Development of novel proximal tubule in vitro models to predict drug-induced nephrotoxicity

• Main Author: Tort Piella, A.
• Other Authors: Antoine, D. ; Murray, P.
• Published: University of Liverpool 2016
• Subjects: 615.1
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.706931

Drug-induced nephrotoxicity is a dose limiting factor and a common side effect of many drugs such as antibiotics, cancer chemotherapeutics or diagnostic agents and represents 20-25% of all cases of acute kidney injury (AKI). Due to a lack of representative animal models and metabolically competent renal cell lines, only 10.5% of the drug-induced nephrotoxicity cases can be predicted in preclinical studies. Furthermore, it has been recently accepted that mitochondria represent an important target for a broad spectrum of renal toxins. However, the supraphysiological glucose concentrations used in many culture conditions favour an energetic metabolic change known as the Crabtree effect, based on using glycolysis as the main route for ATP production and hence producing cell lines highly resistant to mitochondrial impairment. Therefore, there is a need for better physiologically-based renal in vitro models that can better assess drug-induced nephrotoxicity. Conditionally immortalised proximal tubule epithelial cells (ciPTECs) had been previously postulated as a novel renal model with good transporter expression that could produce reliable nephrotoxicity data for pharmacological studies. In order to confirm the applicability of the received ciPTECs for further drug transporter and toxicity studies, the transporter expression was assessed and compared with other cell lines such as HEK 293 cells as well as isolated cells from a proximal tubule (PT) fraction. The functionality of some transporters that are crucial in drug-induced nephrotoxicity such as megalin, MRP2, MRP4 and P-gp were assessed via CMFDA accumulation. Data revealed a very close mRNA expression of ciPTECs compared to the PT fraction and better transporter functionality than HEK 293 cells. Therefore, it was confirmed that ciPTECs hold the necessary transporter functionality to be a valuable nephrotoxicity model. The Crabtree effect had been previously circumvented in other cell lines by replacing glucose for galactose in the medium for 8 weeks. As a result, cells became highly aerobically dependant and more susceptible to mitotoxins. A recently developed strategy with only a 4 h pre-incubation period in galactose conditions was implemented in ciPTECs. As hypothesised, ciPTECs cultured in galactose medium displayed a significant upregulation of mitochondrial OXPHOS compared to glucose-grown cells, which caused a higher susceptibility in front of mitotoxins (rotenone EC50ATPglu / EC50ATPgal > 3125, p < 0.001 and antimycin A EC50ATPglu / EC50ATPgal > 14925, p < 0.001). Furthermore, when exposed to adefovir, a postulated mitotoxin in the clinic, galactose-cultured ciPTECs showed a significantly lower EC50 (EC50ATPglu / EC50 ATPgal > 6.12, p = 0.0049, 3 days). Tenofovir (TFV), a lower toxic analogue of adefovir, was also submitted to the glucose vs. galactose model ATP depletion did not reach 50%. Causes of mitochondrial dysfunction are variable but they all eventually trigger a perturbation of the respiratory parameters. Hence, to further explore the molecular mechanisms behind adefovir and TFV mitochondrial toxicity, ciPTECs bioenergetics were interrogated. After a 24 h adefovir exposure, a decrease in oxygen consumption linked to ATP production was detected as well as a 40% reduction of the coupling efficiency at 100 μM. Electron transport chain (ETC) complexes were also quantified and results revealed a depletion of complex I (45%) and complex IV (20%). These are the only two complexes with an important mtDNA encoded core. Furthermore, reactive oxygen species (ROS) levels increased 2 fold after adefovir exposure (48 h). In conclusion, these data evidenced that adefovir targets either mitochondrial RNA or protein synthesis, impeding a correct protein assembling of ETC complexes and therefore hampering cell respiration. However, other factors such as ROS might also play an important role. No decrease in oxygen consumption was detected after ciPTECs were exposed to TFV for 15 days. In contrast, a dose-dependent increase in extracellular acidification rate (ECAR) was observed, which could be an initial sign of mitochondrial dysfunction. Although ciPTECs exhibit most of the necessary features to investigate nephrotoxicity such as good transporter expression, they cannot take into account different people's pharmacokinetics or transporter polymorphisms that alter drug-induced toxicity. In contrast, urine-derived progenitor cells (UPCs) represent a non-invasive and limitless source of kidney cells that could be differentiated into PT cells. Even though the PT-differentiation was not fully achieved, we successfully isolated three colonies of UPCs that reached passaged 10 with similar clonogenicity values to cancer cell lines (31%) at the exponential growth phase. Therefore, if the differentiation process could be optimised, urine-derived cells could have the potential to be used as a patient-specific nephrotoxicity model as well as being applied in personalised medicine and regenerative studies. In summary, the data presented in this thesis demonstrate that drug-induced mitochondrial dysfunction responsible for nephrotoxic events can be detected using the ciPTECs glucose vs. galactose in vitro model. Furthermore, in addition of screening for mitotoxic compounds, it has been shown that the model can also be used as a tool to investigate the molecular events preceding adefovir toxicity as well as other undisclosed mitotoxicities. Finally, the work undertaken has also demonstrated that UPCs and PT-differentiated UPCs hold promising answers to tackle and prevent drug-induced nephrotoxicity.