Development of a Microphysiological Model of the Kidney Proximal Tubule and Application in Predictive Toxicity Testing of Polymyxin Antibiotics

Abstract

Late-stage attrition of investigational new drugs can be extremely devastating to both the drug developer as well as the patient in regard to money lost and disease left untreated, respectively. A primary reason for a compound’s development to be suspended, if not terminated, can be attributed to the preclinical underestimation of unforeseen toxic events surfacing in human subjects during clinical trials. Existing preclinical models that lack accuracy for predicting toxicity demonstrate the dire need for improving and developing more robust physiologically relevant models. Our goal was to develop an ex vivo 3D microphysiological system (MPS) capable of recapitulating, both structurally and functionally, the microenvironment of the kidney proximal tubule. Using the MPS, we have demonstrated native functions of the human proximal tubule including expression of appropriate marker proteins, biochemical and synthetic activities, as well as secretory and reabsorptive processes. Through concentrative facilitated transport, via secretory and reabsorptive processes, the renal proximal tubule becomes a primary site of drug accumulation and thus drug-induced injury. Our goal was to show that the MPS can now be an accurate preclinical model of nephrotoxicity. Using the MPS, drug-induced nephrotoxicity and sensitivity to prototypical nephrotoxins was assessed using the classic and developing polymyxin antibiotics, Polymyxin B (PMB) and Northern Antibiotics-739/741 (NAB739/NAB741). Polymyxins are highly effective for treating gram-negative bacterial infections yet there remains a high frequency of nephrotoxicity (~30%) which has limited their applications in the clinic and deemed them a last-line therapeutic. Because current mechanisms of polymyxin-induced nephrotoxicity are unclear, we have used the MPS, coupled with global transcriptomics, to elucidate the underlying mechanisms. Furthermore, we have shown that polymyxin-derivatives (NAB739/741), containing a reduced polycationic nature compared to PMB, have markedly reduced indications of injury relative to polymyxin exposure. The experimental approach used was to model and establish a safety profile of polymyxin-induced nephrotoxicity, for both PMB and the NAB compounds, using the MPS via a multiplex assessment of cell-associated biomarker induction, urinary biomarkers induction, and the transcriptional response to exposure. Results of these studies will advance the development of innovative antibiotics and validate the utility of the 3D MPS as a platform for preclinical safety assessment of new molecular entities.