Longitudinal hypertrophic and transcriptional responses to high‐load eccentric‐concentric vs concentric training in males

High‐load eccentric training reputedly produces greater muscle hypertrophy than concentric training, possibly due to greater loading and/or inflammation. We quantified the temporal impact of combined maximal concentric‐eccentric training vs maximal concentric training on muscle cross‐sectional area (CSA), volume, and targeted mRNA expression (93 transcripts). Eight recreationally active males (24 ± 5 years, BMI 23.5 ± 2.5 kg/m2) performed 3 x 30 maximal eccentric isokinetic knee extensions and 2 x 30 maximal concentric knee extensions in dominant limb (ECC + CON) and 5 x 30 maximal concentric contractions (CON) in the non‐dominant limb for 12 weeks (all 90°/s, 3x/wk). Quadriceps muscle CSA and volume were measured at baseline, 28 days (d), and 84 d in both limbs (3T MRI). Resting vastus lateralis biopsies were obtained from both limbs at baseline, 24 hours (h), 7, 28, and 84 d for mRNA abundance measurements (RT‐PCR microfluidic cards). Work output was greater throughout training in ECC + CON vs CON (20.8 ± 9.7%, P < .001). Muscle CSA increased from baseline in both limbs at 28 d (CON 4.3 ± 2.6%, ECC + CON 4.0 ± 1.9%, both P < .001) and 84d (CON 3.9 ± 2.3%, ECC + CON 4.0 ± 3.1%, both P < .001), and muscle volume and isometric strength at 84 d (CON 44.8 ± 40.0%, P < .001; ECC + CON 36.9 ± 40.0%, P < .01), but no between‐limb differences existed in any parameter. Ingenuity Pathway Analysis identified several cellular functions associated with regulation of muscle mass and metabolism as altered by both modalities at 24 h and 7 d, but particularly with ECC + CON. However, mRNA responses waned thereafter, regardless of modality. Initial muscle mRNA responses to training did not reflect chronic training‐induced hypertrophy. Moreover, ECC + CON did not produce greater hypertrophy than CON, despite greater loading throughout and a differential mRNA response during the initial training week.


| INTRODUCTION
High-load eccentric exercise training has been reported to increase muscle hypertrophy, type IIx muscle fiber cross-sectional area (CSA) and strength gains over 10 to 12 weeks when compared to concentric exercise training, 1,2 which has been at least partially attributed to the greater work done during eccentric contractions. 3 However, these reports are equivocal, reflected by reports of no differences in muscle hypertrophy or strength gains when comparing high-load concentric and eccentric exercise training modalities. 4 High-load eccentric exercise is also known to cause shortterm muscle force loss and delayed onset muscle soreness (DOMS) compared to high-load concentric training, 5 both of which diminish with repeated bouts of exercise. 6 Muscle inflammation is commonly reported to accompany high-load eccentric exercise, 5 though is not commonly measured. This has resulted in combined high-load concentric and eccentric training regimens being adopted, 7,8 aimed at improving tolerability and volunteer compliance over high-load eccentric exercise training alone, and also because it is normally difficult to isolate eccentric contractions in a training setting. As far as we are aware no study has examined the temporal response to these two different training modalities. Furthermore, studies comparing muscle adaptations to concentric versus eccentric training have used different training protocols, varying in training loads and number of repetitions, methods of assessment of architectural and functional gains, and have involved both trained and untrained volunteers. 1,3 It is therefore difficult to draw conclusions on the comparative efficacy of the two modalities at inducing muscle training related adaptations, which we aimed to address in this study by adopting a within-subject bilateral exercise study design controlled for contraction number and speed of contractions.
The mechanisms that underpin purported greater muscle adaptation to eccentric exercise training are not well understood. Proposals such as greater neural activation and differing recruitment strategies, greater work done, and increased post-exercise muscle protein synthesis (MPS) have not been definitive and knowledge gaps exist. 1,3,9 Focus has also been directed toward high-load eccentric exercise-induced muscle inflammation being a mechanistic driver of hypertrophy. Application of gene transcript profiling in a study showed that 300 concentric and 300 eccentric contractions (in the form of rising from and controlled lowering to a seated position), induced greater muscle inflammatory mRNA responses after 4-8 h compared with concentric contractions alone. 10 Furthermore, these inflammatory-related mRNA responses were maintained 6h following a second bout of eccentric exercise. 11 It has also been suggested that there is a potentiating effect of inflammation on muscle satellite cell proliferation. 12 In keeping with this, an increase in the serum concentration of the pro-inflammatory cytokine interleukin 6 (IL-6) was accompanied by increased expression of myogenic regulatory factors (MRFs) between 24 to 96 h following maximal electrically evoked eccentric contractions in a rodent model. 13 IL-6 has also been demonstrated to stimulate satellite cell-mediated hypertrophy in isolated cells, 14 and local non-steroidal anti-inflammatory drug (NSAID) infusion following acute maximal eccentric exercise in healthy volunteers has been shown to decrease muscle satellite cell proliferation. 12 McKay et al 15 also showed an increase in satellite cell number 24 h following maximal eccentric exercise, which was associated with an increase in muscle IL-6 mRNA expression and serum IL-6 concentration. However, these studies examined acute responses to acute exercise intervention. In a chronic setting, high-load concentric-eccentric elbow flexor exercise for 6 weeks (5x/wk) combined with daily NSAID ingestion had no differential effect on muscle hypertrophy and strength compared to placebo group. 16 Conversely, 6 weeks of NSAID administration alongside progressive resistance training in older volunteers increased muscle volume and strength above placebo. 17 Importantly, as far as we are aware, no time-course data are available regarding muscle molecular adaptations to combined high-load concentric and eccentric exercise training compared to concentric exercise alone.
We hypothesized that in healthy, young male volunteers who were recreationally active, 12 weeks of unilateral combined maximal isokinetic concentric and eccentric resistance training would produce greater gains in muscle CSA, muscle volume, and isometric strength compared to maximal isokinetic concentric exercise training performed by the contralateral limb. We also hypothesized that these greater architectural and functional gains in the limb performing combined concentric and eccentric training would be preceded by greater exercise-induced changes in the expression of genes involved in inflammation, myogenesis and other cellular functions thought to regulate muscle mass and fuel metabolism.

| Participants
Eight healthy, non-smoking, non-vegetarian males (24 ± 5 years, BMI 23.5 ± 2.5 kg/m 2 ) were recruited to the study. Participants had to have participated in sports such as squash, football, and basketball on at least 2-3 days per week for a minimum of 2 years so were not unaccustomed to regular exercise. The study was approved by the University of Nottingham Medical School Ethics Committee (Ethics Reference No: I 07 2011) in accordance with the Declaration of Helsinki. Prior to taking part, all participants provided informed, written consent, underwent a routine medical screening and completed a general health questionnaire. Participants were asked to refrain from taking anabolic supplements, that is, protein and creatine, for the duration of the study, but were allowed to continue with their regular training regimens apart from lower limb resistance training. Based upon documented changes in muscle CSA at 10 weeks of a maximal isokinetic resistance training protocol in healthy, recreationally active male volunteers 18 and assuming (a) an effect size of 0.7, (b) a significance level of 0.05 (α), and (c) 95% power, the required sample size was calculated (using G*Power, v3.1.2) to be n = 6. We recruited n = 8 volunteers in total for the current study.

| Exercise training protocol
One of the main objectives of the study was to determine whether a greater degree of muscle inflammation caused by high-load eccentric exercise training was associated with greater muscle mass and strength gains. We therefore wished to model a training protocol of sufficient intensity to induce inflammation in the quadriceps, but was also tolerable for the duration of the study. An exercise protocol consisting of a bout of 5 sets of 30 maximal isokinetic concentric knee extensions has been shown by our laboratory to be well-tolerated and effective at increasing muscle mass and strength. 19 However, preliminary protocol development work for this study revealed that five sets of 30 maximal eccentric isokinetic contractions of the quadriceps caused considerable post-exercise discomfort and muscle force loss in volunteers, and to an extent that this would likely deter volunteers from complying with a 12-week training program. Hence, we formulated a combined concentric-eccentric protocol that was likely to induce muscle inflammation while remaining tolerable for the duration of the study. Volunteers trained 3 times per week for 12 weeks and were not in an overnight fasted state for their training sessions ( Figure 1). During each session, participants performed 3 bouts of 30 maximal isokinetic eccentric contractions of the knee extensors with their dominant leg, interspersed with two bouts of 30 maximal isokinetic concentric extensions (ECC + CON) with the same leg. Additionally, volunteers performed 5 bouts of 30 maximal isokinetic concentric contractions of the knee extensors with the non-dominant leg (CON).
Contractions in both legs were performed on an isokinetic dynamometer (HUMAC Norm, CSMi Solutions). For this, the participant was seated securely in the chair of the dynamometer with the knee flexed at 90°, the lever arm attached to the subject's ankle and the actuator axis aligned to the lateral femoral epicondyle. Contractions were performed at an angular velocity of 90°/s with 3 minutes rest between bouts, which ensures extensive quadriceps muscle fiber recruitment. 20 Volunteers were required to exert maximal voluntary force during each isokinetic knee extension (CON) or flexion (ECC) maneuver, and the ergometer lever arm was returned passively to the leg start position between contractions, thereby ensuring minimal muscle loading. Volunteers received real-time visual feedback on their performance, and verbal encouragement throughout each training session to expedite maximum effort being achieved. All volunteers completed 150 contractions per leg per session and 5400 contractions per leg over the 12-week training period. As expected, during each bout of 30 maximal contractions peak torque per contraction decreased substantially in each CON bout with only a partial recovery between bouts, but peak torque per contraction remained approximately the same within and over each bout of ECC exercise.
A within-subject bilateral internally controlled design was used as (a) such a model minimizes the variability of responses to training 21 and (b) it is known that hypertrophic adaptations  22 A potential concern with a within-subject design is the possibility of a cross-education effect. However, such a design has been used in several previous resistance exercise training studies. 9,23,24 Further, since the main targets and outputs of the present study were related to (a) hypertrophy and (b) localized muscular adaptations in mRNA expression measured in muscle samples, we foresaw cross-education having little to no impact.

| Total work output during exercise training
Total work output (both positive and negative) per week of training by each limb was calculated as the work done during the 450 maximal contractions performed each week. Total work output for each limb over the 12-week exercise training protocol was calculated as the sum of work done from all exercise sessions by each limb.

| Muscle strength
Isometric strength was determined (best of three attempts) in the knee extensor muscles, with the knee flexed at 90°, during a static maximal voluntary contraction using an isometric knee extension unit (Medical Physics, University of Nottingham, UK) at baseline (BL), and 24 h, 28 days (28 d) and 84 days (84 d; Figure 1). All measurements were made at the same time of day, at least 72 h after the last training session, and the legs were tested in a random order.

| Quadriceps muscle cross-sectional area and volume
Quadriceps muscle cross-sectional area (CSA) and volume measurements were performed at the BL, 28 d, and 84 d time points (Figure 1). For this, axial plane scans of the thigh were taken using a 3 Tesla MRI scanner (GE 3T 750 Discovery). A T1-Fast Spin Echo protocol was used (repetition time 600 ms, echo time 15 ms, number of excitation 2, field of view 22 mm, slice thickness 10 mm, no gap between slices). Participants were asked to lie supine on the MRI bed. Axial plane scans along the entire length of the quadriceps muscle were obtained to determine anatomical cross-sectional area (ACSA). The ACSA shown represents the changes in the peak value (ACSA peak ) at the middle of the femur length. The number of axial scans (48) taken in each leg of every volunteer was the same for the BL, 28 d, and 84 d time points. Using these axial scans, the contours of the whole quadriceps muscle group of each MRI scan were digitized using Osirix image analysis software (Pixmeo, Geneva, Switzerland). Subsequently, quadriceps muscle volume was calculated as follows 4 :

| Muscle sample collection
Vastus Lateralis muscle biopsies were obtained in the resting state at BL, 24 h, and a minimum of 72 h from the previous exercise session at 7, 28, and 84 d ( Figure 1). Volunteers attended the laboratory in a fasted state, having abstained from alcohol and strenuous exercise for 48 h prior. Muscle samples were obtained from both legs using the Bergstrom percutaneous needle biopsy technique, 25 and samples were immediately snap-frozen and stored in liquid nitrogen for further analysis.

| Muscle mRNA expression
We have previously demonstrated that serial needle muscle biopsy sampling, where repeated biopsies are taken 2.5 cm distal to each other, does not elicit confounding changes in the muscle transcriptome when analyzed by microarray technologies. 26 RNA was extracted from ~30 mg snap-frozen muscle from all but four biopsy samples (ie, 76 biopsies in total due to tissue availability) as previously described. 27 First strand cDNA was synthesized from 1 µg of total RNA, using Superscript III reverse transcriptase (Invitrogen Ltd) and random primers (Promega) and stored at −80°C until analysis. TaqMan low-density arrays were performed using an ABI PRISM 7900HT sequence detection system, and data analyzed using SDS 2.1 software (Applied Biosystems). Data were further analyzed using RQ Manager software (Applied Biosystems), where the threshold level was normalized across all plates before Ct values were calculated for each gene target and sample. Relative quantification of mRNAs of interest was measured using the 2 −ΔΔCt method with hydroxymethylbilane synthase (HMBS) as the endogenous control, with the mean of the baseline sample used as the calibrator. Mean Ct values for HMBS were no different between CON and ECC + CON (data not shown). A total of 93 transcripts were targeted for analysis in the present study (Appendix A). As limited muscle gene array data exist pertaining to chronic eccentric training, muscle transcripts identified as being altered following acute eccentric exercise in array studies involving healthy male volunteers were selected for inclusion in the present study, 1,10 along with muscle transcripts seen to be differentially expressed in healthy male volunteers over the course of 10 weeks of concentric resistance training. 26 Based on these studies, approximately 50 genes were selected that were linked to cell growth and regulation, DNA damage, stress responses, energy metabolism, inflammation, extracellular matrix, and muscle differentiation and signaling. Additionally, mRNA expression of myogenic transcription/regulatory factors such as myoblast determination protein 1 (MyoD), myogenin (MYOG), myostatin (MSTN), myogenic factor 5 (MYF5) were selected for inclusion. Further genes of interest known to be differentially regulated in muscle inflammatory states and linked to muscle mass regulation were also included, 28,29 namely AKT serine/threonine kinase 1 (AKT1), forkhead box protein O 1 (FOXO1), protein tyrosine kinase 2 (PTK2), pyruvate dehydrogenase kinase 4 (PDK4), DNA (cytosine-5)-methyltransferase (Met1A), cathepsin-L, muscle-specific ligases atrophy F-box protein (MAFbx), muscle RING-finger protein-1 (MuRF1) and proteasomal alpha subunit 1 (PSMA1).

| Pathway analysis
To associate a biological function to the identified probe sets, Ct values were uploaded to Ingenuity Pathway Analysis (IPA) software (Redwood City, CA, USA) for pathway analysis of gene expression data. Ingenuity Pathway Analysis is a commercially available online software application that allows for the integration, analysis, and comprehension of data from gene expression, miRNA, and single nucleotide polymorphism (SNP) microarrays, as well as proteomics, metabolomics, and RNA sequencing (RNAseq) experiments. The outcome of IPA analysis, especially for the upstream regulator and cellular function analysis could appear biased because the input was the result of an array analysis of 93 pre-selected genes, as opposed to an unbiased analysis of gene expression on a global level. For example, since several target inflammatory genes were selected for measurement, inflammatory pathways were more likely to be flagged in the IPA analysis. However, the overall outcome of each IPA analysis (eg, upstream regular analysis, cellular function, activation status) was predicted by calculating a regulation Z-score and an overlap p-value, which were based on the number of regulated target genes' function, size change, and direction of expression, and their degree of agreement with the IPA curated database of published literature. Furthermore, irrespective of the genes selected the same battery of targeted genes were compared longitudinally and between CON and CON + ECC trained limbs.

| Statistics
All values are presented as mean ± SD unless otherwise stated, with significance accepted at the P < .05 level. Normality was tested using a Shapiro-Wilks test. Where data were not normally distributed, values were log transformed and parametric statistical tests used. Comparison of mean values between ECC + CON and CON legs was performed using a repeated measures two-way analysis of variance (ANOVA) (time and exercise regimen as independent factors) with Tukey's or Sidak's post hoc test analysis where appropriate (GraphPad Prism 7.03).
To identify the statistical difference in total work output between ECC + CON and CON legs over the entire exercise protocol, a paired sample, two-tailed t test was used (GraphPad Prism 7.03).
For muscle mRNA cellular function analysis, the Benjamini-Hochberg method for multiple hypothesis correction was employed to account for the possibility of false-positive results (type I errors). This is a default function within the IPA platform.

| Baseline comparisons
The mean age of the participants was 24 ± 5 years, and mean BMI was 23.5 ± 2.5 kg/m 2 . There was no difference in isometric strength, muscle CSA or muscle volume when comparing ECC + CON and CON legs at baseline (Table 1).

| Muscle function during training
Total work output per week during training was greater in ECC + CON compared to CON (Figure 2A), such that there was a treatment effect (P < .001), and cumulative work done over the 12-week protocol was greater in ECC + CON than CON (20.8 ± 9.7%, P < .001; Figure 2B).
Isometric strength of the knee extensors increased over time (P < .05), such that it was increased above baseline at 84d (44.8 ± 40.0% in CON, P < .001; 36.9 ± 40.0%. in ECC + CON, P < .01). However, there was no difference between groups in the change in isometric strength from baseline at any time point.
T A B L E 1 Isometric strength and mid-thigh quadriceps muscle cross-sectional area (CSA) and quadriceps muscle volume at baseline are shown for a leg trained using a concentric exercise protocol (CON) and the contralateral leg trained using a combined concentric and eccentric exercise protocol (ECC + CON)

| Muscle cross-sectional area and volume
Quadriceps muscle CSA was increased above baseline in both legs by 28d (CON P < .001, ECC + CON P < .001) and 84d (CON P < .001, ECC + CON P < .001; Figure 3A). There was no increase in muscle CSA in either the CON leg or the ECC + CON leg from 28 d to 84 d. No difference was seen in change in muscle CSA with exercise training at any time point when comparing CON and ECC + CON legs. Thigh muscle volume was also increased above baseline in both legs at 28 d (CON P < .001, ECC + CON P < .001) and 84 d (CON P < .001, ECC + CON P < .001; Figure 3B). Similar to muscle CSA, there was no increase in muscle volume in either the CON leg or the ECC + CON leg from 28 to 84 d. No difference was seen in change in muscle volume with exercise training between CON and ECC + CON legs at any time point.

| Muscle mRNA expression
Heat maps were generated (Figure 4; GraphPad Prism version 7.03) to depict muscle mRNA expression relative to baseline (log 2 ) in both CON and ECC + CON legs at each time point for each volunteer. Figure 4 illustrates the change in mRNA abundance from mean baseline abundance in both CON and ECC + CON legs at 24 h ( Figure 4A), 7d ( Figure 4B), 28d ( Figure 4C), and 84 d ( Figure 4D). A temporal response to resistance training was evident, such that generally there was an apparent increase in mRNA abundance above baseline in both legs at 24 h ( Figure 4A) and 7 d ( Figure 4B), but this response had waned by 28 d ( Figure 4C) and 84 d ( Figure 4D). There was a noticeable pattern of divergence of response between legs, such that there was a greater increase in abundance of muscle mRNA transcripts in the ECC + CON leg compared to the CON leg at 24 h ( Figure 4A) and 7 d ( Figure 4B). This divergence in response between legs had also waned by 28 d ( Figure 5A) and 84 d ( Figure 5B).

| Pathway analysis
Cellular functions most altered from baseline in the CON leg and ECC + CON leg at 24 h, 7, 28, and 84 d are depicted in Figure 5. The x-axis displays cellular functions most affected, while the y-axis displays the -log of the P-values. The P-value associated with each cellular function is a measurement of the likelihood that the association between a set of focus transcripts and a given function is due to random chance. The -log of P-value was calculated using Fisher's exact test (right-tailed). Overall, IPA detected altered cellular functions [-log (P-value) >5.0; Figure 5] associated with organismal injuries and abnormalities, cellular development, skeletal and muscular development and function, tissue development, carbohydrate metabolism, skeletal and muscular disorders, and lipid metabolism. Figure 5 also illustrates that in the main the largest magnitude of response to exercise training occurred at 24 h and 7 d in both limbs, which was invariably greatest in the ECC + CON trained leg and particularly at 7 d. Red indicates a greater abundance and green indicates less abundance relative to baseline F I G U R E 5 Cellular functions identified by IPA as being altered from baseline (n = 8) in vastus lateralis muscle in concentric (CON) and combined concentric + eccentric (ECC + CON) trained legs at 24 h, 7 d, 28 d, and 84 d based on mRNA expression data generated using the lowdensity microarray cards. The x-axis displays cellular functions most affected by training, whereas the y-axis displays the -log of the P-value. The -log of the P-value was calculated by Fisher's exact test right-tailed (P < .05). The threshold line corresponds to a P-value of .05 Figures 6 and 7 are included by way of example to illuminate precisely how any cellular function identified as being altered in Figure 5, in this case skeletal and muscular system development and function, was influenced by training, and to also highlight the clear differences in response between limbs at 24 h and 7 d. Each figure depicts the most differentially regulated mRNAs from baseline (outer ring), and the cellular events predicted to result from these collective mRNA changes in CON and ECC + CON legs at 24 h and 7 d. Similar schematics could be generated for all of the cellular functions highlighted in Figure 5, but in the interest of space this is impracticable. In the case of skeletal and muscular system development and function at 24 h (Figure 6), 16 mRNAs were identified as being altered in abundance relative to baseline in the CON leg ( Figure 6A). The most upregulated mRNAs from baseline were tenascin C (TNC) and IL-6, while myostatin (MSTN), paired box gene 3 (PAX3) and peroxisome proliferator-activated receptor gamma coactivator 1 α (PPARGC1A) were the most downregulated transcripts. Based on the collective changes in mRNA abundance from baseline, IPA predicted among a number of things activation of migration of muscle cells, and inhibition of muscle formation and quantity of muscle with highest confidence at 24 h. In the ECC + CON leg at 24 h, 18 mRNAs were identified as being altered in abundance relative to baseline ( Figure 6B). In contrast to the CON leg however, all mRNAs were increased in abundance. In common with the CON leg, IPA predicted activation of migration of muscle cells with high confidence. Proliferation of myoblasts and muscle cells, and formation of muscle were also predicted to be activated with high confidence. This latter prediction was in direct contrast to the CON leg, where formation of muscle was predicted to be inhibited with high confidence. In addition to this, inhibition of muscle function was predicted in the ECC + CON leg at 24 h, which was not predicted for the CON leg. Figure 7 relates to skeletal and muscular system development and function at 7 d. Fourteen mRNAs were differentially regulated from baseline in the CON leg ( Figure 7A), which were in the main the same transcripts that were changed in abundance at 24 h. Not surprisingly therefore, similar to 24 h, IPA predicted activation of migration of muscle cells, and inhibition of muscle formation with highest confidence at 7 d. Twenty-four mRNAs were found to be differentially expressed from baseline at 7 d in the ECC + CON leg, and similar to 24 h all were increased in expression ( Figure 7B). Also, in keeping with 24 h, IPA predicted activation of migration of muscle cells, proliferation of myoblasts and muscle cells, formation of muscle, and inhibition of muscle function F I G U R E 6 Differentially regulated muscle mRNAs (outer ring) associated with skeletal and muscular system development and function in (A) concentric (CON) and (B) combined eccentric + concentric (ECC + CON) trained legs at 24 h of training compared to baseline, and the cellular events predicted by Ingenuity Pathway Analysis to results from the collective changes in mRNA abundance. IL-6, interleukin 6; MSTN, myostatin; MYC, MYC proto-oncogene; MYF6, myogenic factor 6; MYH7, myosin heavy chain 7; PAX3, paired box 3; PAX7, paired box 7; PDK4, pyruvate dehydrogenase kinase 4; PTK2, protein tyrosine kinase 2; SIRT1, sirtuin 1; PPARGC1A, peroxisome proliferator-activated receptor α; VEGFA, vascular endothelial growth factor A; ATF3, activating transcription factor 3; CYR61, cysteine-rich angiogenic inducer 61; TNC, tenascin C; ANKRD2, ankyrin repeat domain 2; AKT1, AKT serine/threonine kinase 1; PRKAA1, protein kinase AMP-activated catalytic subunit α 1; IGF1, insulin growth factor 1; TNF, tumor necrosis factor; TXNIP, thioredoxin-interacting protein; IL-18, interleukin 18; PTGS2, prostaglandin-endoperoxide synthase 2 with highest confidence at 7 d in the ECC + CON leg. In addition to this, differences between legs in the cellular events predicted at 24 h were still evident at 7 d, that is, formation of muscle was predicted to be inhibited in CON but activated in ECC + CON, while the predicted inhibition of muscle function in ECC + CON was not seen in CON.

| DISCUSSION
The present findings revealed that the greater amount of work done during 84 d of unilateral combined concentric and eccentric maximal voluntary isokinetic exercise training in young, trained males did not result in differential temporal responses or gains in muscle CSA, muscle volume or strength when compared with the contralateral leg which performed only maximal voluntary concentric exercise training. Of importance, mRNA abundance associated with a number of cellular functions was altered from baseline in both limbs throughout the course of training, with changes in abundance being most marked at 24 h and 7 d in the limb that performed combined concentric and eccentric training. However, the early changes in mRNA abundance had waned in both limbs by 28 and 84 d of training. We conclude that initial mRNA abundance responses to maximal load exercise training are unlikely to reflect chronic training-induced muscle hypertrophy, suggesting some caution when extrapolating chronic muscle adaptive responses from acute exercise intervention studies. Moreover, neither the greater amount of work done over the course of 84 d of high-load combined concentric and eccentric training, nor marked differences in mRNA responses during the early phase of training, impacted on muscle CSA, volume or strength gains compared to concentric training alone.
We detected a ~4% increase in muscle CSA in both limbs. This magnitude of change is similar to other studies comparing hypertrophic responses to 10-12 weeks of maximal concentric resistance training versus maximal eccentric resistance training. 4,24 However, some studies have shown that eccentric training produces greater gains in muscle mass than concentric training at low-moderate intensities 30 and high intensities. 1 A meta-analysis by Roig et al 3 considered a number of studies that utilized a variety of exercise protocols, to compare muscle mass and strength gains following concentric versus eccentric training. The authors reported that eccentric training performed at high intensities may have resulted in greater muscle mass gains than concentric training. Of note however, in those studies where workloads were F I G U R E 7 Differentially regulated muscle mRNAs (outer ring) associated with skeletal and muscular system development and function in (A) concentric (CON) and (B) combined eccentric + concentric (ECC + CON) trained legs at 7 d of training compared to baseline, and cellular events predicted by Ingenuity Pathway Analysis to results from the collective changes in mRNA abundance. AKT1, AKT serine/threonine kinase 1; ATF3, activating transcription factor 3; CYR61, cysteine-rich angiogenic inducer 61; FBXO32, F-box protein 32; FOXO1, forkhead box protein O 1; IGF1, insulin growth factor 1; IL-6, interleukin 6; IL-18, interleukin 18; MSTN, myostatin; MYC, MYC proto-oncogene; MYF5, myogenic factor 5; MYF6, myogenic factor 6; PAX3, paired box 3; PAX7, paired box 7; PDK4, pyruvate dehydrogenase kinase 4; PPARGC1A, peroxisome proliferator-activated receptor α; PRKAA1, protein kinase AMP-activated catalytic subunit α 1; PTGS2, prostaglandin-endoperoxide synthase 2; PTK2, protein tyrosine kinase 2; SIRT1, sirtuin 1; SRF, serum response factor; TNC, tenascin C; TNF, tumor necrosis factor; TXNIP, thioredoxininteracting protein; VEGFA, vascular endothelial growth factor A matched between concentric and eccentric training protocols, no difference was noted in muscle mass gains between the two modalities. The authors concluded therefore that the greater muscle forces generated during eccentric training was responsible for the differential gains in muscle mass. Moore et al 21 tested this proposition by having young male volunteers perform lengthening contractions of the biceps brachii using an isokinetic dynamometer with one arm and shortening contractions in the contralateral arm two times per week for nine weeks, but with work output matched between arms. In keeping with the conclusion of Roig et al, 3 a similar increase in muscle CSA was observed in each limb at the end of the training period. Conversely, more recent studies have reported when workloads were unmatched between concentric and eccentric training at 80% of 1-RM (1 repetition-maximum, using modified leg press), still no differences in gains in vastus lateralis muscle volume, 4 or vastus lateralis muscle thickness 31 were demonstrable over 8-10 weeks of training.
Here we show that despite the ECC + CON leg performing more work per week ( Figure 2A) and across the whole training period ( Figure 2B), no differential temporal gains in muscle CSA ( Figure 3A), volume ( Figure 3B) or isometric strength over the CON leg were observed. This observation is supported by Cadore et al 32 who reported no difference in the increase in vastus lateralis muscle thickness, determined using ultrasound, over the course of 6-weeks isokinetic eccentric exercise training of the quadriceps compared to another group of volunteers who performed isokinetic concentric exercise, despite work output being ~40% greater over the course of training in the eccentric exercise trained group. Indeed, several studies have reported no differences in muscle mass adaptation when comparing maximal eccentric vs maximal concentric exercise training, 4,18 and recent studies have shown similar hypertrophic responses following concentric training versus eccentric training at 80% of 1-RM when workloads were not matched between the two training modalities.
Many studies to date have focused on acute muscle responses to eccentric exercise to provide insight on chronic muscle adaptation. However, based on the collective results of this study, it would be misleading to conclude that early mRNA responses to eccentric exercise training will reflect chronic training-induced muscle hypertrophy. This is illustrated by Figure 5 where, with perhaps the exception of lipid metabolism, the heightened effect of combined eccentric and concentric exercise on a number of cellular functions altered by exercise was found to be transitory. Moreover, Figures 6  and 7 clearly demonstrate mRNA responses in the first week of training were very different when comparing exercise modalities, with wholesale differences in mRNA abundance changes from baseline and cellular predictions evident between limbs. It remains unclear why the magnitude of mRNA responses declined from 28 d onwards in the ECC + CON limb despite training workloads remaining different between limbs.
Acute exercise intervention studies have reported greater muscle damage and inflammatory responses to be associated with high-load eccentric exercise, which has been proposed to stimulate myogenesis and cellular processes involved in muscle growth. For example, utilizing gene transcript profiling Chen et al 10 reported a greater muscle inflammatory response 4-8 h following 300 eccentric contractions of one leg when compared with 300 concentric contractions in the contralateral leg. Similarly, Kostek et al 23 identified 51 differentially expressed genes associated with protein synthesis, stress response/early growth, and sarcolemmal structure when comparing concentrically and eccentrically exercised limbs (in the form of rising from and controlled lowering to a seated position) over 24 h following exercise, albeit in n = 3 volunteers. In keeping with this, local NSAID infusion following acute maximal eccentric exercise in healthy volunteers was reported to decrease muscle satellite cell proliferation. 12 On the other hand, Deane et al 33 found the muscle transcriptional response to unilateral concentric and eccentric resistance exercise in young volunteers (using a modified leg press) was highly similar following 5 h of recovery, which may have been attributable to the volunteers being exercise naïve. The point being made here is that studies to date have invariably only documented the acute muscle responses to eccentric and concentric exercise. In the context of this point, although "organismal injury and abnormalities" was identified by IPA as a cellular function significantly altered by training in the present study, which in keeping with the literature was most altered in the ECC + CON limb at 24 h and 7 d of training compared to the CON limb ( Figure 5), the magnitude of this insult had waned in both limbs by 28 d of training, and to an extent that the response between limbs was similar at 28 and 84 d ( Figure 5). In short, the heightened effect of eccentric exercise on the muscle injury was shortlived and may therefore explain why eccentric training did not produce greater gains in muscle CSA and volume in the present study.
As an aside, studies investigating the impact of resistance exercise-induced inflammation per se on myogenesis and other processes linked to regulation of muscle mass appear to be equivocal. Using robust experimental designs involving local and systemic NSAID administration, it has been reported that the presence of inflammatory mediators in the first eight days of post-eccentric exercise recovery stimulated satellite cell proliferation (greater Pax7 + cells per myofiber) Mikkelsen et al. 12 Importantly, few studies have investigated the impact of chronic resistance exercise-induced inflammation on muscle adaptation. Trappe et al 17 reported 6-weeks of progressive resistance exercise training produced a ~25%-50% greater increase in muscle volume in elderly volunteers consuming either acetaminophen or ibuprofen during the training period. The authors suggested that NSAID ingestion, by blocking inflammation, positively impacted on muscle hypertrophy. This may indeed be the case, given a number of cellular functions linked to muscle injury and growth were found to be significantly altered throughout the 12 weeks of exercise training in the present study, irrespective of whether concentric or eccentric exercise was performed ( Figure 5). However, we must contextualize such a conclusion, given our participants were young, healthy males. Aging is associated with low-grade inflammatory state. 34 Hence, the muscle hypertrophy noted by Trappe et al 17 in elderly volunteers in response to a combination NSAID ingestion and progressive resistance training might have been due to an inhibition of this low-grade inflammation, rather than solely of inflammation induced by resistance training.
We examined the temporal effects of concentric versus combined concentric and eccentric chronic training on muscle CSA, muscle volume and strength. We observed that muscle CSA ( Figure 3A) and volume ( Figure 3B) increased by ~4% by 28 d with both training modalities, but perhaps surprisingly was unchanged thereafter despite work done during training being maintained. In keeping with this observation, DeFreitas et al 35 demonstrated the magnitude of increase in muscle CSA from baseline was greatest in the initial 4 weeks (6.7%) of high-intensity resistance training of the quadriceps compared to the increase recorded over 8 weeks (9.6%). Similarly, Brook et al 36 demonstrated a 2.66 ± 0.09 cm increase in vastus lateralis thickness from 0-3 weeks of leg resistance exercise training, which was unchanged (2.72 ± 0.09 cm) after 6 weeks despite a progressive increase in the training workload in this time. In keeping with this observation, by quantifying chronic changes in MPS using deuterated water tracer incorporation, the authors were also able to report an increase in quadriceps MPS over the initial 3 weeks of training, which then waned from 3-6 weeks. Hence, hypertrophic remodeling was most active during the early stages of training. This is in line with evidence indicating that chronic resistance training blunts the MPS response to acute resistance training. 37 Therefore, it seems that the MPS response to resistance training in the present study reached a plateau in both CON and ECC + CON legs between 28 and 84 d, which supports the concept that muscle adaptations to resistance exercise do not occur in a linear fashion.
Clearly, it is important to identify why greater force generation during eccentric training did not translate to greater gains in muscle CSA and volume in the present study. One possible explanation is that work produced by the muscle did not equate to that measured at the joint during the isokinetic maneuvers employed in the present study, 38 which could be viewed as a methodological limitation. An alternative explanation is that the MPS response to resistance exercise reaches an upper ceiling at submaximal exercise intensities. Kumar et al 24 reported that the MPS response to resistance exercise was dose-dependent and plateaued at an intensity of 75%-80% of 1-RM, suggesting that MPS was maximized in both CON and ECC + CON legs in the present study. Alternatively, it is known that the energy cost of eccentric exercise is considerably less than concentric exercise, despite muscle mechanical force generation being greater. 39 This being the case, it is entirely feasible that increases in muscle metabolic stressors known to be linked to muscle protein synthetic and hypertrophic responses, such as the cellular energy charge, the cellular redox state, and muscle lactate and pH, 40 would be less during eccentric compared with concentric exercise. However, is worthy of note that the plateauing of the muscle hypertrophic response to training in both limbs after 28 d in the present study coincided with the waning of muscle mRNA responses in both limbs. Additionally, there was no progression in the training workload after 28 d of training (Figure 2A), suggesting a connection between events. Indeed, as volunteers performed maximal intensity isokinetic contractions throughout training, one possibility is that an over training response occurred at the onset of training in both limbs that blunted subsequent adaptations.
One potential limitation of a within-subject bilateral study design is the possibility of a cross-education effect. However, this invariably applies to muscle functional adaptations. 41 Given primary end-point measurements of the present study were muscle hypertrophy and localized muscle mRNA responses to the two different exercise modalities (rather than functional gains), we foresaw cross-education having little to no impact. Indeed, since both limbs performed maximal intensity contractions, and the ECC + CON limb performed 20% more work throughout training when compared with CON (due to the biomechanical properties of an eccentric contraction; see Figure 2), it is difficult to rationalize how a cross-over effect could have impacted on these primary end-point measurements. This conclusion is corroborated by muscle CSA and volume gains being no different throughout training, and muscle mRNA expression changes from baseline being similar at 28 and 84 d of training when comparing limbs. Moreover, strength gains were different between ECC + CON and CON limbs, which also supports our stance.
To conclude, this study found that while combined maximal, isokinetic concentric and eccentric training produced greater work output during a 12-week training regimen compared to maximal, isokinetic concentric training, this did not translate to more rapid or greater increases in quadriceps muscle CSA, volume or strength. Furthermore, changes in muscle mRNA abundance associated with a number of cellular functions were observed with both training modalities, but was most marked at 24 h and 7 d in the limb that performed combined concentric and eccentric training. However, the early changes in mRNA abundance waned in both limbs by 28 d of training, and it is unclear why this occurred in the face of training workloads being sustained and different between limbs. Thus, neither the greater amount of mechanical work done during combined eccentric and concentric training, nor marked differences in mRNA responses during the early phase of training, was associated with differential gains in muscle CSA, volume or strength compared to concentric training alone. This may be explained by a non-linear relationship existing between mechanical force generation and muscle protein synthetic responses to exercise and/or the blunted metabolic stress of eccentric exercise diminishing hypertrophic stimulation.

| PERSPECTIVE
Whether high-load eccentric exercise training induces greater muscle hypertrophy and strength gains than concentric exercise training alone, possibly due to greater muscle loading and/or inflammation, is controversial. In this study, 12 weeks of combined maximal concentric-eccentric training did not produce greater hypertrophy or strength gains compared to maximal concentric training alone, despite ~20% more work being done in the former. TaqMan lowdensity gene arrays revealed changes in several cellular functions associated with muscle mass regulation/metabolism during the first week of training with both exercise modalities, but particularly during concentric-eccentric training. However, these mRNA responses waned thereafter regardless of exercise modality. Initial mRNA responses to maximal load exercise training did not reflect chronic training-induced muscle hypertrophy, suggesting some caution when extrapolating chronic adaptive responses from acute studies. Moreover, neither greater loading nor differential mRNA responses with combined concentric-eccentric training produced greater hypertrophy or strength gains than concentric training alone.