Abstract
Background: β–Microseminoprotein (MSMB) is an abundant protein in seminal plasma. Most of it is present as a free protein but a small part is bound to cysteine-rich secretory protein 3 (CRISP3) as a non-covalent complex. Even though their physiological function is unknown, both MSMB and CRISP3 have been ascribed roles in prostate carcinogenesis. Thus, several recent experimental studies indicate a tumor-suppressor role for MSMB. The present study was undertaken in order to evaluate, for the first time, the expression of MSMB and CRISP3 in ovaries and in ovarian tumors and to determine if their expression might indicate a role in ovarian tumor development. Materials and Methods: Biopsies from prospectively collected samples from ovaries and benign, borderline and invasive ovarian tumors were analyzed for expression of MSMB and CRISP3 by immunohistochemistry. In patients with ovarian cancer the expression was compared to survival. Results: Both MSMB and CRISP3 were strongly stained in ovarian epithelial cells and weakly stained in the stroma. In ovarian blood vessels, CRISP3 exhibited strong to medium staining, while MSMB was only weakly expressed. In benign and borderline tumors the staining pattern was similar to the one observed in the ovaries. In invasive neoplasms, the expression of MSMB in the tumor cells was significantly reduced. In univariate analysis, decreased expression of MSMB correlated to reduced survival. No correlation was found with stage, the strongest prognostic indicator for ovarian cancer, which supports an independent role of MSMB in ovarian carcinogenesis. For CRISP3, a staining pattern comparable to that for MSMB was observed in all groups, except the fact that decreased expression was not observed in invasive tumor cells. Conclusion: MSMB and CRISP3 were widely distributed in ovaries and in ovarian tumors; the expression of MSMB fits well with a tumor-suppressor function in ovarian carcinogenesis.
Ovarian cancer has the highest mortality rate of all types of cancers of the female reproductive system. In spite of advances in cytoreductive surgery and chemotherapy, the overall 5-year survival has only marginally increased (from 37% to 45%) during the last three decades (1). Surgery has become more centralized and specialized and several drugs in almost any combination have been used in numerous studies. Despite this, according to a recent conference report, progress in survival seems to have stopped, which makes development of new strategies necessary (2). To achieve this, a detailed knowledge on tissue biology during ovarian carcinogenesis is important. However, although reports (3, 4) have described the possible contribution of several growth-regulatory factors, a complete model of the molecular changes that occur during tumor development still seems far away. MSMB, also known as prostatic secretory protein, of 94 amino acids is one of the three major proteins secreted by the prostate gland (5). Remarkably, the functions of MSMB in the gland and in sperm are still unknown. MSMB is produced mainly by the prostate gland but has also been found in other tissues (6, 7). MSMB mRNA has been described in the ovaries (8), but to our knowledge, no further studies concerning its expression or its possible role in ovarian physiology or pathology have been published.
CRISP3 was first isolated from neutrophilic leukocytes (9, 10), but is also found in exocrine secretions such as saliva, pancreatic fluid and seminal plasma. In seminal plasma, it is present as a non-covalent complex with MSMB (11). Like MSMB, the physiological function of CRISP3 is unknown, as is the significance of the binding between MSMB and CRISP3. Patients with prostate cancer positive for CRISP3 in the tumor tissue were less likely to remain recurrence-free (12). In the current work, we studied the expression of MSMB and CRISP3 with immunohistochemistry in ovaries and in benign and malignant ovarian tumors and correlated the presence of these proteins with tumor development.
Materials and Methods
Patient samples. Fresh surgical specimens were collected prospectively and frozen immediately, and kept at −70°C until analyzed. The patient material was anonymized according to the guidelines of the Malmö-Lund Ethical Committee. Biopsies were obtained from 8 normal ovaries, 13 benign, 5 borderline tumors and 37 invasive epithelial ovarian carcinomas. There were 12 in stage I, 4 in stage II, 16 in stage III, 2 in stage IV and for 3 patients the stage was unknown. None of the patients with cancer had been subject to treatment prior to surgery. Survival data were obtained for all patients except three.
Antibodies and immunohistochemistry. The rabbit antisera raised against human MSMB and the aminoterminal part of recombinant human CRISP3 has been described previously (7, 9). Immunostaining was performed as described previously (13), with minor modifications. Six-micrometer-thick cryosectioned tissue sections fixed in 4% paraformaldehyde were treated for 30 min with 0.3% hydrogen peroxide to block endogenous peroxidase activity. Endogenous avidin-binding activity was blocked by sequential incubation with avidin and biotin using a blocking kit (Vector Laboratories, Burlingame, CA, USA). Non-specific binding of the secondary antibody was blocked by incubation with normal goat serum diluted 1:10 with phosphate buffered saline. After incubation with the primary antibody (diluted 1:1000) overnight at 4°C in a humidified chamber, the tissue sections were incubated with biotinylated goat-anti rabbit immunoglobulin (Vector Laboratories). The immunoreaction was visualized with a Vectastain Elite ABC complex (Vector Laboratories) using 0.02% hydrogen peroxide as a substrate and 3,3’-diaminobenzidine tetrahydrochloride as a chromogen (Dako, Glostrup, Denmark). Finally the sections were briefly counterstained for 15 sec in Meyer's hematoxylin. The staining was estimated as being strong if more than two thirds of the cells were immunoreactive, intermediate if one third to two thirds stained positively and weak if less than one third of the cells were positively stained.
Statistical methods. Differences in protein expression between the groups were estimated with the two-tailed Fisher's exact probability test (14). The survival times were assessed from the day of surgery. Cumulative survival curves were constructed according to the Kaplan Meier method (15) and differences in survival were estimated with the log-rank test (16).
Results
MSMB was strongly stained in ovarian epithelium in all eight samples (Table I). Medium and weak staining was observed in the stroma in four cases each and in one case each for blood vessels. In 9 out of 10 benign ovarian tumors, MSMB was strongly expressed in tumor cells while one exhibited intermediate staining. Tumor cells from five borderline tumors all exhibited strong expression. In invasive neoplasms, tumor cells in 25 out of 37 biopsies exhibited a weak to intermediate staining, which deviated significantly from the staining of ovaries (p=0.01) and benign (p=0.005) and borderline tumors (p=0.05) (Table II). Examples of the staining in various samples are shown in Figure 1. Survival related to staining was estimated by dividing the material into three groups, strong staining (9 out of 12 patients with available survival data), intermediate staining (10 patients) and weak or absent staining (15 patients). A Kaplan Meier curve showed that decreased survival correlated to decreased staining, with the shortest survival in the group with most reduced staining and intermediate survival in the group having intermediate staining. A log rank test confirmed the significant difference between the groups having high and low expression (p=0.02) (Figure 3). In the stroma and blood vessels, the expression of MSMB was generally weak or absent, presenting no difference between the groups. Immunostaining for CRISP3 was strong in ovarian epithelial cells, as well as in benign and malignant tumor cells, with a minor non-significant reduction in the invasive neoplasms (Table I and II, Figure 2). In the stromal part, weak or absent staining was observed. In the blood vessels, strong staining was seen in the ovaries, but staining was weak or absent in the tumor groups.
We found no correlation for neither MSMB nor CRISP3 expression to stage or to earlier published expression of p53 (17), Ki67, S-phase fraction or DNA ploidy (18) (results not shown).
Discussion
Our study on the expression of MSMB and CRISP3 in normal and malignant ovarian tissue contributes several new observations. Thus, both proteins were for the first time described in ovaries as well as in ovarian tumors. MSMB was primarily localized to epithelial cells and benign tumor cells with only a weaker or absent staining in stroma and in blood vessels. Furthermore, MSMB immunoreactivity was equally strong in borderline tumors cells. However, in invasive cancer cells, a significantly reduced expression was observed which was correlated to a reduced survival. Although the analysis was univariate, the significance is supported by the observed lack of correlation of MSMB to stage, the strongest prognostic parameter in ovarian cancer. Neither was the expression correlated to other prognostic factors such as cell proliferation and ploidy. Thus, when the malignant cells acquired invasive potential, MSMB expression decreased or did not appear the cells. Although not a proof, this pattern fits well with a suppressor function of MSMB and is in line with recent observations in prostate cancer, as described below.
Initially, MSMB was described as one of three major proteins in seminal plasma with an unknown biological role. Later it was found in other tissues, including genital organs. In recent years, MSMB has been coupled to the development of prostate cancer. Thus, several publications have shown decreased expression of MSMB in prostate cancer cells compared to normal prostate (19-25). Decreased expression has been reported to correlate to shorter recurrence-free time (12) and lower survival (26), although contrary observations have also been reported (27, 28). Furthermore, low concentration of MSMB in serum correlated to a higher risk of having prostate cancer detected at biopsy (29).
Number of patients with tumors exhibiting strong, intermediate or weak to absent expression of β-Microseminoprotein (MSMB) and cysteine-rich secretory protein 3 (Crisp3) in different compartments of ovaries and ovarian tumors, as described in the Materials and Methods.
Beside the pattern of expression, experimental data support a tumor suppressor-like role of MSMB in prostate carcinogenesis. In animal models, naturally-occurring MSMB (30, 31) as well as recombinant MSMB (32) demonstrated an ability to reduce tumor growth. Furthermore, a Canadian group has studied the effect of the amino acid 31-45 region of MSMB (PCH3145), which was found sufficient to elicit MSMB-mediated antitumor effects (33). Studies in a rat model, as well as in humans, have revealed the effects of MSMB on several parameters, with potential tumor-suppressing effects, including vascular endothelial growth factor (34), matrix metalloproteinase-9 (35, 36) and the cell-surface glycoprotein CD44 (35). In vivo studies of rats inoculated with rat prostate cancer Mat Ly Lu cells showed that treatment with PCH3145 resulted in a dose-dependent decrease in tumor volume and a delay in development of skeletal metastases (33). Thus, solid clinical and experimental data indicate a protective role of MSMB in prostate carcinogenesis. Our observations on the protein expression in ovarian tumors are new and support a similar role in ovarian carcinogenesis. The strong expression in ovarian epithelium also indicates a role in ovarian physiology. Ovarian epithelium is a dynamic structure with a determinant role in healing and scar formation after monthly repeated ovulations. Whether MSMB may play a role in this process of repeated proliferative activity is unknown but in light of the proposed involvement in ovarian tumor development, this aspect deserves further study.
Expression of β-Microseminoprotein (MSMB) and cysteine-rich secretory protein 3 (Crisp3) in epithelial cells and tumor cells. The staining was classified as strong, intermediate or weak to absent, as described in the Materials and Methods. In the table, only patients with tumor exhibiting high or low expression are compared.
The biology of benign ovarian tumors at the molecular level is not well-known. Using the proliferation marker Ki-67, we found no signs of cell proliferation (18). The route from ovarian epithelium to invasive epithelial tumor is still unknown and thus, it is an open question as to how many malignant tumors develop through benign and borderline equivalents. The detection of MSMB in benign tumor cells is new and its functional significance unknown.
What regulates the expression of MSMB is not known and thus, the reason for the decreased expression of MSMB correlated to malignancy is not understood. However, recent observations in prostate cancer point to a possible explanation which may fit with the decreased expression in malignant ovarian neoplasms. The loss of expression of MSMB in hormone-refractory prostate cancer correlated with an increased expression of the polycomb protein EZH2 (37). The transcription of the gene coding for MSMB was found to be repressed by EZH2 via trimethylation of histone H3 on Lys27 in androgen-refractory but not in androgen-sensitive prostate cancer, indicating a causal relation (37). Strong expression of EZH2 is associated with aggressive tumor subgroups in other tissues (38). In gene expression analysis of blood vessels from ovaries and ovarian cancer using microarrays, overexpression of EZH2 in tumor-associated blood vessels was observed (39) and, recently, increased expression of EZH2 was found in tumor cells, as well as in tumor vasculature, and was associated with reduced survival in patients with ovarian cancer (40). We found weak to intermediate expression of MSMB in blood vessels in two of eight ovaries, but no corresponding staining in the tumors. Whether the previously described overexpression of EZH2 in ovarian cancer cells and blood vessels may contribute to the decreased expression of MSMB in ovarian cancer is unknown. Another interesting possible regulatory mechanism is the recent observation that EZH2 is suppressed by activated p53 (41). p53 is functionally inactivated in about 50% of malignant neoplasms including malignant ovarian neoplasms, as in our own material (17). Increased EZH2, mediated by decreased p53 activity, may be a new regulatory mechanism of importance in physiology, as well as in carcinogenesis. However, the exact significance of p53 in the regulation of EZH2 must be further studied.
Immunohistochemical expression of β-Microseminoprotein (MSMB) in ovaries and ovarian tumors. a: Ovary showing strong expression in the epithelium; b: benign mucinous tumor with strong staining of the benign tumor cells; c: benign Brenner tumor and strong staining of the epithelial benign tumor cells; d: borderline serous tumour showing strong expression of the tumor cells; e: weak expression in endometrioid cancer (lower part) and in ovarian stroma (upper part), while the epithelial cells stain positively; f: seropapillary cancer with weakly stained tumor cells. Magnification ×200 except in e: ×100.
Immunohistochemical expression of cysteine-rich secretory protein 3 (Crisp3) in ovaries and ovarian tumors. a: Ovary with strongly-stained epithelium; b: ovarian stroma and positively stained blood vessels; c: benign mucinous tumor with strong expression in tumor cells; d: seropapillary cancer showing strong staining of tumor cells. Magnification ×200 in all cases except in d) with magnification ×100.
Survival correlated to expression of β-Microseminoprotein (MSMB) in ovarian cancer. Survival in groups of patients having strong, intermediate or weak expression of MSMB in ovarian cancer cells. Survival was significantly longer in patients with high expression compared to the group with weak expression (p=0.02). No difference was observed between high and intermediate groups nor between groups with intermediate and low expression (p=0.32 for both).
The expression of CRISP3 has been shown to be highly up-regulated in prostate cancer (42, 43) and associated with recurrence after radical prostatectomy (12). In ovaries and ovarian tumors, we found-strong expression in ovarian epithelium, as well as in benign tumor cells, and a minor reduction of staining in malignant tumor cells. Interestingly, the blood vessels in the ovaries were also strongly stained. There was comparatively moderate- to weak-staining in half of the tumors. This is the first association of CRISP3 with blood vessels and again its significance is completely unknown.
In our present work the localizations of the staining for MSMB and CRISP3 were generally overlapping and thus leave the possibility open for interactions as has been described in the prostate. However, as no significant changes in staining of CRISP3 were observed in the malignant tumor cells compared to epithelial or benign tumor cells, our findings do not seem to support a role for CRISP3 in ovarian tumor development.
In conclusion, we have, for the first time, described the existence of MSMB and CRISP3 in the ovaries and in ovarian tumors. In patients with invasive tumors, univariate survival analysis showed a correlation between decreased expression and decreased survival. Our results direct to a tumor-suppressor function of MSMB in ovarian carcinogenesis, corresponding to recently published experimental data for prostate cancer. Therefore, further research in ovarian cancer is imperative to reveal if the MSMB gene product or a fragment of it may allow new options in the treatment of patients with ovarian cancer.
Acknowledgements
Professor Pirkko Härkönen, Department of Clinical Research Center, Lunds University, Malmö, Sweden is deeply acknowledged for support with laboratory facilities and material. The Authors declare that they have no conflict of interest regarding this study.
- Received June 12, 2012.
- Revision received July 2, 2012.
- Accepted July 3, 2012.
- Copyright© 2012 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
