|Home | About | Journals | Submit | Contact Us | Français|
While CD95 is an apoptosis-inducing receptor and has emerged as a potential anticancer therapy target, mounting evidence shows that CD95 is also emerging as a tumor promoter by activating nonapoptotic signaling pathways. Gammaherpesviral infection is closely associated with lymphoproliferative diseases, including B cell lymphomas. The nonapoptotic function of CD95 in gammaherpesvirus-associated lymphomas is largely unknown. Here, we show that stimulation of CD95 agonist antibody drives the majority of sensitive gammaherpesvirus-transformed B cells to undergo caspase-dependent apoptosis and promotes the survival and proliferation of a subpopulation of apoptosis-resistant B cells. Surprisingly, CD95-mediated nonapoptotic signaling induced beta interferon (IFN-β) expression and correlatively inhibited B cell receptor (BCR)-mediated gammaherpesviral replication in the apoptosis-resistant lymphoma cells without influencing BCR signaling. Further analysis showed that IFN-β alone or synergizing with CD95 blocked the activation of lytic switch proteins and the gene expression of gammaherpesviruses. Our findings indicate that, independent of its apoptotic activity, CD95 signaling activity plays an important role in blocking viral replication in apoptosis-resistant, gammaherpesvirus-associated B lymphoma cells, suggesting a novel mechanism that indicates how host CD95 prototype death receptor controls the life cycle of gammaherpesviruses independent of its apoptotic activity.
IMPORTANCE Gammaherpesviruses are closely associated with lymphoid malignancies and other cancers. Viral replication and persistence strategies leading to cancer involve the activation of antiapoptotic and proliferation programs, as well as evasion of the host immune response. Here, we provide evidence that the stimulation of CD95 agonist antibody, mimicking one of the major mechanisms of cytotoxic T cell killing, inhibits B cell receptor-mediated gammaherpesviral replication in CD95 apoptosis-resistant lymphoma cells. CD95-induced type I interferon (IFN-β) contributes to the inhibition of gammaherpesviral replication. This finding sheds new light on the CD95 nonapoptotic function and provides a novel mechanism for gammaherpesviruses that helps them to escape host immune surveillance.
CD95 (also called APO-1 or FAS) is a death receptor belonging to the tumor necrosis factor receptor family that is characterized by the presence of a death domain within its cytoplasmic region (1, 2). Stimulation of CD95 cognate ligand (CD95L) or specific agonistic antibodies results in the assembly of the death-inducing signaling complex (DISC), composed of CD95, the adaptor molecule FADD (FAS associated with a death domain), procaspase 8, procaspase 10, and the caspase 8/10 regulator c-FLIP (3,–7). Activated caspase 8 subsequently cleaves the effector caspases 3 and 7, initiating the apoptotic program. Apoptosis mediated by CD95-CD95L interaction is crucial for the immune system to maintain homeostasis and eliminate virus-infected and cancer cells (6, 8,–10).
A number of cancer cells exhibit high-level surface expression of CD95 but are refractory to CD95-mediated apoptosis. This phenomenon has led to extensive investigation into CD95 nonapoptotic function over the last several decades. Growing evidence demonstrates that the CD95-mediated nonapoptotic signal has evolved diverse roles, such as inducing activation and the proliferation of various cells (11,–14), increasing cancer cell motility and invasiveness (15), and promoting tumor growth and epithelial-to-mesenchymal transition, as well as promoting cancer stem cell survival (16,–20). Additionally, activation of CD95 also triggers the secretion of inflammatory cytokines and plays an important role in inflammation (21,–24). The mechanism underlying the CD95 nonapoptotic function involves the activation of multiple tumorigenic pathways, which include NF-κB; Src/PI3K/AKT/mTOR; Src/PI3K/GSK3β/MMP (matrix metalloproteinase); and three MAP kinases, ERK1/2, JNK1/2, and p38 (18, 25,–27).
Gammaherpesviruses, including Epstein-Barr virus (EBV), Kaposi's sarcoma-associated herpesvirus (KSHV), and murine gammaherpesvirus 68 (MHV68), are characterized by the establishment of latency in B lymphocytes and are closely associated with B cell lymphomas and other malignancies. Virus reactivation from latently infected B cells can be induced in vitro through various stimuli that activate B cells, including phorbol esters, ionophore, butyrate, and anti-immunoglobulin (anti-Ig) (28). Anti-Ig cross-linking-mediated B cell receptor (BCR) signaling mimics the effect of antigen binding to Ig molecules on antigen-specific B cells, which is considered the in vivo stimulation for the disruption of gammaherpesviral latency in B cells that is induced by antigen-driven terminal B cell differentiation. Entry into the lytic replication cycle requires expression of the highly conserved immediate-early gene ORF50 in KSHV and MHV68, which encodes a transcriptional activator referred to as Rta (29). In the case of EBV, the immediate-early transcription activator Zta, encoded by the BZLF1 gene, is required for full expression of the lytic cascade, leading to production of progeny virus (29).
CD95 signaling has been shown to be important for CD4+ T cells to inhibit the growth of EBV-transformed B cells and for CD8+ T cells to control MHV68 infection (30, 31). However, the majority of EBV-positive lymphoma cells appear refractory to CD95-mediated apoptosis (32), and only a few lymphoblastoid cell lines (LCLs), transformed by EBV in vitro or derived from EBV-infected posttransplant lymphoproliferative disorder patients, remain sensitive to CD95-mediated apoptosis (33,–35). The nonapoptotic role of CD95 in gammaherpesviral latency and reactivation remains unknown.
In this study, we performed an analysis of the responses of gammaherpesvirus-associated lymphoma cells to stimulation by a CD95 agonistic antibody and aimed to understand the roles of CD95 nonapoptotic signaling in gammaherpesvirus-associated lymphomagenesis. Here, we report that stimulation with anti-CD95 agonist can induce the majority of CD95-sensitive MHV68- or EBV-associated lymphoma cells to undergo caspase-dependent apoptosis and, simultaneously, also renders a subpopulation of cells resistant to CD95-mediated apoptosis. Anti-CD95 stimulation induced reversible CD95 surface downregulation that is mediated by internalization and endocytosis in CD95-resistant cells. More importantly, we demonstrate that stimulation with anti-CD95 agonist inhibits BCR-mediated viral replication of MHV68 and EBV in apoptosis-resistant B lymphoma cells. The inhibitory effect was not due to blockage of the BCR signaling pathways. We also show that the stimulation with anti-CD95 agonist induces beta interferon (IFN-β) expression. IFN-β alone or synergized with anti-CD95 inhibited the activation of lytic switch proteins and gene expression of MHV68 and EBV, which might contribute to CD95-induced inhibition of BCR-mediated gammaherpesviral replication.
MHV68-transformed SL-1 cells were cultured in RPMI 1640 supplemented with 5% heat-inactivated fetal bovine serum (FBS) and 1% penicillin-streptomycin. The EBV-positive marmoset B cell line B95.8, the EBV-positive Burkitt's lymphoma cell lines Raji and Akata, and the human lymphoblastoid cell lines X50-7 and JY were provided by Samuel Speck (Emory University). EBV-transformed LCLs were generated from healthy peripheral blood mononuclear cells (PBMC) with EBV from the B95.8 cell line as previously described (36). A single clone of LCL-1 was selected from serial dilutions of cell culture. All the cells were cultured in RPMI 1640 containing 10% FBS and 1% penicillin-streptomycin. Primary human B cells were isolated from healthy PBMC with a human naive B cell enrichment kit (Stemcell Technologies; 19254) according to the manufacturer's instructions. Primary murine B cells were isolated from murine splenocytes with a mouse B cell isolation kit (Stemcell Technologies; 19854) according to the manufacturer's instructions.
The EBV Zta promoter sequence, which contains the nucleotide (nt) −221 to +12 region, was amplified from genomic DNA of EBV-positive Akata cells as described previously (37). The forward and reverse primers used for PCR amplification were 5′-GAGGTACCCCATGCATATTTCAACTGGGCTGTCT-3′ and 5′-GTGTAAGCTTGCAAGGTGCAATGTTTAGTGAGTTACC-3′. The PCR fragments were cloned into the KpnI and HindIII restriction sites of the pGL4 basic luciferase vector (Promega) to yield Zta-Luc plasmids. The MHV68 Rta-Luc plasmid containing the 410-bp proximal promoter region was a gift from Samuel Speck (38).
The reagents used were as follows. Healthy PBMC were purchased from Changhai Blood Center, Shanghai, China. Antibodies to GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (sc-32233), Akt1 (sc-5298), NF-κB p65 (sc-372), and Zta (sc-53904) were from Santa Cruz Biotechnology. The Apoptosis Antibody Sampler kit (9915 and 9930), phospho-Akt (Ser473) (193H12) and phospho-Akt (Thr308) (244F9) antibodies, the PhosphoPlus p44/42 mitogen-activated protein kinase (MAPK) (Erk1/2) antibody kit (9100), the PhosphoPlus p38 MAP Kinase Antibody kit (9210), and the PhosphoPlus SAPK/JNK Antibody kit (9250) were from Cell Signaling Technology. Early D antigen (EAD) (ab30541) and anti-mouse CD95 (ab82419) antibodies were from Abcam. Anti-mouse CD95 (555670) and isotype control antibody (554721), anti-mouse CD95-fluorescein isothiocyanate (FITC) (554257), anti-mouse CD95-phycoerythrin (PE) (554258), hamster IgG2 λ1 isotype control (553965), and the Pancaspase inhibitor zVAD-fmk (550377) were from BD Biosciences. Anti-human CD95 CH-11 (SY-001) and isotype control antibody (M079-3) were from MBL international. Goat F(ab′)2 anti-mouse IgG plus IgM, and F(ab′)2 goat-anti-human IgG plus IgM were obtained from Jackson Immuno Research. Staurosporine (19-123) was from Millipore. Etoposide (341205) was from Calbiochem. The NF-κB inhibitor Bay11-7082 (tlrl-b82) was from InvivoGen. Recombinant mouse IFN-β (mIFN-β) (12401-1) was from PBL. Recombinant human IFN-β (hIFN-β) 300-02BC was from PeproTech.
For the apoptosis assay, MHV68-infected SL-1 cells (5 × 105 cells/ml) were cultured in the presence of 50 ng/ml anti-CD95 (555670; BD Biosciences) or isotype-matched control (554721; BD Biosciences), while the EBV-positive LCLs JY and X50-7 were stimulated with 100 ng/ml anti-human CD95 CH-11 (MBL International) or isotype-matched control (M079-3; MBL International). Each sample and experiment had three replicates. Cells were collected for annexin V and propidium iodide (PI) staining with an Annexin V Apoptosis Detection kit APC (88-8007; eBioscience) according to the manufacturer's instructions and subjected to fluorescence-activated cell sorter (FACS) analysis. For proliferation assays, cells were stained with 1 μM carboxyfluorescein succinimidyl ester (CFSE) prior to anti-CD95 stimulation. Briefly, the cells were suspended in 0.1% FBS–phosphate-buffered saline (PBS) at 2 × 106 cells/ml and stained with CFSE at a final concentration of 1 μM. After incubation at 37°C for 15 min, the cells were washed with 1 volume of prewarmed FBS and incubated at 37°C for 10 min. Following washing with 2% FBS in PBS three times, the cells were resuspended and continuously cultured with or without anti-CD95 treatment. The cells were harvested at the indicated times and subjected to FACS analysis. For zVAD-fmk (zVAD) treatment, cells were pretreated with 10 μM zVAD for 1 h prior to stimulation with anti-CD95 for 24 h. Etoposide and staurosporine treatments were performed at concentrations of 25 μg/ml and 200 nM, respectively, for 24 h prior to apoptosis analysis.
Western blot analysis was carried out by standard procedures. For immunofluorescence, SL-1 cells were stimulated with anti-CD95 or isotype control for 72 h, and live cells were isolated by cell sorting and seeded at approximately 1 × 105 cells/well in the cover well (Sigma). After incubation at 37°C for 1 h, the cells were fixed with 4% paraformaldehyde at room temperature for 1 h or 4°C overnight and subjected to staining with the primary antibody NF-κB p65 (sc-372; Santa Cruz) for 1 h. The cells were then washed with PBS and incubated with secondary antibody for an additional 30 min. Images were acquired using a Leica confocal laser microscope system (TCS SP5). For FACS analysis, cells were stored on ice after staining and analyzed on a Becton Dickinson LSR Fortessa. The data were analyzed using Flowjo (Treestar Inc.). When needed, cell sorting was performed on a BD FACSAria II cell sorter.
SL-1 cells (2.5 × 105 cells/ml) were resuspended with culture medium or medium containing 1 M sucrose in the presence of 50 ng/ml anti-CD95 or isotype control. After stimulation at 37°C for 30 min, cells were collected and subjected to surface staining. For endocytosis analysis, SL-1 cells were pretreated with 1 mM chloroquine or dimethyl sulfoxide (DMSO) as a control, followed by incubation with anti-CD95 at 37°C for 1 h. The cells were then harvested and subjected to intracellular staining.
SL-1 cells were stimulated with anti-CD95 agonist or isotype control for 3 days and stained with annexin V-allophycocyanin (APC) and PI (eBiosciences). Live cells were sorted, followed by total RNA isolation with TRIzol reagent (Invitrogen) according to the manufacturer's instructions. Reverse transcription (RT) was carried out using the cDNA first-strand synthesis kit (TianGen) and subjected to analysis with an RT2 Profiler PCR array mouse cancer pathway finder kit (PAMM-033Z; Qiagen) following the manufacturer's instructions.
qRT-PCR was performed by using SYBR green Supermix (Toyobo) with the 7900HT real-time PCR system (Applied Biosystems). Independent experiments were performed in triplicate with each sample. The primers used were as follows: mIFN-β, 5′-AGCTCCAAGAAAGGACGAACAT-3′ and 5′-GCCCTGTAGGTGAGGTTGATCT-3′; hIFN-β, 5′-TGCTCTCCTGTTGTGCTTCTCCACGAC-3′ and 5′-CTCCAGTTTTTCTTCCAGGACTGTCTTCA-3′; MHV68 ORF50, 5′-GGCCGCAGACATTTAATGAC-3′ and 5′-GCCTCAACTTCTCTGGATATCC-3′; MHV68 ORF25, 5′-CAGCGGCGTCTTTGAAACAA-3′ and 5′-GTAGCCGCAGGTATTGTGGT-3′; EBV gp350, 5′-GTCAGTACACCATCCAGAGCC-3′ and 5′-TTGGTAGACAGCCTTCGTATG-3′; mGAPDH, 5′-AACGACCCCTTCATTGACCT-3′ and 5′-ATGTTAGTGGGGTCTCGCTC-3′; and hGAPDH, 5′-GCCATCACTGCCACCCAGAAGACTGTG-3′ and 5′-TTACTCCTTGGAGGCCATGTGGGCCATG-3′. The GAPDH housekeeping gene was used as an internal control for normalization. For viral DNA quantification, genomic DNA was isolated with a Tianamp Genomic DNA kit (DP304-02; TianGen). Real-time DNA PCR for the g50 gene was used to measure the MHV68 genome, while the EBER gene was used to measure the EBV genome. Relative genome copy numbers were calculated based on normalization with the GAPDH housekeeping gene. The primers used were as follows: MHV68 g50, 5′-GGCCGCAGACATTTAATGAC-3′ and 5′-GCCTCAACTTCTCTGGATATGCC-3′; EBV EBER, 5′-CGTCACGGTGACGTAGTCTGTCTTGAG-3′ and 5′-CCCTTCTCCCAGAGGGATTAGAGAATCC-3′.
EBV Zta-Luc promoter and Renilla reporter plasmids were transfected into Raji cells by nucleofection (Lonza) and stimulated with 100 ng/ml anti-CD95 and/or 1,000 U recombinant human IFN-β for 24 h. MHV68 Rta-Luc promoter and Renilla reporter plasmids were transfected into SL-1/R cells cultured with 50 ng/ml anti-CD95 or isotype control and followed by stimulation with 100 U recombinant mouse IFN-β for 48 h. Luciferase assays were carried out in triplicate. Luciferase activity was measured using the Dual Luciferase Reporter Gene Assay kit (Beyotime Biotechnology) according to the manufacturer's instructions and normalized to Renilla activity.
Statistical analysis was performed in Prism (Graph Pad Software). The data are reported as means and standard deviations (SD). Differences between groups of research subjects were analyzed for statistical significance with two-tailed Student t tests. A P value of < 0.05 was considered significant.
Based on the function of CD95 acting as an apoptosis inducer and tumor promoter, we wondered how CD95 functions in gammaherpesvirus-associated lymphomagenesis. Although previous reports demonstrated that EBV-positive B lymphoma cells differentially exhibit sensitivity to CD95-mediated apoptosis, in-depth investigation of CD95 function in gammaherpesvirus-associated lymphomagenesis has not been carried out. We previously reported that MHV68 can transform fetal-liver-derived B cells into LCL-like SL B cells, which can also induce lymphomas in immunodeficient mice (39), and that the growth of MHV68-transformed SL B cells and associated lymphomas can be controlled by both CD4+ and CD8+ T cells (40). To elucidate the CD95 function in gammaherpesvirus-associated lymphomagenesis, we first analyzed CD95 surface expression of MHV68-transformed SL B cells. All tested individual clones, including SL-1, -2, -3, -4, and -5, showed positive surface expression of CD95 compared to murine primary B cells by FACS analysis (Fig. 1A). Next, we tested how prolonged anti-CD95 stimulation would affect MHV68-transformed SL lymphoma cells. We cultured SL-1 cells in the presence of anti-CD95 agonist, which can activate CD95-mediated canonical apoptotic signaling. Stimulation with anti-CD95 agonist rendered SL-1 cells sensitive to apoptosis compared to stimulation with isotype control, as shown by apoptosis detection with annexin V and PI staining (Fig. 1B and andC).C). However, not all cells in a culture died; a subpopulation of cells survived even after 5 days of constitutive stimulation with anti-CD95 (Fig. 1C). To test whether this subpopulation of cells underwent proliferation, we performed a CFSE assay for the live cells gated from annexin V and PI staining. The surviving subpopulation derived from anti-CD95 agonist treatment underwent proliferation, which was even slightly faster than the cells treated with isotype control (Fig. 1D). These data demonstrated that a subpopulation resistant to CD95-mediated apoptosis might exist in SL-1 cells or that anti-CD95 agonist stimulation might render a subpopulation resistant to CD95-mediated apoptosis. A similar observation was also made with EBV-transformed JY and X50-7 LCLs (Fig. 1E).
It has been well demonstrated that CD95-mediated apoptotic signaling is caspase dependent, whereas NF-κB activation is involved in CD95 nonapoptotic function (41, 42). To define CD95-mediated apoptotic and nonapoptotic signaling in gammaherpesvirus-associated lymphoma cells, MHV68-transformed SL-1 cells were stimulated with anti-CD95 agonist or isotype control for 3 days. The live cells were subsequently isolated through flow cytometry, based on annexin V and PI staining, and they continued to grow in culture for a prolonged period in the presence of antibodies (Fig. 2A). Cleaved caspases 8, 3, and 9 were easily detected by immunoblot analyses in SL-1 cells that were stimulated with anti-CD95 agonist for 24 h, but not in isotype-treated cells and the live subpopulation sorted from day 3 poststimulation with anti-CD95, as well as the live cells that continued to grow in culture in the presence of anti-CD95 for 1 month (this population is referred to as SL-1/R below) (Fig. 2B). Comparisons of cleaved caspases were quantitated through normalization with GAPDH expression (indicated below each immunoblot in Fig. 2B). Our data suggested that stimulation with anti-CD95 induces caspase activation in SL-1 cells, but not in the subpopulation of apoptosis-resistant surviving cells. NF-κB activation by CD95 has been linked to antiapoptotic and tumorigenic pathways (15). We therefore tested the localization of NF-κB p65 by confocal immunofluorescent staining. As expected, compared with p65 cytoplasmic localization in the cells that were sorted from day 3 culture with isotype control, p65 was translocated into the nucleus in about 92% of the surviving subpopulation of cells from day 3 culture with anti-CD95 (Fig. 2C), characteristic of activation of NF-κB. These data are in agreement with a CD95 nonapoptotic function that promotes cell growth by activating the NF-κB signaling pathway (42).
To further confirm whether CD95-mediated apoptosis is caspase dependent, SL-1 cells were stimulated with anti-CD95 for 24 h in the presence or absence of the oligospecific caspase inhibitor zVAD. zVAD treatment not only inhibited the cleavage of caspases 9, 8, and 3 and PARP in response to anti-CD95 (Fig. 2D), but also blocked CD95-mediated apoptosis (Fig. 2E). To rule out the possibility of losing apoptosis capability in the surviving apoptosis-resistant cells, SL-1/R cells were stimulated with alternative cell death inducers, the DNA-damaging agent etoposide and the protein kinase inhibitor staurosporine. Both etoposide and staurosporine treatment induced a substantial amount of cell death (Fig. 2F), implying that resistance to CD95-mediated cell death is specific for the surviving SL-1/R cells after prolonged culture with anti-CD95 agonist.
One recent report demonstrated that stimulation by anti-CD95 agonist induces phenotypic conversion of noncancer stem cells to cancer stem cells, along with concomitant reduction in sensitivity to CD95-mediated apoptosis (20), which prompted us to analyze whether stimulation by anti-CD95 agonist could induce any phenotypic change of SL-1/R cells. We analyzed various B cell surface markers through flow cytometry analyses. Surface expression of B220, CD138, IgG, I-Ab, and CD44 was almost the same in SL-1/R and isotype-treated control cells, and only CD25 was slightly decreased in SL-1/R cells (Fig. 3A), indicating that, unlike the previously reported cancer cells, stimulation by anti-CD95 does not induce any phenotypic change in gammaherpesvirus-associated lymphoma cells; other mechanisms could be involved in resistance to apoptosis after anti-CD95 stimulation.
Constant stimulation by CD95L has been shown to induce the internalization of CD95 receptor, which defines cellular fates in different types of cells (43). To assess whether prolonged stimulation by anti-CD95 agonist downregulates the CD95 surface expression of gammaherpesvirus-associated lymphoma cells, we analyzed CD95 surface expression of parental SL-1 cells, live SL-1 cells isolated from day 3 posttreatment with anti-CD95, SL-1/R cells, and the surviving subpopulation naturally selected by constant anti-CD95 stimulation. Compared to the cells treated with isotype control and parental SL-1 cells, anti-CD95 treatment, either short term or long term, strikingly induced surface downregulation of CD95 (Fig. 3B). We next tested the downregulation rate of CD95 by time course flow cytometry analyses. The surface downregulation of CD95 was rapid, and it reached steady state around 30 min after anti-CD95 incubation in SL-1 cells (Fig. 3C). However, this process was much slower in the EBV-transformed LCL JY cell line, reaching a stable level after 4 h of anti-CD95 stimulation (Fig. 3D). The differential downregulation process of CD95 surface expression mediated by anti-CD95 might result from the cellular changes that were differentially mediated by MHV68 and EBV infection and transformation.
To determine whether the surface downregulation of CD95 was due to the internalization mediated by anti-CD95 ligation, SL-1 cells were treated with anti-CD95 for 30 min in the presence of 1 M hypertonic sucrose or 1 mM the endosomal acidification inhibitor chloroquine, followed by cell surface staining and intracellular staining analyses, respectively. Sucrose has been shown to inhibit receptor internalization, while chloroquine is known to inhibit receptor endocytosis. Sucrose treatment blocked the downregulation of CD95 cell surface expression, while chloroquine inhibited the downregulation of CD95 intracellular expression in SL-1 cells (Fig. 3E), suggesting that the ligation of anti-CD95 induces CD95 receptor internalization and endocytosis, leading to CD95 surface downregulation in the surviving subpopulation of gammaherpesvirus-associated lymphoma cells.
To examine whether CD95 surface downregulation was reversible, SL-1/R cells were cultured in the absence of anti-CD95 for 1 week and then assessed for CD95 surface expression. Removal of anti-CD95 in SL-1/R cells restored CD95 surface expression to a level that was similar to that of the parental SL-1 cells and allowed the cells to regain sensitivity to CD95-mediated apoptosis (Fig. 3F). The above-described data demonstrate that anti-CD95 stimulation induces CD95 receptor internalization and endocytosis in the subpopulation of SL-1 cells, leading to reversible downregulation of CD95 surface expression and rendering SL-1/R cells resistant to CD95-mediated apoptosis.
Anti-Ig cross-linking-mediated BCR signaling can efficiently activate the lytic replication of latently EBV- or MHV68-infected B cells (28, 39). To test if anti-CD95 agonist stimulation would affect the activation of viral latency in gammahepesvirus-associated lymphoma cells, which are resistant to CD95-mediated apoptosis, SL-1/R cells were continuously cultured with anti-CD95 or with isotype control or cultured without any treatment after stimulation with 5 μg/ml F(ab′)2 anti-mouse IgG plus IgM for 48 h. Surprisingly, the expression of the viral DNA polymerase processivity factor ORF59 and viral lytic antigens detected with MHV68 antiserum was dramatically blocked upon reactivation by anti-Ig cross-linking in SL-1/R cells continuously cultured with anti-CD95, but not in SL-1/R cells cultured with isotype or without any treatment (Fig. 4A). We did not observe any significant change in the latency-associated vCyclin gene (Fig. 4A). MHV68 DNA replication was further examined by quantitative-PCR analysis. In the absence of anti-Ig stimulation, we did not detect any substantial difference in MHV68 DNA among SL-1/R cells with different treatments. However, upon activation by anti-Ig stimulation, MHV68 DNA levels were significantly lower in SL-1/R cells that were continuously cultured with anti-CD95 than in SL-1/R cells cultured with isotype or without any treatment (Fig. 4B). Gene expression of the MHV68 immediate-early gene ORF50 and late gene ORF25 were analyzed by quantitative RT-PCR and normalized with cellular GAPDH expression. Similarly, mRNA expression of both ORF50 and ORF25 was blocked in SL-1/R cells with anti-CD95 stimulation upon anti-Ig-mediated activation (Fig. 4C). These data suggest that the ligation of anti-CD95 agonist inhibits MHV68 DNA replication and gene expression mediated by anti-Ig cross-linking in apoptosis-resistant SL-1/R cells.
To reveal whether stimulation by anti-CD95 could similarly inhibit anti-Ig induction of the EBV lytic cycle in EBV-positive B cells resistant to CD95-mediated apoptosis, we first tested CD95 surface expression of EBV-transformed LCL-1 and the EBV-positive Burkitt's lymphoma cell lines Akata and Raji, which were resistant to CD95-mediated apoptosis, as demonstrated previously (32, 34). Compared to the 50.7% of primary human B cells that showed positive surface expression of CD95, over 95% of LCL-1, Akata, and Raji cells displayed positive expression of surface CD95 (Fig. 5A). Therefore, LCL-1, Akata, and Raji cells were chosen and treated with F(ab′)2 anti-human IgG plus IgM in the presence of anti-human CD95 agonist or isotype control for 24 h. EBV EAD and the lytic switch protein Zta (also known as ZEBRA or BZLF-1) expression could be efficiently induced by anti-Ig cross-linking in all the tested cells without anti-CD95 agonist treatment, but expression of both EAD and Zta was blocked by anti-CD95 treatment upon anti-Ig-mediated activation in all the tested cell lines (Fig. 5B). To detect when anti-CD95 agonist also inhibited EBV DNA replication and late gene expression, we performed quantitative PCR to detect the EBV genome and quantitative RT-PCR to detect gene expression of the EBV gp350 late gene in Akata cells. Similarly, anti-CD95 agonist treatment blocked EBV DNA replication and gp350 mRNA expression in Akata cells upon anti-Ig stimulation (Fig. 5C and andD).D). These data demonstrate that the stimulation of anti-CD95 agonist inhibits BCR-mediated viral replication in EBV-associated lymphoma cells that are resistant to CD95-mediated apoptosis, in agreement with the observation in MHV68-positive SL-1/R cells, suggesting a potential common characteristic of gammaherpesvirus-associated lymphoma cells.
BCR-mediated signaling molecules, including phosphatidylinositol 3-kinase (PI3K) and MAPK, play essential roles in EBV activation (44, 45). The ability of anti-CD95 stimulation to block BCR-mediated viral replication raises the possibility that CD95-CD95L interaction might disrupt BCR signaling pathways. To test this, SL-1/R cells were continuously cultured with anti-CD95 or with isotype control following anti-Ig cross-linking. We examined the phosphorylated forms of BCR signaling molecules, which act as active signaling pathways, through immunoblot analyses. Constitutive phosphorylation of the BCR signaling molecules Akt, ERK, JNK, p65, and p38 was detected by phosphorylated-protein-specific antibodies in either anti-CD95- or isotype-treated SL-1/R cells in the absence of anti-Ig stimulation (Fig. 6A), which was indicative of constant active BCR signal and provided a potential mechanism underlying apoptosis resistance and survival in SL-1/R cells. This is consistent with the mechanism underlying the nonapoptotic function of CD95, as reported previously (6). Surprisingly, anti-Ig cross-linking decreased phosphorylated Akt, ERK, JNK, NF-κB p65, and p38, but no significant difference was observed between SL-1/R cells cultured with anti-CD95 and with isotype control (Fig. 6A), suggesting that anti-CD95 agonist-induced inhibition of BCR-mediated MHV68 replication is not due to any disruption of the BCR signaling pathway. It should be noted that cross-linking of anti-Ig reduced the phosphorylated form of BCR signaling molecules in SL-1/R cells, which might be due to the secondary effect of anti-Ig on highly constitutive activation of BCR signaling induced by anti-CD95.
To test if a similar observation could be made in EBV-positive cells, LCL-1 cells resistant to CD95-mediated apoptosis were treated with anti-Ig in the presence of anti-CD95 agonist or isotype control for 0 min, 10 min, 30 min, and 24 h. Similarly, treatment with anti-CD95 had no significant effect on the levels of phosphorylated Akt, ERK, JNK, p65, and p38 in LCL-1 cells upon anti-Ig cross-linking compared to the cells treated with isotype control at both early and late time points (Fig. 6B). The above-described findings suggest that the stimulation of anti-CD95 does not inhibit BCR signal transduction in both MHV68-associated and EBV-associated lymphoma cells, where inhibition of viral replication might result from other mechanisms.
Activation of NF-κB is one of the predominant downstream signals mediated by BCR that also figures prominently in the latency program of all gammaherpesviruses, antagonizing their entry into the lytic cycle and thereby stabilizing latency in many cells (29). To test whether CD95-induced NF-κB activation plays a role in inhibiting BCR-mediated gammaherpesviral replication, SL-1/R, LCL-1, Raji, and Akata cells were treated with anti-CD95 in the presence or absence of 2 μM the NF-κB inhibitor Bay11-7082 upon anti-Ig cross-linking. Inactivation of NF-κB by Bay11-7082 did not rescue CD95-mediated inhibition of MHV68 ORF59 and lytic antigen expression upon BCR-mediated activation (Fig. 6C). Similarly, the expression of EBV EAD and Zta was also not restored by Bay11-7082 in LCL-1, Akata, and Raji cells (Fig. 6D). Notably, we performed concentration optimization for Bay11-7082 and did not observe any cytotoxic effect of 2 μM Bay11-7082 on all the tested cell lines. These results demonstrate that the NF-κB activation induced by anti-CD95 does not contribute to CD95-mediated inhibition of viral replication in MHV68-associated and EBV-associated lymphoma cells upon anti-Ig cross-linking. We therefore postulate that another underlying mechanism might be involved in CD95 agonistic antibody-induced inhibition of BCR-mediated viral replication from latency in gammaherpesvirus-associated lymphoma cells.
To reveal the mechanisms involved in anti-CD95-induced inhibition of BCR-mediated gammaherpesviral replication, a cancer pathway qRT-PCR array was performed to compare one subset of the gene expression profile between the live cells isolated from anti-CD95-treated and isotype control-treated SL-1 cells. Through a thorough analysis of the gene expression array, a panel of genes, including Fos, c-Jun, twist, plaur, serpinb2, and IFN-β genes, were significantly induced in anti-CD95-stimulated SL-1 cells (Fig. 7A). Among the genes, those encoding twist and AP-1 transcription factor components Fos and c-Jun were involved in the anti-apoptotic function, supporting the survival signal mediated by anti-CD95 stimulation. The type I interferon IFN-α/β has been observed to play a critical role in inhibiting viral reactivation during MHV68 latency in vivo (46), which prompted us to analyze whether IFN-β expression induced by anti-CD95 contributes to the inhibition of BCR-mediated gammaherpesviral replication in apoptosis-resistant cells. IFN-β expression was confirmed by quantitative RT-PCR assay. SL1/R cells continuously cultured with anti-CD95 showed significant elevation of IFN-β gene expression compared to SL-1 control cells continuously cultured with isotype control (Fig. 7B). Likewise, EBV-positive LCL-1, Akata, and Raji cells were stimulated with anti-CD95 or isotype control, and IFN-β gene expression was also markedly increased among all three tested cell lines upon anti-CD95 agonist stimulation compared to the isotype control (Fig. 7C), indicating that anti-CD95 stimulation effectively induced the upregulation of IFN-β gene expression in apoptosis-resistant, gammaherpesvirus-associated lymphoma cells.
To determine if IFN-β induction contributes to anti-CD95-mediated inhibition of BCR-induced viral replication, we first tested the effects of anti-CD95 and IFN-β on the activation of the MHV68 lytic switch protein Rta and EBV switch protein Zta promoters, which trigger the full lytic cycle from viral latency. The luciferase reporter of the MHV68 Rta promoter and the Renilla reporter were transfected into SL-1/R or isotype control-treated cells in the presence or absence of anti-CD95 and IFN-β stimulation, while the luciferase reporter of the EBV Zta promoter was transfected into EBV-positive Raji cells. The luciferase activities of the Rta and Zta promoters were assayed in triplicate and normalized to Renilla reporter activity. The stimulation of anti-CD95 or IFN-β alone caused substantial reduction of both Rta and Zta luciferase promoter activities, whereas costimulation of anti-CD95 and IFN-β produced the largest reduction of Rta and Zta luciferase activities (Fig. 8A), suggesting that stimulation of anti-CD95 could not only inhibit the promoter activities of MHV68 Rta and EBV Zta, but could also synergize with IFN-β to enhance the reduction of Rta and Zta promoter activities. These results demonstrated that IFN-β expression induced by anti-CD95 might play a role in inhibiting the activation of gammaherpesvirus lytic switch proteins and the subsequent viral replication of gammaherpesviruses in CD95 apoptosis-resistant lymphoma cells.
To further delineate the effect of IFN-β on CD95-mediated inhibition of gammaherpesviral replication, SL-1/R, LCL-1, and Raji cells were cultured with anti-CD95 or isotype control in the presence or absence of recombinant IFN-β after anti-Ig cross-linking. Consistent with the in vivo result described previously (46), IFN-β treatment alone reduced the expression of MHV68 lytic antigens and ORF59 in SL-1/R cells and the expression of EBV EAD and Zta in LCL-1 and Raji cells upon reactivation mediated by anti-Ig cross-linking (Fig. 8B and andC),C), in agreement with IFN-β function in the inhibition of MHV68 Rta and EBV Zta promoter activities. Furthermore, costimulation with anti-CD95 and IFN-β enhanced the reduction of both MHV68 and EBV gene expression in all the tested cell lines (Fig. 8B and andC),C), correlated with the synergistic effect of anti-CD95 and IFN-β on inhibiting the promoter activities of MHV68 Rta and EBV Zta, as described above. We, therefore, conclude that stimulation with anti-CD95 induces IFN-β expression, which possibly contributes to the inhibition of BCR-mediated viral replication in apoptosis-resistant, gammaherpesvirus-associated lymphoma cells.
In this report, we have demonstrated that stimulation with anti-CD95 not only induces the majority of the population of apoptosis-sensitive, MHV68-transformed SL-1 B cells or EBV-transformed LCLs to undergo caspase-dependent cell death, but also renders a small subpopulation of cells resistant to apoptosis. However, this subpopulation of cells was still susceptible to etoposide-induced cell death caused by DNA strand breaks (47) and staurosporine-induced cell death caused by activating caspase 3 (48), suggesting that the surviving cells retain intact cell death pathways mediated by other stimuli. Our observation is consistent with a recent report that constant stimulation by CD95 of breast cancer, colon cancer, and renal cancer cells leads to a subpopulation of cells less sensitive to CD95-mediated apoptosis (20). When we performed a phenotypic analysis of the apoptosis-resistant SL-1 B lymphoma cells after prolonged anti-CD95 stimulation, we did not observe any phenotypic cell surface change, which is different from the recent finding that stimulation by CD95 can transform cancer cells into a more dedifferentiated state and converts apoptosis-resistant noncancer stem cells phenotypically into cancer stem cells (20). Thus, the mechanism underlying apoptosis resistance of SL-1 B lymphoma cells would be the other mode of regulation. It has been reported that the proximal CD95 signaling pathway is impeded in CD95-resistant posttransplant lymphoproliferative-disorder-associated EBV-positive B cell lymphomas (34), whereas reduced CD95 transcription and expression result in CD95 resistance in a variant of the U937 cell line (49). We have attempted to detect the formation of DISC in CD95-resistant SL-1/R cells. However, all the antibodies failed to work in our immunoprecipitation experiments, even though we obtained the same clones from the same companies as reported in other papers. We also detected CD95 transcription but did not observe any notable reduction in SL-1/R cells (data not shown).
The apoptosis-resistant subpopulation of SL-1/R cells exhibits reversible CD95 surface downregulation mediated by internalization and endocytosis (Fig. 3), presumably the cause of CD95 resistance in SL-1/R cells, since the removal of anti-CD95 restores CD95 surface expression and concomitantly renders the cells sensitive to CD95-mediated apoptosis again. This observation could argue for the possibility of a preexisting apoptosis-resistant population. However, we could not rule out the involvement of other CD95-resistant regulation in our system, because we failed to perform some experiments, such as detecting DISC formation, due to the efficacy of the reagents obtained. More delicate experiments are needed to further investigate whether other mechanisms are also involved.
CD95 receptor downregulation in response to CD95L is a common mechanism for a variety of tumor cells to escape destruction by cytotoxic T cells (50). In agreement with this, our unpublished data show that apoptosis-resistant SL-1/R cells with reduced CD95 surface expression display attenuated cytolytic killing by MHV68-specific cytotoxic T cells compared to the parental SL-1 cells. Our finding suggests that, in addition to eliminating the sensitive gammaherpesvirus-associated B lymphoma cells through CD95-mediated apoptosis, the stimulation of anti-CD95 agonist also renders a subpopulation of cells able to evade CD95-mediated apoptosis. As the ability to internalize CD95 and thereby escape immune attack has been correlated with malignancy (51), we need to pay more attention to CD95L cancer therapy, especially antibody-based cancer therapy, in our future studies. It will be important for us to determine the detailed mechanism of CD95 resistance in the resistant subpopulation in order to eliminate them during therapy.
Gammaherpesvirus-associated lymphomas share some common traits with other oncogenic virus-associated cancers, that is, the viruses maintain a latent state in associated cancer cells. Therefore, we primarily concentrated on how CD95-mediated nonapoptosis plays a role in viral latency and reactivation in apoptosis-resistant gammaherpesvirus-associated lymphoma cells, which has not been studied. It is surprising to us that the stimulation of anti-CD95 agonist inhibited BCR-mediated viral replication without any disruption of the BCR signaling pathway (Fig. 4, ,5,5, and and6).6). CD95-induced NF-κB activity in CD95-resistant SL-1/R cells is more consistent with the mechanism underlying the CD95 nonapoptotic survival pathway, as reported previously (41, 42), but it was not involved in CD95-mediated inhibition of MHV68 and EBV replication in our system, which is different from other reports, because NF-κB activation inhibits gammaherpesviral reactivation and stabilizes gammaherpesviral latency (52,–55). BCR determinant signaling molecules like PI3K can activate the EBV transcription activator Zta promoter and subsequent EBV lytic replication (28, 45, 56), but apparently, they do not regulate CD95-mediated MHV68 and EBV replication, which is initiated by anti-Ig cross-linking.
As all KSHV-positive lymphoma cell lines characterized so far lack surface Ig expression, BCR signaling, which is mimicked by anti-Ig cross-linking, has been considered to serve as a more physiologically relevant activator to induce viral reactivation in latently EBV- or MHV68-infected B cells in vivo, while in the present study, we primarily focused on gammaherpesviral replication mediated by anti-Ig treatment. KSHV-positive B lymphoma cells, in which only other chemical stimuli can induce viral reactivation, were not included in this study. Future experiments and investigations will address whether anti-CD95 agonist can also inhibit gammaherpesviral replication that is induced by other stimuli.
The majority of the induced genes from qRT-PCR cancer pathway-related array analysis are involved in antiapoptotic signaling pathways, which might contribute to the survival of the CD95 apoptosis-resistant subpopulation. As a previous report demonstrated that IFN-α/β plays a critical role in inhibiting viral reactivation during latency, based on the observation of increased efficiency of MHV68 reactivation in IFN-α/β receptor knockout mice in vivo (46), we hypothesized that anti-CD95 agonist-induced IFN-β might contribute to the inhibition of gammaherpesviral replication. Direct experiments would involve blocking the IFN-β function with IFN-β or IFN-α/β receptor-blocking antibody. However, we could not obtain efficient blocking effects to perform further analysis, which might be due to the low efficacy of our blocking antibodies. Although either IFN-β or anti-CD95 stimulation could inhibit the activation of lytic switch proteins and MHV68 and EBV gene expression, costimulation by IFN-β and anti-CD95 enhanced the inhibitory effect, but these data are still indirect and are not sufficient to conclude that anti-CD95 agonist-induced IFN-β is the cause of inhibition of MHV68 and EBV replication. However, our findings suggest that IFN-β might contribute to CD95-mediated inhibition of MHV68 and EBV replication upon anti-Ig stimulation. More experiments are needed to address whether CD95-induced IFN-β plays a direct role in CD95-mediated inhibition of gammaherpesviral replication by anti-Ig cross-linking.
The PRDII element of the IFN-β gene promoter is a target for the transcription factor NF-κB (57). NF-κB is also required for virus-induced activation of the IFN-β gene (58). We observed that stimulation by anti-CD95 induced NF-κB activation in apoptosis-resistant B lymphoma cells. However, as discussed above, NF-κB activation did not have a role in inhibiting viral replication in our tested cell lines. It is highly likely that NF-κB activation is not involved in CD95-induced IFN-β gene expression in our system. The mechanism by which stimulation with anti-CD95 induces IFN-β gene expression is still under investigation.
Our data demonstrate for the first time that constant stimulation by anti-CD95 agonist elicits a subpopulation of sensitive gammaherpesvirus-associated lymphoma cells resistant to CD95-mediated apoptosis and inhibits BCR-mediated gammaherpesviral replication in apoptosis-resistant lymphoma cells. This phenomenon suggests that gammaherpesviruses might utilize CD95 nonapoptotic signaling to block viral replication; avoid lytic antigen expression that can be recognized by virus-specific cytotoxic T cells; and, consequently, escape the host immune response, representing a novel mechanism underlying gammaherpesvirus-associated lymphoma cell-host interactions. This raises concerns about the current therapy for gammaherpesvirus-associated lymphomas, since some reports have revealed that radiation and chemotherapy upregulated CD95 through wild-type p53 in different tumor cells (59, 60). Future experiments are needed to investigate the CD95-mediated nonapoptotic function of inhibiting viral reactivation among radiation- and chemotherapy-treated gammaherpesvirus-associated lymphoma patients.
We thank Samuel Speck for providing the MHV68 Rta-Luc plasmid; vCyclin and ORF59 antibodies; and the JY, X50-7, Akata, B95.8, and Raji cell lines.
This work was funded by grants from the National Natural Science Foundation of China (81371825), the Ministry of Science and Technology of China (2016YFA0502100), and the Chinese Academy of Sciences 100 Talents Program (2060299).