A protein kinase A and Wnt-dependent network regulating an intermediate stage in epithelial tubulogenesis during kidney development

Abstract

Genetic interactions regulating intermediate stages of tubulogenesis in the developing kidney have been difficult to define. A systems biology strategy using microarray was combined with in vitro/ex vivo and genetic approaches to identify pathways regulating specific stages of tubulogenesis. Analysis of the progression of the metanephric mesenchyme (MM) through four stages of tubule induction and differentiation (i.e., epithelialization, tubular organization and elongation and early differentiation) revealed signaling pathways potentially involved at each stage and suggested key roles for a number of signaling molecules. A screen of the signaling pathways on in vitro/ex vivo nephron formation implicated a unique regulatory role for protein kinase A (PKA), through PKA-2, in a specific post-epithelialization morphogenetic step (conversion of the renal vesicle to the S-shaped body). Microarray analysis not only confirmed this stage-specificity, but also highlighted the upregulation of Wnt genes. Addition of PKA agonists to LIF-induced nephrons (previously shown to be a Wnt/beta-catenin dependent pathway) disrupted normal tubulogenesis in a manner similar to PKA-agonist treated MM/spinal-cord assays, suggesting that PKA regulates a Wnt-dependent tubulogenesis step. PKA induction of canonical Wnt signaling during tubulogenesis was confirmed genetically using MM from Batgal-reporter mice. Addition of a Wnt synthesis inhibitor to activated PKA cultures rescued tubulogenesis. By re-analysis of existing microarray data from the FGF8, Lim1 and Wnt4 knockouts, which arrest in early tubulogenesis, a network of genes involving PKA, Wnt, Lhx1, FGF8, and hyaluronic acid signaling regulating the transition of nascent epithelial cells to tubular epithelium was derived, helping to reconcile in vivo and in vitro/ex vivo data.

Figure 1. Biochemical and morphological analyses identify distinct stages of in vitro renal tubule development

Metanephric mesenchyme (MM) was induced to form epithelial renal tubules by co-culture with spinal cord for up to a week. The development of the tubules was examined using imaging techniques highlighting several distinct stages of development. (A) Freshly isolated MM. Scale bar is 100μm. (B) The tissue undergoes mesenchymal to epithelial transition and begins epithelial tubule morphogenesis by 2 days of culture. Pax-2 (red) expression is seen throughout the tissue, especially in a band (arrow) adjacent to the spinal cord. E-cadherin (green) expression in nascent epithelium is evident in cell clusters. Scale bar is 100μm. (C) By 3 days, S-shaped bodies have formed. Scale bar is 100μm. (D) At 5 days elongated tubules (E-cadherin, green) are interspersed with developing podocytes (PNA, red). Scale bar is 50μm.

Figure 2. Signaling pathway tubulogenesis screen implicates PKA regulation in tubule formation from renal vesicles

(A) Ex vivo culture of MM results in robust tubule growth, recapitulating much of early nephron development. A tubule is outlined and indicated by an arrowhead. Several pathways identified in the ex vivo culture microarray analysis had no effect on tubule growth, such as inhibition of PI3 Kinase (arrowheads mark representative tubules, one is outlined) (B). Activation of PKA with dibutyryl-cAMP (C) yielded epithelial clusters (circled), with a lesion of tubulogenesis following renal vesicle formation. All scale bars are 100μm.

Figure 3. Renal tubulogenesis is prevented by activation of PKA-2 but not PKA-1

(A) Spinal cord induced MM develops long convoluted epithelial tubules, with apical expression of ZO-1 (red) and lateral expression of E-Cadherin (green). (B) Activation of PKA-1 by isozyme specific cAMP analogues does not impair tubule formation. (C) Activation of PKA-2 by isozyme specific cAMP analogues prevents tubule formation from renal vesicles, resulting in disorganized epithelial clusters and large dilations, recapitulates the developmental lesion by activation of PKA with dbcAMP (shown in Figure 2C). A–C Scale bars are 100μm. (D) Volumetric analysis of control cultures indicates long continuous staining regions of ZO-1 shown as objects, each continuous section with a distinct (arbitrary) color. The control tissue is primarily comprised of two large objects, representing large continuous tubules. (E) Analysis of PKA-1 activated samples indicates many objects with continuous ZO-1 expression, and many of the objects can be seen to connect in the fluorescent mode of the z-stack. (F) Activation of PKA-2 results in a single mass of epithelium, in which only one distinct object can be discerned. D–F, Scale bars are 140μm. The images shown in A–C are the same images for which analysis is shown in D–F.

Figure 4. Inhibition of AKAP function delays tubulogenesis

Inhibition of AKAPs results in a delay of tubule formation. (A) At 72 hours control cultures have developed into S-shaped bodies. (B) At 72 hours, AKAP inhibitor samples have developed renal vesicles (arrowhead). Scale bars are 100μm.

Figure 5. Modulation of the PKA pathway identified PKA-2 as a regulator of tubule morphogenesis

Spinal cord-induced MM was treated with PKA pathway signaling modulators and observed for defects in tubule morphogenesis. (A) Activation of PKA by stimulation of cAMP production (forskolin) or by adding cAMP analogues (dibutyryl-cAMP) prevented tubular morphogenesis. Activation of PKA-1 (8-PIP-cAMP and 8-HA-cAMPS) and Epac (8-CPT-2-O-Me-cAMP) did not inhibit tubulogenesis while activation of PKA-2 (6-MBC-cAMP and Sp-5,6-DCl-cBIMPS) prevented tubule formation. Inhibition of PKA with the Walsh PKA inhibitory peptide, 14–22 stimulated tubulogenesis. (B) Cartoon demonstrating the PKA pathway and the effects of modulation of specific parts on tubulogenesis demonstrates activation of the PKA pathway at different levels prevents tubulogenesis.

Figure 6. PKA agonists produce a stage specific lesion resulting in disorganized or laminar epithelial tissue in spinal cord –induced MM cultures

(A) Control spinal cord-induced MM tubule cultures exhibit epithelial tubules at 7 days of culture, while PKA-2 agonist (6-MBC-cAMP and Sp-5,6-DCl-cBIMPS) treated samples (B) comprise epithelium which is found in disorganized clusters with mislocalized E-cadherin. A and B: E-cadherin (green), ZO-1 (red). (C) By electron microscopy, at 7 days of culture, two rows of epithelial cells can be seen with debris in the luminal space. (D) Electron micrographs of PKA-2 agonist treated cultures show a large dilation. The open area is surrounded by a single layer of epithelial cells. Scale bars: A and B, 25μm; C, 5μm; D, 100μm.

Figure 7. Transcriptomic analysis reveals the rate-limiting step of the lesion to be the renal vesicle

Control and PKA-2 agonist (6-MBC-cAMP and Sp-5,6-DCl-cBIMPS) treated spinal cord induced MM were cultured for 3 days (the control reaches the S-shaped body, refer to figure 1D) and analyzed by cDNA microarray. (A) Highly expressed genes from control and PKA-2 agonist treated cultures were compared to gene signatures of in vivo nephrogenic intermediates. Control samples begin to express genes associated with advanced structures (S-shaped body, renal corpuscle and proximal tubule), not observed in the PKA agonist treated samples. (B) The in vivo genetic markers of the PT which was compared to the genes highly expressed in the control and PKA-2 agonist samples are highly expressed primarily in the PT. Abbreviations: RC, glomerulus/renal corpuscle; mm, metanephric mesenchyme; pt, proximal tubule; rv, renal vesicle; ss, S-shaped body; ub, ureteric bud. (C) Pathways analysis of genes passing t-test (p=0.05) and 5-fold change filtering. The Wnt/beta-catenin canonical pathway (large nodes) is highly represented in the activated PKA-2 condition (red). The pathway includes a nuclear inhibitor of PKA induced transcription (Crem). See Supplementary Tables 2 and 3 for full gene lists.

Figure 8. Canonical Wnt signaling is persistent under activated PKA-2

Spinal cord-induced MM cultures were examined using tissue from Batgal mice, a genetic model of canonical Wnt signaling. Expression of beta-galactosidase (blue x-gal staining) is indicative of beta-catenin induced transcription. (A) Wnt signaling controlled Beta-galactosidase expression in BAT-gal mouse tissue is low in control cultures, and is restricted to several small epithelial clusters (arrow) but is absent from advanced epithelial structures (circled). (B) Beta-galactosidase expression is higher in PKA-2 agonist (6-MBC-cAMP and Sp-5,6-DCl-cBIMPS) treated cultures, and is found throughout all epithelial structures (large mass circled), indicating a direct relationship between PKA and Wnt controlled gene expression. All experiments were stained and imaged at 3 days of culture time. Scale bars are 100μm.

Figure 9. Inhibition of Wnts rescues tubulogenesis in PKA-2 activated cultures

Spinal cord-induced MM tubulogenesis is abolished when PKA-2 is activated (6-MBC-cAMP and Sp-5,6-DCl-cBIMPS). Activation of PKA-2 plus inhibition of Wnt (IWP-2) signaling rescues tubulogenesis at the edges of the epithelial clusters. (A) ZO-1 staining (red) highlights disorganized epithelium in the PKA-2 activated condition. (B) PKA-2 activated disruption of tubulogenesis is partially rescued by the addition of a Wnt inhibitor (IWP-2). ZO-1(red) is localized along the lumen of the tubule, apical from the nuclei (blue). B and C: Scale bars are 20μm. (D) Cartoon demonstrating the effect of PKA activity on Wnt expression and signaling. Diminishing PKA and Wnt activity results in normal tubulogenesis while sustained PKA activity leads to sustained Wnt signaling and disruption of tubule formation.

Figure 10. Network of genes misregulated in lesions which affect early nephron morphogenesis

(A) A diagrammatic description of the developmental defect in nephrogenesis caused by genetic deletions or by treatment of cultured nephrons with signaling pathway modulators. (B–E) A group of core nephrogenesis genes, validated by extensive literature is augmented by molecules found to be differentially expressed in nephron precursors that fail to undergo tubular morphogenesis (Wnt4 null, B; FGF8 null, C; Lim1 (Lhx1) null, D) and in PKA-2 agonist treated spinal cord-induced MM (E). The new nodes link Wnt, PKA and hyaluronic acid, all of which affect nephron formation in vitro/ex vivo, to the existing network. Arrowheads indicate increased expression in vivo, opposite of the direction observed in the expression profile of the mutant, bars indicate reduced expression.