We have a lot more vaccine news to catch up on, and I’ll use this as a new “frontrunning vaccine candidates” post, replacing this one (and updates therein) from about a month back. That one will give you some more general information on most of the projects below; if you haven’t read it, it’ll be a good intro. A general vaccine background post is here, and here’s a recent one one covering some other candidates. The ones below are in no particular order; no oddsmaking is implied, because you’d have to be a lunatic to try to guess which ones will work and which won’t at this point. We can be pretty sure, though, that many of them will fail for one reason or another, just because most drugs and most vaccines fail. We can also be sure that as we go along there will be likely be many hard tradeoffs to consider in efficacy, manufacturing, safety, and other issues – I’ll frankly be amazed if that doesn’t happen. So with those points out in the open, here we go – note that the J&J entry also has some general vaccine news that might be of interest:
Just in the last few days, there has been a very detailed report in The Lancet from the CanSino team on Phase I studies of their vaccine candidate. As you can see from the name, it’s a viral vector using adenovirus-5, a choice that on the one hand speeds up development, but on the other hand raises worries about efficacy and tolerability. The adenovirus in this case is engineered to express the coronavirus Spike glycoprotein, a common target to many of the vaccine efforts. Many people already have antibodies to this sort of adenovirus. Past vaccine development efforts by this route have run into problems because of an immune attack on the vector, which (as you’d figure) can then interrupt the plans of developing immunity to its payload.
This one was an open-label dose-escalation study, with the primary endpoint being adverse events at day 7 post-dosing (intramuscular injection), with follow-up for 28 days. They also determined neutralizing antibody and T-cell responses – valuable data, although it has to be said up front (as the paper does) that until we get into Phase II we won’t know what responses are really needed for protection. There were three dosages, with 36 patients in each group, median age 36 years, almost exactly 50/50 male/female.
75 to 83% of each group reported at least one adverse reaction by day 7, generally pain at the injection site and sometimes general fever, muscle pain, etc. There was more of this at the highest dose, which is no surprise. Keep in mind, with any viral vector you’re looking at giving someone an entirely new viral infection, albeit one that produces your antigens of choice rather than producing more virus. So you can expect it to come on with flu-like symptoms. Importantly, there were no serious events within the 28 day study period in any group – “serious” is a term with regulatory implications, meaning anything leading to hospitalization or the possibility of permanent damage.
How about the immune response? All patients developed neutralizing antibodies, and all of them showed T-cell responses as well (both CD4+ and CD8+). So that’s good – as mentioned, we don’t know what sort of response is needed, but these look significant and dose-responsive, which is all you can ask at this stage. As we get more hard data from the other vaccine efforts, we can compare. But there are some wrinkles in the data. It turns out that 44 to 56% (depending on the patient group) of the people in the study already had high levels of neutralizing antibodies to the Ad5 virus, and the CanSino team found that this definitely hurt the immune response to the spike protein payload. There was also a generally lower antibody response in the older patients (45-60, which prompts me to say “that ain’t so old”). This is something that they’re going to specifically look for as they move into Phase II, especially given how the severity of the disease increases with age.
So this is honestly just about what one would have expected: a real immune response to the Spike protein, but with headwinds from Ad5 antibodies. On to Phase II, where vast amounts of stuff are going to get sorted out.
For this one, I’m going to refer folks to this recent post on the rhesus animal model data. Short version: there are arguments about the way the vaccine did not induce “sterilizing immunity” (i.e. complete protection from infection) in that study. It’s impossible to directly compare these numbers with the SinoVac monkey studies, unfortunately.
The other news on this one has been the statement from Oxford’s Adrian Hill in an interview in the Sunday Telegraph that he thought the vaccine had about a 50% chance of working. This made a lot of headlines, but you know, in a more informed world it wouldn’t have. The vaccine trial failure rate is around 90% overall. That’s not a completely fair comparison, since it goes across all sorts of different pathogens and in this situation we’re only looking at one. But on the downside, if there’s some kind of systemic problem with developing protective immunity to this virus, it’ll be common to all of the vaccine efforts (I should note that I don’t think that’s the case, but until we get to Phase II results we can’t rule it out, either).
Hill’s other point was what should be an obvious one, too: you want to do those Phase II trials in a population where the virus is actively spreading, in order to get the best data about whether your therapy actually protects people. In countries where public health measures (lockdowns, masks, social distancing and all the rest of it) have cut down on viral transmission, this gets harder. We may well have to chase the virus from area to area, country to country, to do the best vaccine trials. Then again, if we have opened back up too quickly, as some fear, we might have plenty of places to test in – we’ll just have to wait and see. The problem is that setting up these trials takes some real time and organization, and it will be hard to guess where the best places to run them will be, so far in advance.
I’ll refer people to this recent post for more details as well. Moderna has taken some criticism for the way that they released these data, and especially for the way that they did so just before a planned sale of more Moderna stock. I also wonder (as do people in that article) about who’s buying the stock at these valuations, considering that uncertainties in vaccine development, in the whole COVID-19 epidemic, and in the mRNA vaccine mechanism in particular. It would be good if we had as comprehensive a look at Moderna’s results as we do at CanSino’s
But I’m still glad that Moderna is in the hunt. We need a number of different mechanisms going to spread around the considerable risk. It’s certainly true that we don’t know how well mRNA vaccines will work in humans – we’re finding out a bit ahead of schedule thanks to the coronavirus. I would not be taking a flyer on Moderna stock, but I’m glad that they’re taking one on the virus.
BioNTech and Pfizer
No recent news here; human dosing has been ongoing but there’s no word on results. Here’s a good background article on the effort, though, from Forbes. Keep in mind that Pfizer and BioNTech are going in with four vaccines rather than just one; they’re taking a massive risk on a massive scale. There will eventually be a lot of data to sort though, and the expectation is that only one of those four (if that!) will be able to move on.
Nothing new here, either, from what I can see. The company did provide one of the earliest looks into animal model data for their inactivated-virus vaccine candidate, but we’re just going to have to wait for some human data (I’ll update here on this post when they do). There was an odd bit of news a few days ago, a press release about how they’d received more funding for the vaccine’s development. But this was only $15 million, a laughably small amount compared to what’s going to be spent in this area by anyone in human trials. Put an asterisk next to anyone you saw rebroadcasting that one as some sort of useful news item; odds are they that don’t know much about the field,
The Wuhan Institute For Biological Products / Sinopharm
Another inactivated-virus candidate here. Latest word is that 96 patients (three age cohorts) have completed a Phase I trial and that the vaccine has moved into Phase II, but that news is about a month old. This one is probably the quietest of all the major vaccine efforts, from what I can see. Will that change?
The company recently published a paper on the results with this DNA vaccine candidate in mice and guinea pigs. The results look good, as far as I can see, but they’re just on immunogenicity. There appears to be a robust antibody response, but we have no data yet on viral challenge in the animal models, and no human data yet on immunogenicity or tolerability. The company has said that it expects a readout in June, and their head of R&D rather informally told Reuters last week that so far they are seeing no safety signals in the human dosing.
Johnson & Johnson (Janssen)
No news in the last couple of weeks on the Ad26 adenovirus vector candidate from J&J. This (like the Oxford vaccine) aims to use an adenovirus type that humans have not really been exposed to, to avoid the problems noted with the CanSino Ad5 candidate (see above). This will all have to be proven in the clinic, of course, and that’s just where all of these candidates are piling up.
In the meantime, the Barouch lab at Harvard (developers of the Ad26 platform) have published a paper with a long list of coauthors, including some at J&J. This takes a look at several DNA-based coronavirus vaccines in rhesus monkeys, and to my knowledge is the first detailed look we’ve had at these in such a model. It’s a lot of work: they tried six variations, namely (1) DNA coding for the full-length spike protein, (2) a variation without the cytoplasmic “tail” region, (3) one that deleted this and the transmembrane domain (leaving only the exposed part of the Spike), (4) one with just the S1 domain of the Spike (and a trimerization tag), (5) one with just the RBD (receptor-binding domain) region of the Spike (and a trimerization tag), and (6) a rather heavily engineered one with deletion of the furin protease cleavage site and two introduced proline mutations (along with the trimerization tag), intended to make the resulting protein more stable overall.
These each gave a somewhat different antibody response, when the team looked at the titer and type of antibodies raised (3 and 6 looked the most similar to each other, and 5 was probably the most distinctive outlier – see the paper’s Figure 2E). They also saw T-cell responses, with the shorter constructs showing weaker activity. After these vaccinations and three weeks after a booster at the 3 week mark, the animals were challenged with live coronavirus. The paper’s Figure 4 has the results (nasal swab and aveolar fluid checks for viral RNA), and the differences between the six candidates are overall not gigantic – all six had efficacy. But it does look like the plain spike protein one (1) and the RBD one (5) perhaps looked strongest overall, with the unstabilized shorter Spike one (3) looking the weakest. That’s good news, most likely, since the vaccine candidates in the clinic are mostly Spike and RBD, from what I can see. There were definite variations among the individual monkeys in each treatment group as well.
The group looked at what markers and assays might correlate best with actual protection against viral challenge. Interestingly, neither the CD4+ nor the CD8+ data were good predictors. The aveolar fluid viral load was a better readout than the nasal swab one, probably due to more variability in the latter. The best reading overall was the titer of neutralizing antibodies. Most animals in this study (as with the Oxford vaccine, see above) did not achieve sterilizing immunity – they still got infected, but in a much milder fashion than unvaccinated animals. But the ones that were completely protected showed significantly higher titers of neutralizing antibodies than the others; that measurement seems to be the way to go. We’ll have to keep all this in mind as the human data pile up.
No recent news on this collaboration, either, from what I can see. That will change!
I just highlighted these folks in a post the other day – they have a recombinant protein vaccine candidate in the works. Word is today that they’ve enrolled their first patients in a Phase I/II trial. Good luck to them! That has sent the company’s stock up, naturally, along with everyone else in the space. Just to emphasize this again, trying to come up with a coronavirus vaccine play in your stock portfolio is a very bad idea, in my opinion: this is going to be a wild, volatile ride and most of the candidates are going to lose out. We are going to need a lot more human data before we have the faintest idea which ones might win, and every fresh wave of news over the summer and fall is going to stir the valuations around chaotically. If that appeals to your investment style, well hey, have at it. But you’ve been warned.
Finally, Merck has come out of stealth mode with some vaccine news of their own. They’re buying Themis, a Vienna-based company that’s been working on a weakened form of the human measles virus as a vector for vaccine delivery (this is already in use as a measles vaccine itself). A coronavirus vaccine candidate using this technology is planned to go into human trials “in a few weeks”, likely in France. And in that same post last week where I mentioned Novavax, I brought up the VSV vector idea, which reached fruition in the Ebola vaccine. Merck has also announced that they’re working with IAVI, a nonprofit that has been focusing on on HIV vaccine using this technique, to try a coronavirus vaccine as well. Merck seems to have made a deliberate decision to focus on viral vectors rather than mRNA, recombinant proteins, etc., in the belief that these have a better shot at potent immunlogic responses. That will get sorted out in the clinic as well!
So that’s the current roundup. I’ll use this post as a platform for updates until there’s enough news to start a new one. And there will be – of that I’m sure. . .