Mechanosensitive calcium signaling in filopodia

Filopodia are ubiquitous membrane projections that play crucial role in guiding cell migration on rigid substrates and through extracellular matrix by utilizing yet unknown mechanosensing molecular pathways. As recent studies show that Ca2+ channels localized to filopodia play an important role in regulation of their formation and since some Ca2+ channels are known to possess mechanosensing properties, activity of filopodial Ca2+ channels might be tightly interlinked with the filopodia mechanosensing function. We tested this hypothesis by monitoring changes in the intra-filopodial Ca2+ level in response to application of stretching force to individual filopodia of several cell types. It has been found that stretching forces of tens of pN strongly promote Ca2+ influx into filopodia, causing persistent Ca2+ oscillations that last for minutes even after the force is released. Most of the known mechanosensitive Ca2+ channels, such as Piezo 1, Piezo 2 and TRPV4, were found to be dispensable for the observed force-dependent Ca2+ influx. In contrast, L-type Ca2+ channels appear to be a key component in the discovered phenomenon. Since previous studies have shown that intra-filopodial transient Ca2+ signals play an important role in guidance of cell migration, our results suggest that the force-dependent activation of L-type Ca2+ channels may contribute to this process. Overall, our study reveals an intricate interplay between mechanical forces and Ca2+ signaling in filopodia, providing novel mechanistic insights for the force-dependent filopodia functions in guidance of cell migration. Significance statement We found that tensile forces of tens of pN applied to individual filopodia trigger Ca2+ influx through L-type Ca2+ channels, producing persistent Ca2+ oscillations inside mechanically stretched filopodia. Resulting elevation of the intra-filopodial Ca2+ level in turn leads to downstream activation of calpain protease, which is known to play a crucial role in regulation of the cell adhesion dynamics. Thus, our work suggests that L-type channel-dependent Ca2+ signaling and the mechanosensing function of filopodia are coupled to each other, synergistically governing cell adhesion and motion in a force-dependent manner. Since L-type Ca2+ channels have been previously found in many different cell types, such as neural or cancer cells, the above mechanism is likely to be widespread among various cell lines.


Introduction
Interestingly, it has been found that filopodia of such cells occasionally produce Ca 2+ 136 bursts (see Movie 1). However, frequency of these events appeared to be very low -137 measurements showed that only ~ 12 % of filopodia (i.e., 11 filopodia out of N = 92 138 monitored) generated 1 or 2 short bursts of transient Ca 2+ signals lasting only for ~ 18 139 ± 5 s (mean ± s.e.m.) during 5-6 min observation period. 140 To check whether attachment of fibronectin-coated polystyrene microbeads to 141 filopodia has any effect on Ca 2+ signal behavior, we used optical tweezers to put 142 microbeads onto the tips of individual filopodia, holding them there for ~ 2-3 sec to 143 initiate beads interaction with filopodia before turning the trap off (see schematic Figure   144 1C). Such simple binding of microbeads to filopodia alone in the absence of applied 145 mechanical load did not change significantly the filopodial Ca 2+ firing rate -only 25% 146 of filopodia with microbeads produced a single transient Ca 2+ signal during ~ 5-6 min 147 observation period (the total number of monitored beads was N = 8), see Movie 1. 148 Interestingly, all the beads attached to filopodia moved in the direction towards the 149 cell body at a rate of 26.9 ± 2.5 nm/s (mean ± s.e.m.), which is similar to the rate of  Typical view of a wild-type (WT) HEK-293 cell transfected with mApple-myosin X construct, which induces filopodia formation. In the figure, mApple-myosin X, which usually clusters at the filopodia tips, is shown in red color; whereas, actin filaments, which are labeled with F-tractin-GFP, are shown in green. C. Schematic illustration of the optical tweezers' experimental setup. Optically trapped microbeads coated with either fibronectin or concanavalin A (ConA) were used to form an adhesion contact with the tip of a filopodium. Filopodium stretching was commenced by moving the microscope stage with the cell body away from the axis of the laser beam. This resulted in generation of a pulling force, F, on the filopodium tip. D. Fibronectin-coated microbead attached to the filopodium tip pulled by the optical trap away from the cell edge. Filopodium stretching-induced Ca 2+ signal indicated by GCaMP6f Ca 2+ sensor (shown in green) as well as mApple-myosin X (shown in red) can be clearly seen from the figure. E. Representative time courses of the pulling force and the length change of a stretched filopodium when the cell is moved away from the optical trap at a speed of ~ 5 nm/s. F. Time-dependent changes in the pulling force applied to the filopodium (data in red) and the Ca 2+ sensor intensity (data in blue) corresponding to the experiment shown in panel E. 158 Next, we investigated the role of extracellular mechanical forces in filopodia- In the presence of mechanical stretching, intense and highly dynamic Ca 2+ signals 166 were observed in ~ 82% of pulled filopodia (9 out of N = 11 tested filopodia from 167 different cells) that lasted for many minutes, which was in strong contrast to rare Ca 2+ 168 signals produced by unperturbed filopodia (~ 12%, 7 out of N = 57 filopodia) or by 169 filopodia bound to microbeads in the absence of applied mechanical load (~ 25%, 2 out 170 of N = 8 filopodia). Figure 1D and Movies 2A and 2B show a typical example of a 171 mechanically stretched filopodium with a clearly visible force-induced Ca 2+ signal. 172 Interestingly, from Movie 2B it can be seen that the Ca 2+ signal was propagating from 173 the tip of the stretched filopodium towards the cell body, which was a general trend in 174 the pulled filopodia.  Figure S1).

183
Once the Ca 2+ signal appeared, it could persist for several minutes regardless of 184 whether the force was retained or released (see Figure 2C, middle panel), suggesting 185 that filopodia-dependent Ca 2+ signaling system has a memory effect, and mechanical  Very similar force-dependent Ca 2+ signals were also observed in MCF-7 and A2058 191 cells co-transfected with GCaMP6f and mApple-myosin X constructs, see Figure S2. 192 Thus, we conclude that mechanical stretching of filopodia strongly promotes the 193 appearance probability of filopodia-generated Ca 2+ signals not only in 194 but also in other cell lines.

195
Force-dependent Ca 2+ signals were also found to be produced by filopodia, whose and not specific to myosin X-induced or Cdc42-induced filopodia. 199 We then checked whether such Ca 2+ signals were induced by bona fide A general view of a mechanically stretched filopodium. In the figure, mApple-myosin X is shown in red color; whereas, green color indicates Ca 2+ sensor, GCaMP6f. B. Sequence of frames demonstrating change of the Ca 2+ sensor intensity in the mechanically stretched filopodium, which is shown in panel A. C. Left 3D graph displays changes in the Ca 2+ sensor intensity along the stretched filopodium as a function of time and distance from the filopodium tip. The middle panel shows a top view on the 3D graph from the left panel. Intra-filopodial Ca 2+ oscillations in the form of periodic vertical strokes can be clearly seen in the graph. Once initiated, such Ca 2+ oscillations keep going for several minutes even without further mechanical stretching of the filopodium when the optical trap is switched off. Right panel shows a representative time trace of the average Ca 2+ sensor intensity at the filopodium tip, which reveals regular Ca 2+ oscillations with a period of ~ 10 s.  Table T1. Altogether, these results suggest 212 that Piezo 1 and most probably Piezo 2 do not make a significant contribution to the 213 observed force-dependent Ca 2+ signals in filopodia.

214
Interestingly, intra-filopodial Ca 2+ signals were also found to be independent from 215 the integrin-mediated adhesion, since they were observed in filopodia stretched by 216 using ConA-coated microbeads that do not activate integrin-related protein assembly at  Finally, it should be noted that in ~ 40% (i.e., 4 out of N = 11) of experiments initial 219 intra-filopodial Ca 2+ signal resulted in strong elevation of the Ca 2+ level in nearby cell Heatmap of the Ca 2+ sensor intensity as a function of time and position on the filopodium, which is shown in panels A and B. Strong Ca 2+ signal caused by filopodia stretching and its time-dependent oscillations can be clearly seen from the graph as well as from the frames shown in panel B. D. View of a mechanically stretched filopodium of a HEK-293 Piezo 1 KO cell in Ca 2+ -free cell culture medium in the presence of 5 mM EGTA. E. Sequence of frames demonstrating changes of the Ca 2+ sensor (GCaMP6f) intensity in the mechanically stretched filopodium, which is shown in panel D. F. Heatmap of the intensity of Ca 2+ sensor as a function of time and position on the filopodium, which is shown in panels D and E. As can be seen from the graph and frames presented in panel E, no Ca 2+ signals have been observed inside the filopodium shaft upon mechanical stretching in the absence of free Ca 2+ ions in the cell culture media. In panels A and D, mApple-myosin X is shown in red color, and Ca 2+ sensor, GCaMP6f, is indicated in green color.     Indeed, qPCR assay performed on HEK-293 Piezo 1 KO cells has shown that the 276 average mRNA level of CACNA1C gene, which encodes the pore-forming α1C subunit 277 of L-type Ca 2+ channels, is more than 10 times higher than that of Piezo 2 (see Table   278 T1), suggesting that L-type Ca 2+ channels may potentially have a stronger effect on the  Heatmaps of the Ca 2+ sensor intensity as a function of time and position on the filopodium that correspond to the cells presented in panels A, D and G, respectively. In panels A, D and G, mApple-myosin X is shown in red color, and Ca 2+ sensor, GCaMP6f, is indicated in green color. Altogether, these results strongly suggest that L-type Ca 2+ channels play a critical role 317 in formation of the force-induced Ca 2+ signals in filopodia.          In summary, our study clearly demonstrates that Ca 2+ influx into filopodia is a basic 530 signaling response to physiological level of tensile forces applied to filopodia tips and 531 therefore can be used by cells in different processes of mechano-orientation and motion guidance, involving response to the matrix rigidity and topography. Moreover, we 533 established that in cells of different types such response is mediated by L-type Ca 2+ 534 channels rather than known mechanosensing Ca 2+ channels like TRPV4, and Piezo 1 535 and Piezo 2. Involvement of L-type Ca 2+ channels in different type of mechanosensory 536 mechanisms is an interesting avenue for the future studies.  Δx is the measured deviation of the bead center from the axis of the optical trap (see 583 Figure 1C). influx into mechanically stretched filopodia.

622
As for X and/or Y microscope stage movements that were necessary to generate a 623 pulling force on filopodia, to minimize potential measurement errors that may arise