Genetic elevation of Sphingosine 1-phosphate suppresses dystrophic muscle phenotypes in Drosophila

Duchenne muscular dystrophy (DMD) is an age-dependent muscle wasting disease caused by the absence of fully functional dystrophin protein. This severe X-linked recessive disease affects 1 in 3500 male births (Deconinck and Dan, 2007). In human dystrophic muscles, rounds of contractions result in degeneration/regeneration cycles. However, regeneration is not sufficient to overcome degeneration in dystrophic muscles, leading to muscle wasting over time. As no effective treatment exists presently, new approaches are imperative (Mendell et al., 2010; Palmieri et al., 2010). Genetically amenable models for DMD might allow identification of new components that interact with the dystrophin glycoprotein complex.

Presently, several mammalian animal models, including mouse, dog and cat, have been established to study the disease (Collins and Morgan, 2003). More recently, genetically tractable DMD models have also been established in zebrafish, Caenorhabditis elegans and Drosophila melanogaster (Collins and Morgan, 2003; Shcherbata et al., 2007). Among the dystrophic phenotypes observed in Drosophila is a visible wing vein defect (Christoforou et al., 2008; Kucherenko et al., 2008), which was previously used to screen for suppressors (Kucherenko et al., 2008). One of the suppressors of the wing vein defect was mutant of wunen, a lipid phosphate phosphatase, which was initially identified as a gene required for proper germ cell migration in the Drosophila embryo (Zhang et al., 1996; Zhang et al., 1997; Starz-Gaiano et al., 2001). There are three lipid phosphate phosphatases in mammals: LPP1, LPP2 and LPP3 (also known as PPAP2A, PPAP2C and PPAP2B, respectively). Rescue analysis revealed that human LPP3 acts as a functional homolog of wunen in germ cell migration (Burnett and Howard, 2003). It is noteworthy that despite having the same broad substrate range in vitro, acting upon phosphatidic acid (PA), sphingosine 1-phosphate (S1P) and their derivatives (Pyne et al., 2005), only Lpp3 knockout mice are embryonic lethal indicating a necessary central role for this enzyme in development (Brindley and Pilquil, 2009). Interestingly, it has been reported that LPP2 preferentially attenuates intracellular PA levels whereas LPP3 specifically attenuates intracellular levels of S1P in human cells (Long et al., 2005), indicating a preference for this substrate.

Human LPP3 is a secreted protein and has its active site either in the extracellular space or in the lumen of intracellular organelles, yet it affects intracellular levels of S1P. This presumably occurs through inter-compartmental movement of S1P via a transporter such as SPNS2, a member of the major facilitator family of transporters that has been shown to specifically transport S1P (Osborne et al., 2008; Kawahara et al., 2009; Hisano et al., 2011).

S1P has been implicated in muscle repair and satellite cell proliferation and myoblast differentiation in vitro (Nagata et al., 2006; Bruni and Donati, 2008; Rapizzi et al., 2008). Here, we show that, in Drosophila, wunen reduction, in addition to suppressing the wing vein defect, also suppresses dystrophic muscle defects as assayed by muscle integrity, Projectin (Bent – FlyBase) protein localization in sarcomeres, and fly movement over time. Furthermore, mutants that increase intracellular S1P levels either biochemically or through the reduction of transport also suppress these dystrophic muscle phenotypes in Drosophila. Finally, elevating S1P or S1P signaling in adult flies using pharmacological agents phenocopied the genetic data.

The fly strains used in this study are: Oregon-R, w¹¹¹⁸, Dys^det1, wun^EMS4/CyO, yw, TubGal4/TM3, cn¹Sply⁰⁵⁰⁹¹/CyO; ry⁵⁰⁶, Df(3R)Exel6184/TM6B (outcrossed to w¹¹¹⁸ at least seven times in this laboratory) and wun^EY11935, all obtained from the Bloomington Stock Center; w; UAS-HA-lace/CyO kindly provided by Dr T. Adachi-Yamada (Gakushuin University, Tokyo, Japan); w; wun^k10201/CyO (stock #111469) obtained from the Kyoto Drosophila Genetic Resource Center; wun^RNAi [w¹¹¹⁸; P{GD1640}v6446/TM3 (stock #6446)] obtained from the Vienna Drosophila RNAi Center (Dietzl et al., 2007); ActGal4:UAS-Dys^N-RNAi/CyO and TubGal4:UAS-Dys^C-RNAi/TM6B, Tb recombinant lines, previously generated in this laboratory (Kucherenko et al., 2008); and SK1 and SK2 RNAi lines obtained from the Vienna Drosophila RNAi Center (stock numbers v32932 and v101018, respectively).

Flies were dipped in 95% ethanol and then dissected in 1× PBS (pH 7.4). Heads and abdomens were removed, and the thoraces opened. Samples were then fixed in 5% paraformaldehyde (Electron Microscopy Sciences) at room temperature (RT) for 1 hour with rocking, rinsed in PBT [PBS with 0.2% Triton X-100 (v/v)] four times (15 minutes each), and then blocked for 1 hour in PBTB [PBT with 0.4% bovine serum albumin (w/v), 5% normal goat serum (v/v)] at RT. Samples were stained with primary antibody pre-diluted in PBTB [rat anti-Projectin (Mac150), Babraham Institute; 1:50] overnight at 4°C with rocking, rinsed in PBT four times (10 minutes each) at RT, and then stained with secondary antibody pre-diluted in PBTB (Alexa Fluor 488-conjugated goat anti-rat at 1:500 and Alexa-Fluor 568-conjugated goat anti-mouse anti-Phalloidin at 1:200; Invitrogen) overnight at 4°C with rocking. Samples were then rinsed in PBT four times (10 minutes each) at RT and stored in 80% glycerol (v/v), 3% n-propyl gallate (w/v), 20% Prolong Gold (Invitrogen) (v/v). Individual indirect flight muscles (IFMs) were mounted on slides using forceps to spread them evenly and then coverslipped. Analysis was carried out using a Leica TCS-SPE confocal microscope with a 40× objective and Leica Software. Myofibrils were quantified as percentage wild type (WT) out of total number analyzed. For each myofibril, a confocal projection was scored WT or non-WT where WT was recorded when Projectin staining continuously spanned the entire width of the myofibril. For each genotype, four to five flies were analyzed, and six to seven myofibrils from three different IFMs (approximately two myofibrils per IFM) per fly were analyzed.

Transverse histological sections of IFMs were prepared from paraffin wax-embedded material. Flies were immobilized in Heisenberg fly collars (Model #10731, 4M Instrument & Tool LLC, New York) between the abdomen and thorax, then fixed in Carnoy’s solution [6:3:1 ethanol (EtOH):chloroform:glacial acetic acid] overnight at 4°C. After fixation, at RT, samples were hydrated/dehydrated to remove the Carnoy’s with the following procedure: 40% EtOH (10 minutes), 75% EtOH (10 minutes), 95% EtOH (10 minutes), then 100% EtOH twice (10 minutes each). The flies were then infiltrated with paraffin (Poly/Fin, Triangle Biomedical Sciences) using the following procedure: samples were placed in methyl benzoate for 30 minutes at 65°C, then transferred to a methyl benzoate:paraffin solution (1:1) and incubated at 65°C for 30 minutes, and finally placed in paraffin alone at 65°C for an additional 30 minutes. Afterwards, samples were placed in casts then filled with melted paraffin (65°C). Once cooled and solidified, the collars were removed, shearing off the fly abdomens. Transverse sections of the fly thoraces were cut with a rotary microtome (Leica 820 Histocut) at 10 μm thickness. Paraffin was removed with xylene (two 4-minute immersions) and the sections were rehydrated [100% EtOH twice (4 minutes each), 95% EtOH (3 minutes), 70% EtOH (2 minutes), H₂O (1 minute)] then stained with Hematoxylin and Eosin using standard protocols. Sections were covered with DPX Mountant (Fluka), coverslipped and analyzed by light microscopy (Leica) using the 20× objective. The middle third (15-20 sections) of the thorax, cut in 10 μm sections, of each fly was scored for degeneration. Each section was rated A-E depending on the severity of the fragmentation phenotype as follows: (A) more than two of the IFMs were unfragmented; (B) at least one IFM was unfragmented; (C) fragmentation in all IFMs; (D) severe degeneration in at least one IFM; (E) severe degeneration in more than one IFM. A muscle integrity index (MII) was then calculated. The MII is a weighted average of sections showing less degeneration over time (A=16, B=8, C=4, D=2 and E=0). Each fly was given an overall score from ∼20 sections. Eight to twelve flies were scored for each experiment; each experiment was repeated three times.

An apparatus made for fractionating a Drosophila population using a countercurrent distribution procedure (Benzer, 1967) was used to measure climbing ability of different fly cohorts. The apparatus was a kind loan from Leo Pallanck (University of Washington, WA, USA). Twenty to thirty flies were tested for their ability to climb at least 10 cm in 30 seconds. Each group of flies was tested for five trials with 3 minutes separating each trial and this denoted a single experiment. Each set of five trials was then given a climbing index (CI) value. The CI is a weighted average of the number of flies that end up in progressively more distant tubes from the first tube of the apparatus. Weighted values were 0, 2, 4, 8, 16, 32 from the first to the last tube, respectively. Each experiment was repeated three times.

A TriKinetics DAM5 activity monitoring system was used to record the movements of individual flies over the course of different lengths of time. The monitors were placed at 38° angles with the food at the bottom of the glass tubes. Seven to eight flies were monitored for each genotype per experiment. Data for the one or two least active flies for each genotype was removed leaving data for the six most active flies per experiment. Each experiment was repeated three times.

For the feeding assay, 4-day-old (or 0-day-old for long term experiments) flies were placed in empty vials and starved for 6 hours. The flies were then provided access to a solution of 0.05 mg/ml (217 μM) 2-acetyl-4(5)-tetrahydroxybutyl imidazole (THI) in 5% light corn syrup (Karo) in either PBS (pH 7.4) or water for 16-18 hours. The flies were placed in new, yeasted vials for 6-7 hours. Flies were then transferred to empty vials with access to a solution of 0.05 mg/ml THI in a 5% light corn syrup solution, as before, for 16-18 hours. This was repeated for a third time (or for 14 days for long term experiments). The flies were then placed in new, yeasted vials and if not analyzed immediately were transferred to new, yeasted vials daily before myofibril or activity analysis. The protocol described above was followed for THI oxime: (E)-1-(4-((1R,2S,3R)-1,2,3,4-tetrahydroxybutyl)-1H-imidazol-2-yl)ethanone oxime (LX2931, Lexicon Pharmaceuticals; a gift from Dr Buchynskyy and Prof. Gelb, University of Washington, WA, USA) as well as Fingolimod (FTY720) (Cayman). In these studies, both THI oxime and FTY720 were delivered at the same molar concentration as used for THI, 217 μM.

Two-tailed Student’s t-test was used to determine statistical significance for all experiments.

Drosophila Dystrophin (Dys) mutants show a variety of dystrophic phenotypes, including muscle, neuronal and heart defects (Shcherbata et al., 2007; Christoforou et al., 2008; Taghli-Lamallem et al., 2008). Here, we extend the muscle analysis to the level of the sarcomere (Perkins et al., 2010; Schnorrer et al., 2010). Two titin-like proteins, Projectin and Kettin (Sallimus – FlyBase), form the connecting filaments in indirect flight muscles (IFMs) that link the Z-bands to thick filaments and function in muscle elasticity during oscillatory work in flight (Lakey et al., 1990; Moore et al., 1999; Ayme-Southgate et al., 2005; Bullard et al., 2006). Because significant degradation of titin has been reported in muscle biopsies of DMD patients (Matsumura et al., 1989) and titin mutations cause muscular dystrophies and cardiomyopathies (Hackman et al., 2002; Udd et al., 2005), we analyzed the expression pattern of a titin-like protein, Projectin, in Dys mutant myofibrils. The data reveal that the Projectin pattern is the most sensitive marker for contraction-induced damage in Drosophila. Projectin is a large linker protein connecting thick filament myosin to thin filament actin, but the antibody used in these studies specifically recognizes an epitope close to the Z-line producing a sharp banding pattern spanning the entire width of the sarcomere on either side of the Z-line (Fig. 1A,B). Whereas the Projectin expression pattern is defective, other sarcomeric proteins, e.g. actin, tropomyosin and troponin C show normal expression patterns in dystrophic flies over the temporal length (5-15 days) of our analyses (Fig. 1; data not shown).

We used the sarcomere expression pattern of Projectin (Lakey et al., 1990; Ayme-Southgate et al., 2005) to assess myofibril integrity in DysDf (a deficiency that removes most of the Dys gene), Dys^det1 (a hypomorphic allele) and Dys^C-RNAi (TubGal4:UAS-Dys^C-RNAi/+) animals along with controls at two different time points. Projectin staining in normal myofibrils results in a uniform and intense band that spans the entire myofibril (Fig. 1A, first upper panel). This normal Projectin pattern was already somewhat disrupted in young Dys mutant myofibrils (Fig. 1A, second-fourth upper panels). However, the phenotype became significantly more penetrant and stronger in older animals (Fig. 1A, second-fourth lower panels). Phenotypes in young animals ranged from a diffuse Projectin staining on either side of the Z-line to a band that only partly stains the myofibril, whereas phenotypes in older (13- to 15-day-old) Dys mutant muscles ranged from highly reduced to punctate, or absent Projectin staining (Fig. 1A, second-fourth lower panels). In quantifying this phenotype at early and late time points (Fig. 1C), wild type is scored when Projectin continuously spans the entire width of the myofibril. However, qualitatively, it appears that we have defined an allelic series of Dystrophin mutants (DysDf/DysDf>DysDf/Dys^det1>Dys^det1/Dys^det1>Dys^C-RNAi/+) in which the deficiency is strongest, showing the greatest reduction of Projectin staining, and the RNAi mutant is weakest (Fig. 1A and Fig. 2A, middle panel). The dramatic defect in the pattern of a structural protein of the muscle contraction machinery in Dys mutants is consistent with contraction-induced damage over time and shows that the Dystrophin protein is an essential component in maintaining the stability of the molecular architecture of myofibrils. Collectively, these defective myofibrils eventually degrade and lead to a significant increase in fragmentation and degeneration in dystrophic compared with control transverse histological sections of IFMs (Fig. 1D; supplementary material Fig. S1A-E).

Previously, we have characterized wunen as a suppressor of the dystrophic wing vein phenotype (Kucherenko et al., 2008). Here, we tested whether reduced Wunen could suppress the muscle defects observed in Dys mutant animals. Unlike mammals, in which dystrophin is on the X chromosome, the gene is autosomic in flies so it is necessary to make the flies homozygous for Dys in addition to reducing the copy number of any gene of interest. However, as we had previously constructed single chromosome Dystrophin RNAi mutants (among which was TubGal4:UAS-Dys^C-RNAi) for screening, we used these animals to more rapidly assess the capacity of wunen mutations to suppress dystrophic muscle phenotypes. A significant suppression of the defective, punctate Projectin staining in Dys^C-RNAi (TubGal4:UAS-Dys^C-RNAi/+) myofibrils was observed when wunen was reduced (Fig. 2A; UAS-wun^RNAi/TubGal4:UAS-Dys^C-RNAi). We observed a fourfold increase in the number of wild-type myofibrils in Dys^C-RNAi flies, reaching 80% of control levels (Fig. 2C). Moreover, to determine whether this suppression was also observed with classical, DNA-lesion mutant alleles, we tested the capacity of wun^k10201, a P-element insertion allele, and wun⁴, an ethyl methanesulfonate (EMS)-induced allele, to suppress the myofibril phenotype seen in DysDf animals. We found that both wunen alleles suppressed the strong loss of Projectin phenotype associated with 5-day-old DysDf animals (wun/CyO; DysDf/DysDf; Fig. 2B,D). These data correlated with the wunen-based suppression of the gross morphological muscle defects seen in dystrophic transverse histological sections of IFMs (Fig. 2E-G) and shows that wunen reduction suppresses muscle defects in Dys mutants.

To determine whether wunen-dependent suppression of the dystrophic muscle phenotypes correlates with the result obtained with elevating S1P, we altered other known components of sphingolipid metabolism using available fly lines to see whether these changes could phenocopy wunen suppression. We tested two mutant backgrounds, which should increase S1P levels, for their ability to suppress dystrophic muscle phenotypes. It has been reported in mice that genetic reduction of S1P lyase, which irreversibly removes S1P from sphingosine metabolism by cleaving the compound to hexadecenaldehyde and phosphorylethanolamine, leads to a significant increase in S1P levels (Bagdanoff et al., 2009). Hence, we first constructed dystrophic animals heterozygous for Sply, the Drosophila S1P lyase gene (Fig. 3A). Interestingly, we found that, like wunen, reduction of Sply significantly suppressed the loss of Projectin in dystrophic myofibrils (Fig. 3B,C; Sply/+; TubGal4:UAS-Dys^C-RNAi/+). Also, like wunen, reduction of Sply suppressed the strong DysDf phenotype in 5-day-old animals (Fig. 3D,E). Additionally, like wunen, Sply reduction suppressed the muscle fragmentation phenotype observed in transverse histological sections of IFMs (Fig. 3G,I).

We next used the UAS-lace transgenic fly line (Adachi-Yamada et al., 1999) to test for suppression of muscle degeneration. We reasoned that, when combined with tubulin-Gal4, the UAS-lace transgene could theoretically increase S1P levels by upregulating the expression of lace, the gene encoding the catalytic subunit of the first protein in the de novo synthesis pathway, serine palmitoyl CoA transferase (Fig. 3A). We placed the UAS-lace-containing chromosome into the tubulin-Gal4-driven Dys^C-RNAi background and found that the Projectin expression pattern in myofibrils at 15 days showed a significant suppression of the dystrophic phenotype (Fig. 3B,C; UAS-lace/+;TubGal4:UAS-Dys^C-RNAi/+). In fact, quantitatively this was the highest level of suppression observed for which the percentage of wild-type myofibrils scored was not significantly different from non-dystrophic control flies (Fig. 3C; TubGal4/+). No suppression was observed in the control dystrophic animals (Fig. 3C; +/CyO; TubGal4:UAS-DysC-RNAi/+). Similar to reduced Sply, the fragmentation defect observed in histological sections of dystrophic flies was significantly suppressed when lace was overexpressed (Fig. 3H,I).

We used tubulin-driven RNAi mutants of Sphingosine Kinases 1 and 2 (SK1 and SK2, respectively) to assess whether reduction of S1P could exacerbate the muscle phenotypes in dystrophic flies. Reduction of kinase activity in wild-type backgrounds resulted in viability for SK1 and lethality for SK2. Interestingly, reduction of either kinase in dystrophic flies was lethal (Fig. 3J).

S1P signaling in mammals is much more complex than that observed in Drosophila. In mammals, S1P is both an extracellular and intracellular signaling molecule and has been shown to bind five different G protein-coupled receptors (Rosen et al., 2009). There do not appear to be S1P G protein-coupled receptors in flies, so it is reasonable to assume that S1P acts intracellularly as it does in the budding yeast Saccharomyces cerevisiae (Strub et al., 2010). Though Drosophila do not appear to have S1P receptors, they do have a putative S1P transporter, spinster. In fact, the vertebrate homolog that is named for spinster, spinster 2 (spns2) has been shown to specifically transport S1P in zebrafish and in human cells (Kawahara et al., 2009; Hisano et al., 2011). This was intriguing to us and we hypothesized that spinster could regulate S1P levels by transporting S1P between cellular compartments. Because S1P is generated intracellularly (Breslow and Weissman, 2010) and S1P lyase had been shown to function on the cytoplasmic side of the endoplasmic reticulum (Ikeda et al., 2004), we investigated whether reduction of spinster could elevate S1P levels in the cytoplasmic compartment and therefore phenocopy reduction of Sply in suppressing the dystrophic myofibril phenotype. We crossed the strong, genetic null spin^11F5 allele into the TubGal4:UAS-Dys^C-RNAi dystrophic background and assessed the Projectin content of myofibrils of flies at both 5 and 15 days of age (Fig. 4A,B). We found that reduction of spinster significantly suppressed the dystrophic myofibril phenotype at the 15-day time point (quantified in Fig. 4B).

Up to this point we have analyzed the structural and morphological phenotypes associated with muscle degeneration in dystrophic flies. To address S1P-based suppression of the dystrophic muscle function phenotypes seen in dystrophic animals, we utilized fly movement assays: climbing via a countercurrent apparatus (Greene et al., 2003; Shcherbata et al., 2007) and automated monitoring of overall activity (Fig. 5A).

Functionally, dystrophic flies have a reduced climbing ability compared with control flies (Shcherbata et al., 2007; supplementary material Fig. S1G). In this assay, the flies are challenged by being gently tapped to the bottom of a sequential series of tubes and their negative geotactic behavior is used to assess muscle function by measuring how far they climb in 30-second intervals (see Materials and methods). In addition to this assay, we used an experimenter-independent assay to test muscle function by recording the overall activity of animals over time using an automated monitoring system (Fig. 5A; Materials and methods). Using this system we show that the overall activity of dystrophic flies of genotypes w; DysDf/DysDf, w; Dys^det1/Dys^det1 and w; Dys^C-RNAi/+ when measured continuously over 6, 2 and 3 days, respectively, is significantly lowered compared with control w flies (Fig. 5B,C; supplementary material Fig. S1F). Activity of dystrophic flies was generally 50% lower than non-dystrophic control flies when tested over different durations of time.

To assess muscle function in dystrophic flies with reduced wunen, we used both the automated activity monitoring system and the countercurrent climbing apparatus. We generated animals with the single chromosome Dystrophin RNAi mutations that we had previously generated [TubGal4:UAS-Dys^C-RNAi described above and ActGal4:UAS-Dys^N-RNAi (Kucherenko et al., 2008)] combined with several wunen alleles (wun^k10201, wun^RNAi and wun⁴) to determine whether there was a suppression of dystrophic muscle function. Fig. 5D shows that all combinations of the Dystrophin mutants with different wunen alleles markedly increased activity compared with the Dystrophin mutants alone. Briefly, over the course of three days, 15-day-old dystrophic flies with reduced wunen were significantly more active (approximately twofold higher) than dystrophic flies alone. This correlated with the climbing analysis performed with the tubulin-Gal4 Dys^C-RNAi dystrophic animals with the wunen alleles mentioned above. This analysis showed that reduced wunen significantly increased the climbing ability of dystrophic flies (Fig. 5D). Additionally, we found that a viable 15-day-old wunen mutant alone (wun^EY11935) was significantly more active than wild-type (OR) flies. At this age, we see significant lethality with Dys^det1 mutant flies (supplementary material Fig. S2).

As with the suppression of the physical muscle defects described above, we investigated whether the genetic elevation of S1P through the reduction of Sply and overexpression of lace would manifest as increased activity in dystrophic animals. As with the muscle structural analysis, we again used the tubulin-Gal4-driven RNAi dystrophic animals to assess activity with reduced Sply. Automated monitoring analysis revealed a significant increase in activity when Sply was reduced in 12-day-old dystrophic flies (Fig. 5F). For the lace overexpression studies, we used both RNAi dystrophic animals (TubGal4:UAS-Dys^C-RNAi and ActGal4:UAS-Dys^N-RNAi) with UAS-lace in the background and found that with the transgene present 15-day-old dystrophic flies were significantly more active than flies without lace overexpression (Fig. 5G).

Finally, to determine whether reduction of spinster expression in dystrophic flies would also yield more active flies, we assayed 7-day-old flies and found that, like the S1P metabolism genes used above, dystrophic flies heterozygous for a strong spin allele showed a nearly twofold increase in activity (Fig. 5H).

In rodents, the genetic reduction of S1P lyase activity can be phenocopied pharmacologically by the delivery of the small molecule 2-acetyl-4(5)-tetrahydroxybutyl imidazole (THI) (Fig. 6F) (Schwab et al., 2005; Bagdanoff et al., 2009). We determined whether this small molecule would have the same effect in Drosophila by testing whether dystrophic mutant flies of differing genotypes (TubGal4:UAS-Dys^C-RNAi/+, an RNAi knockdown mutant; Dys^det1, a hypomorph; DysDf, a Dystrophin deficiency mutant; and Dys^det1/DysDf) given food supplemented with THI would show any suppression of the muscle degeneration phenotypes. After 3 days on a feeding regimen with THI (Fig. 6A), myofibrils from 7- to 10-day-old flies were harvested and analyzed, revealing a significant suppression of muscle wasting in dystrophic flies of all genotypes tested (Fig. 6B,C,E). Longer treatment also appears to be beneficial as after a 14-day feeding regimen we observe similar muscle wasting suppression in myofibrils from 15-day-old flies (supplementary material Fig. S3A,B). Even further suppression was observed when the feeding regimen of THI was combined with genetically reduced Sply (Fig. 6B; Sply/+; TubGal4:UAS-Dys^C-RNAi/+ with THI). Note that the flies with genetically reduced Sply that were not fed food supplemented with THI had more intact myofibrils, which is consistent with elevated S1P suppressing muscle wasting. In addition to muscle wasting suppression, we observe a functional benefit when feeding dystrophic flies food supplemented with THI. The activity analysis revealed that dystrophic flies fed solutions supplemented with THI were significantly more active than flies fed solutions not supplemented with THI (Fig. 6D).

To determine whether other small molecules, particularly drugs that either increase S1P levels or S1P signaling, would produce similar results to THI, we tested THI oxime (Fig. 6F), a candidate for the treatment of rheumatoid arthritis (Bagdanoff et al., 2010). As THI oxime is a derivative of THI we expected and found their beneficial effects to be similar in dystrophic muscles. Myofibrils isolated from 7-day-old Dys^det1 mutant flies after 3 days of THI oxime treatment showed significant suppression of Projectin disruption (Fig. 6G). Finally, we tested FTY720, an analog of S1P and potent elevator of S1P signaling (Napoli, 2000; Bagdanoff et al., 2009; Bagdanoff et al., 2010), for its effect on muscle wasting when fed to Dys^det1/DysDf flies. Using the same 3-day treatment, myofibrils isolated from 7-day-old flies revealed very strong suppression of the myofibril degeneration phenotype (Fig. 6H).

Here, we have established an easy-to-score myofibril phenotype for dystrophic flies that can even be scored in relatively young flies in strong Dystrophin mutants (Fig. 1). This assay complements the histological sections used previously to score gross morphological muscle degeneration. In using both assays, we have shown that a reduction of the LPP3 homolog wunen prevents, to a significant degree, Dystrophin-dependent muscle wasting (Fig. 2). Our data suggest that this is likely to occur through the increase of S1P levels as other avenues used to raise the level of S1P phenocopied this suppression. Both reduction of Sply, encoding the Drosophila S1P lyase, and overexpression of lace, encoding the catalytic subunit of serine palmitoyl CoA transferase in the de novo sphingolipid synthesis pathway, prevent muscle wasting (Fig. 3). We have also shown that muscle function as assayed by fly movement is also improved in dystrophic animals when S1P is genetically elevated (Fig. 5). Furthermore, we have shown that S1P-based suppression is not occurring during development as feeding experiments utilizing adult flies have revealed that S1P elevation in these animals can suppress dystrophic myofibril and activity phenotypes (Fig. 6)

Our data also suggest that minimal levels of S1P are necessary for viability in Drosophila as global reduction of Sphingosine Kinase 2 (SK2) results in lethality. Interestingly, global reduction of Sphingosine Kinase 1 (SK1) is not lethal in non-dystrophic flies yet is lethal in dystrophic flies (Fig. 3J), owing to exacerbation of the phenotype. These data indicate that S1P levels regulated by SK1 are crucial for Dystrophin mutant survival. These data argue for a precise requirement of minimal levels of S1P for muscle development and/or function. Future analyses of the activity of both sphingosine kinases in different tissues and cellular compartments might separate the roles of each kinase in homeostasis and muscle.

In mammals, there are five S1P receptors that share homology with G protein-coupled receptors (Rosen et al., 2009). Though flies do have G protein-coupled receptors, they do not appear to have the S1P receptors seen in vertebrates, suggesting that S1P receptor-mediated signaling might have evolved later in higher organisms. S1P lyase mutants increase intracellular S1P levels (Ikeda et al., 2004) and S1P is generated and has been shown to function inside the cells (Hait et al., 2009; Breslow and Weissman, 2010; Strub et al., 2010; Yonamine et al., 2011), indicating that the suppression of muscle wasting in Drosophila occurs intracellularly. With this in mind, we hypothesized that if spinster, like its mammalian homolog spns2, were an S1P transporter its reduction would prevent S1P from leaving the cytoplasmic compartment and it would then behave like reduced Sply and suppress muscle wasting. Our data support this hypothesis (Fig. 4) yet more work is required to connect this transporter to S1P in Drosophila. Interestingly, Drosophila spinster has been reported to interact with genes of the cell death pathway (Sakurai et al., 2010) and it is known that ceramide in the sphingolipid pathway can induce cell death. Perhaps spinster alters the cell death pathway by perturbing the equilibrium of sphingolipids, particularly S1P, in different subcellular compartments. A recent report has revealed S1P epigenetic regulation of gene expression through direct intracellular interaction with histone deacetylases (HDACs) (Hait et al., 2009). Through this mechanism, perhaps increasing intracellular S1P levels alters gene expression, which ultimately leads to elevated translation of muscle proteins, such as Projectin, which then reduces muscle wasting.

As mentioned, S1P has been shown to be necessary for the proliferation of satellite cells in mammals and is required for differentiation of myoblasts to myotubes (Bruni and Donati, 2008). As there do not appear to be canonical satellite cells in Drosophila, i.e. muscle precursor cells located on the surface of muscle fibers, perhaps S1P-based suppression of muscle wasting occurs as a result of the requirement of S1P for proper differentiation. It has been reported in Drosophila that sarcomeres are formed by an assembly of latent protein complexes (Rui et al., 2010). It would be consistent with this if S1P elevates muscle protein synthesis (in turn increasing the level of latent protein complexes) so that after muscle contraction-induced damage these complexes can assemble and produce new myofibrils. Our work with small molecule effectors of S1P signaling indicate that the abovementioned possibilities for suppression occur in actively contracting adult muscle.

S1P-based suppression of muscle wasting can be dissected further in Drosophila with the abundance of genetic tools available. Furthermore, given our results with THI, THI oxime and FTY720, Drosophila may be used to screen small molecules for their efficacy in suppressing muscle wasting.

We thank members of the Ruohola-Baker laboratory and Prof. Guenter Daum for helpful discussions throughout this work. We thank Prof. Leo Pallanck and Dr Susan McNabb for providing the climbing and activity apparatuses. We also thank Dr Andriy Buchynskyy and Prof. Michael Gelb for THI oxime.