|Home | About | Journals | Submit | Contact Us | Français|
The preclinical optimization and validation of novel treatments for cancer therapy requires the use of laboratory animals. Although in vitro experiments using tumor cell lines and ex vivo treatment of patient tumor samples provide a remarkable first-line tool for the initial study of tumoricidal potential, tumor-bearing animals remain the primary option to study delivery, efficacy, and safety of therapies in the context of a complete tumor microenvironment and functional immune system. In this review, we will describe the use of murine tumor models for oncolytic virotherapy using vesicular stomatitis virus. We will discuss studies using immunocompetent and immunodeficient models with respect to toxicity and therapeutic treatments, as well as the various techniques and tools available to study cancer therapy with Rhabdoviruses.
The lack of options available for patients with chemotherapy-resistant, advanced, aggressive, or systemic disease is pushing the field of cancer therapy to progress rapidly, with novel strategies being developed. Of the many promising approaches that recently emerged, oncolytic viruses (OVs) have the potential of treating disseminated diseases, targeting various tumor types, and inducing durable systemic antitumor immune responses with minimal side effects compared with current therapies (Singh et al. 2012). OVs are selected or designed to specifically destroy cancer cells (Zeyaullah et al. 2012). Although there are many different OVs currently being investigated around the world, the focus of this review is on the Rhabdovirus family member vesicular stomatitis virus (VSV). VSV is a single-stranded, negative-sense RNA virus that can infect both insects and mammals (Kuzmin et al. 2009). These viruses are highly sensitive to the antiviral activity of interferon (IFN), and thus in normal cells, infection is controlled through the induction of the IFN response (Stojdl, Lichty, et al. 2000). However, several tumor cell lines (SK-MEL3, LNCAP, LC80, and OVCA 420), unlike normal cells (human ovarian surface epithelial cells, primary normal human prostate epithelium, and OSF7 cells), have defects in the IFN pathway, which allows Rhabdoviruses to selectively infect and subsequently kill these malignant cells (Hanahan 2000; Stojdl, Lichty, et al. 2000). Various recent reviews focus on the biology, neurotropism, genetic engineering, oncoselectivity, and safety of VSV and thus will not be discussed here (Hastie and Grdzelishvili 2012; Hastie, Cataldi, et al. 2013; Lichty et al. 2004).
Various animal models can be used to study OVs. Classically, safety studies for preclinical data are performed in nonhuman primates because they are genetically closer to humans, but for ethical reasons, tumors cannot be implanted in these animals, and this severely limits the use of this model for efficacy studies (Jenks et al. 2010; Johnson et al. 2007). Companion animals such as canines and felines can also accurately reflect some aspects of human disease because of the natural occurrence of cancer in these animals. However, the limited number of test subjects with similar tumor types constitutes one of the main limitations of this model. Alternatively, purpose-bred canine models have been used in dose-escalation studies, but again, the typically small number of subjects is considered a major limitation (LeBlanc et al. 2013). Importantly, it was reported that canines might be resistant to the neurotoxicity of OVs, which emphasizes the potential differences between species and the necessity of assessing the response to treatment in various models before testing in humans. Smaller animal models like rats and mice confer the undeniable ease of including more animals per experiment in genetically identical backgrounds. The availability of multiple chemical carcinogens and genetically modified mice that predicatively develop specific cancers, as well as tumor cell lines that can be implanted in vivo, provides a powerful set of tools for large-scale, well-controlled studies. Rat tumor models have been used to demonstrate safety and efficacy of OV treatments (Jenks et al. 2010; Kurisetty et al. 2014) but will not be addressed here. Also, promising alternatives for the in vivo study of human samples are chicken embryos (Jayachandran et al. 2015) and zebrafish (Chiavacci et al. 2014; Feitsma and Cuppen 2008), although, to our knowledge, these models have not been reported to be used for OV studies. Although each model has pros and cons, factors such as cost, housing, availability, biology, and ethics should be considered in the selection of the appropriate model that will best answer the specific questions of each study.
Various reviews summarize available mouse models for cancer therapeutics studies (Frese and Tuveson 2007; Olive and Tuveson 2006; Sharpless and Depinho 2006). These reviews describe in great detail the advantages and disadvantages of syngeneic immunocompetent, immunodeficient, and xenograft murine models. A wide variety of tumors can be implanted or induced in these animals in different organs. Many tumor cell lines can be injected subcutaneously to form a primary tumor, intravenously to colonize the lungs, or orthotopically to recapitulate the appropriate tumor microenvironment. In this section, we will discuss some of the more commonly used models in the study of VSV. Details are summarized in Table 1.
Subcutaneous tumor models provide the advantage of easy access to the tumor for treatment and direct monitoring. They are the most common models used to study OVs, with breast (4T1, TS/A, D1-DMBA3, D2F2, EMT6), colon (CT26, MC38), melanoma (B16), and renal (Renca) cancer cell lines having been previously described (Ahmed et al. 2010; Arulanandam, Batenchuk, Varette, et al. 2015; DeRose et al. 2011; Ebert et al. 2005; Edge et al. 2008; Fernandez et al. 2002; Garijo et al. 2014; Janelle et al. 2014; Jha et al. 2013; Kim et al. 2012; Kottke et al. 2008; Le Boeuf et al. 2010; Lemay et al. 2012; McCart et al. 2001; Moussavi et al. 2010; Nanni et al. 1983; Obuchi et al. 2003; Rintoul et al. 2012; Rommelfanger et al. 2012; Stephenson et al. 2012; Stojdl et al. 2003; Wongthida, Diaz, Galivo, et al. 2011). However, to recapitulate closely what occurs in the course of human disease, more clinically relevant sites of implantation are required.
Intravenous administration of tumor cells allows for colonization of the lungs and can also mimic systemic disease. Various cell lines can be injected intravenously as artificial models of lung metastasis. Of these, 4T1, B16, CT26, and TS/A have been used for oncolytic Rhabdovirus studies (Ebert et al. 2005; Fernandez et al. 2002; Lemay et al. 2012; Obuchi et al. 2003; Rintoul et al. 2012; Stephenson et al. 2012; Stojdl et al. 2003; Zhang et al. 2014). Also, hematologic cancer models such as leukemia/lymphoma and multiple myeloma are studied by intravenous administration. Of these, L1210, EL4, and 5TGM1 were used to assess efficacy of VSV virotherapy (Conrad et al. 2013; Goel et al. 2007). Finally, lung cancer cell lines (e.g., Lewis lung carcinoma) are established by intravenous injection as well (Hirasawa et al. 2003; Qiao et al. 2008). Intracardial tumor cell administration is another route that also mimics disseminated disease. Because most tumor cells lodge in the lungs after intravenous administration (Warren and Gates 1963), intracardiac injection allows the initial bypass of the lungs and permits tumor cells to circulate to other organs. Indeed, injection in the left ventricle of the heart can be used to establish tumors in the bones or brain (Campbell et al. 2012). This method allows one to recapitulate late-stage metastatic disease.
An alternative for the generation of systemic disease is the orthotopic engraftment of tumors using syngeneic or xenogeneic models. Indeed, it has been shown that orthotopic xenografts metastasize much more readily than subcutaneous xenografts and are thus better representative of human cancers (Hoffman 2015). This also allows the study of various aspects of cancer biology that may impact OV therapy, such as poor tumor perfusion/hypoxia, tumor-associated stromal elements, and immune cell infiltration (Clark and Vignjevic 2015; Vanharanta and Massagué 2013). To this end, tumors can be implanted directly in the brain, the pancreas, or the mammary fat pad. The injection of cells into the mammary fat pad is easy to perform in breast cancer models such as 4T1, E0771, and EMT6 and allows for a recapitulation of the metastatic route (lungs, liver, and bones) observed in human breast cancer patients (Pulaski and Ostrand-Rosenberg 2001). For other orthotopic sites such as the kidney (Renca), the pancreas (Panc02), the brain (U87, CT2A, CT26 and B16F10), the spleen, and the liver, injections can be more challenging. However, we found that the use of Hamilton syringes and needles specifically tailored for precise injections facilitates the successful delivery of cells (Devaud et al. 2013; Kato et al. 2014; Shevchenko et al. 2013; Tracz et al. 2014; Vonlaufen et al. 2008).
Transgenic models of cancer have the advantages of having predictable timing and known genetics, are syngeneic, and show similar histology and biology to human disease. Ovarian (MISIISTAg, Wv) (Arulanandam, Batenchuk, Varette, et al. 2015; Capo-chichi et al. 2010), prostate (PTEN−/−, TRAMP) (Moussavi et al. 2010; 2013), and pancreatic (KC) (Hastie, Besmer, et al. 2013) transgenic cancer models were previously used for VSV efficacy studies. Importantly, these models allow the study of common mutations in specific cancer types and how these may impact OV therapy (Balachandran et al. 2001).
Finally, chemical induction of tumors recapitulates human disease that results from exposure to mutagenic or inflammatory agents, as is the case with melanoma (UV light), lung carcinoma (tobacco), and hepatocellular carcinomas (chronic alcohol consumption and hepatitis C virus infections). Indeed, these diseases are highly heterogeneous and carry a high mutational load (Alexandrov et al. 2013). A wide variety of carcinogens are available for murine skin, lung, and liver cancer models (Nassar et al. 2015; Simanainen et al. 2015; Westcott et al. 2014). One of the main limitations of this approach is the lengthy period of time required to develop tumors. A thioacetamide-induced hepatocellular carcinoma model was previously used to study VSV in rats (Altomonte et al. 2013), but we could not find any report of chemically induced cancer models used for OV studies in mice.
OVs are more and more appreciated for their ability to induce tumor-specific immune responses, a property that is crucial for efficient control of tumors (Lichty et al. 2014). The use of athymic nude mice that lack T cells or SCID (severe combined immunodeficiency) mice that lack B cells and T cells and antibody-mediated depletion of specific immune cell populations are convenient tools to discriminate the immune compartments involved in the generation of antitumor immunity and OV efficacy (Bergman et al. 2007; Willmon et al. 2011; Wongthida, Diaz, Pulido, et al. 2011; Wongthida et al. 2010).
The in vivo study of human tumors is challenging. Although primary patient samples provide the obvious advantage of being relevant to human disease, their access and numbers are limited. These samples can be used for ex vivo analysis and allow for a wide range of conditions to be tested. We previously published a detailed visual protocol describing the ex vivo study of VSV using tumor cores (Diallo et al. 2011). Unfortunately, the ex vivo study of biopsies does not reflect the conditions present in vivo. In addition, the engraftment of human tumor cell lines into mice presents the challenge of xeno-rejection and thus requires the use of immunocompromised animals. One important factor to consider using immunodeficient mice for OV studies is their increased susceptibility to viral infection. Indeed, the lethal dose of virus was shown to be drastically reduced for immunocompromised animals (Huneycutt et al. 1993; Thomsen et al. 1997). Most human cell lines are grown in nude or SCID mice (Carreno et al. 2009; Choi et al. 2014; Simeoni et al. 2013). These animals have been used for VSV studies using human melanoma (SK-MEL-3), plasmacytoma (KAS 6/1), glioblastoma (U87), colon (SW620, A549, HT29), prostate (DU145, PC3), pancreatic (Miapaca-2, Panc 03.27), ovarian (ES-2, OVCAR8), kidney (ACHN, 786-O), cervical (HeLa), and tongue (SCC25) cancers (Ayala-Breton et al. 2013; Blackham et al. 2014; Breitbach et al. 2007; Edge et al. 2008; Ilkow et al. 2015; Jha et al. 2013; Le Boeuf et al. 2010; 2012; Muik et al. 2014; Stojdl et al. 2003; Stojdl, Lichty, et al. 2000; Zhao et al. 2014).
Other immunodeficient animals, such as NOD (nonobese diabetic) SCID mice are used for tumor models that are more easily rejected, as well as for patient-derived samples (Kim et al. 2009). These mice are highly immunocompromised; they lack T, B, and functional NK cells and thus require special housing and handling under specific pathogen-free conditions. They are more expensive and difficult to work with regarding surgery/anesthesia, and complications such as infections by environmental pathogens can arise (Kim et al. 2009). To increase the chances of successful engraftment of patient samples into NOD SCID mice, matrigel or cultrex can be coinjected (Benton et al. 2011). These products are a mix of growth factors that stimulate tumor growth and basement membrane matrix proteins that structurally form a scaffold to support tumor formation. By providing these key components of the natural tumor microenvironment, these mixtures were shown not only to favor xenograft development but also to increase tumor growth rate and metastatic potential (Fridman et al. 2012). Pancreatic and breast cancer patient samples have been successfully engrafted and passaged for VSV studies using these matrixes (DeRose et al. 2011; Ilkow et al. 2015) (Figure 1).
Many options exist for delivering virus in vivo, with the common routes being intravenous, intratumoral, intranasal, intraperitoneal, and subcutaneous. In efficacy studies, intravenous and intratumoral injections are the most common and clinically applicable methods, whereas intravenous, intranasal, and intracranial injections are most often used for toxicity studies (Stojdl et al. 2003). The intraperitoneal route, which is well suited to treat diseases such as mesothelioma and peritoneal carcinomatosis, and subcutaneous route of delivery are less commonly used and thus will not be discussed here. The injection methods and the models in which they were used are summarized in Table 1.
Direct intratumoral injections are used to administer the virus to subcutaneous and to surgically implanted tumors, such as those in the brain or liver. Immediate local delivery is achieved, with the full dose of virus being delivered to the site of interest. The level of difficulty is low, but precautions must be taken to avoid leakage and subsequent loss of viral dose as a consequence of high interstitial fluid pressure within the tumor (Jain 1990). This route is used to administer OVs in some clinical trials and thus remains relevant to the treatment of human disease (Heo et al. 2013). Although intratumoral injections are easy, they limit delivery of the treatment to the primary site of injection and can be difficult to perform in some anatomical locations such as kidneys, lungs, brain, or ovaries without surgical intervention or the aid of sophisticated equipment (i.e., ultrasound, stereotactic unit).
In mice, systemic delivery of virus by intravenous injection is often performed by the lateral tail vein. Intravenous injection allows for the delivery of virus to primary tumors and also to metastases (Breitbach et al. 2007; Fisher 2006). Unfortunately, drawbacks include limitations in the maximum volume delivered (approximately 100 μl) and the challenge of administering multiple doses, especially when the animals are dehydrated as a result of the disease or side effects of the treatment. Other factors to consider with intravenous delivery are the low efficiency of delivery to the tumor and the increased risk of toxicity to normal tissues because the virus is delivered systemically. Although a single bolus of virus injected intravenously is the simplest approach, gradual delivery, which mimics a clinical scenario, can be performed using various kinds of pumps. Programmable pumps are affordable and allow for slow infusion through the tail vein through a catheter. This technique was previously used to administer various substances to mice and other animals (Matsuo et al. 1998; Nilaver et al. 1995). Another type of device known as an osmotic pump requires surgical implantation and has mostly been used to deliver drugs over a specified time course and rate (Mirandola et al. 2011; Shibata et al. 2008).
Intranasal injections are easy to perform through direct administration into the nares of the animal. Mice must be anesthetized to minimize chances of sneezing, which could result in loss of virus during the procedure (Wollmann et al. 2010). When administered intranasally, the virus will typically infect the brain, olfactory system, and respiratory system (Ozduman et al. 2008). This route has been previously used to study VSV-associated neurovirulence (Edge et al. 2008; Stojdl, Abraham, et al. 2000; Wollmann et al. 2015).
Studying how OVs interact with the host and the tumor can provide valuable insight on their mechanism of action and also reveal important information on how to improve their safety and effectiveness. The ability to monitor virus replication, both quantitatively and qualitatively, is invaluable in this regard. Various techniques are available to analyze the presence of virus in tumors and normal tissues. Virus titration, quantitative polymerase chain reaction (qPCR), bioluminescence/fluorescence imaging, and immunohistochemistry (IHC) are commonly used methods. Whereas titration by plaque assay measures infectious virus particles (plaque forming units [PFUs]), qPCR quantifies genome copies and thus is not a measure of infectivity. Both techniques have been used in several studies using VSV (Breitbach et al. 2007; Eisenstein et al. 2013). On the other hand, bioluminescence and fluorescence imaging requires the use of virus variants that express fluorescent proteins or luciferase. Instead of directly quantifying the virus, this technique visualizes and measures the expression of a reporter transgene and can be performed on anesthetized animals, which confers the advantage of allowing for multiple readings without having to kill the animals, and is therefore ideally suited for kinetic analyses of virus replication. Bioluminescence imaging using the In Vivo Imaging System has been used in various VSV studies (Arulanandam, Batenchuk, Varette, et al. 2015). Figure 2 shows an example of bioluminescence imaging of a VSV-expressing luciferase in various tumor models and anatomical locations. A similar approach is to use a virus variant encoding a sodium iodide symporter gene (NIS). Using this technique, the animals must be injected with radioactive iodine before imaging using a small animal PET/CT scan (Ayala-Breton et al. 2013; Goel et al. 2007; LeBlanc et al. 2013). Another way to visualize VSV is by IHC. This method requires the use of an antibody against the virus and takes more time to perform compared with other methods described but provides the advantage of characterizing the localization of the virus within the tumor or tissue of interest. Virus localization can also be correlated with other markers of interest, such as markers of apoptosis (Breitbach et al. 2007). IHC for VSV was performed in various murine and human tumor models. An example of IHC performed in a xenograft model of human breast cancer treated with VSV is provided in Figure 1. Finally, the window chamber is an apparatus that can be surgically placed on a mouse for the visualization of the formation of tumors and their associated blood vessels in real time. Both dorsal and mammary window chambers have been used and allow for intravital microscopy using fluorescence imaging (Palmer et al. 2012). The window chamber has been used to study various established human and murine tumor cell lines, such as breast (MDA-MB-231 and 4T1) (Jin et al. 2014; Rajaram et al. 2013) and renal (Caki-2) carcinoma models, for visualization of the tumor microvasculature (Biel et al. 2014). Visualization of a fluorescent version of vaccinia, another OV, within the tumor vasculature was previously performed using the window chamber model (Arulanandam, Batenchuk, Angarita, et al. 2015), but to our knowledge, there are no published reports using this technique to study VSV.
One of the main advantages of using animal models over in vitro studies is the presence of a complete tumor microenvironment. The different constituents of the tumor microenvironment, which include tumor blood vessels, cancer-associated fibroblasts, tumor-infiltrating immune cells, and extracellular matrix proteins, all interact with OVs and can dramatically influence the delivery and efficacy of OV therapeutics. For example, it has been shown that intravenous delivery of VSV results in the infection of tumor vasculature, which results in vascular shutdown, a phenomenon that slows down tumor growth (Breitbach, De Silva, et al. 2011). This has also been seen in patients after OV therapy (Breitbach, Burke, et al. 2011). The mechanism of selective infection of tumor endothelium, but not normal endothelium, involves the secretion of vascular endothelium growth factor (VEGF) by the tumor endothelium, which subsequently induces the expression of PRD1-BF1/Blimp1 to increase viral replication in blood vessels (Arulanandam, Batenchuk, Angarita, et al. 2015). Cancer-associated fibroblasts, another cellular component of the tumor microenvironment, were recently demonstrated to sensitize tumors to OV infection through the secretion fibroblast growth factor 2 (FGF2) (Ilkow et al. 2015). Tumors are often infiltrated by several immune cell types, and these can influence virus replication as well as modulate antitumor activity. For example, one study demonstrated that inhibition of chemokine secretion by an engineered VSV resulted in a decrease in the recruitment of neutrophils and NK cells to the tumor after virus infection, and this led to a dramatic increase in virus replication and efficacy (Wu et al. 2008).
Noncellular components of the tumor microenvironment can also impact OV therapy. It has been demonstrated that interstitial collagen restricts the spread of herpes virus within the tumor, and therefore exogenous expression of collagenase was shown to increase virus spreading within the tumor, which improved efficacy (McKee et al. 2006). Subsequent strategies to successfully improve OV cell-to-cell spread or diffusion within tumors, all of which target various extracellular matrix components for degradation, include coadministration with trypsin (various targets) or hyaluronidase (hyaluronan) and encoding matrix metalloproteinases or relaxin (both degrade collagen) within the virus (Cheng et al. 2007; Ganesh et al. 2008; Kim et al. 2006; Kuriyama et al. 2000).
The circulatory system is a hostile environment for VSV. Circulating antibodies, complement nonspecific binding to circulating cells and serum components, as well as uptake by specialized liver macrophages known as Kupffer cells, all of which contribute to minimizing the effective dose of virus delivered to tumors (Evgin et al. 2015; Ferguson et al. 2012). Given the challenges of systemic delivery of OVs in the face of a host immune system that is highly evolved to recognize and eliminate viruses, our group and others have explored different strategies to overcome these barriers. It has been demonstrated that the depletion of Kupffer cells using clodronate liposomes successfully increased systemic delivery of oncolytic adenovirus to tumors (Shashkova et al. 2008). Another approach is to use cells as delivery vehicles, or Trojan horses, for OVs (Bell and Roy 2013; Willmon, Harrington, et al. 2009). Internalization of the virus by the cell carrier not only protects the virus from immune recognition and neutralization, but the virus can also replicate within the cell carrier, thus increasing the OV dose delivered to tumors. Several cell types have been explored as carriers for the delivery of OVs, including tumor cells, T cells, dendritic cells, insect cells, and myeloid-derived suppressor cells (Eisenstein et al. 2013; Ilett et al. 2011; Kottke et al. 2008; Power et al. 2007; Roy et al. 2015). Importantly, we have shown that cell carriers can deliver VSV to tumors in the presence of high levels of VSV-neutralizing antibodies, whereas naked virus was efficiently neutralized (Power et al. 2007). Such strategies allow for multiple doses of virus to successfully be delivered to the tumor, therefore increasing the therapeutic effects.
One of the main concerns regarding systemic administration of virus is the risk of toxicity associated with infection of normal tissues. This risk is even greater using more potent viruses with increased cytotoxic activity. Although virus replication in nontumor cell lines can be tested in vitro, the most relevant way to assess tumor specificity is by determining the biodistribution of the virus after treatment. Typically, organs are harvested after treatment, and the virus is quantified using the techniques described previously. To minimize toxicity to normal cells and increase safety, attenuated versions of VSV have been developed. VSVΔ51, AV1, and AV2 (Stojdl et al. 2003), as well as a VSV variant expressing IFNβ (Willmon, Saloura, et al. 2009), are a few examples. An alternative approach for improving the safety profile of OVs is through microRNA detargeting. By encoding the target sequence of microRNAs that are expressed in normal tissues, but not in tumor tissues, within the viral genome, tumor selectivity can be improved (Edge et al. 2008; Kelly et al. 2008).
Knockout mice can also provide valuable information into the mechanisms of virus clearance by the host. A good example is the PKR-/- mouse, which lacks a gene involved in the induction of the IFN antiviral response (Abraham 1999). These mice are hypersensitive to VSV infection, with a dose of less than 10 PFUs resulting in neurotoxocity, whereas attenuated mutants were tolerated at doses higher than 1 × 107 PFUs (Stojdl et al. 2003). The IFN α and β receptor null mice have also been used for toxicity studies because they are more susceptible to viral infections. Previous results showed that, whereas wild-type mice were unaffected by 50 PFUs of VSV, the knockout mice succumbed to infection within 3 to 6 days (Muller et al. 1994). Similar results were obtained using another mouse strain that is defective in the IFNα/β pathway: the STAT1 knockout (Meraz et al. 1996).
One mechanism by which OVs eliminate tumors is by their capacity to stimulate the immune system. It is now accepted that OVs like VSV induce antitumor immunity, therefore providing a long-lasting systemic protection (Mahoney and Stojdl 2013). In an attempt to improve this aspect of OV therapy, various viruses have been engineered to encode immune-stimulatory molecules. These were recently reviewed elsewhere and will not be listed here (Lichty et al. 2014). Standard immunology techniques such as flow cytometry and ex vivo restimulation with specific antigens can be used to study the antivirus and antitumor immune responses. Interestingly, for some antigens, the specific T cell epitopes are known, which allows for ex vivo restimulation with peptides instead of whole cells or purified molecules. For example, the B16 melanoma cell line expresses the well-characterized dopachrome tautomerase molecule (DCT), for which the exact sequence presented at the cell surface is known. Tumor cell lines can also be engineered to stably express foreign antigens such as ovalbumin (OVA) for which the epitopes presented at the cell surface are known (Rötzschke et al. 1991), thereby facilitating the study of OV-induced immune responses. Similarly, research tools, such as the OTI and OTII mice, which have peripheral T cells that are specific for ovalbumin, are available. The B16-OVA tumor model and OTI mice were previously used to study antitumor immunity after VSV treatment (Kottke et al. 2008).
The methods used to assess efficacy after OV treatment are the same as for other cancer therapies. Tumor growth is assessed by calculating the tumor volume using caliper measurements of length and width of the tumors (Edge et al. 2008; Fernandez et al. 2002). For lung tumors, the quantity and size of lung metastasis can be assessed by weighing the lungs, perfusing them with India ink, or directly visualizing them (Ahmed et al. 2010; Shi et al. 2009; Shibata et al. 2008; Stojdl et al. 2003). Of course, overall survival can also be assessed with Kaplan-Meier survival curves (Qiao et al. 2006; Stojdl et al. 2003). Several studies using these methods are reported in Table 1.
With the objective of improving the efficacy of OV therapy, several groups have combined viruses with other cancer therapies (Ottolino-Perry et al. 2010). The combination of radiotherapy with VSV was shown to be one of many successful approaches (Alajez et al. 2012). Also, various chemotherapeutic agents and other drugs such as sunitinib, doxorubicin, gemcitabine, bortezomib, cyclophosphamide, cisplatin, 5-fluorouracil, triploid, histone deacetylase inhibitors, smac mimetics, and a Bcl-2 inhibitor provided additional survival advantages with VSV treatment in different tumor models, suggesting that these treatment modalities are not incompatible (Ben Yebdri et al. 2013; Beug et al. 2014; Hastie, Besmer, et al. 2013; Jha et al. 2013; Leveille et al. 2011; Nguyên et al. 2008; Porosnicu et al. 2003; Samuel et al. 2013; Schache et al. 2009; Sung et al. 2008; Willmon et al. 2011; Yarde, Nace, et al. 2013, Yarde, Naik, et al. 2013). Alternatively, several compounds dubbed virus sensitizers were shown to increase viral replication and also prolong survival of tumor-bearing mice when administered in combination with VSV (Arulanandam, Batenchuk, Varette, et al. 2015; Diallo et al. 2010). Another strategy that was proven successful is the combination of different OVs. It was shown that coadministration of vaccinia virus, an OV that encodes immunoregulatory genes that can increase VSV replication, led to increased efficacy when compared with either virus alone (Le Boeuf et al. 2010). One factor that was identified as important for this effect is the protein B18R encoded by vaccinia virus. Accordingly, better efficacy was also observed using a VSV encoding B18R as a single agent (Le Bœuf et al. 2013).
Another approach is the combination of VSV with vaccination strategies. VSV is a well-studied vaccine vector and is recognized for its ability to stimulate both innate and adaptive immunity (McKenna et al. 2003). In a previous study, we demonstrated that the administration of VSV-infected B16F10 cells was a successful strategy to vaccinate mice against the same tumor type (Lemay et al. 2012). In another study using the L1210 leukemia cell line, we demonstrated that irradiated virus-infected cells could achieve the same results (Conrad et al. 2013). It was also demonstrated using the B16F10 tumor model that VSV could be used as a vaccination platform. Indeed, use of heterologous viruses that encoded the same tumor-associated antigen to prime and boost the antitumor immune response achieved a significant prolongation of survival (Bridle et al. 2010). Another means to stimulate the immune system is the use of immune checkpoint inhibitors. By depleting/blocking immunosuppressive regulatory T cells using antibodies in combination with VSV, measles virus, Newcastle disease virus, and reovirus, efficacy was improved (Engeland et al. 2014; Gao et al. 2009; Rajani et al. 2016; Zamarin et al. 2014).
With the first OV soon to be approved for use in patients upon the conclusion of a phase III study in advanced melanoma patients (Andtbacka et al. 2015) and various other clinical trial candidates, including VSV and Maraba virus, currently being evaluated, the field of OV therapy is rapidly evolving, and novel viruses and strategies are constantly being developed. The myriad in vivo tumor models and various techniques described herein allow for the validation of these novel treatments, facilitating the generation of preclinical safety data and evaluation of efficacy compared with currently available options.
This work was supported by the (TFF 122868), the Canadian Cancer Society Research Institute, the Ontario Institute for Cancer Research, the Canadian Institute for Health Research (CIHR), the Ottawa Regional Cancer Foundation, and the Ottawa Hospital Foundation. M.C.B.D and D.G.R were supported by the (MFI-140935 and GSD-121796). The animals in this study were used and euthanized humanely in accordance with the Guide for the Care and Use of Laboratory Animals.