Syk Activation of Phosphatidylinositol 3-Kinase/Akt Prevents HtrA2-dependent Loss of X-linked Inhibitor of Apoptosis Protein (XIAP) to Promote Survival of Epstein-Barr Virus+ (EBV+) B Cell Lymphomas*

Background: Syk activation is required for B cell survival. EBV can induce B cell lymphomas. Results: Syk, PI3K/Akt inhibition induces apoptosis of EBV+ B cell lymphomas. Syk PI3K/Akt inhibition results in HtrA2-dependent loss of XIAP protein. Conclusion: Syk activates PI3K/Akt to promote survival by preventing HtrA2-dependent loss of XIAP. Significance: Syk, PI3K/Akt, and XIAP are new therapeutic targets for EBV+ B cell lymphomas. B cell lymphoma survival requires tonic or ligand-independent signals through activation of Syk by the B cell receptor. The Epstein-Barr virus (EBV) protein latent membrane 2a (LMP2a), a mimic of the B cell receptor, provides constitutive survival signals for latently infected cells through Syk activation; however, the precise downstream mechanisms coordinating this survival response in EBV+ B cell lymphomas remain to be elucidated. Herein, we assess the mechanism of Syk survival signaling in EBV+ B cell lymphomas from post-transplant lymphoproliferative disorder (PTLD) to discover virally controlled therapeutic targets involved in lymphomagenesis and tumor progression. Using small molecule inhibition and siRNA strategies, we show that Syk inhibition reduces proliferation and induces apoptosis of PTLD-derived EBV+ B cell lines. Syk inhibition also reduces autocrine IL-10 production. Although Syk inhibition attenuates signaling through both the PI3K/Akt and Erk pathways, only PI3K/Akt inhibition causes apoptosis of PTLD-derived cell lines. Loss of the endogenous caspase inhibitor XIAP is observed after Syk or PI3K/Akt inhibition. The loss of XIAP and apoptosis that results from Syk or PI3K/Akt inhibition is reversed by inhibition of the mitochondrial protease HtrA2. Thus, Syk drives EBV+ B cell lymphoma survival through PI3K/Akt activation, which prevents the HtrA2-dependent loss of XIAP. Syk, Akt, and XIAP antagonists may present potential new therapeutic strategies for PTLD through targeting of EBV-driven survival signals.

vation (18). Along these lines, Akt inhibits the pro-apoptotic molecule Bad (19,20), releasing it from binding the pro-survival Bcl-2 family member Bcl-x L (21). Akt phosphorylates caspase 9 at Ser 196 , inhibiting its cleavage and proteolytic activity (22). Phosphorylation of FOXO transcription factors by Akt inhibits FOXO-mediated transcription of pro-apoptotic genes, including Bim and Fas ligand (23). Akt phosphorylation at Ser 87 stabilizes XIAP, a member of the inhibitor of apoptosis (IAP) family, by preventing its autoubiquitination and degradation (24). XIAP directly binds caspases 3, 7, and 9 and inhibits their activity (25). Thus, Syk signaling may affect a variety of apoptosisrelated proteins in propagation of a survival signal, although the specific mechanisms remain to be elucidated.
In addition to BCR activation, Syk can also be activated by other endogenous receptors and immunoreceptor tyrosinebased activation motif-containing viral proteins, including K1 of Kaposi sarcoma herpesvirus and LMP2a of Epstein-Barr virus (EBV) (26). LMP2a is a 54 kDa protein expressed during the latent cycle of EBV infection. LMP2a has been described as a functional mimic of the BCR, as LMP2a expression restores normal B cell development and survival in peripheral lymphoid organs of Rag2 Ϫ/Ϫ mice, which lack a functional BCR (27). LMP2a has also been shown to activate many of the same signaling pathways as the BCR, including Syk, Lyn, BtK, BLNK, and PI3K/Akt, independently of Ig␣/Ig␤ (28) through self-aggregation in the membrane of latently infected cells (29). These signals, coordinated by Syk activation, function to maintain viral latency (30 -33) and sustain survival (27, 34 -36) in infected cells. LMP2a also blocks normal BCR triggering by sequestering the protein-tyrosine kinases Lyn and Syk (32,33). Finally, LMP2a rescues survival of germinal center B cells with crippling mutations in the BCR (35), suggesting that LMP2a can contribute to the transformation of primary B cells (37,38). However, a direct examination of Syk activation in EBVϩ B cell lymphomagenesis, tumor survival, and proliferation has yet to be undertaken.
LMP2a expression is observed in a variety of EBV-associated malignancies, including Hodgkin lymphoma, nasopharyngeal carcinoma, AIDS-related lymphomas, and post-transplant lymphoproliferative disorder (PTLD) (39). PTLD arises in the setting of immunosuppression when the EBV-specific cytotoxic T lymphocyte response is debilitated, allowing for the proliferation and lymphomagenesis of EBV-infected B cells (40). The aims of this study are to determine whether Syk activation is critical to survival of EBVϩ B cell lymphomas in PTLD and to dissect the mechanism of the Syk survival signal.
Cell Lines-The EBVϩ B cell lines derived from the blood (JB7, JC62, MF4, VB5) or lymph nodes (AB5) of patients diagnosed with PTLD were maintained as previously described (41,42). The Jurkat human T cell line was obtained from ATCC and cultured as recommended. Cell lines were grown in a 5% CO 2 humidified 37°C incubator and cultured in RPMI 1640 media (Mediatech, Inc.) supplemented with 10% heat-inactivated FBS (Serum Source International), 50 units/ml penicillin, and 50 g/ml streptomycin (Invitrogen) unless otherwise noted.
Purification and Stimulation of Human B Cells from Peripheral Blood-Peripheral blood mononuclear cells were obtained from healthy donors by Ficoll density gradient centrifugation. B cells were isolated to Ͼ95% purity from peripheral blood mononuclear cells by negative selection using the MACS B cell Isolation Kit II (Miltenyi Biotec). Purity of isolation was determined by CD19 expression in B-enriched and B-depleted fractions. BCR cross-linking was initiated by the addition of F(ab) 2 fragments of anti-IgM and anti-IgG (final concentration of 10 g/ml).
Syk Knockdown by siRNA-Syk validated Stealth RNAi, and control medium GC oligos were obtained from Invitrogen. Syk or control oligos (200 nmol) was transfected into cells (3-5 ϫ 10 6 ) using the Amaxa nucleofector and reagents from Lonza per the manufacturer's instructions. Briefly, cells were harvested at 900 ϫ g, resuspended in supplemented nucleofection reagent (Solution V) and oligo, and pulsed with program X-001. Pulsed cells were resuspended with 500 l of prewarmed culture medium and transferred to 1 ml of pre-warmed culture medium in a 24-well plate. FITC-conjugated oligo controls indicated Ͼ95% transfection efficiency.
Cellular Proliferation Assay-Cells (0.25-0.5 ϫ 10 6 cells/ml) were plated in triplicate in 96-well flat bottom plates in serial dilutions of small molecule inhibitors R406, PD98059, U029, or LY294002 (0 -10 M) and equivalent amounts of vehicle (DMSO; 0 -1:1000) 48 h after the introduction of siRNA or after BCR cross-linking. Cells were pulsed with 0.5 Ci of [ 3 H]thymidine (PerkinElmer Life Sciences) 18 h before collection, and plates were harvested after a total 48 h in culture at 37°C. Data were converted to a percentage of the cpm incorporated by cells grown in media alone or by cells transfected with control siRNA. IC 50 values or the value at which proliferation is inhibited to 50% of media alone for each line were recorded. Cells were considered sensitive to a small molecule inhibitor if their IC 50 was within the tested range (0 -10 M).
Apoptosis Assays-Cells (0.5-1 ϫ 10 6 cells/ml) were plated in serial dilutions of small molecule inhibitors R406, PD98059, U029, LY294002, or UCF-101 (0 -10 M) or equivalent amounts of vehicle (DMSO; 0 -1:1000) 48 h after the introduction of siRNA. For signaling inhibitor studies, both drug and media were replenished after 48 h of culture at 37°C, and cells were cultured for an additional 48 h. As a positive control, cells were treated with 50 M etoposide or 2 M staurosporine for the indicated amounts of time. The percentage of apoptotic cells was determined by staining with annexin V-EGFP apoptosis detection kits (BioVision) per the manufacturer's instructions. Data were converted to fold-increase in apoptosis by normalizing to cells grown in media alone or by cells transfected with control siRNA. Caspase enzymatic activity was quantified in stimulated lysates using caspase 3 Colorimetric Assay kits (BioVision) according to the manufacturer's instructions. Mitochondrial damage was assessed in cells by DiOC 6 leakage. Briefly, cells were harvested, resuspended in FACS buffer (PBS with 0.1% fetal calf serum and 0.1% NaN 3 ) containing 400 nM DiOC 6 (Sigma), and incubated for 15 min in a 37°C water bath. Data were collected on a FACScan flow cytometry using Cell-Quest software (BD Biosciences), and loss of DiOC 6 fluorescence was analyzed using FlowJo software.
Western Blotting and Immunoprecipitation-Cells (2 ϫ 10 6 cells/ml) were cultured with small molecule inhibitors or equivalent amounts of DMSO for the indicated times or 48 h after introduction of siRNA. Cells were harvested, washed with PBS containing 1 mM sodium orthovanadate, and lysed in 1% Nonidet P-40, 0.5% deoxycholic acid phospholysis buffer (150 mM NaCl, 0.5 mM EDTA, 10 mM NaF, 2 mM PMSF, 50 g/ml aprotinin, 50 g/ml leupeptin, and 5 g/ml pepstatin). Lysates were cleared by centrifugation, and protein concentration was determined by Dc protein assay (Bio-Rad). Samples (20 -40 g/lane) were separated by SDS-PAGE and transferred to a 0.45-m nitrocellulose membrane. Membranes were blocked with nonfat dry milk in TBS/Tween then probed with antibodies of interest in either nonfat dry milk or BSA. Antibody binding was detected with HRP-conjugated secondary antibodies. Blots were developed with SuperSignal West Pico Substrate (Ther-moScientific) according to the manufacturer's instructions. For immunoprecipitation (IP), cells (10 7 /IP) were harvested untreated after 1 h of treatment with 2.5 M R406 or 48 h after the introduction of siRNA and lysed as described above. Lysate was precleared with GammaBind G-Sepharose (Amersham Biosciences) for 1 h at 4°C and immunoprecipitated with 1 g/10 7 cells of the indicated antibody overnight at 4°C. Immune complexes were pulled down with GammaBind G-Sepharose, dissociated with 2ϫ reducing sample buffer, and separated by SDS-PAGE. Gels were transferred to nitrocellulose membranes, and Western blotting was performed as described above. Densitometry was performed using ImageJ (43).

Syk Is Constitutively Active in EBVϩ PTLD-derived B Cell
Lines-To assess Syk activation in EBVϩ B cell lines derived from patients with PTLD, we analyzed the phosphorylation status of BLNK, which both directly binds and is exclusively phosphorylated by Syk (45,46). BLNK was immunoprecipitated from whole cell lysates of three EBVϩ PTLD-derived B cell lines (JC62, AB5, and JB7) with BLNK-specific antibody and analyzed by Western blot with antibodies to pBLNK. Consistent with the requirement of tonic BCR signaling for survival, we observed constitutive pBLNK in each of the three EBVϩ PTLD-derived B cell lines (Fig. 1A, middle lanes). To ensure that this constitutive BLNK activation was indeed the result of Syk signaling, we treated the EBVϩ PTLD-derived B cell lines with the small molecule R406, an ATP-competitive inhibitor of Syk (47). The constitutive activation of BLNK was completely abolished by R406 treatment (Fig. 1A, right lanes), consistent with the exclusive phosphorylation of BLNK by Syk.
Syk Promotes Proliferation of EBVϩ PTLD-derived B Cell Lines-We next asked whether Syk signaling contributes to the proliferation of EBVϩ PTLD-derived B cell lines. To measure the sensitivity of cellular proliferation to Syk inhibition, cell lines were treated with serial dilutions of R406 (Fig. 1B), and the IC 50 values were determined ( Table 1). Proliferation of five of six EBVϩ PTLD-derived B cell lines tested was markedly inhibited by R406 treatment, whereas the JB7 cell line was more resistant to the effects of Syk inhibition on cellular proliferation. Cells treated with equivalent amounts of the vehicle DMSO showed no effect on proliferation (supplemental Fig.  1A). Purified, anti-IgM and anti-IgG stimulated human B cells were used as a positive control for the ability of R406 to inhibit B cell proliferation (IC 50 value Ͻ0.156 M) ( Table 1). Conversely, R406 had no effect on the proliferation of the Jurkat T cell line (Table 1), consistent with the interpretation that R406 shows specificity for Syk and does not affect ZAP-70, a member of the Syk family of tyrosine kinases involved in T cell signaling. R406 has been reported to have minimal off-target effects on the Flt3, Jak1, Jak3, Kit, and Lck signaling pathways (8); therefore, Syk knockdown by siRNA was used to confirm Syk involvement. Syk knockdown by siRNA (Fig. 1C) significantly reduced proliferation of the JC62 and AB5 EBVϩ PTLD-derived B cell lines but not the JB7 EBVϩ PTLD-derived B cell line (Fig. 1D). The extent of the effect of Syk knockdown on proliferation correlated with the sensitivities of these lines to R406 treatment as determined by their IC 50 values. Taken together, this data suggest that constitutive Syk signaling contributes to the autonomous proliferation of EBVϩ PTLD-derived B cell lines.
Syk Augments the Survival of EBVϩ PTLD-derived B Cell Lines-The reduced proliferation of EBVϩ PTLD-derived B cell lines measured after Syk inhibition could be due to apopto-sis of EBVϩ PTLD-derived B cell lines. Indeed, apoptosis was induced in all five R406-sensitive EBVϩ PTLD-derived B cell lines but not the R406-resistant EBVϩ PTLD-derived B cell line JB7 (Fig. 2, A and B). Cells treated with equivalent amounts of the vehicle DMSO showed no effect on apoptosis (supplemental Fig. 1B). Syk knockdown by siRNA (Fig. 1C) also increased apoptosis of the R406-sensitive JC62 and AB5 cell lines, albeit to different extents, but had no affect on the R406-resistant JB7 cell line (Fig. 2C). The extent of the effect of Syk knockdown on apoptosis correlated with the sensitivities of these lines to R406 treatment as determined by their IC 50 values. Similar to the effects on cellular proliferation, pharmacologic inhibition of Syk was more effective in inducing apoptosis of EBVϩ PTLDderived B cell lines than Syk siRNA. However, delivery of siRNA by nucleofection induces significant cellular apoptosis, which complicates quantitative comparison between the two methods of Syk inhibition. Taken together, these data suggest that constitutive Syk signaling is required for survival of EBVϩ PTLDderived B cell lines.
Syk Signaling Results in PI3K/Akt and Erk MAPK Activation in EBVϩ PTLD-derived B Cell Lines-Although constitutive Syk activation is evident in all of the EBVϩ PTLD-derived B cell lines (Fig. 1A), neither proliferation nor survival of the JB7 cell line is affected by Syk inhibition (Figs. 1B and 2, A and B). Therefore, we used the R406-resistant EBVϩ PTLD-derived B cell line JB7 in comparison to the R406-sensitive EBVϩ PTLDderived B cell lines as a tool for dissecting the downstream mechanism of Syk action. Syk is known to use multiple downstream pathways to control cellular functions after BCR ligation. Of these, the PI3K/Akt, Erk MAPK, and NF-B pathways are shared between BCR and LMP2a signaling. As the p38 MAPK pathway is not activated by LMP2a, it was used as a negative control. Syk inhibition markedly decreased activation of both Akt and Erk MAPK (Fig. 3A). Signal transduction by the PI3K/Akt and Erk MAPK pathways was inhibited by low concentrations (0.156 M) of R406 (data not shown), and inhibition occurred as rapidly as 5 min after treatment (data not shown). In contrast, no change was observed in NF-B or p38 MAPK signaling in either the R406-sensitive or the R406-resistant EBVϩ PTLD-derived B cell lines. Cells treated with equivalent amounts of the vehicle DMSO showed no affect on the PI3K/ Akt or Erk MAPK pathways (data not shown). Similarly, Syk knockdown by siRNA abrogated Akt and Erk MAPK phosphorylation, whereas NF-B and p38 MAPK signaling were not affected (Fig. 3B). Taken together, these data suggest that PI3K/ M Syk inhibitor R406 as indicated for 1 h and subsequently lysed. After preclearing, BLNK was immunoprecipitated (IP) by rabbit IgG or total BLNK antibodies as indicated. An aliquot of flow-through was collected. Immunoprecipitated lysates and flow-through were separated by SDS-PAGE, transferred to nitrocellulose, and probed for pBLNK or actin, respectively, by Western blotting. Blots were then reprobed for total BLNK. Densitometry was calculated with ImageJ and is indicated numerically below each set of blots. B, cells were plated in triplicate with the indicated final concentration of the Syk inhibitor R406 for 48 h. After pulsing with [ 3 H]thymidine, plates were harvested, and data were converted to a percentage of the cpm incorporated by cells grown in media alone. Data are shown as the % [ 3 H]thymidine Ϯ S.E. for n Ն 3 independent experiments. C, control (medium GC) or Syk-specific siRNA were introduced by Amaxa nucleofection as described under "Experimental Procedures." Lysates were collected 48 h after nucleofection. Proteins were separated by SDS-PAGE, transferred to nitrocellulose, and probed for Syk. Blots were then reprobed for actin as a loading control. Densitometry was performed as in A. D, control (medium GC) or Syk-specific siRNA was introduced to cells, and proliferation assays were performed 48 after nucleofection as described above. Statistical significance from an unpaired t tests are indicated as follows: p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.0001. Data are representative of n Ն 3 independent experiments in which Syk knockdown was Ն 90%. The Syk Survival Signal Requires PI3K/Akt Activation but Not Erk MAPK Activation-Next we addressed whether the PI3K/Akt or Erk MAPK pathways participate in the Syk survival signal. Autocrine IL-10 secretion in EBVϩ PTLD-derived B cell lines is an PI3K/Akt-, but not Erk MAPK-dependent process (44) and is required for the autonomous proliferation of PTLD lines (48). Whereas the endogenously high secretion of the autocrine growth factor IL-10 was markedly decreased after Syk inhibition in our R406-sensitive, EBVϩ PTLD-derived B cell lines (Fig. 4A), the R406-resistant EBVϩ PTLD-derived B cell line JB7 displayed endogenously low IL-10 secretion (Fig.  4A).
To ask whether the PI3K/Akt or Erk MAPK pathways were involved in proliferation or survival of EBVϩ PTLD-derived B cells, we treated cells with pharmacologic inhibitors of the PI3K/Akt pathway (LY294002) or the Erk MAPK pathway (PD98059 and U029) and assayed both cellular proliferation and apoptosis. PI3K/Akt inhibition (Fig. 4B) reduced proliferation (Fig. 4C) and induced apoptosis (Fig. 4D) of the R406sensitive EBVϩ PTLD-derived B cell lines JC62 and AB5. In contrast, proliferation and survival of the R406-resistant EBVϩ PTLD-derived B cell line JB7 was not affected by PI3K/Akt inhibition (Fig. 4, C and D). Inhibition of Erk MAPK ( Fig. 4B and data not shown) resulted in no change in proliferation ( Fig. 4E and data not shown) or apoptosis ( Fig. 4F and data not shown) in any of the EBVϩ PTLD-derived B cell lines. Taken together, these data suggest that Syk drives survival of the R406-sensitive EBVϩ PTLD-derived B cell lines through the PI3K/Akt, but not the Erk MAPK, pathway.
Syk-and PI3K/Akt-mediated Signals Prevent Loss of the Caspase Inhibitor XIAP-Akt is known to directly phosphorylate the caspase inhibitor XIAP, thereby resulting in protein stabilization. We show that inhibition of Syk results in the complete loss of XIAP protein within 24 h in R406-sensitive, but not R406-resistant, EBVϩ PTLD-derived B cell lines (Fig. 5A). However, no significant changes in XIAP transcript levels were observed after R406 treatment of EBVϩ PTLD-derived B cell lines (supplemental Fig. 2). These data are consistent with regulation of XIAP at the protein rather than transcriptional level.
XIAP is known to inhibit active caspase 3 (25). Therefore, a loss in XIAP should coincide with an increase in caspase 3 activity. Consistent with the loss of XIAP, we observe increased caspase 3 activity after R406 treatment in R406-sensitive, but not R406-resistant, EBVϩ PTLD-derived B cell lines (supplemental Fig. 3A). Moreover, inhibition of Akt results in loss of XIAP (Fig. 5B, first versus fifth lanes) only in the R406-sensitive lines, consistent with the known role of Akt in XIAP stabilization (24). Thus, Akt activity is required to preserve XIAP levels in EBVϩ B lymphoma cells.
Akt has been described as preventing the loss of XIAP by two distinct mechanisms (24,49), both involving regulation of XIAP at the protein level. First, phosphorylation of XIAP Ser 87 by Akt has been reported to prevent autoubiquitination and subsequent degradation of XIAP, thus stabilizing the levels of protein expression (24). To address the differential loss of XIAP in R406-sensitive and R406-resistant EBVϩ B lymphoma cell lines, we considered the possibility that the resistant cell line, JB7, carries a mutation that prevents XIAP autoubiquitination The second mechanism involves the protease HtrA2, a homolog of the bacterial heat-inducible serine protease HtrA/DegP (50). XIAP was one of the first substrates identified for HtrA2, as increased HtrA2 expression results in decreased XIAP expression and vice versa (51,52). Akt phosphorylates HtrA2 at Ser 212 , which attenuates its serine protease activity and, consequently, its ability to degrade XIAP and induce apoptosis (49). We next directly asked if HtrA2 was involved in the loss of XIAP we observed after Syk and Akt inhibition. To address this, we used UCF-101, a small molecule inhibitor of HtrA2. Inhibition of HtrA2 did not affect basal XIAP levels (Fig. 5B, first lanes  versus second lane). HtrA2 inhibition reversed the loss of XIAP after Syk inhibition in R406-sensitive cells (Fig. 5B, third lanes  versus fourth lane). Similarly, HtrA2 inhibition reversed the loss of XIAP after PI3K/Akt inhibition in R406-sensitive cells (Fig.  5B, fifth lanes versus sixth lane). Consistent with our previous observations, XIAP levels remained stable in the R406-resistant JB7 line regardless of inhibitor treatment (Fig. 5B). One of the predominant off-target effects of UCF-101 is an increase in Erk MAPK signaling (53). However, no increase of Erk activation was observed in EBVϩ B cell lymphoma lines treated with UCF-101 (supplemental Fig. 4A). Notably, the dose of UCF-101 we utilized (10 M) is 2-5 times less than the dose used in most other reports. We also did not observe an increase in Akt signaling at the same concentration of UCF-101 (supplemental Fig. 4B). Taken together these data indicate that HtrA2 is involved in the loss of XIAP in EBVϩ PTLD-derived B cell lines.
Although HtrA2 predominately resides in the mitochondria, a small amount can be found in the cytoplasm of unstimulated cells (50). Mitochondrial damage triggers the release of HtrA2 into the cytoplasm. The variable location of HtrA2 raises the question of where XIAP and HtrA2 interact. To address this, we used DiOC 6 , a dye that stains intact mitochondria, such that loss of DiOC 6 staining indicates disruption of mitochondrial inner transmembrane potential. Syk inhibition induced a loss of DiOC 6 staining in the R406-sensitive EBVϩ PTLD-derived B cell lines (supplemental Fig. 3, B and C). Minimal DiOC 6 loss was observed in the R406-resistant line, JB7. Together, this demonstrates that Syk inhibition induces mitochondrial damage and is consistent with release of HtrA2 into the cytoplasm where it can interact with XIAP.
Finally, we sought to determine if blocking HtrA2 function rescued sensitive EBVϩ PTLD-derived B cell lines from Syk inhibition-or PI3K/Akt inhibition-induced apoptosis. EBVϩ PTLD-derived B cell lines were treated with inhibitors of Syk, PI3K/Akt, HtrA2, or a combination of inhibitors and analyzed by flow cytometry for annexin V-EGFP ϩ cells. Syk inhibitioninduced apoptosis was significantly diminished when HtrA2 function was blocked in R406-sensitive cells (Fig. 5C). Similarly, PI3K/Akt inhibition-induced apoptosis was completely prevented by blocking HtrA2 function in R406-sensitive cells (Fig.  5D). Consistent with earlier results, apoptosis was not induced in the R406-resistant JB7 line (Fig. 5, C and D). Taken together Proteins were separated by SDS-PAGE and transferred to a nitrocellulose membrane, and Western blotting was performed for (A, from top to bottom) pAkt, pErk, IB, and pP38. Blots were then reprobed for total Akt, Erk, actin, and P38, respectively. Data are representative of n 3 independent experiments. B, control (medium GC) or Syk-specific siRNA were introduced by Amaxa nucleofection as described under "Experimental Procedures." After 48 h, cells were collected and lysed. Proteins were separated by SDS-PAGE and transferred to a nitrocellulose membrane, and Western blotting was performed for (B, from top to bottom) pAkt, pErk, IB, and pP38. Blots were then reprobed for total Akt, Erk, actin, and P38, respectively. Data are representative of n Ն 3 independent experiments in which Syk knockdown was Ն 90%. Densitometry was calculated with ImageJ and is indicated numerically below each set of blots. these data indicate that HtrA2 is involved Syk-and Akt-mediated survival of EBVϩ PTLD-derived B cell lines.

DISCUSSION
Targeting tonic BCR signals with novel Syk inhibitors has been hotly pursued in the treatment of B cell lymphomas and has highlighted Syk activation as a critical checkpoint in B cell lymphoma survival. Syk controls a variety of cellular functions in addition to survival, however, including reactive oxygen production, cytokine release, and cellular adhesion. Understanding the precise mechanism of the Syk survival signal may highlight specific targets for therapeutic intervention in EBVϩ B cell lymphomas.
The EBVϩ, PTLD-derived B cell lines used in this study express two potential activators of Syk, the BCR and LMP2a. Expression of LMP2a has been reported to block the ability of the BCR to activate the Src family kinases and Syk after ligation (31,32). The EBVϩ PTLD-derived B cell lines used here are similarly unable to phosphorylate BLNK after BCR ligation (data not shown). LMP2a is known to remain constitutively aggregated and phosphorylated in the membrane of infected cells, where it co-localizes with patches of anti-phosphotyrosine activity and associates with the Src family and Syk protein-tyrosine kinases (54,55). LMP2a was more recently shown to reside in the lipid rafts and blocks BCR signaling by excluding the BCR from entering lipid rafts (33). It has been suggested that LMP2a mimics the BCR complex, competing for these protein-tyrosine kinases, and blocks the BCR from entry into lipid rafts (31,33). Taken together these studies support the idea that unlike most B cell lymphomas that achieve Syk activation through BCR tonic signaling or antigen-dependent BCR aggregation, EBVϩ B cell lymphomas activate Syk through LMP2a, the constitutively active BCR mimic encoded by EBV. However, further experiments regarding the precise mecha- Cells were then harvested and lysed. Proteins were separated by SDS-PAGE and transferred to nitrocellulose membranes, and Western blotting was performed for XIAP. Blots were then reprobed for actin as a loading control. Densitometry was calculated with ImageJ and is indicated numerically below each set of blots. C and D, cells were incubated in the presence of 2.5 M Syk inhibitor R406, 5 M PI3K inhibitor LY294002, 10 M HtrA2 inhibitor UCF-101, or an equivalent amount of DMSO as indicated for 96 h with drug and media replaced at 48 h. Cells were collected and stained for annexin V-EGFP, and the percentage of apoptotic cells was measured by flow cytometry. Data are shown as -fold increase of apoptosis normalized to UCF treatment of n 3 independent experiments. Statistical significance from an unpaired t test between R406 treated samples and samples treated with a combination of R406 and UCF-101 are indicated as follows: *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.0001. OCTOBER 28, 2011 • VOLUME 286 • NUMBER 43 nism controlling Syk activation in EBVϩ B cell lymphomas from PTLD are warranted.

Syk Drives Akt-and HtrA2-dependent Survival
Our data support a PI3K-dependent activation of Akt by Syk, but the precise molecular mechanisms of this process are unclear. Syk can directly bind the p85␣ regulatory subunit of PI3K (8). Syk can also phosphorylate tyrosines in LMP2a, creating docking sites for the SH2 domains of PI3K. Either mechanism may result in PI3K activation and subsequent Akt activation (56). Regardless of how Akt is activated, studies in Akt isoform-specific knock-out mice have revealed that Akt isoforms are not created equal, as each isoform has a distinct phenotype. Akt1 deficiency results in delayed pre-and post-natal growth, Akt2 deficiency results in defective glucose homeostasis, and Akt3 deficiency leads to defective brain development (57). Given the universal requirement of PI3K/Akt signaling for B cell lymphoma survival, the mechanism of Akt activation and the utilization of specific Akt isoforms may reveal further survival-specific therapeutic targets.
Although there is support for a general requirement for PI3K/Akt activation in Syk survival signaling, current data support the notion that different B cell lymphomas utilize unique PI3K/Akt anti-apoptotic targets involved in the intrinsic, or mitochondrial, pathway of apoptosis. In EBVϩ B cell lymphomas, loss of the caspase inhibitor XIAP is a unique and critical outcome of Syk and PI3K/Akt survival signaling. Our studies show XIAP protein, but not transcript, levels are reduced after Syk or PI3K/Akt inhibition in the R406-sensitive, but not resistant, EBVϩ PTLD-derived B cell lines. The protease HtrA2 was required for loss of XIAP and apoptosis induced by Syk and Akt inhibition. Reduction of XIAP protein levels after Syk inhibition was also observed in chronic lymphocytic leukemia (12), albeit to a markedly lower extent than in our EBVϩ PTLDderived B cell lines. In chronic lymphocytic leukemia, Syk inhibition by R406 results in the degradation of the anti-apoptotic molecule Mcl-1 (12,15). Syk inhibition by curcumin in diffuse large B cell lymphoma results in decreased expression of the anti-apoptotic molecule Bcl-x L and loss of inhibitory phosphorylation of the pro-apoptotic molecule Bad (9). The differential use of members of the intrinsic pathway of apoptosis in malignancy is not unprecedented. Resistance of malignancies to Apo2L/TRAIL-induced apoptosis also depends on unique members of the mitochondrial pathway of apoptosis (58); for example, Bcl-2 mediates resistance in neuroblastoma and glioblastoma (59), whereas Bcl-x L mediates resistance in pancreatic adenocarcinoma (60). However, the mechanisms controlling differential utilization of unique PI3K/Akt anti-apoptotic targets in Syk inhibition-induced apoptosis of B cell lymphomas are unclear.
Common among the unique PI3K/Akt anti-apoptotic targets in B cell lymphomas is their involvement in, or requirement for, mitochondrial outer-membrane permeabilization (MOMP). MOMP results in the release of apoptogenic factors like HtrA2 from the mitochondria, also known as the "point of no return" in the intrinsic pathway of apoptosis (61). Members of the Bcl-2 family, including Bad, Bcl-x L , and Mcl-1, control MOMP in response to cellular stress, like cytokine deprivation (62). XIAP acts as an important checkpoint in this process by inhibiting caspase activity; MOMP results in the degradation or inhibition of XIAP by the release of HtrA2 or Smac/Diablo, respectively. Here we provide the first data that Syk inhibition induces MOMP, as evidenced by loss of mitochondrial DiOC 6 staining in R406-sensitive EBVϩ PTLD-derived B cell lines. As for the cellular stressor initiating MOMP, loss of PI3K/Akt signaling may trigger apoptosis. Intriguingly, Akt signaling alone maintains mitochondrial membrane potential and integrity (63,64), possibly through its role in regulating Bcl-2 family members. Syk inhibition also induces growth factor withdrawal, a common cellular stressor that initiates the intrinsic pathway of apoptosis. PI3K/Akt-dependent autocrine IL-10 production (44) is required for the autonomous proliferation of EBVϩ B cell lymphomas (48), but the requirement for autocrine IL-10 production for survival has yet to be directly examined. We observed loss of both PI3K/Akt signaling and high, endogenous IL-10 secretion after Syk inhibition in all our R406-sensitive EBVϩ PTLD-derived B cell lines. Studies focusing exclusively on the involvement of IL-10 in PTLD should clarify a requirement for IL-10 secretion in the survival of EBVϩ PTLD-derived B cell lines.
The partial rescue of Syk inhibition-induced apoptosis we observe by HtrA2 inhibition suggests Syk survival signaling is not exclusively executed through the PI3K/Akt pathway. Our data clearly show that Erk MAPK signaling is not required for proliferation or survival and that p38 MAPK and NF-B signaling are not activated downstream of Syk. However, additional novel pathways controlling EBVϩ B cell PTLD survival may be revealed by knockdown of LMP2a or Syk and subsequent proteomic analysis of signaling pathways (65). Erk MAPK activation by Syk may control other Syk-driven cellular responses, separate from apoptosis.
HtrA2 inhibition prevents both loss of XIAP and apoptosis induced by PI3K/Akt inhibition in R406-sensitive EBVϩ PTLD-derived B cell lines. This strongly suggests that loss of XIAP is a crucial outcome of Syk inhibition and the primary outcome of PI3K/Akt inhibition in EBVϩ B cell lymphomas. Although phosphorylation of XIAP by Akt prevents autoubiquitination and subsequent degradation (24), autoubiquitination of XIAP was recently shown as dispensable for degradation of XIAP (66). Our data suggest instead an important role for the protease HtrA2 in the loss of XIAP as well as survival of EBVϩ PTLD-derived B cell lymphomas. This is consistent with findings that phosphorylation of HtrA2 by Akt attenuates its ability to degrade XIAP and induce apoptosis (49). However, a more detailed examination into the interplay between Akt, HtrA2, and XIAP is required.
Collectively, our data support a novel model of Syk survival signaling in EBVϩ B cell lymphomas shown in Fig. 6. In EBVϩ PTLD-derived B cell lines, Syk is activated either by the BCR or LMP2a. Syk directly activates the PI3K/Akt pathway, leading to IL-10 production. Our findings demonstrate involvement of PI3K/Akt and Erk MAPK signaling, but not NF-B signaling, downstream of Syk activation. Whereas we demonstrate a critical requirement for PI3K/Akt signaling, we also clarify that Erk MAPK is not involved in Syk survival signaling. Finally, our studies reveal the requirement of HtrA2 in the loss of XIAP and apoptosis induced by Syk and Akt inhibition.
Given the role of Syk in a wide and diverse set of cellular functions, Syk inhibition in murine models of PTLD will be crucial in determining the preclinical potential of this therapeutic strategy. Our studies highlight Akt, and for this first time, XIAP, as additional promising therapeutic targets for the treatment of EBVϩ B cell lymphomas. Small molecule, peptidic, allosteric mimetic, and substrate mimetic inhibition strategies for both Akt and XIAP are currently in development and under investigation as chemotherapeutics (67,68). These agents may constitute promising therapeutic or chemosensitizing approaches to target EBV-driven lymphomagenesis.