Authors: Claudia Main, Xinyue Chen, Min Zhao, Lawrence W Chamley, Qi Chen
Categories: Mini-Review, endocrinal factors, endometrial cancer, fetal antigens, ovarian cancer, pregnancy, protection, AcademicSubjects/MED00250
Source: Endocrinology
It is well known that many factors, including infertility, obesity, type 2 diabetes, and family history of cancer, increase the risk of developing endometrial and ovarian cancer. However, multiparous women are known to have a lower risk of developing either ovarian or endometrial cancer than nonparous women. The lack of ovulation and shifting of sex hormonal balance, with decreased estrogen levels and increased progesterone levels during pregnancy, has traditionally been thought to be the major contributor to this decreased risk. However, in reality, the mechanisms underlying this phenomenon are relatively unknown. Increasing evidence suggests that endocrine factors are unlikely to completely explain the protective effect of pregnancies, and that multiple other nonendocrine mechanisms including fetal antigens and the newly proposed dormant cells hypothesis may also be involved. In this review, we summarize recent evidence and describe the potential underlying mechanisms that may explain how pregnancy protects against the development of ovarian and endometrial cancers in women's later life.
Keywords: ovarian cancer, endometrial cancer, pregnancy, fetal antigens, endocrinal factors, protection
Ovarian cancer is the eighth most common cancer in women, affecting approximately 300 000 women worldwide, and is responsible for 207 000 deaths every year (1, 2). Ovarian cancer is the leading cause of death among gynecological cancers. Endometrial cancer is the most commonly diagnosed malignancy of the female reproductive tract and ranks as the sixth most common cancer in women, with more than 410 000 new cases confirmed in 2020 (1, 2). The well-known risk factors for developing ovarian and endometrial cancer include increased levels of unopposed estrogen, early-onset menstruation, and late-onset menopause (3, 4). In addition, family history of ovarian or endometrial cancer, carriers of the BRCA genes (5), increasing age, estrogen replacement therapy (6), endometriosis (6, 7), obesity (8), and nulliparity (6, 9) are also considered as risk factors of ovarian and endometrial cancer.
Since the late 1990s, epidemiological studies have reported a negative association between multiple pregnancy and the incidence of ovarian and endometrial cancer (9–13). Furthermore, multiparous women not only experience reduced incidence of endometrial cancer but also delayed time of onset (14). Although the underlying mechanism for this phenomenon is relatively unknown, multiple hypotheses including the incessant ovulation hypothesis and the gonadotropin hypothesis have been reported and updated, most of which revolve around the obvious shift in hormone levels that occur during pregnancy. In this review, we discuss the current evidence underlying this association, while also discussing and proposing that other, nonhormonal mechanisms may contribute to the protection that multiple pregnancies offer against developing ovarian and endometrial cancer.
The mechanism underlying the association between multiparity and protection from developing ovarian and endometrial cancers has traditionally been thought to be due to the lack of ovulation and the shifting of sex hormone balance with decreased estrogen levels and increased progesterone levels during pregnancy (15). It is well accepted that sustained estrogen stimulation without opposing progesterone is the most common risk factor for estrogen-dependent cancer development. Postmenopausal women who have ever used estrogen therapy, for example, have a greater relative risk of developing endometrial cancer than women who have never used exogenous estrogen (16). In addition, the use of estrogen-based hormone replacement therapy in postmenopausal women significantly increases the risk for developing ovarian cancer (17).
Premenopausal women secrete unopposed intracellular estrogen, or 17β-estradiol, primarily from the ovaries, throughout the proliferative phase of the menstrual cycle (18). Leading up to ovulation, estrogen levels increase dramatically and then drop precipitously once ovulation is completed. During pregnancy, however, ovulation is suspended and, as a result, there is no fluctuation in estrogen levels. Instead, there is a rapid increase in progesterone compared with estrogen, which continues to increase 3-fold as the pregnancy reaches term (15, 19, 20). This production of progesterone is first mediated by the corpus luteum in the early stages of pregnancy, followed by the placenta, which takes over as the primary source of progesterone from approximately 8 to 10 weeks of gestation (15, 16, 20). Throughout pregnancy, progesterone and estrogen are in a balance with one another to control uterine activity. Progesterone sustains the state of pregnancy as it inhibits myometrial gap junction formation and contractility, while estrogen promotes myometrial changes and upregulates enzymes responsible for muscle contractions (21, 22). Up until delivery, therefore, the maternal hormone balance shifts to be in a progesterone-dominant state (20, 23).
Estrogen, aside from its regulatory role during pregnancy, has long been considered a mitogen as it strongly increases mitotic rates within hormone-affected tissues such as the breast and endometrial epithelium (23). Estrogen promotes growth of all endometrial cell types, altering both morphology, number, and size of endometrial cells (23). In contrast, progesterone is believed to terminally differentiate epithelium, induce apoptosis, and prevent estrogen-induced proliferation (7, 24–26). Therefore, pregnancy, with its high level of progesterone, is thought to be associated with decreased estrogen-induced mitotic activity and thereby reduces the risk of malignant transformation (7). Women who have never been pregnant have a higher overall level of exposure to unopposed estrogen, and thus have an increased rate of mitogenic activity in epithelial cells. Hence, nulliparity is associated with an increased risk of developing ovarian and endometrial cancer (7, 10, 24, 27).
Recently, a number of studies have reported that complicated pregnancies, such as pre-eclampsia, gestational diabetes mellitus (GDM), multiple miscarriage, and preterm birth, significantly increase the risk of a woman developing ovarian and endometrial cancer (28–33). Despite the increased risk of developing ovarian and endometrial cancer that is associated with pre-eclamptic pregnancies, women with pre-eclampsia have lower levels of estrogen than women with normotensive pregnancies (34–37). In addition, the level of progesterone in multiparous women has been reported to be significantly lower than in primiparous women throughout pregnancy (20). These lines of evidence suggest that reproductive endocrine factors (increased progesterone levels during pregnancy) may not completely be able to explain the negative association of reduced ovarian and endometrial cancer risk and pregnancies.
The “incessant ovulation” hypothesis describes the repeated rupture and subsequent rapid proliferation of the ovarian epithelium that is associated with ovulation (15). In terms of ovarian cancer development, this process is believed to invoke a more inflammatory environment and increase the risk of a spontaneous mutation, thus promoting the chance of neoplastic transformation (38). Indeed, a number of studies suggest that conditions related to inflammation of the ovarian surface epithelium, such as endometriosis and pelvic inflammatory disease, are associated with an increased risk for developing epithelial ovarian cancer (39). In addition, the 9 months of anovulation during pregnancy being protective against developing ovarian cancer is not replicated in other anovulatory conditions such as polycystic ovarian syndrome and female factor infertility. Instead, women with both of these conditions show an increased risk of developing ovarian cancer (19, 40, 41). Furthermore, ovarian cancer subtypes, such as serous carcinomas, are still significantly reduced in multiparous women compared with nulliparous women. The origin of this subtype of ovarian cancer can arise from ovarian surface epithelium or mesenchymal–epithelial transition (42, 43), but it also can arise from neighboring structures such as fallopian tubes (6, 15, 19, 40). A large prospective study (44) reported that the reduction in risk of ovarian cancer was greatest for the first birth (20%) compared with nulliparous women, with subsequent births only reducing the risk by a further 6%. Given this difference in magnitude of risk reduction between the first birth and subsequent births, the authors suggested that 9 months of anovulation may not be able to fully explain the protective effect but that other mechanisms may indeed be at play (44).
Cancer is considered to have a latency period of up to 20 years between the time of the initiation event and clinically apparent disease (19). Ovarian cancer is also typically diagnosed in those who have undergone menopause (44, 45), while many multiparous women who develop ovarian cancer often have their children more than 30 years prior to diagnosis, suggesting that the short-term shift in hormone levels and lack of ovulation during pregnancy may not fully be able to explain the decades-long effect pregnancy has on ovarian cancer risk in postmenopausal women.
A number of studies recently reported that women with incomplete pregnancies, such as spontaneous miscarriage or induced abortion, also have a reduced risk of developing ovarian and endometrial cancer despite progesterone levels only rising modestly compared with full term pregnancies (20, 26, 46). In a pooled analysis of 11 cohort and 29 case–control studies, incomplete pregnancies were found to reduce the risk of endometrial cancer, and to further reduce the risk by 7% to 9% for each additional incomplete pregnancy (26). This risk reduction, when compared with oral contraceptive use, was significantly higher when adjusted for time. The risk reduction observed in incomplete pregnancy (duration approximately 4-8 weeks) was similar in magnitude to the reduction observed with oral contraceptive use for 1 year (∼8% risk reduction per year of oral contraceptive use) (24, 26), suggesting that factors present in pregnancy other than hormonal changes contribute to the protective effect of pregnancy. Another recent pooled study also showed a 16% reduction in ovarian cancer risk in women ever having an incomplete pregnancy (46), and another recent study reported that the reduced risk of ovarian cancer associated with term pregnancy was similar to the reduction with incomplete pregnancy including abortions (47). Taken together, these studies suggested that the incessant ovulation hypothesis may not be able to completely explain the mechanism underlying the protective effect of pregnancies.
Similarly, a separate pooled analysis found that parity was inversely associated with cancer risk across both type I and type II endometrial cancer to a similar degree (23, 26). Traditionally, endometrial cancer has been classified as type I or type II, predominantly based on endocrine observations (23, 48). Type I tumors have been described as estrogen dependent and preceded by endometrial hyperplasia, whereas type II tumors have commonly been believed to be estrogen independent, and are thought to arise from endometrial atrophy (23, 48). Therefore, it can be assumed that estrogenic and antiestrogenic exposures would not be related to type II cancer risk. However, multiple studies have recently found increasing parity as a protective factor against developing both types of endometrial cancer (14, 23, 26, 48). In addition, circulating estradiol and progesterone levels have not been found to differ between cancer subtypes in premenopausal and postmenopausal women, and that parity does not affect sex hormone levels regardless of menopausal status (49). All of this evidence further highlights the suggestion that the protective effect of pregnancy against endometrial cancer may be mediated by more than just endocrine mechanisms (23, 49, 50).
The fetal antigen hypothesis, first described in 1980, suggests a potential mechanism underlying the protective effect of increasing parity against the development of cancer that does not involve reproductive hormones (51, 52). This hypothesis postulates that certain proteins that are released into the maternal circulation from placental/fetal cells during pregnancy are similar to those on some cancer cells. These fetal antigens are believed to stimulate the maternal immune system to produce long-lasting protection or “naturally immunize” the mother against tumors that express the same antigens, producing long-lasting immune responses to proteins expressed by both ovarian and endometrial cancer cells (53, 54). In one particular study, sera from multiparous women, but not nulliparous women or men, were found to contain cross-reactive antibodies that recognized multiple proteins expressed by ovarian tumors. Two of these proteins were characterized as human elongation factor-1α and nucleophosmin/B23 phosphoprotein, both of which are commonly expressed by ovarian cancer cell lines. These proteins have important roles in cell growth, leading the authors to propose that antibodies produced by the maternal immune system in response to pregnancy may prevent the growth of malignant cells expressing these same antigens (54). This study suggested that the unique cross-talk that occurs between the fetus/placenta and the maternal immune system to facilitate survival of the fetal allograft may contribute to the reduction in the risk of developing ovarian, endometrial, and potentially other cancers seen in women following pregnancy.
There is an interesting similarity between the placenta and tumors. Both placental trophoblasts and cancer cells show proliferative, migratory, and invasive properties. Trophoblasts can migrate away from their basement membrane, lose their polarized structure, penetrate the healthy endometrium, remodel the maternal uterine spiral arteries, and avoid detection by the maternal system (55–57). The human placenta therefore grows as a highly invasive tumor–like organ that invades the uterine wall. However, unlike tumors, this invasive process is strictly controlled both spatially and temporally. The factors that control trophoblast invasion are poorly understood. However, the extracellular vesicles (EVs) are a newly emerging physiological control system. EVs are lipid bilayer–enclosed packages of cellular contents that are released from every cell in the body and are known to be involved in cell–cell, cell–organ, and cell–organism communication and signaling. The human placenta is a generous source of EVs, and those EVs released from the placenta have been identified as important shuttle vectors and/or signal transducers during pregnancy that are known to impact the function of target cells by delivery of the cargos they carry (58). These cargos include a range of bioactive and transcriptionally active cargo, including proteins, regulatory RNAs, DNA, and lipids (58, 59). Placental EVs are thought to play a major role in mediating the adaptations of both the maternal immune and cardiovascular systems to pregnancy (60). It is possible that the fetal antigens carried by placental EVs produce the lasting cross-reactive anti-ovarian cancer cell antibody responses described by Shields (54).
While proinflammatory pregnancy conditions such as pre-eclampsia are associated with increased risk of ovarian cancer (29, 61), EVs from normal pregnancies have been shown to have anti-inflammatory effects. For example, an in vitro study showed that phagocytosis of placental EVs by macrophages stimulated the secretion of interleukin (IL)-1Rα and IL-10, both anti-inflammatory cytokines, while decreasing the release of proinflammatory cytokines, IL-1β, IL-8, and IL-12p70 (62). A recent in vitro study also revealed that culturing placental explants in the presence of lipopolysaccharide significantly reduced levels of the proinflammatory cytokine, tumor necrosis factor α, and significantly increased levels of the anti-inflammatory cytokine IL-10 (63). Through further analysis, the authors found that miRNA 519 carried by placental EVs contributed to this reduced proinflammatory response to endotoxin. miRNA 519 is encoded by the chromosome 19 miRNA cluster that is expressed essentially only in the placenta. Taken together, these lines of evidence highlight the ability of normal placental EVs to induce anti-inflammatory responses by the maternal immune system (63).
Moreover, it was recently reported that EVs from pregnancy contained miRNA that was linked to biological pathways associated with cancer. Similar to cancers, placental EVs are able to create a microenvironment that supports immunological privilege and angiogenesis, mediated through cellular reprogramming mechanisms (64). Pillay et al identified 3 miRNAs (miR-302d-3p, miR-223-3p, and miR-451a) in EVs from the maternal circulation that have a role in preventing tumor metastasis and immunosuppression (65). In particular, miR-302d-3p, a known target in estrogen signaling and proteoglycans in cancer pathways, which if aberrantly expressed can promote tumor cell growth, survival, and migration. The identification of these miRNAs in pregnancy suggested a mechanism by which proteoglycan synthesis within the tumor microenvironment might be altered (65). One limitation of these findings (65), however, was that the EVs studied were isolated from the maternal circulation, raising the possibility that these EVs may have a maternal rather than a placental origin. Our recent study (under review) found that up to 7 miRNAs (miR-519a-5p, miR-199b-3p, miR-199a-3p, miR-148a-3p, miR-26a-5p, miR-512-3p, and miR-143-3p) identified in placental EVs inhibited ovarian or endometrial cancer cell proliferation and migration, and promoted cancer cell death. Interestingly, most of these miRNAs have been reported to be downregulated in ovarian (66–70) and endometrial cancer tissues or cell lines (71, 72). Recently, we found that treating ovarian cancer cells with placental EVs isolated from first trimester placentae significantly reduced ovarian cancer cell proliferation due to a delay in cell cycle progression (73). Furthermore, in our recent unpublished proteomics analysis of placental EVs, 2 proteins identified by Sheilds et al (human elongation factor-1α and nucleophosmin/B23 phosphoprotein), as the antigens for placenta/tumor cross-reactive antibodies, were found in the EVs (54). Interestingly, the levels of these 2 proteins are significantly higher in placental EVs derived from healthy placentae than in placental EVs from miscarriage placentae. However, whether placental EVs biologically alter target cells long term has not been fully elucidated in the current literature.
A recent study following Danish women for 50 years found that the duration of pregnancy was not important in the protective effect against the development of ovarian cancer, which the authors interpreted as indicating the protective effect was likely to be associated with factors in early pregnancy and proposed “the dormant cell hypothesis” (47). The authors suggest that rather than the clearance of premalignant cells during term pregnancies, factors in early pregnancy may modify the fallopian tube epithelium, from which ovarian cancer may arise, into a dormant state. The authors speculated that this dormant state is enhanced by up to 3 pregnancies. However, the dormant state was gradually lost over decades. This could be 1 reason that the development of ovarian cancer has a latency of up to 20 years after childbirth. This dormant cell hypothesis is supported by the finding that treating ovarian cancer cells with first trimester placental EVs caused a delay in ovarian cancer cell cycle progression (73). The underlying mechanism of causing the dormant cell state could be due, at least in part, to signals delivered by placental EVs.
The mechanisms underlying the protective effect of pregnancy against the development of ovarian and endometrial cancers in multiparous women are complex and multifactorial. Current evidence suggests that not only reproductive endocrine factors but also nonreproductive endocrine factors, including maternal immune response to fetal material that may be delivered by placental EVs. Further research is required to unravel the protective mechanisms of pregnancy, which, if better understood, may provide a novel new approach for cancer therapy in the future.
Claudia Main is a recipient of the Postgraduate Honours Scholarship from the University of Auckland, New Zealand.
Claudia Main, Department of Obstetrics and Gynaecology, Faculty of Medical and Health Science, The University of Auckland, Auckland 1141, New Zealand.
Xinyue Chen, Department of Obstetrics and Gynaecology, Faculty of Medical and Health Science, The University of Auckland, Auckland 1141, New Zealand.
Min Zhao, Department of Gynecological Cancer, Wuxi Maternity and Child Health Hospital Affiliated to Nanjing Medical University, Nanjing 214002, China.
Lawrence W Chamley, Department of Obstetrics and Gynaecology, Faculty of Medical and Health Science, The University of Auckland, Auckland 1141, New Zealand.
Qi Chen, Department of Obstetrics and Gynaecology, Faculty of Medical and Health Science, The University of Auckland, Auckland 1141, New Zealand.
All authors were involved in the drafting, editing, and approval of the manuscript for publication.
All authors have no conflict of interest to report.
The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.
The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.