Ovarian cancer (OC) has been reported to be the third most common gynecological malignancy worldwide, and the most lethal type of cancer (1). A total of 313,959 newly diagnosed cases and 207,252 associated deaths were reported in 2020 (2). Since OC cells (OCCs) often manifest the disease silently, >75% of cases are diagnosed at the late stage, usually after the tumor has spread throughout the abdomen (3). Currently, the standard treatment for OC comprises maximal cytoreductive surgery followed by platinum-based chemotherapy (4). Although the majority of patients go into clinical remission after conventional chemotherapy, the recurrence rate can be as high as 85% (5). In addition, the overall 5-year survival rate of OC is <50% in numerous countries throughout the world (6).

Almost 90% of ovarian tumors are of the epithelial OC (EOC) type, which is classified into five histological subtypes: Serous tumors (comprising ~80% of EOC), mucinous tumors, endometrioid cancer, ovarian clear cell carcinoma and mixed tumors (7). However, recurrent cases are often chemoresistant, and therefore, these are associated with a high mortality rate (7). Chemoresistance may be driven by three main factors: Pharmacokinetic factors, the tumor microenvironment (TME) and tumor-specific mechanisms (8). Maintenance therapy with poly(ADP-ribose) polymerase inhibitors, bevacizumab and/or drugs targeting homologous recombination deficiency is becoming more widely used in the treatment of OC (9). Nevertheless, a plethora of studies have focused on the intrinsic characteristics of OCCs, while neglecting the role of the TME (10–13).

The TME consists of the blood and lymphoid vessels, nerves, fibroblasts, extracellular matrix (ECM) proteins, endothelial cells, pericytes and immune cells (14). Essentially, communication between OCCs and various components of the TME has a major impact on chemoresistance (15). It is important to understand how OCCs interact with the surrounding matrix to improve our understanding of tumor cell biology, both during oncogenesis and in terms of how chemoresistance develops. The present review offers a summary of the four most vital aspects: Cancer-associated fibroblasts (CAFs), ATP binding cassette (ABC) transporters, extracellular vesicles (EVs) and immune cells. Considered in their entirety, recovery of chemotherapeutic sensitivity and identification of novel anticancer drug targets are of great significance with respect to the treatment of OC.

CAFs, a well-recognized abundant stromal cell population in the TME, steadily nourish the tumor cells by secreting soluble factors (16). The soluble factors derived from CAFs undoubtedly provide an important step in the development of CAF-mediated chemoresistance. Fibroblast growth factor 4 (FGF4) and fibroblast-derived insulin-like growth factor II (IGF2) are respectively able to activate the FGF4-FGF4 receptor 2 and IGF2-IGF1 receptor signaling pathways to induce the OC stem cell (OCSC) niche in CAFs (17,18). OCSCs contain all the particular functionalities of the cell subclasses, such as the ability to self-renew and to differentiate (19). Chemotherapeutic agents usually target fast-dividing cells and act in a cell-cycle specific manner, which confers an advantage on the ability of OCSCs to survive due to their slow proliferation rate (20). OCSCs may stay dormant for long periods of time, but they can self-renew at low seeding concentrations and produce more aggressive metastatic progeny (21). CAFs secrete matrix metalloproteinases (MMPs) to degrade matrix collagens, fibronectins and proteoglycans, facilitating TME structural remodeling and promoting matrix contractility (22,23). Unlike a soft TME, such as the greater omentum, which promotes dispersion of the OCCs (24,25), the increased stiffness of the ECM triggers OCC survival and proliferation (26). In addition, increased mechanical stress may lead to the collapse of blood vessels, leading to hypoxia, thereby promoting more aggressive cancer phenotypes and reducing drug delivery (27). Furthermore, the release of glutathione and cysteine by the CAFs contributes towards the depletion of platinum in the nuclei of the adjacent OCCs, thereby imparting resistance to platinum-based chemotherapies (28,29).

In addition to the resident fibroblasts, CAFs may be derived from five alternative sources: Epithelial cells, endothelial cells, mesothelial cells, bone marrow mesenchymal cells and adipose-derived mesenchymal stem cells (30–32). The levels of surface markers, such as α-smooth muscle actin, fibroblast-specific protein 1 and fibroblast activation protein, differ in different CAFs populations (16,33,34). In breast and lung cancer, CD10 and G protein-coupled receptor 77 have been demonstrated to unequivocally define a subset of CAFs that are associated with chemoresistance due to their ABC transporters (35).

However, relevant therapies in OC have been greatly hindered due to a high level of functional heterogeneity and a lack of a specific subset of markers (36,37). One of the most well-characterized examples is provided by anti-stromal therapy, in which it has proven difficult to precisely target CAFs, thereby increasing the risk of ablating vital stromal components required for tissue homeostasis (38). Therefore, there is an urgent need to classify different CAF phenotypes for improved stratification. With the emergence of single-cell technologies, an increasing array of functional assays has become available, and studies on CAFs are entering a critical stage (39,40). Strategies to ‘normalize’ CAFs (41) or to deprive them of their soluble factors (28,42) may offer feasible methods to complement the existing therapies that target OCCs.

It is well established that the human family of ABC transporters comprises 49 members, which are grouped into 7 distinct subfamilies, termed ABCA through to ABCG (43). In addition to enabling the unidirectional translocation of substrates such as saccharides, lipids, amino acids and proteins, >13 types of ABC transporters are able to permit the extrusion of cytotoxic molecules from cancer cells and reduce the intracellular drug concentration, thereby promoting cell survival and multidrug resistance (MDR) (43,44).

Intrinsically chemoresistant types of cancer (e.g., pancreatic, liver, colon, adrenocortical and kidney cancer) express P-glycoprotein (P-gp; ABCB1) at a high level (44), whereas OC hardly expresses any P-gp at the time of initial presentation (45). The process of acquired chemotherapeutic resistance in OC is often accompanied by a marked overexpression of P-gp, indicating a possible role for P-gp in acquired resistance (45,46). Notably, ascites-induced OC chemoresistance may be mediated by ABC transporters. A previously published study showed that specific MDR-associated protein-1 (MRP1; ABCC1) inhibitors could suppress the ascites-induced resistance to paclitaxel (PTX) in ID8 cells (i.e., a mouse EOC cell line) (47). The expression levels of MRP1 and P-gp were found to be closely associated with the clinical stage and pathological differentiation grade of OC (48,49). Considered together, numerous findings have revealed that ABC transporters are important in facilitating OC drug resistance.

Although a logical approach to overcome MDR would be to inhibit ABC transporters, associated clinical trials that have been conducted have produced disappointing results (50,51). High doses of first-generation P-gp inhibitors (e.g., verapamil) were found to be required to be effective against MDR, resulting in increased levels of toxicity (52). Second-generation inhibitors (e.g., valspodar) have proven to be effective in overcoming the obstacle of high doses, although they still have poor efficacy due to pharmacokinetics (51). To date, no specific, safe and effective third-generation inhibitors have been approved (53). A multiplicity of ABC transporters may be able to contribute to the acquired MDR of these tumors, providing a plausible explanation to explain how inhibiting only one of these ABC transporters is unlikely to reverse chemoresistance (50,54). Furthermore, the majority of clinical trials that have been performed have been small-scale, randomized and single-institution studies (51,55–57). Due to insufficient inclusion criteria, non-specific patients and inconsistent detection criteria, it has proven to be difficult to differentiate valid from invalid data. In addition to these issues, it may not be possible to regard mass-published cell culture model studies (58) and phase I clinical trials (59–61) with too much optimism, since unexpected results are likely to occur in phase II and III clinical trials.

Certainly, novel approaches, such as photodynamic therapy based on mitochondrial oxidative stress (45) and time-of-flight cytometry for the direct quantitation of platinum (62), have aroused great interest. Further developments in positron emission tomography, fluorescence in situ hybridization analysis, RNA sequencing and next-generation sequencing will assist in enabling the selection of a subset of patients for the development of specific ABC transporter inhibitors (63).

EVs, which are classified into exosomes, microvesicles (MVs) and apoptotic bodies, are able to transfer nucleic acids and proteins from donor cells to recipient cells (64,65). MicroRNA (miR/miRNA) fulfills an important role in inducing chemoresistance by targeting various signaling pathways as a major exosomal cargo molecule (66). A particular miRNA that has been widely reported to promote OC chemoresistance is miR-21 (67). Exo-miR-21 released by CAFs induces PTX-resistance in neighboring SKOV3 cells by downregulating apoptotic protease-activating factor-1 (APAF1) (67). APAF1 is able to bind to cytochrome c and dATP, which in turn recruit and activate caspases-9 and −3, as well as the apoptotic pathway (68,69). Additionally, exo-miR-98-5p derived from CAFs enhances cisplatin-resistance in OCCs through the downregulation of cyclin-dependent kinase inhibitor 1A, which serves an important role in cell cycle arrest (70).

Exosomal transmission of proteins also has a crucial role in modulating drug resistance in OC (71). Epithelial-mesenchymal transition (EMT) inducers, such as MMPs, annexin A2 and integrin 3, have been found in tumor-derived exosomes, suggesting that exosomes may promote the EMT process in which epithelial cell characteristics are lost and mesenchymal phenotypes are acquired (72,73). A number of different EMT-driven mechanisms that lead to carboplatin and/or PTX resistance have been identified in OC, including β-tubulin variants (taxane-specific resistance), ABC transporter overexpression, changes in the cell cycle, a greater DNA repair capability, anti-apoptotic effects and changes in stress chaperones (74).

Exosomes have been studied extensively in terms of OC chemoresistance, whereas apoptotic bodies and MVs have not been. Previously published studies showed that A2780/PTX-derived MVs could transport bioactive P-gp to chemosensitive A2780 cells in vitro, which conferred PTX-resistance to the recipient A2780 cells (69,75,76). The same phenomenon had been demonstrated in breast cancer (77); however, much work needs to be completed to improve our understanding of the role of MVs in OC chemoresistance.

There are four widely accepted potential strategies to overcome the pro-tumorigenic effects of exosomes (69): i) The inhibition of exosomal secretion; ii) the inhibition of the uptake of exosomes by target cells; iii) the promotion of exosomal depletion; and iv) the targeting of exosomal cargo. However, all these strategies remain at the preliminary and experimental stages (69). Notably, exosomes and MVs present an appealing platform for delivering drugs, as they are non-toxic and have low immunogenicity (69). In particular, they are able to transport drugs in a specific and targeted manner (78). Bioengineered exosomes are currently in use for the treatment of several different cancer types, including lung, prostate and pancreatic cancer (79–81). By contrast, the progress made using bioengineered exosomes in OC has been limited. Mesenchymal stem cells with a high proliferative capability have been used to produce large quantities of exosomes for therapeutic purposes (82). However, other challenges, such as how to isolate pure exosomes, how to obtain better loading efficiency and how to accurately deliver the targeted drugs, have to be overcome before the use of exosomes in cancer therapy may be successfully implemented (83).

Even though limited numbers of immune cells are able to infiltrate in OC, they exert direct or indirect effects on OC chemoresistance (84). Tumor-associated macrophages (TAMs) are the major population of immune cells that exist in the TME of OC (85), comprising two distinct subsets: Anti-tumorigenic M1-like TAMS and pro-tumorigenic M2-like TAMs (86). A previous study showed that exo-miR223 derived from the M2-like TAMs was effectively internalized by OC cell lines (A2780 and SKOV3 cells), thereby creating a chemoresistant phenotype through activation of the PI3K/AKT signaling pathway (87). In addition to secreting miR-loaded exosomes, in another study, M2-like TAMs were revealed to induce higher expression levels of ABC transporters in A2780 cells (85). Furthermore, TAMs have been shown to occur in close proximity to CAF-populated areas, indicating that a close association may exist between these two cell types (88). A number of previously published studies have established that CAFs are able to actively increase monocyte recruitment and promote their differentiation into M2-like TAMs by secreting multiple soluble factors, including interleukin-6, −8 and −10, and transforming growth factor-β (88,89). More importantly, CAF-induced M2-like TAMs exhibit higher expression levels of programmed cell death protein-1, thereby impairing effector T cell responses and inducing immune suppression of TAMs (90). Reciprocally, M2-like TAMs have been shown to regulate CAF activation as well (86), consequently establishing a positive feedback loop.

Studies that have focused on the influence of other immune cells on OC chemoresistance have been scarce up to the present time. Nevertheless, it should be mentioned that OC-derived EVs have an impact on the adaptive immune escape process (91). For example, EVs stimulate T cell and NK cell proliferation, as well as inhibiting their functional activation (92,93). FAS ligand and TNF-related apoptosis-inducing ligand expressed by OC-derived EVs were shown to inhibit dendritic cell (DC) activation by inducing apoptosis (94). In brief, EVs assist OCCs in acquiring chemoresistance through immune suppression and immune evasion.

Strategies to block macrophage recruitment have been successfully developed (95). It is well established that colony stimulating factor-1 (CSF-1) and chemokine C-C motif ligand 2 are macrophage chemoattractants (96). Anti-CSF-1 receptor agents have been shown to prevent the recruitment of M2-like TAMs to tumor areas in pancreatic ductal adenocarcinoma (PDAC) models (97). However, CSF-1 receptor is not exclusively expressed by M2-like TAMs (98). Other immune cells, including M1-like TAMs and DCs, would be affected too, leading to complex interactions (98). By contrast, repolarizing M2-like TAMs back into the M1-like phenotype appears to be the more attractive option. In the PDAC model, the combination of anti-CD40 antibody and gemcitabine has been demonstrated to repolarize M2-like TAMs back into the M1-like phenotype, leading to increased sensitivity to gemcitabine and a reduced tumor burden (95). However, further clinical trials are required in a range of solid tumors. Additionally, the mechanism through which TAMs interact with CAFs has not been fully investigated to date (86). Future studies are required to delineate the precise mechanisms underlying CAF-TAM interactions in the TME in order to make further advances on the current cancer-targeted therapies.

Low survival rates in patients with OC are considered to mainly result from a late diagnosis, disease recurrence and chemoresistance. Specifically, chemoresistance is emerging as a major hurdle in OC treatment. Rather than focusing on the isolated impact of OCCs, the present review attempted to encompass the dynamic interplay between the TME and OCCs.

The soluble factors derived from CAFs not only induce formation of the OCSC niche, but also increase the stiffness of the ECM, which promotes the development of more aggressive and drug-resistant cancer phenotypes (17,18,22,23). ABC transporters are responsible for the extrusion of cytotoxic molecules from the OCCs and for reducing the intracellular drug concentration, eventually promoting cell survival and MDR (43,44). Since exosomes are used as genetic exchange vectors in the TME, exosomal cargoes activate signaling pathways in recipient cells, thereby facilitating cell proliferation and the EMT process, and inhibiting apoptosis (67,73). In addition, OCCs acquire chemoresistance through immune suppression and immune evasion (90).

Extensive crosstalk occurs among these components in the TME. Soluble factors secreted by CAFs and P-gp proteins can be released in the form of exosomes (70,75). TAMs may secrete EVs and express ABC transporters as well (85,87). TAMs are found close to the CAF-populated areas, and they engage in complex bidirectional interactions with CAFs (88).

The current review briefly presents the most up-to-date roles of the TME in OC chemoresistance and summarizes current research gaps in TME-targeted therapy. Although the role of the TME in fostering OC chemoresistance is becoming more recognized, research into this topic is just beginning and further work is required to advance current TME-targeted OC therapies.