Pexidartinib's Hidden Metabolites Mapped in Mice, Revealing Toxic Pathways

Pexidartinib (PEX, TURALIO®) is a tyrosine kinase inhibitor targeting CSF1R and c-KIT, approved for treating tenosynovial giant cell tumor. Despite its clinical utility, PEX carries an FDA black box warning due to reports of serious and fatal liver injury. The underlying mechanism of this hepatotoxicity remains unknown, and prior human pharmacokinetic studies using radiolabeled PEX left approximately 28% of recovered radioactivity chemically unidentified—suggesting significant gaps in understanding its metabolic fate.

To address this, researchers administered PEX orally to male C57BL/6NJ mice (80 mg/kg) and collected feces, urine, plasma, and liver samples. Using ultra-high-performance LC coupled to a high-resolution Q Exactive Orbitrap mass spectrometer, combined with untargeted metabolomics and OPLS-DA chemometrics, the team performed comprehensive metabolite profiling across all four sample types.

Thirty of 31 phase I metabolites previously characterized in human and mouse liver microsomes were detected in vivo in mice, confirming strong cross-species concordance in phase I PEX metabolism. Critically, the study identified 28 phase II metabolites not previously observed in vivo: 12 glucuronides, 6 sulfates, 1 glucose conjugate, 2 glutathione (GSH) adducts, 1 cysteinyl-glycine adduct, and 6 N-acetylcysteine (NAc) adducts—24 of which had never been reported before. The detection of GSH-PEX adducts and their downstream degradation products (including NAc adducts) provides the first direct in vivo evidence that PEX generates reactive metabolites capable of binding macromolecules, consistent with prior in vitro findings in human hepatocytes.

Feces was the dominant excretion route for unchanged PEX, with fecal PEX abundance 1,574-fold higher per unit volume than urine—mirroring human pharmacokinetic data. In plasma, hydroxylated metabolite M2 (38.7%), ketone M7 (12.5%), and aminopyridine fragment M12 (19.5%) were the dominant circulating species. In the liver, M12 (aminopyridine, 37.1%), M7 (17.3%), and dioxidized PEX M8 (11.9%) predominated, suggesting hepatic accumulation of specific oxidative metabolites. Importantly, products of unusual carbon-carbon bond cleavage reactions—previously identified in microsomes—were confirmed in vivo and their phase II conjugates characterized for the first time.

These findings have significant implications for understanding PEX hepatotoxicity. The reactive intermediates captured as GSH and NAc adducts can covalently modify proteins and DNA, potentially triggering immune-mediated or direct cytotoxic liver injury. The metabolomic approach used here detected metabolites that would be missed by radiolabel-based methods, particularly those losing the labeled carbon during C-C cleavage. This work provides a metabolic roadmap for future mechanistic studies, including use of CYP3A-knockout mouse models, to isolate which metabolites are causally linked to liver damage.