Most newly developed drugs have nephrotoxic potential, and predicting nephrotoxicity during preclinical drug development is essential [19]. Several cell types have been used for in vitro nephrotoxicity assays, and each cell type has advantages and disadvantages (Table 1) [19,20]. Madin-Darby canine kidney cells, which are a well-established cell line, form a polarized tight monolayer, but show differences from those in humans and are poor predictors of nephrotoxicity [20].

Proximal tubular (PT) cells have numerous active transporters to reabsorb filtered drug metabolites, resulting in a high intracellular accumulation of drug metabolites [19-22]. Thus, nephrotoxic drugs are likely to affect PT cells [20]. For this reason, PT cells have been widely used to develop in vitro platforms for drug screening or nephrotoxicity testing [20]. Proximal-like porcine kidney 1 (LLC-PK1) cells (porcine PT cell line), human kidney 2 (HK-2) cells (immortalized human PT cell line), and primary renal PT epithelial cells from different donors are widely used as cell sources for PT cells. In vitro platforms using PT cells have been reported; their sensitivity and accuracy are within the range of 70% to 90% depending on the PT cell type [19,23,24]. However, HK-2 cells have few transporters or proximal tubule characteristics [25,26] and show a tendency to lose their transepithelial transport function once they undergo the epithelial-to-mesenchymal transition [27]. Primary human renal PT epithelial cells have transepithelial transporters and express metabolic enzymes, and are also available for a broad range of biomarker assays. This cell line is suitable for nephrotoxicity testing and show high predictivity [20,28,29]. However, these cells rapidly lose their PT characteristics during prolonged culture, and there is donor-to-donor variability and possible contamination with other renal cell types when the cells are isolated from donors due to heterogeneity of the kidney [20,30].

Differentiating hPSCs into PT epithelial cell-like cells is an alternative and attractive approach to establish an in vitro model. One research group reported an accuracy of 76% for an in vitro model based on embryonic stem cell-derived human renal PT-like cells [31]. However, embryonic stem cell-derived human renal PT-like cells have not been sufficiently characterized [32] and their use is not free from ethical and legal issues.

Thus, human inducible pluripotent stem cell (iPSC)-based models are alternative and attractive in vitro models. A protocol to differentiate human iPSCs into PT-like cells and establish a human iPSC-based model for predicting nephrotoxicity has been published [33]. Using this protocol, PT-like cells with a purity > 90% were obtained from iPSCs. This human iPSC-based model was used to predict the toxicity of 30 compounds to PT cells; the training accuracy of the model was 99.8%, and the test accuracy was 87.0% [33]. Although the model can predict nephrotoxicity with reasonable accurately, it does not mimic the diversity of cells, ECM, or vascularization associated with actual human kidneys. Thus, a human iPSC-based model is required to extrapolate in vitro evidence to the in vivo environment [20].

Cell-to-cell and cell-to-ECM interactions control cell phenotypes and functions in vivo [34-36]. Maintaining these interactions is required to maintain tissue-like responses in an in vitro model in the context of drug efficacy and toxicity screening [37]. Two-dimensional cultures on culture plastic using a single renal cell type may have difficulty in recapitulating these complex conditions. In contrast to conventional kidney cell lines or iPSC-derived PT cells, individual kidney organoids have all of the major components of developing proximal nephrons, namely tubular cells, endothelial cells, nephron progenitors, and podocyte-like cells [1,4,9], and are therefore promising systems for drug screening and nephrotoxicity testing.

Some research groups have tested the nephrotoxicity of gentamycin or cisplatin on kidney organoids derived from hPSCs [1,4,9]. After kidney organoids were treated with gentamycin or cisplatin, increased expression of kidney injury molecule-1, a clinical biomarker of acute kidney injury (AKI) [1,9], or cleaved caspase 3, a marker of apoptosis [4], was observed by immunofluorescence staining or real-time quantitative polymerase chain reaction. These data support the use of kidney organoids as a platform for nephrotoxicity testing. However, more nephrotoxic drugs and non-nephrotoxic compounds have to be evaluated before kidney organoids can be used as a drug testing platform, and the results must be compared with tubular cell-based models or hPSC-based models.

Kidney organoids have multi-lineage nephrogenic progenitor cells, and immunosuppression is not needed to transplant kidney organoids derived from patients. Thus, transplanting kidney organoids is an attractive regenerative therapy. However, this may not be achievable in a short time frame, for several reasons. The evidence of functionality of kidney organoids remains insufficient. It is not easy to generate the urine collecting system and connect it with the host urinary system. The safety of iPSCs and the immune barrier in ESCs also limit transplantation of kidney organoids derived from hPSCs [10].

Nevertheless, some researchers have reported encouraging data [2,54-57]. Taguchi et al. [2] generated metanephric nephron progenitor from mouse ESCs and transplanted them beneath the kidney capsule of immunodeficient mice. Tubulogenesis and formation of glomerular-like structures were observed in the transplanted graft. The transplanted glomeruli vascularized with the host circulation and contained red blood cells.

Another study also reported encouraging data. Sharmin et al. [55] generated human iPSC cell lines expressing green fluorescent protein in the NPHS1 locus using transcription activator-like effector nuclease-mediated homologous recombination, and then induced 3D primordial glomeruli from these iPSCs. They reported that the induced human podocytes derived from iPSCs exhibited apicobasal polarity, by showing the location of an in-basal domain with primary foot process-like structures that were connected with slit diaphragm-like structures. They developed a transplantation technique to effectively form glomeruli from human iPSC-derived nephron progenitors using spacers that released tension from the host kidney capsules. Using this technique, induced glomeruli were transplanted from human iPSCs into the mouse kidney subcapsular space, followed by treatment with vascular endothelial growth factor (VEGF) [55]. The transplanted human glomeruli were vascularized with the host mouse endothelial cells, and the iPSC-derived podocytes had numerous cell processes that accumulated around the fenestrated endothelial cells, forming slit diaphragm-like structures.

Another study reported functional vascularization with progressive morphogenesis of human kidney organoids after renal subscapular transplantation. van den Berg et al. [57] transplanted kidney organoids derived from hPSCs into mouse kidney subcapsular spaces and reported that the transplanted hPSC-derived kidney organoids develop host-derived vascularization in the absence of any exogenous VEGF. They also showed functional glomerular perfusion with connection to host mouse vascular networks by in vivo imaging of the transplanted kidney organoids under the mouse kidney capsule. They revealed that transplantation induced the formation of a slit diaphragm structure consisting of the glomerular basement membrane, fenestrated endothelial cells, and podocyte foot processes in an ultrastructural analysis. They also observed specialized (i.e., more polarized and segmented) tubular epithelium in transplanted kidney organoids compared with non-transplanted organoids. Our laboratory also transplanted kidney organoids derived from hPSCs beneath the renal capsular space of nonobese diabetic/severe combined immunodeficiency (NOD-SCID) mice. The transplanted kidney organoids vascularized with the host mouse and developed nephron-like structures (Fig. 2).

The therapeutic effectiveness of transplanting kidney organoids for kidney diseases has not yet been described but can be predicted based on previous reports [54,56]. Toyohara et al. [56] induced hPSCs into OSR1⁺SIX2⁺ renal progenitors, which were capable of reconstituting 3D proximal renal tubule-like structures in vitro and in vivo. They transplanted these human iPSC-derived renal progenitors cells into the kidney subcapsular space of mice in an ischemia/reperfusion injury AKI model, which resulted in suppression of the elevation of blood urea nitrogen and serum creatinine levels and attenuation of histopathological changes, such as tubular necrosis, tubule dilatation with casts, and interstitial fibrosis. Their data indicate that the renoprotective benefits originated from the trophic effects produced by the transplanted renal progenitor cells.

Another study was performed using a cisplatin injury-induced AKI mouse model [54]. Imberti et al. [54] induced renal progenitors from human iPSCs and infused them intravenously into mice with cisplatin-induced AKI. Injecting human iPSCs derived from renal progenitor cells restored renal function and structure in mice with cisplatin-induced AKI. Interestingly, different from Toyohara et al. [56], these authors reported robust engraftment of human iPSCs derived from renal progenitor cells into damaged tubules of the host mouse kidney. Although it remains unclear whether renal recovery from AKI is due to the integrated human iPSCs derived from renal progenitor cells, or to regenerated tubular cells via paracrine effects arising from infused iPSC-derived renal progenitor cells, these data provide the basis for future applications of kidney organoids derived from hPSCs in regenerative therapy.

Transplantation of a regenerated kidney using kidney scaffolds may be an attractive therapeutic option for patients with ESRD. Recellularization of decellularized kidney scaffolds with renal progenitor cells and endothelial cells is important for generating bioengineered kidneys for clinical use in the future, considering the successful clinical transplantation of recellularized tracheal scaffolds from decellularized cadaveric trachea [58,59].

One group performed experimental orthotopic transplantation of re-populating decellularized kidneys [60]. They seeded decellularized rat kidney scaffolds with rat neonatal kidney cells via the ureter and human umbilical vascular endothelial cells via the renal artery, and then perfused it in a whole-organ bioreactor. The decellularized rat kidney scaffolds were successfully recellularized. The repopulated kidneys produced rudimentary urine in vitro. Furthermore, when these bioengineered kidneys were transplanted in an orthotopic position in rats, the grafts were perfused by the recipient’s circulation and produced urine through the ureteral conduit in vivo [60].

Achieving an appropriate combination of cell types is one of the main issues in recellularization of decellularized kidney scaffolds. Thus, various combinations of cell types, such as renal cells with endothelial cells, including mouse ESCs [61,62], human iPSC-derived endothelial cells with human renal cortical tubular epithelial cells [63], rat aorta endothelial cells with rat epithelial tubular cells [64], and human umbilical vascular endothelial cells with rat neonatal kidney cells [60] have been explored.

Recellularization with renal progenitor cells and endothelial cells derived from human iPSCs results in abundant repopulation of decellularized kidney scaffolds and formation of a vascular network [65]. Du et al. [65] generated PAX2⁺ renal progenitor cells and endothelial cells from human iPSCs, which repopulated decellularized kidney scaffolds. Interestingly, the expression levels of genes related to renal development were higher in the recellularized group that had PAX2⁺ renal progenitor cells and endothelial cells than in the group that only had PAX2⁺ renal progenitor cells. When the recellularized kidney scaffolds were implanted in immunodeficient mice, glomerular assembly and function were achieved by recellularization with renal progenitor cells and endothelial cells. That study provided evidence for the vital role of endothelial cells in the generation of a functional whole kidney from a kidney scaffold, and highlighted the usefulness of iPSC-derived nephrogenic progenitors.

Numerous obstacles must be overcome to clinically apply bioengineered kidneys: optimization of cell-seeding protocols to create human-sized scaffolds, upscaling of biomimetic organ cultures, and differentiation and expansion of the required cell types from clinically feasible sources [60]. Pigs may be suitable for decellularized kidney scaffolds because of the similarity in size and microarchitecture between pig and human kidneys [66]. To achieve repopulation with a broader variety of different kidney-like cell types and appropriate organization, kidney organoids derived from hPSCs may be suitable cell sources for recellularization. However, whether kidney organoids derived from iPSCs can be used as cell sources to seed decellularized kidney scaffolds has not been evaluated.