Discussion
The SKs are becoming increasingly recognized as potential new targets for anticancer drugs; however, the literature provides differing views on the relative importance of SK1 and SK2 in cancer biology. Therefore, it is critical to define the specific roles as well as the “drugability” of the two SK isoenzymes. We previously used siRNAs to selectively deplete SK1 and/or SK2 from cancer cells, and demonstrated that ablation of SK2 results in stronger anticancer effects than does ablation of SK1 [17]. Additionally, that previous work showed that SK1 cannot restore proliferation, migration or invasion activity to cells that lack SK2 activity [17]. The goal of the present study was to use SK inhibitors to determine if selective pharmacologic inhibition of SK1 and/or SK2 activity replicates the findings of the genetic ablation approach. In studies described herein, we show clear differences in the catalytic rates, substrate affinities and structural topologies for SK1 and SK2. Computational modeling suggests that the nucleotide binding site is highly conserved, whereas the lipidFigure 5. Biochemical and phenotypic effects of SK inhibitors. A) Cytotoxicity – After treatment with varying concentrations of the SK inhibitors for the indicated time periods, cell numbers were quantified. The IC50 represents the concentration of the test compound that reduces cell number by 50% compared with DMSO-treated control cultures. In subsequent experiments, cells were incubated with each SK inhibitor at its IC50s (DMS – 5 mM, SKI-II – 20 mM, ABC294735 – 35 mM, CB5468139 – 10 mM and ABC294640 – 40 mM) for 48 hours. B) SK expression – After SK inhibitor treatment, mRNAs for SK1 (open bars) and SK2 (filled bars) compared to the vehicle controls (DMSO-treated cells) were calculated. C) Sphingolipid profiling ?After SK inhibitor treatment, cells were harvested and sphingolipids were analyzed by mass spectrometric. The bars represent the ratio of the amount of the indicated lipids in drug treated cells compared with control cells. Abbreviations are: C26-ceramide (C26-Cer), C24-ceramide (C24Cer), C20-ceramide (C20-Cer), C18-ceramide (C18-Cer), C16-ceramide (C16-Cer), sphingosine (Sph) and dihydrosphingosine (dhSph). D) Cell cycle distribution ?After SK inhibitor treatment, cells were harvested and analyzed by flow cytometry. The bars indicate the percentage of cells in each of the indicated cell cycle phases. E) Cell migration and invasion ?After SK inhibitor treatment, cells were harvested and migration, through unmodified filters (open bars) and invasion through Matrigel-coated filters (filled bars), were analyzed. The bars indicate the percentage of cells that had migrated compared to control cells. Data are the mean 6 SEM of three independent experiments. *p,0.05, **p,0.01, ***p,0.001 versus control.binding sites are divergent between SK1 and SK2. Here, we provide the first comprehensive, side-by-side comparisons of five small molecule SK inhibitors. Each compound was classified as a dual or SK1- or SK2-selective inhibitor, and then the inhibitors were used as pharmacologic probes for several biochemical pathways and cell phenotypes.
It is likely that small molecule inhibitors (SMIs) of the SKs will have advantages over other classes of S1P signaling inhibitors such as monoclonal antibodies. For example, SMIs are more structurally stable, have optimal hydrophobicity to pass through biological membranes to reach the target and are less likely to have immune system tolerance issues. Additionally, many SMIs are orallyFigure 6. Effects of SK inhibitors on signaling proteins. A) Cells were incubated with the indicated SK inhibitors at their respective IC50s for 48 hours and the levels of the indicated proteins were estimated by western blotting. Figures are representative blots of at least three independent experiments. B) Immunoblotting was conducted with the indicated antibodies, and densitometric quantification was done using Image J. Relative expression is represented as the ratio to control after normalization to b-actin. bioavailable, which simplifies the administration and drug formulation systems. On the other hand, SMI may have lower selectivity than antibodies, and in fact we confirm herein that DMS is likely to exert its cellular effects through targets other than or in addition to SKs. CB5468139, described here for the first time, provides an important indication that SK1-selective agents can be developed, but itself is likely to be too non-selective for development. In contrast, the data presented here support the hypothesis that selective targeting of SK2 by ABC294640 has excellent potential for use in cancer chemotherapy [34,36,48]. Importantly, we now show that ABC294640 has similar or greater potency for inhibiting cancer cell proliferation and migration compared to SK1/2-dual and SK1-selective inhibitors. This indicates that the functions of SK2-generated S1P cannot be fully compensated by SK1-generated S1P, possibly due to their different subcellular localizations [17,49,50]. Also, sphingolipid profiling demonstrated significant increases in ceramide species combined with depletion of S1P after ABC294640 treatment which likely intensifies its anti-proliferative activity.
Interestingly,the significant increase of SK2 expression in response to exposure to ABC294640 may also contribute to its anti-cancer activity because SK2 possesses a pro-death BH3 domain. On one hand, ABC294640 treatment inhibits the mitogenic kinase function of SK2; while on the other hand, the overexpression of the BH3 domain could provide a magnified pro-death stimulus. This is consistent with studies that showed that overexpression of SK2 by transfection results in apoptosis [15,51]. Further study of the promoter elements responsible for SK2 transcription would be of considerable interest to elucidate the mechanism for induction by ABC294640. The expression and phosphorylation of pro-survival signaling proteins such as STAT3, AKT, ERK and FAK were markedly impacted by the SK2-selective inhibitor ABC294640, and to a less degree by other SK inhibitory compounds. ABC29460 also disrupted the cell cycle with arrest in G1 and reduced expression of p53 and p21, which mimicked the selective knockdown of SK2 with siRNA [17]. Flow cytometric analyses did not reveal significant increases in apoptosis after treatment with
ABC294640; however elevation of the autophagy markers Beclin1 and LC3 suggest that the cells are dying by excessive autophagy. Although autophagy is recognized as a survival mechanism under most conditions, it is also capable of inducing cell death characterized by extensive digestion of intracellular organelles leading to large numbers of autophagic vacuoles [52,53]. Furthermore, a number of small molecules (including several anticancer drugs) activate autophagy in cancer cells both in vivo and in vitro [54].