Functional transepithelial transport measurements to detect nephrotoxicity in vitro using the RPTEC/TERT1 cell line.

The kidney is a frequent target for organ-specific toxicity as a result of its primary function in controlling body fluids, for example, via resorption of amino acids, peptides, nutrients, ions, xenobiotics and water from the primary urine as well as excretion of metabolic waste products and hydrophilic and amphiphilic xenobiotics. Compounds exhibiting dose-limiting nephrotoxicity include drugs from highly diverse classes and chemical structures, e.g., antibiotics (gentamicin), chemotherapeutics (cisplatin), immunosuppressants (cyclosporine A and tacrolimus) or bisphosphonates (zoledronate). All of these compounds elicit nephrotoxicity primarily by injuring renal proximal tubule epithelial cells (RPTECs). However, prediction of a compound's nephrotoxic potential in humans to support early unmasking of risk-bearing drug candidates remains an unmet challenge, mainly due to the complex kidney anatomy as well as pronounced inter- and intraspecies differences and lack of relevant and validated human in vitro models. Accordingly, we used the recently established human RPTEC/TERT1 cell line to carry out toxicity studies with a focus on impairment of functional characteristics, i.e., transepithelial electrical resistance (TEER), vectorial transport of water, cations, and anions. Results were compared to real-time cytotoxicity assessments using cellular impedance (xCELLigence assay) and the routine cell viability readout (MTT). As expected, most toxins caused exposure time- and concentration-dependent cytotoxicity. However, for some compounds (cyclosporine A and tacrolimus), transport processes were strongly impaired in absence of a concomitant decrease in cell viability. In conclusion, these data demonstrate that functional parameters are important, highly sensitive and meaningful additional readouts for nephrotoxicity assessment in human renal proximal tubule epithelial cells.

PODO/TERT256 - A promising human immortalized podocyte cell line and its potential use for in vitro research at different oxygen levels.

Podocytes are of key interest for the prediction of nephrotoxicity as they are especially sensitive to toxic insults due to their central role in the glomerular filtration apparatus. However, currently, prediction of nephrotoxicity in humans remains insufficiently reliable, thus highlighting the need for advanced in vitro model systems using human cells with improved prediction capacity. Recent approaches for refining in vitro model systems focus on closely replicating physiological conditions as observed under the in vivo situation typical of the respective nephron section of interest. PODO/TERT256, a human immortalized podocyte cell line, were employed in a semi-static transwell system to evaluate its potential use as a human podocyte in vitro system for modelling potential human glomerular toxicity. Furthermore, the impact of routinely employed excessive oxygen tension (21 % - AtmOx), when compared to the physiological oxygen tensions (10 % - PhysOx) observed in vivo, was analyzed. Generally, cultured PODO/TERT256 formed a stable, contact-inhibited monolayer with typical podocyte morphology (large cell body, apical microvilli, finger-like cytoplasmic projections (reminiscent of foot processes), and interdigitating cell-cell junctions) and developed a size-selective filtration barrier. PhysOx, however, induced a more pronounced in vivo like phenotype, comprised of significantly larger cell bodies, significantly enhanced filtration barrier size-selectivity, and a remarkable re-localization of nephrin to the cell membrane, thus suggesting an improved in vitro replication of in vivo characteristics. Preliminary toxicity characterization with the known glomerulotoxin doxorubicin (DOX) suggested an increasing change in filtration permeability, already at the lowest DOX concentrations tested (0.01 µM) under PhysOx, whereas obvious changes under AtmOx were observed as of 0.16 µM and higher with a near all or nothing effect. The latter findings suggested that PODO/TERT256 could serve as an in vitro human podocyte model for studying glomerulotoxicity, whereby culturing at PhyOx tension appeared critical for an improved in vivo-like phenotype and functionality. Moreover, PODO/TERT256 could be incorporated into advanced human glomerulus systems in vitro, recapitulating microfluidic conditions and multiple cell types (endothelial and mesenchymal cells) that can even better predict human glomerular toxicity.

Physiological oxygen and co-culture with human fibroblasts facilitate in vivo-like properties in human renal proximal tubular epithelial cells.

Reliable prediction of compound mediated nephrotoxicity in humans is still unsatisfactory irrespective of the recent advancements in in silico, in vitro and in vivo models. Therefore, current in vitro approaches need refinement to better match the human in vivo situation, specifically with regard to the potential influence of other cell types (e.g. fibroblasts) and to the potential biases introduced by the excessive 21% O2 (AtmOx) as employed in routine cell culturing. We used a transwell co-culture model combining human renal proximal tubule epithelial cells (RPTEC/TERT1) and human fibroblasts (fHDF/TERT166) to compare the functional properties and expression of selected marker proteins at 21% O2 and at the physiologically normal 10% O2 tension (PhysOx) commensurate with in vivo conditions. Culturing at PhysOx and co-culturing with fibroblasts significantly improved epithelial barrier integrity, expression of transporters (e.g. aquaporin 2; OCT-MATE; MRP-OAT) and metabolism. Moreover, beyond culturing these human cells in co-culture for up to 41 days, we were able to demonstrate increased functionality of cation transport, as shown via ASP+ (OCT-MATE axis), and anion transport, as shown via LY (MRP-OAT axis). Thus, adjusting the in vitro system to near physiological conditions had a major impact on functionality and provides the basis for the future development of true flow-through microfluidic renal testing systems with better predictability of human renal proximal toxicity.

RPTEC/TERT1 cells form highly differentiated tubules when cultured in a 3D matrix.

The proximal tubule is the primary site for renal solute reabsorption and secretion and thus a main target for drug-induced toxicity. Current nonclinical methods using 2D cell cultures are unable to fully recapitulate clinical drug responses mainly due to limited in vitro functional lifespan. Since extracellular matrices are known to be key regulators of cell development, culturing cells on classic 2D plastic surfaces inevitably results in loss of differentiation. Hence, 3D models of the human proximal tubule that recapitulate the in vivo morphology would allow for improved drug screening and disease modeling. Here, the development and characterization of a 3D proximal tubule model using RPTEC/TERT1 cells is presented. RPTEC/TERT1 cells self-assembled in matrigel to form highly differentiated and stable 3D tubular structures characterized by a branched network of monolayered cells encircling a cell-free lumen thus mimicking the proximal tubule. In vitro tubuli resembled the polarity of a proximal tubule epithelium as indicated by polar expression of Na+/K+- ATPase and ZO-3. Furthermore, 3D cultured RPTEC/TERT1 cells showed overall increased mRNA expression of xenobiotic transporters e.g. OCTs and MATEs and de novo expression of OAT3 when compared to cultures on plastics or membrane inserts. Finally, this model was used to assess delayed cisplatin-induced nephrotoxicity and demonstrated increased sensitivity when compared to 2D culture. Thus, the easy-to-use model described here may prove to be useful for mechanistic investigations, e.g. in discovery of compounds interfering with tubule formation, differentiation and polarization, as well for the detection and understanding of pharmaceutical induced nephrotoxicity.

Canagliflozin mediated dual inhibition of mitochondrial glutamate dehydrogenase and complex I: an off-target adverse effect.

Recent FDA Drug Safety Communications report an increased risk for acute kidney injury in patients treated with the gliflozin class of sodium/glucose co-transport inhibitors indicated for treatment of type 2 diabetes mellitus. To identify a potential rationale for the latter, we used an in vitro human renal proximal tubule epithelial cell model system (RPTEC/TERT1), physiologically representing human renal proximal tubule function. A targeted metabolomics approach, contrasting gliflozins to inhibitors of central carbon metabolism and mitochondrial function, revealed a double mode of action for canagliflozin, but not for its analogs dapagliflozin and empagliflozin. Canagliflozin inhibited the glutamate dehydrogenase (GDH) and mitochondrial electron transport chain (ETC) complex I at clinically relevant concentrations. This dual inhibition specifically prevented replenishment of tricarboxylic acid cycle metabolites by glutamine (anaplerosis) and thus altered amino acid pools by increasing compensatory transamination reactions. Consequently, canagliflozin caused a characteristic intracellular accumulation of glutamine, glutamate and alanine in confluent, quiescent RPTEC/TERT1. Canagliflozin, but none of the classical ETC inhibitors, induced cytotoxicity at particularly low concentrations in proliferating RPTEC/TERT1, serving as model for proximal tubule regeneration in situ. This finding is testimony of the strong dependence of proliferating cells on glutamine anaplerosis via GDH. Our discovery of canagliflozin-mediated simultaneous inhibition of GDH and ETC complex I in renal cells at clinically relevant concentrations, and their particular susceptibility to necrotic cell death during proliferation, provides a mechanistic rationale for the adverse effects observed especially in patients with preexisting chronic kidney disease or previous kidney injury characterized by sustained regenerative tubular epithelial cell proliferation.