Sall1-dependent signals affect Wnt signaling and ureter tip fate to initiate kidney development

Abstract

Development of the metanephric kidney depends on precise control of branching of the ureteric bud. Branching events represent terminal bifurcations that are thought to depend on unique patterns of gene expression in the tip compared with the stalk and are influenced by mesenchymal signals. The metanephric mesenchyme-derived signals that control gene expression at the ureteric bud tip are not well understood. In mouse Sall1 mutants, the ureteric bud grows out and invades the metanephric mesenchyme, but it fails to initiate branching despite tip-specific expression of Ret and Wnt11. The stalk-specific marker Wnt9b and the beta-catenin downstream target Axin2 are ectopically expressed in the mutant ureteric bud tips, suggesting that upregulated canonical Wnt signaling disrupts ureter branching in this mutant. In support of this hypothesis, ureter arrest is rescued by lowering beta-catenin levels in the Sall1 mutant and is phenocopied by ectopic expression of a stabilized beta-catenin in the ureteric bud. Furthermore, transgenic overexpression of Wnt9b in the ureteric bud causes reduced branching in multiple founder lines. These studies indicate that Sall1-dependent signals from the metanephric mesenchyme are required to modulate ureteric bud tip Wnt patterning in order to initiate branching.

Fig. 1.

Sall1-ΔZn mutants exhibit blind ureters that properly differentiate into urothelium and smooth muscle layers. (A,B) E18.5 urogenital tracts from wild-type and homozygous mouse embryos demonstrate that Sall1-ΔZn mutants do not form kidneys (k), but that morphologically distinct ureters are visible. Testes (t) and bladder (bl) are formed. (C,D) Transverse sections at the level indicated by the arrows are shown in the lower panels. Haemotoxylin and Eosin staining reveals a proper urothelium (u) and smooth muscle (sm) layer in cross sections of both wild-type and mutant ureters. The mutant exhibits decreased luminal space that might be secondary to the lack of urine flow. (E-H) Immunohistochemistry with anti-smooth muscle actin and anti-uroplakin III appears similar in wild-type and mutant sections, and confirms the proper differentiation of mutant ureters.

Fig. 2.

Blind ureters are caused by failure to initiate the first branch. (A-D) Hoxb7GFP-labeled genital tracts from wild-type and mutant mice at E16.5 (A,B) and E11.5 (C,D) reveal that the wild-type ureteric bud (ub) properly forms a T-branch at E11.5 and a branched kidney at E16.5. Sall1-ΔZn homozygous mutant ubs properly grow out from the Wolffian duct (wd), but fail to branch. Blind ureters persist at E16.5 in the mutant.

Fig. 3.

Mutant cap mesenchyme persists and expresses low Gdnf, but this does not account for the blind ureter phenotype. (A-J) Hoxb7GFP expression (A,B) or whole-mount in situ hybridization (C-J) in wild-type or Sall1-ΔZn homozygous mutant E12.5 mouse tissues. Ret is properly tip restricted in the blind ureter. Pax2 and Six2 expression reveal a cap of metanephric mesenchyme (mm) atop the blind ureter that does not express Cited1 (ureter tip is marked by an asterisk in J). (K,L) Anti-Pax2 (green) and anti-laminin (red) staining show that Pax2-positive mm cells encircle the Pax2-positive ureteric bud (ub) outlined with anti-laminin in both heterozygous and homozygous tissues at E11.5. (M,N) Gdnf expression in the mutant mm is reduced compared with wild-type mm at E11.5. (O-R) Organ culture of Hoxb7GFP;Sall1-ΔZn^+/Δ or Hoxb7GFP;Sall1-ΔZn^Δ/Δ tissues at E11.5 with the addition of either BSA- or GDNF-coated beads imaged at 0 and 48 hours of culture. BSA-treated explants recapitulate the in vivo phenotypes with normal branching (O, +/Δ) and a blind ub (P, Δ/Δ). GDNF-treatment causes swollen branch events (Q, +/Δ) or a single branch that does not progress (R, Δ/Δ). (S,T) Recombination of wild-type E12.5 mm with E11.0 ampulla-stage ubs (S, +/Δ) or E12.5 blind ureters (T, Δ/Δ) show that the mutant (Δ/Δ) is capable of multiple branching events, similar to the control (+/Δ) ampulla, when wild-type mm is present.

Fig. 4.

Blind ureters exhibit altered tip fate by ectopic Wnt9b expression and are rescued by reducing β-catenin dose. (A,B) Hoxb7GFP expression at E11.0 (+/+) in mouse ampullas and E12.5 Sall1-ΔZn^Δ/Δ blind ureters. (C,D) Wnt11 is expressed similarly in the wild-type ampulla and mutant blind ureter and is absent from ureteric bud (ub) stalks (outlined). (E,F) Wnt9b expression detected by whole-mount in situ hybridization is upregulated in mutant ub tips (Δ/Δ) at E11.5 (not shown) and E12.5. Wild-type ub tip is outlined. (G,H) In the wild type, Axin2, a canonical Wnt signaling component and β-catenin responsive gene, is more strongly expressed in the Wolffian duct and ub stalk, and is downregulated in the ub tip (outlined) similar to Wnt9b. Axin2 is upregulated in the Sall1-ΔZn^Δ/Δ mutant ub tip. (I-O) Reduction of β-catenin dose rescues Sall1ΔZn^Δ/Δ blind ureter formation. Representative Hoxb7GFP expression (I-K) or whole-mount bright-field photos (L-N) are shown and the results are quantitated (O). Two or more rounds of branching were seen in 50% (seven out of 14) of E12.5 Sall1-ΔZn^Δ/Δ;Ctnnb1^+/− compared with 5% (one out of 22) of Sall1-ΔZn^Δ/Δ;Ctnnb1^+/+ kidneys. Blind ureters were observed in 50% of Sall1-ΔZn^Δ/Δ;Ctnnb1^+/− and 95% of Sall1-ΔZn^Δ/Δ;Ctnnb1^+/+ kidneys. Sall1^+/+;Ctnnb1^+/− kidneys (100%, 18 out of 18) had branched at least eight times by E12.5. Statistical significance was calculated using Fisher's exact test (P<0.0005).

Fig. 5.

Ectopic Wnt9b or increased β-catenin signaling in the ureteric bud (ub) is sufficient to reduce branching. (A-J) Ectopic ub expression of Wnt9b in E12.5 Hoxb7creEGFP^tg/+; ACTWnt9b^tg/+ transgenic mice causes reduced branching and abnormal morphology of the ureter. Four different founders are shown [founders 8 (B,J), 7 (D), 3 (F) and 10 (H)]. Analysis of Pax2 (G,H) and Six2 expression (I,J) shows normal mesenchymal cap formation in both wild-type and transgenic kidneys. (K-T) GFP expression and whole-mount in situ hybridization reveal unbranched ubs in Hoxb7creEGFP;Ctnnb1^ex3fl/+ at E12.5. This constitutively active β-catenin allele would be expected to increase Wnt signaling in the ub tip more significantly than ectopic Wnt9b alone and produces a more dramatic ub arrest. None of the genes assayed (Wnt9b, Wnt11, Pax2 and Sall1) was significantly altered in Hoxb7creEGFP;Ctnnb1^ex3fl/+ embryos, suggesting that Sall1 acts upstream of β-catenin signaling.

Fig. 6.

Model for the effect of mesenchymal Sall1 on ureter stalk expression of Wnt9b and canonical Wnt signaling. When the ureteric bud invades the metanephric mesenchyme (pink), mesenchymal progenitor cells (light gray and blue) coalesce at the ureter tip (brown). Wnt9b and β-catenin signaling (yellow) are downregulated in the ub tip, but continue to be expressed in the ub stalk. Signaling from the mesenchyme causes the ureter tips to branch. Reciprocal signals from the ureter to the mesenchyme induce the mesenchyme to form condensates that differentiate into renal vesicles near the Wnt9b-positive stalk. When Sall1 is not expressed in the mesenchyme (dark gray and blue, below the dotted line), Wnt9b is ectopically expressed in ureter tips and ureter branching arrests before the T-stage.