Human Challenge Trials

Plotkin and Gilbert have defined a correlate of protection (CoP) as “a (bio)marker of immune function that statistically correlates with protection after vaccination” (1). In the 1980s and 1990s, it was widely assumed that CoPs or surrogates of immunity would be established for most infectious diseases following the scientific revolution that accompanied advances in molecular biology. However, a single measurable marker that correlates to functional immunity has all too frequently proved elusive and further research has only served to highlight the complexities of the human immune system in response to antigenic challenge. Many promising serum markers have failed as correlates upon further examination where, for example, positive neutralization or opsonophagocytosis assays in vitro have not translated to in vivo immunity.

A true CoP must possess a number of attributes before it can be considered predictive of vaccine efficacy (2):

1. Protection must be significantly related to the vaccine.
2. The marker must be significantly related to the vaccine.
3. The marker must be significantly related to the clinical endpoint.
4. The marker must explain all of the clinical endpoints.

To further complicate matters, various operator or process-related elements may affect the usefulness of a CoP, including inter/intra-assay variability, the diversity of units of measurement (titer, dose–response curve, and so on) and even inbuilt errors in statistical analysis (point estimates without confidence intervals, poorly defined or inappropriate endpoints, or under-powering for clinical outcomes). Such determinants have led to a high attrition rate for candidate markers (3).

Although antibody responses have been employed as correlates for many licensed vaccines, there are a number of diseases where vaccines have been developed in the absence of any meaningful surrogate, including variola, measles, rotavirus, HPV, and VZV. In fact, for the Top 10 infectious diseases without a licensed vaccine (4), eight also have no established CoP. In response to a lack of de facto evidence, the FDA has recently advised that future licensing of vaccines will be predicated on the effects on symptoms and not just on serum correlates. 

The big question: could human challenge trials be the answer when a CoP cannot be found?

Edward Jenner challenged individuals with live vaccinia virus to promote a protective response to variola. Jenner’s novel smallpox vaccine became accepted practice and pioneered the use of challenge models in exploring immunity to infectious diseases prior to the knowledge of CoPs. Latterly, at the conclusion of World War II, UK physician David Tyrell further developed the human challenge model by inoculating small quantities of rhinovirus into the noses of volunteers at the Common Cold Research Institute, Salisbury, to study the epidemiology of infection and the effectiveness of antiviral therapies. Following the closure of the unit in 1989, the UK and US continued viral challenge studies, utilizing cGMP agents (rhinovirus, influenza, and respiratory syncytial virus) for the evaluation of novel therapies for respiratory disease. 

Such challenge trials are now being applied to a wider range of diseases, to augment – or even replace – community Phase II proof-of-concept studies, and diminish the need for correlates of protection.

The use of human challenge trials to emulate natural infection, and to estimate attack rates and disease severity following vaccination, can help to distinguish between the epidemiological concepts of efficacy (disease incidence in vaccinated and unvaccinated groups) versus effectiveness (preventing negative outcomes of real-world interest; for example, fewer cases of pneumonia). Double-blind, randomized, controlled challenge trials make it possible to determine the impact of vaccines or drug therapies on clinical endpoints, and so optimize drug design, excipients, and delivery systems.

Both human challenge trials and biomarkers can accelerate the identification and licensure of candidate vaccines, reducing total study subject numbers by a factor of 10 and contracting timelines by up to five years (5),(6). Additionally, challenge trials can be employed as an investigative tool to relate humoral markers of disease to symptom scores; to validate vaccines against related families of pathogens or a range of serotypes; or to study multiple clinical endpoints.

With cGMP challenge agents offering increased safety over wild-type infections, we can expect to see challenge trials become increasingly popular, as pressures to reduce costs and shorten timelines continue to drive developmental pipelines. As a researcher in this field, I can foresee a time when viral, bacterial and parasitic diseases are all served by libraries of challenge agents, available to the academic researcher and commercial Pharma alike – funded by centers of excellence. The biggest hurdle to further developing the human challenge model remains GMP challenge agent manufacturing costs. We need governments to recognize this burden and put in place appropriate funding mechanisms to alleviate it.