Ganglioside GM2/GM3 complex affixed on silica nanospheres strongly inhibits cell motility through CD82/cMet-mediated pathway

Abstract

Ganglioside GM2 complexed with tetraspanin CD82 in glycosynaptic microdomain of HCV29 and other epithelial cells inhibits hepatocyte growth factor-induced cMet tyrosine kinase. In addition, adhesion of HCV29 cells to extracellular matrix proteins also activates cMet kinase through "cross-talk" of integrins with cMet, leading to inhibition of cell motility and growth. Present studies indicate that cell motility and growth are greatly influenced by expression of GM2, GM3, or GM2/GM3 complexes, which affect cMet kinase activity of various types of cells, based on the following series of observations: (i) Cells expressing CD82, cultured with GM2 and GM3 cocoated on silica nanospheres, displayed stronger and more consistent motility inhibition than those cultured with GM2 or GM3 alone or with other glycosphingolipids. (ii) GM2-GM3, in the presence of Ca2+ form a heterodimer, as evidenced by electrospray ionization (ESI) mass spectrometry and by specific reactivity with mAb 8E11, directed to GM2/GM3 dimer structure. (iii) Cells expressing cMet and CD82 were characterized by enhanced motility associated with HGF-induced cMet activation. Both cMet and motility were strongly inhibited by culturing cells with GM2/GM3 dimer coated on nanospheres. (iv) Adhesion of HCV29 or YTS-1/CD82 cells to laminin-5-coated plate activated cMet kinase in the absence of HGF, whereas GM2/GM3 dimer inhibited adhesion-induced cMet kinase activity and inhibited cell motility. (v) Inhibited cell motility as in i, iii, and iv was restored to normal level by addition of mAb 8E11, which blocks interaction of GM2/GM3 dimer with CD82. Signaling through Src and MAP kinases is activated or inhibited in close association with cMet kinase, in response to GM2/GM3 dimer interaction with CD82. Thus, a previously uncharacterized GM2/GM3 heterodimer complexed with CD82 inhibits cell motility through CD82-cMet or integrin-cMet pathway.

Fig. 1.

Inhibitory effect of GSLs on cell motility and inhibition of cMet kinase pathway, and interaction of GSLs with CD82. (A) Comparative effect of soluble GSL (a) vs. GSL-coated nanospheres (b) on haptotactic motility. (a) Effect of RPMI medium (column 1) or medium containing 25 nmol/ml of GM2 (column 2), GM3 (column 3), or GM2/GM3 (column 4) on haptotactic motility of HCV29 vs. YTS-1 cells. (b) Effect of noncoated (column 1), GM2-coated (column 2), GM3-coated (column 3), GM2/GM3-cocoated (column 4), Gb3-coated (column 5), Gb3/GM2-cocoated (column 6), LacCer-coated (column 7), or LacCer/GM2-cocoated (column 8) silica nanospheres on cell motility. The type of cell is indicated at the top. Cells (0.5 × 10⁴ per well), in medium with addition of aqueous GSL solution or GSL-coated silica nanospheres, containing 5% FBS and 50 ng/ml HGF, were seed onto gold sol-coated 48-well plates and haptotactic motility was assessed as in Materials and Methods. For methods for preparation of SiO₂ nanospheres and GSL coating, see Materials and Methods. (B) Interaction of GSLs with CD82, probed by binding of CD82 on polystyrene beads alone (lane a), GM2-coated (lane b), GM2-GM3-cocoated (lane c), GM3-coated (lane d), LacCer-coated (lane e), or LacCer-GM2-cocoated (lane f) polystyrene beads. Beads (6.85 × 10⁷, 1-μm diameter) prepared as described in ref. were incubated with 100 μg of YTS-1/CD82 cell lysate overnight at 4°C, washed three times with TBS [140 mM NaCl and 10 mM Tris·HCl (pH 8.0)] (+), resuspended in SDS/PAGE sample buffer, and analyzed by Western blotting with anti-CD82 Ab. (C) Effect of GSL-coated nanospheres on HGF-induced cMet phosphorylation of HCV29 and YTS-1/CD82 cells. Cells were incubated overnight with 1 ml of serum-free RPMI medium 1640, containing 18.9 × 10¹⁶ nanospheres (50-nm diameter) coated with ganglioside GM2 (column 2), GM3 (column 3), GM2/GM3 (column 4) or uncoated nanospheres (column 1), with final concentration of 25 nmol/ml. Cells were then treated with 50 ng/ml HGF for 10 min, washed, and lysed, and 200 μg of protein was immunoprecipitated with anti-Met antibody as described in ref. . Levels of tyrosine phosphate in immunoprecipitated fractions were measured by using anti-phosphotyrosine antibody (Py20) (upper row), and stripped blots were probed by anti-Met antibody (lower row). Ratios of phospho-Met/Met are shown. (D) The difference in ratio of cMet tyrosine phosphorylation relative to total cMet level of YTS-1/CD82 cells pretreated with GM2 (lane 2), GM3 (lane 3), GM2/GM3 (lane 4), or no GSL (lane 1) was compared for cells in suspension (Left) vs. cells adhered to LN5-coated plate (Right). (E) Effect of GSL-coated nanospheres on Src phosphorylation of HCV29 cells. Cells were incubated overnight with 1 ml of serum-free RPMI medium 1640 containing 18.9 × 10¹⁶ nanospheres (50-nm diameter) noncoated (lane 1); coated with ganglioside GM1 (lane 2), GM2 (lane 3), GM3 alone (lane 4), or GM1-GM2 (lane 5); or cocoated with GM2-GM3 (lane 6) at final concentrations of 25 nmol/ml. Cells were washed and lysed. Cell lysate containing 15 μg of protein was Western blotted by using anti-P-Src (Tyr-416), and stripped blots were probed by anti-c-Src antibody. Intensity of Western blot was determined by densitometry, using Scion image program. Ratio of P-Src (top row) relative to total c-Src level (lower row) is shown in each column. (F) Ratio of phospho-p44/42 MAPK (top row) relative to p44/42 MAPK (lower row), analyzed as above. Data in A, and C–F are presented as means ± SD. Significance of difference versus control cells: **, P ≤ 0.01; *, P ≤ 0.001. The results in B–D are representative of three experiments.

Fig. 2.

GM2/GM3 heterodimer indicated by mass spectrometric pattern. (A) Negative ion spray mass spectrum of GM2 (20 nM) plus GM3 (20 nM) with Ca²⁺ (200 nmol) in methanol. Ion m/z = 1,383 (region b) corresponds to [GM2-H⁺]⁻. Ions m/z = 1,208, 1,222, 1,236, 1,250, and 1,264 (region a) correspond to [GM3-H⁺]⁻ and represent a mixture of GM3 molecules containing C14, C18, C19, C20, C21, and C22 fatty acids, respectively. (B) (Region b) Positive ion spray mass spectrum of GM2 (20 nmol) plus GM3 (20 nmol) with Ca²⁺ (200 nmol) in methanol. GM2 complexes with Ca²⁺ generating the ion [(GM2-H+)⁻ + Ca²⁺]⁺ m/z = 1,423. (Region a) The GM3 complex with Ca²⁺ generates the ions [(GM3-H⁺)⁻ + Ca²⁺]⁺ m/z = 1,248, 1,262, 1,276, 1,290, and 1,304. Region c, which is not observed in the negative spectrum in A, corresponds to the heterodimer GM2/GM3 ions [(GM2-H⁺)⁻ + (GM3-H⁺(1237))⁻ + 2Ca²⁺]²⁺ m/z = 1,342, 1,350, 1,356, 1,364, and 1,372. Declustering potential was 80 V. (C) MS/MS spectrum of the GM2/GM3 complex [(GM2-H⁺)⁻ + (GM3-H⁺(1237))⁻ + 2Ca²⁺]²⁺ m/z = 1,350. (D) MS/MS spectrum of the GM2/GM3 complex [(GM2-H⁺)⁻ + (GM3-H⁺(1237))⁻ + 2Ca²⁺]²⁺ m/z = 1,364.

Fig. 3.

GM2/GM3 heterodimer defined by mAb 8E11. (A) Specificity of mAb 8E11 and DH2, determined by ELISA. Various concentrations of gangliosides GM2 (rectangles), GM3 (triangles), and GM2/GM3 (circles) in methanol plus 2 mM CaCl₂ were dried into 96-well flat-bottom polystyrene plates at 37°C and washed in TBS containing 2 mM CaCl₂ (TBS+). Binding of mAb 8E11 (1 μg/ml) was determined by ELISA as described in ref. . Effect of EDTA on binding of 8E11 to GM2/GM3 plus 2 mM CaCl₂ was determined by addition of 50 mM EDTA to a mixture of GM2/GM3 in methanol solution and washing in TBS. 8E11 reactivity is shown by the diamonds. (B) Single GSL (1.25 nmol) and GSL mixtures with GM2 or GM3 (0.65 nmol of each) in methanol plus 2 mM CaCl₂ were tested for reactivity with mAb 8E11, as described for determination of IgG3 Ab reactivity (35). Data are presented as mean ± SD. Significance of difference versus GM2/GM3-treated wells: *, P ≤ 0.001. (C) High-performance TLC (HPTLC) patterns of various gangliosides and their combinations. (Upper) Immunostained with mAb 8E11 (a), DH2 (b), or MK1.8 (c), as described in SI Materials and Methods. (Lower) Detected by 0.2% orcinol in 10% H₂SO₄. (D) Reactivity of GM2/GM3 or GM3 expressed on HCV29 cells was analyzed by flow cytometry, using mAb 8E11 or DH2. Cells were pretreated with P4 for 72 h (1 μM) to deplete endogenous GSLs, washed, and incubated overnight with serum-free medium alone (a) or with 25 μM GM2 (b), 25 μM GM3 (c), or GM2/GM3 (12.5 μM each) (d). Cells were released in 0.01% trypsin and 0.1 mM EDTA; incubated with 1 μg/ml normal mouse IgG (a), 1 μg/ml mAb 8E11 (b–d Left), or 1 μg/ml DH2 mAb (b–d Right); and labeled with Alexa Fluor 488-labeled anti-mouse IgG. The experiments shown were performed multiple times with comparable results.

Fig. 4.

Reactivity of mAb 8E11 and its reversing effect on motility inhibition and signaling by GM2/GM3 heterodimer. (A) Fluorescence-activated flow cytometric pattern of HCV29 and YTS-1 cells. (a) IgG control. (b) Cells with 8E11 by the method as in Fig. 3D. (c) Cells released by 50 mM EDTA. Peak 1, negative cells. Peak 2, 8E11 reactive cells. Note that the peak positive for mAb 8E11 (peak 2) is absent in EDTA-treated cells, suggesting that this peak is endogenous GM2/GM3 heterodimer. (B) Antibody effect on inhibition of haptotactic motility of HCV29 cells induced by GM2/GM3-coated silica nanospheres (open columns), compared with noncoated nanospheres (filled column). GM2/GM3-coated nanospheres were incubated for 2 h with medium, mouse IgG (1 μg/ml), normal mouse IgM (1 μg/ml), mAb 8E11 (IgG3) (1 μg/ml), mAb DH2 (IgG3) (1 μg/ml) or mAb MK1.8 (IgM) (1 μg/ml). HCV29 cells (0.5 × 10⁴ per well), treated as above, were suspended in 0.250 ml of RPMI medium 1640 with 5% FBS, containing noncoated nanospheres or nanospheres/antibody suspension, seeded onto gold sol-coated 48-well plates and incubated, and haptotactic motility was analyzed as in Materials and Methods. (C) Reversing effect of mAb 8E11 on inhibition of p-Src phosphorylation of HCV29 cells induced by 2 h of incubation with GM2/GM3-coated nanospheres in medium containing no IgG (column 2), mouse IgG (1 μg/ml) (column 3), or mAb 8E11 (IgG3) (1 μg/ml) (column 4). Cells were incubated overnight at 37°C in 5% CO₂ and lysed, and phosphorylation at position 416 of Src of each cell lysate was measured by Western blot analysis as described in Materials and Methods and compared with lysate from cells incubated with noncoated nanospheres (column 1). Note that only 8E11 restored Src phosphorylation induced by GM2/GM3. (D) Reversing effect of mAb 8E11 on MAPK phosphorylation inhibition by GM2/GM3-coated nanospheres. Cells were treated as in C. Data are expressed as mean ± SD. and analyzed by one-way ANOVA (Dunnett test). Note that only 8E11 restored MAPK phosphorylation induced by GM2/GM3. Significance of difference versus GM2/GM3-coated nanospheres-treated cells: **, *, P ≤ 0.01; *, P ≤ 0.001.

Fig. 5.

Ganglioside expression patterns of ldlD cells expressing GM2 synthase grown under five different conditions, with or without expression of CD82. (A) Ganglioside expression in ldlD cells grown under five different conditions as in Table 1 determined by HPTLC/immunostaining with anti-GM3 mAb DH2 (a) or anti-GM2 MK1.8 (b). (B) Flow cytometric patterns with mAb 8E11. Shown are differences in ldlD/GM2syn cell variants with vs. without CD82 grown under five different conditions (as in Table 1). Procedures are as described in Materials and Methods and Fig. 4A legend. (D) Haptotactic motility of ldlD/GM2syn cell variants with vs. without CD82, grown under five different conditions (as in Table 1). Cells were cultured as above, and motility was determined as in Materials and Methods. Data were expressed as mean ± SD. Significance of differences: *, P ≤ 0.001.