The mammalian metanephric kidney is a highly complex organ responsible for many functions. The main ones include filtering the blood from waste products, maintaining the electrolyte and pH balance of the body fluids, regulating the bone mineral metabolism, and controlling the blood pressure and blood composition. Many of these functions are performed by the nephron, which is a main functional unit of the kidney.

Therefore, the main goal of the differentiation protocols is to obtain the structure that recapitulates many of the key structural and functional features of the nephron. Accordingly, so much effort has been put into studying in detail all the stages of nephron development.

The first protocols for differentiating kidney organoids from PSCs have been established (6, 9, 19–32) ( Table 1 ). From the outset, the protocols were based on the principle of following the natural processes of embryonic kidney differentiation. Hence, it is necessary to study all of the steps of kidney development and to learn how to repeat these steps sequentially.

Considering the immensely complex organization of the mammalian kidney, reproducing its entire morphology and function in vitro would be an extremely challenging task. Therefore, many projects have focused on the recapitulation in vitro of at least some morphological or functional components of the kidney, e.g., a model of glomerular filtration or the renal tubule representing the re-absorption function, among others. An elegant and logical solution was to build these miniaturized renal “units” on a microfluidic platform.

The use of microfluidic chambers or so-called microfluidic chips, also known by other names such as “lab-on-a-chip,” became widely used in biomedical research about 20 years ago (81). It relies on the use of designs that can process small quantities of liquids using the channels that have microscale dimensions, i.e., typically tens to hundreds of micrometers. The strength of this method is that the microfluidic chambers (or microfluidic chips) can be used to create complex, controlled microenvironments in microchannels. The microfluidic method can provide a sophisticated cell bioanalysis platform by integrating multiple steps including fluid control, cell culture, cell capture, cell–cell and cell–matrix interactions, cell lysis, cell signaling, and biochemical detection in a single device. When research is carried out to study a whole organ or organoid, this method is called organ-on-a-chip (OoC) (78). The main goal of the OoC method is not to recreate an entire living organ with all its functions. The reconstruction of at least one critical function for each model should be considered a great success.

One method for the development of microfluidic systems is photolithography using transparent and gas-permeable polydimethylsiloxane (PDMS) (82). Another widely used method is 3D printing (83). A large list of commercially available microfluidic cameras is also available (82).

Currently, a wide variety of microfluidic devices are used to study the biological processes taking place in the vascular, respiratory, nervous, digestive, and excretory systems (84).

Several studies have aimed to reproduce the renal proximal tubular system in vivo using microfluidic devices (85–88). A bioprinting method to create 3D human proximal tubules in vitro, which are inserted into a perfused tissue microfluidic chip, has been reported (89, 90). In these studies, a 3D kidney tissue allows the proximal tubular epithelium and the vascular endothelium to be co-cultured, making this 3D vascularized proximal tubule model suitable for studying renal reabsorption and the crosstalk between the epithelium and the endothelium. Another type of kidney-on-a-chip platform, which is based on a PDMS mold, has been reported to allow the reconstitution of the human kidney vascular tubule unit in vitro using patient-derived cells (91). The subsequent variant of the microfluidic platform for the co-culture of the epithelium and endothelial tubule was developed, allowing confocal imaging and the study of the tubular morphology of renal proximal tubule epithelial cells and human umbilical vein endothelial cell (HUVEC) cultures (92).

Earlier, an in vitro functional vasculature that mimics the glomerular tuft was reported (93). A glomerulus-on-a-chip could not be developed until functional human podocytes, the cells that regulate selective permeability in the glomerulus, became available. Once a protocol for differentiating the hPSCs into podocytes expressing mature phenotype markers (nephrin⁺, WT1⁺, podocin⁺, and Pax2⁻) and capable of exhibiting primary and secondary foot processes was presented, these podocytes were co-cultured with human glomerular ECs in an OoC microfluidic device (71, 94). In these studies, the function of the kidney glomerular capillary wall was restored on a microfluidic chip. Subsequently, a model for efficient drug screening and nephrotoxicity test using a co-culture of renal proximal tubule epithelial cells and peritubular capillary ECs in a microfluidic kidney chip was proposed (95).

All the models described above can be called renal tissue models. A kidney organoid-on-a-chip, which represents an entire kidney, was described in a study that used a millifluidic device (80). A hiPSC-derived kidney organoid embedded in a gelatin-fibrin (gelbrin) ECM in a 3D-printed millifluidic chip was superfused (flowed around their top surface), which increased the number of EC precursors and then created a vascular network with perfused lumens surrounded by smooth muscle cells and pericytes. This increased the vascularization of the kidney organoid, which will be discussed below.

Currently, two methods of in vivo vascularization are used: the implantation of organoids into an immunodeficient host animal and the grafting on the chorioallantoic membrane (CAM) of a chicken embryo.

All protocols for the differentiation of PSCs into kidney organoids have so far failed to generate functional vasculature. Some protocols have resulted in the generation of PSC-derived kidney organoids with immature ECs that do not form blood vessels or migrate into the glomeruli to form a glomerular tuft (6, 21, 79). Various optimizations of these differentiation protocols, such as the addition of VEGF, only increased the number of ECs that are still unable to initiate invasion into the glomeruli (7).

Takebe et al. (123) presented a method for organ bud formation from different tissues by combining PSC-derived tissue-specific progenitors or relevant tissue samples with ECs and mesenchymal stem cells. They showed that, as well as in the case of kidneys, the mesenchymal stem cells initiated condensation in vitro within these heterotypic cell mixtures, and these kidney buds, after cranial transplantation into immunodeficient mice, revascularized and were able to efficiently self-organize into functional tissues. Mesenchymal-derived cell-driven condensation in vitro with the presence of ECs in the primary mixture of cells demonstrated the ability to self-assemble in vitro from the desired multiple cell/tissue types into a complex tissue. This can be regarded as a step forward in the creation of an in vitro model of a functional vascularized kidney.

Another study proposed a method to reconstitute both the kidney and its vascular architecture in vitro using dissociated and sorted mouse embryonic kidney cells and incorporating ECs into the system and subsequent transplantation (142). This allowed demonstrating that ECs of donor origin make a significant contribution to the formation of arterioles and glomerular capillaries formed after transplantation.

The incorporation of ECs and mesenchymal cells later in a 3D co-culture system with hiPSC-derived kidney precursors and the subsequent vascularization on chicken CAM increased the maturation of renal structures. This demonstrated the possibility of optimizing the in vitro step before inducing vascularization by transplantation, grafting on CAM, or using other approaches (137).

Another potentially very effective strategy is to vascularize the kidney organoids entirely in vitro. The ability to successfully vascularize the kidney organoids in vitro will pave the way for their use in many areas: drug screening, kidney disease modeling, and transplantation, among others. Most importantly, this might elucidate the fundamental mechanism of glomerular vascularization and kidney development in general.

Various attempts have been made to vascularize the kidney in vitro (80, 127–129). Filtration of the blood and the production of urine meant that the kidneys are always exposed to the blood and urine flow. It is therefore not surprising that static in vitro culture conditions do not result in the proper development of kidney organoids. Therefore, an air–liquid interface perfusion system was developed for the organotypic culture of renal organoids derived from hPSCs (127). This system results in the accelerated organization of epithelial cells and tissues in renal organoids in vitro, but not of ECs.

The first promising system was developed as a millifluidic device in which hPSC-derived kidney organoids were cultured under flow (80). The flow conditions and supportive ECM allowed accelerating the maturation of the organoid vasculature. Even the flow along the top surface of the organoids, rather than through it, led to an increase in the number of EC precursors, subsequently allowing them to organize into vessels. The EC vascular structure in the kidney organoids was improved under flow conditions compared to static conditions. Vascular invasion of the glomeruli, organization of the capillary loops, and partial perfusion through the lumen were observed. An important condition was the absence of exogenous VEGF. Only flow particularly enhances the development of the glomerular vasculature.

A similar kidney organoid-on-a-chip system providing fluid flow that mimics shear stress with optimized ECM conditions (Matrigel with VEGF) also demonstrated that kidney organoids cultured in fluid flow conditions exhibit more mature podocytes and vascular structures compared to the static culture condition (128).

A later published study (28) showed that optimizing the differentiation protocol for hiPSC-derived kidney organoids through the precise modulation of Wnt signaling led to the establishment of a specific glomerulus-to-tubule ratio. Thus, the appropriate amount of VEGF was obtained, which coordinated the development of a larger volume of ECs and a higher level of development of the resident vascular network. Functional validation based on the dextran uptake showed that the vasculature developed in vitro correlated with a similar dextran uptake assay performed using kidney aggregates derived from E12.5 mouse embryonic kidneys. In summary, this study successfully induced the vascularization of organoids in vitro, but further development of the blood perfusion system is still needed.

Recently, another important step has been made toward the in vitro vascularization of renal organoids (129). A novel OoC system supports the culture of kidney organoids that enhances the maturation of endothelial populations based on co-localization analysis of the endothelial markers. In addition, the migration and proliferation of HUVECs cultured in the channels of the chip within the organoid tissue was observed. An interconnection has been found between the HUVECs and endogenous ECs. Later, structures were formed that represent hollow endothelial tubules, reminiscent of blood vessels. As it appears, this work accomplished the closest match to the requirement of the development of an in vitro vascularized kidney model with the perfusion organized through endothelial vessels. What is still lacking in this approach is a demonstration of the glomerular vascularization. In our point of view, the formation of the glomerular tuft is the most important hallmark of the development of a functional kidney. Thus, it could be an important goal of future research.

The mammalian vascular system consists of the blood and the lymphatic vascular systems. Blood vessels are lined with blood ECs, and lymphatic vessels are correspondingly lined with lymphatic ECs. Both types of ECs have their own unique structure and gene expression profiles based on specific markers identified for them (143). Blood vessels are composed of ECs and vascular mural cells (VMCs), including pericytes lining the microvessels and vascular smooth muscle cells lining the large diameter vessels such as arteries and veins (144, 145). ECs line the inside of the tube and are in contact with the flowing blood, while vascular mural cells control the vascular tone and provide structural support (146).

Different organs perform different functions, and this is why ECs from different organs acquire their own specialization and specific gene expression profile during development. The kidney has very specialized functions and has a blood vascular system that is different from those of other organs (143). In the embryonic kidney, the VPs, SPs, and NPs are derived from the MM. VPs and SPs co-localize at the E11.5 stage, as it was reported that the endothelial progenitors CD146⁺CD31⁻ are derived from the FOXD1 stroma population (113) ( Figure 2A ). Both the VPs and the invaded vasculature form the endothelium system of the nephron in adult kidney ( Figure 2B ). All three types of cells that form blood vessels—ECs, pericytes, and vascular smooth muscle cells—are closely associated with the renal epithelial system (146) ( Figure 2B ). The ECs of the renal vasculature have a very high level of heterogeneity, depending on which part of the kidney they are found in and, accordingly, what function that part of the kidney performs (143) ( Figure 2B ).

Both the endothelia and VMC progenitors are present in the MM and differentiated by the FOXD1 gene at the mE10.5 stage (44, 147). They vascularize the glomerulus through the synergistic processes of vasculogenesis and angiogenesis, which were described above. All of the major stages of kidney vasculature development occur at mE10.5–mE15. Analysis of the single-cell transcriptome of hPSC-derived organoids generated using two different protocols and compared with human kidneys at 16 weeks and with adult kidneys showed that the organoids contained 12 distinct types of kidney cells, but their level of maturation was closer to that of fetal kidney cells than to adult kidney cells (74). The spatiotemporal heterogeneity and the patterning of the developing kidney blood vessels have been studied to understand when and how endothelial regionalization occurs (148). An in-depth anatomical and molecular analysis of the developing kidney endothelium confirmed that ECs grow synchronously with the UB already at stage mE10.5 and then begin to show signs of specification at stage mE13.5, when the first arteries can be identified. Transcriptional profiling has identified a long list of novel genes enriched in kidney ECs, expressed in regionally defined and organ-specific patterns, and has shown that the renal endothelium exhibits marked heterogeneity as early as mE15.5 and that many of the EC gene expression patterns are highly dynamic over time. The association of ECs with UB was analyzed by immunofluorescence (IF), which showed that, from stage mE14.5, this association started and was maintained for the rest of the embryonic kidney development, except for the renal pelvis, where the ECs were closely associated with the epithelial cells of the renal papilla, but not with those that constitute the ureter. The next step forward in the study of the progression of the molecular pathways that determine kidney vascular heterogeneity has been the study using bulk and single-cell RNA-seq of the kidney vasculature at different stages of development, which revealed distinct transcriptomic signatures associated with the different kidney vasculatures (149). In this study, several genes coding transcription factors and their putative targets that define the ultrastructural, phenotypic, operational, and functional identity of the glomerular capillaries and the surrounding vascular network were identified. The transcription factor Tbx3, which is robustly expressed in glomerular ECs and regulates the transcriptional network that determines the specification and function of glomerular ECs, has been characterized. This work traced the development and specialization of the renal vasculature, starting from the mE14 stage, in the process of maturation into adult vessels. The period between mE10.5 and mE14 was not included in this work. It should also be considered that, in this work, only the EСs were analyzed and that the influence of other cell types was not taken into account. Current knowledge of the phenotypic and molecular heterogeneity of the kidney endothelium provides a very detailed list of the expression markers specific to each defined subpopulation of renal ECs (143).

The glomerulus, the filtering unit of the kidney, consists of four types of cells: the fenestrated ECs, the epithelial cell podocytes, the perivascular mesangial cells, and the parietal epithelial cells (150, 151). The podocytes, GBM, and the fenestrated ECs form the blood-filtering structure called the glomerular capillary tuft. Mesangial cells provide the rigidity and integrity of the glomerular tuft and serve as its structural framework. Parietal epithelial cells form the Bowman’s capsule surrounding the glomerular tuft. Glomerular development is generally described in five steps: the renal vesicle, the comma-shaped body, the S-shaped body, the glomerular capillary loop stage, and the mature glomerulus ( Figure 2C ).

Glomerulogenesis consists of one step before the invasion of ECs (the vesicle stage), then is divided into two steps (the comma-shaped body and the S-shaped body) when the podocyte progenitors begin to express VEGF and the MM-derived glomerular EC precursors expressing Foxd1, VEGFR2 (FLk1/Kdr), and VEGFR1 (FLT) begin to interact with the podocytes. During the last two steps (the glomerular capillary loop stage and the mature glomerulus), all four cell types interact with each other and the mature glomerulus (147, 150, 152) ( Figure 2C ). Among the growth factors, VEGF is a fundamental regulator of EC differentiation, migration, and cell–cell interactions, including the sprouting angiogenesis stimulation and tip cell activation (153). VEGF pathway interactions initiate the migration of ECs into the vascular cleft of an S-shaped glomerulus. The Notch pathway regulates the expression of all the VEGF receptors (VEGFRs). At the same time, VEGFA directly stimulates the Notch pathway during vascular development. Subsequent signaling through VEGFR2 requires the presence of VEGFR3, ephrin-B2, neuropilin-1, the thrombospondin receptor CD47, VE-cadherin, and some integrins (154, 155).

In the first phase, the glomerular EC precursors form a capillary chord without a lumen; subsequently, due to the expression of TGF-β by podocyte precursors, the formation of a capillary lumen occurs. This process requires VEGF to induce apoptosis, through which excess ECs will be eliminated (156, 157). The ECs of the glomerulus at the late stage of the S-shaped body form the fenestrated endothelium along with foot process formation in overlying podocytes ( Figure 2C ).

Glomerular ECs expressing platelet-derived growth factor beta (PDGF-β) attract PDGF receptor beta (PDGFR-β)-expressing mesangial cells, which migrate into the cleft and participate in glomerular maturation (150).

The chemoattractant cytokine C-X-C chemokine ligand 12 (CXCL12) expressed in podocytes during the S-shaped body stage requires its receptor C-X-C chemokine ligand 4 (CXCR4) expressed in ECs, suggesting an intimate crosstalk between these cells by paracrine signaling.

Simultaneous with the formation of the capillary chord in the glomerulus, the angiopoietin–Tie signaling axis determines the formation of a vessel sprouting from remnant glomerular EC precursors. The mesenchymal cells surrounding the developing vasculature express Ang1, which promotes vessel stability and inhibits fibrosis in a paracrine manner (158). ECs are the main Ang2-expressing cells that have an autocrine function and participate in context-dependent regulation. The angiopoietin receptors Tie1 and Tie2 are expressed predominantly in ECs. Following the invasion of ECs to form glomeruli, podocytes and ECs undergo synergistic differentiation to create a filtration barrier.

Immediately after the appearance of blood flow in the capillary system, the luminal surface of the endothelium begins to be constantly exposed to and respond to shear forces. Vascular development occurs through a series of flow-sensitive processes. Shear stress regulates multiple processes in vascular development, including sprouting, EC apoptosis and migration, and arterial fate specification (159). Numerous transcriptome studies of ECs under different flow conditions have shown that shear stress regulates the signaling pathways BMP/TGF-β, Hippo/YAP/TAZ, Notch, Wnt, and HOX (160). All of these pathways participate in kidney nephron development and in endothelial vasculature formation and maturation.

The formation of blood vessels is the result of the interaction of perivascular cells and the extracellular matrix, which has viscoelastic properties and exposes the cells to traction stress (161). The transition from a 2D to a 3D culture using different matrices in the context of organoids not only better resembles the in vivo situation but also ensures the correct cell fate: when using the same differentiation protocol, the initial 2D culture of undifferentiated PSCs has been shown to result into differentiation into cardiac lineages, whereas the initial 3D culture directs differentiation toward the kidney (6). The choice of a 3D matrix for renal organoid culture is critical because the composition of the ECM also influences cell fate (162). Currently, several types of ECM have been developed for 3D culture, including animal-derived ECM and synthetic matrices (163). The ECM obtained from animal tissues (Matrigel) was, for a long time, the only published matrix that allows the expansion, differentiation, and passage of renal organoids (163). However, well-known disadvantages of Matrigel, such as the batch-to-batch variability and the presence of growth factors, make complete control of the ECM impossible. In addition, Matrigel does not reflect the organotypic ECM of the kidney and is not suitable for transplantation studies. Despite this, Matrigel with VEGF has been used to vascularize the kidney organoid-on-a-chip in several reports (128). Tissue-specific types of ECMs isolated through decellularization and digestion of the kidney have also been tested for the optimization of the protocols of renal organoid generation (164, 165). Hydrogels made from natural materials, including gels based on collagen, fibrin, and polyglycans, are most often used for the fabrication of kidney-on-a-chip (166). Gelbrin (80) or fibrin-based hydrogels have been reported to show the best results in supporting renal organoid formation and in in vitro vascularization (129).

The discovery of better treatments for kidney diseases requires developing improved model systems to study the disease mechanisms and examining new therapeutic approaches. A functional in vitro vascularized kidney organoid from patient-derived hPSCs could open new opportunities in the field of regenerative kidney medicine. Depending on how the problems that hinder the achievement of this goal were identified, different strategies must be chosen ( Figure 3 ).

One direction is to optimize the protocol of the differentiation of PSCs into a kidney organoid. More precise tuning of Wnt signaling during kidney organoid differentiation has provided a way to produce the glomeruli and tubules in a specific ratio, allowing podocytes to produce the appropriate amount of VEGF to attract EC precursors, leading to their more proper development and infiltration into the glomeruli (28). It is unclear whether this ratio corresponds to that of the glomeruli and tubules in the embryonic kidney, but it does result in a higher level of development of the resident vasculature in the organoid. Any future improvement in the differentiation protocol, whether it follows a natural process or not, would be beneficial. In general, the method of following the natural process leads to better results, but requires a more detailed study of the embryonic kidney, which is made possible by modern methods.

Optimization of the differentiation protocol also includes the use of exogenous ECs, such as primary ECs (i.e., HUVECs and human glomerular microvascular ECs) (80). In this particular protocol, exogenous cells did not appear to improve vascularity; however, potentially using a different EC type increases the opportunity to determine an effective way of optimizing the protocol. The generation of other non-renal organoids (e.g., cardiac spheroids) involves the step of adding hPSC-derived ECs, which appears to be a potentially very promising method of improving the maturation of other cell types, not just ECs (167). An attempt to co-differentiate the hiPSCs into the renal and endothelial lineage has been made previously (6, 21–23, 28). Some of the important factors in vascularization in engineered tissues include the type of ECs used and their ratio to PSCs or other cell types (143, 168). The recently published first PSC-derived vascular organoid model consisting of ECs and pericytes has shown significant lumen formation (169, 170). Given the common mesodermal origin of the kidney and ECs, attempts at co-differentiation or co-cultivation within the organoid seem quite logical. The ability of modern methods to perform assessment of the cellular heterogeneity and enrichment of the different populations also appears to be a very promising strategy for optimizing the kidney organoid differentiation protocol.

It has been convincingly shown that, between static and flow conditions, the latter significantly increases the level of maturation of the kidney organoids (80). However, the manner in which this flow passes can be designed in different ways (129). The development of new microfluidic or other engineered solutions could be a very promising strategy to enhance the influence of flow on kidney organoid maturation.

The search for the most suitable ECM is ongoing (80, 128, 139). Influencing cell fate using the correct ECM could offer additional ways to optimize the kidney-on-a-chip model.

Various strategies from the organoid field have been used to create a vascularized kidney organoid-on-a-chip, and others should also be applied. The methodological challenge of co-culturing kidney organoids and ECs forming the vasculature lies in the different media required by these two organs. In a generalized method for organ bud formation from diverse tissues through combining embryonic kidney-derived dissociated cells with ECs and mesenchymal stem cells, a medium consisting of an EC medium and a kidney cell medium in a 1:1 ratio was used (123). In a more recent paper on enhancing the maturity of the renal progenitors in renal microtissue derived from the co-culture of renal precursors with endothelial and mesenchymal stem cells, the ratio of the three cell types was optimized, but the ratio of the two different media was the same as that in the above-mentioned article (137). Another existing strategy for generating various organoids is pre-vascularization (171). This allows first using a specific EC medium to create a perfused vasculature and then moving on to an organ-specific medium. Using only one type of medium for both the development of an organoid and the creation of a perfused vascular network seems impractical. A new method of culturing 3D tissue, i.e., placing it directly on a “vascular bed-on-a-chip,” could potentially be applied to a kidney-on-a-chip model as well (172).