Cholesterol depletion impairs contractile machinery in neonatal rat cardiomyocytes

Cholesterol regulates numerous cellular processes. Depleting its synthesis in skeletal myofibers induces vacuolization and contraction impairment. However, little is known about how cholesterol reduction affects cardiomyocyte behavior. Here, we deplete cholesterol by incubating neonatal cardiomyocytes with methyl-beta-cyclodextrin. Traction force microscopy shows that lowering cholesterol increases the rate of cell contraction and generates defects in cell relaxation. Cholesterol depletion also increases membrane tension, Ca2+ spikes frequency and intracellular Ca2+ concentration. These changes can be correlated with modifications in caveolin-3 and L-Type Ca2+ channel distributions across the sarcolemma. Channel regulation is also compromised since cAMP-dependent PKA activity is enhanced, increasing the probability of L-Type Ca2+ channel opening events. Immunofluorescence reveals that cholesterol depletion abrogates sarcomeric organization, changing spacing and alignment of α-actinin bands due to increase in proteolytic activity of calpain. We propose a mechanism in which cholesterol depletion triggers a signaling cascade, culminating with contraction impairment and myofibril disruption in cardiomyocytes.

condition were analyzed. The total fluorescence is a measure of the total amount of caveolin-3 37 within that particular region. No statistical differences were found between analyzed groups. (B) 38 Standard deviation of the mean obtained for the same regions analyzed in (A). The standard 39 deviation informs how uniformly distributed caveolin-3 is within each analyzed region. Statistical 40 differences were found between control and cholesterol depleted groups. Cholesterol depleted 41 groups have a more uniformly distributed fluorescence pattern in comparison to control cells 42 and that corroborates what is depicted in Fig.4  using hydrazine hydrate (Sigma-Aldrich) 3 . Methods for traction force reconstruction have been 128 previously described in the literature 1,2,4-6 . Briefly, the fluorescent beads in the gel were imaged 129 at a frame rate of 2.5 images/second. Following the experiment, cells were detached from the 130 substrate using 0.5% sodium dodecyl sulfate and a reference image of the embedded 131 fluorescent beads was also taken. Images were aligned to correct for drift, and compared with 132 the reference image using particle imaging velocimetry software (http://www. 133 oceanwave.jp/softwares/mpiv/) in MATLAB to produce a displacement field with a grid spacing 134 of 1.43 mm 6 . Displacement vectors were filtered and interpolated using the Kriging interpolation 135 method. Traction stresses were reconstructed from the displacement field via Fourier Transform 136 Traction Cytometry 4,7 , using zeroth-order regularization. The same regularization parameters 137 were used for all datasets. Strain energy per cell area was also calculated. 138 139

Tether extraction from cardiomyocytes 140
Cardiomyocytes from control and cholesterol depleted samples were submitted to tether 141 extraction using an infrared optical tweezers (OT) setup. For this assay, polystyrene beads 142 (radius 1.52±0.02 μm, Polysciences, Warrington) were added to the culture dish containing the 143 cardiomyocytes and the dish was placed on the microscope. The OT captured a single 144 polystyrene bead and was used to press that bead against the surface of a chosen 145 cardiomyocyte for 5 seconds to allow bead attachment. After bead attachment, , the automated 146 microscope stage (Prior Scientific, Rockland, MA) was moved with a controlled and constant 147 speed (1 μm/s). Movies were taken during the tether extraction experiment using a CCD 148 Hamamatsu C2400 camera (Hamamatsu, Japan) coupled with a SCION FG7 frame grabber 149 (Scion Corporation, Torrance, CA) at a 10 frames/second capture rate. The OT setup and force 150 calibration were performed as previously described 8,9 . In order to calculate the bending 151 modulus κ and membrane tension σ we also measured the tethers radii by using Scanning 152 Electron Microscopy (SEM) according to previously published work 8-10 . 153 154

Tether extraction from plasma membrane vesicles 155
In order to separate the contribution from both plasma membrane and cytoskeleton to cortical 156 mechanical properties we decided to make plasma membrane vesicles (PMVs) as previously 157 reported 9,11 . Briefly, after incubating the cardiomyocytes with MβCD, we rinsed the cells and 158 exposed them to PMV solution (25mM formaldehyde, 20mM DTT, 2mM CaCl 2 , 10mM HEPES, 159 0.15M NaCl, pH 7.4). Cells were kept in this solution for 30 minutes in order to make PMVs. 160 After incubation, cells were rinsed carefully and fresh serum free DMEM was added before the 161 samples were submitted to tether extraction using the same setup as described above.

Measurement of Protein Kinase A (PKA) enzymatic activity 181
To measure cAMP-mediated PKA activity, cell extracts obtained from control and cholesterol 182 depleted cardiomyocytes were submitted to enzymatic assay using PKA activity kit (Enzo Life 183 Sciences total fluorescence is a measure of the total amount of caveolin-3 within that particular region. In 218 order to measure how homogeneous the caveolin-3 distribution gets after cholesterol depletion 219 we calculated the standard deviation of the mean for the same regions described previously. 220 The lower the standard deviation values the less heterogeneous the protein distribution 221 becomes. For measuring differences in distribution of the Ca v 1.2 subunit of LTCC between 222 perinuclear regions and the rest of the cell, upon cholesterol depletion, we calculated the ratio 223 between background corrected total fluorescence intensity measured for 5 distinct boxes near 224 the perinuclear region and for 5 other boxes away from the perinuclear region. The ratio values 225 decrease if Ca v 1.2 gets redistributed away from the perinuclear region. In order to measure how 226 heterogeneous the protein distribution gets after cholesterol depletion we also measured the 227 ratio standard deviation. The higher the standard deviation the higher is the discrepancy 228 between Ca v 1.2 distribution between perinuclear and away from the perinuclear regions. 229

Statistical analysis
We performed Student's t-test comparing control and cholesterol depleted groups. Data was 231 represented by mean values ± standard errors unless otherwise stated. Statistical differences 232 were labeled with asterisks. 233