Effects of prostaglandins and COX-inhibiting drugs on skeletal muscle adaptations to exercise

Abstract

It has been ∼40 yr since the discovery that PGs are produced by exercising skeletal muscle and since the discovery that inhibition of PG synthesis is the mechanism of action of what are now known as cyclooxygenase (COX)-inhibiting drugs. Since that time, it has been established that PGs are made during and after aerobic and resistance exercise and have a potent paracrine and autocrine effect on muscle metabolism. Consequently, it has also been determined that orally consumed doses of COX inhibitors can profoundly influence muscle PG synthesis, muscle protein metabolism, and numerous other cellular processes that regulate muscle adaptations to exercise loading. Although data from acute human exercise studies, as well as animal and cell-culture data, would predict that regular consumption of a COX inhibitor during exercise training would dampen the typical muscle adaptations, the chronic data do not support this conjecture. From the studies in young and older individuals, lasting from 1.5 to 4 mo, no interfering effects of COX inhibitors on muscle adaptations to resistance-exercise training have been noted. In fact, in older individuals, a substantial enhancement of muscle mass and strength has been observed. The collective findings of the PG/COX-pathway regulation of skeletal muscle responses and adaptations to exercise are compelling. Considering the discoveries in other areas of COX regulation of health and disease, there is certainly an interesting future of investigation in this re-emerging area, especially as it pertains to older individuals and the condition of sarcopenia, as well as exercise training and performance of individuals of all ages.

pgs produced by the cyclooxygenase (COX) enzyme are ubiquitous in human physiology and regulate numerous processes (105, 108), including muscle protein metabolism (78, 93, 106, 126, 128). As a result, PGs regulate adaptations to muscular exercise, and COX-inhibiting drugs can alter the acute and chronic responses to exercise. Because of the continuum of acute responses and chronic adaptations that exists from lower-intensity, longer-duration exercise to higher-intensity, shorter-duration exercise, it is difficult to adopt a “one-mechanism-fits-all” approach to PG and COX involvement in skeletal muscle exercise adaptations. This is also true when distinguishing between muscle injury and muscular exercise for health and performance.

COX-inhibiting drugs are one of the most commonly consumed classes of drugs in the world. In the United States, the COX inhibitors, acetaminophen, ibuprofen, and aspirin, are the top three consumed drugs of young, middle-aged, and older individuals, with ∼50 million individuals using each of these drugs during any given week (52, 129). Interestingly, two of these drugs (acetaminophen and aspirin) have been used for over 100 yr, originating in the 1800s (4). Understanding the role of PGs in skeletal muscle adaptation to activity is important because of the widespread use and potential influence of COX-inhibiting drugs, for better or worse, and so we can understand more completely the mechanisms that control muscle adaptation.

The focus of this review will be on those studies that have used COX inhibition in human studies of skeletal muscle metabolism and adaptations to exercise. Distinction between studies using exercise paradigms that would be used for health benefits and those more focused on injury will be made when necessary and when appropriate animal and cell-culture studies will be discussed. A historical overview of the PG and COX pathway, with specific reference to skeletal muscle metabolism, will also be presented.

PGs, THE COX PATHWAY, AND SKELETAL MUSCLE

Historical context.

PGs were discovered in the 1930s from extracts of the prostate gland and named by von Euler (136). Subsequent studies, over the next 40 yr, elucidated many of the tenets of PG structural biology and physiology, resulting in the Nobel Prize being awarded to Bergström, Samuelsson, and Vane in 1982 (10, 97, 135). There are numerous PGs (e.g., PGA-J) that are produced from arachidonic acid (53), which is liberated from the membrane of cells by PLA₂ (19, 32, 38) (Fig. 1). The enzyme PG G/H synthase, more commonly known as COX, is a dual-function enzyme that converts arachidonic acid to PGG₂ and then PGH₂ (105, 107), which is then rapidly converted to a specific PG (e.g., PGD₂, PGE₂, PGF_2α, PGI₂) by PG synthases (70, 107, 137). Additionally, there are two well-known isoforms of COX—COX-1 and COX-2 (44, 105, 107, 141)—and a possible third isoform (an intron-retaining variant of COX-1, referred to as COX-1b or COX-3) may exist in some tissues (22, 86, 101, 138).

Fig. 1.

General schematic of the biosynthesis of the primary PGs in the PG/cyclooxygenase (COX) pathway and their associated receptors (53, 107, 108). The thicker arrows reflect conversions that are catalyzed by specific enzymes. Several other PGs are also made from the conversion (enzymatic or nonenzymatic) of those listed here: PGJ₂, Δ¹²-PGJ₂, 15-deoxy-Δ^12,14-PGJ₂, and 11-epi-PGF_2α are derived from PGD₂; PGA₂, PGB₂, and PGC₂ are derived from PGE₂; 15-keto-PGF_2α is derived from PGF_2α; 6-keto-PGF_1α and 6-keto-PGF₁ are derived from PGI₂. PGs are lipid molecules with a general chemical formula of C₂₀H_28–34O_3–6 and molecular weight range of 317–371 Da. See text for specific references related to the numerous aliases that exist for the variants and isoforms of the enzymes and receptors and for further understanding of the nomenclature. Although not a PG, thromboxane A₂ is also derived from the enzymatic conversion of PGH₂. This review focuses on the 2 primary PGs that influence skeletal muscle protein turnover (PGF_2α and PGE₂), which are discussed in the text and presented in further detail (see Fig. 2). DP1–2, EP1–4, FP, and IP, PG receptors.

PGs work in an autocrine and paracrine fashion through receptors specific to each PG, some of which have multiple isoforms (1, 17, 32, 62, 72, 112). PGs are relatively transient molecules with a half-life, typically, of only seconds to minutes (32). For example, ∼90% of PGE₂ and PGF_2α is removed from the blood in one pass through the pulmonary circulation (85). PGs are also potent in relatively small amounts (9, 32). As a result, circulating PG levels are typically low, and in response to stimulation, tissue production can be increased tremendously [e.g., 10 times the total resting tissue content can be generated every minute (135)].

The first studies of PG production and release by skeletal muscle in response to muscular work were published by Herbaczynska-Cedro and colleagues (41–43, 47) in 1974 and 1976 (Table 1). These studies were focused on identifying muscle-produced vasodilators and blood flow regulation during simulated exercise in dogs and showed that more than one PG was produced by exercising skeletal muscle (suggested as at least PGE₂ and PGF_2α), which could be eliminated by the COX inhibitor indomethacin. These authors also speculated that the mechanism of the PG release was the distortion of the muscle cell membrane but showed that the PG release occurred during and after the exercise bouts. Soon thereafter, a study in humans using a static and dynamic exercise forearm model showed similar results (54). Subsequent arteriovenous studies by Nowak and Wennmalm (75) showed no net release or uptake of PGs across the leg at rest, but cycling exercise at ∼75% maximal oxygen consumption substantially increased net release of PGs from the leg. Whereas PGs were reported to be in resting human skeletal muscle as early as 1967 (51), Berlin et al. (11) showed in 1979, with radiolabeled arachidonic acid added to homogenates of skeletal muscle biopsy samples, that skeletal muscle could produce PGD₂, PGE₂, PGF_2α, and PGI₂. PGE₂ was the predominant PG produced (11), but this could be shifted to PGF_2α if the assay environment was altered (74).

Table 1.

Noteworthy studies of PGs, skeletal muscle, and exercise adaptations

Author	Year	Species	Key Findings
Early studies
Karim et al. (51)	1967	Humans	PGs are located in skeletal muscle.
Herbaczynska-Cedro and colleagues (41–43, 47)	1974, 1976	Dogs	Skeletal muscle PG release increased during and after simulated exercise; eliminated with COX inhibitor.
Kilbom and Wennmalm (54)	1976	Humans	Forearm skeletal muscle PG release increased during and after static and dynamic exercise; reduced with COX inhibitor (indomethacin).
Nowak and Wennmalm (75)	1978	Humans	Leg skeletal muscle PG release increased during cycling exercise at 75% maximal oxygen consumption.
Berlin et al. (11)	1979	Humans	Skeletal muscle had the enzymatic capacity to produce PGD2, PGE2, PGF2a, and PGI2 from arachidonic acid.
Young et al. (142)	1981	Monkeys	Aging increased skeletal muscle PG production.
Rodemann and Goldberg (93)	1982	Rats	Arachidonic acid increased skeletal muscle PGF2a and PGE2, which increased muscle protein synthesis and degradation, respectively; reduced or eliminated with COX inhibitors.
Palmer et al. (78)	1983	Rabbits	Simulated exercise (intermittent stretching) increased skeletal muscle PGF2a and protein synthesis; reduced or eliminated with COX inhibitors. PGF2a and protein synthesis levels correlated.
Smith et al. (106)
Gibson et al. (33)	1991	Humans	Reduced skeletal muscle PGF2a levels associated with reduced muscle protein synthesis and type I and II muscle fiber size.
Contemporary studies
Trappe et al. (126, 128)	2001, 2002	Humans	Increased skeletal muscle PGF2a, PGE2, and protein synthesis after resistance exercise; eliminated with over-the-counter, orally consumed COX inhibitors (acetaminophen and ibuprofen).
Karamouzis et al. (49, 50)	2001	Humans	Aerobic exercise increased intramuscular PGE2; increased production with increased workload.
Boushel et al. (15)	2002	Humans	Confirmed findings of Karamouzis et al. (49, 50) and showed that an oral dose of COX inhibitor (indomethacin) reduced intramuscular PGE2 during aerobic exercise by 90%.
Höffner et al. (45)	2003	Humans	An oral dose of COX inhibitor (aspirin) reduced intramuscular PGE2 production at rest by nearly 90% within 1 h.
Trappe et al. (123)	2006	Humans	Both young and old individuals increased intramuscular PGF2a in the hours after resistance exercise; no apparent effect of age on resting or postexercise levels.
Mikkelsen et al. (67)	2008	Humans	Local intramuscular low-dose COX inhibitor (indomethacin) delivery blocked PGE2 production by ∼85% during and after resistance exercise.
Burd et al. (18)	2010	Humans	COX-2-specific inhibitor (celecoxib) did not eliminate or reduce the increase in skeletal muscle protein synthesis after resistance exercise.
Paulsen et al. (79)	2010	Humans	COX-2-specific inhibitor (celecoxib) did not influence intramuscular PGE2 levels or satellite cell activity after resistance exercise.
Petersen et al. (82)	2011	Humans	COX inhibitor (ibuprofen) did not influence skeletal muscle protein synthesis 24 h after aerobic exercise in older osteoarthritic patients.
Kreiner and Galbo (55)	2011	Humans	Resting intramuscular PGE2 levels 20 times higher than plasma.
Standley et al. (110)	2013	Humans	PGE2 stimulated transcription of the skeletal muscle mass regulators IL-6 and muscle RING finger-1.
COX inhibitor effects on chronic exercise adaptations
Krentz et al. (56)	2008	Humans	Duration: 6 wk; Training: resistance exercise 2–3 days/wk; Drug dose: ibuprofen 400 mg, 2–3 days/wk (training days); Participants: 24 yr, men and women, resistance exercise trained (∼6 yr); no effect on muscle mass or strength adaptations compared with placebo-consuming resistance exercise group
Petersen et al. (81)	2011	Humans	Duration: 12 wk; Training: resistance exercise 3 days/wk; Drug dose: ibuprofen 1,200 mg/day; Participants: 50–70 yr, men and women, knee osteoarthritis patients; no effect on muscle mass, increased muscle strength compared with placebo-consuming resistance exercise group
Trappe et al. (125, 127)	2011, 2013	Humans	Duration: 12 wk; Training: resistance exercise 3 days/wk;
			Drug dose: acetaminophen 4 g/day or ibuprofen 1,200 mg/day; Participants: 60–78 yr, men and women, healthy, untrained; enhanced muscle mass and strength gains 25–50% above placebo-consuming resistance exercise group
Jankowski et al. (48)	2012	Humans	Duration: 16 wk; Training: resistance exercise ∼3 days/wk; Drug dose: acetaminophen 1 g, 3 days/wk (training days); Participants: 64 yr, men, healthy, untrained; no effect on fat-free mass or muscle-strength adaptations compared with placebo-consuming resistance exercise group

COX, cyclooxygenase.

The initial reports of PG regulation of skeletal muscle protein turnover were in the early 1980s when Rodemann and Goldberg (93) showed that arachidonic acid supplementation to rat muscle in vitro increased muscle protein synthesis and degradation, which could be replicated with PGF_2α and PGE₂ supplementation, respectively, and reduced or eliminated with COX inhibition. Arachidonic acid supplementation increased muscle production of both PGs, but about twice as much PGE₂ was synthesized compared with PGF_2α. At the same time, Palmer and colleagues (78, 106) showed that with simulated exercise in rabbit muscle, PGF_2α was produced by intermittent muscle stretch (+105%), which in turn, stimulated muscle protein synthesis (+70%), both of which were reduced or eliminated by COX inhibition. In addition, the increase in muscle PGF_2α production and protein synthesis was sustained following the cessation of simulated exercise.

Over the next 10–15 yr following these seminal studies, many investigations in animal and cell-culture models built on these findings (5, 7, 34, 35, 39, 40, 64, 65, 77, 90, 94, 113, 132–134). Although the animal- and cell-culture data were compelling with regard to PG and COX-inhibitor regulation of skeletal muscle protein turnover, very few studies in humans were completed during this time, and none focused on exercise. Gibson et al. (33) showed that women (mean age of 58 yr) with rheumatoid arthritis receiving steroid treatment (which lowers PG production via COX-independent PLA₂ inhibition; mean treatment time: 8 yr) had reduced muscle PGF_2α levels, muscle protein-synthesis rates, and type I and II muscle fiber size compared with a group of controls (mean age of 70 yr). In addition, the rheumatoid arthritis patients had elevated intramuscular PGE₂, resulting in ∼12 times more PGE₂ than PGF_2α compared with approximately three times in the controls. McNurlan et al. (65) showed that a single dose of the COX inhibitor indomethacin was unable to influence whole-body protein synthesis in the fed and fasted states, but given that muscle protein turnover constitutes <30% of whole-body turnover (71), it is unclear if there was an influence on muscle protein turnover (as they had shown in rats).

Contemporary information.

What is known about PG and COX-inhibiting drug regulation of muscle protein turnover with exercise in humans has been obtained since ∼2000 (Table 1). Several studies using animal models of simulated exercise and muscle growth have also been completed during this time period and have provided additional insight. Also during this time frame, the amount of basic information about the PG/COX pathway has increased substantially, including characterization of the structure and enzymology of the PG synthases (107), leading the way for new drug development (8, 96). Much more has also been learned from continued investigation of the COX-specific inhibitors engineered in the 1990s and of many of the “classic” COX inhibitors in other areas of health and disease.

Studies completed during the 1990s using isotopically labeled amino acids showed that muscle protein synthesis increased after resistance exercise for up to 48 h in humans (23, 58, 84), and the accumulation of these acute increases in muscle protein synthesis was generally understood to be, at least in part, the underlying basis of muscle hypertrophy. Perturbations (e.g., pharmaceutical, nutrition, a specific exercise regimen) that altered this response would alter the benefits of strength training. With the consideration of the aforementioned studies showing that PGF_2α was a regulator of muscle protein synthesis, it was realized that orally consumed COX inhibitors might interfere with the normal muscle protein-synthesis response to resistance exercise. This interfering effect might be particularly important after damaging exercise that caused muscle soreness, which is, at least in part, a result of increases in intramuscular PGE₂ (3, 55, 66) and would further increase the likelihood of COX-inhibitor consumption. Indeed, this was shown to be the case when individuals consumed an over-the-counter dose of the COX inhibitor acetaminophen (4 g/day) or ibuprofen (1.2 g/day) in the 24 h following a single bout of excessive resistance exercise [10–14 sets of 10 eccentric (lowering) contractions; thus the work equivalent of ∼70 typical repetitions] that caused muscle soreness (126, 128). Muscle protein synthesis (+76%) and intramuscular PGF_2α (+77%) were increased significantly, 24 h after the exercise bout, and these increases were eliminated with the consumption of the COX-inhibiting drugs. Intramuscular PGE₂ was also increased (+64%), 24 h after the exercise, and this increase was also eliminated with COX inhibition. This study presented some interesting initial findings regarding PG and COX regulation of skeletal muscle responses to exercise and raised several interesting questions. First, the continued, increased production of intramuscular PGF_2α and PGE₂, 24 h after a single bout of resistance exercise, considering the short half-life of PGs, highlighted that it would be important to understand the typical PG levels in human skeletal muscle and how they are affected by exercise (i.e., time course, magnitude, type of exercise). Second, considering the pharmacokinetic and metabolic complexities of orally consumed COX inhibitors, it was interesting that doses of the most commonly consumed COX inhibitors could apparently produce intramuscular drug levels that inhibited intramuscular COX to the extent that PGF_2α and PGE₂ production and muscle protein metabolism after exercise were inhibited. Given the large number of COX-inhibiting drugs available for human consumption, it followed that knowledge regarding the efficacy of these drugs and doses in relation to the COX isoforms found in skeletal muscle at rest and after exercise would be necessary. Third, the obvious question from this investigation was: what are the impacts of chronic COX-inhibitor consumption on exercise-training adaptations? Whereas there is much yet to be discovered in this area, several studies have added to our understanding and started to address some of the key questions.

Intramuscular PGs and exercise.

Muscle biopsy-derived measures of PGs that regulate protein turnover show PGE₂ to be the dominant PG in young and old human muscle (11, 33, 51, 74, 126). Intramuscular PGE₂ levels are three to four times higher than PGF_2α at rest and after resistance exercise in young individuals, suggesting equivalent, relative increases in both PGs after exercise (126). This same ratio of PGE₂ to PGF_2α appears to hold in resting skeletal muscle of older individuals as well (33). However, studies in skeletal muscle of rhesus monkeys show that PG production is approximately twofold higher in older muscle compared with young (142). Several studies in humans using the microdialysis technique to sample the muscle interstitial fluid provide additional information on intramuscular levels of the PGs (primarily PGE₂). Few studies report both intramuscular and plasma levels of PGs, but it has been shown that resting vastus lateralis and trapezius intramuscular PGE₂ levels are ∼20 times higher than plasma levels (55), albeit in older individuals. However, the intramuscular levels reported in that study are similar to those reported in middle-aged and younger individuals (29, 49, 50). In addition, low-level muscular work does not increase PGE₂ levels (30, 50), whereas resistance and aerobic exercise stimulate the muscle to produce PGE₂ and/or PGF_2α during and/or after exercise (15, 49, 50, 67, 123, 126). Furthermore, increasing the workload during aerobic exercise increases the muscle production of PGE₂ (15, 49). Specifically, going from rest to cycling exercise at 100 W and 150 W increased intramuscular PGE₂ levels by ∼400% and ∼600%, respectively (49). No information is available on the possible workload effects on PG production following resistance exercise; however, there does not appear to be an age effect on PGF_2α production during the 24 h following resistance exercise (123).

COX inhibitors and skeletal muscle COX.

There are somewhat limited data on PG synthesis and muscle protein metabolism at rest or after exercise in response to oral consumption of over-the-counter COX inhibitors (82, 126, 128). Mikkelsen et al. (67) have shown that an intramuscular infusion of a relatively low dose of the prescription COX inhibitor indomethacin reduced PGE₂ levels in the muscle during and after resistance exercise by ∼85%. This same COX-inhibitor delivery approach during and for a few hours after resistance exercise did not influence muscle protein synthesis measured near the infusion site 24 h later (69). Furthermore, numerous studies of PG regulation of skeletal muscle blood flow at rest and in response to exercise have used over-the-counter- and prescription-strength COX inhibitors (15, 45, 100), and these studies provide some additional insight. For example, Höffner et al. (45) showed that a single oral dose of 1,000 mg acetylsalicylic acid (aspirin), an irreversible inhibitor of COX, inhibits resting skeletal muscle PGE₂ production by nearly 90% within 1 h. In addition, a single oral dose (100 mg) of indomethacin given to individuals, 16 h before exercise, eliminated 90% of the intramuscular PGE₂ production during aerobic exercise (15). Clearly, doses typical for human consumption can impact intramuscular PG production, but more data are needed in this area and in the context of muscle protein turnover.

The efficacy of these over-the-counter and prescription COX inhibitors should also be considered in the context of the COX enzymes found in skeletal muscle (Table 2). That is, COX-inhibiting drugs are commonly classified based on their specificity toward the two main isoforms of COX—COX-1 and COX-2 (24, 105). The aforementioned drugs that have been shown to reduce intramuscular PG production in humans (acetaminophen, ibuprofen, indomethacin, aspirin) are all generally considered to be nonspecific COX inhibitors (i.e., they block both COX-1 and COX-2 to some degree). Confusion can arise, however, if this classification is considered to hold across all tissues and physiological conditions. That is, different tissues under different stresses express different levels of the COX isoforms and have different cellular environments that appear to affect drug efficacy. In healthy individuals at rest and following exercise, human skeletal muscle expresses COX-1 almost exclusively at the transcript and protein level (18, 125, 138). Interestingly, there is a relatively ignored variant of the COX-1 isoform that is the most abundant transcript in human skeletal muscle, and it is responsive to exercise (18, 125, 138). COX-2 transcript levels in healthy human skeletal muscle at rest or after exercise are very low, and similarly, the amount of detectable, enzymatically active COX-2 protein is questionable (18, 116, 125, 138). However, intramuscular COX-2 transcript levels do increase in response to COX-2-specific and nonspecific COX inhibitors (18, 69). The COX-3 (i.e., COX-1b) isoform in human skeletal muscle has been ruled out as a contributor to PG production (18, 125, 138). Animal models of muscle adaptation (13, 14, 27, 73, 102–104, 109) suggest that COX-2 plays a substantial role in PG production in skeletal muscle, but these models appear to be more reflective of muscle injury and not necessarily human exercise (18, 125). Overall, the COX-2 isoform in skeletal muscle appears to be more responsive to injury-related stimuli, which is consistent with the large induction of skeletal muscle COX-2 protein levels in humans with septic myopathy (87). Further support for this notion comes from two separate studies showing that a COX-2-specific inhibitor designed for human consumption did not influence the skeletal muscle responses to a single bout of eccentric resistance exercise (18, 79). Burd et al. (18) showed that the COX-2 inhibitor was unable to block the normal increase in muscle protein synthesis following exercise, as was shown previously with two different, nonspecific COX inhibitors (128). Similarly, Paulsen et al. (79) were unable to show an effect of the COX-2 inhibitor on intramuscular PGE₂ levels, satellite cell activity, or autologous-radiolabeled leukocyte accumulation.

Table 2.

COX enzymes in healthy human skeletal muscle in relation to exercise and COX-inhibiting drugs

Isoform	Variant(s)	Comments
COX-1	Variant 1 (−1v1)	Relatively abundant at the transcript and protein level at rest and after acute and chronic exercise. The protein product commonly believed to interact with nonspecific COX-inhibiting drugs. Acute exercise increases transcript levels; chronic exercise training increases transcript and protein levels.
COX-1	Variant 2 (−1v2)	A truncated transcript of COX-1v1 (missing 111 bases from exon 9) that may not generate a functional protein product. Most abundant COX transcript but specific role in skeletal muscle unknown. Acute exercise and chronic exercise training increase transcript levels.
COX-1	Variant b (−1b)	Also known as COX-3. An intron-1-retaining version of COX-1 with 3 splice variants: −1b1, −1b2, −1b3. Apparently sensitive to common COX-inhibiting drugs in other tissues. Nondetectable or very low transcript levels at rest and nonresponsive to acute and chronic exercise. Unlikely involved in exercise adaptations or related COX-inhibitor effects.
COX-2		Very low or nondetectable transcript and enzymatically active protein levels at rest and after acute and chronic exercise. Although low, transcript levels increase with ingestion or infusion of COX-2-specific or nonspecific COX inhibitors after acute exercise, as well as with chronic exercise training.

See text and related studies (18, 69, 105, 116, 125, 138) for further discussion of skeletal muscle COX.

Chronic effects of COX inhibitors on exercise adaptations.

Four recent studies (48, 56, 81, 125) provide the initial evidence as to whether chronic consumption of commonly consumed COX inhibitors negatively impacts chronic exercise adaptations (Table 1).

Krentz et al. (56) showed in young men and women that 400 mg ibuprofen taken after six sets of biceps curls [three sets of eight to 10 concentric repetitions at 70% one repetition maximum (1RM) and three sets of four to six eccentric repetitions at 100% 1RM], 2–3 days/wk for 6 wk, did not influence muscle growth or strength adaptations. The discrepancy between the acute suppression of muscle protein synthesis in younger individuals by ibuprofen (128) and these chronic findings is not completely clear. The lower dose of ibuprofen (1,200 mg/day vs. 800–1,200 mg/wk); the somewhat short duration of training, limiting the time for the COX inhibitor to have an effect; and the possible influence on muscle protein breakdown are plausible explanations.

Petersen et al. (81) recently showed that older patients with osteoarthritis (mean age ∼62 yr), taking ibuprofen (1,200 mg/day) and completing 12 wk of progressive resistance training, 3 days/wk (four to five sets of eight to 15 repetitions at 70–80% of 1RM), had no effect on muscle mass gains, but muscle strength was enhanced in those individuals consuming the COX inhibitor. The authors suggest that this effect on muscle function was related to the pain relief obtained from the drug consumption. These findings are corroborated by this same group’s data showing that ibuprofen (1,200 mg/day) does not influence the muscle protein-synthesis response to exercise in older osteoarthritis patients (mean age ∼62 yr) (82). Interestingly, they also reported that 12 wk of training and taking ibuprofen did inhibit the increase in muscle satellite cell number induced with training in the placebo group (81). These results are in accordance with reports of COX-inhibiting drugs interfering with satellite cell activity after exercise, which is apparently mediated through COX-1 (61, 68, 79). These findings raise the question of whether satellite cells are necessary for muscle hypertrophy, at least the amount that is generally elicited with exercise-training paradigms used for health and wellness of older individuals and the treatment of sarcopenia. This general question has been debated recently (63, 76), and the answer is apparently not yet at hand (46, 60, 83).

With the use of doses of acetaminophen (4 g/day) or ibuprofen (1.2 g/day) in healthy, older men and women (60–78 yr) completing resistance-exercise training 3 days/wk (three sets of 10 repetitions at ∼75% of 1RM/day) for 12 wk, Trappe et al. (125) unexpectedly showed an enhancement of muscle mass and strength gains of 25–50% over a placebo-consuming group. Follow-up studies on muscle biopsies obtained from these individuals (125, 127) and subsequent ex vivo studies (110) provide some mechanistic clarity about these unexpected COX-inhibitor effects (Fig. 2). It appears the COX inhibitors reduced the PGE₂ production (126) and resultant stimulation of intramuscular IL-6 and muscle RING finger protein-1 (MuRF-1) production (57, 88, 111), which increased net protein balance in response to each exercise bout throughout the exercise program. This hypothesis is based on the data that show that low-level increases in IL-6 acutely inhibit muscle protein turnover (131), chronically promote muscle atrophy (12, 37), and are associated with a reduction in muscle mass and functional independence in older individuals (6, 25, 28, 99), as well as the proteolytic nature of the ubiquitin ligase MuRF-1 (20, 98). The COX-inhibitor consumption also promoted an upregulation of the PGF_2α receptor in the muscle of the drug groups (127). This increase coupled with a general training increase in COX-1 and the PGF_2α-producing enzymes (PGF_2α synthase and PGE₂-to-PGF_2α reductase) (125, 127) would make the muscle less susceptible to the same, daily COX-inhibiting drug doses and more sensitive to any PGF_2α that was produced following exercise. Whether these responses and muscle adaptations are specific to older individuals and any potential basal inflammatory state or exaggerated response following exercise (21, 26, 36, 80, 89, 117, 118, 124) is unclear and needs further investigation. Interestingly, COX-inhibitor consumption did not promote muscle growth in the nonexercising hamstring muscles, suggesting an exercise-loading and/or stretch-related mechanism. This nonexercise finding is somewhat in conflict with the data from Rieu et al. (92), showing that chronic consumption of ibuprofen limits sarcopenia through restoration of the muscle protein-synthesis response to feeding in older rats. The discrepancy between these studies is possibly due to the longer-term dosing of the animals (20% vs. 0.4% of the lifespan) and the higher dose of the drug (30 vs. 14 mg·kg body wt⁻¹·day⁻¹). However, these collective findings have implications, not only for use of COX inhibitors during resistance-exercise training for the treatment of sarcopenia but also for the potential chronic use of low-level, long-term, exercise-independent consumption of COX inhibitors for the treatment of sarcopenia, as is promoted or contemplated for several other conditions, such as cardiovascular disease, dementia, and certain types of cancer (91, 95, 105, 130).

Fig. 2.

Schematic of the portion of the PG/COX pathway involved in the regulation of skeletal muscle protein metabolism and adaptation. See text for specific references related to the enzymes, intermediates, and receptors of the COX pathway, as well as the studies that have delineated the factors and cellular processes regulated by the PGE₂ and PGF_2α receptors that influence skeletal muscle mass. See Trappe et al. (127) for nomenclature related to the PG synthases, reductase, and receptors. MuRF-1, muscle RING finger protein-1; PI3K/ERK/mTOR, phosphoinositide 3-kinase/ERK/mammalian target of rapamycin (62).

It is interesting to note that acetaminophen is not commonly considered an nonsteroidal, anti-inflammatory drug because of its relative lack of COX-inhibitory or anti-inflammatory effect in many peripheral tissues, yet it is a potent analgesic, fever reducer, and COX inhibitor within the central nervous system (22, 31). Nonetheless, acetaminophen clearly inhibits PG synthesis in human skeletal muscle (126). Chronic acetaminophen consumption in animals has also been shown to influence skeletal muscle fiber size and glucose metabolism (139, 140). The basis for this tissue specificity is not clear but may be related to the cellular environment (16, 138).

Finally, Jankowski et al. (48) showed recently that 16 wk of progressive resistance-exercise training, 3–5 days/wk (three sets of five to 12 repetitions at 60–80% 1RM coupled with stair-climbing and jumping exercises), combined with the COX-inhibitor acetaminophen (1,000 mg only on days of exercise, average of 3 days/wk) had no effect on fat-free mass or muscle-strength gains in older men (mean age 64 yr). The large difference in acetaminophen dosing (28 g/wk vs. 3 g/wk) likely explains the different responses in this study and those reported by Trappe et al. (125).

Collectively, these chronic studies, albeit of a limited number, highlight three main points: 1) chronic consumption of commonly consumed COX inhibitors at over-the-counter doses during exercise training does not appear to interfere with the muscle mass and strength gains expected from typical resistance-exercise-training regimens; 2) there appears to be a threshold of the amount of drug that is needed to influence skeletal muscle metabolism and adaptation; and 3) there may be differences between the acute and chronic COX-inhibitor effects on muscle metabolism between younger and older individuals.

It should also be noted that the animal studies in this area clearly show an interfering effect of COX inhibitors on muscle growth and adaptation (14, 59, 73, 109). The discrepancy between these studies and the aforementioned human exercise-training and COX-inhibitor studies is most likely due to the animal studies not necessarily reflecting typical human exercise stimuli, as eluded to previously (18). The consideration of the stress placed on the muscle in the different models is also important in understanding the potential role of COX isoforms, as well as PG and inflammatory regulation of muscle adaptation. For example, typical and highly effective resistance-exercise programs in humans only need to load the muscle for a few minutes every 2–3 days (2, 115, 119–122, 125), resulting in <10 min of muscle loading/wk. Whole muscle-growth rates typical for these types of training paradigms are ∼0.5%/wk, and the highest reported muscle-growth rate reported in the literature is ∼1%/wk (114). These rates are in contrast to the animal models of hypertrophy—some of which have almost constant loading and elicit edema—that result in muscle growth of 25–40%/wk. Considering the potential use of the animal studies and particularly, the PG/COX-pathway genetically modified animals, development of appropriate animal exercise models could facilitate significant advancements in this area.

CONCLUDING REMARKS

The research field of PG and COX-inhibitor regulation of health and disease has grown enormously over the last 80 yr of existence. That skeletal muscle responses and adaptations are regulated by PGs synthesized by the muscle that produced them is now firmly established. It is also clear that one of the most commonly consumed classes of drugs in the world—COX inhibitors—can alter several cellular processes that regulate skeletal muscle responses to acute exercise loading and chronic exercise training. There is much research yet to be done in this complex area to better understand the role of PGs and COX inhibitors, specifically as they pertain to older individuals and the condition of sarcopenia, as well as exercise training and performance of individuals of all ages. The PG/COX-pathway research being conducted in other areas of health and disease will no doubt continue to add significantly to our understanding of this research area in skeletal muscle.

GRANTS

Support for this research was provided by the National Institute on Aging (AG-020532 and AG-00831).

DISCLOSURES

The authors declare no conflicts of interest.

AUTHOR CONTRIBUTIONS

Author contributions: T.A.T. conception and design of research; T.A.T. and S.Z.L. analyzed data; T.A.T. and S.Z.L. interpreted results of experiments; T.A.T. and S.Z.L. prepared figures; T.A.T. and S.Z.L. drafted manuscript; T.A.T. and S.Z.L. edited and revised manuscript; T.A.T. and S.Z.L. approved final version of manuscript.