The Murine Bladder Supports a Population of Stromal Sca-1+/CD34+/lin- Mesenchymal Stem Cells

Bladder fibrosis is an undesired end point of injury of obstruction and often renders the smooth muscle layer noncompliant. In many cases, the long-term effect of bladder fibrosis is renal failure. Despite our understanding of the progression of this disease, little is known about the cellular mechanisms that lead to a remodeled bladder wall. Resident stem (progenitor) cells have been identified in various organs such as the brain, heart and lung. These cells function normally during organ homeostasis, but become dysregulated after organ injury. Here, we aimed to characterize a mesenchymal progenitor cell population as a first step in understanding its role in bladder fibrosis. Using fluorescence activated cell sorting (FACS), we identified a Sca-1+/ CD34+/ lin- (PECAM-: CD45-: Ter119-) population in the adult murine bladder. These cells were localized to the stromal layer of the adult bladder and appeared by postnatal day 1. Cultured Sca-1+/ CD34+/ lin- bladder cells self-renewed, formed colonies and spontaneously differentiated into cells expressing smooth muscle genes. These cells differentiated into other mesenchymal lineages (chondrocytes, adipocytes and osteocytes) upon culture in induction medium. Both acute and partial obstruction of the bladder reduced expression of CD34 and changed localization of Sca-1 to the urothelium. Partial obstruction resulted in upregulation of fibrosis genes within the Sca-1+/CD34+/lin- population. Our data indicate a resident, mesenchymal stem cell population in the bladder that is altered by bladder obstruction. These findings provide new information about the cellular changes in the bladder that may be associated with bladder fibrosis.


Introduction
Pediatric patients with posterior urethral valves, neurogenic bladder from spina bifida, and other obstructive uropathies are at high risk for kidney injury [1]. In fact, obstructive uropathy is the cause of renal transplantation in up to a quarter of pediatric patients with end stage renal disease [2]. This occurs because after obstructive injury the bladder wall remodels to become either thinner (atonic bladder) or thicker (fibrotic bladder). In either case, urine is not properly voided resulting in retrograde pressure on the ureters and kidney injury. These types of remodeling are reminiscent of other hypertrophic or fibrotic remodeling in injured smooth muscle invested organs. For example, if vascular pressure increases in pulmonary arteries before the typical change in fetal-to-newborn circulation, the vessel walls become abnormally hypertrophied causing pulmonary hypertension [3].
Currently little is understood about which cells proliferate within or migrate to the bladder after obstruction. Studies have demonstrated that bone marrow derived cells are recruited to the mouse bladder after injury [4,5]. However, this is a modest response and it does not appear to lead to robust proliferation of these cells in the bladder. Thus, in cases where bladder injury leads to cell proliferation [6,7], cells intrinsic to the bladder appear to be the primary source of new cell growth after obstruction.
The bladder consists of four layers: a superficial mesothelial layer called the serosa; a smooth muscle layer; a connective tissue layer called the stroma; and an inner layer called the urothelium. Bladder intrinisic progenitors have been identified in the mouse urothelium [6]. Crebased lineage analysis of these cells after bladder infection revealed that urothelial progenitors proliferated in response to injury but their descendants remained in the urothelium. Urothelial progenitors thus appear committed to the urothelial fates, and do not contribute to the stroma, smooth muscle or serosa after injury.
Resident mesenchymal stem cells (MSCs) have been identified in other muscular organs such as in heart [7], lung [8,9], skeletal muscle [10], and vascular smooth muscle [11][12][13][14]. These progenitor cell populations have been shown to play an active role in the deterioration of muscle function in coronary artery vein grafts as well in the restoration of function during muscle remodeling and regeneration [15,16]. Cardiac progenitor cells (CPCs) have been identified in human pediatric patients with congenital heart disease and found to be at higher levels in younger patients. Furthermore, human CPCs were shown to promote a greater degree of wound healing than cardiomyocytes in a cardiac ischemia model [17]. Similar to other muscular organs, the bladder is a plastic organ, demonstrating significant hypertrophic tissue remodeling in response to stimuli such as cystectomy, obstruction or denervation. However, the role of resident MSCs in bladder tissue remodeling has never been described.
Here we set out to determine if MSCs exist in the adult mouse bladder. We focused on cells that express the genes stem cell antigen-1 (Sca-1) and CD34. Co-expression of these genes has been identified as a potential marker for murine progenitor cells in smooth, and skeletal muscle organs. For example, in mouse arterial adventitia, a population of Sca1 + /CD34 + progenitors were identified that spontaneously differentiated into smooth muscle cell in culture [14,18]. Similarly in mouse lung, Sca1 + /CD34 + progenitors were detected that play a role in distal lung homeostasis [8]. Here we show that cells within the stromal layer of the mouse bladder coexpress Sca-1 and CD34 and these cells can be isolated by FACS. FACS sorted bladder Sca-1 + / CD34 + /lincells form colonies in culture and these colonies spontaneously differentiate into cells that express smooth muscle specific genes. We argue that these cells could be intrinsic MSCs that could play a role in bladder remodeling after obstruction.

Experimental Animals
The Institutional Animal Care and Use Committee of the Stanley Manne Children's Research Institute approved all animal studies. For some studies we used CD1 mice (Charles River). In other cases bladders were removed from Rosa mTmG in a mixed B129SV/J/C57BL6 background. This Cre reporter strain was used so that the membrane of each cell was labeled with red fluorescence. Sca-1 egfp (stock # 012643) [19] and Gli1Cre ERT2 (stock # 007913) [20] mice were obtained from Jackson Laboratories (Bar Harbor, Maine).

Acute Bladder Outlet Obstruction Surgery (aBOO)
Adult CD1 mice were used in acute obstruction surgeries. A Matrix Medical Inc. Spartan VMC anesthesia unit was used to induce anesthesia. Mice were placed in the induction chamber, which was filled with 3% vaporized isoflurane (Baxter) at a flow rate of 1.5 L/min. Mice were then transferred to the operating table and kept under anesthesia using a rodent anesthesia mask to deliver a continuous flow of isoflurane. To create acute outlet obstruction, a ligature was tied down over a 22G angiocatheter placed in the urethra. Following removal of the catheter, the urethra was fully occluded. For sham surgeries, an incision was made in the abdomen, and then sutured. Post-operatively, mice were placed in a recovery cage that was placed on a heating pad until normal ambulation was observed. Recovery time was typically one hour or less. For analgesic purposes, 500 mg acetaminophen was pulverized and mixed into 250 mL of water and placed in the cage's water bottle. Mice were subsequently evaluated for pain by observing hunched posturing, decreased ambulation, and/or ruffled fur. If any of these signs were observed, the animal was immediately euthanized. Mice were maintained for 24 hours prior to tissue harvest, at which time they were euthanized by placement in a chamber that was filled with 100% carbon dioxide at a volume replacement rate of 10-30% per minute. Mice were left in the chamber for five minutes. This was followed by cervical dislocation of the animal.

Partial Bladder Outlet Obstruction Surgery (pBOO)
Adult mice (CD1 or Sca-1 egfp ) were used in partial outlet obstruction surgeries. Anesthesia was performed as above. To create a partial outlet obstruction, a 23G angiocatheter was placed alongside the urethra. A 7.0 Prolene suture was then tied around both the catheter and the urethra. The catheter was then removed, leaving the suture tied around the urethra. Abdomen was then closed with a 5.0 vicryl suture. For sham surgeries, an incision was made in the abdomen and then sutured. Post-operative care was performed as above. For pBOO, mice were allowed to recover for 7 days prior to tissue harvest. Periodically, mice were evaluated for outlet obstruction by voiding stain on paper (VSOP). Euthansia was performed as above.

Voiding Stain on Paper
Following partial outlet obstruction surgery, mice were evaluated on post-operative day 7 via VSOP [21]. Mice were separated into individual metabolic cages that contained Whatman paper on the floor and kept there for 2 hours. Whatman paper was then imaged using UV light.

Quantitative PCR
Bladders were removed and placed in ice-cold Trizol reagent (Life Technologies). Samples were either frozen at -80°C or processed immediately. Tissue was thawed and homogenized using either a Polytron PT 1200 Homogenizer (Kinematica AG) or a BeadBug (Benchmark Scientific) with 3mm zirconium beads. Nucleic acid was extracted using chloroform, further purified using an RNeasy kit (Qiagen) and precipitated with LiCl. Sorted cells were sorted into RNALater (Ambion) and RNA was extracted via RNeasy kit (Qiagen). Plated cells were washed twice in PBS and then left to incubate in Trizol for 5 minutes at room temperature. A kit geared toward low volume RNA, Direct-zol RNA MiniPrep kit (Zymo Research), was used to extract RNA from plated cells. cDNA was made using high capacity cDNA kit from Applied Biosystems (Invitrogen). qPCR was performed using the Solaris system from Dharmacon Thermo Fisher or TaqMan on an Applied Biosystems StepOne Plus thermocycler. Primers were obtained from Assays-on-Demand (Applied Biosystems) and the primers we used are shown in Table D in S1 File.

Immunofluorescence of Mouse Bladders
Whole bladders from freshly sacrificed adult and postnatal day 1 (P1) CD1 mice were fixed in 4% formaldehyde (from paraformaldehyde) in phosphate-buffered saline (PBS) overnight at 4°C. They were then transferred sequentially to 20% (wt/v) sucrose in PBS and 30% (wt/v) sucrose in PBS until saturated. The tissue was embedded in OCT (Tissue-Tek) and frozen in a dry ice isopentane slurry. Sections of 10 μm were cut for all experiments. Sections were washed twice in PBS, incubated for 5 minutes at room temperature in PBS + 0.1% Tween (PBT), and blocked for 30 minutes at room temperature with 10% donkey in PBT (or 10% goat serum in PBT, depending on the animal in which the secondary antibody was raised). All primary antibodies were used at 1:100 dilution and left on overnight at 4°C. Slides were then washed 3 times in PBS for 5 minutes. All secondary antibodies were used at a 1:400 dilution and left on for 2 hours at room temperature. Slides were then washed 3X for 5 minutes at room temperature and mounted with Fluorogel (Electron Microscopy Sciences). A complete list of primary and secondary antibodies used can be found in the Tables A and B in S1 File.

Whole Bladder Digestion to Single Cells
Adult mice were sacrificed and whole bladders were transferred to cold, sterile PBS. Bladders were then minced finely with a sterile razor blade and placed into a dispase II (2.7 mg/mL)/collagenase (1mg/mL) solution at 37°C [22]. Tissue was disrupted~20 minutes with gentle pipetting until tissue was digested to single cells. Average time of digestions was~90 minutes. The reaction was stopped using cold FACS buffer (PBS supplemented with 5% fetal bovine serum (FBS) and 10mM HEPES). Cells were filtered through a 70 μm mesh filter, and centrifuged at 150 x g for 5 min at 4°C. The supernatant was removed and the cells were suspended in pure FACS buffer. Cells were then counted using a hemacytometer and volume of FACS buffer was adjusted to yield a cell concentration ranging from 2x10 6 cells/ml-10x10 6 cells/ml for subsequent antibody staining.  [23]. Cells were stained with antibodies from eBioscience at 1:100 dilution. Cells were stained with conjugated antibodies against Sca-1, CD34, Ter119, CD45 and CD31 for 30 minutes in the dark on ice. Conjugated antibody information can be found in Table C of S1 File. Cells were then washed twice in FACS buffer. Single stain and no stain controls were also prepared for proper compensation. The cells were stained with propidium iodide (Sigma-Aldrich, 1:10 dilution) or Sytox Red (1:250 dilution) 15 minutes prior to the sort to screen out dead cells. Compensation was performed using the compensation setup tool in FACSDiva software (BD Biosciences).

Fluorescence Activated Cell Sorting
Live cells were then sorted for Sca-1 + / CD34 + /lin-using a FACSAria II and FACSDiva software. Post analysis of the sort was performed using Flowjo (Tree Star).

Crystal Violet Colony Forming Assays
Cells were stained and sorted as mentioned above. Cells were plated at a density of 2000 cells per 10 cm dish and allowed to grow for 2 weeks at 37°C in 5% CO 2 . The medium (α-MEM supplemented with 2mM Glutamax, 10% FBS, and penicillin/streptomycin) was changed every 3 days. After two weeks, cells were washed twice with PBS, and then fixed with 4% formaldehyde in PBS for 10 minutes at room temperature. After 2 washes in PBS, cells were stained with 0.1% (w/v) crystal violet in ethanol for 10 minutes at room temperature. Plates were washed extensively with deionized water and allowed to dry.

Immunofluorescence of Colony Forming Assays
Sca-1 + /CD34 + /Lincells were stained and sorted as mentioned above. Cells were plated at a density of 700 cells per well in a 6 well plate lined with cover slips and cultured in α-MEM supplemented with 2mM Glutamax, 10% FBS, and penicillin/streptomycin at 37°C in 5% CO 2 . Medium was changed every 3 days. At various time points (24h, 48h, 4days, and 7d), medium was removed, cover slips were washed twice with PBS, fixed with 4% PFA in PBS for 15 minutes at room temperature, incubated with PBT for 5 minutes at room temperature, and blocked with 10% donkey (or goat) serum in PBT for 30 minutes at room temperature.

EdU Assay
Cells were prepared as mentioned above and 5-ethynyl-2'-deoxyuridine (EdU) was added to the medium at a concentration of 1 μm. Medium was changed every 3d. EdU stain was visualized using Click-iT EdU Alexa-Fluor 647 Imaging Kit (Molecular Probes).
Sca-1 + /CD34 + /lin -Tissue Culture for qPCR Analysis Sca-1 + /CD34 + /lincells were stained and sorted as mentioned above. Cells were plated at a density of 10,000 cells per well in 48 well tissue culture plates (BD Falcon) and cultured in α-MEM supplemented with 2mM Glutamax, 10% FBS, and penicillin/streptomycin at 37°C in 5% CO 2 . This medium was considered to be non-induction medium. Medium was changed every 3 days. At 7 days, cells were processed for RNA as mentioned above in the quantitative PCR section.

Differentiation Assays
Cells were stained, sorted by FACS and cultured under specific conditions for adipogenic, chondrogenic, or osteogenic differentiation (Lonza). Note: In addition to TGF-β3 in the chondrogenic assay, cells were grown in the presence of BMP-6 (500 ng/mL, Peprotech).

Oil Red O Assay
Cells were fixed for 30 min. at room temperature in 10% formalin, washed once with milliQ water, incubated with 60% isopropanol for 3 minutes and stained with Oil Red O (Sigma): milliQ Q water (3:2) for 5 minutes. Wells were washed twice with deionized water and coverslips mounted with aqueous mounting medium.

Alizarin Red Stain
Cells were washed in DPBS and then fixed in 10% formaldehyde (v/v). Cells were washed twice with DPBS and incubated in 2% alizarin red stain for 20 minutes. Cells were washed four times with distilled water and imaged on a Leica inverted microscope.

Endogenous Alkaline Phosphatase Stain
Cells were fixed in 4% PFA for 30 seconds and then washed 2X with PBS. 200 μl of Naphothol AS-MX (Sigma) in dimethyl formamide was added to 10mL of 0.1M Tris-HCl and vortexed. 10 mg of Fast blue BB (Sigma) was then added to the solution and vigorously vortexed. The solution was subsequently syringe filtered using a 0.2 μm filter. The solution was placed in the wells to stain the cells for 1 hour at room temperature in the dark. Cells were then washed 2X with PBS.

Toluidine Blue Stain
Chondrogenic pellets were embedded in paraffin blocks and 6 μm sections were cut. Sections were deparaffinized and hydrated with distilled water. Sections were stained in Toluidine Blue O (Sigma) solution at a concentration of 1 mg/mL in water for 30 minutes at room temperature. Sections were then dehydrated once with 95% ethanol for 5 minutes and then 2X for 5 minutes each with 100% ethanol. Sections were then washed 3X for 10 minutes each with xylene and mounted with a coverslip.

Statistical Analyses
All statistical analyses were done using GraphPad Prism version 5 for Mac OSX, (GraphPad Software, La Jolla, CA). For univariate analysis we employed a Student's T-Test. For multivariate analysis we tested for significance using one-way ANOVA and employed Tukey's post-hoc test.

Bladder digests contain a population of Sca-1 + /CD34 + /lin -MSCs
Cell suspensions were made from dispase II, collagenase digests of whole CD1 mouse bladders and analyzed by FACS (Table 1). Unstained (S1 Fig) and single stain controls (S1 Fig) were used for compensation and gating. Propidium iodide or Sytox red uptake was used to gate living cells from dead cells ( Fig 1A). In order to target single cells, side scatter height versus side scatter area was used to screen out doublets ( Fig 1B) and living cells were further limited by gating based on side scatter area versus forward scatter area ( Fig 1C). Bivariate FACS analysis Table 1. Summary of flow cytometric analysis.

Gate profile
Percent living cells Percentages were averaged from 10 different sorting experiments, and represent percentage of total live cells (all Sytox red or propidium iodide negative cells). ± % is standard error of the mean.
To investigate the identity of the four populations, we FACS sorted cells from adult CD1 bladders and immediately isolated RNA. RNA samples were examined for expression of several genes. Here we determined that Sca-1 + /CD34 + /lincells expressed Sca-1 and CD34 transcripts (Fig 2A), but did not express smooth muscle α actin (ACTA2) or smooth muscle myosin (Myh11) (Fig 2B). Sca-1 -/CD34 -/lincells demonstrated the highest levels of expression of ACTA2 and Myh11 (Fig 2B), suggesting that many of these cells are smooth muscle cells. This population also expressed relatively high levels of the epithelial marker keratin 5 (KRT5), suggesting that this population may contain serosal and urothelial cells (S2 Fig). Sca-1 + /CD34 -/linand Sca-1 -/CD34 + /lindemonstrated expression of genes that suggest that these populations are heterogeneous. For example, Sca-1 + /CD34 -/linexpressed the highest levels of Col1α1 (S2 Fig), which is expressed in connective tissue cells. ACTA2 was also expressed at low levels in this population suggesting that some of these cells are smooth muscle cells. This population did not express high levels of Myh11. The Sca-1 -/CD34 + /linpopulation expressed relatively low levels of Col1α1, but expressed relatively high levels of KRT5, suggesting that some of these cells may be serosal or urothelial cells (S2 Fig). Additionally, Sca-1 -/CD34 + /lincells did express some ACTA2, although at low levels ( Fig 2B). This suggests that some of these cells are also smooth muscle cells.
Since Sca-1 + /CD34 + /lincells uniquely expressed Sca-1 and CD34 and did not express genes associated with differentiated cell types, we used FACS to analyze their ability to form colonies in tissue culture. We plated Sca-1 + /CD34 + /lincells at low density on plastic tissue culture plates. Total live cells were sorted and cultured under the same conditions as controls. Here we found that Sca-1 + /CD34 + /lincells robustly formed colonies after two-weeks while control cells did so sporadically (S2 Fig). To investigate if Sca-1 + /CD34 + /lincells were self-renewing we added the nucleotide analog EdU to culture medium for 24-hours. Cells were grown for 48h or 4d in EdU-free medium before fixation and staining for EdU incorporation and Sca-1. At 48h we found Sca-1 cells that had incorporated EdU indicating that sorted Sca-1 + /CD34 + /lincells passed through S-phase and continued to express Sca-1 (Fig 2C). At 4d we found cells expressing Sca-1 and smooth muscle myosin (SMM) proteins that had incorporated EdU (Fig 2D). This finding demonstrated that Sca-1 + /CD34 + /lincells self-renewed during the first 4d of culture.
While the exact sorting profile of MSCs is still a topic debate, a hallmark of MSCs is their ability to form colonies and adhere to tissue culture plastic [24]. Because Sca-1 + /CD34 + /linbladder cells both adhered to tissue culture plastic and formed colonies, we next investigated their ability differentiate into various mesenchymal lineages (Fig 3). Here we found that when bladder derived Sca-1 + /CD34 + /lincells were placed in induction media, they formed bone ( Fig  3A, 3B, 3C and 3D), adipose (Fig 3E and 3F) and cartilage (Fig 3G and 3I) cells. Brightfield images (S2 Fig; Fig 3H and 3J) coupled with staining ( Fig 3B, 3D and 3F) suggest that Sca-1 + / CD34 + lincells cultured in α-MEM alone or Sca-1 -/CD34 -/lincultured in induction medium did not differentiate into these cells types. Thus, EdU incorporation into cells that continued to express Sca-1 and differentiation into bone, cartilage and fat in specialized medium supported the idea that Sca-1 + /CD34 + /linbladder cells could represent a resident MSC population.

Sca-1 + and CD34 + expression is primarily localized in stromal layer cells of the adult mouse bladder
To determine the localization of Sca-1 + /CD34 + /lincells in the adult bladder we examined sectioned mouse bladders from adult (6-8 week old) CD1 mice for cells expressing Sca-1 and CD34 (Fig 5). We found that there was robust expression of Sca-1 in mesenchymal cells of the stromal layer lying in between the urothelial and smooth muscle layers of the bladder. This expression pattern sporadically overlapped with the basal epithelial layer of the urothelium, an area that was previously defined as containing epithelial stem cells of the urothelium [6]. Cells reacting with the Sca-1 antibody extended from the stromal layer into the smooth muscle layer of the bladder (Fig 5A, arrows). CD34 expression was also prominent in the stromal layer ( Fig  5B). This expression was more clearly demarcated from the urothelium as there was no detectable staining near the basal urothelium (Fig 5B and 5C). When Sca-1 and CD34 were co-localized ( Fig 5E) it was clear that CD34 expressing cells are positioned more centrally within the stromal mesenchyme, whereas Sca-1 is found throughout the stroma and in isolated cells of the urothelium. This observation was consistent with our FACS data in which we had observed that approximately 51% of Sca-1 + cells expressed CD34 and that most CD34 + cells express Sca-1.
To investigate if stromal Sca-1 + /CD34 + cells are present during fetal development we stained sections from CD1 embryonic bladders (Fig 6). As early as E14.5, CD34 expression was localized throughout the bladder wall and in portions of the differentiating smooth muscle layer of the bladder (Fig 6A). CD34 was observed throughout the E16.5 bladder, more dispersed through the bladder wall and in what appears to be a condensing stromal layer (Fig 6E). At E18.5, CD34 expressing cells were numerous in what is now clearly the stromal layer of the bladder (Fig 6I). The distribution of these cells narrowed over time to correspond to a compact stromal compartment at P01 (Fig 6M). Sca-1 protein, on the other hand, was not detected in embryonic bladders until E18.5 (Fig 6B, 6F, 6J and 6N) while it was detected in other tissues of the embryo (data not shown). At E18.5, the Sca-1 antibody reacted with a small number of urothelial cells (Fig 6J, arrowheads). In P01 sections, Sca-1 localization overlapped the stromal compartment of the bladder (Fig 6N). It was not until P01 that we observed robust co-localization of CD34 and Sca-1 protein in the stromal layer (Fig 6M and 6N). Thus, CD34 expression precedes Sca-1 expression in the stromal mesenchyme in utero, but both are established by the first day of life.
To investigate whether the Sca-1 promoter was active early in bladder development, we examined tissue from Sca-1 egfp mice [19]. Surprisingly, the Sca-1 gene demonstrated activity at these embryonic stages in the urothelium (Fig 6C, 6G, 6K and 6O). No EGFP was detected in areas where we observed CD34 localization at E14.5 (Fig 6D). However, at E16.5, E18.5 and P01, EGFP was detected in the developing stroma in increasing amounts (arrowheads, Fig 6H,    6L and 6P, arrowheads) and partially overlapped with CD34 expression. In 4-6wk old mice, EGFP was only observed in the stroma and in isolated cells of the detrusor muscle, but not in the urothelium (S3 Fig). This demonstrated that there is a shift in gene expression of Sca-1 from the urothelium in embryos to the stroma and detrusor in adults. The inability to detect Sca-1 protein at earlier stages may reflect a delay in accumulation of the Sca-1 antigenic determinant in the urothelium.

A subset of stromal Sca-1 + /CD34 + /lincells respond to Shh in the adult bladder
The urothelium is a source of sonic hedgehog (Shh) signaling in the bladder [26]. It is also known that other Sca-1 + /CD34 + /linmesenchymal stem cell populations reside in Shh signaling domains [14]. We investigated if the stroma of the adult mouse bladder contained cells responding to Shh signals. We stained sections of adult CD1 bladders with an antiserum to Shh and observed that Shh was largely expressed by the urothelium (Fig 7A). We next used a tamoxifen inducible Cre gene that is driven by the Gli1 promoter, which is activated in response to Shh signaling [20]. We generated mice carrying Gli1Cre ERT2 and the Cre reporter Rosa mTmG . To obtain a "snapshot" of cells responding to Shh we injected one dose of tamoxifen (4mg) three days before bladders were harvested for analysis by fluorescence microscopy. Here we observed two main sites of Cre labeled cells. First we observed cells in the stroma, particularly in the layers directly adjacent to the urothelium (Fig 7B and 7C). Partial overlap between Gli1 driven EGFP, Sca-1 and CD34 was also observed (Fig 7B and 7C). This was consistent with the hypothesis that some Sca-1 + /CD34 + /linmesenchymal stem cells are responding to Shh. We also observed a significant number of cells in the detrusor muscle, primarily in cells closest to the stromal compartment (Fig 7D, arrowheads). Using the same method of tamoxifen administration, we labeled cells in adult bladders and analyzed the cells by FACS. Here we observed that approximately 10% of the Sca-1 + /CD34 + /linpopulation was labeled by recombination of Rosa mTmG (Fig 7E). Together, these observations support a model in which bladder Sca-1 + /CD34 + /lin -MSCs exist in a Shh signaling domain, but are not all directly responding to Shh.

Expression of Sca-1 and CD34 is altered after bladder outlet obstruction
Obstruction is known to cause permanent alterations to the tissue architecture of the bladder wall. This can occur in patients even after the bladder is obstructed for a short period of time and then obstruction is surgically relieved [27]. We used two models of bladder obstruction: acute bladder outlet obstruction (aBOO), in which the bladder was completely obstructed for 24h, and partial bladder outlet obstruction (pBOO) for 7d. Partial obstruction was confirmed by post-mortem bladder size (S3 Fig) and voiding stain on paper assays (S3 Fig). Both acute and partially obstructed bladders were larger and more distended than sham operated bladders (Fig 8A and 8B, S3 Fig). Sca-1 mRNA expression in bladders from animals undergoing acute obstruction vs. sham surgery was decreased (Fig 8C). CD34 was also decreased but by a larger amount (Fig 8C). In the case of either aBOO or pBOO the bladder wall changed after obstruction (Figs 8E, 8G, 8I, 8K, 9B and 9D). Decreased amounts of total SMM mRNA expression were detected through qPCR analysis in the aBOO model ( Fig 8C). Interestingly, we observed Sca-1 expression in the basal urothelium with anti-Sca-1 in aBOO (Fig 8E and 8I) and pBOO mice (Fig 9B). Similarly, a shift in Sca1 egfp expression to the urothelium was detected in pBOO mice (Fig 9C and 9D arrowheads). The urothelium did not express CD34 after either aBOO or pBOO. Stromal coexpression of Sca-1 and CD34 was maintained after pBOO model of obstruction (Fig 9B, arrowheads).
We investigated if the sorting profile of Sca-1 + /CD34 + /lincells changed after pBOO (hereafter we did not pursue aBOO since it is quite damaging to the bladder and mice survive only 24h after surgery). With FACS analysis, we found that the relative proportions of the Sca-1 + / CD34 + /lincells changed after pBOO. In shams, this population constituted roughly 15% of living cells, while in obstructed bladders the population decreased to approximately 5% of living cells (Fig 9E and 9F).
Because we observed a change in the Sca-1 + /CD34 + /linprofile after pBOO, we next analyzed gene expression in this population using qPCR (Fig 9G-9M). We tested if smooth muscle genes were altered in response to pBOO. In sham operated mice neither Myh11 or ACTA2 were detected in the Sca-1 + /CD34 + /linpopulation, but both genes were increased in Sca-1 + / CD34 + /lincells after pBOO (Fig 9G and 9H). Sca-1 -/CD34 -/lincells also expressed Myh11 and ACTA2 in obstructed, although SMM was decreased when compared to the sham (S3 Fig). We observed an approximate 10-fold increase in the expression of uroplakin II in Sca-1 +-CD34lincells (Fig 9M), consistent with our observation that Sca-1 expression is increased in the urothelium after obstruction.
Since obstruction results in a high-pressure bladder that has difficulty emptying, the bladder wall responds by becoming more rigid. This usually involves the production of fibrotic matrices typically marked by the accumulation of collagen fibers. We tested if FACS sorted cells increased their expression of fibrosis associated genes after pBOO. Within the Sca-1 + /CD34 + / linpopulation, we observed significant increased expression of collagen type 3, alpha 1 (COL3α1) and in fibronectin 1 (FN1) between the sham and pBOO model (Fig 9I and 9J). Matrix metalloproteinase 2 (MMP2) and collagen type 1, alpha 1 (Col1α1) were increased but  this difference was not statistically significant (Fig 9K and 9L). Thus, obstructive injury reduced the number of Sca-1 + /CD34 + /linmesenchymal stems cells and caused these cells to express both smooth muscle cytoskeletal genes and genes that are increased in fibrosis.

Discussion
Here we present evidence that the stromal layer of the bladder contains a population of Sca-1 + / CD34 + /linmesenchymal stem cells. This population resembles Sca-1 + /CD34 + /lincells found in the adventitia of arteries and in lung [8,9,14]. Like these other Sca-1 + /CD34 + /lincell-types, mouse bladder mesenchymal stem cells can differentiate into smooth muscle-like cells ex-vivo. Together with urothelial progenitor cells [6], we believe that there are two distinct progenitor lineages in the adult mouse bladder. It is reported that the adult rodent bladder can regenerate de novo [28,29] and bladder augmentations constructed from extracellular matrix become cellularized [30], so it is likely that intrinsic stem-like bladder cells contribute to the re-growth of the bladder after injury.
We do not yet know the source of stromal Sca-1 + /CD34 + /linmesenchymal stem cells. One possibility is that these cells are derived from circulating cells originating from circulating hematopoietic stem cells (HSC). However, Sca-1 + /CD34 + cells are present in the stroma of the P01, suggesting that these cells are derived from bladder development and not from HSCs. Studies have tested if labeled donor cells contributed to the adult bladder after bone marrow replacement and these studies found that a small number of labeled cells arrived in the stromal layer of the bladder [5]. We believe that this small number cannot account for the total number of Sca-1 + /CD34 + /lincells we see in the adult mouse bladder, but our experiments do not rule out the possibility that the stromal mesenchymal cells we observe are derived from embryonic HSCs.
Our results indicate that the mechanism of colonizing the stroma with Sca-1 + /CD34 + /lincells occurs by P01. This means that the original population of stromal mesenchymal stem cells is largely derived during embryonic development. This time was later than observed with adventitial Sca-1 + /CD34 + /lincells, which were observed first in the perivascular space of the great arteries as early as E15.5 [14]. We saw that CD34 was expressed in a large number of cells within the bladder wall including areas where smooth muscle was differentiating. As development continued the CD34+ population appeared to define the bladder stroma before Sca-1 protein was expressed there. However, once formed by E18.5, the stroma began to express Sca-1 and this increased so that two days later, the pattern of expression of Sca-1 reflected the pattern we observed in adult bladders. Since Sca-1 + /CD34 + cells are found in the bladder stroma at birth, we believe that the stromal MSCs we have defined represent an intrinsic or resident population of progenitor cells.
We believe that the stromal layer could create an appropriate environment or niche for mesenchymal stem cells in the bladder. This is supported by the findings of Woo et al., (2011) who observed that non-bladder derived MSCs injected into the tail vein were recruited to the stroma after partial bladder outlet obstruction. In addition, bone marrow replacement experiments demonstrated that circulating stem cells also target the bladder stroma [4,5]. The stroma is an extracellular matrix rich, loose mesenchyme similar to other stem niches such as the hematopoietic [31], hair follicle, intestinal and vascular adventitial niches. Thus, among other things, its role may be to provide an environment that is suitable for both resident stem and circulating HSCs that are targeted to the bladder. had pools of 2 pBOO and 2 sham CD1 mice. RNA was collected 7 days post pBOO surgery. Asterisks represent significance values P<0.05 * and P<0.001 *** after Student's T-Tests. Data is normalized to sham gene expression, and GAPDH was used as the reference gene. (M) qPCR analysis of RNA of Sca-1 + /CD34 -/lincells from sham surgeries (solid red bar) and pBOO surgeries (solid blue bar). doi:10.1371/journal.pone.0141437.g009 It is likely that the adult mouse bladder contains more than one population of progenitor cells. Shin and colleagues (2011) described a population of Shh expressing basal urothelial cells that have progenitor cell properties [6]. These cells proliferated and repopulated the urothelium of bladders challenged with a pathogenic strain of Escherichia coli. FACS sorted cells formed self-renewing organoids in culture and lineage tracing demonstrated that these cells could regenerate all cell types of the urothelium after injury. Interestingly, their findings demonstrated an increase in urothelial expression of Shh and an increase in stromal Wnt and FGF genes after infection. Shh signal originating in the basal epithelial layer of the urothelium was proposed to be received by cells in the stroma, leading to downstream expression of Gli1, Patched, and Axin2. Bladders from Gli1 mutant mice had decreased cell proliferation in the stroma after infection and this could be modulated by pharmacologic reduction or augmentation of Wnt signaling. These findings indicate that, after injury, the urothelium establishes a chemical-based communication with the stroma to regulate cell-proliferation. These signals may be a way that the stroma can respond to injury and promote cell growth therein.
While resident MSCs have been identified in many organs, it is not clear what their functions are within a normal, healthy organ. Even transplantation of sub cultured MSCs into hosts only reveals the potential of such cells in an artificial in vivo context. The true experimental paradigm for understanding the role of these cells is in vivo lineage analysis and this is particularly challenging. For example, if resident MSCs slowly contributed to an organ over the lifetime of an animal, then a lineage tracing experiment would involve investigation for a span of months or years. A few groups have attempted such analyses, but in most cases the true relevance of resident stem cells can only speculated upon. In the case of bladder, stromal MSCs, the appropriate Cre driver that could be used in lineage tracing experiments is still unresolved. The Gli1 CreERT2 mouse line was used to investigate the role of perivascular MSCs in mice [32] and we had considered that this line might be suitable for labeling bladder Sca-1 + /CD34 + /lincells. However, we observed that the Gli1 CreERT2 gene promoted recombination in only about 10% of Sca-1 + /CD34 + /lincells and in numerous cells outside the stroma. Thus, it became clear that better Cre models will need to be developed before the role of bladder Sca-1 + /CD34 + /lin -MSCs in the bladder can be studied in vivo. With respect to the role that Sca-1 + /CD34 + /lincells play in an obstructed bladder, the observed upregulation in expression of genes such as Col3α1 and FN1 might provide insight into this population's contribution to the fibrosis that often accompanies bladder obstruction.
Another unexpected finding in our study was the pattern of expression of the Sca-1 egfp gene. First, it was expressed in the developing urothelium prior to our ability to detect the protein with the Sca-1 antiserum. In adults, Sca-1 egfp was observed in stromal and in isolated cells in the detrusor muscle. However, after obstructive injury, Sca-1 egfp was re-expressed in the urothelium. In this case, urothelial localization was detected by the Sca-1 antiserum. This indicates that obstructive uropathy restores the embryonic pattern of expression of Sca-1, which may be a response reflecting how the urothelium needs to adapt after the injury. Activation of the "fetal gene program" has been observed in other pathologies, such as congestive heart failure [33]. The reasons why embryonically expressed genes are re-expressed is not completely understood, but in some cases the presence of these genes can be used as disease markers. Sca-1 itself would not currently be useful in patients because there is no known human Sca-1 homolog to date. However, urothelial Sca-1 in mice may indicate that other fetal genes are expressed after obstruction and these could be used as indicators of remodeling before other symptomatic changes occur.  Table A, Primary antibodies used in confocal fluorescent microscopy. Table B, Secondary antibodies used for confocal fluorescent microscopy.