OUP user menu

Primary Adenocarcinoma of the Urinary Bladder
Value of Cell Cycle Biomarkers

Payal Kapur MD, Yair Lotan MD, Ellen King MD, Wareef Kabbani MD, Anirban P. Mitra MD, Ahmed Mosbah MD, Hassan Abol-Enein MD, Mohamed Ghoneim MD, Ramy F. Youssef MD
DOI: http://dx.doi.org/10.1309/AJCP76KUVOTBKQRY 822-830 First published online: 1 June 2011

Abstract

Primary adenocarcinomas of the urinary bladder are uncommon, and the molecular pathways are currently not well defined. In this study, we assessed the association between biologic markers and clinicopathologic characteristics in a cohort of 21 patients with primary urinary bladder adenocarcinoma. Immunohistochemical staining for cell cycle–specific markers, including p53, p21, p27, Ki-67, and cyclin E, were performed on sections of a tissue microarray construct. The tumors were high grade in 12 (57%) and pT2 or higher in 18 (86%); lymph nodes were involved in 6 cases (29%); and there was pathologic evidence of schistosomiasis in 14 (67%). The best prognostic combination of markers was combined alterations in p27 and Ki-67 and was associated with stage (P = .012), grade (P = .005), DNA ploidy (P = .005), and lymph node involvement (P = .04). Stage, lymph node involvement, combined alterations of p27 and Ki-67, and combined alterations of all 5 biomarkers were associated with increased probability of disease recurrence and cancer-specific mortality (P < .05).

Key Words:
  • Adenocarcinoma of the urinary bladder
  • Immunohistochemistry
  • Bladder cancer markers

Adenocarcinoma of the bladder is an uncommon malignant neoplasm and accounts for fewer than 2% of all malignant urinary bladder tumors.1 Few series have been reported in literature, but most contain small numbers of cases, with the largest series coming from Egypt.2,3 A study from the United States using Surveillance, Epidemiology, and End Results data identified only 32 patients (0.8%) with adenocarcinoma from a cohort of 4,045 patients with newly diagnosed bladder cancer during a 1-year period.4

Adenocarcinoma occurs more frequently in schistosomiasis-endemic regions and is the most common tumor arising in the bladder of patients with exstrophy.5 The hallmark of this tumor is neoplastic cells forming predominantly glandular structures resembling colonic adenocarcinoma (enteric type), and the tumor may produce intracellular mucin (signet cell type) or extracellular mucin (mucinous type). Other potential primary tumors need to be ruled out to establish a diagnosis of primary adenocarcinoma of the urinary bladder. Radical surgery is considered the best available treatment; however, 5-year survival rates are low (11%–61%).5 Conventional clinicopathologic parameters may not be reflective of clinical outcome in adenocarcinoma of the bladder.6

To the best of our knowledge, prognostic immunohistochemical studies of this neoplasm have not been reported in the English literature. An immunohistochemical panel of predictive markers could potentially identify patients who would benefit from early institution of multimodal therapy and spare the adverse effects of aggressive therapy in patients who do not need it. The combination of clinical and immunohistochemical markers may help improve patient stratification for clinical trials and treatment decisions.

The neoplastic changes in adenocarcinoma of the bladder are not well studied. While one may expect alterations in cell cycle regulators, this finding has not been established, unlike in other types of bladder cancer such as urothelial cell and squamous cell carcinomas. These genetic events probably involve activation of oncogenes, inactivation or loss of tumor suppressor genes, and alteration in the cell cycle gene products.718

Therefore, we evaluated the association of commonly implicated markers of cell cycle alteration in bladder cancer, including p53, p21, p27, cyclin E, and Ki-67, with pathologic features and clinical outcomes of pure adenocarcinoma of the bladder after radical cystectomy (RC) and long-term follow-up. The purpose of the study was to characterize molecular markers that might impact adenocarcinoma development and aggressiveness.

Materials and Methods

Case Selection

The records and pathologic specimens of patients treated by RC and pelvic lymphadenectomy for primary bladder carcinoma in Mansoura, Egypt, from 1997 to 2003 were reviewed. We identified 21 pure nonurachal adenocarcinoma cases with sufficient paraffin-embedded archival material of the cystectomy specimens that were available for extensive immunohistochemical evaluation. For each case, comprehensive clinicopathologic data were obtained and entered into an institutional review board–approved database.

Follow-up

All patients were followed up regularly for disease progression every 2 months in the first 6 months and at 6-month intervals thereafter. Follow-up evaluation routinely comprised a medical history and physical examination, laboratory tests (including measurement of creatinine level, alkaline phosphatase level, liver function parameters, and blood cell count), urinalysis, urinary exfoliative cytology, abdominal ultrasonography, and chest radiographs. Computed tomography, magnetic resonance imaging, and bone scans were done when findings suggested disease progression, defined as any emerging local or distant tumor.

Pathologic Evaluation

Histologic features, tumor grade, and stage were confirmed by blinded review of H&E-stained sections cut from available archival paraffin blocks for each case by experienced pathologists (W.K. and P.K.). To ensure exclusion of urothelial carcinoma with glandular differentiation, only cases with extensive sampling (1 section per centimeter of the maximum tumor dimension) and no evidence of in situ or invasive urothelial carcinoma were included in the study. Tumor grade was assigned according to the 1973 World Health Organization grading system (graded from 1, well-differentiated, to 3, poorly differentiated). Pathologic stage was reassigned according to the 2002 American Joint Committee on Cancer TNM staging system. In addition, DNA ploidy status using flow cytometry19 and lymphovascular invasion (LVI) were evaluated. All available sections were also reviewed to confirm histologic evidence of schistosomal infection and to select representative areas of the tumor from which to acquire cores for tissue microarray (TMA) analysis.

Construction of TMA Blocks

Validation studies of TMA technology have shown that analysis of 2 to 3 cores of 1-mm diameter provides sufficient information to achieve results similar to those from whole sections.20 We identified 3 representative areas for each tumor. The TMA was built using separate core needles for punching the donor and recipient blocks. For each case, three 1.0-mm core diameter representative samples were obtained from the “donor” block and placed on a separate “recipient” TMA block. Tumor samples were randomly arranged on the blocks. Serial 3- to 4-μm sections were obtained from the microarray, and the first slide was stained with H&E to confirm the presence of tumor and to assess the histologic features of the tumor. Sample tracking was based on coordinate positions (eg, column B5, where B refers to the second column and 5 refers to the fifth row) for each tissue spot in the TMA block; the spots were transferred onto TMA slides for staining. This sample tracking system was linked to a Microsoft Excel database (Microsoft, Redmond, WA) containing demographic, clinical, pathologic, and survival data for each patient. The array was read according to the given TMA map, and each core was scored individually. The cores that contained very little tumor were not evaluated. Each case had at least 2 core punches with sufficient tumor (at least 4–6 high-power fields of viable tumor) for immunohistochemical evaluation by image analysis.

Immunohistochemical Analysis and Scoring

We performed p53, p21, p27, cyclin E, and Ki-67 immunohistochemical staining using serial sections from the paraffin-embedded TMA block. Immunostaining was performed on a DAKO Autostainer (DAKO, Carpinteria, CA). Optimum primary antibody dilutions were predetermined, and appropriate positive control samples (tissues known to be positive for the immunohistochemical marker) and negative control samples (test tissue sections without the addition of primary antibody) were used for p21 (monoclonal mouse, SX118, dilution 1:200; DAKO), p53 (monoclonal mouse, DO-7, dilution 1:2,200; DAKO), p27 (monoclonal mouse, SX53G8, dilution 1:150; DAKO), Ki-67 (monoclonal mouse, MIB-1, dilution 1:300; DAKO), and cyclin E (monoclonal mouse, 870P110, dilution 1:400; DAKO). The immunohistochemically stained slides were analyzed using bright-field microscopy imaging coupled with advanced color detection software (Automated Cellular Imaging System, ChromaVision Medical Systems, San Juan Capistrano, CA). This standardized system has been shown to be reproducible and accurate.21

Percentages of positive cells in relation to the total number of cells encountered and the intensity of nuclear staining for each of the markers were quantified in at least 4 to 6 high-power (40× objective) hot spots (representative viable tumor areas with a high level of immunohistochemical staining) in each punch of every case with sufficient tumor. For each marker, nuclei showing more than faint brown staining were considered positive. The percentage of thusly defined positively staining cells was assessed. The mean of the triplicate scores was calculated for data analysis. An investigator (P.K.), blinded to the sample tracking system, confirmed all generated scores manually. Cases in which the manual impression was discrepant from the score generated by the image analyzer were reviewed by a second pathologist (W.K.), and a consensus score was obtained.

Based on published literature, all markers were placed in 1 of 2 categories: altered or normal. Accumulation of p53 protein in 10% or more of the tumor cell nuclei strongly correlates with mutation in the p53 gene.10,22 p53 was, therefore, considered altered when samples demonstrated 10% or more nuclear immunoreactivity. p21 immunoreactivity was considered altered when tumor cells showed very little or no nuclear reactivity (<5%).23 Nuclear p27 and cyclin E immunoreactivity were considered altered when they were less than 30% based on previously published clinically relevant levels.810 Ki-67 was considered altered when samples demonstrated a high proliferation count of more than 10%.24

Statistical Analysis

The Pearson χ2 test was performed to examine the relationships of p53, p21, p27, cyclin E, and Ki-67 expression with clinical and pathologic characteristics. Recurrence and survival probabilities after RC were estimated using the Kaplan-Meier method. Disease recurrence was defined as local failure in the cystectomy bed or regional lymph nodes (LNs) or distant metastasis after RC. The period of disease-free survival (DFS) was defined as the time between the date of RC and the development of local recurrence or distant metastasis. Censored survival values represent patients who were alive without clinical evidence of disease at the last follow-up. The cause of death was determined by chart review or treating physicians. The period of cancer-specific survival (CSS) was defined as the time between the date of RC and death due to cancer. Statistical differences were determined by using the log-rank test. All reported P values are 2-sided, and significance was set at .05. All statistical tests were performed with SPSS, version 17.0 (SPSS, Chicago, IL).

Results

Patient Demographics

The study included 21 cases of primary bladder adenocarcinoma diagnosed at the Urology and Nephrology Center, Mansoura, Egypt. There were 16 men (76%) and 5 women (24%), with a mean ± SD age of 49.4 ± 8.4 years (range, 35–72 years) at the time of RC. Overall, 14 patients (67%) had bilharzial infection, 18 (86%) had stage pT2 or higher tumors, 12 (57%) tumors were high grade, 6 patients (29%) had metastases to regional LNs, and 5 (24%) had LVI. The mean number of LNs removed was 21 (range, 4–68; median, 19; mean, 21). Gross and cystoscopic descriptions for all tumors included nodular or polypoid architecture in 19 cases (90%) and papillary architecture in the remaining cases. Two cases were mucinous adenocarcinomas with signet-ring cells floating within the mucin; the remaining cases were enteric-type nonurachal adenocarcinomas. The tumors involved a single site in 17 cases (81%) and were multicentric in the remaining 4 cases (19%). The commonly involved regions were the posterior wall and dome of the bladder (13 cases [62%]).

Association of Markers With Pathologic Characteristics

Clinicopathologic characteristics of the 21 patients and their association with molecular marker status are shown in Table 1. Owing to the small numbers, we were unable to analyze histologic subtypes with the various immunomarkers. Image 1 shows representative immunohistochemical staining results for p53, p21, p27, cyclin E, and Ki-67 in bladder adenocarcinoma. Each biomarker was altered in at least 70% of cases. Overall, Ki-67 was altered in 20 cases (95%), cyclin E in 18 (86%), p53 in 17 (81%), p21 in 16 (76%), and p27 in 15 (71%). There was no difference in the number of LNs removed or age between patients with normal vs altered expression of any of the markers (P > .05).

Altered p27 expression was associated with stage (P = .012), grade (P = .001), DNA ploidy (P = .018), and nonbilharzial bladder cancer (P = .04) and showed a trend with LN metastasis (P = .06). Altered p21 was associated with grade and DNA ploidy (P = .05). Alterations of p53, cyclin E, and Ki-67 separately were not associated with any of the pathologic parameters (P > .05). The best prognostic combination of markers was combined alterations in p27 and Ki-67 and was associated with stage (P = .012), grade (P = .005), DNA ploidy (P = .005), and LN metastasis (P = .04). Alterations of all 5 markers together were associated with stage (P = .05) and DNA ploidy (P = .04). In node-negative cases (n = 15), p27 and combined alterations of p27 and Ki-67 were associated with high stage, grade, and abnormal DNA ploidy (P < .05); p21 was associated with grade, and p53 was associated with abnormal DNA ploidy (P < .05).

View this table:
Table 1

Clinical Outcome and Its Association With Markers

Follow-up ranged from 0 to 88 months (median, 39 months; mean, 41 months). Recurrence occurred in 6 patients (29%) during a mean ± SD follow-up period of 4.8 ± 1.3 years, and 5 patients (24%) died of adenocarcinoma. The mean ± SD overall 3-year DFS and CSS rates for all patients included in the study were both 76% ± 1%, and the 5-year DFS and CSS rates were 52% ± 2% and 57% ± 2%, respectively Table 2 and Table 3.

Kaplan-Meier analyses showed that stage, DNA ploidy, LN metastasis, LVI, combined alterations of p27 and Ki-67, and combined alterations for all 5 biomarkers were associated with an increased probability of disease recurrence (P < .05) Figure 1. There were no recurrences in patients with T stage lower than T2 or in patients without combined alterations of p27 and Ki-67. Meanwhile, patients with LN metastasis had a dismal prognosis, and none of them exceeded 3 years of recurrence-free survival or CSS. The 5-year mean ± SD DFS dropped significantly to 26% ± 2% in patients with LVI and patients with combined alterations of p27 and Ki-67 and to 31% ± 2% in combined alterations of all 5 biomarkers. Stage, LN metastasis, combined alterations of p27 and Ki-67, and combined alterations for all 5 biomarkers were the only factors associated with an increased probability of cancer-specific mortality (P < .05) Figure 2. Age, sex, tumor grade, and separate alteration of markers were not associated with disease recurrence or cancer-specific mortality (P > .05).

Discussion

In the present study, we developed a TMA of primary urinary bladder adenocarcinoma and analyzed the expression of various cell cycle–related prognostic markers. We studied the correlation of these cell cycle regulators with clinicopathologic parameters and clinical outcomes after RC.

Mutations in cell cycle regulatory genes are one of the most common alterations studied in cancer, and these regulatory genes can lead to tumor growth and progression. In this study, we found that primary bladder adenocarcinoma has frequent alterations in cell cycle markers. Each biomarker was altered in more than 70% of cases. Ki-67 was the most frequently altered marker, and p27 was the least frequently altered marker in our study. Altered p27 expression was associated with nonbilharzial bladder adenocarcinoma (P = .04); however, there was no significant difference in the expression of the other biomarkers with schistosomiasis-associated adenocarcinoma.

Image 1

Representative photomicrographs of primary bladder adenocarcinomas on tissue microarray showing unaltered and altered immunoreactivity of cell cycle markers (×200).

View this table:
Table 2
View this table:
Table 3

One of the most intensively studied markers is the tumor suppressor p53 protein. It has a vital role in regulation of the cell cycle and is important for genetic stability, cell proliferation, apoptosis, and inhibition of angiogenesis.25 Wild-type p53 protein induces p21 protein, a product of the waf1/cip1 gene. p21 and p27 are cyclin-dependent kinase inhibitors that inhibit cells from entering the G1 to S phase.12 p53, p21, and p27 have been implicated in the oncogenesis, progression, and outcome of urothelial bladder cancer.7,8,1318 Many studies have shown that increased p53 expression predicts a poor outcome in urothelial carcinoma.26 Others have shown decreased expression of both p21 and p27 to be predictive of recurrence and disease-specific survival in urothelial carcinoma. We found that altered expression of p53 did not correlate with any of the evaluated clinicopathologic parameters in adenocarcinoma. However, p27 alterations were associated with high stage, grade, and DNA ploidy and showed a trend with LN metastasis. Altered p21 was associated with high grade and DNA ploidy. These markers have not been studied in primary bladder adenocarcinoma, and we recommend larger, multi-institutional analyses to confirm these findings.

Other commonly studied markers in cancers include cyclins and proliferative markers. Cyclin E alterations can enhance tumor progression through cellular growth and proliferation by increasing the rate of the G1 phase.9,12 The Ki-67 proliferative index is prognostic in many cancers, including bladder cancer.7 Tissue growth depends on cell proliferation, and apoptosis and proliferation rate have been investigated extensively for several tumors, including carcinomas of the urinary bladder.27,28 In urothelial bladder cancer, Ki-67 expression was associated with tumor stage, grade, LVI, LN metastases, disease recurrence, and cancer-specific mortality.11 In our study of adenocarcinoma, cyclin E and Ki-67 separately did not correlate with any pathologic parameters, likely because almost all tumors had alterations in these markers. This rationale may also apply to alterations in p53.

The results of this study suggest that no single marker by itself can be relied on to provide a complete prognostic picture. There is growing consensus that multiple markers used in an integrated panel would be required for translation to clinical practice.7,8,12,13,1618 In this study, the best prognostic combination of markers was combined alterations in p27 and Ki-67, which were associated with high stage, high grade, abnormal DNA ploidy, and LN metastasis. Alterations of all 5 markers together were associated with stage and DNA ploidy only.

In this study, there was recurrence in 29% of cases during the 5-year follow-up, and most of the patients died of cancer. The mean ± SD overall 3-year DFS and CSS rates for all patients were both 76% ± 1%, and the 5-year DFS and CSS rates were 52% ± 2% and 57% ± 2%, respectively. These survival rates, although within the range published in literature (11%–61%), are more favorable and similar to rates reported from regions of endemic schistosomiasis.5 The largest series that reported a 5-year survival of 55% is from a region of endemic schistosomiasis and included predominantly low-grade tumors.2 In contrast, studies that reported low survival included nonbilharzial adenocarcinoma with significant numbers of high-grade tumors. These differences in survival likely reflect different tumor biology of bilharzial adenocarcinoma.

Our study could not confirm the observation by Grignon et al29 that patients who died of nonurachal adenocarcinoma tended to do so within the first 2 years of diagnosis. Although the numbers of cases are small, we did not see significant difference in DFS or CSS between bilharzial and nonbilharzial cases. Survival analyses showed that stage and LN metastasis were associated with increased probability of disease recurrence and cancer-specific mortality. Abnormal DNA ploidy and LVI were also associated with increased probability of disease recurrence. Patients with LN metastasis had a dismal prognosis, and none of them exceeded 3 years of recurrence-free survival or CSS. Similarly, 5-year DFS dropped significantly in patients with LVI. el-Mekresh et al2 found that tumor pathologic stage, grade, and LN involvement had a significant impact on survival. Anderström et al30 found tumor differentiation to be associated with survival. In our series, the histologic subtype was not statistically associated with disease recurrence or cancer-specific mortality, a finding shared by others.29

Figure 1

Association of clinicopathologic and molecular markers with disease recurrence in 21 patients treated with radical cystectomy for primary bladder carcinoma.

Survival analyses showed that combined alterations of p27 and Ki-67 and combined alterations for all 5 biomarkers were associated with increased probability of disease recurrence and cancer-specific mortality. There were no recurrences in patients without combined alterations of p27 and Ki-67. The 5-year DFS dropped significantly in patients with combined alterations of p27 and Ki-67 and combined alterations of all 5 biomarkers. Alteration of individual markers was not associated with disease recurrence or cancer-specific mortality (P > .05).

Survival rates for adenocarcinoma of the bladder are poor despite adequate RC, the current “gold standard” for treatment.5 Survival after RC for adenocarcinoma of the bladder in Egyptian patients correlates with tumor pathologic stage and grade and LN involvement.2 However, integration of molecular markers associated with the biologic behavior of primary adenocarcinoma of the urinary bladder may help change the current treatment paradigm and guide incorporation of adjunct neoadjuvant and adjuvant therapies (chemotherapy and radiotherapy). We recommend p27 and Ki-67 be further studied in any panel of markers concerning adenocarcinoma prognostics or predictors of response to therapeutics. Larger studies are necessary to determine if these markers provide independent prediction of outcomes. Ongoing efforts should be directed toward including novel molecular prognostics, aiming at better identification of the behavior of this uncommon type of disease. This effort might add information about individual patient risk for prognostic and therapeutic purposes.

Figure 2

Association of clinicopathologic and molecular markers with cancer-specific mortality in 21 patients treated with radical cystectomy for primary bladder carcinoma.

This study has several limitations. First and foremost is the small number of cases. Unfortunately, primary adenocarcinomas are rare tumors, and no single institution has a large number of cases. Second are limitations inherent in any retrospective data collection and reliability of immunohistochemical techniques. The clinical application of immunohistochemical staining has been limited by discrepancies related to variability in the interpretation and stratification criteria and inconsistency in specimen handling and technical procedures. However, we used an automated immunostainer and a quantitative image analyzer to minimize technique-related variability. Moreover, an investigator who was blinded to the sample tracking system and clinical data confirmed all readings manually.

Cell cycle–related molecular markers are commonly altered and may correlate not only with pathologic parameters, but also with oncologic outcome after RC for bladder adenocarcinoma. Our results suggest a combination of markers may provide more information than single markers alone.

Acknowledgments

We thank the UT Southwestern Immunohistochemistry Laboratory for the excellent immunohistochemical staining.

Footnotes

  • Supported in part by a grant from the Egyptian Ministry of Higher Education via the Egyptian Cultural and Educational Bureau, Washington, DC; and by a grant from the Clinical Development Fund, Department of Pathology, UT Southwestern Medical Center, Dallas.

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
View Abstract