New Study: The Cellular Energy Sensor, AMPK - Switching it on for Training Adaptation

Dec 09, 2021

by Dr Dan Plews

As we have discussed in several previous blogs, and in our courses, adaptations to endurance training are produced through activation of cellular signalling pathways in response to individual training sessions, with these cellular signalling pathways activated through detection of the physiological stresses generated through exercise. In this blog, we are going to focus on one of the most well-known proteins in the adaptive signalling cascade, the cellular energy sensor AMPK.

AMPK, or to give its full name, 5’adenosine monophosphate-activated protein kinase, is an important protein in the detection of the cellular energy stresses generated during endurance exercise (1). Specifically, AMPK within the muscle is activated by the metabolic stresses generated by exercise, and subsequently activates signalling pathways associated with endurance training adaptation (2). This pathway makes a lot of sense; if our muscle is stressed by exercise, we detect it and then send signals to make adaptations that allow us to respond better to that stress the next time it is encountered. To really make it simple, exercise physiologists would consider activating AMPK as one means of generating endurance training adaptations!

The Paper

To that end, my PhD student Jeff Rothschild and I – along with a team of others – did an analysis to try and identify the factors that most contribute to AMPK activation during exercise. To do this we pooled data from 89 studies containing nearly 1000 participants and performed what is called a meta-regression to statistically identify the most important factors influencing the AMPK activation response to exercise. This work has recently been published in the prestigious Journal of Sports Medicine.

The results

As you might expect, the paper has a lot of data in it, so in this blog I’ll try and draw your attention to the most pertinent findings for endurance athletes and coaches. When considering this information, it is worth considering that pre-exercise glycogen availability can be manipulated with prior exercise and diet (so-called low-glycogen training), glycogen depletion during exercise can be increased by manipulating exercise intensity and duration, and the disturbance of muscle metabolic homeostasis is strongly related to exercise intensity, where hard interval sessions above the maximum metabolic steady-state result in greater homeostatic disturbances.

1. The amount of glycogen in muscle at the beginning of exercise was not related to subsequent AMPK activation, but end-exercise glycogen was (where lower post-exercise muscle glycogen was related to more AMPK activation), as was the degree of glycogen depletion during exercise (more depletion, more AMPK activation).

2. The degree of disturbance in muscle metabolic homeostasis (e.g. phosphocreatine depletion and free ADP accumulation) was also related to AMPK activity, with greater disturbances associated with greater AMPK activation.

3. When exercise intensity is high enough, and muscle metabolic homeostasis is disturbed sufficiently, AMPK activation will be pretty high.

4. During exercise below the lactate threshold or VT1, AMPK activation is fairly low.

5. When exercise is above the lactate threshold but below the maximum metabolic steady-state – i.e. in the sweet spot – glycogen and therefore modifiable factors like diet and prior exercise may have a bigger influence on AMPK activation.

 The takeaways: what do these results mean in practice?

Okay, so that was a lot of jargon, but a lot of interesting outputs. It is worth acknowledging here that there is a fair amount of uncertainty around these relationships, so the outcomes we have identified should be considered broad effects of these variables on AMPK activation.

That being said, some interesting patterns and messages do come out of the data that might be used in practice. Firstly, when considering manipulating glycogen availability to up-regulate training adaptations, it is important to understand that it seems to be the end-exercise glycogen that ultimately matters, rather than the pre-exercise glycogen. That’s a subtle distinction. For example, if you are doing an easy, mid-week 45 min spin on the bike, there is likely little glycogen depletion taking place during the exercise. So, restricting carbohydrate intake prior to that session to achieve a moderate reduction in muscle glycogen probably isn’t going to make much difference when it comes to AMPK activation. In order to stimulate the AMPK activation achieved by that session, the pre-exercise reduction in glycogen might have to be quite aggressive, such that the end-exercise glycogen content is really quite low. Therefore, you might be better served by trying to manipulate glycogen in other training sessions like tempo session..

Secondly, when performing demanding, high intensity sessions, it seems to be the effect of glycogen on AMPK activation seems comparatively small vs. the degree of homeostatic disturbance per se, which is achieved by the demands of the exercise. For these sessions, which will achieve a fair amount of glycogen depletion just through the high glycolytic demands of the high exercise intensity, making further efforts to manipulate glycogen may be unnecessary, at least when it comes to AMPK activation.

There may be more scope for glycogen – and therefore manipulating prior exercise and diet – to have an effect during those Ironman-specific sweet-spot sessions taking place between the two thresholds. Glycogen depletion and AMPK activation is observed in these sessions, likely without huge levels of phosphocreatine depletion or ADP accumulation. It might be possible to manipulate glycogen availability prior to training to achieve even lower end-exercise glycogen concentrations and therefore ramp up the stimulus for AMPK activation.

I should point out that AMPK activation is of course only one step in the training adaptation cascade, and there are many other targets of endurance training. The work presented here is, however, a good step in the direction of understanding the mechanisms behind the training process!

Jeff also produced this awesome app to explore the data and the models - Check it out here


1. Gowans GJ, Hardie DG. AMPK – a cellular energy sensor primarily regulated by AMP. Biochem Soc Trans 42: 71–75, 2014. doi: 10.1042/BST20130244.AMPK.

2. Steinberg GR, Kemp BE. AMPK in health and disease. Physiol Rev 89: 1025–1078, 2009. doi: 10.1152/physrev.00011.2008.


Endure IQ's weekly Brew Up newsletter brings you Dr. Dan Plews' weekly top 5. Everything from the latest science and tech, to must read books/podcasts, and weekly workouts.



Get the latest Brew Up newsletter from Endure IQ's founder, Dr. Dan Plews.