Preservation of skeletal muscle mitochondrial content in older adults: relationship between mitochondria, fibre type and high‐intensity exercise training

Abstract

Key points

• Ageing is associated with an upregulation of mitochondrial dynamics proteins mitofusin 2 (Mfn2) and mitochondrial dynamics protein 49 (MiD49) in human skeletal muscle with the increased abundance of Mfn2 being exclusive to type II muscle fibres.

• These changes occur despite a similar content of mitochondria, as measured by COXIV, NDUFA9 and complexes in their native states (Blue Native PAGE).

• Following 12 weeks of high‐intensity training (HIT), older adults exhibit a robust increase in mitochondria content, while there is a decline in Mfn2 in type II fibres.

• We propose that the upregulation of Mfn2 and MiD49 with age may be a protective mechanism to protect against mitochondrial dysfunction, in particularly in type II skeletal muscle fibres, and that exercise may have a unique protective effect negating the need for an increased turnover of mitochondria.

Abstract

Mitochondrial dynamics proteins are critical for mitochondrial turnover and maintenance of mitochondrial health. High‐intensity interval training (HIT) is a potent training modality shown to upregulate mitochondrial content in young adults but little is known about the effects of HIT on mitochondrial dynamics proteins in older adults. This study investigated the abundance of protein markers for mitochondrial dynamics and mitochondrial content in older adults compared to young adults. It also investigated the adaptability of mitochondria to 12 weeks of HIT in older adults. Both older and younger adults showed a higher abundance of mitochondrial respiratory chain subunits COXIV and NDUFA9 in type I compared with type II fibres, with no difference between the older adults and young groups. In whole muscle homogenates, older adults had higher mitofusin‐2 (Mfn2) and mitochondrial dynamics protein 49 (MiD49) contents compared to the young group. Also, older adults had higher levels of Mfn2 in type II fibres compared with young adults. Following HIT in older adults, MiD49 and Mfn2 levels were not different in whole muscle and Mfn2 content decreased in type II fibres. Increases in citrate synthase activity (55%) and mitochondrial respiratory chain subunits COXIV (37%) and NDUFA9 (48%) and mitochondrial respiratory chain complexes (∼70–100%) were observed in homogenates and/or single fibres. These findings reveal (i) a similar amount of mitochondria in muscle from young and healthy older adults and (ii) a robust increase of mitochondrial content following 12 weeks of HIT exercise in older adults.

Key points

• Ageing is associated with an upregulation of mitochondrial dynamics proteins mitofusin 2 (Mfn2) and mitochondrial dynamics protein 49 (MiD49) in human skeletal muscle with the increased abundance of Mfn2 being exclusive to type II muscle fibres.

• These changes occur despite a similar content of mitochondria, as measured by COXIV, NDUFA9 and complexes in their native states (Blue Native PAGE).

• Following 12 weeks of high‐intensity training (HIT), older adults exhibit a robust increase in mitochondria content, while there is a decline in Mfn2 in type II fibres.

• We propose that the upregulation of Mfn2 and MiD49 with age may be a protective mechanism to protect against mitochondrial dysfunction, in particularly in type II skeletal muscle fibres, and that exercise may have a unique protective effect negating the need for an increased turnover of mitochondria.

Abbreviations

BN‐PAGE

Blue Native polyacrylamide gel electrophoresis

BMI

body mass index

CON

control group

COXIV

cytochrome oxidase IV subunit

CS

citrate synthase

HIT

high‐intensity interval training

KO

knockout

Mfn2

mitofusin2

MHCI

myosin heavy chain I isoform

MHCII

myosin heavy chain II isoform

MiD49

mitochondrial dynamics protein 49

NDUFA9

NADH:ubiquinone oxidoreductase (complex I)

Post

after HIT

Pre

before HIT

ROS

reactive oxygen species

RPE

rating of perceived exertion

Introduction

Skeletal muscle mass loss and reduced muscle function are major part of the ageing process and may, at least in part, be related to reduced muscle mitochondria content and function (Joseph et al. 2012; Romanello & Sandri, 2015). Mitochondrial dynamics are critical for the turnover and maintenance of healthy mitochondria and are regulated by both fusion and fission proteins. Mitochondrial fusion allows the exchange of material between mitochondria and dilution of damaged compartments while mitochondrial fission can result in damaged components being segregated for mitophagy (Twig et al. 2008). In skeletal muscle, mitofusin2 (Mfn2, a member of the dynamin superfamily that coordinates the fusion of the outer mitochondrial membrane), is the dominant MFN expressed (Koshiba et al. 2004; Zorzano et al. 2010). Two recently identified proteins, mitochondrial dynamics proteins MiD49 and MiD51, are necessary for recruiting the fission mediator Drp1 to mitochondria and these are specific for mitochondrial fission (Richter et al. 2014). Mitochondrial fission and fusion proteins counterbalance each other, although either event can dominate under circumstances of altered energy balance (Richter et al. 2014).

There are no reports describing MiD49 abundance in skeletal muscle; however, Mfn2 research is developing. Aged mice have lower Mfn2 abundance compared to young mice (Sebastian et al. 2016). Furthermore, young whole body Mfn2 knock‐out (KO) mice have exacerbated sarcopenia through impaired autophagy and mitochondrial quality, linking Mfn2 to skeletal muscle mitochondrial health (Sebastian et al. 2016). In contrast to rodents, in humans the abundance of Mfn2 has been correlated to aerobic fitness and body composition (Distefano et al. 2017). Indeed, muscle Mfn2 content in humans does not seem to be affected by age (Joseph et al. 2012; Konopka et al. 2014; Distefano et al. 2017) while Mfn2 is upregulated by 36% with continuous aerobic training in older adults aged 74 ± 3 years (Konopka et al. 2014). The use of fractionated muscle samples in these studies, however, may confound data interpretation as the entire mitochondrial pool would not have been analysed. In muscle, around 80% of the total cellular pool consists of myofibrillar (contractile) proteins, potentially making it difficult to analyse the less abundant proteins (Murphy & Lamb, 2013). Therefore a 600 g centrifugation step is often undertaken to remove the myofibrillar proteins, and in cardiac muscle this has been shown to result in the loss of up to 80% of the sarcoplasmic reticulum (Murphy et al. 2011). It is likely that such a simple fractionation step would also have adverse ramifications for investigations of the enrichment of skeletal muscle mitochondria, but this is unknown and was explored here.

Another important consideration of mitochondrial and ageing research is the fibre type analysed. Human vastus lateralis muscle from young healthy adults contains approximately equal proportions of fast‐twitch (type II) and slow‐twitch (type I) muscle fibres. This is altered in sedentary older adults aged over ∼60 years, where a shift to an increased proportion of hybrid fibres, that is those expressing more than one myosin heavy chain isoform, is seen (St‐Jean‐Pelletier et al. 2016). Furthermore, with age type II fibres are weakened and demonstrate losses of specific force and of Ca²⁺ sensitivity, reducing contractile force production (Lamboley et al. 2015), whereas type I fibres do not appear to experience these modifications. Therefore, it is important to consider the ageing effect on the mitochondria in individual fibre types.

Exercise training increases mitochondrial content and function in young adults (Holloszy, 1967; Little et al. 2010) and has positive and important implications in the elderly (Joseph et al. 2012). Most training studies in older adults have utilised moderate intensity endurance training (Menshikova et al. 2006; Konopka et al. 2014). Training volume is not the only factor to consider when prescribing exercise, with intensity a critical regulator of exercise‐induced mitochondrial adaptation. High‐intensity interval training (HIT) is superior in upregulating mitochondrial content in healthy young adults compared to moderate‐intensity exercise (MacInnis et al. 2017) but its effectiveness in older adults is unknown.

Therefore this study had three main aims. First, to assess the abundance of mitochondrial proteins lost in samples following centrifugation compared to whole muscle homogenates. Second, to investigate the effects of age and fibre type on the abundance of mitochondrial dynamics proteins and on mitochondrial content, and thirdly to test the hypothesis that HIT will increase skeletal muscle mitochondria content in older adults.

Methods

Participants

In total, 15 young adults and 17 older adults participated in the study, the physical characteristics of these have been described elsewhere (Wyckelsma et al. 2017). Not all of the muscle measurements were conducted from samples from all participants. This study is a follow up study from another and therefore limited tissue was available, the sample sizes are indicated in the methods of each technique used in this study. A subset of 8 participants from the older group completed 12 weeks of HIT. This group comprised 2 females and 6 males; the mean age was 69.9 ± 3.8 years, stature 170.3 ± 8.4 cm, mass 74.0 ± 7.7 kg and ${\dot{V}}_{{\mathrm{O}}_2,\mathrm{peak}}$ 25.1 ± 6.1 ml kg⁻¹ min⁻¹. All participants signed informed consent and the study was approved by the Victoria University Human Research Ethics Committee (HRETH 11/221) and conforms to the Declaration of Helsinki.

Experimental overview

This study is divided in two components; the first examines the effects of age and fibre type on mitochondrial proteins, whilst the second examines the adaptability of the mitochondria to HIT in older‐adults. Before and after 12 weeks of training participants completed a ‘symptom limited’ incremental exercise test to determine the peak heart rate for training, a dual‐energy X‐ray absorptiometry (DXA) scan and a resting muscle biopsy. The post‐training biopsy was taken between 24 and 48 h after the final training session; the incremental exercise test was performed subsequent to that the biopsy.

‘Symptom‐limited’ incremental exercise test

A symptom‐limited incremental exercise test was performed on a cycle ergometer by both young and older adults, as previously described (Wyckelsma et al. 2017). Briefly, the test commenced at 20 W for both males and females and increased 20 W each minute for young adults and older males and by 10 W each minute for older females. For the safety of older adults, the test was discontinued at a symptom‐limited endpoint, defined as when participants advised a rating of perceived exertion (RPE) of 17 (very hard) on the Borg scale (Borg, 1982). For comparison against the older adults the young participants also discontinued the exercise test at an RPE of 17. The HR–work rate relationship was utilised to calculate an exercise intensity corresponding to 90–95% of the HR_peak for use in HIT.

Dual‐energy X‐ray absorptiometry (DXA) scan

Both young and older adults underwent a whole body dual‐energy X‐ray absorptiometry (DXA) scan (DXA, Hologic Discovery W, MA, USA). Each scan was performed by the same technician. Data were collected in body segments and as whole body, with lean body mass (LBM), fat mass (FM) and body fat percentage (% fat) analysed. Data were collected from 16 older adults and 8 young adults.

High‐intensity interval training

Older adults trained under supervision three times per week for 12 weeks on a mechanically braked cycle ergometer (Monark 868, Vansbro, Sweden), as previously described (Wyckelsma et al. 2017). A standardised 3 min warm up was conducted prior to every training session, after which participants were given one minute of passive rest on the cycle ergometer. The training protocol was based on a previous method (Wisloff et al. 2007) and comprised four bouts of 4 min exercise intervals performed at an intensity corresponding to 90–95% of the HR peak; each interval was interspersed by 4 min of active recovery where participants cycled at 50–60% peak HR. Each training session was followed by a 5 min cool down, and all training sessions were fully supervised.

Resting muscle biopsy

A resting muscle biopsy was obtained from each of the young and older participants. For the older participants participating in HIT, the initial biopsy was prior to the commencement of the training programme (Pre) and the final biopsy was conducted 24–48 h following the final training session (Post). After an injection of a local anaesthetic into the skin and fascia (xylocaine 1%, AstraZeneca, Australia) a small incision was made in the vastus lateralis muscle and a sample was taken using a biopsy needle with suction (Bergstrom, 1975). The muscle was blotted on filter paper to remove excess blood and immediately frozen in liquid nitrogen and stored at −80°C until analysis.

Single fibre collection, fibre‐typing and pooling

Segments of single fibres were collected from either fresh muscle or freeze‐dried muscle, as previously described (Murphy, 2011; MacInnis et al. 2017; Wyckelsma et al. 2017) and placed in 10 μl of 3× SDS denaturing solution (0.125 m Tris‐HCI, 10% glycerol, 4% SDS, 4 m urea, 10% 2‐mercaptoethanol and 0.001% Bromophenol Blue, pH 6.8, diluted 2:1 with 1× Tris^.Cl (pH 6.8)). Fibres remained at room temperature for 1 h post collection to aid solubilisation and were thereafter stored at −80°C until dot blotting.

Each single fibre was fibre‐typed using the dot blotting method (MacInnis et al. 2017). Briefly a PVDF membrane was activated in 96% ethanol for 120 s and equilibrated in transfer buffer for 120 s. Following this, 1 μl (10%) of each fibre was spotted on the PVDF membrane. After drying, membranes were re‐activated in 96% ethanol and re‐ equilibrated in transfer buffer (∼120 s). Membranes were then washed in 1× Tris‐buffered saline–Tween (TSBT) for 5 min, blocked for 5 min in 5% skim milk powder in 1× TBST. The membrane was then incubated in primary antibody overnight at 4°C and 2 h at room temperature. After washing and incubating with a secondary antibody and TBST washes, the membrane was coated with chemiluminescent substrate (ECL, Clarity, Bio‐Rad, USA) and imaged using a Chemidoc MP and Image Lab software (Bio‐Rad).

For dot blotting MHCIIa was always identified first (MHCIIa, mouse, monoclonal IgG, A4.74, Developmental Studies Hybridoma Bank (DSHB), 1:200 in 1% bovine serum albumin (BSA) in phosphate‐buffered saline with 0.025% Tween and 0.02% NaN₃ (PBST)). Following imaging of MHCIIa, membranes were stripped for 30 min at 38°C using stripping buffer (Pierce), washed for 5 min in 1× TBST and then incubated in MHCI antibody (MHCI, mouse, monoclonal IgM, A4.840, DSHB, 1:200 in 1% BSA in PBST). Fibres were identified as type II (MHCIIa), type I (MHCI) or hybrid (MHCII and MHCI). Only fibres which identified as type I or II were kept for further analyses.

The remaining 9 μl of each fibre was then pooled with other fibres of the same fibre type from a given participant for each time point. Each pool consisted of 4–11 fibres and was stored at −80°C until Western blotting. Between 85–103 type I fibres and 85–99 type II fibres were analysed from the young cohort (N = 9 participants); from the older adults, between 91–98 type I fibres and 76–101 type II fibres (N = 9 participants) were assessed for COXIV, NDUFA9 and Mfn2.

Whole muscle homogenate preparation

A whole muscle homogenate was prepared as described earlier (Murphy et al. 2006). Briefly, a small portion of whole muscle (15–30 mg), was accurately weighed and homogenised on ice (1:20 w/v) in Na‐EGTA solution (165 mm Na⁺, 50 mm EGTA, 90 mm Hepes, 1 mm free Mg²⁺ (10.3 mm total Mg²⁺), 8 mm total ATP, 10 mm creatine phosphate, pH 7.10) with a protease inhibitor cocktail (PIC, Complete; Roche Diagnostics, Sydney, Australia). Immediately following this, the homogenate was diluted to 33 μg wet weight muscle μl⁻¹ using 3× SDS denaturing solution. Finally, samples were further diluted to 2.5 μg wet weight muscle μl⁻¹ with 1× SDS solution (3× SDS denaturing solution diluted 2:1 with 1× Tris^.Cl (pH 6.8)). A small amount of homogenate from all samples from the older participants (n = 16, Pre and Post samples) was mixed together to generate a suitable sample for 4–5 point calibration curves, spanning 3–60 μg of total muscle wet weight.

For fractionation experiments, 5 μl from each whole muscle homogenate sample from young and old Pre subjects was collected and 1% Triton X‐100 added The samples were kept on ice for ∼60 min then spun at 14,000 × g for 15 min at 4°C. Both the pellet and supernatant were collected, with the pellet suspended in 1× solubilizing buffer and 3× solubilizing buffer added to supernatant, and samples were run side by side on a 4–15% TGX Stain‐Free Gel (details below).

Western blotting

Western blots were performed to determine the abundance of mitochondrial proteins COXIV, NDUFA9, Mfn2 and MiD49 in whole muscle homogenates and pools of single fibres using techniques previously described (Murphy, 2011; Wyckelsma et al. 2015; MacInnis et al. 2017). Briefly, denatured protein in the whole muscle samples and calibration curve samples were separated on 26 well, 10% or 4–15% Criterion TGX Stain‐Free gels (Bio‐Rad, Hercules, CA, USA) and run for 45 min at 200 V. The total protein loaded was visualised using UV activation of the gel and analysed with Image Lab 5.2.1 (Bio‐Rad). Using a wet transfer protocol, protein was transferred to nitrocellulose membrane at 100 V for 30 min. Membranes were incubated in Pierce Miser solution (Pierce, Rockford, IL, USA) and blocked in 5% skimmed milk powder in TBST. Membranes were cut horizontally and each section incubated with antibodies diluted in 1% BSA in PBST (with 0.025% Tween and 0.02% NaN₃). Details of antibodies used are cytochrome oxidase IV (COXIV, Cell Signalling, no. 4844, 1:1000, rabbit, polyclonal), NADH:ubiquinone oxidoreductase (complex I) (NDUFA9, 1:1000, rabbit), Mfn2 (1:500, rabbit), MiD49 (rabbit, 1:500, polyclonal, see Richter et al. 2014, Osellame et al. 2016). Relative to other tissues, MiD51 abundance was lower in muscle (not shown) and so we focused on MiD49 in our muscle analyses, as this is likely to be the most relevant mitochondrial adapter protein in muscle. Membranes were incubated overnight at 4°C and 2 h at room temperature, all with rocking. After washing and incubating with a secondary antibody and following TBST washes, the membrane was coated with chemiluminescent substrate (West Femto, ThermoScientific, IL, USA). Images were acquired using Image Lab software (Bio‐Rad). The positions of molecular mass markers were captured under white light prior to chemiluminescent imaging without moving the membrane.

For data analysis, the density of a given protein was obtained for each muscle sample and expressed relative to the calibration curve and then normalised to the total protein of their respective lane, which was also expressed relative to its standard curve. The same calibration curve was used on all gels and data are expressed relative to the average of a given sample type (i.e. young adult in whole muscle analyses) or sample type (Pre type I in HIT analyses).

Citrate synthase activity assay

For determination of citrate synthase (CS) activity, muscle was accurately weighed (15–25 mg) and homogenised (5 × 5 s, with 5 min on ice between bursts) at 20:1 in buffer containing 70 mm sucrose, 220 mm d‐mannitol, 10 mm Hepes (pH 7.4), 1 mm EGTA. Measurements were made in whole muscle preparations in duplicate or triplicate. The following were placed into a reference cuvette: 825 μl 0.1 m Tris buffer, 100 μl 5′5‐dithiobis (2‐nitrobenzoic acid) (DNTB, 0.5 mg ml⁻¹ made in Tris buffer) and 10 μl acetyl‐coA (6 mg ml⁻¹ made in Tris buffer). Into individual cuvettes containing the same components as the reference cuvette were added 15 μ of homogenate and 50 μl oxaloacetate (6.1 mg ml⁻¹ made in Tris buffer) to initiate the reaction. The cuvette was placed in a spectrophotometer (LKN Novaspec II) and the machine was zeroed at 412 nm. Absorbance at 412 nm was recorded every 15 s for 150 s. The change in absorbance readings were plotted against time (15 s) and linear regression used to determine the slope of the response. The slope between 30 and 90 s was used to calculate CS activity (presented as μmol min⁻¹ g⁻¹). Analyses were conducted on muscle from 5 young and 7 older adults.

Blue native PAGE

For comparison of muscle samples collected from young and older adults, samples were cryo‐sectioned and placed in Na‐EGTA solution (described in the Whole muscle homogenate section) with 1% Triton‐X100 (20 × 10 μm sections in 100 μl buffer). For comparison of Pre and Post HIT samples, muscle homogenates prepared for citrate synthase (described in CS activity) were used with the addition of 1% Triton‐X100. Ten microlitres of each sample were diluted 1:1 in solubilising solution (20 mm Bis‐Tris pH 7.0, 50 mm NaCI, 10% glycerol, 0.5% Coomassie) with 2 μl of loading dye (5% Coomassie Blue G, 500 mm ε‐amino n‐caproic acid, 100 mm Bis‐Tris pH 7.0). Samples were loaded onto 4–13% homemade (McKenzie et al. 2007) or native polyacrylamide gels (Novex, Invitrogen), along with a calibration curve of mixed samples. Gels were run for 10 min at 100 V, 10 min at 400 V and 45 min at 400 V, with a change of buffer between the second and third run. Following electrophoresis, protein was transferred to PVDF membrane for 2 h at 100 V. Membranes were stained (0.02% Coomassie Blue G, 50% methanol (MeOH), 10% acetic acid (HAc)), destained (50% MeOH, 10% HAc) and blocked for 2 h using 5% skimmed milk in TBST. Following blocking, the OXPHOS antibody cocktail (Mouse, Abcam, 1:1000) was applied with subsequent methodology as described for Western blotting. Following imaging the remaining portions of the samples were diluted 1:2 in 3× SDS loading buffer and run on a 4–15% Stain‐Free denaturing gel for confirmation of total protein content. Analyses was performed on muscle from 8 young and 8 younger adults.

Statistical analysis

Fibre‐type specificity across both age groups and the age effect within a particular fibre type for a given protein were analysed using a one‐way ANOVA. For the HIT undertaken by the older participants, a two‐tailed Student's paired t test was used to compare LBM, FM, percentage fat, peak HR, peak work rate, ${\dot{V}}_{{\mathrm{O}}_2,\mathrm{peak}}$, mitochondrial protein abundance in homogenates and pools of single fibres, complex abundances from BN‐PAGE and citrate synthase. Statistics were conducted using GraphPad Prism version 7. All data are presented as means ± SD. Significance was set as P < 0.05.

Results

Mitochondrial protein abundance following fractionation

When examining pellet and supernatant fractions side by side on Western blotting, the amount of mitochondrial proteins retrieved in the supernatant was 60–90% (Fig. 1 A–D). The amounts of Mfn2 and NDUFA9 retrieved in the supernatant tended to be lower in muscle from older compared with young adults (P = 0.06 and P = 0.13, respectively, two‐tailed t tests). In muscle from young adults, the amount of COXIV retrieved from the supernatant was lower than Mfn2 (P < 0.05, two‐tailed t test).

Figure 1. Fractionation of tisue results in a large proportion of the total mitochondrial pool that is not measured.

Human skeletal muscle was homogenised and a portion kept as whole muscle homogenate (W) and a seperate portion centrifuged (14,000 × g, 15 min, 4°C), following which the pellet (P) and supernatant (S) were collected. The P was suspended in the same volume as supernatant and then S, P and W separated on 4–15% Stain‐Free gel and probed for the mitochondrial proteins NADH:ubiquinone oxidoreductase (complex I) (NDUFA9), COXIV and mitofusin 2 (Mfn2). The proportion of each protein present in the supernanant following this fractionation in samples from young and older adults is shown: A, NDUFA9 B, COXIV; C, Mfn2. D, representative blot showing NDUFA9, COXIV and Mfn2 in the fractions. Significant differences indicated by brackets in A–C (P values , Student's unpaired t test). Means ± SD indicated. [Color figure can be viewed at wileyonlinelibrary.com]

Anthropometric and exercise characteristics

There were no differences in body mass, body mass index (BMI) or lean body mass in young and older adults (Table 1). Older adults had a higher fat mass (kg) than young adults (P < 0.05) and tended to have a higher percentage of body fat (P = 0.098, Table 1). Older adults had a lower ${\dot{V}}_{{\mathrm{O}}_2,\mathrm{peak}}$, peak power and higher heart rate than young adults (P < 0.05, Table 1).

Table 1.

Physical characteristics and performance measures of healthy older compared to healthy young adults

Characteristic	Older adults	Young
Body mass (kg)	75.1 ± 12.8	72.8 ± 15.6
BMI (kg m−2)	21.8 ± 2.6	20.8 ± 3.1
Whole body lean mass (kg)	46.9 ± 10.4	48.0 ± 13.5
Whole body fat mass (kg)	24.6 ± 5.0*	19.2 ± 7.1
Whole body fat percentage (%)	33.6 ± 6.0†	28.1 ± 9.6
HRpeak (beats min−1)	138.2 ± 13.0*	172.2 ± 12.7
Peak power (W)	143.5 ± 46.7*	226.6 ± 66.1
${\dot{V}}_{{\mathrm{O}}_2,\mathrm{peak}-\mathrm{RPE}17}$ (ml kg−1 min−1)	23.6 ± 4.9*	36.6 ± 6.5

All data expressed as means ± SD. ^*Significantly different to young P < 0.05, ^†tended to be different to young: P = 0.098. N = 15 young adults and N = 17 older adults for all measures except whole body lean mass, fat mass, fat percentage where N = 8 for young and N = 16 for older.

Mitochondrial characteristics of young and older adults

When measured in whole muscle preparations from young compared with older adults the mitochondrial markers cytochrome oxidase IV (COXIV) and NDUFA9 were unchanged (P > 0.05), whereas, Mfn2 and MiD49 were 1.5‐ to 2.5‐fold greater in older compared with young adults (P < 0.05, Fig. 2).

Figure 2. Mitochondrial proteins in whole skeletal muscle homogenates from young and older adults.

A, Western blots for Mfn2 and NDUFA9 on the left and MiD49 and COXIV on the right, showing total muscle from young (Y) and older (O, Old) adults. Total protein indicated by the abundant muscle protein, myosin, shown in Stain‐Free gels. B, the content of NDUFA9, COXIV, MiD49 and Mfn2 abundances were expressed relative to total protein and the average of young samples present on the same gel. Brackets indicate P < 0.05 (two‐tailed unpaired Student's t test, N = 9 young, N = 8 older adults). [Color figure can be viewed at wileyonlinelibrary.com]

There was a greater abundance of both COXIV and NDUFA9 in type I compared to type II fibres in muscle from young and older adults (P < 0.05, Fig. 3). There was no effect of age on the abundance of COXIV or NDUFA9 proteins in either type I or type II fibres (P > 0.05, Fig. 3). Compared to the young group, older adults had a greater content of Mfn2 in type II fibres (P < 0.05) with no difference in type I fibres (Fig. 3).

Figure 3. Fibre type characteristics of mitochondrial proteins in type I and II fibres from young and older adults.

A, Stain‐Free gel and Mfn2, NDUFA9 and COXIV detected by Western blot. Fibres are classified as either a type I or II based on the presence of myosin heavy chain (MHC) isoform I or IIa present as indicated in the Western blot, with samples from young (Y) and older (O) adults. Plots show data for individual proteins: B, Mfn2, C, NDUFA9 D, COXIV. Symbols of the same colour across the graphs indicate the pooled data from fibres collected from a given participant. Type I fibre pools are denoted with circles and type II fibre pools denoted with triangles. Representative blots are from subjects depicted in black (Y) and light blue (O). Brackets indicate P < 0.05 as shown (one‐way ANOVA with uncorrected Fisher's LSD test). Means ± SD indicated. [Color figure can be viewed at wileyonlinelibrary.com]

Assessment of the mitochondrial complexes in mixed muscle samples in their native states using BN‐PAGE found there were no differences in the abundances of mitochondrial complexes I–V in whole muscle homogenates from young and older adults (P > 0.05, Fig. 4).

Figure 4. Age has no effect on mitochondrial complex contents.

A, representative blot of Blue Native PAGE immunoblotted using the OXPHOS cocktail antibody. B, averaged data from individuals are shown for each of the mitochondrial complexes. On average all samples were run on a minimum of two separate gels, in some instances samples on three different gels. Each band was analysed and normalised to the total protein (Stain‐Free gel). Y represents samples from young adults, O represents samples collected from older adults. Data are presented as means ± SD. [Color figure can be viewed at wileyonlinelibrary.com]

There was no difference in whole muscle citrate synthase activity between young and older adults (6.8 ± 1.4 vs. 6.4 ± 1.9 μmol (g wet weight)⁻¹, respectively, N = 5 young, N = 7 older Pre adults, P > 0.05, Fig. 5).

Figure 5. Citrate synthase activity is similar in muscle form young and older adults, and is increased after high‐intensity training in older adults.

Each colour used for circles (Pre) and squares (Post) on the graph indicates the same participant (consistent across all figures). Bracket indicates P < 0.05 from Pre (one‐way ANOVA, Sidak's multiple comparisons test). Data expressed as means ± SD. [Color figure can be viewed at wileyonlinelibrary.com]

Effects of HIT on older adults on mitochondrial content

HIT increased the abundances of COXIV (26%, P < 0.05) and NDUFA9 (87%, P < 0.05), but had no effect on MiD49 or Mfn2 in whole muscle homogenates (P > 0.05, Fig. 6). When examined in pooled fibres, there were increases after HIT in COXIV in both type I and II fibres and NDUFA9 in type II fibres; Mfn2 decreased in type II fibres (P < 0.05, Fig. 7). BN‐PAGE analyses revealed that HIT increased the abundances of mitochondrial complexes I, II, IV and V (P < 0.05), but not complex III (P > 0.05, Fig. 8). Citrate synthase activity increased ∼55% following HIT in skeletal muscle from the older adults (P < 0.05, Fig. 5).

Figure 6. Mitochondrial abundance is increased with high‐intensity training in older humans.

Each panel shows the representative Western blot and pooled data for Mfn2 (A), MiD49 (B), NDUFA9 (C) and COXIV (D) in human skeletal muscle obtained before (Pre, P) and following (Post, Pt) 12 weeks of high‐intensity interval training (HIT) in older adults. In each panel, the Stain‐Free image indicates the relative amount of tissue loaded in each lane. Calibration curves, loaded with 3–31 μg of protein as indicated, are shown in each panel and were run on all gels analysed. The curves shown are from the same Western blot at the same exposure, with non‐contiguous lanes shown with spaces. Each protein was normalised to the total protein (Stain‐Free gel) and then to the calibration curve, and expressed as relative abundance. Each colour used for Pre and Post symbols on the graphs indicates the same participant (consistent across all figures). Brackets indicate P < 0.05 (Student's paired t test, means ± SD, N = 8). [Color figure can be viewed at wileyonlinelibrary.com]

Figure 7. Mitochondrial abundance in type I or type II muscle fibres before and after 12 weeks of high‐intensity training in older adults.

A–C, representative Western blots of type I and type II fibre pools for COXIV (A), NDUFA9 (B), Mfn2 (C) from human skeletal muscle obtained before (Pre, P) and following (Post, Pt) 12 weeks of high‐intensity interval training (HIT) in older adults. Fibres were classified as either a type I or II based on the MHC isoform (MHC I or MHC IIa) present, as indicated. In each panel, the Stain‐Free image indicates the relative amount of tissue loaded in each lane. Calibration curves, loaded with 3–60 μg of protein as indicated are shown in each panel and were run on all gels analysed. D–I, pooled type I and II fibres for COXIV (D and G, respectively), NDUFA9 (E and H, respectively) and Mfn2 (F and I, respectively). Each colour used for Pre and Post symbols on the graphs indicates the same participant (consistent across all figures), N indicates number of participants, n = indicates number of fibres collected and analysed in fibre pools. Each protein was normalised to the total protein (Stain‐Free gel) and then to the calibration curve, and expressed relative to the type I Pre samples. Brackets indicate P < 0.05 (Student's paired t test, means ± SD). [Color figure can be viewed at wileyonlinelibrary.com]

Figure 8. Mitochondrial complexes I, II, IV and V increase with high‐intensity training in the older adults.

A, representative Western blot of Blue Native PAGE gel showing a Pre and Post whole muscle sample and a calibration curve, probed with anti‐OXPHOS cocktail antibody (see Methods). B, averaged data from each individual are shown for each of the mitochondrial complexes from before (Pre) and after (Post) 12 weeks of High‐intensity interval training. Density of each band normalised to the total protein taken from the denaturing Stain‐Free gel shown at top. Each colour used for Pre and Post symbols on the graphs indicates the same participant (consistent across all figures). Brackets indicate P < 0.05, (Student's paired t test, means ± SD, N = 8). [Color figure can be viewed at wileyonlinelibrary.com]

Discussion

Utilising whole muscle and single fibre segments to analyse mitochondrial content

A significant feature of our biochemical approach is that we only used whole muscle samples, either homogenates (i.e. no centrifugation/purification steps) or segments of individual muscle fibres. Our data demonstrates that the relative yields of various mitochondrial proteins differ when muscle is homogenised in a physiologically based buffer with 1% Triton‐X100 and subsequent fractionation. It is clearly important to measure the entire pool of mitochondria when making any quantitative assessment, as 10–40% of the mitochondria are not present in supernatant, probably remaining in the pellet which is typically discarded from muscle samples in mitochondrial research. Furthermore there is considerable variability in yield between individual samples, which are not typically taken into consideration.

The yield of Mfn2 tended to be lower (P = 0.06) in muscle from older compared with young adults. Studies using fractionated muscle samples show no difference in the abundance of Mfn2 in skeletal muscle between young and older adults (Joseph et al. 2012; Konopka et al. 2014; Distefano et al. 2017) and it is possible that a difference in yield of Mfn2 between the muscles being compared confounded those findings. Some general problems of fractionation are discussed in detail elsewhere (Murphy & Lamb, 2013). Pertinent to the work here, it has been shown that different pools of mitochondria in skeletal muscle, specifically the paravascular, or subsarcolemmal, mitochondria and the intermyofibrillar mitochondria display differing amounts of mitochondrial complexes IV and V (Glancy et al. 2015). Indeed, we report a significant difference in the yield of two mitochondrial proteins, COXIV and Mfn2, in muscle from young adults, namely 74% (COXIV) compared with 91% (Mfn2), indicating the necessity to analyse entire muscle homogenates when analysing mitochondrial proteins. This is likely to extend to other proteins, and we have previously shown substantial loss of sarcoplasmic reticulum proteins during fractionation (Murphy et al. 2011). Thus whenever fractionation is undertaken, the final protein yield should be ascertained, but the better approach is to simply use preparations incorporating the whole sample.

Effects of age and fibre type on mitochondrial content

Using a suite of markers for mitochondrial content – COXIV and NDUFA9 protein contents, citrate synthase activity and mitochondrial complex abundances in native states, –we found that age does not have an influence on skeletal muscle mitochondrial content. Our results support the recent finding that mitochondrial density, assessed by electron microscopy, was not different in skeletal muscle samples from similarly aged groups (Petersen et al. 2015). Whilst, electron microscopy is regarded as the gold standard for measuring mitochondrial content, citrate synthase activity has a close association with mitochondrial content (Larsen et al. 2012). Additionally, complex IV activity has been shown to be strongly associated with mitochondrial oxidative phosphorylation capacity (Larsen et al. 2012). Whilst not measured her, the reporting of COXIV abundance by Western blotting and the abundance of the intact complex IV using BN‐PAGE provides insight into complex IV abundance which may be related to activity. Our finding of no difference in COXIV abundance is similar to that found in two cohorts of older individuals, ∼65 years and 83 years, although in that study the tissue was fractionated (Spendiff et al. 2016). Given the heterogeneity of human vastus lateralis muscle, we examined mitochondrial content using Western blotting of COXIV and NDUFA9 in type I and type II muscle fibre pools from both young and older adults. Similar to a previous report (MacInnis et al. 2017), which found that COXIV and NDUFA9 were more abundant in type I compared with type II fibres from young adults. We extend this finding by showing that there is a similar fibre type dependence of these mitochondrial proteins when muscle fibres are isolated from older adults, and this was a similar ∼40% difference between the fibre types. In line with the whole muscle homogenate data, there was no difference in the abundances of COXIV or NDUFA9 in either type I or type II fibres between young and older adults. Together, the current findings indicate that when expressed relative to the total protein present in whole muscle skeletal muscle homogenates, there is a similar mitochondrial content in young and older (aged 65–76) adults. It is also important to note that the older adults were healthy and recreationally active, as it has been suggested that inactivity may have a more important deleterious effect on mitochondrial content and function than age (Gram et al. 2014, 2015; St‐Jean‐Pelletier et al. 2016); comparison of a less active cohort of older adults, or octogenarians, may yield different results. This would not be surprising, as mitochondria demonstrate clear plasticity in response to exercise training and aerobic capacity (MacInnis et al. 2017).

Effects of age on mitochondrial dynamics

We investigated two key mitochondrial dynamics proteins in whole muscle homogenates and report an increased abundance of both the MiD49 and Mfn2 proteins in muscle from older compared with young adults. Previous research examining Mfn2 protein content found no differences between muscle from young and older adults (Joseph et al. 2012; Konopka et al. 2014; Distefano et al. 2017). In two cases the entire mitochondrial pools were not analysed (Konopka et al. 2014; Distefano et al. 2017). Those studies also looked at fission proteins, Fis1 and/or dynamin related protein 1 (Drp1 or Dnm1L), and found these to be unchanged in muscle from young and older adults (Joseph et al. 2012; Konopka et al. 2014; Distefano et al. 2017). Fis1 has been discovered to not be required for mitochondrial fission (Osellame et al. 2016), while Drp1 is involved in division of inner and outer mitochondrial membranes (Otera & Mihara, 2011), it also regulates fission of peroxisomes, making it difficult to discern a mitochondrial specific function of Drp1. Here we have been able to report that the recently identified protein MiD49, which is a known mitochondrial specific fusion protein (Richter et al. 2014), is more abundant in muscle from older compared with younger adults. Interestingly, there is no difference in the abundance of a related protein, MiD51 in muscle from young and older adults (data not shown). The MiD49 data, combined with the findings on Mfn2, indicate that mitochondria may have an increased capacity to undergo fusion and fission events. A previous study reported a lower level of the mitochondrial‐targeted ubiquitin ligase Parkin in aged muscle, albeit in fractionated tissue (Gouspillou et al. 2014). It is possible that the increase in fusion and fission proteins we report help to overcome the problems of removing damaged mitochondria secondary to the reported reduction in Parkin protein abundance.

Our data show that older adults aged 65–76 years experience a change in mitochondrial dynamics that occurs without a concomitant change in mitochondrial content. We hypothesise that the dysfunction in skeletal muscle mitochondria observed with ageing results in an increase in Mfn2 and MiD49 proteins as a compensatory response to ageing, as previously suggested (Chan, 2006). It needs to be determined if such compensatory responses are maintained in very old adults, such as octogenarians (80–89 years) where normalised power of MHCIIa fibres has been shown to be 63% and 39% higher than in 20 and 80 year old adults, respectively (Grosicki et al. 2016).

Effects of fibre type in mitochondrial dynamics with age

The higher Mfn2 protein in older adults compared with young adults was isolated to type II muscle fibres, suggesting a fibre type specific higher turnover of mitochondria with ageing. Type II muscle fibres are specifically weakened by ageing (Lamboley et al. 2015), have greater reactive oxygen species (ROS) generation (Anderson & Neufer, 2006) and show larger age‐related ATP depletion and uncoupling (Conley et al. 2007), all of which suggests specific mitochondrial dysfunction in type II fibres in muscle from older adults. We speculate that in muscle from older adults up to 76 years, upregulation of Mfn2 may negate the ageing effects within type II fibres thus promoting healthier mitochondria in fibres susceptible to age‐induced weakness. MiD49 wasn't analysed in specific fibre pools due to a reduced efficacy of the antibody in these smaller samples and future work is required to determine whether fibre type differences in this protein simulate those seen for Mfn2.

Effects of HIT on skeletal muscle mitochondria in older adults

As reported previously, the exercise protocol employed in our study improved ${\dot{V}}_{{\mathrm{O}}_2,\mathrm{peak}}$, peak power and increased time to reach RPE 17 during an incremental exercise test and showed a tendency to decrease resting systolic blood pressure (Wyckelsma et al. 2017). Accompanying these physiological adaptations, we now report that HIT increases protein content of COXIV (∼37%) and NDUFA9 (∼48%) and CS activity (55%) in older adults aged 65–76 years old. In these individuals, we did not find a correlation between the protein contents and age (data not shown). The examination of mitochondrial complexes in their native states using BN‐PAGE provides a means to ascertain whether the complexes are likely to be assembled correctly. The current study demonstrates that the abundances of the mitochondrial complexes were also upregulated. Overall, these adaptations are similar to those shown by young adults who underwent intensified single leg HIT training and demonstrated a 24% increase in COXIV, an 11% increase in NDUFA9 and a 39% increase in CS activity (MacInnis et al. 2017). The current study supports other research demonstrating that skeletal muscle of slightly older adults (average age 69 years in current study vs. 73 years in other studies) is highly malleable following exercise training (Konopka et al. 2014; Joseph et al. 2016) and thus exercise is an important and powerful intervention in improving muscle health and function with training. Of note, the study by Joseph et al. (2016) was a cross‐sectional study comparing high‐functioning adults aged 75 ± 1 years and low‐functioning adults aged 81 ± 1 years. Our study provides a unique insight into mitochondrial adaption to HIT in older adults. Our results support recent findings where 24 month old mice who completed acceleration‐based training had greater adaptation in muscle performance, CS and lactate dehydrogenase activity than endurance‐trained mice (Niel et al. 2016). However, another important consideration is the age of these mice, as rats aged 34–36 months show blunted responses to aerobic training (Betik et al. 2009), and therefore it is possible that training responses may be blunted in very old adults, compared to the 65–76 year olds recruited in the current study, although this remains to be determined.

This is the first study to show fibre type specific mitochondrial adaption from HIT in older adults. NDUFA9, the marker of complex I, was upregulated specifically in type II fibres following HIT. In rats, intense exercise increased mitochondrial respiration in tibialis anterior compared to soleus (Ramos‐Filho et al. 2015). The specific upregulation of NDUFA9 in type II fibres may be related to the exercise employed which may recruit more type II fibres (Kristensen et al. 2015), at least in comparison to regular every day activities of the older adults.

Effects of HIT on mitochondrial dynamics

Using whole muscle preparations, despite observing increases in mitochondria abundance, we found no differences in the contents of Mfn2 or MiD49 following HIT. Interestingly, in young adults, Mfn2 protein content increased in skeletal muscle following HIT by 16% (MacInnis et al. 2017). Endurance training in both younger (20 ± 1 years) and older (74 ± 3 years) adults has also been shown to increase Mfn2 protein content (Konopka et al. 2014), although in that study muscle underwent a light centrifugation step to remove cellular debris. When examined in specific fibre types, we found a decrease in Mfn2 protein content in type II fibres following our HIT intervention in the elderly. The decline in Mfn2 in type II fibres following HIT may suggest the mitochondria are becoming more functional with the exercise training and therefore the requirement for increased Mfn2 protein content seen between young and older adults is lower, indicative of a decreased need for mitochondrial turnover.

Conclusions

Ageing has no effect on the amount of mitochondria present in skeletal muscle cells, which show large increases following HIT. Conversely, mitochondrial dynamics proteins are upregulated with ageing but are unaltered following HIT. Mitochondria appear to be combating age‐associated weakness by promoting mitochondrial turnover, whereas HIT in the elderly appears to reduce the requirement for greater abundance of dynamics proteins.

Additional information

Competing interests

The authors declare no competing interests.

Author contributions

All human experimental trials and exercise training was completed at Victoria University. All muscle biochemistry experiments were conducted at La Trobe and Monash Universities. V.L.W., I.L., A.C.P., M.J.M. and R.M.M. contributed to the conception and design of work. V.L.W., I.L., A.C.P., L.E.F., M.T.R., M.J.M., M.J.A. and R.M.M. collected and analysed data. V.L.W., I.L., A.C.P., M.J.M., L.E.F., M.T.R., M.J.A. and R.M.M. drafted and revised the manuscript critically for important intellectual content. V.L.W., I.L., A.C.P., M.J.M., L.E.F., M.T.R., M.J.A. and R.M.M. 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. V.L.W., I.L., A.C.P., M.J.M., L.E.F., M.T.R., M.J.A. and R.M.M. approved the final version of the manuscript and qualify for authorship. All those who qualify for authorship are listed.

Funding

I.L. was supported by Future Leader Fellowship (ID: 100040) from the National Heart Foundation of Australia. A.C.P. was supported by the Australian Government Collaborative Research Networks program. Various biochemical analyses were supported by grants from the Understanding Diseases and Sport, Exercise and Rehabilitation Research Focus Areas of La Trobe University.