The centrosomal protein 83 (CEP83) regulates human pluripotent stem cell differentiation toward the kidney lineage

During embryonic development, the mesoderm undergoes patterning into diverse lineages including axial, paraxial, and lateral plate mesoderm (LPM). Within the LPM, the so-called intermediate mesoderm (IM) forms kidney and urogenital tract progenitor cells, while the remaining LPM forms cardiovascular, hematopoietic, mesothelial, and additional progenitor cells. The signals that regulate these early lineage decisions are incompletely understood. Here, we found that the centrosomal protein 83 (CEP83), a centriolar component necessary for primary cilia formation and mutated in pediatric kidney disease, influences the differentiation of human-induced pluripotent stem cells (hiPSCs) toward IM. We induced inactivating deletions of CEP83 in hiPSCs and applied a 7-day in vitro protocol of IM kidney progenitor differentiation, based on timed application of WNT and FGF agonists. We characterized induced mesodermal cell populations using single-cell and bulk transcriptomics and tested their ability to form kidney structures in subsequent organoid culture. While hiPSCs with homozygous CEP83 inactivation were normal regarding morphology and transcriptome, their induced differentiation into IM progenitor cells was perturbed. Mesodermal cells induced after 7 days of monolayer culture of CEP83-deficient hiPCS exhibited absent or elongated primary cilia, displayed decreased expression of critical IM genes (PAX8, EYA1, HOXB7), and an aberrant induction of LPM markers (e.g. FOXF1, FOXF2, FENDRR, HAND1, HAND2). Upon subsequent organoid culture, wildtype cells differentiated to form kidney tubules and glomerular-like structures, whereas CEP83-deficient cells failed to generate kidney cell types, instead upregulating cardiomyocyte, vascular, and more general LPM progenitor markers. Our data suggest that CEP83 regulates the balance of IM and LPM formation from human pluripotent stem cells, identifying a potential link between centriolar or ciliary function and mesodermal lineage induction.


Introduction
During mammalian embryonic development, the mesoderm forms axial, paraxial, and lateral plate domains that harbor precursor cells for distinct organ systems. Forming as a major part of the lateral plate mesoderm (LPM), the intermediate mesoderm (IM) harbors progenitor cells of all kidney epithelial cells (Davidson et al., 2019), whereas the remaining LPM contributes progenitors of various cell types, including cells of the cardiovascular system (Prummel et al., 2020). The molecular and cellular mechanisms that drive induction of the IM and distinct LPM domains during embryonic development are not fully understood.
The centrosomal protein 83 (CEP83) is a component of distal appendages (DAPs) of centrioles. DAPs are involved in the anchoring of the mother centriole to the cell membrane, an early and critical step in ciliogenesis (Lo et al., 2019;Tanos et al., 2013;Yang et al., 2018;Kurtulmus et al., 2018;Wheway et al., 2015;Bowler et al., 2019;Failler et al., 2014;Shao et al., 2020;Mansour et al., 2021). CEP83 recruits other DAP components to the ciliary base, and loss of CEP83 disrupts ciliogenesis (Tanos et al., 2013). In radial glial progenitors, removal of CEP83 disrupts DAP assembly and impairs the anchoring of the centrosome to the apical membrane as well as primary ciliogenesis (Yang et al., 2018;Shao et al., 2020). Mutations of CEP83 in humans have been associated with infantile nephronophthisis (Failler et al., 2014), an early onset kidney disease that results in end-stage renal disease before the age of 3 years (Hildebrandt, 2004;Luo and Tao, 2018) and additional organ anomalies (Failler et al., 2014). To date, how the loss of CEP83 function contributes to aberrant kidney development remains unclear.
Human-induced pluripotent stem cells (hiPSCs) provide useful tools to study molecular mechanisms of cellular differentiation. Protocols for the induction of kidney organoids from iPSC have been successfully developed (Takahashi et al., 2007;Morizane et al., 2015;Taguchi et al., 2014;Takasato et al., 2015;Freedman et al., 2015;Kumar et al., 2019). The protocol by Takasato et al. uses stepwise exposure of iPSC to WNT and FGF agonists in a monolayer culture system for a 7-day period, which results in the induction of cells with a transcriptional phenotype resembling kidney progenitors in the IM . Transfer of these cells to an organoid culture system followed by another series of WNT and FGF signals results in the differentiation of three-dimensional (3D) kidney organoids composed of different kidney cell types, including glomerular and tubular cells. Genome editing studies have previously been used to study the effects of genetic defects associated with kidney diseases on kidney differentiation in human iPSC systems Tan et al., 2018;Boyle et al., 2008;Kobayashi et al., 2008;Howden et al., 2019;Kuraoka et al., 2020). Here, we studied the effect of an induced knockout of CEP83 in human iPSCs on kidney organoid differentiation. We uncovered a novel role of CEP83 in determining the balance of IM versus LPM differentiation, implicating a centrosomal protein in early mesodermal lineage decisions.

CEP83 is essential for the differentiation of hiPSCs into kidney cells
To investigate the effect of CEP83 loss on the differentiation of hiPSCs into IM kidney progenitors, we applied CRISPR-Cas9 technology to induce a null mutation in the CEP83 gene in hiPSCs ( Figure 1A). Three hiPSCs clones designated CEP83 −/− (KO1, KO2, and KO3) carried deletions within CEP83 exon 7, each of which led to an induction of a premature stop codon resulting in a predicted truncated protein ( Figure 1B-D and Figure 1-figure supplement 1A). These clones exhibited a complete loss of CEP83 protein by immunoblotting ( Figure 1E). Three wildtype clones were derived as controls (WT1, WT2, and WT3). All six clones were morphologically indistinguishable (by brightfield microscopy) and had similar overall gene expression profiles (by bulk RNA-seq and qRT-PCR), including pluripotency and lineage marker expression (Figure 1-figure supplement 1B, C, and Figure 1- figure supplement 2A, B). In KO clones, the anticipated altered transcripts of CEP83 were detectable based on bulk RNA-seq (data not shown). Single nucleotide polymorphism -analysis confirmed identical karyotypes of all six clones (Figure 1-figure supplement 2C).
Together, these findings confirmed the successful deletion of CEP83 in iPSCs without any overt direct cellular phenotypic consequences. We applied a 7-day monolayer protocol using timed application of WNT and FGF agonists as reported by Takasato et al., 2015 to differentiate WT and KO hiPSCs into IM kidney progenitors (Takahashi et al., 2007;Morizane et al., 2015;Taguchi et al., 2014; Figure 2A). Schematic of the experimental approach to induce a deleting mutation in exon 7 of the CEP83 gene (as described in the methods section). (B) DNA extracted from pooled transfected cells was subjected to PCR, targeting the predicted deletion site in the CEP83 gene. In addition to the 182 bp fragment present in untransfected wildtype (WT) cells, an approximately 120 bp fragment was detected in transfected cells, corresponding to the induced deletion in exon 7. (C) Three clones (CEP83 −/− clones KO1, KO2, KO3) carried 62-74 bp deletions within CEP83 exon 7, which led to an induction of premature stop codons or frameshift mutation on both alleles of CEP83. Three wildtype clones (WT1, WT2, and WT3) were used as controls. (D) Quantitative RT-PCR for a fragment corresponding to the deleted region in CEP83 exon 7 produced a detectable signal in RNA extracts from WT clones but not CEP83 −/− clones. (E) Immunoblotting of WT and CEP83 −/− clones using a CEP83 antibody targeting the C-terminal region of the protein (see Methods for details) indicated a complete loss of the 83 KD band corresponding to CEP83 protein in the three KO clones compared with the three WT clones. Data are mean ± SD.*p<0.05 and **p<0.01 vs. WT. See The online version of this article includes the following source data and figure supplement(s) for figure 1: Source data 1. The file contains detailed original PCR gels and immunoblots.
Source data 2. Excel sheet shows RT-qPCR data for mRNA expression of CEP83 in WT and knockout hiPSCs.
Source data 3. File contains uncropped PCR gels and immunoblots.   After 7 days of culture (D7), WT and KO cells exhibited an indistinguishable morphology by bright field microscopy ( Figure 2B and C). Immunostaining for acetylated tubulin, however, indicated abnormal primary cilia formation in CEP83-deficient cells ( Figure 2D and E). The number of ciliated cells was reduced from approximately 30% (in WT clones) to less than 10% (in KO clones) ( Figure 2F). Among ciliated cells, the length of cilia was increased from 2 to 5 µm (in WT clones) to 5-13 µm (in KO clones) ( Figure 2G). This indicated that CEP83 −/− hiPSCs differentiated toward IM progenitors exhibited ciliary abnormalities. To analyze the induced IM kidney progenitor cells functionally, we collected D7 WT and CEP83 −/− cells and placed them into an organoid culture system again applying timed WNT and FGF agonists to foster differentiation of mature kidney cell types, as previously reported  Figure 2A). Organoids harvested from WT clones after a total of 25 days of culture (D25) had formed patterned kidney epithelial-like structures, including NPHS1-positive glomerulus-like structures, Lotus tetragonolobus lectin-positive proximal tubule-like, and E-cadherin (CDH1)-positive distal tubule-like structures ( Figure 3A, C and E, and Kidney epithelial-like structures formed only in WT but not in CEP83 −/− organoids ( Figure 3G). Similar to the findings in day 7 cells reported above, primary cilia were found in fewer cells of CEP83 −/− organoids (<5% of cells) and were abnormally elongated ( Figure 2-figure supplement 1).
Next, bulk RNA sequencing of WT (WT1, WT2, WT3) and CEP83 −/− (KO1, KO2, KO3) organoids was carried out to evaluate differential gene expression on a genome-wide level, and RT-PCR was used to validate selected genes. Hierarchical clustering of the samples indicated strong gene expression differences between WT and CEP83 −/− samples ( Figure 3-figure supplement 2). Several genes associated with kidney development and kidney epithelial differentiation were differentially expressed with high expression in WT organoids but showed comparatively low or absent expression in CEP83 −/− organoids: these genes included kidney-specific lineage genes (PAX2, PAX8), and lineage/differentiation markers of glomerular cells (NPHS1, PODXL, WT1, PTPRO), proximal (HNF1B, LRP2, CUBN), and distal (EMX2, MAL2, EPCAM, GATA3) kidney epithelial cells ( Figure 3H-L, Figure 3-figure supplement 2). This indicated that CEP83 −/− IM progenitors failed to differentiate into kidney cells, suggesting that CEP83 function is necessary to complete essential steps in the process of differentiation from pluripotent stem cells to kidney cells.

CEP83 deficiency associates with molecular defects of nephron progenitor cells
We next aimed to gain molecular insights into the lineage impact of CEP83 deficiency during the course of kidney epithelial differentiation. Since no global transcriptomic differences were detectable between WT and CEP83 −/− hiPSCs prior to differentiation (see above), we focused on mesodermal cell stages induced at D7, which displayed mild overall gene expression differences between WT-and CEP83-deficient cells as detected by bulk RNA sequencing (Figure 4-figure supplement 1).
A marked upregulation of nephron progenitor marker genes (GATA3, HOXB7, HOXD11, EYA1) (Bilous et al., 1992;Grote et al., 2006;Kress et al., 1990;Srinivas et al., 1999;Wellik et al., 2002;Mugford et al., 2008a;Ruf et al., 2004;Sajithlal et al., 2005) was observed in both WT and CEP83 −/− cells at day 7 ( Figure 4-figure supplement 2), suggesting that the differentiation path of pluripotent CEP83 −/− cells to IM nephron progenitors was largely intact. To understand the potential molecular defects at the IM stage in more detail, we performed single-cell RNA (scRNA) sequencing on D7 WT and CEP83 −/− cells (representing two different hiPSC clones for each condition differentiated in two separate experiments). We obtained transcriptomes from 27,328 cells, representing clones WT1     Figure 4A). We combined all cells and generated a uniform manifold approximation and projection plot uncovering 10 different cell states/clusters (0-9; Figure 4B). We identified marker genes for each cluster ( Figure 4C), indicating that clusters 1, 3, and 4 represented kidney progenitors/nascent nephrons (expressing, e.g., PAX8, EYA1, HOXB7) in different phases of the cell cycle. Other clusters represented as-of-yet uncharacterized cell types, which were consistent with previous single-cell transcriptome analyses from iPSC-derived cells induced by the same induction protocol (Subramanian et al., 2019;Low et al., 2019). Each of the four samples (WT1, WT2, KO1, and KO2) contributed to each cluster ( Figure 4D). However, one cluster representing damaged cells (cluster 5) was observed at numerically higher percentages in KO cells compared to WT cells.
Cluster 5 cells expressed high levels of mitochondrial RNAs, and staining for active caspase 3 demonstrated an increased percentage of apoptotic cells in KO samples compared to controls (Figure 4figure supplement 3). We focused on kidney progenitors (clusters 1, 3, and 4) and found that a numerically lower percentage of KO cells (11.9 and 12.5% in KO clones) contributed to cluster 1 when compared with WT cells (25.9 and 36.3% in WT clones) ( Figure 5A). In contrast, similar percentages of WT and KO cells were represented in kidney progenitor clusters 3 and 4 ( Figure 5B and C). Differential gene expression analysis in these three clusters indicated significantly lower expression of kidney progenitor markers PAX8, EYA1, CITED1, and HOXB7 in KO cells from clusters 1, 3, and 4 when compared to WT cells ( Figure 5D, E and F; Figure 5-figure supplements 1-2). Interestingly, scRNA sequencing data also showed downregulated expression of genes encoding ciliary proteins, including OFD1, PCM1, and RAB11A (Ferrante et al., 2006;Dammermann and Merdes, 2002;Knödler et al., 2010; Figure 5-figure supplement 3), consistent with the ciliogenesis defects in CEP83 knockout cells. These results indicate that CEP83 deficiency remained permissive with initial kidney progenitor induction, but that these cells exhibited mild molecular defects detectable by differential expression of kidney progenitor genes, which potentially contributed to the failure of CEP83-deficient cells to further differentiate toward mature kidney cell types.

CEP83 deficiency promotes ectopic induction of lateral plate mesoderm-like cells followed by an expansion of cardiac and vascular progenitors
We next inspected single-cell transcriptomes and bulk RNA sequencing data from D7 cells for genes that were upregulated in CEP83 −/− cells compared to WT cells. From this analysis, we observed a consistent upregulation of genes that are normally expressed in early LPM, including OSR1, FOXF1, The online version of this article includes the following source data and figure supplement(s) for figure 3: Source data 1. The data shows the quantitative analysis of the percent of nephron formation per organoid in knockout organoids versus WT organoids.
Within each cell, we analyzed the expression of LPM markers (FOXF1, HAND1, HAND2, and CXCL12) and of more restricted IM markers (PAX8, EYA1, and HOXB7) ( Figure 6-figure supplement 1). This analysis indicated that WT cells of these clusters exhibited an IM-like phenotype, while KO cells were shifted toward an LPM-like phenotype. The common IM/LPM marker OSR1 was expressed at higher level in KO cells comparing to the WT cells.
Taken together, these observations document that hiPSCs deficient in CEP83 respond to an in vitro differentiation program toward kidney progenitors, yet diverge toward a broader LPM progenitor composition without significant IM instead.

Discussion
This study indicates a novel contribution of CEP83 in regulating the differentiation path from human pluripotent stem cells to kidney progenitors. We pinpoint a stage at day 7 of IM induction where CEP83 loss of function results in a decreased nephron progenitor pool with downregulation of critical kidney progenitor genes (PAX8, EYA1, HOXB7). At the same stage, genes typical of LPM specification (including FOXF1, FOXF2, FENDRR, HAND1, and HAND2) are upregulated (Figure 8). Functionally,         Organoids derived from CEP83-deficient cells fail to induce any detectable nephron structures, suggesting a novel role for CEP83 during the specification of functional kidney progenitors in the mesoderm.
Our findings are relevant to understanding the cellular and molecular functions of CEP83 and might be relevant to the pathophysiology of human genetic diseases. To date, 11 patients with biallelic mutations of CEP83 have been reported, 8 of which displayed kidney phenotypes (Failler et al., 2014;Veldman et al., 2021;Haer-Wigman et al., 2017). Available kidney histologies identified microcystic tubular dilatations, tubular atrophy, thickened basement membranes, and interstitial fibrosis.
Extrarenal phenotypes included speech delay, intellectual disability, hydrocephalus, strabismus, retinal degeneration, retinitis pigmentosa, hepatic cytolysis, cholestasis, and portal septal fibrosis with mild thickening of arterial walls and an increase in the number of the biliary canalicules on liver biopsy. Among individuals with CEP83 mutations, all but one carried at least one missense mutation or short in-frame deletion, suggesting that CEP83 function may have been partially preserved. One individual with presumed full loss of CEP83 displayed a more severe phenotype with multiple organ dysfunction. It will be interesting to await future reports of additional CEP83 mutations in humans and whether complete loss of function alleles will result in broader mesoderm defects or renal agenesis. In this regard, it is interesting that mice with a targeted homozygous loss-of-function mutation of their CEP83 ortholog (Cep83 tm1.1(KOMP)Vlcg ) display midembryonic lethality (at E12.5) with evidence of severe developmental delay as early as E9.5 (https://www.mousephenotype.org/data/genes/MGI: 1924298). These phenotypes are potentially consistent with the role of CEP83 in germ layer patterning and mesoderm development, but a more detailed phenotypical characterization of Cep83 knockout embryos would be required to substantiate this possibility.
The precise molecular and cellular mechanisms underlying our observations remain to be determined. CEP83 is a protein that is necessary for the assembly of DAPs and primary cilia formation in several cell types (Tanos et al., 2013;Yang et al., 2018;Shao et al., 2020;Kumar et al., 2021;Stinchcombe et al., 2015;Joo et al., 2013). A potential involvement of CEP83-mediated primary cilia formation in the findings reported here is suggested by obvious ciliary defects in CEP83-deficient cells at the D7 and the organoid stage ( Figure 2D-G, Figure 2-figure supplement 1). These defects include reduced percentages of ciliated cells and elongated primary cilia in those cells that continue to form a primary cilium. In addition, CEP83-deficient cells displayed downregulated expression of several transcripts encoding ciliary components ( Figure 5-figure supplement 3) and evidence of an activation of several hedgehog signaling associated genes. The primary cilium is critically involved in hedgehog signaling (Ho and Stearns, 2021). Moreover, Hedgehog signaling is important for mesodermal lineage decisions during gastrulation (Guzzetta et al., 2020). This raises the possibility that CEP83 controls mesodermal cell fate decisions by modulating hedgehog signaling in the mesoderm. Nevertheless, additional studies will be necessary to address this possibility. In addition, it remains unknown whether abnormal cilia formation in CEP83-deficient cells causally contributes to the cell fate phenotype.
We observed downregulated expression of the key nephron progenitor genes PAX8, EYA1, and HOXB7 in CEP83 −/− cells at day 7, which might explain their failure to differentiate into kidney cells since each of these genes is essential for normal kidney development (Vincent et al., 1997;Pfeffer *p<0.05, **p<0.01, ***p<0.001, and ****p<0.0001. ns = not significant. See Figure 6-figure supplements 1-3. See also Figure 6-source data 1. Check data availability section for other source data. The online version of this article includes the following source data and figure supplement(s) for figure 6: Source data 1. The sheet shows the plotted TPM values of mRNA sequencing analysis in Figure 6 for the expression of lateral plate mesoderm marker genes between the KO and WT at days 0, 7, and 25 of the differentiation.     Kumar et al., 1998;Bouchard et al., 2002;Xu et al., 1999;Patterson and Potter, 2004;Rojek et al., 2006). Furthermore, inductive signals from HOXB7-positive ureteric bud are known to maintain viability of nephron progenitor cells in the IM (Barasch et al., 1997), which may contribute to increased numbers of apoptotic cells we observed in CEP83 −/− cells. Defects during nephron progenitor differentiation in the IM would be expected to result in severe kidney phenotypes such as renal agenesis or renal hypodysplasia. Defects of centriolar components or cilia have previously been linked to such phenotypes: in mice, centrosome amplification, i.e., the formation of excess centrosomes per cell severely disrupts kidney development, resulting in depletion of renal progenitors and renal hypoplasia (Dionne et al., 2018). In humans, loss of KIF14, a protein necessary for proper DAP assembly and cilium formation, has been associated with kidney malformations, including The online version of this article includes the following source data for figure 7: Source data 1. The sheet shows the plotted TPM values of mRNA sequencing analysis in Figure 7 for the expression of cardiomyocytes and vascular progenitors marker genes between the KO and WT at days 0, 7, and 25 of the differentiation. renal agenesis and renal dysplasia (Filges et al., 2014;Reilly et al., 2019;Pejskova et al., 2020). Furthermore, Kif3a, a ciliary protein involved in intraflagellar transport, is necessary for normal mesoderm formation, and kidney progenitor-specific defects of Kif3a have been associated with reduced nephron numbers (Takeda et al., 1999;Chi et al., 2013). Similarly, mouse genes encoding the ciliary intraflagellar transport proteins IFT25 and IFT27 have been associated with renal agenesis or renal hypoplasia (Desai et al., 2018;Quélin et al., 2018). Together, these studies highlight the importance of molecules involved in ciliogenesis for mesoderm and kidney progenitor development and suggest that CEP83 contributes to such processes by facilitating an early step of ciliogenesis. Nevertheless, the detailed molecular processes that link CEP83 function, cilia formation, and kidney progenitor specification remain to be determined.
The finding of various upregulated LPM markers in CEP83 −/− cells starting from day 7 suggests that CEP83 function may be involved in fine-tuning the balance of LPM and IM, thereby contributing to lineage decisions during mesoderm formation. Crosstalk of LPM and IM has been reported previously in zebrafish, overexpression of LPM transcription factors Scl/Tal1 and Lmo2 induces ectopic vessel and blood specification while inhibiting IM formation (Gering et al., 2003). Furthermore, the LPM transcription factor Hand2 is critical in determining the size of the IM, while natively expressed in the IM-adjacent LPM progenitors that form mesothelia (Barbosa et al., 2007Perens et al., 2016Prummel et al., 2022). Loss of Hand2 in zebrafish results in an expanded IM, whereas Hand2 overexpression reduces or abolishes the IM. Interestingly, HAND2 was among the most strongly induced transcripts in our CEP83 −/− cells at day 7; connecting with the developmental role of Hand2 in IM formation, these observations suggest that HAND2 expression in CEP83-deficient cells may have contributed to the reduced numbers of nephron progenitor cells at this stage. Of note, CEP83deficient cells at D25 expressed increased levels of LPM genes expressed in mesothelial (including OSR1, CXCL12, HAND1/2), cardiopharyngeal (including ISL1, TBX1), and endothelial/hematopoietic (including TAL1, LMO2, GATA2) progenitors (Prummel et al., 2020;Prummel et al., 2022). In sum, we propose a novel role for CEP83 in regulating the development of IM nephron progenitors, which may involve direct effects of CEP83 in the nephron progenitor differentiation program and indirect LPM-mediated effects on the IM. Future studies are warranted to delineate the molecular and cellular mechanisms underlying CEP83 function in LPM and specifically IM patterning.

Materials and methods hiPSCs cell line
We used the human iPSC cell line BIHi005-A, which was generated by the Berlin Institute of Health (BIH). The hiPSCs were maintained in six-well plates (Corning, 353046) coated with Matrigel (Corning, 354277) and cultured in Essential 8 medium (E8, A1517001, Gibco-Thermo Fisher Scientific) supplemented with 10 µM Y-27632 (Rocki,Wako,. Cells were authenticated and tested for the mycoplasma infection.
For PCR genotyping, we isolated genomic DNA from the pool of transfected cells followed by PCR using Phire Tissue Direct PCR Master Mix (Thermo Scientific, F170S) according to the manufacturer's instructions ( Figure 1B). After confirming the editing efficiency in the pool, we generated single-cell clones by the clonal dilution method. We plated 500 single cells per well of a 6 well plate and picked 24 clones using a picking hood S1 (Max Delbrück Centre Stem Cell Core Facility). Then, clones were screened for homozygous deletions of CEP83 by PCR using Phire Tissue Direct PCR Master Mix. Selected knockout clones were further characterized for CEP83 loss of function on the DNA, RNA, and protein level. CEP83 −/− clones (KO1, KO2, and KO3) were registered as (BIHi005-A-71, BIHi005-A-72, and BIHi005-A-73) in the European Human Pluripotent Stem Cell Registry (https://hpscreg.eu).

Single nucleotide polymorphism -karyotype
To assess karyotype integrity, copy number variation (CNV) analysis on the human Illumina OMNI-EXPRESS-8v1.6 BeadChip was used. In brief, genomic DNA was isolated from three WT (WT1, WT2, and WT3) and three KO (KO1, KO2, and KO3) clones using the DNeasy blood and tissue kit (Qiagen, Valencia, CA, United States), hybridized to the human Illumina OMNI-EXPRESS-8v1.6 BeadChip (Illumina), stained, and scanned using the Illumina iScan system according to a standard protocol (LaFramboise, 2009;Arsham et al., 2017;Haraksingh et al., 2017). The genotyping was initially investigated using the GenomeStudio 1 genotyping module (Illumina). Following that, KaryoStudio 1.3 (Illumina) was used to perform automatic normalization and identify genomic aberrations in detected regions by generating B-allele frequency and smoothed Log R ratio plots. To detect CNVs, the stringency parameters were set to 75 kb (loss), 100 kb (gain), and CN-LOH (loss of heterozygosity). KaryoStudio generates reports and displays chromosome, length, list of cytobands, and genes in CNV-affected regions.

DNA isolation and PCR
DNA was isolated from cells using DNeasy Blood & Tissue Kits (Qiagen,69504). CEP83 primers were designed using Primer3 webtool (Supplementary file 1). PCR was done using Phusion high-fidelity DNA polymerase (Biolabs, New England, M0530) according to the manufacturer's instructions. PCR results were visualized on 1.5% agarose gel using a BioDoc Analyze dark hood and software system (Biometra).

RNA isolation, RNA sequencing, and qPCR
Total RNA was isolated from the cells using RNAasy Mini Kit (QIAGEN,Hilden,Germany,74104) following the manufacture's instructions. The concentration, quality, and integrity of the extracted RNA were evaluated using Nanodrop (Thermo Scientific, Waltham, MA; USA), an Agilent 2100 Bioanalyzer, and the Agilent RNA 6000 Nano kit (Agilent Technologies, 5067-1511). 0.4 μg total RNA was used to obtain a poly A-enriched RNA library by Novogene (Cambridge, United Kingdom). Library concentration was performed using a Qubit fluorometer (HS RNA assay kit, Agilent Technologies). Library size was measured by Agilent 2100 bioanalyzer. The libraries were then subjected to 150 bp paired-end next-generation sequencing (Illumina NovaSeq 6000 S4 flow cells). Mutation visualization was performed using the Integrative Genomic Viewer tool . Read counts of the sequenced RNA were normalized to transcripts per million (TPM). The TPM values of the variables were used to plot heatmaps and for principle component analysis (PCA) based on Pearson correlation, using self-written scripts in R (R Development Core Team (2011)) (version 4.0.4).
RNA was reverse transcribed using the RevertAid First Strand cDNA Synthesis Kit (Thermo Scientific). qPCR was performed using the FastStart Universal SYBR Green Master (Rox) mix (Hoffmann-La Roche) according to the manufacturer's instructions. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA expression was calculated according to the ΔΔCt method. All primer pairs were designed using Primer3, purchased at BioTeZ (Berlin, Germany), and sequences are shown in Supplementary file 1.
Single-cell RNA sequencing (scRNA-seq) Cells isolation and preparation Differentiated cells at day 7 were washed twice with 1× DPBS, dissociated with Accumax solution, and resuspended in 1× DPBS. Then, cells were filtered, counted, and checked for viability.

Library preparation and single-cell sequencing
Single-cell 3' RNA sequencing was performed using the 10× Genomics toolkit version v3.1 (Alles et al., 2017), according to the manufacturer's instructions aiming for 10,000 cells. Obtained libraries were sequenced on Illumina NextSeq 500 sequencers.

Single-cell sequencing data analysis and clustering
After sequencing and demultiplexing, fastq files were analyzed using Cellranger version 3.0.2. Gene expression matrices were then imported into R, and Seurat objects were created using the Seurat R package (version 4.0.5) (Stuart et al., 2019). The gene expression matrices were initially filtered by applying lower and upper cut-offs for the number of detected genes (500 and 6000, respectively). The filtered data were then log normalized and scaled according to the number of unique molecular identifiers. The normalized and scaled data derived from the four samples were then merged into one Seurat object. Clustering was performed using the first 20 principal components. We used the Seurat FindAllMarkers function to extract marker gene lists that differentiate between clusters with log fold change threshold ± 0.25 using only positive markers expressed in a minimum of 25% of cells. PCA was done using the first 20 principle components in R using the following libraries factoextra, FactoMineR, and ggplot2.

Protein extraction and immunoblotting
Proteins were extracted from hiPSCs using radioimmunoprecipitation assay (RIPA) buffer (Sigma-Aldrich, R0278) as described in details in supplementary data. 30 µg protein in RIPA buffer were mixed with 1× reducing (10% b-mercaptoethanol) NuPAGE loading buffer (Life Technologies, Carlsbad, CA), loaded on a precast polyacrylamide NuPage 4-12% Bis-Tris protein gel (Invitrogen, Carlsbad, CA, USA) and blotted on 0.45 µm pore size Immobilon-P Polyvinylidene difluoride membrane (EMD Millipore, Billerica, MA; USA). The membrane was blocked in 5% bovine serum albumin for 1 hr at room temperature and incubated overnight at 4°C with primary antibodies: anti-CEP83 produced in rabbit (1:500, Sigma-Aldrich) and anti-α-Tubulin produced in mouse (1:500, Sigma-Aldrich, T9026). Then, the membrane was incubated for 1 hr at room temperature with horseradish peroxidase-conjugated secondary antibodies (1:2000, Sigma-Aldrich, Saint Louis, MO, USA). Chemiluminescent reagent (Super Signal-West Pico; Thermo Scientific, Waltham, MA; USA) was used to detect the proteins. The spectra Multicolor Broad Range Protein Ladder (Thermo Fisher Scientific, USA) was used to evaluate the molecular weight of corresponding protein bands.

Histology and immunofluorescence staining
Cells at different time points were checked regularly under a confocal microscope (Leica DMI 6000 CEL) for differentiation progress. Quantitative analysis of nephron-like structure formation within each organoid (D25) was done on tile scanning images of each organoid by estimating the percentage of the organoid area composed of nephron-like structures using 13 WT and 9 KO organoids. Organoids were fixed in BD Cytofix buffer (554655, BD Biosciences) for 1 hr on ice. Then organoids were gradually dehydrated in increasing ethanol concentrations, cleared in xylene, and embedded in paraffin. Organoids were cut into 3.5-µm thick sections. The sections were deparaffinized, dehydrated, and stained in hematoxylin (Sigma-Aldrich, Saint Louis, MO) for 3 min and in 1% eosin (Sigma-Aldrich) for 2 min. For immunostaining, cells (day 7) and organoids (day 25) were fixed with BD Cytofix, permeabilized with BD Perm/Wash (554723, BD Biosciences), and blocked with blocking solution (1% BSA + 0.3% triton-X-100 in 1× DPBS) for 2 hr. Cells and organoids were incubated overnight at 4°C with primary antibodies (Supplementary file 2), then incubated with fluorescence-labeled secondary antibodies with 1:500 dilution including Cy3, Cy5, Alexa488, and Alexa647 (Jackson ImmunoResearch, Newmarket, UK) and Cy3 Streptavidin (Vector lab, Burlingame, USA) overnight at 4°C. DAPI was then used for nuclear staining (Cell signaling Technology, Danvers, MA, USA) with 1:300,000 dilution for 1 hr at RT. Finally, cells were mounted with Dako fluorescent mounting medium (Agilent Technologies). Images were taken using a SP8 confocal microscope (Carl Zeiss GmbH, Oberkochen, Germany). Quantitative analyses of acquired images were performed using ImageJ software (1.48 v; National Institutes of Health, Bethesda, MD).

Comparison to zebrafish LPM
The upregulated genes in CEP83 −/− cells on day 7 and day 25 were compared with the top 20 orthologous genes identified in subclusters of zebrafish LPM identified by scRNA-seq, as deposited on ArrayExpress (E-MTAB-9727; Prummel et al., 2022).

Statistical analysis
scRNA-seq was done on two biological replicates representing two different clones of CEP83 −/− and control cells, respectively. All other experiments were performed using three biological replicates representing three independent clones of CEP83 −/− and control cells at different time points. A common excel sheet for the genes present in both bulk RNA and scRNA sequencing was generated in R. The sheet includes a total of 20,894 genes and represents the TPM values of both groups (WT and KO) on day 0, day 7, and day 25 of differentiation. The maximum TPM (TPMmax) and the minimum TPM (TPMmin) were calculated for each gene across all samples. HVGs were calculated based on the ratio of TPMmax and TPMmin. For heatmaps and PCA analysis, the top 1000 HVGs were plotted with a selection of TPMmax >2 for each gene. Deregulated (upregulated and downregulated) genes between WT and KO groups were selected with expression criteria of TPM >2, fold change >1.5, and p-value calculated on log10 TPM <0.05. The unpaired two-tailed t-test was used to compare the two groups. All graphs were generated using GraphPad Prism 7.04 (GraphPad Software, San Diego, CA). Data are presented as mean ± SD.
• Supplementary file 1. The table shows the primers list used in the qPCR.
• Supplementary file 2. The table shows the list of primary antibodies used in IF staining.

Data availability
All data supporting the findings of this study are available within the article and its supplementary files. Source data files have been provided for Figures 1-6. Sequencing data have been deposited in GEO under accession code GSE205978.
The following dataset was generated: