Combination of Paclitaxel and MG1 Oncolytic Virus as a Successful Strategy for Breast Cancer Treatment

Marie-Claude Bourgeois-Daigneault; Lauren Elizabeth St-Germain; Dominic Guy Roy; Adrian Pelin; Amelia Sadie Aitken; Rozanne Arulanandam; Theresa Falls; Vanessa Garcia; Jean-Simon Diallo; John Cameron Bell


Breast Cancer Res. 2016;18(83) 

In This Article


PAC Treatment Enhances MG1 Replication and Killing in Different Breast Cancer Cell Lines

In order to determine which concentration of drug and MOI of virus to use in our experiment, we first assessed the sensitivity of our different murine breast cancer cell lines to either treatment. Using a GFP-expressing version of MG1, we infected monolayers of cells at various MOIs and assessed transgene expression as a readout of infectivity. We also evaluated virus-mediated killing of the cells by looking at visible cytotoxicity. Our results demonstrate various sensitivities to the virus with the E0771 cells being completely infected at an MOI of 0.01 and all dead at an MOI of 10, and the EMT6 cells being the most resistant with only a few cells expressing GFP at an MOI of 10 (Fig. 1). The 4 T1 cells displayed an intermediate phenotype. We then investigated the sensitivity of the cell lines to PAC-induced killing. To do so, we incubated monolayers of cells with increasing concentrations of the drug for 48 h and performed a Coomassie Blue viability assay where decreased staining intensity reveals cytotoxicity. Similar to what we observed with the virus, the EO771 cells were the most sensitive to PAC and the EMT6 the most resistant cell line (Additional file 1: Figure S1a). The concentrations required to kill the various tumor cells ranged from 4 uM to 12 uM. To investigate the responsiveness of the cell lines to PAC-mediated inhibition of cellular division, we stained the nucleus of treated cells in order to detect enlarged, polynuclear cells. Interestingly, all three tumor cell lines were responsive to low, sub-lethal concentrations of PAC (0.5 uM) (Additional file 1: Figure S1b).

Figure 1.

Murine breast cancer cell lines display different sensitivities to MG1. Microscopy images of EMT6, 4 T1 and EO771 tumor cells infected with various multiplicities of infection of MG1-green fluorescent protein for 24 h

To evaluate the effect of PAC on viral replication, we assessed the presence of GFP-positive cells following co-treatment with the drug in our murine breast cancer cell lines. We pre-treated the cells for 4 h with 2 uM of PAC, a concentration at which all cell lines displayed polynucleation, but none of them exhibited drug-mediated killing (Additional file 1: Figure S1). Our results indicate the presence of more MG1-infected EMT6 and 4 T1 cells for the co-treatment conditions (Fig. 2a, left panels). This enhancement was confirmed by quantification of the virus in the supernatants where 10-fold to 100-fold more virus was detected in the presence of the drug (Fig. 2a, right panels). For the E0771 cell line, no difference was observed in virus recovery and GFP expression, demonstrating that the drug did not enhance, but did not impair virus production either. We obtained similar results using three different human TNBC cell lines, whereby virus enhancement was observed in two cell lines (MDA-MB-231 and BT-549) but not in the third (Hs578T) (Fig. 2b). We used flow cytometry as a means to quantify the percentage of GFP-expressing cells and the mean fluorescence value (MFV) of infected cells in the presence or absence of the drug. Our results demonstrate that there were more GFP-positive cells when we pre-treated the 4 T1 and EMT6 cells with PAC, with percentages and average fluorescence twofold to threefold higher in the presence of the drug (Fig. 2c and d).

Figure 2.

Paclitaxel (PAC) enhances MG1 in various human and mouse breast tumor cell lines. a Microscopy images of EMT6, 4 T1 and EO771 tumor cells infected with MG1-green fluorescent protein (GFP) after a 4-h pre-treatment with 2 uM PAC. Graphs right represent virus titers obtained 24 h post infection. ND no drug, pfu plaque-forming units. b Microscopy images of MDA-MB-231, BT-549 and Hs578T human tumor cells infected with MG1-GFP after a 4-h pre-treatment with 2 uM PAC. Flow cytometry histograms show the GFP expression of infected EMT6 (c) or (d) 4 T1 cells 24 h post infection with or without PAC pre-treatment. The right graphs show percentage of GFP+ cells and the mean fluorescence values (MFV). Samples were analyzed in triplicates. Statistical significance was tested using the unpaired two-tailed t test with Welch's correction; *p < 0.05, **p < 0.01 ***p < 0.001. Ctrl control

In a recent study we demonstrated that the viral sensitization mediated by colchicine, another drug affecting microtubules and preventing cell division was mediated by a blockade in the secretion of antiviral IFNs.[11] As many tumor cell lines are refractory to antiviral IFNs and would thus be refractory to enhancement involving this mechanism of action,[7] we first assessed the sensitivity of our cell lines to the cytokine. Additional file 2: Figure S2a shows that pre-treating the cells with IFNβ efficiently protected all three cell lines against the virus. Consistent with this, less killing of the cells was observed with IFNβ pre-treatment (Additional file 3: Figure S3). To measure the IFNβ production in response to virus infection, we performed an ELISA on culture supernatants of infected cells. For all three cell lines, the cytokine was detected following infection. Interestingly, and consistent with our virus enhancement data (Fig. 2), our results show that in both EMT6 and 4 T1 cells, the production of IFNβ was impaired in the presence of PAC, while the levels produced by the E0771 cells were unaffected by the drug (Additional file 2: Figure S2b).

As the aim of both MG1 and PAC treatments is ultimately to kill tumor cells, we assessed cell death following co-treatment. We used a concentration of the drug where no cytotoxicity was observed following a 48-h incubation. For all three murine cell lines, we observed more cytotoxicity in the presence of both treatments with almost all the cells being dead, suggesting synergistic rather than cumulative killing (Fig. 3a). This decrease in viability was confirmed by quantification of the staining (Additional file 4: Figure S4). This synergistic killing was also confirmed using the MDA-MB-231, BT-549 and Hs578T human cell lines (Fig. 3b).

Figure 3.

Paclitaxel (PAC) and MG1 synergistically kill breast cancer cell lines. Coomassie Blue staining of EMT6, 4 T1 and EO771 (a) and MDA-MB-231, BT-549 and Hs578T cells (b) infected or not with MG1-green fluorescent protein and co-treated with 2 uM PAC for 48 h. ND no drug

The Virus is Enhanced by PAC Treatment in Vivo

We then sought to confirm our in vitro findings in tumor-bearing animals. We implanted tumors orthotopically in order to recapitulate the natural microenvironment as much as possible. The virus was administered IT and quantified 48 h later by plaque assay. Our results demonstrate that threefold to fourfold more virus was detected in the EMT6 and 4 T1 tumors of the animals that also received PAC treatment compared to those that were treated with MG1 alone (Fig. 4a and b). For the E0771 tumor-bearing animals, we did not observe any difference in the amount of virus we recovered, consistent with the in vitro findings using this cell line (Figs. 2a and 4c). In order to assess if the increased viral replication would also result in increased replication in normal organs and impair the safety of the treatment, we performed a biodistribution experiment of MG1 48 h post injection in EMT6-tumor-bearing animals using the two drug concentrations used in this study. Our data demonstrate that, while both drug concentrations were able to increase viral replication in the tumors, no differences were observed in normal organs following co-treatment (Fig 4d).

Figure 4.

Paclitaxel (PAC) enhances MG1 replication in tumors. EMT6 (a), 4 T1 (b) or EO771 (c) tumor-bearing mice were treated intratumorally with 1 × 108 plaque-forming units (pfu) of MG1-green fluorescent protein (GFP) and intraperitoneally with PAC (10 mg/kg for the EMT6 model and 2 mg/kg for the 4 T1 and EO771 models). Tumors were harvested 48 h later and the viral quantification was obtained by plaque assay. Three or more tumors per condition were analysed. d EMT6 tumor-bearing mice were treated as in a and various organs were collected 48 h post treatment. The virus was quantified by plaque assay. e Tumor cores from human triple-negative breast cancer xenografts were infected ex-vivo with MG1-GFP with or without PAC co-treatment. The viral outputs were quantified by plaque assay. Statistical significance was calculated using the unpaired two-tailed t test with Welch's correction; *p < 0.05, **p < 0.01, ***p < 0.001. ND no drug

To confirm the positive effect of PAC-treatment on MG1 infection of human tumors, we used a breast cancer patient xenograft model that was previously described to recapitulate the human disease.[17,18] We infected tumor cores ex vivo in the presence or absence of PAC. Our results demonstrate that PAC efficiently enhanced viral replication in a TNBC patient-derived xenograft (Fig. 4e).

PAC-MG1 Combination Therapy Demonstrates Greater Tumor Killing in Vivo

Previous work by Lin et al. demonstrates that while both PAC and an oncolytic herpes simplex virus induce apoptosis, the combination of both was more effective in human anaplastic thyroid cancer cell lines.[14] To investigate if this was also the case using MG1 in our tumor models, we performed immunohistochemical analysis against the cleaved pro-apoptotic molecule caspase-3 on EMT6 tumors from mice that received the various treatments. First, the hematoxylin and eosin staining clearly demonstrated the presence of widespread necrotic regions for the MG1 as well as the MG1 and PAC co-treated tumors (Fig. 5, left panels). This was confirmed by caspase-3 staining, which was extensive in these tumors. (Fig. 5, right panels). These regions were larger and more abundant in the co-treated animals compared to those that received the MG1 treatment only. Also, consistent with the virus quantification shown in Fig. 4a, staining of the tumor sections with a virus-specific antibody demonstrated increased virus spread in the presence of PAC (Fig. 5, middle panels).

Figure 5.

The co-treatment with paclitaxel (PAC) and MG1 increases tumor apoptosis. EMT6 tumor-bearing animals were treated intratumorally with MG1-green fluorescent protein (108 plaque-forming units) and intraperitoneally with 10 mg/kg with PAC, and tumor samples were harvested 48 h post treatment. Sections were stained for the presence of virus and apoptotic cells (caspase-3)

The Treatment With MG1 and PAC Demonstrates Greater Efficiency in Murine Tumor Models

Given the improved killing observed with PAC in all three cell lines in vitro, the increased virus recovery from tumors and the greater caspase-3-positive tumor regions observed in the presence of the co-treatment, we sought to determine if these phenomena would translate into slower tumor growth and an improvement in the survival of tumor-bearing animals. The mice were treated with MG1, PAC or both and tumors were measured over time. As expected, we observed that the growth was slower following treatment with the combination of the virus and the drug in all three tumor models (Fig. 6a, b and c, upper panels). Also, as observed in vitro, the E0771 model was the most sensitive to both treatments. Importantly, the enhanced control of tumor growth translated into a significant prolongation of survival in all three models and some animals were even cured for the 4 T1 and E0771 tumor models (Fig. 6a, b and c, bottom panels).

Figure 6.

The combination of Paclitaxel (PAC) and MG1 improves efficacy in various murine breast cancer models. Volumes of EMT6 (a), 4 T1 (b) or EO771 (c) fat-pad tumors in mice treated or not (NT) with MG1-green fluorescent protein (1 × 108 pfu) in combination or not with PAC (10 mg/kg for EMT6 and 2 mg/kg for 4 T1 and EO771). Mice were sacrificed when they reached the endpoint. NS p > 0.05, *p < 0.1, **p < 0.01, ***p < 0.001 (unpaired multiple two-tailed t test). Kaplan-Meier survival curves were generated from the same experiments (bottom panels). For survival experiments, NS p > 0.1, *p < 0.1, **p < 0.01, ***p < 0.001 (Mantel-Cox test)