Urothelial progenitors in development and repair

Abstract

Urothelium is a specialized multilayer epithelium that lines the urinary tract from the proximal urethra to the kidney. In addition to proliferation and differentiation during development, urothelial injury postnatally triggers a robust regenerative capacity to restore the protective barrier between the urine and tissue. Mounting evidence supports the existence of dedicated progenitor cell populations that give rise to urothelium during development and in response to injury. Understanding the cellular and molecular basis for urothelial patterning and repair will inform tissue regeneration therapies designed to ameliorate a number of structural and functional defects of the urinary tract. Here, we review the current understanding of urothelial progenitors and the signaling pathways that govern urothelial development and repair. While most published studies have focused on bladder urothelium, we also discuss literature on upper tract urothelial progenitors. Furthermore, we discuss evidence supporting existence of context-specific progenitors. This knowledge is fundamental to the development of strategies to regenerate or engineer damaged or diseased urothelium.

Introduction

Urothelium is a specialized epithelium that lines the urinary tract from the proximal urethra to the kidney (Figure 1). In addition to proliferation and differentiation during development, urothelial injury postnatally triggers a robust regenerative capacity to restore the protective barrier between the urine and tissue. Mounting evidence supports the existence of dedicated progenitor cell populations that give rise to urothelium during development and in response to injury [1]. Understanding the cellular and molecular basis for urothelial patterning and repair will inform tissue regeneration therapies designed to ameliorate a number of structural and functional defects of the urinary tract. This review will discuss evidence in support of urothelial progenitors and the context dependent mechanisms by which they guide urothelial patterning and repair.

Figure 1.

Diagram shows urothelium from regions of the urinary tract discussed in this review. Hematoxylin and eosin staining demonstrates histologic features of the urothelium at each site, ×40 magnification. Renal urothelium is 1–2 cell layers thick (black arrow). Ureteral and bladder urothelium are 3–5 cells layers thick. Basal cells are indicated with (red arrow), intermediate cells (blue arrow), and superficial cells (green arrows). The urothelial basement membrane is indicated by a yellow dashed line.

Urothelium structure and function

Urothelium is commonly composed of three distinct cell layers (basal, intermediate and superficial) which are discerned based on their morphology and combinatorial expression of protein markers (Figure 2, Table 1). For sake of simplicity, we will describe the structure of mature bladder urothelium, and indicate tissue specific differences as necessary. Basal cells are small, mononuclear and attached to the basement membrane [2]. Basal cells express Keratin (KRT)5, Tumor Protein p63 (TP63), Sonic Hedgehog (SHH), CD44, Cd49f, and β4 Integrin [3–8]. A subset of basal cells express KRT14 [9], which identifies progenitors in other post-natal stratified epithelia such as skin [10] and respiratory epithelium [11]. Intermediate cells comprise the layers formed between basal and superficial cells and exist in either mononucleated or binucleated forms [12]. Intermediate cells may express KRT5, TP63, SHH, or Cd49f or Uroplakins (UPKs) [7, 8, 13–15]. The two major subtypes of intermediate cells are those that express KRT5 (K5-interemediate cells) or UPKs (UPK-intermediate cells); for simplicity’s sake, we will refer to all subtypes as intermediate cells in this review. Superficial cells are large, polyploid apical cells that face the urinary space and are the most differentiated urothelial cell type. They express KRT20 along with UPKs [7, 16, 17].

Figure 2.

Diagram shows organization and marker expression patterns in bladder urothelium. Bladder urothelium consists of a basal, one or more intermediate, and a superficial (umbrella) cell layer. Markers known to be expressed by cells within each layer are depicted on the right.

Table 1.

Urothelial Markers

Maker	ID	Expression	Function
Cytokeratin 5	KRT5/CK5	Basal & Intermediate	• Type II keratin • Dimerizes with Krt14 to form intermediate filament that make up cytoskeleton • Anchors epithelial cells through desmosomes • Anchors urothelium to basal lamina through α6β4 integrin at hemidesmosome
Cytokeratin 14	KRT14/CK14	Subset of Basal	• Type I keratin • Dimerizes with Krt5 to form intermediate filament that make up cytoskeleton • Anchors epithelial cells through desmosomes • Anchors urothelium to basal lamina through α6β4 integrin at hemidesmosome
Tumor Protein 63	TP63	Basal & Intermediate	• Member of p53 family of transcription factors • Maintenance of basal and intermediate cell
Sonic Hedgehog	SHH	Basal & Intermediate	• Secreted protein that serves as a morphogen involved in urothelial patterning
Cluster of differentiation marker 44	CD44	Basal	• Cell-surface glycoprotein • Mediates cell adhesion, cell migration and cell to cell interactions
Integrin β4	ITGB4/ CD104	Basal	• Forms integrin beta chain and complexes with ITGA6 to form α6β4 integrin • Integral cell surface protein Mediates cell-matrix and cell-cell adhesion, and signal transduction
Integrin α6	ITGA6/CD49f	Basal	• Forms integrin alpha chain and complexes with ITGB4 to form α6β4 integrin • Integral cell surface protein • Mediates cell-matrix and cell-cell adhesion, and signal transduction
Uroplakins	UPK1a UPK1b UPK2 UPK3a UPK3b	Intermediate & Superficial	• Transmembrane proteins that form urothelial plaque in superficial cells • Stabilizes apical membrane through urothelial plaque and cytoskeleton interactions • Mediates Signal Transduction
Cytokeratin 20	KRT20/CK20	Superficial	• Type I keratin • Forms intermediate filaments that make up cytoskeleton scaffold of differentiated cells

Ureteral and renal pelvic urothelium are reminiscent of the bladder urothelium, described above [18]. The intrarenal urothelium exhibits significant regional heterogeneity based on the renal compartment that it lines: cortical, outer medulla and inner medulla [18, 19]. Notably, superficial cells in the renal urothelium lack morphologic features of umbrella cells, and their expression of KRT20 is controversial [18, 19]. Conflicting marker expression patterns are reported. These discrepancies may be due to the lack of correlation of mRNA and protein expression, varying detection methods, or transitional nature of urothelium. Therefore, we recommend that investigators report marker expression and positional characteristics of the cell (i.e. basal, intermediate, superficial) for clarification.

A major function of the urothelium is to form a urine permeability barrier. Superficial cells confer barrier function via urothelial plaques (formed from UPKs) and tight junctions [2, 12, 20–22]. Urothelium also serves roles in signal transduction, host response to infection, barrier function, water and solute transport, and compliance [20, 23–29].

Definition of progenitors and methods used for their identification

Progenitor cells have the capacity to proliferate and differentiate. In the relatively quiescent postnatal urothelium, proliferation rates are very low. While other stratified epithelium such as cells that line the gut and epidermis have turnover rates of 1–30 days [30, 31], urothelium has turnover rates of 3–6 months or more [2, 32]. However, in response to damage, urothelium shows significant proliferative and regenerative capacity [7, 33–35]. Interestingly, almost all postnatal urothelial cell types have at least some proliferative capacity that may depend on the type and/or degree of injury.

The search for the cell(s) responsible for urothelial patterning and tissue repair has led to the identification of several context-specific candidate progenitor cells. The literature contains inconsistencies in terminology used to describe these candidates, with some reporting identification of urothelial ‘stem cells’ and others reporting identification of ‘progenitor cells’. For the purposes of this review, we will use the term ‘progenitor’ for urothelial cells that may proliferate and differentiate during normal development and tissue repair.

Cre-LoxP based fate-mapping

The Cre-LoxP recombination strategy is the most widely used tool to investigate progenitor–progeny relationships in vivo [36, 37]. This genetic system relies on cell-specific expression of Cre recombinase to permanently label a population of cells and their progeny. Incorporation of a tamoxifen-inducible Cre recombinase further specifies lineage analysis to a precise developmental or postnatal stage. This technology can also be used ex vivo, supporting the investigation of progenitor–progeny relationships in explanted tissues or organoids. Cre-LoxP based fate mapping strategies have identified many context-dependent progenitor–progeny relationships in urothelium (Table 2). Urothelial markers provide a way to label cell populations. The function and interactions of these markers is outlined in Table 1, but it is not clear whether a common signal transduction pathway regulates these markers.

Table 2.

Urothelial Progenitors Identified by Cre;LoxP Strategies

Mouse Line	Lineage	Development/Homeostasis	Repair
ΔNP63 Cre	TP63	• Gives rise to all urothelial cell types in bladder (Pignon 2013)
Pax2 Cre	PAX2	• Gives rise to all urothelial cell types in ureter (Bohnenpoll 2017)
Shh Cre	SHH	• Gives rise to all urothelial cell types in ureter (Bohnenpoll 2017)
Shh CreERT2	SHH	• E11 Shh-lineage gives rise to all urothelial cell types and 50% of Superficial cells in E18 bladder (Gandhi 2013) • E14 Shh-lineage gives rise to 6% of Superficial cells in E18 bladder (Gandhi 2013) • Adult Shh-lineage gives rise to all urothelial cells in bladder (Shin 2011)	• Adult Shh-lineage gives rise to all urothelial cells in the bladder following chronic bacterial injury (Shin 2011)
Foxa2 CreERT2	FOXA2	• E11 Foxa2-lineage gives rise to Intermediate and Superficial cells in E18 bladder (Gandhi 2013)
Upk3a CGE	UPK	• E11 Upk3a-lineage gives rise to Intermediate and Superficial cells in E18 bladder (Gandhi 2013) • E14, E16, E18 Upk3a-lineages give rise to Intermediate and Superficial cells in ureter explants (Bohnenpoll 2017) • Adult Upk3a-lineage gives rise to Basal, Intermediate and Superficial cells in ureter (Bohnenpoll 2017)	• Adult Upk3a-lineage gives rise to Superficial cells in bladder following a single round of cyclophosphamide injury (Gandhi 2013, Wang 2018)
Upk2 CreERT2	UPK2		• Adult Upk2-lineage gives rise to Intermediate and Superficial cells in bladder following focal mucosal injury (Schafer 2017)
Krt5 CreERT2	KRT5	• E16, P1, and P7 Krt5-lineage gives rise to adult UPK+ UCs in the adult kidney (Jackson 2019)	• P14 Krt5-linaege gives rise to UPK+ UCs in the kidney during UTO-induced remodeling (Jackson 2019) • Adult Krt5-lineage gives rise to all urothelial cells in the bladder following one round of cyclophosphamide injury (Papafotiou 2016) • Adult Krt5-lineage gives rise to Intermediate and Superficial Cells in the bladder following multiple rounds of cyclophosphamide injury (Wang 2018) • Adult Krt5-lineage gives rise to all urothelial cells in the bladder following augmentation cystoplasty (Schafer 2017)
Krt14 CreERT2	KRT14	• Adult Krt14-lineage gives rise to all urothelial cell types in bladder (Papafotiou 2016)	• E16 Krt14-lineage give rise to all urothelial cells in the bladder following a single round of adult cyclophosphamide injury (Papafotiou 2016) •Adult Krt14-lineage gives rise to Intermediate and a small fraction of Superficial cells following a single round of cyclophosphamide, and larger fraction of Superficial cells in the bladder following multiple rounds of cyclophosphamide injury (Papafotiou 2016)

Label retaining cell assays

Label-retaining cell (LRC) assays have been employed to identify potential progenitor cell populations and are especially useful when lineage markers are unavailable. Also known as pulse-chase assays, these strategies rely on the ability to chase a labeled population, following the administration (or pulse) of a nucleoside analog to mark nuclei of mitotic cells. Moreover, the long-term retention of analogs such as 5-bromo-2’-deoxyuridine (BrdU, or 5-ethynyl-2-deoxyuridine [EdU]) is a characteristic observed in stem cells. In adults, stem cells are believed to be slow cycling, while transit amplifying (TA) cells (cells in transition from stem cells to differentiated cells) divide more frequently. Thus, retention of BrdU (or its equivalent) may identify a stem cell, while TA cells may exhibit little or undetectable BrdU [4].

In vitro studies of urothelial progenitors

Organoid forming assays have emerged as a complementary technique to investigate progenitor–progeny relationships in vitro and can also be paired with Cre-LoxP or LRC methodology [8, 9, 13, 38]. This assay relies on the capacity of primary urothelial cells to clonally expand to form organoids (spheres) when cultured in a 3D environment (that can include basement membrane extract/matrix). Phenotyping of the resultant organoids determines whether a putative progenitor can self-renew, self-organize and form differentiated cell types.

The following sections will review literature implementing these techniques (among others) to uncover urothelial progenitor–progeny relationships, highlighting the importance of the contexts through which these discoveries were made.

Primordial urothelial markers

Using a p63Cre line bred with a reporter R26R line (ΔNp63^Cre;Rosa26^EYFP), Pignon et al. demonstrated that all bladder urothelial cell types form from the p63-lineage [14]. In the ureter, Pax2^Cre;R26^mT/mG and Shh^Cre;R26^mT/mG (constitutive Cre) mouse lines showed that the SHH- and PAX2-lineages gave rise to all ureteral urothelium cell types [39]. While these studies suggest important progenitor–progeny relationships, the lack of an inducible Cre-recombinase limited the ability to identify urothelial progenitor populations.

Transient cell progenitor model during development

Elegant use of combinatorial marker expression led the Mendelsohn group to identify a novel urothelial cell type – present in the bladder urothelium from E11-E13 – which they termed “P” cells (FOXA2⁺/P63⁺/UPK⁺/KRT5⁻) [6]. Using a tamoxifen inducible Foxa2^CreERT2;R26^mTmG line, the group showed that P cells gave rise to intermediate and superficial cells, but not KRT5⁺ cells [6]. Similar findings were observed with the UPK3a-lineage, which marks a transient population of embryonic progenitor “P” cells, but also intermediate cells [6].

SHH lineage

Using a tamoxifen-inducible Shh-Cre line bred with the R26^mTmG reporter strain, the Mendelsohn group mapped stage-specific derivatives of the SHH-lineage [6]. Tamoxifen exposure at embryonic day (E)11 revealed that 70% of E18 urothelial cells, including nearly half of the superficial cells, derived from the E11 SHH-lineage [6]. In a separate experiment, only 25% of E18 urothelial cells, and only 6% of superficial cells were derived from the E14 SHH-lineage [6]. However, since SHH is expressed by four different urothelial cell types in the developing bladder (undifferentiated endodermal cells, “P”-cells, intermediate cells & Krt5⁻BCs), the results of those studies could not identify the specific embryonic progenitor subset within the E11 population [6]. Similar to the developing bladder, SHH is expressed across the basal and/or intermediate cell layers in the adult bladder [6, 12, 13]. Using adult induction of lineage labeling, Shin et al. demonstrated that the SHH lineage had the capacity to give rise to other cell types over a 10 month evaluation period, supporting a role for SHH-expressing cells as homeostatic progenitors. This is likely true only during an extended chase due to the fact that the urothelium has a remarkably long turnover rate [13]. Using uropathogenic bacterial infection to trigger a urothelial injury response, Shin et al. also investigated whether SHH-expressing urothelial cells were responsible for tissue repair following UTI [13]. Using the tamoxifen-inducible Shh^CreERT2;R26^mTmG line, the SHH-lineage was found to form all urothelial cell layers, including luminal UPK3A⁺ cells after 3 or 10 rounds of bacterial injury [13]. While, these data support a progenitor role for SHH-expressing cells during tissue repair, the exact SHH-expressing progenitor cell population remains unclear.

UPK lineage

The Mendelsohn and Kispert labs implicated UPK3A-expressing intermediate cells as developmental and tissue repair progenitors [6, 39]. Using the Upk3a^GCE;R26^mCherry line, the Mendelsohn group found that the E11 UPK3A-lineage gave rise to a fraction of E18 intermediate and superficial cells, but not KRT5-expressing basal cells [6]. Furthermore, they showed that in adult bladders, UPK3A⁺ cells give rise to superficial cells after one dose of cyclophosphamide. Using this same line, the Kispert group reported that E14, E16, and E18 UPK3A-lineages gave rise to intermediate and superficial cells in embryonic ureteral explant models and in adult ureters during homeostasis, and also reported that the UPK3A-lineage gave rise to basal cells [39].

KRT5-lineage

In the developing bladder, KRT5-expressing cells are rare prior to E15, but Krt5 mRNA expression is reported at E12. Tamoxifen induction at E14 in Krt5^CreERT2;R26^mTmG mice showed that the KRT5-lineage rarely gave rise to other cell types in the embryonic urothelium [6]. During injury repair, the KRT5-lineage failed to give rise to intermediate or superficial cell derivatives after a single round of cyclophosphamide [6]. However, after repeated cyclophosphamide injury, the KRT5-lineage served as a progenitor for intermediate and superficial cells, likely due to the depletion of the more lineage restricted UPK⁺ progenitors [12]. Further evidence of context-dependent progenitor models is demonstrated through lineage tracing studies in surgical injury paradigms [40]. KRT5-expressing cells regenerate all lineages following augmentation cystoplasty, whereas focal mucosal defects engage basal cells and UPK2-expressing intermediate cells which self-renew and form superficial cells [40]. Taken together, these studies suggest that KRT5⁺ cells have the potential to proliferate after many types of injury to give rise to UPK3⁺ intermediate and superficial cells in the bladder. Furthermore, it appears that the more severe the injury the more likely that KRT5⁺ cells act as the progenitor.

Similar to findings in the developing bladder, Bohnenpoll et al. found that the KRT5-lineage did not give rise to intermediate or superficial cells, but was restricted to the basal cell layer in ureteral explants and in adult ureters during homeostasis [39]. Interestingly, the KRT5-lineage was found to exhibit temporally restricted progenitor capacity in the intrarenal urothelium [19]. Jackson et al. demonstrated that 72% of E16 KRT5-lineage formed adult UPK3⁺ renal urothelial cells [19]. Furthermore, 50% of Postnatal (P)1, 40% of P7, 16% of P14, 7% of the P21 KRT5-lineage formed adult UPK3 cells.

Having identified that the KRT5-lineage formed UPK⁺ renal urothelial cells in a temporally restricted manner, Jackson et al. used unilateral ureteral obstruction (UUO) to evaluate the KRT5-lineage during urothelial repair. Similar to what was observed in the bladder, KRT5-lineage during acute cyclophosphamide injury [6], UUO failed to utilize the adult KRT5-lineage for repair [19]. However, the P14 KRT5-lineage breaks lineage restriction to repair UUO-induced renal urothelial injury [19]. Altogether, these reports demonstrate the importance of varying the timing of tamoxifen administration to discern the contributions of precisely staged urothelial populations on urothelial cell formation/repair. Jackson et al. demonstrated that the adult KRT5-lineage did not contribute to tissue repair – presenting the possibility that adult UPK⁺ renal urothelial cells are responsible for UUO-induced renal urothelial injury repair (as has been shown in the bladder). It is also intriguing to consider that the adult bladder UPK⁺ intermediate cell that replaces superficial cells during cyclophosphamide-induced injury repair may arise from a KRT5-lineage. Given the exceptionally long turnover of urothelial cells, a thorough understanding of the fate of KRT5⁺ cells from numerous developmental stages may resolve these discrepancies.

KRT14-lineage

In 2016, Papafotiou et al. identified a KRT14-expressing subset within the KRT5⁺ basal urothelial cell layer with a remarkable capacity to self-renew and form each urothelial cell type [9]. KRT14-expressing cells were three-times more likely to proliferate one day after cyclophosphamide-induced injury than Krt14-negative cells [9]. In a novel tamoxifen-inducible Krt14^CreERT2;R26^tdTomato line, the KRT14-lineage was found to give rise to all urothelial cell types [9]. This rare basal cell subset was found to support ex vivo expansion in bladder urothelial explants and outgrowths. Subsequent evaluation of KRT14⁺ organoids revealed that the KRT14-lineage exhibited a superior clonogenic (self-renewal) capacity compared to KRT14-negative urothelial cells [9]. Genetic ablation of KRT14⁺ cells abolished explant growth and blocked bladder urothelial proliferation after injury [9]. It remains to be determined whether the Krt14-lineage may account for the progenitor capacity attributed to the Krt5-lineage.

Later studies using various injury models showed that KRT14⁺ cells are the primary cell type that proliferates after injury. In 2017, the Apodaca lab showed that the majority of proliferating urothelial cells one day after spinal cord transection (which leads to chronic urothelial injury from neurogenic bladder) were KRT14⁺ [7]. The Bates lab recently showed that the majority of proliferating bladder urothelial cells from 1–28 days after a single dose of cyclophosphamide (which caused injury to virtually all urothelial cell layers) were KRT14⁺; fewer proliferating cells were KRT14⁻/KRT5⁺, while almost none were UPK3⁺ cells at all timepoints [41]. The group also observed that pretreatment with fibroblast growth factor (FGF)7 led to significant bladder urothelial cytoprotection and a major reduction in KRT14⁺ cell proliferation, suggesting that KRT14⁺ cells are the major cell type leading regeneration after serious injury. KRT14⁺ cells have been shown by multiple groups with potential to give rise to all bladder urothelial cell layers after injury.

CD49f-lineage

CD49f (also known as ITGA6) is enriched in basal cells of the bladder urothelium, and when gated on the basis of CD44 expression (another stem cell marker), the CD49f^high/CD44^high cell population displayed the greatest organoid forming capacity of all urothelial cells in vitro [8]. Importantly, when seeded at single cell density, CD49f^high cells maintained organoid forming capacity, attesting to the superior stemness of this population in vitro. In a 3D in vitro culture system, CD49f^high basal cells were found to clonally expand and give rise to organoids that could differentiate and form UPK-expressing luminal cells [8]. It is unclear whether CD49f^high cells with heightened progenitor capacity represent the KRT14⁺ basal subset or a separate population.

Progenitor cells during urothelial homeostasis and repair

In a one year chase of BrdU-labeled rat bladders, Kurzrock et al. identified β4 integrin-expressing basal cells as LRCs. In vitro analysis confirmed that LRCs demonstrated superior clonogenic (ability to form large colonies) and proliferative capacities compared to unlabeled cells up to 9 months after labeling [4]. In a separate LRC assay, Colopy et al. identified that KRT5⁺ basal cells and intermediate cells were candidate progenitor populations in the mouse bladder. Indeed, BrdU marked basal and intermediate cells layers and those cells were found to give rise to all layers of the adult bladder urothelium during homeostasis and in response to uropathogenic Escherichia coli (UPEC)-induced injury [42]. While the timing of BrdU administration, length of chase and species differed in the aforementioned studies, it is interesting to consider whether the β4 integrin-expressing basal LRC is responsible for tissue restoration.

LRC assays using EdU-labeled neonatal rats failed to correlate LRCs with known epithelial stem cell markers; however, much shorter chase intervals, lack of clonogenic evaluation or absence of injury paradigm limited interpretation of the data [43]. Further evidence of a basal cell progenitor comes from studies in pigs that show a basal cell subset with superior clonal capacities (abilities to self-renew and differentiate in vitro) [44], and from human cystectomy tissue which shows clonal patches of urothelium units (evidenced by naturally occurring somatic mutations in mitochondrial DNA) [45]. Evidence supporting tissue restoration by TA cells is found in response to umbrella cell exfoliation studies in mice, where protamine sulfate-induced exfoliation engages UPK-expressing intermediate cells for repair [46].

Signaling pathways during urothelial development

Signals derived from the mesenchyme of the embryonic bladder and ureter govern urothelial proliferation and differentiation. While the impact of this niche on urothelial development is well known, it is less clear whether these signals target a precise progenitor population. Future efforts to monitor clonal subsets of putative progenitor populations using limited labeling techniques in vivo, or in vitro studies using pure populations may serve to close this knowledge gap. A summary of the signaling pathways involved with urothelial development (discussed below) are shown in Figure 3.

Figure 3.

Diagram showing known signaling pathways that regulate urothelial cell development. FGF7, secreted by mesenchymal cells binds to its urothelial receptor, FGFR2 in urothelium, leading to intermediate cell development. RA from the stroma, binds to its urothelial retinoic acid receptors (RARs) leading to proper intermediate and superficial cell differentiation. PPARG drives expression of FOXA1 and GATA3, both of which cooperate with PPARG to drive superficial cell differentiation. SHH secreted by urothelium activates FOXF1 expression in mesenchyme that drives BMP4 expression; BMP4 is then secreted back into urothelium where it binds to BMP receptors (BMPRs) that activate AKT; all of these signaling factors are critical for urothelial cell differentiation. NOTCH activity in urothelium, including activation of ITGA4, is required for urothelial cell integrity. The abbreviations not defined here are delineated in the main text. Citations for these findings are also found in the main text.

FGF signaling was one of the first pathways identified as having roles in bladder urothelial development. Mesenchymal FGF7 was found to promote stratification of bladder urothelium during embryogenesis. Furthermore, loss of FGF7 results in absence of bladder intermediate cells [47].

Retinoic acid (RA) signaling is critical for embryonic urothelial differentiation (and adult urothelial regeneration after injury) [6]. The Mendelsohn group showed that mesenchyme-derived RA was critical for embryonic P cell differentiation. Mice expressing an inducible, conditional dominant negative RA receptor had ineffective P cell differentiation into UPK3⁺ intermediate and superficial cells. This same strategy implicated RA signaling in differentiation of superficial cells from UPK⁺ intermediate cells following cyclophosphamide-induced injury. Thus, RA signaling appears to have a vital role in bladder urothelial embryonic differentiation and postnatal regeneration after injury.

Given that RA signaling was found to regulate bladder urothelial cell differentiation, the Kispert group examined the role of RA signaling in embryonic ureter explants [48]. A pharmacologic blockade of RA signaling was found to induce differentiation of intermediate cells into superficial cells, while ectopic RA signaling reduced the formation of both basal and superficial cells [48]. These findings suggest that RA signaling is required to prevent premature differentiation of intermediate cells into superficial and basal cells in the ureter.

In addition to RA acid and its receptor, the transcription factor, Peroxisome proliferator-activated receptor gamma (PPARG) also plays a prominent role in urothelial differentiation. Studies have shown that PPARG appears to drive urothelial superficial cell differentiation in vitro [49] and in vivo [49]. Also, the Mendelsohn group recently showed that inducible and conditional loss of Pparg in bladder urothelial cells leads to a failure of superficial cells to properly differentiate by E16 [50]. Furthermore, others showed that in human urothelium, PPARG drives expression of transcription factors, FOXA1 and GATA3, all of which appear critical for urothelial cell differentiation [51]. Finally, a study in human cancer cell lines with basal cell features shows that forced expression of PPARG, FOXA1 and GATA3 cooperate to drive a lumenal urothelial phenotype [52]. Thus, the PPARG/FOXA1/GATA3 network appears to be critical for superficial cell development in urothelium.

During urinary tract morphogenesis, expression of bone morphogenetic protein (BMP) 4 by tailbud mesenchyme serves a critical role in subdivision of the ureteric bud into the intra-renal collecting system and ureter. Ectopic expression of BMP4 drives urothelial fate within renal collecting ducts, whereas deletion of Bmp4 abrogates urothelial differentiation [53]. BMP4/BMP5 also appear to play a key role in development and/or maintenance of superficial cells, as urothelial SHH triggers stromal BMP4/BMP5 expression that drives urothelial differentiation [54]. Subsequent studies found that AKT functions downstream of BMP4 to promote both proliferation and differentiation of ureteral urothelium [55].

The SHH–FOXF1–BMP4–AKT signaling axis supports proliferation and differentiation of the ureteral urothelium. Inactivation of the SHH signaling transducer, SMO, in the ureteral mesenchyme pointed to a role for SHH signaling in ureteral patterning. Ureteral urothelium of Tbx18^Cre;Smo^fl/fl mice was less proliferative, failed to form UPK1B-expressing umbrella cells, and had fewer KRT5 and p63 basal and intermediate cells. On the other hand, constitutive activation of SHH activity in the mesenchymal compartment led to increased urothelial proliferation. Foxf1 is a SHH signaling target expressed by the ureteral inner mesenchyme. Explanted ureters from mice harboring a dominant negative form of Foxf1 lack Bmp4 and thus fail to undergo urothelial differentiation, placing BMP4 downstream of FOXF1 in this pathway. This signaling axis demonstrates the importance of bi-directional communication in regulating urothelial proliferation and differentiation during development, but will also be important for tissue repair.

In addition to mesenchymal–epithelial crosstalk, recent studies demonstrate roles for the NOTCH signaling pathway in bladder urothelium [8, 56]. Using RNA-seq to compare proliferative (P-) and differentiated (D-) bladder urothelium derived organoids, Santos et al. found that the NOTCH signaling pathway was associated with D-organoids [8]. Inactivation of the NOTCH signaling pathway in D-organoids, using γ-secretase inhibitors, prevented lumen formation and led to a reduction in the expression of luminal markers including uroplakins and KRT20 [8]. Remarkably, NOTCH inhibition shifted D-organoids toward a P-organoid phenotype. In a separate study, Paraskevopoulou et al. uncovered a role for NOTCH signaling in maintaining bladder urothelium integrity and linked their findings to humans with interstitial cystitis [56]. Conditional inactivation of NOTCH signaling in basal or superficial urothelial cells led to bladder urothelial hyperplasia, detachment of KRT20⁺ superficial cells, barrier dysfunction, inflammation and downregulation of genes with roles in cell–cell and cell–extracellular matrix interaction, which was ultimately rescued by exogenous reconstitution of NOTCH [56]. Given that NOTCH inactivation led to a reduction in Integrin expression, Paraskevopoulou et al. elegantly demonstrated a role for Integrin signaling in bladder urothelium [56]. Itga4 knockdown reduced the clonogenic potential of bladder urothelial cells in vitro, while the use of cognate ligands to block ITGA4 reduced urothelial proliferation during cyclophosphamide injury [56]. Finally, constitutive exogenous expression of Itga4 increased urothelial outgrowth in Notch-deficient bladder urothelium explants altogether demonstrating a requirement for the Integrin signaling pathway in urothelial proliferation [56].

Signaling pathways during urothelial repair

Several signaling pathways regulate the bladder urothelium injury response. While evidence supports context-specific progenitors, the signaling pathways driving bladder urothelial tissue repair may also be context-dependent. A summary of the signaling pathways involved in bladder urothelial repair after injury (discussed below) are shown in Figure 4.

Figure 4.

Diagram showing known signaling pathways that regulate bladder urothelial repair after injury. Following UPEC infection, BMP4 (through its BMP receptor 1A), SHH and WNT ligands all lead to KRT5⁺ urothelial cell proliferation. Following cyclophosphamide injury, WNT ligands drive KRT14⁺ cell proliferation. Loss of endogenous (Endog) FGFR2IIIB (receptor for FGF7 and FGF10) leads to pathological endoreplication of KRT14⁺ cells with maladaptive repair after cyclophosphamide. Pretreatment with exogenous (Exog) FGF7 (acting through FGFR2IIIB) leads to KRT5⁺ cell proliferation and UPK3⁺, KRT5⁺, and KRT14⁺ cell cytoprotection from cyclophosphamide.

BMP4 signaling supports KRT5⁺ bladder urothelial cell proliferation and tissue repair following UPEC infection [46]. Knockdown of the BMP4 receptor Bmpr1a, significantly diminished KRT5⁺ cell proliferation after infection. In a protamine sulfate tissue injury model where intermediate cells proliferate in response to injury, restoration of the urothelium occurs independent of BMP4 signaling. Thus, BMP4 signaling appears to be required specifically for KRT5⁺ cells to proliferate after injury. It remains to be determined whether BMP4 signaling governs proliferation of other basal cell types, such as the KRT14 subset.

The SHH and Wingless-related integration site (WNT) signaling pathways also support bladder urothelial proliferation and restoration. Shin et al. demonstrated that SHH and WNT signaling were critical for KRT5⁺ cell proliferation following UPEC injury [13]. Both genetic knockdown of Gli1, a transcription factor that mediates Hedgehog signaling, and administration of SHH neutralizing antibodies, significantly reduced SHH⁺ bladder urothelial cell proliferation after infection. Hedgehog signaling blockade (to suppress WNT signaling) or systemic administration of a WNT inhibitor reduced proliferation after bacterial injury. Papafotiou et al. demonstrated the importance of WNT signaling in KRT14⁺ cell proliferation following cyclophosphamide-induced bladder urothelial injury [9]. Thus, both SHH and WNT signaling appear to have (possibly interconnected) roles in basal urothelial cell proliferation following injury.

The FGF7/FGF10/FGFR2IIIB signaling axis also regulates proliferation in adult urothelium during homeostasis and in response to injury [47, 57–59]. Administration of FGF7 or FGF10 to rodents or monkeys induced proliferation in basal urothelial layers in the absence of injury [57]. The Bates lab found that pretreatment with recombinant human FGF7 led to cytoprotection in UPK3⁺/KRT5⁺/KRT14⁺ urothelial cells following cyclophosphamide exposure, accompanied by proliferation of KRT14⁻/KRT5⁺ basal and intermediate cells [41]. In the renal urothelium, FGFR2 activation is critical for FGF-mediated proliferation following unilateral ureteral obstruction and renal ischemia reperfusion injury [58, 59]. Pre-treatment with an FGFR2 antisense probe, or vitamin A (a MAP-kinase dependent FGFR2 pathway inhibitor), significantly reduced renal urothelial proliferation during unilateral ureteral obstruction [58].

Using an inducible, conditional deletion of Fgfr2 in adult urothelium, the Bates group has shown that the FGF7:FGFR2IIIB signaling axis does not regulate KRT14 proliferation during cyclophosphamide-induced urothelial injury [60]. Rather than prevent KRT14 proliferation, urothelial Fgfr2-deficient bladder urothelium exhibited increased cell cycle activity with pathological endoreplication/inappropriate polyploidy of KRT14⁺/KRT5⁺ basal cells. Even after a 6-month recovery period, FGFR2-deficient urothelium failed to form superficial cells and contained many ectopic foci of apical KRT14⁺ cells. While these findings show that FGF7:FGFR2IIIB signaling does not drive proliferation of KRT14⁺ cells, this signaling pathway does regulate KRT14⁺ cell cycle activity (i.e. prevents endoreplication) and has critical roles in urothelial restoration/repair.

Conclusions

This review has discussed the evidence in support of urothelial progenitor cells generated from lineage analysis (and other approaches) conducted in bladder, ureter and renal urothelium, along with the signaling pathways that govern their fate. Given differences in the methodologies employed in each study of urothelial progenitors, these observations should be regarded as stand-alone findings even when conducted in the same tissue. The timing of labeling, analysis, and differences in Cre-drivers and reporters confound the ability to compare or contrast findings. Moreover, up to this point, Cre-LoxP recombination strategies have been limited to evaluation of urothelial progenitors in mice. However, Crispr-Cas9 based strategies are enabling Cre-LoxP lineage tracing applications in other species and model systems, including human immortalized pluripotent stem cells, and as such, these approaches will likely pave the way for identification and further evaluation of human urothelial progenitors. The use of urothelial organoids provides a great opportunity to compare the properties of candidate progenitor populations, as well as the capacity to evaluate the impact of pharmacologic agents and signaling pathways on directing progenitor fate. Ultimately, urothelial progenitors may have applications in regenerative and tissue engineering approaches in benign urologic disorders.