Murine Fig4 is dispensable for muscle development but required for muscle function

Background Phosphatidylinositol phosphates (PIPs) are low-abundance phospholipids that participate in a range of cellular processes, including cell migration and membrane traffic. PIP levels and subcellular distribution are regulated by a series of lipid kinases and phosphatases. In skeletal muscle, PIPs and their enzymatic regulators serve critically important functions exemplified by mutations of the PIP phosphatase MTM1 in myotubular myopathy (MTM), a severe muscle disease characterized by impaired muscle structure and abnormal excitation–contraction coupling. FIG4 functions as a PIP phosphatase that participates in both the synthesis and breakdown of phosphatidylinositol 3,5-bisphosphate (PI(3,5)P2). Mutation of FIG4 results in a severe neurodegenerative disorder in mice and a progressive peripheral polyneuropathy in humans. The effect of FIG4 mutation on skeletal muscle has yet to be examined. Methods Herein we characterize the impact of FIG4 on skeletal muscle development and function using the spontaneously occurring mouse mutant pale tremor (plt), a mouse line with a loss of function mutation in Fig4. Results In plt mice, we characterized abnormalities in skeletal muscle, including reduced muscle size and specific force generation. We also uncovered ultrastructural abnormalities and increased programmed cell death. Conversely, we detected no structural or functional abnormalities to suggest impairment of excitation–contraction coupling, a process previously shown to be influenced by PI(3,5)P2 levels. Conditional rescue of Fig4 mutation in neurons prevented overt muscle weakness and the development of obvious muscle abnormalities, suggesting that the changes observed in the plt mice were primarily related to denervation of skeletal muscle. On the basis of the ability of reduced FIG4 levels to rescue aspects of Mtmr2-dependent neuropathy, we evaluated the effect of Fig4 haploinsufficiency on the myopathy of Mtm1-knockout mice. Male mice with a compound Fig4+/−/Mtm1–/Y genotype displayed no improvements in muscle histology, muscle size or overall survival, indicating that FIG4 reduction does not ameliorate the Mtm1-knockout phenotype. Conclusions Overall, these data indicate that loss of Fig4 impairs skeletal muscle function but does not significantly affect its structural development.

We recently identified and characterized a spontaneous mouse mutant ("pale tremor" or plt) with a homozygous recessive loss-of-function mutation in the Fig4 gene [12]. plt mice exhibit severe and progressive neurodegeneration that involves both neurons and glia of the central and peripheral nervous system [13]. The main subcellular phenotype is increased vacuolization due to defective lysosomal function and impaired autophagy [14,15]. In addition, we (the Meisler group) have identified recessive FIG4 mutations in patients with hereditary peripheral neuropathy (Charcot-Marie-Tooth Disease type 4J) [16][17][18] and motor neuron disease (amyotrophic lateral sclerosis, or ALS) [19]. However, the role of FIG4 in skeletal muscle has yet to be extensively examined in detail.
Studies of myotubularins indicate a potentially important role for PI(3,5)P 2 regulation in skeletal muscle [20][21][22]. The mammalian myotubularin gene family contains 15 members encoding phosphatases that primarily dephosphorylate PI(3)P and PI(3,5)P 2 [23]. Mutation of MTM1, the canonical member of this gene family, results in increased levels of PI(3)P, and possibly PI(3,5)P 2 , in skeletal muscle and causes myotubular myopathy (MTM), a severe congenital muscle disease associated with altered muscle structure and profound muscle weakness [24]. Mutation of MTM1 compromises multiple aspects of muscle function [25], most notably excitation-contraction coupling (EC coupling) [26], the process by which neuronal stimulation to muscle is translated into calcium-dependent muscle contraction. Specifically, MTM1 mutations lead to severe abnormalities in the structure of the EC coupling machinery [27].
MTMR14 is another myotubularin family member that is important for muscle function [26]. Loss of MTMR14 function in zebrafish and mice causes aberrant autophagy and impaired EC coupling [26,28]. Unlike MTM1, loss of MTMR14 alters EC coupling without significantly changing the underlying structure of the EC coupling machinery. Direct application of PI(3,5)P 2 increases calcium release from microsomes containing the intracellular ryanodine receptor 1 calcium release channel (RyR1), a critical component of the EC coupling apparatus. Thus, direct regulation of RyR1-dependent stimulated calcium release may represent one important role of PI(3,5)P 2 and MTMR14 [29].
Given the potential importance of PI(3,5)P 2 in skeletal muscle, we sought to understand the impact of FIG4 mutation on muscle development and homeostasis. To address this issue, we examined skeletal muscle structure and function in the plt-null mouse model of FIG4 dysfunction. We examined plt mice, which die before 2 months of age as a result of progressive neurodegeneration, as well as plt mice with restored FIG4 expression in neurons [12,15]. The latter mice have no overt phenotype and survive for more than 18 months. Our data reveal that Fig4 mutation is associated with skeletal muscle changes (atrophy and increased apoptosis) and impaired muscle force generation, but not with abnormalities in the structure or function of the EC coupling machinery. The changes are likely the consequence of impaired neuronal input because phenotypic rescue is largely provided by neuronal expression of FIG4. In addition, we found that haploinsufficiency of Fig4 does not ameliorate effects of Mtm1 mutation in muscle, in contrast to a previous report that haploinsufficiency of Fig4 rescues the neuropathy associated with Mtmr2 mutation [30,31]. Together, our results support a requirement for FIG4 in skeletal muscle function, but not a cellular autonomous role in either muscle development or EC coupling.

Animal care and husbandry
All animals were cared for per protocol under the guidance of, and with ethical approval from, the University Committee on Use and Care of Animals (UCUCA) and with the assistance of members of the University of Michigan's Unit for Laboratory Animal Medicine (ULAM), who carefully monitored the health of the rodent colonies. ULAM maintained proper environmental regulation, including temperature and light cycles, unlimited access to water, appropriate food supply and clean enclosures. The Fig4-null mutation plt is maintained on two congenic lines, C57BL/6J.plt/+ (N14) and C3H.plt/+ (N10) [32]. Experiments were carried out on homozygous plt/plt F1 mice obtained from crosses between the two congenic strains. Pups were weaned according to standard protocols, and tails were clipped for genotyping.

Western blot analysis
Western blot analyses were performed using the following antibodies: FIG4 (1:1,000 NeuroMab; UC Davis/NIH NeuroMab Facility, Davis, CA, USA) and glyceraldehyde 3-phosphate dehydrogenase (1:1,000 GAPDH; Millipore, Billerica, MA, USA). Mouse multitissue Western blot antibody was obtained from IMGENEX (San Diego, CA, USA). Protein extracts were established from flashfrozen mouse skeletal muscle and brain using T-PER tissue protein extraction reagent and a Dounce tissue homogenizer (Pierce Biotechnology, Rockford, IL, USA). Approximately 50 μg of protein were loaded per sample, resolved by polyacrylamide gel electrophoresis on 11% gels, and transferred to polyvinylidene fluoride. Secondary antibodies were used at 1:2,000 (Santa Cruz Biotechnology, Santa Cruz, CA, USA), blots were developed using electrochemiluminescence reagent (GE Biosciences, Pittsburgh, PA, USA), and bands visualized using the Bio-Rad ChemiDoc XRS+ System illuminator (Bio-Rad Laboratories, Hercules, CA, USA).

Histopathology
Animals were killed by anesthetic injection, which was followed by cervical dislocation. Tissues were then isolated using sterile surgical methods without the use of laminar flow hoods. Muscle tissue from quadriceps and tibialis anterior muscles was dissected and mounted onto small balsa wood pieces that had previously been frozen with drops of Tissue-Tek O.C.T. compound (Sakura Finetek USA, Torrance, CA, USA) and then semithawed with light friction. The mounted muscle tissue was immediately submerged in a −55°C isopentane bath cooled by liquid nitrogen for flash-freezing.
Muscles were cut into 12-μm cross-sections and mounted on Superfrost Plus slides (Thermo Scientific, Waltham, MA, USA) using a Leica cryostat (Leica Biosystems, Buffalo Grove, IL, USA) at −20°C and dried at room temperature before storage at −80°C. Slides were stained with Mayer's hematoxylin and eosin (H&E) or succinate dehydrogenase (SDH) following standard protocols and mounted with Permount mounting medium (Thermo Scientific). Apoptotic fibers were visualized with the ApopTag Plus Peroxidase In Situ Apoptosis Kit (S7101; Chemicon International/EMD Millipore, Billerica, MA, USA). Photomicrographs were captured using an INFIN-ITY1 digital camera with eponymous software (Lumenera Corp, Ottawa, ON, Canada) visualized through an Olympus BX43 light microscope (Olympus America, Center Valley, PA, USA).

Ultrastructural analysis
Immediately following dissection, quadriceps and gastrocnemius muscles were carefully cut into approximately 1-mm × 2-mm fragments and incubated in Karnovsky's fixative overnight at 4°C. Fixed tissue was brought to the Microscopy and Imaging Laboratory (MIL) Core facility at the University of Michigan for processing. Ultrathin sections were analyzed for orientation, and grids were prepared for use on the Philips CM-100 transmission electron microscope (Koninklijke Philips N.V., Amsterdam, The Netherlands).

Myocyte isolation
Muscle was dissected from the shoulders and legs of dead mice and placed immediately into sterile phosphatebuffered saline. Muscle was then minced finely with a sterile razor, fully dissociated with a mixture of collagenase type I (0.1%) and trypsin (0.1%) in Ham's F-12 medium, then incubated at 37°C for approximately 1 h with periodic trituration. Cells were pelleted and resuspended in 1:1 Dulbecco's Modified Eagle Medium:Ham's F-12 Nutrient Mixture (DMEM/F-12) (Gibco/Life Technologies, Grand Island, NY, USA) with 20% fetal calf serum (HyClone Laboratories, Logan, UT, USA), then filtered through 70-and 40-μm meshes and plated onto collagen-coated dishes (BD Biosciences, San Jose, CA, USA). Media were changed after a 1-h incubation at 37°C, and recombinant human fibroblast growth factor-basic (AA 10-155, Publication PHG0026; Gibco/Life Technologies) was added to a final concentration of 10 ng/ml. Cells were maintained at 37°C in a 5% CO 2 atmosphere with daily changes of fresh media. Cells were visualized using Hamamatsu ORCA-R2 camera and software (Hamamatsu Photonics, Hamamatsu-shi, Japan) on a Leica inverted microscope (Leica Microsystems). All experimentwas were performed on passages 2 through 4 myocytes. Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) staining was performed per the manufacturer's recommendations on myocytes grown on coverglass coated with fibronectin and fixed with 4% paraformaldehyde prior to staining.

Muscle force measurement
Extensor digitorum longus (EDL) and soleus muscles were carefully isolated and removed from anesthetized mice. Muscles were immediately placed into a bath of Krebs mammalian Ringer solution with 0.25 mM tubocurarine chloride maintained at 25°C and bubbled with 95% O 2 and 5% CO 2 to stabilize pH at 7.4. Using 5-0 silk suture, the distal tendon of the muscle was attached to a servomotor (model 305B; Aurora Scientific, Aurora, ON, Canada), and the proximal tendon was attached to a force transducer (model BG-50; Kulite Semiconductor Products, Leonia, NJ, USA). Muscles were stimulated by square pulses delivered by two platinum electrodes connected to a high-power biphasic current stimulator (model 701B; Aurora Scientific). A personal computer running custom-designed software (LabVIEW 7.1; National Instruments, Austin, TX, USA) controlled electrical pulse properties and servomotor activity and recorded data from the force transducer. Stimulation voltage and optimal muscle length (L o ) were adjusted to give maximum twitch force [33]. While held at L o , muscles were subjected to trains of pulses 300 ms in duration for EDL muscles and 900 ms for soleus muscles, with increasing stimulation frequency until maximum isometric tetanic force (P o ) was achieved [33]. L o was measured with digital calipers, and muscle fiber lengths (L f ) were determined by multiplying L o by previously established L f -to-L o ratios of 0.44 for EDL muscle and 0.71 for soleus muscle [33]. Total muscle fiber cross-sectional area (CSA) was estimated by dividing the mass of the muscle by the product of L f and 1.06 g/cm 3 , the density of mammalian skeletal muscle. P o was normalized by CSA to give specific P o .
Simultaneous measurement of macroscopic L-type Ca 2+ currents and voltage-gated Ca 2+ transients in myotubes Primary cultures of skeletal myotubes were generated from myoblasts derived from wild-type (WT) and plt mice as previously described [34]. The whole-cell voltage-clamp technique in conjunction with a Ca 2+ -sensitive dye (fluo-4) was used to simultaneously measure voltage-gated L-type Ca 2+ currents (L-currents) and intracellular Ca 2+ transients on individual myotubes from 8-to 11-day-old myotube cultures [34,35]. All voltage-clamp experiments were carried out after an approximately 5-min period of dialysis following establishment of the whole-cell configuration. A 1-s prepulse to −30 mV delivered immediately before each test pulse was used to inactivate voltage-gated Na + and T-type Ca 2+ channels without producing significant L-channel inactivation. L-currents and Ca 2+ release were subsequently elicited by 200-ms test depolarizations from −50 mV to +70 mV in 10-mV increments and a 10-s interval between each test pulse. Capacitative currents were minimized to about 10% using the capacitance cancelation feature of the patch-clamp amplifier. Remaining linear components were leak-subtracted using a P/3 protocol delivered from a holding potential of −80 mV before each test pulse. Peak L-current magnitude was normalized to cell capacitance (pA/pF), which was plotted as a function of membrane potential (V m ) and fitted as where G max is the maximal L-channel conductance, V m is test potential, V rev is extrapolated reversal potential, V G1/2 is the voltage for half-maximal activation of G max , and k G is a slope factor. Relative changes in intracellular Ca 2+ during each test depolarization were measured following dialysis with K 5fluo-4 salt. Fluo-4-dialyzed myotubes were excited at 480 nm and fluorescence emission measured at 535 nm was digitized at 10 kHz. A computer-controlled shutter was used to eliminate dye illumination during intervals between each test pulse. Relative peak changes in intracellular Ca 2+ were expressed as ΔF/F ([F peak -F base ]/F base ) at the end of each test pulse, plotted as a function of V m , and fitted according to the equation where (ΔF/F) max is the calculated maximal change in fluorescence, V F1/2 is the voltage for half-maximal activation of (ΔF/F) max , and k F is a slope factor. Pooled current-voltage (I-V) and fluorescencevoltage (ΔF/F-V) data were expressed as means ± SEM. Statistical significance was determined using a two-tailed Student's t-test.

Statistical analyses
GraphPad Prism software (GraphPad Software, La Jolla, CA, USA) was used to calculate the significance of fiber size differences and TUNEL staining results with unpaired Student's t-tests and one-way analysis of variance with Tukey's post hoc multiple comparison test.

FIG4 is expressed in skeletal muscle
We wanted to establish that FIG4 protein was present in skeletal muscle. We therefore performed Western blot analysis using anti-FIG4 antibody (NeuroMab). We first used a premade multitissue Western blot antibody (IMGENEX) and found that FIG4 was present in most tissues examined, including skeletal muscle ( Figure 1A). To determine if FIG4 expression was regulated with maturation of skeletal muscle, we examined protein extracts from skeletal muscle of various postnatal ages ( Figure 1B) as well as from differentiating C2C12 myocytes ( Figure 1C). We detected no obvious variability in FIG4 levels at different mouse ages or different developmental stages in vitro; thus we conclude that FIG4 is a ubiquitously expressed component of skeletal muscle. Of note, we verified that FIG4 expression was absent in Fig4-null (that is, plt) skeletal muscle as compared to WT littermates ( Figure 1D).

Fig4-null myofibers are reduced in size and exhibit ultrastructural abnormalities
We began our analysis by examining both light and electron microscopic features of skeletal muscle from Fig4null mice. We examined muscle at 5 weeks of age, since these mice do not survive beyond 6 weeks of age [32]. We also studied a limited number of animals with expression of a Fig4 cDNA transgene under control of the neuron-specific enolase (NSE) promoter [15]. These animals were examined at 4, 8 and 20 months of age.
Routine histopathological analysis of quadriceps, gastrocnemius and diaphragm using H&E and SDH stains did not reveal any obvious abnormalities in plt skeletal muscle compared to the WT (Figures 2A through 2D). However, the CSA of plt myofibers was significantly smaller than that of their WT counterparts. Quantitation of cross-sectional fiber area revealed that plt myofibers were 55.3% smaller than WT fibers (n = 400 total fibers counted from quadriceps in three mice per condition) ( Figure 2E). This reduction in size corresponds with the observation that overall quadriceps weight in plt animals was similarly significantly reduced compared to that of age-matched WT littermates (Additional file 1: Figure S1). Of note, overall body weight (which is largely dictated by muscle weight) at 6 weeks of age was also significantly reduced in plt mice (Additional file 1: Figure S2).
Ultrastructural analyses were performed on quadriceps muscle from age-matched WT and plt mice using transmission electron microscopy (n = 4 per condition). Overall, muscle ultrastructure in plt mice resembled that of controls ( Figures 3A and 3B), though swollen and enlarged mitochondria were occasionally observed in muscle from plt mice (M in Figure 3C). We did not, however, detect consistent abnormalities in the triad, the location of the EC coupling machinery, in any plt mouse examined (arrow in Figure 3B).

Increased apoptosis in skeletal muscle and primary myocytes from Fig4-deficient mice
The combination of reduced muscle fiber size and ultrastructurally abnormal mitochondria led us to examine whether there was increased cell death in plt muscle. TUNEL staining of muscle sections revealed few positive cells in the WT tissue but several positive cells in the mutant tissue ( Figure 4A). To better evaluate this finding, we studied isolated myocytes from neonatal WT and plt mice. As with primary cells from other organ systems of the plt mice, cells derived from plt muscle demonstrated abundant vacuolization (Additional file 1: Figure S3). We measured cell death in these primary cell cultures using a TUNEL assay. We observed a significant increase in TUNEL-positive cells in Fig4mutant myocytes, consistent with decreased cell survival ( Figure 4B). Of note, despite the vacuolization and impaired survival, Fig4-null myocytes were able to successfully differentiate into myotubes upon serum withdrawal (data not shown). There was also no change in the ability of plt myocytes to proliferate, as deter-

FIG4-mutant skeletal muscle has impaired force generation
To determine the potential functional effects of Fig4 deficiency in skeletal muscle, we measured force generation from intact muscle fibers. Testing both EDL and soleus muscles revealed statistically significant reductions in specific force generation (12% to 24% decrease when adjusted for CSA), indicating mild muscle weakness and impaired force generation in the plt animals ( Figure 5).

FIG4-mutant myocytes have normal intracellular calcium dynamics
Given the previous association between PI(3,5)P 2 and EC coupling and the fact that FIG4 is a known regulator of PI(3,5)P 2 levels, we interrogated intracellular calcium dynamics. We used whole-cell voltage patch-clamping of isolated neonatal mouse myocytes to simultaneously measure voltage-gated L-type calcium currents and intracellular calcium transients. We did not detect abnormalities in either orthograde or retrograde coupling or in L-type  calcium currents in plt myocytes ( Figure 6 and Table 1). Taken in conjunction with the normal histological appearance of the triad, these data suggest that loss of Fig4 is not associated with impaired bidirectional dihydropyridine receptor (DHPR)-RyR1 coupling.

Neuronal expression of Fig4 in plt mice reduces skeletal muscle pathology
We previously reported that expression of FIG4 under the NSE promoter in plt animals prevents the development of typical phenotypic abnormalities [15]. In fact, the appearance of the transgenic (Tg) animals is indistinguishable from their WT littermates. We examined skeletal muscle in  Figure 7C) and did not exhibit evidence of increased apoptosis.
Last, to determine whether muscle function was altered in Fig4 −/− /TgNSE mice, we measured maximum isometric force in soleus and EDL muscles of 18-month-old animals.
There was a small (approximately 10%) but statistically insignificant (P > 0.1) decline in force generation between Fig4 −/− /TgNSE mice and their WT littermates (Additional file 1: Figure S4). Taking all our data together, we conclude that expression of Fig4 in neurons is sufficient to correct the development of significant muscle atrophy and programmed cell death in plt skeletal muscle, but not to Figure 5 Diminished force production in pale tremor (plt) skeletal muscle. Specific force (normalized to muscle and body weight) was measured for wild-type/littermate (WT) and plt (mutant) skeletal muscle (left panel = extensor digitorum longus, or EDL) (right panel = soleus). Normalized force was significantly reduced in plt muscles (n = 7 animals tested per condition). *P = 0.043 for EDL and P < 0.001 for soleus. prevent subtle but reproducible changes in myofiber size in vivo. Of note, we verified by both RNA (reverse transcriptase polymerase chain reaction) and Western blot analysis that there was no appreciable Fig4 RNA or protein in muscle of Fig4 −/− /TgNSE mice (data not shown and Figure 1D).

Fig4 haploinsufficiency does not improve the Mtm1-knockout mouse phenotype
There is a complex interplay between the enzymes that regulate phosphoinositides. One potential avenue for treatment of diseases related to PIP dysregulation is manipulation of other enzyme levels. This point has been demonstrated for the neurological abnormalities associated with Mtmr2 mutation [30]. Mice with a recessive mutation in Mtmr2 exhibit peripheral neuropathy, and aspects of this neuropathy are reversed in the setting of Fig4 haploinsufficiency. Since MTM1 and MTMR2 are highly homologous, we evaluated the effect of haploinsufficiency of Fig4 (in plt +/− mice) on the severe muscle phenotype in knockout mice lacking expression of Mtm1. Generation of Fig4 +/− /Mtm1 -/Y male mice revealed no difference from Mtm1 -/Y mice, with comparable impairments of weight gain, motor function and survival and no statistically significant difference in weight or survival (Figures 8A  and 8B).
By histopathological analysis, we demonstrated similar alterations between Mtm1 -/Y mice and Mtm1 -/Y /Fig4 +/− mice. We quantitated myofiber size because size is correlated with disease severity in patients with MTM [36]. There was no clear difference in this parameter ( Figure 8B, consistent with the observed lack of functional improvement in the Fig4 +/-/Mtm1 -/Y animals. In all, haploinsufficiency of Fig4 did not appear to alter the Mtm1 phenotype in any meaningful way.

Discussion
FIG4 exhibits an important role in central and peripheral nervous system development and homeostasis. The consequences of FIG4 deficiency on other organ systems are less well delineated. In this study, we defined the consequences of Fig4-null mutation on skeletal muscle. We found that (1) global inactivation of Fig4 in the plt mouse is associated with myofiber atrophy and/or hypotrophy, increased apoptosis and diminished specific force production; (2) loss of   abnormalities. The implications of these findings are discussed below. Perhaps the most striking aspect of this study is the fact that Fig4 mutation did not result in a more deleterious direct effect on muscle development and function. In neurons and glia, loss of FIG4 results in severe structural and functional consequences. Given the previously recognized importance of phosphoinositide regulation in skeletal muscle [26,27], we predicted that plt mice would exhibit pronounced changes in muscle structure and function. However, effects of Fig4 ablation on muscle were minimal and largely rescued following neuron-specific restoration of FIG4 expression (Fig4 −/− /TgNSE mice). Muscle alterations found to persist in Fig4 −/− /TgNSE mice (for example, 11% reduction in fiber size) could be explained by either a minor role for FIG4 expression in muscle fiber size determination or a small amount of residual neurogenic atrophy. It is of interest to note the dichotomy between the relatively normal appearance of adult plt muscle in situ and the extensive vacuolarization of plt myocytes in vitro. Similar effects of culturing have been observed in other cell types, including fibroblasts [12] and osteoblasts (unpublished manuscript, Lenk GM and Meisler MH). The explanation for this difference is unclear and requires further experimentation.
There are several potential explanations for the lack of a more severe phenotype in skeletal muscle of Fig4-null mice. First of all, the partial reduction of PI(3,5)P 2 likely caused by loss of Fig4 may not be enough to result in significant consequences for muscle development and function. Alternatively, a different PIKfyve protein complex may compensate for the loss of FIG4 to provide sufficient levels of PI(3,5)P 2 to maintain myofiber homeostasis. A third possibility is that PI(3,5)P 2 may be generated in muscle by a different or complementary pathway. For example, a three-position kinase may generate PI(3,5)P 2 from PI(5)P. One current barrier to attempting to distinguish these possibilities is the lack of suitable approaches to measure PI(3,5)P 2 in whole tissues such as skeletal muscle. Development of techniques to aid in measuring PI(3,5)P 2 in situ are required to answer these questions more definitively.
Regardless of the explanation, the fact that there is little, if any, muscle-cell autonomous phenotype in Fig4null animals indicates that FIG4 mutations are unlikely to result in primary muscle disease. However, significant secondary neurogenically mediated myopathic features, such as those observed in plt mice, including reduced muscle fiber CSA and specific force generation, suggest that skeletal muscle changes may contribute to disease pathogenesis. In other words, the myopathic changes described herein may influence disease severity in patients with Charcot-Marie-Tooth disease type 4J and other disorders caused by FIG4 gene mutation.
The lack of impairment in the structure of the EC coupling apparatus or of bidirectional triad coupling in Fig4-null mice was unexpected. Data from the MTMR14knockout mice support the hypothesis that increased levels of PI(3,5)P 2 impair calcium release from the ryanodine receptor (the core component of the EC apparatus), though the underlying mechanism is not clear. Our data imply that reduced levels of PI(3,5)P 2 do not acutely impair voltagegated triad calcium release. Furthermore, the chronic loss of FIG4 from muscle (with the potential implication of chronically reduced PI(3,5)P 2 levels) does not alter the ultrastructural appearance of the triad (that is, the location of the EC coupling machinery). Thus, a requirement for normal levels of PI(3,5)P 2 for EC coupling seems unlikely, though the present data do not completely exclude this possibility. In addition to more direct interrogation of EC coupling in FIG4-deficient mice, another potential future direction to address this issue would be to assess the impact on EC coupling of muscle-specific knockout of PIKfyve, the kinase required for PI(3,5)P 2 generation. Again, however, this would necessitate confirmation of a specific reduction in PI(3,5)P 2 levels in skeletal muscle.
The final significant observation in this study is that reduced Fig4 expression via plt haploinsufficiency does not significantly alter the phenotype of Mtm1-knockout mice. Of note, the MTMR2 gene encodes a protein that is highly homologous to MTM1 [8]. In addition, we previously demonstrated that zebrafish mtmr2 functionally compensates (at least in part) for loss of mtm1, suggesting that MTMR2 and MTM1 are functionally quite similar [37]. However, in contrast to our findings in Mtm1-null mice, Bolino and colleagues found that plt haploinsufficiency rescued neuropathy in Mtmr2-knockout animals [30]. The reasons why reduction of FIG4 levels improved the MTMR2-related neuropathology, but not the muscle pathology, seen in Mtm1-null mice are not clear. This distinction may provide another indication of the nonessential role of FIG4 in skeletal muscle or may reflect different quantitative requirements for PI (3,5)P 2 in neurons and muscle or specific differences between mammalian MTMR2 and MTM1.

Conclusions
We present data demonstrating that FIG4 is required for muscle function but is dispensable for muscle development. In addition, most abnormalities associated with Fig4 mutation appears to be secondary to the severe neuropathy documented in plt mice. Our results do not support a role for FIG4 in EC coupling. Future experiments are needed to more firmly establish the relationship between PI(3,5)P 2 and EC coupling.

Additional file
Additional file 1: Figure S1. Reduced muscle mass in plt animals. Figure S2. Reduced body mass in plt animals. Figure S3. Vacuoles are present in plt myocytes. Figure S4. Muscle force is restored to normal in Fig4 -/-; Tg;NSE mice.

Competing interests
The authors declare that they have no competing interests.
Authors' contributions AR performed the majority of experiments, aided with data interpretation and helped generate the manuscript. XL, DW and MB helped perform the experiments. GL helped with mouse husbandry and manuscript generation. SVB performed and interpreted muscle force experimentation. LG and RTD designed, performed, analyzed and interpreted the myotube voltage-clamp experiments. MM aided with experimental design, data interpretation and manuscript generation. JJD conceived the project, helped with all data interpretation and generated the manuscript. All authors read and approved the final manuscript.