Renal toxicity remains a major issue in clinical trials, and stresses the need for more predictive models fit for implementation in early drug development1. Here, we describe the use of a high throughput, microfluidic platform for the detection of drug-induced nephrotoxicity. A microfluidic platform (Mimetas’ OrganoPlate®)2 was combined with renal proximal tubule epithelial cell lines (PTEC) and exposed to fluid shear stress. A 12-compound nephrotoxicity screen across multiple laboratories was performed in collaboration with sponsors and the NC3Rs3. ciPTEC-OAT1 or RPTECs (Sigma) seeded against an ECM gel under perfusion flow were used to establish a proximal tubule-on-a-chip. Tubules with polarized epithelium containing functional transporter expression were obtained. Drug-induced toxicity was assessed by exposing the tubules to 4 benchmark compounds with known clinical effect and 8 blinded compounds supplied by the sponsors for 24 and 48h. Epithelial barrier tightness and drug-transporter interactions were evaluated. Parallel to this, cellular damage and stress were assessed using various read-outs. Finally, gene expression analysis was performed to assess AKI markers. The Nephroscreen revealed that a combination of cell viability, LDH and miRNA release were the most predictive readouts in determining nephrotoxicity. Most of the blinded compounds resulted in toxicity detected by at least one of the functional read-outs. Nephroscreen provides a reliable standardized and automatable system for efficacious identifying nephrotoxicants and revealing their mode of action.