Expression of BCL-2 in primary breast cancer and its correlation with tumour phenotype

Annals of Oncology 5: 409-414, 1994. O 1994 Kluwer Academic Publishers. Printed in the Netherlands.

Original article Expression of BCL-2 in primary breast cancer and its correlation with tumour phenotype For the International (Ludwig) Breast Cancer Study Group B. Nathan,1 B. Gusterson,1 D. Jadayel,2 M. O'Hare,1 R. Anbazhagan,1 H. Jayatilake,1 S. Ebbs,5 K. Micklem,6 K. Price,4 R. Gelber,4 R. Reed,3 H.-J. Senn,3 A. Goldhirsch3 & M. J. S. Dyer2 Sections of Cell Biology and Experimental Pathology, and 2Academic Haematology, The Institute of Cancer Research, London, U.K.; International Breast Cancer Study Group, Bern, Switzerland; 'Division of Biostatistics, Dana-Farber Cancer Institute, Boston, U.S.A.; i Breast Unit, Mayday University Hospital, U.K.; ^Leukaemia Research Fund Immunodiagnostics Unit, John Radcliffe Hospital, Oxford, U.K. 3

Results: In the breast cancer derived cell line MCF-7 BCL-2 is expressed to a level similar to that of the B-lymBackground: The role of apoptosis (programmed cell death) phoma cell line Karpas 231 with t(14;18)(q32.3; q21.3), but in the development and progression of breast cancer is un- no evidence of a rearrangement or gene amplification was known. Recently the bcl-2 gene has been shown to block identified. In a study of 107 breast cancers from the Interapoptosis and thus may promote tumour development. nationa] Breast Cancer Study Group Trials I-IV we have BCL-2 is localised to the luminal cells of the normal breast, demonstrated a very significant inverse correlation of BCL-2 which are considered to be the origin of malignant breast with c-erbB-2 expression (p = 0.002), and a positive correlation with oestrogen receptors (p = 0.001) and progesterone disease. Patients and methods: Immunocytochemistry using anti receptors (p — 0.05). In this study there was no correlation of bcl-2— antibody was performed on 107 breast cancer speci- expression with S-phase fraction in the tumours or with any mens belonging to node-positive patients from the Ludwig stage in the cell cycle as assessed in MCF-7 cells. Breast Cancer Studies I-FV and the results were correlated Conclusion: We conclude that BCL-2 might contribute to with survival, tumour grade, S-phase, oestrogen and proges- the malignant phenotype of breast cancer by modulation of terone receptor status and c-erb B-2 expression. Western and biological behaviour of cancer cells. Southern blotting together with immunofluorescence were performed on the breast cancer cell lines BT-20, BT-474, MDA-MB-361, T47-D and MCF-7. Key words: BCL-2, breast cancer, apoptosis Summary

Background

Breast cancer is responsible for 15,000 deaths per years in the United Kingdom. Of all the prognostic factors that have been examined, lymph node involvement, tumour grade and tumour size remain the most significant [1, 2]. Within these parameters lies the inherent balance between the rate of cell gain and cell loss. It is well known that all tumours exhibit programmed cell death or apoptosis [3], but it is difficult to assess this process accurately at the light microscope level on routinely stained sections. Although a number of biochemical changes have now been identified that are markers of apoptosis none of them have proved to be robust markers to identify apoptotic cells in sections of human tissues. Recently an in situ technique of identifying DNA strand breaks has been developed that may aid the identification of cells before they can be recognised as apoptotic using morphological criteria [4,5].

Molecular cloning of the translocation t(14;18)(q32.3;q21.3) that is found in 85% of follicular lymphomas defined a gene on 18q21.3 termed bcl-2 [6, 7]. This gene encodes a 26kd protein of unknown function which was originally demonstrated biochemically in the inner mitochondrial membrane [8] but more recently has been shown to be specifically localised to the outer mitochondria] membrane and to the perinuclear region [9]. The t(14;18) translocation in B-cells results in a 1000-fold overexpression of the bcl-2 gene and there is very good evidence that expression of this gene protects cells from apoptosis [8]. In bc/-2-immunoglobulin transgenic mice there is B-cell extended survival [10] that later progress to lymphoid hyperplasia and malignant lymphoma [11]. In vitro results also indicate that the bcl-2 gene will promote cell survival of haemopoietic cells and co-operate with c-myc to immortalise pre-B-cells [12]. Recently BCL-2 expression has been detected in a

410 wide range of normal human tissues including breast epithelium [13]. In prostatic cancers BCL-2 is augmented by androgen ablation and correlates with the progression of prostatic cancer from androgen dependence to androgen independence [14]. It is of interest in relation to this steroid effect that during the normal menstrual cycle apoptosis, in both the uterus and breast, is cyclical with a peak of apoptosis in the second half of the cycle [15]. In view of these observations we decided to investigate the expression of BCL-2 in breast carcinomas and breast-derived cell lines. The results reported here indicate that BCL-2 is expressed in the majority of breast carcinomas, with increased expression in relation to the normal breast in some tumours and therefore is likely to have a role in the pathogenesis of this disease, possibly through an inhibition of apoptosis.

Cell lines studied The cell lines BT-20 [21], BT-474 [22], MDA-MB-361 [23] and T47-D [24] were obtained from the American Type Culture Collection. MCF-7 cells were obtained from the Michigan Cancer Foundation at passage 280, and CAL-51 [25] from the Antoine Lacassagne Cancer Center at passage 49. Cells were cultured routinely in Dulbecco's modified Eagle's medium with 10% foetal calf serum and antibiotics. For examination of BCL-2 expression by indirect fluorescence semi-confluent cells grown on coverslips were fixed for 3 minutes in ice-cold methanol, thoroughly rinsed in PBS/BSA (0.22 w/v), and stained for 40 minutes with the mouse monoclonal antiBCL-2 combination described above. After further washing, antibody binding was visualised with FTTC-conjugated sheep antimouse antiserum (Amersham International) at 1:40 (v/v) in PBS/ BSA, the cover slips were rinsed, mounted in 1:1 mixture of Hydromount (National Diagnostics) and AF2 anti-quenching mountant (Amersham International) and examined using a Zeiss fluorescence microscope equipped with a 63X oil immersion (NA 1.25) lens. Controls included a cell suspension of Karpas 231 [26] lymphoma cells carrying the t(14;18) translocation that results in high levels of expression. The breast cancer cell lines were also stained directly using sheep anti-mouse FITC conjugated antibodies as a control.

Patients and methods Patiens and data management

Molecular and protein analyses of BCL-2

Ludwig Breast Cancer Studies I-IV

Southern blotting

The material used in this study was from a cohort of patients treated between July 1978 and August 1981, with either a standard radical, a modified radical, or total mastectomy with axillary clearance. If lymph node positive, patients were entered into one of the Ludwig Breast Cancer Studies I-IV. These trials have all been described in detail previously [16-18]. Details of primary tumour size, number of involved nodes, steroid receptor status, and menopausal status were obtained at diagnosis. Histology review and grading were carried out centrally oestrogen and progesterone results of >10 fmol/mg cytosol protein were considered positive and lesser values were considered negative. These data, S-phase fraction and follow up data are kept at the trial data centre at the Dana Faber Cancer Institute. Formalin fixed, paraffin embedded blocks were availabe on only 107 cases from this trial, of which, 86 had known ER, 79 had known PR, 105 cases had known grade, 69 cases had known S-phase fraction and c-erbB-2 status was known in 105 cases. The median follow-up for these trials is 12 years. The material may have many selection biases indicating availability of paraffin blocks from some trial centres. The primary objective was therefore to study the distribution of BCL-2 expression with biological parameters rather than clinical outcome. We have carried out an analysis on patient follow up but any trends observed require a larger study of cases.

Conventional DNA Blotting was performed using previously described methods [26]. DNA was digested with restriction endonucleases (BamHI, EcoRI, Hind HI and Pst I). Pulsed-field gel electrophoresis (PFGE) was performed using high molecular weight DNA prepared from MCF-7 cells suspended in 1% low melting temperature agarose at 3 x 107 cells/ml"1. The agarose blocks were then incubated at 50*C for 48 hrs in a solution of 0.5M EDTA, pH 9.5 containing 1% sodium lauryl sarcosine and 1 mg/ml"1 proteinase K. Before digestion blocks were washed overnight at 4'C in 10 mM Tris HCL, pH 8, 1 mM EDTA (TE) containing 0.1 mM phenylmethylsulfonylfluoride (PMSF). Blocks were washed for an additional 3 hrs in TE without PMSF and a third of a block containing 106 cells was digested in situ with 30-40 units of enzyme for 6 hrs or overnight. The blocks were then placed in 1% agarose gel and these were subjected to electrophoresis in a clamped homogeneous electric field (CHEF) at 18*C for 24 hrs at 200 V with two different pulse times: 60 s for 15 hrs and 90 s for 9 hrs. Concatamers of X (New England Biolabs) and Saccharomyces cerevisiae were used as size standards. The DNA was then depurinated, by immersing the gel in 0.25M HCL twice for 10 min and then denatured in 0.5M NaOH/1.5M NaCl twice for 15 mins. Transfer of DNA to nylon membranes (Genescreen Plus) was by capillary blotting in 0.4M NaOH. The filters were washed in 0.2M tris-HCL pH 7.5/2xSSc and air dried to immobilise DNA. Hybridization was performed at 65"C in a buffer containing 10% dextran sulfate 0.5M sodium phosphate, pH 7.5, 1% SDS and 100 ug/ml salmon sperm DNA. After hybridization, filters were washed once with 2xSSc at 65'C for 10 mins followed by two washes with 2xSSc 1% SDS for 29 mins. Additional washes with 0.1% SSc 0.1% SDS for 15 mins were performed when necessary. DNA filters were hybridised with the major (pFL-1) or the minor (pFL-2) breakpoint cluster region of the bcl-2 gene. PFGE analysis was performed on DNA digested with rare cutting restriction endonuclease, Not I. This detects a 700 kb DNA fragment which spans the bctl gene.

Pilot study A comparative study was carried out on breast carcinomas frozen, and fixed either in modified methacam or in formalin. 20 examples of histologically normal breast tissue were studied together with 9 cases of extensive ductal carcinoma in situ. The pattern of immunoreactivity was similar in all cases, but staining intensity was reduced in the formalin relative to the other processing schedules.

Immunocytochemistry Staining was performed using the indirect immunoperoxidase Avidin-Biotin Complex technique. One hundred seven breast carcinomas were stained for BCL-2 immunoreactivity using two monoclonal antibodies (100 and 124) raised to the BCL-2 protein as previously described [19, 20]. In a pilot study it was found that the best results were obtained using equal volumes of both antibodies, diluted 1 in 2 in phosphate buffered saline (PBS).

Western blotting Cells from MCF-7 and W133 cell lines (approximately lOVtrack) were solubilised in sample buffer and run on 10% SDS PAGE gels [27]. The separated proteins were transferred to polyvinylidene difluoride membrane (Immobilon P, Millipore) using a semi-dry electroblotter [28]. The membrane was blocked with 5% dried

411 milk in 50 mM Tris-HCl, pH 7.4; 150 mM NaCl containing 0.1% Tween 20 (Sigma) in tris buffered saline (TBS). The membrane was incubated with antibody to the BCL-2 protein for 30 minutes. After a brief wash in TBS the blot was incubated with peroxidase conjugated goat anti-mouse Ig (DAKO, P447; 1:100) for 30 minutes. After three 30 minute washes in TBS the blot was developed with nickel-enhanced diaminobenzidine/hydrogen peroxide substrate. A cell suspension of Karpas 231 lymphoma cells with the t(14;18) translocation was used as a control.

Results MCF-7 cells showed a diffuse granular cytoplasmic staining pattern of intensity and distribution similar to that of Karpas 231 cells. BT-474 cells also showed a similar level of positivity. All of the other breast cancer cell lines examined were negative for BCL-2 immunoreactivity using indirect immunofluorescence. In view of the high intensity of staining of the MCF-7 cell line we examined these cells by conventional and pulsefield gel electrophoresis (PFGE) to determine whether the high expression of the bcl-2 gene product was due to DNA rearrangement or amplification at the bcl-2 locus. No DNA rearrangement or amplification was detected in a region of 700 Kb analysed by PFGE, suggesting that expression is regulated at the level of transcription. Expression of the protein was confirmed by Western blotting where the levels were comparable to those seen in the control Karpas 231 cells (Fig. 1). In view of the possibility that some tumours expressed high levels of BCL-2 a study was carried out on the expression of this protein in a well defined group of breast cancers with immunocytochemistry as the end point. Owing to the possible variability of staining due to fixation, all sections were checked to ensure positive staining of normal lymphocytes or normal breast

luminal epithelial cells in the same section as described previously [20]. High expression was defined as expressing stronger staining than the normal breast luminal cells and/or the infiltrating lymphocytes. Figure 2 shows a high expressing tumour. Table 1 shows the distribution of BCL-2 immunoreactivity with regard to trial and treatment group within the International (Ludwig) Breast Cancer Study Group Trials I-IV. Overall, 70% (76/107) of the pilot study patients had high expression of the BCL-2 protein. Among the patients receiving endocrine therapy (p + T x 12 months.) alone, 12 of the 14 patients in the group had high expression (86%). In the 17 patients who received no adjuvant therapy, 11 showed high expression (65%) (p = N.S). Table 2 shows the association of BCL-2 immunoreactivity with prognostic factors. In this very small sample, BCL-2 high expression appears to be associated with ER, PR, and c-erbB-2 positivity. 90% of ER+ patients are also BCL-2 positive, while 56% of ER— patients are BCL-2 high expressed (p < 0.001). BCL-2

Fig. 2. Histological section of a breast carcinoma immunostained with BCL-2 antibody cocktail (100 and 124) [x 250]. Table 1. Study population by treatment and trial. Trial Population

Randomised treatments

36kD —

I

CMFX12 CMFpxl2

241 250

22 17

17 (77) 13 (76)

29kD — 24kD —

II

CMFpxl2 OxCMFpXl2

161 166

14 10

7 (50) 6 (60)

Observation p+Txl2 CMFp+Txl2 Observation P+TX12

156 153 154 153 167

7 7 13 10 7

5 (71) 5 (71) 10 (77) 6 (60) 7(100)

1601

107

76 (70)

20kD —

Fig. 1. Western blot analysis of cell lines K231 (track 1), W133 (track 2) and MCF-7 (track 3) incubated with anti-human BCL-2 antibody (100, 124). Track 4 is a blank with no primary antibody. The positions of the molecular weight markers are as indicated.

III

IV

Pre- and perimenopausal 1-3 N + Pre- and perimenopausal 4N + Postmenopausal A11N + 65 yrs. Postmenopausal All N + 66-80 yrs Total

Number evaluated

Number BCL-2 studied

No. (%) BCL-2 overexpressed

CMF — cyclophosphamide methotrexate and 5-fluorouracil; CMFp — cyclophosphamide methotrexate and 5-fluorouracil and prednisone; Ox = oophorectomy; N = lymph nodes; T = tamoxifen.

412 Table 2. Association of BCL-2 and prognostic factors. Number of cases BCL-2 +

BCL-2-

ER+ ERER unknown

42 22 12

5 17 9

<0.001

PR+ PRPR unknown

20 40 16

2 17 12

0.05

Grade 1 Grade 2 Grade 3 Unknown

19 38 18 1

4 14 12 1

0.18

S-phase— S-phase+ Unknown

26 24 26

to vo O

p"

0.96

c-erbB-2+ c-erbB-2Unknown

5 69 2

to O to vo

Prognostic factor

0.002

1

chi-squared test excludes unknown cases. ER — oestrogen receptor, PR - progresterone receptor.

high expression is associated with PR positivity. 91% of PR+ patients are also BCL-2 high expressers, while 70% of PR- patients are BCL-2 high expressers (p = 0.05). Patients with low grade tumours tended to have a higher percentage of BCL-2 high expression (83%) than medium grade (73%) or high grade (60%) (p - 0.18). c-erbB-2 positive patients tended to have a relative low level of BCL-2 expression and vice versa (p ~ 0.002). There is no association with S-phase among the 69 patients with both factors known. While considering outcome for the 107 patients in terms of disease free survival (DFS) and overall survival (OS), other factors need to be considered, especially treatment; as 16% of these node positive patients were untreated. Looking just at the influence of BCL-2, patients with tumours expressing high levels of BCL-2 may have a better prognosis in terms of DFS (BCL-2 positive 12-yrs DFS - 45% ± 6; BCL-2 negative 37% ± 9; p = 0.26) and OS (BCL-2 positive 12-yr OS = 51 ± 6; BCL- 2 negative - 41% ± 9; p = 0.08), but the results are not statistically significant.

Conclusions Initial extraction data and biochemical characterisation of the protein used in the study confirmed the identity of the BCL-2 product. Both antibodies to the BCL-2 protein (100 and 124) gave similar results when used individually and also the mRNA levels in a range of cells lines paralleled the levels of expression of the protein as assessed by immunofluorescence. We have also demonstrated the high levels of expression of the BCL-2 protein by Western blotting. As previously reported, the normal breast epithelium and the

majority of breast carcinomas expressed BCL-2 [20]. Whilst a significant proportion of breast cancers expressed higher levels of BCL-2 than normal breast epithelium, we failed to demonstrate any correlation between BCL-2 expression and tumour type. There was a statistically significant correlation between tumour grade and BCL-2 immunoreactivity with low grade tumours tending to be more strongly BCL-2 positive and therefore levels of BCL-2 expression may be an important factor in determining the biological behaviour of breast carcinoma cells. Although our data showed a survival advantage for patients whose tumours expressed BCL-2, the results did not reach statistical significance. Recently, however, Pezzella and colleagues [29] demonstrated that survival at five years was higher in patients with non-small-cell lung carcinomas that expressed BCL-2 due to a similar association of BCL-2 expression with less aggressive tumour behaviour. Whilst it would be predicted that the tumour population with high expression of BCL-2 could expand due to a reduction in cell death, the overexpression of BCL-2 may facilitate differentiation. This has an analogy in lymphomas where transformation of lowgrade lymphoma may be associated with loss of expression or downregulation of BCL-2. Downregulation of BCL-2 expression might explain the high apoptotic indices commonly observed in high grade tumours. The occurrence of a differentiation block by other mechanisms in some tumours may account for the variability in the results observed. The lack of correlation with S-phase fraction and the absence of a correlation between expression of BCL-2 and any particular stage of the cell cycle (unpublished observation) are consistent with the hypothesis that the effect of BCL-2 is not as a major regulator of proliferation. We have also found that the high expression of BCL-2 in MCF-7 cells is not caused by rearrangement or amplification of the bcl-2 locus. As the classic t(14;18) chromosomal translocation occurring in follicular lymphomas has not been identified in breast cancer, it is possible that there is more than one mechanism through which the bcl-2 gene can be deregulated. Similarly, high levels of BCL-2 expression have been seen in haematologica] malignancies of both lymphoid and myeloid leukaemias which lack t(14;18). By reducing apoptosis among normal breast epithelial cells, BCL-2 may cause a shift in tissue kinetics towards the preservation of genetically aberrant cells therefore favouring neoplastic development [30]. The second effect of BCL-2 is on cancer cells preventing their death and allowing differentiation. This effect is akin to that observed in the developing breast. In an immunocytochemical study of BCL-2 expression in human foetal and infant breasts, the basal cell layer of the developing breast bud was found to express BCL-2 only during early breast morphogenesis [31]. Since this layer of cells represent the. progeny of the entire ductal and lobular epithelial systems of the adult breast, we suggest that BCL-2 may behave as a physiological

413 regulator of breast morphogenesis by prolonging the life of these cells and expanding the mammary 'stem cell' population. Breast cancers whose cell populations are largely differentiated, still bear a close resemblance to the normal breast phenotype and would therefore express BCL-2. The inverse correlation between the expression of BCL-2 and c-erB-2, an oncogene overexpressed in tumours that have aggressive features [32, 33], supports this view. A co-operation between BCL-2 and the myc oncogene similar to that described in nonmammary cell lines [12, 34, 35] is also possible but difficult to demonstrate using these data. The extent whereby BCL-2 expression in breast cancer can modulate cell death is, however, difficult to demonstrate or quantify. In an immunocytochemical analysis of breast carcinomas of uncharacterised nodal status, Chan and colleagues [36] found a strong negative correlation between BCL-2 expression and the occurrence of apoptosis quantified by calculating the apoptotic indices and also found a strong positive correlation with p53 expression (p - 0.002). The role of p53 tumour suppressor gene in the execution of some forms of apoptosis has also been recently confirmed by others [37, 38]. Moreover, Ohmori et al. [39] using a BCL-2 transfected human lung cancer cell line found that BCL-2 can modulate the cytotoxicity of some anticancer agents by inhibiting apoptosis. In our study the analysis of survival and disease-free survival was made on a heterogenous group of patients and analysis of a much larger cohort of patients who only received CMF would be of interest to show a similar effect of BCL-2. So far, however, no definite relationship has been established between the BCL-2 oncogene product and apoptosis in mammary epithelial cells and the precise role of BCL-2 in breast carcinogenesis and its clinical and therapeutic implications remain to be defined. The present study now needs expansion to a much larger series in order to clarify this interesting but preliminary finding.

Acknowledgements We acknowledge the Ludwig Institute for Cancer Research for initiating the clinical trials and the Swiss Cancer League, the Cancer League of Ticino, the Swedish Cancer League, the Australia-New Zealand Breast Cancer Trials Group, the Australian Cancer Society, the Frontier Science and Technology Research Foundation, and the Swiss Group for Clinical Cancer Research for continuing support of the trials. This work was supported by the Cancer Research Campaign, Leukaemia Research Fund and the Medical Research Council.

References 1. Carter CL, Allen A, Henson DE. Relation of tumor size, lymph node status, and survival in 24,470 breast cancer cases. Cancer 1989; 63:181-9. 2. Fisher B, Bauer M, Wickerham DL et al. Relation of number of positive axillary nodes to the prognosis of patients with primary breast cancer An NSABP update. Cancer 1983; 52: 1551-7. 3. Sarraf CE, Bowen ID. Proportions of mitotic and apoptotic cells in a range of untreated experimental tumours. Cell Tissue Kinet 1988; 21:45-9. 4. Gavrieli Y, Sherman Y, Ben-Sasson SA. Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 1992; 119:493-501. 5. Gorczyca W, Bruno S, Darzynkiewicz RJ et al. DNA strand breaks occurring during apoptosis- their early in situ detection by the terminal deoxynucleotidyl transferase and nick translation assays and prevention by serine protease inhibitors. Int J Oncology 1992; 1:639-48. 6. Bakhshi A, Jensen JP, Goldman P et al. Cloning the chromosomal breakpoint of the t(14;18) human lymphomas: Clustering around JH on chromosome 14 and near a transcriptional unit on 18. Cell 1985; 41:889-906. 7. deary M, Sklar J. Nucleotide sequence of a t(14;18) chromosomal breakpoint in follicular lymphoma and demonstration of a breakpoint cluster-region near a transcriptionally active locus on chromosome 18. Proc Natl Acad Sci USA 1985; 82: 7439-43. 8. Hockenbury D, Nunez G, Milliman C et al. Bcl-2 is an inner mitochondrial membrane protein that blocks programmed cell death. Nature 1990; 348: 334-6. 9. Monaghan P, Robertson D, Amos TAS et al. Ultrastructural localization of bcl-2 protein. J Histochem Cytochem 1992; 40: 1819-25. 10. McDonnell TJ, Deane N, Plan M et al. Bcl-2-immunoglobulin transgenic mice demonstrate extended B cell survival and follicular lymphoproliferation. Cell 1989; 57: 79-88. 11. McDonnell TJ, Korsmeyer SJ. Progression from lymphoid hyperplasia to high-grade malignant lymphoma in mice transgenic for the t(14;18). Nature 1991; 349: 254-6. 12. Vaux DL, Cory S, Adams JM. Bcl-2 gene promotes haemopoietic cell survival and co-operates with c-myc to immortalize pre-B cells. Nature 1988; 335: 440-2. 13. Hockenbury DM, Zutter M, Hickey W et al. Bcl-2 protein is topographically restricted in tissues characterized by apoptotic cell death. Proc Natl Acad Sci USA 1991; 88: 6961-5. 14. McDonnell TJ, Troncoso P, Brisbay SM et al. Expression of the protooncogene bcl-2 in the prostate and its association with emergence of androgen-independent prostate cancer. Cancer Res 1992; 52:6940-4. 15. Ferguson DJP, Anderson TJ. Morphological evaluation of cell turnover in relation to the menstrual cycle in the 'resting' human breast. Br J Cancer 1981; 44: 177-81. 16. Ludwig Breast Cancer Study Group. Randomized trial of chemoendocrine therapy, endocrine therapy, and mastectomy alone in postmenopausal patients with operable breast cancer and axillary node metastases. Lancet 1984; 1: 1256-60. 17. Ludwig Breast Cancer Study Group. A randomized trial of adjuvant combination chemotherapy, with or without prednisolone in premenopausal breast cancer patients with metastases in one to three axillary lymph nodes. Cancer Res 1985; 45: 4454-9. 18. Ludwig Breast Cancer Study Group. Chemotherapy with or without oophorectomy in high-risk pre-menopausal patients with operable breast cancer. J Clinical Oncology 1985; 3: 1059-67. 19. Pezzella F, Tse AGD, Cordell JL et al. Expression of the bcl-2 oncogene protein is not specific for the 14; 18 chromosomal translocation. Am J Pathol 1990; 137: 225-32.

414 20. Nathan B, Anbazhagan R, Dyer M et al. Expression of bcl-2like immunoreactivity in breast cancer. Breast 1993; 2:134-7. 21. Lasfargues EY, Ozzello L. Cultivation of human breast carcinomas. J Natl Cancer Inst 1958; 21:1131-47. 22. Lasfargues EY, Coutinho WG, Redfield ES. Isolation of two human tumour epithelial cell lines from solid breast carcinomas. J Natl Cancer Inst 1974; 61: 967-8. 23. Cailleau R, Young R, Olive M et al. Breast tumour cell lines from pleural effusions. J Nat] Cancer Inst 1974; 52:661-74. 24. Keydar I, Chan L, Karby S et al. Establishment and characterization of a cell line of human breast carcinoma origin. Eur J Cancer 1979; 15: 659-70. 25. Gioanni J, Le Fancois D, Zanghellini E et al. Establishment and characterisation of a new tumorigenic line with a normal karyotype derived from a human breast adenocarcinoma. Br J Cancer 1990; 62: 8-13. 26. Nacheva E, Dyer MJS, Fisher P et al. C-MYC translocations in de novo B-cell lineage acute leukaemias with t(14;18) (cell lines Karpas 231 and 353). Blood 1993; 82: 231-40. 27. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 277: 280-1. 28. Matsudaira P. Sequence from picomole quantities of proteins electroblotted onto polyvinyl difluoride membranes. J Biol Chem 1987; 262: 10035-8. 29. Pezzella F, Turley H, Kuzu I et al. bcl-2 protein in non-smallceU lung carcinoma. N Engl J Med 1993; 329:690-4. 30. Allan DJ, Howell A, Roberts SA et al. Reduction in apoptosis relative to mitosis in histologically normal epithelium accompanies fibrocystic change and carcinoma of the premenopausal human breast. J Pathol 1992; 167: 25-32. 31. Nathan B, Anbazhagan R, Clarkson P et al. Expression of bcl-2 in the human foetal and infant breast. Histopathol 1994; 24:73-6. 32. Gusterson BA, Machin LG, Gullick WJ et al. c-erbB-2 expres-

33. 34. 35. 36.

37. 38. 39.

sion in benign and malignant breast disease. Br J Cancer 1988; 58:453-7. Gusterson BA, Gelber RD, Goldhirsch A et al. Prognostic importance of c-erbB-2 expression in breast cancer. J Clin Oncol 1992; 10:1049-56. Bissonnette RP, Echeverri F, Mahboubi A et al. Apoptotic cell death induced by c-myc is inhibited by bcl-2. Nature 1992; 359: 552-4. Fanidi A, Harrington EA, Evan GI. Co-operative interaction between c-myc and bcl-2 proto-oncogenes. Nature 1993; 359: 554-6. Chan WK, Poulsom QLL, Patel K et al. Bcl-2 expression in invasive mammary carcinoma: Correlation with apoptosis, hormone receptors and p53 expression. Proceedings of 166th Meeting of the Pathological Society of Great Britain and Ireland. J Pathol 1993; 169(Supplement): 153A. Clarke AR, Purdie CA, Harrison DJ et al. Thymocyte apoptosis induced by '53-dependent and independent pathways. Nature 1993; 362: 849-52. Lowe SW, Schmitt EM, Smith S et al. p53 is required for radiation induced apoptosis in mouse thymocytes. Nature 1993; 362: 847-9. Ohmori T, Podack ER, Nishio K et al. Apoptosis of lung cancer cell caused by some anticancer agents (MMC, CPT-11, ADM) is inhibited by BCL-2. Biochem Biophys Res Commun 1993; 192: 30-6.

Received 1 February 1994; accepted 8 March 1994. Correspondence to: Prof. B. Gusterson Institute of Cancer Research Haddow Laboratories 15 Cotswold Road Sutton, Surrey SM2 5NG, U.K.