Show Me the Human Data
Regina Kelder

Show Me the Human Data

Animal models are really actors re-creating roles about the course of human diseases. But they can only re-enact part of the story.

When a drug performs well pre-clinically only to flop in a pivotal Phase II or III trials or benefit just a tiny percentage of the patient population it was designed to help, it inevitably raises questions about the quality of the preclinical data. Sometimes it’s the limitations of the model. How many times have we heard the phrase “mice are not little people” to explain the disconnection between animal disease models and their human counterparts.

Poor outcomes are also driven by gaps in our understanding of the complexity of the diseases playing out in individuals. For instance, why do checkpoint inhibitors, which harness the immune system to fight tumors, lead to durable remissions in about a quarter of patients, yet deliver short-term benefit or no responses in the remaining 75%? Why has the quest for an AIDS vaccine been so elusive when other common diseases are vaccine preventable? Why do thousands of drug candidates for Alzheimer’s disease, stroke, heart disease and countless other conditions appear robust enough in preclinical studies, yet not pan out in the clinic?

These are not easy questions. Many scientists are trying to answer them by applying reverse translation, a bedside to bench approach using human data to inform new discoveries, identify new pathways and help refine preclinical models of disease.

Better tools and technologies are enabling scientists to collect voluminous amounts of data from a small piece of tissue, and then interrogate multiple parameters of that tissue using sophisticated tools.

Using human genome-wide association studies, scientists can identify genetic variants not implicated previously in the pathogenesis of a certain disease. They can also determine that known genetic variations do not always translate into higher risk of disease. In a different vein, one can analyze diseased tissue and look for molecules that might relate to the disease process, says Alan Daugherty, PhD, Senior Associate Dean for Research and Director of the Saha Cardiovascular Research Center at the University of Kentucky School of Medicine.

“The concept of reverse translation is critical to the way we do preclinical work,” says Daugherty, who will be speaking about reverse translation in atherosclerosis at the World Congress meeting sponsored by Charles River in Boston this Fall. “It is much easier to generate enthusiasm in which the initial target is derived from GWAS studies.

Below are three examples of how human data is refining preclinical drug discovery.

Help for heart disease

Reverse translation strategies have helped identify new treatments for atherosclerosis, notes Daugherty. He says unbiased human genetics approaches have led us to previous unknown pathways and a new class of drugs to combat the major trigger of atherosclerosis.

Atherosclerosis occurs when fat, cholesterol and other substances build up in the walls of arteries and forms structures on the inside walls called plaques.Over time, these plaques precipitated blockage in arteries in specific locations in the body.

Low density lipoprotein (LDL) is generally accepted to be a primary driver of atherosclerosis, and individuals with atherosclerosis are usually treated with cholesterol-lowering statins. However, statins don’t work effectively for everyone, and some people can’t tolerate them due to side effects.

About 15 years ago, Canadian scientists discovered a novel family of proteins that activate other proteins derived from a gene located on chromosome 1. Simultaneously, French scientists had been following a family with familial hypercholesterolemia (FH), a genetic disorder of unusually high LDL that leads to coronary artery disease in 90% of cases and is often fatal. The French group had identified a mutation on chromosome 1 in some patients, but did not know what the gene was. The two labs got together and eventually linked mutations in the gene —dubbed PCSK9—to FH. 

“No one had ever heard of PCSK9 or remotely related it to atherosclerosis or cholesterol metabolism until a human study found there were patients in mutations in this gene that had very high levels of cholesterol and increased risk of atherosclerosis and heart disease,” says Daniel Rader, MD, a heart specialist who directs the Preventive Cardiovascular Medicine and Lipid Clinic at the University of Pennsylvania Health System. “That put this on molecule on the map and other [labs] found human mutations that caused low cholesterol and protected from heart disease. This really opened up the whole biological field about what does PCSK9 do and how does it affect cholesterol and atherosclerosis.

Scientists eventually identified a new therapeutic target which led to a new class of cholesterol-lowering drugs called PCSK9 inhibitors, two of which reached the market last year.

This is not the only example where human data is helping to inform the translation of animal models of heart disease. Using GWAS data, Rader’s colleague, Dr. Muredach Reilly, was the first to associate a gene variant called Adamts7 with coronary disease.

By knocking out the gene in mice, Reilly, Rader and their colleagues were able to show two years ago that it reduced atherosclerosis in mice, and efforts are now underway to develop inhibitors of Adamts7 that reduce atherosclerosis and coronary disease.

“The huge value of studying humans using these unbiased methods is that we find things we didn’t know to look for,” says Rader. “Otherwise we end up doing preclinical studies on things we already know about, hoping we can show that it is ultimately relevant to human disease. When we discover the association with humans we know it is relevant to the question we are asking.”

Checking the checkpoints

With the idea that no two cancer cases are unique, cancer immunotherapy researchers are taking a personalized approach. Clinical, research and biomarker teams are working together to identify who responds to an experiment drug and who does not. For those who do not respond, their unique biology is investigated by bench scientists.

Biomarker and clinical teams then evaluate how that specific tumor evolved and generate hypotheses to explain why it did not respond to treatment. The goal is to try and develop RNA fingerprints or profiles that give a sense of the general phenotypic characteristics of tumors, and that can help in the development of next-generation drugs that might achieve higher response rates. To show how this is playing out on a federal level, the National Health Service in the UK is now asking all cancer patients to provide them with a sample of their DNA so that better, individually tailored therapies (so-called personalized medicine) can be administered.

The quest for an AIDS vaccine

AIDS scientists are adopting certain principles of reverse translation to try and develop a safe and effective preventive vaccine--an unmet goal 36 years into the epidemic. Today, one major objective of HIV vaccine research is determining what vaccine immunogens will induce antibodies that can neutralize a broad swath of the diverse strains of HIV in circulation, so-called broadly neutralizing antibodies (bNAbs). But development of bNAbs is not favored—only a minority of HIV-infected individuals develop them and only after exposure to a rapidly evolving virus.

So, scientists worked backwards to try and solve this problem. They first collected blood samples from a large sample of HIV-infected individuals spanning five continents, and from that isolated a handful of antibodies that exhibited broad and potent neutralizing activity. Such antibodies typically emerge about two years following infection.

Researchers have since been able to design an immunogen that binds to one of these bNAbs. They have also been able to chart, in unprecedented detail, how bNAbs evolve over the course of two and a half years in an infected patient, as well as locate the viruses that developed mutations in response to the maturing antibody. This research could eventually help lead us to a vaccine candidate that elicits a potently protective response against HIV, and, in fact, researchers plan to design immunogens that can be tested in animals and then in humans that might induce bNAbs like the one in this patient.

In the meantime, two large international studies are now recruiting patients for periodic infusions of one particularly potent bNAbs—passive immunization—to see if the antibodies protect against infection following exposure to HIV.

This September, the Inaugural Charles River World Congress on Animal Models in Drug Development will explore how to use clinical data to develop translatable models for drug discovery. There, Dr. Alan Daugherty along with dozens of other industry leaders will present their thoughts on the future of translational medicine. Register today by visiting: