Fstl1 antagonizes BMP signaling and regulates ureter development

Abstract

Bone morphogenetic protein (BMP) signaling pathway plays important roles in urinary tract development although the detailed regulation of its activity in this process remains unclear. Here we report that follistatin-like 1 (Fstl1), encoding a secreted extracellular glycoprotein, is expressed in developing ureter and antagonizes BMP signaling activity. Mouse embryos carrying disrupted Fstl1 gene displayed prominent hydroureter arising from proximal segment and ureterovesical junction defects. These defects were associated with significant reduction in ureteric epithelial cell proliferation at E15.5 and E16.5 as well as absence of subepithelial ureteral mesenchymal cells in the urinary tract at E16.5 and E18.5. At the molecular level, increased BMP signaling was found in Fstl1 deficient ureters, indicated by elevated pSmad1/5/8 activity. In vitro study also indicated that Fstl1 can directly bind to ALK6 which is specifically expressed in ureteric epithelial cells in developing ureter. Furthermore, Sonic hedgehog (SHH) signaling, which is crucial for differentiation of ureteral subepithelial cell proliferation, was also impaired in Fstl1(-/-) ureter. Altogether, our data suggest that Fstl1 is essential in maintaining normal ureter development by antagonizing BMP signaling.

Figure 1. Fstl1 protein expression in developing murine ureter.

Fstl1 immunohistochemistry analysis in sagittal sections of WT (A) and Fstl1 ^-/- ureteric bud (B, negative control for antibody) at E11.5 and transverse sections of ureters at E13.5 (C), E15.5 (D), E16.5 (E) and E17.5 (F). Note that Fstl1 immunostaining was observed in the mesenchyme (um) as well as the ureteric epithelium (ue) from E15.5 to E17.5 (C-F). Scale bars: 40 µm. ub: ureter bud; m: metanephric mesenchyme; um: ureteral mesenchyme; ue: ureteric epithelium.

Figure 2. Histological analysis of the Fstl1^-/- urinary system.

(A-C) Kidneys and ureters from WT and Fstl1^-/- embryos at stages of E15.0 (A), E16.0 (B) and E17.5 (C). Arrows indicate mutant ureters, and arrowheads indicate wild-type ureters. (D) Dye injection experiments detected no complete physical obstruction of the Fstl1 ^-/- ureter at E18.5. (E-J) H&E staining on transverse sections of WT (E, G, I) and Fstl1 ^-/- (F, H, J) ureters at E15.0 (E, F), E16.0 (G, H) and E16.5 (I, J). (E, F) No obvious change was detected at E15.0 in Fstl1 ^-/- ureter. (G, H) Fstl1 ^-/- ureter became dilated from E16.0. (I, J) There are prominent changes in the histological structure of the Fstl1 ^-/- ureter compared to the WT ureter at E16.5. Scale bar: (E-J) 40 µm. k, kidney; p, pelvis; u, ureter; bl, bladder; um: ureteral mesenchyme; ue: ureteric epithelium.

Figure 3. Defects of UV orifice and distal ureter in Fstl1^-/- embryo.

(A-F) Fstl1 ^+/- mice were crossed to Ptch-lacZ ^+/- mice. Fstl1 ^+/- _; Ptch-lacZ ^+/- and Fstl1 ^-/- _; Ptch-lacZ ^+/- ureters were stained for β-galactosidase (blue) at E15.5 (A) and E16.5 (D). The Fstl1^-/- and wild-type ureters were paraffin sectioned and stained with hematoxylin-eosin at E15.5 (distal segment) (B, C), E16.5 (distal segment) (E, F). Note that the Fstl1 ^-/- ureter (A, D, C, F) is narrower than wild-type ureter (A, D, B, E) at both E15.5 and E16.5. The arrows indicated that lacZ staining was reduced in distal part of ureters in Fstl1 ^-/- embryos compared to wild-type embryos. (G, H) Histological analysis of the ureterovesicle orifice at E15.5. Note that the distance between the left and right orifices (asterisk) is shorter in the Fstl1 ^-/- embryo (H), compared with the wild-type embryo (G) at E15.5. (I, J) At E11.0, Pax2 and Pax3 whole mount in situ hybridizations were performed for wild-type (n =  4) (I), Fstl1^-/- (n =  4) (J). The somites, which were labeled by the Pax3 in situ probe, in the caudal region of the embryo were numbered. The initial site of ureteric bud, which was labeled by Pax2 and indicated by a triangle, in both the wild-type (I) and Fstl1^-/- (J) embryos aligned with the 28^th somite level. Scale bar: (B-F) 40µm, (G, H) 100µm.

Figure 4. Proliferation defects in Fstl1^-/- ureteric epithelial cells.

(A-D) Analysis of cell proliferation on transverse sections of wild-type (A, C) and Fstl1^-/- (B, D) ureters by BrdU incorporation at E15.5 (A, B) and E16.5 (distal segment) (C, D). Arrows point to representative proliferating cells. (E) Quantification of cell proliferation by BrdU labeling index. In E15.5 ureter, p<0.001 for epithelial layer (n = 20) and p = 0.49 for mesenchymal inner layer (n = 20). In distal segment of E16.5 ureter, p<0.001 for epithelial layer (n = 20) and p = 0.59 for mesenchymal inner layer (n = 20). (F, G) BrdU incorporation analysis of epithelial cell proliferation on transverse section of cultured E15.0 ureters treated with conditioned media transfected either with pcDNA3.1 vector (Mock) (F) or an Fstl1 overexpress plasmid (G). Arrows point to representative proliferating cells. (H) HEK293 cells were transfected with a plasmid expressing mouse Fstl1 protein or pcDNA3.1 (mock). Western blot analysis using an anti-Fstl1 antibody indicated that Fstl1 was overexpressed in both the cultured media and cell pellet. (I) Quantification of cell proliferation by BrdU labeling index in the cultured E15.0 ureteric epithelium. p<0.005 (n  = 7). Scale bar: (A-D, F-G) 20µm.

Figure 5. Subepithelial ureteral mesenchymal cells were absent in the Fstl1 ^-/- ureter.

(A-L) Immunofluorescence staining of α-SMA (green), pan-Cytokeratin (red), and DAPI (blue) of transverse sections from WT ureters at E16.0 (A, G), E16.5 (C, I) and E18.5 (E, K), and Fstl1 ^-/- ureters at E16.0 (B, H), E16.5 (D, J) and E18.5 (F, L). (G-L) Enlarged views of the boxed area in (A-F). The arrows indicate subepithelial ureteral mesenchymal cells which are negative for both markers. (M) Quantitative real-time PCR of Shh (n = 5, p = 0.0008) from E16.5 ureters. (N-Q) Fstl1 ^+/- mice were crossed to Ptch-lacZ ^+/- mice. Fstl1 ^+/- _; Ptch-lacZ ^+/- and Fstl1 ^-/- _; Ptch-lacZ ^+/- ureters were stained for β- galactosidase (blue) and pan-Cytokeratin (brown) at E16.5 (N, O); β-gal (blue) and nuclear fast red (red) at E18.5 (P, Q). Note that the lacZ staining was reduced in Fstl1 ^-/- ureters (O, Q) compared to Fstl1 ^+/- ureters (N, P). Scale, bar: (A-F) 50µm, and (G-L) 10µm, (N-Q) 40µm. um: ureteral mesenchyme; ue: ureteric epithelium.

Figure 6. Up-regulation of BMP signaling in the Fstl1^-/- ureter and kidney.

(A) Western blots of pSmad1/5/8, Smad1, pAKT (Ser⁴⁷³), AKT, and β-actin from E15.5 (left panel) and E16.5 (second panel) ureter protein, E18.5 kidney protein (third panels), and HEK293 cells treated with the conditioned media containing BMP4 (20 ng/ml) transfected either with the Fstl1 or pcDNA3.1 vector (Mock) for 30 min (right panels). (B) Quantitative real-time PCR of Bmp2 (n = 5, p = 0.35), Bmp4 (n = 4, p = 0.73), Bmp5 (n = 5, p = 0.63), Bmp7 (n = 5, p = 0.50), TGF-β1 (n = 5, p = 0.51), Alk3 (n = 6, p = 0.40), Alk6 (n = 5, p = 0.21), Alk5 (n = 4, p = 0.47), BmprII (n = 6, p = 0.73), ActrIIa (n = 5, p = 0.97), and ActrIIb (n = 5, p = 0.73) from E16.5 ureters. (C) Co-immunoprecipitation of Myc-Fstl1 and HA-tagged BMP/TGF-β receptors in HEK293 cells. Myc-Fstl1 can be immunoprecipitated with the anti-c-Myc antibody. Note that HA-ALK6 was co-immunoprecipitated by the anti-c-Myc and detected by the anti-HA antibody (lane 4). Scale bar: 40µm.

Figure 7. A model for Fstl1 function in urinary tract development.

Fstl1 is mainly expressed in ureteral mesenchymal cells, secreted to the ureteric epithelial layer, and functions as an antagonist regulating BMP signaling. When Fstl1 is disrupted, ureteric epithelium and mesenchymal-epithelial interaction defects down-regulate SHH signaling, interfering with subepithelial mesenchymal cells differentiation and together leading to the hydroureter phenotype.