Vaccinia virus induces programmed necrosis in ovarian cancer cells

Abstract

The mechanisms by which oncolytic vaccinia virus induces tumor cell death are poorly understood. We have evaluated cell death pathways following infection of ovarian cancer cells with both wild-type and thymidine kinase-deleted (dTK) Lister strain vaccinia. We show that death does not rely upon classical apoptosis despite the appearances of some limited apoptotic features, including phosphatidylserine externalization and appearance of sub-G1 DNA populations. Vaccinia infection induces marked lipidation of LC3 proteins, but there is no general activation of the autophagic process and cell death does not rely upon autophagy induction. We show that vaccinia induces necrotic morphology on transmission electron microscopy, accompanied by marked by reductions in intracellular adenosine triphosphate, altered mitochondrial metabolism, and release of high mobility group box 1 (HMGB1) protein. This necrotic cell death appears regulated, as infection induces formation of a receptor interacting protein (RIP1)/caspase-8 complex. In addition, pharmacological inhibition of both RIP1 and substrates downstream of RIP1, including MLKL, significantly attenuate cell death. Blockade of TNF-α, however, does not alter virus efficacy, suggesting that necrosis does not result from autocrine cytokine release. Overall, these results show that, in ovarian cancer cells, vaccinia virus causes necrotic cell death that is mediated through a programmed series of events.

Figure 1

Activity of Lister-dTK vaccinia in ovarian cancer. (a) Deletion of the TK gene in Lister-dTK was confirmed by PCR (left). Cytotoxicity of vaccinia in human (A2780, A2780CP, IGROV1, SKOV3ip1, OVCAR-4, and TOV21G) and murine (MOSEC, MOVCAR7) ovarian cell lines, as well as hTERT-immortalized human ovarian surface epithelial cells (IOSE25) as determined by MTS assay 72 hours postinfection. Mean ± SD IC₅₀ is shown (left). All experiments were performed in triplicate at least three times. (b) Survival of five human ovarian cancer cells 72 hours following infection with Lister-WT and Lister-dTK at MOI 1 (A2780, A2780CP, IGROV1, and SKOV3ip1) and 10 (TOV21G). (c) Ratio of IC₅₀ (mean ± SEM) for Lister-dTK compared with Lister-wt. All experiments were performed in triplicate at least twice. (d) Replication of vaccinia in ovarian cancer cell lines. Cells were infected with Lister-dTK (MOI 1 pfu/cell) and titer determined by TCID₅₀ assay on CV1 cells as described in “Materials and Methods.” Points represent mean titer ± SD, n = 3. (e) Infection and replication of vaccinia virus in nude mice bearing advanced intraperitoneal SKOV3ip1 tumors following a single ip dose of 10⁸ pfu Lister-dTK. Sequential 5 µm sections were stained with hematoxylin & eosin or for vaccinia virus proteins. L, liver, T, tumor, N, necrotic area. Bars represent 500 µm (image i) and 200 µm (images ii, iii and iv). Panels i and ii show an intraperitoneal deposit growing on the surface of the small bowel serosa adjacent to the omentum. hTERT, human telomerase reverse transcriptase.

Figure 2

Vaccinia does not kill ovarian cancer cells using caspase-dependent apoptosis. (a) Externalization of phosphatidylserine on the cell membrane 72 hours after infection with Lister-dTK (MOI 1). Cells were stained with DAPI and annexin V. The percentage viable cells staining positive for annexin V in the lower right quadrant (Q4) are presented. Bar graph shows mean ± SD of three experiments in duplicate. *P = 0.04, **P = 0.005, ***P ≤ 0.0001 for all. (b) DNA fragmentation 96 hours after infection with Lister-dTK (MOI 1). Cells were fixed and stained with propidium iodide. Representative plots from a single sample show percentage gated cells with sub-G1 DNA. Bar graph shows mean ± SD sub-G1 DNA from 2 to 4 experiments, in duplicate. *P = 0.014 (A2780CP) and 0.01 (IGROV1), **P = 0.0055 (IGROV1) and 0.002 (SKOV3ip1), ***P < 0.001 (both). (c) Sensitivity of OVCAR-4 (O4) and Bcl2 over-expressing cells (O4-Bcl2) to Lister-dTK. Cell survival was measured by MTS assay. Dose-response curve shown from one experiment, mean IC₅₀ ± SD is from three triplicate experiments. Bcl2 overexpression, confirmed by western blot, significantly inhibits cell death after exposure to 10 µmol/l cisplatin for 72 hours. Mean ± SD from three experiments, each in triplicate. ***P < 0.0001. (d) Cleavage of PARP and caspase-3 in untreated cells (U) and cells infected with Lister-wt (WT) or Lister-dTK (dTK; both MOI 1) for 72 hours. A2780 cells treated with 2 µmol/l staurosporine overnight were used as a positive control for apoptosis (+). Blots are representative of three independent experiments. (e) Caspase inhibition with zVAD-fmk (10 µmol/l added daily) in Lister-dTK (MOI 1) infection. Cell survival was measured by MTT assay 96 hours postinfection. In combination with 10 µmol/l cisplatin, a single dose of 10 µmol/l zVAD-fmk was used and cell survival determined by MTT assay after 48 hours. Mean ± SD from three experiments, each in triplicate. **P = 0.0055 (A2780CP) and 0.0011 (IGROV1).

Figure 3

Vaccinia does not require autophagy for cytotoxicity. (a) LC3B localization in IGROV1 (i–iii) and A2780CP (iv) cells. Cells were co-infected with Ad GFP-LC3 (MOI 30) and Lister-dTK (MOI 1) and imaged 72 hours postinfection. (iii) is a magnified image from (ii). (b) Appearance of LC3B-II following infection with Lister-wt (WT) or Lister-dTK (dTK; both MOI 1). Protein was extracted 72 hours postinfection and subjected to SDS-electrophoresis. Representative blot of three experiments is shown. Quantification (below) of three independent experiments shows the mean ± SEM LC3B-II:LC3B-I relative expression. (c) Upper blot: accumulation of LC3B-II in A2780CP cells following 24 hours of serum starvation (lane 2, HBSS treatment) or treatment with the rapamycin. Lower blot: accumulation of LC3B-II in A2780 cells following 24 hours treatment with both rapamycin and chloroquine. (d) Lysosomal turnover of LC3B-II. Cells were infected with Lister-dTK (MOI 1) in the presence or absence of lysosomal protease inhibitors (LPI) pepstatin A (10 µg/ml) and E64d (10 µg/ml) for 72 hours. Protein was subjected to SDS-electrophoresis. Bar graph (right) shows quantification of LC3B-II:LC3B-I relative expression. (e) IC₅₀ of Lister-dTK in the presence of the autophagy inhibitor 3-methyladenine relative to the IC₅₀ of Lister-dTK alone. Cell survival was determined by MTS assay and represents the mean ± SD relative IC₅₀. Lines present linear regression of the four IC₅₀ values—the line slope does not differ significantly from 0 in any of the analyses. HBSS, Hanks buffered salt solution.

Figure 4

Vaccinia infection induces necrotic features in ovarian cancer cells. (a) Electron microscopy. A2780 cells were mock infected (i–ii) or infected with Lister-dTK (MOI 10) for 72 hours (iii–viii). Infected cells with cytoplasmic vacuoles (V) flank an already dead cell (iii and iv). Nuclear and cytoplasmic membrane rupture (v). Dense mitochondria (M) in an infected cell (vi). A cell full of infectious virions (electron dense structures) displays a swollen nucleus and ruptured cell membrane (arrowed) (vii). Cytoplasmic vacuoles and mitochondria in a cell harboring both mature and immature virions (viii). (b) Release of HMGB1 from infected cells. Supernatant and cell lysates were collected 72 hours following infection with Lister-dTK (MOI 10). (c) Intracellular ATP levels following infection with Lister-dTK (MOI 10). ATP concentration was quantified using a luciferase based ATP assay. Data show mean ± SD, n = 6. ***P < 0.001 for all analyses. (d) The relative rate of glycolysis and oxidative metabolism in A2780CP cells following Lister-dTK infection (MOI 1). Extracellular acidification rate (ECAR) was used to determine glycolysis and oxygen consumption rate (OCR) to quantify mitochondrial respiration. Rates are displayed relative to control (uninfected) cells. Cells were treated with TNF-α (10 ng/ml) and zVAD-fmk (25 µmol/l) for 4 hours as positive control. Bars represent mean ± SEM, n = 7. *P = 0.049 (OCR, Lister-dTK) and 0.04 (ECAR, TNF/zVAD), ***P < 0.0001 (ECAR, Lister-dTK) and 0.0002 (OCR, TNF/zVAD).

Figure 5

Necrosis mediators following vaccinia infection. (a) Formation of a RIP1/RIP3/caspase 8 complex following infection with Lister-dTK (MOI 10) for 96 hours. Caspase-8 was immunoprecipitated and the presence of RIP1, RIP3, and caspase-8 detected by immunoblotting. Cells were treated with 100 µmol/l etoposide or 10 ng/ml TNF-α in the presence of zVAD-fmk (25 µmol/l) as positive controls for necrosis. (b) Expression of RIP1, RIP3, and caspase-8 was assessed following infection with Lister-dTK (MOI 10) for 72 hours. The proteasome inhibitor MG132 (10 µmol/l) was added 5 hours before cell lysates were collected, and p21 expression assessed as positive control for MG132 activity.

Figure 6

TNF-α blockade, necrostatin, and RNAi knockdown of RIP1 and RIP3. (a) IGROV1 cells with stable shRNA-mediated TNF-α knockdown (IGROV1-shTNF) and control cells encoding scrambled shRNA (IGROV1-shControl) were infected with Lister-dTK in triplicate. Cell survival was assessed 72 hours postinfection (left). Experiments were repeated twice—representative data are presented. TNF-α expression was quantified by quantitative RT-PCR, normalized to 18S. Bars represent mean ± SD, n = 3. *P < 0.05. (b) A2780, SKOV3ip1, and TOV21G were transfected with 10 nmol/l RIP3 siRNA (R3) or nontargeting siRNA (NT). Protein lysates were harvested 24 and 96 hours post-transfection and levels of RIP3 determined by immunoblotting. Cells were infected with Lister-dTK (MOI 10) at 24 hours after siRNA transfection and cell survival determined by MTT assay 72 hours postinfection. (c) Effect of inhibiting RIP1 kinase activity on Lister-dTK induced cell death. Cells were treated with necrostatin-1 (100 µmol/l) ± zVAD-fmk (25 µmol/l) immediately following infection with Lister-dTK and cell survival determined by MTS assay 72 hours postinfection. Representative dose response curves are shown. Mean IC₅₀ ± SEM of a minimum of three experiments (right). (d) A2780 and SKOV3ip1 cells were transfected with 10 nmol/l RIP1 siRNA (R1) or nontargeting siRNA (NT) and treated as in (b). *P < 0.05, **P < 0.005.

Figure 7

Necrosulfonamide and virus replication after necrosis inhibition. (a) Effect of inhibiting MLKL on Lister-dTK induced cell death. Cells were treated with necrosulfonamide (MLKL inhibitor) following infection with virus, and cell survival determined by MTT assay 72 hours postinfection. (b) A2780CP and TOV21G cells were infected with Lister-dTK at MOI 100. Infectivity was assessed by RFP flow cytometry 12 hours postinfection (left). Cell survival was assessed 24 hours postinfection in the presence and absence of 100 µmol/l necrostatin-1 or 10 µmol/l necrosulfonamide.