Figure 4. Structural comparison of SK1 and SK2. Homology models for the catalytic domain of human SK1 and SK2 were created as described in the Materials and Methods section. A) The ribbon model for SK1 is shown in blue and SK2 in yellow. The expanded section highlights the extended loop found in the lipid binding domain of SK2, and bars indicate the distance between residues of SK1 and SK2. B) The nucleotide and lipid binding domains of SK1 (left) and SK2 (right) were compared. Top Row: residues in the surface images of the catalytic domain are colored brown for hydrophobic and blue for hydrophilic. ADP is indicated by the ball-and-stick model, and S1P by the space-filling model. Middle Row: interaction diagrams of ADP binding to SK. Bottom Row: interaction diagrams of S1P binding to SK. for both SK1 and SK2 by ,4-fold, which may indicate an attempt to compensate for inhibition of both SK1 and SK2. In the case of SK1, proteosomal degradation of the protein caused by SKI-II might trigger the increase of mRNA to compensate [42,43]. Neither ABC294735 nor CB5468139 substantially altered the expression of message for either SK1 or SK2. ABC294640treatment strongly increased SK2 mRNA levels suggesting attempted compensation. Also, consistent with the results of SK2-knockdown in A498 cells [17], inhibition of SK2 activity by ABC294640 dramatically increased SK1 mRNA expression by approximately 5-fold.
Sphingolipid metabolism. Because the conversion of sphingosine to S1P by SK is only one step in the dynamic metabolism of sphingolipids [44], an altered flux through SKs may modulate the levels of several sphingolipids. Therefore, we examined the effects of the SK inhibitors on sphingolipids by mass spectrometry (Figure 5C). Treatment with DMS had little impact on the levels of most ceramides, supporting the hypothesis that its effects are mediated by other target proteins. As expected, the SK1/2-dual inhibitor (SKI-II) and SK1-selective inhibitor (CB5468139) increased the total ceramide levels by 30?0%. Treatment with ABC294640 resulted in the greatest increase in total ceramides, reaching 2-times the level of control cells, with significant increases in C26-ceramide, C24-ceramide, C22:1-ceramide, C20:1-ceramide, C18-ceramide and C16-ceramide. The levels of sphingosine were dramatically decreased after treatment with SKI-II or ABC294640. Most importantly, compared to the control group, the levels of intracellular S1P were decreased by over 90% after treatment with SKI-II or ABC294640, while treatment of CB5468139 only decreased S1P by less than 20%, which is similar to the amount of S1P remaining after SK2-selective inhibition by ABC294640. Therefore the cell-based analyses of the sphingolipid profiles support the selectivity of these compounds determined in the in vitro assays. Cell Cycle and migration/invasion. We previously reported that knockdown of either SK1 or SK2 by siRNA promotes cell cycle arrest, but does not induce apoptosis in A498 cells [17]. Treatment of the cells with DMS did not impact the cell cycle, but drove approximately 20% of the cells into apoptosis (Figure 5D). Given that DMS inhibits PKC activity, this apoptotic response is consistent with other studies [22,45]. Interestingly, SKI-II arrested cells in S phase with a concomitant decrease in the G2/M phase. ABC294735 and CB5468139 had no significant affect on the percentage of cells in G1, G2/M or apoptosis, but they did slightly increase the S phase percentage. In contrast, ABC294640 showed a distinct G1 arrest with significant decreases in both S and G2/M phases, similar to changes observed upon genetic ablation of SK2. Taken together, the data show that A498 cells respond differently to the SK2-selective inhibitor than they do to SK1-selective or SK1/2-dual inhibitors. Because S1P is involved in cell migration [46], we investigated the effects of the SK inhibitors on serum-induced migration and invasion (through Matrigel). As shown in Figure 5E, both migration and invasion were more strongly attenuated by treatment with an SK2 inhibitor, either selective (ABC294640) or non-selective (SKI-II), than by the SK1-selective inhibitor (CB5468139).
Signaling pathways. Effects of the SK inhibitors on markers of proliferation and migration were also assessed. We first determined the expression and activation of STAT3, AKT and ERK, all major regulators of cell proliferation (Figure 6). For STAT3, ABC294640 caused the largest decrease in expression (STAT3) and activation (pSTAT3), with no further decrease occurring in the SK1/2-dual inhibitor-treated groups. None of the inhibitors impacted total AKT expression; however, pAKT was decreased dramatically by treatment with ABC294640, and to a lesser degree by CB5468139 and SKI-II. The pattern of effects on ERK expression was not as clear-cut, but decreases of ERK2 were observed after SKI-II or ABC294640 treatment, as were levels of pERK1/2 after treatment of any of the inhibitors. This is consistent with data from SK2-selective ablation in which there was a greater decrease in ERK2 than ERK1 [17]. Taken together, the data suggest that ABC294640 has a greater impact than the other SK inhibitors on regulating the expression and phosphorylation of STAT3, AKT and ERK. To assess the mechanism of the observed cell cycle arrest, levels of p53 and p21 were also measured after treatment with the SK inhibitors. Similar to the changes observed with the proliferation markers, SKI-II- or ABC294640-treatment decreased the protein levels of p53 and p21, while CB5468139 reduced the expression of p53, but not p21. DMS had essentially no affect on these proteins, consistent with its lack of affect on cell cycle distribution. Interestingly, cell migration-related pFAK(Y397) was strongly down-regulated following treatment with SKI-II, ABC294640 or CB5468139. Consistent with inhibition of migration and invasion, these changes were greatest in cells treated with SK2-selective ABC294640. Interestingly, FAK expression and phosphorylation were not altered by DMS, once again indicating that the cellular effects of this compound are markedly different than those of other SK inhibitors. Because we have previously demonstrated that ABC294640 induces autophagy in A498 cells [47], we assessed the effects of the SK inhibitors on beclin1 and LC3. Each of the SK inhibitors elevated the expression of both of these autophagy markers, with ABC294640 having the most pronounced increases.
