Hedgehog-dependent and hedgehog-independent roles for growth arrest specific 1 in mammalian kidney morphogenesis Free

A long-standing tenet of developmental biology is that proper organogenesis requires cells to integrate a multitude of signals in a spatial- and time-dependent fashion. While previous research efforts have identified individual pathway requirements for organ formation, it remains unclear how multiple signaling inputs are synthesized to execute a specific cellular response, and even less clear how subsequent communication between cell types drives appropriate tissue development. One emerging theme that addresses the complex issue of organ formation is the use of individual proteins that impact multiple signaling pathways within a cell or across a tissue. Investigating these proteins may reveal the mechanisms used to ensure proper organ development.

Growth arrest specific 1 (GAS1) is a multi-functional, cell surface-associated protein that plays essential roles throughout mammalian embryogenesis (Allen et al., 2011, 2007; Biau et al., 2013; Lee et al., 2001; Martinelli and Fan, 2007a,b; Seppala et al., 2007). Gas1 was first identified as a gene upregulated in serum-starved NIH/3T3 fibroblasts (Schneider et al., 1988), encoding a 45 kDa glycosylphosphatidylinositol (GPI)-anchored protein that shares structural homology with glial-derived neurotrophic factor (GDNF) receptors (GFRs) (Cabrera et al., 2006). During embryogenesis, GAS1 is broadly expressed, and is best studied as a co-receptor for the hedgehog (HH) pathway, where it is required for the patterning and growth of multiple tissues (Allen et al., 2011, 2007; Biau et al., 2013; Izzi et al., 2011; Lee et al., 2001; Lee and Fan, 2001; Liu et al., 2001; Martinelli and Fan, 2007a).

Initial studies described an antagonistic role for GAS1 in HH signaling in early tooth (Lee et al., 2001) and somite development (Cobourne et al., 2004; Ohazama et al., 2009). However, more recent work demonstrated a role for Gas1 in HH pathway promotion, where Gas1 deletion results in HH-dependent defects in craniofacial development, digit specification, and ventral neural tube patterning (Allen et al., 2011, 2007; Martinelli and Fan, 2007a). Notably, GAS1 coordinates with two additional cell-surface proteins, CAM-related/downregulated by oncogenes (CDON) and brother of CDON (BOC), which together are required for HH signal transduction (Allen et al., 2011, 2007; Echevarria-Andino and Allen, 2020; Echevarria-Andino et al., 2022; Izzi et al., 2011; Martinelli and Fan, 2007a).

In addition to HH signaling, GAS1 regulates other signaling pathways. In particular, the structural similarity of GAS1 to GFRs led to work demonstrating physical interactions between GAS1 and the receptor tyrosine kinase RET (Cabrera et al., 2006). Furthermore, GAS1 negatively regulates RET signaling in cultured cells (Biau et al., 2013; Cabrera et al., 2006; Li et al., 2019; López-Ramírez et al., 2008). More recent work identified roles for GAS1 in modulating RET pathway activation during gastrointestinal development (Biau et al., 2013) and in skeletal muscle stem cell self-renewal (Li et al., 2019). GAS1 also regulates NOTCH-dependent SHH signaling in forebrain morphogenesis (Marczenke et al., 2021).

In the developing kidney, GAS1 is proposed to function independently of HH signaling. Specifically, Gas1 deletion in mice on a mixed genetic background results in renal hypoplasia at later developmental stages – i.e. after embryonic day (E) 15.5 (Kann et al., 2015). Gas1 modulation of FGF signaling during kidney morphogenesis was suggested to explain the renal hypoplasia phenotypes observed in Gas1 mutants (Kann et al., 2015). However, RET signaling has essential and well-studied roles in kidney morphogenesis, while HH signaling also contributes to multiple aspects of both mesonephric and metanephric development (Bohnenpoll et al., 2017; Cain et al., 2009; Chung et al., 2024 preprint; Murashima et al., 2014; Rowan et al., 2018; Yu et al., 2002). To date, potential roles for GAS1 in regulating these signaling pathways during kidney development remain unexplored.

Here, we have investigated GAS1 contributions to kidney morphogenesis in mouse embryonic development. Surprisingly, we found that Gas1 deletion on a congenic C57BL/6J background results in renal agenesis, a phenotype previously undescribed in Gas1 mutant animals. Furthermore, these phenotypes are genetic background dependent, as Gas1 mutant animals on a congenic 129S4/SvJaeJ background do not display renal agenesis but instead display renal hypoplasia as early as E12.5, which is, again, a previously unreported phenotype. Furthermore, Gas1 mutants across multiple genetic backgrounds display ectopic caudal mesonephric tubules and decreased HH target gene expression, consistent with reduced HH pathway activity. Surprisingly, this phenotype was specific to Gas1, as individual or combined deletion of Cdon or Boc did not result in any overt kidney phenotypes.

In Gas1 mutant metanephroi, we observed defects in mesenchyme proliferation and survival in early kidney morphogenesis. Furthermore, Gas1 deletion contributes to decreased RET target gene expression in the metanephros. Mechanistically, our data suggest that GAS1 is secreted from the kidney mesenchyme and physically interacts with RET in the adjacent renal epithelium. Ex vivo kidney explant cultures demonstrate that the branching morphogenesis defects in Gas1 mutant kidneys can be rescued by GDNF-mediated RET pathway activation. Together, these data indicate that GAS1 is a multi-functional regulator of kidney development, employing both HH-dependent and HH-independent mechanisms to drive proper mesonephros formation and kidney branching morphogenesis.

While previous work identified a role for Gas1 in later stages of kidney morphogenesis (Kann et al., 2015), our investigations led to a fundamentally different observation. Specifically, analysis of E15.5 wild-type (Fig. 1A), Gas1^+/− (Fig. 1B), and Gas1^−/− (Fig. 1C-E) embryos revealed a range of developmental defects in kidney morphogenesis in Gas1 mutants on a C57BL/6J background, from renal hypoplasia (Fig. 1C) to unilateral renal agenesis (Fig. 1D) to bilateral renal agenesis (Fig. 1E). Notably, the bilateral renal agenesis observed in a subset of Gas1 mutants phenocopies Ret mutant embryos (Schuchardt et al., 1994) (Fig. 1F). Quantitation of these phenotypes indicated normal kidney development in 100% of wild-type and Gas1^+/− embryos, while Gas1^−/− embryos most frequently display unilateral renal agenesis (64%); renal hypoplasia (14%) and bilateral renal agenesis (22%) occur less frequently (Fig. 1G).

To investigate the onset of these phenotypes, we collected E12.5 Gas1 mutant embryos, a stage that reflects early branching events of the ureteric bud (UB) epithelium, a structure that gives rise to the renal collecting duct system (Fig. 1H-T). Again, wild-type (Fig. 1H) and Gas1^+/− (Fig. 1I) embryos display normal metanephros development; however, Gas1^−/− embryos are perturbed even at this early stage of renal development (Fig. 1J,K), with hypoplasia (26%), unilateral renal agenesis (53%) and bilateral renal agenesis (21%) occurring at similar frequencies to those observed at E15.5 (compare Fig. 1G with Fig. 1T). Immunofluorescent antibody detection of E-cadherin (CDH1) confirmed initial UB branching events in wild-type (Fig. 1L) and Gas1^+/− (Fig. 1M) embryos, while UB branching was observed in hypoplastic Gas1^−/− kidneys (Fig. 1N), but not in Gas1 mutants lacking kidneys (Fig. 1O). Notably, UB branching is delayed in hypoplastic Gas1^−/− kidneys compared with wild-type or Gas1^+/− kidneys (see Fig. 1L-N). Given the importance of epithelial-mesenchymal signaling in metanephros branching morphogenesis (Combes et al., 2015; Costantini and Kopan, 2010; Little and McMahon, 2012), we used antibody detection of Wilms tumor 1 (WT1) to assess the metanephric mesenchyme (MM), which includes both stromal and nephron progenitor cells (SPCs and NPCs, respectively) (Armstrong et al., 1993; Pritchard-Jones et al., 1990; Rackley et al., 1993) in E12.5 wild-type and Gas1 mutant embryos (Fig. 1P-S). Strikingly, WT1 staining indicates a qualitative difference between hypoplastic Gas1 mutant kidneys, where WT1⁺ cells are observed (Fig. 1R), and Gas1^−/− embryos displaying renal agenesis, which lack WT1⁺ cells in the immediate vicinity of the UB (Fig. 1S). To assess Gas1 expression in metanephrogenesis, we used a Gas1^lacZ reporter allele (Martinelli and Fan, 2007a). At initial MM specification (E10.5; Fig. S1A-J) and during the first UB branching event (E11.5; Fig. 1U-D′), we observed Gas1 expression in WT1⁺ MM, whereas Gas1 was not detected in the adjacent CDH1⁺ UB epithelium. To distinguish between SPCs and NPCs, we co-stained for Gas1 expression and SIX2, a NPC-specific marker (Kobayashi et al., 2008) (Fig. S1K-T), and various stromal markers: PBX1 (Fig. S1U-D′), PDGFRα (Fig. S2A-J) and PDGFRβ (Fig. S2K-T). At E11.5, we detected Gas1 in SIX2⁺ NPCs and in a subset of PDGFRα⁺ and PDGFRβ⁺ stromal cells. Furthermore, PBX1 staining revealed distinct PBX1-high and PBX1-low expressing populations that also express Gas1 in the peripheral stromal (PBX1^HIGH) and presumptive NPC regions (PBX1^LOW). These findings are consistent with work suggesting that SPCs express higher levels of PBX1 than NPCs at E11.5 (Schnabel et al., 2003), and indicate that Gas1 is expressed by both NPCs and SPCs. Quantitation of β-GAL⁺ cells within PBX1^HIGH and PBX1^LOW populations revealed no significant differences in the percentage of β-GAL⁺/PBX1⁺ cells in each population at this stage (Fig. S1E′).

Gas1 mutant embryos display genetic background-dependent phenotypes in other tissues during development (Allen et al., 2007; Echevarria-Andino and Allen, 2020; Echevarria-Andino et al., 2022; Martinelli and Fan, 2007a; Seppala et al., 2007, 2014). Thus, we collected kidneys from wild-type (Fig. S3A,D), Gas1^+/− (Fig. S3B,E) and Gas1^−/− (Fig. S3C,F) mouse embryos maintained on a congenic 129S4/SvJaeJ background to investigate kidney morphogenesis. At E15.5, we observed normal kidney development in wild-type (Fig. S3A) and Gas1^+/− (Fig. S3B) embryos. In contrast to the variable renal defects observed in Gas1 mutants on a C57BL/6J genetic background, all Gas1^−/− embryos maintained on a congenic 129S4/SvJaeJ background exhibit bilateral renal hypoplasia (Fig. S3C), which was evident as early as E12.5 (Fig. S3F,G). To confirm that renal hypoplasia is not a result of decreased embryo size in Gas1 mutants, we measured kidney area (normalized to crown-rump length) in wild-type, Gas1^+/− and Gas1^−/− embryos maintained on congenic 129S4/SvJaeJ (Fig. S3H) and C57BL/6J (Fig. S3I) backgrounds. This quantitation revealed similar reductions in kidney size in Gas1 mutants compared to wild-type littermates (39.5% decrease, 129S4/SvJaeJ; 32.7% decrease, C57BL/6J). These data indicate that Gas1 is required for proper kidney development, and that Gas1 loss contributes to variable renal defects in a genetic background-dependent fashion.

Previous work identified a role for HH signaling in mesonephros development (Murashima et al., 2014), which precedes metanephros development (Fig. 2A). Given the role of Gas1 in HH signaling (Allen et al., 2011, 2007; Cobourne et al., 2004; Izzi et al., 2011; Lee et al., 2001; Ohazama et al., 2009; Seppala et al., 2007, 2022) and its expression at the onset of metanephros morphogenesis, we investigated a potential earlier contribution of GAS1 to mesonephros development, along with the HH co-receptors CDON and BOC. We used Gas1^lacZ, Cdon^lacZ and Boc^hPLAP reporter alleles (Cole and Krauss, 2003; Martinelli and Fan, 2007a; Zhang et al., 2011) to define HH co-receptor expression in the developing mesonephros (Fig. 2B-E; Fig. S4A-L). At E12.5, all three co-receptors are expressed in the mesonephros (Fig. 2C-E; Fig. S4D,F,J,L). Gas1 (Fig. 2C) and Boc (Fig. 2D; Fig. S4D,F) are broadly expressed in the mesenchyme, whereas Cdon expression is more restricted in this region (Fig. 2E; Fig. S4J,L). Boc and Cdon expression, but not that of Gas1, is present in the mesonephric tubules and nephric duct (ND) epithelium (Fig. S4D,F,J,L). We also examined HH co-receptor expression in the metanephros in E12.5 embryos, which shows MM-restricted Gas1 expression, whereas Boc is broadly expressed in the UB and MM, and lower levels of Cdon are observed in these areas (Fig. 2C-E; Fig. S4B,H). These data demonstrate partially overlapping, but not identical expression patterns for the HH co-receptors in early kidney development.

To investigate potential contributions of Gas1, Cdon and Boc to mesonephros development, we collected E12.5 wild-type, Gas1^−/−, Boc^−/− and Cdon^−/− mutant embryos, and performed whole-mount in situ hybridization (WISH) for Pax2, a marker expressed in the ND, ureteric epithelium and MM (Dressler et al., 1990) (Fig. 2F-J). Given functional redundancy between CDON and BOC in HH signaling, we generated Cdon;Boc double mutant embryos (Fig. 2K). In all embryos, Pax2 expression is detected in the ND and the metanephros. Notably, in a subset of Gas1 mutant embryos, Pax2 expression in the MM is reduced, consistent with the variable renal defects observed in these animals (Fig. 2H; compare with Fig. 1J,K). Pax2 is also present in the rostral mesonephric tubules in all embryos (Fig. 2F-K; gray arrowheads). Strikingly, in Gas1 mutants we observed ectopic Pax2-expressing mesonephric tubules in the caudal mesonephros (Fig. 2G,H; red arrowheads), phenocopying Shh conditional knockout mouse embryos (Murashima et al., 2014). Additionally, these ectopic Pax2⁺ mesonephric tubules are present in all Gas1^−/− embryos, regardless of the degree of agenesis in the metanephros (Fig. 2G,H), a phenotype confirmed by immunofluorescence for CDH1 and PAX2 in E12.5 wild-type and Gas1^−/− mesonephros sections (Fig. 2L-S). Analysis of Gas1^−/−embryos on a 129S4/SvJaeJ background (Fig. S5A-E) showed similar, but not identical, phenotypes to Gas1 mutants on a C57BL/6J background (Fig. S5F-H). Specifically, Gas1 mutants on a 129S4/SvJaeJ background display ectopic mesonephric tubules that are not fused to the ND (Fig. S5D; red arrowhead), while Gas1 mutants on a C57BL/6J background display ectopic mesonephric tubules that are attached to the ND (Fig. S5G; red arrowhead). In contrast to Gas1 deletion, individual or combined loss of Cdon or Boc results in normal mesonephrogenesis (Fig. 2I-K). Notably, deletion of these co-receptors also results in normal metanephros development in Boc^−/− (Fig. S4N), Cdon^−/− (Fig. S4O) and Cdon^−/−, Boc^−/− (Fig. S4P) embryos on a C57BL/6J background (Fig. S4M,Q). These data suggest that, in contrast to Gas1, Cdon and Boc do not contribute to kidney development.

Given the similar mesonephric abnormalities observed in our Gas1 mutants and the previously reported Shh conditional mutant embryos (Murashima et al., 2014), we examined whether Gas1 deletion correlates with decreased HH pathway activation in the mesonephros. We collected E12.5 wild-type and Gas1^−/−, along with Boc^−/− and Cdon^−/−;Boc^−/− mesonephroi, and performed WISH and section in situ hybridization (SISH) for Gli1, a direct HH pathway transcriptional target (Dai et al., 1999) (Fig. 3A-E,I). In all embryos, Gli1 is detected in the gonads and rostral mesonephros (Fig. 3A-D). Furthermore, Gli1 is expressed in the mesonephric mesenchyme adjacent to the ND epithelium in wild-type, Boc^−/− and Cdon^−/−;Boc^−/− embryos (Fig. 3A,C-E). In contrast, Gli1 expression is reduced in the caudal mesonephros of Gas1 mutants, consistent with the ectopic caudal mesonephric tubules observed in these embryos (Fig. 3B,I). To quantify changes in HH signaling, we performed fluorescent in situ hybridization (FISH) for Gli1 expression. Importantly, Gli1 is not detected in E12.5 Gli1^−/− mesonephroi sections compared to wild type, confirming probe specificity (Fig. S6A-H). Gli1 FISH in E12.5 wild-type (Fig. 3F-H) and Gas1^−/− (Fig. 3J-L) caudal mesonephros sections revealed a significant reduction in Gli1 in the mesenchyme surrounding the ND in Gas1 mutant embryos (Fig. 3M). ND epithelial expression of Shh is not significantly altered compared with wild-type embryos (Fig. S6I-Q), suggesting that reduced mesenchymal Gli1 near the ND in Gas1 mutants is not due to decreased levels of SHH ligand. Notably, abnormal Shh expression is detected in a subset of ectopic mesonephric structures in Gas1^−/− embryos (indicated by red arrowheads; Fig. S6N,P).

HH signaling is required for proper growth and differentiation of the ureteric mesenchyme during metanephros development (Bohnenpoll et al., 2017; Yu et al., 2002). Thus, we analyzed Gli1 expression in the metanephros in E12.5 wild-type (Fig. 3N-Q) and Gas1^−/− embryos (Fig. 3R-U) to assess whether Gas1 contributes to HH signaling in the metanephros. By WISH and FISH, we found that Gli1 is enriched in ureteral regions in E12.5 wild-type kidneys (Fig. 3N,P,Q). In Gas1^−/− mutants, Gli1 is maintained in kidneys that have undergone early UB branching morphogenesis or failed to branch (Fig. 3R, hypoplastic kidney; inset, agenic kidney, Fig. 3T,U). Quantitation of Gli1 FISH in E12.5 ureteric mesenchyme sections confirmed no significant changes in Gli1 in Gas1^−/− embryos (Fig. 3V). These data suggest that GAS1 plays a specific, HH-dependent role in mesonephros morphogenesis, while functioning independently of the HH pathway to drive metanephros development.

To investigate potential mechanisms contributing to renal hypoplasia and agenesis in Gas1 mutants, we collected wild-type and Gas1^−/− metanephroi, and analyzed NPCs (SIX2⁺), and MM (WT1⁺) and UB (CDH1⁺) proliferation or apoptosis by antibody detection of phospho-histone H3 (PHH3) or cleaved caspase 3 (CC3), respectively (Fig. 4A-Y; Fig. S7A-J). At the onset of kidney development (E10.5), the UB and MM are present at the hindlimb level in both wild-type and Gas1^−/− metanephroi; quantitation of MM (WT1⁺CDH1⁻; 34-37 somites) or NPCs (SIX2⁺CDH1⁻; 32-33 somites) revealed no significant differences upon Gas1 loss at this stage (Fig. 4G,H). However, the number of MM cells undergoing mitosis (PHH3+WT1+CDH1−) is significantly reduced in Gas1 mutants (Fig. 4I; Fig. S7A-J), whereas the percentage of proliferating UB cells (PHH3+CDH1+WT1−) remains unchanged (Fig. 4J). In contrast, we did not detect significant changes in MM or UB proliferation at the first UB branching event (E11.5, Fig. 4S,T); however, the metanephros area is reduced in Gas1 mutants (Fig. 4K,N), consistent with a transient decrease in MM cell proliferation at an earlier developmental stage (i.e. E10.5). This observation was confirmed by quantitation of WT1⁺ MM and SIX2⁺ NPCs, where we detected ∼50% (WT1⁺) and 65% (SIX2⁺) fewer cells in Gas1^−/− metanephroi at E11.5 (Fig. 4Q,R). At these earliest stages of metanephros development, few CC3⁺ cells were observed in wild-type (Fig. 4C,M) and Gas1^−/− (Fig. 4F,P) metanephroi. In contrast, we detected a large number of apoptotic NPCs (CC3⁺SIX2⁺) that fail to surround the UB in a subset of E12.5 Gas1 mutant metanephros, indicative of variable renal agenesis (Fig. 4U-Y). CC3⁺ quantitation confirmed no significant changes in apoptotic cell numbers in hypoplastic Gas1^−/− kidneys, whereas a subset of Gas1 mutant kidneys show increased cell death (Fig. 4Y). These results suggest that Gas1 transiently promotes MM proliferation in early kidney development, and that Gas1 loss contributes to MM death in a subset of Gas1^−/− kidneys, resulting in renal agenesis.

As shown previously (Fig. 1F), Gas1^−/− embryos display kidney defects resembling those observed in embryos lacking GDNF/RET signaling (Schuchardt et al., 1994). Therefore, we collected embryos at the onset of kidney morphogenesis to analyze the expression of several key RET pathway components (Fig. 5A-U; Fig. S8A-W). Specifically, we collected E10.5 wild-type and Gas1 mutant metanephroi sections, and performed RET antibody detection coupled with FISH for Gdnf, which encodes a secreted ligand for RET and the GDNF receptors (GFRα1-4) (Durbec et al., 1996; Jing et al., 1996; Treanor et al., 1996; Trupp et al., 1996), all components that are necessary for kidney formation (Cacalano et al., 1998; Moore et al., 1996; Pichel et al., 1996a,b; Sánchez et al., 1996; Schuchardt et al., 1994, 1996) (Fig. 5A-F). Consistent with previous reports (Kume et al., 2000), we observed Gdnf expression in wild-type MM at the onset of metanephros development (Fig. 5B,C). Gdnf expression was significantly reduced upon loss of Gas1 and, notably, to a similar degree in all Gas1^−/− metanephroi analyzed (Fig. 5D-G). Furthermore, we analyzed the expression of two cell-surface receptors of the RET pathway, Ret and Gfra1, in wild-type and Gas1^−/− metanephroi (Fig. S8A-P). Ret is expressed in the ND epithelium before UB outgrowth (Fig. S8A) and in the UB at the first branching event (Fig. S8C). Furthermore, Gfra1 is highly expressed in the UB at E10.5 and E11.0, whereas weak expression is observed in the adjacent MM at E11.0 in wild-type embryos (Fig. S8E,G). In Gas1 mutants, Ret and Gfra1 expression confirm delayed UB branching morphogenesis (Fig. S8D,H). However, we did not detect any overt changes in Ret or Gfra1 expression in these embryos (Fig. S8B,D,F,H), which was confirmed by antibody detection of RET and GFRA1 in the UB of wild type and Gas1 mutants with renal hypoplasia or agenesis (Fig. S8I-P; agenic, inset). Given the maintained expression of Ret and Gfra1, and similar reductions in Gdnf, these data suggest that additional factors contribute to the variable renal agenesis phenotypes observed in Gas1 mutant embryos.

Gas1 can modulate RET signaling in multiple contexts (Biau et al., 2013; Cabrera et al., 2006; Li et al., 2019; López-Ramírez et al., 2008). To determine whether Gas1 deletion alters downstream RET activation, we collected embryos at the onset of branching morphogenesis and performed WISH for Etv5 (Fig. S8Q-S) and Wnt11 (Fig. S8T-V), two transcriptional targets of RET signaling (Lu et al., 2009; Majumdar et al., 2003). Etv5 is expressed in the UB epithelium and surrounding MM (Fig. S8Q). In Gas1^−/− embryos, we observed two categories of Etv5 expression: (1) a mild reduction in metanephric Etv5 expression domain (Fig. S8R); and (2) an apparent loss of Etv5 expression in the UB and MM (Fig. S8S). Similarly, Wnt11 is expressed in the UB in E11.5 wild-type embryos (Fig. S8T), with either a mild reduction in the Wnt11 expression domain (Fig. S8U) or a loss of Wnt11 expression in Gas1^−/− metanephros (Fig. S8V). To quantify Wnt11 and Etv5 expression changes in the metanephros, we performed FISH coupled with RET antibody detection in E11.5 wild-type and Gas1^−/− coronal sections (Fig. 5H-U). Consistent with our WISH data, Gas1^−/− metanephroi display variable reductions in Etv5 expression in both the epithelium and the mesenchyme, as well as variably reduced Wnt11 expression in the UB (Fig. 5N,U; Fig. S8W), suggesting that the spectrum of renal agenesis phenotypes observed upon Gas1 loss results from variable reductions in RET signaling. Interestingly, we identified ectopic Etv5 and Wnt11 expression in Gas1^−/− mesonephroi (Fig. S8R,S, arrowheads; Fig. S8V, inset), suggesting increased RET signaling in an adjacent tissue.

To investigate the mechanisms underlying GAS1 modulation of RET signaling in metanephros branching morphogenesis, we collected E12.0 embryos and performed antibody detection of GAS1, co-stained with RET and HSPG2 (Perlecan) to identify the UB epithelium and basement membrane, respectively (Fig. 6A-J). While Gas1 (β-GAL⁺) appears to be exclusively expressed in the MM in Gas1^lacZ/+ reporter mice, we identified broad GAS1 protein expression that localizes to the MM and UB epithelium (RET⁺) in wild-type embryos (Fig. 6A-E). Notably, these data do not exclude possible low level epithelial Gas1 expression that is below the limits of β-GAL detection. Importantly, GAS1 expression is absent in Gas1^−/− metanephros sections, confirming antibody specificity (Fig. 6G,J). Epithelial GAS1 expression localizes to the cell membrane, with the most robust signal at the basal surface adjacent to the basement membrane (HSPG2⁺) and luminal side of the UB, similar to the regions of RET expression. Given this epithelial GAS1 distribution, we hypothesized that GAS1 is secreted from the MM and accumulates in the UB. To test this, we cultured primary MM cells from E18.5 wild-type and Gas1^−/− metanephros, and performed GAS1 immunoprecipitation from MM cell supernatants, along with NIH/3T3 fibroblasts and Gas1^−/− mouse embryonic fibroblast (MEF) supernatants (Fig. 6K; Fig. S9A). GAS1 was detected in both NIH/3T3 and wild-type MM supernatants, but not in Gas1^−/− MEFs and Gas1^−/− MM supernatants, suggesting that GAS1 can be released from the MM. The presence of multiple GAS1 bands likely indicates GAS1 N-glycosylation as described previously (Stebel et al., 2000) or could result from distinct GAS1 cleavage products. These data do not formally exclude GAS1 release via mechanisms distinct from secretion, including cell death.

Previous studies identified physical interactions between GAS1 and RET (Cabrera et al., 2006; Li et al., 2019); therefore, we performed proximity ligation assay (PLA) analysis coupled with CDH1 antibody detection in E11.5 wild-type and Gas1^−/− metanephros coronal sections to investigate potential interactions between RET and GAS1 in the UB (Fig. 6L-Y). To validate the assay, we performed PLA for RET and GFRA1, where physical interactions have been described previously (Jing et al., 1996; Treanor et al., 1996). We could detect GFRA1-RET PLA⁺ puncta in the UB epithelium in wild-type metanephros sections (Fig. 6L-N). In Gas1 mutant kidneys, we identified similar levels of GFRA1/RET PLA⁺ puncta in the UB compared to wild-type embryos, indicating that Gas1 deletion does not significantly alter GFRA1/RET interactions (Fig. 6O-Q,R). Next, we performed PLA for potential RET and GAS1 physical interactions in wild-type and Gas1 mutant kidneys (Fig. 6S-Y). Importantly, few PLA puncta were detected in Gas1^−/− UBs, whereas the levels of GAS1/RET PLA signal in wild-type embryos was significantly increased (Fig. 6Y). These data suggest a mechanism by which MM-secreted GAS1 can form physical interactions with RET in the UB epithelium to regulate early renal branching morphogenesis.

To test whether RET pathway activation can rescue the branching morphogenesis defects observed in Gas1 mutant kidneys, we cultured E11.5 control (wild-type or Gas1^+/−) and Gas1^−/− kidney explants for 4 days in the presence or absence of GDNF ligand. To confirm effective GDNF stimulation, we performed WISH for Wnt11 on day 4 kidney explants to readout downstream RET pathway activation (Fig. 6Z-D′). In control untreated kidneys, we identified many Wnt11-expressing UB tips (Fig. 6Z,D′), whereas untreated Gas1^−/− kidney explants had very few Wnt11⁺ branches or completely lacked UB branching (Fig. 6B′,D′), indicative of hypoplasia and agenesis, respectively. Addition of GDNF results in increased Wnt11 expression in control explants (1.89-fold increase in Wnt11+ branch tips compared with non-treated explants; Fig. 6A′,D′). Notably, GDNF-mediated RET activation induced UB branching in 100% of Gas1^−/− kidneys (a 12.94-fold increase in Wnt11⁺ branches compared with non-treated Gas1^−/− metanephros, branching to a similar degree as non-treated control kidneys; Fig. 6C′,D′). In contrast, Gas1^−/− kidney explants treated with smoothened agonist (SAG), which stimulates HH signaling downstream of GAS1, did not display a significant increase in UB branching, despite induction of HH pathway activity, as demonstrated by WISH for Gli1 expression (Fig. S9B-F). These data indicate that the UB branching defects and decreased RET signaling observed after Gas1 deletion can be selectively rescued by RET pathway activation.

Here, we have identified multiple contributions of GAS1 to mammalian kidney morphogenesis (Fig. 7). Gas1 is expressed in the early kidney mesenchyme and is selectively required for proper mesonephro- and metanephrogenesis in a genetic background-dependent manner. Specifically, Gas1 deletion results in ectopic mesonephric tubule formation (Fig. 7A), correlating with decreased HH signaling in the mesonephros (Fig. 7B). Furthermore, Gas1 deletion results in a transient decrease in MM cell proliferation and reduced RET target gene expression, contributing to early renal hypoplasia and agenesis. Importantly, the metanephros defects observed in Gas1 mutants can be rescued by increasing RET pathway activity via exogenous GDNF stimulation in Gas1 mutant explant cultures. Overall, these data highlight essential, multifunctional roles for GAS1 in early kidney morphogenesis.

Despite expression of all three HH co-receptors in developing mouse kidneys, we find that Gas1, but not Cdon and Boc, is essential for HH-dependent mesonephrogenesis. Notably, there are precedents for distinct co-receptor contributions to the development of other tissues. For example, during craniofacial development, BOC plays a partially antagonistic role that is distinct from GAS1 and CDON (Echevarria-Andino and Allen, 2020). CDON also restricts HH signaling during optic vesicle patterning in zebrafish (Cardozo et al., 2014). Furthermore, during limb development, only the loss of Gas1, but not Cdon and Boc, results in digit specification defects (Allen et al., 2011).

One explanation for these differential contributions to kidney development is structural differences between GAS1, a GPI-anchored protein resembling GFRs (Cabrera et al., 2006), and CDON and BOC, single pass transmembrane proteins (Kang et al., 1997, 2002). It is also possible that expression levels or cell type-specific co-receptor expression affects their relative contribution (e.g. mesenchymal Cdon expression appears more restricted than either Gas1 or Boc, while Cdon and Boc, but not Gas1, are expressed in the kidney epithelium and MM).

Evaluation of the combined contributions of all three HH co-receptors in later embryogenesis has been hindered by the early embryonic lethality of Gas1;Cdon:Boc germline mutants (Allen et al., 2011). However, the successful development of conditional alleles for Cdon and Gas1 (Bae et al., 2020; Jin et al., 2015) provides an opportunity to assess their combined contribution in the kidney mesenchyme.

GAS1 functions in several signaling pathways, including HH, RET, FGF and NOTCH (Allen et al., 2011, 2007; Biau et al., 2013; Cabrera et al., 2006; Izzi et al., 2011; Kann et al., 2015; Li et al., 2019; López-Ramírez et al., 2008; Marczenke et al., 2021; Martinelli and Fan, 2007a; Seppala et al., 2007). GAS1-mediated regulation of these pathways is thought to occur via direct physical interactions between GAS1 and different signaling molecules, including secreted HH ligands and various membrane receptors, such as PTCH1, RET and NOTCH1 (Cabrera et al., 2006; Izzi et al., 2011; Lee et al., 2001; Li et al., 2019; Marczenke et al., 2021). While GAS1 binds HH ligands with high affinity (Huang et al., 2022; Lee et al., 2001), lower affinity interactions have been reported between GAS1 and RET (Cabrera et al., 2006; Li et al., 2019; Rosti et al., 2015). In regions with overlapping GAS1/RET/HH expression, higher affinity interactions between GAS1 and HH might drive HH pathway activation, while restricting RET signaling. Notably, Shh and Ret are similarly expressed in the kidney epithelium before UB induction but become restricted to the mesonephros/distal ureter (Shh) and UB tips (Ret) around E11.5, while Gas1 maintains broad mesenchymal expression. Consistent with this notion, Gas1 deletion results in decreased HH signaling and ectopic Ret target gene expression in early mesonephrogenesis. Notably, in the metanephros, Gas1 deletion results in reduced Gdnf expression, which can contribute to a range of renal hypoplasia and agenesis phenotypes (Cullen-McEwen et al., 2001; Moore et al., 1996; Pichel et al., 1996a; Sánchez et al., 1996).

Our data indicate that GAS1 is secreted in the metanephros; therefore, the requirements for GAS1 in a specific pathway (e.g. HH versus RET) could depend on whether it acts in a cell-autonomous (membrane-tethered) or non-cell-autonomous (secreted) manner. GAS1 may also act on multiple signaling pathways simultaneously, as suggested for gut development, forebrain development and dentition patterning (Biau et al., 2013; Marczenke et al., 2021; Seppala et al., 2022). Alternatively, GAS1 may function in other signaling pathways. For example, Foxc1, Slit2 and Robo2 mutant mesonephroi resemble Gas1 mutants (Grieshammer et al., 2004; Kume et al., 2000).

Secreted HH-binding proteins are essential for proper HH pathway activity. SCUBE2 promotes HH ligand release and spread through tissues (Creanga et al., 2012; Tukachinsky et al., 2012), while HHIP, a secreted HH pathway antagonist, acts non-cell autonomously to restrict HH signaling in the developing ventral neural tube and lung (Holtz et al., 2015). The HH co-receptor BOC can be released from the cell surface, although membrane tethering of BOC and CDON is essential for HH-dependent neural patterning in the developing chicken neural tube (Song et al., 2015). Here, we identified GAS1 secretion from the kidney mesenchyme during development; however, potential functional roles for secreted GAS1 have not been explored.

Previous studies have described soluble GPI-anchored proteins (GPI-AP) in multiple developmental processes. For example, release of the GPI-AP CRIPTO is required for non-cell-autonomous activation of Nodal signaling and proper axial midline development (Chu et al., 2005; Lee et al., 2016; Parisi et al., 2003; Watanabe et al., 2007; Yan et al., 2002). Furthermore, GDE2-dependent cleavage and inactivation of the GPI-AP RECK regulates motor neuron differentiation through Notch inhibition (Muraguchi et al., 2007; Park et al., 2013; Sabharwal et al., 2011). Interestingly, two additional GDE proteins, GDE6 and GDE3, regulate multiple aspects of neurogenesis (Dobrowolski et al., 2020; McKean et al., 2023), similar to GAS1 promotion of HH-dependent neural patterning. Further reports show that ADAM10 and ADAM17 release GAS1 in cultured mesangial cells (van Roeyen et al., 2013). Whether GDE, ADAM or additional lipases or proteases mediate GAS1 release remains unknown.

This study raises questions about GAS1 presentation to various binding partners (e.g. secreted or membrane tethered), including its presentation to neighboring cells (e.g. cis versus trans). Membrane-bound and soluble GPI-anchored GFRs activate RET through cis and trans interactions, respectively, (Fleming et al., 2015; Jing et al., 1996; Ledda et al., 2002; Paratcha et al., 2001; Yu et al., 1998); however, trans GFRA1 signaling is dispensable for development in mice (Enomoto et al., 2004). Similarly, GAS1 requires membrane tethering to promote DLL1-mediated activation of NOTCH receptors (Marczenke et al., 2021). The development of genetic tools to selectively express membrane-anchored or secreted GAS1 will be essential to determine the contributions of GAS1 variants to kidney development and to a host of developmental signaling pathways.

In summary, our data suggest multiple models for GAS1 regulation of metanephrogenesis. One invokes a HH-dependent role for Gas1 through the regulation of Gdnf levels in the MM (Fig. 7C). Another consists of a HH-independent role, where secreted GAS1 promotes RET signaling during kidney development (Fig. 7D). Finally, GAS1 might regulate RET through trans interactions (Fig. 7E). It is also possible that multiple models may contribute to GAS1-dependent metanephrogenesis.

Gas1^lacZ (Martinelli and Fan, 2007a), Cdon^lacZ-2 (Cole and Krauss, 2003), Cdon^lacZ-1 (Cole and Krauss, 2003), Boc^AP (Zhang et al., 2011), Gli1^lacZ (Bai and Joyner, 2001) and Ret (Schuchardt et al., 1994) mice have been described previously (Table S1). Gas1 mice were maintained on congenic C57BL/6J and 129S4/SvJaeJ genetic backgrounds. Cdon and Boc mice were maintained on a congenic C57BL/6J background. Gli1, Ret and Cdon^lacZ-1 mice used for expression analysis were maintained on a mixed 129S4/SvJaeJ/C57BL/6J background. Noon on the day of the vaginal plugs was considered E0.5. All experiments were conducted using both male and female embryos. All animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Michigan.

Immunofluorescence was performed as described previously (Allen et al., 2011). Kidneys were harvested at various developmental stages and fixed in 4% paraformaldehyde for 20 min on ice, washed three times for 5 min each in 1×PBS (pH 7.4) and cryoprotected in 1×PBS+30% sucrose for 24 h. Tissues were embedded in OCT and cryo-sectioned on a Leica CM1850 cryostat at 12 μm. Sections were washed with 1×PBS and blocked for 1 h at room temperature in blocking buffer (1×PBS, 10% donkey serum, 0.3% Triton-X for goat anti-GAS1 and goat anti-GFRa1 antibodies; all other antibodies were blocked in 1×PBS, 3% bovine serum albumin, 1% heat inactivated sheep serum and 0.1% Triton-X 100). Slides were incubated in the indicated primary antibodies diluted in buffer overnight at 4°C in a humidified chamber. All AlexaFluor secondary antibodies were diluted 1:500 in blocking buffer and applied to the sections for 1 h at room temperature. To label nuclei, DAPI was applied for 10 min at a concentration of 1:30,000 in blocking buffer. Sections were washed three times for 5 min each in 1×PBS and mounted using Immu-mount mounting medium. General reagents (Table S2) and primary antibodies (Table S3) are provided in the Supplementary information. All images were obtained using a Leica SP5 upright confocal. Post-acquisition image processing was limited to brightness adjustments using Adobe Photoshop; when used, these changes were applied to the entire image, and were applied equally between control and experimental images.

Staining of whole kidneys was performed as described previously (Barak and Boyle, 2011). Kidneys were dissected in 1×PBS and fixed for 20 min in 4% PFA. They were washed three times for 5 min each in 1×PBS and blocked for 1 h in blocking solution (1×PBS, 0.1% Triton-X100 and 10% sheep serum). Kidneys were incubated in E-cadherin (1:100; BD Biosciences Cat#610181) antibody diluted in block solution overnight at 4°C. The following day, kidneys were washed four times for 1 h each in 1×PBS+0.1% Triton-X 100. AlexaFluor secondary antibodies were applied to tissues at a concentration of 1:500 and incubated overnight at 4°C. Kidneys were washed as described previously and images were collected using a Nikon SMZ1500 stereomicroscope.

RNA was extracted from E15.5 wild-type mouse kidneys using the Purelink RNA mini-kit (Invitrogen, 12183025) and reversed transcribed using the High-Capacity cDNA Reverse Transcriptase Kit (Applied Biosystems, 4368814). Gene-specific primers containing a T7 promoter sequence were designed for Wnt11 (forward, CCGCCACCATCAGTCACACCAT; reverse, TAATACGACTCACTATAGGGGACGTAGCGCTCCACCGTGC), Gfra1 (forward, CCTCGATGCAGCCAAGGCC; reverse, TAATACGACTCACTATAGGGTGGGAATCTCATTCTCAGAG) and Etv5 (forward, AGCCCACCATGTATCGAGAG; reverse, TAATACGACTCACTATAGGGTCAAAGGGCAAGCTTTAGGA) using Primer3 plus. cDNA was amplified by PCR and subcloned into TOPO 2.1 vector and linearized by restriction enzyme digestion. Digoxigenin-labeled antisense riboprobes were synthesized by in vitro transcription using a DIG RNA labeling kit (Roche, 11277073910), followed by purification using Micro Bio-Spin P-30 gel columns (Bio-Rad, 7326223). The purified RNA was quantified using a Nanodrop and diluted to a concentration of 20 ng/μl in hybridization buffer (50% formamide, 5×SSC (pH 4.5), 50 μg/ml yeast tRNA, 1% SDS and 50 μg/ml heparin).

In situ hybridization was performed as described previously (Allen et al., 2011; Wilkinson, 1992). Briefly, kidneys were harvested and fixed in 4% paraformaldehyde for 24 h at 4°C. Tissues were dehydrated through a 1×PBST+methanol series (25% methanol, 50% methanol and 75% methanol) and stored at −20°C in 100% methanol for up to 3 months. Embryos were washed in 1×PBS+0.1% Tween, bleached for 1 h in 3% hydrogen peroxide and digested with 10 μg/ml Proteinase-K for 5 min. Tissues were hybridized with the indicated DIG-labeled riboprobes diluted to 1 ng/ml in hybridization buffer (50% formamide, 5×SSC (pH 4.5), 50 μg/ml yeast tRNA, 1% SDS and 50 μg/ml heparin) at 70°C for 16-20 h. Kidneys were incubated in alkaline phosphatase-conjugated anti-DIG antibody (1:4000, Roche, 11093274910). For in situ detection, kidneys were incubated in BM purple substrate for up to 48 h depending on the probe. Tissues were post-fixed in 4% PFA+0.1% glutaraldehyde, cleared in 1×PBST+80% glycerol and imaged using a Nikon SMZ1500 stereomicroscope. Etv5, Wnt11 and Gfra1 RNA in situ probes were designed from E15.5 mouse kidney cDNA as indicated above. Pax2, Gli1 and Ret DIG-labelled RNA probes were gifts from Andrew McMahon (McMahon et al., 2008).

Kidneys were harvested and fixed in 4% paraformaldehyde for 24 h at 4°C. The next day, tissues were processed according to the immunofluorescence protocol and sectioned at 20 μm. Sections were thawed and post-fixed in 4% PFA for 10 min, followed by three washes of 5 min each in 1×PBS and digestion for 2 min in 10 μg/ml of proteinase K. Sections were fixed again in 4% PFA for 5 min and incubated in acetylation buffer (water, triethanolamine, hydrochloric acid and acetic anhydride) for 10 min. Sections were washed and dehydrated through a series of solutions (1×PBS, 0.85% NaCl, 70% ethanol and 95% ethanol). Gli1 probe was diluted to 1 ng/μl in hybridization buffer [50% formamide, 5×SSC (pH 4.5), 50 μg/ml yeast tRNA, 1% SDS and 50 μg/ml heparin] and slides were incubated at 70°C for 16-20 h. The next day, sections were washed with 1×SSC (saline sodium citrate at pH4.5)+50% formamide at 65°C for 30 min, followed by a TNE buffer [10 mM Tris (pH 7.5), 500 mM NaCl and 1 mM EDTA] wash for 10 min at 37°C, then incubation in 5 μg/ml RNase A in TNE at 37°C for 15 min. Slides were washed in a series of buffers (TNE, 2×SSC, 0.2×SSC and MBST), blocked in MBST+10% HISS+2% BMB for 2 h at room temperature and incubated in alkaline phosphatase-conjugated anti-DIG antibody (1:4000, Roche, 11093274910) diluted in blocking buffer until signal development. Slides were mounted with Glycergel mounting medium and imaged using a Nikon-E800 microscope.

Embryos were collected in 1×PBS (pH 7.4) and fixed (1% formaldehyde, 0.2% glutaraldehyde, 2 mM MgCl₂, 5 mM EGTA and 0.02% NP-40) for 1 h at room temperature. Tissues were cryoprotected in 1×PBS+30% sucrose overnight at 4°C, embedded in OCT and sectioned using a Leica CM1850 cryostat at 20 μm. Sections were incubated in staining solution [5 mM K₃Fe(CN)₆, 5 mM K₄Fe(CN)₆, 2 mM MgCl₂, 0.01% Na deoxycholate, 0.02% NP-40 and 1 mg/ml X-gal] until signal developed. Slides were washed three times for 5 min each in 1×PBS and counterstained with nuclear Fast Red for 5 min. Sections were washed in 1×PBS, dehydrated through a series of ethanol and xylenes (70% ethanol, 100% ethanol and 100% xylenes), and mounted using Permount mounting medium (Fisher Chemical, Cat#SP15100). Sections were imaged using a Nikon E-800 microscope.

Wild-type and Boc^AP embryos were dissected in 1×PBS and fixed for 1 h in 4% PFA. Embryos were cryoprotected in 1×PBS+30% sucrose for 24 h, embedded in OCT and sectioned on a Leica cryostat at 20 μm. Endogenous alkaline phosphatase was deactivated by incubating slides in 1×PBS for 30 min at 70°C. BM purple substrate was applied to the sections and incubated for 3-4 h at 37°C. After staining, slides were processed as described for X-gal staining and imaged on a Nikon E-800 microscope.

Embryos were dissected in 1×PBS and fixed in 10% neutral buffered formalin (NBF) for 20-24 h at room temperature. The following day, tissues were cryoprotected in 30% sucrose for 24 h, embedded in OCT and sectioned at 10 μm. Fluorescent RNAscope (ACD; Multiplex Fluorescent Assay Kit V2, 323110) was performed according to the manufacturer's instructions. Antigen retrieval was performed for 15 min at 100°C in a steamer and sections were treated with protease plus for 1 min, followed by incubation with the indicated probes. Antibody staining was performed as described previously for section immunofluorescence and slides were imaged on a Leica SP5 upright confocal.

For quantitation, images were manually thresholded using ImageJ and fluorescence signal was measured using the integrated density function normalized to a given area, as indicated in the figure legends. For mesonephric mesenchyme quantitation, Gli1 signal within 30 μm of the nephric duct was measured and normalized to mesenchyme area. The nephric duct and coelomic epithelium were excluded from the quantitation. Each data point represents one kidney. A minimum of three embryos per genotype and two sections from each kidney were analyzed.

Kidney cultures were performed as described previously (Costantini et al., 2011). Briefly, wild-type, Gas1^+/− and Gas1^−/− mutant mouse embryonic kidneys were harvested at E11.5 in ice-cold CO₂-independent medium (Gibco, 18045088). Kidneys were plated on a polyester membrane insert (0.4 µm pore size; 12 mm diameter, Corning, 3460) in a six-well tissue culture dish. 1.5 ml of kidney culture media (DMEM/F-12 supplemented with 10% calf serum and 1% penicillin-streptomycin glutamine) was applied to the bottom of each well below the membrane. For rescue experiments, recombinant murine GDNF was diluted to 75 ng/ml (Peprotech, 4504410UG). Smoothened agonist (SAG, Enzo Life Sciences, ALX-270-426-M001) was diluted to a final concentration of 300 nM in kidney culture media. Kidneys were cultured for 4 days at 37°C in a 5% CO₂ incubator. Fresh kidney culture media was applied after 2 days of culture. For whole-mount in situ hybridization of kidney explant cultures, after 4 days of culture, kidney explants were fixed on the transwell membranes in 4% paraformaldehyde for 24 h at 4°C in a 6-well plate. The next day, kidney explants were washed three times for 5 min each in 1×PBST and processed according to the whole-mount in situ hybridization protocol with a 2 min proteinase-K digestion. Kidneys were imaged using a Leica SMZ1500 Stereomicroscope.

Wild-type and Gas1^−/− kidneys were dissected at E18.5 in CO₂-independent medium, incubated in 2 mg/ml of collagenase diluted in CO₂-independent medium (two kidneys pooled) for 20 min at 37°C and cells were dissociated by gentle pipetting. Cells were pelleted and resuspended in culture media (DMEM/F12+10% calf serum+1% PSG), plated in a six-well dish and cultured for 2-3 days at 37°C and 5% CO₂. NIH/3T3 fibroblasts (ATCC, CRL-1658) and immortalized Gas1^−/− MEFs (generated according to the method of Todaro and Green, 1963) were cultured in DMEM, high glucose+10% calf serum+1% PSG and routinely tested for contamination. Cell supernatants were collected and centrifuged to remove dead cells. For immunoprecipitation, 50 µl of protein-G Dynabeads (Invitrogen, 10-003-D) were washed in 1×PBS then incubated in goat anti-GAS1 (2 mg, R&D Systems, AF2644) antibody diluted in 1×PBS for 2 h at room temperature on a rotating platform. Beads were washed twice with 1×PBST, then incubated with 1 ml of cell supernatant for 16-24 h on a rotator at 4°C. The next day, beads were briefly washed four times in 1×PBS +0.1% Tween, then four times in 1×PBS+0.1% Triton. Proteins were eluted in 30 µl of 0.1 M citrate (pH 2.0) for 15 min on a hula mixer. To neutralize the pH, 3 µl Tris buffer (pH 8.0) was applied to each sample. For denaturation, samples were boiled at 95°C in 6×Laemmli buffer for 10 min. Proteins were separated by SDS-PAGE on a 10% gel, then transferred onto a Immun-Blot PVDF membrane (Bio-Rad, 162-0177). After transfer, membranes were blocked in western blocking buffer [30 g/l bovine serum albumin with 0.2% NaN₃ in TBST (Tris-buffered saline, 0.5% Tween-20)] for 1 h at room temperature. Membranes were incubated in goat anti-GAS1 (1:4000, R&D Systems, AF2644) primary antibody overnight at 4°C. The next day, membranes were washed four times for 10 min each in 1×TBST, followed by a 1 h incubation in peroxidase-conjugated AffiniPure secondary antibodies (1:10,000, Jackson ImmunoResearch) and four 10 min washes in 1×TBST. Membranes were incubated in Amersham ECL Prime western blot detection reagent (Cytiva, RPN2232) for 5 min at room temperature and developed using a Konica Minolta SRX-101A medical film processor.

Embryos were dissected in 1×PBS and processed as described above for section immunofluorescent staining. A Duolink Proximity Ligation Assay (PLA; Sigma, DUO92007) was performed according to the manufacturer's instructions. Sections were blocked (1×PBS+10% donkey serum+0.3% Triton) for 1 h at 37°C, then incubated with rabbit anti-RET (1:200, Cell Signaling, 3223) and goat anti-GAS1 (1:900, R&D Systems, AF2644) or goat anti-GFRA1 (1:500, R&D Systems, AF560) overnight at 4°C. The next day, slides were incubated with PLA Probe Anti-Rabbit PLUS (Sigma, DUO92002) and PLA Probe Anti-Goat Minus (Sigma, DUO92006). After the final washes, slides were processed as described previously for section immunofluorescent staining for detection of mouse IgG2a anti-E-cadherin (1:500, BD Biosciences, 610181) and imaged on a Leica SP5 upright confocal. For quantitation, PLA puncta were manually counted using the cell counter tool on ImageJ and normalized to the total number of CDH1⁺ cells. A minimum of four kidneys per genotype and three sections/kidney were analyzed.

All statistical analyses were performed using GraphPad statistic calculator or GraphPad Prism (www.graphpad.com) and data are reported as mean and standard deviation. Statistical significance was determined using a two-tailed Student's t-test or an ordinary one-way ANOVA with Tukey's multiple comparisons test. For all experiments, a minimum of three kidneys were analyzed for each genotype. The number of replicates and P-values are listed in each figure. P<0.05 was considered to be significant.

Whole-embryos and isolated kidneys were imaged using an SMZ1500 Stereomicroscope. Kidney area was measured using the freehand tool in ImageJ and normalized to the embryo length. Crown-rump length was defined as the top of the crown to the bottom curvature of the embryo at the anterior hindlimb level.

The number of WT1+ (MM), SIX2+ (NPCs) and PHH3+ cells was manually counted using the cell counter tool feature in ImageJ (Schneider et al., 2012). PHH3⁺ cells were normalized to total E-cadherin⁺ (CDH1, UB) or WT1⁺ cells. The coelomic epithelium was not included in the MM or UB quantitation. For cleaved caspase 3 quantitation, SIX2⁺ cells were used to identify the metanephric area. In serial sections from the same animal, the number of CC3⁺ cells per metanephric area was measured using the analyze particle tool in ImageJ. For PBX1 quantitation in E11.5 Gas1^lacZ/+ coronal metanephros sections, images were manually thresholded in ImageJ to only include cells with saturated PBX1-expression (PBX1^HIGH). The remaining PBX1^LOW-expressing MM cells were quantified using the manual cell counter on ImageJ. Within the total PBX1⁺ populations, β-GAL⁺ cells were manually quantified and normalized to either PBX1^LOW- or PBX1^HIGH-expressing cells. The same threshold settings were applied to all images for quantitation. A minimum of three kidneys per genotype and two sections/kidney were analyzed.

We thank Dr Cristina Cebrian for assistance with the kidney explant cultures. We thank Dr Greg Dressler for providing valuable suggestions and reagents. We acknowledge the BRCF Microscopy Core for confocal microscope access.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.203012.reviewer-comments.pdf