| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Chair of Obstetrics and Gynecology (F.M.R., L.C., S.L., W.G., F.P.), University of Siena, 53100 Siena, Italy; Department of Obstetrics and Gynecology, UFMG (F.M.R.) and Hospital Maternidade Odete Valadares (L.C.T.), 30130-100 Belo Horizonte, Brazil; Department of Surgery (G.A.), University of Ferrara, 44100 Ferrara, Italy; Department of Physiology (I.S.B.S.), UFRGS, 90050-170 Porto Alegre, Brazil; and Laboratory of Molecular Biology (A.M.d.B.), Istituto Auxologico Italiano, 20135 Milan, Italy
Address all correspondence and requests for reprints to: Felice Petraglia, M.D., Chair of Obstetrics and Gynecology, University of Siena, Policlinico Le Scotte, Viale Bracci, 53100 Siena, Italy. E-mail: . petraglia{at}unisi.it
Abstract
Activins are growth factors involved in the control of cell proliferation and differentiation. Human breast tissues express immunoreactive activin subunits, and activin A is able to inhibit the replication of mammary cells in vitro. The aim of the present study was to evaluate 1) whether breast cancer expresses activin ßA subunit mRNA, 2) whether serum activin A levels are altered in postmenopausal women with breast cancer, and 3) how circulating activin A levels change after tumor removal. Four groups of women (n = 158) were enrolled for the present prospective study: two groups were composed of postmenopausal women with breast cancer (n = 74) or benign lesions (n = 15); the third was a control group composed of healthy postmenopausal women (n = 62); and the fourth group included healthy fertile women (n = 7) undergoing plastic surgery with removal of non-neoplastic mammary tissue.
RT-PCR showed that ßA subunit mRNA was expressed in breast carcinoma, fibroadenoma, and normal mammary tissue, and the level of expression was higher in carcinoma than in normal tissue (P < 0.05). Dimeric activin A was detectable in homogenates of breast cancer tissue at concentrations twice as high as in non-neoplastic tissue (P < 0.01). In women with breast cancer, median serum activin A levels were significantly higher than in controls (P < 0.001). The high serum activin A levels in patients with breast cancer were not correlated with the presence of lymph node metastasis, tumor grade, or tumor diameter. After tumor excision, a significant decrease of activin A in the first and second postoperative days was observed (P < 0.01; Friedman’s ANOVA). Conversely, activin A levels remained unchanged after plastic surgery in healthy women. The present results suggest that activin A is expressed and secreted in postmenopausal women with breast cancer. The pathophysiological and possible clinical implications of this finding remain to be investigated.
ACTIVINS ARE GLYCOPROTEINS belonging to the transforming growth factor ß superfamily. They are composed of two ß subunits, ßA and/or ßB, which form activin A (ßA-ßA), activin B (ßB-ßB) or activin AB (ßA-ßB). Originally isolated from the ovarian follicular fluid, activins are also produced by extragonadal tissues, such as pituitary gland, adrenal, brain, bone marrow, and placenta (1), where they modulate cell proliferation and differentiation. Recent studies showed the expression of activin subunits in human mammary (2), uterine (3), and prostate (4) tissues.
The expression of activins has been demonstrated in various tumors. The high levels of serum activin A associated with solid tumors (3, 5) suggest that this protein is secreted by the tumors into systemic circulation. The lack of changes between fertile and postmenopausal life excludes a major ovarian contribution to circulating activin A levels (6, 7).
Breast cancer is one of the leading causes of mortality among women in Western countries. In the United States, it is the most common and the second most lethal female cancer (8). Despite a great effort to develop noninvasive methods of screening, staging, and follow-up of patients with breast cancer, a reliable serum marker remains unavailable (9).
After our observation of immunoreactive activin subunits in human breast tissue (2), the breast cancer tissue expression of activin A mRNA and protein was now investigated. Furthermore, the possible secretion of activin A in general circulation was evaluated in postmenopausal women with breast cancer or benign breast nodules. The possible tumoral origin of the circulating activin A was investigated by observing the effect of tumor removal on its serum concentrations. Finally, the sensitivity of activin A measurement in the diagnosis of breast tumors in postmenopausal women was evaluated.
Materials and Methods
Patient selection and study design
Four groups of women (n = 158) were enrolled prospectively for the present study at three different medical centers. The first two groups were composed of postmenopausal women (at least 1 yr of amenorrhea and compatible serum E2 and FSH levels) (Table 1
) consulting for a breast nodule. Inclusion criteria were availability of complete breast investigation, including histological diagnosis of the breast lesion, and absence of other neoplasms. Patients were submitted to thorough clinical and radiological examination and to a biopsy of the lesion, followed by surgical treatment. After histopathological diagnosis of the excised tissues, these patients were classified according with the presence of breast cancer (n = 74) or only benign lesions (n = 15). The third group (n = 62) was a control group composed of postmenopausal women attending a menopause clinic, whose physical and radiological examination ruled out any detectable breast disease. Additional exclusion criteria were neoplasm of any type and current use of hormone therapy. A fourth group composed of healthy fertile women (n = 7) undergoing removal of non-neoplastic mammary tissue for plastic breast reduction was also included.
|
The study was authorized by the review boards of the participant institutions, and all patients provided their informed consent to participate.
Analysis of activin ßA subunit mRNA
The analysis of mRNA for the activin ßA subunit was performed in breast carcinoma (n = 5), fibroadenoma (n = 7), and normal breast tissue (n = 5). Tissue samples were obtained during surgical intervention and were immediately frozen in liquid nitrogen until RNA extraction.
Tissue samples were homogenized in phenol-guanidine isothiocyanate, and total RNA was extracted with chloroform and precipitated with isopropanol by 12,000 x g centrifugation at 4 C. To digest any contaminant genomic DNA, RNA samples were treated with 5 U DNase (Promega RQ1) at 37 C for 10 min, and precipitated with 0.1 M sodium acetate (pH 5.2) and 80% ethanol by centrifugation at 12,000 x g. The RNA pellet was washed twice with 75% ethanol, resuspended in diethylpyrocarbonate-treated water, and quantified by light absorbance at 260 nm.
RT-PCR was carried out using the GeneAmp RNA PCR kit purchased from Perkin-Elmer Corp. (Roche Molecular Systems, Branchburg, NJ). First strand cDNA was synthesized from 2 µg total RNA. After denaturing the template RNA and primers at 70 C for 10 min, 50 U reverse transcriptase was added in the presence of buffer II [50 mM KCl, 10 mM Tris-HCl (pH 8.3)], 2.5 mM MgCl2, 0.5 mM deoxynucleoside triphosphate mix, and 20 U Rnase inhibitor. The mixture (20 µl) was incubated at 42 C for 55 min, then heated at 70 C to stop the reaction, and stored at -20 C.
PCR was carried out in a final volume of 50 µl. Two microliters of the first strand synthesis reaction were incubated with buffer II, 1.5 mM MgCl2, 0.2 µM sense and antisense primers, 0.2 mM deoxynucleotide triphosphate mix, and 1.25 U Taq DNA polymerase. The specific primers used to amplify a cDNA fragment corresponding to activin ßA subunit were: sense, 5' GTT TGC CGA GTC AGG AAC AG 3'; and antisense, 5' GAG GTT GGC AAA GGG GCT ATG GCC CCG CAT 3'. These primers generate a 787-bp fragment corresponding to bases 611-1397 of the human ßA sequence (GenBank accession no. M13436). PCRs consisted of 30 thermal cycles of 94 C for 30 sec, 56.5 C for 30 sec, and 72 C for 60 sec. The number of PCR cycles was chosen to stay in the exponential phase of amplification as determined by pilot experiments ranging from 23–35 cycles. The product of a first strand reaction performed without reverse transcriptase was also submitted to the PCR protocol to serve as negative control. A sample of the PCR mixture (15 µl) was resolved on a 3% agarose gel stained with ethidium bromide and photographed under UV light.
The identity of the PCR product was further confirmed by restriction enzyme digestion. The fragment corresponding to the ßA subunit has a unique restriction site for PstI. The PCR products were precipitated by incubation at -20 C for 10 min, followed by centrifugation at 12,000 x g for 15 min in the presence of dextran, 3 M sodium acetate (pH 5.2), and 99% ethanol. The pellet was washed with 75% ethanol and resuspended in 15 µl sterile water. The endonuclease (20 U PstI, New England Biolabs, Inc., Beverley, MA) was added together with the accompanying buffer and bovine albumin to a final volume of 20 µl and incubated at 37 C for 90 min. The digested products were size fractioned in a 3% agarose gel run at 76 V for 45 min.
Semiquantitative comparison of activin ßA subunit mRNA expression in breast carcinoma, fibroadenoma, and normal breast tissue was obtained by densitometric analysis of the specific bands using an image-processing system (ImageMaster VDS, Pharmacia Biotech, Uppsala, Sweden). Amplification of a 623-bp cDNA fragment corresponding to the ubiquitously expressed protein ß2-microglobulin was used to adjust PCR results to equivalent amounts of cDNA loaded (10).
Tissue activin A associated with breast cancer
The expression of dimeric activin A was evaluated in breast carcinoma (n = 10) and non-neoplastic breast tissue (histologically normal mammary tissue obtained from mastectomy; n = 5). Tissue samples were obtained during surgical intervention and were rinsed with cold PBS, dissected free of gross blood vessels and obvious connective tissue, then cut into 2- to 5-mm pieces, and homogenized in PBS. The homogenate was centrifuged for 2 min at 5000 x g, and the supernatant was stored at -20 C until assayed for activin A and total protein concentrations.
Serum activin A associated with breast cancer
A blood sample was collected from all patients in the morning (between 0800 and 1000 h) at hospital admission the day before the intervention or at a random visit to the menopause clinic (control group). Samples were allowed to clot and centrifuged at room temperature, and the serum was stored at -20 C until the activin A assay.
Changes of serum activin A after surgery for breast cancer
A subgroup of women with breast cancer (n = 21) was submitted to blood sampling also in the first and second days after mastectomy or quadrantectomy. To verify whether these changes could be attributed to the removal of non-neoplastic mammary tissue rather than by the excision of the tumor, the same schedule of blood sampling was applied to healthy women undergoing removal of healthy mammary tissue for plastic breast reduction (n = 7). The sequential serum samples of each patient were assayed for activin A in the same plate, thus eliminating the effect of interassay variation.
Activin A assay
Serum activin A concentrations were measured in duplicate using a specific two-site enzyme immunoassay purchased from Serotec (Oxford, UK) as previously described (3). The limit of detection for activin A was less than 0.1 ng/ml. The intra- and interassay coefficients of variation for quality control samples (
1.0 ng/ml) were 3 and 9%, respectively.
Statistical analysis
Data were tested for normality (symmetry and kurtosis) and for homogeneity of variances. Because the requirements of conventional ANOVA were not met, differences between groups were assessed by nonparametric tests of Mann-Whitney U or Kruskal-Wallis, as appropriate, the latter being followed by Dunn’s multiple comparisons whenever overall differences were present. The results are described as medians and interquartile ranges. Friedman’s ANOVA was used to analyze repeated measures. The performance of activin A measurement as a diagnostic test was evaluated by calculating sensitivity and likelihood ratios with 95% confidence intervals (11). Significance was set at P less than 0.05.
Results
Tissue expression of activin A mRNA and protein
Activin ßA subunit was expressed in breast carcinoma, fibroadenoma, and normal mammary tissue of postmenopausal women. The specificity of these findings was confirmed by the absence of amplification of negative control in the RT-PCR, as well as by the recovery of the expected fragments after restriction enzyme digestion (Fig. 1). A semiquantitative comparison of activin ßA subunit mRNA expression in breast carcinoma, fibroadenoma, and normal breast tissue revealed an approximately 43% increase in the relative abundance of activin ßA mRNA in breast carcinoma compared with normal breast tissue (P < 0.05; Fig. 2). Fibroadenomas expressed the activin ßA mRNA at the same level as normal mammary tissue.
|
|
). Quantitatively, the expression of activin A was significantly higher in carcinoma [2.26 (1.70–3.14) ng/mg protein] than in non-neoplastic breast tissue [1.17 (1.03–1.23) ng/mg protein; P < 0.01].
|
Median serum activin A levels were slightly increased in subjects with benign lesions [0.66 [(0.53–0.85) ng/ml] and significantly higher in those with breast cancer [0.72 (0.53–0.92) ng/ml; P < 0.001] than in healthy controls [0.56 (0.34–0.74) ng/ml; Fig. 4].
|
).
|
The hypothesis that the breast tumor and/or peritumoral tissues were the main sources of activin A in patients with breast cancer was explored by determining the effect of surgical treatment on serum activin A levels in the first and second postoperative days. After the removal of breast cancer, activin A concentrations decreased from 0.89 (0.63–1.09) ng/ml before surgery to 0.57 (0.37–0.92) ng/ml by the second postoperative day (P < 0.01; Friedman’s ANOVA; Fig. 5A). Notwithstanding that most patients (15 of 21) showed a consistent decrease of serum activin A levels after surgery, there was a subgroup of six women whose activin A concentrations did not decrease (Fig. 5A). This subgroup of patients had similar age, tumor staging, and surgical management (proportion of radical vs. conservative interventions) as those patients whose activin A levels decreased consistently. The control group of women submitted to plastic surgery had serum activin A levels that were unchanged (Fig. 5B).
|
After setting an arbitrary definition of high activin A concentration as higher than the 90th percentile of the control group, which yields a 90% specificity, the sensitivity for detection of benign lesions and for detection of malignant lesions was calculated. Similar calculation was performed to test the differential diagnosis of malignant from benign conditions using as a cutoff the 90th percentile of the benign lesion group. As shown in Table 3, activin A had low sensitivity to detect benign [0.20 (0.00–0.40)] or malignant [0.30 (0.21–0.39)] lesions, and also in the differential diagnosis of breast cancer against benign lesions using the 90th percentile of the benign lesion group as cutoff point [0.20 (0.11–0.29)]. The area under the receiving operator characteristics curve for the diagnosis of breast cancer using serum activin A was 0.697.
|
The data presented herein may be summarized as follows: 1) activin ßA subunit mRNA and dimeric activin A are expressed in normal and neoplastic breast tissues; 2) serum levels of activin A are elevated in postmenopausal women with breast cancer; 3) activin A levels frequently decline after surgical excision of the breast tumors, whereas no significant change is associated with the removal of mammary tissue by plastic surgery in healthy women. In addition, the analysis of serum activin data as a screening test to breast cancer in postmenopausal women with clinically detectable disease showed a low sensitivity, at 90% specificity.
The present results suggest that breast cancer is able to produce activin A, as shown by the specific mRNA and protein detection in tumor samples. Other cancers express ßA mRNA and protein, such as ovarian (12), endometrial (3), and prostate (4) carcinomas, and other types of liver (13) and pancreatic (14) tumors. In a previous study using immunohistochemistry, we have not been able to detect ßA antigen in breast carcinoma, whereas Ying and Zhang (15) used an antiserum against a different region of the ßA subunit and successfully detected immunoreactive ßA in MCF-7 cells, a human breast adenocarcinoma cell line. It is possible that our antibody was less sensitive than that used by Ying and Zhang, but it may also be suggested that cultured MCF-7 cells express the ßA subunit to a greater extent than carcinoma cells in vivo. Although these remain merely speculative interpretations, the present demonstration of activin A mRNA and protein in carcinoma tissue homogenates expands previous observations in vitro (15) and further suggests that activin A may be locally produced in breast cancer.
The main source of high circulating levels of activin A in women with breast cancer is probably the tumor itself. In favor of this hypothesis, activin A levels promptly decreased after breast surgery, suggesting that a major source had been removed. In addition, activin A mRNA and protein were more abundant in homogenates of breast carcinoma than in normal breast tissue. One remaining question is whether the increased tissue expression denotes an increased activin A production by neoplastic cells or only reflects the higher concentration of mammary epithelial cells (the only cell type in which activin A has been detected thus far) in carcinoma tissue compared with normal postmenopausal breast tissue. The former hypothesis is not supported by immunohistochemistry (2), thus further studies are needed to assess the mechanism of increased activin A expression in these tumors.
The presence of circulating activin A by the second postoperative day indicates the existence of residual sources, either physiological, as observed in healthy postmenopausal women, or associated with the disease. In addition to the primary tumor, there are other potential sources that might contribute to a rise in activin A production in breast cancer patients, such as inflammatory cells, bone marrow, adrenal cortex, and liver (16), and also metastatic tumor tissue. From the present findings, it appears that some women had only physiological sources remaining after tumor removal because their activin A levels returned to the normal range, whereas others still had elevated activin A levels, suggesting the persistence of some pathological hormone production. However, any possible relationship between elevated postoperative activin A levels and the presence of residual disease still has to be investigated.
Besides malignant cells, there are other potential sites of activin A production in the breast because ßA mRNA and protein have been localized in the epithelial components of non-neoplastic mammary tissues (2) and in selected populations of normal human breast cells (17), whereas the present study shows that ßA mRNA is detectable in homogenates of fibroadenoma as well as normal breast tissue. The present data also suggest that the contribution of healthy mammary tissue to the circulating activin A pool is likely to be modest, because breast reduction by plastic surgery did not induce any significant change in serum activin A levels in healthy women. Because the cohort of plastic surgery patients was composed of cycling women, we cannot exclude the possibility that the presence of the ovary has masked a slight impact that removal of breast tissue might have had on serum activin A levels.
The possible implications of activin A in breast cancer may be as an antiproliferative agent and/or immunomodulatory factor. Indeed, it has been shown that activin A inhibits the proliferation of breast cancer cell lines in vitro (17, 18). Furthermore, the antiproliferative effect of human CG on immortalized human breast epithelial cells is accompanied by the immunocytochemical expression of activin ß subunits (19). This antiproliferative property of activin A has also been demonstrated in human tumoral cell lines derived from cholangiocarcinoma (20) and prostate carcinoma (21) and in primary cell cultures obtained from pituitary tumors (22). Nevertheless, whether activin A plays a physiological role in tumor suppression and whether an imbalance of activin A production is involved in the development of any of these tumors is still unknown.
Regarding the possible immunomodulatory role, it is known that peritumoral breast tissues might respond to the benign and malignant breast tumor by triggering immune responses, including mononuclear infiltration and a complex cascade of cytokines (23). Because activin A affects cell- mediated immunity by modulating monocyte chemotaxis, monocyte migration, and cytokine production (1, 24, 25, 26), it is conceivable that it might play a role in the local immune response to the tumor.
The other possible clinical implication was to analyze whether high activin A levels in postmenopausal women with breast cancer could be used as serum marker for the diagnosis of the disease. An arbitrary cut point was adopted to define a positive test at the 90th percentile of activin A concentrations in unaffected groups, which means a 90% specificity. Using this approach, it resulted that sensitivity was low, suggesting that activin A is unlikely to be a useful screening marker for breast cancer in postmenopausal women. Another potential use of activin A as a serum marker may be for recurrence surveillance, provided the immediate postoperative concentration is recorded to serve as a reference value for each individual. This hypothesis requires further investigation.
In conclusion, the present results suggest that activin A is expressed at both tissue and serum in postmenopausal women with breast cancer. Tissue expression is evidenced by the presence of activin ßA mRNA, and the contribution of tumoral tissue to the circulating protein is suggested by a marked decrease of serum activin A levels immediately after tumor resection. Although on average activin A levels are higher in patients with breast cancer, there is considerable overlapping between concentrations measured in cancer bearers and healthy women, which reduces the potential usefulness of this marker to the screening of breast cancer. The pathophysiological and possible clinical implications of these findings remain to be investigated.
Acknowledgments
We thank Drs. R. Cericatto, M. Stomati, and F. Arcuri for providing breast tissue samples and Dr. P. Ciarmela for her assistance in part of the biochemical procedures.
Footnotes
This work was supported in part by scholarships from Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (0471/98-25) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (301006/99-7), Brazil (to F.M.R.).
Received April 24, 2001.
Accepted February 11, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  J. E Burdette and T. K Woodruff Activin and estrogen crosstalk regulates transcription in human breast cancer cells Endocr. Relat. Cancer, September 1, 2007; 14(3): 679 - 689. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. Burdette, J. S. Jeruss, S. J. Kurley, E. J. Lee, and T. K. Woodruff Activin A Mediates Growth Inhibition and Cell Cycle Arrest through Smads in Human Breast Cancer Cells Cancer Res., September 1, 2005; 65(17): 7968 - 7975. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Cobellis, F. M. Reis, S. Luisi, S. Danero, L. Pirtoli, G. Scambia, and F. Petraglia High Concentrations of Activin A in the Peritoneal Fluid of Women With Epithelial Ovarian Cancer Reproductive Sciences, May 1, 2004; 11(4): 203 - 206. [Abstract] [PDF]
|
|
|
|
|
|
J. S. Jeruss, C. D. Sturgis, A. W. Rademaker, and T. K. Woodruff Down-Regulation of Activin, Activin Receptors, and Smads in High-Grade Breast Cancer Cancer Res., July 1, 2003; 63(13): 3783 - 3790. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
