|Home | About | Journals | Submit | Contact Us | Français|
Oncolytic vesicular stomatitis virus (VSV) is being developed as a novel therapeutic agent for cancer treatment, although it is toxic in animals when administered systemically at high doses. Its safety can be substantively improved by an MΔ51 deletion in the viral genome, and yet VSV(MΔ51) induces a much greater, robust cellular inflammatory response in the host than wild-type VSV, which severely attenuates its oncolytic potency. We have reported that the oncolytic potency of wild-type VSV can be enhanced by vector-mediated expression of a heterologous viral gene that suppresses cellular inflammatory responses in the lesions. To develop an effective and safe VSV vector for cancer treatment, we tested the hypothesis that the oncolytic potency of VSV(MΔ51) can be substantively elevated by vector-mediated expression of M3, a broad-spectrum and high-affinity chemokine-binding protein from murine gammaherpesvirus-68. The recombinant vector rVSV(MΔ51)-M3 was used to treat rats bearing multifocal lesions (1–10mm in diameter) of hepatocellular carcinoma (HCC) in their liver by hepatic artery infusion. Treatment led to a significant reduction of neutrophil and natural killer cell accumulation in the lesions, a 2-log elevation of intratumoral viral titer, substantively enhanced tumor necrosis, and prolonged animal survival with a 50% cure rate. Importantly, there were no apparent systemic and organ toxicities in the treated animals. These results indicate that the robust cellular inflammatory responses induced by VSV(MΔ51) in HCC lesions can be overcome by vector-mediated intratumoral M3 expression, and that rVSV(MΔ51)-M3 can be developed as an effective and safe oncolytic agent to treat advanced HCC patients in the future.
Vesicular stomatitis virus (VSV) is a nonsegmented negative-strand RNA virus with inherent specificity for replication in most mammalian and human tumor cells because of their attenuated antiviral responses, and is being developed as a potent oncolytic virus for cancer treatment (Stojdl et al., 2000). We have previously used VSV vectors to treat an orthotopic model of multifocal hepatocellular carcinoma (HCC) in the livers of syngeneic and immune-competent rats through hepatic artery infusion (Ebert et al., 2003), which led to tumor-selective viral replication, oncolysis, tumor regression, and modest survival prolongation. By incorporating vector-mediated expression of a fusogenic membrane glycoprotein gene (F) from the heterologous Newcastle disease virus (rVSV-F), we were able to amplify the tumoricidal effects of oncolytic VSV through syncytium induction (Ebert et al., 2003, 2004). Although statistically significant survival advantage was achieved in animals bearing multifocal HCC in the liver, long-term survival was not achieved in most treated rats as intratumoral viral replication appeared to be rapidly suppressed by an antiviral inflammatory response in the immune-competent host. In addition, limb paralysis secondary to VSV replication in neurons was observed in some of the animals treated with the vector at doses above the maximum tolerated dose (MTD). These limitations highlight the need for the development of novel rVSV vectors with attenuated replication potential in neurons to improve treatment safety, as well as enhanced replication potency in tumors to improve treatment efficacy.
Cellular inflammatory processes are mediated by chemoattractants called chemokines (Schall et al., 1994), which are a large family of small signaling peptides that bind to G protein-coupled receptors on target immune cells and induce their chemotaxis to the sites of inflammation. Chemokines also play a central role in the host defense against invading viruses and in the pathogenesis of inflammatory diseases (Rollins, 1997; Baggiolini, 1998). A number of viruses have evolved elegant mechanisms to evade detection and subsequent destruction by various inflammatory cells in the host (Alcami, 2003). One such mechanism involves the production of secreted chemokine-binding proteins that bear no sequence homology to host proteins, yet function to competitively bind and/or inhibit the interactions of chemokines with their cognate receptors (Seet and McFadden, 2002), thereby suppressing the chemotaxis of inflammatory cells to the infected sites. The large DNA viruses, such as the poxviruses and herpesviruses, have evolved such mechanisms to undermine the normal functioning of the chemokine network in the host. We have reported the enhancement of oncolytic potency of wild-type VSV by vector-mediated expression of gG, a chemokine-binding protein from equine herpesvirus-1, which inhibited the chemotaxis of natural killer (NK) cells to the lesions (Altomonte et al., 2008).
Most wild-type strains of VSV are known to be relatively poor inducers of interferon (IFN) (Marcus et al., 1998), and the VSV matrix (M) protein is a virulence factor that is capable of inhibiting host gene expression at the level of transcription (Ferran and Lucas-Lenard, 1997; Ahmed et al., 2003), as well as the nuclear–cytoplasmic transport of host RNAs and protein (Petersen et al., 2000; von Kobbe et al., 2000). Stojdl and coworkers (2003) reported that VSV mutants containing either one (M51R) or two (V221F and S226R) amino acid substitutions in the VSVM gene are potent inducers of IFN, and are safe in mice after repeated systemic administrations at high doses. Our group has explored the potential of VSV(MΔ51) as an oncolytic agent for the treatment of breast cancer metastases via intravenous administration in an immune-competent mouse model. The results confirmed that the M-mutant is indeed a much safer oncolytic virus than wild-type VSV, although its intratumoral replication was attenuated as well, which led to significantly reduced oncolytic potency (Ebert et al., 2005). We hypothesized that the oncolytic potency of VSV(MΔ51) can be significantly enhanced by vector-mediated expression of genes from heterologous viruses that counteract the host antiviral inflammatory responses. In this study, we constructed a novel rVSV(MΔ51)-M3 vector that expresses a secreted form of M3 from the murine gammaherpesvirus-68, which is a viral chemokine-binding protein (vCKBP) that binds to a broad range of chemokines (C, CC, CXC, and CX3C) with high affinity (Parry et al., 2000; van Berkel et al., 2000), and tested its treatment efficacy and safety in a multifocal HCC model in the livers of immune-competent and syngeneic rats.
The rat HCC cell line McA-RH7777 was purchased from the American Type Culture Collection (ATCC, Manassas, MA) and maintained in Dulbecco's modified Eagle's medium (DMEM; Mediatech, Herndon, VA) in a humidified atmosphere at 10% CO2 and 37°C. BHK-21 cells (ATCC) were maintained in DMEM in a humidified atmosphere at 5% CO2 and 37°C. All culture media were supplemented with 10% heat-inactivated fetal bovine serum (Sigma-Aldrich, St. Louis, MO) and penicillin–streptomycin (100U/ml; Media-tech).
To generate recombinant VSV with a single methionine deletion at position 51 of the M protein gene (MΔ51), the full-length cDNA VSV clone was digested with XbaI and KpnI and the obtained fragment containing the M protein gene was modified by site-directed polymerase chain reaction (PCR) mutagenesis (QuikChange II XL; Stratagene, La Jolla, CA). Subsequently, the fragment containing MΔ51 was ligated into a similarly digested full-length cDNA clone of VSV encoding the M3 gene constructed as follows. To create recombinant VSV vectors expressing the secreted form of murine gammaherpesvirus M3 (M3), a truncated M3 gene was synthesized chemically in its entirety (GenScript, Piscataway, NJ). To determine the secreted form, a hydrophobicity plot was generated to predict the C-terminal trans-membrane domain. The secreted form of M3 was determined to be the first 1221 bp of the full-length gene, which is consistent with the findings of others. Sequencing of the plasmids was conducted in the DNA Sequencing Core Facility at Mount Sinai School of Medicine (New York, NY).
To rescue the recombinant VSV vector, established methods of reverse genetics were employed (Ebert et al., 2003). Briefly, BHK-21 cells were infected with vaccinia virus expressing the T7 RNA polymerase (vTF-7.3), and subsequently transfected with the full-length rVSV plasmid in addition to plasmids encoding VSV N, P, and L proteins, using Lipofectamine 2000 transfection reagent (Invitrogen, Carlsbad, CA). Seventy-two hours posttransfection, supernatants were centrifuged and filtered through a 0.22-μm (pore size) filter to remove the majority of vaccinia virus, and transferred onto fresh BHK-21 cells. Any remaining vaccinia virus was eliminated by plaque purification, and the titers of recombinant VSV stocks were determined by plaque assays on BHK-21 cells.
Morris (McA-RH7777) cells were plated in 24-well plates at 5×104 cells per well and infected at a multiplicity of infection (MOI) of 0.01 with wt-rVSV-LacZ, rVSV(MΔ51)-LacZ, or rVSV(MΔ51)-M3. After infection at 37°C for 30min, cells were washed twice with phosphate-buffered saline (PBS) to remove any unabsorbed virus, and fresh complete medium was added. At the indicated time points after infection, 500-μl aliquots of culture supernatant were collected and assayed for viral titer by TCID50 (50% tissue culture infective dose) assay.
McA-RH7777 cells were seeded in 24-well plates at 5×104 cells per well overnight and then infected with wt-rVSV-LacZ, rVSV(MΔ51)-LacZ, or rVSV(MΔ51)-M3 at an MOI of 0.01 the next day. Cell viability was measured in triplicate wells at the indicated time points after infection, using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (cell proliferation kit I; Roche, Indianapolis, IN). All cell viability data are expressed as a percentage of viable cells as compared with mock-infected controls at each time point.
Morris (McA-RH7777) cells were plated in 24-well plates at 5×104 cells per well. Anti-PMN, anti-asialo-GM1, or control rabbit IgG was added to cells at 0.1mg/ml and incubated for 24hr. The treated cells were infected with wt-rVSV-LacZ or rVSV(MΔ51)-LacZ at an MOI of 0.0001 plaque-forming units (PFU)/cell. After incubation at 37°C for 60min, cells were washed twice with PBS to remove any unabsorbed virus, and refed with fresh complete medium containing the appropriate antibody. At the indicated time points after infection, 500-μl aliquots of culture supernatant were collected and assayed for viral titer by TCID50 assay.
All procedures involving animals were approved by and performed according to the guidelines of the Institutional Animal Care and Use Committee of the Mount Sinai School of Medicine. Six- to 8-week old male Buffalo rats were purchased from CRL (Indianapolis, IN) and housed in a specific pathogen-free environment under standard conditions. To establish multifocal HCC lesions within the liver, 1×107 syngeneic McA-RH7777 rat HCC cells (in a 1-ml suspension of DMEM) were infused into the portal vein. Twenty-one days after tumor cell implantation, the development of multifocal hepatic tumors of 1–10mm in diameter was confirmed, and TNE buffer control, wt-rVSV-LacZ (5.0×107 PFU/kg), or rVSV(MΔ51)-LacZ or rVSV(MΔ51)-M3 (5.0×107 to 5.0×109 PFU/kg in a total volume of 1ml) was administered via the hepatic artery. To evaluate tumor response to viral treatment, animals were killed 3 days after infusion and tumors were subjected to histological and immunohistochemical staining, as well as TCID50 analysis of tumor extracts for quantification of VSV concentration. In another study, the same groups of animals were monitored for survival.
Neutrophil and NK cell depletion was accomplished by intravenous administration of rabbit anti-rat polymorphonuclear leukocyte (PMN) antiserum (Wako Chemicals USA, Richmond, VA) and polyclonal rabbit anti-asialo-GM1 (Wako Chemicals USA) 24hr before as well as 24hr after vector infusion. Using a defined dose of 1mg/200μl/rat, Buffalo rats harboring multifocal HCC lesions were randomized to receive either rabbit anti-rat PMN antiserum, anti-asialo-GM1 antiserum, or an equal volume of normal rabbit serum (control IgG) in combination with a single hepatic arterial injection of vector. All animals were killed on day 3 after vector administration.
Liver samples containing tumor were fixed overnight in 4% paraformaldehyde and then paraffin-embedded. Thin sections were subjected to either hematoxylin–eosin (H&E) staining for histological analysis or immunohistochemical staining with monoclonal antibodies against VSV-G protein (Alpha Diagnostic, San Antonio, TX) or myeloperoxidase (MPO) (Abcam, Cambridge, UK). Another set of tumor-containing liver samples was fixed overnight in 4% paraformaldehyde and then equilibrated in 20% sucrose in PBS overnight. Frozen sections were subjected to immunohistochemical staining with monoclonal antibodies against NKR-P1A (BD Pharmingen; BD Biosciences, San Jose, CA). Semiquantification of positively stained cells was performed with ImagePro software (MediaCybernetics, Silver Spring, MD), and immune cell index was calculated as a ratio of positive cell number to unit tumor area (10,000 pixels equals 1 unit tumor area).
Blood samples were collected from the left ventricle 3 days after viral infusion, at the time of euthanization, and the levels of serum cytokines were determined by enzyme-linked immunosorbent assay (ELISA) (R&D Systems, Minneapolis, MN). Serum chemistry including alanine aminotransferase (ALT), aspartate aminotransferase (AST), and blood urea nitrogen (BUN) were performed by the chemistry laboratory at Mount Sinai School of Medicine.
For comparison of individual data points, a two-sided Student t test was applied to determine statistical significance. Survival curves were plotted according to the Kaplan–Meier method, and statistical significance between the different treatment groups was compared by log-rank test. Results and graphs were obtained with the GraphPad Prism 3.0 program (GraphPad Software, San Diego, CA).
Buffalo rats bearing multifocal lesions of HCC in the liver were treated with rVSV-LacZ or rVSV(MΔ51)-LacZ at 5×107 PFU/kg, in the presence of rabbit antiserum against rat polymorphonuclear leukocytes (PMNs) to deplete neutrophils, rabbit anti-asialo-GM1 to deplete NK cells, or control rabbit serum. Tumor tissues were obtained from animals killed on day 3 after vector administration, and inflammatory cells were identified by immunohistochemical staining (Fig. 1A). There were ~50 neutrophils per unit tumor area in rVSV-LacZ-treated animals, which was reduced to ~30 with anti-neutrophil treatment (p=0.041). The number of neutrophils was tremendously increased up to ~80 in rVSV(MΔ51)-LacZ-treated rats compared with rVSV-LacZ-treated rats (p=0.044); this number was also reduced to ~30 (p=0.03) with depletion. Lysates of tumor tissues were subjected to TCID50 assays (Fig. 1B), and intratumoral viral titers in rVSV(MΔ51)-LacZ-treated rats were elevated by 2 logs with neutrophil depletion (p=0.018). Intratumoral viral titers were decreased by 1 log when compared with rVSV-LacZ-treated rats (p=0.26), and were restored with neutrophil depletion (p<0.001). Tumor-containing liver sections were stained histologically and the percentages of necrotic areas within tumors were quantified by morphometric analysis (Fig. 1C). The necrotic areas in tumors from rVSV-LacZ-treated animals were increased from 18 to 46% with neutrophil depletion (p=0.049). The necrotic areas were reduced to 18% in tumors from rVSV(MΔ51)-LacZ-treated rats (p=0.042) compared with rVSV-LacZ-treated rats, and restored to 45% with neutrophil depletion (p=0.006). Similarly, there was a statistically significant reduction in intratumoral contents of NK cells after anti-asialo-GM1 antibody treatment, which was associated with an enhancement of intratumoral viral titers and necrotic areas (Fig. 1D–F). Compared with the tumors from rats treated with rVSV-LacZ, there were increased intratumoral NK cell accumulation associated with decreased viral titers and necrotic areas in the tumors of rats treated with rVSV(MΔ51)-LacZ (Fig. 1D–F). The elevated intratumoral viral titers were apparently not caused by direct stimulation by the respective antibodies, which have no effects on viral replication in rat HCC cells in vitro (data not shown). Collectively, these results suggest that there is substantial enhancement of neutrophil and NK cell accumulation in tumors treated with rVSV(MΔ51)-LacZ relative to those treated with rVSV-LacZ, which is correlated with attenuated replication of rVSV(MΔ51)-LacZ and reduced tumor response in HCC tumors. Moreover, neutrophils and NK cells apparently play a major role in suppressing intratumoral VSV replication, especially after attenuated rVSV(MΔ51) infusion, that could be reversed by their antibody-mediated depletion in vivo, leading to substantially enhanced oncolysis and tumor response.
M3 from murine gammaherpesvirus-68 is a broad-spectrum chemokine-binding protein that suppresses the chemotaxis of inflammatory cells in response to C, CC, CXC, and CX3C chemokines with high affinity (Parry et al., 2000; van Berkel et al., 2000). The cDNA corresponding to the secreted form of M3 was cloned into the genome of rVSV containing a single methionine deletion at position 51 of the M protein gene (MΔ51) as a new transcription unit (Fig. 2A). Reverse genetics was employed to generate the corresponding recombinant VSV vector, rVSV(MΔ51)-M3, as previously described (Lawson et al., 1995; Whelan et al., 1995). Rat HCC cells were infected with either rVSV-LacZ, rVSV(MΔ51)-LacZ, or rVSV(MΔ51)-M3 at an MOI of 10, and controls were mock infected with culture medium. Conditioned media were collected after 5hr and analyzed by Western blotting with a mono-specific anti-M3 antibody. Although there was no detectable M3 protein in the mock-, rVSV-LacZ-, or rVSV(MΔ51)-LacZ-infected supernatant, high levels of the protein were present in the supernatant of HCC cells infected with rVSV(MΔ51)-M3 (Fig. 2B). These results indicate that murine gammaherpesvirus M3 can be secreted by cells infected with rVSV(MΔ51)-M3.
One concern about constructing recombinant VSV vectors expressing one or more exogenous genes is that this could be detrimental to viral infectivity and titers. To compare the replication kinetics of rVSV(MΔ51)-M3 with that of rVSV(MΔ51)-LacZ and rVSV-LacZ in vitro, TCID50 assays were performed on culture supernatants collected at various time points after infection of rat HCC cells at an MOI of 0.01 (Fig. 2C). The kinetics of viral replication were similar for all three viruses, with no statistically significant differences at any time point, indicating that the new recombinant viruses suffered no significant changes in replication efficiency or overall yield in rat HCC cells in vitro. To examine the tumor cell-killing potential of the new vector, rat HCC cells were infected with rVSV(MΔ51)-M3, rVSV(MΔ51)-LacZ, or rVSV-LacZ at an MOI of 0.01. The cytopathic effects on the cells were quantified by MTT assays and expressed as a percentage of mock-infected cells at each time point. The kinetic profiles of cell killing caused by all three viruses were similar and without statistically significant differences at all time points, with nearly all of the cells being killed within 72hr postinfection (Fig. 2D). These results demonstrate that rVSV(MΔ51)-LacZ and rVSV(MΔ51)-M3 are able to kill rat hepatoma cells as effectively as rVSV-LacZ in vitro.
To determine secreted M3 expression as well as the chemokine level in tumors after viral infection in vivo, multifocal HCC tumor-bearing rats were infused with rVSV(MΔ51)-LacZ or rVSV(MΔ51)-M3 at 5.0×109 PFU/kg via the hepatic artery. Three days after virus injection, tumors were harvested and homogenized for detection of M3 by Western blotting, and for measurement of monocyte chemoattractant protein (MCP)-1 by ELISA. High levels of secreted M3 protein were present in the tumors infused with rVSV(MΔ51)-M3 but not in those infused with rVSV(MΔ51)-LacZ (Fig. 2E). The intratumoral chemokine MCP-1 protein level was significantly lower in rats administered rVSV(MΔ51)-M3 than in rats administered rVSV(MΔ51)-LacZ (Fig. 2F; p=0.045). These results indicate that elevated intratumoral M3 expression after rVSV(MΔ51)-M3 infusion is associated with reduced intratumoral levels of a chemokine in vivo.
To evaluate whether secretion of the M3 protein by tumor cells infected with rVSV(MΔ51)-M3 could inhibit inflammatory cell accumulation in vivo, rats bearing multifocal HCC lesions ranging from 1 to 10mm in diameter were treated with either buffer, rVSV-LacZ at its MTD (5.0×107 PFU/kg), or rVSV(MΔ51)-LacZ or rVSV(MΔ51)-M3 at the equivalent or higher doses (5.0×107, 5.0×108, and 5.0×109 PFU/kg) via hepatic artery infusion. On day 3 after treatment, animals were killed and tumor-containing liver sections were stained for neutrophils, using anti-MPO (Fig. 3A), and NK cells, using anti-NKR-P1A (Fig. 3C). Semiquantification of marker-positive cells, using ImagePro software, revealed that there was substantial accumulation of neutrophils and NK cells in the lesions of rVSV(MΔ51)-LacZ-versus rVSV-LacZ-treated rats (Fig. 3B, p=0.01 and Fig. 3D, p=0.03, respectively), which was substantially reduced after rVSV(MΔ51)-M3 treatment at the same dose (Fig. 3B, p=0.002 and Fig. 3D, p=0.0046, respectively). In addition, there appeared to be dose-dependent suppression of intratumoral neutrophil and NK cell accumulation in rVSV(MΔ51)-M3-treated rats (Fig. 3B and D). Taken together, these results indicate that the chemotaxis of neutrophils and NK cells to the tumor site is enhanced by VSV(MΔ51), but substantially inhibited by vector-mediated expression of M3.
To assess the in vivo effect of combining the M protein deletion mutant with vector-mediated intratumoral M3 expression on intratumoral viral replication and oncolysis, tumor-bearing rats were treated with either buffer, rVSV-LacZ at its MTD (5.0×107 PFU/kg), or rVSV(MΔ51)-LacZ or rVSV(MΔ51)-M3 at the equivalent or higher doses (5.0×107, 5.0×108, and 5.0×109 PFU/kg) via hepatic artery infusion. Animals were killed on day 3 after treatment and tumor samples were collected and fixed for histological and immunohistochemical staining, as well as snap-frozen for intratumoral viral titer quantification by TCID50 analysis. Whereas rVSV-LacZ infusion resulted in viral titers of less than 104 TCID50/mg of tumor tissue, an equivalent dose of rVSV(MΔ51)-LacZ led to a 1-log attenuation in intratumoral viral titer (Fig. 4; p=0.027). The same dose of rVSV(MΔ51)-M3 resulted in a 3-log enhancement in intratumoral viral titer as compared with rVSV(MΔ51)-LacZ (Fig. 4; p=0.0008). To examine the impact of enhanced intratumoral viral replication on tumor response, tumor-containing liver sections from the animals in the above-described experiment were examined by H&E staining, and the necrotic areas were quantified by morphometric analysis. The extents of tumor necrosis were reduced in the rVSV(MΔ51)-LacZ treatment group compared with the rVSV-LacZ control vector group (Fig. 5, 15 vs. 23%; p=0.03), and a significant enhancement of tumor response was observed in rats treated with rVSV(MΔ51)-M3 versus those treated with an equivalent dose of the rVSV(MΔ51)-LacZ vector (Fig. 5, 50 vs. 15%; p<0.001). There also appeared to be a dose dependence in tumor response to rVSV(MΔ51)-M3 administration, which was further elevated to 80% at the highest dose.
To determine whether the attenuated oncolytic potency of rVSV(MΔ51)-LacZ can be overcome by vector-mediated expression of the M3 gene, rats bearing multifocal lesions of HCC were treated with either buffer, rVSV-LacZ at its MTD, or rVSV(MΔ51)-LacZ or rVSV(MΔ51)-M3 at equivalent or higher doses, via hepatic artery infusion. The animals were monitored daily for survival. Consistent with our previous reports, rVSV-LacZ treatment prolonged median animal survival from 14 to 17 days (Fig. 6; p=0.048 vs. buffer). After treatment with rVSV(MΔ51)-LacZ at doses of 5.0×107, 5×108, and 5.0×109 PFU/kg, median survival was 21, 22, and 23 days, respectively. All animals died by day 35, and there were no statistically significant differences between various dose level cohorts or from rVSV-LacZ-treated animals (Fig. 6). However, rVSV(MΔ51)-M3 treatment resulted in highly significant prolongation of median survival, from 21 to 33 days, when compared with rVSV(MΔ51)-LacZ-treated animals at 5.0×107 PFU/kg (p=0.004), with 3 of 10 animals (30%) achieving long-term survival. The median survival advantage was further increased to 44 and 59 days in HCC-bearing rats given rVSV(MΔ51)-M3 at 5.0×108 and 5.0×109 PFU/kg, with concomitant increases in long-term survival to 40 and 50%, respectively. The surviving rats in the rVSV(MΔ51)-M3 treatment groups were killed on day 130 and evaluated for residual malignancy. There were no visible tumors within the liver or elsewhere, and there was no histological evidence of residual tumor in any of the major organs. These results indicate that the attenuated oncolytic potency of VSV(MΔ51) can be completely overcome by vector-mediated expression of the M3 gene. Importantly, the results indicate that multifocal lesions of up to 10mm in diameter at the time of treatment had undergone complete remission in a significant fraction of the animals treated with rVSV(MΔ51)-M3, which translated into long-term, tumor-free survival.
Safety is of the utmost concern when using genetic strategies to enhance oncolytic virus potency, considering that they are capable of evading the host antiviral inflammatory responses. Consistent with previous reports using VSV(MΔ51)-based vectors, no rVSV(MΔ51)-LacZ- and rVSV(MΔ51)-M3-treated animal showed any significant weight loss, dehydration, piloerection, limb paralysis, or lethality even at doses as high as 5.0×109 PFU/kg, which is 2 logs higher than the MTD of wild-type VSV. To assess potential systemic and organ toxicities, complete blood count (CBC), serum ALT, AST, BUN, creatinine, and serum proinflammatory cytokine levels were measured on day 3 after hepatic artery infusion of buffer, rVSV-LacZ at its MTD, and rVSV(MΔ51)-LacZ or rVSV(MΔ51)-M3 at equivalent or higher doses. There were no abnormal changes in red blood cells (RBCs), white blood cells (WBCs), hemoglobin, and hematocrit after treatment with any of the viruses at all doses used (Fig. 7A, panels a and b), indicating normal hematologic functions. Both AST and ALT were elevated somewhat in the buffer group and all vector-treated groups because of the presence of HCC lesions, and there were no significant differences between any of the treatment groups, indicating that none of these three viruses has any additional toxic effect on liver function (Fig. 7A, panel c). There were also no increases in BUN or creatinine levels, demonstrating that there was no nephrotoxicity (Fig. 7A, panel d). The serum concentrations of the proinflammatory cytokine tumor necrosis factor (TNF)-α were comparable between the buffer and all rVSV vector treatment groups, and were >2 logs below the concentrations associated with systemic toxicity in animals and in human clinical trials (the toxic threshold of TNF-α in clinical trails is 3000pg/ml) (Gaddy and Lyles, 2005) (Fig. 7A, panel e). The serum concentration of another proinflammatory cytokine, IFN-γ, was undetectable in all groups (<31.2pg/ml), indicating that there was no systemic proinflammatory cytokine response in the immune-competent rats. Histological sections of the liver and other major organs including the brain, spinal cord, lung, heart, kidney, spleen, and duodenum were examined 3 days after viral infusion and these tissues were completely normal with no inflammatory cell infiltration (Fig. 7B), indicating that there was no organ toxicity in animals injected with rVSV(MΔ51)-M3.
Although wild-type VSV-based vectors can be effective in oncolysis of tumor cells in animals, they can also cause substantial toxicities including limb paralysis and acute lethality when administered at pharmacological doses (Schneider-Schaulies, 2000; van den Pol et al., 2002; Shinozaki et al., 2005). Because of their vastly improved safety profiles, VSV(MΔ51)-based vectors are particularly attractive candidates for clinical translational applications. The matrix (M) protein of VSV is not only a structural protein necessary for virus assembly, but also a virulence factor of VSV. The VSV-M protein interferes with host cell gene expression in infected cells by blocking mRNA export to the cytosol (Gaddy and Lyles, 2005). It has been reported that deletion of the amino acid at position 51 results in the loss of its ability to block cellular mRNA transport, leading to elevated interferon and cytokine expression in the virus-infected cells. An enhanced IFN response attenuates viral replication in normal cells, thus reducing VSV-related toxicity. Tumor cells with their attenuated IFN responsiveness, however, remain susceptible to VSV(MΔ51) replication and cytolytic killing. The general applicability of VSV(MΔ51) as an effective agent to kill multiple tumor types in vitro has been demonstrated by Bell's group (Stojdl et al., 2003), and it is highly lytic in most of the National Cancer Institute (NCI, Bethesda, MD) panel of 60 human cancer cell lines. Their studies further demonstrated that infection with VSV(MΔ51) could establish an antiviral state in the recipient animals that protects against toxicities normally associated with infection by wild-type VSV (Stojdl et al., 2003). We have confirmed this observation in an immune-competent mouse model of metastatic breast cancer, where the MTD of rVSV(M51R)-LacZ was elevated by at least 100-fold over that of an equivalent virus, rVSV-LacZ (Ebert et al., 2005).
Although VSV(MΔ51) replicates efficiently in most cancer cells in vitro and is a much safer oncolytic virus in animals, its intratumoral viral replication in vivo was decreased by 1 log in rats bearing multifocal lesions of HCC in their livers, which was associated with a significant reduction in the extent of tumor necrosis. There were robust inflammatory cell accumulations in the lesions of VSV(MΔ51)-treated animals that correlated with a logarithmic reduction in intratumoral viral titer, which was not present in virus-infected tumor cells in vitro. Indeed, intratumoral viral titers of VSV(MΔ51) can be restored by antibody-mediated depletion of inflammatory cells in the animals before vector administration. Efficient migration of leukocytes from the blood vessels to the sites of viral infection and inflammation is an important weapon in the host's antiviral defenses. Indeed, it has been reported in a rat brain tumor model that antitumor efficacy of oncolytic herpes simplex virus (HSV) might be limited by the host innate and elicited humoral responses, which could be overcome by cyclophosphamide treatment (Ikeda et al., 1999, 2000). Chemokines play a central role in this process through the activation and mobilization of macrophages, lymphocytes, dendritic cells, natural killer cells, and granulocytes (Luster, 1998). Viruses have evolved mechanisms to interfere with the chemokine system via virus-encoded chemokine receptor homologs (MacDonald et al., 1997; Penfold et al., 1999) or secretion of viral chemokine-binding proteins with novel structures (Lalani and McFadden, 1997). We have reported that suppression of chemotaxis of NK cells to lesions after treatment with wild-type VSV could be achieved by vector-mediated expression of gG, a chemokine-binding protein from equine herpesvirus-1, which in turn led to elevated VSV-mediated oncolysis, tumor response, and survival prolongation (Altomonte et al., 2008). In this study, we tested our hypothesis that the severely attenuated oncolytic potency of VSV(MΔ51), secondary to a much more robust neutrophil and NK cell response to the vector than wild-type VSV, can be overcome by vector-mediated expression of the M3 gene from murine gammaherpesvirus-68, which binds a broad range of chemokines including CC, CXC, C, and CX3C with high affinity (Parry et al., 2000; van Berkel et al., 2000).
During the early phase of viral infection, innate cells (neutrophils and natural killer cells) are the first to infiltrate the infected site after VSV infection. The first phase of chemokine expression corresponds to positive staining for neutrophils (peak, 36hr postinfection) (Bi et al., 1995) and infiltrating NK cells (peak, approximately 3–4 days postinfection) (Chen et al., 2001; Ireland and Reiss, 2006). The second phase of expression corresponds to the infiltration of macrophages (Christian et al., 1996) and CD4+ and CD8+ T cells, which peak after 1 week (Huneycutt et al., 1993). Because intratumoral VSV replication was inhibited after 1–3 days of viral infusion, neutrophil and NK cell recruitment is important in inhibiting virus propagation during early infection (Chen et al., 2001; Ireland and Reiss, 2006). The neutrophil depletion experiment showed a 2-log elevation in viral replication together with enhanced tumor necrosis when animals were treated with PMN-depleting antiserum. M3 is a high-affinity, broad-spectrum secreted vCKBP that binds C, CC, CXC, and CX3C chemokines. Multiple chemokines are capable of recruiting neutrophilic granulocytes (Parry et al., 2000; Bryant et al., 2003), which are known to play an important role in controlling bacterial infections (Mollinedo et al., 1999). However, there is substantial evidence that neutrophils may also contribute to the host defense against viral infections. It is known that certain viral infections can lead to an increase in blood neutrophil counts (Douglas et al., 1996). Indeed, rapid and predominant neutrophil responses have been reported to occur at sites of viral infection (Feigin and Shackelford, 1973; Faden et al., 1984). Thus, viral infections can provide the necessary stimulus for neutrophil migration. In addition, neutrophils have been reported to specifically adhere to virus-infected cells (MacGregor et al., 1980; Ratcliffe et al., 1988), a phenomenon that may be enhanced by the presence of complement and/or virus-specific antibodies (Faden et al., 1984). It has been reported that mature neutrophils do produce a variety of mediators, some of which may have antiviral activity (Lloyd and Oppenheim, 1992; Cassatella, 1995).
There are multiple mechanisms employed by neutrophils to exert their antiviral effect. These cells may directly inhibit virus growth by the synthesis and release of mediators such as defensins (Ganz et al., 1985) or other factors (Rogers and Unanue, 1993). Virus-infected cells may be lysed via the production of reactive oxygen and/or nitrogen species (Smith, 1994). Neutrophils may also act indirectly by producing cytokines (Lloyd and Oppenheim, 1992; Cassatella, 1995), which in turn recruit other effector cells, such as T cells, NK cells, and macrophages, to the site of viral infection. Finally, neutrophils may collaborate with other defensive factors such as antibody (Siebens et al., 1979; Hashimoto et al., 1983; Ihara et al., 1984) or complement (Tsuru et al., 1987) to kill virus-infected cells via cellular cytoxicity reactions. It has been proposed that their recruitment is also important in inhibiting virus propagation early during infection (Schall et al., 1994). In particular, the nitric oxide (NO) produced by neutrophils has potent antiviral effects against VSV both in vitro and in vivo (Bi et al., 1995; Bi and Reiss, 1995). It has been reported that intratumoral neutrophil infiltration correlated with chemokine induction in VSV-treated mice, and neutrophil depletion before VSV administration eliminated uninfected tumor cell apoptosis and permitted extensive intratumoral viral replication and spread (Breitbach et al., 2007). Interestingly, the authors concluded that targeted recruitment of neutrophils to infected tumor beds enhances tumor cell killing and the effectiveness of oncolytic virus therapeutics.
M3 also binds chemokines that are known to be NK cell attractants. Natural killer cells are a major component of the innate immune system, and are crucial in early defense against viral infection. NK cells are found primarily in the blood and spleen, but can also occur in nonlymphoid tissues including lung, intestinal mucosa, and liver (Trinchieri, 1989). These cells represent a distinct population of cytotoxic lymphocytes, characterized by the CD16+ and/or CD56+ phenotype, that act as a first line of defense against invading pathogens and viruses as an integral component of the innate cellular immune response system before the launch of adaptive immune responses (Welsh, 1986; Brutkiewicz and Welsh, 1995; Baraz et al., 1999). NK cells are activated during viral infections (Biron et al., 1999), and they mediate direct lysis of target cells by releasing cytotoxic granules containing lytic enzymes, or by binding to apoptosis-inducing receptors on the target cell (Biron and Brossay, 2001; Orange et al., 2002). Indeed, intratumoral replication of wild-type VSV was enhanced by vector-mediated expression of gG, a vCKBP from equine herpesvirus-1 that reduced the intratumoral accumulation of NK cells (Altomonte et al., 2008). Interestingly, an in vitro study demonstrated that NK cells preferentially lyse human colon adenocarcinoma (Colo-205) cells infected with herpes simplex virus type 1 and vaccinia virus at an early stage of infection, thereby preventing viral dissemination to neighboring cells (Baraz et al., 1999). These data are consistent with our observation that the reduction in intratumoral NK cell contents through vector-mediated expression of a vCKBP gene resulted in enhanced intratumoral VSV replication and its oncolytic potency. On the other hand, the role of cellular immunity has been described in other oncolytic virus systems, particularly with respect to NK cells, T cells, and macrophages. Interestingly, there is evidence to suggest that these cells can augment the tumoricidal effects of oncolytic HSV as viral infection of tumor cells elicits a robust adoptive antitumor immune response (Todo et al., 2001; Thomas and Fraser, 2003; Benencia et al., 2005). These seemingly contradictory results from the current studies could be reconciled by the fact that NK cells have multiple functions in immune-competent animals. First, it is known to be an integral component of the innate cellular immune response system that kills virus-infected cells directly and it is this function that is attenuated by the viral chemokine-binding proteins. Another function of NK cells is to augment an adoptive immune response against cells expressing viral antigens, which has been documented in the oncolytic virus literature. In addition to NK cells and neutrophils, other immune cells such as mononuclear cells and macrophages are also involved in the suppression of oncolytic viral replication in tumors (Fulci et al., 2007; Kurozumi et al., 2007).
In this study, we provide conclusive evidence that suppression of the chemotaxis of antiviral inflammatory cells to lesions after treatment with VSV(MΔ51) can be achieved by vector-mediated expression of M3, a chemokine-binding protein from gammaherpesvirus-68, which in turn led to logarithmically elevated intratumoral viral titer, substantially enhanced oncolysis and tumor response, and survival prolongation. A long-term tumor-free survival rate of 50% was achieved by a single application via hepatic artery infusion of the recombinant oncolytic vector in this challenging multifocal HCC model with lesions that ranged from 1 to 10mm in diameter in the liver of syngeneic and immune-competent rats. This remarkable outcome could be attributed to the ability of M3 to evade host antiviral inflammatory responses, allowing unrestrained viral replication and cell killing within HCC tumors before the onset of neutralizing antibodies. There also appears to be a vector dose–response relationship with intratumoral inflammatory cell contents, viral replication, and tumor necrosis. Systemically administered cyclophosphamide (CPA), a generalized inhibitor of immunoglobulin production by B cells, could also partially suppress the innate antiviral immunity, resulting in an increase in viral replication within the tumors (Ikeda et al., 1999, 2000). The two approaches are distinct in that vector-mediated expression of a vCKBP can give rise to localized suppression of the innate immune response in the tumors as opposed to the systemic suppression achieved with CPA. On the other hand, there is the concern that a “super” virus might be created through recombination of rVSV(MΔ51)-M3 with a more virulent strain of VSV when both are present in the same host. Although VSV is not a pathogen in humans and is not present in the general population, to further minimize this potential risk patients treated with rVSV(MΔ51)-M3 in future clinical trials should be screened negative for existing VSV infection and kept in isolation until they are no longer shedding virus, which took only a few days for the rats treated in this study.
Although it is important to develop recombinant oncolytic virus vectors with greater potency for cancer treatment, strategies focused on inhibition of key players in antiviral defenses could be construed as a potentially dangerous proposition in the development of therapeutic agents for future clinical application. To address this concern, we demonstrated in immune-competent tumor-bearing rats that the rVSV(M?51)-M3 vector introduced no additional toxicities compared with a control rVSV(MΔ51)-LacZ vector, despite the fact that expression of M3 inhibited the chemotaxis of neutrophils and NK cells to virus-infected lesions. This is most likely to be secondary to the exquisite sensitivity of VSV to the type I interferon response in normal cells, which is unaltered in the case of vCKBP-expressing vectors. In addition, there are concerns that rVSV(MΔ51)-M3 might be a more virulent strain in the natural hosts (cattle and pigs). Although these concerns will need to be addressed experimentally in a separate study, the following considerations should also be made: (1) the exquisite sensitivity to type I interferon is not compromised in the new VSV vector and (2) the VSV-Indiana strain used in the current study is a relatively safe laboratory strain and a class II pathogen; there are numerous naturally occurring VSV strains that are class III pathogens and are much more virulent in their natural hosts. In conclusion, we have molecularly engineered a novel, highly effective, and safe rVSV(MΔ51)-M3 vector for cancer treatment, which exploits the antiinflammatory activities of a broad-spectrum chemokine-binding protein, M3. This vector demonstrates much superior intratumoral replication, oncolysis, tumor response, and survival prolongation in multifocal HCC-bearing rats. Thus, rVSV vectors encoding vCKBPs have the potential for development into effective and safe therapeutic agents for the treatment of HCC and possibly other types of cancer in the future, and are suitable for clinical translation. Because antiviral inflammatory responses are key components in the host's defense against most invading pathogens, the concept and technology developed here may be applicable in substantially enhancing the potency and efficacy of a wide spectrum of oncolytic virus agents in cancer treatment as well.
The work was supported in part by an NIH grant (CA100830 to S.L.C.W.), the German Cancer Aid (Max-Eder Research Program), and a Federal Ministry of Education and Research Grant 01GU0505 to O.E. The authors thank Dr. Sergio Lira for sharing the M3 clone and monospecific antibody to M3. The authors also thank Mr. Boxun Xie and Mrs. Yafang Wang for excellent technical assistance, and Ms. Yuemei Zhang for assistance in performing the morphological analyses.
As coinventors in a patent application, J.A., O.E., A.G.S., and S.L.C.W. may receive future royalties in accordance with institutional policies.