Elevated progesterone concentration during pregnancy and the use of progesterone-like contraceptives are known to reduce ovarian cancers. This study was undertaken to decipher whether or not there is any relationship between progesterone (also estrogen)-mediated OSE apoptosis and expression of p53, a cell-cycle arresting protein and potential tumor suppressor. For this, bovine model was chosen due to similarities with human reproductive cycle. Immunohistochemical staining with cytokeratin antibody confirmed epithelial nature of the cells in OSE layer and inclusion cysts that invaginate inside stroma after ovulation takes place. Since serum progesterone concentration in heifers remains elevated after corpus luteum is formed until parturition, the in situ apoptosis index was determined during estrus, and at mid and late-pregnancy stages. Epithelia of both tissues exhibited significantly high nuclear staining, suggesting that these cells are aiming for apoptotic destruction. To further establish the role of progesterone, the OSE cells were cultivated in M199/MCDB 105 under standard conditions and exposed to two concentrations of estrogen and progesterone. It was revealed that progesterone at both concentrations and estrogen only at high concentration converted a large proportion of these cells apoptotic. The stimulatory effect of progesterone (and to some extent estrogen) was also seen on p53 expression in the same cultivated OSE cells. The steroid dosage dependence for apoptosis and p53 expression was also somewhat similar. Assuming that progesterone action is mediated through p53-caused apoptosis as a mechanism to evade malignant transformation of OSE cells, p53 expression at mRNA and protein level was investigated in the OSE layer in proximity to stroma, antrum and corpus luteum (CL). In cycling animals CL produces a large amount of progesterone and also estrogen to maintain the post-ovulatory cycle and to suppress gonadotropin production. Hence, cells undergoing re-epithelilization and which are in contact with CL were expected to undergo maximum apoptotic modification. Indeed we got the maximum p53/p53 gene expression in these cells. We conclude that progesterone during cycling and pregnancy has a role to suppress ovarian cancer by ceasing cell cycle and diverting damaged and mutagenized OSE cells for apoptosis, and the process may be mediated through elevated p53 synthesis. However, it is also possible that progesterone and p53-induced apoptosis may be entirely different cancer suppressive actions but coincidently happening together.
The ovary is surrounded by a layer of stationary peritoneal epithelium called ovarian surface epithelium (OSE) which shows a high degree of metaplastic transformation to differentiated state, and occasionally this leads to epithelial ovarian cancer (EOC). EOC accounts for about 90% of human ovarian cancer in the Western world, and is leading cause of death from gynaecological malignancy. During post-ovulatory repair, OSE cells tend to modulate to fibroblast-like mesenchymal cells with an appearance similar to the underneath stroma. The epithelio-mesenchymal conversion of OSE layer is considered to be a homeostatic mechanism that helps to accommodate the trapped OSE cells after ovulation, and then to incorporate them in stroma as part of stromal fibroblast (Auersperg et al., 2001). Notwithstanding, certain displaced segments of damaged OSE fragments retain the epithelial features and produce surface invaginations (cleft) and inclusion cysts in the ovarian cortex (Murdoch, 1994). The healing process also results in the formation of the crypts in the ovarian surface, which penetrates the ovarian stroma, where they form inclusion cysts lined with OSE cells (Fathalla, 1971; Godwin et al., 1993). The first step in tumorigenesis of the surface epithelium is the formation of epithelial inclusion cysts derived from crypts or invaginations of the OSE (Cramer & Welch, 1983), and these cells and cells covering inclusion cysts are the primary source of EOC’s (Scully, 1977).
The reason as to why otherwise non-proliferative resting OSE cells suddenly become aggressively multiplying and tumorigenic was explained through “incessant ovulation theory”, which projected that repeated ovulation leads to malignancy due to frequent rupture and repair of damaged OSE layer. Under the influence of mitogens and other stress factors, some of the mutagenized cells undergo neoplastic transformation. Women with multiple pregnancies or those who are on oral contraceptives have shown reduced risks of EOC. Later, the validity of this hypothesis was challenged because progestin-only contraceptives, which do not restrict ovulation, were found to be as effective in suppressing malignancy as the oral ovulation-inhibitory contraceptives. Moreover, women with polycystic ovarian syndrome who had decreased ovulatory cycles were shown to have high vulnerability towards ovarian cancer. Administration of progesterone, a dominant hormone during pregnancy, also reduces the EOC risks. During pregnancy progesterone levels are very high. It has been suggested that the protection gained from pregnancy is through the 8-9 months of continuously high progesterone levels (Risch, 1998). Further, both in vitro and in vivo analyses revealed that progesterone inhibits regular cell cycle by inducing apoptosis in normal and malignant human OSE cells (Bu et al., 1997; Hu & Deng, 2000). Hen is a persistent ovulating animal, and like humans it also develops ovarian cancer and reacts with a panel of antibodies specific for human ovarian cancers (Giles, Olson & Johnson, 2006).
Considering a close model to humans, hens treated with progesterone have shown decreased incidences of EOC, supporting the hypothesis that progesterone induces apoptosis of damaged OSE cells (Fredrickson, 1987; Rodriguez et al., 1998). Progesterone has also been shown to reduce the risk of developing ovarian carcinoma in postmenopausal women who have undergone estrogen and progesterone replacement therapy (Schneider & Birkhauser, 1995). Such therapies have also clinically helped the treatment of selected ovarian tumors (Key, 1995). There is also growing evidence indicating etiological role of localized inflammation, which accompanies each ovulation, with an associated release of cytokines, mitogens and intrusion of inflammatory cells, leading to genetic damage in the OSE cells (Landen, Birrer & Sood, 2008). It has been proposed that progesterone most likely has a role in apoptotic elimination of such genetically damaged OSE cells derived from inflammatory and mitogenic responses, although the underlying mechanism is no fully understood.
Among many transcription controlling tumor suppressor proteins, p53, also referred to as “guardian of the genome”, is an important regulator of cell cycle by blocking progression through G1 phase (Livneh & Fishman, 1997). Either it triggers apoptotic response or halts the cell cycle, including cyclin-dependent kinase inhibitor p21. Both p53 and p21 were also suggested to regulate G2/M checkpoint that is transitioned from G2 to M phase. The transforming growth factor-β (TGF-β) family of growth factors inhibit proliferation of normal epithelial cells (Serra & Moses, 1996), and it was thought that TGF-β causes cell-cycle arrest in G1 phase itself. These proteins are produced in the cells in which DNA aberrations and/or mitogenic and inflammatory reactions took place due to some stress conditions like hyporexia and oncogene expression. These cells are then not allowed to grow further, but are programmed to self-destroy by apoptotic process. Apoptosis may serve an important instrument in preventing malignant transformation by specifically eliminating the cells that have undergone mutations. Mutations in p53 or inactivation through interaction with viral or cellular proteins are the most frequent alterations observed in cancer cells (Levine, 1997). P53 gene was found to be over-expressed as a result of repetition of ovulation trauma each cycle (Aunoble et al., 2000).
A question was asked whether steroidal hormones, specifically progesterone, suppress OSE malignancy by activating tumor suppressor genes, in particular p53, and consequently shunting the cells towards apoptotic destruction from the ovulation site before the mutated cells are adhered together by N-cadherin or similar adhesion proteins and proceed to neoplastic growth. A connection between progesterone and p53 was evident from the work on epithelial layer of salivary glands. Adenoid cystic carcinoma in these cells was accompanied with increase in surface progesterone and estrogen receptors and concomitant higher expression of p53 (Barrera et al. 2008). Since OSE cells also express p53 besides estrogen and progesterone receptors, and due to location, these cells are more exposed to steroidal hormones than the salivary gland cells are, we hypothesized that during pregnancy high accumulation of progesterone may be one of the factors responsible for p53-mdiated apoptotic progression of mitogenic and mutationally damaged OSE cells.
In this study we used the cow as animal model for human reproductive system. Cow is a mono-ovular animal that ovulates in a regular pattern, has same size of follicles, and its ovarian cycle is similar to that of the human. The ovarian surface morphology is also same. Even though these animals spend their short life span either gestating or lactating, due to some stray period in the wild, bovine ovarian cancer has also been reported (Marchant, 1980), suggesting that the tumorigenic potential of bovine OSE is similar to human OSE. Therefore, bovine ovaries present a useful model for studying oncology of OSE and inclusion cysts. Normal gestation period in of cattle also averages 278-284 days. Like humans, during pregnancy, the gonadotropin levels remain low but serum progesterone concentration continues to build up and stays high until last month before parturition. After the first trimester of pregnancy, estrogen levels also begin to increase again and climb consistently until the eighth month, when they level off for about two weeks and then increase rapidly until parturition (Henricks & Dickey, 1970; Smith et al., 1973). The most sticking similarity for which bovine OSE can be modelled for humans has been the expression profiles of localized hepatocyte growth factor (HGF) (Parrott & Skinner, 2000), and kit legend stem cell growth factor (KL) and its receptor c-kit (Parrott, Kim & Skinner, 2000) in OSE layer. Moreover, the extent of expression of these proteins in cultured human and bovine OSE cells was also alike. One can expect that the underlying signal transduction mechanism leading to HGF/KL-associated up-regulation of tumor-associated genes would also be similar.
This study was designed to test the hypothesis that pregnancy suppresses EOC by inducing apoptosis within the OSE layer and inclusion cysts, a process mediated through over-expression of p53, which is under the control of high progesterone and estrogen level. This was done by 1) measuring apoptotic cell death in inclusion cysts and OSE, 2) determining direct effect of progesterone and estradiol on cell death and 3) localization of p53 and its expression profile under progesterone influence.
In this study, apoptotic behaviour of OSE and inclusion cyst was examined during the mid and late pregnancy period in cows. The immunohistochemical analysis of the tissues revealed that staining was taken only by the surface epithelia. Anti-cytokeratin antibody used in this test recognizes only the epithelial cells’ cytoplasmic microfilaments and not the mesenchyma, which might arise due to epithelial-mesenchymal conversion after ovulation. Fact that the inclusion cysts were also stained, we rule out any possibility of invasion by fibroblast-like mesenchymal cells in the cysts. The inclusion cyst epithelium is the preferred site for metaplastic transformation and tumorigenesis, and epithelial-mesenchymal transformation is a mechanism to prevent this manifestation. Since the objective was to demonstrate that apoptosis could be a mechanism to avert malignancy in these tissues, it was necessary to verify that the cells were of epithelial origin and not mesenchymal. Using standard TUNNEL assay which recognizes apoptotic nuclei, a very high percentage of epithelial cells from inclusion cysts during the mid pregnancy period (55-80 days) acquired the staining. Although the proportion of apoptosis-positive cells was much lower in the OSE layer than in the inner lining of inclusion cyst, the highest frequency in this tissue was also seen during mid pregnancy period. There were negligible cells with nuclear stain in the cycling animals, suggesting that practically no apoptosis took place in estrous period. The frequency of apoptotic cells started to decrease as the animals approached late pregnancy (90-140 days). This observation was rather unexpected because the serum progesterone level reaches to its maximum during late pregnancy period (Smith et al., 1973). This indicates that serum hormone level may not reflect the actual concentration of progesterone in the vicinity of ovarian surface. Until mid pregnancy, CL produces progesterone, and when it starts to regress, placenta takes over. Because of the proximity, the effective hormonal concentration around OSE is expected to be high under the influence of CL rather than placenta, even though both of the might contribute same progesterone level in the blood stream. Once CL starts to regress that is during mid-pregnancy the effective concentration starts to decline.
To ascertain a direct effect of steroidal hormone on the apoptotic response on OSE, the cells were cultured in M199/MCDB 105 medium with standard additives. For bovine cells there seems to be some flexibility in choice of medium, because for cultivation Dulbecco’s modified Eagle’s medium has also been used by some workers (Parrott & Skinner, 2000). One concern with OSE cell culture is that through growth passages, epithelial cobblestone compact cells tend to convert to spindle shaped mesenchymal phenotype (fibroblast-like shape) (Auersperg et al., 2001). The cultivated bovine OSE cells maintained usual cobblestone morphology of original source cells of OSE layer, and apparently the genetic makeup also remained unchanged. There were two noteworthy findings with the cultivated cells. Estrogen at high and progesterone at both high and low concentrations significantly increased the in vitro apoptotic activity. In fact, high progesterone stimulated apoptosis nearly to extent of positive control (H2O2-treatment). Care was taken not to enumerate the necrosed cells. Besides, both hormones also induced expression of cell-cycle arrester and apoptosis-inducer protein, p53. We propose that progesterone, and to some extent estrogen, through their receptors up-regulate a cascade of downstream pathways leading to cell-cycle arrest and programmed cell death, and further that p53 protein seems to be responsible as a mediator molecule.
The results of in vitro hormonal effects paved way to test the actual in vivo influence. For this, p53 expression at mRNA and protein level was examined in those regions of OSE layer that are directly in contact with stroma, large antrum follicle or CL. CL is the potential endocrine for progesterone synthesis in cycling animals, and one can expect that OSE in its proximity would be under the progesterone influence. Antral follicle and CL are also the source of estrogen and they might also affect the neighbouring OSE layer. In this situation p53 expression should be high only in those regions of OSE which are in contact with CL and antrum. In accordance, we found very high extent of p53 immunoreactivity and PCR product of p53 gene mRNA in OSE layer overlaying CL. The control, GAPDH expression was nearly identical in all the cells. There was some marginal expression of p53 protein in antrum, which is most likely due to estrogen release. Although not tested, we hypothesize that in cow induced apoptosis during mid pregnancy phase is due to elevated progesterone, and perhaps is mediated through p53. The relationship between progesterone induced anti-cancerous effects in OSE layer and p53-mediated apoptosis and thereby evading tumorigenesis may be purely coincidental. This is because the signals inducing p53 expression are mitogenic and/or DNA damaging activities (Corney et al., 2007) rather than steroids.
Wilcox et al. (2007) for the first time elucidated a direct contact effect of progesterone on anti-cancerous properties of cultured human OSE cells. The approach was microarray-based transcriptional profiling of over 22,000 transcripts (cDNA’s generated from total RNA’s) of progesterone treated and un-treated control cells. The most significant up-regulation in treated cells was of the enzymes and transporter proteins of cholesterol biosynthetic pathway, besides the progesterone receptors, PR-A and PR-B. Among the cholesterol biosynthesis genes, TMEM97 gene encoding a trans-membrane protein of unknown function was found to be most expressed. It was concluded that progesterone through its receptors mediates an up-regulation of cholesterol and lipid biosynthesis in OSE cells. This in turn decreases the cell membrane fluidity, and such cells are less vulnerable to inflammatory and physical damage due to ovulation.
Another microarray analysis was carried out to elucidate p53 gene targets in cultured mice OSE cells (Corney et al., 2007). For this, microRNA (miRNA) expression of OSE cells in p53 knock-out mutant and wild type mice was compared. It was found that two genes for miRNA’s, termed miR34-b and miR34-c were over-expressed in wild type cells. Upon induction of the p53 gene these two miRNA’s also over-expressed. These miRNA’s were shown to “silence” several cancer-related genes like Delta-like 1, Notch-1, Met, Myc, Cdk-6, Cyclin D1 etc. Mi-RNA-mediated repression of these downstream target genes was considered to be a mechanism whereby p53 suppresses the tumor progression. In another investigation, Zhang et al. (2008) concluded that other than the epigenetic silencing factors like p53-induced miRNA’s, several human genomic tumor suppressor genes assigned to chromosome #14 also transcribe diverse classes of mi-RNA’s for silencing cancer-related genes. Comparative genome hybridization and microarray profiling of normal and neoplastic OSE cells was carried out to elucidate signature genes for tumour suppression and many of them were found to be involved with p53 pathway (Landen, Birrer & Sood, 2008). Mutation to p53 gene and some other ovarian and breast tumor suppressor genes like BRCA1 and BRCA2 are linked to “high-grade pathway” of malignancy owing to unchecked proliferation, inhibition of apoptosis, angiogenesis, stromal invasion, separation and survival away from the primary tumor and implantation and growth within new tissues.
In conclusion, progesterone and p53 seem to be associated with suppression of bovine EOC in both cycling and pregnant animals. At least two distinct mechanisms may be attributed to this phenomenon, 1) anti-inflammatory activity due to progesterone-mediated increase in cholesterol and lipid biosynthesis, and 2) p53- mediated cell cycle-arrest and apoptosis. Whether these attributes are connected to each other or whether they are two distinct events independently contributing to tumor suppression, needs to be investigated in future.
- Auersperg, N., Wong, A.S.T., Choi, K.C., Kang, S.K. & Leung, P.C.K.(2001). Ovarian Surface Epithelium: Biology, Endocrinology, and Pathology. Endocrine Reviews, 22(2), 255-288.
- Barrera, J.E., Shroyer, K.R., Said, s., Hoernig, G., Melrose, R., Freedman, P.D., Wright, T.A. & Greer, R.O.(2008). Estrogen and Progesterone Receptor and p53 Gene Expression in Adenoid Cystic Cancer. Head and Neck Pathology, 2, 13–18.
- Corney, D.C., Flesken-Nikitin, A., Godwin, A.K., Wang, W. & Nikitin, A.Y.(2007). MicroRNA-34b and MicroRNA-34c Are Targets of p53 and Cooperate in Control of Cell Proliferation and Adhesion-Independent Growth. Cancer Research, 67(18), 8433-8438.
- Giles, J.R., Olson, L.M. & Johnson, P.A.(2006). Characterization of Ovarian Surface Epithelial Cells from the Hen: A Unique Model for Ovarian Cancer. Experimental Biology and Medicine, 231,1718-725.
- Henricks, D.M. & Dickey, J.F.(1970). Serum Luteinizing Hormones and Plasma Progesterone Levels During the Estrous Cycle and Early Pregnancy in Cow. Biology of Reproduction, 2, 346-351.
- Landen, C.N., Birrer, M.J. & Sood, A.K.(2008). Early Events in the Pathogenesis of Epithelial Ovarian Cancer. Journal of Clinical Ontology, 26(8), 995-1005.
- Livneh, E. & Fishman, D.D.(1997). Linking protein kinase C to cell-cycle control. European Journal of Biochemistry, 248(1-9), 257-265.
- Parrott, J.A. & Skinner, M.K.(2000). Expression and Action of Hepatocyte Growth Factor in Human and Bovine Normal Ovarian Surface Epithelium and Ovarian Cancer. Biology of Reproduction, 62, 491–500.
- Parrott, J.A., Kim, G. & Skinner, M.K.(2000). Expression and Action of Kit Ligand/Stem Cell Factor in Normal Human and Bovine Ovarian Surface Epithelium and Ovarian Cancer. Biology of Reproduction, 62, 1600–1609.
- Smith, V.G., Edgerton, L.A., Hajs, H.D. & Convey, E.M.(1973). Bovine Serum Estrogens, Progestins and Glococorticoids During Late Pregnancy, Parturition and Early Lactation. Journal of Animal Sciences, 36(2), 391-396.
- Wilcox, C.B., Feddes, G.O., Willett-Brozick, J.E., Hsu, L.C., DeLoia, J.A. & Baysal, B.E.(2007). Coordinate up-regulation of TMEM97 and cholesterol biosynthesis genes in normal ovarian surface epithelial cells treated with progesterone: implications for pathogenesis of ovarian cancer. BMC Cancer, 7, 223, Web.
- Zhang, L., Volinia, S., Bonome, T., Calin, G.A., Greshock, G., Yang, N., Liu, C.G., Giannakakis, A., Alexiou, P., Hasegawa, K., Johnstone, C.N., Megraw, M.S., Adams, S., Lassus, H., Huang, J., Kaur, S., Liang, S., Sethupathy, P., Leminen, A., Simossis, V.A., Sandaltzopoulos, R., Naomoto, Y., Katsaros, D., Gimotty, P.A., DeMichele, A., Huang, Q., Buetzow, R., Rustgi, A.K., Weber, B.L., Birrer, M.J., Hatzigeorgiou, A.G. Croce, C.M. & Coukos, G.(2008). Genomic and epigenetic alterations deregulate microRNA expression in human epithelial ovarian cancer. Proceedings in National Academy of Sciences U.S.A., 105(19), 7004–7009.