Interplay of SHH, WNT and BMP4 signaling regulates the development of the lamina propria in the murine ureter

The lamina propria (LP) is a loose connective tissue that underlies the inner epithelial lining of most tubular organs. It provides mechanical support by tethering the epithelium to the surrounding mesenchymal wall. Moreover, its abundant and flexible extracellular matrix serves as a bed for various types of immune cells, capillary endothelial cells and nerve endings to fight external pathogens, nurture the epithelium, and sense and respond to stimuli from the luminal content (Andersson and McCloskey, 2014; Roulis and Flavell, 2016).

In the main organs of the urinary drainage system, the ureter and the bladder, the LP is highly compressible and elastic due to a meshwork of thick interwoven and crisscrossing collagen fibers that, upon dilatation, change their orientation to lie parallel to the specialized pseudo-stratified epithelium of the urinary drainage system, the urothelium (Aitken and Bägli, 2009; Macarak and Howard, 1997). The LP does not only play a structural role, it has been recognized as a crucial source of signals for the growth, differentiation and maintenance of the urothelium and the adjacent tunica muscularis (TM) (Andersson and McCloskey, 2014; Chun et al., 2007; Gandhi et al., 2013; Mysorekar et al., 2009).

Despite these important functions, the cellular and molecular mechanisms that control the development and homeostasis of the LP in the excretory system as in other tubular organs, have remained poorly understood. Genetic lineage tracing revealed that in the developing ureter of the mouse, LP fibrocytes arise together with smooth muscle cells (SMCs) of the TM and fibrocytes of the outer tunica adventitia (TA) from a common mesenchymal progenitor pool that surrounds the epithelium of the distal ureteric bud (Bohnenpoll et al., 2013) at embryonic day (E) 11.5. At E12.5, mesenchymal cells adjacent to the ureteric epithelium (UE) acquire a rhomboid shape and start to express the target of canonical WNT signaling Axin2. These Axin2⁺ progenitors activate at E14.5 the SMC regulator Myocd (Trowe et al., 2012; Wang and Olson, 2004). Cells in the direct vicinity of the UE switch Myocd expression off and differentiate into fibrocytes of the LP. The ones further away maintain Myocd expression and activate, in a stepwise fashion, expression of different SMC structural genes. Axin2⁻ cells in the outer region of the ureteric mesenchyme (UM) are lineage segregated from E13.5 onwards and differentiate into adventitial fibrocytes (Bohnenpoll et al., 2017a).

Tissue separation and recombination experiments have shown that development of the early (undifferentiated) UM depends on signals from the UE (Bohnenpoll and Kispert, 2014; Cunha, 1976). Subsequent work primarily focused on the signaling pathways and their transcriptional output that drive the differentiation of SMCs from the mesenchymal progenitor pool (Bohnenpoll and Kispert, 2014). SHH and WNTs (most likely WNT7B) were characterized as epithelial signals important for the proliferation of the mesenchymal progenitors and the differentiation of SMCs (Bohnenpoll et al., 2017c; Trowe et al., 2012; Yu et al., 2002). Downstream or parallel to these epithelial signals, mesenchymal BMP4 is required for mesenchymal expansion and SMC differentiation, whereas retinoic acid (RA) from both the UE and the UM maintains the undifferentiated character of the UM (Bohnenpoll et al., 2017b; Mamo et al., 2017; Wang et al., 2009). Interestingly, increased SHH signaling was associated with expansion of the LP in the ureter and bladder (Ikeda et al., 2017; Yu et al., 2002).

Here, we used time-resolved pharmacological inhibition and activation experiments in explant cultures to decipher the signaling inputs for the development of the LP in the murine ureter. We provide evidence that in the presence of SHH and WNT signaling, it is the lack of BMP4 signaling that favors the LP over the SMC fate.

To characterize LP development in the murine ureter, we performed both histological stainings as well as immunofluorescence analysis of marker proteins on transverse sections of the proximal ureter region in wild-type mice at prenatal and postnatal stages. Hematoxylin and Eosin staining discriminated cell layers due to the differential distribution of acid (Hematoxylin) and basic structures (Eosin) (Fischer et al., 2008) (Fig. 1A). Sirius Red visualized the deposition of collagens in the extracellular matrix (Sweat et al., 1964) (Fig. 1B). CDH1 marked the basolateral membranes of the UE. Combinatorial cytoplasmic expression of ACTA2 and TAGLN identified SMCs of an early differentiation stage; a late stage of differentiation was characterized by additional expression of CKM and SMTN in the cytoplasm of these cells (Bohnenpoll et al., 2017a; Kurz et al., 2022b) (Fig. 1C; Fig. S1). Based on findings that Aldh1a2 expression marks the LP at E18.5 (Bohnenpoll et al., 2017b; Trowe et al., 2012), we used ALDH1A2 as a marker for the cytoplasm of LP fibrocytes (Deuper et al., 2022) (Fig. 1D). COL1A2, CDH2, DES, THY1 and VIM, which have been used as LP markers in other biological contexts (Kuijpers et al., 2014; Poletajew et al., 2015; Powell et al., 2011; Roulis and Flavell, 2016), proved unsuitable in the ureter (Fig. S2).

At the end of the undifferentiated phase, at E14.5, the UM appeared to be histologically subdivided into a dense inner region of large cuboidal cells and an outer region of loosely organized fibrocytic cells. The UE was positive for CDH1; the UM neither expressed SMC markers nor ALDH1A2 at this stage. At E16.5, the cells of the inner region of the UM co-expressed ACTA2 and TAGLN, indicating early SMC differentiation. At E18.5 and postnatal day (P) 0, patches of ALDH1A2⁺ cells were found in between the ACTA2⁺TAGLN⁺CKM⁺SMTN⁺ late SMCs and the CDH1⁺ urothelium. At P7 and P40, the population of SMC marker-negative mesenchymal cells underneath the urothelium appeared to be largely expanded, and expressed ALDH1A2 homogenously. Sirius Red staining indicated a massive deposition of collagens in the LP (Fig. 1; Fig. S1). Quantification of ACTA2⁻CDH1⁻ suburothelial mesenchymal cells at postnatal stages confirmed a strong increase of LP fibrocytes between P0 and P7 (Fig. 1E; Table S1A). The LP layer gained some thickness in this time interval but further expanded until P40 (Fig. 1F; Table S1B). We conclude that ALDH1A2 expression allows robust detection of LP fibrocytes from E18.5/P0 until at least P40. The major expansion of the LP occurs after birth and is associated with a strong increase of LP fibrocytes until P7, and collagen deposition until and after this stage.

Previous work has implicated SHH, WNT, BMP4 and RA signaling in the development of ureteric SMCs (Bohnenpoll et al., 2017b,c; Mamo et al., 2017; Trowe et al., 2012; Yu et al., 2002). To investigate whether these pathways are also associated with LP development, we performed section RNA in situ hybridization analysis to profile the expression of genes encoding the signals as well as targets of their activity during ureter development (Fig. 2).

Shh was expressed at low levels in the UE at E14.5 and E16.5. From E18.5 onwards, increased expression was confined to the basal and intermediate cell layer of the UE. Expression of Ptch1, encoding a receptor for SHH and presenting a target gene of SHH signaling (Ingham and McMahon, 2001), was found in the inner region of the UM at E14.5 and E16.5, and from E18.5 onwards in the LP. Wnt7b was strongly expressed in the UE at E14.5 and E16.5; at subsequent stages expression was found at lower levels in the urothelium, most likely in B- and I-cells. Expression of Axin2, a direct target gene of WNT signaling (Jho et al., 2002), mirrored the spatial pattern of Ptch1. Expression of Bmp4 was found in the inner UM at E14.5 and E16.5. From E18.5 onwards, expression was restricted to the LP. Expression of Id2, a direct target gene of BMP signaling (Hollnagel et al., 1999), occurred at high levels both in the UE and the UM at E14.5. Expression continued at E16.5 in the inner region of the UM, and at E18.5 to P7 in the TM. At P7, expression also occurred in the basal and intermediate layers of the UE. The gene encoding the RA-synthesizing enzyme Aldh1a2 was expressed in the LP at E18.5, P0 and P7. Rarb, encoding a receptor of RA and presenting a direct target gene of RA signaling (Mendelsohn et al., 1991), was expressed in the inner UM at E14.5. From E18.5 to P7, weak expression was found in the UE. We conclude that SHH, WNT, BMP4 and RA signaling occur strongly at E14.5 in the inner region of the undifferentiated UM where the common progenitor of SMCs and LP fibrocytes resides. From E18.5 onwards, the LP becomes a target of SHH and WNT and a source of BMP4 and RA signals.

To characterize the individual and combinatorial function of these signaling pathways in LP development, we wished to employ a pharmacological approach in ureteral explant cultures. Centered on concentrations that are known to affect ureteric SMC development (Bohnenpoll et al., 2017b,c; Trowe et al., 2012), we first tested the efficacy of pathway inhibitors and activators for two developmental stages. For this, we explanted ureters together with kidneys at E12.5 (please note that embryonic ureters grow better in this combination culture and stick better on the membrane during later transfer steps) and ureters only at P0 on polyester membranes, cultured them for 18 h at the air-liquid interface in a range of concentrations of the compounds, and performed whole-mount RNA in situ hybridization for direct target genes as a read-out (see above; Figs S3 and S4). For SHH signaling (read-out: Ptch1), we found robust pathway inhibition by the SMO antagonist cyclopamine at 5 and 10 µM, and activation by the SMO agonist purmorphamine at 2 and 10 µM (Bohnenpoll et al., 2017c; Chen et al., 2002; Cooper et al., 1998). The WNT pathway (read-out: Axin2) was inhibited by the Porcupine inhibitor IWP-2 at 5 and 10 µM and activated by the GSK-3 inhibitor BIO at 10 and 20 µM (Karner et al., 2011; Sato et al., 2004). The BMP4 pathway (read-out: Id2) was inhibited by recombinant Noggin (NOG) at 5 and 10 µg/ml, and activated by recombinant BMP4 at 100 and 250 ng/ml (Bohnenpoll et al., 2017c; Mamo et al., 2017; Zimmerman et al., 1996). Inhibition of RA signaling (read-out: Rarb) was achieved by the pan-RAR antagonist BMS493 at 1 and 5 µM. Overactivation occurred with RA at the same concentrations (Bohnenpoll et al., 2017b; Chazaud et al., 2003). Hence, robust pathway inhibition and activation can be achieved for SHH, WNT, BMP4 and RA signaling in ureteral explant cultures at fetal and early postnatal stages.

To analyze the role of these signaling pathways in the fetal period of LP development (when LP fibrocytes arise together with SMCs from a common mesenchymal progenitor), we explanted E12.5 wild-type ureters (with kidneys) and cultured them for 2 days without drug to reach a stage similar to E14.0 to E14.5 ureters in vivo (Fig. S5), before we added inhibitors and activators in the same concentration range as above for 6 days (Figs S3 and S4). We performed immunofluorescence analysis on sections for expression of the SMC marker ACTA2 and the LP marker ALDH1A2 (Fig. 3, left panel) and quantified the number and ratio of ACTA2⁺ALDH1A2⁻ SMCs, ACTA2⁻ALDH1A2⁺ differentiated LP fibrocytes and ACTA2⁻ALDH1A2⁻ ‘undifferentiated’ mesenchymal cells (Fig. 3, right panel; Tables S2A-S9A).

Treatment with cyclopamine led to a strong reduction of mesenchymal cells and an almost complete loss of differentiated ACTA2⁺ SMCs and ALDH1A2⁺ LP fibrocytes from 1 µM onwards (Fig. 3A, first row; Table S2A). Treatment with purmorphamine led to a dose-dependent increase of both differentiated SMCs and LP fibrocytes (Fig. 3A, second row; Table S3A). Inhibition of WNT signaling by IWP-2 resulted in a reduction of mesenchymal cells, with a loss of differentiated LP fibrocytes and a dose-dependent reduction of SMCs from 1 to 10 µM of the inhibitor (Fig. 3B, first row; Table S4A). BIO treatment led to a dose-dependent increase of the overall mesenchymal cell number largely due to gain of ‘undifferentiated’ LP cells (Fig. 3B, second row; Table S5A). NOG treatment led to a dose-dependent loss of SMCs, accompanied by a gain of undifferentiated mesenchymal cells. The number of ALDH1A2⁺ cells remained largely unchanged (Fig. 3C, first row; Table S6A). Treatment with BMP4 increased the number of SMCs and of undifferentiated LP cells (at 100 and 250 ng/ml) (Fig. 3C, second row; Table S7A). Inhibition of RA signaling did not affect mesenchymal differentiation and cell number in a clear dose-dependent way. However, at 5 µM BMS493 treatment, SMCs and differentiated LP cells were strongly reduced while undifferentiated mesenchymal cells were increased (Fig. 3D, first row; Table S8A). Treatment with RA led to a small reduction of LP cells at the highest dose of 5 µM (Fig. 3D, second row; Table S9A).

To validate that changes of ALDH1A2 expression reflect changes in the number of LP fibrocytes (and not merely the loss or gain of a marker), we additionally analyzed ACTA2 and CDH1 expression, identifying LP fibrocytes as ACTA2⁻CDH1⁻ (unstained) cells in between the TM and the UE (Fig. S6, left panel). The changes of numbers and ratios of LP cells and SMCs were congruent with the above findings (Fig. S6, right panel; Tables S2B-S9B). Together with the fact that the dose response curves fit to the findings of pathway inhibition and activation (Figs S3 and S4), we conclude that in fetal ureter development SHH and WNT signaling are positive regulators of differentiation and expansion of LP fibrocytes and SMCs from the common progenitor. BMP4 signaling is required for differentiation of SMCs but not for LP cells.

We next investigated whether these signaling pathways also affect the early postnatal phase of LP development (when LP fibrocytes largely increase in number) by culturing P0 ureteral explants for 6 days with increasing concentrations of compounds as above. We again performed immunofluorescence analysis on sections for expression of the SMC marker ACTA2 and the LP marker ALDH1A2 (Fig. 4, left panel) and quantified the number and ratio of ACTA2⁺ALDH1A2⁻ SMCs, ACTA2⁻ALDH1A2⁺ differentiated LP fibrocytes and ACTA2⁻ALDH1A2⁻ ‘undifferentiated’ mesenchymal cells (Fig. 4, right panel; Tables S10A-S17A).

Inhibition of SHH signaling with 5 and 10 µM cyclopamine led to a loss of ALDH1A2 expression and a reduction of the mesenchymal cell number, with a stronger reduction of ‘LP cells’ than of SMCs (Fig. 4A, first row; Table S10A). Purmorphamine led to dose-dependent increase of both SMCs and LP fibrocytes; SMCs were disproportionally increased at 10 µM (Fig. 4A, second row; Table S11A). Inhibition of WNT signaling by IWP-2 led to a dose-dependent reduction of LP fibrocytes; the number of SMCs was weakly reduced (Fig. 4B, first row; Table S12A). BIO treatment was without effect (Fig. 4B, second row; Table S13A). Inhibition of BMP4 signaling led to a slight increase of ALDH1A2⁺ LP fibrocytes (peaking at 5 µg/ml NOG) (Fig. 4C, first row; Table S14A). BMP4 treatment increased the ratio of SMCs to LP fibrocytes at 100 and 250 ng/ml, probably due to a slight (but insignificant) increase of the total SMC number. In either case, the effect was weak and lacked a clear dose-response curve (Fig. 4C, second row; Table S15A). Manipulation of RA signaling did not affect the number of LP fibrocytes and SMCs at this stage (Fig. 4D; Tables S16A and S17A).

Analysis of ACTA2 and CDH1 expression, and quantification of SMCs as ACTA2⁺ cells and LP fibrocytes as ACTA2⁻CDH1⁻ (unstained) cells in between the TM and the UE were congruent with the above findings (Fig. S7; Tables S10B-S17B).

These findings suggest that SHH signaling is required for maintenance of ALDH1A2⁺ LP fibrocytes, whereas both SHH and WNT signaling account for the increase of the number of LP fibrocytes after birth. BMP4 signaling seems to slightly favor SMCs over LP fibrocytes, while RA signaling is irrelevant for postnatal LP development.

As both SHH and WNT signaling affected LP expansion in postnatal ureter explants, we characterized their epistatic relationship by combinatorial administration of pathway inhibitors and activators in explant cultures of P0 ureters (Fig. 5A-C; Table S18). Combined treatment with 1 µM of cyclopamine and IWP-2, i.e. concentrations which did not affect LP and SMC development (cyclopamine, Fig. 4A; Tables S10 and S11) or led to a weak reduction of LP fibrocytes only (IWP-2, Fig. 4B; Tables S12 and S13) resulted in a strong reduction of ALDH1A2⁺ LP fibrocytes and of ACTA2⁺ SMCs, enhancing the effects of individual pathway inhibition. Combined treatment with purmorphamine and BIO increased the total cell number but did not affect the ratio of mesenchymal cell types (as did purmorphamine treatment alone). Activation of SHH signaling rescued LP cell number in conditions of WNT inhibition. WNT signaling activation did not rescue the number of ALDH1A2⁺ differentiated LP cells when SHH signaling was inhibited, but did expand the number of undifferentiated LP cells. Analysis of ACTA2/CDH1 co-stainings confirmed these findings. We conclude that in early postnatal ureter development, SHH signaling accounts for differentiation of LP cells, whereas SHH signaling cooperates with WNT signaling in expansion of LP cells and SMCs.

To resolve the stage-dependent interaction of BMP4 and SHH signaling in mesenchymal differentiation, we explanted ureters at E12.5 and P0, and combined BMP4 signaling inhibition with SHH signaling activation. In E12.5 ureteral explants co-treated with purmorphamine (10 µM) and NOG (10 µg/ml), all mesenchymal cells exclusively expressed ALDH1A2 after 6 days of culture. In E12.5 explants precultured for 2 days to resemble E14.5 ureters, the number of ALDH1A2⁺ LP fibrocytes and ACTA2⁺ SMCs was increased; some mesenchymal cells co-expressed both markers after 6 days of compound treatment. In P0 explant cultures, SMCs were increased (Fig. 5D,E; Table S19). This suggests that BMP4 suppresses SHH-induced ALDH1A2 expression in uncommitted mesenchymal progenitors at E12.5. In committed progenitors (E12.5+2 days), and even more in differentiated SMCs (P0), BMP4 signaling does not appear to be required to maintain the SMC phenotype, nor does it have a major repressive effect on ALDH1A2 expression and LP fate.

We next asked whether SHH, WNT, BMP4 and RA signaling affect apoptosis and/or proliferation in postnatal (P0) ureters. The TUNEL assay did not detect apoptotic bodies in P0 ureters cultured for 18 h with inhibitors of SHH, WNT, BMP4 or RA signaling (Fig. 6A). However, the BrdU incorporation assay revealed strongly reduced cell proliferation in the LP upon inhibition of SHH or WNT signaling. The latter also affected proliferation of SMCs. Inhibition of BMP4 and RA signaling did not affect mesenchymal cell proliferation (Fig. 6B-D; Table S20). This suggests that SHH and WNT signaling account for the proliferative expansion of the LP at postnatal stages.

As SHH and WNT signaling act positively on both LP proliferation and/or differentiation, we explored whether they are interdependent and/or affect similar molecular programs. For this, we cultured P0 ureters for 18 h with the respective inhibitors (10 µM cyclopamine or 5 µM IWP-2), and profiled the transcriptional changes by microarray technology. Using filtering by intensity (>100 in control) and fold change (<−1.5), we found 119 genes with reduced expression in cyclopamine-treated ureters (Fig. 7A; Table S21; GSE282233). Functional annotation by DAVID revealed enrichment of terms related to forkhead transcription factors and various types of organ development in this gene set reflecting downregulation of Foxl1, Foxf1 and Foxd1 genes as well as of genes involved in various developmental programs including Bmp4 and Aldh1a2. Ptch1 expression was strongly reduced, validating SHH signaling inhibition (Tables 1A,2A; Table S22).

Using the same filtering conditions, 65 genes with reduced expression were found in IWP-2-treated ureters (Fig. 7B; Table S23; GSE282232). DAVID revealed enrichment of terms related to WNT signaling, reflecting reduced expression of components and/or targets of WNT signaling including Axin2 (Tables 1B,2B; Table S24).

Overlapping both lists identified ten common genes (Fig. 7C; Table 3A). Four additional common genes were found with less stringent conditions of filtering (Table 3B). This suggests that SHH and WNT signaling largely activate unique sets of genes in the LP but also share a (small) set of common targets.

To validate these findings and determine tissue specificity of expression, we performed RNA in situ hybridization analysis for individual candidates on ureter sections. We found a set of genes (Ahr, Aldh1a2, Avpr1a, Axin2, Bmp4, Car3, Foxf1, Itga4, Ptch1, Thy1) that showed specific expression in the LP at P0. Detection of expression was more difficult in cultured ureters but, in agreement with the microarray data, expression of Aldh1a2, Avpr1a, Bmp4, Foxf1, Itga4, Ptch1 and Thy1 was clearly reduced in cyclopamine-treated ureters, whereas in IWP-2-treated ureters expression of Ahr, Aldh1a2 (weakly), Avpr1a, Axin2, Itga4 and Thy1 was reduced. Expression of Ptch1 appeared to be slightly increased in IWP-2-treated ureters, as was Axin2 in cyclopamine-treated ureters (Fig. 7D). We conclude that SHH and WNT signaling have distinct and common targets in the LP at perinatal stages, and may partially attenuate each other's activity.

The lack of specific and robust cytodifferentiation markers has made the characterization of LP fibrocytes difficult (Andersson and McCloskey, 2014; Roulis and Flavell, 2016). Both in the adult bladder and intestine, LP fibrocytes were distinguished from adjacent SMCs by their localization, the lack of specific SMC markers (such as SMTN and DES) and the presence of COL1A2, VIM, CDH2 or in some cases THY1 (Kuijpers et al., 2014; Poletajew et al., 2015; Powell et al., 2011; Roulis and Flavell, 2016). We tested all of these markers in P7 ureters and found that they detected no antigen (COL1A2, THY1), stained most likely nerves (CDH2) and SMCs (DES), and marked the LP and SMCs in a patchy fashion, proving them unsuitable for LP detection in this tissue context.

It has been described that Aldh1a2 mRNA is expressed in the LP of the E18.5 ureter (Trowe et al., 2012). However, this expression is relatively weak, it does not allow the visualization of individual cells and works poorly on adult tissue. We tested the suitability of anti-ALDH1A2 immunofluorescence for LP fibrocyte detection. In fact, ALDH1A2 recapitulated the expression of Aldh1a2 mRNA at E18.5 and P7 in sub-urothelial mesenchymal cells, and allowed sensitive detection of individual cells by their cytoplasmic staining from E18.5 to adult stages. Importantly, ALDH1A2 expression comprised all ACTA2⁻CDH1⁻ suburothelial mesenchymal cells, indicating that LP fibrocytes are a homogenous cell population throughout development and homeostasis, and are clearly demarcated from the adjacent SMCs.

Expression of Aldh1a2 was also reported in the sub-urothelial stroma in the developing and adult bladder (Gandhi et al., 2013), in the stroma of proximal Müllerian ducts (Nakajima et al., 2016), and in the uterine stroma (Ma et al., 2012), while expression of the related Aldh1a3 gene was found in the developing LP of the small intestine (Niederreither et al., 2002). Therefore, expression of RA synthesizing enzymes may more generally mark the LP in different developing and adult organ contexts.

SMC differentiation occurs in the mesenchymal wall of the murine ureter between E15.5 to E18.5, leading to a peristaltically active tube that is able to cope with the urine that is produced in the fetal kidney from E16.5 onwards (Bohnenpoll et al., 2017a; Kurz et al., 2022b). Our histological and immunofluorescent analyses confirmed the presence of a contiguous layer of ACTA2⁺ cells in the mesenchymal wall of the ureter at E16.5 and E18.5, which only marginally gained width at postnatal stages. In contrast, only few LP fibrocytes were detected in late fetal ureters, while a contiguous LP with numerous fibrocytes embedded within a collagen-rich extracellular matrix was found at P7 and gained further width until P40. Hence, LP development is delayed compared to that of TM.

In explant cultures of P0 ureters, LP cells exhibited a proliferation rate more than double that of the adjacent muscle layer. A similar difference in proliferation rates was reported in E18.5 ureters in vivo (Bohnenpoll et al., 2017a). Proliferation rates of LP fibrocytes at P0 resembled those of the undifferentiated UM at E12.5, indicating that they maintain a precursor character to allow proliferative expansion of the tissue at late fetal and early postnatal stages.

Aldh1a2 expression is found in the entire UM at E11.5 and is subsequently downregulated in the inner domain of the UM until E14.5 (Bohnenpoll et al., 2017b). Although this may support the idea that Aldh1a2/ALDH1A2 expression marks the aforementioned precursor state of LP fibrocytes, we deem this is unlikely given that ALDH1A2 expression in the LP is maintained at P40 when all mesenchymal cells in the ureteric wall have become quiescent (Bohnenpoll et al., 2017a).

SMCs and LP fibrocytes develop from a common Axin2⁺Myocd⁺ precursor that resides in the inner layer of the UM at E14.5. While all of these progenitors activate the SMC differentiation program until E15.5, some cells underneath the urothelium switch off the myogenic program, express ALDH1A2 and differentiate into LP fibrocytes from E16.5 onwards (Bohnenpoll et al., 2017a). Our expression analysis of E14.5 ureters revealed that SHH, WNT, BMP4 and RA signaling are all active in the committed SMC/LP progenitor in the inner ring of the UM. We addressed which signaling activity impinges on the differentiation and lineage diversification of these committed progenitors by pharmacological pathway manipulations in explant cultures of E12.5+2 day ureters, which resemble E14.0 to E14.5 ureters in vivo. Inhibition of SHH or WNT signaling compromised mesenchymal expansion and abolished both SMC and LP fibrocyte differentiation, clearly showing that these two epithelial signals are required for the expansion and differentiation of the committed progenitor into both cell types but not for lineage diversification. Moreover, broad and strong activation of SHH signaling or of WNT signaling in the UM did not lead to a switch of SMCs to LP cells but to an expansion of both cell types. These findings contradict the suggestion that segregation of LP fibrocytes and SMCs from the common mesenchymal progenitor is due to differential SHH signaling, in which high doses (in the periureteral region) induce LP development, possibly by repressing the SMC fate, while low doses favor SMC differentiation (Cao et al., 2010; Yu et al., 2002).

In contrast, inhibition of BMP4 signaling compromised SMC differentiation, while differentiation of suburothelial mesenchymal cells into LP fibrocytes was not affected. This shows that positive signals for LP differentiation (SHH and WNTs) are independent of BMP4 signaling, and act onto the adjacent mesenchymal cells only. Moreover, it suggests that lineage diversification is due to a selective requirement for BMP4 signaling in the early SMC program and absence of BMP4 signaling in the LP lineage, respectively. This notion is further supported by expression analysis of the BMP4 signaling target Id2 in ureter development. Id2 expression was found in the inner ring of the UM at E14.5, i.e. in the committed progenitor, but was subsequently lost in the developing LP.

In E12.5 ureters co-treated with purmorphamine and NOG, the largely expanded mesenchymal cell population lost ACTA2 and gained ALDH1A2 expression. This shows that in the undifferentiated UM, i.e. in uncommitted progenitors, SHH is the activator of ALDH1A2 expression, whereas BMP4 signaling represses it.

In the undifferentiated UM, SHH signaling induces expression of the transcription factor FOXF1. FOXF1 is required but not sufficient for Myocd activation and SMC differentiation (Bohnenpoll et al., 2017c). In fact, FOXF1 synergizes with BMP4-induced GATA6 in this program (Kurz et al., 2022a). Absence of BMP4 signaling and GATA6 expression in the LP at perinatal and postnatal stages (this study and Kurz et al., 2022a), provides an explanation that Foxf1 expression in this domain (this study and Fink et al., 2022), does not confer the SMC fate.

We have recently shown that loss of Fgfr2 in the undifferentiated UM results in increased SHH and decreased BMP4 signaling, and premature activation of the LP at the expense of the SMC program (Deuper et al., 2022). These findings are entirely consistent with the conclusion of this study that LP development requires SHH (and WNT) signaling input but absence of BMP4 signaling. They suggest that lack of FGFR2-signaling-induced BMP receptor expression accounts for the loss of BMP4 signaling in LP cells.

High doses of the pan-RA receptor antagonist BMS493 in fetal ureter cultures led to a strong decrease of both SMCs and LP cells, and an increase of undifferentiated cells. This is compatible with the previously described role of RA in expanding the mesenchymal progenitors and/or inhibiting their differentiation (Bohnenpoll et al., 2017b).

Our histological and marker analyses showed that the LP of the murine ureter arises as clutches of ACTA2⁻ALDH1A2⁺ suburothelial mesenchymal cells at late fetal stages and develops at early postnatal stages (from P0 to P7) into a contiguous cell layer with numerous fibrocytes embedded within a collagen-rich extracellular matrix. In the first postnatal week, the LP is a target of SHH and WNT signals, and a source of BMP4 and RA. While BMP4 and RA signaling appear to be dispensable for proliferation and differentiation of LP fibrocytes and SMCs in this phase, WNT and SHH signaling are again highly relevant for these cellular processes. SHH inhibition abolished ALDH1A2 expression and affected the number of LP cells, and had a weak effect on the number of SMCs. Activation of SHH signaling slightly increased the number of LP fibrocytes but strongly expanded the number of SMCs, suggesting that SHH signals induce differentiation and proliferative expansion of LP cells, and are a limiting factor for SMC expansion. WNT signaling inhibition also reduced LP and SMC expansion but did not affect the differentiation of either cell type. Overactivation of WNT signaling did not affect the number of SMCs or LP cells, suggesting that WNTs are required for expansion of LP cells and SMCs but do not represent a limiting factor.

Together with the finding that co-inhibition of both SHH and WNT signaling reduced the number of LP cells and SMCs more strongly than individual inhibition, and that activation of one pathway can rescue the reduction of LP cells by the other pathway, we conclude that both pathways cooperate in the proliferative expansion of both LP cells and SMCs. Interestingly, inhibition of SHH signaling lowered LP proliferation, whereas WNT signaling inhibition reduced LP cell proliferation but also affected more weakly SMC proliferation We conclude that SHH and WNT pathways act in a concentration-dependent manner, with WNT signaling acting in a wider spatial range than SHH signaling.

Our transcriptional profiling uncovered forkhead transcription factor genes including Foxf1/Foxf2, but also Bmp4 and Aldh1a2 as SHH-dependent genes in the early postnatal LP. As the Foxf1/Foxf2-Bmp4 module mediates SHH signaling in the undifferentiated UM at E12.5 (Bohnenpoll et al., 2017c), it is likely that SHH signaling effectors are conserved at both developmental stages. WNT signaling was required for expression of many genes, including a small set of SHH-dependent genes in the LP, showing that the two pathways have common but mostly distinct targets in this tissue. Intriguingly, expression of Ptch1 and Axin2 appeared to be slightly upregulated in postnatal ureters exposed to IWP-2 and cyclopamine, respectively, which may indicate a mutual dampening effect of the two pathways.

Together, our study provides a molecular explanation for the development of the LP in the murine ureter (Fig. 8). SHH and WNT signals from the UE are essential for formation of the common SMC/LP progenitor in the UM, and its subsequent proliferation and differentiation into SMCs and LP fibrocytes. RA signaling expands the progenitors and inhibits their differentiation. BMP4 signals support SMC differentiation of the progenitor. It is the absence of BMP4 signaling which segregates the LP from the SMC fate. In the postnatal phase, SHH signaling maintains the differentiation and proliferation of LP cells, the latter of which is supported by WNT signals.

NMRI mice were used for all experiments. For timed pregnancies, vaginal plugs detected at noon after mating were designated as E0.5. P0 was designated the day of birth. Embryos and urogenital systems were dissected in PBS, ureters for explant cultured in L-15 Leibovitz medium (BS.F1315, BIO&SELL). Specimens were fixed in 4% paraformaldehyde (PFA)/PBS, transferred to methanol and stored at −20°C before immunofluorescence or in situ hybridization analysis.

Mice were housed in rooms with controlled light and temperature. The experiments were performed in accordance with the German Animal Welfare Legislation (TierSchG). They were approved by the local Institutional Animal Care and Research Advisory Committee and authorized by the Lower Saxony State Office for Consumer Protection and Food Safety (reference number 42500/1H).

Ureters were explanted onto a 0.4 µm pore size Transwell polyester membrane (#3450, Corning) and cultured at the air-liquid interface in DMEM/F12 (#21331020, Gibco) supplemented with 10% fetal calf serum (S0615, Merck) and 1% of concentrated stocks of penicillin/streptomycin (#15140122, Gibco), pyruvate (#11360070, Gibco), Glutamax (#35050038, Gibco), NEAA (100×, #11140035, Gibco) with 5% CO₂ at 37°C. For pharmacological manipulation of SHH signaling, we used cyclopamine [#S1146, Selleck Chemicals (SelleckChem)] dissolved in ethanol, and purmorphamine (#S3042, SelleckChem) dissolved in DMSO. For RA signaling manipulation, we used BMS493 (#3509, Tocris; dissolved in DMSO), and RA (#0695, Tocris; dissolved in DMSO). For WNT signaling manipulation, we used IWP-2 (#S7085, SelleckChem; dissolved in DMSO) and BIO (#S7198, SelleckChem; dissolved in DMSO). For BMP4 signaling manipulation, we used NOG (#Z03205, GenScript; dissolved in MilliQ-H₂O) and BMP4 (#5020-BP, R&D Systems; dissolved in 4 mM HCl). Medium containing the respective solvents (control) or compounds was refreshed every second day.

Embryos, urogenital systems and ureters were paraffin wax-embedded and sectioned to 5 µm (immunofluorescence, histological staining) or 10 µm (RNA in situ hybridization) using a Leica RM2155 or a Leica Jung RM2035 manual rotary microtome. Histological staining with Hematoxylin and Eosin (GHS332-1L, Merck/Sigma-Aldrich) and Picrosirius Red (#365548, Sigma-Aldrich) followed standard procedures (Fischer et al., 2008; Junqueira et al., 1979).

Non-radioactive RNA in situ hybridization analysis was performed on whole cultured ureters and on 10 µm transversal paraffin wax sections of the proximal ureter using digoxigenin-labeled antisense riboprobes as previously described (Moorman et al., 2001; Wilkinson and Nieto, 1993).

Immunofluorescence analysis of marker proteins was performed on 5 µm paraffin wax sections with primary and secondary antibodies as listed in Table S25. The signals of ALDH1A2 and BrdU primary antibodies were amplified using the Tyramide Signal Amplification (TSA) system (NEL702001KT, PerkinElmer). Before staining, paraffin sections were deparaffinized and cooked for 15 min in antigen unmasking solution (H-3300, Vector Laboratories). Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI, #6335.1, Carl Roth).

Two independent pools each of five or six male or female P0 ureters cultured for 18 h with solvent (control) or with compounds (10 µM cyclopamine or 5 µM IWP-2) were used for microarray analysis. Total RNA from each pool was extracted using the RNeasy Micro Kit (#74004, Qiagen) and subsequently processed by the Research Core Unit Transcriptomics of Hannover Medical School for hybridization on whole Mouse Genome Oligo v2 (4×44K) microarrays (#G4846A; Agilent Technologies). Normalized expression data were filtered using Microsoft Excel. Functional enrichment analysis was performed using DAVID 6.8 web software (david.ncifcrf.gov); terms were selected based on P-values. Microarray data were submitted to Gene Expression Omnibus (GEO) under accession numbers GSE282232 and GSE282233.

Apoptosis was assessed by the TUNEL assay using the ApopTag Plus Fluorescein Detection Kit (S7111, Merck) on 5 μm paraffin wax sections of the proximal region of P0 ureters cultured for 18 h in the presence of inhibitors of signaling activities. Proliferation rates were assayed by detection of incorporated BrdU on 5 µm paraffin wax sections of the proximal region of P0 ureters cultured for 18 h in the presence of inhibitors of signaling activities. BrdU was added for 3 h at the end of the culture period at a concentration of 3.3 µg/ml (Bohnenpoll et al., 2017b). At least four sections of each specimen (n>4) were analyzed. The TM was defined by ACTA2-staining and the urothelium was marked by CDH1 expression. LP fibrocytes in between both compartments were considered negative for both markers. The BrdU-labeling index was defined as the number of BrdU-positive nuclei relative to the total number of nuclei as detected by (DAPI) counterstaining.

For quantification of cell types, 5 µm ureter sections were co-stained for expression of ACTA2 and ALDH1A2 or for ACTA2 and CDH1. Nuclei were counterstained with DAPI. In the case of the ACTA2/ALDH1A2 co-staining, nuclei with surrounding cytoplasmic ACTA2 staining were considered as SMCs and nuclei with surrounding cytoplasmic ALDH1A2 expression were considered as LP cells. In the case of the ACTA2/CDH1 co-staining, LP cells were recognized as nuclei lying between the ACTA2⁺ SMCs and CDH1⁺ urothelium. Cells were manually counted using the counting tool in Adobe Photoshop CS4 extended.

For the measurement of the LP thickness, sections were co-stained with antibodies against ACTA2 and CDH1. LP cells were recognized as nuclei lying between the ACTA2⁺ SMCs and CDH1⁺ urothelium. Using the measuring tool of the Fiji/ImageJ software, the distance between the ACTA2 and CDH1 staining was measured at five distinct positions per section and averaged. The measurements were converted to micrometers for statistical analysis.

Statistical analysis was performed using GraphPad Prism Version 7a. The Shapiro-Wilk normality test was performed to analyze the distribution within the different data groups [normal distribution (parametric), P-values≥0.05]. The F-test of equality of variances was used to analyze the variances of two data groups (equal variances, P-values≥0.05). Depending on the results of the Shapiro-Wilk and the F-test, different tests were used to compare two different groups. To compare two groups with normal distribution and equal variances the unpaired two-tailed t-test was applied. The two-tailed Welch's t-test was used if two groups with normal distribution had unequal variances. The two-tailed Mann–Whitney U-test was used to compare two groups with abnormal distribution (non parametric). Results were shown as means with standard deviation. A P-value of <0.05 was considered as statistically significant (*), a P-value of P<0.01 as highly significant (**) and P-values of P<0.001 (***) as extremely significant.

Sections were photographed using a DM5000 microscope (Leica Camera) with a Leica DFC300FX digital camera, or a Leica DMI6000B microscope with a Leica DFC350FX digital camera. All images were assembled into figures in Adobe Photoshop 2020, CS3 or CS4.

The Research Core Unit Transcriptomics and Genomics of Hannover Medical School validated the quality of RNAs, generated labeled cRNAs, performed hybridizations on microarrays and provided normalized expression data. We thank Rolf Kemler (Max Planck Institute of Immunobiology and Epigenetics, Freiburg, Germany) for the anti-CDH1 antiserum.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.204214.reviewer-comments.pdf