Large-Scale Engineering and Cryopreservation of hiPSC-Derived Nephron Sheets


 Background The generation of human induced pluripotent stem cells (hiPSCs) has opened a world of opportunities for stem cell-based therapies in regenerative medicine. Currently, several human kidney organoid protocols are available that generate organoids containing kidney structures. However, these kidney organoids are relatively small ranging up to 0.13 cm2 and therefore contain a small number of nephrons compared to an adult kidney, thus defying the exploration of future use for therapy. Method We have developed a scalable, easily accessible, and reproducible to increase the size of the organoid up to a nephron sheet of 2.5 cm2 up to a maximum of 12.6 cm2 containing a magnitude of nephrons. Results Confocal microscopy showed that the subunits of the nephrons remain evenly distributed throughout the entire sheet and that these tissue sheets can attain ~30,000-40,000 glomerular structures. Upon transplantation in immunodeficient mice, such nephron sheets became vascularized and matured. They also show reuptake of injected low-molecular mass dextran molecules in the tubular structures, indicative of glomerular filtration. Furthermore, we developed a protocol for the cryopreservation of intermediate mesoderm cells during the differentiation and demonstrate that these cells can be successfully thawed and recovered to create such tissue sheets. Conclusion The scalability of the procedures, and the ability to cryopreserve the cells during differentiation are important steps forward in the translation of these differentiation protocols to future clinical applications such as transplantable auxiliary kidney tissue.


Introduction
More than 10% of the world population suffers from chronic kidney disease [1] that can result in end-stage renal disease (ESRD) in which kidney function is lost. In 2030, an estimated 5.4 million people globally will have ESRD [2], fuelled by an aging population and increase in obesity, diabetes and cardiovascular disease [3]. To cope, ~5% of national healthcare budgets would need to be allocated to dialysis and expensive renal replacement therapies [4]. Kidney transplantation is currently the best clinical option but is hampered by organ availability and the use of immune suppressive drugs. These limitations drive the eld of regenerative medicine to develop attractive therapies by engineering, replacing or regenerating cells, organs or tissues in order to restore function after damage or loss. The eld has seen tremendous growth with the introduction of human induced pluripotent stem cells (hiPSC) [5] that can differentiate into any cell of the body. Generation of hiPSC-derived human kidney tissues may serve as an future alternative source of kidney tissue that could be used for transplantation purposes to (partially) restore kidney function [6]. In line, we previously demonstrated that transplanting hiPSC derived kidney organoids under the renal kidney capsule of immune de cient mice resulted in vascularisation, maturation, and functional and size selective glomerular ltration [7,8].
However, to take this concept into further development towards clinical application, it is essential to demonstrate that the culture of these hiPSC-derived kidney organoids can be scaled to contain meaningful numbers of nephrons for such clinical application. Current dialysis treatment, on average, delivers 10% of the normal glomerular ltration and clearance. We therefore reasoned that a scaled culture system should be able to deliver at least 50.000 glomeruli to be considered as potential future auxiliary transplant tissue. Here we report a method to culture hiPSC-derived nephron tissue in sheets that contain tens of thousands of glomerular structures, while maintaining their ability to be transplanted and become functional. In addition, we show that the intermediate mesoderm cells in the differentiation culture can be cryopreserved thus enabling a structured and planned manufacturing process of these nephron sheets.
Differentiation of organoids and nephron sheets Differentiation of organoids was described previously [7,11]. Brie y, cells were seeded as single cells at variable densities (8,000 -26,000 cells/cm 2 ) per cell line on 6 well plates. Differentiating cells were dissociated on day 7 and centrifuged at 400xg containing 5 x 10 5 cells per tube. Cell clumps were pipetted on top of a 0.4 µm pore transwell membrane (24 mm, Corning) and further maintained until day 7+18.
For differentiation to hiPSC-derived nephron sheets, hiPSCs were dissociated using TrypLE Select (Thermo Fisher Scienti c) and plated at similar densities as above on a vitronectin coated T75 (Cellstar) in E8 medium supplemented with Revitacell. Differentiation was started the following day (day 0) by replacing medium with STEMdiff APEL2 Medium (APEL2, Stemcell Technologies) containing 1% PFHMII (Life Technologies), 1% Antibiotic-Antimycotic (Life Technologies) and 8 µM CHIR99021 (Tocris). On day 4 medium was switched to APEL2 supplemented with 200 ng/mL rhFGF9 (R&D Systems) and 1 µg/mL Heparin (Sigma-Aldrich). On day 7 intermediate mesodermal cells were dissociated after a 1 h pulse with 5 µM CHIR in APEL2 using Trypsin-EDTA (Thermo Fisher Scienti c). Cells were counted using an automated cell counter (NC-200) and centrifuged at 260xg in 50 mL tubes. Cells were seeded at a density between 19.5 -23 x 10 6 cells/cm 2 on a 0.4 µm pore transwell (75 mm, Corning) using two different templates for cell distribution. Cells were plated either inside a rubber ring (DWK life science) or overlaid with a silicone cover (GRACE BIO-LABS). Nephron sheet differentiation was continued in APEL2 containing FGF9 and heparin with a media change every other day. On day 7+5 the template was removed from the sheet, and medium was changed to plain APEL2. Tissue sheets were maintained until day 7+18 before xation with 2% PFA (Alfa Aesar) diluted in PBS.

Cryopreservation of Differentiating Kidney Organoids
Differentiating intermediate mesoderm cells from 4 independent differentiation experiments were cryopreserved on day 7. Cells were dissociated after CHIR pulse using Trypsin-EDTA, and single cells were counted and resuspended in ice-cold Nutrifreez (Biological Industries). Cryovials containing 8 x 10 6 cells/mL were rate controlled (-1°C/min) frozen to -80°C. 24 h later vials were transferred and stored in liquid nitrogen. For thawing, vials were warmed at 37°C and cells were transferred to a 50 mL tube containing plain DMEM and counted. Cells were centrifuged at the indicated density, resuspended in a small volume of 10% FBS in DMEM and pipetted onto the transwell membrane to allow tissue sheet formation using the templates. The protocol was continued as described above until day 7+18.

Animal Experiments
All animal experimental protocols were approved by the animal welfare committee of the Leiden University Medical Center and the Dutch Animal Experiments Committee. To allow for transplantation of nephron sheets, a hollow punch (Renssteig) was used to create a 2 mm circular biopsy. These biopsies were transplanted at day 7+17 under the renal capsule of both kidneys in 4 seven-week-old recipient mice (non-obese diabetic/severe combined immunode ciency (NOD/SCID), Charles River Laboratories). Before sacri ce two mice were anesthetized with iso urane and injected with low molecular mass 10 kDa dextran labelled with Tetramethylrhodamine (TRITC, TdB Labs). Tissues were collected after 14 days of transplantation and processed for immunohistochemistry or transmission electron microscopy.

Immunohistochemistry
Organoids, untransplanted and transplanted nephron sheets were processed and stained as described previously [7]. In vitro organoids were xed in 2% PFA for 20 min at 4°C before storage in PBS at 4°C. For further analysis biopsies of nephron sheets were either made with a blade or a hollow punch (3 mm, Renssteig). If required, tissues were embedded in TissueTek (Sakura) and stored at -80°C. Transplanted tissues were either immediately snapfrozen in Tissue Tek or xed in 2% PFA overnight. Fixation of tissues containing uorescent dextran was followed by 8 h incubation in 30% sucrose (Sigma) in PBS (B-BRAUN) at 4°C. Transplanted tissues were either stored in PBS or embedded in TissueTek. Tissue sections from embedded samples were made using a cryotome (5 -10 µm thick).
For immuno uorescence analysis of in vitro organoids and nephron sheets (whole mount, biopsies or slides), samples were blocked in 10% Donkey Serum (Sigma-Aldrich) and 0.3% TritonX (Sigma Aldrich) in PBS for 2 h at room temperature. Samples were stained with primary antibody in blocking buffer for 24 -72 h at 4°C. For transplanted tissues Mouse on Mouse kit (MOM; Vector Laboratories) was used. Tissues were stained with primary antibodies for nephron structures: NPHS1 (R&D), NPHS2 (Abcam), ECAD (BD), LTL and DBA (Vector Laboratories), CUBN (Thermo Fisher Scienti c and Abcam); endothelial cells: CD31 (BD) and MECA-32 (BD); and stromal cells: MEIS1/2/3 (Active Motif) and PDGFRα/β (Abcam). Samples were washed 3 times before adding secondary antibody mix for 2 h at room temperature (Supplemental Table 1). Nuclei were occasionally counterstained with Hoechst 33258. Organoids and biopsies from nephron sheets were mounted with ProLongGold (Thermo Fisher Scienti c) in 35 mm glass bottom dish with a 14 mm or 22 mm glass diameter (MatTek corporation, Willco Wells), or 50 mm glass bottom dish with a 30 mm glass diameter (MatTek corporation). Our largest tissue sheet was mounted on a glass bottom plate (MatTek corporation). Nephron sheets were topped with 12, 20 or 55 mm coverslip, and left to dry for 24-48 h. Samples were imaged using a White Light Laser Confocal Microscope TCS SP8 (Leica) with LAS-X software 3.5.5 and an Andor Dragon y Spinning Disk with Fusion 2.2 or higher software. LAS-X Image with 3D module and Imaris 9.5.0 were used to further analyze the data.

Counting glomerular structures
Organoids and biopsies of nephron sheets were stained for glomerular marker NPHS1 and nuclei with Hoechst. During confocal imaging Z-compensation for Excitation and Detector gain were used to create equal uorescence intensity throughout the entire sample. NPHS1-positive structures were visualized using Imaris and further analyzed using the application 'Surfaces' to determine total volume and individual volume of glomerular structures (Supplemental Figure 1). The number of glomerular structures was calculated in individual images from independent experiments. This number was converted to number of glomerular structures in 3-dimensional organoids and nephron sheets.

Transmission Electron Microscopy
For Transmission Electron Microscopy (TEM) analysis, small biopsies of transplanted and nontransplanted nephron sheets were sampled and xed for 1.5 h at room temperature in 1.5% glutaraldehyde (Electron Microscopy Sciences) in 0.1 M sodium cacodylate buffered solution, with pH 7.4. Samples were subsequently rinsed with sodium cacodylate buffer, xed in 1% osmium tetroxide (Electron Microscopy Sciences) in 0.1 M sodium cacodylate buffer for 1 h on ice, then washed with sodium cacodylate buffer and dehydrated in a series of 70%, 80%, 90% and 100% ethanol. Next, samples were in ltrated with a mixture of 1:1 Epon LX-112 (Ladd Research) and propylene oxide (Electron Microscopy Sciences) for 1.5 h, followed by in ltration with pure Epon for 2 h. Afterwards, samples were mounted in BEEM capsules (Agar Scienti c), embedded in pure Epon and polymerized for 48 h at 60°C. Ultrathin sections (100 nm) were collected on copper slot grids (Storck Veco BV), covered with formvar lm and a 6 nm carbon layer. Sections were contrasted with an aqueous solution of 7% uranyl acetate for 20 min, followed by Reynolds's lead citrate for 10 min. Imaging was performed at an acceleration voltage of 120 kV on a FEI Tecnai G 2 Spirit BioTWIN transmission electron microscope (FEI), equipped with an Eagle 4K slow-scan charge-coupled device camera (FEI). Large virtual slides of glomerular and tubular structures were acquired using automated large-scale data collection and stitching software [12] at 13,000x and 18,500x magni cation. Aperio ImageScope software (Leica Biosystems) was used for the visualization of the virtual slides.

Results
Engineering hiPSC-derived nephron sheets using a template The original differentiation protocol creates hiPSC-derived kidney organoids with glomerular, proximal tubular and distal tubular structures comparable to human kidney, but these tissues only range up to 4 mm in diameter ( Figure 1A) [7,11]. Here we focused on designing an easily accessible protocol to generate large hiPSC-derived nephron sheets and aimed to scale-up both the monolayer and 3dimensional phase. The initial phase of monolayer culture was scaled up to a T75 cell culture ask with comparable seeding density as the original protocol.
Expansion of the 3-dimensional phase of differentiation was performed on a transwell membrane of 7.5 cm diameter. On these transwell membranes templates are necessary to enforce a restricted culture site for standardized seeding and reproducibility. We made use of 2 templates: a rubber ring with an inner circle of 18 mm and a silicone cover of 18 mm ( Figure 1B and C). Both templates yielded similar nephron sheets that are indistinguishable by eye, structures observed in bright eld imaging and expression of kidney markers by immuno uorescence analysis ( Figure 1D). We therefore combined results from both templates in this study. With these templates we created nephron sheets with a surface area of ~2.5 cm 2 , while organoids only have a surface area of ~0.13 cm 2 . After 5 days tissue sheets were rm enough to allow for the template to be removed. Sheets further developed in the same manner as the regular organoids. We explored the ability to generate large nephron sheets even further by using a 4 cm silicone cover ( Figure 2). This allowed us to create a tissue sheet with a surface area of 12.6 cm 2 that showed similar characteristics as the 2.5 cm 2 as observed by eye and microscopy ( Figure 2A).

Assessment of hiPSC-derived nephron sheets and counting glomerular structures
Nephron sheet diameter was comparable in LUMC0072, LUMC0099 and iPSC-MAFB, while LUMC0020 yielded slightly smaller sheets ( Figure 3C). We aimed to assess the number of glomerular structures (NPHS1 + ) in these tissue sheets and organoids from multiple independent experiments by using the size of individual glomerular structures and their distribution. The size of individual glomerular structures was determined by immuno uorescence analysis with Imaris. The sizes were comparable between cell lines and both types of organoids ( Figure 3D). Total number of glomerular structures was evaluated and the average number of structures in organoids ranged between ~1,300 -4,000 structures, and betweeñ 30,000 -40,000 in nephron sheets ( Figure 3E).

hiPSC-derived nephron sheets become vascularized and mature upon transplantation
To evaluate whether nephron sheets vascularized and further matured upon transplantation, they were cultured until d7+17 and transplanted in mice. Due to space limitations under the renal capsule, we transplanted a biopsy of 2 mm of the tissue sheet. After two weeks, before sacri ce, 2 mice were intravenously injected with 10 kDa TRITC-labeled dextran. Confocal imaging showed the presence of all kidney structures and vascularization of the nephron sheet ( Figure 4A)

hiPSC-derived nephron sheets can be cryopreserved during differentiation
To explore if nephron sheets could be cryopreserved during differentiation, we froze cells at day 7 of differentiation (intermediate mesoderm phase) in cGMP-manufactured cryopreservation medium ( Figure   5A). Nutrifreez provided excellent recovery and tissue sheets were able to continue differentiation after cryopreservation ( Figure 5B, Supplemental Figure 5). Non-frozen and thawed sheets were compared under identical experimental conditions and immuno uorescence showed the presence of glomerular, proximal and distal tubular structures and endothelial cells ( Figure 5C).

Discussion
Here we engineered a reproducible, scalable hiPSC-derived nephron sheet that can be easily implemented, and show that cryopreservation during differentiation allows for storage of intermediate mesoderm cells, containing progenitors of the kidney. These tissue sheets showed presence and even distribution of kidney structures, high number of glomerular like structures, and vascularization and maturation upon transplantation under the renal capsule in mice. As observed in previous transplantation studies [8,13], ltration in glomerular structures was demonstrated by presence of low molecular mass dextran molecules in tubular structures.
Currently multiple protocols are available for designing iPSC-derived cell or tissue sheets, such as retinal pigment epithelium cells on collagen I, smooth muscle like cells on degradable hydrogel, and various cell types on temperature responsive cell culture surfaces like cardiac tissue and hepatocyte-like cells [14][15][16][17]. Some methods yield a single sheet layer, while others are stacked to create a multi-layered sheets [18,19]. Lawlor et al. explored the use of a bioprinter to generate hiPSC-derived kidney organoids [20]. Their printed organoid patch (6 x 8 mm) consisted of a 3-dimensional layer. Our protocol does not require a bioprinter and it is designed to combine manual pipetting and a template to guide nephron sheet growth and patterning. The template creates an optimal 'cell to square surface' ratio of differentiating cells resulting in a 3-dimensional sheet. Without a pre-de ned culture area, cell density is suboptimal, resulting in an uneven and poor differentiation. Nephron sheets made by either template showed minimal differences in distribution of structures. The templates are easy to use and can be removed without damaging the sheet, allowing further development of the tissue sheet. These nephron sheets can easily be moved from the transwell lter and are strong enough to be transferred without breaking. Additionally, it is important to note that shape and size can be adapted, at least up to 12.6 cm 2 . For clinical purposes, designing an industrialized process for tissue sheet culture could include automated hiPSC-maintenance and differentiation during the monolayer phase, for example in a bioreactor or a multi-layered ask. After harvesting, cells could be further differentiated or cryopreserved. A combination of multiple systems combining hiPSC culture and nephron sheet differentiation could result in an engineered nephron sheet platform.
At present the only treatment for reduced kidney function is dialysis and kidney transplantation. These tissue sheets could potentially ll a gap that alleviates or delays the need for such therapy. We also counted glomerular structures in the tissue sheets as an indication for the amount of nephrons that can be transplanted. We found that the average sheets made with an 18 mm template gave rise to tens of thousands of structures, compared to one tenth of these structures in organoids. The latter deviates from previous ndings [11] where 500 structures were counted in organoids. Takasato et al. counted structures manually in a 2D image while our method relies on 3D-imaging techniques using the total volume of structures providing a more accurate indication of the number of glomerular structures. With these numbers of glomerular structures in nephron sheets, we hypothesize that transplanting multiple sheets could potentially partially restore kidney function in patients.
To further apply these cultures for research and clinical applications, cryopreservation of differentiating cells is of great interest. We show e cient cryopreservation of differentiating kidney cells which alleviates the need for continuous hiPSCs culture and differentiation, and gives rise to an accessible stock of cells. Cryopreservation will also allow for quality control screening on each batch of differentiating organoids. We show that nephron sheets displayed no difference in bright eld imaging and presence of kidney structures upon cryopreservation compared to continuously cultured sheets. Similarly, Mae et al. cryopreserved progenitors of induced ureteric bud organoids and found that they had the same potential to form nephric ducts epithelial aggregates as those who were not cryopreserved [21]. Other groups have shown that cryopreservation of hiPSC-derived cardiomyocytes and hiPSC-blood brain barrier microvascular endothelial cells had little to no adverse effect compared to differentiated cells that did not undergo cryopreservation [22,23]. These and our ndings show great promise for the use of cryopreservation and applicability of these differentiated cells. An additional advantage of the cryopreservation medium used in this study is that it is manufactured under cGMP conditions and allows for clinical translation to a GMP-compliant manufacturing protocol.

Conclusion
We have focused on developing an easily accessible, robust and e cient culture and cryopreservation method for creating large scale nephron sheets of high quality. These results constitute an important step to future application of these sheets as auxiliary tissue in the treatment of kidney disease in patients. While substantial challenges remain for safety, quality control and correct patterning of these structures, our ndings are an advancement in the eld of regenerative medicine.     Cryopreservation of hiPSC-derived nephron sheets A. Schematic of cryopreservation procedure. hiPSCs are differentiated until day 7 of differentiation, dissociated to single cells and cryopreserved. Cells are thawed and differentiation can be continued using both templates for hiPSC-derived nephron sheets.

Abbreviations
B. Bright eld images of nephron sheets without (control) and with cryopreservation (representative image of 4 independent experiments).