Breaking Resistance of Pancreatic Cancer Cells to an Attenuated Vesicular Stomatitis Virus Through a Novel Activity of IKK inhibitor TPCA-1

Abstract

Vesicular stomatitis virus (VSV) is an effective oncolytic virus against most human pancreatic ductal adenocarcinoma (PDAC) cell lines. However, some PDAC cell lines are highly resistant to oncolytic VSV-ΔM51 infection. To better understand the mechanism of resistance, we tested a panel of 16 small molecule inhibitors of different cellular signaling pathways, and identified TPCA-1 (IKK-β inhibitor) and ruxolitinib (JAK1/2 inhibitor), as strong enhancers of VSV-ΔM51 replication and virus-mediated oncolysis in all VSV-resistant PDAC cell lines. Both TPCA-1 and ruxolitinib similarly inhibited STAT1 and STAT2 phosphorylation and decreased expression of antiviral genes MxA and OAS. Moreover, an in situ kinase assay provided biochemical evidence that TPCA-1 directly inhibits JAK1 kinase activity. Together, our data demonstrate that TPCA-1 is a unique dual inhibitor of IKK-β and JAK1 kinase, and provide a new evidence that upregulated type I interferon signaling plays a major role in resistance of pancreatic cancer cells to oncolytic viruses.

INTRODUCTION

The use of oncolytic viruses (OVs) as an anticancer strategy arises from their ability to infect, replicate in and kill cancer cells. Compared to non-malignant cells, cancer cells are generally more susceptible to viral infection due to their defects in type I interferon (IFN)-mediated antiviral responses [reviewed in (Barber, 2005; Hastie et al., 2013; Lichty BD et al., 2004)]. Vesicular stomatitis virus (VSV, a rhabdovirus) is a promising OV successfully used in preclinical models for the treatment of a variety of cancers, and currently in a phase I clinical trial for treatment of hepatocellular carcinoma (clinical trial NCT01628640). Pancreatic ductal adenocarcinoma (PDAC) comprises about 95% of pancreatic cancers and is highly invasive with aggressive local growth and rapid metastases to surrounding tissues. Standard cancer therapies show little efficacy in treating PDAC (Stathis A and Moore, 2010). Our recent studies demonstrated that VSV recombinants are effective against a majority of clinically relevant human PDAC cells lines tested (Murphy et al., 2012). However, out of 11 human PDAC cell lines, 4 were resistant to VSV infection, replication and virus-mediated oncolysis (Murphy et al., 2012). In all VSV-resistant cell lines several interferon stimulated genes (ISGs), including the antiviral genes MxA and OAS, were constitutively expressed at high-level, and inhibition of type I FN signaling pathway using JAK Inhibitor I (JAK Inh. I, a pan-JAK inhibitor) reduced ISG expression and decreased their resistance to VSV (Moerdyk-Schauwecker et al., 2013).

In the present study, to better understand the mechanism of the resistance and find new approaches to overcome it, we tested a panel of 16 inhibitors of different cellular signaling pathways previously shown to affect replication of VSV and other viruses. Our experiments identified one inhibitor of IkB kinase β (IKK-β), TPCA-1, and one selective JAK1/2 inhibitor, ruxolitinib (trade name Jakafi) that decreased levels of ISGs and increased VSV replication and VSV-mediated oncolysis more efficiently than JAK Inhibitor I. Further studies provided evidence that IKK-β inhibitor TPCA-1 also functions as a direct inhibitor of JAK1 kinase. Together, our data show that TPCA-1 is a unique dual inhibitor of IKK-β and JAK1 kinase, and provide a new evidence that the upregulated type I interferon signaling plays a major role in resistance of pancreatic cancer cells to oncolytic viruses.

MATERIALS AND METHODS

Cell lines, viruses and inhibitors

The human PDAC cell lines used in this study were: CFPAC-1 (ATCC CRL-1918), HPAC (ATCC CRL-2119), HPAF-II (ATCC CRL-1997), Hs766T (ATCC HTB-134), Mia PaCa-2 (ATCC CRL-1420), and AsPC-1 (ATCC CRL-1682). Baby hamster kidney BHK-21 fibroblast (ATCC CCL-10) and African green monkey kidney Vero cells (ATCC CCL-81) were used to grow viruses and determine virus titers. CFPAC-1, HPAC, Hs766T, Mia PaCa-2, and Vero cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Cellgro); AsPC-1 in RPMI 1640 (HyClone); HPAF-II, and BHK-21 in modified Eagle’s medium (MEM, Cellgro). All cell culture media were supplemented with 10% fetal bovine serum (FBS, Gibco), 4 mM L-glutamine (Cellgro), 100 IU/ml penicillin-100 µg/ml streptomycin (Cellgro) and 1x MEM Nonessential Amino Acids (Cellgro). MEM was also supplemented with 3.5% glucose. Cells were kept in a 5% CO₂ atmosphere at 37°C. For all experiments, PDAC cell lines were passaged no more than 10 times.

The recombinant VSV-ΔM51-GFP, which has been described previously (Wollmann G et al., 2010), has a deletion of the methionine at amino acid position 51 of the matrix protein, and the green fluorescent protein (GFP) ORF inserted at position 5 of the viral genome (between the VSV G and L genes). The recombinant Sendai virus (SeV) SeV-GFP-F_mut, which has been described previously (Wiegand et al., 2007), has the GFP ORF at position 1 of the viral genome and a mutation in the cleavage site of the fusion (F) protein, allowing F activation and production of infectious virus particles in cells without acetylated trypsin added to the medium. VSV-ΔM51-GFP was grown on BHK-21 cells and SeV-GFP-F_mut was grown in Vero cells. Viral titers for both viruses were determined by standard plaque assay on BHK-21 and expressed as plaque-forming units (PFU) per ml.

JAK Inh. I, IKK-2 Inhibitor VIII, IKK Inhibitor XIII, U-0126, rapamycin, and valproic acid were purchased from EMD Millipore. TPCA-1 was purchased from Tocris Bioscience and Sigma-Aldrich. SC-514, IKK-16 and IMD-0354 were purchased from Tocris Bioscience. Bortezomib, SAHA (vorinostat), celecoxib and ruxolitinib were purchased from Selleck Chemicals. Sulfasalazine and BMS-345541 were purchased from Sigma-Aldrich.

Effect of inhibitors on virus replication and cell viability

Cells were seeded in media with 10% FBS in 96-well plate so that they reached approximately 80% confluence after 24 hours (h). Cells were treated with no drug (here and elsewhere treatment with no drug contained 0.3% DMSO) or with inhibitor (here and elsewhere inhibitor treatment also contained 0.3% DMSO) in culture media with 5% FBS for 48 h prior to infection, or as specified. Media was removed and replaced with fresh inhibitor containing media every 24 h if treatment prior to infection lasted more than 24 h. Cells were then mock infected or infected with VSV-ΔM51-GFP or SeV-GFP-F_mut in DMEM without FBS at the specified multiplicity of infection (MOI), calculated based on virus titration on each cell line. Virus-containing media was aspirated after 1 h absorption period, and replaced with growth media containing 5% FBS and same inhibitor treatment as prior to infection, or as specified. After infection, virus-driven GFP fluorescence was measured at regular intervals (CytoFluor Series 4000, excitation filter of 485/20 nm, emission 530/25 nm, gain=63; Applied Biosystems). Cell viability was analyzed 5 days post infection (p.i.) by a 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) cell viability assay (Sigma-Aldrich). For flow cytometry analysis (Beckman Coulter), cells were seeded in 6-well plate, treated with no drug or inhibitors and infected as described above. Two days p.i. cells were trypsinized, washed with ice-cold phosphate-buffered saline (PBS), and fixed with 4% paraformaldehyde in PBS on ice.

Treatment with multiple inhibitors and Chou-Talalay analysis

HPAF-II and Hs766T cells were seeded as described above, and treated with no drug or two-fold serial dilutions of TPCA-1, ruxolitinib or both inhibitors in combination. Concentrations ranged from 2 to 16 µM for TPCA-1 and 1 to 8 µM for ruxolitinib. After 24 h pre-treatment, cells were mock infected or infected with VSV-ΔM51-GFP in DMEM without FBS at an MOI of 0.001. Virus-containing media was aspirated after 1 h, and replaced with growth media containing 5% FBS and same treatment as prior to infection. After infection, VSV-driven GFP fluorescence was measured every 24 h for 4 days. Dose-effect curves for each inhibitor and Combination Indices (CIs) for each combination of inhibitors were calculated by the Chou and Talalay CI method (Chou, 2006; Chou and Talalay, 1984) using CompuSyn software, version 1.0 (ComboSyn Inc., Paramus, New Jersey) and a non-constant ratio design. The fraction affected was calculated by subtracting background fluorescence (cells treated with no drug), and normalizing each value to the maximum GFP fluorescence value reached for each cell line over the 4 days treatment. For TPCA-1 and JAK Inh. I combination treatment, HPAF-II, HPAC and Hs766T cells were either treated with no drug or with 8 µM TPCA-1, 2.5 µM JAK Inh. I, 4 µM BMS-345541 or TPCA-1 and JAK Inh. I combined, for 2 days before infection with VSV-ΔM51-GFP at an MOI of 15 (based on titration on BHK-21 cells; the Figure 5 legend indicates cell specific MOIs). Cells were harvested 2 days p.i. and cell lysates were prepared and analyzed by western blot as described below.

Fig. 5. Effect of TPCA-1 and JAK inhibitors combination treatment on PDAC cells.

HPAF-II and Hs766T cells were treated with TPCA-1, ruxolitinib, or TPCA-1 and ruxolitinib combined. Treatment was started 1 day before infection with VSV-ΔM51-GFP (cell specific MOI 0.001), and maintained for 4 days p.i.. A) GFP fluorescence was measured and normalized to cells treated with no drug at each time point p.i. Assays were done in triplicate and data represent the mean ± SD of mean. (*) indicates statistical significance (p < 0.05) between treatment and no treated cells (no drug) at 48 and 72 h p.i. B) Combination Indexes (CI) calculated using the method of Chou-Talalay using VSV-driven GFP values at 48h p.i. Range of CI is as described by Chou and Talalay (Chou, 2006). C) HPAF-II, HPAC and Hs766T cells were treated with TPCA-1 (8 µM), JAK Inh. I (2.5 µM), BMS-345541 (BMS) (4 µM), or TPCA-1 and JAK Inh. I combined for 2 days before infection with VSV-ΔM51-GFP at MOI 15 (based on BHK-21 cells). Cell specific MOIs are MOI 0.01 based on HPAF-II, MOI 0.05 based on HPAC, and MOI 0.03 based on Hs766T. Cells lysates were prepared 2 days p.i, and analyzed by western blot for the indicated proteins.

Protein isolation and Western blot analysis

Cells were seeded in a 6-well as described above and treated with no drug or with the specified inhibitor until they were harvested. Where indicated, after 2 h inhibitor treatment, cells were treated with 25 ng/ml of a recombinant human Tumor Necrosis Factor Alpha (TNF-α R&D systems) or 5000 U/ml IFN alpha (IFN-α EMD Millipore) for 4 h. For time-course, cells were first infected with VSV-ΔM51-GFP at MOI of 0.01, and then treated with no drug or with inhibitor until harvested. Media was removed and cells were lysed in lysis buffer containing 0.0625 M Tris-HCl (pH 6.8), 10% glycerol, 2% SDS, 5% 2-mercaptoethanol, and 0.02% (w/v) bromophenol blue. Total protein was separated by electrophoresis on SDS-PAGE gels and electroblotted to polyvinylidene difluoride (PVDF) membranes. Membranes were blocked using 5% non-fat powdered milk in TBS-T [0.5 M NaCl, 20 mM Tris (pH 7.5), 0.1% Tween20]. Membranes were incubated with 1:5000 rabbit polyclonal anti-VSV antibodies (raised against VSV virions), 1:1000 rabbit anti-MxA (Sigma-Aldrich, SAB1100070), 1:200 rabbit anti-OAS (Santa Cruz, sc-99097), 1:1000 rabbit anti-PARP1 (Santa Cruz, sc-25780), 1:500 rabbit anti-p-STAT2 (R&D Systems, MAB2890) and the following antibodies from Cell Signaling Technology (1:1000 or 1:500): STAT1 (9172), p-STAT1 (7649), STAT2 (4594), STAT3 (9139), p-STAT3 (9134), IkBα (4814), p-IkBα (2859), and Caspase 3 (9662) in TBS-T with 5% BSA or milk with 0.02% sodium azide. The 1:2000 goat anti-mouse or 1:2000 goat anti-rabbit horseradish peroxidase-conjugated secondary antibodies (Jackson-ImmunoResearch) were used. The Amersham ECL Western Blotting Detection Kit (GE Healthcare) or Pierce SuperSignal WestPico Detection Kit (Thermo Scientific) was used for detection. To verify total protein in each loaded sample, membranes were re-probed with rabbit anti-GAPDH antibody (Santa Cruz, sc-25778) or stained with Coomassie blue R-250.

RNA isolation and analysis

HPAF-II cells were seeded in a 6-well plate as described above and treated with no drug or with the specified inhibitor for 2 h in serum free-media (SFM). Cells were then treated with TNF-α (25 ng/ml) or IFN-α (5000 U/ml) for 4 h, while inhibitor treatment was maintained. IFN-α and TNF-α induction was performed in SFM to exclude nonspecific NF-κB activation by serum components. Total RNA was extracted using the Quick-RNA Mini Prep kit in accordance with manufacturer instructions (Zymo Research), and reverse transcribed using SMART-Scribe reverse transcriptase (Clontech Laboratories, Inc.) and random hexamer as per manufacturer’s protocol. PCR products were electrophoresed on a 2% agarose gel with ethidium bromide and photographed using a GelDoc-It imager (UVP Imaging). Real-time PCR were run in triplicate using Absolute Blue SYBR Green Rox Mix (Thermo Scientific) in an Applied Biosystems 7500 sequence detection system. Relative gene expression was normalized to GAPDH expression and fold change expression was calculated by the comparative Ct method. The following primers were used for PCR and/or real-time PCR: β-actin: 5’-gcaaagacctgtacgccaaca-3’ (forward), 5’-cctcggccacattgtgaac-3’ (reverse); TNF-α 5’-cccagggacctctctctaatca (forward), 5’-gcttgagggtttgctacaacatg-3’ (reverse); MxA: 5’-gctacacaccgtgacggatatgg-3’ (forward), 5’-cgagctggattggaaagccc-3’ (reverse); OAS2: 5’-tcagaagagaagccaacgtga-3’ (forward), 5’-cggagacagcgagggtaaat-3’ (reverse); GAPDH: 5’-ccatcaccatcttccaggagcgag-3’ (forward), 5’-cacagtcttctgggtggcagtgat-3’ (reverse). IFN-β: 5’-ggcaattgaatgggaggct-3’ (forward), 5’-ggcgtcctccttctggaact-3’ (reverse).

Nuclei isolation and EMSA analysis

HPAF-II cells were seeded in a 6-well plate as described above, and treated with no drug or with 8 µM TPCA, 2.5 µM ruxolitinib, or 2.5 µM JAK Inh. I for 2 h prior to induction with IFN-α (5000 U/ml) for the indicated time. Nuclear protein extracts were isolated as previously described (Holden and Tacon, 2011), and the protein concentration determined by Bradford assay. A double-stranded oligonucleotide corresponding to the consensus ISRE for STAT1/2 binding (5’-ggcttcagtttcggtttccctttcccgagg-3’) was end-labeled with [γ-³²P]ATP using T4 kinase (Promega). Nuclear extracts containing 5 µg of nuclear protein were incubated with radiolabeled ISRE probe and 1 µg poly(dI-dC) in binding buffer (20 mM HEPES pH 7.9, 1 mM DTT, 0.1 mM EDTA, 50 mM KCl, 5% glycerol, and 200 µg/ml BSA) for 20 min at room temperature and subjected to 4% non-denaturing PAGE in 0.5x Tris-borate-EDTA buffer. The gel was vacuum dried and subjected to autoradiography.

JAK1 kinase assay

100 ng/well of purified active recombinant human JAK1 kinase (Life Technologies, cat. No. PV4774) was incubated for 1 h with serial 3-fold dilutions of inhibitors, 5 µM ATP and 0.2 µg/µl Poly E₄Y₁ as substrate (Sigma) in kinase reaction buffer [40 mM Tris (pH 7.5), 20 mM MgCl₂, and 0.1 mg/ml BSA]. Reaction was performed in a total volume of 25 µl in a solid white flat-bottom 96-well plate. Kinase activity was assayed using an ADP-Glo Kinase Assay (Promega). Assay was done in duplicates for each inhibitor, and curve fitting was performed using Graph Pad Prism sigmoidal dose-response (variable slope) software.

Statistical analysis

All statistical analyses were performed using GraphPad Prism, version 5.03 for Windows (GraphPad Software, San Diego, California). Unpaired or multiple t-tests were used for comparison between groups and p values of <0.05 were considered significant.

RESULTS

Identification of TPCA-1 and ruxolitinib as effective enhancers of VSV replication in VSV-resistant HPAF-II cells

We have shown previously that 4 out of 11 tested human PDAC cell lines were resistant to VSV infection (Moerdyk-Schauwecker et al., 2013; Murphy et al., 2012), at least in part due to constitutive high-level expression of ISGs (Moerdyk-Schauwecker et al., 2013). Pretreatment of resistant cell lines with JAK Inh. I (a reversible inhibitor of JAK1, JAK2, JAK3 and TYK2) reduced ISG expression and partially overcame resistance to VSV (Moerdyk-Schauwecker et al., 2013), suggesting potential for further improvement by utilizing other inhibitors and/or targeting additional pathways. Therefore, in the present study we tested a panel of 16 inhibitors targeting different pathways, shown to directly or indirectly affect ISG expression and/or replication of VSV or other viruses in other experimental systems. As a positive control we included JAK Inh. I. In addition, we included ruxolitinib (INCB018424, trade name Jakafi), a selective inhibitor of JAK1 and JAK2. We also tested two histone deacetylase (HDAC) inhibitors, SAHA (also know as Vorinostat), and valproic acid (VPA), both previously shown to inhibit ISG expression and enhance VSV replication in other systems (Chang et al., 2004; Nguyên TL, 2008; Shulak et al., 2014). As the NF-κB signaling pathway was reported to affect IFN regulated gene expression (Pfeffer et al., 2004), the following inhibitors affecting different factors/steps in the NF-κB signaling pathway were included: eight IKK inhibitors (TPCA-1, SC-514, IKK-16, IKK Inh. XIII, IMD-0354, BMS-345541, IKK-2 Inh. VIII, and sulfasalazine); a 20S proteosome inhibitor (bortezomib); a MEK1/2 inhibitor (U-0126); a mTOR inhibitor (rapamycin); and a COX-2 inhibitor (celecoxib).

VSV-ΔM51-GFP, which has a deletion of the methionine at amino acid position 51 of the matrix protein (between the VSV G and L genes) (Wollmann G et al., 2010) was used in this study. The ΔM51 and other M51 mutations in the VSV matrix protein prevent wild type (wt) matrix protein’s ability to shut down expression of antiviral genes (Ahmed et al., 2003; Kopecky et al., 2001; Stojdl DF et al., 2003). Therefore, VSV-ΔM51 is unable to successfully replicate in healthy cells with intact type I I FN responses. However, as many cancer cells have defective type I IFN signaling (Obuchi M et al., 2003), they remain susceptible to VSV-ΔM51 infection. VSV recombinants with M51 mutation are some of the best performing oncolytic VSVs [reviewed in (Hastie and Grdzelishvili, 2012)], and, compared to wt VSV (Bi Z et al., 1995; Reiss et al., 1998; van den Pol et al., 2002), they show a significantly improved oncoselectivity and decreased neurotoxicity (Stojdl DF et al., 2003; Wollmann G et al., 2010).

The screening of the inhibitors was conducted on one of the most VSV-resistant human PDAC cell lines, HPAF-II (Moerdyk-Schauwecker et al., 2013; Murphy et al., 2012). Cells were treated with each inhibitor at different concentrations based on previously reported effective doses. Following inhibitor treatment, cells were infected with VSV-ΔM51-GFP at MOI of 0.001. VSV-driven GFP fluorescence was measured for 5 days p.i. (Fig. 1A and Supplementary Fig. 1A) and cell viability was determined at 5 days p.i. by MTT assay (Fig. 1B and Supplementary Fig. 1B).

Fig. 1. Effect of TPCA-1, ruxolitinib and Jak Inh. I on VSV-infected HPAF-II.

A) Cells were treated with serial dilutions of each inhibitor for 48 h prior infection with VSV-ΔM51-GFP (cell specific MOI 0.001). GFP fluorescence was measured and background fluorescence from uninfected treated cells was subtracted at each time point p.i. (*) indicates statistical significance (p < 0.05) between treatment and no treated cells (0 µM) at 48 and 72 h p.i. B) Cell viability was analyzed by MTT assay at 5 days p.i. and is plotted as percentage of the uninfected treated with no drug control. The assays were done in triplicate and data represent the mean ± SD of mean. (*) indicates statistical significance (p < 0.05) between infected and uninfected cells within the same treatment. C) Cells were infected with VSV-ΔM51-GFP at cell specific MOI 0.01 for 1 h, then treated with TPCA-1 (8 µM), ruxolitinib (2.5 µM) or JAK Inh. I (2.5 µM). Treatment was maintained until the end of the experiment. GFP fluorescence was measured at the specified time point p.i. (*) indicates statistical significance (p < 0.05) between treatment and no treated cells (no drug) at 48 and 72 h p.i. Media from infected cells were collected at 8, 24, 48, and 72 h p.i. and new infectious viral particle production was determined by plaque assay on BHK-21 cells. Titers were determined in duplicate and data represent the mean ± SD of mean. (*) indicates statistical significance (p < 0.05) between inhibitor treated and untreated (no drug) cells within the same time point p.i. D) Cells were treated with no drug or with TPCA-1 (8 µM), ruxolitinib (2.5 µM), or JAK Inh. I (2.5 µM) for 48 h before infection with VSV-ΔM51-GFP (cell specific MOI 0.001). Percentage of GFP positive cells was determined by flow cytometry at 48 h p.i. Gated populations are positive for GFP. The assay was done in triplicate and data represent the mean ± SD of mean.

In agreement with our previous study (Moerdyk-Schauwecker et al., 2013), JAK Inh. I treatment increased VSV-driven GFP fluorescence (Fig. 1A). A similar enhancement of VSV replication was shown for ruxolitinib, which was previously shown to break resistance of human head and neck cancer cells to VSV (Escobar-Zarate et al., 2013), but has never been tested in combination with VSV in PDAC cells. It should be noted that at the highest concentration tested, ruxolitinib was highly toxic to the cancer cells (Fig. 1B). The HDAC inhibitor SAHA (8 µM) showed a small effect, which was statistically significant but 25-fold less effective compared to fluorescence values reached by treatment with JAK Inh. I or ruxolitinib (Supplementary Fig. 1A). Surprisingly, among the inhibitors targeting the NF-κB pathway, only one, TPCA-1, increased VSV-driven GFP fluorescence and matched levels achieved with the JAK inhibitors (JAK Inh. I and ruxolitinib) (Supplementary Fig. 1A). The effect of TPCA-1 treatment on VSV replication was confirmed with TPCA-1 purchased from two different providers (data not show).

Importantly, increase in VSV-driven GFP expression in HPAF-II cells treated with TPCA-1, ruxolitinib or JAK Inh. I directly correlate with increases in new viral particle production (Fig. 1C). Percentage of GFP positive cells measured by flow cytometry at 48 h p.i. showed an increase from 1.7% for cells treated with no drug to 99.1%, 98.7% and 89.2% for TPCA-1, ruxolitinib and JAK Inh. I treatment, respectively (Fig. 1D).

When VSV-mediated cell killing was determined by MTT, striking decreases of 83%, 90%, and 86% in cell viability were observed for JAK Inh. I (5 µM), ruxolitinib (8 µM) and TPCA-1 (8 µM) treatments, respectively, compared to uninfected cells (Fig. 1B). Treatment with SAHA (8 µM) caused a decrease in cell viability comparable to TPCA-1 (8 µM), even though its effect on VSV-driven GFP fluorescence was marginal (Supplementary Fig. 1A and 1B). This suggests epigenetic modifications of chromatin may affect VSV induced cell death independently of viral replication. While treatments with IKK Inh. XIII (0.8 µM), BMS-345541 (4 µM), and rapamycin (80, 8 and 0.8 nM) also showed statistically significant decrease in cell viability of infected compared to uninfected cells (21%, 33%, and up to 26%, respectively), these were not as pronounced as the effect of TPCA-1, ruxolitinib or JAK Inh. I treatment (Supplementary Fig. 1B).

As all inhibitors, except JAK Inh. I, rapamycin, celecoxib and VPA, showed significant toxicity in uninfected cells at the highest tested concentrations compared to uninfected cells treated with no drug, it is unlikely that any of the ineffective inhibitors would enhance VSV replication at even higher concentrations (Supplementary Fig. 1B).

TPCA-1 and ruxolitinib overcome resistance to VSV in all VSV-resistant PDAC cell lines

To determine if the enhancement of VSV replication by TPCA-1 and ruxolitinib was limited only to HPAF-II cells, we tested these inhibitors as well as JAK Inh. I in three additional VSV-resistant PDAC cell lines, Hs766T cells (shows a high resistance to VSV, similar to HPAF-II), CFPAC-1 and HPAC (both show an intermediate resistance to VSV) (Moerdyk-Schauwecker et al., 2013; Murphy et al., 2012). Cells were treated with 4 different concentrations of TPCA-1, ruxolitinib or JAK Inh. I for 48 h prior to infection with VSV-ΔM51-GFP at an MOI of 1.5 (based on titration on BHK-21 cells. Fig. 2 legend indicates cell specific MOIs). TPCA-1, ruxolitinib, and JAK Inh. I, enhanced VSV-ΔM51-GFP replication in all VSV-resistant PDAC cell lines (Fig. 2A). The observed lack of a dose-dependent response in some of the cell lines may be due to the narrow range of drug dilutions used in this experiment. For example, there was no dose-dependent effect for ruxolitinib in HPAF-II cells in Figure 2A (2-fold dilution), but clearly showed a dose dependency when tested at 10-fold dilution (Fig. 1A). In agreement with GFP fluorescence data, for all VSV-resistant PDAC cell lines, treatment with TPCA-1, ruxolitinib and JAK Inh. I at all concentrations tested caused statistically significant decrease in cell viability in all VSV infected cells compared to uninfected cells (Fig. 2B). Together, our results show that TPCA-1, ruxolitinib and JAK Inh. I are effective in overcoming resistance to VSV in all identified VSV-resistant PDAC cell lines.

Fig. 2. Effect of TPCA-1, ruxolitinib and JAK Inh. I on four different VSV-resistant PDAC cell lines.

CFPAC-1, HPAC, Hs766T and HPAF-II cells were treated with TPCA-1, ruxolitinib or JAK Inh. I for 48 h before infection with VSV-ΔM51-GFP at MOI 1.5 (based on BHK-21 cells). Cell specific MOIs are MOI 0.008 based on CFPAC, MOI 0.005 based on HPAC, MOI 0.003 based on Hs766T and MOI 0.001 based on HPAF-II. A) GFP fluorescence was measured and background fluorescence from uninfected treated cells was subtracted at each time point p.i. (*) indicates statistical significance (p < 0.05) between treatment and no treated cells (0 µM) at 48 h p.i. B) Cell viability was analyzed by MTT assay at 5 days p.i., and is expressed as percent of the uninfected treated with no drug control. Results in each row correspond to the same cell line as in A. The assays were done in triplicate and data represent the mean ± SD of mean. (*) indicates statistical significance (p < 0.05) between infected and uninfected cells within the same treatment.

Treatment with TPCA-1, ruxolitinib or JAK Inh. I immediately post infection is sufficient to increase VSV replication

Our previous experiments with JAK Inh. I (Moerdyk-Schauwecker et al., 2013) and those presented above were performed by pre-treating cells with inhibitors for 48 h prior to infection. To examine the treatment schedule required to overcome resistance of PDAC cells to VSV, Hs766T and HPAF-II cells were treated with TPCA-1 (8 µM), ruxolitinib (2.5 µM) or JAK Inh. I (2.5 µM) for either one or two days prior to infection with VSV-ΔM51-GFP at MOI of 1.5 (based on titration on BHK-21 cells. Fig. 3 legend indicates cell specific MOIs), and treatment was maintained for 5 days p.i. or removed right after infection. Alternatively, inhibitors were not added before infection but instead were added right after infection and maintained for 5 days (Treatment Schedule in Fig. 3A).

Fig. 3. Effect of inhibitor treatment timings on VSV infection and oncolysis in VSV-resistant PDAC cells.

A) HPAF-II and Hs766T cells were seeded 3 days before infection (d₋₃) and treated with TPCA-1 (8 µM), ruxolitinib (2.5 µM) or JAK Inh. I (2.5 µM) for 0, 1, or 2 days before infection with VSV-ΔM51-GFP at MOI 1.5 (based on BHK-21 cells). Cell specific MOIs are MOI 0.003 based on Hs766T and MOI 0.001 based on HPAF-II. Virus was removed after 1 h and replaced with either media with inhibitor or no drug (see Treatment Schedule). GFP fluorescence was measured at each time point p.i. Statistical significances (p < 0.05) between treatment schedules and the treatment schedule 6 (no drug) at 48 (*) and 72 (#) h p.i. are indicated. Cell viability was analyzed in HPAF-II cells by MTT assay at 120 h p.i., and is showed as percentage of cells treated with no drug. The assays were done in triplicate and data represent the mean ± SD of mean. (*) indicates statistical significance (p < 0.05) between inhibitor treated and untreated cells (no drug) within the same treatment schedule. B) HPAF-II cells were seeded 2 days before infection (d₋₂) and treated with TPCA-1 (8 µM), ruxolitinib (2.5 µM) or JAK Inh. I (2.5 µM) for 1 day before infection with VSV-ΔM51-GFP (cell specific MOI 0.001). Virus was removed after 1 h and replaced with either media with inhibitor or no drug. Every 24 h p.i. media was removed and replaced with either media with inhibitor or no drug (see Treatment Schedule). GFP fluorescence was measured at each time point p.i. The assays were done in triplicate and data represent the mean ± SD of mean. Statistical significances (p < 0.05) between treatment schedules and the treatment schedule 7 (no drug) at 48 h (*) and 72 h (#) are indicated.

In both Hs766T and HPAF-II, a significant increase in VSV-driven GFP fluorescence was observed only when treatment was maintained after infection, regardless if or when treatment was applied (GFP fluorescence in Fig. 3A; treatment schedules 1, 2 and 5). However, cells remained resistant to VSV infection if treatment was removed before infection (GFP Fluorescence in Fig. 3A; treatment schedules 3 and 4). In Hs766T cell, the absence of treatment before infection (GFP Fluorescence in Fig. 3A; treatment schedule 5) delayed the increase in GFP fluorescence by 16 h for TPCA-1 and JAK Inh. I (but not for ruxolitinib), suggesting the uptake/activation of ruxolitinib is faster than for the other two inhibitors. In contrast, no difference was observed between the three inhibitors in HPAF-II cells. In agreement with these results, MTT assay at day 5 p.i. showed a dramatic decrease in cell viability in VSV-infected cells when inhibitor treatments were maintained after infection, regardless of the presence of inhibitors prior to infection (Cell viability in Fig. 3A). In general, for all three inhibitors, this experiment indicates that treatment before infection is not as critical as treatment immediately after infection.

To determine the optimal duration for treatment p.i., HPAF-II cells were treated with each of the inhibitors for one-day before infection, and then the treatment was either removed immediately after infection, or kept for different periods of time up to 5 days p.i. (Treatment Schedule in Fig. 3B). All inhibitors were ineffective if removed prior to infection (GFP Fluorescence in Fig. 3B; treatment schedule 1). However, when treatment was maintained for 1 day p.i., GFP fluorescence increased up to 1.7-fold with TPCA-1 and ruxolitinib, and up to 1.5-fold with JAK Inh. I. (GFP Fluorescence in Fig. 3B; treatment schedule 2). GFP fluorescence was maximally increased (up to 3-fold with TPCA-1 and ruxolitinib and up to 2.5-fold with JAK Inh. I) when treatment was maintained more than 1 day p.i. (GFP Fluorescence in Fig. 3B; treatment schedules 3–6).

Together our data suggest that TPCA-1, ruxolitinib and JAK Inh. I in HPAF-II and Hs766T cells reversibly inhibit their targets and enhance VSV replication, but are not effective if removed prior to viral exposure.

TPCA-1 and ruxolitinib decrease expression of antiviral ISGs MxA and OAS in uninfected HPAF-II cells

Our screening identified TPCA-1 and ruxolitinib as effective enhancers of VSV replication and VSV-mediated oncolysis in VSV-resistant PDAC cell lines. The identification of ruxolitinib, a JAK1/2 inhibitor, is expected as it acts similarly to the less specific pan-JAK inhibitor, JAK Inh. I. More surprisingly TPCA-1, a selective inhibitor of IKK-β (Birrell et al., 2005; Birrell et al., 2006; Podolin et al., 2005), was the only inhibitor targeting the NF-κB to enhance VSV replication. To determine the impact of inhibitors on constitutive ISG expression, we analyzed the expression of MxA and OAS at 48 h post-inhibitor treatment in uninfected HPAF-II cells (Fig. 4A). As expected, inhibition of the JAK/STAT signaling pathway by JAK Inh. I or ruxolitinib almost completely eliminated MxA protein levels and markedly reduced OAS protein levels compared to cells treated with no drug. Importantly, a similar effect on MxA and OAS was seen in TPCA-1 treated cells. On the other hand, the others IKK-β inhibitors, IKK-16, IKK Inh. XIII, and BMS-345541, did not affect levels of MxA or OAS protein. Interestingly, sulfasalazine (an IKK-αCβ inhibitor) also decreased MxA and OAS protein levels, however, to a degree not sufficient to affect VSV-driven GFP fluorescence or VSV-mediated cell death (Supplementary Fig. 1). Surprisingly, even though treatment with SAHA (an HDAC inhibitor) caused the strongest decrease in MxA and OAS protein level (Fig. 4A), only showed a very minor increase on VSV-driven GFP fluorescence (Supplementary Fig. 1A). It is possible that large epigenetic modifications associated with SAHA treatment, while strongly down regulating MxA and OAS, may also negatively affect other aspects of the virus replication cycle. Future studies will examine the impact of SAHA on ISG expression and VSV replication in PDAC cells.

Fig. 4. Effect of inhibitors on protein and mRNA levels in HPAF-II.

A) Cells were treated with TPCA-1 (8 µM), IKK-16 (0.8 µM), IKK Inh. XIII (0.8 µM), IKK-2 Inh. VIII (8 µM), BMS-345541 (4 µM), Sulfasalazine (4 mM), SAHA (8 µM), Celecoxib (80 µM), Rapamycin (80 nM), ruxolitinib (2.5 µM) or JAK Inh. I (2.5 µM) for 48 h. Cells lysates were prepared and analyzed by western blot for the indicated protein. Protein sizes (kDa) are indicated on the right. B) Cells were treated with TPCA (8 µM), SC-514 (80 µM), IKK-16 (0.8 µM), IKK Inh. XIII (0.8 µM), IKK-2 Inh. VIII (8 µM), IMD-0354 (0.8 µM), or BMS-345541 (4 µM) for 2 h prior to addition of TNF-α (25 ng/ml) or IFN-α (5000 U/ml). Cells were harvested at 4 h post-induction and extracted mRNA was reverse transcribed and analyzed by PCR.

Of the tested IKK inhibitors, TPCA-1 was the only one enhancing VSV replication. This could be due to the lack of uptake or activity of other IKK inhibitors in HPAF-II cells, or if TPCA-1 affects other targets in addition (or instead) of IKKs. To examine if tested IKK inhibitors which did not show any effect (Supplementary Fig. 1) were active in HPAF-II cells, we determined the effect of these inhibitors on TNF-α-mediated induction of TNF-α mRNA (an NF-κB dependent gene) and IFN-α-mediated induction of MxA mRNA (Fig. 4B). Semi-quantitative RT-PCR results showed that at least 5 out of all 7 IKK inhibitors effectively blocked induction of TNF-α-mediated TNF-α mRNA synthesis. Together, these results demonstrate that TPCA-1 is the only one of the tested IKK inhibitors capable of simultaneously inhibit NF-κB pathway and IFN-mediated ISGs expression in PDAC cells, while enhancing VSV replication.

Combining TPCA-1 with ruxolitinib or JAK Inh. I has no cooperative effect on VSV replication or ISG expression

Next, we wanted to examine whether TPCA-1 and JAK inhibitors affect VSV replication and ISG expression via the same or different mechanism. We hypothesized that if TPCA-1 and JAK inhibitors enhance VSV replication via different mechanisms, combining them would have an additive or synergistic effect on VSV replication. HPAF-II and Hs766T cells were treated with no drug or either with TPCA-1 (8 µM), ruxolitinib (2 µM) or both inhibitors together. One day after treatment, cells were infected with VSV-ΔM51-GFP at MOI of 0.001. After infection, treatments were maintained for 4 days and VSV-driven GFP fluorescence was measured every 24 h. As shown in Fig. 5A, VSV-driven GFP fluorescence measured from cells treated with both inhibitors reached intermediate values compared to each inhibitor alone, indicating no additive or synergistic effect of this combined treatment.

To examine the effect of the TPCA-1/ruxolitinib combination in more detail, each inhibitor was used at four different concentrations for a total of 16 combination treatments. One day after treatment, cells were infected with VSV-ΔM51-GFP at an MOI of 0.001, and VSV-driven GFP fluorescence was measured at 2 days p.i. Combination Index (CI) values were calculated for each combined treatment by the method of Chou and Talalay (Chou, 2006; Chou and Talalay, 1984) (Fig. 5B). For all but the lowest concentrations, CI values were greater than 1, indication of an antagonistic effect. Specifically, for the optimal concentration determined for each inhibitor and used throughout our studies (8 µM for TPCA-1 and 2.5 µM for ruxolitinib), the CI values were 2.64 and 1.77 for HPAF-II and Hs766T, respectively. An antagonistic effect suggests that both inhibitors affect the same downstream target(s) essential for VSV replication with the combination causing an effect(s) that negatively impacts virus replication and partially counteracts the positive effects seen during monotherapy.

Similar results were obtained with a TPCA-1 and JAK Inh. I combination in HPAF-II, HPAC, and Hs766T cells (Fig. 5C). Despite targeting two different signaling pathways in the cell, JAK Inh. I and TPAC-1 showed similar decreases in MxA expression and increases in VSV protein expression. BMS-345541 (BMS) was also included as an example of an IKK inhibitor with no effect on VSV replication and ISG expression.

TPCA-1 downregulates type I IFN-mediated JAK/STAT signaling pathway by directly inhibiting JAK1

Next, we examined in more detail how mechanistically TPCA-1 could inhibit ISG expression. First, we analyzed how activation and inhibition of type I IFN and NF-κB pathways affect expression of MxA and OAS. HPAF-II cells were treated with no drug or with TPCA-1 or ruxolitinib for 2 h then activated with either IFN-α or TNF-α for 4 h (Fig. 6A). Gene expression was determined by real-time PCR and normalized to the expression of GAPDH. In cells without TPCA-1 or ruxolitinib treatment, IFN-α induced a strong increase in MxA and OAS expression but showed no effect on TNF-α expression, a gene under control of the NF-κB pathway. Conversely, TNF-α treatment upregulated TNF-α expression, but did not affect MxA or OAS expression. This result shows that activation of the NF-κB pathway does not result in MxA and OAS expression in HPAF-II cells, suggesting that NF-κB pathway does not play a major role in the expression of these ISGs. As expected, TPCA-1 (but not ruxolitinib) inhibited constitutive as well as TNF-α induced expression of TNF-α, and ruxolitinib significantly inhibited constitutive as well as IFN-α induced expression of MxA and OAS. Surprisingly, TPCA-1 also inhibited both constitutive and IFN-α induced expression of MxA and OAS. The ability of TPCA-1 to inhibit MxA and OAS expression even in the presence of exogenously added IFN-α suggests that TPCA-1 affects ISG expression by directly inhibiting IFN signaling, rather than by inhibiting IFN-β expression. It is known that NF-κB can mediate expression of IFN-β, which acts through the IFN-α/β receptor in an autocrine manner to maintain basal expression of ISGs in uninfected cells (Basagoudanavar et al., 2011; Taniguchi and Takaoka, 2001). In agreement with it, our data show that a small but statistically significant decrease in constitutive IFN-β expression was observed in the presence of TPCA-1 (Fig. 6A). This suggests that TPCA-1 could affect ISGs expression by inhibiting NF-κB dependent IFN-β expression. However, this mechanism cannot explain our result demonstrating that TPCA-1 inhibits MxA and OAS expression even in the presence of exogenously added IFN-α. Together, our data suggest that TPCA-1 inhibits IFN-mediated ISG expression mainly via direct inhibition of type I IFN signaling downstream of type I IFN production, although TPCA-1-mediated decrease of the basal IFN-β expression could have a minor contribution to the decreased ISG levels.

Fig. 6. TPCA-1 directly inhibits the JAK/STAT signaling pathway.

HPAF-II cells were treated with TPCA-1 (8 µM), ruxolitinib (2.5 µM), or JAK Inh. I (2.5 µM) for 2 h prior to induction with either TNF-α (25 ng/ml) or IFN-α (5000 U/ml). Cells were harvested at 4 h post-induction. A) Relative gene expression was analyzed by real-time PCR and normalized to GAPDH expression. Fold change expression was calculated by the comparative Ct method. B) Cell lysates were prepared and analyzed by western blot for the indicated protein. Protein sizes (kDa) are indicated on the right. C) HPAF-II cells were treated with TPCA-1 (8 µM), ruxolitinib (2.5 µM), or JAK Inh. I (2.5 µM) for 2 h prior to induction with IFN-α (5000U/ml) for the indicated time. Cells were harvested and nuclear extracts were subjected to EMSA using a radiolabeled ISRE probe. D) In situ titration of TPCA-1, Ruxolitinib and BMS-345541 were performed with recombinant human JAK1 kinase using a luminescent ADP detection assay. Reactions were carried out at 5 µM ATP. The assays were done in duplicate and data represent the mean ± SD of mean. Curve fitting was performed using GraphPad Prism sigmoidal dose-response (variable slope) software.

Type I IFNs (including IFN-α and IFN-β) signal primarily through JAK/STAT pathways, although they can also activate other signaling pathways (Bonjardim CA, 2009). Ligand binding to the IFN-α/β receptor activates phosphorylation of JAK1 and TYK2, which then phosphorylate STAT1 and STAT2, which are also bound to the IFN-α/β receptor. This results in the formation of transcription factor complexes, which translocate to the nucleus and promote the transcription of ISGs. Binding of type I IFNs to their receptor most commonly results in a transcription factor complex of IFN-regulatory factor 9 (IRF-9), p-STAT1 and p-STAT2 (known as ISGF3), which recognizes IFN stimulated response elements (ISRE). To further study the mechanism by which TPCA-1 directly affects the type I IFN pathway, we analyzed expression and phosphorylation of several proteins involved in the signaling pathway (Fig. 6B). HPAF-II cells were treated with no drug or with TPCA-1, ruxolitinib or JAK Inh. I for 2 h and then activated with either IFN-α or TNF-α for 4 h. TPCA-1, but not ruxolitinib or JAK Inh. I, inhibited phosphorylation of IκBα (a target of IKK-β). Phosphorylation of IκBα is required for its proteasome-mediated degradation, resulting in the release and nuclear translocation of active NF-κB (Brown et al., 1995). Importantly, TPCA-1, ruxolitinib and JAK Inh. I prevented induction of MxA and OAS by IFN-α, with TPCA-1 being the most effective of all three. Interestingly, TPCA-1, ruxolitinib and JAK Inh. I were able to reduce low but detectable levels of constitutively phosphorylated STAT1 (which is also an ISG), which would account for the effect of these inhibitors on constitutive expression of ISGs in HPAF-II. Moreover, in the presence of ruxolitinib or JAK Inh. I IFN-α induced activation of STAT1 and STAT 2 was markedly reduced, but it was completely abolished in presence of TPCA-1.

We also directly tested the ability of TPCA-1, ruxolitinib and JAK Inh. I to prevent the functional consequence of activating the type I IFN pathway. If TPCA-1 is a direct inhibitor of IFN signaling, it should prevent ISGF3 transcription factor nuclear re-localization and ISRE binding following activation with IFN-α. HPAF-II cells were treated with TPCA-1, ruxolitinib or JAK Inh. I for 2 h, followed by 0.5 or 1 h induction with IFN-α. Nuclear extracts prepared from these cells were incubated with a radiolabeled ISRE probe (Fig. 6C). Formation of an IFN-α dependent complex was prevented by all three inhibitors, suggesting that TPCA-1 inhibits type I IFN signaling in a manner indistinguishable from ruxolitinib or JAK Inh. I.

To directly examine whether TPCA-1 inhibits phosphorylation of STAT1 and STAT2 via inhibition of JAK1, we performed an in situ kinase activity assay using a purified recombinant human JAK1 in the presence of TPCA-1, ruxolitinib or BMS-345541 (Fig. 6D). ruxolitinib and TPCA-1 were able to inhibit JAK1 kinase activity in a dose dependent manner with an IC₅₀ of 4.33 nM and 43.78 nM, respectively, while BMS-345541 did not affect the kinase enzymatic activity. These data provide biochemical evidence that TPCA-1 directly inhibits JAK1 kinase, which can explain inhibition of STAT1 and STAT2 phosphorylation in vitro (Fig. 6B).

We next investigated the effect of inhibitor treatments on the kinetics of VSV replication, MxA expression and apoptosis induction (Fig. 7). HPAF-II cells were infected with VSV-ΔM51-GFP at an MOI of 0.01, and immediately treated with no drug or with TPCA-1, ruxolitinib or JAK Inh. I. Cell lysates were prepared at 8, 24, 48 and 72 h p.i. and analyzed by Western blot. First, it is important to note that our microscopic observation showed extensive cell death at 48 and 72 h p.i. of infected cells treated with TPCA-1 and ruxolitinib (data not shown), which explains lower levels of total protein isolated from the corresponding cells (see protein staining in Figure 7; an equal fraction of total retrieved cell lysate was loaded for each sample). At all time points after 8 h p.i. (and despite decrease in total protein loaded from 48 and 72 h p.i.), Western blot showed a dramatic increase in accumulation of viral protein in cells treated with TPCA-1, ruxolitinib, or JAK Inh. I. TPCA-1 and ruxolitinib also strongly inhibited STAT1 and STAT2 phosphorylation (at all time points, including 8 and 24 h p.i. where equal amount of total protein were loaded), as wells as total STAT1 and STAT2 levels (after 8 h p.i.), likely accounting for stable decrease of MxA expression (Fig. 7). On the other hand, JAK Inh. I, while inhibiting STAT2 phosphorylation, failed to suppress STAT1 phosphorylation, total STAT1 and STAT2 expression, and MxA expression long term. However, all 3 inhibitors strongly inhibited STAT1 phosphorylation at 8 h p.i., which was likely sufficient to ensure similar high levels of VSV proteins for all 3 inhibitors. Also, despite these differences, all three inhibitors strongly induced VSV-mediated apoptosis, demonstrated by Caspase 3 and PARP cleavage at 24 and 48 h p.i. It should be noted that, despite the induction of apoptosis at 24 and 48 h p.i. in JAK Inh. I treated cells, loss of cells (and cell material) is not observed at these time points (Fig. 7). However, at 120 h p.i. cell viability was clearly reduced in JAK Inh. I treated cells (Fig. 1B).

Fig. 7. Kinetics of VSV replication, MxA expression and apoptosis induction in HPAF-II cells treated with TPCA-1, ruxolitinib or JAK Inh. I.

Cells were infected with VSV-ΔM51-GFP at cell specific MOI 0.01 for 1 h, then treated with TPCA-1 (8 µM), ruxolitinib (2.5 µM) or JAK Inh. I (2.5 µM) immediately after virus removal. Treatment was maintained until the end of the experiment. Cells were harvested at each time point and lysates were prepared and analyzed by western blot for the indicated protein.

TPCA-1, ruxolitinib and JAK Inh. I enhance replication of Sendai virus

Our data suggest that ruxolitinib and TPCA-1-mediated enhancement of VSV-ΔM51-GFP replication in PDAC cell lines occurs via modification of a cellular environment (inhibition of antiviral signaling), rather than virus-specific mechanisms. Therefore, we reasoned that the identified inhibitors should stimulate not only VSV but also other OVs. To test this hypothesis, we examined Sendai virus (SeV, a paramyxovirus), another promising OV (Kinoh H, 2004; Kinoh H, 2008; Komaru A, 2009; Yonemitsu Y, 2008). Importantly, our previous study demonstrated that all VSV-resistant PDAC cell lines were also resistant to SeV (Murphy et al., 2012). In addition, we examined the effect of inhibitors on VSV and SeV in two VSV-permissive PDAC cell lines. Evaluating possible adverse effects of the inhibitors in cells permissive to virus replication would allow us to determine whether OV/ruxolitinib and/or OV/TPCA-1 combinations can be effectively used against all PDACs, regardless of their type I IFN status.

VSV-ΔM51-GFP and SeV-GFP-F_mut (described in Materials and Methods) were tested in HPAF-II, Hs766T, and two VSV-susceptible (AsPC-1 and MiaPaCa-2) PDAC cell lines. Cells were infected with VSV-ΔM51-GFP or SeV-GFP-F_mut at MOI of 1 (based on virus titration on BHK-21 cells, Fig. 8 legend indicates cell specific MOIs), and treated with no drug or with TPCA-1, ruxolitinib or JAK Inh. I. GFP fluorescence was measured through 12 days p.i. due to the slow SeV-F_mut replication kinetics (Fig. 8). TPCA-1, ruxolitinib and JAK Inh. I enhanced VSV and SeV replication in both VSV-resistant PDA cell lines. Interestingly, TPCA-1 and ruxolitinib (and to a less degree JAK Inh. I) also enhance SeV-GFP-F_mut replication, but not VSV-ΔM51-GFP, in MiaPaCa-2 cells. Importantly, none of the inhibitors negatively impact VSV or SeV replication in both susceptible cell lines.

Fig. 8. TPCA-1, ruxolitinib and JAK Inh. I enhance replication of SeV.

Cells were infected with VSV-ΔM51-GFP or SeV-GFP-F_mut at MOI 1 (based on BHK-21 cells), and treated with TPCA-1 (8 µM), ruxolitinib (2.5 µM) or JAK Inh. I (2.5 µM). Cell-specific MOIs are MOI 0.00068 based on HPAF-II, MOI0.0018based on Hs766T, MOI 0.024 based on AsPC-1 and MOI 0.066 based on Mia PaCa-2. GFP fluorescence was measured and background fluorescence from uninfected treated cells was subtracted at each time point p.i. The assays were done in triplicate and data represent the mean ± SD of mean. Statistical significances (p < 0.05) between treatment (T: TPCA-1, R: ruxolitinib, J: JAK Inh. I) and no treated cells (no drug) at 72 h (*) and 148 h (#) are indicated.

DISCUSSION

In this study, we identified two small molecule inhibitors that overcome resistance of PDAC cells to oncolytic VSV. TPCA-1 and ruxolitinib, which have not been previously tested in PDAC in combination with any virus, enhanced VSV replication and virus-mediated cell death. Significantly, these inhibitors were effective in all VSV-resistant PDACs, and also enhanced replication of SeV. Importantly, we demonstrated on a molecular level that TPCA-1, a well characterized IKK-β inhibitor, is also a direct inhibitor of type I FN signaling by directly inhibiting JAK1 kinase activity, down-regulating expression of antiviral ISGs independently from its effect on the NF-κB pathway.

Our data provide a new evidence that the upregulated type I interferon signaling plays a major role in resistance of pancreatic cancer cells to oncolytic viruses, which is in agreement with a growing number of reports indicating that a significant percentage of cancers retain active type I FN signaling, virus-induced antiviral ISG expression, or even constitutive expression of many ISGs (Escobar-Zarate et al., 2013; Linge et al., 1995; Matin et al., 2001; Naik and Russell, 2009; Pfeffer et al., 1996; Saloura et al., 2010; Stojdl DF et al., 2003; Stojdl et al., 2000; Sun et al., 1998; Wong et al., 1997). Our previous studies using a panel of 11 human PDAC cell lines showed that 4 of them (36%) constitutively express several antiviral ISGs, including MxA and OAS (Moerdyk-Schauwecker et al., 2013), and were resistant to VSV and other OVs (Moerdyk-Schauwecker et al., 2013; Murphy et al., 2012). Importantly, a significant subset of tested clinical PDAC tissues and xenografted primary PDAC cells had upregulation of ISGs expression such as MxA (Monsurro et al., 2010), indicating the existence of this phenotype in the PDAC patient population.

In our previous study we showed that a pan-JAK inhibitor, JAK Inh. I, reduced ISG expression in all VSV-resistant PDAC cell lines and partially overcame their resistance to VSV (Moerdyk-Schauwecker et al., 2013). Ruxolitinib, one of the effective inhibitors identified in the present study, is also a JAK inhibitor, but with much greater specificity towards JAK1 and JAK2 (IC₅₀ of 3.3 and 2.8 nM for JAK1 and JAK2 respectively, and >130-fold for JAK3 over JAK1 and JAK2) than JAK Inh. I (Quintas-Cardama et al., 2010). Ruxolitinib was more effective than JAK Inh. I in overcoming resistance of all VSV-resistant PDAC cell lines to VSV. Importantly, ruxolitinib is approved for the treatment of myelofibrosis (a bone marrow cancer) (Mascarenhas and Hoffman, 2013; Quintas-Cardama and Verstovsek, 2013), and is currently being investigated against pancreatic cancer in two different clinical trials (NCT01822756 and NCT01423604). Moreover, ruxolitinib was shown to enhance VSV-GFP and VSV-ΔM51-GFP replication, and virus-mediated oncolysis in human head and neck cancer cells (Escobar-Zarate et al., 2013).

While the identification of ruxolitinib is significant, it is not unexpected considering that its mechanism of action is similar to that of JAK Inh. I. We find more surprising a comparable enhancement of VSV replication in all VSV-resistant PDAC cell lines by TPCA-1, a well-known potent IKK-β inhibitor, but not by any other tested inhibitor of the NF-κB pathway, including other IKK inhibitors. TPCA-1 and other NF-κB inhibitors were included in the inhibitor screen as we hypothesized that upregulation of NF-κB, a hallmark of PDACs, could be responsible for the constitutive ISG expression in VSV-resistant cell lines. Constitutive activation of the canonical NF-κB (RelA/p50) pathway occurs in almost 70% of PDAC specimens (Wang et al., 1999), and it promotes tumorigenesis (Fujioka et al., 2003; Ling et al., 2012). Moreover, specifically for PDACs, a mechanistic link has been demonstrated between a common PDAC mutation, KRAS (G12D), and IKK-β activation (Ling et al., 2012).

Despite this role of NF-κB signaling in PDAC tumorigenesis, our data strongly suggests a direct, NF-κB independent, inhibition of JAK/STAT signaling by TPCA-1 as: 1) all other IKK inhibitors, while inhibiting NF-κB activation, did not inhibit MxA expression or increase VSV replication in resistant PDAC cell lines (Supplementary Fig. 1A and Fig. 4B); 2) TPCA-1 treatment directly inhibited STAT1 and STAT2 phosphorylation induced by exogenously added IFN-α, and blocked ISRE binding upon stimulation with IFN-α (Fig. 6B–C); 3) treatment with TNF-α did not increase ISG expression (Fig. 6A); 4) ruxolitinib and TPCA-1 acted antagonistically at all but the lowest concentrations used, suggesting both inhibitors affect the same downstream target(s) essential for VSV replication (Fig. 5B); 5) despite targeting two different signaling pathways in the cell, TPCA-1 and JAK inhibitors showed surprisingly similar effects on ISG expression and VSV replication (Fig. 1 to 7); 6) in situ kinase assay provided a biochemical evidence that TPCA-1 directly inhibits JAK1 kinase activity (Fig. 6D). To the best of our knowledge, this is the first demonstration that TPCA-1 directly inhibits the formation of an active ISGF3 complex and thus inhibits type I IFN signaling.

Interestingly, Du et. al. previously examined the effects of TPCA-1 and BMS-345541 (another IKK inhibitor) on type I IFN antiviral activity and viral replication in several human glioma cell lines, and showed that both inhibitors selectively inhibited IFN stimulated expression of some ISGs, such as MxA and GBP1, and attenuated IFN-induced antiviral state against VSV and encephalomyocarditis virus (EMCV) (Du et al., 2012). They concluded that since both inhibitors are IKK inhibitors, the IKK complex, which lies upstream of NF-κB activation, must play an important role in IFN induced gene expression and antiviral activity. In agreement with their hypothesis, their treatment of glioma cells with BMS-345541 did not prevent STAT1 and STAT2 phosphorylation in cells treated with IFN-α. Unfortunately, the same experiment with TPCA-1 was omitted. In contrast to the glioma study, our experiments in PDAC cells show that TPCA-1 enhances VSV replication and inhibits ISG expression in NF-κB independent manner, through a direct inhibition on the JAK/STAT signaling pathway. Also, our results showed no effect of BMS-345541 on constitutive and IFN-induced expression of MxA (Fig. 4) and on VSV replication (Supplementary Fig. 1A), suggesting that BMS-345541, as a specific IKK inhibitor, is unable to inhibit type I IFN-mediated antiviral action in PDAC cells. One important difference between PDAC cells used in our study and glioma cells used in the glioma study (Du et al., 2012) is that PDAC cells display constitutive high-level expression of ISGs, while glioma cells did not. This may explain the differences between our results.

Our data clearly show that TPCA-1 inhibits IFN-mediated phosphorylation of STAT1 and STAT2 in PDAC cells, and is able to directly inhibit JAK1 kinase activity in situ. Interestingly, previous studies proposed two alternative activities for TPCA-1, in addition to IKK-β inhibition. Thus, a study using HEK-293T cells proposed TPCA-1 as a direct inhibitor of STAT3 phosphorylation by a mechanism of direct binding of TPCA-1 to the SH2 domain of STAT3 and preventing the docking of STAT3 to the membrane complex (Nan et al., 2014). Another study in hepatocytes also reported inhibition of STAT3 phosphorylation in presence of TPCA-1, but through a mechanism of direct inhibition of JAK2 (the kinase of STAT3) (Saez-Rodriguez et al., 2011). Based on an in vitro kinase activity assay of purified recombinant JAK2 (using Fms-like tyrosine kinase 3 (FLT3) as a JAK2 substrate) and IKK-β (using IκB as a IKK-β substrate), TPCA-1 was shown to inhibit JAK2 (Ki about 9 nM) almost as potently as IKK-β (Ki about 1.6 nM). However, these reports about JAK2 and STAT3 inhibition by TPCA-1 may not exclude each other, as both JAK (Leonard and O’Shea, 1998; Radtke et al., 2005; Saez-Rodriguez et al., 2011) and STAT family members contain SH2 domains (a structurally conserved protein domain in many signal-transducing proteins that allow proteins containing those domains to dock to phosphorylated tyrosine residues on other proteins), and many STAT inhibitors have low specificity due to the high homology of the STAT family members (Szelag et al., 2014). Therefore, we cannot exclude the possibilities that, in addition to direct inhibition of JAK1, TPCA-1 also: 1) inhibits JAK2 in PDAC cells; 2) interacts with STAT1 and STAT2 and prevents their docking to the IFN-α/β receptor. It should be also noted that, although the inhibition of JAK1 kinase activity alone can explain our results, we cannot rule out the possibility that TPCA-1 also inhibits TYK2, a kinase very similar to JAK1 and generally a target for the same inhibitors affecting JAK1 (51–53). Future experiments should examine relative inhibitory activities of TPCA-1 against various JAK family members, which is beyond the scope of the present study.

An important potential advantage of using TPCA-1 over specific JAK inhibitors (such as ruxolitinib) is its dual inhibitory activity. On one hand, TPCA-1 improves efficacy of OV therapies by inhibiting expression of antiviral ISGs and enhancing virus replication and virus-mediated oncolysis, while on the other, inhibits NF-κB pathway, which is an important anticancer approach against PDAC and other cancers [reviewed in (Carbone and Melisi, 2012; Erstad and Cusack, 2013; Kim et al., 2006; Luo et al., 2005)]. Disregulated activation of NF-κB plays a major role in tumor development through maintained expression of target genes that regulate cell growth, proliferation and survival (Basseres and Baldwin, 2006; Karin, 2006).

Ruxolitinib and TPCA-1 also strongly enhanced replication of another RNA virus, SeV, in VSV-resistant PDAC cell lines. Importantly, none of the inhibitors negatively impacted VSV or SeV replication in virus-permissive PDAC cell lines, suggesting that the inhibitors do not have side effects on viral replication in PDAC cells with defective type I IFN signaling, and that the approach combining OVs with TPCA-1, ruxolitinib or JAK Inh. I could enhance OV efficacy against virus-resistant PDACs without compromising it in virus-permissive PDACs.

It is important to be aware that inhibition of innate antiviral responses may result in increase of virulence in normal tissues. However, previously combined treatments of viruses with small molecule inhibiting antiviral responses were examined in vivo and showed promising results. For example, the mammalian target of rapamycin (mTOR) stimulates type I IFN production via phosphorylation of its effectors. VSV-ΔM51 in combination with rapamycin, the inhibitor of mTOR, selectively killed tumor, but not normal cells and increased the survival of immunocompetent rats bearing malignant gliomas. Also, HDAC inhibitors influence epigenetic changes within cells and can alter gene expression affecting antiviral responses. Using VSV-ΔM51 in combination with MS-275 or SAHA reversibly compromised host antiviral responses and enhanced spread of VSV in various cancer types, with no detection of infected normal tissues (Nguyên TL, 2008; Shestakova et al., 2001; Shulak et al., 2014). Interestingly, in our study neither rapamycin nor HDAC inhibitors were able to overcome resistance of PDAC cells to VSV, suggesting that different mechanisms are responsible for ISG expression in PDAC cells compared to other cancer types.

Our future in vivo experiments will address the efficacy and safety of combined treatments of VSV with TPCA-1 or ruxolitinib. Base on our in vitro data, it is likely that the inhibitor would need to be co-administered with the OV and removed after a finite period of time to prevent spread of the virus to normal tissues. JAK 1/2 inhibitors are known to reversibly bind to the ATP-binding site of JAK1 and JAK2 to prevent activation of JAKs and proteins in the JAK/STAT signaling pathway (Mascarenhas et al., 2012; Thompson et al., 2002). Our data are consistent with the reversibility of treatment using JAK inhibitors, since once removed before infection does not enhance VSV replication. Importantly, similar reversible effect was also demonstrated for TPCA-1 in this study.