Interstitial cystitis (IC) is being re-examined within a broader context of chronic pelvic pain and pelvic floor dysfunction with the result that it is being recognized that the prevalence of IC is much higher than previously estimated [1, 2]. It has long been evident that, whatever the cause, IC is a disease of the urothelium and the symptoms are triggered by urine [3]. In general, the bladder exhibits thinning and denudation of urothelium and distinctive changes in the muscle [4], and there is very strong evidence that the urothelium fails to differentiate properly in IC [5, 6]

One of the main therapeutic approaches is so-called "glycosaminoglycan (GAG) replacement therapy." This is partly based upon earlier observation that the so-called "GAG Layer" is lost or damaged in the interstitial cystitis bladder [6, 7]. The "GAG Layer" is a complex set of proteoglycans and glycoproteins on the bladder surface has been isolated and characterized from cow and human bladder [8–11]. Whether or not glycosaminoglycans actually restore impermeability of the bladder in interstitial cystitis is controversial [12–15]. Nonetheless, many patients respond to intravesical instillations of pentosan polysulphate (Elmiron) [16], hyaluronate (Cystistat) [17], chondroitin sulphate (Uracyst) [18]or heparin [19].

In this paper, we use three different models of bladder damage to determine whether GAG instilled intravesically actually binds preferentially to damaged bladder. We prepared a fluorescent-labeled chondroitin sulphate molecule by covalently attaching Texas Red to the reducing terminus of the GAG chain and then observing bladder sections by fluorescence to determine where the labelled GAG was found.

Texas Red-labelled chondroitin sulphate was prepared by mixing 10 μL Texas Red Hydrazide (Molecular Probes, Eugene, OR), 10 mM in dimethyl formamide, 40 μL of 1 mM chondroitin sulphate (Stellar Healthcare, London, Ontario), and 5 μL of 50% acetic acid, The chondroitin sulphate was assumed to be 20,000 molecular weight, yielding a concentration of 20 mg/ml as being equivalent to 1 mM. The reaction was allowed to occur overnight at room temperature. The next morning, the bond was reduced and made permanent by adding 10 μL of 150 mM sodium cyanoborohydride for 2 h. The labelled ChS was separated from unincorporated label by gel filtration on Biogel P-4.

Mice (C57BL/6NHsd, 9–10 weeks old) were anesthetized by intraperitoneal injection of ketamine 40 mg/kg and xylazine 5–10 mg/kg and catheterized with a 24 Ga × 3/4" intravenous catheter (Terumo Medical, Elkton, MD). Bladder damage was produced by 3 methods. (a) Protamine, 1 mg/ml, was instilled into the bladder until fullness and allowed to remain for 10 minutes. The protamine was then washed out with 3 fillings of buffered saline. (b) Dilute HCl, 10 mM was prepared by dilution in saline and was administered in the same way as described for the protamine except that a 0.1 M sodium bicarbonate wash was used as the first rinse;. (c) Trypsin, 1 mg/ml, was dissolved in saline and instilled into the mouse bladder where it was allowed to remain for 30 min. The trypsin was washed out with three fillings of buffered saline. Control animals were treated with buffered saline only. Three animals were in each group. The mouse bladders were then filled with Texas Red-labeled chondroitin sulphate, which was allowed to remain for 1 hr. The bladder was gently rinsed as described above. The animals were then euthanized by cervical dislocation with death being confirmed by opening the chest. The bladders were removed, flash frozen in OCT, and sections were obtained for fluorescence.

The bladders were sectioned (2 μm thick) with a cryostat and then examined first by ordinary fluorescence (2 sections per animal) and then by confocal microscopy using a Leica TCS NT confocal system (1 section). A representative field was selected for confocal microscopy. Texas Red emission was stimulated with excitation at 568 nm and the fluorescence was collected using a 630 nm bandpass filter. The nuclei were visualized by labelling with 20 μM Hoechst 33259. Sections were also stained with hematoxylin and eosin (H&E) and a low-pH Alcian Blue counterstained with Nuclear Fast Red. For analysis of the fluorescence intensities to assess relative binding, areas of urothelium in the image were outlined with the "lasso" tool of PhotoShop. Total pixels within the area were counted. The red colour was then detected within that area and the average brightness and number of pixels was calculated. The integrated intensity was calculated as the average pixel intensity × number of red pixels/total area in pixels. For the controls, the unenhanced image was used so that it would be comparable to the other images. Images were interpreted without knowledge of the treatment.

Although large placebo responses have clouded clinical trials of replacement GAG therapy, thousands of patients have benefited from therapy with one of the agents used for allegedly replacing the bladder "GAG Layer." The mechanism of action could arise from physically coating the bladder surface and restoring impermeability as Parsons has claimed [19], or could be due to a more complex mechanism [20]. The normal urothelium has minimal affinity for chondroitin sulphate and binds it only on the luminal surface of the umbrella cells in miniscule quantities. However, the chondroitin sulphate did not penetrate into the deeper muscle layers of the bladder when the urothelium was damaged or destroyed, but bound extensively to the damaged bladder surface. Two of the bladder models yielded virtually identical results, but the pattern seen with protamine damage was different, and less chondroitin sulphate was bound than was observed in the trypsin and dilute HCl damage models. Lewis and co-workers [21] observed that protamine was inserted into the apical cell membrane, destroying the permeability barrier, but that glycosaminoglycans extracted the protamine, thereby restoring the bladder impermeability. Our findings are not inconsistent with this model, in which case both the extracted protamine and its complexed chondroitin sulphate would be removed with the fluid. The net result could be decreased binding of the label. However, the protamine instillation produced less overt damage as well.

The observed pattern of binding to damaged bladder suggests that its main action is on the urothelial surface and is not due to a pharmacologic effect. Both heparin and pentosanpolysulphate have pharmacologic effects on the coagulation system [22], whereas chondroitin sulphate, a normal constituent of the bladder surface, and hyaluronate do not [23]. All of these glycosaminoglycan molecules have effects on mast cells and other cells within the bladder at high concentrations in vitro [24–26]. The fact that all of these glycosaminoglycan molecules alleviate symptoms in some patients in spite of not penetrating into the bladder and remaining on the surface, even in the face of extensive damage, suggests a physical effect is responsible for clinical efficacy at least of chondroitin sulphate.

We suggest that one possible mechanism for the symptom-relieving effect of this class of compounds is they may exclude urinary proteases rather than salts. Theoretical calculations based on actual measurements of the amount of glycosaminoglycan found normally on the human bladder surface showed the presence of a dense layer of glycosaminoglycan will yield a bound water layer that reduces the concentration of salts at the bladder surface [20]. Lewis and Clausen earlier had demonstrated that urinary proteases degrade bladder ion channels [27], and Negrete and co-workers demonstrated the impermeability of rabbit bladder lay in the apical membrane [15]. Because the highly charged GAG layer with its attendant bound water layer would be a far more significant barrier for macromolecules than small, inorganic ions, we speculate that the GAG layer may normally function to exclude macromolecules rather than salts. Therefore the therapeutic benefit of exogenous glycosaminoglycans may result from exclusion of proteases or other macromolecules from the bladder surface where they could produce additional damage.

Chondroitin sulphate instilled intravesically into three animal models of bladder damage coated the damaged bladder surface and failed to penetrate the bladder. This suggests they act at the bladder surface rather than in deeper layers of the bladder.