Mitochondrial CaMKII causes adverse metabolic reprogramming and dilated cardiomyopathy

Despite the clear association between myocardial injury, heart failure and depressed myocardial energetics, little is known about upstream signals responsible for remodeling myocardial metabolism after pathological stress. Here, we report increased mitochondrial calmodulin kinase II (CaMKII) activation and left ventricular dilation in mice one week after myocardial infarction (MI) surgery. By contrast, mice with genetic mitochondrial CaMKII inhibition are protected from left ventricular dilation and dysfunction after MI. Mice with myocardial and mitochondrial CaMKII overexpression (mtCaMKII) have severe dilated cardiomyopathy and decreased ATP that causes elevated cytoplasmic resting (diastolic) Ca2+ concentration and reduced mechanical performance. We map a metabolic pathway that rescues disease phenotypes in mtCaMKII mice, providing insights into physiological and pathological metabolic consequences of CaMKII signaling in mitochondria. Our findings suggest myocardial dilation, a disease phenotype lacking specific therapies, can be prevented by targeted replacement of mitochondrial creatine kinase or mitochondrial-targeted CaMKII inhibition.


Supplementary
: mtCaMKII mouse data a Summary data for HW (WT n=16, mtCaMKII n=22), BW (WT n=16, mtCaMKII n=22) and HW/TL (WT n=10, mtCaMKII n=14) measurements for WT and mtCaMKII mice. b Representative images and c summary data for quantification of cardiomyocyte cross sectional area in WT (n=4 hearts) and mtCaMKII (n=5 hearts) heart sections. Data are represented as mean ± SEM, significance was determined using two-tailed Student's t test. Source data are provided as a Source Data file.

Supplementary Figure 5: CK expression in transgenic and MI hearts
a Western blot for CKmito and CK-M of cytosolic and mitochondrial fractions from WT mouse hearts. This experiment was repeated independently twice with similar results. b Representative western blot and summary data for Ckmito in sham (n=4) and MI (n=5) hearts. c Summarized echocardiographic measurements from WT (n=7), CK-M (n=12), mtCaMKII (n=5), and mtCaMKII x CK-M (n=3) mice. Data are represented as mean ± SEM, significance was determined using 1 way ANOVA with Tukey's multiple comparison's test. ****P<0.0001, ***P<0.001, **P<0.01. Source data, including exact p values, are provided as a Source Data file.

Supplementary Figure 6: Computational Model Outputs
ATP hydrolysis and cytosolic calcium were simulated using pulsatile functions to simulate cardiac contraction. Mitochondrial creatine kinase flux, myocardial creatine kinase flux, mitochondrial ATP synthesis rate, and oxygen consumption as well as ATP, ADP, PCr, Pi levels, and mitochondrial NADH, membrane potential, H 2 O 2 and Ca 2+ were compared in WT, CKmito, mtCaMKII, and mtCaMKII x CKmito interbred conditions after 5 minutes of simulation. Insets provide zoomed in scale to show small oscillations. Only WT is shown in the inset for clarity.  Summary of the parameters used in the computational model that were taken from the experimental data for WT, CKmito, mtCaMKII, and mtCaMKII x CKmito.

TTC staining and MI sizing
At the end of 24 hours post-MI, mice were sacrificed, the hearts quickly excised, and both atria and right ventricular free wall were removed. Left ventricular tissue was weighed and wrapped with cling film and then frozen at -20°C for 45 min, and then sliced into 6-7 pieces of 1.0 mm thick sections perpendicular to the long axis of the heart. The sections were incubated individually using a 24-well culture plate in 1% TTC in phosphate-buffered saline at pH 7.4 at 37°C for 10 min), and then digitally photographed. For infarct size at 24 hours post-MI, TTCstained area, and TTC-negative staining area (infarct myocardium) were measured using ImageJ. Myocardial infarct size was expressed as a percentage of the total LV area.

Immunofluorescence
In brief, hearts were fixed in 4% PFA and embedded in OCT freezing compound (Fisher Scientific, Waltham MA, USA). 10 µm cardiac two chamber view sections were cut using a Microm HM 550 cryostat. Sections were post-fixed with 4% paraformaldehyde and permeabilized using 0.1% Triton-X in PBS. Sections were washed two times for 5 minutes with 1x HBSS. Sections were incubated with anti-CD45 (Thermo Fisher, Waltham, MA, USA) (#14-0451-82 at 1:50) according to previously published methodology 1 . Samples were subsequently incubated with Alexafluor 555 at 1:1000 (Thermo Fisher A21434). At least 3-5 fields of view per infarct and remote zone were assessed from n=5 hearts per genotype. Images were acquired in 0.63 µm Z-stacks using an Olympus IX83 microscope, 60x/1.42NA PLAPON objective. CD45+, nucleated cells were counted and normalized number of cells/area (mm 2 ).

Myocyte cross-sectional area
In brief, hearts were prepared as above. Sections were incubated with 5 µg/mL in Wheat Germ Agglutinin, Alexa Fluor® 488 conjugate (Molecular Probes, Eugene, OR, USA) for 20 minutes. Samples were washed two times for 5 minutes with 1x HBSS and mounted with Prolong Diamond Anti-Fade Reagent (Molecular Probes, Eugene, OR, USA). Ten fields of view were acquired from each animal. At least four animals were processed per genotype. 0.63 µm Zstacks were collected using an Olympus IX70 microscope, 40x/0.75NA UPLFLN-PH objective, and deconvolved using constrained iterative deconvolution software in the cellSens software suite. Average projections for each field of view were calculated and analyzed using cellProfiler (Cell Profiler Inc. Boston, MA, USA) according to previously published methods to determine cell size.

Computational Model Validation
The model was constructed by integrating two computational models: Kongas and van Beek, 2007 creatine kinase model 2 and Gauthier et al, 2013 cardiac bioenergetic model 3 . The model parameters were directly taken from these two models unless stated otherwise. The metabolite steady-state levels in the control group were matched to the original model ranges as model validation. Simulation codes are available at the URL: https://gitlab.com/MitoModel/mtCaMKII.git. Experimental data for validation is as follows: 1. Mouse heart rate is around 500 bpm (400-600), which is approximated to 120 ms/beat. Ordinary differential equations used are as follows: Mitochondrial ions: (1) [Na ] = − [H ] Citric acid cycle: [ KG] = + − + [NADPH] [PSSG] = , − , High energy phosphates: The ATP hydrolysis pulse 2 : The initial conditions used for calculations are as follows: The creatine kinase models for both mitochondrial IMS and cytosolic compartments are based on a sequential, rapid equilibrium, random bi-bi enzymatic reaction scheme 2 .
(67) ATP + Cr ⇌ ADP + PCr TCA cycle rates were calculated as follows 9 : Citrate synthase (CS): Mitochondrial transporters and ion channels we calculated as follows: Phosphate carrier 9 follows equilibrium random Bi:Bi reaction kinetics.  Mitochondrial electron transport was calculated as follows: Complex I 3 , assuming single electron for each cycle.