Lisa Cheng 1, Po-Chang Chiang 2, Matthew R Wright 3, Cornelis E C A Hop 3, Harvey Wong 1
1 Faculty of Pharmaceutical Sciences, The University Of British Columbia (Vancouver, Canada), 2 Small Molecule Pharmaceutical Sciences, Genentech Inc. (South San Francisco, United States of America), 3 Drug Metabolism and Pharmacokinetics, Genentech Inc. (South San Francisco, United States of America)
Introduction
Gastrointestinal tract physiology is influenced by prandial state, with characteristic differences in gastric pH, gastric emptying time, bile salt secretion among other factors that can alter oral drug absorption. Food effects, a phenomenon that describes the change in drug exposure after the intake of a high-fat meal, can be investigated using physiologically based pharmacokinetic (PBPK) models prior to clinical studies. A notable difference in the stomach compartment of physiological models in the fasted and fed state is the increase in gastric pH from pH 2 to pH 5. Despite successful food effect predictions for various compounds, there remains a challenge in predicting the food effect of ionizable compounds caused by the transient increase in gastric pH followed by reacidification after food consumption. For example, pictilisib is a weak base, and its dissolution is limited in the fed stomach, at pH 5, but improves in a more acidic environment. The co-administration of pictilisib and a proton pump inhibitor, which effectively maintains an elevated gastric pH, reduced drug exposure by approximately 50% in the fasted state [1]. This result highlights the critical role of an acidic environment in governing pictilisib’s dissolution. As such, the process of the stomach returning to a more acidic state over time after meal consumption is a dynamic variable that impacts absorption.
Objectives
This study aims to investigate the value of incorporating gastric reacidification into a PBPK model in the fed state by evaluating the accuracy of food effect predictions for pictilisib.
Methods
A PBPK model consisting of one-compartment stomach, one-compartment small intestine, and one-compartment colon was created using the Simulation, Analysis, and Modeling II software (SAAM II v2.3, Nanomath LLC; Spokane, WA, USA) [2]. Human physiological parameters and compound-specific parameters were retrieved from literature [1,3–5]. The model was first validated using fasted state clinical data and subsequently adjusted to predict the fed state. The current approach to the postprandial stomach involves a static, elevated pH of pH 5 and termed the S1CS model in this study. The goal to incorporate a more physiologically representative fed gastric pH profile was addressed using two different approaches: 1) MCS model: the stomach is represented by four compartments with decreasing pH starting at pH 5, followed by pH 3.5 and pH 2, before finally reaching pH 1.5 and 2) D1CS model: the one-compartment stomach configuration is kept with the addition of a dynamic pH curve over time after food consumption. Postprandial pH data [6] was fitted to a one-phase exponential decay equation using GraphPad Prism Version 8.4.0 (Boston, Massachusetts, USA). Fed to fasted ratios for AUC0-inf (area under the curve from time 0 h to infinity) and Cmax (maximum plasma concentration) were used to determine whether a potential food effect is predicted, then compared against observed clinical data to evaluate model performance regarding food effects.
Results
There were no inherent differences among the S1CS, MCS, and D1CS model predictions in the fasted state. The models were considered validated since the model-predicted exposure parameters were within a two-fold range of fold error when compared to the observed clinical data. The reported fed to fasted fold change for pictilisib suggests no clinical food effect for AUC0-inf (1.18-fold fed to fasted change) and a possible clinically significant negative food effect when evaluating Cmax (0.77-fold change) [1]. The U.S. Food and Drug Administration defines a food effect as a fed to fasted AUC and Cmax outside the range of 80% to 125% [7]. The S1CS model prediction indicates the presence of a negative food effect with a 0.74-fold decrease in AUC0-inf and similarly, a lower Cmax (0.50-fold change). The MCS model suggested no food effect as both AUC0-inf (1.00-fold change) and Cmax (0.92-fold change) are within the 0.80- to 1.25-fold thresholds. Lastly, the D1CS model simulations had a 0.84-fold change in AUC0-inf (no food effect) and 0.58-fold change in Cmax suggesting a possible negative food effect. The MCS model outcome offered a smaller numerical deviation in fed to fasted fold change of observed pictilisib exposure. However, the D1CS model was able to reflect the clinical data by correctly categorizing an absence of a food effect AUC0-inf and a negative food effect for Cmax.
Conclusions
A weakly basic compound in a less acidic environment undergoes reduced dissolution relative to more acidic environment and this may lead to underestimation of drug dissolution in the fed gastric compartment as showcased by pictilisib. Two approaches to include gastric reacidification after a high-fat meal were used to showcase the importance of a more physiologically representative fed stomach to minimize false, negative food effects for weak bases. Although a dynamic stomach pH may not be necessary for all compounds, weakly basic compounds have shown sensitivity to changes in gastric pH. Overall, these results suggest that the MCS and D1CS models provided an improved mechanistic modeling approach in comparison to the static pH 5 (S1CS model) for investigating food effects.
References:
1. Ware JA, Dalziel G, Jin JY, Pellett JD, Smelick GS, West DA, et al. Impact of Food and the Proton Pump Inhibitor Rabeprazole on the Pharmacokinetics of GDC-0941 in Healthy Volunteers: Bench to Bedside Investigation of pH-Dependent Solubility. Mol Pharmaceutics. 2013;10:4074–81. https://doi.org/10.1021/mp4005595
2. Perazzolo S. SAAM II: A general mathematical modeling rapid prototyping environment. CPT Pharmacom & Syst Pharma. 2024;13:1088–102. https://doi.org/10.1002/psp4.13181
3. Davies B, Morris T. Physiological parameters in laboratory animals and humans. Pharm Res. 1993;10:1093–5. https://doi.org/10.1023/A:1018943613122
4. Cheng L, Chiang P-C, Wright MR, Wong H. A multicompartment stomach physiologically-based pharmacokinetic model capturing gastric reacidification can improve food effect predictions of weak bases. Journal of Pharmaceutical Sciences. 2026;115:104155. https://doi.org/10.1016/j.xphs.2026.104155
5. Lu T, Fraczkiewicz G, Salphati L, Budha N, Dalziel G, Smelick GS, et al. Combining “Bottom‐up” and “Top‐down” Approaches to Assess the Impact of Food and Gastric pH on Pictilisib (GDC‐0941) Pharmacokinetics. CPT Pharmacom & Syst Pharma. 2017;6:747–55. https://doi.org/10.1002/psp4.12228
6. Dressman JB, Berardi RR, Dermentzoglou LC, Russell TL, Schmaltz SP, Barnett JL, et al. Upper Gastrointestinal (GI) pH in Young, Healthy Men and Women. Pharmaceutical Research. 1990;07:756–61. https://doi.org/10.1023/A:1015827908309
7. U.S. Food and Drug Administration. Assessing the Effects of Food on Drugs in INDs and NDAs – Clinical Pharmacology Considerations: Guidance for Industry [Internet]. Silver Spring, MD, USA: U.S. Food and Drug Administration; 2022 p. 1–15. https://www.fda.gov/media/121313/download
Reference: PAGE 34 (2026) Abstr 12113 [www.page-meeting.org/?abstract=12113]
Poster: Drug/Disease Modelling - Absorption & PBPK