The Open Philanthropy Project recommended a grant of $11,367,500 over five years to the University of Washington’s Institute for Protein Design (IPD) to support research on the use of computational protein design to develop a universal influenza vaccine. This work will be led by UW Professor David Baker, Ph.D., and UW Assistant Professor Neil King, Ph.D. In addition, part of this funding is intended to improve the Rosetta molecular modeling and design software originally developed in Baker’s lab. UW Assistant Professor Frank DiMaio, Ph.D., and others will work to improve Rosetta to better predict the properties of proteins, which we believe could lead to many helpful applications in both human and animal health, such as facilitating faster creation of antiviral therapies in the event of a pandemic outbreak.
Developing a “universal” influenza vaccine has been a long-standing goal of the infectious disease community. Influenza causes an estimated 290,000 - 650,000 deaths each year,1 in addition to carrying pandemic risks, such as the 1918 flu pandemic, which may have killed more than 50 million people. There can be up to 36 million cases of influenza annually in the United States alone, costing the U.S. economy an estimated $87 billion in medical expenses and lost productivity.2
A universal vaccine would protect against all of the many genetic variants of influenza, thereby substantially reducing or eliminating annual deaths associated with the virus, as well as mitigating or even potentially eliminating the risk of influenza pandemics. By administering a universal vaccine early in life, immunity could reach a majority of the world population and potentially eliminate the need for the “seasonal” flu vaccine, thereby reducing the costs and uncertainty associated with current annual influenza vaccination practices.
We are most interested in supporting this work due to its potential to reduce illness and mortality from influenza, as well as the potential to prevent future harm from influenza pandemics. The grant therefore falls within both our scientific research funding category, and our biosecurity and pandemic preparedness focus area.
About the grant
The project will be led by Dr. David Baker,3 director of the University of Washington’s Institute for Protein Design, founded in 2012. We consider Dr. Baker and his team exceptionally well-qualified to accomplish the universal influenza vaccine design.
Dr. Baker and colleagues will also collaborate with Dr. Barney Graham, deputy director of the National Institute of Allergy and Infectious Diseases’ (NIAID) Vaccine Research Center and chief of the Viral Pathogenesis Laboratory at NIAID. Dr. Graham and his colleagues will be responsible for overseeing evaluation of the safety, immunogenicity, and protective capacity of the universal flu vaccine. NIAID is one of the institutes within the National Institutes of Health, which is funding Dr. Graham’s work.
There are two goals associated with this grant: producing a universal influenza vaccine and improving the Rosetta molecular design and computational chemistry software to better predict properties of proteins.
- Goal 1: Develop a universal flu vaccine
Dr. Baker and his collaborators will test the hypothesis that immunizing with computationally designed, self-assembling protein nanoparticles displaying highly conserved “epitopes” (fragments of proteins from the flu virus) from all known types of influenza viruses will induce expression of broadly neutralizing antibodies that limit the severity of disease symptoms. Although very rare, such antibodies have been preserved for research after being identified in the serum of people who have previously been infected with influenza virus. When the corresponding genes were expressed in transgenic mice, the antibodies protected the mice against all flu viruses tested.4 We consider this evidence that the approach may work, but we expect the challenge to be developing methods to routinely induce such antibodies with a vaccine in humans. Drs. Baker and King previously used a somewhat similar approach to produce broadly neutralizing antibodies against the respiratory syncytial virus (RSV), which we consider a partial proof of principle. However, because RSV is less variable than influenza, there were fewer challenges to producing that vaccine candidate. Our assessment is that there is an acceptable probability that the proposed experiments will produce a useful flu vaccine that is better than anything currently available.
- Goal 2: Improve the Rosetta molecular design and computational chemistry software to better predict properties of proteins
Dr. Baker’s approach to Goal 1 has recently become possible because, during the past several decades, his team and other leading protein chemistry groups have systematically developed sophisticated algorithms that use information about the chemical and physical properties of molecules to facilitate the design and production of proteins with desired three-dimensional structures. This “ab initio” (or “from the beginning”) design approach has made progress but sometimes fails to predict the correct protein structure because some aspects of protein chemistry are not yet completely understood. It is also difficult to make such proteins in a way that accurately resembles the natural three-dimensional structure in the virus, something crucial to a protein’s normal function (for example, generating a robust immune response). Thus, additional research is needed to fully realize the long-term goal of routinely being able to design novel proteins for a wide variety of uses in medicine and industrial applications.
Some of this grant funding will therefore support experiments that may facilitate improvements in the ability of the Rosetta software to predict the properties of proteins from their amino acid sequences. Of particular interest, we envision that in the future it may be possible to rapidly design new vaccines or antiviral drugs from the nucleic acid sequences of a new pathogen and to design vaccines that induce immunity against all strains of a pathogen in a relatively short period of time. Drs. Baker and DiMaio have proposed a series of experiments that are designed to improve the predictive power of Rosetta. In brief, they will expand their work in computationally designing self-folding protein scaffolds with shapes customized for applications including binding small molecules and other proteins. They will then systematically synthesize, test, and redesign until the desired properties are achieved. Dr. Baker believes (and we concur) that this approach has a high probability of leading to improvements in our ability to design proteins. We see this approach as timely because it will potentially generate enough information to take advantage of the power of deep learning algorithms to understand protein structure and function.
This work is related to Goal 1 above because one of Dr. Baker’s goals is to be able to design therapeutic candidates within weeks of determination of the viral sequence in the event of a viral pandemic outbreak. Therefore, the results are expected to be of general utility. We consider Dr. Baker’s previous work constructing small geometric protein nanoparticles to be particularly suited to mimicking viral structure and antigen presentation.
- 1. “Illnesses range from mild to severe and even death. Hospitalization and death occur mainly among high risk groups. Worldwide, these annual epidemics are estimated to result in about 3 to 5 million cases of severe illness, and about 290 000 to 650 000 deaths.” World Health Organization Fact Sheet, 2018
- 2. “In the worst years, in the U.S. alone, seasonal flu causes up to 36 million infections, three-quarters of a million hospitalizations and 56,000 deaths … Annual funding to find a universal vaccine has never approached even $100 million a year, even as the seasonal flu has cost the U.S. economy an estimated $87 billion a year.” Time, Our Complacency About the Flu is Killing Us, 2018
- 3. Archived copy of link: IPD, David Baker Bio, January 2018 [archive only]
- 4. From the abstract “Neutralizing antibodies that target epitopes conserved among many strains of influenza virus have been recently isolated from humans. Here we demonstrate that adeno-associated viruses (AAV) encoding two such broadly neutralizing antibodies are protective against diverse influenza strains. Serum from mice that received a single intramuscular AAV injection efficiently neutralized all H1, H2 and H5 influenza strains tested. After infection with diverse strains of H1N1 influenza, treated mice showed minimal weight loss and lung inflammation. Protection lasted for at least 11 months after AAV injection. Notably, even immunodeficient and older mice were protected by this method, suggesting that expression of a monoclonal antibody alone is sufficient to protect mice from illness. If translated to humans, this prophylactic approach may be uniquely capable of protecting immunocompromised or elderly patient populations not reliably protected by existing vaccines.” Nature Biotechnology, Broad protection against influenza infection by vectored immunoprophylaxis in mice, Balazs, AB., et al (2013)