Food-drug interaction studies are critical to evaluate appropriate dosing, timing, and formulation of new drug candidates. These interactions often reflect prandial-associated changes in the extent and/or rate of systemic drug exposure. Physiologic and physicochemical mechanisms underlying food effects on drug disposition are well-characterized. However, biochemical mechanisms involving drug metabolizing enzymes and transport proteins remain underexplored. Several plant-derived beverages have been shown to modulate enzymes and transporters in the intestine, leading to altered pharmacokinetic (PK) and potentially negative pharmacodynamic (PD) outcomes. Commonly consumed fruit juices, teas, and alcoholic drinks contain phytochemicals that inhibit intestinal cytochrome P450 and phase II conjugation enzymes, as well as uptake and efflux transport proteins. Whereas myriad phytochemicals have been shown to inhibit these processes in vitro, translation to the clinic has been deemed insignificant or undetermined. An overlooked prerequisite for elucidating food effects on drug PK is thorough knowledge of causative bioactive ingredients. Substantial variability in bioactive ingredient composition and activity of a given dietary substance poses a challenge in conducting robust food-drug interaction studies. This confounding factor can be addressed by identifying and characterizing specific components, which could be used as marker compounds to improve clinical trial design and quantitatively predict food effects. Interpretation and integration of data from in vitro, in vivo, and in silico studies require collaborative expertise from multiple disciplines, from botany to clinical pharmacology (i.e., plant to patient). Development of more systematic methods and guidelines is needed to address the general lack of information on examining drug-dietary substance interactions prospectively.

Natural products, including botanical dietary supplements and exotic drinks, represent an ever-increasing share of the health care market. The parallel ever-increasing popularity of self-medicating with natural products increases the likelihood of co-consumption with conventional drugs, raising concerns for unwanted natural product-drug interactions. Assessing the drug interaction liability of natural products is challenging due to the complex and variable chemical composition inherent to these products, necessitating a streamlined preclinical testing approach to prioritize precipitant individual constituents for further investigation. Such an approach was evaluated in the current work to prioritize constituents in the model natural product, grapefruit juice, as inhibitors of intestinal organic anion-transporting peptide (OATP)-mediated uptake. Using OATP2B1-expressing MDCKII cells and the probe substrate estrone 3-sulfate, IC50s were determined for constituents representative of the flavanone (naringin, naringenin, hesperidin), furanocoumarin (bergamottin, 6′,7′-dihydroxybergamottin), and polymethoxyflavone (nobiletin and tangeretin) classes contained in grapefruit juice juice. Nobiletin was the most potent (IC50, 3.7 μM); 6′,7′-dihydroxybergamottin, naringin, naringenin, and tangeretin were moderately potent (IC50, 20–50 μM); and bergamottin and hesperidin were the least potent (IC50, >300 μM) OATP2B1 inhibitors. Intestinal absorption simulations based on physiochemical properties were used to determine ratios of unbound concentration to IC50 for each constituent within enterocytes and to prioritize in order of pre-defined cut-off values. This streamlined approach could be applied to other natural products that contain multiple precipitants of natural product-drug interactions.

Panobinostat (Farydak) is an orally active hydroxamic acid-derived histone deacetylase inhibitor used for the treatment of relapsed or refractory multiple myeloma. Based on recombinant cytochrome P450 (P450) kinetic analyses in vitro, panobinostat oxidative metabolism in human liver microsomes was mediated primarily by CYP3A4 with lower contributions by CYP2D6 and CYP2C19. Panobinostat was also an in vitro reversible and time-dependent inhibitor of CYP3A4/5 and a reversible inhibitor of CYP2D6 and CYP2C19. Based on a previous clinical drug-drug interaction study with ketoconazole (KTZ), the contribution of CYP3A4 in vivo was estimated to be ∼40%. Using clinical pharmacokinetic (PK) data from several trials, including the KTZ drug-drug interaction (DDI) study, a physiologically based pharmacokinetic (PBPK) model was built to predict panobinostat PK after single and multiple doses (within 2-fold of observed values for most trials) and the clinical DDI with KTZ (predicted and observed area under the curve ratios of 1.8). The model was then applied to predict the drug interaction with the strong CYP3A4 inducer rifampin (RIF) and the sensitive CYP3A4 substrate midazolam (MDZ) in lieu of clinical trials. Panobinostat exposure was predicted to decrease in the presence of RIF (65%) and inconsequentially increase MDZ exposure (4%). Additionally, PBPK modeling was used to examine the effects of stomach pH on the absorption of panobinostat in humans and determined that absorption of panobinostat is not expected to be affected by increases in stomach pH. The results from these studies were incorporated into the Food and Drug Administration-approved product label, providing guidance for panobinostat dosing recommendations when it is combined with other drugs.

Sonidegib (Odomzo) is an orally available Smoothened inhibitor for the treatment of advanced basal cell carcinoma. Sonidegib was found to be metabolized primarily by cytochrome P450 (CYP)3A in vitro. The effect of multiple doses of the strong CYP3A perpetrators, ketoconazole (KTZ) and rifampin (RIF), on sonidegib pharmacokinetics (PK) after a single 800 mg dose in healthy subjects was therefore assessed. These data were used to verify a physiologically-based pharmacokinetic (PBPK) model developed to 1) bridge the clinical drugdrug interaction (DDI) study of sonidegib with KTZ and RIF in healthy subjects to the marketed dose (200 mg) in patients 2) predict acute (14 days) versus long-term dosing of the perpetrators with sonidegib at steady state and 3) predict the effect of moderate CYP3A perpetrators on sonidegib exposure in patients. Treatment of healthy subjects with KTZ resulted in an increased sonidegib exposure of 2.25- and 1.49-fold (area under the curve and maximal concentration respectively), and RIF decreased exposure by 72% and 54%, respectively. The model simulated the single- and/or multiple-dose PK of sonidegib (healthy subjects and patients) within ∼50% of observed values. The effect of KTZ and RIF on sonidegib in healthy subjects was also simulated well, and the predicted DDI in patients was slightly less and independent of sonidegib dose. At steady state, sonidegib was predicted to have a higher DDI magnitude with strong or moderate CYP3A perpetrators compared with a single dose. Different dosing regimens of sondigeb with the perpetrators were also simulated and provided guidance to the current dosing recommendations incorporated in the product label.

Ritonavir is one of several ketoconazole alternatives used to evaluate strong CYP3A4 inhibition potential in clinical drug-drug interaction (DDI) studies. In this study, four physiologically based pharmacokinetic (PBPK) models of ritonavir as an in vivo time-dependent inhibitor of CYP3A4 were created and verified for oral doses of 20, 50, 100 and 200 mg using the fraction absorbed (F ) and oral clearance (CL ) values reported in the literature, because transporter and CYP enzyme reaction phenotyping data were not available. The models were used subsequently to predict and compare the magnitude of the AUC increase in nine reference DDI studies evaluating the effect of ritonavir at steady-state on midazolam (CYP3A4 substrate) exposure. Midazolam AUC and C ratios were predicted within 2-fold of the respective observations in seven studies. Simulations of the hepatic and gut CYP3A4 abundance after multiple oral dosing of ritonavir indicated that a 3-day treatment with ritonavir 100 mg twice daily is sufficient to reach maximal CYP3A4 inhibition and subsequent systemic exposure increase of a CYP3A4 substrate, resulting in the reliable estimation of f . The ritonavir model was submitted as part of the new drug application for Kisqali® (ribociclib) and accepted by health authorities.