Reciprocal inhibition of Wnt signaling pathways pattern the interconnection of epithelial tubules in the regenerating zebrafish kidney
Caramai N. Kamei, William G. B. Sampson, Carolin Albertz, Oliver Aries, Amber Wolf, Rohan M. Upadhyay, Samuel M. Hughes, Heiko Schenk, Frederic Bonnet, Bruce B. W. Draper, Kyle W. McCracken, Denise K. Marciano, Leif Oxburgh, Iain A. Drummond
TL;DR
The study shows how Wnt signaling pathways work together and in opposition to guide kidney tubule connections during zebrafish kidney regeneration.
Contribution
The novel contribution is the discovery that a single Wnt ligand can activate both canonical and non-canonical pathways in the same cells to control tubule interconnection.
Findings
Canonical Wnt signaling induces invasive cell protrusions in new nephrons.
Non-canonical Wnt signaling via frizzled9b restricts canonical Wnt and drives apical constriction.
Wnt ligands wnt9b and wnt4 are essential for new nephron formation and tubule interconnection.
Abstract
In the adult zebrafish kidney, nephrogenesis occurs as a regenerative response to injury and provides a model to explore cell signaling pathways required for nephron formation and engraftment. Differentiating kidney tubules interconnect with collecting system epithelia to generate a pathway for fluid excretion. We show that canonical Wnt signaling induces a mesenchymal, invasive cell phenotype and is required, along with Src kinase and Rac1, to generate basal cell protrusions on new nephrons. The Wnt ligands wnt9b and wnt4 are both required for new nephron formation after injury. Mutation in wnt4 and wnt9b, or treatment with the canonical Wnt inhibitor IWR1 blocks the formation of basal protrusions in forming nephrons. Mutation in the Wnt receptor frizzled9b reveals a fusion-associated non-canonical Wnt pathway that acts to (1) restrict canonical Wnt gene expression, (2) drive Rho…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8- —National Institutes of Health
- —National Institute of Diabetes and Digestive and Kidney Diseaseshttp://dx.doi.org/10.13039/100000062
- —National Science Foundationhttp://dx.doi.org/10.13039/100000001
- —National Institute of General Medical Scienceshttp://dx.doi.org/10.13039/100000057
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsRenal and related cancers · Developmental Biology and Gene Regulation · Wnt/β-catenin signaling in development and cancer
INTRODUCTION
Embryonic development and epithelial organogenesis require fusion of cell sheets or tubules to create polarized transport and secretory structures that are crucial for tissue function (Palmer and Nelson, 2020). Epithelial fusion is a complex process involving regionally specialized cells that partially uncouple from cell neighbors to undergo an incomplete mesenchymal transformation with the formation of invasive or exploratory actin-based cell protrusions (Cote and Feldman, 2022). These transient changes in cell adhesion and polarity allow cell rearrangements and integration with new cell neighbors, followed by resolution of cell junctions, stabilization of cell adhesion and restoration of a uniform epithelial cell polarity (Cote and Feldman, 2022). Examples of epithelial sheet fusion include the morphogenetic process of neural tube closure (Pai et al., 2012), optic cup fusion in the developing eye (Bernstein et al., 2018; Gestri et al., 2018) and palate fusion (Yu et al., 2009). These fusion events require cell movement, lamellipodia extension, initiation of new cell adhesions and, finally, stabilization of uniform cell polarity. The fusion of two epithelial tubules has the added requirement of invasion into tubule basement membranes, as basal to basal cell apposition must be overcome to allow tubules to join lumens (Kao et al., 2012; Yu et al., 2009). Currently, the cell signaling processes that mediate vertebrate tubule fusion are not well understood.
Epithelial tubule fusion is observed during development of the Drosophila tracheal system (Gervais et al., 2012), in the vertebrate pancreas (Slack, 1995) and in the avian lung (Palmer and Nelson, 2020). In Drosophila, FGF signaling plays a prominent role in guiding the fusion of tracheal tip cells and supporting lumenal interconnection between single cell thick tubules (Sutherland et al., 1996). In the chick lung, an elongating branched multicellular epithelium forms a continuous network of airways by a mass fusion event involving changes in epithelial shape at apposing tubule tips, formation of bridging cytoskeletal protrusions, removal of intervening basement membrane and cell apoptosis at tubule junctions, making a path for lumen fusion (Palmer and Nelson, 2020). Signals driving avian epithelial budding and fusion are currently not known.
In the embryonic mammalian kidney, epithelial tubule fusion is required for the integration of newly forming epithelial nephron tubules into an arborized network of filtration units (Georgas et al., 2009; Kao et al., 2012). This process is characterized by a thinning of tubule basement membranes, invasion of new nephron cell basal protrusions, and formation of a connecting lumen between the distal end of S-shaped body nephron intermediates and the branched tips of the ureteric epithelium that will form the collecting system (Hiremath et al., 2023; Kao et al., 2012; Yang et al., 2013). This fusion event requires distal patterning of the invading S-shaped body, implying that invasive interconnecting cell behavior is a differentiated phenotype of the distal tubule (Kao et al., 2012).
In the adult zebrafish kidney, nephron addition and tubule fusion occur throughout life as the kidney grows (Diep et al., 2011; Reimschuessel, 2001; Zhou et al., 2010). Nephron addition also occurs in response to acute kidney injury where adult renal progenitor cells migrate to form rosette-like aggregates on distal tubules prior to new nephron outgrowth, patterning and interconnection (Gallegos et al., 2019; Kamei et al., 2019). We have previously shown that cell aggregation and rosette formation require FGF signaling (Gallegos et al., 2019), while tubule outgrowth requires canonical Wnt signaling (Kamei et al., 2019). Here, we demonstrate that the final step of new nephron integration requires both canonical and non-canonical Wnt signals acting in a single layer of invading tubule cells. We also identify downstream morphogenetic processes involved in the generation of Src- and Rac1-dependent basal protrusions, and show that non-canonical frizzled9b signaling is required to orient invasive processes and coordinate apical and basal constriction of invading tubule cells. Our results illuminate multiple signaling events required for normal nephron tubule interconnection and reveal how mutually repressive signals pattern the tubule fusion interface.
RESULTS
Regenerating new kidney nephrons form lumenal connections by invasion of existing distal tubules
We took advantage of the synchronous addition of up to 100 new nephrons at 7-8 days post-acute kidney injury (Kamei et al., 2015) to dissect cell signaling and morphogenetic processes guiding tubule fusion. A diagram of adult kidney structure and the regenerative process is depicted in Fig. 1A. After acute injury, Tg(lhx1a:eGFP)+ zebrafish kidney progenitor cells first aggregate to form dome-shaped rosettes Fig. 1A and then undergo a proliferative burst to extend a newly forming kidney tubule that ultimately becomes a fully functional nephron (Fig. 1A) (Diep et al., 2011). To acquire function, initially blind-ended new nephrons must fuse with existing distal tubules to allow fluid filtration. Tg(lhx1a:eGFP)^+^ new nephrons stained with phalloidin to visualize tubule lumens were imaged before (Fig. 1B; Movie 1) and after tubule fusion (Fig. 1C; Movie 2) 7-8 days after gentamicin-induced acute kidney injury. New nephron epithelial cells are apically constricted before fusion (Fig. 1B,D,F) and transition to a basally constricted morphology during the process of tubule fusion (Fig. 1C,E,G). Preceding interconnection, the basal surfaces of newly formed lhx1a:eGFP+ nephrons adjacent to a distal tubule segment exhibit multiple basal protrusions that invade the target distal tubule, preceding interconnection of new and existing tubule lumens (Fig. 1H; Movie 3). These basal protrusions are composed of a central core of filamentous actin and form at the periphery of the rosette, as well as directly under the forming lumen of new nephrons (Fig. 1B, inset, H).
Stages of epithelial tubule interconnection. (A) Diagram of the adult zebrafish kidney and stages of regeneration. Single lhx1a:eGFP+ mesenchymal cells (green) aggregate into dome-shaped rosettes on distal tubules (blue) and then proliferate to extend a new kidney nephron (red; glomerulus). Lumenal fusion (white) to the existing distal tubules enables new nephron function. (B) Prior to tubule fusion, phalloidin-stained (magenta) apical cell actin highlights a new tubule epithelial lumen (arrow indicates the direction of fusion) in the process of nephron fusion to the distal tubule (DT; dotted outline). Tg(lhx1a:egfp) (green) marks newly forming new nephrons adjacent to target distal tubules (DTs). Junctional nephron cells are constricted apically. Inset in B shows a higher-magnification view of a basal protrusion (arrowhead) with a phalloidin-positive (magenta) actin core. (C) After tubule fusion, phalloidin staining (magenta) outlines the apical cell actin, and lumen continuity between old and new nephron tubules (arrow indicates direction of fusion). DT, distal tubule (outlined). (D,E) New nephron cells in rosettes are apically constricted (D) and become basally constricted in the process of lumen fusion (E) (a, apical surface; b, basal surface). (F,G) Diagram of cell boundaries drawn from D and E (a, apical surface; b, basal surface). (H,I) Basal protrusions (H, arrowheads) invade the distal tubule (DT) at the point of tubule fusion; these cells are non-proliferating (I, arrowheads). (J,K) New nephrons lack laminin-positive basement membranes at their basal surface point of fusion, marked by Tg(lhx1a:egfp) expression (K, arrowheads). Blue fluorescence is Hoechst-stained nuclei. Arrows indicate the direction of fusion. Scale bars: 5 µm in C-K; 10 µm in B.
New nephron cells closest to the target distal tubule do not proliferate, as revealed by the lack of EdU uptake in invading cells (Fig. 1I). Proliferating, EdU-positive cells in the new nephron extend the nascent tubule (Kamei et al., 2019) and develop a laminin-containing basement membrane, while invasive cells closest to the point of future interconnection are laminin negative (Fig. 1J,K). To confirm the observed invasive phenotype with a molecular signature, we compared a previously reported microarray transcriptome of Tg(lhx1a:eGFP)^+^ cells with the gene set represented in the human cancer metastasis database (Diep et al., 2011; Zheng et al., 2018). Zebrafish orthologs of genes associated with metastatic cell invasive behavior were expressed at points of nephron connection, including invadopodia-associated metalloproteases mmp14a and mmp14b, the SH3 domain protein tks5 (sh3pxd2aa), the Wnt ligand wnt4, canonical Wnt targets lef1 and cdh11 implicated in invasive cell behavior, and transcriptional regulators of invasiveness, id1 and jun (Lambert et al., 2017; Zheng et al., 2018) (Fig. S1). Taken together, the data indicate that a narrow band of new nephron cells at the point of tubule fusion maintain a non-proliferating, invasive mesenchymal character, while new nephron cells further from the point of interconnection develop an epithelial basement membrane and proliferate, extending the nascent epithelial tubule.
Canonical Wnt signaling stimulates tubule invasion
Wnt signaling is required for zebrafish new nephron outgrowth and Wnt ligands wnt9a and wnt9b are induced after acute kidney injury specifically in distal tubules where new nephrons form (Kamei et al., 2019). To better characterize expression domains of known canonical Wnt signaling ligands and targets*,* we performed confocal 3D imaging of wnt4 and lef1 expression by fluorescent in situ hybridization. Both wnt4 and lef1 were expressed in a restricted pattern in cells at the point of tubule fusion (Fig. 2A,B), consistent with high canonical Wnt signaling in this domain (Carroll et al., 2005; Kuure et al., 2007). The zebrafish canonical Wnt reporter Tg(tcflef:dGFP) also showed highest expression at the point of contact with the distal tubule (Fig. 2C; Shimizu et al., 2012).
*Canonical Wnt signaling highlights the point of tubule invasion. (A,B) Canonical Wnt target genes lef1 (A) and wnt4 (B) are expressed in new nephron tubules (marked by EdU incorporation) at the tubule fusion junction with the distal tubule (DT; dotted outline) marked by Tg(lhx1a:eGFP) expression. (C) The Tg(tcflef:dGFP) Wnt reporter transgene is expressed at the point of tubule fusion. (D,E) The canonical Wnt inhibitor IWR1 blocks formation of invasive basal protrusions (arrowheads). (F) Quantification of basal protrusion phenotype by Kappa curvature analysis in ImageJ reveals loss of protrusions after IWR1 treatment (Mary and Brouhard, 2019 preprint; see Materials and Methods). Green in B is autofluorescence. Blue fluorescence is Hoechst-stained nuclei. Scale bars: 5 µm. *P=0.0012 (Mann-Whitney test). Data are mean±s.d. (control: 0.6986±0.2750, n=9; IWR1 median: 0.2204±0.1136, n=7).
Previous work from our lab has demonstrated that blocking canonical Wnt signaling for 4 days (3-7 dpi) blocked proliferative outgrowth of new nephron aggregates and also indicated that inhibiting canonical Wnt signaling completely blocked invasive tubule interconnection, giving rise to small, quiescent aggregates (Kamei et al., 2019). To more closely examine whether canonical Wnt signaling was required to maintain the invasive characteristic of new nephron cells, we treated gentamicin-injured zebrafish with the canonical Wnt inhibitor IWR1 for 24 h at 7 days post-injury (dpi) and imaged the new nephron basal surfaces at 8 dpi for invasive basal protrusions. Twenty-four hour IWR1 treatment eliminated basal protrusions (Fig. 2D,E). The average basal surface curvature of multiple new nephrons (n=7) measured using ImageJ Kappa (see Materials and Methods; Mary and Brouhard, 2019 preprint) showed a significant reduction in basal protrusions (P<0.0014) in IWR1-treated fish (Fig. 2F), indicating a requirement for canonical signaling for invasive activity.
Wnt ligands wnt9b and wnt4 are required for new nephron tubule formation
wnt9a and wnt9b are both induced by acute kidney injury in the adult collecting system and distal tubules, which are the site of new nephron addition (Kamei et al., 2019), while wnt4 is expressed in new nephrons at the point of contact between new nephrons and target distal tubules (Fig. 2B), making these Wnts candidate signaling ligands for regulating new nephron formation. Mutation in wnt4 (Kossack et al., 2019) (wnt4^uc55^, referred to as wnt4^−/−^ hereafter) (Fig. 3B,E,G) or wnt9b (Liu et al., 2022) (wnt9b^fb207^, referred to as wnt9b^−/−^ hereafter) (Fig. 3C,F,H) markedly reduced new nephrogenesis following acute gentamicin injury compared to control (Fig. 3A), with Fig. 3G,H representing quantification of in situ hybridization results. However, adult mesonephric nephron number, as determined by counting nephrin^+^ glomeruli in uninjured adult kidneys, was not affected by either mutation (Fig. S2A,B). The results show that both wnt9b and wnt4 are not required for initial mesonephric kidney development, and that adult kidney morphology is not significantly affected. wnt9b and wnt4 are instead required for new nephron formation following injury and represent regeneration-specific, inducible Wnt ligands. Normal mesonephric development and residual new nephrogenesis in wnt4 and wnt9b mutants may be due to low level expression of Wnt paralogs in the mutants (Fig. S3A,B). We also find that wnt9b expression was unaffected in 7 dpi wnt4^−/−^ mutants (Fig. S3C,D).
*wnt9b and wnt4 are required for new nephron formation and invasive cell behavior. (A) Wild-type adult kidney shows a robust induction of lhx1a-expressing new nephrons in response to gentamicin acute injury 7 days post-injection. (B,C) Homozygous mutation in wnt4 (B) or wnt9b (C) significantly reduces new nephron formation after acute gentamicin injury. (D,E) Lumen formation and basal cell protrusions (arrowhead in D; DT, distal tubule, dotted outline) seen in wild-type lhx1a+ new nephrons are not observed in wnt4 mutant new nephrons (E, arrowhead). (F) New nephrons in wnt9b mutants lack extensive lhx1a:eGFP fluorescence and do not show basal protrusions (arrowhead). Blue fluorescence indicates Hoechst-stained nuclei. (G,H) Quantification of lhx1a+ new nephron aggregates (by in situ hybridization) in wnt4 (G) and wnt9b (H) heterozygous and homozygous mutants reveals a significant reduction in nephrogenesis after injury. G: *P=0.0242, **P=0.0033; H: *P=0.027, *P=0.0061 (ordinary one-way ANOVA). Data are mean±s.d. (wnt4 +/+: 5.86±4.44, n=13, wnt4 +/−: 6.96±4.64, n=17; wnt4 −/−: 2.34±3.31, n=15; wnt9b +/+: 13.22±8.12, n=7; wnt9b +/−: 7.5±3.83, n=7; wnt9b −/−: 3.38±2.61, n=8). Scale bars: 0.2 mm in A-C; 10 µm in D; 5 µm in E,F.
Rare nephron cell aggregates that did form in 7 dpi wnt4^−/−^ mutant kidneys remained small, and confocal sections of image stacks showed no invasive basal protrusions or significant tubule outgrowth (Fig. 3E; n=7) compared with wild type (Fig. 3D, Movie 4), although variable EdU incorporation was detected in wnt4^−/−^ new nephrons (Movie 5). wnt4 mutant new nephrons did not exhibit tubule lumen formation, as observed in 3D maximum intensity projections (Movie 5). Injury-induced wnt9b expression was not affected by mutation in wnt4 mutants (Fig. S3). wnt9b mutation strongly reduced the generation of new nephrons (Fig. 3C) but in those that formed, lumen formation did occur with some tubule extension, but no invasive basal protrusions were detected (Fig. 3F; n=6). Overall, we observed that, for both wnt4 and wnt9b mutants, new nephrogenesis was strongly reduced, basal protrusions were not generated in the new nephrons that did form, and proliferation and tubule outgrowth was highly variable. We also note that Tg(lhx1a:eGFP) expression was reduced in wnt9b mutants and restricted to cells immediately adjacent to the distal tubules in the wnt9b mutant (Fig. 3F) compared to wild type (Fig. 3D), consistent with a potential requirement for canonical Wnt signaling in regulation of lhx1a gene expression (Guo et al., 2021).
Src kinase and Rac1 signaling are required for basal protrusion formation
Cellular protrusions, including invadopodia, are often initiated by localized Src family kinase activity (Barbayianni et al., 2023). To determine whether Src kinases were involved in tubule interconnection, we treated gentamicin-injured zebrafish at 7 dpi with the Src inhibitor PP2 and assessed basal surface curvature in confocal image projections. Twenty-four hour treatment with PP2 at 7-8 dpi eliminated basal protrusions on new nephron basal surfaces, while the negative control PP3 had little effect (Fig. 4A-C). Assessment of new nephron average basal surface curvature in PP2-treated zebrafish using ImageJ Kappa revealed a significant inhibition of protrusive activity (n=13, P<5 E-05) (Fig. 4C). Basal protrusions have also been shown to require the GTPase Rac1 to direct actin-based protrusive activity (Sun et al., 2017). Involvement of Rac1 was tested using the inhibitor EHT1864 (Wang et al., 2020). EHT1864 treatment disrupted the organization of actin filaments in new nephron basal protrusions (Fig. 4D,E). The normally central core of actin filaments in protrusions (Fig. 4D,D′) was dispersed to the plasma membrane and basal protrusions had a swollen, bleb-like appearance (Fig. 4E,E′). Basal protrusion cross-sectional area of EHT1864-treated fish was significantly increased (Fig. 4F). These results reveal that invasive cell behavior associated with tubule interconnection requires Src kinases and Rac1 activity.
*Src kinases and Rac1 function in basal protrusion formation. (A,B) Control-treated kidney new nephrons (A; PP3) show invasive basal protrusions (arrowhead) that are missing in Src family inhibitor-treated kidneys (B; PP2; arrowhead indicates basal surface). (C) Quantification of surface curvature revealed PP2 caused a significant loss of basal protrusions (DT, distal tubule). Data are mean±s.d. (PP3: 0.47±0.17, n=19; PP2: 0.23±0.09, n=13). (D) Basal protrusions in control new nephrons (arrowhead) are formed with a central actin core (phalloidin; magenta) and maintain a narrow invasive morphology. (D′) Enlarged magnification of a slice from the same image stack as in D highlights the actin core structure (phalloidin; magenta) of protrusions. (E) Rac1 inhibition with EHT1864 caused mislocalization of actin to basal protrusion cell membranes, distension of protrusions and a ‘blebbing’ morphology (arrowhead). (E′) Enlarged magnification of a slice from the same image stack as in E. (F) EHT1864-induced distension of basal protrusions quantified by protrusion cross-sectional area. Data are mean±s.d. (control: 2.21±1.13, n=31; Rac1 inhib: 5.262±3.13, n=34). Blue fluorescence is Hoechst-stained nuclei. Scale bars: 5 µm in A,B,D,E; 2 µm in D′,E′. ***P<0.0001 (unpaired t-test).
frizzled 9b mediates a non-canonical Wnt signaling pathway in tubule fusion
Frizzled 9b (fzd9b) is a Wnt receptor expressed in undifferentiated adult zebrafish kidney progenitor cells (Fig. 5A) and in newly forming nephron cell aggregates directly adjacent to the distal tubule (Kamei et al., 2019). As aggregates polarize and grow in size, fzd9b expression is downregulated in cells closest to the distal tubule (Fig. 5C), which is a domain of high canonical Wnt signaling (Fig. 2A-C). Notably, fzd9b was not downregulated in wnt4 mutant new nephrons (Fig. 5D; Fig. S4), implying Wnt4 signaling represses fzd9b expression. We next assayed fzd9b Wnt pathway activation using the Tg(tcflef:dGFP) canonical Wnt reporter and endogenous canonical Wnt target gene expression in fzd9b homozygous mutant kidneys. Mutation in fzd9b resulted in expanded expression of the Tg(tcflef:dGFP) reporter (Fig. 5E,F) and the endogenous canonical Wnt target gene wnt4 (Fig. 5G,H). Similarly, the expression domains of canonical Wnt target transcription factors lef1 and lhx1a, and Wnt-induced negative-feedback inhibitors notum1a and wif1 were also expanded in fzd9b mutant sections (Fig. S5I,J, quantification of wnt4 and lef1 expression domains), suggesting that fzd9b signaling acts to antagonize canonical Wnt signaling at the point of tubule interconnection. We have previously shown that canonical Wnt signaling is required for cell proliferation in new tubule aggregates (Kamei et al., 2019). Consistent with an antagonistic function for fzd9b, cells at the point of interconnection in new tubules that normally do not proliferate (Figs 1I and 6A), exhibit active cell proliferation in the fzd9b mutant (Fig. 6B, quantified in E; Movies 6-9 illustrating the approach). We also found that ectopic cell proliferation in fzd9b mutants was blocked by treatment with the canonical Wnt inhibitor IWR1 (Fig. 6E), further supporting the idea that signaling via Frizzled9b suppresses canonical Wnt signaling. Taken together, the data indicate that fzd9b antagonizes canonical Wnt signaling while canonical Wnt signaling via Wnt4 acts to repress fzd9b expression, suggesting that mutually repressive Wnt signaling pathways pattern the interconnection of epithelial tubules. In support of a non-canonical role for fzd9b signaling, we found additional non-canonical Wnt signaling components, prickle2b and ptk7a, were expressed in newly forming kidney tubules (Fig. S6).
Mutual repression by frizzled9b and wnt4 signaling patterns tubule junctions. (A) fzd9b is expressed in single nephron progenitor cells prior to aggregation. (B,C) After acute injury, fzd9b expression persists in new cell aggregates (B) but is downregulated in cells adjacent to the distal tubule (C; DT, dotted outline), a domain of high canonical Wnt signaling (asterisk in C). (D) fzd9b expression in wnt4 mutant new nephrons escapes downregulation and remains uniformly expressed in new nephrons. (E) Expression of the canonical Wnt reporter Tg(tcflef:dGFP) is restricted to new nephron cells adjacent to the distal tubule (DT) and is not expressed in cells several cell diameters away (asterisk). (F) Mutation in fzd9b results in expanded expression of Tg(tcflef:dGFP) where the entire new nephron can show GFP expression (asterisk). (G) wnt4, an endogenous canonical Wnt target gene, is expressed in new nephron cells adjacent to the distal tubule. (H) Mutation in fzd9b significantly expands wnt4 expression. Blue fluorescence in E and F indicates Hoechst-stained nuclei. Scale bars: 20 µm.
*frizzled9b limits basal cell proliferation and directs basal protrusions to generate orthogonal tubule interconnections. (A,B) Basal new nephron invading cells are non-proliferating in wild type (A, asterisk) but show ectopic proliferation in fzd9b mutants (B, asterisk). (C,D) Basal protrusions (arrowheads) are directed orthogonally at the target distal tubule (DT) in wild type (C, arrowhead) but show a lateral displacement in fzd9b mutants (D, arrowhead). (E) Quantification of the number of EdU+ basal cells (in the bottom 6 µm) per nephron reveals a significant increase in fzd9b mutants that can be blocked by IWR1 treatment. Data are mean±s.d. (wt: 0.55±0.93, n=11; fzd9b−/−: 5.75±3.2, n=16; fzd9b−/− + IWR1: 0.38±0.52, n=8). (F) Quantification of lateral displacement of basal protrusions in fzd9b mutants. Data are mean±s.d. (wt: 0.44±0.25, n=11; fzd9b−/−: 1.06±0.67, n=11). Blue fluorescence indicates Hoechst-stained nuclei. Arrows indicate direction of fusion with the distal tubule (DT). Scale bars: 5 µm. E: ****P<0.0001; F: *P=0.0095 (ordinary one-way ANOVA).
Non-canonical Wnt signaling pathways drive planar cell polarity and modulate the cytoskeleton to regulate cell shape and orient collective cell movements during tissue morphogenesis and gastrulation (Tada et al., 2002). During interconnection, new tubules manifest several cellular features associated with non-canonical Wnt signaling, including apical constriction of new tubule epithelial cells (Fig. 1B,D,F) and oriented cell protrusions associated with cell movement that precedes lumen fusion (Fig. 1H) (Hu et al., 2021). In wild-type kidneys, basal protrusions in new tubule aggregates are distributed across lhx1a:eGFP^+^ cell aggregate basal surfaces and oriented parallel to the invading lumen (Fig. 6C) to ultimately create an orthogonal junction between two epithelial tubes (Fig. 1C). Mutation in fzd9b disrupted this organization and mispolarized cell protrusions laterally such that they were misdirected away from the target distal tubule lumen (Fig. 6B,D,F; Movie 10, compare with Fig. 1H; Movie 3). Overall protrusive activity was not significantly impaired by fzd9b mutation (Fig. S7). Consistent with lateral orientation of basal protrusions in the fzd9b mutant (Fig. 6D), new nephron lumen growth was directed laterally in the fzd9b mutant (compare Fig. 7A with B; Movie 11) instead of toward the target distal tubule lumen.
*frizzled9b-dependent apical constriction guides orthogonal interconnection of tubules and functional fluid filtration. (A) New nephron lumens are narrow and composed of apically constricted cells with small apical surfaces (arrowhead). DT, distal tubule (dotted outline). (B) fzd9b mutants show distended lumens but do not show apical constriction and they fail to make orthogonal interconnections (arrowhead). (C) Apical constriction is disrupted in wnt9b−/− mutants, phenocopying fzd9b mutants. (D) Apical constriction requires Rho kinase signaling. ROCK inhibitor Y-27632 treatment phenocopies the fzd9b mutant phenotype. (E) Quantification of apical surface area in fzd9b and wnt9b mutants: ROCK/Y-27632 and Rac1/EHT1826 inhibited new nephrons. Data are mean±s.d. (control: 0.5±0.3, n=105; fzd9b−/−: 3.32±1.45, n=37; wnt9b−/−: 3.03±1.26, n=58; Y-27632: 1.80±1.77, n=104; EHT1826: 0.56±0.32, n=38). (F,G) fzd9b mutant new nephron cells fail to undergo cell shape change from apical to basal constriction (see Fig. 1) and instead maintain a primarily rectangular shape with roughly equivalent apical (a) and basal (b) surface areas. (H,I) Compared to wild-type tubule interconnections (H, arrowhead), fzd9b mutant interconnections (I) are delayed and narrow (arrowhead in I). (J) In wild-type kidneys at 14 dpi, intraperitoneally injected 10 K fluorescent dextran (blue) marks both proliferating EdU+ (magenta) new nephrons (asterisk) as well as non-proliferating pre-existing nephrons (hashtag). (K) In fzd9b mutants, pre-existing nephrons take up injected dextran (hashtag); however, EdU+ new nephrons show reduced frequency of dextran uptake (asterisk). White fluorescence indicates Hoechst-stained nuclei. Scale bars: 5 µm. Arrows indicate the direction of fusion with the distal tubule (DT). ***P<0.0001; ns, not significant (ordinary one-way ANOVA).
Apical constriction of new tubule epithelial cells was also disrupted in the fzd9b mutant. Comparison of the apical cell surface in wild-type (Fig. 7A) and fzd9b mutant (Fig. 7B) tubules prior to interconnection revealed that the apical cell area in fzd9b mutants was increased more than sixfold (quantified in Fig. 7E) leading to cystic expansion of the new nephron lumen abutting on the distal tubule. Interestingly, mutation in wnt9b phenocopied the loss of apical constriction seen in fzd9b mutants (Fig. 7C,E), suggesting that, in addition to its role in signaling a canonical Wnt pathway to generate basal cell protrusions (see above; Fig. 3F), Wnt9b also activates a non-canonical pathway involving Frizzled9b.
In other developing tissues, non-canonical Wnt signaling drives apical constriction via Rho and Rho kinase, leading to myosin light chain phosphorylation and contraction of apical actomyosin to generate cell shape change (Martin and Goldstein, 2014). We found that treating 7 dpi fish with the ROCK inhibitor Y-27632 for 24 h phenocopied the fzd9b mutant and caused a greater than threefold increase in new nephron apical cell surface area (Fig. 7D,E). The Rac1 inhibitor EHT1826 had no effect on apical cell surface area (Fig. 7E). Invading cells of fzd9b mutants exhibited neither apical nor basal constriction (Fig. 7F,G; compare Fig. 7F to Fig. 1E,F), indicating a loss of dynamic cell shape change in the mutant. Final lumenal connections in wild-type tubules were orthogonal and patent (Fig. 7H), whereas in fzd9b mutants, connections were constricted when they did occur and lumens were circuitous, often migrating around a target distal tubule before making a connection (Fig. 7I, Movie 12).
To functionally assess tubule continuity, we performed intraperitoneal injections of 10 kDa fluorescent dextran in adult fish 12 days after gentamicin injury and identified newly formed tubules using EdU incorporation. In this protocol, only newly forming tubules are marked by EdU since EdU is injected at 12 and 14 dpi, well after older tubules have responded to injury but during the period of new nephron growth. Dextran uptake was readily apparent in older, pre-existing tubules in both wild-type and fzd9b mutant kidneys (Fig. 7J,K; non-proliferating tubules marked with a hashtag). In wild type, EdU^+^ new tubules were also positive for dextran (Fig. 7J; 10/12 tubules; 83%), indicating fluid flow and successful connections had been established. In contrast, fzd9b mutant tubules showed minimal dextran uptake (Fig. 7K; 6/17 tubules; 29%), consistent with reduced or absent lumenal perfusion due to constricted or failed tubule connections. In the kidney, tubule obstruction results in tubule swelling and distension due to blocked outflow and lumenal pressure (Kida and Sato, 2007). Consistent with failed interconnection and constricted flow, we observed large, distended tubule lumens in fzd9b mutant kidneys but not wild type (Fig. S8). Overall, these findings suggest that fzd9b enables tubule invasion and orthogonal lumen interconnection by transducing non-canonical Wnt signaling that controls both spatially restricted formation of invasive protrusions and the concerted apical to basal cell constriction associated with efficient epithelial tubule integration. Our studies also reveal a previously unreported feature of morphogenetic cell signaling, where mutually inhibitory canonical and non-canonical Wnt pathways are sequentially engaged in the same cells to drive programs of proliferation versus invasion (Fig. 8).
Model of tubule interconnection patterned by reciprocal inhibition of Wnt signaling pathways. (A-C) fzd9b expression patterns cell proliferation and cell polarization in new nephrons. (A) At early stages of new nephron formation, fzd9b (green) is uniformly expressed in non-proliferating cell aggregates. (B) Early fzd9b expression (green) establishes a persistent non-proliferating cell state (red; EdU) in basal cells destined to invade the distal tubule. fzd9b simultaneously establishes apical constriction and orients basal protrusions in new nephron cells. (C) A mutation in fzd9b relieves the block in canonical Wnt-dependent basal cell proliferation, disrupts apical constriction, resulting in an enlarged lumen, and misorients basal protrusions to the lateral edges of the new nephron basal surface. (D-F) Reciprocal inhibition of canonical and non-canonical Wnt signaling component expression. (D) As new nephron formation progresses, fzd9b expression becomes excluded from the domain closest to the distal tubule, which is a domain of high canonical Wnt target gene expression, including wnt4. (E) Mutation in wnt4 permits strong uniform expression of fzd9b in new nephrons. wnt4 mutant cells remain rounded and non-polarized. (F) Mutation in fzd9b expands the expression domain of wnt4 and other canonical Wnt target genes in new nephrons.
DISCUSSION
Mechanisms of tubule fusion in the regenerating zebrafish kidney reveal a complex interplay of Wnt signaling systems that mediate the formation of patent orthogonal tubule interconnections and ensure fluid flow in newly engrafted kidney nephrons. We found that canonical and non-canonical Wnt signaling pathways operate simultaneously and antagonistically in the same cells to orient and drive invasive cell behavior and signal cell shape changes that facilitate tubule interconnection.
Canonical Wnt signaling in tubule interconnection
We have previously shown that canonical Wnt signaling is required for new nephron cell proliferation during regeneration (Kamei et al., 2019). Here, we show that canonical Wnt signaling is also required to induce a mesenchymal invasive cell phenotype at the new nephron/distal tubule point of interconnection. This invasive phenotype was supported by expression of multiple genes typically associated with cancer metastasis. Generation of invasive protrusions was blocked by the canonical Wnt inhibitor IWR1, while mutant analysis revealed that wnt9b expression in adjacent distal tubules and autocrine wnt4 signaling in new nephrons was required for basal protrusion formation. In human pathology, Wnt4 is known to drive invasive lobular carcinoma and colorectal cancer cell proliferation; these cancers are associated with invasive cell behavior or epithelial mesenchymal transformation (Sikora et al., 2016; Yang et al., 2020). Similar to what we find in the zebrafish kidney, Wnt4 signaling in mammary carcinoma and thymic epithelial tumors is largely autocrine (Hétu-Arbour et al., 2021; Vouyovitch et al., 2016) and, in some cases, occurs independently of typical pathways for Wnt secretion (Rao et al., 2019).
Invasive cell behavior also required Src kinase activity. Src family kinase activity is observed in the context of multiple Wnt-dependent invasive tumors (Shojima et al., 2015; Villarroel et al., 2020). Src can be activated in a Wnt-dependent fashion by docking with Dishevelled 2 (Yokoyama and Malbon, 2009) and Lrp5/6 or Ror co-receptors, subsequently phosphorylating Frizzled 2 and recruiting Fyn kinase, leading to invasive cell behavior (Villarroel et al., 2020). Activated Src is also a prominent feature of invadopodia, invasive structures that are characterized by Mmp-14 and Tks5 expression (Balzer et al., 2010; Courtneidge et al., 2005), and are similar to what we see expressed in cells with protrusions in new nephrons.
Basal protrusions maintain a dense core actin organization that was Rac1-dependent. Rac1 inhibition caused protrusions to become cytoplasmic blebs that may also support cell migration (Ridley, 2015). Rac1-dependent morphogenetic processes driven by basal protrusions are often observed in migratory developmental contexts, including Drosophila germband extension (Sun et al., 2017), mammalian endothelial cell rearrangement during angiogenesis (Paatero et al., 2018), epithelial intercalation of the C. elegans embryonic epidermis (Walck-Shannon et al., 2015) and vertebrate gastrulation (Habas et al., 2003). Our results on tubule invasion in the adult zebrafish kidney present a useful model for studying invasive epithelial morphogenesis in the context of a functioning adult organ.
Despite our previous findings that canonical Wnt signaling drives epithelial tubule outgrowth (Kamei et al., 2019), invading cells adjacent to the target distal tubule, which is a source of Wnt9b, do not proliferate in the regenerating zebrafish kidney. It is therefore likely that inhibitors of canonical Wnt proliferation act in these cells. In fzd9b mutants, EdU incorporation is observed in this basal cell layer, and we conclude that fzd9b acts as one of these proliferation inhibitory signals. In support of this idea, ectopic basal cell proliferation in fzd9b mutants is blocked by the canonical Wnt inhibitor IWR1 in fzd9b mutant fish. Additional canonical Wnt signaling inhibitors, wif1 and notum1, are also expressed in the basal, invading cell layer of new nephrons. These factors are Wnt induced, negative-feedback inhibitors of canonical Wnt signaling, affecting extracellular processing of Wnt ligands (de Almeida Magalhaes et al., 2024; Tang et al., 2009). The signaling system that emerges reveals a complex interplay of regulatory inputs that promotes distinct cellular and subcellular phenotypes, and generates a restricted layer of invasive protrusions. In this context, invasion and cell proliferation may be mutually exclusive phenotypes. Cell cycle arrest linked to invasive behavior has been noted in both developmental contexts and in cancer metastasis (Evdokimova et al., 2009; Matus et al., 2015; Medwig-Kinney et al., 2023). Studies in C. elegans show that cell invasion is a differentiated cell state that requires G1 arrest and is under genetic control by the transcription factor NHR-67/TLX (nuclear hormone receptor family) and HDAC-mediated changes in gene expression (Matus et al., 2015). Our results point to a fzd9b pathway that maintains cell quiescence, possibly similar to non-canonical Wnt pathways that maintain quiescence in neural stem cells (Chavali et al., 2018), mouse muscle satellite cells (Eliazer et al., 2019) and Xenopus kidney (McCoy et al., 2011). This effect is likely to be transient during injury responses, given the changing dynamics of fzd9b and wnt4 expression we observe during interconnection. We propose that injury induction of wnt9b, and subsequently wnt4, must overcome fzd9b-induced quiescence to promote proliferation in new nephrons.
Non-canonical fzd9b signaling mediates morphogenetic transitions required for tubule interconnection
In addition to its function in limiting cell proliferation and patterning canonical Wnt target gene expression in new nephrons, fzd9b was required for apical constriction of new nephron epithelial cells at the fusion interface and for a transition to basal constriction as interconnection progressed. Somewhat paradoxically, lumen distension and failed apical constriction was phenocopied in the wnt9b mutant, implying Wnt9b may be a ligand for Frizzled9b. Both canonical and non-canonical signaling activities have been ascribed to Wnt9b in mouse kidney development, indicating that the activities of Wnt9b depend on the available combinations of Wnt receptors and co-receptors (Dickinson et al., 2019; Karner et al., 2009). Apical and basal constriction are universal developmental mechanisms in epithelial sheets undergoing shape change or invagination, as occurs in gastrulation (Keller, 1981; Popov et al., 2018), optic cup formation (Nicolas-Perez et al., 2016) and brain morphogenesis (Gutzman et al., 2018). A recent report on Xenopus neurulation highlights a role for a Wnt4/ephrinB2 non-canonical Wnt pathway that activates Rho GTPase dependent acto-myosin contraction to effect apical constriction in the neural plate (Yoon et al., 2023). Our data show that a non-canonical Wnt pathway involving fzd9b signaling can similarly act in tubular epithelia to activate Rho pathway-dependent apical constriction. Cells abutting the target distal tubule in fzd9b mutant kidneys also showed a mislocalization of basal protrusions to the periphery of the connection interface, reflecting an overall failed apical-basal polarization of new nephron cells in the fzd9b mutant. In Xenopus neurectoderm and mouse trophectoderm, the planar cell polarity component Prickle2 mediates epithelial apical-basal polarity and apical junction reorganization by a Rho GTPase dependent mechanism (Matsuda and Sokol, 2025; Tao et al., 2012). Expression of prickle1b that we observe in new nephron cells is consistent with a similar role for non-canonical Wnt signaling in establishing apical-basal polarity during tubule interconnection.
The result of mis-localized protrusions and failed cell shape change was that new nephrons did not fully penetrate the target distal tubule, and no orthogonal connections were made. Instead, the growing lumen was displaced laterally and, in some cases, projected around the target distal tubule to form a partially patent connection. Our results are consistent with studies implicating planar cell polarity mechanisms in cell rearrangement and rosette formation during tubule elongation as a crucial component of vertebrate kidney development (Lienkamp et al., 2012).
Reciprocal Wnt inhibition patterns the new nephron interconnection interface
The co-existence of antagonistic Wnt pathways acting in the same cells raises questions about the spatial and temporal dynamics of signaling during tubule interconnection. If fzd9b signaling suppresses canonical Wnt target gene expression while Wnt signaling via wnt4 represses fzd9b mRNA expression in new nephrons, how can these systems co-exist in the same cells? How can Fzd9b block basal cell proliferation if its mRNA expression is becoming repressed in these basal cells, which are the domain of high canonical Wnt signaling? Our view is that tubule interconnection is a highly dynamic process and that there must be a temporal shift in signaling as interconnection proceeds. We note that in stem cells and early aggregates, fzd9b was uniformly expressed; however’ wnt4 downregulation of fzd9b mRNA only occurred in larger, presumably further differentiated, polarizing cell aggregates. Persistence of an earlier, *fzd9b-*dependent inhibition of cell proliferation in cells no longer expressing fzd9b mRNA may be due to an extended fzd9b protein half-life or a persistent state of fzd9b signaling, after fzd9b mRNA has been reduced. Cell proliferation at a distance from the distal tubule, where fzd9b mRNA is waning but remains expressed, may be due to the dominance of a canonical Wnt signal that shifts the balance of signaling toward proliferation.
Together, these data demonstrate that canonical Wnt and non-canonical fzd9b signaling function antagonistically, patterning the new nephron into a distal layer of mesenchymal invasive cells, while preserving an epithelial phenotype in the newly forming nephron tubule. Similar systems of reciprocal signaling inhibition have been observed for Wnt signaling in patterning hair follicle development (Matos et al., 2020), for Wnt and BMP signaling in Drosophila ovary stem cell niche (Borday et al., 2012), and for Wnt and Hedgehog signaling in patterning post-embryonic retinal proliferation in the Xenopus retina (Mottier-Pavie et al., 2016). Additional Wnt ligands, receptors and signaling mediators such as wif1 and notum1, which we find are expressed in new nephrons, are likely to further refine gradients of Wnt activity and contribute to the robustness of patterning mechanisms in new nephrons.
Beyond providing insights into how epithelial cells and tubules interconnect in vivo, our data may have implications for driving fusion of stem cell-derived kidney organoid nephrons in vitro and in transplant settings (Chuang et al., 2023; Naved et al., 2022). Attempts to engraft and integrate organoid-derived nephrons in an adult mouse kidney have been challenging due to the lack of patent nephron interconnections with the host tubular architecture (Chuang et al., 2023). Our results suggest signaling systems that could be engineered at an organoid/host tubule interface to promote tubule interconnection and function of engrafted nephrons.
MATERIALS AND METHODS
Fish care, injury and drug treatments
Wild-type TuAB zebrafish were maintained according to established protocols (Westerfield, 2000). All animal experiments were performed in compliance with MDIBL IACUC animal welfare regulations. Each experiment was performed with age-matched siblings reared together to minimize background genetic variation, when possible. All adult experiments were performed with male and female fish between 6 and 18 months of age. Acute kidney injury was induced by intraperitoneal injection of gentamicin, as previously described (Kamei et al., 2015). Fish weighed between 0.25 g and 1.0 g. Gentamicin (Sigma Aldrich) diluted in PBS was injected at 80 mg/kg. Drug treatments were performed by keeping fish in either 5 μM (IWR1 and IWP2) or 10 μM (PP2 and PP3) or 100 μM (Y-27632) (Tocris) or DMSO (Sigma) dissolved in system water starting at 7 dpi until harvesting kidneys at 8 dpi, except for EHT1864, which was injected intraperitoneally at a dose of 50 mg/kg. Water changes were performed every other day by replacing half of the volume with fresh drug-treated water. For short-term proliferation studies, 20 μl of 0.5 mg/ml EdU (Invitrogen) dissolved in HBSS (Sigma Aldrich) was delivered by intraperitoneal injection 2 h before euthanization. For longer term interconnection studies, EdU was injected at 6 dpi, drug treatments were carried out at 9 dpi and kidneys were harvested at 10 dpi. For dextran filtration experiments, EdU was injected at 10 and 12 dpi, 20 µl 10 kDa rhodamine dextran (1 mg/ml in PBS) (Thermo Fisher) was injected at 12 dpi and kidneys were harvested at 14 dpi.
In situ hybridization and immunofluorescence
Whole-mount single in situ hybridization was performed as described previously (Thisse et al., 2004) with some modifications (Kamei et al., 2015). Briefly, fish were fixed with the head and internal organs removed, leaving the kidneys attached to the dorsal body wall, overnight with rocking in 4% paraformaldehyde (Electron Microscopy Sciences). After washing five times with PBST (PBS with 0.1% Tween-20), fixed kidneys were removed from the body using forceps and permeabilized with proteinase K (10 µg/ml Roche) in PBST for 1 h with rocking, postfixed in 4%PFA overnight and washed five times with PBST. Probes for mmp14a, mmp14b, tks5, cdh11, jun and notum1a were cloned from 2 dpf TuAB cDNA using the indicated primers using a hot start PCR with Phusion polymerase (NEB) and cloned into PCR II Blunt TOPO vector (Thermo Fisher). These and lhx1a, fzd9b, lef1, wnt4, wnt9b, nephrin (nphs1), id1, prickle1b, ptk7a, notum1 and wif1 (Jezewski et al., 2008; Swanhart et al., 2010) were synthesized using DIG RNA labeling mix (Sigma Aldrich). After staining, kidneys were fixed with 4% PFA, cleared with dimethylformamide, depigmented with hydrogen peroxide, transferred into PBS:glycerol (1:1) and imaged on a Leica MZ12 microscope equipped with a Spot Image digital camera. Dehydrated kidneys were embedded in JB-4 plastic resin (Polysciences) and then sectioned to a thickness of 7 µm using a LEICA 2065 rotary microtome and mounted using Permount (Fisher Scientific) with #1 coverslip (Electron Microscopy Sciences). Sections were imaged on a Nikon E800 microscope equipped with Plan Apo 60×1.4 NA oil objective with a Spot Insight CCD digital camera.
Quantification of lhx1a^+^ aggregates was performed in a blinded manner using ImageJ. Briefly, the number of aggregates in a single 5× image taken of the widest section of each kidney was divided by the total kidney area measured in ImageJ using the freehand tool to calculate aggregates per mm^2^. Pictures were relabeled before counting such that the person carrying out the analysis had no knowledge of treatment conditions for each sample.
For immunostaining, the initial fixation step was 3 h instead of overnight, but otherwise the same. Kidneys were stained for GFP [chick anti-GFP, 1:5000 (Abcam ab13970) and goat anti-chick Alexa Fluor 488, 1:3000 (Thermo Fisher A11039)]. After antibody staining, kidneys were treated for EdU detection with the Click-iT EdU kit (Invitrogen) according to manufacturer's instructions with a 1 h incubation and stained with Hoechst 33342 (1:2000, Invitrogen) overnight to detect nuclei. For double in situ hybridization and antibody staining, probe was detected using the Tyramide Super Boost kit (Invitrogen) followed by antibody staining for GFP using rabbit anti-GFP antibody (A11122, 1:500, Invitrogen) and goat anti-rabbit Alexa Fluor 488 (A11008, 1:3000, Thermo Fisher). Phalloidin labeled with Alexa Fluor 594 (A12381, 1:400, Thermo Fisher) was added to overnight Hoechst 33342 staining to visualize F-actin. Stained kidneys were mounted in PBS:glycerol (1:1) on a #1 coverslip high bridge slide using #1.5 coverslips glued down with Krazy glue (WB Mason) for imaging.
Laminin (L9393, 1:30, Sigma), goat anti-rabbit Alexa Fluor 546 (A11010, 1:3000, Thermo Fisher) and GFP staining were performed on 20 μm cryosections. Briefly, kidneys were infiltrated with 30% sucrose/PBS overnight, embedded in OCT medium (TissueTek) and sectioned on a Leica CM1860UV cryostat, stained in a humidified chamber then mounted with Vectashield medium (Vectashield) under a #1.5 coverslip for imaging.
Confocal imaging was performed on a Zeiss LSM980 point scanning confocal microscope with Airyscan 2 (Carl Zeiss Microscopy) equipped with a Zeiss Axio Examiner Z1 upright microscope stand (409000-9752-000) and a Plan/Apochromat 10×/0.45 (420640-9900-000) or Plan/Apochromat 63×/1.4 oil (420782-9900-799) objective controlled with Zen Blue software (Zen Pro 3.1). Z stacks taken with the 10× objective were acquired using Nyquist criterion mode. Z-stack images at 63× were collected with the Super Resolution mode (SR) at either 1.7× or 3.4× zoom, with optimal interval using the motorized scanning stage 130×85 PIEZO mounted on the Z-piezo stage insert WSB500 (Carl Zeiss Microscopy). Airyscan images were processed using the ‘auto’ mode and saved in CZI format. 3D reconstructions of z stacks were created using Imaris 9.6 Software.
Quantification of dextran filtration was carried out by two separate observers using both Imaris 3D reconstructions and stacks of slices in ImageJ, and results were reconciled. Newly regenerated nephrons were identified by searching stacks for EdU^+^ tubules (identified by epithelial nuclear morphology and EdU^+^ nuclei) with the rhodamine dextran channel turned off, then counting the number of tubules that had dextran-containing vesicles on the apical/luminal side of nuclei.
Cell membrane curvature analysis
Basal protrusions on new nephrons were quantified by measuring average basal surface curvature of lhx1a:egfp+ new nephrons. Zeiss .czi files were converted to Imaris format (.ims) and free rotated to be orthogonal to the viewing axis and exported as an image sequence. The ImageJ Kappa plugin (Mary and Brouhard, 2019 preprint) was used to analyze individual slices every 2 µm of stack depth using tracings of new nephron bottom surfaces outlined by lhx1a:egfp fluorescence. Final ImageJ kappa measurements were exported to Excel for averaging and to Prism for statistical analysis and plotting.
Basal protrusion position analysis
Zeiss .czi confocal stack files were converted to Imaris format (.ims) and free rotated to present an en face view of new nephron tubule basal surfaces. A region of interest was drawn, encompassing the entire basal surface then scaled to 90%. The ImageJ process ‘find maxima’ (prominence 15.0) was used to generate a ratio of lhx1a:egfp fluorescence intensity maxima inside (central) versus outside (peripheral) the ROI and quantified as a proxy for basal protrusion position on new nephron basal surfaces. Results were analyzed by unpaired t-test (wild type versus fzd9b mutant) and plotted using Prism software.
Quantification of basal cell proliferation
Zeiss .czi confocal stacks of EdU incorporation in new nephrons were converted to Imaris format (.ims) and free rotated to present an orthogonal view of lhx1a:eGFP^+^ new nephrons. Image stacks were then 3D cropped to include only the bottom 6 μm of each new tubule. EdU-positive nuclei were counted, statistically analyzed by one-way ANOVA (control versus experimental) and plotted using Prism software.
Quantification of canonical Wnt target expression domains
JB-4 glycolmethacrylate sections (7 µm thick) of whole-mount in situ hybridization experiments with Wnt target gene probes were photographed with DIC optics on a Nikon Eclipse microscope using a Nikon Plan APO Lambda D 60× oil immersion lens. Expression domains were quantified by circling reaction product regions of interest and quantified in ImageJ/Fiji after setting the magnification scale.
Apical surface area analysis
Apical surface area was measured in individual confocal slices of phalloidin-stained new nephron confocal stacks. Phalloidin^+^ lumenal cell apical perimeters were traced and measured in ImageJ using the area measurement tool. Apical surface areas were statistically analyzed by one-way ANOVA (control versus experimental) and plotted using Prism software.
Supplementary Material
10.1242/develop.205074_sup1Supplementary information
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Balzer, E. M., Whipple, R. A., Thompson, K., Boggs, A. E., Slovic, J., Cho, E. H., Matrone, M. A., Yoneda, T., Mueller, S. C. and Martin, S. S. (2010). c-Src differentially regulates the functions of microtentacles and invadopodia. Oncogene 29, 6402-6408. 10.1038/onc.2010.36020956943 PMC 4073667 · doi ↗ · pubmed ↗
- 2Barbayianni, I., Kanellopoulou, P., Fanidis, D., Nastos, D., Ntouskou, E. D., Galaris, A., Harokopos, V., Hatzis, P., Tsitoura, E., Homer, R. et al. (2023). SRC and TKS 5 mediated podosome formation in fibroblasts promotes extracellular matrix invasion and pulmonary fibrosis. Nat. Commun. 14, 5882. 10.1038/s 41467-023-41614-x 37735172 PMC 10514346 · doi ↗ · pubmed ↗
- 3Bernstein, C. S., Anderson, M. T., Gohel, C., Slater, K., Gross, J. M. and Agarwala, S. (2018). The cellular bases of choroid fissure formation and closure. Dev. Biol. 440, 137-151. 10.1016/j.ydbio.2018.05.01029803644 PMC 7177177 · doi ↗ · pubmed ↗
- 4Borday, C., Cabochette, P., Parain, K., Mazurier, N., Janssens, S., Tran, H. T., Sekkali, B., Bronchain, O., Vleminckx, K., Locker, M. et al. (2012). Antagonistic cross-regulation between Wnt and Hedgehog signalling pathways controls post-embryonic retinal proliferation. Development 139, 3499-3509. 10.1242/dev.07958222899850 · doi ↗ · pubmed ↗
- 5Carroll, T. J., Park, J. S., Hayashi, S., Majumdar, A. and Mc Mahon, A. P. (2005). Wnt 9b plays a central role in the regulation of mesenchymal to epithelial transitions underlying organogenesis of the mammalian urogenital system. Dev. Cell 9, 283-292. 10.1016/j.devcel.2005.05.01616054034 · doi ↗ · pubmed ↗
- 6Chavali, M., Klingener, M., Kokkosis, A. G., Garkun, Y., Felong, S., Maffei, A. and Aguirre, A. (2018). Non-canonical Wnt signaling regulates neural stem cell quiescence during homeostasis and after demyelination. Nat. Commun. 9, 36. 10.1038/s 41467-017-02440-029296000 PMC 5750230 · doi ↗ · pubmed ↗
- 7Chuang, T., Bejar, J., Yue, Z., Slavinsky, M., Marciano, D., Drummond, I. and Oxburgh, L. (2023). In Vivo assessment of laboratory-grown kidney tissue grafts. Bioengineering (Basel) 10, 1261. 10.3390/bioengineering 1011126138002385 PMC 10669198 · doi ↗ · pubmed ↗
- 8Cote, L. E. and Feldman, J. L. (2022). Won't you be my neighbor: how epithelial cells connect together to build global tissue polarity. Front. Cell Dev. Biol. 10, 887107. 10.3389/fcell.2022.88710735800889 PMC 9253303 · doi ↗ · pubmed ↗