Expression of a Novel Factor, com1, in Early Tumor Progression of Breast Cancer¹

After malignant transformation, tumor cells and their surrounding stroma produce a variety of growth factors and proteolytic enzymes within the tumor cell environment to induce tumor growth, angiogenesis,and degradation of the extracellular matrix to promote tumor progression (1, 2). Moreover, present evidence suggests that metastasis formation primarily is a result of the ability of disseminated tumor cells to initiate and continue growth in the target organ (3, 4). We recently reported a novel approach,comparing the phenotypes of human breast carcinoma cells acquired from an early step in clinical tumor progression and the cell population isolated from experimental metastases formed by these tumor cells, to identify properties that might prevail in the metastatic cells. A factor termed com1 was identified as up-regulated in the fully metastatic cell population. com1 presumably represents a helix-turn-helix type DNA-binding protein and may participate in the response of breast carcinoma cells to growth conditions offered by the target organ upon formation of distant metastases (5).

Tumor cell invasion is dependent on the action of several proteolytic enzymes, including factors of the urokinase pathway. Plasmin is formed from its precursor, plasminogen, by uPA.³ The activity of uPA is regulated by PAI-1 as well as the specific cell surface receptor uPAR (6). Experimental evidence strongly implies that besides the proteolytic activity, uPA in concert with uPAR exerts biological effects characteristic for regulatory molecules involved in tumor cell proliferation and migration during invasion and metastasis (2, 7). In line with this, clinical studies have shown that high expression of uPA and uPAR, and also of PAI-1, by primary breast carcinomas is associated with reduced disease-free and overall survival (6, 8, 9).

The aim of the present study was to further explore the presumably biological role of com1 in tumor growth and metastasis. We examined the expression of com1 mRNA in relation to standard clinicopathological features of primary breast carcinomas to evaluate whether com1 may contribute to the phenotype of tumor aggressiveness defined by elevated expression levels of factors of the urokinase pathway.

Eighty-one patients who had been admitted to two cancer centers in São Paulo, A. C. Camargo Hospital and Instituto Brasileiro de Contrôle do Câncer to undergo mastectomy for breast cancer were randomly selected for this study. Histologically confirmed samples of the primary tumors (of which 94% were infiltrating ductal carcinomas and the remaining lobular or medullar carcinomas) as well as samples of uninvolved breast tissue adjacent to the malignant lesions from 27 of the patients were obtained. The excised specimens were immediately dissected before freezing and storage in liquid nitrogen. Clinical and histopathological characteristics were recorded at the time of primary surgery according to standard diagnostic classification. Patient age ranged from 30 to 86 years, and tumor size ranged from 0.2 to 11 cm. Follow-up information was obtained from patient files at the hospitals. Development of metastatic disease was diagnosed by conventional procedures. Median survival time of the whole group was 4.9 years.

Total RNA was extracted and analyzed by Northern blot technique. Briefly, the samples were pulverized under liquid nitrogen and subsequently lysed in 4 m guanidine isothiocyanate before the lysates were centrifuged through 5.7 m CsCl cushions and purified by phenol-chloroform extraction. Samples of 10μg of RNA were resolved by gel electrophoresis before transfer onto nylon membranes. Integrity of the RNA and equal loading in each lane were verified by ethidium bromide staining. The full-length com1 cDNA was recently described (5). The uPAR cDNA probe was a 1.4-kb Xho fragment (10), whereas the uPA cDNA probe was a 1.5-kb PstI fragment (11), both provided by Dr. R. Mira-y-Lopez at Mount Sinai School of Medicine (New York, NY). The PAI-1 cDNA fragment (12) was provided by Dr. P. A. Andreasen at Aarhus University (Århus,Denmark). The DNA fragment complementary to 18S rRNA (13)was obtained from Dr. N. Arnheim at the State University of New York(New York, NY). The cDNA probes were labeled with[α-³²P]dCTP (Amersham Pharmacia Biotech,Little Chalfont, United Kingdom) by random priming technique, and standard hybridization conditions (50% formamide, 42°C) were used. Absorbance was measured and quantified (in arbitrary units) by laser densitometry of suitably exposed autoradiographs.

Statistical analyses were performed using the Stata Statistical software program (Stata Corporation, College Station, TX)with a significance level of P < 0.05. The distribution of mRNA expression levels of com1, uPAR, uPA, and PAI-1 in tumor samples and normal tissue specimens as well as the correlation of tumor expression values to conventional prognostic variables were calculated by the Mann-Whitney U test. A log-rank test was applied to compare correlations between expression levels of the different mRNAs and the Kaplan-Meier technique to calculate overall survival probability.

Eighty-one specimens of breast cancer and nonneoplastic tissue samples from 27 of the mastectomies were analyzed for expression of com1 mRNA as well as mRNAs for uPAR, uPA, and PAI-1. Representative Northern signals from paired tumors and the adjacent normal breast tissues are displayed in Fig. 1. The mRNA for com1 (0.7 kb) was expressed in all tumor samples but not in every normal tissue sample analyzed. This expression pattern applied also to uPAR mRNA (1.4 kb), whereas mRNAs for uPA (2.5 kb) and PAI-1 (3.3 and 2.4 kb) were detected in all tissue samples, either malignant or nonneoplastic. For analyses of expression levels, the sum of values(densitometry of signal intensities) obtained from both PAI-1 mRNA species was used.

The expression levels were calculated relative to the level of 18S rRNA for each individual sample, and the distribution of values from the 81 tumor samples was compared to the values from the 27 normal tissue specimens. As indicated by Fig. 1, the expression of com1 mRNA was consistently lower in the normal tissues than in the tumors, and the distribution of values between the nonneoplastic and malignant tissues was statistically different (P <0.0001; Fig. 2). Similar expression patterns were observed for the mRNAs for uPAR and uPA (Fig. 1),resulting in significant differences in the distribution of these mRNA values as well (P = 0.0001 and P =0.01, respectively; Fig. 2). In contrast, no difference was found when comparing PAI-1 mRNA values in tumors and in the normal tissue specimens (Fig. 2).

In view of these findings, correlations between different mRNA levels in the tumors were compared to test whether com1 expression might essentially covariate with the others. But as seen from Fig. 3, A-C, the expression of com1 mRNA in the individual tumors did not correlate with the mRNAs for either uPAR, uPA, or PAI-1. For comparison, the individual mRNA levels for uPAR and uPA in the tumors were analyzed, and these showed parallel variation (r = 0.73, P < 0.0001; Fig. 3,D). Similar but less stringent correlations in expression levels were also found between PAI-1 mRNA and the mRNAs for uPAR and uPA (r = 0.40, P = 0.0002 and r = 0.41, P = 0.0001, respectively;Fig. 3, E and F).

Next, the mRNA expression levels of com1, uPAR, uPA, and PAI-1 in the tumors were correlated to several clinicopathological features. As seen from Table 1, there were no significant correlations between these mRNA levels and patient age or menopausal status. Neither were any associations found with histological and biochemical characteristics of the tumors such as size, the presence of vascular infiltrate or necrosis, or (for all mRNAs except uPAR) steroid receptor status, nor with disease parameters such as lymph node status, stage of the disease at diagnosis, or development of metastatic disease during follow-up. In contrast, the levels of mRNAs for uPAR and uPA were significantly lower in the tumors from patients who were alive at the end of follow-up compared with those who were not (P = 0.04 and P =0.03, respectively). A similar correlation to survival was, however,not found for the mRNAs for com1 or PAI-1.

Finally, the impact of tumor expression of these factors on overall survival probability was evaluated. The 5-year overall survival rate was 50% (Fig. 4,A). The patients were divided into two groups with mRNA expression levels of each factor below and above the respective median values. A high uPAR mRNA level was significantly associated with shorter overall survival(P = 0.01; Fig. 4,C). A similar but statistically nonsignificant display was observed for uPA mRNA (Fig. 4,D). Yet again, differences in mRNA expression of com1 or PAI-1 among the tumors did not have any impact on survival (Fig. 4, B and E).

In the present study, we showed that the expression levels of com1 mRNA in the breast tumors were consistently and significantly higher than in the adjacent normal breast tissue, whereas no evidence of correlations between tumor levels of com1 mRNA and conventional prognostic factors was found. On the other hand, the mRNA levels of uPAR and uPA were elevated in the tumors from patients who subsequently had poor outcome, in agreement with reports by others (6, 8, 9) and implying that our material is a reliable tool to study the regulatory biology of other potential biomarkers, such as com1.

We recently identified com1 as a factor that may mediate the growth response of breast carcinoma cells upon establishment in secondary organs during experimental metastasis formation (5). Based on these experimental results, it is rather intriguing that the tumor level of com1 did not correlate with outcome or any other prognostic variables in this clinical study. But because com1 mRNA values in tumors were significantly higher than in normal breast tissues, com1 may still be assumed to mediate some growth stimulatory signal(s) from the stroma to the carcinoma cells in the early primary tumor, but not in later stages of tumorigenesis or metastasis. Actually, com1 levels seemed to be lower, rather than higher, in the more aggressive tumors (e.g., primary tumors with vascular infiltrate), although not significantly different by statistical measurement (P = 0.08). This might indicate that tumors with high propensity for angiogenesis do not need elevated com1 expression or, vice versa, that com1 may somehow compensate for the lack of other features of tumor aggressiveness.

The gene sequence for com1 is earlier identified as the human gene p8 by Vasseur et al. (14) with a localization within chromosome 16 at position p11.2. This chromosomal region is occasionally amplified in primary breast tumors (15, 16, 17). This supports the assumption that com1 may function as a growth-promoting factor in the development of malignant breast tumors, a notion also suggested by Vasseur et al.(14).

Breast cancer patients with high levels of uPAR, uPA, and/or PAI-1 in the primary tumors have an increased risk of relapse and death (6, 8, 9). In our material, the mRNA levels of uPAR and uPA, but not of PAI-1, were elevated in tumors from the patients who subsequently had poor outcome. Some investigators have shown that a high level of tumor PAI-1 may identify the subset of patients with increased risk of relapse among node-negative breast cancer patients (18, 19) and that the PAI-1 level may display a time-dependent variation in prognostic strength (19, 20). We did not, however, distinguish between node-negative and node-positive patients or between different periods of follow-up in our outcome analyses.

Early exploratory studies are mainly conducted to evaluate a presumable association between a potentially useful biomarker and disease characteristics and not to discriminate between patients at high and low risk of disease progression (21). For this reason, and because the number of cases available (n =81) was rather low, we did not analyze specific subgroups of the patients in this preliminary study. Nevertheless, the following methodological considerations can be made. First, although the material was randomly selected, a bias on the background of selection of patients cannot be excluded. For instance, we have no information about outcome for 14 of the patients included from one of the hospitals because some subsets of patients admitted there unfortunately do not follow the general instructions for proper follow-up of their malignancies. Next, Northern blot analysis does not distinguish between functionally different cell types of the tumor. com1 was originally identified as specifically expressed by lobular breast carcinoma cells with a complete lack of expression by the surrounding stroma (5). In the present study, we have not addressed the question of whether ductal cells also preferentially express com1 and if therefore the apparent up-regulation of com1 mRNA in the breast tumors is attributable to the selective proliferation of this epithelium in the ductal carcinomas. Finally, data that are statistically reliable may need the definition of optimal cutoff points to discriminate between functionally low and high levels of the factors studied. In our view, however, it would be incorrect to determine cutoff points for continuous mRNA values that are measured in arbitrary units, a precaution strongly recommended to undertake (21).

In conclusion, this report showed significant up-regulation of a novel factor, termed com1, in primary breast carcinomas, presumably as a result of a growth response of the tumor cells following malignant transformation. No evidence of correlation between tumor com1 expression and prognosis was found, implying that com1 may participate in early stages of breast carcinoma development. Future studies of com1 function will include the identification of growth factors that induce com1 expression in breast tumors.

We acknowledge the contributions of cDNAs for uPAR and uPA from Dr. R. Mira-y-Lopez at Mount Sinai School of Medicine (New York,NY), the PAI-1 cDNA from Dr. P. A. Andreasen at Aarhus University(Århus, Denmark), and the 18S rRNA probe from Dr. N. Arnheim at the State University of New York (New York, NY).