Spleen Tyrosine Kinase (Syk), a Novel Target of Curcumin, Is Required for B Lymphoma Growth¹

Non-Hodgkin’s lymphoma (NHL)³ is the most abundant lymphoma with extensive heterogeneity over 25 subtypes. NHL is the fifth most common cancer in the United States, however, only ∼30–60% of these cases can be cured through modern therapies such as radiation therapy, chemotherapy, or biological therapy (Lymphoma Research Foundation). Diffuse large B cell lymphoma (DLBCL) is the most common of all the NHL accounting for 30% of all the cases. DLBCLs arise mainly due to chromosomal translocation of Bcl-6, a transcriptional repressor (1).

Curcumin (diferuloylmethane) is a natural yellow pigment isolated from the plant Curcuma longa (2, 3). Curcumin has been shown in recent studies to have therapeutic potential in the treatment of cancers, diabetes, arthritis, and osteoporosis (4, 5, 6, 7). Curcumin has exhibited proapoptotic and antiproliferative properties in vitro and has been able to suppress tumors in vivo on the skin, fore stomach, breast, and colon (5, 8). It was shown that in non-small cell lung carcinoma cells, curcumin down-regulated NF-κB, a potent antiapoptotic factor, and led to growth arrest of these cells (9). Recently, curcumin has been shown to cause apoptosis of multiple myeloma cells via inhibition of NF-κB and STAT3 activation (10, 11).

Previously, we have shown that low doses of curcumin inhibit the proliferation of murine B lymphoma, BKS-2, but not that of normal B cells (12). The effect of curcumin was partially reversed by stimulation with CpG oligonucleotides or by ligation of CD40. NF-κB, a potent survival factor, was shown to be almost completely down-regulated by curcumin. Curcumin has also been shown to cause growth arrest of EBV-transformed lymphoblastoid cell lines (13). The initial in vitro studies in two murine B lymphoma cell lines suggest that curcumin may have a potential therapeutic benefit in treating human B lymphomas. However, the biochemical mechanisms underlying curcumin-induced B lymphoma growth inhibition have not yet been determined.

Curcumin has been shown to down-regulate Akt in prostate cancer cells (14). Akt, which is a serine/threonine kinase, regulates cell survival and growth through a PI3K-dependent manner. One of the mechanisms by which Akt promotes cell survival is through phosphorylation of a proapoptotic protein Bad. Phosphorylated Bad loses its proapoptotic function (15, 16, 17). Akt is activated by upstream kinases like spleen tyrosine kinase (Syk) and PI3K in response to BCR cross-linking and growth factor stimulation (18, 19, 20, 21, 22). Curcumin has also been shown to down-regulate Bcl-x_L in prostate cancer cells grown in vitro (23). Bcl-x_L, an antiapoptotic protein, is overexpressed in many cancer cells in an NF-κB-dependent manner. Although much work has been done about the effects of curcumin on the NF-κB pathway, the direct targets of curcumin so far have not been identified (24).

In this study, we investigated the ability of curcumin to inhibit the growth of B-lymphomas of both murine and human origin (DLBCL). The main focus of the present study is to explore the effects of curcumin in multiple B lymphoma cells and understand the pathways that are upstream of NF-κB critically modulated by curcumin including Akt and Syk. We report that curcumin inhibits B lymphoma growth in vitro and in vivo. Although curcumin inhibited the NF-κB pathway and its upstream regulator Akt in B lymphoma cells, we made the novel observation that Akt is not a direct target of curcumin. Instead, we show here for the first time that curcumin directly inhibits activation of Syk in vitro as well as in vivo, an enzyme essential for BCR signal transduction. Moreover, we find that Syk is constitutively activated in a variety of B lymphoma cell lines and primary B lymphoma samples but not in primary B cells. Thus, we demonstrate a novel growth promoting role for Syk, which has been considered a tumor suppressor in certain nonlymphoid cancers.

Curcumin (diferuloylmethane) obtained from Sigma-Aldrich was dissolved in DMSO and was then diluted to 2 mM in culture medium as needed. Phosphospecific Abs against Syk (525/526), Lyn, Akt (Ser⁴⁷³), glycogen synthase kinase-3 (GSK3β) and Bad, and Ab against Akt were obtained from Cell Signaling Technologies. Ab to Bcl-x was obtained from BD Pharmingen. Monoclonal anti-β-actin Ab was obtained from Sigma- Aldrich. Caspase inhibitor ZVAD-fmk was obtained from Kamiya Biomedical. Substrates for immunoprecipitation studies including recombinant Akt and GSK-3β were obtained from Upstate Biotechnology. Immobilized anti-mammalian target of rapamycin (mTOR) affinity-purified Ab was obtained from Bethyl Laboratories.

The panel of B cell lymphomas included cells of murine and human origin. Of murine origin were BKS-2, CH31, WEHI-231, and the WEHI-231 Bcl-x_L; those of human origin included Ramos and the human DLBCL cell lines OCI-Ly7 and OCI-Ly10. The immature B lymphoma cell line BKS-2 was isolated and maintained in vivo as a splenic tumor in our laboratory (25). Female CBA/N (Xid) mice were obtained from The Jackson Laboratory. Mice were kept in microisolator cages in our American Association for Laboratory Animal Accreditation and Certification-approved rodent facility. BKS-2 B lymphoma cells obtained from the spleens of CBA/N mice and normal spleen cells were depleted of T cells with a mixture of anti-T cell Abs and complement as described (26). B cells were isolated from human PBMC by previously described procedures (26, 27, 28).

WEHI-231 and CH31 cells were cultured in RPMI 1640 (Invitrogen Life Technologies) supplemented with 10% FCS. DLBCL cell lines were cultured in IMDM plus 20% normal human plasma. To measure proliferation, 2 × 10⁴ cells were cultured in 200 μl of medium supplemented with 10% FCS. Cultures were treated with varying doses of curcumin or an equivalent concentration of DMSO. After 44 h, the cultures were pulsed for 4 h with 1 μCi [³H] thymidine (DuPont/NEN) and the cells were harvested onto filter mats using a cell harvester (Packard Instrument). The levels of radionucleotide incorporation were measured with a Matrix 96 β-radiation counter (Packard Instrument). Results were presented as arithmetic mean of triplicate cultures ± SE and statistical significance of different treatments was evaluated by the Student t test.

B lymphoma cells were treated with curcumin at concentrations from 2.5 to 20 μM in 1 ml cultures of 1 × 10⁶ cells in 6-well plates (Corning/Costar). Cell lysates were prepared in 1× SDS sample buffer or 1% Triton X-100 as described earlier and were subjected to SDS-PAGE and Western blot analysis (26). The Western blots were analyzed by probing the membrane using various primary Abs followed by HRP-conjugated secondary Abs (Santa Cruz Biotechnology). The blots were developed with Pico Chemiluminescence substrate (Pierce Biotechnology) and exposed to Kodak X-O mat film which was scanned with a flat-bed scanner (UMAX Technologies). Alternatively, the blots were scanned by a Kodak Image Station 2000RT (Eastman Kodak). For reprobing, membranes were stripped using a solution containing 62.5 mM Tris-HCl, 2% SDS, and 100 mM 2-ME at 65°C for 20 min. For flow cytometry, 1 × 10⁶ cells were stained with B220-PE and Annexin V^FITC (BD Pharmingen).

For immunoprecipitation, cell lysates were precleared for 1 h at 4°C by protein G-Sepharose (Sigma-Aldrich) and then incubated with Abs to Akt, mTOR, Syk, or Igα. For Akt kinase assays, immunoprecipitated Akt or mTOR was incubated with 10 μM ATP, rGSK3β (direct substrate for Akt), or rAkt (for mTOR-mediated Akt phosphorylation) and then the reaction was stopped after 30 min by adding 1× SDS sample buffer. For Syk kinase assay, immunoprecipitated Syk was incubated with ATP and different doses of curcumin or vehicle for 30 min at room temperature and then the reaction was stopped by adding 1× SDS sample buffer. Western blotting of the lysates was performed as described earlier in this section.

The cell cycle status and apoptosis were analyzed using propidium iodide (PI) and annexin V staining. Cultured B lymphoma cells (1 × 10⁶/2 ml) were treated with varying doses of curcumin and then fixed in ethanol for at least 1 h at 4°C, after which cells were incubated in a mixture of 1 μg/ml PI (Sigma-Aldrich) and 25 μg/ml RNase A (Sigma-Aldrich) at 37°C for at least 30 min. The level of PI fluorescence was measured with a MoFlo flow cytometer (DakoCytomation). Cells in the sub-G₁ region were considered to be apoptotic.

For myristoylated Akt (MyrAkt) overexpression studies, WEHI-231 B lymphoma cells were transfected with GFP alone (pEGFP-N1) or cotransfected with GFP and MyrAkt expression vector (29) by electroporation at 250 mV, 960 μF, and 200 Ohm with a Gene Pulser electroporator (Bio-Rad). Two days later, cells were treated with 5 and 10 μM curcumin. Forty-eight hours after treatment, cells were analyzed for apoptosis by Annexin V staining. For RNA interference studies, lymphoma cells were cotransfected with a GFP plasmid and control or Syk specific siRNA (Dharmacon). Forty-eight hours after transfection, cells were treated with curcumin and were analyzed for cell cycle 48 h after treatment by Hoechst staining.

Female CBA/N mice were administered i.v. with 10 × 10⁶ BKS-2 B lymphoma cells on day 0. From day 1, mice were injected i.p. either with 40 mg/kg body weight curcumin or the vehicle (DMSO) alternate days for 10 days. On day 11, mice were sacrificed and the number of nucleated cells was determined.

Previously, we reported that curcumin inhibited the growth of murine B lymphoma BKS-2. Here, we tested the effect of curcumin on multiple cell lines including those of human origin to study the importance of curcumin as a potential antilymphoma agent and to identify its intracellular target. We demonstrate here that curcumin causes a dose-dependent decrease in proliferation of WEHI-231 (mouse B lymphoma) and OCI-Ly-10 (human DLBCL) cells as measured by [³H]thymidine incorporation. DMSO is used as vehicle control and the concentration of DMSO used here has no toxicity for these cells (Fig. 1, A and B). Similarly, growth of DLBCL cells OCI-Ly-7 and SUDHL-4 was also inhibited by 10 μM curcumin by 90% (data not shown). We investigated whether the decrease in proliferation was due to growth arrest, apoptosis, or both. Cell cycle analysis performed by staining DNA with PI demonstrated an increase in sub-G₁ population, an indicator of apoptotic cells and a decrease in cells in the S phase of cell cycle (Fig. 1,C for OCI-Ly10 and Table I for WEHI-231). The apoptotic population increased from 11 ± 1% to 23 ± 2% in Ly-10 B lymphoma cells (DLBCL) when treated with 10 μM curcumin for 48 h whereas the cells in S phase decreased from 54 ± 2% to 26 ± 3%. Because the sub-G₁ population identified by PI staining could represent apoptotic or necrotic cells, we performed annexin V staining, which distinguishes apoptotic and necrotic cells. Apoptotic population increased from 16 ± 2% to 74 ± 4% in WEHI-231 B lymphoma cells and from 4 ± 1% to 16 ± 1% in OCI-Ly10 B lymphoma cells treated with 10 μM curcumin (Tables II and III). Thus, curcumin inhibits the growth and survival of both murine and human B lymphoma cells. The effects of curcumin are not limited to cell lines, since curcumin induced a dose-dependent decrease in proliferation of primary human malignant Burkitt’s cells (Fig. 1 D).

Caspases are cysteine proteases which upon activation can induce apoptosis (30). Using thymidine incorporation and cell cycle analysis, we show here that curcumin-induced apoptosis is caspase dependent. At 10 μM concentration of a pan-caspase inhibitor, Z-VAD-fmk, the proliferation of OCI-Ly10 DLBCL lymphoma cells was not inhibited by curcumin. PI analysis revealed that curcumin-induced apoptosis was inhibited by the caspase inhibitor (Fig. 1,E). Moreover, curcumin-induced cell cycle arrest was also reversed by ZVAD-fmk (Fig. 1, E and F) suggesting that cell cycle arrest is a secondary consequence of curcumin-induced apoptosis. Moreover, we demonstrate that curcumin-induced apoptosis is caspase dependent because poly(ADP-ribose) polymerase (PARP), a substrate of caspase, is cleaved in the presence of curcumin in addition to caspase-9 cleavage (Fig. 1 G).

Expression of Bcl-x_L, an antiapoptotic factor overexpressed in many B lymphomas, is regulated by NF-κB. Curcumin has been shown to inhibit the NF-κB pathway in many cancer cells including multiple myeloma (11, 12). Accordingly, we found that Bcl-x_L is down-regulated in OCI-Ly10 and WEHI-231 cell lines upon treatment with curcumin suggesting that curcumin targets the NF-κB pathway in B lymphoma cells (Fig. 2, A and B). Then, we determined whether overexpression of Bcl-x_L, an NF-κB target gene, overcomes curcumin-induced apoptosis in B lymphoma cells. Using a WEHI-231 cell line, which is stably transfected with Bcl-x_L (31), we observed no difference in apoptosis between untreated and curcumin-treated cells. But WEHI-231 cells which express normal levels of Bcl-x_L were susceptible to curcumin-induced apoptosis (Fig. 2 C). Thus, these results demonstrate that down-regulation of Bcl-x_L is critical for apoptosis of B lymphoma induced by curcumin and its overexpression rescues cells from apoptosis.

Akt is a serine/threonine kinase which critically regulates cell survival and cell growth in many cell types (17, 32). Akt is activated by phosphorylation at serine 473 and threonine 308 by upstream kinases. Activated Akt has been shown to phosphorylate at least 20 different substrates within the cell including activation of key survival proteins, inactivation of proapoptotic proteins like Bad, and activation of key enzymes like GSK-3β (15). Curcumin and its analog have been shown to inhibit phosphorylation of Akt in prostate cancer cells but the effect of curcumin on Akt in most cancers susceptible to curcumin so far has not been investigated (14). We found that curcumin inhibited phosphorylation of Akt at serine 473 (pAkt) as early as 30 min in a dose-dependent manner and the Akt inhibition persisted even at 24 h (Fig. 3 A). Similar results were obtained when an Ab to phosphothreonine 308 was used to determine Akt phosphorylation (data not shown). Total Akt levels were not changed by curcumin treatment.

Bad is a proapoptotic protein which binds to antiapoptotic proteins like Bcl-2 and Bcl-x_L and inhibits their prosurvival activity. Akt is known to promote cell survival by phosphorylation of Bad and its subsequent degradation by the proteasomes (17). Western Blot analysis showed that the levels of phosphorylated Bad (pBad) were almost undetectable in OCI-Ly10 (DLBCL) cells treated with curcumin supporting the notion that down-regulation of Akt phosphorylation also leads to a reduction in its activity (Fig. 3,B). Akt phosphorylates GSK3β, another substrate, whose activity is inhibited by phosphorylation. GSK3β facilitates glucose to glycogen conversion in the cells. There is a dose- and time-dependent decrease in phosphorylation of GSK-3β in curcumin-treated OCI-Ly10 cells further confirming that curcumin inhibited Akt activity (Fig. 3 C).

Because curcumin inhibits Akt phosphorylation, we sought to determine whether introduction of a constitutively active form of Akt (MyrAkt) can overcome curcumin-induced apoptosis in B lymphoma cells. MyrAkt has been used extensively as a way to overexpress active Akt (33, 34, 35). In the presence of 5 and 10 μM concentrations of curcumin, WEHI-231 B lymphoma cells transiently transfected with MyrAkt are resistant to apoptosis compared with cells transfected with vector alone at both 24 and 48 h time points (Fig. 4,A or Tables II and Table IV). Expression of MyrAkt was confirmed by Western blot analysis (Fig. 4 B). These findings demonstrate that inhibition of Akt pathway is critical for curcumin to induce apoptosis in B lymphoma cells.

To see whether Akt is a direct target of curcumin, we performed an in vitro kinase assay using GSK3β as substrate and immunoprecipitated Akt. Surprisingly, curcumin did not inhibit Akt-mediated GSK3β phosphorylation in vitro suggesting that it acts upstream of Akt (Fig. 5,A). Preliminary experiments suggested that curcumin did not affect the activity of phosphoinositide-dependent kinase-1 (PDK1) also, an upstream regulator of Akt (data not shown). Recently, it has been shown that mTOR-rictor complex is involved in Akt activation (36). Using immobilized anti-mTOR Ab to isolate mTOR and its associated complex from B lymphoma cells, we studied mTOR-mediated phosphorylation of rAkt in vitro and found that curcumin had no effect in this assay also (Fig. 5,B). In B cells, Akt activation has been shown to be dependent on Syk, as Syk-negative B lymphoma cells have a dramatic reduction in Akt phosphorylation (22, 37, 38). Activated Syk has been shown to activate PI3K either directly or indirectly through adaptor proteins like B cell adaptor protein (BCAP) (39). Active PI3K generates phosphatidylinositol 3,4,5-trisphosphate, a lipid necessary for membrane recruitment of pleckstrin homology (PH) domain containing proteins like Akt. Membrane recruitment of Akt is absolutely essential for its activation by kinases like PDK1 and integrin linked kinase-1 (ILK1) and for its autophosphorylation (36, 40). Syk has been shown to become activated upon phosphorylation of Tyr⁵²⁵ and Tyr⁵²⁶ (41). Hence, we hypothesized that the inhibitory effect of curcumin on Akt could be due to its ability to prevent Syk activation. Consistent with our hypothesis, curcumin induced a dose- and time-dependent reduction in Syk phosphorylation in several B lymphoma cell lines of both murine (WEHI-231) and human origin (LY-10, Ramos) (Fig. 5,C). Because Syk appears to be required for B lymphoma growth, we questioned whether it is constitutively active in B lymphoma cells. Accordingly, constitutive Syk phosphorylation was observed in both primary B lymphoma cells and B lymphoma cell lines with little or no phosphorylation in resting splenic B cells and PBMC B cells from healthy donors (Fig. 5, C and D).

Because BCR cross-linking leads to activation of Syk in a Lyn- and Igα-dependent manner, we investigated whether curcumin-mediated down-regulation of Syk is due to targeting of Lyn or Igα activity. Interestingly, curcumin did not inhibit phosphorylation of Lyn as shown in Fig. 5,E. Moreover, the phosphorylation of the tyrosine motifs of the Igα cytoplasmic tail remains unaffected in the presence of varying concentrations of curcumin in DLBCL cells (Fig. 5 F).

To test whether Syk is a direct target of curcumin, we immunoprecipitated Syk and measured its autophosphorylation activity by probing for tyrosine-phosphorylated Syk using a p-tyr Ab. As shown in Fig. 5,G, indeed curcumin strongly inhibited in vitro kinase activity of Syk immunoprecipitated from the DLBCL line Ly10 (Fig. 5,G) or from the mouse B lymphoma line CH31 in a dose-dependent manner (data not shown). When BKS-2 (mouse) and OCI-Ly10 (human DLBCL) B lymphoma cell lines were treated with piceatannol, a Syk inhibitor, there was a dose-dependent reduction in the basal proliferation compared with vehicle-treated cells (Fig. 6,A). Partially knocking down Syk with a specific siRNA decreases the growth of SUDHL-4 (DLBCL) B lymphoma cells accompanied by a decrease in the phosphorylation of Akt and a decrease in Bcl-x_L, an NF-κB target gene (Fig. 6 B). These results demonstrate that constitutively active Syk is required for B lymphoma growth and that blocking its activity by multiple agents including curcumin, piceatannol, and siRNA decreases lymphoma growth.

Because all the studies on the growth inhibitory effects on B lymphoma have been in vitro, we investigated the in vivo effectiveness of curcumin using our mouse model of B lymphoma, BKS-2. There was a significant reduction in the tumor burden of curcumin-treated but not vehicle (DMSO)-treated mice carrying the BKS-2 B lymphoma as measured by the total splenocyte numbers (p < 0.01) (Fig. 7,A). Representative spleens from both vehicle and curcumin-treated mice are shown in Fig. 7 B.

NHL poses a serious challenge in terms of cure and recovery. Current therapies include chemotherapy and a mAb that targets CD20 Ag expressed by B cells leading to depletion of B cells from the patients (42, 43). Naturally occurring compounds derived from plants have been long shown to have antitumorigenic and anti-inflammatory properties. Curcumin, a spice used in Asian foods, has strong anticancer properties both in vitro and in vivo with little or no toxicity to normal cells at a wide range of doses (2, 44). We report here that curcumin induces a dose- and time-dependent apoptosis in B lymphoma cells of both murine and human origin in vitro and murine B lymphoma in vivo.

Previous studies from our laboratory demonstrated that curcumin induced both apoptosis and growth arrest (12). Current studies using caspase inhibitor and the Bcl-x_L overexpressing cell line suggests that apoptosis is the primary mechanism of curcumin-induced growth inhibition. The proximal signaling pathways that are targeted by curcumin are poorly understood. One of the well-characterized targets for curcumin is the NF-κB pathway (10). Curcumin was shown to inhibit IκB kinase (IKK), an upstream component of the NF-κB pathway. Here, we showed that in B lymphoma cells curcumin inhibited the NF-κB pathway as evident by down-regulation of Bcl-x_L, an antiapoptotic protein regulated by NF-κB. Moreover, overexpression of Bcl-x_L partially overcame curcumin-induced apoptosis in B lymphoma cells.

The upstream signaling components that regulate NF-κB pathways are fairly well-characterized. Two such components are Akt and Syk. Akt is a serine/threonine kinase which was shown to promote survival and proliferation of a variety of cell types by modulating the activity or expression of a number of downstream effectors including Bad, p27, caspase-9, and Bcl-x_L. Akt is shown to be dysregulated in a variety of tumors due to its overexpression, mutation of the oncogene Ras and inactivation of a key phosphatase, phosphatase and tensin homolog deleted on chromosome 10 (PTEN) (45). Two independent studies demonstrated that Akt regulates NF-κB in an IKK-dependent manner in cells treated with platelet-derived growth factor (PDGF) or TNF-α (46, 47). We observed a dose-dependent reduction in constitutive Akt phosphorylation in B lymphoma cells treated with curcumin. Inhibition of Akt is essential for curcumin-mediated growth inhibition of B lymphoma as overexpression of a constitutive form of Akt (MyrAkt) overcomes curcumin-induced apoptosis of lymphoma cells. This is consistent with a previous report where curcumin was shown to block TNF-induced Akt activation in a myeloid leukemic cell line (48). Moreover, we showed that phosphorylation of Bad and GSK-3β, two well-defined substrates of Akt were down-regulated in curcumin-treated B lymphoma cells. Thus, curcumin induces apoptosis of B lymphoma cells by inhibiting activation of a key survival kinase Akt.

Although curcumin-induced apoptosis is Akt dependent, surprisingly, curcumin did not directly inhibit Akt (in vitro kinase assay) or its recently identified key regulator mTOR/rictor complex. So, we hypothesized that curcumin-mediated inhibition of Akt is indirect. A variety of cytoplasmic tyrosine kinases are activated in response to BCR cross-linking and growth factor stimulation in B cells. Syk is a nonreceptor tyrosine kinase expressed by all hemopoietic cells including B, T, and NK cells (49). In the absence of Syk, immature B cells can be detected in the T cell zones of the spleen but fail to undergo maturation into recirculating B cells (50, 51). These results suggest that Syk is an essential transducer of BCR signals required for the transition of immature into mature recirculating B cells. Syk is shown to regulate Akt activation through activation of PI3K (20, 22, 52, 53). PI3K generates phosphatidylinositol 3,4,5-trisphosphate, a lipid necessary for membrane recruitment of Akt. Presently, the role of Syk in B lymphoma growth is not known. Curcumin inhibited constitutive Syk phosphorylation of both murine and human B lymphoma cell lines in a dose-dependent manner and this process is direct as shown by the in vitro kinase assay. Previous studies suggested the IKK complex as a target for curcumin in cancer cells (10, 11, 54). But none of these studies showed that curcumin directly affects the target as in these experiments lysates from curcumin-treated cells were analyzed for the target protein change. Hence, it is unclear whether IKK is inhibited directly by curcumin or was inhibited due to the effect of curcumin on an upstream component. It has been shown that IKK is phosphorylated and NF-κB is activated in an Syk- and Akt-dependent manner in response to BCR cross-linking, TNF stimulation, and oxidative stress (22, 37, 38, 46, 47, 55, 56). These findings suggest a critical link between the upstream kinases and the NF-κB pathway. For the first time, here we demonstrate under in vitro conditions, curcumin can directly inhibit phosphorylation of Syk and thus identify Syk as a novel target for curcumin in B lymphoma cells. Consistent with the concept that Syk is a direct target of curcumin, Lyn activity and Igα phosphorylation, which are upstream of Syk activation, were not affected by curcumin in B lymphoma cells.

Curcumin-mediated growth inhibition of B lymphoma cells via Syk led us to hypothesize that Syk may be constitutively active in primary lymphoma cells. Accordingly, we showed that a variety of B lymphoma cells (cell lines and primary tumor from Eμ-Myc-transgenic mice) expressed constitutively active Syk with little or no active Syk in normal splenic and PBMC B cells. Also, we found that Syk is constitutively active in primary B lymphoma cells from human patients (67). Moreover, for the first time we demonstrate that the constitutively activated Syk is required for B lymphoma growth. Syk induces multiple signaling pathways including PI3K, NF-κB, phospholipase C-γ, and JNK (20, 57, 58). Curcumin has been shown to target NF-κB, Akt, JNK, JAK, STAT3, and Src kinases in tumor cells. In multiple myeloma cells, it was shown that IL-6-induced STAT3 phosphorylation is blocked by curcumin resulting in growth inhibition (10, 11). Similar observation was made in brain microglia (59). In addition, curcumin was shown to target Src activity in vitro in murine fibroblast cell lines (60). In light of these findings, we propose that curcumin-mediated targeting of some of the above-mentioned pathways is Syk dependent in B lymphoma cells. Thus, by targeting Syk, all downstream components are down-regulated which is a potential way to modulate pathways that are diversified from the proximal components. Using a Syk-selective inhibitor and by RNA interference, we also demonstrated that constitutive Syk activity is critical for B lymphoma growth. Unlike breast cancer cells, where Syk was shown to be a tumor suppressor, we show here that Syk has a prosurvival role in B lymphoma suggesting multiple roles for Syk depending on the cellular context. Loss of Syk expression in breast tumors as a result of DNA hypermethylation promotes tumor cell proliferation and invasion and predicts shorter survival of breast cancer patients (61, 62, 63, 64).

Presently, we do not know the mechanisms that lead to constitutive activation of Syk in B lymphoma cells. One possible mechanism is that lymphoma cells secrete a cytokine, which can activate Syk. Another possibility is suggested by the link between oxidative stress and Syk activation observed in many cell types including B cells. Stimulation of a B cell line with hydrogen peroxide led to an increase in Syk phosphorylation followed by Akt activation and protection of cells against oxidative stress-induced apoptosis (37, 55, 65). Although our studies suggest that Syk is a direct target of curcumin, it will be interesting to see whether curcumin also affects reactive oxygen species production and thereby modulates Syk activity in an indirect manner. Another possibility is that BCR expressed on lymphoma cells constitutively activates Syk, because a prosurvival function for BCR has been demonstrated in normal B cells (66). Consistent with this hypothesis, we observed a significant block in B lymphoma growth when BCR expression was down-regulated by knocking down Igα or Igβ expression. This was accompanied by a decrease in Syk phosphorylation (67).

In summary, our data demonstrate for the first time that B lymphoma cells express Syk constitutively and that targeting Syk with curcumin-induced growth arrest and apoptosis in vitro and in vivo. Our observations indicate that curcumin modifies key signaling pathways including Syk, Akt, and NF-κB. Overall, these findings suggest that curcumin may have some important therapeutic benefits in the treatment of B lymphomas as there is a significant reduction in splenic tumor burden in our in vivo mouse model.

We thank Jennifer Strange and Greg Bauman for performing flow cytometry. We also thank Drs. Richard May and Petr Starostik for OCI-Ly7 and OCI-Ly10 cells and Dr. Gabriel Nunez for WEHI-231 Bcl-x_L cells.

The authors have no financial conflict of interest.