Extracellular Matrix: A Treasure Trove in Ovarian Cancer Dissemination and Chemotherapeutic Resistance

Late presentation and resistance to chemotherapeutic agents make a deadly combination for ovarian cancer patients. The treatment of these patients is thus challenging. This study explores the possible molecular mechanisms by which tumor cells interact with the extracellular matrix (ECM) constituents, forming metastatic implants and enhancing patients' sensitivity to drugs. For the literature review, PubMed was used as a database. The standard search was done using keywords "collagen, ovarian cancer, extracellular matrix, drug resistance" in different combinations, which finally yielded 32 studies meeting the inclusion/exclusion criteria. The studies included were published in the English language in the past seven years. After analyzing, we found all of them to be histopathological studies. Nine studies also used murine cell lines besides human cell lines and tissue samples from ovarian cancer patients. One study has a retrospective analysis done. Eight studies demonstrate the role of hypoxia and matrix remodeling enzymes in ovarian cancer dissemination. Genetics playing a crucial role in cancer metastasis is demonstrated in eight studies. Ten studies included shows receptors, enzymes, and spheroid organization in disease progression. Six studies address chemotherapeutic resistance. Intraperitoneal dissemination of ovarian cancer and the development of chemotherapeutic resistance depends on certain molecular interactions, and they can be targeted to improve patients' overall survival.


Introduction And Background
Ovarian cancer remains the most lethal of all gynecological cancers, with approximately 14,000 deaths each year in the United States [1]. The Annual Incidence reported in the year 2018 was 22,240 [1]. Ovarian cancer patients are often asymptomatic and are not recognized until late in the disease course. The patients usually present with distant metastasis and malignant ascites hurting the overall survival. The patients initially respond to debulking surgery and chemotherapeutic regimen usually consisting of a platinum complex (cisplatin or carboplatin) and a taxane (paclitaxel or docetaxel), nearly 80% eventually relapse into chemoresistant cases resulting in a five-year survival rate of 30% [2].
Unlike most solid tumors, which spread via lymphatic or hematogenous routes, ovarian cancer cells disseminate through direct extension into the peritoneum by shedding either as a single cell or multicellular aggregates (MCA), which interact with mesothelial cells lining the peritoneum, later invading the underlying basement membrane and spreading across the extracellular matrix (ECM) forming metastatic implants [3]. The extracellular matrix is mainly composed of collagen, laminin, fibronectin, vitronectin, proteoglycans, and gelatin, which also play a vital role in tumor invasion.
Despite all the studies, little is known regarding the molecular mechanisms of tumor cell-mesenchymal cell interactions underlying metastasis. The studies described in this review mainly address the studies that have been done on tissue samples obtained from post-operative patients with ovarian cancer or the cell lines purchased from repositories. Nevertheless, the phenomenon of tumor seeding onto the peritoneum and peritoneal organs causing bowel obstruction, ascites leading to impaired circulation, and impaired drug delivery need further exploration in in-vivo experiments on human subjects.
The following research paper aims to review all the possible mechanisms of tumor proliferation, invasion, and metastasis, the role of ECM in intraperitoneal dissemination of tumor cells, and enhancing the sensitivity to chemotherapeutic agents discovered in the year 2013-2020. Our study reviews the role of the tumor microenvironment, receptors, pathways/regulators that may serve as therapeutic targets in the disease progression of ovarian cancer patients.

Review Methods
This traditional review was done without following PRISMA guidelines. The PRISMA flow diagram was however, included to explain the search strategy.

Search Strategy
A detailed literature review was done using the PubMed database for studies published from January 2013 to November 2020. The search for the studies was done manually using the regular keywords "collagen AND ovarian cancer, extracellular matrix AND ovarian cancer, drug resistance AND ovarian cancer".

Study Records
The relevant clinical, pathological, and pathophysiological data was stored and organized in a word document in Microsoft Word. An independent reviewer (SV) screened the studies using the title and abstracts based on eligibility criteria and relevance to the research question. In case of uncertainty, full articles were reviewed to determine the eligibility for inclusion and framed into this traditional review.

Ethical Issue
All the data in the following article was collected ethically and legally. All the included studies in this review had full-text links available freely on PubMed.

Inclusion/Exclusion Criteria
Types of patients and conditions: We reviewed the studies that included tissue samples from women of any age who suffered from ovarian cancer. The reviewed study also includes cell lines prepared in the laboratory or obtained from repositories.
Types of outcomes: We searched for the mechanism underlying distant metastasis in ovarian cancer patients, which is the major reason for patients' suffering and decreased survival, the mechanisms responsible for chemotherapeutic resistance to the drugs to improve the sensitivity for a better prognosis and survival.
Types of studies: Mixed human studies relevant to the research question published in peer-reviewed journals were included. There were no geographical restrictions, and articles published only in the English language were included.

Results
Using the PubMed database, a literature search was done with four keywords, which yielded 9010 studies; after removing 333 duplicates using Mendeley citation manager, 8677 studies were left. On applying inclusion/exclusion criteria, 2781 articles remained, which were screened using title and abstracts. As a result, 2641 studies were excluded. One hundred and forty studies were left. Full-text articles were assessed, which led to the exclusion of 103 due to non-relevance to the research question, and five studies were excluded due to the unavailability of full-text links. In this review, 32 studies were finally included. The number of studies selected with each keyword is collagen (seven studies), ovarian cancer (14 studies), extracellular matrix (five studies), drug resistance (six studies), making a total of 32 studies . Figure 1 below represents the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) flow diagram.

Analysis (PRISMA) Flow Diagram
Thirty-two histopathological studies assessing the role of extracellular matrix in ovarian cancer dissemination using a total of 672 female patient's tissue samples, 103 human cell lines, and three murine cell lines were studied. Peritoneal xenografts were used in three studies.
Study characteristics: A total of 32 studies were included, from which information was extracted. A summary of the involved studies is given in Table 1  .

Hypoxic Tumor Microenvironment and Matrix Metalloproteinases
Hypoxia has been shown to promote ovarian cancer cell invasion. Natarajan et al. showed enhanced expression of proline-lysine hydroxylases and lysyl oxidase in mesothelial cells leading to increased collagen deposition and increased collagen cross-linking respectively in the hypoxic tumor microenvironment, and this collagen remodeling subsequently leads to tumor invasion as hypoxia-inducible factor-1 and factor-2 (HIF-1 and 2) gets stabilized [4]. Zhang et al. showed increased mRNA and protein levels of HIF-α and matrix metalloproteinase-13 (MMP-13) in response to hypoxia, both of which decreased after transfection with small interfering RNA (siRNA) for HIFα. Matrix metalloproteinases are enzymes responsible for degrading ECM proteins, making tumor cells invade [5]. In contrast, Sun et al. showed hypoxia-induced ovarian cancer cell invasion using hydroxyproline levels to detect collagen degradation by membrane-type 1-matrix metalloproteinase (MT1-MMP). They also showed increased mRNA and protein levels of MT1-MMP in response to hypoxia, which aids in invasions like Zhang et al. Hypoxia also causes increased Snail (a zincfinger transcription factor) expression, which leads to enhanced expression of MT1-MMP responsible for three-dimensional collagen invasion [19].
The glycoproteins expressed on the surface of tumor cells mediate cell adhesion and subsequent invasion. CD147, a highly glycosylated protein on the tumor cell surface, having Lewis antigen on its surface promotes CD147 mediated tumor cell adhesion and expression of matrix metalloproteinase-2 (MMP-2). CD147 and Lewis-Y antigen expression was associated with a higher adhesive ability to ECM proteins collagen and laminin, higher tumor grade, metastasis to lymph nodes, and decreased survival [21]. Another glycoprotein, Mucin 16 (MUC 16)/cancer antigen (CA-125), expressed on tumor cells binds mesothelin of mesothelial cells of the peritoneum, which is the initial event in tumor adhesion and spreading across the ECM. Matrix metalloproteinases (MT1-MMP, MMP-14) were seen catalyzing the degradation of MUC-16/CA-125, eventually leading to its decreased surface expression, and attachment to peritoneum was decreased [20].

The cadherin composition and expression of matrix metalloproteinases (MMP) govern collagen matrix invasion and peritoneal metastasis. Klymenko et al. highlighted neural cadherin's (Ncad) presence on MCA
to have significant depth penetration, rapid lateral migration through the mesothelium before disrupting the basement membrane to invade ECM compared to epithelial cadherin (Ecad) expressing cells. Inhibition of MMP showed decreased invasion of cancer cells [22]. Integrins play a role in the adhesion of tumor cells to ECM proteins. Oct4A is a transcription factor, suppression of which is associated with decreased expression of integrin β1, α5, and α2 subunits leading to diminished adhesion of tumor cells to collagen and fibronectin, decreased levels of MMP-2, and certain markers. Upregulation of Oct4A was linked to the formation of stable, highly compact MCA [9]. Another study showing the formation of more cohesive MCA was coupled with the expression of integrin-linked kinase (ILK), co-expressed with MT1-MMP. ILK catalyzes the phosphorylation of cytoplasmic tails of MMP, which leads to the formation of stable MCA, which causes metastasis. ILK also regulates Integrin-mediated cell adhesion causing mesothelial cells to retract, uncovering the ECM [16]. Table 2 below highlights the objectives and potential therapeutic targets of the included studies [4,5,9,16,[19][20][21][22].

Authors
Year   [24]. Another study showed the role of tumor suppressor MicroRNAs in the invasion of ovarian cancer. MicroRNA overexpression leads to increased filamentous actin (F-actin) levels, which decreases the invasive potential of cancer cells through collagen matrices. Increased cell size and reduced deformability were seen with tumor suppressor MicroRNA transfection, making it useful as bigger cells have lesser invasive potential, and less deformable cells take longer to transit [25].
CTHRC1 is a cancer-related gene involved in certain signaling molecules' phosphorylation, enhancing the migration and invasion of ovarian cancer. A study by Ye et al. observed decreased phosphorylation of epidermal growth factor receptor (EGFR), extracellular signal-regulated kinase 1/2 (ERK1/2), protein kinase (AKT), and reduced migration and invasion of ovarian cancer cells on silencing the CTHRC1. They also observed that the EGFR inhibitors blocked the effect of CTHRC1. Hence it was suggested that CTHRC1, by activating EGFR signaling, promotes metastasis in ovarian cancer, which is mediated through ERK1/2, PI3K/AKT [26]. CTHRC1 was also found to be increasing the expression of Integrinβ3 and phosphorylating focal adhesion kinase (FAK), which led to invasion and migration of ovarian cancer cells, enhanced adhesion to ECM proteins, peritoneal and lymph node metastasis [12].
Cheon et al.'s gene signature analysis identified 10 collagen remodeling genes regulated by transforming growth factor-β (TGFβ1) signaling. COL11A1 was among the genes discovered, which showed significantly higher expression in metastatic tumors than primary tumors. These genes were found to promote metastasis and decreased survival [15]. Identification of downstream targets upon activation of NOTCH3 Intracellular domain (NOTCH3IC) showed induction of genes encoding ECM proteins and adhesion molecules. Collagen and Integrin genes were identified to promote cancer cell attachment to the peritoneum, causing ascites leading to poor survival [27]. An epithelial to mesenchymal transcription factor TWIST1 induces the expression of a mesenchymal gene discoidin domain receptor 2 (DDR2), which recognizes collagen as its ligand. DDR2 controls ECM remodeling enzymes' expression, which could lead to enhanced migration and invasion of tumor cells, cleavage of fibronectin, mesothelial cell clearance facilitating metastasis, and decreased survival [13]. Table 3 below highlights the objectives and potential therapeutic targets of the included studies [12,13,15,[23][24][25][26][27].

Year of publication Purpose of the study Potential therapeutic target
Li et al. [23] 2020 To investigate the role of anticancer gene miR-30b-3p on the biological activity of ovarian cancer cells and its association with the CTHRC1 gene MicroRNA miR-30 family Wahab et al. [24] 2020 To study the effect of differential expression of microRNAs and their target genes in ovarian cancer growth, migration, and invasion MicroRNA Price et al. [27] 2020 To assess the effect of NOTCH3 signaling on tumor cell adhesion to the peritoneum, tumor cell proliferation, and patient survival NOTCH3 Grither et al. [13] 2018 To study the TWIST1 induced expression of DDR2 leading to mesothelial cell clearance and tumor cell invasion DDR2 Guo et al. [12] 2017 To investigate the role of CTHRC1 in ovarian cancer cell migration, invasion, and adhesion to vitronectin, peritoneal metastasis, and metastasis to distant organs

Metastatic Cascade: Receptors, Enzymes, and Spheroids Explained
Interaction of receptors was shown to increase ovarian cancer dissemination. Urokinase plasminogen activator receptor (uPAR) interacts with formyl peptide receptor 1 (FPR1) and promotes tumor cell adhesion to mesothelial cells of peritoneum and vitronectin. Higher expression of uPAR and FPR1 was correlated with metastasis and poor clinical outcome [14].
In addition to receptors, enzymes also promote ovarian cancer proliferation and dissemination. Shen et al. showed how histone deacetylase-4 (HDCA-4) enhanced cancer cell proliferation. Higher expression of HDCA-4 corresponds to a higher stage of cancer. HDCA-4 gets co-localized in the nucleus and PP1 in response to collagen matrices, which leads to altered transcription and translation of p21, promoting proliferation and migration of cancer cells [17]. Another study by Dai et al. showed mitogen-activated protein kinase 7 (MAPK7) associated with ovarian cancer cell proliferation, migration, and invasion. Type II collagen expression also increased with MAPK7 overexpression [8]. Flate and Stalvey showed that tumor cells interact with ECM proteins like collagen and fibronectin, which affects cell migration and invasion. They also highlighted p21 activated kinase (PAK) mediating the process of migration of cancer cells [28].
Cancer cells form spheroids by making cell-to-cell contact, affecting ovarian cancer cell behavior in the tumor microenvironment. Fogg et al., in their study, identified alternatively activated macrophages (AAM) in the ascitic fluid to be secreting certain soluble factors, e.g., Fms related tyrosine kinase 3 (FLT3), interleukin-2 (IL-2), interleukin-8 (IL-8), leptin, heparin binding-epidermal growth factor-like factor (HB-EGF). These factors were acting through a common pathway of JAK2/STAT3 activation, inducing MMP-9. This pathway leads to disaggregation of spheroids making single cells spread across the ECM [7]. Klymenko et al. showed increased lateral dispersion of spheroids (aka MCA) in response to ascites-induced compression. They also studied the effect of compression on MCA gene expression [29]. The expression of cadherins also influences metastatic success. Ncad+ cells form highly compact stable MCA, whereas Ecad+ cells tend to form loosely adherent cell clusters. Acquisition of Ncad by Ecad+ cells, making a hybrid increased migration, adhesion to mesothelial cells, invasion of ECM developing into secondary metastatic lesions [30]. Another study delineating the role of Cadherins showed high levels of E-cadherin in cancerous cysts. They also showed that E-cadherin repression led to cyst disruption and inhibited collective cancer cell migration. Hence, it was suggested that E-cadherin is important for cancer cells to migrate collectively in collagen matrices [31]. A study by Pettee et al. discovered a Rho GTPase effector mDia formin, which controls F-actin assembly required for cell-cell junction and tight spheroid formation. Inhibiting mDia led to the transition of cells acquiring an amoeboid configuration and invasive single cell dissemination. Rhoassociated protein kinase (ROCK) is another Rho effector. Inhibiting both mDia and ROCK completely blocked invasion, which suggested that single-cell dissemination on inhibiting mDia is ROCK dependent [32]. Vallen et al. showed the overexpression of a 4,6 sulfated glycosaminoglycan (chondroitin sulfate E aka CSE) enhanced the adhesiveness between cells, which is seen in the spheroid formation and cancer metastasis [33]. Table 4 below highlights the objectives and potential therapeutic targets of the included studies [7,8,14,17,[28][29][30][31][32][33].

Combating the Chemotherapeutic Resistance
The amount of collagen in the tumor microenvironment regulates the response to chemotherapeutic drugs. Yeung et al. showed that cancer-associated fibroblasts (CAFs) secretes microfibril associated protein 5 (MFAP5), which causes increased expression of fibrosis-related genes thus, increasing the amount of collagen, which together with decreasing microvessel stability (more leakiness of vessels) leads to hindrance in the delivery of drugs [10]. Another study showed improved response to chemotherapy in patients who were given Losartan along with the standard therapy. Losartan decreases the number of fibroblasts, upregulates antifibrotic miRNA decreasing the ECM content (collagen type I levels), and thus reduce the solid stress, which improves drug delivery [6]. In another study, doxorubicin diffusion was impaired due to increased collagen accumulation in the tumor microenvironment due to decreased susceptibility to proteolytic enzymes caused by a serine protease inhibitor Maspin. Tumor growth was significantly inhibited when the anti-Maspin antibody was given along with doxorubicin compared to doxorubicin alone [35].
Resistance to cisplatin was found to be governed by a tumor suppressor microRNA. In 2017, Deng et al. showed decreased expression of miR-199a-3p led to enhanced Dicoidin domain receptor-1 (DDR1) expression, which was responsible for resistance to cisplatin [34]. Dexamethasone (DEX) enhances fibronectin (FN) expression and mucin 16, responsible for its pro adhesive, pro-survival effects and protects cancer cells from chemotherapeutic agents. Inhibition of fibronectin with FN-siRNA and mucin 16 with MUC-siRNA attenuated the effects [18]. Ovarian cancer cells up-regulate the expression of a Microtubuleassociated protein, which causes resistance to paclitaxel. Tau protein binds with microtubules and competes with paclitaxel for the binding site. Thus, enhanced Tau protein expression in ovarian cancer leads to resistance [11]. Table 5 below highlights the objectives and potential therapeutic targets of the included studies [6,10,11,18,34,35].

Authors
Year of

Strengths and Limitations
To the best of our knowledge, this is to date the only traditional review gathering all the information regarding molecular mechanisms underlying the ovarian cancer metastasis addressing the interaction between tumor cells, peritoneal mesothelial cells, and the extracellular matrix discovered in the last seven years published in the English language in PubMed. The details of how tumor cells spread are described in these studies; however, what happens to the tumor cell structure, any configurational change in the enzymes, receptors, and other signaling molecules involved is not explained. The structural component of metastasis can be a topic of research in the future. Tumor cell interaction with peritoneum is well described in this review, but how do tumor cells initially travel from ovary to peritoneum have not been detailed in the studies and can be researched upon. The studies included in this review are done on human tissue samples and cell lines. Animal models were also used. This warrants the need to corroborate the findings in randomized control trials.

Conclusions
The signaling molecules and pathways encountered in disseminating ovarian cancer and drug response lie in the peritoneum and extracellular matrix. Tumor cells detach from their primitive location on the ovaries and travel through the matrix to reach metastatic sites such as the peritoneum. They interact with peritoneal mesothelial cells and establish new sites for tumor growth and proliferation. Based on the reviewed articles, we found a substantial relationship between hypoxia, matrix remodeling enzymes, and disease progression. Several genes, their mRNAs, and subsequent proteins have been identified to alter patients' overall survival. There is a cascade of enzymes, receptors working in harmony to promote metastasis. The emerging chemotherapeutic resistance imposes great difficulties in treating patients, and targeting these molecular mechanisms has shown an improved response to the therapy. There is no good screening method available for early diagnosis before the patient develops metastasis. Studies focusing on finding a suitable marker for screening should be done. If studied in detail with more experimental studies and possibly targeted, the mechanisms highlighted in this review may enhance the quality of care and reduce the disease burden.

Conflicts of interest:
In compliance with the ICMJE uniform disclosure form, all authors declare the following: Payment/services info: All authors have declared that no financial support was received from any organization for the submitted work. Financial relationships: All authors have declared that they have no financial relationships at present or within the previous three years with any organizations that might have an interest in the submitted work. Other relationships: All authors have declared that there are no other relationships or activities that could appear to have influenced the submitted work.