THE BIOLOGICAL FUNCTION OF HUMAN EPIDIDYMIS PROTEIN 4 IN EPITHELIAL OVARIAN CANCER

Epithelial ovarian cancer (EOC) is the most common gynecologic malignancy worldwide. EOC has a notably poor prognosis, owing to the fact that patients are frequently diagnosed at a late stage after the disease has significantly progressed. While many patients typically respond well to frontline platinum-based chemotherapy, the tumor becomes chemoresistant when a recurrence follows within five years. Therefore, there is an urgent need for the discovery of non-invasive early detection biomarkers and novel targeted therapies. Human Epididymis Protein 4 (HE4) is a secretory protein that is encoded by the gene whey acidic protein (WAP)-four disulfide core domain protein 2. The WAP domain family is a conserved motif that is inherit of many antiproteases. HE4 was initially found to be a component of the innate immune defenses of multiple epithelia and to function in epithelial host defense, through the promotion of mucosal surfaces first line of defense. HE4 is highly overexpressed in EOC and has been identified as a novel clinical biomarker. Clinical and translational studies have established HE4 as a contributor to tumorigenesis and chemoresistance in EOC. However, the exact processes in which HE4 promotes pathogenesis is unclear. The driving hypothesis of this thesis is that HE4 represents a novel targeted therapy due to its established role EOC tumorigenesis and suggested function in innate immunity. This evidence underlies the goals of this dissertation which are to elucidate the precise mechanisms of HE4’s contribution in EOC pathogenesis and establish HE4’s role in tumor immune invasion. It is hoped that results from this investigation will ultimately aide in the development of a novel targeted therapy against HE4 that can modulate tumor pathogenesis as well as the tumor immune response. In manuscript I, subtractive hybridization revealed that HE4 significantly suppresses expression of osteopontin (OPN) in peripheral blood mononuclear cells (PBMCs) which ultimately compromised their cytotoxicity against ovarian cancer cells. Ovarian cancer cells exhibited enhanced proliferation in conditioned media from HE4-exposed PBMCs and this effect was attenuated by the addition of recombinant OPN and OPN inducible cytokines (IL-12 and IFN-y). In addition, ovarian cancer cells and PBMCs with HE4 downregulation via short hairpin RNA (shRNA) were found to be increasingly more susceptible to cell death. In manuscript II, subtractive hybridization identified dual specificity phosphatase 6 (DUSP6) as the most upregulated gene upon treatment with recombinant HE4 in PBMCs. Flow cytometry revealed that recombinant HE4 significantly upregulated DUSP6 levels specifically in CD8+ (cytotoxic T cell) and CD56+ (NK cell) populations. Exposure of these cells to HE4 led to an increase in ERK 1⁄2 phosphorylation, which was subsequently decreased upon DUSP6 inhibition. These results show that DUSP6 suppression of CD8+ and CD56+ lymphocyte toxicity is strongly enhanced by HE4. In co-culture of PBMCs and ovarian cancer cells, DUSP6 inhibition attenuated the enhanced proliferation noted upon stimulation with HE4. The effect of DUSP6 inhibition was obliterated in CD8+ and CD56+ devoid PBMCs. In manuscript III, the role of DUSP6 and its relationship to HE4 in EOC was further elucidated. Increased DUSP6 levels were observed in ovarian cancer cells overexpressing HE4. siRNA-mediated downregulation of both HE4 and DUSP6 revealed a corresponding decrease of either factor. Treatment with an allosteric DUSP6 inhibitor in combination with chemotherapeutic agents produced synergistic effects on the reduction of cell viability. These effects correlated with alterations in expression of ERK pathway mediated genes. Finally, it was found that DUSP6 was significantly overexpressed in serous EOC patient tissue compared to normal adjacent tissue. In manuscript IV it was determined from a small-scale proteomics study that 63 proteins were found to interact more strongly with HE4, in HE4 overexpressing clones compared to null vector control. The protein found to exhibit the highest interaction in the HE4 clones was Septin-2, a GTP binding protein. Immunohistochemical analysis of Septin-2 in EOC patient tissue revealed that levels were overexpressed in cancer compared to normal and benign controls. To identify Septin-2’s role in EOC, stable knockdown cell lines were constructed using the ovarian cancer cell line SKOV3. Septin-2 knockdown cells demonstrated a significantly lowered proliferation rate compared wild-type (WT) and Plasmid C control cells. To better define the role of Septin-2 in EOC, proteomics was employed. Pathway analysis showed an enrichment in autophosphorylation, citric acid cycle, acetyl CoA/energy, and proteasomal/ubiquitin processes in Septin-2 knockout cells.

the tumor becomes chemoresistant when a recurrence follows within five years.
Therefore, there is an urgent need for the discovery of non-invasive early detection biomarkers and novel targeted therapies.
Human Epididymis Protein 4 (HE4) is a secretory protein that is encoded by the gene whey acidic protein (WAP)-four disulfide core domain protein 2. The WAP domain family is a conserved motif that is inherit of many antiproteases. HE4 was initially found to be a component of the innate immune defenses of multiple epithelia and to function in epithelial host defense, through the promotion of mucosal surfaces first line of defense. HE4 is highly overexpressed in EOC and has been identified as a novel clinical biomarker. Clinical and translational studies have established HE4 as a contributor to tumorigenesis and chemoresistance in EOC. However, the exact processes in which HE4 promotes pathogenesis is unclear. The driving hypothesis of this thesis is that HE4 represents a novel targeted therapy due to its established role EOC tumorigenesis and suggested function in innate immunity. This evidence underlies the goals of this dissertation which are to elucidate the precise mechanisms of HE4's contribution in EOC pathogenesis and establish HE4's role in tumor immune invasion. It is hoped that results from this investigation will ultimately aide in the development of a novel targeted therapy against HE4 that can modulate tumor pathogenesis as well as the tumor immune response.
In manuscript I, subtractive hybridization revealed that HE4 significantly suppresses expression of osteopontin (OPN) in peripheral blood mononuclear cells (PBMCs) which ultimately compromised their cytotoxicity against ovarian cancer cells. Ovarian cancer cells exhibited enhanced proliferation in conditioned media from HE4-exposed PBMCs and this effect was attenuated by the addition of recombinant OPN and OPNinducible cytokines (IL-12 and IFN-y). In addition, ovarian cancer cells and PBMCs with HE4 downregulation via short hairpin RNA (shRNA) were found to be increasingly more susceptible to cell death.
In manuscript II, subtractive hybridization identified dual specificity phosphatase 6 (DUSP6) as the most upregulated gene upon treatment with recombinant HE4 in PBMCs. Flow cytometry revealed that recombinant HE4 significantly upregulated DUSP6 levels specifically in CD8+ (cytotoxic T cell) and CD56+ (NK cell) populations. Exposure of these cells to HE4 led to an increase in ERK ½ phosphorylation, which was subsequently decreased upon DUSP6 inhibition. These results show that DUSP6 suppression of CD8+ and CD56+ lymphocyte toxicity is strongly enhanced by HE4. In co-culture of PBMCs and ovarian cancer cells, DUSP6 inhibition attenuated the enhanced proliferation noted upon stimulation with HE4. The effect of DUSP6 inhibition was obliterated in CD8+ and CD56+ devoid PBMCs.
In manuscript III, the role of DUSP6 and its relationship to HE4 in EOC was further elucidated. Increased DUSP6 levels were observed in ovarian cancer cells overexpressing HE4. siRNA-mediated downregulation of both HE4 and DUSP6 revealed a corresponding decrease of either factor. Treatment with an allosteric DUSP6 inhibitor in combination with chemotherapeutic agents produced synergistic effects on the reduction of cell viability. These effects correlated with alterations in expression of ERK pathway mediated genes. Finally, it was found that DUSP6 was significantly overexpressed in serous EOC patient tissue compared to normal adjacent tissue.
In manuscript IV it was determined from a small-scale proteomics study that 63 proteins were found to interact more strongly with HE4, in HE4 overexpressing clones compared to null vector control. The protein found to exhibit the highest interaction in the HE4 clones was Septin-2, a GTP binding protein. Immunohistochemical analysis of Septin-2 in EOC patient tissue revealed that levels were overexpressed in cancer compared to normal and benign controls. To identify Septin-2's role in EOC, stable knockdown cell lines were constructed using the ovarian cancer cell line SKOV3.
Septin-2 knockdown cells demonstrated a significantly lowered proliferation rate compared wild-type (WT) and Plasmid C control cells. To better define the role of Septin-2 in EOC, proteomics was employed. Pathway analysis showed an enrichment in autophosphorylation, citric acid cycle, acetyl CoA/energy, and proteasomal/ubiquitin processes in Septin-2 knockout cells.  Worldwide ovarian cancer has an incidence of 240,000 cases per year and an annual mortality rate of 152,000 [1]. This high mortality rate is largely due to that fact that in many cases ovarian cancer is detected at an advanced disease state. In addition, while the initial response rate to frontline chemotherapy is 60-80%, when the tumor recurs it eventually becomes unresponsive to traditional platinum-based chemotherapeutics [2].

LIST OF TABLES
Unfortunately, only a minority of patients with advanced stage disease achieve long term survival, as many patients will develop a recurrence within 12-18 months of completion of their primary treatment regimen [3]. Currently, the five year survival rate for ovarian cancer is only 35% [4] , and these dire statistics have not improved significantly in the last 30 years [5].

Ovarian Cancer Subtypes
Ovarian Cancer is divided into two major subtypes that depend on the tissue of origin.
Non-epithelial ovarian cancer includes sex cord stromal, germ cell and non-specified ovarian cancers. Non-epithelial ovarian cancers only represent 10% of all ovarian cancer, [6]while the remaining 90% of cancer comprises epithelial ovarian cancer (EOC). EOC encompasses serous, transitional cell, mucinous, endometrioid, and clear cell ovarian cancer [7]. EOC is generally divided into two subtypes. Type 1 EOC are considered more genetically stable, exhibit a slower tumor growth, and have disease contained within the ovary upon initial presentation. These cancers respond well to surgical intervention [7]. In contrast, type 2 EOC are characterized by an aggressive growth rate and are usually detected at an advanced stage of IIIC. High grade serous ovarian cancer (HGSOC) is the most common histological subtype of Type II EOC, representing nearly three quarters of all patients diagnosed with ovarian cancer [5].
Seventy percent of the time, HGSOC is diagnosed at an advanced stage, leading to a poor prognosis [5]. Therefore, efforts have been made to develop novel prognostic/diagnostic methods and treatments to combat chemoresistance and improve overall survival for HGSOC.

Current Ovarian Cancer Therapies
Many women who present with elevated tumor markers and abnormal imaging typically proceed with primary debulking surgery. Initial surgery has three goals: diagnosis, staging and cytoreduction. Diagnosis is important as needle biopsies are not indicated for larger ovarian masses to prevent inadvertent spreading of the disease [8]. If a patient presents with significant comorbidities, clinicians will favor neoadjuvant chemotherapy over surgery. This approach minimizes surgical side effects for patients, as the tumor will be reduced following chemotherapy [8].
For the past 20 years, standard of care for women diagnosed with EOC is a primary frontline regimen of carboplatin and paclitaxel [9]. Carboplatin binds to DNA forming a platinum adduct and causes cell death [10]. Paclitaxel's mechanism of action involves enhancing polymerization of tubulin, which stabilizes microtubules.
This stabilization results in the protection of the microtubule polymer from disassembly, and chromosomes are unable to achieve proper metaphase spindle organization. Ultimately, cells are halted in the G2/M phase of the cell cycle [11]. The overall response rate (ORR) for this combinational first line therapy is greater than 75%. However, the majority of patients experience a recurrence and progression of disease. Once a recurrence occurs post-frontline therapy, the chemotherapy chosen for the patient is based on the platinum-free interval (PFI), which represents the time between the completion of the last platinum-based treatment and the detection of relapse [12]. Patients that have a PFI of six months or less are considered to be platinum-resistant, while patients that have a PFI greater than six months mark are considered platinum-sensitive. This distinction determines the second-line chemotherapy regimen used for the patient. [13]For platinum-sensitive patients experiencing recurrence, doxil or gemcitabine is added to a platinum regimen [12].
Doxorubicin is an antitumor antibiotic that promotes cell death by intercalation into DNA, disrupting DNA repair mediated by topisomerase II, and generating free radicals [15], Gemcitabeine is a pyrimidine antimetabolite that is inhibits tumor cell progression through the G1/S phase, halting DNA synthesis [14]. While platinumsensitive patients undoubtedly survive longer than patients who are initially platinum refractory, prognosis for these patients is still dismal. Platinum combinatorial therapies with doxil and gemcitabine exhibit a progression-free survival (PFS) of only 11.3 and 8.6, respectively [16].
For platinum-resistant ovarian cancer, a non-platinum monotherapy is used with a non-curative goal of toxicity management, as prognosis in this group is poor. Patients in this group are frequently enrolled in clinical trials as a last attempt to control disease [17]. Topotecan, which works through inhibition of topoisomerase I, is a typical example of a salvage chemotherapy that is used in platinum resistant ovarian cancer [18]. The response rate of patients to this treatment is only 12-18%, and PFS is around 3-4 months [19,20]. Other typical monotherapies for platinum-resistant second line EOC include doxil and bevacizumab [8]. Bevacizumab is a monoclonal antibody against vascular endothelial growth factor (VEGF), a major regulator of angiogenesis. Bevacizumab is approved in the recurrent setting, however it's overall efficacy continues to be studied clinically in different chemotherapy lines and in combination with various treatment regimens [21]. While many large phase III trials report an increase in PFS for patients, this response does not correlate with an increased overall survival [21]. Other approved therapies in the maintenance setting are PARP inhibitors. While these inhibitors are approved for all patients, within this setting it has shown the most substantial benefit for patients who harbor the BRCA mutation-about 20-25% of the patient population [22,23]. Current clinical trials for EOC have largely focused on the immune checkpoint inhibition of programmed death receptor (PD-1) and it's ligand PD-L1, however clinical trial results have suggested only a modest benefit [24]. Therefore, there is still a crucial treatment need for the non-BRCA patient population.

Detection Methods of EOC
Early detection for EOC is difficult as many symptoms reported by patients, such as bloating and pelvic pain, are common symptoms of benign disease [25]. In addition, the sensitivity and specificity of pelvic examinations for EOC screening purposes within an asymptomatic population are poor. Therefore, diagnosis relies heavily on tumor markers and radiologic imaging [25]. Currently, there has not been an official recommendation for routine screening of asymptomatic women who are not high risk for development of an ovarian malignancy [26].
Cancer antigen 125 (CA 125) is the most commonly used and validated tumor marker for the detection of EOC [27]. However, recently there has been sufficient research dedicated to an improvement of serum biomarkers for early detection of EOC. One biomarker that represents such improvement is Human Epididymis Protein 4 (HE4), which has been shown to have a higher specificity and comparable sensitivity to CA 125 [28]. From these results, the risk of ovarian malignancy algorithm (ROMA) was established, which takes into account a woman's menopausal status and incorporates preoperative serum levels of CA 125 and HE4. The ROMA score exhibits both a higher sensitivity and specificity than CA 125 alone [29]. As HE4 has been extensively studied clinically, its prognostic capabilities have also begun to be examined translationally.

Molecular Functions of HE4
HE4 is encoded by the Whey Acidic Protein (WAP) 4-disulphide core domain (WFDC2) gene. The WFDC2 transcript was thought to be exclusively expressed in the epididymis and hence was originally proposed to be a specific marker for this tissue type [30]. WFDC2 is a member of the WAP domain, which is a conserved motif of 50 amino acids, including eight cysteine residues arranged as a 4-disulphide core [31]. While WAP proteins can display a variety of functions, the most comprehensively studied members of this family are the antiproteinases secretory leukocyte protease inhibitor (SLPI) and elafin. In addition to antiproteinase activities, both exhibit anti-inflammatory activities [32,33]. Due to the familial similarity of HE4, it has been proposed to function similarly to SLPI and elafin; however, this role has not been fully defined. In addition to HE4 overexpression in EOC tissue compared to normal and benign ovarian tissues [29,34], it is also readily expressed in the oral cavity, nasopharynx and respiratory tract [35]. It was suggested that HE4 functions in concert with other WAP domain family members to promote epithelial host defenses of the lung, nasal, and oral cavity; supporting the claim that HE4 plays a role in innate immune defenses [35]. HE4's known molecular functions in EOC pathogenesis, particularly its role in promotion of cell proliferation, chemoresistance, metastasis and steroid biosynthesis, are comprehensively discussed in Chapter 2.

Problem Statement
Challenges in both treatment and diagnosis of patients has led to strong efforts to elucidate new mechanisms of ovarian cancer pathology that can be used to develop novel targeted therapies, which are so desperately needed for this patient population.
HE4 is a secretory protein that is overexpressed in EOC serum and tissue. Extensive studies have also shown that HE4 promotes EOC growth and chemoresistance.
However, the exact mechanisms of HE4 functions in EOC pathogenesis are not completely understood. In addition, while HE4 was initially found to play a role in innate immunity, its function in tumor immunity has yet to be defined. Therefore, further investigation of HE4's mechanistic promotion of tumorigenesis is an important step to determine potential efficacy of a targeted anti-HE4 therapy for the treatment of EOC.
The aim of this thesis research is to: 1. Determine genes most suppressed by HE4 in immune cell populations and determine their involvement in muting the cytotoxic ability of immune cells toward ovarian cancer cells.
2. Determine genes most induced by HE4 in immune cell populations and determine their involvement in muting the cytotoxic ability of immune cells toward ovarian cancer cells.
3. Establish the significance of HE4 regulated genes in EOC pathogenesis.
4. Define novel roles of proteins with an identified association with HE4 in EOC pathogenesis.

Hypothesis
The overall driving hypothesis of this investigation is that HE4 represents a novel therapeutic target due to its role in the promotion of EOC pathogenesis. While it is known that HE4 has a profound role in EOC diagnosis, its therapeutics capabilities have been largely undefined due to an incomplete identification of its signaling network in EOC. Although the precise mechanism is unknown, it has been established that HE4 promotes tumorigenesis, chemoresistance, and metastasis in EOC. It has been previously proposed that HE4 plays a role in innate immunity; however, its immune functions in EOC have not been explored. However, many patients experience a chemoresistant recurrence within the first 2 years following treatment (2). Therefore, there is an urgent need for tools to aid in the early diagnosis of ovarian cancer when the disease is fundamentally curable, as well as improved treatment options for later stage disease.
Human epididymis protein 4 (HE4) is a secretory protein that is member of the whey acidic protein domain family, bearing a conserved motif found in a number a protease inhibitors (3). HE4 was initially suggested to be involved in the innate immune defense of multiple epithelia and has also been found to function in epithelial host defense (4). In ovarian tissue, HE4 is highly overexpressed in EOC compared normal tissue (5,6). Clinically, HE4 has been identified as a novel therapeutic biomarker for EOC and has also proven useful in detection of recurrent disease (7)  Recently, it has been reported that HE4 can be detected in EOC patient urine, indicating the possibility that it may be utilized as a non-invasive biomarker (8).
While HE4 has been well studied in the clinical setting, less is known regarding its specific molecular and biological roles in EOC. Several studies have investigated its effect on gene expression in EOC cells, as well as on events associated with aggressive disease. This review will summarize HE4's effect on cell proliferation and tumor growth; invasion, migration, and adhesion; chemoresistance; and steroid biosynthesis ( Figure 1). Each section will detail associated pathways and factors that are reported to be involved in these HE4-mediated effects, with the goal of revealing Several in vitro studies suggest that HE4 promotes proliferation through its involvement in cell cycle regulation (11). Silencing of HE4 causes G0/G1 cell cycle arrest and blocks the transition from the G1 to the S phase of the cell cycle.
Conversely, when cells are stimulated with recombinant HE4, the number of cells in the G2/M phase is increased, while the number of cells in the G0/G1 phase is reduced (9). These results indicate that HE4 may mediate the cell cycle by promoting the G0/G1 transition. In addition, in vivo tumorigenicity studies using HE4 knockdown clones revealed a marked inhibition in the growth of ovarian tumors in nude mice (14), while injection of HE4-overexpressing cells led to more aggressive tumor growth and an overall higher tumor volume compared with controls (10,15). Taken together, results from numerous in vitro and in vivo studies provide compelling evidence that HE4 plays a role in cell proliferation and the promotion of tumorigenesis. A full list of factors associated with HE4-mediated cell proliferation and tumor growth can be found in Table 1A and is outlined in greater detail below.

Associated Pathways and Factors-Cell Proliferation and Tumor Growth
Human epididymis protein 4 has been connected to several oncogenic signaling cascades that play key roles in ovarian cancer progression, including the PI3K/AKT pathway, HIF1α, and ERK/mitogen-activated protein kinase (MAPK) signaling.
Evidence of HE4's effect on activation of each of these pathways is discussed below.
Protein Kinase B Signaling AKT has been established as a strong promoter of tumorigenesis, and the PI3K/AKT pathway is one of the most commonly hyperactivated pathways in many types of human cancers (16). Its diverse signaling regulates proliferation, growth, survival, motility, angiogenesis, and glucose metabolism (17). HE4-overexpressing OVCAR3 ovarian cancer cells were found to have a marked increase in activation of protein kinase B (AKT) compared with control cells, while HE4 knockdown in OVCAR3 cells reduced AKT activation (12). Moreover, it was found that HE4-overexpressing SKOV3 clones had naturally higher gene levels of AKT3 compared with the nullvector control (18), bolstering the claim that HE4 affects the PI3K/AKT pathway.
Hypoxia-Inducible Factor-1 Alpha (HIF1α) Adaptation of malignant cells to hypoxic conditions is a key step in the promotion of tumorigenesis and angiogenesis (19)(20)(21), a process that is regulated by the transcription factor HIF1α. Co-immunoprecipitation revealed an interaction between HIF1α and HE4 in HE4-overexpressing SKOV3 xenografts. There was also strong colocalization of HE4 and HIF1α in SKOV3 ovarian xenograft tissue. In addition, when SKOV3 cells were treated with HIF1α siRNA or 2-methoxyestradiol (a HIF1α inhibitor), there was a marked decrease in HE4 protein levels (15). It is important to note that 2-methoxyestradiol is not a specific HIF1α inhibitor as it primarily causes the depolymerization of microtubules, which in turn prevents HIF1α expression (22).
Thus, the specificity of the effect of HIF1α inhibition on HE4 levels may require further investigation. Although the exact mechanism and significance of the HE4-HIF1α interaction is not understood, this evidence suggests that HE4 could play a role in regulating HIF1α functions in angiogenesis.

MAPK Signaling
The MAPK pathway is composed of a family of conserved kinases that mediate essential cellular processes such as migration, growth, proliferation, differentiation, and apoptosis (23 Interestingly, silencing of HE4 in ovarian cancer cells led to a decrease in protein levels of MMP-9, MMP-2, and Cathepsin B, suggesting these factors may be involved in HE4-mediated tumor promoting effects (11).

Interleukin-1 alpha (ILIA)
Interleukin-1 alpha is a pro-inflammatory cytokine that is involved in angiogenesis and metastasis. ILIA can directly stimulate the synthesis of VEGF (69) and fibroblastic pro matrix metallic proteinase I (70, 71). IL1A causes resistance to EGFR inhibitors in both colon and head and neck cancers (72, 73). IL1A was also found to be differentially expressed in three separate microarray studies involving HE4. In two microarrays, IL1A levels positively associated with HE4 levels (10,74), while in one study their levels were inversely associated (18). While there may be some ambiguity as to how HE4 and IL1A are mechanistically linked, the consistent connection between IL1A with HE4 merits further investigation.

Extracellular Matrix Proteins
Integrins are a family of transmembrane proteins that are vital to ECM adhesion and play important roles in wound healing as well as the pathogenesis of cancer (75-77).
Integrin β5 (ITGβ5) gene expression was differentially regulated by HE4 in ES-2 and CaOV3 cells, which was confirmed by positive correlation of ITGB5 and HE4 staining in paraffin embedded ovarian tissue samples (10). This finding suggests that integrin signaling is one mechanism by which HE4 can promote increased adhesion of ovarian cancer cells. However, further research is needed to clarify the mechanisms involved.

Microtubule Stabilization
Microtubule-associated protein tau, which has been associated with paclitaxel resistance in ovarian (116), breast (117), and gastric cancer (118), was upregulated in SKOV3 cells overexpressing HE4 compared with null-vector cells (18). In addition, HE4-overexpressing cells were found to express significantly higher levels of SEPT3 (Septin 3) mRNA compared with null-vector controls (18). Septins are a family of conserved GTP binding proteins that are associated with microtubules and actin filaments and have an important role in cytoskeletal organization (119). Furthermore, recombinant HE4 treatment of SKOV3 cells increased β-tubulin levels, indicating that HE4 might promote microtubule stability, leading to paclitaxel resistance.

Kinase Signaling Pathways
Human epididymis protein 4 knockdown has also been shown to lead to a reduction in cell growth and the resensitization of ovarian cancer cells to both cisplatin and paclitaxel (12). Lee et al. found that this effect was due to corresponding decreases of ERK and AKT in HE4 knockouts. Activation of these pathways suppresses apoptotic signaling in tumors, suggesting that HE4's regulation of these pathways may be an important mechanism of chemoresistance (120).

Steroid Biosynthesis
Evidence suggests an association between sex steroids and EOC pathogenesis, which is explained by processes that take place during the menstrual cycle. The ovarian surface epithelium (OSE) plays a critical role in ovulation and postovulatory wound repair. During the menstrual cycle, the OSE proliferates during the pro-estrus/estrus transition. After, ovulation the proliferation rate decreases (121). It is hypothesized that when the OSE is repeatedly exposed to high doses of luteinizing hormone and follicle stimulating hormone during the menstrual cycle, this can promote cell proliferation and increase the likelihood of tumor growth over time (121).  Table 1D and is outlined in detail below.

Steroid Biosynthesis Gene Expression
Two separate microarray pathway analyses identified steroid biosynthesis as a pathway affected by HE4 (10,74). Important genes that were differentially expressed between HE4-overexpressing and HE4 knockdown cell lines were Finally, NSDHL is also involved in cholesterol biosynthesis and produces metabolites that are essential in the conversion of squalene to cholesterol (130 Although HE4 is well established as a clinical biomarker for ovarian cancer, it has been largely understudied for its therapeutic targeting potential. However, ongoing research continues to support that HE4 is profoundly involved in the pathogenesis of EOC. The individual studies mentioned in this review provide evidence that HE4 promotes EOC progression through pathways associated with cell proliferation, tumor growth, metastasis, chemoresistance, and steroid biosynthesis. These pathways, along with specific genes that have been shown to be associated with HE4, are summarized in Table 1. This compilation of HE4 regulated factors and pathways will serve as a starting point for scientists to further elucidate specific mechanisms by which HE4 ultimately drives tumorigenesis. In addition, a comprehensive summary of clinical, in vivo, and in vitro studies related to each facet of EOC progression and HE4 can be seen in Figure 1. This diagram highlights the progress that has been made to establish HE4 as an attractive therapeutic target, while simultaneously denoting areas of research that are still lacking. The results discussed here suggest that inhibition of HE4 via a neutralizing antibody or small molecule inhibitor could provide viable treatment options for patients in dire need of more effective therapies.      In these biopsy specimens, the number of OPN + T cells correlates positively with progression free survival (PFS) and inversely with serum HE4 level. Taken together, these findings show that HE4 enhances ovarian cancer tumorigenesis by compromising OPN-mediated T cell activation.

I.2 Introduction
Human epididymis protein 4 (HE4) is a member of the whey acidic domain family of proteins (WAP), which are generally regarded as protease inhibitors (1-3). HE4 was first identified in the male reproductive tract but has since been found in select other tissues, such as the kidney, female reproductive tract, breast, and lungs (4,5). In addition, it is highly overexpressed in several human malignancies, including ovarian and endometrial cancer (5-8). HE4's role in normal and malignant tissue is still unclear; however, as a known negative prognostic factor in women with epithelial ovarian cancer, its serum levels correlate with chemoresistance and reduced survival (9)(10)(11). Our previous work with HE4 has led to the development of a USFDA approved biomarker tool for evaluation of pelvic masses, coined the Risk of Ovarian Malignancy Algorithm (ROMA) (12)(13)(14)(15). The ROMA score incorporates HE4, CA-125, and menopausal status into a calculation to estimate ovarian cancer risk. As a biomarker, HE4 detection and monitoring is already improving patient care. However, it is imperative that we learn more about its function in order to better understand ovarian tumorigenesis and ultimately develop effective therapies for this fatal cancer.
In this present study, we begin to elucidate HE4's role in the interplay between tumor cells and the immune system. We generated cDNA-subtracted libraries of HE4 treated Together, our data demonstrates that HE4 inhibits the immune function of PBMCs, most prominently T cells, via suppression of OPN production.

HE4-mediated IL-12 and IFN- reduction in PBMCs is reversible with supplementation of OPN
In lipopolysaccharide-stimulated macrophages, OPN has been shown to enhance IL-12 production and suppress IL-10 production, thereby promoting Th1 activity (17,18).
In   Figure 3A). Next, a cell migration assay was employed to determine whether conditioned media from rHE4-exposed PBMCs affects ovarian cancer migration as a surrogate of metastatic capability. The SKOV3 cells that were incubated with HE4-exposed PBMC media showed more extensive migration than control cells (RFU of 1147.21 ± 365.09 vs. 3138.14 ± 419.66, p < 0.01, Figure 3B). Immunohistochemistry using anti-Ki67 was performed to evaluate the proliferation of SKOV3 cells in the presence of rHE4-exposed PBMC media or vehicle-exposed conditioned media. The proliferation rate of tumor cells in HE4exposed PBMC conditioned media was higher than control media (63.8 ± 18.1 vs 39.9 ± 7.6 %, p < 0.01, Figure 3C). These findings suggest that PBMCs alter their soluble factor release under the influence of rHE4, thus enhancing the viability, proliferation and migration capabilities of the cultured ovarian cancer cells.

HE4 inhibition increases ovarian cancer susceptibility to PBMC-mediated cytotoxicity
In order to evaluate the impact of native (tumor-cell produced) HE4 on PBMCs, SKOV3 cells were co-cultured with PBMCs after stable transfection with HE4 specific shRNA (shHE4) or a scrambled oligonucleotide control plasmid (SO). Clones of shRNA transfected cells were tested for their phenotype by western blotting and ELISA ( Figure S5). After a 2-hour incubation at 37 °C, the effector cells were washed away and the target cells were analyzed by flow cytometry. The silencing of HE4 in SKOV3/PBMC co-cultures led to a significant increase in IL-12 and IFN concentrations ( Table 2). As shown in Figure 4, HE4 silencing also increased tumor cell susceptibility to PBMC cytotoxicity, an effect that was reversed by the addition of rHE4. Furthermore, this "rescue" by rHE4 was at least partially abrogated by the addition of recombinant IL12 (rIL-12) or recombinant IFN- (rIFN-) to the culture conditions. These findings suggest that the native HE4 production by ovarian cancer cells is critical to cell-mediated cytotoxicity resistance.
Ovarian cancer patient prognosis correlates to the number of intra-and peri-tumoral

CD3 + T cells and stromal OPN-producing cells
Twenty biopsies from high-grade serous ovarian cancer patients were evaluated by dual fluorescent stain with antibodies against CD3 and OPN (Table 3 Figure 5D). These findings suggest that tissue infiltrating T cells play a critical role in the suppression of ovarian cancer progression.

I.5 Discussion
HE4 is known to be highly overexpressed in ovarian cancer, but its causal relationship to ovarian tumorigenesis has not been firmly established. Emerging studies suggest that HE4 overexpression promotes ovarian tumor growth and imparts strong resistance against the most commonly used chemotherapeutics (20-24). Accordingly, serum HE4 level is an early predictor of platinum resistance (9,23), and ovarian cancer patients that experienced greater HE4 reduction during neoadjuvant chemotherapy exhibited improved overall survival (24). Our study has shown a novel role for HE4 in the In summary, this study is the first to implicate HE4 in ovarian cancer immune escape and provide the rationale for targeting HE4 to restore normal tumor immune editing.
We are currently working to identify small molecules and/or neutralizing antibodies to further validate the utility of HE4 inhibition as a novel immunotherapeutic in the treatment of ovarian cancer. However, several barriers remain in the achievement of this objective. For example, PBMCs in ovarian cancer patients may already be exposed to a chronically high level of HE4, which may have differing effects than the acute exposure performed in our study. Secondly, due to multiple complicated steps in the subtractive hybridization procedure, this study stands on the data from a single donor. The benefit of this experimental strategy lies in perspicuous outcomes; however, it also introduces inherent limitations in interpretation of the results. To begin to circumvent this pitfall, we validated the HE4-mediated downregulation of OPN using flow cytometry, qPCR, and ELISA in PBMCs from four healthy donors.
This issue will be further addressed in subsequent studies on HE4. Lastly, it is important to note that OPN is known to play a role in humoral immunity (34)(35)(36).
Further studies are required to fully understand the role of HE4 and OPN in humoral immunity in relation to ovarian cancer. Additionally, as we showed in Table 1 that PBMCs modulated a variety of genes in response to HE4 exposure. It is therefore very likely that other factors, besides osteopontin, are also contributing to in the inhibitory effect of HE4 on the immune system. Further analysis of the functions of these genes, and how they are associated with HE4, is warranted.

II.1 Abstract
Objective Selective overexpression of Human epididymal secretory protein 4 (HE4) points to a role in ovarian cancer tumorigenesis but little is known about the role the HE4 gene or the gene product plays. Here we examine the role of the secretory glycoprotein HE4 in ovarian cancer immune evasion.

Methods
Through the modified subtractive hybridization analyses of human peripheral blood mononuclear cells (PBMCs), we have characterized gene targets of HE4 and established a preliminary mechanism of HE4-mediated immune failure in ovarian tumors.

Results
Dual specificity phosphatase 6 (DUSP6) emerged as the most upregulated gene in PBMCs upon in vitro exposure to HE4. CD8 + cells and CD56 + cells found to be sources of the upregulated DUSP6. The HE4 exposure enhanced Erk1/2 phosphorylation specifically in these cell populations and the effect was erased by coincubation with DUSP6 inhibitor, (E)-2-benzylidene-3-(cyclohexylamino)-2,3dihydro-1H-inden-1-one (BCI). In co-culture with PBMC, HE4-silenced SKOV3, a human ovarian carcinoma cell line, exhibited enhanced proliferation with exposure to the external HE4; this effect was partially attenuated by adding BCI to the culture.
Additionally, the reversal effects of BCI were erased in the co-culture with CD8 + / CD56 + cell deprived PBMC.

Conclusion
Taken together, these findings show that DUSP6 is a suppressor of the cytotoxicity of the CD8 + and CD56 + lymphocytes and HE4 enhances tumorigenesis of ovarian cancer through the compromised cytotoxicity of the CD8 + and CD56 + cells by upregulation of self-produced DUSP6, which acts as an autocrine factor.

II.2 Introduction
Epithelial ovarian cancer (EOC) is the fifth leading cause of cancer death in women, and the deadliest gynecologic cancer. The American Cancer Society estimates that in 2017, there will be an estimated 22,440 new cases of EOC and 14,080 deaths in the United States [1]. Only 15% of patients are diagnosed at an early stage when the disease is fundamentally curable, keeping the 5-year survival rate at a dismal 46% [2].
Recurrence following initial treatment is common, occurring in approximately 80% of cases, and all patients with recurrent disease eventually succumb to their illness [3].
These dire statistics highlight the need for continued research into improved diagnostic and treatment options for EOC.
Despite continued efforts, there remains a lack of effective treatments for EOC.
Standard first-line therapy consists of debulking surgery followed by taxane-platinum chemotherapy [3]. Other targeted therapies are also employed, including the antiangiogenic drug bevacizumab and the PARP inhibitor olaparib; however, these treatments have not led to an improvement in overall survival [4]. One promising new area of investigation lies in understanding how tumors develop immune tolerance and evade elimination by cytotoxic lymphocytes. Immune checkpoint molecules such as PD-1, CTLA4, TIM3, IDO, and others, suppress T cell activation and help tumor cells escape targeting and elimination by the immune system [5]. Nivolumab, a monoclonal antibody against PD-1, is expressed on T cells and suppresses their activation upon binding of its tumor cell associated ligands, PDL1/PDL2, has greatly improved survival for metastatic melanoma patients [6]. PD-1 has also been studied in relapsed platinum-resistant EOC; however, overall response rates for EOC do not exceed 15% [7]. This inefficacy of immune checkpoint inhibitors is likely due to compensatory immune suppressive pathways [8,9], or activation of oncogenic pathways that promote immune tolerance [5]. Overall, we require a greater understanding of factors that contribute to immune evasion in EOC in order to develop treatments that reactivate the body's immune response to tumors.
Human epididymis protein-4 (HE4) is a member of the whey acidic four-disulfide core protein family [10]. It is elevated in tumor tissue and serum of EOC patients, and is used as part of the Risk of Ovarian Malignancy Algorithm (ROMA)-along with CA125 and menopausal status-for the diagnosis and management of EOC [11,12].
We hypothesized that HE4 may also promote immune evasion in EOC. We began to test this hypothesis by determining HE4-mediated gene expression in peripheral blood mononuclear cells (PBMCs), and went on to evaluate the effect of HE4 and one of its targets, DUSP6, on immune cell function and cytotoxicity against ovarian cancer cells.

II.3 Methods
Subtractive hybridization and TA-cloning 5 x 10 7 PBMCs from single donor were suspended in 5 mL of serum free RPMI1640 medium (Invitrogen, 31800) and incubated with or without 0.01 g/mL of rHE4 (Abcam, ab184603) for 6 hours. Then, total RNA was isolated using TRIzol TM Reagent (Invitrogen, 15596018). Next, mRNA was purified using Magnosphere TM

Statistics
Data ware expressed as average ± SEM of at least four independent experiments. An unpaired, two-tailed Student t-test was used to determine significance. Multiple treatments were analyzed by using one-way ANOVA followed by Ryan's multiple comparison test. Differences between groups were considered statistically significant when p < 0.05.

Differential expression of PBMC genes after HE4 exposure
To identify differentially expressed genes after HE4 exposure, modified subtractive hybridization was performed. PCR products of the differentially expressed genes were cloned into pUC19-TA vectors to create a differential cDNA library. PCR products from 250 each of HE4-induced and HE4-suppressed gene colonies were sequenced resulting in the identification of 209 induced genes and 206 suppressed genes. Among the identified genes, 20 induced and 13 suppressed sequences showed no significant similarity (NSS) to known genes in available nucleotide databases. Among the 209 induced genes, dual specificity phosphatase 6 (DUSP6) emerged as one of the most frequently identified genes (3 times out of 250 sequences, 1.2%; Table 1).

HE4 attenuates ovarian cancer susceptibility to PBMC mediated cytotoxicity
In order to evaluate the impact of HE4 on PBMC cytotoxicity against cancer cells, the human ovarian tumor cell line, SKOV3, was co-cultured with PBMCs (5 x 10 6 / mL density). To minimize the effect of native HE4 produced by tumor cell, the SKOV3 cells were stably transfected with HE4 specific shRNA (shHE4). The effector cells (PBMCs) were washed away, and the target cells (SKOV3) were analyzed by three independent modalities: cell viability, Ki67 immunostaining, and flow cytometry for propidium iodide (PI) and annexin V. First, SKOV3 cells co-cultured with PBMC suspensions containing 0.01 g/mL of rHE4 showed significantly higher viability than cells cultured with the rHE4 free suspensions at 24 Figure 4A). Second, immunohistochemistry using anti-Ki67 was performed to evaluate the proliferation activities of SKOV3 cells in the presence of PBMCs with or without rHE4 and BCI for 24 hours. The number of Ki67 positive tumor cells in rHE4-containing PBMC suspension was higher than the cells in rHE4-free suspension, and the increased activity was partially attenuated by adding BCI to the culture (27.6 ± 1.7 %, 68.5 ± 2.6 % and 48.9 ± 2.3 %, respectively; Figure 4B). Finally, after a 6 -hour incubation at 37 degrees, the effector cells were washed away and the target cells were analyzed by 2color flow cytometry using PI and Alexa Fluor ® 488 labeled annexin V. As shown in Figure 4C, SKOV3 / PBMC co-cultures with rHE4 led to a significant decrease in populations of PI / annexin v double positive dying cells (24.3 ± 1.2 % vs. 13.4 ± 0.8 %, p < 0.01), and the tolerance of the target cells was partially reversed by adding BCI to the culture (18.1 ± 0.6 %, p < 0.01 vs. CTR and HE4). These findings suggest that HE4 enhances tolerance of cancer cells against immunocompetent mononuclear cells via up-regulation of DUSP6 in PBMCs. In order to confirm involvement of CD8 + / CD56 + cytotoxic lymphocytes in the HE4 induced immunomodulation, the co-culture study was repeated using PBMCs deprived of CD8 + / CD56 + cells. As shown in Figure   5A-C, all the effects of BCI shown in Figure 4 were erased in the CD8 + / CD56 + cell free co-cultures, suggesting that the cytotoxic lymphocytes play a pivotal role in the immunoediting by DUSP6 up-regulation in response to exposure to HE4.

II.5 Discussion
Several studies from our laboratory and elsewhere have revealed multidimensional roles for HE4 in the pathogenesis of ovarian cancer, including the promotion of tumor growth, chemoresistance, anti-estrogen resistance, invasion, migration, and adhesion [14][15][16][17][18][19][20][21][22][23]. In this present study, we have begun to delineate another vital function of HE4 in disrupting immune cell function, which has implications for immune system targeting of tumor cells. DUSP6, which we found to be upregulated by rHE4 treatment in CD8 + T cells and CD56 + NK cell subsets of PBMCs, is likely one key mediator of this effect in these immune cell subsets.
DUSP6 is a member of the DUSP family that dephosphorylates threonine and tyrosine residues on MAPK substrates. It specifically dephosphorylates ERK, a member of the MAPK family that also includes p38 and JNK. MAPKs are activated by growth factors, cytokines, integrin ligands, and stress signals to regulate growth, survival, apoptosis, and immune response in diverse cell types. Interestingly, DUSP6 is expressed at low levels in resting cells and is actually stimulated by ERK activation, promoting a negative feedback loop on ERK activity [27]. This early response of DUSP6 to ERK activation could explain the apparently contradictory activation of ERK by HE4 in cancer cells [14,16,17,23] and our results showing that HE4 upregulation of DUSP6 expression leads to suppression of ERK phosphorylation in PBMC subsets.
Several reports reveal a role for DUSP6 in development, organogenesis, and cancer [27]. However, its effect on cancer progression is highly dependent upon the type of cancer and even the stage. For example, in pancreatic cancer, it is upregulated in early stages but is often completely diminished as the tumor progresses towards the invasive ductal carcinoma state [28]. In lung cancer, it has been shown to act as a tumor suppressor [29]. Conversely, it is upregulated in glioblastoma and HER2-positive breast cancer [30,31]. One report found that its downregulation in ovarian cancer results in hyperactivation of ERK and subsequent chemoresistance [32]. These discrepancies are likely due to variable deregulation of ERK signaling and compensatory pathways that are highly context dependent [27]. In contrast to the roles of the tumor producing DUSP6 on the tumorigenesis, the functions of DUSP6 originated from immune cells have rarely been evaluated.
Even less is known regarding the role of DUSP6 in immune cell function. Other members of the DUSP family, including DUSP1, DUSP2, and DUSP10, are known to have roles in immune response [27], and a few reports suggest that DUSP6 does as well. Elevated DUSP6 was shown to cause downregulation of ERK phosphorylation in CD4 + T cells in elderly individuals, who have suppressed immune responses [33].
Another report confirmed this age associated rise in CD4 + T cell DUSP6 expression, and found that young immunosuppressed patients with end stage renal disease have DUSP6 levels comparable to elderly healthy individuals [34]. One study found that DUSP6 downregulates ERK activity in CD4 + T cells and increases their regulatory T cell functions [35]. Together, these reports suggest that higher levels of DUSP6 contribute to immune suppression. It has also been shown that DUSP6 is downregulated in T cells upon IL-2 withdrawal [36], and IL-2 was found to upregulate DUSP6 gene expression in T cells [37]. Since IL-2 stimulates cytotoxic T cell expansion and activation as well as that of immune suppressive regulatory T cells [38], it remains to be determined how the IL-2 responsiveness of DUSP6 plays into its apparent effect on immune suppression, and how this relates to tumor immune response.
Although much remains unknown regarding the specific effects of DUSP6 on cancer progression and tumor immunity, our findings begin to reveal some novel insights. We report for the first time that HE4-mediated upregulation of DUSP6 in CD8 + T cell and CD56 + NK cell subsets of PBMC cells leads to the inhibition of their cytotoxic activity against SKOV3 ovarian cancer cells. While DUSP6 has been connected to immune function of CD4 + T cells, our results reveal that the subsets of lymphocytes affected by DUSP6 are context dependent. Further investigation into the inhibitory effects of DUSP6 in these different populations will be illuminating. Moreover, we have begun to establish HE4 as a critical regulator of immune cell function, which deepens our understanding of the mechanistic role HE4 plays in ovarian cancer pathogenesis.         15.62 ± 0.97*** 0.77 ± 0.10 1.43 ± 0.14*** *I 2.5 mg/mL of total protein(ng/mL) **in 5 mL media of 5 x 10 6 PBMC culture The mean ± are shown, n = 10 / each group, ***< 0.01 vs CTR

III.1 Abstract
Dual Specificity phosphatase 6 (DUSP6) is a phosphatase that deactivates extracellular-signal-regulated kinase (ERK). Since the ovarian cancer clinical biomarker human epididymis protein 4 (HE4) has been shown to interact with the ERK pathway, the objective of this study was to determine the relationship between DUSP6 and HE4 and begin to elucidate the role of DUSP6 in epithelial ovarian cancer (EOC). Western blot and quantitative PCR following knockdowns showed that HE4 and DUSP6 levels were reduced with knockdown of the other protein in SKOV3 and OVCAR8 ovarian cancer cells. Furthermore, DUSP6 levels were upregulated in cells overexpressing HE4. Since HE4 has been shown to promote chemoresistance in EOC, the effect of DUSP6 on chemotherapeutic response was evaluated. MTS assay revealed a significant decrease in cell viability with pharmacological inhibition of DUSP6 using BCI in cells treated with carboplatin or paclitaxel, compared to treatment with single-agent chemotherapy alone. Quantitative PCR was used to evaluate gene expression responses to BCI, recombinant HE4 (rHE4), carboplatin, paclitaxel, and combinatorial treatments. DUSP6 inhibition with BCI altered expression of ERK pathway response genes, including early growth response protein 1 (EGR1) and c-Jun. Expression of EGR1, a strong promotor of apoptosis, was higher in ovarian cancer cells co-treated with BCI and paclitaxel or carboplatin than in cells treated with chemotherapeutic agent alone. Alternatively, the expression of c-Jun, a proto-oncogene, decreased with co-treatment of BCI and paclitaxel or carboplatin. The effect of BCI on the expression of these two genes opposed the effect of rHE4 on their expression. Finally, expression levels of DUSP6 in EOC tissue were evaluated by

III.2 Introduction
Epithelial ovarian cancer (EOC) remains the most common and deadly gynecologic cancer, responsible for 240,000 diagnoses and 152,000 deaths worldwide each year [1]. The 5-year survival rate remains at 35% [2], which is largely due to difficulty with early diagnosis, coupled with the frequency of chemoresistant recurrences [3].
Although a majority of EOC is initially responsive to chemotherapy, once the disease recurs, chemoresistance inevitably develops and the patient eventually will succumb to their illness [4]. Therefore, there is a need for improved diagnostic approaches, as well as novel treatment targets to combat chemoresistance.
Human epididymis protein 4 (HE4) has been established as a novel clinical biomarker for EOC. Inclusion of preoperative levels of HE4 into the diagnostic Risk of Ovarian Malignancy Algorithm (ROMA) results in demonstrably improved specificity and sensitivity in detection and monitoring of the disease over Cancer Antigen 125 (CA 125), pelvic sonography, and menopausal status [5]. Research has also shown its mechanistic involvement in promoting EOC pathogenesis, including the promotion of proliferation, chemoresistance, anti-estrogen resistance, adhesion, invasion, and migration [6][7][8][9][10][11][12][13][14][15][16]. One oncogenic pathway that has been shown to interact with HE4 in several studies is the extracellular signal regulated kinase (ERK) pathway. Several reports indicate that ERK activation is enhanced with HE4 treatment or overexpression, while ERK activation is reduced with HE4 knockdown [8,14,15].
Our lab has revealed a more complicated response of ERK to recombinant HE4 treatment; specifically, we have observed downregulation of ERK phosphorylation at early time points, and upregulation at later time points [8]. Although the exact mechanism of HE4 interaction with the ERK pathway is not clarified, it is well established that HE4 mediates ERK activation in EOC.
Dual specificity phosphatase 6 (DUSP6) is a key negative regulator of ERK signaling via dephosphorylation of ERK at serine/tyrosine residues. ERK activation upregulates gene expression of DUSP6, which promotes a negative feedback loop on ERK activation [17]. DUSP6 has been shown to have differing effects on tumor progression depending on the tumor type. In pancreatic cancer, it is initially upregulated, but diminished at later stages, and is considered a tumor suppressor [18]. It is also considered a tumor suppressor in lung cancer [19]. However, in glioblastoma and HER-2 positive breast cancer, it has been shown to be upregulated [20,21]. In gastric cancer, DUSP6 inhibition can overcome chemoresistance [22], and it has also been characterized as a therapeutic target in acute lymphoblastic leukemia [23]. One study in ovarian cancer suggested that it may act as a tumor suppressor [24]. The goal of the present study was to determine the relationship between HE4 and DUSP6 in EOC and begin to elucidate the role of DUSP6 in EOC.

III.3 Methods
Cell Culture, Treatments, and siRNA Knockdowns

Western Blot
Western blot was performed as previously described [9]. GAPDH was used as a loading control. Antibodies and dilutions used are as follows: Densitometry Image J "analyze gel" function was used to perform densitometry analysis of western blot images in 8-bit TIFF format. Band densities were normalized to GAPDH, and the lowest value was set to 1 for plotted graphs.
Quantitative RT-PCR Quantitative RT-PCR was performed as previously described [9].

HE4 and DUSP6 Levels Are Co-Dependent in Ovarian Cancer Cell Lines
We first confirmed the upregulation of DUSP6 by HE4 by examining mRNA and protein levels in SKOV3 and OVCAR8 ovarian cancer cells stably overexpressing HE4 (clone 1 and clone 5, respectively) or their null vector (NV) counterparts.

Inhibition of DUSP6 Sensitizes Ovarian Cancer Cells to Chemotherapeutic Drugs
Next, we wanted to begin to determine the function of DUSP6 in ovarian cancer cells.
Since one well-known role of HE4 in EOC is the promotion of chemoresistance, we treated SKOV3 and OVCAR8 cells with a DUSP6 inhibitor (BCI) alone or in combination with paclitaxel or carboplatin, the standard of care chemotherapeutic agents in EOC. Treatment of cells with BCI alone resulted in a small but significant reduction in cell viability as determined by MTS assay -86.3% and 84.7% in OVCAR8 and SKOV3, respectively. In both cell lines, co-treatment with BCI and carboplatin resulted in a synergistic effect on cytotoxicity compared to either treatment alone. Carboplatin alone treatment resulted in 89.8% and 86.8% survival in OVCAR8 and SKOV3 cells, respectively, while BCI with carboplatin resulted in 33.9% and 50.2% survival in OVCAR8 and SKOV3 cells, respectively. In OVCAR8 cells, a synergistic effect was noted with BCI and paclitaxel treatment as well, with survival reducing from 51.4% with paclitaxel alone to 25.3% with BCI and paclitaxel ( Figure   2A-B).

DUSP6 Inhibition Alters Expression of ERK Pathway Responsive Genes
In order to determine how regulation of ERK signaling by BCI versus rHE4 might affect downstream gene expression, we treated cells with BCI alone or in combination with rHE4, paclitaxel, or carboplatin, and examined expression of the ERK pathway response genes EGR1 and c-Jun. EGR1 is a transcription factor involved in promoting apoptosis in many cancers [25][26][27][28], and has been shown to be involved in cisplatin resistance in esophageal and ovarian cancers [29,25]. We have previously shown that HE4 suppresses EGR1 gene upregulation in response to cisplatin treatment of SKOV3 cells [8]. On the other hand, c-Jun is an AP-1 transcription factor involved in promoting cell survival and growth [30,31]. Treatment with BCI modestly upregulated EGR1 expression in both cell lines, while treatment with rHE4 downregulated EGR1 expression-a result that is in agreement with our previous study showing HE4 suppresses cisplatin-mediated upregulation of EGR1. The effect of BCI on EGR1 expression was more apparent with rHE4 co-treatment, where it reversed the downregulation of EGR1 by rHE4. Furthermore, co-treatment with BCI and either paclitaxel or carboplatin upregulated expression of EGR1 compared to treatment with either chemo drug alone. These results show that BCI opposes the effects of HE4 on EGR1 expression and promotes EGR1 expression while suppressing c-Jun expression in cells exposed to chemotherapy drugs ( Figure 3A-D). Together, these results suggest that DUSP6 may be involved in promoting tumorigenesis in EOC, and corroborate our results indicating a relationship between HE4 and DUSP6.

III.5 Discussion
In this study, we have determined that HE4 and DUSP6 levels are co-dependent in ovarian cancer cells, and that these two proteins interact and are correlated in patient tissue. Future studies are needed to elucidate the exact mechanistic relationship between DUSP6 and HE4. Studies by us and others have confirmed that HE4 activates ERK in ovarian cancer cells [8,14,15], while DUSP6 is a known negative regulator of ERK signaling [17]. Interestingly, despite the fact that HE4 and DUSP6 have opposing roles on ERK activation, they appear to produce similar effects on biological function of tumor cells. Our results show that activation of ERK by the DUSP6 inhibitor BCI as opposed to HE4 produces very different effects on gene expression and cellular functions such as chemotherapy response.
The two ERK responsive genes we have characterized show opposite expression patterns with BCI treatment. EGR1 is activated by ERK via the transcription factor ELK-1, and EGR1 is itself a transcription factor that activates expression of proapoptotic genes [32]. A previous study by our lab showed that HE4 overexpression in SKOV3 cells suppresses cisplatin-mediated upregulation of EGR1 [8]. Here, we observe that HE4 downregulates EGR1 expression, which is consistent with these previous results. Conversely, BCI treatment opposes the effect of rHE4 on EGR1 expression, indicating differing effects downstream of ERK activation by these two treatments. C-Jun, which is also an ERK responsive gene, is regulated oppositely as EGR1. rHE4 treatment upregulates expression of c-Jun, which is consistent with its role as a promoter of tumor growth and proliferation [6,12,13,33,34]. Meanwhile, BCI again opposes this effect in BCI and rHE4 co-treated cells. Furthermore, BCI suppresses chemotherapy-mediated increases in c-Jun levels. The effects of BCI on EGR1 and c-Jun together may contribute to the overall increased efficacy of BCI and chemotherapy treatment over chemotherapy alone.
The role of DUSP6 in EOC is not well studied. One report showed that DUSP6 appears to function as a tumor suppressor in EOC [24], but our results suggest the opposite effect. Therefore, further study is needed to fully elucidate the role of DUSP6 and determine if its function is context dependent. In general, DUSP6 remains an interesting protein, in that it has opposing roles in different tumor types. In some cancers, it appears to act as a tumor suppressor, while in others it acts to promote tumorigenesis and aggressive behavior [19][20][21][22][23][24]. Our results are consistent with a recent study by Wu et al. (2018) showing its involvement in cisplatin resistance in gastric cancer [22]. The authors observed an increase in phospho-ERK with BCI treatment, but a downregulation of the ERK-response genes RPS6KA1, EGR1, MMP2, MMP9, MYC, and ELK3. Furthermore, they found that BCI treatment enhanced cisplatin sensitivity in gastric cancer cells and in vivo xenografts. In our study, we observed different effects of DUSP6 inhibition on ERK-response genes depending upon gene function-namely, upregulation of the tumor suppressor EGR1 and downregulation of the proto-oncogene c-Jun. Collectively, our study and the one by Wu et al. illustrate that the relationship between ERK activation and downstream gene activation is not straightforward and appears to be highly context-dependent.
Therefore, although BCI serves to increase ERK activation, it has different effects on ERK response genes, which serve to enhance chemotherapy efficacy.
In conclusion, this study highlights a novel function of DUSP6 in EOC and reveals that it may be involved in regulating chemoresponse. Targeting HE4 and/or DUSP6 in EOC may be an effective method of reversing chemoresistance and improving longterm response rates in select patient populations.

IV.2 Introduction
Epithelial Ovarian Cancer (EOC) is the most lethal gynecologic malignancy [1]. In 2018, there will be an estimated 22,240 new cases of EOC diagnosed and 14,070 deaths in the United States. While EOC accounts for only 2.5 % of all female cancers, it is responsible for 5% of all cancer deaths due to low disease survival rates [2].
These dire statistics are attributed to the fact that the majority of patients are diagnosed at an advanced stage. In addition, while patients generally respond well to frontline platinum-based chemotherapy, chemoresistant recurrences are common [3]. Therefore, there is a strong need for novel early detection methods and targeted therapies for EOC patients.
Septin-2 is a member of the septin family, a conserved family comprised of 13 GTP binding proteins [4]. Septins, which are structurally observed as rods and filaments, are vital to a number of cellular processes, including cytokinesis, vesicle trafficking, and exocytosis [5]. They are considered to be a fourth component of the cytoskeleton due to their association with actin, microtubules, and membranes [6]. Septins have been identified as having a role in neurodegenerative disease, since they were detected in brain tissue from patients with Alzheimer disease [7]. In addition, they have been reported to be involved in bacterial infections, Parkinson's disease, and male infertility [8].
In more recent years, emphasis has been placed on investigating the role of septins in tumorigenesis [9]. Due to their natural function in scaffolding and membrane compartmentalization, it is plausible that they could also play a role in the organization of membrane associated proteins involved in diverse tumorigenic signaling pathways [6]. Septin-9 is the best studied septin family member in relationship to cancer, and its methylation status is utilized as a biomarker in colorectal cancer [10]. However, there have also been numerous studies linking septin-2 to neoplasia. Thus far, septin-2 has been specifically implicated in Hodgkin's lymphoma and biliary tract, gastric, hepatocellular, and breast cancer [11][12][13][14][15], but its role in EOC has not yet been investigated.
In this study, we begin to elucidate septin-2's function in EOC. As septins have been shown to have diverse roles in tumorigenesis, this is the first step in specifically defining septin-2's contribution to EOC pathogenesis. To establish the clinical relevance of septin-2 in EOC, we first sought to compare levels of septin-2 in various histological pathologies of EOC versus benign disease. Furthermore, we present for the first time a global analysis of septin-2 mediated proteomics in EOC and describe signaling pathways most affected by septin-2 depletion. The results from this study lay the framework for future mechanistic studies to determine the precise role of septin-2 in EOC.  The LC-MS/MS was performed on a fully automated proteomic technology platform [16,17] that includes an Agilent 1200 Series Quaternary HPLC system (Agilent Technologies, Santa Clara, CA) connected to a Q Exactive Plus mass spectrometer (Thermo Fisher Scientific, Waltham, MA). The LC-MS/MS set up was used as described earlier [18]. Briefly, the peptides were separated through a linear reversedphase 90 min gradient from 0% to 40% buffer B (0.1 M acetic acid in acetonitrile) at a flow rate of 3 µl /min through a 3 µm 20 cm C18 column. The electrospray voltage of 2.0 kV was applied in a split flow configuration, and spectra were collected using a top-9 data-dependent method. Survey full scan MS spectra (m/z 400-1800) were (Matrix Science, Ltd, London, UK). A concatenated database containing "target" and "decoy" sequences was employed to estimate the false discovery rate (FDR) [19].

Cell
Msconvert from ProteoWizard (v. 3.0.5047), using default parameters and with the MS2Deisotope filter on, was employed to create peak lists for Mascot. The Mascot database search was performed with the following parameters: trypsin enzyme cleavage specificity, 2 possible missed cleavages, 10 ppm mass tolerance for precursor ions, 20 mmu mass tolerance for fragment ions. Search parameters permitted variable modification of methionine oxidation (+15.9949 Da) and static modification of carbamidomethylation (+57.0215 Da) on cysteine. The resulting peptide spectrum matches (PSMs) were reduced to sets of unique PSMs by eliminating lower scoring duplicates. To provide high confidence, the Mascot results were filtered for Mowse Score (>20). Peptide assignments from the database search were filtered down to a 1% FDR by a logistic spectral score as previously described [19,20].

Relative quantitation of the identified peptides
Relative quantification of peptide abundance was performed via calculation of selected ion chromatograms (SIC) peak areas. Retention time alignment of individual replicate analyses was performed as previously described [21]. Peak areas were calculated by inspection of SICs using in-house software programmed in R 3.0 based on the Scripps Center for Metabolomics' XCMS package (version 1.40.0). This approach performed multiple passes through XCMS' central wavelet transformation algorithm (implemented in the centWave function) over increasingly narrower ranges of peak widths and used the following parameters: mass window of 10 ppm, minimum peak widths ranging from 2 to 20 seconds, maximum peak width of 80 seconds, signal to noise threshold of 10 and detection of peak limits via descent on the nontransformed data enabled. SIC peak areas were determined for every peptide that was identified by MS/MS. In the case of a missing MS/MS for a particular peptide, in a particular replicate, the SIC peak area was calculated according to the peptide's isolated mass and the retention time calculated from retention time alignment. A minimum SIC peak area equivalent to the typical spectral noise level of 1000 was required of all data reported for label-free quantitation. Individual SIC peak areas were normalized to the peak area of the standard synthetic peptide DRVYHPF that was exogenously spiked prior to reversed-phase elution into the mass spectrometer.
Quantitative analysis was applied to replicate experiments. To select peptides that show a statistically significant change in abundance between control vs treatment cells, q-values for multiple hypothesis tests were calculated based on p-values from two-tailed unpaired Student's t tests using the R package QVALUE as previously described [22,23].

Septin-2 is overexpressed in EOC
A preliminary proteomic study determined interacting partners of the clinical EOC biomarker HE4. It was noted that septin-2 was the most upregulated HE4-interacting protein (13-fold) in SKOV3 ovarian cancer cells overexpressing HE4 compared to null vector cells (data not shown). This finding prompted us to begin to characterize septin-2's role in EOC, as it had not been previously documented in the literature. To establish the clinical relevance of septin-2 in EOC, we evaluated its levels in EOC samples of a variety of histopathologies and compared these to levels in benign controls. Immunohistochemical analysis of septin-2 levels in a human ovarian tissue microarray comprising normal, serous, mucinous, clear cell, and dysgerminoma histopathologies revealed that mean intensity of the septin-2 staining was statistically significantly greater in serous EOC (703.3889 pixels) than in adjacent normal tissue (539 pixels) (p=0.0037) (Fig.1a). While all other histopatholgies exhibited higher mean intensity levels of septin-2-mucinous (603 pixels), clear cell (821 pixels), and dysgerminoma (744 pixels)-compared to the normal adjacent tissue, none where considered statistically significant possibly due to low numbers of samples available.

Stable knockdown of septin-2 influences cell proliferation
In order to study septin-2's function in EOC, stable septin-2 knockout shRNA clones were generated in human serous ovarian SKOV3 wild type (WT) cells. Two clonal populations were employed for these studies-knockout 9 (KO9) and knockout 11 (KO11)-based on confirmation of successful septin-2 downregulation. A stable line was also generated by clonal expansion of cells transfected with control shRNA, designated Plasmid C. To confirm the efficacy of knockdowns at the genomic level, qPCR was employed. Septin-2 levels in KO9 were 1.93-and 4.16-fold lower than WT and Plasmid C cells, respectively. Septin-2 levels in KO11 were 1.67-and 3.88-fold lower than WT and Plasmid C cells, respectively (Fig 2a).
To further validate successful knockdown of septin-2, protein levels were detected by western blot. We observed substantial decreases septin-2 levels in KO9 and KO11 compared to the WT and Plasmid C controls (Fig 2b). Septin-2 levels in KO9 were decreased by 72% compared to WT and by 62.3% compared to Plasmid C. Septin-2 levels in KO11 were reduced by 76.4% and 67.7% compared to WT and Plasmid C, respectively (Fig.2c).
To begin to determine the consequence of septin-2 knockdown in SKOV3 cells, proliferation of the shRNA clones was evaluated. WT, Plasmid C, KO9, and KO11 cells were seeded at equal cell densities and allowed to expand. The cells were trypsinized at 72 and 96 hours, and numbers of live cells in each clonal population were quantified (Fig 2d). At 72 hours, KO9 clones exhibited a 67.5% decrease in cell proliferation compared to WT, and a 60.4% decrease compared to Plasmid C. KO11 clones demonstrated a 66.4% and 59.1% decrease in proliferation from respective WT and Plasmid C cell numbers. The 96-hour timepoint revealed a 51.1% reduction in KO9 cells compared to WT and a 39.3% reduction compared to Plasmid C. KO11 cells showed a 62.6% and 53.6% decrease compared to WT and Plasmid C cells, respectively. All decreases in cell counts displayed by KO9 and KO11 at both timepoints were determined to be statistically significant (p<0.02). This finding strongly suggests that the downregulation of septin-2 has a profound impact on cell proliferation in EOC cells.

Proteomic analysis of septin-2 knockdown in EOC cells
A comparative label-free proteomic analysis was performed to examine global protein expression level differences resulting from the knockdown of septin-2. Interestingly, significant differences in protein-peptide levels between control cells and septin-2 knockouts was observed only in KO11 populations, even though our proliferation results demonstrated that KO9's phenotype was similar to that of KO11. We concluded that it was possible that the knockdown resulted in less significant effects on protein levels, but still enough to affect proliferation, or that spontaneous loss of the knockdown had occurred during cell culture. Therefore, we proceeded with analysis using KO11 cells. As expected, a principal component analysis of three biological replicates of WT, Plasmid C, and KO11 revealed separate clusters when comparing principal component 1 and principal component 2 scores (Fig.3). In contrast, for KO9 sample, the 3 biological replicates were very scattered (Data not shown). Therefore, for any further analysis or validation process KO9 was not included.
Mass spectrometry of the control and knockdown cells identified 19976 unique peptides corresponding to 3565 unique proteins. Of those, only one peptide/protein in Plasmid C exhibited an absolute fold change greater than 1 with a q-value < 0.05 compared to WT (Fig 4a). This result allowed us to conclude that there was no significant difference between both control cell populations. Conversely, 5% of all peptides in KO11 cells revealed relative fold change greater than 1 (q<0.05) compared to WT cells. In addition, 93.5% of those peptides identified as exhibiting substantial expression differences displayed a lower peak area in KO11 than WT, indicating a majority of peptides was downregulated (Fig 4b). Representative examples of peakarea of four peptide sequences from the proteins galetin-3 binding protein (LFALS3BP), transketolase (TKT), poly(A) binding protein (PABPC4), and enolase-

IV.5 Discussion
For the first time, we have characterized septin-2 function in EOC and examined its proteomic effects on a global level. Several biological pathways were found to be differentially regulated in septin-2 knockout ovarian cancer cells, exemplified by representative proteins from (Fig 4c.) Galectin-3 is a member of the β-galactoside binding protein family that is involved in diverse functions inherent to cancer, such as metastasis, immune surveillance, inflammation, apoptosis, molecular trafficking, and mRNA splicing [28]. Transketolase is a pentose phosphate pathway enzyme essential for cancer growth due to its ability to control NADPH production and counteract oxidative stress [26]. Poly(A) binding protein is a highly conserved protein that plays an important role in mRNA stabilization and translation [29], which controls cell growth, proliferation, and differentiation [30]. Enloase1, found to be differentially expressed in cancer, is a key glycolytic enzyme that catalyzes 2-phosphogylgerate to phosphoenolpyruvate in the last steps of the glycolytic catabolic pathway [31].
Of these pathways identified, it was most expected that autophosphorylation and proteasomal/ubiquitin protein functions were affected by septin-2 knockdown. It has been previously established that proper control of septins' phosphorylation status is required for the completion of cytokinesis [32]. In fungus, Meseroll et. al (2013) discovered that changes in specific phosphorylation sites on septins (Cdc3p and Cdc11p) leads to the disruption of higher order septin structures, indicating septin phosphorylation is also a vital regulator of their own structure formation [33].
Similar to phosphorylation, ubiquitination represents another important septin posttranslational modification. Septins have an established role interacting with proteins involved in degradation pathways, such as ubiquitin ligases and de-ubiquitylating enzymes, which modulates protein turnover [12,34,35]. Recently, it has also been reported that SUMOylation of human septins is a critical process contributing to proper septin filament bundling and cytokinesis [36]. Unlike ubiquitin, SUMO (small ubiquitin-like modifiers) modification does not always lead to protein degradation, as SUMOylation can also modulate localization, interaction, and activity of the target protein [37]. Ribet et.al (2017) reported that septin-7 is constitutively SUMOylated throughout the cell cycle, and septin variants that are unable to be SUMOylated halt septin bundle formation and lead to defects in cytokinesis, highlighting its crucial role in septin filament bundling and cell division [36].
GO analysis revealed that septin-2 is also involved in post-transcriptional modifications, as the spliceosome pathway was found to be enriched among septin-2 regulated proteins (Fig 6). This result suggests that septin-2 plays a major role in the editing of both precursor messenger RNA (pre-mRNA) and proteins. The spliceosome, a large molecular complex involved in the removal of non-coding introns from pre-mRNA, represents a potential oncogenic target as evidence has shown that tumors rely on normal spliceosome function for cell survival [38,39]. In addition, Poly(A) binding protein, which we reported as an example of a differentially expressed protein (Fig. 4c), is a translation initiation factor that binds to the mRNA 3'poly(A) tail [30] and also influences cell growth and survival. Since we have shown that the knockdown of septin-2 promotes irregular expression of a multitude of pathways related to mRNA and protein modifications, it seems reasonable that its downregulation would also affect tumor cell growth.
As the depletion of septins can lead to cytokinesis failures, it is logical that cellular proliferation would subsequently be affected. [36] In this study, we observed a reduction in proliferation with septin-2 knockdown. (Fig.2d) Corroborating results from our study in EOC, Zhang et. al (2016) treated breast cancer cells with the broad septin inhibitor forchlorfenuron(FCF) and also observed a decrease in cell proliferation [15], which they attributed to the suppression of MEK and ERK1/2 (extracellular signal-regulated kinase 1/2) signaling [15]. Another study showed that septin-8 interacts with MAPK5 (mitogen activated protein kinase 5), further suggesting that septins play a role in the MAPK/ERK pathway [40]. Septin-9 has also been implicated in cell proliferation, as a septin-9 variant SEPT9_i1 binds to c-Jun-N terminal kinase (JNK), preventing its degradation and therefore promoting tumor cell proliferation [4]. In addition, another septin-9 variant SEPT9_i3 has been found to be phosphorylated by cell-cycle-dependent kinase 1 (CDK1), controlling entry into mitosis and promoting cell survival and proliferation [41]. These investigations highlight that septin-2, and septins in general, play an important role in cellular proliferation and potentially promote tumor growth.
Interestingly, the most novel conclusion drawn from this investigation was the robust enrichment seen in cellular metabolism and energy dynamics in proteins affected by septin-2 downregulation. This novel finding regarding septin-2 is in agreement with previous studies reporting on septin functions related to energy metabolism. One study identified that fungal septins FaCdc3 and FaCdc12 are required for lipid metabolism [42]. In addition, septin-9 was found to induce lipid droplet growth through binding to phosphatidylinositol-5-phosphate(PtdIns5P), a phospholipid with a well-established role in dynamics and intracellular membrane trafficking [43]. PtdIns5P binding in turn controls septin-9 filament formation and its interaction with microtubules [44]. Furthermore, septin-11 was found to be expressed in human adipocytes and upregulated in obese individuals. SEPT11 mRNA was positively correlated with markers of insulin resistance in adipose tissue, and silencing of septin-11 muted insulin signaling and insulin-induced lipid accumulation in adipocytes [45].
Our findings, however, represent the first time a septin family member has been implicated in cellular metabolism as it relates to tumorigenesis. Acetyl-CoA, one of the pathways most differentially expressed by septin-2, is a key metabolic player that links glycolysis, fatty acid oxidation, ketogenesis, amino acid metabolism, the TCA cycle, and lipid synthesis [46]. In normoxic conditions, acetyl-CoA is derived from glucose. However, under hypoxic conditions like in cancer, acetyl-CoA has been to found to derive from acetate, suggesting that targeting the acetyl-CoA pathway in cancer could represent a viable treatment option [47]. The TCA cycle, another important metabolic pathway, was also deregulated in the septin-2 knockdown clones.
While previous dogma stated that tumor cells do not utilize the TCA cycle for energy, it has now been found that some cancer cells with deregulated oncogenes and tumor suppressor genes actually do rely on the TCA cycle [48]. In addition, the metabolic proteins transketolase and enolase, which are involved in glycolysis and the pentose phosphate pathway, respectively, were found to be differentially expressed by septin-2 inhibition (Fig 4c), demonstrating that septin-2 is involved in various facets of cellular metabolism within EOC. Pathways related to metabolism and energy production have previously been found to contribute to EOC tumorigenesis, as it has been shown that glycolysis drives chemoresistance in EOC and that high levels of fatty acid synthase (FASN) contribute to tumor cell growth through the promotion of human epidermal growth factor [49,50]. Therefore, it can be hypothesized that the inhibition of septin-2 would exhibit a therapeutic effect in EOC via suppression of tumor metabolic pathways.
Overall, our study demonstrates the novel finding that septin-2 is involved in EOC pathogenesis. This investigation represents a springboard for future studies to determine the efficacy of septin-2 inhibition, in addition to more clearly elucidating its diverse mechanistic pathways in EOC tumorigenesis. While our proteomics study was performed in a serous ovarian cancer cell line, it would be interesting to repeat the stable knockdown experiment in a clear cell EOC line, since septin-2 was also found to be overexpressed in this histopathology. Additionally, both in vitro and in vivo studies could be performed to confirm that inhibition of septin-2 affects cell viability and tumor growth in order to determine if targeting of septin-2 synergizes with platinum-based chemotherapeutics.    Gene ontology (GO) analysis using DAVID (https://david.ncifcrf.gov/). Proteins with differential expression (n = 231, q < 0.05, in KO11 versus WT is compared with proteins (n = 3334) that showed no differential expression. Former showed enrichment for proteasomal/ubiquitin related GO terms (q << 0.05, Bonferroni) in biological process (BP) category. In cellular component (CC) and molecular function (MF) categories, differentially expressed proteins showed enrichment for ribonucleoprotein and RNA related terms. No enrichment was seen in molecular function category.
Differentially expressed proteins showed enrichment for KEGG pathways relating to citrate cycle/energy and spliceosome.

CONCLUSION
Epithelial ovarian cancer (EOC) is such a deadly disease largely owing to the two major challenges of diagnosis and treatment. Ovarian cancer is detected at a late stage when tumor cells have already detached and metastasized directly into the peritoneal cavity, making it challenging for all lesions to be removed surgically [1]. Therefore, extensive disease remains in the body even after surgery. While treatment has evolved to include PARP inhibitors and anti-angiogenic therapies, prognosis remains poor.
Immunotherapies for the treatment of cancer have recently garnered much attention, as it has been observed that the number of intratumoral T-cell numbers correlate to a better clinical outcome [2]. However, establishing a breakthrough immune target for ovarian cancer has been met with challenges, as the response rate remains low [3]. Therefore, a critical need for novel therapies for EOC still exists.
HE4 plays a unique role in EOC as it has been implicated in both diagnosis and prognosis of the disease. As a clinical biomarker, HE4 represents a promising early detection method. Compared to the more established biomarker CA-125, it is less frequently elevated in benign disease and is potentially able to identify patients that are at high risk for primary platinum resistance [4]. While much is known about HE4 clinically, far less is known about its biological functions in EOC. The goal of this investigation was to determine the mechanisms in which HE4 drives ovarian pathogenesis, and to ultimately provide evidence as to whether HE4 should be recommended as a therapeutic target for EOC.
As HE4 was initially suggested to have a potential role in innate immunity, [5] these studies aimed to better understand HE4's function in tumor immunity. For the first time, this investigation has shown that HE4 is involved in promoting ovarian tumor immune evasion, through influencing expression of two proteins, osteopontin (OPN) and dual specificity phosphatase 6 (DUSP6). Subtractive hybridization revealed that when peripheral blood mononuclear cells (PBMCs) were treated with recombinant HE4, OPN was the most downregulated protein, and DUSP6 was the most upregulated.
OPN is a secreted glycoprotein that has been identified as having important T helper 1 (Th1) cytokine functions. [6]Specifically, it was discovered that HE4 suppresses OPN in CD3+ T cells, while also impairing the secretion of IL-12 and IFN-γ, two important cytokines downstream of OPN that promote T-cell survival [6,7]. Furthermore, when ovarian cancer cells were cultured with media from PBMCs cultured with recombinant HE4, those cells were less susceptible to cell death, which was reversed upon silencing of HE4. Also, in human EOC patient tissue, serum HE4 levels inversely correlated to the number of OPN positive T cells in patient tumors.
The second objective in defining HE4's role in tumor immunity was to delineate the effect of HE4's upregulation of DUSP6. DUSP6 is an extracellular signal-regulated kinase (ERK) phosphatase that has been found to regulate CD4+ T cell activation and differentiation through the inhibition of T-cell receptor (TCR) dependent ERK activation [8]. Interestingly, upon testing HE4's upregulation of DUSP6 in specific subsets of cells within PBMCs, the upregulation was found to be restricted to CD8+ T-cells and CD56+ natural killer (NK) cells, and not CD4+ T cells. It was also discovered that HE4 promotes ERK ½ phosphorylation in these cell populations. Upon co-culture of PMBCs with ovarian tumor cells it was found that adding recombinant HE4 enhanced cell proliferation. However, this effect was attenuated by the addition of an allosteric DUSP6 inhibitor (BCI). PBMCs devoid of CD8+/CD56+ cells did not produce the same result, proving that CD8+ and CD56+ populations were solely responsible for the observed effects. This result was particularly interesting in light of HE4's hypothesize role in innate immunity, since NK cells, as part of the innate immune response, have been found to play an important role in helping tumor cells escape immune surveillance [9].
These two studies indicate that through targeting of HE4, it may be possible to restore a normal tumor immune response. To confirm this, future directions include testing the inhibition of HE4 via a neutralizing antibody in an immune competent mouse model to see how this affects tumor burden. In addition, testing HE4 inhibition in vivo in combination with platinum-based chemotherapeutics and immune checkpoint inhibitors to determine synergistic effects is important. Results from these studies will be valuable, as many successful EOC regimens are combination therapies that produce higher response rates and lower resistance rates compared with monotherapies [10].
Before HE4 can truly be recommended as a novel therapy that can remedy tumor immune evasion, results from these in vivo experiments should be obtained.
The study of DUSP6 and HE4 in immune cells lead to an additional investigation that examined DUSP6's role in epithelial ovarian cells. This was of particular interest since DUSP6 has not been well defined in cancer, and it has been published that HE4 interacts with the ERK signaling pathway in EOC [11][12][13]. This study confirmed that DUSP6 functions similarly to HE4 in EOC pathogenesis, as the inhibition of both factors promotes apoptosis in EOC cells. Furthermore, DUSP6 is overexpressed in serous EOC patient tissue and intratumoral levels of HE4 and DUSP6 correlate. Since it has been published that HE4 promotes chemoresistance in EOC [14], the effect of DUSP6 on platinum response was also evaluated. When DUSP6 was inhibited with BCI in combination with carboplatin and paclitaxel, it produced a synergistic response over single-agent chemotherapeutic. To assess downstream effects of this inhibition, it was shown that BCI altered genomic levels of the ERK related response genes early growth response protein 1 (EGR1), a strong promoter of apoptosis and proto-oncogene c-Jun [15]. EGR1 was upregulated in cells co-treated with BCI and either paclitaxel or carboplatin, compared to a single-agent treatment, while c-Jun expression was decreased upon co-treatment. This study was able to define a new role for DUSP6 within EOC, indicating that targeting this factor is important both to restore proper tumor immune function and to overcome chemoresistance in EOC cells.
Moreover, as HE4 has the ability to be detected in patient serum, it would be interesting to determine if DUSP6 could also be detected in patient blood. Additional future directions include establishing stable DUSP6 knockdown and overexpressing clones to test cancer related phenotypes. Furthermore, as HE4 overexpressing and stable knockout cell lines have been previously established, global genomic arrays could then be performed to establish similarities and differences between the overexpression and knockout populations of each factor. Finally, the last part of this thesis sought to characterize the novel protein septin-2 in EOC. Septin-2 is a member of the septins protein family, which comprises 13 GTP binding proteins that play important roles in various cellular processes including cytokinesis [16,17]. Septin-2 was identified in a small proteomics study as most enriched with HE4 immunoprecipitation in HE4 overexpressing cells versus null vector cells. For the first time, this study revealed that septin-2 is overexpressed in both serous and clear cell EOC patient tissue. Establishment of stable knockout clones in an ovarian cell lines showed that proliferation was drastically decreased in septin-2 knockout clones. Global proteome analysis was employed to determine the relevant pathways in which septin-2 is involved with in EOC, revealing that down regulation of septin-2 produced differential expression of major metabolic and cellular energy pathways.
As this was a pilot study with the simple goal of defining septin-2 in EOC, more research needs to be completed in order to understand its mechanistic role in ovarian tumorigenesis. Future directions involve an in vivo study to determine if septin-2 knockout lead to a decrease in tumor growth, alike to the reduction of cell proliferation observed in vitro. Furthermore, it will also be important to elucidate the mechanistic relationship between septin-2 and HE4, in addition to determining how septin-2 and HE4 interact with metabolic and cellular energy pathways. This is an especially original finding as both proteins have not been previously found to interact with cellular metabolism and may lead to new novel therapeutic targets for EOC.
As a reputable clinical biomarker, HE4 is valuable in the diagnosis and prognosis of EOC; however, knowledge of its role in treatment of EOC is deficient in comparison.
Overall, this thesis compilation improves the understanding of HE4's diverse biological function in EOC, through highlighting its role in the promotion of tumor immune dysfunction and characterizing novel interacting proteins. As there is a dire need for innovative targeted therapies for EOC patients, this thesis presents new evidence that inhibiting HE4 represents promise not only in downregulating molecular mechanisms that promote tumorigenesis, but also in restoration of normal tumor immune function. Furthermore, global genomic and proteomics analysis of differential HE4 levels revealed its relationship to novel factors that had not previously been characterized in EOC prior to this investigation. Taken as a whole, this dissertation offers original insights that emphasize the importance of HE4's role in the pathogenesis of EOC.