The Pathogenesis of Epithelial Ovarian Cancer

Leading Opinions, 26.12.2019

Dr. Julie Terzic
Dr. Intidhar Labidi-Galy, MD-PhD
Department of Oncology
Geneva University Hospitals
E-Mail: intidhar.labidi-galy@hcuge.ch

Onkologie | Gynäkologie & Geburtshilfe

Pathogenesis of epithelial ovarian carcinomas (EOC) is complex, with five different cancer subtypes each having their particularities and arising from distinct cells of origin. A deep understanding of disease initiation and development is crucial for future development of screening tools and targeted therapies.

EOC is the leading cause of death among patients with gynecologic malignancies and represents the fifth most frequent cause of all cancer-related deaths in women. In most patients, the disease is diagnosed at an advanced stage, having spread beyond the ovaries to the peritoneum or distant organs (stage FIGO[Fédération Internationale de Gynécologie et d'Obstétrique] IIIIV). Patients diagnosed at advanced stages have a 5-year overall survival of 29% only.1

Five distinct histotypes

EOC is a heterogeneous disease with five histotypes: serous (high and low grade), endometrioid, clear cell and mucinous carcinomas. EOC represents 90% of ovarian cancers. Several hypotheses based on different etiopathological models have been proposed to explain the pathogenesis of EOC: incessant ovulation, incessant menstruation and gonadotropins. However, these hypotheses failed to provide a unified explanation.2–5
New evidences suggest that the majority of EOC are of extra-ovarian origin, arising from Müllerian derived epithelial cells outside the ovary. Indeed, high-grade serous ovarian carcinomas are very likely to arise from the fallopian tube epithelium (FTE),6–8 whereas clear cell and endometrioid carcinomas are related to endometriosis and mucinous ovarian carcinomas originating from germ cells.
Investigating the prevalence of occult ovarian and fallopian tube carcinomas in women carrying germline mutation of breast cancer (BRCA)1/2 genes who underwent prophylactic salpingo-oophorectomy played an important role in the understanding of ovarian cancer pathogenesis. Indeed, dysplastic changes and early invasive cancers that resemble high-grade serous ovarian cancer were observed within the fimbrae of BRCA carriers but not in the ovaries.7 Occult tubal carcinomas might shed malignant cells that then implant and grow on the ovary, simulating primary ovarian cancer.9 The first step in tubal transformation from benign to precursor lesion is called p53 signature, characterized by the appearance of benign secretory cells that exhibit evidence of DNA damage, TP53 mutation and p53 protein stabilisation.8 The next recognizable step is the development of serous tubal intraepithelial carcinoma (STIC), a multilayered epithelium that lacks polarity and is composed of malignant secretory cells with evidence of DNA damage and p53 stabilisation.6, 10, 11 Genetic analysis comparing microdissected paired samples of STIC and HGSOC (high grade serous ovarian carcinoma) revealed a clonal evolution, supporting the hypothesis that these lesions were the precursors of ovarian carcinomas.12
Low-grade serous carcinomas develop from serous borderline tumors (cystadenoma or papillary tubal hyperplasia). They present similar cytomorphology and the coexistence of both borderline and low-grade components in the same tumors are very frequent. They show overexpression of cyclin dependent kinase inhibitor 1A (CDKN1A) and p53 modulated genes.13
Ovarian endometrioid and clear-cell carcinomas (CCOC) are likely to originate from ectopic endometrium implanted on the ovary (endometriosis). The identification of identical mutations of AT-rich interactive domain-containing protein 1A (ARID1A), a tumor suppressor gene mutated in 50% of CCOC, in atypical endometriosis and the contiguous cancerous lesions suggested a clonal relationship.14, 15 Chronic inflammation and hyperoestrogenism in the pelvis caused by endometriosis seems to be responsible for oxidative stress and specific molecular alterations leading to malignant transformation.16
Mucinous ovarian carcinoma (MOC) is the rarest EOC and the cell of origin remains unclear.17 Recent gene-expression profiling of MOC and single-cell RNA sequencing showed that MOC are more likely to evolve from primordial germ cells than the eutopic tubal or ovarian epithelium,18 supporting the idea that less than 15% of ovarian cancer would be of primary ovarian origin.

The dualistic model of EOC

A dualistic model of EOC that took into account the molecular, histopathologic and clinical presentations of the different subtypes of ovarian carcinomas histology was proposed in 2004 by Robert Kurman and Ie-Ming Shih.19 Type I represents 25% of cases and comprises low-grade serous, low-grade endometrioid, clear-cell and mucinous carcinomas. These tumors are mainly diagnosed at early stages. They are slow-growing, treated by surgery and chemoresistant. Next-generation sequencing revealed that these tumors are chromosomally stable and TP53 mutations are unfrequent.15, 20 In low-grade serous and mucinous carcinomas, mutations in KRAS, BRAF and amplification of ERBB2 resulting in constitutive activation of mitogen-activated protein kinase (MAPK) pathway have been underlined.21, 22 Frequent inactivating mutations in ARID1A as well as mutations of genes in the phosphoinositide 3-kinases (PI3K) signaling pathway are characteristic of endometrioid and clear cell carcinomas.15 Type II represents 75% of EOC (HGSOC, high grade endometrioid and undifferenciated carcinomas). HGSOC show marked chromosomal disruption with DNA copy number alterations and TP53 mutations.21 In 20% of HGSOC mutations, germline and somatic mutations of BRCA1/2 are found with somatic loss of heterozygoty (loss of the second allele).21, 23 The Cancer Genome Atlas revealed that homologous recombination is deficient in half of the tumors.21 These findings support the development of poly(ADP-ribose)- polymerase (PARP) inhibitors as maintenance therapy in platinum-sensitive patients regardless of BRCA status.24

Why does ovarian cancer develop at all?

Several hypotheses have been suggested to explain ovarian cancer pathogenesis. The first hypothesis, called “incessant ovulation” was first proposed by Fathallah2 in 1971, based on the observation that humans are the only animals that have persistent ovulatory cycles from menarche to menopause. Disruption of the ovarian surface epithelium (OSE) with subsequent exposure to a surge of estrogen damages the OSE, leading to increased proliferation of the cells in order to repair the destructed part of the epithelium. High rates of proliferation enhance the risk of genomic instability and could lead to carcinogenesis.25 The process of ovulation is characterized by local release of the follicular fluid that bathes OSE and fallopian tube. Follicular fluid contains high levels of hormones, in particular estrogens. Concentration of estrogens in follicular fluid can reach up to 1000-fold that of serum levels. Estrogens are considered as genotoxic carcinogens generating free radicals that induce DNA damage.26 It is well established that estradiol induces DNA double-strands breaks and BRCA1 is required to repair these breaks.27 BRCA1 mutation carriers have higher risk of developing breast and ovarian cancer suggesting that these DNA damages are crucial events in tumorigenesis.
The main epidemiologic argument to support the “incessant ovulation” theory is the decrease in the incidence of EOC when ovulation is suppressed by parity or contraceptive pills.28–30 Similarly, lactation has a protective effect for ovarian cancer. Situations where women are exposed to high levels of reproductive hormones such as hormone replacement therapy of infertility treatments have been associated with an increased risk of ovarian cancer development.31, 32
Excessive exposure to gonadotrophins (follicle-stimulating hormone [FSH] and luteinizing hormone [LH]), regulatory hormones secreted by the anterior pituitary gland during menopause, ovulation and infertility therapy, raised the question of their role in ovarian cancer pathogenesis. The “gonadotrophins hypothesis” is supported by the observations on mice with FSH receptor knockout (FSHRKO). These mice are sterile with atrophic ovaries and display high levels of FSH and LH with low levels of estradiol.33 They develop changes in the OSE with formation of cysts and after a delay, they present serous papillary cystadenoma.33, 34 In EOC, FSH and LH receptors are mostly expressed in serous cystadenoma and low-grade serous carcinomas, but not in HGSOC.35, 36 Thus, gonadotrophins could play a role in the development of low-grade serous carcinoma.
A third hypothesis called “incessant menstruation” has recently emerged.4 Ovaries and the FTE are exposed to blood monthly through retrograde menstruation.37 Blood would cause reactive oxygen species and oxidative iron, which itself would induce oxidative stress and carcinogenic DNA mutations or loss. Retrograde menstruation into the peritoneal cavity is physiological in all menstruating women. It is also a well-established model of endometriosis.38 Epidemiological arguments supporting this hypothesis are the observation that hysterectomy without oophorectomy and tubal ligation plays a protective role against EOC, particularly clear- cell and endometroid carcinomas.39, 40 The younger the woman at the time of hysterectomy, the better it was in terms of ovarian cancer risk reduction, supporting the role of chronic exposure to inflammation in promoting EOC.40