Interview: Dr Satyajit Rath

‘India needs to spread its bets on vaccines’: Dr Satyajit Rath

Print edition : June 18, 2021

Dr Satyajit Rath.

Interview with Dr Satyajit Rath, Indian Institute of Science Education and Research, Pune.

In the wake of the pandemic, countries and companies have raced to develop and deliver COVID-19 vaccines at an unprecedented pace. Vaccines have been developed across several technology platforms, each of which has disadvantages as well as advantages. As the pace of India’s vaccine drive stalls dangerously, and even as the possibility of “vaccine escape” variants cloud the prospects for the future, there is an urgent need for countries to stay at the cutting edge of vaccine design to counter the pandemic.

Navigating through the apparent clutter of vaccine technologies, their design methodologies, and their potential drawbacks, Dr Satyajit Rath, eminent immunologist, physician and pathologist, spoke to Frontline. Rath is now associated with the Indian Institute of Science Education and Research (IISER), Pune. Excerpts:

There are several ways to make vaccines. What determines the choice of technology platform? How do they reflect the disease perspectives that a specific vaccine seeks to address?

In the first place, all vaccines are biological products. There is a fundamental difference between traditional drug manufacturing and small molecule drug manufacturing and vaccine manufacturing. That is why many of the biological pharmaceutical manufacturers, rather than small molecule drug producers, find it easier to manufacture vaccines.

The second point is that at the manufacturing level what the disease is is really immaterial to the manufacturer. That is true of small molecule drug manufacturing too. It is the integrity of the chemical synthesis and purification processes that are important. What the drug will eventually be used for is immaterial as far as the manufacturer is concerned.

I understand… that there is an element of being agnostic…

Yes, to the ultimate usage. There is a whole generic suite of manufacturing technologies involving biologics that are brought into play. But the specific nature would depend not on the disease they seek to address but on the vaccine design platform they [the manufacturers] seek to use.

What drives the choice of a particular platform?

Ah! This is where we come to a set of specific decisions in which the biomedical science and the manufacturing design and processes intersect with each other. Let me explain this. Vaccine design platforms fall into two broad categories. Essentially, all vaccines in some form are mimics of the original infectious agent. They are basically designed to fool the body’s immune system into making a response. We know that the vaccine is not going to cause any serious damage, but you are still providing enough cues to the immune system for it to think that it is worth responding to. Essentially, you aim to deliver to the immune system target molecules that resemble the original infectious agent. Whether it is one target or many is immaterial.

Also read: Misplaced optimism as COVID numbers go down

The vaccine design falls into two categories. In one design you make the target in the manufacturing process and introduce it into the body. In the other, during the manufacturing process you make the genetic code that specifies the target, then you introduce the genetic code into the body and let the body make the target and respond to it.

Can you please explain this process?

Hopefully, if I explain using the specific platforms that fall into the categories as examples, it may be easier to understand. The first is manufacturing the target outside and then introducing it. Let me give you two different COVID-19 vaccines in two different forms. One is Covaxin, in which you are making the target, in this case the whole virus, in the manufacturing process and then introducing it into the body. The other vaccine in the same category is the Novavax vaccine design. In this case, the idea is not to make the whole virus but just the spike protein as a recombinant protein and introduce it in an appropriate formulation into the body. In both these cases, although the manufacturing processes are radically different, what you are doing is manufacturing the target, formulating it and then injecting it.

Completely different from this targeting process are the RNA and DNA-based COVID-19 vaccines. The first are the ones by Moderna and Pfizer-BioNTech—the mRNA vaccines. They make the RNA chain that codes for the spike protein. The mRNA is then packaged so that when it is injected and gets to the cells, the cells will simply translate it into proteins. Basically, this process is about manufacturing the genetic code, not the final product: introducing the genetic code into the body and letting the body make the target and respond.

This is also the case with the Zydus Cadila vaccine under trial. In this case, the attempt is to use DNA instead of RNA, but the principle remains the same. The DNA molecules are first introduced into the cell. It is then read by the RNA molecules, which convert it into a specific protein.

The third platform in this category is the widespread adenovirus vectored vaccines. (Adenoviruses are double-stranded DNA viruses which cause mild respiratory and gastrointestinal tract infections. They are considered excellent vehicles for delivering target antigens in mammals). Essentially, you are introducing the SARS-CoV-2 spike protein genetic sequence into the adenoviral genetic sequence. Then you are using the adenovirus as a delivery vehicle to get into cells. Once it gets into the cells, the same process repeats itself, as with other platforms. In all these three cases, what is introduced into the body—-that is, what is being manufactured—is not the target but the genetic code of the target. The vaccine consists of the genetic code introduced into the body.

However, if you take the protein subunit or viral particle or the whole inactivated virus, or even, eventually, infectious viruses—all of these methodologies use a target that is made in the manufacturing process and then introduced into the body. The reason I am making this categorisation arises from the fact that the manufacturing constraints are quite different in the two categories.

What in the nature of the biologics makes the manufacturing platforms different or difficult?

They are different for an interesting biological reason. When you make the code, it does not have a shape, it only has a sequence. The exact shape of the RNA or the DNA really does not matter because it is going to be read as a sequence, one after another. Therefore, all you have to ensure is that the sequence manufactured is right.

Also read: The fiasco that is India’s COVID-19 vaccine policy

On the other hand, to make the target in a manufacturing process and then introduce it into cells, we need to make sure that the target, a protein, is folded into a three-dimensional structure which resembles the original virus as closely as possible. Otherwise, you are going to create an immune response that does not recognise the virus. This has been Novavax’s problem. Recall that it has twice delayed its clinical trials.

One of the reasons is that it has had protein-folding problems. That problem is shared with all other biologics manufacturers such as those manufacturing monoclonal antibodies. All of them use protein manufacturing technologies. All of them share the problem of having to make certain that the final folded shape of the protein is accurate and consistent.


Getting that right is problematic. Is that what you are saying?

Exactly. You now begin to see why the RNA vaccines came up much earlier than the protein vaccines. The only way you can be reasonably certain—in a manufacturing sense—that the folded shape is good is by actually growing the virus! If you take a cell line and grow the virus — SARS-CoV-2 grown in a cell line—in a manufacturing process, then you know it is correctly folded because it is made as part of the virus construction. This is the reason why the inactivated virus vaccines were designed as such and this is why they, along with the mRNA vaccines, arrived earlier. Note that the Sinopharm and Covaxin inactivated viral vaccines thus arrived earlier than the others.

Where does the Oxford AstraZeneca vaccine stand in this categorisation?

This vaccine is an adenovirus vaccine. It falls in exactly the same category as the RNA, DNA and the adenoviral vector vaccines. Instead of chemically packing the RNA—which is what the mRNA vaccines are—here it is biologically packaged into the carrier adenoviral vector. But here, you don’t have to worry about whether the SARS-CoV-2 target that you have introduced is correctly folded or not. In a manufacturing sense, the RNA, DNA and adenoviral vector vaccines are much more robust than the protein subunit manufacturing process.

Why would someone choose such a challenging option at such a time?

That depends on whether you are asking why protein folding is a problem.

No, I am asking why would someone choose such an obviously more difficult path…

That is a great question because the likes of Pfizer and AstraZeneca keep hoping that nobody will ask this. If you look at the amount of spike-directed antibodies generated by the Novavax vaccine candidate in people, those amounts are larger than the amounts generated by the mRNA and adenoviral vaccines.

How much more in terms of quantities?

(Laughing). I am laughing because I knew you would ask me this. The point that I keep making is that we have adopted a neoliberal solution to a public health problem. Thus, none of them [vaccine manufacturers] is willing to conduct comparative trials with each other. As a result, we do not have any comparative figures at all. I am basing this simply on the preclinical data—what happens when you immunise mice—with mRNA versus DNA versus whole protein vaccines. In mice there can be 4-5-fold higher levels in antibody levels. This is perhaps why a company like Novavax, whose forte is protein-based design, knows that if it can solve the protein folding problem, it will have a vaccine that can generate antibody levels that are far higher than the others.

Also read: COVID-19 vaccines: Trials & errors

Also, keep in mind what the ICMR [Indian Council of Medical Research] is saying about the so-called Indian variant, that 50 per cent of the vaccine-generated antibodies bind to the “Indian variant”. This is a complex matrix, based on how much of the antibodies bind to the variant and how much are the total antibodies generated. If the total is very high, then even if a fraction binds to the variant, it will be enough to provide protection. So, there is a lot of utilitarian complexity.


We are in a situation of an acute shortage of vaccines. What do you think is really the problem for companies in scaling up capacity after having achieved breakthroughs in developing the vaccines?

Let me start by being rude. I do not think there has been any breakthrough in any of this. Keep in mind that the mRNA and the adenoviral vaccines, the major vaccines being used the world over, are tweaked versions of vaccines that these exact same groups had developed, not today, but 10-15 years ago, against SARS-CoV-1 and MERS-Cov (Middle East Respiratory Syndrome). All of them had been to some extent road-tested. So, it is not as if we have miraculously made vaccines within months. It is just that we had been doing this sort of stuff for a long time but we had decided to let the market decide whether to complete the trials. And, because those viruses were no good at person-to-person transmission, simple physical separation, care and isolation terminated those little outbreaks. Thus, advanced vaccine prototypes were sitting on the shelf. All that Sarah Gilbert’s group at Oxford did was to take it from the shelf, dust it off and tweak the sequence from the SARS-CoV-1 spike protein to the SARS-CoV-2 spike protein. And, hey presto, you have the vaccine. It is important for all of us to recognise that this was not a major breakthrough. This is the outcome of cumulative science. This is the outcome of sustained global investment in open-ended efforts.

But that said, the scale of constraints would be different for each of these platforms. But this is where we get into separating the matrices. So, let me try a two-dimensional matrix or a kind of SWOT analysis, where we are talking about advantages and limitations.

Think of the Covaxin technology platform. You grow the virus in tissue culture and inactivate the virus. Essentially this is what Louis Pasteur’s students used to do in animals while developing the rabies virus—a century-old technology. It was actually grown in animals. Even during our childhoods, the rabies vaccine was developed in infant mice. But the principle remains the same: you grow the virus, you purify it and then inactivate it and then package it for injection. Growing the virus in animals is easy because it is low-tech maintenance for an assembly line. But the difficulty is that you get a lot of animal product contamination in the virus preparation; this was the problem in our childhood when we used to have 21-shot vaccines for dog bites. The answer to that was to clean up the virus growing technology. The second step in cleaning up the virus growing technology is what is done even today. Every year the world makes completely new influenza vaccines. And, those are made in the extraordinarily ancient, stable, low-tech method of manufacture which is this: grow the virus, inactivate it and package the virus for injection. But in the case of influenza vaccines, you don’t grow it in animals; instead, you grow it in chicken eggs, which is cleaner and easier to maintain than in growing the virus in live animals. That is the technology.

What are the constraints to scaling up?

You can see that the growing of the virus is fairly low tech as long as you are growing it in eggs. But as soon as you bring it to cell culture in bioreactors, you up the ante as far as the technology is concerned because growing cells in bioreactors is a much more finicky process. They have to be grown in sterile conditions, your nourishing fluid has to be available, it has to be just so, it must have just that many ingredients, all of which must be readily available. As a result, supply chain problems multiply because you need ancillary industries that ensure reliable supplies of materials and equipment. Sterility becomes a problem in the bioreactors because the contents are nutrient-rich. If sterility is breached, all sorts of fungal and bacterial stuff begin to grow, and upset the whole thing.

Also read: What are the criteria to compare vaccines?

As a result of all this the whole thing becomes essentially a biologics manufacturing technology. Remember that monoclonal antibodies are made in cell lines exactly like these. Any company that makes biologics can make vaccines.

But here is your problem: what you are growing is an infectious virus. If I grow infectious viruses in very small amounts it is not a problem. But if I grow a larger amount in the same place, an even larger amount, and a still larger amount, the potential catastrophic biosafety problem begins to amplify. Between growing the virus, harvesting, cleaning it up and inactivating it, you are left handling live infectious and dangerous viruses. And, the problem is that biosafety containment needs to be properly and robustly backed up. I don’t mean to create panic with an analogy that I may be careful to use in public, but you are creating exactly the same kind of a situation that a nuclear reactor poses. It is a low-likelihood event, but the consequences can be catastrophic. This is the crucial difference between growing infectious SARS-CoV-2 virus in the manufacture of inactivated vaccines and growing adenovirus in the adenoviral vector vaccines because the latter virus is not dangerous. So, it does not need the containment that the vaccine manufacturing process from actual infectious virus requires. This is a great limitation of the infectious virus technology in scaling up.

How have the other manufactures of infectious virus-based vaccines, such as Sinopharm, coped with this problem?

They have not put all their bioreactors in one location. They have spread them out. Basically, you spread them out and limit the amounts handled in a single location in order to reduce risk. Geographically spreading out the production makes it possible to reduce risk in a single location.

This is the reason why many of us, including people like me, have been for long puzzled. Here is the problem: the Indian Council of Medical Research made this vaccine and gave it to BBIL. Why did the government not give the vaccine to a dozen different biologics manufacturers? All it needed to have done was to ensure geographical decentralisation, while ensuring large-scale production. What was the problem in doing this? It is only now that they are talking about this.

But what about the other vaccine platforms, what are their problems in scaling up?

As far as the mRNA vaccines are concerned there is really no biological processes involved. There are no live cell processes involved. It is much more of a chemical technology. As a chemical technology, it in some ways makes manufacturing easier. You do not have to worry much about nutrient fluids, sterility so much—you always do, but not so much. But on the other hand, the number of building units you need is enormous. You are building mRNA from scratch, you are building massive molecules from scratch, using enzymatic as well as non-enzymatic processes, none of which is easy and straightforward. You can use biologic pathways but then you come to biologics technologies, and those are not easy. In all this a fair amount of detailed know-how in terms of tricks of the trade—I am very reluctant to use the term intellectual property—is involved. But tricks of the trade can be understood and assimilated if there is a level of upskilling already available in the manufacturing technologists human resource pool. All of these become potential roadblocks.

In fact, quite apart from what Bill Gates has said about vaccine manufacture and intellectual property rights and manufacturing in the Global South, I see it to some extent as the white man’s burden kind of perspective. But, apart from that politics, he has a point: it is not simply patents, it is these tricks of the trade, ability to assimilate and operationalise the tricks of the trade. This depends on whether you have sufficiently upskilled human resources in the manufacturing sector.

You have to learn to do it…

You have to be able to learn to do it. You must have a sufficiently sophisticated background and capacity already in order to be able to learn to do it. This is a limitation in building the mRNA platform.

Also read: ‘Variant tracking an essential part of long-term strategy’: Satyajit Rath

Moreover, there is a short time-window problem. As mRNA manufacturing base expands, it is in the nature of things that the manufacturing technologies will spread to the Global South in a relatively short time. But there is also the issue of QA-QC (Quality Assurance and Quality Control) that for mRNA sequencing have to be extremely finicky because you want the exact arrangement. And, because you are building that sequence batch by batch, unless you use biological processes you need a lot of step-side quality control. This is because you are building a long chain of the RNA brick by brick. So you need to run the QA-QC protocol every few bricks at a time. That makes for a slow and cumbersome component of the manufacturing process. These are the issues with the mRA technologies.

Let me go further. When the RNA is introduced into your body, there are so many enzymes that damage the RNA. Remember, the RNA is a very delicate molecule. This requires that the RNA is packaged. Packaging requires two properties: one, it must protect the RNA, and two, it must help to deliver the RNA inside cells. This is why the nano lipid formulation that Pfizer and others talk about—that their vaccine is a nano lipid formulation—is essentially encasing the RNA in a fat droplet. This is to ensure that water and water-based damaging agents do not get to the RNA. And, since all our cell linings are made up of fat, this fat and that fat can sort of stick together and allow the RNA to get absorbed. The fat packaging is highly specific and with high degree of purity in terms of components used. This again causes you to run into supply chain problems. You need not only a highly specialised, but a diverse supply chain to run the manufacturing process.


Are mRNA platforms easier to manage because the supply chains are more closely integrated spatially as well as qualitatively?

I would think that the mRNA platform would depend much more on a diverse supply of fine chemicals than biologics technologies. It is in the nature of the development of these technologies that, inevitably, there has been a geographical focus in the availability of these intermediate fine chemicals. But I suspect that this, like the tricks of the trade advantage I mentioned earlier, is temporary. What does it take to set up a fine chemicals ancillary industry? The Global South does not set it up because it does not see a credible market for it. If there is a credible local market for using mRNA technologies that source materials from them, they will set it up. The advantages that the entities in the West enjoy now is likely to be transient, a few years at the most.

Based on the description I have given of the manufacturing platforms—between using infectious viruses at one end and making the RNA at the other end—you will appreciate that the major technologies that are being used and spreading rapidly are those based on adenoviral vaccines. This is because the adenoviral platform does not involve a chemical manufacturing process. Once the innovator has designed and built the adenovirus and the cell line for it to grow in, all that a generic biologicals manufacturer has to be given are the cell line and the adenovirus. So, someone like the Serum Institute of India [SII] can keep upscaling its capacity, unless interrupted by fire. There is no real difficulty, pretty much anybody can do it. There are manufacturers in South Africa, Argentina, Malaysia, Thailand, Korea and many other countries who can do this. This has become a widespread technology platform because it is built on the basic manufacturing skeleton of widespread biological drug manufacturing of cell line and of products in bioreactors.

So, is it just the exclusive licence given by AstraZeneca to SII that has been a hindrance to the wider dispersal of vaccine capacities in the Global South?

Aha! It is not a matter of exclusive licence given to SII. It is a matter because SII is a subcontractor of AstraZeneca. Basically, AstraZeneca is holding the IP and has contracted SII to manufacture it in India. Every dose that goes out of SII’s doors has the approval from AstraZeneca.

Let me rephrase. If not for this legal stipulation, would there not have been many more Indian companies manufacturing the vaccines?

It is a mystery why ICMR, which is controlling the IP on Covaxin, chose to give it exclusively to BBIL. From a public health perspective, that would have been the sensible thing to do. The so-called AstraZeneca vaccine, which I never refer to as the AstraZeneca vaccine—I always call it the Oxford-AstraZeneca vaccine just as I always refer to the so-called Moderna vaccine as the NIH-Moderna vaccine—was developed at Oxford. Sarah Gilbert’s team has been working on adenovirus vector vaccines for 25 years, funded entirely by U.K. taxpayers. In fact, Sarah Gilbert and Andrew Pollard made the announcement last March that they were developing the vaccine. They also said it would be available in the public sector.

Also read: Ramanan Laxminarayan: ‘Vaccination the only way out of the pandemic’

Why was the property of the people of the United Kingdom sold to AstraZeneca? Once the British government sold it to AstraZeneca, why would AstraZeneca contract more manufacturers to produce the vaccine than it needs? After all, for the for-profit sector it is always preferable to have a desired product in short supply. That is the logic of the marketplace. Why hasn’t AstraZeneca contracted another 35 different companies which have proven capacities the world over by simply giving them the protocol, the cell line and the adenovirus? That is all they need to make the vaccine. The short answer to that is if this had happened AstraZeneca would not be as important, where people go begging to them. The public sector, the people and governments threatening companies is paradoxically a measure of the pivotal importance these companies have acquired. That will translate into non-financial clout.

So, we have two models—one implemented by AstraZeneca with Covishield and the other by the Indian government with Covaxin—which are mirror images of each other.

Pretty much. Can BBIL manufacture the AstraZeneca product? Of course, it can. If it can make a cell line and grow a virus from it, it can grow another cell line and grow another virus. In fact, BBIL has a separate deal with the Washington University School of Medicine in St. Louis to make an adenoviral nasal vaccine. It is already growing an adenoviral cell line. This is just empirical evidence to support my contention.

What may be BBIL’s constraints?

I suspect the increasing requirements for biosafety of an infectious virus places an additional constraint on upscaling. This is why, I think, BBIL has very reluctantly allowed ICMR to license other companies to produce Covaxin. But that is going to take time. There is a learning curve involved in all this. For pity’s sake, Haffkine Bio-Pharmaceutical Corporation Ltd is, I am sorry to say, an utterly moribund company, like many other vaccine PSUs. And, this has happened because of the way governments have treated them over the last 30 years.

Regarding Covishield, I suspect what Adar Poonawala has been saying, about not having the money, may probably be correct. Why did the U.K. government, the WHO, the Government of India, the United Nations, not intervene and tell AstraZeneca: hey, Zydus Cadila can manufacture, so can Cipla, Torrent, Dr. Reddy’s, Biological E, Shanta Biotech, all these companies can manufacture, so why don’t you distribute the licences? Similarly, South African, Korean, Argentinian companies and others around the world could have been asked to produce. Why was this not done? The short answer to that question is that this is what happens when you apply a neoliberal solution to a public health problem.

The scale of human suffering puts more and more pressure on the companies. But the companies are not pressed by politics alone, they are pressed by actual suffering. Until the suffering is visible, pressure does not grow on the companies. And, that is what we are seeing.

Even if they want to scale up, the supply chains will take at least 2-3 months to set up. The bioreactors have to be imported. All this will take time. Setting up and upscaling the infrastructure in PSU units will take time, money and human resources. Equipment, setting up protocols, testing, etc., will take time and money.

Also read: India’s vaccine saga continues with COVID-19 providing an opportunity to boost vaccine literacy

The mRNA technologies appear to be succeeding and they have an interesting future manufacturing trajectory. They are going to be useful for not just vaccines but for other drug-like products. Now that there is proof-of-principle out there, the mRNA technologies are going to be part of the ecosystem. I think over the next three to five years, we are going to see supply chain components emerging in the Global South. We will begin to see fine chemical components that go into mRNA platforms develop as ancillaries to the supply chain in these countries.


Given the overhang of neglect of the public sector in biologicals in general and vaccine manufacturing in particular, what would it take to revive them? And, how long would it take?

That can be done in a year. Because we have public sector vaccine manufacturing entities that already exist, even if in a moribund state, it would take only political will and financial muscle to revive them. Enthusiastic recruitment of skilled professionals, re-equipping and revitalising them and expanding them should be possible within a year.

If you had a 3-5 year perspective plan for these units, as well as involving private companies operating in this field, how much would you need to spend for building such an ecosystem?

I honestly have no idea about this. But if SII was paid about Rs.3,000 crore for ramping up its capacity, it may perhaps cost about twice as much to expand capacities in public sector units.

Expansion of capacities would also give leverage to the government in its dealings with SII...

Exactly. Not just with SII, it will give leverage with AstraZeneca or with any other vaccine company.

Would some platforms be better suited to address the pressures from new vaccine escape variants?

In a sense, the way we are measuring vaccine effectiveness is like the man who lost his ring in the darkness but is searching for it under the light because that is where the light is. What we are trying to detect is what we already know how to detect, which are the so-called neutralising antibodies. This is how we are detecting and correlating the effectiveness of all the vaccines. But neutralising antibodies actually interrupt the transmission cycle. So, it is logical that if the virus changes to improve its transmissibility, it is almost as a byproduct that it will have some effect on the neutralising antibody mechanism.

Does India need to spread its bets on vaccines?

Absolutely. No question about that. We badly need to spread our bets—platform-wise as well as capability-wise. Let me give you an example. I do think that the adenoviral vector vaccines have a problem. If we generate very high levels of antibody responses to the adenoviruses, it is possible that later the vaccines based on the same design might run into difficulties, particularly in populations that have been recently immunised with vaccines of the same platform.

Also read: India and COVID-19: Government's vaccination drive a total mess

This is because we are injecting a whole adenovirus. The body is generating response to not just the target but also the adenovirus. As a result, when the next time an adenovirus injection is given, adenoviral antibodies may prevent the target from getting into cells. We ought to be mindful of this problem and also invest in mRNA vaccines.

Let me point out that the Novavax approach, which is based on the protein subunit approach, is a much more focussed approach. It is a much more traditional biologics approach in the sense that you make a functional protein, mix it with an adjuvant and then inject it. It generates very targeted antibodies and does not generate these toxic cells. It is a very narrow platform, but it is not something we should ignore.

This is partly exemplified by what my friend Raghavan Varadarajan (biophysicist at IISc, Bangalore) is doing. What Mynvax is doing (and published in a journal recently) is making a Novavax kind of spike protein on a very biologics-friendly technology platform, but generating extremely high level of antibodies and is temperature stable. So, the vaccine formulation may not even need refrigeration. So, there are different advantages to different formulations. I am giving you this example in order to underline the fact that we need to hedge our bets by building as large and diverse a portfolio as possible.

Full disclosure: Satyajit Rath is an adviser to Mynvax, a company incubated at the Indian Institute of Science, Bengaluru. Mynvax was founded by Raghavan Varadarajan in 2017. It is developing a COVID-19 vaccine.