Loss of myotubularin function results in T-tubule disorganization in zebrafish and human myotubular myopathy

Abstract

Myotubularin is a lipid phosphatase implicated in endosomal trafficking in vitro, but with an unknown function in vivo. Mutations in myotubularin cause myotubular myopathy, a devastating congenital myopathy with unclear pathogenesis and no current therapies. Myotubular myopathy was the first described of a growing list of conditions caused by mutations in proteins implicated in membrane trafficking. To advance the understanding of myotubularin function and disease pathogenesis, we have created a zebrafish model of myotubular myopathy using morpholino antisense technology. Zebrafish with reduced levels of myotubularin have significantly impaired motor function and obvious histopathologic changes in their muscle. These changes include abnormally shaped and positioned nuclei and myofiber hypotrophy. These findings are consistent with those observed in the human disease. We demonstrate for the first time that myotubularin functions to regulate PI3P levels in a vertebrate in vivo, and that homologous myotubularin-related proteins can functionally compensate for the loss of myotubularin. Finally, we identify abnormalities in the tubulo-reticular network in muscle from myotubularin zebrafish morphants and correlate these changes with abnormalities in T-tubule organization in biopsies from patients with myotubular myopathy. In all, we have generated a new model of myotubular myopathy and employed this model to uncover a novel function for myotubularin and a new pathomechanism for the human disease that may explain the weakness associated with the condition (defective excitation-contraction coupling). In addition, our findings of tubuloreticular abnormalities and defective excitation-contraction coupling mechanistically link myotubular myopathy with several other inherited muscle diseases, most notably those due to ryanodine receptor mutations. Based on our findings, we speculate that congenital myopathies, usually considered entities with similar clinical features but very disparate pathomechanisms, may at their root be disorders of calcium homeostasis.

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Figure 1. Abnormal morphology in myotubularin morphant embryos.

(A) Live embryos at 24 hpf injected with either control (CTL) or myotubularin (MTM) morpholinos. MTM morphants are of equivalent size, but are bent or U-shaped in appearance. (B) Live embryos at 72 hpf injected with control (CTL MO) or myotubularin (MTM MO) morpholinos. MTM morphants are mildly dysmorphic in appearance, and display selective thinning of the muscle compartment (brackets) as well as foreshortening of their tails.

Figure 2. Abnormal motor function in myotubularin morphants.

(A) Quantitation of spontaneous embryo coiling at 24 hpf (see also Supplemental Videos 1 and 2). On average, CTL morphants coiled 10.2 times in 15 seconds, while MTM morphants coiled only 5.2 times. (B) Quantitation of chorion hatching in 60 hpf morphants. 87.2% of CTL morphants are hatched from their protective chorions by 60 hpf, as opposed to only 35.3% of MTM morphants. (C) Touch-evoked swimming was video captured in 72 hpf morphants. As expected, CTL morphants responded to tactile stimuli with a rapid escape response contraction followed by swimming. Conversely, MTM morphants displayed a weak escape contraction followed by “twittering” movements (example at 33 ms) but never normal swimming. Scale bar = 1 mm.

Figure 3. Abnormal histopathology in 72 hpf myotubularin morphants.

(A) H/E stained longitudinal myofibers from myotubular myopathy (MTM) and age matched control (CTL) human muscle biopsies. Arrows point to abnormal nuclei. (B) H/E stained longitudinal myofibers from control (CTL MO) and myotubularin (MTM MO) morphant 72 hpf embryos. Myonuclei are abnormally rounded (arrows), and there is increased space between fibers (*). (C) Toluidine blue stained semi-thin sections from 72 hpf morphants. Myonuclei from myotubularin morphants are large, abnormally rounded, and contain discrete nucleoli (arrows). Sarcomeric units, however, are normal in appearance. Scale bar = 20 mm.

Figure 4. Abnormal perinuclear ultrastructure in 72 hpf myotubularin morphants.

Comparison of the perinuclear area from control (CTL MO) and myotubularin (MTM MO) morphants. (A, B) Control injected embryos had thin myonuclei (N) with well organized perinuclear organelles (M = mitochondria). Myotubularin injected embryos had large, rounded nuclei (N) and disorganized perinuclear compartments. Three embryos from three independent injections were examined. Higher magnification (C, D) of the perinuclear compartment revealed abnormal mitochondria (M), areas nearly devoid of organelles (**), and several tubule-like structures (arrows). (E) Example of an unusual membranous perinuclear structure. Such structures were observed in multiple myofibers. Scale bars: A, B (2 mm), C–E (500 nm).

Figure 5. Myofiber hypotrophy in myotubularin morphants.

(A) Representative myofibers from control (CTL) and myotubularin (MTM) morphant embryos at 72 hpf. Fibers were immunostained with an antibody to myosin heavy chain (a-MHC). MTM fibers have normal MHC staining, but appear thinner. Scale bar = 20 mm. (B) Quantitation of myofiber size. Control myofibers averaged 7000 pixels, while myotubularin morphant fibers were only 4000 pixels.

Figure 6. Increased PI3P levels in myotubularin morphant myofibers.

(A) Representative myofibers immunostained with anti-PI3P. Perinuclear staining of PI3P in myotubularin morphant myofibers is much more abundant than in control myofibers (arrows). There is also a modest increase in membrane localized PI3P (*). Scale bar = 10 mm. (B) Quantitation of PI3P immunofluorescence. PI3P intensity measured over a uniform perinuclear area (see methods for details) and was 83.7+/−7.8 pixels for control morphants and 135.5+/−3.3 for myotubularin morphants (3 trials; p = 0.0027). This represented a 1.6× increase in PI3P staining intensity.

Figure 7. MTMR rescue of the myotubularin morphant phenotype.

(A) Whole mount in situ hybridization of 24 hpf embryos reveals muscle staining for myotubularin (MTM1 AS) and not for MTMR1 or MTMR2 (data not shown). A sense probe to MTM1 (MTM1 S) was used as a background control. (B) RNA rescue experiment. RNA to MTM1, MTMR1 or MTMR2 was co-injected with myotubularin morpholino. Rescue was determined by the % of hatched embryos at 60 hpf. Values were: No RNA (35.5%+/−3.3%, N = 201), MTM1 RNA (71%+/−4.5, N = 100, p (when compared to No RNA)

Figure 8. Myotubularin localizes to T-tubules.

Myotubularin and DHPRa1 co-localize. Double label immunofluorescence was performed on isolated myofibers. As demonstrated using confocal microscopy, myotubularin (red) and DHPRa1 (green) signal significantly overlap (orange, panel 3). Scale bar = 10 mm.

Figure 9. T-tubule structural abnormalities in myotubularin morphant muscle.

T-tubule (vertical arrows) and sarcoplasmic reticulum (angled arrows) abnormalities as demonstrated by electron microscopy. Control morphant (panel 1; CTL): normal T-tubule triad with accompanying thin, well-organized SR network (arrow). Myotubularin morphants (panels 2–4): Panel 2 shows mildly dilated triads and SR networks. Panel 3 shows severely dilated and dysmorphic triads and widely looped SR. Panel 4 illustrates severe disorganization, with unrecognizable T-tubule triads and aberrant adjacent SR networks. Scale bar = 500 nm.

Figure 10. Excitation-Contraction Coupling abnormalities in myotubularin morphants.

Excitation-contraction (E–C) coupling. Control morphant myofibers could respond to 15 ms depolarizing current injections to 0 mV from 0 to 30 Hz (average max frequency = 27.0+/−0.9 Hz, N = 5). In contrast, myotubularin morphant muscle progressively failed to contract to stimuli above 10 Hz (average maximum frequency = 11.5+/−1.5 Hz, N = 5, p

Figure 11. Alteration of T-tubule/SR component localization in myotubular myopathy.

Muscle from 3 myotubular myopathy patients and an age matched control were immunostained with DHPRa1 to mark T-tubules (A) and RYR1 to mark the adjacent sarcoplasmic reticulum (B). Abnormal distribution of DHPRa1 and RYR1 was observed in numerous fibers in all 3 myotubularin patients but in none of the control fibers (arrows). Scale bar = 20 mm.

Figure 12. Ultrastructural changes in T-tubules in myotubular myopathy.

Electron microscopic analysis of muscle from 3 myotubular myopathy patients (MTM) and one age-matched control (CTL). Control T-tubule triads are discretely formed (arrow), and the adjacent SR network is thin and well organized. Triads and adjacent SR from patient biopsies are dilated and disorganized. Scale bar = 500 nm.