Disrupted glucocorticoid receptor cell signalling causes a ciliogenesis defect in the fetal mouse renal tubule

Abstract

Primary cilia are cell signalling and environment sensing organelles and have important roles during embryogenesis and homeostasis. We demonstrate glucocorticoid signalling is essential for normal cilia formation in mouse and human renal tubules. RNA sequencing of E18.5 kidneys from glucocorticoid receptor (GR) null mice identified significant reductions in key ciliogenesis-related genes including Ccp110, Cep97, Cep290 and Kif3a. Confocal microscopy reveals abnormal, stunted cilia on proximal tubules, podocytes, and collecting duct cells in mice with global or conditional deletion of GR. In contrast, activation of GR signalling with dexamethasone in human kidney organoids or mouse IMCD3 cells increases cilia length, an effect blocked by the GR antagonist RU486. Analysis of GR-null kidney extracts demonstrates reduced levels of pERK and SUFU identifying potential cell pathway crosstalk with GR signalling that coordinately regulate ciliogenesis in the renal tubule. Finally, dexamethasone reduces Aurora kinase A levels, a factor driving cilia disassembly and implicated in the pathogenesis of polycystic kidney disease.

Keywords: Ciliogenesis; Ciliopathies; Glucocorticoid Receptor Signaling; Glucocorticoids; Kidney Development.

Figure 1. Glucocorticoid receptor gene expression in the fetal kidney during embryonic development.

(A) Average expression of GR across the cell clusters at E13.5, E15.5 and E18.5. (B) tSNE plot showing cell clusters within the kidney at E13.5. (C) tSNE plot showing the expression of GR, purple within the cell clusters at E13.5. (D) tSNE plot showing cell clusters within the kidney at E15.5. (E) tSNE plot showing the expression of GR, purple within the cell clusters at E15.5. (F) tSNE plot showing cell clusters within the kidney at E18.5. (G) tSNE plot showing the expression of GR, purple within the cell clusters at E18.5. (H) Immunofluorescence of glucocorticoid receptor (GR) in the fetal kidney during embryonic development at E18.5. Sections were stained Hoechst (blue, nucleus), Dolichos Biflorus Agglutinin (DBA) (green, collecting duct), Lotus Tetragonolobus Lectin (LTL) (green, proximal tubule), nephrin (NPHS1) (green, podocyte), MEIS123 (green, stroma) and GR (red, GR). Slides were imaged with a Zeiss LSM 980 confocal microscope (×63 objective, 2× digital zoom), scale bar represents 20 μm. All images are representative of n = 4 animals per experimental group. Source data are available online for this figure.

Figure 2. Transcriptome analysis of E18.5 fetal kidney RNA from GR-null and control mice.

(A) NGS RNA-seq was performed on total RNA isolated from control and GR-null mouse kidneys at E18.5. Heatmaps generated from differentially expressed genes (LogFC >1 and FDR < 0.05) log2-CPM values. The number of suppressed genes is four times more than augmented genes in GR-null compared to control. Data from control (n = 4) and GR-null (n = 3) animals per experimental group. (B) Principal component plot showing separation of control (red) and GR-null (blue) kidney at E18.5. Data from control (n = 4) and GR-null (n = 3) animals per experimental group. (C) Volcano plot of top differentially expressed genes with a significant FDR < 0.05 and log-fold change >1.5 and < −2. Red dots represent gene mRNA levels that are significantly increased, and blue dots represent genes with significantly decreased. Dashed lines show the LogFC 0.5 and FDR 0.05 cut-offs. Data from control (n = 4) and GR-null (n = 3) animals per experimental group. (D) Volcano plot of differentially expressed ciliary genes with a significant FDR < 0.05. Red dots represent gene mRNA levels that are significantly increased, and blue dots represent genes with significantly decreased. Dashed lines show the LogFC 0.5 and FDR 0.05 cut-offs. Data from control (n = 4) and GR-null (n = 3) animals per experimental group. (E) mRNA levels of eleven target genes identified from NGS RNA-seq in fetal GR-null mouse kidney at E18.5. Crisp1 (fold −9.6, P = 0.00030), Gbp7 (fold −3.7, P = 0.012), Ifi203 (fold −5.6, P = 0.00073), Ifit1 (fold −8.6, P = 0.024), Kap (fold −9.2, P = 0.0024), Mndal (fold −3.8, P = 0.00046), Oas2 (fold −2.6, P = 0.016), Oasl2 (fold −4.6, P = 0.0062) and S100g (fold −2.7, P = 0.00094) were downregulated. Consistent with RNA sequencing Hba-A1 (fold 3.9, P = 0.049) and Igf2 (fold 1.7, P = 0.071) were upregulated. The mRNA levels are expressed relative to mRNA levels of the housekeeping gene Rps29. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null samples, n = 3 animals per experimental group for each gene. (F) mRNA levels of six primary cilia structural genes in fetal GR-null mouse kidney at E18.5. Ccp110 (fold −2.2, P = 0.011), Cep97 (fold −1.8, P = 0.048), Cep290 (fold −2.9, P = 0.012), Kif3a (fold −1.8, P = 0.026), Rab8a (fold 1.1, P = −0.80) and Rpgr (fold −1.9, P = 0.026). The mRNA levels are expressed relative to mRNA levels of the housekeeping gene Rps29. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control (n = 3–4) and GR-null (n = 3) animals for each gene. (G) mRNA level of three primary cilia regulatory genes in fetal GR-null mouse kidney at E18.5. Atat1 (fold −1.19, P = 0.57), Aurka (fold 1.32, P = 0.19) and Nedd9 (fold 1.16, P = 0.79). The mRNA levels are expressed relative to mRNA levels of the housekeeping gene Rps29. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null, n = 3 animals for each gene. (H) Expression of target genes within different cell types of the kidney at E18.5. Scale represents average expression, generated from single cell dataset. Source data are available online for this figure.

Figure 3. Glucocorticoid regulation of primary cilia length on GR-null fetal kidney proximal tubule cells at E18.5.

(A) Immunofluorescence of primary cilia morphology in control and GR-null proximal tubules at E18.5. Sections were stained with Hoechst (blue, nucleus), acetylated tubulin (AceTub) (green, microtubules), pericentrin (PCNT) (red, basal body) and Lotus Tetragonolobus Lectin (LTL) (grey, proximal tubule). White arrows indicate primary cilia. Slides were imaged with a Zeiss LSM 980 confocal microscope (×63 objective, 2× digital zoom), scale bar represents 20 µm. All images are representative of n = 4 animals per experimental group. (B) Proximal tubule primary cilia length was measured using Imaris software. Four images were taken per an animal, 574 primary cilia were measured in the control and 712 primary cilia were measured in the GR-null mouse kidneys. Lines represent median and quartiles. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null (P = 0.0001), n = 4 animals per experimental group. (C) Percentage of proximal tubule primary cilia greater than 5 µm in control and GR-null proximal tubules. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null (P = 0.027), n = 4 animals per experimental group. (D) Percentage of ciliated proximal tubule cells. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null (P = 0.11), n = 4 animals per experimental group. (E) Scanning electron microscopy of primary cilia morphology in control and GR-null kidneys at E18.5. Kidneys were imaged with Nova NanoSEM 450 scanning electron microscope (Thermo Fisher Scientific) at a voltage of 2 kV and a spot size of 2. Black arrows indicate primary cilia. Scale bar represents 4 µm (control) and 5 µm (GR-null). All images are representative of n = 5 animals per experimental group. Source data are available online for this figure.

Figure 4. Glucocorticoid regulation of primary cilia length on GRcdKO postnatal kidney collecting duct cells at P11.

(A) Immunofluorescence of primary cilia morphology in control and GRcdKO collecting ducts at P11. Sections were stained with Hoechst (blue, nucleus), acetylated tubulin (AceTub) (green, microtubules), pericentrin (PCNT) (red, basal body) and Dolichos Biflorus Agglutinin (DBA) (grey, collecting duct). White arrows indicate primary cilia. Slides were imaged with a Zeiss LSM 980 confocal microscope (×63 objective, 2× digital zoom), scale bar represents 20 µm. All images are representative of n = 4 animals per experimental group. (B) Collecting duct primary cilia length was measured using Imaris software. Four images were taken per an animal, 1048 primary cilia were measured in the control and 957 primary cilia were measured in the GRcdKO mouse kidneys. Lines represent median and quartiles. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control vs GRcdKO (P = 0.0001), n = 3 animals per experimental group. (C) Percentage of collecting duct primary cilia greater than 5 µm in the control and GRcdKO mouse kidney at P11. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GRcdKO (P = 0.66), n = 3 animals per experimental group. (D) Percentage of ciliated collecting duct cells. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GRcdKO (P = 0.94), n = 3 animals per experimental group. Source data are available online for this figure.

Figure 5. Glucocorticoid regulation of primary cilia length on GR-null fetal kidney podocyte cells at E18.5 and induced pluripotent stem cell kidney organoid podocyte cells after dexamethasone treatment.

(A) Immunofluorescence of primary cilia morphology in control and GR-null podocytes at E18.5. Sections were stained with Hoechst (blue, nucleus), acetylated tubulin (AceTub) (green, microtubules) and Nephrin (NPHS1) (grey, podocyte). White arrows indicate primary cilia. Slides were imaged with a Zeiss LSM 980 confocal microscope (×63 objective, 2× digital zoom), scale bar represents 20 µm. All images are representative of n = 4 animals per experimental group. (B) GR-null podocyte primary cilia length was measured using Imaris software. Four images were taken per animal, 235 primary cilia were measured in control and 212 primary cilia were measured in GR-null. Lines represent median and quartiles. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null (P = 0.030). Data from control (n = 4) and GR-null (n = 3) animals per experimental group. (C) Percentage of primary cilia greater than 5 μm in control and GR-null podocyte cells. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null (P = 0.32). Data from control (n = 4) and GR-null (n = 3) animals per experimental group. (D) Percentage of ciliated podocyte cells. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null (P = 0.71). Data from control (n = 4) and GR-null (n = 3) animals per experimental group. (E) Immunofluorescence of primary cilia morphology in induced pluripotent stem cell (iPSC) kidney organoids. iPSC kidney organoids were treated with vehicle (veh) or dexamethasone (dex). Sections were stained with Hoechst (blue, nucleus), acetylated tubulin (AceTub) (green, microtubules), nephrin (NPHS1) (red, podocyte) and Lotus Tetragonolobus Lectin (LTL) (grey, proximal tubule). White arrows indicate primary cilia. Slides were imaged with a Zeiss LSM 980 confocal microscope (×63 objective, 2× digital zoom), scale bar represents 20 µm. All images are representative of n = 4 animals per experimental group. (F) Podocyte primary cilia length was measured using Imaris software. Four images were taken per animal, 1685 primary cilia were measured in vehicle (veh) treated organoids and 1638 primary cilia were measured in dexamethasone (dex) treated organoids. Lines represent median and quartiles. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between veh and dex treated cells (P = 0.0001), n = 3 biological replicates per experimental group. (G) Percentage of primary cilia greater than 4 μm in veh and dex treated organoids. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between veh and dex treated cells (P = 0.0093), n = 3 biological replicates per experimental group. (H) Percentage of ciliated podocyte cells. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between veh and dex treated cells (P = 0.99), n = 3 biological replicates per experimental group. Source data are available online for this figure.

Figure 6. Glucocorticoid regulation of primary cilia in IMCD3 cells after dexamethasone treatment.

(A) Immunofluorescence images of primary cilia morphology in IMCD3 cells. IMCD3 cells were treated with vehicle (veh), dexamethasone (dex), vehicle + RU486 (veh + RU486) or dexamethasone + RU486 (dex + RU486). Sections were stained with Hoechst (blue, nucleus), acetylated tubulin (AceTub) (green, microtubules) and gamma tubulin (γ-Tub) (red, basal body). White arrows indicate primary cilia. Slides were imaged with a Lecia SP8 confocal microscope (×63 objective, 2× digital zoom), scale bar represents 20 µm. All images represent n = 3 biological replicates per experimental group. (B) Primary cilia length was measured using Imaris software. Four images were taken per biological replicate, the number of individual cilia measured per a treatment group was 1698 vehicle, 1514 dexamethasone, 1966 vehicle + RU486 and 2143 dexamethasone + RU486. Lines represent median and quartiles. All data presented as mean ± SEM, significant differences were analysed by one-way ANOVA with multiple comparisons indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant. Veh vs dex (P = 0.0001), veh vs veh + RU486 (P = 0.0001), veh vs dex + RU486 (P = 0.0001), dex vs veh + RU486 (P = 0.0001), dex vs dex + RU486 (P = 0.0001), veh + RU486 vs dex + RU486 (P = 0.28). Data from n = 3 biological replicates per experimental group. (C) Percentage of primary cilia greater than 4 µm. All data presented as mean ± SEM, significant differences were analysed by one-way ANOVA with multiple comparisons indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant. Veh vs dex (P = 0.0026), veh vs veh + RU486 (P = 0.073), veh + dex + RU486 (P = 0.11), dex vs veh RU486 (P = 0.0001), dex vs dex + RU486 (P = 0.0002), veh + RU486 vs dex + RU486 (P = 0.99). Data from n = 3 biological replicates per experimental group. (D) Percentage of ciliated collecting duct cells. All data presented as mean ± SEM, significant differences were analysed by one-way ANOVA with multiple comparisons indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant. Veh vs dex (P = 0.59), veh vs veh + RU486 (P = 0.29), veh vs dex + RU486 (P = 0.046), dex vs veh + RU486 (P = 0.049), dex vs dex + RU486 (P = 0.0081), veh + RU486 vs dex + RU486 (P = 0.57). Data from n = 3 biological replicates per experimental group. Source data are available online for this figure.

Figure 7. Analysis of cell signalling pathways and cilia structural proteins by western blot analysis as targets for glucocorticoid regulation of primary ciliogenesis.

Western blot analysis of signaling pathway proteins (A) P-AKT and AKT (P = 0.067), (B) P-AMPKα and AMPKα (P = 0.22), (C) P-β-catenin and β-catenin (P = 0.96), (D) P-ERK and ERK (P = 0.022), (E) JNK (P = 0.06), (F) P-S6 and S6 (P = 0.64), and (G) SUFU (P = 0.049) in control and GR-null fetal kidney at E18.5. Western blot analysis of (H) Aurora kinase A (AURKA) (P = 0.017) in IMCD3 cells treated with vehicle (veh) or dexamethasone (dex). Western blot analysis of cilia structural proteins (I) acetylated tubulin (P = 0.090), (J) CEP290 (P = 0.52), (K) IFT88 (P = 0.15), (L) KIF3A (P = 0.34) in control and GR-null fetal kidney at E18.5. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null, n = 4 animals per experimental group or veh and dex, n = 3 biological replicates per experimental group. Western blots from P-AKT and AKT (A) were re-probed with IFT88 (K) and acetylated tubulin (I), respectively, and share the same β-actin blot. Western blots from P-ERK and ERK (D) was re-probed with P-S6 and S6 (F), respectively, and share the same β-actin blot. KIF3A (L) was re-probed with CEP290 (J) and share the same β-actin blot. The same β-actin panels were included in each figure for ease of comparison. Source data are available online for this figure.

Figure EV1. Single cell analysis and localisation of target gene expression in the fetal mouse kidney at E13.5, E15.5 and E18.5.

Expression of cell type-specific marker genes within single cell clusters at E13.5 (A), E15.5 (B) and E18.5 (C). Colour scale represents average expression, dot size represents percent of cells within cluster that express the gene.

Figure EV2. Glucocorticoid receptor localisation and deletion in the GR-null fetal mouse kidney.

(A) Immunofluorescence of glucocorticoid receptor (GR) localisation and deletion in control and GR-null fetal kidney at E18.5. Sections were stained with Hoechst (blue, nucleus), Dolichos Biflorus Agglutinin (DBA) (green, collecting duct), GR (red, GR) and Lotus Tetragonolobus Lectin (LTL) (yellow, proximal tubule). White arrow indicates GR localisation. Slides were imaged with a Zeiss LSM 980 confocal microscopy (63x objective, 2x digital zoom), scale bar represents 20 μm. All images are representative of n = 4 animals per experimental group. (B) Western blot analysis of GR protein in GR-null fetal kidney at E18.5. All data presented as mean ± SEM, significant differences were analysed by one-way ANOVA with multiple comparisons indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control, heterozygous (het) and GR-null. Control vs het (P = 0.0001), control vs GR-null (P = 0.0001), het vs GR-null (P = 0.0012). Data from n = 3 animals per experimental group. Source data are available online for this figure.

Figure EV3. Gene set enrichment analysis and RNA-seq deconvolution.

(A) Gene set enrichment analysis highlights impacted pathways between control and GR-null bulk RNA-seq datasets. RNA-seq was performed on total RNA isolated from control (n = 4) and GR-null (n = 3) mouse kidneys at E18.5. P value for each term is plotted on the x axis. (B) Bar plot of single cell deconvolution analysis performed to estimate changes in cell signatures in the bulk RNA-seq profiles of control (C1-4) and GR-null (N1-N3) samples. Cell type signatures based on the E18.5 single cell reference are colour coded as per the legend. Analysis of cell proportions between replicates from control and GR-null groups identified small but significant reductions in signatures related to the early proximal tubule (P = 0.0143) and medullary stroma (P = 0.055) in the GR-null group (two-tailed t test with unequal variance). (C) Bar plot of single cell deconvolution analysis performed to estimate changes in cell signatures in the bulk RNA-seq profiles of control (AVG. Cont, n = 4) and GR-null (AVG.KO, n = 3) samples. Bars illustrate the percentage average in each group tagged with standard error, significant differences were analysed by two-tailed t test with unequal variance, indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant. Analysis of cell proportions between replicates from control and GR-null groups identified small but significant reductions in signatures related to the early proximal tubule (P = 0.0143) and medullary stroma (P = 0.055) in the GR-null group.

Figure EV4. Glucocorticoid regulation of primary cilia length in GR-null kidney at E18.5.

(A) Immunofluorescence of primary cilia morphology in control and GR-null fetal kidney at E18.5. Sections were stained with Hoechst (blue, nucleus), acetylated tubulin (AceTub) (green, microtubules) and ARL13B (red, cilia axoneme). White arrows indicate primary cilia morphology. Slides were imaged with a Zeiss LSM 980 confocal microscopy (×63 objective, 2× digital zoom), scale bar represents 20 µm. All images are representative of n = 3 animals per experimental group. (B) Immunofluorescence of primary cilia morphology in control and GR-null collecting ducts at E18.5. Sections were stained with Hoechst (blue, nucleus), acetylated tubulin (AceTub) (green, microtubules), pericentrin (PCNT) (red, basal body) and Dolichos Biflorus Agglutinin (DBA) (grey, collecting duct). White arrows indicate primary cilia morphology. Slides were imaged with a Zeiss LSM 980 confocal microscope (×63 objective, 2× digital zoom), scale bar represents 20 µm. All images are representative of n = 4 animals per experimental group. Source data are available online for this figure.

Figure EV5. Cell proliferation localisation in the GR-null fetal mouse kidney proximal tubule cells at E18.5 and primary cilia expression in IMCD3 cells after serum starved media conditions.

(A) Immunofluorescence of proliferation localisation in control and GR-null fetal kidney at E18.5. Sections were stained with Hoechst (blue, nucleus), acetylated tubulin (AceTub) (green, microtubules), KI67 (red, proliferative cells) and Lotus Tetragonolobus Lectin (LTL) (grey, proximal tubule). White arrow indicates proliferative cells. Slides were imaged with a Zeiss LSM 980 confocal microscopy (×63 objective, 2× digital zoom), scale bar represents 20 μm. All images are representative of n = 4 animals per experimental group. (B) Percentage of proliferating cells. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null (P = 0.67), n = 3 animals per experimental group. (C) Percentage of proliferating proximal tubule cells. All data presented as mean ± SEM, significant differences were analysed by unpaired T tests indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between control and GR-null (P = 0.70), n = 3 animals per experimental group. (D) Percentage of cells with a primary cilium in complete media versus serum starved vehicle (veh) and serum starved dexamethasone (dex) in IMCD3 cells. All data presented as mean ± SEM, significant differences were analysed by one-way ANOVA with multiple comparisons indicated by *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001, ns=not significant, between complete media, serum starved veh and serum starved dex. Complete media vs serum starved veh (P = 0.0001), complete media vs serum starved dex (P = 0.0001), serum starved veh vs serum starved dex (P = 0.54). Data from n = 3 biological replicates per experimental group. Source data are available online for this figure.