Exercise Training Enhances Platelet Mitochondrial Bioenergetics in Stroke Patients: A Randomized Controlled Trial

Exercise training (ET) may impact physical fitness by affecting mitochondrial functions. This study aimed to elucidate the effect of ET on aerobic capacity and platelet mitochondrial bioenergetics (MTB) in stroke patients. Among the 30 stroke patients who underwent the traditional rehabilitation program (TRP), 15 were randomly assigned to have ET (50–60% VO2peak for 30 min/day, 5 days/week for 4 weeks), and those remaining received only the TRP (control group). The peak exercise capacity (VO2peak) and platelet MTB, including oxidative phosphorylation (OXPHOS) and the electron transport chain (ETC), were measured through automatic gas analysis and high-resolution respirometry, respectively. The results demonstrated that ET significantly increased the VO2peak (17.7%) and O2 uptake efficiency slope (31.9%) but decreased the ventilation versus CO2 production slope (−7.65%). Patients who underwent ET also had significantly enhanced platelet mitochondrial OXPHOS and ETC by activating the FADH2 (Complex II)-dependent pathway, but depressed plasma myeloperoxidase (−28.4%) and interleukin-6 levels (−29.9%). Moreover, changes in VO2peak levels were positively correlated with changes in platelet OXPHOS and ETC capacities. In conclusion, ET increases the platelet MTB by enhancing Complex II activity in stroke patients. The exercise regimen also enhances aerobic fitness and depresses oxidative stress/pro-inflammatory status in stroke patients.


Introduction
Stroke, caused by cerebral vascular atherosclerosis/thrombosis or intracranial hemorrhage, is the leading cause of long-term disability, which produces enormous global health and economic burdens [1,2]. Stroke-induced functional impairment reduces exercise capacity, which negatively affects the ability of patients to perform their daily activities. This limitation further decreases their independence and quality of life [1,2]. Although a randomized clinical study for 61 chronic stroke patients showed exercise training (ET) improved cardiovascular fitness [3] and a scientific overview encouraged post-stroke survivors to consider the importance of ET in regaining activities of daily living [4], controversy persists regarding how ET influences the hemostasis in stroke patients or stroke prevention [1,5].

Participants
A total of 52 stroke patients diagnosed by the neurologist were surveyed from August 2017 to July 2018 in a tertiary care hospital. Brain computed tomography was used to diagnose the hemorrhagic or ischemic stroke. The ischemic stroke subtype was determined according to the TOAST classification system [21]. The inclusion criteria were listed as follows: (I) ≥20 years old; (II) stable stroke events ≥3 months; (III) mini-mental state examination (MMSE) >24; (IV) no acute coronary syndrome; (V) lower extremity function > Brunnstrom stage III with active voluntary movement. Those who had unstable angina, systolic blood pressure >200 mmHg or diastolic blood pressure >110 mm Hg, symptomatic orthostatic hypotension, severe aortic stenosis (peak systolic pressure gradient >50 mmHg, or an aortic valve opening area <0.75 cm 2 ), inflammatory disease within the previous 3 months, uncontrolled cardiac dysrhythmias, uncompensated heart failure (HF), third-degree atrioventricular block, pericarditis or myocarditis within the previous 3 months, embolic disease within the previous 3 months, ST segment displacement ≥2 mm at rest, or uncontrolled diabetes (blood glucose ≥300 mg/dL or ≥250 mg/dL with ketone bodies) were excluded from the study.
All our eligible candidates did not engage in regular physical exercise except 30-45 min of supervised in-hospital traditional rehabilitation programs (TRP), which included balance, neurofacilitation, and therapeutic exercise, as instructed by their rehabilitation physicians, before starting intervention. They were randomly assigned to the intervention (under ET and TRP) and control (under TRP only) groups ( Figure S1) based on the computer-generated random allocation sequence. An independent physician performed the randomization procedure, which was concealed to the members performing the intervention and the enrolled participants before starting the experiment. All subjects provided informed consent after the experimental procedures were explained. The Institutional Review Board of a tertiary care hospital approved this study (IRB No. 201602058A3) and the clinical trial registration number was NCT03960918.

Graded Exercise Test (GET)
The physician performing the GET was blinded to the subject grouping. He assessed their cardiopulmonary fitness 2 days before and 2 days after 4-week intervention. Moreover, the data collector was isolated from the analytic specialist. Each participant underwent the GET at an incremental work-rate of 10 W/min on a bicycle ergometer (Ergoselect 150P, ergoline GmbH, Bitz, Germany) with a continuous monitoring heart rate (HR), brachial blood pressure, and arterial O 2 saturation, until the stopping conditions described previously [13,14]. O 2 consumption (VO 2 ), minute ventilation (V E ), and carbon dioxide production (VCO 2 ) were measured breath by breath by a cardiopulmonary measurement device (MasterScreen CPX, CareFusion Corp., Hoechberg, Germany), and the VO 2peak was defined as the guideline for exercise testing suggested by the American College of Sports Medicine (ACSM) [22]. V E and VCO 2 responses acquired from the initiation of exercise to the peak values were used to calculate the V E -VCO 2 slope using least-square linear regression [23]. The O 2 uptake efficiency slope (OUES), an estimation of the efficiency of O 2 consumption during exercise, was derived from the slope of a natural logarithm plot of V E vs. VO 2 . A greater slope indicated a higher ventilation efficiency [23,24]. Additionally, a 6-min walk test (6 MWT) was used to screen functional capacity and exercise endurance in stroke patients [25].

Cardiac Hemodynamic Measurements
A noninvasive continuous cardiac output monitoring system (NICOM, Cheetah Medical, Wilmington, DE, USA) was used to evaluate cardiac hemodynamic response to exercise, which analyzes the phase shift (∆Φ) created by alternating electrical current across the chest of the subject, as described in our previous study [26].

Health-Related Quality of Life
Quality of life (QoL) was measured using a Short Form-36 Health Survey questionnaire (SF-36), the findings of which were related to the co-morbidities of stroke patients [27].

Exercise Training (ET) Protocol
A 30-45 min of TRP was implemented for all included stroke survivors, which was judged by physician discretion. The control subjects only engaged in the TRP. In addition to the TRP, the subjects in the intervention group underwent supervised hospital-based training 20 times (5 session/week for 4 weeks) on a bicycle ergometer (Ergoselect 150P, Bitz, Germany) as our previous protocol describes [26]. The ET comprised a warm-up at 30% of VO 2peak for 3 min, followed by continuous 60% of VO 2peak for 30 min, then a cool-down at 30% of VO 2peak for 3 min. The training was terminated when the subject had symptoms/signs suggested by the ACSM guideline [22]. The compliance rate of all our subjects was 100% ( Figure S1).

Blood Cell Count
Based on our previous work [28], an amount of 25 mL of blood was sampled just prior to each exercise test performed. The first 2 mL was discarded, then the remaining blood sample was used for the measurements of hematological parameters and platelet function. Each blood sampling was performed after holding medication (anti-platelet or anticoagulant) for 48 h, stroke prevention medication was restarted after the procedure. The blood analysis was performed initially and after 4 weeks of ET. Blood cells were enumerated using a Sysmax SF-3000 cell counter (GMI Inc., Ramsey, MN, USA) and were repetitively analyzed twice to ensure the reproducibility of the results.

Platelet Isolation
Samples, processed as the above description to avoid platelet activation [28], were collected in polypropylene tubes containing a sodium citrate concentration of 3.8 g/dL. Platelet rich plasma (PRP) was prepared through centrifugation at 300× g for 10 min at approximately 20 • C. Platelets were sedimented through centrifugation of the PRP at 1500× g for 10 min at approximately 20 • C and then were washed once with PBS containing 4 mM ethylenediaminetetraacetic acid (EDTA) (Sigma-Aldrich, St. Louis, MO, USA) to inhibit platelet activation [14,15]. They were mixed with mitochondrial respiration medium (MiR05, containing sucrose 110 mM, HEPES 20 mM, taurine 20 mM, K-lactobionate 60 mM, MgCl 2 3 mM, KH 2 PO 4 10 mM, EGTA 0.5 mM, BSA 1 g/L, pH 7.1) to a final concentration of 2 × 10 8 cells/mL. All platelet fractions were analyzed within 2 h after cell purification.

High-Resolution Respirometry
Platelet mitochondrial respiration was measured by a high-resolution respirometry (Oxygraph-2K, Oroboros Instrument, Innsbruck, Austria) with a stirrer speed of 750 rpm at a constant temperature of 37 • C. Data were acquired and recorded every 2 s by DatLab software version 6 (Oroboros Instrument, Innsbruck, Austria). Platelets of 2 × 10 8 cells/mL were added to the glass chamber filled with 2 mL mitochondrial respiration medium (MiR05) for measurement [14,15]. The O 2 concentration was automatically calculated from barometric pressure and the solubility factor was 0.92 for MiR05.

Mitochondrial ETC and OXPHOS in Platelets
The substrate-uncoupler-inhibitor titration reference protocol (SUIT-RP), consisting of two mitochondrial substrate-controlled experiments (RP1 and RP2), was applied to acquire the platelet mitochondrial bioenergetics ( Figure S2A-C). All the chemicals were purchased from the Sigma-Aldrich Corporation.
The SUIT-RP1 was used to measure the mitochondrial ETC capacity ( Figure S2A). The cell membrane was permeabilized with digitonin. Malate (2 mM) and pyruvate (5 mM), NADH-linked (N-linked) substrates were subsequently added [14,15]. O 2 consumption was only driven by uncoupling proton leakage (LEAK state, PM L ) because the ADP was absent. The OXPHOS capacity (PM P ) was then evaluated by 1 mM ADP (Calbiochem, St. Louis, MO, USA) titration. The ETC capacity driven by malate and pyruvate (PM E ) was obtained by mitochondrial protonophore carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP) titration (0.5 µM/step) until no further respiration increased. Glutamate (10 mM) was added to access the maximal ETC capacity driven by NADH or mitochondrial complex I (CI)-related resources (MPG E ). Thereafter, 10 mM succinate was used to complete the convergent input of CI and complex II (CII) (MPGS E ). Octanoyl-carnitine (Oct) titration (0.5 mM) was performed to evaluate the additional effect of fatty acid oxidation (FAO) (MPGSOct E ). The N-linked substrate-dependent respiration and FAO pathways (S E ) were blocked by 0.5 µM rotenone (CI inhibitor). The additional contribution of the mitochondrial complex of glycerophosphate dehydrogenase (CGpDH) was measured with 10 mM glycerophosphate (SGp E ). At last, the mitochondrial respiration was inhibited by 2.5 µM antimycin A, the mitochondrial complex III inhibitor ( Figure S2A).
The mitochondrial ETC capacities in the permeabilized platelets were obtained from the following equations: The SUIT-RP2 was to measure the ATP synthase-dependent OXPHOS state ( Figure S2B). Following routine respiration and digitonin permeabilization, 1 mM ADP was titrated to accelerate the depletion of residual endogenous substrates. Then, 0.5 mM OC and 0.1 mM malate were added to evaluate the FAO. An amount of 2 mM Malate was used to reduce the flux from succinate dehydrogenase. This procedure could also activate the mitochondrial malic enzyme and led to higher O 2 consumption (OctM P ). An amount of 5 mM pyruvate (OctMP P ) and 10 mM glutamate (OctMPG P ) were sequentially titrated to acquire the N-linked respiration. A consecutive titration of 10 mM succinate and 10 mM glycerophosphate stimulated the mitochondria respiration through CII (OctMPGS P ) and CGpDH (OctMPGSGp P ) input. To evaluate the limitation of the OXPHOS system, FCCP (0.5 µM/step) was titrated to access the maximal ETC capacity (OctMPGSGp E ). An amount of 0.5 µM rotenone was added to reach the identical SGp E state in the RP1 and RP2 protocols. Antimycin A (2.5 µM) was added in the final step to block the whole mitochondria respiration.
The capacities of mitochondrial OXPHOS in permeabilized platelets were calculated from the following equations ( Figure S2B): OXPHOS FAO+CI+CII (OctMPGS P ) = ADP + Oct + Malate + Pyruvate + Glutamate + Succinate (8) This study defined the maximal capacities of OXPHOS and ETC in platelets as the OctMPGSGp P and OctMPGSGp E states in the RP2 protocol, respectively ( Figure S2B).

Biomarkers of Oxidative Stress and Pro-Inflammation in Plasma
A 5-mL blood sample was placed in a cold centrifuge tube containing 4 mM EDTA, and was immediately centrifuged at 3000× g for 10 min at 4 • C. Plasma myeloperoxidase (MPO) (Immunology Consultants Lab., Newberg, OR, USA) and interleukin-6 (IL-6) (eBioscience, SanDiego, CA, USA) levels were quantified by commercially available ELISA kits [26]. Prominent increases of TNF α and IL-1 occur within few hours in the inflammatory cascade and may be inadequate for the study for chronic stroke patients [29]. Therefore, IL-6 and MPO were reviewed in the study because they have been reported to be associated with chronic low-grade inflammation and good responses to exercise related anti-inflammatory effects [29,30].

Statistical Analysis
Data were expressed as mean ± SEM and were analyzed using the statistical software package StatView. The experimental results were analyzed by 2 (groups) × 2 (time sample points; i.e., preand post-interventions) repeated measures ANOVA. Bonferroni's post hoc was used to compare cardiopulmonary fitness, biomarkers of plasma oxidative stress and inflammation, and platelet mitochondria functions at the beginning of this study and after 4 weeks of observation in both groups. Pearson correlation analysis was used to determine the correlation between the VO 2peak and platelet mitochondrial OXPHOS and ETC in stroke patients. p values lower than 0.05 were considered statistically significant.

Cardiopulmonary Fitness and Health-Related QoL
A total of 30 subjects completed the study in the intervention (n = 15) and control (n = 15) groups ( Figure 1). No adverse vascular or thrombotic event occurred in the two groups during the investigation. Differences of baseline anthropometric and clinical parameters between the two groups were not significant (Table 1). Among the included subjects, baseline clinical information and comorbidities of the hemorrhagic stroke patients in the two groups is shown in the Table S1. All our hemorrhagic stroke patients suffered from hypertension. The initial cardiopulmonary fitness of the two groups was also similar ( Table 2). post-interventions) repeated measures ANOVA. Bonferroni's post hoc was used to compare cardiopulmonary fitness, biomarkers of plasma oxidative stress and inflammation, and platelet mitochondria functions at the beginning of this study and after 4 weeks of observation in both groups. Pearson correlation analysis was used to determine the correlation between the VO2peak and platelet mitochondrial OXPHOS and ETC in stroke patients. p values lower than 0.05 were considered statistically significant.

Cardiopulmonary Fitness and Health-Related QoL
A total of 30 subjects completed the study in the intervention (n = 15) and control (n = 15) groups ( Figure 1). No adverse vascular or thrombotic event occurred in the two groups during the investigation. Differences of baseline anthropometric and clinical parameters between the two groups were not significant (Table 1). Among the included subjects, baseline clinical information and comorbidities of the hemorrhagic stroke patients in the two groups is shown in the Table S1. All our hemorrhagic stroke patients suffered from hypertension. The initial cardiopulmonary fitness of the two groups was also similar ( Table 2).    Following 4 weeks of interventions, the stroke patients underwent ET displayed significantly increased work-rate, heart rate, stroke volume (SV), cardiac output (CO) as well as, V E , VCO 2 , OUES, VO 2peak , and 6 MWT distance (p < 0.05). In addition, a significant decrease of V E -VCO 2 slope (p < 0.05) was detected in stroke patients with ET. However, significant changes in cardiopulmonary fitness were not identified in those with 4 weeks of traditional rehabilitation training. In our observation, the TRP did not influence indicators of ventilatory efficiency and functional capacity in stroke patients. Detailed information is shown in Table 2.
In health-related QoL, 4 weeks of ET significantly increased the subclass scores of the physical (scores from 40.1 to 44.2, p < 0.05) and mental (scores from 36.1 to 44.9, p < 0.05) dimensions (Table 2). However, QoL in the stroke patients who underwent 4 weeks of TRP remained unchanged (Table 2).

Hematologic Parameters and Oxidative Stress/Inflammatory Biomarkers
There were no significant changes in hemograms (i.e., erythrocyte, hemoglobin, hematocrit, leukocyte, and platelet) following 4 weeks of TRP with or without ET (Table 1). However, ET associated with TRP considerably reduced the plasma levels of MPO (p < 0.05) and IL-6 (p < 0.05). The oxidative stress and pro-inflammation in plasma were unchanged following the TRP alone (Table 1).

Mitochondrial ETC and OXPHOS in Platelets
Neither TRP + ET nor TRP alone changed the routine respiration and uncoupling proton leakage of platelets (Figures 1 and 2). Moreover, the two therapeutic regimens were also insufficient to influence platelet ETC driven by NADH (CI)-dependent pathway (PM E and PMG E , Figure 1A,B). The OXPHOS driven by NADH (CI)-and FAO-dependent pathways (Oct p , OctM p , OctMP p , and OctMPG p ) were not affected (Figure 2A  On the other hand, TRP + ET also enhanced succinate-involved platelet OXPHOS levels ( Figure 2A, p < 0.05), such as OctMPGS P , rather than OctMPG P . Furthermore, the exercise regimen also augmented the capacities of the maximal OXPHOS (OctMPGSGp P , p < 0.05) and ETC (OctMPGSGp E , p < 0.05) in platelet (Figure 2A). Additionally, changes in VO 2peak levels were positively associated with changes in OctMPGSGp P ( Figure 4A, r = 0.587, p < 0.001) and OctMPGSGp E in platelet ( Figure 4B, r = 0.581, p < 0.001). However, no significant changes in platelet mitochondrial bioenergetics, including ETC and OXPHOS capacities, were observed in the control subjects (Figures 1 and 2). Effects of TRP with (A, TRP + ET) and without (B, TRP) exercise training in platelet mitochondrial O2 consumption rate (OCR) using the reference protocol 2 (RP2) protocol in patients with cerebral vascular accident. P, pyruvate; M, malate; G, glutamate; S, succinate; Oct, octanoylcarnitine; Gp, glycerophosphate; Rot, rotenone; Ama, antimycin; ROX, residual O2 consumption; OXPHOS or P, oxidative phosphorylation; ETC or E, electron transport system; LEAK or L, uncoupling proton leakage. Pre-intervention vs. post-interve ntion, * p < 0.05. Values were mean ± SEM.

Discussion
The present work exhibited that 4 weeks of moderate-intensity ET improved exercise capacity and ventilatory/hemodynamic efficiency, and physiological adaptation was accompanied by reduced oxidative stress and pro-inflammatory status in stroke patients. The ET-induced relief of oxidative stress in both ischemic and hemorrhagic stroke patients in the study might reverse the systemic coagulopathy and may be helpful in ischemic and hemorrhagic stroke prevention. In addition to the above interesting clinical findings, two novel reference protocols (SUIT-RP1 and SUIT-RP2) were

Discussion
The present work exhibited that 4 weeks of moderate-intensity ET improved exercise capacity and ventilatory/hemodynamic efficiency, and physiological adaptation was accompanied by reduced oxidative stress and pro-inflammatory status in stroke patients. The ET-induced relief of oxidative stress in both ischemic and hemorrhagic stroke patients in the study might reverse the systemic coagulopathy and may be helpful in ischemic and hemorrhagic stroke prevention. In addition to the above interesting clinical findings, two novel reference protocols (SUIT-RP1 and SUIT-RP2) were developed in this study to determine the platelet mitochondrial OXPHOS and ETC capacities. To our knowledge, this human study is the first to demonstrate that moderate-intensity ET effectively enhanced platelet mitochondrial OXPHOS and ETC capacities through increasing Complex II activity in stroke patients.

Exercise Capacity
In comparison with healthy individuals of the same age, reduced exercise capacity in stroke patients, especially about one month after onset of stroke, decreased their ability to perform daily activity independently, thereby further worsening their QoL [4,5]. The ventilatory parameters obtained from the GET may convey information regarding prognosis of circulatory disorders [23]. The VO 2peak and OUES are indicators of cardiorespiratory fitness and O 2 metabolic efficiency, respectively [22,23]. The V E -VCO 2 slope has long been known to be a useful survival predictor in patients with circulatory disorders [23]. Our previous study reported that these indices of ventilatory efficiency were correlated with exercise-induced central and peripheral hemodynamic changes in HF patients. We also found that ET effectively improved ventilation-perfusion matching during exercise, and this improvement was accompanied by better QoL in HF patients [25]. In this study, VO 2peak and OUES, as well as the cardiac pumping capability (such as CO and SV) of stroke patients, improved significantly after 4 weeks of moderate-intensity ET. These post-exercise benefits associated with a decreased trend in the V E -VCO 2 slope suggest that ET effectively improves ventilatory and hemodynamic efficiencies, leading to an increase of exercise capacity in stroke patients. In addition, the exercise regimen also increased the capacities of the maximal platelet OXPHOS and ETC in stroke patients. Moreover, improved VO 2peak was positively correlated with platelet mitochondrial bioenergetics in the stroke patients. Hence, platelet mitochondrial bioenergetics may be an ideal surrogate for determination of systemic aerobic capacity in stroke patients.

Mitochondrial Functionalities in Platelets
Emerging findings indicate that neuron death in stroke patients is associated with an impaired mitochondria respiratory capacity, which may be caused by the inability to maintain energy homeostatic to compensate for the acute cerebrovascular event. Oxygen deprivation after stroke onset contribute to mitochondria dysfunction, resulting in overproduction of reactive oxidative species (ROS) [31]. Systemic inflammatory response occurs subsequently and further changes of pro-or anti-inflammatory cytokines are associated with the severity of neurological deficit [32]. Additionally, mitochondria play a pivotal role in platelet-mediated aggregation and are essential in non-ATP-mediated thrombosis [8,9]. Hence, platelet mitochondrial bioenergetics can be a marker reflecting the oxidative stress of stroke.
Since mitochondria are highly sensitive and respond dynamically to oxidative stress in their microenvironment [33], mitochondria dysfunctions are frequently observed in the early stage of a typical ischemia/reperfusion tissue injury, such as acute myocardial infarction or stroke. Overproduction of mitochondria-derived ROS has been considered to be the consequence of the interaction of the malfunctional respiratory chain with O 2 during reperfusion [34]. The present investigation demonstrated that moderate-intensity ET significantly decreased plasma MPO and IL-6 levels, inflammatory cytokines, in stroke patients. Thus, it is considered that the ET may blunt stroke-induced oxidative stress possibly incited by mitochondrial dysfunction.
The accumulation of intracellular succinate after an ischemic insult, resulting in elevated mitochondrial ROS production, has been reported [35]. Oxidative stress may further induce succinate overproduction by decreasing succinate dehydrogenase (SDH) activity [36]. Our previous study on sedentary healthy young adults found that ET under O 2 tension of 12% for 30 min significantly enhanced platelet mitochondria SDH activity and Complex II-related respiration [14]. In another clinical investigation, we further identified that ET elevated the platelet mitochondrial O 2 consumption rate against oxidative stress by increasing Complex II activity in HF patients [14]. In the present study, 4 weeks of ET enhanced platelet mitochondrial OXPHOS and ETC capacities by increasing Complex II activity, but lowered plasma MPO and IL-6 concentrations in the stroke patients. In this investigation, we further detected that an improved aerobic capacity was positively correlated with changes in FADH2 (CII)-related OXPHOS and ETC activities. Therefore, ET-induced elevation in platelet Complex II activity may rapidly eliminate succinate and further reduce ROS production from platelet mitochondria. This ET-induced physiologic adaptation may eventually depress circulatory oxidative stress and pro-inflammatory status in stroke patients.

Health-Related Quality of Life
In addition to an increase in aerobic capacity, this study demonstrated that ET significantly improved mental health in stroke patients. These findings imply that the moderate-intensity exercise regimen effectively facilitated the ability of patients to cope with the physical demands of daily activity, subsequently improving psychosocial status in stroke patients. Furthermore, an improved health-related QoL might improve survival probability in stroke patients and simultaneously reduce the financial burden in their health care system [1,2].

Limitations of This Study
Although limited stroke patients were included in the study, the results regarding platelet mitochondrial bioenergetics have high values of statistical power ranging from 0.902 to 1.000 (Table S2). Additionally, this study mainly focused on the effects of ET on platelet mitochondrial bioenergetics rather than platelet reactivity. Our previous studies investigated the effect of different aerobic exercise regimens on underlying mechanisms of platelet reactivity in healthy people and patients with cardiovascular disorders [11][12][13]. A recent study further reported that aerobic exercise training markedly suppressed the hypoxia-induced oxidative damage of platelet mitochondria and consequently, reduced platelet-mediated thrombin generation in healthy sedentary individuals [14]. However, the role of ET-mediated improvement of platelet mitochondrial function on platelet reactivity and coagulation in stroke patients requires further investigation.
Six hemorrhagic strokes, 3 in each group, suffered from hypertension in this study. It is well-known that hypertension is highly prevalent and an important risk factor for hemorrhagic stroke [37]. Increased platelet aggregation has been observed in aneurysmal subarachnoid hemorrhagic patients and was possibly related with abnormal platelet ADP secretion [38]. Moreover, elevated oxidative stress may impair mitochondria function and could be a potent trigger of consumptive coagulopathy to result in abnormal cerebral hemostasis [39]. Therefore, we recruited hemorrhagic stroke patients because both hypertension and abnormal platelet function can be expected in them.

Conclusions
In the present study, 4 weeks of ET improved systemic aerobic capacity through increased ventilatory and hemodynamic efficiencies in stroke patients. The exercise regimen also increased the capacities of OXPHOS and ETC by enhancing mitochondrial Complex II activity in platelets. These experimental findings may provide insight into the identification of adequate exercise regimen to increase physical performance and improve platelet mitochondrial bioenergetics in stroke patients.