SPHINGOSINE-1-PHOSPHATE SIGNALING AND ITS ROLE IN DISEASE

HDAC

Although SphK2 is present in the nucleus of many cells, its function there was unknown. Recently, it was shown for the first time that nuclei contain significant amounts of both S1P and sphingosine. Moreover, SphK2 is in a repressor complex with histone H3 and histone deacetylases (HDACs), producing S1P that regulates histone acetylation at specific lysine residues and gene transcription ⁷. Although S1P has no effect on histone acetyl transferases, it binds to and inhibits both HDAC1 and its close homolog, HDAC2. Inhibition of HDAC1/2 by S1P is physiologically relevant, as SphK2 was detected at the promoters of the cyclin-dependent kinase inhibitor p21, and the transcriptional regulator c-fos, where it regulates histone acetylation and enhances their transcription ⁷. These results indicate that S1P produced in the nucleus by SphK2 influences the dynamic balance of histone acetylation and thus the epigenetic regulation of specific target genes.

TRAF2

TNF receptor-associated factor 2 (TRAF2) is an adaptor protein containing a RING domain that is implicated in the regulatory ubiquitination of RIP1, a critical event in activation of NF-κB in response to TNF-α. All attempts to demonstrate that TRAF2 was the E3 ligase for RIP1 failed, however. TRAF2 is known to bind SphK1 and stimulate its activity ³⁷, though the role of S1P in the canonical NF-κB pathway was unclear. Recently, it was discovered that S1P is a missing cofactor required for the E3 ligase activity of TRAF2 and, consequently, Lys-63-linked polyubiquitination of RIP1 and NF-κB activation ³⁸. Interestingly, only S1P, and not dihydro-S1P, which lacks the double bond in S1P, binds to and activates TRAF2. This explains why, although S1P and dihydro-S1P are equally potent ligands for the five S1P receptors, only S1P suppresses apoptosis ³⁸. These data provide the first mechanistic explanation for the numerous observations of the importance of SphK1 and S1P in cytoprotective, inflammatory, and immune responses.

PKCδ

SphK1 plays a key role in a number of immune responses, including sepsis ³⁹. Investigation of the molecular mechanism revealed a role for protein kinase C delta (PKCδ) in mediating endotoxin-induced activation of NF-κB. Moreover, S1P, or recombinant SphK1 incubated with sphingosine and ATP to produce S1P, activates recombinant human PKCδ in vitro ³⁹. Although direct binding was not determined, these data suggest that S1P binds and activates PKCδ in vivo, and implicate SphK1 as a critical component and therapeutic target for septic shock.

PHB2

Prohibitin 2 (PHB2), a highly conserved protein that regulates mitochondrial assembly and function, was recently shown to bind S1P in vitro and in vivo ⁸. PHB2 predominantly localizes to the inner mitochondrial membrane where it is thought to form a large, macromolecular complex with PHB1 that is involved in mitochondrial biogenesis and metabolism ⁴⁰. This study showed for the first time that SphK2 also localizes to the mitochondria, where it produces the majority of mitochondrial S1P. Mitochondria respiration was reduced in SphK2 knockout mice due to aberrant assembly and reduced activity of complex IV (cytochrome oxidase) of the electron transport chain, suggesting that interaction of S1P with PHB2 is important for cytochrome-c oxidase assembly and mitochondrial respiration. Further work is needed to elucidate the role of mitochondria S1P in mitochondrial function during oxidative stress and aging.

Cancer

Numerous studies have shown that SphK1 and production of S1P promotes tumor growth, resistance to apoptosis, tumor angiogenesis and metastasis (reviewed in ⁴³). SphK1 message and protein levels are often upregulated in cancerous tissue and expression is correlated with chemo- and radio-resistance and poor prognosis. Consistent with these observations, overexpression of SphK1 promotes, while its inhibition reduces, tumor growth, angiogenesis and chemoresistance in numerous xenograft models ^{43, 44}.

S1P secreted from tumor cells can act through S1PRs either in an autocrine manner to promote growth, survival, motility, and metastasis ^{45, 46} or in a paracrine manner to induce endothelial adhesion molecules, angiogenesis and regulate tumor-stromal interactions as well as immune cells ⁴⁷ (Figure 3). As discussed above, S1P may also promote cancer progression via intracellular targets, such as HDAC1/2 and NF-κB. Finally, S1P can promote resistance of cancer cells to therapy by counteracting the pro-apoptotic effects of ceramide ² (Figure 1). An intriguing study recently identified S1PR1 as a key element involved in persistent activation of signal transducer and activator of transcription-3 (STAT3) in tumor cells and the tumor microenvironment ⁴⁸. STAT3 is a transcription factor for S1pr1 and, conversely, enhanced S1pr1 expression activates STAT3 and upregulates IL-6 expression, a pro-inflammatory cytokine crucial for STAT3 activation and inflammatory cell-mediated transformation and tumor progression. Thus, this is a new feed-forward mechanism that explains persistent STAT3 activation in cancer cells and the tumor microenvironment that is important for malignant progression and metastasis ⁴⁸.

Open in a separate window

Figure 3

Involvement of S1P in cancer

As described in the text, S1P regulates many processes in cancer cells and the tumor microenvironment that are important for malignant progression, including growth, survival and metastasis of the tumor, migration of stromal cells, and angiogenesis. S1P is also involved in recruitment of inflammatory cells and secretion of cytokines and chemokines that are important for inflammation and tumorigenesis.

Because of the potential importance of SphK1 in cancer and other diseases, much emphasis is now on development of isotype-specific inhibitors of SphK1 (Table 1). First-generation SphK inhibitors lacked isoform specificity, although one of these, safingol, is in phase 1 clinical trials ⁴⁹. One SphK1-specific inhibitor, known as SK1-I, inhibits the growth of acute myelogenous leukemia ⁵⁰ and glioblastoma tumors ⁵¹ in xenografts by affecting multiple survival signaling pathways. Another inhibitor, SKI-II, enhances the sensitivity of non-small cell lung cancer cells to chemotherapeutics both in vitro and in vivo ⁵². Surprisingly, SKI-II also induces proteosomal degradation of SphK1 ⁵³ and binds directly to the antagonist ligand-binding domain of the estrogen receptor, inhibiting estrogen receptor-stimulated transcriptional activity in human breast cancer cells ⁵⁴. Recently, progress was reported in the design of specific SphK1 inhibitors based on a homology model of SphK1 trained with a library of amidine-based compounds. Inhibitors with nM Ki’s for SphK1 were developed and found to significantly reduce endogenous S1P levels in leukemia U937 cells ⁵⁵. It will be interesting to see the results of in vivo studies with these compounds.

In contrast, the connection between SphK2 and cancer is still not well defined. SphK2 downregulation was more effective than SphK1 downregulation in inhibiting growth of glioblastoma cells ⁵⁶. Downregulation of SphK2 in MCF7 cells also decreased G₂-M arrest and markedly enhanced apoptosis induced by doxorubicin, probably due to effects on p21 expression ⁵⁷. SphK2-deficient breast cancer cells have impaired growth in a mouse tumor model ⁵⁸. A proposed SphK2-specific inhibitor, ABC294640, inhibited the proliferation of a variety of cancer cells in culture and reduced the S1P content and growth of mammary tumors in nude mice ⁵⁹. Interpretation of these results is complicated by the recent demonstration that this inhibitor also has anti-estrogenic effects and binds to the estrogen receptor ⁶⁰. The demonstration that S1P produced by SphK2 is an endogenous HDAC inhibitor ⁷ also suggests that the role of SphK2 in cancer progression may be dependent on tissue-specific and perhaps tumor-specific properties.

Atherosclerosis

Atherosclerosis is a chronic inflammatory condition, a hallmark of which is the extensive accumulation of lipids and immune cells, called plaques, in subendothelial regions, causing the arterial lumen to narrow and thrombose after rupture of the plaque. Interest in S1P in the pathogenesis of atherosclerosis stems from observations that S1P is in high concentrations in the circulation, where it is largely bound to high density lipoprotein (HDL) and albumin. S1P has diverse effects on a variety of cell types central to the development of atherosclerosis, including monocyte attachment and migration, proliferation of smooth muscle cells, vascular tone, and stimulation of the NF-κB pathway leading to production of pro-inflammatory cytokines ⁶¹. S1P actions are complex and it is still not clear whether S1P is pro- or anti-atherogenic. The stimulatory pathway is mediated mainly by NF-κB activation and adhesion molecule expression, probably via S1PR3, and the inhibitory pathway is mediated mainly by activation of endothelial nitric oxide synthase and nitric oxide-dependent vasorelaxation by activation of S1PR1 ⁶² (Figure 4). Although there are no major differences in atherosclerotic lesions and lipid volume in the aorta of apolipoprotein E (ApoE)^−/− S1PR3^−/− double knockout mice, there is significant reduction in macrophage and smooth muscle content of the lesions ⁶³. These results suggest that S1PR3 promotes inflammatory monocyte/macrophage recruitment and alters smooth muscle cell behavior. Because proliferation of smooth muscle cells in vitro was unaffected by S1PR3 deletion, further studies with targeted deletions of this receptor are needed to confirm its role. Both pharmacologic and genetic approaches to silence S1PR2 in ApoE−/− mice attenuated atherosclerotic lesion formation as shown by reduced plaque area, inhibition of macrophage accumulation in the aorta and increased cholesterol efflux in macrophages ^{64, 65}. These effects were attributed to lower S1PR2-dependent macrophage retention and/or transmigration ⁶⁵ and decreased ROCK and NF-κB activities, leading to decreased expression of pro-inflammatory cytokines, adhesion molecules and the chemokine MCP-1 ⁶⁴ (Figure 4). Nevertheless, the anti-atherogenic actions of HDL independent of cholesterol metabolism are suggested to be mediated by HDL-associated S1P ⁶². Indeed, a recent study showed that HDL-associated S1P is bound solely to apolipoprotein M (ApoM), which is known to have anti-atherogenic effects ⁶⁶. Hence, in addition to increasing cholesterol efflux from macrophage foam cells, and pre-β-HDL formation and its antioxidative effects, ApoM is vasculoprotective by delivering S1P to S1PR1 on endothelial cells ⁶⁶, and S1P-ApoM may well be anti-atherogenic. It is possible that binding of S1P to ApoM in HDL is one factor that determines the specificity of the physiological effects of S1P.

Open in a separate window

Figure 4

Role of S1P and its receptors in atherogenesis

Binding of S1P to distinct S1P receptors regulates vascular tone and barrier function, monocyte attachment and migration, macrophage infiltration and retention, proliferation of smooth muscle cells, and activation of NF-κB leading to production of pro-inflammatory cytokines and adhesion molecules. S1PR1, red; S1PR2, green; S1PR3, blue.

Interestingly, FTY720 markedly reduced atherosclerotic lesions in both low density lipoprotein receptor-null and ApoE-null mice ^{67, 68}. This is likely due to FTY720-mediated downregulation of S1PR1 on lymphocytes and resulting lymphopenia, as well as to shifting classical inflammatory M1 macrophages to anti-inflammatory M2 types ⁶⁷. FTY720 conferred atheroprotective effects in human primary macrophages by reducing transport of scavenged lipoprotein cholesterol to the ER and facilitating its release ⁶⁹. Surprisingly, however, this effect was independent of S1PRs. Rather, FTY720 stimulated production of 27-hydroxycholesterol from cholesterol, providing an additional explanation for its atheroprotective effects and identifying a novel mechanism for the actions of FTY720 ⁶⁹.

Diabetes and obesity

Diabetes is a chronic disease affecting hundreds of millions of people worldwide. Sphingolipid metabolism is altered in diabetic conditions, however, focus to date in this area has been on ceramide ^{70, 71} and only a few studies have examined the involvement of S1P. Fatty acid (palmitate)-induced stimulation of de novo ceramide synthesis is required, but not sufficient, for the onset of insulin resistance and ceramide is linked to insulin resistance, in part due to inhibition of Akt ^{70, 71}. Moreover, inhibition of sphingolipid synthesis increased insulin sensitivity, resolved hepatic steatosis, and prevented the onset of diabetes in obese rodents.

Inflammatory responses in obesity exacerbate insulin resistance and development of diabetes. Activation of toll-like receptor 4 (TLR4) by saturated fatty acids is one pathway of innate immunity contributing to increased inflammation associated with obesity. An intriguing study recently demonstrated that TLR4 stimulates IKKβ and NF-κB, which increases the transcription of genes involved in ceramide synthesis ⁷². These genes drive ceramide elevation, inhibiting Akt, which is required for insulin signaling (Figure 5). Interestingly, increased ceramide production was dispensable for TLR4-dependent induction of inflammatory cytokines, but required for TLR4-dependent insulin resistance ⁷². Altogether, this work suggests that ceramide is a key mediator that links lipid-induced inflammatory pathways to insulin resistance and diabetes.

Open in a separate window

Figure 5

Role of S1P in diabetes and obesity

Saturated fatty acid-induced insulin resistance is mediated by the proinflammatory receptor TLR4, which stimulates IKKβ and leads to NF-κB activation. This results in the upregulation of biosynthetic genes involved in de novo ceramide formation. Ceramide, in turn, inhibits Akt to induce apoptosis and suppress insulin function. Binding of adiponectin to its receptors enhances deacylation of ceramide to sphingosine, which can then be phosphorylated to form S1P. Inside-out signaling via S1PRs can activate AMPK, important for mitochondria biogenesis. S1P also stimulates Akt and prevents apoptosis.

Another intriguing link was recently uncovered between adipocyte-derived adiponectin – a hormone with insulin sensitizing, anti-inflammatory and anti-apoptotic functions – and ceramide and its metabolites with insulin sensitivity ⁷³. Binding of adiponectin to its two receptors, AdipoR1 and AdipoR2, stimulates ceramidase activity (deacylation to sphingosine) and formation of its anti-apoptotic metabolite S1P. S1P may then bind to its receptors, leading to activation of AMPK. These actions of adiponectin via S1P might promote survival of pancreatic beta cells, nutrient uptake, nutrient utilization and mitochondrial proliferation ⁷³ (Figure 5).

Plasma S1P levels are elevated in two animal models of type 1 diabetes (streptozotocin-induced diabetic rats and Ins2 Akita diabetic mice), yet no changes in levels were detected in the livers of these animals ⁷⁴, suggesting other sources of S1P. An S1PR2 antagonist prevented the onset of diabetes in a streptozotocin diabetes mouse model and S1PR2^−/− mice displayed lower blood glucose levels and reduced beta cell apoptosis together with higher insulin/glucose ratios (an index of relative insulin deficiency) ⁷⁵. Glucose induces a rapid and sustained increase in islet sphingosine kinase activity that is dependent on glucose phosphorylation, but independent of ATP generation and protein biosynthesis. Moreover, glucose-supported beta cell growth appears to be, in part, mediated by this increased activity. Although several studies suggest that S1P might be pro-survival for pancreatic beta cells, it is still not clear whether this is mediated by intracellular or extracellular actions of S1P, nor do we know the mechanism or which S1PRs are involved.

We apologize to authors whose work has not been cited here owing to space limitations. This work was supported by grants from the US National Institutes of Health R01CA61774, R37GM043880, R01AI50094, 1U19AI077435 (to S.S.).