Two hiPSC-derived proximal tubule models show how transporter localization relates to directional drug transport and inhibitor responses. Transport activity and longer-term cell viability do not necessarily favor the same model.
The proximal tubule contributes to drug elimination and is also a site of drug-induced injury. A team led by Kyoto University reported a proximal tubule-on-chip study in iScience, using cells isolated from kidney organoids produced through two human induced pluripotent stem cell (hiPSC) differentiation approaches. Both were compared with a chip containing immortalized RPTEC/TERT1 proximal tubule cells.
The organoid-derived preparations contained Lotus tetragonolobus lectin (LTL)-positive cells and were labeled #1-LTL+ and #2-LTL+. They formed epithelial layers on porous membranes, with separately accessible basal and apical compartments. This arrangement allows measurement of transport from the blood-facing side toward the tubular lumen and in the reverse direction.
Both preparations expressed drug transporters including organic anion transporters OAT1 and OAT3 and organic cation transporter OCT2. OAT1/3 localization was more clearly basolateral in #1-LTL+, whereas #2-LTL+ showed a more dispersed distribution. The immortalized comparator had weaker OAT expression and polarization. The study paired these localization findings with functional measurements of drug movement.
Adefovir and para-aminohippurate were used as OAT1-associated substrates, with rosuvastatin as an OAT3-associated substrate. The researchers measured transport in both directions across each cell layer. The ratio of basal-to-apical transport to apical-to-basal transport exceeded two for all three substrates in both organoid-derived models, although absolute secretory transport was generally greater in #1-LTL+.
At the lower tested probenecid concentration, transport of adefovir and para-aminohippurate toward the apical compartment decreased, while rosuvastatin transport changed relatively little. A higher probenecid concentration inhibited all three substrates. The efflux inhibitor MK571 reduced transcellular transport of adefovir and para-aminohippurate but had no significant effect on rosuvastatin in the reported comparison.
The immortalized model showed smaller changes in bidirectional transport and inhibitor tests. The two organoid-derived models could have similar transport ratios despite substantially different absolute permeability. Transporter localization, barrier properties and the size of the transport response therefore all contributed to the comparison.
The team extended the analysis to metformin and quinidine, probing transport associated with OCT2/MATE and P-glycoprotein. The #1-LTL+ model had the highest basal-to-apical transport for both drugs. Performance of #2-LTL+ relative to the immortalized model varied by substrate, alongside differences in transporter expression and polarization.
Cisplatin was introduced from the basal side, and injury was assessed by lactate dehydrogenase (LDH) release and live/dead measurements. Both organoid-derived models showed injury. Cimetidine, an OCT2 inhibitor, reduced LDH release in #1-LTL+ at 24 hours; a clearer effect in #2-LTL+ emerged at 48 hours.
At 48 hours, #1-LTL+ LDH values no longer differed between conditions. Live/dead observations showed that many cells had already lost viability after 24 hours of cisplatin exposure.
Aristolochic acid was also tested as a nephrotoxic reference, adding an OAT-associated injury comparison to the transport and cisplatin measurements.
The work combines transporter localization, bidirectional permeability, inhibitor responses and injury in the same human-derived systems. It demonstrates why transporter expression alone is insufficient: the protein must be appropriately localized, and the epithelial layer must support the expected movement of substrates.
The two models had different trade-offs. The #1-LTL+ preparation showed stronger transport but also higher passive permeability and baseline LDH release. The #2-LTL+ preparation formed a tighter barrier and was better suited to longer observation periods. Each differentiation approach used a different starting hiPSC line, so the study did not isolate protocol effects from cell-line effects.
The authors also identify incomplete maturation and drug sorption by polydimethylsiloxane (PDMS) as areas for improvement. In particular, the high cimetidine concentration warrants further study with lower-sorption materials. Broader use in drug development would require testing across additional compounds, cell lines and exposure conditions.
A four-platform comparison combined lipid imaging, liver-function measurements, expression profiling and CYP assays. Lipid accumulation did not uniformly reduce drug metabolism, and differences depended on the model and donor.
Dendritic-cell/T-cell co-cultures and drug-specific clones revealed separate responses to mitiperstat and four metabolites. Most clone activation was consistent with direct pharmacological interaction.
