Normal urothelium serves as the main source of a major class of soluble, urine proteins including urokinase, tissue-type plasminogen activator, and a potent serine protease inhibitor, PP5 [1]; their synthesis requires differentiation of the vesical urothelium. It is speculated that these urinary proteins and their inhibitors play a crucial role in regulating desquamation of urothelium, which constitutes a vital defence mechanism against bacterial attachment. Thus the mammalian urothelium, besides acting as a permeability barrier, may serve important functions in prevention of bacterial cystitis.

Urothelium is characterised by terminally differentiated superficial cells ("umbrella cells"), that express uroplakins in their luminal plasma membrane and cytokeratin 20 in their cytoplasm. Basal and intermediate cells are cytokeratin 20 negative. [2] In contrast to normal urothelium, cultured human stratified urothelium does not undergo complete terminal differentiation of its superficial cells. Using immunohistochemical staining with specific antibodies, the superficial layer of reconstructed stratified urothelium has been shown to express uroplakins but not cytokeratin 20 [3–5].

We studied cytokeratin-20 immunostaining of bladder biopsies taken from spinal cord injury patients. It is possible that injury to the spinal cord, and consequent lack of trophic effect upon the urothelium, may lead to incomplete maturation and differentiation. [6]. As normal urothelium is characterised by cytoplasmic cytokeratin 20 expression in terminally differentiated superficial cells, cytokeratin 20 was taken as a marker of complete terminal differentiation of urothelial cells in bladder biopsies. If spinal cord injury does indeed affect urothelial differentiation or induces squamous or other metaplastic change undetected by histological analysis, the superficial urothelial cells in the neuropathic bladder of SCI patients may not show positive immunostaining for cytokeratin 20.

Bladder biopsies were obtained from 63 spinal cord injury patients after obtaining written informed consent. The North Sefton Local Research Ethics Committee approved this study. All were adults with spinal cord injury and neuropathic bladder. They were registered with the Regional Spinal Injuries Centre, Southport, England; were not suffering from acute urinary infection and were undergoing an elective therapeutic procedure in the urinary tract such as endoscopic lithotripsy of bladder stone, insertion of a ureteric stent, or diagnostic cystoscopy.

After routine prophylactic intravenous gentamicin, cold cup biopsies of the bladder mucosa were taken from the trigone of the urinary bladder. Thereafter, the biopsy site was fulgurated with diathermy to achieve haemostasis. Indwelling urinary catheter drainage was maintained after the procedure, and all patients remained in hospital for at least 24 hours. Biopsies were fixed in neutral buffered formaldehyde 4%, and then embedded in Paraplast. (Paraplast is a commercial paraffin wax incorporating plasticisers, used for embedding formalin-fixed tissues for histology.)

A Consultant Cellular Pathologist recorded the histopathological features of each bladder biopsy after examining sections stained with haematoxylin and eosin. None of the 63 biopsies from spinal cord injury patients showed evidence of dysplasia or neoplasia. The classical morphological features of dysplasia in the urothelium comprise: increased nucleo-cytoplasmic ratio, nuclear hyperchromatism, abnormal chromatin pattern, prominent nucleoli, and nuclear pleomorphism; and loss of normal polarity leading to a haphazard spatial arrangement of urothelial cells. The affected urothelium is, by definition, flat rather than papillary. In the most severe form of dysplasia (carcinoma-in-situ), the above nuclear characteristics are present to a marked degree, comparable in severity to those in a grade 3 Transitional Cell Carcinoma. Lesser degrees are termed 'dysplasia not amounting to carcinoma-in-situ'. At the mild end of the spectrum, distinguishing dysplasia from inflammatory or regenerative atypia can be difficult; the abnormal chromatin pattern in dysplasia is usually the most reliable criterion.

In two biopsies, epithelium was scarce. Eight biopsies showed squamous metaplasia with no transitional epithelium present. Umbrella cells were identifiable in the remaining 53 biopsies at least in one microscopic field. Umbrella cells were identified by their morphological characteristics, namely: large, elongated cells with abundant eosinophilic cytoplasm, their long axes parallel to the basement membrane, covering several underlying intermediate urothelial cells. The umbrella cells may occasionally be binucleate.

Immunostaining for cytokeratin 20 was carried out on tissue sections in one lot using a mouse monoclonal antibody, clone Ks20.8 (Novocastra Laboratories Ltd, Balliol Business Park West, Benton Lane, Newcastle upon Tyne NE12 8EW, UK). This is an IgG antibody (IgG2a, kappa), with specificity for human cytokeratin 20 intermediate filament protein, raised against cytoskeletal preparations isolated from microdissected villi of human duodenal mucosa. The normal positive staining pattern for cytokeratin 20 is cytoplasmic. The manufacturer's instructions were followed for immuno-staining. Sections were pre-treated with proteinase-K, to unmask cytokeratin epitopes masked by the cross-linking effect of formaldehyde during fixation and tissue processing. The antibody was used at a working dilution of 1:25. Immunoperoxidase technique was employed, with primary antibody incubation at 25°C for 60 minutes. Each sample included a negative control, by omitting the primary antibody. Avidin Biotin Complex (ABC) kit was purchased from Vector Laboratories. The version of the kit used was Elite ABC.

The presence of immunostaining in the vesical epithelium was recorded either as negative or positive, and the location of any positively staining cells within the transitional epithelium was noted.

In an earlier study, we observed strong immunostaining for secretory immunoglobulin A (sIgA) in the urothelium of all biopsies taken from non-neuropathic bladders, but in only 18 of 37 biopsies from neuropathic bladders. [7]. This suggested that interruption of the nerve supply to the urinary bladder might be associated with alterations in urothelial differentiation and maturation. In the present study, we observed positive staining for cytokeratin 20, a marker of terminal urothelial differentiation, in only 10 of 53 biopsies taken from spinal cord injury patients, thereby providing further supporting evidence to that hypothesis.

Harnden and associates [8] noted positive immunostaining for cytokeratin 20 in the umbrella cells of all control biopsies, with staining restricted to the umbrella cells in 24 of 30 biopsies; in the remaining six cases, occasional dispersed intermediate cells also stained positively. Basal cells were consistently negative. A similar pattern of immunostaining was observed in the presence of hyperplasia and inflammation. In this study, immunostaining for cytokeratin 20 was absent in 43 of 53 biopsies taken from the neuropathic bladders of SCI patients. This absence of cytokeratin 20 immunostaining raises the question as to whether the transitional epithelium of the neuropathic bladder in SCI patients does not achieve complete functional differentiation, despite apparently complete morphological differentiation as judged by the presence of umbrella cells. Terminal differentiation of the urothelium plays a crucial role in desquamation of urothelium and hence in prevention of bacterial cystitis; "defective" or "incomplete" differentiation of bladder urothelium may explain why some SCI patients are at greater risk of developing bacterial cystitis.

There appears to be an absence of cytokeratin 20 in a high proportion of urothelium from neuropathic bladders. The reasons for this could be either that there is an underlying metaplasia or that changes in the neuropathic bladder affect urothelial differentiation. This question could be addressed by further analysis of other differentiation markers of urothelium, including cytokeratins and uroplakins. Taken with evidence from other systems, such as loss of cytokeratin 20 expression from static organ cultures of urothelial tissue [3], this might suggest that other factors, such as impairment of normal voiding function in SCI patients, could affect expression of differentiation markers such as cytokeratin 20.