Endometriosis: Where are We and Where are We Going?

Abstract

Endometriosis currently affects ∼5.5 million reproductive-aged women in the U.S. with symptoms such as painful periods (dysmenorrhea), chronic pelvic pain, pain with intercourse (dyspareunia), and infertility. It is defined as the presence of endometrial tissue outside the uterine cavity and is found predominately attached to sites within the peritoneal cavity. Diagnosis for endometriosis is solely made through surgery as no consistent biomarkers for disease diagnosis exist. There is no cure for endometriosis and treatments only target symptoms and not the underlying mechanism(s) of disease. The nature of individual predisposing factors or inherent defects in the endometrium, immune system, and/or peritoneal cavity of women with endometriosis remains unclear. The literature over the last 5 years (2010-2015) has advanced our critical knowledge related to hormones, hormone receptors, immune dysregulation, hormonal treatments, and the transformation of endometriosis to ovarian cancer. In this review, we cover the aforementioned topics with the goal of providing the reader an overview and related references for further study to highlight the progress made in endometriosis research, while concluding with critical areas of endometriosis research that are urgently needed.

Introduction

Endometriosis is an estrogen-dependent gynecological condition characterized by the presence and growth of ectopic endometrial tissue, often associated with inflammation, severe and chronic pain, and infertility (Hickey et al. 2014). Lesions identified during laparoscopy are categorized as superficial peritoneal lesions, endometriomas, or deep infiltrating nodules, with high degree of individual variability in lesion color, size, and morphology. Histopathological analysis requires the presence of at least two features for a diagnosis of endometriosis, the features being endometrial epithelium, endometrial glands, endometrial stroma, and hemosiderin-filled macrophages (Hsu et al. 2010). Retrograde menstruation, in which uterine epithelial and stromal cells are disseminated and implanted into the peritoneal cavity via the fallopian tubes, is the most accepted mechanism for the pathogenesis of endometriosis (Sampson 1927b, Ahn et al. 2015a). Greater than 90% of women undergo retrograde menstruation; however, the prevalence of endometriosis in the general population is 6-10% (Sampson 1927a, Syrop & Halme 1987). Such a discrepancy between these two values suggests women who develop endometriosis are likely to have other genetic, biochemical, and pathophysiological factors contributing to development of the disease (Ahn et al. 2015a).

The goal of this review is to provide a broad overview of the advancements in endometriosis research over the last 5 years (2010-2015). First, we delve into animal models often used in endometriosis research. After which, we cover critical areas of endometriosis study, including basic and clinical research, and the transformation of endometriosis into ovarian cancer. Within basic research, we focus on angiogenesis, cytokine/chemokine expression, and hormones and their receptors, and the significance they may play in the pathogenesis of endometriosis. This review is a synopsis of important findings for researchers to quickly find relevant sources of interest to his/her studies.

Animal Research Models

The use of animal models in the study of endometriosis allows for the control of numerous variables related to pathogenesis and disease progression, including angiogenesis, inflammation, and hormonal response. Non-human primate and rodent models are the most common animal models used, while the chicken chorioallantoic membrane model has limited use.

Non-human Primate Models

Non-human primates (baboons and rhesus macaques) are often used to study pathogenesis, progression, and treatment of endometriosis. While primates can spontaneously develop endometriosis at a low prevalence (D'hooghe et al. 1996, Zondervan et al. 2004, King et al. 2015), techniques have been developed to increase disease incidence. Cervical occlusion to promote retrograde menstruation (Scott et al. 1953, D'Hooghe et al. 1994) and a homologous model, in which endometrial tissue is excised from a donor primate and surgically transplanted or injected into a recipient primate, are used (Te Linde & Scott 1950, D'Hooghe et al. 1995, Sillem et al. 1996). Primate models, including advantages and disadvantages, have been previously described (Tirado-Gonzalez et al. 2010, Grummer 2012, King et al. 2015).

Rodent Models

Rodents are often used in endometriosis research due to quick generation time, ability for genetic manipulation, and relatively low cost, especially in comparison to non-human primate models. Rodent models of endometriosis are divided into two main groups: heterologous or homologous/autologous models. Heterologous models use human tissue transplanted into immunocompromised mice, while homologous models involve transferring endometrial tissue from one animal to a syngeneic animal (Tirado-Gonzalez et al. 2010, King et al. 2015).

Heterologous models involve the transfer of human endometrial tissue into an immunocompromised rodent, such as athymic nude, severe combined immunodeficient (SCID), or Rag2γ(c) mice, to prevent the rodent immune system from attacking the foreign tissue (Zamah et al. 1984, Aoki et al. 1994, Greenberg & Slayden 2004). Once human tissue is collected, it is disseminated via intraperitoneal or subcutaneous injection into the immunocompromised rodent. Heterologous rodent models with associated advantages and disadvantages have been described (Tirado-Gonzalez et al. 2010, Bruner-Tran et al. 2012, Grummer 2012, King et al. 2015).

Several homologous rodent models are utilized in endometriosis research, and the generation of these models involves several important considerations regarding the reproductive status of the donor and recipient, transplantation method, and potential genetic manipulation (King et al. 2015). Often, the recipient rodents are ovariectomized and treated with estrogens to promote lesion growth (Cummings & Metcalf 1995, Somigliana et al. 1999, Styer et al. 2008, Burns et al. 2012). Critically valuable for the study of endometriosis is that the homologous model maintains an intact immune system. A large difference between homologous models is the method of transplantation and tissue dissemination. Various models exist for the development of ectopic lesions, including: 1) suturing uterine tissue into the peritoneal wall or intestinal mesentery, 2) injecting minced uterine tissue intraperitoneally to disperse freely and attach at sites within the peritoneal cavity, 3) using entire uterine tissue or endometrial tissue, and 4) using minced “menstruated” tissue for intraperitoneal injection (Vernon & Wilson 1985, Somigliana et al. 1999, Burns et al. 2012, Greaves et al. 2014). For models used to study endometriosis, it is critically important to remember the definition of and requirements for an endometriotic lesion. Discouragingly, some models inherently do not fulfill these criteria and are suboptimal for the study of endometriosis, ultimately occluding comprehensive comparison and interpretation of data in the scientific literature.

Chicken Chorioallantoic Membrane Model

The chicken chorioallantoic membrane (CAM) assay is used to study molecular processes involved in adhesion, invasion, and angiogenesis of developing endometriotic lesions. This assay involves culturing human endometrial tissue on the CAM of fertilized chicken embryos (Maas et al. 2001). The CAM has a dense microvasculature, useful for examining angiogenesis and for experimentation with anti-angiogenic agents (Nap et al. 2005). This method has been used to study the impact of matrix metalloproteinase (MMP) expression and activity on adhesion and invasion (Nap et al. 2004, Juhasz-Boss et al. 2010). However, it is not suited for studying immunological or inflammatory aspects of lesion development or for potential effects of systemic treatments.

Pathogenesis and Progression of Endometriosis

Animal models and human samples are paramount in the study of pathogenesis and progression of endometriosis. They allow for in-depth analysis of factors involved in this disease, including inflammation, angiogenesis, cytokine/chemokine expression, and endocrine alterations such as steroid and steroid receptor expression. These components also form a complex, interacting system greatly impacting the development of endometriosis. While we understand that several other factors are involved in the pathogenesis and progression of this disease, including genetics and epigenetics, and significant advances in these components have been made, covering them in depth is beyond the scope of this review. For researchers interested in these topics, an elegant and comprehensive review by Bulun et al. (2015) recently addresses the molecular biology, genetics, and epigenetics of endometriosis and covers 25 years of research (1990-2015).

Inflammation - Angiogenesis

Angiogenesis is the formation of new blood vessels, and subsequently, is a key process to form functional blood vessels to ectopic menstrual tissue for the establishment/maintenance of endometriotic lesions. Theorized is that women with endometriosis respond to retrograde menstrual tissue as a “wound” that must be healed and not as “self” that must be removed (Herington et al. 2011). Examining key players involved in angiogenesis, both in women with endometriosis and in animal models, similarities between angiogenesis in endometriotic lesions and angiogenesis in wound healing exist. A variety of growth factors and genes related to angiogenesis have been studied in endometriosis.

The VEGF protein family is well known for roles in angiogenesis, vasculogenesis, and lymphangiogenesis. Human peritoneal fluid (PF) from women with endometriosis show inconsistent protein levels of VEGF, but this may be due to sample size, dilution of PF, or true variability among women. For example, some studies show increased VEGF levels in the PF (Bourlev et al. 2010, Xu et al. 2013, Szubert et al. 2014); however, other studies show no increase in VEGF levels in women with endometriosis compared to healthy women (Barcz et al. 2012, Bersinger et al. 2012, Rathore et al. 2014). Interestingly, more consistency is found in animal models of endometriosis, most likely because of controlled onset of experimental conditions. A variety of rodent models of endometriosis show VEGF levels increase in endometriosis-like lesions (Machado et al. 2010, Ricci et al. 2011, Kumar et al. 2014, Lu et al. 2014a, Machado et al. 2014, Zhao et al. 2015). Data are inconsistent when attempting to target VEGF in the mouse to treat endometriosis (Xu et al. 2011, Novella-Maestre et al. 2012, Virani et al. 2013, Kumar et al. 2014); however, the data suggest VEGF drives angiogenesis in endometriosis. Furthermore, these results from human and animal models demonstrate challenges of clearly deciphering VEGF as an appropriate marker for endometriosis.

Other angiogenic factors play important roles in the adhesion and maintenance of endometriosis lesions, including hypoxia factors (ie. HIF1A), MMPs (ie. MMP9), and microRNAs (miRNA). As mentioned, the peritoneal microenvironment of women with endometriosis is often different from healthy women. In a heterologous mouse model, hypoxic conditions promote angiogenesis and proliferation of endometriosis demonstrated by larger lesions, higher levels of VEGF, HIF1A, Ki67, and PECAM1 (Lu et al. 2014b). Concomitantly, the same group shows lesion location affects adhesion and angiogenesis when comparing intraperitoneal versus subcutaneous endometriotic tissue injection (Lu et al. 2014a), suggesting the microenvironment of the peritoneal cavity plays a crucial role in lesion adhesion and angiogenesis. MMPs are proteases required for reorganizing existing blood vessels during budding angiogenesis (Page-McCaw et al. 2007). Recently, the role of MMPs in endometriosis were not studied in-depth; but MMPs play a known role in endometriosis (Machado et al. 2010). For example, Mmp9^-/- uterine tissue does not grow in a mouse suture endometriosis model (Han et al. 2012); however, this model does not account for actual tissue attachment. An emerging field in endometriosis is the function of miRNA in angiogenesis. Primary eutopic and ectopic endometrial stromal cells exposed to PF from women with endometriosis have downregulated miRNAs known to regulate VEGF expression compared to cells exposed to PF from control women (Braza-Boils et al. 2013, Braza-Boils et al. 2014, Braza-Boils et al. 2015). Future in-depth analysis of the interplay between inflammation and angiogenesis in the early stages of endometriosis development is needed to determine which molecules could potentially be targeted therapeutically.

Inflammation - Cytokine and Chemokine Expression

Cytokines and chemokines are emerging as key players in endometriosis pathobiology. Cytokines are a broad group of secreted proteins important in cell signaling, while chemokines are a family of cytokines important in inducing chemotaxis in nearby cells. A complete overview of chemokines and cytokines in endometriosis is too exhaustive; however, these proteins are altered in PF, ectopic lesions, eutopic endometrium, and serum. To demonstrate the growing role of cytokine research in the study of endometriosis, Table 1 lists cytokines and chemokines that appear in > two endometriosis research papers between 2010-2015. Additional to the dysregulation of cytokines/chemokines, altered levels of a large number of cytokines/chemokines are found in cyst fluid removed from endometriomas/chocolate cysts (Chen et al. 2013b). Before the interplay and implications of chemokines and cytokines can be elucidated in endometriosis, large scale controlled human studies or meta-analyses will need to be conducted to fully encompass cytokine dysregulation. Most likely, with the signaling complexity of the immune system and endometriosis as a disease, a single chemokine/cytokine will not diagnose disease, but instead, a disease profile of altered cytokines may be used to establish disease diagnosis. Furthermore, as nicely outlined in Fassbender et al., international standardized methods for BioBanking endometriosis samples needs to be implemented (Fassbender et al. 2013).

Table 1. Cytokine or Chemokine Alterations in Endometriosis.

Cytokine/chemokine	Model	Results	Reference
CCL2 (MCP-1)	hPF	+, ++, +++	(Mier-Cabrera et al. 2011, Margari et al. 2013), (Tao et al. 2011),(Bersinger et al. 2012)
CCL5 (RANTES)	hPF	NS, +, ++	(Bersinger et al. 2012, Margari et al. 2013), (Mier-Cabrera et al. 2011),(Beste et al. 2014)
	hEctopic	qualitative +, ++	(Wang et al. 2010),(Yang et al. 2013)
CCL11 (Eotaxin)	hPF	++, +	(Bersinger et al. 2012), (Mier-Cabrera et al. 2011)
CXCL1 (GROa)	hPF	Nearing	(Bersinger et al. 2012)
CXCL8 (IL8)	hPF	NS,+, ++	(Velasco et al. 2010, Bersinger et al. 2012), (Mier-Cabrera et al. 2011, Beste et al. 2014), (Milewski et al. 2011, Malhotra et al. 2012)
	hSerum	++	(Carmona et al. 2012)
	hEESCs	++	(Delbandi et al. 2013)
	hPF mRNA Cells	NS	(Yeo et al. 2013)
CXCL10 (IP-10)	hLesion	NS trend –	(Bellelis et al. 2013)
	hPF	++	(Bersinger et al. 2012)
CXCL12 (SDF1)	hEctopic	+	(Bellelis et al. 2013)
	hPF	+	(Leconte et al. 2014)
IL1B	hPF	+, +++	(Mier-Cabrera et al. 2011, Beste et al. 2014), (Sikora et al. 2012)
	hEctopic	+	(Chen et al. 2013)
	hPF cells	NS	(Yeo et al. 2013)
IL4	hPF	NS, ++	(Mier-Cabrera et al. 2011, Wickiewicz et al. 2013), (Beste et al. 2014)
IL6	hPF	NS, +, ++, +++	(Rathore et al. 2014), (Mier-Cabrera et al. 2011, Khan et al. 2015), (Velasco et al. 2010, Kang et al. 2014), (Milewski et al. 2011, Bersinger et al. 2012, Podgaec et al. 2012, Wickiewicz et al. 2013)
	hSerum	+, ++, +++	(Kinugasa et al. 2011), (Carmona et al. 2012), (Elgafor El Sharkwy 2013)
	hEESC	++	(Delbandi et al. 2013)
	hPF mRNA Cells	NS	(Yeo et al. 2013)
IL10	hSerum/PF	NS/NS	(Andreoli et al. 2011)
	hPF	NS, +	(Bersinger et al. 2012, Podgaec et al. 2012), (Mier-Cabrera et al. 2011, Wickiewicz et al. 2013)
	hPF mRNA Cells	NS	(Yeo et al. 2013)
IL12	hSerum/PF	NS/NS	(Andreoli et al. 2011)
	hPF	NS	(Mier-Cabrera et al. 2011)
	hPF mRNA Cells	NS	(Yeo et al. 2013)
IL17A	hEndo, Serum	**/+	(Ahn et al. 2015)
	hSerum/PF	NS/NS	(Andreoli et al. 2011)
	hPF	NS	(Podgaec et al. 2012)
	hFF/Serum	+++/+++	(Sabbaghi et al. 2014)
IL18	hPF	+, ---	(Bersinger et al. 2012), (Sikora et al. 2012)
IL22	hEctopic	**	(Guo et al. 2013)
	hSerum	--	(Santulli et al. 2013)
IFNG	hPF	NS, +	(Wickiewicz et al. 2013),(Mier-Cabrera et al. 2011)
	hPF mRNA Cells	NS	(Yeo et al. 2013)
MIF	hPF	++	(Beste et al. 2014)
	hEctopic	++	(Lin et al. 2010)
TGFB	hPF	+++	(Podgaec et al. 2012)
	mKO w/human	- size	(Hull et al. 2012)
TNFA	hPF	NS, +, ++	(Tao et al. 2011, Wickiewicz et al. 2013),(Mier-Cabrera et al. 2011, Beste et al. 2014, Young et al. 2014a, Young et al. 2014b), (Khan et al. 2015)
	hEctopic cd56+	+	(Chen et al. 2013)
	pfNKcell	+	(Funamizu et al. 2014)
	bEctopic	NS	(Ilad et al. 2010)
	hPF mRNA Cells	NS	(Yeo et al. 2013)

Increased levels:

⁺= p<0.05,

⁺⁺= p<0.01,

⁺⁺⁺= p<0.001,

^**= qualitative IHC,

Decreased levels: - = p<0.05; NS = non-significant

h=human, b=baboon, Ectopic=ectopic endometriosis lesion, PF=peritoneal fluid, FF=follicular fluid, EESC=ectopic endometrial stromal cells

Hormones and Hormone Receptors

Endometriosis is intimately associated with steroid metabolism and associated pathways, corresponding to the paramount roles estrogen receptors (ESRs) and progesterone receptors (PGRs) play in uterine biology. Both human and animal model studies show endometriosis is estrogen (E2) dependent and is regulated through the ESRs alpha and beta (ESR1 and ESR2) (Burns et al. 2012, Pellegrini et al. 2012, Wu et al. 2012, Han et al. 2015, Zhao et al. 2015). An increased ratio of ESR2 to ESR1 mRNA is observed in endometriomas compared with endometriosis implants and eutopic endometrium (Bukulmez et al. 2008). Knockout studies in mice show lesion attachment, size, and proliferation are closely associated with the presence or absence of Esr1 and Esr2 (Burns et al. 2012). The Bulun Laboratory have focused efforts on ESR2 and demonstrate ESR2 expression is highly increased in endometriotic tissue due to hypomethylation of the promoter region (Dyson et al. 2014). They also identify RAS-like estrogen-regulated growth inhibitor (RERG) as a key enzymatic target of estradiol signaling through ESR2. This enzyme regulates numerous factors involved in the progression of endometriosis, including cell proliferation and apoptotic resistance (Monsivais et al. 2014). Additionally, they have nicely detailed multiple studies on the role of ESR2 in endometriosis in a comprehensive review (Bulun et al. 2012). Use of estrogen receptor ligands, inhibitors, and agonists also support the role of ESRs in endometriosis (Colette et al. 2011, Kulak et al. 2011, Han et al. 2012, Chen et al. 2014, Naqvi et al. 2014, Palmer et al. 2015, Zhao et al. 2015). Specifically, selective estrogen receptor modulators (SERMs) are synthetic molecules which bind to ESRs and act either as antagonists or agonists. Two compounds, chloroindazole (CLI) and oxabicycloheptene sulfonate (OBHS), have strong ER-dependent anti-inflammatory effects on endometriosis lesions in vivo in a suture mouse model of endometriosis and in vitro, with primary human endometriotic stromal cells (Zhao et al. 2015). Their data suggests that both CLI and OBHS inhibit the establishment of new lesions and reduce the size of already established lesions; however, important next studies using these inhibitors will be to examine lesion attachment without a suture endometriosis model, as suturing alone creates an unnecessary inflammatory response similar to any reaction towards a foreign body, (Carr et al. 2009) and, in some respects, negates the use of homologous tissue.

Progesterone (P4) and its receptor isoforms, PGR-A and –B, also have established roles in endometriosis. The endometrium of women with endometriosis demonstrates an attenuated response to P4 because PGR responsive genes are not suppressed in the eutopic endometrium of women with endometriosis compared to healthy women in the early secretory phase of the menstrual cycle, suggesting the presence of a progesterone resistance phenotype in these women (Burney et al. 2007). A more recent study to discriminate between the PGR isoforms finds elevated levels of PGR-A in endometriosis lesions and eutopic endometrium from women with endometriosis and shows a PGR-A-dominant state, regardless of menstrual phase (Bedaiwy et al. 2015). While the data is from a small cohort of women, their findings suggest a PGR-A-dominant menstrual efflux in the peritoneal cavity may mirror the growth and invasive properties known about cancers overexpressing PGR-A.

Aromatase is the enzyme responsible for the aromatization of androgens into estrogens. Aromatase protein level is increased in vaginal septum lesions and decreased in intestinal lesions in women with endometriosis (Goncalves et al. 2015). Ovarian endometriomas express higher levels of aromatase and peroxisome proliferator-activated receptor gamma co-activator 1 alpha (PPARGC1A) than associated ectopic lesions and eutopic endometrium (Suganuma et al. 2014). Activation of peroxisome proliferator-activated receptor gamma (PPARG) inhibits the growth and survival of human endometriotic cells by suppressing E2 biosynthesis and prostaglandin E2 (PGE2) signaling (Lebovic et al. 2013). The use of AIs for the treatment of endometriosis is becoming more common and is discussed below.

The last 5 years have expanded our knowledge of hormones, hormone receptors (HRs), and associated co-regulators. These studies are important for integrating dysregulation found in ectopic lesions, but also have allowed for the design of more targeted areas to be studied. More in-depth studies with targeted HR uterine knockouts, co-regulator knockouts, and/or with the recently synthesized SERMS will lead to greater understanding of the role of HR in disease. The results of these future experiments will allow for even more targeted experiments and hopefully development and use of more targeted therapeutic paradigms.

Interactions between Inflammation and the Endocrine System

Cross-talk between the immune/inflammatory and endocrine systems can significantly impact pathogenesis and progression of endometriosis. The sex hormone receptors can markedly alter the immune response in ectopic tissue. Both ESR1 and ESR2 have distinct roles in regulating the immune response, as discovered through the use of multiple animal models. Signaling of E2 through ESR1 appears to have both an anti- and pro-inflammatory role, as observed by increased mitogenesis and decreased IFNG, TNF, and IL12 transcript expression (Burns et al. 2012). Overexpression of ESR2 activates the inflammasome and modulates TNF-induced apoptosis, as observed with increases in IL1B and cleaved caspase-1 levels and decreases in cleaved caspase-8 levels in ectopic lesions (Han et al. 2015).

Hormones themselves also directly alter the immune system. Monocyte chemotactic factor-1 (MCP1/CCL2) is an example of a chemokine significantly affected by sex hormones. In human endometrial endothelial cells from women with endometriosis, both E2 and P4 increase MCP1 mRNA and protein expression; this effect is not observed in cells from healthy women (Luk et al. 2010). After treating monocytes with control peritoneal fluid (cPF) or endometriotic peritoneal fluid (ePF), the addition of E2 to the culture suppresses MCP1 release from cPF-treated monocytes. However, E2 does not suppress MCP1 release from ePF-treated monocytes (Lee et al. 2012). E2 promotes a pro-inflammatory environment by increasing secretion of IL6 and TNF from peritoneal macrophages from women with endometriosis compared to control women. This effect is further enhanced by co-treatment with lipopolysaccharide (Khan et al. 2015). Other chemokines, CXCR4 and CXCL12, are downregulated by sex hormones in human epithelial endometrial cells and human endometrial stromal cells respectively (Ruiz et al. 2010). All of these findings provide evidence that the immune environment and its response to sex hormones is altered in women with endometriosis; however, a definitive mechanism for these differences is largely unknown and is a major area that future research needs to address.

A third aspect to the endocrine-immune crosstalk involves aromatase expression. Macrophage migration inhibitory factor (MIF) increases aromatase mRNA and protein expression in ectopic endometrial stromal cells via posttranscriptional stabilization (Veillat et al. 2012). Interestingly though, E2 treatment in the same cells increases MIF mRNA and protein expression, suggesting a positive feedback loop between the endocrine and immune systems in women with endometriosis (Veillat et al. 2012). Potential for a continuous positive feedback loop between these systems is an area for further exploration to understand the dynamic and altered environment in women with endometriosis.

Clinical Symptoms and Diagnosis of Endometriosis

Endometriosis is often characterized by pelvic pain that manifests in a variety of ways; most commonly, patients present with dysmenorrhea, noncyclical pelvic pain, and dyspareunia, but other common symptoms are dyschezia, dysuria, and infertility (Fritz MA 2011, Practice Committee of the American Society for Reproductive 2014). Definitive diagnosis of endometriosis is by visualization or excision of lesions via laparoscopy. The American Society for Reproductive Medicine (ASRM) grading system for endometriosis guides surgeons in determining the severity of disease (1997) and was created to help predict pregnancy with fertility treatment. The grading system does not correlate with pain level, and has limited reproducibility to predict pregnancy; however; it remains the best objective way to communicate disease severity between physicians and surgeons.

Accuracy of visual diagnosis increases with disease severity (Fernando et al. 2013). While the European Society of Human Reproduction and Embryology (ESHRE) requirements suggest surgical diagnosis by visualization alone is appropriate, ASRM stresses that biopsies be taken when diagnosis is unclear (Fernando et al. 2013). Importantly, poor correlation exists between clinical symptoms and disease burden (Dunselman et al. 2014, Practice Committee of the American Society for Reproductive 2014). Diagnosing endometriosis by pelvic pain alone is not sufficient, as pelvic pain is also a symptom of many other diseases, including pelvic adhesions, adenomyosis, and gastrointestinal urologic disorders (Bulun 2009, Practice Committee of the American Society for Reproductive 2014). This vast differential diagnosis for pelvic pain can complicate the diagnosis of endometriosis.

Treatments

Several different treatment modalities, including medical, surgical, and alternative, exist for endometriosis. First-line medical management includes options that have a favorable safety and cost profile, are well tolerated by the patient, and are effective in treatment (Zito et al. 2014). If medical therapy fails, surgical therapy to remove endometriotic lesions and endometriomas is performed. Finally, alternative therapies are being used to supplement conventional treatments.

Medical Therapy

Combined oral contraceptive pills (OCPs), which include ethinyl estradiol (EE) and various progestins, are used to treat endometriosis, particularly in women not trying to conceive (Practice Committee of the American Society for Reproductive 2014, Zito et al. 2014). Historically, OCPs have been first-line therapy, but most studies are decades old and the pills contained higher doses of EE. Based on a more recent randomized control trial (RCT) in 100 patients, low dose OCPs decrease pain more significantly than placebo on the Visual Analog Scale (VAS) (Harada et al. 2008). Continuous OCPs decrease recurrence rates of dysmenorrhea after surgical therapy when compared to cyclic OCPs (Muzii et al. 2015, Zorbas et al. 2015). Of the progestins, the 19-nortestosterone derivatives are less androgenic and offer better side effect profiles (Angioni et al. 2014). In a RCT, dienogest significantly decreases endometriosis-related pain similar to gonadotropin releasing hormone agonists (GnRHa), both as initial and post-operative therapy, without the negative side effect profile of GnRHa (Angioni et al. 2014, Andres Mde et al. 2015, Granese et al. 2015, Strowitzki et al. 2015). Levonorgestrel, delivered through an intrauterine system after conservative surgery, significantly decreases dysmenorrhea, dyspareunia, and non-cyclic pelvic pain compared to expectant management in a RCT of 55 patients (Tanmahasamut et al. 2012, Imai et al. 2014).

Several therapies aim to create a hypoestrogenic state in women with endometriosis. Examples of these treatments include GnRHa, GnRH antagonists (GnRH-ant), synthetic androgens, and AIs. GnRHa therapy downregulates gonadotropin receptors and desensitizes the body to gonadotropins. It decreases pain and endometriotic nodules in comparison to placebo (Leone Roberti Maggiore et al. 2014, Brown & Farquhar 2015). A multi-center RCT comparing GnRHa to OCPs as post-surgical therapy reports both groups increase quality of life scores (Granese et al. 2015). Although GnRHa is proven effective, a severe side effect is decreased bone mineral density (BMD); therefore, estrogens or progestins are given for bone protection (Leone Roberti Maggiore et al. 2014, Zito et al. 2014). In contrast, GnRH-ant inhibit gonadotropin receptors. Elagolix improves dysmenorrhea and dyspareunia compared to placebo in a phase 2 RCT (Ezzati & Carr 2015, Munoz-Hernando et al. 2015) and, comparing BMD profiles of elagolix with depot medroxyprogesterone acetate, both minimally impact BMD (Carr et al. 2014, Ezzati & Carr 2015).

Danazol, a synthetic androgen, inhibits the luteinizing hormone (LH) surge; however, it also increases free testosterone, causing undesired side effects including hirsutism, deepening of voice, weight gain, and acne. Danazol effectively decreases pelvic pain compared to placebo, and is as effective as other hormonal therapies, but the numerous side effects limit use (Practice Committee of the American Society for Reproductive 2014, Zito et al. 2014). AIs are currently a second line treatment in women refractory to first line treatments (Abu Hashim 2014). AIs such as letrozole decrease estrogen stimulation of endometriosis and, when used in combination with GnRHa, improve pelvic pain more than GnRHa alone. Additionally, letrozole with norethindrone acetate add-back has improved endometriosis symptoms, and high dose aromatase inhibition reduces ovarian endometrioma size (Agarwal et al. 2015). In contrast, a small RCT investigating post-surgical endometriosis pain comparing OCPs alone and in combination with letrozole reports similar pain scores, suggesting no benefit with letrozole addition (Almassinokiani et al. 2014).

Surgical Therapy

Surgical therapy for endometriosis is typically necessary for intractable pelvic pain despite medical therapy. Several different surgical techniques are performed (Table 2), including excision/removal of endometriosis, uterosacral nerve ablation, presacral neurectomy, and hysterectomy with bilateral salpingo-oophorectomy (BSO) (Daniels et al. 2010, Healey et al. 2014, Posadzka et al. 2015), and some techniques provide better symptomatic control than others. For symptom improvement and preventing disease recurrence, endometrioma removal is superior to drainage (Duffy et al. 2014, Practice Committee of the American Society for Reproductive 2014). Hysterectomy without BSO is less effective because of continued hormonal stimulation of microscopic endometriotic lesions. Hysterectomy with BSO leads to surgical menopause, which negatively impacts bone and cardiac health. Extreme surgical management is reserved for patients who fail conservative management (Duffy et al. 2014, Practice Committee of the American Society for Reproductive 2014).

Table 2. Endometriosis Surgical Treatments and Associated Efficacy.

Surgical Treatment	Surgical Technique	Compared Treatment	Efficacy
Laparoscopic Ablation	Ablate, or apply heat, to lesion	CO2 laser vs. electric cautery	Decreased pain with ablation (NRS 3) vs. CO2 laser (NRS) (Posadzka et al. 2015)
		Diagnostic Laparoscopy	Decreased overall pain OR 5.63 (Duffy et al. 2014)
Laparoscopic Excision	Remove lesion with scissor or laser	Ablation	No difference in overall pain, dyspareunia, or dyschezia at 1 year (Bulun 2009, Duffy et al. 2014)
			Excision decreased dyspareunia (VAS 3.2) vs. ablation (VAS 6.0) at 5 years (Healey et al. 2014)
			Ablation required more medical treatment (31%) vs. excision (20%) (Healey et al. 2014)
Conservative Laparoscopy	Ablate or excise lesions, restore anatomy, adhesiolysis	Diagnostic Laparoscopy	Decreased overall pain OR 6.58 (Duffy et al. 2014)
Laparoscopic Uterosacral Nerve Ablation	Ablate nerve fibers responsible for pain pathway	Conservative Laparoscopy	No difference in pain level (Daniels et al. 2010)
Endometrioma Removal	Separate cyst wall from ovary and excise cyst	Cyst drainage	Decreased recurrence of cyst (Dunselman et al. 2014)
			Decreased recurrence of dysmenorrhea (OR 0.15) (Brown & Farquhar 2015)
Presacral Neurectomy	Disrupts sympathetic innervation of uterus at level of superior hypogastric plexus	Conservative Laparoscopy	1 RCT: decrease midline dysmenorrheal (Practice Committee of the American Society for Reproductive 2014)(Dunselman et al. 2014)
			1 RCT: no additional benefit (Practice Committee of the American Society for Reproductive 2014)
Hysterectomy + BSO	Debulking to place in surgical menopause	Hysterectomy without BSO	Improved symptoms (Practice Committee of the American Society for Reproductive 2014)(Dunselman et al. 2014)(Duffy et al. 2014)

BSO = bilateral salpingo-oophorectomy, VAS = Visual Analog Scale, NRS = Numeric Rating Scale, OR = Odds Ratio

Alternative Therapy

Given that endometriosis is such a difficult disease to treat, alternative therapies are welcomed in addition to conventional therapy. Comparing Chinese medicine (CM) to GnRHa as post-surgical treatment for endometriosis found no differences in recurrence rates on follow-up (Weng et al. 2015). In contrast, Chinese herbal enemas decrease dysmenorrhea comparable to danazol (Kong et al. 2014), and CM and herbal enema combination is superior to danazol in decreasing pain symptoms (Flower et al. 2012). An acupuncture study in addition to conventional medical therapy significantly decreases pelvic pain by 5 to 6 points on the 10-point VAS (Rubi-Klein et al. 2010). Pelvic physical therapy includes internal manual treatment to stretch pelvic floor muscles, myofascial release, biofeedback, and trigger point release. In those with myofascial chronic pelvic pain, 63% report significant pain improvement after at least 6 sessions (Bedaiwy et al. 2013). Exercise can provide pain relief, based on questionnaire studies composed of 50-2730 women with endometriosis and 400-4000 control women; however, other survey studies correlate exercise with increased pelvic pain. Unfortunately, not all of these studies are controlled and all are from self-reporting (Bonocher et al. 2014). Large prospective cohort or case-control studies demonstrate increased risk of endometriosis with diets high in trans-fatty acids and decreased risk with diets containing high levels of long-chain omega 3 fatty acids (Hansen & Knudsen 2013). More high quality studies are needed in these areas and importantly, a positive publication selection bias likely exists with alternative therapies, exaggerating true effectiveness (Kong et al. 2014).

Association between Endometriosis and Cancer

The potential association between endometriosis and cancer has been theorized for decades. This association is based upon observational case-control and cohort studies that propose malignant transformation occurs within endometriotic lesions, giving rise to cancer. Our molecular-genetic understanding of both endometriosis and ovarian cancer continues to rapidly evolve; yet, a definitive mechanism for malignant transformation remains elusive.

Risk and Prognosis

The 10% prevalence of endometriosis and an even higher prevalence for women with infertility or chronic pelvic pain, makes the establishment of an absolute “cause-and-effect” relationship problematic. Lifetime risk of developing ovarian cancer in the general population is ∼1.4%, with a median age of onset in the early 60s (Schorge et al. 2010). Epithelial ovarian cancer is no longer seen as a single disease, but rather a constellation of multiple diseases based upon histologic subtypes and unique molecular signatures (Galic et al. 2013). The risk of ovarian cancer increases for women who incur fewer pregnancies and/or suffer from infertility. The possibility of confounding when assessing associative risk between these two entities must be considered because infertility is related to both conditions.

Nonetheless, a number of epidemiologic and clinical features lead investigators to propose an association between endometriosis and cancer. The establishment of an association was reported 90 years ago (Sampson 1925) and was refined in 1953, proposing that benign endometriosis should be observed in close anatomic proximity to the arising endometriosis-associated cancer (Scott 1953). Chief among the observations are that both entities produce tissues that can metastasize, invade, and destroy normal surrounding tissues. Furthermore, cancers often are identified in endometriotic lesions or in tissues that are contiguous with endometriosis, and there are often findings of candidate precursor lesions exhibiting histologic atypia in these surrounding tissues (Wei et al. 2011). Finally, endometriosis in younger women, which persists into older age, creates a long window for malignant transformation.

Several retrospective studies initially document the increased rate of endometriosis in women with ovarian cancer. A Swedish study containing over 20,000 patients that cross-matched inpatient endometriosis diagnosis and any cancer diagnosis (Brinton et al. 1997) found a small increased risk of any cancer, but the risks were not confirmed upon longer-term follow up (Brinton et al. 1997). The risk of ovarian cancer, however, is significantly increased in both the initial and long-term analyses. In patients with a history of prolonged endometriosis, the statistical risk for the development of ovarian cancer is even higher.

A linkage analysis of over 99,000 women from Denmark shows an endometriosis-related increase in ovarian cancer occurs in two histologic subtypes, clear cell and endometrioid (summarized in Table 3) (Brinton et al. 2005). Recent evidence also suggests a correlation between endometriosis and high-grade serous histologic type ovarian cancer (Lee et al. 2016). A large case-control study confirms an approximate 3-fold increased risk of clear cell or endometrioid ovarian cancer in association with endometriosis (Rossing et al. 2008). Malignant transformation risk to ovarian cancer from ovarian endometriosis is reportedly 0.2-2.5% (Gadducci et al. 2014). Recent studies also show the association between endometriosis and different forms of ovarian cancer: serous, mucinous, clear-cell, and endometrioid, with the predominant cell types being clear cell and endometrioid (Table 4).

Table 3. Summary of Risks Associated with Endometriosis and Cancer from Registry Studies by Brinton et. al. (Brinton et al. 1997, Brinton et al. 2005).

Population	Risk	SIR* or RR	95% CI
History of endometriosis admission (HEA)	Any cancer	1.2*	1.1-1.3*
HEA	Ovarian Cancer	1.9*	1.3-2.8*
HEA & Prolonged endometriosis	Ovarian cancer	4.2*	2.0-7.7*
HEA	Endometrial cancer	1.1*	0.6-1.9*
HEA Long term F/U	Any cancer
HEA Long term F/U	Ovarian Cancer	1.43*	1.19-1.71*
Long term F/U & Prolonged endometriosis	Ovarian Cancer	2.23*	1.36-3.44*
HEA Denmark cohort	Clear cell Ovarian Cancer	3.37	1.24-9.14
HEA Denmark cohort	Endometrioid Ovarian Cancer	2.53	1.19-5.38

HEA-History of endometriosis admission, SIR- Standardized incidence ratio, RR-Relative Risk, CI – Confidence interval, F/U- Follow up

Table 4. Ovarian Cancer Types Arising from Endometriosis Transformation.

Population (# Patients EAOC/Total in Study)	Ovarian Cancer Type in EAOC	Age (Mean ± SD) years	Reference
Quebec, BC (41/2854)	Serous 19.51% Mucinous NR Clear-Cell 21.9% Endometrioid 24.4%	OC 53.9 ± 11.4 EAOC 48.3 ± 10.8	(Aris 2010)
Belegrade, Serbia (23/210)	Serous 3.5% Mucinous NR Clear-Cell 36.8% Endometrioid 31.6%	NR	(Dzatic-Smiljkovic et al. 2011)
Michigan, USA (42/184)	Serous 55% Mucinous 10% Clear-Cell 21% Endometrioid 14%	OC 59 EAOC 52	(Kumar et al. 2011)
Athens, Greece (17)	Serous 5.9% Mucinous NRClear-Cell 58.8% Endometrioid 35.3%	EAOC 58 (27-76)	(Kondi-Pafiti et al. 2012)
Ovarian Cancer Association Consortium (738/7911)	Serous 7.1% Mucinous 6.0% Clear-Cell 20.2% Endometrioid 13.9%	OC 56.1 EAOC 56.3	(Pearce et al. 2012)
Ankara, Turkey (45/1086)	Serous 13.3% Mucinous 8.9% Clear-Cell 37.8% Endometrioid 33.3%	EAOC 55 (35-77)	(Boyraz et al. 2013)
Massachusetts, USA (67/134)	Serous 0% Mucinous NR Clear-Cell 38.8% Endometrioid 61.2%	OC 56.6 EAOC 51.7	(Davis et al. 2014)
Milano, Italy (27/73)	Serous NR Mucinous NRClear-Cell 76.1%Endometrioid NR	OC 58.4 ± 11.2 EAOC 51.4 ±10	(Scarfone et al. 2014)
San Juan, Puerto Rico (20/192)	Serous 2.2% Mucinous 2.7% Clear-Cell 23% Endometrioid 50%	OC 56.1 ± 14.9 EAOC 48.8 ± 11.6	(Acien et al. 2015)
Shiraz, Iran (28/110)	Serous 14.5% Mucinous 0% Clear-Cell 14.5% Endometrioid 39%	OC 50.18 ± 12.8 EAOC 49.93 ±9.36	(Akbarzadeh-Jahromi et al. 2015)

OC=Ovarian Cancer, EAOC=Endometriosis Associated Ovarian Cancer, NR=not reported

A meta-analysis conducted by Kim et al. (2014) evaluates the risk and prognosis of ovarian cancer in ∼445,000 women with or without endometriosis. Based on 35 studies, women with endometriosis are significantly at risk of developing ovarian cancer; however, stage is more likely to be early and low-grade, suggesting the cancer is slow growing and less invasive. Endometrioid and clear cell are common in women with endometriosis, with the serous subtype occurring less frequently and the mucinous subtype displaying no differences between control women and women with endometriosis (Kim et al. 2014). Endometriosis does not affect prognosis, and overall survival in women with endometriosis-associated ovarian cancer (EAOC) and in women with non-EAOC are similar when accounting for histology, disease status, assessment of endometriosis, and potential confounding factors. Unfortunately, the effect of endometriosis on a successful debulking surgery is not analyzed (Kim et al. 2014), so it is unknown if there is a benefit in survival in women with EAOC.

Proposed Mechanisms of Malignant Transformation

Complex hormonal, genetic, and immunologic interactions must be considered when assessing the interplay between endometriosis in the development of epithelial primary peritoneal or ovarian carcinomas. Chronic inflammation, autocrine and paracrine effects, hormonal interactions, and micro-environmental alterations caused by endometriosis in the pelvic region could be relevant mechanisms for malignant transformation. Aberrant immune function, stimulated by estrogens, may create a positive feed-forward loop, enhancing growth and invasiveness of endometriosis and promoting malignant transformation (Ness 2003). Zanetta et al. report a role for a hyper-estrogenic state in stimulating endometriosis and promoting malignant transformation (Zanetta et al. 2000).

A permissive microenvironment and accumulation of genetic mutations is suggested to cause malignant change of endometriosis (Wei et al. 2011). Distinct molecular events may occur in early stages of tumorigenesis of endometriosis-associated carcinoma. Recent studies focus on genetic alterations such as phosphatase and tensin homolog (PTEN), tumor protein p53 (TP53), and B-cell lymphoma (BCL) gene mutations that lead to malignant changes of endometriosis (Nezhat et al. 2008, Munksgaard & Blaakaer 2012, Lai et al. 2013, McConechy et al. 2014). An interplay of genetics and oxidative stress, with decreased expression of interleukin 1 receptor type 2 (IL1R2), is a common signature between endometrioid ovarian cancer and endometriosis (Kobayashi et al. 2009, Keita et al. 2010, Keita et al. 2011). IL1 ligands are expressed by all endometriosis-associated ovarian cancer subtypes and endometrial cells. A decrease in IL1R1 levels, a protector against the tumorigenic effects of IL1, occurs in endometrioid carcinoma (Keita et al. 2010, Keita et al. 2011).

Multiple tumor-associated somatic mutations, detected by examining single gene or by whole genome sequencing, have revealed a signature of mutations. Mutations in catenin beta 1 (CTNNB1) are seen in 60% of ovarian endometrioid carcinomas (Matsumoto et al. 2015). Mutations in AT Rich Interactive Domain 1A (ARID1A) and phosphatidylinositol-4,5-bisphosphate 3-kinase, catalytic subunit alpha (PIK3CA) appear most consistently in clear cell ovarian carcinomas (Gadducci et al. 2014, Anglesio et al. 2015, Matsumoto et al. 2015). Mutations in ARID1A, involved in chromatin remodeling, are present in both clear cell (15-75%) and endometrioid carcinomas (30-55%) (Wiegand et al. 2010, Gadducci et al. 2014). Associated with malignant transformation, mutations in ARID1A lead to the loss of its product, BAF250a, which correlates strongly with ovarian clear-cell carcinoma and endometrioid carcinoma subtypes, as well as with high-grade endometrial carcinomas (Wiegand et al. 2010, Wiegand et al. 2011, Ayhan et al. 2012, Lowery et al. 2012, Samartzis et al. 2012, Chene et al. 2015). ARID1A mutations and BAF250a loss are also observed in tumors and contiguous atypical endometriosis, but not in distant endometriotic lesions. The loss of ARID1A expression usually coexists with PI3K-Akt pathway activation and/or zinc finger protein 217 (ZNF217) amplification in ovarian clear cell cancers and may indicate an early event in the malignant transformation of endometriosis into the various histotypes of ovarian cancer (Ayhan et al. 2012, Huang et al. 2014).

Loss of PTEN is observed in clear cell-associated endometriosis and cancers, including a significant increase in expression levels of X-ray repair cross-complementing protein 5 (XRCC5), patched 2 (PTCH2), elongation factor 1-alpha 2 (EEF1A2) and protein phosphatase 1 regulatory subunit 14B (PPP1R14B). However, these changes are not observed in benign endometriosis (Worley et al. 2015). PTEN loss is proposed as an early and permissive event in endometriosis development, while loss of ESR1 and polycomb-mediated transcriptional factor cause ultimate malignant transformation (Worley et al. 2015).

Future research will clarify the likely complex interaction between genetic alterations, estrogen exposure, inflammatory cytokines, and the immunologic microenvironment in the transformation of endometriosis to endometrioid and clear cell ovarian and primary peritoneal cancers. Treatment of these cancers will hopefully improve with use of targeted and immunologic therapies that address underlying causes of malignant transformation.

Concluding Remarks: Where are we going?

While the studies reviewed from the last 5 years demonstrate a deeper understanding of endometriosis as dysregulations pertain to hormones, hormone receptors, immune function, and transformation to ovarian cancer, endometriosis still remains mysterious from many facets. Critically needed for this enigmatic disease are mechanistic understandings of disease initiation and perturbation that will hopefully lead to the development of non-invasive disease diagnosis and the development of treatments that do not negate hormonal cyclicity or have other undesired side effect profiles and decrease the need for surgical extirpation. To allow for this to happen, the following areas of need are identified:

• Establish clear limits to animal models and clarify what the model may and may not reveal

• Establish international standards for collection of patient information and samples as outlined by (Fassbender et al. 2013)

• Establish disease profile through clearer understanding of cytokines and the potential association with autoimmune disorders

• Characterization of interplay between the hormonal milieu and immune system

• Focus on lifetime exposures, acute and chronic, to endocrine disrupting chemicals that may interfere with uterine development, immune system regulation, and ultimately endometriosis development

• Full recognition that this disease is truly multifaceted with pain, psychology, infertility, immunity, etc.

• Transformation of endometriosis to ovarian cancer through characterization of the lag between endometriosis found on the ovary to an ovarian cancer diagnosis

• Determine if age, parity, weight, hormonal regulators (oral contraceptives) contribute to transformation to cancer diagnosis