Drug-drug interactions (DDIs) are a primary threat to the safety and efficacy of clinical practice. Clinically relevant drug interactions are primarily due to drug-induced alterations in the activity and quantity of metabolic enzymes and transporters. Indeed, DDIs that cause unmanageable, severe adverse effects have led to restrictions in clinical use and even, drug withdrawals from the market. There is a real need to be able to predict variability in the extent of DDI for individuals that could be missed in a clinical trial.
While there has been notable success in predicting potential DDIs caused by alterations in cytochrome P-450 (CYP) metabolism, there are still major challenges in assessing “Complex DDIs.” DDIs become more complex when they involve more than two drugs, multiple CYP, non-CYP and active transport processes and metabolites of any of the drugs involved. Moreover, the prediction of DDIs is further complicated by “Complex Patients”—patients with organ impairment, pediatrics, or other special populations where the conduct of clinical trials is either not ethical or feasible. In this blog post, I’ll discuss how mechanistic predictions using physiologically-based pharmacokinetic (PBPK) modeling and simulation (M&S) are increasingly being used to address these complex scenarios. Ultimately, further refining these approaches will enable personalized assessment of DDI potential during drug development and also, potentially by clinicians at the point of care.
Developing mechanistic models for complex DDIs
DDIs are often caused by inhibition or induction of drug metabolizing enzymes, particularly CYPs. But there are other significant drug metabolic pathways, in addition to those mediated by CYPs. Liver cells also express transporter proteins like the organic anion transporting polypeptides (OATPs) that facilitate the uptake of many drugs. A number of clinically relevant DDIs can be attributed to drug-induced inhibition of hepatic transporters. A complete consideration of complex DDIs requires incorporating transporter-enzyme dynamics in the liver.
Historically, pharmacokinetics research has focused on the liver. The intestine and kidney also express drug metabolizing enzymes and drug transporters. Thus, for orally administered drugs, intestinal drug disposition can significantly impact the amount of drug released systemically. This explains the increased attention on modeling the contribution of intestinal metabolism to DDIs.
Role of metabolites in causing DDIs
While most research has examined the role of parent drugs in causing DDIs, drug metabolites can also contribute to the risk of DDIs. In recent years, various regulatory agencies, including the FDA, EMA, and PMDA, have issued guidance documents encouraging sponsors to consider metabolites during DDI risk assessment. The need to evaluate metabolites stems from the fact that the majority of clinically significant CYP inhibitors have metabolites that circulate at concentrations greater than 25% of the parent drug. Yet, there are a lack of in vitro data on these metabolites that would allow evaluation of their contribution to DDI risk. Resolving this issue will depend on developing methods that can predict which metabolites appreciably add to DDI risk.
Regulatory agencies advocate use of modeling to predict untested clinical scenarios
Clearly, comprehensive assessment of DDI risk is complicated by a number of factors including multiple biological mechanisms that influence drug clearance as well as variability within the patient population. Since it is not feasible to evaluate all potential scenarios in clinical trials, PBPK modeling can aid predicting untested, clinically relevant scenarios. This methodology is increasingly being used in drug development and regulatory review— including informing the drug label. For a more in-depth examination of this topic, I recommend reading the Biopharmaceutics & Drug Disposition review by Varma and colleagues, Dealing with the complex drug–drug interactions: Towards mechanistic models.
Moving towards personalized medicine
The future of quantitative pharmacology is in personalized drug dosing. Virtual Twin™ technology will be an important step towards making this vision a reality. The idea is to match the characteristics of a real patient with his or her virtual twin in order to predict individual risk of complex DDIs. This matching would happen at several levels:
- Age, weight, height, sex, and ethnicity
- Current drug dosage and co-medications
- Activity of metabolic enzymes and transporters
- Level of organ function
Realization of this technology would allow clinicians to first try different drug doses, schedules, and combinations in the virtual twin to determine an optimal dosing regimen for the patient. For a deeper dive into this topic, I’ve included the slides from a recent talk “Use of Modeling and Simulation to Assess and Manage Individualized Risk of Drug-Drug Interaction.” I was fortunate to be able to present this talk at a meeting hosted by The Wellcome Trust. The meeting was a very interesting discussion with key clinicians and research scientists in the UK on the potential future for complex DDI prediction at the point-of-care.
Ready to learn more about the applications of physiologically-based pharmacokinetic modeling?
Drug-metabolizing enzymes and drug transporters play an important role in hepatic drug metabolism and disposition and therefore have major implications on the fate of drugs in the human body. Recently, an increasing number of studies have reported proteomic expression data for pharmacologically relevant enzymes and transporters providing rich data that can be used for simulation and modeling. However, systematic analysis and understanding of the expression of these proteins in human tissue and commonly used in vitro systems are lacking.
New LC-MS techniques in quantitative proteomics are transforming the approach to quantitative pharmacology and in vitro-in vivo extrapolation (IVIVE) for PBPK models. My colleague, Dr. Amin Rostami, and Dr. Brahim Achour (University of Manchester) recently gave a webinar where they discussed the scientific and economic implications of this approach. I hope that you’ll watch it and let me know what you think in the comments section.