Abstract
An important characteristic of vertebrate CNS development is the formation of specific amounts of insulating myelin membrane on axons. CNS myelin is produced by oligodendrocytes, glial cells that extend multiple membrane processes to wrap multiple axons. Recent data have shown that signaling mediated by the mechanistic target of rapamycin (mTOR) serine/threonine kinase promotes myelination, but factors that regulate mTOR activity for myelination remain poorly defined. Through a forward genetic screen in zebrafish, we discovered that mutation of fbxw7, which encodes the substrate recognition subunit of a SCF ubiquitin ligase that targets proteins for degradation, causes hypermyelination. Among known Fbxw7 targets is mTOR. Here, we provide evidence that mTOR signaling activity is elevated in oligodendrocyte lineage cells of fbxw7 mutant zebrafish larvae. Both genetic and pharmacological inhibition of mTOR function suppressed the excess myelin gene expression resulting from loss of Fbxw7 function, indicating that mTOR is a functionally relevant target of Fbxw7 in oligodendrocytes. fbxw7 mutant larvae wrapped axons with more myelin membrane than wild-type larvae and oligodendrocyte-specific expression of dominant-negative Fbxw7 produced longer myelin sheaths. Our data indicate that Fbxw7 limits the myelin-promoting activity of mTOR, thereby serving as an important brake on developmental myelination.
SIGNIFICANCE STATEMENT Myelin, a specialized, proteolipid-rich membrane that ensheaths and insulates nerve fibers, facilitates the rapid conduction of electrical impulses over long distances. Abnormalities in myelin formation or maintenance result in intellectual and motor disabilities, raising a need for therapeutic strategies designed to promote myelination. The mTOR kinase is a powerful driver of myelination, but the mechanisms that regulate mTOR function in myelination are not well understood. Our studies reveal that Fbxw7, a subunit of a ubiquitin ligase that targets other proteins for degradation, acts as a brake on myelination by limiting mTOR function. These findings suggest that Fbxw7 helps tune the amount of myelin produced during development and raise the possibility that Fbxw7 could be a target of myelin-promoting therapies.
Introduction
By tightly ensheathing axons, myelin membrane facilitates rapid conduction of electrical impulses and maintains axon health. In the CNS, myelin is produced by oligodendrocytes, glial cells that extend multiple membrane processes to wrap multiple axons. The unique physiological properties of myelin membrane result from combination of myelin-specific proteins and enrichment of certain lipids such as cholesterol. Therefore, the formation of myelin membrane requires mechanisms that coordinate the production of large amounts of myelin proteins and lipids.
One key driver of CNS myelination is the Akt and mTOR kinase signaling pathway. Pharmacological inhibition of mTOR in vitro using rapamycin reduced the expression of myelin gene transcripts and proteins (Tyler et al., 2009; Guardiola-Diaz et al., 2012) and diminished the levels of proteins necessary for cholesterol synthesis (Tyler et al., 2011). Consistent with this, spinal cords of mice in which mTOR gene function was inactivated in oligodendrocytes were hypomyelinated (Wahl et al., 2014). Conversely, oligodendrocyte-specific expression of constitutively active Akt, a serine/threonine kinase that can activate mTOR, drove formation of excess myelin (Flores et al., 2008). Hypermyelination was blocked by rapamycin, indicating that Akt promotes myelin formation by activating mTOR (Narayanan et al., 2009).
What factors influence mTOR signaling activity in myelination? Akt and mTOR are activated by signals transduced from receptor tyrosine kinases via formation of phosphatidylinositol 3,4,5-triphosphate (PIP3) from phosphatidylinositol 4,5 biphosphate (PIP2) by phosphoinositide 3-kinase (PI3K). Candidate activating pathways include those mediated by insulin growth factor (Carson et al., 1993; Goddard et al., 1999; Ye et al., 2002) and Neuregulin-1 (Brinkmann et al., 2008; Makinodan et al., 2012). Pathway activity is antagonized by the phosphatase and tensin homolog (PTEN), which converts PIP3 to PIP2, and oligodendrocyte-specific inactivation of PTEN caused hypermyelination (Goebbels et al., 2010; Harrington et al., 2010). Therefore, the amount of myelin formed in the CNS may be determined, in part, by the balance of positive and negative regulators of the PIP3/Akt/mTOR pathway. Additional regulators of pathway activity that modulate CNS myelination have not yet been described.
In a forward genetic screen in zebrafish, we identified a mutation disrupting Fbxw7, the substrate recognition subunit of a SCF ubiquitin ligase, which caused the formation of excess oligodendrocyte lineage cells (Snyder et al., 2012). We attributed specification of excess oligodendrocytes to elevated activities of Notch receptors, known targets of Fbxw7-mediated degradation (Hubbard et al., 1997; Gupta-Rossi et al., 2001; Oberg et al., 2001). We noted that fbxw7 mutant larvae appeared also to have ectopic and excess myelin gene expression. Because Notch signaling is inhibitory to myelin gene expression (Wang et al., 1998; Genoud et al., 2002; Givogri et al., 2002; John et al., 2002; Park and Appel, 2003), we reasoned that Fbxw7 must target a positive regulator of myelination for degradation. Because Fbxw7 targets mTOR for degradation in cancer cells (Mao et al., 2008), we hypothesized that Fbxw7 limits myelination by limiting the amount of mTOR available to promote myelination. Here, we present tests of this hypothesis and provide evidence that Fbxw7 modulates myelination by negatively regulating mTOR signaling activity.
Materials and Methods
Ethics statement.
The animal work in this study was approved by the Institutional Animal Care and Use Committees of the University of Colorado School of Medicine.
Zebrafish lines and husbandry.
Embryos were raised at 28.5°C in egg water or embryo medium and staged according to hours postfertilization (hpf), days postfertilization (dpf), and morphological criteria (Kimmel et al., 1995). Zebrafish lines used for this study included fbxw7vu56 (Snyder et al., 2012), mtorxu015Gt (Ding et al., 2011), and Tg(olig2:EGFP)vu12 (Shin et al., 2003).
In situ RNA hybridization and immunohistochemistry.
plp1a, mbp (Brösamle and Halpern, 2002), and cldnk (Takada and Appel, 2010) RNA probes were generated using digoxigenin RNA labeling kits (Roche). In situ RNA hybridization was performed as described previously (Hauptmann and Gerster, 2000). For immunohistochemistry, larvae were fixed using 4% paraformaldehyde, embedded, frozen, and sectioned using a cryostat microtome as described previously (Park and Appel, 2003). We used rabbit anti-Sox10 (Park et al., 2005), rabbit anti-Mbp (Kucenas et al., 2009), rabbit anti-phospho-S6 (Cell Signaling Technology, catalog #2215), mouse anti-GFP (Life Technologies, catalog #A-11120, 1:100), and anti-rabbit activated Caspase 3 (Abcam, catalog #13847) as primary antibodies. For fluorescent detection of antibody labeling, we used Alexa Fluor 568 goat anti-rabbit and goat anti-mouse conjugates (1:1000; Life Technologies). To stain myelin, we used FluoroMyelin Red (Life Technologies, catalog #F34652). In situ hybridization images were collected using a QImaging Retiga Exi color CCD camera mounted on a compound microscope and imported into Adobe Photoshop. Image manipulations were limited to levels, curve, and contrast adjustments. Fluorescence images were collected using a Zeiss Axiovert 200 microscope equipped with a PerkinElmer spinning disk confocal system and Volocity software and imported into Adobe Photoshop.
Genotyping.
mtor mutant embryos were genotyped by digesting a small piece of tail tissue in lysis buffer with 1 μg/μl proteinase K at 55°C, followed by PCR amplification using the following primers: forward primer: 5′-ATAAGAAAAGAAACCACATGTCATACC-3′; reverse primer: 5′-CTTACCACTCAGAGAGACCAAAG-3′; 5′LTR primer: 5′-CCCTAAGTACTTGTACTTTCACTTG-3′.
Plasmid construction.
The plasmid pEXPR-Tol2-sox10:DN-fbxw7-2A-EGFP-CaaX;Cryaa:Cerulean was constructed using the oligonucleotide primers 5′-TGAGGAGAGGAGACCTCAAGTC-3′ and 5′-CTTCATGTCCACGTCAAAGTCC-3′ to amplify from zebrafish cDNA a fbxw7 fragment encoding aa 419–761 of XP_005171005.1 (predicted F-box/WD repeat-containing protein 7 isoform X3). This fragment includes the last five coding exons and includes the WD40 repeat domain but excludes the F-box. This was cloned into the Gateway pME vector (Kwan et al., 2007) and then used to create a fusion proprotein with EGFP-CaaX following the viral 2A peptide linker, combined with an ∼8.2 kb fragment of sox10 regulatory DNA (Carney et al., 2006). The plasmid pEXPR-Tol2-sox10:EGFP-CaaX was also created with the Gateway system and the same sox10 fragment.
Quantitative PCR.
RNA was isolated from 20 pooled larvae for each control or experimental condition. RNA isolation for each experiment was performed in triplicate. Reverse transcription was performed using iScript Reverse Transcriptase Supermix (Bio-Rad, catalog #170-8840). Real-time qPCR was performed in triplicate for each cDNA sample using an Applied Biosystems StepOne Plus machine and software version 2.1. Taqman Gene Expression assays (Applied Biosystems) were used to detect mpz (Dr03131917_m1), 36k (Dr03438574_g1), hmgcs1 (Dr03107119_m1), fdps (Dr03424631_g1), hmgcra (Dr034228703_m1), and rpl13a (Dr03101115_g1) as the endogenous control. A custom-designed Taqman assay to detect mbp included the primers mbp-“A” forward, 5′GTTCTTCGGAGGAGACAAGAAGAG-3′; mbp “A” reverse, 5′-GTCTCTGTGGAGAGGAGGATAGATGA-3′; and mbp-“a” probe, 5′-AAGGGAAAGGGTTCATT-3′.
Drug inhibitor and rescue experiments.
Rapamycin (R-5000; LC Laboratories) was resuspended in dimethyl-sulfoxide (DMSO) to a stock concentration of 10 mm, aliquoted, and stored at −20°C. Embryos were treated from 72 to 96 hpf for qPCR experiments and from 3 to 7 dpf for Western blot analysis and incubated at 28.5°C in a water bath of 10 μm rapamycin and 0.1% DMSO in embryo medium. DMSO control groups were treated with 0.1% DMSO in embryo medium. For assessment of oligodendrocyte lineage cell number, fbxw7−/−;Tg(olig2:EGFG) larvae were imaged at 72 or 96 hpf on a confocal microscope using a 20× objective, placed individually in dishes, treated for 24 h with DMSO or rapamycin, and then imaged again. Dorsal olig2:EGFP+ cells were counted in confocal stacks spanning three or six somite hemisegments at the level of the trunk spinal cord.
EM.
At 8 dpf, zebrafish larvae were anesthetized with Tricaine and fixed in a solution of paraformaldehyde, glutaraldehyde, and sodium cacodylate, followed by osmium fixation using osmium tetroxide and sodium cacodylate as described previously (Langworthy and Appel, 2012). Reagents were purchased from Electron Microscopy Sciences. Sixty-nanometer sections were imaged on an FEI Technai Biotwin microscope with a Gatan Ultrascan camera.
Western blotting.
Protein lysates were prepared in RIPA buffer, protease inhibitors, and phosphatase inhibitors (Roche) from 7 dpf pooled larvae from each drug treatment and genotype. 20 μg samples were loaded into a 4–20% gel (Bio-Rad). After transfer, the membrane was blocked in 5% BSA in TBST, followed by overnight incubation at 4°C with rabbit anti-pS6 antibody at 1:1000 (Cell Signaling Technology, catalog #2215). Membranes were stripped and reprobed for total S6 using mouse anti-S6 antibody (Cell Signaling Technology, catalog #1217, 1:1000). Goat anti-rabbit HRP (Pierce, catalog #1858415) and goat anti-mouse HRP secondary antibodies (Pierce, catalog #1858413) were applied at 1:10,000 for 1 h at room temperature. Blots were developed using an ECL kit (Pierce, catalog #34076) and imaged on film.
Results
Fbxw7 inhibits myelin gene expression and myelination
We reported previously that expression of constitutively active Notch1a produced excess oligodendrocyte progenitor cells (OPCs) (Park and Appel, 2003). Consistent with this finding, we subsequently showed that fbxw7 mutant zebrafish larvae produce excess oligodendrocyte lineage cells as a consequence of elevated Notch signaling activity (Snyder et al., 2012). However, whereas expression of constitutively active Notch1a after OPC formation blocked myelin gene expression (Park and Appel, 2003), fbxw7 mutant larvae expressed myelin genes at a high level (Snyder et al., 2012). If Fbxw7 limits Notch signaling activity and Notch inhibits myelin gene expression, how do fbxw7 mutant larvae express myelin genes? One possibility is that Fbxw7 also inhibits activity of positive regulators of myelination and that, in the absence of Fbxw7, these factors overcome the inhibitory effect of Notch. To begin to test this idea, we first characterized myelination in fbxw7vu56 mutant larvae in greater detail. At 4 dpf, wild-type larvae expressed plp1a, cldnk, and mbp near the pial surface of ventral and dorsal spinal cord (Fig. 1A–C). The spatial distributions of transcripts in fbxw7 mutant larvae were mostly similar to that of wild-type (Fig. 1D–F) except that, as we noted previously (Snyder et al., 2012), cells within the medial portion of spinal cord sometimes expressed plp1a and cldnk (Fig. 1D,E; see Fig. 3B,F). In addition, although processed similarly and simultaneously, staining of hybridization products appeared more intense in mutant larvae compared with wild-type, suggesting that myelin genes were expressed at higher levels in the absence of fbxw7 function. qPCR supported this observation, revealing significantly higher levels of mbp and CNS-specific mpz and 36k transcripts (Fig. 1G). Because myelin membrane is enriched for particular lipids, especially cholesterol (Horrocks, 1967), we also measured expression of genes that encode enzymes of the cholesterol biosynthesis pathway. Similar to myelin protein-encoding transcripts, fdps, hmgcra, and hmgcs1 transcript levels were elevated in fbxw7 mutant larvae compared with wild-type (Fig. 1H). Consistent with these results, levels of Mbp, detected by immunohistochemistry, and myelin, detected by Fluoromyelin staining, appeared elevated in fbxw7 mutant larvae compared with wild-type (Fig. 1I–L). Furthermore, EM revealed that myelinated axons in the ventral spinal cord of 8 dpf fbxw7 mutant larvae averaged ∼2 more turns of myelin membrane than those of similar size in wild-type larvae (Fig. 1M–O). Because larger axons typically have thicker myelin, we also determined the cross-sectional area of myelinated axons in wild-type and fbxw7 mutant larvae. Myelinated ventral spinal cord axons of fbxw7 mutant larvae were larger than those of wild-type, averaging 0.73 μm2 compared with 0.49 μm2 (n = 67 axons in 3 larvae for each genotype; p = 0.0032, two-tailed Mann–Whitney test). Although the size distributions of axons overlapped in larvae of both genotypes, fbxw7 mutant larvae had fewer small axons and more large axons that were myelinated than wild-type (Fig. 1P). We conclude that fbxw7 mutant larvae produce excess oligodendrocytes accompanied by elevated expression of myelin genes and thicker myelin sheaths.
Fbxw7 limits myelination via negative regulation of mTOR
What myelin-promoting factors could be negatively regulated by Fbxw7? Among known Fbxw7 targets is mTOR (Mao et al., 2008), a positive regulator of myelination in rodent models (Tyler et al., 2009; Guardiola-Diaz et al., 2012; Bercury et al., 2014; Lebrun-Julien et al., 2014; Wahl et al., 2014). To investigate the possibility that Fbxw7 regulates mTOR in myelination, we first tested whether mTOR promotes myelination in zebrafish as it does in mice. To do so, we examined oligodendrocyte development and myelination in larvae homozygous for the mtorxu015Gt mutant allele, which nearly eliminates mTOR protein (Ding et al., 2011). We first assessed oligodendrocyte lineage cell formation and distribution by labeling tissue sections obtained from 4 dpf larvae with antibody to detect Sox10, which marks both OPCs and myelinating oligodendrocytes. Sox10+ cells occupied similar positions in wild-type and mtor mutant larvae (Fig. 2A,B), indicating that mtor function is not necessary for OPC specification and migration. However, 4 dpf mutant larvae had slightly fewer Sox10+ cells than wild-type larvae (Fig. 2C).
We next examined myelin gene expression using in situ RNA hybridization. In contrast to wild-type (Fig. 2D–F) and fbxw7 mutants (Fig. 1D–F), mtor mutant larvae appeared to express plp1a, cldnk, and mbp transcripts at a low level (Fig. 2G–I). Quantitative PCR confirmed these observations, revealing >2-fold reductions in transcript levels compared with wild-type (Fig. 2J). Because mTOR promotes cholesterol biosynthetic pathway gene expression via activation of the SREBP2 transcription factor, we also quantified fdps, hmgcs1, and hmgcra transcript levels and found them to be substantially lower in mtor mutant larvae relative to wild-type (Fig. 2K). Consistent with the low level of myelin gene transcripts, anti-Mbp and fluoromyelin staining revealed barely detectable levels of myelin proteins in mtor mutant larvae (Fig. 2L–O). Furthermore, EM revealed that myelin sheaths on CNS axons were thinner in mtor mutant larvae relative to axons of similar size in wild-type (Fig. 2P,Q), averaging 2.5 myelin membrane wraps compared with the wild-type average of 3.2 wraps (n = 3 larvae and 30 axons for wild-type and 3 larvae and 8 axons for mtor mutants; p < 0.0001, two-tailed unpaired t test; analysis limited to axons with a cross-sectional area of 0.401–3.0 μm2). Myelinated axons of mtor mutant larvae were smaller than those of wild-type larvae, averaging 0.29 μm2 compared with 0.49 μm2 (n = 3 larvae and 67 axons for wild-type and 3 larvae and 66 axons for mutant; p = 0.0002, two-tailed Mann–Whitney test). Analysis of size distribution revealed that mtor mutant larvae had more small, myelinated axons and fewer large, myelinated axons than wild-type (Fig. 2R). Altogether, these data indicate that mTOR is a positive regulator of myelination in zebrafish, as it is in rodents.
If the loss of Fbxw7 function results in excess myelination because of elevated mTOR activity, then genetic or pharmacological reduction of mTOR activity should reverse the hypermyelination phenotype of fbxw7 mutant larvae. To test this, we first examined oligodendrocyte lineage cell number in larvae lacking both Fbxw7 and mTOR functions. Whereas fbxw7 mutant larvae had an ∼2-fold excess of Sox10+ cells relative to wild-type, fbxw7−/−;mtor−/− mutant larvae had only a slight excess of oligodendrocyte lineage cells (Fig. 3A). fbxw7−/− mutant larvae treated with rapamycin from 3 to 4 dpf also had fewer Sox10+ cells than untreated mutants, but more than untreated wild-type larvae (Fig. 3A). To investigate the reason for this difference in number, we counted dorsally migrated olig2:EGFP+ oligodendrocyte lineage cells in living 3 dpf fbxw7−/− larvae, incubated them individually in rapamycin for 24 h, and recounted the cells. Whereas the number of oligodendrocyte lineage cells increased 42% in untreated control larvae, the number in rapamycin-treated larvae increased 15% (Fig. 3B). Treating fbxw7−/− larvae with rapamycin from 4 to 5 dpf did not change oligodendrocyte lineage cell number (Fig. 3B). To determine whether the smaller increase between 3 and 4 dpf resulted from cell death in rapamycin-treated larvae, we again incubated fbxw7−/− larvae with rapamycin from 3 to 4 dpf and performed immunohistochemistry to detect activated Caspase 3. This revealed no difference in oligodendrocyte lineage cell death in control and rapamycin-treated larvae (n = 10 larvae for each condition).
We also tested myelin gene expression in larvae lacking both Fbxw7 and mTOR functions. By in situ RNA hybridization, expression levels of plp1a, cldnk, and mbp in fbxw7−/−;mtor−/− mutant larvae appeared to be substantially reduced relative to fbxw7 single mutants and more similar to levels detected in mtor single mutants (Fig. 3B–M). In addition, double mutant larvae did not ectopically express plp1a and cldnk, providing evidence that the ectopic expression in fbxw7 mutant larvae results from misregulated mTOR activity. Consistent with these observations, immunohistochemistry revealed that Mbp expression also was much lower in fbxw7−/−;mtor−/− larvae than in fbxw7−/− larvae (Fig. 3N–Q).
To validate these observations, we performed qPCR using rapamycin treatments to interfere with mTOR activity. Wild-type larvae treated with rapamycin from 72 to 96 hpf expressed the myelin genes 36k and mbp at lower levels than wild-type controls, although mpz levels were not changed (Fig. 3S). fbxw7 mutant larvae treated with rapamycin expressed all three genes at lower levels than untreated mutant larvae, although the difference in mbp RNA levels did not reach statistical significance (Fig. 3S). We also tested expression of the cholesterol pathway genes fdps, hmgcra, and hmgcs1 and found that both wild-type and fbxw7 mutant larvae treated with rapamycin expressed these genes at a lower level than corresponding untreated control larvae (Fig. 3T). These data provide evidence that the elevated myelin and cholesterol pathway gene expression of fbxw7 mutant larvae requires mTOR function, consistent with the idea that Fbxw7 limits myelination by negatively regulating mTOR.
If Fbxw7 negatively regulates myelination by limiting the myelin-promoting function of mTOR, then fbxw7 mutant larvae should have elevated levels of mTOR signaling. As an initial test of this prediction, we performed Western blotting of whole larval protein extracts to detect the phosphorylated form of ribosomal protein S6 (phospho-S6), a target of mTOR kinase (Inoki et al., 2002). Compared with wild-type control larvae, phospho-S6 levels were low in extracts from mtor mutant larvae and in extracts from wild-type and fbxw7 mutant larvae treated with rapamycin (Fig. 4A), indicating that this method provides a reliable assessment of mTOR signaling activity. In contrast, phospho-S6 levels in fbxw7 mutant larvae appeared to be higher than in wild-type larvae (Fig. 4A). To investigate whether Fbxw7 regulates mTOR signaling in oligodendrocyte lineage cells, we performed immunohistochemistry to detect phospho-S6 in Tg(olig2:EGFP) larvae in which OPCs and oligodendrocytes express EGFP under control of olig2 regulatory DNA (Shin et al., 2003). phospho-S6 was evident in numerous cells of 4 dpf wild-type and fbxw7 mutant spinal cords (Fig. 4B–E′′). In contrast, wild-type larvae treated with rapamycin had little detectable phospho-S6 (Fig. 4F–F′′), validating this method as an assay of mTOR signaling activity. As an initial measurement of mTOR signaling in oligodendrocyte lineage cells, we determined the proportion of olig2:EGFP+ OPCs and oligodendrocytes labeled by anti-phospho-S6 antibody by examining single optical sections from confocal image stacks. Whereas ∼36% of olig2:EGFP+ cells were phospho-S6+ in wild-type larvae, this fraction was increased to ∼52% in fbxw7 mutant larvae and decreased to ∼9% in wild-type larvae treated with rapamycin (Fig. 4H). To determine whether the increased percentage of phospho-S6+ oligodendrocyte lineages cells in fbxw7 mutants resulted from mTOR activity, we treated them with rapamycin. This reduced the proportion of phospho-S6+ olig2:EGFP+ cells to 9%, similar to the fraction in wild-type larvae treated with rapamycin (Fig. 4G,H). We also calculated the relative fluorescence intensity of anti-phospho-S6 antibody labeling using similarly processed and imaged samples. Although the ranges of intensity values overlapped for individual cells in wild-type and fbxw7 mutant larvae, the average intensity was greater in mutant larvae than in wild-type and in wild-type larvae treated with rapamycin (Fig. 4I). We conclude that the loss of Fbxw7 function results in elevated mTOR signaling activity in oligodendrocyte lineage cells.
Fbxw7 function in oligodendrocytes limits myelin membrane growth
Our data show that loss of Fbxw7 function results in elevated levels of mTOR signaling in oligodendrocytes and elevated levels of mTOR-dependent myelination. As a direct test of whether Fbxw7 function in oligodendrocytes limits myelin membrane growth, we expressed either membrane-tethered EGFP (EGFP-CaaX) alone or a dominant-negative form of Fbxw7 (dnFbxw7) (Hubbard et al., 1997; Mao et al., 2008) and EGFP-CaaX under control of sox10 regulatory DNA, which drives expression in oligodendrocyte lineage cells. Injection into newly fertilized zebrafish eggs using the Tol2 transgenesis system results in expression by a subset of oligodendrocytes, permitting measurement of individual myelin sheath lengths as an indicator of myelin sheath growth (Fig. 5A–D). Oligodendrocytes expressing dnFbxw7 and EGFP-CaaX in wild-type larvae formed myelin sheaths that were, on average, ∼23% longer than those of oligodendrocytes expressing only EGFP-CaaX (Fig. 5A,B,E). In contrast, myelin sheaths marked by EGFP-CaaX expression were shorter in mtor mutant larvae (Fig. 5C,E), indicating that mTOR promotes sheath extension. To determine whether the excess length of myelin sheaths resulting from dominant-negative Fbxw7 expression requires mTOR function, we expressed dnFbxw7 in oligodendrocytes of mtor mutant larvae. Average sheath length was similar to that of oligodendrocytes in mtor mutants that expressed only the EGFP-CaaX reporter (Fig. 5D,E), consistent with the idea that Fbxw7 limits myelin sheath length by negatively regulating mTOR activity.
Discussion
Oligodendrocyte number, myelin thickness, myelin sheath length, and selection of specific axons for ensheathment are key variables in developmental myelination and remyelination after disease or injury. Identification of molecular mechanisms that regulate these variables is an important goal because mechanistic knowledge should provide insights into the basis of myelin disease and reveal potential targets for remyelinating therapy. Numerous positive and negative regulators of CNS myelination are known (Mitew et al., 2014; Wood et al., 2013), but how these are coordinated to control distinct features of myelination remains poorly understood. Here, we provide new evidence that the myelin-promoting function of mTOR kinase is restricted by the F-box protein Fbxw7.
mTOR activity coordinately controls protein translation and lipid synthesis and thereby promotes cell growth (Laplante and Sabatini, 2012). Therefore, mTOR is well suited to promote formation of large amounts of protein and lipid-rich myelin membrane. Accordingly, oligodendrocyte-specific mutation in mice of Mtor or Raptor, which encodes a subunit of the mTORC1 complex, reduced spinal cord myelination (Bercury et al., 2014; Wahl et al., 2014). Conversely, oligodendrocyte-specific expression of constitutively active Akt drove formation of excess myelin dependent on mTOR function (Flores et al., 2008; Narayanan et al., 2009). Together, these studies indicate that, although mTOR signaling is not solely responsible for CNS myelination, it may be rate limiting. Because hypermyelination resulting from excessive Akt/mTOR activity is pathogenic (Flores et al., 2008), mechanisms that limit pathway activity in oligodendrocytes might be an important feature of myelination control.
One well characterized negative regulator of mTOR pathway activity is PTEN, a lipid phosphatase that antagonizes PIP3/Akt signaling by converting PIP3 to PIP2. Oligodendrocyte-specific mutation of Pten in mice caused hypermyelination, which appeared to result primarily from increased myelin thickness and not from an increase in oligodendrocyte number (Goebbels et al., 2010; Harrington et al., 2010). Therefore, by limiting PIP3 accumulation, PTEN dampens Akt activation upstream of mTOR and thus limits the formation of myelin membrane.
Our work indicates that Fbxw7 provides an additional level of negative regulation of myelination by limiting both oligodendrocyte number and myelin synthesis. We reported previously that fbxw7 mutant larvae have a nearly 2-fold excess of oligodendrocyte cells, which we attributed to elevated Notch signaling (Snyder et al., 2012). The intracellular signaling domain of Notch is a known target of Fbxw7 for ubiquitin-mediated degradation (Gupta-Rossi et al., 2001; Oberg et al., 2001; Wu et al., 2001) and conditional expression of constitutively active Notch produced an excess of OPCs (Park and Appel, 2003) that was similar to that of fbxw7 loss of function. Time-lapse imaging showed that excess OPCs migrated from their ventral spinal cord origin but did not subsequently divide more frequently than OPCs in wild-type larvae, indicating that overproduction of OPCs in fbxw7 mutant larvae results from specification of excess neural precursors for oligodendrocyte fate rather than from excessive proliferation of oligodendrocyte progenitors (Snyder et al., 2012). Pharmacological inhibition of Notch signaling suppressed formation of excess OPCs in fbxw7 mutant larvae (Snyder et al., 2012), providing evidence that misregulated Notch signaling is primarily responsible for the excess OPC phenotype. In this study, we found that loss of mTOR function also had an effect on oligodendrocyte lineage cell number whereby both mtor mutant larvae and wild-type larvae treated with rapamycin had slightly fewer oligodendrocyte lineage cells than wild-type control larvae. Notably, loss of mTOR function had a more pronounced effect on fbxw7 mutant larvae, reducing oligodendrocyte lineage cell numbers to nearly normal levels. We think that these observations reveal a role for mTOR in OPC proliferation and not OPC specification. Treatment with rapamycin beginning at 72 hpf after OPC formation resulted in a smaller increase in cell number 1 d later than in control larvae. We found no evidence for a difference in cell death using activated Caspase 3 immunohistochemistry, suggesting that mTOR promotes OPC proliferation between 3 and 4 dpf. Consistent with this, mTOR can promote OPC proliferation in response to growth factors in culture (Gomez et al., 2015). In addition, Fbxw7 mutation sensitizes cancer cells to rapamycin, apparently by increasing the efficiency of cell killing by rapamycin (Mao et al., 2008). The mechanism by which this occurs is not known. In rodents, oligodendrocyte number is regulated by limiting amounts of growth factors, particularly PDGF (Calver et al., 1998; Fruttiger et al., 1999). We speculate that elevated mTOR function in fbxw7 mutant larvae enhances cell proliferation and possibly viability when growth factors are limiting, supporting an enlarged oligodendrocyte population.
Our work also provides evidence that Fbxw7 regulates myelin gene expression and myelin membrane growth by modulating mTOR signaling. Our observation that oligodendrocytes of fbxw7 mutant larvae express myelin genes was initially confounding because Notch signaling inhibits myelination (Wang et al., 1998; Genoud et al., 2002; Givogri et al., 2002; John et al., 2002; Park and Appel, 2003) and most known Fbxw7 targets function in cell proliferation (Welcker and Clurman, 2008). However, the identification of mTOR as a Fbxw7 target in cancer cells (Mao et al., 2008) and the mounting evidence that mTOR drives CNS myelination (Tyler et al., 2009; Bercury et al., 2014; Wahl et al., 2014) raised the possibility that elevated mTOR activity overcomes Notch-mediated inhibition of myelination in fbxw7 mutant larvae. In support of this, mTOR function was required for the elevated levels of myelin gene expression in fbxw7 mutant larvae, the mTOR signaling pathway was active at higher level in more oligodendrocyte lineage cells of fbxw7 mutant larvae relative to control larvae, fbxw7 mutant larvae formed more myelin membrane wraps on axons than wild-type, and oligodendrocytes expressing a dominant-negative Fbxw7 had longer myelin sheaths.
Our data also raise the possibility that Fbxw7 control of mTOR signaling influences the differentiation status of nonmyelinating OPCs. fbxw7 mutant larvae ectopically expressed the myelin-associated genes plp1a and cldnk in medial spinal cord, which is normally occupied by Sox10+ oligodendrocyte lineage cells that do not express markers of mature oligodendrocytes. Ectopic plp1a and cldnk expression was dependent on mTOR function because it was absent from medial spinal cords of fbxw7 mutant larvae treated with rapamycin or homozygous for a mtor mutation. One potential explanation for these observations is that Fbxw7 inhibition of mTOR helps to maintain the nonmyelinating state of a subset of oligodendrocyte lineage cells during development. However, we never detected ectopic expression of mbp in fbxw7 mutant larvae, indicating that reduction of Fbxw7 function is not sufficient for OPCs to fully differentiate as myelinating oligodendrocytes.
Could Fbxw7 tune mTOR activity to modulate the amount of myelin formed during development? Fbxw7 binds to its substrates at a conserved, phospho-threonine containing motif called the Cdc4 phospho-degron (CPD) (Nash et al., 2001). Substrate binding and degradation requires phosphorylation of the CPD, raising the possibility that developmentally controlled phosphorylation of the mTOR CPD regulates the myelin-promoting activity of mTOR. The central threonine of most, if not all, CPDs is phosphorylated by glycogen synthase kinase 3 (GSK3) (Welcker and Clurman, 2008). Therefore, GSK3-mediated phosphorylation of the mTOR CPD might suppress myelination. Notably, inhibition of GSK3β in rodents stimulated OPC proliferation, supported oligodendrocyte lineage cell survival, and promoted myelination (Azim and Butt, 2011). These effects were attributed to changes in cAMP response element binding and Notch signaling. Our study now indicates that GSK3β inhibition might also promote myelination by decreasing the amount of mTOR targeted for degradation by Fbxw7.
Footnotes
This work was supported by the National Institutes of Health (NIH Grant RO1 NS046668 to B.A., National Cancer Institute Grant 5T32CA08208613 to A.M.R.) and the Gates Frontiers Fund. The University of Colorado Anschutz Medical Campus Zebrafish Core Facility is supported by NIH Grant P30 NS048154. The electron micrographs were generated in the EM core facility of the Department of Cell and Developmental Biology, which is supported by NIH Grant P30 NS048154. We thank Xiaolei Xu (Mayo Clinic College of Medicine) for the gift of mtor mutant fish, Kristen Kwan (University of Utah) for the gift of Gateway plasmids, Dot Dill for EM technical support, and members of the Appel laboratory and Wendy Macklin (University of Colorado School of Medicine) for critical feedback on this project.
The authors declare no competing financial interests.
- Correspondence should be addressed to Bruce Appel, Department of Pediatrics, University of Colorado Anschutz Medical Campus, MS 8108, Aurora, CO 80045. bruce.appel{at}ucdenver.edu