One-legged dynamic knee-extension exercise (DKE) has been one of the most used models to study the effects of acute exercise on human skeletal muscle since its introduction in mid 1980s’ (Andersen et al., 1985). Early electromyography studies suggested that DKE causes quite even activation of the superficial quadriceps femoris (QF) muscles with no or minor activation in other leg muscles (Andersen et al., 1985). Later, it was shown with magnetic resonance imaging (MRI) that other than QF muscles remain fairly inactive during DKE even with relatively high exercise intensity (90% of VO2max; Richardson et al., 1998). On the other hand, it was shown that the different QF muscles are perhaps not so evenly activated during DKE as previously anticipated (Richardson et al., 1998). In a more recent series of experiments, it has been shown by use of measurements of muscle heat accumulation in the early phase of exercise (within the first 90 s), that all QF muscles are relatively equally activated during intense DKE (Gonzalez-Alonso et al., 2000; Krustrup et al., 2004, 2009), but not during moderate 50% intensity of max workload (Krustrup et al., 2009).

The findings of considerable heterogeneity in muscle blood flow during DKE measured with positron emission tomography (PET) conforms with variable muscle activation determined by heat production (Laaksonen et al., 2003; Heinonen et al., 2007). On the other hand, the rectus femoris (RF) was the most activated muscle in the study by Krustrup et al. (2009), whereas blood flow has usually been found to be the highest in vastus intermedius (VI) or medialis (VM; Laaksonen et al., 2006). This discrepant finding can partly be explained by the different time frame in the studies as Krustrup and colleagues studied the early response after commencing the exercise, while muscle blood flow has been measured during more prolonged steady-state exercise. However, further studies are warranted to more clearly elucidate how tightly muscle metabolism and blood flow are linked.

Glucose uptake measured with PET and [¹⁸F]-fluoro-deoxy-glucose ([¹⁸F]-FDG) has been recently applied to provide an estimate for metabolic activity and thus muscle activation in humans (Fujimoto et al., 1996; Tashiro et al., 1999; Pappas et al., 2001; Oi et al., 2003; Bojsen-Moller et al., 2006; Kalliokoski et al., 2007). Due to the fact that this particular tracer accumulates into the tissues where it has been taken up, it can be administered before PET scanning, for example during exercise. PET scanning can then be performed after the exercise and the obtained images show the tracer accumulation during the exercise (Figure 1). Thus, this method provides an excellent way to investigate the activation of different muscles during long-term steady-state DKE.

In the present study we hypothesized that individual muscles of QF are not evenly activated during long-term steady-state moderate-intensity DKE. To test this we measured [18F]-FDG uptake with PET in five healthy men during DKE.

On average glucose uptake was highest during exercise in VI and lowest in VL muscle but the individual variation was large (Figure 2). In the resting muscle glucose uptake was also highest in VI but lowest in RF (Figure 2). The coefficient of variation in the four resting muscles was 21 ± 3% and it increased (p = 0.044) to 35 ± 9% between exercising muscles showing heterogeneous activation of QF muscles. The same phenomenon from another point of view was observed in the relative increases in glucose uptake caused by exercise (11.5-fold in VL, 15.1-fold in RF, 14.6-fold in VM, and 18.4-fold in VI). Only the difference between VL and VI was significant (p = 0.04).

The largest variation in glucose uptake between the subjects in the exercising muscle was observed in the VL (69%) while in the other three muscles variation was almost the same (38% in RF, 38% in VM, and 36% in VI) as in the entire QF (41%). Compared to the mean glucose uptake in QF (=100%), glucose uptake was similar in RF (84% of the QF value, VM 115%, but in VL (73%) and in VI (142%) it differed significantly.

Other than QF muscles were also variably activated between the subjects. In the posterior thigh region biceps femoris was markedly activated in four out of the five subjects (Table 1; Figure 1). Some activation was also seen in the sartorius muscle as seen in the posterio-medial side of the thigh in Figure 1. Also, lower leg muscles were metabolically active. Three of the subjects activated the tibialis anterior muscle (Table 1). The same subjects had also increased activation in the popliteus muscle (Table 1). Soleus muscle was not significantly activated in any of the subjects, but there was increased activation of lateral gastrocnemius muscle in all subjects and of medial gastrocnemius muscle in two subjects.

Since its introduction in the mid-eighties the one-legged DKE model has been widely applied in different settings to study acute and chronic responses to exercise of an isolated skeletal muscle group (Andersen et al., 1985). It has been assumed that the DKE model activates all QF muscles with limited contribution from other leg and thigh muscles. In the present study we used FDG–PET to study activation of QF and other thigh and leg muscles during this same exercise model and found several interesting novel findings. First, QF muscles have an uneven glucose uptake during moderate-intensity long-term steady-state one-leg DKE. Second, the pattern in QF muscles differs significantly between subjects and the variation between the subjects is largest in VL. Third, glucose uptake in RF and VM is closest to the mean glucose uptake of whole QF, while glucose uptake in VL is lower than the value for the whole QF. Finally, variable activation was also observed during knee-extension in the posterior thigh and lower leg muscles. These findings are in concordance with the recent findings by Krustrup et al. (2009) and emphasize that metabolism in the leg muscles is not homogenous during knee-extension exercise.

The discrete differences in the patterns of metabolic activity in RF between the study of Krustrup et al. (2009) likely stem from differences in muscle recruitment patterns over time. They studied the early response after commencing exercise, while we studied the activation during a longer period of steady-state exercise (cumulative activity during the period of 10–35 min of steady-state exercise).

It has generally been assumed that muscle glucose uptake distribution among QF muscles varies in parallel with blood flow distribution during exercise (Laaksonen et al., 2003; Heinonen et al., 2007). However, there was large inter-individual variation in the activation pattern (see Figure 2). In two subjects (S3 and S4), VL was minimally activated compared to VM and VI, while one (S2) had almost even activation of VL, VM, and VI. Subjects 1 and 5 had more similar metabolic activation of all muscles. On average the largest increase from rest to exercise was observed in VI and the smallest was observed in VL, which displayed large individual variation (3- to 26-fold-increase). The implications of this study are that local measures of metabolism of VL muscle for example, from biopsy analysis may not be representative of the whole muscle. Although we measured only glucose uptake and nothing is known regarding the use of other substrates (blood FFA, glycogen, muscle triglycerides), our findings could at least partly implicate variation in the mechanical efficiency of muscular work and that it may be larger in moderate-intensity exercise than previously estimated during high-intensity exercise (Bangsbo, 1996).

Muscle activity was lowest in VL relative to the whole QF and highest in VI. RF and VM were somewhat closer to the mean QF values, but still ∼15–20% under- and overestimated, respectively. Within individuals, variation was large both in VL and RF. In the VL of one subject glucose uptake was 60% lower than the mean of whole QF, while another had almost exactly the same value in VL as in the whole QF. Similarly in one subject, glucose uptake in the RF was 40% lower and in another it was 40% higher than the mean of whole QF.

Our results showed also a variable activation of other than QF muscles. In the biceps femoris and tibialis anterior muscles of the exercising limb glucose uptake was on average about 4.5-fold higher than that of the resting contralateral muscle and roughly half of the glucose uptake in the exercising VL and RF. In contrary to our results, Andersen et al. (1985) found minimal EMG activity in these particular muscles. This could imply that, although these muscles seem not to be voluntarily activated, glucose uptake was elevated due to passive movement. The experimental approach is not entirely comparable to this study since an occlusion cuff was placed above the knee to isolate the thigh region. Yet it has been recently shown that muscle blood flow during knee-extension is approximately tripled during passive leg movement (Hellsten et al., 2008). However, Hellsten et al. (2008) did not observe statistically significant increase in oxygen uptake in the passively moved leg, though on average oxygen uptake was twice as high in the passively moved leg. Taken together, these findings could imply that glucose taken up by the biceps muscle was not used for energy production, but stored for future use as muscle glycogen and warrants further investigation.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.