Hope in the Long Road Toward the Development of a Therapeutic Human Papillomavirus Vaccine

In this issue of Clinical Cancer Research, van Poelgeest and colleagues report the clinical results of a therapeutic human papillomavirus (HPV) vaccine for the treatment of premalignant anogenital lesions (1). Currently, two prophylactic vaccines against oncogenic HPV16 and HPV18, the genotypes predominantly found in cancers, and a new vaccine directed against 9 HPV serotypes have demonstrated their efficacy in non-HPV–infected individuals. However, these vaccines failed to demonstrate any clinical activity in subjects already infected with HPV with 14 million new HPV infections each year in the world. In a subgroup of patients who do not spontaneously clear their infection, premalignant lesions will ultimately be followed by the development of invasive cancer of the cervix, anus, vagina, penis, and oropharynx. In addition, only a few countries have achieved vaccine coverage exceeding 50%. Therapeutic HPV vaccines (THV), therefore, correspond to a currently unmet medical need. van Poelgeest and colleagues report the ability of ISA 101 vaccine [13 HPV16 E6 and E7 synthetic long peptides mixed with the Montanide adjuvant] without additional treatment to induce objective partial or complete histologic regression of the lesions in more than 50% of patients with high-grade vulvar (VuVIN3) and vaginal intraepithelial neoplasia (VaIN3) with a follow-up of 12 months (1). Interestingly, 7 of 29 patients displayed complete response, which coincided with clearance of HPV16 likely to indicate complete cure. The only potential bias of this clinical trial was the absence of a control arm, but, as recalled by the authors, spontaneous regression of the lesions is observed in less than 1.5% of cases, a much lower rate than the clinical response observed after vaccination (>50%). This report therefore confirms the efficacy of ISA 101 to elicit durable complete regression in vulvar intraepithelial neoplasia (2). These results extend recent findings from a randomized control trial based on an HPV DNA vaccine, which induced regression of HPV16- or HPV18-positive CIN2/3 in 48.2% of cases compared with a spontaneous regression rate of 30% in the placebo arm (3).

A different HPV DNA vaccine administered in 9 CIN3 patients also induced complete regression of lesions and viral clearance in 7 patients (4). A recombinant Modified Vaccinia Ankara (MVA) HPV16 E6 and E7 cDNA vaccine achieved cytologic clearance of CIN2/3 lesions and clearance of HPV16 virus in previously infected tissue in 7 of 10 patients (5).

The first administration of THV was 1996 and, after many clinical failures, encouraging clinical results have now been observed in recent clinical trials. Many THV formats have been developed and the most advanced formats include long synthetic peptides, DNA vaccine, recombinant E6/E7 protein fused to Bordetella pertussis cyclase A, HSP65, B subunit of Shiga toxin, Pseudomonas aeruginosa exotoxin, as well as attenuated bacteria (Salmonella, Listeria monocytogenes) or viral vectors (alphavirus replicon particles, adenovirus, MVA; ref. 5). Most of these vaccines were designed to target the vaccine in dendritic cells and/or promote cross-presentation of HPV antigen to CD8⁺ T cells. The vaccine itself did not appear to be the critical element to explain the differential clinical activity of THV, as the clinical impact of HPV vaccines was only observed in patients with premalignant HPV-associated lesions, whereas none of the vaccines displayed any clinical activity in established cancer. In addition, the same vaccines (long synthetic peptide and MVA) which failed in HPV-associated cancer-induced significant clinical responses in premalignant lesions (5). It has been hypothesized that higher levels of immunosuppression and immune effector-evading mechanisms of tumor cells may explain the resistance of cancer to HPV vaccines.

In contrast to prophylactic vaccines, based on a humoral response, preclinical models support the main role of cytotoxic CD8⁺ T cells for THV. van Poelgeest and colleagues clearly demonstrated a positive correlation between the clinical effects of the HPV vaccine and the strength of the induction of IFNγ- and TNFα-producing T cells in the blood of vaccinated patients (1). Surprisingly, they did not observe a correlation between the ex vivo detectable HPV16-specific CD8⁺ T-cell response and clinical outcome. This absence of correlation of CD8⁺ T-cell response against HPV in the blood and clinical response has already been reported (5). It may mean that both CD4⁺ and CD8⁺ T-cell responses are required to achieve regression of the lesions. Other studies strongly support that intratumoral CD8⁺ T-cell infiltration is correlated with spontaneous regression of HPV lesions and the efficacy of the cancer vaccine. CIN2/3 lesions with CD8⁺ T-cell infiltrates in the dysplastic epithelium are more prone to clinical regression. In murine models and in an approved prostate cancer vaccine, only the induction of intratumoral CD8⁺ T cells was correlated with vaccine efficacy and not with CD8⁺ T-cell counts in blood (6, 7).

Results from the van der Brug group in this current issue of Clinical Cancer Research and from the Trimble group convincingly demonstrate the clinical efficacy of THV (1, 3). However, only a complete response would prevent surgical mutilation and such a response is observed in only about one-quarter of patients exhibiting vulvar/genital CIN3 lesions in the van Poelgeest report. These results may not be sufficient enough to modify clinical practice, except if predictive factors of response can be identified to allow better selection of patients likely to benefit from the vaccine. Prior to clinical implementation of these vaccines, strategies need to be developed to improve their clinical efficacy (Fig. 1). Cutaneous administration of imiquimod by van Poelgeest and colleagues did not increase the T-cell immune response elicited by the vaccine, but its impact on clinical response was not described (1). In murine models, peripheral vaccination with HPV16 E7 long peptides, followed by intravaginal administration of imiquimod, induced a 5-fold increase in the infiltration of vaccine-induced CD8⁺ T cells in the vaginal mucosa (8). The Wu group reported that cancer vaccine boosts via the cervicovaginal rather than the intramuscular route of immunization appears to be crucial to induce genital CD8⁺ T cells and tumor regression (9). In humans, a DNA HPV vaccine administered by cervical intralesional injections induced more intense recruitment of intraepithelial CD8⁺ T-cell infiltrates than the other routes of administration (10). These examples emphasize the critical role of mucosal administration of adjuvant or vaccine to specifically recruit resident memory T cells at the mucosal site of the lesions (11).

Strategies designed to counteract immunosuppressive mechanisms such as regulatory T cells, myeloid-derived suppressive cells or checkpoint inhibitors expressed by T cells in the tumor microenvironment may overcome clinical resistance to THV. HPV-positive tumors often present upregulated levels of PD-1 ligand on their surface and decreased levels of costimulatory molecules such as OX40 and 4-1BB on T cells. In preclinical models, blockade of these cells or molecules synergized with HPV vaccine, but it was not always the same antagonist or agonist of costimulatory molecules that optimally synergized with selected HPV vaccines (12–14). Several clinical trials are currently examining the effects of checkpoint inhibitors such as anti–CTLA-4 and/or anti–PD-1/PD-L1 in advanced stage HPV-associated disease (NCT01711515, NCT01693783, NCT01975831). Second-generation THV integrating the requirement of mucosal route of immunization to elicit protective resident memory T cells and combined with an antagonist of costimulatory inhibitory molecules or an agonist of positive cosignaling pathways may finally lead to the clinical success of THV after a long, winding road.

No potential conflicts of interest were disclosed.

Conception and design: S. Karaki, E. Tartour

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Karaki, C. Badoual, E. Tartour

Writing, review, and/or revision of the manuscript: S. Karaki, H. Pere, C. Badoual, E. Tartour

Study supervision: S. Karaki, E. Tartour

E. Tartour was supported by grants from Institut National du Cancer (INCA), Ligue Contre le Cancer, Université Sorbonne Paris Cité, LabEx Immuno-Oncology, Site Intégré de Recherche en cancerologie (SIRIC CARPEM), and Agence Nationale de la Recherche (ANR).