1252 AJVR, Vol 73, No. 8, August 2012

Novel cancer treatments are needed to minimize deleterious effects and improve outcomes. Target-

ed lysis of cancerous cells by viruses without lysis of healthy cells (ie, oncolytic virotherapy) is a promising treatment option that may be more effective and have fewer adverse effects than traditional cancer treatments. A key factor in the success of oncolytic virus treatments is the susceptibility of neoplastic cells to viral infection because of mutations in genes required for function of antiviral defense mechanisms.1 Some of these muta-tions also alter cell signaling pathways, which confer

Oncolysis of canine tumor cells by myxoma virus lacking the serp2 gene

Ashlee S. Urbasic, DVM; Stacy Hynes, DVM; Amy Somrak, DVM; Stacey Contakos, DVM, MS; Masmudur M. Rahman, PhD; Jia Liu, PhD; Amy L. MacNeill, DVM, PhD

Objective—To determine the oncolytic efficacy of an attenuated form of myxoma virus lacking the serp2 gene in canine tumor cells.Sample—Primary cells were isolated from tumors that were surgically removed from dogs and from connective tissue obtained from the cadaver of a dog. Cells of various established cell lines from tumors and nontumorous tissues were obtained.Procedures—Experiments were performed with cells in monolayer culture. Cell cultures were inoculated with wild-type myxoma viruses or myxoma viruses lacking the serp2 gene, and measures of cytopathic effects, viral growth kinetics, and cell death and apoptosis were determined.Results—Myxoma viruses replicated in cells of many of the primary and established canine tumor cell lines. Canine tumor cells in which expression of activated protein kinase B was upregulated were more permissive to myxoma virus infection than were cells in which expression of activated protein kinase B was not upregulated. Myxoma viruses lacking the serp2 gene caused more cytopathic effects in canine tumor cells because of apoptosis than did wild-type myxoma viruses.Conclusions and Clinical Relevance—Results of the present study indicated myxoma viruses lacking the serp2 gene may be useful for treatment of cancer in dogs.Impact for Human Medicine—Results of the present study may be useful for development of novel oncolytic treatments for tumors in humans. (Am J Vet Res 2012;73:1252–1261)

selective growth advantages on cells.2,3 For example, cells of many tumors have upregulated expression of phosphorylated Akt, which contributes to altered gene

Received February 28, 2011.Accepted August 30, 2011.From the Department of Pathobiology, College of Veterinary Medi-

cine, University of Illinois, Urbana, IL 61802 (Urbasic, Hynes, Som-rak, Contakos, MacNeill); and the Department of Molecular Genet-ics and Microbiology, College of Medicine, University of Florida, Gainesville, FL 32611 (Rahman, Liu).

Supported by University of Illinois College of Veterinary Medicine Department of Pathology, the University of Illinois College of Vet-erinary Medicine Companion Animal Grants Memorial Fund, the Wayne D. and Josephine Spangler Fund, and the National Center for Research Resources (National Institutes of Health, T35 RR020292).

Presented as an oral presentation at the American College of Veteri-nary Pathologists–American Society for Veterinary Clinical Pathol-ogy Annual Meeting, Monterey, Calif, June 2010.

The authors thank Drs. Gail Scherba and Joanna Shisler for assistance with reporting of data and Karen Kelly for performance of transmis-sion electron microscopy.

Address correspondence to Dr. MacNeill (almac@illinois.edu).

AbbreviAtions7-AAD 7-aminoactinomycin DAkt Protein kinase BDEVD-AMC Aspartic acid-glutamic acid-valine- aspartic acid–7-amino-4- methylcoumarin fluorogenic caspase 3 substrateCPE Cytopathic effectFBS Fetal bovine serumFITC Fluorescein isothiocyanateMEM Minimum essential mediumMOI Multiplicity of infectionMYXV Myxoma virusMYXV∆serp2 Attenuated myxoma virus lacking a viral serine proteinase inhibitorMYXV-gfp Recombinant myxoma virus expressing green fluorescent proteinMYXV-red Recombinant myxoma virus expressing tandem dimer tomato red fluorescent proteinserp2 Serine proteinase inhibitor 2 from myxoma virusVACV Vaccinia virus

11-02-0072r.indd 1252 7/20/2012 1:25:39 PM

AJVR, Vol 73, No. 8, August 2012 1253

regulation, decreased apoptosis, and increased prolif-eration of tumor cells4 and enables productive infection of tumor cells by some oncolytic viruses.5

Poxviruses have characteristics that make them excellent candidates for use in oncolytic virotherapy. These viruses preferentially infect cells in tissues with permeable vasculature,6–8 which is a feature of many ag-gressive tumors. Additionally, they can be genetically engineered to express antitumor agents without con-straints on gene size, which is a limitation of many oth-er viral vectors used for oncolysis.9 Poxvirus infection elicits strong cell-mediated immune responses, which cause lysis of infected cells,10–12 and effective humoral immune responses, which prevent establishment of la-tent or recurrent viral infections.11,13

One poxvirus, MYXV, causes a lethal disease called myxomatosis in European rabbits (Oryctolagus cunicu-lus).14 Myxoma virus is nonpathogenic in all other ver-tebrate species investigated,15–21 including humans and dogs. Myxoma virus productively infects cells of mul-tiple human tumor cell lines in vitro.22 Its effectiveness as an oncolytic virus is supported by findings of other studies in which MYXV treatment was determined to be efficacious in mice with xenografts of human glioma23 and medulloblastoma.24 In immunocompetent ani-mals, MYXV oncolysis is more efficacious when used in combination with the chemotherapeutic drug rapamycin.25,26

Administration of chemotherapeutic drugs to ani-mals with tumors might not be necessary if effective oncolytic viruses were available. As with all poxvirus-infected cells, MYXV-infected cells express several immunomodulatory proteins, some of which inhibit apoptosis. Although beneficial for viral pathogenesis during natural infections, expression of such genes is likely to hinder MYXV-based oncolysis. Thus, modifi-cation or deletion of antiapoptotic virulence factors in viruses may enhance apoptosis of tumor cells infected by those viruses. Indeed, VACV vectors that do not have the antiapoptosis genes SPI-1 and SPI-2 have en-hanced oncolytic specificity and efficacy in mice with adenocarcinoma.27 However, because VACV is a patho-gen of humans and mice, such mutant VACVs replicate in nontumorous cells in several organs, including the brain.27 Unlike VACV, MYXV does not cause disease in any species of animal except rabbits, which makes pro-duction and use of a recombinant MYXV with low rep-lication efficiency unnecessary. Therefore, we wanted to determine the oncolytic properties of a recombinant MYXV that lacks expression of serp2, an antiapoptotic virulence factor. This virus causes milder disease in rabbits than wild-type MYXV but replicates at the same rate in rabbit cells in vitro and in vivo.28,29

Testing of viruses for oncolytic efficacy is typically performed with mice that have induced or implanted neoplasms. This may not be the best method to evalu-ate oncolytic viruses for treatment of naturally develop-ing tumors in humans. Investigation of oncolytic effica-cy of viruses in dogs may be a superior method because dogs, like humans, naturally develop neoplasms. Fur-thermore, because dogs live in close contact with hu-mans, neoplasms in dogs and humans that are caused by environmental factors may have similar biological

behaviors. The purpose of the study reported here was to determine the oncolytic efficacy of MYXV in cells of established canine tumor cell lines and in primary ca-nine tumor cells obtained from tumor explants and to compare results with those for MYXV∆serp2 to deter-mine whether MYXV-mediated oncolysis is enhanced by removal of the serp2 gene.

Materials and Methods

Sample—Cells from the following established cell lines were obtained: rabbit kidney epithelial RK-13,a hu-man malignant glioma U-118 MG,b canine osteosarco-ma Abrams,c canine transitional cell carcinoma Bliley,c canine osteosarcoma D-17,c canine hemangiosarcoma Den,c and canine hemangiosarcoma Fitz.c In addition, primary tumor cells were isolated from naturally devel-oping tumors surgically removed from dogs that were patients at the University of Illinois Veterinary Teach-ing Hospital. These tumors included 2 mast cell tumors and 1 each of hemangiosarcoma, soft tissue sarcoma, gastrointestinal stromal tumor, hemangioma, perianal adenocarcinoma, perianal adenoma, mixed mammary gland tumor, and renal carcinoma. Cells were also iso-lated from noncancerous fibrous connective tissue ob-tained from the cadaver of a dog undergoing necropsy for reasons unrelated to this study. Unfixed portions of each tumor surgically removed from dogs were dissect-ed free of fat and necrotic tissue. These tumor sections were then minced into 1-mm cubes, placed in a 25-cm2 cell culture flask,d and incubated at room temperature (approx 22°C) for 10 minutes to allow tissue to adhere to bottoms of the flasks. Then, minimal essential cell culture media with Earle salts, 2mM glutamine, 50 U of penicillin G/mL, 50 µg of streptomycin/mL, 1mM sodium pyruvate, and 0.1mM nonessential amino ac-ids (MEM)e containing 10% FBSf were added and flasks were incubated (37°C, 5% CO

2, and 100% humidity).

Cells from noncancerous fibrous connective tissue (ie, fibroblasts) obtained from the cadaver of a dog were isolated in a similar manner. Trypsinization and pas-sage of cultured cells was performed when cells were 95% to 100% confluent or sooner if tissue explants were degrading. Canine primary tumor and fibroblast cell cultures and cells from established cell lines were maintained in MEM with 10% FBS at 37°C, 5% CO

2,

and 100% humidity.

Identification of tumor and cell types—Portions of each tumor surgically removed from dogs were fixed in neutral-buffered 10% formalin and submitted to the University of Illinois Veterinary Diagnostic Laboratory for histologic analysis. Tumors were paraffin embed-ded, sectioned, stained with H&E, and evaluated to identify tumor types.

Primary tumor cells isolated from tumors of dogs were cultured in 35-mm-diameter platesd until they were 90% to 100% confluent. Cells were trypsinized, collected in MEM with 10% FBS, and pelleted by cen-trifugation at 400 X g for 5 minutes. Cell pellets were suspended in 1 mL of PBS solution, and 100-µL ali-quots of cells were cytocentrifuged at 400 X g for 3 min-utes onto charged glass slides. To ensure cultured cells were representative of the types of tumors from which

11-02-0072r.indd 1253 7/20/2012 1:25:39 PM

1254 AJVR, Vol 73, No. 8, August 2012

they were isolated (rather than adjacent nonneoplas-tic tissue), cells adhering to slides were evaluated after staining with Wright-Giemsa and cytochemical stains.

Virus isolation—Isolation and characteriza-tion of MYXV∆serp2 (previously referred to as MYX∆SERP2::lacZ) was performed as described.29,30 Two recombinant MYXVs (MYXV-gfp and MYXV-redg [previously referred to as vMyxgfp31 and vMyx-tdTr,32 respectively]) were also isolated; regulation of each fluorescent protein in these recombinant viruses was under the control of a synthetic vaccinia virus early-late promoter.

Determination of CPEs of viruses—Primary ca-nine tumor cells and fibroblasts were inoculated with viruses. For comparison with results for canine pri-mary tumor cells, RK-13 (fully permissive for virus replication)33 and U-118 MG (semipermissive for virus replication)34,b cells were inoculated with viruses. Cells were grown to confluence in 35-mm-diameter platesd and inoculated with MYXV, MYXV-gfp, MYXV-red, or MYXV∆serp2 at an MOI of 0, 0.1, 1, 5, or 10 infectious virus particles/cell. A low MOI of 0.1 infectious virus particles/cell was included to simulate likely in vivo virus particle-to-tumor cell ratios because sufficient numbers of infectious virus particles cannot be deliv-ered homogenously in a solid tumor in vivo to ensure a high MOI. The MYXV-gfp and MYXV-red (recombinant viruses that express fluorescent proteins throughout the life cycle) were used to aid identification of virus-infected cells. Every 24 hours after inoculation, phase and fluorescent microscopic images of cells were ac-quired with a microscopeh (magnification, 100X) and software.i Fluorescent images of MYXV-gfp–inoculated cells were acquired with a 480/40-nm bandpass excita-tion filter and an exposure time of 1,500 milliseconds. Fluorescent images of cells inoculated with MYXV-red were acquired with a 560/40-nm bandpass excitation filter and an exposure time of 400 milliseconds. Images were evaluated, and CPE scores of 1+ (≥ 50% of cells adherent to culture plate), 2+ (10% to 50% cells adher-ent to culture plate), or 3+ (< 10% of cells adherent to culture plate) were determined for each cell culture.

Akt expression in cells—Because increased ex-pression of phosphorylated Akt is associated with greater susceptibility to MYXV-induced CPEs in human cancer cells,35 western immunoblot analysis of lysates of cells of some of the canine primary tumors (hem-angiosarcoma and soft tissue sarcoma) and established canine tumor cell lines (hemangiosarcoma Den and transitional cell carcinoma Bliley) was performed to determine relative expression of Akt phosphorylated at serine 473; these cells were not inoculated with viruses. For comparison, protein extracted from human glioma U-118 MG cells, a cell line that moderately expresses phosphorylated Akt,34 was also assayed. Cells were grown in 35-mm-diameter plates to 95% confluency and then collected in cell lysis buffer with protease and phosphatase inhibitor.j Total protein concentration was determined with a standard Bradford assay, and 50 µg of protein from each cell lysate was analyzed via SDS-PAGE (10%). The SDS-PAGE–separated proteins were transferred to a nitrocellulose membrane.k Membranes

were incubated in blocking buffer overnight (approx 16 hours) at 4°C, then incubated with rabbit antiphospho-Akt (serine 473) monoclonal antibodyl in accordance with the manufacturer’s protocol. Membranes were in-cubated for 1 hour at room temperature with a horse-radish peroxidase–conjugated goat anti-rabbit IgG polyclonal antibodym diluted 1:2,000 in blocking buf-fer. A chemiluminescent western blot detection kitk was used to detect antigen-antibody complexes. Membranes were stripped in accordance with the protocol provided by the manufacturer of the detection kit,k incubated with rabbit anti-Akt monoclonal antibodyl following the manufacturers’ instructions, and antigen-antibody complexes were detected.k Experiments were repeated 3 times. Immunoblots were scanned and analyzed with photo editing software.n Images were inverted (ie, gray-scale values were reversed), and integrated density mea-surements were determined for standardized regions of interest including each protein band. Density values for phosphorylated Akt were normalized relative to density values for total Akt to determine mean percentage of phosphorylated Akt in each cell sample.

As a measure of virus maturation, expression of an MYXV protein (M130R) that is expressed during late stages of viral infection of cells was determined. Cells were collected 24 hours after inoculation with MYXV (MOI, 5 infectious virus particles/cell) and processed as described for determination of Akt expression, except that 25 µg of protein from each sample was separated by SDS-PAGE, transferred to nitrocellulose membranes, and probed with rabbit polyclonal anti-M130R primary antibodies.g

Virus growth curves—Growth curve experiments were performed to determine whether MYXV could replicate in canine tumor cells and produce infectious viruses. To determine 1-step virus growth curves, cells of established canine tumor cell lines (hemangiosar-coma Den and transitional cell carcinoma Bliley) and primary canine tumor (hemangiosarcoma), primary ca-nine fibroblast, and RK-13 cells were grown to conflu-ence in 35-mm-diameter plates, inoculated with MYXV (MOI ≥ 5 infectious virus particles/cell), and incubated at 37°C in 5% CO

2 for 1 hour. Virus inoculum was

removed, and cells were washed with PBS solution 3 times. Then, 1.5 mL of MEM with 10% FBS was added to each plate. Cells were harvested with a sterile hand-held flexible natural-rubber scraper attached to a glass rod (ie, rubber policeman) 4, 8, 12, 24, and 48 hours later and centrifuged (400 X g for 5 minutes). Cell pel-lets were resuspended in 1 mL of MEM without serum and stored at –80°C until determination of virus titers. Multistep growth curve experiments were performed to determine whether MYXV could spread from infected cells to adjacent noninfected cells in cultures. Multistep growth curve experiments were performed in a manner similar to that used for determination of 1-step growth curves, except growth curves were not determined for canine fibroblasts but were determined for canine pri-mary soft tissue sarcoma cells, an MOI of 0.1 infectious virus particles/cell was used, and cells were harvested 0 (ie, 1 hour after virus inoculation), 12, 24, 48, 72, and 96 hours after MEM with 10% FBS was added to each plate.

11-02-0072r.indd 1254 7/20/2012 1:25:40 PM

AJVR, Vol 73, No. 8, August 2012 1255

To determine virus titers for 1-step and multistep MYXV growth curves, tumor cell culture samples were frozen and thawed 3 times, sonicated for 1 to 5 min-utes in a water batho at room temperature, and a series of 10-fold dilutions were prepared. Four hundred mi-croliters of each diluted sample was added to conflu-ent RK-13 cells in 35-mm-diameter plates. Cells were incubated (37°C in 5% CO

2 for 1 hour) with agitation

every 15 minutes, and a 3-mL overlay (1:1 mixture of 1% high–gelling temperature agarosep solution and 2X MEM with 20% FBS; heated to 45°C) was added to the plates. Plates were incubated at 37°C in 5% CO

2 for 5

to 7 days. Plaques formed by MYXV infection of cells (small white foci) were counted. Mean ± SEM virus ti-ters were calculated for results of ≥ 3 experiments and expressed as the number of plaque-forming units per milliliter.

Transmission electron microscopy—To determine whether MYXVs in all stages of morphogenesis were present in cells, MYXV-inoculated RK-13, hemangio-sarcoma Den, transitional cell carcinoma Bliley, osteo-sarcoma Abrams, osteosarcoma D-17, and primary ca-nine tumor hemangiosarcoma and renal carcinoma cells were evaluated via transmission electron microscopy. Twenty-four hours after cultured cells were inoculated with MYXV (MOI, 5 infectious virus particles/cell), media and nonadherent cells were collected. Then, adherent cells were removed by trypsinization, pooled with nonadherent cells, and centrifuged (400 X g for 5 minutes). Supernatants were removed, cell pellets were suspended in Karnovsky fixative, and cell samples were stored at 4°C until imaging. Cell samples were shipped on ice packs (approx 4°C) to the University of Florida Interdisciplinary Center for Biotechnology Research, and cell samples were prepared and imaged with a transmission electron microscopeq and digital camera.r

Cell death and apoptosis assays—Because CPEs in cells infected with viruses may be caused by sever-al factors, including cell death, trypan blue exclusion assays were performed to quantify cell death after in-oculation with viruses. Confluent RK-13, hemangiosar-coma Den, transitional cell carcinoma Bliley, and pri-mary canine hemangiosarcoma and soft tissue sarcoma tumor cells in 35-mm-diameter plates were mock in-oculated (by replacing the media with 400 µL of MEM without viruses) or inoculated with the Lausanne strain of MYXVa or MYXV∆serp2a (diluted in 400 µL of MEM at a high MOI [10 or 5 infectious virus particles/cell, respectively] or a low MOI [0.1 infectious virus par-ticles/cell]). Cells were incubated for 1 hour, then the inoculum was removed and 1.5 mL of MEM with 10% FBS was added to each well. At 0, 16, 24, and 48 hours after adding MEM with 10% FBS to each well, medium containing nonadherent cells was collected. Adherent cells were washed in PBS solution, and wash solution from each culture was pooled with medium containing nonadherent cells. Then, adherent cells were collected by trypsinization and pooled with the medium contain-ing nonadherent cells. Pooled cells were diluted 1:10 in trypan blue dye solution.f Blue (dead) and refractile, unstained (live) cells were manually counted by use of a hemocytometer. Proportion of dead cells was ex-

pressed as a percentage of total cells counted. Means of values for 3 to 12 replicates of each experiment were calculated.

Flow cytometry was performed to detect cells with annexin V binding to phosphatidylserine on cell membranes; this is a sensitive and specific method for identification of cells undergoing this early-stage event during apoptosis.36 Confluent RK-13, hemangiosarco-ma Den, transitional cell carcinoma Bliley, and primary canine hemangiosarcoma and soft tissue sarcoma cells were mock inoculated or inoculated with MYXV or MYXV∆serp2 (MOI, 5 infectious virus particles/cell). Cells were collected 48 hours after inoculation. Ali-quots of 105 cells were incubated with FITC-conjugated annexin Vs and 7-AADt and assayed with a flow cytome-teru in accordance with the manufacturer’s instructions. Compensation control samples were evaluated for each cell line; these included unstained cells (negative con-trol cells), cells treated with 1µM staurosporine over-night and then stained with FITC-conjugated annexin V (FITC positive control cells), MYXV-red–inoculated cells (phycoerythrin positive control cells), and cells fixed in buffered 2.5% formalin and then stained with 7-AAD (7-AAD positive control cells). A gate based on cell size (forward scatter-H) and complexity (side scatter-A) was applied to these data to exclude cell frag-ments from analysis. Data were analyzed with software.v Percentage of cells with positive results for annexin V staining (ie, apoptotic cells) at 48 hours after inocu-lation was determined; cells with positive results for 7-AAD staining were eliminated from analysis because those cells may have undergone cell death by necrosis rather than by apoptosis.

Caspase-3 activation, a terminal event during apoptosis, was also quantified in those same cells via a caspase substrate-fluorochrome cleavage assay as previ-ously described.29 Uninoculated cells incubated for at least 4 hours at 37°C in media with 1µM staurosporine were used as positive control cells for caspase-3 activa-tion. The rate of DEVD-AMC cleavage was determined.w

Statistical analysis—Data were analyzed via 1-way ANOVAs and pairwise multiple comparisons (Holm-Sidak method) with software.x Values of P ≤ 0.05 were considered significant.

Results

Histologic and cytologic characterization of pri-mary tumor cells—Histopathologic diagnoses deter-mined for tumors surgically removed from dogs were summarized (Table 1). Cell morphology and cyto-chemical staining results were consistent with tumors from which cells had been isolated, except for cells of one of the mast cell tumors, which unexpectedly lacked staining for mast cell tryptase.

CPEs and early-stage virus-encoded protein ex-pression in MYXV-inoculated tumor cells—The CPEs observed after MYXV infection of tumor cells included a change from a spindiloid or polygonal morphology to a round and highly condensed appearance with detach-ment of cells from culture plates (Figure 1). The CPE scores for tumor cells 48 hours after inoculation with

11-02-0072r.indd 1255 7/20/2012 1:25:40 PM

1256 AJVR, Vol 73, No. 8, August 2012

MYXV (MOI, 5 infectious virus particles/cell) were summarized (Table 2). Most cells of established canine tumor cell lines and those obtained from primary tu-mors had CPEs by 48 hours after inoculation, but few CPEs were observed in primary canine fibroblasts ob-tained from noncancerous connective tissue.

Extensive CPEs were observed in hemangiosarcoma Den, transitional cell carcinoma Bliley, and primary canine soft tissue sarcoma cells 48 hours after virus inoculation, whereas few CPEs were observed in primary canine hem-

angiosarcoma cells at that time (Figure 1). Expression of recombinant virus–encoded green (MYXV-gfp) and red (MYXV-red) fluorescent proteins was observed in all ca-nine tumor cell cultures evaluated 48 hours after virus in-oculation. In contrast, faint red fluorescence was observed in extremely low numbers of primary canine fibroblasts at 48 hours after inoculation with MYXV-red (MOI, 0.1 in-fectious virus particles/cell), which indicated that viral in-fection was inhibited in these noncancerous cells. Growth curve kinetics of MYXV, MYXV-gfp, and MYXV-red in RK-

Histopathologic Positive results for Negative results forCell line diagnosis cytochemical staining cytochemical staining

HSA-1 Hemangiosarcoma Vimentin Cytokeratin CD62E STSA-1 Soft tissue sarcoma Vimentin Alkaline phosphatase CD18 CytokeratinGIST-1 Gastrointestinal S100 CD18 stromal tumor CytokeratinHG-1 Hemangioma Vimentin CD18 CytokeratinMCT-1 Mast cell tumor Giemsa CD18 Cytokeratin Mast cell tryptase VimentinMCT-2 Mast cell tumor Giemsa NA Mast cell tryptase PACA-1 Perianal Cytokeratin CD18 adenocarcinoma VimentinPA-1 Perianal adenoma Cytokeratin CD18 VimentinMMT-1 Mixed mammary Cytokeratin CD18 gland tumor Vimentin RCA-1 Renal carcinoma ND NDFibroblast ND Vimentin Alkaline phosphatase

NA = Not applicable. ND = Not determined.

Table 1—Histologic diagnoses and cytochemical staining results for tumors surgically removed from 10 dogs, connective tissue obtained from the cadaver of 1 dog, and canine primary cells obtained from explants of those tissues.

Figure 1—Representative photomicrographic images of monolayer cultures of various types of cells (cells of established cell lines rabbit kidney epithelial RK-13, human malignant glioma U-118 MG, canine osteosarcoma Den [HSA-Den], and canine transitional cell carcino-ma Bliley [TCC-Bliley]; canine primary hemangiosarcoma [HSA-1] and soft tissue sarcoma [STSA-1] cells obtained from tumors surgically removed from dogs; and canine primary fibroblasts obtained from connective tissue of the cadaver of a dog) 48 hours after inoculation with viruses. A—Phase photomicrographic images of mock-inoculated cells. B—Phase photomicrographic images of MYXV-inoculated cells (MOI, 5 infectious virus particles/cell). Notice CPEs in cells. C—Fluorescent photomicrographic images of cells inoculated with MYXV-gfp (cells with green staining) or MYXV-red (cells with red staining; MOI, 0.1 infectious virus particles/cell). D—Representative phase photomicrographic images of cells inoculated with MYXV-gfp or MYXV-red (MOI, 0.1 infectious virus particles/cell). Bar = 0.1 µm.

11-02-0072r.indd 1256 7/20/2012 1:25:40 PM

AJVR, Vol 73, No. 8, August 2012 1257

13 cells were indistinguishable, which indicated the ad-dition of fluorescent markers did not alter in vitro MYXV replication (data not shown).

Relationship between Akt phosphorylation and CPEs in MYXV-inoculated cells—Slightly greater ex-pression of phosphorylated Akt was detected in heman-giosarcoma Den, transitional cell carcinoma Bliley, and primary soft tissue sarcoma cells than in human glioma U-118 MG cells (Figure 2), although results were not significantly different. These cells had moderate to se-vere CPEs after inoculation with MYXV (Table 2; Fig-ure 1). In contrast, primary canine hemangiosarcoma cells had the lowest expression of phosphorylated Akt and had few CPEs after inoculation with MYXV. These findings indicated that the level of expression of phos-phorylated Akt may be related to susceptibility of ca-nine tumor cells to infection with MYXV.

Production of infectious virus particles in ca-nine tumor cells versus that in RK-13 cells—One-step growth curves (MYXV MOI, ≥ 5 infectious virus particles/cell) indicated logarithmic increases in virus titers during 24 hours after inoculation of RK-13, hem-angiosarcoma Den, and primary canine hemangiosar-coma cells (Figure 3). However, virus titers typically decreased after that time. High rates of MYXV replica-tion were not detected in transitional cell carcinoma Bliley cells or primary canine fibroblasts. One-step growth curves were not determined for canine primary soft tissue carcinoma cells because of the large amount of virus particles needed to inoculate cells at an MOI ≥ 5 infectious virus particles/cell.

Multistep growth curves (MYXV MOI, 0.1 infec-tious virus particles/cell) indicated moderate increas-es in virus titers after inoculation of cells (Figure 3). Moderate virus production was maintained in RK-13 and canine primary hemangiosarcoma cells through 96 hours after virus inoculation, indicating cell-to-cell spread of viruses in these cell cultures. Although an in-crease in virus yield was detected for hemangiosarcoma Den, transitional cell carcinoma Bliley, and canine pri-mary soft tissue sarcoma cells, virus titers for these cells decreased after 24 hours; this finding was attributed to either rapid reduction in the number of viable cells or minimal cell-to-cell spread of viruses after initial virus replication.

CPECell name Cell origin score

Established cell lines RK-13 Rabbit kidney epithelium 2+ U-118 MG Human glioma 2+ HSA-Den Canine hemangiosarcoma 3+ TCC-Bliley Canine transitional cell carcinoma 3+ OSA-Abrams Canine osteosarcoma 2+ OSA-D-17 Canine osteosarcoma 2+ HSA-Fitz Canine hemangiosarcoma 3+

Canine primary cells HSA-1 Canine hemangiosarcoma 1+ STSA-1 Canine soft tissue sarcoma 2+ MMT-1 Canine mixed mammary gland tumor 3+ GIST-1 Canine gastrointestinal stromal tumor 3+ PACA-1 Canine perianal adenocarcinoma 1+ PA-1 Canine perianal adenoma 3+ MCT-1 Canine mast cell tumor ND MCT-2 Canine mast cell tumor 2+ HG-1 Canine hemangioma 3+ RCA-1 Canine renal carcinoma 1+ Fibroblast Canine fibroblast 1+

Cytopathologic effects were determined on the basis of the se-verity of cellular damage scores (1+ [≥ 50% cells adherent to culture plate], 2+ [10% to 50% cells adherent to culture plate], or 3+ [≤ 10% cells adherent to culture plate]) observed 48 hours after inoculation of cell cultures with MYXV (MOI, 5 infectious virus particles/cell).

ND = Not determined because results of cytochemical staining were inconsistent with type of tumor from which cells had been obtained.

Table 2—Cytopathic effect scores for MYXV-inoculated cultures of cells of established cell lines, cells obtained from explants of tumors surgically removed from 10 dogs, and fibroblasts obtained from connective tissue collected from the cadaver of 1 dog.

Figure 2—Total and phosphorylated Akt protein concentrations in uninfected cells detected by immunoblotting. A—Photo-graphs of a representative western immunoblot of lysates of cells of various types that were not inoculated with viruses. The immunoblot was stained with primary antibodies for to-tal Akt and phosphorylated Akt (P-Akt). Numbers on the left side are molecular weight in kilodaltons. B—Mean ± SEM percentage phosphorylated Akt for lysates of cells of various types. Measurements were performed to determine total AKT and phosphorylated Akt western immunoblot band densities. Phosphorylated Akt band density was calculated as a percent-age relative to total Akt band density. Values are results for 3 immunoblots. See Figure 1 for remainder of key.

11-02-0072r.indd 1257 7/20/2012 1:25:41 PM

1258 AJVR, Vol 73, No. 8, August 2012

Virion maturation in MYXV-infected cells—Because the lower yield of infectious viruses from canine tumor cells, compared with that from RK-13 cells, may have been attributable to arrest of the virus life cycle because of aber-rant virus morphogenesis, virion maturation in cells was determined. Characteristics of cells of 3 MYXV-inoculated canine tumor cell cultures (hemangiosarcoma Den, tran-sitional cell carcinoma Bliley, and primary hemangiosarcoma) were determined by evaluation of transmission electron micro-graphic images, and results were compared with those for RK-13 cells (Figure 4). Vi-ral crescents, circular immature virions, immature virions with nucleoids, and in-tracellular mature virions were detected at 24 hours after virus inoculation in all cell cultures examined by transmission elec-tron microscopy. This finding indicated MYXV replication completed a full cycle of maturation within canine tumor cells and the limited infectivity of lysates of MYXV-inoculated cells was not attributable to ab-errant virus morphogenesis. In addition, the late-stage MYXV protein M130R was detected in inoculated RK-13 cells (positive control cells) and canine tumor cells, indicating virus maturation devel-oped in canine tumor cells.

Cell death after MYXV∆serp2 inoculation versus that after MYXV inoculation—To determine wheth-

er deletion of the antiapoptotic viral serp2 protein from MYXV would result in a microscopically detect-able increase in signs of cell damage, CPE results for MYXV∆serp2-inoculated cells were compared with those for cells inoculated with MYXV-gfp or MYXV-red. More severe CPEs were observed in RK-13, U-118

Figure 3—One-step (A) and multistep (B) MYXV (MOI, ≥ 5 and 0.1 infectious virus particles/cell, respectively) growth curves in cultures of cells of various types (RK-13 [black diamonds], HAS-Den [black triangles], TCC-Bliley [white squares], HSA-1 [black circles], and canine fibroblasts [asterisks; panel B only]). At the indicated times after inoculation of cells with MYXV, cells were collected, frozen and thawed 3 times, and sonicated. Serial dilu-tions of sonicated cell lysates were used to inoculate RK-13 cells for determination of virus titers. Values are mean ± SEM of re-sults for ≥ 3 experiments. PFU = Plaque-forming unit. See Figure 1 for remainder of key.

Figure 4—Myxoma virus maturation in canine cell cultures. A—Representative transmission electron micrographic images of MYXV in various types of cells collected 24 hours after inocu-lation (MOI, 5 infectious virus particles/cell). Notice the various virus particle morphologies (immature virions [IV], immature viri-ons with nucleoids [IVN], and intracellular mature virions [IMV]). B—Photograph of a representative immunoblot for various types of cells not inoculated with MYXV (negative signs) or collected 24 hours after inoculation with MYXV (MOI, 5 infectious virus particles/cell; positive signs). Immunoblots were stained with pri-mary antibody for M130R, a protein expressed during the late stages of MYXV infection of cells. Numbers on the right side are molecular weight in kilodaltons. Bars = 1 µm (left panels for RK-13, TCC-Bliley, and HSA-1), 2 µm (left panel for HSA-Den), 200 µm (right panels for TCC-Bliley and HSA-1), or 500 µm (all other panels). See Figure 1 for remainder of key.

Figure 5—Representative phase photomicrographic images of monolayer cultures of vari-ous cell types 48 hours after mock inoculation, inoculation with MYXV-gfp or MYXV-red (MOI, 0.1 infectious virus particles/cell), or inoculation with MYXV∆serp2 (MOI, 0.1 infec-tious virus particles/cell). Notice that with the exception of HSA-1, CPEs in cell cultures inoculated with MYXV∆serp2 are more severe than those in the other cell cultures. Bar = 0.1 µm.

11-02-0072r.indd 1258 7/20/2012 1:25:42 PM

AJVR, Vol 73, No. 8, August 2012 1259

MG, hemangiosarcoma Den, transitional cell carcinoma Bliley, and canine primary soft tissue sarcoma cells in-oculated with MYXV∆serp2, compared with CPEs in those cells inoculated with MYXV (Figure 5). In con-trast, CPEs in canine primary hemangiosarcoma cells inoculated with those 2 viruses were similar. More se-vere CPEs were detected in osteosarcoma Abrams and canine primary perianal adenosarcoma and perianal ad-enoma cells inoculated with MYXV∆serp2, compared with CPEs in those cells inoculated with MYXV (data not shown). Because 1-step growth curves for MYXV and MYXV∆serp2 in RK-13 cells were indistinguish-able, MYXV∆serp2-enhanced CPEs in these cells were likely not attributable to differences in replication ki-netics (data not shown).

Percentage of dead cells gradually increased from 0 to 48 hours after mock or virus inoculation in all cell cultures evaluated (data not shown). Significant dif-ferences in percentage of dead cells were not detected when results for mock-inoculated cells were compared with those for cells inoculated with a low virus MOI at any time point (data not shown). The percentage of trypan blue–stained (dead) cells in cultures of 5 types of cells (RK-13, hemangiosarcoma Den, transitional cell carcinoma Bliley, and canine primary hemangiosar-coma and soft tissue sarcoma) at 48 hours after mock,

MYXV, or MYXV∆serp2 inoculation (high MOI) were compared (Figure 6). Percentage of dead RK-13, hem-angiosarcoma Den, and canine primary hemangiosar-coma cells in cultures inoculated with a high MOI of MXYV was slightly lower than that in mock-inoculated cultures, but differences were not significant. A sig-nificantly higher number of dead cells was detected in RK-13, hemangiosarcoma Den, transitional cell carci-noma Bliley, and canine primary hemangiosarcoma cell cultures 48 hours after inoculation with MYXV∆serp2 (high MOI) than in cultures of those cells 48 hours af-ter inoculation with MXYV (high MOI).

Hemangiosarcoma Den, transitional cell carcinoma Bliley, and canine primary hemangiosarcoma and soft tissue sarcoma cell cultures had higher percentages of cells in the early stages of apoptosis (ie, cells with positive results for annexin V staining and negative results for 7-AAD staining) 48 hours after inoculation with MYXV∆serp2 than did mock-inoculated cell cul-tures (Figure 6). Percentage cells with positive results for annexin V staining and negative results for 7-AAD staining in MYXV∆serp2-inoculated hemangiosarcoma Den cell cultures was significantly higher than that in MYXV-inoculated cultures of those cells. These results indicated MYXV∆serp2 induced apoptosis in canine tu-mor cells more often than did MYXV (Figure 6).

Figure 6—Results of assays to determine death or apoptosis in monolayer cultures of various types of cells 48 hours after mock inoculation or inoculation with MYXV or MYXV∆serp2. A—Re-sults of trypan blue dye exclusion assay for detection of cell death. An MOI of 10 (MYXV) or 5 (MYXV∆serp2) infectious virus particles/cell was used. Results are mean ± SEM percentage dead cells for ≥ 3 experiments. B—Representative flow cytom-etry scatterplots. An MOI of 5 infectious virus particles/cell was used for both viruses. Cells were stained with annexin V and 7-AAD. Cells with positive results for annexin V staining and negative results for 7-AAD staining (ie, apoptotic cells) are shown in the upper left quadrants of the plots. Numbers in quadrants of the plots are the percentage of cells in each of those quadrants. C—Results of flow cytometry analy-sis indicating percentage of apoptotic cells (ie, cells with positive results for annexin V staining and negative results for 7-AAD staining). An MOI of 5 infectious virus particles/cell was used for each virus. Values are mean ± SEM for results of 2 to 4 experiments. D—Results of DEVD-AMC cleavage assay to detect caspase-3 activity. An MOI of 5 infectious virus particles/cell was used for each virus. Notice that staurosporine-treated (positive control) cultures had higher caspase-3 activity than did mock-inoculated (negative control) cultures for TCC-Bliley and HAS-1 cells. Values are mean ± SEM. *Values that are significantly (P ≤ 0.05) different are indicated by brackets. †Positive control sample values are significantly (P ≤ 0.05) higher than negative control sample values. See Figure 1 for remainder of key.

11-02-0072r.indd 1259 7/20/2012 1:25:43 PM

1260 AJVR, Vol 73, No. 8, August 2012

A much higher percentage of mock-inoculated RK-13 cells had positive results for annexin V staining than was expected on the basis of the finding that CPEs had not been detected in those cells (Figure 5). Therefore, caspase-3 activation assays were performed. Although results were not significantly different, slightly higher rates of DEVD-AMC cleavage by caspase-3 were detect-ed for tumor cells inoculated with MYXV∆serp2, com-pared with that for mock- or MYXV-inoculated cells; this finding suggested that MYXV∆serp2 increased apoptosis of canine tumor cells.

Discussion

The biological behavior of naturally developing tu-mors in dogs is more similar to the biological behavior of human neoplasms than is that of induced tumors in mice.37 Similar to human populations, dog populations are outbred and there are differences in genetics and immune responses among dogs that directly impact neoplastic transformation and immunoevasion by tu-mor cells. In addition, tumors of dogs can be caused by environmental factors to which humans can also be exposed. Results of the present study indicated canine tumor cells can be permissive to MYXV infection, as can human tumor cells.22 Interestingly, all of the canine tumor cell types evaluated in the present study were permissive to MYXV infection, as indicated by the find-ing that intracellular virus-controlled fluorescent pro-tein expression was consistently observed after inocula-tion of cells with MYXV-gfp or MYXV-red (even at low MOIs). Nontumorous canine primary fibroblasts were resistant to MYXV-red infection. Importantly, there seemed to be a relationship between increased severity of CPEs and upregulated expression of phosphorylated Akt in canine tumor cells, which was similar to find-ings for human cancer cells reported by other au-thors.35 Thus, use of canine primary tumor cells for evaluation of the efficacy of MYXV as an oncolytic treatment seemed to be valid, and results of the pres-ent study may be relevant to oncolytic treatment of human cancers.

The primary objective of oncolytic virotherapy is to eliminate cancerous cells. For in vivo viral oncolysis, apoptotic cell death may be preferred to necrotic cell death because necrosis causes severe inflammation that can damage adjacent healthy tissues. Because MYXV encodes multiple proteins that can prevent cell death, we hypothesized that a mutant MYXV without ≥ 1 of these proteins would increase the rate of death of host cells. Although MYXV inoculation of cultures induced death in some cells, results were not significantly dif-ferent from those for mock-inoculated cells for most canine tumor cell types evaluated. Percentage of dead cells was significantly increased in all but 1 (soft tis-sue sarcoma) of the canine tumor cell cultures inocu-lated with MYXV∆serp2, compared with results for cell cultures inoculated with MYXV. Because the serp2 pro-tein is antiapoptotic in RK-13 cells,29 it was anticipated that the lack of serp2 expression during MYXV∆serp2 infection would increase the rate of apoptosis in ca-nine tumor cells. As anticipated, results of annexin V staining and DEVD-AMC cleavage assays suggested MYXV∆serp2 inoculation of canine tumor cell cultures

increased the number of apoptotic cells, compared with results for MYXV-inoculated or mock-inoculated cell cultures.

Results of recent studies38,39 indicate the efficacy of an attenuated form of the zoonotic poxvirus VACV for oncolysis of canine mammary gland tumor xenografts in mice. In those studies,38,39 administration of 1 IV dose of an attenuated VACV (GLV-1h68) to mice resulted in significant reduction in the rate of tumor growth. In addition, the same canine primary soft tissue sarcoma cells used in the present study have been used by a col-laborating investigator40 to evaluate oncolytic efficacy of another strain of VACV for treatment of canine soft tissue sarcoma xenografts in nude mice. Results of that study40 indicate systemic administration of poxviruses may be efficacious for treatment of dogs with tumors; systemic administration of oncolytic viruses may en-able concurrent targeting of primary tumors and me-tastases. However, immunocompromised mice were used in those studies38–40; such mice do not develop the complex immune responses that develop during viral infection of immunocompetent hosts. Indeed, humoral immune responses to poxvirus infections are strong in susceptible hosts41; such humoral immune responses may limit the effectiveness of oncolytic viruses. Despite this possibility, results of clinical trials in which attenu-ated VACVs were administered to humans with cancer indicate that circulating antibodies to VACV (the small-pox vaccine virus) do not substantially affect treatment outcomes.42,43 Similarly, we anticipate that IV admin-istration of modified poxviruses to dogs with cancer would be efficacious, despite clearance of the virus via cell and humoral immune responses.

Results of the present study indicated the poxvirus MYXV, with specific pathogenicity for rabbits, can cause damage and induce expression of viral-encoded pro-teins in canine tumor cells. In addition, increased ex-pression of phosphorylated Akt seemed to correspond with increased CPEs in canine tumor cells inoculated with MYXV, as has been found for human tumor cells.35 Furthermore, inoculation of canine tumor cell cultures with MYXV∆serp2 increased cell death by apoptosis, compared with results for cells inoculated with MYXV. Additional studies are warranted to determine effective-ness of MYXV∆serp2 as an oncolytic agent in vivo. Re-sults of the present study may be relevant to oncolytic virotherapy for humans with cancer.

a. Provided by Dr. Richard Moyer.b. HTB-15, American Type Culture Collection, Manassas, Va.c. Provided by Dr. Douglas Thamm.d. Nunc, Roskilde, Denmark.e. Mediatech, Manassas, Va.f. HyClone, Logan, Utah.g. Provided by Dr. Grant McFadden.h. DMI 4000 B, Leica Microsystems GmbH, Wetzlar, Germany.i. ImagePro Express, Media Cybernetics, Bethesda, Md.j. Halt protease and phosphatase inhibitor cocktail, Pierce Bio-

technology, Rockford, Ill.k. GE, Minnetonka, Minn.l. Cell Signaling Technology, Danvers, Mass.m. Millipore, Temecula, Calif.n. Photoshop, version CS4, Adobe Systems, Mountain View, Calif.o. Heat Systems Ultrasonics, Farmingdale, NY.p. SeaKem, Lonza, Rockland, Me.

11-02-0072r.indd 1260 7/20/2012 1:25:43 PM

AJVR, Vol 73, No. 8, August 2012 1261

q. H-7000, Hitachi, Tokyo, Japan.r. MegaView III, Olympus Soft-Imaging Solutions, Tokyo, Japan.s. BD Biosciences, FITC Annexin V Apoptosis Detection Kit II,

San Jose, Calif.t. BD Biosciences, 7-AAD, San Jose, Calif.u. LSR II Flow Cytometer, BD Biosciences, Franklin Lakes, NJ.v. FCS Express, version 3, DeNovo Software, Los Angeles, Calif.w. SPECTRAmax GeminiEM, Molecular Devices, Sunnyvale, Calif.x. SigmaPlot, version 12.0, Systat Software, Chicago, Ill.

References1. Stojdl DF, Lichty B, Knowles S, et al. Exploiting tumor-specific

defects in the interferon pathway with a previously unknown oncolytic virus. Nat Med 2000;6:821–825.

2. Cahill DP, Kinzler KW, Vogelstein B, et al. Genetic instability and darwinian selection in tumours. Trends Cell Biol 1999;9:M57–M60.

3. Vaha-Koskela MJ, Heikkila JE, Hinkkanen AE. Oncolytic virus-es in cancer therapy. Cancer Lett 2007;254:178–216.

4. Jiang BH, Liu LZ. PI3K/PTEN signaling in angiogenesis and tu-morigenesis. Adv Cancer Res 2009;102:19–65.

5. Soares JA, Leite FG, Andrade LG, et al. Activation of the PI3K/Akt pathway early during vaccinia and cowpox virus infections is required for both host survival and viral replication. J Virol 2009;83:6883–6899.

6. Ricketts CR. Molecular permeability of the burned area. Proc Soc Exp Biol Med 1966;121:770–772.

7. McCart JA, Ward JM, Lee J, et al. Systemic cancer therapy with a tumor-selective vaccinia virus mutant lacking thymidine kinase and vaccinia growth factor genes. Cancer Res 2001;61:8751–8757.

8. Thorne SH, Bartlett DL, Kirn DH. The use of oncolytic vaccinia viruses in the treatment of cancer: a new role for an old ally? Curr Gene Ther 2005;5:429–443.

9. Merchlinsky M, Eckert D, Smith E, et al. Construction and characterization of vaccinia direct ligation vectors. Virology 1997;238:444–451.

10. Karupiah G, Blanden RV. Anti-asialo-GM1 inhibits vaccinia vi-rus infection of murine ovaries: asialo-GM1 as an additional vi-rus receptor? Immunol Cell Biol 1990;68:343–346.

11. Selin LK, Santolucito PA, Pinto AK, et al. Innate immunity to viruses: control of vaccinia virus infection by gamma delta T cells. J Immunol 2001;166:6784–6794.

12. Xu R, Johnson AJ, Liggitt D, et al. Cellular and humoral im-munity against vaccinia virus infection of mice. J Immunol 2004;172:6265–6271.

13. Chaudhri G, Panchanathan V, Bluethmann H, et al. Obligatory requirement for antibody in recovery from a primary poxvirus infection. J Virol 2006;80:6339–6344.

14. Fenner F, Woodroffe GM. The pathogenesis of infectious myxo-matosis: the mechanism of infection and the immunological re-sponse in the European rabbit (Oryctolagus cuniculus). Br J Exp Pathol 1953;34:400–411.

15. Andrewes CH, Harisijades S. Propagation of myxoma virus in one-day old mice. Br J Exp Pathol 1955;36:18–21.

16. Jackson EW, Dorn CR, Saito JK, et al. Absence of serological evidence of myxoma virus infection in humans exposed during an outbreak of myxomatosis. Nature 1966;211:313–314.

17. Gorski J, Mizak B, Chrobocinska M. Control of rabbit myxoma-tosis in Poland. Rev Sci Tech 1994;13:869–879.

18. Fenner F. Adventures with poxviruses of vertebrates. FEMS Mi-crobiol Rev 2000;24:123–133.

19. McCabe VJ, Tarpey I, Spibey N. Vaccination of cats with an at-tenuated recombinant myxoma virus expressing feline calicivi-rus capsid protein. Vaccine 2002;20:2454–2462.

20. McCabe VJ, Spibey N. Potential for broad-spectrum protection against feline calicivirus using an attenuated myxoma virus express-ing a chimeric FCV capsid protein. Vaccine 2005;23:5380–5388.

21. Pignolet B, Boullier S, Gelfi J, et al. Safety and immunogenicity of myxoma virus as a new viral vector for small ruminants. J Gen Virol 2008;89:1371–1379.

22. Sypula J, Wang F, Ma Y, et al. Myxoma virus tropism in human tumor cells. Gene Ther Mol Biol 2004;8:103–114.

23. Lun X, Yang W, Alain T, et al. Myxoma virus is a novel oncolytic

virus with significant antitumor activity against experimental human gliomas. Cancer Res 2005;65:9982–9990.

24. Lun XQ, Zhou H, Alain T, et al. Targeting human medulloblas-toma: oncolytic virotherapy with myxoma virus is enhanced by rapamycin. Cancer Res 2007;67:8818–8827.

25. Stanford MM, Shaban M, Barrett JW, et al. Myxoma virus on-colysis of primary and metastatic B16F10 mouse tumors in vivo. Mol Ther 2008;16:52–59.

26. Lun X, Alain T, Zemp FJ, et al. Myxoma virus virotherapy for glioma in immunocompetent animal models: optimizing administration routes and synergy with rapamycin. Cancer Res 2010;70:598–608.

27. Guo ZS, Naik A, O’Malley ME, et al. The enhanced tumor selectiv-ity of an oncolytic vaccinia lacking the host range and antiapopto-sis genes SPI-1 and SPI-2. Cancer Res 2005;65:9991–9998.

28. Messud-Petit F, Gelfi J, Delverdier M, et al. Serp2, an inhibitor of the interleukin-1beta–converting enzyme, is critical in the pathobiology of myxoma virus. J Virol 1998;72:7830–7839.

29. MacNeill AL, Turner PC, Moyer RW. Mutation of the Myxoma virus SERP2 P1-site to prevent proteinase inhibition causes apoptosis in cultured RK-13 cells and attenuates disease in rab-bits, but mutation to alter specificity causes apoptosis without reducing virulence. Virology 2006;356:12–22.

30. Nathaniel R, MacNeill AL, Wang YX, et al. Cowpox virus crmA, myxoma virus serp2 and baculovirus P35 are not functionally interchangeable caspase inhibitors in poxvirus infections. J Gen Virol 2004;85:1267–1278.

31. Johnston JB, Barrett JW, Chang W, et al. Role of the serine-threonine kinase PAK-1 in myxoma virus replication. J Virol 2003;77:5877–5888.

32. Liu J, Wennier S, Reinhard M, et al. Myxoma virus expressing interleukin-15 fails to cause lethal myxomatosis in European rabbits. J Virol 2009;83:5933–5938.

33. Duteyrat JL, Gelfi J, Bertagnoli S. Ultrastructural study of myx-oma virus morphogenesis. Arch Virol 2006;151:2161–2180.

34. Barrett JW, Alston LR, Wang F, et al. Identification of host range mutants of myxoma virus with altered oncolytic potential in hu-man glioma cells. J Neurovirol 2007;13:549–560.

35. Wang G, Barrett JW, Stanford M, et al. Infection of human can-cer cells with myxoma virus requires Akt activation via interac-tion with a viral ankyrin-repeat host range factor. Proc Natl Acad Sci U S A 2006;103:4640–4645.

36. Kylarova D, Prochazkova J, Mad’arova J, et al. Comparison of the TUNEL, lamin B and annexin V methods for the detection of apop-tosis by flow cytometry. Acta Histochem 2002;104:367–370.

37. Porrello A, Cardelli P, Spugnini EP. Oncology of companion ani-mals as a model for humans. an overview of tumor histotypes 2. J Exp Clin Cancer Res 2006;25:97–105.

38. Gentschev I, Stritzker J, Hofmann E, et al. Use of an oncolytic vaccinia virus for the treatment of canine breast cancer in nude mice: preclinical development of a therapeutic agent. Cancer Gene Ther 2009;16:320–328.

39. Gentschev I, Ehrig K, Donat U, et al. Significant growth inhi-bition of canine mammary carcinoma xenografts following treatment with oncolytic vaccinia virus GLV-1h68. J Oncol 2010;2010:736907.

40. Gentschev I, Adelfinger M, Josupeit R, et al. Preclinical evalua-tion of oncolytic vaccinia virus for therapy of canine soft tissue sarcoma. PLoS One [serial online] 2012; 7:e37239. Available at: www.plosone.org/article/fetchObjectAttachment;jsessionid=EB9DA36616175442A5E5F9C7E105C6FD?uri=info%3Adoi%2F10.1371%2Fjournal.pone.0037239&representation=PDF. Ac-cessed May 15, 2012.

41. Kennedy RB, Ovsyannikova IG, Jacobson RM, et al. The immunol-ogy of smallpox vaccines. Curr Opin Immunol 2009;21:314–320.

42. Liu TC, Hwang T, Park BH, et al. The targeted oncolytic pox-virus JX-594 demonstrates antitumoral, antivascular, and anti-HBV activities in patients with hepatocellular carcinoma. Mol Ther 2008;16:1637–1642.

43. Amato RJ, Shingler W, Goonewardena M, et al. Vaccination of renal cell cancer patients with modified vaccinia Ankara deliv-ering the tumor antigen 5T4 (TroVax) alone or administered in combination with interferon-alpha (IFN-alpha): a phase 2 trial. J Immunother 2009;32:765–772.

11-02-0072r.indd 1261 7/20/2012 1:25:43 PM