This study is not particularly interesting as a measurement of vaccine effectiveness (because of the extremely low sample count of 12 patients). So don't read too much into the particular numbers. Instead, there's two useful and valuable results in it:
1. It gives us information on the mechanism that the variants are using for vaccine evasion - the E484K substitution which was the suspected mechanism for B.1.351 ("south african variant") was tested on its own and did not evade the vaccine-induced antibodies as effectively as the full B.1.351 mutation set. This tells us that there is some other mechanism in B.1.351 that makes it more successful at evading antibodies. This is important and valuable information.
2. It demonstrates a really fast, safe, and clever method of testing vaccine effectiveness against a new variant - what they did was to take an entirely unrelated virus, graft a modified S protein on it, expose it to blood serum from vaccinated patients, and then try to infect a cell culture with it. By counting infected cells compared to the same count without the serum exposure, you can measure how effective the neutralization is, without having to measure infection counts in a large population. This is awesome and can easily be done with other vaccines and any new variant that shows up, as long as it's sequenced, even if it doesn't have much prevalence in the population. In fact, you can use this method to test hypothetical variants that only exist in animal viruses or not at all, and be able to redeploy vaccines that are particularly effective against those in regions where they happen to emerge. This, to me, is the absolute highlight of this paper, and I suspect it will be extremely useful.
1. It gives us information on the mechanism that the variants are using for vaccine evasion - the E484K substitution which was the suspected mechanism for B.1.351 ("south african variant") was tested on its own and did not evade the vaccine-induced antibodies as effectively as the full B.1.351 mutation set. This tells us that there is some other mechanism in B.1.351 that makes it more successful at evading antibodies. This is important and valuable information.
2. It demonstrates a really fast, safe, and clever method of testing vaccine effectiveness against a new variant - what they did was to take an entirely unrelated virus, graft a modified S protein on it, expose it to blood serum from vaccinated patients, and then try to infect a cell culture with it. By counting infected cells compared to the same count without the serum exposure, you can measure how effective the neutralization is, without having to measure infection counts in a large population. This is awesome and can easily be done with other vaccines and any new variant that shows up, as long as it's sequenced, even if it doesn't have much prevalence in the population. In fact, you can use this method to test hypothetical variants that only exist in animal viruses or not at all, and be able to redeploy vaccines that are particularly effective against those in regions where they happen to emerge. This, to me, is the absolute highlight of this paper, and I suspect it will be extremely useful.