Abstract
Background: Urothelial carcinoma (UC) variants can be difficult to differentiate from carcinoma metastatic to the bladder. Materials and Methods: We examined immunostaining for uroplakin III in 43 cases of primary bladder UC variants including micropapillary UC (n=19), nested variant of UC (n=2), pleomorphic giant-cell carcinoma (n=8), plasmacytoid UC (n=4), lymphoepithelioma-like carcinoma (n=2), large cell undifferentiated carcinoma (n=2), UC with abundant myxoid stroma (n=3) and lipid cell variant (n=3) and in 11 tumors from other organs metastatic to the bladder. These tumors included invasive ductal carcinoma of the breast (n=2), colorectal adenocarcinoma (n=4), endometrioid adenocarcinoma (n=1) and serous papillary carcinoma of the uterus (n=1) melanoma (n=1), embryonal carcinoma of the testis (n=1), and renal clear cell carcinoma (n=1). Results: Out of the 43 UC variants, 35 (81%) were positive for uroplakin III, including micropapillary, lipid cell variant and UC with abundant myxoid stroma. Pleomorphic giant cell carcinoma, plasmacytoid UC and nested variant of UC were less commonly positive. Of the 11 metastatic tumors, six were found to be positive for uropIakin III: metastatic colorectal adenocarcinoma, clear cell carcinoma of the kidney and embryonal carcinoma of testis. Conclusion: UP III Positivity for uroplakin III is not found only in primary bladder UC variants, but in some tumors that have metastatized to the bladder. Staining for uroplakin III alone should not be taken as evidence of UC.
Cancer of the urinary bladder is the ninth most common type of cancer worldwide and the 13th most common cause of cancer-related death (1).The most frequently encountered type of bladder cancer in developed countries is urothelial carcinoma (UC), accounting for more than 90% of bladder malignancies (2). UC has the propensity to undergo divergent differentiation, resulting in a number of morphological subtypes or variants, and this has been shown to occur in approximately 20% of cases (3, 4). The morphological patterns observed in these subtypes overlaps with those of many types of primary tumors of other organs that can metastasize to the bladder. In view of this and unless there is an associated component of typical UC, accurate diagnosis can be difficult. Many of these cases of variant UC require clinicopathological correlation, including imaging, in order to exclude the presence of a primary tumor originating from outside the bladder.
A wide variety of antigens have been investigated as immunohistochemical markers of UC (5). Out of these, much emphasis has been placed on the uroplakins. Within this group, five specific proteins have been identified. These range in size from 27 kDa to 47 kDa, and are designated uroplakin Ia, Ib, II, IIIa and IIIb. These are transmembrane proteins associated with urothelial cells and are expressed in the plasmalemma of the luminal surface of urothelial superficial or umbrella cells. They appear to have a key role in the structure of plaques found on the surface of umbrella cells that are integral to urothelial impermeability (6, 8).
Uroplakin III, the first uroplakin with a commercially available antibody, has been found to be a highly specific and moderately-sensitive immunohistochemical marker for primary UC. Although the diagnostic value of uroplakin III has been investigated for typical UC in a variety of studies, there are few data regarding the utility of the antibody with respect to uroplakin expression in UC variants (5, 9). In previous studies, the staining of uroplakin III in small numbers of cases of UC variants was investigated, however the results were too limited to be conclusive (5, 9, 10) .
In addition to investigating the value of uroplakin staining as a diagnostic tool for UC localized to the bladder, a number of studies have evaluated uroplakin expression in metastatic deposits of UC. In these studies, uroplakin staining was shown to discriminate between extravesical metastatic deposits of typical UC and secondary tumors of non-urothelial origin, metastatic to the lung, bone and lymph nodes (10). In view of the paucity of data with respect to uroplakin expression in the various morphotypes of UC, we investigated uroplakin III expression in a wide variety of UC variants. In this investigation, we also determined expression in a number of non-urothelial tumors metastatic to the bladder. We undertook this component of the study to determine if uroplakin may inform patient management, as many of the morphotypes of UC may resemble bladder metastases from primary extravesical malignancies.
Materials and Methods
We searched the surgical pathology files of Aquesta Pathology, Brisbane, Queensland, Australia for variants of invasive high-grade UC of the bladder, as well as tumors metastatic to the bladder, submitted between 2009 and 2011. Tumor sections were reviewed to confirm the diagnosis and as well as pT staging category, and to select a representative tissue block for immunohistochemical investigation. From this series, material was selected from 43 cases of primary bladder UC variants, from 33 patients. There were 25 males and 8 females, with a mean age 69 years and an age range of 44 to 89 years. The variants included micropapillary UC (n=19), nested variant of UC (n=2), pleomorphic giant cell carcinoma (n=8), plasmacytoid UC (n=4), lymphoepithelioma-like carcinoma (n=2), large cell undifferentiated carcinoma (n=2), UC with abundant myxoid stroma (n=3) and lipid cell variant of UC (n=3). The UC variants were pT2 in 19 cases and pT1 in 14 cases according to the Union Internationale Contre le Cancer TNM staging classification (seventh edition, 2010) (11).
We also found 11 tumors originating in other organs that were metastatic to the bladder. These patients were seven males and four females, with a mean age of 69 years and an age range of 44 to 99 years. The primary tumors were infiltrating ductal carcinoma of the breast (n=2), colorectal adenocarcinoma (n=4), endometrioid adenocarcinoma of the uterus (n=1), serous papillary carcinoma of the uterus (n=1) cutaneous melanoma (n=1), embryonal carcinoma of the testis (n=1), and clear cell renal cell carcinoma of the kidney (n=1).
All immunohistochemical staining was performed on formalin-fixed paraffin-embedded tissue sections cut at 3 μm. A MACH2 micro-polymer detection method was used according to the manufacturer's instructions (Biocare Medical, Concorde, California, USA), with appropriate positive and negative controls. Sections were dried at 60°C for a minimum period of 15 min prior to de-paraffinization and rehydration. Slides were then treated with a solution of 3% hydrogen peroxide for a period of 3 min to block endogenous peroxide activity. These were then heated in a conventional microwave oven at 550 W for a period of 30 min at 30% power, achieving boiling, in citrate buffer (pH 6.0) that was pre-warmed for a period of 5 min. Immediately upon removing from the microwave, slides were left to sit overnight at room temperature in a solution of primary mouse antibody to human uroplakin III (Clone AU-1; Supplier, Cell Marque, Rocklin, California, USA) diluted 1:25. This was for a minimum period of 15 h. Following incubation in primary antibody, slides were treated with a MACH2 (Biocare Medical, Concorde, California, USA) secondary antibody for 45 minutes. Detection was a one-step detection system using DAB chromogen (Biocare). Slides were then weakly counterstained in Lillie-Mayer's hematoxylin.
In this study, we assessed all tumors for all cytoplasmic, membranous and nuclear expression of uroplakin III. Cytoplasmic and membranous staining was assessed simultaneously and reported as 0 (negative staining), 1+ (focal and weakly positive), 2+ (moderately positive) or 3+ (strong and diffusely positive). Nuclear staining was reported as positive or negative.
Results
Out of the 43 UC variant tumors, 35 (81.4%) exhibited positive membranous and cytoplasmic staining of uroplakin III, with 26 (60.5%) tumors having membranous staining alone. The highest expression rates were found in micropapillary UC (18 out of 19 cases, 95%; Figure 1A), lipid cell variant (three out of three; Figure 1B), and UC with abundant myxoid stroma (three out of three; Figure 1C). Pleomorphic giant cell carcinoma (Figure 1D) and plasmacytoid UC and nested variant of UC were less frequently positive. The cases of large cell undifferentiated carcinoma and lymphoepithelioma-like carcinoma were entirely negative (Tables I and II).
Out of the 11 metastatic tumors examined, six (54.4%) exhibited positive membranous and cytoplasmic staining, with for (36.4%) having membranous staining alone. All cases of metastatic colorectal adenocarcinoma (n=4; Figure 2A), clear cell carcinoma of the kidney (n=1) and embryonal carcinoma of testis (n=1) (Figure 2B) were positive. Metastatic infiltrating ductal carcinoma of the breast, endometrioid and serous papillary carcinoma of the uterus, and melanoma were entirely negative. The staining intensity and the proportion of positively stained cells were variable in positive cases and subjectively ranged from moderate (2+) to strong (3+) staining.
In addition to membranous and cytoplasmic staining, positive nuclear staining was seen in 21 out of 43 (49%) UC variants and three out of 11 (27%) metastatic carcinomas (27.3%). All the cases with nuclear positivity also demonstrated cytoplasmic and membranous positivity for uroplakin III.
Discussion
UC of the bladder is known to have a great propensity for divergent differentiation, resulting in a number of subtypes or variants (3, 4). These variants show a significant morphological overlap with carcinoma arising in other organs, such that differentiation of variant UC from carcinoma metastatic to the bladder can be difficult. This differentiation is of paramount importance as the treatment and prognosis of primary bladder UC variants are entirely different from that of metastatic carcinoma. While primary bladder cancer may be amenable to curative resection, a diagnosis of malignancy metastatic to the bladder removes this option (12, 13). In some cases, the presence of an in situ UC component can help to resolve this situation and identify the primary nature of the tumor. If this is not seen, then clinicopathological correlation and immunohistochemical staining become necessary.
Representative sections showing uroplakin III immunostaining. A: Micropapillary urothelial carcinoma with 3+ staining intensity; B: lipid cell variant with 2+ staining intensity; C: urothelial carcinoma with abundant myxoid stroma with 2+ staining intensity; D: pleomorphic giant cell carcinoma with 3+ staining intensity (magnification ×400).
Metastatic carcinoma showing uroplakin III immunostaining. A: Colorectal adenocarcinoma with 3+ intensity staining; B: embryonal carcinoma of testis with 2+ staining intensity (magnification ×400).
Expression of uroplakin III (UPIII) in variants of urothelial carcinoma (UC).
Expression of uroplakin III (UPIII) in carcinomas metastatic to the bladder.
Previous investigations of typical UC have shown varying degrees of immunostaining for uroplakin III. In an earlier study, uroplakin III was detected in 53% of primary and 66% of metastatic UC using a polyclonal antibody (14). In that report, a number of non-urothelial carcinomas were all shown to be uroplakin III-negative. Using tissue microarrays, Parker et al. found uroplakin III in 57% of urothelial neoplasms, including 25 metastatic urothelial carcinomas. Of the metastatic UC in this study, 52% were positive for uroplakin III and this compared with an absence of staining for the metastatic non-urothelial tumors investigated (10). There are only limited studies, with small numbers of cases, that have assessed the diagnostic utility of uroplakin III immunostaining in variants of UC. Paner et al. found UP III immunostaining in 20% of bladder UC and in 33% of the UC variants that they examined. In their series none of the undifferentiated bladder cancers, such as lymphoepithelioma-like carcinoma, sarcomatoid carcinoma and carcinoma with rhabdoid or giant cells, showed positive uroplakin III expression (5). In an additional study, it was found that uroplakin III was positively expressed in 21.8% of cases of UC and that this was not seen in any of the tumors that exhibited squamous differentiation (9). In separate studies, two out of eight cases of large cell undifferentiated carcinoma of the bladder, 10 out of 13 cases of micropapillary UC and two out of eight cases of pleomorphic giant cell carcinoma exhibited positive expression for uroplakin III (15-17). In the cases of pleomorphic giant-cell carcinoma that were positive for uroplakin III, staining was seen in only 5-10% of cells (17). The only other study to have investigated immunostaining of variant UC by uroplakin III showed positivity in all three studied cases of UC with glandular differentiation; two out of three cases with micropapillary UC and in the one case of UC with plasmacytoid features included (10). These studies demonstrate the utility of uroplakin III immunostaining in the diagnosis of UC and, to a more limited degree, for UC variants, and show that in a metastatic location, it is useful for the determination of a urothelial origin.
In the present study we have demonstrated that some variants of UC often exhibit positive expression of uroplakin III, with staining having a membranous/cytoplasmic distribution. This staining, however, was not universal amongst all morphotypes of UC. Specifically, micropapillary UC, lipid cell variant, and UC with abundant myxoid stroma frequently exhibited positive staining. Pleomorphic giant cell carcinoma, plasmacytoid UC and nested variant of UC were less commonly positive, while large cell undifferentiated carcinoma and lymphoepithelioma-like carcinoma were negative. It is clear from this that those variants of UC that exhibit a lesser degree of de-differentiation and more closely resemble typical UC were more likely to express uroplakin III. This would suggest that as UC undergoes morphological evolution to a more poorly differentiated morphotype, there is a loss of the umbrella cell phenoptype and this limits the value of staining for uroplakin III in a diagnostic setting.
Our study also demonstrated immunostaining for uroplakin III in some carcinomas metastatic to the bladder which are non-urothelial. While two cases of ductal adenocarcinoma of the breast, one case of endometrioid adenocarcinoma of the uterus, one case of serous papillary carcinoma of the uterus and one case of melanoma were negative for uroplakin III, all four cases of colorectal adenocarcinoma, one case of embryonal carcinoma of the testis and one case of clear cell carcinoma of the kidney exhibited positive staining. This is of particular significance as colorectal cancer may involve the bladder by either direct or metastatic spread. In some clinical settings, the true origin of this tumor may not be apparent and it would appear that immunostaining for uroplakin III could have a confounding influence. Embryonal carcinoma of the testis metastatic to the bladder would also appear to be associated with false-positive expression of uroplakin III. Again, this has the potential to provide conflicting diagnostic information as embryonal carcinoma can resemble UC with glandular differentiation. From our results, it is apparent that uroplakin III is not entirely specific for UC, although the reason for this is unclear. Release of antigen by injured native tissue with diffusion and uptake by metastatic tumor cells may explain this finding, as has been previously reported for metastatic renal cell carcinoma in the thyroid displaying thyroglobulin positivity (18). This would appear to reduce the value of uroplakin III as a specific urothelial marker in bladder cancer variants when there is a need to differentiate these from tumors metastatic to the bladder. In view of this, it is important to be aware that a diagnosis of UC should not be made solely on uroplakin III staining in situations when this differential diagnosis is considered.
Footnotes
-
Conflicts of Interest
The Authors state that there are no conflicts of interest to disclose.
- Received July 9, 2014.
- Revision received August 2, 2014.
- Accepted August 5, 2014.
- Copyright© 2014 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.