Dale E. Bjorling, Department of Surgical Sciences, School of Veterinary Medicine, 2015 Linden Drive, Madison, WI 53706, USA. e-mail: email@example.com
To develop a method of direct intravesical administration of acrolein and evaluate the severity of cystitis in response to increasing doses of acrolein in female C57BL/6N (C57) mice, with further studies to compare the severity of acute acrolein-induced cystitis among C57, C3H/HeJ (HeJ), and C3H/OuJ (OuJ) strains of mice, as chemical cystitis produced by the systemic administration of cyclophosphamide is thought to result from renal excretion of hepatic metabolites, particularly acrolein.
MATERIALS AND METHODS
Doses of acrolein (0–1000 µg, 15 µL total volume) were instilled into the bladders of C57 female mice; the bladders were removed 4 or 24 h later, weighed, and processed for histology. Acrolein (6 or 10 µg; 15 µL) was instilled into the bladders of C57, HeJ and OuJ female mice, the bladders removed 4 or 24 h later, weighed, and processed for standard histology and immunohistochemical detection of uroplakin.
Increasing doses of acrolein up to 100–200 µg caused a linear increase in bladder weight and greater histological evidence of inflammation. Doses of >200 µg caused submaximal increases in bladder weight, apparently due to structural damage of the bladder. Bladder weight and submucosal oedema were consistently greater in C57 and HeJ than OuJ mice. Treatment with acrolein caused loss of urothelium along with uroplakin in some areas of all bladder sections 4 h after treatment. Bladders from C57 mice had some loss of urothelium 24 h after instillation of 6 or 10 µg acrolein, but urothelium and uroplakin covered nearly all the surface of bladders of HeJ and OuJ mice 24 h after treatment. There were significantly more white blood cells in bladders from C57 or HeJ mice than in bladders from OuJ mice 24 h after an instillation of 6 or 10 µg acrolein.
Intravesical instillation of acrolein produces dose-dependent cystitis in mice. OuJ mice appear relatively more resistant to irritant effects of intravesical acrolein than C57 or HeJ mice, and future studies will be directed at identifying genetic causes for these differences.
Systemic injection with cyclophosphamide (CP) is a widely used method of inducing experimental chemical cystitis in mice and rats. CP is an antineoplastic alkylating agent that creates cross-links in DNA, resulting in strand breaks. Haemorrhagic cystitis is a common complication of the use of this drug in patients with cancer, and it was shown that acrolein is responsible for cystitis associated with CP [1,2]. Acrolein is a highly reactive aldehyde derived from precursors formed by the metabolism of CP in the liver and possibly the kidney and bladder, but the mechanism by which acrolein reaches the bladder is unclear . The remainder of the urinary tract is relatively unaffected by CP, and its effect on the bladder has been attributed to the duration of contact with acrolein resulting from the urinary storage function of the bladder. Effects of acrolein on the bladder wall include oedema, ulceration, neovascularization, haemorrhage, and necrosis [1,2].
Functionally, CP-induced cystitis results in increased frequency of voiding, decreased urine volume/void, and increased permeability of the bladder wall [4,5]. It was reported that rats develop progressive bladder inflammation within 4 h of one i.p. injection with CP (100 mg/kg) that is accompanied by lacrimation, piloerection, loss of general mobility characterized by a rounded back (or ‘hunched’) posture and evidence of abdominal discomfort, as indicated by tail hyperextension, abdominal retraction, and licking the lower abdomen . Behavioural changes have been attributed to the effects of cystitis, and CP-induced cystitis has been considered a model of visceral pain [6,7]. Although i.p. injection with CP does not appear to cause generalized abdominal inflammation, as indicated by gross inspection and lack of extravasation of Evan’s blue dye within abdominal visceral other than the bladder , it remains unclear as to whether or not i.p. injection with CP might affect the innervation or structure of abdominal viscera other than the bladder.
Previous studies indicated that the severity of cystitis induced by CP can vary among strains of mice [6,8,9]. It is unclear whether or not the relative sensitivity of strains of mice to CP-induced cystitis is due to a differential response to intravesical acrolein or if this is the result of variations in the efficiency with which CP is metabolized to acrolein and metabolites are excreted in the urine. At least one study showed substantially different rates of CP metabolism and acrolein excretion in two strains of mice .
Many inflammatory disorders of the bladder have been shown to be accompanied by increased permeability of the bladder mucosa. Although the health of umbrella cells (apical mucosal cells) and the tight junctions between these cells probably are the most crucial elements in maintaining the relative impermeability of the bladder mucosa, it was suggested that uroplakins might also contribute to the relative permeability of the bladder mucosa [10,11]. Uroplakins are expressed by differentiated urothelial cells and consist of 16-nm protein particles packed hexagonally to form two-dimensional crystals of asymmetric unit membranes found on the luminal surface of mature urothelial cells . Although the function(s) of uroplakin is not entirely clear, at least three have been proposed; physical stabilization of umbrella cells, surface area adjustment, and enhancement of the barrier function of urothelium [10,11].
The acute effects of the intravesical administration of acrolein in doses of 0–1000 µg on bladder tissue in mice were evaluated to establish a more tightly controlled model of cystitis that would eliminate potential confounding effects of i.p. injection with CP, and allow a critical assessment of therapeutic intervention or factors contributing to the severity of cystitis. Based on the results of these studies, we selected two doses of acrolein (6 and 10 µg) to compare the response of the bladder independent of variations in CP metabolism and excretion among three strains of mice. In this model, we also investigated whether or not uroplakins were lost from the apical surface of the mucosa concurrent with the development of inflammation.
C3H/HeJ (HeJ) and C3H/OuJ (OuJ) are co-isogenic strains of mice differing in that a dominant negative proline-to-histidine substitution in the gene for Toll-like receptor 4 (Tlr4) is present in HeJ . Tlr-4 is a member of the Toll-like receptor family that forms a key element of innate immunity. Tlr-4 is particularly important in the recognition of lipopolysaccharide found in the cell wall of Gram-negative bacteria, and lack of functional Tlr-4 results in decreased cytokine formation in response to Gram-negative bacterial infection, and increased susceptibility to resultant morbidity and mortality. C57BL/6N (C57) mice are derived from a different strain than are HeJ or OuJ mice, and C57 mice have functional Tlr-4. Recent reports indicate that Tlr-4 might bind to a variety of endogenous ligands associated with the initiation of inflammation from nonbacterial causes [14–16]. Although absence of functional Tlr-4 receptors has been shown to decrease the severity of inflammation associated with Escherichia coli- or lipopolysaccharide-induced cystitis in mice [17,18], the relative significance of the presence or absence of functional Tlr-4 receptors in chemical cystitis has not been assessed. These three strains of mice were therefore chosen to compare the relative importance of functional Tlr-4 receptors in acrolein-induced cystitis.
MATERIALS AND METHODS
C57 female mice (19–21 g body weight) were anaesthetized with isoflurane in oxygen. Polyethylene (PE10) tubing was passed into the bladder, gentle pressure was applied to the abdomen to empty urine from the bladder, and varying amounts of acrolein (Ultra Scientific, North Kingston, USA; RI2, stored at − 20 °C; fresh dilutions made in PBS each day and used immediately; 6, 10, 100, 200, 400 and 1000 µg; 15 µL total volume) or a similar volume of sterile PBS (0.9% saline) were instilled into the bladders. Mice remained anaesthetized for 30 min after acrolein instillation. Immediately before killing the mice, and 8, 12 and 16 h after acrolein instillation in mice killed 24 h after treatment, the mice were observed for activity (moving about the cage, interacting with other mice, etc.) and the individual appearance of mice recorded (posture, hair ruffled or smooth, eyes open, excessive lacrimation, etc.), by one author, to subjectively assess the systemic response to intravesical instillation of acrolein. At 4 or 24 h after instillation of acrolein, mice were weighed, killed by exposure to 100% CO2, and the bladders removed, blotted, weighed, and processed for histological evaluation. Bladders for histological evaluation were fixed in 4% buffered neutral formalin, sectioned, and stained with haematoxylin and eosin (H&E). Sections were viewed by light microscopy.
In separate experiments, 10–12-week-old C57, HeJ and OuJ female mice (18–20 g body weight) were anaesthetized, and acrolein (6 or 10 µg; 15 µL total volume) or a similar volume of sterile PBS was instilled into the bladders (eight mice for all treatments in each strain). Mice were killed by exposure to 100% CO2 at 4 and 24 h after treatment, the bladders removed, blotted, weighed, and fixed in 4% buffered neutral formalin. Tissues were sectioned and processed for standard histological evaluation using H&E staining or immunohistochemical (IHC) staining for uroplakin.
Sections to be stained for IHC were deparaffinized in xylene and alcohol. Antigen was retrieved by incubation in citrate buffer in a de-cloaking chamber. Slides were rinsed in Tween/buffer solution (TBS) and extraneous proteins were blocked by incubation with normal goat serum for 1 h. Slides were incubated overnight at 4 °C with primary pan-anti-uroplakin antibody (recognising all subtypes of uroplakin; kindly provided by Dr TT Sun, New York University School of Medicine). Slides were rinsed with TBS, incubated with secondary antibody (MACH 3 rabbit probe, BioCarta US, San Diego, CA, USA) for 15 min, rinsed, and incubated with secondary antibody (MACH 3 R-polymer HRP, BioCarta) for 15 min, rinsed, incubated with diaminobenzidine reaction solution to reveal horse-radish peroxidase staining, and counterstained with haematoxylin. Control slides were treated in a similar manner, but the primary antibody was replaced with normal rabbit serum.
The presence or absence of uroplakin and urothelium was determined by visual inspection of sections stained for uroplakin and counterstained with haematoxylin. The entire circumference of the bladder lumen was traced, and the length of areas devoid of uroplakin, urothelium, or both was determined using Image Pro Plus (Media Cybernetics, Inc., Silver Spring, MD, USA). Urothelium was considered present if any cells remained between the submucosa and lumen. No areas could be found where urothelium was present and uroplakin was absent, i.e. absence of staining for uroplakin consistently indicated absence of urothelium.
An oedema index was determined by examining three sections of each bladder at 100 µm intervals, with scoring determined as: 0, no oedema evident; 1, oedema present, limited to submucosal region, and the width of the submucosal region not exceeding the combined width of urothelium and detrusor; 2, oedema present in the bladder wall but not detrusor, and the width of the submucosal region greater than the combined width of urothelium and detrusor but less than twice this distance; 3, oedema present in the bladder wall, possibly including occasional areas of detrusor, the width of the submucosal region 2–4 times the width of urothelium and detrusor; and 4, oedema present throughout bladder wall, including detrusor, and the width of the submucosal region more than four times the width of urothelium and detrusor (but urothelium typically lost in these tissues due to the severity of injury).
The mean number of white blood cells (WBCs) per 50 000 µm2 was determined by obtaining four images of each bladder stained with H&E at × 40. Total WBCs were counted, and field areas measured using Image Pro Plus® v.5.0) The abundance of WBCs was then expressed as the mean number of WBCs per 50 000 µm2.
Comparisons over time or within groups were made using Student’s paired t-test or anova. Multiple groups were compared using anovapre hoc and the Scheffe F test to determine post hoc significance. Linear regression analysis was used to compare acrolein doses and associated bladder weights. In all tests the significance was set at 95% CIs.
C57 mice recovered quickly from isoflurane anaesthesia, and activity patterns appeared to be normal 4 h after acrolein instillation; acrolein had no effect on body weight at 4 or 24 h after treatment. Instillation of 0–600 µg had no effect on the appearance or activity patterns during the 24 h after instillation, but mice that received 1000 µg were less active and the hair over the neck and dorsal thorax appeared ruffled compared with other mice. These mice also generally adopted a ‘hunched’ posture, and their eyes appeared half closed. There was no apparent increase in lacrimation in any mice.
Intravesical acrolein consistently produced cystitis in a dose-dependent manner that was characterized primarily by submucosal oedema (Fig. 1A,B). The bladder mucosa remained intact in mice that received 6 µg acrolein at both 4 and 24 h, but doses of ≥ 10 µg stimulated increasing loss of urothelium at both times. There were widely scattered nucleated cells (WBCs) in the submucosal space. There was haemorrhage sporadically in mice treated with 6 or 10 µg at both 4 and 24 h after treatment, but it was more consistent in mice treated with 100 or 200 µg at both sample times. Instillation of acrolein caused a significant increase in bladder weight at both 4 and 24 h for all doses examined, and the increase in bladder weight was maximal at 100–200 µg. At higher doses, bladder weights declined and this was associated with loss of the mucosa and submucosa and thinning of the detrusor. (Fig. 2). At doses of >100–200 µg, bladder weights began to decline, and this was also associated with loss of the mucosa and submucosa and thinning of the detrusor. Linear regression analysis of doses of acrolein from 0 to 200 µg gave R = 0.81 for the comparison of bladder weight to acrolein dose 4 h after acrolein instillation, and R = 0.93 up to 24 h after instillation, showing a significant relationship between acrolein dose and bladder weight.
Having characterized the response of C57 female mice, we then compared the dependence of the response on strain of mouse. At 4 h after instillation of 6 µg acrolein, bladder weight was significantly increased in all mice (Fig. 3A); at 24 h after instillation of 6 µg acrolein, there were significant differences in bladder weights among the three strains, being greatest in C57 mice and lowest in OuJ mice (Fig. 3B). Similarly, at 4 and 24 h after instillation of 10 µg acrolein, bladder weights in all groups were significantly greater in all strains but were highest in C57 mice and lowest in OuJ mice (Fig. 3A,B). As expected, instillation of saline had no effect on bladder weight (data not shown).
Oedema indices were increased in C57 and HeJ mice at 4 h after instillation of 6 µg acrolein (Fig. 4A); they were also increased in all three strains at 4 h after instillation of 10 µg acrolein. At 24 h after instillation of either 6 or 10 µg acrolein, oedema indices of C57 and HeJ mice were significantly greater than OuJ mice, and the oedema indices of OuJ mice were not significantly different from those in saline-treated controls (Fig. 4B).
Urothelium and uroplakin were consistently present on the entire luminal surfaces of bladders of all three strains of mice that received intravesical saline. Treatment with acrolein caused loss of uroplakin and urothelium in some areas of the bladder sections (Fig. 5). Intact urothelium was never seen in the absence of staining for uroplakin, suggesting that uroplakin is not shed by the urothelium before detachment of the urothelium from underlying tissues. At 4 h after instillation of 6 µg acrolein there was minimal loss of urothelium and uroplakin (Fig. 6A). Urothelium and uroplakin covered 97%, 87% and 89% of the surface of bladders from OuJ, HeJ, and C57 mice, respectively. Whereas the urothelium and uroplakin were almost completely restored in OuJ and HeJ mice 24 h after instillation (100 and 99%, respectively), C57 mice had a further loss of urothelium and uroplakin (75% coverage of luminal surface; Fig. 6B). At 4 h after instillation of 10 µg acrolein, there was significantly greater loss of urothelium and uroplakin in C57 than OuJ or HeJ mice (61% coverage of luminal surface, vs 83% and 91%, respectively; Fig. 6A), and OuJ and HeJ mice recovered significantly more urothelium and uroplakin expression at 24 h after instillation of 10 µg acrolein (100%, 93% and 81%, respectively; Fig. 6B).
In general, there were very few WBCs within the urothelium, submucosal tissue, or detrusor after intravesical instillation of acrolein, and haemorrhage was also relatively uncommon. There were differences in the number of WBCs in the strains of mice after acrolein instillation, but these were not consistent. There were fewer WBCs in the bladders of C57 mice than either HeJ or OuJ at 4 h after instillation of 6 µg acrolein, while there were fewer WBCs in bladders from OuJ mice after instillation of 10 µg acrolein (Fig. 7A). There were significantly more WBCs in bladders from C57 or HeJ mice than OuJ mice at 24 h after instillation of either 6 or 10 µg acrolein (Fig. 7B).
The primary response of the mouse bladder to an instillation of acrolein was the formation of submucosal oedema. This increased in a concentration-dependent manner and progressed to include oedema of the detrusor. The linear relationship between bladder weight and acrolein dose indicates that this model of cystitis can be relatively tightly controlled to produce the degree of inflammation desired for a particular study. Complete erosion of the mucosa and significant intramural haemorrhage only occurred at doses of ≥ 100 µg. Bladder weight declined at doses of >200 µg, and this appeared to be the result of a substantial disruption of the architecture of the bladder wall. Infiltration of the bladder wall by nucleated cells (predominantly WBCs) was not a prominent feature of the histological response of the mouse bladder to acrolein at the sample times assessed. Instillation of acrolein into the bladder was only accompanied by behavioural changes at a dose of 1000 µg.
The use of two co-isogenic strains, and one non-isogenic strain of mice in the present study allowed a comparison of the response of closely related stains of mice to that of a substantially different strain of inbred mice. There were strain-specific differences in the response to direct intravesical administration of acrolein. This strongly suggests that there are genetic determinants that influence the sensitivity and/or response of the different strains to direct chemical injury of the urothelium. The present studies indicate that C57 mice are the most sensitive of the three strains tested to the effects of intravesical acrolein on the urothelium, having both a greater loss of urothelium and uroplakin, and a slower recovery. C57 mice also had the greatest increase in weight after intravesical administration. However, it is unclear whether this is a function of the greater urothelial injury or indicative of a greater inflammatory response, or both. OuJ mice appeared to be the least sensitive to chemical injury, while HeJ mice appeared to have intermediate sensitivity.
As noted, it was reported that Tlr-4 receptors might mediate the inflammatory response to non-infectious stimuli [14–16]. Butylated hydroxytoluene stimulated more severe chronic pulmonary inflammation, and 3-methylcholanthrene stimulated more pulmonary tumour formation, in HeJ than OuJ mice . Bilateral sterile femur fracture resulted in liver damage, accompanied by increased hepatic interleukin-6, -10, and TNF-α levels, that was more severe in OuJ than HeJ mice . The present results support the concept that the inflammatory response of the bladder to irritant compounds varies among strains of mice, but do not support the conclusion that the response to acrolein is dependent upon functional Tlr-4. Indeed, the effects of intravesical acrolein on bladder weight, oedema index, and WBC infiltration were more severe in HeJ than OuJ mice, suggesting the novel possibility that functional Tlr-4 might actually diminish inflammation as opposed to augmenting it in this model. Additional experiments are needed to conclusively confirm this, but if it can be substantiated, this would be the first evidence to support the participation of Tlr-4 in suppressing inflammation.
Other investigators have identified differences among strains of mice in their relative sensitivity to the effects of CP [6,9]. Treatment of female C57 and DBA/2 mice with CP (300 mg/kg, i.p.) resulted in oedema, haemorrhage, invasion of the bladder wall with WBCs, and erosion of the bladder mucosa that appeared to be of similar severity in both strains, and reached maximum intensity 48 h after treatment with CP . The normal histological appearance of the bladder was restored in both strains within 7–10 days after CP treatment. However, 30–100 days after CP treatment, 86% of DBA/2 mice had significant bladder pathology, including oedema and more WBCs (particularly mast cells) within the bladder wall, as well as erosion of the mucosa, while only 4% of treated C57 mice had similar lesions .
Another study showed that ICR mice developed more intense cystitis than C57 mice in response to treatment with CP . These investigators also described differential metabolism of CP to acrolein, with greater quantities of CP metabolites being present in the urine of ICR than C57 mice at 2 and 3 h after CP treatment. Although the urinary concentrations were similar at 6 h after CP treatment, the total area under the curve calculated for CP metabolites was significantly less for urine from C57 mice than for urine from ICR mice. To determine whether or not strain differences were simply the result of differential metabolism of CP, these investigators infused acrolein directly into the bladder, by needle puncture, and the severity of cystitis in response to acrolein was again found to be more severe in male ICR than male C57 mice. It is difficult to make direct comparisons between that study and the present because these investigators instilled acrolein into the bladders of male mice by direct percutaneous puncture, and the total volume injected was not indicated.
The volume instilled into the bladders of mice in the present study (15 µL) was smaller than that typically used in other studies (75–200 µL). Although it could be argued that this volume was insufficient to provide uniform contact with the urothelium, this volume was selected intentionally to avoid distension of the bladder and potential resultant damage to the urothelium. As the bladder fills, it has been proposed that cytoplasmic vesicles fuse with the cell membrane of umbrella cells to increase the surface area of umbrella cells . Although this process accommodates normal filling of the bladder, the exact point at which distension of the bladder compromises the normal barrier function of the urothelium is unclear.
Whether or not uroplakin is critical in protecting the urothelium from damage is unclear. It was suggested that the three-dimensional structure of the uroplakin plaque covering the umbrella cells might be crucial in maintaining the relative impermeability of the urothelium. Using electron cryo-microscopy, it was found that specific subunits of the normally assembled uroplakin plaque maintained contact with cytoplasmic and exoplasmic cellular components, and that these interactions with specialized lipids might be essential for preventing leakage of urinary components across the urothelium . There are many limitations to the present results using IHC to detect the presence or absence of uroplakins. The most significant of these limitations is that, while this technique detects the presence of uroplakin proteins, it does not evaluate the structure or function of the uroplakins present. Thus it is possible that exposure to acrolein might disrupt the three-dimensional structure of the uroplakin plaques, causing an increase in permeability while leaving the component proteins in situ. Alternatively, acrolein, a product of the intravesical metabolism of acrolein, or some other signalling molecule stimulated by the presence of acrolein, might cross the urothelial barrier in the presence of intact and fully functional uroplakin. It is interesting that we detected no areas of intact urothelium that were devoid of staining for uroplakin.
The biochemical causes for differential susceptibility to the effects of intravesical acrolein remain unclear. Acrolein readily forms adducts with thiols, and it was suggested that the biological effects of acrolein might be the result of this reaction . Thiols are found in high quantities in the apical portion of umbrella cells, and total protein thiol levels were greater in bladders from relatively resistant than relatively sensitive mice . However, thiol levels did not correlate with bladder damage or permeability, and changes in thiol concentrations might be the result of conformational alterations in proteins due to alkylation by acrolein . It remains highly probable that fundamental differences in the severity of the response of the bladder to irritant substances among individuals have a genetic basis.
In conclusion, systemic illness due to the administration of CP might confound the interpretation of the effects of cystitis. This is particularly true for studies investigating behavioural responses to visceral pain (i.e. cystitis), but decreased activity, loss of appetite, decreased water consumption, and weight loss might confound observations on the pathogenic mechanisms of cystitis and physiological responses to bladder inflammation. Intravesical instillation of acrolein in mice stimulated a consistent, reproducible inflammatory response that could be tightly controlled by the dose administered. Female C57 mice had more extensive urothelial injury, greater oedema and what appeared to be slower urothelial repair than either HeJ or OuJ mice. The fundamental biochemical mechanisms underlying the response of the bladder to acrolein, and the differential effects among strains of mice, remain unknown. However, acrolein-induced cystitis in mice appears to be highly reproducible and dose-dependent, and it is likely that genetic linkage studies can be used to identify specific genes that might influence the sensitivity to intravesical irritation.
Supported by NIH awards R01DK57258 and R01DK066349
CONFLICT OF INTEREST
None declared. Source of funding: National Institutes of Health.