As antibodies are being developed and used to treat SARS-CoV-2 in the current COVID-19 pandemic, scientists are becoming concerned about the risk that certain mutations can cause the virus to escape from this inhibition. A new study maps all known mutations in the virus, showing how each contributes to escape from individual antibodies.
Image Credit: https://www.biorxiv.org/content/10.1101/2020.11.30.405472v1.full.pdf
Since the SARS-CoV-2 virus depends on the binding of its receptor-binding domain (RBD), on the spike glycoprotein, with the angiotensin-converting enzyme 2 (ACE2) receptor on the target host cell, inhibition of this binding can prevent the infection from establishing itself.
Such a mechanism of inhibition has been observed in other viruses, that have evaded existing antibodies by selecting mutations that disrupt the binding epitope. This prevents binding, which depends on antigen-antibody specificity.
Most antibodies being used against SARS-CoV-2 today use the RBD, so the researchers developed a new method of examining the different ways in which RBD mutations affect the ability of antibodies to recognize and bind to it. This is called deep mutational scanning. The study was published in November 2020 in the preprint server bioRxiv*.
Here, the researchers gathered together all the known RBD mutants, stimulating their production on the surface of yeast cells. They then used known methods such as fluorescence-activated cell sorting (FACS) along with deep sequencing, where the sequences of the nucleotides were read many thousands of times.
This allowed them to find out the exact extent to which each mutation affected the folding of the RBD protein, which is essential to achieve the right shape for antibody attachment. They also found out how these changes would affect the affinity of the RBD for the ACE2 receptor.
The researchers selected the two antibodies in the Regeneron REGN-COV2 cocktail, namely, REGN-10933 and -10987, and the Eli Lilly antibody CB6 or LY-CoV016, for testing. The former was recently given an emergency use authorization for the treatment of COVID-19, and the latter is in phase II trials.
Individual and Dual Antibody Evasion
Their findings show the RBD mutations that avoid binding by each of the antibodies and by the Regeneron cocktail. The two REGN antibodies are evaded by distinct sets of mutations in the receptor-binding motif (RBM) of the RBD, and they bind to different epitopes in this location.
However, they also found one mutation, E406W, that powerfully evades the double-antibody cocktail as well. These findings were confirmed in actual viral neutralization experiments.
This shows the importance of complete mutation mapping in demonstrating escape mutations, since both by structural analysis and by selecting for viral escape mutations, the cocktail had previously been thought to be immune to any single mutation.
The CB6 antibody can also be evaded by several mutations in the RBD at different epitopes, only some of which affect the ability of the RBD to bind the receptor or its functionally important protein structure. Thus, the RBD has considerable tolerance to mutation.
Identification of ‘Missed’ Escape Mutations
Secondly, they explored whether the mutational map could detect the direction of viral evolution in human beings, they performed deep sequencing of viral sequences from a persistently infected immunocompromised patient who was given the Regeneron cocktail at day 145. The virus had enough time to go through a multitude of mutations, making it an ideal subject for the study of viral genetic mutations.
The administration of the cocktail was found to result in the rapid development of five mutations in the RBD, three causing the virus to evade one antibody and one leading to evasion of the second antibody. The mutant strains competed with each other, and thus the counts of each showed rises and falls, as seen in other viruses where adaptive evolution occurred within the host. This could be caused by the proximity of two or more mutations causing them to be transmitted together to a new viral particle, along with competition between viral strains carrying different escape mutations.
Here again, three of the four identified escape mutations were missed in the manufacturer’s experiments using cell cultures to detect viral selection. The problem with this type of setup is its ability to identify only whatever mutations happen to be selected under those conditions, while mutational mapping records the full range of mutations even when unrelated to treatment but which may impact virus-antibody binding.
Prediction of Evolutionary Path
The researchers used their complete maps to predict the most probable paths the virus would take over its evolution. This would involve mutations that are completely functional in terms of ACE2 affinity and which come about via a single nucleotide substitution, and which disrupt antibody binding.
Of course, over longer periods, epistatic interactions could change the limits of mutational tolerance, as when the virus is challenged by drugs or antibodies.
What are the Implications?
The study used all human-derived viral sequences up to November 12, 2020, and found multiple RBD mutations that allow antibody escape, but at very low frequencies. They set a frequency threshold of above 0.1% of sequences, above which was the REGN10933 escape-mutant Y453F, at 0.2%, and the REGN10987 escape-mutant N439K at 1.2%.
The former was found in many mink infections in European mink farms, but the mink virus sequences also have other escape mutations. The latter is very common in sequences from Scotland and Ireland.
Another important finding from this study is that structural studies cannot provide a full idea of the escape mutation map. While most such mutations occur at the interface of the viral RBD and the antibody, the Eli-Lilly antibody binding was evaded by mutations at the RBD amino acids that come into contact with the heavy chain CDRs, even though both heavy and light chains are used to engage a large region overlapping the surface that interacts with the ACE2 receptor.
This is different from those seen with the REGN cocktail, occurring at the RBD residues that engage the heavy/light chain interface. And the single mutation that escapes both REGN antibodies occurs, interestingly, at an epitope that is not bound by heavy or light chains. Thus, no obvious rules apply to escape mutations concerning epitope location.
The study concludes with the observation that the presence of escape mutations does not imply that SARS-CoV-2 will develop extensive antibody resistance since it rarely persists for long in the same patient.
These maps demonstrate that prior characterization of escape mutations was incomplete: for instance, overlooking a single amino-acid mutation that escapes both antibodies in the REGN-COV2 cocktail, and failing to identify most mutations that arose in a persistently infected patient treated with the cocktail.”
As scientists wait for mutations to arise at higher frequencies in the circulating viral strains, such studies will help to interpret the effects of such mutations immediately.
bioRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.
- Starr, T. N. et al. (2020). Prospective mapping of viral mutations that escape antibodies used to treat COVID-19. bioRxiv preprint. doi: https://doi.org/10.1101/2020.11.30.405472. https://www.biorxiv.org/content/10.1101/2020.11.30.405472v1