Deletion of Aurora kinase A prevents the development of polycystic kidney disease in mice

Abstract

Aurora Kinase A (AURKA) promotes cell proliferation and is overexpressed in different types of polycystic kidney disease (PKD). To understand AURKA's role in regulating renal cyst development we conditionally deleted the gene in mouse models of Autosomal Dominant PKD (ADPKD) and Joubert Syndrome, caused by Polycystin 1 (Pkd1) and Inositol polyphosphate-5-phosphatase E (Inpp5e) mutations respectively. We show that while Aurka is dispensable for collecting duct development and homeostasis, its deletion prevents cyst formation in both disease models. Cross-comparison of transcriptional changes implicated AKT signaling in cyst prevention and we show that (i) AURKA and AKT physically interact, (ii) AURKA regulates AKT activity in a kinase-independent manner and (iii) inhibition of AKT can reduce disease severity. AKT activation also regulates Aurka expression, creating a feed-forward loop driving renal cystogenesis. We find that the AURKA kinase inhibitor Alisertib stabilises the AURKA protein, agonizing its cystogenic functions. These studies identify AURKA as a master regulator of renal cyst development in different types of PKD, functioning in-part via AKT.

Fig. 1. Aurka knock-out prevents cystogenesis.

a, j Images of mouse kidneys from indicated genotypes (bar = (A) 1 cm (whole organs); j = 0.5 cm; 2 mm (for H&E sections); 50 μm (for immunostains)). b, k cystic index (Inpp5e p values for pairs indicated from left to right; p = 8.7 × 10^-7, p = 9.3 × 10^-7 and Pkd1 p values; p = 3.9 × 10^-4, p = 4.1 × 10^-6). c, l average cyst number per cross-section number of cysts (Inpp5e p values for pairs indicated from left to right; p = 4.6 × 10^-5, p = 7.5 × 10^-6, and Pkd1 p values; p = 2.6 × 10^-4, p = 0.0122). d, m average cyst cross-sectional size (Pkd1 p values for pairs indicated from left to right; p = 0.0419 and p = 0.0094). e, n Kidney sections of indicated genotypes of increasing age. e bars = 0.4 mm and n = 4 mm. f–h and o–q Plot of kidney-to-body-weight ratio, cyst number and cystic index of indicated mice over time. i, r Blood Urea Nitrogen (BUN) assessment of kidney function with age (Inpp5e p values indicated = 0.0003, Pkd1 p values for pairs indicated from left to right p = 0.0005, p = 0.0490). All graph data indicate mean ± S.E.M. a–e Mice were P17–P21 days of age. j–o Mice were P11 days of age. Graph abbreviations that apply throughout I Inpp5e, P Pkd1, A Aurka, ns not significant. a–d n = 4–9, e–h n = 3–24, i n = 4–12, j–m n = 3–4, n, p–q n = 3–16, o n = 3–34, r n = 4–9. Cysts were not present in the kidneys of other mouse genotypes and are omitted from cyst size, number and index graphs. Additional data in Figs. S1 and S2. Exact n and data point values provided in supplementary material- Source Data File.

Fig. 2. Aurka knock-out normalises ciliation, proliferation and DNA-damage responses.

a, e P4 kidney sections immunostained for DAPI, acetylated-tubulin (AcTub), Pericentrin (Pcent) and DBA. Arrowheads indicate collecting duct cells with cilia. Bar = 12 μm (Inpp5e), 10 μm (Pkd1). b, f P4 kidney sections immunostained for DAPI, Ki67 and DBA. Arrowheads demonstrate collecting duct cells positive for Ki67. Bar = 25 μm. c, g Quantification of collecting duct cilia length at P21 in Inpp5e model and P11 in Pkd1 model, respectively (ns not significant). d, h-left Quantification of the proportion of collecting duct cells exhibiting cilia across the time points indicated. Inpp5e p values for pairs indicated from left to right; p = 0.0021, p = 0.0047, p = 0.0020, p = 0.0016, p = 0.0056 and p = 0.0039. Pkd1 p values for pairs indicated from left to right; p = 0.0006, p = 0.0017, p = 0.0008, p = 0.0008, p = 0.0002, p = 0.00001. d, h-right Quantification of the proportion of collecting duct cells stained for Ki67 across time. Inpp5e p values for pairs indicated from left to right; p = 0.0002, p = 0.0007, p = 0.0031, p = 0.0013, p = 0.0002 and p = 0.0005. Pkd1 p values for pairs indicated from left to right; p = 0.0092, p = 0.0171, p = 0.0096, p = 0.0180, p = 0.0045, p = 0.0010. i, k P4 Kidney sections immunostained for DAPI, γ-H2AX and DBA; arrowhead indicate CD cells expressing γ-H2AX reactivity. Quantification of P4 γ-H2AX expression in collecting ducts (Inpp5e p values for pairs indicated from left to right; p = 0.0043, p = 0.0069, Pkd1 p values; p = 0.0042, p = 0.0016). j Western blot for relative γ-H2AX expression at P21, re-probed with Actin, Bar = 25 μm. All graph data indicates mean ± S.E.M. n = 3 all subpanels. Additional data in Figs. S1 and S2. Full vertical lane scans of WB image (j) in Source Data File. Exact data point values provided in supplementary material- Source Data File.

Fig. 3. Aurka deletion reduces the severity of adult-onset ADPKD.

a Representative images of whole mouse kidney sections stained for either Haematoxylin and Eosin (H&E, left), or DAPI, DBA, Uromodulin (UMOD) and LTL (middle and right). Kidneys were collected at 20 weeks of age from mice of the indicated genotypes. Scale bar = 3 mm (left and middle), 50 μm (right). The staining intensity of DBA was low in Pkd1 f/f; Aurka f/f non-cystic animals compared to cystic animals. b The average cyst size per cross-section for overall cyst analysis, significant p values indicated from left to right: 0.0054; 0.0282, 0.0027, 0.0084, 0.0077, 0.0001. c Analysis of cystic index measured overall; p values indicated from left to right: 0.0170, 0.0445, 0.0055, 0.0295. d Quantitation of the average cyst number per cross-section overall; p values indicated from left to right: 0.0158, 0.0272. e Kaplan–Meier curve indicating survival of Pkd1^Δ/Δ, Pkd1^Δ/Δ;Aurka^Δ/+ and Pkd1^Δ/Δ;Aurka^Δ/Δ over time. Comparing all three genotypes gives p = 0.0491 (Mantel-Cox test, Chi-square 6.029, df 2). Pairwise comparison p value between Pkd1^Δ/Δ and Pkd1^Δ/Δ;Aurka^Δ/Δ = 0.0166. Assessment of Blood Urea Nitrogen (BUN) as a measure of kidney function mice of indicated genotypes at experimental end point (either that the conclusion of 20-week time course or at the point of humane euthanasia if earlier, left to right p = 0.0108 and p = 0.0493). All graphs indicate mean ± S.E.M. (a–d) n = 4–8 at all time points for all genotypes, see exact n and data point values provided in supplementary material- Source Data File. Kaplan–Meier curve data displays n = 20 (Pkd1^Δ/Δ), n = 19 (Pkd1^Δ/Δ;Aurka^Δ/+) and n = 16 (Pkd1^Δ/Δ;Aurka^Δ/Δ). Cysts were not present in the kidneys mouse genotypes not indicated. Additional data in Figs. S3 & S4.

Fig. 4. AURKA alters the AKT signalling pathway.

Venn diagrams show the number of genes whose expression varies by ±1.5-fold in P4 Inpp5e (a) and Pkd1 (b) mouse models comparing the genotypes indicated. c Venn diagram comparing altered KEGG Pathways in P4 kidneys of indicated genotypes. d, e Staining of P4 kidney sections for DAPI, pAKT (T308) and DBA (arrowheads indicate pAKT T308 expression; bar = 25 μm) and their quantifications. Inpp5e p values from left to right; p = 0.0035, p = 0.0130 and Pkd1 p values; p = 0.0008, p = 0.0008, p = 0.0085. f, g Expression and co-localisation of AURKA-V5 and AURKA in control and cystic kidneys of both models at P15 and P11 respectively (bar = 25 μm). h Co-immunoprecipitation of endogenous AKT with AURKA-V5 from P11 Pkd1^Δ/Δ mouse kidneys and competition with V5-peptide (blocking IP) (* = mouse endogenous IgG). i Co-immunoprecipitaton of endogenous AKT as in (h) from Inpp5e^Δ/Δ kidneys at P15. Additional data in Fig. S5. j Co-immunoprecipitation of endogenous AKT from testis extracts. All graphs display mean ± S.E.M. a–c n = 6, d, e n = 3–4, f–j n = 3. Exact n and data point values provided in supplementary material- Source Data File.

Fig. 5. AURKA regulates pAKT (T308) status and co-localise.

a–e Western blots and densitometry of mIMCD3 cell lysates transfected with control siRNA or Aurka siRNA#1 (cultured for 24 h). P values from left to right; p = 0.0001, p = 0.0069, p = 0.0360. f–j Western blots and densitometry of mIMCD3 cell lysates treated with Alisertib for 48 h under growth conditions and probed as indicated. P values from left to right; p = 0.0033, p = 0.0039, p = 0.0223. k–o Western blots and densitometry of mIMCD3 cell lysates transfected with HA, HA-AURKA or HA-AURKA KD expression plasmids (cultured 24 h) and probed as indicated. Densitometry values are sum of multiple bands observed. P values from left to right for pairs indicated; p = 1.8 × 10^-5, p = 0.0007, p = 0.0313, p = 0.0014, p = 0.019. p Immunostaining for DAPI, AURKA, and pAKT (T308) in mIMCD3 cells showing co-localisation at the basal body and the distribution of protein localisation. q Immunostaining for DAPI, AURKA, and PDK1 in mIMCD3 cells showing co-localisation at the basal body and the distribution of protein localisation (bar = 10 μm). r Immunostaining for DAPI, AURKA, and pAKT (S473) in mIMCD3 cells showing no co-localisation at the basal body (bar = 10 μm). All graph data indicate mean ± S.E.M. Control references defined as 1. pT308 pAKT (T308). a–e n = 7, f–j n = 6, k–o n = 8, p–q n = 4, r n = 3. Additional data in Figs. S6 and S7. Western blots represent biological replicates with each independent experiment set run on a separate gel or two replicates on a gel. Each control sample within a replicate set was then defined as 1. The western membranes were cut to 40−60 kDa interval before antibody probing, repeated stripping and reprobing of the same membrane to generate datasets. Individual western blot scans, n and data point values provided in supplementary material- Source Data File. All graph data indicate mean ± S.E.M.

Fig. 6. Alisertib increases cyst number in vivo in JS model.

a Mice, whole kidneys and sections following treatment with vehicle or Alisertib (bar = 1 cm for kidneys; 2 mm for sections). b–e Quantification of combined kidney weight over total body weight, cyst number, cyst size and cystic index for vehicle and Alisertib-treated mice. P values are 0.0003 and 0.0059 (b), 0.0247 (c) and 0.0117 (e). f Ciliary co-localisation of AURKA and pAKT (T308) in Inpp5e^Δ/Δ kidneys (bar = 10 μm). g Quantification of ciliated collecting duct cells following treatment with vehicle and Alisertib. P values from left to right; p = 0.0012, p = 0.0162. h, i Active-p53 in collecting ducts of Inpp5e^Δ/Δ kidneys (arrow = nuclear p53; bar = 25 μm). P values from left to right; p = 0.0123, p = 0.0022. j, k Staining and quantification of AURKA and pAKT (T308) +ve collecting duct cells. P values from left to right; p = 0.0173, p = 0.0035. l Quantification of the proportion of dual AURKA/Ki67+ve collecting duct cells. P values from left to right; p = 0.0072, p = 0.0009. All graph data indicate mean ± S.E.M. pT308 pAKT (T308). ns not significant. White bars = vehicle; Grey bars = Alisertib treatment. a–b n = 3–5, c–l n = 3. All data from P15. Additional data in Fig. S8. Exact n and data point values provided in supplementary material- Source Data File.

Fig. 7. Inhibition of AKT constrains cyst formation.

a, i Whole kidneys from mice treated with vehicle or MK2206 (bar = 1 cm (Inpp5e); 0.5 cm (Pkd1)). b, j Quantification of the combined kidney weight over total body weight percentage (2 K/BW%). P values as indicated are p = 0.0075 (b), p = 1.8 × 10^-5 for 37.5 mg/kg; p = 0.0006 for 75 mg/kg (j). c, k Cystic index quantification, p = 0.0233 (c), p = 0.0096 for 37.5 mg/kg; p = 0.0226 for 75 mg/kg (k). d, l Quantification of average cyst number per section, p = 0.0323 (d), p = 0.0314 for 10 mg/kg; p = 0.0080 for 37.5 mg/kg; p = 0.0306 for 75 mg/kg (n). e, m Average cyst cross-sectional size, p = 0.0267 (e), p = 0.0102 for 37.5 mg/kg; p = 0.0046 for 75 mg/kg (m). f, n Ciliation, p = 0.0249 and 0.0118 for (f) and p = 0.0035, p = 0.0269, p = 0.0327 for (n). o Staining for AURKA, DBA and pAKT (T308) in mice exposed to MK2206 or vehicle controls (bar = 10 μm). g Quantification of pAKT (T308) +ve co-labelled collecting duct cells. P values from left to right; p = 0.0008, p = 0.0062. h, p Quantification of pAKT (T308)/AURKA +ve co-labelled collecting duct cells. P values from left to right; p = 0.0018, p = 0.0034 for (h) and p = 0.0263, 0.0024 and 0.0017 for (p). q, r Staining and quantification of Ki67/AURKA collecting duct cells. P values from left to right; p = 0.0061 and 0.0075, bar = 10 μm. All graph data indicate mean ± S.E.M. pT308 pAKT (T308). White bars = vehicle; Grey bars = MK2206 treatment. a–h n = 3–5, i–j n = 4–12, k–m 4–10, n–r n = 3. Inpp5e model used MK2206 at 75 mg/kg and data is at P15. Pkd1 model used MK2206 doses as indicated and data is at P11. Additional data in Fig. S8. s A model of the role of AURKA and AKT in regulating cystogenesis in PKD. Deletion of either Pkd1 (Major pathway driver) or Inpp5e (Minor pathway driver) result in upregulation of AURKA and enhanced AKT activity through phosphorylation of T308 to form a self-reinforcing signalling loop driving cystogenesis. Co-deletion of Aurka prevents AKT activation and stops cyst formation. Inhibiting AKT activity with MK2206 breaks the feedback loop and also reduces cyst formation. Exact n and data point values provided in Source Data File.