Urinary Retention and Cystitis Associated with Subcutaneous Estradiol Pellets in Female Nude Mice

Abstract

Unexpected deaths occurred in studies involving a nude mouse model of mammary cancer that required subcutaneous implantation of 0.5 mg twenty-one–day release estrogen pellets for growth of the estrogen-dependent mammary tumor xenograft BT474c. Early deaths occurred in female nude mice and were associated with urinary retention, frequently with cystitis. Drug treatment had no effect on the incidence or severity of cystitis. Histological findings did not alter significantly over various time points following pellet implantation. Changes were not seen in males or in females receiving lower doses of estradiol even when the duration of administration was prolonged, suggesting that a threshold level was required for the onset of urinary retention. Because of the influence of estrogen on micturition, immunohistochemistry for estrogen receptor alpha (ERα) in the urinary bladder was carried out, which did not demonstrate any differences between females implanted with 0.5 mg twenty-one–day release estrogen pellets and nonimplanted females.

Although previous publications have concentrated on possible mechanisms of action, this paper describes the histopathological changes seen in the urinary bladder of female nude mice resulting from exposure to high levels of estradiol.

Introduction

Nude mice (nu/nu) are commonly used in the development of anticancer drugs as an animal model for human neoplasia using human tumor xenografts against which potential therapies can be tested (Voskoglou-Nomikos et al. 2003). The assessment of drug therapies targeted against estrogen-responsive tumors, such as the estrogen antagonist Tamoxifen, requires the development of animal models using estrogen-sensitive mammary tumors (Wu and Tannock 2005). One model of an estrogen-responsive mammary cancer cell line in nu/nu mice involves the implantation of slow-release estrogen to sustain the growth of the tumor (Detre et al. 2003).

In developing an in-house model of an estrogen-dependent mammary cancer in nu/nu mice (ONU-Foxn1nu-Alpk), unexpected deaths, associated with urinary retention and grossly enlarged urinary bladders, were noted in female mice implanted subcutaneously with 0.5 mg, twenty-one–day release estradiol pellets. In an attempt to identify the cause of the urinary retention and death of these animals, microscopic examination of the urinary tract was carried out. This paper describes the histopathological changes found in the urinary bladder in association with estradiol pellets. Although there are many references in the literature covering the possible mechanism of action, description of the microscopic lesions encountered are difficult to find. Because of the reported influence of estrogen on the control of micturition, we also immunohistochemically stained the urinary bladder for estrogen receptor alpha (ERα).

Materials and Methods

The studies from which the samples were obtained were conducted in accordance with the provisions of the Animals (Scientific Procedures) Act 1986. A total of sixty-three female nu/nu mice (ONU-Foxn1nu-Alpk), on reaching a threshold body weight of 18 g, were subcutaneously implanted with twenty-one–day release, 0.5 mg estradiol pellets (Innovative Research of America, Sarasota, FL, USA), and on the following day they were injected subcutaneously with 1 × 10⁷ human breast cancer cells, of the BT474c cell line. These mice were either untreated controls (35/63) or received anticancer treatment (28/63). In addition, ten females received breast cancer cells and 0.36 mg sixty-day release estradiol pellets. Six unpelleted females were included in the examination as negative controls.

Breast cancer cells were also injected into six males subcutaneously implanted with 0.5 mg, twenty-one–day release estradiol pellets.

Mice were killed with rising concentrations of CO₂ followed by exsanguination. A midline incision was made, and both kidneys were removed, cut into 3–5 mm slices and fixed for twenty-four hours in 10% neutral buffered formalin. The urinary bladder was removed, with the urethra and vagina to maintain anatomical relationships, and fixed intact in 10% neutral buffered formalin for twenty-four hours. To avoid artifactual distention of the lumen, no fixative was injected into the lumen of the bladder. Following fixation, the tissues were dehydrated, cleared in toluene, and embedded in paraffin wax prior to sectioning at 4 μm. Sections were dewaxed and brought through a graded series of ethanols to distilled water and either stained with hematoxylin and eosin (H&E) or processed further for immunohistochemical localization of ERα. In the first twenty-two animals examined, step sections were made at intervals of approximately 100 μm through the urinary bladder, urethra, and vagina to examine whether a point of urinary obstruction could be identified. Subsequently, single sections were cut, stained with H&E, and examined by light microscopy. In a proportion of cases where bacteria were identified on H&E, sections were also stained with Gram’s stain (Stevens and Francis 1996).

Immunohistochemistry of ERα

Sections of urinary bladder and urethra were dewaxed in xylene, passed through graded alcohols, and rehydrated with water. Heat-mediated antigen retrieval was performed in a Milestone RHS-2 microwave (Milestone, Sorisole, Italy) at 110°C for two minutes in 10 mM EDTA buffer, pH 8.0, and immunohistochemical staining was performed, at room temperature, on a Lab Vision Autostainer 720 (Lab Vision, New-market, UK). The reagent and wash buffer was 0.05 M Tris buffered saline plus 0.05% Tween 20 (TBST). Endogenous peroxidase activity was quenched by pre-incubation in 3% hydrogen peroxide in TBST for ten minutes. Slides were then washed and incubated for twenty minutes with 5% normal goat serum in TBST. Excess blocking serum was removed, and the slides were incubated with polyclonal rabbit anti-ERα (Affinity BioReagents Inc., Golden, CO, USA) diluted 1:200 in TBST for sixty minutes. Following washing, the slides were incubated for thirty minutes in rabbit-specific EnVision + System–HRP (Dako UK Ltd., Ely, UK), and antibody binding was visualized by incubation in diaminobenzidine (DAB), from the EnVision+ kits, for ten minutes. The slides were counterstained for one minute using Carazzi’s hematoxylin (Clin-Tech, Guildford, UK). A negative control was produced by substituting rabbit Ig fraction, diluted to the same concentration, as the primary antibody. Sections of mouse prostate, fixed and processed in the same way, were used as positive controls.

Attempts were made to develop a method for the immunohistochemical demonstration of ERβ on formalin-fixed, paraffin-embedded mouse tissues, but these attempts proved unsuccessful owing to a lack of suitable antibody.

Results

Characteristics of the animals examined are summarized in Table 1. Because this was an investigation into an ongoing problem and not a specifically designed time course study, only a proportion of animals were examined histologically, and different numbers of animals were examined at different time points. The urogenital tracts of forty-five female nu/nu mice, subcutaneously implanted with 0.5 mg, twenty-one–day release estradiol pellets, were examined. Macroscopically, bladders were variably distended with clear, pale yellow urine, or white, amorphous, sand-like material. In one animal, the lumen contained blood. Bladder walls were notably thickened in some cases. In total, thirty-three of forty-five animals had macroscopic and/or microscopic abnormalities of the urinary bladder.

A summary of the main histopathological findings is given in Figure 7. Histological evidence of cystitis, which varied in severity from minimal to severe, was seen in the majority of female nu/nu mice, subcutaneously implanted with 0.5 mg, twenty-one–day release estradiol pellets (thirty-one of forty-five). Little difference was seen in the nature or severity of urinary bladder lesions examined between twenty-one and forty days following estrogen pellet implantation. Changes were seen in implanted mice receiving no drug treatment (control animals) as well as those receiving anticancer treatment. Urinary bladder pathology was therefore considered to be a direct effect of the estrogen, rather than an effect of treatment.

Microscopically, there was variable luminal distention, with or without evidence of intraluminal basophilic, amorphous, granular/crystalline material; bacterial colonies; or hemorrhage. Of the thirty-one bladders that showed an inflammatory response, twenty were distended, which was graded according to the degree of attenuation of the bladder wall. In two cases, the bladder was distended with no microscopic evidence of cystitis. There was no apparent correlation with the degree of bladder distention and the microscopic evidence of intraluminal granular/crystalline material. In addition, examination of step sections of the bladder and urethra revealed no evidence of physical obstruction of urinary outflow, except in a single female in which the urethra was distended with this material (Figure 2b). In one case, in which there was a significant amount of intraluminal hemorrhage, crystals were seen in negative relief (Figure 5c). These crystals were of variable size and shape, often with jagged outlines, and were consistent with the morphology of calcium phosphate crystals (Cohen et al. 2007). Intraluminal bacterial colonies of Gram-positive coco bacilli were seen in a few cases (eleven of thirty-two) of urinary bladder distention, the majority of which also had loss of the urothelium (not shown). Ulceration was seen in eleven cases. This ulceration could be extensive and was often accompanied by formation of granulation tissue in the lamina propria (Figure 4). In a proportion of cases, sub-epithelial granulation tissue formation indicated previous or adjacent ulceration, which could not be seen in the plane of section. There was no evidence of squamous metaplasia in surviving epithelium in the H&E-stained sections examined. Inflammatory cell infiltrates consisted of perivascular, focal, and more diffuse accumulations of polymorphonuclear leukocytes that expanded the serosa, separated and replaced detrusor muscle fibers, and occasionally infiltrated the urothelium (Figure 3). In one animal there was severe, diffuse necrohemorrhagic cystitis characterized by extensive ulceration, coagulative necrosis of the bladder wall, and intraluminal hemorrhage, crystals, and bacterial colonies (Figure 5).

In addition, the urinary bladder from ten females implanted subcutaneously with 0.36 mg, sixty-day–release estradiol pellets (Figure 1a) and six normal, cycling (unimplanted) females was examined. No evidence of cystitis was seen in either of these groups. Six male nu/nu mice were also implanted subcutaneously with 0.5 mg, twenty-one–day release estrogen pellets (killed at day 31 postimplantation). No microscopic abnormalities were seen in the urinary bladder of these animals (Figure 1b). Immunohistochemical staining for ERα was negative in the urinary bladder and urethra from females with 0.5 mg pellets and unpelleted controls. Nuclear staining in the bladder and urethral epithelium and stroma was negligible in these animals (Figure 6). In contrast, strong nuclear staining for ERα was seen in the uterine endometrium and myometrium, in the cervical and vaginal epithelium, and in the stroma (internal positive control). Nuclear staining for ERα was also seen in the control epididymal epithelium. Unfortunately, ERβ expression could not be investigated because of the difficulty in obtaining this antibody for use in mice.

Discussion

Unexpected deaths associated with urinary retention, bladder distention, and cystitis were seen in female nu/nu mice subcutaneously implanted with 0.5 mg, twenty-one–day release estradiol pellets. Changes were seen in implanted mice receiving no drug treatment (control animals) as well as those receiving anticancer treatment. In addition, no evidence of cystitis was seen in unimplanted females or in male mice implanted subcutaneously with 0.5 mg, twenty-one–day release estrogen pellets. Urinary bladder pathology was therefore considered to be associated with the effect of the subcutaneously implanted estradiol pellets on the lower urinary tract of the female mice. As similar changes did not occur in females given lower doses of estrogen even when the duration of administration was prolonged (0.36 mg for sixty days), we concluded that a threshold level of estrogen was required for the onset of urinary tract pathology. However, Elson et al. (2000) reported unexpected mortality associated with urinary stasis and bladder distention in one-month-old female transgenic mice expressing HPV16 oncogenes, treated with 0.25 and 0.10 mg/sixty-day 17 β–estradiol subcutaneous pellets. No description of the histopathology of the urinary tract was given, however.

In our study, no effects were seen on the urinary bladder of male nu/nu mice with subcutaneous estrogen implants (0.5 mg twenty-one–day release), but results in other studies vary. Walker et al. (1992) reported that adult male mice (New Zeal-and Black x New Zealand White mice) chronically treated with estrogens developed bladder output obstruction with urinary retention, an enlarged bladder with occasional bladder stones, and hydronephrosis. However, Streng et al. (2005) found that low concentrations of estrogens may be needed for effective voiding function in male mice.

The reported effects of estrogen on the urinary tract are often conflicting (Aikawa et al. 2003; Fleishmann et al. 2002; Liang et al. 2002; Longhurst, 2002; Longhurst and Levendusky 2000; Yono et al. 2000), which is likely associated with the varying effects of estrogen according to the sex of the animal, level of exposure, and age of initial exposure (Nilsson et al. 2001). In addition, strain differences in the sensitivity to estradiol are also said to exist (Elson et. al., 2000). Estrogenic action is mediated through ERα and ERβ, which are members of the nuclear receptor family of ligand-regulated transcription factors (Matthews and Gustafsson 2003). However, others have postulated that many of the rapid estrogen-induced changes of animals are non-receptor mediated (Nilsson et al. 2001).

Micturition is a highly complex process involving the orchestrated interaction of the autonomic and central nervous systems and various hormonal control systems (Andersson and Arner 2003). It is well documented that disturbance of estrogen levels can affect the function of the lower urinary tract, although the mechanism is complex and remains unresolved. Andersson and Wein (2004) consider that estrogen can influence lower urinary tract structure and function and modify the responses of the detrusor muscle to autonomic nervous influences, thereby interfering with bladder contraction and emptying (Wallace 2001). In the animals examined in this study, only one of sixty-three had any evidence of a physical obstruction to urinary output. Thus, a physiological effect is more likely to be associated with the urinary retention seen. Low numbers of bacteria were also occasionally seen within the bladder lumen, and the presence of these organisms was considered secondary to the urinary stagnation. Any anatomical or functional abnormalities of the urinary tract disable innate host defenses, since incomplete urine flow and bladder emptying cause urine stagnation and will promote infection (Lee and Nield 2007).

No evidence of bladder urothelial squamous metaplasia was seen on H&E-stained sections from any of the animals examined in this study. In contrast, Kuroda et al. (1985) induced urinary retention in both female and castrated male mice with subcutaneous estrogen injections, but increased resistance to urinary flow, and dilatation of the posterior urethra was seen only in the males. Histological changes involved the urethra of both sexes and consisted of urethral epithelial stratification and cornification and periurethral fibrosis. In man, bladder urothelial keratinization occurs heterogeneously, with fully keratinized areas adjacent to areas lined with apparently normal urothelium, and is thought to be associated with different embryonic origins of the epithelium in different regions of the bladder (Liang et al. 2005). Similar foci of metaplasia have been noted in rats and mice with vitamin A deficiency (Gijbels et al. 1992; Liang et al. 2005).

Immunohistochemical staining for ERα, on sections of urinary bladder and urethra from both 0.5 mg estradiol pelleted and unpelleted females was examined. Staining of the lower urinary tract for ERα was negligible in both groups, and there was no evidence of altered expression in the 0.5 mg pelleted females when compared with the unpelleted females. This finding suggests that ERα is not the predominant receptor subtype in the bladder or urethra of nu/nu mice and that estrogen supplementation of 0.5 mg estradiol for twenty-one days does not result in altered regulation of ERα at these sites. This finding conflicts with the findings of Streng et al. (2005) using transgenic mice in which ERs are inactivated. They reported that ERα was the subtype mediating the estrogen effects on lower urinary tract function, whereas the role of ERβ was unclear. However, inter-species variation may occur since Elicevik et al. (2006) report that in rats of both sexes, ERβ is the predominant subtype found in the epithelium of the urinary bladder, whereas ERα has been identified in bladder smooth muscle and fibroblasts.

In conclusion, with respect to high levels of estrogen and urinary retention, the nu/nu mouse appears to behave in a way similar to conventional animals. Therefore when using female nu/nu mouse models of estrogen-dependant tumors, urinary retention and histopathological changes of cystitis may be encountered. In this study chronic subcutaneous administration of estradiol pellets, above a threshold concentration, resulted in urinary stasis in the female but not the male nu/nu mouse. Reducing the level of estrogen administered controlled these effects. Cystitis varied in severity from minimal to ulcerative and hemorrhagic and in some cases was evident up to nineteen days after withdrawal of estrogen. The results of immunohistochemistry also suggest that this effect is independent of ERα.