Enhancement of Epidermal Growth Factor Signaling and Activation of Src Kinase by Gangliosides*

In a recent study, inhibition of cellular ganglioside synthesis blocked growth factor-induced fibroblast proliferation. Conversely, enrichment of cell membrane gangliosides by ganglioside preincubation enhanced growth factor-elicited cell proliferation. In the absence of serum and growth factors, NeuNAcα2–3Galβ1–3GalNAcβ1–4(NeuNAcα2–3)Galβ1–4Glcβ1–1Cer (GD1a) acted like a growth factor when cells were pretreated with the ganglioside, stimulating proliferation of normal human dermal fibroblasts and Swiss 3T3 fibroblasts. In contrast, growth inhibition was observed when high concentrations of gangliosides were continuously present in the culture medium during incubation of fibroblasts with growth factors (Li, R., Manela, J., Kong, Y., and Ladisch, S. (2000) J. Biol. Chem. 275, 34213–34223). Here, we investigated the mechanisms whereby gangliosides elicit proliferation-coupled signaling in normal human dermal fibroblasts. Incubation of the fibroblasts with GD1a enhanced epidermal growth factor (EGF) receptor autophosphorylation and Ras and MAPK activation in a dose-dependent manner. Exposure of the cells to GD1a also enhanced the phosphorylation of Elk-1 by the activated MAPK. Brief pretreatment of the cells with PD98059 blocked the enhancing effect of gangliosides on EGF-induced MAPK activation. In the absence of serum and growth factors, GD1a incubation induced phosphorylation of Src kinase, Ras activation, and phosphorylation of MAPK and Elk-1 in a dose-dependent manner. The activation of Src kinase was confirmed by enhanced Src kinase activity. Brief treatment of the cells with PP1 blocked the activation of Src kinase and MAPK. Again, PD98059 treatment inhibited ganglioside-elicited MAPK phosphorylation. Among the gangliosides tested, GD1a, was the most active molecule, whereas lactosylceramide was the least active one, indicating relative structural specificity of the ganglioside action. In conclusion, gangliosides promote fibroblast proliferation through enhancement of growth factor signaling and activation of Src kinase.

Protein-tyrosine kinases are critical components of signaling pathways that control cell proliferation and differentiation. The receptors for most growth factors such as epidermal growth factor (EGF) 1 are transmembrane tyrosine-specific pro-tein kinases. They all share a crucial signaling pathway that runs from the cell surface to the nucleus. The binding of a growth factor to its tyrosine kinase receptor on the cell surface results in receptor trans autophosphorylation and activation of the Ras/Raf/MAPK cascade (1). This cascade passes the signal to the nucleus through phosphorylation of transcription factors that regulate gene expression, eventually leading to DNA synthesis and cell division (2,3). In addition, the activated receptor tyrosine kinase may also activate other intracellular signaling proteins, including phospholipase C␥, phosphatidylinositol 3-kinase, and Src-like non-receptor tyrosine kinase (1). Nonreceptor tyrosine kinases, e.g. Src, have no extracellular or transmembrane domains but possess modular domains that are responsible for subcellular targeting and regulation of catalytic activity (4,5). Both receptor tyrosine kinases and nonreceptor Src family tyrosine kinases have been reported to be located in lipid rafts (5)(6)(7) or caveolae membrane domains (8,9). Thus, one important issue is to investigate how membrane lipids interact with receptor molecules at the cell surface and initiate signaling for cells to proliferate.
Cell surface gangliosides exist in glycosphingolipid-enriched domains and modulate transmembrane signaling (10 -12). Our recent study provides a new perspective on the biological effect of gangliosides on cell proliferation (13). Specific inhibition of cellular ganglioside synthesis that conceivably abolishes ganglioside domain formation blocks growth factor-mediated cell proliferation. Conversely, enrichment of cell membrane gangliosides by ganglioside incubation that conceivably promotes ganglioside domain formation enhances growth factor-elicited cell proliferation. Strikingly, gangliosides themselves exert a growth factor-like effect in the absence of serum and growth factors, enhancing fibroblast proliferation. However, the mechanisms remain to be investigated.
Cell Culture-Normal human dermal fibroblasts (NHDF) were purchased from Clonetics (San Diego, CA) and cultured in fibroblast growth medium (FGM) supplemented with 2% fetal bovine serum (FBS, HyClone, Logan, UT), 0.5 ml of insulin, 0.5 ml of human FGF, and 0.5 ml of GA1000 (Clonetics). For serum-free culture, fibroblast basal medium (FBM) was used. Cells from passages 3-10 were used for this study (13). The culture medium was changed every 3 days. Cell viability was assessed by trypan blue dye exclusion.
Preparation of Cell Lysate-NHDF were seeded at the density of 2 ϫ 10 5 cells/dish (100 ϫ 20 mm; area ϭ 55 cm 2 ) or 1.5 ϫ 10 5 cells/well in six-well plates (area ϭ 9.4 cm 2 ) in FGM with 2% FBS. After reaching ϳ70% of confluence, NHDF were incubated with gangliosides in FGM with 2% FBS for 18 h. After removal of the culture medium by aspiration, the cells were washed twice with serum-free medium and starved overnight (13,16). The cells were then exposed to 2 ng/ml EGF in serum-free FBM for 5 min at 37°C. In some experiments, cells were incubated with gangliosides either in FGM with 2% FBS or in serumfree FBM for 6 or 18 h, washed, and exposed to 2 ng/ml EGF in serum-free FBM without starvation. The cells were immediately washed twice with ice-cold phosphate-buffered saline and lysed. The lysis buffer contained 20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerophosphate, 1 mM Na 3 VO 4 , 1 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride. The lysate was transferred to microcentrifuge tubes, sonicated briefly on ice, and microcentrifuged for 10 min at 4°C. The supernatant (the cell lysate) was adjusted to 1 g/l and used for kinase assays. Proteins were quantified by the Lowry method, using bovine albumin as a standard (17).
EGF Receptor Activation Assays-200 l of cell lysate (ϳ200 g of total protein) was mixed with 100 l of washed Protein G-Sepharose agarose bead slurry (50-l packed beads) and stirred by a rotary shaker for 2 h at 4°C to preclear nonspecific binding. After microcentrifugation at 14,000 ϫ g for 5 s, the supernatant was transferred to a new microcentrifuge tube and mixed with 4 g of anti-EGF receptor antibody (sheep polyclonal IgG). The mixture was incubated at 4°C overnight with stirring. The immune complexes were recovered by adding 100 l of washed Protein G-Sepharose agarose bead slurry (50-l packed beads) and gently rocking the mixture for 2 h at 4°C. After microcentrifugation at 14,000 ϫ g for 5 s and removal of the supernatant, the beads were washed three times with ice-cold lysis buffer, resuspended in 50 l of 1ϫ SDS sample buffer, and boiled for 5 min. After microcentrifugation, 20 l of the supernatant of each sample (ϳ40 g) was loaded onto a 7.5% SDS-polyacrylamide gel. Phosphorylation of EGF receptor was detected by Western blot analysis using an anti-phosphotyrosine antibody p-Tyr (PY99). Alternatively, total cell lysate was used to determine the phosphorylation of EGF receptor, using either a phospho-EGF receptor antibody (Y1173, mouse monoclonal IgG) or an anti-phosphotyrosine antibody p-Tyr (PY99). Total EGF receptor was detected by an anti-EGF receptor antibody.
Phospho-p44/42 MAP Kinase Assays-Phosphorylation of p44/42 MAPK was determined by Western blot analysis, using phospho-p44/42 MAPK antibody. MAPK antibody was used to detect the total MAPK in each sample. MAPK activity was further determined by measuring Elk-1 phosphorylation. In these experiments, phospho-p44/42 MAPK was first immunoprecipitated by phospho-p44/42 MAPK antibody, as described in the protocol of the Bio-Rad MAPK assay kit. Briefly, 200 l of cell lysate (ϳ200 g of total protein) was transferred to a 1-ml microcentrifuge tube, and 15 l of resuspended immobilized phospho-p44/42 MAPK (Thr-202/Tyr-204) monoclonal antibody was added to each tube. The mixture was incubated with gentle rocking overnight at 4°C and microcentrifuged for 30 s at 4°C. The pellet was washed twice with 500 l of lysis buffer and 500 l of kinase buffer. Active MAPK (Erk2, 20 ng/200 l of control cell lysate) was used as a positive control. The pellet prepared above was resuspended in 50 l of kinase buffer with 200 M ATP and 2 g of Elk-1 fusion protein. The resulting suspension was incubated for 30 min at 30°C, and the reaction was terminated with 25 l of SDS sample buffer. Each sample was boiled for 5 min, vortexed, and microcentrifuged for 2 min, subjected to SDS-PAGE electrophoresis and Western blot analysis.
Western Blotting-After transfer, the polyvinylidene fluoride membrane was incubated in 25 ml of blocking buffer for overnight at 4°C. The membrane was incubated with primary phospho-Elk-1 (Ser-383) antibody (rabbit polyclonal IgG, 1:1000 dilution) with gentle agitation overnight at 4°C. After being washed three times, the membrane was incubated with horseradish peroxidase-conjugated anti-rabbit second antibody (1:2000) and proteins were detected using the LumiGLO chemiluminescence reagents. Proteins with different molecular weights were used as standards.
Src Phosphorylation Assay-NHDF were seeded at the density of 2 ϫ 10 5 cells/dish (100 ϫ 20 mm, area ϭ 55 cm 2 ) or 1.5 ϫ 10 5 cells/well in six-well plates (area ϭ 9.4 cm 2 ) in FGM with 2% FBS and were cultured until their density reached ϳ70% confluence. The cells were then incubated with ganglioside in FGM with 2% FBS for 18 h. After removal of the culture medium by aspiration, the cells were washed twice with serum-free medium and starved overnight (13), followed by the lysis of the cells as described above. In some experiments, cells were incubated with ganglioside either in FGM with 2% FBS or in serum-free FBM for 6 or 18 h, washed, and lysed without serum starvation. The phosphorylation of Src was detected by Western blot analysis, using phospho-Src kinase antibody. Total Src was measured by Western blot analysis using c-Src antibody.
Src Kinase Activity Assay-Src kinase activity was determined by measuring the phosphorylation of a Src substrate peptide, using a Src kinase assay kit (Upstate Biotechnology Inc.). Briefly, Src was immunoprecipitated from the cell lysate (ϳ200 g of protein). One-fifth of the immunoprecipitated protein (in 10 l) was transferred to a microcentrifuge tube that contained 10 l of Src substrate peptide (final concentration, 120 M) and 10 l of Src kinase reaction buffer (100 mM Tris-HCl, pH 7.2, 125 mM MgCl 2 , 25 mM MnCl 2 , 2 mM EGTA, 250 M sodium orthovanadate, 2 mM dithiothreitol). After gentle mixing, 10 l of [␥-32 P]ATP (10 Ci) was added to the tube. The mixture was incubated at 30°C for 10 min with agitation, followed by addition of 20 l of 40% trichloroacetic acid and incubation for 5 min at room temperature to precipitate the phosphorylated peptide. 25 l of the mixture was then slowly spotted onto the center of a numbered P81 paper square. The paper squares were washed five times for 5 min with 0.75%phosphoric acid and then once with acetone for 3 min. Finally, the paper squares were transferred to scintillation vials, and 5 ml of scintillation mixture was added for ␥-counting. 32 P incorporated into the substrate peptide was calculated by subtraction of the nonspecific binding of [␥-32 P]ATP and binding of phosphorylated endogenous proteins in the sample extracts (B) from the total count (A). The picomoles of phosphate incorporated into substrate peptide per minute was calculated by the for-mula, (A Ϫ B) ϫ 2.4/(specific radioactivity ϫ 10 min), where specific radioactivity ϭ 2760 cpm/pmol of ATP.
Analysis of Csk-Csk in total cell lysate and plasma membrane was determined by Western blot analysis. Plasma membrane was isolated as previously described (19). NHDF at 90% confluence were washed twice with buffer consisting of 0.25 M sucrose/1 mM EDTA/20 mM Tricine, pH 7.8. The cells were scraped off the dishes in 3 ml of buffer. After being centrifuged at 1400 ϫ g for 5 min, cells from two culture dishes (100 ϫ 20 mm; area ϭ 55 cm 2 ) were resuspended in 1.0 ml of buffer and homogenized in a 2-ml tissue grinder with 20 strokes. The suspension was then transferred to a 1.5 ml-centrifuge tube and centrifuged at 1000 ϫ g for 10 min. The postnuclear supernatant fraction was removed and stored on ice. The pellet from each tube was resuspended in 1.0 ml of buffer, homogenized, and centrifuged at 1000 ϫ g for 10 min again. The two postnuclear supernatant fractions were combined, layered on top of 2.3 ml of 30% Percoll in buffer, and centrifuged at 84,000 ϫ g for 30 min. The plasma membrane fraction was collected using a Pasteur pipette, mixed with 1 ml of buffer in a 1.5-ml centrifuge tube, and centrifuged at 17,530 ϫ g for 20 min. The pellet (plasma membrane) was lysed, and Csk was detected by Western blot analysis, using an anti-Csk antibody.
Data Analysis-The data presented in each figure are the results obtained from a typical experiment. Each experiment was performed at least twice, and essentially the same results were obtained. The optical density of protein bands from Western blot analysis was quantified using Scion image software from Scion Corp. The -fold induction or percent increase was calculated to compare the difference between control and treatment groups.

Enhancement of EGF Receptor-mediated
Signaling by Gangliosides-Ganglioside preincubation caused a dose-dependent enhancement of EGF-induced fibroblast proliferation (13). This finding points to the possibility that gangliosides enhance the EGF receptor-mediated signaling pathway. Thus, in our initial experiments we determined the effect of gangliosides on EGF receptor autophosphorylation in NHDF. We used the same conditions that we previously used (13). Following 18-h incubation of NHDF with ganglioside G D1a in FGM with 2% FBS, the cells were washed and starved in serum-free medium overnight, followed by exposure of the cells to ϮEGF (2 ng/ml) for 5 min in serum-free medium (16).
Under these conditions, G D1a was shown to enhance EGFinduced EGF receptor phosphorylation by three experimental approaches (Fig. 1). In the first approach, the cell lysate was directly used for the detection of EGF receptor phosphorylation by an anti-phosphotyrosine antibody (Fig. 1A). Preincubation of the cells with 10 and 20 M G D1a caused 28 and 56% increase in EGF-induced EGF receptor phosphorylation, respectively ( Fig. 1A), whereas the total EGF receptor level was nearly equal in each sample. The EGF receptor (170 kDa) was identified by Western blot analysis using an anti-EGF receptor antibody, as well as by the standard protein marker (Fig. 1A). Additional separate experiments confirmed the enhancing effect of ganglioside G D1a on EGF receptor phosphorylation (Fig. 1C).
In the second approach, the EGF receptor phosphorylation of the same cell lysate was detected by an anti-phospho-EGF receptor antibody (Fig. 1A). G D1a preincubation clearly enhanced EGF receptor phosphorylation. Although exerting a strong enhancing effect on EGF receptor phosphorylation, gan-glioside G D1a alone did not cause EGF receptor phosphorylation (Fig. 1A).
In the third approach, EGF receptors were first immunoprecipitated from the cell lysate with an anti-EGF receptor antibody. A specific anti-phosphotyrosine antibody, p-Tyr (PY99), was then used to detect EGF receptor phosphorylation (Fig.  1B). Again, ganglioside G D1a enhanced EGF receptor phosphorylation in the tested range of 10 -50 M. The detection of EGF receptor phosphorylation using immunoprecipitated EGF receptors was a more sensitive approach than simply using the total cell lysate. For example, preincubation of NHDF with 10 M G D1a essentially doubled the EGF-stimulated EGF receptor phosphorylation. Thus, ganglioside G D1a enhances EGF-induced EGF receptor phosphorylation.
To determine whether gangliosides affect the time course of EGF receptor activation, we exposed fibroblasts to 2 ng/ml EGF FIG. 1. Enhancement of EGF receptor autophosphorylation by G D1a preincubation. NHDF were seeded at 2 ϫ 10 5 cells in a culture dish (100 ϫ 20 mm) in FGM with 2% FBS. After the cell density reached 70% of confluence, the cells were incubated with G D1a in FGM with 2% FBS for 18 h. after removal of the culture medium by aspiration, the cells were washed twice and starved in serum-free FBM overnight. The cells were then cultured in serum-free medium Ϯ EGF (2 ng/ml) for 5 min and immediately washed with ice-cold phosphate-buffered saline and lysed. In A, 25 g of the total cell lysate was loaded onto each lane and subjected to SDS-PAGE and Western blot analysis. EGF receptor phosphorylation was detected either by an anti-phosphotyrosine antibody or by an anti-phospho-EGF receptor (Y1173) mouse monoclonal IgG 1k . In B, the cell lysate containing 100 g of total protein was immunoprecipitated using an anti-EGF receptor sheep polyclonal IgG. 25 g of the immunocomplex was subjected to SDS-PAGE and Western blot analysis. EGF receptor phosphorylation was detected by an antiphosphotyrosine antibody. In both panels, the total EGF receptor was detected using an anti-EGF receptor sheep polyclonal IgG. The data presented in each panel are the results obtained from a typical experiment. Five separate experiments for A and three separate experiments for B were performed, and similar results were obtained. C, relative optical intensity of the phosphorylated EGF receptor bands from Western blot (A, top). The results represent the mean Ϯ S.D. of densitometric scanning of three gels from three separate experiments.
for 0 -60 min, following G D1a treatment and starvation. As expected, stimulation of the cells with EGF rapidly activated the EGF receptor. Within 5 min, EGF receptor phosphorylation reached the maximal level and, after 60 min, it had returned to the basal level (Fig. 2). When cells were preincubated with G D1a , EGF receptor phosphorylation was clearly enhanced over the entire time course of EGF receptor activation. The greatest enhancing effect of G D1a was observed at the peak of EGF receptor phosphorylation (5 min). Therefore, gangliosides enhance EGF receptor activation rather than alter the time course of EGF receptor activation induced by EGF.
We next asked whether ganglioside incubation activated Ras, which is an important component of EGF receptor-mediated signaling. Ras functions as a molecular switch by cycling between an active GTP-bound state and an inactive GDP-bound state. Upon activation, Ras exchanges GDP for GTP, becomes associated with the N-terminal region of Raf, and causes activation of Raf. In turn, Raf activates MEK and MAPK (22). As expected, exposure of the cells to EGF activated Ras. Incubation of the cells with 10 and 20 M G D1a increased the level of Ras-GTP by 41 and 83%, respectively, whereas the total Ras remained at the same level (Fig. 3A). Additional separate experiments confirmed the enhancing effect of G D1a on Ras-GTP level (Fig. 3C).
Because MAPK plays a central role in cell proliferation (23), we examined the effect of G D1a on activation MAPK (p44/p42 MAPK or Erk1/Erk2). In these studies, we used a phospho-p44/42 MAPK antibody and a p44/42 MAPK antibody. The phospho-MAPK antibody recognizes only the phosphorylated MAPK, whereas the MAPK antibody recognizes total MAPK (both phosphorylated and non-phosphorylated MAPK). Thus, the MAPK antibody was used to detect the total MAPK in each sample for Western blot analysis. Using this method, we detected the enhanced phosphorylation of MAPK by G D1a in a dose-dependent manner (Fig. 4) under the same experimental conditions as described above. Exposure of NHDF to 10, 20, and 50 M G D1a caused a 19%, 55%, and 65% increase in MAPK phosphorylation.
To further assess the effect of gangliosides on MAPK activity, we measured the phosphorylation of Elk-1, a transcription factor and a substrate of MAPK. Following G D1a ganglioside incubation, the cells were lysed and immunoprecipitated using a specific anti-phospho-p44/42 MAPK antibody. The activity of MAPK was determined by measuring the phosphorylation of Elk-1. As shown in Fig. 4, G D1a enhanced the Elk-1 phosphorylation induced by EGF. For example, 20 M G D1a increased the phosphorylation of Elk-1 by 68%.
Brief pretreatment of the cells with PD98059, a specific inhibitor of MEK (14), blocked MAPK activation by EGF, and the enhancing effect of ganglioside G D1a on EGF-induced MAPK activation was also diminished (Fig. 5). This suggests that G D1a promotes fibroblast proliferation at least in part through enhancement of the EGF receptor-mediated Ras/Raf/ MAPK signaling pathway.
To investigate the influence of ganglioside structure on EGFinduced signaling, we tested four highly purified gangliosides: one monosialoganglioside, G M3 ; two disialogangliosides, G D1a and G D3 ; and one trisialoganglioside, G T1b . One neutral glycosphingolipid, lactosylceramide, was also tested. As shown in Fig. 6, NHDF contain G M3 and G D3 as their major ganglioside components. G D1a and G T1b are present only in a trace amount. Among the gangliosides tested, G D1a was the most active molecule in enhancing MAPK activation. On the other hand, lactosylceramide, a neutral glycosphingolipid was the least active one (Fig. 7), indicating relative structural specificity of the ganglioside effect on MAPK activation.
All the experiments above were performed under the same conditions. That is, NHDF at 70% of confluence was incubated with ganglioside G D1a in FGM with 2% FBS for 18 h, washed, and starved in serum-free medium overnight, then the cells were exposed to 2 ng/ml EGF for 5 min in serum-free medium. To investigate whether the results obtained from these exper- iments are culture condition-dependent, we further assessed the potential influence of serum or serum starvation on the enhancing effect of G D1a on EGF receptor activation. In these experiments, cells were incubated with G D1a in either FGM with 2% FBS or in serum-free medium for 18 h, washed, and immediately exposed to EGF for 5 min without serum starva-tion. As shown in Fig. 8, following incubation of NHDF with ganglioside G D1a in FGM with 2% FBS for 18 h, EGF receptor phosphorylation and MAPK phosphorylation were strikingly enhanced by ganglioside G D1a in a clear dose-dependent manner. Exposure of the cells to 10 and 20 M G D1a caused 1-and 3-fold enhancement in EGF receptor phosphorylation, and 22 and 52% enhancement in MAPK phosphorylation. Similarly, phosphorylation of the EGF receptor and MAPK was also enhanced when NHDF were incubated with G D1a in serum-free medium for 18 h (Fig. 8). Finally, when exposed to G D1a for 6 h instead of 18 h, EGF-induced EGF receptor phosphorylation and MAPK phosphorylation were also enhanced, albeit to a lesser degree, suggesting a time-dependent ganglioside effect. For example, exposure of the cells to 20 M G D1a in serum-free medium caused 21% increase in MAPK phosphorylation (not shown). Together, these results clearly demonstrate that exogenous addition of gangliosides under various conditions en- FIG. 4. The enhancing effect of G D1 preincubation on EGFinduced MAPK activation. NHDF were cultured and treated, as described in Fig. 1. 20 g of the cell lysate was subjected to SDS-PAGE and Western blot analysis, using either phospho-p44/42 MAPK or control MAPK antibodies. In addition, the MAPK activity was determined by measuring Elk-1 phosphorylation. The cell lysate containing 200 g of total protein for each sample was immunoprecipitated using a specific anti-phospho-p44/42 MAPK antibody. The kinase assay was performed using Elk-1 as the substrate. One-third of the total product was subjected to SDS-PAGE and Western blot analysis to visualize the phosphorylated Elk-1 bands, using an anti-phospho-Elk-1 antibody. The data are the results obtained from a representative experiment. Six separate experiments were performed, and similar results were obtained.

FIG. 5. Inhibition of MAPK activation by PD98059.
NHDF were seeded at 2 ϫ 10 5 cells in six-well plates in FGM with 2% FBS and cultured until the cell density reached 70% of confluence. The cells were then cultured in FGM with 2% Ϯ G D1a for 18 h, the medium was then removed, and the cells were starved in serum-free medium overnight. During the final 3 h, PD98059 (dissolved in Me 2 SO) was added to the culture medium. After washing twice with serum-free medium, the cells were exposed to EGF (2 ng/ml) in serum-free medium for 5 min. 20 g of the total cell lysate was subjected to SDS-PAGE and Western blot analysis, using either phospho-p44/42 MAPK or MAPK antibodies. The lysate of cells cultured without G D1a incubation and EGF exposure was used as a negative control (NC), whereas the lysate of cells cultured with G D1a and EGF (without Me 2 SO) was used as a positive control (PC). The experiment was repeated once and similar results were obtained.

FIG. 6. HPTLC analysis of gangliosides isolated from NHDF.
Total cellular gangliosides were purified from NHDF cultured in FGM with 2% FBS, and stained as purple bands on the HPTLC plate with resorcinol hydrochloric acid reagent. HBG, human brain gangliosides (4 nmol of lipid bound sialic acid) used as standard. Both G M3 and G D3 migrated as double bands on the HPTLC plate due to the ceramide heterogeneity.
FIG. 7. Influence of ganglioside structure on activity of gangliosides in enhancing MAPK phosphorylation. NHDF were seeded in six-well plates in FGM with 2% FBS. When the cell density reached 70% of confluence, the cells were incubated with 20 M G M3 , G D1a , G D3 , G T1b , or lactosylceramide (Lac-cer) for 18 h. The medium was then removed, and the cells were starved in serum-free medium overnight. Finally, the cells were cultured in serum-free medium Ϯ EGF (2 ng/ml) for 5 min. 20 g of the total cell lysate was subjected to SDS-PAGE and Western blot analysis, using either phospho-p44/42 MAPK or MAPK antibodies. The data presented are the results obtained from a typical experiment. Four separate experiments were performed, and similar results were obtained. hances EGF receptor-mediated proliferation-coupled signaling.
Although preincubation of cells with gangliosides enhanced EGF-dependent proliferation, earlier studies in the literature show high concentrations of gangliosides inhibited cell proliferation when continuously present in the culture medium during incubation of fibroblasts with growth factors (24 -31). Thus, we further evaluated the effect of the continuing presence of ganglioside G D1a in the culture medium during incubation with EGF on MAPK phosphorylation. In this experiment, G D1a was present in the culture medium throughout the assay. Consistent with our proliferation data (13), a low concentration (5 M) caused an increase in the phosphorylation of MAPK when compared with the level of MAPK phosphorylation stimulated by EGF alone (Fig. 9). On the other hand, a high concentration of G D1a (20 M) inhibited EGF-induced MAPK phosphorylation (Fig. 9), which is in accordance with the literature (30).
Activation of Src Kinase by Gangliosides-In our recent study (13), we observed that gangliosides enhanced cell proliferation in the absence of serum and growth factors. Here, we investigated how gangliosides themselves elicit proliferation signaling. Because Src kinase is located on the cytoplasmic side of the plasma membrane, and because it has been shown that Src kinase exists in glycosphingolipid-enriched domains (5, 7, 12), we determined the role of Src kinase in ganglioside-induced cell proliferation.
Following 18-h incubation of NHDF with G D1a in FGM with 2% FBS, the cells were washed and starved in serum-free medium overnight, followed by the lysis of the cells. Under these conditions, G D1a induced Src kinase phosphorylation in a dose-dependent manner (Fig. 10, A and D). Western blot analysis using an anti-Src antibody indicates that the total Src (60 kDa) in each sample was equal. Because activation of EGF receptors also activates Src kinase (1), there is the possibility that Src activation was due to the activation of EGF receptors by gangliosides. However, despite the enhancing effect on EGFinduced EGF receptor phosphorylation, G D1a alone did not cause activation of the EGF receptor (Fig. 1A). In addition, brief treatment of the cells with AG1478 (an EGF receptor inhibitor) did not block Src kinase phosphorylation induced by G D1a (Fig. 11), suggesting that G D1a activates Src kinase via another pathway.
Because of the importance of Ras in normal cell proliferation, we determined whether activation of Src by G D1a acts through a Ras-dependent pathway. As shown in Fig. 3B, in the absence of EGF, G D1a exposure clearly activated Ras. Whereas the total Ras remained at the same level, 20 M G D1a caused a striking increase (119%) in the level of Ras-GTP, which is nearly equivalent to the effect of 2 ng/ml EGF (Figs. 3, B and D). G D1a preincubation also activated MAPK in a dose-dependent manner. As shown in Fig. 10B, exposure of the cells to 10, 20, and 50 M G D1a induced 27-, 39-, and 74-fold stimulation in MAPK phosphorylation, respectively. In this experiment, 50 M G D1a had an effect equivalent to ϳ80% of the effect of 2 ng/ml EGF (Fig. 10B). The MAPK phosphorylation, in turn, resulted in Elk-1 phosphorylation (Fig. 10C). For example, exposure of the cells to 20 and 50 M G D1a induced 10-and 23-fold stimulation in Elk-1 phosphorylation. Pretreatment of the cells with PD98059 for 3 h effectively blocked MAPK activation (Fig. 12). These results suggest that gangliosides, in the absence of serum, activate the Src3 Ras3 MEK3 MAPK pathway.
To determine whether the structure of gangliosides influences the activation of Src kinase and MAPK, we tested four highly purified gangliosides: G M3 , G D1a , G D3 , and G T1b . Lactosylceramide, a neutral glycosphingolipid, was also tested. At 20 M, all four gangliosides were active in stimulating Src kinase phosphorylation, with G D1a and G M3 being the most active molecules (Fig. 13A). Lactosylceramide, on the other hand, was inactive, suggesting that the intact ganglioside structure is important in activating Src kinase. In accordance with these observations, G D1a , G M3 , and G D3 were also active in enhancing MAPK activation, whereas lactosylceramide was not active (Fig. 13B), again indicating the relative specificity of the ganglioside effect.
In the above experiments to assess the ganglioside effect on FIG. 8. Enhancement of EGF-induced EGF receptor autophosphorylation and MAPK phosphorylation by G D1a . NHDF were seeded at 2 ϫ 10 5 cells in culture dishes (100 ϫ 20 mm) and cultured in FGM with 2% FBS until their density reached ϳ70% of confluence. The cells were then cultured in either serum-free FBM or FGM with 2% FBS, with or without G D1a . After 18 h, the medium was removed and the cells were washed, and immediately exposed to 2 ng/ml EGF in serum-free medium for 5 min. 25 g of the total cell lysate was subjected to SDS-PAGE and Western blot analysis. EGF receptor phosphorylation was detected by a phosphotyrosine-specific antibody, whereas the total EGF receptor was detected using an anti-EGF receptor sheep polyclonal IgG. For determination of MAPK phosphorylation status, 20 g of the total cell lysate was subjected to SDS-PAGE and Western blot analysis, using either phospho-p44/42 MAPK or MAPK antibodies. The data presented are the results obtained from a typical experiment. Four separate experiments were performed, and similar results were obtained. FIG. 9. Effect of G D1a continuous incubation on MAPK phosphorylation. NHDF were seeded in a six-well plate in FGM with 2% FBS. When the cell density reached 70% of confluence, the cells were incubated with G D1a in FGM with 2% FBS for 18 h. After the medium was removed, the cells were cultured in serum-free medium Ϯ G D1a overnight. Finally, the cells were cultured in serum-free medium Ϯ G D1a Ϯ EGF (2 ng/ml) for 5 min. 20 g of the total cell lysate was subjected to SDS-PAGE and Western blot analysis, using either phospho-p44/42 MAPK or MAPK antibodies. The experiment was repeated once, and similar results were obtained.
Src kinase activation, NHDF were incubated with gangliosides in FGM with 2% FBS for 18 h, washed, and starved in serumfree medium overnight. To investigate whether the induction of Src kinase activation by gangliosides is culture condition-dependent, we examined the potential influence of serum or serum starvation on ganglioside-induced Src kinase activation. In these experiments, NHDF were incubated with ganglioside G D1a in either FGM with 2% FBS or in serum-free FBM for 18 h, washed, and immediately lysed. As shown in Fig. 14, whether serum (2% FBS) was present or not during the incubation of cells with G D1a , phosphorylation of Src kinase and MAPK was strikingly enhanced in a clear dose-dependent man-ner. For example, when NHDF were incubated with G D1a in serum-free medium for 18 h and without serum starvation, 10 and 20 M G D1a induced an 11-and 12-fold increase phosphorylation of Src kinase, and a 33-and 68-fold increase in phosphorylation of MAPK, respectively. The stimulation of Src and MAPK by G D1a was independent of the specific assay procedures used. Together, these results demonstrate that exogenous addition of gangliosides in the culture medium induces proliferation-coupled signaling via Src kinase and MAPK activation.
The prolonged activation of Src kinase caused by ganglioside treatment is unusual. Thus, we further investigated the activation of Src kinase using three approaches: an Src kinase inhibitor, measurement of plasma membrane-bound Csk, and assessment of Src kinase activity. When PP1, a Src kinase inhibitor (15), was added to the culture medium during the final 3 h of the ganglioside incubation, phosphorylation of both Src kinase and MAPK (Fig. 15) was clearly reduced at 1 M PP1 and abolished at 50 M PP1. The inhibition of Src kinase by PP1 confirmed that gangliosides act to stimulate Src kinase.
It is known that when Src is activated, a Csk-binding protein, which is anchored in the membrane, binds to Csk and FIG. 10. Activation of Src kinase and MAPK by G D1a . NHDF were seeded at 2 ϫ 10 5 cells in culture dishes (100 ϫ 20 mm) and were cultured in FGM with 2% FBS until their density reached ϳ70% of confluence. The cells were then cultured in FGM with 2% FBS medium Ϯ G D1a . After 18 h, the medium was removed, and the cells were washed and cultured in serum-free medium overnight, followed by the lysis of the cells to determine Src (A) and MAPK activation (B and C). 20 g of the total cell lysate was subjected to SDS-PAGE and Western blot analysis, using either a phospho-Src kinase antibody or a c-Src antibody (A). For determination of MAPK phosphorylation status (B), 20 g of the total cell lysate was subjected to SDS-PAGE and Western blot analysis, using either phospho-p44/42 MAPK or MAPK antibodies. The MAPK activity was determined by measuring Elk-1 phosphorylation (C). The cell lysate containing 200 g of total protein for each sample was immunoprecipitated using a specific anti-phospho-p44/42 MAPK antibody. The kinase assay was performed using Elk-1 as the substrate. One-third of the total reaction product was subjected to SDS-PAGE and Western blot analysis to visualize the phosphorylated Elk-1 bands, using anti-phospho-Elk-1 antibody. The lysate of cells stimulated by EGF was used as a positive control. A phosphorylated Elk-1 protein was used as a standard. The data presented in each panel are the results obtained from a typical experiment. Five separate experiments for A, four separate experiments for B, and two separate experiments for C were performed, respectively, and similar results were obtained. D, relative optical intensity of the phosphorylated Src bands from Western blot analysis (A, top). The results represent the mean Ϯ S.D. of densitometric scanning of four gels from four separate experiments.
FIG. 11. Lack of an effect of AG1478 on G D1a -induced Src kinase phosphorylation. NHDF at 70% of confluence were washed and incubated in serum-free medium Ϯ 20 M G D1a for 18 h. During the final 3 h, 1 M AG1478, 10 M PP1, or Me 2 SO (the solvent vehicle) were added to the culture medium. After washing the cells twice with serumfree medium, the cells were immediately lysed to determine the kinase phosphorylation. The data presented are the results obtained from a typical experiment. Three separate experiments were performed, and similar results were obtained.
FIG. 12. Inhibition of MAPK activation by PD98059. NHDF were seeded at 2 ϫ 10 5 cells in six-well plates in FGM with 2% FBS and cultured until they reached 70% of confluence. The cells were then cultured in fresh FGM with 2% FBS Ϯ 20 M G D1a for 18 h. After the medium was removed, the cells were starved in serum-free medium overnight. During the final 3 h, PD98059 was added to the culture medium. After washing twice with serum-free medium, the cells were lysed to determine the MAPK phosphorylation. The data are the results obtained from one of duplicate experiments.
brings Csk to the membrane from the cytosol. The recruited Csk then adds an inhibitory phosphate to Src kinase and removes an activating phosphate (5). To investigate whether ganglioside incubation results in any changes in the level of Csk, we determined the Csk level in total cell lysate and plasma membrane. Following 18-h incubation with 0 -50 M G D1a and overnight serum starvation, the level of total cellular Csk remained constant, whereas the plasma membrane-bound Csk level increased (Fig. 16).
To confirm the enhancement of Src kinase phosphorylation detected by Western blot analysis, we evaluated Src kinase activity by measuring the phosphorylation of a substrate peptide. NHDF were cultured in FGM containing 2% FBS medium and 0 -50 M G D1a for 18 h and starved in serum-free medium overnight, followed by the determination of Src kinase activity. G D1a preincubation resulted in a significant increase in Src kinase activity (Fig. 17). For example, the activity following 20 M G D1a preincubation was nearly double that of the control or basal Src kinase activity (3.2 versus 1.8 pmol of phosphate incorporated into the substrate peptide/min/100 g of total cellular protein; p ϭ 0.038). This confirmed the results of Western blot analysis (Figs. 10 and 13).
Finally, we determined the time course of Src activation in response to EGF and G D1a in four parallel experiments. In the first experiment (Fig. 18A), NHDF were preincubated with 20 M G D1a in FGM with 2% FBS for 18 h, starved in serum-free medium for 18 h, and stimulated with 2 ng/ml EGF for 5 min. The cells were then washed and further cultured in serum-free medium for up to 6 h, during which the Src activity was determined. As expected, stimulation of the cells with EGF rapidly activated Src kinase. Within 5 min, Src activity reached the maximal level, and after 30 min, it had returned to the basal level (Fig. 18A). In the second experiment (Fig. 18B), the experimental approach was the same as that of the first experiment, except NHDF was stimulated with a second dose of 20 M G D1a for 1 h instead of being stimulated by EGF. As shown in Fig. 18B, Src activity was highly elevated (indicated by "a") following G D1a preincubation and 18-h starvation. Upon stimulation by a second dose of G D1a , Src activity further increased to a higher level (indicated by "b"). Src remained activated at the higher level over the 6-h period of serum-free culture. In the third experiment (Fig. 18C), NHDF were incubated with 20 M G D1a in FGM with 2% FBS for 18 h, followed by incubation of the cells with 20 M G D1a in serum-free medium for 18 h. The cells were then washed and cultured in serum-free medium for up to 6 h, during which Src activity was measured. The prolonged incubation of NHDF with G D1a resulted in an increase in Src activity in a time-dependent manner (Fig. 18C). Similarly, 6-h incubation of the cells with 20 M G D1a in serum-free medium in experiment 4 also caused a time-dependent increase FIG. 13. Influence of ganglioside structure on activity of gangliosides in inducing Src and MAPK phosphorylation. NHDF at 70% of confluence were incubated in FGM with 2% FBS containing 20 M G M3 , G D1a , G D3 , G T1b , or lactosylceramide (Lac-cer). After overnight incubation, the cells were starved in serum-free medium for 18 h, followed by the lysis of the cells to determine Src and MAPK phosphorylation. The data presented in each panel are the results obtained from a typical experiment. Three separate experiments were performed, and similar results were obtained.
FIG. 14. Dose-dependent stimulation of Src and MAPK phosphorylation by G D1a . NHDF were seeded at 2 ϫ 10 5 cells in culture dishes (100 ϫ 20 mm) and were cultured in FGM with 2% FBS until their density reached ϳ70% confluence. The cells were washed twice with warm serum-free medium to remove serum and then cultured in either serum-free FBM or FGM with 2% FBS, with or without G D1a for 18 h. After the medium was removed by aspiration, the cells were washed and lysed. Phosphorylation of Src and MAPK was determined, as described in Fig. 10. The data presented in each panel are the results obtained from a typical experiment. Three separate experiments were performed, and similar results were obtained.
in Src activity (Fig. 18D). Together, these studies show that the response of Src to G D1a is different from that to EGF; EGF causes a transient stimulation, whereas G D1a causes a sustained level of activation. DISCUSSION Our recent study has delineated a stimulatory role of membrane gangliosides in fibroblast proliferation (13). Inhibition of cellular ganglioside synthesis blocked growth factor-induced fibroblast proliferation. Conversely, enrichment of cell membrane gangliosides by ganglioside preincubation enhanced growth factor-elicited cell proliferation. In the absence of serum and growth factors, G D1a acted like a growth factor when cells were pretreated with the ganglioside, stimulating proliferation of fibroblasts. Here, we demonstrate two independent mechanisms whereby gangliosides elicit proliferation-coupled signaling in normal human dermal fibroblasts. First, gangliosides enhance EGF receptor-mediated signaling. Incubation of the fibroblasts with G D1a enhanced EGF receptor autophosphorylation and Ras/MAPK activation in a dose-dependent manner. Although enhancing EGF receptor phosphorylation, G D1a itself did not cause EGF receptor phosphorylation. Pretreatment of the cells with PD98059 blocked the enhancement of gangliosides on EGF-induced MAPK activation. Second, gangliosides stimulate Src kinase activity. In the absence of serum and growth factors, ganglioside G D1a incubation induced the phosphorylation of Src kinase, Ras activation, and phosphorylation of MAPK in a dose-dependent manner. Brief treatment of the cells with PP1 blocked the activation of Src kinase and MAPK. Again, PD98059 treatment inhibited ganglioside-elicited MAPK phosphorylation. Thus, gangliosides enhance growth factor signaling and activate Src kinase. In addition, the apparent paradox regarding the effect of gangliosides on cell proliferation was resolved in this study. That is, consistent with our proliferation data (13), when continuously present in the culture medium during incubation of fibroblasts with EGF, a low concentration of G D1a (5 M) caused an increase in the phosphorylation of MAPK. In contrast, when continuously present in the culture medium during incubation of fibroblasts with EGF, a high concentration of G D1a (20 M) inhibited EGFinduced MAPK phosphorylation. This observation is consistent with the literature that high concentrations of gangliosides inhibited cell growth when continuously present in the culture medium during incubation of cells with growth factors (24 -31). A study by Saqr et al. (32) reports that G M1 , G D1a , and G T1b stimulate DNA synthesis in human glioma U-1242 cells and the ganglioside effect on DNA synthesis was more prominent in quiescent, confluent cells than in quiescent, sparse cells. The authors suggest that the exogenous ganglioside effect depends on state of confluency (cell contact) and that gangliosides have a bimodal effect on DNA synthesis (32).
How do gangliosides enhance EGF receptor-mediated signaling and stimulate non-receptor Src kinase activity? The ceramide portion of gangliosides anchors the molecules in the plasma membrane, and gangliosides themselves are generally believed to be devoid of enzymatic features. Hence, gangliosides transduce signals possibly by association with growth factor receptors. Presumably, aggregation of growth factor receptors with gangliosides in the membrane domains brings the growth factor signaling to a threshold of activity and optimal signaling and thus promotes receptor dimerization and phosphorylation, leading to enhanced MAPK activation and cell proliferation. Although this study focused on EGF receptors, the conclusion here is likely applicable to other proliferation-coupled receptor tyrosine kinases, such as the platelet-derived growth factor receptor and fibroblast growth factor receptor. How can gangliosides stimulate Src kinase? One possibility is that gangliosides inserted in the plasma membrane interact with the nonreceptor Src tyrosine kinase that is located in the inner leaflet via N-terminal lipid modification (palmitoylation and myristoylation). Another possibility is that ganglioside molecules inserted in the plasma membrane interact with Src kinase negative regulators, removing inhibition of Src kinase and consequently causing activation of Src kinase (5). In Neuro2a cells, the Csk/c-Src ratio was decreased in glycosphingolipid-en-riched domains following G M3 treatment (12). What is found here is an increase in the level of plasma membrane-bound Csk. Why ganglioside incubation increased membrane Csk and how such an increase correlates with the prolonged activation of Src kinase are unknown. Future studies are needed to address these issues.
Lipid membrane domains have been implicated in signal transduction. Glycosphingolipids, sphingomyelin, and cholesterol cluster together to form lipid microdomains within the fluid lipid bilayer (6,7,33,34). Glycosylphosphatidylinositolanchored proteins and other signaling molecules are enriched in these microdomains (35). Upon cross-linking, B cell receptors rapidly translocate into ganglioside G M1 -enriched domains (36,37). T cell receptor cross-linking causes aggregation of raft-associated proteins, which in turn promotes tyrosine phosphorylation and signaling protein recruitment (38). Targeting of T cell receptor signaling molecules to glycosphingolipid-enriched microdomains is considered critical for T cell activation (39). Incubation of the cells with gangliosides causes insertion of the molecules in the plasma membrane (40,41), conceivably enhances the formation of lipid microdomains, and promotes EGF-receptor mediated signaling, as observed in the present study. Conversely, depletion or reduction of membrane gangliosides (13,42) may adversely affect the ability of EGF receptor to properly localize to lipid microdomains before stimulation or to exit the microdomains upon stimulation with EGF.
Recently, the concept of "glycosignaling domains" has been proposed, to emphasize that clustered glycosphingolipids themselves may initiate signal transduction through the interaction of their lipid portion with aliphatic chains of transducer molecules (7,10). Association of Src family tyrosine kinase Lyn with ganglioside G D3 was reported in rat brain (43). Subsequently, the glycosphingolipid-containing membrane domains isolated from mouse melanoma B16 cells (7) and mouse neuroblastoma Neuro2a cells (12) were found to contain c-Src. In Neuro2a cells that are responsive to gangliosides but not to nerve growth factor, addition of G M3 to the cells caused enhanced c-Src phosphorylation and MAPK activation leading to neuritogenesis. Brief stimulation of isolated glycosphingolipid-enriched domains by G M3 also resulted in enhanced c-Src phosphorylation, suggesting that ganglioside induction of neuritogenesis in Neuro2a cells is mediated by glycosphingolipid-enriched domains (12). In the present study, ganglioside G D1a and other gangliosides, G M3 , G D3 , and G T1b , were found to activate Src, Ras, and MAPK, leading to enhanced fibroblast proliferation (13). These findings clearly provide evidence on the role of gangliosides in proliferation-coupled signal transduction.
In conclusion, gangliosides elicit proliferation-coupled signaling through enhancement of growth factor signaling and activation of Src kinase. The delineation of ganglioside-mediated signaling may provide a new approach to regulate cell proliferation. Further studies are needed to determine how gangliosides interact with the growth factor receptor in the plasma membrane and how gangliosides activate Src kinase, with the goal of elucidating the biological functions of gangliosides in cell proliferation and signal transduction.