Background
Oncolytic virotherapy, which takes advantage of selective viral infection and apoptosis in cancer cells due to robust viral replication, is emerging as an important alternative to conventional cancer treatment modalities [
1‐
4]. Evidence indicates that concurrent use of a repertoire of different oncolytic viruses (with different modes of action) may produce more efficacious therapeutic response. Human respiratory syncytial virus (RSV) is a respiratory tract-specific enveloped non-segmented negative sense single stranded RNA (NNS) virus of the paramyxovirus family [
5,
6]. We have recently identified RSV as an oncolytic virus by demonstrating that RSV can cause apoptosis of PC-3 human prostate cancer cells in culture and in a xenograft tumor environment as a consequence of excessive viral replication in PC-3 cells that led to apoptotic cell death [
7]. Specificity of the virus-induced oncolysis of cancer cells was evident from the lack of significant viral burden and apoptosis of non-tumorigenic human prostate epithelial cells, such as RWPE-1.
Metastatic prostate cancer is a leading cause of cancer deaths in men in the United States. The steroid hormone androgen is a potent mitogen for normal and cancerous prostate epithelial cells. The cognate androgen receptor (AR) mediates nuclear responses to androgen signaling [
8,
9]. Although initially androgen-sensitive, metastatic prostate cancer evolves to a state of androgen independence for growth and proliferation, despite continued expression of AR at all stages of the disease. AR was shown to activate the transcriptional program of a distinct set of gene networks, including many genes involved in cell cycle regulation, during progression of the cancer cells to androgen independence [
9]. As noted above, RSV can induce oncolysis of androgen-independent PC-3 prostate cancer cells and RSV caused regression of PC-3 cell derived xenograft tumors in immune-deficient nude mice [
7]. Extending this study, we examined susceptibility of the androgen-sensitive LNCaP human prostate cancer cells and LNCaP xenograft tumors to RSV-induced oncolysis, and the impact of host immune-competence on the oncolytic activity of RSV. Innate immunity is the first line of antiviral defense for restricting virus growth and spread. Since both NF-κB and type-I interferon (IFNα/β)-mediated JAK/STAT signaling [
10] is required for innate antiviral response; we also examined the activation status of NF-κB and IFN-induced JAK-STAT pathways in RSV-infected prostate cancer cells.
Herein we report that RSV is potently oncolytic against androgen-sensitive LNCaP human prostate cancer cells in vitro and in vivo, and aberrant IFN-regulated signaling accounts for LNCaP cell susceptibility to RSV. While RWPE-1 non-tumorigenic prostate cells were protected against RSV infection by activation of JAK-STAT and NF-κB signaling, a lack of sustained NF-κB activation was associated with the susceptibility of PC-3 androgen-independent prostate cancer cells to RSV-induced oncolysis, although IFN-mediated signaling was functional in these cells.
Methods
Virus, Cells, Culture Conditions, Chemicals
RSV (A2 strain, ATCC, Manassas, VA) was propagated in CV-1 cells. Viral titer was monitored by plaque assay [
7]. LNCaP, PC-3 and RWPE-1 (source: ATCC) were cultured as per the supplier's instructions. RM1 murine prostate cancer cells were cultured in DMEM [
11,
12]. RM1 cells (a gift from Dr. Timothy Thompson, Baylor College of Medicine, TX) originated from stable expression of Ha-Ras and c-Myc oncogenes in mouse prostate epithelial cells [
11,
12]. RM1 cells express androgen receptor but they are androgen-independent. Interferon-α (IFN-α) and Tumor necrosis factor-α (TNF-α) were purchased from R&D Systems (Minneapolis, MN).
RSV infection, interferon (IFN) treatment, ELISA for IFN-β, IFN-γ and IL-10
RSV (1 MOI) was added to cells for adsorption at 37°C for 1.5 h. After washing, infection continued for additional 0 h-48 h. At various time points post-infection, virus yields in culture supernatants were assayed by plaque assay of the monolayer of CV-1 cells (African green monkey kidney cells) covered with a nutrient medium in methyl-cellulose [
7]. Crystal violet staining of living cells allowed clear visualization of the plaques. Cell morphology was visualized microscopically. In some experiments, cells were pre-treated with 1000 units/ml IFN-α for 16 h, followed by infection with RSV for 24 h in the presence of IFN. Medium supernatants were used to measure viral titer (plaque assay).
Medium supernatants from RSV-infected cells at 16 h post-infection were analyzed for IFN-β levels using a human IFN-β specific ELISA kit (PBL Interferon Source, NJ). In some experiments, spleen homogenate and tumor homogenate obtained from mice were analyzed for IFN-γ and IL-10 levels using FACS based cytokine bead array (BD Biosciences, San Jose, CA).
Apoptosis, cell viability
Cells infected with RSV for 24 h-36 h, were examined for apoptosis by annexinV labeling, using annexin V/propidium iodide apoptosis detection kit (BioVision, CA) and for cell viability using MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay [
13].
Xenograft prostate tumors in nude mice and allograft tumors in syngeneic C57BL/6J mice
4-6-week-old athymic nude male mice (Harlan Laboratories Inc.) were injected subcutaneously with 3 × 10
6 LNCaP cells at a site below the ear [
7]. When tumors reached palpable size (sizes ranging from 100-300 mm
3 for individual mice), RSV (at 1 × 10
6 pfu per animal) or Opti-MEM (Medium, carrier control) was injected repeatedly into the tumor mass (intratumoral administration) at 2-day intervals over a two-week period. Tumor volumes were measured by a caliper (4/3 × 3.14 × r
1
2 × r
2 with r
1 < r
2) and normalized to the tumor volume for the corresponding mouse at day-1 when the first injection was administered. Tumor samples from euthanized mice were excised and processed for histology and for TUNEL assay to evaluate apoptosis.
For imaging of subcutaneous xenograft tumors, which were generated using luciferase-expressing LNCaP-Luc-2 cells, palpable tumors were injected with RSV (1 × 106 pfu per animal via the intratumoral (I.T.) route) or Opti-MEM (Medium, carrier control). At various days post-infection, luciferin was injected and real-time tumor bioluminescence (reflecting tumor growth) was monitored in live animals using the Xenogen IVIS imaging system (Caliper Life Sciences, Hopkinton, MA).
To analyze prostate tumors in syngeneic C57BL/6J mice, 4-6 week-old mice were injected subcutaneously with RM1 murine prostate cancer cells (1 × 106 cells) in the right dorsal flank. Tumors (tumor volumes ranging from 56-164 mm3 for individual mice) were injected intra-tumorally with either RSV (at 1 × 106 pfu, each animal) or Opti-MEM. The tumor volume for each animal was normalized to the corresponding volume recorded just prior to the first injection.
Housing and all procedures involving animals were performed according to protocols approved by the Institutional Committee for Animal Care and Use of the University of Texas Health Science Center San Antonio.
Electrophoretic gel mobility shift (EMSA)
RSV-infected (0-24 h) or mock-infected cells were processed for nuclear fraction enriched total cell extracts [
14]. In some cases, cells were treated with IFN-α (1000 units/ml, 4 h) or TNF-α (30 ng/m, 2 h). Nuclear extracts were incubated with
32P-labeled oligonucleotide duplex for cis-elements of NF-κB (IL6 promoter) or STAT-1 (hSIE/m67, synthetic c-sis-inducible element). Protein-DNA complex was analyzed as before [
15]. NFκB element: 5'-GGGAATTCCCCATCTACGCTA; STAT-1 element: 5'-GTCGACA-TTTCCCGTAAATC.
Histopathology and TUNEL
Part of tumor tissues (excised immediately after euthanasia) was fixed in 10% neutral buffered formalin, embedded in paraffin, sectioned at 5 micron and analyzed for histology after staining with hematoxylin and eosin. TUNEL staining in situ was performed using DeadEnd Colorimetric kit (Promega, WI). For each tumor sample, pathologic changes were graded by a blinded board certified veterinary pathologist (GBH). Grading criteria were based on invasion of adjacent normal tissue, necrosis, mitotic figures per 20X high power field, edema, congestion, compression of the surrounding tissue, mineralization, inflammation (granulocytes, lymphocytes, mixed), hemorrhage, neovascularization and hemosiderin deposition.
Statistics
Differential tumor growth was determined utilizing STATA Analysis and Statistical Software (College Station, Texas, USA). Results are mean values ± Standard Error (SE) of the mean. For each mouse, the tumor volume at every time point was normalized to its tumor volume at the start of RSV (or Medium) injection, which was set as 100%. Significance of the difference between Medium- and RSV-injected tumor growth trajectories was evaluated using linear mixed model that incorporated a random intercept for each mouse and used normalized mean tumor volume measurements as the dependent variable. Wald tests assessed statistical significance of differences between growth rates in two mouse groups.
Discussion
Our study has been the first to demonstrate RSV is an oncolytic virus, and this oncolytic activity is functional
in vivo both in immune-deficient nude mice and in an immune-competent host environment, since RSV inhibited prostate tumor growth in syngeneic C57BL/6J mice. RSV infectivity and the virus-induced apoptotic index
in vitro were much higher in androgen-dependent LNCaP cells compared to non-tumorigenic RWPE-1 prostate cells. Aberrant type I interferon (IFN)-dependent antiviral defense response [
10], which culminated in impaired activation of the STAT-1 transcription factor (STAT-1 is required for expression of IFN-dependent antiviral genes), associated with the high RSV burden in infected LNCaP cells. We conclude that blockade in STAT-1 activation, leading to inhibition in the expression of critical IFN-dependent antiviral genes, accounts for excessive RSV replication leading to apoptosis of LNCaP cells. This is unlike PC-3 androgen-independent prostate cancer cells for which RSV-induced oncolysis was associated with failure in a sustained NF-κB activation, which would cause failure in the induction of NF-κB dependent antiviral genes.
Using IFN-neutralizing antibody, we also provide the first direct evidence (Figure
9) that protection of non-malignant epithelial cells against virus-induced oncolysis is due to IFN-mediated antiviral defense response. Our results further revealed that the oncolytic function of RSV may remain active even when the IFN-regulated antiviral pathway is functional, provided a second defense arm involving NF-κB signaling is deregulated.
Innate immunity is the first line of defense mounted by the host to combat virus infection before an orchestrated adaptive immune response is launched [
10,
20]. IFN-mediated activation of the JAK/STAT antiviral pathway is recognized as a major antiviral innate immune defense mechanism [
24]. In this regard, we have recently demonstrated that RSV-infected lung epithelial cells and immune cells (e.g. macrophages) utilize Nod2 protein as a molecular sensor to induce production of IFN-α/β from infected cells after interacting with viral single-stranded RNA genome and subsequently triggering innate antiviral response [
25]. IFN-α/β, which are potent antiviral cytokines produced in infected cells, bind to cognate cell surface receptors on uninfected cells (via autocrine/paracrine action) to induce the JAK/STAT antiviral pathway; which helps promote nuclear translocation and activation of the transcription factors STAT-1 and STAT-2 that in turn activate antiviral genes [
24]. We also reported that an IFN-independent innate defense mechanism involving TNF-α -induced activation of NF-κB can restrict virus replication in infected cells due to induction of antiviral genes [
15,
26]. These two antiviral pathways mediated by IFN (via the JAK/STAT pathway) and TNF-α (via the NF-κB pathway) are activated in infected cells either individually or together to coordinate the transcriptional induction of the antiviral gene network.
A large number of cancer cells are deficient in the IFN signaling cascade [
27,
28], making many types of cancer cells susceptible to apoptosis by oncolytic viruses. In the context of prostate cancer, our results suggest that both androgen-dependent prostate cancer cells (such as LNCaP cells) and androgen-independent prostate cancer cells (such as PC-3 and RM1 cells) are susceptible to RSV-induced oncolysis. We show that, while IFN production from infected LNCaP cells was normal, IFN failed to activate STAT-1 in LNCaP cells. In fact, it was reported earlier that LNCaP cells fail to express JAK1 [
29]. On the other hand, RSV infection and IFN treatment of non-tumorogenic RWPE-1 cells and PC-3 cells was associated with robust STAT-1 activation and protection against RSV-induced apoptosis. We also show that while PC-3 cells respond to IFN and induce DNA-binding activity of STAT-1 (in agreement with previous reports that IFN-treated PC-3 cells are activated for antiviral signaling; [
30,
31]), impaired NF-κB activation is associated with apoptosis in RSV-infected PC-3 cells. LNCaP cells, on the other hand, were competent to activate NF-κB in response to RSV infection. We speculate that androgen dependence and/or the androgen receptor expression status of prostate cancer cells may influence RSV-mediated modulation of the innate antiviral apparatus (NF-κB activation vs. IFN-mediated JAK/STAT activation). Our results (Figures
8,
9,
10,
11) [
7] lead us to conclude that deregulation of the IFN pathway in androgen-sensitive LNCaP prostate cancer cells accounts for loss of STAT-1 activation (and non-expression of antiviral factors), higher RSV replication, induction of apoptosis and reduced cell viability, whereas deregulation of the NF-κB-dependent antiviral defense in androgen-independent PC-3 prostate cancer cells accounts for susceptibility of these cells to RSV-induced apoptosis.
Advanced-stage cancer cells, which continue to express the androgen receptor in a majority of tumor specimens, are resistant to apoptosis from androgen ablation or from the cytotoxicity induced by chemotherapeutic agents. Development of treatment protocols that would promote prostate cancer cell apoptosis and prevent cancer cell progression to androgen independence has remained a major challenge in prostate cancer therapy. To this end, it is tempting to speculate that complete ablation of prostate cancer cells at an early stage, when the cells are still androgen-sensitive, is likely to prevent clonal emergence of androgen-independent prostate cancer cells. The anti-tumor, oncolytic activity of RSV against androgen receptor-negative prostate tumors has additional clinical significance, since reduced or non-detectable androgen receptor expression has been observed in a small percent of metastatic neoplastic foci at distant organ sites from castrate resistant prostate cancer patients [
32,
33]. The observation that RSV-induced oncolysis of prostate tumors can occur in immune-competent C57BL/6J mice has obvious clinical significance. It is important to mention that systemic delivery of RSV represent a clinically feasible route for therapy. In that regard, we have previously demonstrated that intraperitoneal (i.p.) injections of RSV are effective in causing regression of PC-3 xenograft tumors [
7] which are more aggressive than LNCaP xenograft tumors with regard to tumor growth. Thus, studies are underway to examine oncolytic efficacy of RSV against LNCaP tumors following systemic administration. The results from our study suggest that RSV-based therapy has the potential to be a viable strategy for prostate cancer treatment.
Acknowledgements
We would like to thank Jesus Segovia for technical assistance. This work was supported in part by the American Lung Association National Biomedical Research grant (RG-49629-N) (S. Bose), NIH grants AI069062 (SB), AI083387 (SB), CA129246 (SB; BC), AG10486 & AG19660 (BC), UTHSCSA-UTSA SALSI Grant (SB; BC), VA Merit Review Grant (BC). The work was also supported by Translational Technology Resources Award (to IE) from UTHSCSA IIMS. AS was supported by NIH NIDCR grant DE14318 for the COSTAR program. BC is a VA Senior Research Career Scientist. FACS at the UTHSCSA Core Facility was supported by the Cancer Center Program grant CA54174 (NIH-P30). Financial assistance was also provided to SB and BC by the Cancer Therapy and Research Center at the University of Texas Health Science Center-San Antonio through the NCI Cancer Center Support Grant (2 P30 CA054174-17).
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
All authors have read and approved the final manuscript. The authors made the following contributions to this work. IE responsible for the study design, experimental work, data evaluation and analysis and helped write the manuscript. TH was involved in experimental design, experimental work and critique. AS performed the interferon assays and viral titer assays. IB was involved in the animal experiments. YK and GBH performed the histology analysis. BC and SB were the research supervisors and participated in the study design, assessment of the results, and writing the manuscript.