ON Regeneron Pharmaceuticals’ COVID-19 treatment, which consists of a pair of monoclonal antibodies, sometimes fails to bind to antigens produced by the relevant B.1.351 variant of SARS-CoV-2 bioRxiv.
In laboratory experiments, the researchers found that nine times fewer antibodies in the cocktail bind to the antigens of B.1.351 than to antigens from the most frequently circulating version of the virus. This means that a treatment for B.1.351 would have to be nine times as large to achieve the same degree of virus neutralization.
“There is certainly cause for concern,” says Nathaniel (Ned) Landau, a microbiologist at New York University’s Grossman School of Medicine and lead author of the study. “If the titer drops 9 times it can cause it to not work as well.” Given that his study was conducted in vitro, Landau notes that the only way to know for sure is to determine the clinical effects of the cocktail on patients infected with the B.1.351 variant, which first appeared in South Africa.
“You could treat patients with nine-fold more antibodies, but it becomes too expensive and very unhealthy to inject such a large amount of a protein,” Landau added in a follow-up email.
Regeneron’s drug is a single infusion of a combination of the antibodies REGN10987 (casirivimab) and REGN10933 (imdevimab). In November 2020, it received Emergency Use (EUA) approval for those with mild to moderate COVID-19 who are at risk of progression to serious illness. In the clinical trial in support of the EEA that enrolled 799 patients, 3 percent of those who received the infusion ended up hospitalized compared to 9 percent of those who received a placebo.
If one antibody in a cocktail with two antibodies can neutralize SARS-CoV-2, the overall effect is neutralization, albeit with less potency.
– Pamela Bjorkman, Caltech
On February 25, Regeneron stopped the placebo arm of its outpatient study on the effectiveness of the cocktail after an independent review committee stated that the intervention protected well against various SARS-CoV-2 variants. The company’s announcement did not provide specific details on B.1.351. The trial takes place in the US, where B.1.351 is not widely used.
Treatment is based on previous forms of SARS-CoV-2, and new variants raise questions about how effective the cocktail will continue to be. Landau and colleagues tested how well the Regeneron antibodies, either individually or in combination, controlled the infectivity of several SARS-CoV-2 variants.
They developed artificial viruses containing the spike proteins, the projections on the outside of the virus that grab the cells’ receptors to trigger infection, of five different variants including B.1.351 and the benchmark variant D614G, one of the first mutations to show up in the pandemic that quickly dominated. In the cell culture, they exposed the viruses to the same antibodies as those contained in the Regeneron cocktail and measured the antibody serum concentration required to reduce virus activity by 50 percent.
REGN10987 held its own against all variants, but with less success against the top of B.1.351. His companion REGN10933 found it much more difficult to neutralize this variant, as he required a serum concentration 76 times higher than to neutralize D614G. Together, the two were nine times less effective than against D614G.
“If an antibody in a cocktail with two antibodies can neutralize SARS-CoV-2, the overall effect is a neutralization, albeit with less effectiveness,” writes Pamela Bjorkman, a structural biologist at Caltech who was not involved in either study -Mail to The scientist.
Landau and colleagues attribute the decrease in the effectiveness of REGN10933 to two different point mutations in the spike protein that could help the virus evade the antibody, namely K417N and E484K.
The results are supported by Björkman’s work. Her group used 3D models to categorize the results of various antibody / spike protein combinations. In a study published last fall in natureShe and her colleagues predicted that SARS-CoV-2 variants with point mutations as in B.1.351 would successfully evade antibodies such as REGN10933 – just as Landau’s team observed in the laboratory.
In an email to The scientistRegeneron spokeswoman Alexandra Bowie says: “The values reported in this article do not appear to be consistent with other results obtained so far, both from Regeneron’s internal data and from publications by external researchers.”
The study, which Bowie cites, finds that the antibodies in the Regeneron cocktail bind to the antigens of B.1.351 with sufficient effectiveness to reduce virus activity by 50 percent. The authors report that one portion of the two-part cocktail retained its ability to neutralize the antigens of B.1.351 while the other did not. Landau’s team came to a similar conclusion about the effectiveness of the different parts of the cocktail, saying that the cocktail offered some protection against B.1.351, but not as much as against other variants. So there isn’t much of a disagreement between the two studies.
Adapting treatments to the pandemic will be crucial. According to Bowie, Regeneron is aware of the risks of emerging variants and the company has more antibodies on deck. “There is never an antibody that only protects against everything,” says Landau. Over time, however, it is possible “to make it more and more difficult for the virus to find a way out”.