Oncolytic Activity of Vesicular Stomatitis Virus Is Effective against Tumors Exhibiting Aberrant p53, Ras, or Myc Function and Involves the Induction of Apoptosis

Abstract

We have recently shown that vesicular stomatitis virus (VSV) exhibits potent oncolytic activity both in vitro and in vivo (S. Balachandran and G. N. Barber, IUBMB Life 50:135–138, 2000). In this study, we further demonstrated, in vivo, the efficacy of VSV antitumor action by showing that tumors that are defective in p53 function or transformed with myc or activated ras are also susceptible to viral cytolysis. The mechanism of viral oncolytic activity involved the induction of multiple caspase-dependent apoptotic pathways was effective in the absence of any significant cytotoxic T-lymphocyte response, and occurred despite normal PKR activity and eIF2α phosphorylation. In addition, VSV caused significant inhibition of tumor growth when administered intravenously in immunocompetent hosts. Our data indicate that VSV shows significant promise as an effective oncolytic agent against a wide variety of malignant diseases that harbor a diversity of genetic defects.

Vesicular stomatitis virus (VSV) is a negative-stranded RNA virus and prototypic member of the family Rhabdoviridae that is extremely sensitive to the antiviral actions of the interferons (IFNs), a family of cytokines produced in response to viral infection that act by inducing the expression of more than 30 genes (9, 14). Indeed, the importance of the IFNs in controlling VSV infection has been underscored by research demonstrating that embryonic fibroblasts and mice lacking a functional IFN system or the IFN-inducible double-stranded RNA-dependent protein kinase PKR are extremely susceptible to VSV infection (2, 6a, 7a, 7b, 11). In addition to these studies, it has become apparent that although VSV replicates inefficiently in primary cells that contain a functional IFN/PKR system, this virus can replicate to high titers in a majority of immortalized and transformed tissue culture cell lines. It currently remains unclear whether aspects of IFN signaling and PKR action may be compromised in such malignant cells, thus affording a cellular environment that would facilitate viral replication. Nevertheless, we and others recently exploited these observations and demonstrated that VSV could also selectively inhibit, in vivo, the growth of tumors derived from transformed cells (1, 12). Our findings indicate that VSV could provide a potentially novel antitumor therapy.

Rat C6 glioblastoma cells are permissive to VSV.

As a start to analyzing the mechanisms of VSV-induced oncolysis, we examined the kinetics of VSV replication in the rat p53-defective C6 glioblastoma (C6) cell line (4). We selected this cell line for further study because we had previously shown that VSV causes the potent inhibition of C6-derived tumor growth in athymic nude mice (1). Accordingly, C6 cells were infected with VSV (multiplicity of infection [MOI] = 1), and lysates, prepared at various time points postinfection (p.i.), were analyzed for viral replication by immunoblotting for VSV proteins. As shown in Fig. 1A, viral protein synthesis was readily detectable within 24 h of infection and persisted for up to 48 h. Figure 1B shows that VSV G protein synthesis can be detected as early as 4 h p.i. Supernatants taken from infected cells at 24 and 48 h p.i. were examined for viral progeny yield by a standard plaque assay with BHK-21 cells and revealed mean titers of 5 × 10⁸ and 1 × 10⁹ PFU/ml, respectively. These data indicate that C6 cells are very permissive to VSV replication and yield high levels of progeny virus.

FIG. 1.

VSV replicates and induces apoptosis in C6 cells despite PKR activation and eIF2α phosphorylation. C6 cells were infected with VSV at an MOI of 1 and examined by immunoblotting for total VSV protein synthesis at the indicated time points p.i. using polyclonal antiserum to VSV (A) or analyzed for kinetics of viral replication by following synthesis of VSV G protein at the indicated time points (B). (C) C6 cells were infected with VSV (lane 2) for 4 h in the presence of [³²P]orthophosphate and subsequently analyzed by autoradiography or immunoblotting for phosphorylated PKR, total eIF2α, and serine 51-phosphorylated eIF2α. Equivalent levels of tubulin show that approximately equal amounts of protein were loaded in each lane. (D) C6 cells were infected with VSV and subsequently treated with 100 μM concentrations of each of the caspase inhibitors zVAD.fmk (zVAD), zIETD.fmk (zIETD), zLEHD.fmk (zLEHTD), and zDEVD.fmk (zDEVD). Forty-eight hours p.i., cells were assayed for viability by trypan blue exclusion. (E) Cells treated as in panel C were examined for apoptosis using TUNEL followed by fluorescence microscopy (magnification, ×80).

PKR is activated and phosphorylates eIF2α in response to VSV infection.

Since the previous work of members of our group with primary murine embryonic fibroblasts revealed that VSV infection induces the activation of PKR (2), which then potently inhibits viral protein synthesis by phosphorylating eIF2α on serine 51, we speculated that C6 cells might be permissive to VSV replication because of defective PKR function. To examine this hypothesis, we infected C6 cells with VSV (MOI = 100) in the presence of [³²P]orthophosphate for 4 h and analyzed PKR activity and subsequent eIF2α phosphorylation by autoradiography and immunoblotting, respectively. However, as shown in Fig. 1C, PKR was able to autophosphorylate and to phosphorylate eIF2α in response to VSV infection of C6 cells, implying that elements of cellular signaling downstream of, or parallel to, PKR activation and eIF2α phosphorylation are compromised in these cells.

VSV induces apoptosis in C6 cells.

To determine the nature of VSV-induced cytolysis, C6 cells were infected with VSV (MOI = 1) and treated with a 100 μM concentration of the broad-specificity caspase inhibitor zVAD.fmk or with 100 μM concentrations each of relatively specific inhibitors of caspases 8 (zIETD.fmk), 9 (zLEHD.fmk), and 3 (zDEVD.fmk). After 48 h p.i., viability was assessed by trypan blue exclusion analysis. As shown in Fig. 1D, zVAD.fmk was completely able to inhibit VSV-triggered cytolysis, indicating that this virus induces caspase-dependent programmed cell death in C6 cells. Ninety-six hours p.i., however, a significant percentage (∼20%) of infected cells stained positive for trypan blue despite the continuous presence of zVAD, indicating that VSV can trigger caspase-independent lysis, perhaps as a direct result of viral replication, as well. Terminal deoxynucleotidyltransferase-mediated fluorescein isothiocyanate-dUTP nick end labeling (TUNEL) (Fig. 1E) was used to confirm apoptosis. However, the fact that none of the specific caspase inhibitors was able to protect C6 cells from VSV-induced apoptosis implies that VSV triggers the activation of multiple caspase-dependent pathways in these cells. Indeed, members of our group had previously found that in immortalized murine fibroblasts, VSV-induced apoptosis was significantly dependent on the Apaf-1/caspase 9 pathway (3). Interestingly, while zVAD.fmk was completely able to inhibit VSV-induced apoptosis, it was incapable of preventing cell rounding following viral infection (Fig. 1E), which occurred with the same kinetics in the presence or absence of the caspase inhibitor. Furthermore, despite the block in apoptosis, titers of virus in zVAD-treated cells were similar to those in untreated cells (data not shown), implying that apoptosis is not required by the virus as a mechanism to facilitate its replication.

In vivo induction of apoptosis by VSV in C6 tumors mediates viral oncolysis.

We next examined whether VSV triggered oncolysis in vivo through the direct induction of apoptosis, similar to its effects in vitro. For these studies, athymic nu/nu mice were implanted with 2 × 10⁶ C6 cells subcutaneously (s.c.), and tumors were allowed to grow to a mean size of 25 mm² (approximately 8 to 10 days postimplantation). These tumors were then infected intratumorally (i.t.) with a single injection of 2 × 10⁷ PFU of VSV or with an equivalent amount of heat-inactivated (HI) VSV. As a control for injection, some tumors were injected with saline alone. Tumors were excised at various times p.i., and sections prepared from these tumors were examined histologically following hematoxylin and eosin (H&E) staining or for apoptosis by TUNEL (Fig. 2). VSV-infected tumors showed marked areas of cell death, characterized by shrunken cells with condensed, densely staining nuclei, within 48 h of VSV treatment (Fig. 2, top panels). In contrast, mock-infected or control tumors showed no evidence of any cell death during this period. To establish whether in vivo VSV-mediated C6 cell cytolysis was apoptotic in nature or not, sections similar to those analyzed by H&E staining were examined by TUNEL (Fig. 2, middle panels). Our results showed large numbers of TUNEL-positive nuclei in the VSV-infected tumors but not in HI VSV-treated tumors. Areas of TUNEL-positive cells corresponded to the patches of cell death evident in H&E-stained sections (Fig. 2). These results support the idea that the mechanism of VSV-mediated inhibition of C6 tumor growth is indeed through the induction of apoptosis of infected cells. To confirm that VSV was replicating in the same areas of the tumor that contained apoptotic cells, paraffin-embedded sections of VSV- or HI VSV-treated tumors were stained for VSV antigens using a polyclonal antiserum that recognizes all VSV proteins. As shown in Fig. 2 (bottom panels), areas staining positive for viral replication also contained several apoptotic nuclei and large areas of cell death. Some positive staining was observed in HI VSV-treated tumors 24 h p.i., but such staining could not be detected by 2 days p.i. and did not colocalize with any apoptotic cells (Fig. 2, bottom panels). Additionally, these sections were analyzed for the presence of inflammatory infiltration. H&E staining showed mild neutrophil infiltration in peritumoral areas by day 6 p.i. Staining with leukocyte common antigen revealed the absence of any lymphocytes in these peritumoral infiltrates or in the tumors themselves (data not shown). Coupled with the fact that the animals used in this study were athymic nude (nu/nu) mice which are almost completely deficient in T cells, our data would imply that the induction of apoptosis was not because of a CD8⁺ cytotoxic T-cell response. Collectively, our results indicate that the observed apoptotic effect was a direct result of viral replication and was unlikely to be mediated by an immune response to the virus.

FIG. 2.

VSV induces apoptosis in C6 tumors in vivo. Athymic nu/nu mice were implanted s.c. with 2 × 10⁶ C6 cells and subsequently infected i.t. with 2 × 10⁷ PFU of VSV/dose (or an equivalent amount of HI VSV as a control) i.t. after palpable tumors had formed. One, two, and six days p.i., tumors were excised, fixed in 4% paraformaldehyde, and sectioned. Paraffin-embedded sections were then stained with hematoxylin and eosin (H/E) and photographed by bright-field microscopy (magnification, ×89) (top panels), assayed for apoptosis using TUNEL, and photographed on a fluorescence microscope (magnification, ×89) (middle panels) or stained for VSV replication using an anti-VSV polyclonal antiserum and photographed by bright-field microscopy (magnification, ×178) (bottom panels).

VSV inhibits the growth of myc- and ras-transformed tumors in vivo.

To examine whether VSV oncolytic activity could be effective against other types of tumors, we examined the ability of VSV to induce the cytolysis of BALB/3T3 Ras or BALB/3T3 Myc cells in vitro. These cells were generated by transforming immortalized BALB/3T3 cells with the c-myc or K-ras oncogene, respectively (8). As shown in Fig. 3A, VSV was able to destroy the majority of these cells within 48 h following infection (MOI = 1). In fact, the immortalized BALB/3T3 parental cell line, while not tumorigenic, was also quite permissive to VSV, similar to other immortalized cells examined (3). High titers of progeny virus were detected in the supernatants from these cells, and the nature of cell death was confirmed as apoptotic by TUNEL (data not shown). Next, athymic nu/nu mice were implanted with 2 × 10⁶ BALB/3T3 Myc or BALB/3T3 Ras cells s.c. and subsequently inoculated with a single i.t. injection of 2 × 10⁷ PFU of VSV after palpable tumors had formed (6 to 10 days postimplantation). Tumor growth was monitored daily. As shown in Fig. 3B and C, administration of VSV but not HI VSV was able to markedly inhibit the growth of both myc- and ras-transformed tumors in vivo. Virtually no virus (<10 PFU/organ) was present in the brain, spleen, lungs, kidneys, or liver of infected animals. Residual virus (2 × 10⁴ to 5 × 10⁵ PFU/g) was, however, detectable in the tumors themselves at the end point of the experiment. These results highlight the potential efficacy of VSV as an anticancer therapy against a variety of tumors, irrespective of their genetic backgrounds or the oncogenic events that led to their transformation. Indeed, the genetic lesions of the tumors described thus far in this study (p53 deficiency, Myc overexpression, or Ras overexpression) are found in >90% of all human malignancies, indicating the potential application of VSV against a wide range of malignant disorders (15).

FIG. 3.

VSV inhibits growth of myc- and ras-transformed tumors in nu/nu mice, can repress tumor growth when administered distally, and inhibits growth of syngeneic tumors in immunocompetent mice. (A) BALB/3T3, BALB/3T3 Myc, and BALB/3T3 Ras cells were treated with or without 1,000 U of alpha/beta murine IFN/ml for 18 h and subsequently infected with VSV at an MOI of 10. Viability was assessed 24 h p.i. by trypan blue exclusion. (B) nu/nu mice with orthotopic s.c. tumors derived from myc-transformed BALB/3T3 cells (n = 5) were injected i.t. with 2 × 10⁷ PFU of VSV/dose. Control tumors (n = 5) received equivalent amounts of HI VSV. Tumor volumes were measured daily for a period of 2 weeks. (C) nu/nu mice with orthotopic s.c. tumors derived from ras-transformed BALB 3T3 cells (n = 5) were injected i.t. with 2 × 10⁷ PFU of VSV/dose. Control tumors (n = 5) received equivalent amounts of HI VSV. Tumor volumes were measured for a period of 10 days, at which time the tumor burden of control-infected animals became excessive. Means ± standard errors of the means (S.E.M.) are given. (D) nu/nu mice were implanted with 2 × 10⁶ C6 cells/flank s.c. into both the right and left rear flanks of each mouse. After palpable tumor formation, the right flank tumor (n = 5) was infected with VSV i.t. (2 × 10⁷ PFU/tumor) or with HI VSV, and all tumor volumes were measured for a period of 15 days. Results are given as means ± S.E.M. (E) nu/nu mice were implanted with 2 × 10⁶ C6 cells s.c. and injected after palpable tumors had formed (n = 5) i.v. with VSV in three serial doses of approximately 2.5 × 10⁷ PFU/dose every 2 days (arrows), and tumor growth was monitored daily. Control tumors (n = 5) received equivalent amounts of HI VSV. Tumor volumes were measured daily for a period of 13 days. (F) C3H mice bearing syngeneic Ag104 sarcoma-derived orthotopic s.c. tumors (n = 5) were injected with three doses of 2.5 × 10⁷ PFU of VSV/dose 3 days apart. Control tumors (n = 5) received equivalent amounts of HI VSV. Tumor volumes were measured for a period of 2 weeks. Results are given as means ± S.E.M.

VSV represses neoplastic growth when administered at sites distal to the tumor.

Accessibility of the tumor to a therapeutic agent is a major limitation of a successful antitumor therapy. We therefore examined whether VSV introduced into one tumor was capable of spreading beyond the infected tumor and infecting the contralateral tumor without replicating in normal tissue. Accordingly, 2 × 10⁶ C6 cells/site were implanted s.c. bilaterally into the rear flanks of nu/nu mice. After palpable tumors (25 mm²) had formed, the tumor on the right flank of each mouse was inoculated i.t. with a single injection of 2 × 10⁷ PFU of VSV/tumor or with an equivalent amount of control heat-inactivated virus, and bilateral tumor volumes were monitored daily. As shown in Fig. 3D, VSV, but not HI VSV, administered to the right tumor was able to cause a significant, albeit not complete, repression of growth of the left flank tumor, indicating that VSV was circulated from tumor to tumor. In agreement with this observation, virus (mean titer, ∼3 × 10⁵ PFU/g) was detectable in the left flank tumors of all VSV-inoculated mice. Significantly, no virus (<10 PFU/organ) was detectable in any of the organs (brain, spleen, kidneys, liver, lungs, and heart) examined. This study shows the potential of VSV as a therapeutic strategy against metastatic disease. We next examined whether VSV was capable of repressing tumor growth when administered intravenously (i.v.). Accordingly, nu/nu mice were implanted with 2 × 10⁶ C6 cells s.c., and tumors were allowed to grow to a mean size of 25 mm². At this time, mice were injected i.v. through a tail vein with VSV or HI VSV with three serial doses of approximately 2.5 × 10⁷ PFU/dose every 2 days, and tumor growth was monitored daily. Tumors in VSV-treated mice initially grew at the same rate as tumors in control-treated animals. However, 48 h after the last dose of i.v.-administered VSV, tumors in virus-treated mice showed markedly repressed growth and did not significantly increase in size over the next 7 days (Fig. 3E). The experiment was terminated because VSV-infected animals began to show hind-limb paralysis, which is characteristic of VSV disease in these immunodeficient mice (7) following inoculation with high doses of virus. Recent studies in our laboratory, however, have shown that immunocompetent mice of several strains remain disease free despite receiving high doses of VSV i.v. Importantly, moderately high titers of virus (mean titer, ∼2.5 × 10⁵ PFU/g) were detected in all tumors from the i.v.-infected mice, indicating that VSV is capable of reaching and replicating in tumor tissue from distal sites of administration. In this experiment, significant virus was detected in the brains of infected mice (mean titer, ∼4 × 10⁴). No virus was detected in any of the other organs (liver, lungs, kidneys, spleen, and heart) obtained from these animals.

VSV can inhibit growth of syngeneic tumors in immunocompetent mice.

To date, all data demonstrating the potential of VSV as an oncolytic agent have come from studies performed with immunodeficient animals. Since the host immune response to viral therapeutics can seriously affect the efficacy of such treatment by eliminating the virus prior to killing of target cells, we also examined the effects of VSV in syngeneic tumors grown in immunocompetent mice. For this investigation, we used Ag104 cells, which, following inoculation, form sarcomas in syngeneic C3H mice (14a). In vitro studies confirmed that VSV induced the cytolysis of Ag104 cells (data not shown). Ag104 cells (3 × 10⁶) were subsequently implanted into the rear flank of the animals, and palpable tumor formation occurred within 8 to 10 days (approximately 0.25 cm²). Approximately 2.5 × 10⁷ PFU of VSV/tumor/dose was inoculated i.t. in three serial administrations 3 days apart. Control mice were injected with equivalent amounts of HI VSV, and tumor growth was monitored daily. This experiment revealed significant growth arrest of all tumors treated with viable VSV (Fig. 3F). As was the case with i.t.-inoculated nu/nu mice, no sickness or overt symptoms of VSV disease were evident in any of the inoculated animals, and no virus could be detected in the examined organs. These data show that VSV has potential as an oncolytic agent in the treatment of neoplastic disease in immunocompetent hosts.

Previous attempts at using viral agents in antitumor therapy have been limited by restricted tropism, the transforming potential of the virus itself, inefficient replication in the tumor, or neutralization by the host immune response. VSV, however, replicates rapidly to high titers in virtually all transformed cells, has no known transforming potential, is nonhazardous to humans, and has been extensively characterized immunobiologically. Besides, the genetic malleability of this virus makes it an attractive vector for the delivery of proapoptotic and immunostimulatory genes into the tumor, thus potentially enhancing its oncolytic activity. The studies presented herein demonstrate that VSV is capable of repressing tumor growth when administered i.v. Furthermore, VSV can spread to and inhibit the growth of distal tumors when injected i.t. and can induce the oncolysis of tumors in immunocompetent hosts without replicating to any significant extent in normal tissue. Collectively, our data show that VSV offers significant promise as a novel therapy for cancer.