Tart Cherry Supplement Enhances Skeletal Muscle Glutathione Peroxidase Expression and Functional Recovery after Muscle Damage

ABSTRACT Introduction Montmorency cherry concentrate (MCC) supplementation enhances functional recovery from exercise, potentially due to antioxidant and anti-inflammatory effects. However, to date, supporting empirical evidence for these mechanistic hypotheses is reliant on indirect blood biomarkers. This study is the first to investigate functional recovery from exercise alongside molecular changes within the exercised muscle after MCC supplementation. Methods Ten participants completed two maximal unilateral eccentric knee extension trials after MCC or placebo (PLA) supplementation for 7 d before and 48 h after exercise. Knee extension maximum voluntary contractions, maximal isokinetic contractions, single leg jumps, and soreness measures were assessed before, immediately, 24 h, and 48 h after exercise. Venous blood and vastus lateralis muscle samples were collected at each time point. Plasma concentrations of interleukin-6, tumor necrosis factor alpha, C-reactive protein, creatine kinase, and phenolic acids were quantified. Intramuscular mRNA expressions of superoxide dismutase 1 (SOD1), SOD3, glutathione peroxidase 1 (GPX1), GPX3, GPX4, GPX7, catalase, and nuclear factor erythroid 2–related factor 2 and relative intramuscular protein expressions of SOD1, catalase, and GPX3 were quantified. Results MCC supplementation enhanced the recovery of normalized maximum voluntary contraction 1-s average compared with PLA (postexercise PLA, 59.5% ± 18.0%, vs MCC, 76.5% ± 13.9%; 24 h PLA, 69.8% ± 15.9%, vs MCC, 80.5% ± 15.3%; supplementation effect P = 0.024). MCC supplementation increased plasma hydroxybenzoic, hippuric, and vanillic acid concentrations (supplementation effect P = 0.028, P = 0.002, P = 0.003); SOD3, GPX3, GPX4, GPX7 (supplement effect P < 0.05), and GPX1 (interaction effect P = 0.017) gene expression; and GPX3 protein expression (supplementation effect P = 0.004) versus PLA. There were no significant differences between conditions for other outcome measures. Conclusions MCC supplementation conserved isometric muscle strength and upregulated antioxidant gene and protein expression in parallel with increased phenolic acid concentrations.

I ntense exercise may induce muscle damage, resulting in muscle soreness and associated reductions in force generating capacity of the muscle. This damage occurs through a complex combination of mechanisms, including structural damage to the contractile apparatus, as well as disruption to biochemical pathways such as those governing skeletal muscle calcium handling. This is in part due to high intramuscular forces and increased reactive oxygen species (ROS) exposure, generated during exercise (1,2).
ROS are generated during exercise (1,2) and are understood to play an important role in maintaining homeostasis; when levels exceed the capabilities of the endogenous antioxidant defense mechanisms, cellular redox balance is altered resulting in oxidative stress (3). This in turn causes further disruption and damage to cellular processes and structures (1)(2)(3)(4). For example, structures within the sarcoplasmic reticulum are sensitive to ROS, such that increased exposure to ROS impairs muscle calcium handling and sensitivity (5). This leads to decrements in muscle contractile force development and consequently exercise performance (2,4). Furthermore, disruptions in skeletal muscle calcium handling are also likely to impair recovery, for example, by elevating muscle protein breakdown and reducing phosphorylation of protein kinase B (Akt) and mammalian target of rapamycin (mTOR) (6), thereby reducing protein synthesis required for repair.
Because of this involvement of oxidative damage, there has been an abundance of research investigating the use of exogenous antioxidant supplements as a means of reducing exerciseinduced muscle damage and the associated recovery time (7,8). Numerous studies have shown that polyphenol supplementation reduces blood markers of oxidative damage and inflammation (9)(10)(11)(12)(13). Montmorency cherries contain high concentrations of polyphenols (14), and there is evidence that their consumption in supplement form may attenuate oxidative stress, inflammation, and muscle soreness, aiding muscular recovery from multiple exercise modalities (9)(10)(11)13,15,16). However, to date, research in this area has relied on proxy markers of intramuscular oxidative stress and inflammation within blood plasma or serum, rather than analysis of the exercised muscle tissue itself. This approach is unlikely to comprehensively elucidate the effects of intensive exercise and supplementation strategies to support recovery because these proxy measures have been shown to respond differently in recovery from intensive exercise to direct muscle measures (17).
The mechanisms by which supplementation may enhance recovery are unclear, and the limited evidence available is equivocal. Initial theories proposed radical scavenging as the primary mechanism because of the ability of phenolic compounds to donate electrons via hydrogen atom transfer from a hydroxyl unit. However, the low concentrations of polyphenols and phenolic metabolites present in the plasma suggest they are unlikely to act as direct antioxidants in vivo (18). We and others have hypothesized that it is more likely that antioxidant effects in vivo arise from nuclear translocation and activation of nuclear factor erythroid 2-related factor 2 (Nrf2) signaling after exposure to phenolic metabolites (7,19,20).
This study is the first to investigate the potential biological mechanisms within human skeletal muscle that underpin improvements in exercise recovery in response to Montmorency cherry concentrate (MCC) supplementation. We quantify antioxidant mRNA and protein expression within exercised muscle tissue. We also quantify the phenolic metabolites of MCC in plasma throughout the MCC loading and exercise recovery phases. It was hypothesized that supplementation would amplify gene and protein expression of endogenous antioxidant enzymes thus enhancing functional recovery.

METHODS
This study used a double-blind crossover design in which participants completed two trials separated by a 2-wk washout period in line with previous literature (10,15,21,22). The study received ethical approval from the Sport and Health Sciences ethics committee at the University of Exeter and Human Research Ethics committee at the University of Queensland, and all testing conformed to the guidance set out by the Declaration of Helsinki (see additional information for details).

Participants
Ten recreationally active male participants (age = 23.4 ± 5.4 yr, weight = 78.0 ± 21.9 kg, height = 178.4 ± 6.9 m), asymptomatic of illness and injury, completed the study. Twelve participants were recruited and consented, but two withdrew before completion of the study. One participant was unable to participate following consent and familiarization because of commitment to another study, which required dietary manipulation, and one participant completed one arm of the study but was unresponsive to all subsequent contact attempts. Participants completed a physical activity readiness questionnaire and medical and exclusion screening questionnaire before providing written informed consent. The physical activity readiness questionnaire was also used to exclude sedentary individuals. Exclusion criteria comprised individuals age below 18 or above 40 yr, females, individuals allergic to fruit, and highly trained individuals. Females were excluded to avoid any confounding influence of low-grade inflammation caused by menstrual cycle symptoms (23), and highly trained individuals were excluded because of their familiarity with high-intensity exercise, which may have dampened the effects of the exercise damage protocol (24). Individuals who are highly trained will have chronic training adaptations, such as increased antioxidant and buffering capacity, as well as increased monocarboxylate transporters, which can improve the speed of postexercise recovery (25). Furthermore, trained individuals' familiarity with eccentric exercise may be protective against muscle damage from subsequent eccentric exercise during the damage protocol (26). Trained individuals were excluded by verbal questioning before giving informed consent, and their responses were subsequently confirmed by visual assessment of their activity diary. Trained individuals were defined as completing more than 3 h·wk −1 of deliberate planned exercise outside of normal physical activity as defined by Caspersen et al. (27).
Sample size was calculated as 10 participants to provide 80% power to detect a 5% difference between trials, based on the expected difference between MCC and placebo (PLA) (effect size of 1). Calculations were based on the maximum voluntary contraction (MVC) force recovery data from Bowtell et al. (10) and the anticipation of a curvilinear relationship between dosage and functional effects on performance markers. Twelve participants were initially recruited to account for dropouts.

Supplementation Period
Trials were counterbalanced for trial order and leg dominance, with participants and investigators blinded to treatment to avoid potential bias. A researcher, who had no further involvement with data collection, prepared the supplement packs for the participants. The participants were randomized to supplement code "A" or "B" using a sealed envelope system. The MCC and the PLA supplements were then provided to investigators by a member of the research team not involved in data collection, in opaque bags coded "A" or "B." During each trial, participants ingested two 30-mL daily doses (morning and evening) of either MCC or PLA, for 10 d. MCC supplement was a commercially available product (CherryActive ® , ActivEdge, Oldham, UK) produced from US Montmorency cherries. PLA was a commercially available fruit concentrate (Morello Cherry Cordial, Blossom Cottage, Gloucester, UK) with additional carbohydrate added to ensure it was isoenergetic. Analysis of the phenolic content of the MCC by high-performance liquid chromatography (HPLC) (28) was conducted by Atlas Bioscience, Inc. (Tucson, AZ). Total content was 20.167 mg·mL −1 polyphenolics and 7.211 mg·mL −1 total anthocyanins, with pelargonidin (3.319 mg·mL −1 ) and delphinidin (1.299 mg·mL −1 ), the most prevalent anthocyanins (for details on polyphenol composition of MCC supplement, see Table A, Supplemental Digital Content, Appendix, http://links.lww.com/MSS/C455). This supplementation protocol has previously been shown to enhance the recovery of MVC knee extensors after exercise-induced muscle damage (10). Participants were also asked to maintain any normal exercise habits throughout the trial, but to refrain from high-volume and high-intensity exercise, such as resistance, interval, or unaccustomed endurance type exercise, for 48 h before experimental visits. Participants were also asked to maintain their normal diet, but to avoid any increase in consumption of foods with high polyphenol concentrations during the supplementation and wash out periods, in addition to alcohol and caffeine 48 h before the test. To this end, participants were provided with diet and exercise diaries, to record all food and beverages consumed in the final 6 d of the supplementation period and exercise completed over the final 5 d of each supplementation period. Avoidance of an increase in consumption of polyphenol-rich foods, rather than avoidance in toto, was advised to maintain ecological validity. Participants were also asked to refrain from eating or drinking anything apart from water for 10 h before the laboratory visits during both supplementation periods, and all laboratory tests were conducted at the same time of day for each participant. On days where experimental testing occurred, morning doses of MCC were consumed by participant before arrival at the laboratory.

Experimental Design
Familiarization. Participants' baseline measures of height and weight were assessed, before familiarization with all experimental procedures, and measurement of MVC. Performance of the warm-up during familiarization was force limited to 100 N for all participants. Familiarization for all tests occurred for both dominant and nondominant legs on the same day. A maximum of one set of the damage protocol was performed submaximally. Chair and dynamometer arm settings for the Biodex were determined during the familiarization visit and recorded for use in all subsequent tests. Leg dominance was determined by defining the nondominant leg as the stabilizing leg during single leg movements such as kicking.
Damage protocol and functional testing. Overnight fasted participants returned to the laboratory on day 8 of supplementation at which point resting venous blood samples were collected from an antecubital vein. Serum samples were collected in tubes containing clot activator and gel for serum separation and kept at room temperature for 30 min, and plasma samples were collected in tubes containing lithium heparin. Tubes were centrifuged at 4500 rpm for 15 min at 4°C to fractionate samples and remove the cellular components. Serum and plasma were distributed into microcentrifuge tubes before storage at −80°C until analysis. Vastus lateralis muscle biopsy samples were taken using the suction-modified percutaneous Bergstrom needle technique (29). The leg from which the biopsy was taken was sterilized using iodine and anesthetized locally with 2% lidocaine. An incision of approximately 0.8 cm was then made, before a biopsy needle was used to collect a sample of muscle (~150 mg). The incision was then closed with butterfly stitches and covered with a waterproof dressing. Eight biopsies were taken from each participant, with four per experimental trial, before and after exercise, at 24 and 48 h. All biopsies were taken from the same leg during the course of each trial, each time from a new incision.
Muscle soreness was assessed with participants seated and knee extensors in a stretched position, with a knee angle of 90°via the use of a (200 mm) visual analog scale (VAS) and pain pressure threshold (PPT) using a handheld algometer (FDX 50; Wagner, Greenwich, CT) (10,16). Algometer measures were taken by application of increasing pressure, with a handheld algometer, to the participant determined point of being "uncomfortable but not painful," at the vastus lateralis, vastus medialis, and rectus femoris. One measure was taken at each site by the same investigator before, after, 24 h, and 48 h after completion of the damage protocol, as in the study of Bowtell et al. (10). For VAS analysis, participants were instructed to mark their level of soreness on a line from 0 to 10, with 0 being no pain and 10 being extreme pain. VAS values of this length have previously been shown to have good reliability for measuring acute pain and detecting changes in pain intensity (30).
The exercise protocol ( Fig. 1) consisted of a warm-up, muscle function measures (single-leg maximal isokinetic knee extension and flexion repetitions [IK Max ], MVC, and single leg jumps [SLJ]), and a muscle damage protocol, using a Biodex Isokinetic Dynamometer (Biodex System 3 Medical Systems; Shirley, NY). Functional measures and muscle damaging exercise via eccentric contractions of knee extension exercise were selected in line with previous research to allow for direct comparisons (10,31).
Before beginning experimental measures (IK Max , MVC, and SLJ), participants completed the warm-up protocol using the leg from which the preexercise biopsy had been taken. The warm-up consisted of 5 sets of 5 single-leg submaximal isokinetic knee extension and flexion repetitions with a force limit set at 50% of familiarization MVC for that leg, separated by 1 min of rest. After the warm-up protocol, participants completed 3 sets of 3 IK Max repetitions and 3 MVC, separated by 1 min of rest, and 3 SLJ performed consecutively with a rest period of at least 10 s. Jumps were performed on a mat (Jump Mat Pro; SL Electronics Ltd., Cookstown, UK) with hands on hips and a single leg takeoff to isolate performance as much as possible to the limb of interest. A two-footed landing was used to account for discrepancies in participants ability to balance upon landing, especially with fatigue after the exercise protocol. Jump height was recorded in mm.
Warm-up and IK Max repetitions were performed over a range of motion of 80°from full flexion at the knee, at 60°·s −1 for both the concentric and eccentric phases. MVC repetitions were performed at a knee angle of 90°, with the Biodex arm stationary, as this has previously been shown to be the angle at which the maximum amount of force can be produced, due to the optimal overlap of sarcomeres (32).
The damaging exercise protocol was performed 5 min after completion of SLJ and consisted of 10 sets of 30 maximal eccentric knee flexion contractions (EC Max ), with each set separated by a period of 1 min. Repetitions were performed over the same range of motion of 80°as that used for the warm-up and IK Max repetitions, with a passive (no contraction) concentric phase at 180°·s −1 and a maximal eccentric phase at 60°·s −1 . There were no significant differences in joint angles between legs ( P = 0.528). Performance tests were then repeated after the damage protocol. Throughout the experimental measures and damage protocol, participants were given verbal encouragement. A further muscle biopsy and blood sample were taken immediately after completion of the postdamage protocol performance tests, in addition to further measures of muscle soreness.
Participants returned to the laboratory after an overnight fast 24 and 48 h later. During both visits, resting venous blood samples, a further assessment of muscle soreness, and a further muscle biopsy were taken before repetition of the warm-up, functional performance measures ( Fig. 1). All biopsies were taken from the same leg during the course of a trial. After a 2-wk supplement washout period, this protocol was repeated with the functional measures and damage protocol performed using the contralateral leg to minimize any repeated bout effects. All visits for each participant were performed at the same time of day.
Force recordings. Force produced during knee extension exercise was measured using a Biodex isokinetic dynamometer. Torque was displayed in Newton-meters (N·m). Force data were recorded and analyzed using a custom written script in Spike2 version 6 software (CED, Cambridge, UK).
Work done during the damaging exercise protocol was determined via calculation of the area under the force time curve: Force data from MVC were analyzed to calculate both peak force output and the highest average value over a 1-s period, occurring within the plateau of each contraction. The reported MVC value for each respective time point was represented by the highest value achieved across the three MVC contractions for each measurement time point to ensure the maximal possible peak values was recorded.
Force data for maximal isokinetic contractions were assessed by measurement of peak force during each set for the three individual concentric and eccentric contractions, respectively, completed during each set at each measurement time point. Three sets of contractions were completed at each time point, from which the highest values of peak concentric and eccentric force were taken.

Sample Analyses
Blood sample analysis. Serum samples were analyzed for activity of interleukin-6 (IL-6), C-reactive protein (CRP), and tumor necrosis factor alpha (TNF-α) via ELISA (IL-6, HS600C; TNF-α, HSTA00E; CRP, DCRP00; R&D Systems Quantikine High Sensitivity ELISA, R&D Systems, Minneapolis, MN), according to the manufacturer's instructions to assess muscle damage and systemic inflammation. Creatine kinase (CK) analysis was performed by the Royal Devon and Exeter NHS Trust on the 702 module of the Cobas 8000 automated platform according to the manufacturers recommended protocol (Roche Diagnostics, Basel, Switzerland).

Measurement of Plasma Phenolic Metabolites Profile by High-Resolution Accurate-Mass Mass Spectrometry
Phenolic metabolite analysis were performed at the Bioanalytical Facility, University of East Anglia. Plasma concentrations of protocatechuic acid, 4-hydroxybenzoic acid, hippuric acid, vanillic acid, ferulic acid, and isoferulic acid were quantified using an Orbitrap Velos Linear Trap Quadropole high-resolution accurate-mass mass spectrometry system coupled with an Accela autosampler and ultra high-pressure liquid chromatography pump (Thermo Scientific, Cheshire, UK). The Orbitrap system was operated in Fourier transform MS mode at the resolution of 30,000 in negative electrospray ionization (ESI) mode.
To prepare the samples for analysis, 200 μL of plasma/ calibration stands/quality controls and 20 μL of internal standard containing ferulic acid-[2H3] (100 nmol·L −1 ) and hippuric acid [13C6] (200 μmol·L −1 ) (Toronto Research Chemicals, Ontario, Canada) in 0.1% formic acid (Merck, Germany) were pipetted into a microcentifuge tube and mixed. To this, 1 mL of methanol was added slowly with gentle mixing, the mixture was then incubated at room temperature for 15 min, followed by centrifugation at 14,000 rpm for 7 min. The supernatant was transferred to borosilicate glass tube and placed in an evaporator to dry under a constant stream of nitrogen at a temperature of 60°C. To the dried supernatant, 200 μL of methanol (Merck, Germany) with 0.1% formic acid was added into each tube and vortex mixed for 30 s, followed by 2.5 mL of ethylacetate (Merck, Germany) and vigorously mixed for 10 min. After centrifugation at 4000 rpm for 10 min, 2 mL of the ethylacetate in the upper layer was transferred to a fresh set of borosilicate glass tubes and again evaporated to dryness as described above. The dried residue was resuspended in 250 μL of LCMS grade deionized water with 1% acetic acid (Merck, Germany), then vortex mixed followed by centrifugation at 4000 rpm for 10 min. The final mixture was transferred into polypropylene autosampler vials, and 50 μL was injected into the liquid chromatography high-resolution mass spectrometry system for analysis.
Chromatographic separation was achieved using a ModusCore C18 reverse phase column (2.1 m Â 50 mm, 2.7 μm) (Chromatography Direct, Runcorn, UK) maintained at a temperature of 40°C. Mobile phases A consisted of 1% acetic acid in LCMSgrade deionized water with LCMS grade methanol as mobile phase B. The binary gradient program was as follows: 0 min 1% B, 0-1 min 1% B with a linear increase to 45% B at 10 min, 10-10.5 min 95% B and held to 12 min, returned to 1% B at 12.5 min to reequilibrate with a cycle time of 15 min. Mobile phase flow rate was 0.5 mL·min −1 throughout the run.
The mass scan range used to quantify each phenolic metabolite was determined by the direct infusion of pure standards into the ion source via a T-connector. European Pharmacopoeia reference standards used to prepare the calibration standards were obtained via Merck (Germany). The quantitation mass range (Da) values were as follows: protocatechuic acid, 152.99800-153.00300; 4-hydroxybenzoic acid, 137.005-137.010; hippuric acid, 178.026-178.032; vanillic acid, 167.012-167.017; ferulic acid and isoferulic acid, 193.023-193.030. Xcalibur software version 2.1 (Thermo Scientific) was used for system control, data acquisition, baseline integration, and peak quantification (for summary of assay performance, see Table B, Supplemental Digital Content, Appendix, http://links.lww. com/MSS/C455).
Subsequently, RNA concentration and purity of samples were analyzed by spectrophotometry (NanoDrop Lite Spectrophotometer; ThermoFisher Scientific, Waltham, MA), before cDNA transcription of RNA was performed with Primerdesign Precision nanoScript 2 Reverse Transcription kit, according to manufacturer's instructions (Primerdesign, Southampton, UK).
After this, 5 μL of TaqMan fast advanced master mix (Applied Biosystems, Waltham, MA) and 5 μL of each sample containing 2.5 ng cDNA were added to custom TaqMan Gene Expression Array 96 well fast plates (Applied Biosystems). Plates were then sealed and vortexed briefly to ensure contents were mixed before centrifuging for 1 min at 1200g at 4°C (Sorvall ST 16 Centrifuge Series, ThermoFisher Scientific). Plates were then loaded into the real-time quantitative polymerase chain reaction (RTqPCR) instrument for analysis (QuantStudio 6 Flex Real-Time PCR System, ThermoFisher Scientific). Samples underwent 1 cycle for enzyme activation at 95°C for 20 s and then underwent 40 cycles of sequential denaturing at 95°C for 1 min, and annealing/extending at 60°C for 20 s. The internal control used was 18 s rRNA (data available in Supplemental Digital Content, Appendix, Table C, http://links.lww.com/MSS/C455).
RTqPCR fold change was calculated using the Pfaffl formula (33) for quantification relative to the preexercise PLA condition. These values were log 10 (x) transformed before analysis to linearize data. Primer efficiency was assumed to be 2.

Protein Content Analysis
Protein extraction. The 25-mg muscle was placed in microcentrifuge tubes with 400 μL of radioimmunoprecipitation assay (RIPA) buffer (1 Pierce A32961 Protease and Phosphatase Inhibitor EDTA-free mini tablet, ThermoFisher Scientific, dissolved in 10 mL Pierce 89900 RIPA buffer, ThermoFisher Scientific), before bead homogenization (Speedmill Plus, Analytik Jena AG) for 30 s and 1 min sequentially. Homogenized muscle and RIPA buffer samples were then aspirated and placed into clean microcentrifuge tubes and vortexed thoroughly (FB15012 TopMix Vortex Mixer, ThermoFisher Scientific), before incubation on ice for 30 min, with occasional vortexing. Samples were then centrifuged for 10 min at 8000g at 4°C (Sorvall ST 16 Centrifuge Series, ThermoFisher Scientific). The supernatant was retained, and the pellet was discarded.
Western blotting. Gels were loaded with prestained protein molecular weight ladder (Pierce 26612 Prestained Protein MW Marker, ThermoFisher Scientific), a preprepared protein standard of pooled positive control sample, produced by combining samples of a dropout participant, and participant lysate samples for analysis. All gel electrophoresis was run at 120 V constant (Mini-PROTEAN Tetra Cell System Tank and PowerPac Basic Power Supply; Bio-Rad, Hercules, CA) until the dye reached the bottom of gels. Membranes were then incubated in a blocking solution of either 5% milk powder (Marvel Dried Skimmed Milk; Premier Foods, Hertfordshire, UK) or 5% BSA (BP9702-100 BSA; Fisher BioReagents, Waltham, MA), for 1 h at room temperature before overnight incubation of at least 12 h in a primary antibody dilution at 4°C (see Table D, Supplemental Digital Content-Appendix, for details on individual protocols by protein target, http://links.lww.com/MSS/C455).
Membranes were incubated in the appropriate secondary antibody dilution at room temperature for 1 h. The membrane was washed with Tris-buffered saline (S5886 Sodium Chloride, Sigma-Aldrich Company Ltd., and Tris-Base BP152-1, Fisher BioReagents) (TBS) and Tween 20 (Tween BP337-100, Fisher BioReagents) (TBS-T) solution at least three times after every cycle of antibody incubation. Protein detection was conducted on the membrane using enhanced chemiluminescence (ECL) detection reagent (RPN2232 Amersham ECL Prime Membranes were destained using TBS-T before colormetric imaging using automatically determined exposure (ChemiDoc XRS+ System, Bio-Rad).
Densitometry analysis. Blots were analyzed for optical density using ImageJ (Rasband, W.S., ImageJ, U.S. National Institutes of Health, Bethesda, MD). Band intensity was normalized to total protein load (Coomassie blue, intensity of entire lane) and normalized across gels using the positive control sample.

Statistical Analysis
All data (except total work during the muscle damage protocol for which a paired t-test was used) were analyzed by a two-way repeated-measures ANOVA. Where data were missing for one time point in a participant trial arm, z-scores were calculated for the missing data point. In cases where there were multiple data points missing from a participant trial arm, that participant was excluded for the corresponding analysis. MVC data are presented for each time point normalized to preexercise MVC to control for differences in preexercise MVC between legs and between participants. For transparency, absolute MVC data are also presented along with results of those statistical analyses. Fold change from RTqPCR analysis was linearized by log 10 (x) transformation before analysis. In cases where a significant interaction effect was detected, post hoc pairwise tests were conducted with Bonferroni corrections. Throughout analyses, values that did not meet the assumption of sphericity as measured by Mauchly's test were Greenhouse-Geisser corrected. Where data were not normally distributed and could not be normalized by standard approaches to data transformation, a nonparametric Friedman's test was conducted for comparison with the results of the two-way repeated-measures ANOVA. All statistical analyses were performed using IBM SPSS Statistics (Version 26). For ease of reading, main effects (supplement, time, and interaction) and post hoc differences are only reported in text, tables, and figures where statistical significance was achieved.

RESULTS
Knee extension MVC 1-s average decreased to 59.5% ± 18.0% and 76.5% ± 13.9% of preexercise for PLA and MCC conditions, respectively, after completion of the intensive exercise protocol ( Fig. 2A). There was no significant difference between conditions in work performed throughout the damage protocol (PLA, 44,722.4 ± 14,535.7 J, vs MCC, 46,812.0 ± 12,341.6 J). A visual inspection was conducted on food and activity diaries to ensure participants replicated dietary intake and activity. There was no effect of trial order on the MVC measures (MVC 1-s average, P = 0.467; MVC peak, P = 0.394).
MCC supplementation significantly enhanced the force recovery of normalized MVC 1-s average (supplementation effect, P = 0.024; interaction effect, P = 0.043) (Fig. 2A). Post hoc testing revealed a significantly higher force recovery for MCC immediately postexercise ( P = 0.033), but no significant differences at any other time point.
The recovery of normalized peak MVC force was enhanced in the MCC versus PLA condition (supplementation effect, P = 0.032; interaction effect, P = 0.049). Post hoc testing revealed no significant differences in normalized peak MVC between conditions preexercise and immediately, 24-, and 48-h postexercise, although postexercise was close to significance ( P = 0.054) (Fig. 2B). Nonnormalized peak and 1-s average MVC force were not significantly higher in MCC than PLA condition; however, there were significant interaction effects for both 1-s average and peak MVC force (Table 1).
There was no significant difference in the recovery of normalized IK Max (combined eccentric and concentric phases), or during separated concentric (IKCon Max ) (Fig. 2C), or eccentric contraction phases (IKEcc Max ) (Fig. 2D).
There was no significant difference between supplement conditions in jump height (Fig. 2E) or soreness measures (VAS and PPT VL , PPT VM , PPT RF , and PPT SUM ) between conditions (Table 1). Soreness, as measured by VAS and PPT RF , was significantly elevated after the damaging exercise (time effect, P < 0.05).
There was a significant increase in protein expression of GPX3 after MCC supplementation (3.0-fold increase; supplementation effect, P = 0.004) (Fig. 5C). There was no significant difference in SOD1 and CAT protein expression between MCC and PLA (Fig. 5A, B). GPX4 and GPX7 proteins were undetectable and are not presented.
There were significant increases in serum concentrations of IL-6, TNF-⍺, and CK (time effect, P < 0.05), but not CRP ( P = 0.130), after the damage protocol (Fig. 6). No significant differences were found between supplementation conditions for serum concentrations of IL-6, TNF-⍺, CRP, and CK (Fig. 6). IL-6 data were not normally distributed and could not be normalized by standard approaches to data transformation. However, a nonparametric Friedman's test produced results that were consistent with the results of our two-way RM ANOVA analysis of these data (time effect, P < 0.001; supplement effect P = 0.145).

DISCUSSION
This study presents the first evidence demonstrating a significant upregulation of antioxidant gene and protein expression in human skeletal muscle after 7 d preload and 3 d postload supplementation with MCC containing a complex blend of polyphenols. Crucially, these effects occurred in parallel with significantly improved functional recovery after intensive exercise, as observed previously (9,10,13,15,16). Furthermore, enhancements in antioxidant expression profile and functional recovery were accompanied by a corresponding augmentation in plasma concentrations of phenolic acids. These novel findings shed new light on the mechanisms that underpin functional changes after natural polyphenol blend supplementation such as MCC.
Plasma concentrations of 4-hydroxybenzoic acid, hippuric acid, and vanillic acid were significantly elevated after MCC supplementation, which demonstrates that supplementation increased circulating exogenous antioxidant concentrations. However, direct ROS scavenging by these compounds is not thought to be the primary mechanism of reduced oxidative damage observed after polyphenol supplementation (18). Indeed, the greatest mean value of plasma phenolic acids measured in the current study did not surpass 40 μmol·L −1 (vanillic acid), which is 4-to 10-fold lower than the values observed for endogenous extracellular antioxidants such as plasma urate, for which concentrations range between 150 and 450 μmol·L −1 (7). Although not sufficient to elicit direct antioxidant effects, the increase in plasma concentrations of phenolic compounds appears to have been sufficient to elicit a significant upregulation of endogenous antioxidant gene and protein expression in skeletal muscle in the MCC condition. The mechanism hypothesized to underpin these changes is an upregulation of endogenous antioxidant production via the Nrf2 antioxidant response element pathway after exposure to the aforementioned elevation in phenolic metabolites (19,20). Nrf2 is widely accepted as the "master regulator" of antioxidant defense, and upregulation induces an expression profile protective against Under normal homeostatic conditions, Nrf2 is repressed through binding to Keap1 within the cytoplasm, where it is ubiquitinated and subsequently proteolysed (35). The proposed activation of Nrf2 after supplementation of MCC is hypothesized to occur via exposure of Keap1 to phenolic metabolites. Indeed, plasma levels of phenolic acids were augmented after MCC supplementation in the present study, supporting previous research which has observed bioavailability of phenolic compounds after consumption of other polyphenol-rich supplements (36,37). These phenolic compounds are then hypothesized to undergo conversion to quinones, semiquinones, and superoxide radicals via dismutation of phenoxyl radicals and redox complexes produced during radical scavenging (38,39). The literature suggests that cellular exposure to these compounds then causes oxidative modification of Keap1 cystine residues via alkylation (40). Consequently, Nrf2 dissociates from Keap1, enabling nucleic accumulation of Nrf2 and the observed upregulation of endogenous antioxidant gene and protein expression (34,38,40).
Previous in vitro and rodent model research has shown evidence that polyphenol exposure can induce Nrf2 gene expression and translocation, as well as augment activity of antioxidant enzymes, including GPX and SOD (19,41); however, to the authors' knowledge, this study presents the first in vivo evidence of upregulation of antioxidant enzyme gene and protein expression in human skeletal muscle after polyphenol supplementation. These novel data provide strong evidence that increased expression of endogenous antioxidant and cytoprotective genes after exposure to phenolic acids confers protection against oxidative stressors, such as intensive exercise and the resulting inflammatory response, thereby contributing to observed improvements in functional recovery (20,34) (Fig. 7).
This study is the first to demonstrate an increase in both GPX3 mRNA and protein expression after MCC supplementation. One of the primary postulated mechanisms for reductions in force generating capacity after intensive exercise is altered myofibrillar Ca 2+ sensitivity (1,42). Oxidized troponin I cysteine residues have been shown in vitro to bind with glutathione (43). This glutathionylation has a protective effect on troponin I molecules against oxidative stressors and increases the Ca2+ sensitivity of the contractile apparatus (41). These mechanisms have in turn been suggested to beneficially affect exercise performance (43). GPX enzymes catalyze the reduction of H 2 O 2 and organic hydroperoxides by glutathione, and thus their induction (MCC-induced induction of GPX3 in this instance) may contribute to the MCC-induced attenuation of FIGURE 7-Hypothesized mechanism underlying increased expression of GPX after MCC supplementation. Mechanisms demonstrated by the current study are demarcated from hypothesized mechanisms. Keap1, kelch-like ECH-associated protein 1; sMAF, small musculoaponeurotic fibrosarcoma. Nrf2, created with BioRender.com. the MVC force reduction identified immediately postexercise ( Fig. 2A); this requires considerable further study.
We describe an ergogenic effect of MCC on the recovery of maximal isometric force production in accordance with previous research (9,10,13,15,16), demonstrating that the aforementioned biological changes occur in parallel with significant functional effects. Importantly, the intensive bout of eccentric exercise induced significant muscle damage as indicated by the significant impairment of all functional measures, measures of perceived soreness, PPT at the rectus femoris, and all blood inflammatory markers. Therefore, an experimental paradigm was created in which the favorable effects of MCC supplementation were detectable. Although there were no significant differences between supplementation groups for soreness, some of these measures may have been affected by the use of lidocaine local anesthetic, which, while required for the vastus lateralis biopsies, may have influenced participants' ability to detect pressure pain. The measure most likely to have been affected in the current study was the PPT measure at the vastus lateralis, as this location was at the closest proximity to the biopsy site. Indeed, there were no significant effects detected for supplement or time point at the vastus lateralis or vastus medialis. However, there was a significant time effect at the rectus femoris, suggesting that the damage protocol did elicit significant muscle soreness, but pain sensitivity was reduced in the areas closest to the biopsy site. It must also be noted that a similar phenomenon has been observed previously, where functional enhancements in recovery are not necessarily reflected in soreness or PPT measures (10,11,15).
The lack of difference between conditions in measures of circulating inflammatory cytokines may be due to the methodological limitations associated with the ability of proxy measures from blood to detect subtle changes occurring at the intramuscular level. Indeed, previous research of MCC supplementation that has found that significant functional effects have demonstrated equivocal results for blood measures of inflammatory cytokines (9)(10)(11)13). This further highlights the importance of conducting analysis on the exercised tissue, a major limitation of previous research in the area. Unfortunately, because of limited muscle tissue availability, we were unable to characterize inflammatory processes within muscle in the present study. A further limitation in the current study is the lack of quantified dietary intake, without which we cannot be certain of the extent to which background diet, including polyphenol intake, was replicated between trial arms. This may have influenced the observed functional and molecular responses to MCC supplementation. However, visual inspection of dietary logs suggested that intake was faithfully replicated between study arms. Notably, such logs, whether quantified or not, are prone to participant reporting errors (44).

CONCLUSIONS
In conclusion, this study showed for the first time that supplementation with US MCC, a polyphenol-rich fruit concentrate, significantly increased expression of antioxidant genes and proteins in human skeletal muscle, in parallel with a significant increase in plasma concentrations of phenolic acids. This study also confirmed previous findings that MCC supplementation improved functional muscle recovery from exerciseinduced muscle damage. This study provides new and compelling evidence to support an upregulation of the antioxidant response element pathway, perhaps due to the increased nuclear translocation of Nrf2 after exposure to elevated phenolic metabolites, as the primary mechanism underpinning enhanced functional recovery after polyphenol supplementation.
This work was supported by the Cherry Research Committee (grant awarded November 2018). The Cherry Research Committee was not involved in study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the article for publication. J. T. W. was supported by a Ph. D. scholarship via the QUEX Institute (University of Queensland and University of Exeter).
Drafting the work or revising it critically for important intellectual content (J. T. W., M. F. O., V. G. K., J. C. Y. T., J. L. B.). All authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All persons designated as authors qualify for authorship and are listed.
The individual responsible for research governance at the University of Exeter Sport and Health Sciences Ethics Committee was Dr. Melvyn Hillsdon.
The individual responsible for research governance at the University of Queensland's Human Research Ethics Committees A and B was Chris Rose'Meyer.
Institutional Ethics Committee Approval Numbers were as follows: University of Exeter, 180314/B/05; University of Queensland, 2018000928.
Data are available within the figures and tables of this manuscript and are also included in the statistical summary document.