The nature of infectious agents is rarely static. The replication machinery of viruses such as influenza ensures a gradual antigenic drift. This antigenic modification is the driver behind the annual change to the strains included in the seasonal influenza vaccine. The same is true of influenza challenge agents used to assess drug and vaccine efficacy.
Challenge agents require to be updated over time if they are to represent circulating strains and accurately emulate efficacy in the community. Although it is rare for circulating strains to alter wildly, recently the common H3N2 influenza virus has changed its attachment proteins to the extent that serological tests to detect its presence are failing.

The rate and degree of drift is difficult to predict and measuring evidence of serological relatedness for predominant strains may not always be accurate. There have been several seasons of mismatched vaccines from 2013-15, but this is not the only issue facing vaccine developers. Vaccines can become adapted over time to the eggs or cell lines in which they are grown, resulting in viral strains that are less effective at binding onto human cells and causing infection. The result is a strain that has a low attack rate, poor replication and low seroconversion rates.

Viral challenge testing of influenza vaccines with latest challenge agents, which are manufactured to current Good Manufacturing Practice (cGMP) guidelines, represent a useful technique to evaluate whether a new vaccine is likely to be effective against the currently circulating strains. Ideally, the challenge agent will be derived from a pediatric source, as the patient will be less likely to be co-infected with other viruses and will have a clear and simple medical history. The agent should also show a strong and consistent symptom profile and not cause any serious adverse events (sequelae). Another important characteristic of the challenge agent is that it should have a epidemic strain in its lineage, in order to confer some cross-protection should that season’s strain become epidemic too.

The manufacturing process for challenge agents, as for vaccines, is heavily regulated. Production of live virus must be carried out in line with cGMP guidelines, within accredited premises, and using accredited laboratory testing, compliant with Good Laboratory Practice (GLP). Although there are currently no specific guidelines for developing challenge agents, possibilities to lighten the burden of proofs for safety are under discussion, in the hope that it may be possible to develop new challenge agents faster and at a lower cost.

The manufacturing process
The first step in the process for identifying a suitable challenge virus (i.e. a virus causing acute disease with a low mortality rate, limited morbidity and little or no evidence for chronicity, sequelae or serious long term disease, for example many respiratory viruses) is to isolate and identify a candidate strain from the community, preferably in a pediatric patient with a well-documented medical history. Preferably the candidate virus is then sequenced to ensure it is representative of circulating strains, and to look for resistance to available therapeutics. Pilot tests may be undertaken on a range of candidate viruses to see how well each grows on an appropriate cell line, such as Madin-Darby Canine Kidney (MDCK), and to check whether the virus can infect the cells easily, without losing infectivity or showing signs of attenuation, and if it may easily be grown on, or amplified in, other cell lines.

At this stage it may also be prudent to check the antigenic properties, comparing them for similarities (homology) to current vaccine strains.

With a suitable viral candidate identified, further pilot studies may be performed to determine whether the virus can multiply successfully in the Working Cell Bank or the chosen egg stock during the manufacturing process. Pilot studies are used to identify the relative suitability of differing cell-lines and embryonated eggs from differing specific-pathogen free (SPF) chicken flocks by investigating dosing titers and incubation temperatures and times. Cell lines are increasingly being preferred to egg stock in the light of egg adaptation and the increase of egg albumin allergies. If eggs are ultimately the medium of choice, it is advisable to test out more than one supplier of eggs to check that there is no “egg effect”; it is not unknown for eggs to be resistant to a particular virus.

The optimal titer for inoculating the medium in the manufacturing process is important. Too many rounds of amplification is not ideal, as the virus will be more likely to become adapted to its new host medium and thus become less effective at infecting humans. So-called egg or cell-line adapted virus will have a reduced capacity for causing symptomatic disease and may exhibit truncated or decreased shedding, thus compromising its usefulness as a successful challenge agent.

Once it is certain that the candidate agent grows well in pilot studies, the viral seed-stock can be produced. Seed-stock or the starting inoculum to grow on into the final product is grown either in high-yield cell lines or a small number of embryonated eggs from SPF flocks. Virus is injected or inoculated at a dilution established during the pilot study as optimal for rapid expansion: this might be 1:100, 1:10 or even neat. Seed virus is harvested after a minimal number of amplification rounds (2-3), centrifuged, filtered to remove cellular material and any extraneous contaminants and frozen down in aliquots.

Small quantities of harvested seed-stock are used to inoculate the manufacturing cell-line (which may differ from the seed-stock cell-line) or a large number of embryonated eggs. To obtain a reasonable volume of virus it is typical to utilize up to 300-400 eggs, or to propagate the virus in a small bioreactor dosed with cells from the cell bank.

Inoculated eggs are typically incubated at 35°C for three days before being harvested and centrifuged down. Virus suspended in the supernatant is further filtered at 0.4µm and subsequently through a 0.22 µm filter to remove the majority of the remaining egg residue, proteinaceous clumps and any bacteria, leaving the supernatant containing only challenge virus. Finally, virus is aliquoted out into 2ml vials before snap freezing in liquid nitrogen.

A manufacturing process as described should generate sufficient virus to fill about 2,500 vials; these are stored at –80°C to ensure long-term viability.

The entire manufacturing process is designed to ensure the minimal number of passages that yields the highest possible viral titer. As the virus has not been passaged or amplified multiple times, the final agent should remain highly infectious to humans, and not have undergone attenuation. The ideal challenge agent should be very close to the wild type (WT) virus, as the aim of a human challenge trial is to reflect real-life infections in the community.

Testing the virus
It is important to note that, as well as the requirements to grow the virus under cGMP conditions, a number of tests must be performed before it can be used to infect human subjects in a challenge trial. These are similar to those involved in a vaccine testing program: for example, the presence of any adventitious agents must be excluded, which can be carried out by further incubation in embryonated eggs, in mouse studies and potentially in other cell lines that support contaminant human pathogens.

Such studies should show up whether any other infectious agents are present, which may include human viruses such as HIV, herpes and hepatitis and also, if it has been grown in eggs, the avian leucosis virus. Other pathogens that might be present include spiroplasma and mycoplasma, which are a type of bacterium without a cell wall that are common contaminants in cell lines. Toxin tests should also be performed to rule out bacterial, fungal and other toxins that may have leached into the product.

Genetic testing is employed to ensure there is nothing present other than challenge virus DNA. Next generation sequencing (NGS) can confirm the purity of the virus and assist in determining its homology to WT. Even with 100% homology with the original candidate virus on sequencing, there is always a chance that fragments of DNA from the egg or from adventitious agents may be present.

Before the virus can be given to humans (first-in-human or FIH studies), it must undergo animal testing to check for unusual, unexpected or unwanted pathological changes, as well as the severity of the disease induced. Influenza animal studies usually involve female ferrets, which represent a good model for human influenza. The ferrets are given the virus via intranasal inoculation and the course of the resultant disease is monitored.

Warning signals of unusual pathogenicity in mammals may include a very high temperature, vomiting, diarrhea, decreased or altered visual acuity and systemic problems such as falling blood pressure. The ideal challenge agent should only cause limited symptoms, akin to a bad cold or a mild dose of influenza, with a zero or vanishingly small mortality rate, even in vulnerable populations. If nothing severe or worrying is seen, the ferrets are sacrificed after eight or 10 days and pathological sections are taken to examine for virus in the lung, brain, heart and other major organs. For the virus to be deemed fit for use in human studies it should not spread widely from the nose and throat and cause respiratory illness only, with no dissemination into bodily organs.

Under normal circumstances, if the challenge virus being tested is a seasonal H3N2 or H1N1 strain, there should be no problems with only low-grade infection likely to develop. In contrast, if the virus is a wild type H1N1, pandemic strain, such as the H1N1/2009pdm, ferrets and humans may develop lung lesions and a profound lower respiratory tract infection. Such complications rarely occur with normal, circulating H3N2 and H1N1 strains. Unattenuated, pandemic H1N1 or H3N2 strains or potentially pandemic H5, H7 and H9 should not be used as challenge agents due to the risk of triggering fatal cytokine events.

The recently-developed SGS 3C H2N2 challenge agent (A/Belgium/4217/2015) has passed all of the above criteria following extensive testing and its profile appears to match up to the ideal virus. A/Belgium does not cause lower lung lesions or other adverse events, and gives a classic influenza-like illness, with a raised temperature, shivering, decreased movement, and reduced appetite in ferrets. On pathological examination of animals, there were no lesions other than those usually associated with a moderate upper respiratory tract infection.

Human dosing
With the evidence from the initial work in cell lines, the pilot and scaled-up manufacture in eggs (or a cell line), the testing for adventitious agents and other contaminants and the animal model safety test, a dossier can then be submitted to the regulatory authorities for its use in human subjects.

Viral challenge testing in humans is a new way to obtain proof-of-concept for the effectiveness of an anti-viral drug or vaccine early in Phase I, after initial safety and PK/PD assessments, and can assist in faster go-no-go decision making. Viral challenge testing clinical trials must obviously be conducted according to the Good Clinical Practices (GCP) within a dedicated clinical pharmacology unit which has hospitalization beds within a biosafety level 2 Q compliant quarantine facility.

The final part of the testing process is a titration study in human subjects to determine the titer that gives the optimum attack rate and symptomologies. This characterization study typically involves three different titers being dosed intranasally to cohorts of between 12 and 24 subjects in three arms to determine the optimal dosage level. The consistency of viral shedding is assessed via nasopharyngeal swabs taken twice a day, and symptoms that develop are recorded using a symptom scorecard (SSC).

The optimal symptomology reflects the usual influenza cycle of nasal congestion, coughing, raised temperature and lethargy, none of which should be too severe. If all of these develop and there is a consistent viral shedding profile after inoculation (with a good area under the curve) this indicates an effective infection. A consistent shedding pattern with few ‘blips’ and a high peak titer allows the effect of a drug or vaccine against the virus to be assessed in subsequent challenge trials.

There are ongoing discussions about how symptom scores should be taken and interpreted, and it is likely that there will be some changes in human challenge trial endpoints and outcomes in the near future. But, however the studies are performed, and regardless of the way the results are assessed, the ultimate aim is to emulate the infection that happens in the community, so the challenge virus can be used to estimate the effects of novel interventions.