3.1 Participant Characteristics
Participant characteristics for each FTO rs9939609 genotype were similar with no significant differences in total body mass, height, BMI, hip and waist circumference, fat mass, lean muscle mass, bone mineral content, blood pressure or VO2peak noted (p > 0.05) (Table 1). A genotype effect was detected for age (p = 0.038), with AT genotypes significantly older than TT genotypes (p = 0.019). A trend towards significance for a genotype effect was detected for fasting plasma glucose concentrations (p = 0.057).
Table 1. Participant characteristics when separated by FTO genotype of the rs9939609 polymorphism.
|
Genotype
|
|
|
AA
|
|
AT
|
|
TT
|
P value
|
n
|
10
|
|
9
|
|
9
|
/
|
Sex
|
5F / 5M
|
|
4F / 5M
|
|
6F / 3M
|
/
|
Age (yr)
|
24.4 ± 1.7
|
|
29.3 ± 2.2
|
|
22.7 ± 1.2
|
0.038
|
Total Body Mass (kg)
|
74.5 ± 3.9
|
|
72.0 ± 2.6
|
|
72.6 ± 4.2
|
0.875
|
Height (cm)
|
176.1 ± 2.8
|
|
169.9 ± 2.4
|
|
172.1 ± 1.9
|
0.212
|
BMI (kg/m2)
|
24.0 ± 1.0
|
|
25.0 ± 1.1
|
|
24.4 ± 1.1
|
0.791
|
Hip Circumference (cm)
|
99.6 ± 2.4
|
|
99.6 ± 2.4
|
|
101.4 ± 1.8
|
0.813
|
Waist Circumference (cm)
|
81.8 ± 3.8
|
|
79.6 ± 2.8
|
|
77.8 ± 3.0
|
0.674
|
Fat Mass (%)
|
24.9 ± 1.6
|
|
22.6 ± 3.3
|
|
27.2 ± 2.0
|
0.416
|
Fat Mass (kg)
|
18.1 ± 1.0
|
|
16.1 ± 2.7
|
|
19.6 ± 2.0
|
0.486
|
Lean Muscle Mass (kg)
|
53.4 ± 3.5
|
|
52.3 ± 2.7
|
|
52.1 ± 3.3
|
0.689
|
Bone Mineral Content (kg)
|
2.6 ± 0.1
|
|
2.5 ± 0.1
|
|
2.5 ± 0.1
|
0.737
|
Systolic BP (mmHg)
|
129.0 ± 5.1
|
|
128.7 ± 4.2
|
|
122.7 ± 1.9
|
0.480
|
Diastolic BP (mmHg)
|
76.4 ± 3.2
|
|
75.8 ± 2.5
|
|
75.1 ± 3.0
|
0.952
|
Fasting Plasma Glucose (mmol.L-1)
|
5.0 ± 0.1
|
|
5.4 ± 0.1
|
|
4.9 ± 0.1
|
0.057
|
VO2peak (ml.kg.min-1)
|
40.0 ± 1.4
|
|
39.0 ± 2.6
|
|
37.9 ± 2.3
|
0.771
|
Values are expressed as mean ± SEM. F, female; M, male; VO2peak, peak oxygen uptake.
3.2 Participant’s Physiological Reponses to HI and LO Exercise Trials
Workloads (W) performed during the HI and LO intensity exercise protocols were similar between genotypes (p > 0.05) (Table 2). Both HI and LO intensity exercise trials elicited an increase in heart rate, with higher elevations during the HI trial compared to the LO trial (data not shown). Heart rate was similar between genotypes before, during and following HI and LO intensity exercise within each trial (data not shown). Additionally, RPE (considered on a numerical scale and presented as median (interquartile range)) was similar between genotypes at the completion of each exercise protocol (HI: AA, 16 (14 – 17), AT 16 (15 – 18), TT, 17 (15 – 19) (representing “Hard – Very Hard”) (p = 0.254); LO: AA, 12 (11 – 13), AT 11 (11 – 12), TT, 13 (11 – 14) (representing “Fairly Light – Somewhat Hard”) (p = 0.456)). There were no significant differences between genotypes in time to expend 400 kcal during the HI (p = 0.511) and LO intensity (p = 0.472) exercise protocol, or for average RER (HI, p = 0.323; LO, p = 0.603), glucose utilisation (g.kgLBM-1.T.I-1) (HI, p = 0.740; LO, p = 0.310) and fat utilisation (g.kgLBM-1.T.I-1) (HI, p = 0.709; LO, p = 0.498) measured during each exercise protocol (Table 2). The transition in substrate utilisation during exercise (RER AUC) was not significantly different between genotypes for each exercise protocol (HI, p = 0.206; LO, p = 0.410) (Table 2). The absence of an effect of age or sex on respiratory gas exchange measurements between FTO genotypes was confirmed by ANCOVA (p > 0.05).
Table 2. Respiratory gas exchange measurements, and calculated fat and glucose utilisation, between FTO rs9939609 genotypes after isocaloric HI (80% VO2peak) and LO (40% VO2peak) intensity exercise.
|
Genotype
|
P value
|
|
AA
|
AT
|
TT
|
HI
|
|
|
|
|
Workload (W)
|
127.6 ± 13.1
|
126.0 ± 7.5
|
113.9 ± 7.6
|
0.767
|
T.I (min:sec)
|
36:45 ± 2:00
|
39:29 ± 3:07
|
41:21 ± 3:02
|
0.511
|
Av. VO2 (ml.kgbw.min-1)
|
29.0 ± 1.4
|
28.7 ± 2.1
|
27.0 ± 1.7
|
0.674
|
Av. RER
|
0.96 ± 0.07
|
0.99 ± 0.02
|
0.96 ± 0.01
|
0.323
|
RER AUC
|
62.7 ± 4.4
|
58.9 ± 5.8
|
71.9 ± 5.0
|
0.206
|
Fat utilisation (g.kgLBM-1.T.I-1)
|
0.12 ± 0.04
|
0.10 ± 0.04
|
0.14 ± 0.03
|
0.709
|
Glucose utilisation (g.kgLBM-1.T.I-1)
|
1.58 ± 0.13
|
1.68 ± 0.04
|
1.69 ± 0.12
|
0.740
|
Fat utilisation (%)
|
14.5 ± 4.5
|
10.2 ± 3.8
|
15.8 ± 3.2
|
0.593
|
Glucose utilisation (%)
|
85.5 ± 4.5
|
89.8 ± 3.8
|
84.2 ± 3.2
|
0.593
|
|
|
|
|
|
LO
|
|
|
|
|
Workload (W)
|
63.8 ± 6.6
|
62.9 ± 7.5
|
57.1 ± 7.6
|
0.779
|
T.I (min:sec)
|
54:28 ± 2:58
|
57:59 ± 4:06
|
61:04 ± 4:19
|
0.472
|
Av. VO2 (ml.kgbw.min-1)
|
20.0 ± 0.7
|
21.3 ± 1.2
|
19.0 ± 0.9
|
0.289
|
Av. RER
|
0.89 ± 0.01
|
0.91 ± 0.02
|
0.90 ± 0.01
|
0.603
|
RER AUC
|
82.6 ± 3.4
|
77.8 ± 6.5
|
86.4 ± 6.5
|
0.410
|
Fat utilisation (g.kgLBM-1.T.I-1)
|
0.31 ± 0.04
|
0.24 ± 0.05
|
0.32 ± 0.05
|
0.498
|
Glucose utilisation (g.kgLBM-1.T.I-1)
|
1.15 ± 0.08
|
1.33 ± 0.08
|
1.29 ± 0.10
|
0.310
|
Fat utilisation (%)
|
36.6 ± 4.0
|
30.1 ± 5.6
|
34.5 ± 4.3
|
0.607
|
Glucose utilisation (%)
|
56.4 ± 5.5
|
69.9 ± 5.6
|
57.2 ± 6.2
|
0.206
|
Values expressed as mean ± SEM. AUC, area under the curve; LBM, lean body mass; Av. RER, average respiratory exchange ratio; T.I, Time interval required to expend 400kcal; Av. VO2 average oxygen uptake.
3.3 Metabolite Analysis
3.3.1 Skeletal Muscle Metabolites: Multivariate Analysis
Analysis of the chromatogram resulted in the detection of 48 identifiable metabolites (see Supplementary Table S-2 for metabolite identification details). Unpaired multivariate data models, O2PLS-DA with Pareto data scaling, were used to determine participant variation during the HI and LO intensity exercise trials, regardless of time (Figure 1 A + C). The O2PLS-DA modelling method demonstrated a similar metabolic signature between genotypes in the HI intensity exercise trial (p = 0.999), with very good validation metrics for data goodness of fit, R2X(cum) = 0.914, and very poor validation metrics for goodness of prediction, Q2 = 0.084 (Figure 1A). Orthogonal variation in metabolites (X) accounted for 44% of the variation, and orthogonal variation between genotypes (Y) accounted for 32% of the variation. The O2PLS-DA modelling method also demonstrated a similar metabolic signature between genotypes in the LO intensity exercise trial (p = 0.982), with moderate validation metrics for data goodness of fit, R2X(cum) = 0.511, and very poor validation metrics for goodness of prediction, Q2 = 0.017 (Figure 1C). Orthogonal variation in metabolites (X) accounted for 27% of the variation, and orthogonal variation between genotypes (Y) accounted for 17% of the variation.
The metabolites associated with each genotype can be extracted from the loadings scatter plot (Figure 1 B & D). Distribution of metabolites in the direction of each genotype signifies their contribution to model variation due to the respective genotype, whilst metabolites with the least importance are clustered in the centre. The metabolites likely to contribute most to each genotype in the HI intensity exercise trial model were, AA: alanine, glutamate and glycine; AT: proline, adenosine monophosphate (AMP), urea and myoinositol; TT: glycerol-3-phosphate (glyercol-3-P), glycerate-3-phosphate (glycerate-3-P) and pyrophosphate. Data from the LO intensity exercise trial did not provide sufficient power to differentiate metabolite variation in relation to genotype using loading plot observations, or to generate a secondary predictive component. AUC of the ROC curve showed a poorer fit in the LO intensity exercise trial compared to HI intensity exercise trial, with the AA genotype better described by the model than the TT genotype (see Supplementary Figure S-1).
Correlation coefficient scores based on the weighted sum of the PLS regression were used to identify the top 10 metabolites with the greatest influence on the components at each time point, regardless of exercise intensity (Supplementary Figure S-2). PLS-DA cross validation determined 27 metabolites in total with VIP scores ≥ 1. These VIP metabolites were used for subsequent univariate analysis to determine metabolic changes over time and between genotypes for the HI and LO intensity exercise trials.
3.3.2 Skeletal Muscle Metabolites: Univariate Analysis
HI Intensity Exercise: A significant main effect for time was observed for skeletal muscle alanine, erythronate, fumarate, gamma hydroxybutyric acid (GHB), glucose, glutamate, glycine, glycolate, lactate, leucine, malate, maltose, mannose, monopalmitoglycerol, nicotinamide, phenylalanine, proline, tyrosine and uric acid following HI intensity exercise (p < 0.05) (see Supplementary Table S-3). Time as a main effect approached significance for muscle ß-alanine (p = 0.052) and glycerate-3-P (p = 0.056) following HI intensity exercise. At 10 mins post HI intensity exercise, muscle alanine, erythronate, fumarate, GHB, glucose, glycolate, lactate, malate, maltose, mannose, monopalmitoglycerol and tyrosine were significantly elevated compared to pre-exercise (p < 0.05), whereas muscle glutamate and proline were significantly decreased (p < 0.05). A trend for lower muscle nicotinamide was detected at 10 mins post HI intensity exercise compared to pre-exercise (p = 0.065). At 90 mins post HI intensity exercise, muscle erythronate and maltose were significantly elevated compared to pre-exercise (p < 0.05), with a trend towards significance for higher levels for glucose (p = 0.066), glycolate (p = 0.089) and uric acid (p = 0.060). Conversely, muscle fumarate, glutamate, glycine, leucine, phenylalanine and proline were significantly lower at 90 mins post HI intensity exercise compared to pre-exercise (p < 0.05). No main effect for genotype was identified for any of the VIP muscle metabolites (p > 0.05). A significant genotype by time interaction was observed for muscle glucose (p = 0.036), with subsequent analysis revealing a significantly higher level of muscle glucose in AA genotypes compared to TT genotypes at 10 mins following HI intensity exercise (p = 0.021).
LO Intensity Exercise: A significant main effect for time was observed for skeletal muscle alanine, erythronate, fumarate, glucose, glutamate, glycolate, glycerate-3-P, lactate, malate, maltose, monopalmitoglycerol, pyrophosphate and tyrosine following LO intensity exercise (p < 0.05) (see Supplementary Table S-3). Time as a main effect approached significance for muscle mannose (p = 0.068), uric acid (p = 0.074) and phenylalanine (p = 0.086). At 10 mins post LO intensity exercise, muscle alanine, erythronate, fumarate, glucose, glycolate, lactate, malate, monopalmitoglycerol and tyrosine were significantly elevated compared to pre-exercise (p < 0.05), with a trend for elevated muscle maltose (p = 0.060). At 90 mins post LO intensity exercise, muscle erythronate, glutamate, glycerate-3-P, glycolate, lactate, monopalmitoglycerol and pyrophosphate were significantly elevated compared to pre-exercise (p < 0.05), while only a trend towards significance for elevated fumarate (p = 0.064), maltose (p = 0.068) and glucose (p = 0.074) were observed compared to pre-exercise. No main effect for genotype was identified for any of the VIP muscle metabolites (p > 0.05). Similar to the HI intensity exercise trial, a genotype by time interaction was observed for muscle glucose (p = 0.035), with subsequent analysis revealing a significantly higher level of muscle glucose in AA genotypes compared to AT (p = 0.028) and TT (p = 0.033) genotypes at 10 mins post LO intensity exercise.
The absence of an effect of age or sex on muscle metabolite responses between FTO genotypes for both the HI and LO exercise trial was confirmed by ANCOVA (p > 0.05).
3.3.3 Plasma Metabolites: Univariate Analysis
Exercise-induced changes in plasma albumin concentrations were similar between genotypes at all observed time points during both exercise trials (data not shown) (p > 0.05). A significant main effect for time for plasma glucose was observed in the HI intensity exercise trial (p < 0.001). Subsequent pairwise comparisons revealed significantly higher plasma glucose at 10 mins (p = 0.001) and 90 mins (p = 0.042) post HI intensity exercise compared to pre-exercise (Figure 2 A). No main effect for time for plasma glucose was detected in the LO intensity exercise trial (p = 0.533) (Figure 2 B). No genotype main effect (HI, p = 0.656; LO, p = 0.196), or genotype by time interaction (HI, p = 0.681; LO, p = 0.932) was identified for either exercise trial.
3.4 Skeletal Muscle mRNA Expression Analysis
HI Intensity Exercise: A significant main effect for time was observed for FTO (p = 0.002), AMPK (p = 0.009) and mTOR (mammalian target of rapamycin) (p = 0.001) mRNA expression following HI intensity exercise at 80% VO2peak (Figure 3). Time as a main effect approached significance for GLUT4 (glucose transporter type 4) mRNA expression (p = 0.054). Subsequent pairwise comparisons revealed a significant decrease in FTO (p < 0.001) and mTOR (p = 0.001) mRNA expression from pre-exercise to 10 mins post-exercise, and in mTOR (p = 0.002) from pre-exercise to 90 mins post-exercise. A significant increase in AMPK mRNA expression was observed from pre-exercise to 90 mins post-exercise (p = 0.009).
A weak trend for a genotype by time interaction was observed for FTO (p = 0.095). No genotype by time interactions were identified for the mRNA expression of AMPK (p = 0.304), GLUT4 (p = 0.366) or mTOR (p = 0.377). No genotype main effects were identified for FTO (p = 0.894), AMPK (p = 0.606), GLUT4 (p = 0.310) or mTOR (p = 0.611) mRNA expression. The absence of an effect of age or sex on FTO mRNA expression during the HI intensity exercise trial was confirmed by ANCOVA (p > 0.05).
LO Intensity Exercise: A significant main effect for time was observed for AMPK mRNA expression following LO intensity exercise at 40% VO2peak (p = 0.009), with subsequent pairwise comparisons revealing a significance increase in AMPK mRNA from pre-exercise to 10 mins post-exercise (p = 0.005) and from pre-exercise to 90 mins post-exercise (p = 0.004) (Figure 3). Time as a main effect approached significance for GLUT4 (p = 0.093) mRNA expression following LO intensity exercise, with no main effect for time observed for FTO (p = 0.505) or mTOR (p = 0.642) mRNA expression.
No genotype main effects (FTO, p = 0.931; AMPK, p = 0.804; GLUT4, p = 0.164; mTOR, p = 0.280), or genotype by time interactions (FTO, p = 0.970; AMPK, p = 0.803; GLUT4, p = 0.277; mTOR, p = 0.528) were identified. The absence of an effect of age or sex on FTO mRNA expression during the LO intensity exercise trial was confirmed by ANCOVA (p > 0.05).
3.4.1 Regression Analysis of mRNA
Regression analysis was used to determine if a relationship between skeletal muscle FTO mRNA and muscle glucose existed, as glucose was the only metabolite to demonstrate a genotype by time interaction. Further regression analyses between FTO mRNA and mRNA of other metabolic genes are in Supplementary Figure S-3.
A negative correlation was detected between skeletal muscle glucose levels and mRNA expression of FTO during the HI (r = -0.234, p = 0.033) and LO intensity exercise trial (r = -0.264 p = 0.017), regardless of time and genotype (see Supplementary Figure S-4). When accounting for genotype, the negative correlation remained between skeletal muscle glucose levels and mRNA expression of FTO in AA genotypes (HI, r = -0.370, p = 0.044; LO, r = -0.395, p = 0.031), during both exercise trials. No relationship between skeletal muscle glucose levels and mRNA expression of FTO was observed in AT genotypes (HI, r = -0.205, p = 0.306; LO, r = -0.291, p = 0.141) or TT genotypes (HI, r = -0.100, p = 0.621; LO, r = 0.027, p = 0.899), during either the HI or LO intensity exercise trials. Further regression analyses between mRNA of other metabolic genes (AMPK, mTOR and GLUT4) and muscle glucose are in Supplementary Figure S-4.
3.5 Skeletal Muscle Protein Expression Analysis
No main effect for time (p = 0.128), genotype main effect (p = 0.181), or genotype by time interaction (p = 0.485) was detected for skeletal muscle FTO protein expression in response to HI intensity exercise (Figure 4B). Similarly, LO intensity exercise did not have a significant effect on skeletal muscle FTO protein expression, with no main effect for time (p = 0.544), genotype main effect (p = 0.378) or genotype by time interaction (p = 0.650) observed (Figure 4C).
A significant main effect for time was observed for pAS160Ser588 (p = 0.049; Figure 4D) and pAMPK (p = 0.035; Figure 4E) in skeletal muscle following HI intensity exercise. Subsequent pairwise comparisons revealed a significant increase in pAMPK relative to total AMPK from pre-exercise to 10 mins (p = 0.010) post-exercise, and a significant increase in pAS160Ser588 relative to total AS160 from pre-exercise to 10 mins (p = 0.011) and 90 mins (p = 0.046) post-exercise. No genotype main effects (pAMPK, p = 0.563; pAS160Ser588, p = 0.252), or genotype by time interactions (pAMPK, p = 0.490; pAS160Ser588, p = 0.386) were identified.
The absence of an effect of age or sex on FTO and pAMPK protein expression between FTO genotypes for both the HI and LO exercise trial was confirmed by ANCOVA (p > 0.05). However, there was a significant sex effect for pAS160Ser588, with females demonstrating a greater increase at 10 min recovery (p = 0.042) compared to males. No effect of age was evident. No relationships were detected between skeletal muscle levels of pAS160Ser588, pAMPK or FTO expression and the mRNA expression of FTO, AMPK, mTOR or GLUT4 during the exercise trials, regardless of time and genotype (see Supplementary Figure S-5).