Messenger RNA Determination of Estrogen Receptor, Progesterone ...

Vol. 5, 1497–1502, June 1999

Clinical Cancer Research 1497

Messenger RNA Determination of Estrogen Receptor, Progesterone Receptor, pS2, and Plasminogen Activator Inhibitor-1 by Competitive Reverse Transcription-Polymerase Chain Reaction in Human Breast Cancer1 Dan Tong, Christian Schneeberger, Klaus Czerwenka, Rita K. Schmutzler, Paul Speiser, Elisabeth Kucera, Nicole Concin, Ernst Kubista, Sepp Leodolter, and Robert Zeillinger2 Division of Gynecology, Molecular Oncology Group [D. T., P. S., E. Kuc., N. C., S. L., R. Z.], Division of Gynecological Endocrinology and Reproductive Medicine [C. S.], and Division of Senology [E. Kub.], Department of Obstetrics and Gynecology, and Department of Clinical Pathology [K. C.], and Ludwig-Boltzmann Institute for Oncology and Fertility Treatment [S. L.], University of Vienna, A1090 Vienna, Austria; and Department of Obstetrics and Gynecology, University of Bonn Medical Center, University of Bonn, D-53105 Bonn, Germany [R. K. S.]

ABSTRACT Estrogen receptor (ER), progesterone receptor (PR), the estrogen-inducible protein pS2, and plasminogen activator inhibitor-1 (PAI-1) are important prognostic factors in primary breast cancer. The protein concentrations of these factors in breast tumors have been well documented. However, few data about the mRNA expression of ER, PR, pS2, and PAI-1 in breast cancer are available, which is mostly due to the limitations of conventional techniques for mRNA analysis. We have described a competitive reverse transcription-PCR system for the simultaneous quantification of ER, PR, pS2, and PAI-1 mRNA in tumor samples. Here, we evaluated 100 tumor biopsies from breast cancer patients for the mRNA expression of ER, PR, pS2, and PAI-1. The results were analyzed for correlations with protein status and with clinical data. Significant correlations between mRNA expression levels and protein concentrations of all tested markers were found. In only a few cases was there an obvious discordance between the measurable amounts of mRNA and protein, especially for ER and PR. In addition, ER, PR, and pS2 mRNA levels correlated significantly with each other. No correlation between PAI-1 mRNA amount

Received 8/26/98; revised 2/25/99; accepted 3/15/99. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by Austrian Science Fund Project P10032Med. 2 To whom requests for reprints should be addressed, at Department of Obstetrics and Gynecology, University of Vienna, Wa¨hringer Gu¨rtel 18-20, EBO 05, A-1090 Vienna, Austria. Phone: 43-1-40400-7831; Fax: 43-1-40400-7832; E-mail: [email protected].

and the expression of the other markers was found. With respect to clinical data, ER and PR mRNA levels were found to be inversely correlated to tumor size and histological grade but not to the lymph node status. pS2 and PAI-1 mRNA expression were not correlated with tumor size, grade, or lymph node involvement. In conclusion, competitive reverse transcription-PCR may be used as an alternative for the study of prognostic factors in human breast cancer and other malignancies. However, before mRNA expression is measured for diagnostics, a presumed concordance of mRNA and protein expression must be evaluated very carefully for every gene.

INTRODUCTION ER3 and PR are important regulators of growth and differentiation in the mammary gland and the female reproductive tract. Both are also involved in the development of malignant tumors (1). pS2 is an estrogen-inducible protein, the function of which is still unknown. However, it is assumed that pS2 is associated with tissue differentiation (2). Urokinase-type plasminogen activator catalyzes the conversion of inactive plasminogen to active plasmin, which contributes to the degradation of basement membranes in tumor invasion and metastasis (3). PAI-1 inhibits the activity of urokinase-type plasminogen activator by forming an enzyme-inhibitor complex (4). ER, PR, pS2, and PAI-1 have been reported to be prognostic as well as predictive parameters in primary breast cancer patients. Negative correlations of ER and PR protein levels to tumor size, number of axillary lymph nodes, and histological grade have been reported frequently (5– 8). ER and PR are also important predictors of longer disease-free intervals and better overall survival for patients with primary breast cancer (9). pS2 protein is highly correlated with ER and PR. Like the hormone receptors, pS2 is often reported to be inversely correlated with tumor size and histological grade (5, 10, 11). However, there are also contradictory reports describing pS2 as an independent parameter (12, 13). pS2 is not related to lymph node involvement, according to most of the publications (5, 10 –13). Moreover, it was also described as a marker for good prognosis (13). PAI-1 has been reported to be expressed at significantly higher levels in malignant or lymph node-positive breast cancers than in lymph node-negative ones or benign lesions (14). Breast

3

The abbreviations used are: ER, estrogen receptor; PR, progesterone receptor; PAI-1, plasminogen activator inhibitor-1; RT-PCR, reverse transcription-PCR; cRT-PCR, competitive RT-PCR; GAPDH, glyceraldehyde 3-phosphate dehydrogenase.

1498 mRNA Expression of Prognostic Markers in Breast Cancer

cancer patients with a high content of PAI-1 in their primary tumors have an increased risk of relapse and death (15, 16). Thus far, most of the studies on ER, PR, pS2, and PAI-1 in breast cancer have been based on the protein measurement of these factors. The protein amount of a specific gene represents only one aspect of the altered gene expression in tumor cells. The mRNA content is, likewise, an important indicator of the gene expression changes. At present, there is a growing number of reports on the differential gene expression in normal and tumor tissues using high-density microarray technology. To apply this technology in diagnostics, we need investigations on the correlation of the mRNA expression with protein levels and, eventually, clinical data. In addition, the relationship between the protein levels and the amount of mRNA is also a substantial element for the understanding of changes in gene expression, especially in the process of tumor development. It has been confirmed that the ratio of ER protein to ER mRNA in ERpositive patients is significantly related to risk of relapse (17). Conventional methods used for mRNA evaluation, such as the Northern blot technique, dot blot analysis, and RNase protection assay, are laborious, require large amounts of starting material, and have low detection limits (18). In practice, small breast tumor biopsies are usually available for RNA analysis. Therefore, it is difficult to perform experiments on gene expression in large series of patients. Actually, there have been only a few examinations of ER, PR, and pS2 mRNA status in large numbers of breast cancer specimens using the Northern blot technique (17, 19, 20) or dot blot analysis (21, 22). cRT-PCR has been widely used in the investigation of gene expression. It is more sensitive than blot hybridization methods or RNase protection assay (18) and allows the quantification of various mRNA species with a very low amount of starting material (23). Thus far, there are only a few studies concerning mRNA expression of ER, PR, and PAI-1 in breast cancer using RT-PCR techniques (24 –26). In two of these studies, the results have been presented qualitatively, and in the other, the mRNA content has been determined by the intensities of the PCR products on agarose gel normalized to that of a coamplified “housekeeping” gene. No internal standard was used for the mRNA quantification, and the relationship of ER, PR, and PAI-1 mRNA to clinical data has not been addressed here. A cRT-PCR system has been developed and validated in our laboratory (27). It allows the simultaneous quantitative determination of ER, PR, pS2, and PAI-1 mRNA in tissue samples. Using this technique, we evaluated 100 breast tumor biopsies for the mRNA expression of ER, PR, pS2, and PAI-1. We compared the results with the corresponding protein status and analyzed the relationship of mRNA levels to clinical data.

MATERIALS AND METHODS Tumor Samples and Patient Characteristics. This study was carried out with a group of 100 fresh tumor tissues from breast cancer patients. Sixty-two of the tissues were from the Department of Obstetrics and Gynecology, University of Vienna, and 38 were from the Department of Obstetrics and Gynecology, University of Bonn. Patient and tumor characteristics are shown in Table 1. Patient age ranged from 32 to 87 years; the median age was 58 years.

Table 1

Patient and tumor characteristics

Characteristic Age ,50 yr $50 yr Histological type Infiltrating ductal carcinoma Infiltrating lobular carcinoma Intraductal carcinoma (DCISa) Sarcoma Unknown Nodal status pN0 pN1 pN2 Unknown Tumor stage pT1 (#2 cm) pT2 (.2–5 cm) pT3 (.5 cm) pT4 Unknown Differentiation grade Well (G1) Moderate (G2) Poor (G3) Unknown (including DCIS) a

No. of samples 33 67 73 19 2 1 5 44 35 7 14 24 56 4 7 9 1 50 43 6

DCIS, ductal carcinoma in situ.

Quantitative Determination of ER, PR, pS2, and PAI-1 Protein Levels. Tumor samples from the University of Vienna were homogenized at 0 – 4°C with an Ultra-Turrax (3 bursts for 10 s each) in phosphate buffer containing 0.1% monothioglycerol. The homogenate was centrifuged at 50,000 3 g for 1 h. The supernatant was used for steroid receptor measurement and stored at 220°C. ER and PR in samples from the University of Vienna were determined by means of the dextran-coated charcoal technique. Cytosol protein was measured with Coomassie Brilliant Blue G-250 using a commercially available kit (Bio-Rad Laboratories, Munich, Germany; Ref. 28). Calculation of affinity constants and binding capacities was carried out according to Scatchard (29). Protein level of the steroid receptors in tumor samples from the University of Bonn was measured by monoclonal antibody assay (30). For the measurement of pS2 protein in cytosol, a radioimmunometric assay kit (CIS Bio Int., Gif-Sur-Yvette, France) was used. PAI-1 protein level was determined with a commercial ELISA kit TintEliza (Biopool, Umea, Sweden). pS2 and PAI-1 protein were not measured in the samples from the University of Bonn Medical Center because no cytosols were available. cRT-PCR and Quantification of Gene Expression of ER, PR, pS2, and PAI-1. Total RNA was extracted from tumor tissue lysates by isopyknic centrifugation, as described previously (31). For the quantification of several genes that play important roles in the pathogenesis of breast cancer, a cRT-PCR system has been developed and validated previously (27). Briefly, a RNA multistandard and a GAPDH RNA standard were constructed. They contain sequences that are homologous to primer pairs for in vitro amplification of human ER, PR, PAI-1, pS2, and GAPDH. The amplification of the standard


Fig. 1 Correlation of ER, PR, pS2, and PAI-1 mRNA expression with the corresponding protein concentrations calculated by the Spearman Rank Correlation Test (ER: r 5 0.65, P , 0.00001, n 5 51; PR: r 5 0.67, P , 0.00001, n 5 51; pS2: r 5 0.88, P , 0.00001, n 5 34; PAI-1: r 5 0.40, P 5 0.0147, n 5 37). mRNA amount was given as relative expression of each gene determined by cRT-PCR normalized to the corresponding GAPDH expression (27). Protein concentrations were determined in tumor cytosol.

sequences with each primer pair generated a standard product differing from its related endogenous wild-type product by size, allowing the discrimination between wild-type and standard PCR products by agarose gel electrophoresis. A known amount of tumor RNA was then mixed with 10 different amounts of the synthesized RNA multistandard and GAPDH RNA standard. Analysis of band intensities was performed following reverse transcription and PCR. To calculate the amount of mRNA from each gene, we plotted the ratio of band intensity of standard to that of wild-type PCR product within each lane against the amount of the standard initially used in RT reaction. The quantity of the target mRNA was determined, where this ratio was equal to 1. To correct the possible sample variations caused either by quantitative determination of total RNA or by RNA degradation, we normalized the quantitative RT-PCR results of the mRNAs to that of GAPDH mRNA. Statistical Analysis. Spearman rank correlation coefficients were used to study the association between continuous variables. For the analysis of the correlation between nonparametric data, the Mann-Whitney U test was used. Differences were considered statistically significant when Ps were ,0.05.

RESULTS AND DISCUSSION Correlations between mRNA and Protein Levels of ER, PR, pS2, and PAI-1. Using the Spearman Rank Correlation Test, we observed significant correlations between the mRNA level and the protein concentration of ER (r 5 0.65, P , 0.00001, n 5 51), PR (r 5 0.67, P , 0.00001, n 5 51), pS2 (r 5 0.88, P , 0.00001, n 5 34), and PAI-1 (r 5 0.40, P 5 0.0147, n 5 37; Fig. 1). Similar results were also obtained using the Mann-Whitney U Test comparing the mRNA level and the protein status of ER (P 5 0.0020, NER1 5 24, NER2 5 11) and PR (P 5 0.0015, NPR1 5 24, NPR2 5 11) assessed by immunohistochemistry. Our results are in agreement with those obtained from several previous studies regarding the concordance of mRNA and the protein levels of ER and PR. One of these studies was carried out on the regulation of ER mRNA and protein levels by steroid hormones, their antagonists, and growth factors (32). Using Northern and Western blotting, it was demonstrated that, in breast cancer cell lines, the ER mRNA and protein levels were down-regulated to a comparable extent when treated with estradiol. This indirectly indicates that the ER


Table 2

Correlation of the expression of ER, PR, pS2, and PAI-1 with clinical data

Nodal statusa

ER mRNA Mean SD n Proteinc Mean SD n PR mRNA Mean SD n Proteinc Mean SD n pS2 mRNA Mean SD n Protein Mean SD n PAI-1 mRNA Mean SD n Protein Mean SD n a b c

Tumor sizea

Differentiation gradea

Negative

Positive

P

#2 cm

.2 cm

P

Moderately/well

Poorly

0.951 1.483 44

0.618 1.055 42

NSb

1.008 1.397 24

0.653 1.133 60

0.0464

1.016 1.381 51

0.567 1.150 43

88.000 153.844 22

35.182 50.335 22

NS

143.333 160.622 9

45.108 97.857 37

0.0141

96.840 144.181 25

50.652 98.200 23

NS

0.203 0.414 44

0.193 0.418 42

NS

0.491 0.662 24

0.086 0.166 60

0.0004

0.286 0.451 51

0.061 0.234 43

,0.0001

127.318 276.147 22

145.909 348.246 22

NS

482.667 555.052 9

42.541 92.204 37

0.0172

178.440 365.071 25

99.130 224.963 23

NS

2.819 5.138 44

4.399 8.772 42

NS

2.736 4.699 24

3.861 7.957 60

NS

4.594 9.216 51

3.048 5.374 43

NS

15.568 37.504 18

37.659 67.509 11

NS

18.958 45.453 7

21.233 51.345 24

NS

21.667 45.012 15

28.943 59.673 18

NS

0.082 0.124 44

0.121 0.173 42

NS

0.105 0.138 24

0.112 0.167 60

NS

0.128 0.190 51

0.110 0.160 43

NS

2.863 2.989 19

2.623 3.049 13

NS

3.300 2.200 7

3.778 6.154 27

NS

2.228 1.982 18

4.856 7.271 18

NS

P

0.0048

Means and SDs are given in units as described in the legend to Fig. 1. NS, not significant. For statistical analysis, only quantitative results obtained by the dextran-coated charcoal technique were included.

protein level correlates with its mRNA level. In another study, comparison of the results generated by a semiquantitative RTPCR system with the results from enzyme immunoassays revealed a significant correlation between mRNA and protein levels of ER and PR (24). Similar findings were reported using Northern blot (17), dot-blot (21, 22), or slot-blot (33) hybridization. With respect to pS2 and PAI-1, no indication for a correlation of mRNA with protein expression could be found in the literature. In a few cases, there was an obvious discordance between the measurable amounts of mRNA and protein, especially with ER and PR expression. However, this was not observed for all analyzed markers in a certain specimen. For example, the specimens showing discordant results for ER showed concordant results for the other markers. Thus, this is probably not a result caused by tumor heterogeneity, tissue handling, or RNA extraction failure. More likely, this reflects biological differences between mRNA and protein in these few samples. In all assays for the analysis of tumor markers, using tissue specimens containing varying portions of tumor cells is

very critical for the results. Enriching the tumor cell fracture by microdissection might improve the correlations in results. For clinical purposes, the results from each of the protein assays were dichotomized. This dichotomization is undoubtedly also needed for mRNA assays. In principal, RNA-based methods, including an in vitro amplification step, are more sensitive than protein-based methods. Therefore, there will be situations in which positive results of the RNA expression of a certain gene can be obtained but the corresponding protein cannot be detected. In such cases, a cutoff point is especially important to interpret the low-level gene expression results that might be without clinical relevance. However, in this study, too few cases have been analyzed to define a cutoff point of clinical relevance. Further studies will be needed to understand the significance of low levels of RNA expression of these markers in breast cancer. Correlation between mRNA levels of ER, PR, pS2, and PAI-1. The mRNA levels of ER, PR, and pS2 in malignant breast tumors (n 5 100) were significantly correlated to each other (ERpPR: r 5 0.49, P , 0.00001; ERppS2: r 5 0.56, P ,


0.00001; PRppS2: r 5 0.45, P , 0.0001; Spearman Rank Correlation Test). But none of them was associated with the mRNA expression of PAI-1. Thus far, direct correlations between ER, PR, and pS2 in all combinations (5, 12, 34, 35) and inverse correlations of PAI-1 with ER and PR (14 –16, 36) at the protein level were reported frequently for breast tumors. We confirmed the correlation of ER, PR, and pS2 to each other at the mRNA level but failed to present the negative correlation of PAI-1 with ER, PR, and pS2. On the one hand, the amount of protein in a cell might not always reflect the amount of mRNA expression. On the other hand, although a correlation of PAI-1 mRNA expression with the protein concentration was found in this study, it was weaker than those observed for ER, PR, and pS2. These might be reasons for the lack of a correlation between PAI-1 and the hormone receptors that was reported at the protein level. In addition, differences in the intracellular stability of different mRNAs could also account for our result. Correlation of Gene Expression to Patient’s Ages. Both ER mRNA (r 5 0.39, P , 0.0001, n 5 100) and protein levels (r 5 0.51, P 5 0.00014, n 5 51) were significantly correlated to the age of the patients (Spearman Rank Correlation Test). A weak correlation between age and PR protein levels but not mRNA levels was also found (r 5 0.33, P 5 0.0199, n 5 51). Correlation of Expression Results and Clinical Data. Both mRNA and protein levels of ER, PR, pS2, and PAI-1 were independent of axillary lymph node involvement (Mann-Whitney U Test, Table 2). In tumor samples of #2 cm, 79.2% (19 of 24), 70.8% (17 of 24), and 70.8% (17 of 24) had detectable ER, PR, and pS2, respectively. In tumors of .2 cm, ER, PR, and pS2 could only be detected in 56.7% (34 of 60), 31.7% (19 of 60), and 58.3% (35 of 60) of the samples, respectively. A statistically significant inverse correlation between tumor size and the expression of ER and PR at both mRNA and protein level was demonstrated by the Mann-Whitney U Test (Table 2). No such correlation was found for pS2 and PAI-1. mRNA expression of hormone receptors also correlated inversely with histological grade. In well or moderately differentiated tumor samples 78.4% (40 of 51) and 64.7% (33 of 51) had detectable ER and PR mRNA, respectively, whereas in the poorly differentiated samples, only 46.5% (20 of 43) and 18.6% (8 of 43) had detectable ER and PR mRNA, respectively. These differences were statistically significant (Mann-Whitney U Test, Table 2). Similarly, there were also more samples with detectable pS2 mRNA in well or moderately differentiated tumors (66.7%; 34 of 51) than in poorly differentiated tumors (55.8%; 24 of 43), However, this was not statistically significant. No correlation of PAI-1 mRNA or protein levels to histological grade was found. In several previous studies, ER and PR were shown to be inversely correlated to tumor size, histological grade, and the number of involved lymph nodes (5–7). However, lack of such correlations has also been demonstrated (37). Our data are in agreement with reports on the correlation between ER, PR, and tumor size and grade. But no correlation of ER and PR mRNA to nodal status was found. Regarding pS2, there were conflicts in reports on the correlation of the protein concentration in tumor cytosols with clinical data (5, 10 –13). However, our data are in agreement with a study using Northern blot analysis that

revealed no correlation between pS2 mRNA expression and tumor size, grade, and nodal status (20). PAI-1 was found to be independent of tumor size, grade, and lymph node involvement, which is in concordance with previous approaches (14 –16). There were evident differences between the Ps for the comparison of PR mRNA and protein levels in regard to tumor size and differentiation grade and for the comparison of ER mRNA and protein levels in regard to differentiation grade (Table 2). These differences are probably due to the small sample size. We show that the investigations of prognostic factors in breast cancer could be performed not only at the protein level but also at the mRNA level. However, before mRNA expression is measured for diagnostics, a presumed concordance of mRNA and protein expression must be evaluated very carefully for every gene because discordant results may be obtained in a small percentage of specimens. For some genes, such as pS2, RNA-based techniques may be sufficient for high throughput analysis of expression in microdissected breast tissue.

ACKNOWLEDGMENTS We are grateful to Ulrike Pruckner and Sabine Eder for their excellent technical assistance.

REFERENCES 1. Katzenellenbogen, B. S. Dynamics of steroid hormone receptor action. Annu. Rev. Physiol., 42: 17–35, 1980. 2. Henry, J. A., Bennett, M. K., Piggott, N. H., Levett, D. L., May, F. E., and Westley, B. R. Expression of the pNR-2/pS2 protein in diverse human epithelial tumours. Br. J. Cancer, 64: 677– 682, 1991. 3. Liotta, L. A., Goldfarb, R. H., Brundage, R., Siegal, G. P., Terranova, V., and Garbisa, S. Effect of plasminogen activator (urokinase), plasmin, and thrombin on glycoprotein and collagenous components of basement membrane. Cancer Res., 41: 4629 – 4636, 1981. 4. Van Mourik, J. A., Lawrence, D. A., and Loskutoff, D. J. Purification of an inhibitor of plasminogen activator (antiactivator) synthesized by endothelial cells. J. Biol. Chem., 259: 14914 –14921, 1984. 5. Stonelake, P. S., Baker, P. G., Gillespie, W. M., Dunn, J. A., Spooner, D., Morrison, J. M., Bundred, N. J., Oates, G. D., Lee, M. J. R., Neoptolemos, J. P., Chan, S. Y., and Baker, P. R. Steroid receptors, pS2 and cathepsin D in early clinically node-negative breast cancer. Eur. J. Cancer, 30: 5–11, 1994. 6. Stal, O., Brisfors, A., Carstensen, J., Ferraud, L., Hatschek, T., Nordenskjo¨ld, B., and The South-East Sweden Breast Cancer Group. Relationships of DNA ploidy, S-phase fraction and hormone receptor status to tumor stage in breast cancers detected by population screening. Int. J. Cancer, 51: 28 –33, 1992. 7. Alghanem, A. A., and Hussain, S. The effect of tumor size and axillary lymph node metastasis on estrogen and progesterone receptors in primary breast cancer. J. Surg. Oncol., 31: 218 –221, 1986. 8. Silva, J. S., Cox, C. E., Wells, S. A., Jr., Paull, D., Dilley, W. G., McCarty, K. S., Sr., Fetter, B. F., Glaubitz, L. C., and McCarty, K. S., Jr. Biochemical correlates of morphologic differentiation in human breast cancer. Surgery (St. Louis), 92: 443– 449, 1982. 9. Clark, G. M., and McGuire, W. L. Steroid receptors and other prognostic factors in primary breast cancer. Semin. Oncol., 15: 20 –25, 1988. 10. Soubeyran, I., Wafflart, J., Bonichon, F., De Mascarel, I., Trojani, M., Durand, M., Avril, A., and Coindre, J. M. Immunohistochemical determination of pS2 in invasive breast carcinomas: a study on 942 cases. Breast Cancer Res. Treat., 34: 119 –128, 1995. 11. Racca, S., Conti, G., Pietribiasi, F., Stramignoni, D., Tampellini, M., Valetto, M. R., Ghezzo, F., and Di Carlo, F. Correlation between


pS2 protein positivity, steroid receptor status and other prognostic factors in breast cancer. Int. J. Biol. Mark., 10: 87–93, 1995. 12. Predine, J., Spyratos, F., Prud’homme, J. F., Andrieu, C., Hace`ne, K., Brunet, M., Pallud, C., and Milgrom, E. Enzyme-linked immunosorbent assay of pS2 in breast cancers, benign tumors and normal breast tissues. Correlation with prognosis and adjuvant hormone therapy. Cancer (Phila.), 69: 2116 –2123, 1992. 13. Foekens, J. A., Rio, M. C., Seguin, P., Van Putten, W. L., Fauque, J., Nap, M., Klijn, J. G., and Chambon, P. Prediction of relapse and survival in breast cancer patients by pS2 protein status. Cancer Res., 50: 3832–3837, 1990. 14. Foucre´, D., Bouchet, C., Hace`ne, K., Pourreau-Schneider, N., Gentile, A., Martin, P. M., Desplaces, A., and Oglobine, J. Relationship between cathepsin D, urokinase, and plasminogen activator inhibitors in malignant vs. benign breast tumours. Br. J. Cancer, 64: 926 –932, 1991. 15. Foekens, J. A., Schmitt, M., Van Putten, W. L. J., Peters H. A., Kramer, M. D., Ja¨nicke, F., and Klijn, J. G. M. Plasminogen activator inhibitor-1 and prognosis in primary breast cancer. J. Clin. Oncol., 12: 1648 –1658, 1994. 16. Ja¨nicke, F., Schmitt, M., Pache, L., Ulm, K., Harbeck, N., Ho¨fler, H., and Graeff, H. Urokinase (uPA), and its inhibitor PAI-1 are strong and independent prognostic factors in node-negative breast cancer. Breast Cancer Res. Treat., 24: 195–208, 1993. 17. May, E., Mouriesse, H., May-Levin, F., Contesso, G., and Delarue, J. C. 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: 1037–1042, 1989. 18. Ferre´, F., Marchese, A., Pezzoli, P., Griffin, S., Buxton, E., and Boyer, V. Quantitative PCR: an overview. In: K. B. Mullis, F. Ferre´, and R. A. Gibbs (eds.), The Polymerase Chain Reaction. pp. 67– 88. Boston: Birkhauser Boston, 1994. 19. Gotteland, M., May, E., May-Levin, F., Contesso, G., Delarue, J. C., and Mouriesse, H. Estrogen receptors (ER) in human breast cancer. The significance of a new prognostic factor based on both ER protein and ER mRNA contents. Cancer (Phila.), 74: 864 – 871, 1994. 20. Thompson, A. M., Hawkins, R. A., Elton, R. A., Steel, C. M., Chetty, U., and Carter, D. C. pS2 is an independent factor of good prognosis in primary breast cancer. Br. J. Cancer, 68: 93–96, 1993. 21. Barrett-Lee, P. J., Travers, M. T., McClelland, R. A., Luqmani, Y., and Coombes, R. C. Characterization of estrogen receptor messenger RNA in human breast cancer. Cancer Res., 47: 6653– 6659, 1987. 22. Nagai, M. A., Marques, L. A., Yamamoto, L., Fujiyama, C. T., and Brentani, M. M. Estrogen and progesterone receptor mRNA levels in primary breast cancer: association with patient survival and other clinical and tumor features. Int. J. Cancer, 59: 351–356, 1994. 23. Scheuermann, R. H., and Bauer, S. R. Polymerase chain reactionbased mRNA quantification using an internal standard: analysis of oncogene expression. Methods Enzymol., 218: 446 – 473, 1993. 24. Chevillard, S., Mu¨ller, A., Levalois, C., Laine´-Bidron, C., Vielh, P., and Magdele´nat, H. Reverse transcription-polymerase chain reaction

(RT-PCR) assays of estrogen and progesterone receptors in breast cancer. Breast Cancer Res. Treat., 41: 81– 89, 1996. 25. Amoils, K. D., and Bezwoda, W. R. TGF-b1 mRNA expression in clinical breast cancer and its relationship to ER mRNA expression. Breast Cancer Res. Treat., 42: 95–101, 1997. 26. Ishikawa, N., Endo, Y., and Sasaki, T. Inverse correlation between mRNA expression of plasminogen activator inhibitor-2 and lymph node metastasis in human breast cancer. Jpn. J. Cancer Res., 87: 480 – 487, 1996. 27. Tong, D., Schneeberger, C., Leodolter, S., and Zeillinger, R. Quantitative determination of gene expression by competitive reverse transcription-polymerase chain reaction in degraded RNA samples. Anal. Biochem., 251: 173–177, 1997. 28. Grill, H. J., Manz, B., Belovsky, O., and Pollow, K. Criteria for the establishment of a double-labeling assay for simultaneous determination of estrogen and progesterone receptors. Oncology, 41: 25–32, 1984. 29. Scatchard, G. The attraction of proteins for small molecules and ions. Ann. N. Y. Acad. Sci., 51: 660 – 665, 1949. 30. Remmele, W., Hildebrand, U., Heinz, H. A., Klein, P. J., Vierbuchen, M., Behnken, L. J., Heicke, B., and Scheidt, E. Comparative histological, histochemical, immunohistochemical and biochemical studies on oestrogen receptors, lectin receptors, and Barr bodies in human breast cancer. Virchows Arch., 409: 127–147, 1986. 31. Kury, F. D., Schneeberger, C., Sliutz, G., Kubista, E., Salzer, H., Medl, M., Leodolter, S., Swoboda, H., Zeillinger, R., and Spona, J. Determination of HER-2/neu amplification and expression in tumor tissue and cultured cells using a simple, phenol free method for nucleic acid isolation. Oncogene, 5: 1403–1408, 1990. 32. Read, L. D., Greene, G. L., and Katzenellenbogen, B. S. Regulation of estrogen receptor messenger ribonucleic acid and protein levels in human breast cancer cell lines by sex steroid hormones, their antagonists, and growth factors. Mol. Endocrinol., 3: 295–304, 1989. 33. Fuqua, S. A. W., Falette, N. F., and McGuire, W. L. Sensitive detection of estrogen receptor RNA by polymerase chain reaction assay. J. Natl. Cancer Inst. (Bethesda), 82: 858 – 861, 1990. 34. Horiguchi, J., Iino, Y., and Takei, H. Expression of pS2 estrogen-inducible protein in primary breast cancer. Oncology, 53: 12–15, 1996. 35. Speiser, P., Stolzlechner, J., Haider, K., Heinzl H., Jakesz, R., Pecherstorfer, M., Rosen, H., Sevelda, P., and Zeillinger, R. pS2 protein status fails to be an independent prognostic factor in an average breast cancer population. Anticancer Res., 14: 2125–2130, 1994. 36. Grøndahl-Hansen, J., Christensen, I. J., Rosenquist C., Bru¨nner, N., Mouridsen, H. T., Danø, K., and Blichert-Toft, M. High levels of urokinase-type plasminogen activator and its inhibitor PAI-1 in cytosolic extracts of breast carcinomas are associated with poor prognosis. Cancer Res., 53: 2513–2521, 1993. 37. Aaltomaa, S., Lipponen, P., Eskelinen, M., Kosma, V. M., Marin, S., Alhava, E., and Syrjanen, K. Hormone receptors as prognostic factors in female breast cancer. Ann. Med., 23: 643– 648, 1991.