RR:C19 Evidence Scale rating by reviewer:
Reliable. The main study claims are generally justified by its methods and data. The results and conclusions are likely to be similar to the hypothetical ideal study. There are some minor caveats or limitations, but they would/do not change the major claims of the study. The study provides sufficient strength of evidence on its own that its main claims should be considered actionable, with some room for future revision.
Zhou et al. investigate how SARS-CoV-2 adapted for the infection of farmed mink and lab ferrets, and conclude with reliable evidence that publicly known mink-adapted viruses are unlikely to pose an increased threat to human beings. However, they emphasize the need to continue the prevention of cross-species transmissions of the virus, particularly because SARS-CoV-2 outbreaks in mink have been reported in at least eleven countries: the Netherlands, the USA, France, Spain, Denmark, Italy, Sweden, Canada, Greece, Lithuania, and Poland; are ongoing; and incidents of human beings infected by mink have been reported.
This work focuses on mutations enriched after circulation in mink and ferrets, particularly in the RBD: Y453F and N501T, which have both been associated with increased binding to human ACE2. The authors hypothesize that these mutations may non-specifically increase binding to several groups of mammalian ACE2 proteins but did not test this in the study. They very nicely demonstrate that Y453F, N501T, or F486L each allowed SARS-CoV-2 spike expressing pseudoviruses to enter ferret ACE2 expressing cells with much greater efficiency. In addition, they show that Y453F-bearing SARS-CoV-2 viruses are outcompeted by their 453Y counterparts in human airway epithelial cells, and readily neutralized by convalescent first wave antisera and antisera from health care workers vaccinated with two doses of the Pfizer-BioNtech-BNT162b2 vaccine. Zhou et al. suggest that this loss of fitness in human cells may explain why Y453F is not very commonly observed in human SARS-CoV-2 samples. I confirmed the low prevalence of Y453F among human SARS-CoV-2 genomes on covidcg.org: it turned out to be only 0.04% (present in 1131 out of 2780099 high-quality human SARS-CoV-2 genome sequences on September 9, 2021). It would be great if similar experiments could be performed to characterize N501T, F486L/I, as well as other mink-enriched variants to see if any might not lead to a loss of fitness or impact antibody neutralization in human beings.
The first set of results presented in this study draws from experiments in their previously published ferret SARS-CoV-2 infection and transmission experiments: two pairs of ferrets; each pair consists of a donor and a direct contact (infected by the donor). Zhou et al. sequenced the viruses extracted from the ferret pairs and corroborated others’ findings that the two mutations, Y453F and N501T, are associated with experimental ferret adaptation. It’s a little unclear what the experimental setup was because the paper points to a previous publication for details: “Ferret (Mustela putorius furo) infection studies with SARS-CoV-2 virus were performed as described previously.” So, I had to go to their previous publication where the experiment from which this data derives is described: “Four ferrets per group were each infected intranasally with 105 p.f.u. of clonal WT or ΔCS mutant SARS-CoV-2. After 24 h, naïve contact ferrets were cohoused with each donor. Ferrets were nasal-washed daily for the following 2 weeks and virus shedding was titrated by quantitative PCR with reverse transcription (RT–qPCR) and by median tissue culture infectious dose (TCID50).” My recommendation #1 is for this detail to be contained within this manuscript as well so readers do not need to read about the experiment in a separate publication.
Recommendation #2 Their figure 1A (and also figure 2A) could be improved by showing the % Y453F and N501T in a different way. Rather than the pie chart, it might be clearer to just state the numerical % value because the two mutations do not lie at the same coordinate. Currently, it’s not clear to a reader how much the two mutations co-occur at each time point. It would be good to have these numbers shown for day 0 (initial virus inoculum) in the figure as well. Figure 1B is a useful depiction of the genetic diversity exhibited in mink SARS-CoV-2 across several countries. One minor recommendation #3 is to change the colors used in the phylogenetic tree because these overlap with the colors used to indicate different mutations on the left and right of the tree, e.g., orange indicates both N501T and the Netherlands; blue indicates both Y453F and France.
The next ferret experiment in the study compares the P2 results with the WT results from the previous experiment; this means that the ferret infection experiments were run at different times. I think this avoids sacrificing animals but still needs to be noted as a caveat in directly comparing the outcomes of the ferret infection experiments. Recommendation #4 Since there are only 4 replicates per experiment (figures 2B-E), it might help readers if the data points show the individual replicates as opposed to the mean with error bars. What does the # indicate in figure 2F?
Recommendation #5 Figure 3A-C (and figure 6) could benefit from not being normalized to the human ACE2 expressing cells. From the normalized figures, it is not possible to tell how each spike mutation affects pseudovirus entry into human ACE2 expressing cells. In other words, it would be better to see the actual entry rate per ACE2 expressing cell per spike variant. For figure 3B, the non-normalized data is available in extended figure 2A. There, readers see that it’s not necessarily that Y453F has dramatically increased pseudovirus entry into cells expressing ferret ACE2, but that Y453F pseudovirus entry into cells expressing human ACE2 has taken a hit; when normalizing to the entry into cells expressing human ACE2, this makes Y453F look like it has very much improved compared to WT. For this reason, I think it is better for the non-normalized data to be presented in all main figures.
There is a minor error in the caption of extended data Figure 2B. It’s probably the non-normalized data from Figure 3C, not extended data in Figure 3C.
Minor typo: “We found that nearly all variants of concern tested could better utilize mink ACE2” where mink ACE2 should be ferret ACE2.
Zhou et al. conclude the work by showing that several VOC-associated mutations such as L452R, E484K, and N501Y may help SARS-CoV-2 to infect ferrets or mink. As they write in their discussion, their findings emphasize the importance of continuing to closely survey, sequence, and share mink SARS-CoV-2 data in case new VOCs infect farmed mink and adapt in unexpected ways. They express concern rightfully that despite ongoing mink outbreaks, only sequences from 2020 have been shared. They also emphasize that ferrets and mink may not be optimal models for investigating spike-humanACE2 interactions because of the different adaptations required for effective infection and transmission in each species.