Single‐cell profiling reveals various types of interstitial cells in the bladder

Abstract Clarifying the locations, molecular markers, functions and roles of bladder interstitial cells is crucial for comprehending the pathophysiology of the bladder. This research utilized human, rat and mouse bladder single‐cell sequencing, bioinformatics analysis and experimental validation. The main cell types found in human, rat and mouse bladder tissues include epithelial cells, smooth muscle cells, endothelial cells, fibroblasts, myofibroblasts, neurons and various immune cells. Our study identified two significant types of interstitial cells (PTN+IGFBP6+PI16 (CD364)+ CD34+) and myofibroblasts (STC1+PLAT+TNC+). These two types of interstitial cells are mainly located in the subepithelial lamina propria, between muscles and between muscle bundles. In the CYP (cyclophosphamide)‐induced bladder injury mouse model, the interaction types and signals (MK, MIF, GDF and CXCL) of fibroblasts and myofibroblasts significantly increased compared with the normal group. However, in the aging mouse model, the signals CD34, LAMININ, GALECTIN, MK, SELPLG, ncWNT, HSPG, ICAM and ITGAL‐ITGB2 of fibroblasts and myofibroblasts disappeared, but the signals PTN and SEMA3 significantly increased. Our findings identified two crucial types of interstitial cells in bladder tissue, fibroblasts and myofibroblasts, which play a significant role in normal bladder physiology, CYP‐induced bladder injury and aging bladder development.


| INTRODUCTION
The bladder's filling and voiding of urine are controlled by a combination of complex neural systems in the spinal cord and peripheral neurons, as well as the muscle cells and urothelium. 1 Smooth muscle cells are crucial for bladder function, and increased excitation-contraction coupling in the detrusor muscle results in the voiding of urine from the bladder. The urothelium also acts as a barrier between the blood and urine, preventing reabsorption of harmful compounds from the urine, playing a role in sensory function. This is accomplished through the expression of numerous receptors/ion channels that serve as neuronal sensor molecules. 2 In contrast, the role of interstitial cells in the bladder is still a subject of debate. 3 Further research is needed to fully understand the locations, molecular markers, functions and roles of interstitial cells, as well as their contribution to spontaneous contractions of the bladder.
In previous studies, various molecular markers such as c-Kit, PDGFRα, Desmin, Vimentin, cGMP and CD34 have been utilized to label interstitial cells. [4][5][6][7] Moreover, the identified bladder interstitial cells were named interstitial cells (ICs), Cajal interstitial cell-like cells (ICCs), Cajal interstitial cells (ICCs), pacemaker cells and fibroblast-like cells (FLCs), leading to confusion in the field. [4][5][6][7][8][9] Recent research has used immunohistochemistry and electron microscopy to identify three types of interstitial cells in the bladder: telocytes, fibroblasts and myofibroblasts. [4][5][6][7][8][9] Despite these advancements, the exact locations, molecular markers, functions and roles of bladder interstitial cells are yet to be fully understood. The development of single-cell RNA-seq (scRNA-seq) technology, specifically the 10x Genomics platform, has allowed for an unbiased classification of cell types from single-cell suspensions, and this method has been applied to bladder tissue research with great success. [10][11][12] Additionally, the utilization of nextgeneration sequencing (NGS) enables easy handling and analysis of substantial amounts of RNA-seq data, leading to improved visualization. This study aims to explore the various cell types present in bladder tissue, with the ultimate goal of gaining new insights into bladder organization and disease occurrences. Five wild-type C57BL/6J mouse bladder tissues were citied from published paper. 12 The mouse scRNA-seq data (detailed information in Table S5) were downloaded from the published database (accession number GSE153562) for analysis. Bladder tissues of newly collected samples were collected from female mouse.

| Tissue preparation, handling and enzymatic isolation
Rats were sacrificed by cervical dislocation. Bladder tissues were isolated from the sacrificed rats under aseptic conditions and placed in sterile PBS (phosphate buffered saline) (Gibco, pH 7.4 basic, lot #8119170, China). With sterile forceps and either small sterile scissors or a sterile scalpel, the bladder tissue was minced into small pieces.
The minced bladder tissue was transferred to a 50 ml tube for enzymatic digestion.
Donor bladder tissues from humans (size 1 cm Â 1 cm Â 1 cm) were collected in DMEM-high glucose medium (Gibco, lot #8119079, China) containing 10% FBS (Gibco, lot #2017490C, Australia) and 5 μg/ml gentamicin (Gibco, lot #2023926, USA). The wells of a sterile 6-well plate were prefilled with 10 ml of prewarmed tissue washing medium (PBS and 5 μg/ml gentamicin). The tissue was gently agitated in the well using sterile forceps for 5-10 s; successive washing of the tissue was continued through each unused well until all six wells were used. With sterile forceps and either small sterile scissors or a sterile scalpel, the bladder tissue was minced into small pieces. Minced bladder tissue was transferred to a 50 ml tube for subsequent enzymatic digestion. Tissues from rats and humans were digested for approximately 4 h at 37 C with agitation in a digestion solution containing 2 mg/ml papain (Worthington, lot #LS003120), 100 U/ml Collagenase II (Sigma, CAS No: 9001-12-1, USA) and 100 U/ml Collagenase IV (Sigma, CAS No: 9001-12-1, USA) in DMEM-high glucose medium. 11 The digested suspension was passed through a 60 μm Steriflip (Millipore, CAS No: SCNY00060) and washed twice with washing medium (PBS and 0.04% BSA). Enriched live cells were washed and counted using a haemocytometer with trypan blue. Overall, it took 6-7 h from obtaining the tissues to generating single-cell suspensions run on the Chromium 10Â device.

| Library Preparation and Sequencing
After digestion, the single-cell suspension was used for the quality assessment and counting; the cell survival rate was generally above 80%. Cells were loaded onto the 10Â Chromium Single Cell Platform (10Â Genomics, https://www.10xgenomics.com/resources/supportdocumentation/) at a concentration of 1000 cells/μl (Single Cell 3 0 library and Gel Bead Kit v.3) following the manufacturer's protocol.
The generation of gel beads in emulsion (GEMs), barcoding, GEM-RT clean-up, complementary DNA amplification and library construction were all performed according to the manufacturer's protocol. Qubit was used for library quantification before pooling. The final library pool was sequenced on the Illumina HiSeq instrument using 150-base-pair paired-end reads. Briefly, 6-10 billion base calls were generated for each sample. PCR was performed with the same PCR primer cocktail used in TruSeq DNA Sample Preparation. The pipeline of raw data processing and calling was as follows. Quality control of raw data was conducted with Fast QC software (http://www. bioinformatics.babraham.ac.uk/projects/fastqc/). Reads were mapped to reference the genome build using the Burrows-Wheeler Alignment tool (BWA). The duplicate reads were flagged using Picard-tools  Table S5) were downloaded from the published database (accession number GSE153562) for analysis.

| Data processing tools
All data processing was based on the R platform (version 3. 6

| Immunohistochemistry and Immunofluorescence Staining
All antibodies applied in the current assays are listed in Table S6.   3 | RESULTS

| Identification of cell types in the bladder tissues of humans, rats and mouse
The cells were collected from different zones of the human bladder and from the overall bladder of rats, and were then processed into a single-cell suspension for scRNA-seq analysis using the 10x Genomics platform version 3.0 ( Figure 1A). scRNA-seq data from mouse bladders was obtained from the Short Read Archive (accession number GSE153562). 12 Unsupervised clustering analysis (UMAP, uniform manifold approximation and projection) showed 36, 22 and 28 clusters in human, rat and mouse cells, respectively. Based on bladder histological structure and marker gene characterization (Table S1)

| Characterization of different cell types in the bladder
After strict filtering and quality control (average gene level ≥ 1000), a total of 27,798 cells from two human (one male and one female), 14,891 cells from two rat (one male and one female) and 33,622 cells from five mouse (female, different ages) bladder tissues were obtained for clustering ( Figure 2A and Table S5). The number and type of cells showed minimal differences among species (

| Characterization of fibroblasts and myofibroblasts in the bladder
The marker genes for distinguishing fibroblasts were PTN, IGFBP6,  Figure 3G). To locate myofibroblasts in the bladder, we used antibodies against TRPA1 (in this study, we found that TRPA1 specifically expressed in human myofibroblast), DES and TGFBI were used for double immunofluorescence staining. The results showed that myofibroblasts were located mainly between the urothelium and detrusor ( Figure 4C). However, the expression level of TRPA1 was found to be higher in human myofibroblasts, while it was undetectable in myofibroblasts from rat and mouse tissue ( Figures 2H-J).
These findings suggest that the function of myofibroblasts in the bladder tissue of humans, rats and mice may still differ and requires further investigation.

| The role of fibroblasts and myofibroblasts in CYP (cyclophosphamide)-induced bladder injury
To gain a comprehensive understanding of fibroblast and myofibroblast heterogeneity and dynamics in the presence of injury, we analysed published scRNA-seq data. 14  were highly active, albeit at different levels, in acute injury bladder tissues ( Figure 5L-N). We found that highly correlated cytokines and receptors, including Mif, Cd44, Sdc1, Sdc4 and Ncl, dominated the signalling patterns ( Figure 5L). A closer look at the signalling shows its higher signalling redundancy and high target promiscuity in fibroblasts and myofibroblasts in acute injury bladder tissues ( Figure S3I, J).

| The role of fibroblasts and myofibroblasts in the aged bladder
To better understand fibroblasts and myofibroblasts' change in aging mouse bladder tissues, we divided them in mouse bladder ( Figure 3C) into two groups ( Figure 6A, normal, age ≤1 years old; aged, age ≥ 18 months).
No significant difference in subcategory cell proportions was found between the aged and normal groups ( Figure 6B). In terms of cell-cell interaction signalling, aged mouse demonstrated a lower interaction number and strength than normal one ( Figure 6C). Additionally, the number and strength of cell-cell interactions decreased in fibroblasts and myofibroblasts ( Figure 6D). Analysis of the overall signalling pattern revealed that CD34, LAMININ, GALECTIN, MK, SELPLG, ncWNT, HSPG, ICAM and ITGAL-ITGB2 signalling disappeared in aged mouse; while PTN and SEMA3 signalling was significantly increased in aged mouse ( Figure 6E).
We also compared the relative information flow between aged and normal mouse groups, showing the same changes as the overall signalling pattern ( Figure 6F). More specifically, signalling changes in fibroblasts included incoming signalling of APP, increased signalling of Thbs2-Cd47 and decreased signalling of Col4a2/Col4a1/Col1a2/Col1a1-Sdc1 ( Figure 6G, H). Signalling changes in myofibroblasts included outgoing signalling of MIF ( Figure 6G), increased signalling of App-Cd47 and decreased signalling of Col4a2/Col4a1/Col1a2/Col1a1-Sdc1 ( Figure 6I).

| DISCUSSION
In the current study, scRNA-seq technology was employed to map the transcriptional landscape of human and rat bladder cells. A total of 12, 8 and 10 major cell types were identified in human, rat and mouse bladder tissues respectively, after capturing and filtering cells from the tissues for clustering. In addition to the commonly reported cell types (epithelial cells, smooth muscle cells, endothelial cells and immune cells), we found two types of interstitial cells, that is, fibroblasts and myofibroblasts, which participate in the development of bladder pathology, including drug-induced cystitis and bladder aging.
Although the major cell clusters did not have a one-to-one correspondence across human, rat and mouse tissues, they were coincident with each other. Interstitial cells have been of growing interest in bladder research because they have been linked to the spontaneous contraction activity of the bladder. Interstitial cells have been proven to exist in the bladder of humans and guinea pigs. 7 Along with immune-related cells, two types of suburothelial interstitial cells, including fibroblasts and myofibroblasts, were identified. 15 Buoro et al. first reported the existence of myofibroblasts in the bladder in 1993, and these cells stain for vimentin and α-smooth muscle actin but not for desmin. 16 More importantly, the scRNA-seq data show that myofibroblasts express the specific CD markers CD362 and ultrastructure of fibroblasts in the human bladder, which was coincident with the current findings. 19 We found that fibroblasts have the specific markers PI16 + PDGFRα + /CD34 + and the specific CD marker However, more mechanistic studies need to be performed in the future.
The current study has some limitations. First, due to difficult process of tissue digestion, some types of cells were captured disproportionately between human, rat and mouse tissues, such as the urothelium cells, fibroblasts and smooth muscle cells. Many connective tissues are present in the bladder, so biases are likely introduced by the different tissue dissociation protocols used. 27,28 We used a combination of enzymatic dissociation that can avoid biases of the cell populations captured for scRNA-seq. However, bladder detrusor cells within a bundle are connected to form a functional syncytium, which determines digestion's difficulty. Only a small part of smooth muscle cells was attributed to the detrusor, because a large ratio of captured muscle cells may come from vascular, and even from the muscularis mucosae where smooth muscle fascicles exist freely. 27,29 Secondly, due to the limitation of cost and sample source, the sample size of this study is small, and only the anterior wall of bladder tissue is captured and digested for analysis; thus, we could not ensure that all cell types were present in the human bladder tissues obtained.
Further, the structure of the bladder neck and urethral sphincter differs from that at the anterior wall, which may lead to the absence of some important cell types related to urination control. However, the whole bladder from rats was analysed for clustering, and we did not find any significant difference between the rat and human samples. Thirdly, only two samples from humans and rats were enrolled for scRNA-seq analysis, though enough cell numbers were captured using a 10Â genomic platform. But current research captured all cell types that emerged in previous studies using scRNA-seq technology.
In addition, in order to exclude interference from other bladders path- ACKNOWLEDGEMENT Special thanks give for Genergy Technology Co., Ltd., Shanghai to do 10Â genomic sequencing for us.

FUNDING INFORMATION
The project was supported partly by grants from National Natural Science Foundation of China (NSFC81873628, 81974101, 81873629).