An Ad26‐MVA‐BN‐based therapeutic vaccine targeting HPV16 and HPV18 related disease is immunogenic in preclinical models and in women with persistent HPV infections

Abstract

Objectives

To develop a therapeutic human papillomavirus (HPV) vaccine capable of intercepting HPV‐related cervical disease to reduce the number of new cancer cases and associated deaths.

Methods

We previously reported that replication‐deficient adenovirus type 26 and 35‐based vectors (Ad26, Ad35) expressing a fusion protein of E2, E6 and E7 (E2SH) from HPV16 and HPV18 induced robust HPV‐specific T‐cell immunity capable of reducing HPV16+ tumors in mice. Here, we assessed immune responses induced by Ad26 and Modified Vaccinia Virus Ankara—strain Bavarian Nordic (MVA‐BN) viral vectors expressing E2SH in heterologous vaccination regimens in mice, macaques and a limited number of persistently HPV‐infected women (N = 5).

Results

The magnitude and breadth of cellular immune responses were higher in mice that received Ad26.HPV16/18‐MVA‐BN.HPV16/18 than in those that received Ad26.HPV16/18‐Ad35.HPV16/18, all expressing the fusion protein E2SH of HPV16 and HPV18. Sustained and broad antigen‐specific systemic T‐cell responses against HPV16 and HPV18 proteins were induced in macaques. In the Phase 1/2a trial, a 2‐dose heterologous immunisation with Ad26.HPV16 or Ad26.HPV18 in Week 0 and MVA‐BN.HPV16/18 in Week 8 induced long‐lived antigen‐specific cytokine‐producing T cells. After Week 8, HPV16/18 infection was no longer detected in 5/5 active vaccines and 1/4 placebo recipients during the study period of approximately 1 year. The small number of vaccinated participants did not allow concluding on the relationship between immune responses and viral clearance.

Conclusion

These data show the ability of Ad26.HPV16/18‐MVA‐BN.HPV16/18 heterologous regimens to elicit HPV‐specific T‐cell responses and warrant further clinical studies to evaluate vaccine efficacy.

Immunisation with adenovirus (Ad) vectors followed by Modified Vaccinia Virus Ankara—strain Bavarian Nordic (MVA‐BN) vectors expressing HPV16 and HPV18 engineered antigens induced antigen‐specific polyfunctional T cells in mice, rhesus macaques and persistent HPV16/HPV18 infected human females vaccines (n = 5). These data showed the promise of the Ad26‐MVA‐BN combination for therapeutic vaccination against persistent HPV16/HPV18 infections. Further clinical studies are needed to evaluate the efficacy of the Ad26‐MVA‐BN therapeutic HPV vaccine candidate to stop progression towards more severe disease.

Introduction

Infections with so‐called high‐risk human papillomaviruses (hrHPVs) are the leading cause of cervical cancer despite the widespread use of an effective preventive vaccine, with HPV types HPV16 and 18 accounting for approximately 70% of all cases. ¹ HrHPV has a global prevalence of 6.1% in women with a normal cervix, although there is substantial geographic variability (from 1.4% in Spain to 25.6% in Nigeria). ² In 2020, there were an estimated 640 000 new cases of cervical cancer worldwide and 342 000 deaths. ³ , ⁴ , ⁵

Moreover, hrHPV, and especially HPV16, plays a causal role in various other anogenital cancers with HPV16 and HPV18, encompassing around 50% of penile cancers, 88% of anal cancers, 43% of vulvar cancers and 70% of vaginal cancers. ⁶ , ⁷

HrHPV infections are also the leading cause of oropharyngeal cancer (OPC) ⁸ , ⁹ with an estimated proportion of OPCs driven by HPV ranges from 31% to 43%, globally. ¹⁰ HPV‐associated OPC affects both sexes, but with higher incidence in men than in women. ⁹ , ¹¹ In contrast to cervical cancer, HPV‐associated OPCs are more prevalent in high‐income countries than in low‐ and middle‐income countries. ⁸

Diagnosis of early‐stage cervical neoplasia through regular screening and surgical treatment to prevent progression has significantly reduced the incidence of cervical cancers in high‐income countries. ¹² Furthermore, prevention of a substantial proportion of cervical cancers is now possible through prophylactic vaccination based on the L1 antigen prior to the onset of sexual activity. ¹³ , ¹⁴ However, the uptake of prophylactic HPV vaccination has lagged behind other immunisations, ¹⁵ , ¹⁶ and women and men who were not vaccinated prior to infection with a hrHPV type remain at risk of developing HPV‐related cancers.

The majority of HPV‐induced low‐grade cervicovaginal lesions resolve spontaneously, ¹¹ , ¹⁷ but infected women are at increased risk for progression towards high‐grade neoplasia, especially if the infection becomes persistent. A therapeutic HPV vaccine that would intercept early in the disease cascade could reduce the number of surgical interventions for high‐grade disease as well as new cancer cases and associated deaths. Importantly, such a vaccine could reduce anxiety in HPV‐positive individuals who would otherwise be relegated to a watch‐and‐wait approach after being found positive after screening.

We have previously reported on the development of a vaccine candidate consisting of replication‐deficient adenovirus type 26 (Ad26) and 35 (Ad35) vectors encoding HPV16‐ and HPV18‐specific antigens consisting of fusion proteins of modified E2, E6 and E7 proteins (E2SH). ¹⁸ Notably, E6 and E7 are present in the candidate vaccine as re‐ordered fragments to abrogate the transforming activity of these two proteins. ¹⁸ The vaccine candidate aims to induce strong cellular immune responses targeting HPV‐infected cells expressing the HPV antigens E2, E6, and/or E7, leading to their eradication and hence, to clearance of the viral infection.

A candidate vaccine based on Ad26 and Ad35 viral vectors elicited potent HPV16‐ and HPV18‐specific T‐cell immune responses in murine models ¹⁸ and therapeutic efficacy was demonstrated in the TC‐1 mouse model which expresses E6 and E7 protein of HPV16. Combining two different (vector‐based) vaccine platforms in a heterologous prime‐boost immunisation schedule shows higher potency compared with homologous immunisation ¹⁹ ; for instance, individuals who received a 2‐dose heterologous Ad26‐Modified Vaccinia Virus Ankara—strain Bavarian Nordic (MVA‐BN) Ebola vaccine induced stronger durable immune responses compared with those who received a 2‐dose homologous Ad26‐Ad26 vaccine regimen. ²⁰ Similarly, a significantly larger increase was seen in SARS‐CoV‐2 spike binding antibodies in individuals who received a prime with Ad26.COV2.S followed by a booster with either mRNA‐1272 or BNT16b2 compared with individuals who received a homologous Ad26.COV2.S prime‐boost regimen. ²¹ Here, we explored a heterologous 2‐dose immunisation strategy using adenoviral vectors Ad26 in combination with MVA‐BN, and we characterised vaccination‐induced HPV‐specific T‐cell responses in mice, macaques and humans.

Results

Increased magnitude and breadth towards HPV16/18 antigens after Ad26.HPV16/18‐MVA‐BN.HPV16/18 dosing compared to Ad26.HPV16/18‐Ad35.HPV16/18 dosing in mice, while prevention of TC‐1 tumor outgrowth control remains comparable

We have previously shown that Ad26 and Ad35 vectors encoding HPV16 and HPV18 antigens induce a strong T‐cell response in mice that is able to increase the survival of tumor bearing mice. ¹⁸ Here, we investigated whether the HPV‐specific immune response elicited by one dose of Ad26‐based vectors followed by an MVA‐BN vector, both encoding HPV antigens, would induce a similar or higher immune response compared with that induced by Ad26 followed by an Ad35 vector in mice.

Mice (n = 11/group) were vaccinated with Ad26 in combination with MVA‐BN or Ad35 as outlined in Figure 1a. and cellular immune responses were measured through IFNγ ELISpot and ICS on splenocytes 2 weeks after the boost immunisation (Figure 1a).

Figure 1.

Heterologous immunisations with Ad26‐MVA‐BN viral vectors expressing HPV16 or HPV18 antigens increase transgene‐specific immunity in mice. CB6F1 mice (n = 11 per group) were immunised with Ad26.HPV16 and Ad26.HPV18, or with Ad26 not encoding an antigen (Empty). Eight weeks later mice were dosed with MVA‐BN.HPV16/18 or Ad35‐vectors encoding the same transgene as during the first immunisation, or MVA‐BN.control (a). At Week 10, mice were euthanised and the induction of HPV16‐E2, HPV16‐E7, HPV18‐E2 and HPV18‐E6‐specific IFNγ‐producing cells were determined by IFNγ ELISpot (b, c). Displayed are splenocytes producing IFNγ in SFU/10⁶ cells in response to the different peptide stimuli for each animal. The horizontal dotted line represents the lower limit of detection in the ELISpot assay. A Student's t‐test with unequal variance was performed on log₁₀ transformed ELISpot data (b). The breadth of the immune response was calculated per animal per group, IFNγ ELISpot responses above 100 SFU/10⁶ cells were considered positive and groups were compared using a Wilcoxon rank sum test (c). Splenocytes were stimulated with HPV16‐E7 or HPV18‐E6 peptides, and processed for ICS as previously described. Cells were gated on CD3+CD8+ lymphocytes, and induction of IFNγ+TNF‐α+ CD8 T cells determined. A Student's t‐test with unequal variance was performed on log₁₀ transformed ICS data. The horizontal dotted line represents the lower limit of detection (0.037%) (d). The red horizontal lines represent the group mean.

Significant increases in IFNγ‐producing cells specific for HPV16‐E2, HPV16‐E7, HPV18‐E2 and HPV18‐E6 were observed in animals boosted with MVA‐BN.HPV16/18 compared to the cohort administered Ad35 vectors (Figure 1b) (HPV16‐E2: MVA‐BN.HPV16/18 170 ± 26 SFU/10⁶ vs Ad35.HPV16/18 37 ± 5 SFU/10⁶; HPV16‐E7: MVA‐BN.HPV16/18 1274 ± 110 SFU/10⁶ vs Ad35 549 ± 78 SFU/10⁶; HPV18‐E2: MVA‐BN.HPV16/18 72 ± 20 SFU/10⁶ vs Ad35.HPV16/18 16 ± 3 SFU/10⁶; HPV18‐E6: MVA‐BN.HPV16/18 1554 ± 38 SFU/10⁶ vs Ad35.HPV16/18 914 ± 110 SFU/10⁶; P < 0.05, Student's t‐test with unequal variance on log₁₀‐transformed SFU/10⁶ cells).

The HPV16‐E6 and HPV18‐E7 specific responses were at the lower limit of detection (data not shown), and in line with our published data ¹⁸ and observations by others.

The breadth of the immune response, quantified as the number of antigens to which each animal mounted a response, was higher in animals that received the Ad26.HPV16/18‐MVA‐BN.HPV16/18 combination (where 6 out of 11 animals had a response to 4 antigens) compared with animals receiving Ad26 and Ad35 vectors (where 0 animals out of 11 had a response to four antigens) (Figure 1c, P < 0.05, Wilcoxon Rank Sum test).

To further characterise the immune response, induction of cytokine‐producing CD4 and CD8 T cells was assessed by intracellular cytokine analysis. The cytokine responses within the CD4+ population were low for all antigens (data not shown). In contrast, a similar number of IFN‐γ+ TNF‐α+ positive CD8 T cells specific for HPV16‐E7 were detectable in animals boosted with either Ad35.HPV16/18 or MVA‐BN.HPV16/18 vectors (Figure 1d, representative FACS plots can be found in Supplementary figure 2), while a significantly higher frequency of HPV18‐E6 specific T cells was measured in animals boosted with MVA‐BN.HPV16/18 compared with Ad35.HPV16/18 (P < 0.0006, Student's t‐test). The frequency of IFNγ+ single cytokine‐producing cells were lower in both vaccine groups compared with the IFN‐γ+ TNF‐α+ frequency (Supplementary figure 2), and the frequency of TNF‐α+ CD8 T‐cell responses were in the same range as those induced by the negative vaccine control.

We compared the efficacy of Ad26.HPV16/18‐MVA‐BN.HPV16/18 vaccine to that induced by Ad26.HPV16/18‐Ad35.HPV16/18 vaccine schedule, using the TC‐1 mouse model. ²² We shortened the vaccination interval from 8 to 2 weeks to adjust for rapid tumor growth in this model. Both vaccination strategies led to a significant delay in the initial tumor outgrowth compared with the control mice (median survival time: 27 days, range: 18–31 days) (Figure 2) while the overall median survival time for animals receiving Ad26.HPV16/18‐MVA‐BN.HPV16/18 (55 days, range: 38–90 days) was comparable with that of Ad26.HPV16/18‐Ad35.HPV16/18 (58 days, range 40–90) (P ≤ 0.0001, Mantel‐Cox test).

Figure 2.

Ad26.HPV16/18‐MVA‐BN.HPV16/18 prime‐boost vaccination reduces established tumor outgrowth. C57BL/6 mice (n = 12) received 50 000 TC‐1 cells and 6 days later were immunised with Ad26.HPV16 and Ad26.HPV18 or with Ad26.Control. Booster immunisations followed 14 days later with MVA‐BN or with Ad35 encoding the same transgene used for prime immunisation. Animals were dosed at 1 × 10¹⁰ vp/adenovector/immunisation or with 8.9 × 10⁷ Inf.U MVA‐BN/immunisation. Tumor growth was measured over time (a), and survival curves in different treatment groups (b). Statistical analysis was conducted using a log‐rank test on survival duration. P and B refer to Prime and Boost immunisation.

Ad26.HPV16/18‐MVA‐BN.HPV16/18 (HPV16/18) vaccination induces cellular immune responses in rhesus macaques

Next, the Ad26‐MVA‐BN vaccine combination was further evaluated in non‐human primates, a model that is regarded as a relevant immunogenicity model for Ad26‐and MVA‐BN‐based vaccines designed for humans. ²³ Animals (n = 7) were dosed as outlined in Figure 3a and peripheral blood mononuclear cells (PBMCs) were collected from animals at defined time points to longitudinally assess the cellular immune responses by IFNγ ELISpot using peptide pools spanning the wild‐type sequence of E2, E6 and E7 of HPV16 and HPV18. Two weeks after the Ad26 immunisation four out of seven animals elicited a response to one to two antigens (Figure 3 and Supplementary figure 3). By Week 8, prior to MVA‐BN dosing, all animals had a response to at least one antigen, albeit no response was seen towards HPV16‐E7 or HPV18‐E6, and a very low but detectable HPV18‐E7‐response was observed in one animal (Supplementary figure 3F). Of note approximately 58% of total IFNγ‐producing cells were directed against HPV16 antigens versus 42% towards HPV18. Post‐MVA‐BN dosing, a significant enhancement of > sevenfold (P < 0.05, paired t‐test) in the total immune response was seen at Week 13 (3153 SFU/10⁶ cells ±1192 SEM) compared with pre‐MVA‐BN dosing at Week 2 (401 SFU/10⁶ cells ±151 SEM) (Figure 3b). Post‐MVA‐BN dosing all animals elicited a response to two or more of the antigens (average response 3.5 antigen per animal) (Figure 3c and Supplementary figure 3). Approximately 70% of the total IFNγ‐producing cells were directed against HPV16 antigens versus 30% towards HPV18. Of the animals having a response at Week 13, the hierarchy of the immune response was HPV16‐E6 (1396 SFU/10⁶ ± 363 SEM) > HPV18‐E2 (905 SFU/10⁶ ± 320 SEM SFU/10⁶) > HPV16‐E2 (587 SFU/10⁶ ± 164 SEM) > HPV16‐E7 (170 SFU/10⁶ ± 62 SEM) > HPV18‐E6 (94 SFU/10⁶ ± 25 SEM), while no animals had a response to HPV18‐E7 (Supplementary figure 3). Notably, HPV16‐ and HPV18‐specific IFNγ‐producing cells were still detectable at 16 weeks after the MVA‐BN.HPV16/18 immunisation.

Figure 3.

Induction of HPV16 and HPV18 specific IFNγ‐producing cells after Ad26.HPV16/18‐MVA‐BN.HPV16/18 immunisations in rhesus macaques. Seven female rhesus macaques were immunised at Week 0 with Ad26.HPV16+ Ad26.HPV18 (10¹¹ VP/vector). At Week 8, all animals received MVA‐BN.HPV16/18 (1.8 × 10⁸ Inf.U) (a). PBMCs were collected in Weeks −4, 2, 8, 13, 16 and 24. The cellular immune response was assessed using ELISpot analysis, by measuring the induction of IFNγ‐producing cells after stimulation of PBMCs with HPV16 and HPV18 E2‐, E6‐ or E7‐specific peptide pools. The total mean group response per antigen is depicted and error bars are standard error of means. A paired t‐test on log₁₀ transformed ELISpot data was conducted on the total responses of Week 2 vs Week 13 and Week 13 vs Week 24 (b). The breadth of the immune response was calculated as indicated in the material and method section to determine the number of antigens to which an animal responded (c). A Wilcoxon signed rank test with a Bonferroni correction was applied comparing Week 13 with Week 2, and Week 24 with Week 13.

Induction of durable cellular immune response following Ad26.HPV16/18‐MVA‐BN.HPV16/18 vaccination in women with persistent HPV16 or HPV18 infections

Based on these preclinical results, a first‐in‐human clinical trial was initiated to evaluate the safety, reactogenicity and immunogenicity of a two‐dose regimen of Ad26.HPV16 or Ad26.HPV18 and MVA‐BN.HPV16/18 (NCT03610581). The time interval between the two doses was 8 weeks, the same as used in the animal studies. Initial vaccine dose levels and sequence of first dosing Ad26 followed by MVA were based on previous experience using Ad26 and MVA in human studies for other indications. In these studies, we observed that MVA priming was inferior in non‐human primates and in clinical studies, while an Ad26 prime followed by MVA boost produced more favourable cellular immune responses that were protective in macaques ²⁰ , ²⁴ , ²⁵ and we have therefore not investigated a first dose of MVA in this study. Depending on whether individuals had HPV16 or HPV18 (persistent) infection, participating women were either dosed with Ad26.HPV16 or Ad26.HPV18, respectively, for induction of optimal antiviral immunity. A total of nine participants with confirmed HPV16 (N = 5) or HPV18 (N = 4) infection at baseline were enrolled (Figure 4a). For HPV16+ confirmed participants, two completed the full Ad26.HPV16 (5 × 10¹⁰ vp) and MVA‐BN.HPV16/18 (2 × 10⁸ Inf.U) vaccination regimen, one participant only received a single dose of Ad26.HPV16 (and was excluded from the immunogenicity assessment), and two received two doses of placebo. For HPV18+ confirmed participants, two completed the full Ad26.HPV18 (5 × 10¹⁰ vp) and MVA‐BN.HPV16/18 (2 × 10⁸ Inf.U) vaccination regimen, and two received two doses of placebo.

Figure 4.

Study CONSORT diagram and characterisation of HPV16‐specific T‐cell responses induced by Ad26.HPV16/18‐MVA‐BN.HPV16/18 vaccination in women with persistent HPV16 infection. CONSORT flow diagram showing disposition of trial participants (a). Four HPV16‐infected women received Ad26.HPV16 and MVA‐BN.HPV16/18 (N = 2), or 2 doses of placebo (N = 2): (b) IFNγ ELISpot and (c, d) ICS assays were performed using HPV16‐E2 and HPV16‐E6/E7 peptide pools to stimulate PBMCs collected at pre‐vaccination (Day 1), Day 57, Day 78, Day 239 and Day 366 post‐vaccination. For ICS, CD4 and CD8 T‐cell responses were evaluated by flow cytometry measuring IFNγ and/or IL‐2 expression. The individual responses for ELISpot (SFU per 10⁶ PBMCs) and ICS (% IFNγ+ and/or IL‐2+ CD4 or CD8 T cells) are shown for the separate and combined E2 and E6/E7 peptide pools for the HPV16 type. Individual patients are indicated by Pt and number.

Safety and reactogenicity were the primary objectives of the trial. Overall, both Ad26‐ and MVA‐BN‐based vaccination were well tolerated, with no Grade 3 or higher adverse events observed. The most frequent solicited adverse events in the active group were fever, headache, myalgia, nausea and injection‐site pain. A detailed evaluation of the safety and tolerability is shown in Supplementary table 1 listing solicited adverse events and Supplementary table 2 listing unsolicited and serious adverse events.

Cellular immunogenicity was a secondary objective of the study and was evaluated by IFNγ ELISpot and ICS using PBMCs isolated pre‐and post‐vaccination and stimulated ex vivo with peptide pools covering E2 and E6/E7 of HPV16 or HPV18.

No HPV‐specific IFNγ responses were detected in any group at pre‐vaccination when measured by ELISpot (Figure 4 and Supplementary figure 4). By Week 8 following Ad26.HPV16 vaccination, participants had detectable IFNγ responses to both HPV16 E2 and E6/E7 peptide pools as measured by ELISpot (Figure 4b), with higher E2 responses. Administration of MVA‐BN.HPV16/18 (Week 8) in previously Ad26.HPV16‐administered participants elicited a robust increase in both HPV16 E2 (2‐ to 3.6‐fold) and E6/E7 (2.3‐ to 6‐fold)‐specific IFNγ responses compared to pre‐MVA‐BN vaccination, with the HPV16‐E2 and HPV16‐E6/E7 peptide responses being at either 713 and 260 SFU/10⁶ PBMCs (participant #5) or 1525 and 693 (participant #7) SFU/10⁶ PBMCs, respectively. IFNγ production by PBMCs was detectable up to at least 6 months (N = 2) and 1 year post‐dose 1 (N = 1), with sample not being available from the other participant. Participants who received Ad26.HPV18/MVA‐BN.HPV16/18, or placebo, had no detectable HPV18‐specific IFNγ responses as measured by IFNγ ELISpot (Supplementary figure 4A).

ICS showed similar trends for participants primed with Ad26.HPV16 and boosted with MVA‐BN.HPV16/18 as for the ELISpot output in terms of response to vaccination and antigen hierarchy. HPV16‐E2 and HPV16‐E6/E7‐specific IFNγ‐ or IL2‐producing CD4 and CD8 T‐cell responses were low post‐Ad26.HPV16 priming and the responses were boosted post‐MVA‐BN, with higher frequency of CD8 than CD4 T cells in both HPV16‐positive vaccines (Figure 4c and d). Induction of HPV16‐E2‐specific CD107a+ CD8 T cells was detected in these two HPV16 vaccines that received both Ad26.HPV16 and MVA‐BN.HPV16/18 and one individual of them also had E6/E7‐specific CD107a+ CD8 T cells (Supplementary figure 5A). Participants who received Ad26.HPV18 and MVA‐BN.HPV16/18, or two doses of placebo, had no detectable HPV18− specific cytokine‐producing CD4 T‐cell responses (Supplementary figure 4B); however, low cytokine‐producing CD8 T‐cell responses were observed in one participant which was dosed with Ad26.HPV18 and MVA‐BN.HPV16/18 (participant #8). A transient HPV18 E2 and E6/E7‐specific CD107a+ CD8 T‐cell response was seen in participant #8 at day at 239 (Supplementary figure 5B).

HPV quantification and genotyping was performed at multiple time points throughout the study on either physician‐taken or self‐collected samples using two independent assays (Supplementary data [Link], [Link]). Except for 1 participant (#6) who had a non‐persistent HPV16 infection, all participants had persistent HPV16 or HPV18 infection at study entry (Figure 5). Seven of eight participants with an evaluable baseline sample were still HPV16 or HPV18 positive by either qualitative or quantitative HPV test at the time of first vaccination. Three of four placebo‐treated participants remained persistently HPV16 or HPV18 positive during the study. All four persistently infected participants in the active regimen groups who were HPV16 (n = 2) or HPV18 (n = 2) positive at baseline became HPV16‐ or HPV18‐negative during the study.

Figure 5.

PCR‐based quantitative and qualitative HPV infection testing. Vertical dotted and dashed lines represent the time point of study vaccination with placebo (a, c), Ad26.HPV16 and MVA.HPV16/18 at Day 1 and Week 8, respectively (b) or Ad26.HPV18 and MVA.HPV16/18 at Day 1 and Week 8, respectively (d). Patient 6 did not receive MVA.HPV16/18 at Week 8 and Patient 2 withdrew from the study prior to Week 52. Results shown in green and blue represent the analysis of physician‐taken cervical sample (obtained at Day 1 and Week 52/Early Withdrawal) and self‐samples (obtained at Screening, Day 1, Week 8, Week 11, Week 16, Week 26, Week 34 and Week 52/Early Withdrawal), respectively. Quantitative HPV16 or HPV18 levels, as determined by the AML HPV assay (left y‐axis), are shown with a line graph over time for the self‐samples or with a ○ for the physician‐taken sample. Qualitative HPV16 or HPV18 levels, as determined with the Roche COBAS HPV assay, are shown on top (positive result indicated with ‘+’ and negative result with ‘−’, right y‐axis). When the symbol is missing, no sample data were available.

Discussion

Here, we show that heterologous HPV immunisation with Ad26 and MVA‐BN (encoding E2, E6 and E7 of HPV16 and HPV18) is immunogenic in mice and induces a higher frequency and antigen breadth of cytokine‐producing cells compared to an earlier described Ad26‐Ad35 vaccination schedule. ¹⁸ In rhesus macaques, the immune responses after a full Ad26.HPV16/18‐MVA‐BN.HPV16/18 regimen were primarily directed against the E2 antigen of HPV16 or HPV18 and against E6 of HPV16. Approximately 70% of total IFNγ‐producing cells were directed against HPV16 antigens versus 30% towards HPV18. Although the number of participants was limited, the results from the Phase 1/2a clinical study indicated that a single dose of Ad26.HPV16 followed by MVA‐BN.HPV16/18 mounted IFNγ+ CD8 T‐cell responses and cytotoxic CD8 T‐cell responses (CD107a+) post‐MVA‐BN.HPV16/18 administration. Notably, these responses were detectable up to at least 1 year after the last dosing suggesting induction of a durable cellular immune response. Analogous to the results in the rhesus macaques, responses against HPV16 antigens in human participants were higher than against HPV18 antigens.

So far, no licensed therapeutic HPV vaccine exists, and multiple different vaccine platforms (peptide, nucleic‐acid and viral vector) ²² , ²⁶ , ²⁷ , ²⁸ , ²⁹ are being tested in the clinic for treatment of HPV‐induced disease and cancers. ³⁰ , ³¹ Most of these vaccines are focused on the treatment of high‐grade lesions and cancers. They typically contain E6/E7 antigens and have as stand‐alone treatment with up to 50% efficacy, possibly because of an immune‐suppressing environment. Therefore, for the treatment of HPV‐positive cancer, a combination treatment with immune‐checkpoint inhibitors is likely required to overcome the immune suppression. ³⁰ This was recently demonstrated by Massarelli et al. ³² showing that combining nivolumab (a PD‐1 targeting antibody) with a synthetic long peptide based HPV vaccine in patients with incurable HPV16‐positive cancers led to an approximately doubling of the median overall survival (17.5 months) compared with PD‐1 inhibition alone. Targeting of low‐grade lesions or persistent infection does not warrant the use of immune checkpoint inhibitors because of the extensive toxicity associated with these immune modulators, ³³ including liver toxicity. ³⁴ Here, immune intervention by vaccines that are sufficiently strong to induce an immune response able to clear the lesion and persistent infection could be a solution and might be more successfully used than for treatment of high‐grade lesions. During persistent HPV infection and in the presence of only low‐grade lesions (CIN1), the E2 protein is expressed at higher levels compared to the E6 and E7 oncogenes. The majority of the HPV vaccine modalities assessed in clinical trials lack the E2 component. Consequently, it is anticipated that these vaccines may exhibit reduced efficacy in clearing persistent infection and low‐grade lesions. Interestingly, recently van der Burg et al. reported the presence of HPV16 E2‐specific polyfunctional cytokine‐producing T cells in ex vivo expanded tumor infiltrating lymphocyte cultures isolated from HPV16+ oropharyngeal squamous cell carcinoma. Although the authors indicate, they did not measure the expression of the E2 antigen in the tumors of the patients, these data suggest that a vaccine encoding E2 antigen, like ours, may be of benefit not only for treatment of persistent HPV infection but also HPV‐positive cancers. ³⁵

Interestingly, several groups recently reported preclinical data using additional HPV antigens next to E6 & E7 documenting superior potency. ³⁶ , ³⁷ , ³⁸ , ³⁹ Hancock et al. ³⁷ reported on the combination of a Chimpanzee adenovirus vector and MVA based multi‐genotype HPV vaccine encoding selected conserved regions from E1, E2, E4, E5, E6 and E7 from HPV genotypes 16, 18, 31, 52 and 58, which induced systemic and mucosal immune responses in mice. Moreover, T‐cell‐specific for the vaccine‐transgene insert were present in PBMCs isolated from women previously exposed to hrHPV. ³⁷ While we have not detected such E2/E6/E7 specific T‐cell responses prior to dosing in our persistently HPV‐infected study subjects, we detected HPV‐specific immune responses against E2, E6 and E7 after vaccination. Although the sample size was low, none of the vaccine dosed women was positive for HPV from 10 weeks after vaccination onwards, in contrast to three out of four placebo subjects. To draw definite conclusions of the vaccine efficacy, a longer follow up and a larger number of participants would be required.

Our study demonstrates that heterologous vaccination with adenoviral vectors and MVA‐BN encoding fusion proteins of the E2 combined with transformation‐deficient forms of E6 and E7 early antigens from HPV16 and HPV18 induces cellular immune responses in preclinical models and humans. The encouraging, albeit limited data from the first‐in‐human study warrant further clinical evaluation to understand the potential of this vaccine to mediate viral HPV clearance in a larger cohort.

Methods

Cell lines

TC‐1 cells [Mouse lung] (RRID: CVCL_4699; John Hopkins Medical Institution, Baltimore, Maryland) were cultured exactly as previously described. ¹⁸ All experiments were performed with mycoplasma‐free cells.

Generation of replication‐deficient ad‐HPV16 and HPV18 vectors

E1/E3‐deleted replication‐deficient Ad26 and Ad35 expressing the HPV16 E2SH or HPV18 E2SH sequences under control of a CMV promoter were constructed, amplified, purified and characterised as described. ⁴⁰ In our previous study, we annotated the candidate vaccine inserts as HPV16 E2SH /HPV18 E2SH. For simplicity, we have since updated the nomenclature to HPV16 and HPV18. Henceforth, the Adeno vector‐based vaccine will be referred to as Ad26/Ad35.HPV16 and/or Ad26/Ad35.HPV18 throughout this article.

Generation of modified Ankara vaccinia HPV16/HPV18 vector

MVA‐BN encoding HPV16 E2SH and HPV18 E2SH antigens (abbreviated as MVA‐BN.HPV16/18) was generated under serum free conditions and characterised as previously described ⁴¹ , including conventional PCR to confirm the absence of wild‐type MVA‐BN, along with insert and adjacent MVA‐BN backbone sequence integrity analysis. Expression of the HPV16‐E2SH and HPV18‐E2SH fusion protein was confirmed by Western blot using a HPV16‐E7 or HPV18‐E7 specific antibody, respectively (Supplementary figure 1).

The cDNAs encoding the HPV16 and HPV18 E2SH fusion proteins ¹⁸ were synthesised at GeneArt (Thermo Fisher Scientific), Regensburg, Germany. For optimal expression, the individual cDNAs were combined with suitable vaccinia early or early/late promoters (expression of the HPV16 fusion protein is driven by the PrMVA13.5‐long promoter), ⁴² and HPV18 fusion protein expression by the synthetic pHyb promoter. ⁴³ The MVA‐BN vector (abbreviated as MVA‐BN.control) is deposited at the European Collection of Cell Cultures, Salisbury, UK under number V00083008. Primary chicken embryo fibroblast (CEF) cells were cultured in virus production‐serum free medium VP‐SFM (Gibco) and used during all steps of recombinant MVA‐BN‐based vaccine generation and production of purified research grade stocks used during preclinical studies.

Animals and housing

Six‐to‐eight‐week‐old specific pathogen‐free female CB6F1 (C57BL/6 × Balb/c) or C57BL/6 mice were purchased from Charles River and kept at the institutional animal facility under specified pathogen‐free conditions.

A total of seven 3‐ to 4‐year‐old, healthy female rhesus macaques (Macaca mulatta) of Indian origin (body weight 4–6 kg at study start) were purchased from Covance (Alice, TX, USA). Food and water were provided ad libitum.

Rhesus macaques were specific pathogen‐free by serology for the following pathogens: B virus, STLV, SRV and SIV. Animals underwent physical examination, laboratory testing, TB testing and quarantine and received standard husbandry for primates as per institution protocol. Moreover, all animals were negative for TB. The Ad neutralising antibodies were not screened pre‐study enrolment. Food and water were provided ad libitum.

Animal procedures

Mice were immunised intramuscularly (IM) with two replication‐deficient Ad26.HPV16 or/and Ad26.HPV18 vectors, at doses of 10¹⁰ virus particles (vp) per vector type. Eight weeks later a second immunisation was given at a dose of either 10¹⁰ vp of replication‐deficient Ad35 encoding the same inserts as during first immunisation, or 8.9 × 10⁷ Infectious Units (Inf.U) MVA‐BN.HPV16/18. Control mice received Ad26 and MVA‐BN vectors not encoding an antigen Ad26.Empty (10¹⁰ vp) and MVA‐BN.control 8.9 × 10⁷ Inf.U, respectively. Vaccines were formulated in injection buffer and administered in the quadriceps of both hind legs (50 μL/leg) immunisation was performed under anaesthesia with isoflurane. For the tumor challenge study, TC‐1 cells (5 × 10⁴) were injected SC in the flank of naïve C57BL/6 mice exactly as previously described. ¹⁸ When tumors were palpable (~9 mm³), mice were vaccinated with Ad26.HPV16 and Ad26.HPV18 vectors. Two weeks later, a second immunisation was administered with Ad35 encoding the same insert as during the first immunisation, or with MVA‐BN.HPV16/18. The doses of the vaccines were described above. Tumor growth was monitored one to three times per week using digital calliper measurement in two dimensions. Tumor volume was calculated as [(width² × length)/2]. Animals were sacrificed for ethical reasons when tumors reached a maximum of 1000 mm³.

The rhesus macaques received 1011 vp of Ad26.HPV16 and 1011 vp of Ad26.HPV18 at Week 0 and 1.79 × 10⁸ Inf.U of MVA‐BN.HPV16/18 at Week 8. All vaccines were administered IM in a 0.5 mL volume in the quadriceps. Venous blood for peripheral blood mononuclear cells (PBMCs) isolation was collected from the femoral vein and were used for IFNγ ELISpot assay, as described previously. Blood volumes taken did not exceed 12 mL/kg within 30 days, and a maximum of 9 mL/kg at each individual bleeding time point.

All animal procedures were performed under anaesthesia either with ketamine (10–15 mg/kg IM) or Domitor (0.015 mg/kg IM).

Processing of peripheral blood (NHP study)

Peripheral blood mononuclear cells (PBMCs) were harvested and used for IFNγ ELISpot assay, as previously described. ⁴⁴ , ⁴⁵ Briefly, PBMCs were isolated from whole blood drawn into anticoagulant‐containing tubes (EDTA) by Ficoll density gradient centrifugation, followed by lysis of residual red blood cells (RBCs). Viable cell numbers were subsequently determined by Trypan Blue exclusion. Cells were kept on ice until analysis.

Peptides

Peptide pools used for ELISpot or ICS were > 90% pure and obtained from JPT Peptide Technology. Each peptide pool consisted of 15‐mers overlapping by 11 amino acids covering the wild‐type amino acid sequence of E2, E6 and E7 of HPV16 and HPV18. In total, six peptide pools were generated, with one pool representing each individual antigen, and these were used across both the mouse and NHP studies as previously described. ¹⁸

Mouse and non‐human primate IFNγ ELISpot

Mouse IFNγ ELISpot assay was performed on mouse splenocytes as previously described, ¹⁸ with the only exception that a mouse IFNγ ELISpot‐plus kit (Mabtech, Cincinnati, OH) was used. Freshly isolated splenocytes were incubated with a pool of 15‐mer peptides (overlapping by 11 amino acids) spanning the entire sequence of the HPV16 or HPV18 of E2, E6 or E7 proteins. All samples were run in duplicates. Plates were counted on an AELVIS ELISpot reader, and the numbers of spot‐forming units (SFU) per 10⁶ cells were calculated. Background was defined as the 95th percentile of values from the 0.4% DMSO in medium.

Antigen‐specific, IFNγ‐secreting T cells were determined in isolated PBMCs isolated from rhesus macaques using an ELISpot kit specific for monkey IFNγ (Monkey IFNγ ELISpot^PRO; MabTech, Cincinnati, OH) according to manufacturer instructions. PBMCs were seeded at 2 to 5 × 10⁵ cells/well and stimulated with peptide pools reconstituted in dimethylsulfoxide (DMSO), consisting of 15‐mers overlapping by 11 amino acids at a final concentration of 2 μg/mL for 18–20 h at 37°C and 5% CO₂, in a final volume of 200 μL.

Rhesus macaques PBMC samples were analysed in triplicate. Mean SFU per 10⁶ cells were calculated from the replicate measurements, followed by individual background subtraction of the mean medium control values from the mean peptide‐stimulated values. Based on historical data, the background/threshold was empirically set at 50 SFU/10⁶ PBMC. Values below the threshold of 50 SFU/10⁶ PBMC were set at half that threshold (25 SFU/10⁶ PBMC) for the purpose of graphical representation. Reported are background‐subtracted responses induced by each of the different peptide pools.

For calculation of the breath, a response to any of the E2, E6 or E7 of HPV16 or HPV18 antigen (above 50 SFU/10⁶ PBMC) was assigned the value of 1, and the values were summed up to determine the breath.

Mouse intracellular cytokine staining

Mouse ICS was performed on splenocytes exactly as previously described. ¹⁸ Briefly, splenocytes were subjected to red blood cell lysis and then incubated with the HPV16 or HPV18 peptide pools described above or with no peptide (background) before Golgiplug (containing Brefeldin A; BD Pharmingen) was added. The next day, cells were stained to detect live/dead cells, followed by cell surface staining, and thereafter, cells were subjected to intracellular staining. Samples were acquired with a BD FACSCanto™ II (Becton Dickson B.V.) flow cytometer, and results were analysed using the FlowJo software (Tree Star). Gate boundaries were set based on medium or/or mice vaccinated with non‐antigen encoding vectors.

First‐in‐human study design, participant cohort and study design

A randomised, double‐blind, placebo‐controlled Phase 1/2a study to evaluate safety, reactogenicity and immunogenicity of monovalent HPV16 and HPV18 Ad‐vectored vaccine component with an MVA‐BN‐vectored HPV16/18 vaccine component in otherwise healthy women with HPV16 or HPV18 infection of the cervix. The trial is registered with clinicaltrials.gov, identifier NCT03610581. Only individuals with persistent HPV16 or HPV18 infections were initially recruited, after an amendment one participant was enrolled that did not have a persistent infection according to the following definition. To be considered as persistent infection, participants had to have a 12‐month persistent cervical infection with HPV types 16 or 18, defined as at least two positive cervical HPV PCR tests with an interval of at least 11 months as follows: two positive HPV16 (or two positive HPV18) tests or, hrHPV positivity followed by HPV16 (or HPV18) positivity. Because of COVID‐19 pandemic restrictions and challenges with recruitment. The study was stopped after nine participants were enrolled. Notably, eight out of nine participants were positive or co‐infected with other hrHPV types than HVP16/18 at screening and/or baseline of which five participants (including in three active treated participants) remained persistently positive. The ages of the nine Participants were 30–59 years. Body mass index ranged from 20.5 to 40.3 kg/m². Eight participants received a complete vaccination scheme (Ad26‐MVA‐BN or placebo). Five participants were confirmed with a (persistent) HPV16 infection, two of which received Ad26.HPV16 (5 × 10¹⁰ vp) and MVA‐BN.HPV16/18 (2 × 10⁸ Inf.U), and one received Ad26.HPV16 (5 × 10¹⁰ vp) only and two received placebo. Four participants were confirmed with an HPV18 persistent HPV18 infection, of which two received Ad26.HPV18 (5 × 10¹⁰ vp) and MVA‐BN.HPV16/18 (2 × 10⁸ Inf.U), and two received placebo. Whole blood was collected and PBMCs were isolated at Day 1 (pre‐Ad26 vaccination), Day 57 (pre‐MVA‐BN vaccination), 78, 239 and 366 post‐primary Ad26 vaccination.

IFNγ ELISpot on human PBMC from clinical trial participants

Cryopreserved and thawed PBMCs were adapted with OpTmizer CTS medium (Life Technologies) for more than 6 h at 37 °C, 5% CO₂, and subsequently PBMCs (2 × 10⁵ cells per well) were stimulated with 2 μg/mL of E2 or E6/E7 peptides of HPV16 or HPV18 (15‐mer with 11 amino acids overlapping) for 48 h. Medium served as negative controls. After stimulation, spots indicating IFNγ‐γ‐secreting cells were developed, and the number of spots was analysed with an automated ImmunoSpot Analyzer (Cellular Technology Ltd.). The HPV‐specific responses were calculated by subtracting the mean number of spots in the medium control from the mean number of spots in experimental wells, which were expressed as SFUs per 10⁶ PBMCs. The samples were plated in duplicate. Antigen‐specific T‐cell responses were considered to be positive when the mean number of spots in the well with the antigen was above the lower limit of detection of the assay (211 SFU/10⁶ cells).

ICS on human PBMC from clinical trial participants

Cryopreserved and thawed PBMCs were resuspended in OpTimizer CTS, and rested overnight at 37°C, 5% CO₂. PBMCs were plated in duplicate and stimulated with E2 or a combination of E6/E7 peptides of HPV16 and HPV18 (15‐mer with 11 amino acids overlapping) at a concentration of 2 μg/mL, α‐CD3 (positive control, 10 μg/mL, UCHT1, BD Bioscience) or the medium alone (negative control) in the presence of 1 μg/mL of α‐CD28 (L293, BD Bioscience) and α‐CD49d (L25, BD Bioscience) for 13 h. Secretion inhibitors (monensin/brefeldin A, BD Bioscience) were added 90 min after initial stimulation. After stimulation, cells were washed with PBS for subsequent immunostaining and polychromatic flow cytometric analysis; gate boundaries were determined by FMO controls. Antibodies for staining cells were CD19‐APC‐R700 (Biolegend), CD4‐BUV395 (Biolegend), CD8‐BUV737 (Biolegend), CD3‐FITC (BD Bioscience), CD14‐AF700 (BD Bioscience), CD56‐AF700 (BD Bioscience), Live/dead‐APC‐Cy7 (Life technologies), TIM‐3‐PE‐Cy5 (BD Bioscience), PD‐1 BV650 (BD Bioscience), IFNγ‐γ‐BV421 (Biolegend), TNF‐α‐BV786 (Biolegend), IL‐2‐PE‐Cy7 (BD Horizon), CD107a‐BV510 (BD Bioscience), IL‐5‐PE (BD Bioscience), IL‐17‐BV605, Perforin‐BV711 (BD Bioscience) and GranzymeB‐PE‐CF594 (BD Bioscience). PBMCs flow cytometry was accomplished by Fortessa flow cytometer (BD Bioscience), and the data were analysed using the FlowJo software (Tree Star). Gating strategy is shown in Supplementary figure 6. Boolean gating was used to determine simultaneous cytokine production from CD4 and CD8 T cells. No positivity threshold was set for CD4 and CD8 T cells expressing IFNγ and/or IL‐2, and a threshold of 0.065% was set for CD8 T cells expressing CD107a.

Statistical analysis

The magnitude of the immune responses (i.e. ELISpot SFU and ICS data) were logarithmically transformed. Groups of animals were compared using Student's t‐tests with unequal variance. Within group comparisons (i.e. comparisons between timepoints) were analysed using paired t‐tests. The breadth of immune response, quantified as the number of antigens to which each animal mounted a response, was treated as an ordinal outcome and analysed using Wilcoxon rank sum tests when comparing between groups and Wilcoxon signed rank tests when comparing within groups. Tumor survival data were compared between treatment groups using a log‐rank test. A Bonferroni adjustment was used to correct for multiple testing. Differences were considered significant when P ≤ 0.05. Statistical analyses were performed with SAS version 9.4 and JMP version v10.

Author contributions

Selina Khan: Conceptualisation; data curation; formal analysis; project administration; supervision; validation; visualisation; writing – original draft, writing – reviewing and editing. Satish Boedhoe and Miranda Baert: Data curation; formal analysis; investigation; validation; visualisation. Jan Serroyen: Formal analysis; writing – reviewing and editing. Ariane Volkmann, Markus Kalla and Katrin Weidner: MVA‐BN based vaccine conceptualisation; writing – reviewing and editing. Thierry Verbinnen: Data curation; formal analysis; validation; visualisation; writing – reviewing and editing. Mathieu Le Gars: Data curation; formal analysis; investigation; validation; writing – reviewing and editing. Jerry Sadoff: Conceptualisation; supervision. Michal Sarnecki: Investigation; supervision. Jort Vellinga and Jerome Custers: Conceptualisation; project administration; supervision. Gert Scheper: Conceptualisation; project administration; writing – reviewing and editing. Hanneke Schuitemaker: Conceptualisation; supervision. Roland Zahn: Conceptualisation; supervision; writing – reviewing and editing. All authors have read and agreed to the published version of the manuscript.

Conflict of interest

The authors declare the following competing interests: Selina Khan, Satish Boedhoe, Miranda Baert, Jan Serroyen, Mathieu Le Gars, Jerry Sadoff, Michal Sarnecki, Jort Vellinga, Jerome Custers, Gert Scheper, Hanneke Schuitemaker and Roland Zahn are or were employees of Johnson and Johnson while engaged in the research project. Thierry Verbinnen is an employee of Janssen Pharmaceutica NV. Selina Khan, Satish Boedhoe, Miranda Baerts, Jan Serroyen, Mathieu Le Gars, Jerry Sadoff, Jort Vellinga, Thierry Verbinnen, Jerome Custers, Gert Scheper, Hanneke Schuitemaker and Roland Zahn held or still hold stock in Johnson & Johnson. Jerome Custers, Gert Scheper and Selina Khan are inventors on patent applications related to this work (WO2016/071306, PCT EP2016/069618), which are held by Janssen Pharmaceutical Companies. Ariane Volkmann, Markus Kalla and Katrin Weidner are or were employees of Bavarian Nordic GmbH while engaged in the research project.

Ethics statement

The mouse studies were performed with approval of the appropriate Institutional Animal Care and Use Committees and according to the Dutch law and Guidelines on the Protection of Experimental Animals by the Council of the European Committee (EU Dir. 86/609) after approval by the Dier Experimenten Commissie of Crucell (permit number: 21300).

The non‐human primate study was conducted in facilities assured by the National Institutes of Health Office of Animal Welfare and accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care (AALAC). The study was performed at Advanced BioScience Laboratories Inc. (Rockville, MD, USA); with animals purchased by BIOQUAL Inc., Rockville, MD, USA; BIOQUAL.

All animal research protocols were approved by the Institutional Animal Care and Use Committees at each centre, and the studies were conducted in compliance with the Animal Welfare Act, Public Health Service Policy on Humane Care and Use of Laboratory Animals, and other federal statutes and regulations relating to animals and experiments involving animals. Import and export permits for vectors and NHP biospecimens were obtained in compliance with US federal regulations and in accordance with the Convention on International Trade in Endangered Species of Wild Fauna and Flora, as overseen by the US Fish and Wildlife Service.

The human clinical trial was conducted in accordance with the ethical principles that have their origin in the Declaration of Helsinki and that are consistent with Good Clinical Practices and applicable regulatory requirements. The study protocol and amendments were reviewed by an Independent Ethics Committee or Institutional Review Board. Subjects or their legally acceptable representatives provided their written consent to participate in the study after having been informed about the nature and purpose of the study, participation/termination conditions and risks and benefits of treatment. The trial is registered with clinicaltrials.gov, identifier NCT03610581.

Supporting information

Supplementary figure 1

Supplementary figure 2

Supplementary figure 3

Supplementary figure 4

Supplementary figure 5

Supplementary figure 6

Supplementary table 1

Supplementary table 2

Data availability statement

The data that support the findings of our study are available from the corresponding author upon reasonable request.