Comparing Acute, High Dietary Protein and Carbohydrate Intake on Transcriptional Biomarkers, Fuel Utilisation and Exercise Performance in Trained Male Runners

This study was conducted in accordance with the Declaration of Helsinki (2013), and the protocol was approved by the University of Hertfordshire Life and Medical Sciences Ethics Committee (protocol number LMS/PGR/UH/02227). Following a study briefing and suitability interview, participants were invited to take part. Power calculation assessment for sample size (G *power3, Dusseldorf) using α = 0.05; 1 − β = 0.80; based on observed data of PGC-1α mRNA expression [9] indicated that a total of 6 participants were required per group. All participants were required to be male and between the age of 18–45 years, and were recruited through local running clubs, and research team network. From an initial pool of twenty individuals, three withdrawals were unrelated to the trial, and one did not complete the second trial and was excluded from subsequent analyses. Therefore, sixteen healthy, trained male volunteers completed this study (participant characteristics are shown in Table 1 according to allocated test group). Written informed consent was obtained from all participants included in the study. All participants were required to be free from any medical condition which would conflict with the study parameters. Additionally, all participants were required to refrain from following a restrictive dietary regime, not suffer from food allergies or intolerances, and be prepared to consume a controlled, prescribed dietary intake for each trial period. For training suitability, all participants needed to have completed 10-km in sub-38 min or 5-km in sub-18 min in the previous 12 month period, with a running history >2 years, training ≥4 times per week with no significant change in training load or body mass in the previous 3 months.

In an investigator-blinded, parallel-groups, repeated-measures design, participants were randomly assigned to either: PRO (n = 8; total energy intake macronutrient distribution of 40% protein, 30% carbohydrate, and 30% fat) or energy-matched CHO (n = 8; macronutrient distribution of 10% protein, 60% carbohydrate, and 30% fat). Participants were not informed of group allocation; however, due to the nature of dietary intervention studies it was not possible to blind food items. Participants attended the Human Performance Laboratory (GlaxoSmithKline, Brentford, UK) for baseline testing and a familiarisation session, followed by three experimental sessions (T1, T2, T3). Participants followed their habitual dietary intake for 7-days prior to completing T1, after which they followed their prescribed intervention diet (PRO or CHO) for 7-days prior to T2, and finally returned back to their habitual dietary intake for a further 7-days prior to T3. Participants were requested to complete the same training sessions each week throughout the duration of the study and rest the day prior to each test visit. All experimental procedures were completed in the same order and time of the day.

Baseline and Familiarisation Session: All testing was completed in a temperature- controlled room (TISs Sport Science environmental chamber) set at 20 °C and 40% humidity (averaging 20.7 ± 0.8 °C, 41.3 ± 5.0% humidity and 1010.0 ± 13.6 mmHg air pressure across the trials) using the same treadmill (Cosmed T170, Cosmed, Rome, Italy), which was not moved throughout the experimental period. Fan cooling was provided throughout all testing, with a floor fan positioned 50 cm in front of the treadmill at a 45° angle and set to an air speed of ~5.8 m·s⁻¹. Prior to assessment of maximal oxygen consumption (V˙O_2max), participants rested for 5-min, after which height (mechanical stadiometer, Seca 216, seca Ltd., Hamburg, Germany) and weight (digital column scale, Seca 704) were recorded and a heart rate (HR) belt fitted (Polar T31 Coded, Polar Electro Ltd., Warwick, UK). A 20 µL resting, capillarised, fingertip blood sample was then collected, and participants then undertook a standardised 5 min warm up at a running speed of 9 km·h⁻¹.

The protocol for the assessment of V˙O_2max was the same for all participants. A breath-by-breath online gas analyser (Metalyzer 3B, Cortex, Leipzig, Germany) was used to capture continuous respiratory measures. Criteria for the attainment of V˙O_2max complied with that outlined by the British Association of Sport and Exercise Sciences (BASES—[24]). Following a 20 min recovery period, participants completed a 5 km steady state (SS) familiarisation treadmill run at 70% vV˙O_2max (~75% V˙O_2max) in a similar manner to the 10 km SS exercise trial detailed below. Once completed, participants had 5 min recovery before completing a maximal sustainable effort at 95% V˙O_2max (calculated from the previous V˙O_2max test) following the protocol detailed below. Prior to departure, participants were familiarised with all other testing procedures, and provided with a food diary and a GPS watch (Garmin Vivo Active, Garmin Ltd., Schaffhausen, Switzerland) to log dietary intake and training volume.

Assessment of Habitual Dietary Intake: A 3-day food diary based on estimated household measures was used [25] to collect habitual dietary intake prior to T1 and T3. During the study briefing, the participants were instructed how to complete the food diary and were provided with a comprehensive example. The importance of accuracy and detail was emphasised (e.g., item weight, portion size, meal breakdown, and fluid intake) as was the importance of maintaining current dietary habits and documenting all food and drink consumed. Participants were requested to record dietary intake on 2 weekdays and 1 weekend day in the 7-day period preceding T1 and T3 to estimate habitual intake. All food diaries were analysed by the same researcher using Nutritics Professional Dietary Analysis software (Nutritics Ltd., Co., Dublin, Ireland).

Assessment of Training Volume: It was important that any observed changes were due to the intervention diet and not due to change in training volume. Therefore, participants were requested to maintain the same weekly training programme throughout the study and were requested to replicate, as closely as possible, each training session on the same day and time each week, logging all training sessions with the GPS watch provided. The training sessions were automatically uploaded online (Garmin Connect, Garmin Ltd., Schaffhausen, Switzerland) for the investigative team.

Caloric Intake and Intervention Dietary Prescription: Individual caloric intake for the intervention diet was prescribed to match energy expenditure, and was calculated using basal metabolic rate (BMR) multiplied by a physical activity factor (PAF). The equation of Henry [26] was used to estimate BMR. The PAF was quantified based on occupational and exercise level factors using a modified Harris and Benedict [27] equation, incorporating estimated non-exercise activity thermogenesis.

A total of 34 diet plans (17 CHO, 17 PRO) were formulated by the investigative team, starting at 1800 kcal·d⁻¹ and increasing by 150 kcal·d⁻¹ for each plan up to 4200 kcal·d⁻¹. Participants were assigned the diet closest to their calculated interventional energy intake. The greatest difference between estimated intervention diet and prescribed diet was 75 kcal per day. All meals were produced by Soulmatefood^® (Waterfoot, Lancashire, UK) and delivered to each individual participant in two batches (a 3-day then a 4-day supply). Daily food packages consisted of 5 pre-packaged/cooked meals (breakfast, morning snack, lunch, afternoon snack and dinner). Matching energy intake to energy expenditure is challenging [25] and it was important that the participants were not in a negative energy balance for this study. Therefore, in addition to the calculated dietary intake, a further 500 kcal·day⁻¹ macronutrient-matched meal was provided. Included with the delivery was a daily menu with consumption instructions and participants were briefed that the additional meal was only to be consumed if still hungry after consuming the standard diet and were required to document if the additional meal was consumed. Participants were requested to remain hydrated through the day, only consume foods provided and requested to drink water ad libitum or drinks free from caffeine or additional energy. The mean dietary intake for both prescribed diets is shown in Table 2.

Experimental Laboratory Sessions (T1–T3): After a rest day from exercise and following a standardised pre-testing protocol, participants arrived at the laboratory at the same time of day for each visit. Participants were advised to arrive at the laboratory using the same mode of transport and requested to refrain from morning physical exertion. Upon arrival, body composition assessment (dual energy x-ray absorptiometry (DXA) scan) was undertaken, followed by a resting micro muscle-biopsy. Participants then completed a 10-km treadmill run at 75% V˙O_2max, had 5-min rest, then completed a time to exhaustion (TTE) running trial at 95% V˙O_2max.

Body Composition Assessment: Following measurement of fasted body mass (BM), participants were requested to remove any metal piercings prior to having a full body DXA scan (GE Lunar iDXA, GE Healthcare, Bucks, UK). The DXA scan was performed in accordance with manufacturer’s guidelines for patient positioning and analysed using enCORE Software, version 14.10 (GE Healthcare, Bucks, UK). All scans were performed by the same operator. Results were analysed for fat mass (FM) and fat free mass (FFM).

Micro Muscle-Biopsy: Muscle biopsies were obtained from the midpoint on the lateral aspect of the right vastus lateralis muscle. The biopsy site was cleaned using Betadine (Pardue Products, Stamford, CT, USA) and samples were obtained under local anaesthesia (2 mL of 1% v/v without adrenaline, Lidocaine Hydrochloride (Hameln Pharmaceuticals: cat. no PL01502) injected into the subcutaneous fat of the selected biopsy site). Following insertion of a 14-gauge co-axial needle ~2 cm into the muscle, a disposable core biopsy instrument (TSK Stericut Biopsy Needle 14 Gauge, TSK Laboratories, Tochigi-Shi, Tochigi-Ken, Japan) was inserted through the co-axial and activated. A single biopsy pass was used collecting ~15–20 mg of muscle tissue. The biopsy instrument was immediately removed, and within 10 s the muscle tissue was removed from the biopsy instrument using a sterile scalpel and flash frozen in liquid nitrogen. The muscle sample was placed in an individual vial and stored at −80 °C until analysis (HTA license number 12202).

10-km Steady State Treadmill Run: Following a resting 20 µL capillary blood sample, participants completed a standardised 5 min warm up at 10 km·h⁻¹ on the treadmill, followed by self-directed stretching. Along with a HR monitor, participants were fitted with an oro-nasal facemask in the same manner as the V˙O_2max test, connected to an online gas analyser for breath-to-breath assessment of respiratory measures. The steady state 10 km run was calculated to be performed at 75% V˙O_2max, with a rolling start (calculations post-trial indicated an average of 75.6 ± 2.6% V˙O_2max with no difference between groups observed: PRO = 76.4 ± 3.2%; CHO = 74.7 ± 2.4%, p > 0.05). After each kilometre, standardised prompts were provided to participants along with collection of a capillary blood sample and assessment of HR and rated perceived exertion [28]. Rates of carbohydrate and fat oxidation were calculated with the attained V˙O₂ and V˙CO₂ data using stoichiometric equations [29], with protein oxidation assumed negligible.

Time to Exhaustion (TTE) Trial: Following a 5 min recovery period, participants then performed a maximal sustainable effort (MaxSE) test at a speed equivalent to 95% V˙O_2max. No visual or audible encouragement was permitted throughout the trial. Participants ran until volitional fatigue, upon which a final capillary blood sample was collected. The same investigator timed all TTE efforts throughout the study.

RNA Isolation: Human muscle biopsies were homogenised in 700 µL MagNA Pure LC RNA Isolation Tissue Lysis buffer (Roche, Mannheim, Germany) in Roche MagNA Lyser Green bead tubes at 6500 rpm for 30 s. After homogenisation, the tubes were centrifuged for 10 min at 13,000× g before 350 µL of each homogenate was used for RNA extraction. Total RNA was extracted using the MagNA Pure 96 Cellular RNA Large Volume Kit on a MagNA Pure 96 (Roche, Mannheim, Germany), in an elution volume of 50 µL according to the manufacturer’s protocol. RNA was stored at −80 °C. RNA concentrations were determined (A260) using a NanoQuant plate on a Tecan Infinite 200PRO. RNA quality (RNA integrity number equivalent, RINe) was assessed using RNA ScreenTape on a 2200 TapeStation (Agilent, Santa Clara, CA, USA).

Reverse Transcription (RT): As a result of the large sample number and the automated process, RNA concentration was not taken into account for each individual reaction. Fourteen µL of each RNA sample was reverse transcribed in a total volume of 20 µL using the High capacity cDNA reverse transcription kit (without RNase inhibitor, Applied Biosystems, Thermo Fisher, Loughborough, UK). The average RNA concentration across the 142 samples was 32.4 ng·µL⁻¹ equating to an average of 453 ng RNA/reverse transcription reaction. Reactions were performed in 96 well PCR plates on a PTC-225 Peltier thermal cycler (MJ Research, St. Bruno, Quebec, Canada) using the following profile: 25 °C for 10-min, 37 °C for 60-min, 85 °C for 5-min, 4 °C hold. Minus RT control reactions were set up for 14 samples, in which the RNA component was replaced with nuclease free water (Ambion, Fisher Scientific, Loughborough, UK).

qPCR Analysis: For quantitative real-time PCR (qPCR), Human TaqMan^® gene expression assays were purchased as a 20× assay ready stock from Life Technologies (Carlsbad, USA) (primers 18 mM and probes 5 mM) (Table 3). One µL cDNA was added to each qPCR reaction mixture which also contained gene expression assay mix (primers 900 nM final, probe 250 nM final), LightCycler 480 Probes Master and nuclease free water to give a 5 µL total reaction volume. Reactions were prepared in white multiwell 384 plates (Roche, Mannheim, German) using a Mosquito HV (TTP Labtech, Melbourn, UK). The plates were sealed using optical seals (Roche, Mannheim, Germany) and centrifuged at 290× g for 2 min before being run on a LightCycler^® 480 instrument (Roche, Mannheim, Germany) with the following thermal cycling parameters: initial de-naturation step 95 °C for 10 min, followed by 45 cycles of denaturation at 95 °C for 10 s and primer annealing/extension at 60 °C for 30 s. A cooling step at 40 °C for 30 s was the final stage of the run. The crossing point (Cp) value for each sample was calculated using the second derivative maximum method applied directly by the Roche software to the real-time amplification curves. This value represents the cycle at which the increase of fluorescence is highest and where the logarithmic phase of a PCR begins.

Amplification Efficiency of qPCR Assays: The efficiency of each of the Taqman^® assays on demand was confirmed by performing standard curves in a 384 well qPCR assay on the LightCycler^® 480 under the same conditions described above. Six point standard curves with 1 in 10 serial dilutions were prepared in nuclease free water for human plasmid DNA for each of the genes being tested alongside Human genomic DNA with a top concentration of 3 × 10⁵ copies/µL and 1 × 10⁵ copies/µL, respectively. qPCR reactions were performed in triplicate and the amplification efficiency calculated for each Taqman^® assay on the basis of the equation E = (10 − 1/slope − 1) × 100 with the logarithm of the template concentration on the x axis and the average Cp plotted on the y axis.

Data Analysis of the qPCR Assays: The muscle biopsy qPCR raw Cp values were exported and normalised to the housekeeper gene (GAPDH) using an analysis of covariance method. Relative fold change in expression levels were calculated using the 2-ΔΔCT method.

Statistical analysis was conducted using the statistical package for the Social Science software program (SPSS; version 22, IBM, Armonk, NY, USA). Dependent variable distributions were assessed for Kolmogorov–Smirnov test of normality, along with visual inspection of histograms and boxplots. Mauchly’s sphericity test was used to assess the assumption of sphericity within repeated-measures effects. Unless stated, Mauchly’s test was not significant, therefore the assumption of sphericity was accepted. Group and time point effects were assessed using a two factorial mixed design ANOVA (4 groups, 2 time points) with dietary intake (between) and time point (within) as the main variables. To determine the location of the time-point differences and group interactions, a Bonferroni post hoc test was used. Statistical significance was set at p ≤ 0.05. Data presented as mean ± SEMs (unless stated).

No significant group interactions or time point differences were observed in habitual dietary intake, body mass, lean mass or fat mass (p > 0.05), demonstrating energy intake matched energy requirement during both the dietary intervention and habitual intake weeks. No within- or between-group differences were observed in total training distance covered each week (training distance PRO vs. CHO, p = 0.83) across the study period. However, a significant reduction in average running speed was observed between T1 (14.7 km·h⁻¹) and T2 (13.8 km·h⁻¹) in the PRO group (p = 0.047, Table 4). Once reverted back to habitual diet, average training speed for PRO in T3 was 14.2 km·h⁻¹.

No change was observed in the resting mRNA content of PGC-1α, SIRT1 and SIRT3 at any time point in either group (Figure 1A,D,E). A within group difference main effect (p = 0.027) for AMPK mRNA was observed in the CHO group (Figure 1B) and AMPK mRNA expression was significantly elevated at T2 (1.44 fold increase relative to T1, p = 0.042), and maintained at T3 (1.57 fold increase relative to T1, p = 0.011). A within group difference (p = 0.018) was also observed in the PRO group where AMPK mRNA expression was significantly increased at T2 (1.37 fold increase relative to T1, p = 0.016); however, this returned to pre-intervention levels by T3 (p = 0.18). PPAR expression in the CHO group exhibited the same trend as AMPK, with a 1.32 and 1.29 fold increase at T2 (p = 0.029) and T3 (p = 0.044) relative to T1, respectively. No time point difference in PPAR expression was observed in the PRO group. No changes were observed for SIRT1 or SIRT3 mRNA expression within or between groups (p > 0.05).

In the CHO group, both carbohydrate and fat oxidation remained similar across all trials during the 10-km SS trial (p > 0.05, Figure 2B,D). In the PRO group a significant time point difference in fat (p < 0.001) and carbohydrate oxidation (p < 0.001) was observed, where fat oxidation was significantly increased and carbohydrate oxidation was significantly decreased at each km (1–10 km) collection point in TP2 compared to TP1 (Figure 2A,C, p < 0.0001). Mean fat oxidation following PRO increased from 0.29 ± 0.05 to 0.59 ± 0.05 g·min⁻¹ (p < 0.0001), an increase of 101% relative to T1. Correspondingly, mean carbohydrate oxidation following PRO decreased by 27% from 2.96 ± 0.18 to 2.15 ± 0.19 g·min⁻¹ (p < 0.0001). Both fat and carbohydrate oxidation returned to pre-intervention levels once reverted back to habitual dietary intake.

No differences were observed between trials during the 10 km SS trial in the CHO group for perceived exertion, blood glucose or lactate (Figure 3A–E). A between group difference in blood glucose at rest was observed in the PRO group (p = 0.031), with a significant difference observed between the intervention diet and T3 (4.1 ± 0.6 and 4.9 ± 0.7 mmol·L⁻¹, p = 0.034), and a trend to difference between T1 and T2 (4.7 ± 0.4 vs. 4.1 ± 0.6 mmol·L⁻¹, p = 0.055). No other differences were observed for PRO, or between groups.

The results of the 95% MaxSE trial are shown in Figure 4A,B, expressed as relative change (%) from T1. Following PRO, TTE was significantly reduced within group by 23.3% at T2, with all participants exhibiting decreased ability to perform (p = 0.0003). Upon subsequent return to habitual dietary intake, TTE for PRO at T3 was comparable to T1 (T1: 128.3 ± 29.3 s, T2: 98.4 ± 31.8 s, T3: 125.3 ± 32.39 s). In contrast, significant (p = 0.05) within group change of +6.5% in TTE at T2 was noted in the CHO group with 7 of the 8 participants improving relative to TP1. TTE performance for CHO also returned to pre intervention levels by T3 (T1: 182.2 ± 44.4 s, T2: 194 ± 45.5 s, T3: 184 ± 38.9 s). A between group effect in TTE was observed (p = 0.017), with no relative difference reported at T1 or T3, but at T2 there was a significant relative difference in performance (p < 0.0001).

Endurance exercise is a potent stimulator of PGC-1α via the AMPK-SIRT1 pathway [40,41], and increases in PGC-1α expression have been observed following a single bout of endurance exercise [13]. This response can be further amplified when in a single exercise bout in a low glycogen state [42,43]. Furthermore, it has been demonstrated in vitro [8] and in vivo [10] that reduced glucose and/or nutrient availability promote increased PGC-1α, AMPK and SIRT1 mRNA expression. In untrained individuals over a 14-day training period, Perry et al. [13] found a transient post-exercise response in PGC-1α expression, observing a 10-fold increase in PGC-1α mRNA expression subsequent to the first bout of endurance cycling, which was reduced to a 4-fold increase after the seventh training session. In addition, with adequate glycogen availability, Psilander et al. [43] found a 2.5-fold increase in PGC-1α mRNA expression following a single bout of endurance cycling in well-trained cyclists (V˙O_2max = 65.4 ± 0.9 mL·kg⁻¹·min⁻¹), and with the addition of glycogen restriction to the endurance session, an 8.1-fold increase in expression; similar to that observed in untrained individuals [13]. These data sets suggest that the transcriptional adaptive response is blunted when skeletal muscle becomes accustomed to the exercise stress, and increasing the metabolic stress by restricting glycogen availability may result in enhanced post exercise transcriptional adaptation to a single one-off exercise bout. However, these studies examine the impact of carbohydrate restriction after a single bout of exercise. The results from the current study provide a different perspective on the adaptive response and performance outcomes using such nutrient manipulation models, as carbohydrate restriction was prolonged for 7 days. The data does not support the use of repeated days of carbohydrate restriction on training adaptation and exercise performance, demonstrating that reduced carbohydrate availability, without repletion, during a typical training week resulted in a significant decrease in exercise performance and no discernible improvement in resting PGC-1α mRNA expression. Thus, it is likely the PRO intervention utilised in this trial is not a positive method to implement to improve endurance performance.

An important consideration is that post exercise muscle biopsies were not collected in the current study. However, the results from our previous research [23], along with others [10], demonstrate that mRNA expression is sensitive to dietary manipulation at rest. In addition, Baar et al. [12] found that PGC-1α protein responds similarly to mRNA expression, thus an assumption can be made that an increase in mRNA expression is accompanied by increased protein content. Additionally, despite initial power calculation, the effect size of the mRNA expression is fairly small, thus a larger sample size may be warranted to further validate these findings. As such these aspects of the results should be interpreted with due consideration.

Furthermore, muscle glycogen was not measured, however seminal work from Bergstrom et al. [44] demonstrated that consuming a very low carbohydrate diet for 7-days without exercising resulted in a 32% reduction in muscle glycogen. Thus, it is fair to assume that the high protein intervention, in combination with endurance training, resulted in significant reductions in muscle glycogen. The significant reduction in resting blood glucose, and increase in fat oxidation, during the sub-maximal trial in the high protein group corroborate this further.

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to ethical considerations, in accordance with consent provided by participants on the use of confidential data.