WT1, Estrogen Receptor, and Progesterone Receptor as Markers for Breast or Ovarian Primary Sites in Metastatic Adenocarcinoma to Body Fluids

Transcription

1 Anatomic Pathology / WT1, ESTROGEN RECEPTOR, AND PROGESTERONE RECEPTOR IN CYTOLOGY OF BODY FLUIDS WT1, Estrogen Receptor, and Progesterone Receptor as Markers for Breast or Ovarian Primary Sites in Metastatic Adenocarcinoma to Body Fluids Benjamin H. Lee, MD, PhD, Jonathan L. Hecht, MD, PhD, * Jack L. Pinkus, PhD, and Geraldine S. Pinkus, MD Key Words: WT1; Estrogen receptor; Progesterone receptor; Cell block; Cytology; Body fluids; Adenocarcinoma; Estrogen; Progesterone; Gross cystic disease fluid protein Abstract In tissue sections, detection of the Wilms tumor susceptibility gene 1 (WT1) protein, the hormonal receptors for estrogen (ER) and progesterone (PR), and gross cystic disease fluid protein (GCDFP) are useful for diagnosing ovarian and breast adenocarcinomas. We evaluated these markers for cytology cell-block preparations from 96 effusion specimens (metastases from 29 breast, 22 ovarian, and 45 adenocarcinomas from other sites). WT1 protein was reactive in 19 cases metastatic from ovary (86%), 2 from breast (7%), and none from other sites (specificity, 97%). Of the metastatic breast carcinomas, 21 (72%) were reactive for ER, 15 (52%) for PR, and 13 (45%) for both (combined specificity, 84%). GCDFP was reactive in only 4 breast cancer cases (14%). Ovarian tumors also were frequently positive for ER (19 [86%]), PR (11 [50%]), or both (10 [45%]). WT1 protein is an effective marker for ovarian adenocarcinoma, especially in ascites. The detection of ER and PR in metastatic adenocarcinoma from pleural or pericardial effusions can distinguish breast from lung primary sites. Reactivity for ER and PR did not distinguish between breast and ovarian metastases; however, studies for WT1 protein and GCDFP may aid in making this distinction. Malignant pleural effusion or ascites is not an uncommon clinical manifestation of adenocarcinoma. 1 Elucidating the origin of these malignant neoplasms can pose a considerable diagnostic challenge to both clinicians and pathologists and often may have therapeutic consequences for the patient. Immunohistochemical analysis has become standard practice in the evaluation of nodal and soft tissue metastasis, and a number of highly effective tissue-specific tumor markers are available. When used in specific panels, immunohistochemical studies can be extremely useful for the determination of tumor type, particularly in cases with an unknown primary site. These markers also can be used in the identification of primary tumors in patients with malignant effusions from various body cavities. In this immunohistochemical study we examined the effectiveness of the Wilms tumor susceptibility gene 1 (WT1) protein, estrogen receptor (ER), progesterone receptor (PR), and gross cystic disease fluid protein (GCDFP) as markers for ovarian and breast adenocarcinomas in cytologic material prepared from malignant effusion specimens derived from tumors involving a variety of primary sites. Materials and Methods A total of 96 cell blocks containing various malignant tumors were identified in the files of the Cytology Division of Brigham and Women s Hospital, Boston, MA. Cell blocks were prepared from pleural (56), peritoneal (27), or pericardial fluids (13) containing tumors of various types with approval from the institutional review board. All fluids were diagnosed as positive for metastatic adenocarcinoma (96 cases). Biopsy-documented primary sites of metastatic tumors included lung (33 American Society for Clinical Pathology Am J Clin Pathol 2002;117:

2 Lee et al / WT1, ESTROGEN RECEPTOR, AND PROGESTERONE RECEPTOR IN CYTOLOGY OF BODY FLUIDS Table 1 Immunoreactivity for ER, PR, WT1 Protein, and GCDFP in Malignant Effusions of Various Types of Adenocarcinoma No. (%) Immunoreactive for Various Markers Primary Adenocarcinoma Site No. of Cases ER PR ER and PR WT1 GCDFP Breast (72) 15 (52) 13 (45) 2 (7) 4 (14) Endometrium 2 1 (50) 2 (100) 1 (50) 0 (0) ND Esophagus 2 0 (0) 2 (100) 0 (0) 0 (0) ND Gallbladder 1 0 (0) 0 (0) 0 (0) 0 (0) ND Stomach 3 0 (0) 0 (0) 0 (0) 0 (0) ND Lung 33 0 (0) 0 (0) 0 (0) 0 (0) ND Ovary (86) 11 (50) 10 (45) 19 (86) ND Pancreas 2 0 (0) 0 (0) 0 (0) 0 (0) ND Prostate 2 0 (0) 0 (0) 0 (0) 0 (0) ND ER, estrogen receptor; GCDFP, gross cystic disease fluid protein; ND, not done; PR, progesterone receptor; WT1, Wilms tumor susceptibility gene 1 protein. adenocarcinomas), breast (29, including 2 cases of lobular carcinoma), endometrium (2), esophagus (2), gallbladder (1), ovary (22), pancreas (2), prostate (2), and stomach (3). The body cavity fluids were processed into cell blocks using thromboplastin and plasma. The cell blocks were fixed in 10% neutral buffered formalin and embedded in paraffin. Immunoperoxidase studies for WT1 protein, ER, PR, and GCDFP-15 were performed manually on 5-µm paraffin sections of the cell blocks following heat-induced epitope retrieval (0.001-mol/L concentration of EDTA, ph 8.0). Deparaffinized sections were placed in a container of preheated retrieval solution, steamed for 30 minutes (steamer model HS80, Black & Decker, Shelton, CT), then cooled for 20 minutes, washed under running water, placed in distilled water, then treated with methanolic peroxide (5 parts methanol/1 part 3% hydrogen peroxide) for 20 minutes. Slides then were washed under running water, placed in distilled water, then in a 0.05-mol/L concentration of tris(hydroxymethyl)aminomethane (Tris) buffer, ph 7.6, supplemented with 3% porcine serum. Slides then were incubated for 1 hour with monoclonal antibodies specific for WT1 protein (clone 6F-H2, DAKO, Carpinteria, CA) at a dilution of 1:100, ER (clone 6F-11, Novocastra Laboratories, Newcastle upon Tyne, England) at a dilution of 1:50, PR (clone 1A6, DAKO) at a dilution of 1:40, and GCDFP-15 (BRST-2, clone D6, Signet Laboratories, Dedham, MA) at a dilution of 1:200, then washed and incubated with horseradish peroxidase labeled polymer conjugated to goat antimouse immunoglobulin antibodies (Envision + detection system, K4006, DAKO). Antibody localization was effected using a peroxidase reaction with 3,3'-diaminobenzidine tetrahydrochloride (DAB+) as the chromogen (Envision + detection system). Staining intensity was enhanced with DAB enhancer (Zymed Laboratories, San Francisco, CA). Slides were counterstained with methyl green solution or hematoxylin, dehydrated, and coverslipped. For WT1 protein, ER, and PR studies, only nuclear staining was regarded as a positive result. Cells immunoreactive for GCDFP exhibited diffuse granular cytoplasmic staining. Intensity of staining was graded on a 0 to 3+ scale. A positive control slide for each marker evaluated was included in each run. Control slides substituting Tris buffer for primary antibody were run as negative controls. Results Immunoreactivity for WT1 protein, ER, and PR was determined for neoplastic cells in a total of 96 cases of metastatic adenocarcinoma in cell blocks prepared from pleural, peritoneal, or pericardial fluids. Immunophenotypic analysis for GCDFP also was examined in the subset of breast cancer cases (n = 29). For WT1 protein or ER, all cases interpreted as positive exhibited nuclear reactivity in the majority of neoplastic cells, while the proportion of tumor cells in cases staining positively for PR and GCDFP were more variable, although readily demonstrable. Overall results are summarized in Table 1. ER, PR, and GCDFP Immunoreactivity for ER was detected in 41 (43%) of 96 cases, including 21 (72%) of 29 malignant effusion specimens metastatic from breast adenocarcinoma cases Image 1, 1 (50%) of 2 endometrial carcinoma cases, and 19 (86%) of 22 ovarian cancer cases Image 2. Strong nuclear reactivity was noted for most breast cancer cases Table 2. The remainder of the cases evaluated, including cell-block specimens from 33 lung adenocarcinomas, had no discernible nuclear staining for ER. Cytoplasmic staining for ER was noted in a small number of cases (7 total: 5 lung, 1 ovary, 1 pancreas) Image 3. The latter cases did not display nuclear staining for ER and were regarded as negative. Detection of PR was observed in 30 (31%) of 96 cases, including 15 (52%) of 29 breast (Image 1), 11 (50%) of Am J Clin Pathol 2002;117: American Society for Clinical Pathology

3 Anatomic Pathology / ORIGINAL ARTICLE A B C Image 1 Cell block sections from pleural fluids containing breast carcinoma. Nuclear immunoreactivity for estrogen receptor (ER; A) and progesterone receptor (PR; B) and cytoplasmic reactivity for gross cystic disease fluid protein (GCDFP; C) are demonstrated (immunoperoxidase study, methyl green counterstain for all, A, ER, 400; B, PR, 400; C, GCDFP, 400). ovarian (Image 2), 2 (100%) of 2 endometrial, and 2 (100%) of 2 esophageal carcinomas. The remaining cytologic specimens displayed no reactivity for PR, including the 33 effusion specimens of pulmonary origin. Cytoplasmic reactivity for PR was detected in a number of cases (12 total) including 3 breast, 1 endometrial, 1 esophageal, and 7 ovarian cases. Of these, only the single endometrial and 3 of the 7 ovarian cases were considered positive, since nuclear reactivity also was detected in each of these instances. Cases that exhibited dual positivity for both ER and PR included 13 (45%) of 29 breast, 1 (50%) of 2 endometrial, and 10 (45%) of 22 ovarian tumors. We also examined GCDFP reactivity in our subset of cell-block specimens derived from metastatic breast adenocarcinomas and detected positivity for most neoplastic cells in 4 (14%) of 29 cases, all of which exhibited a diffuse granular cytoplasmic staining pattern (Image 1). Interestingly, 2 of the 4 GCDFP-positive breast cancer cases had no staining for either ER or PR, while the remaining 2 GCDFP-positive breast cancer cases displayed positive staining for both ER and PR. In an additional 3 cases, rare neoplastic cells exhibited GCDFP positivity. WT1 Of 96 cases of metastatic adenocarcinoma, neoplastic cells in 21 cases (22%) exhibited nuclear reactivity for WT1 protein, and the majority of reactive cases (19/21 [90%]) represented ovarian metastases (Table 1). Of the 22 ovarian cases evaluated, 19 (86%) were positive for WT1 protein. Primary ovarian tumors in these cases included 17 papillary serous cystadenocarcinomas (16/17 positive), 2 mixed serous and endometrioid (both cases positive), 1 endometrioid (positive), and 2 clear cell carcinomas (both negative). Mucinous ovarian tumors metastatic to body fluids were not included in the study since no cases could be found. Two of American Society for Clinical Pathology Am J Clin Pathol 2002;117:

4 Lee et al / WT1, ESTROGEN RECEPTOR, AND PROGESTERONE RECEPTOR IN CYTOLOGY OF BODY FLUIDS A B C Image 2 Cell block sections from a peritoneal fluid containing ovarian carcinoma. Strong nuclear immunoreactivity for estrogen receptor (ER; A), progesterone receptor (PR; B), and Wilms tumor susceptibility gene 1 protein (WT1; C) is seen (immunoperoxidase study, methyl green counterstain for all, A, ER, 400; B, PR, 400; C, WT1 protein, 400). 29 cell blocks containing metastatic breast carcinomas were positive for WT1 protein. Nuclear staining intensity was strong in the majority of positive cases Table 3, and the majority of tumor cells were reactive. The remaining cases Table 2 Adenocarcinoma of Breast: Staining Intensity for Various Markers in Cell Blocks of 29 Malignant Effusions Staining Intensity * Marker Estrogen receptor Progesterone receptor Gross cystic disease fluid protein * Staining intensity was scored on a 0 to 3+ scale, with 3+ representing the strongest staining. Cytoplasmic staining only in 3 cases. Cytoplasmic reactivity. with metastatic tumor were regarded as negative, including 33 cases of pulmonary derivation. Cytoplasmic staining only was present in a subset of adenocarcinomas (14 cases: 9 lung, 2 breast, 1 endometrium, 1 esophagus, 1 gallbladder) and was regarded as a negative result. Nuclear staining was not seen in any of the latter cases. Tris control slides of all WT1 protein positive cases with nuclear reactivity in neoplastic cells, as well as cases exhibiting cytoplasmic staining, were negative. Cases with cytoplasmic staining only may reflect reactivity of a non WT1 protein crossreacting epitope. In some cases, staining of numerous mesothelial cells created difficulty identifying tumor cells, particularly if small numbers of tumor cells were present. In these cases, detection of tumor cells was facilitated by using a hematoxylin counterstain (rather than methyl green) or by performing an additional immunostain for calretinin to verify the identification of the mesothelial cells. 748 Am J Clin Pathol 2002;117: American Society for Clinical Pathology

5 Anatomic Pathology / ORIGINAL ARTICLE Discussion Breast cancer is the most common cause of malignant pleural effusions in women, followed closely by lung adenocarcinoma. Approximately half of the patients with disseminated breast cancer develop a malignant pleural effusion at some time during their illness, while the most common tumor to initially manifest as a malignant pleural effusion is lung carcinoma. 1 As with nodal and soft tissue metastases, the determination of the primary tumor site in cases of malignant effusion can be facilitated by immunoreactivity for various markers. In tissue sections, the identification of metastatic breast or ovarian carcinoma has been greatly enhanced by the use of antibodies directed against ER and PR, which have been demonstrated to be useful markers for breast carcinoma. 2 Our results indicate that they also are useful as part of a panel in the distinction between breast and lung adenocarcinoma in pleural- or pericardial-derived effusion specimens, a frequent clinical differential diagnosis. Although positive staining for ER and PR in our study demonstrated individual sensitivities in discriminating breast from other tumors of 72% and 52%, respectively, dual positive staining for both ER and PR increased the specificity of identifying a breast primary to 84%. Of importance, none of the lung carcinoma based effusion specimens demonstrated positivity for ER and PR. These results suggest that both ER and PR, when added to an immunohistochemical panel that includes thyroid transcription factor-1, a sensitive marker for tumors of pulmonary origin in tissue sections 3 and in cytologic preparations, 4 can contribute to the distinction between a tumor of a breast or lung primary site. It also should be noted that in our study, tumors of ovarian derivation, which infrequently enter the differential diagnosis of tumor type in pleural effusions, also are frequently positive for ER (86%) or PR (50%) or reactive for both ER and PR (45%). In cases in which this distinction may be relevant, additional studies for WT1 protein (reactive in most ovarian adenocarcinomas) may be helpful. Previous studies have demonstrated GCDFP to be a relatively sensitive (74%) and highly specific (up to 95%) marker for breast carcinoma in tissue sections. 5,6 Analyses of this marker in cytologic specimens have been lacking to date. Our study demonstrated GCDFP reactivity in 4 of 29 breast cases, a sensitivity of only 14%. In 3 additional cases, rare GCDFPpositive cells were identified. These data indicate that while detection of GCDFP can aid in confirming a tumor of breast origin, the lack of positive staining for GCDFP is of limited value, particularly when used in isolation. Similar to pleural- or pericardial-based effusions, which may be related to occult malignant neoplasms, malignant ascites may be due to undiagnosed gastrointestinal or gynecologic primary tumors. The use of cytokeratins 7 and 20, as well as carcinoembryonic antigen, for defining the sites of Image 3 Cell block section from a pleural fluid containing lung carcinoma. Cytoplasmic reactivity for estrogen receptor, as noted in this case, was considered a negative result (immunoperoxidase study, methyl green counterstain, 400). Table 3 Adenocarcinoma of the Ovary: Staining Intensity for Various Markers in Cell Blocks of 22 Malignant Effusions Nuclear Staining Intensity * Cyto- Marker plasmic Estrogen receptor Progesterone receptor WT1 protein WT1, Wilms tumor susceptibility gene 1. * Staining intensity was scored on a 0 to 3+ scale, with 3+ representing the strongest staining. No nuclear reactivity. This includes 3 of the cases with nuclear reactivity (see text). origin in these cases has been well documented WT1, a tumor suppressor gene initially identified as contributing to the development of Wilms tumor, 14 is expressed preferentially in the urogenital system and mesoderm-derived tissues. 15 Relatively specific reactivity for WT1 protein has been shown in ovarian papillary serous carcinoma and in tissue metastases. 16,17 Our study demonstrates that nuclear reactivity for WT1 protein is highly sensitive (86%) and specific (97%) for ovarian carcinoma in malignant ascites (particularly papillary serous cystadenocarcinomas). If one considers only strong staining (2+-3+), the sensitivity and specificity of WT1 protein for ovarian carcinoma become 82% and 99%, respectively, making it a highly specific marker for ovarian carcinoma. As expected, ER and PR were detected in 86% and 50% of ovarian tumors respectively, with 45% of the ovarian American Society for Clinical Pathology Am J Clin Pathol 2002;117:

6 Lee et al / WT1, ESTROGEN RECEPTOR, AND PROGESTERONE RECEPTOR IN CYTOLOGY OF BODY FLUIDS tumors demonstrating dual staining for PR and ER. However, whereas the majority of ovarian tumors stained positively for WT1 protein, only a small percentage (7%) of breast cancer cases were reactive for WT1 protein. Although breast carcinoma is rarely metastatic to ascites fluid, WT1 protein may have a helpful role in the distinction between ovarian and breast carcinomas in problematic cases. Immunoreactivity for WT1 protein is also a feature of benign and malignant mesothelial cells in tissue sections and cytologic material, 16,18-20 potentially creating a problem in interpretation in some cases. In cases containing few metastatic carcinoma cells and abundant mesothelial cells, accurate assessment may be facilitated by the use of other immunostains such as calretinin, a marker for mesothelial cells. For cases with WT1 protein positive and ER- and PR-negative tumor cells in which mesothelioma is a diagnostic consideration, the use of a larger panel of markers (eg, calretinin, cytokeratin 5/6, HBME-1, thrombomodulin, carcinoembryonic antigen, and others) may be helpful for definitive evaluation. 19,20 Conclusions We studied the usefulness of WT1 protein, ER, PR, and GCDFP as diagnostic discriminators for tumors in cytologic material from a broad range of primary sites. For metastatic adenocarcinoma, strong nuclear reactivity for WT1 protein was observed nearly exclusively (90%) for neoplasms of ovarian origin, providing a helpful marker for identification of this tumor type. Ovarian carcinomas also were frequently positive for ER, PR, or both. In pleural- or pericardial-based effusions, the addition of both ER and PR to a panel of other immunohistochemical markers, including thyroid transcription factor-1 as a marker for pulmonary adenocarcinoma, may be particularly helpful in the distinction between breast and pulmonary adenocarcinomas. Reactivity for GCDFP, while very specific for tumors of breast origin, seems to have a limited sensitivity and should be used in conjunction with a panel to detect other tissue markers. From the Department of Pathology, Brigham and Women s Hospital, Boston, MA. Address reprint requests to Dr G.S. Pinkus: Dept of Pathology, Brigham and Women s Hospital, 75 Francis St, Boston, MA * Dr Hecht is now with the Department of Pathology, Beth- Israel Deaconess Medical Center, Boston. References 1. Monte SA, Ehya H, Lang WR. Positive effusion cytology as the initial presentation of malignancy. Acta Cytol. 1987;31: Kaufmann O, Kother S, Dietel M. Use of antibodies against estrogen and progesterone receptors to identify metastatic breast and ovarian carcinomas by conventional immunohistochemical and tyramide signal amplification methods. Mod Pathol. 1998;11: Ordonez NG. Thyroid transcription factor-1 is a marker of lung and thyroid carcinomas. Adv Anat Pathol. 2000;7: Hecht JL, Pinkus JL, Weinstein LJ, et al. The value of thyroid transcription factor-1 in cytologic preparations as a marker for metastatic adenocarcinoma of lung origin. Am J Clin Pathol. 2001;116: Wick MR, Lillemoe TJ, Copland GT, et al. Gross cystic disease fluid protein-15 as a marker for breast cancer: immunohistochemical analysis of 690 human neoplasms and comparison with alpha-lactalbumin. Hum Pathol. 1989;20: Raab SS, Berg LC, Swanson PE, et al. Adenocarcinoma in the lung in patients with breast cancer: a prospective analysis of the discriminatory value of immunohistology. Am J Clin Pathol. 1993;100: Chu P, Wu E, Weiss LM. Cytokeratin 7 and cytokeratin 20 expression in epithelial neoplasms: a survey of 435 cases. Mod Pathol. 2000;13: Ramaekers F, van Niekerk C, Poels L, et al. Use of monoclonal antibodies to keratin 7 in the differential diagnosis of adenocarcinomas. Am J Pathol. 1990;136: Moll R, Lowe A, Laufer J, et al. Cytokeratin 20 in human carcinomas: a new histodiagnostic marker detected by monoclonal antibodies. Am J Pathol. 1992;140: van Niekerk CC, Jap PH, Ramaekers FC, et al. Immunohistochemical demonstration of keratin 7 in routinely fixed paraffin-embedded human tissues. J Pathol. 1991;165: Brown RW, Campagna LB, Dunn JK, et al. Immunohistochemical identification of tumor markers in metastatic adenocarcinoma: a diagnostic adjunct in the determination of primary site. Am J Clin Pathol. 1997;107: DeYoung BR, Wick MR. Immunohistologic evaluation of metastatic carcinomas of unknown origin: an algorithmic approach. Semin Diagn Pathol. 2000;17: Wang N, Zee S, Zarbo R, et al. Coordinate expression of cytokeratins 7 and 20 defines unique subsets of carcinoma. Appl Immunohistochem. 1995;3: Call KM, Glaser T, Ito CY, et al. Isolation and characterization of a zinc finger polypeptide gene at the human chromosome 11 Wilms tumor locus. Cell. 1990;60: Hastie ND. Wilms tumour gene and function. Curr Opin Genet Dev. 1993;3: Kumar-Singh S, Segers K, Rodeck U, et al. WT1 mutation in malignant mesothelioma and WT1 immunoreactivity in relation to p53 and growth factor receptor expression, celltype transition, and prognosis. J Pathol. 1997;181: Shimizu M, Toki T, Takagi Y, et al. Immunohistochemical detection of the Wilms tumor gene (WT1) in epithelial ovarian tumors. Int J Gynecol Pathol. 2000;19: Hecht JL, Lee BH, Pinkus JL, et al. The value of WT1 in cytologic preparations as a marker for malignant mesothelioma. Cancer. In press. 19. Ordonez NG. The immunohistochemical diagnosis of epithelial mesothelioma. Hum Pathol. 1999;30: Ordonez NG. Value of thyroid transcription factor-1, E-cadherin, BG8, WT1, and CD44S immunostaining in distinguishing epithelial pleural mesothelioma from pulmonary and nonpulmonary adenocarcinoma. Am J Surg Pathol. 2000;24: Am J Clin Pathol 2002;117: American Society for Clinical Pathology