Proving variant-specific efficacy will be a logistic nightmare, very expensive: Gagandeep Kang

Influenza vaccines are decided based on strains recommended by WHO twice a year for northern and southern hemisphere vaccines. They require animal studies in ferrets and small safety and immunogenicity studies, based on measurement of antibodies. If we had a reasonable antibody measurement which could reflect protection, then we could potentially follow a similar strategy for SARS-CoV-2 variants too. And not every variant requires a new vaccine, says Dr. Gagandeep Kang.

The emergence of three variants — B.1.1.7 first found in Britain, B.1.1.248 first seen in Brazil, and B.1.351 first detected in South Africa — less than two months after the first COVID-19 vaccine was approved for emergency use by the U.S. FDA, all companies have begun making booster shots or tweaking the existing vaccine based on the new variants. Dr. Gagandeep Kang, Professor of Microbiology at CMC Vellore and a vaccine expert explains in an email the challenges in developing and testing newer vaccines to keep pace with the emergence of new variants.

What are the main challenges in developing and testing booster vaccine doses to address the new variants?

Each variant (and many more beyond the first three that have been widely reported in the last couple of months) will need to be evaluated to see whether the variant has mutations in sites known to affect the antibody response. Because we have good structural models now, this is feasible in silico (or studying structures on a computer). If variants are found that might affect the immune response, then we will need to validate this information in real life. Doing this in humans in phase-3 trials is going to be very difficult if not impossible, so we will have to develop methods that can be done faster. One example of an approach would be to challenge studies in animals, where animals immunised with the version of the vaccine to be tested are challenged with the variant virus.

There have been experiments in primates, but not with the variant strains, where it was shown that a low level of antibodies is required for protection, but also that if antibody responses are less than needed then cellular immune responses may also contribute to protection. So we could hypothesise that in case of a variant, what was an effective antibody response could become less effective in terms of antibodies, but the cellular immune system could come to the rescue, but this will need to be tested.

How easy is to produce and test a booster dose using the mRNA platform? Will the booster dose be required to be tested on thousands of participants?

Designing the vaccine requires the sequences of the variants. While we can predict which variants are likely to matter for antibody binding and neutralisation, there will be a need to validate the results. The design of the vaccine can be done in a few days, and production of the vaccine is a matter of a few weeks. But after that, knowing whether a vaccine works will require a testing strategy that has not been defined yet. It is highly unlikely that we will be repeating phase-3 studies for every new vaccine based on a variant strain, but we do need alternative regulatory pathways, and most likely these will involve hundreds rather than tens of thousands of people.

If we had an immune correlate of protection, where a biological specimen, most likely blood, from a vaccinated person could be tested to show that the person was protected from disease, that would make it easier. But would we need a separate correlate for each variant virus? At the moment, we do not know.

Can AstraZeneca/Serum’s vector-based vaccine be redesigned to develop a booster dose? How easy will it be to develop a booster dose using the platform?

Yes, the AstraZeneca vaccine is easy to redesign as a booster based on a variant strain, and the time taken will not be much longer than for mRNA vaccines.

Redesigning and making the vaccine is not the major hurdle. We need to figure ways of being reasonably sure that the vaccine will work in humans, and preferably against old and new variants.

This is feasible to do — in many vaccines the first dose of the vaccine induces a response to the virus on which the vaccine is based, called a homologous response. With further doses, the immune response becomes more able to recognise a wider range of viruses — this is called a heterologous response. What we would like to induce with the smallest number of doses is a protection against a wide range of viruses — old and new variants.

Will a separate booster dose be required for each variant in the case of Covaxin? Will the inactivated whole virus platform be better at protecting against variants and be required to go through all stages of clinical testing?

We do not know. What is known is that vaccines that are based on the whole spike protein work, and work well. For inactivated vaccines, we expect that immune response that is induced will be against many proteins, including the spike. Does this mean that the vaccine will be more protective? It is feasible that additional immune response might induce a T cell response that could contribute to further protection, but that has not been shown at this time.

At the moment, we do not know what protects and how well, but we do know that vaccines based on the spike protect against disease, so it is not essential to have an immune response against other parts of the virus to induce protection.If the most important protein for inducing protection is the spike, then if an inactivated vaccine was made using an older variant with one kind of spike structure, why should we expect it to protect against a virus with a different spike?

When a vaccine is based on a whole virus, and we need a new version of the vaccine, is that a new vaccine and does it require a full evaluation through all the phases of testing? So far the only vaccine for which new versions, including inactivated vaccines, have been permitted to be licensed based on limited human testing are vaccines where we have a good understanding of what constitutes protective immunity or influenza vaccines — where annually updated vaccines are made.

Will regulators have to make new guidance for SARS-CoV-2 vaccines? I think this is very likely and many stringent regulatory authorities, like the U.S. FDA, are already in discussion about the testing requirements which will help the regulators to adequately assess the vaccines.

In general, will booster vaccines be required to undergo phase-3 testing?

I do not think new phase-3 trials will be required — proving variant-specific efficacy will be a logistic nightmare and very expensive, so we will need to have better ways for regulators to evaluate new vaccines on any platform.

I think we will need phase-1 and phase-2 studies of safety and immunogenicity. It is likely that regulators will require inclusion of standards such as those developed by the National Institute of Biological Standards (which is derived from people who have been infected and recovered, but may need to be periodically revised or added to ensure that the standard or panel of standards cover the breadth of viral variants).

Can vaccine development and testing keep pace with variant emergence? 

Not every variant requires a new vaccine. The D614G mutation emerged in early 2020 and became globally dominant, but vaccines based on older variants protect, and protect well.

We can design a new vaccine in days, produce small amounts in weeks and test in weeks or months. The duration of testing requires a regulatory pathway — how much testing will be needed to approve a new vaccine. Manufacturing at scale can take several months, but again when a vaccine is a modification going into an established process, this is not impossible.

Will it become inevitable to update COVID-19 vaccines each year or frequently to keep it highly efficient against new variants?

It is not inevitable, but it is possible. I think it is more likely that with a primary set of two immunisations we might need to take boosters perhaps in a couple of years. After that depending on whether the virus settles down to be endemic and less severe or continues to cause severe disease flare-ups, we may not need a vaccine or need subsequent boosters.

Influenza vaccines developed each year do not undergo fresh trials? Why is it so and can the same strategy be deployed for vaccines produced against SARS-CoV-2 variants?

Influenza vaccines are decided based on strains recommended by WHO twice a year for northern and southern hemisphere vaccines. They require animal studies in ferrets and small safety and immunogenicity studies, based on measurement of antibodies. If we had a reasonable antibody measurement which could reflect protection, then we could potentially follow a similar strategy for SARS-CoV-2.

How well will a fully vaccinated person respond to a vaccine against a new variant? Will the original antigenic sin come into play and not be effective against the new variant?

Original antigenic sin depends on the antigen which induced the immune response, and as far as I know should not differ between mRNA vaccines or spike-based protein vaccine (unless the adjuvant differs).

With the mRNA vaccines, we know that the antibodies induced by older variants have decreased but not absent activity against the variants. Unlike other infections, where a worry is that the original response might trap the immune system into an ineffective response, that does not appear to be the case at this time for [SARS-CoV-2] infection or vaccination, but we will need to do the evaluations.

How much better will vaccines that target multiple sites on several viral proteins be? For example, vaccines that target not just the spike but also the nucleocapsid protein?

Based on what we know about the other coronaviruses, and now about SARS-CoV-2, the spike is the most important protein certainly for the antibody response. It is not clear what a nucleocapsid protein might contribute to humoral and cellular immunity. We cannot treat all variants as being equivalent and make predictions on what might or might not work. We will need experimental evidence from the laboratory in animals and in humans.

There are several approaches, including adoptive transfer experiments that might help us with answers to these questions, but to my knowledge, we are only at the start of working with variants.

How well will the strategy of combining two completely different vaccines for the first and second dose be against variants?

At the moment, we can only hypothesise and then proceed to test. The first vaccine to combine two different vaccines for a single pathogen was the Janssen vaccine for Ebola, which uses an adenovirus vector for the first dose and a different poxvirus vector (modified Vaccinia Ankara) for the second dose and this was licensed in 2020. So we have proof-of-principle, but are new to this. We now have the Sputnik V or the Gamaleya vaccine which usesAd26, the same vector as the Janssen Ebola and COVID-19 vaccines, for the first dose and the Ad 5 vector, same as used by CanSino, for the second dose. We have data showing high vaccine efficacy, and there are also studies combining the Ad26 with the AstraZeneca vaccine.

In the UK, studies have started combining the Pfizer mRNA and the AstraZeneca vaccines. These studies will only evaluate immunogenicity and not efficacy.  However, these types of studies are very valuable, because from the sera of individuals participating in studies of combinations of vaccines it will be possible to evaluate the ability of antibodies produced in vaccinated individuals to bind to and neutralise different types of variants which will give us a better understanding of approaches that may work well in vaccination programmes.

With SARS-CoV-2 deemed to become endemic in many countries, will the emergence of new variants at regular intervals become common?

Yes, we will continue to have variants. All RNA viruses have higher rates of mutation than DNA viruses, and among the RNA viruses coronaviruses mutate relatively slowly. Nonetheless, SARS-CoV-2 has been accumulating mutations at a rate of one every two weeks or so. Recently, from California, we have also had another form of change, called recombination, where two viruses can combine with each other to produce a hybrid virus. The only reason that we are able to find these variants now, is because we are looking for them and sequencing more than we have ever done before. Now that we have a sense of how and how fast change happens in SARS-CoV-2, we should expect to continue to have variants and have the laboratories ready to sequence so that we stay a step ahead of virus spread.

The more important question is will variants matter and why? In my view, there are five reasons to track variants and understand what, if anything, we need to do about them. These are:

  1. Will the mutation result in a change in our ability to detect the virus? In other words, will our tests fail to detect infection? Since the most widely used test is RT-PCR, we might need to change the test if this happens.
  2. Will the mutation result in increased transmission? We know that this is happening with the new variants. The new viruses are getting better at sticking to their receptor protein, and that makes it easier for them to get into the host cell. This kind of evolution might continue to happen, since it makes it easier for the virus spread, which is an advantage for the virus.
  3. Will the mutation result in worse disease? Usually, as viruses evolve, they tend to cause milder disease and not more severe disease. It is in the interest of the virus to spread easily but not kill its host, because spread from a host that dies is not possible. There are some data that indicate that some of the new SARS-CoV2 variants cause more severe disease, but that has not been definitively proven yet. We need to track patients carefully to understand what is happening.
  4. Will the mutation allow the virus to escape treatments? At the moment with monoclonal antibodies, it has been shown in the laboratory that variant viruses can change enough to prevent antibodies from binding. We do not know whether the variants will escape treatments, because we do not have any direct antivirals that work well. Remdesivir is an antiviral, but not very good even against the older forms of the virus.
  5. Will the mutations allow the virus to avoid the immune responses induced by vaccines? Not all vaccines — we have data showing that most vaccines work well against the 501Y.V1, which is also known as the UK variant. On the other hand, we also already have data that show us that vaccines designed against older versions of the spike protein have lower efficacy against the new variants. There are data showing that the Janssen, Novavax and Astra-Zeneca vaccines all work less well against the 501Y.V2, or the South African variant.

Even if we quickly achieve herd immunity through vaccination, will the emergence of new variants in some part of the world threaten all the gains?

I do not think that with the new variants we will lose all the benefits of protection acquired through infection or vaccination. A variant virus is exactly that — a variant, a mostly related, slightly different virus, not an entirely new one.

These are early days of variants that partially escape the immune response. We know people can be re-infected and that immune responses induced by the older variants do not prevent mild or moderate illness, but we still need data on the more important outcomes of severe disease. When we have those data, we will be able to make better predictions and think about the future. At the moment, all we can offer is opinions based on very limited information.

Published in The Hindu on February 20, 2021