Vet Pathol 46:211–221 (2009)
Immunohistochemical Evaluation of Vascular Urinary Bladder Tumors from Cows with Enzootic Hematuria T. CARVALHO, D. NAYDAN, T. NUNES, C. PINTO,
AND
M. C. PELETEIRO
CIISA, Faculdade de MedicinaVeterina´ria, Universidade Te´cnica de Lisboa, Lisboa, Portugal (TC, TN, MCP); Veterinary Medical Teaching Hospital, School of Veterinary Medicine, University of California, Davis, CA (DN); and Servic¸o de Desenvolvimento Agra´rio de Sa˜o Miguel, Ponta Delgada, Ac¸ores, Portugal (CP) Abstract. Fifty-six endothelial-derived urinary bladder tumor samples collected from 26 animals with bovine enzootic hematuria were selected for immunohistochemical studies. Expression of factor VIII–related antigen (FVIIIra), CD31, muscle-specific actin, uroplakin III (UPIII), and the cell cycle– related proteins cyclin D1 and p53 was evaluated in hemangiomas, ‘‘hemangioendotheliomas’’ (a vascular tumor that histologically is intermediate in appearance between a hemangioma and a conventional hemangiosarcoma), and hemangiosarcomas. Although CD31 expression was seen in all endothelial tumors tested, FVIIIra was not expressed in poorly differentiated endothelial tumor cells from solid areas or in 7 muscle-invasive hemangiosarcomas. Cyclin D1 overexpression was seen in 53% of hemangiomas, 82% of hemangioendotheliomas, and 95% of hemangiosarcomas. P53 immunoreactivity was only seen in muscle-invasive hemangiosarcomas. The UPIII staining pattern, normally very intense on the apical aspect and cytoplasm of superficial urothelial cells, was altered in the urothelium in an estimated 25% of hemangiomas, most hemangioendotheliomas, and most hemangiosarcomas. In conclusion, CD31 is a better marker than FVIIIra in the characterization of bovine endothelial tumors. The cell cycle regulatory pathways involving cyclin D1 and p53 seem to be impaired in endothelial urinary bladder tumors, p53 immunoreactivity positively correlating with enhanced invasion. Key words: bovine enzootic hematuria; immunohistochemistry; urinary bladder cancer; cyclin D1; p53; uroplakin III; vascular tumor.
Introduction Chronic ingestion of bracken fern (Pteridium spp.), the only plant known to cause cancer naturally,14,45,46 produces a neoplastic syndrome in cows known as bovine enzootic hematuria (BEH), which is characterized by the occurrence of hematuria and urinary bladder tumors.10 Moreover, in approximately 62% of the cases, multiple tumors occur in the same bladder.10 This syndrome occurs worldwide and is very common in cows on Sa˜o Miguel Island, Azores, Portugal, where bracken grows endemically. As a complexing factor, bovine papillomavirus type 2 (BPV-2) is frequently associated with bracken-induced tumors of the urinary bladder,4,7 and 84 of 200 bladder tumors collected at the slaughterhouse on Sa˜o Miguel Island displayed BPV-2 DNA.35 Exposure to the BEH carcinogen(s) causes the development of bladder cancer with diverse morphologies.10,34 Metastases are known to occur, namely to the lumboaortic lymph node,38 but no extensive study on the frequency of metastatic disease has been performed so far. Histologically,
among the epithelial-derived tumors, transitional cell carcinoma is the most common neoplasm.10,34 Endothelial-derived urinary bladder tumors are also quite common in BEH,8,10 ranging from welldifferentiated cavernous hemangiomas to solid hemangiosarcomas. This is a very heterogeneous group of tumors, and although some tumors present as clear-cut hemangiomas and hemangiosarcomas, others possess nearly all of the features of a hemangioma but also display increased cellularity and mitotic activity or contain well-formed vascular channels not well-demarcated and confined to the lamina propria, as is the case in most hemangiomas. Hence, in BEH, some authors use the designation ‘‘hemangioendothelioma’’ for those vascular tumors that histologically are intermediate in appearance between a hemangioma and a conventional hemangiosarcoma.10,32,33 Two of the antibodies most commonly used in the diagnosis of endothelial tumors are directed to FVIIIra and CD31. FVIIIra, also referred to as von Willebrand factor, is a soluble plasma protein, carrier of factor VIII, and present in the Weibel-Pallade bodies of endothelial cells.2 It is commonly used as a marker for endothelial cells in
211
212
Carvalho, Naydan, Nunes, Pinto, and Peleteiro
humans but was seldom expressed in some poorly differentiated vascular tumors.23,36 The platelet/ endothelial cell adhesion molecule CD31 (PECAM-1) is a transmembrane glycoprotein, expressed in platelets, monocytes, and endothelial intercellular junctions. Angiosarcomas are among the rarest malignancies in human pathology, and only 12 cases have been described in the bladder.40 Hemangioendothelioma is a rare vascular tumor of low-grade or borderline malignancy that arises in several organ systems, skin, and soft tissue,6,12 but to our knowledge, in humans no cases have been described in the urinary system. Hemangioma is the most common vascular lesion described in the World Health Organization classification of urinary bladder tumors in humans and is supposedly of congenital origin.11 Spontaneous bladder tumors of urothelial type are common in dogs but rare in all other species.28 In dogs, hemangiomas, ‘‘hemangioendotheliomas,’’ and hemangiosarcomas are mostly limited to other organ systems, including the spleen, right atrium, skin, and subcutis; they are less common in other species.17,18 Bracken fern carcinogens, often acting synergistically with BPV-2, interfere with multiple pathways encompassing DNA damage/repair mechanisms,3 activation of specific oncogenes,39,41 and growth factor receptors,5 which eventually perturb control of cell proliferation. Cellular proliferation is, in part, regulated by D-type cyclins,43 and cyclin D1, one of its members, is specifically involved in the G1/Sphase transition. Two recently published studies report cyclin D1 expression in human breast hemangiomas and hemangiosarcomas44 and in hemangiosarcomas of the spleen, liver, kidney, and soft tissue of dogs.53 In the former study, no correlation was seen between cyclin D1 immunoreactivity and malignancy grade.44 In 2007, Yonemaru et al. found, using immunohistochemistry, that all canine hemangiosarcomas tested displayed cyclin D1 expression, while none of the canine subcutaneous hemangiomas were found to express this protein. Furthermore, control of cell growth and maintenance of genome stability are controlled by the nuclear phosphoprotein p53. Mutations of p53 occur in 50% of human cancers,22,51 including softtissue angiosarcoma,27,50 and usually lead to p53 protein accumulation in the nucleus. The current study of BEH endothelial urinary bladder tumor specimens included the investigation of 1) the expression levels of structural proteins, 2) the possible correlation of structural proteins with pathologic parameters, and 3) the identification of
Vet Pathol 46:2, 2009
altered expression of the cell-cycle regulatory factors cyclin D1 and p53. Materials and Methods Sample collection From December 2002 to March 2003, a total of 4,026 Friesian cows (after calving and .2 years of age) were slaughtered in the abattoir of Sa˜o Miguel Island. Rejection due to bladder lesions occurred in 612 animals (15.2%), from which 26 displaying vascular tumors were selected, at random, for the present study. Up to 90% of the animals had hematuria at or before slaughter, and all were known to have grazed on bracken-infested pasture. Their ages ranged from 3 to 12 years (mean, 6). Multiple samples were collected from bladders exhibiting more than 1 tumor (18/26), each sample corresponding to discrete, nonconfluent tumor. Hence, 56 bovine urinary bladder tumors of both endothelial origin were available for immunohistochemical studies. Regional lymph nodes and internal organs were not accessed for metastasis. Tumors were formalin fixed and paraffin embedded, and 3-mm sections were cut for routine histologic diagnosis (hematoxilin and eosin, HE) and immunohistochemical studies. One histologic section was performed in tumors ,3 cm in size, and 2 sections in tumors .3 cm. Histologic typing and immunohistochemical studies Histologic typing of vascular tumors was based on World Health Organization established criteria for the diagnosis of urinary bladder tumors in domestic animals, recent publications on BEH-associated bladder lesions,10,28,34 and diagnostic histopathology criteria for human blood vessel tumors.6 Endothelial tumors were classified as hemangiomas, hemangioendotheliomas, and hemangiosarcomas. Although the designation hemangioendothelioma is infrequently used in veterinary pathology,10,16,26,32,33,52 this category was adopted in the present study to deal with the heterogeneity within this group and assist the correlation between the expression levels of selected proteins and tumor grade. This category includes lesions with microscopic features intermediate between hemangiomas and hemangiosarcomas in terms of cellularity, mitotic activity, presence of well-formed vascular channels, and degree of invasion. Hemangioendotheliomas show focal invasion of deeper layers of the lamina propria and increased celullarity, opposed to hemangiomas, but have welldifferentiated vessels and lack moderate to high mitotic activity, anastomosing vascular spaces, and increased cellular atypia, characteristics of the hemangiosarcoma. Classification was based on HE staining for all endothelial tumors except 5 solid hemangiosarcomas, in which CD31-positive immunohistochemical staining was contributory to the final diagnosis. Endothelial tumors were immunohistochemically stained for FVIIIra, PECAM-1, MSA, cyclin D1, p53, and UPIII, as indicated in Table 1.
Immunohistochemistry of Vascular Bovine Urinary Bladder Tumors
Vet Pathol 46:2, 2009
213
Table 1. Antibodies used and technique details. Antigen
Clone/Source
Retrieval
Dilution
Detection System/Source
Factor VIII– A0082/Dako, Glostrup, Proteinase K 10 minutes 60 minutes @ 1 : 300 related Denmark @ room temperature antigen CD31 PECAM-1 (M-20)/Santa Steam 20 minutes 60 minutes @ 1 : 200 Cruz Biotechnology, @ 98uC in citrate Inc, Santa Cruz, CA
Cyclin D1 p53 Uroplakin III
18-0220, clone AM29/ Zymed Laboratories, San Francisco, CA MCA1701, clone DO1/ Serotec Inc, UK
Microwave 20 minutes in EDTA pH 8
Overnight @ 1 100
Microwave 15 minutes in citrate
Overnight @ 1 : 300
PRO651108/clone AU1 Proteinase K 10 minutes 60 minutes (ready to Research Diagnostics, @ room temperature use) Inc, Concord, MA
Antisera validation
Histostain-plus/Zymed Laboratories, San Francisco, CA Mouse anti-goat IgG and 4plus immunoperoxidase fetection dystem/ Biocare Medical, Concord, CA ImmPRESS/Vector Laboratories, Burlingame, CA Histostain-plus/Zymed Laboratories, San Francisco, CA ImmPRESS/Vector Laboratories, Burlingame, CA
parameters, namely tumor grade/stage, was evaluated using 2-sided Fisher’s exact test. Statistical analyses were performed using SPSS Version 14.0 (Chicago, IL, USA). The associations were considered significant when P values were less than 0.05.
To determine possible cross-reaction of the antiUPIII antibody with nonurothelial tumors, 8 hemangiomas, 6 hemangioendotheliomas, and 8 hemangiosarcomas were subject to immunohistochemistry with this antibody. Normal blood vessels served as internal positive controls for FVIIIra/CD31, and urothelium served as internal positive controls for UPIII. All tumors were tested for each antibody within a short period. Hence, variations in antigen retrieval technique overtime were minimal and unimportant to the outcome of the technique. Human mantle-cell lymphoma was included as a positive control for human cyclin D1 antibody. Heatinduced epitope retrieval was essential for immunodetection of cyclin D1 and was carried out by placing sections in a high pH buffer solution (pH 5 8.0) in a microwave (20 minutes, 750 W). The aggressiveness of the technique occasionally compromised cell/tissue morphology, and for this reason it was not possible to assess cyclin D1 staining in 9 endothelial tumors. Bovine ocular squamous cell carcinoma was used as a positive control for p53 immunodetection. The DO-1 antibody utilized in the study recognizes an epitope present in the N terminus of p53 and cross-reacts with both wild-type and mutant human protein (amino acids 20–25). In this study, a 20% cutoff value (.100 labeled tumor cells per 500 tumor cells randomly counted) was accepted as positive staining for p53 and cyclin D1.25,49 Negative controls for immunohistochemistry were processed identically to test slides, but the primary antibody was omitted.
Summarized results are shown on Table 2. Based on evaluation of the HE-stained tissue sections, the 56 endothelial-derived bovine urinary bladder tumors were diagnosed as follows: 21 hemangiomas, 23 hemangiosarcomas, and 12 neoplasms that had microscopic features suggestive of a low-grade tumor that were classified as hemangioendotheliomas (Fig. 1). Growth patterns of vascular tumors ranged from vasoformative to solid, and tumor vessels varied from well formed with a well-defined lumen in hemangiomas to more poorly defined in hemangiosarcomas. The proportion of vasoformative and solid components varied, and only 6 hemangiosarcomas were found to display a full solid growth pattern and epithelioid morphology. In 3 hemangioendotheliomas and 2 other hemangiosarcomas, neoplastic cells were also predominantly epithelioid, with rare spindled areas. Muscle invasion was only seen in 14 hemangiosarcomas. Lymphoplasmacytic and neutrophilic inflammatory infiltrates (often found together, present within the tumor stroma and tumor margins) varied from prominent to slight or absent.
Statistical analysis
CD31 and FVIIIra
Association between presence/absence of immunoreactivity for the antibodies tested and pathologic
Vascular tumors were further characterized through the use of endothelial markers. CD31
Results
214
Carvalho, Naydan, Nunes, Pinto, and Peleteiro
Vet Pathol 46:2, 2009
Table 2. Histologic and immunohistochemical characterization of bladder tumors.
Animal*
1 2 3 4
5 6 7
8 9
10
11 12
13 14
15 16 17
18 19 20 21 22
23
Immunohistochemistry{
Tumor Type
Muscle Invasion
p53
Cyclin D1
UPIII
FVIIIra
CD31
Hemangiosarcoma Hemangioma Hemangioma Hemangiothelioma Hemangiothelioma Hemangiothelioma Hemangioma Hemangioma Hemangiosarcoma Hemangiothelioma Hemangiosarcoma Hemangiothelioma Hemangiosarcoma Hemangioma Hemangiosarcoma Hemangiosarcoma Hemangiosarcoma Hemangioma Hemangioma Hemangiothelioma Hemangioma Hemangiosarcoma Hemangiosarcoma Hemangiosarcoma Hemangiosarcoma Hemangiothelioma Hemangioma Hemangioma Hemangioma Hemangiosarcoma Hemangioma Hemangiosarcoma Hemangiothelioma Hemangiothelioma Hemangioma Hemangioma Hemangiosarcoma Hemangioma Hemangioma Hemangiosarcoma Hemangioma Hemangiosarcoma Hemangioma Hemangiothelioma Hemangiosarcoma Hemangiosarcoma Hemangiothelioma Hemangiosarcoma Hemangiosarcoma Hemangiosarcoma
No No No No No No No No No No No No Yes No Yes Yes Yes No No No No Yes Yes Yes No No No No No Yes No Yes No No No No Yes No No No No Yes No No No No No Yes No Yes
2 2 2 + 2 + 2 2 2 2 2 2 2 2 + + + 2 2 2 2 + + + 2 2 2 2 2 + + 2 2 + 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2
+ cm + + + 2 2 2 + + + cm + + cm cm cm cm cm + cm cm + + + + + + 2 + + + + + + 2 + 2 + + 2 2 2 2 + + + + + +
{ 2 2 { { { 2 { { { { { { 2 { { nd 2 2 { { nd { 2 nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd nd
+ + + + + + + + + + + + + + 2 2 2 + + + + 2 2 2 + + + + + + + + + + + bk + + + + + 2 + + + + + + + +
+ + + + bk bk + bk + + + + + + + + + + bk bk bk + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
Vet Pathol 46:2, 2009
215
Immunohistochemistry of Vascular Bovine Urinary Bladder Tumors
Table 2. Continued.
Animal*
24
25 26
Immunohistochemistry{
Tumor Type
Muscle Invasion
p53
Cyclin D1
UPIII
FVIIIra
CD31
Hemangioma Hemangiothelioma Hemangiosarcoma Hemangioma Hemangioma Hemangiosarcoma
No No No No No Yes
2 2 2 2 2 2
+ + + + 2 +
nd nd nd nd nd nd
+ + bk + + +
+ + + + + +
* Multiple samples were collected from animals with urinary bladders exhibiting more than 1 tumor (18/26), each sample corresponding to discrete, nonconfluent lesions. { UPIII 5 uroplakin III; FVIIIra 5 factor VIII–related antigen; 2 5 tumor cells showed no immunostaining; + 5 some/all tumor cells showed positive immunostaining; bk 5 excessive background staining within the vascular lumen that impeded proper interpretation; cm 5 compromised tissue morphology due to aggressive high pH antigen retrieval technique; nd 5 not determined. { Presence of atypical UPIII pattern in the urothelium covering the endothelial tumor.
expression was detected in all endothelial tumors examined, including muscle-invasive hemangiosarcomas (Fig. 2). All neoplastic cells within the tumor consistently displayed CD31 immunostaining, which was intense at the cell membrane level. FVIIIra expression was seen in all hemangiomas (20/20) and hemangioendotheliomas (12/12), and in 15 of 20 hemangiosarcomas (67%). The hemangiosarcomas negative for FVIIIra (6 solid and 1 poorly vasoformative) were all muscle-invasive lesions, and the endothelial origin of these tumors was confirmed by their positive reaction with CD31. Undifferentiated endothelial tumor cells, generally found in areas of solid growth, did not express FVIIIra. The staining was diffusely granular within the cytoplasm of endothelial cells, with occasional staining of plasma within vascular space (Fig. 3). It was not possible to assess FVIIIra and CD31 expression in 2 and 6 endothelial tumors, respectively, because of high general background within the vascular space.
Fig. 1. Histologic classification of urinary bladder tumors of endothelial origin.
Cyclin D1
Normal urothelium surrounding the tumor lesion demonstrated nuclear staining for cyclin D1 in ,5% of basal cells. Moderate to strong nuclear staining was identified in 53% hemangio-
Fig. 2. Hemangiosarcoma in bovine urinary bladder. Streptavidin-peroxidase stain with Mayer’s hematoxylin counterstain. Intense CD31 staining in a solid hemangiosarcoma, with immunoreactivity in all neoplastic cells. Bar 5 100 mm. Inset: Detail of the staining, intense in the cytoplasmic membrane (arrow). Bar 5 30 mm.
216
Carvalho, Naydan, Nunes, Pinto, and Peleteiro
Vet Pathol 46:2, 2009
Fig. 3. Hemangiosarcoma in bovine urinary bladder. Factor VIII–related antigen reaction in a vasoformative hemangiosarcoma. Distinct granular cytoplasmic staining (arrow). Immunohistochemistry using streptavidinperoxidase and counterstained with Mayer’s hematoxylin. Bar 5 30 mm. Fig. 4. Hemangiosarcoma in bovine urinary bladder. Strong nuclear cyclin D1 reaction in the majority of neoplastic cells in a poorly vasoformative hemangiosarcoma. Immunohistochemistry using ImmPRESS-peroxidase and counterstained with Mayer’s hematoxylin. Bar 5 100 mm. Inset: Detail of the staining. Bar 5 30 mm.
mas (9/17), 82% hemangioendotheliomas (9/11), and 100% hemangiosarcomas (19/19). Cyclin D1 expression positively correlates with malignancy grade in vascular tumors. The staining was mainly nuclear, although occasional weak cytoplasmic immunoreactivity was detected (Fig. 4). p53
Immunoreactive p53 was detected in 30% of the hemangiosarcomas (7/23, 6 solid and 1 poorly
vasoformative) and 25% of the hemangioendotheliomas (3/12) (Table 3). Nuclear p53 immunoreactivity was seen in more than 90% of hemangiosarcoma tumor cells (Fig. 5). All p53-positive hemangiosarcomas were muscle-invasive lesions. Staining intensity was moderate to strong, and cytoplasmic p53 staining was never observed. Nuclear p53 positivity graded high in 90% of cells was also seen in 1 hemangioma. In this particular case, not only the tumor but also all capillary
Table 3. Factor VIII–related antigen (FVIIIra), CD31, uroplakin III (UPIII), p53, and cyclin D1 immunoreactivity in relation to pathologic parameters in endothelial-derived bladder tumors.
Parameters
UPIII1 FVIIIra Aberrant CD31 Positive n/ P Positive n/ P Pattern n/ P Total (%) Value* Total (%) Value Total (%) Value
p53 Positive n/Total (%)
Cyclin D1 Positive P n/Total P Value (%) Value
Classification Hemangioma 20/20 (100) .002 18/18 (100) NS{ 2/8 (25) ,.01 1/21 (5) NS 9/17 Hemangioendothelioma 12/12 (100) 9/9 (100) 6/6 (100) 3/12 (25) 9/11 Hemangiosarcoma 15/22 (68) 23/23 (75) 7/8 (87.5) 7/23 (30) 18/19 Staging{ Suburothelial 8/8 (100) ,.03 7/7 (100) NS 3/3 (100) NS 0/9 (0%) ,.02 9/9 Muscle invasive 7/14 (50) 2/5 (40) 4/5 (80) 7/14 (50) 9/10
(53) ,.02 (82) (95) (100) NS (90)
* Association between presence/absence of immunoreactivity for the antibodies tested and pathologica parameters, namely tumor grade/stage, was evaluated using 2-sided Fisher’s exact test. { NS 5 not significant. { Only hemangiosarcomas were considered for analysis of markers’ expression versus tumor stage. 1 These values refer to the presence of atypical UPIII pattern in the urothelium covering the endothelial tumor.
Vet Pathol 46:2, 2009
Immunohistochemistry of Vascular Bovine Urinary Bladder Tumors
217
Fig. 5. Hemangiosarcoma in bovine urinary bladder. Strong nuclear p53 reaction in the majority of neoplastic cells in a solid hemangiosarcoma. Streptavidin-peroxidase stain with Mayer’s hematoxylin counterstain. Bar 5 100 mm. Inset: Detail of the staining. Bar 5 30 mm. Fig. 6. Hemangiosarcoma in bovine urinary bladder. Urothelium covering endothelial tumor with strong nuclear p53 reaction in the majority of basal cells. Streptavidin-peroxidase stain with Mayer’s hematoxylin counterstain. Bar 5 100 mm. Inset: There is marked atypia of basal cells. Bar 5 30 mm. Fig. 7. Hemangioendothelioma in bovine urinary bladder. Urothelium covering endothelial tumor displays atypical uroplakin III reaction, with intense and diffuse staining of cells in the deeper layers (arrow). ImmPRESSperoxidase with Mayer’s hematoxylin counterstain. Bar 5 50 mm. Fig. 8. Hemangiosarcoma in bovine urinary bladder. Heterogenous uroplakin III staining of urothelial cells in von Bru¨nn Nests within a hemangiosarcoma, with patchy intracytoplasmic positivity (asterisk); superficial cells display normal uroplakin III expression, more intense in the apical border of the cytoplasmic membrane (arrow). ImmPRESS-peroxidase with Mayer’s hematoxylin counterstain. Bar 5 30 mm. Inset: Detail of the staining. Bar 5 10 mm.
218
Carvalho, Naydan, Nunes, Pinto, and Peleteiro
endothelial cells exhibited p53 immunoreactivity. All p53-positive tumors were also positive for cyclin D1, except for 1 hemangioendothelioma. P53 staining was also found in dysplastic urothelium covering in the vicinity of 2 endothelial malignancies, 1 hemangioendothelioma, and 1 hemangiosarcoma (Fig. 6). UPIII
UPIII expression was not detected in the neoplastic endothelial cells of the 22 vascular tumors evaluated for this marker. A noteworthy finding in this series was that atypical UPIII expression pattern was seen specifically in the urothelium, covering 25% hemangiomas (2/8), 100% hemangioendotheliomas (6/6), and 88% hemangiosarcomas (7/8). In contrast with the normal UPIII staining pattern, very intense on the apical aspect and cytoplasm of superficial urothelial cells, this atypical staining was characterized by discontinuity of the apical pattern, heterogeneous staining of cells in the deeper layers, and presence of clustered intermediate cells with irregular membranous staining and a diffuse or patchy stain in the cytoplasm (Figs. 7, 8). Only 1 of these tumors displayed altered uroplakin distribution within the urothelium distant from the vascular tumor. No overlap in staining pattern was observed between p53-positive urothelial cells and atypical UPIII expression. Discussion The current study documents the use of immunohistochemistry as a tool to characterize endothelial-derived urinary bladder tumors developing as a result of bracken fern carcinogens. No loss of CD31 expression was registered in endothelial tumors, and generally all neoplastic cells within these lesions were CD31 positive, as has previously been reported by other authors.36 CD31 has an extracellular domain that mediates endothelial cellto-cell contact and a cytoplasmic domain with potential sites for phosphorylation after cellular activation,30 and it has been proposed to be involved in interactive events during angiogenesis. CD31 expression by tumor cells may therefore represent a selective advantage for tumor progression and invasion. In contrast to CD31, FVIIIra expression was lost in high-grade, muscle-invasive hemangiosarcomas of solid phenotype, negatively correlating with tumor grade and stage. Through ultrastructural studies in soft-tissue angiosarcomas, Weibel-Palade bodies, which store FVIIIra, were not observed in neoplastic endothe-
Vet Pathol 46:2, 2009
lium, whereas they were readily identified in the endothelium of capillaries of the surrounding fibrous stroma.27 This is in agreement with our findings. Previous reports also noted that FVIIIra is seldom expressed in poorly differentiated vascular tumors.23,36 Hence, although FVIIIra and CD31 are both specific for the characterization of a vasoformative phenotype, only the latter was found to be consistently expressed by neoplastic endothelial cells, even in poorly differentiated/solid neoplasms. FVIIIra immunoreactivity was occasionally seen in epithelial tumor cells, in a subset of urothelial carcinomas. This was unexpected because this rabbit-anti human FVIIIra antibody is supposed to react exclusively with endothelial cells, megakaryocytes, and platelets in both human and bovine tissue; production of this glycoprotein by cells other than endothelial cells and megakaryocytes has not been described.48 It is possible that an FVIIIra-like molecule is being coexpressed in bladder carcinoma cells. Alternatively, partial cross-reactivity with another epitope in the urothelium may explain the spurious staining. Urothelium overlaying endothelial malignancies frequently displayed atypical UPIII expression, quite different from that of normal bladder urothelium and urothelium adjacent to the endothelial tumors. According to other authors, this pattern of expression is diagnostic of severe dysplasia/carcinoma in situ.1,37 It is unclear whether urothelial lesions are carcinogen induced and develop together with the endothelial tumor or if a specific microenvironment in the underlying suburothelial space favors the onset of dysplastic changes in the superimposed urothelium. Evans and Mason demonstrated, for the first time in 1965, undeniable evidence for bracken carcinogenicity.15 Since then, several studies have shown increased numbers of chromosome aberrations in cattle maintained on natural bracken fern29 or with overt enzootic hematuria.24,29 From the various toxic compounds that have been isolated from bracken fern, unequivocal urinary bladder carcinogenesis has been proved only for ptaquiloside.19,20 At sublethal or subtoxic levels, ptaquiloside causes genotoxicity through the formation of a number of different DNA adducts, the main labile adducts occurring at the N-3 of adenine, and to a minor extent with the N-7 of guanine.21,31,42,47 H-ras activation is also known to be an early event in a rat model of ptaquilosideinduced carcinogenesis.41 Overexpression of this protein has been documented in cases of bladder neoplasia and cystitis in animals naturally exposed to bracken fern.39 Cyclin D1 and p53 overexpres-
Vet Pathol 46:2, 2009
Immunohistochemistry of Vascular Bovine Urinary Bladder Tumors
sion was seen in endothelial-derived bovine urinary bladder tumors. Regarding cyclin D1, this overexpression was positively associated with high tumor grade and occurred in 53% of hemangiomas (9/17), 82% of hemangioendotheliomas (9/11), and 95% of hemangiosarcomas (18/19). Cyclin D1 expression was recently reported in breast vascular tumors in humans, but no correlation with malignancy grade was seen.44 On the other hand, expression of this protein was seen in 39 of 39 canine hemangiosarcomas of the spleen, liver, kidney, and soft tissue, but none of the subcutaneous hemangiomas tested (n 5 10) displayed cyclin D1 immunoreactivity.53 In our study, cyclin D1 expression positively correlates with malignancy grade occurring in benign, low-grade, and high-grade vascular tumors. It is possible that this pathway is being differentially regulated in endothelial tumors arising in distinct organ systems. Abnormal p53 expression was seen in bovine urinary bladder cancer, being present only in muscle-invasive hemangiosarcomas. P53 nuclear accumulation and consequent detection by immunohistochemistry can be due to gene mutation or activation of mechanisms that influence p53 stability.54 Mutation and deletion of the p53 tumor suppressor gene are the most frequent genetic alterations of epithelial-derived bladder cancer in humans and are associated with advanced disease.13 Within soft-tissue angiosarcomas, also in humans, 20–50% display p53 overexpression.27,54 The results presently obtained are in agreement with previous studies, and a significant correlation occurred between strong positive reactions with p53 and an invasive phenotype in both endothelial and urothelial bladder cancer. BPV-2 status of the tumors was not accessed in this study, and hence no further details can be add to our understanding of its relevance in these cellular specific cellular pathways, in BEH. In humans, within the spectrum of hemangioendothelioma, several distinct entities are recognized, differing not only histologically but also in terms of epidemiology and biologic behavior.6,12 Here, although this term refers exclusively to a histologic pattern, it was seen that tumors within this category are also immunohistochemically distinct from hemangiomas and hemangiosarcomas. FVIIIra expression is maintained, similar to hemangiomas, but the percentage of tumors with cyclin D1 and p53 overexpression resembles that of hemangiosarcomas. It seems that, phenotipically, these lesions are well differentiated, but the cellular pathways involving certain oncogenes and tumor suppressor genes are already impaired.
219
In BEH no association between histologic grade/stage of bladder lesions and their biologic behavior can be made, because samples are obtained at slaughter and there is, therefore, no treatment or follow up. Nevertheless, there is unquestionable merit in the use of animal models for studying the pathogenesis of endothelial tumors, and, in fact, molecular studies in bovine urinary bladder tumors already provided evidence, for the first time in spontaneous tumors, of the importance of the angiogenesis/chemokine axis in endothelial and urothelial malignancies.9 Hence, results presented here can be used to the benefit of other species (other domestic animals and man) with the goal of further predicting the outcome of the host. In conclusion, the data presented in this study show that while CD31 expression was retained by all endothelial tumors examined, FVIIIra was lost in hemangiosarcomas of solid phenotype and was only seldom expressed in poorly differentiated vascular tumors, being therefore less sensitive than CD31 for the characterization of bovine endothelial tumors. The cell cycle regulatory pathways involving cyclin D1 and p53 seem to be altered in urinary bladder tumors in BEH. As an extention of this work, immunohistochemical studies will further be applied to epithelial-derived bovine urinary bladder tumors, and the distinctive features between the 2 different tumor histotypes will be further addressed. Acknowledgements The authors thank Professor Verena K. Affolter for her critical reading of the manuscript. We also acknowledge Dr. Iolanda Fernandes and Mrs. Maria do Rosa´rio Luis for expert technical assistance. This research was supported by the Foundation for Science and Technology, project POCTI 34320, and by the Interdisciplinary Centre of Research in Animal Health (CIISA).
References 1 Ambrosio V, Borzacchiello G, Bruno F, Galati P, Roperto F: Uroplakin expression in the urothelial tumors of cows. Vet Pathol 38:657–660, 2001 2 Arnout J, Hoylaerts MF, Lijnen HR: Haemostasis. Handb Exp Pharmacol 176:1–41, 2006 3 Beniston RG, Morgan IM, O’Brien V, Campo MS: Quercetin, E7 and p53 in papillomavirus oncogenic cell transformation. Carcinogenesis 22:1069–1076, 2001 4 Borzacchiello G, Iovane G, Marcante ML, Poggiali F, Roperto F, Roperto S, Venuti A: Presence of bovine papillomavirus type 2 DNA and expression of the viral oncoprotein E5 in naturally occurring
220
5
6
7 8
9
10 11
12
13
14 15 16 17
18
19
Carvalho, Naydan, Nunes, Pinto, and Peleteiro
urinary bladder tumors in cows. J Gen Virol 84:2921–2926, 2003 Borzacchiello G, Russo V, Gentile F, Roperto F, Venuti A, Nitsch L, Campo MS, Roperto S: Bovine papillomavirus E5 oncoprotein binds to the activated form of the platelet-derived growth factor [beta] receptor in naturally occurring bovine urinary bladder tumors. Oncogene 25:1251–1260, 2006 Calonje E, Fletcher DM: Tumors of blood vessels and lymphatics. In: Diagnostic Histopathology of Tumors, ed. Fletcher CDM, vol. 1, pp. 43–77. Churchill Livingstone, Edinburgh, 1995 Campo MS: Bovine papillomavirus and cancer. Vet J 154:175–188, 1997 Campo MS, Jarrett WF, Barron R, O’Neil BW, Smith KT: Association of bovine papillomavirus type 2 and bracken fern with bladder cancer in cattle. Cancer Res 52:6898–6904, 1992 Carvalho T, Elias AP, Nunes T, Peleteiro MC, Dias S: Chemo-angiogenic profile of bovine urinary bladder tumors distinguishes urothelial carcinomas from hemangiosarcomas. Vet Immunol Immunopathol 121:344–358, 2008 Carvalho T, Pinto C, Peleteiro MC: Urinary bladder lesions in bovine enzootic hematuria. J Comp Pathol 134:336–346, 2006 Eble JN, Sauter G, Epstein JI, and Sesterhenn IA, eds.: World Health Organization Classification of Tumors. Pathology and Genetics of Tumors of the Urinary System and Male Genital Organs. IARC Press, Lyon, France, 2004 Enzinger FM, Weiss SW: Hemangioendothelioma: vascular tumors of intermediate malignancy. In: Soft Tissue Tumors, ed. Enzinger FM and Weiss SW, 3rd ed., pp. 627–639. CV Mosby, St. Louis, 1995 Erill N, Colomer A, Verdu M, Roman R, Condom E, Hannaoui N, Banus JM, Cordon-Cardo C, Puig X: Genetic and immunophenotype analyses of TP53 in bladder cancer: TP53 alterations are associated with tumor progression. Diagn Mol Pathol 13:217–223, 2004 Evans IA: Bracken carcinogenicity. Res Vet Sci 26:339–348, 1979 Evans IA, Mason J: Carcinogenic activity of bracken. Nature 208:913–914, 1965 Finnie JW, Leong AS, Swift JG: Two hemangioendotheliomas in a bovine brain. J Comp Pathol 108:393–398, 1993 Goldschmidt MH, Hendrick MJ: Tumors of the skin and soft tissues. In: Tumors in Domestic Animals, ed. Meuten DJ, 4th ed., pp. 99–101. Iowa State Press, IA, 2002 Hendrick MJ, Mahaffey EA, Moore FM, Vos JH, Walder EJ: WHO Histological Classification of Mesenchimal Tumors of Skin and Soft Tissues, ed. Schulman FY, 2nd series, Vol. II, pp. 22–24. Armed Forces Institute of Pathology, Washington, DC, 1988 Hirono I, Kono Y, Takahashi K, Yamada K, Niwa H, Ojika M, Kigoshi H, Niiyama K, Uosaki Y:
20
21
22 23 24
25
26 27 28
29
30 31
32 33
Vet Pathol 46:2, 2009
Reproduction of acute bracken poisoning in a calf with ptaquiloside, a bracken constituent. Vet Rec 115:375–378, 1984 Hirono I, Ogino H, Fujimoto M, Yamada K, Yoshida Y, Ikagawa M, Okumura M: Induction of tumors in ACI rats given a diet containing ptaquiloside, a bracken carcinogen. J Natl Cancer Inst 79:1143–1149, 1987 Kushida T, Uesugi M, Sugiura Y, Kigoshi H, Tanaka H, Hirokawa J, Ojika M, Yamada K: DNA damage by ptaquiloside, a potent bracken carcinogen: detection of selective strand breaks and identification of DNA cleavage products. J Am Chem Soc 116:479–486, 1994 Lane DP: p53 and human cancers. Br Med Bull 50:582–599, 1994 Leong ASY, Cooper K, Leong FJWM: Manual of Diagnostic Antibodies for Immunohistology, p. 167, Oxford University Press, Oxford, 1999 Lioi MB, Barbieri R, Borzacchiello G, Dezzi S, Roperto S, Santoro A, Russo V, Roperto F: Chromosome aberrations in cattle with chronic enzootic haematuria. J Comp Pathol 131:233–236, 2004 Lu ML, Wikman F, Orntoft TF, Charytonowicz E, Rabbani F, Zhang Z, Dalbagni G, Pohar KS, Yu G, Cordon-Cardo C: Impact of alterations affecting the p53 pathway in bladder cancer on clinical outcome, assessed by conventional and array-based methods. Clin Cancer Res 8:171–179, 2002 Machida N, Arimura T, Otagiri Y, Kiryu K, Oka T: Epithelioid hemangioendothelioma of the lung in a dog. J Comp Pathol 119:317–322, 1998 Meis-Kindblom JM, Kindblom LG: Angiosarcoma of soft tissue: a study of 80 cases. Am J Surg Pathol 22:683–697, 1998 Meuten DJ, Everitt J, Inskeep W, Jacobs RM, Peleteiro M, Thompson KJ: Urinary bladder tumors. In: WHO Histological Classification of Tumors of the Urinary System of Domestic Animals, ed. Schulman FY, 2nd series, Vol. XI, pp. 26–37. Armed Forces Institute of Pathology, Washington, DC, 2004 Moura JW, Stocco dos Santos RC, Dagli ML, D’Angelino JL, Birgel EH, Becak W: Chromosome aberrations in cattle raised on bracken fern pasture. Experientia 44:785–788, 1988 Newman PJ: The biology of PECAM-1. J Clin Invest 100:S25–S29, 1997 Ojika M, Wakamatsu K, Niwa H, Yamada K: Ptaquiloside, a potent carcinogen isolated from bracken fern Pteridium aquilinum var. laticulum. Structure elucidation based on chemical and spectral evidence and reactions with amino acids, nucleosides and nucleotides. Tetrahedron 43:5261–5274, 1987 Ozkul IA, Aydin Y: Tumours of the urinary bladder in cattle and water buffalo in the Black Sea region of Turkey. Br Vet J 152:473–475, 1996 Pamukcu AM, Price JM, Bryan GT: Naturally occurring and braken-fern-induced bovine urinary
Vet Pathol 46:2, 2009
34 35
36
37
38
39
40 41
42
43
Immunohistochemistry of Vascular Bovine Urinary Bladder Tumors
bladder tumors. Clinical and morphological characteristics. Vet Pathol 13:110–122, 1976 Peixoto PV, Franc¸a TN, Barros CSL, Tokarnia CH: Histopathological aspects of Bovine Enzootic Hematuria in Brazil. Pesq Vet Bras 23:65–81, 2003 Pinto C, Peleteiro C, Carvalho T, Lobo M, Machado J: Slaughterhouse monitoring of cattle urinary bladder tumours in Sa˜o Miguel Island (Azores). Proc Int Symp Intractable Weeds Plant Invasers Ponta Delgada, Azores 17, 2006 Poblet E, Gonzalez-Palacios F, Jimenez FJ: Different immunoreactivity of endothelial markers in well and poorly differentiated areas of angiosarcomas. Virchows Arch 428:217–221, 1996 Ramos-Vara JA, Miller MA, Boucher M, Roudabush A, Johnson GC: Immunohistochemical Detection of Uroplakin III, Cytokeratin 7, and Cytokeratin 20 in Canine Urothelial Tumors. Vet Pathol 40:55–62, 2003 Roperto S, Ambrosio V, Borzacchiello G, Galati P, Paciello O, Russo V, Roperto F: Bovine papillomavirus type-2 (BPV-2) infection and expression of uroplakin IIIb, a novel urothelial cell marker, in urinary bladder tumors of cows. Vet Pathol 42:812–818, 2005 Sardon D, de la Fuente I, Calonge E, Perez-Alenza MD, Castano M, Dunner S, Pena L: H-ras immunohistochemical expression and molecular analysis of urinary bladder lesions in grazing adult cattle exposed to bracken fern. J Comp Pathol 132:195–201, 2005 Seethala RR, Gomez JA, Vakar-Lopez F: Primary angiosarcoma of the bladder. Arch Pathol Lab Med 130:1543–1547, 2006 Shahin M, Moore MR, Worrall S, Smith BL, Seawright AA, Prakash AS: H-ras activation is an early event in the ptaquiloside-induced carcinogenesis: comparison of acute and chronic toxicity in rats. Biochem Biophys Res Commun 250:491–497, 1998 Shahin M, Smith BL, Worral S, Moore MR, Seawright AA, Prakash AS: Bracken fern carcinogenesis: multiple intravenous doses of activated ptaquiloside induce DNA adducts, monocytosis, increased TNF alpha levels, and mammary gland carcinoma in rats. Biochem Biophys Res Commun 244:192–197, 1998 Sherr CJ: Cancer cell cycles. Science 274:1672–1677, 1996
221
44 Shin SJ, Lesser M, Rosen PP: Hemangiomas and angiosarcomas of the breast: diagnostic utility of cell cycle markers with emphasis on Ki-67. Arch Pathol Lab Med 131:538–544, 2007 45 Smith BL: The toxicity of bracken fern (genus Pteridium) to animals and its relevance to man. In: Handbook of Plant and Fungal Toxicants, ed. D’Mello JPF, pp. 63–76. CRC Press, Boca Raton, FL, 1997 46 Smith BL, Embling PP, Agnew MP, Lauren DR, Holland PT: Carcinogenicity of bracken fern (Pteridium esculentum) in New Zealand. N Z Vet J 36:56–58, 1988 47 Smith BL, Shaw G, Prakash AS, Seawright AA: Studies on DNA adduct formation by ptaquiloside, the carcinogen of bracken ferns (Pteridium spp.). In: Plant Associated Toxins, ed. Colgate SM and Dorling PR, Chap. 31, pp. 167–172. CAB International, Wallingford, UK, 1994 48 Smith R: Evaluation of cross species reactivity of antibodies to human antigens in animal models using immunoperoxidase techniques. J Histotechnol 13: 255–269, 1990 49 Thomazy VA, Luthra R, Uthman MO, Davies PJ, Medeiros LJ: Determination of cyclin D1 and CD20 mRNA levels by real-time quantitative RT-PCR from archival tissue sections of mantle cell lymphoma and other non-Hodgkin’s lymphomas. J Mol Diagn 4:201–208, 2002 50 Vardar E, Gunlusoy B, Minareci S, Postaci H, Ayder AR: Evaluation of p53 nuclear accumulation in low- and high-grade (WHO/ISUP classification) transitional papillary carcinomas of the bladder for tumor recurrence and progression. Urol Int 77:27–33, 2006 51 Vousden KH, Lu X: Live or let die: the cell’s response to p53. Nat Rev Cancer 2:594–604, 2002 52 Warren AL, Summers BA: Epithelioid Variant of Hemangioma and Hemangiosarcoma in the Dog, Horse, and Cow. Vet Pathol 44:15–24, 2007 53 Yonemaru K, Sakai H, Murakami M, Kodama A, Mori T, Yanai T, Maruo K, Masegi T: The significance of p53 and retinoblastoma pathways in canine hemangiosarcoma. J Vet Med Sci 69:271–278, 2007 54 Zietz C, Rossle M, Haas C, Sendelhofert A, Hirschmann A, Sturzl M, Lohrs U: MDM-2 oncoprotein overexpression, p53 gene mutation, and VEGF up-regulation in angiosarcomas. Am J Pathol 153:1425–1433, 1998
Request reprints from Taˆnia Carvalho, Faculdade de Medicina Veterina´ria, Avenida da Universidade Te´cnica, 1300-477 Lisboa (Portugal). E-mail:
[email protected].
