Evaluating alterations in drug-metabolizing enzymes in cancer: insights from in vivo, in vitro, and biomarker studies.

Drug-metabolizing enzymes (DMEs) are critical determinants of drug elimination, yet hepatic and intestinal DME changes between cancer and healthy subjects remain controversial. To address this uncertainty, a physiologically based pharmacokinetic (PBPK) approach was employed to evaluate how reported DME alterations influence drug pharmacokinetics by comparing predicted-to-observed clinical exposures of substrates for cytochrome P450 (P450) enzymes, including CYP1A2, 2C8, 2C9, 2C19, 2D6, and 3A4. A systematic review of clinical, biomarker, and in vitro studies identified key DME changes in cancer. Four P450 abundance scenarios were applied in a PBPK model: baseline (no change in P450 abundance), in vivo clearance-based changes, IL-6-mediated P450 suppression, and in vitro proteomic predictions from healthy and cancerous livers. Model predictions of Cmax and area under plasma concentration-time curve for P450 major substrates were compared with clinical data, with accuracy assessed by average fold error, absolute average fold error, and Δ absolute average fold error relative to the baseline model. PBPK simulations showed no need to adjust most P450 enzyme levels in patients with cancer to match clinical data. However, intravenous and oral simulations indicated reduced intestinal CYP3A4 expression, without hepatic enzyme level reductions, was required for oral CYP3A major substrate pharmacokinetics to align with observations. Model comparison suggests a 50%-90% reduction in intestinal CYP3A4 abundance in cancer, with 90% best fitting the broader oral dataset. In conclusion, current clinical, biomarker, and PBPK evidence do not support substantial DME changes in patients with cancer, particularly for non-CYP3A enzymes. Additional clinical and mechanistic studies are needed to improve understanding and guide personalized oncology pharmacotherapy. SIGNIFICANCE STATEMENT: This work explores how cancer alters drug-metabolizing enzymes by combining clinical, biomarker, and proteomic data with physiologically based pharmacokinetic modeling to test competing hypotheses on P450 abundance. It shows that most P450 can be left unchanged in cancer models, whereas extensive reduction of 50%-90% in intestinal CYP3A4 is needed to capture oral CYP3A substrate exposure, providing practical guidance for dosing and drug-drug interaction prediction in oncology.