Controlling nephron precursor differentiation to generate proximal-biased kidney organoids with emerging maturity

Abstract

The kidney maintains fluid homeostasis by reabsorbing essential compounds and excreting waste. Proximal tubule cells, crucial for reabsorbing sugars, ions, and amino acids, are highly susceptible to injury, often leading to pathologies necessitating dialysis or transplants. Human pluripotent stem cell-derived kidney organoids offer a platform to model renal development, function, and disease, but proximal nephron differentiation and maturation in these structures is incomplete. Here, we drive proximal tubule development in pluripotent stem cell-derived kidney organoids by mimicking in vivo proximal differentiation. Transient PI3K inhibition during early nephrogenesis activates Notch signaling, shifting nephron axial differentiation towards epithelial and proximal precursor states that mature to proximal convoluted tubule cells broadly expressing physiology-imparting solute carriers including organic cation and organic anion family members. The "proximal-biased" organoids thus acquire function, and on exposure to nephrotoxic injury, display tubular collapse and DNA damage, and upregulate injury response markers HAVCR1/KIM1 and SOX9 while downregulating proximal transcription factor HNF4A. Here, we show that proximally biased human-derived kidney organoids provide a robust model to study nephron development, injury responses, and a platform for therapeutic discovery.

Fig. 1. Comparative morphological and transcriptomic analyses of proximal nephron formation in vivo and in iPSC-derived kidney organoids.

a, b Immunofluorescent antibody stains at defined stages of nephrogenesis. Human kidneys from week 16.1 of development. a represents stages before the onset of HNF4A, while b represents the commencement and elongation of the HNF4A⁺ domain. White arrowheads indicate earliest detectable HNF4A⁺ cells. S-shaped body-stage channels are split. Scale bars: 10 microns. c, d Whole-mount immunofluorescent stains of day 10 kidney organoids. Boxed regions are magnified and split into individual channels. Scale bars: 10 microns. e, f Whole-mount immunofluorescent stains of kidney organoid nephrons. Data in (e) match proteins and stages in (a), while (f) corresponds to (b). White arrowheads in (f) indicate earliest detectable HNF4A⁺ cells (Day 13) and earliest emerging HNF4A⁺ domain (Day 14). Scale bars: 10 microns. g Hierarchical clustering from bulk RNA-sequencing of kidney organoids at differentiation days 10, 12 (control) and 12 (treated for 48 h with 10 µM LY294002). Genes shown are those that are highlighted in Supplementary Fig. 1j. Legend categorizes genes and timepoints. Timeline of organoid differentiation and samples profiled is shown in the top right. h, i Gene set enrichment analysis plots from Day 12 LY294002-treated / Day 12 control kidney organoids. Plots depict pathway effects from treatment with the PI3K inhibitor LY294002: downregulated (PI3K-Akt signaling pathway, h and upregulated (Notch signaling pathway, (i). Normalized enrichment scores (NES), false discovery rates (FDR), and p-values are shown for each.

Fig. 2. PI3K inhibitor-treated kidney organoid nephrons exhibit a shift toward early tubular nephron cell fates.

a UMAP of single-cell RNA-sequencing from day 10 (pre-treatment), day 12 control, and day 12 PI3K inhibitor (LY294002)-treated kidney organoid nephrons. Cells are colored by sample with marker gene annotations. b Heatmap of Z-scores for hierarchically clustered differentially expressed genes across samples. Representative genes are listed, with a Z-score legend. c Feature plots of genes enriched in each sample group from the differentially expressed gene list. d Feature plots of selected genes for each group based on sample characteristics. For c, d human nephron single-cell detection (Supplementary Fig. 1a) is shown on the left, organoid plot on the right. e UMAP of single-cell RNA-sequencing from day 12 control and day 12 LY294002-treated kidney organoid nephrons. Cells match a but exclude day 10 sample. Cells are colored by sample with marker gene annotations. f Heatmap of log₂ fold change (log₂FC) for differentially expressed genes, calculated as the ratio of LY294002-treated to control aggregate expression. Representative genes are listed with a log₂FC legend. g Feature plots of genes enriched in each sample from the differentially expressed gene list. Human nephron single-cell (Supplementary Fig. 1a) detection is on the left, organoid plot on the right. h Split violin plots of genes from the differentially expressed gene list. Top two rows show genes enriched in day 12 control cells; bottom two rows show genes enriched in day 12 LY294002-treated cells. Plot colors match sample colors from (e). i Gene ontology terms for the top 50 differentially expressed genes per sample. Top graph: day 12 control nephron cells; bottom graph: LY294002-treated nephron cells. j–l Whole-mount immunofluorescent stains of day 12 control and LY294002-treated kidney organoids. Boxed regions are magnified. Inset in j shows absence of detectable HNF4A protein (orange) in day 12 samples. Scale bars: 10 microns. m Whole-mount immunofluorescent stain of day 12 control and LY294002-treated kidney organoids. Insets highlight JAG1 and HNF1B protein detection. Scale bars: 500 microns. n Quantification of JAG1⁺ and HNF1B⁺ nephron size (µm²) and HNF1B⁺ intensity (RFU) for n = 3 day 12 organoids each, across all positive segments. SEM error bars shown. Statistical significance determined by two-sided Student’s t-test.

Fig. 3. Kidney organoid nephrons shift toward proximal-biased cell fates following early PI3K inhibitor treatment.

a Single-cell RNA-seq UMAP from day 14 control and PI3K inhibitor-treated (LY294002 from days 10–12, proximal-biased, PB) kidney organoid nephrons. Cells colored by sample with marker gene annotations. b Log₂ fold change (log₂FC) heatmap for differentially expressed genes, calculated as the ratio of proximal-biased to control aggregate expression. Representative genes listed with a legend. c Feature plots of genes enriched in each sample from the differentially expressed gene list. d Split violin plots of differentially expressed genes. Top two rows: genes enriched in day 14 control cells; bottom two rows: genes enriched in day 14 proximal-biased cells. Plot colors match sample colors from (a). e Feature plots of selected genes for each group based on sample characteristics. For c, e human nephron (Supplementary Fig. 1a) detection is on the left, organoid plot on the right. f Immunofluorescent stains of day 14 control and proximal-biased kidney organoids. Boxed regions split and magnified. White arrowheads mark early HNF4A⁺ nephron regions, yellow arrowheads indicate autofluorescence. Scale bars: 10 microns. g Bulk RNA-sequencing (TPM) of select genes on days 10, 12, 14, and 18 from n = 2 whole organoids. SEM error bars shown. h, i FACS of day 18 kidney organoids from an HNF4A-YFP iPSC reporter line for control (h) and proximal-biased (i) conditions. Brightfield and YFP channels are shown as merged and split panels. FACS plots in duplicates with YFP⁺ percentages in bottom right. Scale bars: 500 microns. j Immunofluorescent stains of week 16.1 human kidneys during HNF4A⁺, HNF4G⁺ proximal tubule elongation. White arrowheads mark endothelial autofluorescence. Scale bars: 10 microns. k Immunofluorescent stain of control and proximal-biased day 18 organoids. Boxed regions magnify HNF4A detection. Scale bars: 200 microns. l Total HNF4A⁺ nephrons in n = 4 whole day 18 organoids. SEM error bars shown. Statistical significance by two-sided Student’s t-test. m Average area (µm²) of HNF4A⁺ nephrons from n = 4 whole day 18 organoids, across all positive segments. SEM error bars shown. Statistical significance by two-sided Student’s t-test. n Sum total area (µm²) of HNF4A⁺ nephron segments from n = 3 whole day 18 organoids.

Fig. 4. Single-cell transcriptomic and morphological analyses of nephron segment development in proximal-biased kidney organoids across differentiation stages.

a Schematic of single-cell RNA-sequencing timepoints with UMAP reduction of kidney organoid nephrogenic cells, colored by samples. b UMAP reduction of nephrogenic cell clusters in kidney organoids, colored by clusters, with identifying marker gene annotations. UMAPs split by sample type of origin (untreated day 10; control days 12, 14, and 18; or proximal-biased (PB) days 12, 14, and 18) are shown at the top. c Feature plots for select nephron marker genes with insets separating control from proximal-biased cells, and scaled percent quantification of total cells expressing the gene. Whole-mount immunofluorescent stains of control (d) and proximal-biased (e) kidney organoid nephrons from differentiation days 11, 12, 14, 15, and 18, with the bottom row highlighting HNF4A protein detection. Scale bars: 10 microns. Differentiation model for control kidney organoids (f) based on (d), and for proximal-biased kidney organoids (g) derived from (e). h Time-lapse video captures of kidney organoids differentiated from a human HNF4A-YFP iPSC reporter line from differentiation days 15–18, showing the emergence and progression of HNF4A-YFP detection in control and proximal-biased conditions. Brightfield and YFP channels are merged. Scale bars: 200 microns. Movies are included as supplementary files. i Whole-mount immunofluorescent stains of control and proximal-biased kidney organoids from differentiation days 14 and 21. Boxed regions are magnified. Scale bars: 10 microns. j Whole-mount immunofluorescent stains of control and proximal-biased kidney organoids from differentiation days 14, 18, and 21. Boxed regions are magnified. Scale bars: 10 microns. k Quantification of nephron segments positive for indicated markers per 15X field of view at differentiation days 14 and 21. Data are from n = 3 independent organoids, including those shown in (i, j). SEM error bars shown. Statistical significance is determined using two-sided Student’s t-test. l Quantification of the average segment length (µm) for indicated proteins in day 18 control and proximal-biased kidney organoids. Data are from n = 4 independent organoids (15X), including those shown in Supplementary Fig. 7c, d. SEM error bars are shown. Statistical significance is determined using two-sided Student’s t-test.

Fig. 5. Single-cell transcriptomics-driven comparison of proximal-biased kidney organoids and proximal tubule-enhanced kidney organoids.

a UMAP reduction of single-nucleus RNA-sequencing of differentiation day 21 proximal-biased and day 27 proximal-biased (PB) kidney organoids. Cells are colored by cluster, with marker gene annotations. UMAPs split by sample timepoint of origin are shown at the top. b Dot plot with identifying genes for each cluster. c Feature plots for select genes, showing a matching comparison to Supplementary Fig. 1c, f. d UMAP reduction of the nephrogenic subset of day 21 proximal-biased and day 27 proximal-biased kidney organoid single-nucleus RNA-sequencing from (a). Cells are colored by sample timepoint of origin, with marker gene annotations. e Feature plots of select proximal nephron organic cation, organic anion, solute, and drug transporters. Detection of the gene in human nephron single-cell RNA-sequencing (Supplementary Fig. 1a) data is shown on the left, with the organoid plot on the right. In vivo proximal tubule segment specificity is listed. f UMAP reduction of single-cell RNA-sequencing of differentiation day 13 + 14 (27) proximal tubule-enhanced kidney organoids, with marker gene annotations, acquired from ref. . g Dot plot with identifying genes for each cluster. h Feature plots for select genes, showing a matching comparison to Supplementary Fig. 1cf. i Diagram of differentiation and sampling timelines for day 21 proximal-biased, day 27 proximal-biased, and day 13 + 14 (27) proximal tubule-enhanced kidney organoids. j Schematic of merging single-cell RNA-sequencing datasets: day 21 proximal-biased and day 27 proximal-biased kidney organoids (a) were integrated with differentiation day 13 + 14 (27) proximal tubule-enhanced kidney organoids in (f). UMAP reduction of resulting integration is shown, with cells colored by cluster. Split UMAP reductions showing sample of origin are shown on the right. k Feature plots of select nephron, proximal tubule, and off-target genes, split by sample of origin: proximal-biased cells (left plot) or proximal tubule-enhanced cells (right plot).

Fig. 6. Functional characterization and injury response in proximal-biased kidney organoids.

a Live imaging of control and proximal-biased HNF4A-YFP kidney organoids on day 21 co-cultured with Alexa 647-Dextran and Alexa 594-LRP2 antibody. Split channels shown. Arrowheads indicate autofluorescence (yellow) and true signal (white). Scale bars: 100 microns. b Quantification of HNF4A-YFP⁺ segments co-detected with 594-LRP2 and 647-Dextran. Averaged from 3 images/organoid, n = 3 organoids. SEM error bars shown. Statistical significance by two-sided Student’s t-test. c Whole-mount immunofluorescence of uninjured and cisplatin-injured kidney organoids on day 21, with HNF4A⁺ nephron regions magnified. Scale bars: 10 microns. d Timeline of kidney organoid injury strategy. e Quantification of HAVCR1⁺ signal in control and proximal-biased nephrons on day 21 post-injury (n = 3 replicate organoids). SEM error bars shown. Statistical significance by two-sided Student’s t-test. f Quantification of HNF4A⁺ nephron segments with HAVCR1 expression from (c). Data from 4 images/organoid (20X), n = 2 replicates. SEM error bars shown. Statistical significance by two-sided Student’s t-test. g Immunofluorescence of uninjured and cisplatin-injured proximal-biased kidney organoid nephrons on day 21. Dashed white line marks uninterrupted tubule lumen. Scale bars: 10 microns. h Individual continuous lumen length quantifications from (g), n = 4 organoids per condition. SEM error bars shown. Statistical significance by two-sided Student’s t-test. i Immunofluorescence of uninjured and cisplatin-injured proximal-biased kidney organoid nephrons on day 21. Yellow arrowheads: HNF4A⁺ tubule cells with low HNF4A and high γH2AX. White arrowheads: cells with high HNF4A and low γH2AX. Split channels for HNF4A and γH2AX shown. Scale bars: 10 microns. j Immunofluorescence of uninjured and cisplatin-injured kidney organoid nephrons on day 21. White arrowheads: SOX9⁺ interstitial cells. Scale bars: 10 µm. k HNF4A and SOX9 intensities from (j). Intensities from individual cells in duplicates, min-max normalized. l Kidney organoid single-injury timeline for organoids in (m). m Immunofluorescence of uninjured and cisplatin-injured proximal-biased kidney organoid nephrons on days 19, 21, and 25. Arrowheads indicate cells with high HNF4A and low SOX9 (white), and low HNF4A and high SOX9 (yellow). Scale bars: 10 microns. n Schematic model of single-injury data, depicting gradual SOX9 activation, collapsing lumens, and loss of HNF4A in organoid nephrons.