Skip to main content

Mapping Which Coronavirus Variants Will Resist Antibody Treatments

Posted on by Dr. Francis Collins

Antibodies Binding to RBD
Caption: The antibody LY-CoV016 (purple) is bound to RBD. This “escape map” indicates where in the viral RBD new mutations are most likely to make LY-CoV016 less effective (red). It also shows places where mutations are least likely to affect antibody binding (white) and where mutations can’t persist because they’d disrupt RBD’s ability to function (gray). Credit: Adapted from Starr TN, Science, 2021.

You may have heard about the new variants of SARS-CoV-2—the coronavirus that causes COVID-19—that have appeared in other parts of the world and have now been detected in the United States. These variants, particularly one called B.1.351 that was first identified in South Africa, have raised growing concerns about the extent to which their mutations might help them evade current antibody treatments and highly effective vaccines.

While researchers take a closer look, it’s already possible in the laboratory to predict which mutations will help SARS-CoV-2 evade our therapies and vaccines, and even to prepare for the emergence of new mutations before they occur. In fact, an NIH-funded study, which originally appeared as a bioRxiv pre-print in November and was recently peer-reviewed and published in Science, has done exactly that. In the study, researchers mapped all possible mutations that would allow SARS-CoV-2 to resist treatment with three different monoclonal antibodies developed for treatment of COVID-19 [1].

The work, led by Jesse Bloom, Allison Greaney, and Tyler Starr, Fred Hutchinson Cancer Center, Seattle, focused on the receptor binding domain (RBD), a key region of the spike protein that studs SARS-CoV-2’s outer surface. The virus uses RBD to anchor itself to the ACE2 receptor of human cells before infecting them. That makes the RBD a prime target for the antibodies that our bodies generate to defend against the virus.

In the new study, researchers used a method called deep mutational scanning to find out which mutations positively or negatively influence the RBD from being able to bind to ACE2 and/or thwart antibodies from striking their target. Here’s how it works: Rather than waiting for new mutations to arise, the researchers created a library of RBD fragments, each of which contained a change in a single nucleotide “letter” that would alter the spike protein’s shape and/or function by swapping one amino acid for another. It turns out that there are more than 3,800 such possible mutations, and Bloom’s team managed to make all but a handful of those versions of the RBD fragment.

The team then used a standard laboratory approach to measure systematically how each of those single-letter typos altered RBD’s ability to bind ACE2 and infect human cells. They also measured how those changes affected three different therapeutic antibodies from recognizing and binding to the viral RBD. Those antibodies include two developed by Regeneron (REGN10933 and REGN10987), which have been granted emergency use authorization for treatment of COVID-19 together as a cocktail called REGN-COV2. They also looked at an antibody developed by Eli Lilly (LY-CoV016), which is now in phase 3 clinical trials for treating COVID-19.

Based on the data, the researchers created four mutational maps for SARS-CoV-2 to escape each of the three therapeutic antibodies, as well as for the REGN-COV2 cocktail. Their studies show most of the mutations that would allow SARS-CoV-2 to escape treatment differed between the two Regeneron antibodies. That’s encouraging because it indicates that the virus likely needs more than one mutation to become resistant to the REGN-COV2 cocktail. However, it appears there’s one spot where a single mutation could allow the virus to resist REGN-COV2 treatment.

The escape map for LY-CoV016 similarly showed a number of mutations that could allow the virus to escape. Importantly, while some of those changes might impair the virus’s ability to cause infection, most of them appeared to come at little to no cost to the virus to reproduce.

How do these laboratory data relate to the real world? To begin to explore this question, the researchers teamed up with Jonathan Li, Brigham and Women’s Hospital, Boston. They looked at an immunocompromised patient who’d had COVID-19 for an unusually long time and who was treated with the Regeneron cocktail for 145 days, giving the virus time to replicate and acquire new mutations.

Viral genome data from the infected patient showed that these maps can indeed be used to predict likely paths of viral evolution. Over the course of the antibody treatment, SARS-CoV-2 showed changes in the frequency of five mutations that would change the makeup of the spike protein and its RBD. Based on the newly drawn escape maps, three of those five are expected to reduce the efficacy of REGN10933. One of the others is expected to limit binding by the other antibody, REGN10987.

The researchers also looked to data from all known circulating SARS-CoV-2 variants as of Jan. 11, 2021, for evidence of escape mutations. They found that a substantial number of mutations with potential to allow escape from antibody treatment already are present, particularly in parts of Europe and South Africa.

However, it’s important to note that these maps reflect just three important antibody treatments. Bloom says they’ll continue to produce maps for other promising therapeutic antibodies. They’ll also continue to explore where changes in the virus could allow for escape from the more diverse set of antibodies produced by our immune system after a COVID-19 infection or vaccination.

While it’s possible some COVID-19 vaccines may offer less protection against some of these new variants—and recent results have suggested the AstraZeneca vaccine may not provide much protection against the South African variant, there’s still enough protection in most other current vaccines to prevent serious illness, hospitalization, and death. And the best way to keep SARS-CoV-2 from finding new ways to escape our ongoing efforts to end this terrible pandemic is to double down on whatever we can do to prevent the virus from multiplying and spreading in the first place.

For now, emergence of these new variants should encourage all of us to take steps to slow the spread of SARS-CoV-2. That means following the three W’s: Wear a mask, Watch your distance, Wash your hands often. It also means rolling up our sleeves to get vaccinated as soon as the opportunity arises.


[1] Prospective mapping of viral mutations that escape antibodies used to treat COVID-19.
Starr TN, Greaney AJ, Addetia A, Hannon WW, Choudhary MC, Dingens AS, Li JZ, Bloom JD.
Science. 2021 Jan 25:eabf9302.


COVID-19 Research (NIH)

Bloom Lab (Fred Hutchinson Cancer Center, Seattle)

NIH Support: National Institute of Allergy and Infectious Diseases


  • Monika G. says:

    Very informative article. Thank you.
    I live in Monroe County, Florida. Specifically Key Largo, Florida. I am eighty years old and it is impossible to receive an appointment for the vaccine. I have tried since the beginning, appointments cannot be had. I try at the twice-a-week openings for the vaccinations. I have done everything possible. I cannot get an appointment for this vaccine. What more can I do?

    • Gail says:

      I would contact my primary-care physician. In Pinellas County Florida, physicians can refer their eligible patients to the hospitals they are affiliated with. The hospitals as they receive doses then contact the patients and give appointments. There is also a website my where you can pre-register. There is a drop down menu where you can indicate your county. . . Good luck!!

  • Neal Dante Castagnoli says:

    Why does a great article have to end with this speculation, asserted as fact?

    While it’s possible some COVID-19 vaccines may offer less protection against some of these new variants—and recent results have suggested the AstraZeneca vaccine may not provide much protection against the South African variant, there’s still enough protection in most other current vaccines to prevent serious illness, hospitalization, and death. And the BEST way to keep SARS-CoV-2 from finding new ways to escape our ongoing efforts to end this terrible pandemic is to double down on whatever we can do to prevent the virus from multiplying and spreading in the first place.

    “The best?” How do you know? Here is a rejoinder:

    Would you not agree partial herd immunity, by vaccination, illness., is a member of “whatever we can do to prevent the virus from multiplying and spreading in the first place?”

    Yet, consider the potential danger of partial herd immunity. It will favor those variants that can infect those who have already been inoculated or have natural antibodies. A natural filter, that favors mutation, adds a new dimension to “fitness”: those viruses that can infect those inoculated or previously infected. Life finds the path.

    The filter is NOT present in virgin populations, and nor in populations with full herd immunity. It is ONLY present during the transition from virgin to full herd immunity. That means shortening the period to get to full herd immunity is highly desirable, WORLD WIDE.

    This means do the opposite of what you suggest. It means passing as quickly as practical through the partial herd immunity stage. It means changing policy to encourage those with small chances of serious disease worldwide to take off the mask and eschew 2 meter and time based distancing. It means opening restaurants for those at low risk, keeping jobs open for those at low risk, allowing children to live children’s lives, for mothers to live mother’s lives, for workers below 55 or whatever is considered safe, to work and to remove the fear instilled by the horror stories promoted by the press. Putting things into perspective, there is only a 2% increase in mortality for those under 55. I suspect many, if not most, of the 2% have pre-existing conditions and could be protected.

    I could go on, but the response to SARS-COV-2 has been damaging in the extreme, and for reasons that make little sense. The “Curve” did not need to be lengthened shortly after the initial surge in NY, in which doctors learned better ways to treat, reducing deaths by a factor of four. That approach tapped out in April. It’s uncertain when the numbers were known that this is an old people’s disease/people with pre-existing conditions (hey, I’m in a reasonably high risk group), but once this was known, the answer was NOT to treat everyone equally, but to sequester only those of unacceptable risk, and to allow those who are at low risk to live their lives. Heck, the whole approach only flattened the curve, but did not reduce the area under the curve in any event (well, depending on which 95% CI published by the CDC you believe: I use the 1/1/2021 numbers).

    So how about this? NO sequestration except for those at real risk. Your vaccine is too late, and you aren’t “preventing” deaths. Building up more immunity in the population especially those at low risk will reduce chances for those higher at risk of contracting the disease. What is not to like? It boggles my mind.

    And, for goodness sake. Leave something for the next generation and stop destroying the economy. And let children live normal lives: they have almost NO risk. And get them back to school. College kids too. These are important formative years, and the stupid policies are destroying futures. Some cannot succeed with remote learning. Socialization skills are lost.

    As if cell phones weren’t bad enough, now you have added a government layer to isolation and anti-humanness to our technologically driven society. And for what? I suspect you will find “Very little.” This concept that “The Vaccine will save us” in a population, using 1/1/2021 CDC using the CDCs own 95% confidence mean of 7.2X puts the already infected US population at 59%. Given how fast new infections are dropping, I would say “Almost over.”

    Please tighten up your science, tighten up your statements, and STOP promoting policies that damage the economy, put lives at risk, put young people’s futures at risk, and may damage the economy for decades.

    A “Safe” pablum filled statement “The Best” is not worthy of an institution like the CDC. Unless you have data to back it up, that is. If you do, show it.

    • G says:

      Thank you … for some common sense in all of this! I have been saying the same thing you are saying since this all began. The costs for “slowing the spread” are astronomical in so many ways. We have not saved anyone by slowing the spread. We have just postponed it. If someone is at risk of dying from this virus, they will die whenever they get it. And, as this article explains, it will continue to mutate, so there is no real safety if you are at that risk anyway. All while destroying our freedom, our mental health, our economy, our social networks, our children’s futures, and so much more. It will take DECADES to recover from the damage caused by this failed strategy. Lives have literally been destroyed because of this strategy. But who will take responsibility for THAT?

    • wolf h. says:

      And what about the long-term symptoms, which nobody knows if they will remain with you? Who are those willing to live the rest of their life with brain fog, heart and lung issues and possibly Parkinson and or Alzheimer’s? Plenty of studies have been done about these possibilities. And another tidbit here, most of those with long-term complications are the ones who had a mild Covid-19 infection.

    • Nelson says:

      That was a lengthy response. What medical school did you graduate from? Thanks.

  • Trudi T-U says:

    I am not a scientist, but having lived with nearly dying a thousand times, immune deficient, a simple cold virus turns to lung infection for me quicker than you can spit – I warned last April that covid19 was nothing compared to the mutations we will see and not one government, not one scientist had the brains to stop travellers. Well, it is here and still travellers are carrying mutations – frankly – we get exactly what we deserve.

    • ndcsr says:

      Sorry to hear that. In this interconnected world, I don’t think you can stop SARS-CoV-2 coming in. We have a porous southern border, for one. The only way is to treat those with low risk as the front line, let them get infected, and help protect the elderly and others such as yourself. I am curious.

      Would you feel insulted if you were ASKED to self quarantine? In the big sense, it’s your risk, and your decision. If you knew that others were out there building up herd immunity, that would help protect you?

      Because that’s precisely what our leaders did not do. It seems to me it would be far better to warn those at risk to be very careful, and let those of us at low or tolerable risk live their lives, and keep the economy going.

  • Kim D.B. says:


  • T.E.M. says:

    I don’t want any virus that is given on a trial basis. The trial ends in 2023.

  • Kathleen says:

    thank you for your tireless work. I get so angry and a bit depressed watching Super Bowl celebrations of yelling and spreading spit. Why of why don’t they get it? or have we become so self centered they just don’t care? You are our heroes.

  • Andrew Goldstein says:

    “…maps can indeed be used to predict likely paths of viral evolution.”

    How fantastic this sounds, that scientists can actually predict the future outcomes of Darwinian evolution. Darwin observed that Nature exerts a survival selection pressure on all living things but I don’t think we have ever been able to predict it so precisely at the DNA-RNA level using our detailed knowedge of the immune system as both metric and crystal ball.

    Thank you for sharing this information.

    • Neal Dante Castagnoli says:

      Upon discussion further, it turns out that there are (n|k) [n choose k] variation, with “k” in the twenties and “n” in the thousands.

      So long as the escaping variation has a population to exploit, it’s usual, as I understand it, for clusters of additional mutations to pop up. So, unfortunately, it’s hard to predict the future of the virus. The impressive study kept “k” at “1”.

  • robert Dukes says:

    But at some point in this epidemic – and certainly in the countries that are most affected, like Italy and Spain – there will be saturation, because according to predictions, up to 40% percent of the Spanish and 26% of the Italian population are or have been infected already. And, of course, when you go over 50% or so, even without doing anything else, the virus just has fewer people to infect – and so the epidemic will come down naturally. And that’s what happened in in all the previous epidemics when we didn’t have any [treatments]. The rate of infection and the number of those susceptible will determine when that happens.

  • Dr. Sethuraman Subramanian says:

    Mapping mutations synthetically and predicting how they will be effective or not in escaping the antibodies is an academic exercise which may not be useful in the real world situation. Mutations occur randomly and not at all directed. The virus mutates in a given way not to escape detection but mutations occur randomly and those species in which mutations help the virus survive will continue to propagate. So synthetic mapping will not be of much use. Assume you get a real life mutation in the virus which escapes targeting by antibody, what do you do then? It is a whole new ballgame to devise a vaccine for that. All your academic exercise will be in vain.

    • Andrew Goldstein says:

      A “…mutation in the virus [glycoprotein] which escapes targeting by antibody” means that existing antibodies are no longer able to recognize the spike protein as non-self. Is this the most likely outcome of a mutation? I thought that mutations alter the structure of the spike protein such that antibodies raised by existing vaccines or post-infection antibodies no longer recognize the target but that a modified, mutation-specific vaccine could be made that would work as well as the original vaccine.

    • ndcsr says:

      I’m not certain that’s clear. It seems to me if you could assign probability to an “escape,” and you could quickly use contact tracing, you could have small stores of vaccine that could be provided around the escape, you might slow the spread enough to provide time to build up production of the necessary vaccine.

      Note, the “ifs.”

  • G. says:

    Has anyone looked at the role of vitamin A status in COVID-19 patients? About 20 years ago, there was a lot of discussion about the effects of supplementation for treating children with measles and pneumonia, and a follow-up about whether vitamin A was helpful with treating all forms of pneumonia or just measles specific pneumonia. (The outcome was that vitamin A was helpful in treating measles specific pneumonia.) Not all of us convert vitamin A optimally. About 45% of us are poor converters. Also, I see that in the past few years, it’s been found that a deficiency of vitamin A affects pancreatic beta cells which decreases their output of insulin. Would this be why diabetics are prone to more complications from Covid-19? I see that Dr. Trasino, who did one of the two studies on vitamin A and the functioning of beta cells published a letter advocating for a consideration of the use for retinoids in treating Covid-19. It may not be a magic bullet, but maybe it could reduce complicatons such as pneumonia.

  • Carol Rae says:

    It was notable that the inordinately rude posted comment by N.D. Castagnoli on Feb. 9 was more than inappropriately sent to a such a distinguished physician and scholar as Dr. Collins, who has earned the confidence of three U.S. Presidents, and who has recently been reconfirmed to his post by President Biden. He was, and is, attempting to educate in this Blog post. One might wonder what was YOUR motivation? Perhaps many of you readers have noticed that it is individuals with an ax to grind, and a paucity of real data, without the decades invested in professional education, service, research, and publication, who zero into comment sections like this, do not cite their professional qualifications, or lack thereof, and toss out every conceivable argument that they can think of, no matter whether or not they are building a rational line of logical reasoning, hoping to lure in the untutored? This method is the essence of the aggressive and dangerous providers of Covid vaccine and public policy disinformation, less than roughly 5,000 individuals total worldwide, according to Western intelligence services and IT disinformation profilers, who, nevertheless, have found these rants reproduced through Social Media some several hundred million times worldwide, amidst the untutored — inordinately complicating efforts to save lives, and economies, worldwide. It behooves us all not to opine, but to use our specialized knowledge to help, not to distract, and worse, to deceive!

    FYI: the simple answer, the cost of action or non-action which Mr. Castagnoli’s derides, lies in the number of lives damaged and lost, families without government safety nets bereft, despite every effort to save lives via medicine and public policy: In this 21st. Century Pandemic there have been 146,000,000 recorded cases worldwide through 4/24/21 (likely many more), with 3,000,000 recorded deaths (likely many more), and 585,000 unnecessary deaths in the U.S., again through 4/24/21. Blessedly, despite the early chaos in the worldwide responses, the lack of PPP and vaccine access, the overloaded hospital capacity, this time medical, research, and public policy learning have grown substantially, and we are NOT facing the abyss of the 1918-1919 H1N1 Pandemic of avian origin. The 20th. Century Influenza Pandemic infected an estimated 500 million, with 20-40, perhaps 50 million deaths worldwide, and 670,000 Americans lost their lives in the Pandemic. In that Pandemic, ALL prophylactic interventions were NON-pharmaceutical, the ones that Mr. Castagnoli ridicules: isolation, quarantine, masks, good personal hygiene, disinfectants, and limited gatherings. Without those proven methods, the 1918-1919 losses would have been much worse! Correspondingly, on the way to “Herd Immunity” (a bovine veterinary term inappropriately applied to humans) the numbers would have been much worse. We are so rich, so blessed with industry, technology, and medical science, that we can suffer some minor inconvenience, with our social safety nets and business support through Congressional Stimulus, and keep our deaths and illness to a minimum, while also building the capacity to vaccinate and treat the world. It may take five years, but no other country can do it. On this issue, we either put our head in the sand, or raise it proudly to help. FYI: I have two two-year Masters in Public Policy/Government from Johns Hopkins and Harvard University, and a Ph.D. from Harvard in Government (with Behavioral Science). I have served the U.S. public, the government, the public policy and philanthropic community, educational bodies, and academia since 1977.

Leave a Reply to wolf h.Cancel reply