Early signs of architectural and biomechanical failure in isolated myofibers and immortalized myoblasts from desmin-mutant knock-in mice

In striated muscle, desmin intermediate filaments interlink the contractile myofibrillar apparatus with mitochondria, nuclei, and the sarcolemma. The desmin network’s pivotal role in myocytes is evident since mutations in the human desmin gene cause severe myopathies and cardiomyopathies. Here, we investigated skeletal muscle pathology in myofibers and myofibrils isolated from young hetero- and homozygous R349P desmin knock-in mice, which carry the orthologue of the most frequent human desmin missense mutation R350P. We demonstrate that mutant desmin alters myofibrillar cytoarchitecture, markedly disrupts the lateral sarcomere lattice and distorts myofibrillar angular axial orientation. Biomechanical assessment revealed a high predisposition to stretch-induced damage in fiber bundles of R349P mice. Notably, Ca2 +-sensitivity and passive myofibrillar tension were decreased in heterozygous fiber bundles, but increased in homozygous fiber bundles compared to wildtype mice. In a parallel approach, we generated and subsequently subjected immortalized heterozygous R349P desmin knock-in myoblasts to magnetic tweezer experiments that revealed a significantly increased sarcolemmal lateral stiffness. Our data suggest that mutated desmin already markedly impedes myocyte structure and function at pre-symptomatic stages of myofibrillar myopathies.


R349P desmin knock-in mice
In this disease model, in which the expression of the mutant desmin is controlled by the endogenous gene regulation sites, the het and hom mice develop phenotypes that correspond to the autosomal-dominant and -recessive human desminopathies, respectively. As controls, wild-type (hereafter termed wt) littermates were used. All animal-related work was performed in accordance with the German Animal Welfare Act reference number TS-14/2015). All applicable international, national, and institutional guidelines for the care and use of animals were followed. Mice were 17 -23 weeks of age.

Small fiber bundles, single muscle fiber, and myofibrillar bundle preparations
After inhalation anesthesia with isoflurane, mice were killed by cervical dislocation. The tail tip was stored at -20 °C for genotyping. The hind limbs were cut off and immersed in Ringer solution (Ri; in mM: NaCl, 145, Hepes, 10, glucose, 10, KCl, 5, CaCl2, 2.5, MgCl2, 1, pH 7.4), and the soleus muscle (SOL), the extensor digitorum longus muscle (EDL)  For single myocyte biomechanics recordings, immortalized (p53-deficient) mouse myoblasts homozygous and heterozygous for the Des R349P mutation and controls carrying the wt desmin were used. Myoblasts were cultured in growth medium. Fluor 594 F(ab`)2 fragment (A11020, Molecular Probes Life Technologies) was added at 1:10,000 in 5 % BSA-TBS for 1 h followed by a three times 10 min washing step.

Second Harmonic Generation (SHG) and multiphoton fluorescence (MPF) imaging
The recorded images had a size of 150 x 150 µm and consisted of 1072 x 1072 pixels.
Two scans of each pixel at 600 Hz were averaged to increase the signal-to-noise ratio.
The average laser power at the sample was around 16 mW, and pulse duration was The intact muscle fibers were then separated from the debris by transferring them in a Pasteur pipette to a dish coated with Matrigel (BD Biosciences; diluted 1:100 in DMEM).
The plated fibers were allowed to settle and attach for 3 min to the Matrigel substrate, and then 1 ml plating medium (DMEM containing 10 % horse serum and 0.5 % chick embryo extract) was slowly added to each dish. Plates were returned to the incubator at 37° C and 5 % CO2 for 24 h. Myoblasts were split, pre-plated on uncoated culture dishes for up to 2 h (to remove contaminating fibroblasts, performed at every splitting procedure), and finally cultivated in Ham's F10 medium supplemented with 20 % FCS, 2.5 ng/ml basic fibroblast growth factor (bFGF, Promega), and 1 % penicillin/streptomycin on collagen-coated (0.01 % collagen in PBS) culture dishes.
Immortalized myoblast cell lines with passage numbers of up to 40 were used for experiments.

Image processing and morphometric analysis of SHG and MPF data
Y-shaped deviations from the sarcomere pattern in a z-stack of SHG images were defined as verniers 23 . The density of verniers (VD) 46 was weighted according to the single fiber area per slice (ai) and was presented as weighted number of verniers within a total area (A) of 100 μm 2 . The normalized VD was given by: The cosine angle sum (CAS) 24 was used as a direct measure for the coherency and structural integrity of single fibers, i.e. reflecting the degree of local angular deviation of myofibrillar bundles from the main trunk axis 45 . The CAS was calculated as weighted mean according to the fiber area of all slices of a z-stack (Ω) by using as number of pixels representing the surface of the fiber, Φ (x,y) as local direction, and the median [Φ (x,y)] as the main direction of the fiber: The two-and three-channel (SHG, Hoechst and Alexa Fluor 594) z-stacks were displayed with Fiji based on ImageJ (National Institutes of Health, Bethesda, MD, USA).
The segmentation operation by thresholding extracted the nucleus volume (NV) and myosin volume (MV) by counting the segmented voxels (nHoechst, nSHG). Using these parameters, the nuclear-myosin-quotient (NMQ) was calculated as: The corresponding biomotoric efficiency (BE) was derived from NMQ as described before as NMV 47 . The nuclear density and the 3D morphology of the nuclei was analyzed using IMARIS software (Bitplane AG, Zurich, Switzerland). The number of nuclei per fiber volume (nuclear density), the volume per nucleus (µm³), and the nucleus sphericity in single fibers were calculated using the Surpass viewcontour surface menu of IMARIS.

SDS PAGE analysis of myosin heavy chain distributions in soleus muscle
homogenates.

Assessment of active and passive biomechanics in small fiber bundles
Soleus fiber bundles were subjected to a series of active and passive force recordings.
After attaching a bundle to the force transducer and voice coil actuator pin, the preparation was chemically permeabilized in high relaxing ( were immediately stretched in 10 % bins at very fast velocity and kept for 5 s at the new length before proceeding to the next extension. The force response consisted of an instantaneous restoration force followed by an exponential relaxation to a steady-state force level Fss at the given stretch bin. For analysis, the peak restoration force FR, the difference between FR and Fss, F, and the time constant of exponential relaxation relax were determined in each bundle.

Myofibrillar bundle biomechanics
Force measurements in isolated myofibrillar bundles were performed in relaxing solution (pCa 8) containing 3 mM K4Cl2EGTA, 10 mM imidazole, 1 mM K2Cl2Na2MgATP, 3 mM MgCl2, 47.7 mM Na2CrP, 2 mM DTT, pH 7.0 at 10 °C using the experimental setup described in 45,46 . Thin subcellular myofibrillar bundles (diameters of 3 -4 µm) were mounted between the tip of an atomic force cantilever and the tip of a length-driving stiff tungsten needle. After mounting, the slack sarcomere length, the overall slack length L0, and the diameter of the bundles were determined. To determine the relation of the passive steady-state tension versus the sarcomere length, the bundles were rapidly stretched by moving the tungsten needle with a piezo motor using variable length steps with sizes of 8 %-multiples of L0. After stretch, an image of the bundle was taken using an ORCA-ER camera (Hamamatsu, Hamamatsu Photonics, Germany) and 60x objective (60x/0.70 Ph2 LCPlanFl, Olympus) for evaluating the actual sarcomere length using the software Aquacosmos (Hamamatsu Photonics). After holding the bundle for 12 s at stretched length, the bundle was rapidly slackened to determine its passive force from the drop of force to zero. Passive tension was calculated by normalizing passive force to the cross-sectional area and then plotted against the actual sarcomere length.

Magnetic tweezer compliance recordings in single myoblasts
For each experiment, 8 x 10 4 myoblasts were seeded overnight in a culture dish. 30 min before experiments, cells were incubated with fibronectin-coated paramagnetic beads of 4.5 µm diameter (Invitrogen). A magnetic field was generated using a solenoid with a needle-shaped core (HyMu80 alloy, Carpenter, Reading, PA). The needle tip was placed at a distance of 20 -30 µm from a bead bound to the cell using a motorized micromanipulator (Injectman NI-2, Eppendorf). During measurements, bright-field images were taken by a CCD-camera (ORCAER, Hamamatsu) at a rate of 40 fps. The bead position was tracked using an intensity-weighted center-of-mass algorithm.
Measurements on multiple beads per well were performed at 37 o C for 1 h, using a heated microscope stage on an inverted microscope at 40x magnification (NA 0.6) under bright-field illumination. The bead displacement (d) after a step increase of the force (F; 10 nN for 3 s to fibronectin-coated superparamagnetic beads attached to integrin receptors on the surface of myoblasts) followed a power law with time (t). The cell's lateral compliance , which is inversely proportional to its stiffness, was determined from the creep response of the cells by fitting the displacement with the typical power law response , where .