PI-002 - APPLICATION OF VIRTUAL TWIN PBPK MODELS IN OBESE WITH CYP3A4 PHENOTYPING USING ENDOGENOUS BIOMARKER DATA.

N. Izat¹, H. Kangne¹, I. Robertsen², A. Rostami-Hodjegan^1,3, A. Galetin¹; ¹The University of Manchester, Manchester, United Kingdom, ²University of Oslo, Oslo, Norway, ³Certara Predictive Technologies, Sheffield, United Kingdom.

Background: Developing virtual twins within physiologically-based pharmacokinetic modelling (VT-PBPK) is an emerging strategy to predict individual drug exposure and facilitate model-informed precision dosing [1]. Plasma 4β-hydroxycholesterol (4β-OHC), CYP3A metabolite of cholesterol, is an endogenous biomarker (EB) for CYP3A activity. In this study, it was utilised for development of VT-PBPK models for obese individuals.

Methods: 4β-OHC/cholesterol ratio and midazolam plasma concentration data were available from obese and morbidly obese individuals before (n=36) and 2 years after (n=15) the gastric bypass surgery [2]. CYP3A4 abundance was either predicted from the EB phenotyping data or measured values obtained from the liver biopsies during surgery were used. Real patients were ‘virtualized’ as shown in Figure 1 and the systemic exposure of midazolam was simulated using Simcyp v.23 R package.

Results: Predicted CYP3A4 abundance using the EB data was within 2-fold of the measured abundance in 90% of the individuals. VT models informed with EB data predicted in vivo AUCinf,iv of midazolam within 2-fold of observed value in 83% of individuals, with geometric mean fold error (gmfe) of 1.57 (n=36). VT models using CYP3A4 actual protein measurements showed similar performance (86% prediction success, gmfe=1.56), verifying the EB approach. Two years after the surgery, BMI of obese individuals was decreased, whereas 4β-OHC levels increased over time. VT models successfully predicted midazolam PK after 2 years in the same individuals following weight loss (gmfe=1.44; n=15).

Conclusion: The study highlights the utility of 4β-OHC for predicting midazolam exposure within VT-PBPK modelling concept. This biomarker-informed approach can be used for individual dose requirements of intravenously administered CYP3A substrates in obese individuals.

[1] Mostafa et al. CPT Pharmacometrics Syst Pharmacol. 2024; 13(3):424-436.
[2] Kvitne et al. Eur J Clin Pharmacol. 2022; 78(8):1289-1299.