L-type calcium channels regulate filopodia stability and cancer cell invasion downstream of integrin signalling

ARTICLE Received 7 Mar 2016 |Accepted 19 Sep 2016 |Published 2 Dec 2016 L-type calcium channels regulate filopodia stability and cancer cell invasion downstream of integrin signalling Guillaume Jacquemet1, Habib Baghirov1,w, Maria Georgiadou1, Harri Sihto2, Emilia Peuhu1, Pierre Cettour-Janet1, Tao He3,w, Merja Pera ¨la ¨3,w, Pauliina Kronqvist4, Heikki Joensuu2,5 & Johanna Ivaska1,6 Mounting in vitro,in vivo and clinical evidence suggest an important role for filopodia in driving cancer cell invasion. Using a high-throughput microscopic-based drug screen, we identify FDA-approved calcium channel blockers (CCBs) as potent inhibitors of filopodia formation in cancer cells. Unexpectedly, we discover that L-type calcium channels are functional and frequently expressed in cancer cells suggesting a previously unappreciated role for these channels during tumorigenesis. We further demonstrate that, at filopodia, L-type calcium channels are activated by integrin inside-out signalling, integrin activation and Src. Moreover, L-type calcium channels promote filopodia stability and maturation into talin-rich adhesions through the spatially restricted regulation of calcium entry and subsequent activation of the protease calpain-1. Altogether we uncover a novel and clinically relevant signalling pathway that regulates filopodia formation in cancer cells and propose that cycles of filopodia stabilization, followed by maturation into focal adhesions, directs cancer cell migration and invasion. DOI: 10.1038/ncomms13297 OPEN 1Turku Centre for Biotechnology, University of Turku, FIN-20520 Turku, Finland. 2Laboratory of Molecular Oncology, Translational Cancer Biology program, University of Helsinki, FIN-00290 Helsinki, Finland. 3VTT Medical Biotechnology, Technical Research Centre of Finland, FIN-20520 Turku, Finland. 4Department of Pathology, University of Turku and Turku University Hospital, FIN-20520 Turku, Finland. 5Department of Oncology, Helsinki University Hospital, FIN-00290 Helsinki, Finland. 6Department of Biochemistry, University of Turku, FIN-20520 Turku, Finland. wPresent addresses: Norwegian University of Science and Technology, NO-7491 Trondheim, Norway (H.B.); Radiometer Turku Oy, 20750 Turku, Finland (T.H.); Natural Resources Institute Finland, FI-31600 Jokioinen, Finland (M.P.). Correspondence and requests for materials should be addressed to G.J. (email: guillaume.jacquemet@utu.fi) or to J.I. (email: Johanna.ivaska@utu.fi). NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications 1 Cell motility is involved at every stage of tumorigenesis and contributes to primary tumour growth, cancer cell dissemination and metastasis formation1,2. As metastasis remains the leading cause of cancer-related morbidity in patients with solid tumours3, there is an immediate need to gain a more comprehensive understanding of the cellular structures and signalling pathways that drive cancer cell migration. To migrate, cells interact and sense the surrounding extracellular matrix (ECM) via transmembrane adhesion receptors such as integrins4–6. Integrin function is controlled by a conformational switch between active and inactive states that determine ECM ligand interaction and subsequent receptor signalling5. Integrin activation from within the cell (integrin inside-out signalling) is promoted by several mechanisms including the Rap1-RIAM-talin pathway and leads to integrin-ECM engagement (integrin outside-in signalling) and the recruitment and activation of a large number of proteins including the oncogenic kinases focal adhesion kinase (FAK) and Src to the integrin4,7. Filopodia are actin-rich finger-like protrusions that extend from the plasma membrane and have been implicated in cell migration and invasion both in vitro and in vivo8. Filopodia are assembled at the front of invading cancer cells8–10 and filopodia-like structures promote cancer cell survival at metastatic sites11,12. Several filopodia-inducing proteins such as the molecular motor myosin-X (MYO10) or the actin-bundling protein fascin promote cancer cell invasion both in vitro and in mouse models and are associated with poor patient prognosis in multiple carcinoma types8,13,14. Thus, interfering with filopodia formation could be a viable strategy to inhibit cancer metastasis in vivo.MYO10 is a homodimeric molecular motor which is upregulated in breast cancer where its expression correlates with mutant p53, poor prognosis and increased metastatic potential13,15. Monomeric MYO10 is inactive and localizes to the cytosol or to Rab7-positive vesicles16.MYO10 activation, promoted by PI(3,4,5)P3, results in motor dimerization and drives filopodia formation by transporting actin regulators, cell– cell adhesion receptors and integrins to filopodia tips. Here, we describe a novel druggable and clinically relevant pathway regulating MYO10-positive filopodia formation and stability. Unbiased high-throughput microscopy screens reveal that L-type calcium channels, through regulation of calcium entry at filopodia tips, drive filopodia stabilization. Unexpectedly, L-type calcium channels are expressed and frequently altered in many human cancers and contribute to cancer cell invasion by regulating filopodia downstream of b1 integrin and Src activation. Results L-type calcium channel blockers inhibit filopodia formation. To identify novel regulators of filopodia formation, cancer cells expressing MYO10-GFP (to induce and visualize filopodia) were treated with a library comprising over 500 compounds for 1 h and imaged using high-throughput microscopy. The number of MYO10-positive spots was automatically quantified to determine the average number of filopodia per cell (Supplementary Fig. 1A–D, see methods for details). From this screen, several L-type calcium channel blockers (CCBs) were identified as compounds that consistently inhibit filopodia formation (Supplementary Figs 1D and 2A). In validation experiments, four structurally distinct CCBs (amlodipine besylate, felodipine, manidipine dichloride and cilnidipine) were demonstrated to significantly reduce the number of MYO10-induced filopodia in breast cancer cells as efficiently as a PI3K inhibitor (positive control to block MYO10 activity16), whereas a treatment with zonisamide (inhibits t-type calcium channels, voltage-gated sodium channels and carbonic anhydrase) or bumetanide (inhibits the Naþ/Kþ/2Cl cotransporter) failed to affect filopodia number (Fig. 1a). Similar results were obtained in pancreatic cancer cells following CCB treatment (Supplementary Fig. 2B). In addition, overall inhibition of calcium entry into cells by EGTA-mediated chelation of extracellular calcium dramatically reduced the number of MYO10-induced filopodia (Supplementary Fig. 2C). Together, these data indicate that calcium entry into cells via L-type calcium channels positively regulates filopodia formation in cancer cells. L-type calcium channel expression in cancer cell lines. The identification of L-type calcium channels as regulators of filopodia formation in cancer cells was unexpected as their expression and activity is principally thought to be restricted to excitable cells17,18. L-type calcium channels are composed of multiple subunits (a1, a2d,band g), of which the a1 subunit forms the core channel transporting calcium across the plasma membrane and the other subunits form regulatory components17,19. Four different genes (CACNA1C,CACNA1D, CACNA1F and CACNA1S) encode the a1 subunit that is targeted by CCBs. Interestingly, all four genes were found to be widely expressed at variable levels across cancer cell lines20 (Fig. 1b) regardless of their tissue of origin (Supplementary Fig. 3A–C). Furthermore, we found that L-type calcium channels are functional in cancer cell lines MDA-MB-231 and PDAC p53R172H as treatment with a specific L-type calcium channel activator (BAY K8644) triggered a rapid and transient increase in intracellular calcium throughout the cell body (Fig. 1c; Supplementary Fig. 3D; Supplementary Movies 1 and 2) (3–4 fold increase at 1 min post stimulation; detected with a GFP-based calcium probe21) and at filopodia tips (Fig. 1d). Importantly, calcium entry mediated by the L-type calcium channel activator was inhibited in the presence of an L-type calcium channel inhibitor (Fig. 1e). L-type calcium channels are clinically relevant in cancer. Analysis of public datasets using cBioPortal22,23 revealed that L-type calcium channels are commonly altered in patient samples of different cancer types (Supplementary Fig. 4A). In particular, over 29% of patient samples in the Breast Invasive Carcinoma24 data set displayed alterations in CACNA1C,CACNA1D, CACNA1F or CACNA1S and these alterations showed a significant association with unfavourable patient survival (Supplementary Fig. 4B–D). Interestingly, the worst survival rates were observed when alterations in CACNA1C,CACNA1D and CACNA1S were analysed together (Supplementary Table 1). Furthermore, while all four L-type calcium channel a1 subunits are expressed at low levels in both healthy breast and breast carcinoma samples (IST Online), CACNA1D is the most commonly overexpressed a1 subunit in breast carcinoma while CACNA1F is often downregulated (Supplementary Fig. 4E). CACNA1D expression was also found to be upregulated in breast cancer Oncomine data sets25 and was the most commonly expressed L-type calcium channel a1 subunit in breast and pancreatic cancer cell lines (Supplementary Fig. 3B,C). Taken together, these data indicate that L-type calcium channels are frequently altered in breast cancer samples and that alteration in these genes may correlate with poor prognosis. The expression of the individual L-type calcium channel a1 subunit in clinical samples will require further studies using specific antibodies. L-type calcium channels, cancer cell migration and invasion. Filopodia support three-dimensional (3D) cell migration and cancer invasion, particularly in cancers harbouring p53 ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 2NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications mutations8,13. Importantly, similar to PI3K inhibition, CCB treatment of p53 mutant breast and pancreatic cancer cells (MDA-MB-231, P53R280K; PDAC, P53R172H; Su.86.86, P53G245S) significantly impaired cancer cell invasion (Fig. 2a–c), in a concentration-dependent manner (Supplementary Fig. 5A,B). In contrast, treatment with zonisamide or bumetanide had no effect on cancer cell invasion (Fig. 2a,b). In addition, upregulation of MYO10 expression lead to filopodia formation and was sufficient a Zonisamide Bumetanide MYO10-GFP Actin MYO10-GFP Actin MYO10-GFP Actin MYO10-GFP Actin DMSO Manidipine Amlodipine Zonisamide Bumetanide LY294002 MYO10-GFP Actin MYO10-GFP Actin MYO10-GFP Actin MYO10-GFP Actin 0 20 40 60 80 100 120 DMSO Manidipine dichloride Cilnidipine Felodipine Amlodipine LY294002 Number of MYO10-positive filopodia per cell T-type calcium channel blocker Sodium channel inhibitor L-type calcium channel blockers DMSO PI3K inhibitor *** *** *** *** *** CACNA1S CACNA1C CACNA1F CACNA1D MCF10DCIS.com MCF-7 MDA-MB-453 Caco 2 MDA-MB-468 QGP-1 SNU-719 MDA-MB-361 PANC-1 Capan-2 Capan-1 DMS-53 Hs-683 SUM-1315M02 U-937 A2780 U-698-M BT-474 HPAC NCI-H889 NCI-H2171 COR-L279 U2OS SW-480 A-375 SU.86.86 HeLa K-562 NCI-H82 MDA-MB-231 A-431 COLO-704 b ND Gene expression read counts Cell lines: CACNA1S CACNA1C CACNA1F CACNA1D c –1 min +1 min +10 min Calcium probe BAY K8644 Calcium probe Fold increase in calcium probe intensity 0 2 4 6 8 1 min 10 min de Fold increase in calcium probe intensity DMSO Amlodipine 0 2 4 6*** ROI MYO10-mCherry– 1 min –1 min BAY K8644 +1min Calcium probe MYO10-mCherry –1 min BAY K8644 +1min 11 Calcium probe BAY K8644 FelodipineCilnidipine NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 ARTICLE NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications 3 to drive cancer cell invasion in non-invasive cell lines in a CCB sensitive manner (Fig. 2d; Supplementary Fig. 5C,D). Functional L-type calcium channels and PI3K activity were also important for regulation of directionality during 3D cell migration on cell-derived matrices (Fig. 2e). Thus, MYO10 expression facilitates filopodia formation and invasion whereas inhibition of L-type calcium channels using CCBs or inhibition of PI3K impedes filopodia formation, directional cell migration and cancer cell invasion. Of the L-type calcium channel a1 subunits, expression of CACNA1D,CACNA1S, but not CACNA1C, was detected in MDA-MB-231 cells at the mRNA level (Fig. 3a; CACNA1F was not tested as it appeared to be often downregulated in breast cancer clinical samples). Importantly, individual silencing of CACNA1D or CACNA1S gene expression using multiple siRNA oligos (Supplementary Fig. 6A,B) decreased both filopodia formation and cancer cell invasion (Fig. 3b,c), indicating that CACNA1D and CACNA1S are likely to synergistically support filopodia formation and cancer cell invasion in MDA-MB-231 cells. The cell biological reagents to study L-type calcium channels are rather limited and thus we were only able to analyse CACNA1D localization (Supplementary Fig. 6B). Importantly, endogenous CACNA1D localized with MYO10 at filopodia tips (Fig. 3d). In addition, CACNA1D protein expression could be detected in all major breast cancer subtypes (cancer type, positive expression; luminal, 3/6; triple-negative, 7/8; Her2-positive, 4/4; healthy breast, 0/1) from a set of breast carcinoma patient samples (Fig. 3e). Therefore, L-type calcium channels, in particular CACNA1S and CACNA1D are expressed in breast cancer and contribute to filopodia formation and cancer cell invasion. Calcium concentration and filopodia stability. Next we sought to further study the relationship between calcium and filopodia using two distinct approaches: a GFP-based calcium probe (GCaMP6s) and a dye-based calcium indicator (Fluo4-AM). Livecell TIRF imaging of cells transiently co-expressing the GFP-based calcium probe and MYO10-mCherry revealed that stable filopodia (41min MYO10 spot lifetime) display higher levels of calcium than unstable filopodia (o1min MYO10 spot lifetime) (Fig. 4a; Supplementary Movies 3) while no difference in the fluorescence signal was observed between stable and unstable filopodia in cells expressing GFP as a control (Fig. 4b). Correspondingly, Fluo4-AM demonstrated higher intensity in stable filopodia (Supplementary Fig. 6C). These data suggest that increased calcium levels at filopodia tips correlates with filopodia stability. In line with this notion, CCB-treated cells rapidly lost most of their calciumpositive stable filopodia and displayed a higher proportion of unstable filopodia compared with DMSO-treated cells (Fig. 4c,d; Supplementary Movies 4–6). As a second read-out of filopodia stability, the average velocity of MYO10 spots was measured and was found to be much higher in CCB-treated cells indicative of faster filopodia turnover following L-type calcium channel inhibition (Fig. 4e). CCB treatment also decreased the number and the stability of endogenous filopodia in both breast and pancreatic cancer cell lines (see method for details; Fig. 4f,g; Supplementary Fig. 6D). Taken together, calcium entry at filopodia via L-type calcium channels contributes to filopodia stability as well as filopodia formation. Active integrins, filopodia and L-type calcium channels. Integrins are one of the established cargos of MYO10 that are transported to filopodia tips to mediate cell–ECM adhesion26.In addition to the filopodia tip localization we observed active b1 integrins in the shafts of MYO10-positive filopodia (Fig. 5a). Moreover, and congruent with a study describing Rap1-RIAM localization to filopodia27, the integrin activator talin-1 was found to localize with MYO10 at filopodia tips (Fig. 5b). In addition, the Rap1/talin axis was important for filopodia formation as inhibition of Rap1 (Fig. 5c; Supplementary Fig. 7A) or silencing of talin-1 expression (Fig. 5d; Supplementary Fig. 7B,C) significantly reduced filopodia number. Conversely, overexpression of a constitutively active mutant of Rap1 (CA-Rap1, Fig. 5e; Supplementary Fig. S7D) or of the FERM domain of talin-1 (talin head, Fig. 5f; Supplementary Fig. 7E), known to promote integrin activity28, significantly increased filopodia formation. Given that Rap1 can be activated by increases in calcium levels29, we tested whether L-type calcium channels regulate filopodia formation through Rap1. Interestingly, the expression of CA-Rap1 or talin head did not restore filopodia numbers following CCB treatment (Fig. 5e,f), suggesting that L-type calcium channels could act downstream of the Rap1-RIAM-talin pathway. In addition, CCB treatment did not inhibit overall integrin activity in cells suggesting that L-type calcium channels do not promote filopodia by inducing integrin activation (Supplementary Fig. 7F). To further study the relationship between integrin activity, filopodia number and calcium concentration at filopodia tips, cells expressing a calcium probe and MYO10-mCherry were plated on conformation-specific anti-b1 integrin antibodies which lock b1 integrin in either an active or inactive conformation30. Using this system, b1 integrin activation significantly increased filopodia number as well as calcium levels at filopodia tips (Fig. 5g; Supplementary Fig. 8). These integrin-mediated effects were fully inhibited by CCB treatment (Fig. 5g; Supplementary Fig. 8), further demonstrating that integrin activation acts upstream of L-type calcium channels to regulate filopodia formation (Fig. 5h). Integrins promote filopodia formation and stability via Src.As the activation of FAK and Src are downstream events following b1 integrin-ECM engagement, we next assessed a potential role for these kinases in filopodia formation. The Src inhibitor Figure 1 | L-type calcium channel blockers (CCBs) inhibit filopodia formation and L-type calcium channels are functional in cancer cells. (a) MDA-MB231 cells transiently expressing MYO10-GFP and adhering to fibronectin (FN) were treated with various compounds (10 mM) for 1 h, fixed, stained for actin and imaged on a TIRF microscope (scale bar, 20 mm). The number of MYO10-positive filopodia was counted for each cell and displayed as a box plot (three biological repeats, n4100 cells, ***Pvalueo8.3 1017). (b) Relative expression of the four genes encoding the L-type calcium channel a1 subunit across 676 commonly used cancer cell lines19. Gene expression read counts are displayed. The value 7.99 corresponds to non-detected (ND). Selected cell lines are annotated. (c) MDA-MB-231 cells transiently expressing the calcium probe (pGP-CMV-GCaMP6s) and adhering to FN were treated with an L-type calcium channel activator (BAY K8644; 1 mM) while being imaged on a TIRF microscope (63 objective). The relative increase in the intracellular intensity of the calcium probe was measured at 1 and 10 min post stimulation. Cell boundaries are indicated by dotted lines (three biological repeats, n¼74 cells; scale bar, 20 mm). (d) MDA-MB-231 cells transiently expressing the calcium probe (pGP-CMV-GCaMP6s) and MYO10-mCherry were seeded on FN and treated with an L-type calcium channel activator (BAY K8644; 1 mM) while being imaged on a TIRF microscope (100 objective; scale bar, 10 mm). The inset shows a representative MYO10-positive filopodia tip delineated by a dotted line. ROI: region of interest. (e) MDA-MB-231 cells transiently expressing the calcium probe (pGP-CMV-GCaMP6s) and adhering to FN were treated with an L-type calcium channel activator (BAY K8644; 1 mM) in combination with DMSO or amlodipine besylate (1 mM). The relative increase in the intracellular intensity of the calcium probe was measured at 1 min (three biological repeats, n421 cells; ***Pvalueo1.39 10 4). Pvalues were calculated using Student’s t-test (unpaired, two-tailed, unequal variance). ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 4NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications dasatinib decreased filopodia number in the original microscopybased drug screen (60% fewer filopodia than DMSO at 20 mM). This result was validated using a small molecule inhibitor of Src (PP2) which triggered a significant loss in filopodia compared with controls (PP3 and DMSO). In contrast, inhibition of FAK had no effect on filopodia formation (Fig. 6a; a LY294002 DMSO Felodipine Amlodipine Su.86.86 Invasion (>45 μm) 0.8 1.2 0 0.4 *** *** *** c 0 μm 150 μm DMSO Cilnidipine Felodipine Manidipine Amlodipine Zonisamide Bumetanide 45 μm Untreated DMSO Cilnidipine Felodipine Amlodipine Zonisamide Carbamazepine Bumetanide LY294002 Invasion (>45 μm) 0 0.1 0.2 0.3 *** *** *** *** *** Manidipine dichloride MDA-MB-231 NS NS NS LY294002 DMSO Cilnidipine Felodipine Amlodipine Zonisamide Bumetanide Manidipine dichloride DMSO Bumetanide PDAC P53–/– PDAC P53R172H Invasion (>45 μm) 0.8 1.2 0 0.4 ND *** *** *** *** *** *** b d DMSO Amlodipine DMSO Amlodipine U2OS GFP U2OS MYO10-GFP 0 0.4 0.2 0.6 0.8 Invasion (>45 μm) *** PDAC P53–/–, DMSO Cilnidipine Felodipine Manidipine Amlodipine Zonisamide Bumetanide 0 μm 150 μm45 μm DMSO PDAC P53R172H Velocity (μm min–1) Untreated DMSO Cilnidipine Felodipine Amlodipine Carbamazepine Bumetanide LY294002 0 4 8 12 Manidipine dichloride ********* Directionality 0 0.2 0.4 0.6 0.8 Untreated DMSO Cilnidipine Felodipine Amlodipine Carbamazepine Bumetanide LY294002 Manidipine dichloride *** *** *** *** *** e L-type calcium channel blockers DMSO T-type calcium channel blocker Sodium channel inhibitors PI3K inhibitor ND *** Figure 2 | L-type calcium channels regulate filopodia formation and cancer cell invasion. (a) MDA-MB-231 cells were seeded into an inverted invasion assay in the presence of various compounds (10 mM) where indicated for 48 h. The relative invasion over 45 mm was quantified (n¼three biological repeats, ***Pvalueo1.3 10 5). (b) P53/and P53R172H PDAC cells were seeded into an inverted invasion assay in the presence of various compounds (10 mM) for 4 days. The relative invasion over 45 mm was quantified (n¼three biological repeats, ***Pvalueo4.1 108). (c) Su.86.86 pancreatic carcinoma cells were seeded into an inverted invasion assay and allowed to invade for 4 days in the presence of various compounds (10 mM). Relative invasion over 45 mm was quantified (n¼three biological repeats, ***Pvalueo9103). (d) U2OS cells stably expressing either GFP or MYO10-GFP were seeded into an inverted invasion assay and allowed to invade for 4 days in the presence of amlodipine besylate (10 mM) or DMSO. Relative invasion over 45 mm was quantified (n¼three biological repeats, ***Pvalueo4.05 106). (e) MDA-MB-231 cells were seeded on fibroblast-generated cell derived matrices (representative image is shown) in the presence of various compounds (10 mM), and cell migration was recorded over 24 h. Over 65 cells were manually tracked for each condition and migration speed and directionality were measured (n¼two biological repeats, scale bar, 200 mm; ***Pvalueo2.6 105). Pvalues were calculated using Student’s t-test (unpaired, two-tailed, unequal variance). All error bars represent s.e.m. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 ARTICLE NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications 5 Supplementary Fig. 9A). Similar to CCB treatment, PP2-mediated inhibition of Src also promoted filopodia instability (Supplementary Fig. 9B). Importantly, overexpression of a constitutively-active mutant of Src (CA-Src) increased filopodia number while overexpression of a dominant-negative mutant of Src (DN-Src) inhibited filopodia formation (Fig. 6b; Supplementary Fig. 9C). Immunofluorescence analyses further revealed active Src (pSrcY416) localization to filopodia (Fig. 6c) suggesting a central role for Src in filopodia induction. These Srcdependent effects appeared to be downstream of integrin signalling as treatment with a Src inhibitor (Fig. 6d) or overexpression of DN-Src (Fig. 6e) significantly reduced filopodia number and calcium levels at filopodia tips in cells plated on the anti-active b1 integrin antibody (Fig. 6d,e). Correspondingly, overexpression of CA-Src was sufficient to bypass the requirement for integrin activation and promoted filopodia formation and calcium increase at filopodia tips in cells plated on the anti-inactive b1 integrin antibody (Fig. 6e). Taken together, these data indicate that Src activity plays a key role in filopodia formation and stability downstream of integrin signalling. Importantly, as CA-Src-induced filopodia remained sensitive to CCB treatment (Fig. 6b) and as CCBs did not affect overall Src activity (Supplementary Fig. 9D), these results indicate that Src acts upstream of L-type calcium channels (Fig. 6f). However, how Src promotes L-type calcium channel activation remains to be determined. CACNA1D MYO10-mCherry CACNA1D MYO10-mCherry d cba 0 20 40 60 Number of MYO10-positive filopodia per cell siCTRL siCACNA1C siCTRL siCACNA1D #1 *** siCACNA1S #1 *** siCACNA1D #2 *** siCACNA1S #2 *** Invasion (>45 μm) 0.4 0 0.2 0.6 siCTRL siCACNA1C siCTRL siCACNA1D #1 siCACNA1S #1 siCACNA1D #2 siCACNA1S #2 *** *** *** *** CACNA1D CACNA1S mRNA expression relative to 106 copies of GAPDH CACNA1C 0 0.05 0.10 0.15 0.20 0.25 ND MDA-MB-231 CACNA1D Luminal #1 Luminal #2 Her2-positive eTriple-negative Figure 3 | CACNA1D and CACNA1S regulate filopodia formation and cancer cell invasion. (a) Relative expression of CACNA1C,CACNA1D and CACNA1S in MDA-MB-231 cells as determined by Q-RT-PCR (n¼three biological repeats). (b) MDA-MB-231 cells previously silenced for CACNA1D or CACNA1S using multiple RNAi oligos as indicated and transiently expressing MYO10-GFP were plated on FN for 2 h, fixed and the number of MYO10-positive filopodia per cell was quantified (n495 cells, three biological repeats ***Pvalueo2.4 1010). A smartpool of oligos targeting CACNA1C was used as an additional control as no CACNA1C expression was detected in these cells. (c) MDA-MB-231 cells previously silenced for CACNA1D or CACNA1S using multiple oligos were seeded into an inverted invasion assay and allowed to invade for 48 h. Relative invasion over 45 mm was quantified (n¼three biological repeats, ***Pvalueo8.9 10 3). (d) MDA-MB-231 cells transiently expressing MYO10-mCherry were plated on FN for 2 h, stained for endogenous CACNA1D and imaged on a TIRF microscope (scale bar, 20 mm). (e) Representative images of various types of breast cancer tissue samples stained for CACNA1D (scale bar, 200 mm). Pvalues were calculated using Student’s t-test (unpaired, two-tailed, unequal variance). All error bars represent s.e.m. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 6NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications Increased MYO10 and Src protein expression and/or activity have been reported independently in breast cancer and are linked to metastasis and poor patient survival13,15,31. Given that Src is critical for the formation of MYO10-positive filopodia, we investigated a possible association between MYO10 and Src protein levels and/or activity in patient samples. Using cBioPortal, we determined that increased MYO10 levels strongly correlates with increased Src mRNA levels in the Breast Invasive Carcinoma data set (Log Odds ratio: 0.826; Pvalue, o0.001). Moreover, phosphoproteomic analyses indicated that patient Average number of endogenous filopodia per frame per cell 0 10 20 30 DMSO Felodipine A mlodipine DMSO Felodipine A mlodipine Average lifetime of endogenous filopodia (s) 30 20 10 0 40 50 dMYO10 spot velocity DMSO Felopidine Amlodipine Average MYO10 spot velocity (μm s –1 ) * 0 0.04 0.08 0.12 *** e 0 min 1 min 2 min 0 min 1 min 2 min MYO10-mCherryCalcium probe DMSO Felodipine t 0 t 2 min t 4 min t 0 t 2 min t 4 min F Amlodipine t 0 t 2 min t 4 min a Calcium probeMYO10-mCherry Calcium probe intensity at filopodia tips 0 0.2 0.4 0.6 0.8 *** <1 min >1 min Myo10 spot lifetime c DMSOFelopidineAmlodipineDMSOFelopidineAmlodipine f <1 min MYO10 spots (% per cell) MYO10 spot lifetime 0 20 40 60 80 Felodipine DMSO Amlodipine *** *** *** *** Calcium probe Myo10 spot lifetime 0 min 1 min 2 min *** *** ** GFP intensity at filopodia tips NS <1 min >1 min 0 0.2 0.4 0.6 0.8 GFP b g 1–2 min 2–3 min 3–4 min >4 min NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 ARTICLE NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications 7 samples with high MYO10 expression (130 patients out of 1,080; 12%) also exhibited higher Src activity (mean of alteration in altered MYO10,0.05; mean of alteration in unaltered group, 0.23; Pvalue, 1.408e 3; qvalue, 0.0162). Furthermore, re-analysis of a tissue microarray comprising 1,212 samples from a Finnish nationwide population-based breast cancer series (FinProg cohort), previously stained for MYO10 (ref. 13) and pSrcY416 (refs 13,32), revealed that MYO10 protein levels correlate with Src activity in patient samples (Fig. 6g; Table 1). Therefore, MYO10 expression correlates with both Src expression and Src activity in breast cancer patient samples. Calpain 1 and filopodia formation. Next we sought to identify the calcium-regulated pathway(s) that is/are involved in filopodia formation. Cells expressing MYO10-GFP were treated with inhibitors targeting multiple calcium-regulated proteins and, among these inhibitors, only the compound targeting calpain activity significantly diminished filopodia formation (Fig. 7a; Supplementary Fig. 10A). More specifically, silencing of calpain-1, and not calpain-2, expression decreased filopodia number compared with control (Fig. 7b; Supplementary Fig. 10B–E). In addition, calpain-1 and calpain-2 appeared to have cooperative functions during the invasive process as silencing of both calpain-1 and calpain-2 inhibited cell invasion more efficiently than the individual depletion of either isoform (Fig. 7c). Immunofluorescence analyses revealed that calpain-1 localizes to filopodia (Fig. 7d) and ratiometric FRET analyses of cells co-expressing a calpain sensor33 and MYO10-mCherry demonstrated that calpains are active within filopodia (Fig. 7e; controls related to the use of the calpain FRET probe are shown in Supplementary Fig. 11). To assess whether calpains regulate filopodia formation downstream of integrin activation, cells expressing the calcium probe and MYO10-mCherry were plated on anti-b1 integrin antibodies and treated with a calpain inhibitor. Blocking calpain activity decreased filopodia number but not calcium levels at the remaining filopodia tips in cells plated on the anti-active b1 integrin antibody (Fig. 7f). Altogether, these data indicate that calpain-1 regulates filopodia formation downstream of integrin activation and calcium accumulation at filopodia tips. Filopodia stabilization leads to focal adhesion formation. The importance of filopodia in the process of tumorigenesis has been described in several cancers. However, the significance of a mechanism that could regulate filopodia stability during cell migration has never been explored. CCB treatment only weakly impacted focal adhesion dynamics as their overall lifetime distribution was unaffected (Fig. 8a). CCB treatment appeared to decrease both the assembly and disassembly rate of focal adhesions as well as their maximal size (Fig. 8a). Using live-cell TIRF imaging we found that talin-1-GFP localizes to both focal adhesions and MYO10-positive filopodia and in many instances was observed to move together with MYO10 spots along filopodia shafts (Fig. 8b; Supplementary Movies 7). In stabilized filopodia, both MYO10 and talin-1 were found to accumulate at filopodia tips (Fig. 8b; Supplementary Movies 7). Subsequent advancement of the plasma membrane coincided with MYO10 leaving the growing talin-1-positive structure and the formation of new MYO10-positive filopodia followed by accumulation of a talin-positive patch resembling a classical focal adhesion. These observations are in agreement with previous studies demonstrating a role for filopodia in directing lamellipodia formation in fibroblasts34 and maturation of filopodial shaft adhesions into focal adhesions upon lamellipodia advancement35. Taken together, these data allow us to propose a model by which filopodia formation and stabilization via an integrin/Src/L-type calcium channel/calpain pathway, described here, contributes to directional cell motility and cancer cell invasion (Fig. 8c). Discussion Here, we define the foundation of a druggable and clinically relevant signalling pathway (Fig. 8c) that regulates filopodia formation and stability in cancer cells. Notably, L-type calcium channels are expressed in human cancer and targeting their function with FDA-approved CCBs impairs filopodia formation and blocks cancer cell invasion. Mechanistically, we establish a link between a localized increase in calcium concentration at filopodia tips, mediated by L-type calcium channels and filopodia stability. Moreover, integrin inside-out activation and ligand binding is indispensable for filopodia formation and appears to be the first critical step in this process following MYO10-dependent delivery of integrins to filopodia tips. Subsequent steps require integrin outside-in signalling to trigger Src activation and the spatially restricted calcium entry at filopodia tips and the activation of the calcium-regulated protease calpain-1. Finally, we demonstrate that filopodia stabilization appears to precede focal adhesion maturation and propose that cycles of filopodia stabilization and focal adhesion maturation direct cell migration and invasion. The central position of L-type calcium channels in cancer cell invasion was very unexpected as these voltage-gated channels (expression and activity) are considered to be restricted to excitable cells (neuronal and muscle cells). Nevertheless, several studies report roles for L-type calcium channels in many other cell types including fibroblasts, kidney cells and endometrial and prostate cancer cells36–38. In addition, analyses of available gene expression data sets by us (this study) and others25 revealed a wide expression of L-type calcium channels in many cancer cell Figure 4 | Calcium at filopodia tips regulates filopodia stability. (a) MDA-MB-231 cells transiently expressing the calcium probe (pGP-CMV-GCaMP6s) and MYO10-mCherry were plated on FN and imaged live using a TIRF microscope (1 picture every 5 s; scale bar, 20 mm). The intensity of the calcium probe at MYO10-positive filopodia tips was measured and compared between transient (o1 min lifetime) and stable filopodia (41 min lifetime) (n¼276 filopodia, three biological repeats, ***Pvalueo1.8 109). (b) MDA-MB-231 cells transiently expressing GFP and MYO10-mCherry were plated on FN and imaged live using a TIRF microscope. The intensity of GFP at MYO10-positive filopodia tips was measured and compared between transient and stable filopodia (n¼293 filopodia, three biological repeats). (c–e) MDA-MB-231 cells transiently expressing the calcium probe (pGP-CMV-GCaMP6s) and MYO10-mCherry were plated on FN, treated with DMSO, felodipine or amlodipine besylate (10 mM) and imaged live using a TIRF microscope (1 picture every 5 s; scale bar, 20 mm). Representative images are shown (c). For each condition, MYO10-positive particles were automatically tracked and MYO10 spot lifetime (calculated as a percentage of the total number of filopodia generated per cell) (d) and average MYO10 spot velocity (e) were plotted (see method for details; three biological repeats, n45,600 particles tracked in more than 16 cells, *Pvalue ¼0.015, ***Pvalueo2.98 105). (f,g) MDA-MB-231 cells transiently expressing lifeact-GFP were plated on FN, treated with DMSO, felodipine or amlodipine besylate (10 mM), and imaged live on a TIRF microscope. Movies were segmented and endogenous filopodia automatically identified using CellGeo51 (See method for details). Average filopodia number per frame and per cell (f) and average filopodia lifetimes are displayed (g)(n416 cells, two biological repeats; *Pvalue ¼0.035, ***Pvalueo5.6 105). Pvalues were calculated using Student’s t-test (unpaired, two-tailed, unequal variance). All error bars represent sem. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 8NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications lines and clinical samples. In particular, CACNA1D, which localizes to filopodia tips and contributes to cancer cell invasion in vitro (shown in this study), was found to be expressed in patient samples in all major breast cancer subtypes. The pathway delineated here as fundamental for filopodia formation in cancer cells shares some similarities with calcium-regulated processes identified in normal excitable cells. In rat smooth muscle cells a5b1 integrin signalling has been g Felodipine A mlodipine Anti-active β1 integrin mAb Anti-inactive β1 integrin mAb Number of MYO10-positive fi lopodia per cell DMSO DMSO Felodipine A mlodipine 0 100 200 Calcium probe intensity at filopodia tips DMSO DMSO 0 1 2 ec Rap1 inhibitor DMSO Number of MYO10-positive filopodia per cell d siCTRL siTalin-1 #1 siTalin-1 #2 f Felodipine Amlodipine DMSO Felodipine Amlodipine DMSO GFP 0 20 40 60 10 30 50 Felodipine Amlodipine DMSO Felodipine Amlodipine DMSO 0 100 20 40 60 80 GFP CA-Rap1 b aActive β1 integrin Talin-1 MYO10-mCherry Active β1 integrin MYO10-mCherry Talin-1 AmlodipineFelodipineDMSODMSO DMSO DMSO AmlodipineFelodipine Anti-inactive β1 integrin mAb Anti-active β1 integrin mAb Calcium probe Calcium probe MYO10-mCherry Number of MYO10-positive filopodia per cell Number of MYO10-positive filopodia per cell Number of MYO10-positive filopodia per cell Inactive integrins Active integrins Talin Rap1 Ca 2+ L-type Ca2+ channels ? ? Filopodia stability h MYO10-mCherry MYO10-mCherry 0 100 20 40 60 80 *** 0 20 40 10 30 50 *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** Talin head *** NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 ARTICLE NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications 9 Hs00913730_m1; CACNA1S: Hs00163885_m1) and were analysed with the 7900HT fast RT-PCR System (Applied Biosystems). Data were studied using RQ Manager Software (Applied Biosystems). Clinical sample analyses.Analyses of publicly available data sets were performed using IST online (MediSapiens Ltd; http://ist.medisapiens.com/) or cBioPortal (http://www.cbioportal.org/index.do). Interrogation of the Breast Invasive Carcinoma (TCGA, Cell 2015)24 which contains 1,105 samples was performed using cBioPortal. The genomic profiles selected were mutations, putative copy-number alterations from GISTIC and mRNA expression data (mRNA Expression z-Scores (RNA Seq V2 RSEM)). Survival analyses were performed using the Survival tab and the co-expression analyses were performed using the Co-expression or the Enrichment tabs. Potential correlation between MYO10 and pSrcY416 levels in patient samples was assessed in the FinProg series, a Finnish nationwide breast cancer cohort. Patient and sample inclusion criteria for FinProg are described in detail elsewhere56. The FinProg series (1,212 samples) was previously stained for pSrcY416 and MYO10 expression as described earlier13,32. From these data a potential association between the expression groups was assessed using the w2-test. Permission to use patient samples for research purposes was provided by the Ministry of Social Affairs and Health, Finland (permission 123/08/97). CACNA1D staining of patients samples.The use of breast cancer tissue samples was performed with approval from the ethical committee of Turku University Hospital (149/6/2002) and Auria Biobank (AB15-9859) and accordingly informed consent was obtained from all human participants. Breast tissue material was prepared according to standard histology practice, that is, fixed in buffered formalin (pH 7.0) and embedded into paraffin blocks. TMAs were prepared by collecting tissue cores (4 mm in diameter) from the representative tumour area of each patient. Tissue sections (3 mm thick) were cut and stained with anti-CACNA1D antibody (1/600) using Lab Vision Autostainer 480 (Thermo-Fisher Scientific, Fremont, CA, USA) and detected with PowerVision þpolymer kit, according to standard protocols (DPVB þ110HRP; Immunovision Technologies, Vision Biosystems, Norwell, MA, USA), using diaminobenzidine as the chromogen. Statistical analysis.Statistical analyses were performed when appropriate, and Pvalues indicated by an asterisk in the figure legends. Unless otherwise indicated, the Student’s t-test was used (unpaired, two-tailed, unequal variance). Frequency tables were analysed by using two-tailed w2-test. Data availability.The authors declare that the data supporting the findings of this study are available within the article and from the authors on request. References 1. Waclaw, B. et al. A spatial model predicts that dispersal and cell turnover limit intratumour heterogeneity. Nature 525, 261–264 (2015). 2. Gundem, G. et al. The evolutionary history of lethal metastatic prostate cancer. Nature 520, 353–357 (2015). 3. Gupta, G. P. & Massague ´, J. Cancer Metastasis: Building a Framework. Cell 127, 679–695 (2006). 4. Legate, K. R., Wickstro ¨m, S. A. & Fa ¨ssler, R. Genetic and cell biological analysis of integrin outside-in signaling. Genes Dev. 23, 397–418 (2009). 5. Askari, J. A., Buckley, P. A., Mould, A. P. & Humphries, M. J. Linking integrin conformation to function. J. Cell Sci. 122, 165–170 (2009). 6. Jacquemet, G., Humphries, M. J. & Caswell, P. T. Role of adhesion receptor trafficking in 3D cell migration. Curr. Opin. Cell Biol. 25, 627–632 (2013). 7. Hynes, R. O. Integrins: bidirectional, allosteric signaling machines in their roles as major adhesion receptors, integrins. Cell 110, 673–687 (2002). 8. Jacquemet, G., Hamidi, H. & Ivaska, J. Filopodia in cell adhesion, 3D migration and cancer cell invasion. Curr. Opin. Cell Biol. 36, 23–31 (2015). 9. Jacquemet, G. et al. RCP-driven a5b1 recycling suppresses Rac and promotes RhoA activity via the RacGAP1-IQGAP1 complex. J. Cell Biol. 202, 917–935 (2013). 10. Paul, N. R. et al. a5b1 integrin recycling promotes Arp2/3-independent cancer cell invasion via the formin FHOD3. J. Cell Biol. 210, 1013–1031 (2015). 11. Shibue, T., Brooks, M. W., Inan, M. F., Reinhardt, F. & Weinberg, R. A. The outgrowth of micrometastases is enabled by the formation of filopodium-like protrusions. Cancer Discov. 2, 706–721 (2012). 12. Shibue, T., Brooks, M. W. & Weinberg, R. A. An integrin-linked machinery of cytoskeletal regulation that enables experimental tumor initiation and metastatic colonization. Cancer Cell 24, 481–498 (2013). 13. Arjonen, A. et al. Mutant p53-associated myosin-X upregulation promotes breast cancer invasion and metastasis. J. Clin. Invest. 124, 1069–1082 (2014). 14. Li, A. et al. Fascin is regulated by slug, promotes progression of pancreatic cancer in mice, and is associated with patient outcomes. Gastroenterology 146, 1386 (2014). 15. Cao, R. et al. Elevated expression of myosin X in tumours contributes to breast cancer aggressiveness and metastasis. Br. J. Cancer 111, 539–550 (2014). 16. Plantard, L. et al. PtdIns(3,4,5)P3 is a regulator of myosin-X localization and filopodia formation. J. Cell Sci. 123, 3525–3534 (2010). 17. Catterall, W. A. Voltage-gated calcium channels. Cold Spring Harb. Perspect. Biol. 3, a003947 (2011). 18. Lipscombe, D., Helton, T. D. & Xu, W. L-type calcium channels: the low down. J. Neurophysiol. 92, 2633–2641 (2004). 19. Tyson, J. R. & Snutch, T. P. Molecular nature of voltage-gated calcium channels: structure and species comparison. WIREs Membr. Transp. Signal 2, 181–206 (2013). 20. Klijn, C. et al. A comprehensive transcriptional portrait of human cancer cell lines. Nat. Biotechnol. 33, 306–312 (2015). 21. Chen, T.-W. et al. Ultrasensitive fluorescent proteins for imaging neuronal activity. Nature 499, 295–300 (2013). 22. Cerami, E. et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2, 401–404 (2012). 23. Gao, J. et al. Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci. Signal. 6, pl1 (2013). 24. Ciriello, G. et al. Comprehensive molecular portraits of invasive lobular breast cancer. Cell 163, 506–519 (2015). 25. Wang, C.-Y., Lai, M.-D., Phan, N. N., Sun, Z. & Lin, Y.-C. Meta-analysis of public microarray datasets reveals voltage-gated calcium gene signatures in clinical cancer patients. PLoS ONE 10, e0125766 (2015). 26. Zhang, H. et al. Myosin-X provides a motor-based link between integrins and the cytoskeleton. Nat. Cell Biol. 6, 523–531 (2004). 27. Lagarrigue, F. et al. A RIAM/lamellipodin-talin-integrin complex forms the tip of sticky fingers that guide cell migration. Nat. Commun. 6, 8492 (2015). 28. Calderwood, D. A. et al. The Talin head domain binds to integrin beta subunit cytoplasmic tails and regulates integrin activation. J. Biol. Chem. 274, 28071–28074 (1999). 29. Bos, J. L. Linking Rap to cell adhesion. Curr. Opin. Cell Biol. 17, 123–128 (2005). 30. Byron, A. et al. Anti-integrin monoclonal antibodies. J. Cell Sci. 122, 4009–4011 (2009). 31. Elsberger, B. Translational evidence on the role of Src kinase and activated Src kinase in invasive breast cancer. Crit. Rev. Oncol. Hematol. 89, 343–351 (2014). 32. Vassilev, B. et al. Elevated levels of StAR-related lipid transfer protein 3 alter cholesterol balance and adhesiveness of breast cancer cells: potential mechanisms contributing to progression of HER2-positive breast cancers. Am. J. Pathol. 185, 987–1000 (2015). 33. Stockholm, D. et al. Imaging calpain protease activity by multiphoton FRET in living mice. J. Mol. Biol. 346, 215–222 (2005). 34. Johnson, H. E. et al. F-actin bundles direct the initiation and orientation of lamellipodia through adhesion-based signaling. J. Cell Biol. 208, 443–455 (2015). 35. Hu, W., Wehrle-Haller, B. & Vogel, V. Maturation of filopodia shaft adhesions is upregulated by local cycles of lamellipodia advancements and retractions. PLoS ONE 9, e107097 (2014). 36. Hao, J. et al. Ca2 þchannel subunit a1D promotes proliferation and migration of endometrial cancer cells mediated by 17b-estradiol via the G protein-coupled estrogen receptor. FASEB J. 29, 2883–2893 (2015). 37. Chen, R. et al. Cav1.3 channel a1D protein is overexpressed and modulates androgen receptor transactivation in prostate cancers. Urol. Oncol. 32, 524–536 (2014). 38. Yang, S. & Huang, X.-Y. Ca2 þinflux through L-type Ca2 þchannels controls the trailing tail contraction in growth factor-induced fibroblast cell migration. J. Biol. Chem. 280, 27130–27137 (2005). 39. Wu, X. et al. Modulation of calcium current in arteriolar smooth muscle by avb3 and a5b1 integrin ligands. J. Cell Biol. 143, 241–252 (1998). 40. Wu, X., Davis, G. E., Meininger, G. A., Wilson, E. & Davis, M. J. Regulation of the L-type calcium channel by alpha 5beta 1 integrin requires signaling between focal adhesion proteins. J. Biol. Chem. 276, 30285–30292 (2001). 41. Gui, P. et al. Integrin receptor activation triggers converging regulation of Cav1.2 calcium channels by c-Src and protein kinase A pathways. J. Biol. Chem. 281, 14015–14025 (2006). 42. Robles, E., Huttenlocher, A. & Gomez, T. M. Filopodial calcium transients regulate growth cone motility and guidance through local activation of calpain. Neuron 38, 597–609 (2003). 43. Gomez, T. M., Robles, E., Poo, M. & Spitzer, N. C. Filopodial calcium transients promote substrate-dependent growth cone turning. Science 291, 1983–1987 (2001). 44. Lautermilch, N. J. & Spitzer, N. C. Regulation of calcineurin by growth cone calcium waves controls neurite extension. J. Neurosci. 20, 315–325 (2000). 45. Wen, Z., Guirland, C., Ming, G.-L. & Zheng, J. Q. A CaMKII/calcineurin switch controls the direction of Ca(2 þ)-dependent growth cone guidance. Neuron 43, 835–846 (2004). ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 16 NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications 46. Yang, M. & Brackenbury, W. J. Membrane potential and cancer progression. Front. Physiol. 4, 185 (2013). 47. Fraser, S. P. et al. Voltage-gated sodium channel expression and potentiation of human breast cancer metastasis. Clin. Cancer Res. 11, 5381–5389 (2005). 48. Hein, M. Y. et al. A human interactome in three quantitative dimensions organized by stoichiometries and abundances. Cell 163, 712–723 (2015). 49. Paul, N. R., Jacquemet, G. & Caswell, P. T. Endocytic trafficking of integrins in cell migration. Curr. Biol. 25, R1092–R1105 (2015). 50. Byron, A. et al. A proteomic approach reveals integrin activation statedependent control of microtubule cortical targeting. Nat. Commun. 6, 6135 (2015). 51. Sbalzarini, I. F. & Koumoutsakos, P. Feature point tracking and trajectory analysis for video imaging in cell biology. J. Struct. Biol. 151, 182–195 (2005). 52. Tsygankov, D. et al. CellGeo: a computational platform for the analysis of shape changes in cells with complex geometries. J. Cell Biol. 204, 443–460 (2014). 53. Berginski, M. E., Vitriol, E. A., Hahn, K. M. & Gomez, S. M. High-resolution quantification of focal adhesion spatiotemporal dynamics in living cells. PLoS ONE 6, e22025 (2011). 54. Jacquemet, G. et al. Rac1 is deactivated at integrin activation sites through an IQGAP1-filamin-A-RacGAP1 pathway. J. Cell Sci. 126, 4121–4135 (2013). 55. Hennigan, R. F., Hawker, K. L. & Ozanne, B. W. Fos-transformation activates genes associated with invasion. Oncogene 9, 3591–3600 (1994). 56. Joensuu, H. et al. Amplification of erbB2 and erbB2 expression are superior to estrogen receptor status as risk factors for distant recurrence in pT1N0M0 breast cancer: a nationwide population-based study. Clin. Cancer Res. 9, 923–930 (2003). Acknowledgements We thank J. Siivonen and P. Laasola for technical assistance, M. Saari for help with the microscopes and H. Hamidi for scientific writing and editing of the manuscript. This study has been supported by the Academy of Finland, ERC Starting Grant, ERC Consolidator Grant, the Sigrid Juselius Foundation, and the Finnish Cancer Organization. G.J. and M.G. are supported by an EMBO Long-Term Fellowship. E.P. is supported by an Academy of Finland postdoctoral fellowship. Author contributions G.J designed, carried out and analysed the majority of the experiments with help from J.I, M.G, E.P and P.C. G.J wrote the manuscript with help from J.I. H.B performed the original drug screen with crucial help from T.H and M.P. H.S, H.J and P.K provided the clinical samples and the tissue sample analyses. J.I supervised the study. Additional information Supplementary Information accompanies this paper at http://www.nature.com/ naturecommunications Competing financial interests: The authors declare no competing financial interests. Reprints and permission information is available online at http://npg.nature.com/ reprintsandpermissions/ How to cite this article: Jacquemet, G. et al. L-type calcium channels regulate filopodia stability and cancer cell invasion downstream of integrin signalling. Nat. Commun. 7, 13297 doi: 10.1038/ncomms13297 (2016). Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/ rThe Author(s) 2016 NATURE COMMUNICATIONS | DOI: 10.1038/ncomms13297 ARTICLE NATURE COMMUNICATIONS | 7:13297 | DOI: 10.1038/ncomms13297 | www.nature.com/naturecommunications 17