3D organoid-derived human glomeruli for personalised podocyte disease modelling and drug screening

Abstract

The podocytes within the glomeruli of the kidney maintain the filtration barrier by forming interdigitating foot processes with intervening slit diaphragms, disruption in which results in proteinuria. Studies into human podocytopathies to date have employed primary or immortalised podocyte cell lines cultured in 2D. Here we compare 3D human glomeruli sieved from induced pluripotent stem cell-derived kidney organoids with conditionally immortalised human podocyte cell lines, revealing improved podocyte-specific gene expression, maintenance in vitro of polarised protein localisation and an improved glomerular basement membrane matrisome compared to 2D cultures. Organoid-derived glomeruli retain marker expression in culture for 96 h, proving amenable to toxicity screening. In addition, 3D organoid glomeruli from a congenital nephrotic syndrome patient with compound heterozygous NPHS1 mutations reveal reduced protein levels of both NEPHRIN and PODOCIN. Hence, human iPSC-derived organoid glomeruli represent an accessible approach to the in vitro modelling of human podocytopathies and screening for podocyte toxicity.

Fig. 1

Intact glomeruli can be isolated from human iPSC-derived kidney organoids. a Top left: Pure populations of whole glomeruli can be isolated from kidney organoids at scale by sieving. Scale bar 200 µm. Right: These organoid-derived glomeruli (OrgGloms) when examined in detail (scale bar 100 µm) show similarity to adult human glomeruli (Bottom left, scale bar 100 µm). b OrgGloms are formed in the appropriate size range, comparable to both human glomeruli at 32 weeks gestation, and adult human glomeruli. One-way ANOVA p < 0.0001; error bars = SEM; significant difference assessed by Tukey’s multiple comparisons test; F-value = 176.9; DF = 2; biological replicates n = 30 adult glomeruli, n = 30 neonatal glomeruli, n = 155 OrgGlom. c Glomeruli developed within kidney organoids show comparable morphology to capillary loop stage human glomeruli, highlighted by in situ immunofluorescent staining of serial sections with the podocyte protein NEPHRIN, which shows infolding of the glomerular surface layer. Scale bar 10 µm, central cross-section (left), surface (right). d Quantitative PCR analysis of isolated glomerular fractions compared to whole organoid show a significant enrichment in NPHS1 gene expression in sieved glomeruli. Error bar = SEM; significant difference assessed by Student’s unpaired t-test p = 0.0002, t-value = 13, DF = 4, n = 3 biological replicates shown by symbols. e Transmission electron microscopy of kidney organoid glomeruli show podocyte cell bodies connected by primary and secondary processes (inset), surrounded by an outer layer of parietal epithelial cells (arrowhead). Scale bar 10 µm. f Immunostaining of whole OrgGloms shows appropriate apicobasal cell polarity. A single cross-sectional plane is shown on the left, and a 3D reconstruction using all Z-stack images acquired (444 sections) on the right. Expression levels of single channels shown in greyscale to preserve maximum contrast, merged images shown in colour. Full 3D reconstruction video of the Z-stack can be found as Supplementary Video 1

Fig. 2

Primary podocytes can be cultured from kidney organoid glomeruli. a Isolated organoid glomeruli show evidence of podocyte cell migration (OrgPods) displaying thin arborized projections (TAPs) (inset). Inverted image shown to provide maximum contrast, scale bar 100 µm. b TAPS from newly emerged podocytes are composed of F-actin shown by phallodin immunofluorescent staining. Inverted image, scale bar 50 µm. c Immunostaining at 36 h post-plating shows a strong positively stained 3D OrgGlom with a migrating 2D OrgPod population. Left panel 2D images, right panel 3D reconstruction of Z-stack. Scale bars 50 µm. d At 48 h post-plating OrgPods display a flattened, arborized morphology with processes connecting adjacent cells (arrow), scale bar 50 µm. e Immunostaining of ciPods for SYNAPTOPODIN showed expression is absent in undifferentiated cells (ciPod: Un), only becoming evident following 14 days induced differentiation at 37 °C (ciPod: Diff). OrgPods also display strong SYNAPTOPODIN protein expression, aligned with F-actin stress fibres. Scale bars 100 µm. f OrgPods express the neonatal Fc receptor (FcRN) and actively endocytose fluorescein isothiocyanate (FITC)-labelled albumin at 37 °C resulting in FITC-accumulation in endosomes on the cell surface. This is process halted when performed at 4 °C. Scale bars 50 µm. g OrgPods stimulated with insulin (10 mg/ml) for 10 min showed cortical reorganisation of their actin cytoskeleton with GLUT4 translocation from a vesicular to plasma membrane localisation. Scale bars 50 µm. All representative images reflect a minimum of three biological replicates. For immunofluorescence, images are shown in greyscale for single channels, and merged images in colour

Fig. 3

Organoid-derived glomeruli display improved podocyte identity. a Principle component analysis of RNA sequencing (RNA-Seq) data was performed on three biological replicates of OrgGloms, OrgPods and ciPods in both differentiated and undifferentiated states and compared to previously published data (Lab 2). b Venn diagram displaying the intersections of each comparison, upregulated genes are shown in red and downregulated genes are shown in blue. This shows the greatest number statistically significant upregulated genes were identified in the OrgGloms versus differentiated ciPods. There was little to no overlap of differentially expressed genes in differentiated ciPods versus undifferentiated ciPods compared with all other pairwise comparisons. c Heatmap showing the top 50 differentially expressed genes of each sample triplicate. A stark contrast in expression levels between the 3D OrgGloms and the 2D podocyte cultures is observed, particularly in genes associated with the podocyte. Log-normalised gene expression levels depicted. d Gene Ontology (GO) enrichment analysis of the top 100 upregulated genes differentially expressed between OrgGloms and differentiated ciPods found enrichment of GO terms associated with developmental processes, and slit diaphragm components including genes associated with podocyte foot processes and those of the podocyte actin cytoskeleton. The top 10 most statistically significant GO term categories are shown. P-value of 0.05 depicted as dotted line. e Heatmap representation of key podocyte-associated genes showed low expression levels in both undifferentiated and differentiated ciPods, regardless of the laboratory in which they were cultured. By contrast, significantly elevated gene expression levels were displayed in the OrgGlom samples. Log-normalised gene expression levels depicted. f Quantitative PCR analysis of key podocyte-associated genes validated the RNA-seq data showing enhanced gene expression in OrgGloms compared to 2D podocyte models. Two-way ANOVA p < 0.0001; error bars = SEM; significant difference was assessed by Tukey’s multiple comparisons test; n = 3 biological replicates shown by symbols; F-value (Interaction) = 58.22, DF (Interaction) = 21; F-value (Gene) = 56.82, DF (Gene) = 7; F-value (Cell Type) = 544.4, DF (Cell Type) = 3. g A glomerular expression score was determined for each sample by calculating the average log expression across the top 100 upregulated genes identified from a human renal glomerulus-enriched gene expression dataset (GSE21785). OrgGlom samples were found to have the highest scores for this gene set, showing greatest congruence to human glomerular isolates

Fig. 4

Organoid glomeruli display a maturing GBM matrisome. a Serial chemical fractionation of organoid glomeruli (OrgGlom) and organoid proximal tubules (OrgPT) to derive fractions enriched for cellular and extracellular matrix (ECM) components, where C1 and C2 are predominantly cellular proteins, C3 nuclear proteins, and C4 enriched for ECM proteins. Representative blot showing one of the three biological replicates for each cell type. b Principal component analysis of mass spectrometry data from C1 and C4 fractions of both isolated glomeruli and tubules. c Mapping of organoid proteomic data onto the human matrisome database and the identification of 60 ECM proteins (Human matrisome resource: MatrisomeDB). d, e The expression profile of matrix proteins detected in OrgGloms and OrgPTs in the ECM enriched fraction (C4) and cellular fraction (C1). Gene ontology (GO) classifications (GO terms are based on the MatrisomeDB Gene Ontology divisions and categories) were used to group those detected into core matrix proteins (d), or proteoglycans and matrix-associated proteins (e). Error bars = SEM, n = 3 biological replicates. f Immunostaining of isolated OrgGloms shows basal expression of the GBM protein Laminin-α5. Representative image from >3 biological replicates. Expression levels of single channels shown in greyscale to preserve maximum contrast, merged image shown in colour. Scale bar 10 µm. g A comparison of matrix proteins identification in human organoid glomeruli (OrgGlom) versus: human glomerular matrix (GM), immortalised human glomerular endothelial cells (ciGEnC) (doi: 10.1681/ASN.2013070795), immortalised human podocytes (ciPod) (doi: 10.1681/ASN.2013030233) and GEnC and podocyte co-culture (doi: 10.1681/ASN.2013070795) based on previously reported data. Red bar denotes presence of this protein in the sample

Fig. 5

Temporal analysis of MAFB-expressing cells within organoids. a Differentiation of MAFB-BFP2 iPSC into kidney organoids was successful with blue fluorescent protein 2 (BFP2) expression observed in live organoids from d7 + 7 onwards. Scale bar 1000 µm. b FACS plot showing the BFP2-positive cell population isolated at d7 + 10. Representative plot of three biological replicates. c RT-PCR analysis of BFP2-positive and negative organoid cell fractions showed MAFB expression is only found in BFP2-positive cells. d Brightfield and BFP2-fluorescent live imaging of glomeruli isolated from d7 + 18 MAFB-reporter organoids at the time of plating and after 24 h culture. Strong BFP2 signal is observed within glomeruli when in suspension (0 h and 24 h arrow), but does not remain active when the glomeruli are adhered for culture, nor is it expressed in the migrating podocyte population (inset). Scale bar 200 µm. e A principle component analysis of RNA sequencing (RNA-Seq) data was performed on three biological replicates for d7 + 10, d7 + 14 and d7 + 19 BFP2-positive cell populations and compared to OrgPod samples. A clear separation between the 2D cultured organoid podocytes and MAFB BFP2-positive cells isolated from whole organoids was observed in dimension 1; in dimension 2 the variable of time was evident in the separation of the triplicates. f Venn diagram displaying the intersections of each comparison, upregulated genes are shown in red and downregulated genes are shown in blue. This shows the greatest number of statistically significant upregulated genes was identified in the OrgGlom (d7 + 19) vs immature podocyte (d7 + 10) MAFB-BFP2 population. g Heatmap showing the top 50 upregulated differentially expressed genes between d7 + 19 and d7 + 10, with many enriched genes found to transcribe proteins in the human glomerular ECM proteome. Fold change of log-normalised gene expression levels for each of the triplicate samples presented. h Gene Ontology (GO) enrichment analysis of the top 100 upregulated genes differentially expressed between d7 + 19 and d7 + 10 found enrichment of GO terms associated with extracellular matrix (ECM) maturation, collagen maturation and cell adhesion. The top 10 most statistically significant GO term categories are shown. P-value of 0.05 depicted as a dotted line. i Top 15 most-significantly upregulated genes with time, identified using a human renal glomerulus-enriched gene expression dataset. Significant upregulation of genes associated with the maturation of additional glomerular cells, including endothelium and mesangial cells alongside specific GBM components and associated proteoglycans. Two-way ANOVA p < 0.0001; error bars = SEM; significant difference between time points assessed by Tukey’s multiple comparisons test; n = 3 biological replicates shown by symbols; F-value(Interaction) = 27.16, DF (Interaction) = 28; F-value (Gene) = 37.58, DF (Gene) = 14; F-value(Time) = 495.7, DF (Time) = 2

Fig. 6

Organoid Glomeruli model of congenital nephrotic syndrome in vitro. a Description of the NPHS1 variants identified in the patient modelled, diagnosed with congenital nephrotic syndrome (CNS). b–d Immunostaining of OrgGloms isolated from control organoids and CNS patient organoids show reduced NEPHRIN and PODOCIN protein levels in the organoids derived from patient-iPSC, representative images shown of >3 biological replicates. Scale bars 10 µm. e Higher power immunofluorescent images show the polarised co-localisation of NEPHRIN with NEPH1 (solid white arrowheads) and PODOCIN in control OrgGloms. This polarisation is lost in CNS OrgGloms due to the absence of NEPHRIN (white arrows). Scale bars 10 µm. f Quantitative analysis of fluorescence intensities from independent OrgGlom biological replicates performed using one control and two distinct patient-derived CNS iPSC clones. Organoid glomeruli generated from both patient-derived iPSC clones show significant reduction in NEPHRIN and PODOCIN protein levels. Two-way ANOVA p < 0.0001; error bars = SEM. Biological replicates. NEPHRIN (controls, n = 20; CNS, n = 56); PODOCIN (controls, n = 14; CNS, n = 22); CD2AP (controls, n = 8; CNS, n = 15); NEPH1 (controls, n = 10; CNS, n = 17). Significant difference assessed by Sidak’s multiple comparisons test between cell lines; F-value = 112; DF = 1. NEPHRIN: control vs CNS#1, p < 0.0001; control vs CNS#2, p < 0.0001; CNS#1 vs CNS#2, p > 0.9999. PODOCIN: control vs CNS#1, p < 0.0001; control vs CNS#2, p < 0.0001; CNS#1 vs CNS#2, p = 0.9995. CD2AP: control vs CNS#1, p = 0.0007; control vs CNS#2, p = 0.0016; CNS#1 vs CNS#2, p = 0.9980. NEPH1: control vs CNS#1, p = 0.5320; control vs CNS#2, p = 0.9994; CNS#1 vs CNS#2, p = 0.9992. g Quantitative western blot analysis of NEPHRIN and PODOCIN protein levels within independent biological replicates confirms the significant depletion of these proteins in OrgGloms derived from CNS iPSCs

Fig. 7

Toxicity screening using cultured organoid glomeruli. a MAFB-BFP2 OrgGloms can be cultured in isolation in a 96-well format. Live imaging of glomeruli at 48 h post-treatment with doxorubicin showed a dose-dependent decrease of BFP2-reporter intensity. b Fixed glomeruli immunolabelled with Caspase-3 showed activation of this pro-apoptotic pathway following doxorubicin treatment. Scale bar 100 µm. c Semi-quantitative analysis of BFP2-reporter intensity (blue) alongside OrgGlom diameter (red) showed a dose-dependent reduction in both glomerular diameter and BFP2-reporter intensity. One-way ANOVA p < 0.0001; error bars = SEM; significant difference between sequential doses assessed by Tukey’s multiple comparisons test. BFP2-reporter intensity: F-value = 80.3; DF = 5; biological replicates n = 16 per dose. OrgGlom Diameter: F-value = 60.94; DF = 5; biological replicates n = 16 per dose