Understanding Preclinical Vaccine Safety: Regulatory Guidelines Explained

Vaccines are one of the few medical interventions which are regularly undertaken by healthy individuals. As a result, the bar for ensuring the safety of a vaccine is extremely high. It would be difficult to justify administering a preventative vaccine for a disease an individual might contract if that vaccine regularly causes serious side effects or – worse – giving that individual the disease the vaccine is designed to prevent.

We’ve previously taken a broad view of preclinical testing for vaccines, including the key properties which a vaccine possess before human testing can be considered. Here, we’re going to focus specifically on preclinical studies designed to establish the safety of a vaccine. These studies, conducted in animal models and laboratory settings, identify risks, evaluate immune responses, and confirm that the vaccine doesn’t cause harmful side effects. But what exactly does this process involve, and how do regulators like the FDA, ICH, and WHO ensure every vaccine meets strict safety criteria before it enters First In Human (FIH) trials? In this article, we’ll examine preclinical safety testing for vaccines, exploring how it differs from other drug products and why it’s crucial in safeguarding public health.

All vaccines require some form of human testing before they can be licensed. Even if the efficacy of the vaccine can only be ethically established in animal models due to the severity of the target, the safety of the vaccine must be formally established in human trials. But, before human safety trials can take place, it must be established that the vaccine is safe enough to expose to human volunteers. Trials without these checks in place would be unethical due to the undue risk posed to the volunteers. While no trial will ever be 100% safe for participants – particularly as safety data from animal models does not always translate to humans – it is the duty of researchers to mitigate the risk of complications as far as possible.

As such, there are stringent safety requirements a vaccine must pass before it can be used in human trials. The FDA guidance document S6 Preclinical Safety Evaluation for Biotechnology-Derived Pharmaceuticals recognises three main goals for preclinical safety testing:

1. “To identify an initial safe dose and subsequent dose escalation scheme in humans”
2. “To identify potential target organs for toxicity and for the study of whether such toxicity is reversible”
3. “To identify safety parameters for clinical monitoring”

This guidance document is far from specific to vaccines – indeed, it is only strictly applicable to certain types of vaccines, specifically those based on recombinant DNA proteins. These three goals, however, provide an excellent framework for the broad aims of preclinical safety testing for vaccines.

A key component of preclinical safety testing for vaccines is the repeat-dose toxicity (RDT) study. This is a study in animal models which aims to establish a safe dose of a vaccine in humans and detect any toxicities to be monitored in further trials. Unlike for small molecule drugs, for which studies in multiple species are usually required, a RDT study in a single species is often sufficient, provided it is adequately predictive. All RDT studies should be performed to Good Laboratory Practice (GLP) standards.

A RDT study administers several doses of a vaccine to a population of animal subjects, which are then examined for a range of endpoints which indicate the safety profile of the vaccine. This is compared against results from a control group, who are usually administered with saline, though the vaccine formulation without antigen can also be used. The aim is to maximise the immune response generated by the vaccine: the dose administered should be the highest dose proposed to be used in a clinical trial or an equivalent. The number of doses administered should be at least as many as proposed for human use.

The toxicology of the vaccine is assessed over a wide range of endpoints. During the trial period, this includes routine monitoring of subjects for signs of illness, such in changes in behaviour and food and water consumption, as well as for injection site reactions. Serological studies, comparing blood chemistry before, during, and after the treatment phase  are also often included. At the conclusion of the trial, a full necropsy should take place, with the histopathology and weight of relevant organs assessed. The choice of tissues to be examined will depend on the nature of the vaccine  but, generally, pivotal organs such as the brain, kidneys, and liver are included along with immune organs and the site of injection.

Once the data has been collected, the results are summarised and compared between the treatment and control groups. For example, one important endpoint might be the concentration of cytokine proteins near the site of injection. Cytokines are signalling proteins associated with inflammation, so a high concentration could indicate a strong reaction near the site of injections. So, the cytokine concentration near the injection site is measured for both test and control groups. From this, a mean concentration is found, alongside a standard deviation across the sample. The latter can be combined with the sample size to calculate a confidence interval on the mean. If the data follows a log-normal distribution, these calculations would be performed after first taking a log transformation.

Once these summary statistics have been calculated, there are several approaches to come to a conclusion about the safety implications of the results. One could simply compare the (or another property of the data, such as the median) for the test group to that of the control to establish whether there is a statistically significant difference between two. This would be an appropriate approach if the safety profile of the vaccine demands that there be minimal or no injection site reaction. More commonly, however, some degree of injection site reaction is deemed safe, as it can typically be easily managed with painkillers. One might, therefore, consider comparing the test group mean to a critical concentration which indicates a severe injection site reaction. Each key endpoint should be evaluated against an appropriate criterion, with the results determining whether the study is indicative of vaccine toxicity.

Typical approaches to comparing results from different groups include an ANOVA on raw or transformed data. A non-parametric counterpart, the Kruskal-Wallis test, can also be used. These tests would be followed by appropriate post-hoc tests. More statistically advanced techniques (e.g., time series) can be applied, too, if the data structure allows it.

While, ideally, a vaccine candidate would perform without any toxicity and side effects, it is not the case that there need be no side effects observed at all for a vaccine to be deemed suitable to proceed to FIH trials. WHO guidance outlines that “scientific judgement should be applied to the interpretation of data from preclinical studies, regarding the risk–benefit ratio, animal model, dosing etc” and that, if mild side effects are observed, this “may indicate the necessity for careful monitoring of a particular clinical parameter” rather than precluding further trials.

Further, evidence from a RDT study can help determine a safe initial dose for FIH trials. Generally, vaccine toxicity testing takes place at or slightly above the expected human dose, or an equivalent adjusted to the size of the animal. If there are no serious adverse events observed at these dose levels, then the vaccine is usually considered safe to move to FIH trials.

Developmental toxicity is the potential for a substance to cause damage to a developing foetus, including birth defects and miscarriage. As many vaccines are given in early childhood, developmental toxicity testing is often not required for vaccine candidates. For vaccines whose target population includes adolescents or adults, FDA guidance indicates that developmental toxicity should form part of the testing of the vaccine.

However, specific developmental toxicity studies need only form a part of a preclinical safety screening if the vaccine is specifically intended to be given to patients who are currently pregnant. These studies are typically conducted in animal models, where pregnant female subjects are dosed with the vaccine during pregnancy. The mothers and offspring are then monitored through birth and onto weaning to observe for normal growth and development.

Carcinogenicity – the potential for a substance to increase the likelihood e development of cancer – and genotoxicity – the potential for a substance to damage a subject’s DNA – are generally not required to be tested for a vaccine. Vaccines are short-duration exposures, meaning the potential to cause cancer is limited. Similarly, most vaccine components are not DNA reactive, so there is little chance for DNA to be damaged. If a vaccine formulation includes a novel adjuvant or additive, however, the new component may require testing for carcinogenicity and genotoxicity.

Should a carcinogenicity study in animal models be required, a common approach is to use the Peto analysis . This is a method which combines time-to-death and prevalence analysis into one test, giving a more in-depth evaluation of the carcinogenicity of an adjuvant or other component than either analysis alone.

In Vitro vs In Vivo Potency Assays: Modern Vaccine Batch Release

Choosing a Path: Natural History Studies for Vaccine Development

Understanding GLP Carcinogenicity Studies: OECD 451 & 116

Modern vaccines deploy several different techniques to present the vaccine antigen to the immune system. These range from the traditional live or attenuated virus platforms, through to modern mRNA vaccines. So far, we’ve focused on the safety testing which are required for all vaccines regardless of platform choice. There are, however, different safety testing recommendations for each platform due to their diverse mechanisms by which they perform their role. We’re going to examine each in turn, and highlight any specific guidance where appropriate.

mRNA vaccines – among the greatest success stories to come out of the COVID-19 pandemic – contain messenger RNA (mRNA) which encodes for a target antigen. These are encapsulated in lipid nanoparticles LNPs which protect the mRNA and allow it to enter cells.

A key additional safety concern for mRNA vaccines is the biodistribution of the LNPs and the mRNA they contain. It is important to determine where in the body these components travel, how long they persist, and whether they cause any toxicity as a result. For example, for Moderna’s mRNA-based COVID-19 vaccine, the distribution of mRNA and LNPs was assessed by measuring the amount of these components in the blood and other organs of animal models at a series of timepoints after dosing. The subjects were also examined for any toxicity, and this was correlated against the component concentration.

A further concern specific to the LNPs often used in mRNA vaccines is hypersensitivity. There is potential for the ingredients used in the LNPs to cause local inflammation or even more serious allergic reactions. Animal subjects should be examined to assess whether there are indications of hypersensitivity either at the injection site or local organs. If these are present, then it is prudent to monitor for hypersensitivity closely in any further trials.

A commonly used vaccine platform is the viral-vector vaccine. These use a harmless virus – such as an adenovirus – which has been genetically modified to express a key antigen from the target pathogen. Several COVID-19 vaccines were based on viral-vector technology, as well as vaccines for Ebola and Smallpox.

As with mRNA vaccines, studies of biodistribution are a key component of safety testing of a viral-vectored vaccine. EMA guidance for preclinical studies of viral-vector vaccines specifically notes the importance of testing the brain of animal models for the presence of the virus, as crossing of the blood-brain barrier could be an indication of neurovirulence.

Unlike mRNA vaccines, however, the presence of live viruses in the formulation means that, while the viruses are attenuated, unintentional infection by those viruses is a possibility. Therefore, it is important to confirm that the virus remains attenuated and does not return to a virulent form. For example, studies in animal models must assess whether there is evidence of viral shedding. This is not only a sign of infection, but could lead to unintentional transmission of the viral vector if it persists when used in humans.

Finally, it’s important to establish the immune response to the viral vector itself. The vaccine is intended to generate immunogenicity to the target antigen, but there will also inevitably be some response to the virus itself. Preclinical studies should, therefore, test for the presence of anti-vector antibodies in serum samples, as a strong immune response could lead to adverse side effects or affect dosing regimens.

Many vaccines contain the pathogen which they are intended to prevent in either a weakened – attenuated – form or fully inactivated. Examples of the former include the MMR and yellow fever vaccines, while Salk’s polio vaccine was among the first to use an inactivated virus.

These vaccine platforms share many of the same safety concerns as viral-vector vaccines, such as ensuring the virus is fully inactivated or attenuated. Testing for these is particularly stringent for live-attenuated viruses – signs of a return to a virulent form have significant implications for further development of the vaccine candidate. Another test which is often required for live-attenuated viruses is for the possibility of genetic exchange with wild strains of the virus. This could result in a reversion to virulence or – worse – a more virulent strain of the virus than even the wild type.

Many vaccine formulations include an adjuvant – a component which boosts the immune response to the target antigen. While adjuvants can be used with any delivery platform, protein subunit vaccines almost always require an adjuvant to be effective. These are vaccines which contain only the purified antigen proteins rather than full pathogens – examples include the HPV and Meningitis vaccines.

Adjuvants used in vaccine formulations are often tested in isolation, particularly if the adjuvant is new. This helps establish whether any adverse effects are associated with the adjuvant or other components of the vaccine formulation. Important safety endpoints include the concentration of inflammation markers, both throughout the body of the subject and particularly at the injection site. These provide guidance for aspects to monitor in later studies.

As we’ve examined, the broad philosophy of preclinical safety testing for vaccines is similar to that for other drug products. For example, the repeat-dose toxicity study remains core to demonstrating the safety of both classes of products. There are, however, some key differences here. For small molecules, the focus is often on quantitative toxicology, such as identifying a safe starting dose in humans using the No Observed Adverse Events Level. For vaccines, a greater emphasis is placed on qualitative outcomes – whether concerning pathology is observed or not – as guidance for the safety of further trials and important clinical endpoints to be monitored.

Further, many tests which are required for small molecule drug products are not required for vaccines due to their different use cases. For example, small molecule submissions must include testing for carcinogenicity. This is not required for vaccines, in part due to the limited period over which a patient is exposed to a vaccine.  Similarly, genotoxicity studies do not typically form part of a preclinical testing programme for vaccines, while they are usually necessary for other drug products.

These differences highlight that, while there are many aspects of vaccine development which follow a similar process to other classes of drug products, vaccines form a unique sector of the drug development landscape. They are among the very few products which are given prophylactically, meaning their risk-benefit profile is unusual and that safety is of paramount importance. So, while efficacy can be proven in animal models in some cases, all vaccines require human testing for safety, meaning there is a strong ethical mandate to prove that a candidate is safe enough to expose to human testing. That’s why preclinical safety testing, with its myriad aspects and requirements for the wide range of different vaccine platforms, is so crucial.

Many of the trials which make up preclinical safety testing are required to be performed to GLP standards. We’ve looked into some of the guidance which covers GLP statistical testing in previous blogs. Quantics are the only independent GLP-certified statistical consultancy, with more than 20 years of experience providing expert analysis for vaccine development and testing. Don’t hesitate to reach out to find out how we can help you with your latest project.