Cytologic evaluation of cyclin D1 expression in primary breast carcinoma

CANCER 211 CYTOPATHOLOGY Cytologic Evaluation of Cyclin D1 Expression in Primary Breast Carcinoma Kyeongmee Park, M.D.1 BACKGROUND. Preoperative assessment of the biologic characteristics of primary Sehwan Han, M.D., Ph.D.2 breast carcinoma is important because neoadjuvant medical therapy is being used Hong-Yong Kim, M.D., Ph.D.2 increasingly. In the current study, the authors attempted to evaluate the validity of Ilhyang Ko, M.D., Ph.D.1 cyclin D1 assay in fine-needle aspiration (FNA) samples from patients with primary breast carcinoma. 1 Department of Pathology, Inje University Sanggye METHODS. FNA samples were obtained prior to therapy and multiple slides were Paik Hospital, Seoul, South Korea. stored at -80 °C for subsequent immunocytochemical analysis (ICA). ICA for cyclin 2 Department of Surgery, Inje University Sanggye D1 protein was performed on FNA samples from 51 breast carcinoma patients and Paik Hospital, Seoul, South Korea. 20 samples from patients with benign breast disease. In 45 breast carcinoma patients who had undergone surgery, sections were taken from paraffin blocks and stained by ICA for cyclin D1 validation. Possible correlations between cyclin D1 expression in the FNA samples and the biologic data of the patients also were analyzed. RESULTS. Cyclin D1 expression was detected in 37 FNA samples from 51 breast carcinomas (72.5%) whereas expression of cyclin D1 was detected in 8 FNA sam- ples from 20 patients with benign breast disease (40%). In histologic sections after surgery, 26 cases of breast carcinoma (65%) showed a positive reaction to cyclin D1. Concordance for the presence of cyclin D1 between FNA samples and histo- logic samples was 75%. Cyclin D1 expression was high in patients with the tumors that expressed estrogen receptor (ER) (30 of 34 vs. 5 of 11; P 5 0.028) and progesterone receptor (PR) (33 of 38 vs. 2 of 7; P 5 0.007). There was no significant relation found between cyclin D1 expression and tumor size or lymph node metastasis. Cyclin D1 expression within invasive ductal carcinoma was observed in . 80% of low or intermediate nuclear grade tumors but its expression decreased to 61.5% (8 of 13 cases) in tumors with high nuclear grade (P 5 0.023). All 14 breast carcinomas in which the S-phase fraction was # 15% showed a high degree of cyclin D1 expression whereas 4 of 6 tumors (66.7%) in which the S-phase fraction was . 15% showed cyclin D1 expression. Cyclin D1 expression was found to be correlated inversely with proliferative activity in breast carcinoma (P 5 0.023). CONCLUSIONS. The results of the current study show that cyclin D1 expression can be measured by ICA in FNA samples with reasonable concordance with the results of histologic section. Cyclin D1 expression was found to be associated with ER/PR status and cell differentiation. The results of the current study indicate that the Supported by a grant from the Inje Research Foun- measurement of novel molecular markers could be performed adequately in FNA dation. samples as well as in histologic sections. Cancer (Cancer Cytopathol) 2001;93:211–215. © 2001 American Cancer Society. Address for reprints: Kyeongmee Park, M.D., De- partment of Pathology, Inje University Sanggye Paik Hospital, 761-1 Sanggye-Dong, Nowon-Gu, KEYWORDS: breast carcinoma, cyclin D1, differentiation, fine-needle aspiration, Seoul 139-707, South Korea; Fax: (011) 82-2-950- immunocytochemical assay. 1266; E-mail: [email protected] Received May 2, 2000; revisions received January 23, 2001, and February 5, 2001; accepted Febru- ary 8, 2001. T he cyclin D1 gene, CCND1, is amplified in approximately 20% of human breast carcinomas,1 and is overexpressed in approximately 50% of breast tumors.2 To our knowledge, a number of studies have © 2001 American Cancer Society 212 CANCER (CANCER CYTOPATHOLOGY) June 25, 2001 / Volume 93 / Number 3 been conducted to determine whether cyclin D1 is a fluorescence was analyzed using FACScan (Becton- useful biologic marker in the treatment of breast car- Dickinson, Bedford, MA) as described previously.6 cinoma patients.3–5 To our knowledge the majority of Data acquisition was performed using Cell Fit soft- studies have used tissue sections for cyclin D1 assay ware (Becton-Dickinson) and data analysis was per- but no report to date has examined cyclin D1 expres- formed using the Phoenix Flow System Multicycle AV sion using fine-needle aspiration (FNA) samples. software. FNA has become a useful method with which to make an accurate diagnosis of breast carcinoma pre- operatively. Wide acceptance of breast conservation Immunocytochemical and Immunohistochemical Assay surgery has increased the use of preoperative chemo- FNA samples were taken from all patients prior to therapy to downstage locally advanced breast carci- treatment using a 23-gauge needle and a 10-mL sy- noma. Tumor biology may be altered after chemo- ringe. Multiple FNA slides were fixed in 10% formalin- therapy because of clonal selection. Pretreatment phosphate-buffered saline (PBS) for 15–30 minutes at assessment of tumor biology may be useful in deter- -4 °C and rinsed with pH 7.0 PBS. Slides were fixed in mining a treatment plan for individual patients with methanol for 5 minutes at -20 °C and rinsed with PBS. breast carcinoma. If the biologic characteristics of the Slides were placed in acetone for 2 minutes at 20 °C tumor could be assessed more precisely, it might be and rinsed with PBS. ICA subsequently was performed possible to provide a more reasonable treatment plan on these samples. FNA slides and 5-mm tissue sections to breast carcinoma patients while minimizing ad- were microwaved in 10 mM ethylenediamine tetraace- verse side effects. tic acid for 25 minutes in a 600-watt microwave oven The current study investigated whether cyclin D1 for pretreatment before immunostaining. These sam- expression might be detected accurately in tumor cells ples were examined for expression of cyclin D1 protein obtained by FNA. Immunocytochemical assay (ICA) using the avidin-biotin complex-immunoperoxidase for cyclin D1 also was undertaken in paraffin-embed- method. FNA slides and 5-mm tissue sections were ded tumor sections and the results were correlated incubated in commercially available monoclonal an- with those from FNA samples. The correlation be- tibody (NCL-CYCLIN D1-GM [1:20 dilution] for cyclin tween cyclin D1 expression in FNA with clinical/his- D1 protein assay [Novocastra Laboratories Ltd., New- tologic variables was analyzed to estimate the clinical castle-upon-Tyne, UK]). The reaction was detected implications of cyclin D1 expression in FNA samples using diaminobenzidine with hydrogen peroxide. of primary breast carcinoma. Slides were counterstained with hematoxylin. MATERIALS AND METHODS Materials Slide Review ICA for cyclin D1 was performed in 51 patients with FNA slides were reviewed by two cytopathologists primary breast carcinoma using FNA samples. Three (K.P. and I.K.) without knowledge of the patients’ out- patients with ductal carcinoma in situ (DCIS) were come data. Cellularity was regarded as adequate when included. Cyclin D1 expression also was assayed in 20 there were . 30 neoplastic cells per slide. Histologic cases of FNA samples from patients with benign breast sections also were reviewed by two pathologists (K.P. disease. The mean age of the patients was 47 years and I.K.) with had no prior knowledge of the original (range, 19 – 82 years). There was one male patient with cytologic diagnosis. Three separate blocks containing invasive ductal carcinoma (IDC). Six patients did not breast carcinoma were stained and scored by calcu- undergo tumor excision and were excluded from data lating the stained cells from 500 observed carcinoma analysis of tissue cyclin D1 expression. Disease stage cells as a percentage. Breast carcinoma sections that was determined according to the TNM classification of were observed to express homogeneous and intense the American Joint Committee on Cancer. Five pa- nuclear immunostaining for cyclin D1 protein in . 5% tients had Stage I disease (11.1%), 20 had Stage II of cells in the observed field were considered to be disease (44.4%), 16 had Stage III disease (35.6%), and 1 positive (Fig. 1). This cutoff level in the current study had Stage IV disease (2.2%). was found to correlate well with clinical outcome based on the tumor marker study undertaken at the Cell Cycle Analysis study institution. With regard to determining nuclear Flow cytometric analysis was performed on cell sus- grade, we used the Nottingham modification of the pensions obtained from 20 patients with breast carci- Bloom-Richardson system for breast carcinoma.7 Nu- noma. Only 20 of 51 FNA samples were found to have clei were graded as 1 for Scores 2 and 3, 2 for Scores 4 adequate cellularity for flow cytometric analysis. DNA and 5, and 3 for Score 6.  Cyclin D1 Expression in Breast Carcinoma/Park et al. 213 FIGURE 1. Fine-needle aspiration of ductal carcinoma exhibiting positive FIGURE 2. Histologic section of invasive ductal carcinoma showing intense nuclear immunoreactivity to cyclin D1 (3200). nuclear immunoreactivity to cyclin D1 (3200). TABLE 1 TABLE 2 Cyclin D1 Expression in Fine-Needle Aspiration Specimens Cyclin D1 Expression in Fine-Needle Aspiration and Biopsy Specimens of Breast Carcinoma Cyclin D1 expression (%) Cyclin D1 in cytology (%) Positive Negative Cyclin D1 in tissue Positive Negative Benign disease 8 (40.0) 12 (60.0) DCIS 3 (100.0) 0 (0.0) Positive 23 (57.5) 3 (7.5) IDC 37 (77.1) 11 (22.9) Negative 7 (17.5) 7 (17.5) DCIS: ductal carcinoma in situ; IDC: invasive ductal carcinoma. Concordance: 75% Statistical Analysis FNA or on histologic tissue sections. The concordance Data for the ICA were merged with the clinicopatho- rate of cyclin D1 expression in the FNA and histologic logic characteristics of the patients to evaluate the sections was 75% (30 of 40 cases). Results of cyclin D1 clinical implication of cyclin D1 protein expression in expression in FNA and histologic sections of breast FNA samples. The correlation between cyclin D1 ex- carcinoma are described in Table 2. pression in FNA and histologic samples was analyzed The results of the cyclin D1 assay in FNA samples to estimate the accuracy of cyclin D1 assay using FNA. were analyzed according to the histologic and biologic Statistical analyses were performed using the SPSS characteristics of the primary breast carcinoma. In package program (SPSS Inc., Chicago, IL). A possible estrogen receptor (ER)-positive tumors, cyclin D1 ex- correlation between clinical/biologic parameters and pression was detected in 88.2% of cases (30 of 34 cyclin D1 expression was estimated using the chi- tumors) whereas ER-negative tumors showed cyclin square test. D1 expression in 45.5% of cases (5 of 11 tumors). The difference in cyclin D1 expression according to ER RESULTS status was found to be statistically significant (P 5 In FNA samples, cyclin D1 expression was observed in 0.028). This finding was similar when cyclin D1 ex- 37 of 48 IDC cases (77.1%) whereas cyclin D1 expres- pression was analyzed according to progesterone re- sion was detected in 8 of 20 patients with benign ceptor (PR) status. Thirty-three of 38 PR-positive tu- breast disease (40.0%). Three cases of DCIS were mors (86.8%) showed a strong reaction to the cyclin found to demonstrate positive immunoreaction to cy- D1 antibody, whereas only 2 of 7 PR-negative tumors clin D1 antibody (Table 1). In histologic sections, 26 showed a positive reaction (P 5 0.007). cases (65.0%) showed a positive reaction to cyclin D1 There was a significant difference in cyclin D1 (Fig. 2). Cyclin D1 was expressed in 23 cases in both expression according to the nuclear grade of the IDC FNA samples and histologic sections. Cyclin D1 ex- (P 5 0.042). The frequency of cyclin D1 expression pression was not observed in seven cases either by was . 80% in low or intermediate nuclear grade tu- 214 CANCER (CANCER CYTOPATHOLOGY) June 25, 2001 / Volume 93 / Number 3 mors. The frequency of cyclin D1 expression de- TABLE 3 creased to 61.5% (8 of 13 cases) in high nuclear grade Relation between Cyclin D1 Expression in Fine-Needle Aspiration Specimens and Clinical-Biologic Parameters tumors (P 5 0.023). Cyclin D1 expression appears to increase in better differentiated breast carcinomas. Cyclin D1 expression (%) Results of cell cycle analyses using flow cytometry Chi-square showed that 7 patients (28.0%) had diploid tumors Negative Positive value P value whereas 18 patients (72.0%) had nondiploid tumors. Menopause 0.97 0.516 No significant association was found between tumor Pre 6 (20.0) 24 (80.0) ploidy and cyclin D1 expression in the FNA samples. Post 5 (27.8) 13 (72.2) All 14 tumors in which the S-phase fraction was , 15% Size (cm) 0.024 0.683 showed high cyclin D1 expression whereas 4 of 6 tu- ,2 1 (20.0) 4 (80.0) mors with a high S-phase fraction (66.7%) had in- 2–5 8 (24.2) 25 (75.8) .5 1 (16.7) 5 (83.3) creased levels of cyclin D1 expression. Cyclin D1 ex- AJCC Stage 4.869 0.104 pression was found to be associated with S-phase DCIS 0 (0.0) 3 (100.0) fraction in the tumors. I 2 (40.0) 3 (60.0) We also analyzed cyclin D1 expression in FNA II 4 (20.0) 16 (60.0) samples according to the size and lymph node status III 3 (18.8) 13 (81.2) IV 0 (0.0) 1 (100.0) of the breast tumor. No significant correlation was LN metastasis 1.762 0.168 found between cyclin D1 expression in FNA samples Absent 6 (33.3) 12 (66.7) and tumor size or axillary lymph node involvement Present 4 (16.0) 21 (84.0) (Table 3). ER 9.094 0.028 Negative 6 (54.5) 5 (45.5) Positive 4 (11.8) 30 (88.2) DISCUSSION PR 11.994 0.007 Concordance between cytologic and histologic sam- Negative 5 (71.4) 2 (28.6) ples with regard to cyclin D1 expression was 75%. This Positive 5 (13.2) 33 (86.8) result indicates that cyclin D1 expression can be tested NG 4.117 0.042 adequately in cytologic samples. Makris et al. and 1 5 (38.5) 8 (61.5) 2 4 (14.3) 24 (85.7) Nizzoli et al. reported that the cytologic evaluation of 3 0 (0.0) 4 (100.0) biologic prognostic markers such as ER, PR, p53, and Ploidy 0.048 0.645 Ki-67 could be performed successfully on FNA sam- Diploid 1 (14.3) 6 (85.7) ples.8,9 The concordance rate of their studies was ap- Aneuploid 2 (11.1) 16 (88.9) proximately 75%, a rate that is within the same range S-phase 5.185 0.023 # 15% 0 (0.0) 14 (100.0) of the current study. FNA has been used mainly for . 15% 2 (33.3) 4 (66.7) diagnostic purposes in patients with breast carci- noma. The current study was conducted to evaluate AJCC: American Joint Committee on Cancer; DCIS: ductal carcinoma in situ; LN: lymph node; ER: whether cyclin D1 expression could be assayed suc- estrogen receptor; PR: progesterone receptor; NG: nuclear grade. cessfully in FNA samples. Accurate assessment of the biologic properties of the breast tumor preoperatively may provide patients with more optimal therapy. The cyclin D1 quantitatively.11–15 A similar finding was wider acceptance of preoperative chemotherapy for demonstrated in the current study. Cyclin D1 expres- breast carcinoma patients mandates the development sion in FNA samples was found to be increased sig- of methods to measure biologic markers to predict nificantly in ER-positive or PR-positive breast tumors. tumor response and prognosis. Although to our If cyclin D1 expression is linked closely to ER expres- knowledge cyclin D1 expression has not been estab- sion in breast tissues, response to tamoxifen treatment lished as a valid prognostic or predictive marker in could be predicted by assaying cyclin D1 expression in breast carcinoma, the results of the current study in- FNA samples. dicate the clinical utility of measuring molecular Recently, Collecchi et al. reported that cyclin D1 markers in FNA samples. expression showed different patterns according to ax- In breast carcinoma cell lines, the expression of illary lymph node status.16 This finding suggests that cyclin D1 precedes estrogen-induced progression cyclin D1 may be regulated differently during tumor through the G1 phase of the cell cycle.10 A number of progression. However, we could not find a correlation reports have examined cyclin D1 expression among between cyclin D1 expression and tumor size or axil- panels of human breast carcinoma cell lines. These lary lymph node involvement. In a certain proportion studies found that ER-positive cell lines overexpressed of patients with benign breast disease, cyclin D1 ex-  Cyclin D1 Expression in Breast Carcinoma/Park et al. 215 pression was evident in FNA samples. However, the expression of p27Kip1 protein is associated with poor clin- mechanism by which cyclin D1 was expressed in non- ical outcome of the breast cancer patients treated with systemic chemotherapy and is linked to cell proliferation cancerous breast epithelia must be investigated fur- and differentiation. Breast Cancer Res Treat 1999;55:161–7. ther. 7. Elston CW, Ellis IO. Pathological prognostic factors in breast The majority of studies concerning breast carci- cancer. I. The value of histological grades in breast cancer. noma have found that high levels of cyclin D1 expres- Experience from a large study with long-term follow-up. sion are associated with well-differentiated tu- Histopathology 1991;19:403–10. mors.17–19 A similar finding was observed based on the 8. Makris A, Allred DC, Powles TJ, Dowsett M, Fernando IN, Trott PA, et al. Cytological evaluation of biological prognos- results of the current study. Cyclin D1 expression was tic markers from primary breast carcinomas. Breast Cancer found to be decreased markedly in high-grade tumors Res Treat 1997;44:65–74. compared with tumors of low or intermediate nuclear 9. Nizzoli R, Bozzetti C, Naldi N, Guazzi A, Gabrielli M, Michi- grade (Table 3). The results of the current study sug- ara M, et al. Comparison of the results of immunocytochem- gest that cyclin D1 expression in primary breast car- ical assays for biologic variables on preoperative fine needle aspirates and on surgical specimens of primary breast car- cinoma may have a differentiation signal. Cyclin D1 cinomas. Cancer (Cancer Cytopathol) 2000;90:61– 6. and its kinase partners are involved in cell-cycle reg- 10. Prall OW, Rogan EM, Musgrove EA, Watts CK, Sutherland ulation and the overexpression of cyclin D1 acceler- RL. C-Myc or cyclin D1 mimics estrogen effects on cyclin ates and shortens the G1 phase.20 –22 However, we ob- E-Cdk2 activation and cell cycle reentry. Mol Cell Biol 1998; served an inverse correlation between cyclin D1 18:4499 –508. expression and S-phase in primary breast carcinoma. 11. Zwijsen RM, Klompmaker R, Wientjens EB, Kristel PM, van der Burg B, Michalides RJ. Cyclin D1 triggers autonomous The relation between cyclin D1 expression and the growth of breast cancer cells by governing cell cycle exit. proliferative activity of breast carcinoma must be de- Mol Cell Biol 1996;16:2554 – 60. termined by further study with a larger sample size. It 12. Bartkova J, Lukas J, Muller H, Lutzhoft D, Strauss M, Bartek also is possible that cyclin D1 plays a role in processes J. Cyclin D1 protein expression and function in human other than cell proliferation of breast carcinoma (e.g., breast cancer. Int J Cancer 1994;57:353– 61. 13. Lukas J, Pagano M, Staskova Z, Draetta G, Bartek J. Cyclin D1 cellular differentiation or genetic instabilities). protein oscillates and is essential for cell cycle progression In the current study, cyclin D1 expression was in human tumour cell lines. Oncogene 1994;9:707–18. measured successfully in FNA samples and the validity 14. Lukas J, Bartkova J, Bartek J. Convergence of mitogenic of this measurement was confirmed by high concor- signalling cascades from diverse classes of receptors at the dance of the study results with those from histologic cyclin D-cyclin-dependent kinase-pRb-controlled G1 check- sections. Results of the current study indicate that the point. Mol Cell Biol 1996;16:6917–25. 15. Han EK-H, Sgambato A, Jiang W, Zhang Y-J, Santella R, Doki measurement of novel molecular markers may be per- Y, et al. Stable overexpression of cyclin D1 in a human formed adequately in FNA samples as well as in his- mammary epithelial cell line prolongs the S-phase and in- tologic sections. hibits growth. Oncogene 1995;10:953– 61. 16. Collecchi P, Passoni A, Rocchetta M, Gnesi E, Baldini E, REFERENCES Bevilacqua G. Cyclin-D1 expression in node-positive (N1) 1. Barnes DM. Cyclin D1 in mammary carcinoma. J Pathol and node-negative (N-) infiltrating human mammary carci- 1997;181:267– 8. nomas. Int J Cancer 1999;84:139 – 44. 2. Barnes DM, Gillett CE. Cyclin D1 in breast cancer. Breast 17. Simpson JF, Quan DE, O’Malley F, Odom-Maryon T, Clarke Cancer Res Treat 1998;52:1–15. PE. Amplification of CCND1 and expression of its protein 3. Zuckerberg L, Yang W, Gadd M, Thor A, Koerner F, Schmidt product, cyclin D1, in ductal carcinoma in situ of the breast. E, et al. Cyclin D1 (PRAD1) protein expression in breast Am J Pathol 1997;151:161– 8. cancer: approximately one third of infiltrating mammary 18. Gillett C, Smith P, Gregory W, Richards M, Millis R, Peters G, carcinomas show overexpression of the cyclin D1 oncogene. et al. Cyclin D1 and prognosis in human breast cancer. Int J Mod Pathol 1995;8:560 –7. Cancer 1996;69:92–9. 4. Michalides R, Hageman P, van Tinteren H, Houben L, 19. van Diest P, Michalides RJAM, Jannink I, van der Valk P, Wientjens E, Klompmaker R, et al. A clinicopathological Meijer CJLM, Baak JPA. Cyclin D1 expression in invasive study on overexpression of cyclin D1 and of p53 in a series breast cancer. Am J Pathol 1997;150:705–11. of 248 patients with operable breast cancer. Br J Cancer 20. Baldin V, Lukas J, Marcote MJ, Pagano M, Draetta G. Cyclin 1996;73:728 –34. D1 is a nuclear protein required for cell cycle progression in 5. Barbareschi M, Pelosio P, Caffo O, Buttitta F, Pellegrini S, G1. Genes Develop 1993;7:812–21. Barbazza R, et al. Cyclin D1-gene amplification and expres- 21. Hartwell LH, Kastan MB. Cell cycle control and cancer. sion in breast carcinoma: relation with clinicopathologic Science 1994;266:1821–3. characteristics and with retinoblastoma gene product, p53 22. Jiang W, Kahn SM, Zhou P, Zhang YJ, Cacage AM, Infante and p21waf immunohistochemical expression. Int J Cancer AM, et al. Overexpression of cyclin D1 in rat fibroblasts 1997;74:171– 4. causes abnormalities in growth control, cell-cycle progres- 6. Han S, Park K, Kim HY, Lee MS, Kim HJ, Kim YD. Reduced sion and gene expression. Oncogene 1993;8:3447–57.