Role of Cancer Side Population Stem Cells in Ovarian Cancer Angiogenesis

Tumor angiogenesis aids tumor growth and development, as well as metastatic spread. However, due to the rapid growth of the tumor and changes within the tumor environment, the resultant vessels are often immature and not fully functional. Several therapeutics have been developed to try and inhibit tumor angiogenesis, but they appear most effective in combination with more traditional chemotherapeutics. To develop better chemotherapeutics aimed at tumor angiogenesis, we must first fully understand this process and the cell types and molecules involved in facilitating it. This review will discuss the contribution of cancer stem cells (CSCs), in particular the side population (SP) stem cells, to angiogenesis in ovarian cancer with the view to identifying gaps in knowledge and potential therapeutic targets for further research.

Ovarian cancer (OC) is one of the most common gynecological malignancies. OC can be divided into three types: epithelial tumor, germ cell tumor and specialized stromal cell tumor [1]. Epithelial OC, the most malignant ovarian tumors (∼90%), is heterogeneous and can be divided into four main subtypes: serous, mucinous, endometrioid, and clear cell [2]. The different histotypes of OC can be stratified into two main classes: type I and type II. Type I includes low-grade and borderline serous, low-grade endometrioid, clear cell, mucinous, and transitional carcinomas, which are generally confined to the ovary and are characterized by microsatellite and genomic instability. Type II includes high-grade endometrioid ovarian carcinoma, high-grade serous carcinoma, undifferentiated carcinoma, and carcinosarcoma, most at an advanced stage and with highly aggressive behavior, the most frequent form of which is high-grade serous carcinoma [3, 4]. The 5-year survival rate for OC is about 47%. Patients with OC often have no specific clinical symptoms in the early stages of disease, and it is often not diagnosed until it is already at an advanced stage [5]. Transvaginal ultrasound and serum CA-125 levels are the most common methods for the diagnosis of OC. Although CA-125 has made a great contribution to diagnosis and prognosis of OC, the lack of biomarkers that are highly sensitive to early diagnosis increases the mortality rate of this disease. Recently, HE4 and interleukin-6 have been identified as potential biomarkers for the early diagnosis of OC [6]. In addition, genetic testing for BRCA and HRR deficiency may aide treatment decisions for women with newly diagnosed advanced disease [7]. Despite these recent advances, early diagnosis of OC is still a challenge and needs to be improved through OC awareness campaigns and biomarker discovery research [8].

Many patients with OC find relief after initial treatment, but most patients relapse, and over time may well go on to develop multidrug-resistant (MDR) disease, leading to remission and eventually to death [9]. MDR in cancers is often associated with the expression of members of the ABC transporter family that are involved in efflux of a range of chemotherapeutics. These transporters are often expressed at elevated levels by stem cells conferring on these cells a means of protection against a range of structurally unrelated xenobiotics that could cause cellular damage [10]. Advanced OC is a highly metastatic disease, with secondary sites generally being located within the abdomen [11]. Although most advanced patients respond well to surgical treatment and chemotherapy, high recurrence rate and chemotherapy resistance are still the main factors affecting the low survival rate of OC patients [12]. Progress in tumor stem cell research has helped deepen our understanding of OC metastasis and chemotherapy resistance and may lead to the development of better treatment plans for patients with OC [13, 14].

Currently, OC is routinely treated with surgery and chemotherapy based on paclitaxel and platinum. However, most OC becomes less sensitive to the same chemotherapy drug after repeat treatments and eventually becomes chemoresistant [15]. It is now generally accepted that BRCA mutation status, histological subtypes, and previous use of bevacizumab or other therapies need to be considered when choosing further treatment for advanced OC. For this reason, patients with advanced OC are classified into two groups: those who can receive new platinum treatments and those who cannot [16]. Patients who can continue to receive platinum therapy could be treated with a different strategy combined with biological drugs that target the genes related to angiogenesis or DNA repair. Patients who have latent platinum responses and are unable to take secondary platinum treatment could receive the most effective therapy, a combination of aspergillin and pegylated liposomal doxorubicin. To date, poly(adenosine diphosphate-ribose)polymerase inhibitors and antiangiogenic agents have been shown to be effective as maintenance and concurrent therapy after chemotherapy, respectively [17]. However, there are small populations of tumor cells, including CSCs that may be resistant to these therapies. Breast cancer CSCs isolated from triple-negative breast cancers are relatively resistant to poly(adenosine diphosphate-ribose)polymerase inhibitors [18]. There is some evidence that CSCs may be susceptible to antiangiogenic therapies [19]. Proangiogenic signaling by tumor-associated mesenchymal stem cells is impacted by tetrac, a promising antiangiogenic therapeutic [19]. However, more research is required on the impact of different therapies on CSCs compared to “normal” tumor cells, where it is likely that combination therapies targeting different cell types will be required. In addition, the mechanism by which CSCs are more susceptible to different therapies also requires further investigation. Several recent review articles have addressed the potential therapeutic targets available in CSCs in general [20, 21] and ovarian cancer stem cells (OCSCs) in particular [22, 23].

Ovarian germ cell and epithelial ovarian tumors are different in biology and genetics, requiring different treatment approaches [24]. Like patients with testicular germ cell tumors, patients with platinum resistance display alterations in TP53‐MDM2, Wnt/β-catenin, PI3K, and MAPK signaling pathways that may be targeted as potential treatment options. It still needs to be determined whether ovarian germ cell tumors would also benefit from this approach. In germ cell tumors, BRAF mutations and lack of regulation of the p53 signaling pathway are associated with cisplatin resistance. Blocking programmed death ligand 1 (PD-1/PD-L1) may be an effective treatment for germ cell tumors [25]. In addition, immunotherapy may be effective in patients with late platinum resistance of advanced germ cell tumors, but further studies are required to determine the optimal immunotherapeutic targets [26, 27]. To date, there has been limited clinical application of immunotherapy for OC, and most of the published studies have used PD-1 or PD-L1 inhibitors in early-stage OC patients [28‒30]. In addition, women who have severe menopausal symptoms following OC treatment can safely take hormone-replacement therapy, which may confer benefits in terms of OC as well as improving quality of life for these women [31].

In antitumor therapy, the success of antiangiogenic therapy remains limited [32]. Anti-VEGF therapy with bevacizumab reduced the number of blood vessels and blood supply within the glioblastoma xenograft but increased the invasive ability of the tumor cells [33]. Therefore, antiangiogenic therapy alone is not sufficient to improve patient survival. In fact, in patients with breast cancer, colon cancer, small-cell lung cancer, and other tumors, the combination of immunotherapy and antiangiogenesis agents often shows better results than either strategy alone [34‒37]. Although this approach has not been well studied in OC animal models, in other tumor animal models, animals treated with the combination therapy showed greater tumor regression, less ascites development, and lower metastasis rates [37]. It may therefore be assumed that targeting of OC angiogenesis in combination with other chemotherapeutics may be a viable strategy for treatment. However, we need a greater understanding of the drivers of tumor angiogenesis and the different cell types involved in this process.

CSCs are a small population of cells with the properties of stem cells such as self-renewal and multidirectional differentiation in tumor tissues that differentiate into multiple different cell types including tumor cells, endothelial cells, and vascular smooth muscle-like cells (Fig. 1) [38]. CSCs can divide asymmetrically, produce heterogeneous tumor cells, and regulate the occurrence and development of tumors [39]. CSC theory suggests that a few drug-resistant tumor cells are closely related to the occurrence, spread, metastasis, and recurrence of tumors [40]. The targets of chemotherapy and radiotherapy are cells that are actively proliferating, but CSCs can remain dormant and then start proliferating again; therefore, many CSCs are resistant to chemotherapy, which plays an important role in drug resistance and high tumor recurrence rates [41, 42]. OCSCs were first isolated from the ascites of OC [43]. In vitro and in vivo studies provided evidence of OCSCs with self-renewal and differentiation capabilities [44]. In addition, different populations of OCSCs have been identified, including the CD44+, CD24+, CD117+, CD133+, or ALDH1+ stem cells and SP stem cells [14]. OCSCs are associated with the occurrence of OC as well as the development of drug resistance [43]. Therefore, drugs targeting OCSCs could improve responses to chemotherapy and prevent disease recurrence.

The cell cycle of CSC is in a static state with inactive DNA replication; therefore, chemotherapy directed at rapidly dividing cells is not effective against CSCs. In addition, CSCs have a high level of expression of MDR proteins, including the ATP-binding cassette (ABC) transporter family [45]. ABC transporters, as a large family of transmembrane proteins, have 49 members and have been classified into seven subfamilies, termed ABCA-G. ABC transporters can exude various substrates from cells, including proteins, amino acids, and drugs [46]. Some members are termed MDR proteins or MDRs as they can expel cytotoxic chemotherapeutics out of cancer cells. Increasing evidence demonstrated that some MDRs are more highly expressed in CSCs than in normal cells or cancer cells; for example, ABCG2 overexpression is found in breast CSCs and ABCB1 overexpression in OCSCs [47, 48]. Therefore, understanding the role of ABC transporters and their mechanisms in CSCs will help us assess their significance in cancer progression and develop a strategy for targeting more aggressive and resistant cancer cells.

SP cells are a subset of stem cells, characterized by a high expression of some ABC transporters and the ability to efflux lipotropic dyes (such as Hoechst 33342) out of cells [49]. SP cells are usually sorted and detected by flow cytometry after Hoechst 33342 staining, and this method has been used to screen SP cells from a variety of cells or tissues. SP cells are not only isolated and identified in bone marrow, liver, kidney, skeletal muscle, brain, heart, and lung but also isolated from hematological tumors, liver cancer, breast cancer, OC, gastrointestinal tumors, and other tumor types [50‒59]. SP cells have strong stem cell properties, including self-renewal and multidifferentiation potential like stem cells, but they also possess unique phenotypic markers and biological characteristics [60]. SP have a stronger capacity than non-SP for regeneration and clone formation. Szotek et al. [57] used the Hoechst 33342 method to isolate SP cells from OC. Compared with NSP cells, the same number of SP cells had stronger tumorigenicity in nude mice. These results suggest that SP cells play an important role in tumor development.

Cancer SP cells have the stem cell-like characteristics of clonal ability, asymmetric division, and high tumorigenicity in vivo [60]. Several studies have shown that SP cells have stronger drug resistance than non-SP cells, and that the drug resistance of SP cells is closely related to the expression of ABC transporter proteins, which can actively pump chemotherapeutic drugs out of the cell and reduce the concentration of drugs in the cell and the cytotoxic effect, thus conferring resistance to these antitumor drugs. Britton et al. [61] showed that MCF7 (a breast cancer cell line) SP displays high expression of ABCG2 and increased resistance to mitoxantrone compared to non-SP cells. Hu et al. [62] demonstrated that OC SP cells are tumorigenic and chemoresistant and that ABCG2 plays an important role in this chemoresistance. Moreover, downregulating the expression of ABC members or inhibiting their functions could reduce the percentage of SP cells and reverse drug resistance [63]. A relationship between SP and ABC transporters in OC has also been shown [63]. SP cells could be detected in both OC cell lines and ascitic fluid samples, which possess stem cell and drug resistance properties [63]. ABCB1 is the major functioning ABC transporter in OC cell lines HeyA8MDR and IGROV1, and its silencing could inhibit the SP phenotype and increase the sensitivity of OC cells to paclitaxel [63]. These results suggest that ABC transporters could serve as a good therapeutic target in OC. Current approaches targeting a single ABC transporter have not been successful due to the redundancy that exists in their functions and therapies targeting multiple transporters need to be developed and further research is required to define the role of the different ABC transporters in different types of OC cells. In addition, SP cells can be isolated from ovarian ascitic fluid from some but not all patients, but it should be noted that ABC transporter expression can be detected in ascitic fluid even when SP cannot be identified [63]. This could indicate limitations of detection of SP based on the Hoechst SP assay or the presence on non-SP cells that can also express ABC transporters and could include other stem cell types such as immature MSCs that have also been reported to express ABC transporters [64].

Under physiological conditions, angiogenesis rarely occurs in adults [65]. However, due to periodic changes in the female reproductive system and pregnancy, a dynamic process of tissue remodeling, vascular germination, and subsequent vascular regression can be observed in the ovary [66]. The physiology of the normal ovary changes conditionally after sexual maturity in female mammals, a process accompanied by periodic angiogenesis and degeneration. Changes in the blood vessels of the normal ovary involve changes in the vasculature of the follicles or the corpus luteum and are required for normal ovarian function [67].

Angiogenesis is the growth of new blood vessels from existing ones and has an essential role not just in development, reproduction, and tissue repair but also in tumor formation. Tumor growth requires angiogenesis to provide oxygen and nutrients, as well as a route for metastatic spread, for the tumor. Tumor angiogenesis can be induced by hypoxia, signals in the tumor microenvironment, or secretion of proangiogenic growth factors by tumor or mesenchymal cells (Fig. 2) [68]. Oxygen and nutrient depth of penetration into the surrounding tissue of an existing capillary is 100–500 microns; therefore, the rapidly growing tumor becomes hypoxic and triggers a proangiogenic response. However, the resultant vessels are often immature, with chaotic structure, limited mural coverage, and increased permeability. This can then lead to vascular collapse, greater egress for metastatic spread, and limited penetration of chemotherapeutic agents. Therefore, angiogenesis is an important step for tumor growth, maintenance, and metastasis, and angiogenesis inhibition was seen as an attractive strategy for the treatment of solid tumors [69‒71]. However, clinical trials have shown that this approach was not as effective as initially predicted. As a consequence, the idea of establishing vascular normalization as an adjuvant therapy has gained traction [72, 73]; the theory is that this process would lead to a more mature stable vasculature capable of delivering other chemotherapeutic agents.

In OC, as in other solid tumors, angiogenesis is a key process. Increased tumor angiogenesis is related to the clinical status of the tumor and associated with promoting epithelial OC to be more aggressive [74]. Physiological angiogenesis is a gradual process, in order to ensure that the process develops in the correct order, and the mutual inhibition or promotion of many angiogenic growth factors plays an important role in this process. Vascular endothelial growth factor (VEGF)-A and angiopoietins (Ang) are expressed in a variety of tissues and cells and are important for new blood vessel growth and maturation [75]. There is evidence that VEGF-R2 (also called fetal liver kinase-1, Flk-1) mediates the proangiogenic activity of VEGF-A, and the combination of VEGF-A and VEGF-R2 can induce the proliferation, migration, and germination of endothelial cells, as well as promote the formation of the vessel lumen [76, 77]. VEGF-A has been shown to mediate the differentiation of breast CSCs into endothelial cells, and we may postulate that the same situation exists in OC, although this hypothesis remains to be investigated [78]. On the other hand, Ang2 can block binding of Ang1 to the receptor-tyrosine-kinase (Tie2), resulting in endothelial cell apoptosis, vascular degradation, and vascular structure loss. This Ang2-dependent vascular wall disintegration makes it easier for VEGF-A to bind to endothelial cells and the receptor VEGF-R2 [79].

In patients with advanced OC, production of ascitic fluid is common [80]. The concentration of proangiogenic factors in the ascites is a marker of tumor invasiveness and correlates with poor prognosis for these OC patients [81, 82]. VEGF, transforming growth factor β, and epidermal growth factor are all proangiogenic factors. VEGF-A, as an abundant proangiogenic factor, is detected in these ascites [82, 83]. VEGF-A expression in tumor tissues is higher than that in normal tissue and participates in the chemotherapy resistance of OC cells through autophagy, and VEGF-A expression levels are positively correlated with chemoresistance and are also associated with tumor progression and poor prognosis [84, 85]. Blocking VEGF-A can temporarily normalize the tumor’s convoluted vascular system [73].

Angiogenesis is one of the most important steps in tumorigenesis and development. SP cells have the characteristics of stem cells, so do SP cells play an important role in tumor angiogenesis? Iohara et al. [86] isolated CD31-/CD146-SP cells from the dental pulp. These cells were CD34, VEGFR2, and Flk1 positive, suggesting that they were endothelial progenitor cells. They had high proliferative and migratory ability and also had multidirectional differentiation potential, including angiogenic ability. These cells can form blood vessels and express various angiogenic factors, such as VEGF-A, G-CSF, GM-CSF, and MMP3. SP cells could migrate to the ischemic area, secrete many angiogenic and neurotrophic factors in the ischemic area, and form blood vessels (Fig. 3). A similar population of endothelial cell SP cells has also been identified in lung cancer, demonstrating chemoresistance which would lead to a greater capacity for angiogenesis [87]. Cao et al. [88] used the Hoechst33342 method to isolate SP cells from non-small cell lung cancer. Compared with non-SP cells, SP cells expressed high levels of mRNA for VEGF-A, VEGF-B, Ang1, Ang2, fibroblast growth factor-2, cyclooxygenase-2, interleukin-8, ABCG2, and MDR1. Supernatants of non-small cell lung cancer SP cells could significantly induce endothelial cell migration and promote angiogenesis. In hepatocellular cancer, SP cells have been shown to preferentially downregulate several miRNA species, of note the downregulation of miR-148b by hepatocellular cancer SP cells promotes tumor angiogenesis by targeting neuropilin 1, a key VEGF co-receptor [89]. The role of SP cells in angiogenesis may also be related to their numbers, which can be modulated by factors in the tumor microenvironment. For example, in lung cancer cell lines, the expression of tissue inhibitor of matrix metalloproteinases 2 is inversely proportional to number of SP cells and therefore likely impacts angiogenesis not only directly but also by modulating SP function [90]. In addition, tumor hypoxia may expand or diminish SP cell numbers depending on the tumor type as evidenced by two different breast cancer cell lines [91].

Recent studies have demonstrated that tumor endothelial cells are a specialized cell type [92]; these are unique cells and can be distinguished from normal endothelial cells based on their larger nuclei, increased aneuploidy, abnormal centrosomes, and increased proangiogenic gene expression. They communicate with tumor cells to allow transendothelial migration and metastatic spread [92]. Their origins are diverse including bone marrow progenitors, normal endothelial cells, and CSCs [92]. Whether the same population of CSCs contributes to endothelial cell growth and tumor cell growth requires further examination.

Recent studies have also proposed that stem cell-derived exosomes, via their cargo of both miRNA species and specific proteins, may regulate angiogenesis [93]. However, there are little data on whether this is also a mechanism by which CSCs regulate tumor angiogenesis and opens up a potentially exciting area of research and novel therapeutic targets [94].

While it is clear that cancer SP cells contribute to tumor angiogenesis in several different cancer types, it is not yet clear whether the same function can be attributed to OC SP cells. Therefore, the OC SP population warrants further investigation in terms of their angiogenic potential in the progression of OC. It is possible that there is also redundancy with the other tumor-resident stem cell populations, and the coordinated roles of these cells should also be examined.

In addition to facilitating tumor growth, tumor angiogenesis also establishes a route for metastatic spread of tumor cells. However, a balance needs to be established as complete inhibition of angiogenesis creates a hypoxic environment, which in turn can promote a more invasive phenotype in the tumor cells [95]. Therefore, therapies that can target both angiogenesis and tumor metastasis are more likely to be effective than either therapy alone. Indeed, early anti-VEGF therapy trials showed greater efficacy in combination with traditional chemotherapy due to stabilization of the vasculature and better delivery of the chemotherapeutic agents to the tumor than when chemotherapy was used alone. Similar approaches could be taken in targeting CSCs, with a need to target both their angiogenic and metastatic potential.

CSCs in OC may contribute to tumor angiogenesis not only through secretion of activating factors but also through their own differentiation into endothelial cells via a mechanism associated with CCL5 signaling [96]. In addition, new promising therapeutics such as ALM201, an FKBPL-based peptide, have been shown to target both CSCs and angiogenesis [97]. Given the presence of different populations of CSCs within each tumor, combined therapies against the different cell populations and their modes of action need to be considered. Therefore, much more research on the different resident tumor stem cell populations, and how they might be targeted therapeutically, is urgently required.

In addition to the different populations of tumor resident stem cells, stem cells within the tumor microenvironment may also contribute to tumor metastasis, adding a further layer of complexity. Zhang et al. [98] demonstrated that adipose-derived stem cells from the ovarian microenvironment could contribute to increased metastasis and chemoresistance.

Tumor stem cells, including SP cells, play important roles in tumor angiogenesis and metastasis. The relationship between SP cells and angiogenesis and its related molecular mechanisms still needs to be fully elucidated but will likely provide a new direction for finding effective targets. So, do SP cells play a role in the angiogenesis of OC, and if so, how? Current studies in our laboratory are investigating these important questions.

The authors have no conflicts of interest to declare.

Funding for this study was provided by Guangzhou City Science and Technology Program (Grant No. 202102010292).

Yue Pan, XueFen Yang, and Miaojuan Chen wrote the first draft of the manuscript; Kun Shi, Yuan Lyu, Annette P. Meeson, and Gendie E. Lash edited the manuscript. All authors approved the final version of the manuscript for submission.

Yue Pan and XueFen Yang contributed equally to this work.