VSV based virotherapy in ovarian cancer: the past, the present and …future?

The standard approach to treating patients with advanced epithelial ovarian cancer (EOC) after primary debulking surgery remains taxane and platinum-based chemotherapy. Despite treatment with this strategy, the vast majority of patients relapse and develop drug-resistant metastatic disease that may be driven by cancer stem cells (CSCs) or cancer initiating cells (CICs). Oncolytic viruses circumvent typical drug-resistance mechanisms, therefore they may provide a safe and effective alternative treatment for chemotherapy-resistant CSCs/CICs. Among oncolytic viruses vesicular stomatitis virus (VSV) has demonstrated oncolysis and preferential replication in cancer cells. In this review, we summarize the recent findings regarding existing knowledge on biology of the ovarian cancer and the role of ovarian CSCs (OCSCs) in tumor dissemination and chemoresistance. In addition we also present an overview of recent advances in ovarian cancer therapies with oncolytic viruses (OV). We focus particularly on key genetic or immune response pathways involved in tumorigenesis in ovarian cancer which facilitate oncolytic activity of vesicular stomatitis virus (VSV). We highlight the prospects of targeting OCSCs with VSV. The importance of testing an emerging ovarian cancer animal models and ovarian cancer cell culture conditions influencing oncolytic efficacy of VSV is also addressed.

Introduction

Despite significant improvements in therapy, cancer is still one of the leading causes of death worldwide. Epithelium ovarian carcinoma (EOC) ranks fifth in lethal tumors among women, accounting for more deaths than any other cancer of the female genital tract. The American Cancer Society estimates that in the United States, in 2016 approximately 22,000 women received a new diagnosis of ovarian cancer, and near 14,000 women succumbed to this disease 1. Ovarian cancer can be classified into multiple types: serous, endometrioid, clear cell, mucinous, transitional cell, or any combination of these (mixed) with each type having widely different clinicopathologic properties. Among them, high grade serous ovarian cancer (HGSOC) is the most frequent histological type representing 70% of all EOC 2. Metastasis of EOC involves shedding of malignant cells from the primary tumor into the peritoneal cavity in the form of multicellular 3D spheroids. The cells within spheroid exhibit up-regulated expression of CSC markers suggesting that ovarian CSCs are enriched in this population 3. Emerging evidence supports the concept that during chemotherapy treatment the majority of differentiated ovarian cancer cells are initially chemosensitive and are eliminated. However platinum-based therapies are unable to eliminate OCSCs that have developed chemoresistance 4, 5. Therefore, development of novel therapeutic strategies for ovarian cancer therapy is of increasing interest. In particular successful clinical eradication of ovarian CSCs or targeting signaling pathways that are unique to OCSCs would ideally provide effective therapeutics that can potentially eradicate tumors entirely and lead to the sustained remission for patients with ovarian cancer. More and more evidence is accumulating to support the idea that ovarian CSCs can be an excellent target for oncolytic viruses 6, 7. Oncolytic viruses are natural or modified viruses that are highly selective to tumor cells.

The anti-tumor activity of oncolytic viruses may rely on direct lysis of neoplastic cells or induction of anti-tumor immunity. New generation of oncolytic viruses has been engineered to target CSCs and the CSCs' environment 8. Vesicular stomatitis virus (VSV) has demonstrated oncolysis or preferential replication in cancer cells and the topic has been addressed in several excellent reviews 9, 10. Nonetheless, the selective ability of this virus to kill cancer cells while sparing the normal cells remains debatable among some researchers. Also the potential of VSV in targeting of CSCs, specifically ovarian CSCs has not yet been explored. Here we will briefly summarize the current state of our knowledge regarding ovarian cancer, some characteristics of ovarian cancer stem cells, mechanism of their chemoresistance, and current state of oncolytic virotherapy for ovarian cancer [Table 1]. We also review the key components of the innate immune system and signaling pathways that might be involved in VSV susceptibility. In recent years several ovarian cancer animal models and pathway-specific engineered mouse allograft models that functionally recapitulate human serous epithelial ovarian cancer have been developed 11. Thus, future challenges and the potential of VSV in ovarian cancer treatment is discussed.

Table 1

Summary of article content

Section	Key points	References
1	Current diagnostic tools in early detection and treatment options for ovarian cancer patients are still very limited	[12-16]
	High recurrence rate of chemoresistant ovarian cancer has been associated with self-renewing ovarian cancer stem cells (OCSCs)	[27-40]
	Multiple mechanisms have been identified for OCSCs associated chemoresistance	[41-66]
2	Several clinical trials to treat ovarian cancer patients with oncolytic viruses have been initiated in recent years	[67-82]
3	Viruses that have been shown to have the potential of eradicating CSCs include adenovirus (Ads), herpes simplex virus-1 (HSV-1), vaccinia virus (VV), myxoma virus (MYXV), reovirus, measles virus (MV), Newcastle disease virus (NDV), and Maraba virus (MRBV)	[8, 83-98]
4	Vesicular stomatitis virus (VSV) is newly emerging and promising oncolytic agent for cancer treatment.Defects in innate immune responses involving the interferon (IFN) system and dysregulated translation machinery may contribute to VSV's intrinsic, oncolytic propertiesSeveral approaches have been undertaken to improve safety and selectivity of VSV	[99-137]
5	Efficacy of VSV in ovarian cancer mouse models varied considerably. Discrepancies in the outcome in these models may be related to the differences within tumors	[139-140, 142, 145-152]
6	The potential oncolytic effectiveness of VSV should be evaluated in newly emerged experimental models (in vivo or in vitro) that recapitulate different aspects of human ovarian cancer such as fallopian tube origins and ovarian cancer spheroids	[153-171]

1. Ovarian Cancer

Detection and Current State of Ovarian Cancer Treatment

The best currently available protocol for early detection of ovarian cancer is transvaginal ultrasound (TVS) and the presence of elevated CA-125 12, 13. Among other biomarkers used for the detection of ovarian cancer is HE4 (human epididymis secretory protein 4), which together with CA125 is a part of the Risk of Ovarian Malignancy Algorithm (ROMA), a scoring model proven to be useful in excluding malignant diagnosis in premenopausal women 14. New screening methods, including multiple serum markers are also being investigated 15, 16. All newly diagnosed epithelial ovarian cancer (EOC) patients undergo surgery, which is required for diagnosis and staging, in addition to providing cytoreduction. Nevertheless conventional combinations of primary cytoreductive surgery and paclitaxel-platinum based chemotherapy have not had a significant impact on overall survival of ovarian cancer in last several decades 17. The overall five-year survival rates of patients diagnosed at stage III and IV of this disease are 32% and 18%, respectively 18. Patients with platinum-resistant/refractory ovarian cancer are good candidates for novel investigational approaches. Recently it has been shown that introducing an antiangiogenic agent, such as, bevacizumab may significantly prolong the disease free survival time and the response rate to both front-line chemotherapy and recurrent disease treatment of EOC patients 19, 20. The strategy of antiangiogenic agents actions concentrates on the inhibition of the proangiogenic ligands with antibodies and receptors with tyrosine kinase inhibitors or receptors antibodies which should lead, in combination with chemotherapy, to antiangiogenic-induced tumor vasculature normalization and better chemotherapeutic agents elivery 21. Recently hyperthermic intraperitoneal chemotherapy (HIPEC) has been investigated as a treatment option for EOC 22, 23. The development of poly (adenosine diphosphate-ribose) polymerase (PARP) inhibitors also represents a novel therapeutic strategy, especially in patients with advanced, recurrent ovarian cancer who have mutations of the breast cancer 1 gene (BRCA1) or breast cancer 2 gene (BRCA2) 24.

Ovarian Cancer Stem Cells (OCSCs)

Relapse occurring after aggressive treatments such as surgical resection, chemotherapy, and radiotherapy has been attributed to a distinct tumor subpopulation of cancer stem cells (CSCs) or cancer initiating cells (CICs). The idea that cancer is driven by a smaller population of stem cells is not a new one. In the last century it was suggested that tumors contain a cellular subcomponent that retains key stem cell properties 25, 26. In recent years the clonal evolution of cancer formation in which tumors are thought to be homogeneous masses of proliferating cells with identical genetic alterations was replacing by “cancer stem cell hypothesis”. According to the CSC theory, solid tumors, similarly to many adult tissues may also be hierarchically organized and contain cancer stem cells that sustain tumor growth and relapse after therapy 27, 28. American Association for Cancer Research defined CSC as a malignant cancer cell with a stem cell phenotype 29. CSCs are characterized by their ability to self-renew, asymmetric division and their capacity to initiate tumor formation 30. Several excellent reviews provide comprehensive discussion of the presence of CSCs in many cancers 31, 32. Ovarian cancer stem cells were first reported in the ascites of ovarian cancer patients 33. Taking into a consideration the heterogeneity of ovarian tumor, it is possible that each of the types of ovarian cancer has its own unique ovarian CSC phenotype. The characteristic of tumor associated stem cells promises a challenge. Firstly, the expression of cell surface markers in the tissue of interest has to be known, secondly, the presence of normal adult stem cells in the tissue has to be taken into consideration as well. Currently there is no single, specific marker that definitively identify and enrich CSCs, rather a combination of multiple markers depending on the particular tumor type is used. However, with advancement of knowledge in this concept ovarian CSCs have been associated with the cell surface markers such as CD24, CD44, CD133, CD117, EpCAM and differential biochemical properties such as high expression of aldehyde dehydrogenase 1 (ALDH1) 34.

Among them expression rate of CD24, a membrane glycoprotein, positively correlated with higher histopathologic grade, clinical stage and omental metastasis in patients with epithelial ovarian cancer. Gao et al. revealed that CD24+ cells possess stem cell-like characteristics with elevated expression of related stem cell genes and proteins (including Nestin, β-catenin, Bmi-1, Oct4, Oct3/4, Notch 1, Notch 4, CD133, CD44 and CD117). Moreover, CD24+ cells had stronger resistance to chemotherapy drugs with higher self-renewal ability 35. Another surface transmembrane glycoprotein highly expressed in ovarian CSCs is CD44, a receptor for hyaluronic acid (HA) involved in cell-cell and cell-matrix interactions. Meng et al. demonstrated that CD44+/CD24- cells in ovarian cancer cells exhibit cancer stem cell-like properties of enhanced differentiation, invasion, and resistance to chemotherapy 36. Another widely described ovarian CSCs marker is CD133. It is a transmembrane glycoprotein associated with cell migration and invasion. It has been demonstrated that ovarian CD133+ CSCs express other stem cell markers, including Oct-4 and Nanog 37. Chau et al. revealed that CD117, a type III receptor tyrosine kinase, is involved in cell signal transduction in ovarian carcinoma and its presence was associated with poor response to chemotherapy 38. Nevertheless, it should be noted that ovarian cancer stem cell population may be composed of small overlapping cell fractions defined by mixed stem-like markers. There are several excellent reviews that address the more recent developments in ovarian CSCs research 39, 40.

Mechanisms of Ovarian CSCs Resistance to Chemotherapy

Standard treatment of epithelial ovarian cancer relies on the combination of cytoreductive surgery and combination chemotherapy with taxane and platinum. Recently, ovarian cancer stem cells (CSCs) population is accounted for temporary effects of such therapy and subsequent cancer recurrence in ovarian cancer patients 41, 42. Multiple mechanisms have been identified for CSCs associated chemoresistance. Yang et al. hypothesized that chemotherapy not only enriches the cancer stem cells due to their intrinsic resistance but induces CSCs in hypoxic region of the tumor 43. It is still unclear how CSCs survive DNA-damaging agent treatment, acquiring multidrug resistance (MDR) and enhanced the migration potential. Several pathways could be involved in these mechanisms. One of them is an increased rate of efflux of anti-cancer drugs which is mediated mainly by a group of transmembrane transporter glycoproteins known as ATP-binding cassette (ABC) proteins 44, 45. Glutathione system (GSH), involved in cellular detoxification by conjugation with toxic chemicals and certain chemotherapeutic agents such as cisplatin promote their efflux from cells 46. Moreover, the glutathione system is one of the intracellular scavenger systems for neutralizing free oxygen radicals (ROS) and has been found to maintain the lower levels of ROS in CSCs, which is essential to sustain CSC self-renewal and stemness 47. Recently the aldehyde dehydrogenase (ALDH) has also been recognized for CSC associated chemoresistance. Landen et al. demonstrated that expression and activity ALDH which are responsible for oxidizing intracellular aldehydes was significantly higher in taxane and platinum-resistant ovarian cancer cell lines 48. Deregulation of β-catenin, the key protein in the Wnt signalling pathway may also play a role in the pathogenesis of HGSOC. Furthermore, strong membranous β-catenin expression was associated with resistance to platinum-based chemotherapy 49. Another marker that has been associated with chemoresistance of ovarian CSCs is a transcriptional receptor Bmi1 (B cell-specific Moloney murine leukemia virus integration site 1). Bmi1 belongs to the polycomb group (PcG) gene family involved in several cellular processes including cell cycle regulation, cell immortalization and senescence of adult, and neoplastic stem cells 50. Zhang et al. revealed that ovarian tumor-derived spheroid cells overexpressed Bmi1 and furthermore, they were resistant to cisplatin 39. Additionally it has been demonstrated that Bmi-1 is essential to promote epithelial to mesenchymal transition (EMT) and tumor-initiating capability 51, 52. EMT is a crucial event for cell dissemination of epithelial tumors. Through this process epithelial cells lose their differentiated characteristics, acquire mesenchymal features, and finally become mesenchymal phenotype which facilitates migration and resistance to chemotherapy. Chiu et al. revealed that chemoresistant ovarian cancer cells exhibit an epithelial-mesenchymal transition (EMT) phenotype and high invasion ability 53. Ip et al. pointed out the significance of hydrodynamics in ovarian cancer progression. With the use of a microfluidic model of the peritoneum and three-dimensional (3D) spheroids for mimicking tumor behavior, the researchers showed that shear stress enhanced stemness and chemoresistance of ovarian cancer spheroids through downregulation of miR-199a-3p expression and activation of PI3K/Akt signaling 54. The NFκB pathway which is a major source of pro-inflammatory cytokines may also contribute to ovarian cancer chemoresistance 55. Recently a higher expression of the survival protein Bcl-2, and a signaling kinase PKCδVIII in cisplatin-resistant ovarian cells TOV-112D has been reported 56. Another key player which is influencing apoptosis is the p53 protein. Mutation of the p53 gene is a frequent event in tumorigenesis and the majority of high-grade serous carcinomas harbor mutations in this gene 57. Yang-Hartwich et al. demonstrated that impaired p53 degradation in a population of ovarian cancer cells with cancer stem cell properties led to the imbalance of p53 turnover that promoted the formation of p53 aggregates and consequently to p53 inactivation and platinum resistance 58. Another pathway, important in maintaining the cancer stem cell in a variety of cancers is Notch signaling. The Notch pathway regulates cell proliferation, cell fate, differentiation, and cell death. It has been found that particularly Notch3 contributes to the tumorigenesis in ovarian cancer. Notch3 is overexpressed in 66% of ovarian serous carcinomas 59. The study of McAuliffe et al. demonstrated that Notch3 overexpression in ovarian CSCs results in their expansion and increased resistance to cisplatin 60. Steg et al. has reported other signaling pathways that were overexpressed in recurrent ovarian tumors. Among them were members of the TGF-β superfamily, mainly ENG, and the primary mediators of Hedgehog (Hh) pathway (GLI1 and GLI2) 41. Additionally, Zeng et al. demonstrated that Hedgehog signaling is also involved in the regulation of CD24 expression, which is one of the most widely described ovarian CSCs markers 61. Liao et al. pointed out that resistance of ovarian cancer stem cells to conventional chemotherapies may be due to their metabolic properties, mainly their resistance to oxygen deprivation. His group observed that spheroid cells have defects in complete glucose oxidation in the tricarboxylic acid (TCA) cycle and utilize glucose via anaerobic glycolysis. According to the authors, such cells are more likely to survive in severely hypoxic microenvironments and have an increased chemoresistance 62. Another mechanism involved in acquiring drug resistance is regional gene activation as was demonstrated in taxane-selected human ovarian cancer cell lines 63. Srivastava et al. showed recently that ovarian CSCs have intrinsically enhanced Pol η-mediated translesion DNA synthesis (TLS) which permitting CSCs to survive cisplatin treatment and leading to tumor relapse 64. Despite these findings however, successful cytoreduction and chemotherapy is currently a first, and sometimes the only line of defense. It may destroy the CSCs environment preventing tumor recurrence. Paclitaxel for example has been shown to induce cell cycle arrest in the G2-M phase, and activating proapoptotic signaling in ovarian cells 65. Jia et al. showed that ovarian cell line- SKOV3 treated with paclitaxel, acquired a benign fibroblast-like phenotype and have decreased proliferative potential and impaired capacity to form colonies 66. It is crucial to fully understand the basic biology of CSCs and mechanisms causing chemoresistance. Furthermore, identification of specific therapeutic targets in key signaling pathways of CSCs may improve the development of novel treatment strategies involving oncolytic viruses.

2. Oncolytic Virotherapy for Ovarian Cancer - Current Clinical Testing

Oncolytic viruses (OVs) are defined as genetically engineered or naturally occurring viruses that selectively replicate and kill cancer cells, spreading within the tumor while sparing surrounding normal, healthy tissue. OVs can be injected intratumorally or intravenously. Systemic distribution of the oncolytic virus allows for the opportunity to treat even distant, metastatic lesions. Another hallmark of oncolytic viruses, besides oncolysis is initiating or augmenting existing anti-tumor immunity. Several virus families have served as backbones for the development of OVs for treatment ovarian cancer 67, 68. Currently 9 active phase I or II oncolytic clinical trials (as of end of 2016) have been conducted for cancers affecting ovary, fallopian tube and peritoneal cavity. Three study employed measles virus encoding thyroidal sodium iodide symporter IP (NCT02364713, NCT00408590, NCT02068794), no results are yet available 69-71. The trial NCT02017678 with vaccinia virus (JX-594) with inactivated viral thymidine kinase gene (TK) and encoding human granulocyte-macrophage colony-stimulating factor (GM-CSF) has been withdrawn prior to enrollment, however promising results of yet another study using JX-594 in unresectable primary hepatocellular carcinoma (NCT00554372) gives hope for the treatment of ovarian carcinoma 72. GL-ONC1- vaccinia virus with the insertion of beta-galactosidase and beta-glucuronidase in place of TK and hemagglutinin, and vaccinia virus Ankara expressing p53 (p53MVA) are also currently evaluated in oncolytic immunotherapy in patients with recurrent ovarian cancer (NCT02759588 and NCT02275039) 73, 74.

Adenovirus (Enadenotucirev, ColoAd1), a chimeric unarmed Ad11p/Ad3 group B - is currently under evaluation in patients with recurrent, platinum-resistant ovarian cancer (NCT02028117), no results are yet available 75. The safety, potential efficacy, and possible gene transfer imaging capacity of adenovirus expressing a therapeutic thymidine kinase suicide gene -Ad5.SSTR/TK.RGD was confirmed in patients with recurrent gynecologic cancer (NCT00964756) 76. Another genetically modified adenovirus Ad5Delta24RGD binds to integrin αvβ3 and αvβ5, and it has been shown to be well tolerated by ovarian cancer patients (NCT00562003) 77. Several replication defective adenoviral vectors (Adp53) to restore of wild-type p53 function have been conducted in various clinical trials. However, clinical studies fail to provide an adequate safety and antitumor efficacy for adenoviral vectors 78. There are new, promising studies underway utilizing adenoviral vectors in targeting ovarian cancer (NCT02435186 and NCT02140996) 79, 80. An additional prominent example of OV is reovirus. Used already in many clinical trials as monotherapy currently is also evaluated in combination with chemotherapeutic drug in patients with ovarian epithelial cancer (NCT01199263) 81. Although reovirus oncotherapy has been shown to have a potential in treatment of ovarian cancer, researchers have to combat the problem of an antiviral host immune responses induced by reovirus and refine delivery of the virus to cancer cells 82.

3. Targeting Cancer Stem Cells with Oncolytic Viruses

A growing number of researchers worldwide question whether CSCs can be eradicated by oncolytic virus therapy. Viruses that have been shown to have the potential of eliminating CSCs include adenovirus (Ads), herpes simplex virus-1 (HSV-1), vaccinia virus (VV), myxoma virus (MYXV), reovirus, measles virus (MV), Newcastle disease virus (NDV) and Maraba virus (MRBV) Tong 8. Recent study utilizing GP73-regulated oncolytic adenoviruses demonstrated their effectiveness in destroying human liver cancer stem-like cells 83. Yano et al. have shown that genetically engineered, oncolytic adenovirus, OBP-301 efficiently killed CD133+ GCSCs resistant to chemoradiotherapy 84. Lv et al. revealed that adenovirus expressing 11R-P53 and GM-CSF targets hepatocellular carcinoma and teratoma stem cells 85. Among herpesvirus family, HSV G47Δ was found to be highly effective to the CD44+CD24-/low breast cancer stem-like cells population in vitro 86. An oncolytic HSV-1 mutant rQNestin34.5 has been shown kill neuroblastoma CSCs 87. Belonging to the poxvirus family-oncolytic vaccinia virus (CVV) effectively suppressed stem cell-like colon cancer cells (SCC) 88. Recent study of Wang et al. demonstrated that oncolytic VV, GLV-1h68, replicated more efficiently in breast cancer stem-like cells that were characterized by higher ALDH1 activity 89. Myxoma virus (MYXV), another virus from the Poxviridae family has been shown to effectively infect neuroblastoma CIC cultures and human brain tumor-initiating cells (BTICs) 90, 91. Moreover, MYXV was potent ex vivo to remove CICs from samples obtained from AML patients 92. CSCs of breast cancers expressing CD24- CD44+ and ALDH1+ have been susceptible to oncolytic reovirus in vitro and in vivo 93. Reovirus has been also shown to be effective in human glioblastoma cell line grown in spheroid cultures and in glioblastoma stem-like cells (GSC) 94, 95. The attenuated strains of measles virus (MV), MV-141.7 and MV-AC133 have been shown to infect and selectively lyse CD133+ tumor cells of primary glioma and colon cancer 96. Another promising oncolytic virus is Maraba virus (MRBV) belonging to Rhabdoviridae family. Different point mutations in the strain of MRBV were introduced to increase its oncolytic efficiency. The specific tropism of MRBV for cancer cells remains unclear, however MRBV has been shown to be a promising oncolytic agent for cancer virotherapy in a broad range of cancer cells, including EOC 97. Moreover, Maraba virus effectively replicates in epithelial ovarian cancer (EOC) tumor spheroids cells suggesting that VSV might also be a potential therapeutic agent for eradication of ovarian CSCs 98.

4. Oncolytic properties of vesicular stomatitis virus (VSV)

Vesicular stomatitis virus (VSV) of the Rhabdoviridae family as an oncolytic agent has been studied recently by numerous investigators 99. VSV's single-stranded RNA genome is relatively simple, fully understood, and easy to manipulate. The 11 kb VSV genome encodes five genes; these include the nucleocapsid (N) protein, phosphoprotein (P), matrix (M) protein, G protein and large polymerase (L). Bullet-shaped virion is enveloped and encodes all five virus-encoded proteins 100. VSV causes a relatively harmless nature of oral diseases in livestock and symptomatic human infection is usually rare and is characterized by mild, flu-like symptoms 101. VSV replicates within the cytoplasm of infected cells, and does not undergo genetic recombination; moreover, it does not integrate any part of its genome into the host, and it is not known to have transformation potential. Another advantage of VSV is fast growth to very high titers in vitro, which eases a purification of large amounts of virus and viral proteins in infected cell lysates 102. Although VSV rarely causes symptomatic human infection, seroconversion with no obvious signs of illness was reported among laboratory workers, in human populations in endemic regions, as well as animal handlers 103. VSV as an oncolytic agent is currently being considered by a number of insightful research groups in the USA and Canada. Recent findings revealed that several obstacles for successful oncolytic virotherapy with VSV must be overcome. Although preferential tropism of VSV to the tumor cells has been examined in the case of melanoma, sarcoma, glioblastoma, cervical cancer, breast cancer and leukemia [104 and Ref. within], some scientists are not fully convinced in regards to VSV oncoselectivity 105. We have been exploited VSV in the study of innate immunity in humans for many years. In our extensive research on the population of 300 people we determined the ability of VSV to replicate within human peripheral blood leukocyte populations obtained from healthy individuals as well as from people with immunocompromised immune system. We have shown that sensitivity among people to VSV is differentiated and depends on the condition of the immune system 106-109. Thus health status of blood donors as well as their age contributes to the sensitivity of PBL to VSV infection 110. In addition, we have also shown that VSV replication capacity reflects the progress of HIV infection in peripheral blood leukocytes 111. Oncolytic virotherapy exploits live viruses and the fact that VSV has been shown to cause viral encephalitis in animal models make the safety the highest priority, especially in individuals with immune system compromised by the disease or chemotherapy 112. Therefore the current key challenge for successful virotherapy is to generate VSV which is resistant to the host's antiviral immune responses, with better tumor specificity, enhanced oncolytic efficacy, and the most important- has an excellent safety profile. Furthermore, a therapeutic synergy with currently available cancer therapeutics would be also desirable. An additional problem associated with the use of VSV in oncolytic therapy is the fact that individual cancers may vary in their sensitivity to VSV even when these cancers arise from the same tissue type. Carey et al. discovered that the resistance to VSV is related primarily to the constitutive expression of antiviral genes, mainly IFN-inducible genes, prior to infection 113. To improve safety and selectivity of VSV an attenuated matrix protein mutant (M51RVSV) was generated. Tumor cells often harbor defects in the IFN system, therefore VSV replicates to high levels in many transformed cells. Moreover, continuously proliferating cancer cells often have dysregulated translation machinery conducive to VSV replication 9, 114. Lyles at al. discovered that wild type VSV inhibits hosts antiviral responses through interfering of the VSV matrix (M) protein with the functions of all the three RNA polymerases (RNAP) 115. Moreover, M protein interacts with the mRNA export factor Rae1 and Nup98 protein in nucleus inhibiting the export of a subset of mRNAs and triggering a shutoff of host gene expression 116. However, this M- protein mediated inhibition of mRNA export can be antagonized by IFN, which upregulates the expression of Nup98, Nup96 and Rae1 117. In contrast to wild type VSV, M51RVSV is defective for inhibition of the host antiviral response, however it can efficiently induce type I IFN production by normal cells. M51RVSV can produce progeny virions and is capable of inducing apoptosis facilitating selective killing of IFN- sensitive cancer cells, therefore M protein mutants have been proposed as strong candidates for oncolytic viral therapy 104. To further improve selectivity of VSV different approaches have been undertaken. Recent studies have shown that VSV can be efficiently retargeted to different cellular receptors using measles virus envelope glycoproteins, non-neurotropic envelope glycoprotein of lymphocytic choriomeningitis virus (LCMV), Lassa virus, or human immunodeficiency virus type 1 (HIV-1) gp160 118-121. Another engineering strategy focused on controlling post-entry replication of VSV. Kelly et al. demonstrated ablated neurotoxicity of recombinant VSV which had incorporated neuron-specific microRNA target sequences without negatively affecting viral oncolysis 122. Neuroattenuation of VSV, without perturbing its oncolytic potency has been achieved also through engineering of IRES elements from human rhinovirus type 2 (HRV2) and foot-and-mouth disease virus (FMDV) into VSV to control the translation of the matrix gene (M) 123. Several studies were also conducted on attenuation of VSV through deleting viral genes or disruption of their order 124-126. The safety and oncoselectivity of VSV was significantly enhanced in VSVs expressing murine interferon beta (VSV-mIFNβ) or human IFN-β (VSV-hIFNβ) and furthermore, such viruses retained their oncolytic activity against tumor cells in vitro and against tumor xenografts in mice 127, 128. Safety studies on intravenous administration of oncolytic recombinant VSV confirmed that intravenous administration of VSV-mIFNβ-NIS is well tolerated in C57BL/6 mice as well as in purpose-bred dogs 129, 130. Moreover, a phase I clinical study is currently underway in patients with liver cancer (NCT01628640), mainly to evaluate the safety of intratumoral administration of VSV expressing human IFN-β 131. Recently Westcott et al. demonstrated that IFN-2α rather than IFN-β may be more effective for improving VSV selectivity in cancer cells 132. Other interesting studies on improving therapeutic synergy have shown the oncolytic effect of WT VSV or (VSV)-MΔ51 engineered to express suicide enzyme, cytosine deaminase/uracil phosphoribosyltransferase which allow the conversion of non-toxic prodrug (5-fluorocytosine) into toxic, chemotherapeutic 5-fluorouracil (5-FU), which exhibited a considerable bystander effect at the tumor site 133, 134. Another important problem associated with oncolytic virus therapy is generation of undesirable host immune response. For example neutralizing antibodies can be raised through prior systemic administration of VSV during therapy. Diverse techniques can be applied to protect the oncolytic virus from inactivation by pre-existing antibodies in the blood. Among such techniques one of them is combination of virotherpy with pharmacological immunosuppression to control immune responses 135, 136. Other methods involve protecting the VSV from neutralization through virus shielding strategies such as serotype switching when the entire VSV-G glycoprotein gene sequence is replaced with that of another serotype or chemical modification 137.

5. Current Status of VSV Based Virotherapy in Ovarian Cancer Animal Models

Anticancer activity of VSV has been shown across a spectrum of malignancies in the preclinical setting 104. Because significant progress has been made to understand the mechanisms of VSV oncolytic selectivity and activity, therefore a strong rationale exists for the use of VSV in ovarian cancer therapy. In ovarian cancer research one of the major challenges is lack of suitable animal model systems targeting different cell types that truly replicate human ovarian disease. Multiple histological subtypes of ovarian cancer are being treated with similar surgical and therapeutic approaches, even though they originate from different cells and are characterized by various genetic and genomic alterations. Most researchers utilize murine xenograft and transgenic models of ovarian carcinoma. Xenografts of human or mouse ovarian cancer cells in immunodeficient mice allow for investigation of growth, metastasis, and treatment response of ovarian tumors in a live animal. Perspectives for personalized cancer therapy has led to, an increased interest in patient-derived xenografts (PDX) engrafted into immune-compromised mice for preclinical modeling. Tumor cells together with parenchymal and stromal tumor components, can be transplanted into ovarian bursa (intrabursally, IB) or into mouse peritoneal space (intraperitoneally, IP) generating “the Avatar”, an orthotopic PDX model allowing further study of tumor development and metastases 138. However, xenografts do not illustrate the initial transforming events leading to tumorigenesis. Apart from problems with species differences between host and graft, another major disadvantage of immunodeficient models for oncolytic virotherapy studies is their increased susceptibility to viral infection. Since an inherent neurotropism of wild type VSV poses a severe threat to immunocompromised animals, a recombinant, attenuated version of virus has become a favored approach. A recent study of Dold et al. revealed that wild type VSV caused neurotoxicity at 102 PFU per mice while doses up to 109 PFU of VSV pseudotyped with the glycoprotein of the lymphocytic choriomeningitis virus (LCMV)-VSV-GP were well tolerated even in immmunodeficient, tumor-bearing NOD/SCID mice. However, treatment of subcutaneous xenografts of human ovarian carcinoma with VSV-GP was not very effective. Authors concluded that recurrence of tumors in vivo was related to the intact antiviral mechanisms in the tumor cells and that inhibition of VSV-GP activity was mediated by IFN as application of interferon modulators improved oncolysis 139. Since mouse xenograft models lack a functional immune system, the development of immunocompetent, genetically engineered mouse models of epithelial ovarian cancer that more accurately recapitulate metastatic ovarian carcinoma is crucial 140. The cellular origin of ovarian cancer has remained disputable therefore a variety of transgenic animal models have been developed for the various histotypes of ovarian cancer. Genetically engineered mouse models of ovarian cancer faithfully recapitulate many aspects of human disease and have greater clinical relevance because tumor development, neovasculaturization and immune responses take place within the context of normal tissues. Consequently, cancer in transgenic mouse models may be more difficult to treat. There are some excellent mouse transgenic models of EOC 141 but only some were exploited in studies on oncolytic efficacy of VSV. In a first transgenic model of spontaneous EOC (SV40) large T antigen, including the small and large T antigen (TAg) was expressed genes under transcriptional control of the Müllerian inhibiting substance type II receptor promoter (MISIIR). Female TgMISIIR-Tag DR26 transgenic mice spontaneously develop bilateral poorly differentiated serous tumors derived from the ovarian surface epithelium (OSE) 142. Lesions resemble high-grade serous ovarian carcinoma (HGSOC), the most common form of ovarian cancer. Expression of Simian virus 40 (SV40) large T antigen causes inactivation of p53 and pRB which is a crucial feature as HGSOC is characterized by a high frequency of TP53 mutations, observed in 96% cases. Dysregulation of Rb1 are seen in about 67% cases 143, 144. Another model is represented by a naturally occurring mutant, white spotting variant (Wv) mice which harbor a point mutation in the kinase domain of the c-kit gene. Mutations at the mouse W/c-kit locus lead to intrinsic defects in stem cells of the melanocytic, hematopoietic, and germ cell lineages. The Wv/Wv mice develop tubular adenomas, a benign epithelial neoplasm that can become invasive in older Wv mice 145.

The results of oncolytic cancer therapy with VSV in these two mouse models are quite different. Initial testing the therapeutic efficacy of an attenuated strain VSV-Δ51 in a TgMISIIR-Tag DR26 mouse was not satisfactory 146. Nonetheless this group reported later that replication and spread of the virus was augmented within tumors arising on the ovaries by the histone deacetylase inhibitor (HDI) treatment which weakened the innate antiviral responses 147. Arulanandam et al. study demonstrated that treatment with VSV-Δ51-GM-CSF and microtubule-destabilizing agents (MDAs) such as colchicine both delivered i.p. led to significantly prolonged survival of TgMISIIR-TAg mice as compared with either monotherapy 148. On the other hand, a study utilizing Wv mice, demonstrated the high activity of VSV in the monotherapy. VSV was particularly efficient in clearing tumor lesions without significant replication in any other organs and tissues 149. Discrepancies in outcome in these models may be related to the differences within tumors. The ovarian tubular adenomas in the ovary of the Wv mice resemble rather the low-grade ovarian cancer tumors, which usually have a wild-type Trp53 and KRas or Raf-1 activating mutation while lesions in TgMISIIR-TAg animals are similar to HGSOC with inactivated p53 150. The differences in two models are highlighted by Cai et al. The study revealed that even the addition of an oncogenic mutation, such as Trp53 deletion or reduction of Pten gene dosage in Wv mice did not convert the benign ovarian epithelial tumors into malignant cancer resembling HGSOC 151. Nevertheless the white spotting variant (Wv) mouse model can provide important insights into the etiology and pathogenesis of ovarian cancer and is an excellent research model for the menopause study. There are also several disadvantages of using mouse models. The researchers pointed out that some mouse models do not accurately mimic human molecular mechanisms of inflammatory responses 152. For example unlike human ovaries, murine ovaries have a covering bursa that protects the peritoneal cavity from epithelial cell shedding in the OSE, which could inhibit metastasis. Additionally there are differences in the cycle between human and mice. Humans are a mono-ovulatory species and have menstrual cycles rather than estrous cycles 18. All of these make the interpretation of mouse models extremely challenging.

6. VSV virotherapy in ovarian cancer-future study directions

The potential oncolytic effectiveness of VSV in ovarian cancer should be more precisely evaluated in newly emerged experimental models (in vivo or in vitro) that recapitulate different aspects of human ovarian cancer. Such experimental models include mouse model of fallopian tube origins of HGSOC and mouse models with altered signaling pathways, which are not only involved in ovarian cancer development, but are also crucial for the innate immune responses. Not to mention that animal models should address a metastatic potential of ovarian cancer spheroids, presence and biology of CSCs, and their role in chemoresistance. Table 2 presents a simplified overview of the development of various mouse models addressing the issue of the origin of high grade serous ovarian cancer (HGSOC) as well as potential mechanisms of VSV oncoselectivity. In the contexts of recent findings that HGSOC is considered a müllerian origin resembling the epithelium of the fallopian tube, mogp-TAg transgenic mouse has been developed 153. In this model the SV40 large T-antigen (TAg) is expresses under the control of the mouse müllerian-specific murine oviduct-specific glycoprotein promoter (Ovgp-1) that is highly active in the fallopian tube. Transgenic mogp-TAg mice develop neoplastic lesions in the fallopian tube resembling human serous tubal intraepithelial carcinoma (STIC) with disrupted p53 and Rb pathways 154. Therefore mogp-TAg mice model is valuable not only for elucidating the mechanisms of serous ovarian tumorigenesis, but may also serve as more accurate model for testing the efficacy of VSV as an oncolytic virus.

Table 2

New experimental mouse models of high-grade serous ovarian carcinoma (HGSOC)

Promoter	Targeted gene	Mechanism	Tumorigenesis	Reference
Ovgp-1	SV40 TAg	Inactivation of p53 and pRb	Neoplastic lesions in the fallopian tube resembling human serous tubal intraepithelial carcinoma (STIC) a potential precursor of ovarian HGSC	Miyoshi et al., 2002 [153]Sherman-Baust et al., 2014 [154]
MISIIR (Amhr2-Cre)	Dicer-/-& Pten-/-	Disruption of PI3K and AKT pathways (increased phosphorylation of AKT, STMN1- stathmin) and BIRC5 -survivin)	High grade serous carcinomas from the oviduct. Tumors spread to the ovary and metastasize throughout the abdominal cavity. Upregulated expression of cytokeratin 14, 17, and 8, E-cadherin, CA125	Kim et al., 2012 [155]
Pax8-Cre	Brca 1 or Brca 2-/-, p53-/-, Pten-/-	Alteration in PTEN/PI3K pathway	Serous tubal intraepithelial carcinoma (precursor lesion to ovarian HGSC and peritoneal carcinomas). Upregulated expression of including cytokeratin-8, STMN1, PAX2, P53, Ki-67, and CA-125	Perets et al., 2013 [160]
AdCre	Pik3ca & Pten-/-	Disruption of PI3K and AKT pathways (increased phosphorylation of AKT)	Serous papillary hyperplasia of the ovaries	Kinross et al., 2012 [156]
AdCre	Pten-/- & Apc-/-& p53-/-	Disruption of PI3K/AKT/mTOR pathway	High grade metastatic ovarian carcinomasCD24+ cells showed greater tumor initiation rates	Burgos-Ojeda et al., 2015 [161]

STIC, serous tubal intraepithelial carcinoma; HGSC, high-grade serous carcinoma; MISIIR, Müllerian inhibiting substance type II receptor; PAX2 and PAX8, transcription factors; CA-125, cancer antigen 125; Ki-67, a nuclear protein expressed in proliferating mammalian cells; PI3K, phosphatidylinositol 3-kinase; AKT, serine/threonine kinase, mTOR, mammalian target of rapamycin; PTEN, phosphatase and tensin homolog; Dicer, endoribonuclease

In another transgenic model of EOC a conditional knock out of Dicer and Pten led to activation of the phosphatidylinositol 3-kinase (PI3K) and AKT pathway, and resulted in the development of high grade serous adenocarcinomas from the oviduct. Authors suggest that the lesions consist of cells of a mesenchymal lineage since they originate in the stroma of the fallopian tube. Moreover, in this model the primary tumors metastasize throughout the abdominal cavity 155. Kinross et al. confirmed development of HGSOC in mice with yet another alteration of PI3K/AKT/mTOR pathway with deletion of Pten and mutation in PIK3CA gene enhancing PI3K activity 156. PTEN is the tumor-suppressor phosphatase with phosphatidylinositol 3,4,5 trisphosphate (PIP3) as its primary substrate. Because PIP3 is the main cellular product of PI3K, PTEN is capable of antagonizing PI3K activity and negatively regulates the PI3K-AKT cell-survival signaling pathway 157 (Figure 1). PTEN loss and simultaneous activation of AKT and mTOR signaling is frequently observed in human ovarian carcinomas. Martins et al. revealed that PTEN not only is frequently deleted in HGSOC but reduced PTEN expression is associated with significantly worse survival 158. Another important role of PTEN is associated with its role as the p53 regulator which is a key factor as the majority of HGSOC also harbor mutations of p53. It has been shown that PTEN and p53 interact and regulate each other at the transcription as well as protein level. PTEN regulates p53 protein levels through AKT-dependent and independent mechanisms which could be one of the switching mechanism between cell death and survival 159. Perets et al. established serous tubal intraepithelial carcinoma as the precursor lesion to high-grade serous ovarian and peritoneal carcinomas in mouse models by targeting the Brca, Tp53, and Pten genes 160. Burgos-Ojeda et al. utilized yet another transgenic murine model of ovarian cancer with conditional deletion of Pten additionally to Apc and Trp53 in the ovarian surface epithelium (OSE), which results in the generation of high grade metastatic ovarian carcinomas. Although these tumors have endometrioid histology, in the presence of a p53 mutation they, have a high grade metastatic phenotype similar to that seen in patients with high grade serous cancer. What is interesting is that cell lines and primary tumors derived from this model had higher expression of CD24 and CD44 markers that have been associated with cancer stem cells (CSCs). Moreover these CD24+ cells possessed higher tumorigenic ability and demonstrated a greater ability to passage to form secondary spheres. CD24+ cells had also increased pSTAT3 and expressed the stem cell genes Nanog, cyclin-D1, and c-Myc 161. It is not surprising since Schubbert et al. demonstrated recently that PTEN loss leads the development of CSCs that share properties with somatic stem cells, including the capacity for self-renewal and multi-lineage differentiation 162. Moreover another study revealed that PTEN is also crucial for stem cell maintenance as PTEN loss can lead to the emergence and proliferation of cancer stem cell (CSC) clones 163. Therefore it would be worth testing oncolytic efficacy of VSV in ovarian cancer mouse model based on fallopian tube transformation or with deletion of PTEN in particular because tumor suppressor PTEN has a critical role in antiviral innate immunity. Recently, Li et al. demonstrated that PTEN is involved in IRF3-mediated, innate immunity and can be responsible for the inhibition of VSV replication. VSV-induced induction of type I interferon was severely impaired in PTEN-mutant mice and this effect was accompanied with higher VSV titers in the livers of PTEN-mutant mice and higher VSV-triggered mortality than in PTEN-wild-type mice.

Figure 1

A schematic overview of the molecular network involved in VSV oncoselectivity. Binding of growth factors to the receptor tyrosine kinase (RTK) activates the receptor complex, which in turn recruits and activates PI3K. Activated PI3K converts PIP2 to PIP3, which subsequently mediates the phosphorylation of AKT. Tumor suppressor PTEN negatively regulates the pathway by removing the 3-phosphate from PIP3, converting it back to PIP2. PTEN is crucial for the activation of IRF3, its import into the nucleus and production of IFN-β. Loss of PTEN leads to over-activation of AKT and subsequently mTOR, which is associated with uncontrolled cell growth, proliferation, and survival. PTEN loss leads also to the development of cancer stem cells (CSCs) and an impaired cellular responses to viral infection. Up-regulation of LDLR in ovarian carcinomas enables VSV to enter cells with altered PTEN function through LDLR. Abbreviations: PI3K, phosphatidylinositol 3-kinase; PIP2, phosphatidylinositol (4,5)-bisphosphate); PIP3, phosphatidylinositol (3,4,5)-trisphosphate); AKT, serine/threonine kinase; IRF3, interferon regulatory factor-3; mTOR, mammalian target of rapamycin; LDLR, low-density lipoprotein receptor. See text for further details.

Moreover, deficiency in PTEN and its phosphatase activity increased the replication of VSV in human prostate cancer cells PC-3. Enhancement of VSV replication and induction of type I interferon was independent of the PI3K/Akt signaling pathway and was related rather to the impairment import of IRF3 (a master transcription factor responsible for IFN-β production) into the nucleus. 164 (Figure 1). However, it is worth mentioning that regardless of the suppressive effects on the antiviral response of acute deletion of PTEN, its loss does not always account for the difference in susceptibility to VSV infection. Recently, Yu et al. study revealed that deletion of Pten in murine prostate epithelial progenitor (MPE) caused poor response to interferon and increased susceptibility to VSV infection. On the other hand, tumor-derived Pten-/- cells were resistant to VSV infection. Authors speculate that sensitivity to VSV infection observed in early tumor development, following Pten deletion is lost during tumor evolution due to other mutations or heterogeneity in the tumor microenvironment 165. The role of PTEN in participating in VSV oncoselectivity should then be further elucidated.

One of the reasons that EOC is difficult to treat is due to the unique mechanism of ovarian cancer metastasis. EOC dissemination involves local invasion of pelvic and abdominal organs through aggregates of malignant cells (spheroids). Intra-peritoneal metastatic lesions are initiated during proteinase-mediated shedding of single cells and spheroids that subsequently adhere to peritoneal mesothelial cells 166. In vitro cultured spheroids possess high tumorigenic ability. Pease et al. revealed that spheroids can be formed by budding directly as adherent clusters from a monolayer of ovarian cancer cell lines. This phenomenon was accompanied by expression of vimentin and lack of cortical E-cadherin suggesting epithelial to mesenchymal transition (EMT). Furthermore, cells grown as spheroids acquired progressive resistance to the chemotherapy drugs paclitaxel and cisplatin in comparison to their adherent counterparts 167. The main problem associated with the current developments of new antitumor agents is that they do not address the anchorage- and vascular-independent ovarian cancer spheroids. Therefore studying ovarian cancer development in a three-dimensional (3D) system has many advantages. Currently there is no stable ovarian CSC model, nevertheless, hanging drop culture method, 3D cultures with three-dimensional matrices or co-culture of ovarian cancer cells with mesothelial cells or multiple cell types provide for now an excellent platform to study adhesion, invasion, as well as tumor proliferation and differentiation 168, 169. These techniques require highly specified spheroid culture conditions with media containing growth factors such as LIF (leukaemia inhibitory factor), FGF (fibroblast growth factor), EGF (epidermal growth factor) and Insulin. 3-dimensional spheroids rather than monolayer cultures should be used to screen for anticancer drugs or oncolytic efficacy of viruses. In an in vitro model of epithelial ovarian cancer metastasis Tong et al. demonstrated different kinetics of Maraba virus (MRBV)-mediated killing of epithelial ovarian cancer (EOC) between monolayer and tumor spheroids 98. Both VSV and Maraba virus are classified as vesiculoviruses and they require low-density lipoprotein receptor (LDLR) for binding and entry of host cells 170. MRBV entry into EOC spheroid cells was significantly reduced in comparison to adherent cells and this effect was associated with decreased expression of LDLR protein in spheroids. However, after initial delay MRBV replicated within spheroids generating significant cell killing effects 97 Pampalakis et al. has found up-regulated of LDLR in ovarian carcinomas 171 that implies that VSV might targets ovarian cancer spheroids as well.

Concluding remarks

EOC metastasis occurs via the shedding of malignant cells from the primary tumor into the peritoneal cavity. Ovarian cancer stem cells (CSCs) within spheroids, commonly present in cancer associated ascites are believed to be a source of both chemotherapy resistance and metastases. Cancer cells multicellular spheroids are less sensitive to apoptosis. Therefore virotherapy with oncolytic virus penetrating deep into the hypoxic and acidic region of the 3D spheroids seems a quite appealing idea. Vesicular stomatitis virus (VSV) is a is an attractive oncolytic agent for cancer virotherapy. Since VSV is able to replicate in healthy cells, several approaches have been used to engineer a VSV that replicates preferentially in tumor cells. The potential of VSV to eradicate ovarian CSCs is not confirmed however more and more studies point out the key elements of antiviral response and signaling pathways that may improve VSV oncolysis. Therefore employing newly emerged, genetically engineered mouse models of human serous epithelial ovarian cancer (HGSOC), or with alteration of PI3K/AKT/mTOR pathway, and 3D tumor spheroid models for screening VSV oncolytic potential is necessary in order to be able to determine its real role as a promising approach to ovarian cancer therapy.