Introduction Cell migration is a dynamic process that involves the coordination of multiple cellular events, which include the disassembly of focal adhesions at the trailing edges and the assembly of new focal adhesions at the migrating fronts (Lauffenburger and Horwitz, 1996; Caswell et al., 2009). Constitutive integrin turnover, internalization, and recycling have been demonstrated under basal cell migration conditions (Pellinen and Ivaska, 2006; Mosesson et al., 2008). In recent years, clathrin-mediated endocytosis has been shown to play a pivotal role in the internalization of surface integrins at focal adhesions that are undergoing basal turnover (Chao and Kunz, 2009; Ezratty et al., 2009). However, few studies have examined dynamic integrin disassembly, redistribution, and reassembly in highly motile cells (Webb et al., 2002). In fact, in vivo cell migration is often significantly increased by growth factor up-regulation under physiological and pathological conditions, such as inflammation, wound healing (Ross et al., 1986), and cancer (Price et al., 1999). It is unknown whether the mechanisms of integrin redistribution from the trailing edge to the migrating front are the same as in basal cell migration. Unexpectedly, we found that growth factor–stimulated cell migration is achieved by using a special circular dorsal ruffle (CDR) macropinocytosis mechanism that recruits, internalizes, and recycles integrins. CDRs are massive actin cytoskeletal remodeling structures that form within minutes at the dorsal cell surface after stimulation by growth factors, such as PDGF, EGF, and VEGF, in various cell types (Chinkers et al., 1979; Mellström et al., 1988; Wu et al., 2003; Orth and McNiven, 2006). Although the function of these structures is largely unknown, they have been suggested to be part of an initial step leading to massive macropinocytosis (Orth et al., 2006). Here, we delineate the pathway by which focal adhesions rapidly disassemble as integrins translocate to CDRs, are internalized by macropinocytosis, and then distribute to newly forming focal adhesions at the leading edge of cells during stimulated cell migration. This pathway was found to be entirely distinct from the clathrin-dependent or caveolin-dependent constitutive pathway of integrin turnover at focal adhesions in basal cell migration. Results and discussion Growth factor stimulation induces integrin focal adhesion disassembly at the ventral cell surface and massive CDR formation with the accumulation of integrins at the dorsal cell surface Stimulation of fibroblasts by PDGF is a model system to study stimulated cell migration (Ballestrem et al., 2001; Roberts et al., 2001). Examining integrin β3 in these cells, we detected that integrins concentrate at focal adhesions (Fig. 1 A). Remarkably, after the addition of PDGF for 5 min, the integrins accumulated at actin-rich circular structures (Fig. 1 A). According to our previous findings in PDGF-stimulated actin cytoskeleton remodeling (Gu et al., 2007), such actin-enriched circular structures are CDRs. Comparing the distribution of integrin β3 with two markers of CDRs, F-actin and cortactin (Buccione et al., 2004), we found that all three molecules showed colocalization. 3D analysis showed integrin β3, F-actin, and cortactin concentrating at cup-shaped structures that were raised upward from the dorsal cell surface (Fig. 1 B, Fig. S1 A, and Video 1). As a control, actin-independent membrane protein major histocompatibility complex (MHC) class I did not translocate to CDRs under the same conditions (Fig. S1 B). Kinetic quantification showed that 33, 41, 25, 15, 11, and 5% of cells had integrin β3 at CDRs at 5, 10, 15, 20, 25, and 30 min after PDGF stimulation, respectively (Fig. 1 C). This temporal profile was concordant with previous lifetime studies on CDRs (Buccione et al., 2004; Orth et al., 2006). Besides integrin β3, we also observed that integrin β1 redistributes to CDRs (Fig. S1 C). After surface integrin β1 antibody labeling in live cells, we confirmed surface integrin β1 translocation to CDRs within as short as 3 min after PDGF stimulation (Fig. S1 D). Such fast and massive surface integrin redistribution suggests that surface integrins follow a direct cell surface route rather than the slow surface–cytosol–surface endocytosis and recycling route. Figure 1. Integrin β3 localizes at CDRs after PDGF-BB stimulation and integrin β3–GFP translocation 4D tracing in a live cell. (A) Primary mouse fibroblasts were stimulated with or without PDGF-BB, fixed, and IF stained. Arrows denote integrin β3 at focal adhesions, and arrowheads denote integrin β3 at CDRs. (B) Primary mouse fibroblasts were stimulated with PDGF-BB, fixed, and IF stained. Confocal image stacks were scanned along the z axis. Arrows 1 and 2 denote remnant integrin β3 at focal adhesions. Arrowheads 3 and 4 denote integrin β3 at a large (early) CDR and a condensed (late) CDR, respectively. Crossed lines denote the orthogonal position. The z distance in the orthogonal views was exaggerated five times. (C) Primary mouse fibroblasts were stimulated with PDGF-BB for various times, fixed, and IF stained. The number of cells forming integrin β3 CDRs per 100 cells was counted (n = 5). (D) NIH3T3 cells stably expressing integrin β3–GFP were stimulated with PDGF-BB. The temporal and spatial translocation of integrin β3–GFP was traced by 4D time-lapse confocal live-cell imaging. Sections of confocal images were scanned along the cell z axis every 1 min. The ventral cell surface position was defined as z = 0 µm. A positive z distance defined the position distance above the ventral cell surface. t defined the time after PDGF-BB addition. Arrows denote integrin β3–GFP at focal adhesions. Large arrowheads denote integrin β3–GFP at early CDRs, late condensed CDRs, macropinosomes, and the perinuclear region. (E) The number of focal adhesions, diameter of CDRs or macropinosomes, and number of macropinosomes at various times after PDGF-BB stimulation were counted to determine the temporal and spatial translocation of integrin β3–GFP in D and Video 2. (F) The lifetimes of ventral surface integrin β3–GFP focal adhesions were quantified in PDGF-BB–stimulated cells and unstimulated control cells as in Video 2 and Video 3, respectively. A focal adhesion lifetime 30 min was recorded as 30 min. The lifetimes of 30 focal adhesions per cell were recorded (n = 3). (G) Primary mouse fibroblasts or NIH3T3 cells stably expressing integrin β3–GFP were stimulated with PDGF-BB for various times. Integrin β3 protein levels from whole-cell lysates were analyzed by SDS-PAGE and Western blotting. Error bars represent means ± SD. ***, P 13% after PDGF stimulation. In contrast, WAVE1 down-regulation blocked this process (Fig. 4, D and E). Furthermore, CHC or CAV1 down-regulation did not block integrin β3 macropinosome formation or its subsequent recycling in the endocytosis and recycling assay, whereas WAVE1 down-regulation did (Fig. 4 F and Fig. S2, G−I). Finally, we found that PDGF stimulation still enhanced colocalization of integrin β3 with EEA1 at perinuclear regions after macropinocytosis in CHC and CAV1 down-regulated cells but not in WAVE1 down-regulated cells (Fig. 5, C–F). Quantification of the number of integrin β3 focal adhesions per cell showed CHC or CAV1 down-regulation did not block focal adhesion disassembly or subsequent integrin β3 recycling to form new focal adhesions, whereas WAVE1 down-regulation did block focal adhesion disassembly, and thus, the number of integrin β3 focal adhesions per cell remained almost unchanged over time (Fig. S2 C). Next, we examined fast cell migration using cells seeded in the upper chamber of Transwells with vitronectin-coated membranes. After 2 h of cell adhesion to the membrane, PDGF was added to the lower chamber, and cell migration across the membrane for 2 h was assessed to quantitate stimulated fast cell migration. As a control, we targeted integrin β3 or both integrin β3 and integrin β1 with blocking antibodies. These results confirmed that stimulated cell migration on vitronectin is integrin dependent. As previous experiments showed that PDGF-stimulated recycling of integrin β3 required Rab4 (Roberts et al., 2001), we also found that siRNA against Rab4 decreased this cell migration. We found that PDGF-stimulated fast cell migration was significantly delayed by siRNA against BARS or WAVE1, whereas such a fast cell migration was only slightly delayed by siRNA against CHC or CAV1 (Fig. 5, G and H; and Fig. S2 D). These results suggest that clathrin- or CAV1-dependent constitutive integrin turnover plays only a minor role in growth factor–stimulated fast cell migration. In contrast, WAVE1-dependent CDR formation, BARS-dependent integrin macropinocytosis, and subsequent Rab4-dependent integrin recycling contributed significantly to growth factor–stimulated fast cell migration. In conclusion, we found that cell surface integrins in focal adhesions undergo internalization by macropinocytosis after stimulation by growth factors (summarized in Fig. 5 I). This result is in contrast to those previously observed for surface integrins at disassembling focal adhesions, which have been observed to undergo clathrin- or CAV1-mediated endocytosis. As the previous observations were made under conditions that did not involve the addition of growth factors, a likely explanation is that integrins use clathrin- or CAV1-mediated endocytosis under basal conditions and BARS-dependent macropinocytosis under the growth factor–stimulated conditions studied here. Subsequently, the internalized integrins undergo a recycling itinerary similar to that previously documented for how integrins recycle in stimulated cells (Powelka et al., 2004). Notably, this itinerary also explains why integrins were observed to move both along the cell surface and also through internal vesicular routes during cell migration (Regen and Horwitz, 1992). Integrins were previously noted at cell edge membrane ruffles (Bretscher, 2008; Sung et al., 2008). CDRs are distinct from cell edge membrane ruffles in terms of their location, formation, signaling, and especially their link to macropinocytosis. Here, we report that CDRs also rapidly recruit a majority of the cell surface integrins and internalize them through macropinocytosis for subsequent fast recycling. These results provide direct evidence to support the hypothesis that CDRs function as indicators of cellular transition from static to motile states (Buccione et al., 2004). Macropinocytosis is known to mediate the bulk uptake of membranes, fluid, and signaling receptors. Our findings suggest that macropinocytosis also can participate in the very rapid turnover of cell surface integrins, a pathway that is critical for stimulated cell migration. Materials and methods Mice and cell culture Wild-type C57BL/6 mice were maintained as homozygous inbred lines in the Dana-Farber Cancer Institute animal facility. All mice were male and 8–12 wk old at the time of the study. Mouse experiments followed Institutional Animal Care and Use Committee guidelines. Primary wild-type mouse synovial fibroblasts were recovered from normal mouse joints and were cultured in DME supplemented with 10% FBS (Hyclone), l-glutamine, penicillin/streptomycin, 2-mercaptoethanol, and essential and nonessential amino acids at 37°C under 10% CO2. Cells had fibroblast-like morphology and were VCAM-1 positive but lacked F4/80 and were CD45 negative by flow cytometry (Lee et al., 2007; Agarwal et al., 2008). Cell culture reagents were purchased from Invitrogen unless otherwise indicated. Retrieved mouse synovial fibroblasts were used between passages 5 and 10. NIH3T3 cells were purchased from American Type Culture Collection and cultured in DME supplemented with 10% bovine calf serum (Hyclone), l-glutamine, and penicillin/streptomycin at 37°C under 10% CO2. MDA-MB-231 human breast cancer cells were purchased from American Type Culture Collection and cultured in American Type Culture Collection–formulated Leibovitz’s L-15 medium supplemented with 10% FBS, l-glutamine, and penicillin/streptomycin at 37°C without CO2. HUV-EC-C human umbilical vein endothelial cells were purchased from American Type Culture Collection and cultured in American Type Culture Collection–formulated F-12K medium supplemented with 10% FBS (Hyclone), 0.1 mg/ml heparin (Sigma-Aldrich), 0.03–0.05 mg/ml endothelial cell growth supplement (Sigma-Aldrich), l-glutamine, and penicillin/streptomycin at 37°C under 5% CO2. Plasmids, siRNA, and transfection cDNA encoding the full-length mouse integrin β3–EGFP fusion protein in a pcDNA3 expression vector was a gift from B. Wehrle-Haller (University of Geneva, Geneva, Switzerland; Ballestrem et al., 2001). siRNA against BARS (Yang et al., 2005), CHC (Li et al., 2007), and WAVE1 (Suetsugu et al., 2003) were previously described. siRNAs against CAV1, Rab4A, and Rab4B were obtained from Santa Cruz Biotechnology, Inc. Nonsilencing negative control siRNA was purchased from QIAGEN. Plasmids were transfected with Lipofectamine LTX reagent (Invitrogen), and siRNAs were transfected with Lipofectamine RNAiMAX reagent (Invitrogen). Antibodies Armenian hamster anti–integrin β3 (clone 2C9.G2), mouse anti-EEA1, mouse anti-Rab5, mouse anti-Rab4, mouse anti-Rab11, mouse anti-WAVE1, mouse anti-CAV1, and rat anti–integrin β1 (clone Ha2/5) antibodies were purchased from BD. Nonblocking rat anti–integrin β3 antibody (clone 1–55-4) was obtained from MBL International. The rat anti–integrin β1 antibody (clone MB1.2) was obtained from Millipore. The mouse anti–integrin β1 antibody (clone 12G10) was obtained from Abcam. The rabbit anticortactin antibody was purchased from Sigma-Aldrich. The rabbit anti-BARS antibody (clone p50-2) was raised against GST-BARS and purified as described previously (Spanò et al., 1999). The mouse anti-CHC antibody was obtained from American Type Culture Collection. The mouse anti–β-actin antibody was obtained from Sigma-Aldrich. The mouse anti-MHC class I antibody was purchased from Santa Cruz Biotechnology, Inc. Alexa Fluor–conjugated anti–mouse, anti–rat, and anti–rabbit secondary antibodies and Alexa Fluor–conjugated phalloidin were purchased from Invitrogen. FITC-conjugated anti–Armenian hamster IgG secondary antibody and peroxidase-conjugated anti–mouse, anti–rat, and anti–rabbit secondary antibodies were obtained from Jackson ImmunoResearch Laboratories, Inc. Immunofluorescence (IF), dextran endocytosis assay, cell surface antibody-binding, acid-stripping endocytosis and recycling assay, and transferrin endocytosis assay For IF microscopy experiments, cells were cultured on 10-mm glass coverslips coated with 0.1 µg/cm2 vitronectin (Sigma-Aldrich) for integrin β3 experiments or 1 µg/cm2 fibronectin (Sigma-Aldrich) for integrin β1 experiments. PDGF-BB and VEGF were obtained from Sigma-Aldrich. EGF was obtained from Invitrogen. For IF staining, cells were serum starved for 16 h and stimulated with 20 ng/ml PDGF-BB, 30 ng/ml EGF, or 30 ng/ml VEGF in serum-free medium for various times as indicated in the following paragraphs. Cells were fixed in 4% paraformaldehyde (Electron Microscopy Sciences) for 15 min. For intracellular protein staining, cells were permeabilized with 0.3% Triton X-100 (Thermo Fisher Scientific) for 10 min. Permeabilized cells were labeled with primary antibodies for 1 h at RT or 16 h at 4°C followed by secondary antibody labeling for 1 h and mounted on slides with FluorSave reagent (EMD). Tetramethylrhodamine (TMR)-conjugated, 10,000-MW fixable dextran was obtained from Invitrogen. For the PDGF-BB– and TMR-dextran–chasing endocytosis assay, cells were serum starved for 16 h and then stimulated with 20 ng/ml PDGF-BB together with 200 µg/ml TMR-dextran in serum-free medium for various times. The cells were then washed three times to remove all cell surface–remaining TMR-dextran and fixed, and IF staining was performed. For the cell surface antibody-binding, acid-stripping endocytosis and recycling assay, surface integrin β3 on live cells was stained by a nonblocking rat anti–integrin β3 monoclonal antibody (clone 1–55-4) in serum-free medium containing 1% BSA as blocking reagent for 1 h at RT. Cells were washed in antibody-free medium three times. Then, cells were treated with 20 ng/ml PDGF-BB for 15 min to allow cells time to internalize the antibody-bound integrin β3 by CDR macropinocytosis. Then, cells were equilibrated to 4°C followed by a 1-min ice-cold acid (0.5% acetic acid and 0.5 M NaCl, pH 3.0)-washing step to remove leftover cell surface integrin β3–bound antibodies. After acid washing, cells were washed in ice-cold medium to bring the pH back to 7.4 as indicated by phenol red. Cells were then incubated at 37°C in 20 ng/ml PDGF-BB–containing medium for another 60 min to allow cells time to recycle the antibody-bound integrin β3 back to the cell surface. Cells were then fixed, and without cell membrane permeabilization, the recycled cell surface antibody-bound integrin β3 was detected by IF staining with an Alexa Fluor 488 goat anti–rat secondary antibody. After PBS wash, cell membranes were permeabilized with 0.3% Triton X-100 for 5 min. F-actin was stained by Alexa Fluor 568 phalloidin and visualized by confocal microscopy. For transferrin endocytosis assays, nonsilencing negative control siRNA or CHC siRNA-transfected primary mouse fibroblasts were incubated with Alexa Fluor 488 transferrin (Invitrogen) for 1 h at 4°C. Unbound transferrin was washed, and cells were then incubated at 37°C for various times. After incubation, cells were equilibrated to 4°C followed by a 1-min ice-cold acid (0.5% acetic acid and 0.5 M NaCl, pH 3.0)-washing step to remove leftover cell surface transferrin. After acid washing, cells were washed in ice-cold medium to bring the pH back to 7.4 as indicated by phenol red. Cells were then fixed and subjected to confocal microscopy to detect internalized Alexa Fluor 488 transferrin. 3D confocal microscopy and 4D time-lapse confocal live-cell imaging IF images were captured on an inverted microscope (TE2000-U; Nikon) equipped with a confocal system (C1; Nikon) controlled by EZ-C1 software (Nikon) using a Plan Apochromat 60×/1.40 NA oil objective or a Plan Apochromat 20×/0.75 NA objective (Nikon). 3D confocal microscopy was performed by scanning multiple confocal layers along the z axis. For 4D time-lapse confocal live-cell imaging experiments, cells were grown in 35-mm glass-bottomed Petri dishes (World Precision Instruments). These dishes were mounted onto the confocal microscope with a heating chamber at 37°C and were superfused with 10% CO2. 4D confocal microscopy was performed by 3D confocal microscopy scanning over time. Fluorescent 3D or 4D image reconstructions and protein colocalization analysis were performed in MetaMorph Imaging software (version 7.6.1.0; Universal Imaging). Flow cytometry Cells were detached by a brief incubation with 0.05% trypsin-EDTA (Invitrogen) at RT to minimize endocytosis. Trypsin was quenched by a trypsin inhibitor from Glycine max (soybean; Sigma-Aldrich). Cells were then washed by 2% FBS in PBS and stained with anti–integrin β3 monoclonal antibody (1–55-4) or isotype control antibody for 1 h at RT followed by Alexa Fluor 488 secondary antibody staining. After staining, cells were fixed in 1% paraformaldehyde in PBS and subjected to flow cytometry analysis by a flow cytometer (FACS Canto; BD). FlowJo software (Tree Star) was used to analyze the flow cytometry data. Cell migration assay 56 h after siRNA transfection, cells were serum starved for another 16 h and detached by a brief incubation with 0.05% trypsin-EDTA at RT. Trypsin was quenched by a trypsin inhibitor from Glycine max (soybean). Cells were washed and seeded with or without antiintegrin-blocking antibodies into the upper wells of vitronectin-coated Transwell inserts (8.0-µm pore size; Corning) in serum-free media. After a 2-h incubation at 37°C, 50 ng/ml PDGF-BB–containing serum-free media were then added into the lower well to drive cell migration. After a 2-h migration, the upper filter membrane surface was wiped to remove cells that had not migrated through the filter, and then the filter was fixed and stained to detect cells on the lower filter membrane surface using a stain set (Diff-Quik; Dade Behring). The number of cells that had migrated through a 0.8-mm2 Transwell membrane was counted. SDS-PAGE and Western blotting For SDS-PAGE experiments, cells were lysed and scraped at 4°C in a cell lysis buffer of the following composition: 1% Triton X-100, 50 mM Hepes, 150 mM NaCl, 5 mM EDTA, 5 mM EGTA, 20 mM NaF, 20 mM sodium pyrophosphate, 1 mM PMSF, and 1 mM Na3VO4, pH 7.4. DC protein assays (Bio-Rad Laboratories) were performed on cell lysate samples. Equal amounts of protein from each sample were run on each lane of 7.5% SDS-PAGE gels. After gel electrophoresis, proteins were transferred to a polyvinylidene fluoride membrane for Western blotting analysis. Proteins on the membranes were labeled with primary antibodies overnight at 4°C and then labeled by peroxidase-conjugated secondary antibodies and visualized by ECL detection reagents (Thermo Fisher Scientific). β-Actin was blotted as a protein loading control. Statistics Numerical data are presented as means ± SD. Student’s t test was used for the comparison of two means (P < 0.05 was considered significant as marked by asterisks in the figures). Online supplemental material Fig. S1 shows that integrin β3 and integrin β1 localize at CDRs in various cell types after growth factor stimulation. Fig. S2 shows that PDGF-BB–stimulated integrin β3 macropinocytosis is BARS dependent but clathrin and CAV1 independent. Video 1 shows a 3D view of CDRs with integrin β3, F-actin, and cortactin colocalization. Video 2 shows integrin β3–GFP translocation 4D tracing in a live NIH3T3 cell after PDGF-BB stimulation. Video 3 shows integrin β3–GFP translocation 4D tracing in a live NIH3T3 cell without PDGF-BB stimulation. Video 4 shows integrin β3–GFP translocation 4D tracing in a BARS siRNA-transfected live NIH3T3 cell after PDGF-BB stimulation. Online supplemental material is available at http://www.jcb.org/cgi/content/full/jcb.201007003/DC1.
