Freshly isolated primary human proximal tubule cells as an in vitro model for the detection of renal tubular toxicity.

ABSTRACT

Drug induced kidney injury (DIKI) is a common reason for compound attrition in drug development pipelines with proximal tubule epithelial cells (PTECs) most commonly associated with DIKI. Here, we investigated freshly isolated human (hPTECs) as an in vitro model for assessing renal tubular toxicity. The freshly isolated hPTECs were first characterized to confirm gene expression of important renal transporters involved in drug handling which was further corroborated by confirming the functional activity of organic cation transporter 2 and organic anion transporter 1 by using transporter specific inhibitors. Additionally, functionality of megalin/cubilin endocytic receptors was also confirmed. A training set of 36 compounds was used to test the ability of the model to classify them using six different endpoints which included three biomarkers (Kidney Injury Molecule-1, Neutrophil gelatinase-associated lipocalin, and Clusterin) and three non-specific injury endpoints (ATP depletion, LDH leakage, and barrier permeability via transepithelial electrical resistance) in a dose-dependent manner across two independent kidney donors. In general, biomarkers showed higher predictivity than non-specific endpoints, with Clusterin showing the highest predictivity (Sensitivity/Specificity - 65.0/93.8 %). By using the thresholds generated from the training set, nine candidate internal Takeda compounds were screened where PTEC toxicity was identified as one of the findings in preclinical animal studies. The model correctly classified four of six true positives and two of three true negatives, showing validation of the in vitro model for detection of tubular toxicants. This work thus shows the potential application of freshly isolated primary hPTECs using translational biomarkers in assessment of tubular toxicity within the drug discovery pipeline.