The primary finding of this study was that electrical stimulation of the hemidiaphragm attenuated integrative and intrinsic mitochondrial respiratory function in the diaphragm muscle. Specifically, leak respiration, and OXPHOS and ETS capacities were lower, and spare capacity was reduced after electrical stimulation in muscle biopsy samples from the stimulated side.
The current findings contrast with our previous work, where hourly hemidiaphragm stimulation over the course of prolonged cardiovascular surgery significantly increased state 3 (PCI) and state 4 (leak) respiration . In the same subjects presented here, this supramaximal phrenic stimulation regimen offset contractile dysfunction and atrophy of slow diaphragm fibers (Bresciani et al., under review). A precise intraoperative quantification of the stimulation is not feasible, but the supramaximal parameters were targeted to preserve maximal force production. Although we are not aware of existing data on mitochondrial capacity after hemidiaphragm electrical stimulation, there is evidence that other types of intense exercise bouts could acutely disrupt mitochondrial function in skeletal muscle. For example, acute high-intensity exercise (ergometer 5-km time-trial) in young healthy subjects diminished ATP-linked respiratory capacity and spare capacity immediately after an exercise bout, without differences in CS activity and proton leak . Repeated exercise bouts with recovery periods (exercise training) induce adaptive responses which lead to performance and strength improvement; however, excessive workload during particular exercise sessions blunts adaptive response to exercise training, such as mitochondrial respiration[19–20]. For example, in response to an 8-week lower-limb maximal strength training in older adults, muscle strength was improved, however the mitochondrial respiration was impaired, suggesting qualitative mitochondrial adaptations to accommodate increase in force vs aerobic capacity. In another study, after exercise training including high-intensity ergometer sprints performed on 7 consecutive days over 15 days, mitochondrial respiration was profoundly diminished by up to 65%, driven by attenuation of aconitase activity by up to 72%.
We then investigated whether the observed overt mitochondrial dysfunction was associated with mitochondrial stress and general cellular oxidative damage. Expression of 4-HNE, a marker of lipid peroxidation, was unaffected by electrical stimulation. Mitochondrial aconitase, which has been suggested to be a sensitive indicator of mitochondrial oxidative stress , displayed a lower enzymatic activity in samples from the stimulated side. Interestingly, there was a positive correlation of (full-length) PINK-1 with PCI+II and ECI+II in samples from the stimulated side, which was absent in those from the unstimulated side. Subcellular localization of full-length PINK-1 to the mitochondria is regulated by the mitochondrial membrane potential, and increases with mitochondrial depolarization . Once stabilized at the mitochondrial membrane, PINK-1 recruits Parkin to the damaged mitochondria and the process of mitophagy, the removal of damaged mitochondria through autophagy ensues . The association and correlation of PINK-1 accumulation with higher respiratory activity that we observed only in electrically stimulated muscle fibers, together with less functional mitochondrial aconitase, could be indicative of increased oxidative stress in the stimulated muscle fibers. We did not, however, detect an increase in Parkin or p62 protein, the latter a selective cargo receptor associated with autophagy and mitophagy , and both downstream of PINK-1 recruitment to the outer mitochondrial membrane. It is possible that the observed PINK-1 accumulation indicates an early stage of mitochondrial stress. In our previous pillow work, we did not investigate markers of mitophagy, but found an increase in autophagy markers Beclin-1 and LC3II/I ratio, suggesting induction of autophagy following the stimulation protocol . While autophagy per se could be beneficial by facilitating removal of damaged cellular material, it could as such also be consistent with cellular damage.
While we did not find a similar benefit for mitochondrial function in the current study, some key differences between the current study and our pilot project may have contributed to the discrepancy. These include differences in the volume/dose of stimulation, the severity of intraoperative cooling, and the baseline respiratory function of participants. Across a similar intraoperative time, an average of 3.4±0.6 hourly stimulations were administered in the previous study, as compared to 6.2±1.9 stimulations every 30 minutes in the current study. Despite similar stimulation intensities (17±4 vs 18 ±5 mA) in both studies, the number of stimulation bouts was > 80% greater in the current study. Since the “dose” of stimulation necessary to optimize mitochondrial respiration has not yet been determined, we cannot verify whether the increased stimulation volume was detrimental to mitochondrial function.
Besides the actual stimulation protocol, intraoperative temperature was another difference between the previous and current study that could explain the different outcomes. Although diaphragm biopsies were acquired after rewarming, the surgical procedures in the previous study were associated with more prolonged and extensive intraoperative cooling (31.3oC ± 2.6 degrees, versus 34.8oC ± 1.1 degrees in the current trial. In animal studies, heating skeletal muscle by 5oC decreased mitochondrial respiration in response to a fatty acid substrate without changes in CS enzyme activity . In another study in isolated resting rat skeletal muscle, increasing temperature increased mitochondrial oxidative capacity and proton leak, which could explain decreased efficiency of OXPHOS .
Additionally, we speculate that differences in the baseline pulmonary function of the current study participants could contribute to the study findings. Despite similarities in age and BMI of subjects, average spirometry and maximal inspiratory pressures (a clinical estimate of strength) were normal in the current sample (FVC: 87±14% predicted and maximal inspiratory pressure: 91±23) cm H2O) and considerably higher than our prior reports (Ahn et al: FVC: 54±19% predicted, MIP: 62±3 cm H2O; Mankowski et al: FVC: 78±17% predicted, MIP: 72±15 cm H2O) [9, 27]. Animal  and human  studies reveal a reduced skeletal muscle mitochondrial density, impaired mitochondrial respiration, and excessive oxidative stress in obstructive lung disease. Thus, the contraction stimulus may have elicited relatively smaller gains in muscle fiber function, when starting from normal baseline diaphragm function. Since acute, intense exercise can transiently decrease mitochondrial function in skeletal muscle of both control and lung disease subjects  the observed differences between the studies are likely more complex, reflecting combined influences of both pre-existing patient function and the phrenic stimulation protocol.
While intraoperative phrenic stimulation led to differences in mitochondrial function, compared to the non-stimulated hemidiaphragm, mitochondrial respiration does not appear to significantly contribute to diaphragm contractile dysfunction in critically ill adults with prolonged mechanical ventilation . However, intraoperative mechanical ventilation does not represent a pure model of VIDD, since other aspects of the surgical environment including cardiopulmonary bypass, anesthesia, and hypothermia[31–33] may independently affect mitochondrial respiration, and the surgical process itself can alter the mechanics of the respiratory muscles. Taken together, our data on mitochondrial stress and/or damage, possibly in an early stage, could have been influenced by the specific electrical stimulation protocol and the clinical management employed in this study. Future investigations need to clarify whether this protocol leads to mitochondrial and cellular oxidative damage.
Strengths and limitations
A strength of the study design is that it incorporates a clinically-relevant model of mechanical ventilation that can lead to prolonged ventilatory support. One limitation of this research model is some necessary variability of the surgical environment required to meet each subject’s clinical needs. In result, we note differences in the total amount of anesthetic, minimal core body temperature, and duration of the surgical procedure, which also impacted the total number of stimulations. Considerable heterogeneity was also observed in the study dependent metrics. We evaluated the strength of association between surgical factors and subsequent mitochondrial respiration and protein expression, but did not identify any significant associations that could account for the variable subject responses. Similarly, no subject characteristics could account for the observed variability in mitochondrial respiration, protein expression, mitophagy, and enzyme activity.
While most subjects were recruited in advance of open cardiothoracic surgery, two participants in the study (subject 8 and 9) underwent bilateral lung transplantation due to end-stage restrictive pulmonary disease. The remaining study participants were relatively young and active with few pulmonary comorbidities and thus considered to have a lower risk for post-operative ventilatory failure. We compared the study dependent measures of the full sample, to those with the transplant patients excluded, and excluding these subjects did not change the primary study findings.
Unilateral phrenic stimulation reduced variability by using each subject’s non-stimulated hemidiaphragm as an inactive control. Supramaximal stimulation of one hemidiaphragm could induce passive stretch of the unstimulated side via force transduction from the central tendon . Indeed, intermittent passive stretch elicited by unilateral diaphragm denervation elicits high passive stretch and titin-mediated fiber hypertrophy (both cross-sectional and longitudinal) . Passive stretch also increases mitochondrial calcium concentrations, which stimulates mitochondrial respiration . Thus, stimulation-induced compensatory changes in mitochondrial respiration, mitophagy, and protein expression may have been underestimated. Further studies of bilateral stimulation may indicate more clinically-relevant benefits of phrenic stimulation.