Moderna, Inc. (MRNA) Earnings Call Transcript & Summary

Good afternoon, everyone. Thanks for joining us. I'm Salveen Richter, biotechnology analyst at Goldman Sachs, and we're pleased to have Moderna with us today. And with us is Stéphane Bancel, CEO of the company.

So Stéphane, to start here, just given the COVID-19 pandemic, Moderna has the opportunity here to demonstrate the potential of the mRNA platform as it relates to vaccine development over the course of a few months versus what would have typically taken 2 to 5 years or maybe even more. Why is Moderna positioned to successfully develop a vaccine and supply it to the global population?

Yes. So Salveen, good afternoon, and thank you for having us. I think a few reasons. I mean first, we believed since we started the company, that mRNA could be potentially the best technology that we've since Monsieur Pasteur started doing vaccines 100-plus years ago. And the platform -- the piece that is fascinating about this technology is you use the sequence, genetic sequence of a virus, to then put into our platform to make a vaccine. And because mRNA is an information molecule, once you figure out to do 1 vaccine, the next vaccine, you take the sequence as raw material, you drop it into your cassette, and then you can be flying. So I think the platform nature of what we have is very powerful. The second piece is this was not our first rodeo. This was our tenth vaccine to get into the clinic, and I feel much comfortable now because of all the things we learned in the clinic. We've dosed around 2,000 people. And we've improved the technology. We've improved the formulation system. We improved the manufacturing processes to try to find the best balance we can. Because in a vaccine, safety is priority #1. You give vaccine to healthy people. You want a well-tolerated product. And then, of course, is efficacy, like in any pharmaceutical products. And so these 10 vaccines, so there's actually more than 4.5 years of clinical experience with the technology around vaccine. The third piece, I think, is the manufacturing investments we've made. We decided years ago when we saw the first clinical data on our flu vaccine, if you recall, to build our manufacturing plant in Massachusetts. And we decided to do that because it was clear to us that given the first vaccine data were positive, because it's a platform, many more vaccine were going to work. Nothing else made scientific sense to us. And so we decided to invest in manufacturing because, like with any new technology, there's a lot of things you have to invent. Actually, you have to invent everything. And we thought that being able to do it ourselves with our own teams, that we are 100% committed on the upside and on the downside of making this work the best we could was very important to us. I mean think about the same company if we have had contract manufacturers, we might not be in the clinic yet, and we announced this morning that we are very close, in weeks away to start a Phase III study. The other piece where we got lucky, but again, as Monsieur Pasteur will say, only for prepared minds, is we had worked with NIAID with Dr. Tony Fauci's team on the coronavirus vaccine before, for MERS, the Middle East Respiratory Syndrome. So we've done extensive work for around 2 years, looking at a lot of different protein and antigen, subsegment versus full-length, full protein, and this was, as you can appreciate, very informative to us when we realized it was a coronavirus, early days of January. And then around January 10, 11, when we got the sequence published online. That gave us the scientific conviction that we had learned enough about coronavirus with Moderna's mRNA science and technology, to be confident that picking the full-length spike (S), not a fragment, given us a high probability of having good translation into the clinic. And the other piece, I think, was the team, which is -- that's, I think the beauty of biotech and smaller company with a focus, with kind of a work ethic. I mean if you think about the team, we have been 7 days on since early January. So many team members are pulling all-nighters to get things moving because we know every day matters. We know every day matters. And we've tried to think about how you can you shrink this development timeline as much as you can without taking risk on safety. The key for us has been taking business risk, investing, making material at risk ahead of data. That was a smart investment we thought because we know we have a platform, and we've seen all those vaccines before. And given the circumstances, taking a few million dollars of material risk when you have a $1-plus billion on our now a much bigger balance sheet, we thought it was a calculated risk to make because we knew MERS data preclinical look great. We knew across all 9 vaccines before it, very good succession from preclinical data to clinical data. So all that body of data help us make those decisions around business risks. And I think that's one of the things that people don't always appreciate about Moderna, is we have a true platform. And there has -- there is really no -- been not many platform in our industry, if you think about it, no. I would say [ RNA LAM ] on RNAi, what they do is, I think -- why the velocity of [ RNA LAM ] is I think to open the eyes of people that they got the technology right. They had ups and downs. We had ups and downs. Most people didn't see them because we are a private company or investors saw them, but the general public didn't see them because science never go in a straight line. But I think this is the best example and analogy I have to what Moderna can do is using genetic information, using molecular biology platform technology, that you invest a lot on the science to make your technology-ready, and then you have a velocity that is undoable using recombinant or small molecule.

Great. And assuming success here of the COVID-19 vaccine, what does this mean for the platform and the remaining mRNA medicines in your portfolio? You do have a pretty vast portfolio here. So can you put this in the context of technology and biology risk across the different modalities that you have?

Yes, that's a great question. So as we've always said, we think our 6 modalities or 6 application of a technology, not correlated from a technology risk standpoint because we use different lipid systems, different chemical matter for the lipids and a lot of time, different route of administration. The vaccine is intramuscular, we have intratumoral in cancer, we have IV products. I hope soon, we will have a respiratory product into the lung because of our Vertex partnership. And so those things are not correlated. But within an application, so vaccine now, we think that either COVID vaccine mRNA-1273 gets to the finish line for an FDA approval, we'll have very important read across the portfolio, because it will give a good sense for the safety of a platform because it's always the same chemistry for the mRNA, always the same chemistry for the lipid. And so we're going to run a 30,000 subject Phase III as we announced this morning, with that safety database on the Moderna vaccine platform. When we show you early human data, there should be great translation that, that product, even as a Phase I, which I think a lot of time, people who don't really understand vaccine because vaccine are basically buried in 4 big pharma companies: Pfizer, Merck, and Sanofi, GSK. But if you look at vaccine, as you are aware, because in Phase I, it's the same population, healthy people as Phase II, as Phase III, as commercial, you don't have kind of a surprise you see in the treatment side of the world when you go from healthy volunteers to patients. And as you know, from the data we shared from Andrew Lo from MIT, the probability of a vaccine, when you look at many, many platforms, is 42% after a positive Phase I to launch, which is the highest of any therapeutic area. But if you look then within the same platform, my belief, it's going to be much, much higher, which I think Phase I are going to be very important value inflection point for vaccine. If we can show high neutralizing antibody of, let's say, Zika vaccine, or CMV vaccine, and we've shown with COVID the safety database across 30,000 people, knowing it's the same technology, the priority of success of those vaccine are going to be very high. And the piece for us that is very powerful, which makes us we think as you -- as we do investments, in the capital allocation that we do is we've seen data from mouse translate to human. So think about what it means for us, when Dr. Hoge, Stephen Hoge, showed us with his team a new animal challenge data with high titer working in the mice, I believe, because what I've seen over those 9 vaccine, it will translate into human. And from Phase I in human, you will have 42% chance of approval. Assuming the industry average across all the platforms that have very different efficacy profile and across high-priority risk product because, as we know, all the CMV vaccine that have failed before us are in that 42%. All the HIV vaccine that have failed before are also in that failure rate, leading to 42%. And so I believe that the ability for investors who can connect the dots and look across the platform and do the work, will realize that when we show, like we showed for EBV, of our product that are still not in the clinic, a good titer of a right antigen in preclinical model, they will translate to Phase I. And if COVID is approved, they have a very high chance to translate to approval, which I think is such a paradigm change. I'm not aware of any other technology, where you can look at a mice study, again, of course, the challenges in animal model study, and you can say, okay, this one has a high chance to get approved.

Great. And given the resources focused here on 1273, how is that impacting the rest of your 23 product candidate portfolio? And if you were to call out any of those additional pipeline assets, what would be the major ones that you highlight?

Yes. So what we've done to ensure that we do both things right 1273, and you've seen the pace at which this program is moving and the focus it needs from the team, both from development, but also manufacturing scale up, which I'm sure we'll talk about, and the rest of the portfolio, we basically divided the teams. So we have a dedicated team that is just on COVID, very experienced development, late-stage commercial professionals. And the rest of the development team focused on what we call the non-COVID portfolio because we don't want to slow down CMV. We don't want to slow down Zika or the personalized cancer vaccine and so on. So that's how we organized, and we, of course, hired a lot of people recently to help us have the right capacity as well as capabilities to do those 2 things in parallel. In terms of the key problem I will highlight, I would say the CMV vaccine, obviously, because I would just talk across -- read across the portfolio. I believe CMV is a $2 billion to $5 billion annual big sell product. I believe we will launch CMV because, as you know, the CMV vaccine from Sanofi at 50% efficacy in its Phase II with only the gB antigen. We have gB and the pentamer, which we were not able to make. And if you ask any scientists, we need both. So do I believe we're going to have more than 50% efficacy with gB and pentamer, I do. And given there's no product on the market and the FDA before has approved vaccine with 50% efficacy like ZOSTAVAX from Merck, we will launch this product. So of course, the sales will depend on the efficacy, obviously, in the commercial ramp. But I believe, and we're getting the company ready to launch the CMV vaccine. It's a very important one. I think the EBV vaccine is another big one. EBV, Epstein-Barr virus is the virus that causes mononucleosis. There is more and more epidemiology data showing it is 1 of the leading potential root cause for MS. 90% of people that have MS are EBV-positive. I mean having had infection of EBV in their life and have had mononucleosis. The cancer opportunity for us is also very big. In cancer, we have 5 immuno-oncology drugs in the clinic. They either combine to KEYTRUDA or combined to Diovan from AZ. It's a complementary mechanism from those drugs, which are, of course, wonderful drugs for patients. And our goal there is to see, can we improve the efficacy rate of those drugs versus monotherapy. Our personalized cancer vaccine is in Phase II. It's ahead to where the CIS KEYTRUDA monotherapy, and the biology will tell us, is it better or not? And if it's better, we'll take it forward. If it's not, we'll learn the consequences and deploy our capital somewhere else. I think the next big wave is going to be the rare genetic disease. A bit, I will make the analogy to [ RNA LAM ] here. We are not competing. We are doing things that are very complementary. With RNAi, they turn things off, we turn things on with mRNA. We had a good human data with chikungunya antibody. We use the same formulation system, the same delivery route, IV, then our rare genetic disease, MMA and PA. And so I'm cautiously optimistic that when those get into the clinic, we should get some positive signal. And if one works, because we are careful at this stage, not to have 10 development candidates, there are a few dozen rare genetic disease coming from a protein deficiency inside the liver cell. If one works, all the rest is the same biology. There's low biology risk, 1 gene deficiency in novo gene. Again, we use DNA as raw material. We put in the same cassette and we go again. So what we have done in a vaccine, we're having those 10 vaccines in the clinic in a 4-year time line. I could see the same thing happen in rare disease, and we could literally build an Alnylam if MMA or PA are positive in the clinic within Moderna, just in the rare disease in liver space.

Great. So moving on to the COVID-19 vaccines. So 1273 for SARS-CoV-2. Last month, you actually disclosed qualitative data on this program where neutralizing antibodies were at or above what you've seen with convalescent sera. Could you briefly recap this for us? And then [indiscernible], I think, I guess, there were some questions about the lack of data included in the release and why you didn't -- why you released it in that form, perhaps you can just give us some color.

Sure. So let me give you a few data points on a few dates. So as you know, NIAID, Dr. Tony Fauci's team has been running the Phase I, and so we are running the study with some of them, the vials, they're running the study, collecting the blood, analyzing the blood, running the assay. And so on May 14, on the first day, we got a data update with a neutralizing antibody, then we got a PDF report, we spent a lot of time looking at it. We're, of course, very pleased with the data. And we had the Board on the Friday, a meeting to share the data with the Board. And we were advised, after a lot of discussions that, obviously, I won't go into it here, that we should release an interim top line data because of the maturity of the data. And one of the way we had was that this data, it was going to be very hard for us given the number of people outside the company aware of the data to keep the data in check. And obviously, given our legal obligation of Reg FD to share with investors, all investors, the same data at the same time, we were advised by inside counsel and outside counsel that we could not keep the data under wrap. And because, again, we don't own the data and NIAID, which I understand and fully respect, obviously, wanted to preserve integrity and the confidentiality of the data to enable a peer-reviewed publication in a top-tier journal. And as you know, top-tier journalists don't like to publish data that have been public before. What we agreed to do is to have our outside counsel and our medical team write the press release. The press release was reviewed by NIAID, we're back and forth the whole weekend, so that before the market open Monday, we can release the top line data so that if that data were to leak in the following days or weeks, we have had our obligation dealt with because it was very important for us to make sure that everybody had that data at the same time. So to recap the data, basically, what we had with this data cohort as of May 14 is we had the binding antibody for all the subjects. We were at or above sera. Those are interesting, but the one scientists always care about, and clinician is, of course, neutralizing antibodies, those that bind to the virus, at the part of a virus that goes into infecting the cells, if you want. And what we shared is what we had at the time, which is a small data set, but it is the first 4 subject on the 25-low microgram dose and the middle dose, 100-microgram. That's all we had. Why do we have more binding than neutralizing is because a binding assay is a quick ELISA. It's run very fast. And neutralizing assay takes a long time because it's a live virus assay, so we have to grow in petri dish, the virus, to see that you neutralize virus replication with the antibody that you got from the blood of those subjects. And the reason was it only 4 and 4 is because as it's typically done for vaccine by everybody, when you start a vaccine study in healthy people, you test a few people in what is called a sentinel's cohort to make sure they're safe. And then you pause it to make sure the safety of vaccine is good. And then when you have good safety, then you finish enrollment of that cohort. And so that's why if you play all the timelines, the data we had as of May 14 was only those first 4 and 4, so 8 subjects. We got very excited about it and very happy, and I said on the conference call when we walk people for the top line in the morning that we could not have been happier is because what we've learned about technology over the years in vaccines through the 9 vaccines before is that if you look at the data, the error bar between subjects is very tight. And I invite people that are not familiar, like you, Salveen, about our data, is to look, for example, at the CMV data, or the most recent, the Zika data. And what you see is that every dose level, the error bars are very, very tight. If you look at the adenovirus technology, which is going to be part of our competitive space with Inovio, AstraZeneca, Oxford and CanSino and with J&J and so on. If you look at that data, a lot of time, you see a bimodal distribution of the data where some people respond and some people, as it's believed scientifically that have already been exposed to a virus, don't respond as well because of the antibody response to the carrier that brings the information into the body. And so there, you see a very high -- very, very big error bars. And so given we know we have smaller error bars and we saw dose response. And as you said already, the 25 within that band or above and the 100 by definition, given dose response, were higher than that. We were very, very happy. We put the 250-microgram dose just to not miss the dose because, as you know, we went in the clinic with 2 days of work in silico, only on computers. And so we basically guessed the dose based on the previous vaccines we have in the clinic but as we've learned over the years, and as you have seen, the antigens react differently and created a different immune response because some antigen are more immunogenic to the human immune system. And so we put the 250-microgram just to not miss the dose in the Phase I because we could have been in a world we had nonneutralizing antibody at 25 and a few at very low titer at 100. And given what we saw at 25 and 100, we decided to drop the 250-microgram dose and the news of this morning, of course, where we shared publicly the update of the Phase III design now that we have alignment with the FDA through formal meetings in the last few days is that we are going to Phase III with a 100-microgram dose, which is well-tolerated. And our goal with picking that dose and has been, as you can imagine, many long discussions with our advisers and our teams because picking a dose is always a complicated process, is we believe this is the optimal dose to have very high efficacy at a very well-tolerated profile. Do I believe a 50-microgram could have worked? I do. Do I believe a 25-microgram could have worked in healthy people? I do. But when you are going after a virus that we don't know a lot of biology yet because it's still new, and every day, we're seeing new papers that inform us of new things, sometimes it goes in the other direction. We are like, "Look, this first vaccine of Moderna is going to be a big read across the platform, and so we need to solve for efficacy." I believe, based on the data I have seen, that we have a chance to maybe have a best-in-class vaccine in the field, and we don't want to be too cute about the dose. 50-microgram could have worked. I know I made the life of our manufacturing team harder because we need to make twice as much mass of product, and when you're talking about so many hundreds of millions of doses, billions of doses, it's a big deal. But again, solving for efficacy, I believe, in this business, is the most important thing we can do.

Great. And congrats here on the quick advancements. So when you look at that, when you look at the Phase II, which is ongoing now, how quickly can you roll? And how quickly can you -- can that data set from the Phase II read out? Maybe we'll start there and then talk about the Phase III.

Sure. So the Phase I, as we said, we're going to publish very soon. Doctor Tony Fauci has said publicly, it's weeks away that we have a paper out not months away. And I anticipate, like everybody in the field, we will put the submitted paper online right away, while it is peer-reviewed, so the whole community has access to the raw data and can make their own opinion before the peer review process. The Phase II, so we started at end of May. We announced this morning that the adult cohort, so 18 to 54 years old, 300 people is fully enrolled. The elderly, we enrolled the sentinel first 50 because we have less safety experience, and we agreed with FDA to do it this way. We're going to wait a little bit. I'm talking a couple of weeks, not 6 months, to then fully enroll that cohort very quickly after because we are screening people as we speak. So it is as soon as we get safety monitoring committee agreeing that the 100-microgram and 50-microgram in the elderly is safe to proceed, we'll go to 300 in a matter, I think, of days. I mean, looking across 13 days, I think, if I remember correctly, from starting dosing to fully enrolling. So if you play the movie forward to your question, it's a boost at 28 days, so a boost is going to be a month from now, I mean, from when they started. So we're basically talking a kind of early July for the boost. And then it's typical, you wait a few weeks to a month to take the blood draw, to look at the neutralizing antibody, you run the assay in a big batch, so you reduce assay variability. So it's worldwide in August, early September time frame, I think, we should be able to get the data out. We'll want, of course, also to publish it as well. But we know it's important data with an N of 600 is going to be very telling in terms of where we are landing kind of directionally for the Phase III. And to see also, can we keep, which is what I believe, the error bar super tight at an N of 600. So we'll share that data, I think, around the time frame, August, September.

Great. And then this morning, as you mentioned, you did announce on alignment with the FDA on the Phase III trial. Can you discuss the trial, including design and site selection enrollment? And then also talk about challenges that could occur with running a Phase III trial given the attack rate of the virus and the potential for geographic hotspots to occur.

Yes. All great questions. So it's a randomized placebo-controlled study, obviously, 1:1 ratio, active and placebo. 30,000 healthy subjects, mostly in the U.S., we might have a few sites outside the U.S. as we're mostly U.S. driven. We're going to put sites all across the country because of the attack rate question that you just raised, which is none of us know and won't try to be too smart about where are the infection and outbreak is going to be in the September, October time frame. Because again, if you think about the study starting in July, prime and boost, you should have most of the people boosted by August as you go into September, October, November. But I hope September, October is where we're going to get read out on efficacy, and as we said, this is going to be an event-driven study. So as soon as we have the event rate, we are good to go for filing a BLA. And so the way we think about it, and I think the 30,000 subject study is a good dialogue we've had with the agency to try to solve for 2 things. First is safety. Given the very high demand there is for this vaccine, not only in the U.S., but as you know, many countries follow the FDA approval process to get local approval processes, and so this vaccine could be made available for several hundred million people next year, maybe up to 1 billion next year. And so the safety database is critical for us as sponsor because, of course, we want to help people. We don't want to hurt anybody. And of course, for the FDA, in their duty as a regulator to ensure that they believe that the sample size is big enough to not have missed a very rare event that could happen and be a public health problem when you go to hundreds of millions of people. The other piece that the size solve for us is the attack rate, which is you don't need a very high attack rate in the 1:1 placebo-controlled study to get to a number of event that is statistically significant that you can have a good sense, both for us as a sponsor and the FDA in their duty as a regulator to determine the efficacy of a product, to talk of a vaccine as x efficacy, [ exposed ] efficacy. So basically, you don't need to have an outbreak all over the country. I mean if in San Francisco, we have an outbreak in September, October, and we have enough cases on a placebo-controlled basis, that will be enough to declare success in terms of efficacy. But because in July, August time frame, we'll have enrolled all those 30,000 participants across the U.S., we'll have a safety data base. And if you look about the technology, given the mRNA is made of natural components, the nucleotide, even modify uridine, which we do in 100% of our product for 100% of uridine exist in human cells. It is used in tRNA. So it's a natural molecule that our cells have every day since we are born. And if you look at lipid, those are biodegradable lipid, the lipid that we designed, and it was a lot of investment, it was not easy. And our first lipid were not like that. Our first lipid were not biodegradable. We had 6 day half-life in human. The lipid that we're using for this product or CMV or Zika has a 2-hour half-life. So it degrades very quickly once it gets into a cell, so 2 hours half-life into human cell. And so we believe that evades any safety issues there within the first hours or day from injection, the prime and the boost. We do not believe that there's any scientific reason to believe you'd have safety event 6 months or a year after people have been dosed. And so we think that this is important in thinking about the safety of this product, which is its safety around the 24, 48 hours after injection. And then the size is going to help us deal with the attack rate. If the attack rate in several cities is great, if it's only in 1 or 2 hospitals, we're going to be fine to get to the event rate.

Great. And just digging back into the dose here, I mean, was it really all the data from the Phase I and then some of the early data from the Phase II that really informed you as to 100 being better just to give you a delta on protective levels?

Yes. So on the neutralization, it is mostly, if not all the Phase I plus the preclinical work. As we said in our May 18 press release, we've shown full protection in mice. That data should be published literally very, very soon and in days. Now we're talking in days now. We have a nonhuman primate challenge, which also ongoing. The challenge has not happened yet, but we've got antibody titer in the monkeys. As I told you before, we believe animal model is a very good predictor for Moderna's mRNA platform. So between the Phase I data of mouse and monkey data and the safety data on Phase 2, we don't have a neutralization assay on Phase II yet because of the time it'd take, as we just discussed about the Phase I. But we have a good sense for the safety because the first cohort is fully-enrolled. And so the totality of that data made us really, I think, spend a lot of time going back and forth, I think, maybe between the 50 and 100, which is the 2 doses we have in the Phase II. And at the end of the day, our determination was because the 100-microgram is well-tolerated, there's no safety issue with it. Given we are solving for efficacy, which is how much antibody you make, which is dose-dependent, so you can assume 50 is -- 100 is better than 50. And then duration, the piece that, I don't think, that is talked about enough is duration of vaccination. And what we have seen is if you start, of course, with a much higher titer, we will have longer duration because of degradation of the response. And so we think that by getting at 100 is the best chance we are giving this product to get the highest efficacy rate we can. And because, again, of the opportunity again ahead of us, but so given where the other vaccines might be, it's still too early to know, of course, and I think the next month or 2 are going to be very informative to see how all the vaccines shape up as people share their data and so on. And while you cannot compare precisely because the assays are different, we believe, from an order of magnitude standpoint, the assays are meaningful. Meaning if we have 20 and AstraZeneca has 25 or vice versa, we don't think it's meaningful. But if you see a lot of difference in the data, we think that is meaningful. And so I think the adenovirus technology, again, we have the bimodal type of distribution in terms of response, which, of course, is going to have an impact on efficacy. And the other piece that I think we're going to all be looking for as a society is duration and what do you do when you cannot boost? People have got those technology, because as we know, because of the immune response, you cannot boost or you cannot boost well. So it's another opportunity we're also thinking about for us, which is in addition to the vaccines, we might be providing to people for prime and boost initially. Are we going to use mRNA-1273 to boost a year after, 2 years after an elderly that got the Astra vaccine if it goes to the finish line or a J&J vaccine or a CanSino vaccine? So I think that's going to be interesting for us to think about is not only what could be our market share, let's say, in the pandemic setting in '21 and in 2022. But in '23 and '24, what could be actually our market share? Because I think very few vaccine might end up being able to boost people.

And what would be required to inform an emergency use authorization potentially by year-end?

Yes. So obviously, this is an FDA decision and prerogative, our understanding from the discussions we've had is definitely Phase II readouts that we see on the bigger N, where are you on the antibody titer or neutralizing antibody tighter. Obviously, I would say a nonhuman primate challenge model completed, which we should be very soon, as we just discussed. We just need to run the challenge now. The prime and boost have happened. And then safety data from a Phase III, because if you think about it, by the time we read the Phase II, as we said, in August, September time frame, we will have already thousands and thousands of people, maybe tens of thousands of people that have been dosed with prime and boost. So that safety data, given, again, as we discussed, the safety period, we should be all careful about is the first 24 to 48 hours after dosing the prime and dosing the boost. So you could see a world in which, let's say, in the September time frame, we have all those 3 things in terms of data sets, the Phase II neutralization data on the bigger N, the nonhuman primate challenge data and the safety -- the short-term safety data on the Phase III. And I think it will also depend on what's happening in the country in terms of outbreak, and so we will be ready. We are making as much product as we can, and we're ramping up manufacturing. So in the fall, we'll have millions of doses available, tens of millions of doses available in Q4. And so we're just going to ramp up, ramp up and ramp up. Every month, we will have a bigger output of vaccines. And we're not waiting for approval to start making vaccines, and we've been making as much vaccines as we can now for a while. And so if the government were to decide that it's the right thing to do to give Moderna's 1273 an emergency use approval so that we can supply the CDC for them to decide the allocation across the country where it's most needed. Of course, we'll be there to help, and we'll do the right thing.

And then Stéphane, just checking on a couple of time line items. So when the NIH publishes the first full Phase I data set, that will be on the original 45 patients. But what about all these animal challenge models that are also going to be quite important in informing your study?

Yes. So the mice challenge data should be out very soon. And again, so I talked about days, maybe a week or so, very, very soon now. The primate data, as I told you, we have already done prime and boost. So I would say within a month or 2, the challenge model should be complete. When it's complete, the team is going to work with NIH to write a paper. So I would say, in the summer, we should also get the nonhuman primate data. So I think the way we think about the timelines is we should get first the mouse challenge data out published, days. Weeks, the Phase I data, cohort 1, 2 and 3, so 2,500 on a 50-microgram in healthy adults. The elderly data will be published at a later date just because we know everybody is waiting for the healthy adult data and then the nonhuman primate data and then most probably the Phase II data. I think this is the order of 4 publication that should be coming out in the next few months over summer.

And do you still think, based on the Phase III, you could potentially -- the earliest you could potentially file for commercial approval would be 1Q of next year?

So I think it will really depend on the event rate. I think the best-case scenario is there is a world where we get efficacy data by Thanksgiving. Is it 100% sure, of course, not. I said it's the best-case scenario. And we just want to be very transparent, as you know, Salveen, we don't share best-case scenario with any of our program. But given the public suffering in terms of the public health crisis we are going through, which is a total human tragedy, and the economic suffering, we think it is our responsibility to share our best-case scenario. And we'll do our best, like we've done so far, to move as fast as we can without compromising safety. But there's a best case now that by Thanksgiving-ish, we have efficacy data. So we're going to be working, as we've done so far, very closely with the FDA. I have to give them a lot of credit. They have been so responsive and so collaborative and so helpful. They are a big part of the reason why we've been able to move so fast. And we'll be, of course, filing a rolling BLA, so that we have a good dialogue with them. So we educate them on the BLA as much as we can on the CMC front, and I think a big derisking is we have the Type C meeting with the FDA on the Phase III of CMV a month or 2 ago. We reported the outcome of that meeting at our last quarterly call. It was a very positive meeting, very -- responded positively to all of our CMC question, so we believe we have a good line of sight of how to get the right process to get a check on the CMC side of a BLA. We'd be filing all the data on a rolling basis, so we have time to learn about the product as fast as we learn about the product. We have a global CIO that's going to be fully integrating the data from all the sites in real-time from a digital standpoint. And we're going to be sharing that with the FDA on a very regular basis. So that the last piece of data that will come is what they will have to spend their time focusing on as they go into the BLA approval process, so that we all together have shorten that time as much as possible, again, without compromising the safety of the drug, without compromising the processes from FDA, which are very important to ensure public safety. But we believe that by collaboration, sharing data in real time, we can do something quite unprecedented here. And so again, this is a best-case scenario, but is it possible that before the end of the year we file a BLA? It is possible. It is the base plan now, but we are working on the best-case scenario and everything we lose we'll deal with it. But it is our moral obligation and our duty to just do the best work we can without compromising safety.

And Stéphane one last question here. So, you previously stated that in partnership with Lonza, you plan to scale production up to 1 billion doses per year on the 50-microgram dose. Now with 100 micrograms, you're still saying you can potentially get up to 1 billion. So how do you expect to achieve this? And given supply constraints, how do you expect -- how should we be thinking about capacity in year 1 and then what it takes to get to the 1 billion?

Yes. That's a great question. And as I said, this was a big struggle between the 50 or 100-microgram choice for Phase III. And you can imagine Juan and his team, reminding us all the time that the math is you get half of the mass out, and again, because efficacy was our #1 priority. So what we said this morning is that because of the math, the base plan is 500 million dose now. But what is different now versus a month ago, when we announce the Lonza partnership, is the team has continued to make a lot of progress on process scale-up and has find new ways that we didn't even know a month ago because the team is working so collaboratively and so creatively to debottleneck even further the process. So what we are saying today is we have a very good line of sight to 500 million dose per year output, which is consistent with 1 billion dose at 50-microgram before. But now we have pretty good idea how we're going to get to 1 billion. So I don't want to commit to 1 billion today because there's still work to be done. But pick a number in between, I feel pretty good about the number in between. We just wanted to connect the dots for people to be very transparent and to help people understand with the change or with the selection of the dose, what it meant manufacturing wise. So the way I think about it is, I think 500 million dose is a good line of sight pointing between, which is 750 million, I think, is highly doable. Do we think we have a few ways to massage things around to get to 1 billion? Yes. Are we going to get to 1 billion capacity output in January? No. But as you see the time progressing, the way I think about it, Salveen is we're going to scale our most more linearly, which is every month, the output from Moderna plus Lonza will increase -- will increase from literally now in couple of hundred thousand doses a month to a few million a month to low 10s million a month to 20 million, 30 million, 40 million a month. And so I think that's what you want to think about the ramp. I think also, as we learned about the clinical data of the other vaccines, we might decide to be more aggressive in investments. As you know, we raised $1.3 billion in the follow-on placement mid-May. And as we said in [ ES3 ], the use of proceeds is mostly around CapEx, raw material, people for manufacturing for 1273. Of course, we can use that capacity for any other product, CMV and so on. But we think even the demand is so large. And that's a piece, I think that most of us don't appreciate sometimes, and sometimes I even have to walk my team over it, which is we have 7 billion people or 8 billion people who need a vaccine. We have never seen, in the history of medicine, a product, where you need to vaccine the whole planet almost at the same time. And so I have no idea what the other vaccines are going to look like. I don't know what our vaccine is going look like in terms of efficacy. We'll have to learn this in the fall. But do I believe we have a good chance now to launch this product? I do believe so. The FDA has already approved vaccine with 50% efficacy rate. ZOSTAVAX from Merck was approved with 50% efficacy rate. Will it be a good thing for the country to have a vaccine approved at 50% efficacy rate? I believe so, because at the public health level, you will reduce the spread of a virus tremendously. If in addition to all the people that are already naturally infected, already 2 million people confirmed by testing, most probably much higher that we are not aware of, and by the time we launch the vaccine in '21, many more Americans will be infected. Naturally, especially, as we go back into the fall season, as Dr. Tony Fauci has said. And so a 50% vaccine, if you have 3%, 5%, 7%, 10% of the U.S. population naturally infected, will get you to herd immunity. And so we will be better off as a country with a vaccine with 50% than with no vaccine? Of course. And so that is why I think that now we have a very high chance of getting this product launch. So why -- efficacy might get us not only first-in-class, it might get us best-in-class performance. And if we can get, given all the read across the platform that we started this discussion about, if we're going to get this vaccine, if we have very high efficacy compared to what others will be able to provide because of the limitation of our technology with the bipolar (sic) [ bimodal ] distribution of outcome, we might have a best-in-class product. Again, we'll know more in the fall. We are definitely in the first wave of companies with AstraZeneca and Oxford and Pfizer. The other companies are behind us in that people that have critical mass in terms of production. And so efficacy for us is the #1 priority of this product.

Great. Well, with that, Stéphane, thank you so much for your time. Good luck with everything.

Thank you.

Take care.

Thank you. Thanks for having us.

Bye. Thank you.

Bye.