Translational methods for quantitative prediction of metabolic herbal product-drug interactions: case study with milk thistle

Brantley, Scott J.

Download PDF

Request Version for Screen Reader

Last Modified

March 22, 2019

Creator

Brantley, Scott J.
- Affiliation: Eshelman School of Pharmacy, Division of Pharmacotherapy and Experimental Therapeutics

Abstract

The misperception that herbal products are safe has perpetuated multibillion dollar sales of these products, exposing the public to potentially harmful herb-drug interactions when constituents in the herbal supplement inhibit drug metabolizing enzymes. Regulation of herbal products is not as rigorous as drug products. Consequently, evaluation of inhibitory properties of an herbal product typically is not requested before marketing. Traditional drug-drug interaction evaluation methods often are inadequate to evaluate herbal product interaction liability due to the mixture of bioactive constituents, high inherent variability between batches and manufacturers, and limited pharmacokinetic knowledge of constituents. Milk thistle was selected as an exemplar herbal product due to high usage rates in patient populations, particularly the hepatically-impaired; availability of isolated, purified constituents; and disparate effects between previous clinical interaction studies. Initial screens of inhibitory activity against the clinically relevant drug metabolizing enzymes, cytochrome P450 (CYP) 2C9 and CYP3A4, prioritized milk thistle constituents for further evaluation. The main constituents, silybin A and silybin B, inhibited CYP2C9 in a reversible manner (Ki, 10 and 4.8 μM, respectively) and CYP3A4 in an irreversible manner (KI, 110 and 89 μM, respectively). Incorporation of these in vitro kinetic parameters into a physiologically based pharmacokinetic (PBPK) model facilitated predictions of the interaction liability of milk thistle administration with FDA-recommended probe substrates of CYP2C9 (warfarin) and CYP3A4 (midazolam). Administration of large doses of the milk thistle product silibinin (1440 mg/day) was predicted to increase the peak concentration and systemic exposure of both warfarin and midazolam by roughly 5%. Proof-of-concept clinical evaluation of these silibinin-drug interactions confirmed the low interaction potential of the selected milk thistle product, as midazolam and warfarin exposure was increased modestly (9 and 13%, respectively). This mechanistic modeling and simulation approach facilitated prospective evaluation of interactions between a well-characterized herbal product and two widely used and clinically relevant probe substrates. This framework could be applied to other herbal products to predict the magnitude and likelihood of interactions with conventional drugs, guide pharmacotherapeutic decisions, and improve patient care.

Date of publication