Sphingosine Kinase 2 (SPHK2)
Sphingolipids are highly abundant within the cell and have long been known as important structural components of cell membranes. Within the last few decades, however, their roles as critical signaling molecules have come to light (Pitson 2011; Newton et al. 2015). Sphingosine and its precursor ceramide are now known to mediate cell death through modulation of multiple cellular targets. In contrast, S1P has an extensive functional repertoire in promoting cell survival, proliferation, migration, differentiation, as well as mediating angiogenesis and inflammatory responses, through both intracellular targets as well as extracellular roles in acting as a ligand to five G protein-coupled S1P receptors (S1P1–5). Hence, the sphingosine kinases have gained recognition as being critical regulators of cell fate through regulating the balance between proapoptotic sphingosine and ceramide, and prosurvival S1P, deemed the “sphingolipid rheostat” (Newton et al. 2015).
Since its discovery in 2000, SK2 has remained somewhat enigmatic and poorly understood compared to its extensively characterized and well-defined counterpart, SK1. Despite catalyzing the formation of the same product, S1P, functional and expression analyses have suggested that SK1 and SK2 may have differing roles. SK2 is a substantially larger protein (618 amino acids for SK2a compared with 384 amino acids for SK1a), as it contains an extended N-terminus and central domain in addition to the homologous regions it shares with SK1 (Fig. 1). Many studies have shown that SK1 promotes cell survival, proliferation, and neoplastic transformation through the production of S1P (Pitson 2011). However, the initial characterization of SK2 function revealed that, surprisingly, it seems to have a completely opposite role in facilitating cell cycle arrest and cell death (Liu et al. 2003; Okada et al. 2005). Both isoforms are ubiquitously expressed in all cells and tissues, but human SK2 is most highly expressed in kidney, liver, and brain, whereas the highest expression of human SK1 is in lung, spleen, and leukocytes. They also have different expression profiles throughout mouse embryonic development, with SK1 being more highly expressed at earlier stages (E7–E11) and SK2 expression peaking later in development (E15–E17). However, there does appear to be at least some functional overlap between SK1 and SK2, as the genetic ablation of either gene individually in mice does not give rise to any gross phenotypic abnormalities, suggesting that one can compensate for the other. Notably, double knockout of both genes results in embryonic lethality due to severe defects in neurological and vascular development.
In line with this, and despite a number of studies demonstrating a proapoptotic role for SK2, some earlier reports emerged showing that targeting the expression of SK2 in some cancer cell types could in fact induce cell death and sensitize these cells to chemotherapy. In 2010 the first SK2-selective inhibitor, ABC294640, was reported, and since then many others have been developed and characterized (Pitman et al. 2016). These inhibitors have provided the field with new tools to interrogate the functions of SK2, and somewhat surprisingly, targeting SK2 in vitro as well as in vivo has been found to provide significant anticancer effects.
A third putative human SK2 variant (SK2c; 761 amino acids) was reported in 2007, containing a further extension of the N-terminus as well as an additional sequence at the C-terminus (Neubauer and Pitson 2013). This variant has been detected at the mRNA level in two human cell lines; however its expression at the protein level has not been reported. Most databases refer to the SK2b variant as the canonical human SK2 sequence, and as such, references to specific SK2 residues will generally utilize numbering from this sequence. There is only one known SK2 variant in both mouse and rat, containing 617 and 616 amino acids (82.4% and 82.6% identity with human SK2a), respectively.
Functions of SK2
Cell Cycle Arrest and Cell Death
Many of the initial studies characterizing SK2 function revealed that it can promote cell cycle arrest and apoptosis. This is in complete contrast to SK1, and it was suggested that the opposing functions of these two enzymes is a result of their differential localization in the cell and consequent contribution to different pools of S1P (Neubauer and Pitson 2013). Exogenous expression of SK2 can inhibit DNA synthesis and enhance apoptosis, even in the absence of cellular stress (Liu et al. 2003; Okada et al. 2005; Neubauer et al. 2016). Endogenous SK2 can also function in a similar manner, as small interfering RNA (siRNA)-mediated knockdown of SK2 in both human and mouse cells rescued the cells from tumor necrosis factor α (TNFα)-induced apoptosis. Complete genetic knockout of SK2 in murine mesangial cells also recapitulated these findings, as these cells were more resistant to staurosporine-induced apoptosis than wildtype or SK1−/− cells.
SK2 in Cancer and Disease
It has been known for some time that SK1 can promote cancer initiation and progression, as it was first demonstrated in 2000 that overexpression of SK1 alone can result in full neoplastic transformation of cells (Xia et al. 2000). However, comparable overexpression of SK2 promotes cell death and so it was generally believed that SK2 did not have the same roles as SK1 in promoting cancer. Despite this, many groups have found that reducing SK2 levels by siRNA in different cancer cell types results in decreased cell proliferation and survival, and can also decrease their resistance to chemotherapy. In some instances, this anticancer effect observed with decreased SK2 levels can be more prominent than the effects observed with decreased SK1 levels. Most importantly, targeting SK2 either by genetic knockout of the gene, or by inhibition of its activity using pharmacological agents, can significantly reduce the size and progression of tumors from a range of different human cancer models in mice (Neubauer and Pitson 2013). The expression level of SK2 has been found to be elevated, but only modestly, in a number of human cancers. Notably, however, recent studies have shown that when SK2 is ectopically overexpressed to only low-levels, similar to those observed in these cancers, it can induce full transformation of cells and promote tumor formation in mice (Neubauer et al. 2016). These findings therefore confirm that SK2 can directly mediate oncogenesis, finally providing a rationale for the observed benefits of targeting SK2 in cancer.
SK2 may be elevated in some cancers as a result of the downregulation of negative regulators such as microRNAs (miRs). In papillary thyroid cancer (PTC), miR-613 is significantly downregulated compared to normal thyroid tissues, and overexpression of miR-613 in PTC cells reduced their growth as tumors in vivo by targeting the SK2 3′ untranslated region and decreasing SK2 mRNA and protein expression. This regulation of SK2 expression may contribute to the seemingly opposite roles SK2 plays in cancer versus noncancer cells, as even small increases in cellular SK2 levels can result in it becoming oncogenic (Neubauer et al. 2016).
A few mechanisms have been reported that may explain how SK2 mediates this procancer role. Through S1P-mediated inhibition of HDAC1/2 activity, SK2 can facilitate an increase in the transcription of MYC (Wallington-Beddoe et al. 2014), and S1P produced by nuclear-SK2 can also bind to and stabilize telomerase, promoting telomere maintenance and replicative potential (Panneer Selvam et al. 2015). There is also evidence of a role for SK2 in promoting metastasis, as SK2 is critical for both transforming growth factor (TGF) β- and epidermal growth factor (EGF)-induced migration of cancer cells, but not for EGF-mediated migration of noncancerous cells. Recently, it was reported that S1P released from apoptotic tumor cells, which was found to be largely dependent on SK2, could signal to tumor-associated macrophages to promote lymphangiogenesis and tumor metastasis (Jung et al. 2016).
Recently, the SK2-selective inhibitor ABC294640 (Yeliva™) completed a Phase I clinical trial for advanced solid tumors and is now in Phase I/II clinical trials for refractory/relapsed diffuse large B-cell lymphoma, and will soon commence Phase II clinical trials for both multiple myeloma and advanced hepatocellular carcinoma. ABC294640 is therefore a promising anticancer agent, but its effects may not be entirely attributed to the inhibition of SK2, as it has been shown to also inhibit other targets such as dihydroceramide desaturase (Des1) and estrogen receptor, as well as promote the degradation of SK1, Des1, Myc, androgen receptor, and Mcl-1 (Pitman et al. 2016). This emphasizes the need for more specific SK2 inhibitors in the field in order to truly interrogate the benefits of targeting SK2 in cancer, and it also highlights the importance of validating data generated through the use of pharmacological agents with genetic approaches.
There is also evidence of SK2 playing a role in other diseases (Neubauer and Pitson 2013). Unlike SK1, which is generally proinflammatory, targeting SK2 by siRNA or genetic approaches results in an increase in proinflammatory factors in various models, such as collagen-induced arthritis, inflammatory bowel disease (IBD), or cancer xenograft models. However, the overall outcomes of targeting SK2 in these disease models can vary, as where an increased inflammatory response coincided with decreased tumor burden in the xenograft model, loss of SK2 in the arthritis and IBD models resulted in enhanced disease. Counterintuitively, a number of studies have found that the genetic knockout of SK2 in mice, unlike SK1, results in increased levels of circulating S1P, and this was found to be protective against cardiac dysfunction in a murine model of sepsis (Coldewey et al. 2016). SK2 has also been shown to have protective roles in ischemia-reperfusion (IR) injury models, which model the injury caused by reoxygenation of a tissue after oxygen starvation, such as that caused by myocardial infarction or stroke. SK2 played a beneficial role in an IR injury model of the kidney, and in ischemic preconditioning in IR injury models of the brain and heart. Notably, SK2 is responsible for phosphorylating FTY720 (Fingolimod; Gilenya), an FDA-approved immunosuppressive drug used for the treatment of multiple sclerosis, converting it to its active form, FTY720-P. SK2 may also play a role in promoting Alzheimer’s disease, as SK2 activity was found to be upregulated in the brain of Alzheimer’s disease patients, and SK2 promotes the activity of BACE1, an enzyme involved in amyloid-β peptide production. Therefore, there may be some benefit in targeting SK2 as a therapeutic in diseases other than cancer.
SK2 Cellular Localization
SK2 can localize to many different cellular organelles, with its role differing at these various compartments (Fig. 3) (Neubauer and Pitson 2013). SK2 can be present in the cytoplasm, but is found generally associated with membranes, at least in part due to the lipid binding domain present within its N-terminus. SK2 has been reported to localize to the plasma membrane under various conditions and in many cell types. In apoptotic cells, SK2 can be cleaved at the N-terminus by caspase-1 and transported to the plasma membrane, where it is then released from the cell in an active form to produce extracellular S1P. This SK2-mediated extracellular S1P produced by dying cells may play a role in mediating inflammatory and immune responses, and wound healing processes. However, when overexpressed to low-levels, similar to that observed in many human cancers, SK2 was also found at the plasma membrane, coinciding with an increase in extracellular S1P formation, oncogenic signaling, and neoplastic transformation (Neubauer et al. 2016). In agreement, SK2 was strongly localized to the plasma membrane in breast cancer cells, where SK2 was found to be important for EGF-induced migration of these cells.
SK2 can also shuttle between the cytoplasm and the nucleus, as it contains both a nuclear localization sequence (NLS) and a nuclear export sequence (NES). S1P produced by SK2 in the nucleus can regulate epigenetic marks to alter the transcription of a range of different genes (discussed further below). Nuclear SK2 has been reported to induce cell cycle arrest in multiple cell lines, yet in lung cancer cells S1P generated by SK2 in the nucleus can bind to and stabilize telomerase, resulting in delayed senescence and increased tumor growth (Panneer Selvam et al. 2015). It is unclear how SK2 can have opposing functions within the same organelle, and the mechanisms regulating this are yet to be elucidated. As outlined earlier, SK2 can also localize to other internal organelles such as the mitochondria and ER, and production of S1P at these locations seems to mediate the proapoptotic functions of SK2. There have also been reports of both SK1 and SK2 colocalizing with centrosomes to play a possible role in regulating cell division, although further work is required to characterize their role at this organelle.
It is evident that the localization of SK2 within the cell is critical in determining the specific roles it performs, as S1P can have vastly different intracellular functions depending on where it is generated. Despite the importance of localization on this complex enzyme, surprisingly little is known about how it is regulated.
Protein Domains and Modifications
The sphingosine kinases possess basal catalytic activity even in the absence of any activating stimuli, allowing them to play a “housekeeping” role in maintaining cellular sphingosine and ceramide levels. However, they can be rapidly activated by a number of external factors, such as EGF, TNF-α, interleukin 1β (IL-1β), phorbol esters, as well as crosslinking of the IgE receptor FcεR1 upon antigen binding. SK2 can also be activated by hypoxia and via interaction with eukaryotic elongation factor 1A (eEF1A). This rapid increase in SK2 catalytic activity is generally mediated by phosphorylation, possibly on Thr614 (Thr578 of SK2a) by extracellular signal-regulated kinases 1/2 (ERK1/2), which results in up to a six-fold increase in SK2 activity in cells (Hait et al. 2007).
SK2 is also a substrate of protein kinase D (PKD), which phosphorylates Ser419 and/or Ser421 within the NES to promote the nuclear export of SK2 (Ding et al. 2007). Many large phosphoproteomic studies have detected the presence of other phosphorylation sites on SK2, most of which occur within the extended proline-rich central domain unique to SK2 (Fig. 4). These sites currently remain uncharacterized but they provide a hint that there is much more to discover about SK2 regulation and function. Unlike SK1, the crystal structure of SK2 has not yet been solved. Although the regions within SK2 that are conserved in SK1 can be predicted using the SK1 crystal structure, the two additional domains present in SK2 have no similarities to any other known proteins. Thus, virtually nothing is known about how these regions, and the modifications they carry, influence, or regulate SK2 function.
SK2 as an Epigenetic Regulator
The first report of SK2 acting as a regulator of gene transcription was from Hait et al., who demonstrated that S1P generated by nuclear-localized SK2 could interact with and inhibit HDAC1/2 activity, resulting in an increase in histone acetylation and increased transcription of downstream target genes encoding cyclin-dependent kinase inhibitor p21 and transcriptional regulator c-fos (Hait et al. 2009). This seminal finding uncovered an important novel role for SK2 and strengthened the notion that SK2 has physiological roles in promoting cell cycle arrest. More recently, additional target genes that can be epigenetically regulated by SK2 have been identified, confirming that SK2 is in fact a broad regulator of transcription. MYC gene transcription and protein expression were increased by SK2-mediated inhibition of HDAC1/2 in acute lymphoblastic leukemia (Wallington-Beddoe et al. 2014). Furthermore, hepatic SK2 was also found to mediate HDAC1/2 inhibition and the subsequent transcriptional upregulation of a number of genes involved in hepatic lipid metabolism, including sterol regulatory element-binding protein 1 (SREBP-1c) and fatty acid synthase (FAS), in response to bile acids in the liver (Nagahashi et al. 2015). In a mouse model of muscular dystrophy, increasing levels of nuclear S1P in muscle tissue by inhibition of the enzyme that catalyzes its irreversible degradation, S1P lyase, resulted in decreased HDAC activity, increased histone acetylation, and increased transcription of many genes involved in chromosome condensation, fatty-acid metabolism and muscle regeneration. However, although SK2 is thought to be the main source of nuclear S1P, it was not specifically demonstrated in this study that these findings were a direct result of SK2 function.
The synthetic sphingosine analog FTY-720, used clinically for the treatment of multiple sclerosis, can also enter the nucleus where it can be phosphorylated by SK2 to FTY720-P, and subsequently act like S1P to inhibit Class 1 HDACs and enhance histone acetylation (Hait et al. 2014). Broad gene expression analyses have been performed in multiple models to examine specific genes regulated by FTY720-P, and variations in the expression of genes involved in pathways such as memory, associative learning, and behavior, as well as transcription and lipid metabolism were reported. Specifically, estrogen receptor-α (ERα) was found to be transcriptionally upregulated downstream of FTY720-P-mediated HDAC inhibition and increased histone acetylation. It remains to be determined if endogenous S1P generated by nuclear-SK2 also functions similarly to FTY720-P to target these genes and pathways.
Notably, these genes reported to be targeted by SK2-associated epigenetic regulation are all involved in rather varied, somewhat opposing pathways. This novel role of SK2 is still not well characterized, and hence, the mechanisms involved in localizing SK2 to the nucleus and the specific conditions required to target it to specific promoters are still largely unknown.
Almost two decades since its discovery, there is still much to learn about the complex functions and regulation of SK2. It currently appears that SK2 can act to promote both cell survival and proliferation, as well as cell death. But the mechanisms regulating how SK2 can switch between these opposing functions, and how they become biased toward a prosurvival role in some cancers, are virtually unknown. It is clear that the localization of SK2 within the cell, and hence the regions of S1P production, influences its functions. However, SK2 seems to have opposing roles even within the same organelles, and it possesses a number of uncharacterized post-translational modifications that may act to regulate its function. Clearly, substantial work is still required to better understand this complex enzyme, and given the promising effects of targeting SK2 in cancer, interest in studying SK2 and generating more potent, specific SK2 inhibitors is likely to gain momentum within the field. It will remain to be determined if the opposing functions of SK2 result in any complications in targeting this enzyme in cancer, or whether they in fact provide a beneficial therapeutic window for cancer treatment.
- Neubauer HA, Pham DH, Zebol JR, Moretti PA, Peterson AL, Leclercq TM, et al. An oncogenic role for sphingosine kinase 2. Oncotarget. 2016;7:64886–99.Google Scholar