Images courtesy of the Virtual Physiological Rat Project, Medical College of Wisconsin and the laboratory of Aron Geurts, Ph.D.
Researchers around the world are spearheading an ambitious initiative aimed at understanding and predicting physiological function in health and disease. What's unusual about this initiative is that the team—a network of investigators from the United States, Europe and New Zealand—isn't simply running thousands of in vitro and in vivo experiments as they try to unearth the mechanisms underlying disease and dysfunction. Instead, they are working with computational tools and computer simulations based on the physiology of a rat.
The project, aptly named the Virtual Physiological Rat (VPR), "will fill a major gap in the understanding of the interactions among multiple genetic and environmental causes of diseases," says principal investigator Daniel Beard, Ph.D., professor of physiology at the Medical College of Wisconsin (MCW). "In the process, we are likely to uncover genes and proteins that are potential drug targets, as well as physiological information that would influence pharmacokinetics and pharmacodynamics. That is why we are especially eager to partner with pharma. If industry scientists tell us what computational models they can use for specific applications, then we can really help each other."
The VPR project got its start in 2011 with a five-year, $13 million grant from the U.S. National Institutes of Health that established MCW's National Center for Systems Biology. "We are working toward the grand challenge of biomedical research: understanding the complex interplay between physiological, genetic and environmental factors," says Beard. Genome-wide association studies typically explain only a small percentage of the variation in complex diseases, he explains. Therefore, "our ability to predict the outcome of genetic variability is negligible. We need to figure out how the heritable information in an organism's genome encodes that organism's phenotype, and what mapping this information tells us about disease."
The VPR project is tackling this challenge by integrating computational modeling of a rat's molecular genetic and physiological systems with experimental studies on strains with cardiovascular phenotypes relevant to human disease. They also will engineer new rat strains to see if their computer models can predict the physiological characteristics of those strains, and will investigate how the new strains respond to various physiological and environmental inputs.
The researchers started by studying the physiology of healthy rats with known genomes to learn how the animals' hearts, kidneys, skeletal muscles and blood vessels work together on a molecular level to maintain normal function. They then created computer models of normal function, both to help analyze data from animal experiments and to generate hypotheses to test in the lab. Now they are in the process of measuring cardiovascular function in rats whose genes have been linked to complex diseases such as high blood pressure, and adding this information to the computer models. The goal is to be able to predict—by examining links between molecular functions and specific genes and the environmental factors that affect their expression—the state of a rat's (or human's) cardiovascular health and eventually, what might be done to ameliorate it.
Beard's MCW team is collaborating on the project with researchers at UW-Madison, University of Washington-Seattle, University of California-San Diego, North Carolina State University, King's College London, the Norwegian Life Sciences University and the University of Auckland in New Zealand.
Underlying their efforts, Beard explained in a presentation at the National Centers for Systems Biology Annual Meeting in July 2012, are the project's "governing principles"—that "complex diseases manifest on the background of physiological control," and that "computational physiology is a vehicle for genotype-to-phenotype mapping."
Why develop a virtual rat to investigate complex human diseases instead of a virtual human? Well, for one thing, the Virtual Physiological Human (VPH) project already exists! But VPH, which is a European Union endeavor, "is a much bigger program than VPR and has different goals," Beard explains. VPH is creating virtual humans based on data (biological, imaging, clinical, genomic) from actual patients. The aim is to improve the understanding of human physiology and pathology and, thereby, to develop and test new therapies. "The focus of VPH is more on clinical translation and commercialization, whereas VPR is really about basic science." That said, "ultimately, we will connect the two projects with our software and databases, and that will help us to translate our findings from rats to humans." Moreover, the VPR acronym was chosen specifically to show the project's intellectual connection to VPH—both of which are derived from the IUPS (International Union of Physiological Sciences) Physiome Project, which had its genesis at the IUPS 32nd World Congress in Glasgow in 1993.
Maintaining that connection with collaborators in the UK and Norway (as well as New Zealand) can only benefit the VPR, Beard says. "Europe has moved more conservatively in terms of fashionable trends in science, and as a result their scientists have a stronger foundation in traditional integrative physiology—which is akin to what we call ‘systems biology' today—than we do in the United States. That's because our physiology departments, responding to recent trends, mostly became molecular genetics departments, whereas in Europe, systems physiology has remained a strong discipline. So those researchers bring that heritage and knowledge to the table—although, of course, our aim is to share information and expertise across countries and across disciplines."
Rats and humans share about 90% of their genes, including those for most known diseases. Even specific genetic sequences are about 85% identical between the two, according to Beard. In addition, both live rat models and computer models of rat physiology already have been used extensively to study various aspects of cardiovascular disease—but not the ways in which multiple genes and environmental factors interact to cause them.
"The virtual rat model allows us to address some fundamental questions we simply can't answer directly with a human model, because obviously we can't genetically engineer or experiment on humans," Beard says. And though the virtual rat won't do away with the need to use actual animal models, "when we use the virtual model to analyze data and draw insights from it, we certainly are in a position to use animals more efficiently by orders of magnitude. And to the degree that the rat model is accurate enough, it shortens the experiment time frame and lowers considerably the cost of pre-compound drug testing."
VPR projects are underway in the following areas: cardiovascular dynamics, heart, kidney, metabolism and transport and whole-body function. The cardiovascular dynamics team is developing models of circulatory system dynamics and respiratory gas exchange in different phenotyping experiments—for example, response to acute occlusions of major blood vessels, or to blood withdrawal. Since previous work in rats has used models that spanned a wide range of species and experimental conditions, the heart group is developing canonical species and robust condition-specific cardiac models to bolster the model's predictive capabilities. The kidney team is developing models of renal blood-flow regulation and renal-solute transport to help inform the role of renal function in hypertension.
Because heart failure, diabetes and related diseases show impaired glucose and fatty acid oxidation, as well as diminished energy states in various tissue/organ systems, the metabolism and transport team is developing models of solute transport and energy metabolism in those tissue/organ systems, as well as in the body as a whole. The whole-body function group is working with open-source physiological and cellular model simulation packages to simulate the integrated operation of the cardiovascular, heart, metabolism and transport and renal models to predict and understand whole-body phenotypes, such as blood pressure. This knowledge will allow the novel rat strains team to create animal models that will be used to test what happens when a particular gene or gene product is dysfunctional in tissues of the heart or kidney.
What does all this mean to drug discovery and laboratory automation scientists? "The low-hanging fruit for pharma is in physiologically based pharmacokinetics, which actually means we're looking at both pharmacokinetics and pharmacodynamics together, in an integrated way," says Beard.
"Our computational models allow us to predict whole-body kinetics, rather than fitting some kinetic transport parameters to data. By generating predictions of blood flows and blood volumes throughout the body, and with a bit of information about partitioning that can come from independent in vitro data, we are in the position of being able to predict the whole-body pharmacokinetics of any drug. In addition, our kidney models will give us information about filtration, and—as we go further with the metabolic and transport models—we'll be able to predict transformation. All this will allow us to translate pharmacokinetics/pharmacodynamics from model organisms to humans with an unprecedented level of sophistication."
The VPR also will enable scientists to do in vitro to in vivo extrapolation "with a lot less tweaking and maneuvering," Beard asserts. "If, for example, we have targets for particular enzymes in mitochondria, and we're measuring how specific compounds change the rates of oxidation of fatty acids or substrate utilization, we can almost immediately and in parallel probe how those changes affect whole-body energy metabolism in skeletal muscle. Or, if we're looking at a compound that might have an effect in Type 2 diabetes, we can look at a model of normal energetics and a model of diabetic energetics and see what that compound would do in our model of the whole system." From a laboratory automation perspective, he adds, "it would be great to be able to do many of the requisite in vitro experiments in some automated way."
Although the project is underway, partners and collaborators are welcome, Beard emphasizes. In terms of skills, "there's no knowledge that's not useful—but we particularly need people skilled in computational biology and computational physiology. We've got biochemists, we've got chemists, but also engineers, mathematicians and physicists and it is often a real struggle to get everybody to be able to communicate with each other. No matter what your Ph.D. is in, the important thing is to be able to think in an integrative, multidisciplinary way. You can't be an engineer who doesn't know any biochemistry and you can't be a biochemist who's afraid of numbers and computers."
Meanwhile, Beard and his colleagues are in the process of creating software that would enable other scientists—partners or not—to take advantage of the knowledge that is being gained through the VPR. "We continue to publish our methods and models in the traditional way, and that information is available in scientific journals right now. But for this kind of systems biology project to take hold, the traditional way of disseminating information is not efficient," Beard says. Although he acknowledges a penchant to be "overly optimistic," he envisions having open-source software available for downloading on the VPR website within a year or so.
At the same time, Beard will be presenting at various scientific meetings and looking to make connections. "We especially want to create applications that are useful to people working in drug discovery and development," he stresses. "Tell us what you need and we will work with you."
February 15, 2013