Voltage dependent anion channel-1 regulates death receptor mediated apoptosis by enabling cleavage of caspase-8
© Chacko et al. 2010
Received: 20 May 2010
Accepted: 20 July 2010
Published: 20 July 2010
Skip to main content
© Chacko et al. 2010
Received: 20 May 2010
Accepted: 20 July 2010
Published: 20 July 2010
Activation of the extrinsic apoptosis pathway by tumour necrosis factor related apoptosis inducing ligand (TRAIL) is a novel therapeutic strategy for treating cancer that is currently under clinical evaluation. Identification of molecular biomarkers of resistance is likely to play an important role in predicting clinical anti tumour activity. The involvement of the mitochondrial type 1 voltage dependent anion channel (VDAC1) in regulating apoptosis has been highly debated. To date, a functional role in regulating the extrinsic apoptosis pathway has not been formally excluded.
We carried out stable and transient RNAi knockdowns of VDAC1 in non-small cell lung cancer cells, and stimulated the extrinsic apoptotic pathway principally by incubating cells with the death ligand TRAIL. We used in-vitro apoptotic and cell viability assays, as well as western blot for markers of apoptosis, to demonstrate that TRAIL-induced toxicity is VDAC1 dependant. Confocal microscopy and mitochondrial fractionation were used to determine the importance of mitochondria for caspase-8 activation.
Here we show that either stable or transient knockdown of VDAC1 is sufficient to antagonize TRAIL mediated apoptosis in non-small cell lung cancer (NSCLC) cells. Specifically, VDAC1 is required for processing of procaspase-8 to its fully active p18 form at the mitochondria. Loss of VDAC1 does not alter mitochondrial sensitivity to exogenous caspase-8-cleaved BID induced mitochondrial depolarization, even though VDAC1 expression is essential for TRAIL dependent activation of the intrinsic apoptosis pathway. Furthermore, expression of exogenous VDAC1 restores the apoptotic response to TRAIL in cells in which endogenous VDAC1 has been selectively silenced.
Expression of VDAC1 is required for full processing and activation of caspase-8 and supports a role for mitochondria in regulating apoptosis signaling via the death receptor pathway.
The voltage dependent anion channel 1 (VDAC1) is a conserved beta barreled pore forming protein integral to the outer mitochondrial membrane where it regulates ATP/ADP exchange and respiratory control . The functional role of the VDAC proteins VDAC1, 2, and 3 in the regulation of apoptosis remains controversial. Different VDAC proteins exhibit distinct apoptosis regulating functions as evidenced by the antagonism of BAK-induced apoptosis by VDAC2 . A pro-apoptotic role for VDAC1 has been implicated in some cell death models. For example, knockdown of VDAC1 has been reported to abrogate BAX activation and apoptosis following cisplatin treatment in NSCLC cells . VDAC1 has been shown to be required for endostatin-induced endothelial cell apoptosis . Knockout of all three VDAC isoforms does not indicate a direct role in regulating calcium or BID induced mitochondrial apoptosis . However, hexokinase II binds VDAC1 and this interaction has been implicated in regulating cell survival downstream of AKT , and blocking ion transport through VDAC1 following toxic insult has been shown to reduce subsequent apoptosis .
The potential involvement of VDAC1 in regulating death receptor mediated apoptosis has not been determined. The extrinsic death pathway involves binding of ligands such as TRAIL  or FAS  to receptors of the Tumour Necrosis Factor Receptor family. This results in recruitment and activation of initiator caspase-8 at the death-inducing signaling complex (DISC), resulting in cleavage of the 53/55 kDa procaspase to catalytically active p43 and p18 forms . Cleaved caspase-8 then directly activates the executioner caspases 3 and 7 , and the mitochondrial apoptosis pathway through cleavage of the 23 kDa BID protein to its truncated form tBID, promoting oligomerisation of BAX and BAK [12, 13]. Mitochondrial cardiolipin has been proposed to regulate translocation and activation of caspase-8, implicating this organelle in extrinsic death pathway regulation [14, 15], and a number of studies have shown that procaspase-8 and p18-caspase-8 localise to the mitochondria [16–18]. Non-small cell lung cancer (NSCLC) cells express relatively high levels of procaspase-8 and are sensitive to induction of apoptosis by TRAIL compared with normal cells both in vitro and in vivo [19, 20]. Accordingly, there is interest in the potential clinical application of TRAIL, and TRAIL receptor agonists in NSCLC and other tumour types [20, 21]. In this study we show that VDAC1 is necessary for full caspase-8 activation and apoptosis following activation of death receptors by TRAIL, FAS or FLIP siRNA knockdown in NSCLC cells, implicating a novel functional role for mitochondria in regulating death ligand induced apoptosis.
To confirm whether the VDAC1 requirement for caspase-8 activation could be generalised to other cells, VDAC1 was silenced in the MOR NSCLC cell line. Following 10 ng/ml TRAIL, strong accumulation of caspase-8-p18 was observed in NT siRNA treated cells, but this was abolished by knockdown of VDAC1 (Figure 2D). A similar effect was observed in SKMES (Figure 2E), although only p43-caspase-8 was readily detectable in these cells. These results indicate that TRAIL-induced caspase-8 cleavage and apoptosis are VDAC1-dependent.
Silencing of c-FLIP has been shown to induce caspase-8 activation in a number of NSCLC cell lines, including H460 . Accordingly, the role of VDAC1 in regulating apoptosis induced by cFLIP knockdown was also examined. FLIP silencing by siRNA induced accumulation of caspase-8-p18 in sh-NT control cells, but this was not observed in sh-VDAC1 cells (Figure. 5C).
To determine whether re-expression of exogenous VDAC1 could restore sensitivity to TRAIL in cells in which endogenous VDAC1 had been selectively silenced by siRNA targeting the 3' untranslated region (3'UTR) of VDAC1, H460 clones stably expressing the full length coding region of VDAC1 fused to Myc-tag were generated. Clones 5A and 5B exhibited moderate increases in overall expression of VDAC1 compared to empty vector control cells (Figure 5E). Cells stably expressing empty vector (EV) or VDAC1-myc were treated with NT siRNA or a VDAC1 3'UTR siRNA. Following treatment with TRAIL, VDAC1-myc transfected cells exhibited increased activation of caspase-3 compared to control cells (Figure 5D). Control cells also showed a lack of TRAIL induced caspase-8 processing to p18 (Figure 5F). In VDAC1-myc overexpressing cells, however, PARP processing occurred to a greater degree compared to control (Figure 5G), and was unaffected by silencing of endogenous VDAC1.
The role of VDAC1 in regulating apoptosis has been the subject of considerable debate. Knockout of all three isoforms of VDAC was shown to have no effect on mitochondrial apoptosis in mouse embryonic fibrosis , whereas conflicting data has indicated that the N-terminal of VDAC1 is essential for release of cytochrome C following various apoptotic stimuli . VDAC2 has been shown to inhibit BAK activation in a similar manner to a BCL2 family protein ; however recent data suggests that VDAC2 also has a proapoptotic role associated with regulation of BAK and BAX [23, 24]. VDAC1 promotes aerobic glycolysis in cell lines through its interaction with hexokinase at the outer mitochondrial membrane [25, 26]. Other studies have shown that hexokinase binds to mitochondria via the N-term of VDAC1 and that this is associated with resistance to mitochondrial apoptosis [27, 28]. These reported interactions between VDAC1 and hexokinase imply a pro-survival rather than pro-apoptotic role for VDAC1 activity in cancerous cells.
VDAC1 interacts directly with, and may be functionally regulated by the anti-apoptotic BCL-2 family protein BCL-XL [1, 29–32] or the proapoptotic BH3-only proteins BAD  and BID . The mitochondrion has been previously implicated in BAX/BAK independent regulation of extrinsic death pathway signaling. BCL-XL can inhibit cleavage of caspase-8 at the mitochondrial surface following treatment with FAS ligand . Confocal microscopy showed that caspase-8 partially colocalised with a mitochondrial marker in H460 cells, and western blot indicated that pro-caspase-8 was divided between mitochondrial and cytosolic fractions. This is consistent with previous studies which have demonstrated pro-caspase-8 and p18-caspase-8 localization to the mitochondria [16–18]. We observed that TRAIL induced cleavage of caspase-8 to the active p18 protein, which was detected predominantly in the mitochondrial compartment, and that this was dependent upon expression of VDAC1 in various NSCLC cell lines.
This study is the first to implicate VDAC1 in the regulation of extrinsic pathway mediated apoptosis. Using both stable and transient knockdown of VDAC1 we have shown that death receptor dependent apoptosis relies on the expression of this outer mitochondrial membrane protein for efficient processing of caspase-8. VDAC1 did not appear to directly affect BAX/BAK dependent activation by tBID, in agreement with a recent finding that mitochondria from VDAC1(-/-) MEFs remained sensitive to tBID . Accordingly, the mitochondrial VDAC1 directly regulates the extrinsic apoptotic pathway via caspase-8 and indirectly regulates the mitochondrial death pathway via BID processing, implicating a previously unknown proapoptotic, and physiological function for VDAC1.
Although our findings suggest that downstream of death ligand binding, VDAC1 acts as a facilitator of caspase-8 processing, it is not known whether VDAC1 requires cooperation with additional proteins in the outer mitochondrial membrane or is sufficient to activate caspase-8. Recent evidence has shown that caspase-8 localisation occurs at contact sites where cardiolipin, a predominantly inner mitochondrial membrane phospholipid, is exposed at the mitochondrial surface; however, the binding interactions of caspase-8 at the mitochondria remain undefined . VDAC1 is also present at contact sites on the outer mitochondrial surface raising the possibility that loss of VDAC1 could in some way alter contact site architecture and the cardiolipin platform for caspase-8 docking. One argument against this from our own and previous data however, is that significant levels of procaspase-8 are present in the mitochondrial outer membrane in the absence of TRAIL. It is therefore unclear by what molecular mechanism VDAC1 may regulate caspase-8 cleavage. Caspase-8 processing to p18 may occur at the mitochondria in a VDAC1-dependent manner leading to a signal amplification loop following binding of death ligand and DISC assembly. Such an amplification loop could be reliant on active caspase-3, which accumulates in the mitochondria following etoposide-induced cell death .
The structural requirements for VDAC1 regulation in regulating caspase-8 activation are unknown. The E3 ligase Cullin-3 polyubiquitinates caspase-8 prior to its cleavage at the DISC following death receptor stimulation , and VDAC1 could play a regulatory role in this process. We cannot rule out the possibility that VDAC1 may promote caspase-8 cleavage by associating with DISC components at the plasma membrane; redox activity of VDAC1 at the plasma membrane has been shown to regulate apoptosis [7, 36]. Further studies are required to delineate the domains and protein interactions which are essential for VDAC1 to regulate caspase-8 activation; such studies are ongoing and will investigate whether this function is specific to VDAC1 alone, or shared by other VDAC isoforms due to local homology in critical domains.
In summary, we have shown that VDAC1 is required for activation of caspase-8 following stimulation of the extrinsic apoptosis pathway in a panel of NSCLC cells. TRAIL is currently being investigated in the clinical setting as a potential anticancer therapy; trials are ongoing in NSCLC [20, 21]. Our findings suggest that expression levels of VDAC1 could be a factor predictive for tumour sensitivity to therapeutic TRAIL receptor agonists in a subset of cancers.
ATCC-NCI-H460 Non-Small Cell Lung Cancer Cells, and MOR cells, were grown in RPMI 1640 medium (PAA) with 10% fetal calf serum and penicillin/streptomycin at 37°C with 5% CO2. SKMES cells were grown in EMEM media. Plasmids encoding shRNA sequences to VDAC1 and Non-Targeting (NT) shRNA (SA Biosciences) were transfected into H460 NSCLC cells using FuGene6 reagent (Roche) according to manufacturer's instructions. Plasmids bore G418 antibiotic resistance selection marker and transfected cells in 12-well plate were incubated with 0.4 mg/ml G418 until discrete colonies visible to the naked eye could be selected for transferral to a fresh plate. Three rounds of selection were preformed and clones which grew steadily in the presence of antibiotic were tested for knockdown of protein of interest by western blot. Cells were cultured continuously in the presence of appropriate antibiotics, but antibiotics were removed when plating for drug treatment.
Cells were transfected with 50 nM of VDAC1 on-target siRNA smartpool (Dharmacon), VDAC1_11 siRNA sequence (Qiagen) or Non-Targeting siRNA using Dharmafect-1 reagent. Following transfection, transfection media was replaced with RPMI 1640 media, cells incubated for 24 h, trypsinized and seeded for experiments described in text (5000 cells/well for 96-well plate). Endpoint was taken 24 h post TRAIL treatment and 48 h post siRNA transfection.
Cells typically 90% confluent were detached with a cell scraper into culture media, pelleted by centrifugation to include non-adherent cells, if any, and washed three times in Mitochondrial Isolation Buffer (200 mM Mannitol, 70 mM Sucrose, 1 mM EGTA, 10 mM HEPES, 0.5 mg/ml BSA, pH7.4). Mitochondria were then isolated by Dounce homogenization followed by centrifugation at 800 g for 10 minutes to remove debris and heavy membranes, then centrifugation at 10,000 g for 10 minutes to separate mitochondria from cytosolic fraction. For measurement of mitochondrial polarization, approximately 1 mg mitochondria were resuspended in 800 μl of respiration buffer (125 mM KCl, 5 mM HEPES, 1 mM EGTA, 1 mM KHP04, 2.5 mM MgCl2, 0.4% BSA, pH 7.4) with 5 μM Rotenone. Mitochondria were then loaded with JC-1 (10 nM) (Invitrogen), incubated for 5 minutes at 37°C, and washed and centrifuged 3 times in 800 μl MRB. 50 μl of mitochondria solution was applied to each well of a 96 well plate, with final volume made to 100 μl by applying treatment as described in text. Mitochondria were incubated at 37°C and JC-1 fluorescence measured at 530 nm (green) and 580 nm (red) with excitation at 480 nm. Results were summarized as mean ± standard deviation.
25 μg of cell lysate or mitochondrial fractions were loaded into a 12% SDS-PAGE gel. Proteins were separated by electrophoresis and transferred to PVDF membrane for blocking and incubation in 5% milk with primary and HRP-conjugated secondary antibodies. Images were developed with ECL-plus (GE healthcare). Alternatively, fluorescently tagged secondary antibodies were used for analysis of Western blot on Licor Odyssey Imaging system. Primary antibodies used were; anti-VDAC1 (Abcam ab14734), anti-PARP (E-Bioscience 14-6666-92), anti-Caspase-8 (Alexis 804-242) anti-BAX (Cell Signaling #2774), anti-COX4 (Cell Signaling #4850), anti-BID (Cell Signaling #2002), anti α-tubulin (Abcam ab15246). Secondary antibodies used were; goat anti-mouse HRP (DAKO P0447), goat anti-rabbit HRP (DAKO P0448), anti-mouse Dy-Lite800 (Rockland 610-145-121), anti-rabbit Dy-Lite680 (Rockland 611-144-122).
Cells were seeded at 5000/well into 96-well plates, (minimum n = 5). After 24 h cells were treated with 10 ng/ml recombinant human TRAIL (Calbiochem) or vehicle control to final volume of 150 μl. Following 24 h treatment, luminescent ATPase assay (Lonza) was carried out according to manufacturer's instructions and 1 second luminescence readings taken on a Berthold Tristar plate reader. p-values were generated using Student's t-test.
For DEVDase activity assays, Casp3/7-glo reagent (Promega) was mixed 1:1 with RPMI and 50 μl applied to cells after removal of culture media. Plates were incubated at 21°C for 30 minutes and luminescence read as before.
Cells were seeded at a density of 1 × 105cells/well in 6-well plates. Following treatment, DNA content was evaluated by propidium iodide (PI, Sigma) staining of cells by re-suspending them in 360 μL 0.1%FCS, 10 μg/mL PI, 0.25 mg/mL RNaseA in 1 xPBS. Samples were incubated for 30 minutes prior to analysis on a FacsScalibur Flow Cytometer (Becton Dickinson), in the FL-2 channel, to determine percentage of cells with DNA content <2N. For cell surface expression of receptor proteins, cells were seeded as before, harvested, and rinsed twice in PBS/0.2% BSA/0.2% Sodium Azide. Cells were resuspended in PBS/BSA/sodium azide and incubated at 4°C for 60 minutes with PE-conjugated antibodies to DR4, DR5, or isotype control antibody. Staining intensity was measured on flow cytometer in FL-2 channel.
Cells were seeded onto a glass coverslip inside a 6-well plate and fixed in 2% paraformaldehyde. Coverslips were incubated in PBS/0.1% Triton-X with 1:1000 dilutions of anti-Caspase-8 and anti-COX-4 antibodies, for 1 h at 37°C with agitation, each followed by secondary antibodies; goat anti-mouse AlexaFluor488 and goat anti rabbit AlexaFluor594 (Invitrogen). Slides were mounted with ProLong gold mountant (Invitrogen) and imaged on Leica SP5 confocal microscope.
DAF is a recipient of a Cancer Research UK Clinician Scientist Fellowship. IP is a recipient of a NI R&D research fellowship. AC was supported by Cancer Research UK.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.