Overexpression of caveolin-1 in a human melanoma cell line results in dispersion of ganglioside GD3 from lipid rafts and alteration of leading edges, leading to attenuation of malignant properties

Abstract

Caveolin-1 is a component of lipid rafts, and is considered to be a tumor suppressor molecule. However, the mechanisms by which caveolin-1 functions in cancer cells are not well understood. We generated caveolin-1 transfectant cells (Cav-1(+) cells) using a human melanoma cell line (SK-MEL-28) and investigated the effects of caveolin-1 overexpression on the GD3-mediated malignant properties of melanomas. Cav-1(+) cells had decreased cell growth and motility, and reduced phosphorylation levels of p130Cas and paxillin relative to controls. In floatation analysis, although GD3 was mainly localized in glycolipid-enriched microdomain (GEM)/rafts in control cells, it was dispersed from GEM/rafts in Cav-1(+) cells. Correspondingly, GD3 in Cav-1(+) cells stained uniformly throughout the membrane, whereas control cells showed partial staining of the membrane, probably at the leading edge. p130Cas and paxillin were stained in the leading edges and colocalized with GD3 in the control cells. In contrast, these molecules were diffusely stained and no definite leading edges were detected in Cav-1(+) cells. These results suggest that caveolin-1 regulates GD3-mediated malignant signals by altering GD3 distribution and leading edge formation. These results reveal one of the mechanisms by which caveolin-1 curtails the malignant properties of tumor cells.

Figure 1

Expression analysis of caveolin‐1 and establishment of stable transfectants. a, Immunoblotting to examine the expression levels of caveolin‐1 in SK‐MEL‐28 and N1, a mutant line of SK‐MEL‐28, was performed using a rabbit anti‐caveolin‐1 antibody. b, After transfection of SK‐MEL‐28 cells with a caveolin‐1 synthase gene expression vector (pCMV‐Tag 3/caveolin‐1), or pCMV‐Tag 3, two transfectant lines (Cav12, Cav35) and two vector controls (V1 and V2) were established. Total cell lysates were subjected to immunoblotting with anticaveolin‐1 (upper panel) or antic‐myc (middle panel) antibodies. The location of the fusion protein is indicated by arrows.

Figure 2

Effects of caveolin‐1 expression on cell proliferation, motility and invasion. a, Analysis of cell proliferation. Cells (2 × 10³) were seeded onto 96‐well plates. On days 1, 3, 6, and 9 of culture, the MTT assay was performed. Data are presented as relative absorbance, with absorbance on day 1 set as 1.0. *: P < 0.05, **: P < 0.001. b and c, Cell motility of the transfectants was analyzed using wound‐healing scratching motility assays. B, Photomicrographs of the scratched regions of vector control (V1) and Cav‐1⁺ transfectant (Cav12) cultures at times 0, 12, and 24 h. C, Results of the motility assay. Wound areas are presented as a percentage of the initial wound size (100%). The time course of the wound areas at times 0, 6, 12, 18, and 24 h is shown. Mean values (n = 12) were plotted for individual points. *: P < 0.05. d, Result of invasion assay. No significant difference in invasion activity, as analyzed by using the Boyden chamber, was found between control and Cav‐1⁺ transfectant cells.

Figure 3

Reduced phosphorylation of p130Cas and paxillin after fetal calf serum (FCS) treatment of transfectant cells. a, Time course of the phosphorylation levels of p130Cas and paxillin in vector control (V1 and V2) and Cav‐1⁺ (Cav12 and Cav35) cells. Cells (2 × 10⁵) were treated with FCS after serum starvation for 12 h, and the phosphorylation levels of p130Cas and paxillin were observed up to 60 min after addition of FCS. Immunoblotting was carried out using PY20. b, Relative intensities of bands in a were plotted. c, Identification of bands at 130 kDa and 68 kDa as p130Cas and paxillin, respectively. Immunoprecipitation/immunoblotting (IP/IB) was performed to identify the two tyrosine‐phosphorylated components using PY20 and antip130Cas (a), and PY20 and antipaxillin antibodies (b) in vector control (V1) and Cav‐1⁺ cells (Cav35). Immunoprecipitates with PY20 (PY20 IP) and those with antip130Cas (p130Cas IP) together with total lysate were immunoblotted using PY20 (Ca, Cb, left) or anti‐p130Cas (Ca, Cb, right). d, Immunoprecipitates with PY20 or antipaxillin antibody were immunoblotted as performed for p130Cas in a to identify the 68 kDa band as paxillin (Da, Db).

Figure 4

Alteration in the floatation pattern of GD3 in sucrose density gradient fractionation. Fractionation patterns for control and Cav‐1⁺ cells are shown. Triton X‐100 cell extracts were fractionated by sucrose density gradient ultracentrifugation, and each fraction was used for immunoblotting. Reagents used for the detection of individual molecules are described in the Materials and Methods. GD3 was detected with mAbR24. a, Results for the vector control cells (V1 and V2). b, Results for the Cav‐1⁺ cells (Cav12 and Cav35). Note that GD3 was widely distributed inside and outside of the GEM/rafts in the Cav‐1⁺ cells.

Figure 5

Localization of caveolin‐1 and GD3 with immunocytostaining. a, Staining pattern of caveolin‐1 and GD3 in vector control (V1 and V2) and Cav‐1⁺ (Cav12 and Cav35) cells as analyzed by confocal laser microscopy. Cells were cultured on a cover glass and fixed with 4% paraformaldehyde in PBS. GD3 was stained with mAbR24 and anti‐mouse IgG3‐Alexa555. Caveolin‐1 was stained with rabbit anti‐caveolin‐1 antibody and anti‐rabbit IgG‐Alexa488. b, GD3 staining patterns on the membranes facing the scratched areas. To compare the directions of the leading edges of the migrating cells (indicated by →), a wound was generated by scratching the culture with a pipette tip. After 3 h, the wounded monolayer was fixed, and GD3 was stained using mAbR24 and anti‐mouse IgG3‐Alexa555. Caveolin‐1 was stained with rabbit anti‐caveolin‐1 antibody and anti‐rabbit IgG‐Alexa488. The left side is the scratched area.

Figure 6

Intracellular localization of p130Cas and paxillin changed in parallel with GD3. Cells were cultured on a cover glass and fixed, then stained as described in the Materials and Methods. a and b, GD3 was stained with mAbR24 and antimouse IgG3‐Alexa555. p130Cas and paxillin were stained with rabbit antip130Cas antibody and antirabbit IgG‐Alexa488 (a) or antipaxillin mouse mAb and antimouse IgG1‐Alexa488 (b), respectively. The staining pattern was analyzed by confocal laser microscopy. The results for two control lines (V1 and V2) and two Cav‐1⁺ transfectant lines (Cav12 and Cav35) are shown. Note that no clear leading edges were observed for the transfectant cells.

Figure 7

TLC and mass spectrometry analysis of GD3 in the GEM/raft and non‐GEM/raft fractions. a, Gangliosides were extracted from GEM/raft (2 and 3) and non‐GEM/raft (5–7 and 8–10) fractions from vector control (V1) and Cav‐1⁺ (Cav12) cells, then separated by TLC. The solvent used was chloroform/methanol/2.5 N NH₄Cl (55:45:10). Resorcinol spray was used for the detection of bands. B, Molecular species of GD3 were analyzed by ESI‐MS using a Q‐Tof micro^TM quadrupole time‐of‐flight hybrid mass spectrometer (Micromass, Manchester, UK) with an Ultimate HPLC system as described in the Materials and Methods. The results of samples from GEM/raft (2 and 3) and non‐GEM/raft (5–7) fractions are shown. Peaks at 1442.90–1442.98 (m/z) were assigned d18:1–16:0, those at 1555.07–1555.11 were d18:1–24:0, and those at 1553.08–1553.11 were 18:1–24:1.