Mechanistic Representation of Clusterin, a Damage Biomarker for Early Detection of Drug-Induced AKI

Background: Biomarkers have the potential to address several challenges in acute kidney injury (AKI). Novel biomarkers such as clusterin have emerged as promising candidates to address limitations of conventional biomarkers in early detection of acute kidney injury (AKI). To fully leverage the clinical potential of these biomarkers, a mechanistic understanding of the biochemical processes that lead to biomarker release is essential. Methods: We developed a mechanistic model of clusterin release within the framework of RENAsym, a QST model of drug-induced AKI that incorporates key cellular injury mechanisms and renal hemodynamics. After tubular injury, clusterin starts to upregulate on dedifferentiated tubular epithelial cells and appear in the urine. The clusterin model in RENAsym was used to predict urinary clusterin following cisplatin administration to rats in connection with proximal tubular cell necrosis and regeneration. Results: The clusterin submodel parameters were calibrated using urinary clusterin data following cisplatin administration into rats: 3 mg/kg and 6 mg/kg single dose, and 1 mg/kg daily dose for two weeks. Clusterin release was modeled to be linked with cellular necrosis to capture timing of observed peak. Simulated urinary clusterin peaked on day 5 matching the peak of necrotic flux and remained elevated for a few more days, in accordance with preclinical data. Representing clusterin during regeneration to replicate the observed delayed clusterin resolution was required to accurately represent clusterin dynamics. Conclusion: A mechanistic submodel of clusterin was developed in RENAsym that captures the kinetics of urinary clusterin in rats dosed with cisplatin. This modeling effort informed us that signals from necrotic tubular cells predicted the peak timing of urinary clusterin, and clusterin release during regeneration was required to capture clusterin delayed resolution observed in cisplatin-treated rats. This effort demonstrates the ability of QST modeling to provide mechanistic insight into the behavior of novel kidney injury biomarkers. Funding: Private Foundation Support