So now we have some clinical data on yet another category of vaccine: SinoPharm’s inactivated coronavirus candidate. This is one of the classic vaccine techniques, where an infectious virus is altered by some sort of protein-denaturing treatment (heating or reactive chemistry) to make it noninfectious. But such particles can retain enough of their protein surfaces to set off a useful immune response – the tricky part is inactivating the virus enough so that it can’t infect cells and replicate, but not so much that it presents totally different proteins to the immune system and raises a response that won’t help against the real virus.
In SinoPharm’s case, they inactivated the coronavirus with beta-propiolactone, which is a classic protein-alkylating compound. BPL is a strained four-membered ring that is ready to be attacked and opened by pretty much any sort of nucleophile, including protein side chains from amino acids such as Cys or Lys. The compound is used for chemical disinfection (surgical instruments and the like), but that’s not a casual application, because it’s carcinogenic by itself. It works out for such applications, though, because it’s very volatile (and thus easy to remove by vacuum or heating), much like another small reactive and toxi) strained ring compound, ethylene oxide. So there’s no danger in using BPL to inactivate a virus – the question, as mentioned, is going to be whether you’ve inactivated it too much.
Patients in the Phase I trial got 2.5, 5, or 10 micrograms of this agent at Day 0, Day 28, and a third time at Day 56. There were 24 patients in each group, plus an equal-sized placebo group that just got alum adjuvant injections. In the Phase II trial, the 5 microgram dose was chosen, and there were two groups: injection at Day 0 and Day 14, or injection at Day 0 and Day 21, with 84 patients in each group and a 28-patient placebo group for each. Median ages were around the early 40s, slightly more men than women. Adverse reactions appear to have been nothing remarkable – pain at the injection site mostly, with very little systemic stuff like fever or fatigue, which certainly appears to be the mildest profile of the vaccines that we’ve seen so far.
As for neutralizing antibodies, it looks like the three-dose Phase I trial had an odd dose-response. The medium dose was actually slightly worse than either the low or high one. Meanwhile, in the Phase II, which was done with that medium five-microgram dose, the antibody response (measured two weeks after the second dose) was not as strong as with the full three-dose schedule, but the 0/21 day dosing schedule led to a better response than the 0/14 one. It appears from the Phase II data that one of the 42 patients who were tested for antibody response in that group did not seroconvert at all. The geometric mean titer values for the neutralizing antibodies (247 for the 0/21 group) appear to be in the range of other Phase I data reported, although it’s not easy to make a head-to-head comparison with any certainty. There is no comparison in the study with a convalescent plasma group, but as we’ve been seeing, those samples tend to be pretty variable themselves. There are also no data on T-cell responses.
So this is a rather preliminary report (as the authors themselves note), but it’s the first one we have on an inactivated vaccine. Like all of the others so far except the J&J Ad26 one, this candidate will also need a booster shot. The small and mild adverse-event reactions here are really the main thing that stands out – if you’re a glass half full person, then you can be glad about that, but if you’re a glass-half-empty one, you might wonder about the overall robustness of the immune response (update: see here for more on this issue). We’re going to need more data to make any calls about that, and (just as with every other vaccine under development!) the real numbers we’re waiting on for efficacy. How many people will this (or any) vaccine protect, and how well? Stay tuned.