Notch signaling is a novel regulator of visceral smooth muscle cell differentiation in the murine ureter

Smooth muscle cells (SMCs) are found in the mesenchymal wall of many visceral tubular organs but also as an ensheathment of endothelial cells in the vascular system. Owing to their contractile activity, they play a decisive role in maintaining the flexibility and rigidity of these tubes, and in mediating the unidirectional transport of their luminal content. SMCs arise from a diverse range of progenitors and show a high phenotypic plasticity, yet their specialized contractile phenotype seems universally transcriptionally controlled by a complex of the DNA-binding protein serum response factor (SRF) and the co-activator myocardin (MYOCD) (Norman et al., 1988; Wang and Olson, 2004; Yoshida et al., 2003). Expression of Myocd and of SMC structural genes occurs in SMC progenitors as a response to a multitude of extrinsic and intrinsic signals. The nature of these signals seems fundamentally different in vascular and visceral SMC progenitors, probably owing to their specific association with endothelial and epithelial primordia, respectively (Creemers et al., 2006; Donadon and Santoro, 2021; Mack, 2011; Shi and Chen, 2016).

Owing to its simple design, its pharmacological and genetic accessibility, and its relevance for congenital anomalies in humans, the murine ureter is an attractive model with which to unravel the regulatory network that drives visceral SMC differentiation during organogenesis (Bohnenpoll and Kispert, 2014; Woolf and Davies, 2013; Woolf et al., 2019). Previous work has shown that visceral SMCs of the mouse ureter arise at embryonic day (E)11.0 from a Tbx18⁺ mesenchymal progenitor pool that surrounds the distal aspect of the ureteric bud: an epithelial diverticulum of the nephric duct (Bohnenpoll et al., 2013). Until E14.5, two signals from the ureteric epithelium (UE), SHH and WNTs, act on the undifferentiated ureteric mesenchyme (UM) to maintain its proliferative expansion and trigger SMC differentiation. SHH activates the expression of the transcription factor gene Foxf1 in the UM, which, in turn, induces and synergizes with the signaling molecule BMP4 in activation of Myocd and SMC structural genes (Bohnenpoll et al., 2017c; Mamo et al., 2017; Yu et al., 2002). WNTs act, at least partly, through the transcription factors TBX2 and TBX3 to maintain BMP4 signaling and suppress an outer adventitial fate (Aydogdu et al., 2018; Trowe et al., 2012). Retinoic acid (RA) synthesized in both the UM and UE inhibits SMC differentiation possibly by counteracting WNT signaling (Bohnenpoll et al., 2017b). As a consequence of a poorly understood interplay of these and most likely additional signals, Myocd is precisely activated in the inner layer of the proximal UM at E14.5, expression of SMC structural genes starts at E15.5, and a peristaltically active SMC layer is established concomitantly with the onset of urine production in the kidney around E16.5 (Bohnenpoll et al., 2017a).

Notch signaling is an evolutionarily conserved pathway that mediates contact-dependent cell-to-cell communication in a variety of developmental contexts. In mammals, four Notch receptors (NOTCH1-4) and five ligands [jagged 1 (JAG1), jagged 2 (JAG2), delta-like 1 (DLL1), delta-like 3 (DLL3) and delta-like 4 (DLL4)] are described, which are all type I transmembrane proteins. Ligand-receptor interaction triggers proteolytic cleavages that release the intracellular domain of the receptor (NICD) from the membrane. NICD translocates to the nucleus where it forms an active transcriptional complex with the transcription factor RBPJ and several co-activators (Henrique and Schweisguth, 2019; Kopan, 2012; Kovall et al., 2017). Notch signaling has been characterized as a crucial pathway for vascular SMC differentiation (Baeten and Lilly, 2017; Fouillade et al., 2012), whereas its potential role in visceral SMC development has remained unexplored.

Here, we set out to analyze a possible role of Notch signaling in visceral SMC differentiation in the murine ureter. We show that the pathway is essential for the timely activation of expression of Myocd and of a group of late SMC structural genes, and, hence, for achieving and maintaining proper peristaltic activity in this organ.

To determine the abundance of Notch signaling components in ureter development, we analyzed expression of genes encoding Notch ligands and receptors by RNA in situ hybridization on transverse sections of the proximal ureter region of E12.5 to E18.5 wild-type embryos (Fig. S1). Jag1 was homogenously expressed in the UE and the UM, with lower levels at E16.5 and E18.5. Jag2 was robustly expressed in the UE at E14.5. At E16.5 and E18.5, expression in this tissue was predominantly found in the basal cell layer. Expression was also found in endothelial cells of vessels in the outer UM from E12.5 to E18.5. Dll1 and Dll3 expression was not detected in ureter development. Dll4 expression was found in endothelial cells of larger vessels at all stages (Fig. S1A). Notch1 was weakly expressed in the UE from E12.5 to E16.5, and in some basal cells at E18.5. Expression also occurred in endothelia in the outer UM from E12.5 to E18.5. Notch2 and Notch3 were strongly expressed in the UM at E12.5 and E14.5, and more weakly at E16.5 and E18.5. Notch3 expression was additionally found in the UE, and more strongly in perivascular cells in the outer UM from E12.5 to E18.5. Notch4 expression was associated with endothelia in the outer UM throughout ureter development (Fig. S1B). Expression of Rbpj encoding the unique intracellular mediator of this signaling pathway (Jarriault et al., 1995) occurred homogenously at low level at all stages both in the UM and UE (Fig. S1C).

We next used immunohistochemistry to analyze protein expression of those pathway components for which we had detected mRNA expression in the UM and/or the UE. As we did not find a suitable antibody for JAG2, we omitted it from this analysis. JAG1 and NOTCH1 showed low level expression both in the UM and UE at 12.5 and E14.5. Expression in the UE was increased at E18.5. NOTCH1 was additionally found in endothelial linings of vessels. NOTCH2 was detected in the UM and UE at all analyzed stages. NOTCH3 and RBPJ expression was weak in the UM and UE at E12.5 and E14.5. Expression was strongly increased in the UM and decreased in the UE at E16.5 and E18.5. NOTCH3 was also strongly expressed in vessel walls, including that of the dorsal aorta, at all stages (Fig. 1A).

To test for direct interaction of JAG1 with any of the NOTCH receptors, we performed a proximity ligation assay. We found proximity signals for JAG1 with all three NOTCH receptors in the UM and UE at all stages except E16.5. The JAG1-NOTCH3 interaction appeared strongest at all stages but was still markedly lower than that observed in the dorsal aorta (Fig. 1B).

To test for Notch pathway activity, we analyzed expression of direct transcriptional targets of RBPJ, namely Hes and Hey genes, and Nrarp in ureter development. Hes1 was strongly expressed in the UE at E12.5, E14.5 and E18.5. Hes2 and Hes5 expression was found in both the UE and UM at these stages. Hes6 and Hes7 were weakly found in the UM at E12.5 (Fig. S2A). Weak expression of Hey1 was detected in the UM at E14.5, of Hey2 in the UM at E12.5 and of Heyl in the UM at E12.5, E14.5 and E18.5. Higher levels of Hey1 and Hey2 expression were found in endothelial cells of larger vessels, and of Heyl in the surrounding vascular SMCs (Fig. S2B). Nrarp expression was widespread, but increased in the UM at E14.5 and E18.5. Expression in the UE was strong at E16.5 and E18.5, and in vessels at all stages (Fig. 1C). This dataset points to a biphasic wave of Notch signaling activity in the UM, peaking at E14.5 and E18.5. Signaling activity is reduced compared with that in the mesenchymal wall of vessels but may similarly be promoted by JAG1 interaction with NOTCH2 and NOTCH3.

To investigate the role of canonical Notch signaling in the UM, we employed a tissue-specific gene inactivation approach using a Tbx18^cre line generated in our laboratory (Airik et al., 2010) and a floxed allele of Rbpj (synonym Rbpj^fl) (Tanigaki et al., 2002), the unique intracellular mediator of this signaling pathway (Jarriault et al., 1995). Tbx18^cre mediates recombination in precursors of all differentiated cell types of the UM: fibroblasts of the inner lamina propria and the outer tunica adventitia, SMCs of the medial tunica muscularis (Bohnenpoll et al., 2013) and vascular SMCs but not endothelial cells (Fig. S3). Absence of RBPJ expression in the UM of Tbx18^cre/+;Rbpj^fl/fl (Rbpj-cKO) embryos confirmed the suitability of our approach (Fig. S4).

We started our phenotypic analysis at the end of embryogenesis, at E18.5, when all differentiated cell types of the ureter are established. At this stage, the urogenital system of Rbpj-cKO embryos was morphologically unaffected with the exception of the ureter, which appeared more translucent than in the control (Fig. 2A). The kidney was histologically normal but the tunica muscularis of the ureter appeared less condensed (Fig. S5, Fig. 2B). Expression of the SMC proteins ACTA2, TAGLN and NOTCH3 was unchanged in the tunica muscularis of the mutant ureter but was reduced in large adventitial vessels (Fig. 2C). Expression of the SMC structural genes Cnn1 and Myh11 appeared unaffected, whereas Tagln was weakly reduced and Tnnt2 was strongly reduced in the ureteric muscle layer. The lamina propria marker Aldh1a2 and the adventitial marker Dpt were unchanged (Fig. 2D). The distribution of endomucin (EMCN) (Morgan et al., 1999) and of KRT5, ΔNP63 and UPK1B (Bohnenpoll et al., 2017a) reflected normal vascular endowment and urothelial differentiation, respectively (Fig. S6A).

To profile transcriptional changes in E18.5 Rbpj-cKO ureters in a global and unbiased fashion, we used microarray analysis. Using a threshold of at least a 1.5-fold change and an expression intensity robustly above background (>100), we detected 93 genes with reduced expression and 45 with increased expression in Rbpj-cKO ureters (Table S1A,B; deposited in GEO under accession number GSE169662). Functional annotation using the DAVID software tool (Huang da et al., 2009) revealed a highly significant enrichment of gene ontology (GO) terms and clusters related to ‘muscle contraction’ for the pool of downregulated genes, whereas variable terms and clusters with low significance were found for the pool of upregulated genes (Fig. 2E, Tables S2, S3). Manual inspection of the list of downregulated genes detected Rbpj (−2.4) and the Notch effector gene Heyl (−2.9), confirming the loss of Notch signaling activity. Tnnt2 expression was strongly reduced (−2.1), Tagln (−1.1), Cnn1 (−1.3) and Myh11 (−1.3) were weakly reduced, and Acta2 was unchanged, largely confirming our in situ hybridization analysis (Fig. 2F, Table S1A).

We validated expression of a subset of the downregulated genes by in situ hybridization analysis. We found strongly reduced expression of Pcp4, Ckm, Myl4, Pcp4l1, Mfap4, Rhoa and Synpo2 in the muscle layer of the mutant ureter. Tpm2 appeared weakly affected; other candidates were not detected by this method (Fig. 2G, Fig. S7). Reverse transcriptase-quantitative PCR (RT-qPCR) confirmed slightly reduced (Tagln and Tpm2) and strongly reduced (Ckm, Pcp4, Pcp4l1 and Tnnt2) expression of SMC genes at this stage (Fig. 2H, Table S4A). We conclude that Rbpj-cKO ureters exhibit defects in visceral SMC differentiation shortly before birth.

To investigate whether expression of SMC genes is merely delayed and normalizes after birth, we analyzed Rbpj-cKO ureters at postnatal day (P) 4. At this stage, compartmentalization of the ureter was histologically unaffected but the muscle layer appeared less condensed (Fig. 3A). Immunofluorescence analysis detected normal expression of the SMC proteins ACTA2, TAGLN and NOTCH3 (Fig. 3B), and of the epithelial markers KRT5, ΔNP63 and UPK1B (Fig. S6B). In situ hybridization and/or RT-qPCR analysis uncovered that transcripts of SMC genes were differentially affected in their expression: Cnn1, Myh11, Tagln and Tpm2 appeared unaffected; expression of Ckm, Pcp4, Pcp4l1, Myl4 and Tnnt2 was strongly reduced (Fig. 3C,D; Table S4B). Transcriptional profiling by microarray analysis detected 141 transcripts with reduced expression and 88 with increased expression (>1.5-fold change, expression intensity >100) in P4 Rbpj-cKO ureters (Table S5; deposited in GEO under accession number GSE184597). Functional annotation revealed a significant enrichment of GO terms related to both ‘lipid and glucose metabolism’ and ‘muscle’ for the pool of downregulated genes, whereas GO terms related to ‘extracellular matrix’ were enriched in the pool of upregulated genes (Tables S6 and S7), indicating possible metabolic and structural compensatory mechanisms. Notably, 25 of the genes with decreased expression at P4 also showed decreased expression at E18.5, yet with fold changes that were reduced compared with that at E18.5 (Fig. 3E). Functional annotations showed an enrichment of terms related to ‘muscle’, ‘Z-disc’, ‘sarcolemma’ and ‘heart muscle’, indicating that common downregulated genes mainly relate to SMC differentiation (Table S8). Expression changes of additional SMC genes fell below the threshold at P4 (Fig. 3E).

To determine whether the SMC defects further decrease with time, we analyzed urogenital systems at P14. At this stage, the mutant ureter was invariably dilated at the proximal level. Some SMC genes seemed unchanged (Cnn1, Myh11 and Tpm2), others were strongly reduced (Ckm, Pcp4, Pcp4l1, Tagln and Tnnt2) in their expression (Fig. 3F). Together, this shows that SMC differentiation defects, although decreasing after birth, cannot be functionally compensated for and lead to hydroureter formation in early adolescence.

To define the onset of SMC defects in Rbpj-cKO ureters, we performed histological and molecular analyses at stages (E14.5 to E16.5) when the SMC phenotype is progressively established. Histological analysis showed that the UM of the mutant was subdivided into an inner layer with rhomboid-shaped condensed cells and an outer layer with loosely organized fibroblast-like cells at all analyzed stages, as in the control, but the inner layer appeared less condensed at E15.5 and E16.5 (Fig. 4A). In the control, expression of ACTA2, TAGLN, Cnn1, Myh11, Tagln and Tpm2 commenced at E15.5; that of NOTCH3 and Tnnt2 began at E16.5 in the inner layer of the UM. In Rbpj-cKO ureters, expression of Cnn1, Myh11 and Tpm2 occurred normally from E15.5 onwards. ACTA2, TAGLN and Tagln expression was reduced at E15.5 and at E16.5; NOTCH3 and Tnnt2 expression was not observed in the mutants at E16.5 (Fig. 4B,C). Ckm, Myl4, Pcp4 and Pcp4l1 mRNA expression was neither detected in the control nor in the mutant ureter in the analyzed time window (Fig. S8). We conclude that loss of Rbpj in the UM affects the staggered activation of SMC genes in the fetal ureter.

We next investigated whether the observed changes in visceral SMC differentiation are accompanied by functional deficits in ureter contractility in explant cultures. Explants of E14.5 Rbpj-cKO ureters exhibited a 1.5-day delay in onset of peristaltic activity (Fig. 5A,B, Table S9A). The contraction frequency was significantly decreased until day 6 and reached control levels only at day 7 and 8 of the culture (Fig. 5C, Table S9B). The contraction intensity was strongly reduced at all analyzed levels throughout the entire contraction wave at day 4 of the culture. At the endpoint, day 8, the initial contraction velocity in the proximal and the medial part was normal but the contraction intensities remained lower throughout the contraction wave (Fig. 5D, Table S9C).

Mutant ureters explanted at E18.5 exhibited a significantly reduced contraction frequency at day 1 and 2 of culture but reached the level of the control from day 3 onwards (Fig. 5E,F, Table S10A). At day 1, the contraction occurred less rapidly and reached lower intensities; the relaxation wave was, however, unaffected. At day 3, the mutant ureters reached the contraction intensity of the control, albeit with a slight but significant delay. At day 6, the mutant ureters reached higher contraction intensities and maintained them for longer. This was most prominent at the medial position (Fig. 5G, Table S10B).

Hence, loss of Rbpj affects onset and progression of the peristaltic activity of the fetal and perinatal ureter. At perinatal stages, ureter contractility appears weakly affected, indicating that, in the absence of hydrostatic pressure, deficits in the SMC program can be functionally compensated for.

To identify molecular changes that may cause delayed and reduced SMC differentiation in Rbpj-cKO ureters in an unbiased fashion, we performed microarray-based gene expression profiling of E14.5 ureters. Using an intensity threshold of 100 and fold changes of at least 1.5 in the two individual arrays, we detected 30 genes with increased and 16 with decreased expression in mutant ureters (Fig. 6A; Table S11A,B; deposited in GEO under accession number GSE169661).

Functional annotation revealed an enrichment of GO terms related to the differentiation of secretory cells, dopaminergic neurons and/or chromaffin cells in the pool of upregulated genes, but in situ hybridization did not detect expression of any of the selected candidates in control and mutant ureters (Table S12A, Fig. S9). In the pool of downregulated genes, GO terms related to protein binding and negative regulation of WNT signaling (Mdfi, Shisa2 and Wif1) were found (Table S12B). Manual inspection of the list identified Rbpj (−1.9), confirming the functionality of our genetic approach, and Myocd (−2.0), the key regulator of SMC differentiation (Fig. 6A). In situ hybridization detected reduced expression of Mdfi, Car3, Shisa2 and Myocd in the UM of mutant embryos (Fig. 6B). Other candidates showed unspecific or no expression in control and mutant ureters (Fig. S10).

In agreement with our microarray data, we did not detect expression changes for genes encoding cellular signals, signaling targets and transcription factors that have previously been implicated in Myocd activation and SMC differentiation in the ureter by in situ hybridization analysis (Fig. S11A,B). These findings validate that reduced expression of the WNT antagonist Shisa2 does not translate into changes in WNT signaling, and that known regulators of Myocd expression are unchanged in E14.5 Rbpj-cKO ureters.

To determine whether Myocd expression is delayed in Rbpj-cKO ureters, we analyzed its expression at subsequent stages. In situ hybridization detected normal expression at E15.5, E16.5, E18.5 and P4 (Fig. 6C). RT-qPCR analysis confirmed strongly reduced expression of Myocd at E14.5, whereas expression of Foxf1, an activator of Myocd expression, was unchanged (Fig. 6D, Table S4C). Expression of Myocd was unchanged in this assay at E18.5 and P4, supporting the in situ hybridization results (Fig. 6E, Table S4D). We conclude that Rbpj-dependent Notch signaling is required for precise activation of Myocd at E14.5 but not for its further maintenance at fetal and postnatal stages.

To exclude the possibility that RBPJ acts independently of Notch receptors in the context of the UM, and to distinguish early from late requirements of this pathway, we performed time-controlled pharmacological Notch pathway interference experiments with the γ-secretase inhibitor DAPT (Cheng et al., 2003) in ureter explant cultures. Administration of 1 µM and 2.5 µM DAPT (Cheng et al., 2003) to E12.5 ureter explants led to a dose-dependent delay in the onset of the peristaltic activity and a reduction of contraction frequency, similar to the situation observed in explants of E14.5 Rbpj-cKO ureters (Fig. 7A,B, Table S13).

We next explanted wild-type ureters at E18.5 and treated them with 1 µM of DAPT. These ureters showed a normal peristaltic onset but a reduced contraction frequency until day 3 of culture, again similar to Rbpj-cKO ureters (Fig. 7C, Table S14). After 18 h in culture, expression of Ckm and Tnnt2 was significantly reduced; Pcp4l1 and Tpm2 showed a strong trend towards reduction. Expression of Myocd, Pcp4 and Tagln appeared unaffected (Fig. 7D, Table S4E).

Wild-type ureter explanted at P4 and treated with 1 µM of DAPT exhibited normal peristaltic contractions in a 6-day culture period (Fig. 7E, Table S15). After 18 h in culture, expression of Ckm, Pcp4, Tagln and Tnnt2 was reduced, while Myocd, Pcp4l1 and Tpm2 appeared unchanged (Fig. 7F, Table S4F).

Hence, loss of Notch signaling affects onset and progression of the peristaltic activity of the fetal and perinatal ureter. Reduced expression of SMC genes at perinatal and postnatal stages is independent of Myocd expression.

We finally asked whether Notch signaling is sufficient to induce SMC relevant genes in ureter development. For this, we combined our Tbx18^cre driver line with a Rosa26 knock-in allele (Rosa26^NICD) (Murtaugh et al., 2003), allowing conditional expression of the Notch1 intracellular domain (N1ICD) in the undifferentiated UM. As Tbx18^cre/+;Rosa26^NICD/+ embryos died around E13.5 (Grieskamp et al., 2011), we used E12.5 ureters for section in situ hybridization analysis. We did not find ectopic and/or precocious expression of the SMC regulators Foxf1 and Myocd, of SMC structural genes, or of Car3 and Shisa2, indicating that Notch signaling is required but not sufficient to activate SMC regulatory and structural genes in the developing ureter (Fig. S12). To analyze the effect of N1ICD overexpression in the UM at late fetal stages, we combined a tamoxifen-inducible variant of Tbx18 (Tbx18^creERT2) with the Rosa26^NICD allele, explanted mutant ureters at E13.5 and cultured them for 4 days in the presence of tamoxifen. Enhanced expression of Heyl showed the suitability of the approach. Similar to induced Heyl expression, SMC markers appeared more patchy, but levels were decreased rather than increased in some cases (Fig. S13) indicating that a (low) level of Notch signaling in the UM is important for ureteric SMC integrity.

Previous genetic work provided compelling evidence that Notch signaling is a crucial regulator of vascular SMC differentiation (for reviews, see Baeten and Lilly, 2017; Fouillade et al., 2012). This applies both to neural crest cells, from which SMCs of the great vessels, including the aorta, are derived (Feng et al., 2010; High et al., 2007; Manderfield et al., 2012), as well as to mesothelial cells and other progenitors of arterial SMCs in different organ systems (Etchevers et al., 2001; Grieskamp et al., 2011; Volz et al., 2015). In either case, loss of Notch signaling (components) was associated with severely reduced expression of SMC structural genes, including early differentiation markers ACTA2 and TAGLN, and with subsequent vessel dilatation.

To unravel the function of Notch signaling in the development of the UM, we used a combination of genetic and pharmacological pathway inhibition experiments. Loss of the Notch signaling mediator Rbpj affected neither ureter shape and length nor the subdivision of its mesenchymal wall at fetal and postnatal stages, excluding a role of the pathway in survival, proliferation and patterning of the UM. At newborn stages, a set of important SMC proteins/genes (ACTA2, TAGLN, Myh11 and Cnn1) was correctly expressed in the medial region of the UM, indicating that visceral SMC specification and early differentiation of SMCs has occurred normally. However, we observed a delayed expression onset of Myocd and of ‘early’ SMC genes around E14.5 and E16.5, as well as a delayed onset and reduced expression of a set of ‘late’ SMC genes in Rbpj-cKO ureters at fetal and postnatal stages. At the physiological level, these changes translated into a delayed onset of peristaltic activity, reduced contraction frequency and intensity at fetal stages. Time-controlled pharmacological Notch pathway inhibition experiments largely recapitulated these phenotypic changes, confirming a role for Notch signaling and RBPJ in timing, modifying and/or fine-tuning visceral SMC differentiation in the ureter.

Rbpj deficiency or Notch pathway inhibition did not affect peristaltic activity of P4 ureters ex vivo, suggesting that the SMC defects at this stage are minor. However, and to our surprise, Rbpj-cKO ureters exhibited ureter dilatation (hydroureter) at P14, indicating that the mutant SMC layer has reduced capacity to withstand the hydrostatic pressure of the urine with time. The phenotypic burden of Rbpj-cKO mice prevented analysis at later stages in adults. However, it is likely that, under the permanent pressure exerted by the urine, even a weak reduction of SMC structural proteins will cause further deficits of SMC contractility and rigidity that will translate in progressive ureter dilatation, hydronephrosis and end-stage renal disease. Mutations that affect expression of Notch components may therefore underlie human congenital anomalies of the kidney and ureteric tract (CAKUT), a group of diseases for which the genetic cause has only partly been resolved (Kohl et al., 2021).

Although not analyzed in any detail, we noted reduced ACTA2 and TAGLN expression in cells surrounding endothelial linings in the adventitial layer of Rbpj-cKO ureters, indicating that Notch signaling is essential for vascular SMC differentiation in the ureter as in many, if not all, other organs.

Work in the vascular system characterized JAG1, NOTCH2 and NOTCH3 as the major Notch components involved in SMC differentiation (for a review, see Baeten and Lilly, 2017). Endothelial JAG1 serves as the initial cue to activate Notch signaling in surrounding perivascular cells (High et al., 2008). This leads to increased expression of NOTCH3, as well as JAG1, in these cells, which in turn promotes, by homotypic interaction, their differentiation into SMCs (Hoglund and Majesky, 2012; Liu et al., 2009). Whereas NOTCH3 is expressed in all mural cells, NOTCH2 is mainly found in large vessels, such as the cardiac outflow tract (High et al., 2007; Joutel et al., 2000). Accordingly, Notch2 and Notch3 are redundantly required for vascular SMC differentiation in larger vessels, whereas in pericytes and small vessels NOTCH3 is the dominant player (Liu et al., 2010; Volz et al., 2015; Wang et al., 2012). Our expression analysis confirmed the relevance of this JAG1-NOTCH2/NOTCH3 network in vascular SMC development in the dorsal aorta, which we used as ‘control’ tissue. JAG1 was found in the endothelium and surrounding perivascular cells, where co-expression occurred with NOTCH2 and NOTCH3. Strong JAG1-NOTCH2 and particularly JAG1-NOTCH3 interaction in these perivascular cells was validated by the proximity ligation assay.

A similar JAG1-NOTCH2/NOTCH3 module may operate in the UM during development. Our immunohistochemical analysis revealed low level expression of JAG1 and high level expression of NOTCH2 in the UM at all developmental stages. NOTCH3 was strongly upregulated at E16.5 in the UM, reaching levels of expression similar to that seen in the perivascular cells of the dorsal aorta. RBPJ also showed an upregulation in the UM, suggesting that translation of both mRNAs is similarly controlled. Our proximity ligation assay uncovered interaction of JAG1 with NOTCH2 and NOTCH3 from E12.5 to E14.5, and at E18.5, and increased expression of the direct NOTCH target genes Heyl and Nrarp in the UM at these stages. Although the overall activation of Notch signaling in the UM is clearly much weaker compared with that in perivascular cells of the dorsal aorta, our findings point to a biphasic activation of NOTCH2 and/or NOTCH3 by JAG1 in the UM: first, from E12.5 and E14.5, and then concomitant with the upregulation of NOTCH3 after E16.5. It is noteworthy that, unlike JAG1 and NOTCH2, expression of NOTCH3 is largely cytosolic, raising the possibility that JAG1-NOTCH3 interaction occurs cell-autonomously with the UM. This would be reminiscent of the situation found in pulmonary artery vascular SMCs in the developing lung. There, NOTCH3 is expressed and activated in late fetal to early postnatal life, dependent on SMC-derived JAG1 (Ghosh et al., 2011). Although the presence of a JAG1-NOTCH2/NOTCH3 signaling module in the UM seems highly plausible, we cannot exclude the possibility that epithelial Notch ligands (JAG1 or the strongly expressed JAG2) contribute to activation of Notch receptors in the UM. We deem it unlikely as the presence of an intervening basement membrane and, later, also that of the lamina propria layer should prevent the direct contact between the ligand- and receptor-bearing cells required for Notch pathway activation. Conditional gene targeting experiments for individual Notch components may provide conclusive data in the future.

Our molecular analysis found that a group of ‘early’ SMC structural genes, including ACTA2 and TAGLN are activated with a delay of 1-2 days in Rbpj-cKO ureters but reached normal levels of expression at E18.5. Compatible with this expression pattern, we observed a delayed activation of the regulator of the SMC differentiation, Myocd, at E15.5 in Rbpj-cKO ureters. Importantly, we did not detect changes in the activity of signaling pathways (SHH, BMP4, WNT and RA) and transcription factors (Foxf1, Tshz3 and Sox9) that have been implicated in the regulation of Myocd at E14.5 (Airik et al., 2010; Bohnenpoll et al., 2017b,c; Caubit et al., 2008; Mamo et al., 2017; Trowe et al., 2012). Hence, Myocd may be a direct target of RBPJ or of HES/HEY bHLH proteins that mediate the activity of this pathway in many contexts (Bray and Bernard, 2010; Fischer et al., 2004). Irrespective of the precise mode of action, we posit that Notch signaling provides an important input for precise temporal activation of Myocd transcription in the fetal ureter.

Our expression analyses uncovered that a group of SMC structural genes that are normally activated between E17.5 and E18.5 in the UM (including Ckm, Pcp4, Pcp4l1 and Tnnt2) were not present at E18.5 and showed reduced expression at P4 in Rbpj-cKO ureters. Pharmacological Notch signaling inhibition of E18.5 ureters resulted in similar changes. Given unchanged Myocd expression in mutant ureters from E15.5 onwards, we suggest that MYOCD/SRF is not sufficient to activate expression of these ‘late’ SMC genes but that Notch signaling provides a crucial second input for their timely activation. Importantly, misexpression of NICD neither prematurely activated nor enhanced expression of any of the SMC genes tested. In fact, enforced Notch signaling at late fetal stages led to a decreased and patchy SMC gene expression. This confirms that Notch is a modulator and not a driver of the visceral SMC program, and that the level of signaling at late fetal stages is tightly controlled to assure integrity of the tunica muscularis.

‘Late’ SMC genes affected in Rbpj-cKO ureters have been implicated in constriction (Tnnt2), relaxation (Pcp4) and energy conservation (Ckm) of cardiomyocytes, and in cardiomyopathies when deficient (Kim et al., 2014; Rentschler et al., 2012; Walker et al., 2021; Wei and Jin, 2016). Therefore, reduced expression of these genes/proteins may affect constriction and/or relaxation of ureteric SMCs, and contribute to hydroureter formation in Rbpj-cKO mice.

In the vascular system, Notch signaling regulates and synergizes with PDGFRB and TGFβ signaling in activation of early SMC genes (for reviews, see Baeten and Lilly, 2017; Fouillade et al., 2012). We did not find changes in expression of components or targets of these pathways in our transcriptional profiling experiments, suggesting that the molecular circuits regulated by Notch signaling in the control of visceral SMC differentiation are different from those in the vascular context.

In summary, we suggest that JAG1-activated NOTCH2 and/or NOTCH3 signaling regulates visceral SMC differentiation in the UM in a bimodal and biphasic manner. First, the module enhances Myocd expression to a critical level at E14.5; second, it enhances from around E17.5 the expression of a set of ‘late’ SMC genes that are crucial for long-term maintenance of ureter peristaltic activity (Fig. 8).

All alleles used in this study were maintained on an NMRI outbred background: Rbpj^tm1.1Hon (synonym Rbpj^fl) (Tanigaki et al., 2002), Gt(ROSA)26Sor^{tm1(Notch1)Dam} (synonym Rosa26^NICD) (Murtaugh et al., 2003), Tbx18^tm4(cre)Akis (synonym Tbx18^cre) (Trowe et al., 2010), Tbx18^{tm3.1(cre/ERT2)Sev} (synonym Tbx18^creERT2) (Guimaraes-Camboa et al., 2017) and Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo} (synonym Rosa26^mTmG) (Muzumdar et al., 2007). Embryos for ureter explant cultures and for expression analysis of genes encoding Notch components were obtained from matings of NMRI wild-type mice. Tbx18^cre/+;Rbpj^fl/+ males were mated with Rbpj^fl/fl females, and Tbx18^cre/+ or Tbx18^creERT2/+ males were mated with Rosa26^NICD/NICD females to obtain embryos for phenotypic characterization. Littermates without the cre/creERT2 allele were used as controls. Pregnancies were timed as embryonic day (E) 0.5 by the presence of vaginal plugs the morning after mating. Embryos and urogenital systems were dissected in PBS. Specimens were fixed in 4% PFA/PBS, transferred to methanol and stored at −20°C prior to further processing. PCR genotyping was performed on genomic DNA prepared from liver biopsies or yolk sacs.

Mice were housed in rooms with controlled light and temperature at the central animal laboratory of the Medizinische Hochschule Hannover. The experiments were in accordance with the German Animal Welfare Legislation and approved by the local Institutional Animal Care and Research Advisory Committee and permitted by the Lower Saxony State Office for Consumer Protection and Food Safety (AZ 33.12-42502-04-13/1356, AZ42500/1H).

Ureters were explanted on 0.4 µm polyester membrane Transwell supports (3450, Corning) and cultured in DMEM/F12 supplemented with 1% of concentrated stocks of penicillin/streptomycin, sodium pyruvate, glutamax, non-essential amino acids and IST-G (insulin-transferrin-selenium) (Thermo Fisher Scientific) at the air-liquid interface as previously described (Bohnenpoll et al., 2013). DAPT (GSI-IX) (S2215, Selleckchem) was used at final concentrations of 1 or 2.5 µM. To induce recombination with the Tbx18^creERT2 line, 4-hydroxytamoxifen (#H7904, Sigma-Aldrich) was used at a final concentration of 500 nM. Culture medium was replaced every 48 h. Analysis of frequencies and intensities of ureter contractions in explant cultures was performed by videomicroscopy as recently described (Weiss et al., 2019).

Embryos, urogenital systems or explant cultures were embedded in paraffin wax, and 5 μm sections were cut. Hematoxylin and Eosin staining was performed according to standard procedures.

For immunofluorescence and immunohistochemistry, the following primary antibodies and dilutions were used: polyclonal rabbit-anti-JAG1 (1:200; sc-8303, Santa Cruz Biotechnology), monoclonal mouse-anti-JAG1 (1:200; sc-390177, Santa Cruz Biotechnology), monoclonal rabbit-anti-NOTCH1 (1:200; 3608, Cell Signaling Technology), monoclonal rabbit-anti-NOTCH2 (1:200; 5732, Cell Signaling Technology), polyclonal rabbit-anti-NOTCH3 (1:200; ab23426, Abcam), monoclonal rat-anti-RBPJ (1:200; SIM-2ZRBP3, Cosmo BIO USA), polyclonal rabbit-anti-KRT5 (1:250; PRB-160P, Covance), polyclonal rabbit-anti-ΔNP63 (1:250; 619001, BioLegend), monoclonal mouse-anti-UPK1B (1:250; WH0007348M2, Sigma-Aldrich), polyclonal rabbit-anti-TAGLN (1:200; ab14106, Abcam), monoclonal mouse-anti-ACTA2 (1:200; A5228, Sigma-Aldrich), monoclonal rat-anti-EMCN (1:5, a kind gift from D. Vestweber, MPI Münster; Germany), polyclonal rabbit-anti-CD31 (1:400, 50408-T16, Sino Biological), monoclonal mouse-anti-GFP (1:200, 11814460001 Roche, Sigma-Aldrich) and polyclonal rabbit-anti-GFP (1:250, ab290, Abcam).

Primary antibodies were detected using the following secondary antibodies: biotinylated goat-anti-mouse IgG (1:400; 115-065-003, Dianova), biotinylated goat-anti-rabbit IgG (1:400; 111-065-033, Dianova), biotinylated goat-anti-rat IgG (1:400; 112-065-003, Dianova), biotinylated donkey-anti-goat IgG (1:400; 705-065-003, Dianova), Alexa 488-conjugated goat-anti-rabbit IgG (1:500; A11034, Thermo Fisher Scientific), Alexa 488-conjugated donkey-anti-mouse IgG (1:500; A21202, Thermo Fisher Scientific), Alexa 555-conjugated goat-anti-mouse IgG (1:500; A21422, Thermo Fisher Scientific) and Alexa 555-conjugated goat-anti-mouse IgG (1:500; A21428, Thermo Fisher Scientific).

The signals of NOTCH3, ΔNP63 and EMCN were amplified using the Tyramide Signal Amplification system (NEL702001KT, Perkin Elmer). For detection of JAG1, NOTCH1-3 and RBPJ, the DAB substrate solution (NEL938001EA, Perkin Elmer, Waltham, MA, USA) was used.

For antigen retrieval, paraffin wax-embedded sections were deparaffinized, pressure-cooked for 15 min in antigen unmasking solution (H3300, Vector Laboratories), treated with 3% H₂O₂/PBS for blocking of endogenous peroxidases, washed in PBST (0.05% Tween-20 in PBS) and incubated in TNB Blocking Buffer (NEL702001KT, Perkin Elmer). Sections were then incubated with primary antibodies at 4°C overnight. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI, 6335.1, Carl Roth).

Analysis of direct protein interactions on 5 µm transverse sections of the ureter and dorsal aorta region of wild-type embryos was performed with the proximity ligation assay (Bellucci et al., 2014; Gústafsdóttir et al., 2002) using the Duolink In Situ Red Starter Kit Mouse/Rabbit (DUO92101, Sigma-Aldrich/Merck) applying minor modifications of the manufacturer's instructions, as published before (Lüdtke et al., 2021). Antibody combinations were used and antibody retrieval was performed as described for immunofluorescence analysis. The primary antibody reaction was performed in blocking buffer from the tyramide signal amplification (TSA) system (NEL702001KT, PerkinElmer) overnight at 4°C, containing both corresponding primary antibodies for JAG1 and NOTCH1, NOTCH2 or NOTCH3 in a 1:100 dilution. After three washing steps with PBS/0.1% Tween20 for 5 min, the sections were blocked for 60 min with blocking buffer from the PLA kit and washed three times for 5 min in buffer A from the PLA kit before Duolink PLA probes were applied. Polymerase amplification reaction was performed for 150 min at 37°C.

In situ hybridization was carried out on 10 µm paraffin wax-embedded sections essentially as described (Moorman et al., 2001).

Ureters were dissected from male and female control and Tbx18^cre/+;Rbpj^fl/fl embryos. Forty specimens for each sex and genotype were pooled for analysis at E14.5, 12 specimens each for analysis at E18.5 and 10 specimens each for analysis at P4. Total RNA was extracted using peqGOLD RNApure (732-3312, 30-1010; PeqLab Biotechnologie) and subsequently sent to the Research Core Unit Transcriptomics of Hannover Medical School, where RNA was Cy3-labelled and hybridized to Agilent Whole Mouse Genome Oligo v2 (4×44K) microarrays (G4846A; Agilent Technologies). To identify differentially expressed genes, normalized expression data were filtered using Excel (Microsoft) based on an intensity threshold of 100 and a more than 1.4-fold change in all pools. Microarray data have been deposited in GEO under accession numbers GSE169661, GSE169662 and GSE184597.

RNA extraction and RT-PCR analysis for Myocd and Foxf1 expression was performed on pools of 10 ureters each of E14.5 control and Tbx18^cre/+;Rbpj^fl/fl embryos, as previously described (Weiss et al., 2019). For all other analyses, we isolated total RNA using TRIzol (15596-018, Thermo Fisher Scientific) and synthesized cDNA from 2.5 µg total RNA applying RevertAid H Minus reverse transcriptase (EP0452, Thermo Fisher Scientific) as described previously (Thiesler et al., 2021). The NCBI tool Primer3 version4.1 (Untergasser et al., 2012; Ye et al., 2012) was used to design specific primers (Table S16). RT-qPCR of mouse genes was performed in 10 µl 1:2 diluted BIO SyGreen Lo-ROX mix (PCR Biosystems) with 400 nM primers and 1 ng/µl cDNA using a QuantStudio3 PCR system fluorometric thermal cycler (Thermo Fisher Scientific). Each of the three biological replicates represents the average of four technical replicates. Data were processed by QuantStudio data analysis software (version1.5.1, Thermo Fisher Scientific) using the comparative threshold cycle (ΔΔC_T) method with Gapdh and Ppia as reference genes (Werneburg et al., 2015).

Statistical analysis was performed using the unpaired, two-tailed Student's t-test (GraphPad Prism version 7.03). Data are mean±s.d. P<0.05 was considered significant.

Sections and organ cultures were photographed using a Leica DM5000 microscope with a Leica DFC300FX digital camera or using a Leica DM6000 microscope with Leica DFC350FX digital camera. Urogenital systems were documented using a Leica M420 microscope with a Fujix HC-300Z digital camera (Fujifilm Holdings, Minato/Tokyo, Japan). All images were processed in Adobe Photoshop CS4.

We thank Doug Melton and Tasuku Honjo for mice; Dietmar Vestweber, Anja-Münster Kühnel, Henry Wedekind and Achim Gossler for antibodies; the Research Core Unit Transcriptomics of Hannover Medical School for microarray analyses; and Patricia Zarnovican for excellent technical support. F.Q. thanks the Hannover Biomedical Research School (HBRS) and the MD/PhD program Molecular Medicine for support. Some data in this study form part of a PhD thesis (J. Kurz, Gottfried Wilhelm Leibniz Universität Hannover, 2021).