Extracellular matrix-specific Caveolin-1 phosphorylation on tyrosine 14 is linked to augmented melanoma metastasis but not tumorigenesis

Abstract

Caveolin-1 (CAV1) is a scaffolding protein that plays a dual role in cancer. In advanced stages of this disease, CAV1 expression in tumor cells is associated with enhanced metastatic potential, while, at earlier stages, CAV1 functions as a tumor suppressor. We recently implicated CAV1 phosphorylation on tyrosine 14 (Y14) in CAV1-enhanced cell migration. However, the contribution of this modification to the dual role of CAV1 in cancer remained unexplored. Here, we used in vitro [2D and transendothelial cell migration (TEM), invasion] and in vivo (metastasis) assays, as well as genetic and biochemical approaches to address this question in B16F10 murine melanoma cells. CAV1 promoted directional migration on fibronectin or laminin, two abundant lung extracellular matrix (ECM) components, which correlated with enhanced Y14 phosphorylation during spreading. Moreover, CAV1-driven migration, invasion, TEM and metastasis were ablated by expression of the phosphorylation null CAV1(Y14F), but not the phosphorylation mimicking CAV1(Y14E) mutation. Finally, CAV1-enhanced focal adhesion dynamics and surface expression of beta1 integrin were required for CAV1-driven TEM. Importantly, CAV1 function as a tumor suppressor in tumor formation assays was not altered by the Y14F mutation. In conclusion, our results provide critical insight to the mechanisms of CAV1 action during cancer development. Specific ECM-integrin interactions and Y14 phosphorylation are required for CAV1-enhanced melanoma cell migration, invasion and metastasis to the lung. Because Y14F mutation diminishes metastasis without inhibiting the tumor suppressor function of CAV1, Y14 phosphorylation emerges as an attractive therapeutic target to prevent metastasis without altering beneficial traits of CAV1.

Keywords: Caveolin-1; cancer; dual role; invasion; migration.

Figure 1. CAV1-enhanced B16F10 cell migration and invasion are dependent on tyrosine 14

A. Mutated versions of CAV1 in Y14, (CAV1/Y14F and CAV1/Y14E; non phosphorylatable and phosphomimetic, respectively) generated by site-directed mutagenesis are depicted in a scheme. B. B16F10 cells were transfected with empty vector pLacIOP, pLacIOP-(CAV1/wt), pLacIOP-(CAV1/Y14F) or pLacIOP-(CAV1/Y14E) (see Materials and Methods for details) to generate stably transfected B16F10(mock), B16F10(CAV1/wt), B16F10(CAV1/Y14F) and B16F10(CAV1/Y14E) cells, respectively. Post-selection with hygromycin B, cells were induced with 1 mM IPTG for 48 h and treated with 5 mM H₂O₂ for 20 min to induce CAV1-phosphorylation on Y14, for analysis by Western blotting. Relative CAV1 levels normalized to β-Actin by scanning densitometry are shown as the fold-increase with respect to the (mock) condition. C. and D. B16F10 cells (5×10⁴) were added to transwell inserts pre-coated on the lower side with fibronectin (2 μg/ml). Cells were allowed to migrate for 2 h and then detected after fixation on the lower side of the membrane by crystal violet staining. (C) Images of the transwell inserts viewed at 400X magnification are shown (scale bar 100 μm). (D) Data averaged from 6 different fields in three independent experiments and normalized to values for mock cells are shown (mean ± S.E.M, **p<0.01). E. and F. B16F10(mock), (CAV1/wt), (CAV1/Y14F) and (CAV1/Y14E) cells (5×10⁴) were added to matrigel inserts, allowed to invade for 22 h and then detected and quantified in the same manner as in D. (E) Images of the matrigel inserts viewed at 200X magnification are shown (scale bar 200 μm). (F) Data averaged from 6 different fields in three independent experiments were normalized to values obtained for B16F10(mock) cells (mean ± S.E.M, *p<0.05).

Figure 2. CAV1-increased migration on fibronectin and laminin requires tyrosine 14 in B16F10 cells

B16F10(mock), (CAV1/wt), (CAV1/Y14F) and (CAV1/Y14E) cells were induced with IPTG (1 mM) for 48 h. Then, 1×10⁶ cells were seeded in migration micro-devices, pre-coated in the side-channels with fibronectin (50 μg/ml) or laminin (50 μg/ml). Cells were allowed to attach for 2 h in the central chamber. Then, side-channels were filled with culture media and migration was recorded by time-lapse video microscopy for 7 h at 15-min time intervals. Cell tracks were determined using the Image J Software (“Manual Tracking” plug-in). A. The Instant velocity (μm/min) at any given time point was analyzed for individual cells during tracking on fibronectin. B. The Average velocity was obtained as the quotient between the Euclidean distance (μm) and the total time of migration while tracking the cells on fibronectin. C. Persistency of migration was calculated as the ratio between the net distance and the total distance of migration on fibronectin. D. Individual cell tracks on fibronectin are shown in a Cartesian coordinate system for each cell type. E. Directionality of migration (% of cells) on fibronectin was obtained from D, whereby tracks within a 60° angle with respect to the direction of cell movement were considered as oriented (shaded region). Migration on laminin; F. Instant velocity, G. Average velocity, and H. Persistency of migration on laminin are shown. I. Individual cell tracks and J. Directionality of migration on laminin were obtained from I, as described above. Graphs show values of each parameter averaged from three independent experiments (mean ± S.E.M, n = 3, ***p<0.001; **p<0.01 and *p<0.05).

Figure 3. CAV1-phosphorylation on tyrosine 14 during cell spreading on pure ECM surfaces

In spreading assays, B16F10(CAV1/wt) cells (1,5×10⁶) were allowed to attach to A. fibronectin, B. laminin, C. vitronectin and D. collagen IV-coated plates (2 μg/ml) for different periods of time (0, 5, 15, 30 and 60 min), with time 0 representing cells in suspension. Then, whole cell lysates were prepared and pY14-CAV1 levels were determined by Western blotting. Upper graphs show the densitometric analysis of relative pY14-CAV1 levels during cell spreading. Lower images show pY14-CAV1, CAV1 and Actin (control) expression by Western Blotting. Lower panels show cells in phase contrast and stained with phalloidin during spreading. The average area per cell is indicated in μm². Data shown are the averages from three independent experiments (mean ± S.E.M, n=3,***p<0.001; **p<0.01 and *p<0.05).

Figure 4. CAV1 distribution during spreading in B16F10 cells

B16F10(mock), (CAV1/wt), (CAV1/Y14F) and (CAV1/Y14E) cells were induced with 1mM IPTG for 48 h. Cells were seeded on fibronectin-coated chambered slides (2 μg/ml) and grown in the presence of IPTG (1 mM) for 24 h. Thereafter, cells were serum-starved for 60 min, pulsed with 3% serum and fixed at 15, 30 and 45 min of spreading for CAV1 detection in immunofluorescense experiments. Samples were analyzed with the Fiji Software. A–C. Images of B16F10(mock), (CAV1/wt), (CAV1/Y14F) and (CAV1/Y14E) are shown at 15 (A), 30 (B) and 45 (C) min of spreading. Red stain corresponds to CAV1 expression. D. Image of B16F10(CAV1/wt) showing ROI definition in the cell periphery. Total fluorescence and ROI (border fluorescence) were quantified at 30 and 45 min of spreading. E. In the graph the distribution of CAV1 in the cell periphery is shown in percent (%), calculated as (border fluorescence*100)/whole cell fluorescence). For each experimental condition, at least 5 individual cells were analyzed. Data shown are the averages from three independent experiments (mean ± S.E.M, n=3,***p<0.001; **p<0.01 and *p<0.05).

Figure 5. CAV1-enhanced FA dynamics is dependent on tyrosine 14

B16F10(mock), (CAV1/wt), (CAV1/Y14F) and (CAV1/Y14E) cells were induced with 1mM IPTG for 48 h. Cells were transfected with pEGFP-vinculin 24 h prior to the experiment, then seeded on fibronectin-coated chambered slides (2 μg/ml) and grown in the presence of IPTG (1 mM) for 24 h. Thereafter, cells were serum-starved for 60 min, pulsed with 3% serum and recorded by time-lapse video microscopy for 60 min (2-min time intervals). A. FA assembly after cells begin to attach to the substrate (time of FA formation); B. and C. FA disassembly (time of FA disappearance) were measured for at least 10 structures per experiment (scale bar, 50 μm). Note that the kinetics reported in (A) and (B) were obtained from the same set of time-lapse video microscopy experiments. Digital zoom areas in C are shown at selected time points for each cell type. Focal adhesions (FAs, arrows) were defined by size. Images are representative of four independent experiments. Statistically significant differences are indicated (mean ± S.E.M; ***p<0.001 and *p<0.05).

Figure 6. CAV1-enhanced wound closure and transendothelial migration require tyrosine 14 and beta1 surface expression in B16F10 cells

B16F10(mock), (CAV1/wt), (CAV1/Y14F) and (CAV1/Y14E) cells were induced with 1mM IPTG for 48 h. Cells were then trypsinized, fixed and immunostained for beta1 and alpha5 integrins and analyzed by flow cytometry. A. Beta1 integrin fluorescence intensity. B. alpha5 integrin fluorescence intensity. C. Confluent monolayers of B16F10 cell lines were wounded with a pipette tip, incubated with anti-beta1 or anti-alpha5 integrin antibodies (5 μg) and images were recorded at 0 and 7 h post-wounding. As a control (CTRL), a non-related anti-GFP antibody was used. The wounded area was measured with the Adobe Photoshop software and the percentage (%) of wound closure in 7 h is plotted for the indicated condition. D. EA.hy926 cells (2,5 × 10⁵) were seeded on 24-well plates and impermeable cell monolayers were allowed to form for 72 h. B16F10(mock), (CAV1/wt), (CAV1/Y14F) and (CAV1/Y14E) cells (5×10⁴), previously stained with CellTracker green and incubated for 1 h with anti-HA (CTRL), anti-beta1 or anti-alpha5 integrin antibodies, were added to the EA.hy926 monolayer. Then, B16F10 cells were allowed to adhere to the EA.hy926 monolayer for 1 h (scale bar, 100 μm). E. The graph represents the average for adhesion (cells per field) following incubation of the B16F10 cells with the antibodies mentioned above. F. EA.hy926 cells (2,5 × 10⁵) were seeded on the Transwell inserts and impermeable cell monolayers were allowed to form for 72 h. B16F10 cell lines (5×10⁴), previously stained with CellTracker green and incubated for 1 h with the neutralizing antibodies using the same procedure described above, were added to the EAhy monolayer in the inserts. Then, B16F10 cells were allowed to penetrate the EA.hy926 monolayer for 6 h. B16F10 cells observed by epifluorescence microscopy with a 40X objective on the lower side of the Transwell membrane are shown (scale bar, 50 μm). G. Values in the graph represent the average of TEM (cells per field) following incubation of the in B16F10 cells with the different antibodies. Data were normalized to values obtained for control (mock) cells. Adhesion and TEM was quantified as cells per field from 10 different fields in three independent experiments (mean ± S.E.M, n=3, ***p<0.001; **p<0.01 and *p<0.05).

Figure 7. CAV1-enhanced lung metastasis of B16F10 melanoma cells is dependent on tyrosine 14

C57BL/6 mice were intravenously injected with B16F10(mock), (CAV1/wt), (CAV1/Y14F) or (CAV1/Y14E) cells (5×10⁵), previously grown for 48 h in the presence of IPTG (1 mM). A. The images show the black metastatic lung mass after sacrificing the animals at day 21. B. The graph shows the results of 44 mice in total (11 per group). The lung tumor mass in C57BL/6 mice for B16F10(mock), B16F10(CAV1/wt), B16F10(CAV1/Y14F) and B16F10(CAV1/Y14E) cells was 7%, 35%, 13% and 34%, respectively. C. and D. C57BL/6 mice were subcutaneously injected with B16F10(mock), (CAV1/wt), (CAV1/Y14F) or (CAV1/Y14E) cells (1×10⁵) previously grown for 48 h, in the presence of IPTG (1 mM). (C) Tumor volume (mm³) was monitored in each animal. Results shown are the average from data obtained with 8 mice per group between day 7 and day 15. (D) The average tumor volumes (mm³) measured on day 15 for the 4 groups of animals (32 in total) are shown. For CAV1/wt, CAV1/Y14F, CAV1/Y14E expressing cells and mock-transfected controls, tumor volumes were 1693 mm³ (S.D ± 655); 323 mm³ (S.D ± 252); 455 mm³ (S.D ± 298) and 418 mm³ (S.D ± 284), respectively. Statistically significant differences are indicated (***p<0.001; **p<0.01 and *p<0.05).

Figure 8. Schematic summary of data showing that CAV1-enhanced migration, invasion, TEM and metastasis are dependent on tyrosine 14 and membrane expression of beta1 (and alpha5) integrins

CAV1 is shown as a dimer associated with the plasma membrane. For simplicity, higher states of oligomerization were not considered. Specific ECM-integrin interactions stimulate phosphorylation of CAV1 on Y14 mediated by Src family kinases (SFK), which favors FA dynamics, migration and invasion by melanoma cells. Note that alpha5 is shown here as the beta1 binding partner, although our results indicate that other relevant binding partners must exist. On the other hand, the expression of CAV1 in melanoma cells is shown to enhance beta1 and alpha5 surface expression, both necessary for adhesion of melanomas to EA.hy926 vascular endothelial cells, possibly via interactions with CAMs (not shown). Furthermore, phosphorylation on Y14 and the surface expression of beta1 integrin are depicted as critical elements for CAV1-enhanced TEM. The presence of CAV1 enhances melanoma adhesion to the endothelium (1 and 2 in TEM scheme) and favors extravasation (3), as well as further colonization into the lung matrix (lung colonization). Taken together, these observations provide a rationale to explain CAV1-enhanced lung metastasis of B16F10 cells.