Lipotoxicity has been considered the main cause of pancreatic beta-cell failure during type 2 diabetes development. Lipid droplets (LD) are believed to regulate the beta-cell sensitivity to free fatty acids (FFA), but the underlying molecular mechanisms are largely unclear. Accumulating evidence points, however, to an important role of intracellular sphingosine-1-phosphate (S1P) metabolism in lipotoxicity-mediated disturbances of beta-cell function. In the present study, we compared the effects of an increased irreversible S1P degradation (S1P-lyase, SPL overexpression) with those associated with an enhanced S1P recycling (overexpression of S1P phosphatase 1, SGPP1) on LD formation and lipotoxicity in rat INS1E beta-cells. Interestingly, although both approaches led to a reduced S1P concentration, they had opposite effects on the susceptibility to FFA. Overexpression of SGPP1 prevented FFA-mediated caspase-3 activation by a mechanism involving an enhanced lipid storage capacity and prevention of oxidative stress. In contrast, SPL overexpression limited LD biogenesis, content, and size, while accelerating lipophagy. This was associated with FFA-induced hydrogen peroxide formation, mitochondrial fragmentation, and dysfunction, as well as ER stress. These changes coincided with the upregulation of proapoptotic ceramides but were independent of lipid peroxidation rate. Also in human EndoC-βH1 beta-cells, suppression of SPL with simultaneous overexpression of SGPP1 led to a similar and even more pronounced LD phenotype as that in INS1E-SGPP1 cells. Thus, intracellular S1P turnover significantly regulates LD content and size and influences beta-cell sensitivity to FFA.

Astrocytes are critical players in brain health and disease. Brain pathologies and lesions are usually accompanied by astroglial alterations known as reactive astrogliosis. Sphingosine 1-phosphate lyase (SGPL1) catalysis, the final step in sphingolipid catabolism, irreversibly cleaves its substrate sphingosine 1-phosphate (S1P). We have shown that neural ablation of SGPL1 causes accumulation of S1P and hence neuronal damage, cognitive deficits, as well as microglial activation. Moreover, the S1P/S1P-receptor signaling axis enhances ATP production in SGPL1-deficient astrocytes. Using immunohistochemical methods as well as RNA Seq and CUT&Tag we show how S1P signaling causes activation of the astrocytic purinoreceptor P2Y1 (P2Y1R). With specific pharmacological agonists and antagonists, we uncover the P2Y1R as the key player in S1P-induced astrogliosis, and DDX3X mediated the activation of the NLRP3 inflammasome, including caspase-1 and henceforward generation of interleukin-1ß (IL-1ß) and of other proinflammatory cytokines. Our results provide a novel route connecting S1P metabolism and signaling with astrogliosis and the activation of the NLRP3 inflammasome, a central player in neuroinflammation, known to be crucial for the pathogenesis of numerous brain illnesses. Thus, our study opens the door for new therapeutic strategies surrounding S1P metabolism and signaling in the brain.

Simplified analogues of the myxobacterial polyketide ajudazol were obtained by synthesis and evaluated for their biological activities. Potent simplified 5-lipoxygenase inhibitors were identified. Moreover, strong antiproliferative and apoptotic activities were observed in brain cancer cell lines at low nano- to micromolar concentrations.

Astrocytes are critical players in brain health and disease. Sphingosine-1-phosphate (S1P), a bioactive signaling lipid, is involved in several vital processes, including cellular proliferation, survival, and migration. It was shown to be crucial for brain development. Its absence is embryonically lethal, affecting, inter alia, the anterior neural tube closure. However, an excess of S1P due to mutations in S1P-lyase (SGPL1), the enzyme responsible for its constitutive removal, is also harmful. Of note, the gene SGPL1 maps to a region prone to mutations in several human cancers and also in S1P-lyase insufficiency syndrome (SPLIS) characterized by several symptoms, including peripheral and central neurological defects. Here, we investigated the impact of S1P on astrocytes in a mouse model with the neural-targeted ablation of SGPL1. We found that SGPL1 deficiency, and hence the accumulation of its substrate, S1P, causes the elevated expression of glycolytic enzymes and preferentially directs pyruvate into the tricarboxylic acid (TCA) cycle through its receptors (S1PR(2,4)). In addition, the activity of TCA regulatory enzymes was increased, and consequently, so was the cellular ATP content. The high energy load activates the mammalian target of rapamycin (mTOR), thus keeping astrocytic autophagy in check. Possible consequences for the viability of neurons are discussed.

Lipids are essential structural and functional components of the central nervous system (CNS). Sphingolipids are ubiquitous membrane components which were discovered in the brain in the late 19th century. In mammals, the brain contains the highest concentration of sphingolipids in the body. Sphingosine 1-phosphate (S1P) derived from membrane sphingolipids evokes multiple cellular responses which, depending on its concentration and localization, make S1P a double-edged sword in the brain. In the present review we highlight the role of S1P in brain development and focus on the often contrasting findings regarding its contributions to the initiation, progression and potential recovery of different brain pathologies, including neurodegeneration, multiple sclerosis (MS), brain cancers, and psychiatric illnesses. A detailed understanding of the critical implications of S1P in brain health and disease may open the door for new therapeutic options. Thus, targeting S1P-metabolizing enzymes and/or signaling pathways might help overcome, or at least ameliorate, several brain illnesses.

Mitochondrial dysfunction can either extend or decrease Caenorhabditis elegans lifespan, depending on whether transcriptionally regulated responses can elicit durable stress adaptation to otherwise detrimental lesions. Here, we test the hypothesis that enhanced metabolic flexibility is sufficient to circumvent bioenergetic abnormalities associated with the phenotypic threshold effect, thereby transforming short‐lived mitochondrial mutants into long‐lived ones. We find that CEST‐2.2, a carboxylesterase mainly localizes in the intestine, may stimulate the survival of mitochondrial deficient animals. We report that genetic manipulation of cest‐2.2 expression has a minor lifespan impact on wild‐type nematodes, whereas its overexpression markedly extends the lifespan of complex I‐deficient gas‐1(fc21) mutants. We profile the transcriptome and lipidome of cest‐2.2 overexpressing animals and show that CEST‐2.2 stimulates lipid metabolism and fatty acid beta‐oxidation, thereby enhancing mitochondrial respiratory capacity through complex II and LET‐721/ETFDH, despite the inherited genetic lesion of complex I. Together, our findings unveil a metabolic pathway that, through the tissue‐specific mobilization of lipid deposits, may influence the longevity of mitochondrial mutant C. elegans. Mitochondrial function influences survival and stress resilience in Caenorhabditis elegans. cest‐2.2 overexpression promotes mobilization of lipid deposits and extends lifespan of complex I deficient mutants. Mitochondrial function influences survival and stress resilience in Caenorhabditis elegans. cest‐2.2 overexpression promotes mobilization of lipid deposits and extends lifespan of complex I deficient mutants.

Sphingosine-1-phosphate (S1P), a bioactive signaling lipid, is involved in several vital processes, including cellular proliferation, survival and migration, as well as neovascularization and inflammation. Its critical role in the development and progression of cancer is well documented. The metabolism of S1P, which exerts its effect mainly via five G protein-coupled receptors (S1PR1-5 ), is tightly regulated. S1P-lyase (SGPL1) irreversibly cleaves S1P in the final step of sphingolipid catabolism and exhibits remarkably decreased enzymatic activity in tumor samples. In this study, we used SGPL1-deficient (Sgpl1-/- ) mouse embryonic fibroblasts (MEFs) and investigated the impact of S1P on glucose metabolism. Accumulated S1P activates, via its receptors (S1PR1-3 ), hypoxia-inducible factor 1 and stimulates the expression of proteins involved in glucose uptake and breakdown, indicating that Sgpl1-/- cells, like cancer cells, prefer to convert glucose to lactate even in the presence of oxygen. Accordingly, their rate of proliferation is significantly increased. Activation of the Akt/mTOR pathway and hence down-regulation of autophagy indicate that these changes do not negatively affect the cellular energy status. In summary, we report on a newly identified role of the S1P/S1PR1-3 axis in glucose metabolism in SGPL1-deficient MEFs.

Cancer development is a multistep process in which cells must overcome a series of obstacles before they can become fully developed tumors. First, cells must develop the ability to proliferate unchecked. Once this is accomplished, they must be able to invade the neighboring tissue, as well as provide themselves with oxygen and nutrients. Finally, they must acquire the ability to detach from the newly formed mass in order to spread to other tissues, all the while evading an immune system that is primed for their destruction. Furthermore, increased levels of inflammation have been shown to be linked to the development of cancer, with sites of chronic inflammation being a common component of tumorigenic microenvironments. In this Review, we give an overview of the impact of sphingolipid metabolism in cancers, from initiation to metastatic dissemination, as well as discussing immune responses and resistance to treatments. We explore how sphingolipids can either help or hinder the progression of cells from a healthy phenotype to a cancerous one.

Lipids play an important role in neurodegeneration, neuroinflammation, and psychiatric disorders and an imbalance in sphingolipid levels is associated with disease. Although early diagnosis and intervention of these disorders would clearly have favorable long-term outcomes, no diagnostic tests currently exist that can accurately identify people at risk. Reliable prognostic biomarkers that are easily accessible would be beneficial to determine therapy and treatment response in clinical trials. Recent advances in lipidomic investigation methods have greatly progressed the knowledge of sphingolipids in neurodegenerative and psychiatric disorders over the past decades although more longitudinal studies are needed to understand its exact role in these disorders to be used as potential tools in the clinic. In this review, we give an overview of the current knowledge of sphingolipids in neurodegenerative and psychiatric disorders and explore recent advances in investigation methods. Finally, the potential of sphingolipid metabolism products and signaling molecules as potential biomarkers for diagnosis, prognostic, or surrogate markers of treatment response is discussed.

It is an honor for us to dedicate this Special Issue to our dearest friend Lina Obeid, who was not only a pioneer in the field of sphingolipids, but also a remarkable personality [...].

Bird feather lipids are usually attributed to the oily secretion product of the uropygial (preen) gland. We have observed, however, that feathers exhibit a strong reaction with osmium tetroxide (OsO 4 ), even after treatment with detergents. This leads us to postulate the existence of endogenous feather lipids distinct from preen gland lipids. In order to substantiate our hypothesis, we investigated down feathers from a 1-day-old chicken as their uropgygial gland is not functionally active. The results confirmed the osmiophilic reaction, which was concentrated in the center of barbs and strongly reduced after lipid extraction. In these lipid extracts, we identified using thin layer chromatography, cholesterol, various ceramides, glycolipids, phospholipids, and fatty acids, which closely resembled the lipid composition of the water barrier in the chicken-cornified epidermal envelope. This composition is clearly distinct from chicken uropygeal gland secretion (UGS) known to consist of fatty alcohols as part of aliphatic monoester waxes and of free, predominantly saturated, fatty acids. A filter assay showed a strong reactivity between OsO 4 and the fatty acids C18:1 and C18:2 and with feather lipid extracts, but not with UGS. These observations were confirmed by gas chromatography detecting unsaturated fatty acids including C18:1 and C18:2 as well as cholesterol exclusively in chicken feathers. Our results indicate that (1) endogenous lipids are detectable in chicken feathers and distinct from UGS and (2) in analogy to the morphogenesis of the cornified envelope of chicken feather lipids that may have derived from cellular feather-precursors, apparently enduring the specific cell death during developmental feather cornification.

We have shown that sphingosine 1-phosphate (S1P) generated by sphingosine kinase 2 (SK2) is toxic in neurons lacking S1P-lyase (SGPL1), the enzyme that catalyzes its irreversible cleavage. Interestingly, patients harboring mutations in the gene encoding this enzyme (SGPL1) often present with neurological pathologies. Studies in a mouse model with a developmental neural-specific ablation of SGPL1 (SGPL1(fl/fl/Nes)) confirmed the importance of S1P metabolism for the presynaptic architecture and neuronal autophagy, known to be essential for brain health. We now investigated in SGPL1-deficient murine brains two other factors involved in neurodegenerative processes, namely tau phosphorylation and histone acetylation. In hippocampal and cortical slices SGPL1 deficiency and hence S1P accumulation are accompanied by hyperphosphorylation of tau and an elevated acetylation of histone3 (H3) and histone4 (H4). Calcium chelation with BAPTA-AM rescued both tau hyperphosphorylation and histone acetylation, designating calcium as an essential mediator of these (patho)physiological functions of S1P in the brain. Studies in primary cultured neurons and astrocytes derived from SGPL1(fl/fl/Nes) mice revealed hyperphosphorylated tau only in SGPL1-deficient neurons and increased histone acetylation only in SGPL1-deficient astrocytes. Both could be reversed to control values with BAPTA-AM, indicating the close interdependence of S1P metabolism, calcium homeostasis, and brain health.

Sphingosine-1-phosphate (S1P) is not only a catabolic intermediate of all sphingolipids but also an evolutionary conserved bioactive lipid with critical functions in cell survival, differentiation, and migration as well as in immunity and angiogenesis. S1P-lyase (SGPL1) irreversibly cleaves S1P in the final step of sphingolipid catabolism. As sphingoid bases and their 1-phosphates are not only metabolic intermediates but also highly bioactive lipids that modulate a wide range of physiological processes, it would be predicted that their elevation might induce adjustments in other facets of sphingolipid metabolism and/or alter cell behavior. We actually found in a previous study that in terminally differentiated neurons SGPL1 deficiency increases sphingolipid formation via recycling at the expense of de novo synthesis. We now investigated whether and how SGPL1 deficiency affects the metabolism of (glyco)sphingolipids in mouse embryonic fibroblasts (MEFs). According to our previous experiments in neurons, we found a strong accumulation of S1P in SGPL1-deficient MEFs. Surprisingly, a completely different situation arose as we analyzed sphingolipid metabolism in this non-differentiated cell type. The production of biosynthetic precursors of complex glycosphingolipids including ceramide, glucosylceramide and also ganglioside GM3 via de novo synthesis and recycling pathway was substantially increased whereas the amount of more complex gangliosides dropped significantly.

The compatible solute ectoine is one of the most abundant and powerful cytoprotectant in the microbial world. Due to its unique ability to stabilize biological membranes and macromolecules it has been successfully commercialized as ingredient of various over-the-counter drugs, achieving primarily epithelial protection. While trying to elucidate the mechanism of its cell protective properties in in-vitro studies, a significant anti-inflammatory effect was documented for the small molecule. The tissue protective potential of ectoine considerably improved organ quality during preservation. In addition, ectoine and derivatives have been demonstrated to significantly decrease inflammatory cytokine production, thereby alleviating the inflammatory response following organ transplantation, and launching new therapeutic options for pathologies such as Inflammatory Bowel Disease (IBD) and Chronic Obstructive Pulmonary Disease (COPD). In this review, we aim to summarize the knowledge of this fairly nascent field of the anti-inflammatory potential of diverse ectoines. We also point out that this promising field faces challenges in its biochemical and molecular substantiations, including defining the molecular mechanisms of the observed effects and their regulation. However, based on their potent cytoprotective, anti-inflammatory, and non-toxic properties we believe that ectoines represent promising candidates for risk free interventions in inflammatory pathologies with steeply increasing demands for new therapeutics.

Article Figures and data Abstract Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Sphingolipids are membrane and bioactive lipids that are required for many aspects of normal mammalian development and physiology. However, the importance of the regulatory mechanisms that control sphingolipid levels in these processes is not well understood. The mammalian ORMDL proteins (ORMDL1, 2 and 3) mediate feedback inhibition of the de novo synthesis pathway of sphingolipids by inhibiting serine palmitoyl transferase in response to elevated ceramide levels. To understand the function of ORMDL proteins in vivo, we studied mouse knockouts (KOs) of the Ormdl genes. We found that Ormdl1 and Ormdl3 function redundantly to suppress the levels of bioactive sphingolipid metabolites during myelination of the sciatic nerve. Without proper ORMDL-mediated regulation of sphingolipid synthesis, severe dysmyelination results. Our data indicate that the Ormdls function to restrain sphingolipid metabolism in order to limit levels of dangerous metabolic intermediates that can interfere with essential physiological processes such as myelination. Introduction The sphingolipid metabolic pathway is a fundamental feature of all eukaryotic cells (Merrill, 2011). It is required to produce complex sphingolipids, such as sphingomyelin and the expansive glycosphingolipid family, that are plasma-membrane building blocks. It also generates bioactive metabolites (such as ceramide, sphingosine, and sphingosine-1-phosphate) that alter cell activities, including growth regulation and apoptosis, through interactions with receptors and enzymes. Some segments of the pathway have been specialized in a cell- and tissue-specific manner to supply essential sphingolipids that have unique properties needed for key physiological functions (e.g., the production and transport of ultra-long chain ceramides for creating the skin permeability barrier, or specific glycosphingolipids for forming the tightly packed myelin membrane that insulates neuronal axons) (Dunn et al., 2019). The de novo sphingolipid biosynthetic pathway begins with the condensation of an amino acid (usually serine) and a fatty acyl-CoA (usually palmitoyl-CoA) by the serine palmitoyltransferase (SPT) enzyme complex (Figure 1A) (Merrill, 2011). The first product, 3-keto-dihydrosphingosine, is then reduced to form dihydrosphingosine, which is subsequently acylated with fatty acids of different chain lengths to produce dihydroceramide. Introduction of a double bond into the sphingoid base generates ceramide, a central metabolic intermediate in the pathway. Ceramide, and its degradation product sphingosine, are bioactive and can cause cell death at elevated levels (Hannun and Obeid, 2018). Complex sphingolipids – sphingomyelin and glycosphingolipids – are generated by addition of hydrophilic head groups to the ceramide anchor. Figure 1 with 2 supplements see all Download asset Open asset Ormdl3, but not Ormdl1 or Ormdl2, single knockout (KO) mice exhibit significantly increased levels of sphingolipids in the brain. (A) Schematic of the de novo sphingolipid biosynthetic pathway and its feedback inhibition by ORMDLs through the sensing of ceramide levels. SPT, serine palmitoyltransferase; 3KDHSph, 3-keto-dihydrosphingosine; DHSph, dihydrosphingosine; DHCer, dihydroceramide; Cer, ceramide; Sph, sphingosine; S1P, sphingosine-1-phosphate. (B–D) Generation of Ormdl KO mice. Panels show the intron-exon organizations of the Ormdl genes and the protein coding regions (white). (B, C) Ormdl1 and Ormdl2 KO mice were produced by CRISPR/Cas9-induced mutations, resulting in frameshifts and premature stop codons. The locations of sgRNA sequences (red), PAM sites (green), as well as the changes in DNA and protein are indicated. The base insertion in the CRISPR/Cas9 modified Ormdl2 gene is underlined. (D) Ormdl3 KO mice were generated by germline Cre-LoxP recombination to excise exons 2, 3, and part of exon 4, resulting in the deletion of the entire protein-coding sequence. (E) RT-qPCR of Ormdl WT RNA in brain of Ormdl KO mice relative to that in WT mice. The mice were 8 weeks old. Probes detect the WT Ormdl sequences. Data are expressed as means ± SD. Unpaired Student’s t test; ***p<0.001. nd, not detectable. n = 4 for all genotypes. (F) Levels of dihydrosphingosine, total dihydroceramide, total ceramide, and sphingosine were determined by HPLC-tandem MS on lipid extracts of whole brains harvested from 8-week-old WT, Ormdl1 KO, Ormdl2 KO, Ormdl3 KO, Ormdl1/2 double KO, Ormdl1/3 double KO, and Ormdl2/3 double KO mice (Figure 1—source data 1). Data are expressed as means ± SD. One-way ANOVA with Bonferroni correction; *p<0.05, ***p<0.001. n = 8 for all genotypes. DKO, double knockout. Figure 1—source data 1 Levels of dihydrosphingosine, total dihydroceramide, total ceramide, and sphingosine from brains of WT, Ormdl1 KO, Ormdl2 KO, Ormdl3 KO, Ormdl1/2 double KO, Ormdl1/3 double KO, and Ormdl2/3 double KO mice. https://cdn.elifesciences.org/articles/51067/elife-51067-fig1-data1-v1.xlsx Download elife-51067-fig1-data1-v1.xlsx Like sterols and glycerolipids, cellular sphingolipid levels are tightly regulated (Breslow and Weissman, 2010; Brown et al., 2018; Harayama and Riezman, 2018). The discovery that the Ormdl gene family mediates feedback inhibition of de novo sphingolipid synthesis has provided insight into a homeostatic mechanism that controls sphingolipid generation in mammals (Breslow et al., 2010; Davis et al., 2019). Three mammalian Ormdl genes exist (Ormdl1, Ormdl2, and Ormdl3), and they encode small transmembrane endoplasmic reticulum proteins with amino-acid identities of around 80% (Hjelmqvist et al., 2002). The ORMDL proteins, in complex with SPT, have the capacity to sense elevated ceramide levels and to inhibit SPT enzymatic activity, thereby blocking entry of de novo synthesized sphingolipid substrate into the sphingolipid biosynthetic pathway (Davis et al., 2019; Han et al., 2019; Siow et al., 2015) (Figure 1A). The physiological contexts in which the ORMDLs are required are not well understood, but could inform us about when and why the de novo sphingolipid biosynthetic pathway requires negative regulatory control. Here, we investigated this issue by establishing knockout (KO) mice for each of the Ormdl genes. While the single KO mice were without overt phenotypes, we found that Ormdl1/3 double KO mice exhibited a conspicuous phenotype, with elevated levels of sphingolipid metabolites, severe myelination defects and neurologic abnormalities. The results indicate that ORMDLs, functioning redundantly, are essential for maintaining control of the de novo sphingolipid biosynthetic pathway in a specific physiological context – myelination – where there is a high demand for sphingolipids. Results Ormdl3, but not Ormdl1 or Ormdl2, single KO mice exhibit significantly increased levels of sphingolipids in brain In order to identify physiologic functions that are associated with the Ormdl genes, we established whole-body Ormdl1 KO, Ormdl2 KO, and Ormdl3 KO mice (Figure 1B–D, Figure 1—figure supplement 1). Ormdl1 and Ormdl2 KOs were created by CRISPR/Cas9 genome editing in embryos (Wang et al., 2013). In both cases, small deletions were introduced into exon 2, the first protein-coding exon of the two genes, to cause frameshifts in the coding sequences. These frameshifts produced premature terminations that eliminated more than 80% of the native protein-coding sequences (Figure 1B and C). For Ormdl3, gene targeting using homologous recombination at the Ormdl3 locus in embryonic stem cells was applied. This introduced LoxP sequences flanking the region encompassed by Ormdl3 exons 2 to 4, which contains the entire protein-coding region (Figure 1—figure supplement 1). Mice carrying the floxed Ormdl3 allele were bred with mice expressing Cre recombinase under the control of the ubiquitous EIIA promoter element (Lakso et al., 1996) to generate mice with a germline deletion of the entire Ormdl3 protein-coding region (Ormdl3 KO) (Figure 1D). For each of the single-gene KOs, mice carrying homozygous mutant alleles were obtained. RT-qPCR mRNA expression assays using brain RNA with probes corresponding to the deleted coding sequences indicated a deficiency of the individual wild-type (WT) Ormdl mRNA expression in the corresponding KO mouse, consistent with the introduced genomic changes (Figure 1E). We measured levels of dihydrosphingosine and dihydroceramide, two intermediates generated early in the sphingolipid biosynthetic pathway exclusively through de novo synthesis, as well as levels of ceramide and sphingosine, two bioactive metabolites generated through both de novo synthesis and recycling (Merrill, 2011) (Figure 1A), in the brains of the Ormdl1 KO, Ormdl2 KO, and Ormdl3 KO mice (Figure 1F). The Ormdl1 KO and Ormdl2 KO mice had levels of these four types of sphingolipids that were generally similar to those measured in WT mice; however, the Ormdl3 KO mice had significantly elevated brain levels of each of these sphingolipids compared with WT mice. This result suggests that amongst the Ormdls, Ormdl3 has the greatest influence on the levels of these sphingolipids, including those generated solely through de novo synthesis, consistent with ORMDL inhibitory action on SPT (Figure 1A). Elevated sphingolipids, neurologic phenotype and reduced viability when multiple Ormdls are deleted Ormdl1 KO, Ormdl2 KO, and Ormdl3 KO mice appeared overtly normal, with body weights not significantly different from those of WT mice at 8 weeks of age (Figure 2A). Intercross mating between respective single Ormdl KO mice produced litter sizes at weaning similar to those obtained through WT matings, suggesting normal fecundity and perinatal survival of Ormdl1 KO, Ormdl2 KO, and Ormdl3 KO mice (Figure 2B). Figure 2 Download asset Open asset Elevated sphingolipids, neurologic phenotype and reduced viability when multiple Ormdls are deleted. (A) Body weight of 8-week-old male (n = 7–14) and female mice (n = 7–10) of the indicated genotypes. Each circle represents the weight of an individual mouse. Unpaired Student’s t test; ***p<0.001 versus WT mice. (B) Litter size was determined at weaning for offspring of the indicated matings. Each circle represents the number of weanlings from an individual litter (n = 19–25). bnd, breeding not done. One-way ANOVA with Bonferroni correction. (C) Image of an 8-week-old Ormdl1/3 double KO mouse (right) showing characteristic clasping of the hindlimbs upon tail suspension, a sign of neurodegeneration. An Ormdl1–/– Ormdl3+/– littermate is shown (left). (D) Wire hang behavioral test. Eight-week-old mice were allowed to hang from a suspended wire using the forelimbs and the latency time to fall was recorded. The maximum hanging time was 180 s. Circles represent the mean of three determinations for each mouse (n = 16–29 mice for each genotype). One-way ANOVA with Bonferroni correction; ***p<0.001 versus WT mice. DKO, double knockout. To determine whether functional redundancy of the Ormdls may have masked the expression of a phenotype in the single Ormdl KO mice, we cross-bred mice to produce Ormdl double KO mice in each of the three possible combinations – Ormdl1/2 double KO, Ormdl1/3 double KO, and Ormdl2/3 double KO mice – and then characterized those offspring. As with the Ormdl single KO mice, dihydrosphingosine, dihydroceramide, ceramide and sphingosine levels were measured in the brains of the Ormdl double KO mice (Figure 1F). Total amounts of these four sphingolipid groups were not increased above WT levels in Ormdl1/2 double KO brain (Figure 1F). However, levels of dihydrosphingosine, dihydroceramide, ceramide and sphingosine in Ormdl1/3 double KO brain substantially exceeded those of the Ormdl3 single KO mice (Figure 1F). Ormdl2/3 double KO brain dihydrosphingosine and dihydroceramide levels were significantly elevated above WT levels, but not to the extent observed in Ormdl1/3 double KO mouse brain (Figure 1F, Figure 1—figure supplement 2). At 8 weeks of age, Ormdl1/2 and Ormdl2/3 double KO males and females had body weights similar to those of WT mice (Figure 2A). However, both male and female Ormdl1/3 double KO mice weighed significantly less than WT mice. Ormdl1/2 double KO and Ormdl2/3 double KO mice were fertile and, when intercrossed, produced litter sizes similar to those of WT mating pairs (Figure 2B). Owing to their compromised physical and neurological status (below), breeding of Ormdl1/3 double KO mice was not attempted. The Ormdl1/3 double KO mice exhibited abnormal hindlimb clasping upon tail suspension, a sign of possible neurodegeneration (Figure 2C) (Lalonde and Strazielle, 2011). We used the forelimb wire hang behavioral test (Aartsma-Rus and van Putten, 2014), which measures strength and coordination, to assess the neuromuscular status of the single and double KO Ormdl mice at 8 weeks of age. The performance of Ormdl1, Ormdl2, and Ormdl3 single KO mice and of Ormdl1/2 and Ormdl2/3 double KO mice was not significantly different from that observed for the WT mice (Figure 2D). However, the latency time to fall from the wire was significantly shorter for Ormdl1/3 double KO mice than for WT mice, indicating that Ormdl1/3 double KO mice had a behavioral deficit related to limb strength and coordination. To determine the phenotypic consequences when more than one Ormdl gene was deleted, we crossed Ormdl 1–/– 2 +/– 3 +/– mice and determined the genotype of the pups produced at weaning (Table 1). The expected distribution of genotypes of offspring from these mating pairs, based on Mendelian considerations, was significantly different from the genotype distribution actually obtained from the 156 offspring (p<0.0001, Chi-square analysis). Notably, no triple Ormdl1/2/3 KO mice were identified, although eight were predicted from the total number of offspring produced. Moreover, we also obtained much lower than predicted numbers of mice carrying only one WT Ormdl allele; that is, although 17 mice each were predicted for the Ormdl 1–/– 2–/– 3+/– and Ormdl 1–/– 2+/– 3–/– genotypes, only 5 and 0, respectively, were actually obtained. The results suggest that the absence of all six WT Ormdl alleles causes either embryonic or neonatal lethality prior to weaning. Supporting this conclusion is the observation that inheritance of only one WT Ormdl2 or Ormdl3 allele substantially reduced viability. Table 1 Analysis of offspring from Ormdl1–/– Ormdl2+/– Ormdl3+/– intercrosses. Mouse genotypes were determined by PCR of tail-snip DNA of 156 pups at weaning from 38 litters derived from Ormdl1–/– Ormdl2+/– Ormdl3+/– intercrosses. The genotype distribution frequency of offspring, predicted by Mendelian considerations and actually obtained, is shown. The Chi-square test was used to determine whether the obtained distribution of genotypes was statistically different from the predicted Mendelian ratios; p<0.0001. Ormdl1–/–Ormdl2+/+Ormdl3+/–Ormdl1–/–Ormdl2+/+Ormdl3+/+Ormdl1–/–Ormdl2+/+Ormdl3–/–Ormdl1–/–Ormdl2+/–Ormdl3+/–Ormdl1–/–Ormdl2+/–Ormdl3+/+Ormdl1–/–Ormdl2+/–Ormdl3–/–Ormdl1–/–Ormdl2–/–Ormdl3+/+Ormdl1–/–Ormdl2–/–Ormdl3+/–Ormdl1–/–Ormdl2–/–Ormdl3–/–Observed #16209336201150Observed %10.2612.825.7721.1539.7407.053.210Predicted #9.7519.59.7519.53919.59.7519.59.75Predicted %6.2512.56.2512.52512.56.2512.56.25 Myelination is disrupted in Ormdl1/3 double KO mice Sphingolipids are major components of the lipid-rich myelin membrane that surrounds axons in the nervous system and are essential for proper myelin function (Coetzee et al., 1996). We hypothesized that the ORMDLs may be especially critical when sphingolipid synthesis demand is high, such as during the formation of myelin membranes after birth. If de novo sphingolipid synthesis is poorly regulated under these circumstances by the absence of ORMDLs, increased amounts of sphingolipids might be generated, thereby interfering with orderly myelination in Ormdl1/3 double KO mice. To evaluate the impacts of the ORMDLs on myelination, we examined sciatic nerves at 6 weeks of age in WT, Ormdl1 KO, Ormdl3 KO, and Ormdl1/3 double KO mice by transmission electron microscopy (EM) (Figure 3A). Ormdl1/3 double KO mouse sciatic nerve exhibited a highly abnormal morphology when compared with sciatic nerve obtained from WT or single Ormdl1 and Ormdl3 KO mice. The nerves from the Ormdl1/3 double KO mice had a significantly lower frequency of myelinated axons per field compared with WT sciatic nerves (Figure 3B). No significant difference in the numbers of myelinated axons was observed between sciatic nerves from WT and Ormdl1 or Ormd3 single KO mice. A significantly higher frequency of redundant myelin, generally appearing as myelin ‘outfoldings’ (Golan et al., 2013), was also observed in sciatic nerves from Ormdl1/3 double KO mice compared with WT mice (Figure 3C and D). Figure 3 with 1 supplement see all Download asset Open asset Myelination is disrupted in Ormdl1/3 double KO mice. (A) Representative transmission EM images of sciatic nerve of 6-week-old WT, Ormdl1 KO, Ormdl3 KO, and Ormdl1/3 double KO mice. Scale bars, 10 μm. (B) Axon density in sciatic nerve of 6-week-old WT, Ormdl1 KO, Ormdl3 KO, and Ormdl1/3 double KO mice was determined by quantifying the number of myelinated axons in 7–12 EM fields per genotype. Data are expressed as means ± SD. One-way ANOVA with Bonferroni correction; ***p<0.001. n = 3 for WT, n = 2 for Ormdl1 KO, n = 2 for Ormdl3 KO, n = 3 for Ormdl1/3 double KO mice. (C) Example images of redundant myelin figures in sciatic nerve axons of 6-week-old Ormdl1/3 double KO mice. Scale bars, 5 μm. (D) Percentage of myelinated axons in sciatic nerve of 6-week-old WT, Ormdl1 KO, Ormdl3 KO, and Ormdl1/3 double KO mice showing redundant myelination was quantified in 7–12 EM fields per genotype. Data are expressed as means ± SD. One-way ANOVA with Bonferroni correction; ***p<0.001. n = 3 for WT, n = 2 for Ormdl1 KO, n = 2 for Ormdl3 KO, n = 3 for Ormdl1/3 double KO mice. (E) Levels of dihydrosphingosine, total dihydroceramide, total ceramide, sphingosine and hexosylceramide were determined by HPLC-tandem MS on lipid extracts of sciatic nerve from 8-week-old WT, Ormdl1 KO, Ormdl3 KO, and Ormdl1/3 double KO mice (Figure 3—source data 1). Data are expressed as means ± SD. One-way ANOVA with Bonferroni correction; ***p<0.001. n = 8 for all genotypes. DKO, double knockout. Figure 3—source data 1 Levels of dihydrosphingosine, total dihydroceramide, total ceramide, sphingosine and total hexosylceramide from sciatic nerves of WT, Ormdl1 KO, Ormdl3 KO, and Ormdl1/3 double KO mice. https://cdn.elifesciences.org/articles/51067/elife-51067-fig3-data1-v1.xlsx Download elife-51067-fig3-data1-v1.xlsx We measured levels of sphingolipid metabolites, dihydrosphingosine, dihydroceramide, ceramide, and sphingosine in sciatic nerves from these same genotypes. Sciatic nerve of Ormdl1 KO mice had levels of these sphingolipids that were not significantly different from those from WT mice, whereas sciatic nerve of Ormdl3 KO mice had significantly elevated levels of both dihydroceramide and ceramide compared with nerve from WT mice. However, the Ormdl1/3 double KO sciatic nerve had significantly elevated levels of all four of these sphingolipid species compared with levels observed in the Ormdl3 KO mice (Figure 3E). We also measured sciatic nerve levels of hexosylceramide, comprising glucosyl and galactosylceramide, the major glycosphingolipids in myelin (Wattenberg, 2019). Hexosylceramide levels in the Ormdl1 and Ormdl3 single KO sciatic nerve were significantly increased compared to those in WT sciatic nerve (Figure 3E, Figure 3—figure supplement 1C). However, sciatic nerve from Ormdl1/3 double KO mice had decreased hexosylceramide levels, even though they had an 8-fold increase in the level of ceramide, the direct precursor for hexosylceramide synthesis. This result suggests that the large increase in precursor sphingolipid substrate ceramide may have saturated the terminal glycosylation step for hexosylceramide production. Increased de novo sphingolipid biosynthesis in myelin-producing cells mimicked the phenotype of Ormdl1/3 KO mice Our results are consistent with the conclusion that elevated sphingolipid synthesis resulting from Ormdl1/3 deficiency causes dysmyelination and motor-function abnormalities in mice. Although reductions in the synthesis of specific sphingolipids have been associated with defects in myelination (Coetzee et al., 1996), it has not been shown previously that increased de novo sphingolipid synthesis can cause myelination defects. To address this issue, we constructed a mouse model that allows for conditional expression of SPT activity (Figure 4—figure supplement 1A). The transgenic mouse carries a single-chain version of the SPT enzyme with its three core subunits (SPTLC2, SPTSSA and SPTLC1) genetically fused (fusion[f]SPT) and under the control of a promoter that is activated by Cre recombinase expression (Stop-fSPT) (Alexaki et al., 2014; Gable et al., 2010). The Stop-fSPT cassette was introduced into the Rosa26 locus using the phiC31 integrase system (Tasic et al., 2011). In order to validate the model, we first generated mice carrying Stop-fSPT along with the Mx1-Cre gene, which can be activated by polyinosinic-polycytidylic acid (pIpC) administration to induce Cre expression in the liver (Kühn et al., 1995). After administration of pIpC, the fSPT polypeptide was detected by Western blotting, along with significantly increased SPT enzymatic activity, in Stop-fSPT/Mx1-Cre mouse liver compared with the liver of mouse controls carrying only Stop-fSPT (Figure 4—figure supplement 1B and C). Ceramide levels were significantly increased in the liver of mice expressing fSPT compared with that of control mice (Figure 4—figure supplement 1D). Next, we specifically expressed the fSPT gene in cells that produce myelin in order to determine whether elevated sphingolipid synthesis causes dysmyelination, as was observed in the Ormdl1/3 double KO mice. We generated mice carrying Stop-fSPT and the Cre gene under the control of the tamoxifen-inducible Plp1 promoter (Doerflinger et al., 2003), which would allow for conditional fSPT expression in Schwann cells and oligodendrocytes upon Cre activation. Cre expression was induced at 4 weeks of age by use of a tamoxifen diet. After 2 weeks, the Stop-fSPT/iPlp1-Cre mice that had been fed the tamoxifen diet exhibited severe hindlimb paralysis (Figure 4A). Sciatic nerve expression of the fSPT protein was confirmed by Western blotting (Figure 4B). Transmission EM analysis of sciatic nerves from fSPT-overexpressing mice revealed reduced axon density (Figure 4C and D) and a significantly increased frequency of redundant myelin structures (Figure 4E and F), similar to those found in Ormdl1/3 double KO mice (Figure 3C and D). Levels of dihydrosphingosine, total dihydroceramide, total ceramide, and sphingosine were significantly elevated in the sciatic nerves of the fSPT-expressing mice compared with those of control mice (Figure 4G, Figure 4—figure supplement 2A and B). However, as in the Ormdl1/3 double KO mice, hexosylceramide levels were decreased in sciatic nerves of fSPT-expressing mice even though these mice had highly elevated ceramide levels (Figure 4G – Figure 4—figure supplement 2C). Figure 4 with 2 supplements see all Download asset Open asset Increased de novo sphingolipid biosynthesis in myelin-producing cells mimicked the phenotype of Ormdl1/3 KO mice. (A) Six-week-old fSPT/iPlp1-Cre (fSPT) mouse after treatment with tamoxifen for 2 weeks. (B) Western blot of fSPT expression in sciatic nerve of 6-week-old mice carrying Stop-fSPT without and with iPlp1-Cre (fSPT) after treatment with tamoxifen for 2 weeks. Bottom panel represents the same blot reprobed with an antibody against β-actin as a loading control. n = 3 for both genotypes. (C) Representative transmission EM images of sciatic nerve of 6-week-old Stop-fSPT and Stop-fSPT/iPlp1-Cre (fSPT)mice after treatment with tamoxifen for 2 weeks. Scale bars, 5 μm. (D) Axon density in sciatic nerve of 6-week-old Stop-fSPT and fSPT/iPlp1-Cre (fSPT) mice after treatment with tamoxifen for 2 weeks was determined by quantifying the number of myelinated axons in five and six EM fields for Stop-fSPT and fSPT/iPlp1-Cre mice (fSPT), respectively. Data are expressed as means ± SD. One-way ANOVA with Bonferroni correction; ***p<0.001. n = 2 for both genotypes. (E) Example images of redundant myelin figures in sciatic nerve of 6-week-old fSPT/iPlp1-Cre (fSPT) mice after treatment with tamoxifen for 2 weeks. Scale bars, 2 μm. (F) Percentage of myelinated axons showing redundant myelination in sciatic nerve of 6-week-old Stop-fSPT and fSPT/iPlp1-Cre (fSPT) mice after treatment with tamoxifen for 2 weeks, quantified in five and six EM fields for Stop-fSPT and fSPT mice, respectively. Data are expressed as means ± SD. One-way ANOVA with Bonferroni correction; ***p<0.001. n = 2 for both genotypes. (G) Levels of dihydrosphingosine, total dihydroceramide, total ceramide, sphingosine and hexosylceramide were determined by HPLC-tandem MS on lipid extracts of sciatic nerve from 6-week-old Stop-fSPT and fSPT/iPlp1-Cre (fSPT) mice after treatment with tamoxifen for 2 weeks (Figure 4—source data 1). Data are expressed as means ± SD. One-way ANOVA with Bonferroni correction; ***p<0.001. n = 5 for Stop-fSPT, n = 7 for fSPT/iPlp1-Cre (fSPT) mice after treatment with tamoxifen for 2 weeks. Figure 4—source data 1 Levels of dihydrosphingosine, total dihydroceramide, total ceramide, sphingosine and hexosylceramide from sciatic nerves of mice carrying Stop-fSPT, without or with iPlp1-Cre (fSPT), after treatment with tamoxifen. https://cdn.elifesciences.org/articles/51067/elife-51067-fig4-data1-v1.xlsx Download elife-51067-fig4-data1-v1.xlsx Discussion The ORMDLs have been identified as regulators that mediate feedback inhibition of de novo sphingolipid synthesis to control sphingolipid levels (Breslow et al., 2010; Davis et al., 2019; Hagen-Euteneuer et al., 2012; Zhakupova et al., 2016). They are believed to act by sensing elevations in ceramide levels and by inhibiting SPT through direct protein–protein interactions within a multi-subunit enzyme complex (Davis et al., 2019), although elevated free and phosphorylated sphingoid bases have also been suggested to inhibit SPT via ORMDLs (Hagen-Euteneuer et al., 2012). The results presented here establish ORMDLs as functionally important modulators of in vivo sphingolipid levels in the nervous system. By evaluating mice in which either one or combinations of two of the three Ormdl genes were deleted, we found that mice lacking both Ormdl1 and Ormdl3 exhibit abnormal myelination in the sciatic nerve, as well as highly elevated sphingolipid levels in the nervous system. When de novo synthesis of sphingolipids was directly increased in myelin-producing cells by overexpression of fSPT, phenotypic manifestations similar to those seen in the Ormdl1/3 double KO mice occurred. These findings indicate that ORMDLs suppress sphingolipid levels that, when elevated, cause dysmyelination. After birth, during the acute phase of myelination, the Schwann cell membrane surface area expands several thousand times (Webster, 1971). The myelin membrane is approximately 70–80% lipid by weight, with much of the non-sterol lipid made up of sphingolipids (Wattenberg, 2019). Thus, the formation of the myelin membrane requires an extremely high influx of new substrate through the sphingolipid biosynthetic pathway to produce the mature sphingolipids, largely galactosylceramide and sulfated galactosylceramide, that are needed for proper myelin formation (Coetzee et al., 1996). Under conditions of very high sphingolipid synthesis, a lack of feedback inhibition could lead to high levels of ceramide if ceramide production exceeded the capacity of the downstream pathway to convert it into mature sphingolipids. Indeed, we found that the Ormdl1/3 double KO mouse sciatic nerve contained high levels of ceramide together with lower levels of the mature sphingolipid hexosylceramide. The lower density of myelinated axons in the sciatic nerve of Ormdl1/3 double KO mice may be a consequence of cell death due to this build-up of toxic sphingolipid precursor metabolites (Hannun and Obeid, 2018; Riley and Merrill, 2019). Our findings are consistent with the idea that a key physiological function of the ORMDLs is to keep de novo sphingolipid synthesis in check so that ceramide production does not exceed the metabolic capacity of the cell to convert ceramide into mature, non-toxic sphingolipids that are destined for physiological processes (Davis et al., 2019). In this study, Ormdl1/3 double KO mice exhibited a striking increase in the frequency of axons featuring redundant myelination. Redundant myelination, often observed as loops of myelin growing into the Schwann cell cytoplasm or surrounding the axon outside the normal myelin sheath, is believed to occur because of excessive myelin membrane growth or improper lipid composition (Golan et al., 2013). Unbalanced sphingolipid synthesis, caused either by deletion of the Ormdls in the Ormdl1/3 double KO mice or by the overexpression of fSPT in myelin-producing cells, may have promoted the formation of myelin with abnormal lipid composition through the excessive production of sphingolipid precursors relative to mature sphingolipids. Incorrect sphingolipid composition, caused by a deficiency of galactosyceramide, has been shown to produce redundant myelin profiles (Dup

Microglia mediated responses to neuronal damage in the form of neuroinflammation is a common thread propagating neuropathology. In this study, we investigated the microglial alterations occurring as a result of sphingosine 1-phosphate (S1P) accumulation in neural cells. We evidenced increased microglial activation in the brains of neural S1P-lyase (SGPL1) ablated mice (SGPL1fl/fl/Nes ) as shown by an activated and deramified morphology and increased activation markers on microglia. In addition, an increase of pro-inflammatory cytokines in sorted and primary cultured microglia generated from SGPL1 deficient mice was noticed. Further, we assessed autophagy, one of the major mechanisms in the brain that keeps inflammation in check. Indeed, microglial inflammation was accompanied by defective microglial autophagy in SGPL1 ablated mice. Rescuing autophagy by treatment with rapamycin was sufficient to decrease interleukin 6 (IL-6) but not tumor necrosis factor (TNF) secretion in cultured microglia. Rapamycin mediated decrease of IL-6 secretion suggests a particular mechanistic target of rapamycin (mTOR)-IL-6 link and appeared to be microglia specific. Using pharmacological inhibitors of the major receptors of S1P expressed in the microglia, we identified S1P receptor 2 (S1PR2) as the mediator of both impaired autophagy and proinflammatory effects. In line with these results, the addition of exogenous S1P to BV2 microglial cells showed similar effects as those observed in the genetic knock out of SGPL1 in the neural cells. In summary, we show a novel role of the S1P-S1PR2 axis in the microglia of mice with neural-targeted SGPL1 ablation and in BV2 microglial cell line exogenously treated with S1P.

In mammals, the brain exhibits the highest lipid content in the body next to adipose tissue. Complex sphin-golipids are characteristic compounds of neuronal membranes. Vital neural functions including information flux and transduction occur along these membranes. It is therefore not surprising that neuronal function and survival is dependent on the metabolism of these lipids. Autophagy is a critical factor for the survival of post-mitotic neurons. On the one hand, it fulfils homeostatic and waste-recycling functions and on the other hand, it constitutes an effective strategy to eliminate harmful proteins that cause neuronal death. A growing number of experimental data indicate that several sphingolipids as well as enzymes catalyzing their metabolic transformations efficiently but very differently affect neuronal autophagy and hence survival. This review attempts to elucidate the roles and mechanisms of sphingolipid metabolism with regard to the regulation of autophagy and its consequences for brain physiology and pathology.

Sphingosine-1-phosphate (S1P) lyase irreversibly cleaves S1P, thereby catalysing the ultimate step of sphingolipid degradation. We show here that embryonic fibroblasts from S1P lyase-deficient mice (Sgpl1(-/-)-MEFs), in which S1P and sphingosine accumulate, have features of Niemann-Pick disease type C (NPC) cells. In the presence of serum, overall cholesterol content was elevated in Sgpl1(-/-)-MEFs, due to upregulation of the LDL receptor and enhanced cholesterol uptake. Despite this, activation of sterol regulatory element-binding protein-2 was increased in Sgpl1(-/-)-MEFs, indicating a local lack of cholesterol at the ER. Indeed, free cholesterol was retained in NPC1-containing vesicles, which is a hallmark of NPC. Furthermore, upregulation of amyloid precursor protein in Sgpl1(-/-)-MEFs was mimicked by an NPC1 inhibitor in Sgpl1(+/+)-MEFs and reduced by overexpression of NPC1. Lysosomal pH was not altered by S1P lyase deficiency, similar to NPC. Interestingly, lysosomal Ca2+ content and bafilomycin A1-induced [Ca2+](i) increases were enhanced in Sgpl1(-/-)-MEFs, contrary to NPC. These results show that both a primary defect in cholesterol trafficking and S1P lyase deficiency cause overlapping phenotypic alterations, and challenge the present view on the role of sphingosine in lysosomal Ca2+ homeostasis.