Pharmacometrics uses mathematical models of biology, pharmacology, disease, and physiology to describe and quantify interactions between drugs and patients, including beneficial effects and adverse effects. I recently had the pleasure of talking to a thought leader, Dr. Lawrence Lesko, about the history of pharmacometrics and how it will continue to shape drug development in the future.
Dr. Lesko was Director of the Office of Clinical Pharmacology in the Center for Drug Evaluation and Research (CDER) at the Food and Drug Administration (FDA) for 16 years until his retirement in July 2011. He currently serves as Clinical Professor and Director of the Center for Pharmacometrics and Systems Pharmacology in the University of Florida, College of Pharmacy at Lake Nona in Orlando, FL.
My trends in drug development Q&A with Dr. Lesko
Q. You’ve probably participated in countless interviews over the years. So, I wanted to start our conversation with something a little bit different: If you had one superpower, what would it be and why?
A. Nobody has ever asked me this before! [Laughter] This question made me think of the recent 30th anniversary of the film, “Back to the Future.” The film is based the ability to travel back through time. It occurred to me that it’d be really cool to be able to be able to travel back in time to historical health care game-changing moments that occurred over the past century. The first place that I’d go would be to 1953. I’d love to be able to visit James Watson and Francis Crick when they discovered the structure of DNA. Also, I’d want to ask them what their vision for health care for the next 50 years was based on their discovery, especially in terms of precision medicine.
Q. Speaking of going back in history, can you discuss how you become involved in the field of pharmacometrics?
A. My favorite quote that I like to give to my students is that “I never thought that I’d get into this business.” That’s because my training is not in pharmacometrics. It wasn’t even around when I was a grad student!
My involvement in the field began in 1991. This was right before I joined the FDA. I was attending a conference called “The Integration of PK (pharmacokinetics), PD (pharmacodynamics), and TK (toxicokinetics): Principles of Rational Drug Development.” Sitting in the room and listening to the talk, it occurred to me that this approach was a blueprint for the development of quantitative pharmacology, or pharmacometrics.
The leaders of pharmacometrics at the time were Carl Peck, who was the director of the FDA Center Drug Evaluation and Research (CDER) and Bob Temple, who was the Chief Medical Officer at the FDA. They were both champions of the “learning elements” of pre-clinical and early clinical drug development, based on the concepts of pharmacometrics. At the time, pharmacometrics was only based on concepts. There was no significant use of pharmacometrics at either the FDA or in the pharmaceutical industry.
That’s when I began to pick up the banner of quantitative pharmacology. As an FDA officer, I endorsed the “learn/confirm paradigm” which was published by Lewis Sheiner in 1997. I also advocated for model based drug development in the Critical Path Initiative. And I installed a very proactive pharmacometrics group when I was [CDER Office of Clinical Pharmacology] director. So, that’s how I got into the field.
Q. You mentioned getting the dose right. From your vantage points of academia and the agency, how else do you think that biosimulation has influenced drug development?
A. I think that we need to first define biosimulation. It’s a term that it used in many different domains. I’ll define it as a discipline that uses computer-based mathematical simulations of biological processes. Biosimulation is generally applied to pharmacokinetics and pharmacodynamics.
More recently, biosimulation has been used in the context of systems pharmacology. I define “systems” to mean the response of a target (a cell signaling pathway, organ, tumor, etc) to a drug treatment. Biosimulation has impacted getting the dose right by informing the analysis of PK/PD data.
But, the reason why we use biosimulation is to improve our ability to predict outcomes. Let me give you a few examples. Getting the dose right is a universal tenet of all modern drug development programs.
Four examples further illustrate the benefits of biosimulation. First, fifteen years ago, 20% of drug candidate attrition was due to poor ADME (absorption, distribution, metabolism, and excretion) characteristics. Now, the attrition rate is down to 1-3% because of ADME problems. That’s due to the use of modeling and simulation. Because, we can simulate not only structure-activity relationships, we also can simulate dissolution, pharmacokinetic profiles, and more.
Secondly, there are numerous pediatric indications that were approved on the basis of PK/PD models that bridged human adult data to pediatrics. Modeling and simulation benefits pediatric drug development by reducing the regulatory burden in conducting pediatric efficacy trials.
My next example focuses on the use of physiologically-based pharmacokinetics (PBPK) to predict clinical drug-drug interactions (DDIs) and drug-gene interactions in untested scenarios. Biosimulation approaches enable these predictions to be made from a limited number of clinical trials. Such information is now almost routinely being included in the labels of FDA-approved drugs.
And the final example is that biosimulation software has become an integral part of drug development. The Simcyp Simulator is one example. The emergence of biosimulation software is de facto evidence that it’s important to be able to precisely represent biological phenomena such as drug metabolism, drug-drug interactions (DDIs), and the exposure-response relationship.
Q. If you were to gaze into a crystal ball, what would you predict as the biggest emerging trends in the pharmaceutical industry in the next 10 years?
A. Now you’re asking a really tough question! There are many emerging trends, but I’ll touch on a few. First, I think that we’ll find diminishing returns, at a huge cost in drug development, for new monotherapies for important diseases. I predict that we’ll move away from monotherapies towards combination therapies, or ‘drug cocktails,’ for the treatment of cancer and Alzheimer’s disease, to name a few. The reason is that diseases are complex. And therefore, drug cocktails will be important to effectively treat complex diseases.
In addition, PhRMA [the Pharmaceutical Research and Manufacturers of America] is beginning to include environmental factors in drug development. What’s your diet like? How much do you exercise? Do you eat bacon? What’s your energy turnover based on your body composition? A shift away from monotherapies towards combination therapies will better address the multifactorial nature of disease pathology. There’s going to be more and more reliance on biosimulation to integrate combination therapies and environmental factors into clinical trials to optimize medical product treatments.
Q. That makes a lot of sense to me. Do you think that the trend toward drug cocktails will make PBPK approaches for predicting DDIs even more important?
A. Yes, exactly! As you add drugs to a patient’s regimen, the potential combinations grow exponentially. Biosimulation is going to answer the following questions:
- What drugs should I add together to maximize the benefits and minimize toxicity?
- What doses of each drug should I give?
- How frequently should I give the drugs?
- Should I give the drugs during the morning hours or at night?
Another big trend is for rare diseases and orphan drugs. Half of the drugs that the FDA approved in 2014 were for orphan drugs. In the future, drugs are going to be developed for smaller and smaller groups. Why is this? For one, orphan drugs have significant financial returns. The top 10 most expensive drugs in the United States right now are orphan drugs. Secondly, breakthrough designations by the FDA are multiplying. These are generally being awarded to drugs for smaller and smaller patient populations. Finally, working with smaller patient subgroups aids a better understanding of disease pathophysiology because you have removed a lot of the interindividual variability of large population disease states.
For example, in cancer, genomics has given us the ability to characterize tumors on a molecular basis, rather than on an organ basis. Right now, we can sequence an entire tumor genome at a reasonable cost. It wouldn’t surprise me if many currently fatal cancers, like pancreatic or lung, become chronic diseases in 10 years.
Likewise, I see a trend towards repurposing older approved medications. Some medications may not have worked for very well for the general population. These ‘failed’ drugs might be revitalized for genomic sub-groups. A lot of these drugs can also attain an improved risk-benefit profile through new dosage delivery systems such as liposomes.
Our current approach to treating high cost patients will also change. For example, consider patients with heart failure. They can take drugs for their condition, but they don’t work very well. So, they end up being frequently re-admitted to the hospital. This imposes a very high cost to taxpayers and society in general. There will be a greater future focus on developing drugs that can prevent readmission to the hospital for costly conditions like heart failure.
An increased use of data from wearable devices (a form of eHealth) will complement drug development. I wear a Fitbit® device and my Nike® sneakers have a device in the sole that tells me how many steps I take in a day. I can also use devices that monitor my heart rate, my sleep, and the amount of calories I burn on a walk or run. In ten years, I imagine that we’ll be able to measure almost any physiological data that we want to know about ourselves using wearable devices. Biosimulation plays into this trend because you cannot make sense of all these data without computer modeling. For more information, I recommend that you read Larry Smarr’s work about the “quantified self” in Biotechnology Journal (2012, 7, 980-991).
My last prediction from the “crystal ball” is that there will be a shift away from treating disease to developing drugs that maintain health. In the next decade or so, people won’t wait until they are sick to visit a doctor. They will have wearable device e-health technology to monitor their own health and warn them of impending health crises. This knowledge will empower them to seek medical help, take their medications, and initiate lifestyle changes that will keep them healthy. We’ll use this information to turn fatal diseases into chronic diseases and reduce the societal impact of devastating diseases like Alzheimer’s.
The bottom line for all these trends is that the pharmaceutical industry is using technology to bring increased value to patients.
Q. What would be your advice for a young pharmacometrician just starting her career?
A. The importance of being a strong communicator cannot be overstated. The ability to communicate clearly can be especially difficult for scientists for whom English is not their native language. I try to instill this skill in my students and post-doctoral fellows at the University of Florida.
A few years ago while I was at the FDA, I hired a speech coach to work with regulatory scientists. The reviewers were videotaped giving a scientific presentation, and then they had follow-up 1:1 sessions with the speech coach. It made a huge and noticeable difference in their presentation skills.
Chatting with a thought leader in drug development
I’d like to thank Dr. Lesko for pulling up a seat at the Roundtable and sharing his insights with us. Learn more about his work at the University of Florida Center for Pharmacometrics and Systems Pharmacology.
Join the biosimulation revolution
Biosimulation can influence every phase of the drug development process. It has the power to help develop new treatments for deadly diseases. Read our white paper, “The Benefits of Biosimulation in Drug Development” to learn how biosimulation transforms data into information and information into knowledge.