Introduction

Primary mitochondrial myopathies (PMMs) lead to muscle fatigue and weakness. Resistance exercise has been shown to increase mitochondrial content in skeletal muscle in PMMs.^1,2 Outcome measures in clinical trials focus on increased oxidative capacity. Typically, this involves a muscle biopsy, which is invasive, samples a tiny volume of muscle and gives limited information on muscle function.³ Motor Unit Magnetic Resonance Imaging (MUMRI) has shown promise in measurement of muscle twitch dynamics.⁴ We used MUMRI to measure tibialis anterior (TA) twitch dynamics in PMM participants before and after a 12-week exercise programme.

Methods

Data Acquisition: Left lower leg muscles of 10 healthy controls (mean ± SD, age: 45.8±11.4) and 9 participants with single deletion PMM (mean ± SD, age: 59.6±10.7) were scanned with a pulsed gradient spin echo (PGSE) sequence with echo planar imaging readout (field of view=160 x 160, in-plane resolution=1.5x1.5 mm, slice thickness=8 mm, slices=2, TR=1000 ms, TE=36 ms, b value=20 s/mm², Δ/δ=16.9/2.2 ms). Their foot was strapped into an MR compatible force plate, with stimulating electrodes placed over the common peroneal nerve⁴ (figure 1A). First, the stimulation (I_muscle) was determined, i.e. the current that produced a visible muscle twitch and sufficient image contrast by gradually increasing the stimulation current from 0 mA to ~ 20 mA within ~20 acquisitions. The temporal shape of the muscle twitch of the TA was captured by repeating the PGSE sequence while stimulating the nerve with I_muscle and the stimulus gated between 400 ms before to 45 ms after the 90° radiofrequency pulse (in steps of 5 ms, 90 acquisitions; figures 1B-D). Repeatability of the technique was tested by repeating the scan twice in 5 controls.

Data Analysis: Images were registered to the first image within the time series using a non-rigid registration in Matlab. The TA was delineated using a manual segmentation in ITKSnap. Voxel-wise, analysis was performed in Matlab, the time-series data from the region were smoothed and interpolated to a time step of 0.5ms. For each time-series four points were automatically calculated: the beginning of the signal change (T_start), defined as the first time point at which the signal decreased to a value 3 times less than the standard deviation of the baseline, the first signal minima (T_min1), in between the first and second signal minima (T_middle) and the second signal minima (T_min2), figure 2. Voxel-wise twitch rise time (T_rise), contraction time (T_contract) and half-relaxation time (T_half-relax) of the TA were then calculated and the average of all voxels taken (figure 3).

Exercise Programme: PMM participants underwent a 12-week home-based exercise programme (Giraffe Healthcare) with fortnightly remote monitoring. Exercises were predominantly lower limb, with mandatory exercises to activate TA. Participants recorded completed exercises in a diary, used to calculate compliance to the exercise programme and compared to the percentage change in T_rise, T_contract and T_half-relax between baseline and post exercise programme.

Statistics: T_rise, T_contract and T_half-relax were compared between controls and PMM participants at baseline using unpaired parametric students t-tests. Repeatability was assessed for each variable using the absolute average percentage difference between the two scans. T_rise, T_contract and T_half-relax were compared for PMM participants between baseline and post exercise programme using paired parametric students t-tests. Results are presented as mean ± standard deviation.

Results

Baseline: No significant differences in T_rise, T_contract and T_half-relax were detected between control and PMM participants, for T_rise: controls-39.1±1.3 vs. PMM participants-39.4±2.5ms; p=0.761, for T_contract: 115.8±11.3 vs. 110.6±10.0ms; p=0.300 and for T_half-relax: 101.9±13.8 vs. 104.8±21.9ms; p=0.729 (figure 4A).

Repeatability: Data were reproducible between the two scans, absolute percentage differences were (T_rise=3.4%, T_contract=6.4% and T_half-relax=7.1%), figure 4B.

PMM participants: Two participants were lost to follow-up. T_contract was significantly longer: baseline-108.7±7.9 vs. post-119.3±10.4ms; p=0.018. T_rise was longer but not significant post exercise programme: 39.2 ± 2.8 vs. 40.6 ± 1.5ms; p=0.159. There was no difference in T_half-relax: 103.0±23.1 vs. 102.1±16.0ms; p=0.811 (figure 5A-B). PMM participants who had the highest overall increase in T_contract demonstrated the highest adherence to the exercise programme (figure 5C).

Discussion & Conclusions

Contraction times were significantly longer in PMM participants post-exercise programme. This may be due to the increased activity in type I fibres due to the exercise, leading to an increase in mitochondria and overall slower muscle twitch in the TA. Participants with the highest adherence to the exercise programme demonstrated the greatest increase in T_contract, supporting the objective changes.
In future, it will be important to compliment the observed changes in muscle twitch dynamics with ³¹P MRS to measure the metabolic recovery post exercise and use a muscle biopsy measure accompanying fibre type changes post exercise.

Acknowledgements

The authors are grateful to the participants who dedicated their time to take part in the study and the radiographers: Tim Hodgson, Dorothy Wallace and Louise Ward for scanning the participants. PMM participants were recruited through the Wellcome Centre for Mitochondrial Research Patient Cohort: A Natural History Study and Patient Registry (REC Ref: 13/NE/0326) and the Newcastle NHS Highly Specialised Services (HSS) for rare mitochondrial disorders.