Kidney Organoids Generated from iPS Cells Obtained from the Erythroid Progenitors of Patients with Autosomal Dominant Polycystic Kidney Disease

Background. Kidney organoids have been broadly obtained from commercially available induced pluripotent stem cells (iPSCs); however, it has been a great challenge to eciently produce renal organoid models from patients with autosomal dominant polycystic kidney disease (ADPKD) that recapitulate both embryogenesis and the mechanisms of cystogenesis. Methods. Blood erythroid progenitors (EPs) from two ADPKD patients and one healthy control (HC) donor were reprogrammed by an episomal vector into iPSCs, which were differentiated into renal tubular organoids and then stimulated by forskolin to induce cysts formation. Results. iPSCs derived from EPs exhibited all characteristics of pluripotency and were able to differentiate into all three germ layers. 3D tubular organoids were generated from single cells after 28 days in Matrigel. HC and ADPKD organoids did not spontaneously form cysts, but upon forskolin stimulation, cysts-like structures were observed in the ADPKD organoids but not in the HC-derived organoids. Conclusion. The ndings of this study showed that kidney organoids were successfully generated from the blood EP cells of ADPKD patients and a healthy control donor. This approach should contribute as a powerful tool for embryonic kidney developing model which is able to recapitulate the very early pathophysiological mechanisms involved in cytogenesis. heterogeneous cell types biopsy in contrast to the erythroid progenitors, a more homogeneous cell population. these indicate that despite the variability and differences in the eciency of different iPSC clones to form kidney organoids, obtaining the ADPKD patients may constitute a unique model for studying the multiple forms of phenotypical manifestations of this disease. Moreover, it is important to note that if there is a considerable variability of mutations in the ADPKD-affected population, the patient-specic iPSCs, they might be used for individual regenerative therapies in the to allowing the development of an in vitro model to verify whether the very early phenotypical


Introduction
Autosomal dominant polycystic kidney disease (ADPKD) is the most frequent genetic cause of chronic kidney disease (CKD) worldwide, and it is one of the most prevalent inherited monogenic disorders, affecting a population estimated at 1:400-1000 1 . ADPKD is caused by mutations in the PKD1 and PKD2 genes, coding for polycystin-1 (PC1) and polycystin-2 (PC2), respectively, resulting in cilia alterations and cyst formation. Both proteins modulate numerous molecular pathways as well as tissue morphogenesis and function. However, the exact molecular mechanisms underlying cystogenesis remain unclear. It has been well accepted that a germline mutation followed by a somatic mutation (second-hit) in the normal allele is necessary for cystogenesis 2,3 . Moreover, based on results from orthologous mouse models, it has been demonstrated that an additional renal insult (third-hit) is required for rapid and severe cyst growth in mature kidneys [4][5][6] . It seems that in humans, most second hits occur during renal development, leading to cystic formation and growth even in utero. Therefore, the severity and the rate of progression of kidney disease resulting from PKD1 and PKD2 mutations may be in uenced by several factors, including the timing of inactivation of the genes, kidney developmental stage, environmental in uence, presence of concomitant renal lesions, and diversity of genetic mutations.
Although there has been innumerable data concerning the pathophysiological mechanisms of ADPKD collected from studies employing orthologous and nonorthologous mouse models, only a handful of studies focusing on human cellular models of cystogenesis are available 7 . In 2007, Takahashi et al 8 successfully reprogrammed somatic cells into pluripotent stem cells with embryonic properties, naming them "induced pluripotent stem cells" (iPSCs). Since then, many somatic cell sources have been reprogrammed into iPSCs from either animal or human tissues (hiPSCs) 8  Over the last few years, there have been signi cant advances in the generation of 3-dimensional organoids from iPSCs 8,11−13 . Primary kidney tubular epithelial organoids can be derived from kidney tissue, as well as from urine, and these organoids are an important tool for personalized disease modeling, as shown by Schutgens et al 14 . In subsequent studies, Freedman et al 15 were able to generate kidney organoids derived from a commercially available hiPS cell line, establishing a protocol to generate the cell types present in the metanephric mass that give rise to all structures of the nephron. Then, by knocking down the PKD1 or PKD2 genes, these investigators observed cyst formation from kidney tubules. In view of all these advanced and innovative technologies, we aimed to obtain iPSCs from the circulating erythroid progenitor cells of ADPKD patients and test the development of kidney organoids in vitro to recapitulate the steps involved in organogenesis, trying to evoke some pathophysiological events related to early cystogenesis.

Methods
Blood samples were collected from two unrelated female patients (PT1 and PT2) clinically diagnosed with ADPKD (PT1: 66 years old, eGFR 43.8 mL/min per 1.73 m2 and PT2: 56 years old, eGFR 60.2 mL/min per 1.73 m2), followed up at the Polycystic Kidney Disease Outpatient Clinic of the Nephrology Division of the Federal University of São Paulo (UNIFESP), and another sample was collected from one healthy female, which served as a control. The diagnosis of ADPKD was based on a positive family history and renal cysts according to the ultrasonographic diagnostic criteria by Pei et al 16 (family pedigrees are presented in supplementary Figure S1A). Both patients had cystic kidneys and livers (Magnetic Resonance Images are presented in supplementary Figure S1B). The study was reviewed and approved by the Ethics Advisory Committee of the University (Nr.1199.-0079-10/2018). After an interview to explain the purpose of the study, both patients and the healthy control signed the informed consent form.
Expansion of erythroid progenitor cells isolated from whole blood Blood samples were collected in a vacutainer tube containing EDTA (as an anticoagulant) at room temperature. Erythroid progenitor (EP) cells were then separated through a RosetteSep method (15216; Stem Cell Technologies), which removed unwanted cells, such as mature RBCs and platelets, while retaining the desired EP population. Due to the low concentration of EPs in whole blood, the cells were expanded and cultured in a speci c culture medium containing cytokines and supplements (X-vivo medium, Lonza). EP cells were identi ed by ow cytometry using the transferrin receptor (CD71) and glycophorin A (GlyA) as markers labeled with GFP (Suppl Figure S2).

Karyotyping
To verify whether the reprogramming process preserved chromosomal integrity, a karyotype analysis was carried out at the Genetic Department of UNIFESP. Cells were treated with 100 µl of colchicine (0.08 µg/ml) in 5 ml of iPS cell culture medium to induce cell cycle arrest, which was followed by the addition of a 0.075 M hypotonic KCl solution to induce nuclear swelling; cells were then xed with a mixture of methanol and acetic acid (3:1). Metaphase chromosomes were harvested, and Wright staining was used for cytogenic analysis of G-bands (Suppl Figure S3). iPSC colony characterization EP cells from the healthy control (HC) and the two ADPKD patients reprogrammed by episomal vectors formed typical iPSC colonies and presented a normal karyotype ( Figure S3). The iPSC colony characteristics were compared with those of an established control (H9 embryo cells) (Supp Figure S4).
As the iPSCs presented typical morphology, the pluripotency markers were identi ed by immuno uorescence. Cells were xed with 4% paraformaldehyde for 30 min at room temperature. After xing, samples were washed with PBS three times, blocked in 5% PSA and 0.3% Triton-X-100/DPBS, and incubated overnight with primary antibodies (Cell Signaling, Ma, EUA), namely, OCT4 (1:400 dilution) and NANOG (1:400 dilution), as a test of pluripotency. On the next day, a new slide was prepared, washed in PBS and incubated with an Alexa Fluor secondary antibody (1:200 dilution, Cell Signaling), and DAPI (1:1000, Invitrogen, MA, USA) was added for nuclear identi cation.
Capacity of iPSCs to differentiate into all three germ layers iPSCs were grown in 6-well plates coated with Matrigel (1:10, BD Biosciences, NY, EUA) in mTeSR1 culture medium (Stem Cell Technologies, BC, Canada). To prevent the iPSCs from differentiating, the colonies were passaged manually by mechanical fragmentation with a pipet tip under a microscope, or they were dissociated by enzymatic digestion by Gentle Cell (Life Technologies) (Supp Figure S5). Using these methods, iPSCs were maintained for 22 serial passages to ensure that cells properly incorporated the reprogramming factors. The iPSCs were then stimulated to differentiate into the three germ layers using a STEMdiff Trilineage Differentiation kit (Stem Cell Technologies). After ve days, cells were harvested for analysis of lineage-speci c markers of mesoderm and endoderm, and after seven days they were analyzed for ectoderm, according to the STEMdiff Trilineage Differentiation protocol (Supp Figure S6).

Molecular characterization of all three germ lines by qRT-PCR
Total RNA from cultures of the three germ layers was extracted using a GE RNA Spin Mini kit (GE Healthcare, USA), and it was converted to cDNA using a High Capacity Reverse Transcription Kit (Applied Biosystems, USA) according to the manufacturer's instructions. Quantitative real-time PCR (qRT-PCR) was carried out using a SYBR Green PCR Kit (Applied Biosystems, USA) according to the manufacturer's protocol. mRNA expression levels were calculated using the 2 ∆∆ Ct method. Hydroxymethylbilane synthase (HMBS) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were used as internal controls, and the expression of OCT4 was used as a quantitative control. The expression of SOX17 was used to identify differentiation into endoderm, CXCR4 used to identify differentiation into mesoderm, and PAX6 used to identify differentiation into ectoderm (Supp Figure S7). All primer sequences are speci ed in Table S1, and the e ciency curve was evaluated for each primer pair.
Immuno uorescence for the different germ layers iPSCs were induced to differentiate on coverslips in six-well plates. After differentiation, the coverslips were removed from the original wells and transferred to a new plate, where the cells were xed with 4% paraformaldehyde and then washed with PBS, as described above. Primary antibodies against OCT4

Kidney organoid differentiation from iPSCs
Renal organoids were generated following the protocol described by Freedman et al 7 with some adaptations. Brie y, 50,000 iPSCs from the HC and two ADPKD patients were used to produce suspensions of single cells with Accutase detachment solution (Stem Cell Technologies). Single cells were supplemented with 10 µM Rho-kinase inhibitor (Y27632, Stem Cell Technologies or Stemgent). After 16 h, the culture medium was replaced with 1 mL of mTeSR1 + 2.5% Matrigel + 10 µM Y27632. After 20 h, the medium was replaced with 1 mL of mTeSR1 only; 14 h later, the culture medium was replaced, and cells were induced with 6 µM CHIR99021 (AT104; Stem Cell or Stemgent GSK-3b inhibitor/Wnt pathway agonist) + Advanced RPMI (Thermo Fisher) + Glutamax (Life Technologies) + 10 µg/ml antibiotic (Ampicillin; 1:1000; Gibco). Finally, 36 h later, the culture medium was replaced with DRB medium containing Advanced RPMI (Thermo Fischer) + Glutamax (Thermo Fischer) + B27 Supplement (17504-044; Life Technologies) + antibiotic. The DRB medium was changed every three days for 28 days. This period (28 days) was enough for the organoids to be visible by microscopy. Molecular characterization of organoids by RT-PCR Semiquantitative gene expression was evaluated by RT-PCR during the embryogenic processes with total RNA being puri ed from the cell structures formed at days 0 (iPSCs) and at 14, 21, 28, and 35 days after differentiation induction. The expression of WT1 (developing kidney), AQP1 (proximal tube), and ECAD (cyst epithelium marker) was evaluated. GAPDH was used as an internal control. The expression of OCT4 was used as a quantitative standard.
Cyst induction Forskolin (10 µM; Sigma), an adenyl cyclase inhibitor, was used to induce cyst formation. Forskolin was added to adherent cultures on the 28th day of differentiation, since this period was coincident with the onset of tubular protein expression.

Obtaining iPSCs from erythroid progenitors (EPs)
The protocol employed to obtain kidney organoids from erythroid progenitors is summarized in Fig. 1A.
EPs expansion resulted in 1 × 10 6 cells after approximately 20 days ( Figure S2A), and approximately two months later, EPs were identi ed by their expression of the markers CD71 and GlyA. The e ciency of the separation method revealed that 38.6% of the cells were appropriately sorted, as shown by GFP ( uorescent green protein) expression; 65.1% of the cells were CD71 positive, and 70.9% were GlyA positive ( Figure S2B).
After verifying that the cells were EPs, they were transfected with reprogramming factors (Oct-4, Sox2, Lin28, L-Myc and klf4) and then were maintained in a speci c medium for reprogramming (ReproTeSR, Stem Cell Technologies) for approximately ve days. iPSCs developed typical cellular characteristics by forming colonies with a morphology distinct from that of non-reprogrammed cells present in the same cell culture. Cells were monitored until we observed typical iPSC-like colonies that were similar to the colonies of commercial H9 iPSCs (Suppl Figure S4). iPSC colonies were also recognized by their relatively larger size, tight cell packing, well-de ned homogeneous shape and regular edges. Identi cation was facilitated by the low degree of colony overgrowth from differentiated cells in feeder-free reprogramming medium (Fig. 1B). No detectable differences were found among iPSC colonies from the HC and ADPKD patients (PT1 and PT2) in terms of colony morphology and pluripotency markers, as shown in Fig. 1C.
iPSCs were further characterized by their ability to differentiate into the three germ layers. Typical images of the progression of germ layer development in cells from the HC are shown in supplementary Figure S6. Development of the three germ layers was veri ed by typical morphology (Fig. 2A) and by immuno uorescence labeling for differentiation markers, namely, FOXA2 for endoderm, smooth muscle actin (SMA) for mesoderm and SOX2 for ectoderm (Figs. 2B-D). Figure 2E shows the gene expression levels of speci c markers of endoderm (SOX17), mesoderm (CXCR4) and ectoderm (PAX6), as detected by RT-PCR.

Generation of kidney organoids from iPSCs
Three-dimensional renal structures were generated from iPSCs derived from the HC and both ADPKD patients following steps that recapitulate embryonic development. Figure 3 shows the steps that are necessary to obtain the respective embryonic structures. iPSC colonies (Fig. 3a) were dissociated with Accutase (3b) to obtain single iPSCs (3c). Considering the normal embryogenic process, the colonies would be expected to differentiate into a blastula and then a gastrula, but the spontaneous differentiation of iPSCs in culture is still a challenge since cells spontaneously undergo apoptosis 17 . To circumvent this undesired outcome, a ROCK inhibitor of signaling (10 µM Rho-kinase inhibitor Y27632) was added 16 h after the single cell stage. Inhibition of the ROCK pathway prevents cells from triggering the death pathways, allowing them to survive and adjust to proceed towards the differentiation process. The iPSCs then grew until reaching the epiblast stage, which is characterized by the presence of spheroid cavities (3d). Next, a potent GSK-3b inhibitor/Wnt pathway agonist (CHIR99021) was added, which was followed by the addition of DRB medium to induce epithelial-mesenchymal transition (3e-f) and nally mesenchyme-epithelial transition (3 g); these steps enabled epithelialization and sequential formation of pretubular aggregates and renal vesicles (3 h), which ultimately enabled the formation of renal tubular organoids (3i). All these in vitro embryogenesis steps were similar for cells derived from both HC and ADPKD patients.
The presence of tubular structures was veri ed by the expression of speci c markers by immuno uorescence: NHE3 and AQP1 for proximal tubules (Fig. 4A) and NPHS2 for glomeruli (Fig. 4B). In contrast, AQP2 labeling was not detected (Fig. 4B), con rming that this methodology allowed only the development of the metanephric mass but not of the ureteric bud. The marker ECAD was not detected in the absence of forskolin (4B).
To verify the presence of speci c markers expressed throughout the process of progenitor kidney development, all samples (HC, PT1 and PT2) were collected at ve points of the whole process of differentiation: day 0 (iPSCs), days 14, 21, 28 and then at day 35 (7 days after cyst induction). The time course of gene expression revealed differentiation during the embryogenesis process, as shown in Fig. 5A-C. The pluripotency marker OCT4 was expressed by iPSCs (day 0), but its levels were reduced over 14 days, reaching very low levels following differentiation. In contrast, the expression of the speci c markers of metanephric mesenchyme (WT1) and proximal tubule (AQP1) progressively increased when the differentiation process was initiated by day 14, reaching a plateau by day 21. Figure 5D shows the expression of ECAD, a cystogenesis marker, at day 35, i.e. 7 days after cyst induction by forskolin. In the absence of forskolin, the expression of ECAD was very low, even in ADPKD patients. However, when forskolin was added (day 28), both patients exhibited increased ECAD expression; but the same pattern was not observed in HC organoid which remained with low levels of ECAD expression even in the presence of forskolin.
The presence of cysts formed from the organoids stimulated with forskolin was evident for ADPKD patients, as shown in Fig. 6. The upper panel shows images of the organoids at day 28, before the addition of forskolin. (Fig. 6Aa). Figures 6Ab (panoramic view-4X) and 6Ac (20X viewing) show the organoids at day 35 in the presence of forskolin showing the presence of cyst-like structures in samples from both patients (PT1 and PT2) but not in the HC sample. The presence of cysts was also veri ed by immuno uorescence staining (Fig. 6B). ECAD labeling was not detected in the HC organoid, but it was clearly observed in the organoids from both patients. Interestingly, the presence of a cyst lumen can also be observed.

Discussion
In the present study, we described in detail a procedure to generate kidney organoids from human iPSCs obtained from the blood erythroid progenitor cells of patients with ADPKD. Many somatic cell types have been successfully reprogrammed into iPSCs, including blood cells 8 . Here, we opted to use blood erythroid progenitors as the starting cell type for a number of reasons, such as their genomic integrity, epigenetic memory and e cient reprogramming 12,18,19 . The e ciency by which erythroid progenitors had been converted into iPSCs was as much as two orders of magnitude better (7%-28%) than the e ciency of differentiated blood cell types (0.02%-0.60%) or broblasts (0.74%) 5 . Despite the low abundance of these cells in the peripheral blood, they can easily be isolated and expanded in vitro to produce a su cient number for reprogramming. One of the most important challenges in establishing suitable iPSCs is their genotype and phenotypic variability 7 , but there are some technical strategies proposed to circumvent or reduce such variability. In the present experiments, we used a non-integrating episomal system for reprogramming that does not require genome integration for the reprogramming genes to be expressed 11 , allowing less immunogenicity and rendering more genetically stable cells 20,21 . Moreover, all samples were derived from the same source, which helps minimize epigenetic variability 13 , assuming that differentiation of the starting cell might in uence the e ciency of the end results. Analysis of the pluripotency markers and the ability of iPSCs to differentiate into all three germ layers were compared side-by-side, and they were also compared to the H9 embryonic cell line, which was used as a positive control. All these comparisons showed no signi cant differences between ADPKD patients, HCs and H9 cells.
Several protocols have been developed for inducing the differentiation of iPSCs into kidney organoids 22,23 . Based on protocols described by Cruz et al 7 , we obtained structures with the characteristics of proximal tubules, including the expression of NHE3 and AQP1. In addition, the organoids expressed NPHS2, indicating the presence of podocyte-like cells. Conversely, AQP2 was not expressed in these tubular organoids. This nding is in agreement with previous studies pointing to several limitations in developing structures derived from the ureteric bud (UB) in vitro, with the development of UB requiring speci c protocols, since the timing window for the addition of the speci c inductor is very limited 7,24,25 .
In the process of generating kidney organoids from iPSCs, all of renal embryogenesis is recapitulated in vitro, initially involving the formation of cavitated spheroids from single cells 15 . Indeed, we found that undifferentiated iPSCs from both ADPKD patients and healthy control formed epiblast-like spheroid structures that were morphologically similar. Although DNA sequencing of the PKD1 and PKD2 genes was not performed in the present study, the current ndings indicate that the type of mutation carried by the patients had no in uence on typical epiblast morphogenesis. Further studies are still needed to more properly evaluate the expression of pivotal molecules involved in this phase of embryonic development.
According to the protocol developed by Freedman et al 15, we observed that spheroids did differentiate into tubular organoids 26 days after induction of differentiation, and the timing was similar for both ADPKD patients and HC (Fig. 3). Nevertheless, we noticed some variability regarding the timetable of expression and the levels of differentiation markers. On the other hand, the morphology of the tubular structures was similar and comparable among samples from the ADPKD patients and HC. In contrast, using iPSCs originating from renal biopsies of ADPKD patients, Freedman's group found a higher variability in the ability of these cells to form organoids 7 . In addition to the huge diversity of mutations inherent to this disease, these differences in the e ciency of obtaining organoids may also be explained by the heterogeneous cell types present in biopsy samples, in contrast to the erythroid progenitors, which constitute a more homogeneous cell population. Taken together, these results indicate that despite the variability and differences in the e ciency of different iPSC clones to form kidney organoids, obtaining the latter from ADPKD patients may constitute a unique model for studying the multiple forms of phenotypical manifestations of this disease. Moreover, it is important to note that if there is a considerable variability of mutations in the ADPKD-affected population, the patient-speci c iPSCs, being totally immune compatible, they might be used for individual regenerative therapies in the future, in addition to allowing the development of an in vitro model to verify whether the very early phenotypical manifestations are also as diverse as the clinical presentation.
Unstimulated organoids derived from HC or ADPKD patients did not spontaneously form cysts. The typical morphology of cysts and the expression of cystogenesis markers were not detectable in any sample, even extending the organoid differentiation in culture until completing 35 days. It has been demonstrated that cell proliferation alone does not produce cysts 26 . Moreover, a third independent event (known as third-hit) is required to accelerate cyst formation and growth in vivo 25 . Conversely, cyst formation induced by forskolin did occur in the organoids from ADPKD patients but not in those from HC.
Of note, although present, cysts were observed at a low percentage in the ADPKD samples even in the presence of forskolin, as has also been observed by Freedman et al 15 in organoids formed from the CRISPR/Cas9 model of ADPKD disease. However, when adherent organoids were transferred to a 3D suspension in their model, cystogenesis increased 10-fold 7 , indicating that the microenvironment is directly implicated in cystogenesis. The low percentage of cysts in the ADPKD organoids in the current study could be accounted for by the use of adherent organoids.
Aside from the role of the microenvironment in enabling e cient cyst formation in vitro, the experimental protocol employed in the present study formed proximal tubule organoids but not more distal portions of the nephron 25 . Notably, it is well established that in vivo, the proximal tubule does not represent the main site of cyst formation, corresponding to approximately 1.8% 27 , whereas the distal nephron (particularly the collecting duct) is the main site of cyst origin.

Conclusion
In conclusion, the present ndings showed that it is feasible to generate organoids with three-dimensional structures similar to the kidney progenitor architecture using iPSCs reprogrammed from erythroid progenitors that were easily isolated from ADPKD patients and that these organoids are able to form cysts under forskolin stimulation. This promising methodology unveils a method for increasing understanding of the very early genetic, phenotypical and physiological steps in cyst formation in the ADPKD setting.   indicates pluripotency of iPSCs, SOX17 was a marker for endoderm, CXCR4 was a marker for mesoderm and PAX6 was a marker for ectoderm. OCT4 expression by iPSCs was used as an internal quantitative control. OCT4 expression was undetectable in cells of the three germ layers.

Figure 3
Generation of tubular organoids. Phase contrast images of the generation of epithelial tubular kidney organoids at 28 days from iPSCs generated from the HC, PT1 and PT2 groups. In the rst 36 h, the iPSCs were maintained in 3D culture; then, they were dissociated, transformed into single cells and treated with Y-27632, which allowed cellular reorganization and survival, producing epiblast spheroids (d). Then, CHIR99201 and DRB were added, activating the Wnt/β-catenin pathway and allowing differentiation into tubular organoids (i). The morphology among spheroids from ADPKD patients and a HC donor is similar. Kidney organoids. Representative immunostaining of epithelial tubular kidney organoids at 28 days for HC, PT1 and PT2 patients. Typical morphology of organoid structures ( rst columns). Cell nuclei were stained with DAPI (second columns A-B). A, each structure was characterized by immuno uorescence staining using speci c antibodies against NHE3 (Abcam 1:500) and AQP1 (Sigma 1:500). B, NPHS2 (Abcam 1:100) was detected. AQP2 (Sigma 1:100) was not detectable, proving that only the metanephric mass of the structures was performed. ECAD (Sigma 1:100) was not detectable in the absence of forskolin. A-B: Images taken at 20X; scale bar 1000 μm.

Figure 5
Relative gene expression levels of renal structures. A-C. Sequential expression of pluripotency and organogenesis markers. Day 0 (D0) refers to iPSCs and OCT4 expression in D0 was used as a quantitative standard for each sample. In this set of experiments, forskolin was added at day 28. D. Expression levels ECAD in the organoids at day 35 in the presence and absence of forskolin (added at day 28). GAPDH was used as an internal control and OCT4 expression by iPSCs from healthy donor (HD) was used as the quantitative standard. WT1 and AQP1 were used as markers of renal structures, and ECAD as the cyst marker.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.