Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2023 Dec 5;603(1):7–15. doi: 10.1113/JP285616

Current limitations and future opportunities of tracer studies of muscle ageing

Colleen L O'Reilly 1, Sue C Bodine 1,2, Benjamin F Miller 1,2,✉
PMCID: PMC11150331  NIHMSID: NIHMS1962231  PMID: 38051758

Current focus of muscle protein turnover with age

The overarching question that we address here is whether current tracer approaches focused on the measurement of protein synthesis are leading us in a fruitful direction to combat muscle loss with age. Because ageing is largely a loss of protein mass over time, logically, the field has focused on improving anabolism. To assess an improvement in anabolism, tracer‐based measurement of protein synthesis is often a primary outcome. It is possible that the relative ease of measuring protein synthesis in comparison to breakdown has biased research efforts towards the protein synthesis side of the equation. An additional reason is that some argue that the slowing of protein synthesis and/or anabolic resistance is the primary determinant of muscle loss with age (Burd et al., 2013; Nunes et al., 2022). The statement that ‘ageing decreases muscle protein synthesis’ often appears without citation because it is a generally accepted ‘fact’ in the skeletal muscle literature. As far as we can tell, the studies that first led to this conclusion were those studies from the Nair lab that looked at skeletal muscle in ageing humans (Balagopal et al., 1997; Rooyackers et al., 1996). A point of consideration is that these studies were performed with labelled amino acid infusions, the importance of which will be discussed here. The conclusion that ageing decreases muscle protein synthesis has led to interventions for aged muscle, such as amino acid feeding, that focus on increasing protein synthesis through activation of the mechanistic target of rapamycin (mTOR). However, a rigorous evaluation of the literature indicates that there are several contradictory findings to this ‘known’. There are studies showing that protein synthesis does not decrease in skeletal muscle with age and, conversely, showing increases in protein synthesis compared to adult muscle (Fuqua et al., 2023; Kimball et al., 2004; Miller et al., 2019). Furthermore, studies in rodent models (Fuqua et al., 2023; Joseph et al., 2019; Tang et al., 2019; White et al., 2016) and humans (Guillet et al., 2004; Markofski et al., 2015) show that mTOR activity is increased in aged muscle and that treatments that inhibit mTOR or slow protein synthesis over the long‐term, such as rapalogs (Joseph et al., 2019) and protein or branch chain amino acid restriction (Richardson et al., 2021), improve muscle function with age. Instead of basing interventions and future studies on the premise that protein synthesis uniformly decreases with age, an approach that focuses on proteostatic mechanisms that slow muscle ageing could open new possibilities for interventions.

The current overall (but admittedly generalized) framework paints protein synthesis as good and protein breakdown as bad. In other words, it focuses the attention on increasing muscle protein synthesis in aged muscle at the same time as reducing muscle protein breakdown. However, these conclusions are based on a body of literature using approaches that may miss crucial aspects regarding protein turnover, and are also based on the foundation that accumulation of protein mass is the most important target for maintaining muscle function with age. Here, we aim to highlight some methodological considerations to shift the focus of what is important for muscle protein turnover with age. At a minimum, we hope to provide tools that help readers evaluate the literature with a critical eye. Ideally, we hope that the field will adapt some non‐traditional isotope approaches to better direct therapeutic efforts.

A focus on proteostasis

The loss of skeletal muscle function often precedes and exceeds loss of mass, indicating that protein quality is as important as protein quantity. Although muscle mass, and hence protein mass, has an impact on overall function, it is equally important to have the correct proteins for cellular tasks and for those proteins to be assembled properly and function well. The matching of well‐functioning proteins to the demands of the cell is referred to as protein homeostasis, or proteostasis. Proteostasis is a self‐regulating process through which systems maintain protein functionality at the same time as adjusting to changing conditions. As such, maintaining proteostasis in one cell type may not be the same as in another cell type, or even at different times within a cell type. The dynamic mechanisms through which proteostasis is maintained is a network of complex interrelated cellular activities such as protein biogenesis, folding, transport and degradation that collectively determine proteome structure and function (Balch et al., 2008). This dynamic feature of proteostasis is important when assessing mechanisms because the proteome continually adjusts to the current environment. When these mechanisms fail, there is a loss of proteostasis.

The mechanisms that maintain proteostasis become dysregulated with advancing age (Hipp et al., 2019; Taylor & Dillin, 2011). The accumulation of dysfunctional proteins compromises cellular function and responsiveness to cellular stresses leading to age‐related deterioration, and more than 50 diseases (Walker & LeVine, 2012). Conversely, the ability to maintain proteostasis slows the ageing process (Balch et al., 2008; Jayaraj et al., 2020; Pride et al., 2015). It is well established in a variety of tissues, including skeletal muscle, that mechanisms of proteostatic quality control fail with advancing age such that decreased proteostasis is a hallmark of ageing (López‐Otín et al., 2013). One well described feature of proteostatic decline is the accumulation of protein aggregates that are resistant to turnover. An additional consequence of a decline in proteostasis is increased energetic costs of maintaining the proteome. A decline of protein function in organelles such as the mitochondria compounded by increased energetic costs of maintaining the proteome can result in an increasing strain on the cell to maintain proteostasis. It is important first to distinguish proteostasis from simple changes in protein content.

Here, we focus on the proteostatic mechanisms of protein synthesis and protein breakdown. A helpful construct for understanding proteostasis is to consider pools of protein. The proteins entering (synthesis) and proteins leaving (breakdown) constitute the protein ‘pool’ with synthesis and breakdown together determining the composition of the pool and the rate at which the pool turns over. Furthermore, the balance of synthesis and breakdown determines the size of the protein pool (increases, decreases or stays the same). Protein pools (e.g. skeletal muscle protein pool, extracellular matrix pool, mitochondrial pool or myosin) have progressively become more specific over time because our measurement techniques have become more sensitive. With improved sensitivity there is also improved understanding of the mechanisms of proteostasis.

Considerations and opportunities of deuterium oxide (D2O) labelling

Dufner et al. (2005), Gasier et al. (2009) and Busch et al. (2006) started using deuterium oxide (D2O) for long‐term measurements of protein synthesis in the mid 2000s. Our group published our first paper using D2O in 2011 (Robinson et al., 2011), with others thereafter (Wilkinson et al., 2014). The groups that pioneered this work all brought slightly different approaches to its use. Although D2O shares the same precursor‐product method of measuring protein synthesis as labelled amino acids, the long‐term nature of the labelling introduced new considerations as well as new opportunities (Miller et al., 2020). Early on, we learned that there is not a one‐size fits all approach for using D2O. However, each failure taught us something new that informs the current perspective. Here, we use our studies to address what we consider to be crucial considerations and opportunities for the use of D2O for understanding proteostatic mechanisms. We provide these insights in four brief sections: (1) ‘muscle’ protein synthesis is an oversimplification; (2) the period of measurement is important; (3) the non‐steady state must be accounted for and protein breakdown matters; and (4) the number of proteins that are turning over is often overlooked. Although we have focused on muscle ageing, the principles laid out here are relevant to all long‐term tracer studies and period of protein loss or gain.

Muscle protein synthesis is an oversimplification

Published papers often refer to ‘muscle’ protein synthesis where muscle is a homogeneous entity with a given protein synthesis rate. However, similar to all tissues, muscle is composed of multiple cell types and, within each cell type, there are thousands of proteins that are independently regulated (Fig. 1). With single cell (which do not include myofibers) and single nuclei (which do include myofibers) sequencing, it is possible to estimate the abundance of cell types. Adult skeletal muscle is composed primarily of mature myofibers and satellite cells, with 55–88% of measured nuclei in mice being myonuclei and 1–4% being satellite cells (Borowik et al., 2023; Dos Santos et al., 2020; Orchard et al., 2021; Perez et al., 2022). Single nuclei studies of mice show that skeletal muscle has 2–4% immune cells, 2–4% smooth muscle cells, 2–7% endothelial cells and 10–19% fibro‐adipogenic progenitors (Dos Santos et al., 2020; Orchard et al., 2021; Perez et al., 2022). In humans, the percentage of myonuclei may be slightly higher. What is often overlooked is that the abundance and turnover of proliferative cell types contributes to bulk protein synthesis rates because cell replication requires the doubling of protein content (Miller et al., 2014). Therefore, even though supportive cell types are less abundant than myofibers, they have an important influence on bulk protein synthesis rates.

Figure 1.  Muscle protein synthesis is an oversimplification.

Figure 1

Open in a new tab

A, schematic of the complexity of skeletal muscle. Within skeletal muscle there are various cell types that will have different abundance and bulk turnover rates based on the primary functions of the cells. Additionally, each cell type has multiple components, made up of thousands of proteins that are all independently regulated. B, example representation of data from the different fractions and components of a muscle sample. In this example, when looking at bulk differences (i.e. a mixed muscle fraction), there were no differences between control at treatment. However, when only the mitochondrial fraction was measured in bulk the treatment had a lower synthesis rate than the control group. Finally, when looking at synthesis of individual proteins, there will be some proteins that increase synthesis with treatment and some that decrease with treatment. Created with BioRender.

Similar to the differences in the abundance of cell types, there is a large distribution of protein abundances from highly abundant myofibrillar components to less abundant cell surface proteins. The measurement of muscle protein synthesis distills this complexity into a single value, which is largely determined by rapidly turning over proteins and highly abundant proteins (Miller et al., 2015). Protein synthesis is one of the most energy costly processes in the cell. Because there is a finite amount of energy on demand in a cell, increases in the synthesis of some proteins inevitably results in the slowing of the synthesis of others. Translation of proteins is highly regulated as in transcription. It is inconceivable that a single value of ‘mRNA transcription’ would be published to apply to all individually regulated mRNA in a tissue, yet it is common practice to publish a single value of ‘protein synthesis’ to represent the making of independently regulated proteins. Admittedly, measuring the synthesis rates of individual proteins is challenging, although the advent of tracer‐based proteomic approaches is helping to rectify this need. Using tracer‐based proteomics has shown that, even when a specific treatment does not change bulk protein synthesis, the majority of individual proteins within the bulk measurement are changing synthesis rates (Wolff et al., 2021). Ironically, what makes tracer‐based proteomics challenging stems from the topic of this section, namely the large differences in abundances and turnover of individual proteins. Newer technology and algorithms for liquid chromatography‐tandem mass spectrometry analysis show great promise for capturing a larger number of muscle proteins so that the reporting of ‘muscle’ protein synthesis could become a thing of the past.

The period of measurement is important

One of the primary advantages of using D2O is that it can be used over prolonged periods. This long‐term use is facilitated by the ease of administration in drinking water that does not require tethering a human to an i.v. pole or using surgery for an indwelling catheter in rodents. An advantage of the long‐term assessment is that it integrates responses over time to understand cumulative responses. With the increased use of D2O for muscle protein turnover measurements, it became clear that rates obtained with D2O were often lower than those from infused amino acids (Gasier et al., 2009). We hypothesized that the length of labelling, not the choice of tracer, was impacting these results. We demonstrated, via a mathematical model, that a short labelling period (e.g. 20 min) could result in 29% higher rates of synthesis than labelling the same protein pool for a period of 6 weeks (Miller et al., 2015). The reason for the difference is the integrated nature of the protein synthesis measurements. As mentioned in the previous section, synthesis values are the integrated average of thousands of individual proteins that have different abundances and synthesis rates. At shorter periods, faster proteins are over‐represented in an average synthesis value. As time goes on, the faster proteins reach 100% new (fully turned over), whereas slower proteins increase their contribution to the average value. Therefore, with longer labelling periods, the integrated rates of synthesis are lower because they have increasing representation of slower proteins (Fig. 2).

Figure 2. The period of measure impacts protein synthesis.

Figure 2

Open in a new tab

A, the black line represents the sum of all proteins measured in a sample (slow and fast) and is used to calculate the fraction new of a sample. This line is the sum of the fast (red line) and slow (blue line) fraction new accumulation over time. In early time points (identified by the box on the left), the contribution of slow turning over proteins (blue) to the total fraction new is low compared to fast proteins (red) as represented by the one blue protein. However, in later time points, the overall contribution of slow proteins is higher as shown in the box on the right. B, There are many benefits to a time course approach. In this example, the black line and the blue line represent the regression lines calculated from a time course where data was taken every 10 days. The boxes above the line describe what would be calculated when using a single time point at any of these 3 day points (10, 30 or 50). This shows that the shorter time (10 days) would represent both groups still rising to plateau with higher but similar synthesis rates for both groups (old and young). However, at 30 and 50 days, both groups have reached a plateau at a different fraction. If the time course data were not available, it would not be apparent when either group has reached a plateau, making the 50 day data an underestimation of what is happening as shown by the difference in numbers of the 30 day and 50 day bar graphs. Created with BioRender.

There are several practical issues associated with the choice of labelling period. The first is that the proteins of primary interest by the investigator should inform the period of labelling. In muscle, many investigators are interested in muscle mass, which is primarily determined by the amount of myofibrillar proteins. Others might be interested in mitochondrial proteins aiming to study mitochondrial adaptations. In both cases, these pools of proteins are relatively slow to turn over and are therefore best captured with longer labelling periods. It is worth mentioning that, out of convenience, many studies in the skeletal muscle field use short labelling periods. An extreme example is the 20 min labelling period used in the puromycin labelling method. Our observation of studies that have used the puromycin method is that, when an image of the bands is included, there is sometimes a lack of bands at the higher molecular weight (200+ kDa) where one would expect the myosin heavy chains. This lack of higher molecular weight bands is a concern when one is interested in the contributions of protein synthesis to changes in mass because the method is not measuring the rate of synthesis of the proteins that primarily contribute to mass. In our paper that modelled the impact of labelling period (Miller et al., 2015), we initially recommended 4 weeks of labelling in skeletal muscle. However, as our experience has grown, we have learned that this period can be too long and that there is no one‐size‐fits‐all labelling period. Therefore, each infusion period should be tailored to the specific question and outcome of interest.

A second practical consideration is the use of repeated biopsies with long‐term labelling studies in human. In studies with amino acid infusions, it was common practice to take repeated biopsies over a 4–12 h period to look at early and late responses. When using D2O, these periods can be days to weeks. However, when using D2O and comparing an outcome from an early biopsy with that of a later biopsy (e.g. early in training and later in training), this approach is flawed. As discussed above, early sampling biases to rapidly turning over proteins, whereas later sampling biases towards proteins with a slower turnover rate. Therefore, the protein pools represented in the synthesis rates of repeated biopsies include different proteins and cannot be compared. We recommend avoiding studies with test/re‐test designs when using D2O labelling and limiting the approach to cross‐sectional studies.

Non‐steady state must be accounted for and protein breakdown matters

An assumption for isotope studies is that the size of the protein pool of interest is constant over the experimental period. However, the goal of many studies is to understand changes in protein synthesis and breakdown that contribute to a change in muscle protein mass. Traditional approaches using labelled amino acids were typically conducted over a period of hours, where changes in protein mass were minimal. When performing measurements over days to weeks, there are many conditions, such as disuse‐induced atrophy and exercise‐induced hypertrophy, where the protein product pool is not constant over time. Our observation is that most published studies using D2O do not account for this lack of steady state. A relatively simple way to point out the error is that in a steady state, by definition, protein synthesis equals protein breakdown. When these two are equal, it would not be possible to have a change in protein mass. Therefore, during a period of muscle loss or gain, assuming a constant pool size and steady state is inappropriate.

We recently demonstrated that not accounting for a changing product pool size can substantially alter synthesis rate calculations and interpretation of the contributions of synthesis or breakdown to a change in protein mass. To account for changes in protein product pool size during non‐steady state conditions, we have developed a model (Miller et al., 2018, 2015), which is similar to others (Bederman et al., 2015; Samarel, 1991), that can express mass changes (e.g. mg day–1) as well as fractional synthesis rate. Expressing rates as percent of absolute mass can be meaningful during changes in pool size. For example, when a protein pool increases from 100 proteins to 200 proteins, an fractional synthesis rate of 10% per day differs in an absolute sense (i.e. mg proteins day–1) as the size of the pool changes. Both rates can provide valuable, but different, insight of the regulation of muscle mass (Fig. 3).

Figure 3. Accounting for changing product pool size can substantially alter interpretations of the contribution of synthesis or breakdown to change in protein mass.

Figure 3

Open in a new tab

During hypertrophy there is a gain of muscle mass and an increased product pool as designated by the larger circles on the right. In this scenario, a similar protein mass could accrue (10 proteins day–1) with differing fractional synthesis rates (yellow panel) or similar fractional synthesis rates could be measured (10% day–1) but the outcome of protein mass accrued would be larger (purple panel). These results depend on the product pool as illustrated with the pie shapes of each panel. Created with BioRender.

An additional benefit of non‐steady state equations is that rates of protein breakdown can be calculated to understand individual contributions of changes in synthesis and degradation during gain or loss of mass. There are established protocols using labelled amino acids to measure muscle protein breakdown (Biolo et al., 1994; Zhang et al., 2002). These approaches are somewhat limited for widespread use because of the need for arterial catheterization or multiple muscle biopsies. Furthermore, these approaches captured relatively short periods of time (hours) and relied on a steady state. There was a notable effort to measure protein breakdown using D2O (Holm et al., 2013). This approach required a long pre‐labelling period (over 40 days) and a loss of enrichment over time. As a practical point, the approach necessitates a steady state, but, if in a steady state, the breakdown value should equal synthesis, so why not just measure synthesis?

We provide two examples of why the consideration of a change in pool size and the ability to calculate synthesis and breakdown are important. The first has to do with our studies of disuse atrophy using the hindlimb unloading model (Lawrence et al., 2020). One primary outcome in this study was that the period of disuse impacted ribosomal content. Adult rats that were hindlimb suspended for 7 days had ∼13% lower ribosome concentrations (as measured by RNA) in gastrocnemius muscle compared to weightbearing controls (Lawrence et al., 2020). Ribosomal biogenesis was 5% lower during hindlimb unloading compared to weightbearing control, but the rate of ribosome degradation was 600% greater during hindlimb unloading. These results highlight a major, and underappreciated, potential role for ribophagy during periods of disuse.

As a second example, we highlight collagen protein turnover during denervation of mice (Kobak et al., 2021). At the end of a 7 day period of denervation, collagen concentration was higher in the denervated leg compared to control. The collagen protein synthesis rate over this period was minimally changed while collagen breakdown went to zero. Therefore, decreases in collagen protein breakdown was almost solely responsible for the greater accumulation of collagen over this period. Thus, these findings point to collagen breakdown, rather than synthesis, as a target to minimize muscle fibrosis.

In the study by Kobak et al. (2021), myofibrillar protein turnover was measured during denervation, demonstrating the importance of considering the non‐steady state. Using the steady‐state fractional synthesis rate calculations employed by others in the literature, it was calculated that denervated muscle had lower protein synthesis rates than sham or non‐surgical controls. Because a steady state means that synthesis equals breakdown, the calculated breakdown rates were lower than sham and the non‐surgical controls by the same amount. However, as mentioned above, this approach is flawed because an equal rate of myofibrillar protein synthesis and breakdown could not result in a loss of protein mass. By using non‐steady state equations, it was calculated that protein synthesis is higher in denervated vs. controls as a percent (% day–1), but less than controls when expressed as absolute mass (mg day–1). Furthermore, rates of breakdown were greater in denervated muscle than controls whether expressed as a percent or as a rate constant. These differences from the steady state assumption demonstrate that the individual contributions of protein synthesis and breakdown to a change in muscle mass requires non‐steady state equations and can be misleading when the steady state is assumed.

No approach is perfect, and our approach still has limitations. The first is that we are currently restricted to bulk turnover that is calculated by a change in mass. This approach has hindered our calculation of the turnover of subfractions such as mitochondria. We (Groennebaek et al., 2020) and others (Larsen et al., 2012) have commented on the issues of various measures of mitochondrial mass. Without a valid and reliable measure of mitochondrial protein mass, we cannot adequately calculate changes in mitochondrial protein mass; an issue we are currently working on rectifying. Furthermore, we have not yet been able to use our tracer methods to determine the separate rates of autophagy or proteasomal breakdown. In vitro, we were able to estimate the contribution of autophagy to breakdown (more correctly, half‐life) by comparing breakdown rates with and without an autophagy inhibitor (Wolff et al., 2020). However, this was imperfect in that the treatment inhibits the process we are interested in and cannot be adapted to in vivo. Long‐term in vivo measurements of the rates of autophagy and proteasomal breakdown using tracers would be impactful for the muscle field.

The number of proteins that are turning over is overlooked

We have commented on the importance of using direct measures, rather than changes in protein content, transcription, or markers of protein synthesis, to increase scientific rigor of protein synthesis assessment (Miller & Hamilton, 2012; Miller et al., 2014, 2020). For example, simply measuring changes in protein content hides the underlying dynamic processes that are important for conditions such as ageing. If protein synthesis increases by a 100‐fold and was matched by a 100‐fold increase in breakdown, there would be no change in protein concentration even though there were dramatic changes to the proteome. As important as it is to understand protein turnover in the context of proteostasis, it is not the whole story or maybe not even the most important story for some proteins during ageing.

A hallmark of the failure to maintain proteostasis with ageing is protein aggregation. Aggregates form when there is unintended congregation of proteins because of exposed hydrophobic regions from protein misfolding. Aggregates are a well‐described phenomenon in the CNS for such diseases as Alzheimer's disease and amyotrophic lateral sclerosis. Protein aggregates also form in muscle with ageing (Fuqua et al., 2023, 2019). Given the resistance of aggregates to breakdown, when aggregates form and accumulate, the proteins in the aggregates have essentially exited the pool of proteins that are turning over. There are other examples of exiting the dynamic protein pool, such as that associated with extracellular matrix and fibrosis (Abbott et al., 2021). Under these conditions, collagen cross‐linking makes the collagen resistant to protein breakdown. Importantly, standard tracer protocols where protein synthesis is calculated from product protein enrichment, the product enrichment does not distinguish pools that are resistant to tracer incorporation. As we explain below, the occurrence of protein pools resistant to turnover can hinder our understanding of protein turnover.

An assumption of protein turnover studies, which is particularly important during long‐term labelling, is that a protein pool fully renews. Zhou et al. (2015) showed that liver collagen had incomplete renewal of proteins in mice. Around the time of this publication, we began to use a time course approach (Reid et al., 2020; Wolff et al., 2021) to understand the incomplete renewal of protein pools. We considered that the time course approach could show us the fraction of a pool that is resistant to turnover as an indicator of changes in proteostasis. Our study, also in mice, used several indicators of fibrosis to validate that changes in the plateau of the fraction new of skeletal muscle collagen protein over time is indicative of proteostatic changes (Abbott et al., 2021). As suspected, the fraction of collagen that was resistant to turnover increased with age, which confirmed the decline in proteostatic maintenance with ageing. However, a surprising finding from this study was how little of the collagen in adult skeletal muscle renewed. In this study, we showed that the fraction of new collagen plateaus at a value of 18–21%, indicating that 79–82% of the protein pool is not renewing. Our assumption is that this resistance to turnover may represent an adaption for structural stability.

In the study by Abbott et al. (2021), the plateau values of collagen protein fraction new were reached before 30 days of labelling. For illustrative purposes of why this is important, the common approach was employed of calculating protein synthesis rates using a single timepoint at both 15 and 60 days. At 15 days, the collagen protein synthesis rates of the gastrocnemius were the same between the adult and old group (∼0.5% day–1). At 60 days, the adult was slightly higher than old. Therefore, the conclusion would differ depending on which labelling time point was used for the calculations. Furthermore, there was a much lower collagen synthesis rate (approximately halved) when measured at 60 days compared to 15 days. However, this apparent slowing of synthesis was simply because the fraction new (numerator) did not change (i.e. plateaued) but was divided by a greater number of days (denominator). It was only by using a time course design that the true nature of the collagen protein pool could be determined. Thus, although a study with a single timepoint provides a result, there is uncertainty about the validity of the finding. Not accounting for the fraction of the pool that is resistant to turnover can have profound effects on the interpretation of physiological mechanisms.

To illustrate how the dynamic pool can be as important as the rates of synthesis, we use a hypothetical protein pool that contains 100 proteins (Fig. 4). In this example, we show that if 50% of the proteins are synthesized at 20% day–1 (‘Aged’), you would get the same result as if 100% of the proteins were synthesized at 10% day–1 (‘Adult’). Both protein pools would have a measured synthesis rate of 10% day–1, but the proteostatic implications of the two conditions would be profoundly different. However, if an experiment was to use appropriate methods to capture both the change in protein pool size and synthesis, there are several important findings from this hypothetical case. The most obvious difference is that the number of proteins that are dynamic and adapt to stress decreases in the ‘aged’. Second, the proteins that are still turning over in the aged example are turning over at a very high rate. In this example, it may be inappropriate to say that protein synthesis is lower in the aged compared to adult, but rather that the number of proteins turning over is lower. In addition, in the context of maintaining adaptability of muscle with age, the decrease in the dynamic protein pool could be as important as changes in synthesis. Therefore, solely focusing on measuring protein synthesis to improve skeletal muscle with age may be limited for some proteins, whereas understanding the deficiencies that led to a decrease in the dynamic pool could be more helpful.

Figure 4. The dynamic pool is as important as the rates of synthesis.

Figure 4

Open in a new tab

Example of 100 proteins of adult and 100 proteins of aged muscle where purple proteins are dynamic (turning over), whereas grey proteins are resistant to turnover. The fractional synthesis rates of both adult and aged muscle, as indicated on the right, would be the same even though 50% of the aged muscle is resistant to turnover because the rates of synthesis are higher in the proteins that are dynamic. The fractional synthesis rate alone would not provide information on what is occurring during the ageing process because it does not consider the dynamic pool. Created with BioRender.

Where does this get us?

We are interested in how to maintain healthy muscle in aged individuals. When we started this line of research, we focused on muscle protein synthesis and efforts to increase synthesis rates. However, when we started using D2O labelling, our data in pre‐clinical models indicated that we perhaps were not seeing the whole picture. The first indication was our study in which we looked at rats between the ages (24–28 months) where muscle mass and function, as well as the ability to recover from the disuse atrophy, declined (Miller et al., 2019). That study showed that in some muscles the 28‐month‐old rats had higher rates of protein synthesis and RNA synthesis than the 24‐month‐old rats. In addition, there were no deficiencies in synthesis rates of the older muscles during recovery from atrophy. Over time, we developed approaches using D2O that facilitated understanding changes in proteostatic maintenance. Recently, we used these approaches, including a time course approach, tissue fractionation, proteomic analysis and separation of soluble and insoluble proteins, to understand response to reloading in older rats (Fuqua et al., 2023). The results of the study indicate that the failure of older muscle to fully recover after a period of disuse is not due to limitations in the ability to synthesize myofibrillar proteins, but rather to other proteostatic mechanisms. There is still work to be done in this area, including a further understanding of the gradual loss of mass over time (sarcopenia) and how to target alternative mechanisms to protein synthesis.

In the last 10–15 years, there has been an increase in research groups using D2O. Our impression is that sometimes the template for studies using labelled amino acids were applied to studies using D2O without consideration of how using a tracer over prolonged periods of time may introduce important considerations. It is critical to understand these important points that are inherent in tracer studies. Furthermore, it is important to realize that different questions require different approaches to address them correctly. For example, we are now of the opinion that studies of collagen turnover should always use a time course approach. Understanding the appropriate use of long‐term labelling can also be an advantage to understanding the physiology of muscle protein remodelling. It is our hope that this perspective sheds light on these important considerations for future studies.

Additional information

Competing interests

No competing interests declared.

Author contributions

C.O.'R., S.B. and B.M. were responsible for the conception or design of the work, as well as drafting the work or revising it critically for important intellectual content. All authors approved the final version of the manuscript submitted for publication. All authors agree to be accountable for all aspects of the work.

Funding

Veterans Association: Benjamin F Miller, VA I01 BX005592; Veterans Association: Sue C Bodine, VA I01 BX005626; HHS | NIH | National Institute on Aging (NIA): Colleen L O'Reilly, T32 AG052363

Supporting information

Peer Review History

TJP-603-7-s001.pdf (852.7KB, pdf)

Acknowledgements

We acknowledge Patrick Shipman PhD for deriving our non‐steady state equations. We thank Drs Matthew Bubak, Jordan Fuqua, Agnieszka Borowik and Paulo Henrique Caldeira Mesquita for their critical feedback. We acknowledge salary support from VA I01 BX005592 (BFM) and VA I01 BX005626 (SCB), and T32 AG052363 (CLO'R), and the support from our funders that led to the work described herein.

Handling Editors: Paul Greenhaff & Christopher Sundberg

The peer review history is available in the Supporting Information section of this article (https://doi.org/10.1113/JP285616#support‐information‐section).

References

  1. Abbott, C. B. , Lawrence, M. M. , Kobak, K. A. , Lopes, E. B. P. , Peelor, F. F. , Donald, E. J. , Van Remmen, H. , Griffin, T. M. , & Miller, B. F. (2021). A novel stable isotope approach demonstrates surprising degree of age‐related decline in skeletal muscle collagen proteostasis. Function (Oxford), 2(4), zqab028. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Balagopal, P. , Rooyackers, O. E. , Adey, D. B. , Ades, P. A. , & Nair, K. S. (1997). Effects of aging on in vivo synthesis of skeletal muscle myosin heavy‐chain and sarcoplasmic protein in humans. American Journal of Physiology, 273, E790–E800. [DOI] [PubMed] [Google Scholar]
  3. Balch, W. E. , Morimoto, R. I. , Dillin, A. , & Kelly, J. W. (2008). Adapting proteostasis for disease intervention. Science, 319(5865), 916–919. [DOI] [PubMed] [Google Scholar]
  4. Bederman, I R. , Lai, N. , Shuster, J. , Henderson, L. , Ewart, S. , & Cabrera, M. E. (2015). Chronic hindlimb suspension unloading markedly decreases turnover rates of skeletal and cardiac muscle proteins and adipose tissue triglycerides. Journal of Applied Physiology (1985), 119(1), 16–26. [DOI] [PubMed] [Google Scholar]
  5. Biolo, G. , Gastaldelli, A. , Zhang, X. J. , & Wolfe, R. R. (1994). Protein synthesis and breakdown in skin and muscle: A leg model of amino acid kinetics. American Journal of Physiology, 267, E467–E474. [DOI] [PubMed] [Google Scholar]
  6. Borowik, A. K. , Davidyan, A. , Peelor, F. F. , Voloviceva, E. , Doidge, S. M. , Bubak, M. P. , Mobley, C. B. , Mccarthy, J. J. , Dupont‐Versteegden, E. E. , & Miller, B. F. (2023). Skeletal muscle nuclei in mice are not post‐mitotic. Function, 4(1), zqac059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Burd, N. A. , Gorissen, S. H. , & Van Loon, L. J. C. (2013). Anabolic resistance of muscle protein synthesis with aging. Exercise and Sport Sciences Reviews, 41(3), 169–173. [DOI] [PubMed] [Google Scholar]
  8. Busch, R. , Kim, Y. , Neese, R. , Schadeserin, V. , Collins, M. , Awada, M. , Gardner, J. , Beysen, C. , Marino, M. , & Misell, L. (2006). Measurement of protein turnover rates by heavy water labeling of nonessential amino acids. Biochimica Et Biophysica Acta, 1760(5), 730–744. [DOI] [PubMed] [Google Scholar]
  9. Dos Santos, M. , Backer, S. , Saintpierre, B. , Izac, B. , Andrieu, M. , Letourneur, F. , Relaix, F. , Sotiropoulos, A. , & Maire, P. (2020). Single‐nucleus RNA‐seq and FISH identify coordinated transcriptional activity in mammalian myofibers. Nature Communications, 11(1), 5102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dufner, D. A. , Bederman, I R. , Brunengraber, D. Z. , Rachdaoui, N. , Ismail‐Beigi, F. , Siegfried, B. A. , Kimball, S. R. , & Previs, S. F. (2005). Using 2H2O to study the influence of feeding on protein synthesis: Effect of isotope equilibration in vivo vs. in cell culture. American Journal of Physiology‐Endocrinology and Metabolism, 288(6), E1277–E1283. [DOI] [PubMed] [Google Scholar]
  11. Fuqua, J. D. , Lawrence, M. M. , Hettinger, Z. R. , Borowik, A. K. , Brecheen, P. L. , Szczygiel, M. M. , Abbott, C. B. , Peelor, F. F. , Confides, A. L. , Kinter, M. , Bodine, S. C. , Dupont‐Versteegden, E. E. , & Miller, B. F. (2023). Impaired proteostatic mechanisms other than decreased protein synthesis limit old skeletal muscle recovery after disuse atrophy. Journal of Cachexia Sarcopenia Muscle, 14(5), 2076–2089. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Fuqua, J. D. , Mere, C. P. , Kronemberger, A. , Blomme, J. , Bae, D. , Turner, K. D. , Harris, M. P. , Scudese, E. , Edwards, M. , Ebert, S. M. , Sousa, L. G. O. , Bodine, S. C. , Yang, L. , Adams, C. M. , & Lira, V. A. (2019). ULK2 is essential for degradation of ubiquitinated protein aggregates and homeostasis in skeletal muscle. Federation of American Societies for Experimental Biology Journal, 33(11), 11735–12745. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gasier, H. G. , Riechman, S. E. , Wiggs, M. P. , Previs, S. F. , & Fluckey, J. D. (2009). A comparison of 2H2O and phenylalanine flooding dose to investigate muscle protein synthesis with acute exercise in rats. American Journal of Physiology‐Endocrinology and Metabolism, 297(1), E252–E259. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Groennebaek, T. , Nielsen, J. , Jespersen, N. R. , Bøtker, H. E. , De Paoli, F. V. , Miller, B. F. , & Vissing, K. (2020). Utilization of biomarkers as predictors of skeletal muscle mitochondrial content after physiological intervention and in clinical settings. American Journal of Physiology‐Endocrinology and Metabolism, 318(6), E886–E889. [DOI] [PubMed] [Google Scholar]
  15. Guillet, C. , Prod'homme, M. , Balage, M. , Gachon, P. , Giraudet, C. , Morin, L. , Grizard, J. , & Boirie, Y. (2004). Impaired anabolic response of muscle protein synthesis is associated with S6K1 dysregulation in elderly humans. Federation of American Societies for Experimental Biology Journal, 18(13), 1586–1587. [DOI] [PubMed] [Google Scholar]
  16. Hipp, M. S. , Kasturi, P. , & Hartl, F. U. (2019). The proteostasis network and its decline in ageing. Nature Reviews Molecular Cell Biology, 20(7), 421–435. [DOI] [PubMed] [Google Scholar]
  17. Holm, L. , O'rourke, B. , Ebenstein, D. , Toth, M. J. , Bechshoeft, R. , Holstein‐Rathlou, N.‐H. , Kjaer, M. , & Matthews, D. E. (2013). Determination of steady‐state protein breakdown rate in vivo by the disappearance of protein‐bound tracer‐labeled amino acids: A method applicable in humans. American Journal of Physiology‐Endocrinology and Metabolism, 304(8), E895–E907. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Jayaraj, G. G. , Hipp, M. S. , & Hartl, F. U. (2020). Functional modules of the proteostasis network. Cold Spring Harbor Perspectives in Biology, 12(1), a033951. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Joseph, G. A. , Wang, S. X. , Jacobs, C. E. , Zhou, W. , Kimble, G. C. , Tse, H. W. , Eash, J. K. , Shavlakadze, T. , & Glass, D. J. (2019). Partial inhibition of mTORC1 in aged rats counteracts the decline in muscle mass and reverses molecular signaling associated with sarcopenia. Molecular and Cellular Biology, 39(19), e00141. 19. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Kimball, S. R. , O'malley, J P. , Anthony, J. C. , Crozier, S. J. , & Jefferson, L. S. (2004). Assessment of biomarkers of protein anabolism in skeletal muscle during the life span of the rat: Sarcopenia despite elevated protein synthesis. American Journal of Physiology‐Endocrinology and Metabolism, 287(4), E772–E780. [DOI] [PubMed] [Google Scholar]
  21. Kobak, K. A. , Lawrence, M. M. , Pharaoh, G. , Borowik, A. K. , Peelor, F. F. , Shipman, P. D. , Griffin, T. M. , Van Remmen, H. , & Miller, B. F. (2021). Determining the contributions of protein synthesis and breakdown to muscle atrophy requires non‐steady‐state equations. Journal of Cachexia Sarcopenia Muscle, 12(6), 1764–1775. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Larsen, S. , Nielsen, J. , Hansen, C. N. , Nielsen, L. B. , Wibrand, F. , Stride, N. , Schroder, H. D. , Boushel, R. , Helge, J. W. , Dela, F. , & Hey‐Mogensen, M. (2012). Biomarkers of mitochondrial content in skeletal muscle of healthy young human subjects. The Journal of Physiology, 590(14), 3349–3360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lawrence, M. M. , Van Pelt, D. W. , Confides, A. L. , Hunt, E. R. , Hettinger, Z. R. , Laurin, J. L. , Reid, J. J. , Peelor, F. F. , Butterfield, T. A. , Dupont‐Versteegden, E. E. , & Miller, B. F. (2020). Massage as a mechanotherapy promotes skeletal muscle protein and ribosomal turnover but does not mitigate muscle atrophy during disuse in adult rats. Acta Physiology (Oxford), 229(3), e13460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. López‐Otín, C. , Blasco, M. A. , Partridge, L. , Serrano, M. , & Kroemer, G. (2013). The hallmarks of aging. Cell, 153(6), 1194–1217. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Markofski, M. M. , Dickinson, J. M. , Drummond, M. J. , Fry, C. S. , Fujita, S. , Gundermann, D. M. , Glynn, E. L. , Jennings, K. , Paddon‐Jones, D. , Reidy, P. T. , Sheffield‐Moore, M. , Timmerman, K. L. , Rasmussen, B. B. , & Volpi, E. (2015). Effect of age on basal muscle protein synthesis and mTORC1 signaling in a large cohort of young and older men and women. Experimental Gerontology, 65, 1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Miller, B. F. , Baehr, L. M. , Musci, R. V. , Reid, J. J. , Peelor, F. F. , Hamilton, K. L. , & Bodine, S. C. (2019). Muscle‐specific changes in protein synthesis with aging and reloading after disuse atrophy. Journal of Cachexia, Sarcopenia and Muscle, 10(6), 1195–1209. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Miller, B. F. , Drake, J. C. , Naylor, B. , Price, J. C. , & Hamilton, K. L. (2014). The measurement of protein synthesis for assessing proteostasis in studies of slowed aging. Ageing Research Reviews, 18, 106–111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Miller, B. F. , & Hamilton, K. L. (2012). A perspective on the determination of mitochondrial biogenesis. American Journal of Physiology‐Endocrinology and Metabolism, 302(5), E496–E499. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Miller, B. F. , Hamilton, K. L. , Majeed, Z. R. , Abshire, S. M. , Confides, A. L. , Hayek, A. M. , Hunt, E. R. , Shipman, P. , Peelor, F. F. , Butterfield, T. A. , & Dupont‐Versteegden, E. E. (2018). Enhanced skeletal muscle regrowth and remodelling in massaged and contralateral non‐massaged hindlimb. The Journal of Physiology, 596(1), 83–103. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Miller, B. F. , Reid, J. J. , Price, J. C. , Lin, H.‐J. L. , Atherton, P. J. , & Smith, K. (2020). CORP: The use of deuterated water for the measurement of protein synthesis. Journal of Applied Physiology (1985), 128(5), 1163–1176. [DOI] [PubMed] [Google Scholar]
  31. Miller, B. F. , Wolff, C. A. , Peelor, F. F. , Shipman, P. D. , & Hamilton, K. L. (2015). Modeling the contribution of individual proteins to mixed skeletal muscle protein synthetic rates over increasing periods of label incorporation. Journal of Applied Physiology (1985), 118(6), 655–661. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Nunes, E. A. , Stokes, T. , Mckendry, J. , Currier, B. S. , & Phillips, S. M. (2022). Disuse‐induced skeletal muscle atrophy in disease and nondisease states in humans: mechanisms, prevention, and recovery strategies. American Journal of Physiology‐Cell Physiology, 322(6), C1068–C1084. [DOI] [PubMed] [Google Scholar]
  33. Orchard, P. , Manickam, N. , Ventresca, C. , Vadlamudi, S. , Varshney, A. , Rai, V. , Kaplan, J. , Lalancette, C. , Mohlke, K. L. , Gallagher, K. , Burant, C. F. , & Parker, S. C. J. (2021). Human and rat skeletal muscle single‐nuclei multi‐omic integrative analyses nominate causal cell types, regulatory elements, and SNPs for complex traits. Genome Research, 31(12), 2258–2275. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Perez, K. , Ciotlos, S. , McGirr, J. , Limbad, C. , Doi, R. , Nederveen, J. P. , Nilsson, M. I. , Winer, D. A. , Evans, W. , Tarnopolsky, M. , Campisi, J. , & Melov, S. (2022). Single nuclei profiling identifies cell specific markers of skeletal muscle aging, frailty, and senescence. Aging (Albany NY), 14, 9393–9422. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Pride, H. , Yu, Z. , Sunchu, B. , Mochnick, J. , Coles, A. , Zhang, Y. , Buffenstein, R. , Hornsby, P. J. , Austad, S. N. , & Pérez, V. I. (2015). Long‐lived species have improved proteostasis compared to phylogenetically‐related shorter‐lived species. Biochemical and Biophysical Research Communications, 457(4), 669–675. [DOI] [PubMed] [Google Scholar]
  36. Reid, J. J. , Linden, M. A. , Peelor, F. F. , Miller, R. A. , Hamilton, K. L. , & Miller, B. F (2020). Brain protein synthesis rates in the UM‐HET3 mouse following treatment with rapamycin or rapamycin with metformin. Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 75(1), 40–49. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Richardson, N. E. , Konon, E. N. , Schuster, H. S. , Mitchell, A. T. , Boyle, C. , Rodgers, A. C. , Finke, M. , Haider, L. R. , Yu, D. , Flores, V. , Pak, H. H. , Ahmad, S. , Ahmed, S. , Radcliff, A. , Wu, J. , Williams, E. M. , Abdi, L. , Sherman, D. S. , Hacker, T. A. , & Lamming, D. W. (2021). Lifelong restriction of dietary branched‐chain amino acids has sex‐specific benefits for frailty and lifespan in mice. Nature Aging, 1(1), 73–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Robinson, M. M. , Turner, S. M. , Hellerstein, M. K. , Hamilton, K. L. , & Miller, B. F. (2011). Long‐term synthesis rates of skeletal muscle DNA and protein are higher during aerobic training in older humans than in sedentary young subjects but are not altered by protein supplementation. Federation of American Societies for Experimental Biology Journal, 25(9), 3240–3249. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rooyackers, O.  E. , Adey, D.  B. , Ades, P. A. , & Nair, K.  S. (1996). Effect of age on in vivo rates of mitochondrial protein synthesis in human skeletal muscle. The Proceedings of the National Academy of Sciences, 93(26), 15364–15369. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Samarel, A. M. (1991). In vivo measurements of protein turnover during muscle growth and atrophy. Federation of American Societies for Experimental Biology Journal, 5(7), 2020–2028. [DOI] [PubMed] [Google Scholar]
  41. Tang, H. , Inoki, K. , Brooks, S. V. , Okazawa, H. , Lee, M. , Wang, J. , Kim, M. , Kennedy, C. L. , Macpherson, P. C. D. , Ji, X. , Van Roekel, S. , Fraga, D. A. , Wang, K. , Zhu, J. , Wang, Y. , Sharp, Z. D. , Miller, R. A. , Rando, T. A. , Goldman, D. , Guan, K.‐L. , & Shrager, J. B. (2019). mTORC1 underlies age‐related muscle fiber damage and loss by inducing oxidative stress and catabolism. Aging Cell, 18(3), e12943. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Taylor, R. C. , & Dillin, A. (2011). Aging as an event of proteostasis collapse. Cold Spring Harbor Perspectives in Biology, 3(5), a004440. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Walker, L. C. , & Levine, H. (2012). Corruption and spread of pathogenic proteins in neurodegenerative diseases. Journal of Biological Chemistry, 287(40), 33109–33115. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. White, Z. , White, R. B. , Mcmahon, C. , Grounds, M. D. , & Shavlakadze, T. (2016). High mTORC1 signaling is maintained, while protein degradation pathways are perturbed in old murine skeletal muscles in the fasted state. International Journal of Biochemistry & Cell Biology 78, 10–21. [DOI] [PubMed] [Google Scholar]
  45. Wilkinson, D. J. , Franchi, M. V. , Brook, M. S. , Narici, M. V. , Williams, J. P. , Mitchell, W. K. , Szewczyk, N. J. , Greenhaff, P. L. , Atherton, P. J. , & Smith, K. (2014). A validation of the application of D(2)O stable isotope tracer techniques for monitoring day‐to‐day changes in muscle protein subfraction synthesis in humans. American Journal of Physiology‐Endocrinology and Metabolism, 306(5), E571–E579. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Wolff, C. A. , Lawrence, M. M. , Porter, H. , Zhang, Q. , Reid, J. J. , Laurin, J. L. , Musci, R. V. , Linden, M. A. , Peelor, F. F. , Wren, J. D. , Creery, J. S. , Cutler, K. J. , Carson, R. H. , Price, J. C. , Hamilton, K. L. , & Miller, B. F. (2021). Sex differences in changes of protein synthesis with rapamycin treatment are minimized when metformin is added to rapamycin. GeroScience, 43(2), 809–828. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Wolff, C. A. , Reid, J. J. , Musci, R. V. , Bruns, D. R. , Linden, M. A. , Konopka, A. R. , Peelor, F. F. , Miller, B. F. , & Hamilton, K. L (2020). Differential effects of rapamycin and metformin in combination with rapamycin on mechanisms of proteostasis in cultured skeletal myotubes. Journals of Gerontology. Series A, Biological Sciences and Medical Sciences, 75(1), 32–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Zhang, X.‐J. , Chinkes, D. L. , & Wolfe, R. R. (2002). Measurement of muscle protein fractional synthesis and breakdown rates from a pulse tracer injection. American Journal of Physiology‐Endocrinology and Metabolism, 283(4), E753–E764. [DOI] [PubMed] [Google Scholar]
  49. Zhou, H. , Wang, S.‐P. , Herath, K. , Kasumov, T. , Sadygov, R. G. , Previs, S. F. , & Kelley, D. E. (2015). Tracer‐based estimates of protein flux in cases of incomplete product renewal: evidence and implications of heterogeneity in collagen turnover. American Journal of Physiology‐Endocrinology and Metabolism, 309(2), E115–E121. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Peer Review History

TJP-603-7-s001.pdf (852.7KB, pdf)

Articles from The Journal of Physiology are provided here courtesy of Wiley

RESOURCES