Jeff Saucerman temporarily sets modeling aside for high-throughput drug discovery

Associate Professor of Biomedical Engineering Jeffrey Saucerman’s work illustrates the versatility of computational methods. Saucerman is dedicated to finding ways to treat heart failure, an irreversible, inevitably fatal condition that afflicts five million Americans. Saucerman and the members of his Cardiac Systems Biology Lab strive to identify molecular networks and control points that decide how the heart responds to stress, with the overall objective of rewiring these molecular networks to prevent or reverse heart failure.

Saucerman has used computation both to build virtual models of how these networks work and to screen thousands of compounds for those that might stimulate heart tissue to regenerate. “Computation is giving us a variety of starting points to attack this problem,” Saucerman said.

open Jeff Saucerman's Portrait

Jeffrey Saucerman, PhD, Professor of Biomedical Engineering, combines computational models and high-throughput experiments to discover molecular networks that control cardiac remodeling and regeneration.

Using Computation to Explore Heart Regeneration

The root cause of heart failure is a loss of cardiac myocytes, often induced by heart attack or hypertension. For years, scientists assumed that mammals, unlike some other organisms, have very limited ability to regenerate cardiac muscle, but recent studies have identified genes implicated in generating new heart tissue during early development. “These findings suggest that there is a basis for pursuing therapies that amplify this untapped capacity to repair damaged heart tissue,” Saucerman said.

The problem, as Saucerman describes it, is that we are at the very early stages of the biology. “We can’t model the process of cardiac myocyte regeneration because we don’t have a lot of data to begin with,” he said. As a result, Saucerman set modeling aside, at least temporarily, for high-throughput drug discovery.

In collaboration with AstraZeneca, the multinational pharmaceutical company, Saucerman assembled a group of approximately 5,000 compounds from the company’s library of molecules and screened it for those stimulating cardiac myocyte growth in the lab. This approach serves the practical end of identifying candidates for drug development. At the same time, the nature of these candidates can provide clues to underlying mechanisms of cardiac myocyte regeneration.

“One of the exciting things about this process is that the hits we get could open up exciting opportunities for computational exploration by suggesting pathways that may not have been known before,” Saucerman said.

Automating and Accelerating the Process

A decade ago, sorting through thousands of potential compounds was a laborious, painstaking process. Fortunately, Saucerman and his team could take advantage of advances in laboratory automation and image analysis software to prepare hundreds of samples at a time, image them under a microscope and examine these images to identify instances of cell growth. “Once the samples are ready, you can proceed quite quickly,” he said. “You can image a plate with 384 wells in 30 minutes.” The first pass identified 48 candidates.

Commercially available image analysis software that Saucerman used for this initial screen has its limitations. It cannot distinguish between actual cell division and other processes that might increase the amount of DNA and the number of nuclei in a sample. Saucerman’s team developed a more discriminating assay, creating algorithms that examined each image for a greater number of characteristics, and applied these methods to the top hits from the initial screen. “On the basis of our secondary assay, we narrowed our candidate list to 28 compounds,” he said.

A Foundation for Collaboration

Saucerman’s results have paved the way for a series of collaborations with other researchers at the University of Virginia. He is conducting genomic and proteomic experiments to profile the molecular responses of a number of his top candidates and combining this information with bioinformatic analysis from Stefan Bekiranov, an associate professor of biochemistry and molecular genetics. Their goal is to develop computer models of the underlying signaling networks.

At the same time, he is collaborating with Dr. Matthew Wolf, associate professor in the Division of Cardiovascular Medicine, to take the next step in drug discovery by testing these compounds in vivo. Wolf has developed mouse models with fluorescent reporters that measure cardiac myocyte proliferation. They are co-mentoring a post-doctoral fellow, Cody Narciso-Widmer, who is using engineering analysis to improve the process and applying their image analysis algorithms to scan heart tissue from these mice.

“We are using computational methods to streamline and refine virtually every aspect of this project,” Saucerman said. “It is one of the advantages that engineers bring to biomedical research.”

Saucerman Lab 2020

The Cardiac Systems Pharmacology Group at the University of Virginia