Vaccinology: Getting our modernization act together

Vaccines are the most effective biomedical intervention, but the path to their development and authorization or approval remains slow and has a high failure rate. The typical approach to the discovery and development of immunomodulators and vaccines typically involves a one-size-fits-all concept. While this would be ideal, demographic features such as age, sex, comorbidity, genetics, and epigenetics affect the immune response as well as the safety and efficacy of immunomodulators and vaccines (van Dorst et al., 2024). For example, even for the highly successful mRNA vaccines, the need for multiple/repeated dosing, especially in vulnerable populations, such as older adults and immunocompromised individuals, remains a significant limitation.

Considering all the infections for which we need new (e.g., HIV) or improved (e.g., influenza, SARS-CoV-2, tuberculosis , and pertussis) vaccines, the range of potential vaccinal antigens, a growing pipeline of vaccine adjuvants, as well as wide population variability, it will simply not be possible to pursue phase 3 trials for each combination of antigen, adjuvant, and population (Singleton, 2023). Vaccines to prevent noninfectious diseases, including allergy, asthma, cancer, and drug overdose, further expand the range of agents to be considered.

Up until recently, the United States Food and Drug Administration (FDA) was operating under the Federal Food Drug and Cosmetic Act of 1938, requiring use of animal models (often mice) for preclinical safety and activity assessment of new therapeutics prior to proceeding to phase 1 human trials. Considering the high failure rate and low approval rate for new drugs and vaccines entering phase 1 clinical trials, in part reflecting the distinct genetics, pharmacokinetics, and pharmacodynamics in animals, there has been an unmet need for more efficient biomedical translation. In this context, passage and signing into law of the FDA Modernization Act 2.0 on December 29, 2022, authorizes the use of novel alternatives to animal testing. This major development paves the way to realizing the power of in vitro and in silico modeling as well as systems biology, artificial intelligence, and machine learning as applied to drug and vaccine discovery and development (Zushin et al., 2023). Such technologies can focus on human biology to identify actionable biomarkers to enhance translation, including assessment of drug safety and efficacy, in the case of vaccines, as linked to immunogenicity (e.g., induction of antibody and T cell responses).

In vitro modeling approaches are potentially powerful and impactful as they (a) use biosample from the same individual as control and test conditions, thereby inherently controlling for genetic and epigenetic features, which is not possible in a clinical trial in vivo; (b) enable testing of multiple vaccine components—e.g., adjuvant and antigen at multiple concentrations; (c) provide insight into various formulations (aqueous, oil-in-water, liposomal, etc.); (d) allow benchmarking of novel adjuvants and vaccines against licensed products of known safety, immunogenicity, and efficacy; (e) successfully model the action of mRNA vaccines as well; and (f) can be subjected to targeted molecular assays or multi-omic profiling generating biomarkers and molecular signatures matching those observed in vivo (Oh et al., 2016; Dowling et al., 2017; Sanchez-Schmitz et al., 2018; Brook et al., 2024; Jeger-Madiot et al., 2024) (Fig. 1).

As recently reviewed, multiple human in vitro systems have been described, including fluid two-dimensional (2D) systems, such as whole blood assays and peripheral blood mononuclear cells, as well as three-dimensional (3D) systems with extracellular matrix, such as age-specific tissue constructs and tonsillar organoids, and multi-modular systems that can measure antigen presenting cell–induced CD4 and CD8 T cell activation (Morrocchi et al., 2024). 2D systems, such as whole blood and peripheral blood mononuclear cell assays, are more straightforward and can provide insight into clinically relevant innate responses, such as those to adjuvants. 3D systems can more accurately reflect immunologic events involving autonomous cell migration, such as those occurring in lymph nodes, but are more complex and require extensive training. Capturing both 2D and 3D approaches from the same individual likely provides a fuller picture of immune responses. For in vitro modeling, use of fresh autologous plasma enables more accurate modeling than use of xenologous (e.g., fetal bovine serum) or artificial media (England et al., 2021; Pettengill et al., 2014).

In addition to human in vitro modeling, an alternative or complementary strategy to innovate, enhance, and accelerate preclinical testing involves development and use of humanized mice, an approach which has become increasingly sophisticated (Akkina, 2013; Kaushik et al., 2024). Humanized mice can support engraftment of human hematopoietic stem cells. SCID or RAG1/2^−/− mice have been genetically engineered to express key human cytokines to further support immune responses that are similar to those in humans. Such humanized mice have been used to model responses to vaccines against dengue, Hepatitis C virus, HIV, and EBV, as well as malaria. In principal, humanized murine immune system can be constituted from the cells of different human populations to model population-specific immunity.

A challenge in modern vaccinology is definition of the correlate of protection, which then enables subsequent studies that are designed and powered on this more practical endpoint, rather than on disease. For example, despite the resources and effort invested into COVID vaccines, we still do not have a clear correlate of protection for those vaccines, hampering progress in designing next-generation vaccines that provide more durable protection against this virus. With the goal of better defining biomarkers predictive of safety, immunogenicity, and protection. Moving forward, clinical trials should include collection and processing of in vivo and in vitro (vaccine stimulated) samples from the same study participants collected pre-treatment for downstream multi-omic analysis and bioinformatic integration of clinical, immunologic, and systems biology data to define basal and vaccine-induced biomarkers predictive of vaccine safety and immunogenicity.

To advance human health, each component of the biomedical ecosystem has a role to play in embracing the vision inherent to FDA Modernization Act 2.0. Academics can build the multidisciplinary collaborations needed to design and conduct experiments employing these new tools. Government, including the National Institutes of Health, can support programs that leverage humanized mice and human in vitro systems coupled with robust measurement and analysis of multi-omic biomarkers. Finally, with respect to human clinical trials, industry should be expected to provide study results beyond simple antibody titers or a few biomarkers. If the costs of applying these methodologies across an entire large clinical trial become prohibitive, sub-cohorts should be included for which samples are collected amenable to downstream multi-omics to ensure that if a trial succeeds, we define actionable biomarkers that correlate with protection, and if it fails, we learn why.

Many will argue that such innovative approaches to translational medicine will be cumbersome and costly. However, the question is whether it is not more cumbersome and costly to conduct multiple clinical trials that fail for reasons we do not understand, resulting in a long and inefficient process that exposes study participants to agents that often fail or are sub-optimal? Indeed, time is a major factor. For example, rigorous economic analysis indicates that had a safe and effective SARS-CoV-2 vaccine been released half a day—a mere 12 h—earlier than it had been, the economic benefits accrued from reopening the global economy sooner would have paid for the entire $12 billion U.S. dollar Warp Speed initiative (Pecetta et al., 2022)!

Before us lies tremendous opportunity to advance human health by embracing new preclinical approaches, including human in vitro modeling and systems biology. Realizing the benefit of these technologies will require innovative partnerships between academia, government, and industry. Considering the threats we face, the scale of potential benefit is such that substantial and urgent investment is called for.