Body maintenance and repair: how food and exercise keep the musculoskeletal system in good shape


  • Michael J. Rennie

    1. University of Nottingham, School of Biomedical Sciences, Graduate Entry Medical School, Derby City General Hospital, Uttoxeter Road, Derby DE22 3DT, UK
    Search for more papers by this author

  • This is the 2004 G. L. Brown Prize Lecture, which was given by Professor M. J. Rennie at the Universities of Aberdeen, Bristol, Coventry, Glasgow, Loughborough and Oxford, and at The Queen's University of Belfast, King's College London and Tapton Beacon School, Sheffield.

Corresponding author M. J. Rennie: University of Nottingham, School of Biomedical Sciences, Graduate Entry Medical School, Derby City General Hospital, Uttoxeter Road, Derby DE22 3DT, UK. Email:


This article provides a personal view of how feeding and exercise acutely modify protein metabolism of human skeletal muscle, with discussion of the anabolic signalling mechanisms involved and some new findings on the metabolism of the turnover of collagen, tendon and bone.

In general terms it is very easy to understand how we can maintain the musculoskeletal mass of the body: eat well and keep active. The underlying mechanisms by which the lean body mass is maintained are of course hidden, but modern metabolic techniques have advanced so much that now we can delve beneath the surface of the body to examine what happens in terms of gene expression, cell signalling and tissue protein turnover and so provide a rather good description of some of the previously hidden mechanisms. The purpose of this article is to review some of the recent discoveries concerning how the musculoskeletal system responds to food and exercise, and how different parts of the system (e.g. those with mechanical functions and others with support functions) seem to show an integration of metabolic responses which suggest that the whole system does to some extent behave in a unitary and teleologically sensible fashion. Lastly there will be a discussion of the effects of ageing.

The musculoskeletal system and protein metabolism: why are we interested?

The musculoskeletal system comprises 75% of the lean body mass of a healthy person (Forbes, 1987) and, although it is obviously important in terms of our everyday life, many questions remain about its metabolic functions, particularly those involving protein. To some extent muscle can be regarded as a store of amino acids which can be used for wound healing and, after suitable conversion to glucose and ketone bodies, as fuels. Even bone appears to be a storage site for zinc, the need for which is closely related to its involvement in protein metabolism, particularly anabolism. Also, despite the fact that we know that there are many different varieties of collagen in the body, we know very little about the physiology of collagen turnover in tissues such as bone, tendon and muscle and the extent to which it is controlled by feeding, exercise, etc. We should expect that improved knowledge will help us understand the major changes which occur in the musculoskeletal system during growth, development and ageing. The question of ageing is particularly important since, owing to increased longevity, a larger proportion of our population, especially in the developed countries, comprises older people than in previous eras, but as yet there are very few successful drug-related strategies to prevent the wasting of muscle and bone that is so damaging to the quality of life and even to life itself. If we wish to pay more than lip service to the concept that we can better preserve mass and function of the musculoskeletal system by appropriate diet and exercise regimes, we need to uncover the evidence for this that will influence health and education policy-makers to provide educational and physical resources in order to allow individuals to manage their lives in the most healthful ways.

The relative roles of genes and the environment

Since the sequencing of the human genome there has been an expectation that we will be able to unveil many of the secrets underlying ways in which the human body is put together, the differences that exist between individuals in muscle and bone mass and composition, and how adaptable they are to physical activity. Although there have been some successes in identifying genes that are associated with particular musculoskeletal functions (Rankinen et al. 2004; Thompson et al. 2004), it seems that, as for many other human attributes, human body size and composition are as much a matter of environment as of natural endowment, with each having about 50% influence. The gentlemen pictured in Fig. 1 are in fact identical twins who chose to sculpt their bodies by different training regimes to completely different results, in order to pursue athletic careers in distance running and field events. Obviously the scope for environmental effects is large.

Figure 1.

Identical twins of different body size and composition due to different training regimes.

Most of what I will discuss concerns relatively short-term effects of food and exercise, i.e. those which occur within a time frame of up to 72 h, and I am going to say very little about alterations of gene transcription, since this has not been the focus of our work until recently. Nevertheless, it did come as a surprise to me and other workers to realize that it was possible to see marked alterations in gene expression within 2 h of finishing a bout of exercise (Pilegaard et al. 2003) or infusing insulin (Rome et al. 2003); given the much slower metabolic rate of human organs compared to that of a rat or a mouse, it was to be expected that these changes would take much longer. However, the biopsy procedure itself may influence gene expression (Vissing et al. 2005), and some of the reports in the literature may be questioned as not having taken this fully into account in experimental design.

Tools for the investigation of protein metabolic responses in muscle

For almost 30 years I have worked on mammalian protein and amino acid metabolism, initially in both animals and man but now exclusively in humans, using techniques specifically developed to investigate the dynamic nature of protein metabolism.

The concept of protein turnover is a difficult one for many outside the field to grasp, and the analogy I often use is a financial one. Most people can understand that there is a larger throughput of money through their current account than the amount of money it contains at any given time and this is affected by the amounts paid in income, spent and any deposits into a savings account or withdrawals from it to keep the current account in balance. This is exactly analogous to the flux of amino acids through the metabolic pool, in which dietary intake is analogous to income, amino acid oxidation (to ammonia and bicarbonate and then urea) is analogous to spending, protein synthesis (i.e. the movement of amino acids into protein in the carcass) is analogous to saving and protein breakdown (i.e. the liberation of amino acids into the free metabolic pool) is analogous to withdrawals from the savings account. Obviously the sizes of the current account (the metabolic pool) and the deposit account (the protein in the lean body mass) can be modulated by changes in diet and rates of amino acid oxidation, protein synthesis and protein breakdown.

Simply describing what happens to these processes in muscle with, for example, feeding, fasting or exercise, has taken the best part of 30 years, using stable isotope-labelled amino acid tracers. Stable isotopes are useful because by definition they do not spontaneously decay to release damaging ionization, and stable isotopes of nitrogen and oxygen are useful as biological tracers because they do not exist in a long-lived form as radioactive isotopes. For example, we have given stable isotope-labelled tracers to pregnant women to measure amino acid flux across the placenta and into babies just before birth (Chien et al. 1993), something that would be unethical with radioactive tracers.

The essence of the tracer technique is to measure the dilution of the tracer into the metabolic pool of the tracee of interest, whether it is the intracellular, free amino acid pool, the plasma pool, any tissue protein pool or indeed a specific protein. The dilution of the tracer into that pool gives a measure of metabolic flux into the pool. We have developed methods that allow us to measure the turnover of protein in the body as a whole and, simultaneously, using arteriovenous tracer labelling techniques, to measure processes of amino acid oxidation, protein synthesis and breakdown in limbs (which are mostly muscle), as well as synthesis of specific components of muscle, e.g. myofibrillar or sarcoplasmic proteins or collagen isolated from muscle biopsies taken during studies in which stable isotope tracer amino acids were infused. Details of exactly how measurements are made of the components of whole body and tissue amino acid metabolism are reviewed and detailed elsewhere (Smith & Rennie, 1996; Rennie, 1999).

Using these methods we have attempted to relate what happens when people eat or exercise to the processes of muscle protein building and wasting, which contribute to the maintenance of muscle made under steady-state conditions.

The effects of food on muscle protein turnover

We showed over 20 years ago that feeding doubled the amount of amino acid converted into protein in human muscle (Rennie et al. 1982) and we have since gone on to investigate which components of a mixed meal are important in causing this response, tackling the problem both by the use of limb arteriovenous tracer flux measurements and by incorporation of tracers into muscle protein sampled by muscle biopsy. What we found was that the major component of food responsible for stimulation of muscle protein synthesis was amino acids present as protein in food. No other component or normal response turns out to be as powerful as blood amino acid concentration. The evidence for this is that a very similar response in muscle protein synthesis to that obtained by feeding a mixed meal could be elicited by infusion of a mixture of all metabolically important amino acids (Bennet et al. 1989) and indeed that over a short period of time individual essential amino acids (especially the branched chain amino acids and particularly leucine amongst these) were stimulatory (Smith et al. 1998). Now this is interesting but also puzzling because in order to make most proteins all 20 amino acids are required. Foods insufficient in specific essential amino acids (e.g. haemoglobin or wheat protein) are not capable of sustaining the lean body mass or supporting growth without other protein sources (which is presumably why we add oatmeal to blood in making black pudding and the Masai add milk to the blood they bleed for food from their cows).

In fact it appears that the effect of individual essential amino acids, especially leucine, in stimulating protein synthesis occurs not by a mass action effect, i.e. protein synthesis increasing simply because there is more substrate, but due to a more subtle mechanism. Rather surprisingly, when low doses of essential amino acids are infused into humans (Alvestrand et al. 1990; Bohe et al. 2003) the intracellular free amino acid pool actually diminishes, suggesting that amino acids are being ‘sucked out’ of the intracellular pool, into protein. This is doubly surprising because it might have been expected that the processes of amino acid transport would have maintained the size of the intracellular pool, and certainly our knowledge of the capacity of amino acid transport systems (Taylor et al. 1999) would reinforce this; so this remains a problem to be solved. Nevertheless, this behaviour is reproducible and it also suggests that there is, somewhere on the outside of cells, an amino acid-sensing protein, possibly a leucine-sensing protein, since leucine alone can apparently cause the stimulation of synthesis. The sensor has to be outside rather than inside cells, where it would be of little use in detecting a rise in the availability of blood-borne amino acids after a meal coupled with a fall in the intracellular pool size.

What about the other major process of protein turnover, protein breakdown; does this not pay an important regulatory role in the control of the size of the muscle mass? In our studies it does contribute, but the extent of its contribution is much less: during feeding of a protein-rich meal muscle protein breakdown decreases about 20%, whereas muscle protein synthesis may double. Regulation of muscle bulk protein also appears to be attuned not to a direct effect of protein but to a secondary effect of the stimulation of insulin by components of a meal, including glucose and amino acids. The nature of the involvement of insulin, classically thought of as the most important anabolic hormone, is bizarre because it appears that in adult humans insulin does not do what it does in the growing animals traditionally used for metabolic research, such as rats, mice and even pigs. In people, it appears that it is possible to stimulate muscle protein synthesis by supplying exogenous amino acids alone while maintaining (using the insulin clamp techniques) basal blood insulin concentration at the overnight fasted level (Cuthbertson et al. 2004; Fig. 2). Furthermore, if amino acids in amounts capable of causing a maximal response in protein synthesis are given, adding further insulin has no further stimulatory effect on protein synthesis but does sharply reduce protein breakdown (Greenhaff et al. 2005). When we use the insulin clamp technique, utilizing octreotide to inhibit insulin secretion and then replacing insulin to desired levels, we also inhibit the secretion of growth hormone and any subsequent changes in insulin-like growth factor 1 (IGF-1). Thus it appears that the amino acid stimulation of muscle protein synthesis is also independent of the increases in growth hormone or IGF-1 which would normally occur in the unclamped situation.

Figure 2.

Dose–response relationships between rates of protein synthesis in myofibrillar and sarcoplasmic protein and oral dose of essential amino acids (EAA; A and B) and plasma leucine availability expressed as area under the curve (C and D) during insulin clamped studies in young and elderly men (from Cuthbertson et al. 2005, with permission).

It may be that the common link between the processes of synthesis and breakdown is the concentration of amino acids in the plasma, which stimulate synthesis directly and inhibit breakdown via stimulation of insulin; this would make mechanistic sense because the plasma is in connection with all tissues in the body and would therefore be capable of integrating the separate rates of tissue and whole body protein synthesis if, as we might suppose, the effects of amino acids on protein turnover are common to a variety of tissues.

The relationship between the availability of amino acids and the extent of stimulation of muscle protein synthesis is curvilinear (Bohe et al. 2003), suggesting that the process is saturated at high concentrations of amino acids, about five times those which are normally found in the bloodstream even after a rather large meal. Nevertheless, in the normal prebreakfast-to-postdinner range of blood amino acids, about a two-fold range, the responsiveness of the protein synthetic system appears to be almost linear and therefore rather sensitive to the protein content of meals. However, there is one puzzling feature about the behaviour of muscle in response to amino acids: after a short, latent period of about half an hour, muscle protein synthesis is initially rapidly stimulated but then there appears to be a switch off – a tachyphylaxis – in response to amino acids (Bohe et al. 2001), which results in the synthetic rate falling. It must also, presumably, result in diversion of amino acids away to catabolism in the liver, by ureagenesis and gluconeogenesis. This ‘muscle-full’ behaviour, together with the inability to stimulate muscle amino acid synthesis continuously by pouring in endogenous amino acids, explains why it is impossible to increase muscle size simply by eating (although the burden of weight carried by obese individuals does help to stimulate muscle hypertrophy!).

How does muscle know to increase muscle building when amino acids are available?

Of course this is an important question: understanding the sensing and signalling mechanisms involved has been made possible by rapid progress over the last 20 years by the pioneering work of Jefferson and Kimball, who have made great strides in developing this area (Jefferson & Kimball, 2001).

We are still ignorant about the nature of the initial amino acid sensor, but it is well recognized that the availability of amino acids causes changes in the activity of protein kinases in muscle and, as a result of phosphorylation of the kinase itself, there is a cascade of anabolic signalling molecules which ultimately results in stimulation of the translation of pre-existing mRNA and also the translation of specific mRNA coding for regulatory proteins themselves. We have shown that ingestion of essential amino acids causes a marked increase in the activity of the phospho forms of the proteins mTOR and p70s6 kinase and diminution in the phosphorylation of eIF2b (Cuthbertson et al. 2005; Fig. 3). These changes activate the signalling proteins and are an indication that the anabolic signalling pathways are stimulated by amino acids alone, since these experiments were carried out in the absence of any change in insulin, growth hormone or IGF-1 and without changes in other substrates, such as glucose or fatty acids.

Figure 3.

Response of phosphoproteins in human muscle to essential amino acids
As in Fig. 2 (from Cuthbertson et al. 2005, with permission).

We are now hoping to go on to investigate the basis of the muscle ‘muscle-full’ phenomenon and the dose responsiveness of the changes in the anabolic signalling molecule and their activation.

So far I have implied that mixed muscle protein is an indivisible entity, but of course it is made up of thousands of proteins, with actin and myosin of the myofibrillar apparatus making up about two-thirds of the total, with substantial contributions from some soluble proteins, such as creatine kinase and carbonic anhydrase. It is relatively easy to make measurements of the turnover of subgroups of mixed proteins, such as the myofibrillar and the sarcoplasmic proteins (Bohe et al. 2001; Louis et al. 2003), and some workers have been able to make measurements of the rate of turnover of individual proteins, such as actin and myosin; however, in our experience the directions of change in synthesis of these proteins is identical under every circumstance of stimulation or challenge in which we have examined them. Thus, when we feed an individual after a period of overnight fasting, the synthetic rates of all of the proteins we have measured goes up although the relative increase of the slowest turning over proteins is greater, presumably in order to maintain the composition of muscle cells. A similar phenomenon is seen with exercise (Louis et al. 2003; Moore et al. 2005).

However, there is one class of proteins within muscle tissue, i.e. collagen, one of the extracellular matrix proteins, which behaves somewhat differently. Collagen is a large protein made up of repeating subunits of triple helices of polypeptides which are comprised of about 20% of the imino acids proline and its post-translationally hydroxylated derivative, hydroxyproline. It is made by fibroblasts in muscle, not by muscle cells themselves, and it forms a basket-like, three-dimensional network around muscle fibres, allowing the muscle fibres to move independently of one another (because of the elasticity of this system), but also allowing the transmission of force to tendons and ligaments (see Kjaer, 2004).

We have been very interested in the behaviour of collagen in skeletal muscle because of the obvious requirements for bigger muscles to have a bigger support network. We therefore predicted that if we were to measure muscle collagen turnover it ought to show nutritional regulation similar to that of muscle per se. In fact, we were totally wrong. Muscle collagen turnover is about half as fast as myofibrillar turnover and it shows no nutritional modulation whatsoever; in fact its main physiological regulation appears to be via stimuli arising from contractile activity (Kjaer et al. 2005; Figs 4 and 5). Similarly, tendon collagen appears to be nutritionally insensitive, although the turnover of tendon collagen is about twice as fast as that of muscle collagen, for reasons that we do not understand.

Figure 4.

Response of muscle and tendon protein synthetic rates to preceding strenuous exercise.

Figure 5.

Scheme suggesting how collagen synthesis of muscle and intracellular matrix may be coordinated
ECM, extracellular matrix.

Of course, the other tissue in which there is a substantial amount of collagen is bone, of which 95% of the total organic content is type I collagen. Bone has traditionally been thought of as having very slow turnover characteristics, but although there is good objective quantitative evidence of the rates of turnover of bone calcium (about 1 g day−1) there have been no direct measurements of the rates of human bone collagen turnover, so we cannot be sure. In my view (perhaps controversially) the popular so-called markers of bone metabolism, such as the pro-collagen peptides or C and N terminal telopeptide fragments, do not give adequate quantitative information and may not reflect the extent of physiological changes when appropriately stimulated. We therefore developed a method for measurement of bone collagen synthesis in humans using bone biopsy to sample bone from the hip; we have been extremely surprised by the results. First, it appears that bone collagen synthesis is about as fast as that of mixed muscle protein (i.e. about 1–2% day−1) and, secondly, astonishingly, is able to double in response to stimulation by an intravenously supplied meal (Babraj et al. 2005). This information is potentially important because it suggests that bone turnover is fast enough to respond quickly to imposed stimuli (such as change in feeding and pattern of exercise) and may contribute more to whole body protein turnover than has been hitherto realized.

Effects of exercise and immobilization

At the end of the nineteenth century there was a substantial controversy about the possibility that protein was used as a fuel for exercise. When we examined this question again about 20 years ago we showed that indeed there was a substantial increase in amino acid oxidation as a result of exercise and that leucine (and probably the other branched chain amino acids) were oxidized in muscle during moderate exercise (Rennie et al. 1980). However, the total amount of fuel supplied from this route was small, certainly less than 15% of the total fuel utilized. It may be that pari passu amino acid oxidation provides anaplerotic substrate for the Krebs cycle to replace intermediates drained off elsewhere (although recent work suggest this is more unlikely than previously thought) and, possibly, amino acids help regenerate adenine nucleotides in muscle in the postexercise period (Rennie, 1996). However, it appears that this chapter is effectively closed in terms of research.

Much more interesting are the effects of strenuous exercise in stimulating muscle growth and muscle protein synthesis (and indeed collagen synthesis in muscle, tendon and bone). Paradoxically, after I became interested in this topic, we approached it in a rather back-to-front fashion. We started by looking at the effects of immobilization in a long leg cast and were able to show that, relative to the uncasted leg, the rate of muscle protein synthesis fell by about half over a period of 7 weeks (Gibson et al. 1987) and that muscle protein breakdown appeared to fall also, in an adaptive response which presumably had the advantage of lessening net muscle loss. We showed, in a separate study (Gibson et al. 1989), that stimulating quadriceps muscle electrically at a very low rate (∼5% maximum voluntary contraction for 1 h day−1) caused an increase in muscle protein synthesis so that muscle wasting was abolished over the period of the casting. We also showed that the adaptive fall in muscle protein breakdown did not occur with stimulation, and indeed this pattern of increased turnover, with the synthesis exceeding breakdown, is what is seen in other circumstances of chronic stimulation, such as in the non-classic studies of Geoff Laurent on the muscles of the chicken wing (Laurent et al. 1978). Since Alan Chesley (Chesley et al. 1992) carried out the first study of the effects of acute exercise on muscle protein synthesis (generating samples which were analysed in our laboratory by my long-time coworker, Ken Smith), a number of workers in Canada and the USA and indeed in our laboratory have delineated the acute response of muscle protein synthesis to strenuous exercise (see for reviews Rennie & Tipton, 2000; Rennie et al. 2004). It appears that when muscle protein synthesis is measured by incorporation of tracer into the protein there is a latency of about 90 min after exercise when the synthetic rate is similar to that at rest, but afterwards there is a very substantial increase, possibly up to 4–5 times the basal rate by 12 h, this rate being maintained for a further 12 h before falling over the subsequent 48 h. If protein-containing food or amino acids are delivered either immediately before exercise or in the postexercise period then the rise is greater. If no amino acids are delivered then protein breakdown will exceed protein synthesis and there will be no net accretion of protein. Other workers have demonstrated that the rise in protein breakdown may be ameliorated by supply of substrates which are insulin secretagogues, such as glucose, so feeding helps to maximize the postexercise response.

We have made some preliminary studies aimed at delineating the dose–response relationship between the intensity of exercise and the rates of muscle protein synthesis and have shown that when the same total amount of work is done, i.e. when the same total amount of ATP is turned over, exercise at 60, 75 and 90% of the one-repetition-maximum force results in exactly the same stimulation of muscle protein synthesis, suggesting that once all muscle fibres are recruited (as they were in our study) increases in tension above 65% cause no further stimulation in muscle protein synthesis (Bowtell et al. 2003). As with feeding (Cuthbertson et al. 2005), the increases in myofibrillar protein synthesis appear to be greater than the increases in sarcoplasmic protein synthesis, so that after exercise the rates of protein synthesis in the two pools appear to be virtually identical (Cuthbertson et al. 2002).

What role does the anabolic signalling network play in the postexercise period?

We have shown that both protein kinase B (PKB) and P70s6 kinase rise soon after exercise and in fact the rise appears to antedate the rise in protein synthesis caused by the exercise; feeding in the postexercise period maximizes the anabolic response.

Only resistance exercise causes changes in muscle mass in the longer term, and the mechanism underlying the dichotomy between the increases in muscle mass with resistance exercise and the changes in muscle composition (expression of slow myosin phenotypes, increased mitochondrial capacity, etc.) have been rather mysterious. In experiments initiated by colleagues (Henning Wackerhage and Phil Atherton), it has been possible to gain some insight into signalling mechanisms underlying the process (Atherton et al. 2005). My colleagues arranged to stimulate isolated rat muscles in vitro either with low-frequency (10 Hz) electrical stimulation, mimicking a pattern seen in an endurance exercise, or with trains of high-frequency stimulation (100 Hz), which caused a much greater increase in force, but with only 18 such trains in total. We measured myofibrillar and sarcoplasmic protein synthesis and my colleagues measured the phosphorylation of a protein called TSC2, found upstream of mTOR but which is differentially phosphorylated by PKB and AMP-dependent protein kinase. What we found was that high-intensity contraction stimulated PKB activity, which was associated with an increase in the phosphorylation of TSC2 and the activation of the mTOR pathway, resulting in markedly increased myofibrillar protein synthesis. However, the low-intensity stimulation was associated with decreased ATP concentrations, activation of AMP kinase and phosphorylation of TSC2 on an alternative site, which had the effect of inhibiting any stimulation of TSC2 via PKB. The result was that bulk protein synthesis was not stimulated and mTOR was not activated. This switch mechanism may help to explain why the two gentlemen in Fig. 1 ended up with the kinds of bodies they did.

Responses of tendon and muscle collagen to exercise

In studies carried out with colleagues in Copenhagen, we investigated the relative relationships between possible alterations in muscle myofibrillar proteins, collagen within muscle (mainly perimysial collagen) and tendon collagen (sampled by percutaneous biopsy of the patella tendon). The results were gratifying because they demonstrated that the relative extents of the stimulation of perimysial collagen and myofibrillar protein synthesis were almost superimposable, and that tendon collagen synthesis, although not responding as much as perimysial collagen, nevertheless showed a 30% rise within 6 h and a 50% rise within 24 h after exercise, remaining elevated for up to 48 h after the exercise bout (Miller et al. 2004). The similar pattern of response between collagen and the extracellular matrix in skeletal muscle protein is fascinating and leads us to hypothesize that there may be either mechanical or humoral factors linking skeletal muscle and the extracellular matrix (e.g. through integrins or through growth factors such as transforming growth factor-β (TGFβ) or mechano growth factor (MGF)) which co-ordinate the responses of the perimysial network and muscle fibre hypertrophy (Fig. 5). We have not yet managed to carry out any studies on bone, but the literature suggests to us that physical activity is a potent stimulator of bone anabolism, and we believe that our newly developed methods would be capable of revealing the effects of stimulating bone collagen synthesis in a way which would be much more sensitive and precise than alternative methods, such as measurement of bone mineral density by Dual Energy X-ray Absorptiometry, ultrasound or computerized tomography. We are certainly looking forward to carrying out such studies.

The effects of ageing

It is a matter of common experience that as people age muscle and bone tend to waste, leading to decreased strength, increased fatigability and an increased propensity to fall and break bones. This has a devastating effect on the ability to carry out a normal independent life and is a matter of substantial concern. There are almost no good pharmacological treatments for sarcopenia (the name coined by workers in the area to ‘badge’ the condition, to help attract research funding); the few drugs available for bone wasting are inconvenient to take and seem to have very little effect on qualities of bone such as resilience and toughness. However, it seems that there is reasonably good epidemiological evidence that people who are fit and active maintain a higher level of bone and muscle mass for longer than sedentary individuals. When we started to think about studying the effects of ageing, the literature portrayed a substantial amount of controversy in relation to muscle (Dorrens & Rennie, 2003). It was agreed by all workers that there was evidence of muscle wasting in individuals over, say, 70 years of age but the mechanisms involved in terms of alteration of muscle protein synthesis and breakdown were not clear. Some workers claimed that there were diminutions in the rate of muscle protein synthesis in the overnight fasted state, but the extent of the reported deficit was substantial, so much so that the observed rate of muscle wasting should have been faster than measured. There was certainly no good evidence of increased rates of muscle protein breakdown in the fasted state. In our view, the balance of evidence suggested that in the fasted state there were indeed very few decrements in muscle protein turnover in elderly subjects.

How then to explain the wasting? Our hypothesis was that there was some kind of difficulty in taking advantage of nutrients, and we set out to test this by examining the responses to oral essential amino acids in a group of young, healthy (28-year-old) men and a group of healthy, elderly (70-year-old) men who had similar lean body masses as a percentage of their body weight (∼80% on average) but who had different leg muscle masses, the elderly subjects showing a slight decrement from 20 to 16% of body weight. We studied these subjects under conditions in which we clamped the secretion of insulin at postabsorptive values (i.e. fasting). We were pleased to discover that our hypothesis was supported. There were no decrements of muscle protein synthesis in the basal state but elderly men appeared to be less able to take advantage of essential amino acids in stimulating muscle protein synthesis. Indeed, the extent on their inability to do so was reflected in the fact that, paradoxically, they had more amino acid in the blood after ingestion of the oral essential amino acid than the younger subjects. When we related the response of muscle protein synthesis not to the dose given but to the availability of amino acids in the plasma (in fact the area under the curve of leucine concentration versus time), the differences between the young and the old subjects were exacerbated. Thus, it appeared that the elderly subjects exhibited what might be called ‘amino acid resistance’ of muscle (Cuthbertson et al. 2005). The elderly subjects did have slightly less muscle and also slightly less capacity for protein synthesis within them (indicated by the RNA:DNA ratio), so that the efficiency of protein synthesis, i.e. protein synthesis per unit RNA, was substantially less. The major mechanistic differences appeared to lie in the anabolic signalling pathways in the young and old subjects. We found that the total amounts of anabolic signalling proteins were diminished in the elderly subjects, and in addition their responsiveness in terms of phosphorylation of the proteins after essential amino acid ingestion was markedly less. Thus, the responsiveness and the sensitivity of the response were both diminished in the elderly subjects. Note that these effects were entirely to do with the responses to essential amino acids, since we had taken insulin responsiveness out of the equation by using the insulin clamp.

We can say immediately that any idea that elderly people require more dietary protein to maintain muscle which is wasting is unlikely to be true, because they are unable to take advantage of it. However, they do show some responsiveness of muscle protein synthesis to diet, so it may be that maximizing the protein:energy ratio of food, while keeping the total amount of energy intake down to minimize weight gain, etc. is a reasonable strategy. We also predict that because there is a synergistic effect between exercise (especially resistance exercise) and feeding in stimulating muscle protein synthesis, elderly subjects who have been shown to be capable of muscle hypertrophy after appropriate resistance training programmes could maximize their muscle maintenance by a suitable programme of resistance exercise, with feeding concentrated in the immediate postexercise period. The latter appears to be an efficient way of increasing the synergy between exercise and amino acid delivery according to recent work from Copenhagen and the USA (Esmarck et al. 2001; Levenhagen et al. 2002).


The past 20 years have shown a substantial increase in our ability to peer inside muscle at the molecular mechanisms involved in control of the size of the muscle mass, and we are now extending these techniques to other tissues, such as extracellular matrix and bone. I predict that we will be able to provide a firm rational basis for alterations of lifestyle, principally change of diet and exercise activities, which will be shown objectively to underpin increases in the ability to maintain the muscle mass. I sincerely hope I live long enough to see this prediction fulfilled.



I thank The Physiological Society for electing me to the G. L. Brown Prize Lectureship, the extremely generous hosts throughout the country at whose institutions I delivered the lectures and of course my many colleagues who contributed invaluable hard work and insight to the studies I have described here. The work described here was supported from a number of sources, principally United Kingdom MRC and BBSRC, The University of Dundee, The University of Nottingham, The Wellcome Trust, US NIH and the Shriners' Burns Institute of the University of Texas Medical Branch, Galveston.