Pharmacokinetic profile that reduces nephrotoxicity of gentamicin in a perfused kidney-on-a-chip.

Nephrotoxicity is often underestimated because renal clearance in animals is higher compared to in humans. This paper aims to illustrate the potential to fill in such pharmacokinetic gaps between animals and humans using a microfluidic kidney model. As an initial demonstration, we compare nephrotoxicity of a drug, administered at the same total dosage, but using different pharmacokinetic regimens. Kidney epithelial cell, cultured under physiological shear stress conditions, are exposed to gentamicin using regimens that mimic the pharmacokinetics of bolus injection or continuous infusion in humans. The perfusion culture utilized is important both for controlling drug exposure and for providing cells with physiological shear stress (1.0 dyn cm(-2)). Compared to static cultures, perfusion culture improves epithelial barrier function. We tested two drug treatment regimens that give the same gentamycin dose over a 24 h period. In one regimen, we mimicked drug clearance profiles for human bolus injection by starting cell exposure at 19.2 mM of gentamicin and reducing the dosage level by half every 2 h over a 24 h period. In the other regimen, we continuously infused gentamicin (3 mM for 24 h). Although junctional protein immunoreactivity was decreased with both regimens, ZO-1 and occludin fluorescence decreased less with the bolus injection mimicking regimen. The bolus injection mimicking regimen also led to less cytotoxicity and allowed the epithelium to maintain low permeability, while continuous infusion led to an increase in cytotoxicity and permeability. These data show that gentamicin disrupts cell-cell junctions, increases membrane permeability, and decreases cell viability particularly with prolonged low-level exposure. Importantly a bolus injection mimicking regimen alleviates much of the nephrotoxicity compared to the continuous infused regimen. In addition to potential relevance to clinical gentamicin administration regimens, the results are important in demonstrating the general potential of using microfluidic cell culture models for pharmacokinetics and toxicity studies.