In vivo imaging of skeletal muscle in mice highlights muscle defects in a model of myotubular myopathy

ABSTRACT

Skeletal muscle structure and function are altered in different myopathies. However, the understanding of the molecular and cellular mechanisms mainly rely on in vitro and ex vivo investigations in mammalian models. In order to monitor in vivo the intracellular structure of the neuromuscular system in its environment under normal and pathological conditions, we set-up and validated non-invasive imaging of ear and leg muscles in mice. This original approach allows simultaneous imaging of different cellular and intracellular structures such as neuromuscular junctions and sarcomeres, reconstruction of the 3D architecture of the neuromuscular system, and video recording of dynamic events such as spontaneous muscle fiber contraction. Second harmonic generation was combined with vital dyes and fluorescent-coupled molecules. Skin pigmentation, although limiting, did not prevent intravital imaging. Using this versatile toolbox on the Mtm1 knockout mouse, a model for myotubular myopathy which is a severe congenital myopathy in human, we identified several hallmarks of the disease such as defects in fiber size and neuromuscular junction shape. Intravital imaging of the neuromuscular system paves the way for the follow-up of disease progression or/and disease amelioration upon therapeutic tests. It has also the potential to reduce the number of animals needed to reach scientific conclusions.

Introduction

Skeletal muscle represents nearly half of the dried body weight and is essential for movement, energy homeostasis and thermogenesis. Skeletal muscle fibers are syncytia containing hundreds of nuclei, span up to 20cm in human and are innervated through the neuromuscular junction. The sarcomere is the main unit of the contractile apparatus and is mainly composed of actin and myosin filaments.¹ At the neuromuscular junction (NMJ), the nerve action potential triggers the release of acetylcholine that binds to the acetylcholine receptors to create a depolarization of the muscle sarcolemma. Depolarization is sensed by voltage sensor channels located at the triad, the membrane structure sustaining excitation-contraction coupling, and induces the release of calcium from the sarcoplasmic reticulum store to the cytoplasm. Calcium binds and activates troponin followed by sliding of the actin and myosin filaments on each other.

Skeletal muscle structure and function are altered in a plethora of genetic diseases, as myopathies,¹ while progressive muscle loss as cachexia and sarcopenia are hallmarks of AIDS and cancer and a main sign of aging. Mutations in about 150 genes were implicated in myopathies and muscular dystrophies to date.² Among the most severe myopathies, myotubular myopathy (also called X-linked centronuclear myopathy, XLMTM, OMIM #310400) is associated with neonatal hypotonia, muscle weakness and breathing difficulties in male patients and is due to loss-of-function mutations in the MTM1 gene.^3-7 Histopathological hallmarks comprise fiber hypotrophy and abnormal organelle positioning.⁸ Different potential causes of the muscle weakness have been proposed for XLMTM, as triad structural defects,^9-11 unbalanced autophagy and protein homeostasis,^12-14 satellite cells alterations¹⁵ or anomalies of the neuromuscular junction.^16,17 However, monitoring cellular pathological hallmarks in living mammalian models of XLMTM was not achieved to date, and the knowledge of the cellular pathology of XLMTM mainly relies on in vitro or ex vivo observations in mammalian models,^9,13,18-21 or on data from more evolutionary-distant models as zebrafish,^10,17 drosophila,²² C. elegans²³ or yeast.^18,24,25

Monitoring the structure of muscles in place is thus of importance to follow disease progression and potential amelioration upon treatments. A few studies previously reported in vivo microscopy of skeletal muscle.^26,27 While some disease hallmarks can be recapitulated in vitro, intravital imaging (IVM) offers a window to the progression of the disease, and its associated cellular events such as subcellular membrane trafficking,²⁸ in a living animal.^29-31 Owing to its superior imaging property and its compatibility with non-linear optical processes, Two-Photon-Excitation-Microscopy (2PEM) allows simultaneous tracking of fluorescent events in vivo as well as detection of non-centrosymetric molecules by second harmonic generation (SHG)such as myosin in muscle,³² and collagen in the extracellular matrix.^33,34 2PEM and SHG were previously used to assess the structure of the contractile apparatus and tubular system in isolated mouse muscle fibers,³⁵ in preserved muscles from myopathic mice,^36,37 or in muscle in place in the animal after removal of the skin.³⁸ Minimally invasive imaging of muscle sarcomeres was achieved in mouse and human using microendoscopes.^39,40 However, non-invasive imaging of skeletal muscle under normal and pathological conditions in living mice was not extensively reported.

In order to monitor the intracellular structure of muscle fibers and the surrounding environment under normal and pathological conditions, we set-up and validated non-invasive imaging of ear and leg muscles, allowing simultaneous imaging of different cellular structures, reconstruction and visualization of the 3D architecture of the neuromuscular system in its environment, and video recording of dynamic events as muscle fiber contraction. Intravital analysis of a XLMTM murine model highlighted several pathological hallmarks of the disease.

Materials and methods

Animals

BALB/c and 129/SvPas mouse strains were used. 129/SvPas Mtm1-/y mice modeling XLMTM were previously described.⁹ Animal experimentations were approved by IGBMC/ICS institutional review board (N° 2012-133).

Confocal macroscope imaging

Prior to imaging, the mice were either anesthetized via intra-peritoneal injection of a ketamine (100 mg/kg) and xylazine (10 mg/kg) mix, or sacrificed. When the skin was removed, the following dyes were applied in a 50% glycerol-PBS solution on the bare muscle: Hoechst 33342 (100 ug/ml), DiOC6 (100 uM), Cell Mask deep red (500 ug/ml), cresyl violet (0.3%), and α-Bungarotoxin (1 ug/ml). Anesthetized mice were kept at a constant temperature of 28°C for preserving the mouse body temperature around 37°C and positioned under the Leica Macrofluo confocal macroscope A rectal probe may also be used to monitor body temperature of the experimental mouse throughout the anesthesia and imaging session. To image muscle contraction, the macroscope was equipped with a Leica PLAN APO 5x lwd objective (NA: 0.5), a spinning disk head (Yokogawa CSU22) and an EMCCD camera (Andor, iXon+). Time-lapse acquisitions at 64–73 frames/s displaying significant muscle contraction were analyzed by tracking single sarcomeres over time (Fig. 4, movies S2-3). Relative muscle contraction was assessed by measuring the distance spanning 50 sarcomeres over time (Fig. 4).

Figure 4.

Dynamic monitoring of single muscle fiber contraction. (A) Schematic drawing of the muscle fiber contraction. (B) Image at the start of the video recording; myofiber sarcomeric organization was labeled with the vital dye DIOC6. (C-D) Total movement of regions of interest during the whole spontaneous contraction. (E) Positions of regions of interest through time. (F) Repetitive organization and contraction of sarcomeres assessed through time based on the total length of 50 contiguous sarcomeres. (G) Longitudinal scan line depicting the distance between several contiguous sarcomeres and highlighting their regular spacing. (H) Variation of sarcomere length through the contraction period. The full movie is shown in Video S2.

2-Photon-excitation microscopy (2PEM) imaging

Before 2PEM imaging, the mice were first anesthetized through intra-peritoneal injection of a mixture of ketamine (100 mg/kg) and xylazine (10 mg/kg). The NMJ were stained by subcutaneous injection of  α-bungarotoxin conjugate to green fluorochrome CF488A (Biotium, 00005) at 2,5 µg/ml. Evans Blue (100 µl , 10 mg/ml in PBS, E-2129 Sigma-Aldrich) was administrated via retro-orbital injection in order to allow imaging of the blood vasculature. Alternatively, fluorescent dextran can also be administrated as described previously.⁴¹ The mouse was maintained anesthetised during the whole imaging process. The mouse ear was imaged as previously described.^41,42 The anesthetized mouse was mounted on a home-made holder.^41,43 The mice were then disposed on the stage of an upright Leica SP5X MP (Leica microsystems) and the temperature was set at 28°C for preserving the mouse body temperature around 37°C. A single excitation wavelength at 940 nm (Chameleon Ultra II, Coherent) was used for 2PEM and recorded using a 25x 0.95 N.A. or 20x 1 N.A. water immersion objectives. Non-descanned detectors (NDD) were used to collect the emission signal of the CF488A at 510 nm and the Evans blue at 680 nm. Second harmonic generation was detected with a trans-detector through a single-band bandpass filter 470/22 nm. As described previously, second harmonic generation occurs mostly from 2PEM of non-centrosymetric structures, and is restricted to collagen fibers and striated muscle myosin rod domains⁴⁴ in mammalian soft tissues. Consecutive images of 512 × 512 pixels (pixel size: 0.3 µm in x/y, 1 µm in z) were acquired at 0.4 frames/s.

Image processing, single particle tracking and analysis

The 2PEM datasets were analyzed using Fiji.⁴⁵ The module Analyze Particles was used to assess the morphology of the NMJ in 3D. Confocal acquisitions restricted to planes containing the NMJ were stacked onto a maximum z projection and circularity was analyzed on n = 11 NMJ for the WT and n = 12 for the KO mice. The measurements of the sarcomere length were determined by plotting the intensity profiles along a longitudinal line and measuring the distance between 2 consecutive SHG intensity peaks corresponding to myosin.

Results

Set-up for imaging different skeletal muscles

Two different muscles were used for imaging. The tibialis anterior (TA) in the lower hind limb is a mixed muscle containing oxidative and glycolytic fibers, is easily injectable and dissectable, and has been extensively studied in various murine myopathy models (Fig. 1B). As movement artifacts due to breathing significantly challenge IVM of the TA, we analyzed in parallel muscles of the outer ear. These three muscles stripes, which can be seen by backward lighting, are easily accessible and almost insensitive to breathing movements (Fig. 1A). They are thus suitable for non-invasive IVM of subcutaneous tissues and can be used for tracking tumor cell behavior in vivo,^41,43 where non-linear 2PEM can be used to track collagen-rich ECM fibers using SHG. We designed a specific holder to constrain the ear for imaging without interfering with SHG signal detection (Fig. S1). Muscle imaging was performed with several microscopy approaches such as 2PEM for in-depth imaging, spinning disc microscopy for high speed acquisition, or macroconfocal microscopy allowing a good intracellular resolution with wide field of view and large distance between objective and sample (Fig. 1C and 1D). Hematoxilin-eosin staining revealed the general organization of TA and ear muscles (Fig. 1E and 1F), and electron microscopy highlighted the characteristic ultrastructure of muscle fibers with the contractile unit (sarcomere) composed of bands of different contrasts (Fig. 1G and 1H). The ear muscles, poorly described to date, are located on the outside surface of the ear close to the cartilage and underneath the skin and connective tissue. Overall, both ear and TA muscles display a classical skeletal muscle intracellular organization and are easily accessible for imaging. They represent an excellent model for tracking muscle by intravital imaging.

Figure 1.

Assessing muscles with 2-photon excitation microscopy and confocal macroscopy. Imaging skeletal muscle in the ear (A, C, E, G), and in the anterior leg (B, D, F, H) of anesthetized mice. (A, B) Position of the ear and TA muscles. (C, D) Set-up for imaging ear muscles with a microscope (C) and TA muscles with a confocal macroscope (D). (E, F) Haematoxilin-eosin staining of transversal sections showing muscle fibers in pink and nuclei in blue. (G, H) Ultrastructure of myofibrils in longitudinal EM sections. The major sarcomeric components have been identified (M and Z lines, I and A bands). The A line is responsible for the SHG pattern.

3D multi-color imaging of muscle fibers in their environment

In order to simultaneously label and record different tissues and intracellular structures, we used either non-invasive SHG muscle imaging (2PEM), or imaging upon injection of vital dyes and fluorescently labeled small molecules, or a combination of both approaches.

For ear imaging, blood vessels were visualized by an intravenous injection of Evans blue, neuromuscular junctions were stained with CF488-coupled bungarotoxin, and sarcomeres and collagen fibers were monitored with 2PEM-SHG (Fig. 2A and B). As expected, the NMJ are located at similar depth as the muscle fibers (Fig. 2C). 3D volume reconstruction offers access to the inner architecture of the muscle fibers and the relative position of the NMJ and the vasculature (Fig. 2D and Movie S1). Deep intracellular resolution can be achieved up to hundred micrometers.

Figure 2.

Muscle fibers and pre-labeled neuromuscular junctions can be imaged simultaneously. (A) Light microscopy picture of a section (toluidine blue staining) through the mouse ear. Muscle fibers are colored in pink. (B-D) Imaging of mouse ear structures across the skin. Evans Blue labels the blood flow (red), non-linear 2PEM (second harmonic generation) shows mainly collagen fibers and muscle sarcomeres (blue), and CF488-bungarotoxin labels the neuromuscular junctions (green). (B) Optical section at different Z-depths. The SHG signal (left panel) is color coded according to the z depth. The collagen rich layer of the skin appears in blue and the muscle fibers in yellow. The cartilage is the deepest structure that can be imaged. (C) Transversal imaging with 2PEM-SHG displaying the position of ear muscles relative to neuromuscular junctions and blood vessels. (D) 3D reconstruction of the architecture of the neuromuscular system in its environment. The neuromuscular junctions are in close proximity with the layer of muscle fibers. As the SHG signal drops at significant depth the muscle fiber connected to the upper right NMJ is barely visible.

For TA muscle, the cytosol was labeled with the cresyl violet, the plasma membrane with CellMask Deep Red, the nuclei with Hoechst, and mitochondria with DIOC6 (Fig. S2). Note that mitochondria labeling highlights the sarcomeric organization as they are enriched in the I-band between the Z-disc and the triad. All these structures can be imaged upon minimal surgery to expose the muscle and make it amenable for microscopy. Co-labeling was used to assess the position of several structures in the same fiber.

Taken together, we show that different tissues and intracellular structures can be labeled simultaneously, imaged within a tissue-depth of up to 150 µm (with the Leica TCS LSI macroconfocal or with the upright Leica SP5X MP microscope), and visualized in 3D, thus offering a wide range of structural information.

Impact of skin pigmentation on imaging

We next assessed the impact of the skin and its pigmentation on the resolution and depth of intravital imaging. Intravital imaging of the ear muscles was performed as above using Evans blue for blood vessels (in case the skin was not peeled), 2PEM-SHG for sarcomeres and collagen fibers, and CF488-coupled bungarotoxin for the neuromuscular junctions (Fig. 3). The skin was either removed (Fig. 3A) or left intact (Fig. 3B and 3C). In addition, we performed imaging through the skin of albino BALB/c (white fur and light skin) or 129/SvPas (agouti fur and skin) mice. As expected, the overall image quality was better when performed after skin removal, which can be related to light absorption by pigmented layers of the skin. Of note, skin removal induces bleeding of small capillaries and is therefore incompatible with Evans blue injection. In contrast, image quality and resolution with intact skin was adequate to assess the position and shape of blood vessels and muscle fibers, the sarcomere organization, and the position and complexity of neuromuscular junctions. For example, 2PEM imaging of intact-skin ears allows monitoring of the mentioned structures to a depth of 150 µm (Fig. 2). The higher pigmentation of the agouti skin had a slight impact on image quality, albeit the same structures could be visualized. We noticed that strong skin pigmentation in C57BL/6 mice significantly impaired depth and quality of imaging across the skin, making non-invasive transcutaneous intravital imaging difficult in this specific genetic background (not shown).

Figure 3.

Skin pigmentation does not prevent imaging of muscle fibers and neuromuscular junctions. Intravital imaging of muscle sarcomeres (2PEM-SHG, white), neuromuscular junctions (CF488-bungarotoxin, green) and blood vessels (Evans Blue, red). Optical sections and merge photos on the last right column. (A) Imaging after skin removal of the dorsal surface of the mouse ear. (B) Imaging through the skin of a BALB/c mouse. The skin absorbed and scattered the excitation laser without preventing the imaging of the muscle fibers and labeled structures. (C) Imaging through the skin of an Agouti 129PAS mouse. The pigmentation of the skin has a main impact on the SHG signal noise.

The impact of the presence of the skin was also assessed on the TA leg muscle. To label the muscle fibers, we injected and electroporated a plasmid expressing a cytoplasmic GFP protein (Fig. S3). Individual muscle fibers can be clearly identified and imaged over the whole muscle length with the confocal macroscope both with and without the skin.

In conclusion, the murine neuromuscular system can be imaged and studied with a non-invasive protocol, and different labeling/detection methods as fluorescent dyes, SHG and recombinant fluorescent protein could be recorded through the skin, offering a wide palette for studying muscle function in vivo.

Recording and analysis of muscle fiber contraction

In vivo imaging can in principle allow recording of dynamic events as muscle contraction encompassing movements in space and through time, and representing a primary role of skeletal muscle. Quite serendipitously, we found single fibers contracting spontaneously and independently of neighboring fibers within the same muscle. The regular positioning of mitochondria within sarcomeres, and thus sarcomeric organization, was labeled with DIOC6 in the TA leg muscle and time recording performed with a spinning disc macroscope (Fig. 4 and Movie S2-3). A whole cycle of a spontaneous contraction lasted for a few seconds, based on the observation of several contractions of the same fiber. The displacement of specific intracellular tractable structures was followed through time for different fibers (Fig. 4C-E, Movie S2-3) and quantification of the total displacement distance confirmed that a single fiber significantly contracts compared to adjacent fibers that display no or weaker contraction. Pixel intensity along a longitudinal lane within a muscle fiber highlights the repetitive organization of sarcomeres (Fig. 4F and 4G). We measured the length of 50 successive sarcomeres through time and calculated the maximal contraction to be 5% (Fig. 4H). The substantial contraction of the whole fiber is thus due to the fact that it contains a large number of sarcomeres.

Intravital analysis of neuromuscular defects in a murine model of myotubular myopathy

Using the different imaging modes, we assessed whether a neuromuscular phenotype can be detected in a mouse model for myotubular myopathy. The Mtm1-/y mouse displays a progressive muscle atrophy and weakness starting at 3 w, and leading to death by 7-12 w.^9,20 Fiber hypotrophy and organelle mis-positioning were previously found in the TA leg muscle, and here we investigated the ear to assess whether this more accessible muscle can be used for disease phenotyping. Upon inspection with back lighting, the 3 muscle stripes of the back ear appeared thinner in the Mtm1-/y mouse compared to wild type mice.

Histological analysis with hematoxylin-eosin staining of transversal sections highlighted a typical CNM phenotype comprising thinner ears mainly due to a decrease in total muscle bulk, while other tissues like cartilage were not significantly affected. We also observed fiber size heterogeneity and an increased number of atrophic fibers, as well as abnormal nuclear centralization (Fig. 5A). Measuring the muscle fiber area confirmed that Mtm1-/y fibers are overall significantly thinner in the ear compared to WT (Fig. 5B).

Figure 5.

Intravital imaging highlights neuromuscular defects in a murine model of myotubular myopathy. Haematoxilin-eosin staining (A-B) or intravital imaging (C-E) to assess fiber size and neuromuscular junction shape in 5 week old 129/SvPas WT and Mtm1-/y mice. (A) The Mtm1-/y genotype displays decreased muscle mass, fiber size heterogeneity, a bias toward smaller fibers, and abnormal centralization of nuclei that are normally at the periphery of fibers. On the right, muscle fibers are color coded. (B) Measurement of the fiber area (WT n = 14; mean = 470,2. KO n = 14; mean = 167,7). (C) Intravital imaging of muscle sarcomeres (2PEM-SHG, white), neuromuscular junctions (CF488-bungarotoxin, green) and blood vessels (Evans Blue, red). Optical sections and merge photos in the right column. (D) To assess the shape of the neuromuscular junctions, a circularity coefficient was calculated with 0 representing a lane and 1 a circle (WT n = 10; mean = 0,20, KO n = 11; mean = 0,32). (E) Measurement of the fiber width (WT n = 16; mean = 16,54µm, KO n = 13; mean = 14,77µm). Mtm1-/y mice display smaller average fiber width both by histology and intravital imaging, and have rounder neuromuscular junctions. Distribution of the data is presented as whisker plots displaying both minimal and maximal values obtained over a single animal for each genotype.

Co-imaging of the sarcomeres with 2PEM-SHG, the neuromuscular junction with CF488-coupled bungarotoxin and blood vessels with Evans blue showed that the general organization of the tissues and the position of the muscles were conserved (Fig. 5C). Using SHG, we measured an average fiber width of 16.5 µm for wildtype and 14.7 µm for Mtm1-/y mice (Fig. 5E). Notably, there was a large distribution of fiber size in wildtype in the ear muscle, while the Mtm1-/y mice lacked the larger fibers. Taken together, both histological and IVM approaches revealed similar structural defects, suggesting that IVM is suitable for disease phenotyping. The neuromuscular junctions were further scrutinized (Fig. 5C-D). Their position compared to the neighboring tissue did not appear different in the Mtm1-/y or wildtype genotypes. In addition, we could not observe any obvious change in the density and distribution of neuromuscular junctions in the ears from Mtm1-/y mice. To get insight into their complexity, the shape was assessed and was found different between the 2 genotypes; a significant number of neuromuscular junctions was rounder in the Mtm1-/y genotype.

Altogether, these data show that we can detect and quantify significant alterations in the myotubular myopathy Mtm1-/y mouse model, especially concerning fiber size and neuromuscular junction shape, supporting that monitoring the ear muscle is a good marker to potentially follow the disease progression or amelioration.

Discussion

Here we describe a non-invasive analysis of the intracellular organization of skeletal muscle and the neuromuscular system in mice by multimodal microscopy combining SHG and fluorescence detection. We demonstrate that this approach is suitable to reveal disease hallmarks in a murine model of myotubular myopathy.

A versatile toolbox for intravital muscle imaging.

In order to analyze the architecture of the murine neuromuscular system in vivo, we focused on 2 main structures, the sarcomere and the neuromuscular junction. We monitored the tibialis anterior (TA) in the lower hind limb, as well as muscles in the outer ear. The TA muscle is well described and often used for ex vivo analysis, but we noted significant motion artifacts resulting from breathing and/or blood pressure, challenging the construction of image stacks and video recording. The ear muscle was seldomly investigated, but has the advantage to be easily accessible and insensitive to motion artifacts when the ear is constrained. As TA and ear muscles displayed similar structural neuromuscular defects in the myotubular myopathy mouse, we propose the ear muscle as new paradigm for intravital imaging of skeletal muscle.

Multimodal imaging, especially if achievable at the same excitation wavelength, is a powerful approach to combine the analysis of diverse tissues and even subcellular structures at different depths. Here we imaged blood circulation, neuromuscular junctions, collagen, cartilage, and muscle fibers at the same time. This versatility enables 3D reconstruction of specific cells and structures in situ, and paves the way to the study of the surface between 2 cell types like the neuromuscular junction in contact or close proximity with a myofiber. The intracellular organization of myofibers can be followed through detection of myosin molecules with SHG, and mitochondria, membranes and nuclei or other organelles that can be labeled fluorescently. It is possible to measure the distance between each sarcomere and thus to deduce the sarcomere length and organization, or the surface occupied by the neuromuscular junction.

In addition, we validated the use of a versatile imaging toolbox by combining SHG with chemical dyes, fluorescence-coupled molecules as bungarotoxin, and recombinant DNA constructs expressing fluorescent proteins. SHG does not require exogenous labeling as it is based on the fluorescence of non-centrosymmetric structures as myosin rod domains that are parallelly aligned in the sarcomere.^32,44,46 Chemical dyes have the advantage to cover a very large spectrum of structures and organelles, are easy to use, and are partially applied in clinical practice. The fluorescence-coupled molecules offer the possibility to tag and follow nearly any desirable molecule, and recombinant DNA constructs open the perspective to utilize advanced imaging modes as photoactivation and optogenetics.⁴⁷

The depth of analysis is an important parameter for intravital imaging. The infrared light used in 2PEM deeply penetrates into the sample, allowing the visualization of muscle fibers beneath other tissues as the skin, collagen and blood vessels. Infrared wavelengths also have the advantage of reducing phototoxicity and photobleaching. Imaging the neuromuscular system is still feasible through an intact skin, even when using Agouti mice with colored skin, allowing fully non-invasive imaging. However, the mice need to be anesthetized before imaging and some labelings require the injection of dyes.

Monitoring dynamic events provides functional information in addition to intravital structure analysis. As a proof-of-concept, we followed spontaneous myofiber contraction. Sarcomere length decreased by about 5%. Given the large number of sarcomeres per fiber, this decrease represents a substantial contraction of the whole myofiber. Our approach provides the possibility to study various parameters and kinetics of muscle contraction as time to maximal contraction or relaxation time under normal and pathological conditions.

Intravital monitoring of a muscle disease

To validate the intravital muscle imaging toolbox for the characterization of a disease state, we focused on a mouse model for myotubular myopathy. Previous ex vivo investigations of this model revealed different histopathological features including myofiber hypotrophy and structural abnormalities of the neuromuscular junction.^17,20,21 Using intravital imaging, we were indeed able to detect an increased proportion of atrophic fibers, and an abnormal shape of the neuromuscular junctions in the Mtm1 knockout mice. Our in vivo observations support the notion that the muscle weakness correlates with hypotrophy, and that functional defects of the neuromuscular transmission represent one of the disease causes. Previous histopathological examinations mainly focused on the leg muscles, and here we describe similar defects in the ear muscle. This suggests that most skeletal muscles are affected in myotubular myopathy, whether or not they drive locomotor activities. To our knowledge, this is a first evidence that histological hallmarks of myotubular myopathy can be illustrated without muscle sectioning in a mammalian model.

Further in-depth analyses of this mouse model during pre-symptomatic age (before 3 w) might potentially reveal primary structural or/and functional defects preceding clinical manifestations and thus allows a better understanding of the pathological mechanisms underlying myotubular/centronuclear myopathies. Moreover, the possibility to monitor non-muscle tissues as blood vessels or connective tissues may reveal additional defects not considered before. For instance, it is known that patients with myotubular myopathy display non-muscle features as fatal liver hemorrhages that might relate to smooth muscle defect.³

In conclusion, multimodal intravital imaging of the neuromuscular system in time and space in mice, under healthy and diseased conditions, paves the way for the follow-up of disease progression or/and disease amelioration upon therapeutic tests, and has the potential to significantly reduce the number of animal needed to reach scientific conclusions.

Abbreviations

2PEM

2-photon microscopy

IVM

intravital imaging

NMJ

neuromuscular junction

SHG

second harmonic generation

XLMTM

X-linked myotubular myopathy

Acknowledgments

We thank Stéphane Vassilopoulos for suggestions, Christine Kretz for mouse handling and genotyping, Hichem Tasfaout and Ivana Prokic for technical assistance, and Serge Taubert for the mouse ear holder.

Funding

This study was supported by INSERM, University of Strasbourg, CNRS, the Agence Nationale de la Recherche (ANR-08-GENOPAT-005, ANR-10-LABX-0030-INRT), the program Investissements d’Avenir (ANR-10-IDEX-0002-02, Idex starting grant), the Association Française contre les Myopathies (AFM 14204) and the Fondation pour la Recherche Médicale (FRM DEQ2007). L.M. is supported by an INSERM/Région Alsace Ph.D fellowship.