Investigation of ifosfamide and chloroacetaldehyde renal toxicity through integration of in vitro liver-kidney microfluidic data and pharmacokinetic-system biology models.

We have integrated in vitro and in silico data to describe the toxicity of chloroacetaldehyde (CAA) on renal cells via its production from the metabolism of ifosfamide (IFO) by hepatic cells. A pharmacokinetic (PK) model described the production of CAA by the hepatocytes and its transport to the renal cells. A system biology model was coupled to the PK model to describe the production of reactive oxygen species (ROS) induced by CAA in the renal cells. In response to the ROS production, the metabolism of glutathione (GSH) and its depletion were modeled by the action of an NFE2L2 gene-dependent pathway. The model parameters were estimated in a Bayesian context via Markov Chain Monte Carlo (MCMC) simulations based on microfluidic experiments and literature in vitro data. Hepatic IFO and CAA in vitro intrinsic clearances were estimated to be 1.85 x 10(-9) µL s(-1) cell(-1) and 0.185 x 10(-9) µL s(-1) cell(-1) ,respectively (corresponding to an in vivo intrinsic IFO clearance estimate of 1.23 l h(-1) , to be compared to IFO published values ranging from 3 to 10 l h(-1) ). After model calibration, simulations made at therapeutic doses of IFO showed CAA renal intracellular concentrations ranging from 11 to 131 µM. Intracellular CAA concentrations above 70 µM induced intense ROS production and GSH depletion. Those responses were time and dose dependent, showing transient and non-linear kinetics. Those results are in agreement with literature data reporting that intracellular CAA toxic concentrations range from 35 to 320 µM, after therapeutic ifosfamide dosing. The results were also consistent with in vitro CAA renal cytotoxicity data.

Investigation of acetaminophen toxicity in HepG2/C3a microscale cultures using a system biology model of glutathione depletion.

We have integrated in vitro and in silico information to investigate acetaminophen (APAP) and its metabolite N-acetyl-p-benzoquinone imine (NAPQI) toxicity in liver biochip. In previous works, we observed higher cytotoxicity of HepG2/C3a cultivated in biochips when exposed to 1 mM of APAP for 72 h as compared to Petri cultures. We complete our investigation with the present in silico approach to extend the mechanistic interpretation of the intracellular kinetics of the toxicity process. For that purpose, we propose a mathematical model based on the coupling of a drug pharmacokinetic model (PK) with a systemic biology model (SB) describing the reactive oxygen species (ROS) production by NAPQI and the subsequent glutathione (GSH) depletion. The SB model was parameterized using (i) transcriptomic data, (ii) qualitative results of time lapses ROS fluorescent curves for both control and 1-mM APAP-treated experiments, and (iii) additional GSH literature data. The PK model was parameterized (i) using the in vitro kinetic data (at 160 µM, 1 mM, 10 mM), (ii) using the parameters resulting from a physiologically based pharmacokinetic (PBPK) literature model for APAP, and (iii) by literature data describing NAPQI formation. The PK-SB model predicted a ROS increase and GSH depletion due to the NAPQI formation. The transition from a detoxification phase and NAPQI and ROS accumulation was predicted for a NAPQI concentration ranging between 0.025 and 0.25 µM in the cytosol. In parallel, we performed a dose response analysis in biochips that shows a reduction of the final hepatic cell number appeared in agreement with the time and doses associated with the switch of the NAPQI detoxification/accumulation. As a result, we were able to correlate in vitro extracellular APAP exposures with an intracellular in silico ROS accumulation using an integration of a coupled mathematical and experimental liver on chip approach.

Investigation of omeprazole and phenacetin first-pass metabolism in humans using a microscale bioreactor and pharmacokinetic models.

A new in vitro microfluidic platform (integrated insert dynamic microfluidic platform, IIDMP) allowing the co-culture of intestinal Caco-2 TC7 cells and of human primary hepatocytes was used to test the absorption and first-pass metabolism of two drugs: phenacetin and omeprazole. The metabolism of these drugs by CYP1A2, CYP2C19 and CYP3A4 was evaluated by the calculation of bioavailabilities and of intrinsic clearances using a pharmacokinetic (PK) model. To demonstrate the usefulness of the device and of the PK model, predictions were compared with in vitro and in vivo results from the literature. Based on the IIDMP experiments, hepatic in vivo clearances of phenacetin and omeprazole in the IIDMP were predicted to be 3.10 ± 0.36 and 1.46 ± 0.25 ml/min/kg body weight, respectively. This appeared lower than the in vivo observed data with values ranging between 11.9-19.6 and 5.8-7.5 ml/min/kg body weight, respectively. Then the calculated hepatic and intestinal clearances led to predicting an oral bioavailability of 0.85 and 0.77 for phenacetin and omeprazole versus 0.92 and 0.78 using separate data from the simple monoculture of Caco-2 TC7 cells and hepatocytes in Petri dishes. When compared with the in vivo data, the results of oral bioavailability were overestimated (0.37 and 0.71, respectively). The feasibility of co-culture in a device allowing the integration of intestinal absorption, intestinal metabolism and hepatic metabolism in a single model was demonstrated. Nevertheless, further experiments with other drugs are needed to extend knowledge of the device to predict oral bioavailability and intestinal first-pass metabolism.

Quantitative in vitro to in vivo extrapolation of tissues toxicity.

Predicting repeated-dosing in vivo drug toxicity from in vitro testing and omics data gathering requires significant support in bioinformatics, mathematical modeling and statistics. We present here the major aspects of the work devoted within the framework of the European integrated Predict-IV to pharmacokinetic modeling of in vitro experiments, physiologically based pharmacokinetic (PBPK) modeling, mechanistic models of toxicity for the kidney and brain, large scale dose-response analyses methods and biomarker discovery tools. All of those methods have been applied to various extent to the drug datasets developed by the project's partners. Our approach is rather generic and could be adapted to other drugs or drug candidates. It marks a successful integration of the work of the different teams toward a common goal of predictive quantitative in vitro to in vivo extrapolation.

Systems biology modeling of omics data: effect of cyclosporine a on the Nrf2 pathway in human renal cells.

BACKGROUND: Incorporation of omic data streams for building improved systems biology models has great potential for improving their predictions of biological outcomes. We have recently shown that cyclosporine A (CsA) strongly activates the nuclear factor (erythroid-derived 2)-like 2 pathway (Nrf2) in renal proximal tubular epithelial cells (RPTECs) exposed in vitro. We present here a quantitative calibration of a differential equation model of the Nrf2 pathway with a subset of the omics data we collected. RESULTS: In vitro pharmacokinetic data on CsA exchange between cells, culture medium and vial walls, and data on the time course of omics markers in response to CsA exposure were reasonably well fitted with a coupled PK-systems biology model. Posterior statistical distributions of the model parameter values were obtained by Markov chain Monte Carlo sampling in a Bayesian framework. A complex cyclic pattern of ROS production and control emerged at 5 µM CsA repeated exposure. Plateau responses were found at 15 µM exposures. Shortly above those exposure levels, the model predicts a disproportionate increase in cellular ROS quantity which is consistent with an in vitro EC50 of about 40 µM for CsA in RPTECs. CONCLUSIONS: The model proposed can be used to analyze and predict cellular response to oxidative stress, provided sufficient data to set its parameters to cell-specific values. Omics data can be used to that effect in a Bayesian statistical framework which retains prior information about the likely parameter values.

Integration of pharmacokinetic and NRF2 system biology models to describe reactive oxygen species production and subsequent glutathione depletion in liver microfluidic biochips after flutamide exposure.

We present a systems biology analysis of rat primary hepatocytes response after exposure to 10 µM and 100 µM flutamide in liver microfluidic biochips. We coupled an in vitro pharmacokinetic (PK) model of flutamide to a system biology model of its reactive oxygen species (ROS) production and scavenging by the Nrf2 regulated glutathione production. The PK model was calibrated using data on flutamide kinetics, hydroxyflutamide and glutathione conjugates formation in microfluidic conditions. The parameters of Nrf2-related gene activities and the subsequent glutathione depletion were calibrated using microarray data from our microfluidic experiments and literature information. Following a 10 µM flutamide exposure, the model predicted a recovery time to baseline levels of glutathione (GSH) and ROS in agreement with our experimental observations. At 100 µM, the model predicted that metabolism saturation led to an important accumulation of flutamide in cells, a high ROS production and complete GSH depletion. The high levels of ROS predicted were consistent with the necrotic switch observed by transcriptomics, and the high cell mortality we had experimentally observed. The model predicted a transition between recoverable GSH depletion and deep GSH depletion at about 12.5 µM of flutamide (single perfusion exposure). Our work shows that in vitro biochip experiments can provide supporting information for complex in silico modeling including data from extra cellular and intra cellular levels. We believe that this approach can be an efficient strategy for a global integrated methodology in predictive toxicology.

Application of integrated transcriptomic, proteomic and metabolomic profiling for the delineation of mechanisms of drug induced cell stress.

High content omic techniques in combination with stable human in vitro cell culture systems have the potential to improve on current pre-clinical safety regimes by providing detailed mechanistic information of altered cellular processes. Here we investigated the added benefit of integrating transcriptomics, proteomics and metabolomics together with pharmacokinetics for drug testing regimes. Cultured human renal epithelial cells (RPTEC/TERT1) were exposed to the nephrotoxin Cyclosporine A (CsA) at therapeutic and supratherapeutic concentrations for 14days. CsA was quantified in supernatants and cellular lysates by LC-MS/MS for kinetic modeling. There was a rapid cellular uptake and accumulation of CsA, with a non-linear relationship between intracellular and applied concentrations. CsA at 15µM induced mitochondrial disturbances and activation of the Nrf2-oxidative-damage and the unfolded protein-response pathways. All three omic streams provided complementary information, especially pertaining to Nrf2 and ATF4 activation. No stress induction was detected with 5µM CsA; however, both concentrations resulted in a maximal secretion of cyclophilin B. The study demonstrates for the first time that CsA-induced stress is not directly linked to its primary pharmacology. In addition we demonstrate the power of integrated omics for the elucidation of signaling cascades brought about by compound induced cell stress.