The Lipid Products of Phosphoinositide 3-Kinase Increase Cell Motility through Protein Kinase C

Phosphoinositide 3-kinase has been implicated as an activator of cell motility in a variety of recent studies, yet the role of its lipid product, phosphatidylinositol 1,4,5-trisphosphate (PtdIns-3,4,5-P 3 ), has yet to be eluci- dated. In this study, three independent preparations of PtdIns-3,4,5-P 3 were found to increase the motility of NIH 3T3 cells when examined utilizing a microchemotaxis chamber. Dipalmitoyl L - (cid:97) -phosphatidyl- D -myo-ino-sitol 3,4,5-triphosphate (Di-C 16 -PtdIns-3,4,5-P 3 ) also pro- duced actin reorganization and membrane ruffling. Cells pretreated with 12- O -tetradecanoylphorbol-13-ac- etate to cause down-regulation of protein kinase C (PKC) exhibited complete inhibition of cell motility in- duced by Di-C 16 -PtdIns-3,4,5-P 3 . These results are con- sistent with previous observations that PtdIns-3,4,5-P 3 activates Ca 2 (cid:49) -independent PKC isoforms in vitro and in vivo and provide the first demonstration of an in vivo role for the lipid products of the phosphoinositide 3-ki-nase. PtdIns-3,4,5-P 3 appears to directly initiate cellular motility via activation of a PKC family member.

Initiation of cellular motility has been demonstrated with multiple growth factors, including platelet-derived growth factor (PDGF) 1 (1), hepatocyte growth factor (HGF) (2), and insulin (3). The mechanisms whereby cells undergo chemotaxis (directional cell movement) and chemokinesis (random cell movement) are complex, requiring dissolution of cell-cell contacts (such as tight junctions in epithelial cells) and cell-surface contacts, formation of lamellipodia, actin filament severing and nucleation, and finally contraction of the actin filament network leading to movement of the cell body (4). An understanding of the signaling pathways required to orchestrate these cellular events should provide critical new insights into numerous biological events such as cell migration and organization during organ development and wound healing, tumor cell metastasis, and progression of arterial atherosclerotic plaques.

MATERIALS AND METHODS
Cell Culture and Reagents-The majority of experiments were performed with NIH 3T3 fibroblasts, using PDGF as the positive control. Selected experiments were repeated with mIMCD-3 cells, a murine renal tubular epithelial cell line that expresses the c-met receptor and exhibits striking chemotaxis to a gradient of HGF (17)(18)(19). All cells were cultured in Dulbecco's modified Eagle's medium with 5% fetal calf serum using standard techniques. PDGF (Upstate Biotechnology, Inc., Lake Placid, NY) and HGF (Institute of Immunology, Tokyo, Japan) were used in concentrations of 10 and 40 ng/ml, respectively, based on previous dose response curves for maximal chemotaxis (19).
PtdIns-4,5-P 2 was obtained from Upstate Biologicals, and phosphatidylserine (PtdSer) was from Avanti Polar Lipids. Diacylglycerol (DAG) and horseradish anti-mouse conjugate were purchased from Boehringer Mannheim. 12-O-Tetradecanoylphorbol-13-acetate (TPA) was obtained from Life Technologies, Inc., and wortmannin was from Sigma. Calphostin C was obtained from Calbiochem, and P81 phosphocellulose paper was purchased from Whatman. Thin layer chromatography plates (Silica Gel 60) were obtained from EM Separations.
Chemotaxis Assay-Chemotaxis was evaluated using a modified Boyden chamber assay with a 48-well microchemotaxis chamber as described previously (Neuro Probe Inc., Cabin John, MD) (19,23). Lipids * This work was supported in part by National Institutes of Health Grant DK48871. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
FIG. 1. PtdIns-3,4,5-P 3 enhances motility of NIH 3T3 fibroblasts (A and B) and IMCD epithelial cells (C). Cell motility was evaluated using a modified Boyden chamber assay with a 48-well microchemotaxis chamber. A, the lower section of the Boyden chamber was filled with media alone or media containing PDGF, PtdSer/PtdIns-4,5-P 2 lipid substrate, or enzymatically generated PtdIns-3,4,5-P 3 . B and C, either chemically were dried in a stream of nitrogen and then sonicated for 5 min in serum-free media. The lower section of the Boyden chamber was filled with media alone or media containing either PDGF (10 ng/ml) or PtdSer/PtdIns-4,5-P 2 (100 M/25 M) or PtdSer/PtdIns-4,5-P 2 /PtdIns-3,4,5-P 3 (100 M/25 M/5 M). A polycarbonate filter (Nucleopore Corp., Pleasanton, CA) coated with rat tail collagen type I (Collaborative Biomedical, Bedford, MA) was placed over the lower compartment, and 1.5 ϫ 10 4 cells were added to the upper compartment. In some experiments, wortmannin was diluted 1:1000 in serum-free media immediately prior to use and added at the appropriate concentration to both the upper and lower chambers at time 0. Control experiments were performed with Me 2 SO vehicle alone. After 4 h of incubation at 37°C, filters were removed, the cells were fixed and stained with Diff-Quik (Baxter Healthcare Corp., Miami, FL), and the upper surface was wiped with a cotton applicator to remove nonchemotaxing cells. For each well, cells that had passed through the pores were counted, and the mean value of cells/mm 2 was calculated.
Electrophoresis and Western Blotting-Confluent plates of cells were serum-starved overnight in the presence of either 0.3% Me 2 SO or 300 nM TPA followed by a wash with phosphate-buffered saline and lysis in ice-cold lysis buffer (137 mM NaCl, 20 mM Tris, 1 mM MgCl 2 , 1 mM CaCl 2 , 10% glycerol, 1% Nonidet P-40, 2 mM sodium vanadate, 1 mM phenylmethylsulfonyl fluoride, pH 7.5). The suspension was centrifuged for 10 min at 12,000 ϫ g. Equal aliquots of supernatant determined by protein assay (Bio-Rad) were resolved by SDS-polyacrylamide gel electrophoresis and transferred to Immobilon (Millipore). Expression of protein kinase C⑀ was detected by a monoclonal antibody specific for this enzyme (Transduction Laboratories) and quantified using a Molecular Dynamics PhosphorImager.
In Vivo Labeling of Lipids-In vivo levels of D3 phosphoinositides in response to stimuli were measured as described previously (8). Briefly, 3T3 cells maintained in Dulbecco's modified Eagle's medium, 5% fetal calf serum were grown to 80% confluence and then placed in Dulbecco's modified Eagle's medium, 0.1% fetal calf serum for 12-16 h. For labeling purposes, monolayers were placed in phosphate-free Dulbecco's modified Eagle's medium in the absence of serum for 1 h, followed by 2 mCi/ml of [ 32 P]orthophosphate for 3 h. Cells were then stimulated with PDGF (20 ng/ml), Di-C 16 -PtdIns-3,4,5-P 3 , or vehicle control for 10 min. Following stimulation, cells were washed twice with ice-cold phosphatebuffered saline and lysed in 750 l of methanol, 1 M HCl (1:1). 20 g of crude brain phosphoinositides (Sigma) were added as carrier. Lipids were extracted by the addition of 380 l of chloroform, and the organic phase was washed twice with 400 l of methanol, 0.1 M EDTA. Phospholipids were then deacylated and prepared for Sepharose A exchange-high pressure liquid chromatography analysis as described previously (8).
Membrane Ruffling-F-actin was localized in coverslip adherent cells. Quiescent 3T3-fibroblasts or cells exposed to 40 ng/ml of PDGF, 5 M Di-C 16 -PtdIns-3,4,5-P 3 , or 5 M PtdIns-4,5-P 2 for 10 -60 min were fixed by the addition of an equal volume of 3.7% formaldehyde in phosphate-buffered saline at 37°C for 30 min. Fixed cells were permeabilized with 0.1 volume of 1% Triton X-100 containing 2 M tetramethylrhodamine B isothiocyanate-phalloidin at 37°C for 60 min (24), washed three times with phosphate-buffered saline for 5 min each, and magnified in a Zeiss IM45 Inverted microscope.
Statistical Analysis-Results were averaged, and statistical relevance was determined by Student's t test. Data are presented as mean Ϯ S.E.

RESULTS
Enzymatically Generated and Synthetic PtdIns-3,4,5-P 3 Initiate Cell Motility and Ruffling-Since exogenously added lipid vesicles and micelles are known to fuse with the plasma membrane of live cells, we investigated the role of PtdIns-3,4,5-P 3 in cell motility by directly adding this lipid to the cells in a Boyden chamber. Three independent preparations of PtdIns-3,4,5-P 3 were employed to evaluate the motility response of this putative second messenger (Fig. 1). PtdIns-3,4,5-P 3 was enzymatically generated by adding purified PI 3-kinase to a 1:4 mixture of PtdIns-4,5-P 2 and PtdSer followed by chloroform extraction of the lipid products. This resulted in conversion of 20% of the PtdIns-4,5-P 2 to PtdIns-3,4,5-P 3 to give a final mixture of 100 M PtdSer, 25 M PtdIns-4,5-P 2 , 5 M PtdIns-3,4,5-P 3 . The addition of this PtdSer/PtdIns-4,5-P 2 /PtdIns-3,4,5-P 3 mixture to the bottom well of the chemotaxis chamber resulted in a 10-fold increase in motility of NIH 3T3 fibroblasts compared with vehicle control and a 3-fold increase compared with the PtdSer/PtdIns-4,5-P 2 mixture alone (Fig. 1A). The small but reproducible motility response to 100 M PtdSer, 25 M PtdIns-4,5-P 2 may be due to either activation of PKC isoforms by the high concentrations of PtdIns-4,5-P 2 (13) or impurities in these lipids not seen when lower concentrations of lipids were investigated individually (see "Discussion").

Exogenously Added PtdIns-3,4,5-P 3 Does Not Activate Endogenous PI 3-Kinase and Is Not Inhibited by Wortmannin-It
was conceivable that a contaminant or a breakdown product of PtdIns-3,4,5-P 3 might initiate the observed effects via activation of a cell surface receptor (as has been shown for lysophosphatidic acid). Although this seemed unlikely, since PtdIns-3,4,5-P 3 made by three different procedures stimulated cell motility and comparable concentrations of PtdIns-4,5-P 2 and/or PtdSer failed to stimulate cell motility, we searched for evidence that exogenously added PtdIns-3,4,5-P 3 might act via cell surface receptor activation by examining intracellular production of PtdIns-3,4-P 2 and PtdIns-3,4,5-P 3 in 32 PO 4 3Ϫ -labeled NIH 3T3 cells following the addition of extracellular Di-C 16 -PtdIns-3,4,5-P 3 . This approach was chosen because the receptors known to initiate chemotaxis (PDGF receptor, insulin receptor, c-met receptor, lysophosphatidic acid receptor) have also been found to activate the PI 3-kinase (8, 26 -28). While stimulation with PDGF produced a dramatic rise in intracellular [ 32  FIG. 3. Membrane ruffling by cells exposed to Di-C 16 -PtdIns-3,4,5-P 3 mimics the response seen with PDGF. The top panels show that quiescent cells and cells exposed to 5 M PtdIns-4,5-P 2 (PIP 2 ) are no different, whereas significant membrane ruffling can be seen in the cells exposed to either 40 ng/ml PDGF or 5 M PtdIns-3,4,5-P 3 (PIP 3 ) at 10 (middle panels) and 30 min (bottom panels).
The PI 3-kinase inhibitor wortmannin binds irreversibly to the catalytic subunit of the enzyme and prevents production of the D3 phosphorylated lipid products of the enzyme. 10 nM wortmannin, the lowest dose that produces reliable inhibition of the PI 3-kinase in vivo in 3T3 fibroblasts and mIMCD-3 cells (19), caused a 60% inhibition of PDGF-and HGF-dependent cell motility but had no effect on Di-C 16 -PtdIns-3,4,5-P 3 -stimulated cell movement in 3T3 cells or in IMCD cells (Fig. 1, B  and C). These results demonstrate that wortmannin at a dose that inhibits PDGF receptor-mediated activation of the PI 3-kinase does not prevent PtdIns-3,4,5-P 3 -initiated motility, further supporting the hypothesis that these lipids are inserting into the membrane and directly initiating downstream signaling events.
100 nM wortmannin, a concentration where effects on several other kinases have been observed, caused essentially complete inhibition of both PDGF and Di-C 16 -PtdIns-3,4,5-P 3 -stimulated cell motility (data not shown). In light of the observation by Kundra et al. (1) that selective activation of phospholipase C␥ by the PDGF receptor resulted in a substantial chemotactic response, even in the absence of PI 3-kinase activation, this result suggests that other targets of wortmannin that are likely to be inhibited at the higher concentration, such as myosin light chain kinase (29) or PtdIns 4-kinase (30), may be critical for cell motility as well.
To examine this possibility, the TPA-activable PKC family members were down-regulated by overnight preincubation of NIH 3T3 cells with 300 nM TPA. Under these conditions, there was a 78% decline in the concentration of PKC⑀ by Western analysis (Fig. 4), an effect comparable with that seen in human dermal fibroblasts (34). PKC⑀ was chosen because it shows the greatest activation to PtdIns-3,4,5-P 3 in vitro and in vivo. The migratory response to PtdIns-3,4,5-P 3 was completely eliminated in NIH 3T3 cells pretreated with TPA (Fig. 5A), while PDGF-mediated cell movement was inhibited by 70%, a finding similar to that seen with exposure to 10 nM wortmannin. In addition, the specific PKC inhibitor calphostin C was tested (35,36). NIH 3T3 cells exposed to 100 nM calphostin C (IC 50 ϭ 75-100 nM) for 30 min demonstrated a 90% reduction in PtdIns-3,4,5-P 3mediated cell movement (Fig. 5B). These results suggest that activation of the PI 3-kinase mediates cell motility via the local generation of PtdIns-3,4,5-P 3 and subsequent activation of PKC. DISCUSSION The PI 3-kinase has been clearly implicated in cell motility by several laboratories (1, 6, 7, 19), yet the actual mechanism of this effect is poorly understood. The p85 subunit of the PI 3-kinase has a BCR homology domain, which is capable of binding GTP-Rac (37,38) and may therefore act to recruit activated Rac or associated family members to the membrane where these signaling proteins have been shown to initiate motility (39,40). In addition, PtdIns-3,4,5-P 3 has been shown to be capable of activating the actin-severing protein, gelsolin (28). The present results demonstrate that PtdIns-3,4,5-P 3 , the lipid product of the PI 3-kinase, is capable of directly initiating cell motility and that this effect is mediated by activation of PKC. The greatest effect on cell motility occurred as directional movement toward a gradient of PtdIns-3,4,5-P 3 (i.e. chemotaxis), while nondirectional movement increased as well.
The model we propose to explain the ability of exogenously added PtdIns-3,4,5-P 3 to stimulate cell motility requires that some fraction of the lipid fuse with the plasma membrane and arrive at the inner leaflet over the 4-h assay period. There it can act similarly to endogenous PtdIns-3,4,5-P 3 . This might occur by different mechanisms for the different PtdIns-3,4,5-P 3 preparations. The enzymatically produced PtdIns-3,4,5-P 3 , sonicated in the presence of excess PtdSer, is expected to be distributed in both the inner and outer leaflets of the newly formed vesicle. Fusion of these vesicles with the plasma membrane would presumably result in PtdIns-3,4,5-P 3 in both leaflets. The Di-C 16 -PtdIns-3,4,5-P 3 forms micelles when sonicated 2 and is likely to cause local detergent-like effects when fusing with the plasma membrane, resulting in distribution of this lipid on both leaflets of the plasma membrane as well.
Recently, PKC⑀ as well as the atypical PKC isoform have been found to be activated downstream of the PtdIns 3-kinase in PDGF-and epidermal growth factor-stimulated cells (15,16). In addition, several researchers have demonstrated that PtdIns-3,4,5-P 3 can directly activate calcium-insensitive PKC family members in vitro (12)(13)(14). Our results demonstrate that activation of the phorbol-sensitive PKC family members enhances cell motility in a fashion similar to that seen with PtdIns-3,4,5-P 3 , while both down-regulation of PKC by overnight treatment with TPA and inhibition of PKC with calphostin C completely blocked the PtdIns-3,4,5-P 3 response. The present data cannot distinguish which PKC family member or members are directly involved, although in vitro data suggest that the PKC⑀ is strongly up-regulated by these lipids. The observation that high concentrations of PtdIns-4,5-P 2 caused a modest increase in cell motility (Figs. 1A and 2) is consistent with this hypothesis, since this polyphosphoinositide has also been found to weakly activate PKC in vitro (13).
In addition to PKC activation, there are several other targets for the lipid products of the PI 3-kinase. pp70 S6k was the first target shown to be downstream of PI 3-kinase (42), and recently, Akt (41,43) and Rac (38) have also been implicated. The latter is of particular interest, since microinjection of constitutively active forms of rac leads to membrane ruffling (39). The recent availability of synthetic forms of these phosphoinositides should help identify their targets and determine the pathways that lead to the motile response.
Previous results from our laboratory and others have shown that activation of the PI 3-kinase is essential for PDGF and HGF-dependent cell movement. The present experiments demonstrate that the lipid products of the PI 3-kinase act directly as second messengers in cell motility and provide the first indication that PKC family members are required for the motility effects of this lipid in vivo.