CANCER 55 CYTOPATHOLOGY Cytologic Differential Diagnosis among Reactive Mesothelial Cells, Malignant Mesothelioma, and Adenocarcinoma Utility of Combined E-Cadherin and Calretinin Immunostaining Hiromi Kitazume, C.T.1 Kazuhisa Kitamura, C.T.1 Katsuhiko Mukai, M.D.1 Yoshiaki Inayama, M.D.1 Naomi Kawano, M.D.1 Nobuo Nakamura, M.D.2 Jinyu Sano, M.D.2 Kunihiro Mitsui, C.T.2 Sachiko Yoshida, M.D.3 Yukio Nakatani, M.D.1 1 Division of Anatomic and Surgical Pathology, Hospital of Yokohama City University, Yokohama City University School of Medicine, Yokohama, Japan. 2 Department of Pathology, Yokohama Citizens Hospital, Yokohama, Japan. 3 Department of Pathology, Yokohama Kowan Hospital, Yokohama, Japan. BACKGROUND. The differential diagnosis between reactive mesothelial cells (RMs), malignant mesotheliomas (MMs), and adenocarcinomas (ACs) is often difficult in cytologic specimens, and the utility of various immunohistochemical markers have been explored. Because recent immunohistologic studies have suggested that E-cadherin (E-cad) and calretinin (Cal) may be useful markers for epithelial and mesothelial differentiations, respectively, the authors investigated their utility in cytologic diagnosis. METHODS. In this retrospective study, immunostaining was performed on smears retrieved from Papanicolaou-stained slides of effusions using the labeled streptavidin-biotin method. Sixteen cases of RM, 9 cases of MM, and 52 cases of AC from various sites, including 13 pulmonary primaries, were examined with primary antibodies against E-cad and Cal. RESULTS. The positive rates for E-cad and Cal, respectively, were as follows: RM, 0/16 (0%) and 16/16 (100%); MM, 9/9 (100%) and 8/8 (100%); and AC, 45/52 (86.5%) and 0/51 (0%). The E-cad expression by neoplastic cells was strongest in the intercellular junctions, and poorly differentiated neoplastic cells in the single cell form showed the weakest expression. CONCLUSIONS. In contrast to the results of previous immunohistochemical studies, the current study indicates that MMs constantly express E-cad, whereas RMs lack its expression in cytologic specimens, which would be useful in the differential diagnosis between the two. On the other hand, E-cad expression is not reliable for distinguishing AC from MM. The Cal expression can be a very useful marker for the distinction between AC and the mesothelial lineage. The combined immunostaining for E-cad and Cal has utility in differential diagnosis among RM, MM, and AC. Cancer (Cancer Cytopathol) 2000;90:55– 60. © 2000 American Cancer Society. KEYWORDS: immunostaining, mesothelial cell, mesothelioma, adenocarcinoma, Ecadherin, calretinin, cytologic diagnosis. Presented in part at the 39th General Meeting of the Japanese Society of Clinical Cytology, Sapporo, Japan, June 19, 1998. Address for reprints: Hiromi Kitazume, M.T., Division of Anatomic and Surgical Pathology, Hospital of Yokohama City University, Yokohama City University School of Medicine, 3-9 Fukuura, Kanazawa-ku, Yokohama 236-8566, Japan. Received February 1, 1999; revision received June 11, 1999; accepted June 18, 1999. © 2000 American Cancer Society I t has long been a major diagnostic challenge to distinguish among reactive mesothelial cells (RMs), malignant mesothelioma (MM), and adenocarcinoma (AC) in both cytologic and surgical pathologic specimens. For differential diagnosis between MM and AC, previous immunohistochemical and immunocytochemical studies established panels of useful antibodies against immunodeterminants, including carcinoembryonic antigen (CEA), CD15 (recognized by the monoclonal antibody Leu M1), tumor-associated glycoprotein 72 (recognized by the monoclonal antibody B72.3), BerEP4 glycoprotein, and MOC-31 glycoprotein.1–5 Most of these are markers for AC and can be 56 CANCER (CANCER CYTOPATHOLOGY) February 25, 2000 / Volume 90 / Number 1 TABLE 1 Summary of Cases Cytodiagnosis Pleural fluids Peritoneal fluids Total Reactive mesothelial cells Malignant mesothelioma Adenocarcinoma Lung Breast Stomach Bile duct Pancreas Colon Endometrium Ovary Total 16 9 31 13 4 4 0 3 1 2 4 56 0 0 21 0 0 4 2 2 3 4 6 21 16 9 52 13 4 8 2 5 4 6 10 77 useful in the setting of differential diagnosis between RM and AC as well. For differentiation between RMs and MMs, on the other hand, only a few markers have been reported to be of utility. Among these are epithelial membrane antigen (EMA) and p53 protein, which are often expressed in MMs but usually not in RMs.6,7 Recently, two new immunohistochemical markers have come into notice: E-cadherin for ACs8,9 and calretinin for the mesothelial lineage.10 Because there are only a few reports of experience with the immuonostaining of these new markers in cytologic specimens,11,12 we investigated their utility in cytologic differential diagnosis. MATERIALS AND METHODS A total of 77 cytologic specimens of pleural or peritoneal fluids collected between 1991 to 1997 were retrieved from the cytology files of our laboratories (Table 1). They included 16 cases of RM, 9 cases of MM, and 52 cases of AC. All cases of RM and MM were from specimens of pleural fluids. Of 16 cases of RM, the direct causes of the effusions were infectious in 9 cases, combined infectious and malignancy (lung involvement) in 3 cases, malignant (with lung involvement) in 1 case, traumatic in 1 case, and unknown in 2 cases. Of these, 12 cases had underlying malignant disease at some site, but none of them showed malignant cells in the pleural effusions during their clinical courses. The primary sites for the 52 cases of AC included 13 lungs, 4 breasts, 8 stomachs, 2 bile ducts, 5 pancreases, 4 colons, 6 endometria, and 10 ovaries. The diagnoses of all neoplastic cases had been confirmed histopathologically. MMs were of either epithelial or biphasic type. Immunocytochemical staining was performed using smears on the slides originally stained by Papani- colaou dye and later decolorized with 0.5% hydrochloric acid in 95% ethanol for this study. The primary antibody against E-cadherin (mouse monoclonal HECD-1; Takara, Otsu, Japan) was diluted 1:400 in 20 mM Tris-buffered saline (TBS), pH 7.4, with 1.0% bovine albumin and 10mM CaCl2. The anticalretinin antibody (rabbit polyclonal; Swant, Bellinzona, Switzerland) was diluted 1:8000 in 10 mM phosphatebuffered saline (PBS) with 1.0% bovine albumin. The decolorized slides were rinsed in tap water for 5 minutes. The slides were then incubated with 0.3% hydrogen peroxide for 30 minutes to quench the endogenous peroxidase activity. Following overnight incubation with the primary antibodies at room temperature, the slides were rinsed twice in TBS for 5 minutes each. The labeled streptavidin-biotin method (LSAB2 Kit; Dako, Carpinteria, CA) was used for the subsequent immunostaining procedure. Chromogen 3, 3⬘diaminobenzidine solution was used for detection of the antigens. The smears were finally counterstained with hematoxylin and coverslipped for examination. Negative controls were run in parallel by replacing the primary specific antibody with mouse immunoglobulin G1 (Dako, Glostrup, Denmark) or rabbit immunoglobulins. The immunostaining was graded on a sliding scale of ⫺ to 3⫹ according to the percentage of positive cells (⫺, no positive cell; ⫹, 1–9%; ⫹⫹, 10 – 89%; ⫹⫹⫹, 90 –100%). In the final evaluation of the staining for calretinin, there were not enough cells on the slides in two cases (one mesothelioma and one ovarian adenocarcinoma), and these two cases were excluded from the analysis of calretinin immunostaining. The relation between cytologic differentiation and the degree of expression for E-cadherin in ACs was analyzed by the Mann–Whitney U test. For this purpose, all ACs of the digestive organs were classified as poorly, moderately, or well differentiated based on the degree of cohesiveness of neoplastic cells in the cytologic specimens, and also were scored from 0 to 3, with these scores corresponding to grades of immunoreactivity. RESULTS For the immunostaining of E-cadherin (Table 2), the positivity rate in each category of the cases was as follows: RM, 0 of 16 cases (0%); MM, 9 of 9 cases (100%); and AC, 45 of 52 cases (87%). All 13 ACs of the lung primary were positively stained. The positive rates for ACs of the other sites varied from 63% (stomach primary) to 100% (bile duct and colon primaries). Three negative cases of the gastric ACs were all poorly differentiated adenocarcinoma or signet-ring cell carcinoma. When the cytologic differentiation was eval- E-Cadherin and Calretinin Immunostaining/Kitazume et al. 57 TABLE 2 Immunoreactivity for E-Cadherin E-cadherin immunoreactivity Cytodiagnosis No. of positive cases/ no. of cases (%) Reactive mesothelial cells Malignant mesothelioma Adenocarcinoma Lung Breast Stomach Bile duct Pancreas Colon Endometrium Ovary 0/16 (0) 9/9 (100) 45/52 (87) 13/13 (100) 3/4 (75) 5/8 (63) 2/2 (100) 4/5 (80) 4/4 (100) 5/6 (83) 9/10 (90) ⴙⴙⴙ ⴙⴙ 7 20 11 2 9 ⴙ ⴚ 16 1 3 1 4 3 1 2 2 1 16 2 3 2 7 1 2 2 4 1 1 3 1 1 ⫺: no positive cell; ⫹: positive cell ⬍10%; ⫹⫹: 10% ⱕ positive cell ⬍ 90%; ⫹⫹⫹: positive cell ⱖ 90%. TABLE 3 A Correlation Analysis between Cytologic Differentiation and Ecadherin Immunoreactivity in Adenocarcinomas of Digestive Organs E-cadherin immunoreactivity mesothelial cells in the background of adenocarcinomas were consistently positive for calretinin (Fig. 1I). DISCUSSION Cytologic differentiation ⴙⴙⴙ ⴙⴙ ⴙ ⴚ Total Poor Moderate Well Total 0 3 16 19 1 5 1 7 3 6 3 12 3 1 0 4 7a,b 15a,c 20b,c 42 ⫺: no positive cell; ⫹: positive cell ⬍10%; ⫹⫹: 10% ⱕ positive cell ⬍ 90%; ⫹⫹⫹: positive cell ⱖ 90% Mann–Whitney U test: aP ⫽ 0.0376; bP ⫽ 0.0006; cP ⫽ 0.0035. uated in relation to the expression of E-cadherin in ACs of the digestive organs (stomach, bile duct, and pancreas primaries combined) (Table 3), cytologically well-differentiated ACs showed significantly higher expression of E-cadherin than moderately differentiated ACs (P ⫽ 0.0035) or poorly differentiated ACs (P ⫽ 0.0006). The localization of the E-cadherin staining was generally membranous and was especially intense along the intercellular junctions (Fig. 1E, H; Fig. 2A, B). Poorly differentiated ACs in the single cell form, including signet-ring cell carcinoma, often showed dotlike cytoplasmic staining, which often appeared to be perinuclear (Fig. 2C). Reactive mesothelial cells in the backgrounds of adenocarcinomas were consistently negative for E-cadherin. As shown in Table 4, the positivity rate of calretinin expression was 16 of 16 cases (100%) for RM, 8 of 8 cases (100%) for MM, and 0 of 51 cases (0%) for AC. All cases of RM and MM showed many positive cells with strong staining intensity in the cytoplasm and often in the nucleus as well (Fig. 1C, F). Reactive E-cadherin is a member of the Ca2⫹-dependent cell adhesion molecules that is specifically expressed on all epithelial cells.13 Soler et al. studied the expression of E-cadherin (and also N-cadherin and catenins) in MMs and pulmonary ACs using immunohistochemical methods on frozen tissue sections.8 They found that all 19 ACs expressed E-cadherin, whereas a minority of the cells in some MMs expressed weak and mostly cytoplasmic staining for E-cadherin, leading them to advocate the utility of E-cadherin in differentiating between pulmonary ACs and MMs. A few years later, the same group reported similar results using formalin fixed, paraffin embedded tissues.9 In an immunohistochemical study reported by Leers et al.,14 20 of 21 metastatic ACs of the pleura were positive for E-cadherin, whereas only 3 of 20 MMs were immunoreactive for E-cadherin; and most were focal, with less than 5% of tumor cells positive. In contrast to these previous studies using tissue sections, the current study indicates that E-cadherin is constantly and clearly expressed in MMs and cannnot serve as a reliable marker for distinguishing ACs from MMs in cytologic specimens. The discrepancy between the results of our study and the previous report by Leers et al.14 may be due partly to the difference in the state of the antigen preservation, because both studies used the same monoclonal antibody (HECD-1) and immunostaining method (labeled streptavidin-biotin method). This view is supported by our recent observation that most MMs expressed E-cadherin in formalin fixed, paraffin embedded tissues when a highly sensi- 58 CANCER (CANCER CYTOPATHOLOGY) February 25, 2000 / Volume 90 / Number 1 FIGURE 1. (A) Papanicolaou staining of reactive mesothelial cells (RMs). (B) Immunostaining of RMs for E-cadherin. Note complete absence of staining. (C) Immunostaining of RMs for calretinin. The cells show intense cytoplasmic staining. (D) Papanicolaou staining of malignant mesothelioma (MM) cells. (E) Immunostaining of MM cells for E-cadherin. Intercellular junctions of a cell ball are strongly positive. (F) Immunostaining of MM cells for calretinin. Note diffuse cytoplasmic staining. (G) Papanicolaou staining of adenocarcinoma (AC) cells. (H) Immunostaining of AC cells for E-cadherin. Note strong positivity localized in the intercellular junctions. (I) Immunostaining of AC cells for calretinin. Note the sharp contrast between the negatively stained cluster of AC cells and the strongly positive cluster of RMs. tive detection method, the Catalyzed Signal Amplification (CSA) System (Dako, Carpinteria, CA), was used in the immunohistochemical staining (Mukai et al., unpublished data). Since the original study by Solar et al.8 used frozen tissues, it is hard to attribute the difference in the results to the state of the antigen preservation in this case. It may be of note that these authors do admit that MMs sometimes express E- cadherin and that its staining property alone should not be relied upon in the differentiation between ACs and MMs.8 Our study indicates that the expression of E-cadherin is a very useful marker for differentiating ACs or MMs from RMs in cytologic specimens. Schofield et al. demonstrated that E-cadherin immunostaining is useful in the former setting, ACs versus RMs.12 Our study E-Cadherin and Calretinin Immunostaining/Kitazume et al. 59 FIGURE 2. Immunostaining for E-cadherin in various adenocarcinomas (ACs). (A) Well-differentiated pancreatic AC cells. (B) Ovarian clear cell AC cells. (C) Poorly differentiated gastric AC cells. Note the dotlike cytoplasmic staining. TABLE 4 Immunoreactivity for Calretinin Calretinin immunoreactivity Cytodiagnosis No. of positive cases/ no. of cases (%) Reactive mesothelial cells Malignant mesothelioma Adenocarcinoma Lung Breast Stomach Bile duct Pancreas Colon Endometrium Ovary 16/16 (100) 8/8 (100) 0/51 (0) 0/13 (0) 0/4 (0) 0/8 (0) 0/2 (0) 0/5 (0) 0/4 (0) 0/6 (0) 0/9 (0) ⴙⴙⴙ ⴙⴙ ⴙ ⴚ 16 8 51 13 4 8 2 5 4 6 9 ⫺: no positive cell; ⫹: positive cell ⬍10%; ⫹⫹: 10% ⱕ positive cell ⬍ 90%; ⫹⫹⫹: positive cell ⱖ 90%. further suggests that it would be especially useful in the latter setting as well, as there have been only a few markers available, such as EMA and p53, to distingusish MMs from RMs.6,7,15 It was reported by van der Kwast et al. that of 25 serous fluid smears from patients with MM, 23 cases showed varying numbers of EMA-stained cells, whereas reactive mesothelial cells in 56 smears were virtually all negative.6 The finding that EMA and E-cadherin, both normally expressed in epithelial cells, are detected in MMs suggests that mesothelial cells, which are of mesodermal origin, may acquire epithelial characters when they are transformed. Our study also demonstrates that poorly differentiated ACs, in the single cell form or in a few cell clusters, show less expression of E-cadherin than welldifferentiated ACs in tight, large clusters. Because Ecadherin plays a central role in epithelial cell– cell adhesions, this observation is not unexpected and actually is in accordance with the results of immunohis- tochemical examination.16 E-cadherin immunostaining thus may be more informative and of practical diagnostic value in a situation where it is necessary to distingush well-differentiated ACs, rather than poorly differentiated ACs, from RMs. It is noteworthy that some poorly differentiated gastric AC cells showed dotlike cytoplasmic immunostaining for E-cadherin. Some previous immunohistochemical studies17,18 have described the “disorganized” pattern of cytoplasmic staining for E-cadherin in carcinomas of advanced stage or of invasive nature. Our finding also suggests that the disorganized cytoplasmic staining pattern may be related to poor cohesiveness of the neoplastic cells. In this regard, it is interesting that E-cadherin gene mutations have been detected in gastric signet-ring cell carcinomas.19 Calretinin is a 29 kD calcium-binding protein that normally is expressed in neuronal cells. Recent immunohistochemical studies10,20 have suggested that it is a very useful marker for cells of mesothelial lineage, both reactive and neoplastic. Barberis et al. have reported recently that calretinin is just as useful in evaluating cytologic specimens as it is in the histologic differential diagnosis between cells of the mesothelial lineage and ACs.9 Observations in our study support the results of these studies, indicating the utility of calretinin as a good mesothelial marker. Although the current series of ACs in cytologic specimens included no positive cases, a small number of calretinin positive ACs have been noted in some previous studies.10,11,21 Our own experience also suggests the presence of a few calretinin positive ACs in immunohistochemical staining (Mukai et al., unpublished data). Thus, just as is the case with other immunocytolochemical markers, the results of immunostaining for calretinin should be evaluated in the whole context of examination with a panel of antibodies. 60 CANCER (CANCER CYTOPATHOLOGY) February 25, 2000 / Volume 90 / Number 1 TABLE 5 Typical Immunoreactivity for E-cadherin and Calretinin 9. Cytodiagnosis E-cadherin Calretinin Reactive mesothelial cell Malignant mesothelioma Adenocarcinoma ⫺ ⫹ ⫹ ⫹ ⫹ ⫺ 10. 11. In summary, the current study indicates that the combined immunostaining of E-cadherin and calretinin is a useful adjunct to the panels of antibodies for immunocytochemical differentiation among RMs, MMs, and ACs (Table 5). REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. Sheibani K, Esteban JM, Bailey A, Battifora H, Weiss LM. Immunopathologic and molecular studies as an aid to the diagnosis of malignant mesothelioma. Hum Pathol 1992;23: 107–16. McCaughey WTE, Colby TV, Battifora H, Churg A, Corson JM, Greenberg SD, et al. Diagnosis of diffuse malignant mesothelioma: experience of a US/Canadian mesothelioma panel. Mod Pathol 1991;4:342–53. Ordonez NG. Value of the Ber-EP4 antibody in differentiating epithelial pleural mesothelioma from adenocarcinoma. Am J Clin Pathol 1998;109:85–9. Osborn M, Domagala W. Immunocytochemistry. In: Bibbo M. Comprehensive cytopathology. Chicago: Harcourt Brace Jovanovich, Inc., 1991:1011–51. Riera JR, Astengo-Osuna C, Longmate JA, Battifora H. The immunohistochemical diagnostic panel for epithelial mesothelioma. Am J Surg Pathol 1997;21:1409 –19. van der Kwast TH, Versnel MA, Delahaye M, Zondervan AJPE, Hoogsteden H. Expression of epithelial membrane antigen on malignant mesothelioma cells: an immunocytochemical and immunoelectron microscopic study. Acta Cytol 1988;32:169 –74. Henderson DW, Shilkin KB, Whitaker D. Reactive mesothelial vs. mesothelioma, including mesothelioma in situ. Am J Clin Pathol 1998;110:397– 404. Soler AP, Knudsen KA, Jaurand M-C, Johnson KR, Wheelock MJ, Klein-Szanto AJP, et al. The differential expression of N-cadherin and E-cadherin distinguishes pleural mesotheliomas from lung adenocarcinomas. Hum Pathol 1995;26: 1363–9. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. Han AC, Peralta-Soler A, Knudsen KA, Wheelock MJ, Johnson KR, Salazar H. Differential expression of N-cadherin in pleural mesotheliomas and E-cadherin in lung adenocarcinomas in formalin-fixed, paraffin-embedded tissues. Hum Pathol 1997;28:641–5. Doglioni C, Dei Tos AP, Laurino L, Iuzzolino P, Chiarelli C, Celio MR, et al. Calretinin: a novel immunocytochemical marker for mesothelioma. Am J Surg Pathol 1996;20:1037– 46. Barberis MCP, Faleri M, Veronese S, Casadio C, Viale G. Calretinin: a selective marker of normal and neoplastic mesothelial cells in serous effusions. Acta Cytol 1997;41:1757– 61. Schofield K, D’Aquila T, Rimm DL. The cell adhesion molecule, E-cadherin, distinguishes mesothelial cells from carcinoma cells in fluids. Cancer (Cancer Cytopathol) 1997;81: 293– 8. Takeichi M. Cadherins: a molecular family important in selective cell-cell adhesion. Annu Rev Biochem 1990;59:237– 52. Leers MPG, Aarts MMJ, Theunissen PHMH. E-cadherin and calretinin: a useful combination of immunochemical markers for differentiation between mesothelioma and metastatic adenocarcinoma. Histopathology 1998;32:209 – 16. Walts AE, Said JW, Shintaku IP. Epithelial membrane antigen in the cytodiagnosis of effusions and aspirates. Diagn Cytopathol 1987;3:41–9. Takeichi M. Cadherin cell adhesion receptors as a morphogenetic regulator. Science 1991;251:1451–5. Bongiorno PF, Al-Kasspooles M, Lee SW, Rachwal WJ, Moore JH, Whyte RI, et al. E-cadherin expression in primary and metastatic thoracic neoplasms and in Barrett’s esophagus. Br J Cancer 1995; 71:166 –72. Hidaka N, Nagao T, Asoh A, Kondo Y, Nagao K. Expression of E-cadherin, ␣-catenin, ␤-catein, and ␥-catenin in bronchioloalveolar carcinoma and conventional pulmonary adenocarcinoma: an immunohistochemical study. Mod Pathol 1998;11:1039 – 45. Muta H, Noguchi M, Kanai Y, Ochiai A, Nawata H, Hirohashi S. E-cadherin gene mutations in signet ring cell carcinoma of the stomach. Jpn J Cancer Res 1996;87:843– 8. Vogt VGP, Celio MR. The calcium binding protein Calretinin is a selective marker for malignant pleural mesotheliomas of the epithelial type. Pathol Res Pract 1996;192:137– 47. Ordóñetz NG. Value of calretinin immunostaining in differentiating epithelial mesothelioma from lung adenocarcinoma. Mod Pathol 1998;11:929 –33.