One of the cool things about being a medical scientist is going to international conferences in places all over the world. I’m not saying it’s a good reason to become one, but when you end up in this line of work, conferences abroad are certainly the cherries on the pie (or the “raisins in the porridge” as we would say in the Netherlands). Just now, I’m flying back from San Diego where I attended the annual meeting of the American Society for Clinical Pharmacology and Therapeutics. There was a keynote lecture by dr Allen Shuldiner, professor of medicine and physiology at the University of Maryland, Baltimore and Vice President of Regeneron Genetics, a biotech company that aims to find new disease targets based on extensive genome wide screens in large populations before developing targeted therapies. Dr. Shuldiner started his talk by saying that humans are the best model for human disease, which was his way to make it clear that we should rather study the genetic factors contributing to disease in humans than to do animal studies, if we want to learn more about the etiology. I couldn’t agree more and I would even go further and argue that we should also rather use humans than animals as models for human disease when we test our new drugs. My own PhD research was focused on finding new treatments for amyotrophic lateral sclerosis (ALS) and included several animal experiments using the G93A SOD1 mouse model for ALS. Some of the experiments were positive, some were negative, but that’s not the issue. The problem is that the predictive value of a positive G93A SOD1 mouse experiment is nearing zero! There has only been one drug that had a positive effect in this animal model that also worked in humans -causing a very modest but statistically significant prolongation of life expectancy of several months- while the number of drugs that worked in the diseased mice that didn’t work in humans is steadily increasing to probably over 100 by now, and counting. Thinking about the waste of time and money is already discouraging, but if you consider that the reverse might also be true, that there will be ample compounds that tested negative in animal experiments, which might have worked in humans, should really make you feel utterly depressed!
And that is only considering efficacy issues. The recent disaster in France with the FAAH inhibitor of Bial has again shown that the safety of drugs can also not be guaranteed or predicted based on animal experiments. While the real problem with that particular study might have been related to the unnecessary urge of companies to elevate dose levels to what is maximally tolerated in humans instead of to the levels where the pharmacodynamic effects no longer increase -a different issue that that I may decide to cover in a future blog- , it does also point out that animal drug studies are far removed from human reality.
And why not use humans as models for human disease? We have developed a wide array of pharmacological and non-pharmacological challenges that will lead to temporary states resembling human disease in healthy subjects. Anti-muscarinic and anti-nicotinic challenges cause temporary cognitive disturbances and can be used to show effects of compounds that enhance cognition, inhaled Δ9-tetrahydrocannabinol or low doses of iv ketamine lead to symptoms reminiscent of psychosis and can be used to show effects of antipsychotics, electrical, heat, cold or pressure stimuli lead to diverse types of pain and are used to show effects of analgesics and we even infuse very low quantities of lipopolysaccharide derived from bacteria to induce symptoms related to systemic inflammation to show effects of anti-inflammatory drugs. Positive studies with new compounds that affect these symptoms, occurring in humans and not in animal models, have a very high predictive value of success in patients and can even predict the correct dose levels. In our experience, therefore, real live humans are indeed the best model for human disease, which is why we will remain the Centre for Human Drug Research and will advocate the use of our highly predictive human disease models in early phase drug development.