The First World War, one of the world’s deadliest conflicts, claimed approximately 20 million lives. But in the year that it ended, an even deadlier calamity swept the globe. The 1918 influenza pandemic is estimated to have killed between 50 million and 100 million people1. Within months, a simple virus took a greater toll on human life than had the brutal four-year war. Although flu vaccines now save countless lives and have undoubtedly helped to avert other worldwide pandemics, their manufacture must vary annually to match the current circulating viral strains. Flu continues to pose a threat to human health, and there is a pressing need to develop countermeasures that can confer broad protection against the different flu strains. Moreover, vaccines are usually less effective at providing protection for children and for older individuals than they are for the rest of the population2.
Writing in Science, Laursen et al.3 report that they have engineered antibodies that protect against diverse flu viruses in mice, and that notably provide protection against most viral strains from the two major types of virus that cause human disease: influenza A and influenza B. Obtaining such broad protection has been a challenge because both influenza A and influenza B are composed of diverse strains, and developing ‘universal’ protection has been an elusive goal. If the authors’ approach could be adapted for effective use in humans, it might help to prevent or contain the spread of new and evolving flu infections worldwide.
During the 1918 pandemic, the cause of the disease was unknown. Had a vaccine been available, it would probably have limited the global catastrophe. However, developing an effective flu vaccine is not straightforward because flu viruses can mutate rapidly4. The high level of mutation results in continuous variation in two key viral proteins over time. One of these is haemagglutinin, which is located on the surface of the virus (Fig. 1) and recognizes a molecule on host cells that serves as a receptor for viral attachment and entry.
Haemagglutinin also associates with a viral protein called neuraminidase. There are 18 distinct subtypes of haemagglutinin and 11 versions of neuraminidase. These two proteins form the basis for how flu strains are named. For example, the designation H1N1 indicates that a flu virus has haemagglutinin subtype 1 and neuraminidase subtype 1.
A breakthrough in attempts to offer protection against diverse flu strains came with the identification of antibodies, termed broadly neutralizing antibodies, that can bind to a highly evolutionarily conserved and invariant structure in a region of haemagglutinin termed the stem5,6. Such antibodies battle the flu virus by binding haemagglutinin and inhibiting the virus’s ability to enter cells. They can also boost an antiviral response, for example by engaging immune cells that promote the killing of virus-infected cells. However, these antibodies usually do not recognize all flu viruses. For instance, broadly neutralizing antibodies that recognize haemagglutinin of one major genetic subgroup of influenza A, group 1, typically do not react against the second group, group 2, and also do not recognize influenza B (ref. 7)7.
To try to target influenza A and influenza B viruses, Laursen and colleagues had the idea of engineering an antibody by ‘stitching’ together influenza-recognition domains from different antibodies that bind to evolutionarily conserved regions of haemagglutinin, especially in the stem region of this protein. The authors vaccinated llamas (Lama glama) with a flu vaccine or haemagglutinin proteins, and used in vitro tests to identify resulting antibodies that had the greatest potency and breadth of neutralization against diverse flu viruses. They found that specific combinations of these antibodies could target nearly all of the flu-virus strains tested. Llama antibodies have a simpler structure and are smaller than human antibodies, and therefore aid an engineering approach that seeks to combine protein regions from more than one antibody.
By engineering antibodies in which several influenza-recognizing regions were connected by protein linkers, the authors were able to create antibodies that targeted multiple viruses. And fusing such structures to an antibody structure called the Fc region enabled such chimaeric proteins to interact with and activate immune cells.
When mice received either engineered antibodies or the gene encoding such an antibody — delivered by means of an adeno-associated virus (AAV) into cells of the nasal passage — they were protected against a flu virus that would usually have been lethal. The gene-delivery approach ensured production of the antibody for weeks to months, providing sustained protection without the need for multiple rounds of antibody injections over time.
Whether this approach could be used to prevent flu in humans is uncertain. Mice do not serve as optimal models for investigating human influenza because the receptor used by viral strains to infect mouse cells is a different version from that needed for cellular entry into human cells. In addition, the patterns of tissue infection and virus in the bloodstream often differ between mice and humans8. Protection in mice can involve a pathway that is mediated by a receptor protein called FcγR-III on immune cells that recognizes antibodies bound to targets9, but whether this type of immune mechanism has relevance for humans is unknown. Antibodies that target the stem region of haemagglutinin have so far failed to alleviate symptoms in humans who are already infected, and the ability of these antibodies to prevent infection is being tested in clinical trials10.
Another worry regarding this approach in humans is whether an immune response might be triggered against the non-human antibodies. Although an engineered llama antibody has been approved for clinical use to treat a blood-clotting condition11, whether an immune response is generated against the anti-flu multi-domain antibodies will become clear only with clinical testing. Llama antibodies can be ‘humanized’ (engineered to closely resemble the related domains of human antibodies), yet the efficacy of such modifications would need to be evaluated in humans.
Also of concern is the use of AAV, because there are limitations to achieving sufficient and sustained levels of gene expression when using this virus in gene-therapy treatments12. Other safety and regulatory concerns regarding AAVs relate to their use to drive continuous gene expression, because this raises the possibility that complexes of human antibodies bound to the engineered antibodies might form over time. That said, certain groups, such as older people, might especially benefit from engineered antibodies, given the high mortality rates from flu in such individuals, and the fact that their immune responses tend to be less robust than those of younger adults.
The expression of engineered antibodies through gene-delivery approaches might offer a way of preventing or treating diverse types of infectious disease. Moreover, the outcomes of such treatments might help to confirm useful targets for the development of antiviral drugs or vaccines. For example, if broadly neutralizing antibodies that target the stem region of haemagglutinin can prevent flu infection in vivo in humans, it would encourage efforts to generate such antibodies by vaccination approaches. Antibodies targeting the stem region of haemagglutinin have previously been generated using structure-based approaches for vaccine design, and have shown promise in preclinical tests using animal models13–15.
Laursen and colleagues’ approach of generating an antibody that can target more than one site is reminiscent of earlier work in which an antibody was developed16 from broadly neutralizing antibodies to target three independent sites on the HIV virus. Such antibodies can neutralize more than 99% of circulating HIV strains. This antibody blocked infection from viruses that were unaffected if single antibody components of the trispecific antibody were used. The era of multi-specific target engagement by engineered antibodies has begun, and might lead to new countermeasures to protect human health.
Nature 565, 29-31 (2018)
Competing Financial Interests
G.J.N. and J.W.S. are employees of Sanofi, whose Sanofi-Pasteur vaccine subsidiary develops and manufactures influenza vaccines and whose Ablynx subsidiary develops nanobody medicines. G.J.N. is chief scientific officer and J.W.S. is director of Sanofi-Pasteur vaccine research and development.
G.J.N. is listed on intellectual property developed and owned by the National Institutes of Health that relates to the development of novel influenza vaccines.