Protein science 2.0: Amping up antibodies

Institute for Protein Innovation antibody libraries
The Institute for Protein Innovation, launching next week with $15 million in grants and philanthropy, aims to develop comprehensive, open-source libraries of antibodies targeting human proteins.

It began with the proteins. Before Watson and Crick unraveled DNA’s double helix in the 1950s, biochemists snipped, ground and pulverized animal tissues to extract and study proteins, the workhorses of the body.

Then, in 1990, the Human Genome Project launched. It promised to uncover the underpinnings of all human biology and the keys to treating disease. Funding for DNA and RNA tools and studies skyrocketed. Meanwhile, protein science fell behind.

While genomics unveiled a wealth of information, including the identity of genes that lead to disease when mutated, researchers still do not fully understand what all the genes really do and how mutations change their function and cause disease.

Now proteins are promising to provide the missing link.

Biological chemist, immunologist and structural biologist Timothy Springer, PhD, is reinventing protein science with new high-tech technologies. A seven-time entrepreneur, Springer is launching the Institute for Protein Innovation (IPI). The one-of-a-kind nonprofit aims to create a massive, openly available resource of protein technologies to accelerate drug discovery and development.

Institute for Protein Innovation founder Timothy Springer, PhD

“Proteins lag behind DNA and RNA in institutional research support and funding,” says Springer, an investigator at the Program in Cellular and Molecular Medicine at Boston Children’s Hospital. “Yet proteins serve as the targets of almost all drugs and in many cases as therapeutic drugs themselves.”

Accelerating antibodies

Nearly half of all drugs on the market today are proteins, mostly monoclonal antibodies that home in on specific targets in the body, just as our own antibodies target proteins made by disease-causing organisms. Researchers spent decades producing antibodies in the lab that blocked or activated proteins’ function. This helped investigators figure out what genes and proteins do in the body.

Seeing their success, drug developers began seeking to mass-produce antibodies that would target protein culprits in diseases such as cancer. However, the development of antibodies for use as therapeutics faced – and still faces – major hurdles.

First, antibodies are traditionally produced by immunizing animals like mice with a target human protein. That process is often slow and not always effective: If the mouse and human proteins are too similar, the mouse immune system won’t recognize the human protein as foreign and won’t make an antibody. This leaves many extremely important molecules without targeting antibodies.

Institute for Protein Innovation antibody libraries
A mouse antibody

Second, many antibodies used in published research studies have not been properly validated through extra testing. This makes it difficult to reproduce the results of pivotal antibody-based studies, slowing scientific progress.

Probing for proteins

To circumvent these problems, researchers including British biochemist Greg Winter invented robust “molecular display” technologies. They involve creating immense libraries of unique antibodies or antibody fragments.

Researchers insert the genes for antibodies into one-celled organisms such as yeast, modified so that the antibodies made by the yeast are displayed on their surface. Billions of yeast cells, each bearing a unique antibody, are then mixed with the protein of interest. The final step is fishing out the yeast that bind to the target proteins most strongly, through a two-step process:

Institute for Protein Innovation fishing for antibodies
Fishing for antibodies from yeast libraries: To screen for antibodies with high affinity for their target protein, researchers first coat the protein with magnetic nanoparticles, then use a magnet to fish for the yeast whose surface antibodies bind to it. These yeast then enter a second fishing step to assess the strength of antibody binding, using fluorescent labeling of the proteins and fluorescence-activated cell sorting (FACS). The brightest yeast cells, with the most fluorescently tagged protein, are those whose antibodies bind most strongly. They are separated from the dimmer cells through flow cytometry. (More on this technique)

Because the yeast make the antibodies as instructed by inserted genes, the organisms can be used to create antibodies to molecules that a mouse immune system wouldn’t respond to. In addition, the technique can be scaled up for the high-throughput approaches of genomics and proteomics.

But there are other problems yet to solve. Many proteins that would be used as targets in the antibody selection process are still extremely difficult to produce. Expertise in antibody manufacture is also lacking.

Springer believes the solution is an infrastructure to make, develop and validate synthetic antibodies and their target proteins at a scale never before attempted. Partnering with co-founder Andrew C. Kruse, PhD, at Harvard Medical School (HMS), Springer will formally launch the IPI on May 10. Its first major effort will be to develop open-source libraries of well-validated antibodies targeting every human protein found outside cells.

Academic entrepreneurship

Unlike Springer’s other commercial start-ups, the IPI is an academic-entrepreneurial hybrid. By straddling the traditional divide, Springer hopes to create alliances between leaders in academic research, biotechnology, pharma and biomedical investing, while providing open-source resources to the scientific community at little or no charge. The institute, currently housed at HMS, will publish its technical processes and share its know-how freely, equipping the next generation of protein scientists.

Springer intends for the IPI to stay afloat through a mix of public and philanthropic investment (including a $10 million gift from Springer himself). With the 2018 NIH budget poised to be cut 18 percent, the effort could not have come at a better time.

And there is probably no better person to launch it than Springer. He conducted his postdoctoral studies with Nobel Laureate César Milstein, who co-invented monoclonal antibody technology in 1975. This invention has led to far more drugs than have come from genomics. Springer sees IPI as the embodiment of Milstein’s vision of using antibodies to target disease.

“César would love this,” says Springer. “He really wanted to have the promise of antibodies fulfilled.”