Final Report Summary - IMPROVE (Improving Prostate Cancer Outcome with Vectored Vaccines)
Executive Summary:
IMPROVE was a project funded by the European Commission through its Seventh Framework Programme (FP7). It brought together two academic institutions and five industrial partners from across Europe with the common objective of developing a novel prostate cancer immunotherapy.
Cancer immunotherapy has recently emerged as a viable and attractive treatment strategy. The only licensed therapeutic prostate cancer vaccine, Sipuleucel-T, provided a modest survival benefit of 4 months and showed low immunogenicity. In this project, we have clinically evaluated the most potent vaccine technology available for inducing cellular immunity to target prostate cancer. This highly immunogenic vaccination platform deploys the replication-deficient simian adenovirus, ChAdOx1, as a priming agent and attenuated poxvirus, MVA, as a boosting vaccine. Both viral vectors encode the tumour associated antigen 5T4 overexpressed in prostate tumours.
The output of the IMPROVE project was remarkable, and the most significant achievements are highlighted below. For the first time, a heterologous vaccination platform based on ChAdOx1 and MVA vectors was deployed clinically in a cancer setting (a phase I clinical trial – VANCE, NCT02390063). As a result, T cell immune responses against the tumour-specific antigen, 5T4, encoded by the vaccine were detected in the majority of vaccinated patients ex vivo, without prior expansion in culture. This is a prominent result given that 5T4 is a self-antigen against which the immune system is tolerized. A detailed immunomonitoring programme was undertaken that allowed for identification of significant number of CD8- and CD4- restricted T cell epitopes with the vaccine antigen and their HLA restriction elements which will potentially inform the design of a next generation vaccine. Importantly, T cell immune responses were induced not only in the blood but also the immune cell infiltration was detected in the target organ, the prostate, thus challenging previously held belief that prostate cancer is a “cold” tumour which is not amenable to immunotherapies. The objectives of this first-in human clinical study were fully achieved, having demonstrated that the vaccine was safe and induced vaccine antigen specific immune responses in the majority of patients.
One of the major impediments to tumour protective vaccine efficacy is an up-regulation of inhibitory receptors on effector T cells and tumour cells induced by immunisation. In the second clinical trial within this programme, we aimed to evaluate the clinical efficacy of the vaccine in combination with an agent that blocks inhibitory receptor, and thus disrupts the pathways that attenuate the vaccine induced immune response. This phase I/II clinical trial (ADVANCE, NCT03815942) recruited 23 patients with metastatic disease with progression after hormone therapy, where the vaccine is combined with an anti PD-1 checkpoint inhibitor Nivolumab. The preliminary data from this trial demonstrated that the vaccine in combination with nivolumab is safe in this patients cohort. The clinical efficacy as measured by PSA reduction of 50% or greater compared to baseline have been achieved in 26% of patients which is significantly higher than the efficacy of nivolumab as a single agent in the reported KEYNOTE-199 study. A reduction in the number of circulating tumour cells was also observed in some of the PSA responders.
In parallel, we have undertaken a preclinical programme on comparative immunogenicity and efficacy assessment of a range of new and old cancer antigens in mouse models of prostate cancer to identify the most effective antigens for clinical development as an immunotherapeutic, with an output of the ChAdOx1 and MVA multi-antigen vectors encoding three prostate-specific antigens, PSA, STEAP1 and 5T4, being available for early clinical development.
A further part of the IMPROVE project was communications with the public and dissemination of general knowledge about prostate cancer vaccines. IMPROVE gained visibility through a number of presentations at the prestigious international conferences and publications in scientific journals. After funding of the project the website will be still active and summaries of the results of the publications will be disseminated on the website.
This five year programme has combined the most potent vaccine technology available for inducing cellular immunity in humans with leading cancer antigens to target prostate cancer, for which there is substantial evidence that vaccine immunotherapy is feasible.
Project Context and Objectives:
Summary description of the project context and the main objectives.
Prostate cancer (PCa) has been under investigation as a target for antigen-specific immunotherapies in metastatic disease settings. Neither of the two clinically most advanced PCa vaccines, Sipuleucel-T and ProstVac, induced strong T cell immunity and their clinical efficacy was modest. Prostate cancer has not been considered a good target for checkpoint inhibitor therapies, given its low mutational load and lack of pre-existing intratumoural immune cell infiltration, until recent data from KEYNOTE-199 study (Antonarakis, J. Clin Oncol. 2019) demonstrated a 9% response rate to pembrolizumab monotherapy.
The major clinical development goal of IMPROVE was the evaluation of a new very potent vector-based immunisation strategy in the context of a design that allows the measurement of multiple sensitive early endpoints that may represent vaccine efficacy. The immunisation approach is based on two viral vectors that, in a heterologous prime-boost immunisation regime, have yielded the highest measured CD8 T cell responses in humans across a range of prophylactic and therapeutic vaccine indications, following on from pre-clinical studies that demonstrated exceptional potency.
IMPROVE has developed a novel vaccination platform based on two replication-deficient viruses, chimpanzee adenovirus and MVA, targeting an oncofetal self-antigen 5T4 and evaluated this vaccine alone and in combination with anti-PD-1 in mouse tumour models.
Next, we have tested this vaccine in a first-in-human trial, VANCE (NCT02390063), in early stage prostate cancer patients. The patients, either newly diagnosed with early stage PCa and scheduled for radical prostatectomy or patients with stable disease on active surveillance protocol, were randomised to a “standard” immunisation regimen to receive 3 vaccinations four weeks apart, or to an “accelerated” immunisation protocol to receive 2 vaccinations at one week interval. Study primary endpoints were vaccine safety and immunogenicity.
Extensive studies of immune down-regulatory mechanisms in cancer have now identified a range of mechanisms that attenuate the immunogenicity of candidate cancer vaccines and also reduce the efficacy of the induced immune responses in the target tumour. Among those, an up-regulation of inhibitory receptors on effector T cells and tumour cells induced by immunisation is one of the major impediments to tumour protective vaccine efficacy. The second part of the clinical programme was the evaluation of this immunisation strategy in combination with a checkpoint inhibitor therapy in a phase II trial, ADVANCE (NCT03815942), to test the vaccine safety and efficacy in combination with PD-1 blockade in intermediate risk diseases and metastatic prostate cancer. This trial is designed to ask if early evidence of efficacy in combination with immune checkpoint blocker can be detected using sensitive established biochemical, histological and radiological indicators of vaccine efficacy as well as novel methods of genomic analyses in advanced metastatic prostate cancer. The patients will be followed for one year with regular PSA, circulating tumour cells (CTC) and circulating tumour DNA (ctDNA) measurements to determine the duration of biochemical relapse-free survival and to monitor tumour burden and correlate the early measures of efficacy with this established surrogate measure of relapse free survival and also clinical measures. The preliminary data from ADVANCE trial demonstrated that the vaccine in combination with nivolumab has good safety profile. The clinical efficacy as measured by PSA reduction of 50% or greater compared to baseline have been achieved in 26% of patients which is significantly higher than the efficacy of nivolumab as a single agent in the reported KEYNOTE-199 study. A reduction in the number of circulating tumour cells and reduction in tumour lesions were also observed in some of the PSA responders.
The goal of a preclinical component of IMPROVE was development of a novel multi-antigen prostate cancer vaccine. Extensive antigen discovery efforts have led to the identification of several older and new candidate antigens for prostate cancer vaccines, but these have never been compared head to head. Therefore, in parallel with preclinical and clinical programme on the development of ChAdOx1 and MVA vaccines expressing the 5T4 antigen, we have undertaken a preclinical development of a second generation vaccine vectors encoding 3 prostate-associated antigens – 5T4, PSA and STEAP1.
The main objectives of the IMPROVE project are listed below.
• Preclinical evaluation of ChAdOx1-MVA with 5T4 antigen
• GMP manufacturing of viral vectored vaccines for phase I and II clinical trials
• Assessment of safety and immunogenicity of the novel prostate cancer immunisation approach in intermediate risk i prostate cancer in a phase I clinical trial
• Assessment of safety and efficacy of the prostate cancer vaccine in combination with checkpoint inhibitors in intermediate risk and advanced metastatic prostate cancer in a phase I/II clinical trial
• Detailed characterization of the vaccine-induced immune responses in terms of breadth and functionality
• Analysis of correlates of immunogenicity and MCHC pre-treatment biomarker
• Comparative preclinical antigen evaluation in terms of immunogenicity and efficacy for a next generation multi-antigen vaccine.
Dissemination and exploitation
In addition to this technical work there was a programme of dissemination activity:
• Dissemination of project results to clinicians, industry, the public and investment communities
• Conference presentations and peer-reviewed journal publications
• Newsletters, public website and other dissemination material
The project has achieved success across all these target areas.
This report will present the project results and provide contact information for those desiring further information.
Project Results:
Description of main S & T results/foregrounds.
The project has achieved most of its objectives and technical goals. The main objectives of the project are as follows:
• To evaluate a novel viral vectored prostate cancer vaccine in a phase I first-in-human study
• To evaluate a novel viral vectored prostate cancer vaccine in combination with checkpoint inhibitor drugs in the phase II clinical trial
• To evaluate preclinically a next generation prostate cancer vaccine encoding several cancer antigens
To accomplish its objectives, IMPROVE based the organization of the necessary management and RTD around 9 work-packages (WPs) as shown in Figure 1.
WP1 Pre-clinical evaluation of ChAdOx1-MVA with 5T4 antigen
The objectives of WP1 comprised generation of the two recombinant viral vectors, ChAdOx1 and MVA, expressing the tumour-associated antigen 5T4 and their evaluation in a mouse model prior to testing these viruses as a prostate cancer immunotherapy in a phase I clinical trial. Recombinant viral vectors expressing a murine homologue of the tumour-associated antigen 5T4, ChAdOx1.m5T4 and MVA.m5T4 were designed, manufactured and tested in a proof-of-concept study to evaluate their immunogenicity and efficacy in a mouse model. A heterologous prime boost vaccination regimen based on ChAdOx1.m5T4 priming immunisation and MVA.m5T4 boost was demonstrated to break immunological tolerance to this antigen in that both cellular and humoral 5T4 specific immune responses were detected following immunisation. A modest tumour protective efficacy resulting in delayed tumour growth was also observed in a transplantable tumour model following vaccination.
Further, the clinical batch of the ChAdOx1.5T4 vaccine was tested in mice and the treatment was well tolerated and was not associated with any adverse effects when administered at the high dose of 5x10^9 vp (virus particles).
WP2 GMP Vector Manufacture
The objective of WP2 was to manufacture a clinical batch of GMP grade ChAdOx1.5T4 and to release to trial this batch of the vaccine. GMP compatible starting material, process development and GMP manufacture of a batch of material suitable for stability study and toxicology has been completed first, followed GMP manufacture of ChAdOx1.5T4 clinical batch resulted in 300 vials available for clinical use in planned phase I and phase II clinical trials. Post QC testing (internal and external) and QA batch review, ChAdOx1.5T4 was QP certified as having met the requirements of EU GMP.
WP3: Phase I clinical trial (acronym - VANCE)
The objectives of WP3 was to perform a first-in-man clinical trial of a novel prostate cancer (PCa) vaccine comprised of the ChAdOx.5T4 and MVA.5T4 viral vectors delivered in a heterologous prime-boost regimen to low- and intermediate- risk prostate cancer patients in order to assess the safety and immunogenicity of the vaccine. 40 patients, either newly diagnosed with early stage PCa and scheduled for radical prostatectomy or patients with stable disease on active surveillance protocol, were randomised to a “standard” immunisation regimen to receive 3 vaccinations four weeks apart, or to an “accelerated” immunisation protocol to receive 2 vaccinations at one week interval. Half of the patients were also treated with low dose cyclophosphamide prior to each immunisation to deplete regulatory T cells. Study primary endpoints were vaccine safety and immunogenicity. Secondary endpoints included immune infiltration into the prostate and PSA level change secondary to treatment. As exploratory endpoints, phenotype and functionality of antigen-specific T cells and breadth of induced T cell responses were assessed. The study main findings are as follows.
39 patients completed the study and were eligible for analysis. The vaccine had an excellent safety profile, with the majority of vaccine-related adverse events graded as mild. No additional safety findings were apparent following vaccination in combination with low dose cyclophosphamide. There have been two serious adverse events (SAE) which were deemed unlikely to be related to the investigational medicinal products employed in this study. Vaccination-induced 5T4-specific T cell responses were measured in blood by ex vivo IFNγ ELISPOT and were detected in the majority of patients. Cyclophosphamide pre-conditioning has not improved frequency or magnitude of the T cell responses. Flow cytometry analysis demonstrated the presence of both CD8 and CD4 poly-functional 5T4-specific T cells in the circulation. 5T4-reactive tumour-infiltrating lymphocytes (TILs) were isolated from post-treatment prostate tissue. Some of the patients had a transient PSA level increase 2-8 weeks following vaccination possibly indicating an inflammatory response in the target organ.
Analysis of ex vivo cellular immune responses to 5T4.
The main readout of the vaccine immunogenicity was the induction of cellular and humoral 5T4-specific immune responses. To this end, blood samples from each patient have been collected at baseline and at each subsequent clinic visit. Peripheral blood mononuclear cells (PBMCs) and plasma have been separated out and used to assess the 5T4-specific T cell responses by an ex vivo IFNγ ELISPOT assay and the 5T4-specific antibody titres by semi-quantitative ELISA. In the ELISPOT assay, PBMCs have been exposed to overlapping 15-mer peptides covering the whole 5T4 protein for 18-20 hours, individual cells secreting IFNγ have been enumerated and presented as frequencies per 10^6 PBMCs. An example of the response magnitude and kinetics is shown in Figure 2, demonstrating the cellular immune responses against 5T4 peptide pools in 2 surgical patients randomised either to the standard or accelerated vaccination regimen and one active surveillance patient (accelerated regimen). Overall, detected cellular ex vivo immune responses have been encouraging taking into account that 5T4 is a self-antigen against which an immunological tolerance is likely to exist, with the majority of patients mounting relatively high T cell response to 5T4 (Figure 3).
Analysis of humoral immune responses to 5T4.
In order to evaluate the capacity of the vaccine to induce an antibody response against 5T4 a validated semi-quantitative ELISA assay has been performed on patient sera before vaccination and at different time points post-vaccination. For each patient, 5T4-specific antibody levels were assessed. The 5T4 sero-conversion rates by treatment group are displayed in Table 1. Patients who received at least 2 injections of MVA.5T4 showed a 100% sero-conversion rate for 5T4 antibodies. In contrast, patients who only received 1 injection of MVA.5T4 as part of an accelerated vaccination schedule had a lower sero-conversion rate of <50%. It is impossible to conclude from these results whether the reduced number of MVA.5T4 injections or the accelerated vaccination regimen is the cause of the reduced antibody response rate. However, based upon the kinetics of antibody responses in the homologous vaccination regimen, it is likely that the reduced number of MVA.5T4 vaccinations is the cause of the reduced antibody responsiveness. To conclude, robust 5T4-specific antibody responses are induced in patients enrolled into the standard vaccination regimen (but only weak responses are detected in patients enrolled into the accelerated vaccination groups
Study secondary endpoints.
• To address the study secondary outcomes, the patients’ PBMC samples collected at different timepoints have been cultured for 12-14 days either in medium alone or in the presence of the total 5T4 peptide pool to expand the relatively infrequent vaccine antigen-specific T cells for further analysis by cytokine flow cytometry, which allows a more detailed characterisation of vaccine-induced immune responses but has lower sensitivity and thus requires a higher number of antigen-specific cells for the assay. Following in vitro stimulation, cell cultures were stained with fluorochrome-labelled antibodies against CD4 and CD8 T cell surface markers and intra-cellular cytokines characteristic for effector T cells (i.e. IFNγ and TNFalfa). Representative data from 3 patients are shown in Figure 4. Typically, after in vitro 5T4 antigen-specific stimulation and expansion, between 0.2% and 2% of cultured cells are specific for the vaccine antigen as measured by secretion of cytokines.
• One of the study secondary endpoints is the detection of antigen-specific immune cells in the prostate, as this is where they are expected to exert their effector function and destroy tumour cells expressing the 5T4 antigen. Fresh tissue from the resection specimens or biopsies was processed to expand the immune cells resident in the prostate prior to analysis of their 5T4 specificity. In the majority of cases, high numbers of TILs were expanded from fresh tissue (with a range of 6-80 million immune cells per tissue core/biopsy). Antigen-specific tumour-infiltration lymphocytes could be detected in ~70% of the patients. Flow cytometric analysis of the TILs from one of the patients is shown in Figure 5.
• Immunohistochemical (IHC) analysis of immune cell tumour infiltration following vaccination was performed on fixed tissue. Tissue blocks from each patient were formalin-fixed and paraffin-embedded (FFPE) for IHC analysis of immune cells density in post-vaccination tissue compared to available archival tissue from pre-study diagnostic biopsy. Available paired pre- and post-treatment FFPE tissue sections have been stained for a panel of pre-selected immune cell markers including CD3 and CD8. Batch analysis was performed to quantify the density of immune cells by digital image analysis in order to assess whether the vaccination course leads to trafficking of immune cells to the prostate. The density of cells positive for these markers before and after study treatment was assessed by digital image analysis. In summary, for 12 cases diagnostic pre-treatment biopsies were compared with post-treatment surgical specimens (samples from patients in the surgical arms), and for 13 cases diagnostic pre-treatment biopsies were compared with post-treatment on-study biopsies (patients in active surveillance arms). Cell densities were calculated to the number of marker positive cells per unit area (mm2). The comparison of CD3 and CD8 cell densities in surgical patients revealed a decrease when comparing biopsies (pre-treatment) to resections (post-treatment). Interestingly, when the spatial distribution of marker positive cells was analysed the Ripley’s K index, high values for K-Ripley were observed in resections compared to pre-treatment biopsies. This finding can be explained by potential re-distribution of the immune cells following the study treatment, i.e. the immune cells in prostate tissue normally diffusely distributed in the tissue become activated due to vaccination and as a result traffic to the areas where tumour lesions are to mediate their function of killing tumour cells. This is an interesting observation that needs to be further investigated by studying matched biopsies and resection tissue of untreated control patients.
• PSA reduction in peripheral blood of advanced-stage prostate cancer patients secondary to interventions has been widely used as a surrogate marker of treatment efficacy before In our study, PSA level change was also included as one of the study endpoints. PSA levels in the blood of cancer patients at baseline (before treatment) was compared to PSA levels post vaccination course. Unexpectedly, instead of PSA level drop we have observed a transient PSA increase in some patients following vaccination. If baseline PSA values in all patients in the active surveillance arm are compared to their respective highest PSA values over a period of one year post vaccination, the difference between a baseline and the peak value is not significant (Figure 6). However, there is clearly a trend that can imply that the vaccination induced T cells have trafficked to the prostate and acted upon 5T4-expressing prostate cells leading to the leakage of PSA into the circulation.
WP4: Phase I/II clinical trial (acronym - ADVANCE)
The objective of WP4 was to undertake a phase II safety and efficacy clinical trial with the vectors encoding 5T4 with the trial design informed by results from the phase I clinical trial. In the phase II study, ChAdOx1-MVA 5T4 vaccine was tested in combination with the anti-PD-1 checkpoint blocker Nivolumab in low- or intermediate-risk localized or locally advanced prostate cancer and metastatic castration resistant prostate cancer (mCRPC). The study synopsis is shown in Table 2. Recruitment to the study started in February 2019 and by October 2019 23 out of 24 patients in the metastatic cohort have been enrolled. As of January 2020, all recruited patients either completed the study treatment and were withdrawn due to disease progression. The full data set for this patient cohort will be analysed by Q4 2020. The recruitment to the early stage prostate cancer cohort is ongoing.
• Safety. To date, administrations of ChAdOx1.5T4 and MVA.5T4 vaccines in combination with nivolumab were found to be safe and well tolerated, in agreement with our previous studies with these vectors.
• PSA response rate. Serum PSA levels were measured through the study a t each clinic visit. 5 of 19 patients demonstrated 50% or greater PSA decrease from baseline (Fig.7). This response rate is significantly higher than the response rate reported in the KEYNOTE-199 study (Antonarakis et al., 2019) of anti-PD-1 antibody alone in advanced metastatic prostate cancer. A maximal decrease in PSA concentration compared to baseline in each of 5 responders is shown in Fig.8.
• CTC enumeration. The quantification of circulating tumour cells (CTCs), which is one of the study exploratory endpoints, has been performed. Encouragingly, in some of the patients a decrease in serum PSA correlated with decrease in a number of CTCs (Fig.9).
• T cell immune responses. The primary immunogenicity measure in the study is the IFN-γ ELISpot response of PBMCs stimulated with overlapping peptide pools covering the entire length of 5T4 proteins. Responses were measured before vaccination and at different time points post-vaccination. Out of 23 patients who completed the vaccination course, 2 patients mounted 5T4-specific T cell response that was detected at several timepoints during the study (Fig. 10). The immune response rate to the vaccine is lower that the one observed in the VANCE study, but this is not surprising given the patient baseline characteristics such advanced metastatic disease and multiple potentially immunosuppressive treatment modalities prior to enrolment to the study.
WP5 Immunogenicity studies
The aim of WP5 was to carry out a detailed analysis of the immune responses both in the blood generated by the 5T4-targeting prostate cancer vaccine deployed in the phase I and II clinical trials. Dissection of CD8 and CD4 T cell immune responses was performed by in vitro expansion and flow cytometry. The main achievements are as follows:
Test and distribution of protocol for tumour-infiltration lymphocytes (TILs) expansion from prostate cancer biopsies
• Phenotype of TILs derived from Prostatic carcinoma to a non-cancer control was completed.
• The ex-vivo TILs phenotype was investigated and the phenotypic changes evaluated after in vitro expansion.
• We observed high level of similarity between expanded TILs from PC and BPH samples. T cells are the predominant population in both pre-REP and REP expanded condition and NK cells are completely lost only after REP expansion.
• In BPH samples the expansion with pre-REP and REP protocols generated a predominant T cell population that progressively upregulates activation/exhaustion and cytotoxic markers in comparison to TILs directly analyzed from digested disuse ex-vivo.
Development of a flow cytometry panel to monitor regulatory T (T-reg) cells in peripheral blood
• T-reg flow cytometry panel with 14 markers was developed
Mapping of T cell responses in vaccinated patents
• In vitro culture protocol and mapping of CD4 and CD8 T cell responses in 25 cancer patients vaccinated for 5T4 protein.
• Several CD4 and CD8 T cell epitopes were identified. Interestingly, many of those were never published previously.
Development and harmonization of protocol for testing immunogenicity and T cell phenotype longitudinally in prostate cancer patients enrolled into the study.
• Ex-vivo and In-vitro culture, stimulation and intracellular staining protocols were developed and harmonized between UNIL and OX (analysis performed in 3 healthy donors).
• Harmonization resulted successful. Test of the developed flow cytometry panel on PBMC of 3 healthy donors gave comparable results.
An extended flow cytometry panel was validated for use in the phase II clinical trial for detailed phenotypic and functional characteristics of induced T cell responses (Table 3).
WP6 Correlates of Immunity and efficacy
The objective of WP6 is to build on the phase I and phase I/II clinical trials. In phase I trial, VANCE, Blood samples recovered from patients recruited to the trial have been analysed for antigen-specific responses and the results used to investigate whether there are any potential associations between immune responses and clinical benefit. Analysis of correlates of immunogenicity and efficacy is dependent on the provision of data from WP5 and suitable markers of clinical efficacy. Given that the patients treated in the phase I clinical trial were at a relatively early-stage in their disease course, no recognised markers of clinical benefit (e.g. response rate, survival etc) were available within the timeframe of this work package. As such, the key analytic testing looked to identify predictors of immunogenicity.
An analysis of potential correlates with immunogenicity was undertaken using antibody response data. Previously, a pre-treatment biomarker was identified which predicted 5T4 specific immune responses induced in patients following vaccination with MVA-5T4 (TroVax; Harrop et al., Cancer Immunology Immunotherapy 2011). The same pre-treatment biomarker (consisting of baseline levels of haemoglobin, haematocrit and 5T4 antibody level) was used to determine if this biomarker was associated with the fold-increase in 5T4 or MVA antibodies induced following vaccination with ChAdOX.5T4 and MVA-5T4. These exploratory analyses suggested a significant relationship between the pre-treatment biomarker and the 5T4, but not MVA, antibody responses induced post vaccination. An assessment of correlation coefficient between the pre-treatment biomarker and 5T4- and MVA- specific antibody responses at weeks 1, 2, 4 and 8 are shown in Table 4. As can be seen, there is no significant correlation between the pre-treatment biomarker and antibody responses detected at week 1; this is not surprising as it normally takes >1 vaccination before antibody responses against the target antigen are detected. However, at weeks 2, 4 and 8 post-vaccination there are strong trends (weeks 2 and 8) or significant correlations (week 4) between the pre-treatment biomarker and the 5T4-specific antibody response, but not the MVA-specific antibody response. These data replicate what have been reported previously i.e. significant correlations between pre-treatment biomarker and antibody responses against the target antigen (5T4), but not against the viral vector (MVA).
These data are of potential significance because 8 clinical trials using MVA-5T4 have demonstrated a link between the magnitude of the 5T4-specific immune response and clinical benefit (tumour shrinkage or overall survival). The ability to select the patients who are most likely to benefit from this class of therapy prior to treating them would be a huge advantage. Further confirmatory prospective studies would be required to develop a test which predicted treatment benefit.
For correlation between immune response and clinical efficacy in the phase I/II study, ADVANCE, blood samples from 23 recruited patients were processed to isolated PBMCs for IFN-γ ELISPOT assays and to measure the number of antigen-specific T cells induced by vaccination. Patients’ blood and serum samples were also collected for CTC enumeration and PSA concentration.
Out of 23 patients, 2 patients mounted vaccine-specific T cell immune response and 5 patients had a PSA decrease from baseline of 50% or more and 5 patients had a CTC decrease from baseline. One of 2 patients with immunological response had also a decrease in the number of CTC but no PSA reduction. So there clearly was no detectable correlation between our primary immunogenicity read-out (ELISpot) and the available and most information efficacy read-out (PSA reduction). However, this correlation analysis is very underpowered because of the small number of responders in the ELISpot assay. This low frequency contrast with the findings in our phase I VANCE trial where the same vaccines produced a >65% response rate. The difference may reflect a difference in the general immune competence of early and late-stage prostate cancer patients. A correlation analysis was also performed on the patients who had either PSA or CTC decrease or both and although there is as yet no significant difference on small numbers we will extend this analysis once CTC results are available from more subjects (Fig. 11).
WP7 Pre-clinical antigen evaluation – immunogenicity
WP7 comprises the characterisation of novel tumour-associated antigens and the assessment of their immunogenicity following expression from the viral vectors. The original focus was on the characterization of three antigens, namely CRISP3, CNPY2 and DPY19L3 discovered by Externautics partner (EXT). Overall, the study reinforced the association of these antigens with prostate cancer and provided a first indication of the antigen expression in early PCa and pre-cancerous stages. The expression of the three antigens was confirmed in different human prostate cell lines representing different PCa subtypes. Moreover, the murine counterparts of the three antigens were found expressed in murine prostate cell lines, including TRAMP cell lines, thus justifying the use of the TRAMP mouse model for efficacy studies. Finally, the three protein antigens showed an overall marginal expression in human normal tissues, thus allowing to hypothesize an acceptable toxicity profile of vaccine based on these antigens. Results specifically related to each antigen are summarized below:
• CNPY2 (Canopy FGF Signalling Regulator 2)
Human CNPY2 protein is annotated in two isoforms of 182 and 84 amino acids. The 182 aa isoform is predicted to be associated with the endoplasmic reticulum while the 84 aa isoform is secreted or intracellular.
Expression in PCa: IHC analysis of 100 prostate cancer samples and 47 PIN cases showed that CNPY2 was detected in all tested samples with the highest expression in 17% of PCa and 17% of tested PIN samples.
Expression in normal human tissue: IHC analysis of 33 normal human tissues showed weak cytoplasmic staining only in stomach, colon and chorionic villi and negligible reactivity in the other 31 tissues. Western blot (WB) and flow cytometry (FACS) analyses of PBMC purified from healthy individuals showed moderate CNPY2 expression, which was confined to the intracellular compartments.
Endogenous expression in prostate cell lines: CNPY2 was clearly expressed in all tested human prostate cell lines (R22V1, VCaP, PC3, PNT1, LNCaP, DU145) and in the murine cell lines TRAMP-C1 and TRAMP-C2, as judged by Q-RT-PCR and Western blot using anti-CNPY2 antibody. Western blot analysis allowed detection of a protein band of 20 kDa, compatible with the 182 amino acid isoform. CNPY2 was also well expressed in murine prostate TRAMP cell lines. FACS analysis showed that the protein is confined to the intracellular compartment.
• DPY19L3 (probable C-mannosyltransferase)
DPY19L3 is the least characterized protein of the group. The protein has been annotated in at least 3 overlapping isoforms with predicted MW of 83, 63 and 12 kDa, respectively. Only one isoform is predicted for the murine DPY19L3 orthologues, showing 88% amino acid identity with the 716 amino acid human isoform and a predicted MW of 83 kDa.
Expression in PCa and PIN: IHC analysis of 100 PCa showed that the protein was detected in 44.4% of PCa, among which 6.1% showed the highest expression. Analysis of 47 PIN cases showed that the protein is detected in 74% of the samples, 13% of which showed the highest expression. Staining was generally cytoplasmic and in some samples a membranous staining was visible. Staining of normal prostate was negligible.
Expression in normal human tissue: IHC analysis of the 33 normal tissues by IHC did not show any significant staining. In addition, DPY19L3 expression was not detected in human PBMC or in neutrophils by Western blot or by FACS
Endogenous expression in prostate cell lines: DPY19L3 expression analysis in human prostate cell lines by Q-RT-PCR revealed the presence of the three annotated DPY19L3 transcripts. Western blot revealed two major protein bands of approximately 90 and 60 kDa, compatible with the two longer DPY19L3 isoforms. FACS analysis showed a moderate surface staining of VCaP and R22V1 cells suggesting that it is at least partially surface exposed in these cells. DPY19L3 expression in murine prostate TRAMP cell lines was confirmed at transcript level. In these cells, Western blot revealed a unique band of approximately 60 kDa that might result from proteolytic processing of the annotated 83kDa protein.
• CRISP3 (Cystein-rich secretory protein 3)
Human CRISP3 protein is annotated with at least two variants of 258 and 268 amino acids and is predicted to be secreted. In prostate carcinoma, it has been reported as associated with a subset of prostate carcinoma carrying the TMPRSS2-ERG fusion, pT3 disease stage. Two potential CRISP3 orthologues are annotated in the mouse, namely CRISP1 and CRISP3, sharing 54% amino acid identity with the human protein, whose role in cancer has not been previously described.
Expression in PCa and PIN: IHC analysis of 100 PCa cases showed that the protein was detected in 29% of PCa samples, among which 19% showed the highest expression. Analysis of 47 PIN cases showed that the protein is detected in 23% of PIN, 13% of which showed the highest expression. Staining was generally cytoplasmic, except for a few samples that also showed a membranous staining.
Expression in normal human tissue: CRISP3 was not detected in human PBMC, as assessed by WB and FACS. Instead, in agreement with published data, the protein was detected in the neutrophil population. IHC analysis of the 35 normal human tissues did not give any relevant expression of the protein.
Endogenous expression in prostate cell lines: among the different human cell lines currently tested CRISP3 was exclusively detected in the VCaP cell line, having the TMPRSS2-ERG fusion. In these cells WB analysis with the anti-CRISP3 antibody revealed the presence of a band of expected size (approximately 30 kDa). The antibody also bound to the surface of VCaP cells in FACS staining assay, indicating that CRISP3, besides being secreted, is at least partially exposed on the cell surface. Concerning CRISP3 expression in the murine TRAMP cells, transcription profile analysis by Q-RT-PCR of the tumours excised from TRAMP mice showed that mCRISP1 was detected in two of the three tested biopsies whereas mCRISP3 cDNA was not detected, indicating that mCRISP1 could be the protein form expressed in murine prostate cancer.
Construction of recombinant viral vectors ChAdOx1 and MVA expressing novel prostate-associated antigens, CRISP1, CNPY2, DPY19L3, and comparative immunogenicity of these novel antigens and well defined prostate-associated antigens 5T4 and STEAP1 that also represent potential antigenic targets for a prostate cancer vaccine.
Recombinant viral vectors ChAdOx1 and MVA, expressing the novel prostate associated antigens CNPY2, DPY19L3 and CRISP3 were designed and generated for immunogenicity testing in mice.
These vectors delivered in a heterologous ChAdOx1 prime – MVA boost vaccination regime have been tested for immunogenicity in C57Bl/6 mouse strain. The cellular immune responses following immunisation have been assessed by an ex vivo IFNg ELISPOT assay against pools of 15-mer peptides overlapping by 10 amino acids spanning the full length of the proteins encoded by the vector transgenes. As a result, there have been no detectable T cell reactivity against CRISP1 and CNPY2 antigens, however, there was potentially a weak response induced against the DPY19L3. These results are in striking contrast to the immunogenicity of another murine prostate-specific antigen, STEAP1, induced by ChAdOx1-MVA immunisation (Cappuccini et al 2016). Although disappointing, these results were not unexpected as the central and/or peripheral immunological tolerance usually prevents an induction of immune responses against self-antigens. We have demonstrated that exceptionally strong T cell reactivity against STEAP1 is likely to be due to the lack of its expression in the thymus, although the mRNA transcripts corresponding to the m5T4, CRISP1, CNPY2 and DPY19L13 are detectable in the murine thymus by reverse transcription PCR.
The novel prostate-associated antigens identified by EXT have not been taken further into efficacy testing due to paucity of the vaccine-induced immune responses against these antigens. Instead the Steering Committee decided to proceed with the pre-clinical testing of the panel of human antigens, PSA (prostate-specific antigen), STEAP1 (six transmembrane antigen of the prostate 1) and ERG (ETS transcription factor) as a polyvalent vaccine with the aim of taking them into the clinic in the future. To this end, firstly the single antigen vectors expressing these antigens in native form or fusions to the invariant chain have been made for immunogenicity testing. With further clinical development in mind, fusion constructs have been made with the transmembrane domain of the shark invariant chain for the following reasons. Firstly, the shark and human invariant chain share only 30% homology so the potential risk of autoimmunity will be greatly reduced. The 25 amino acid long transmembrane domain is the minimal sequence that is required to maintain the enhancer effect, so the risk of autoimmunity will be reduced even further. Strong immune responses against PSA and STEAP1 have been induced in a mouse model but tolerance to ERG antigen could not be broken (Figure 12) therefore this antigen has been excluded from the multi-antigen construct developed on WP8.
WP8 Pre-clinical antigen evaluation - efficacy
Objectives of WP8 comprise the efficacy evaluation of 5T4 encoding vaccines compared to the vectors expressing novel antigens, assessment of immune correlates of efficacy in mouse models and generation of new vectors encoding the most promising vaccine targets to be available for early stage clinical development.
According to the original project plan, the novel EXT antigens identified and described in WP7 should have been taken into further pre-clinical development in order to create the vectors available for early stage clinical development at the end of the project. As the antigens CRISP-1, CNPY-2 and DPY19L3 appeared to be of low immunogenicity or non-immunogenic at all, as an alternative, we have chosen the three well-defined prostate-associated antigens, prostate-specific antigen (PSA), six transmembrane epithelial antigen of the prostate – 1 (STEAP-1), and 5T4 antigen for further pre-clinical and clinical development and expressed a string of these antigens from the same vector thus creating multi-antigen ChAdOx1 and MVA vectors. Of note, the ERG antigen originally planned as a component of the multi-antigen vaccine, was replaced by 5T4 as the phase I clinical trial (WP3) results clearly indicated that both CD8 and CD4 T cell responses against 5T4 can be induced in prostate cancer patients.
PSA
Human PSA, a chymotrypsin-like serine protease, has a highly restricted tissue distribution and is expressed in the epithelial cells of the prostate gland, the same cell type from which most prostate tumours arise. Indeed, PSA is widely used as a serum marker for prostate cancer. Its expression is regulated by androgen, and itis present at extremely low levels in the circulation of adult men. Most prostate tumours, even the poorly differentiated ones, continue to express PSA. This cell type-specific expression of PSA makes it a potential target antigen for antitumor CTL. In fact, PSA is a target antigen of one of the two most clinically advanced prostate cancer vaccines - ProstVac.
STEAP1
The six-transmembrane epithelial antigen of prostate protein was identified in advanced prostate cancer. STEAP1 is highly expressed in human prostate cancer and is up-regulated in various cancers, including lung, bladder, colon, ovarian, and Ewing cancers. Immunohistochemical analysis of clinical specimens demonstrates significant STEAP1 expression at the cell–cell junctions of the secretory epithelium of prostate and prostate cancer cells. Little to no staining was detected at the plasma membranes of normal non-prostate human tissues, except for bladder tissue, which expressed low levels of STEAP1 at the cell membrane. Its cell-surface localization, together with its six-transmembrane topology, suggests STEAP1 may function as a channel/transporter protein in cell–cell junctions. Given its increased expression in cancer tissues, STEAP1 could be a promising target for T-cell based or antibody immunotherapy. In our previous experiments, we have expressed murine STEAP1 from ChAdOx1 and MVA viral vectors and investigated STEAP1-specific T cell response in mice vaccinated with these vaccines. Despite being a self-antigen, STEAP1 appeared to be highly immunogenic in mice, most likely because of absence of central tolerance to this antigen. STEAP1 has been evaluated in early stage prostate cancer clinical trials and delivered in the form of RNA as part of a multi-antigen vaccine with encouraging preliminary data.
Construction of ChAd and MVA viral vectors.
Both mono-cistronic ChAdOx1 and MVA vectors encoding individual antigens and multi-antigen vectors expressing all three antigens were constructed for immunogenicity testing in mice. Poly-cistronic vectors were constructed to express a string of three antigens from the least immunogenic to the most immunogenic antigen, i.e. with STEAP1 cDNA downstream of the promoter followed by 5T4 and PSA cDNA. Flexible polypeptide linkers composed of glycine and proline (GGG-P-GGG) were inserted between cDNAs to create a fusion protein.
Immunogenicity testing of mono- and poly- cistronic viral vectors in a mouse model.
To assess the immunogenicity of mono- and poly- cistronic viral vectors encoding human antigens PSA, STEAP1 and 5T4 in a mouse model, outbred mice (CD1 strain) were randomised into 5 groups. Mice were primed with ChAdOx1 virus encoding each individual antigen or with the vector expressing a string of antigens and were boosted with the MVA vectors accordingly. The induction of T cell immune responses against the vaccine transgenes was tested by ex vivo IFN-γ Elispot assay after priming and boosting immunisations. As evident from the Figure 13, T cell responses of relatively high magnitude are detectable in blood after a single immunisation with a polcistronic ChAdOx1 vector against all three antigens . However, surprisingly mono-cistronic ChAdOx1 vectors appeared to be less immunogenic in the groups having received either an individual vaccine or a combination of three mono-cistronic ChAdOx1 vectors. The MVA mono-cistronic vaccines significantly increased the magnitude of the immune responses (Figure 14). The increase in antigen-specific immune responses was also observed in mice that received a polycistronic MVA vaccine.
Efficacy testing of mono- and poly-cistronic viral vectors in a mouse model.
Having demonstrated that the novel polycistronic viral vectors encoding a string of these 3 human prostate cancer associated antigens, are immunogenic in mice, next we moved to testing its tumour protective efficacy. To this end, we required a murine tumour cell line that express these three human antigens to perform tumour challenge experiments. The experimental design is to inoculate mice subcutaneously with a syngeneic tumour cell line that expressed human PSA, STEAP1 and 5T4 and, after the tumours are established, to treat the mice with the novel poly-cistronic vaccine in order to test whether tumour growth is delayed compared to mice in the control group that are challenged with the tumour and left untreated. As such murine cell line does not exist, we had to create it. We have chosen the method of lipofectamine transfection using plasmid DNA encoding all three antigens. The work of making stably transfected B16 and CT26 murine tumour cell lines (C57Bl/6 mouse strain and BALB/c mouse strain respectively) is ongoing.
Work Package 9: Project management
WP9 ensured the proper overall management of the project in order to strengthen and support the participants to achieve the objectives, complete the milestones in time and deliver the deliverables.
The management structure of the project ensured that:
• the consortium’s contractual duties were carried out
• advise and guidance were provided to the participants to comply with the EU regulations and their contractual and legal requirements.
• an effective communication infrastructure was set-up and an integrative process within the consortium was fostered.
• knowledge produced within the project was disseminated to the relevant target groups through publications and training
• both phase I and phase II clinical trials were conducted in full compliance with Good Clinical Practice
Description of main S & T results/foregrounds
The project has achieved most of its objectives and technical goals. The main objectives of the project are as follows:
• To evaluate a novel viral vectored prostate cancer vaccine in a phase I first-in-human study
• To evaluate a novel viral vectored prostate cancer vaccine in combination with checkpoint inhibitor drugs in the phase II clinical trial
• To evaluate preclinically a next generation prostate cancer vaccine encoding several cancer antigens
To accomplish its objectives, IMPROVE based the organization of the necessary management and RTD around 9 work-packages (WPs) as shown in Figure 1.
WP1 Pre-clinical evaluation of ChAdOx1-MVA with 5T4 antigen
The objectives of WP1 comprised generation of the two recombinant viral vectors, ChAdOx1 and MVA, expressing the tumour-associated antigen 5T4 and their evaluation in a mouse model prior to testing these viruses as a prostate cancer immunotherapy in a phase I clinical trial. Recombinant viral vectors expressing a murine homologue of the tumour-associated antigen 5T4, ChAdOx1.m5T4 and MVA.m5T4 were designed, manufactured and tested in a proof-of-concept study to evaluate their immunogenicity and efficacy in a mouse model. A heterologous prime boost vaccination regimen based on ChAdOx1.m5T4 priming immunisation and MVA.m5T4 boost was demonstrated to break immunological tolerance to this antigen in that both cellular and humoral 5T4 specific immune responses were detected following immunisation. A modest tumour protective efficacy resulting in delayed tumour growth was also observed in a transplantable tumour model following vaccination.
Further, the clinical batch of the ChAdOx1.5T4 vaccine was tested in mice and the treatment was well tolerated and was not associated with any adverse effects when administered at the high dose of 5x10^9 vp (virus particles).
WP2 GMP Vector Manufacture
The objective of WP2 was to manufacture a clinical batch of GMP grade ChAdOx1.5T4 and to release to trial this batch of the vaccine. GMP compatible starting material, process development and GMP manufacture of a batch of material suitable for stability study and toxicology has been completed first, followed GMP manufacture of ChAdOx1.5T4 clinical batch resulted in 300 vials available for clinical use in planned phase I and phase II clinical trials. Post QC testing (internal and external) and QA batch review, ChAdOx1.5T4 was QP certified as having met the requirements of EU GMP.
WP3: Phase I clinical trial (acronym - VANCE)
The objectives of WP3 was to perform a first-in-man clinical trial of a novel prostate cancer (PCa) vaccine comprised of the ChAdOx.5T4 and MVA.5T4 viral vectors delivered in a heterologous prime-boost regimen to low- and intermediate- risk prostate cancer patients in order to assess the safety and immunogenicity of the vaccine. 40 patients, either newly diagnosed with early stage PCa and scheduled for radical prostatectomy or patients with stable disease on active surveillance protocol, were randomised to a “standard” immunisation regimen to receive 3 vaccinations four weeks apart, or to an “accelerated” immunisation protocol to receive 2 vaccinations at one week interval. Half of the patients were also treated with low dose cyclophosphamide prior to each immunisation to deplete regulatory T cells. Study primary endpoints were vaccine safety and immunogenicity. Secondary endpoints included immune infiltration into the prostate and PSA level change secondary to treatment. As exploratory endpoints, phenotype and functionality of antigen-specific T cells and breadth of induced T cell responses were assessed. The study main findings are as follows.
39 patients completed the study and were eligible for analysis. The vaccine had an excellent safety profile, with the majority of vaccine-related adverse events graded as mild. No additional safety findings were apparent following vaccination in combination with low dose cyclophosphamide. There have been two serious adverse events (SAE) which were deemed unlikely to be related to the investigational medicinal products employed in this study. Vaccination-induced 5T4-specific T cell responses were measured in blood by ex vivo IFNγ ELISPOT and were detected in the majority of patients. Cyclophosphamide pre-conditioning has not improved frequency or magnitude of the T cell responses. Flow cytometry analysis demonstrated the presence of both CD8 and CD4 poly-functional 5T4-specific T cells in the circulation. 5T4-reactive tumour-infiltrating lymphocytes (TILs) were isolated from post-treatment prostate tissue. Some of the patients had a transient PSA level increase 2-8 weeks following vaccination possibly indicating an inflammatory response in the target organ.
Analysis of ex vivo cellular immune responses to 5T4.
The main readout of the vaccine immunogenicity was the induction of cellular and humoral 5T4-specific immune responses. To this end, blood samples from each patient have been collected at baseline and at each subsequent clinic visit. Peripheral blood mononuclear cells (PBMCs) and plasma have been separated out and used to assess the 5T4-specific T cell responses by an ex vivo IFNγ ELISPOT assay and the 5T4-specific antibody titres by semi-quantitative ELISA. In the ELISPOT assay, PBMCs have been exposed to overlapping 15-mer peptides covering the whole 5T4 protein for 18-20 hours, individual cells secreting IFNγ have been enumerated and presented as frequencies per 10^6 PBMCs. An example of the response magnitude and kinetics is shown in Figure 2, demonstrating the cellular immune responses against 5T4 peptide pools in 2 surgical patients randomised either to the standard or accelerated vaccination regimen and one active surveillance patient (accelerated regimen). Overall, detected cellular ex vivo immune responses have been encouraging taking into account that 5T4 is a self-antigen against which an immunological tolerance is likely to exist, with the majority of patients mounting relatively high T cell response to 5T4 (Figure 3).
Analysis of humoral immune responses to 5T4.
In order to evaluate the capacity of the vaccine to induce an antibody response against 5T4 a validated semi-quantitative ELISA assay has been performed on patient sera before vaccination and at different time points post-vaccination. For each patient, 5T4-specific antibody levels were assessed. The 5T4 sero-conversion rates by treatment group are displayed in Table 1. Patients who received at least 2 injections of MVA.5T4 showed a 100% sero-conversion rate for 5T4 antibodies. In contrast, patients who only received 1 injection of MVA.5T4 as part of an accelerated vaccination schedule had a lower sero-conversion rate of <50%. It is impossible to conclude from these results whether the reduced number of MVA.5T4 injections or the accelerated vaccination regimen is the cause of the reduced antibody response rate. However, based upon the kinetics of antibody responses in the homologous vaccination regimen, it is likely that the reduced number of MVA.5T4 vaccinations is the cause of the reduced antibody responsiveness. To conclude, robust 5T4-specific antibody responses are induced in patients enrolled into the standard vaccination regimen (but only weak responses are detected in patients enrolled into the accelerated vaccination groups
Study secondary endpoints.
• To address the study secondary outcomes, the patients’ PBMC samples collected at different timepoints have been cultured for 12-14 days either in medium alone or in the presence of the total 5T4 peptide pool to expand the relatively infrequent vaccine antigen-specific T cells for further analysis by cytokine flow cytometry, which allows a more detailed characterisation of vaccine-induced immune responses but has lower sensitivity and thus requires a higher number of antigen-specific cells for the assay. Following in vitro stimulation, cell cultures were stained with fluorochrome-labelled antibodies against CD4 and CD8 T cell surface markers and intra-cellular cytokines characteristic for effector T cells (i.e. IFNγ and TNFalfa). Representative data from 3 patients are shown in Figure 4. Typically, after in vitro 5T4 antigen-specific stimulation and expansion, between 0.2% and 2% of cultured cells are specific for the vaccine antigen as measured by secretion of cytokines.
• One of the study secondary endpoints is the detection of antigen-specific immune cells in the prostate, as this is where they are expected to exert their effector function and destroy tumour cells expressing the 5T4 antigen. Fresh tissue from the resection specimens or biopsies was processed to expand the immune cells resident in the prostate prior to analysis of their 5T4 specificity. In the majority of cases, high numbers of TILs were expanded from fresh tissue (with a range of 6-80 million immune cells per tissue core/biopsy). Antigen-specific tumour-infiltration lymphocytes could be detected in ~70% of the patients. Flow cytometric analysis of the TILs from one of the patients is shown in Figure 5.
• Immunohistochemical (IHC) analysis of immune cell tumour infiltration following vaccination was performed on fixed tissue. Tissue blocks from each patient were formalin-fixed and paraffin-embedded (FFPE) for IHC analysis of immune cells density in post-vaccination tissue compared to available archival tissue from pre-study diagnostic biopsy. Available paired pre- and post-treatment FFPE tissue sections have been stained for a panel of pre-selected immune cell markers including CD3 and CD8. Batch analysis was performed to quantify the density of immune cells by digital image analysis in order to assess whether the vaccination course leads to trafficking of immune cells to the prostate. The density of cells positive for these markers before and after study treatment was assessed by digital image analysis. In summary, for 12 cases diagnostic pre-treatment biopsies were compared with post-treatment surgical specimens (samples from patients in the surgical arms), and for 13 cases diagnostic pre-treatment biopsies were compared with post-treatment on-study biopsies (patients in active surveillance arms). Cell densities were calculated to the number of marker positive cells per unit area (mm2). The comparison of CD3 and CD8 cell densities in surgical patients revealed a decrease when comparing biopsies (pre-treatment) to resections (post-treatment). Interestingly, when the spatial distribution of marker positive cells was analysed the Ripley’s K index, high values for K-Ripley were observed in resections compared to pre-treatment biopsies. This finding can be explained by potential re-distribution of the immune cells following the study treatment, i.e. the immune cells in prostate tissue normally diffusely distributed in the tissue become activated due to vaccination and as a result traffic to the areas where tumour lesions are to mediate their function of killing tumour cells. This is an interesting observation that needs to be further investigated by studying matched biopsies and resection tissue of untreated control patients.
• PSA reduction in peripheral blood of advanced-stage prostate cancer patients secondary to interventions has been widely used as a surrogate marker of treatment efficacy before In our study, PSA level change was also included as one of the study endpoints. PSA levels in the blood of cancer patients at baseline (before treatment) was compared to PSA levels post vaccination course. Unexpectedly, instead of PSA level drop we have observed a transient PSA increase in some patients following vaccination. If baseline PSA values in all patients in the active surveillance arm are compared to their respective highest PSA values over a period of one year post vaccination, the difference between a baseline and the peak value is not significant (Figure 6). However, there is clearly a trend that can imply that the vaccination induced T cells have trafficked to the prostate and acted upon 5T4-expressing prostate cells leading to the leakage of PSA into the circulation.
WP4: Phase I/II clinical trial (acronym - ADVANCE)
The objective of WP4 was to undertake a phase II safety and efficacy clinical trial with the vectors encoding 5T4 with the trial design informed by results from the phase I clinical trial. In the phase II study, ChAdOx1-MVA 5T4 vaccine was tested in combination with the anti-PD-1 checkpoint blocker Nivolumab in low- or intermediate-risk localized or locally advanced prostate cancer and metastatic castration resistant prostate cancer (mCRPC). The study synopsis is shown in Table 2. Recruitment to the study started in February 2019 and by October 2019 23 out of 24 patients in the metastatic cohort have been enrolled. As of January 2020, all recruited patients either completed the study treatment and were withdrawn due to disease progression. The full data set for this patient cohort will be analysed by Q4 2020. The recruitment to the early stage prostate cancer cohort is ongoing.
• Safety. To date, administrations of ChAdOx1.5T4 and MVA.5T4 vaccines in combination with nivolumab were found to be safe and well tolerated, in agreement with our previous studies with these vectors.
• PSA response rate. Serum PSA levels were measured through the study a t each clinic visit. 5 of 19 patients demonstrated 50% or greater PSA decrease from baseline (Fig.7). This response rate is significantly higher than the response rate reported in the KEYNOTE-199 study (Antonarakis et al., 2019) of anti-PD-1 antibody alone in advanced metastatic prostate cancer. A maximal decrease in PSA concentration compared to baseline in each of 5 responders is shown in Fig.8.
• CTC enumeration. The quantification of circulating tumour cells (CTCs), which is one of the study exploratory endpoints, has been performed. Encouragingly, in some of the patients a decrease in serum PSA correlated with decrease in a number of CTCs (Fig.9).
• T cell immune responses. The primary immunogenicity measure in the study is the IFN-γ ELISpot response of PBMCs stimulated with overlapping peptide pools covering the entire length of 5T4 proteins. Responses were measured before vaccination and at different time points post-vaccination. Out of 23 patients who completed the vaccination course, 2 patients mounted 5T4-specific T cell response that was detected at several timepoints during the study (Fig. 10). The immune response rate to the vaccine is lower that the one observed in the VANCE study, but this is not surprising given the patient baseline characteristics such advanced metastatic disease and multiple potentially immunosuppressive treatment modalities prior to enrolment to the study.
WP5 Immunogenicity studies
The aim of WP5 was to carry out a detailed analysis of the immune responses both in the blood generated by the 5T4-targeting prostate cancer vaccine deployed in the phase I and II clinical trials. Dissection of CD8 and CD4 T cell immune responses was performed by in vitro expansion and flow cytometry. The main achievements are as follows:
Test and distribution of protocol for tumour-infiltration lymphocytes (TILs) expansion from prostate cancer biopsies
• Phenotype of TILs derived from Prostatic carcinoma to a non-cancer control was completed.
• The ex-vivo TILs phenotype was investigated and the phenotypic changes evaluated after in vitro expansion.
• We observed high level of similarity between expanded TILs from PC and BPH samples. T cells are the predominant population in both pre-REP and REP expanded condition and NK cells are completely lost only after REP expansion.
• In BPH samples the expansion with pre-REP and REP protocols generated a predominant T cell population that progressively upregulates activation/exhaustion and cytotoxic markers in comparison to TILs directly analyzed from digested disuse ex-vivo.
Development of a flow cytometry panel to monitor regulatory T (T-reg) cells in peripheral blood
• T-reg flow cytometry panel with 14 markers was developed
Mapping of T cell responses in vaccinated patents
• In vitro culture protocol and mapping of CD4 and CD8 T cell responses in 25 cancer patients vaccinated for 5T4 protein.
• Several CD4 and CD8 T cell epitopes were identified. Interestingly, many of those were never published previously.
Development and harmonization of protocol for testing immunogenicity and T cell phenotype longitudinally in prostate cancer patients enrolled into the study.
• Ex-vivo and In-vitro culture, stimulation and intracellular staining protocols were developed and harmonized between UNIL and OX (analysis performed in 3 healthy donors).
• Harmonization resulted successful. Test of the developed flow cytometry panel on PBMC of 3 healthy donors gave comparable results.
An extended flow cytometry panel was validated for use in the phase II clinical trial for detailed phenotypic and functional characteristics of induced T cell responses (Table 3).
WP6 Correlates of Immunity and efficacy
The objective of WP6 is to build on the phase I and phase I/II clinical trials. In phase I trial, VANCE, Blood samples recovered from patients recruited to the trial have been analysed for antigen-specific responses and the results used to investigate whether there are any potential associations between immune responses and clinical benefit. Analysis of correlates of immunogenicity and efficacy is dependent on the provision of data from WP5 and suitable markers of clinical efficacy. Given that the patients treated in the phase I clinical trial were at a relatively early-stage in their disease course, no recognised markers of clinical benefit (e.g. response rate, survival etc) were available within the timeframe of this work package. As such, the key analytic testing looked to identify predictors of immunogenicity.
An analysis of potential correlates with immunogenicity was undertaken using antibody response data. Previously, a pre-treatment biomarker was identified which predicted 5T4 specific immune responses induced in patients following vaccination with MVA-5T4 (TroVax; Harrop et al., Cancer Immunology Immunotherapy 2011). The same pre-treatment biomarker (consisting of baseline levels of haemoglobin, haematocrit and 5T4 antibody level) was used to determine if this biomarker was associated with the fold-increase in 5T4 or MVA antibodies induced following vaccination with ChAdOX.5T4 and MVA-5T4. These exploratory analyses suggested a significant relationship between the pre-treatment biomarker and the 5T4, but not MVA, antibody responses induced post vaccination. An assessment of correlation coefficient between the pre-treatment biomarker and 5T4- and MVA- specific antibody responses at weeks 1, 2, 4 and 8 are shown in Table 4. As can be seen, there is no significant correlation between the pre-treatment biomarker and antibody responses detected at week 1; this is not surprising as it normally takes >1 vaccination before antibody responses against the target antigen are detected. However, at weeks 2, 4 and 8 post-vaccination there are strong trends (weeks 2 and 8) or significant correlations (week 4) between the pre-treatment biomarker and the 5T4-specific antibody response, but not the MVA-specific antibody response. These data replicate what have been reported previously i.e. significant correlations between pre-treatment biomarker and antibody responses against the target antigen (5T4), but not against the viral vector (MVA).
These data are of potential significance because 8 clinical trials using MVA-5T4 have demonstrated a link between the magnitude of the 5T4-specific immune response and clinical benefit (tumour shrinkage or overall survival). The ability to select the patients who are most likely to benefit from this class of therapy prior to treating them would be a huge advantage. Further confirmatory prospective studies would be required to develop a test which predicted treatment benefit.
For correlation between immune response and clinical efficacy in the phase I/II study, ADVANCE, blood samples from 23 recruited patients were processed to isolated PBMCs for IFN-γ ELISPOT assays and to measure the number of antigen-specific T cells induced by vaccination. Patients’ blood and serum samples were also collected for CTC enumeration and PSA concentration.
Out of 23 patients, 2 patients mounted vaccine-specific T cell immune response and 5 patients had a PSA decrease from baseline of 50% or more and 5 patients had a CTC decrease from baseline. One of 2 patients with immunological response had also a decrease in the number of CTC but no PSA reduction. So there clearly was no detectable correlation between our primary immunogenicity read-out (ELISpot) and the available and most information efficacy read-out (PSA reduction). However, this correlation analysis is very underpowered because of the small number of responders in the ELISpot assay. This low frequency contrast with the findings in our phase I VANCE trial where the same vaccines produced a >65% response rate. The difference may reflect a difference in the general immune competence of early and late-stage prostate cancer patients. A correlation analysis was also performed on the patients who had either PSA or CTC decrease or both and although there is as yet no significant difference on small numbers we will extend this analysis once CTC results are available from more subjects (Fig. 11).
WP7 Pre-clinical antigen evaluation – immunogenicity
WP7 comprises the characterisation of novel tumour-associated antigens and the assessment of their immunogenicity following expression from the viral vectors. The original focus was on the characterization of three antigens, namely CRISP3, CNPY2 and DPY19L3 discovered by Externautics partner (EXT). Overall, the study reinforced the association of these antigens with prostate cancer and provided a first indication of the antigen expression in early PCa and pre-cancerous stages. The expression of the three antigens was confirmed in different human prostate cell lines representing different PCa subtypes. Moreover, the murine counterparts of the three antigens were found expressed in murine prostate cell lines, including TRAMP cell lines, thus justifying the use of the TRAMP mouse model for efficacy studies. Finally, the three protein antigens showed an overall marginal expression in human normal tissues, thus allowing to hypothesize an acceptable toxicity profile of vaccine based on these antigens. Results specifically related to each antigen are summarized below:
• CNPY2 (Canopy FGF Signalling Regulator 2)
Human CNPY2 protein is annotated in two isoforms of 182 and 84 amino acids. The 182 aa isoform is predicted to be associated with the endoplasmic reticulum while the 84 aa isoform is secreted or intracellular.
Expression in PCa: IHC analysis of 100 prostate cancer samples and 47 PIN cases showed that CNPY2 was detected in all tested samples with the highest expression in 17% of PCa and 17% of tested PIN samples.
Expression in normal human tissue: IHC analysis of 33 normal human tissues showed weak cytoplasmic staining only in stomach, colon and chorionic villi and negligible reactivity in the other 31 tissues. Western blot (WB) and flow cytometry (FACS) analyses of PBMC purified from healthy individuals showed moderate CNPY2 expression, which was confined to the intracellular compartments.
Endogenous expression in prostate cell lines: CNPY2 was clearly expressed in all tested human prostate cell lines (R22V1, VCaP, PC3, PNT1, LNCaP, DU145) and in the murine cell lines TRAMP-C1 and TRAMP-C2, as judged by Q-RT-PCR and Western blot using anti-CNPY2 antibody. Western blot analysis allowed detection of a protein band of 20 kDa, compatible with the 182 amino acid isoform. CNPY2 was also well expressed in murine prostate TRAMP cell lines. FACS analysis showed that the protein is confined to the intracellular compartment.
• DPY19L3 (probable C-mannosyltransferase)
DPY19L3 is the least characterized protein of the group. The protein has been annotated in at least 3 overlapping isoforms with predicted MW of 83, 63 and 12 kDa, respectively. Only one isoform is predicted for the murine DPY19L3 orthologues, showing 88% amino acid identity with the 716 amino acid human isoform and a predicted MW of 83 kDa.
Expression in PCa and PIN: IHC analysis of 100 PCa showed that the protein was detected in 44.4% of PCa, among which 6.1% showed the highest expression. Analysis of 47 PIN cases showed that the protein is detected in 74% of the samples, 13% of which showed the highest expression. Staining was generally cytoplasmic and in some samples a membranous staining was visible. Staining of normal prostate was negligible.
Expression in normal human tissue: IHC analysis of the 33 normal tissues by IHC did not show any significant staining. In addition, DPY19L3 expression was not detected in human PBMC or in neutrophils by Western blot or by FACS
Endogenous expression in prostate cell lines: DPY19L3 expression analysis in human prostate cell lines by Q-RT-PCR revealed the presence of the three annotated DPY19L3 transcripts. Western blot revealed two major protein bands of approximately 90 and 60 kDa, compatible with the two longer DPY19L3 isoforms. FACS analysis showed a moderate surface staining of VCaP and R22V1 cells suggesting that it is at least partially surface exposed in these cells. DPY19L3 expression in murine prostate TRAMP cell lines was confirmed at transcript level. In these cells, Western blot revealed a unique band of approximately 60 kDa that might result from proteolytic processing of the annotated 83kDa protein.
• CRISP3 (Cystein-rich secretory protein 3)
Human CRISP3 protein is annotated with at least two variants of 258 and 268 amino acids and is predicted to be secreted. In prostate carcinoma, it has been reported as associated with a subset of prostate carcinoma carrying the TMPRSS2-ERG fusion, pT3 disease stage. Two potential CRISP3 orthologues are annotated in the mouse, namely CRISP1 and CRISP3, sharing 54% amino acid identity with the human protein, whose role in cancer has not been previously described.
Expression in PCa and PIN: IHC analysis of 100 PCa cases showed that the protein was detected in 29% of PCa samples, among which 19% showed the highest expression. Analysis of 47 PIN cases showed that the protein is detected in 23% of PIN, 13% of which showed the highest expression. Staining was generally cytoplasmic, except for a few samples that also showed a membranous staining.
Expression in normal human tissue: CRISP3 was not detected in human PBMC, as assessed by WB and FACS. Instead, in agreement with published data, the protein was detected in the neutrophil population. IHC analysis of the 35 normal human tissues did not give any relevant expression of the protein.
Endogenous expression in prostate cell lines: among the different human cell lines currently tested CRISP3 was exclusively detected in the VCaP cell line, having the TMPRSS2-ERG fusion. In these cells WB analysis with the anti-CRISP3 antibody revealed the presence of a band of expected size (approximately 30 kDa). The antibody also bound to the surface of VCaP cells in FACS staining assay, indicating that CRISP3, besides being secreted, is at least partially exposed on the cell surface. Concerning CRISP3 expression in the murine TRAMP cells, transcription profile analysis by Q-RT-PCR of the tumours excised from TRAMP mice showed that mCRISP1 was detected in two of the three tested biopsies whereas mCRISP3 cDNA was not detected, indicating that mCRISP1 could be the protein form expressed in murine prostate cancer.
Construction of recombinant viral vectors ChAdOx1 and MVA expressing novel prostate-associated antigens, CRISP1, CNPY2, DPY19L3, and comparative immunogenicity of these novel antigens and well defined prostate-associated antigens 5T4 and STEAP1 that also represent potential antigenic targets for a prostate cancer vaccine.
Recombinant viral vectors ChAdOx1 and MVA, expressing the novel prostate associated antigens CNPY2, DPY19L3 and CRISP3 were designed and generated for immunogenicity testing in mice.
These vectors delivered in a heterologous ChAdOx1 prime – MVA boost vaccination regime have been tested for immunogenicity in C57Bl/6 mouse strain. The cellular immune responses following immunisation have been assessed by an ex vivo IFNg ELISPOT assay against pools of 15-mer peptides overlapping by 10 amino acids spanning the full length of the proteins encoded by the vector transgenes. As a result, there have been no detectable T cell reactivity against CRISP1 and CNPY2 antigens, however, there was potentially a weak response induced against the DPY19L3. These results are in striking contrast to the immunogenicity of another murine prostate-specific antigen, STEAP1, induced by ChAdOx1-MVA immunisation (Cappuccini et al 2016). Although disappointing, these results were not unexpected as the central and/or peripheral immunological tolerance usually prevents an induction of immune responses against self-antigens. We have demonstrated that exceptionally strong T cell reactivity against STEAP1 is likely to be due to the lack of its expression in the thymus, although the mRNA transcripts corresponding to the m5T4, CRISP1, CNPY2 and DPY19L13 are detectable in the murine thymus by reverse transcription PCR.
The novel prostate-associated antigens identified by EXT have not been taken further into efficacy testing due to paucity of the vaccine-induced immune responses against these antigens. Instead the Steering Committee decided to proceed with the pre-clinical testing of the panel of human antigens, PSA (prostate-specific antigen), STEAP1 (six transmembrane antigen of the prostate 1) and ERG (ETS transcription factor) as a polyvalent vaccine with the aim of taking them into the clinic in the future. To this end, firstly the single antigen vectors expressing these antigens in native form or fusions to the invariant chain have been made for immunogenicity testing. With further clinical development in mind, fusion constructs have been made with the transmembrane domain of the shark invariant chain for the following reasons. Firstly, the shark and human invariant chain share only 30% homology so the potential risk of autoimmunity will be greatly reduced. The 25 amino acid long transmembrane domain is the minimal sequence that is required to maintain the enhancer effect, so the risk of autoimmunity will be reduced even further. Strong immune responses against PSA and STEAP1 have been induced in a mouse model but tolerance to ERG antigen could not be broken (Figure 12) therefore this antigen has been excluded from the multi-antigen construct developed on WP8.
WP8 Pre-clinical antigen evaluation - efficacy
Objectives of WP8 comprise the efficacy evaluation of 5T4 encoding vaccines compared to the vectors expressing novel antigens, assessment of immune correlates of efficacy in mouse models and generation of new vectors encoding the most promising vaccine targets to be available for early stage clinical development.
According to the original project plan, the novel EXT antigens identified and described in WP7 should have been taken into further pre-clinical development in order to create the vectors available for early stage clinical development at the end of the project. As the antigens CRISP-1, CNPY-2 and DPY19L3 appeared to be of low immunogenicity or non-immunogenic at all, as an alternative, we have chosen the three well-defined prostate-associated antigens, prostate-specific antigen (PSA), six transmembrane epithelial antigen of the prostate – 1 (STEAP-1), and 5T4 antigen for further pre-clinical and clinical development and expressed a string of these antigens from the same vector thus creating multi-antigen ChAdOx1 and MVA vectors. Of note, the ERG antigen originally planned as a component of the multi-antigen vaccine, was replaced by 5T4 as the phase I clinical trial (WP3) results clearly indicated that both CD8 and CD4 T cell responses against 5T4 can be induced in prostate cancer patients.
PSA
Human PSA, a chymotrypsin-like serine protease, has a highly restricted tissue distribution and is expressed in the epithelial cells of the prostate gland, the same cell type from which most prostate tumours arise. Indeed, PSA is widely used as a serum marker for prostate cancer. Its expression is regulated by androgen, and itis present at extremely low levels in the circulation of adult men. Most prostate tumours, even the poorly differentiated ones, continue to express PSA. This cell type-specific expression of PSA makes it a potential target antigen for antitumor CTL. In fact, PSA is a target antigen of one of the two most clinically advanced prostate cancer vaccines - ProstVac.
STEAP1
The six-transmembrane epithelial antigen of prostate protein was identified in advanced prostate cancer. STEAP1 is highly expressed in human prostate cancer and is up-regulated in various cancers, including lung, bladder, colon, ovarian, and Ewing cancers. Immunohistochemical analysis of clinical specimens demonstrates significant STEAP1 expression at the cell–cell junctions of the secretory epithelium of prostate and prostate cancer cells. Little to no staining was detected at the plasma membranes of normal non-prostate human tissues, except for bladder tissue, which expressed low levels of STEAP1 at the cell membrane. Its cell-surface localization, together with its six-transmembrane topology, suggests STEAP1 may function as a channel/transporter protein in cell–cell junctions. Given its increased expression in cancer tissues, STEAP1 could be a promising target for T-cell based or antibody immunotherapy. In our previous experiments, we have expressed murine STEAP1 from ChAdOx1 and MVA viral vectors and investigated STEAP1-specific T cell response in mice vaccinated with these vaccines. Despite being a self-antigen, STEAP1 appeared to be highly immunogenic in mice, most likely because of absence of central tolerance to this antigen. STEAP1 has been evaluated in early stage prostate cancer clinical trials and delivered in the form of RNA as part of a multi-antigen vaccine with encouraging preliminary data.
Construction of ChAd and MVA viral vectors.
Both mono-cistronic ChAdOx1 and MVA vectors encoding individual antigens and multi-antigen vectors expressing all three antigens were constructed for immunogenicity testing in mice. Poly-cistronic vectors were constructed to express a string of three antigens from the least immunogenic to the most immunogenic antigen, i.e. with STEAP1 cDNA downstream of the promoter followed by 5T4 and PSA cDNA. Flexible polypeptide linkers composed of glycine and proline (GGG-P-GGG) were inserted between cDNAs to create a fusion protein.
Immunogenicity testing of mono- and poly- cistronic viral vectors in a mouse model.
To assess the immunogenicity of mono- and poly- cistronic viral vectors encoding human antigens PSA, STEAP1 and 5T4 in a mouse model, outbred mice (CD1 strain) were randomised into 5 groups. Mice were primed with ChAdOx1 virus encoding each individual antigen or with the vector expressing a string of antigens and were boosted with the MVA vectors accordingly. The induction of T cell immune responses against the vaccine transgenes was tested by ex vivo IFN-γ Elispot assay after priming and boosting immunisations. As evident from the Figure 13, T cell responses of relatively high magnitude are detectable in blood after a single immunisation with a polcistronic ChAdOx1 vector against all three antigens . However, surprisingly mono-cistronic ChAdOx1 vectors appeared to be less immunogenic in the groups having received either an individual vaccine or a combination of three mono-cistronic ChAdOx1 vectors. The MVA mono-cistronic vaccines significantly increased the magnitude of the immune responses (Figure 14). The increase in antigen-specific immune responses was also observed in mice that received a polycistronic MVA vaccine.
Efficacy testing of mono- and poly-cistronic viral vectors in a mouse model.
Having demonstrated that the novel polycistronic viral vectors encoding a string of these 3 human prostate cancer associated antigens, are immunogenic in mice, next we moved to testing its tumour protective efficacy. To this end, we required a murine tumour cell line that express these three human antigens to perform tumour challenge experiments. The experimental design is to inoculate mice subcutaneously with a syngeneic tumour cell line that expressed human PSA, STEAP1 and 5T4 and, after the tumours are established, to treat the mice with the novel poly-cistronic vaccine in order to test whether tumour growth is delayed compared to mice in the control group that are challenged with the tumour and left untreated. As such murine cell line does not exist, we had to create it. We have chosen the method of lipofectamine transfection using plasmid DNA encoding all three antigens. The work of making stably transfected B16 and CT26 murine tumour cell lines (C57Bl/6 mouse strain and BALB/c mouse strain respectively) is ongoing.
Work Package 9: Project management
WP9 ensured the proper overall management of the project in order to strengthen and support the participants to achieve the objectives, complete the milestones in time and deliver the deliverables.
The management structure of the project ensured that:
• the consortium’s contractual duties were carried out
• advise and guidance were provided to the participants to comply with the EU regulations and their contractual and legal requirements.
• an effective communication infrastructure was set-up and an integrative process within the consortium was fostered.
• knowledge produced within the project was disseminated to the relevant target groups through publications and training
• both phase I and phase II clinical trials were conducted in full compliance with Good Clinical Practice
Potential Impact:
Potential impact and main dissemination activities and exploitation results.
Prostate cancer is the second most common cancer in North American and European men and the second leading cause of male cancer related death after lung cancer. Only in the UK Over 47,000 men are diagnosed with prostate cancer every year – that's 129 men every day, and every 45 minutes one man dies from prostate cancer – that's more than 11,000 men every year. Although the number of therapeutic options for patients with advanced stage prostate cancer has increased recently, the treatments remain largely palliative and often have severe side-effects.
Immunotherapy has been evolving as a potential new treatment modality offering the promise of long-term treatment effect with a minimal side-effect profile. Despite, some early failures of immunotherapy products in phase III trials, a key milestone was achieved in 2010 when the FDA approved Provenge (Sipuleucel-T) as the first therapeutic cancer vaccine for the treatment of patients with advanced prostate cancer. This individualised treatment costs over $90,000 per patient and provides a modest survival benefit of 4 months. Taking into account the cost-benefit ratio and lack of a clear mechanistic explana¬tion of the clinical benefit, a less expensive more effective standard immunisation approach, which avoids the multiple rounds of leukapheresis associated with this treatment, would be highly desirable. A major impact of this proposal is expected to be a new, significantly less expensive and more efficacious prostate cancer vaccine available as a monotherapy and as a combinatorial treatment with a PD-1 blockade for late stage clinical development within the next few years.
A further potential impact of this program arises from the identification and testing of a new, more rapid, clinical efficacy protocol for prostate immunotherapies. We have assess a 4 week and 12 week intervention strategy and utilised histological and biochemical measures of efficacy within this time period, along with longer term monitoring of biochemical and clinical impact. If markers of efficacy can be detected within the 12 week window, this would considerably accelerate the testing of new vaccination approaches.
Class Leading Immunological Potency
By definition, an immunotherapy product must act through the induction of an efficacious immune response. In cancer, the target molecule is usually a tumour associated antigen (TAA) which is derived from “self”; therefore the induction of an immune response against the TAA requires the breaking of tolerance. As such, any vaccination strategy needs to be sufficiently potent to ensure that tolerance to the self-antigen is broken. Various strategies have been utilized to achieve this goal ,and to date, results with various antigen/delivery methods have been mixed with only a few reporting strong immune responses against the tumour antigen in the majority of treated patients. Where immune responses have been detected, an association between the magnitude of the response and indicators of clinical benefit has been frequently reported. A key goal of this program was to induce strong cytotoxic T cell a well as antibody responses against the target tumour antigen (5T4) in the great majority of treated patients. The aim of this program was to combine the experience derived from the use of MVA.5T4 in a homologous prime-boost regimen with the experience derived from use of a heterologous prime-boost regimen in which MVA as a booster has been combined with simian adenovirus as a potent priming agent. This novel heterologous prime-boost regimen has been tested in over 40 trials in infectious disease settings and shown to be highly immunogenic across all studies. Indeed, we have demonstrated that the combination of simian adenovirus and MVA in a heterologous prime-boost regimen consistently generates the strongest recorded sustained CD8+ T cell immune responses in humans of any subunit vaccine approach previously deployed. The immunological basis of the ChAd-MVA regime potency is quite well understood. In humans, non-human primates and mice adenoviral vectors outperform all other technologies at priming a potent CD8 T cell response with a single immunisation, likely due to a preferred combination of innate receptor ligands. MVA is a particularly good boosting agent, but only a modest priming vector. This is due to the 180 genes in MVA that compete with the transgene when priming new response; however, on boosting the MVA-induced response uses natural immunodominance mechanisms to amplify the primed immune response to the transgene product which outcompete new primary responses. The consistent consequence is a CD8 T cell response amplified 5 -10 fold compared to the ChAd (or MVA vector) used alone. As expected, the delivery of 5T4 using 2 different vectors (simian adenovirus and MVA) in a heterologous prime boost regimen resulted in substantially enhanced immune responses against the target tumour antigen in terms of frequency and magnitude of responses. In the context of the mediocre immune responses previously reported for the majority of cancer vaccines, it is believed that the results from this study will provide a significant step forward.
Safety and Tolerability
The majority of approved cancer therapies currently in use have a significant side-effect profile which can usually be managed, but can have a significant impact on a patients’ quality of life. The therapeutic vaccine approach proposed here has demonstrated its safety and tolerability, compatible with prophylactic vaccine standards, in multiple clinical trials. As such, the patient compliance rate was very high and there were no toxicity issues.
As more treatments become available, clinicians will increasingly look to combinatorial approaches to enhance clinical benefit to the patient. Indeed, we recognise that the optimal therapies for most cancers will often comprise multimodality treatments e.g. chemotherapy, radiotherapy and various types of immunotherapy. However, our aim here was to improve a leading form of vaccine therapy, prior to assessment of combination strategies. As we expected, the safety profile of the conducted trials was excellent, and therefore, the vaccination approach has the potential to be used in combination with many different standard therapies without impacting on the side-effect profile or diminishing the efficacy of the standard treatment. Such flexibility broadens the potential therapeutic window of a vaccine approach beyond niche indications or settings where no standard treatments are available.
Cost of Goods
Sipeuleucel-T is currently the only immunotherapy approved for the treatment of prostate cancer. The treatment is patient specific and requires leukapheresis of the patient in order to provide the active treatment modality. Such a treatment is both cumbersome and expensive ($90,000 per patient). The vaccination strategy proposed in this application represents an “off the shelf” approach (i.e. not patient specific) and can be manufactured at low cost (with costs of goods of each of the two components at scale of less than a euro). For example, ChAd vectors have been manufactured for clinical use on a variety of cell lines, the widely available HEK293 cells, the new Procell cell line that allows tetracycline repression of the transgene, and the PER C6 cell line from Crucell. There are also three duck and quail-derived cell lines available for the manufactured of MVA that lead to low costs of goods and avoid the need for chick embryo fibroblast for large scale biomanufacture. Therefore, this product configuration has potential for widespread use and low health care system costs, in contrast to most other novel cancer interventions.
New Targets for Therapeutic Intervention
The identification of tumour antigens which are suitable for immune targeted therapeutic strategies remains a challenge. The preclinical part of the proposed work assessed novel prostate specific antigens as potential candidates for therapeutic intervention, aiming for a multi-antigen second generation vaccine. Combining several partially protective antigens in a vaccine will provide synergistic efficacy. Another advantage of targeting several antigens is the increased breadth of the immune response and the reduced likelihood of both escape mutations and selection of more aggressive tumour phenotypes. As anticipated, pre-clinical evaluation of the novel vaccine candidates translated into a new multi-antigen prostate cancer vaccine available for early stage clinical development by the end of this project.
Additionally, while this proposal aims to test the vaccine in prostate cancer patients, the strategy is broadly applicable to the design and development of vaccines against other types of cancer since 5T4 and some of the antigens that we plan to test are present in other tumour types.
A Platform for Therapeutic Vaccination
This project opened new research avenues for therapeutic cancer vaccination in general as well as prostate cancer immunotherapy specifically. If good immunogenicity and significant efficacy can be achieved for prostate cancer in this programme there are many opportunities for applying the ChAd-MVA vectored prime-boost approach to many other cancers. Furthermore, success with ChAd-MVA immunisation in prostate cancer immunotherapy would open the way to assessment on the same vaccines in several other cancers. For example, 5T4 is known to be well expressed on colorectal, breast and renal tumours so that the same vectors could also be assessed in these cancers.
Impact on the Vaccine Field as a Whole
Some of the potential impacts of this programme on the vaccine field as a whole are worth noting. The interest in the ability of vectored vaccines to induce very potent T cell responses is of potentially even greater impact in the area of therapeutic vaccination where CD8 T cell response are known to play a particularly important role, and antibody-based vaccines are less likely to be effective. This project provided the first key test of concept of this approach in cancer immunotherapy.
The dissemination of results
Main dissemination activities of the consortium included scientific publications in academic journals, presentations at the national and international conferences and workshops, press releases to the media and public awareness campaigns.
The IMPROVE partners participated in local, national and international level workshops and meetings that are relevant in all areas relating to the project. Results from the project were presented at scientific meetings for the benefit of the wider scientific community. Regular seminars organized by the partners in their own institutions were implemented.
Engagement of patients and public into discussion of the clinical trial design and study related documents constituted important part of dissemination activities. The patients were made aware of the project objectives and ongoing clinical trials through the participant information sheets distributed through the clinical teams participating in the trials
Experience has shown that when patients were engaged right from the onset in the research planning process, they were more committed to applying the research findings in real life and policy settings.
The project website played an important role in targeting the general public.
Any dissemination activities and publications in the project, including the project website specified that the project had received Community research funding and displayed the European emblem.
List of Websites:
http://www.project-improve.eu/home(opens in new window)
Project co-ordinator: Professor Adrian Hill, adrian.hill@ndm.ox.ac.uk
IMPROVE was a project funded by the European Commission through its Seventh Framework Programme (FP7). It brought together two academic institutions and five industrial partners from across Europe with the common objective of developing a novel prostate cancer immunotherapy.
Cancer immunotherapy has recently emerged as a viable and attractive treatment strategy. The only licensed therapeutic prostate cancer vaccine, Sipuleucel-T, provided a modest survival benefit of 4 months and showed low immunogenicity. In this project, we have clinically evaluated the most potent vaccine technology available for inducing cellular immunity to target prostate cancer. This highly immunogenic vaccination platform deploys the replication-deficient simian adenovirus, ChAdOx1, as a priming agent and attenuated poxvirus, MVA, as a boosting vaccine. Both viral vectors encode the tumour associated antigen 5T4 overexpressed in prostate tumours.
The output of the IMPROVE project was remarkable, and the most significant achievements are highlighted below. For the first time, a heterologous vaccination platform based on ChAdOx1 and MVA vectors was deployed clinically in a cancer setting (a phase I clinical trial – VANCE, NCT02390063). As a result, T cell immune responses against the tumour-specific antigen, 5T4, encoded by the vaccine were detected in the majority of vaccinated patients ex vivo, without prior expansion in culture. This is a prominent result given that 5T4 is a self-antigen against which the immune system is tolerized. A detailed immunomonitoring programme was undertaken that allowed for identification of significant number of CD8- and CD4- restricted T cell epitopes with the vaccine antigen and their HLA restriction elements which will potentially inform the design of a next generation vaccine. Importantly, T cell immune responses were induced not only in the blood but also the immune cell infiltration was detected in the target organ, the prostate, thus challenging previously held belief that prostate cancer is a “cold” tumour which is not amenable to immunotherapies. The objectives of this first-in human clinical study were fully achieved, having demonstrated that the vaccine was safe and induced vaccine antigen specific immune responses in the majority of patients.
One of the major impediments to tumour protective vaccine efficacy is an up-regulation of inhibitory receptors on effector T cells and tumour cells induced by immunisation. In the second clinical trial within this programme, we aimed to evaluate the clinical efficacy of the vaccine in combination with an agent that blocks inhibitory receptor, and thus disrupts the pathways that attenuate the vaccine induced immune response. This phase I/II clinical trial (ADVANCE, NCT03815942) recruited 23 patients with metastatic disease with progression after hormone therapy, where the vaccine is combined with an anti PD-1 checkpoint inhibitor Nivolumab. The preliminary data from this trial demonstrated that the vaccine in combination with nivolumab is safe in this patients cohort. The clinical efficacy as measured by PSA reduction of 50% or greater compared to baseline have been achieved in 26% of patients which is significantly higher than the efficacy of nivolumab as a single agent in the reported KEYNOTE-199 study. A reduction in the number of circulating tumour cells was also observed in some of the PSA responders.
In parallel, we have undertaken a preclinical programme on comparative immunogenicity and efficacy assessment of a range of new and old cancer antigens in mouse models of prostate cancer to identify the most effective antigens for clinical development as an immunotherapeutic, with an output of the ChAdOx1 and MVA multi-antigen vectors encoding three prostate-specific antigens, PSA, STEAP1 and 5T4, being available for early clinical development.
A further part of the IMPROVE project was communications with the public and dissemination of general knowledge about prostate cancer vaccines. IMPROVE gained visibility through a number of presentations at the prestigious international conferences and publications in scientific journals. After funding of the project the website will be still active and summaries of the results of the publications will be disseminated on the website.
This five year programme has combined the most potent vaccine technology available for inducing cellular immunity in humans with leading cancer antigens to target prostate cancer, for which there is substantial evidence that vaccine immunotherapy is feasible.
Project Context and Objectives:
Summary description of the project context and the main objectives.
Prostate cancer (PCa) has been under investigation as a target for antigen-specific immunotherapies in metastatic disease settings. Neither of the two clinically most advanced PCa vaccines, Sipuleucel-T and ProstVac, induced strong T cell immunity and their clinical efficacy was modest. Prostate cancer has not been considered a good target for checkpoint inhibitor therapies, given its low mutational load and lack of pre-existing intratumoural immune cell infiltration, until recent data from KEYNOTE-199 study (Antonarakis, J. Clin Oncol. 2019) demonstrated a 9% response rate to pembrolizumab monotherapy.
The major clinical development goal of IMPROVE was the evaluation of a new very potent vector-based immunisation strategy in the context of a design that allows the measurement of multiple sensitive early endpoints that may represent vaccine efficacy. The immunisation approach is based on two viral vectors that, in a heterologous prime-boost immunisation regime, have yielded the highest measured CD8 T cell responses in humans across a range of prophylactic and therapeutic vaccine indications, following on from pre-clinical studies that demonstrated exceptional potency.
IMPROVE has developed a novel vaccination platform based on two replication-deficient viruses, chimpanzee adenovirus and MVA, targeting an oncofetal self-antigen 5T4 and evaluated this vaccine alone and in combination with anti-PD-1 in mouse tumour models.
Next, we have tested this vaccine in a first-in-human trial, VANCE (NCT02390063), in early stage prostate cancer patients. The patients, either newly diagnosed with early stage PCa and scheduled for radical prostatectomy or patients with stable disease on active surveillance protocol, were randomised to a “standard” immunisation regimen to receive 3 vaccinations four weeks apart, or to an “accelerated” immunisation protocol to receive 2 vaccinations at one week interval. Study primary endpoints were vaccine safety and immunogenicity.
Extensive studies of immune down-regulatory mechanisms in cancer have now identified a range of mechanisms that attenuate the immunogenicity of candidate cancer vaccines and also reduce the efficacy of the induced immune responses in the target tumour. Among those, an up-regulation of inhibitory receptors on effector T cells and tumour cells induced by immunisation is one of the major impediments to tumour protective vaccine efficacy. The second part of the clinical programme was the evaluation of this immunisation strategy in combination with a checkpoint inhibitor therapy in a phase II trial, ADVANCE (NCT03815942), to test the vaccine safety and efficacy in combination with PD-1 blockade in intermediate risk diseases and metastatic prostate cancer. This trial is designed to ask if early evidence of efficacy in combination with immune checkpoint blocker can be detected using sensitive established biochemical, histological and radiological indicators of vaccine efficacy as well as novel methods of genomic analyses in advanced metastatic prostate cancer. The patients will be followed for one year with regular PSA, circulating tumour cells (CTC) and circulating tumour DNA (ctDNA) measurements to determine the duration of biochemical relapse-free survival and to monitor tumour burden and correlate the early measures of efficacy with this established surrogate measure of relapse free survival and also clinical measures. The preliminary data from ADVANCE trial demonstrated that the vaccine in combination with nivolumab has good safety profile. The clinical efficacy as measured by PSA reduction of 50% or greater compared to baseline have been achieved in 26% of patients which is significantly higher than the efficacy of nivolumab as a single agent in the reported KEYNOTE-199 study. A reduction in the number of circulating tumour cells and reduction in tumour lesions were also observed in some of the PSA responders.
The goal of a preclinical component of IMPROVE was development of a novel multi-antigen prostate cancer vaccine. Extensive antigen discovery efforts have led to the identification of several older and new candidate antigens for prostate cancer vaccines, but these have never been compared head to head. Therefore, in parallel with preclinical and clinical programme on the development of ChAdOx1 and MVA vaccines expressing the 5T4 antigen, we have undertaken a preclinical development of a second generation vaccine vectors encoding 3 prostate-associated antigens – 5T4, PSA and STEAP1.
The main objectives of the IMPROVE project are listed below.
• Preclinical evaluation of ChAdOx1-MVA with 5T4 antigen
• GMP manufacturing of viral vectored vaccines for phase I and II clinical trials
• Assessment of safety and immunogenicity of the novel prostate cancer immunisation approach in intermediate risk i prostate cancer in a phase I clinical trial
• Assessment of safety and efficacy of the prostate cancer vaccine in combination with checkpoint inhibitors in intermediate risk and advanced metastatic prostate cancer in a phase I/II clinical trial
• Detailed characterization of the vaccine-induced immune responses in terms of breadth and functionality
• Analysis of correlates of immunogenicity and MCHC pre-treatment biomarker
• Comparative preclinical antigen evaluation in terms of immunogenicity and efficacy for a next generation multi-antigen vaccine.
Dissemination and exploitation
In addition to this technical work there was a programme of dissemination activity:
• Dissemination of project results to clinicians, industry, the public and investment communities
• Conference presentations and peer-reviewed journal publications
• Newsletters, public website and other dissemination material
The project has achieved success across all these target areas.
This report will present the project results and provide contact information for those desiring further information.
Project Results:
Description of main S & T results/foregrounds.
The project has achieved most of its objectives and technical goals. The main objectives of the project are as follows:
• To evaluate a novel viral vectored prostate cancer vaccine in a phase I first-in-human study
• To evaluate a novel viral vectored prostate cancer vaccine in combination with checkpoint inhibitor drugs in the phase II clinical trial
• To evaluate preclinically a next generation prostate cancer vaccine encoding several cancer antigens
To accomplish its objectives, IMPROVE based the organization of the necessary management and RTD around 9 work-packages (WPs) as shown in Figure 1.
WP1 Pre-clinical evaluation of ChAdOx1-MVA with 5T4 antigen
The objectives of WP1 comprised generation of the two recombinant viral vectors, ChAdOx1 and MVA, expressing the tumour-associated antigen 5T4 and their evaluation in a mouse model prior to testing these viruses as a prostate cancer immunotherapy in a phase I clinical trial. Recombinant viral vectors expressing a murine homologue of the tumour-associated antigen 5T4, ChAdOx1.m5T4 and MVA.m5T4 were designed, manufactured and tested in a proof-of-concept study to evaluate their immunogenicity and efficacy in a mouse model. A heterologous prime boost vaccination regimen based on ChAdOx1.m5T4 priming immunisation and MVA.m5T4 boost was demonstrated to break immunological tolerance to this antigen in that both cellular and humoral 5T4 specific immune responses were detected following immunisation. A modest tumour protective efficacy resulting in delayed tumour growth was also observed in a transplantable tumour model following vaccination.
Further, the clinical batch of the ChAdOx1.5T4 vaccine was tested in mice and the treatment was well tolerated and was not associated with any adverse effects when administered at the high dose of 5x10^9 vp (virus particles).
WP2 GMP Vector Manufacture
The objective of WP2 was to manufacture a clinical batch of GMP grade ChAdOx1.5T4 and to release to trial this batch of the vaccine. GMP compatible starting material, process development and GMP manufacture of a batch of material suitable for stability study and toxicology has been completed first, followed GMP manufacture of ChAdOx1.5T4 clinical batch resulted in 300 vials available for clinical use in planned phase I and phase II clinical trials. Post QC testing (internal and external) and QA batch review, ChAdOx1.5T4 was QP certified as having met the requirements of EU GMP.
WP3: Phase I clinical trial (acronym - VANCE)
The objectives of WP3 was to perform a first-in-man clinical trial of a novel prostate cancer (PCa) vaccine comprised of the ChAdOx.5T4 and MVA.5T4 viral vectors delivered in a heterologous prime-boost regimen to low- and intermediate- risk prostate cancer patients in order to assess the safety and immunogenicity of the vaccine. 40 patients, either newly diagnosed with early stage PCa and scheduled for radical prostatectomy or patients with stable disease on active surveillance protocol, were randomised to a “standard” immunisation regimen to receive 3 vaccinations four weeks apart, or to an “accelerated” immunisation protocol to receive 2 vaccinations at one week interval. Half of the patients were also treated with low dose cyclophosphamide prior to each immunisation to deplete regulatory T cells. Study primary endpoints were vaccine safety and immunogenicity. Secondary endpoints included immune infiltration into the prostate and PSA level change secondary to treatment. As exploratory endpoints, phenotype and functionality of antigen-specific T cells and breadth of induced T cell responses were assessed. The study main findings are as follows.
39 patients completed the study and were eligible for analysis. The vaccine had an excellent safety profile, with the majority of vaccine-related adverse events graded as mild. No additional safety findings were apparent following vaccination in combination with low dose cyclophosphamide. There have been two serious adverse events (SAE) which were deemed unlikely to be related to the investigational medicinal products employed in this study. Vaccination-induced 5T4-specific T cell responses were measured in blood by ex vivo IFNγ ELISPOT and were detected in the majority of patients. Cyclophosphamide pre-conditioning has not improved frequency or magnitude of the T cell responses. Flow cytometry analysis demonstrated the presence of both CD8 and CD4 poly-functional 5T4-specific T cells in the circulation. 5T4-reactive tumour-infiltrating lymphocytes (TILs) were isolated from post-treatment prostate tissue. Some of the patients had a transient PSA level increase 2-8 weeks following vaccination possibly indicating an inflammatory response in the target organ.
Analysis of ex vivo cellular immune responses to 5T4.
The main readout of the vaccine immunogenicity was the induction of cellular and humoral 5T4-specific immune responses. To this end, blood samples from each patient have been collected at baseline and at each subsequent clinic visit. Peripheral blood mononuclear cells (PBMCs) and plasma have been separated out and used to assess the 5T4-specific T cell responses by an ex vivo IFNγ ELISPOT assay and the 5T4-specific antibody titres by semi-quantitative ELISA. In the ELISPOT assay, PBMCs have been exposed to overlapping 15-mer peptides covering the whole 5T4 protein for 18-20 hours, individual cells secreting IFNγ have been enumerated and presented as frequencies per 10^6 PBMCs. An example of the response magnitude and kinetics is shown in Figure 2, demonstrating the cellular immune responses against 5T4 peptide pools in 2 surgical patients randomised either to the standard or accelerated vaccination regimen and one active surveillance patient (accelerated regimen). Overall, detected cellular ex vivo immune responses have been encouraging taking into account that 5T4 is a self-antigen against which an immunological tolerance is likely to exist, with the majority of patients mounting relatively high T cell response to 5T4 (Figure 3).
Analysis of humoral immune responses to 5T4.
In order to evaluate the capacity of the vaccine to induce an antibody response against 5T4 a validated semi-quantitative ELISA assay has been performed on patient sera before vaccination and at different time points post-vaccination. For each patient, 5T4-specific antibody levels were assessed. The 5T4 sero-conversion rates by treatment group are displayed in Table 1. Patients who received at least 2 injections of MVA.5T4 showed a 100% sero-conversion rate for 5T4 antibodies. In contrast, patients who only received 1 injection of MVA.5T4 as part of an accelerated vaccination schedule had a lower sero-conversion rate of <50%. It is impossible to conclude from these results whether the reduced number of MVA.5T4 injections or the accelerated vaccination regimen is the cause of the reduced antibody response rate. However, based upon the kinetics of antibody responses in the homologous vaccination regimen, it is likely that the reduced number of MVA.5T4 vaccinations is the cause of the reduced antibody responsiveness. To conclude, robust 5T4-specific antibody responses are induced in patients enrolled into the standard vaccination regimen (but only weak responses are detected in patients enrolled into the accelerated vaccination groups
Study secondary endpoints.
• To address the study secondary outcomes, the patients’ PBMC samples collected at different timepoints have been cultured for 12-14 days either in medium alone or in the presence of the total 5T4 peptide pool to expand the relatively infrequent vaccine antigen-specific T cells for further analysis by cytokine flow cytometry, which allows a more detailed characterisation of vaccine-induced immune responses but has lower sensitivity and thus requires a higher number of antigen-specific cells for the assay. Following in vitro stimulation, cell cultures were stained with fluorochrome-labelled antibodies against CD4 and CD8 T cell surface markers and intra-cellular cytokines characteristic for effector T cells (i.e. IFNγ and TNFalfa). Representative data from 3 patients are shown in Figure 4. Typically, after in vitro 5T4 antigen-specific stimulation and expansion, between 0.2% and 2% of cultured cells are specific for the vaccine antigen as measured by secretion of cytokines.
• One of the study secondary endpoints is the detection of antigen-specific immune cells in the prostate, as this is where they are expected to exert their effector function and destroy tumour cells expressing the 5T4 antigen. Fresh tissue from the resection specimens or biopsies was processed to expand the immune cells resident in the prostate prior to analysis of their 5T4 specificity. In the majority of cases, high numbers of TILs were expanded from fresh tissue (with a range of 6-80 million immune cells per tissue core/biopsy). Antigen-specific tumour-infiltration lymphocytes could be detected in ~70% of the patients. Flow cytometric analysis of the TILs from one of the patients is shown in Figure 5.
• Immunohistochemical (IHC) analysis of immune cell tumour infiltration following vaccination was performed on fixed tissue. Tissue blocks from each patient were formalin-fixed and paraffin-embedded (FFPE) for IHC analysis of immune cells density in post-vaccination tissue compared to available archival tissue from pre-study diagnostic biopsy. Available paired pre- and post-treatment FFPE tissue sections have been stained for a panel of pre-selected immune cell markers including CD3 and CD8. Batch analysis was performed to quantify the density of immune cells by digital image analysis in order to assess whether the vaccination course leads to trafficking of immune cells to the prostate. The density of cells positive for these markers before and after study treatment was assessed by digital image analysis. In summary, for 12 cases diagnostic pre-treatment biopsies were compared with post-treatment surgical specimens (samples from patients in the surgical arms), and for 13 cases diagnostic pre-treatment biopsies were compared with post-treatment on-study biopsies (patients in active surveillance arms). Cell densities were calculated to the number of marker positive cells per unit area (mm2). The comparison of CD3 and CD8 cell densities in surgical patients revealed a decrease when comparing biopsies (pre-treatment) to resections (post-treatment). Interestingly, when the spatial distribution of marker positive cells was analysed the Ripley’s K index, high values for K-Ripley were observed in resections compared to pre-treatment biopsies. This finding can be explained by potential re-distribution of the immune cells following the study treatment, i.e. the immune cells in prostate tissue normally diffusely distributed in the tissue become activated due to vaccination and as a result traffic to the areas where tumour lesions are to mediate their function of killing tumour cells. This is an interesting observation that needs to be further investigated by studying matched biopsies and resection tissue of untreated control patients.
• PSA reduction in peripheral blood of advanced-stage prostate cancer patients secondary to interventions has been widely used as a surrogate marker of treatment efficacy before In our study, PSA level change was also included as one of the study endpoints. PSA levels in the blood of cancer patients at baseline (before treatment) was compared to PSA levels post vaccination course. Unexpectedly, instead of PSA level drop we have observed a transient PSA increase in some patients following vaccination. If baseline PSA values in all patients in the active surveillance arm are compared to their respective highest PSA values over a period of one year post vaccination, the difference between a baseline and the peak value is not significant (Figure 6). However, there is clearly a trend that can imply that the vaccination induced T cells have trafficked to the prostate and acted upon 5T4-expressing prostate cells leading to the leakage of PSA into the circulation.
WP4: Phase I/II clinical trial (acronym - ADVANCE)
The objective of WP4 was to undertake a phase II safety and efficacy clinical trial with the vectors encoding 5T4 with the trial design informed by results from the phase I clinical trial. In the phase II study, ChAdOx1-MVA 5T4 vaccine was tested in combination with the anti-PD-1 checkpoint blocker Nivolumab in low- or intermediate-risk localized or locally advanced prostate cancer and metastatic castration resistant prostate cancer (mCRPC). The study synopsis is shown in Table 2. Recruitment to the study started in February 2019 and by October 2019 23 out of 24 patients in the metastatic cohort have been enrolled. As of January 2020, all recruited patients either completed the study treatment and were withdrawn due to disease progression. The full data set for this patient cohort will be analysed by Q4 2020. The recruitment to the early stage prostate cancer cohort is ongoing.
• Safety. To date, administrations of ChAdOx1.5T4 and MVA.5T4 vaccines in combination with nivolumab were found to be safe and well tolerated, in agreement with our previous studies with these vectors.
• PSA response rate. Serum PSA levels were measured through the study a t each clinic visit. 5 of 19 patients demonstrated 50% or greater PSA decrease from baseline (Fig.7). This response rate is significantly higher than the response rate reported in the KEYNOTE-199 study (Antonarakis et al., 2019) of anti-PD-1 antibody alone in advanced metastatic prostate cancer. A maximal decrease in PSA concentration compared to baseline in each of 5 responders is shown in Fig.8.
• CTC enumeration. The quantification of circulating tumour cells (CTCs), which is one of the study exploratory endpoints, has been performed. Encouragingly, in some of the patients a decrease in serum PSA correlated with decrease in a number of CTCs (Fig.9).
• T cell immune responses. The primary immunogenicity measure in the study is the IFN-γ ELISpot response of PBMCs stimulated with overlapping peptide pools covering the entire length of 5T4 proteins. Responses were measured before vaccination and at different time points post-vaccination. Out of 23 patients who completed the vaccination course, 2 patients mounted 5T4-specific T cell response that was detected at several timepoints during the study (Fig. 10). The immune response rate to the vaccine is lower that the one observed in the VANCE study, but this is not surprising given the patient baseline characteristics such advanced metastatic disease and multiple potentially immunosuppressive treatment modalities prior to enrolment to the study.
WP5 Immunogenicity studies
The aim of WP5 was to carry out a detailed analysis of the immune responses both in the blood generated by the 5T4-targeting prostate cancer vaccine deployed in the phase I and II clinical trials. Dissection of CD8 and CD4 T cell immune responses was performed by in vitro expansion and flow cytometry. The main achievements are as follows:
Test and distribution of protocol for tumour-infiltration lymphocytes (TILs) expansion from prostate cancer biopsies
• Phenotype of TILs derived from Prostatic carcinoma to a non-cancer control was completed.
• The ex-vivo TILs phenotype was investigated and the phenotypic changes evaluated after in vitro expansion.
• We observed high level of similarity between expanded TILs from PC and BPH samples. T cells are the predominant population in both pre-REP and REP expanded condition and NK cells are completely lost only after REP expansion.
• In BPH samples the expansion with pre-REP and REP protocols generated a predominant T cell population that progressively upregulates activation/exhaustion and cytotoxic markers in comparison to TILs directly analyzed from digested disuse ex-vivo.
Development of a flow cytometry panel to monitor regulatory T (T-reg) cells in peripheral blood
• T-reg flow cytometry panel with 14 markers was developed
Mapping of T cell responses in vaccinated patents
• In vitro culture protocol and mapping of CD4 and CD8 T cell responses in 25 cancer patients vaccinated for 5T4 protein.
• Several CD4 and CD8 T cell epitopes were identified. Interestingly, many of those were never published previously.
Development and harmonization of protocol for testing immunogenicity and T cell phenotype longitudinally in prostate cancer patients enrolled into the study.
• Ex-vivo and In-vitro culture, stimulation and intracellular staining protocols were developed and harmonized between UNIL and OX (analysis performed in 3 healthy donors).
• Harmonization resulted successful. Test of the developed flow cytometry panel on PBMC of 3 healthy donors gave comparable results.
An extended flow cytometry panel was validated for use in the phase II clinical trial for detailed phenotypic and functional characteristics of induced T cell responses (Table 3).
WP6 Correlates of Immunity and efficacy
The objective of WP6 is to build on the phase I and phase I/II clinical trials. In phase I trial, VANCE, Blood samples recovered from patients recruited to the trial have been analysed for antigen-specific responses and the results used to investigate whether there are any potential associations between immune responses and clinical benefit. Analysis of correlates of immunogenicity and efficacy is dependent on the provision of data from WP5 and suitable markers of clinical efficacy. Given that the patients treated in the phase I clinical trial were at a relatively early-stage in their disease course, no recognised markers of clinical benefit (e.g. response rate, survival etc) were available within the timeframe of this work package. As such, the key analytic testing looked to identify predictors of immunogenicity.
An analysis of potential correlates with immunogenicity was undertaken using antibody response data. Previously, a pre-treatment biomarker was identified which predicted 5T4 specific immune responses induced in patients following vaccination with MVA-5T4 (TroVax; Harrop et al., Cancer Immunology Immunotherapy 2011). The same pre-treatment biomarker (consisting of baseline levels of haemoglobin, haematocrit and 5T4 antibody level) was used to determine if this biomarker was associated with the fold-increase in 5T4 or MVA antibodies induced following vaccination with ChAdOX.5T4 and MVA-5T4. These exploratory analyses suggested a significant relationship between the pre-treatment biomarker and the 5T4, but not MVA, antibody responses induced post vaccination. An assessment of correlation coefficient between the pre-treatment biomarker and 5T4- and MVA- specific antibody responses at weeks 1, 2, 4 and 8 are shown in Table 4. As can be seen, there is no significant correlation between the pre-treatment biomarker and antibody responses detected at week 1; this is not surprising as it normally takes >1 vaccination before antibody responses against the target antigen are detected. However, at weeks 2, 4 and 8 post-vaccination there are strong trends (weeks 2 and 8) or significant correlations (week 4) between the pre-treatment biomarker and the 5T4-specific antibody response, but not the MVA-specific antibody response. These data replicate what have been reported previously i.e. significant correlations between pre-treatment biomarker and antibody responses against the target antigen (5T4), but not against the viral vector (MVA).
These data are of potential significance because 8 clinical trials using MVA-5T4 have demonstrated a link between the magnitude of the 5T4-specific immune response and clinical benefit (tumour shrinkage or overall survival). The ability to select the patients who are most likely to benefit from this class of therapy prior to treating them would be a huge advantage. Further confirmatory prospective studies would be required to develop a test which predicted treatment benefit.
For correlation between immune response and clinical efficacy in the phase I/II study, ADVANCE, blood samples from 23 recruited patients were processed to isolated PBMCs for IFN-γ ELISPOT assays and to measure the number of antigen-specific T cells induced by vaccination. Patients’ blood and serum samples were also collected for CTC enumeration and PSA concentration.
Out of 23 patients, 2 patients mounted vaccine-specific T cell immune response and 5 patients had a PSA decrease from baseline of 50% or more and 5 patients had a CTC decrease from baseline. One of 2 patients with immunological response had also a decrease in the number of CTC but no PSA reduction. So there clearly was no detectable correlation between our primary immunogenicity read-out (ELISpot) and the available and most information efficacy read-out (PSA reduction). However, this correlation analysis is very underpowered because of the small number of responders in the ELISpot assay. This low frequency contrast with the findings in our phase I VANCE trial where the same vaccines produced a >65% response rate. The difference may reflect a difference in the general immune competence of early and late-stage prostate cancer patients. A correlation analysis was also performed on the patients who had either PSA or CTC decrease or both and although there is as yet no significant difference on small numbers we will extend this analysis once CTC results are available from more subjects (Fig. 11).
WP7 Pre-clinical antigen evaluation – immunogenicity
WP7 comprises the characterisation of novel tumour-associated antigens and the assessment of their immunogenicity following expression from the viral vectors. The original focus was on the characterization of three antigens, namely CRISP3, CNPY2 and DPY19L3 discovered by Externautics partner (EXT). Overall, the study reinforced the association of these antigens with prostate cancer and provided a first indication of the antigen expression in early PCa and pre-cancerous stages. The expression of the three antigens was confirmed in different human prostate cell lines representing different PCa subtypes. Moreover, the murine counterparts of the three antigens were found expressed in murine prostate cell lines, including TRAMP cell lines, thus justifying the use of the TRAMP mouse model for efficacy studies. Finally, the three protein antigens showed an overall marginal expression in human normal tissues, thus allowing to hypothesize an acceptable toxicity profile of vaccine based on these antigens. Results specifically related to each antigen are summarized below:
• CNPY2 (Canopy FGF Signalling Regulator 2)
Human CNPY2 protein is annotated in two isoforms of 182 and 84 amino acids. The 182 aa isoform is predicted to be associated with the endoplasmic reticulum while the 84 aa isoform is secreted or intracellular.
Expression in PCa: IHC analysis of 100 prostate cancer samples and 47 PIN cases showed that CNPY2 was detected in all tested samples with the highest expression in 17% of PCa and 17% of tested PIN samples.
Expression in normal human tissue: IHC analysis of 33 normal human tissues showed weak cytoplasmic staining only in stomach, colon and chorionic villi and negligible reactivity in the other 31 tissues. Western blot (WB) and flow cytometry (FACS) analyses of PBMC purified from healthy individuals showed moderate CNPY2 expression, which was confined to the intracellular compartments.
Endogenous expression in prostate cell lines: CNPY2 was clearly expressed in all tested human prostate cell lines (R22V1, VCaP, PC3, PNT1, LNCaP, DU145) and in the murine cell lines TRAMP-C1 and TRAMP-C2, as judged by Q-RT-PCR and Western blot using anti-CNPY2 antibody. Western blot analysis allowed detection of a protein band of 20 kDa, compatible with the 182 amino acid isoform. CNPY2 was also well expressed in murine prostate TRAMP cell lines. FACS analysis showed that the protein is confined to the intracellular compartment.
• DPY19L3 (probable C-mannosyltransferase)
DPY19L3 is the least characterized protein of the group. The protein has been annotated in at least 3 overlapping isoforms with predicted MW of 83, 63 and 12 kDa, respectively. Only one isoform is predicted for the murine DPY19L3 orthologues, showing 88% amino acid identity with the 716 amino acid human isoform and a predicted MW of 83 kDa.
Expression in PCa and PIN: IHC analysis of 100 PCa showed that the protein was detected in 44.4% of PCa, among which 6.1% showed the highest expression. Analysis of 47 PIN cases showed that the protein is detected in 74% of the samples, 13% of which showed the highest expression. Staining was generally cytoplasmic and in some samples a membranous staining was visible. Staining of normal prostate was negligible.
Expression in normal human tissue: IHC analysis of the 33 normal tissues by IHC did not show any significant staining. In addition, DPY19L3 expression was not detected in human PBMC or in neutrophils by Western blot or by FACS
Endogenous expression in prostate cell lines: DPY19L3 expression analysis in human prostate cell lines by Q-RT-PCR revealed the presence of the three annotated DPY19L3 transcripts. Western blot revealed two major protein bands of approximately 90 and 60 kDa, compatible with the two longer DPY19L3 isoforms. FACS analysis showed a moderate surface staining of VCaP and R22V1 cells suggesting that it is at least partially surface exposed in these cells. DPY19L3 expression in murine prostate TRAMP cell lines was confirmed at transcript level. In these cells, Western blot revealed a unique band of approximately 60 kDa that might result from proteolytic processing of the annotated 83kDa protein.
• CRISP3 (Cystein-rich secretory protein 3)
Human CRISP3 protein is annotated with at least two variants of 258 and 268 amino acids and is predicted to be secreted. In prostate carcinoma, it has been reported as associated with a subset of prostate carcinoma carrying the TMPRSS2-ERG fusion, pT3 disease stage. Two potential CRISP3 orthologues are annotated in the mouse, namely CRISP1 and CRISP3, sharing 54% amino acid identity with the human protein, whose role in cancer has not been previously described.
Expression in PCa and PIN: IHC analysis of 100 PCa cases showed that the protein was detected in 29% of PCa samples, among which 19% showed the highest expression. Analysis of 47 PIN cases showed that the protein is detected in 23% of PIN, 13% of which showed the highest expression. Staining was generally cytoplasmic, except for a few samples that also showed a membranous staining.
Expression in normal human tissue: CRISP3 was not detected in human PBMC, as assessed by WB and FACS. Instead, in agreement with published data, the protein was detected in the neutrophil population. IHC analysis of the 35 normal human tissues did not give any relevant expression of the protein.
Endogenous expression in prostate cell lines: among the different human cell lines currently tested CRISP3 was exclusively detected in the VCaP cell line, having the TMPRSS2-ERG fusion. In these cells WB analysis with the anti-CRISP3 antibody revealed the presence of a band of expected size (approximately 30 kDa). The antibody also bound to the surface of VCaP cells in FACS staining assay, indicating that CRISP3, besides being secreted, is at least partially exposed on the cell surface. Concerning CRISP3 expression in the murine TRAMP cells, transcription profile analysis by Q-RT-PCR of the tumours excised from TRAMP mice showed that mCRISP1 was detected in two of the three tested biopsies whereas mCRISP3 cDNA was not detected, indicating that mCRISP1 could be the protein form expressed in murine prostate cancer.
Construction of recombinant viral vectors ChAdOx1 and MVA expressing novel prostate-associated antigens, CRISP1, CNPY2, DPY19L3, and comparative immunogenicity of these novel antigens and well defined prostate-associated antigens 5T4 and STEAP1 that also represent potential antigenic targets for a prostate cancer vaccine.
Recombinant viral vectors ChAdOx1 and MVA, expressing the novel prostate associated antigens CNPY2, DPY19L3 and CRISP3 were designed and generated for immunogenicity testing in mice.
These vectors delivered in a heterologous ChAdOx1 prime – MVA boost vaccination regime have been tested for immunogenicity in C57Bl/6 mouse strain. The cellular immune responses following immunisation have been assessed by an ex vivo IFNg ELISPOT assay against pools of 15-mer peptides overlapping by 10 amino acids spanning the full length of the proteins encoded by the vector transgenes. As a result, there have been no detectable T cell reactivity against CRISP1 and CNPY2 antigens, however, there was potentially a weak response induced against the DPY19L3. These results are in striking contrast to the immunogenicity of another murine prostate-specific antigen, STEAP1, induced by ChAdOx1-MVA immunisation (Cappuccini et al 2016). Although disappointing, these results were not unexpected as the central and/or peripheral immunological tolerance usually prevents an induction of immune responses against self-antigens. We have demonstrated that exceptionally strong T cell reactivity against STEAP1 is likely to be due to the lack of its expression in the thymus, although the mRNA transcripts corresponding to the m5T4, CRISP1, CNPY2 and DPY19L13 are detectable in the murine thymus by reverse transcription PCR.
The novel prostate-associated antigens identified by EXT have not been taken further into efficacy testing due to paucity of the vaccine-induced immune responses against these antigens. Instead the Steering Committee decided to proceed with the pre-clinical testing of the panel of human antigens, PSA (prostate-specific antigen), STEAP1 (six transmembrane antigen of the prostate 1) and ERG (ETS transcription factor) as a polyvalent vaccine with the aim of taking them into the clinic in the future. To this end, firstly the single antigen vectors expressing these antigens in native form or fusions to the invariant chain have been made for immunogenicity testing. With further clinical development in mind, fusion constructs have been made with the transmembrane domain of the shark invariant chain for the following reasons. Firstly, the shark and human invariant chain share only 30% homology so the potential risk of autoimmunity will be greatly reduced. The 25 amino acid long transmembrane domain is the minimal sequence that is required to maintain the enhancer effect, so the risk of autoimmunity will be reduced even further. Strong immune responses against PSA and STEAP1 have been induced in a mouse model but tolerance to ERG antigen could not be broken (Figure 12) therefore this antigen has been excluded from the multi-antigen construct developed on WP8.
WP8 Pre-clinical antigen evaluation - efficacy
Objectives of WP8 comprise the efficacy evaluation of 5T4 encoding vaccines compared to the vectors expressing novel antigens, assessment of immune correlates of efficacy in mouse models and generation of new vectors encoding the most promising vaccine targets to be available for early stage clinical development.
According to the original project plan, the novel EXT antigens identified and described in WP7 should have been taken into further pre-clinical development in order to create the vectors available for early stage clinical development at the end of the project. As the antigens CRISP-1, CNPY-2 and DPY19L3 appeared to be of low immunogenicity or non-immunogenic at all, as an alternative, we have chosen the three well-defined prostate-associated antigens, prostate-specific antigen (PSA), six transmembrane epithelial antigen of the prostate – 1 (STEAP-1), and 5T4 antigen for further pre-clinical and clinical development and expressed a string of these antigens from the same vector thus creating multi-antigen ChAdOx1 and MVA vectors. Of note, the ERG antigen originally planned as a component of the multi-antigen vaccine, was replaced by 5T4 as the phase I clinical trial (WP3) results clearly indicated that both CD8 and CD4 T cell responses against 5T4 can be induced in prostate cancer patients.
PSA
Human PSA, a chymotrypsin-like serine protease, has a highly restricted tissue distribution and is expressed in the epithelial cells of the prostate gland, the same cell type from which most prostate tumours arise. Indeed, PSA is widely used as a serum marker for prostate cancer. Its expression is regulated by androgen, and itis present at extremely low levels in the circulation of adult men. Most prostate tumours, even the poorly differentiated ones, continue to express PSA. This cell type-specific expression of PSA makes it a potential target antigen for antitumor CTL. In fact, PSA is a target antigen of one of the two most clinically advanced prostate cancer vaccines - ProstVac.
STEAP1
The six-transmembrane epithelial antigen of prostate protein was identified in advanced prostate cancer. STEAP1 is highly expressed in human prostate cancer and is up-regulated in various cancers, including lung, bladder, colon, ovarian, and Ewing cancers. Immunohistochemical analysis of clinical specimens demonstrates significant STEAP1 expression at the cell–cell junctions of the secretory epithelium of prostate and prostate cancer cells. Little to no staining was detected at the plasma membranes of normal non-prostate human tissues, except for bladder tissue, which expressed low levels of STEAP1 at the cell membrane. Its cell-surface localization, together with its six-transmembrane topology, suggests STEAP1 may function as a channel/transporter protein in cell–cell junctions. Given its increased expression in cancer tissues, STEAP1 could be a promising target for T-cell based or antibody immunotherapy. In our previous experiments, we have expressed murine STEAP1 from ChAdOx1 and MVA viral vectors and investigated STEAP1-specific T cell response in mice vaccinated with these vaccines. Despite being a self-antigen, STEAP1 appeared to be highly immunogenic in mice, most likely because of absence of central tolerance to this antigen. STEAP1 has been evaluated in early stage prostate cancer clinical trials and delivered in the form of RNA as part of a multi-antigen vaccine with encouraging preliminary data.
Construction of ChAd and MVA viral vectors.
Both mono-cistronic ChAdOx1 and MVA vectors encoding individual antigens and multi-antigen vectors expressing all three antigens were constructed for immunogenicity testing in mice. Poly-cistronic vectors were constructed to express a string of three antigens from the least immunogenic to the most immunogenic antigen, i.e. with STEAP1 cDNA downstream of the promoter followed by 5T4 and PSA cDNA. Flexible polypeptide linkers composed of glycine and proline (GGG-P-GGG) were inserted between cDNAs to create a fusion protein.
Immunogenicity testing of mono- and poly- cistronic viral vectors in a mouse model.
To assess the immunogenicity of mono- and poly- cistronic viral vectors encoding human antigens PSA, STEAP1 and 5T4 in a mouse model, outbred mice (CD1 strain) were randomised into 5 groups. Mice were primed with ChAdOx1 virus encoding each individual antigen or with the vector expressing a string of antigens and were boosted with the MVA vectors accordingly. The induction of T cell immune responses against the vaccine transgenes was tested by ex vivo IFN-γ Elispot assay after priming and boosting immunisations. As evident from the Figure 13, T cell responses of relatively high magnitude are detectable in blood after a single immunisation with a polcistronic ChAdOx1 vector against all three antigens . However, surprisingly mono-cistronic ChAdOx1 vectors appeared to be less immunogenic in the groups having received either an individual vaccine or a combination of three mono-cistronic ChAdOx1 vectors. The MVA mono-cistronic vaccines significantly increased the magnitude of the immune responses (Figure 14). The increase in antigen-specific immune responses was also observed in mice that received a polycistronic MVA vaccine.
Efficacy testing of mono- and poly-cistronic viral vectors in a mouse model.
Having demonstrated that the novel polycistronic viral vectors encoding a string of these 3 human prostate cancer associated antigens, are immunogenic in mice, next we moved to testing its tumour protective efficacy. To this end, we required a murine tumour cell line that express these three human antigens to perform tumour challenge experiments. The experimental design is to inoculate mice subcutaneously with a syngeneic tumour cell line that expressed human PSA, STEAP1 and 5T4 and, after the tumours are established, to treat the mice with the novel poly-cistronic vaccine in order to test whether tumour growth is delayed compared to mice in the control group that are challenged with the tumour and left untreated. As such murine cell line does not exist, we had to create it. We have chosen the method of lipofectamine transfection using plasmid DNA encoding all three antigens. The work of making stably transfected B16 and CT26 murine tumour cell lines (C57Bl/6 mouse strain and BALB/c mouse strain respectively) is ongoing.
Work Package 9: Project management
WP9 ensured the proper overall management of the project in order to strengthen and support the participants to achieve the objectives, complete the milestones in time and deliver the deliverables.
The management structure of the project ensured that:
• the consortium’s contractual duties were carried out
• advise and guidance were provided to the participants to comply with the EU regulations and their contractual and legal requirements.
• an effective communication infrastructure was set-up and an integrative process within the consortium was fostered.
• knowledge produced within the project was disseminated to the relevant target groups through publications and training
• both phase I and phase II clinical trials were conducted in full compliance with Good Clinical Practice
Description of main S & T results/foregrounds
The project has achieved most of its objectives and technical goals. The main objectives of the project are as follows:
• To evaluate a novel viral vectored prostate cancer vaccine in a phase I first-in-human study
• To evaluate a novel viral vectored prostate cancer vaccine in combination with checkpoint inhibitor drugs in the phase II clinical trial
• To evaluate preclinically a next generation prostate cancer vaccine encoding several cancer antigens
To accomplish its objectives, IMPROVE based the organization of the necessary management and RTD around 9 work-packages (WPs) as shown in Figure 1.
WP1 Pre-clinical evaluation of ChAdOx1-MVA with 5T4 antigen
The objectives of WP1 comprised generation of the two recombinant viral vectors, ChAdOx1 and MVA, expressing the tumour-associated antigen 5T4 and their evaluation in a mouse model prior to testing these viruses as a prostate cancer immunotherapy in a phase I clinical trial. Recombinant viral vectors expressing a murine homologue of the tumour-associated antigen 5T4, ChAdOx1.m5T4 and MVA.m5T4 were designed, manufactured and tested in a proof-of-concept study to evaluate their immunogenicity and efficacy in a mouse model. A heterologous prime boost vaccination regimen based on ChAdOx1.m5T4 priming immunisation and MVA.m5T4 boost was demonstrated to break immunological tolerance to this antigen in that both cellular and humoral 5T4 specific immune responses were detected following immunisation. A modest tumour protective efficacy resulting in delayed tumour growth was also observed in a transplantable tumour model following vaccination.
Further, the clinical batch of the ChAdOx1.5T4 vaccine was tested in mice and the treatment was well tolerated and was not associated with any adverse effects when administered at the high dose of 5x10^9 vp (virus particles).
WP2 GMP Vector Manufacture
The objective of WP2 was to manufacture a clinical batch of GMP grade ChAdOx1.5T4 and to release to trial this batch of the vaccine. GMP compatible starting material, process development and GMP manufacture of a batch of material suitable for stability study and toxicology has been completed first, followed GMP manufacture of ChAdOx1.5T4 clinical batch resulted in 300 vials available for clinical use in planned phase I and phase II clinical trials. Post QC testing (internal and external) and QA batch review, ChAdOx1.5T4 was QP certified as having met the requirements of EU GMP.
WP3: Phase I clinical trial (acronym - VANCE)
The objectives of WP3 was to perform a first-in-man clinical trial of a novel prostate cancer (PCa) vaccine comprised of the ChAdOx.5T4 and MVA.5T4 viral vectors delivered in a heterologous prime-boost regimen to low- and intermediate- risk prostate cancer patients in order to assess the safety and immunogenicity of the vaccine. 40 patients, either newly diagnosed with early stage PCa and scheduled for radical prostatectomy or patients with stable disease on active surveillance protocol, were randomised to a “standard” immunisation regimen to receive 3 vaccinations four weeks apart, or to an “accelerated” immunisation protocol to receive 2 vaccinations at one week interval. Half of the patients were also treated with low dose cyclophosphamide prior to each immunisation to deplete regulatory T cells. Study primary endpoints were vaccine safety and immunogenicity. Secondary endpoints included immune infiltration into the prostate and PSA level change secondary to treatment. As exploratory endpoints, phenotype and functionality of antigen-specific T cells and breadth of induced T cell responses were assessed. The study main findings are as follows.
39 patients completed the study and were eligible for analysis. The vaccine had an excellent safety profile, with the majority of vaccine-related adverse events graded as mild. No additional safety findings were apparent following vaccination in combination with low dose cyclophosphamide. There have been two serious adverse events (SAE) which were deemed unlikely to be related to the investigational medicinal products employed in this study. Vaccination-induced 5T4-specific T cell responses were measured in blood by ex vivo IFNγ ELISPOT and were detected in the majority of patients. Cyclophosphamide pre-conditioning has not improved frequency or magnitude of the T cell responses. Flow cytometry analysis demonstrated the presence of both CD8 and CD4 poly-functional 5T4-specific T cells in the circulation. 5T4-reactive tumour-infiltrating lymphocytes (TILs) were isolated from post-treatment prostate tissue. Some of the patients had a transient PSA level increase 2-8 weeks following vaccination possibly indicating an inflammatory response in the target organ.
Analysis of ex vivo cellular immune responses to 5T4.
The main readout of the vaccine immunogenicity was the induction of cellular and humoral 5T4-specific immune responses. To this end, blood samples from each patient have been collected at baseline and at each subsequent clinic visit. Peripheral blood mononuclear cells (PBMCs) and plasma have been separated out and used to assess the 5T4-specific T cell responses by an ex vivo IFNγ ELISPOT assay and the 5T4-specific antibody titres by semi-quantitative ELISA. In the ELISPOT assay, PBMCs have been exposed to overlapping 15-mer peptides covering the whole 5T4 protein for 18-20 hours, individual cells secreting IFNγ have been enumerated and presented as frequencies per 10^6 PBMCs. An example of the response magnitude and kinetics is shown in Figure 2, demonstrating the cellular immune responses against 5T4 peptide pools in 2 surgical patients randomised either to the standard or accelerated vaccination regimen and one active surveillance patient (accelerated regimen). Overall, detected cellular ex vivo immune responses have been encouraging taking into account that 5T4 is a self-antigen against which an immunological tolerance is likely to exist, with the majority of patients mounting relatively high T cell response to 5T4 (Figure 3).
Analysis of humoral immune responses to 5T4.
In order to evaluate the capacity of the vaccine to induce an antibody response against 5T4 a validated semi-quantitative ELISA assay has been performed on patient sera before vaccination and at different time points post-vaccination. For each patient, 5T4-specific antibody levels were assessed. The 5T4 sero-conversion rates by treatment group are displayed in Table 1. Patients who received at least 2 injections of MVA.5T4 showed a 100% sero-conversion rate for 5T4 antibodies. In contrast, patients who only received 1 injection of MVA.5T4 as part of an accelerated vaccination schedule had a lower sero-conversion rate of <50%. It is impossible to conclude from these results whether the reduced number of MVA.5T4 injections or the accelerated vaccination regimen is the cause of the reduced antibody response rate. However, based upon the kinetics of antibody responses in the homologous vaccination regimen, it is likely that the reduced number of MVA.5T4 vaccinations is the cause of the reduced antibody responsiveness. To conclude, robust 5T4-specific antibody responses are induced in patients enrolled into the standard vaccination regimen (but only weak responses are detected in patients enrolled into the accelerated vaccination groups
Study secondary endpoints.
• To address the study secondary outcomes, the patients’ PBMC samples collected at different timepoints have been cultured for 12-14 days either in medium alone or in the presence of the total 5T4 peptide pool to expand the relatively infrequent vaccine antigen-specific T cells for further analysis by cytokine flow cytometry, which allows a more detailed characterisation of vaccine-induced immune responses but has lower sensitivity and thus requires a higher number of antigen-specific cells for the assay. Following in vitro stimulation, cell cultures were stained with fluorochrome-labelled antibodies against CD4 and CD8 T cell surface markers and intra-cellular cytokines characteristic for effector T cells (i.e. IFNγ and TNFalfa). Representative data from 3 patients are shown in Figure 4. Typically, after in vitro 5T4 antigen-specific stimulation and expansion, between 0.2% and 2% of cultured cells are specific for the vaccine antigen as measured by secretion of cytokines.
• One of the study secondary endpoints is the detection of antigen-specific immune cells in the prostate, as this is where they are expected to exert their effector function and destroy tumour cells expressing the 5T4 antigen. Fresh tissue from the resection specimens or biopsies was processed to expand the immune cells resident in the prostate prior to analysis of their 5T4 specificity. In the majority of cases, high numbers of TILs were expanded from fresh tissue (with a range of 6-80 million immune cells per tissue core/biopsy). Antigen-specific tumour-infiltration lymphocytes could be detected in ~70% of the patients. Flow cytometric analysis of the TILs from one of the patients is shown in Figure 5.
• Immunohistochemical (IHC) analysis of immune cell tumour infiltration following vaccination was performed on fixed tissue. Tissue blocks from each patient were formalin-fixed and paraffin-embedded (FFPE) for IHC analysis of immune cells density in post-vaccination tissue compared to available archival tissue from pre-study diagnostic biopsy. Available paired pre- and post-treatment FFPE tissue sections have been stained for a panel of pre-selected immune cell markers including CD3 and CD8. Batch analysis was performed to quantify the density of immune cells by digital image analysis in order to assess whether the vaccination course leads to trafficking of immune cells to the prostate. The density of cells positive for these markers before and after study treatment was assessed by digital image analysis. In summary, for 12 cases diagnostic pre-treatment biopsies were compared with post-treatment surgical specimens (samples from patients in the surgical arms), and for 13 cases diagnostic pre-treatment biopsies were compared with post-treatment on-study biopsies (patients in active surveillance arms). Cell densities were calculated to the number of marker positive cells per unit area (mm2). The comparison of CD3 and CD8 cell densities in surgical patients revealed a decrease when comparing biopsies (pre-treatment) to resections (post-treatment). Interestingly, when the spatial distribution of marker positive cells was analysed the Ripley’s K index, high values for K-Ripley were observed in resections compared to pre-treatment biopsies. This finding can be explained by potential re-distribution of the immune cells following the study treatment, i.e. the immune cells in prostate tissue normally diffusely distributed in the tissue become activated due to vaccination and as a result traffic to the areas where tumour lesions are to mediate their function of killing tumour cells. This is an interesting observation that needs to be further investigated by studying matched biopsies and resection tissue of untreated control patients.
• PSA reduction in peripheral blood of advanced-stage prostate cancer patients secondary to interventions has been widely used as a surrogate marker of treatment efficacy before In our study, PSA level change was also included as one of the study endpoints. PSA levels in the blood of cancer patients at baseline (before treatment) was compared to PSA levels post vaccination course. Unexpectedly, instead of PSA level drop we have observed a transient PSA increase in some patients following vaccination. If baseline PSA values in all patients in the active surveillance arm are compared to their respective highest PSA values over a period of one year post vaccination, the difference between a baseline and the peak value is not significant (Figure 6). However, there is clearly a trend that can imply that the vaccination induced T cells have trafficked to the prostate and acted upon 5T4-expressing prostate cells leading to the leakage of PSA into the circulation.
WP4: Phase I/II clinical trial (acronym - ADVANCE)
The objective of WP4 was to undertake a phase II safety and efficacy clinical trial with the vectors encoding 5T4 with the trial design informed by results from the phase I clinical trial. In the phase II study, ChAdOx1-MVA 5T4 vaccine was tested in combination with the anti-PD-1 checkpoint blocker Nivolumab in low- or intermediate-risk localized or locally advanced prostate cancer and metastatic castration resistant prostate cancer (mCRPC). The study synopsis is shown in Table 2. Recruitment to the study started in February 2019 and by October 2019 23 out of 24 patients in the metastatic cohort have been enrolled. As of January 2020, all recruited patients either completed the study treatment and were withdrawn due to disease progression. The full data set for this patient cohort will be analysed by Q4 2020. The recruitment to the early stage prostate cancer cohort is ongoing.
• Safety. To date, administrations of ChAdOx1.5T4 and MVA.5T4 vaccines in combination with nivolumab were found to be safe and well tolerated, in agreement with our previous studies with these vectors.
• PSA response rate. Serum PSA levels were measured through the study a t each clinic visit. 5 of 19 patients demonstrated 50% or greater PSA decrease from baseline (Fig.7). This response rate is significantly higher than the response rate reported in the KEYNOTE-199 study (Antonarakis et al., 2019) of anti-PD-1 antibody alone in advanced metastatic prostate cancer. A maximal decrease in PSA concentration compared to baseline in each of 5 responders is shown in Fig.8.
• CTC enumeration. The quantification of circulating tumour cells (CTCs), which is one of the study exploratory endpoints, has been performed. Encouragingly, in some of the patients a decrease in serum PSA correlated with decrease in a number of CTCs (Fig.9).
• T cell immune responses. The primary immunogenicity measure in the study is the IFN-γ ELISpot response of PBMCs stimulated with overlapping peptide pools covering the entire length of 5T4 proteins. Responses were measured before vaccination and at different time points post-vaccination. Out of 23 patients who completed the vaccination course, 2 patients mounted 5T4-specific T cell response that was detected at several timepoints during the study (Fig. 10). The immune response rate to the vaccine is lower that the one observed in the VANCE study, but this is not surprising given the patient baseline characteristics such advanced metastatic disease and multiple potentially immunosuppressive treatment modalities prior to enrolment to the study.
WP5 Immunogenicity studies
The aim of WP5 was to carry out a detailed analysis of the immune responses both in the blood generated by the 5T4-targeting prostate cancer vaccine deployed in the phase I and II clinical trials. Dissection of CD8 and CD4 T cell immune responses was performed by in vitro expansion and flow cytometry. The main achievements are as follows:
Test and distribution of protocol for tumour-infiltration lymphocytes (TILs) expansion from prostate cancer biopsies
• Phenotype of TILs derived from Prostatic carcinoma to a non-cancer control was completed.
• The ex-vivo TILs phenotype was investigated and the phenotypic changes evaluated after in vitro expansion.
• We observed high level of similarity between expanded TILs from PC and BPH samples. T cells are the predominant population in both pre-REP and REP expanded condition and NK cells are completely lost only after REP expansion.
• In BPH samples the expansion with pre-REP and REP protocols generated a predominant T cell population that progressively upregulates activation/exhaustion and cytotoxic markers in comparison to TILs directly analyzed from digested disuse ex-vivo.
Development of a flow cytometry panel to monitor regulatory T (T-reg) cells in peripheral blood
• T-reg flow cytometry panel with 14 markers was developed
Mapping of T cell responses in vaccinated patents
• In vitro culture protocol and mapping of CD4 and CD8 T cell responses in 25 cancer patients vaccinated for 5T4 protein.
• Several CD4 and CD8 T cell epitopes were identified. Interestingly, many of those were never published previously.
Development and harmonization of protocol for testing immunogenicity and T cell phenotype longitudinally in prostate cancer patients enrolled into the study.
• Ex-vivo and In-vitro culture, stimulation and intracellular staining protocols were developed and harmonized between UNIL and OX (analysis performed in 3 healthy donors).
• Harmonization resulted successful. Test of the developed flow cytometry panel on PBMC of 3 healthy donors gave comparable results.
An extended flow cytometry panel was validated for use in the phase II clinical trial for detailed phenotypic and functional characteristics of induced T cell responses (Table 3).
WP6 Correlates of Immunity and efficacy
The objective of WP6 is to build on the phase I and phase I/II clinical trials. In phase I trial, VANCE, Blood samples recovered from patients recruited to the trial have been analysed for antigen-specific responses and the results used to investigate whether there are any potential associations between immune responses and clinical benefit. Analysis of correlates of immunogenicity and efficacy is dependent on the provision of data from WP5 and suitable markers of clinical efficacy. Given that the patients treated in the phase I clinical trial were at a relatively early-stage in their disease course, no recognised markers of clinical benefit (e.g. response rate, survival etc) were available within the timeframe of this work package. As such, the key analytic testing looked to identify predictors of immunogenicity.
An analysis of potential correlates with immunogenicity was undertaken using antibody response data. Previously, a pre-treatment biomarker was identified which predicted 5T4 specific immune responses induced in patients following vaccination with MVA-5T4 (TroVax; Harrop et al., Cancer Immunology Immunotherapy 2011). The same pre-treatment biomarker (consisting of baseline levels of haemoglobin, haematocrit and 5T4 antibody level) was used to determine if this biomarker was associated with the fold-increase in 5T4 or MVA antibodies induced following vaccination with ChAdOX.5T4 and MVA-5T4. These exploratory analyses suggested a significant relationship between the pre-treatment biomarker and the 5T4, but not MVA, antibody responses induced post vaccination. An assessment of correlation coefficient between the pre-treatment biomarker and 5T4- and MVA- specific antibody responses at weeks 1, 2, 4 and 8 are shown in Table 4. As can be seen, there is no significant correlation between the pre-treatment biomarker and antibody responses detected at week 1; this is not surprising as it normally takes >1 vaccination before antibody responses against the target antigen are detected. However, at weeks 2, 4 and 8 post-vaccination there are strong trends (weeks 2 and 8) or significant correlations (week 4) between the pre-treatment biomarker and the 5T4-specific antibody response, but not the MVA-specific antibody response. These data replicate what have been reported previously i.e. significant correlations between pre-treatment biomarker and antibody responses against the target antigen (5T4), but not against the viral vector (MVA).
These data are of potential significance because 8 clinical trials using MVA-5T4 have demonstrated a link between the magnitude of the 5T4-specific immune response and clinical benefit (tumour shrinkage or overall survival). The ability to select the patients who are most likely to benefit from this class of therapy prior to treating them would be a huge advantage. Further confirmatory prospective studies would be required to develop a test which predicted treatment benefit.
For correlation between immune response and clinical efficacy in the phase I/II study, ADVANCE, blood samples from 23 recruited patients were processed to isolated PBMCs for IFN-γ ELISPOT assays and to measure the number of antigen-specific T cells induced by vaccination. Patients’ blood and serum samples were also collected for CTC enumeration and PSA concentration.
Out of 23 patients, 2 patients mounted vaccine-specific T cell immune response and 5 patients had a PSA decrease from baseline of 50% or more and 5 patients had a CTC decrease from baseline. One of 2 patients with immunological response had also a decrease in the number of CTC but no PSA reduction. So there clearly was no detectable correlation between our primary immunogenicity read-out (ELISpot) and the available and most information efficacy read-out (PSA reduction). However, this correlation analysis is very underpowered because of the small number of responders in the ELISpot assay. This low frequency contrast with the findings in our phase I VANCE trial where the same vaccines produced a >65% response rate. The difference may reflect a difference in the general immune competence of early and late-stage prostate cancer patients. A correlation analysis was also performed on the patients who had either PSA or CTC decrease or both and although there is as yet no significant difference on small numbers we will extend this analysis once CTC results are available from more subjects (Fig. 11).
WP7 Pre-clinical antigen evaluation – immunogenicity
WP7 comprises the characterisation of novel tumour-associated antigens and the assessment of their immunogenicity following expression from the viral vectors. The original focus was on the characterization of three antigens, namely CRISP3, CNPY2 and DPY19L3 discovered by Externautics partner (EXT). Overall, the study reinforced the association of these antigens with prostate cancer and provided a first indication of the antigen expression in early PCa and pre-cancerous stages. The expression of the three antigens was confirmed in different human prostate cell lines representing different PCa subtypes. Moreover, the murine counterparts of the three antigens were found expressed in murine prostate cell lines, including TRAMP cell lines, thus justifying the use of the TRAMP mouse model for efficacy studies. Finally, the three protein antigens showed an overall marginal expression in human normal tissues, thus allowing to hypothesize an acceptable toxicity profile of vaccine based on these antigens. Results specifically related to each antigen are summarized below:
• CNPY2 (Canopy FGF Signalling Regulator 2)
Human CNPY2 protein is annotated in two isoforms of 182 and 84 amino acids. The 182 aa isoform is predicted to be associated with the endoplasmic reticulum while the 84 aa isoform is secreted or intracellular.
Expression in PCa: IHC analysis of 100 prostate cancer samples and 47 PIN cases showed that CNPY2 was detected in all tested samples with the highest expression in 17% of PCa and 17% of tested PIN samples.
Expression in normal human tissue: IHC analysis of 33 normal human tissues showed weak cytoplasmic staining only in stomach, colon and chorionic villi and negligible reactivity in the other 31 tissues. Western blot (WB) and flow cytometry (FACS) analyses of PBMC purified from healthy individuals showed moderate CNPY2 expression, which was confined to the intracellular compartments.
Endogenous expression in prostate cell lines: CNPY2 was clearly expressed in all tested human prostate cell lines (R22V1, VCaP, PC3, PNT1, LNCaP, DU145) and in the murine cell lines TRAMP-C1 and TRAMP-C2, as judged by Q-RT-PCR and Western blot using anti-CNPY2 antibody. Western blot analysis allowed detection of a protein band of 20 kDa, compatible with the 182 amino acid isoform. CNPY2 was also well expressed in murine prostate TRAMP cell lines. FACS analysis showed that the protein is confined to the intracellular compartment.
• DPY19L3 (probable C-mannosyltransferase)
DPY19L3 is the least characterized protein of the group. The protein has been annotated in at least 3 overlapping isoforms with predicted MW of 83, 63 and 12 kDa, respectively. Only one isoform is predicted for the murine DPY19L3 orthologues, showing 88% amino acid identity with the 716 amino acid human isoform and a predicted MW of 83 kDa.
Expression in PCa and PIN: IHC analysis of 100 PCa showed that the protein was detected in 44.4% of PCa, among which 6.1% showed the highest expression. Analysis of 47 PIN cases showed that the protein is detected in 74% of the samples, 13% of which showed the highest expression. Staining was generally cytoplasmic and in some samples a membranous staining was visible. Staining of normal prostate was negligible.
Expression in normal human tissue: IHC analysis of the 33 normal tissues by IHC did not show any significant staining. In addition, DPY19L3 expression was not detected in human PBMC or in neutrophils by Western blot or by FACS
Endogenous expression in prostate cell lines: DPY19L3 expression analysis in human prostate cell lines by Q-RT-PCR revealed the presence of the three annotated DPY19L3 transcripts. Western blot revealed two major protein bands of approximately 90 and 60 kDa, compatible with the two longer DPY19L3 isoforms. FACS analysis showed a moderate surface staining of VCaP and R22V1 cells suggesting that it is at least partially surface exposed in these cells. DPY19L3 expression in murine prostate TRAMP cell lines was confirmed at transcript level. In these cells, Western blot revealed a unique band of approximately 60 kDa that might result from proteolytic processing of the annotated 83kDa protein.
• CRISP3 (Cystein-rich secretory protein 3)
Human CRISP3 protein is annotated with at least two variants of 258 and 268 amino acids and is predicted to be secreted. In prostate carcinoma, it has been reported as associated with a subset of prostate carcinoma carrying the TMPRSS2-ERG fusion, pT3 disease stage. Two potential CRISP3 orthologues are annotated in the mouse, namely CRISP1 and CRISP3, sharing 54% amino acid identity with the human protein, whose role in cancer has not been previously described.
Expression in PCa and PIN: IHC analysis of 100 PCa cases showed that the protein was detected in 29% of PCa samples, among which 19% showed the highest expression. Analysis of 47 PIN cases showed that the protein is detected in 23% of PIN, 13% of which showed the highest expression. Staining was generally cytoplasmic, except for a few samples that also showed a membranous staining.
Expression in normal human tissue: CRISP3 was not detected in human PBMC, as assessed by WB and FACS. Instead, in agreement with published data, the protein was detected in the neutrophil population. IHC analysis of the 35 normal human tissues did not give any relevant expression of the protein.
Endogenous expression in prostate cell lines: among the different human cell lines currently tested CRISP3 was exclusively detected in the VCaP cell line, having the TMPRSS2-ERG fusion. In these cells WB analysis with the anti-CRISP3 antibody revealed the presence of a band of expected size (approximately 30 kDa). The antibody also bound to the surface of VCaP cells in FACS staining assay, indicating that CRISP3, besides being secreted, is at least partially exposed on the cell surface. Concerning CRISP3 expression in the murine TRAMP cells, transcription profile analysis by Q-RT-PCR of the tumours excised from TRAMP mice showed that mCRISP1 was detected in two of the three tested biopsies whereas mCRISP3 cDNA was not detected, indicating that mCRISP1 could be the protein form expressed in murine prostate cancer.
Construction of recombinant viral vectors ChAdOx1 and MVA expressing novel prostate-associated antigens, CRISP1, CNPY2, DPY19L3, and comparative immunogenicity of these novel antigens and well defined prostate-associated antigens 5T4 and STEAP1 that also represent potential antigenic targets for a prostate cancer vaccine.
Recombinant viral vectors ChAdOx1 and MVA, expressing the novel prostate associated antigens CNPY2, DPY19L3 and CRISP3 were designed and generated for immunogenicity testing in mice.
These vectors delivered in a heterologous ChAdOx1 prime – MVA boost vaccination regime have been tested for immunogenicity in C57Bl/6 mouse strain. The cellular immune responses following immunisation have been assessed by an ex vivo IFNg ELISPOT assay against pools of 15-mer peptides overlapping by 10 amino acids spanning the full length of the proteins encoded by the vector transgenes. As a result, there have been no detectable T cell reactivity against CRISP1 and CNPY2 antigens, however, there was potentially a weak response induced against the DPY19L3. These results are in striking contrast to the immunogenicity of another murine prostate-specific antigen, STEAP1, induced by ChAdOx1-MVA immunisation (Cappuccini et al 2016). Although disappointing, these results were not unexpected as the central and/or peripheral immunological tolerance usually prevents an induction of immune responses against self-antigens. We have demonstrated that exceptionally strong T cell reactivity against STEAP1 is likely to be due to the lack of its expression in the thymus, although the mRNA transcripts corresponding to the m5T4, CRISP1, CNPY2 and DPY19L13 are detectable in the murine thymus by reverse transcription PCR.
The novel prostate-associated antigens identified by EXT have not been taken further into efficacy testing due to paucity of the vaccine-induced immune responses against these antigens. Instead the Steering Committee decided to proceed with the pre-clinical testing of the panel of human antigens, PSA (prostate-specific antigen), STEAP1 (six transmembrane antigen of the prostate 1) and ERG (ETS transcription factor) as a polyvalent vaccine with the aim of taking them into the clinic in the future. To this end, firstly the single antigen vectors expressing these antigens in native form or fusions to the invariant chain have been made for immunogenicity testing. With further clinical development in mind, fusion constructs have been made with the transmembrane domain of the shark invariant chain for the following reasons. Firstly, the shark and human invariant chain share only 30% homology so the potential risk of autoimmunity will be greatly reduced. The 25 amino acid long transmembrane domain is the minimal sequence that is required to maintain the enhancer effect, so the risk of autoimmunity will be reduced even further. Strong immune responses against PSA and STEAP1 have been induced in a mouse model but tolerance to ERG antigen could not be broken (Figure 12) therefore this antigen has been excluded from the multi-antigen construct developed on WP8.
WP8 Pre-clinical antigen evaluation - efficacy
Objectives of WP8 comprise the efficacy evaluation of 5T4 encoding vaccines compared to the vectors expressing novel antigens, assessment of immune correlates of efficacy in mouse models and generation of new vectors encoding the most promising vaccine targets to be available for early stage clinical development.
According to the original project plan, the novel EXT antigens identified and described in WP7 should have been taken into further pre-clinical development in order to create the vectors available for early stage clinical development at the end of the project. As the antigens CRISP-1, CNPY-2 and DPY19L3 appeared to be of low immunogenicity or non-immunogenic at all, as an alternative, we have chosen the three well-defined prostate-associated antigens, prostate-specific antigen (PSA), six transmembrane epithelial antigen of the prostate – 1 (STEAP-1), and 5T4 antigen for further pre-clinical and clinical development and expressed a string of these antigens from the same vector thus creating multi-antigen ChAdOx1 and MVA vectors. Of note, the ERG antigen originally planned as a component of the multi-antigen vaccine, was replaced by 5T4 as the phase I clinical trial (WP3) results clearly indicated that both CD8 and CD4 T cell responses against 5T4 can be induced in prostate cancer patients.
PSA
Human PSA, a chymotrypsin-like serine protease, has a highly restricted tissue distribution and is expressed in the epithelial cells of the prostate gland, the same cell type from which most prostate tumours arise. Indeed, PSA is widely used as a serum marker for prostate cancer. Its expression is regulated by androgen, and itis present at extremely low levels in the circulation of adult men. Most prostate tumours, even the poorly differentiated ones, continue to express PSA. This cell type-specific expression of PSA makes it a potential target antigen for antitumor CTL. In fact, PSA is a target antigen of one of the two most clinically advanced prostate cancer vaccines - ProstVac.
STEAP1
The six-transmembrane epithelial antigen of prostate protein was identified in advanced prostate cancer. STEAP1 is highly expressed in human prostate cancer and is up-regulated in various cancers, including lung, bladder, colon, ovarian, and Ewing cancers. Immunohistochemical analysis of clinical specimens demonstrates significant STEAP1 expression at the cell–cell junctions of the secretory epithelium of prostate and prostate cancer cells. Little to no staining was detected at the plasma membranes of normal non-prostate human tissues, except for bladder tissue, which expressed low levels of STEAP1 at the cell membrane. Its cell-surface localization, together with its six-transmembrane topology, suggests STEAP1 may function as a channel/transporter protein in cell–cell junctions. Given its increased expression in cancer tissues, STEAP1 could be a promising target for T-cell based or antibody immunotherapy. In our previous experiments, we have expressed murine STEAP1 from ChAdOx1 and MVA viral vectors and investigated STEAP1-specific T cell response in mice vaccinated with these vaccines. Despite being a self-antigen, STEAP1 appeared to be highly immunogenic in mice, most likely because of absence of central tolerance to this antigen. STEAP1 has been evaluated in early stage prostate cancer clinical trials and delivered in the form of RNA as part of a multi-antigen vaccine with encouraging preliminary data.
Construction of ChAd and MVA viral vectors.
Both mono-cistronic ChAdOx1 and MVA vectors encoding individual antigens and multi-antigen vectors expressing all three antigens were constructed for immunogenicity testing in mice. Poly-cistronic vectors were constructed to express a string of three antigens from the least immunogenic to the most immunogenic antigen, i.e. with STEAP1 cDNA downstream of the promoter followed by 5T4 and PSA cDNA. Flexible polypeptide linkers composed of glycine and proline (GGG-P-GGG) were inserted between cDNAs to create a fusion protein.
Immunogenicity testing of mono- and poly- cistronic viral vectors in a mouse model.
To assess the immunogenicity of mono- and poly- cistronic viral vectors encoding human antigens PSA, STEAP1 and 5T4 in a mouse model, outbred mice (CD1 strain) were randomised into 5 groups. Mice were primed with ChAdOx1 virus encoding each individual antigen or with the vector expressing a string of antigens and were boosted with the MVA vectors accordingly. The induction of T cell immune responses against the vaccine transgenes was tested by ex vivo IFN-γ Elispot assay after priming and boosting immunisations. As evident from the Figure 13, T cell responses of relatively high magnitude are detectable in blood after a single immunisation with a polcistronic ChAdOx1 vector against all three antigens . However, surprisingly mono-cistronic ChAdOx1 vectors appeared to be less immunogenic in the groups having received either an individual vaccine or a combination of three mono-cistronic ChAdOx1 vectors. The MVA mono-cistronic vaccines significantly increased the magnitude of the immune responses (Figure 14). The increase in antigen-specific immune responses was also observed in mice that received a polycistronic MVA vaccine.
Efficacy testing of mono- and poly-cistronic viral vectors in a mouse model.
Having demonstrated that the novel polycistronic viral vectors encoding a string of these 3 human prostate cancer associated antigens, are immunogenic in mice, next we moved to testing its tumour protective efficacy. To this end, we required a murine tumour cell line that express these three human antigens to perform tumour challenge experiments. The experimental design is to inoculate mice subcutaneously with a syngeneic tumour cell line that expressed human PSA, STEAP1 and 5T4 and, after the tumours are established, to treat the mice with the novel poly-cistronic vaccine in order to test whether tumour growth is delayed compared to mice in the control group that are challenged with the tumour and left untreated. As such murine cell line does not exist, we had to create it. We have chosen the method of lipofectamine transfection using plasmid DNA encoding all three antigens. The work of making stably transfected B16 and CT26 murine tumour cell lines (C57Bl/6 mouse strain and BALB/c mouse strain respectively) is ongoing.
Work Package 9: Project management
WP9 ensured the proper overall management of the project in order to strengthen and support the participants to achieve the objectives, complete the milestones in time and deliver the deliverables.
The management structure of the project ensured that:
• the consortium’s contractual duties were carried out
• advise and guidance were provided to the participants to comply with the EU regulations and their contractual and legal requirements.
• an effective communication infrastructure was set-up and an integrative process within the consortium was fostered.
• knowledge produced within the project was disseminated to the relevant target groups through publications and training
• both phase I and phase II clinical trials were conducted in full compliance with Good Clinical Practice
Potential Impact:
Potential impact and main dissemination activities and exploitation results.
Prostate cancer is the second most common cancer in North American and European men and the second leading cause of male cancer related death after lung cancer. Only in the UK Over 47,000 men are diagnosed with prostate cancer every year – that's 129 men every day, and every 45 minutes one man dies from prostate cancer – that's more than 11,000 men every year. Although the number of therapeutic options for patients with advanced stage prostate cancer has increased recently, the treatments remain largely palliative and often have severe side-effects.
Immunotherapy has been evolving as a potential new treatment modality offering the promise of long-term treatment effect with a minimal side-effect profile. Despite, some early failures of immunotherapy products in phase III trials, a key milestone was achieved in 2010 when the FDA approved Provenge (Sipuleucel-T) as the first therapeutic cancer vaccine for the treatment of patients with advanced prostate cancer. This individualised treatment costs over $90,000 per patient and provides a modest survival benefit of 4 months. Taking into account the cost-benefit ratio and lack of a clear mechanistic explana¬tion of the clinical benefit, a less expensive more effective standard immunisation approach, which avoids the multiple rounds of leukapheresis associated with this treatment, would be highly desirable. A major impact of this proposal is expected to be a new, significantly less expensive and more efficacious prostate cancer vaccine available as a monotherapy and as a combinatorial treatment with a PD-1 blockade for late stage clinical development within the next few years.
A further potential impact of this program arises from the identification and testing of a new, more rapid, clinical efficacy protocol for prostate immunotherapies. We have assess a 4 week and 12 week intervention strategy and utilised histological and biochemical measures of efficacy within this time period, along with longer term monitoring of biochemical and clinical impact. If markers of efficacy can be detected within the 12 week window, this would considerably accelerate the testing of new vaccination approaches.
Class Leading Immunological Potency
By definition, an immunotherapy product must act through the induction of an efficacious immune response. In cancer, the target molecule is usually a tumour associated antigen (TAA) which is derived from “self”; therefore the induction of an immune response against the TAA requires the breaking of tolerance. As such, any vaccination strategy needs to be sufficiently potent to ensure that tolerance to the self-antigen is broken. Various strategies have been utilized to achieve this goal ,and to date, results with various antigen/delivery methods have been mixed with only a few reporting strong immune responses against the tumour antigen in the majority of treated patients. Where immune responses have been detected, an association between the magnitude of the response and indicators of clinical benefit has been frequently reported. A key goal of this program was to induce strong cytotoxic T cell a well as antibody responses against the target tumour antigen (5T4) in the great majority of treated patients. The aim of this program was to combine the experience derived from the use of MVA.5T4 in a homologous prime-boost regimen with the experience derived from use of a heterologous prime-boost regimen in which MVA as a booster has been combined with simian adenovirus as a potent priming agent. This novel heterologous prime-boost regimen has been tested in over 40 trials in infectious disease settings and shown to be highly immunogenic across all studies. Indeed, we have demonstrated that the combination of simian adenovirus and MVA in a heterologous prime-boost regimen consistently generates the strongest recorded sustained CD8+ T cell immune responses in humans of any subunit vaccine approach previously deployed. The immunological basis of the ChAd-MVA regime potency is quite well understood. In humans, non-human primates and mice adenoviral vectors outperform all other technologies at priming a potent CD8 T cell response with a single immunisation, likely due to a preferred combination of innate receptor ligands. MVA is a particularly good boosting agent, but only a modest priming vector. This is due to the 180 genes in MVA that compete with the transgene when priming new response; however, on boosting the MVA-induced response uses natural immunodominance mechanisms to amplify the primed immune response to the transgene product which outcompete new primary responses. The consistent consequence is a CD8 T cell response amplified 5 -10 fold compared to the ChAd (or MVA vector) used alone. As expected, the delivery of 5T4 using 2 different vectors (simian adenovirus and MVA) in a heterologous prime boost regimen resulted in substantially enhanced immune responses against the target tumour antigen in terms of frequency and magnitude of responses. In the context of the mediocre immune responses previously reported for the majority of cancer vaccines, it is believed that the results from this study will provide a significant step forward.
Safety and Tolerability
The majority of approved cancer therapies currently in use have a significant side-effect profile which can usually be managed, but can have a significant impact on a patients’ quality of life. The therapeutic vaccine approach proposed here has demonstrated its safety and tolerability, compatible with prophylactic vaccine standards, in multiple clinical trials. As such, the patient compliance rate was very high and there were no toxicity issues.
As more treatments become available, clinicians will increasingly look to combinatorial approaches to enhance clinical benefit to the patient. Indeed, we recognise that the optimal therapies for most cancers will often comprise multimodality treatments e.g. chemotherapy, radiotherapy and various types of immunotherapy. However, our aim here was to improve a leading form of vaccine therapy, prior to assessment of combination strategies. As we expected, the safety profile of the conducted trials was excellent, and therefore, the vaccination approach has the potential to be used in combination with many different standard therapies without impacting on the side-effect profile or diminishing the efficacy of the standard treatment. Such flexibility broadens the potential therapeutic window of a vaccine approach beyond niche indications or settings where no standard treatments are available.
Cost of Goods
Sipeuleucel-T is currently the only immunotherapy approved for the treatment of prostate cancer. The treatment is patient specific and requires leukapheresis of the patient in order to provide the active treatment modality. Such a treatment is both cumbersome and expensive ($90,000 per patient). The vaccination strategy proposed in this application represents an “off the shelf” approach (i.e. not patient specific) and can be manufactured at low cost (with costs of goods of each of the two components at scale of less than a euro). For example, ChAd vectors have been manufactured for clinical use on a variety of cell lines, the widely available HEK293 cells, the new Procell cell line that allows tetracycline repression of the transgene, and the PER C6 cell line from Crucell. There are also three duck and quail-derived cell lines available for the manufactured of MVA that lead to low costs of goods and avoid the need for chick embryo fibroblast for large scale biomanufacture. Therefore, this product configuration has potential for widespread use and low health care system costs, in contrast to most other novel cancer interventions.
New Targets for Therapeutic Intervention
The identification of tumour antigens which are suitable for immune targeted therapeutic strategies remains a challenge. The preclinical part of the proposed work assessed novel prostate specific antigens as potential candidates for therapeutic intervention, aiming for a multi-antigen second generation vaccine. Combining several partially protective antigens in a vaccine will provide synergistic efficacy. Another advantage of targeting several antigens is the increased breadth of the immune response and the reduced likelihood of both escape mutations and selection of more aggressive tumour phenotypes. As anticipated, pre-clinical evaluation of the novel vaccine candidates translated into a new multi-antigen prostate cancer vaccine available for early stage clinical development by the end of this project.
Additionally, while this proposal aims to test the vaccine in prostate cancer patients, the strategy is broadly applicable to the design and development of vaccines against other types of cancer since 5T4 and some of the antigens that we plan to test are present in other tumour types.
A Platform for Therapeutic Vaccination
This project opened new research avenues for therapeutic cancer vaccination in general as well as prostate cancer immunotherapy specifically. If good immunogenicity and significant efficacy can be achieved for prostate cancer in this programme there are many opportunities for applying the ChAd-MVA vectored prime-boost approach to many other cancers. Furthermore, success with ChAd-MVA immunisation in prostate cancer immunotherapy would open the way to assessment on the same vaccines in several other cancers. For example, 5T4 is known to be well expressed on colorectal, breast and renal tumours so that the same vectors could also be assessed in these cancers.
Impact on the Vaccine Field as a Whole
Some of the potential impacts of this programme on the vaccine field as a whole are worth noting. The interest in the ability of vectored vaccines to induce very potent T cell responses is of potentially even greater impact in the area of therapeutic vaccination where CD8 T cell response are known to play a particularly important role, and antibody-based vaccines are less likely to be effective. This project provided the first key test of concept of this approach in cancer immunotherapy.
The dissemination of results
Main dissemination activities of the consortium included scientific publications in academic journals, presentations at the national and international conferences and workshops, press releases to the media and public awareness campaigns.
The IMPROVE partners participated in local, national and international level workshops and meetings that are relevant in all areas relating to the project. Results from the project were presented at scientific meetings for the benefit of the wider scientific community. Regular seminars organized by the partners in their own institutions were implemented.
Engagement of patients and public into discussion of the clinical trial design and study related documents constituted important part of dissemination activities. The patients were made aware of the project objectives and ongoing clinical trials through the participant information sheets distributed through the clinical teams participating in the trials
Experience has shown that when patients were engaged right from the onset in the research planning process, they were more committed to applying the research findings in real life and policy settings.
The project website played an important role in targeting the general public.
Any dissemination activities and publications in the project, including the project website specified that the project had received Community research funding and displayed the European emblem.
List of Websites:
http://www.project-improve.eu/home(opens in new window)
Project co-ordinator: Professor Adrian Hill, adrian.hill@ndm.ox.ac.uk