Periodic Reporting for period 4 - Inno4Vac (Innovations to accelerate vaccine development and manufacture)
Periodo di rendicontazione: 2024-09-01 al 2025-08-31
Vaccines are a huge public health achievement, saving an estimated 2.5 million lives every year and protecting millions more from illness and disability. However, developing new vaccines is extremely time-consuming and financially risky, averaging 10+ years and more than €800 million to bring a vaccine to the market.
Recently, researchers from academia, pharma and biotech companies have made huge strides in the fields of immunology, big data, and artificial intelligence. These advances could potentially speed up new vaccine development and make the entire development process more efficient.
Inno4Vac aims to harness these advances: incorporating them into the vaccine industry pipeline. The project brings together experts in clinical research, immunology, microbiology, systems biology, mathematical models, and regulatory affairs.
The project comprises four subtopics (STs). Two of these focus on in silico tools. One builds models for immune response induction by applying machine learning algorithms to lab-based data. The models and predictive tools developed are combined into an open-access, cloud-based platform to predict post-vaccination immune response characteristics. The second in silico area is developing a modular computational platform to model the manufacture and stability testing of vaccines.
The other two STs focus on lab-based and clinical tools. One is developing new controlled human infection models (CHIMs) for influenza, Respiratory Syncytial Virus (RSV) and Clostridioides difficile, with the aim to provide better ways to study vaccine efficacy at early stage of clinical developments. The other ST aims to deliver in vitro complex model mimicking human mucosa to offer an innovative and reliable view of the immune protection a vaccine could offer.
The consortium is developing strategies and roadmaps to ensure the various models meet the needs of medicines regulators for integration into vaccine development processes.
Ultimately, the Inno4Vac models should help to make vaccine development both faster and more efficient.
Recently, researchers from academia, pharma and biotech companies have made huge strides in the fields of immunology, big data, and artificial intelligence. These advances could potentially speed up new vaccine development and make the entire development process more efficient.
Inno4Vac aims to harness these advances: incorporating them into the vaccine industry pipeline. The project brings together experts in clinical research, immunology, microbiology, systems biology, mathematical models, and regulatory affairs.
The project comprises four subtopics (STs). Two of these focus on in silico tools. One builds models for immune response induction by applying machine learning algorithms to lab-based data. The models and predictive tools developed are combined into an open-access, cloud-based platform to predict post-vaccination immune response characteristics. The second in silico area is developing a modular computational platform to model the manufacture and stability testing of vaccines.
The other two STs focus on lab-based and clinical tools. One is developing new controlled human infection models (CHIMs) for influenza, Respiratory Syncytial Virus (RSV) and Clostridioides difficile, with the aim to provide better ways to study vaccine efficacy at early stage of clinical developments. The other ST aims to deliver in vitro complex model mimicking human mucosa to offer an innovative and reliable view of the immune protection a vaccine could offer.
The consortium is developing strategies and roadmaps to ensure the various models meet the needs of medicines regulators for integration into vaccine development processes.
Ultimately, the Inno4Vac models should help to make vaccine development both faster and more efficient.
After 4 years, the Inno4Vac subtopics have progressed as follows:
ST1 has developed pipelines for bulk and antigen specific sequencing of BCR/TCR repertoires pre- and post-vaccination of healthy volunteers. The various data are fed into development and refinement of in silico tools, modelling and predicting post-vaccination immune formation. Currently, the partners develop in silico strategies to quantify the heterogeneous baseline of human adaptive immunity, integrated specific antigens into in silico germinal center modelling, and developed predictive algorithms based on machine learning for identification of B- and T-cell epitopes. Further data from the BCR/TCR in vivo experiment will be needed to develop a model predicting TCR-pMHCI interactions, and refine predictions of B cell epitopes. The current refined models and tools were also added to a first blueprint model of the in silico predictive platform, interconnecting different predictive tools and simulations.
ST2 advanced the establishment of CHIMs for Influenza, RSV and C. difficile. The selection and characterisation of the H3N2 influenza strain was completed, and GMP production is underway, with clinical activities slated to start in the 2nd half of 2026. The RSV-B and C. difficile challenge agents were manufactured following GMP guidelines, in sufficient quantities for clinical activities. The dose-escalation clinical study for RSV-B started in Mid-2025, with good preliminary results for safety, infection rate and symptomatology, and trial completion is expected in late 2025. The C. difficile challenge agent has been administered to the first study cohort and was shown to be safe allowing to enrol study participants into the next cohort. Additionally, several stakeholder engagement activities accompanied the clinical activities to position CHIMs in the evolving regulatory, ethical, and scientific landscape.
ST3 developed next-generation human in vitro 3D models for gastro-intestinal, respiratory and urovaginal mucosae that include immune system components. 14 models were designed, tested and optimised under a variety of conditions in combination with selected pathogens. In the recent reporting period a detailed Stage Gate 2 evaluation process was completed. While no models advanced from Phase 2 (“Exploration”) to final use and validation, significant positive progress was made and 6 models were prioritised for further development based on a critical review process. Continuing efforts focus on scientific and logistical alignment with ST2 to ensure the models build on the CHIM trials and optimally utilise participant samples. ST3 also organised several workshops with external stakeholders to discuss and plan implementation of in vitro models in the vaccine development and regulatory approval processes.
ST4 developed several in silico models for biomanufacturing, bioreactor computational fluid dynamics (CFD) models, faster compartment models (CM), kinetic metabolic models (KM) of host cell metabolism, centrifugation, chromatography, filtration, crystallization, and control models, as well as models on forecasting vaccine product stability. Several models were developed based on industry partner data. For upstream processing (USP), the CM, the KM, and the combination of these (hybrid model) are being implemented in the online software platform CADET-Hub. For downstream processing (DSP), capture chromatography, dead-end filtration, and crystallization models are implemented in CADET-Hub, and the centrifugation model is currently being implemented. In vaccine stability forecasting, design of experiments and Bayesian models were deployed and are being implemented in CADET-Hub. Digital twins for model-predictive control were released for USP and DSP. All approaches were presented to regulatory agencies at a workshop, allowing ample time for discussion with developers. Several papers were published and the overall open-access online platform is currently being published (manuscript submitted, web-pages being prototyped).
The subtopics initiated and continued discussions with regulatory bodies about the project stage at which regulatory considerations were relevant (ST1), a strategy for integration of CHIMs into pharmaceutical vaccine development (ST2), acceptance of next-generation human in vitro 3D models in de-risking vaccine development (ST3) and acceptance of in silico models into CMC dossier (ST4).
ST1 has developed pipelines for bulk and antigen specific sequencing of BCR/TCR repertoires pre- and post-vaccination of healthy volunteers. The various data are fed into development and refinement of in silico tools, modelling and predicting post-vaccination immune formation. Currently, the partners develop in silico strategies to quantify the heterogeneous baseline of human adaptive immunity, integrated specific antigens into in silico germinal center modelling, and developed predictive algorithms based on machine learning for identification of B- and T-cell epitopes. Further data from the BCR/TCR in vivo experiment will be needed to develop a model predicting TCR-pMHCI interactions, and refine predictions of B cell epitopes. The current refined models and tools were also added to a first blueprint model of the in silico predictive platform, interconnecting different predictive tools and simulations.
ST2 advanced the establishment of CHIMs for Influenza, RSV and C. difficile. The selection and characterisation of the H3N2 influenza strain was completed, and GMP production is underway, with clinical activities slated to start in the 2nd half of 2026. The RSV-B and C. difficile challenge agents were manufactured following GMP guidelines, in sufficient quantities for clinical activities. The dose-escalation clinical study for RSV-B started in Mid-2025, with good preliminary results for safety, infection rate and symptomatology, and trial completion is expected in late 2025. The C. difficile challenge agent has been administered to the first study cohort and was shown to be safe allowing to enrol study participants into the next cohort. Additionally, several stakeholder engagement activities accompanied the clinical activities to position CHIMs in the evolving regulatory, ethical, and scientific landscape.
ST3 developed next-generation human in vitro 3D models for gastro-intestinal, respiratory and urovaginal mucosae that include immune system components. 14 models were designed, tested and optimised under a variety of conditions in combination with selected pathogens. In the recent reporting period a detailed Stage Gate 2 evaluation process was completed. While no models advanced from Phase 2 (“Exploration”) to final use and validation, significant positive progress was made and 6 models were prioritised for further development based on a critical review process. Continuing efforts focus on scientific and logistical alignment with ST2 to ensure the models build on the CHIM trials and optimally utilise participant samples. ST3 also organised several workshops with external stakeholders to discuss and plan implementation of in vitro models in the vaccine development and regulatory approval processes.
ST4 developed several in silico models for biomanufacturing, bioreactor computational fluid dynamics (CFD) models, faster compartment models (CM), kinetic metabolic models (KM) of host cell metabolism, centrifugation, chromatography, filtration, crystallization, and control models, as well as models on forecasting vaccine product stability. Several models were developed based on industry partner data. For upstream processing (USP), the CM, the KM, and the combination of these (hybrid model) are being implemented in the online software platform CADET-Hub. For downstream processing (DSP), capture chromatography, dead-end filtration, and crystallization models are implemented in CADET-Hub, and the centrifugation model is currently being implemented. In vaccine stability forecasting, design of experiments and Bayesian models were deployed and are being implemented in CADET-Hub. Digital twins for model-predictive control were released for USP and DSP. All approaches were presented to regulatory agencies at a workshop, allowing ample time for discussion with developers. Several papers were published and the overall open-access online platform is currently being published (manuscript submitted, web-pages being prototyped).
The subtopics initiated and continued discussions with regulatory bodies about the project stage at which regulatory considerations were relevant (ST1), a strategy for integration of CHIMs into pharmaceutical vaccine development (ST2), acceptance of next-generation human in vitro 3D models in de-risking vaccine development (ST3) and acceptance of in silico models into CMC dossier (ST4).
We expect to develop several innovative methods to de-risk vaccine development and shorten vaccine development timelines.
•Open-access, cloud-based platform to predict post-vaccination immune formation, to enable early selection of vaccine candidates more likely to produce efficient immune responses and succeed through later stages of vaccine development.
•New and improved CHIMs for influenza, RSV and C. difficile will enable early vaccine efficacy evaluation.
•Cell-based human in vitro 3D models simulating mucosal infections to more reliably predict immune protection. These will be combined with development of related functional immune assays for clinically relevant (surrogate) endpoints. 3D models will help reduce animal experiments and de-risk pre-clinical development by complementing traditional assays.
•Modular one-stop computational platform for in silico modelling of vaccine bio-manufacturing and stability testing will help optimise vaccine production.
•Open-access, cloud-based platform to predict post-vaccination immune formation, to enable early selection of vaccine candidates more likely to produce efficient immune responses and succeed through later stages of vaccine development.
•New and improved CHIMs for influenza, RSV and C. difficile will enable early vaccine efficacy evaluation.
•Cell-based human in vitro 3D models simulating mucosal infections to more reliably predict immune protection. These will be combined with development of related functional immune assays for clinically relevant (surrogate) endpoints. 3D models will help reduce animal experiments and de-risk pre-clinical development by complementing traditional assays.
•Modular one-stop computational platform for in silico modelling of vaccine bio-manufacturing and stability testing will help optimise vaccine production.