Stiff matrix promotes lung cancer cell migration through down-regulating the Piezo1 channel expression to facilitate Ca-dependent filopodia formation.

1
Introduction
Lung cancer is the leading cause of cancer-related death worldwide, among which non-small cell lung cancers (NSCLC) account for about 85 % [1]. Despite the great advances in medical treatments, the 5-year survival rate of lung cancer remains less than 10 % [2]. Similar to other solid tumors, the biophysical and biochemical cues in the tumor microenvironments in concerted actions influence lung cancer cell functions, including proliferation, migration, invasion as well as epithelial to mesenchymal transition (EMT), and thereby drive cancer progression [[3], [4], [5]]. Therefore, clarifying the molecular mechanism, by which the microenvironmental cues regulate lung cancer cell functions, would assist optimization and development of therapeutic strategies.
Extracellular matrix (ECM) is an important part of the tumor microenvironments, a complex network assembled by a variety of matrix components. Solid tumors are in common characterized by stiffened ECM, with the elastic modulus of cancerous lung (15–100 kPa) being much higher than that of normal lung parenchyma (0.5–5 kPa) [6,7]. It is well established that ECM stiffness strongly influences cancer cell functions, cancer progression and even drug resistance. In general, cells including cancer cells sense, transmit and respond to the changes in matrix stiffness using multiple mechanosensitive machineries, such as integrins, ion channels and cytoskeletal proteins [8]. Piezo1, as an intrinsically mechanosensitive ion channel, is critically engaged in these processes [[9], [10], [11]]. In most cancers, the expression of Piezo1 channel is up-regulated and positively related to cancer cell migration [12]. For example, its knockout inhibited gastric cancer cell invasion and xenograft tumor growth in BALB/c nude mice [13]. Intriguingly, the Piezo1 expression in cancerous lung tissue is lower than that in normal lung tissue, and its knockdown promoted cell migration in both NSCLC and small cell lung cancers (SCLC) models [[14], [15], [16]]. Given the importance of ECM stiffness, interesting questions arise, namely, does the Piezo1 channel respond differently in lung cancer migration, does it still play an inhibitory regulatory role, and what molecular mechanism is responsible for such a unique role?
Usually, cells on stiff substrates reorganize their actin-cytoskeleton for spreading or moving [17]. In some cases, actin-cytoskeletons are assembled and polymerized into special plasma membrane protrusion, named lamellipodia or filopodia [18]. Filopodia are thin finger-like, actin-rich protrusions and play a crucial role in probing ECM stiffness and directing cancer cell migration [19]. There is increasing evidence that filopodia formation and increases in its length and number facilitate cancer cell migration, invasion and metastasis [20,21]. Specifically, it has been demonstrated that substrate stiffness regulates filopodial activities in lung cancer cells [22]. Filopodia formation is tightly controlled by dynamic actin polymerization and depolymerization, which is tightly regulated by various actin-binding proteins (ABPs) [23]. Cofilin, an actin depolymerizing factor, is a small ABP (19–21 kDa) that severs actin polymerization, regulates filopodia formation and participates in cancer cell migration and invasion [[24], [25], [26], [27]]. Cofilin is inactivated by phosphorylation at Ser-3 by kinases such as RhoA/ROCK (Rho-associated kinase) and LIMK (Lin11, Isl-1 and Mec-3), and reactivated by dephosphorylation by phosphatases such as Slingshot [24]. It is well known that the Piezo1 channel functions mainly through mediating extracellular Ca2+ influx to raise the [Ca2+]i [9,[28], [29], [30]]. Of more interest, it has been reported that a rise in the [Ca2+]i can promote cofilin dephosphorylation/activation through calcineurin (CaN)/Slingshot (SSH) signaling pathway [31]. In addition, it has been shown that mechanical stress-induced activation of the Piezo1 channel in chondrocytes tightly associates with actin polymerization through the RhoA/LIMK/cofilin pathway [32]. Therefore, it is interesting to examine whether the Piezo1 channel is engaged in regulating lung cancer cell migration through the Ca2+-dependent cofilin/actin polymerization/filopodia formation pathway.
In this study, we investigated the role of Piezo1 channel in regulating the [Ca2+]i, cofilin phosphorylation and filopodial formation and thereby mediating matrix stiffness regulation of lung cancer cell migration in A549 cells, a NSCLC line, growing on soft and stiff polyacrylamide (PA) hydrogels with stiffnesses close to these of healthy and cancerous lung tissues, respectively, and aim to gain mechanistic understanding of how the Piezo1 channel is involved in the regulation of lung cancer cell migration by ECM.

2
Materials and methods
2.1
Bioinformatics data analysis
The public databases were analyzed to compare the Piezo1 expression between cancerous and normal tissues. The clinical data of cancer patients were collected from Gene Expression database of Normal and Tumor tissues 2 (GENT2) and GEO database (NCSLC and normal lung tissues, GEO: GSE18842, GSE19188, GSE63074; pancreatic tumor and normal pancreas tissues, GEO: GSE16515, GSE15471, GSE2361, GSE19650, GSE22780, GSE32688, GSE15932; breast cancer and normal breast tissues, GEO: GSE31448, GSE7904, GSE45827, GSE2109, GSE54002, GSE5764, GSE7307, GSE73613, GSE42568, GSE65216; brain cancer and normal brain tissues, GEO: GSE68848, GSE7307; liver cancer and normal liver tissues, GEO: GSE41804, GSE40873, GSE62232, GSE45436; ovarian cancer and normal ovary tissues, GEO: GSE14764, GSE23603, GSE26712, GSE38666, GSE7307, GSE14407, GSE27651, GSE28044, GSE52460, GSE69428, GSE33805).

2.2
Fabrication of polyacrylamide (PA) gels
PA gels were fabricated as previously described [33]. Briefly, mixed solutions were prepared containing 40 % acrylamide stock solution, 2 % bis-acrylamide stock solution, 10 % ammonium persulfate (APS), and tetramethylethylenediamine (TEMED, #HC001L, YTHX, China). The mechanical properties of the gels were modulated by altering the volume of 40 % acrylamide and 2 % bis-acrylamide in the pre-polymerization solution. Gels with distinct stiffness values were fabricated by adjusting the final concentrations of acrylamide (4 %, 10 % and 8 %) and bis-acrylamide (0.3 %, 0.1 % and 0.264 %) to achieve elastic moduli of 3.24 kPa, 10.61 kPa and 19.66 kPa (3 kPa, 10 kPa and 20 kPa in short), which correspond to the stiffness ranges of normal lung parenchyma (0.5–5 kPa), pulmonary fibrosis or early-stage tumors (∼10 kPa) and solid lung tumors (20–30 kPa), respectively [7,33,34]. Then, the PA hydrogels were incubated with 1 mg/mL sulfo-SANPAH (#C1111, ProteoChem, USA) and were activated by exposure to UV (365 nm). Before use, the PA gels were coated with 0.12 mg/mL rat tail collagen type I (#354236, CORNING, USA) overnight at room temperature.

2.3
Cell culture and treatment
Two human non-small lung cancer cell lines, A549 and H460, were used in the study and they were obtained from the American Type Culture Collection (ATCC) and cultured in 1640 medium (#11875119, Gibco, USA) containing 10 % fetal bovine serum (FBS, #11011–8611, Sijiqing, China), 100 U/ml penicillin and 100 μg/ml streptomycin (#15140122, Gibco, USA) at 37 °C in a 5 % CO2 incubator.
Cells were seeded on PA substrates for the following determinations. In some cases, cells were cultured in medium containing 2 μM Yoda1 (#M9372, AbMole, USA) to activate the Piezo1 channel, 2.5 μM GsMTx4 (#HY-P1410, MedChemExpress, USA) to block the Piezo1 channel, 25 μM BAPTA-AM (#M4973, AbMole, USA) to chelate intracellular Ca2+ or adding 2 mM CaCl2 to observe extracellular Ca2+ influx, 4 μM cyclosporine A (CsA) (#M1831, AbMole, USA) to inhibit CaN, 0.5 μM dasatinib (Das, #9052S, Cell signaling technology) to inhibit YAP nuclear translocation, as well as 10 μM PF-573228 (PF, #HY-10461, MCE) to inhibit focal adhesion kinase (FAK) activation, respectively.

2.4
F-actin staining and confocal microscopy
Cells were seeded into 6‐well plates with PA hydrogels-covered glass coverslips at a density of 2 × 104 cells/well, and cultured further for 48 h. Following fixation with 4 % paraformaldehyde and permeabilization with 0.1 % Triton-X, cells were incubated with rhodamine-labeled phalloidin (1:100 dilution, #C2207S, Beyotime, China) for 40 min. Hoechst 33342 (1:1000, #H3570, ThermoFisher, USA) was used to stain nuclei. Images were captured using a confocal laser scanning microscope (ZEISS LSM 880, Germany). The length and number of filopodia were calculated by Image J.

2.5
Transwell assay
Cells were seeded on PA substrates at a density of 1 × 105 cells/well in 6-well plates and cultured for 48 h, and cell migration was determined using 24-well transwell chambers (8.0-μm pores; Corning, New York, USA) as previously described [35]. Briefly, 1 × 104 cells in 200 μl serum-free medium were added into the upper transwell chamber, the lower chamber was filled with 500 μl 10 % FBS-containing medium. After incubation for 18 h, the inserts were fixed with 4 % paraformaldehyde and stained with 1 % Crystal Violet (#C616390, Aladdin, China). The cells on upper surfaces of the inserts were swabbed gently. The migrated cells were imaged using a microscope (Nikon Ti2) in 4 random view fields and quantified by counting the number of cells with Image J. The experiments were repeated three times. In each time, there were two replicated samples for each group. Cell migration was calculated by the migrated cell number in the test group normalized to that in control group (as defined in figures).

2.6
Western blotting
Cells were seeded on PA substrates as mentioned above. After 48 h in culture, cells were collected and lysed in RIPA buffer (#P0013B, Beyotime, China) containing 1 % protease inhibitor ST507 (#P1050, Beyotime, China). Protein concentrations were determined using a bicinchoninic acid assay kit (#U8915, TIANGEN, China). Twenty micrograms of proteins were separated in 10 % SDS-PAGE (#HC015, YTHX, China) and transferred to PVDF membranes (#BS-PVDF-45, Biosharp, China). Membranes were blocked with 5 % bovine serum albumin (BSA, #ST025, Beyotime, China) in Tris-buffered saline and Tween 20 (TBST, #HE011, YTHX, China) for 1 h and then incubated with primary antibody at 4 °C overnight. The following primary antibodies were used: rabbit anti-Piezo1 antibody (1:100 dilution, #DF12083, Affinity Biosciences, China) and rabbit anti-β-actin antibody (1:500 dilution, #GB11001-100, Servicebio, China), anti-p-cofilin and anti-cofilin antibodies (1:100 dilution, #SC365882 and #SC376476, Santa, USA). After three washes with TBST, membranes were incubated with secondary horseradish peroxidase-conjugated anti-rabbit IgG antibody (1:5000 dilution, #ZB2301, Origene, China) for 1 h at room temperature. After extensively washed in TBST, protein bands were visualized using enhanced chemiluminescence (#AR1197, BOSTER, USA) and images captured by MiniChemi (Tanon-4800, Tanon, China) and analyzed with Image J. β-actin was used as loading control.

2.7
Flow cytometry
The protein expressions of Piezo1 on cell surface and the level of phospho-SSH1 (p-SSH1) were determined by flow cytometry as previously described [36]. Briefly, cells were harvested and fixed with 4 % paraformaldehyde for 30 min, then permeabilized with 0.3 % Triton X-100 for 10 min. After that, cells were incubated with the primary polyclonal antibody recognizing the extracellular domain of Piezo1 protein (1:100 dilution, #15939–1-AP, Proteintech, USA) or p-SSH1 (p Ser978) (1:100 dilution, #NBP3-23411, Novus Biologicals) for 40 min. Then, cells were incubated with FITC-labeled goat anti-rabbit secondary antibody solution (1:100 dilution, #A0562, Beyotime, China) at room temperature for 1 h away from light. Cells were incubated with the same secondary antibody as negative control (NC). The fluorescence intensity was detected by flow cytometer (BD Accuri C6). The protein expression was derived by subtracting the fluorescence intensity of NC from the fluorescence intensity of specific binding. All data were analyzed using FlowJo software.

2.8
Calcium imaging
Changes in the [Ca2+]i were monitored by single-cell Ca2+ imaging as previously described [37]. Briefly, 2 × 104 cells were seeded on 32-mm coverslips with PA hydrogels and cultured in 10 % FBS-containing RPMI 1640 medium with or without 2.5 μM GsMTx4 for 48 h. After that, cells were washed twice with Ca2+-containing buffer (10 mM HEPES, 140 mM NaCl, 2 mM CaCl2, 1.13 mM MgCl2, 4.7 mM KCl, 10 mM glucose) and loaded with 2.5 μM Fluo4-AM (#F14201, ThermoFisher, USA) for 30 min at room temperature. To investigate Piezo1-mediated extracellular Ca2+ influx, cells was imaged firstly in Ca2+-free buffer for 1 min and then in buffer containing 2 mM CaCl2. Images of fluorescence excited at 488 nm and emitted at 515 nm–3 random fields were captured using a confocal microscope (Andor, Dragonfly). The Fluo4 fluorescence intensity in individual cells was analyzed using Image J.

2.9
CaN activity determination
The changes in the CaN activity in A549 cells after indicated treatments were determined using CaN Phosphatase Activity Assay Kit (#ab139461, Abcam), according to the instruction's protocol. Briefly, the cell lysates were mixed with the CaN assay buffer, followed by addition of CaN substrates and incubation for further 10 min. Then, the green assay reagent was added into the mixture, and the OD value at 620 nm was recorded using a microplate reader (Thermo Fisher Scientific). The CaN activity in each group was calculated relative to that in the 3 kPa group.

2.10
Data presentation and statistical analysis
All data are presented as means ± SD, where appropriate. Student's t-test was performed for comparison of two groups, and one-way analysis of variance (ANOVA) followed by post hoc Fisher's test for comparison of three or more groups, with p < 0.05 being considered significant.

3
Results
3.1
The Piezo1 channel expression profile in lung cancer is opposite to that in other cancers
There are only three publications directly researching the association between Piezo1 channel and lung cancer cell migration [[14], [15], [16]]. We analyzed the Piezo1 expression in six types of solid tumors using the GENT2 public datasets. As shown in Fig. 1, the Piezo1 expression was significantly up-regulated in pancreas cancer, breast cancer, brain cancer, liver cancer and ovary cancer. In contrast, the Piezo1 expression was significantly down-regulated in lung cancer. These results show that the expression of the Piezo1 channel in lung cancer is opposite to that in other types of solid tumors.

3.2
Stiff matrix promotes cell migration and downregulates the Piezo1 channel expression
To investigate the effects of matrix stiffness on lung cancer cell migration and the Piezo1 channel expression, we prepared PA gels with stiffness close to that of reported healthy (3 kPa, soft) or cancerous (10 and 20 kPa, stiff) lung tissues and cultured A549 or H460 cells on them (Fig. S1). Transwell assay showed that A549 cells on the 10 and 20 kPa substrates migrated significantly faster than those on the 3 kPa substrate (Fig. 2A and C). Flow cytometry (Fig. 2E and G) and Western blotting (Fig. S2) demonstrated that the levels of both membrane-bound and total Piezo1 channel were significantly lower in A549 cells on the 10 and 20 kPa substrate than on the 3 kPa substrate. Furthermore, H460 cells showed a similar increase in cell migration but decrease in the Piezo1 protein expression on the 10 and 20 kPa substrates in comparison with that on 3 kPa substrate (Fig. 2B–D, F and H). Interestingly, cell migration or the Piezo1 expression exhibited substrate stiffness-dependent increase or decrease. Collectively, these results suggest that stiff substrates promote cell migration but downregulates the Piezo1 channel expression.

3.3
Downregulation of Piezo1 channel expression is crucial for stiff substrate-induced cell migration
To investigate the linkage between the increased cell migration and the decreased Piezo1 expression, we next examined cell migration in the absence or presence of Piezo1 channel blockade or activation. Treatment with GsMTx4 to block the Piezo1 channel significantly promoted A549 cell migration on both soft and stiff substrates (Fig. 3A and D). In contrast, exposure to Yoda1 induced Piezo1 channel activation significantly inhibited cell migration, with a greater inhibition in cells on the stiff substrates (Fig. 3B and E). We further investigated the effect of Piezo1 knockdown on matrix stiffness-regulated cell migration. Treatment with Piezo1-specific siRNA strongly reduced the Piezo1 expression (Fig. S3), and promoted A549 cell migration on both soft and stiff substrates (Fig. 3C and F). Similarly, the channel blockade with GsMTx4 treatment promoted, but the channel activation with Yoda-1 treatment inhibited H460 cell migration (Fig. S4). These results indicate that the Piezo1 channel plays a negative regulatory role in substrate stiffness-induced lung cancer cell migration.

3.4
Downregulation of Piezo1 channel expression facilitates stiff substrate-induced filopodia formation
To clarify how the downregulated Piezo1 channel contributes to stiff substrate-induced increase in A549 cell migration, we focused on filopodium formation, due to its implication in probing ECM stiffness and directing cancer cell migration [20]. More importantly, it has been reported that filopodia formation is negatively associated with the Piezo1 channel expression in Hela cells [38]. We firstly examined the filopodia formation of A549 cells seeding on the 3, 10 and 20 kPa substrates, respectively. Similar to cell migration and Piezo1 expression, the filopodia formation were promoted in a substrate stiffness-dependent manner, which A549 cells on the 10 and 20 kPa substrates exhibited longer filopodia and more numbers of filopodia than those on the 3 kPa substrate (Fig. 4). We further investigated the role of the Piezo1 channel in such stiff substrate-induced enhancement of filopodia formation by blockade with GsMTx4 or activation with Yoda1. Piezo1 channel blockade did not abolish but promoted filopodia formation on both soft (3 kPa) and stiff (20 kPa) substrates (Fig. 5A–C and E). In contrast, Piezo1 channel activation inhibited filopodia formation on the stiff substrate, without effect on the soft substrate (Fig. 5B–D and F). Consistently, Piezo1 channel blockade or activation led to a promotive or inhibitory effect on stiffness-regulated filopodia formation in H460 cells (Fig. S5). These results indicate that the Piezo1 channel mediates substrate stiffness modulation of filopodia formation, with the down-regulated Piezo1 channel facilitating stiff substrate-induced filopodia formation.

3.5
The lower [Ca2+]i caused by the down-regulated Piezo1 channel is required for stiff substrate-induced cell migration and filopodia formation
Both A549 and H460 cells followed the same patterns in the stiffness regulation of the Piezo1 channel expression and the role of Piezo1 in the stiffness regulation cell migration and filopodia formation and, thus, further experiments were carried out using A549 cells to gain the underlying mechanisms. The Piezo1 channel mainly functions as a Ca2+-permeable channel mediating extracellular Ca2+ influx to raise intracellular [Ca2+]i [9,[28], [29], [30]]. Therefore, to probe the mechanisms by which the down-regulated Piezo1 channel regulates stiff substrate stimulation of A549 cell migration and filopodia formation, we first examined the role of the Piezo1 channel in substrate stiffness-induced changes in the [Ca2+]i in A549 cells. As shown in Fig. 6A and B, the [Ca2+]i in cells on the stiff substrate was lower than that in cells on the soft substrate. Piezo1 channel blockade reduced the [Ca2+]i in A549 cells on both soft and stiff substrates and, as a result, abolished the difference in the [Ca2+]i in A549 cells on both substrates (Fig. 6C and D). To further investigate whether this difference in [Ca2+]i between cells on soft and stiff substrates, especially the lower [Ca2+]i in cells on the stiff substrate, was cause by the Piezo1 channel, we compared extracellular Ca2+ influx in cells on the soft and stiff substrates. As shown in Fig. 7A and B, introduction of Ca2+ in extracellular solutions induced extracellular influx that was higher or more noticeable in cells on the soft substrates compared to that in cells on the stiff substrates. Such difference in extracellular Ca2+ influx was largely abolished by prior treatment with GsMTx4 to block the Piezo1 channel (Fig. 7C and D). These results indicate that downregulation of the Piezo1 channel expression induced by stiff substrate restrains extracellular Ca2+ influx, resulting in a lower [Ca2+]i in cells on the stiff substrates.
We next investigated whether such Piezo1-mediated reduction in the [Ca2+]i contributes to stiff substrate-induced regulation of A549 cell migration and filopodia formation. Treatment with BAPTA-AM, a Ca2+ chelator, to reduce the [Ca2+]i significantly promoted A549 cell migration on both soft or stiff substrates (Fig. 8A). Similarly, treatment with BAPTA-AM promoted filopodia formation in cells on both soft and stiff substrates (Fig. 8B). Collectively, these results suggest that the [Ca2+]i negatively regulates stiff matrix-induced A549 cell migration and filopodia formation, and the lower [Ca2+]i due to downregulation of the Piezo1 channel is required for stiff substrate-induced regulation of cell migration and filopodia formation.

3.6
The lower [Ca2+]idue to the down-regulated Piezo1 channel is indispensable for stiff substrate-induced phosphorylation of cofilin
It is reported that the phosphorylated cofilin loses its ability to bind to and cut off F-actin polymerization, thereby facilitating filopodia formation and cancer cell migration [21]. In addition, the status of cofilin phosphorylation is regulated by the [Ca2+]i [39,40]. Therefore, to further decipher the downstream molecular mechanisms by which Piezo1 channel-mediated reduction in [Ca2+]i modulates stiff substrate-induced filopodia formation and cell migration, we investigated the phosphorylation level of cofilin (p-cofilin) in cells on the soft and stiff substrates with or without treatment with BAPTA-AM, Piezo1 channel blockade and activation. As shown in Fig. 9A–C, compared to the soft substrates, stiff substrates significantly promoted the p-cofilin level. In addition, Piezo1 channel blockage increased, whereas Piezo1 channel activation attenuated, the p-cofilin levels in cells on both soft and stiff substrates, (Fig. 9D–G). Furthermore, the p-cofilin levels in cells on both soft and stiff substrates were further enhanced by treatment with BAPTA-AM, suggesting that decrease in the [Ca2+]i favored cofilin phosphorylation (Fig. 9H and I). These results suggest that the [Ca2+]i negatively regulates the p-cofilin level and the lower [Ca2+]i resulting from the down-regulated Piezo1 expression is indispensable for stiff substrate-induced phosphorylation of cofilin.

3.7
Attenuation of CaN/SSH phosphatase by lower [Ca2+]idue to the down-regulated Piezo1 channel is required for stiff substrate-induced phosphorylation of cofilin
As mentioned in the Introduction, the p-cofilin level is also regulated by the CaN/SSH phosphatase, which is activated by a rise in the [Ca2+]i, thereby promoting cofilin dephosphorylation and a consequent decrease in the p-cofilin level [31]. Therefore, to further clarify how the lower [Ca2+]i mediated by the down-regulated Piezo1 channel on stiff substrates increased the p-cofilin level, we firstly examined the p-cofilin level in cells on the soft and stiff substrates with or without treatment with CsA to inhibit CaN. As shown in Fig. 10A and B, the p-cofilin levels in cells on both soft and stiff substrates were significantly enhanced by treatment with CsA, suggesting an engagement of CaN in stiffness regulation of increase in the p-cofilin level. Next, we compared the CaN activity in cells on the soft and stiff substrates with or without Piezo1 channel blockade or intracellular Ca2+ chelation, respectively. As shown in Fig. 10C and D, the CaN activity in cells on the stiff substrate was lower than that in cells on the soft substrate. Piezo1 channel blockade or intracellular Ca2+ chelation reduced the CaN activity in cells on both soft and stiff substrates and, as a result, abolished the difference in the CaN activity in A549 cells on both substrates, suggesting linkage of the lower CaN activity with the lower [Ca2+]i and the Piezo1 expression on the stiff substrates. We further examined the CaN-dependent SSH1 dephosphorylation by comparing the p-SSH1 level in A549 cells on the soft and stiff substrates with or without CaN inhibition. As shown in Fig. 10E and F, the p-SSH1 level in cells on the stiff substrate was higher than that in cells on the soft substrate, and CaN inhibition reduced the p-SSH1 level in cells on both soft and stiff substrates and, furthermore, abolished the difference in the p-SSH1 level in cells on both substrates, suggesting correlation between substrate stiffness regulation of CaN activity and SSH activity. Collectively, these results suggest that the lower [Ca2+]i resulting from the down-regulated Piezo1 expression attenuated the activity of CaN/SSH1, which is required for stiff substrate-induced phosphorylation of cofilin.

4
Discussion
Different from what has been observed in most solid tumors, the Piezo1 channel plays a negative regulatory role in lung cancer cell migration and lung cancer metastasis. The present study shows that ECM stiffness is implicated in such negative regulation and, mechanistically, stiff substrates down-regulate the Piezo1 channel expression and restrain the [Ca2+]i to favor cofilin phosphorylation and facilitates filopodia formation, thereby promoting lung cancer cell migration.
As a common characteristic of solid tumors, stiffening substrates promotes cancer cell migration, as shown by previous [3,4] and present studies (Fig. 2A–D). However, it remains unknown how the Piezo1 channel in lung cancer cells respond to substrate stiffness and related to stiff substrate-induced lung cancer cell migration. The present study showed that blockade of the Piezo1 channel or siRNA-mediated knockdown of its expression promoted cell migration, whereas activation resulted in an inhibitory effect (Fig. 3 and Fig. S4). These results demonstrate that the Piezo1 channel is engaged in stiff substrate-induced lung cancer cell migration in a negative manner. Similar to our findings, Pathak and colleagues reported that knockout of the Piezo1 channel expression promoted, whereas the Piezo1 channel activation attenuated keratinocyte migration in skin wound healing, where the skin wound and skin scar are stiff compared to normal skin tissues [41].
Filopodia is crucial in probe probing ECM stiffness and directing cancer cell migration [42]. In this study, we showed that the downregulated Piezo1 channel expression in A549 and H460 cells on the stiff substrate enhanced filopodia formation. In contrast, the higher Piezo1 channel expression in cells on the soft substrates resulted in less filopodia formation (Fig. 2, Fig. 4). Furthermore, Piezo1 channel blockade exerted a stimulatory effect and, by contrast, channel activation exerts an inhibitory effect on filopodia formation on the stiff substrate (Fig. 5 and Fig. S5). Yang et al. have demonstrated that Piezo1 overexpression in HEK239T cells inhibited filopodia formation and, as reported in their study, the Piezo1 channel is depleted in highly curved membrane protrusions such as filopodia and enriched to nanoscale membrane invaginations [38]. Our results are consistent with those by Yang et al. and together highlights an important role of the Piezo1 channel in substrate-induced negative regulation of stiff lung cancer cell migration through modulating filopodia formation.
Piezo1 channel functions by mediating extracellular Ca2+ influx to raise the [Ca2+]i [9,[28], [29], [30]] and the [Ca2+]i is critical in regulating cell migration through modulating filopodia formation [40,43,44]. We showed that Piezo1 channel blockade attenuated the [Ca2+]i in cells on both soft and stiff substrates (Fig. 6). Particularly, we showed that it is through downregulating the Piezo1 channel expression that resulted in the lower [Ca2+]i in cells on the stiff substrates; there were higher extracellular Ca2+ influx in cells on the soft substrate than in cells on the stiff substrate, and Piezo1 channel blockade abolished the difference (Fig. 7). Accordingly, treatment with BAPTA-AM to reduce the [Ca2+]i produced a stimulatory effect on lung cancer migration on both soft and stiff substrate (Fig. 8A). Similar to the effects on cell migration, reduction in the [Ca2+]i also promoted filopodia formation in cells on both soft and stiff substrates (Fig. 8B). Our results are consistent to that from a previous report that Ca2+ overload blunts filopodia formation and inhibits cell migration in endothelial cells [40] and provide evidence to suggest that the Piezo1 channel negatively regulates stiff substrate-induced lung cancer cell migration through reducing the [Ca2+]i and facilitating filopodia formation.
It is well established that the status of cofilin phosphorylation is crucial for actin reorganization, especially filopodia formation, and thereby cancer cell migration [[45], [46], [47], [48], [49], [50], [51]]. It has also been reported that the status of cofilin phosphorylation is regulated by the Ca2+-dependent CaN/SSH phosphatase pathway [31]. However, the relationships remain unknown among the Piezo1 channel, [Ca2+]i, CaN, SSH, cofilin phosphorylation and filopodia formation in cancer cell migration, especially in substrate stiffness regulation of cell migration. Our results show that stiff substrate downregulated the Piezo1 channel expression and thereby reduced the [Ca2+]i to attenuate CaN/SSH1-mediated cofilin dephosphorylation, promote A549 cell migration, filopodia formation and cofilin phosphorylation (Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8, Fig. 9, Fig. 10). Our finding is consistent with a recent study showing that Piezo1 overexpression in HEK239T cells promoted cofilin dephosphorylation and F-actin filaments severance [52]. However, further investigations are required to clarify how Piezo1 channel-dependent reduction in [Ca2+]i promotes cofilin phosphorylation. In addition, the current results and mechanistic explanation are from A549 and H460 cells and it is interesting to evaluate the findings from this study in another lung cancer lines or patient-derived cells in future. We also try to ascertain how stiff substrates downregulate the Piezo1 channel expression, focused on the canonical mechano-transduction pathways, integrin-FAK and Hippo-YAP pathway, and found that the two pathways are not engaged in stiffness downregulation of the Piezo1 channel expression in A549 cells (Fig. S6). These results are to some extent consistent with the previous findings, which depletion of the Piezo1 expression promoted lung cancer cell migration in an integrin-independent manner [15], and Piezo1 acts as the upstream but not downstream molecular of YAP [53,54]. Therefore, it is also interesting to identify the signaling pathway engaged in stiff substrate downregulation of the Piezo1 channel expression in future work.

5
Conclusions
In summary, our study has revealed a negative regulatory role of the Piezo1 channel in mediating stiff matrix-induced lung cancer cell migration (Fig. 11). Mechanistically, stiff substrates downregulate the Piezo1 channel expression and thereby reduces Piezo1 channel-mediated extracellular Ca2+ influx to restrain an increase in [Ca2+]i, which favors phosphorylation/inactivation of cofilin through attenuating CaN/SSH1-mediated cofilin dephosphorylation and facilitates filopodia formation, leading to lung cancer cell migration on stiff substrates. Clinical data from Public datasets (Fig. 1) and experimental data from limited researches [[14], [15], [16]] suggest that the expression of Piezo1 channel is decreased in lung cancer and negatively correlated with lung cancer metastasis, but the underlying mechanism is unknown. The findings described in this study provide a previously unrecognized mechanistic explanation and broaden our understanding of the molecular mechanism how the Piezo1 channel functions in lung cancer and provided opportunities for the development of new lung cancer treatment.

CRediT authorship contribution statement
Xiaoling Jia: Writing – original draft, Supervision, Resources, Methodology, Conceptualization. Lin Zhao: Writing – review & editing, Formal analysis, Data curation. Juncheng Bai: Methodology, Data curation. Lu Wen: Methodology, Data curation. Qianyu Meng: Methodology. Haikun Wang: Data curation. Junqi Men: Methodology. Hui Shao: Methodology. Yingying Guo: Methodology. Xinlan Chen: Data curation. Xing Chen: Resources. Lin-Hua Jiang: Writing – review & editing, Supervision. Yubo Fan: Writing – review & editing, Supervision. Huawei Liu: Conceptualization, Funding acquisition, Resources.

Statement of significance
This is the first study that reports a negative regulatory role of the Piezo1 channel in matrix stiffness regulation of lung cancer cell migration and the molecular mechanism involved. We have revealed that stiff matrix down-regulates the Piezo1 channel expression and thereby maintain a low [Ca2+]i to favor cofilin phosphorylation and facilitate filopodia formation, consequently, promoting cell migration. These findings broaden our understanding of the molecular mechanism how the Piezo1 channel functions differently in lung cancer from in most cancers, and provide new strategies for lung cancer treatment.

Declaration of competing interest
The authors declare no competing financial interests or personal relationships that influence the work reported in this paper.