Cell mechanics control rapid transitions between blebs and lamellipodia during migration

Sublines of Walker Cells Can Form Either Lamellipodia or Blebs During Migration.

By selecting for or against adhesion we obtained two sublines of Walker cells: a suspension subline (suspSL) and an adherent subline (adhSL) (Fig. 1A, (15, 16)). Cells of both sublines displayed spontaneous polarization and formed protrusions at their leading edge. In order to characterize the nature of these protrusions, we expressed Lifeact, a marker of filamentous actin (17), and analyzed protrusion dynamics in living cells. We found that adhSL cells predominantly formed flat, actin-filled protrusions (Fig. 1 B and C). The presence of filamentous actin and the localization of components of the Arp2/3 complex to the leading edge of these protrusions (Fig. S1A) identified them as lamellipodia. In contrast, suspSL cells almost exclusively formed spherical membrane blebs initially devoid of filamentous actin (Fig. 1 B and C). Thus, Walker cells can be induced to form either lamellipodia or blebs by varying their culture conditions.

We then analyzed the ability of the cells from the two sublines to migrate in different environments. AdhSL cells could migrate efficiently on two-dimensional (2D) substrates (Fig. 1D and Movie S1), as suggested by previous studies (16). In contrast, suspSL cells were nonadherent and displayed uncoordinated, random movements when placed on 2D substrates (Movie S2), but were able to migrate in confined environments, such as inside a micropipette or when placed between glass and agarose (Fig. 1D and Movies S3 and S4). Thus, the two Walker sublines represent a good model system to investigate how transitions between protrusion types can be induced during cell migration.

Lamellipodia-Forming Cells Have a Lower Cortical Tension than Bleb-Forming Cells.

Cortical tension generates the intracellular pressure driving bleb growth, and a threshold tension is required for bleb expansion (18). We thus asked if adhSL cells did not form blebs because their cortical tension was too low. We measured tension in the sublines by micropipette aspiration (19) and found that suspSL cells had a cortical tension about two times higher than adhSL cells (Fig. 2A and Table S1). Tension measurements were performed on detached, rounded adhSL cells (19). Earlier work on fibroblasts showed that spreading leads to a slight decrease in cortical tension (20), therefore the difference in tension measured here is likely to be an underestimate of the actual difference. Treating cells with the Rho kinase (ROCK) inhibitor Y27632 or with blebbistatin, an inhibitor of myosin II activity, significantly decreased tension in both sublines (Table S1). Upon myosin inhibition, tension decreased to the same basic value in the two sublines, indicating that differences in myosin activity could explain the difference in cortical tension between adhSL and suspSL cells. In summary, we found that myosin-driven cortical tension was significantly lower in adhSL than in suspSL cells.

Increasing Cortical Tension Induces Bleb Formation in Lamellipodia-Forming adhSL Cells.

We then investigated if the observed difference in cortical tension was responsible for the absence of bleb formation in adhSL cells. To this aim, we first performed laser ablations of the cortex, which induce bleb expansion if cortical tension exceeds a threshold value (18). Cortex ablation in the front part of polarized cells induced bleb formation in 90% of suspSL cells and in 14% of adhSL (Fig. 2B), indicating that most adhSL cells cannot form blebs even upon induced rupture of the actin cortex. However, ablation may fail to trigger bleb formation in adhSL cells because of their spread morphology, as lamellipodia extension and substrate adhesions might limit cortical contractions necessary for bleb formation. We thus repeated the ablation experiments on adhSL detached from the substrate, which displayed a round shape with no protrusions, morphologically identical to unpolarized, round suspSL. Cortex ablation effectively induced bleb formation in round suspSL cells but not in detached, round adhSL cells (63% vs. 15%, Fig. 2B). Taken together, our ablation experiments suggest that adhSL Walker cells cannot form blebs.

We then investigated whether increasing cortical tension is sufficient to induce bleb formation in adhSL cells or if other factors limit bleb extension in these cells. We transfected adhSL cells with a constitutively active version of the kinase ROCK, ROCK-Δ3, which increases myosin regulatory light chain phosphorylation, and thus myosin activity, and enhances blebbing in the parental Walker cell line (21). Expression of ROCK-Δ3 resulted in a rounder morphology of adhSL cells. Some cells spontaneously formed blebs and almost all displayed bleb formation upon ablation (Fig. 2 B and C). Thus, the limited ability of adhSL cells to form blebs appears to be due to their low cortical tension, and increasing contractility is sufficient to trigger bleb formation in these cells.

Expression of ROCK-Δ3 in adhSL cells also reduced lamellipodia formation (Fig. 2C). Therefore, we asked whether contractility limits lamellipodia extension. We reduced cortical tension in adhSL cells with the ROCK inhibitor Y27632, or with the myosin inhibitor blebbistatin and found that the size of lamellipodia was significantly increased (Fig. 2 D and E). Thus, cortical tension appears to limit lamellipodia extension, suggesting that the protrusive mechanisms leading to the formation of blebs and lamellipodia are mutually exclusive.

Decreasing Arp2/3 Activity Induces Switching from Lamellipodia to Blebs.

We next investigated the contribution of protrusive actin polymerization to protrusion formation in the two sublines. The Arp2/3 complex is a major nucleator of actin filaments in the lamellipodium (22). We thus inhibited Arp2/3 activity using the small molecule inhibitor CK-666 (23). Treatment with CK-666 led to an instantaneous and dose-dependent decrease of the percentage of adhSL cells forming lamellipodia, and strongly increased bleb formation (Fig. 3 A and B and Movie S5). To rule out nonspecific effects of the inhibitor, we decreased Arp2/3 levels using siRNA. Upon depletion of the Arp3 subunit in adhSL cells, we observed a switch from lamellipodia to blebs, similar to what was observed after treatment with CK-666 (Fig. S1 B and C, and Movie S6). These results indicate that inhibiting Arp2/3 function is sufficient to induce rapid transitions from lamellipodia to blebs in adhSL cells. Interestingly, interference reflection microscopy, which can be used to estimate the distance between the cell and the substrate (24), indicated that the transitions from lamellipodia to blebs occurred with no apparent change in adhesion to the substrate (Fig. 3A). Moreover, blebs formed at the locations where lamellipodia were previously formed (Movie S5), suggesting that the mechanisms underlying cell polarity are similar in bleb- and lamellipodia-forming cells.

The increase in blebbing upon Arp2/3 inhibition seemed inconsistent with the hypothesis that cortical tension was too low for bleb formation in adhSL cells (Fig. 2). We therefore measured tension in CK-666 treated adhSL cells and found that it was higher than in control cells (Table S1). Interestingly, previous studies in Caenorhabditis elegans show a cortical localization of the Arp2/3 complex, and depletion of Arp2/3 subunits results in bleb formation (25, 26). Similarly, an Arp2 mutant strain of Dictyostelium shows enhanced blebbing (27). These data suggest that in addition to controlling actin protrusivity and lamellipodia extension, Arp2/3 activity influences bleb formation via controlling the cellular actin cortex and cortical tension.

As tension appears to limit lamellipodia extension (Fig. 2 D and E), we asked whether the decrease in lamellipodia upon CK-666 treatment was the result of decreased actin protrusivity or increased tension. We decreased contractility in adhSL cells with Y27632 prior to CK-666 treatment. We found that blocking Arp2/3 activity in Y27632 treated cells rapidly decreased lamellipodia outgrowth and resulted in round cells without any protrusions (Fig. 3C). Taken together, our results indicate that Arp2/3 inhibition decreases lamellipodia formation independently of the increase in tension, while the tension increase is essential for bleb formation.

We then asked whether the Arp2/3 complex plays any role in bleb formation and blebbing-based migration. We quantified the protrusions and analyzed the migration of suspSL cells treated with CK-666, and found no significant differences between treated and untreated cells (Fig. S2). These observations indicate that the activity of the Arp2/3 complex is not required for bleb formation and blebbing-based migration of suspSL cells.

Rac1 Activation Triggers Switching from Blebs to Lamellipodia.

Next, we investigated if enhancing actin polymerization could induce switching from bleb to lamellipodia formation in suspSL cells. To address this question, we expressed a constitutively active, photoactivatable version of the small RhoGTPase Rac1 (PA-Rac1) in Walker cells (28). Rac1 is a major activator of the Arp2/3 complex and enhances actin polymerization (28, 29). Activation of Rac1 in adhSL cells increased lamellipodia formation and membrane ruffling, demonstrating that this construct promotes actin protrusivity in Walker cells (Movie S7). In suspSL cells, global activation of Rac1 led to two related phenotypes: 50% of the cells with intermediate expression levels of PA-Rac1 immediately stopped blebbing (Fig. 3D top and Table S2 and Movie S8), the remaining 50% of the cells stopped blebbing and formed one or more lamellipodia at the location where blebs were previously formed (Fig. 3D middle and Table S2 and Movie S9). Lamellipodia formation was dependent on Arp2/3 activity as Rac1 activation in suspSL cells pretreated with the Arp2/3 inhibitor CK-666 did not result in lamellipodia growth (Table S2 and Movie S10). Substrate adhesions were not required for Rac1-induced switching from blebs to lamellipodia, as lamellipodia formation could also be triggered in suspSL cells placed between agarose and PEG-coated glass, which precludes substrate attachment (Fig. 3D bottom and Table S2). In summary, PA-Rac1 activation was sufficient to trigger the formation of lamellipodia in suspSL cells that normally form blebs, and the formation of lamellipodia occurred without changes in cell polarity or substrate adhesiveness.

Interestingly, Rac1 activation resulted in an increase of the cell cross-sectional area, which then rapidly decreased when activation was stopped (Movie S9). This observation suggests that active Rac1 may decrease cortical tension, which would in turn lead to a reduction in bleb formation. Such a tension decrease could be caused by a reduction in RhoA activity, as PA-Rac1 activation has been shown to globally inhibit RhoA (28). As contractility limits lamellipodia extension in adhSL cells (Fig. 2 D and E), we checked whether decreasing contractility alone would be sufficient to switch from blebs to lamellipodia in suspSL cells. We inhibited RhoA activity with a cell permeable version of the exoenzyme C3 Transferase, or blocked Myosin II activity with blebbistatin. Both treatments induced a strong decrease in blebbing but no significant increase in lamellipodia formation (Fig. S3). Thus, inhibition of RhoA or Myosin II was not sufficient to change the protrusion type of Walker cells from blebs to lamellipodia.

Taken together, our results indicate that activation of Rac1 causes both an increase in Arp2/3-driven actin protrusivity and a decrease in contractility. This dual effect results in an immediate switch from bleb to lamellipodia formation in suspSL cells.

Changes in Substrate Adhesion Can Induce Dynamic Changes in Protrusion Formation.

Our data suggest that modulating the relative magnitudes of actin protrusivity and actomyosin contractility is sufficient to induce transitions between blebs and lamellipodia in migrating cells (Fig. 4). These transitions could be triggered without any changes in cell adhesion to the substrate (Fig. 3 A and D). However, substrate adhesiveness has been proposed to be a major factor influencing cell shape and migration mode (1, 10, 11). We thus asked whether abrupt changes in substrate adhesion could induce switches between blebs and lamellipodia in Walker cells, and on what time scales such changes would occur. Detaching lamellipodia-forming adhSL cells from the substrate did not result in bleb formation even upon cortex ablation (Fig. 2B). Therefore, we focused on exploring the effect of increasing substrate attachment in bleb forming suspSL cells. We first analyzed the protrusions formed by supSL cells migrating in confinement on adhesive, fibronectin coated, surfaces. We found an increased proportion of cells forming both blebs and lamellipodia compared to suspSL cells on glass or a nonadhesive substrate (Fig. 5 A and B), indicating that adhesive substrates trigger lamellipodia formation in these cells. We then investigated how such transitions occur at the single cell level and observed cells migrating over substrates with changing adhesiveness. To this aim, we engineered micropatterned surfaces with alternating adhesive and nonadhesive regions, using microcontact printing (for details see SI Text). We observed that blebbing suspSL cells crossing region boundaries formed lamellipodia immediately after contacting adhesive areas (50% of the cells, n = 18, Fig. 5 C–E and Movie S11). Lamellipodia disappeared quickly when cells moved onto a nonadhesive region. Interestingly, bleb formation seemed unaffected by changes in substrate adhesion (Fig. 5D), suggesting that adhesion mainly influences actin polymerization. Taken together, our data indicate that changes in substrate adhesiveness can induce immediate and reversible changes in protrusion types.

Culture, Transfection, and Treatment of Cells.

Walker 256 carcinosarcoma cells were a gift from V. Niggli. Cells were grown in RPMI 1640 medium supplemented with 10% FCS, 1% penicillin-streptomycin and 2 mM glutamine (all GIBCO, Invitrogen) at 37 °C and 5% CO₂. Selection of the sublines was done as described in ref. 16, except that nonadherent suspSL cells were grown in ultra-low-attachment surface flasks (Corning Life Sciences) to prevent any substrate adhesion. Reagents and plasmids, cell preparation for microscopy, cell transfection, treatments, and microcontact printing are described in the SI Text.

Laser Ablation, Photoactivation, Imaging, and Tension Measurements.

Laser ablation experiments were performed on a scanning confocal microscope (Olympus FV1000, UPlanSApo 60x NA 1.35 Oil Objective) using a 405 nm picosecond pulsed laser as described previously (18). Photoactivation of the activatable Rac1 construct (28) was done using the same system, either with a 458 nm laser for global activation or with the 405 nm ablation laser (set to minimal power) to locally activate a spot in the cell. For differential interference contrast (DIC) and fluorescence imaging a 488 nm or a 561 nm laser was used as a light source. Interference reflection microscopy was performed as described previously (24) using the Olympus FV1000 system described above. Phase contrast observations and Histone2B-GFP imaging were performed on an inverted Zeiss Axiovert 200M wide-field microscope using a Zeiss LD Achroplan 20x 0.40 Ph2 or a Zeiss Ph2 Plan Neofluar 40x 0.75 objective. Environmental control (37 °C, 5% CO₂) was applied for all imaging setups. Image processing and cell tracking are described in the SI Text. Cortical tension was measured by micropipette aspiration as described previously (18).