A COVID‐19 human viral challenge model. Learning from experience

Abstract The controlled human infection model and specifically the human viral challenge model are not dissimilar to standard clinical trials while adding another layer of complexity and safety considerations. The models deliberately infect volunteers, with an infectious challenge agent to determine the effect of the infection and the potential benefits of the experimental interventions. The human viral challenge model studies can shorten the time to assess the efficacy of a new vaccine or treatment by combining this with the assessment of safety. The newly emerging SARS‐CoV‐2 virus is highly contagious, and an urgent race is on to develop a new vaccine against this virus in a timeframe never attempted before. The use of the human viral challenge model has been proposed to accelerate the development of the vaccine. In the early 2000s, the authors successfully developed a pathogenic human viral challenge model for another virus for which there was no effective treatment and established it to evaluate potential therapies and vaccines against respiratory syncytial virus. Experience gained in the development of that model can help with the development of a COVID‐19 HVCM and the authors describe it here.

Consequently, there is a tremendous need for effective vaccines, mAbs, immunomodulators, antivirals and therapeutic interventions which must be developed quickly, efficiently, and with high scientific and ethical rigour.

Historically, controlled human infection studies (CHIMS) and
specifically human viral challenge model (HVCM) have rapidly advanced the development of interventions for many infectious diseases. Recognising their potential advantages for controlling the COVID-19 pandemic, the World Health Organization has recently issued guidelines on such studies. 9 A COVID-19 HVCM would be an important tool to rapidly evaluate the several hundred candidate interventions in early development, 10 selecting the most promising candidates while also improving our critical understanding of viral pathogenesis and defining correlates of protection.

| CONTROLLED H UMAN INFEC TI ON MODEL AND HUMAN VIR AL CHALLENG E S TUD IE S
Human challenge studies have existed for over 200 years since Jenner and the development of the smallpox vaccine in 1793. 11 In the modern era, the CHIM has become established for many pathogens in particular as the HVCM. 12 Standard clinical trials carry a risk which must be balanced with the potential benefit. Likewise, an optimised risk-benefit balance must be considered in all clinical studies. Despite the inherent risk of standard clinical trials, thousands of such studies are conducted successfully and safely each year, allowing the development of treatments and in some cases, the control or even eradication of diseases.
HVCM studies are not dissimilar to standard clinical trials while adding another layer of complexity and safety considerations. The model ( Figure 1) deliberately infects volunteers, with a challenge agent (CA) to determine the effect of the infection and the potential benefits of the experimental interventions. In standard clinical trials, it is extremely difficult if not impossible, to study the events occurring early during human infections. This is especially true in SARS-CoV-2 infections which are known to begin without recognisable symptoms. Therefore, HVCM studies can also be extremely valuable to understand better the pathogenicity of the virus before symptoms are apparent and the correlates of subsequent protection.
HVCM studies can shorten the time to assess the efficacy of a new vaccine or treatment by combining the assessment of safety, with potential efficacy end points, as shown in Figure 2. HVCM studies may be used to accelerate the evaluation of vaccines and therapies against COVID-19, leading to possibly earlier licensure. 13 COVID-19 is highly contagious and can cause severe illness and death. The understanding of the pathogenesis of this new virus is still developing, literally daily. Also, there is no standard therapy or vaccine available, although remdesivir has antiviral and clinical benefit even when started relatively late in the infection and has thus been given emergency use authorisation by the FDA. 14 Previous successful human challenge studies have usually had a therapy available for treatment should a complication occur during an HVCM study, but this has not always been the case, as will be discussed later. A COVID-19 HVCM study presents multiple ethics, safety, feasibility and regulatory hurdles.
F I G U R E 1 An outline of a HVCM study, specifically the Human viral challenge model. The study typically consists of inputs, such as the volunteers, their selection criteria, isolation in quarantine and exposure to a GMP virus. There are two treatment options; a vaccination/ prophylaxis with an antiviral or b treatment with an antiviral. Outputs from the study summarised on the right, such as virus symptoms and virus shedding. X is the number of days before virus exposure vaccination may occur. Y is the number of days post-virus exposure that a volunteer may be followed for 12 However, HVCM studies, conducted ethically and safely, can have two significant benefits. Firstly, they can speed the development of vaccines and treatments, and secondly, possibly and more importantly, they can prioritise the most promising candidates amongst the hundreds currently being researched for more extensive studies.
On 6 May, the WHO issued its guidance on the conduct of COVID-19 CHIM/HVCM studies, outlining "Key Criteria for the ethical acceptability of COVID-19 human challenge studies". 9 The specified eight criteria are shown in Table 1.
Like all clinical studies, the criteria ensure that the optimised balance between risk and benefit is evaluated. A notable addition in the WHO guidance is criteria seven, "a specialised independent ethics committee" consisting of those with experience of COVID-19 and HVC studies which is eminently sensible.
The choice of the challenge agent and the participants is particularly important concerning HVC studies where no effective treatment might exist. The authors describe their experience in addressing these two issues specifically, in the context of setting up F I G U R E 2 The role of the HVC model in the clinical development pathway. Short duration proof-of-concept studies, which incorporate the HVC model, typically include small numbers of subjects. The resulting safety and, particularly, efficacy data can more accurately guide decisions on whether to expose a larger number of subjects to promising candidate therapeutics in community-based 14 field studies than conventional phase 1 safety data alone might otherwise 12

1
There should be a strong scientific justification For COVID-19 however, several animal models have been proposed, including ACE2 mice, ferrets, rabbits, non-human primates and golden Syrian hamsters that may simulate human infection and disease. 22 Whether or not they should be used to "safety test" a potential challenge agent (CA) should be debated.

| PARTI CIPANT S ELEC TI ON
Assuming that a suitable CA can be selected and a GMP lot produced, it is imperative to select a population with the lowest possible risks of developing severe complications of the infection.
As RSV can be a severe lower respiratory tract illness, in the authors' development of the RSV HVCM participating volunteers were carefully selected to be healthy males and females 18-45 years of age and who had no history during adulthood of asthma of any aetiology or any use of a bronchodilator within the past year as well as other factors that may have complicated the analysis of the end points, for example hay fever or allergic rhinitis.
Importantly, contact with people at risk of severe RSV infections, for example steroid use in the past month, chronic sinusitis and the presence of known immunosuppressive conditions, would also exclude a volunteer.
Similarly, the selection of volunteers would need careful consideration for a COVID-19 HVCM. As an example, the age range would likely be limited to adults <30 years of age as the risk of complications increases in older age groups. 23 The SARS-CoV-2 infection in younger age groups causes less severe disease, perhaps due to lower expression of ACE-2 receptors in the respiratory tract. 24 Multiple additional factors appear to place prospective participants volunteering at risk of serious COVID-19 disease, including, but not limited to, race, 25

| RELE VANT E XPERIEN CE FROM THE E S TAB LIS HMENT OF THE R S V HVCM
Respiratory syncytial virus (RSV) is a paramyxovirus that infects more than 60% of children during the first year of life. 31 This virus is associated with significant morbidity and mortality. RSV is particularly severe in immune naïve (infants), 32 and amongst the frail elderly, RSV disease burden is similar to that of influenza. 33  CA was manufactured to a high titre due to the sensitivity of the virus to freeze/thaw cycles and that we were attempting to infect RSV immune experienced adults and therefore perhaps needed a high inoculum to infect.
After production, the fill-finish was aliquoted in sterile glass vials and stored at −70 to −80°C. The final CA had been passaged only five times, thus limiting the number of mutations that could be introduced during manufacturing and hopefully ensuring the virus was typical of a wild-type circulating strain. As with all live biologics administered to human volunteers in clinical studies, tests for purity and adventitious agents were extensive and followed FDA guidance documents for live viral vaccine production for human use.
The selection of the challenge agent is important; in this case, the patient from whom Memphis-37 was originally isolated was a 4-month-old, 5.9 kg, African-American male. He was previously healthy, and a full medical history was available for both mother and child, and repository consent was obtained. RSV exists as multiple different circulating strains of the virus; Memphis-37 was shown to be broadly relevant to those currently circulating.

| PRELIMINARY CLINIC AL A SS E SS MENT OF CHALLENG E AG ENT
Thirty-five volunteers pre-selected to have low RSV Memphis-37 neutralising titres were divided into five cohorts and in a sequential manner were inoculated with increasing quantities (3.0-5.4 log PFU/ person) of RSV Memphis-37 intranasally. 40 Each volunteer in a cohort received an aliquot of the same inoculum. Inoculation was by intranasal drops (0.5 mL/nare).
Between each cohort, the data were evaluated, including symptoms, signs and the viral load from daily nasal washes. After review by the safety committee, a decision was made as to whether or not to administer a higher inoculum of challenge agent to the next volunteer group. The lowest inoculum that reliably produced infection and symptoms was chosen.
Overall, 77% of volunteers consistently shed the virus. Infection rate, viral loads, disease severity and safety were similar between cohorts and appeared unrelated to the quantity of RSV received.
Symptoms began near the time of initial viral detection, peaked in severity near when viral load peaked and subsided as viral loads (measured by RT-PCR) slowly declined as shown in Figure 3. Interestingly, in the experience of both authors, the desire of volunteers to leave once admitted to isolation is rare, and usually caused by an unexpected social/family event, rather than boredom.
Indeed, the team caring for the volunteers ensures they are "occupied" and the provided social media access means they rarely feel alone. However, if a volunteer needed to leave the facility consideration would be given to isolation at home, available antiviral treatments and appropriate support and home monitoring as has been accomplished in other HVCMs.
There is an important difference between long-term PCR positivity, and long-term shedding of infectious viral particles, 47  There is also recent precedent for vaccine licensure based on data from HVCM studies. Examples include cholera and typhoid vaccines. [49][50][51] The use of the COVID-19 HVCM may enable emergency use authorisation/licensure far more rapidly than a conventional approach, without replacing it. allowing for safety and manufacturing considerations, such that the data generated from the model is applicable to naturally occurring infections. Ten years ago, we developed an RSV HVCM that is now frequently used, and believe that a similar properly designed and orchestrated safe and reproducible COVID-19 HVCM is possible and would be of substantial global benefit.

ACK N OWLED G EM ENTS
The authors would like to thank all of those involved in the development of the RSV HVCM, including those colleagues who assisted with the study, the clinical trial subjects and those providing regulatory and ethical oversight.

CO N FLI C T O F I NTE R E S T
This is a review article, and all various declarations have been made

DATA AVA I L A B I L I T Y S TAT E M E N T
This manuscript is a commentary piece and any raw data; therefore, this section is not applicable.