KSHV vFLIP is essential for the survival of infected lymphoma cells

Abstract

Primary effusion lymphomas (PELs) associated with infection by the Kaposi's sarcoma-associated herpesvirus (KSHV/HHV-8) have constitutive nuclear factor (NF)-kappaB activity that is essential for their survival, but the source of this activity is unknown. We report that viral FADD-like interleukin-1-beta-converting enzyme [FLICE/caspase 8]-inhibitory protein (FLIP) activates NF-kappaB more potently than cellular FLIP in B cells and that it is largely responsible for NF-kappaB activation in latently infected PEL cells. Elimination of vFLIP production in PEL cells by RNA interference results in significantly decreased NF-kappaB activity, down-regulation of essential NF-kappaB-regulated cellular prosurvival factors, induction of apoptosis, and enhanced sensitivity to external apoptotic stimuli. vFLIP is the first virally encoded gene shown to be essential for the survival of naturally infected tumor cells.

Figure 1.

Correlation of NF-κB activity and vFLIP expression in PEL cell lines. (top) EMSA using a NF-κB–specific oligonucleotide probe, demonstrating constitutive activity in three PEL cell lines. Competition with cold NF-κB and random oligonucleotides has confirmed the specificity of this binding (not depicted). The level of NF-κB activity was in the following order: BC1>BC3>BC2. In all three PEL lines, activity was higher than in BJAB, a virus-negative lymphoma cell line. (middle) Northern blot analysis was performed on polyA-selected RNA extracted from the BC-1, -2, and -3 cell lines, as well as from BC3 after induction of lytic replication by treatment with tetradecanoyl phorbol acetate (TPA) for 48 h. BJAB cells were used as negative control. A probe that recognizes vFLIP and vCYC was used, and filters were reprobed with β-actin to ensure even loading. (bottom) Immunoblot analysis with an antibody to vFLIP showed vFLIP protein in BC1, -2, and -3 cells, but not in KSHV-negative BJAB. Equal protein loading was found in all four lanes with an antibody to β-actin.

Figure 2.

Activation of NF-κB by KSHV-encoded vFLIP. (A) vFLIP induces NF-κB transcriptional activity in human lymphoma cells. Namalwa cells were cotransfected by electroporation with 5 μg NF-κB luciferase reporter plasmid, 0.5 μg of control renilla luciferase reporter (pRL-CMV), and vFLIP expression vector (vFLIP/pcDNA3.1) in the indicated amounts, or empty vector (pcDNA3.1). Total DNA content was maintained at 25 μg with empty vector. In all experiments, luciferase activities were measured 48 h after transfection, and firefly luciferase was normalized to renilla luciferase (control reporter) activity. Values shown are averages (± SEM) of one representative experiment out of three in which each transfection was performed in triplicate. As an experimental control, cells were transfected with reporter constructs and treated with 100 ng/ml recombinant human TNF-α for 3 h (at which maximal activation was identified in previous time course experiments) before cell lysis (top). (B) Namalwa cells were cotransfected with 5 μg NF-κB luciferase reporter plasmid, 0.5 μg of control reporter (pRL-CMV), and 5 and 10 μg of vFLIP, cFLIP_L, or cFLIP_S expression vectors. (C) BC-3 cells were cotransfected with 5 μg NF-κB luciferase reporter plasmid, 0.5 μg of control reporter (pRL-CMV), and 5 and 10 μg of 3X-FLAG-tagged vFLIP and cFLIP_L, or empty vector (P3X-FLAG-CMV14). NF-κB activation was measured using a luciferase reporter assay. No significant difference was found between FLAG-tagged and untagged proteins (not depicted). (D) Immunoblot analysis of extracts from transfected BC-3 cells using an anti-FLAG monoclonal antibody is shown (top), indicating lower protein levels for vFLIP than cFLIP_L in spite of higher NF-κB–inducing activity. Because the sizes of vFLIP and cFLIP_L are very different, the bands in this figure were cut from a single gel to show only the specific band.

Figure 3.

The KSHV-encoded FLIP induces NF-κB and expression of NF-κB–regulated cellular genes in stably transfected lymphoma cells. (A) NF13 cells were transfected with 5 μg NF-κB luciferase reporter plasmid and 0.5 μg of control reporter (pRL-CMV), divided, and cultured at 5 × 10⁵ cells/ml in the absence and presence of 2 μg/ml Doxy. As a control, the parental cell line, Namalwa, was similarly treated with 2 μg/ml Doxy. After 48 h, luciferase activity was measured. Values shown are averages (± SEM) of one representative experiment out of three in which each transfection was performed in triplicate. (B) Proteins extracted from untreated and Doxy-induced NF-13 cells were assessed for vFLIP (to confirm induction), cFLIP, cIAP-1, cIAP-2, and BCL-X_L expression by Western blotting. BC-3 was used as a positive control because it has constitutive NF-κB activity and expresses these proteins. The Namalwa cells were also treated with Doxy to exclude drug-related effects and evaluated for expression of vFLIP and cFLIP. Actin reprobing was performed to assure even protein loading. Doxy induction and analysis of protein expression was confirmed in three independent experiments of which one representative is shown.

Figure 4.

Inhibition of endogenous vFLIP by siRNA results in depletion of constitutive NF-κB activity in PEL cells. (A) BC-3 cells were transfected with a vFLIP siRNA (v) and scramble siRNA (s) and compared with mock-transfected cells (–). Protein extracts were prepared 48 h after transfection. Actin reprobing was performed to assure even protein loading. This experiment has been performed at least 10 times, and a representative immunoblot is shown. (B) BC-3 cells were transfected with an NF-κB luciferase reporter plasmid and either scramble siRNA as a control or siRNA for vFLIP. Luciferase assays were performed 48 h after transfection. Values shown are averages (± SEM) of one representative experiment out of three in which each transfection was performed in triplicate. (C) Electrophoretic mobility shift assays using a radiolabeled probe containing an NF-κB–binding site. BC-1 and BC-3 cells were transfected with vFLIP siRNA (v) or scramble siRNA (s) and compared with mock-transfected cells (–). Cold competition using 50-fold molar excess of an unlabeled NF-κB oligonucleotide (c) demonstrated the specificity of the protein–DNA-binding complexes. Nuclear extracts were prepared 48 h after transfection. In the lane labeled H₂O, water was used instead of nuclear extract. Binding to a radiolabeled oligonucleotide containing an octamer (Oct) motif was similarly examined as a control for specificity and for nuclear protein amount (bottom), where cold competition (c) was performed with nuclear extracts from mock-transfected cells and excess unlabeled oligonucleotide containing an octamer motif. This experiment was performed three times with similar results.

Figure 5.

siRNA to vFLIP inhibits vCYC, but this protein can be reconstituted and its expression does not affect NF-κB activity. (A) BC-3 cells were transfected with a vFLIP siRNA (v) and scramble siRNA (s) with or without 2 μg of a viral cyclin (vCYC) expression vector at days 0, 2, 4, and 6, and compared to mock-transfected cells (–). Protein extracts were prepared at day 8. This experiment was performed three times, and a representative immunoblot is shown. (B) 293T cells were transfected with 5 μg of NF-κB luciferase reporter plasmid in addition to 5 μg of vFLIP/pcDNA3.1, vCYC/pcDNA3.1, both, or vector alone. Luciferase activities were measured 48 h after transfection and firefly luciferase was normalized to renilla luciferase (control reporter) activity. Values shown are averages (± SEM) of one representative experiment out of three in which each transfection was performed in triplicate.

Figure 6.

Inhibition of endogenous vFLIP by siRNA results in down-regulation of NF-κB–dependent proteins in PEL cells. BC-3 cell lines were transfected with a vFLIP siRNA (v), scramble siRNA (s), or mock-transfected (–) at days 0, 2, 4, and 6, and protein extracts were prepared at the indicated time points. Western blot analysis was performed on these extracts using antibodies to the indicated proteins. Actin reprobing was performed to confirm even protein loading. Similar results were obtained in three independent experiments.

Figure 7.

Inhibition of endogenous vFLIP by siRNA results in apoptosis of PEL cells. (A) BC-3 cells were transfected with a vFLIP siRNA (v), scramble siRNA (s), or mock-transfected (–) at days 0, 2, 4, and 6, with or without vCYC, and assessment of apoptosis was performed by annexin V staining at the indicated time points. BC-3 cells were also evaluated for annexin V positivity at day 8 after being transfected once at day 0 (*) and twice at days 0 and 4 (**). Bars represent the average number of Annexin V positive cells (± SEM) of one representative experiment out of three in which each transfection and analysis was performed in triplicate. Inset shows a representative flow cytometric histogram plot at day 8 for annexin V analysis of untreated BC-3 cells (dashed line), cells transfected with scrambled siRNA as a negative control (solid line), and cells treated with siRNA to vFLIP (gray line and filled area). (B) Induction of apoptosis was confirmed by cleavage of caspase 3 and a caspase substrate, PARP, by Western blot analysis. Actin reprobing was performed to confirm even protein loading. Three independent experiments were performed, and a representative Western blot is shown.

Figure 8.

Only KSHV-positive PEL cell lines are sensitive to vFLIP siRNA. The indicated cell lines were transfected with a vFLIP siRNA (v), scramble siRNA (s), or mock-transfected (–) at days 0, 2, 4, and 6, and flow cytometry using antibodies to annexin V was performed at day 8. Bars represent the average number of annexin V positive cells (± SEM) of an experiment in which each transfection and analysis was performed in triplicate. Efficiency of transfection as documented by flow cytometry analysis after transfection with a fluoresceinated siRNA to luciferase, was comparable in all cell lines (not depicted). KSHV-positive PEL cell lines (BC-1, BCBL-1, BC-2, and BC-5) and KSHV-negative B cell lymphoma (BJAB, Namalwa, and IBL) or lymphoblastoid cell lines (LCLs) were evaluated.

Figure 9.

Inhibition of vFLIP in BC-3 cells sensitizes them to FAS-mediated apoptosis. BC-3 cells were transfected with a vFLIP siRNA (v) or scramble siRNA (s) at days 0, 2, and 4, and flow cytometry using antibodies to annexin V was performed at day 6. An agonistic antibody to Fas was added as indicated 24 h before flow cytometry at a concentration of 800 ng/ml. Bars represent the average number of annexin V positive cells (± SEM) of a representative experiment out of three, in which each transfection and analysis was performed in triplicate. A Student's t test showed that the difference between the untreated and anti-FAS–treated cells in conjunction with vFLIP siRNA was statistically significant (P = 0.001).

Figure 10.

Model of vFLIP effect of cellular survival. We propose that vFLIP may inhibit apoptosis via a direct mechanism, by disrupting activation of the DISC complex, and an indirect pathway by forming part of a signaling complex that induces IKK phosphorylation (via an unknown kinase or kinases), IκB degradation, and NF-κB activation. In turn, activation of NF-κB up-regulates expression of genes that inhibit apoptosis, which include cFLIP, cIAP1, and cIAP2. We do not know if TNF receptors (active or inactive) are part of the vFLIP-containing complex and whether this complex is located in membrane lipid rafts in proximity to TNF receptors.