Comparative Pathology of Proliferative Lesions of the Urinary Bladder

Transcription

1 TOXICOLOGIC PATHOLOGY, vol 30, no 6, pp , 2002 Copyright C 2002 by the Society of Toxicologic Pathology DOI: / Comparative Pathology of Proliferative Lesions of the Urinary Bladder SAMUEL M. COHEN Department of Pathology and Microbiology, University of Nebraska Medical Center, Omaha, Nebraska , USA ABSTRACT Bladder neoplasia in humans consists of 2 diseases, a low-grade papillary tumor that does not invade or metastasize, and a high-grade lesion that usually invades and metastasizes. Bladder carcinogenesis in rats is most like the low-grade, papillary tumor, although it eventually does progress and invade. In the mouse, models are available that mimic each of these disease processes. Preneoplastic lesions in humans and rodents include various types of hyperplasia, proliferative cystitis, and dysplasia. These preneoplastic and neoplastic lesions arise throughout the urothelium, from the renal pelvis to the urethra, although most commonly in the bladder. Rarely, benign and malignant mesenchymal lesions occur in rats and mice, with a unique submucosal mesenchymal lesion present in some strains of mice. In addition, eosinophilic and clear inclusions in the super cial layer of urothelium in mice, which do not appear to be associated with toxicity or carcinogenesis, have been reported. An approach to evaluation of carcinogenic mechanisms involved in the urothelium is presented. It focuses on distinguishing between DNA reactive carcinogens vs those that act by increasing cell proliferation. Although rodent models do not precisely mimic the human disease, they have provided useful models for furthering our understanding of the carcinogenic process in the urothelium as it pertains to human diseases. Keywords. Bladder; pathology; hyperplasia; papilloma; carcinoma; dysplasia; urothelium. UROTHELIAL TUMORS IN HUMANS The urothelium lines the lower urinary tract extending from the renal pelvis through the ureters, urinary bladder, and into the urethra. Most tumors arising from this epithelium are referred to as urothelial or transitional cell tumors, although squamous cell and glandular tumors can also arise. In addition, occasional mesenchymal tumors arise in the lower urinary tract in humans and animals (1, 2). Urothelial (transitional) cell tumors in humans actually comprise 2 basic disease processes (3, 4). The separation of bladder tumors into these 2 entities is increasingly supported by clinical, biological, and molecular observations. The rst of these types of bladder neoplasms is the lowgrade, well-differentiated papillary tumor that does not invade or metastasize. However, these lesions frequently recur, either as recurrences of the previous lesion or as additional similar neoplasms arising in the same bladder (or elsewhere in the lower urinary tract that is lined by urothelium). These tumors were previously classi ed as papillary transitional cell carcinomas, grade 1. However, based on the increasing evidence that these are essentially benign tumors that do not in and of themselves progress, the World Health Organization International Society of Urologic Pathology (WHO-ISUP) reclassi ed these lesions in 1999 as papillary bladder tumors with low-grade malignant potential (5). These low-grade urothelial tumors frequently have abnormalities in chromosomes 9 and 5 and occasionally in other chromosomes (4, 6). They infrequently have mutations in the p53 gene. As indicated before, these lesions do not progress to invasive or metastatic lesions. However, individuals who Address correspondenc e to: Samuel M. Cohen, Havlik-Wall Professor of Oncology, University of Nebraska Medical Center, Nebraska Medical Center, Omaha, Nebraska , USA. Presented at the satellite symposium of the National Toxicology Program at the annual meeting of the Society of Toxicologic Pathology, Phoenix, Arizona, June 24 25, have these tumors have an increased risk of developing the second type of bladder neoplasia that clearly behaves malignantly (1, 3). The second type of bladder neoplasia is a high-grade lesion (1, 5). These primarily progress from at dysplastic and carcinoma in situ changes to invasive lesions. Most commonly, they present as invasive tumors, frequently already with nodal or distant metastases. Occasionally, invasive lesions begin as high-grade papillary tumors. The high-grade urothelial tumors frequently have areas of squamous cell or glandular differentiation or areas of undifferentiated tumor, including small cell or spindle cell variations. These highgrade tumors frequently have mutations in the p53 gene or in genes related to the p53 pathway (4, 6). Numerous chromosomal abnormalities and mutations in other key growth control genes are frequently detected in these tumors, even at their early noninvasive stages. These high-grade, malignant urothelial cell tumors of the lower urinary tract represent a minority of the neoplasms arising from the urothelium in humans. Proliferative cystitis in humans comprises three related entities: Brunn s nests (Figure 1), cystitis glandularis, and cystitis cystica (1). These have the appearance of submucosal growths, but in reality are down growths extending from the overlying epithelium that on serial section can be shown to be in continuity with the overlying epithelium. They are commonly seen in the human bladder, and beyond the age of 50 are present in all bladders, as demonstrated in large autopsy series in which the entire bladder mucosa was mapped for alterations (7, 8). It appears that these lesions have no greater risk of neoplastic development than the normal appearing epithelium. Inverted papillomas are an uncommon, distinct entity arising from the human urothelium (1). They are polypoid masses with an inverted growth pattern of the urothelium, giving a super cial appearance of invasive growth (Figure 2). However, these nodular epithelial growths are similar to Brunn s nests in that they are extensions of the overlying epithelium into /02$3.00 $0.00

2 664 COHEN TOXICOLOGIC PATHOLOGY FIGURE 1. Brunn s nests in a human urinary bladder. These resemble nodular hyperplasia in rodents (Provided by S. L. Johannson). FIGURE 2. Human inverted papilloma. These resemble the polypoid noninvasive papilloma and low-grade papillary carcinomas in rodents (Provided by Dr. S. L. Johansson).

3 Vol. 30, No. 6, 2002 COMPARATIVE BLADDER PATHOLOGY 665 the underlying submucosa. These lesions are benign and do not have a tendency to recur. They are similar in appearance to the usual polypoid papillomas and low grade, noninvasive carcinomas commonly produced in the rat and occasionally in the mouse (2). UROTHELIAL LESIONS IN THE RAT AND MOUSE Urothelial carcinogenesis in the rat proceeds through a sequence of morphologic changes beginning as simple hyperplasia, an increase in the number of cell layers in the urothelium (2, 9, 10). It then progresses to nodular and papillary hyperplasia, which resemble the Brunn s nests and papillary neoplasms with low-grade malignant potential in humans. These progress to papillomas, frequently giving an appearance similar to inverted papillomas in humans. In contrast to the low-grade papillary lesions in humans, these lesions can eventually progress to higher-grade, noninvasive carcinomas and ultimately to invasive neoplasms. Rarely they metastasize to lymph nodes, lung, and other tissues. Nodular hyperplasia has occasionally been misdiagnosed in the rat as an invasive neoplasm. However, nodular and papillary hyperplastic lesions are entirely reversible if the inciting agent is removed (11 13). Morphologically they do not show cellular atypia, although they may have mitotic gures present. Serial sections demonstrate their connection to the overlying epithelium. A similar sequence of events can also occur in the mouse in response to treatment with a variety of agents. However, nodular hyperplasia is considerably more common than papillary proliferations, and nodular hyperplasia frequently occurs with papillary hyperplasia entirely absent (2, 14). In contrast to this sequence of events proceeding through low-grade papillary lesions, a disease process has been described in the mouse that resembles the morphology and biology of high-grade urothelial neoplasia of humans. It has been best studied in mice treated with N-butyl-N-(4- hydroxybutyl)nitrosamines (BBN) administered in the drinking water or by gavage (15, 16). This sequence of events begins as a at, nonpapillary dysplasia and carcinoma in situ, usually without an accompanying hyperplastic response. It is produced rapidly, within 8 to 10 weeks, followed quickly by invasive urothelial carcinoma within 12 to 20 weeks from the beginning of treatment. These lesions are highly malignant and frequently metastasize, particularly to the lungs but also to regional lymph nodes and occasionally other organs. Similar to the low-grade papillary lesions in the human, lesions in the rat infrequently have mutations in the p53 gene, although this is somewhat dependent on the speci c carcinogen that is administered (17, 18). In contrast, the high grade dysplastic and carcinomatous lesions produced in the mouse frequently have mutations in the p53 gene and in other cancerrelated genes (16). Thus, the rat model strongly resembles low-grade papillary urothelial neoplasia in humans, whereas the mouse model resembles the high-grade, highly malignant disease of humans. OTHER BLADDER LESIONS IN RODENTS Spindle cell sarcomas occasionally occur in the rat and in the mouse (2). In the rat, these frequently arise in the polypoid papillomatous lesions. Frequently, the overlying epithelium is either normal or mildly hyperplastic, but it is not dysplastic. In the underlying submucosal connective tissue is a highly proliferative spindle cell lesion, occasionally benign with few mitoses but usually appearing to be malignant with anaplasia, numerous mitotic gures, and invasion into the muscle wall. The spindle cell lesion, whether benign or malignant, usually does not in ltrate the overlying mucosa. Occasionally, immunohistochemistry is required to demonstrate that these are mesenchymal lesions rather than spindle differentiation of a high-grade carcinoma. A mouse speci c lesion, which has smooth muscle differentiation, has been described in the submucosa of the bladder (19, 20). The speci c cell of origin has not been identi ed, and consequently, the lesions have recently been classi ed as mesenchymal lesions of the mouse bladder submucosa. It is unclear whether these arise from a regenerative process or whether they represent true neoplasms. They consist of clusters of large polygonal cells with a highly pleomorphic epithelioid appearance surrounded by small, spindle-like cells. These lesions appear to have a vascular appearance in some areas. A detailed description of these lesions is presented by Dr. William Halliwell elsewhere in this symposium (20). In the super cial cells of the mouse bladder, we occasionally see eosinophilic (Figure 3) or clear cytoplasmic inclusions. The clear appearance may be due to the dissolution of the inclusion during tissue processing. The inclusions are FIGURE 3. Granular eosinophilic (A) and clear (B) inclusions (arrows) in the super cial cell layer of the mouse.

4 666 COHEN TOXICOLOGIC PATHOLOGY FIGURE 4. Trichrome stain of a normal bladder (A) for comparison to the bladder showing thickened wall due to scar formation following healing of an ulcer (B). Both shown at the same magni cation. not accompanied by any other evidence of cellular degeneration or necrosis, nor are they associated with regeneration, hyperplasia, or malignancy. There is some evidence that they represent lipid inclusions and may represent incorporation of the administered test chemical or metabolite into the super - cial cells. These inclusions appear to be speci c to the mouse, and they do not appear to be associated with preneoplasia or neoplasia. UROTHELIAL TOXICITY AND REGENERATION Injury to the urothelium results in regenerative hyperplasia (11 13). If the injury is severe, including complete ulceration of the epithelium, the regeneration is more signi cant than if there is merely super cial erosion. Ulceration of the bladder epithelium can be produced by a variety of means (11 13, 21), including by the administration of highly cor- rosive chemicals (or their metabolites), such as cyclophosphamide administered by intraperitoneal injection, tributyl phosphate fed in the diet, or intravesical instillation of acetic acid or formalin; by mechanical injury, such as surgical incision of the bladder or freeze ulceration; or, most commonly, by the production of urinary tract calculi or implantation of pellets composed of various materials. These treatments produce extensive ulceration of the bladder epithelium with its associated acute and chronic in ammatory reaction, hemorrhage, necrosis, and granulation tissue formation. It can either heal completely, with the epithelium and bladder wall returning to normal, or it can be associated with scar formation that is readily demonstrated by trichrome stain (21) (Figure 4). Scar formation can lead to the appearance of retraction of the epithelium with a crater-like appearance as observed by scanning electron microscopy (21) (Figure 5). FIGURE 5. Scanning electron micrograph of bladder showing retraction of urothelium due to underlying scar formation. Bar 100 l m.

5 Vol. 30, No. 6, 2002 COMPARATIVE BLADDER PATHOLOGY 667 TABLE 1. Orally administered chemicals that can produce urinary calculi at high doses. Uric acid Cystine Calcium oxalate Calcium phosphate Homocysteine Cystine Magnesium ammonium phosphate Sulfonamides Triamterene Acetazolamide Adenine Uracil Fosetyl-al Silicate Terephthalic acid Dimethyl terephthalate Glycine Orotic acid Oxamide Melamine Diethylene glycol Biphenyl Note: This is not a complete list, but an indication of the variety of chemicals that can lead to the production of urinary calculi in humans and/or experimental animals. Nevertheless, the bladder epithelium returns to normal with or without scar formation. The regenerative hyperplastic response is usually extensive and involves multifocal or diffuse nodular and/or papillary hyperplasia and even diffuse papillomatosis. However, if the inciting agent is removed, the bladder epithelium returns to normal within a few weeks (11 13, 21). Calculi are the most common cause of this severe type of toxicity and regeneration. Urinary tract calculi can be formed following administration of a variety of chemicals (22, 23) (Table 1). The calculi are either formed from the administered chemical and/or metabolite, or they are produced by precipitation of normal urinary constituents, such as calcium, phosphate, oxalate, magnesium, or uric acid, by altering the normal urinary physiology and composition. Calculi can even be produced in rats by surgically performing a portacaval shunt, which leads to signi cant alteration of uric acid metabolism and eventual production of uric acid calculi, ulceration, regenerative proliferation, and carcinoma (24). If the urinary tract calculi remain for a prolonged period of time, the proliferation is extensive and ultimately results in the production of carcinomas. Jull (25) demonstrated that implantation of paraf n wax pellets into the mouse bladder produced an incidence of bladder carcinoma in approximately 10% after 1 year, 25% after 18 months, and 50% after 2 years. Calculi represent a true physicochemically de ned threshold process of carcinogenesis (22, 23). To develop the bladder tumors, a dose of chemical is required that produces suf - cient concentration of the offending substance in the urine to lead to precipitation and formation of calculi. If a dose is administered that is inadequate to generate the calculi, there is no urothelial toxicity, no regeneration, and no bladder tumor formation. The concept of threshold carcinogenesis for calculi holds true only for those chemicals which are non- DNA reactive, such as all those listed in Table 1. A threshold, high-dose response must be taken into consideration for human risk assessment of such chemicals. Consideration must also be made for decreased likelihood of formation of calculi in human urine compared to rodent urine (interspecies extrapolation factor < 1), and humans appear to be signi cantly less susceptible to the carcinogenic effects of calculi, once formed, than are rodents (26 28). Numerous other agents produce a mild cytotoxicity, producing necrosis only of the super cial cell layer (Figure 6), or occasionally also including the intermediate cell layers. As the epithelium is merely eroded, not ulcerated, there is no hemorrhage or in ammatory reaction. This super cial cytotoxicity can be produced either directly by the administered chemical or a metabolite, as with orally administered orthophenylphenol (30), by the formation of the calcium phosphate-containing precipitate generated by administration of high doses of sodium salts to rats (31), or by the generation of various types of microcrystals (22, 23). The sodium salts include such chemicals as saccharin, chloride, bicarbonate, ascorbate, glutamate, aspartate, and a variety of other chemicals which are essential to viability or are FIGURE 6. Necrosis of super cial layer of bladder with exfoliation, exposing underlying intermediate cells. Bar 100 l m.

6 668 COHEN TOXICOLOGIC PATHOLOGY produced as intermediary metabolites (31). Administration of these high doses of sodium salts results in alterations of the urinary composition in the rat which leads to a calcium phosphate-containing precipitate (31). It is a high-dose, ratspeci c phenomenon. The regenerative hyperplasia following such mild cytotoxicity, regardless of what caused it, is generally mild, may not be detectable by light microscopy (requires scanning electron microscopy), and results in a low incidence of urothelial tumors in a standard 2-year bioassay (31, 32). EXAMINATION OF BLADDER AND URINE Numerous methods are useful in detecting alterations in the urine and in the urothelium that are helpful in determining the mode of action of a given agent. Analysis of urine is essential (26). This requires measurement at least of osmolality and ph, and possibly other parameters depending on speci c circumstances. It is essential that fresh-voided urine is used for the analysis and that specimens are collected during the feeding period of the animal (during the dark for rodents) or shortly thereafter (within 2 hours of the lights going on) (26). Detection of precipitate or crystals in the urine is most sensitively observed by scanning electron microscopy. Elemental composition is readily assayed by attached X-ray dispersive spectroscopy. Processing of bladder tissues for light microscopy, immunohistochemistry, such as following bromodeoxyuridine (BrdU) labeling, and scanning electron microscopy can be readily achieved on individual specimens (29). Scanning electron microscopy is particularly useful in detecting the super cial cytotoxicity that is not usually visible by light microscopy. BrdU labeling has essentially replaced tritiated thymidine for measurement of increased cell proliferation. Proliferating cell nuclear antigen (PCNA) and other endogenous substrates have also proven useful for detecting cell proliferation, although they are not as speci c for S-phase as BrdU or tritiated thymidine incorporation. In ation with x- ative of the bladder in situ with the animal under anesthesia is strongly encouraged since electron microscopic alterations of autolysis are apparent within 1 minute of death. In ation also makes interpretation of early changes by light microscopy easier by avoiding the artifacts created by the numerous folds of the bladder and by tangential sectioning. Bouin s xative is excellent as long as no large lesions (greater than a few millimeters) are present, giving adequate preservation for scanning electron microscopy but allowing for quality light microscopic and immunohistochemical evaluations. Uroplakins The urothelium is a unique epithelium in mammals, with differentiation toward the luminal surface and terminal differentiation to large, polygonal, super cial (umbrella) cells (33). The luminal membrane has been referred to as the asymmetric unit membrane because by transmission electron microscopy the outer lea et of the plasmalemma is signi cantly thicker than the cytoplasmic lea et. This unique asymmetric membrane is formed because of the presence of 4 interrelated proteins referred to as uroplakins (34). These proteins combine to form plaques that are incorporated into the fusiform vesicles and ultimately into the luminal plasma membrane of the super cial cells. Antibodies to these proteins have been developed and can be utilized for the study of terminal differentiation during a variety of processes in the urothelium, including toxicity, regeneration, and carcinogenesis (35). Urothelial Carcinogens Numerous agents have been identi ed as carcinogens toward the urothelium in humans (36). Most of these have also been demonstrated as bladder carcinogens in rodent models. Additional chemicals have been identi ed in rodent bioassays that might be related to the disease in humans. The most common etiologic factor in bladder carcinogenesis in humans is cigarette smoke (36). This accounts for nearly half of the bladder tumors in males in countries in which schistosomiasis is not a factor. It accounts for a smaller proportion of the disease in females. Schistosoma hematobium is a parasitic infestation of the lower urinary tract, associated with an increased risk of bladder cancer (37). However, schistosomiasis, like other chronic in ammatory processes of the lower urinary tract, is associated with a disproportionate risk of squamous cell carcinomas rather than urothelial cell carcinomas (27). Other chronic in ammatory processes that appear to be related to an increased risk of developing bladder cancer, although signi cantly less than that seen with schistosomiasis, are calculi, neurogenic bladder, diverticuli, and chronic bacterial cystitis (27). Speci c chemicals have also been identi ed as bladder carcinogens in humans (36). Most notably, these include a variety of aromatic amines such as 2-naphthylamine, benzidine, and 4-aminobiphenyl. Various aromatic amines, such as 4-aminobiphenyl, have been identi ed in cigarette smoke and appear to account for at least some of the bladder carcinogenicity of cigarette smoke. Several other aromatic amines have been identi ed as possible human bladder carcinogens, including a variety of heterocyclic amines generated by pyrolysis of food. In addition, certain aromatic amides, such as phenacetin, have been associated with increased risk of urothelial tumors of the lower urinary tract, especially of the renal pelvis, but also including the ureters and urinary bladder. Chlornaphazine was a potent bladder carcinogen in humans when it was administered as a chemotherapeutic agent for the treatment of polycythemia vera (36). Chlornaphazine is a nitrogen mustard that is rapidly metabolized to the bladder carcinogen 2-napthylamine, but it produced bladder tumors signi cantly faster than usually seen with 2-naphthylamine exposure. Another chemotherapeutic agent, cyclophosphamide, also produces bladder cancer in humans, particularly high-grade, highly aggressive, malignant lesions. Other phosphoramides have been associated with an increased risk of bladder cancer. Intravesical administration of mitomycin or thiotepa in association with pelvic radiation have also been associated with increased risks of developing bladder cancer. Numerous occupations have been associated with the development of bladder cancer, including the chemical, rubber, textile, and paper industries (36). More recently, exposure to combustion products, such as diesel exhaust or aluminum

7 Vol. 30, No. 6, 2002 COMPARATIVE BLADDER PATHOLOGY 669 smeltering processes, also appears to be strongly related to an increased risk of developing bladder cancer, possibly related to exposure to polycyclic aromatic hydrocarbons or polycyclic nitroaromatic compounds (38). Tryptophan metabolites have been suggested as bladder carcinogens because some are aromatic amines or aminophenols (36). However, recent evidence does not support such an association (39). Similarly, cyclamate and saccharin have been found to be bladder carcinogens in rodents, but are unlikely to be carcinogenic to humans because the relationship in rodents appears to be high-dose and rat-speci c (31, 40). This has recently led to the removal of the chemical saccharin from the National Toxicology Program s list of carcinogens (41) and the down-classi cation of saccharin from 2B to 3 by the International Agency for Research on Cancer (42). Most of the chemicals and agents associated with bladder carcinogenesis are known as genotoxic or DNA reactive agents, including the aromatic amines and amides, polycyclic aromatic hydrocarbons and nitroaromatics, phosphoramides, and nitrosamines (36). However, certain chemicals or processes are associated with non-dna reactive chemicals and involve an increase in cell proliferation, resulting ultimately in the development of bladder tumors (43, 44). These include chronic in ammatory processes, such as schisotomiasis and exposure to calculi. However, speci c chemical agents also can act by an increased proliferative process. This appears to be likely for the relationship of inorganic arsenic exposure leading to bladder carcinogenesis in humans, and also related to a slight increased incidence of bladder tumors in rats administered dimethylarsinic acid (DMA) in the diet or drinking water (45, 46). EVALUATION OF MODE OF ACTION OF BLADDER CARCINOGENS A relatively rapid evaluation process can be performed that narrows the possible modes of action of a given agent in the production of bladder tumors in rodents and humans (44). It is essential to rst evaluate the potential of DNA reactivity of the chemical and its metabolites. This can be evaluated by various in vitro and in vivo assays, such as the Ames Salmonella mutation assay, and by various computerized chemical structure activity relationship analyses. If the agent or its metabolites are DNA reactive, it is classi ed as genotoxic and will produce bladder cancer by a direct mutational process involving DNA adduct formation, mutation, and carcinogenesis. The potency of the chemical can be greatly increased if doses are utilized that also produce an increase in proliferation (43). If the agent is non-dna reactive, it produces bladder cancer by increasing cell proliferation (43, 44). Rarely, this can be by a direct mitogenic effect as seen following the dietary administration of propoxur (47). However, bladder carcinogenesis by non-dna reactive agents most commonly involves toxicity and regeneration. This can either involve the production of urinary solids (precipitate, microcrystals, and/or calculi) or be secondary to the chemical and/or its metabolites. To evaluate the possible mode of action of the non-dna reactive agents, the chemical should be administered at doses and conditions identical to those of the chronic bioassay in which tumors were detected. During the course of the in-life portion of the experiment, fresh-voided urine is collected for examination by light and scanning electron microscopy for the presence of urinary solids and for chemical analyses (26). At a minimum, this should include evaluation of urinary ph, osmolality, creatinine, calcium, magnesium, and phosphate. Other chemical measurements can be made as appropriate. After 10 to 13 weeks, the animals are sacri ced. BrdU can be injected intraperitoneally 1 hour prior to sacri ce if labeling index is to be measured, and this is strongly encouraged. At sacri ce, the bladder is in ated with Bouin s xative, and within a few hours is washed with 70% ethanol (29). After at least 24 hours xation, the bladder is divided sagitally, with one half being divided into several slices longitudinally for processing, paraf n-embedding and for light microscopy and immunohistochemistry. The other half of the bladder is processed for examination by scanning electron microscopy. This light and scanning electron microscopic evaluation combined with labeling index determination can rapidly determine whether super cial or deep cytotoxicity is involved and if there is increased proliferation. An interim sacri ce after 4 to 6 weeks is frequently useful in detecting the very earliest changes in response to chemical treatment. SUMMARY AND CONCLUSIONS Bladder neoplasia in humans is essentially 2 diseases, a low-grade papillary lesion that does not progress or metastasize, and a second, highly malignant neoplastic process that progresses through dysplasia, carcinoma in situ, and invasion. In rats, bladder carcinogenesis usually proceeds through a process similar to the low-grade neoplastic lesion, although these eventually progress to high-grade, invasive lesions in the rat. In contrast, the mouse commonly proceeds through a process of at, nonpapillary dysplasia, carcinoma in situ, and invasive carcinoma with frequent metastases. Numerous agents have been identi edthat produce super - cial or deep cytotoxicity and regeneration, and are associated with increased incidences of bladder tumors in rodents. Similar toxic and regenerative processes appear to be involved with bladder carcinogenesis in humans related to chronic in- ammation, such as schistosomiasis and calculi. Given the techniques of modern pathology, biochemistry, and chemistry, including immunohistochemistry, light microscopy, scanning electron microscopy, urinalyses, urinary microanalyses, and X-ray dispersive spectroscopy, the possible mode of action of bladder carcinogens can be rapidly ascertained. Distinguishing between DNA reactive and non- DNA reactive agents is essential in rationally extrapolating the observations in rodent models to the human disease and exposures, particularly in ascertaining species speci city and the dose response relationship. Although rodent models have proven useful for studying bladder carcinogenesis because of numerous similarities to humans, rats and mice have distinct differences compared to humans, which can result in species speci c phenomena, such as seen with saccharin in the rat. Cognizance of these similarities and differences compared to humans are essential to provide an understanding of the chemical s mode of action and a more rational extrapolation and appropriate risk assessment for humans.

8 670 COHEN TOXICOLOGIC PATHOLOGY ACKNOWLEDGMENTS I gratefully acknowledge the assistance of Dr. Sonny Johansson, Marty Cano, Lora Arnold, and Earline Wayne- Titsworth in the preparation of this manuscript. REFERENCES 1. Johansson SL, Cohen SM (1996). Lower urinary tract. In: Anderson s Pathology, Vol 2, Damjanov I, Linder J (eds). Mosby, New York, pp Cohen SM (1983). Pathology of experimental bladder cancer in rodents. In: The Pathology of Bladder Cancer, Vol II, Cohen SM, Bryan GT (eds). CRC Press, Boca Raton, FL, pp Droller MJ (1998). Bladder cancer: State of the art care. CA-A Cancer J For Clin 48: Spruck CH III, Ohneseit PF, Gonzalez-Zuluet a M, Esrig D, Miyao N, Tsai YC, Lerner SP, Schmutte C, Yang AS, Cote R, Dubeau L, Nichols PW, Hermann GG, Steven K, Horn T, Skinner DG, Jones PA (1994). Two molecular pathways to transitional cell carcinoma of the bladder. Cancer Res 54: Epstein JI, Amin MB, Reuter VR, Mosto FK (1998). Bladder consensus conference committee: The World Health Organization/International Society of Urologic Pathology consensus classi cation of urothelial (transitional cell) neoplasms of the urinary bladder. Am J Surg Pathol 22: Cordon-Cardo C, Shenfeld J, Dalbagni G (1997). Genetic studies and molecular markers of bladder cancer. Sem Surg Oncol 13: Shirai T, Fukushima S, Hirose M, Ohshima M, Ito N (1987). Epithelial lesions of the urinary bladder in three hundred and thirteen autopsy cases. Jpn J Cancer Res 78: Wiener DP, Koss LG, Sablay B, Freed SZ (1979). The prevalence and significance of Brunn s nests, cystitis cystica and squamous metaplasia in normal bladders. J Urol 122: Ito N, Fukushima S (1986). Carcinogenesis, urinary tract, rat. In: Urinary System, Jones TC, Mohr U, Hunt RD (eds). Springer-Verlag, Berlin, pp Kunze E, Schauer A, Schatt S (1976). Stages of transformation in the development of N-butyl-N-(4-hydroxybutyl)-nitrosamine-induce d transitional cell carcinomas in the urinary bladder of rats. Z Krebsforsch 87: Fukushima S, Arai M, Cohen SM, Jacobs JB, Friedell GH (1981). Scanning electron microscopy of cyclophosphamide-induce d hyperplasia of the rat urinary bladder. Lab Invest 44: Fukushima S, Cohen SM, Arai M, Jacobs JB, Friedell GH (1981). Scanning electron microscopic examination of reversible hyperplasia of the rat urinary bladder. Am J Pathol 102: Shirai T, Ikawa E, Fukushima S, Masui T, Ito N (1986). Uracil-induced urolithiasis and the development of reversible papillomatosis in the urinary bladder of F344 rats. Cancer Res 46: Sakata T, Masui T, St. John M, Cohen SM (1988). Uracil-induced caculi and proliferative lesions of the mouse urinary bladder. Carcinogenesis 9: Becci PJ, Thompson HJ, Strum JM, Brown CC, Sporn MB, Moon RC (1981). N-butyl-N-(4-hydroxybutyl)-nitrosamine-induce d urinary bladder cancer in C57BL/6 XDBA/2F 1 mice as a useful model for study of chemoprevention of cancer with retinoids. Cancer Res 41: Ogawa K, Uzvolgyi E, St. John MS, deoliveria ML, Arnold LL, Cohen SM (1998). Frequent p53 mutations and occasional loss of chromosome 4 in invasive bladder carcinoma induced by N-butyl-N(4-hydroxybutyl) nitrosamine in B6D2F 1 mice. Mol Carcinogenesi s 21: Asamoto M, Mann AM, Cohen SM (1994). p53 Mutation is infrequent and might not give a growth advantage in rat bladder carcinogenesis in vivo. Carcinogenesis 15: Masui T, Don I, Takada N, Ogawa K, Shirai T, Fukushima S (1994). p53 mutations in early neoplastic lesions of the urinary bladder in rats treated with N-butyl-N-(4-hydroxybutyl)-nitrosamine. Carcinogenesis 15: Jacobs JB, Cohen SM, Arai M, Friedell GH, Bulay O, Urman HK (1976). Chemically induced smooth muscle tumors of the mouse urinary bladder. Cancer Res 36: Halliwell W (in press). Proliferative mesenchymal lesions (submucosal mesenchymal tumor) of the urinary bladder. Env Hlth Persp. 21. Arnold LL, Christenson R, Cano M, St. John MK, Wahl BS, Cohen SM (1997). Tributyl phosphate effects on urine and bladder epithelium in male Sprague-Dawley rats. Fund Applied Toxicol 40: Cohen SM (1998). Cell proliferation and carcinogenesis. Drug Met Rev 30: Clayson DB, Fishbein L, Cohen SM (1995). The effect of stones and other physical factors on the induction of rodent bladder cancer. Fd Chem Toxicol 33: Engelmann U, Schramek P, Baum HP, Wertmann B, Grun M, Jacobi GH (1987). Bladder carcinogenesi s in portacaval shunt rats. Urol Int 42: Jull JW (1979). The effect of time on the incidence of carcinomas obtained by the implantation of paraf n wax pellets into mouse bladder. Cancer Lett 6: Cohen SM (1995). The role of urinary physiology and chemistry in bladder carcinogenesis. Fd Chem Toxicol 33: Burin GJ, Gibb HJ, Hill RN (1995). Human bladder cancer: evidence for a potential irritation-induced mechanism. Fd Chem Toxicol 33: Rodent Bladder Carcinogenesi s Working Group (1995). Urinary bladder carcinogenesis: implications for risk assessment. Fd Chem Toxicol 33: Cohen SM, Fisher MJ, Sakata T, Cano M, Schoenig GP, Chappel CI, Garland EM (1990). Comparative analysis of the proliferative response of the rat urinary bladder to sodium saccharin by light and scanning electron microscopy and autoradiograph y. Scan Microsc 4: Smith RA, Christenson WR, Bartels MJ, Arnold LL, St. John MK, Cano M, Wahle BS, McNett DA, Cohen SM (1998). Urinary physiologic and chemical metabolic effects on the urothelial cytotoxicity and potential genotoxicity of o-phenylpheno l in male rats. Toxicol Appl Pharmacol 150: Cohen SM (1999). Calcium phosphate-containin g urinary precipitate in rat urinary bladder carcinogenesis. International Agency for Research on Cancer, Lyon, France, IARC Sci Publ 147: Ellwein LB, Cohen SM (1990). The health risks of sacchharin revisited. Crit Rev Toxicol 20: Pauli BU, Alroy J, Weinstein RS (1983). The ultrastructure and pathobiology of urinary bladder cancer. In: The Pathology of Bladder Cancer, Vol II, Cohen SM, Bryan GT (eds). CRC Press, Boca Raton, FL, pp Wu XR, Lin JH, Walz T, Haner M, Yu J, Aebi U, Sun TT (1994). Mammalian uroplakins. J Biol Chem 269: Ogawa K, Sun T-T, Cohen SM (1996). Analysis of differentiation-associate d proteins in rat bladder carcinogenesis. Carcinogenesis 17: Johansson SL, Cohen SM (1997). Epidemiology and etiology of bladder cancer. Sem Surg Oncol 13: El-Bolkainy MN (1983). Schistosomiasis and bladder cancer. In: The Pathology of Bladder Cancer, Vol I, Cohen SM, Bryan GT (eds). CRC Press, Boca Raton, FL, pp Steineck G, Plato N, Alfredsson L, Norell SE (1989). Industry-related urothelial carcinogens: Application of a job-exposure matrix to census data. Am J Ind Med 16: Birt DF, Julius AD, Hasegawa R, St. John MS, Cohen SM (1987). Effect of L-tryptophan excess and vitamin B6 de ciency on urinary bladder cancer promotion. Cancer Res 47: IARC Working Group (1999). Consensus Report. International Agency for Research on Cancer, Lyon, France, IARC Scienti c Publications 147: th Report on Carcinogens (2000). U.S. Department of Health and Human Services, National Toxicology Program, Research Triangle Park, NC, pp B3 B IARC Monographs on the Evaluation of Carcinogenic Risks to Humans (1999). Some chemicals that cause tumours of the kidney or urinary bladder in rodents and some other substances. International Agency for Research on Cancer, Lyon, France, 73:

9 Vol. 30, No. 6, 2002 COMPARATIVE BLADDER PATHOLOGY Cohen SM, Ellwein LB (1990). Cell proliferation in carcinogenesis. Science 249: Cohen SM (1998). Urinary bladder carcinogenesis. Toxicol Pathol 26: US Environmental Protection Agency (1997). Report on the expert panel on arsenic carcinogenicit y. National Center for Environmental Assessment, US Environmental Protection Agency, Washington, DC. 46. Arnold LL, Cano M, St. John MK, Eldan M, van Gemert M, Cohen SM (1999). Effects of dietary dimethylarsinic acid on the urine and urothelium of rats. Carcinogenesis 20: Cohen SM, Cano M, Johnson LS, St. John MK, Asamoto M, Garland EM, Thyssen JH, Sangha GK, Van Goethem DL (1994). Mitogenic effects of propoxur on male rat bladder urothelium. Carcinogenesis 15: