• Open Access

GM3 suppresses anchorage-independent growth via Rho GDP dissociation inhibitor beta in melanoma B16 cells

Authors


4To whom correspondence should be addressed.
E-mail: tcyamagata@gmail.com

Abstract

Ly-GDI, Rho GTPase dissociation inhibitor beta, was found to be expressed parallel to the GM3 level in mouse B16 cells whose GM3 contents were modified by B4galt6 sense, B4galt6 antisense cDNA, or St3galt5 siRNA transfection. Ly-GDI expression was increased on GM3 addition to these cells and decreased with D-PDMP treatment, a glucosylceramide synthesis inhibitor. Suppression of GM3 or Ly-GDI by RNAi was concomitantly associated with an increase in anchorage-independent growth in soft agar. These results clearly indicate that GM3 suppresses anchorage-independent growth through Ly-GDI. GM3 signals regulating Ly-GDI expression was inhibited by LY294002, siRNA against Akt1 and Akt2 and rapamycin, showing that GM3 signals are transduced via the PI3K/Akt/mTOR pathway. Either siRNA towards Rictor or Raptor suppressed Ly-GDI expression. The Raptor siRNA suppressed the effects of GM3 on Ly-GDI expression and Akt phosphorylation at Thr308, suggesting GM3 signals to be transduced to mTOR-Raptor and Akt-Thr308, leading to Ly-GDI stimulation. siRNA targeting Pdpk1 reduced Akt phosphorylation at Thr308 and rendered the cells insensitive to GM3 stimulation, indicating that Akt-Thr308 plays a critical role in the pathway. The components aligned in this pathway showed similar effects on anchorage-independent growth as GM3 and Ly-GDI. Taken together, GM3 signals are transduced in B16 cells through PI3K, Pdpk1, AktThr308 and the mTOR/Raptor pathway, leading to enhanced expression of Ly-GDI mRNA, which in turn suppresses anchorage-independent growth in melanoma B16 cells. (Cancer Sci 2011; 102: 1476–1485)

Ly-GDI, a Rho GTPase dissociation inhibitor beta, is also known as RhoGDI2 or D4-GDI. It belongs to a family of Rho GDP dissociation inhibitors (RhoGDI) including RhoGDI1 and RhoGDI3. The family is named for its ability to inhibit the dissociation of bound GDP from its partner Rho GTPase, which regulates interactions with regulatory guanine nucleotide exchange factors, GTPase-activating proteins, and effector targets.(1) Rho GTPases, such as RhoA, Rac1 and CDC42, are important regulators of cellular migration and adhesion,(2,3) and regulate the dynamics of β-actin, resulting in stress fibers, lamellipodia, and filopodia formation. In contrast to other GDIs, Ly-GDI interacts poorly with RhoA, Rac and Cdc42 in vivo.(4) Although Ly-GDI differs substantially at the structural level from the other two GDIs, RhoGDI-1 and RhoGDI-3, these three GDI exhibit very similar biological regulatory activities on Rho family proteins.(5–7) Ly-GDI is regarded as an invasion and metastasis suppressor gene in human bladder cancer cells.(8,9) For example, the truncated Ly-GDI has been shown to promote metastasis of colon cancer.(10) However, Ly-GDI was reported to be highly expressed in human breast cancer cells and its involvement in tumor cell invasion has been suggested.(11) Thus, function of Ly-GDI is controversial, and furthermore, it remains unknown as to how Ly-GDI is regulated in cells and how its regulation is related to tumor behavior, such as invasion and metastasis.

Mouse melanoma B16 cells are characterized by predominant ganglioside GM3 expression, which has been implicated as a modulator of signal transduction at lipid rafts on the plasma membrane(12) particularly associated with growth factor receptor, such as EGF and FGF receptors, which in turn modulate cell growth.(13,14) Since then, a number of studies have supported the idea that gangliosides inhibit or promote activity of growth factor receptor tyrosine kinase, which in turn regulates cell growth via signal transduction.(15) Specifically, stimulation or suppression of GM3 directly changes the signal transduction in B16 cells.

Recently, GM3 signals were shown to be transferred to downstream molecules via the PI3K pathway. In the human keratinocyte-derived squamous carcinoma cell line, SCC12F2, GM3 depletion stimulated phosphorylation of Akt at the Ser473 and Thr308 sites.(16) Anti-GM3 antibody increased phosphorylation of Akt at Thr308 but not at Ser473,(16) suggesting a key role for Thr308 of Akt in GM3 signaling cascade. These findings are also consistent with the known ability of GM3 to regulate downstream signaling through PI3K by EGF receptor phosphorylation inhibition.(13) In addition, GM3 has the ability to stimulate PTEN expression, a dual specificity phosphotase that antagonizes PI3K/Akt signaling.(17) The PI3K/Akt pathway has also been shown to be involved in GM3 signaling, resulting in the stimulation of TNF-α production in B16 cells.(18,19)

In addition to the effects of GM3 on cell growth,(13,14) reduced expression of GM3 and GM3 synthase as a result of v-Jun transformation resulted in the enhanced ability of mouse fibroblast cells’ anchorage-independent growth, and re-expression of GM3 by introducing the GM3 synthase gene to the transfectants correlated with a reduced ability of the cells to form colonies in nutrient agar.(20) Contrary to this observation, expression of GM3 in 3LL Lewis lung carcinoma cells endowed cells with the ability of anchorage-independent growth.(21) In view of the important role of GM3 in anchorage-independent growth, there is an indirect potential reason to believe in the involvement of Ly-GDI in mediating GM3 signals to B16 cells’ anchorage-independent growth as follows: (i) c-Src has been recently identified to be recruited in the GM3-enriched microdomain on the cell surface to mediate signal transduction in B16 cells,(22) (ii) GM3 indirectly regulates Ly-GDI phosphorylation by modulating activity of c-Src(23) via EGFRTyr845;(24) (iii) Ly-GDI phosphorylation at Tyr153 decreased the amount of Rac1 in the Ly-GDI complex;(23) and (iv) relieving and activating form of Rac1 from Ly-GDI will result in the formation of colony in soft agar.(25) However, there is no direct evidence to show this mechanism. Thus, the effects of GM3 expression on anchorage-independent growth, one of the hallmarks of transformation, which is considered most accurate and stringent in in vitro assay for detecting the malignant transformation of cells, are controversial in different cell lines and the mechanism still remains unknown.

During our investigation of ganglioside functions on cell behavior,(18,19) we obtained cells (CSSH-1) that overexpressed B4galt6 cDNA and cells (CAH-3) that suppressed its expression. In the CSSH-1 cells, GM3 expression was doubled, but in the CAH-3 cells, GM3 expression was halved. GM3-deficient CAH-3 cells showed anchorage-independent growth in soft agar, but neither the vector control cells nor the GM3-enriched cells were able to propagate. To determine whether the anchorage independent growth capacity of CAH-3 cells was due to the deficiency of GM3 on the cell surface, St3gal5-silenced cells were produced by transfecting B16 cells with St3gal5 siRNA. It was found that the introduction of St3gal5 siRNA to B16 cells resulted in the survival and proliferation of the cells in soft agar. This result led us to seek the target molecules through which GM3 exerts its regulation of cell malignancy. Ly-GDI was found to be a candidate for the mediation of cell proliferation in soft agar after gene screening. Thus far, the level of Ly-GDI expression was always parallel to the determined GM3 contents in all of the B16 cells. The following results clearly indicate that GM3 induces Ly-GDI expression via PI3K, Pdpk1, AktThr308 and the mTOR/Raptor pathway, through which it exerts its function in anchorage-independent growth.

Materials and Methods

Cell lines and culture.  Mouse melanoma B16 cells were kindly provided by Dr Kiyoshi Furukawa at Nagaoka University of Technology, Japan. The mouse melanoma cell lines, CSSH-1 and CAH-3, were produced from B16 cells by transfection with a vector containing B4Galt6 cDNA in a sense or antisense direction, respectively, as reported elsewhere.(18,19) The cells were maintained in media containing DMEM (Gibco, Grand Island, NY, USA) supplemented with 10% fetal bovine serum (FBS) (TBD-TianJin Hao Yang Biological Company, TianJin, China), 100 units/mL penicillin, and 100 μg/mL streptomycin, and were incubated in a humidified (37°C, 5% CO2 and 95% air) incubator (Sanyo, Tokyo, Japan).

Chemicals and antibodies.  Ganglioside GM3 from bovine, LY294002 and LY303511 were purchased from Sigma-Aldrich (St Louis, MO, USA). D-PDMP was from Matreya LLC (Pleasant Gap, PA, USA). Rabbit anti-Akt, antiphospho-Akt (Ser473), antiphospho-Akt (Thr308) antibodies and horseradish peroxidase (HRP)-linked anti-rabbit secondary antibody were obtained from Cell Signaling (Danvers, MA, USA). The RNeasy mini kit used to extract total RNA was from Qiagen (Hilden, Germany). The RT-PCR kit was from the Takara Biotechnology Corporation (Dalian, China).

RNA extraction and RT-PCR.  RNA extraction and analysis of amplified DNA have been described previously.(26) The primers used in the present study were designed by Primer 3 software (funded by Howard Hughes Medical Institute and by the National Institutes of Health, http://frodo.wi.mit.edu/primer3/) and synthesized by Invitrogen (Shanghai, China). Primer sequences were as follows: for β-actin, sense: 5′-ACACTGTGTGCCCATCTACGAGG-3′ and antisense: 5′-AGGGGCGG-ACTCGTCGTCATACT-3′; for Pdpk1, sense: 5′-CCCGGTTTTACACGGCTGAG-3′ and antisense: 5′-GAGGCCCGTACCCTTCCATC-3′; sense and antisense primers for St3gal5, Akt1, Akt2, Raptor, Rictor, Ly-GDI and Eef can be found in our previous publications.(18,19,26) RT-PCR was used to semi-quantitatively determine the levels of mRNA of the genes under consideration.(26) Eef or β-actin mRNA served as a control.

Western blot analysis.  2 × 106 cells were lysed in 1 mL sample buffer (0.125 M Tris–HCl, pH 6.8, 4% SDS, 20% glycerine, 5%β-mercaptoethanol, 0.01% bromophenol blue) at 37°C for 30 min and boiled at 100°C for 5 min. An aliquot of the lysate was subjected to SDS-PAGE, transferred to a membrane (Protran BA85; Schleicher and Schuell, Dassel, Germany), and probed with a panel of specific antibodies. All western hybridizations were performed at least in triplicate using a different cell preparation each time.

siRNA and transfection.  Target sequences were designed, synthesized and transfected as previously described.(18,19) Effective siRNA sequences targeting St3gal5, Pdpk1, Akt1, Akt2, Raptor, Rictor and Ly-GDI can be found in our previous publications.(18,19) Transfectants were selected with G418 at an active strength of 1.5 mg/mL 3 days later. The medium with G418 was changed every third day until colonies appeared. The colonies were either pooled (G418-resistant cells) or isolated (siRNA knockdown clones) to establish stable cell lines. The cell lines used in the current investigation include G418-resistant cells and siRNA knockdown clones. Once we found that the corresponding gene had been significantly suppressed in the G418-resistant cells, we would not further screen the siRNA knockdown clones because they are acceptable for anchorage-independent growth experiments.(27,28)

GM3 incorporation, extraction and HPTLC.  B16 cells were incubated with 25 μM GM3 in the absence of serum for 4 h and cultured with media containing 5% serum for an additional 20 h as previously reported.(18,19) After washing three times with PBS(−) and dissociating from dishes with 2.5% trypsin (PBS[−]/0.02% EDTA) solution to remove the gangliosides not inserted in the membrane bilayer,(29) GM3 was extracted according to the method of Svennerholm and Fredman with minor modification.(30) Briefly, treated and untreated cells were extracted once with 1 mL of chloroform/methanol (2:1, v/v) and once with chloroform/isopropanol/methanol (7:11:2, v/v/v) with sonication for 1 h. The supernatants were evaporated at 60°C, and lipid fractions were dissolved in chloroform/methanol (2:1, v/v); the gangliosides equal to 1 × 106 cells were applied to HPTLC and developed in chloroform/methanol/0.25% KCl (5:4:1, v/v/v), and stained with orcinol/sulfuric acid reagent.

Soft agar assay.  The soft agar assay was performed essentially as previously described.(31) In brief, parental B16 control and siRNA-transfected cells growing in monolayer culture were trypsinized and resuspended in 0.33% agar medium containing 10% fetal bovine serum. In triplicate 60 mm dishes, 1.5 mL of the cell suspension containing 104 cells was plated over the bottom 0.5% agar layer and examined immediately after plating to show that there was no cell colony in the dish. Soft agar colonies of B16 cells were monitored every day and scored after 28 days as previously described.

Statistic analysis.  All values are given as mean ± SD of at least three independent experiments and statistical significance is indicated in the figures. * indicates significance < 0.05 with respect to no treatment control. bsl00066 indicates significance < 0.05 with respect to the different treated group.

Results

Suppression of either GM3 or Ly-GDI expression increases anchorage-independent growth.  As B4galt6 has been identified in lactosylceramide synthesis,(32) we obtained cells (CSSH-1) that overexpressed B4galt6 cDNA and cells (CAH-3) that suppressed its expression. In the CSSH-1 cells, GM3 expression doubled, whereas in the CAH-3 cells, GM3 expression was halved (Fig. 1A). As GM3 has been previously reported to be involved in anchorage-independent growth in different cell lines, such as mouse fibroblast cell line C3H 10T1/2, chicken fibroblast cell line DF1 and 3LL Lewis lung carcinoma cells,(20,21) anchorage-independent growth was determined in B4galt6 cDNA sense- and antisense-transfected cells. The results demonstrated that the B4galt6 antisense transfected B16 cells with a low level of GM3 induced remarkable anchorage-independent growth in soft agar compared with the cells with higher GM3 contents (Fig. 1B). This observation might underline the possible involvement of GM3 in the regulation of anchorage-independent growth in soft agar, one of the hallmarks of transformation, which is considered the most accurate and stringent in in vitro assay for detecting malignant transformation of cells. Thus, we were prompted to seek out genes that were upregulated in CSSH-1 cells and downregulated in CAH-3 cells. Using RT-PCR techniques, the Ly-GDI gene was found as one of candidate genes regulated by GM3 (Table S1). As shown in Figure 1(C), expression of Ly-GDI was proportional to GM3 expression of the cells. The following work was performed to confirm that GM3 is regulating cell behavior through Ly-GDI.

Figure 1.

 GM3 is proportional to Ly-GDI expression and inversely related to anchorage-independent growth. B16 and its variants were generated by transfecting B4galt6 cDNA in a sense or antisense direction (CSSH-1, SM-1, CAH-3 and CM-1 are B16 cells transfected with sense B4galt6 cDNA, its vector control, antisense B4galt6 cDNA and its vector control cells, respectively). GM3 content (A), anchorage-independent growth ability (B) and Ly-GDI (C) expression was determined by HPTLC, soft agar or RT-PCR assays.

To determine if regulation of Ly-GDI was dependent on the GM3 level, siRNA targeting GM3 synthase, St3gal5, was transfected into B16 cells to decrease GM3 expression. Suppression of St3gal5 mRNA production was successful, and monoclonal transfectants with St3gal5 siRNA were obtained as described in the Materials and Methods. Two clones, A7 and B11, were analyzed with a control scrambled sequence transfectant. Both clones showed significant reduction of St3gal5 mRNA expression (Fig. 2A) as well as the GM3 level (Fig. 2B), with a concomitant suppression of Ly-GDI (Fig. 2C). These A7 and B11 cells were able to grow in soft agar (Fig. 2D) as CAH-3 cells (Fig. 1B), which demonstrated a decrease in GM3 expression generated by B4galt6 antisense transfection, as shown above (Fig. 1A).

Figure 2.

 Ly-GDI expression is suppressed in GM3-decreased cells, which in turn induce anchorage-independent growth. St3gal5 (A–D) or Ly-GDI (E) knockdown of monoclonal cell lines were obtained by G418 screening as described in the Materials and Methods. St3gal5 mRNA (A), GM3 (B) and Ly-GDI mRNA (C) levels were determined by RT-PCR and HPTLC, respectively. Anchorage-independent growth was determined by soft agar experiments (D,E).

Thus, GM3 might be related to the regulation of anchorage-independent cell growth through Ly-GDI, although it remains unknown if Ly-GDI was really involved in this process. Thus, Ly-GDI-silenced monoclonal cell lines, A10G, A3F and A10C were obtained by introducing Ly-GDI siRNA to B16 cells followed by cultivation in G418, as described in a previous paper.(19) Obviously, any of the Ly-GDI-suppressed cells acquired the ability to grow in soft agar medium (Fig. 2E). These results indicate that GM3 might regulate cell behavior via Ly-GDI at the transcriptional level.

GM3 positively regulates Ly-GDI expression.  To further confirm the notion that GM3 is responsible for Ly-GDI regulation, 25 μM GM3 was added in the culture medium of GM3-depleted cells, such as B11 clone and CAH-3 cells, as well as B16 parental cells. Exogenous GM3 significantly bound to biological membranes following GM3 incubation for 24 h (Fig. 3A, upper panel), which resulted in upregulation of Ly-GDI expression by 30% for any cell types (Fig. 3A, lower panel). Consistent with our results, Garofalo et al.(33) also reported that exogenous GM3 rapidly and significantly bound to biological membranes in lymphocytes by HPTLC experiments. Reciprocally, GM3 depletion was achieved by incubating cells with 12.5 μM D-PDMP up to 3 and 6 days (Fig. 3B, upper panel). Ly-GDI expression was suppressed in association with GM3 depletion (Fig. 3B, lower panel). Taken together, these results demonstrated that the level of GM3 plays a critical role in the expression of Ly-GDI, through which cell behavior such as anchorage-independent growth is regulated.

Figure 3.

 Treatment of cells with GM3 or D-PDMP further confirmed the notion that GM3 positively regulates the expression of Ly-GDI. B16, CAH-3 and B11 cells were incubated with 25 μM GM3 for 24 h and GM3 contents as well as Ly-GDI expression were determined by HPTLC (upper panel) or RT-PCR (lower panel) in (A). In select experiments, B16 and CSSH-1 cells were cultured with 12.5 μM D-PDMP up to 3 or 6 days and GM3 contents as well as Ly-GDI expression were determined by HPTLC (upper panel) or RT-PCR (lower panel) in (B).

GM3 regulates Ly-GDI expression via the PI3K/Akt pathway.  As previously reported, GM3 treatment of diverse B16-derived cells elicited fast activation of Akt,(34) and we further found that GM3 rapidly stimulated Akt phosphorylation within 5 min of GM3 stimulation before returning to basal levels at the 6 h time point (Fig. S1A). In addition, phosphorylation of Akt was inhibited in association with GM3 knockdown (Fig. S1B). These results, along with our previous investigations,(18,19) suggest that GM3 regulates Ly-GDI expression via the PI3K/Akt pathway. When B16 and CSSH-1 cells were incubated with PI3K inhibitors, LY294002 (Fig. 4A) and LY303511 (Fig. 4B), Ly-GDI expression in both cell lines was impaired and Akt phosphorylation at both Ser473 and Thr308 sites were completely inhibited as revealed by western blotting (Fig. 4C). It is noteworthy to mention that the most apparent difference between B16 and CSSH-1 cell lines was a GM3-enriched transfectant obtained from the B16 cell line. However, the difference in GM3 levels did not affect the exertion of the PI3K inhibitors’ function on Ly-GDI, suggesting that PI3K might possibly be located downstream of GM3 (Fig. 4A,B). To verify the hypothesis that GM3 is located upstream of the PI3K pathway leading to Ly-GDI regulation, the effect of GM3 on Ly-GDI mRNA expression was assayed in the presence or absence of PI3K inhibitors. LY294002 (25 μM) or LY303511 (100 μM) treatment thoroughly blocked the effects of GM3 on inducing Ly-GDI expression (Fig. 4D) by inhibiting Akt phosphorylation (Fig. 4E). This observation strongly demonstrates that GM3 positively regulates Ly-GDI expression via the PI3K/Akt pathway, although the involvement of an additional pathway through which GM3 exerts its function could not be fully excluded.

Figure 4.

 GM3 affects Ly-GDI expression through the PI3K pathway. B16 (A,C–E) and CSSH-1 (A,B) cells were treated with 25 μM LY294002 (A,C–E) or with 25 μM and 100 μM LY303511 (B,D,E) for 24 h in the presence (D,E) or absence (A–E) of GM3. Ly-GDI was determined by RT-PCR and Eef served as an internal control (A,B,D). Phosphorylated Akt (Ser473 and Thr308) is shown by immunoblotting using specific Ab. Equal loading in each lane is ensured by the similar intensities of total Akt.

Akt is a key molecule for transducing GM3 signals to Ly-GDI-mediated anchorage-independent growth.  To further assess the role of Akt in Ly-GDI-mediated anchorage-independent growth, B16 cells were transfected with Akt1 or Akt2 siRNA. G418-resistant cells were isolated from B16 transiently transfected cells following 3 days of transfection. Akt1 or Akt2 siRNA effectively reduced mRNA expression of corresponding genes to 50% or 70%, respectively (Fig. 5A). The suppression of Akt1 or Akt2 was shown to decrease Ly-GDI expression by 70% or 50%, respectively (Fig. 5B). As shown in Figure 5(B), suppression of Ly-GDI was more profound in Akt1/2 doubly silenced cells. To further understand the mechanism by which GM3 affects the expression of Ly-GDI via Akt, we determined the effect of GM3 (25 μM) on Akt1-, Akt2- or Akt1/2-silenced cells. Enhancement of Ly-GDI expression was observed via incubating Akt1 or Akt2 knockdown cells with GM3. However, when both Akt1 and Akt2 were concomitantly disrupted by both siRNA, GM3 signals were unable to stimulate Ly-GDI any longer (Fig. 5B).

Figure 5.

 Akt is a key molecule for GM3-stimulated Ly-GDI synthesis, resulting in inhibition of anchorage-independent growth in B16 cells. B16 cells were transfected with Akt1 or Akt2 siRNA to obtain G418-resistant transfectants and either Akt1 or Akt2 expression was assayed by RT-PCR (A). In select experiments, B16 cells were transfected with Akt1, Akt2, or Akt1 plus Akt2 siRNA to screen G418-resistant transfectants, and the cells were incubated in the presence or absence of 25 μM GM3. Ly-GDI was determined by RT-PCR and Eef served as an internal control in (B). Anchorage-independent growth was determined in Akt1 or Akt2 siRNA or both, and Akt-silenced cells (C).

In light of our results showing that GM3 associated with Ly-GDI are able to inhibit anchorage-independent growth and the pivotal role of Akt1/2 in mediating GM3 signals to Ly-GDI expression, we next examined the effects of key signaling molecules, Akt1 and Akt2, on anchorage-independent growth in soft agar medium. As shown in Figure 5(C), the number of colonies in soft agar medium was remarkably induced by Akt1 or Akt2 gene knockdown. In addition, Akt1 and Akt2 were further found as cumulative genes to regulate anchorage-independent growth (Fig. 5C), which demonstrated a similar manner in regulating Ly-GDI mRNA expression as shown above (Fig. 5B). The signals of GM3 would be blocked only after double knockdown of the expression of Akt1 and Akt2, or alternatively they will transduct GM3 signals to Ly-GDI. It is noteworthy to mention that Akt1/2 suppression is also capable of inducing colony formation in CSSH-1 cells, although the relative number of colonies is lower than Akt1/2 knockdown of SM-1 cells for the relative higher contents of GM3 in CSSH-1 cells (Fig. S2). All these results clearly support the notion that Akt is the unique molecule through which GM3 exerts its function in B16 cells.

Involvement of Pdpk1 in GM3 signaling and regulation of Ly-GDI mediated anchorage-independent growth.  Pdpk1 is one of the components recruited in the signaling pathway activated by PIP3 to serve as a multifunctional effector downstream of PI3K. Activated Pdpk1 is shown to phosphorylate Akt at Thr308.(35) Given the important role of Pdpk1 in the PI3K/Akt pathway along with our recent observation showing that Pdpk1 was suppressed in B11 cells indicated that Pdpk1 was involved in GM3 signals to Ly-GDI expression (Fig. 6A). To evaluate the importance of Pdpk1 in the PI3K/Akt pathway, G418-resistant cells were isolated from Pdpk1 transfectants (Fig. 6B). As shown in Figure 6(C), GM3 stimulation of Ly-GDI expression was not observed in the cells treated with Pdpk1 siRNA, suggesting that Pdpk1 is a mediator of GM3 signals. Phosphorylation of Akt was determined by western blotting. siRNA of Pdpk1 severely impaired the level of Akt phosphorylation at Thr308, whereas Akt phosphorylation at Ser473 was not affected. Akt phosphorylation was enhanced by GM3 as shown above, but phosphorylation at either Ser473 or Thr308 was not stimulated by GM3 in the cells treated with Pdpk1 siRNA (Fig. 6D), although we could not decipher the potential reason why phosphorylation at AktSer473 was also insensitive to GM3 treatment in Pdpk1 G418-resistant cells. Similarly, the formation of colonies was remarkably induced in Pdpk1-suppressing cells (Fig. 6E). These results indicate that GM3 signals are controlled by Pdpk1 in terms of Akt phosphorylation at both Thr308 and Ser473.

Figure 6.

 Pdpk1 induces the expression of Ly-GDI through AktThr308 phosphorylation, which in turn inhibits anchorage-independent growth in soft agar. B16 cells were transfected with St3gal5 (A) or Pdpk1 (B–E) siRNA to obtain G418-resistant transfectants and Pdpk1 (A,B), Raptor (A), Rictor (A), as well as Ly-GDI expression (B,C) were determined by RT-PCR. Eef served as an internal control. Phosphorylated Akt (Ser473 and Thr308) is shown by immunoblotting using specific Ab. Equal loading in each lane is ensured by the similar intensities of total Akt (D). Anchorage-independent growth was determined in Pdpk1-silenced cells (E).

mTOR complexes differentially exert their effects on Ly-GDI expression.  Given the reason that mTOR complexes have been elucidated as effectors of PI3K/Akt(19) and Raptor or Rictor was oppositely regulated by GM3 in B11 cells (Fig. 6A), we next assessed the role of mTOR complexes as the potential downstream effectors of Akt to mediate GM3 signals to Ly-GDI expression. Incubation of B16 cells with rapamycin, an inhibitor of mTOR, resulted in suppression of Ly-GDI (Fig. 7A). To reveal the mechanism of the mTOR complexes in detail, G418-resistant cells were isolated from Raptor or Rictor siRNA transfectants as described in the Materials and Methods. As shown in Figure 7(B,C), both siRNA efficiently suppressed the corresponding genes’ expression, with simultaneous suppression of Ly-GDI expression. To further substantiate the notion that mTOR complexes mediate the Akt signals leading to Ly-GDI stimulation, we incubated these siRNA-treated cells with GM3 (25 μM). In the Raptor-silenced cells, but not the Rictor-silenced cells, Ly-GDI mRNA expression was not stimulated in the presence of GM3 (Fig. 7D,E), indicating that the mTOR/Raptor complex plays an important role in GM3 signals. Ly-GDI mRNA expression in the Rictor-silenced cells was lower than that in the control as well (Fig. 7C), but this expression recovered to the level of the control when the cells were incubated with GM3 (Fig. 7E). This result suggests that Rictor is involved in Ly-GDI regulation, but the GM3 signal does not predominantly go through Rictor.

Figure 7.

 Raptor, but not Rictor, mTOR complex mediates Ly-GDI expression via AktThr308, which in turn suppresses anchorage-independent growth in B16 cells. B16 cells were incubated with rapamycin (20 and 40 nM) (A) for 24 h or transfected with either Raptor (B,D,F–H) or Rictor siRNA (C,E). In selected experiments, cells were treated with 25 μM GM3 in the presence or absence of transfection (D–F). Pdpk1 (D), Raptor (B,G), Rictor (C) and Ly-GDI mRNA (A–E) expression was determined by RT-PCR and Eef served as an internal control. Phosphorylated Akt (Ser473 and Thr308) is shown by immunoblotting using specific Ab. Equal loading in each lane is ensured by the similar intensities of total Akt (F). Anchorage-independent growth was determined in Raptor-silenced cells (G).

It is worthwhile to note that the mTOR/Rictor complex was able to phosphorylate Akt at Ser473 in our previous work.(19) GM3 stimulation of the Rictor siRNA-treated cells restored the phosphorylation of Akt at both Ser473 and Thr308 to the level observed in the control B16 cells (Fig. S3A), suggesting that GM3 signals transduced by Akt are not interrupted by Rictor. In line with the mTOR/Rictor complex, we further found that GM3 stimulation of the Raptor siRNA-treated cells restored the phosphorylation of Akt at Ser473 to the level observed in the control B16 cells (Fig. 7F), implying that GM3 signals transduced by Akt phosphorylation at Ser473 are not interrupted by Raptor. Contrary to this result, phosphorylation of Akt at Thr308 was no longer affected by GM3 stimulation, just as Pdpk1-silenced cells had shown in the previous section (Fig. 6D). An additional experiment revealed that Raptor silencing by siRNA decreased Pdpk1 gene expression significantly, explaining why Raptor suppression can affect Akt phosphorylation at Thr308 (Fig. 7G). Taken together, these results provide further compelling evidence that the mTOR/Raptor complex is predominately located downstream of Akt to mediate GM3 signals to Ly-GDI expression.

Ly-GDI is a pivotal molecule to inhibit B16 cell transformation.  Since both Raptor and Rictor siRNA have the ability to inhibit Ly-GDI expression, we assessed the role of mTOR complexes in anchorage-independent growth. Surprisingly, both Rictor and Raptor siRNA-treated cells showed the ability to proliferate in soft agar medium once Ly-GDI was suppressed (Figs 6E,7H and S3B). These results concretely support the notion that Ly-GDI is a pivotal molecule in inhibiting B16 cell growth in soft agar as summarized in Table S2 and the number of colonies is basically correlated with the mRNA level of Ly-GDI. In contrast, the cell proliferating rate was faster in St3gal5, Ly-GDI, Pdpk1 and Raptor siRNA-treated cells compared with that of the control B16 cells under serum-deprived medium (Fig. S4), although they showed a similar proliferating rate under serum-containing medium (data not shown). All these results demonstrated that Ly-GDI is a key molecule in inhibiting B16 cell transformation to malignant cancer.

Discussion

One of the most important oncogenic properties of tumor cells is anchorage-independent growth. In vitro transformed cells and tumor cells can survive and grow without attachment to the ECM, whereas normal adhering cells are fully dependent on integrin ECM-mediated signals for their survival. GM3 was recently identified to indirectly mediate anchorage-independent growth via Ly-GDI.(23,24) On the basis of these studies, we here further revealed that GM3 induces Ly-GDI expression via PI3K, Pdpk1, AktThr308 and the mTOR/Raptor pathway, through which it exerts its function in anchorage-independent growth, leading to the onset of melanoma (Fig. 8).

Figure 8.

 Proposed cascade of signaling events regulating Ly-GDI expression by GM3, which in turn inhibits anchorage-independent growth in B16 cells. GM3 signals are transduced in B16 cells through PI3K, Pdpk1, Akt and the mTOR Raptor pathway, leading to the enhanced expression of Ly-GDI mRNA, which in turn suppresses B16 cell proliferation in soft agar or serum-deprived medium.

Numerous studies suggest that GM3 could inhibit cell proliferation.(13,36,37) Miura et al.(20) concluded that expression of the v-Jun induces transformed cell clones with greatly reduced levels of GM3 and GM3 synthase (St3gal5), which in turn increased the ability of anchorage-independent growth in the mouse fibroblast cell line C3H 10T1/2 and the chicken fibroblast cell line DF1. In contrast, Uemura et al.(21) recently reported that GM3 reconstituted 3LL Lewis lung carcinoma cells formed a significantly higher number of colonies in soft agar compared with mock-transfected cells, indicating that endogenous GM3 is essential for maintaining these fundamental properties of malignant cells. Although GM3 reconstitution increased anchorage-independent growth in 3LL Lewis lung carcinoma cells, all the aforementioned previously published results need to be interpreted with caution, as it is just shown in one specific cloning cell line. To validate our findings, we used B4galt6 antisense and St3gal5 siRNA to decrease GM3 content in B16 cells. These interventions interfere with GM3 contents, which in turn regulate anchorage-independent growth in B16 cells (Figs 1A,2D).

The PI3K pathway has been recently identified to mediate GM3 signals to downstream molecules. Compelling evidence showed that GM3 negatively regulates PI3K expression by the techniques of GM3 depletion or GM3 antibody treatment in SCC12F2 cells.(16) In contrast, Wang et al.(38) concluded that overexpression of GM3 activated PI3K by inhibiting the ERK1/2 signaling pathway via binding with EGFR in uPA-treated SCC12F2 cells. In agreement with these observations, our data demonstrated that inhibition of the PI3K/Akt pathway by pharmacological inhibitors or Akt siRNA intervention thoroughly blocked GM3 signals to activation of Akt as well as Ly-GDI synthesis (Fig. 4). Importantly, although phosphorylation of Thr308 is the key for expression of Ly-GDI, we could not negate the contribution of phosphorylation of Ser473 on Ly-GDI expression by GM3 stimulation. However, on the basis of these findings, only AktThr308 and the mTOR/Raptor complex are predominantly located downstream of PI3K to mediate GM3 biological signals to cell proliferation in soft agar or serum deprived medium in B16 cells.

The biological actions of Ly-GDI in anchorage-independent growth appear to be controversial. For instance, Ly-GDI has been reported to exert its effects on inducing anchorage-independent growth in MDA-MB-231 cells, as evidenced by its ability to induce anoikis under anchorage-independent conditions.(39) In contrast, other studies suggest that Ly-GDI phosphorylation at Tyr153 decreased the amount of Rac1 in the Ly-GDI complex and resulted in relieving Rac1 from inhibition by Ly-GDI,(23) which in turn induced the formation of colony in soft agar.(25) As there is no direct evidence to show the effects of GM3 on anchorage-independent growth in B16 cells, our data suggest that Ly-GDI knockdown increased the ability of B16 cells to propagate in soft agar or serum-deprived medium, which is characteristic of tumor cell transformation (Figs 2E and S4D). In addition, experiments were carried out to show the effects of Ly-GDI on the expression of the Rho family genes. As shown in Figure S5, the results demonstrated that Rho B is significantly affected by Ly-GDI. It is possible that Ly-GDI regulates B16 anchorage-independent growth via Rho B since Rho B has been identified as a suppressor of transformation, invasion and metastasis in B16-F10 cells.(40) These experiments are underway in our laboratory and beyond the range of the current study.

In summary, we have shown that GM3 regulates Ly-GDI mRNA expression via PI3K, Pdpk1, AktThr308 and the mTOR/Raptor pathway, which in turn suppresses cell proliferation in soft agar or serum-deprived medium in melanoma B16 cells. Reconstructing the signaling network regulating GM3 induced Ly-GDI expression in B16 cells might provide insights for developing therapeutic strategies for melanoma.

Acknowledgment

This work was supported by funding from the Mizutani Foundation for Glycoscience 080029.

Disclosure Statement

The authors have no conflict of interest.

Abbreviations
B4galt6

UDP-Gal:betaGlcNAc beta 1,4-galactosyltransferase, polypeptide 6

Eef

Eef1a1 (eukaryotic translation elongation factor 1 alpha 1)

EGF

epidermal growth factor

EGFR

epidermal growth factor receptor

FGF

fibroblast growth factor

GDI

GDP dissociation inhibitor

HPTLC

high-performance TLC

Ly-GDI

Rho GDP dissociation inhibitor beta

mTOR

mammalian target of rapamycin

Pdpk1

3-phosphoinositide dependent protein kinase-1

PI3K

phosphatidylinositol 3-kinase

Raptor

regulatory associated protein of mTOR

Rictor

RPTOR independent companion of mTOR

Src

Rous sarcoma oncogene

SCC12F2

human keratinocyte-derived squamous carcinoma cell line

St3gal5

ST3 beta-galactoside alpha-2,3-sialyltransferase 5

TNF-α

tumor necrosis factor α

uPA

urokinase plasminogen activator

v-Jun

viral oncoprotein Jun

Ancillary