Vascular Endothelial Growth Factor-induced Migration of Multiple Myeloma Cells Is Associated with (cid:1) 1 Integrin- and Phosphatidylinositol 3-Kinase-dependent PKC (cid:2) Activation*

In multiple myeloma (MM), migration is necessary for the homing of tumor cells to bone marrow (BM), for expansion within the BM microenvironment, and for egress into the peripheral blood. In the present study we characterize the role of vascular endothelial growth factor (VEGF) and (cid:1) 1 integrin (CD29) in MM cell migration. We show that protein kinase C (PKC) (cid:2) is translocated to the plasma membrane and activated by adhesion of MM cells to fibronectin and VEGF. We identify (cid:1) 1 integrin modu- lating VEGF-triggered MM cell migration on fibronectin. We show that transient enhancement of MM cell adhesion to fibronectin triggered by VEGF is dependent on the activity of both PKC and (cid:1) 1 integrin. Moreover, we dem- onstrate that PKC (cid:2) is constitutively associated with (cid:1) 1 integrin. These data are consistent with PKC (cid:2) -dependent exocytosis of activated (cid:1) 1 integrin to the plasma mem- brane, where its increased surface expression mediates binding to fibronectin; conversely, catalytically active PKC (cid:2) -driven internalization of (cid:1) 1 integrin results in MM cell de-adhesion. We show that the regulatory subunit of phosphatidylinositol (PI) 3-kinase (p85) is constitutively associated with FMS-like tyrosine kinase-1 (Flt-1). VEGF stimulates activation of PI 3-kinase, and both MM cell adhesion and migration are PI 3-kinase-dependent. More-over, both VEGF-induced PI 3-kinase activation and (cid:1) 1 integrin-mediated binding to fibronectin are required for the recruitment and activation of PKC (cid:2) . Time-lapse phase contrast video microscopy (TLVM) studies confirm the importance of these signaling components in VEGF-triggered MM cell migration on fibronectin.

In multiple myeloma (MM) 1 migration is necessary for the homing of tumor cells to bone marrow (BM), for expansion of malignant plasma cells within the BM microenvironment, and for the egress into peripheral blood. It has been reported that the extracellular matrix (ECM) proteins laminin, microfibrillar collagen type VI, and fibronectin are strong adhesive components for MM cells and that adhesion to laminin and fibronectin is ␤ 1 integrin (CD29)-mediated (1). ␤ 1 integrins are typically expressed on MM cells, specifically ␣ integrins VLA-4 (␣ 4 ␤ 1 ) and VLA-5 (␣ 5 ␤ 1 ) (2, 3). ␤ 1 integrin-mediated adhesion of MM cells to fibronectin confers protection against drug-induced apoptosis and triggers NFB-dependent transcription and secretion of interleukin-6, the major MM growth and survival factor (4,5). Interestingly, chimeric mice (␤ 1 Ϫ/Ϫ 3 wild-type chimeras) lack ␤ 1 -null cells in blood and in hematopoietic organs such as spleen, thymus, and BM as a consequence of the inability of ␤ 1 -null cells to invade the fetal liver (6). In addition to upregulation of cell surface expression and induction of surfaceclustering, integrin activity can be triggered by multiple agonists through "inside-out" signaling independent of changes in integrin expression levels (e.g. ligand binding to growth factor receptors is associated with changes in the way in which adhesion receptors on the cell surface engage the ECM). This concept is illustrated in human umbilical vein endothelial cells in which VEGF stimulates ␤ 1 integrins and leads to markedly enhanced movement (7).
Although VEGF induces migration as a key step in angiogenesis, the interplay between VEGF and integrins is not restricted to angiogenesis. VEGF and VEGFR are expressed by many tumor cell lines; moreover, elevated levels of VEGF are found in cancer patients, and inhibition of VEGF can suppress tumor growth (8). Indeed, clinical studies are underway investigating VEGF as a novel therapeutic target (9). In MM VEGF is expressed and secreted by tumor cells as well as BM stromal cells (10,11); moreover, binding of MM cells to BM stromal cells enhances both interleukin-6 and VEGF secretion (11). We recently showed that in addition to stimulating angiogenesis, VEGF directly induces MM cell proliferation via a protein kinase C (PKC)-independent mitogen-activated protein kinase/ extracellular signal-regulated kinase (MEK/ERK) pathway and triggers MM cell migration on fibronectin via a PKC-dependent pathway (12). Members of the PKC family mediate multiple physiological functions (13)(14)(15)(16)(17)(18) including integrin-mediated cell spreading and migration (19 -22). To date, 11 isoenzymes of the serine/threonine kinase PKC have been identified and classified into three subgroups based on structure and cofactor regulation: conventional PKC, novel PKC, and atypical PKC (15,18). The conventional PKC isoforms participate in the inside-out signaling activation of cell adhesion mediated by ␤ 1 , ␤ 2 , and ␤ 3 integrins and are also required for cell spreading (23)(24)(25). Moreover, PKC␣ associates with ␤ 1 integrins, thereby regulating cell trafficking (26,27).
In the present study, we describe the close interrelationship between integrin and growth factor-induced signaling pathways in MM. We identify PKC␣ as the primary PKC isozyme involved in VEGF-induced MM cell migration. By showing that VEGF-mediated MM cell migration is associated with ␤ 1 integrin-and PI 3-kinase-dependent PKC␣ activation, we further confirm the importance of tumor cell-BM microenvironment interaction as a pivotal process in the pathogenesis of MM. Moreover, our studies identify several potential targets for novel therapies to improve outcome in MM.

EXPERIMENTAL PROCEDURES
Materials-Recombinant human VEGF 165 was purchased from R&D Systems (Minneapolis, MN). Human plasma fibronectin was obtained from Invitrogen. ␤ 1 integrin-specific mAb was purified from P4C10 ascites (Chemicon, Temecula, CA). PKC isoforms were purchased from Transduction Laboratories. The goat polyclonal Ab raised against the carboxyl terminus of PYK2, the mouse mAb raised against full-length ␤ 1 integrin (4B7R), and the rabbit polyclonal Ab directed against amino acids within the extracellular domain of Flt-1 were purchased from Santa Cruz. Anti-phosphotyrosine 4G10 antibody was kindly provided by Dr. Tom Roberts (Dana-Farber Cancer Institute, Boston, MA).
Stimulation of Cells-Cells were starved for 15-18 h in medium with 3% FBS overnight and then for 3 h without FBS prior to stimulation with indicated VEGF concentrations (recombinant human VEGF, R&D Systems) or 50 nM/300 nM PMA for 20 -30 min at 37°C.
Immunocomplexes were collected following overnight incubation at 4°C with 10 -20 l of 100% protein A-Sepharose CL-4B beads (Amersham Biosciences, Inc.). For Western blotting, cell lysates (30 -100 g/lane) or immunoprecipitates (500 g-1.5 mg total proteins) were separated by 8 or 10% SDS-PAGE prior to electrophoretic transfer onto Hybond TM -C super nitrocellulose membranes (Amersham Biosciences, Inc.). After blocking with 5% nonfat milk in phosphate-buffered saline-Tween 20 buffer at room temperature for 1 h, membranes were sequentially blotted with the indicated specific primary Abs and then with horseradish peroxidase-conjugated secondary mouse, rabbit, or goat Abs and were developed using chemiluminescence (Amersham Biosciences, Inc.).
Cell Fractionation-After washing three times with phosphate-buffered saline, cells were transferred into 200 l of hypotonic lysis buffer (HLB: 10 mM Tris-HCl, 1 mM EDTA, 1 mM sodium vanadate, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and protease inhibitor mixture) and incubated for 20 min on ice. The cells were then lysed by 80 strokes in a Dounce homogenizer and subjected to centrifugation at 1500 ϫ g to pellet nuclei and unbroken cells followed by centrifugation of the supernatant at 100,000 ϫ g for 20 min. The supernatant was collected (S100 fraction), and the pellet was resuspended in 70 l of HLB containing 0.1% Triton X (P100 fraction).
Measurement of PKC␣ Activity-PKC␣ activity was measured with a PKC assay kit (Upstate Biotechnology, Lake Placid, NY) according to the manufacturer's instructions. The phosphotransferase activity of PKC␣ was quantitated in a scintillation counter (Beckman LS 6500, multi-purpose scintillation counter) measuring the amount of the ␥-phosphate of [␥-32 P]ATP incorporated in a specific PKC substrate peptide (QKRPSQRSKYL) bound to P81 phosphocellulose paper. Endogenous phosphorylation of proteins in the sample was determined by substituting the assay dilution buffer for the substrate mixture. To assure that equal amounts of PKC␣ were used in the assay, immunoprecipitates were denaturated, eluted, separated by 10% SDS-PAGE, electrophoretically transferred, and immunoblotted with PKC␣.
Phosphoinositide Kinase Assays-PI 3-kinase assays were performed as described previously (29) in a total volume of 50 l. The radioactivity was visualized and quantitated on a PhosphorImager (Amersham Biosciences, Inc.).
Cell Adhesion Assays-These were performed using the Vybrant TM cell adhesion assay kit (Molecular Probes, Eugene, OR) according to the manufacturer's instructions. Briefly, after an 18-h starvation in RPMI/2% FBS, MM.1S cells (4 ϫ 10 6 cells/ml) were harvested and labeled with calcein-acetoxymethyl ester (5 M per cell loading) for 30 min. After washing with prewarmed (37°C) RPMI 1640 (without serum), the cells were preincubated with or without blocking ␤ 1 integrin neutralizing mAb, bisindolylmaleimide I (BIM I, Calbiochem), or LY294002 (Calbiochem), respectively, and then stimulated with VEGF. The cell suspensions were immediately added to fibronectin-or polylysine (1 g/ml)-coated or non-coated wells. After 90 -120 min, nonadherent calcein-labeled cells were removed by gently washing twice with RPMI 1640 by inversion of the plates. Adherent cells were quantitated in a fluorescence multi-well plate reader (Molecular Devices, Sunnyvale, CA) and examined microscopically. All experiments were done in triplicate.
Transwell Migration Assay-Cell migration was assayed as described previously (12,30,31). After 2-5 h, cells that had migrated into the lower compartment of a Boyden-modified chamber were counted using a Coulter counter ZBII (Beckman Coulter).
Time-lapse Video Microscopy-MM.1S cells were starved in RPMI medium containing 2% FBS for 16 h and plated to uncoated or fibronectin-coated tissue culture plates (35 ϫ 10-mm plates, BD PharMingen), respectively, in the presence or absence of VEGF (100 ng/ml). For image capturing, an Olympus IX70 inverted microscope (Olympus, Lake Success, NY) with Hoffman optics (ϫ10/ϫ20/ϫ40) equipped with a temperature-(Therm-Omega-Tech, Warmington, PA) and CO 2 (5%)controlled chamber was connected to an Optronics Engineering DEI-750 3CCD digital video camera (Optronics, Galeta, CA). Animation and export to a Quick Time movie were performed using the QED Camera Standalone 145 software at 2-min intervals. Images were analyzed with the NIH Image 1.62 software. To generate migration tracks, the position of the centroid of individual cells on each image were marked. The migratory speed was calculated based on the sum of distances divided by the time of observation. Migration of at least 15 cells was analyzed for each experimental condition.
Statistical Analysis-Statistical significance of differences observed in VEGF-treated versus control cultures was determined using an unpaired Student's t test. The minimal level of significance was p Ͻ 0.05.

RESULTS AND DISCUSSION
Previously we showed that VEGF-triggered MM cell migration on fibronectin is mediated via a PKC-dependent signaling pathway (12). In the present study we define and characterize the interrelationship of VEGF-and integrin-signaling in activating PKC that ultimately leads to MM cell migration. PKC Isoform Expression in MM Cells-Several PKC isoforms, including PKC␣ (32), PKC␦ (33), and PKC (34), have been implicated in a cell migratory phenotype. Additionally, overexpression of PKC␣ was shown to enhance cell motility/ invasiveness of breast cancer cells (26). We have shown previously that MM cell migration is PKC-dependent because it can be selectively inhibited by the PKC inhibitor BIM I (12). As a first step to identify the class of PKC mediating VEGF-induced migration in MM cells, we examined the expression of the PKC isoforms in MM cell lines and patient cells (Fig. 1). Immunoblot analysis revealed that PKC␣, PKC␥, PKC⑀, and PKC are significantly expressed in all human MM cell lines and MM patient cells investigated, in contrast to PKC␤ (low expression), PKC␦ (variable expression), and PKC (no expression, data not shown). As in our previous studies we chose the MM.1S human MM cell line (28) as a representative model system for this study.

VEGF-induced Signaling Pathways and Adhesion to Fibronectin
Mediate PKC␣ Activation-Because the activation of PKC occurs concomitantly with recruitment to the plasma membrane, we next performed immunoblotting of cell fractions with isoform-specific Abs against PKC␣, PKC␥, PKC⑀, and PKC to delineate the intracellular distribution after VEGF stimulation (Fig. 2). In non-treated MM cells all PKC isoforms were detected primarily in the cytosolic fraction. After VEGF stimulation of cells seeded on fibronectin, PKC␣ translocated into the membrane fraction after 30 min (Fig. 2a), whereas no significant changes in distribution of PKC isoforms as PKC␥, PKC⑀, and PKC were observed. The activation of PKC␣ after VEGF treatment of cells attached to fibronectin was further supported by a 2-fold increase in PKC␣-IP kinase activity; PKC␣-IP kinase activation by PMA served as a positive control (Fig. 2b). Although a cAMP-dependent protein kinase/calmod-ulin kinase (PKA/CaMK) inhibitor mixture was used, this assay (Upstate Biotechnology) may not exclude the phosphorylation of a specific substrate peptide (QKRPSQRSKYL) by unknown PKC␣-coprecipitated kinases. Neither plating of cells on fibronectin alone nor stimulation of suspended cells with VEGF alone significantly changed the subcellular distribution of PKC␣ (Fig. 2, c and d). In contrast, treatment with PMA, a major PKC activator, induced translocation of responsive PKC (PKC␣, PKC␥, and PKC⑀) into the detergent-soluble membrane fraction as expected (Fig. 2d) (35,36) VEGF Enhances MM Cell Adhesion to Fibronectin-Laminin, the microfibrillar collagen type VI, and fibronectin bind MM cells, and adhesion to laminin and fibronectin is ␤ 1 integrin (CD29)-mediated (1). ␤ 1 integrins expressed on MM cells include VLA-4 (␣ 4 ␤ 1 ) and VLA-5 (␣ 5 ␤ 1 ) (2-5), which mediate adherence to both the ECM and BM stromal cells. In human umbilical vein endothelial cells, VEGF stimulates ␤ 1 integrins via inside-out signaling, leading to significantly increased motility (7). Because migration is a dynamic process of cell adhesion formation and release, we next investigated whether VEGF can modulate ␤ 1 integrin-mediated MM cell adhesion. As shown previously, MM cells spontaneously adhered to fibronectin, and this adhesion was increased upon stimulation with VEGF (Fig. 3a). Maximal increments of VEGF-mediated cell adhesion were observed at fibronectin concentrations of 25-30 g/ml, whereas adhesion decreased to baseline levels at higher fibronectin concentrations (Fig. 3b). Notably, VEGFmediated increments in adhesion were time-dependent, with maximal binding observed after 75-90 min of VEGF treatment and decreasing after 120 min (Fig. 3c).
We next examined whether the increment of adhesion observed at 90 min of VEGF stimulation was PKC-and ␤ 1 integrindependent. As shown in Fig. 3, d-e, incubation with the PKC inhibitor BIM I (as well as with the ␤ 1 -neutralizing mAb) blocked dose-dependent VEGF-induced cell adhesion to fibronectin. This involvement of PKC in MM cell adhesion to fibronectin was further confirmed by a dose-dependent increment of adhesion triggered by PMA stimulation (Fig. 3d) similar to that induced by VEGF. Taken together, our results show that VEGF transiently enhances MM cell adhesion to fibronectin, dependent on both PKC and ␤ 1 integrin activity.
␤ 1 Integrins Modulate VEGF-triggered Migration on Fibronectin-We next sought to determine whether ␤ 1 integrin modulates VEGF-mediated MM cell migration on fibronectin. As seen in Fig. 3f, ␤ 1 integrin neutralizing mAb (but not irrelevant IgG) mediated dose-dependent inhibition of VEGF-trig- Immunoprecipitates were analyzed by SDS-PAGE followed by immunoblotting with the indicated antibodies. Panel d, for the PI 3-kinase assay, equal amounts of lysates were immunoprecipitated with anti-phosphotyrosine mAb, and immunocomplexes were assayed for the ability to phosphorylate PIP 2 . PIP 3 , phosphatidylinositol (3,4,5)P 3 . C, control; indicates a PI 3-kinase assay performed on protein A alone. gered MM cell migration in a Boyden-modified microchemotaxis chamber. These data confirm that ␤ 1 integrin (CD29) is the integrin primarily associated with VEGF-triggered MM cell migration on fibronectin.
PKC␣ Associates with ␤ 1 Integrin and CADTK-The control mechanisms leading to the various stages of the integrin receptor life cycle are largely unknown. Propagation of cell movement is thought to be regulated by the distribution and redistribution of integrins through surface diffusion, internalization, clustering at the leading front, and ripping release from the cell rear (37). In mammary epithelial cells Ng et al. (26) recently have found that PKC␣ interacts with activated ␤ 1 integrin, which regulates its exocytosis to the plasma membrane; moreover, catalytically active PKC␣ is responsible for ␤ 1 integrin internalization through a Ca 2ϩ -and PI 3-kinase-dependent, dynamin I-controlled endocytic pathway. PKC␣ induced up-regulation of the integrin-dependent cell migration of these cells, which was blocked under conditions that prevented the internalization of the receptor complex. We therefore next determined whether PKC␣ and ␤ 1 integrin are associated in MM cells. Constitutive complex formation between these two proteins was demonstrated by co-immunoprecipitation (Fig. 4a). Upon activation and translocation, conventional PKCs associate with proteins of the transmembrane-4 superfamily (TM4SF or tetraspans) linking PKC to several subsets of integrins (38) including ␣ 4 ␤ 1 (39) and ␣ 5 ␤ 1 (40). Specificity of binding to TM4SF is dependent on the extracellular domain of the integrin ␣ chain, and the integrin-TM4SF-PKC complex formation results in phosphorylation of the integrin ␣ cytoplasmic tail (41). In ongoing studies, we are investigating the regulatory role of these integrin-associated proteins on VEGF-induced MM cell migration.
VEGF Stimulates Tyrosine Phosphorylation of CADTK in Adherent MM Cells-In addition to influencing cell motility via controlling ␤ 1 integrin trafficking, catalytically active PKC also regulates other components of the focal complexes, including the small GTPases (Cdc42, Rho, Rac) and/or actin cytoskeletonbinding proteins. Calcium-dependent tyrosine kinase (CADTK), also known as proline-rich tyrosine kinase 2 (PYK2), calcium-dependent tyrosine-kinase ␤ (CAK␤), and related adhesion focal tyrosine kinase (RAFTK), is a cytoplasmic tyrosine kinase homologous to focal adhesion kinase (FAK). Like FAK, CADTK is a platform kinase site for the coalescence of signaling and adaptor molecules, thereby facilitating the transmission of surface signals to the cytoskeleton and signaling pathways associated with cell growth, apoptosis, and migration. A number of studies have demonstrated tyrosine phosphorylation of CADTK in cells of hematopoietic origin (e.g. T and B cells, monocytes, natural killer cells, granulocytes, bone marrow progenitors, mast cells, megakaryocytes, and platelets). Stimuli that activate CADTK in these cells are associated with cell motility or at least cytoskeletal rearrangement (e.g. CADTK activation is required for cytoskeletal reorganization and monocyte motility) (42).
In MM.1S cells, we have shown previously that CADTK is activated upon dexamethasone treatment, suggesting its role in dexamethasone-induced apoptosis (47). Notably, CADTK has been reported to link calcium-and integrin-mediated signaling to the cytoskeleton in brain and hematopoietic cells (43)(44)(45)(46). Our results show that VEGF induces the association of CADTK with the ␤ 1 integrin/PKC␣ protein complex (Fig. 4a). VEGFR immunoprecipitation and immunoblotting studies using anti-Flt-1 and anti-CADTK Abs showed specific constitutive association (Fig. 4b). We next investigated whether the VEGF-induced formation of a membrane complex containing PKC␣, ␤ 1 integrin, and Flt-1 changes the activity of the associated CADTK in adherent MM cells. As shown in Fig. 4d, VEGF increased (2-fold) fibronectin-induced CADTK tyrosine phosphorylation in MM.1S cells; equivalent loading of proteins was confirmed by stripping and reprobing the membrane with anti-CADTK Ab. The PKC activator PMA enhanced constitutive phosphorylation of CADTK (Fig. 4c). Finally, VEGF-triggered CADTK activation was blocked by both the PKC inhibitor BIM I and a ␤ 1 integrin neutralizing mAb (Fig. 4e). These results indicate that the activation of CADTK in adherent MM cells is influenced by both PKC-and ␤ 1 integrin-mediated signaling pathways. Moreover, a recent report shows that CADTK in hematopoietic cells is also associated with the cytoskeletal protein paxillin (48). We therefore postulate that VEGF may facilitate signal transduction to the cytoskeleton in MM cells and modulate MM cell migration and/or adhesion on fibronectin via CADTK. Ongoing studies are directed toward understanding this role of CADTK in MM cells in more detail.
p85 Interacts with Flt-1-We next sought to identify the inside-out signaling components mediating regulation of ␤ 1 integrin function by VEGF in MM cells. PI 3-kinases phosphorylate PI(4,5)P 2 , and PLC␥ hydrolyzes PI(4,5)P 2 to diacylglycerol and inositol 1,4,5-trisphosphate, thereby mediating intracellular calcium release and PKC activation (49). PLC␥ activity is enhanced by 3-phosphoinositides, both indirectly via Btkrelated tyrosine kinases and directly through the binding of PI(3,4,5)P 3 to the amino-terminal pleckstrin homology (PH) domain and the tandem Src homology (SH) 2 domains of PLC␥ (50 -53). We have shown recently that Flt-1 VEGFR is expressed in the human MM cell line and patient MM cells and is tyrosine-phosphorylated upon VEGF stimulation (12). Because Flt-1 and the p85 subunit of PI 3-kinase were associated when overexpressed in yeast cells (54), we next determined whether this interaction is biologically significant in MM cells. Flt-1 was immunoprecipitated from VEGF treated and non-treated MM.1S cells followed by immunoblotting with anti-p85 Ab. As illustrated in Fig. 5a, p85 is constitutively associated with Flt-1, suggesting that PI 3-kinase participates in Flt-1-mediated VEGF signal transduction.
VEGF Mediates PI 3-Kinase Activation in MM Cells-To confirm this hypothesis, we next examined whether PI 3-kinase activation is an early event in VEGF-induced signaling in MM cells. VEGF induced tyrosine phosphorylation of the regulatory p85 subunit of PI 3-kinase, evidenced by immunoblotting of anti-p85 immunoprecipitates with an anti-pY mAb (Fig. 5b). Moreover, immunoblotting of anti-pY immunoprecipitates with specific p85 mAb showed an increased association of p85 with tyrosine-phosphorylated proteins, which was maximal after 5 min of VEGF treatment (Fig. 5c). To determine whether VEGFinduced p85 phosphorylation regulates p110, the catalytic counterpart of PI 3-kinase, we performed in vitro kinase assays using PI 4,5-bisphosphate as a substrate. As shown in Fig. 5d, VEGF induced peak activation (6-fold) of PI 3-kinase activity at 5 min. Taken together, these data show that activation of PI 3-kinase is an early event in the VEGF-triggered Flt-1 signaling cascade.
VEGF-induced PI 3-Kinase Activation and Adhesion to Fibronectin Mediate PKC␣ Translocation to the Plasma Membrane-We next examined whether activation of PI 3-kinase is required for activation of PKC␣. When PI 3-kinase activation in MM cells seeded on fibronectin was inhibited by LY294002, PKC␣ recruitment to the plasma membrane was also abrogated (Fig. 6). These results indicate that VEGF-induced PI 3-kinase activation along with ␤ 1 integrin binding to fibronectin activate PKC␣.
MM Cell Migration and Cell Adhesion Are PI 3-Kinase-dependent-To determine whether activation of PI 3-kinase is necessary for MM cell migration, we next used a Boydenmodified chamber to assay the effect of VEGF on the transfilter migration activity of MM.1S cells seeded on membranes precoated with fibronectin or polylysine as a control for ECM specificity. As seen in Fig. 7a, VEGF added to the conditioned medium in the lower chamber induced a dose-dependent migration of growth factor-deprived MM.1S cells seeded on fibronectin in the upper chamber, which was totally inhibited by the PKC inhibitor BIM I. Preincubation with the PI 3-kinase inhibitor LY294002 (20 M) at 37°C for 45 min prior to VEGF stimulation similarly inhibited MM cell migration. In contrast, VEGF did not induce migration of MM.1S cells seeded on polylysine in the upper chamber. We next examined whether VEGF-induced increments in adhesion were also PI 3-kinasedependent. As shown in Fig. 7b, incubation with the PI 3-ki-  nase inhibitor LY294002 (20 M) blocked VEGF-induced MM cell adhesion to fibronectin. In contrast, VEGF did not effect MM cell adhesion toward polylysine. Taken together, these results demonstrate that PI 3-kinase activation during MM cell migration on fibronectin regulates PKC. These observations are similar to the Ca 2ϩ -and PI 3-kinase-dependent PKC␣catalyzed endocytosis reported in breast cancer cells (26).
Time-lapse Phase Contrast Video Microscopy-Migration is a complex dynamic process of cell adhesion formation and release organized and coordinated both in time and space. To enhance our understanding of the above results and to link them with the dynamic changes in MM cell morphology that ultimately mediate cell migration, we used time-lapse phase contrast video microscopy (TLVM) (Figs. 8 and 9). During migration filopodia and lamellipodia together with new adhesions are formed at the leading edge, whereas detachment of adhesions are concomitantly observed at the trailing cell edge. In timecourse experiments, MM.1S cells were seeded on either fibronectin precoated or control tissue culture plates in the presence or absence of VEGF. Representative cells were investigated for 30 min starting at 3 h after plating, and composite time-lapse phase contrast micrographs (magnification, ϫ40) were acquired at 2-min intervals tracked with the NIH Image 1.62 software (Fig. 8). MM cells adherent to non-coated tissue culture plates failed to polarize, and the random movement was slightly increased by VEGF (Fig. 8, a and b). In contrast, MM cells adherent to fibronectin rapidly polarized and exhibited increased continuous membrane ruffling (Fig. 8c). Although additional stimulation with VEGF did not significantly change membrane ruffling, the migration rate was markedly increased (Fig. 8d). Importantly, in the presence of the PKC inhibitor BIM I, migration rate was again decreased to levels obtained in MM cells adherent to fibronectin alone (Fig. 8e).
The average speed of MM cell migration was next measured as described under "Materials and Methods" (Fig. 9). The baseline migration rate of MM cells seeded on tissue culture plates (not precoated with fibronectin) corresponded to the spontaneous motility of resting cells. The average speed remained steady for the whole period of observation and never exceeded 15-20 m/hr. Similar results were observed using VEGFtreated cells in the absence of fibronectin. Importantly, binding of MM cells to fibronectin induced a marked acceleration of cell motility, peaking 2-3 h after seeding and remaining higher compared with the conditions described above. VEGF induced an additional increase in motility peaking at 3 h at 100 m/hr and remaining significantly higher compared with non-stimulated cells seeded to fibronectin. This acceleration was inhibited to baseline migration rates in the presence of 2 M BIM I.
This study therefore demonstrates that VEGF-mediated MM cell migration is associated with ␤ 1 integrin-and PI 3-kinasedependent PKC␣ activation. To further verify the role of PKC␣ as a potential new therapeutic target in MM, ongoing studies are investigating the effect of PKC␣-antisense both in vitro and in murine MM models.