Interplay of Epigenetics with Gynecological Cancer

Recent data on the cell deregulation that occurs during the progression to cancer un‐ derlines the cooperation between genetic and epigenetic alterations leading to a ma‐ lignant phenotype. Unlike genetic alterations, the epigenetic changes do not affect the DNA sequence of the genes, but determine the regulation of gene expression acting upon the genome. Moreover, unlike genetic changes, epigenetic ones are reversible, making them therapeutic targets in various conditions in general and in cancer dis‐ ease in particular. The term epigenetics includes a series of covalent modifications that regulate the methylation pattern of DNA and posttranslational modifications of histo‐ nes. Gene expression can also be regulated at the posttranscriptional level by micro‐ RNAs (miRNAs), a family of small noncoding RNAs that inhibit the translation of mRNA to protein. miRNAs can act as ‘oncomiRs’, as tumor suppressors, or both. In this chapter, we will (1) summarize the current literature on the key processes respon‐ sible for epigenetic regulation: DNA methylation, histone modifications and posttran‐ scriptional gene regulation by miRNAs; (2) evaluate aberrant epigenetic modifications as essential players in cancer progression; (3) establish the roles of microenvironmentmediated epigenetic perturbations in the development of gynecological neoplasia; (4) evaluate epigenetic factors involved in drug resistance.


Key processes responsible for epigenetic regulation
Epigenetics could be broadly defined as the sum of cellular and physiological trait variations that are not caused by changes in the DNA sequence. Epigenetic mechanisms are essential for the normal development and maintenance of tissue-specific gene expression patterns in mammals. Disruption of epigenetic processes can lead to altered gene function resulting in imprinting disorders, developmental abnormalities and cancer. The epigenetic mechanisms that will be presented in this chapter are (1) DNA methylation, (2) chromatin and histone modifications, and (3) regulatory noncoding RNAs.

DNA methylation
DNA methylation is a biochemical process characterized by the addition of a methyl group especially at the C5 position of cytosine from CpG dinucleotides and is accomplished by two classes of DNA methyltransferases involved in maintenance and de novo methylation [1]. CpG dinucleotides are not randomly distributed across the human genome but are found in short CpG-rich DNA sequences called 'CpG islands.' CpG islands are found in regions of large repetitive sequences (e.g. centromeric repeats, retrotransposon elements, rDNA) [2,3] and in 60% of human gene promoters [4]. Some CpG islands are methylated, whereas the majority of them usually remain unmethylated during development and in differentiated tissues [5]. CpG islands' promoters become methylated during development (imprinted genes, chromosome X inactivation) [2]. Another role of CpG island methylation is to silence noncoding DNA and transposable DNA elements to prevent chromosomal instability by heavy methylation of repetitive sequences [5]. DNA methylation leads to gene silencing by either preventing or promoting the recruitment of regulatory proteins to DNA. Methylation of CpG islands can block the access of transcription factors to the transcription sites [6,7], or by recruiting methylbinding domain proteins (MBDs), which can mediate gene repression through interactions with histone deacetylases (HDACs) [8,9]. This epigenetic modification does not change the DNA sequence, but enhances the stability and chromosome integrity and promotes genome organization into transcriptionally active or silenced regions. DNA methylation at the whole genome level provides a specific global methylation pattern [2,10] that plays an important role in regulating gene expression (e.g. development and cell-specific gene expression) in association with chromatin-associated proteins. The maintenance of a cell-specific methylation pattern after every cellular DNA replication cycle provides a stable gene-silencing mechanism that plays an important role in regulating gene expression. The maintenance methyltransferase DNMT1 is responsible for copying DNA methylation patterns to the daughter strands during DNA replication, whereas DNMT3a and DNMT3b are de novo methyltransferases that establish the methylation patterns early in development [11]. DNMT3L, a homologous protein to other DNMT3s, increases the ability of DNM3a and 3b to bind to DNA, stimulating their activity. Some problems in the establishment of methylation biomarkers in gynecologic cancers, especially in cervical cancer [12], come from the fact that: (1) the extent of methylation across the various CpG sites in a promoter can be rather heterogeneous and consequently, the assay outcome is likely to be influenced by the region of CpGs that is targeted; (2) the distinct levels of background methylation due to differences in cell type composition between cervical tissue samples that can contain substantial amounts of nonepithelial (stromal) cells and cervical scrapings that are enriched in superficial epithelial cells. For this reason, the methylation results obtained from tissue samples may not be directly extrapolated to cervical scrapings [13]. In addition, while the methylation of tumor suppressor' promoters is an early and frequent alteration in carcinogenesis [14] and, on the other hand, is widespread in the human genome, only a subset of affected loci play critical roles in tumorigenesis [15]. CpG hypermethylation is gene-and cancer type-specific [16,17,18,19], providing a useful signature for tumor diagnosis and prognosis [18] that must be established accurately.

Covalent histone modifications
Mammalian genome represents a highly structured complex comprised of compacted DNA and proteins that can adopt different three-dimensional conformations dependent of nuclear context and biochemical changes present in the genome and at the histone level [20]. At first glance, the chromatin is present in two forms: transcriptionally active euchromatin and more condensed and transcriptionally inactive heterochromatin. In the genome, there are some structural regions (such as centromeres) containing constitutive heterochromatin; others may go through an open conformation to a compact one-optional heterochromatin. These transitions, vital to the establishment of necessary transcriptional various models of embryonic development, growth, and adult life, are under epigenetic control. Nucleosomes form the repetitive fundamental units of the chromatin and are designed to pack the huge eukaryotic genome in the nucleus (mammalian cells contain approximately 2 m of linear DNA wrapped in a core size of 10 µm in diameter) [20]. The nucleosomes in turn are compacted and form the chromosomes. The nucleosomal core consists of approximately 147 base pairs wrapped around a histone octamer made up of two copies of the histones H2A, H2B, H3, and H4. Histone H1 (linker histone) and its isoforms are involved in chromatin compaction underlying nucleosome condensation. Decondensed nucleosomes look like a bead wrapping a DNA molecule [21]. Histone covalent modifications (epigenetic changes) represent important regulatory elements that influence chromatin interactions by structural changes either by electrostatic interactions and recruitment of nonhistone proteins [22].
Histones can undergo a variety of posttranslational modifications at the N-terminus (like acetylation, methylation, phosphorylation, sumoylation, ubiquitination, and ADP-ribosylation) that can alter the DNA-histone interaction, with a major impact on chromatin structure and key cellular processes such as transcription, replication, and repair [20]. The histone code may be transient or stable. The mechanism of inheritance of this histone code is not fully understood. The patterns of histone modifications are specific to each cell type and play a key role in determining cellular identity [23,24]. In contrast with stem cells, differentiated cells acquire a more rigid chromatin structure, which is important for maintaining cell specialization [23]. Epigenetic regulation mediated by histone modification is a dynamic process. Lysine residue methylation using histone methyltransferase (HMT) is correlated either with transcriptional activation or repression, whereas lysine acetylation correlates with transcriptional activation [25]. Histone methyltransferases (HMTs) and demethylases (HDMs) work in tandem to determine the degree of methylation of the lysine residue [26]. Histone H3 lysine 4 trimethylation (H3K4me3) correlates with euchromatin and gene transcription activation. Histone H3 lysine 27 trimethylation and/or lysine 9 (H3K27me3/H3K9me3) is correlated with the transcriptional repression of heterochromatin and H3K27me3 modification is critical for stem cells; demethylation at this level is correlated with differentiation [27, 28, 29, 30, 31]. These two modifications represent the main silencing mechanisms in mammalian cells, H3K9me3 working in concert with DNA methylation and H3K27me3 largely working exclusive of DNA methylation [32]. Histone acetylation is one of the histone modifications that have been studied extensively. The two homonymous enzymes that are involved in maintaining a specific profile are histone acetyltransferases (HATs) and histone deacetylases (HDACs) [26]. Generally, the level of histone acetylation correlated with transcriptional activation and deacetylation correlates with transcriptional repression. H3 histone acetylations at lysine 9 (H3K9ac) and lysine 4 to 16 are characteristic euchromatin changes located in regions where genes are actively transcribed. Although histone modifications act mainly by altering the architecture of some modifications (H3K4me3 and H3K9ac) mediates gene regulation by recruiting other proteins involved in chromatin remodeling [33,34]. Histone modifications and DNA methylation interact with each other at multiple levels to determine gene expression status, chromatin organization, and cellular identity [35]. Several HMTs, including G9a, SUV39H1, and PRMT5, methylate DNA to specific genomic targets recruiting DNA methyltransferases (DNMTs) [36,37,38]. In addition, DNMTs may recruit HDACs and methyl-binding proteins to achieve gene silencing and chromatin condensation [8,9]. DNA methylation can also be established via H3K9 methylation, such as MeCP2, thereby establishing a repressive chromatin state [39]. Recent studies showed that the main chromatin changes that occurs during tumorigenesis are characterized by a global loss of acetylated H4 lysine 16 (H4K16ac) and H4 lysine 20 trimethylation (H4K20me3) [40]. HDACs were found overexpressed in various types of cancer [41,42] (becoming a major target for epigenetic therapy), along with HATs, whose expression can also be altered in cancer. MOZ, MORF, CBP, and p300 (HATs) may be targets for chromosomal translocations, especially in leukemia [43]. Changes in histone methylation patterns (deregulation of HMTs) are associated with aberrant gene silencing in cancer, and an effective cancer treatment strategy targeting HDMs represents a promising treatment option.

Posttranscriptional gene regulation by noncoding RNAs
Noncoding RNAs are involved in fundamental processes, such as chromatin dynamics and gene silencing, and their transcripts outnumber the group of protein transcripts. It is well known that the initiation of X-chromosome inactivation is regulated by noncoding RNAs (Xist function) and the noncoding RNAs molecules are also involved in imprinting, suggesting that antisense RNA can induce transcriptional silencing [44,45,46]. The characterized noncoding RNA family consists of a large group of small regulatory microRNAs (about 1400 microRNAs in humans) [47].MicroRNAs (miRNAs) are short noncoding RNAs of 20-24 nucleotides that play important roles in virtually all biological pathways in mammals like differentiation and growth control. Based on computer predictions, it was proposed that miRNAs may regulate many cell cycle control genes [48]. miRNAs influence numerous cancer-relevant processes such as proliferation, cell cycle control, apoptosis, differentiation, migration, and metabolism.
The key processes of miRNA biogenesis pathways have been characterized. Primary miRNA transcripts are transcribed from separate transcriptional units or embedded within the introns of protein coding genes by RNA polymerase II. Primary miRNA transcripts are processed by a complex formed by RNase III enzyme and Drosha, resulting in a pre-miRNA hairpin that is subsequently exported from the nucleus to the cytoplasm by exportin 5 (XPO5). Further pre-miRNA molecules are processed by another protein complex, including DICER and TRBP, to produce the single-stranded mature miRNA (ssmiRNA). ssmiRNA is subsequently incorporated in RNA induced silencing complex (RISC), along with key proteins such as AGO2 and GW182. The role of mature miRNA (as part of the RISC) is to induce posttranscriptional gene silencing by complementary sequence motifs to the target mRNAs predominantly found within the 3′ untranslated regions (UTRs) [47,49,50]. One specific miRNA may target up to several hundred mRNAs; therefore, a miRNA may silence various genes while a specific mRNA may be targeted by several miRNAs. Aberrant miRNA expression may interfere with gene transcription and influence cancer-related signaling pathways [51,52,53].New data are added to decipher the role of miRNAs in normal physiology and pathology. Several microarray expression studies performed on a wide spectrum of cancer types have proved that deregulated miRNAs expression is the rule rather than the exception in cancer [54,55,56,57]. Animal models featuring miRNA overexpression or knock-down have demonstrated the relation between miRNAs and cancer development, thus proposing miRNAs as potential biomarkers and putative therapeutic targets [58]. In addition, since miRNAs were discovered, many researchers focused their interest on identifying miRNAs generated by viruses. Several data support this hypothesis mainly based on miRNA size, which allows them to avoid the immune system but also to be supported by the small size of viral genome. It is not unexpected that many miRNAs encoded by viruses have been discovered, most of them transcribed from double-stranded DNA viruses [59]. miRNAs can regulate the expression of viral genes that are involved in controlling viral replication. It is supposed that these miRNAs might influence viral gene expression in a differentiation-dependent manner by targeting viral transcripts. On the other hand, different hrHPV types have different oncogenic potentials, viral miRNA being considered one of the factors involved in oncogenic regulation; some conserved miRNAs are involved in the switch from HPV productive to transforming infections.

Evaluation of aberrant epigenetic modifications as essential players in cancer progression
Normally, evolution and morphological state of genital organs are in close interdependence with hormonal status that is different in different periods: childhood, sexual maturity, climacterium, and menopause. On the other hand, there is an increasing interest in the identification of diagnostic biomarkers and biomarkers able to predict both response to treatment and survival. For an optimal planning of therapeutic strategy in high-risk patients, a close association between biological variables and (epi)genetic profiles associated with aggressive clinical behavior could be useful. Therefore, many cellular changes should be analysed in this context.
Benign tumors of the vulva can be developed from epithelial components (papillomas and warts) mezenchimatos tissue (fibroma, leiomyoma, lipoma, hemangioma, and lymphangioma), and local glands (Bartholin gland cysts or cysts of sweat glands). Vulvar cancer is a rare malignant disease accounting for less than 5% of gynecological malignancies [60,61,62]. The most common vulvar cancers are epidermoid carcinoma and rarely adenocarcinomas that are developed in the Bartholin glands or sweat glands. Approximately 20%-40% of vulvar squamous carcinomas are often associated with papilloma virus infection [60 -66] and are more frequent in young people. Non-HPV vulvar cancers occur in the elderly and are associated with somatic mutations, especially in TP53 [60 -63, 65, 66]. Tumors harbouring a mutation have a worse prognosis than vulvar squamous cancers without (epi)genetic changes [67 -70]. However, allelic imbalances seem to occur in both groups and the cumulative number of epigenetic changes increases from dysplasia to cancer [71]. The data with respect to epigenetic changes in vulvar cancer progression is limited to a few articles on DNA hypermethylation but not to chromatin remodeling or histone modifications. This data is presented in Table 1. Hypermethylation seems to be more frequent in vulvar squamous cancers than in vulvar intraepithelial neoplasia, but more studies are needed. Taking into account the existence of two etiological categories of vulvar carcinomas (related or not to HPV), the miRNA signature in these two types of vulvar carcinomas were evaluated [72]. Some miRNAs had lower expression in HPV-positive tumors (miR-1291, miR-342-3p, miR-193a-5p, miR-29c#-, miR-106b#, miR-22#, miR-365, miR-151-5P, miR-144#, miR-125b-1#, miR-519b-3p, miR-26b, miR-19b-1, and miR-1254) and other microRNAs had higher expression in HPV-positive tumors (miR-1274B, miR-142-3p, miR-21, miR-708, miR-16, miR-660, miR-29c, miR-1267, miR-454, and miR-186) [72]. In HPV-negative samples, we observed an association between lymph node metastases with decreased expression of miR-223-5p and miR-19b-1-5p, vascular invasion with decreased expression of miR-100-3p and miR-19b-5p-1, and advanced tumor staging (FIGO IIIA, IIIB, and IIIC) with expression of microRNAs miR-519b-1-5p and miR-133a. In addition, de Melo Maia and collaborators (2013) built a network between miRNA expression profiles and putative target mRNAs (TP53, RB, PTEN, and EGFR) based on prediction algorithms, demonstrating that the evaluated miRNAs can be involved in vulvar cancer progression, thereby providing biomarkers for the establishment of prognostic and predictive values of response to novel targeting therapies in vulvar cancer [72].
The vagina is a fibromuscular tubular organ, which histologically consists of three layers of tissue: (1) an outer layer consisting of fibro-elastic connective tissue; (2) vaginal muscles with a longitudinal outer layer and an inner layer of fibers circularly arranged in a spiral; and (3) Malpighian mucosa, covered by squamous epithelium. The vaginal epithelium undergoes changes in relation to the period of the woman's life and depending on hormonal stimulus. Histological changes are reflected in vaginal cytology. Vaginal epithelium responds to ovarian stimuli through proliferation, differentiation and desquamation. Thus, in adult women, under the action of estrogen during the proliferative phase, vaginal mucosa proliferates and differentiates morphologically and functionally, and later, during the luteal phase, under the action of progesterone, superficial cell layer desquamation occurs. The action of estrogen on the vaginal mucosa is exercised on the epithelium as well as on the subjacent stroma.
Vaginal cancer is also a rare malignancy, accounting for about 2%-3% of all gynecologic cancers [73,74]. The squamous cell carcinomas (SCC) are more frequent (80%-90%) than adenocarcinomas. If the risk factors linked to vaginal squamous cell carcinoma are smoking, immunosuppression, high number of sexual partners, papillomavirus and history of cervical precancerous and cancerous lesions [75,76,77], in the case of the vaginal adenocarcinomas, particularly clear cell adenocarcinomas, exposure to an antiabortive drug diethylstilbestrol (DES) was incriminated [78,79,80]. On the other hand, if squamous vaginal cancer tends to occur more commonly in the proximal third of the vagina, especially the posterior vaginal wall, the adenocarcinomas are mostly seen in the anterior upper vaginal wall [74]. Human papillomaviruses have been also linked to vaginal cancers, HPV prevalence in 2/3 lesions of vaginal intraepithelial neoplasia and invasive vaginal cancer being over 90% and 70%, respectively [81,82]. The HPV oncogenic transformation has been associated with high levels of E6 and E7 viral oncoproteins in the epithelia that can be achieved by two mechanisms: (1) increased production of E6 and E7 after the loss of E2 (the normal regulator of E6 and E7 expression) during viral integration [83]; (2) methylation of the E2-binding sites (E2BS) in the viral LCR in the region close to the early promoter that could inhibit E6 and E7 transcription The uterus is a hollow organ, in which the product of conception is developed. It consists of three parts: body, isthmus, and cervix. The corpus presents a mucosa (endometrium), muscular wall (myometrium), and serous peritoneal surface. The endometrium is a specialized tissue, particularly receptive to the influence of sex hormones that differs from a histological point of view at prepubertal periods, sexual maturity, and menopause. Also, the uterine mucosa is in constant transformation during menstrual cycles, sexual maturity, growth processes, functional maturation, and regression. Similar risk factors for endometrial cancers were incriminated: adult obesity [95], first-degree family history of endometrial cancer, or colorectal cancer [96]. Nulliparity and infertility appeared to independently contribute to endometrial cancer risk [97]. The endometrium is extremely sensitive to hormones, the estrogen and progesterone being two key regulators of proliferation and differentiation in reproductive tissues [98]. The two isoforms of the progesterone receptor, PRA and PRB, required for endometrial differentiation [99], are generated by alternative transcription and translation from the same gene with the addition of 164 amino acids in the N-terminus sequences of PRB [98] that makes them functionally different [99]. A shift in the estrogen-progesterone balance is the major cause for the development of endometrial cancer [100]. Progesterone is an important inducer of endometrial differentiation and an inhibitor of tumorigenesis because the addition of progestin (synthetic progesterone) can prevent endometrial cancer induced by an excess of estrogens from endogenous sources (e.g., adipose tissue storage of estrogen and with polycystic ovarian syndrome) or from exogenous sources in therapeutic administration [100]. While progestin therapy achieves promising outcomes with early stage endometrial cancer, advanced and recurrent disease has only minor effects. This is due to the fact that in advanced endometrial cancer, the progesterone receptor is lost but it has been demonstrated that reestablishing progesterone signaling in these cells can inhibit endometrial cancer cell proliferation and invasion and increase sensitivity to apoptotic stimuli [100]. The epigenetic restoration of progesterone receptor expression could result in resensitization of endometrial tumors to progestin therapy. The functional role of epigenetic factors in endometrial cancer development began to be evaluated. A study by Jones and collaborators (2013) emphasizes the role of HAND2 hypermethylation, which is a key step in endometrial carcinogenesis [101]. HAND2 is a basic helix-loop-helix transcription factor and developmental regulator [102], expressed in the normal endometrial stroma. The physiological function of HAND2 is to suppress the production of fibroblast growth factors that mediate the paracrine mitogenic effects of estrogen on the endometrial epithelium [103]. HAND2 is under progesterone regulation [104,105], entering in the progesterone-mediated suppression of estrogen-induced pathways. Consequently, the methylation of HAND2 is able to predict the response to progesterone [101]. HAND2 methylation is the most common molecular alteration in endometrial cancer and, on the other hand, is an early event in endometrial carcinogenesis that makes it a sensitive test to correctly identify endometrial cancer patients amongst those women who present with postmenopausal bleeding [101].
Histologically, the cervix shows mucosa, muscle wall, and the peritoneal serosa. The mucosa of the cervix has an exocervical portion (covered by squamous epithelium, nonkeratinized) and one endocervical (covered by a single-layered cylindrical epithelium, mucus secreting, which contains a small number of ciliated cells, basal stem cells and racemic, tubular, or branched type glands). Cancer of the uterine cervix is the major cause of death from gynecological cancers and in over 90% of cases is associated with high-risk human papilloma virus (hrHPV). Etiological factors include cigarette smoking, impairment of cell-mediated immunity, and long-term estrogen-progestin use [106,107,108]. But the main etiological factor of squamous cell carcinoma (that accounts for about 80% of the cases) as well as adenocarcinoma are human papilloma virus infections [109]. The role of other sexually transmitted infections (Chlamydia trachomatis and herpes simplex virus) is still unclear [108,110].In cervical cancer, tumorigenesis of both squamous cell carcinoma and adenocarcinoma is HPV-related [109]. The transforming potential of E6 and E7 viral oncoproteins is based on their numerous actions on cellular proteins, mainly on p53 and pRB tumor suppressors, which are degraded and inactivated, respectively. In addition to the already reported genomic alterations in cervical cancer development by hrHPV, many studies underline the involvement of epigenetic alteration in host cell genes or at the levels of RNA. In order to find some diagnostic and prognostic biomarkers, the methylation of host cell genes and methylation of viral genes were evaluated [12]. The CpG hypermethylation of promoters of tumor suppressor genes, an early and frequent alteration in carcinogenesis, affects all important pathways: cell adhesion (cell adhesion molecule 1 (CADM1)) [13], E-cadherin [111,112], apoptosis (DAPK, a proapoptotic serine/threonine kinase [113,114]), cell cycle (cyclin A1 methylation [114,115]), fragile histidine triad (FHIT) [116], cell signaling pathways (retinoic acid receptor [117], Ras association domain family 1 isoform A (RASSF1) [118]), Wnt/β catenin pathway (adenomatous polyposis coli (APC) [119] and PTEN [120]), p53 signaling pathway (p73 [121]), and DNA repair (O6 methylguanine DNA methyltransferase (MGMT) [113,122]).For cervical scrapings, some methylation marker panels of host genes, with sensitivities of over 80% for CIN3+ were evaluated: SOX1/PAX1, SOX1/LMX1A, SOX1/NKX6-1, PAX1/LMX1A; PAX1/NKX6-1, LMX1A/NKX6-1 [123], JAM3/EPB41L3/TERT/C13ORF18 [124], and CADM1/MAL [13, 125], etc. Host gene methylation analysis might be an alternative for hrHPV DNA detection because aberrant methylation can be detected in cervical smears up to 7 years prior to the diagnosis of cervical cancer [126]. On the other hand, for methylation analysis, cervical scrape samples as well as self-collected cervico-vaginal lavage samples can be used [127].As accurate predictor tests, the measurement of DNA methylation in HPV genomes, in certain early (E) and late (L) open reading frames (ORF) as well as in parts of the upstream regulatory region (URR), may have diagnostic value. The hypermethylation in the L1 region was a common feature of cervical cancer but not of CIN induced by HPV16 [128], or HPV18 [129]. But the DNA methylation on multiple CG sites in the L1, L2, E2, and E4 ORFs were significantly associated with CIN2+ after accounting for multiple testing [130]. Some studies have contradictory results because most were quite small and heterogeneous and did not always include (1) comparable sets of specimens (cancer, high-grade CIN, cell lines), (2) exactly the same CG sites, or (3) the same methodology [12]. Overall, as cervical cancer prevention moves to DNA testing methods, DNA-based biomarkers, such as HPV methylation could serve as a reflex strategy to identify women at high risk for cervical cancer [131], but the region with the best predictive value must be established.In addition to the already reported genomic alterations in cervical cancer development by hrHPV, many studies underline the involvement of viral or cellular miRNAs, mainly based on the fact that some RNA micromolecules target transcriptional factors that modulate both cellular and viral gene expression [132,133].In HPV infection, E6 decreases miR-34a [132,134], which is a target of p53, thus the effect of E6 on miRNA-34a is mediated by decreased p53 [132,134]. On the other hand, one of the targets of miR-34a is p18Ink4c [135], an inhibitor of CDK4/6 that promotes the cell cycle. E7 decreases miR-203 during keratinocyte differentiation, which is a tumor suppressor and thus increases carcinogenesis [136] through an increase of cell survival targeting antiapoptotic protein bcl-w [137], induction of G1 cell cycle arrest targeting survivin [138], inhibition of migration and invasion targeting LIM and SH3 protein [139]. E7 upregulates miR-15a, miR-15b, and miR-15b through E2F1 and E2F3 [140,141] and in turn, these miR decrease cyclin E1, leading to cell cycle arrest [142]. A lot of other miRs are upregulated or decreased by virus oncogenes inducing changes in cellular signaling pathways, some of these have not yet been elucidated [143].
Ovaries, paired organs, constitute the female sexual gland with endocrine lunette and also produce ova. The ovary is covered by germinal epithelium (formed from cuboid or cylindrical cells) and subjacent is a thin layer of dense connective tissue. The ovary presents a cortical area (comprised of follicles, corpus luteum, and stroma) and a medulla. Starting from puberty till menopause, there is a growth and maturation of one ovarian follicle during each menstrual cycle and the formation of one corpus luteum after rupture of the follicle and oocyte removal. If the egg is not fertilized, the corpus luteum regresses, undergoes progressive sclerosis forming a hyaline. If the egg is fertilized, the corpus luteum becomes more voluminous and luteal cells increase, constituting the corpus luteum of pregnancy. Ovarian stroma is formed from fibroblastic and mesenchymal cells. Stromal cells present both characters of connective cells and steroid activity (secreting androgens and estrogens). Ovarian medulla consists of lax connective tissue containing blood and lymph vessels, nerves, and embryonic elements. The growth and development of the follicle during the ovarian cycle are driven by two gonadotrophic hormones, secreted by the anterior pituitary: follicle-stimulating hormone (FSH) and luteinising hormone (LH). Both FSH and LH are under the control of gonadotrophin-releasing hormone (GnRH) secreted by the hypothalamus through negative feedback carried out by estrogens that are secreted by thecal cells of the follicle.
Ovarian cancer ranks second after cervical cancer worldwide. On the other hand, ovarian cancer is in seventh place in terms of incidence among malignant tumors in women and eighth with respect to death due to malignant tumors in women worldwide [144]. If approximately 90% of ovarian cancers arise from epithelial cells, 3% are from germ cells and 7% from granulosa-theca cells. Ovarian cancer comprises different types of tumors with widely differing clinicopathologic features and behaviors. Based on clinicopathologic and molecular genetic studies, two histologic types of epithelial ovarian serous carcinomas were established: low-grade serous carcinomas (LGSCs) and high-grade serous carcinomas (HGSCs) [145]. Although they are developed independently along different molecular pathways, both types develop from fallopian tube epithelium and involve the ovary secondarily. Type I tumors (LGSCs) are comprised of low-grade serous, low-grade endometrioid, mucinous, and clear cell carcinomas; typically present as large cystic masses confined to one ovary; have a relatively indolent course; and are relatively genetically stable being associated with mutations in KRAS, BRAF, PTEN, PIK3CA, CTNNB1, ARID1A, and PPP2R1A [146,147] that perturb signaling pathways. Type II tumors (HGSCs) are composed of high-grade serous, high-grade endometrioid, undifferentiated carcinomas and malignant-mixed mesodermal tumors; clinically aggressive and typically present at an advanced stage, which contributes to their high fatality [148]; at the time of diagnosis, they demonstrate marked chromosomal aberrations but over the course of the disease these changes remain relatively stable [149]; approximately 60% of HGSC have the fallopian tube as the origin of serous tumors [150], because the expression profiles of ovarian HGSCs more closely resemble fallopian tube epithelium than the ovarian surface epithelium [151]; they harbor TP53 mutations in over 95% of cases [152,153], but rarely harbor the mutations detected in the low-grade serous tumors; another possible origin of HGSC is from inclusion cysts through a process of implantation of tubal (müllerian-type) tissue rather than by a process of metaplasia from ovarian surface epithelium (mesothelial). Hypermethylation has been found to be associated with the inactivation of almost every pathway involved in ovarian cancer development, including DNA repair, cell cycle regulation, apoptosis, cell adherence, and detoxification pathways [154]. Complete or partial inactivation of the BRCA1 gene through hypermethylation of its promoter has been reported in 15% of sporadic ovarian tumors [155,156], 31% of carcinomas but not in the benign or borderline tumors [157], or in the hereditary type of the disease, nor in samples from women with a germ line BRCA1 mutation [158,159]. On the other hand, hypermethylation of BRCA1 was detected at a significantly higher frequency in serous carcinomas than in tumors of the other histological types [160]. The homeobox genes (HOX), a family of transcription factors that function during embryonic development and control pattern formation, differentiation, and proliferation [161] was associated with ovarian cancers [162]. In addition, based on the high percentage of methylation of the HOXA9 gene observed in 95% of patients with high-grade serous ovarian carcinoma [163,164], it has been suggested that the methylation status of HOXA9 and HOXAD11 genes may serve as potential diagnostic and prognostic biomarkers [163,164]. Some other genes found hypomethylated were associated with progression towards cancer: LINE-1 elements [165], SNGG (synucelin-γ), encoding an activator of the MAPK and Elk-1 signaling cascades [166,167], etc. Overall, DNA hypomethylation may promote tumorigenesis by transcriptional activation of proto-oncogenes and on the other hand loss of imprinting or genomic instability. DNA hypermethylation predisposes to gene mutation because the methylated cytosines are often deaminated and converted to thymine leading to inactivation of tumor suppressor genes. However, these phenomena deregulate the main functions of gynecological cancer cells (Figure 1 and Table 1).  Gynecologic Cancers -Basic Sciences, Clinical and Therapeutic Perspectives miRNA as key players in cell fate decisions are strongly linked to gynecological cancer. But, although the methods to discover miRNA were improved, research is still in progress. Some of these miRNA that have been associated with gynecologic cancers are shown in Figure 2 and Table 2. pecific miRNAs have effects on various molecular pathways, and specific miRNA expression signatures i ynecological cancers can be associated with diagnosis, prognosis, and therapy response. miRNAs can regulate a larg umber of target genes and Table 2 lists the estimated targets.

The roles of microenvironment-mediated epigenetic perturbations in the development of gynecological neoplasia
The complexity that governs the tumor phenotype cannot be explained only at the genetic level, as genetic abnormalities occur with low frequency. Therefore, major attention was focused on the study of the role of tumor microenvironment (TME) not only in tumor initiation but also in progression and metastasis. The hypothesis of cancer cell development and proliferation only in a conducive environment has been made by Paget since 1889 [308]. While Paget suggested that the microenvironment facilitates or inhibits metastasis through growth-promoting/ inhibiting factors, recent research sustains that the tumor is directed into one or several possible molecular evolution pathways by signals originating in native and/or modified microenvironmental factors [309]. The tumor microenvironment consists of epithelial cells, vascular endothelial cells, fibroblasts and myofibroblasts, macrophages, leukocytes, and the extracellular matrix (ECM). Together with the ECM, these nonmalignant cell types constitute the stromal tissue of the tumor that secretes ECM components, cytokines, and growth factors involved in tumor growth and invasion. All these components are dynamically interconnected around the tumor. In the tumorigenesis process, studies have shown the critical role of chronic inflammation by hyperexpression of the inflammatory mediators in the microenvironment. The inflammatory microenvironment is both the result of genetic alterations in cancer cells and of the tumor-infiltrating cells that produce inflammatory mediators [310].
While normal fibroblasts prevent tumor progression, cancer-associated fibroblasts (CAFs) that display a different secretory pattern generate an environment that favors tumor growth and invasiveness. Tumor formation is characterized by changes in cell behavior, like accelerated growth with loss of tissue architecture and epithelial dysfunction, angiogenesis, stromal activation, and migratory and invasive features. Therefore, dysfunction in the tumor microenvironment, in addition to epithelial dysfunction, is crucial for carcinogenesis as altering its components leads to impaired immune response. TME promotes tumorigenesis through new blood vessel formation. Although studies have suggested that some cells in TME contained mutations, recent data pointed, first, to the presence of mutations only in tumorigenic cells and second, to the contribution of these mutations to epigenetic changes in both nontumorigenic cells and TME. In turn, the cells in the microenvironment produce epigenetic changes in tumor cells reflected in their pattern of differentiation [311] and animal models demonstrate that the tumor microenvironment can induce epigenetic alterations and changes in gene expression in tumors [312].
It was suggested that the epigenome serves as the interface between the genome and the environment [313,314]. The epigenetic role of TME in growth induction seems to be linked with transforming growth factor (TGF)-β and its receptor, whose expressions are regulated through chromatin remodeling [315], although no research on stromal fibroblasts was performed. TGFβ pathways are involved in the oncogenesis process, acting either as tumor suppressor or as tumor promotor, depending on TME crosstalk in the tumor microenvironment [316]. In malignant progression, epigenetic changes in the expression of 12 genes responsive to the TME stress suggest that coordinated transcriptional response of eukaryotic cells to microenvironment might be correlated with chemotherapy resistance of solid tumors [317]. Since tumor development is lead by physiological responses to an aberrant stromal environment, the interaction between the tumor and stromal cells determines tumoral progression [318]. In the chemokine network, epigenetic silencing of CXCR4 in SDF-1α/CXCR4 signaling of tumor microenvironment of cervical cancer cell lines and primary biopsy samples limited the cell response to the paracrine source of SDF-1α, which lead to loss of cell adhesion and disease progression [319]. Other authors reported miRNA's contribution to cancer progression and metastasis. While extracellular miRNAs are involved in cell-cell communication and stromal remodeling [320], specific intracellular ones lead to cell proliferation through cancer-associated fibroblast activation [321].
The acquisition of invasive properties in tumor cells seems to be partially linked to epithelialmesenchymal transition by abrogation of homotypic cell-cell adhesion due to the absence of E-cadherin expression. Starting from the important role of transient E-cadherin expression in neoplasia, DesRoches and collaborators investigated its regulation by the microenvironment. Using 3D human tissue constructs, the authors suggested the role of epigenetic changes (DNA methylation, chromatin remodeling, and specific miRNA regulation) in the plasticity of Ecadherin-mediated adhesion in different tissue microenvironments during tumor cell invasion and metastasis [322]. The entry of the epithelial cells into the stroma is promoted through the E-cadherin intercellular junction disruption by MMP-3 and break down of the ECM collagen fibers by MMP-2 and MMP-9 [323]. MicroRNA suppression also influences the changes involved in epithelial-mesenchymal transition [324]. Reexpression of E-cadherin might reestablish cell-cell adhesion and may result in a mesenchymal-epithelial transition that might lead to proliferative growth of metastases.
Metastasis, as a multistage process (tumor cell migration from primary tumor, invasion of the surrounding tissues, intravasation into the circulation or the lymphatic system metastasis) involves communication with surrounding nonneoplastic cells [325] that can be epigenetically modulated to lead to ECM remodeling. Also, the epigenetic changes in the microenvironment have a significant impact on distant metastasis. In order to create a favorable local environment for cell proliferation in the metastatic sites, carcinoma cells induce epigenetic changes in both the stromal cells and bone marrow-derived cells [326]. The bone marrow cells are mobilized by the primary tumors to the metastatic sites before the actual metastasis creating a suitable microenvironment for metastasis [315,327].
Due to their reversal character, epigenetic changes of TME might be targeted for controlling diseases and for therapeutic approach as drug resistance seems to also depend on TME. But, chemotherapeutic drug resistance depends at least partly on the TME rather than the tumor itself [328] and the combined treatment of both the tumor and the TME may be more efficient in the fight with cancer [315].

Molecular and epigenetic factors involved in drug resistance
Chemotherapy success is challenged by a multitude of intrinsic or acquired, molecular, genetic and epigenetic factors involved in drug transport, detoxification, signal transduction, gene expression, DNA repair, and programmed cell death. Drug resistance is a major challenge that chemotherapy should overcome. Even if the drug itself is efficient in destroying cancer cells, it is much more complicated to avoid triggering resistance than might appear at different levels of interaction between the drug and its cellular components.
The efflux mechanism is considered to be mainly responsible for the multiple drug resistance phenotypes in gynecologic cancers as well as in all types of cancers [329]. The process may be managed by cancer cells at the genetic and/or epigenetic level. While the genetic modifications of MDR1 and related multidrug resistance proteins were intensely explored over the past few decades, the contribution of epigenetic modification to the expression of MDR1 remains insufficiently explored in human gynecological cancers. It was observed that MDR1 was hypermethylated in 100% of ovarian cancer cell lines, and in 5 out of 13 (38%) primary ovarian cancers associated with loss of MDR1 mRNA expression in ovarian cancer cell lines, sustaining the importance role of epigenetic regulation in the expression of MDR1 and clinical treatment outcomes in human ovarian cancer [330]. However, in six ovarian cancer cell lines-W1MR, W1CR, W1DR, W1VR, W1TR, and W1PR that are respectively resistant to methotrexate, cisplatin, doxorubicin, vincristine, topotecan, and paclitaxel, P-gp is responsible for chemoresistance and, in the case of methotrexate, was found to have a relation between the MRP2 transcript level and drug resistance [331]. Among inhibitors of Pgp MDR, valspodar, an analog of cyclosporine A, showed no clinical benefit in a phase III trial with paclitaxel and carboplatin [332], because while these agents can block drug efflux at the cellular level, the effects are not tumor specific, requiring a reduction in dosage for minimizing the side effects but also the therapeutic advantage. On the other hand, miRNA was involved in resistance through the regulation of MDR proteins at a posttranscriptional level. The interaction of miRNAs with the targeted mRNA can downmodulate MDR proteins improving the response to anticancer drugs. It was described [329] that miR-223 can downregulate ABCB1 and mRNA levels. miR-124a and miR-506 significantly decreased the protein level of MRP4 (ABCC4), which is another efflux membrane transporter; however, these miRNAs did not change the gene transcription levels [333]. In addition, although there are many modalities acting on efflux proteins in order to circumvent drug resistance, their effective action can be compromised due to the diversity of signal transduction pathways involved in transporter-mediated MDR, such as MAPK, JNK, PI3K, among others; as well as some transcription factors, like NF-κB, TNFα, and PTEN that could influence the levels of carrier proteins in different conditions [334].
Also, the signal transduction pathways can be involved in drug resistance. The Wnt signaling pathway, which is regulated by a multiprotein complex consisting of, among others, members of β-catenin, adenomatous polyposis coli APC, Axin, and GSK-3β [335], are involved in calcium-dependent cell adhesion due to the interaction between β-catenin and cadherin [336]. Different mutations in APC, promotes β-catenin proteolysis and reduces its transcriptional activity. PTEN, a lipid and protein phosphatase that is a negative regulator of phosphatidylinositol 3 (PI-3) kinase-dependent signaling interacts with the WNT pathway by impeding activation of integrin-linked kinase (ILK), which inhibits GSK-3β and thus causes accumulation of β-catenin [337]. The WNT signaling pathway is the most frequently altered pathway in the majority of cancers; therefore, individual components of the pathway are interesting targets for epigenetic inactivation. PI3K/Akt is another signaling pathway that is involved in acquired resistance of many cancers including gynecological ones. All of its isoforms (Akt1, Akt2, and Akt3) are activated (phosphorylated) by phosphatidylinositol 3-kinase (PI3-K) in response to growth factors and promote cell survival. It was demonstrated that the Akt pathway is directly related to the resistance of cancers against different drugs like sorafenib, trastuzumab, and erlotinib [329]. The epigenetic control of Akt and NF-κB is important for the establishment of drug resistance. RUNX3 suppresses Akt1 transcription by directly binding to the Akt1 promoter, and methylation of RUNX3 induces activation of the Akt signaling pathway [329].
Acquired resistance may develop additionally as blockage of apoptotic pathways or defective apoptotic signaling, often associated with loss of tumor suppressor protein p53, but also independent of p53, alteration of the control points of the cell cycle, increased ability to repair DNA, increased DNA damage tolerance, oncogene induction, and downmodulation of tumor suppressor genes. Eluding the normal process of programmed cell death is already known as a crucial strategy for cancer development and progression, but even more importantly, its participation in the intrinsic or acquired resistance of cancer cells to chemotherapy and radiation. Identification of the points of therapeutic intervention could potentially open up more efficient treatment opportunities. Epigenetic strategies might also be a feasible strategy to reactivate apoptosis or on the contrary to inactivate apoptosis-related genes that inhibit the process. However, it has now been demonstrated that inhibitors of DNA methylation and histone deacetylases can reactivate expression of tumor suppressor genes and induce histone hyperacetylation in the tumors of patients with cervical cancer after treatment with these agents. Preclinical studies have suggested a multitude of strategies to prevent or overcome resistance, but these approaches have not successfully translated to clinical practice yet [338].

Conclusions
This chapter underlined the importance of epigenetic events in gynecological cancer. Deciphering the relevant epigenetic changes associated with each step of tumor development might improve molecular diagnostic and cancer risk assessment. Advances in elucidating epigenetic regulation in cancer disease, as well as in the development of technology, lead to the identification of potential biomarkers for diagnostic screening. As epigenetic changes occur early in neoplastic process, epigenetic biomarkers seem to be more sensitive and specific in cancer detection and some have already been tested for several types of cancer, alone or in combination with traditional biomarkers. Unlike genetic changes, epigenetic alterations are essentially reversible and allow plasticity. These features are exploited and new therapeutic agents targeting epigenetic processes have been developed. The epigenetic changes of the transformed cells or TME can be modified by chemotherapeutic drugs and this epigenetic reversal therapy has potential in the future. In addition, miRNAs should be heavily explored as they might represent future alternatives for combined therapy of cancer. Many epigenetic targets are druggable and in order to overcome drug resistance, epigenetic therapy might also be a feasible strategy for induced cell death. Moreover, epigenetic patterns might be useful tools for therapy response prediction.