HPV16.E6/7 was expressed in stable CHO-S transfectants and purified by protein A column chromatography. Endotoxin levels were less than 0.02 units/mL. The sequence of IgG4 of the control IgG4-HPV16.E6/7 has been described (31).

All human samples were acquired in accordance with the protocol that was approved by the Institutional Review Boards at Baylor Research Institute (BRI) and the Icahn School of Medicine at Mount Sinai School of Medicine. PBMCs were enriched from peripheral blood by density gradient centrifugation with Ficoll Paque PLUS (GE Healthcare). Complete RPMI 1640 medium (Gibco; Supplementary Materials and Methods) supplemented with 10% heat-inactivated human AB serum (Gemini) was used in human in vitro experiments.

Six- to 10-week-old female wild-type (C57BL/6) and hCD40Tg mice (C56BL/6 background; ref. 30) were used. All mouse experiments were conducted with the approval of the Institutional Animal Care and Use Committee at BRI. Animals were immunized subcutaneously (s.c.) on days 0, 14, and 28, with 100 μL phosphate-buffered solution (PBS) containing 30 μg mAb-HPV16.E6/7 fusion proteins (αCD40-HPV16.E6/7 or IgG4-HPV16.E6/7) plus 50 μg poly(I:C). In the experiment testing the effect of Montanide ISA 51 VG, water-in-oil (1:1, v/v) emulsions were prepared. All animals were euthanized 7 days after the third immunization. Spleens and blood were collected for assessing HPV16.E6/7-specific T-cell responses.

Mouse IFNγ ELISpot^Plus precoated plates and reagents were obtained from Mabtech. ELISpot assays were performed according to the manufacturer's instructions. Details can be found in Supplementary Materials and Methods.

A total of 2 × 10⁴ TC-1 tumor cells (kindly provided by Dr. T.C. Wu at Johns Hopkins University in May 2013) were authenticated by detecting HPV16.E6/7 proteins with Western blot analysis. Cell lines used in these experiments were passaged a maximum of 4 times before implanted s.c. in the left flank of the mice (marked as day 0). TC-1 cells were also tested for Mycoplasma contamination. Animals were treated s.c. with 30 μg αCD40-HPV16.E6/7 plus 50 μg poly(I:C) in 100 μL PBS via the indicated routes. In separate experiments, animals bearing tumors (100–200 mm³) were treated s.c. or by intratumoral (i.t.) injection with 5 doses (at 6-day intervals) of 30 μg αCD40-HPV16.E6/7 plus 50 μg poly(I:C). For testing the preventative efficacy of the vaccine, animals were immunized s.c. with 30 μg αCD40-HPV16.E6/7 plus 50 μg poly(I:C) on days 0, 14, and 28. TC-1 tumor cells (2 × 10⁴) were implanted s.c. on indicated days. Tumor volumes were measured twice every week. Removed tumors were processed with Mouse Tumor Dissociation Kit (Miltenyi Biotec) to generate single-cell suspensions.

Primary methods of data analysis included descriptive statistics (means ± SD). Differences between the two groups were detected using the Student t test or ANOVA. Statistical significance was set at P < 0.05.

A recombinant fusion protein of humanized αCD40 antibody (12E12) and HPV16.E6/7 (αCD40–HPV16.E6/7) was generated as previously described (31, 33, 34). The fusion protein αCD40–HPV16.E6/7 carries HPV16.E6_1–120/E7_1–60 appended to the heavy chain C-terminus (Fig. 1A and B, left). The parental humanized CD40 antibody, αCD40-HPV16.E6/7, and IgG4-HPV16.E6/7 were analyzed by SDS-PAGE (Fig. 1B, right). All proteins were expressed by CHO-S cells and purified by protein A affinity chromatography (30, 31, 33, 34).

Human CD40-transfected CHO cells were stained with different amounts (0.01–10 μg/mL) of αCD40–HPV16.E6/7 or control IgG4–HPV16.E6/7. αCD40–HPV16.E6/7 bound to CHO-hCD40 cells in a dose-dependent manner, whereas IgG4–HPV16.E6/7 did not (Supplementary Fig. S2A). Neither αCD40–HPV16.E6/7 nor IgG4–HPV16.E6/7 bound to CHO cells transfected with a mock plasmid (data not shown). αCD40–HPV16.E6/7 also bound to blood CD11c⁺ myeloid DCs, while IgG4–HPV16.E6/7 did not. Similarly, αCD40–HPV16.E6/7 could bind to blood B cells and plasmacytoid DCs, but not to T cells (Supplementary Fig. S2B). Taken together, αCD40–HPV16.E6/7 can bind to CD40 and can thus efficiently target APCs, including DCs, at the concentrations (0.02–2 μg/mL) tested.

To test whether αCD40–HPV16.E6/7 can activate E6/7-specific T cells from HPV16⁺ head-and-neck cancer patients, PBMCs were cultured for 7 days in the presence of αCD40–HPV16.E6/7 (0.2 μg/mL) or an equimolar amount of E6/7 peptide pool. Cells were then restimulated for 48 hours with E6/7 peptide pool or control peptides. Both αCD40–HPV16.E6/7 and an E6/7 peptide pool could expand E6/7-specific IFNγ-producing T cells (Fig. 1C). They produced significant amounts of IFNγ in response to E6/7 peptide pool, but not to control peptides. These data were from experiments performed with PBMCs from 10 patients (Fig. 1D).

We next examined E6/7-specific CD4⁺ and CD8⁺ T-cell responses by intracellular IFNγ staining (Fig. 1E). Compared with HPV16.E6/7 whole protein, αCD40–HPV16.E6/7 expanded greater numbers of E6/7-specific IFNγ⁺CD8⁺ and IFNγ⁺CD4⁺ T cells, demonstrating that αCD40–HPV16.E6/7 could efficiently activate both E6/7-specific CD8⁺ and CD4⁺ T cells. There was no significant response by E6/7-specific CD8⁺ T cells without targeting the protein antigens. Therefore, αCD40–HPV16.E6/7 can efficiently activate HPV16.E6/7-specific CD8⁺ and CD4⁺ T cells from the blood of head-and-neck cancer patients.

We assessed the immunogenicity of αCD40–HPV16.E6/7 with hCD40Tg mice (30). Three s.c. doses of 30 μg αCD40–HPV16.E6/7 alone did not elicit significant HPV16.E6/7-specific CD8⁺ or CD4⁺ T-cell responses, as assessed by measuring the frequency of H-2D^b–E7_49-57 tetramer⁺ CD8⁺ T cells in the blood (Fig. 2A) and IFNγ ELISpot assays using CD8⁺ and CD4⁺ T cells from spleens (Fig. 2B). However, a combination of 30 μg αCD40–HPV16.E6/7 plus 50 μg poly(I:C) (35) resulted in significant HPV16.E6/7-specific CD4⁺ and CD8⁺ T-cell responses. Most CD8⁺ T cells primed with αCD40–HPV16.E6/7 plus poly(I:C) were specific to peptide cluster 5 from the HPV16.E7 (Supplementary Fig. S3A), as reported (36), whereas CD4⁺ T cells were mainly specific to peptide cluster 2 from HPV16.E6 protein (Supplementary Fig. S3B).

We chose poly(I:C) as an adjuvant based on our in vitro data (Supplementary Fig. S3) and published data (35). Human monocyte-derived DCs (MoDC) were loaded with 1 μg/mL αCD40–Flu.M1 (matrix protein 1 of influenza virus, PR8, H1N1; ref. 30) and then activated overnight with poly(I:C) or CL075, a TLR7/8 agonist. Purified autologous CD8⁺ T cells were cocultured for 7 days with the DCs. Compared with CL075, poly(I:C) resulted in enhanced Flu.M1-specific CD8⁺ T-cell responses, as assessed by measuring the frequency of HLA-A2-Flu.M1_58–66 tetramer⁺ CD8⁺ T cells (Supplementary Fig. S4A) and IFNγ⁺CD8⁺ T cells (Supplementary Fig. S4B). Poly(I:C), compared with CL075 or no adjuvant, resulted in similar or increased expression of granzyme B (Supplementary Fig. S4C) and perforin (Supplementary Fig. S4D) in the CD8⁺ CTLs.

Thus, αCD40-HPV16.E6/7 plus poly(I:C) can prime HPV16.E6/7-specific CD4⁺ and CD8⁺ T-cell responses in vivo.

We tested whether montanide could further enhance the αCD40–HPV16.E6/7-induced T-cell responses in vivo. Two groups of hCD40Tg animals (4 mice per group) were immunized s.c. with αCD40–HPV16.E6/7 plus poly(I:C) in either PBS or an emulsion of PBS and montanide (1:1 ratio, v/v). Animals were boosted twice at 2-week intervals. IFNγ ELISpot data (Fig. 2C, left) showed that montanide did not promote HPV16.E6/7-specific CD8⁺ T-cell responses. However, it significantly enhanced HPV16-specific IFNγ-producing CD4⁺ T-cell responses (Fig. 2C, right). αCD40–HPV16.E6/7 alone emulsified in montanide did not induce a significant HPV-specific CD4⁺ or CD8⁺ T-cell response (data not shown). Therefore, montanide, along with poly(I:C), can promote αCD40–HPV16.E6/7-induced HPV16-specific IFNγ-producing CD4⁺ T-cell responses, but not CD8⁺ T-cell responses.

We investigated whether the HPV16.E6/7-specific CD8⁺ T-cell responses elicited with αCD40–HPV16.E6/7 were the result of its specific targeting to human CD40 in vivo. αCD40–HPV16.E6/7 plus poly(I:C) elicited HPV16.E6/7-specific CD8⁺ T-cell responses in hCD40Tg, but not in wild-type animals, as assessed by measuring the frequency of H-2D^b–E7_49–57 tetramer⁺CD8⁺ T cells in the blood (Fig. 3A). IFNγ ELISpot assay with splenocytes (Fig. 3B) also showed that αCD40–HPV16.E6/7 plus poly(I:C) elicited HPV16.E6/7-specific T-cell responses only in the hCD40Tg animals. αCD40–HPV16.E6/7 plus poly(I:C) was also more efficient than control IgG4–HPV16.E6/7 plus poly(I:C) at eliciting HPV16-specific T-cell responses in hCD40Tg animals (Fig. 3C and D). We concluded that αCD40–HPV16.E6/7 targets human CD40 and can thus efficiently evoke HPV16-specific CD8⁺ T-cell responses in hCD40Tg mice.

We assessed the efficacy of αCD40–HPV16.E6/7 plus poly(I:C) in the prevention of TC-1 tumor growth in hCD40Tg animals. Animals were first immunized s.c. with a combination of 30 μg αCD40–HPV16.E6/7 and 50 μg poly(I:C) three times at 2-week intervals. Animals were then divided into two groups (5 mice per group). One group of animals was implanted s.c. with TC-1 tumor cells 1 week after the third immunization (day 35). The other group was implanted with TC-1 tumor cells 2 months after the final immunization (day 90). Unimmunized animals were used as controls. All control animals died within 35 days after TC-1 tumor cell implantation (Fig. 4A). In contrast, all animals in the two experimental groups survived until the end of the experiment. Consistent with these survival data, all control animals showed rapid progression of tumor growth that reached a tumor volume of 1,000 mm³ within 35 days after TC-1 tumor cell implantation, whereas animals in the two experimental groups showed no tumor growth or significantly delayed progression of TC-1 tumor growth (Fig. 4B).

In a separate experiment, we assessed the frequency of HPV16.E6/7-specific T-cell memory in the spleens of animals that received three doses of αCD40–HPV16.E6/7 plus poly(I:C). One group of mice was euthanized 1 week after the final immunization (day 35), the other 2 months after the final immunization (day 90). The frequency of HPV16.E7-specific CD8⁺ T cells in the spleens of the two groups were not significantly different, although the HPV16.E7-specific CD8⁺ T-cell memory tended to decay over time (Fig. 4C). The frequencies of HPV16.E6/7-specific IFNγ⁺CD8⁺ and CD4⁺ T cells in the spleens of the two groups were also not significantly different (Fig. 4D). Thus, the combination of αCD40–HPV16.E6/7 plus poly(I:C) can mount preventative immunity against TC-1 tumor growth and HPV16.E6/7-specific T-cell memory in hCD40Tg animals.

To assess the therapeutic efficacy of αCD40–HPV16.E6/7 plus poly(I:C), hCD40Tg mice were first implanted s.c. with TC-1 tumor cells and then immunized s.c., intraperitoneally (i.p.), or intramuscularly (i.m.) with αCD40–HPV16.E6/7 plus poly(I:C). Ten mice in each group were treated once every 6 days until day 36. Treatment began on day 6 after TC-1 tumor implant when the mice had palpable-sized tumors (Fig. 5A). Regardless of the injection routes, αCD40–HPV16.E6/7 plus poly(I:C) substantially increased overall survival, with 90% of mice of each immunized group still alive at the end of the experiment, compared with survival of only 10% of mice in the unimmunized group (Fig. 5B). Tumor progression in the majority of immunized mice was also suppressed, with some mice showing tumor rejection after the fourth, fifth, or sixth immunization (Fig. 5C). Tumor rejection was confirmed with surgeries. Thus, αCD40–HPV16.E6/7 plus poly(I:C) could mount therapeutic immunity against TC-1 tumors. The TC-1 tumors started regrowing, particularly in the animals that did not reject the tumors, when the treatment was stopped, although the tumor growth rate in these mice was slower than that in unimmunized animals. Poly(I:C) alone did not result in enhanced survival or suppression of tumor growth (data not shown).

To further understand the therapeutic immunity evoked with αCD40-HPV16.E6/7 plus poly(I:C), we assessed the frequencies of tetramer⁺CD8⁺ T cells in the tumors as well as in the peripheral blood of immunized and unimmunized mice. TC-1 cells were implanted on day 0 and then animals received 3 s.c. doses of αCD40–HPV16.E6/7 plus poly(I:C) once every 6 days starting from day 6. Mice were euthanized whenever they had large tumor volumes (>1,500 mm³), significant body weight loss (>20%), hunched backs, or significantly diminished activity. Based on these criteria, all mice in the unimmunized group were euthanized within 35 days after TC-1 tumor implantation. In contrast, all mice in the immunized group were euthanized between days 40 and 75—mostly on day 75 when the experiment was finished. Figure 5D (top) shows that animals immunized with the vaccine had significantly higher frequencies of HPV16.E7-specific CD8⁺ T cells in the tumors than the unimmunized mice. The frequencies of blood HPV16.E7-specific CD8⁺ T cells between the two groups of animals, however, were not significantly different (Fig. 5D, bottom). Importantly, the frequencies of tetramer⁺CD8⁺ T cells in the tumors inversely correlated with the tumor volumes only in the immunized group (Fig. 5E, top)—but not in the unimmunized group (Fig. 5E, bottom). The frequencies of tetramer⁺CD8⁺ T cells in the blood and tumor volumes were not significantly associated in either group of animals (Fig. 5F). Collectively, we concluded that αCD40–HPV16.E6/7 plus poly(I:C) could mount HPV16.E6/7-specific CD8⁺ T-cell–mediated therapeutic immunity to control HPV16.E6/7-expressing TC-1 tumor progression.

The efficacy of αCD40–HPV16.E6/7 plus poly(I:C) was further investigated using animals bearing tumors (100–200 mm³; Fig. 6). In this experiment, hCD40Tg animals were treated with 30 μg αCD40–HPV16.E6/7 plus 50 μg poly(I:C) per dose once tumor volumes reached at least 100 mm³. All control animals (without treatment) died within 20 days (Fig. 6A). In contrast, treatment with αCD40–HPV16.E6/7 plus poly(I:C) via either s.c. or i.t. route resulted in enhanced survival rates, although s.c. resulted in a greater survival rate than the i.t. route. Thus, αCD40-HPV16.E6/7 plus poly(I:C) could effectively inhibit further progression of TC-1 tumors (Fig. 6B).

Safe and effective therapeutic vaccines for HPV-associated malignancies will be highly beneficial for patients, including those who receive modern multidisciplinary therapeutic approaches. Currently available FDA-approved preventative vaccines do not provide therapeutic immunity against HPV-associated cancers. We therefore tested a new prototype vaccine, αCD40–HPV16.E6/7, for HPV16-associated malignancies in this study. Although there is still much work that needs to be done to bring this vaccine model to patients, we could demonstrate that this protein-based prototype vaccine αCD40–HPV16.E6/7 is highly immunogenic in patients in vitro and in hCD40Tg mice in vivo. These data strongly support our continuing investigation of this new vaccine model in different settings, which can lead us to test this vaccine model in patients in the near future.

αCD40–HPV16.E6/7 carries both E6_1–120 and E7_1–60. It could thus elicit broad repertoires of E6/7-specific CD4⁺ and CD8⁺ T-cell responses in patients. This may overcome one of the major limitations of HLA epitope-based peptide vaccines, which is that a specific vaccine is useful for only patients who express the corresponding HLA types. In line with this notion, αCD40–HPV16.E6/7 could activate HPV16.E6/7-specific T cells in the blood of head-and-neck cancer patients tested in this study. Peptide vaccines composed of overlapping peptides covering the majority of TAA could overcome such limitations of HLA epitope-based peptide vaccines. However, a long peptide-based immunotherapy by itself does not exert an effective therapeutic action in patients with established HPV-associated cervical cancer (14, 24). This could be due to an increased induction of FoxP3-positive HPV-specific (regulatory) T cells, immunoediting, including downregulation of MHC class I molecules or processing molecules, and failure of effector T cells to properly home into cancer tissues. Myeloid-derived suppressor cells (MDSC) in these patients can also preclude the efficacy of long peptide vaccine (37). Such MDSC-mediated immune suppression could be alleviated by depleting MDSC by timed vaccination during standard carboplatin and paclitaxel chemotherapy (37). It was also notable that αCD40–HPV16.E6/7 was more efficient than an HPV16.E6/7 peptide pool (15-mers overlapping 11 amino acids) at activating HPV16.E6/7-specific IFNγ-producing T cells in some of the patients tested in this study. Both patients in vitro and hCD40Tg mice in vivo data also confirmed that targeted delivery of HPV16.E6/7 to CD40 was far more efficient than nontargeted delivery at eliciting HPV16.E6/7-specific T-cell responses, as we have previously demonstrated (30). This protein-based vaccine model would be better suited for immune-compromised patients, including patients who are coinfected with HIV, than a live vector–based vaccine model. HPV-associated cancers, particularly cervical cancer in women and anal cancer in men who have sex with men (38, 39), are major HIV-related malignancies. Thus, the development of an effective and safe therapeutic vaccine against HPV-associated cancers in HIV patients is of importance.

The ability of a vaccine to evoke strong cellular immunity, particularly CD8⁺ CTL-mediated immunity, is one of the key components that could determine the efficacy of the vaccine against cancers, including HPV-associated malignancies. For the purpose of evoking strong CD8⁺ CTL-mediated immunity, this prototype vaccine was designed to target CD40, which was based on our previous studies (30, 31). After receptor-mediated antigen uptake, CD40 mainly localizes to the early endosomes and the plasma membrane, where the stability of the antigens it carries can be enhanced, leading to prolonged antigen presentation and enhanced CD8⁺ T-cell responses, as we and others have previously reported (30, 40, 41). αCD40–HPV16.E6/7 used in this study could not induce DCs to secrete increased amounts of cytokine and chemokine or to induce surface phenotype maturation (30). This was in line with the previous observation (40) that the enhanced antigen cross-presentation by CD40-targeted DCs was not due to the CD40-mediated activation signals. As shown in Supplementary Fig. S5A, the addition of an agonistic anti-CD40 (clone 12B4), did not significantly promote αCD40–HPV16.E6/7-induced HPV16.E6/7-specific CD8⁺ T-cell responses. The specificity of clone 12B4 was confirmed by staining CHO cells transfected with hCD40 (Supplementary Fig. S4B). It can bind to MoDCs (Supplementary Fig. S4C) and activate human B cells (Supplementary Fig. S4D). We have previously demonstrated that DCs are the major APCs that lead to antigen-specific CD4⁺ and CD8⁺ T-cell responses when antigens were targeted via CD40, although B cells and monocytes express CD40 (30, 31).

This study provides a proof-of-concept for the immunogenicity and efficacy of our prototype vaccine made with humanized αCD40 mAb. However, a few experiments might need to be performed before testing this vaccine model in patients. First, it would be valuable to explore methods that can further improve the immunogenicity and efficacy of this prototype vaccine. In this study, we used poly(I:C) as an adjuvant. This was based on our data as well as on the data from previous studies (35). As a part of our efforts to enhance T-cell responses, we tested whether montanide could further promote the vaccine-induced T-cell responses. It promoted CD4⁺ T-cell responses but not CD8⁺ T-cell responses. If this is the case in humans, montanide might improve efficacy of this vaccine model in patients because of the critical roles of CD4⁺ T cells in the maintenance of CD8⁺ CTLs (42–44). Other DC activators, including TLR7/8 ligands, particularly in the form of conjugates to the prototype vaccine, might also promote CD8⁺ T-cell–mediated immunity, as described previously (45, 46). Second, the route of immunization could be important in patients, although s.c., i.p., and i.m. delivery of αCD40-HPV16.E6/7 plus poly(I:C) resulted in similar outcomes in mice. Alternatively, an experimental protocol using non-human primates (NHP) may provide us with better insights for the selection of an optimal immunization route for evoking strong CD8⁺ T-cell responses. Such experiments in NHPs could also provide us with better insights for E6/7-specific CD4⁺ and CD8⁺ T-cell responses by assessing more than one single peptide epitope-specific CD4⁺ and CD8⁺ T-cell response, respectively, if possible. Although we and others use the TC-1 tumor model in C57BL/6 mice, it is also important to note that CD8⁺ CTL responses elicited in these mice are limited mainly to one epitope (E7_RAHYNIVTF). This will not be the case in humans, indicating that the efficacy of vaccines tested in the TC-1 tumor model does not necessarily reflect the efficacy in patients. The efficacy of this vaccine model needs to be investigated in patients once the vaccine regimen and immunization route are optimized. Third, it is also important to test whether this vaccine model is capable of evoking mucosal CD8⁺ CTL-mediated immunity. This might be one of the most important and relevant tasks for the successful development of vaccines against HPV-associated cancers in the mucosa. This question will need to be addressed in the context of the selection of an optimal immunization route as well as choosing appropriate adjuvants. Lastly, the efficacy of this vaccine model might be improved by harnessing tumor immune evasion mechanisms. Although HPV-induced tumors in different mucosal tissues might possess distinct inhibitory mechanisms in patients (and these need to be further studied), recent data suggest that antibodies specific for immune checkpoint inhibitors (47–50), including CTLA-4, PD-1, and PD-L1, might also improve the efficacy of this vaccine model in patients with HPV-associated malignancies.

In summary, we have presented a new immunotherapeutic vaccine model for HPV16-associated malignancies in this study. Although several key experiments remain to be performed, data from this study strongly suggest that this vaccine model has a high potential to provide patients with improved clinical benefits in the future. In addition, data from this and previous studies (30, 31, 40, 41) strongly support the development of CD40-targeting vaccines for the treatment of other cancers.

S. Zurawski and G. Zurawski have ownership interest as inventors on Baylor Research Institute patents covering test vaccine. A.G. Sikora reports receiving a commercial research grant from Advaxis. S. Oh is one of the owners of the patent on α-CDKO, which was used in the reported study. No potential conflicts of interest were disclosed by the other authors.

Conception and design: W. Yin, A. Clark, C. Beauregard, S. Zurawski, G. Zurawski, S. Oh

Development of methodology: W. Yin, Y. Xue, A. Clark, S. Kim-Schulze, G. Zurawski, S. Oh

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): W. Yin, D. Duluc, Y. Xue, C. Gu, R. Ouedraogo, L. Oxford, A. Clark, F. Parikh, S. Kim-Schulze, L. Thompson-Snipes, S. Zurawski, A.G. Sikora

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): W. Yin, D. Duluc, H. Joo, A.G. Sikora, S. Oh

Writing, review, and/or revision of the manuscript: W. Yin, H. Joo, R. Ouedraogo, L. Thompson-Snipes, S.-Y. Lee, C. Beauregard, A.G. Sikora, G. Zurawski, S. Oh

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C. Gu, Z. Wang, L. Wang, A. Clark, S. Oh

Study supervision: G. Zurawski, S. Oh

Other (provide quality study material for this study): J.-H. Woo

Other (design, development and production of vaccine used in study and acquisition of transgenic animals): S. Zurawski

We thank the Sample, Flow Cytometry, and Luminex Cores at BIIR. We thank Dr. Carson Harrod (BIIR) for reading this manuscript, Dr. Luz Muniz (BIIR) for helping us acquire patient samples, and Dr. T.C. Wu (Johns Hopkins University) for providing TC-1 cells.

This study was financially supported by the Baylor Health Care System Foundation (S. Oh), the American Cancer Society (S. Oh), Roche (S. Oh), and Tisch Cancer Institute at the Icahn School of Medicine at Mount Sinai (A.G. Sikora).

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.