SEARCH

SEARCH BY CITATION

Keywords:

  • bladder;
  • claudin;
  • differentiation;
  • idiopathic detrusor overactivity;
  • interstitial cystitis;
  • stress urinary incontinence;
  • uroplakin

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

OBJECTIVE

To develop a novel in vitro approach to test the hypothesis that failure of urothelial differentiation underlies the aetiopathology of interstitial cystitis (IC), where there is evidence of compromised urinary barrier function, as benign dysfunctional bladder disease encompass several poorly understood clinically defined conditions, including IC, idiopathic detrusor overactivity (IDO) and stress urinary incontinence (SUI).

MATERIALS AND METHODS

Biopsy-derived urothelial cells from dysfunctional bladder biopsies were propagated as finite cell lines and examined for their capacity to differentiate in vitro, as assessed by the acquisition of a transitional cell morphology, a switch from a cytokeratin (CK)13lo/CK14hi to a CK13hi/CK14lo phenotype, expression of claudin 3, 4 and 5 proteins, and induction of uroplakin gene transcription.

RESULTS

Two of 12 SUI cell lines showed early senescent changes in culture and were not characterized further; one of seven IC, one of five IDO and a further three SUI cell lines had some evidence of senescence at passage 3. Of the seven IC-derived cell lines, four showed a near normal range of differentiation-associated responses, but the remainder showed little or no response. Most IDO cell lines (four of five) showed a normal differentiation response, but at least three of the 10 SUI cell lines showed some compromise of differentiation potential.

CONCLUSION

This study supports the existence of a subset of patients with IC in whom a failure of urothelial cytodifferentiation might contribute to the disease, and provides a novel platform for investigating the cell biology of urothelium from SUI and other benign dysfunctional conditions.


Abbreviations
CK

cytokeratin

EGF(R)

epidermal growth factor (receptor)

IDO

idiopathic detrusor overactivity

IC

interstitial cystitis

(N)HU

(normal) human urothelial

PPARγ

peroxisome proliferator activated receptor γ

RXRα

retinoid X receptor α

(S)UI

(stress) urinary incontinence

RT

reverse-transcription

GAPDH

glyceraldehyde phosphate dehydrogenase.

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The urothelium is the highly specialized transitional epithelium found lining the urinary bladder and associated urinary tract. Organized into basal, intermediate and terminally differentiated superficial cell zones, the urothelium provides a highly effective barrier that prevents urine from penetrating the underlying tissues. The urinary barrier properties are primarily attributable to the specialization of the terminally differentiated superficial urothelial cells, which have molecular features that limit permeability via transcellular and paracellular routes.

The apical surface of the superficial cell membrane is covered with multiple thickened plaques of asymmetric unit membrane, a unique ultrastructural feature of urothelial cells that forms as a result of interactions between integral transmembrane proteins, known collectively as the uroplakins. There are four major uroplakin proteins that interact, to form uroplakin-1a/uroplakin-2 and uroplakin-1b/uroplakin-3a pairs; disruption of asymmetric unit membrane formation, e.g. by germline deletion of the uroplakin-3a gene [1], leads to major alterations in urothelial morphology and function, including increased transcellular permeability.

The intercellular tight junctions constitute the main paracellular barrier. The capacity for different epithelia to show a range of paracellular permeability functions and to modulate these in response to (patho)physiological signals is dictated by differential expression of the claudins, which form the primary seal-forming fibrils of the tight junction. Claudins constitute a family of ≈ 24 proteins and in urothelium show a pattern of expression related to the stage of differentiation, with claudin 7 expressed in all but the superficial layer of the urothelium, claudins 4 and 5 expressed at basolateral junctions of superficial cells, and claudin 3 restricted to the ‘kissing points’ between adjacent superficial cells [2].

Interstitial cystitis (IC) is a chronic and often debilitating inflammatory disorder of the urinary bladder, characterized by urinary urgency, frequency and bladder pain, in the apparent absence of any infectious agent. The aetiology and pathophysiological mechanisms of IC remain undetermined [3,4] and it is considered unlikely that one causal mechanism is responsible. Several reports suggested that a compromised urothelial barrier is a feature of the disease [5–7]. However, it is unclear whether this is due to an inherent dysfunction of the urothelium itself, or an indirect influence of the local (e.g. cytokine) environment on urothelial differentiation and function.

There are several other benign dysfunctional conditions in the bladder, including urge urinary incontinence (UI) secondary to idiopathic detrusor overactivity (IDO) and stress UI (SUI) associated with urodynamic stress incontinence. Although some cases of DO are neurogenic in origin, involving dysregulation of bladder function secondary to changes in the peripheral or CNS, in many cases the cause remains unidentified and the condition is classified as IDO [8]. Due to its intimate association with the suburothelial afferent nerve fibres and a potential role in bladder sensation, it was suggested that the urothelium might be involved in the aetiopathology of IDO [9], although urothelial barrier dysfunction has not been associated with the condition. SUI is a complaint of involuntary leakage on effort or exertion, or on sneezing or coughing. The urothelium has not been implicated in the aetiopathology of SUI, for which the principal mechanisms involve anatomical changes to the pelvic floor, resulting in a loss of support at the bladder neck, and compromised neuromuscular function of the urethral sphincter [10].

We have developed methods to isolate and propagate normal human urothelial (NHU) cells as finite cell lines from surgical resection specimens [11,12]. In a low (0.09 mm) calcium, serum-free culture system, NHU cells have a proliferative phenotype driven by autocrine/paracrine activation of the epidermal growth factor receptor (EGFR) [13] and can be induced to undergo differentiation in response to activation of the nuclear receptor, peroxisome proliferator activated receptor γ (PPARγ), when downstream signalling through EGFR is blocked [2,14,15]. The in vitro cytodifferentiation of NHU cells is accompanied by specific changes in the transcription and/or translation of cytokeratins, claudins and uroplakins that relate to urothelial differentiation in situ, and serve as objective markers of differentiation response [2,14,15]. Thus, we developed the tools to assess the differentiation potential of HU cells in culture.

In the present study, we established a novel method for the growth of finite HU cell lines from three small cystoscopic cold-cut biopsies retrieved from patients with IC (diagnosed according to strict guidelines [16]), IDO and SUI. In addition to investigating the morphology and growth properties, we examined whether urothelial cells from the different patient groups could be induced to undergo molecular differentiation in terms of a switch from a cytokeratin (CK)13lo/CK14hi to a CK13hi/CK14lo phenotype, expression of claudin 3, 4 and 5 proteins, and induction of uroplakin gene transcription.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The high-affinity PPARγ ligand, troglitazone, was a kind gift from Parke-Davis Pharmaceutical Research (Ann Arbor, MI, USA). The EGFR tyrosine kinase inhibitor, PD153035, was obtained from Calbiochem-Novabiochem Biosciences Ltd. (Nottingham, UK). All urothelial specimens were collected with appropriate local research ethics committee approval and full-informed patient consent. A bladder-derived NHU cell line established from a surgical resection specimen from a patient with no history of atypia or malignancy was included in all experiments as a positive control [11,12]. Four cold-cut biopsies were obtained away from the trigone of the bladder of patients with IC, diagnosed according to published specifications [17]. Biopsies were similarly obtained from patients with clinically confirmed urge UI secondary to IDO and SUI secondary to urodynamic stress incontinence, as defined by the ICS [18]. There was no statistically significant difference in the mean (range) age of patients for each group, i.e. for IC, IDO and SUI, at 51 (25–67), 47 (27–71) and 52 (38–80) years, respectively. Biopsies were collected in 10 mL of transport medium, consisting of sterile Hanks’ Balanced Salt Solution with Ca2+ and Mg2+ (Invitrogen Ltd, Paisley, UK) containing 10 mm HEPES (pH 7.6) and 20 KIU/mL aprotinin (Trasylol; Bayer plc., Newbury, UK). One biopsy was fixed for 1 h in 10% (v/v) formalin and processed into paraffin wax for immunohistology; the remaining three biopsies were pooled and used to establish primary urothelial cell cultures by scaling down and adapting methods developed for resection specimens (see below).

Haematoxylin and eosin-stained sections from all three patient groups were assessed to exclude carcinoma in situ and other inflammatory conditions causing identical symptoms [19], and to assess urothelial integrity for immunolabelling.

For immunohistochemistry, cold-cut biopsies were fixed in 10% (v/v) formalin in PBS for 1 h, dehydrated in ethanol to isopropanol then xylene, before careful orientation and embedding in paraffin wax. This protocol was strictly observed, to best preserve superficial cells and avoid potential processing artefacts caused by over-fixation. Haematoxylin and eosin-stained sections were screened for areas of full-thickness urothelium for analysis of differentiation antigen expression.

Immunohistochemistry was performed on dewaxed 5 µm sections using the StreptABComplex/horseradish peroxidase system from Dako Cytomation (Ely, UK), as previously described [15]. Primary antibodies are listed in Table 1. Blocking steps were included to neutralize any endogenous peroxidase and avidin-binding activities, and to prevent non-specific binding of secondary antibody. For most antibodies, sections were boiled for 10 min in 10 mm citric acid buffer, pH 6.0 to retrieve antigens lost during tissue processing. The exception was for retrieval of the CK20 antigen, where sections were digested for 10 min in 0.1% (w/v) trypsin in 0.1% (w/v) CaCl2, pH 7.6 at 37 °C. After overnight incubation in primary antibody at 4 °C, slides were washed, incubated sequentially in biotinylated secondary antibodies and a streptavidin biotin-horseradish peroxidase complex (Dako Cytomation), and visualized using a diaminobenzidine substrate reaction (Sigma-Aldrich Ltd, Poole, UK). The sections were counterstained with haematoxylin, dehydrated and mounted in DPX (Sigma-Aldrich). Negative and positive antibody specificity controls were included in all experiments.

Table 1.  The antibodies used
AntigenAntibodyHostConcentration, µg/mL or ratio Antigen retrieval for IHCSource (catalogue no.)
WBIIFIHC
  1. WB, Western blot; IIF, indirect immunofluorescence; IHC, immunohistochemistry; MW, microwave retrieval; ICRF, Imperial Cancer Research Fund.

Ki67MIB-1 (clone MM1)Mouse 0.5MWNovocastra (NCL-Ki67-MM1)
Uroplakin-3aAU1Mouse 1/40MWProgen Biotechnik (651108)
Occludinanti-occludinRabbit  2.5 0.25MWZymed (71–1500)
Claudin 1JAY.8Rabbit0.25  2.5Zymed (51–9000)
Claudin 3Z23.JMRabbit0.22  2.2 1.1MWZymed (18–7340)
Claudin 43E2C1Mouse0.5  2.0 1.0MWZymed (32–9400)
Claudin 54C3C2Mouse0.5  4.010MWZymed (18–7364)
Claudin 7ZMD.241Rabbit0.25  2.0 2.5MWZymed (34–9100)
CK7LP1KMouse 1/2000MWGift from ICRF, London
CK13IC7Mouse2 12.5Abcam (ab22685)
CK14LL002Mouse1200Serotech (MCA890)
CK20Ks20.8Mouse 0.07trypsinNovocastra (NCL–CK20)
CK20IT–Ks20.3Mouse  0.125Cymbus (61032)
PPARγP&A53.25Mouse  4.8Gift from GSK
β-actinAC15Mouse0.048Sigma (A5441)
RXRαD20Rabbit  4Santa Cruz (sc−553)

For urothelial biopsy cell culture, biopsies were centrifuged in the original transport medium for 4 min at 61 g to salvage any shed urothelial cells. To separate the urothelium from the stroma, pellets were resuspended in 3 mL of Hanks’ Balanced Salt Solution (with no Ca2+ and Mg2+) containing 10 mm HEPES, pH 7.6, 20 KIU/mL aprotinin and 0.1% (w/v) EDTA, and incubated for 3 h at 37 °C. Pellets were collected by centrifugation, flicked to resuspend, and any stromal pieces removed with sterile forceps. The remaining urothelium was incubated in 1 mL collagenase type IV (200 U/mL, Sigma-Aldrich) for 20 min at 37 °C. After centrifugation at 61 g for 4 min, media were aspirated and pellets resuspended in 2 mL Keratinocyte Serum-Free Medium-complete with bovine pituitary extract and EGF, at the manufacturer’s recommended concentrations (Invitrogen), with 30 ng/mL of cholera toxin (Sigma-Aldrich). Cell suspensions were seeded into a 3-cm Primaria-coated dish (BD Biosciences, Oxford, UK) and the medium was renewed every 2–3 days.

Biopsy-derived patient urothelial cell cultures were maintained at 37 °C in a humidified atmosphere of 5% CO2 in air and passaged at imminent confluence, exactly as detailed previously [11,12]. Cultures were transferred from a 3-cm Petri-dish (passage 0) to a 1 × 25 cm2 flask (passage 1), to 4 × 25 cm2 flasks (passage 2), of which one flask was cryopreserved and the other three transferred into two flasks each and used for analysis of differentiation potential (passage 3). All experiments were performed with a bladder NHU cell line (Y607) which was established from a resection specimen and included as a positive control.

The differentiation assay was based on the optimized conditions for inducing uroplakin gene expression in NHU cell cultures, as described previously [14]. Control NHU cell lines were grown to 70% confluence and treated with or without 1 µm troglitazone for 24 h, followed by incubation with or without 1 µm PD153035 for up to 6 days. Due to the limited availability of cells from biopsy-derived urothelial cell lines, experiments were constrained to: (a) 1 µm troglitazone for 24 h followed by 1 µm PD153035; and (b) no treatment control (0.01% v/v DMSO). Media were changed every 3 days and cells were harvested for RNA (day 3) or protein (day 6), as with the NHU cell controls. These sample times were selected as optimal from our previous work [14]. Cells seeded on 12-well PTFE-coated slides and treated as above were used for immunofluorescence experiments.

For indirect immunofluorescence, slides were fixed in methanol and acetone (v/v), air-dried and incubated with titrated primary antibody (Table 1), or no primary antibody as control, for 16 h at 4 °C. After washing in PBS, the slides were incubated with secondary antibodies conjugated to Alexa 488 (Molecular Probes, Invitrogen). To visualize nuclei, Hoechst 33258 (0.1 µg/mL; Sigma-Aldrich) was included in the penultimate wash. Slides were examined by light microscopy under epifluorescence illumination.

For Western blot analysis, cells grown in 25 cm2 flasks were treated in situ with lysis buffer (25 mm HEPES pH 7.4, 125 mm NaCl, 10 mm NaF, 10 mm Na3VO4, 10 mm Na4P2O7, 0.2% (w/v) SDS, 0.5% (w/v) sodium deoxycholate, 1% (w/v) Triton X-100, 1 µg/mL aprotinin, 10 µg/mL leupeptin and 100 µg/mL phenylmethylsulphonyl fluoride) and the lysates were sheared by passing three times through a 21-G needle. After 30 min on ice, lysates were microcentrifuged at 10 000 g for 30 min at 4 °C before the protein concentrations of supernatants were determined by Coomassie assay (Pierce, supplied by Perbio Science UK Ltd, Cheshire, UK). Cell extracts were resolved on 12.5% SDS polyacrylamide gels, transferred electrophoretically onto nitrocellulose membranes, and membranes were incubated with titrated primary antibodies (Table 1) for 16 h at 4 °C. Bound antibody was detected with anti-mouse Alexa Fluor® 680 (Molecular Probes) and anti-rabbit LI-COR IRDyeTM 800 (Rockland, supplied by Tebu-Bio Ltd, Peterborough, UK) and quantified using the OdysseyTM Infrared Imaging System (LI-COR Biosciences UK Ltd, Cambridge).

For reverse-transcribed (RT) PCR, RNA was extracted from cell monolayers using Trizol (Invitrogen Ltd) and isolated by chloroform extraction and isopropanol precipitation, as recommended by the manufacturer. The RNA was treated with DNase I to remove any DNA contamination (DNA-freeTM kit, Ambion Europe Ltd, Huntingdon, UK) and cDNA was synthesized using 1 µg of total RNA and the SuperscriptTM first-strand synthesis system (Invitrogen), as outlined by the manufacturer. PCR was performed as previously described, using Surestart Taq polymerase (Stratagene Europe, Amsterdam, the Netherlands) and primer pairs designed to amplify specific claudin products, as described previously [2]. Based on previous work [20], glyceraldehyde phosphate dehydrogenase (GAPDH) was included as the internal transcript control, using a forward primer: 5′-GTCGGAGTCAACGGATTTGG-3′, and reverse primer: 5′-ATTGCTGATGATCTTGAGGC-3′. PCR reactions were as follows: 95 °C for 12 min, then 35 cycles of 95 °C for 1 min, optimum annealing temperature for 1 min and 72 °C for 1 min, and a final extension at 72 °C for 10 min in a PCR Express Thermal Cycler (Hybaid Ltd, Ashford, UK). All experiments were done in parallel with no-template and no-RT controls. PCR products were run on a 2% (w/v) agarose gel and visualized with ethidium bromide.

For real-time RT-PCR, cDNA was synthesized as outlined above. Quantitative real-time PCR was performed using TaqMan real-time PCR primers and probes designed for uroplakin genes using the Primer Express Software (Applied Biosystems UK, Warrington) (Table 2). The reactions were done in TaqMan Universal PCR master mix (Applied Biosystems) with 200 nm of probe and 300 nm of primers on an ABI Prism 7700 Sequence Detector system (Applied Biosystems) for 10 min at 95 °C, followed by 40 cycles of 15 s at 95 °C and 60 s at 60 °C. The data were analysed using the ABI Prism 7700 SDS software. Data were normalized against GAPDH, used as the internal control [20].

Table 2.  The Taqman PCR primers and probes
GeneDirectionSequence (5′−3′)
  1. UPK, uroplakin.

UPK1aforwardCATTCTTGCTGAACCGTTTGT
reverseGTGACCGTGACAGAACTCTCATG
probeTTGCCATTTGAGCTCTGGAAGCCTCTATT
UPK1bforwardCGCTTGCCTTCAGCTTGTG
reverseGGCCCTGGAAGCAACGA
probeCCCGAAGATGGCCAAAGACAACTC
UPK2forwardCAGTGCCTCACCTTCCAACA
reverseTGGTAAAATGGGAGGAAAGTCAA
probeTCCATTATTCCTCTCACCCCACTCCTGTC
UPK3aforwardCGGAGGCATGATCGTCAT
reverseCAGCAAAACCCACAAGTAGAAAGA
probeCTTCCATCCTGGGCTCCCTGCC
UPK3bforwardCCTCCTGCTTCACTCTCTCTGTCT
reverseGAAACTGACAATCACGGCAGAA
probeCCAGAAACGTGCCTGCTTCCCCTT
GAPDHforwardCAAGGTCATCCATGACAACTTTG
reverseGGGCCATCCACAGTCTTCTG
probeACCACAGTCCATGCCATCACTGCCA

Means and medians were used as descriptive statistics and plotted as solid and dashed lines, respectively. Nonparametric methods (two-tailed Mann–Whitney U-test or the two-tailed Wilcoxon matched-pairs signed-ranks test, as appropriate) were used for tests of statistical significance. On graphs, the level of significance is indicated between marked groups as *P ≤ 0.05; **P ≤ 0.01 and ***P ≤ 0.001.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The urothelial thickness on biopsies was two to six cell layers. Focal denudation of the urothelium was not accompanied by underlying granulation tissue or oedema, and was regarded as artefactual. The IC samples had variable dense non-specific chronic inflammation in the lamina propria. No granulomatous or eosinophilic inflammation was seen in any patient group.

Assessment of the in situ differentiation status of urothelium from patient biopsies was limited to biopsies that on histological examination had an intact, full-thickness urothelium. Sections from IC, IDO and SUI biopsy specimens were screened with a panel of differentiation-associated antigens (Fig. 1). In keeping with adult NHU (ureter control tissue), the urothelia from all patients were mitotically quiescent, with very few cells in the cell cycle, as assessed by an absence of Ki67 (MIB-1) labelling.

image

Figure 1. Immunohistochemistry; immunoperoxidase labelling of paraffin wax-embedded sections of IC, IDO and SUI, compared to normal ureter, as control. A panel of markers was used to assess the proliferation and differentiation status of urothelium in situ. Note the patchy presence of superficial cells in biopsy specimens. Two different IC biopsies for claudin 5 are shown, illustrating the variability of the labelling. In addition to urothelial reactivity, claudin 5 also labelled the vascular endothelium, which served as an internal positive control. Note that the stromal reactivity apparent with the rabbit anti-claudin 7 antibody is artefactual. Scale bar 50 µm.

Download figure to PowerPoint

Two CK isotypes were investigated; CK7, which is expressed in all urothelial cell layers, and CK20, which labels superficial and occasional intermediate cells in normal urothelium [21]. In most biopsies, CK7 showed a heterogeneous labelling pattern, with patches of intense full-thickness CK7 labelling, interspersed by patches of weak labelling; this phenomenon was not related to clinical derivation and was described previously in normal bladder urothelium [22]. CK20-positive cells were identified in all specimens, primarily located at the superficial edge, but often with the morphology of late intermediate cells, suggesting either a loss or absence of mature superficial cells. Uroplakin-3a was localized discretely along the apical edge of superficial urothelial cells in intact urothelium, but the extent of uroplakin-3a labelling in patient biopsies varied from near normal to absent and, as with the CK20 labelling, reflected either the absence or artefactual shedding of superficial cells.

The expression and location of tight junction-associated proteins, occludin and the claudins 4, 5 and 7, was very similar to that in ureteric urothelium [2], once allowance was made for the absence or loss of some superficial cells (Fig. 1). Claudins 4 and 5 were most intensely expressed at basolateral junctions of superficial cells, and were useful discriminatory markers for the presence or absence of superficial cells (Fig. 1). One IC sample showed no claudin 5 expression in the urothelium, despite claudin 5 expression in the vascular endothelium and the confirmed presence of uroplakin- 3a-positive superficial cells in that sample (Fig. 1). Claudin 3 is restricted to the terminal junction between superficial urothelial cells, but appeared negative in all biopsy specimens, although this interpretation was confounded by the loss/absence of superficial cells.

The results from the immunohistochemistry studies indicated no gross differences between IC urothelium and the other benign conditions (IDO and SUI), with the possible exception of claudin 5 down-regulation in one IC specimen. However, the differentiated phenotype could only, by definition, be assessed in areas of full-thickness urothelium. To determine more specifically the differentiation and function potential of urothelium from IC patients, a cell culture system was developed.

Urothelial cells isolated from each of three biopsies were used to establish successful primary urothelial cell cultures from six IC, five IDO and 12 SUI patients (Table 3). In a further two cases of IDO and one case of SUI, primary culture was unsuccessful because too few urothelial cells were isolated. One additional IC cell line was established from tissue obtained at cystectomy (referred to as IC1). In all cases, primary cultures had a typical epithelioid pavement morphology, comparable to NHU cell cultures established from bladder resection specimens (Fig. 2). In most cases the primary cultures grew to confluence and were successfully subcultured as finite cell lines through a minimum of three passages, generating adequate cell numbers to assess differentiation and functional potential (Table 3, and below). The exceptions were two SUI cell lines, which were growth compromised and failed to grow beyond passage 2; these were not characterized further. By passage 3, cell lines from one IC, one IDO and three further cases of SUI had begun to show some evidence of senescence, with reduced cell growth and the appearance of larger, mitotically quiescent cells (Table 3 and Fig. 2).

Table 3.  A summary of the results for individual cell lines
SampleCell line IDGrowthResponse to troglitazone + PD153035
CK13/CK14 switchUPK gene up-regulationClaudin protein change*
  • *

    Claudins; PR, partial response (defined as no up-regulation of claudin 3); NR, no response (defined as no up-regulation of claudins 3, 4 and 5).

  • Cell line derived from cystectomy specimen; +, normal response; LR limited response; NR no response; –, no data; r, troglitazone-induced rosettes; S2, cells senesced and failed to grow beyond passage 2; S3, some evidence of senescence at passage 3.

NHUcontrol+r+++
ICIC1+r+++
IC2S3LR+PR
IC3+r+++
IC4+NR
IC5++++
IC6+NRNRNR
IC7+LRPR
IDOIDO1+r+++
IDO2+r+++
IDO3S3LRPR
IDO4++
IDO5+rLR++
SUISUI1S3LR+PR
SUI2S3LRLRPR
SUI3+rNR++
SUI4S2
SUI5S2
SUI6+LRLRPR
SUI7+rLR+PR
SUI8+LR++
SUI9+r+++
SUI10+r+++
SUI11S3LRNR
SUI12++LRPR
image

Figure 2. Phase-contrast micrographs showing typical examples of cultures at passage 3 from the bladder-derived NHU cell line control, compared to biopsy-derived IDO, IC and SUI cultures in control growth medium (left panel) and after treatment for 3 days with troglitazone (TZ, 1 µm) and PD153035 (1 µm) (right panel). Arrows indicate the differentiation-induced formation of rosettes. The third panel (IC2) shows a typical senescent culture containing larger quiescent cells. Scale bar 30 µm.

Download figure to PowerPoint

We showed previously that cultured NHU cells undergo a programme of differentiation-associated gene expression changes in response to activation of PPARγ, and that this response is enhanced when signalling downstream of the EGFR is blocked [14]. Immunofluorescence labelling of the cystectomy-derived IC1 cell line confirmed that PPARγ and retinoid X receptor α (RXRα) showed an equivalent location and intensity to that in the NHU bladder control cell line, with both antigens showing a predominantly nuclear location (Fig. 3).

image

Figure 3. Localization of PPARγ and RXRα in NHU and IC urothelial cells. Bladder-derived Y607 NHU cells and cystectomy-derived IC1 urothelial cells were seeded at 3 × 105 cells/mL onto glass slides, grown to near-confluence and fixed in methanol : acetone. Indirect immunofluorescence was used for PPARγ and RXRα and showed a predominantly nuclear location pattern. Scale bar 10 µm.

Download figure to PowerPoint

To determine whether urothelial cells from IC, IDO and SUI patients retained the capacity to differentiate, cells were treated with the PPARγ agonist, and the EGFR-specific tyrosine kinase inhibitor, PD153035. The NHU bladder cell line control and a proportion of cultures from all three patient groups showed morphological evidence of response to treatment with troglitazone and PD153035, forming rosettes of tear-shaped cells within 3 days of treatment (Table 3 and Fig. 2, arrows), as described previously for NHU cell cultures [14].

Treatment with troglitazone and PD153035 was shown previously to switch NHU cells from a squamous to a transitional differentiation programme, as assessed by the shift in CK expression from a CK13lo/CK14hi to a CK13hi/CK14lo profile [15]. This was confirmed in the bladder-derived NHU cell line by immunoblotting (Fig. 4 and Table 3). Characterization of IC-derived cell lines showed induction of CK13 expression and decrease in CK14 expression in response to troglitazone/PD153035 in three of six lines, but a much reduced (two) or absent (one) response in the remaining three cell lines tested. Of the IDO cell lines tested, two of four showed an induction of CK13 and decrease of CK14. In most SUI-derived cell lines (seven of 10), the expression of CK13 was either not induced or induced only marginally, although in eight there was a reduction in CK14 expression in response to treatment (Fig. 4 and Table 3).

image

Figure 4. Influence of differentiation on CK13 and CK14 expression in cultured IC, IDO and SUI urothelial cells. A, urothelial cell cultures from IC (six), IDO (four) and SUI (10) lines were incubated in the presence or absence of troglitazone (1 µm) for 24 h, followed or not by PD153035 (1 µm) for 6 days. A normal bladder NHU cell line was included as a positive control. Media were changed every 3 days. Protein was extracted and 20 µg was resolved on 12.5% SDS polyacrylamide gels and transferred to nitrocellulose membranes. Bound primary antibody was detected with fluorescent-conjugated secondary antibodies and detected using the Li-CoR system. A representative Western blot is presented in the figure and illustrates both normal and atypical responses in terms of a switch from a CK13lo/CK14hi to a CK13hi/CK14lo phenotype. B, Plots of individual cell lines analysed by Western blot analysis for expression of CK13 and CK14 proteins from undifferentiated (U) cells, and cells after differentiation-inducing (D) treatment with troglitazone and PD153035 for IC, IDO and SUI cultured urothelial cells. The cystectomy-derived IC1 cell line is indicated by a dashed line. Means are indicated by solid lines and medians by broken lines. The lysate from bladder NHU cell line (Y607) was included in each Western blot and was used as an internal normalization control. Troglitazone and PD153035-treated Y607 lysate was taken as 1.0. There was no statistical significance between IC and IDO or SUI groups.

Download figure to PowerPoint

The induction of differentiation in NHU cells with troglitazone and PD153035 is associated with de novo expression of genes associated with urothelial terminal differentiation, including the uroplakins [14]. The expression of uroplakin-1a, uroplakin-1b, uroplakin-2 and uroplakin-3a transcripts was quantified by real-time PCR in the IC, IDO and SUI cell lines, and compared to the control bladder NHU cell line (Fig. 5). In all three patient groups there was significant greater uroplakin gene expression in the differentiated than undifferentiated conditions, and there was a significantly greater expression of uroplakin-1b in differentiated SUI than the IC groups (Fig. 5). When assessed in terms of individual samples, two of the IC cell lines showed a consistent lack of induction of uroplakin-1a, uroplakin-2 or uroplakin-3a gene expression in response to troglitazone and PD153035, whereas the remaining IC lines showed induction of uroplakin-1a and uroplakin-2 to at least half that of the NHU cell line control (Fig. 5 and Table 3). All of the IDO cell lines responded to some degree, but three of nine SUI cell lines showed only limited up-regulation of uroplakin-1a, uroplakin-2 and uroplakin-3a expression after treatment (Fig. 5).

image

Figure 5. The effect of troglitazone and PD153035 on uroplakin (UPK) expression in cultured IC, IDO and SUI urothelial cells. Urothelial cell cultures from IC (six), IDO (four) and SUI (nine) lines were treated with or without troglitazone (1 µm) and fresh medium containing PD153035 (1 µm) was added after 24 h. After a further 3 days, RNA was extracted, cDNA generated and quantitative PCR used, using GAPDH as the internal control. The cystectomy-derived IC1 cell line is indicated by a dashed line. The results are shown as individual line plots and expressed relative to the treated bladder-derived Y607 NHU cell line, which was included as a control. There was a significant induction of uroplakin expression in differentiated (D) vs undifferentiated (U) conditions for all three patient groups, although note the individual non-responder cell lines. The induction of uroplakin-1b expression was significantly greater in the SUI than the IC group.

Download figure to PowerPoint

One of the features of urothelial cytodifferentiation is the formation of intercellular tight junctions associated with changes in the transcription and protein stability of the claudins [2]. In agreement with our previous study using ureter-derived NHU cells, the bladder-derived NHU cell line expressed claudins 1, 2, 4 and 5, and treatment with troglitazone and PD153035 resulted in induction of claudin 3 (Fig. 6A). Claudin transcript expression was examined by RT-PCR in three IC and two IDO cell lines. All showed constitutive expression of claudins 1, 2 and 4, and baseline expression of claudin 3 was detected in two IC and both IDO cell lines. There was some evidence of up-regulation of claudin 3 in response to treatment with troglitazone and PD153035 in one of two IC cell lines and in both IDO cell lines examined by RT-PCR (Fig. 6A).

image

Figure 6. The influence of troglitazone (TZ) and PD153035 on claudin expression in cultured NHU, IC and IDO urothelial cells. A , Claudin transcript expression. Cultures of IC (three) and IDO (two) cells were treated for 24 h in the absence or presence of troglitazone (1 µm) and then in medium with or without PD153035 (1 µm). RNA was extracted at 3 days, cDNA was generated and RT-PCR used, with claudin primers and GAPDH as the internal control. The PCR products were electrophoresed on a 2% agarose gel and visualized using ethidium bromide. B , Western blot analysis. Urothelial IC, IDO and SUI cell cultures were treated with or without troglitazone (1 µm) for 24 h and then in medium with or without PD153035 (1 µm) for 6 days. Medium was changed every 3 days with PD153035, as appropriate. Protein (20 µg) was resolved on 12.5% SDS polyacrylamide gels and transferred onto nitrocellulose membranes. Bound primary antibody was detected with fluorescent-conjugated secondary antibodies and quantified using the LI-COR system. The bladder-derived Y607 NHU cell line was included in all blots as a control. A representative Western blot is shown, which for comparison purposes has the same samples shown in Fig. 4. C, Individual line plots of Western blot data for claudin protein expression, indicating relative change in claudin expression in response to treatment with troglitazone and PD153035 for IC, IDO and SUI cultured urothelial cells. The result for the Y607 bladder NHU cell line was taken as 1.0. The cystectomy-derived IC1 cell line is indicated by a dashed line. Means are indicated by solid lines and medians by broken lines. Induction of claudin 4 and 5 expression with differentiation was significant for claudin 4 and 5 in all three patient groups and for claudin 3 in the SUI group, but note the presence of individual poor responder cell lines. There was no statistical significance between IC and IDO or SUI groups.

Download figure to PowerPoint

At the protein level, treatment of control bladder NHU cells with troglitazone and PD153035 resulted in large increases in claudins 4 and 5, with a small induction of the late differentiation marker claudin 3 (Fig. 6B,C); this was in agreement with our previous study of ureteric NHU cell lines [2]. In addition, claudins 1 and 7 showed some change in expression, which were not significant in ureteric NHU cell lines [2]. The pattern was similar in the patient groups, with a significantly greater expression of claudins 4 and 5 in response to differentiating conditions, although claudin 5 was poorly induced in some individual cell lines (Tables 3 and 4). Induction of claudin 3 was more variable and was absent in three of six IC, one of four IDO and six of 10 SUI cell lines (Fig. 6B,C; Table 3). In samples where there was a failure to induce claudin 3, the induction of claudin 5 was correspondingly weak.

Table 4.  Changes in claudin (cl) expression after induction by troglitazone and PD153035
SampleCell line IDcl1cl3cl4cl5cl7Summary
  • *

    Cell line derived from cystectomy specimen; =, no change; – negative; [UPWARDS ARROW]up-regulation; [DOWNWARDS ARROW]decrease; ( ) minor change.

NHUcontrol[DOWNWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]=Normal
ICIC1*[DOWNWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]Normal
IC2=[UPWARDS ARROW]([UPWARDS ARROW])[UPWARDS ARROW]weak cl5 no cl3
IC3=[UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]Normal
IC5=[UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]Normal
IC6=([UPWARDS ARROW])=No response
IC7=[UPWARDS ARROW]([UPWARDS ARROW])[UPWARDS ARROW]weak cl5 no cl3
IDOIDO1=[UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]Normal
IDO2[DOWNWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]Normal
IDO3=[UPWARDS ARROW]([UPWARDS ARROW])[UPWARDS ARROW]weak cl5 no cl3
IDO5=[UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]=Normal
SUISUI1=[UPWARDS ARROW]([UPWARDS ARROW])[UPWARDS ARROW]weak cl5 no cl3
SUI2=[UPWARDS ARROW]([UPWARDS ARROW])[UPWARDS ARROW]weak cl5 no cl3
SUI3=[UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]=Normal
SUI6=[UPWARDS ARROW]([UPWARDS ARROW])=weak cl5 no cl3
SUI7([UPWARDS ARROW])[UPWARDS ARROW]([UPWARDS ARROW])[UPWARDS ARROW]weak cl5 no cl3
SUI8=[UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]Normal
SUI9([UPWARDS ARROW])[UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]=Normal
SUI10=[UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]Normal
SUI11===No response
SUI12=[UPWARDS ARROW][UPWARDS ARROW][UPWARDS ARROW]no cl3

In the IC1 cell line, which showed a normal claudin expression pattern, the response to troglitazone and PD153035 was associated with the relocation of tight junction-associated proteins to intercellular borders, as seen in the control bladder NHU cells (Fig. 7).

image

Figure 7. Immunofluorescence showing the effect of differentiation on the expression and localization of CK13 and claudin proteins in cultured urothelial cells. Bladder-derived Y607 NHU and cystectomy-derived IC1 urothelial cells were seeded at 2 × 105 cells/mL onto glass slides, allowed to adhere and then treated with troglitazone (TZ, 1 µm) for 24 h, then PD153035 (1 µm), vs a no-treatment control; slides were fixed after 6 days. Media were replaced every 3 days. Immunofluorescence was performed using the antibodies indicated and nuclei were counterstained with Hoechst 33258 (shown for the claudin 7 micrograph). Note the nuclear location of claudin 1 in untreated cultures, which might be a consequence of low E-cadherin expression by proliferating cell cultures [23]. Scale bar, 20 µm.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

This is the first report of the application of an in vitro cell culture system for NHU cells to the study of dysfunctional bladder syndromes. Because clinical considerations limit the amount of urothelium that can be harvested by cystoscopic biopsy, a modified method for the isolation and in vitro expansion of urothelial cells from very small amounts of starting material has been developed. We considered it critical that control groups for the study should be derived from similarly harvested tissues, and due to ethical constraints preventing the biopsy of normal bladders, this led us to a decision to compare IC to two other benign dysfunctional bladder syndromes, IDO and SUI.

Objective comparison of the in vitro differentiation potential of urothelial cells from different patient groups, especially when selected by stringent clinical criteria, might reveal disease-specific differences and hence provide a path to identifying the underlying mechanisms. However, diseases such as IC are defined primarily by symptoms, and might reflect convergent progression pathways from more than one causal mechanism. Several causal mechanisms for IC can be postulated. A leaky urinary barrier giving rise to symptoms of cystitis could be caused by an inherent failure of urothelial cytodifferentiation, but alternatively could reflect a response to exogenous factors derived from, e.g. the immunological microenvironment, having an influence on urothelial tissue integrity. To determine the influence of intrinsic factors, we used an in vitro approach wherein extrinsic factors are absent, or introduced experimentally under controlled conditions. Although the present study did not address all questions and will need to be extended to include larger sample groups, nevertheless, we think it offers a new platform for understanding the pathogenesis of benign bladder diseases, and will open the door to identifying biomarkers to differentiate disease subsets based on mechanism. This in turn will lead to the development of therapies aimed at causes rather than symptom relief.

We described previously the culture of NHU cells from resection specimens. Despite being derived from a mitotically quiescent tissue in situ, cultures of NHU cells have a remarkable, albeit finite, proliferative capacity, reflecting the regenerative capacity of native urothelium [13]. Although the biopsies were derived from diseased bladders, there was little evidence of proliferation being driven in situ, as seen from the lack of Ki67-positive cells by immunohistochemistry. This was even the case in the IC-derived specimens, where chronic inflammatory-mediated tissue damage might be expected to initiate regenerative field changes. The small starting cell population obtainable from biopsies limited ultimate expansion capacity compared to either the bladder resection-derived NHU cell line included as control, or the cystectomy-derived IC1 cell line. Nevertheless, urothelial cells derived from most biopsies showed a proliferative phenotype in culture, enabling adequate expansion through limited serial passage to enable controlled differentiation studies. An unexpected finding was that cell lines showing premature senescence were mostly derived not from the IC, but from the SUI group. This difference was not due to donor age per se, as there was no age-related pattern of senescence (mean age of donors for non-senescent and senescent cultures was 50 and 51 years, respectively).

In our previous studies we identified a PPARγ-mediated pathway that, in the absence of EGFR signalling, initiates the programmed expression of gene and protein changes associated with urothelial cytodifferentiation, including characteristic changes to the CKs [15], claudins [2] and uroplakins [14]. From this programme, we selected several components to use as objective markers of differentiation in response to PPARγ activation, i.e (a) morphology (b) switch from a CK14 squamous to a CK13 transitional phenotype; (c) induction of claudin 3, 4 and 5 protein expression; and (d) induction of uroplakin gene transcription.

A morphological change from a homogeneous squamous pavement morphology to the formation of rosettes of ‘tear-shaped’ cells of characteristic transitional morphology provided an indicator of a PPARγ-mediated response and occurred in half the biopsy-derived cultures that showed no senescent change. However, it was not an absolute predictor of response to PPARγ, as there was at least one cell line (IC5) that showed no morphological change, but still showed responses at the molecular level. Most biopsy-derived cell lines showed some molecular evidence of response to PPARγ, although this response was limited in the case of three of the IC-derived cell lines (IC4, IC6 and IC7).

As markers, the claudins provided particular insight. In situ, the claudins show a stage-related pattern of expression in urothelium, with claudins 4 and 5 expressed at a later stage of differentiation than claudin 7 and occludin, but before claudin 3, which, together with ZO-1, is localized in the terminal junction between adjacent superficial cells [2]. In vitro, PPARγ activation results in de novo expression of claudin 3 and the stabilization of claudins 4 and 5 proteins [2]. Except for one IC and one SUI cell line, all biopsy-derived cell lines showed up-regulation of claudins 4 and 5, but detection of claudin 3 was limited to cases where there was a strong up-regulation of claudin 5 protein. Claudin 3 and ZO-1 can be difficult to locate in tissue sections, and the absence/loss of superficial cells from biopsies restricted the use of these markers. However, at least one of the informative IC cases was negative for claudin 5 in situ. Given the probability that IC has no single cause, we suggest that there might be a subset of IC cases in which there is aberrance in the formation of tight-junction structures during urothelial differentiation, which would be predicted to affect paracellular permeability. Future studies using a more sophisticated three-dimensional differentiation model [24] are proposed, as these could combine functional paracellular permeability assessment with claudin expression profiling, and hence might provide functional evidence to support claudin dysregulation in the aetiopathology of a subset of IC. However, we also emphasize caution for follow-up studies, as we have found many tissue blocks from hospital archives to show artefactual loss of claudin antigenicity, possibly due to poorly controlled fixation conditions (unpublished observations).

Although urothelial dysfunction has been implicated in IC and IDO, there has been no previous suggestion of urothelial involvement in the aetiopathology of SUI, which was therefore presumed to represent a ‘normal’ urothelium. The unexpected finding from the present study was that urothelial cells from several patients with SUI showed poor growth and differentiation characteristics. Previous research has mainly focused on the diagnosis and treatment of SUI, but the role of nerve damage in the aetiology of the disorder is attracting interest, as it offers a potential therapeutic target. It was suggested that the purinergic signalling pathway might be involved in functional motor and sensory bladder disorders [25]. It is reasonable to hypothesise that interaction between the neuromuscular elements of the bladder wall and the urothelium might play a role in more than just bladder dynamics and bladder sensation, and that dysregulation of this intimate relationship might underlie the pathophysiology of bladder dysfunction syndromes that do not appear to include a primary sensory element. Further work to characterize the purinergic receptors present in human urothelium in normal and diseased states would assist in defining the precise relationship between urothelial sensation and function.

In summary, a novel in vitro approach was developed to investigate the growth and differentiation potential of urothelium from dysfunctional bladder syndromes, and provides a platform for investigating the causal mechanisms in IC, IDO and SUI. Our results suggest that there might be a subset of IC cases in which the differentiation capacity of the urothelium is compromised, possibly through derangement of tight-junction structure, which would give rise to a leaky urothelium and associated symptoms of chronic cystitis.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The authors thank Professor J.J. Walker for helpful discussions. They are extremely grateful to Drs C. Ramage, L. Rogerson and G. Urwin, who assessed patients and provided clinical specimens for the study. J.S. holds a Research Chair funded by York Against Cancer. Contract grant sponsor: NIH; Contract grant number: IR21 DK066075.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES