Kinase Activity is Required for Growth Regulation but not Invasion Suppression by Syk Kinase in Pancreatic Adenocarcinoma Cells

Syk (spleen tyrosine kinase) is a nonreceptor tyrosine kinase containing two tandem amino-terminal SH2 domains, followed by an extended linker region and a carboxyterminal kinase domain (Sada et al., 2001). Tyrosine-352 (Y352) in the linker region is trans-phosphorylated by src family members, promoting the activation of syk (Kimura et al., 1996; Sada et al., 2001). Activation further involves the autophosphorylation of syk on tyrosines 525 and 526 (YY525/6) in the activation loop of the kinase domain, which promotes substrate-specific catalytic activity and is required for signaling by syk (Sada et al., 2001; Zhang et al., 1998); this modification is thus indicative of the functional enzyme. The active form of syk then localizes to appropriate substrates or bridging molecules primarily through interactions with its SH2 domains and linker tyrosines (Kimura et al., 1996; Sada et al., 2001).


Introduction
Syk (spleen tyrosine kinase) is a nonreceptor tyrosine kinase containing two tandem amino-terminal SH2 domains, followed by an extended linker region and a carboxyterminal kinase domain (Sada et al., 2001).  in the linker region is trans-phosphorylated by src family members, promoting the activation of syk (Kimura et al., 1996;Sada et al., 2001).Activation further involves the autophosphorylation of syk on tyrosines 525 and 526 (YY525/6) in the activation loop of the kinase domain, which promotes substrate-specific catalytic activity and is required for signaling by syk (Sada et al., 2001;Zhang et al., 1998); this modification is thus indicative of the functional enzyme.The active form of syk then localizes to appropriate substrates or bridging molecules primarily through interactions with its SH2 domains and linker tyrosines (Kimura et al., 1996;Sada et al., 2001).
Syk has been identified as a putative tumor suppressor in human breast cancer since reintroduction of syk retarded the growth of syk-negative breast cancer cells, and suppression of endogenous syk enhanced the tumorigenic phenotype of the resulting cells (Coopman et al., 2000).Loss of syk correlates with poor survival and metastasis of breast cancer in patients (Toyama et al., 2003), and syk regulates breast cancer cell mitosis (Zyss et al., 2005) and transcription (Wang et al., 2005).Previously we identified syk as being a pancreatic ductal adenocarcinoma (PDAC) tumor suppressor.Syk is uniformly expressed by normal pancreatic ductal epithelium and well-differentiated (grade1; G1) PDAC; however moderately-differentiated (grade2; G2) PDAC demonstrates progressive loss of syk, and high-grade (grade3; G3), poorly-differentiated lesions are essentially devoid of syk in situ (Layton et al., 2009).In fact, syk expression is a strong positive indicator of patient survival (Layton et al., 2009).Mechanistically, we demonstrated that syk is a central mediator of phenotypic changes regulating PDAC progression, including anchorage-independent growth, cellular invasion, and gene expression changes responsible for epithelialmesenchymal transition (Layton et al., 2009).To assess the utilization of syk in pancreatic ductal cells, we examined phosphorylation states of syk that correspond to upstream www.intechopen.comPancreatic Cancer -Molecular Mechanism and Targets 104 activation (tyrosine-352) and catalytic activity (tyrosines-525/6).We further examined the activity of syk and the requirement for syk kinase activity in regulating cell growth and invasion.Herein we show that although blockade of syk activity suppressed growth of endogenously syk-positive PDAC cells in vitro, kinase activity is not required for sykdependent regulation of PDAC cell invasion.Phosphorylation of tyrosine-352 and tyrosines-525/6 is detectable in pancreatic ductal epithelial cells in situ and in vitro, and inhibition of syk kinase activity specifically retards the growth of endogenously syk-positive PDAC cells in vitro.In contrast to our previous demonstration that syk regulates the invasion of v3positive Panc1 PDAC cells by attenuation of the matrix metalloproteinase-2 axis (Layton et al., 2009), invasion of v3-negative MIAPaCa2 and BxPC3 PDAC cells is metalloproteinaseindependent and involves the urokinase/plasminogen system, which is regulated by syk in these cells.Thus we demonstrate that endogenous syk is active in pancreatic ductal epithelial cells, and that syk kinase activity is required for growth regulation, but not invasion suppression, by syk in this cell type.
MIAPaCa2 cells were transfected using Lipofectamine2000 (Invitrogen, Carlsbad, CA).BxPC3 cells were electroporated as described previously (Layton et al., 2009).Stable populations were hygromycin-selected and assessed for protein expression.For transient studies, pEF4-LacZ reporter (Invitrogen, Carlsbad, CA) was cotransfected, and cells stained with x-gal to identify and specifically quantitate transfected cells; average MIAPaCa2 transfection efficiency >95%.In all cases, serum-free (SF-) media consisted of all components except serum, as appropriate for the cell line, supplemented with 0.5% bovine serum albumin (BSA).

Immunohistochemistry
Patient tissue samples were obtained under approved UCSD Institutional Review Board protocol from the UCSD Dept. of Pathology archives and stained essentially as described previously (Layton et al., 2009).Briefly, samples were deparaffinized with xylenes, rehydrated through sequential alcohols (100%, 95%, 70%, 50%), and incubated with 1% H 2 O 2 to inactivate endogenous peroxidases.Slides were quenched with 50 mM glycine, and blocked with 2% horse serum/5% BSA/phosphate-buffered saline (PBS), pH7.4,before renaturing for 20min in a steamer using Target Retrieval Solution (DAKO North America; Carpinteria, CA).Slides were allowed to cool and then incubated with the appropriate primary antibody overnight at 4˚C.Slides were washed and biotinylated-anti-rabbit applied according to the VectaStain Elite ABC Kit (Vector Labs; Burlingame, CA, USA).Sections were developed with DAB, counterstained with hematoxylin, dehydrated through sequential alcohols and mounted.Brightfield images were acquired on a Nikon TE600 microscope with a Model 3.2.0CCD camera (Diagnostic Instruments, Sterling Heights, MI, USA) using SpotBasic software at the Moores UCSD Cancer Center.

Flow cytometry (FACS)
FACS was performed on a FACScalibur (BD Biosciences, Bedford, MA) at the Moores UCSD Cancer Center Flow Cytometry Shared Resource as described previously (Chen et al., 2010).Cells were harvested with 0.1% trypsin/versene, inactivated with 0.1% soybean trypsin inhibitor and resuspended in FACS buffer (1mM MgCl 2 , 1mM CaCl 2 , 0.1% NaN 3 0.5% BSA in PBS pH 7.4) before sequential labelling with primary and FITC-conjugated secondary antibodies.Gates were set with secondary alone, and 5μg/ml propidium iodide was included to exclude dead and dying cells.

Proliferation assays
Cell growth was assessed with the CellTiter96 Aqueous One Solution Cell Proliferation Assay (MTS) kit (Promega, Madison, WI), or as follows: cells (5x10 2 /well) were seeded into a 48-well plate.After 24 hours (and every 72 hours thereafter), fresh growth medium was replaced and the initial time point fixed with 1% paraformaldehyde/PBS, pH7.4.Additional triplicate wells were fixed at 24 hour intervals.All wells were stained with 1% Crystal Violet, which was subsequently extracted with 10% acetic acid, quantitated at 550nm and 106 compared to a standard curve of cells.For proliferation assays in the presence of the syk inhibitor, piceatannol, a titration of cells was plated and allowed to adhere for 24 hours before fresh growth medium (full serum) containing the indicated concentrations of piceatannol or DMSO vehicle was replaced.Medium was replenished in the same manner every other day for a total of 3 treatments, and the cells were fixed and stained with Crystal Violet 24h after the last treatment.Dye was extracted with 10% acetic acid and quantitated at 550nm.

Anchorage-independent growth
Anchorage-independent growth was assessed as described previously (Layton et al., 2009).Briefly, a top layer containing 5x10 3 cells in 0.5% agar/DMEM/10%FBS was seeded onto a base layer of 0.7% agar/DMEM containing 10% FBS in a 6 well plate.Cultures were incubated at 37°C, media was replaced every 3rd day, and the assay stopped on day 10.Cultures were stained with 0.01% Crystal Violet and colonies were enumerated on a Bio-Rad GelDocXR using QuantityOne Software (Sensitivity=8.1,Average=5).

Invasion assay
Invasion assays were performed as described previously (Layton et al., 2009) using BioCoat Growth Factor-Reduced Matrigel Invasion Chambers (BD Biosciences, Bedford, MA, USA).Briefly, cells were applied in SF-media to the upper chamber with or without 15min preincubation with TAPI1 (40μg/mL), aprotinin (100μg/mL), anti-uPA monoclonal antibody 3689 (25μg/mL), nonspecific IgG 1 control antibody MOPC-21 (25g/ml), or an equal volume of DMSO or PBS control included in both chambers.SF-media or growth media was provided in the lower chamber and cells were allowed to invade for 24h before removal of uninvaded cells, and enumeration of invaded cells.

Reverse Transcription (RT)-PCR
cDNA was synthesized from 1μg of total RNA using oligo-dT primer.PCR was performed on 1μL of resulting cDNA using primers described previously (Layton et al., 2009;Leissner et al., 2006).GAPDH primers were from Stratagene (San Diego, CA) and served as internal controls.Densitometry was performed on unadjusted images using NIH Image 1.61 software and GAPDH as reference control.

Statistics
Experiments were performed in triplicate and independently repeated at least twice.Data shown are mean + standard deviation unless otherwise indicated.Colony-formation and cellular invasion were analyzed by two-tailed Students t-Test.

Endogenously-expressed syk is functional in pancreatic ductal cells in situ and in vitro
Previously we demonstrated the expression of syk in normal pancreatic ductal epithelial cells in situ (Layton et al., 2009).In order to assess whether this endogenous syk is functionally active, we performed immunohistochemistry with antibodies specific for the phosphorylated form of Y352, which is representative of syk activation by upstream mediators, or YY525/6, which is indicative of substrate-specific activity of the syk kinase domain (Kimura et al., 1996;Sada et al., 2001;Zhang et al., 1998).These antibodies react with both isoforms of syk and do not recognize more distantly related src family members (Cell Signaling Technology, Beverly, MA).Strong staining for both phospho-Y352 (Fig. 1A-C) and phospho-YY525/6 (Fig. 1D-F) was observed in individual cells of independent ducts as well as ductules associated with acinar clusters.No staining was observed in the absence of primary antibody or when purified rabbit IgG was used as a control (not shown).Nuclear staining was observed in a subset of the duct cells (arrowheads in Fig. 1C,F), consistent with the detection of nuclear syk in pancreatic ductal cells in situ (Layton et al., 2009) and previous reports of syk nuclear translocalization (Wang et al., 2003) and regulation of cell division (Zyss et al., 2005) and transcription (Wang et al., 2005).Of the 20 samples analyzed, all showed evidence of syk activation in a subset of ductal cells and none showed widespread constitutive phosphorylation at either site, demonstrating the active regulation of syk activation and activity in ductal cells of the normal human pancreas and suggesting an active role for syk in regulating the phenotype of this cell type.
We also previously established the expression of syk in well-to moderately-differentiated PDAC cells (CAPAN2-G1, CFPAC1-G2, BxPC3-G2, AsPC1-G2), to the exclusion of poorlydifferentiated PDAC cells (Panc1-G3, MIAPaCa2-G3+) (Layton et al., 2009).In order to establish the utilization of syk by these endogenously syk-positive PDAC cells, we analyzed the activation surrogate readout of Y352 and YY525/6 phosphorylation in BxPC3 cells.Serum-starved cells were stimulated with insulin and then lysates immunoblotted with the phospho-syk-specific pAbs used for IHC.Interestingly, both Y352 and YY525/6 are 108 constitutively phosphorylated under starvation conditions, and dephosphorylated in response to insulin (Fig. 2).Similar results were obtained with serum stimulation (not shown), and with endogenously syk-positive CAPAN2 cells (not shown).These data demonstrate the active regulation of syk activity in response to specific signalling pathways in endogenously syk-positive, well-differentiated PDAC cells, consistent with the restricted syk activation observed in situ and suggesting that syk is an active participant in regulating the phenotype of these cells.

Syk-dependent regulation of cell growth requires syk kinase activity
Syk regulates proliferation of immune cells in response to specific receptor activation events (Kimura et al., 1996;Sada et al., 2001;Wieder et al., 2001), and ectopic expression of an RFPtagged syk caused anomalous cell division and mitotic catastrophe in breast cancer cells, where it was observed to interact with -tubulin of the mitotic spindle (Zyss et al., 2005).We observed localization of syk to the perinuclear region of dividing PDAC cells (not shown), therefore we assessed whether syk might be involved in regulating the proliferation of endogenously syk-positive PDAC cells in a manner analogous to that described in breast cancer (Coopman et al., 2000).Consistent with such a role, the in vitro growth of endogenously syk-positive CAPAN2, CFPAC1, BxPC3 and AsPC1 cells was suppressed in a dose-dependent manner by piceatannol, a phytochemical that inhibits the kinase activity of syk, suppressing downstream phosphorylation events by a largely unknown mechanism (Geahlen et al., 1989;Ferrigni et al., 1984;Wieder et al., 2001)(Fig.3A).Significantly, the only G1 cells in this group, CAPAN2, were completely eradicated by the highest dose of inhibitor, suggesting cytotoxicity in addition to any cytostatic effect of this compound in these highly differentiated cells.The specificity of this effect to the inhibition of syk is demonstrated by the following facts.First, trans-stilbene, which is the base molecule from which piceatannol is derived, had no effect on the proliferation of any of these cells (not shown).Second, syk-negative MIAPaCa2 and Panc1 cells demonstrated no growth response to piceatannol in this assay (Fig. 3A).Third, although piceatannol has been reported to inhibit FAK, src, PI3K and IB/NF-B kinases (Ashikawa et al., 2002;Choi et al., 2010;Law et al., 1999), inhibition of these kinases has been shown to retard the proliferation of both syk-positive and syknegative lines used in this study (Hering et al., 2007;Hochwald, et al., 2009;Ito et al., 2003;Perugini et al., 2000).Together, these points demonstrate that the antiproliferative effect of piceatannol reported here is likely restricted to the inhibition of syk itself.It should be noted that the concentrations of piceatannol used in this study are based on prior established parameters designed to minimize non-specific effects (Ashikawa et al., 2002;Choi et al., 2010;Geahlen et al., 1989;Law et al., 1999;Seow et al., 2002;Wieder et al., 2001).To extend these pharmacological inhibitor studies, we stably transfected endogenously sykpositive BxPC3 cells with a kinase-dead syk construct harbouring a point mutation (K402R) in the ATP binding site of the kinase domain (BxPC3/KD cells).This construct has been shown to act as a dominant-negative in endogenously syk-positive cells through an incompletely understood mechanism (Coopman et al., 2000).BxPC3/KD cells demonstrated a consistently reduced growth rate versus mock-transfectants (Fig. 3B).Stable expression of wildtype syk did not affect the proliferation rate of these cells (not shown).It should be noted that these engineered cells are non-clonal populations, being the result of a bicistronic system that links transgene expression to drug resistance through an IRES sequence.Moreover, repeated generation of stable populations using this system resulted in stable "lines" that behaved similarly, demonstrating that these data are not artifactual in nature.
Since stable reexpression of syk in G3 Panc1 cells reduced their growth rate in vitro (Layton et al., 2009), we questioned whether stable reexpression of syk in MIAPaCa2 (MP2/syk wt ) cells would have a similar effect on these G3+ PDAC cells.Unlike Panc1 cells, no growth rate difference was observed between MP2/mock and MP2/syk wt cells in culture (Fig. 3C).Moreover, piceatannol did not suppress the growth of MIAPaCa2 or Panc1 cells stably reexpressing syk (Fig. 3D), suggesting that PDAC cells that have progressed to syk-negative status are able to bypass the growth requirement for endogenous syk when it is ectopically reintroduced.It should be noted, however, that piceatannol was functional in this assay, since it caused a slight dose-dependent increase in the growth rate of Panc1/syk cells (Fig. 3D) that is commensurate with overcoming the minor growth rate reduction that resulted from the original stable reexpression of syk in these cells (Layton et al., 2009).In contrast, however, we did observe an effect of stable syk reexpression on overall MIAPaCa2 morphology similar to that observed in Panc1/syk cells (Layton et al., 2009).As such, MP2/syk wt cells demonstrate increased cell-cell interactions, resulting in more of a traditional monolayer characteristic of epithelial cells in culture, and a reduced propensity for cells to remain separate in between cell clusters (Fig. 3E).This suggests the reestablishment of a more differentiated phenotype by syk in these cells.Accordingly, MP2/syk wt cells demonstrate a dramatically reduced ability to grow in an anchorageindependent growth assay (Fig. 3F), exhibiting a plating efficiency of 1.7%, versus 9.0% for MP2/mock cells, less dramatic but similar to the effect observed previously in Panc1/syk cells (Layton et al., 2009).

Stable expression, but not kinase activity, is required for syk-dependent invasion suppression
We previously observed a negative regulation of Panc1 invasion by stable reexpression of syk (Layton et al., 2009).Therefore, we assessed the in vitro invasion capabilities of MP2/mock and MP2/syk wt cells.Stable MP2/syk wt cells exhibit 90% less invasion towards serum-containing media compared to MP2/mock, and essentially no invasion in the absence of serum attractant (Fig. 4A).To determine whether the kinase activity of syk is required for this effect, we stably expressed the K402R kinase-dead (KD) mutant syk in MIAPaCa2 cells (MP2/syk KD ); both wildtype and KD proteins were expressed at equal levels (Fig. 4A, inset).Importantly, stable MP2/syk KD cells exhibit nearly identical invasion suppression to MP2/syk wt cells (Fig. 4A), demonstrating that kinase activity is not required for syk's invasion-suppressor function in these cells.Consistent with this finding, 48h pretreatment with piceatannol did not affect the invasion of either endogenously sykpositive BxPC3 cells or stable MP2/syk wt cells (Fig. 4B).Additionally, transient transfection with syk wt or syk KD did not retard MIAPaCa2 invasion (Fig. 4C), suggesting that long-term expression is required for syk to influence invasion, likely through gene expression changes as described previously in Panc1 cells (Layton et al., 2009).

MIAPaCa2 invasion is regulated by urokinase-type plasminogen activators
We previously demonstrated that Panc1 invasion is dependent upon the matrix metalloproteinase (MMP)-2 axis, and that syk specifically attenuates the expression of MMP2 and its inhibitor, TIMP2, in these cells (Layton et al., 2009).In contrast to Panc1 cells, the MMP inhibitor TAPI1 did not suppress MIAPaCa2/mock invasion (Fig. 5A), and RT-PCR (Fig. 5B) and zymography (not shown) showed that MIAPaCa2 cells do not produce MMP2 in culture.MMP2 and MMP9 products were detected in parallel reactions run at the same time on unrelated samples (not shown), demonstrating that lack of signal in these samples is not the result of failed amplification reactions.These results demonstrate that MIAPaCa2 invasion is MMP-independent in vitro.
Another key regulator of epithelial cell invasion is the urokinase-type plasminogen activator (uPA)/uPA receptor (uPAR) axis (Leissner et al., 2006;McMahon and Kwaan, 2009), whose components are expressed by MIAPaCa2 cells in culture (Fig. 5B).Consistent with the potential involvement of this system, MP2/mock invasion was completely suppressed by the serine protease inhibitor aprotinin (Fig. 5C).This effect was not due to toxicity, as 112 aprotinin-treated cells remained 100% viable after 24h (Fig. 5C, inset), which is the duration of the invasion assay.Specifically demonstrating the involvement of the uPA/uPAR axis, MP2/mock invasion was almost completely inhibited by a function-blocking anti-uPA antibody (Fig. 5C).An isotype-matched (IgG 1 ) control antibody (MOPC-21) had no effect (not shown).Importantly, densitometry demonstrated >75% reduction in uPA mRNA levels by both syk wt and syk KD in MIAPaCa2 cells (Fig. 5B), suggesting a potential mechanism for syk's effect on MIAPaCa2 invasion.Interestingly, expression of the uPA inhibitor, PAI-1, was also suppressed >30% by both wildtype and kinase-dead syk (Fig. 5B).However, the uPA receptor, uPAR, was actually increased ~50% by syk wt , but suppressed by >60% by syk KD , suggesting that kinase activity may be required for appropriate regulation of uPAR, but not uPA or PAI-1.Previous studies have demonstrated the regulation of the MMP2 axis by the v3 integrin (Deryugina et al., 2001;Nisato et al., 2005), and an association of the uPA/uPAR axis with the v5 integrin (Yebra et al., 1995;1996).Moreover, MMP2 activation has been linked to v3 in PDAC (Hosotani et al., 2002), and v3 engagement suppresses the expression of uPA/uPAR components (Hapke et al., 2001).Therefore, to assess whether the different protease dependencies observed in this and our former study (Layton et al., 2009) are related to differential integrin expression, FACS analysis was performed on live cells from standard culture.Consistent with MMP2-dependence, Panc1 cells proved to be strongly v3-positive (Fig. 6A).In contrast, MIAPaCa2 cells are v3-negative (Fig. 6B), commensurate with their MMP2-independent/uPA-dependent phenotype.These results on cell surface integrin expression were corroborated at the total expression level by immunoblotting of whole cell lysates (not shown).Further demonstrating the integrindependent phenotypes of these cells, MIAPaCa2 invasion could be fully suppressed by a v5 integrin-specific function-blocking mAb, while Panc1 invasion also involves v3 and could only be fully blocked by combination of v5and v3-specific function blocking mAbs (Fig. 6C).To further test the role of syk in regulating PDAC invasion, we examined the BxPC3/KD cells, which stably express the dominant-negative kinase-dead syk in addition to their endogenous wildtype syk.While BxPC3/mock cells show marginal invasion towards serum-containing media and essentially no invasion in the absence of serum attractant, stable BxPC3/KD cells demonstrate a >75% increase in invasion towards serum-containing media and significant de novo invasion in the absence of serum attractant (Fig. 7A).FACS demonstrated that BxPC3 cells are also v3-negative (not shown) and, consistent with this finding, the MMP inhibitor TAPI1 had no effect on serum-free BxPC3/KD invasion (Fig. 7B).However, both aprotinin and anti-uPA mAb treatment reduced serum-free BxPC3/KD invasion significantly (Fig. 7B), demonstrating a role for the uPA/uPAR axis in the sykregulated invasion of these endogenously syk-positive PDAC cells.An isotype-matched (IgG 1 ) control antibody (MOPC-21) had no effect (not shown).It should be stressed that BxPC3/mock cells do not invade under these conditions, thereby allowing us to assess the role of uPA in a phenotype that was dependent upon syk inhibition and allowing us to exclude confounding issues derived from the endogenous phenotype.

Discussion
Previously we identified syk as being expressed not only in normal pancreatic ductal epithelium, but also in well-to moderately-differentiated PDAC in situ (Layton et al., 2009).The expression of syk in normal ductal epithelium of the breast has also been reported, as has a role for syk as a tumor suppressor in that tissue (Coopman et al., 2000).As such, loss of syk expression in primary breast tumors is associated with a poor prognosis (Toyama et al., 2003).We found that syk similarly correlates with patient survival in PDAC patients (Layton et al., 2009).However, we observe consistent loss of syk expression in poorlydifferentiated PDAC, and have not observed loss of syk in well-differentiated PDAC samples.This differs from breast cancer, where syk is absent from a subset of welldifferentiated lesions and expressed normally in many poorly-differentiated samples (Toyama et al., 2003), suggesting potentially dramatic differences in the regulation or function of syk in the ductal epithelium of these two glandular tissues.Indeed, we present evidence that syk functions to facilitate growth of PDAC cells that express it.This is perhaps not entirely unexpected since syk promotes immune cell proliferation in response to antigenic stimulation (Kimura et al., 1996;Sada et al., 2001), and vascular defects observed in syk-deficient mice are attributable to a reduction in endothelial cell number (Yanagi et al., 2001).Indeed, syk was subsequently shown to be required for endothelial cell proliferation in vitro (Yanagi et al., 2001).
Mechanistically, syk has been shown to regulate mitosis through direct interactions with tubulin and catalytic activity within the centrosome (Zyss et al., 2005).Morphological examination of piceatannol-treated PDAC cells stained with DAPI revealed that aberrant/dysfunctional mitosis may be responsible for the reduced cell numbers observed in this assay (not shown).This finding is consistent with prior ectopic expression studies that demonstrated localization of a carboxy-terminally RFP-tagged syk to centrosomes (Zyss et al., 2005); this chimera caused aberrant mitosis that likely resulted from steric interference of the RFP molecule on the directly adjacent kinase domain, further supporting our contention that syk kinase activity is required for syk-dependent regulation of cell division, and hence growth.
Similarly, since pharmacological inhibition of syk does not affect BxPC3 invasion, but stable expression of syk KD does, this suggests that this effect of syk KD is not due to direct suppression of endogenous syk signaling in these cells.As such, this construct may function by sequestering binding partners that would normally be phosphorylated by catalytically active syk, or a similar mechanism, since this construct can still be phosphorylated by upstream mediators such as src.Indeed, phosphorylation of Y352 couples syk to binding partners such as phospholipaseC and vav through the SH2 domains of these proteins (Sada et al., 2001).Mutation of Y352 impairs signaling in immune cells (Sada et al., 2001), demonstrating that Y352 is an important regulator of syk function in these cells.The presence of phosphorylated Y352 in pancreatic cells in situ and in vitro suggests that these or similar pathways are functional in syk-mediated processes in this cell type as well.
Proteolytic degradation of the extracellular matrix during tumor progression often involves uPA, its inhibitor, PAI-1, and its receptor, uPAR (Binder and Mihalay, 2008;McMahon and Kwaan, 2009).This system regulates the activation of the serine protease plasmin, and in the pancreas uPAR has been shown to regulate PDAC phenotype through interaction with integrin v5 and PKC signalling (Yebra et al., 1995;1996).More importantly, studies have shown a correlation between uPA and PAI-1 expression and tumor aggressiveness (Hansen et al., 2003), and PAI-1 has been shown to be necessary for tumor invasion (Binder and Mihaly, 2008).We recently reported that stable reexpression of syk dramatically reduced the invasion of endogenously syk-negative Panc1 cells at least partly by attenuating the MMP2 axis (Layton et al., 2009).Herein, we demonstrate that MIAPaCa2 and BxPC3 invasion is MMP-independent and mediated by the urokinase/plasminogen system, and that syk attenuates the expression of both uPA and PAI-1.This difference in protease utilization may be due to the differential expression and/or involvement of specific integrins since Panc1 cells are v3-positive while MIAPaCa2 and BxPC3 cells are v3-negative.Integrin v3 suppresses uPA/uPAR expression in ovarian cancer cells (Hapke et al., 2001) and, as noted previously, the MMP2 axis has been associated with expression of the v3 integrin and v3 is required for MMP2 activation in some cell types (Deryugina et al., 2001);Nisato et al., 2005), including PDAC (Hosotani et al., 2002).Reciprocally, the uPA/uPAR axis is linked to v5 (Yebra et al., 1995;1996) and other integrins including 31 (Zhu et al., 2009), which is also expressed by both MIAPaCa2 and BxPC3 cells (S.Silletti, unpublished data).Importantly, the differential integrin utilization by these cells extended to the regulation of invasion as well.Therefore, this differential use of protease pathways in an integrin-specific manner by PDAC cells may be useful from a diagnostic and/or prognostic standpoint if further studies bear out this relationship in this tumor type.
In summary, we have further characterized the activity of syk in the growth and invasion regulation of pancreatic ductal epithelial cells.The expression of syk in these cells is associated with phosphorylation of tyrosines indicative of syk activation and activity, both in vitro and in situ, and syk activity is involved in regulating the proliferation/survival of syk-positive PDAC cells in vitro.This effect may be related to syk's role in transducing signals from growth factor receptors, or from syk's regulation of the mitotic spindle during mitosis.Irrespective, there has been a recent surge of interest in inhibitor-based strategies to target syk for conditions such as asthma, rheumatoid arthritis, and other immune disorders (Scott, 2011;Ulanova et al., 2005;Wong et al., 2004).Therefore, we propose that, aside from being a potential regulator of PDAC biology and biomarker of more differentiated PDAC tumors in situ, syk may be a viable target for therapeutic intervention in the clinic, since all grades of PDAC are uniformly fatal (NCI PDQ Database [http://www.cnacer.gov/cancertopics/pdq];NCI-PANC-PRG) and syk expression is maintained in G1 and G2 PDAC in situ (Layton et al., 2009).

Conclusion
In this report we demonstrate that upstream activated and catalytically active syk kinase states can be detected in pancreatic ductal epithelial cells, both in vitro and in situ.The activity of syk is further demonstrated with regard to the regulation of cellular growth and invasion.Importantly, we have found that syk-dependent regulation of invasion is via modulation of the urokinase/plasminogen system in v3-negative PDAC cells.This is in contrast to the MMP-dependent invasion-suppressor effect of syk in v3-positive PDAC cells, which are the exception, rather than the norm in this tumor type.These data provide further evidence of a central role for syk in regulating pancreatic ductal epithelial cell phenotype, and support the potential utility of targeting syk as a potential therapeutic modality in well-to moderately-differentiated PDAC patients, which are the majority of clinical cases.The significance of this is highlighted by the fact that PDAC is characterized by extensive dissemination at the time of diagnosis, irrespective of grade of disease.

Acknowledgement
S.Silletti is an American Cancer Society Research Scholar supported by ACS RSG-05-116-01-CSM and NIH grants CA130104 and CA109956.Funding agencies had no role in study design/performance or manuscript preparation.Moores UCSD Cancer Center is an NCIsponsored Comprehensive Cancer Center supported by NCI Specialized Support Grant P30 CA23100.

Fig. 2 .
Fig. 2. Syk is functional in PDAC cells in vitro.Immunoblot of phospho-Y352 and phospho-YY525/6 in BxPC3 cells that had been serum-starved for 24h prior to stimulation with 10ng/ml insulin for the indicated times.NT, no treatment.Total syk levels determined with the 4D10 mAb are shown as a control.

Fig. 3 .
Fig. 3. Kinase activity is required for syk-dependent regulation of cell growth.(A) Cell growth of syk-positive versus syk-negative PDAC cell lines after 7 days of piceatannol treatment, plotted as percent of untreated control.(B) Proliferation of stable BxPC3/mock and BxPC3/syk KD cells after 5 days.(B,inset) Immunoblot of total syk (syk) and the myc-tag (myc) of syk KD .Erk2, loading control.(C) Proliferation of MP2/mock and MP2/syk wt cells after 5 days.(D) MP2/syk wt and Panc1/syk cell growth assay after 7 days of piceatannol treatment, as in (A).(E) Phase contrast images demonstrating the effect of stable syk expression on MIAPaCa2 cell morphology.Both populations were maintained under identical culture conditions including maintenance doses of the selectable marker hygromycin.(F) Effect of syk on anchorage-independent growth of MIAPaCa2 cells.A representative example of replicate plates is shown with quantitation points superimposed.

Fig. 4 .
Fig. 4. Stable expression, but not kinase activity, is required for invasion suppression by syk.(A) The indicated stable MIAPaCa2 populations were provided serum-containing (FBS) or serum-free (SF) medium in the lower compartment of invasion chambers and allowed to invade for 24h before removal of uninvaded cells and enumeration of invaded cells.(A,inset) Immunoblot of ectopic syk expression with the 4D10 mAb (syk).Erk2, loading control.(B) Cells were pretreated for 48h with 25M piceatannol (PC) or DMSO vehicle (NT) prior to seeding into invasion chambers with serum-containing media in the lower compartment.(C) MIAPaCa2 cells were transiently cotransfected with lacZ reporter and empty vector (mock), or wildtype (WT) or kinase-dead (KD) syk and 48h later seeded into invasion chambers.After 24 hours cells were stained with x-gal and invaded transfected cells (blue) were enumerated and plotted as percent of mock.

Fig. 6 .
Fig. 6.The invasion mechanism of MIAPaCa2 and Panc1 cells is related to their integrin expression profiles.(A,B) FACS analysis of integrin v3 and v5 surface expression in Panc1 (A) and MIAPaCa2 cells (B).(C) Invasion of MIAPaCa2 (left) and Panc1 (right) cells in the presence or absence of 50g/ml of function-blocking antibodies directed against the indicated integrins.

Fig. 7 .
Fig. 7. BxPC3 invasion is regulated by syk and mediated by the uPA/uPAR axis.(A) Invasion of stable BxPC3/mock (Bx/mock) and BxPC3/KD (Bx/KD) cells towards serum-containing (FBS) or serum-free (SF) medium.(B) Serum-free BxPC3/KD invasion +/-TAPI1 (left), or aprotinin or function-blocking anti-uPA antibody (right).BxPC3/mock cells were not assessed as they exhibit no invasion in the absence of serum in the lower chamber.