Our preclinical and clinical trials using a replication-defective adenoviral vector expressing IFN-β have shown promising results for the treatment of malignant mesothelioma. Based on the hypotheses that a replication-competent vesicular stomatitis virus (VSV) oncolytic vector would transduce more tumor cells in vivo, that coexpression of the immunostimulatory IFN-β gene would enhance the immune-based effector mechanisms associated both with regression of mesotheliomas and with VSV-mediated virotherapy, and that virus-derived IFN-β would add further safety to the VSV platform, we tested the use of IFN-β as a therapeutic transgene expressed from VSV as a novel treatment for mesothelioma. VSV-IFN-β showed significant therapy against AB12 murine mesotheliomas in the context of both local and locoregional viral delivery. Biologically active IFN-β expressed from VSV added significantly to therapy compared with VSV alone, dependent in part on host CD8+ T-cell responses. Immune monitoring suggested that these antitumor T-cell responses may be due to a generalized T-cell activation rather than the priming of tumor antigen–specific T-cell responses. Finally, IFN-β also added considerable extra safety to the virus by providing protection from off-target viral replication in nontumor tissues and protected severe combined immunodeficient mice from developing lethal neurotoxicity. The enhanced therapeutic index provided by the addition of IFN-β to VSV therefore provides a powerful justification for the development of this virus for future clinical trials. [Cancer Res 2009;69(19):7713–20]

Malignant mesothelioma is an aggressive neoplasm of the pleura or peritoneum associated with asbestos exposure (1). Due to a lack of effective therapy, few patients survive beyond 2 years from onset (1), highlighting the need for new therapeutic approaches. In this respect, oncolytic viruses are being developed to replicate selectively in tumor cells leading to tumor cell lysis while sparing normal cells. In theory, use of a replication-competent oncolytic virus would require only low levels of seeding in a tumor to initiate spreading infections to cover the tumor comprehensively (2, 3). We have previously shown that vesicular stomatitis virus (VSV), which replicates in the cytoplasm and is highly lytic, is an effective oncolytic virus in various tumor models (4, 5). The potential of VSV as an oncolytic agent was suggested because VSV infection of normal cells induces type-I IFN responses (IFN-α/β), blocking viral replication and extinguishing infection. However, many tumor cells have defects in their IFN response (68), allowing free-ranging infection and lysis (911). VSV is indeed a potent oncolytic agent against a variety of both human and murine tumors (5, 917). In addition, we have shown that the efficacy of VSV-mediated virotherapy in immunocompetent mice is not solely attributable to direct viral replication and tumor cell lysis but is also dependent on host-derived CD8+ and natural killer (NK) cells (4).

The accessibility of malignant pleural mesothelioma makes it a good candidate for oncolytic viral therapy with the need for only locoregional delivery of the virus (18). Use of an oncolytic virus extends our previous work in which inclusion of the IFN-β gene into a replication-defective adenoviral vector led to significant therapy following intratumoral (i.t.) injections (19, 20). Expression of IFN-β generated tumor regressions and cures, dependent on its pleiotropic antitumor effects including enhancement of innate immune responses, its antiproliferative activities (21), and the priming of T-cell responses (19, 22, 23), and our resulting clinical trials have generated promising results (24).

Taken together, these data (19, 20, 24, 25) suggest that expression of IFN-β (21) will further enhance the antitumor immunotherapeutic effects of the replication-competent oncolytic VSV. Moreover, inclusion of IFN-β within the VSV vector should also increase the safety of VSV in the event that it were to infect normal cells (21). Thus, because normal cells have intact responses to IFN-α/β (unlike many tumor cells), which shut down VSV replication rapidly both in vitro and in vivo, by adding further levels of expression of IFN-β from the virus, we predicted that the toxicity of VSV-IFN-β would be further diminished without major effect on the efficacy against tumor. This is an important issue given the findings that at high doses, VSV infection can lead to lethal neurotoxicity in some mouse models (especially immunodeficient mice; refs. 10, 21).

Here, we tested IFN-β as a therapeutic transgene expressed from VSV as a novel treatment for mesothelioma based on the potential of VSV-IFN-β to increase immunotherapeutic efficacy, enhance tumor cell transduction, and increase the safety of previous approaches using VSV alone. We show that VSV-IFN-β efficiently lyses the murine mesothelioma AB12 cell line, and that added expression of IFN-β does not significantly interfere with VSV replication. Locoregional delivery of VSV-IFN-β generated tumor regressions, dependent, at least in part, on virus-induced activation of CD8+ T cells. Finally, we confirmed that IFN-β expressed from VSV added considerable extra safety to the virus and protected severe combined immunodeficient (SCID) mice from lethal neurotoxicity. These data support the development of the VSV-IFN-β vector for use in patients with malignant mesothelioma.

Cell lines. The human mesothelioma line MSTO-211H was purchased from the American Tissue Type Collection. The murine mesothelioma cell line AB12 was provided by Dr. Bruce Robinson (University of Western Australia, Perth, Australia; ref. 26). B16ova cells, a murine melanoma cell line, were derived from the parental cell by transduction of cDNA encoding chicken ovalbumin gene (27).

Viral strains. The VSV-mIFN-β and VSV-hIFN-β vectors were originally described by Obuchi and colleagues (21). VSV-green fluorescent protein (GFP) was generated as previously described (4). Viral stocks were manufactured by the Core Viral Facility of Mayo Clinic. VSV-GFP is referred to as VSV.

MTT assays. MTT assays were done with VSV-mIFN-β and VSV-GFP. Cells were plated in quadruplicate on 96-well plates and infected with different multiplicities of infection (MOI). Cell viability was assessed at the indicated time points per manufacturer's instruction (Promega).

ELISpot analysis for IFN-γ secretion. Spleens or tumor-draining lymph nodes were harvested from mice at the indicated times. For ELISpot assays, 1 × 105 cells were plated in triplicate on a 96-well plate and restimulated for 48 h at 37°C under different conditions (all peptides were at 5 μg/mL). Peptide-specific, IFN-γ–positive spots were detected per manufacturer's protocol (Mabtech, Inc.) and quantified by a computer-assisted image analyzer.

The following synthetic peptides were synthesized at the Mayo Foundation Core Facility: EGSRNQDWL, mouse gp100; SIINFEKL, chicken OVA; SVYDFFVWL, mouse TRP-2; and RGYVYQGL, VSV-N protein.

In vivo studies. All procedures were approved by the Mayo Foundation or the University of Pennsylvania Animal Care and Use Committee in compliance with the Guide for the Care and Use of Laboratory Animals. To establish s.c. tumors, 5 × 105 B16ova cells, 1 × 106 AB12 cells, or 1 × 106 MSTO-211H cells in 100 μL of PBS were injected into the flank of C57Bl/6, BALB/c, or SCID mice, respectively. Once s.c. tumors reached ∼200 mm3 in size, i.t. injections were done with saline, 5 × 108 plaque-forming units (pfu; C57Bl/6) or 6.6 × 108 pfu in 100 μL of each vector (VSV-mIFN-β and VSV-hIFN-β), once weekly for 2 (SCID, C57Bl/6) or 3 consecutive weeks (BALB/c). Tumors were measured twice a week, and mice were euthanized if toxicity was evident or tumor burden exceeded 1,500 mm3. To establish i.p. tumors, 3.5 × 105 cells were injected i.p. On day 4, growth of the i.p. tumors was confirmed, and injections were done with saline or 6.6 × 108 pfu in 100 μL of each virus (VSV-mIFN-β and VSV-hIFN-β).

In vivo depletion of CD8+ T cells. BALB/c mice received i.p. injections of 200 μg of purified monoclonal antibodies purified from the anti-CD8+ hybridoma 53-6.7 (American Type Culture Collection). Injections were administered 3 d and 1 d before inoculation with AB12 cells. Thereafter, a maintenance dose of antibody was injected i.p. every 7 d throughout the entire experimental period to ensure depletion. CD8+ T-cell depletion was confirmed by flow cytometry of splenic suspensions at the time of tumor injection and weekly afterward. On days 11 and 18, mice received 6.6 × 108 pfu VSV-mIFN-β or PBS. Tumors were measured twice a week and mice were euthanized if toxicity was evident or tumor burden exceeded 1,500 mm3.

Flow cytometry and IFN-γ intracellular staining assay. For analysis of phenotype, 1 × 106 cells were washed in PBS containing 0.1% bovine serum albumin (wash buffer), resuspended in 50 μL of wash buffer, and exposed to conjugated primary antibodies for 30 min at 4°C. The OVA-iTag-H-2Kb-SIINFEKL-PE conjugated tetramer was used per manufacturer's protocol (Beckman Coulter). Cells were washed and resuspended in 500 μL PBS containing 4% formaldehyde and analyzed by flow cytometry, and data were analyzed using FlowJo software. For intracellular staining, single-cell suspensions were prepared from tumors harvested (three mice per group) at the indicated times. IFN-γ production in response to antigen was measured by incubation with peptides (5 μg/mL) in the presence of Golgi Plug for 4 h. Cells were stained, fixed, and permeabilized for intracellular staining using a Cytofix/Cytoperm kit (BD Biosciences) per manufacturer's instructions.

Statistical analyses. For comparison of two individual data points, two-sided Student's t test was applied to determine statistical significance. ANOVA with post hoc testing was used for groups of three or more. Survival curves were plotted according to the Kaplan-Meier method, and statistical significance in the different treatment groups was compared using the log-rank test.

VSV-mIFN-β is potently cytotoxic to mesothelioma cells in vitro. Both VSV (Fig. 1A) and VSV-mIFN-β (Fig. 1B) induced rapid and extensive cell killing following infection of AB12 cells at MOIs ranging from 0.01 to 10, associated with ongoing replication of the viruses (Fig. 1C and D). We observed a moderate but significant reduction in the rate of viral replication of VSV-mIFN-β compared with VSV at the lowest MOI of infection (Fig. 1C and D), additionally reflected in a slightly slower rate of tumor cell killing (Fig. 1A and B), suggesting that AB12 cells may retain a slight degree of responsiveness to IFN-β–mediated inhibition of viral replication. By 72 hours after infection, however, more than 99% of AB12 cells had been eradicated by both viruses (Fig. 1A and B).

Figure 1.

VSV-IFN-β is potently cytotoxic to mesothelioma cells in vitro. AB12 cells plated in 96-well plates were infected at various MOIs with either (A) VSV or (B) VSV-mIFN-β. At specific time points, an MTT assay was done per manufacturer's instructions. Data are plotted as a fraction of control cell survival with SD. AB12 cells were infected with VSV (C) or VSV-mIFN-β (D) for 1 h at 37°C at the indicated MOIs. After incubation, cells were washed and, at the indicated times points, samples were taken for a plaque-forming unit assay.

Figure 1.

VSV-IFN-β is potently cytotoxic to mesothelioma cells in vitro. AB12 cells plated in 96-well plates were infected at various MOIs with either (A) VSV or (B) VSV-mIFN-β. At specific time points, an MTT assay was done per manufacturer's instructions. Data are plotted as a fraction of control cell survival with SD. AB12 cells were infected with VSV (C) or VSV-mIFN-β (D) for 1 h at 37°C at the indicated MOIs. After incubation, cells were washed and, at the indicated times points, samples were taken for a plaque-forming unit assay.

Close modal

VSV-IFN-β has antitumor activity against established tumors. Consistent with the in vitro results of Fig. 1, direct injection of VSV-mIFN-β into established s.c. AB12 tumors in immunocompetent mice also generated significant antitumor activity compared with controls (P < 0.01 at day 32; Fig. 2A). This reduction in the rate of tumor growth translated into significantly increased survival in VSV-mIFN-β–treated mice, with four of eight mice cured of their tumors (Fig. 2A), compared with zero of eight in the PBS-treated group. In addition, all of these long-term survivors rejected a later challenge of live AB12 tumor cells. We predict that the antitumor immunity observed in Fig. 2A is specific against AB12 cells, although we did not have sufficient numbers of survivors to rechallenge these tumor-cured animals with a different, nonmesothelioma tumor cell line.

Figure 2.

In vivo therapy is dependent on biological activity of IFN-β. Groups (n = 8) of BALB/c mice bearing either s.c. (A) or i.p. (B) AB12 tumors were treated with saline, VSV-mIFN-β, or VSV-hIFN-β at a dose of 6.6 × 108 pfu, once a week for 3 wk. Administration of VSV-mIFN-β led to significant tumor regressions in the s.c. model (P < 0.01) in comparison with the saline group, whereas VSV-h-IFN-β does not show any significant therapy. However, both VSV-mIFN-β and VSV-hIFN-β significantly prolong survival in the i.p. model over saline (P < 0.0001 and P < 0.01, respectively).

Figure 2.

In vivo therapy is dependent on biological activity of IFN-β. Groups (n = 8) of BALB/c mice bearing either s.c. (A) or i.p. (B) AB12 tumors were treated with saline, VSV-mIFN-β, or VSV-hIFN-β at a dose of 6.6 × 108 pfu, once a week for 3 wk. Administration of VSV-mIFN-β led to significant tumor regressions in the s.c. model (P < 0.01) in comparison with the saline group, whereas VSV-h-IFN-β does not show any significant therapy. However, both VSV-mIFN-β and VSV-hIFN-β significantly prolong survival in the i.p. model over saline (P < 0.0001 and P < 0.01, respectively).

Close modal

VSV-mIFN-β was also therapeutically effective in a model of locoregional virus delivery to treat i.p. seeded AB12 tumors (Fig. 2B). Unlike the s.c. model, there were no long-term cures. However, survival of tumor-bearing mice treated with VSV-IFN-β (median survival, 54 days) was extended significantly over control-treated animals (median survival, 26 days; P < 0.01).

Taken together, these data show that VSV-mIFN-β is an effective agent against AB12 mesothelioma tumors both in vitro and in vivo in the contexts of both local and locoregional delivery.

In vivo therapy is dependent on biological activity of IFN-β. Based on our previous studies using a replication-defective adenovirus expressing IFN-β (19, 20, 24, 25), our hypothesis was that addition of the IFN-β gene to the replication-competent VSV would further enhance the immunostimulatory activity of virus-mediated tumor cell killing. To test this hypothesis, we constructed a VSV expressing the human IFN-β gene, which is not active in the mouse. Therefore, the efficacy of VSV-mIFN-β and VSV-hIFN-β in immunocompetent mice would provide a closely matched comparison between immune-based components of antitumor therapy and the effects of direct VSV-mediated oncolysis. We confirmed that both the VSV-mIFN-β and the VSV-hIFN-β viruses have very similar profiles of replication in, and cytotoxicity to, AB12 cells in vitro (not shown). As before, direct i.t. injection of VSV-mIFN-β into s.c. AB12 tumors significantly improved antitumor therapy compared with PBS alone (Fig. 2A; P < 0.01). In contrast, i.t. injection of VSV-hIFN-β gave no significant therapy over PBS (Fig. 2A; P = 0.27), suggesting that a major component of the therapy in vivo associated with VSV-IFN-β is contributed by immune reactivity of local IFN-β expression, as opposed to direct oncolysis by the virus.

We also observed a trend toward increased efficacy of VSV-mIFN-β compared with VSV-hIFN-β in the model of locoregional delivery of virus to i.p. tumors, although this difference did not reach statistical significance (P = 0.54; Fig. 2B). Both VSV-hIFN-β and VSV-mIFN-β were, however, significantly more therapeutic than PBS (P < 0.01 and P < 0.0001, respectively).

VSV-mediated expression of IFN-β enhances therapy through CD8+ T cells. To further define the immune-mediated mechanisms by which IFN-β enhances the therapy associated with i.t. VSV injection, the experiments of Fig. 2 were repeated in SCID mice. Intratumoral injection of VSV-IFN-β significantly reduced the rate of tumor growth in SCID mice compared with control injections (P < 0.001; Fig. 3A). However, AB12 tumors develop at a faster rate in SCID mice than in BALB/c mice. Therefore, by measuring the percentage of growth inhibition between control- and VSV-IFN-β–treated groups, we observed that there was a significant difference in the potency of VSV-IFN-β between immunocompetent and immunodeficient strains (P < 0.0001 for all days; Fig. 3B). Thus, on day 14 after AB12 injection, whereas SCID mice showed a growth inhibition of ∼35% of tumor growth, immunocompetent BALB/c mice had a growth inhibition of 60% of tumor growth. Therefore, the absence of T and/or B cells in SCID mice was responsible for a loss of ∼50% of the therapy associated with i.t. injection of VSV-IFN-β in the AB12 model from 14 days onward.

Figure 3.

VSV-mediated expression of IFN-β enhances therapy through the activity of CD8+ T cells. Groups (n = 8) of BALB/c SCID (A) or CD8+ T-cell–depleted (B) mice bearing s.c. tumors were treated with saline or VSV-mIFN-β (6.6 × 108 pfu), once a week for 2 wk. Because AB12 tumors develop at a faster rate in the SCID and CD8-depleted mice than in the BALB/c mice, a percentage of growth inhibition between control- and VSV-IFN-β–treated groups was calculated as [(VcontrolVVSV-mIFN-β)/Vcontrol] × 100%.

Figure 3.

VSV-mediated expression of IFN-β enhances therapy through the activity of CD8+ T cells. Groups (n = 8) of BALB/c SCID (A) or CD8+ T-cell–depleted (B) mice bearing s.c. tumors were treated with saline or VSV-mIFN-β (6.6 × 108 pfu), once a week for 2 wk. Because AB12 tumors develop at a faster rate in the SCID and CD8-depleted mice than in the BALB/c mice, a percentage of growth inhibition between control- and VSV-IFN-β–treated groups was calculated as [(VcontrolVVSV-mIFN-β)/Vcontrol] × 100%.

Close modal

Consistent with the results of Fig. 3A and B, tumors grown in mice depleted of CD8+ T cells and treated with VSV-IFN-β were ∼25% smaller than control-treated tumors at day 14, compared with a 60% growth inhibition of tumors treated identically in immunocompetent mice (P < 0.0001; Fig. 3C and D). These data suggest that about half of the therapy induced by i.t. injection of VSV-IFN-β is dependent on an intact CD8+ T-cell compartment.

Expression of IFN-β does not prime detectable tumor antigen–specific T-cell responses. Given this T-cell dependence, it was of interest to further study this IFN-β–induced immune response. Because the AB12 cell line has no known specific tumor antigens, we conducted further mechanistic studies using a more well-immunologically characterized cell line (B16ova) that additionally expressed a neo-antigen (chicken ovalbumin) that allowed the use of tetramers. We have previously analyzed the immune responses of this model to wild-type VSV treatment (4). We first showed that B16ova tumors responded to VSV-mIFNβ in a similar fashion as AB12 cells (Fig. 4A). Next, we assayed splenocytes of VSV- or VSV-IFN-β–treated mice for the magnitude and specificity of responses against viral or tumor-associated antigens. To track antigen responses, we took advantage of the fact that C57Bl/6 mice bearing B16ova tumors generate T-cell–specific responses against the well-characterized SIINFEKL epitope of the model OVA tumor-associated antigen, which can readily be monitored.

Figure 4.

Expression of IFN-β does not prime detectable tumor antigen–specific T-cell responses. A, groups (n = 8) of C57Bl/6 mice bearing s.c. tumors were treated with saline or VSV-mIFN-β (5 × 108 pfu) on days 8 and 12 after tumor inoculation. Administration of VSV-mIFN-β led to significant tumor regressions in the s.c. model (P < 0.05) in comparison with the saline group. B to E, groups of C57Bl/6 mice (n = 4) bearing s.c. B16ova tumors were i.t. treated twice with heat-inactivated VSV, VSV, or VSV-mIFN-β at 5 × 108 pfu. Mice were sacrificed 7 d after the final injection and organs were harvested. B, lymph nodes were pulsed with no peptide or with ova, VSV-N, TRP-2, or gp100 peptides and assayed for IFN-γ–producing cells by ELISpot. Points, average of duplicates or triplicates from each individual mouse; bars, SD. C, tumor cells were incubated for 4 h in the presence of Golgi Plug with or without the indicated peptides and analyzed by flow cytometry for intracellular IFN-γ. Additionally, tumor cells (D) and splenocytes (E) were analyzed by flow cytometry for H2-Kb–restricted SIINFEKL tetramer–positive CD8+ T cells.

Figure 4.

Expression of IFN-β does not prime detectable tumor antigen–specific T-cell responses. A, groups (n = 8) of C57Bl/6 mice bearing s.c. tumors were treated with saline or VSV-mIFN-β (5 × 108 pfu) on days 8 and 12 after tumor inoculation. Administration of VSV-mIFN-β led to significant tumor regressions in the s.c. model (P < 0.05) in comparison with the saline group. B to E, groups of C57Bl/6 mice (n = 4) bearing s.c. B16ova tumors were i.t. treated twice with heat-inactivated VSV, VSV, or VSV-mIFN-β at 5 × 108 pfu. Mice were sacrificed 7 d after the final injection and organs were harvested. B, lymph nodes were pulsed with no peptide or with ova, VSV-N, TRP-2, or gp100 peptides and assayed for IFN-γ–producing cells by ELISpot. Points, average of duplicates or triplicates from each individual mouse; bars, SD. C, tumor cells were incubated for 4 h in the presence of Golgi Plug with or without the indicated peptides and analyzed by flow cytometry for intracellular IFN-γ. Additionally, tumor cells (D) and splenocytes (E) were analyzed by flow cytometry for H2-Kb–restricted SIINFEKL tetramer–positive CD8+ T cells.

Close modal

Intratumoral injection of VSV into B16ova tumors seemed to prime increased numbers of T cells specific for the immunodominant epitope of the viral N protein, compared with mice in which tumors were injected with control heat-inactivated VSV (Fig. 4B,, column 2 versus column 7), although there is no statistically significant increase in the N-specific T-cell response compared with the control in mice injected with VSV (Fig. 4B,, column 6 versus column 7). Coexpression of IFN-β from the injected VSV generally enhanced the frequencies of T cells specific for the viral N protein compared with mice treated with VSV alone (Fig. 4B,, column 12). However, we also consistently observed that splenocytes from mice treated with either VSV or VSV-mIFN-β exhibited a generalized, non–antigen-specific activation, as evidenced by the elevated frequencies of IFN-γ–producing spots generated in response to stimulation with no added peptide (Fig. 4B,, column 11). Relative to the frequencies of antigen nonspecific T-cell responses (no stimulating peptide), we did not observe a significant increase in the frequency of OVA-(SIINFEKL)–specific T cells in the spleens or lymph nodes of mice treated with VSV-IFN-β (Fig. 4B,, column 13). Moreover, frequencies of OVA-specific T cells were not significantly different in mice whose tumors were treated with VSV-IFN-β compared with those treated with VSV, although between experiments there was a trend toward increased levels of SIINFEKL-specific T cells (Fig. 4B , column 8 versus column 13).

Consistent with these data, although i.t. injection of VSV significantly enhanced the magnitude of the antiviral T-cell response seen within tumors (Fig. 4C,, column 6 versus column 7), there was no significant increase in the T-cell response to either the nonself OVA–specific or the self TRP-2– or gp100-specific tumor-associated antigens (Fig. 4C,, column 6 versus columns 810). Coexpression of IFN-β from the virus did not enhance either the antiviral T-cell response (relative to treatment with VSV) or the anti-OVA response within the tumor (Fig. 4C,, column 8 versus column 13). Similarly, although we observed a trend toward increased number of tetramer-positive OVA-specific T cells infiltrating tumors treated with VSV compared with control-treated animals, this was not significant (Fig. 4D). Addition of IFN-β expression from the virus did not increase the frequency of tumor antigen–associated T-cell responses in either the tumor or the spleens of treated mice compared with control- or VSV-treated mice (Fig. 4D and E).

Taken together, these data indicate VSV i.t. injection primes viral-specific T-cell responses, which can be enhanced in their frequency by the coexpression of IFN-β from the virus. In addition, VSV i.t. injection is associated with enhanced levels of T-cell activation but through a non–antigen-specific mechanism. Finally, we could not show the priming of T-cell responses specifically either against the endogenous TRP-2 or gp100 tumor antigens or against the artificial nonself OVA tumor antigen at levels higher than controls.

IFN-β provides enhanced safety against VSV-induced neurotoxicity. A principal rationale for the inclusion of the IFN-β gene within the VSV platform was to increase the safety of this virus should it become disseminated within a patient and infect normal cells—in which case, the expression of IFN-β would enhance the antiviral response of normal cells and extinguish extraneous viral replication. The potential toxicities of VSV, including neurovirulence, are most potent when the virus is administered to immunodeficient mice. For this reason, human MSTO-211H mesothelioma tumors were grown in SCID mice, which were then treated either with VSV-hIFN-β (in which the human IFN-β confers no protection against viral replication in a mouse) or with VSV-mIFN-β (in which the IFN-β will be active in the host mouse at protecting normal tissues from viral replication). Significantly, all tumor-bearing mice treated with VSV-hIFN-β showed significant tumor regressions compared with mice with control-treated tumors (Fig. 5), but these mice also exhibited unacceptable neurotoxicity by day 40, necessitating termination of the experiment in that group. In contrast, treatment of tumors with VSV-mIFN-β led to similar levels of tumor growth inhibition by day 40 after tumor seeding (Fig. 5) but was not associated with any overt viral-associated toxicities. This lack of toxicity enabled the experiment to be continued, under which circumstances tumor regressions continued and long-term survivors were generated (Fig. 5). These results confirm that inclusion of a biologically active IFN-β in the VSV platform adds a significant safety benefit of protection against VSV-associated neurologic toxicity and allows for increased efficacy to be manifested by protecting the host from adverse side effects.

Figure 5.

IFN-β provides enhanced safety against VSV-induced neurotoxicity. Groups (n = 8) of SCID mice bearing s.c. MSTO-211H tumors were i.t. treated with saline, VSV-mIFN-β, or VSV-hIFN-β at a dose of 6.6 × 108 pfu, once weekly for 2 wk. Administration of both viruses significantly reduces tumor growth in the mice compared with control (P < 0.0001 for both). However, mice treated with VSV-hIFN-β displayed neurotoxicity and were sacrificed around day 40.

Figure 5.

IFN-β provides enhanced safety against VSV-induced neurotoxicity. Groups (n = 8) of SCID mice bearing s.c. MSTO-211H tumors were i.t. treated with saline, VSV-mIFN-β, or VSV-hIFN-β at a dose of 6.6 × 108 pfu, once weekly for 2 wk. Administration of both viruses significantly reduces tumor growth in the mice compared with control (P < 0.0001 for both). However, mice treated with VSV-hIFN-β displayed neurotoxicity and were sacrificed around day 40.

Close modal

Based on the hypothesis that expression of IFN-β would enhance the immunotherapeutic efficacy of VSV while decreasing its systemic toxicity, we tested a replication-competent oncolytic VSV expressing IFN-β in a murine model of mesothelioma. We show here that AB12 cells are highly sensitive to both VSV and its IFN-β–expressing derivative in vitro (Fig. 1). This sensitivity to VSV-mIFN-β is also reproduced in vivo after both i.t. and loco-regional delivery (Fig. 2). Consistent with our original rationale, a significant proportion of the in vivo therapy is contributed by IFN-β–dependent immune-based mechanisms as shown by comparing VSV-mIFN-β with a closely matched VSV-hIFN-β virus in which the IFN-β is not biologically active in the murine host (Fig. 2). This immune component was most marked in the s.c. model as opposed to the i.p. model of tumor growth (Fig. 3A and B). We believe that this is probably because tumors growing in the s.c. and i.p. sites may differ significantly in the access and activation of the immune effectors that we believe are critical to these antitumor mechanisms, such as macrophages, NK (4) cells, and, as shown in Fig. 3, CD8+ T cells. However, further studies will be required to confirm this hypothesis.

We previously showed that inclusion of the IFN-β gene into a replication-defective adenoviral vector led to significant therapy following i.t. injections through priming of immune-mediated antitumor responses (19, 20, 24, 25). Therefore, we hypothesized that addition of IFN-β may lead to similar enhancements in the immunotherapeutic effects of oncolytic VSV. To monitor or track potential tumor antigen–specific responses, we used the B16ova model in which immunologic assays are available for both the model nonself OVA antigen and several endogenous self-melanoma-associated antigens. To justify the similarities between the mesothelioma and melanoma models, we have shown that VSV-mIFN-β provides effective therapy against B16ova, and that addition of IFN-β to VSV has similar effects as seen in the AB12 model (Fig. 4A). Significantly, however, our results indicate that addition of IFN-β to oncolytic VSV did not enhance priming of tumor antigen–specific T-cell responses (Fig. 4). Whereas VSV injection into tumors primed specific antiviral T-cell responses (Fig. 4), we were unable to detect antigen-specific responses against either OVA or the endogenous TRP-2 or gp100 tumor antigens. Mice injected i.t. with VSV developed generally elevated levels of T-cell activity, although these were not specific for any stimulating peptide other than virally derived epitopes (Fig. 4B–E and data not shown). Nonetheless, depletion studies showed that these apparently antigen nonspecific CD8+ T-cell responses are important for therapy. This may be due to the fact that the nonspecific immune activation induced by either virus or IFN-β induces a series of relatively infrequent, but polyclonal, antigen responses against tumor-associated antigens but that these responses are not strong enough to be detected without some sort of amplification. This is consistent with our previous reports that effective anti-OVA T-cell responses can be primed optimally in the B16ova model when the tumor antigen (OVA) is expressed from the VSV itself (4). Alternatively, it may be possible that the strong viral responses predominate and overshadow T-cell responses against antigens expressed within the tumor, and dependence of the therapy on CD8+ T cells may be associated more with the observed generalized T-cell hyperactivity induced by a potently immunogenic adjuvant rather than by antigen-specific T-cell effectors recognizing tumor-associated antigens. Favoring the first possibility is that BALB/c mice cured of established AB12 tumors following i.t. VSV-IFN-β therapy rejected a subsequent challenge with tumor. Therefore, it will be important to elucidate whether these rejection responses are associated with the concept of nonspecific concomitant tumor immunity (28)—compatible with the VSV-induced generalized T-cell reactivity—or with low-frequency, genuinely tumor antigen–specific T-cell responses.

Viral-directed expression of IFN-β would be expected to counteract antitumor efficacy, compared with VSV alone, by inhibiting viral replication if the tumor cells retain any levels of sensitivity to type-I IFN signaling. However, our data, as well as those of other studies,5

5

Galivo et al. submitted.

suggest that immune effectors are the predominant mechanism of antitumor therapy in the immunocompetent model. Therefore, the immune-activating benefits of addition of IFN-β to VSV may outweigh the negative therapeutic effect of any reduced viral replication. Further studies on the levels of viral replication in tumors with time will be required to confirm this.

Consistent with this, our data clearly show a predominantly immune-based component of antitumor therapy with VSV-IFN-β in the immunocompetent AB12 model (Fig. 2). However, in the human MSTO/SCID model (Fig. 5), therapy was principally associated with oncolyis because both VSV-hIFN-β and VSV-mIFN-β viruses gave equivalent antitumor therapy. These results are consistent with the hypothesis that the antitumor mechanisms of oncolytic virotherapy represent a balance between immune effectors and direct tumor cell destruction by viral replication and oncolysis. In the fully immunocompetent setting, we believe that immune effectors predominate in both viral clearance and antitumor efficacy. In contrast, in the context of SCID mice, the role of direct oncolysis is likely to be both more dominant and more apparent.

We also confirmed that inclusion of IFN-β into VSV as an additional safety feature protects normal cells or tissues from VSV-induced replication and cytotoxicity (Fig. 5). Thus, inclusion of the biologically inactive human IFN-β gene into VSV unmasked very significant neurotoxicity of i.t. injected VSV in the context of SCID mice, in which toxicity of VSV is more readily apparent, compared with the immunocompetent counterpart hosts. In contrast, replacing the human IFN-β gene with the murine IFN-β gene led to a more prolonged therapeutic window of opportunity without any apparent toxicity, during which i.t. injections of virus were able to achieve significant tumor cures. These data are consistent with our ongoing toxicology studies to support implementation of clinical trials of the VSV-IFN-β virus in patients. In this respect, we have observed that incorporation of a biologically active IFN-β gene into VSV significantly increases the maximum tolerated dose in excess of 2 orders of magnitude when administered in rodent models.

In summary, we have shown that VSV-IFN-β has significant therapeutic potential against mesothelioma tumors in the context of both local and locoregional viral delivery. Addition of IFN-β provides both added efficacy through immune-mediated mechanisms, which include activation of CD8+ T-cell responses, and added safety by providing protection from off-target viral replication in nontumor tissues. Our original hypothesis was that the use of a replication-competent viral vector expressing IFN-β would be therapeutically more effective than a replication-defective adenoviral vector against established mesothelioma tumors. This was based on the assumption that the replication-competent VSV-IFN-β virus would transduce more tumor cells and, correspondingly, express higher levels of IFN-β, leading to higher levels of direct tumor cell killing as well as improved IFN-β–mediated antitumor immunity. Our studies here suggest that this initial hypothesis may not necessarily prove to be correct. Thus, we observe only very limited replication of the VSV vector in vivo, which is almost certainly associated with the potent innate immunogenicity of this virus.5 Moreover, this immunogenicity also resulted in an effect in which adaptive T-cell responses to the virus seem to dominate over those to tumor-associated antigens. In the light of our current results, therefore, a direct head-to-head comparison between the two vector platforms (VSV-IFN-β and Ad-IFN-β) is warranted, and these studies are now under way in our laboratory.

No potential conflicts of interest were disclosed.

Note: C.L. Willmon and V. Saloura contributed equally to this work. S.M. Albelda and R.G. Vile are joint senior authors.

Grant support: Richard Schulze Family Foundation, the Mayo Foundation, and NIH grants CA107082-02, CA130878-01, and CA66726-12.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Toni Higgins for expert secretarial assistance.

1
Robinson BW, Lake RA. Advances in malignant mesothelioma.[see comment].
N Engl J Med
2005
;
353
:
1591
–603.
2
Kirn D, Martuza RL, Zwiebel J. Replication-selective virotherapy for cancer: biological principles, risk management and future directions.
Nat Med
2001
;
7
:
781
–7.
3
Parato KA, Senger D, Forsyth PA, Bell JC. Recent progress in the battle between oncolytic viruses and tumours.
Nat Rev Cancer
2005
;
5
:
965
–76.
4
Diaz RM, Galivo F, Kottke T, et al. Oncolytic immunovirotherapy for melanoma using vesicular stomatitis virus.
Cancer Res
2007
;
67
:
2840
–8.
5
Fernandez M, Porosnicu M, Markovic D, Barber GN. Genetically engineered vesicular stomatitis virus in gene therapy: application for treatment of malignant disease.
J Virol
2002
;
76
:
895
–904.
6
Barber GN. Vesicular stomatitis virus as an oncolytic vector.
Viral Immunol
2004
;
17
:
516
–27.
7
Barber GN. VSV-tumor selective replication and protein translation.
Oncogene
2005
;
24
:
7710
–9.
8
Lichty BD, Power AT, Stojdl DF, Bell JC. Vesicular stomatitis virus: re-inventing the bullet.
Trends Mol Med
2004
;
10
:
210
–6.
9
Balachandran S, Barber GN. Vesicular stomatitis virus (VSV) therapy of tumors.
IUBMB Life
2000
;
50
:
135
–8.
10
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
–5.
11
Stojdl DF, Lichty BD, tenOever BR, et al. VSV strains with defects in their ability to shutdown innate immunity are potent systemic anti-cancer agents.
Cancer Cell
2003
;
4
:
263
–75.
12
Balachandran S, Porosnicu M, Barber GN. Oncolytic activity of vesicular stomatitis virus is effective against tumors exhibiting aberrant p53, Ras, or myc function and involves the induction of apoptosis.
J Virol
2001
;
75
:
3474
–9.
13
Ebert O, Harbaran S, Shinozaki K, Woo SL. Systemic therapy of experimental breast cancer metastases by mutant vesicular stomatitis virus in immune-competent mice.
Cancer Gene Ther
2005
;
12
:
350
–8.
14
Lichty BD, Stojdl DF, Taylor RA, et al. Vesicular stomatitis virus: a potential therapeutic virus for the treatment of hematologic malignancy.
Hum Gene Ther
2004
;
15
:
821
–31.
15
Lun X, Senger DL, Alain T, et al. Effects of intravenously administered recombinant vesicular stomatitis virus (VSV(δM51)) on multifocal and invasive gliomas.
J Natl Cancer Inst
2006
;
98
:
1546
–57.
16
Shinozaki K, Ebert O, Woo SL. Eradication of advanced hepatocellular carcinoma in rats via repeated hepatic arterial infusions of recombinant VSV.
Hepatology
2005
;
41
:
196
–203.
17
Stojdl DF, Abraham N, Knowles S, et al. The murine double-stranded RNA-dependent protein kinase PKR is required for resistance to vesicular stomatitis virus.
J Virol
2000
;
74
:
9580
–5.
18
Sterman DH, Albelda SM. Advances in the diagnosis, evaluation, and management of malignant pleural mesothelioma.
Respirology
2005
;
10
:
266
–83.
19
Odaka M, Sterman DH, Wiewrodt R, et al. Eradication of intraperitoneal and distant tumor by adenovirus-mediated interferon-β gene therapy is attributable to induction of systemic immunity.
Cancer Res
2001
;
61
:
6201
–12.
20
Odaka M, Wiewrodt R, DeLong P, et al. Analysis of the immunologic response generated by Ad.IFN-β during successful intraperitoneal tumor gene therapy.
Mol Ther
2002
;
6
:
210
–8.
21
Obuchi M, Fernandez M, Barber GN. Development of recombinant vesicular stomatitis viruses that exploit defects in host defense to augment specific oncolytic activity.
J Virol
2003
;
77
:
8843
–56.
22
Ahmed CM, Wills KN, Sugarman BJ, et al. Selective expression of nonsecreted interferon by an adenoviral vector confers antiproliferative and antiviral properties and causes reduction of tumor growth in nude mice.
J Interferon Cytokine Res
2001
;
21
:
399
–408.
23
Brin E, Atencio I, Helmich BK, Maneval D, Laface D. Adenovirus delivery provides extended interferon-α exposure and augments treatment of metastatic carcinoma.
Cancer Gene Ther
2006
;
13
:
664
–75.
24
Sterman DH, Recio A, Carroll RG, et al. A phase I clinical trial of single-dose intrapleural IFN-β gene transfer for malignant pleural mesothelioma and metastatic pleural effusions: high rate of antitumor immune responses.[erratum appears in Clin Cancer Res. 2007 Sep 1;13:5226 Note: Kanther, Michelle [added]].
Clin Cancer Res
2007
;
13
:
4456
–66.
25
Kruklitis RJ, Singhal S, Delong P, et al. Immuno-gene therapy with interferon-β before surgical debulking delays recurrence and improves survival in a murine model of malignant mesothelioma.
J Thorac Cardiovasc Surg
2004
;
127
:
123
–30.
26
Davis MR, Manning LS, Whitaker D, Garlepp MJ, Robinson BW. Establishment of a murine model of malignant mesothelioma.
Int J Cancer
1992
;
52
:
881
–6.
27
Linardakis E, Bateman A, Phan V, et al. Enhancing the efficacy of a weak allogeneic melanoma vaccine by viral fusogenic membrane glycoprotein-mediated tumor cell-tumor cell fusion.
Cancer Res
2002
;
62
:
5495
–504.
28
Prehn RT. Two competing influences that may explain concomitant tumor resistance.
Cancer Res
1993
;
53
:
3266
–9.