Mice with renal-specific alterations of stem cell-associated signaling develop symptoms of chronic kidney disease but surprisingly no tumors

Abstract

Previously, we found that Wnt and Notch signaling govern stem cells of clear cell kidney cancer (ccRCC) in patients. To mimic stem cell responses in the normal kidney in vitro in a marker-unbiased fashion, we have established tubular organoids (tubuloids) from total single adult mouse kidney epithelial cells in Matrigel and serum-free conditions. Deep proteomic and phosphoproteomic analyses revealed that tubuloids resembled renewal of adult kidney tubular epithelia, since tubuloid cells displayed activity of Wnt and Notch signaling, long-term proliferation and expression of markers of proximal and distal nephron lineages. In our wish to model stem cell-derived human ccRCC, we have generated two types of genetic double kidney mutants in mice: Wnt-β-catenin-GOF together with Notch-GOF and Wnt-β-catenin-GOF together with a most common alteration in ccRCC, Vhl-LOF. An inducible Pax8-rtTA-LC1-Cre was used to drive recombination specifically in adult kidney epithelial cells. We confirmed mutagenesis of β-catenin, Notch and Vhl alleles on DNA, protein and mRNA target gene levels. Surprisingly, we observed symptoms of chronic kidney disease (CKD) in mutant mice, but no increased proliferation and tumorigenesis. Thus, the responses of kidney stem cells in the tubuloid and genetic systems produced different phenotypes, i.e. enhanced renewal versus CKD.

Fig 1. Long-term tubuloid cultures from adult mouse kidneys.

(A) Brightfield images acquired on consecutive days of an tubuloid formed from a single freshly seeded cell. (B) H&E staining showing freshly seeded tubuloids with cystic (left), solid (middle) and alveolar (right) morphology. (C) Brightfield images of a freshly seeded (left) and a once passaged (right) tubuloid culture grown from 10⁵ single cells. (D) Quantification of tubuloid numbers in a freshly seeded and passage 1 culture. (E) Brightfield image of a serially passaged tubuloid culture after 5 passages. Data information: scale bars, 100 μm in A, 50 μm in B, 500 μm in C and E. 20 freshly seeded tubuloid cultures were established in total. In A, one independent replicate was examined. For quantification in B, 180 tubuloids were counted in total. One independent replicate was examined. For quantification in C and D, 10⁵ cells were seeded and tubuloids with diameters ≥ 100 μm were counted after 14 (freshly seeded cells) or 7 (passaged cells) days of culture (6 wells, technical replicates). One independent replicate was examined. In D, the graph shows mean ± SD (error bars). Data passed the Shapiro-Wilk normality test (α = 0.05). The unpaired, two-tailed Student’s t-test was performed; P-value, < 0.0001, ****p < 0.0001. In E, three independent replicates were examined.

Fig 2. Kidney tubuloids closely resemble adult mouse kidney epithelia, are phenotypically stable over time, and display enhanced proliferation and stem cell-associated Wnt and Notch signaling.

(A) Heatmap for the Pearson correlation matrix (coefficient) of the expression levels of 9,000 proteins between all samples of mouse kidney epithelia (control kidney, CK), early passage tubuloids (OEP) and long-term passage tubuloids (OLP). The experimental columns were correlated against each other. (B) Heatmap for the Pearson correlation matrix (coefficient) of the phosphorylation levels of 16,000 corresponding phosphorylation sites between all samples of mouse kidney epithelia (control kidney, CK), early passage tubuloids (OEP) and long-term passage tubuloids (OLP). The experimental columns were correlated against each other. (C) Functional annotation clustering of the most upregulated proteins in tubuloids using DAVID bioinformatic tool. Shown are 10 most enriched clusters, based on the overall enrichment scores, with their representative terms (biological processes), based on the adjusted P-values. (D) Proteomic heatmap for the typical proliferation markers Ki67, Cyclin D1 (Ccnd1) and Pcna in both early (OEP) and long-term (OLP) passage tubuloids in comparison to mouse kidney epithelia (control kidney, CK). (E) Proteomic heatmap for components of Wnt signaling in both early (OEP) and long-term (OLP) passage tubuloids in comparison to mouse kidney epithelia (control kidney, CK). (F) Proteomic heatmap for components of Notch signaling in both early (OEP) and long-term (OLP) passage tubuloids in comparison to mouse kidney epithelia (control kidney, CK). Data information: in D-F, the heatmaps show normalized log2 intensity values for three independent replicates of CK, OEP and OLP. A 5% FDR (adjusted P-value < 0.05) cutoff and a log2 fold change cutoff of > 0 were applied for both OEP over CK and OLP over CK. The values were scaled (z-score by row) with breaks from ≤ -2 to ≥ 2.

Fig 3. Kidney tubuloids exhibit markers of differentiated and polarized tubular epithelial cells.

(A) Immunohistochemistry for Pax8 (brown) of an tubuloid culture. (B) 2D immunofluorescence for E-cad (orange) of an tubuloid culture. (C) Proteomic heatmaps for Krt8 and Cldn4 in both early (OEP) and long-term (OLP) passage tubuloids in comparison to mouse kidney epithelia (control kidney, CK; upper part), and real-time qPCR analysis of gene expression of Krt18 in both early (OEP) and long-term (OLP) passage tubuloids in comparison to whole mouse kidney cells (control kidney, CK; lower part). (D) 2D immunofluorescence for Laminin (green) and Zo-1 (red) of an tubuloid culture. (E) 2D immunofluorescence for LTL (red) and PNA (green) of an tubuloid culture. (F) Proteomic heatmap for 48 membrane transporters in both early (OEP) and long-term (OLP) passage tubuloids in comparison to mouse kidney epithelia (control kidney, CK). (G) Proteomic heatmaps for Wnk1 and Gata3 in both early (OEP) and long-term (OLP) passage tubuloids in comparison to mouse kidney epithelia (control kidney, CK). (H) 2D immunofluorescence for Abcb1b (cyan), LTL (red) and PNA (green) of an tubuloid culture. (I) 2D immunofluorescence for Aqp3 (magenta), LTL (red) and PNA (green) of an tubuloid culture. Data information: scale bars in A/B/D/E/H/I, 50 μm. In A, nuclei are counterstained with haematoxylin; in B/D/E/H/I, nuclei are counterstained with DAPI. In C, the graph shows mean ± SD (error bars). The ordinary one-way ANOVA followed by the Dunnett’s multiple comparison to CK were performed; both P-values, 0.0001, ****p < 0.0001. In C/F/G, the heatmaps show normalized log2 intensity values for three independent replicates of CK, OEP and OLP. A 5% FDR (adjusted P-value < 0.05) cutoff and a log2 fold change cutoff of > 0 were applied for both OEP over CK and OLP over CK. Values were scaled (z-score by row) with breaks from ≤ -2 to ≥ 2. In A/B/D/E/H/I, three independent replicates were examined. For the real-time qPCR analysis of Krt18 in C, three independent replicates in technical triplicates were examined.

Fig 4. Wnt signaling controls kidney tubuloid formation, growth and differentiation.

(A) Tubuloid numbers per well with indicated growth factor combinations (withdrawal or addition). (B) Diameters (μm) of tubuloids (single dots) from wells with indicated growth factor combinations (withdrawal or addition). (C) Scheme of Wnt signaling with a common target for the small-molecule inhibitors ICG-001 and LF3. (D and F) Brightfield images of tubuloid cultures at indicated ICG-001 (D) and LF3 (F) concentrations (μM) on day 6 of treatment. (E and G) Curves showing the dependence of tubuloid viability (fold change ATP luminescence) on log10 ICG-001 (E) and LF3 (G) concentration (μM) on day 6 of treatment. (H) H&E staining of tubuloid cultures upon R-spondin-1 removal or treatment with ICG-001 at IC50. (I) Real-time qPCR analysis of gene expression of differentiation markers Krt8, Krt18 and Cldn4 in tubuloids upon R-spondin-1 removal or treatment with ICG-001 at IC50. (J) Real-time qPCR analysis of gene expression of markers of proximal tubular cells, Abcb1b, Slc3a1 and Slc40a1, in tubuloids upon R-spondin-1 removal or treatment with ICG-001 at IC50. (K) Real-time qPCR analysis of gene expression of markers of distal nephrons, Aqp3, Atp6ap2 and Wnk1, in tubuloids upon R-spondin-1 removal or treatment with ICG-001 at IC50. (L) 2D immunofluorescence for Aqp3 (magenta) in tubuloids upon R-spondin-1 removal or treatment with ICG-001 at IC50. Data information: in A and B, culture medium containing DMEM/F12 with GlutaMax, HEPES, N-acetylcysteine, B27, N2, nicotinamide, hydrocortisone (HC) and prostaglandin E2 (PGE2) is common to all the conditions. For quantification, 10⁵ cells were freshly seeded and tubuloids with diameters ≥ 200 μm were counted after 14 days of culture. The graphs show mean ± SD (error bars). In A, data passed the Shapiro-Wilk normality test (α = 0.05). The ordinary one-way ANOVA followed by the Dunnett’s multiple comparison to complete tubuloid medium (Ctrl, first from the left) were performed; P-values from left to right, 0.0001, 0.2132; ****p < 0.0001, ns: non-significant. In B, data did not pass the Shapiro-Wilk normality test (α = 0.05). The alternative non-parametric Kruskal-Wallis test followed by the Dunn’s multiple comparison to complete tubuloid medium (Ctrl, first from the left) were performed; P-values from left to right, 0.0090, 0.0129; **p < 0.01, *p < 0.05. In A and B, three independent replicates with 6 technical replicates (wells) were examined and collectively shown. Ctrl in D-G, tubuloids cultured in complete medium and treated with DMSO at re-spective concentrations. Scale bars in D and F, 500 μm. In E and G, the curves show mean ± SD (error bars). In E, ICG-001 produced the IC50 of 14.66 μM; in G, LF3 produced the IC50 of 15.72 μM. In D-G, three independent replicates in technical triplicates (wells) were examined. Ctrl in H-L, tubuloids cultured in complete medium and treated or not with DMSO at a respective concentration. Scale bars in H, 100 μm. Three independent replicates were examined. In I-K, the graphs show mean ± SD (error bars). The ordinary one-way ANOVA followed by the Dunnett’s multiple comparison to Ctrl were performed; P-values from left to right in I; 0.0004, 0.0001 for Krt8; 0.0153, 0.0002 for Krt18; 0.0002, 0.0003 for Cldn4; in J: 0.0001, 0.0001 for Abcb1b; 0.0005, 0.0164 for Slc3a1; 0.0910, 0.0001 for Slc40a1; in K: 0.0001, 0.0001 for Aqp3; 0.0014, 0.0001 for Atp6ap2; 0.001, 0.0001 for Wnk1; ****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05, ns: non-significant. Three independent replicates in technical triplicates were examined. Scale bars in L, 50 μm. Three independent replicates were examined.

Fig 5. Generation of double kidney mutants in mice.

(A) Scheme of the genetic system to generate mutant kidneys. (B) Time course and pattern of recombination in kidneys of Pax8-rtTA(+); LC1-Cre(+); LacZ(+) reporter mice. (C) PCR for DNA recombination of the loxP-flanked exon 3 sequence of β-catenin gene in β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant kidneys (700 bp) versus controls (900 bp). (D) PCR for DNA recombination of the loxP-flanked promoter-exon 1 sequence of Vhl gene in Vhl-LOF; β-catenin-GOF mutant kidneys (260 bp) versus controls (460 bp). (E) Immunoblotting for recombination of β-catenin in β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant kidneys (72 kDa) versus controls (95 kDa). (F) Immunohistochemistry for nuclear N1icd in β-catenin-GOF; Notch-GOF mutant cortical kidneys versus controls. (G) Fold change of expression of selected β-catenin target genes in β-catenin-GOF; Notch-GOF mutant kidneys versus controls. (H) Fold change of expression of selected β-catenin target genes in Vhl-LOF; β-catenin-GOF mutant kidneys versus controls. (I) Fold change of expression of selected Notch target genes in β-catenin-GOF; Notch-GOF mutant kidneys versus controls. (J) Fold change of expression of selected Hif-1α/2α target genes in Vhl-LOF; β-catenin-GOF mutant kidneys versus controls. Data information: abbreviation in A; dox, doxycycline. Scale bars in B, 100 μm. In the upper panel, nuclei are counterstained with haematoxylin; in the lower panel, nuclei are counterstained with nuclear fast red. Glomeruli are marked by black lines. Two independent replicates of Pax8-rtTA(+); LC1-Cre(+); LacZ(+) mice and one of Pax8-rtTA(-); LC1-Cre(+); LacZ(+) controls were examined. Abbreviations in C-E; wt, wild-type; ex3 del, exon 3 deletion; prom-ex1 loxP/del, promoter-exon 1 loxP/deletion. Scale bars in F, 100 μm. Nuclei are counterstained with haematoxylin. In C-F, three independent replicates per line were examined. In G-J, graphs show mean + SD (error bars). Three independent replicates (mice, dots) in technical triplicates were analyzed. No statistical test was performed because of high inter-mouse variance. However, increase trends were observed.

Fig 6. No tumorigenesis was observed in double mutant kidneys.

(A) Body weight of β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant mice versus controls. (B) Overview of kidneys and spleens of β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant mice versus controls. (C) Kidney weight to body weight ratio of β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant mice versus controls. (D) Spleen weight to body weight ratio of β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant mice versus controls. (E) PAS staining in β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant kidneys versus controls (in the cortex and medulla) after 8 months after the doxycycline pulse. (F) PAS-stained outer part of the cortex with nuclear crowding (arrowheads) in some tubules in β-catenin-GOF; Notch-GOF, but not in Vhl-LOF; β-catenin-GOF mutant kidneys. (G) Immunohistochemistry for nuclear Ki67 in β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant cortical kidneys versus controls after 8 months after the doxycycline pulse. (H) Quantification of Ki67-positive nuclei in β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant cortical kidneys versus controls after 8 months after the doxycycline pulse. Data information: in A, C and D, 5 independent replicates (mice, dots) per line were examined. Graphs show mean ± SD (error bars). Data passed the Shapiro-Wilk normality test (α = 0.05). The ordinary one-way ANOVA followed by the Dunnett’s multiple comparison to controls was performed; P-values from left to right in A; 0.0001, 0.0002; ****p < 0.0001, ***p < 0.001; in C; 0.4879, 0.8354; ns: non-significant; in D; 0.0001, 0,0229; ****p < 0.0001, *p < 0.05. Scale bar in B, 1 cm. 20 independent replicates (all mice induced) per line were examined. Scale bars in E-G, 100 μm. Nuclei are counterstained with haematoxylin. In E and G, insets are enlarged on the right. 20 independent replicates (all mice induced) per line were examined. In H, three independent replicates (mice, dots) with 10 technical replicates (fields per kidney) each per line were examined. Graphs show mean + SD (error bars). Data passed the Shapiro-Wilk normality test (α = 0.05). The ordinary one-way ANOVA followed by the Dunnett’s multiple comparison to controls was performed; P-values from left to right, 0.6019, 0.9871; ns: non-significant.

Fig 7. Mutant mice displayed phenotypes of chronic kidney disease (CKD).

(A) Hematocrit of β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant mice versus controls. (B) Concentration of plasma EPO in β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant mice versus controls. (C) PAS staining showing extramedullary hematopoiesis in spleens of β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant mice versus controls. Megakaryocytes are marked by arrowheads. (D) Fold change of gene expression of kidney injury markers in β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant kidneys versus controls. (E) Concentration of blood urea nitrogen (BUN) in β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant mice versus controls. (F) Fold change of gene expression of inflammation markers in β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant kidneys versus controls. (G) Fold change of gene expression of fibrotic markers of myofibroblasts and extracellular matrix in β-catenin-GOF; Notch-GOF and Vhl-LOF; β-catenin-GOF mutant kidneys versus controls. Data information: in A, 5 independent replicates (mice, dots) with one technical replicate each per line were examined. Graphs show mean ± SD (error bars). Data passed the Shapiro-Wilk normality test (α = 0.05). The ordinary one-way ANOVA followed by the Dunnett’s multiple comparison to controls was performed; P-values from left to right, 0.0001, 0.0001; ****p < 0.0001. In B, 5 independent replicates (mice, dots) in technical triplicates per line were examined. Graphs show mean ± SD (error bars). No statistical test was performed because of high inter-mouse variance. However, increase trends were observed. Scale bars in C, 100 μm. Nuclei are counterstained with haematoxylin. An undisrupted follicular B-cell compartment in the white pulp in controls is marked by white lines. Insets are enlarged on the right. Three independent replicates per line were examined. In D, three independent replicates (mice, dots) in technical triplicates were examined. Graphs show mean + SD (error bars). No statistical test was performed because of high inter-mouse variance. However, increase trends were observed. In E, 5 independent replicates (mice, dots) with one technical replicate each per line were examined. Graphs show mean ± SD (error bars). Data passed the Shapiro-Wilk normality test (α = 0.05). The ordinary one-way ANOVA followed by the Dunnett’s multiple comparison to controls was performed; P-values from left to right, 0.0109, 0.0888; *p < 0.05, ns: non-significant. Although the difference in average concentration of BUN between the Vhl-LOF; β-catenin-GOF mutant mice and controls was non-significant (p < 0.05), an increase trend was observed. In F and G, three independent replicates (mice, dots) in technical triplicates were examined. Graphs show mean + SD (error bars). No statistical test was performed because of high inter-mouse variance. However, increase trends were observed in F and no increase trends were observed in G.