Breast CancerResearchand Treatment20: 25-32, 1991. © 1991KluwerAcademicPublishers. Printedin the Netherlands. Report
Cytosol protein content and prognosis in operable breast cancer Correlations with steroid h o r m o n e receptors and other prognostic factors J.A. S0reide, 1 0 . A . Lea 2 and S. Kvinnsland 3
1Dept. of Surgery, Rogaland Central Hospital, Stavanger; 20ncology Research Laboratory, Haukeland University Hospital, Bergen; 3Dept. of Oncology, University Hospital, Trondheim, Norway
Key words: cytosol protein, primary breast cancer, prognostic factors Summary
Several biochemical parameters quantitated in tumor cytosols from malignant breast tumors have been evaluated as possible prognostic factors. Cytosol protein content has always been regarded as a reference parameter, to correct for cellularity and representativity of tumor samples. But recent studies have suggested an altered protein distribution in malignant tissues. The present study on 382 women with histologically proven breast cancer, Stage I and Stage II, therefore evaluates whether cytosol protein content by itself may add information as a prognostic factor in the clinical management of breast cancer. Cytosol protein content was found to be significantly correlated (p < 0.001) to tumor size, and inversely correlated to progesterone receptor (PgR) content (p = 0.015) and age at operation (p = 0.021). Using the median value of protein (4.15 mg/ml) as a cut-off value, two groups could be constructed. The number of node-positive patients in the protein-poor group was significantly decreased (p = 0.018) compared to the protein-rich group, which also contained a significantly (p < 0.001) lower number of patients with estrogen receptor (ER) positive tumors (i.e. E R >- 10 pmol/g). An increased number of events was observed in the protein-rich group (p < 0.001), with a great contribution to the number of deaths due to breast cancer. In a multivariate analysis of the likelihood to predict axillary nodal involvement, protein category was found to be a significant (p < 0.031) independent predictive factor. As to relapse free survival (RFS), protein category did not reveal any prognostic power. A subgroup containing Stage II patients with E R + tumors was included in a prospective, randomized study on the role of tamoxifen as an adjuvant endocrine therapy. In a multivariate analysis, treatment option (tamoxifen vs. control) (p = 0.0008) and axillary nodal tumor burden (p = 0.009) were significant independent prognostic factors for RFS in protein-rich patients. In protein-poor patients, only tumor size showed a positive association with RFS, but without reaching statistical significance (p = 0.062). The present observations make cytosol protein content interesting in the context of tumor biology and breast cancer prognosis. Further studies are needed to evaluate cytosol protein content as a possible prognostic factor useful in the clinical management of breast cancer.
Introduction
Breast cancer is characterized by a long natural history and by a wide variation in its clinical course
[1]. Different prognostic factors have been used in the clinical management of breast cancer [2]. Still, axillary lymph node involvement is the 'gold standard' which all other prognostic factors should be
Addressfor offprints: S. Kvinnsland, Dept. of Oncology,UniversityHospital, N - 7006Trondheim, Norway
26
J A Screide et al.
compared to [2, 3]. As some node-negative breast cancer patients are still at high risk for recurrent disease [4], other parameters of the primary tumor can be of importance. Several biochemical components such as steroid hormone receptors [5], epidermal growth factor receptor (EGFR) [6], cathepsin D [7], or oncogene expression or amplification [8], have been analysed to obtain more information as to the malignant potential of a tumor. Based on a specific biochemical profile in addition to clinical and demographic variables, our goal should be to offer an efficacious adjuvant therapy to patients at high risk for recurrence of the disease. The amount of protein is quantitated in cytosols intended for steroid hormone receptor (ER = estrogen receptor, PgR = progesterone receptor) analysis, to serve as a reference parameter [9] and to correct for cellularity and representativity of tumor samples [10, 11]. Plasma protein uptake in different malignant cells without prior extracellular hydrolysis of the protein was suggested by Ghose et al. [12]. The question of an altered protein distribution in malignant tissue has recently been addressed by others [13, 14]. The present study was done to evaluate if cytosol protein content could add information as a prognostic factor in the clinical management of breast cancer. Furthermore, we have extended our evaluation to analyse if cytosol protein content may yield information as a possible predictive factor for response to adjuvant endocrine therapy with tamoxifen given to stage II patients with E R positive (-> 10 pmol/g protein) tumors.
Materials and methods Patients and treatment
The study includes 382 women with histologically verified breast cancer diagnosed between 1983 and 1987. Only stage I and stage II patients who underwent primary surgical treatment were included. A modified radical mastectomy was performed in 359 patients (94%), and 23 patients (6%) had a simple mastectomy. Among these 23 patients, the axillary nodal status was unknown in 10 cases, node-nega-
rive in six patients, and node-positive in seven patients. Steroid hormone receptors (i.e. estrogen receptor [ER] and progesterone receptor [PgR]) as well as total protein were quantitated in the tumor cytosols. Patient characteristics are given in Table 1. All patients younger than 70 years (79% of the study population) received perioperative adjuvant chemotherapy treatment: cyclophosphamide 400 mg, vincristine 1 mg, and fluorouracil 500 mg given intravenously on the day of operation, repeated on day seven giving methotrexate 50mg instead of fluoruracil. As a part of the primary treatment, 27 patients (8.7%) received postoperative radiation. Within the study population, 130 node-positive patients with ER-positive tumors were included in a prospective randomized trial on the role of adjuvant treatment with tamoxifen. Tamoxifen was given as 20mg tablets once daily. Relapse-freesurvival (RFS) was defined as the time from diagnosis to first relapse, to death, or to end of followup. The median follow-up time was 40 months (range 3-75 months).
Methods
Tumor tissue from mastectomy specimens was immediately chilled on ice and frozen at - 70 ° C upon receipt in the laboratory. Frozen tissue (0.2-0.5 g) was finely sliced with a scalpel and homogenized at
Table 1. Patient characteristics (n = 382)
Age at diagnosis median range Age distribution < 50 yrs ->50 yrs Menopausal status Pre-menopausal Post-menopausal Unknown Clinical stage at diagnosisa Stage I Stage II
: 61.5 yrs : 21-89 yrs : 106 (27.7%) : 276 (72.8%) : 98 (25.6%) : 276 (72.3%) : 8(2.1%) : 136 (35.6%) : 246 (65.4%)
"Stages (I-II) are used as defined by UICC [19].
Cytosol protein and prognosis 0°C in an adjusted volume (5 vol) of buffer using 3 x 20 seconds burst from an Ultra Turrax homogenizer. The buffer, intended for hormone receptor assay, consisted of 5 mM phosphate buffer, pH 7.4, containing 10% glycerol and l m M thioglycerol. The homogenate was centrifuged for 1 hour at 105,000 g and the resulting supernatant, 'cytosol', was used for protein analysis. Total protein was measured by a dye-binding method [15] using a 1 : i mixture of bovine albumin and IgG as reference standard. Estrogen receptor (ER) and progesterone receptor (PgR) were analysed by single point dextrancoated charcoal assay using saturating concentrations of tritiated estradiol and promegesterone (R5020) as ligands [16].
Statistics The M E D L O G program for personal computers was used for statistical analysis. The variables were categorized as displayed in the tables. Only patients with known values of all variables in a particular analysis have been included in the computation. Chi-square analysis with correction for continuity was used for the univariate analysis of dis-
Table 2. Correlation matrix R
Protein Protein Protein Protein Protein Protein ER ER ER ER ER PgR PgR PgR PgR
vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs.
pT-size PgR Age ER involvedLN examinedLN Age PgR pT-size involvedLN examinedLN involvedLN Age pT-size examinedLN
P
(+)0.119 4.15 mg/ml), treatment option was shown to be a highly statistical significant independent factor (p = 0.0008) for relapse-free survival. Tumor burden in the axillary lymph nodes emerged as another significant (p = 0.009) prognostic factor for RFS. Neither age at operation, nor PgR positivity, nor tumor size revealed significant independent prognostic power. In the other group, with a low protein content ( median (n = 71)
(+)0.105 (+)0.612 (+)0.062 (-)0.367 (+)0.570
(+)0.0008 (+)0.009 (+)0.381 (-)0.109 (-)0.295
d = randomized adjuvant treatment with taN. tamoxifen 20 mg daily for 2 years or no endocrine adjuvant treatment; 1 = cut-off level 50 yrs; 2 = cut-off level 10 pmol/g; c = UICC criteria [19]; (+) or ( - ) = positive or negative association.
30
JA SOreide et al.
in biological specimens. Whether this is an optimal choice for a standard reference has been disputed [11]. Nevertheless, the ease of methods for protein quantification, and the need for an ubiquitous standard variable, have made 'cytosol protein' an appropriate parameter. In our study only cytosols containing at least lmg/ml protein have been included. Below this level of protein, the representativity of the specimen intended for quantification of biochemical parameters in breast tumor tissue should be questioned, and the results of steroid hormone receptor quantification are considered unfit for use in the clinical management of breast cancer patients. According to patient characteristics (Table 1) and distribution of steroid hormone receptors, the study population is regarded as representative for patients with operable breast cancer. A large contribution of stage II patients with ER positive tumors may explain a rather high percent (61.1%) of patients being both E R and PgR positive. According to other reports [20, 21], 45-50% of breast cancer patients are expected to be both ER and PgR positive. The correlation analysis reveals a well known and highly significant positive association between ER and PgR content, and between ER content and age. The significant positive correlation between protein content and tumor size has until now not been focused on in the literature. Furthermore, the negative associations between protein content and PgR content as well as age at operation have, to our knowledge, not been emphasized previously. According to a recent study [10], extraceUular proteins constitute about 50% of total cytosol protein. This indicates an exceptional capillary leakage in breast tumors possibly related to abnormal hormonal influences. From hemoglobin measurements it can be concluded that blood contamination contributes less than 10% to the cytosol measured [10]. A simple explanation for the tumor size and protein association could be that large tumors may provide large tumor specimens intended for biochemical analysis. This bias, however, should in part be circumvented in our study by the precaution of homogenizing the specimens in a fixed volume of buffer related to weight of specimen [16].
This should at least partly correct for variations in protein content measured per milliliter in relation to size of specimen. Another possibility could be that larger tumors may have an increased protein content [12] which in part may be due to an increased capillary permeability in poorly developed vessels in rapidly growing tissues [22]. It remains speculative whether the observed inverse association between protein and age indicates a more aggressive or rapid tumor growth in younger patients. According to a recent report [23] the prognosis in breast cancer patients aged 34 years or younger was found to be poorer compared to all other age groups. It is hard to explain this observation, and the authors speculate on the possibility of a more malignant tumor biology in younger women with breast cancer. The cut-off value of protein (median = 4.15 m~/ ml) to split up the study population in two subgroups is arbitrarily chosen. An interesting observation concerning lymph node metastases is that the two groups may be regarded as equal according to the distribution of patients with different nodal involvement (Table 3). On the other hand, protein category (low versus high content) emerges as an independent significant factor for predicting axillary nodal involvement (Table 4). This is in accordance with numbers from Table 3, which indicate a significantly increased number of node-positive patients in the subgroup B with a higher protein content. Tumor size, ER status, and protein category are all independent significant factors which add information as to the likelihood of axillary nodal involvement. Whether this information may influence the surgical approach to the axilla is presently unclear. In the analysis of relapse-free survival protein category did not provide any prognostic information. It is not surprising that tumor size and nodal status are significant independent prognostic factors in patients not offered endocrine treatment. In node-positive patients with ER-positive tumors, the contribution of tumor size and treatment option as prognostic factors was as expected. The stratification of patients included in the adjuvant trial on the role of tamoxifen provides, nevertheless, further interesting prognostic information
Cytosol protein and prognosis (Table 5). Relapse-free survival was found to be positively associated with adjuvant tamoxifen treatment in the protein-rich subgroup of ER-positive patients. The biological significance of this observation is unclear, but we may again speculate on an alteration of growth rate or metastatic potential in protein-rich tumors. Accordingly, the benefit of tamoxifen is shown first of all in patients supposed to have rapidly growing tumors or tumors with an increased biological agressiveness. Tumor size has been regarded as an important prognostic factor in breast cancer [24]. In the protein-poor subgroup (Table 5) of ER-positive tumors, only the tumor size approaches a borderline positive association with prognosis, but without reaching the level of statistical significance in the multivariate analysis. Several tumor variables and patient characteristics have been evaluated to provide prognostic information useful in breast cancer management [2, 3, 5-8, 23-25]. In the clinical situation, there is still a need for valid and feasible factors to improve recognition of patients with an increased risk of recurrent disease. Our results indicate that the importance of cytosol protein content may be extended beyond its role as a reference parameter. The ease of quantification and our present observations and correlations to known prognostic factors suggest this tumor variable to be of interest for further evaluation as a possible prognostic factor in the clinical management of breast cancer.
3. 4.
5.
6.
7.
8.
9.
10. 11.
12. 13.
14.
Acknowledgements
15.
The financial support from the Norwegian Cancer Society is highly appreciated.
16.
References
17.
1. Harris JR, Henderson IC: Natural history and staging of breast cancer. In: Harris JR, Hellman S, Henderson IC, Kinne DW (eds) Breast Diseases. JB Lippincott Company, Philadelphia, 1987, pp 233-258 2. McGuire WL, Clark GM: Prognosis in breast cancer. In: Senn H-J, Goldhirsch A, Gelber RD, Osteiwalder B (eds)
18. 19. 20.
31
Adjuvant Therapy of Primary Breast Cancer. Springer Verlag, New York, 1989, pp t70-174 Jotti GL: New prognostic indicators in resectable breast cancer (review). Anticancer Res 9: 1227-1232, 1989 Van de Velde CJH, Gallager HS, Giacco GG: Prognosis in node negative breast cancer. Breast Cancer Res Treat 8: 189--196, 1986 Chevallier B, Heintzmann F, Mosseri V, Dauce JP, Bastit P, Graic Y, Brunelle P, Basuyau JP, Comoz M, Asselain B: Prognostic value of estrogen and progesterone receptors in operable breast cancer. Results of an univariate and multivariate analysis. Cancer 62: 251%2524, 1988 Sainsbury JRC, Farndon JR, Harris, Sherbet GV: Epidermal growth factor receptor on human breast cancer. Br J Surg 72: 186-188, 1985 Tandon AK, Clark GM, Chamness GC, Chirgwin JM, McGuire WL: Cathepsin D and prognosis in breast cancer. N Engl J Med 322: 297-302, 1990 Slamon D J, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL: Human breast cancer: Correlation of retapse and survival with amplification of HER-2/neu oncogene. Science 235: 177-182, 1987 Koenders A, Thorpe SM: Standardization of steroid receptor assays in human breast cancer-IV. Long-term withinand between-laboratory variation of estrogen and progesterone receptor assays. Eur J Clin Onco122: 945-953, 1986 Lea OA, Thorsen T, Kvinnsland S: Extracellular proteins in breast tumor cytosol. Anticancer Res 7: 113-118, 1987 Fiszer-Szafarz B, Szafarz D: DNA and protein content as cellular biochemical parameters. Anal Biochem 138: 255258, 1984 Ghose T, Nairn RC, Fothergill JE: Uptake of proteins by malignant cells. Nature 196: 1108-1109, 1962 Etzrodt A, Bigalke LH, Jung H, Totxdorff T: Occurrence and distribution of plasma proteins in malignant and benign breast tissue. Tumor Diagnostik & Therapie 4: 109-113, 1983 Miller FN, Anderson GL, Terman DS, Harris PD, Savage H: Macromolecular leakage from the vasculature of rat [breast] adenocarcinoma. Fed Proc 43:599 (Abstr. # 599), 1984 Sedmark J J, Grosberg SE: A rapid, sensitive and versatile assay for protein using Coomassie Brilliant Blue G250. Anal Biochem 10: 544-552, 1977 Thorsen T: Occupied and unoccupied nuclear oestradiol receptor in human breast tumors: Relation to oestradiol and progesterone cytosot receptors. J Steroid Biochem 10: 661-668, 1979 Kalbfleisch JD, Prentice RL: The Statistical Analysis of Failure Time Data. Wiley, New York, 1980 Cox DR: Analysis of Binary Data. Chapman and Hall, London, 1970, Section 6.6 UICC: TNM Classification of Malignant Turnouts (ed 4). Springer Verlag, New York, 1987 Clark GM, McGuire WL: Progesterone receptors and hu-
32
J A SOreide et al.
man breast cancer. Breast Cancer Res Treat 3: 15%163, 1983 21. S0reide JA, Lea OA, Kvinnsland S: Progesterone binding cyst protein (PBCP) in primary breast cancer. A new prognostic factor? Breast Cancer Res Treat 9: 123-128, 1987 22. Haraldsson B, Regner L, Hultborn L, Weiss L, Rippe B: Transcapillary passage of albumin in mammary tumours and in normal lactating mammary glands of the rat. Acta Physiol Scand 122: 487-505, 1984
23. HOst H, Lund E: Age as a prognostic factor in breast cancer. Cancer 57: 2217-2221, 1986 24. Carter CL, Allen C, Henson DE: Relation of tumor size, lymph node status, and survival in 24,740 breast cancer cases. Cancer 63: 181-187, 1989 25. Merkel DE, Osborne CK: Prognostic factors in breast cancer. Hematol/Oncol Clin N Am 3: 641-652, 1990
Cytosol protein content and prognosis in operable breast cancer Correlations with steroid h o r m o n e receptors and other prognostic factors J.A. S0reide, 1 0 . A . Lea 2 and S. Kvinnsland 3
1Dept. of Surgery, Rogaland Central Hospital, Stavanger; 20ncology Research Laboratory, Haukeland University Hospital, Bergen; 3Dept. of Oncology, University Hospital, Trondheim, Norway
Key words: cytosol protein, primary breast cancer, prognostic factors Summary
Several biochemical parameters quantitated in tumor cytosols from malignant breast tumors have been evaluated as possible prognostic factors. Cytosol protein content has always been regarded as a reference parameter, to correct for cellularity and representativity of tumor samples. But recent studies have suggested an altered protein distribution in malignant tissues. The present study on 382 women with histologically proven breast cancer, Stage I and Stage II, therefore evaluates whether cytosol protein content by itself may add information as a prognostic factor in the clinical management of breast cancer. Cytosol protein content was found to be significantly correlated (p < 0.001) to tumor size, and inversely correlated to progesterone receptor (PgR) content (p = 0.015) and age at operation (p = 0.021). Using the median value of protein (4.15 mg/ml) as a cut-off value, two groups could be constructed. The number of node-positive patients in the protein-poor group was significantly decreased (p = 0.018) compared to the protein-rich group, which also contained a significantly (p < 0.001) lower number of patients with estrogen receptor (ER) positive tumors (i.e. E R >- 10 pmol/g). An increased number of events was observed in the protein-rich group (p < 0.001), with a great contribution to the number of deaths due to breast cancer. In a multivariate analysis of the likelihood to predict axillary nodal involvement, protein category was found to be a significant (p < 0.031) independent predictive factor. As to relapse free survival (RFS), protein category did not reveal any prognostic power. A subgroup containing Stage II patients with E R + tumors was included in a prospective, randomized study on the role of tamoxifen as an adjuvant endocrine therapy. In a multivariate analysis, treatment option (tamoxifen vs. control) (p = 0.0008) and axillary nodal tumor burden (p = 0.009) were significant independent prognostic factors for RFS in protein-rich patients. In protein-poor patients, only tumor size showed a positive association with RFS, but without reaching statistical significance (p = 0.062). The present observations make cytosol protein content interesting in the context of tumor biology and breast cancer prognosis. Further studies are needed to evaluate cytosol protein content as a possible prognostic factor useful in the clinical management of breast cancer.
Introduction
Breast cancer is characterized by a long natural history and by a wide variation in its clinical course
[1]. Different prognostic factors have been used in the clinical management of breast cancer [2]. Still, axillary lymph node involvement is the 'gold standard' which all other prognostic factors should be
Addressfor offprints: S. Kvinnsland, Dept. of Oncology,UniversityHospital, N - 7006Trondheim, Norway
26
J A Screide et al.
compared to [2, 3]. As some node-negative breast cancer patients are still at high risk for recurrent disease [4], other parameters of the primary tumor can be of importance. Several biochemical components such as steroid hormone receptors [5], epidermal growth factor receptor (EGFR) [6], cathepsin D [7], or oncogene expression or amplification [8], have been analysed to obtain more information as to the malignant potential of a tumor. Based on a specific biochemical profile in addition to clinical and demographic variables, our goal should be to offer an efficacious adjuvant therapy to patients at high risk for recurrence of the disease. The amount of protein is quantitated in cytosols intended for steroid hormone receptor (ER = estrogen receptor, PgR = progesterone receptor) analysis, to serve as a reference parameter [9] and to correct for cellularity and representativity of tumor samples [10, 11]. Plasma protein uptake in different malignant cells without prior extracellular hydrolysis of the protein was suggested by Ghose et al. [12]. The question of an altered protein distribution in malignant tissue has recently been addressed by others [13, 14]. The present study was done to evaluate if cytosol protein content could add information as a prognostic factor in the clinical management of breast cancer. Furthermore, we have extended our evaluation to analyse if cytosol protein content may yield information as a possible predictive factor for response to adjuvant endocrine therapy with tamoxifen given to stage II patients with E R positive (-> 10 pmol/g protein) tumors.
Materials and methods Patients and treatment
The study includes 382 women with histologically verified breast cancer diagnosed between 1983 and 1987. Only stage I and stage II patients who underwent primary surgical treatment were included. A modified radical mastectomy was performed in 359 patients (94%), and 23 patients (6%) had a simple mastectomy. Among these 23 patients, the axillary nodal status was unknown in 10 cases, node-nega-
rive in six patients, and node-positive in seven patients. Steroid hormone receptors (i.e. estrogen receptor [ER] and progesterone receptor [PgR]) as well as total protein were quantitated in the tumor cytosols. Patient characteristics are given in Table 1. All patients younger than 70 years (79% of the study population) received perioperative adjuvant chemotherapy treatment: cyclophosphamide 400 mg, vincristine 1 mg, and fluorouracil 500 mg given intravenously on the day of operation, repeated on day seven giving methotrexate 50mg instead of fluoruracil. As a part of the primary treatment, 27 patients (8.7%) received postoperative radiation. Within the study population, 130 node-positive patients with ER-positive tumors were included in a prospective randomized trial on the role of adjuvant treatment with tamoxifen. Tamoxifen was given as 20mg tablets once daily. Relapse-freesurvival (RFS) was defined as the time from diagnosis to first relapse, to death, or to end of followup. The median follow-up time was 40 months (range 3-75 months).
Methods
Tumor tissue from mastectomy specimens was immediately chilled on ice and frozen at - 70 ° C upon receipt in the laboratory. Frozen tissue (0.2-0.5 g) was finely sliced with a scalpel and homogenized at
Table 1. Patient characteristics (n = 382)
Age at diagnosis median range Age distribution < 50 yrs ->50 yrs Menopausal status Pre-menopausal Post-menopausal Unknown Clinical stage at diagnosisa Stage I Stage II
: 61.5 yrs : 21-89 yrs : 106 (27.7%) : 276 (72.8%) : 98 (25.6%) : 276 (72.3%) : 8(2.1%) : 136 (35.6%) : 246 (65.4%)
"Stages (I-II) are used as defined by UICC [19].
Cytosol protein and prognosis 0°C in an adjusted volume (5 vol) of buffer using 3 x 20 seconds burst from an Ultra Turrax homogenizer. The buffer, intended for hormone receptor assay, consisted of 5 mM phosphate buffer, pH 7.4, containing 10% glycerol and l m M thioglycerol. The homogenate was centrifuged for 1 hour at 105,000 g and the resulting supernatant, 'cytosol', was used for protein analysis. Total protein was measured by a dye-binding method [15] using a 1 : i mixture of bovine albumin and IgG as reference standard. Estrogen receptor (ER) and progesterone receptor (PgR) were analysed by single point dextrancoated charcoal assay using saturating concentrations of tritiated estradiol and promegesterone (R5020) as ligands [16].
Statistics The M E D L O G program for personal computers was used for statistical analysis. The variables were categorized as displayed in the tables. Only patients with known values of all variables in a particular analysis have been included in the computation. Chi-square analysis with correction for continuity was used for the univariate analysis of dis-
Table 2. Correlation matrix R
Protein Protein Protein Protein Protein Protein ER ER ER ER ER PgR PgR PgR PgR
vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. vs.
pT-size PgR Age ER involvedLN examinedLN Age PgR pT-size involvedLN examinedLN involvedLN Age pT-size examinedLN
P
(+)0.119 4.15 mg/ml), treatment option was shown to be a highly statistical significant independent factor (p = 0.0008) for relapse-free survival. Tumor burden in the axillary lymph nodes emerged as another significant (p = 0.009) prognostic factor for RFS. Neither age at operation, nor PgR positivity, nor tumor size revealed significant independent prognostic power. In the other group, with a low protein content ( median (n = 71)
(+)0.105 (+)0.612 (+)0.062 (-)0.367 (+)0.570
(+)0.0008 (+)0.009 (+)0.381 (-)0.109 (-)0.295
d = randomized adjuvant treatment with taN. tamoxifen 20 mg daily for 2 years or no endocrine adjuvant treatment; 1 = cut-off level 50 yrs; 2 = cut-off level 10 pmol/g; c = UICC criteria [19]; (+) or ( - ) = positive or negative association.
30
JA SOreide et al.
in biological specimens. Whether this is an optimal choice for a standard reference has been disputed [11]. Nevertheless, the ease of methods for protein quantification, and the need for an ubiquitous standard variable, have made 'cytosol protein' an appropriate parameter. In our study only cytosols containing at least lmg/ml protein have been included. Below this level of protein, the representativity of the specimen intended for quantification of biochemical parameters in breast tumor tissue should be questioned, and the results of steroid hormone receptor quantification are considered unfit for use in the clinical management of breast cancer patients. According to patient characteristics (Table 1) and distribution of steroid hormone receptors, the study population is regarded as representative for patients with operable breast cancer. A large contribution of stage II patients with ER positive tumors may explain a rather high percent (61.1%) of patients being both E R and PgR positive. According to other reports [20, 21], 45-50% of breast cancer patients are expected to be both ER and PgR positive. The correlation analysis reveals a well known and highly significant positive association between ER and PgR content, and between ER content and age. The significant positive correlation between protein content and tumor size has until now not been focused on in the literature. Furthermore, the negative associations between protein content and PgR content as well as age at operation have, to our knowledge, not been emphasized previously. According to a recent study [10], extraceUular proteins constitute about 50% of total cytosol protein. This indicates an exceptional capillary leakage in breast tumors possibly related to abnormal hormonal influences. From hemoglobin measurements it can be concluded that blood contamination contributes less than 10% to the cytosol measured [10]. A simple explanation for the tumor size and protein association could be that large tumors may provide large tumor specimens intended for biochemical analysis. This bias, however, should in part be circumvented in our study by the precaution of homogenizing the specimens in a fixed volume of buffer related to weight of specimen [16].
This should at least partly correct for variations in protein content measured per milliliter in relation to size of specimen. Another possibility could be that larger tumors may have an increased protein content [12] which in part may be due to an increased capillary permeability in poorly developed vessels in rapidly growing tissues [22]. It remains speculative whether the observed inverse association between protein and age indicates a more aggressive or rapid tumor growth in younger patients. According to a recent report [23] the prognosis in breast cancer patients aged 34 years or younger was found to be poorer compared to all other age groups. It is hard to explain this observation, and the authors speculate on the possibility of a more malignant tumor biology in younger women with breast cancer. The cut-off value of protein (median = 4.15 m~/ ml) to split up the study population in two subgroups is arbitrarily chosen. An interesting observation concerning lymph node metastases is that the two groups may be regarded as equal according to the distribution of patients with different nodal involvement (Table 3). On the other hand, protein category (low versus high content) emerges as an independent significant factor for predicting axillary nodal involvement (Table 4). This is in accordance with numbers from Table 3, which indicate a significantly increased number of node-positive patients in the subgroup B with a higher protein content. Tumor size, ER status, and protein category are all independent significant factors which add information as to the likelihood of axillary nodal involvement. Whether this information may influence the surgical approach to the axilla is presently unclear. In the analysis of relapse-free survival protein category did not provide any prognostic information. It is not surprising that tumor size and nodal status are significant independent prognostic factors in patients not offered endocrine treatment. In node-positive patients with ER-positive tumors, the contribution of tumor size and treatment option as prognostic factors was as expected. The stratification of patients included in the adjuvant trial on the role of tamoxifen provides, nevertheless, further interesting prognostic information
Cytosol protein and prognosis (Table 5). Relapse-free survival was found to be positively associated with adjuvant tamoxifen treatment in the protein-rich subgroup of ER-positive patients. The biological significance of this observation is unclear, but we may again speculate on an alteration of growth rate or metastatic potential in protein-rich tumors. Accordingly, the benefit of tamoxifen is shown first of all in patients supposed to have rapidly growing tumors or tumors with an increased biological agressiveness. Tumor size has been regarded as an important prognostic factor in breast cancer [24]. In the protein-poor subgroup (Table 5) of ER-positive tumors, only the tumor size approaches a borderline positive association with prognosis, but without reaching the level of statistical significance in the multivariate analysis. Several tumor variables and patient characteristics have been evaluated to provide prognostic information useful in breast cancer management [2, 3, 5-8, 23-25]. In the clinical situation, there is still a need for valid and feasible factors to improve recognition of patients with an increased risk of recurrent disease. Our results indicate that the importance of cytosol protein content may be extended beyond its role as a reference parameter. The ease of quantification and our present observations and correlations to known prognostic factors suggest this tumor variable to be of interest for further evaluation as a possible prognostic factor in the clinical management of breast cancer.
3. 4.
5.
6.
7.
8.
9.
10. 11.
12. 13.
14.
Acknowledgements
15.
The financial support from the Norwegian Cancer Society is highly appreciated.
16.
References
17.
1. Harris JR, Henderson IC: Natural history and staging of breast cancer. In: Harris JR, Hellman S, Henderson IC, Kinne DW (eds) Breast Diseases. JB Lippincott Company, Philadelphia, 1987, pp 233-258 2. McGuire WL, Clark GM: Prognosis in breast cancer. In: Senn H-J, Goldhirsch A, Gelber RD, Osteiwalder B (eds)
18. 19. 20.
31
Adjuvant Therapy of Primary Breast Cancer. Springer Verlag, New York, 1989, pp t70-174 Jotti GL: New prognostic indicators in resectable breast cancer (review). Anticancer Res 9: 1227-1232, 1989 Van de Velde CJH, Gallager HS, Giacco GG: Prognosis in node negative breast cancer. Breast Cancer Res Treat 8: 189--196, 1986 Chevallier B, Heintzmann F, Mosseri V, Dauce JP, Bastit P, Graic Y, Brunelle P, Basuyau JP, Comoz M, Asselain B: Prognostic value of estrogen and progesterone receptors in operable breast cancer. Results of an univariate and multivariate analysis. Cancer 62: 251%2524, 1988 Sainsbury JRC, Farndon JR, Harris, Sherbet GV: Epidermal growth factor receptor on human breast cancer. Br J Surg 72: 186-188, 1985 Tandon AK, Clark GM, Chamness GC, Chirgwin JM, McGuire WL: Cathepsin D and prognosis in breast cancer. N Engl J Med 322: 297-302, 1990 Slamon D J, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL: Human breast cancer: Correlation of retapse and survival with amplification of HER-2/neu oncogene. Science 235: 177-182, 1987 Koenders A, Thorpe SM: Standardization of steroid receptor assays in human breast cancer-IV. Long-term withinand between-laboratory variation of estrogen and progesterone receptor assays. Eur J Clin Onco122: 945-953, 1986 Lea OA, Thorsen T, Kvinnsland S: Extracellular proteins in breast tumor cytosol. Anticancer Res 7: 113-118, 1987 Fiszer-Szafarz B, Szafarz D: DNA and protein content as cellular biochemical parameters. Anal Biochem 138: 255258, 1984 Ghose T, Nairn RC, Fothergill JE: Uptake of proteins by malignant cells. Nature 196: 1108-1109, 1962 Etzrodt A, Bigalke LH, Jung H, Totxdorff T: Occurrence and distribution of plasma proteins in malignant and benign breast tissue. Tumor Diagnostik & Therapie 4: 109-113, 1983 Miller FN, Anderson GL, Terman DS, Harris PD, Savage H: Macromolecular leakage from the vasculature of rat [breast] adenocarcinoma. Fed Proc 43:599 (Abstr. # 599), 1984 Sedmark J J, Grosberg SE: A rapid, sensitive and versatile assay for protein using Coomassie Brilliant Blue G250. Anal Biochem 10: 544-552, 1977 Thorsen T: Occupied and unoccupied nuclear oestradiol receptor in human breast tumors: Relation to oestradiol and progesterone cytosot receptors. J Steroid Biochem 10: 661-668, 1979 Kalbfleisch JD, Prentice RL: The Statistical Analysis of Failure Time Data. Wiley, New York, 1980 Cox DR: Analysis of Binary Data. Chapman and Hall, London, 1970, Section 6.6 UICC: TNM Classification of Malignant Turnouts (ed 4). Springer Verlag, New York, 1987 Clark GM, McGuire WL: Progesterone receptors and hu-
32
J A SOreide et al.
man breast cancer. Breast Cancer Res Treat 3: 15%163, 1983 21. S0reide JA, Lea OA, Kvinnsland S: Progesterone binding cyst protein (PBCP) in primary breast cancer. A new prognostic factor? Breast Cancer Res Treat 9: 123-128, 1987 22. Haraldsson B, Regner L, Hultborn L, Weiss L, Rippe B: Transcapillary passage of albumin in mammary tumours and in normal lactating mammary glands of the rat. Acta Physiol Scand 122: 487-505, 1984
23. HOst H, Lund E: Age as a prognostic factor in breast cancer. Cancer 57: 2217-2221, 1986 24. Carter CL, Allen C, Henson DE: Relation of tumor size, lymph node status, and survival in 24,740 breast cancer cases. Cancer 63: 181-187, 1989 25. Merkel DE, Osborne CK: Prognostic factors in breast cancer. Hematol/Oncol Clin N Am 3: 641-652, 1990
