Background
Breast cancer is a highly aggressive disease with most of the deaths resulting from metastases within the first three years upon diagnosis [
1]. Metastatic human triple-negative breast cancer (TNBC) has the worst prognosis among all types of breast cancer, with high risk of rapid recurrence and shortened survival [
2]. TNBC is deficient in the expression of estrogen receptor, progesterone receptor and human epidermal growth factor receptor 2, and thus, is refractory to conventional breast cancer hormonal therapy such as Tamoxifen. The main therapeutic option with surgery is chemotherapy, but some subsets of tumor are resistant and the prognosis for these patients is poor [
3]. The standard of care for TNBC is the administration of anthracyclines and/or taxanes [
2]. Paclitaxel (PAC), also called Taxol, is a cancer chemotherapeutic agent of the taxane family that acts by stabilizing microtubules and thus preventing cell division [
4]. PAC is commonly used as monotherapy or in combination with different agents. Significant effort is currently being directed toward improving its efficacy and developing alternate strategies for the treatment of chemotherapy-resistant and recurrent disease.
A novel strategy being explored for the treatment of metastatic diseases such as TNBC is the use of oncolytic viruses (OV). Several candidates are currently undergoing clinical trials and are considered promising approaches for the treatment of various cancers including TNBC [
5]. At the forefront of this field is T-Vec, a herpes virus that was successfully tested in a phase III study in melanoma and was approved in 2015 by the Food and Drug Administration for clinical use. OVs specifically replicate in and destroy tumor cells by several mechanisms including direct oncolysis [
6]. The rhabdovirus family members, vesicular stomatitis virus (VSV) and maraba, were first identified as oncolytic agents by our group [
7,
8]. The tumor specificity of these viruses is conferred by the capacity of normal cells, but not tumor cells, to respond to antiviral interferons (IFN) [
7,
8]. Variants with a greater therapeutic index, VSVΔ51 and Maraba MG1, were subsequently developed for clinical use [
8,
9]. Importantly, enrolling recently began for a clinical trial using MG1 both as a stand-alone therapy and in a vaccination strategy in patients with late-stage disseminated disease (NCT02285816).
A means to further improve the efficacy of the virus is to augment its replication in the tumor. In a previous study, we identified drugs, so-called virus sensitizers (VSe), that enhanced VSV replication in a tumor-specific manner [
10]. The compound identified as VSe12 in that study is PAC and it demonstrated the ability to substantially increase viral replication in vitro. Another VSe, colchicine, affects microtubule dynamics and was also the subject of a recent detailed study [
11]. As opposed to PAC, which stabilizes microtubules, colchicine has a destabilizing effect, which also results in the blockade of cell division [
12]. Colchicine-mediated enhancement of VSV was attributed in part to a defect in IFN secretion by infected cells, thus preventing the cytokine-conferred antiviral protection [
11].
The combination of PAC with OV treatment has been tested for vaccinia virus and herpes virus for other indications [
13,
14]. This study focuses on the efficacy of MG1 for breast cancer treatment and investigates the co-treatment with PAC. Here, using three different murine breast cancer models, we demonstrate that MG1 can be enhanced by PAC both in vitro and in vivo and that the co-treatment improves efficacy better than either treatment on its own without impairing the safety profile of the virus.
Methods
Cell lines and culture
Vero kidney epithelial, 4 T1, EMT6 and EO771 murine mammary carcinoma and Hs578T, BT-549 and MDA-MB-231 human mammary carcinoma cell lines (American Type Culture Collection (Manassas, VA, USA)) were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Corning cellgro, Manassas, VA, USA) supplemented with 10 % fetal bovine serum (FBS) (Sigma life science, St-Louis, MO, USA) and maintained at 37 °C with 5 % CO2.
Virus amplification and purification
MG1-green fluorescent protein (GFP) was purified as previously described [
8]. Briefly, Vero cells were infected for 24 h at a multiplicity of infection (MOI) of 0.01. Supernatants were then filtered using a 0.2-um bottle top filter (Millipore, MA, USA) prior to 1.5-h centrifugation at 30100 g. The pellet was resuspended in Dulbecco’s phosphate-buffered saline (DPBS) (Corning cellgro, Manassas, VA, USA) and aliquots were stored at −80 °C.
PAC treatment
PAC was purchased from Accord healthcare Inc. (Durham, NC). The cells were pre-treated at a concentration of 2 uM in culture media for 4 h prior to infection, unless specified otherwise. For in vivo experiments, animals were treated intraperitoneally (IP) with 2 mg/kg or 10 mg/kg of PAC as specified (see figure legends).
Virus titration
Titers were obtained by plaque assay. Briefly, serial dilutions of the samples were transferred to monolayers of Vero cells. Following an incubation of 1 h, cells were overlaid with 0.5 % agarose/DMEM supplemented with 10 % FBS. Plaques were counted 24 h later. For in vivo experiments, tumors and organs were collected 48 h post treatment, homogenized in PBS using a tissue homogenizer, then serially diluted and virus quantified as described above.
In vitro IFNβ treatment and quantification
Monolayers of tumor cells were treated with 250 U/mL of murine IFNβ (PBL interferon source, Piscataway, NJ, USA) 4 h prior to virus infection. The production of IFNβ by tumor cells was quantified using the ELISA mouse IFNβ kit (R&D systems, Minneapolis, MN, USA) following the manufacturer’s protocol. The samples were generated by pre-treating the cells with PAC as described above and infecting them for 24 h at an MOI of 0.1.
Coomassie Blue staining/viability assay
At 72 h post infection, cells were fixed for 30 minutes using fixative solution (3:1 methanol-acetic acid). The fixative was then replaced by the Coomassie Blue staining solution (3:1 methanol-acetic acid, 0.1 % Commassie Blue dye) for 30 minutes. The plates were washed and dried overnight prior to scanning. For quantification, the Coomassie Blue staining was solubilized using 10 % SDS, and serial dilutions were performed and transferred to a 96-well plate for reading using a Fluoroscan plate reader at 450 nm.
Microscopy
For nuclear staining, cells were cultured and treated on coverslips for 72 h. Cells were then washed with cold PBS and fixed using ice-cold methanol-acetone (1:1). Nuclei were stained using 4',6-diamidino-2-phenylindole (DAPI) included in the Prolong gold anti-fade (Molecular Probes) used to mount the coverslips onto slides. Live images of MG1-GFP infected cells were acquired using an EVOS Fl cell imaging system (ThermoFisher Scientific) microscope 24 h post infection.
Flow cytometry
For quantification of virus infection, cells were processed as previously described [
15]. Briefly, cells were harvested and fixed using IC fixation buffer (eBioscience) 24 h after PAC treatment and infection with MG1-GFP at an MOI of 0.01. Cells were then washed twice and resuspended in FACS buffer (3 % FBS, PBS) for analysis using a Cyan ADP 9 flow cytometer (Beckman Coulter, Mississauga, ON, Canada).
In vivo experiments and tumor models
Balb/c mice were used (Charles River Laboratories) for the 4 T1 and EMT6 murine tumor models. For orthotopic implantation of the tumors, 2 × 105 cells were injected into the second left mammary fat pad. For the EO771 tumor model, 1 × 106 cells were implanted into the second left mammary fat pad of C57/Bl6 mice. For treatments, the virus and drug preparations were diluted to the appropriate concentration in a total volume of 100 uL of PBS and injected IP or intratumorally (IT) using insulin syringes (The Stevens Co, Montreal, QC, Canada). All experiments were performed in accordance with the University of Ottawa animal care and veterinary services guidelines.
Histological analysis
Tumors were collected 48 h after treatment and fixed in 10 % buffered formalin phosphate (Fisher Scientific, Waltham, MA, USA) for 48 h. Paraffin-embedded sections were stained using hematoxylin and eosin or the specified antibodies. For antibody staining, the sections were rehydrated through graded alcohol and heat-mediated antigen retrieval was performed in citrate buffer (sodium citrate 10 mM, pH 6). Tissue sections were stained as described previously [
16] using a rabbit anti-VSV (made in house) and rabbit anti-caspase-3 (Cell signalling technology) antibodies.
TNBC ex-vivo samples
Patient-derived TNBC xenografts were grown into NOD/SCID mice as described previously [
17,
18]. When the tumors reached 1500 mm
3 in size they were collected and cores were generated as described previously [
19]. The cores were treated ex-vivo with MG1 (10
3 plaque-forming units (pfu)) and PAC and culture supernatant was collected 48 h later to titer the virus output.
Tumor measurements and survival experiments
The length and width of the tumors were measured using digital calipers (Fowler). The formula (length × width2)/2 was used to calculate tumor volumes. The mice were sacrificed when they displayed respiratory distress, significant weight loss, ulceration, or discomfort, or when the tumor volume reached 1500 mm3 in size.
Statistical analysis
Statistical analyses were performed using GraphPad Prism 6.0 software (see figure legends). Error bars represent standard error of the mean.
Discussion
In this study, we demonstrated the compatibility of PAC, a standard-of-care chemotherapeutic agent for breast cancer, and MG1, an OV that is considered a promising and novel strategy for treating disseminated diseases like breast cancer. Our results not only show that treatments do not interfere with one another, but they can also perform even better when co-administered. Using three different syngeneic murine breast cancer models, we show a prolongation of survival for animals that received both treatments compared to either treatment alone (Fig.
6). These findings have potential implications for the future treatment of patients. Our data support clinical testing of the combination. Even if the patient’s cancer has become resistant to the drug, it might still effectively enhance MG1 and at the very least, should not impair the viral treatment. Second, because the beneficial effects we observed were achieved using sub-lethal concentrations of PAC and knowing the various side effects of the drug in patients with cancer, it is tempting to suggest that using a lower concentration of PAC in combination with the virus would be a suitable strategy.
Interestingly, we found that two out of the three murine tumor cell lines and the human tumor cell lines that we tested were sensitized to viral infection by PAC (Fig.
2). Indeed, while EMT6 and 4 T1 cells produced more virus when pre-treated with the drug, the E0771 cell line was not affected. Our model is that PAC pre-treatment would block the secretion of antiviral factors like IFNβ by infected cells (Additional file
2: Figure S2B), thereby increasing virus infection. In line with this idea, both EMT6 and 4 T1 cells in which we see sensitization to the virus, but not the E0771 cells, which are refractory to this effect, demonstrated impaired production of the antiviral cytokine (Additional file
2: Figure S2B). Importantly, our results also show that while ex-vivo infection and treatment of patient breast cancer xenografts can inform us on the potential enhancement of the virus by the drug in specific patient samples (Fig.
4e), the lack of enhancement would not necessarily imply that both treatments would not be compatible. Indeed, increased production of virus is desirable, but the killing of the target tumor cells is what ultimately matters. Our data show that although the E0771 cell line does not demonstrate sensitization to the virus in the presence of PAC, complete killing of the cells was still observed at concentrations of PAC and MG1 that did not affect cell viability alone. Notably, the E0771 tumor model was the one in which we observed the highest percentage of cures with the treatment combination (Fig.
6c). This is most likely due to the greater sensitivity of the E0771 tumors to both single treatments alone compared to the two other tumor models (Fig.
6). Remarkably, while the EMT6 and 4 T1 tumors are slightly smaller compared to the control animals when treated with either the virus or the drug, the combination of both treatments was the only condition that was significantly different in terms of the ability to control tumor growth and prolong survival.
An interesting idea to explain the improved killing observed in vitro in the E0771 cell line would be that virus-induced factors may promote PAC killing. Indeed, it has been reported by Thorne and colleagues that oncolytic vaccinia virus induces the secretion of factors, including type I IFNs, which sensitize tumor cells to taxol [
13]. Also, a similar mechanism was observed for the colchicine and VSV co-treatment [
11]. This scenario would provide a mechanism by which even tumors that are refractory to the PAC-mediated sensitization to MG1 could still benefit from the combination. Of note, infection with vaccinia virus produces various anti-inflammatory factors [
20], including B18R which inhibits the activity of IFNs and could potentially minimize this effect, while MG1 and VSV do not encode these inhibitors, suggesting that the virus-mediated sensitization of tumor cells to PAC-mediated killing could be even greater [
7,
8].
Another interesting application of our combination strategy would be to potentiate the production of virally-encoded transgenes. The engineering of oncolytic viruses has been shown to be a successful strategy for tumor-targeted gene delivery and improvement of treatment efficacy. Indeed, rhabdoviruses encoding antiviral suppressors to increase viral replication [
21], suicide genes to improve killing [
22] or immune-stimulating cytokines to induce a greater anti-tumor immune response [
23] have all been shown to control tumor growth more efficiently compared to the parental virus. Given that, along with the increased virus production, we also observed more GFP-positive cells and higher MFV (Fig
2b and c) and thus, greater transgene production, the co-treatment could potentially be even more beneficial using viruses that encode transgenes that mediate greater control of the tumors.
Conclusions
With the urgent need for novel strategies for breast cancer treatment, especially in patients with TNBC, which are refractory to the limited available treatment options, our work provides a rational alternative to improve outcomes. Our data demonstrate that the combination of PAC and MG1 is effective at controlling tumor progression. Because PAC is a standard of care for breast cancer treatment and MG1 is undergoing clinical testing, we believe the findings included in this study are of great importance and that the translation of this work to the clinic could be rapid.
Abbreviations
DAPI, 4',6-diamidino-2-phenylindole; DMEM, Dulbecco’s modified Eagle’s medium; ELISA, enzyme-linked immunosorbent assay; FBS, fetal bovine serum; GFP, green fluorescent protein; IFN, interferon; IP, intrperitoneally; IT, intratumorally; MFV, mean fluorescence value; MOI, multiplicity of infection; OV, oncolytic virus; PAC, paclitaxel; PBS, phosphate-buffered saline; pfu, plaque-forming units;TNBC, triple-negative breast cancer; VSe, virus sensitizers; VSV, vesicular stomatitis virus
Acknowledgements
We would like to thank Dr. Alana Welm for kindly providing the patient-derived TNBC xenograft samples.