Normal hAK samples were retrieved from the borders of renal cell carcinoma tumors from patients undergoing partial or total nephrectomy at Sheba Medical Center. This procedure was performed after approval by the local ethical committee and after obtaining signed informed consent from the patients (smc-5574-18). The samples were minced in Hanks’ balanced salt solution, soaked in collagenase for 2 h, and cultured in a growth medium consisting of Iscove’s modified Dulbecco’s medium supplemented with 10% fetal bovine serum (FBS), 1% L-glutamine, 1% penicillin–streptomycin, and the following growth factors: 50 ng/mL of bFGF, 50 ng/mL of EGF, and 5 ng/mL of SCF (R&D Systems, Minneapolis, MN, United States) at 37 °C in a humidified atmosphere containing 5% CO₂. The cells were passaged using trypsin–EDTA (Gibco) upon reaching 70%–80% confluency and used between passages 2 and 7 for all experiments to ensure consistency in cellular phenotype while avoiding senescence-associated changes.

Lentiviral vectors encoding human OSR1 and SIX2 cDNA under the control of the cytomegalovirus promoter were purchased from Vector Builder. The OSR1 construct included an mCherry fluorescent reporter for visual confirmation of expression, whereas the SIX2 construct contained only the gene of interest. A lentiviral vector encoding mCherry alone was used as a control to account for any effects of viral transduction or fluorescent protein expression. For transduction, the hAK cells were seeded at a density of 5 × 10⁴ cells per well in 6-well plates and allowed to adhere for 24 h. The cells were then transduced with lentiviral particles at a multiplicity of infection (MOI) of 5 in the presence of 8 μg/mL polybrene (Sigma-Aldrich) for enhanced transduction efficiency. After 24 h of incubation with the viral particles, the medium was replaced with fresh renal epithelial cell growth medium containing 10% FBS to remove any residual viral particles and polybrene. The transduced cells were selected with puromycin (2 μg/mL) for 7 d to establish stable cell lines expressing the genes of interest. To generate double-transduced cells, OSR1-expressing cells were additionally transduced with SIX2-encoding virus and selected with appropriate antibiotics (puromycin and neomycin) to ensure the expression of both factors. The resulting cell lines were designated as OSR1-hKEpCs, SIX2-hKEpCs, OSR1+SIX2-hKEpCs, and mCherry-hKEpCs (Supplementary Table S1). Successful transduction was confirmed through quantitative polymerase chain reaction (PCR), Western blotting, and fluorescence microscopy for the mCherry reporter.

The total RNA was extracted from the hAK cells using the RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. The RNA concentration and purity were then assessed using a NanoDrop spectrophotometer (Thermo Fisher Scientific), and samples having 260/280 ratios between 1.8 and 2.0 were considered suitable for further analyses. Next, cDNA was synthesized using the iScript cDNA Synthesis Kit (Bio-Rad) with 1 μg of RNA as the template. Real-time PCR was performed using the CFX96 RT-PCR Detection System (Bio-Rad) with SYBR Green PCR Master Mix (Bio-Rad). The thermal cycling conditions consisted of initial denaturation at 95 °C for 3 min, followed by 40 cycles of denaturation at 95 °C for 15 s each, annealing at 60 °C for 30 s, and extension at 72 °C for 30 s. A melting curve analysis was performed to confirm the specificity of the amplified products. Primers for all the target genes were designed using Primer-BLAST (NCBI) and validated for specificity and efficiency prior to use. The gene expressions were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) using the 2^-ΔΔCt method. Three independent experiments were performed with triplicate samples for each condition to ensure reproducibility.

For the proliferation assays, the cells were seeded at a density of 2 × 10⁴ cells per well in 12-well plates with complete growth medium. The cells were then counted at 24-h intervals for 5 d using a hemocytometer after trypsinization and trypan blue dye exclusion to assess viability. The growth rate was calculated as the number of cell divisions per day normalized to the control cells to determine the relative proliferation rate. For the clonogenic assays, the cells were seeded at ultralow densities (0.3, 1, and 10 cells per well) in 96-well plates via limiting dilution. After 14 d of culture with medium changes every 3 d, the colonies were fixed with 4% paraformaldehyde for 15 min at room temperature, stained with 0.5% crystal violet for 30 min, and washed thoroughly with phosphate-buffered saline (PBS). Colonies containing more than 50 cells were then counted under a light microscope, and the clonal efficiency was calculated as the percentage of wells containing colonies relative to the total number of wells seeded. For each cell line, at least 96 wells were analyzed for each density condition across three independent experiments.

All animal procedures employed in this study were approved by the Institutional Animal Care and Use Committee (IACUC-15-2018) at Sheba Medical Center and performed in accordance with the institutional guidelines. For subcutaneous transplantation, engineered human kidney epithelial cells (hKEpCs) were harvested at passages 3–6 using trypsin–EDTA (0.25%, Gibco), washed twice with PBS, and resuspended in growth-factor-reduced Matrigel (BD Biosciences) to provide a supportive extracellular matrix environment. Two injection protocols were then employed: 1.5 × 10⁶ cells in 200 μL of Matrigel for the AK83 cell line; 2 × 10⁶ cells in 50 μL of Matrigel mixed with 50 μL of PBS for the AK87 cell line. The cell suspensions were injected subcutaneously into the dorsal flank of 6–8 week old male NOD-SCID mice (n = four to six per group) under isoflurane anesthesia (2%–3% in oxygen). The injected cell masses were harvested either immediately (T0) or after 14 d (T14) for analyses. The animals were maintained under specific pathogen-free conditions at the animal facility of Sheba Medical Center with ad libitum access to food and water; further, they were monitored daily for signs of discomfort or health issues, and their body weights were recorded weekly.

The tissue samples were collected at 14 d and 28 d post-injection, fixed in 4% paraformaldehyde overnight at 4 °C, processed through graded ethanol, embedded in paraffin, and sectioned with a thickness of 5 μm. For the immunofluorescence analysis, the sections were deparaffinized, subjected to antigen retrieval in citrate buffer (pH 6.0) for 20 min at 95 °C, and blocked with 5% normal goat serum for 1 h at room temperature to reduce non-specific binding. The sections were next incubated overnight at 4 °C with the following primary antibodies: anti-human leukocyte antigen (HLA; 1:100, Abcam), anti-CD13 (1:200, Abcam), anti-epithelial membrane antigen (EMA; 1:100, Dako), anti-multifunctional/mesenephric nephric marker (MNF; 1:100, Sigma), anti-vimentin (1:200, Abcam), anti-pancytokeratin (1:200, Invitrogen), anti-E-cadherin (1:200, BD Biosciences), anti-SIX2 (1:200, Proteintech), and anti-Ki67 (1:100, Cell Signaling). After washing with PBS, the sections were incubated with appropriate Alexa-Fluor-conjugated secondary antibodies (1:500, Invitrogen) for 1 h at room temperature, and the nuclei were counterstained with DAPI (Sigma). Lastly, the slides were mounted with ProLong Gold Antifade (Invitrogen) and imaged using a Zeiss LSM 700 confocal microscope with ×20 and ×40 objectives. For the immunohistochemical analyses, the sections were processed with the Ventana automated immunostainer using DAB as the chromogen, followed by hematoxylin counterstaining. Sections from the human fetal kidney (14–16 weeks) and adult kidney were used as the positive controls for marker expression patterns, whereas primary antibody omission served as the negative control. Images were then acquired using a Zeiss LSM 700 confocal microscope or an Olympus BX51 light microscope equipped with a digital camera.

The total RNA was extracted from the engineered hKEpC lines (SIX2-hKEpCs, OSR1-hKEpCs, OSR1+SIX2-hKEpCs, and naïve control hKEpCs) at passage 4–6 using the RNeasy Mini Kit (Qiagen) with on-column DNase digestion (RNase-Free DNase Set, Qiagen) to eliminate genomic DNA contamination. Three independent biological replicates were prepared for each cell line to ensure statistical robustness. Then, the RNA concentrations were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific), where 260/280 ratios between 1.8 and 2.0 were considered acceptable. RNA integrity was assessed using an Agilent 2100 Bioanalyzer with the RNA 6000 Nano Kit (Agilent Technologies), and only samples with RNA integrity number (RIN) values exceeding 8.0 were selected for library preparation and sequencing.

The RNA sequencing libraries were prepared from 1 μg of total RNA using the TruSeq Stranded mRNA Library Prep Kit (Illumina) according to manufacturer protocols. Briefly, the mRNA was isolated using poly-A selection before being fragmented and reverse-transcribed to generate first-strand cDNA using random hexamer primers. Then, the second-strand cDNA synthesis was performed using dUTP to maintain strand specificity. Following end repair, A-tailing, and adapter ligation, the libraries were amplified by PCR for 12–15 cycles. The library quality and concentration were assessed using the Agilent 2100 Bioanalyzer with the DNA 1000 Kit and quantified by qPCR using the KAPA Library Quantification Kit (KAPA Biosystems). The libraries were sequenced on an Illumina NovaSeq 6000 platform using paired-end 150-bp reads with a target depth of >30 million read pairs per sample to ensure adequate coverage for differential expression analysis. The raw sequencing reads (FASTQ files) were assessed using FastQC (v0.11.9) to evaluate the read quality metrics, adapter contamination, and level of sequence duplication. Quality filtering and adapter trimming were then performed using Trimmomatic (v0.39) with the following parameters: ILLUMINACLIP:TruSeq3-PE.fa:2:30:10, LEADING:3, TRAILING:3, SLIDINGWINDOW:4:15, and MINLEN:36. The high-quality trimmed reads were aligned to the human reference genome (GRCh38/hg38) using STAR aligner (v2.7.3a) with the default parameters optimized for mammalian genomes. The alignment quality was assessed using samtools (v1.10) and RSeQC (v3.0.1) to evaluate the mapping rates, read distributions, and potential biases.

Next, gene expression quantification was performed using featureCounts (v2.0.1) from the Subread package by counting reads mapping to protein-coding genes annotated in GENCODE v32; the raw count matrices were imported into R (v4.0.3) for downstream analyses. The differential gene expression analysis was conducted using DESeq2 (v1.30.0) with the default normalization and dispersion estimation procedures. Genes with low expressions (fewer than 30 counts across all samples) were filtered prior to the analyses. The differential expressions were assessed using the Wald test, with genes showing adjusted p-value <0.05 (Benjamini–Hochberg correction) and absolute log₂(fold change) >1 being considered as significantly differentially expressed.

To benchmark the overexpression levels against physiological ranges, we compared the SIX2 and OSR1 expressions in our engineered cell lines with published RNA sequencing data from human fetal kidney tissues. The expression data were obtained from O'Brien et al. (2016), which provided the transcriptomic profiles of 17-week-old human fetal kidney cortex, including whole tissue, cortex-enriched fractions, and purified ITGA8+ NPCs isolated by FACS. The transcripts per million (TPM) values for SIX2 and OSR1 were extracted from published datasets and compared with our engineered cell lines to evaluate the biological relevance of the overexpression levels relative to developmental kidney cell populations. Principal component analysis (PCA) was performed using the plotPCA function in DESeq2 with variance-stabilizing transformation (VST) to visualize the global transcriptional relationships between the samples. Hierarchical clustering analysis was then conducted using the pheatmap package (v1.0.12), with Euclidean distance and complete linkage clustering of the top-1,000 most variable genes. Pearson correlation coefficients were calculated between the log₂-normalized SIX2 or OSR1 counts and expressions of all genes (VST scale) across the samples (control and overexpressing). For each gene, the p-values were computed by permutation and adjusted for multiple testing. Genes with strong positive (r > 0.7) or negative (r < −0.7) correlations were visualized in color-coded bar plots (red: positive; blue: negative; gray: weak or non-significant). To identify genes whose expressions varied with SIX2 or OSR1 dosage across all samples, we fit a linear model for each gene using the limma package (lmFit, eBayes) by modeling the gene expression as a function of the log₂ expression of SIX2 or OSR1 as well as applying empirical Bayes moderation and array weights to handle the sample outliers; genes with significant associations (adjusted p-value <0.05, |log₂FC| > 0.1) were then selected for the downstream analyses. Gene ontology (GO) biological process (BP) enrichment was performed for the upregulated (positive slope) and downregulated (negative slope) genes using clusterProfiler enrichGO; bar plots were used to summarize the most enriched GO-BP terms (blue: upregulated; red: downregulated). The TPM values were calculated and visualized as mean ± standard error of the mean (SEM) bar plots for selected genes across different conditions. The data were visualized using ggplot2 (v3.3.2) for the PCA plots and volcano plots, pheatmap for heatmaps, and custom R scripts for the correlation plots and pathway visualizations. All bioinformatics analyses were performed using R/Bioconductor packages, along with documentation of the reproducible code for transparency and replication.

The OSR1-hKEpCs were cultured to 70%–80% confluence and treated with colcemid (0.1 μg/mL) for 2 h to arrest the cells in metaphase; the cells were then trypsinized, treated with a hypotonic solution (0.075 M of KCl) for 20 min at 37 °C, and fixed with freshly prepared methanol/acetic acid (3:1) through three sequential changes of the fixative. The fixed cell suspensions were placed dropwise onto clean glass slides and air-dried. Then, the chromosomes were stained with Giemsa (4% in phosphate buffer, pH 6.8) for 10 min and analyzed under a light microscope with a ×100 oil-immersion objective. At least 20 metaphase cells were analyzed for each sample to identify the numerical and structural chromosomal abnormalities. The karyotyping was performed in accordance with the guidelines of the International System for Human Cytogenetic Nomenclature.

Genomic DNA was extracted from the OSR1-hKEpC-E09 cells using the DNeasy Blood & Tissue Kit (Qiagen). Then, lentiviral integration sites were identified using custom biotinylated probes targeting the LTR regions, followed by targeted sequence capture and next-generation sequencing on an Illumina platform (Miyazato et al., 2016; Ustek et al., 2012). The sequencing reads were processed using a bioinformatics pipeline for integration site analysis, including quality filtering, genome alignment (GRCh38), and junction identification (Park et al., 2015). The insertion site was mapped to an intronic region of HECTD4 on chromosome 12, with no integration events detected in the known oncogenes or tumor suppressor genes, indicating that the malignant transformations were not due to insertional mutagenesis.

The data are presented as mean ± SEM unless indicated otherwise in the figure legends. For the proliferation assays, experiments were performed using cells from three independent primary kidney cell isolations (n = 3 biological replicates), where each experiment was conducted in three technical replicates and repeated three times. For the clonogenic assays, three independent primary cell sources were used (n = 3 biological replicates), where each density condition analyzed in at least 96 wells per experiment and repeated across three independent experiments. For the in vivo transplantation studies, three different primary cell sources were used (hAK83, hAK86, and hAK87; n = 3 biological replicates), where each cell line transplanted into 4–6 mice per group. Statistical significance was determined using appropriate tests based on the data distribution and experimental design. For two-group comparisons, unpaired two-tailed Student’s t-test was used after confirming normal distribution using the Shapiro–Wilk test. For multiple groups, we performed one-way ANOVA followed by Tukey’s post hoc test for multiple comparisons. For the RNA sequencing data, differential expressions were assessed using DESeq2 with Benjamini–Hochberg correction for multiple testing, where adjusted p-value <0.05 and |log₂FC| >1 were considered to be statistically significant. The exact p-values are reported in the figures where space permits, given the following significance thresholds: *p < 0.05, **p < 0.01, and ***p < 0.001. All statistical analyses were performed using GraphPad Prism 10 (GraphPad Software, San Diego, CA, United States).

The sample size was determined using the resource equation method (Festing and Altman, 2002) that is recommended when precise estimates of the effect size and variance are not available, as in exploratory mechanistic in vivo tubulogenesis studies. In our design, three experimental groups (naïve, SIX2-overexpressing, and OSR1-overexpressing cells) were each generated from three independent human donors (biological replicates) and transplanted into 4–6 recipient mice per group. For the resource equation E = N−G, where N is the total number of experimental units (mice) and G is the number of groups, our experimental range (total N = 12–18; G = 3) yielded E = 9–15, which is within the recommended range of 10–20 for adequately powered exploratory studies. This approach balances the need for statistical validity with ethical considerations of animal use, in line with ARRIVE guidelines. The sample sizes were determined through preliminary experiments and a power analysis to ensure adequate statistical power (>0.8) while minimizing animal usage in accordance with the ethical guidelines. All experiments included appropriate controls and were performed by investigators blinded to the experimental conditions where feasible.

To investigate the differential effects of renal developmental gene induction on hAK cells, we established primary hAK cell lines overexpressing OSR1 and SIX2 individually and in combination (OSR1+SIX2) using lentiviral vectors (Supplementary Figure S1A). The control groups included naïve hKEpCs as well as hKEpCs expressing mCherry. Successful transduction was confirmed by qRT-PCR, which showed significant overexpressions of SIX2 (approximately 50,000-fold) and OSR1 (approximately 3,000-fold) compared to the control cells (Supplementary Figure S1A). SIX2 overexpression was validated at the protein level by Western blotting, whereas OSR1 expression was visualized through the mCherry reporter (Supplementary Figure S1B,C). The transcription factors showed appropriate nuclear localizations, confirming their functional expressions. The resultant cell lines displayed distinct morphological characteristics; the control cells (naïve hKEpCs and mCherry-hKEpCs) exhibited typical epithelial morphology characteristic of primary cultured kidney cells (Figure 1B; Supplementary Figure S1D), whereas the SIX2-hKEpCs maintained their epithelial appearance with minimal morphological alterations and showed only a slightly more compact hyperchromatic phenotype (Figure 1B). In contrast, OSR1 induction triggered significant morphological changes, with the cells developing a dense hyperchromatic morphology that was distinctly different from the surrounding cells (Figure 1C). After several weeks in culture, the OSR1-hKEpCs formed well-defined colonies with small and cuboidal cells that aggregated in colony-like structures. Notably, at higher passages, these cells adopted a spindle-shaped morphology characteristic of senescent hKEpCs, suggesting the transient nature of morphological reprogramming. The OSR1+SIX2-hKEpCs exhibited an intermediate morphology with features characteristic of overexpression of both of the primary individual genes (Figure 1B). These morphological alterations, particularly in the OSR1-hKEpCs, are consistent with activation of partial epithelial-to-mesenchymal transition (EMT), which is associated with developmental reprogramming and is observed during the early stages of nephrogenesis when mesenchymal condensation occurs prior to epithelialization.

The key phenotypic differences among the NPC-gene-expressing hAK cells are their proliferative and clonogenic capacities. SIX2 overexpression significantly enhanced the proliferative capacity of adult kidney cells, where the SIX2-hKEpCs demonstrated a 2.39-fold higher growth rate than the control cells (p < 0.05) (Figure 1D). Additionally, the SIX2-hKEpCs exhibited significantly improved clonogenic capacity, with a 2.07-fold increase in single-cell-derived colony formation compared to the control cells (p < 0.05) (Figure 1E). This enhanced self-renewal capacity is consistent with the established role of SIX2 in maintaining nephron progenitor proliferation during kidney development. In contrast, the OSR1-hKEpCs did not demonstrate a significant difference in the overall proliferation rate compared to the control cells (0.94-fold vs. control) (Figure 1D). Cells coexpressing OSR1 and SIX2 (OSR1+SIX2-hKEpCs) demonstrated enhanced proliferation (3.15-fold, p < 0.05) even beyond that of the SIX2-hKEpCs (Figure 1D), suggesting a synergistic effect between the two factors on proliferation. The clonogenic capacity of these cells was also comparable to that of SIX2-hKEpCs, with a slight additional improvement that is possibly attributable to the OSR1-induced stemness properties. These findings indicate the distinct functional roles of SIX2 and OSR1 in regulating adult kidney cell behaviors, where SIX2 primarily enhances the proliferation and colony formation capacities while OSR1 confers novel clonogenic potential at ultralow densities without significantly affecting the overall proliferation rate. The combination of both factors appears to have a synergistic effect, particularly on proliferation.

The tubulogenic capacities of the engineered kidney cells were assessed through subcutaneous injection in immunodeficient mice. To assess the functional implications of renal progenitor gene overexpression, we evaluated the capacities of the engineered cell lines to form tubule-like structures in vivo. When injected subcutaneously into immunodeficient NOD-SCID mice, all three cell populations (control hKEpCs, SIX2-hKEpCs, and OSR1-hKEpCs) from the three different hAK lines initially exhibited comparable engraftment at T0, as demonstrated by HLA staining (Figure 2B, upper panels). After 14 d (T14), striking differences were observed in the organization, with the SIX2-hKEpCs assembling into distinct tubular structures with clear luminal formation, while the control hKEpCs formed limited disorganized cell clusters and OSR1-hKEpCs primarily remained as scattered cells with minimal aggregation (Figure 2B, lower panels).

All three cell types initially expressed both vimentin and cytokeratin at T0, indicating their epithelial–mesenchymal characteristics (Figure 2C, upper panels). By T14, the SIX2-hKEpCs developed well-defined tubular structures that maintained vimentin/cytokeratin coexpression, with a representative high-magnification image showing a clear tubular lumen (Figure 2C, lower panels). After 14 d, the control hKEpCs formed mainly EMA-positive aggregates with limited tubular organization, whereas the SIX2-hKEpCs developed tubular structures with distinct luminal architecture comprising CD13⁺, EMA⁺, or occasionally CD13⁺EMA⁺ cells, with the latter possibly representing a transitional cell state (Figure 2D). The tubular structures formed by the SIX2-hKEpCs continued to express SIX2 but showed decreased proliferation compared to T0, as indicated by Ki67 staining (Figure 2D; Supplementary Figure S2). These findings establish SIX2 as a key driver of tubulogenic potential in adult kidney cells and promoter for formation of organized tubular structures in vivo, whereas OSR1 expression alone was insufficient to confer robust tubulogenic capacity.

RNA sequencing analysis revealed distinct transcriptional signatures among the cell lines overexpressing SIX2, OSR1, or both. PCA demonstrated clear separation between the cell types, with the SIX2-hKEpCs and naïve control cells forming well-defined clusters (Figure 3A). Notably, one OSR1-overexpressing sample clustered distantly from the other samples, potentially reflecting the heterogeneity in OSR1-induced cellular responses or extreme overexpression levels that may have affected transcriptional stability. Analysis of the gene expression patterns revealed coordinated changes in nephron-segment-specific markers across the overexpressing cell lines (Figure 3B). Both SIX2-hKEpCs and OSR1-hKEpCs demonstrated significant downregulation of the proximal tubule markers, including SLC3A1, AQP1, GPX3, ANPEP, and HNF1A, suggesting a shift away from mature proximal tubule identity. Concurrently, these cells showed upregulation of epithelialization markers correlated with distal nephron segments, such as EPCAM, CDH1, MUC1, KRT8, and TFAP2B, indicating reprogramming toward a more distal-tubule-like state. The SIX2-hKEpCs alone or in combination with OSR1 (SIX2+OSR1-hKEpCs) exhibited pronounced overexpression of proliferation and cell-cycle genes, including CDK1, MKI67, TOP2A, and CCNA2, which are directly correlated with the enhanced clonogenic and tubulogenic capacities observed in our functional analyses.

Gene correlation analysis revealed distinct transcriptional networks associated with SIX2 and OSR1 overexpressions (Figures 4A–C; Supplementary Figure S4; Supplementary Tables S1–S4). SIX2 expression showed strong positive correlations with the epithelial markers and developmental regulators, most notably EPCAM (highest correlation), followed by SLC12A3, GATA3, FOXD1, TFAP2B, CLDN1, LHX1, IRX2, ALK6, CDK1, MMP7, GPC3, and WNT11 (Figure 4A); these positively correlated genes suggest that SIX2 overexpression drives epithelialization with enhanced proliferative capacity. Conversely, SIX2 showed negative correlations with some progenitor markers like SALL1 and POU3F3, proximal tubule markers like ANPEP and GPX3, and mesenchymal markers like VCAN and CDH6; this pattern indicates that SIX2 overexpression suppresses specific progenitor programs and proximal tubule differentiation while promoting distal epithelial identity. The OSR1-correlated genes (Figure 4B) demonstrated positive associations with developmental factors like OCIAD2, GDNF, and EPCAM, while showing negative correlations with progenitor markers like PAX2 and PAX8 as well as the proximal tubule genes HNF1A and SLC3A1, indicating distinct regulatory networks activated by each transcription factor.

Pathway enrichment analysis revealed differential activation of the BPs between the transcription factors (Figure 4C). SIX2 overexpression was enriched for pathways related to mitotic nuclear division, microtubule cytoskeleton organization, and cell-cycle regulation while being downregulated for cellular responses to decreased oxygen levels and hormone metabolic processes. In contrast, OSR1 overexpression was upregulated for pathways involved in epithelial tube morphogenesis, renal system development, and canonical Wnt signaling while being downregulated for metabolic processes like cellular responses to hypoxia and amino acid metabolism. Given the superior functional performance of the SIX2-hKEpCs in both in vitro proliferation assays and in vivo tubulogenic studies, we conducted focused transcriptional analysis of these cells (Figure 4D). Hierarchical clustering analysis revealed distinct expression patterns across key functional categories, including kidney development, nephron segments, EMTs, and metabolic processes. Detailed analyses of the nephron segment markers confirmed that the SIX2-hKEpCs specifically downregulated proximal tubule genes (SLC3A1, AQP1, GPX3, ANPEP, and HNF1A) while maintaining or upregulating certain distal tubule markers (EPCAM, CDH1, MUC1, KRT8, and TFAP2B). The kidney developmental genes, including SIX2, SIX1, LHX1, and IRX2, showed enhanced expressions while the proliferation/cell-cycle markers like CDK1, MKI67, TOP2A, and CCNA2 were significantly upregulated, consistent with the enhanced proliferative capacity observed functionally. In summary, the addition of SIX2 or OSR1 to adult kidney cells can reprogram their gene expression profile away from a mature proximal tubule phenotype toward an epithelialized state with the features of distal nephron segments. SIX2 alone or with OSR1 can concomitantly enhance the proliferative genes to support improved self-renewal and tubulogenic functions.

A strikingly unexpected finding of this study was that OSR1 overexpression led to malignant transformation in approximately one in a hundred OSR1-hKEpC-transduced clones. This discovery was based on the identification of a unique clone, OSR1-E09-hKEpCs, which exhibited distinct morphological characteristics that prompted further investigation (Figures 5A,B). Unlike the other OSR1-hKEpCs clones, this particular clone displayed exceptional growth and markedly different morphological features, prompting us to conduct a more detailed characterization. Karyotype analysis of the transformed cells revealed multiple chromosomal abnormalities affecting both the structure and numbers across multiple chromosomes, confirming cellular transformation (Supplementary Figure S4H). Importantly, sequencing analysis of the viral insertion site showed no disruption of any known oncogenes or tumor suppressor genes, suggesting that the transformation was likely due to the OSR1 overexpression itself rather than insertional mutagenesis (Supplementary Figure S4I). When injected subcutaneously into NOD-SCID mice, the OSR1-E09-hKEpCs clone formed tumors within approximately 3 weeks. Histological examination of these tumors showed a structure characterized by dense sheets of poorly differentiated cells with a hyperchromatic morphology, high nucleus-to-cytoplasm ratios, and evident mitotic activity (Supplementary Figure S4J) that closely resembled the blastemal component of WT.

Immunohistochemical characterization revealed that these tumors exhibited a strikingly similar molecular profile to WT. Both the OSR1-E09-hKEpC-derived masses and WT samples expressed key renal developmental markers, including SIX2, NCAM1, and WT1 (Figure 5D). A comparison between WT, OSR1-E09-hKEpC-derived masses, and human fetal kidney tissue revealed further similarities in the phenotypic profiles (Figure 5E). The tumors were exclusively mesenchymal, expressing vimentin but not cytokeratin (Figure 5E, left column), and lacked mature renal epithelial markers like EMA and CD13 (Figure 5E, middle column). Additionally, the tumors showed robust positive staining for the proliferation marker Ki67 (Figure 5E, right column), comparable to the proliferative activity observed in the blastemal component of WT. Importantly, this malignant transformation phenomenon was not observed in either the SIX2-hKEpCs or OSR1+SIX2-hKEpCs, suggesting that it is specific to OSR1 overexpression and may be mitigated by concomitant SIX2 expression. The relatively low transformation frequency (approximately one in 100 OSR1-transduced clones) suggests that additional factors may influence susceptibility to malignant transformation, possibly including stochastic genetic or epigenetic alterations that cooperate with OSR1 overexpression to drive tumorigenesis.