Breast Cancer Research and Treatment 17: 83-89, 1990. © 1990 Kluwer Academic Publishers. Printed in the Netherlands.
Report
Epidermal growth factor receptors and prognosis in primary breast ea cer Fr6d6rique Spyratos I, Jean-Claude Delarue 5, Catherine Andrieu 1, Rosette Lidereau 2, Marie-H616ne Champ6me 2, Kamel Hac~ne 3, Maurice Brune¢, 4 1Laboratoire des rOcepteurs horrnonaux; 2Laboratoire d'oncovirologie; 3DOpartement de statistiques mOdicales; 4DOpartement de chirurgie," Centre Rend Huguenin, 92211 Saint-Cloud, France," and 5Institut Gustave Roussy, 94800 Villejuif, France
Key words: breast cancer, prognosis, epidermal growth factor receptor Abstract
A retrospective study was performed on 109 human breast tumors stored in liquid nitrogen in order to assess the prognostic value of epidermal growth factor receptor (EGF-R) (median patient follow-up 5 years). A significant inverse relationship was observed between EGF-R and both estrogen (ER) and progesterone receptors (PR). Univariate analysis showed a trend towards a shorter metastasis-free survival both in the overall population and in node-negative patients with EGF-R positive tumors. Multivariate analysis of the overall population showed that lymph-node involvement and PR status were the only significant variables in predicting metastasis-free survival. However, in patients receiving no adjuvant treatment (hormone therapy or chemotherapy), EGF was the only significant variable in the multivariate Cox analysis. No c-erbB-1 amplification was detected in these tumors.
Introduction
Several clinical, morphological, and biological features are commonly used to classify breast tumors. Using histological grade [1, 2] and estrogen and progesterone receptors (ER, PR) [3-7], attempts have been made to identify patients at high risk of early disease recurrence and shortened survival. However, potent prognostic factors independent of lymph node invasiveness are still required in order to target aggressive adjuvant therapy more effectively. Oncogene products [8], FCM-DNA parameters [9], proteases [10], and epidermal growth factor receptors (EGF-R) could constitute the next generation of prognostic factors.
In breast cancer, EGF-R has been found to show an inverse relationship with ER [11, 12] and PR [12] and is associated with higher histologic grade; however, few published studies concern the prognostic value of EGF-R [11, 13, 14]. We have evaluated the prognostic value of EGF-R in 109 primary breast cancers in comparison with nine other factors and in subgroups defined according to adjuvant treatments. In 62 cases the amplification of c-erbB-1 was also studied, since the amino acid sequence of EGF-R shows large homologies with this oncogene [15].
Address for offprints: Fr6d6rique Spyratos, Laboratoire des r6cepteurs hormonaux, Centre Ren6 Huguen~n, 35 rue Dailly, 92211 Saint-Cloud, France
84
F Spyratos et al.
Materials and methods
Patients
The 109 cases included in this retrospective study based on specimens from our liquid nitrogen tumor bank were selected according to the following criteria: 1) primary and unilateral breast cancer, 2) estrogen and progesterone receptors determined on the primary tumor, 3) no other cancer, 4) full treatment and follow-up at the Centre Ren6 Huguenin. Patient data are shown in Table 1. All clinical, histological, and biological information was regularly entered into a computerized data base. Patients have a maximum follow up of 8 years (median 5 years). All patients were staged according to the UICC criteria [16]. Histologic grade was determined according to Scarff, Bloom, and Richardson [17]. Thirty-four patients (31%) regarded as high risk (on classical criteria) received postoperative adjuvant therapies (chemotherapy and/ or hormone therapy). Tumor samples were frozen immediately after surgery and kept in liquid nitrogen. ER and PR assays were performed according to the EORTC recommendations [18], with a cut-off set at 10 fm/ mg protein. FCM-DNA analyses were performed after mechanical dissociation of the samples [19].
EGF receptor assay
EGF-R assays were performed on specimens remaining after steroid receptor assays. After pulverization, samples were homogenized in 10mM TRIS-HCI buffer, pH 7.4 containing 1 mM EDTA, 0.5 mM DTT, and 10% glycerol. In 62 of the 109 cases, the homogenate was aliquoted in two parts and centrifuged at 100,000g for one hour. One pellet was kept at - 2 0 ° C in 10 mM TRIS-HC1 buffer, pH 8, containing 10 mM NaC1 and 10 mM EDTA, for c-erbB-i analysis by Southern blot [20]. The other pellet was kept in liquid nitrogen in 0.025 M phosphate buffer pH 7.4, containing NaC1 (0.15M), MgC1z (0.002M), and 10% glycerol for EGF-R assays using Scatchard plots according to the method previously described [12]. In the re-
maining 47 cases, only one pellet was prepared and used for EGF-R assay. Briefly, after homogenization by Turrax for 5 sec at 0°C and by glass-glass Potter homogenizer (30 sec), the homogenate was centrifuged for 5 min at 800 g and the supernatant was used for the assay. Fifty/zl containing 100200/xg of membrane protein were incubated for 2 hours at 25°C with increasing quantities (60240 fm) of 125I-EGF in the presence or absence of a large excess of unlabelled EGF (200/zl final volume). The reaction was stopped by the addition of 0.5 ml of chilled phosphate buffer containing i rag/ ml of BSA. After 5min of centrifugation at 15,000 g, the radioactivity of the pellet was counted. The results calculated according to the ScatchTable 1. Characteristics of the overall population according to EGF-R status
Age (years) 64 UICC stage 1 2 3 Macroscopic tumor size - 10 fm/mg) Pre-menopausal Post-menopausal Total
EGF-R
EGF-R+
30 48 31
21(8) 30 (13) 21 ( 6 )
9(4) 18 ( 9 ) 10 (4)
4 75 30
1(0) 48 (18) 23 (9)
3(1) 27 (11) 7 (5)
55 37 17
37 (14) 26 ( 9 ) 9 (4)
18 (7) 11 ( 5 ) 8 (5)
4 61 44
3(1) 44 (16) 25 (10)
1(0) 17 ( 8 ) 19 (9)
55 54
34 ( 7 ) 38 (20)
21 ( 8 ) 16 (9)
31 78 42 67 67 42 43 66 109
22 (8) 49 (19) 20(9) 52 (18) 38 (16) 34 (11) 31 (12) 41 (15) 72 (27)
9 (5) 29 (12) 22(9) 15 ( 8 ) 29 (14) 8 (3) 12 ( 7 ) 25 (10) 37 (17)
*
Chi-square p values: * p = 0.002, ** p = 0.02. Numbers in parentheses: numbers of relapses
Epidermal growth factor receptor in breast cancer ard method were expressed as femtomoles (fro) EGF fixed/mg protein. A tumor was considered positive for EGF-R (EGF-R +) if it contained more than 10 fm/mg of membrane protein. The amount of membrane protein obtained was between 1 and 2 mg. 125I-EGF was prepared by labeling EGF (Sigma) with 125Iby the iodogen method [21] (specific activity 100-150/zCi//xg).
DNA extraction and Southern hybridization Genomic DNAs were extracted [22] and 10/zg of each were digested with the Eco RI restriction enzyme. The resulting fragments were fractionated in 0.8% agarose gels and transferred to Genatran 45 (Plasco, Woburn, MA) nylon membranes. The membranes were baked at 80°C for 2-3 h, prehybridized, and hybridized with 32p-labeled D N A probes by standard procedures. Probes were labeled with 32p-dCTP to a specific activity of 1-2 x 109 cpm//zg using the random primers system [23]. After hybridization, filters were washed under stringent conditions (15mM sodium chloride, 1.5mM sodium citrate, pH7, 65°C, 30min), exposed for 1-4 days to X-ray films, and the results quantitated by scanning densitometry.
DNA probes The c-erbB-1 cDNA cloned by GT Merlino (plasmid pE7) contained a 2.4 kilobase (kb) pair sequence encoding 2/3 of the E G F receptor, provided from the ATCC (number 57346). The c-erbB-1 probe recognized a 8.7 kb fragment as published elsewhere [24, 25]. The 2.5kb pair Eco RI fragment subcloned in pBR322 and containing all vmos-specific sequences has been described by Watson et al. [26]. This probe recognizes a frequent Eco RI 2.5 kb allele and rarely a 5 kb Eco RI fragment [27]. The amount of D N A transferred on Eco RI blots was standardized by hybridization with a cmos probe.
85
Statistical methods Differences in the distribution of patient characteristics between EGF-R ÷ and EGF-R- subgroups were analyzed using Chi-square tests. Actuarial metastasis-free survival (MFS) rates were compared using the Log-Rank test [28]. Multivariate analyses based on the Cox proportional hazards model [29] were performed to identify in a stepwise manner variables associated with metastasis-free survival. The variables included in the Cox model are listed in Table 1. Forty-four patients presented with metastases during the follow-up period.
Results
EGF-R concentration and other prognostic factors Characteristics of the patient population according to EGF-R status are presented in Table 1. EGF-R was detected in 37 of the 109 samples (34%) at a mean concentration of 47fm/mg protein and a mean Kd of 4.6riM. An example of a Scatchard plot is shown in Fig. 1. Only one type of binding site was observed. There was a strong inverse relationship between EGF-R and both E R (p = 0.002) and PR (p = 0.02). Histologic grade III tumors were more frequently EGF-R + (43%) than grade II (28%) tumors, although the difference was not significant. There were only four grade I tumors in our population. No significant link was observed between EGF-R status and other factors, including FCM-DNA content.
Prognostic value of EGF-R in the overall population Univariate analysis showed that patients with EGF-R + tumors had a worse prognosis than patients with EGF-R- tumors, but the difference was not significant (relative risk = 1.38; p = 0.29). Analyses were also performed in subgroups defined by the factors listed in Table 1. In nodenegative patients, a shorter disease-free survival was observed in those with EGF-R + tumors, but
86
F Spyratos et al. 8/F
T I
Kd=3.1nM
0.02
2's
s'o7 ~
~m/m!
Fig. t. Scatchard plot of EGF binding data from tumor membrane with a single class of binding sites.
the difference was not significant (relative risk 2.2; p = 0.14). Multivariate analysis (Table 2) indicated that nodal status and PR were the only significant variables in predicting metastasis-free survival. EGF-R status was not identified as a prognostic factor.
Prognostic value of EGF-R according to adjuvant treatments in the 75 patients (mainly node-negative), receiving no adjuvant treatment (hormone therapy or chemotherapy), EGF-R status was a significant
prognostic factor by univariate analysis (p -- 0.05). EGF-R status was the only significant factor in the multivariate Cox analysis (Table 3).
c-erb B-1 amplification No amplification of c-erbB-1 was observed in the 62 tumors for which sufficient material was available for testing, as shown in Fig. 2 which represents 14 different specimens. This variable was therefore not included in the prognostic analyses.
~2~bie 2. Mukivariate analysis of metastasis-free survival in 109 patients Significant variables
Regression coefficient 13
Standard error o
13/o
p value
Nodal status Progesterone receptors
0.83 - 0.57
0.273 0.283
3.05 2.02
0.01 0.04
Variables included in the Cox model were: age, menopausal status, UICC stage, macroscopic tumor size, histologic grade, nodal status, ER and PR status, EGF-R status and FCM-DNA content.
Epidermal growth factor receptor in breast cancer
87
Fig. 2. Southern blot hybridized with c-erbB-1 and c-mos probes. 14 different tumor DNAs were separated on agarose gel and hybridized after transfer with c-erbB-1 and c-mos probes. The size of each fragment is indicated on the left. 8.7kb corresponds to c-erbB-1, 2.5 and 5.0 kb to c-mos alleles (the 5.0 kb is a rare c-mos allele); the c-mos is necessary to evaluate the amount of DNAs for each specimen. We do not detect any amplification of c-erbB-1; the variation in intensity is due only to varying quantities of DNAs as verified by c-mos allele intensity.
Discussion E p i d e r m a l g r o w t h factor receptors w e r e f o u n d in 34% o f o u r patients with p r i m a r y breast cancer. This is in the range described in the literature, i.e. 3 0 - 5 0 % of p r i m a r y breast cancer [11-14, 23-25]. T h e strong inverse relationship f o u n d in previous studies b e t w e e n E G F - R and b o t h E R [11, 12, 3 0 -
32] and P R [12] was confirmed here. H o w e v e r , the relationship b e t w e e n E G F - R and higher histologic grade was non-significant, p r o b a b l y because of the small n u m b e r of grade I tumors. O u r results are in g o o d a g r e e m e n t with the a b o v e studies with regard to the m e a n K d and the p e r c e n t a g e of E G F - R positive tumors. T h e results of o u r first screening study [33],
Table 3. Multivariate analysis of metastasis-free survival in the 75 patients receiving no adjuvant treatment Significant variables
Regression coefficient 13
Standard error o
fl/o
p value
EGF-R
1.12
0.476
2.35
0.03
Variables included in the Cox model were: age, menopausal status, UICC stage, macroscopic tumor size, histologic grade, nodal status, ER and PR status, EGF-R status and FCM-DNA content.
88
F Spyratos et al.
which detected 2% of c-erbB-1 gene alteration in human primary breast tumors, were confirmed here. Increased EGF-R expression does not therefore reflect c-erbB-1 amplification, suggesting instead a probable dysregulation of the EGF-R gene. Retrospective studies based on specimens from tumor banks may be biased in two ways. First, large tumors may be predominantly selected; however, the main patient characteristics in the present study correspond well with the distribution generally described in primary breast cancer. Second, the EGF receptor may be altered during storage; however, our results for EGF-R distribution are similar to those classically described in prospective studies. An increased risk of recurrence in EGF-R positive tumors was observed, particularly in nodenegative patients, but neither univariate nor multivariate analysis demonstrated an independent prognostic value of EGF-R in the overall population. The apparent discrepancy between our results and those of Sainsbury [11] and Costa [13] could be partly explained by differences in patient management. In effect, no adjuvant chemotherapy or hormone therapy was administred to the patients in the study of Sainsbury [11], while 31% of our patients did receive adjuvant therapies. When we excluded patients receiving adjuvant treatments, EGF-R was the only significant prognostic factor in the multivariate analysis. Recent studies have shown the usefulness of EGF-R status as a predictor of response to hormone therapy in addition to hormone receptor status [34, 35]. Recently, Grirnaux et al. showed that the presence of EGF-R enabled subgroups of ER-negative patients with a particularly bad prognosis to be identified [36]. Hence, the prognostic value of EGF-R could be modulated according to adjuvant treatments as is probably the case for steroid receptors [6]. As previously noted [37], differences in the cut-off values and frequency distributions of EGF-R exist in the literature, and standardization is required before direct comparison can be made between different studies. If the clinical importance of EGF-R is in fact confirmed, a more reproducible assay with lower cost and requiring smaller specimens will be necessary.
Acknowledgements This work was supported by La Ligue Nationale Fran~aise Contre le Cancer et ses comit6s r6gionaux des Hauts-de-Seine et des Yvelines.
References 1. Contesso G, Mouriesse H, Friedman S, Genin J, Sarrazin D, RouEss6 J: The importance of histologic grade in long term prognosis of breast cancer. A study of 1010 patients uniformly treated at the Institut Gustave Roussy. J Clin Oncol 5: 1378-1386, 1987 2. Le Doussal V, Tubiana-Hulin M, Friedman S, Hac6ne K, Spyratos F, Brunet M: Prognostic value of nuclear components of the histologic grade of Scarff-Bloom-Richardson (SBR). An improved score modification based on a multivariate analysis of 1262 invasive ductal carcinomas. Cancer, 64: 1914--1921, 1989 3. Knight WA, Livingstone RB, Gregory EJ, McGuire WL: Estrogen receptors as an independent prognostic factor for early recurrence in breast cancer. Cancer Res 37: 46694671, 1977 4. Pichon MF, Pallud C, Brunet M, Milgrom E: Relationship of presence of progesterone receptors to prognosis in early breast cancer. Cancer Res 40: 3357-3360, 1980 5. McGuire WL, Clark GM, Dressier LG, Owens M: Role of steroid hormone receptors as prognostic factors in primary breast cancer. NCI Monogr 1: 19-23, 1986 6. Spyratos F, Hac6ne K, Tubiana-Hulin M, Pallud C, Brunet M: Prognostic value of estrogen and progesterone receptors in primary infiltrating ductal breast cancer. A sequential multivariate analysis of 1262 patients. Eur J Cancer Clin Oncol 8: 1233-1240, 1989 7. May E, Mouriesse H, May-Levin F, Contesso G, Delarue JC: A new approach allowing an early prognosis in breast cancer. The ratio of estrogen receptor (ER) ligand binding activity to the ER-specific mRNA level. Oncogene 4: 10371042, 1989 8. Callahan R: Genetic alterations in primary breast cancer. Breast Cancer Res Treat 13: 191-203, 1989 9. McGuire WL, Meyer JS, Barlogie B, Kute TE: Impact of flow cytometry on predicting recurrence and survival in breast cancer patients. Breast Cancer Res Treat 5: 117-128, 1985 10. Spyratos F, Maudelonde T, Brouillet JP, Brunet M, Defrenne A, Andrieu C, Ha6ene K, Desplaces A, Rou~ss6 J, Rochefort H: Cathepsin D: an important marker predicting metastasis in primary breast cancer. Lancet ii: 1115-1118, 1989 11. Sainsbury JRC, Malcolm A, Appleton D, Farndon JR, Harris AL: Presence of epidermal growth factor receptors
Epidermal growth factor receptor in breast cancer
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
as an indicator of poor prognosis in patients with breast cancer. J Clin Pathol 38: 1225-1228, 1985 Delarue JC, Friedman S, Mouriesse H, May-Levin F, Sancho-Garnier H, Contesso G: Epidermal growth factor receptor in human breast cancers: correlation with estrogen and progesterone receptors. Breast Cancer Res Treat 11: 173-178, 1988 Costa S, Stamm H, Almendral A, Ludwig H, Wyss R, Fabbro D, Ernst A, Takahashi A, Eppenberger U: Predictive value of EGF receptor in breast cancer. Lancet ii: 1258, 1988 Rios MA, Macias A, Perez R, Lage A, Skoog L: Receptors for epidermal growth factor and estrogen as predictors of relapse in patients with mammary carcinoma. Anticancer Res 8: 173-176, 1988 Downward J, Yarden Y, Mayes E, et al.: Close similarity of epidermal growth factor receptor and v-erbB oncogene protein sequence. Nature 307: 521, 1984 UICC (International Union Against Cancer): TNM Classification of malignant tumors, 3rd edn. Geneva, UICC, pp 47-54, 1979 Bloom I-IJC, Richardson WW: Histological grading and prognosis in breast cancer. A study on 1409 cases of which 359 have been followed for 15 years. Br J Cancer 11: 35-77, 1957 EORTC Breast Cooperative Group revision of the standards for the assessment of hormone receptor in human breast cancer. Report of the second EORTC workshop, held on 16-17 March, 1979, in the Netherlands Cancer Institute. Eur J Cancer 16: 1513-1515, 1980 Spyratos F, Briffod M, Gentile A, Brunet M, Brault C, Desplaces A: Flow cytometric study of DNA cytopunctures of benign and malignant breast lesions. Anal Quant Cytol Histol 9: 485-486, 1987 Southern EM: Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98: 503-517, 1975 Fraker PJ, Speck JC: Protein and cell membrane iodination with a sparingly soluble chloramide, 1, 3, 4, 6-tetrachloro-3a,6a-diphenyl glycoluril. Biochem Biophys Res Corn° mun 80: 849-857, 1978 Gross-Bellard M, Oudet P, Chambon P: Isolation of high molecular weight DNA from mammalian cells. Eur J Biochem 36: 32-38, 1973 Feinberg AP, Vogelstein B: A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal Biochem 131: 6-13, 1983 Ullrich A, Coussens L, Hayflick JS, Dull TLJ, Gray A, Tam AW, Lee J, Yarden Y, Libermann TA, Schlessinger J, Downward J, Mayes EMV, Whittle N, Waterfield MD, Seeburg PH: Human epidermal growth factor receptor cDNA sequence and aberrant expression of the amplified gene in A431 epidermoid carcinoma cells. Nature 309: 418425, 1984 Merlino GT, Sunsuke I, Whang-Peng J, Knutsen T, Xu
26.
27.
28.
29. 30.
31.
32.
33.
34.
35.
36.
37.
89
YH, Clarck AJL, Stratton RH, Wilson RK, Ma DP, Roe BA, Hunts JH, Shimizu N, Pastan I: Structure and localization of genes encoding aberrant and normal epidermal growth factor receptor RNAs from A431 human carcinoma cells. Mol Cell Biol 5: 1722-1734, 1985 Watson R, Oskarsson M, Vande Woude GF: Human DNA sequence homologous to the transforming gene (mos) of Moloney sarcoma virus. Proc Natl Acad Sci USA 79: 40784082, 1982 Lidereau R, Mathieu-Mahul D, Theillet C, Renaud M, Mauchauffe M, Gest J, Larsen CJ: Presence of an allelic Eco RI restriction fragment of the c-mos locus in leukocyte and tumor cell DNAs of breast cancer patients. Proc Natl Acad Sci USA 82: 7068--7070, 1985 Peto R, Pike MC, Armitage Pet al.: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 35: 1-39, 1977 Cox DR: Regression models and life tables. J R Stat Soc 34: 187-220, 1972 Perez R, Pascual M, Macias A, Loge A: Epidermal growth factor receptor in human breast cancer. Breast Cancer Res Treat 4: 189-193, 1984 Sainsbury JRC, Farndon JR, Sherbet GV, Harris AL: Epidermal growth factor receptors and estrogen receptors in human breast cancer. Lancet i: 364-366, 1985 Battaglia F, Scambia G, Rossi S, Panici PB, Bellantone R, Polizzi G, Querzoli P, Negrini R, Iacobelli S, Crucitti F, Mancuso S: Epidermal growth fractor receptor in human breast cancer: correlation with steroid hormone receptors and axillary lymph node involvement. Eur J Cancer Clin Oncol 24: 1685-1690, 1988 Ali IU, Lidereau R, Callahan R: Heterogeneity of genetic alterations in primary human breast tumors. Lippman ME, Dickson RB (eds) Breast Cancer: Cellular and Molecular Biology. Kluwer, Boston, 1988, pp 25-48 Nicholson S, Sainsbury JRC, Holerow P, Chambers P, Farndon JR, Harris AL: Expression of epidermal growth factor receptors associated with lack of response to endocrine therapy in recurrent breast cancer. Lancet i: 182184, 1989 Nicholson S, Halcrow P, Sainsbury JRC, Angus B, Chambers P, Farndon JR, Harris AL: Epidermal growth factor receptor (EGFr) status associated with failure of primary endocrine therapy in elderly postmenopausal patients with breast cancer. Br J Cancer 58: 810-814, 1988 Grimaux M, Romain S, Remvikos Y, Martin PM, Magdel6nat H: Prognostic value of epidermal growth factor receptor in node-positive breast cancer. Breast Cancer Res Treat 14: 77-90, 1989 Nicholson S, Sainsbury JRC, Needham GK, Chambers P, Farndon JR, Harris AL: Quantitative assays of epidermal growth factor receptor in human breast cancer: cut-off points of clinical relevance. Int J Cancer 42: 36--41, 1988
Report
Epidermal growth factor receptors and prognosis in primary breast ea cer Fr6d6rique Spyratos I, Jean-Claude Delarue 5, Catherine Andrieu 1, Rosette Lidereau 2, Marie-H616ne Champ6me 2, Kamel Hac~ne 3, Maurice Brune¢, 4 1Laboratoire des rOcepteurs horrnonaux; 2Laboratoire d'oncovirologie; 3DOpartement de statistiques mOdicales; 4DOpartement de chirurgie," Centre Rend Huguenin, 92211 Saint-Cloud, France," and 5Institut Gustave Roussy, 94800 Villejuif, France
Key words: breast cancer, prognosis, epidermal growth factor receptor Abstract
A retrospective study was performed on 109 human breast tumors stored in liquid nitrogen in order to assess the prognostic value of epidermal growth factor receptor (EGF-R) (median patient follow-up 5 years). A significant inverse relationship was observed between EGF-R and both estrogen (ER) and progesterone receptors (PR). Univariate analysis showed a trend towards a shorter metastasis-free survival both in the overall population and in node-negative patients with EGF-R positive tumors. Multivariate analysis of the overall population showed that lymph-node involvement and PR status were the only significant variables in predicting metastasis-free survival. However, in patients receiving no adjuvant treatment (hormone therapy or chemotherapy), EGF was the only significant variable in the multivariate Cox analysis. No c-erbB-1 amplification was detected in these tumors.
Introduction
Several clinical, morphological, and biological features are commonly used to classify breast tumors. Using histological grade [1, 2] and estrogen and progesterone receptors (ER, PR) [3-7], attempts have been made to identify patients at high risk of early disease recurrence and shortened survival. However, potent prognostic factors independent of lymph node invasiveness are still required in order to target aggressive adjuvant therapy more effectively. Oncogene products [8], FCM-DNA parameters [9], proteases [10], and epidermal growth factor receptors (EGF-R) could constitute the next generation of prognostic factors.
In breast cancer, EGF-R has been found to show an inverse relationship with ER [11, 12] and PR [12] and is associated with higher histologic grade; however, few published studies concern the prognostic value of EGF-R [11, 13, 14]. We have evaluated the prognostic value of EGF-R in 109 primary breast cancers in comparison with nine other factors and in subgroups defined according to adjuvant treatments. In 62 cases the amplification of c-erbB-1 was also studied, since the amino acid sequence of EGF-R shows large homologies with this oncogene [15].
Address for offprints: Fr6d6rique Spyratos, Laboratoire des r6cepteurs hormonaux, Centre Ren6 Huguen~n, 35 rue Dailly, 92211 Saint-Cloud, France
84
F Spyratos et al.
Materials and methods
Patients
The 109 cases included in this retrospective study based on specimens from our liquid nitrogen tumor bank were selected according to the following criteria: 1) primary and unilateral breast cancer, 2) estrogen and progesterone receptors determined on the primary tumor, 3) no other cancer, 4) full treatment and follow-up at the Centre Ren6 Huguenin. Patient data are shown in Table 1. All clinical, histological, and biological information was regularly entered into a computerized data base. Patients have a maximum follow up of 8 years (median 5 years). All patients were staged according to the UICC criteria [16]. Histologic grade was determined according to Scarff, Bloom, and Richardson [17]. Thirty-four patients (31%) regarded as high risk (on classical criteria) received postoperative adjuvant therapies (chemotherapy and/ or hormone therapy). Tumor samples were frozen immediately after surgery and kept in liquid nitrogen. ER and PR assays were performed according to the EORTC recommendations [18], with a cut-off set at 10 fm/ mg protein. FCM-DNA analyses were performed after mechanical dissociation of the samples [19].
EGF receptor assay
EGF-R assays were performed on specimens remaining after steroid receptor assays. After pulverization, samples were homogenized in 10mM TRIS-HCI buffer, pH 7.4 containing 1 mM EDTA, 0.5 mM DTT, and 10% glycerol. In 62 of the 109 cases, the homogenate was aliquoted in two parts and centrifuged at 100,000g for one hour. One pellet was kept at - 2 0 ° C in 10 mM TRIS-HC1 buffer, pH 8, containing 10 mM NaC1 and 10 mM EDTA, for c-erbB-i analysis by Southern blot [20]. The other pellet was kept in liquid nitrogen in 0.025 M phosphate buffer pH 7.4, containing NaC1 (0.15M), MgC1z (0.002M), and 10% glycerol for EGF-R assays using Scatchard plots according to the method previously described [12]. In the re-
maining 47 cases, only one pellet was prepared and used for EGF-R assay. Briefly, after homogenization by Turrax for 5 sec at 0°C and by glass-glass Potter homogenizer (30 sec), the homogenate was centrifuged for 5 min at 800 g and the supernatant was used for the assay. Fifty/zl containing 100200/xg of membrane protein were incubated for 2 hours at 25°C with increasing quantities (60240 fm) of 125I-EGF in the presence or absence of a large excess of unlabelled EGF (200/zl final volume). The reaction was stopped by the addition of 0.5 ml of chilled phosphate buffer containing i rag/ ml of BSA. After 5min of centrifugation at 15,000 g, the radioactivity of the pellet was counted. The results calculated according to the ScatchTable 1. Characteristics of the overall population according to EGF-R status
Age (years) 64 UICC stage 1 2 3 Macroscopic tumor size - 10 fm/mg) Pre-menopausal Post-menopausal Total
EGF-R
EGF-R+
30 48 31
21(8) 30 (13) 21 ( 6 )
9(4) 18 ( 9 ) 10 (4)
4 75 30
1(0) 48 (18) 23 (9)
3(1) 27 (11) 7 (5)
55 37 17
37 (14) 26 ( 9 ) 9 (4)
18 (7) 11 ( 5 ) 8 (5)
4 61 44
3(1) 44 (16) 25 (10)
1(0) 17 ( 8 ) 19 (9)
55 54
34 ( 7 ) 38 (20)
21 ( 8 ) 16 (9)
31 78 42 67 67 42 43 66 109
22 (8) 49 (19) 20(9) 52 (18) 38 (16) 34 (11) 31 (12) 41 (15) 72 (27)
9 (5) 29 (12) 22(9) 15 ( 8 ) 29 (14) 8 (3) 12 ( 7 ) 25 (10) 37 (17)
*
Chi-square p values: * p = 0.002, ** p = 0.02. Numbers in parentheses: numbers of relapses
Epidermal growth factor receptor in breast cancer ard method were expressed as femtomoles (fro) EGF fixed/mg protein. A tumor was considered positive for EGF-R (EGF-R +) if it contained more than 10 fm/mg of membrane protein. The amount of membrane protein obtained was between 1 and 2 mg. 125I-EGF was prepared by labeling EGF (Sigma) with 125Iby the iodogen method [21] (specific activity 100-150/zCi//xg).
DNA extraction and Southern hybridization Genomic DNAs were extracted [22] and 10/zg of each were digested with the Eco RI restriction enzyme. The resulting fragments were fractionated in 0.8% agarose gels and transferred to Genatran 45 (Plasco, Woburn, MA) nylon membranes. The membranes were baked at 80°C for 2-3 h, prehybridized, and hybridized with 32p-labeled D N A probes by standard procedures. Probes were labeled with 32p-dCTP to a specific activity of 1-2 x 109 cpm//zg using the random primers system [23]. After hybridization, filters were washed under stringent conditions (15mM sodium chloride, 1.5mM sodium citrate, pH7, 65°C, 30min), exposed for 1-4 days to X-ray films, and the results quantitated by scanning densitometry.
DNA probes The c-erbB-1 cDNA cloned by GT Merlino (plasmid pE7) contained a 2.4 kilobase (kb) pair sequence encoding 2/3 of the E G F receptor, provided from the ATCC (number 57346). The c-erbB-1 probe recognized a 8.7 kb fragment as published elsewhere [24, 25]. The 2.5kb pair Eco RI fragment subcloned in pBR322 and containing all vmos-specific sequences has been described by Watson et al. [26]. This probe recognizes a frequent Eco RI 2.5 kb allele and rarely a 5 kb Eco RI fragment [27]. The amount of D N A transferred on Eco RI blots was standardized by hybridization with a cmos probe.
85
Statistical methods Differences in the distribution of patient characteristics between EGF-R ÷ and EGF-R- subgroups were analyzed using Chi-square tests. Actuarial metastasis-free survival (MFS) rates were compared using the Log-Rank test [28]. Multivariate analyses based on the Cox proportional hazards model [29] were performed to identify in a stepwise manner variables associated with metastasis-free survival. The variables included in the Cox model are listed in Table 1. Forty-four patients presented with metastases during the follow-up period.
Results
EGF-R concentration and other prognostic factors Characteristics of the patient population according to EGF-R status are presented in Table 1. EGF-R was detected in 37 of the 109 samples (34%) at a mean concentration of 47fm/mg protein and a mean Kd of 4.6riM. An example of a Scatchard plot is shown in Fig. 1. Only one type of binding site was observed. There was a strong inverse relationship between EGF-R and both E R (p = 0.002) and PR (p = 0.02). Histologic grade III tumors were more frequently EGF-R + (43%) than grade II (28%) tumors, although the difference was not significant. There were only four grade I tumors in our population. No significant link was observed between EGF-R status and other factors, including FCM-DNA content.
Prognostic value of EGF-R in the overall population Univariate analysis showed that patients with EGF-R + tumors had a worse prognosis than patients with EGF-R- tumors, but the difference was not significant (relative risk = 1.38; p = 0.29). Analyses were also performed in subgroups defined by the factors listed in Table 1. In nodenegative patients, a shorter disease-free survival was observed in those with EGF-R + tumors, but
86
F Spyratos et al. 8/F
T I
Kd=3.1nM
0.02
2's
s'o7 ~
~m/m!
Fig. t. Scatchard plot of EGF binding data from tumor membrane with a single class of binding sites.
the difference was not significant (relative risk 2.2; p = 0.14). Multivariate analysis (Table 2) indicated that nodal status and PR were the only significant variables in predicting metastasis-free survival. EGF-R status was not identified as a prognostic factor.
Prognostic value of EGF-R according to adjuvant treatments in the 75 patients (mainly node-negative), receiving no adjuvant treatment (hormone therapy or chemotherapy), EGF-R status was a significant
prognostic factor by univariate analysis (p -- 0.05). EGF-R status was the only significant factor in the multivariate Cox analysis (Table 3).
c-erb B-1 amplification No amplification of c-erbB-1 was observed in the 62 tumors for which sufficient material was available for testing, as shown in Fig. 2 which represents 14 different specimens. This variable was therefore not included in the prognostic analyses.
~2~bie 2. Mukivariate analysis of metastasis-free survival in 109 patients Significant variables
Regression coefficient 13
Standard error o
13/o
p value
Nodal status Progesterone receptors
0.83 - 0.57
0.273 0.283
3.05 2.02
0.01 0.04
Variables included in the Cox model were: age, menopausal status, UICC stage, macroscopic tumor size, histologic grade, nodal status, ER and PR status, EGF-R status and FCM-DNA content.
Epidermal growth factor receptor in breast cancer
87
Fig. 2. Southern blot hybridized with c-erbB-1 and c-mos probes. 14 different tumor DNAs were separated on agarose gel and hybridized after transfer with c-erbB-1 and c-mos probes. The size of each fragment is indicated on the left. 8.7kb corresponds to c-erbB-1, 2.5 and 5.0 kb to c-mos alleles (the 5.0 kb is a rare c-mos allele); the c-mos is necessary to evaluate the amount of DNAs for each specimen. We do not detect any amplification of c-erbB-1; the variation in intensity is due only to varying quantities of DNAs as verified by c-mos allele intensity.
Discussion E p i d e r m a l g r o w t h factor receptors w e r e f o u n d in 34% o f o u r patients with p r i m a r y breast cancer. This is in the range described in the literature, i.e. 3 0 - 5 0 % of p r i m a r y breast cancer [11-14, 23-25]. T h e strong inverse relationship f o u n d in previous studies b e t w e e n E G F - R and b o t h E R [11, 12, 3 0 -
32] and P R [12] was confirmed here. H o w e v e r , the relationship b e t w e e n E G F - R and higher histologic grade was non-significant, p r o b a b l y because of the small n u m b e r of grade I tumors. O u r results are in g o o d a g r e e m e n t with the a b o v e studies with regard to the m e a n K d and the p e r c e n t a g e of E G F - R positive tumors. T h e results of o u r first screening study [33],
Table 3. Multivariate analysis of metastasis-free survival in the 75 patients receiving no adjuvant treatment Significant variables
Regression coefficient 13
Standard error o
fl/o
p value
EGF-R
1.12
0.476
2.35
0.03
Variables included in the Cox model were: age, menopausal status, UICC stage, macroscopic tumor size, histologic grade, nodal status, ER and PR status, EGF-R status and FCM-DNA content.
88
F Spyratos et al.
which detected 2% of c-erbB-1 gene alteration in human primary breast tumors, were confirmed here. Increased EGF-R expression does not therefore reflect c-erbB-1 amplification, suggesting instead a probable dysregulation of the EGF-R gene. Retrospective studies based on specimens from tumor banks may be biased in two ways. First, large tumors may be predominantly selected; however, the main patient characteristics in the present study correspond well with the distribution generally described in primary breast cancer. Second, the EGF receptor may be altered during storage; however, our results for EGF-R distribution are similar to those classically described in prospective studies. An increased risk of recurrence in EGF-R positive tumors was observed, particularly in nodenegative patients, but neither univariate nor multivariate analysis demonstrated an independent prognostic value of EGF-R in the overall population. The apparent discrepancy between our results and those of Sainsbury [11] and Costa [13] could be partly explained by differences in patient management. In effect, no adjuvant chemotherapy or hormone therapy was administred to the patients in the study of Sainsbury [11], while 31% of our patients did receive adjuvant therapies. When we excluded patients receiving adjuvant treatments, EGF-R was the only significant prognostic factor in the multivariate analysis. Recent studies have shown the usefulness of EGF-R status as a predictor of response to hormone therapy in addition to hormone receptor status [34, 35]. Recently, Grirnaux et al. showed that the presence of EGF-R enabled subgroups of ER-negative patients with a particularly bad prognosis to be identified [36]. Hence, the prognostic value of EGF-R could be modulated according to adjuvant treatments as is probably the case for steroid receptors [6]. As previously noted [37], differences in the cut-off values and frequency distributions of EGF-R exist in the literature, and standardization is required before direct comparison can be made between different studies. If the clinical importance of EGF-R is in fact confirmed, a more reproducible assay with lower cost and requiring smaller specimens will be necessary.
Acknowledgements This work was supported by La Ligue Nationale Fran~aise Contre le Cancer et ses comit6s r6gionaux des Hauts-de-Seine et des Yvelines.
References 1. Contesso G, Mouriesse H, Friedman S, Genin J, Sarrazin D, RouEss6 J: The importance of histologic grade in long term prognosis of breast cancer. A study of 1010 patients uniformly treated at the Institut Gustave Roussy. J Clin Oncol 5: 1378-1386, 1987 2. Le Doussal V, Tubiana-Hulin M, Friedman S, Hac6ne K, Spyratos F, Brunet M: Prognostic value of nuclear components of the histologic grade of Scarff-Bloom-Richardson (SBR). An improved score modification based on a multivariate analysis of 1262 invasive ductal carcinomas. Cancer, 64: 1914--1921, 1989 3. Knight WA, Livingstone RB, Gregory EJ, McGuire WL: Estrogen receptors as an independent prognostic factor for early recurrence in breast cancer. Cancer Res 37: 46694671, 1977 4. Pichon MF, Pallud C, Brunet M, Milgrom E: Relationship of presence of progesterone receptors to prognosis in early breast cancer. Cancer Res 40: 3357-3360, 1980 5. McGuire WL, Clark GM, Dressier LG, Owens M: Role of steroid hormone receptors as prognostic factors in primary breast cancer. NCI Monogr 1: 19-23, 1986 6. Spyratos F, Hac6ne K, Tubiana-Hulin M, Pallud C, Brunet M: Prognostic value of estrogen and progesterone receptors in primary infiltrating ductal breast cancer. A sequential multivariate analysis of 1262 patients. Eur J Cancer Clin Oncol 8: 1233-1240, 1989 7. May E, Mouriesse H, May-Levin F, Contesso G, Delarue JC: A new approach allowing an early prognosis in breast cancer. The ratio of estrogen receptor (ER) ligand binding activity to the ER-specific mRNA level. Oncogene 4: 10371042, 1989 8. Callahan R: Genetic alterations in primary breast cancer. Breast Cancer Res Treat 13: 191-203, 1989 9. McGuire WL, Meyer JS, Barlogie B, Kute TE: Impact of flow cytometry on predicting recurrence and survival in breast cancer patients. Breast Cancer Res Treat 5: 117-128, 1985 10. Spyratos F, Maudelonde T, Brouillet JP, Brunet M, Defrenne A, Andrieu C, Ha6ene K, Desplaces A, Rou~ss6 J, Rochefort H: Cathepsin D: an important marker predicting metastasis in primary breast cancer. Lancet ii: 1115-1118, 1989 11. Sainsbury JRC, Malcolm A, Appleton D, Farndon JR, Harris AL: Presence of epidermal growth factor receptors
Epidermal growth factor receptor in breast cancer
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
as an indicator of poor prognosis in patients with breast cancer. J Clin Pathol 38: 1225-1228, 1985 Delarue JC, Friedman S, Mouriesse H, May-Levin F, Sancho-Garnier H, Contesso G: Epidermal growth factor receptor in human breast cancers: correlation with estrogen and progesterone receptors. Breast Cancer Res Treat 11: 173-178, 1988 Costa S, Stamm H, Almendral A, Ludwig H, Wyss R, Fabbro D, Ernst A, Takahashi A, Eppenberger U: Predictive value of EGF receptor in breast cancer. Lancet ii: 1258, 1988 Rios MA, Macias A, Perez R, Lage A, Skoog L: Receptors for epidermal growth factor and estrogen as predictors of relapse in patients with mammary carcinoma. Anticancer Res 8: 173-176, 1988 Downward J, Yarden Y, Mayes E, et al.: Close similarity of epidermal growth factor receptor and v-erbB oncogene protein sequence. Nature 307: 521, 1984 UICC (International Union Against Cancer): TNM Classification of malignant tumors, 3rd edn. Geneva, UICC, pp 47-54, 1979 Bloom I-IJC, Richardson WW: Histological grading and prognosis in breast cancer. A study on 1409 cases of which 359 have been followed for 15 years. Br J Cancer 11: 35-77, 1957 EORTC Breast Cooperative Group revision of the standards for the assessment of hormone receptor in human breast cancer. Report of the second EORTC workshop, held on 16-17 March, 1979, in the Netherlands Cancer Institute. Eur J Cancer 16: 1513-1515, 1980 Spyratos F, Briffod M, Gentile A, Brunet M, Brault C, Desplaces A: Flow cytometric study of DNA cytopunctures of benign and malignant breast lesions. Anal Quant Cytol Histol 9: 485-486, 1987 Southern EM: Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98: 503-517, 1975 Fraker PJ, Speck JC: Protein and cell membrane iodination with a sparingly soluble chloramide, 1, 3, 4, 6-tetrachloro-3a,6a-diphenyl glycoluril. Biochem Biophys Res Corn° mun 80: 849-857, 1978 Gross-Bellard M, Oudet P, Chambon P: Isolation of high molecular weight DNA from mammalian cells. Eur J Biochem 36: 32-38, 1973 Feinberg AP, Vogelstein B: A technique for radiolabeling DNA restriction endonuclease fragments to high specific activity. Anal Biochem 131: 6-13, 1983 Ullrich A, Coussens L, Hayflick JS, Dull TLJ, Gray A, Tam AW, Lee J, Yarden Y, Libermann TA, Schlessinger J, Downward J, Mayes EMV, Whittle N, Waterfield MD, Seeburg PH: Human epidermal growth factor receptor cDNA sequence and aberrant expression of the amplified gene in A431 epidermoid carcinoma cells. Nature 309: 418425, 1984 Merlino GT, Sunsuke I, Whang-Peng J, Knutsen T, Xu
26.
27.
28.
29. 30.
31.
32.
33.
34.
35.
36.
37.
89
YH, Clarck AJL, Stratton RH, Wilson RK, Ma DP, Roe BA, Hunts JH, Shimizu N, Pastan I: Structure and localization of genes encoding aberrant and normal epidermal growth factor receptor RNAs from A431 human carcinoma cells. Mol Cell Biol 5: 1722-1734, 1985 Watson R, Oskarsson M, Vande Woude GF: Human DNA sequence homologous to the transforming gene (mos) of Moloney sarcoma virus. Proc Natl Acad Sci USA 79: 40784082, 1982 Lidereau R, Mathieu-Mahul D, Theillet C, Renaud M, Mauchauffe M, Gest J, Larsen CJ: Presence of an allelic Eco RI restriction fragment of the c-mos locus in leukocyte and tumor cell DNAs of breast cancer patients. Proc Natl Acad Sci USA 82: 7068--7070, 1985 Peto R, Pike MC, Armitage Pet al.: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 35: 1-39, 1977 Cox DR: Regression models and life tables. J R Stat Soc 34: 187-220, 1972 Perez R, Pascual M, Macias A, Loge A: Epidermal growth factor receptor in human breast cancer. Breast Cancer Res Treat 4: 189-193, 1984 Sainsbury JRC, Farndon JR, Sherbet GV, Harris AL: Epidermal growth factor receptors and estrogen receptors in human breast cancer. Lancet i: 364-366, 1985 Battaglia F, Scambia G, Rossi S, Panici PB, Bellantone R, Polizzi G, Querzoli P, Negrini R, Iacobelli S, Crucitti F, Mancuso S: Epidermal growth fractor receptor in human breast cancer: correlation with steroid hormone receptors and axillary lymph node involvement. Eur J Cancer Clin Oncol 24: 1685-1690, 1988 Ali IU, Lidereau R, Callahan R: Heterogeneity of genetic alterations in primary human breast tumors. Lippman ME, Dickson RB (eds) Breast Cancer: Cellular and Molecular Biology. Kluwer, Boston, 1988, pp 25-48 Nicholson S, Sainsbury JRC, Holerow P, Chambers P, Farndon JR, Harris AL: Expression of epidermal growth factor receptors associated with lack of response to endocrine therapy in recurrent breast cancer. Lancet i: 182184, 1989 Nicholson S, Halcrow P, Sainsbury JRC, Angus B, Chambers P, Farndon JR, Harris AL: Epidermal growth factor receptor (EGFr) status associated with failure of primary endocrine therapy in elderly postmenopausal patients with breast cancer. Br J Cancer 58: 810-814, 1988 Grimaux M, Romain S, Remvikos Y, Martin PM, Magdel6nat H: Prognostic value of epidermal growth factor receptor in node-positive breast cancer. Breast Cancer Res Treat 14: 77-90, 1989 Nicholson S, Sainsbury JRC, Needham GK, Chambers P, Farndon JR, Harris AL: Quantitative assays of epidermal growth factor receptor in human breast cancer: cut-off points of clinical relevance. Int J Cancer 42: 36--41, 1988
