Prostaglandin E₂/cyclooxygenase pathway in human skeletal muscle: influence of muscle fiber type and age

Abstract

Prostaglandin E₂ (PGE₂) produced by the cyclooxygenase (COX) pathway regulates skeletal muscle protein turnover and exercise training adaptations. The purpose of this study was twofold: 1) define the PGE₂/COX pathway enzymes and receptors in human skeletal muscle, with a focus on type I and II muscle fibers; and 2) examine the influence of aging on this pathway. Muscle biopsies were obtained from the soleus (primarily type I fibers) and vastus lateralis (proportionally more type II fibers than soleus) of young men and women (n = 8; 26 ± 2 yr), and from the vastus lateralis of young (n = 8; 25 ± 1 yr) and old (n = 12; 79 ± 2 yr) men and women. PGE₂/COX pathway proteins [COX enzymes (COX-1 and COX-2), PGE₂ synthases (cPGES, mPGES-1, and mPGES-2), and PGE₂ receptors (EP1, EP2, EP3, and EP4)] were quantified via Western blot. COX-1, cPGES, mPGES-2, and all four PGE₂ receptors were detected in all skeletal muscle samples examined. COX-1 (P < 0.1) and mPGES-2 were ∼20% higher, while EP3 was 99% higher and EP4 57% lower in soleus compared with vastus lateralis (P < 0.05). Aging did not change the level of skeletal muscle COX-1, while cPGES increased 45% and EP1 (P < 0.1), EP3, and EP4 decreased ∼33% (P < 0.05). In summary, PGE₂ production capacity and receptor levels are different in human skeletal muscles with markedly different type I and II muscle fiber composition. In aging skeletal muscle, PGE₂ production capacity is elevated and receptor levels are downregulated. These findings have implications for understanding the regulation of skeletal muscle adaptations to exercise and aging by the PGE₂/COX pathway and related inhibitors.

over the last decade it has become clear that the cyclooxygenase (COX) pathway, which produces prostaglandins (PGs) and is regulated by common COX inhibiting drugs (38), is a major regulator of skeletal muscle adaptations to aging and exercise (49). PGE₂ is one of the key PGs produced by this pathway and is intricately involved in regulating these adaptations (40, 49). As a result, this portion of the pathway has been referred to as the PGE₂/COX pathway (49).

While some of the downstream molecular and metabolic signals of this pathway have been clarified (20, 40, 49, 51), the actual enzymes that direct the synthesis of PGE₂, as well as the receptors for PGE₂, have not been elucidated in human skeletal muscle. In addition, evidence from animal muscle studies (34, 43) and recent findings in older exercising individuals suggest that this pathway may be differentially expressed or regulated between the two major fiber types found in human skeletal muscle (i.e., type I and II) (47, 50). This potential fiber type specificity is important because of the large difference in functional properties and adaptability of type I and II muscle fibers (7, 11, 32, 35, 44, 45).

The purpose of this investigation was twofold: 1) define which PGE₂/COX pathway enzymes and receptors are present in human skeletal muscle, with an additional focus on type I and II muscle fibers; and 2) examine the influence of aging on this pathway in a cohort with well-documented sarcopenia. The overarching hypothesis of the investigation was that the PGE₂/COX pathway enzymes and receptors would generally be higher in type I muscle fibers and aging skeletal muscle.

MATERIALS AND METHODS

Subjects.

A total of 28 subjects were studied for this investigation. For the first part of the investigation that examined the PGE₂/COX pathway and muscle fiber type, muscle biopsies (3) were obtained from the soleus (primarily type I fibers) and vastus lateralis (proportionally more type II fibers than soleus) of recreationally active young men and women (n = 8: 6 male, 2 female; 26 ± 2 yr; 175 ± 4 cm; and 75.2 ± 5.7 kg). For the second part of the investigation that examined the PGE₂/COX pathway and aging, muscle biopsies were obtained from the vastus lateralis of a separate group of sedentary young (n = 8: 4 male, 4 female; 25 ± 1 yr; 172 ± 3 cm; and 71.5 ± 3.4 kg) and old (n = 12: 6 male, 6 female; 79 ± 2 yr, range: 70–93 yr; 166 ± 2 cm; and 69.8 ± 4.0 kg) men and women. The older group had well-documented sarcopenia, including an impaired ability (increased time) to climb stairs (80%), rise from a chair (56%), and walk (44%), as well as lower quadriceps muscle volume (−29%), muscle strength (−35%), muscle power (−48%), and strength (−17%) and power (−33%) normalized to muscle size (16). Data on both of these cohorts have been reported previously (16, 22, 33, 44, 48). The muscle biopsies were immediately frozen and stored in liquid nitrogen until analyzed as described below. All procedures, risks, and benefits associated with the experimental testing were explained to the subjects before they signed a consent form adhering to the Institutional Review Board guidelines.

PGE₂/COX pathway protein quantification.

The following PGE₂/COX pathway proteins, based on identification primarily from other tissues (4, 15, 18, 25, 27–29, 39, 41, 49), were quantified in each muscle sample using Western blot analysis: two COX enzymes (COX-1 and COX-2), three PGE₂ synthases (cPGES, mPGES-1, and mPGES-2), and four PGE₂ receptors (EP1, EP2, EP3, and EP4) (Fig. 1).

Fig. 1.

The general prostaglandin E₂/cyclooxygenase (PGE₂/COX) pathway. The specific enzyme isoforms and receptor subtypes listed in the blue offset boxes have been identified primarily in other tissues and were the focus of the current investigation. The yellow components represent lipid components of the pathway. Downstream effects of this pathway in human skeletal muscle are also shown (49).

Muscle samples (∼10 mg) were weighed at −25°C (Cahn C-35; Orion Research, Beverly, MA) and homogenized using a motorized ground-glass homogenizer (Duall; Kimble Chase, Vineland, NJ) in 40 vol of cold RIPA buffer (25 mM Tris·HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, and 0.1% SDS; Pierce, Rockford, IL) with Halt protease and phosphatase inhibitors (Pierce). Homogenized samples were spun for 5 min at 14,000 g and 4°C. Protein concentration was determined on an aliquot of each supernatant with a BCA assay (Pierce) using bovine serum albumin (BSA) as the standard. Samples were diluted in SDS sample buffer, heated one time for 5 min at 95°C, and further diluted to a common concentration (1 μg/μl) before gel loading.

Proteins were resolved on a 4–20% gradient gel (Pierce) via SDS-PAGE for ∼90 min at 100 V (Mini Protean 3 system; Bio-Rad Laboratories, Hercules, CA) and then transferred to a polyvinylidene fluoride membrane (Immobilon-P; Millipore, Bedford, MA) for 90 min at 40 V and ∼4°C (Table 1). Membranes were blocked with either 5% milk or BSA (EP1 only) diluted in Tris-buffered saline with 1% Tween-20 (TBS-T) for 60 min at room temperature. Incubation with human primary antibodies diluted in 1× TBS-T was completed overnight at 4°C (Table 1). Membranes were incubated with secondary antibody (Table 1) in 5% milk or BSA (EP1 only) for 60 min at room temperature and exposed to an enhanced chemiluminescent substrate [Amersham ECL Prime, GE Healthcare, Piscataway, NJ; or ECL 2 Western blotting Substrate, Pierce (EP1 only)] for protein identification and quantification with a chemiluminescent imaging system (FluorChem SP; Alpha Innotech, San Leandro, CA).

Table 1.

PGE₂/COX Pathway Western blot parameters

	Protein Loaded, μg	Primary Antibody Cat. No.	Primary Antibody Dilution	Secondary Antibody Cat. No.	Secondary Antibody Dilution	Positive Control Cat. No.	Negative Control Cat. No.
COX-1	20	160110	1:200	10004302	1:10,000	60100	—
COX-2	20	160112	1:500	10004302	1:10,000	10009624	360107
cPGES	20	160150	1:200	10004301	1:10,000	10010498	—
mPGES-1	20	160140	1:200	10004301	1:10,000	10009734	—
mPGES-2	8	160145	1:200	10004301	1:10,000	SC-2254	—
EP1	20	101740	1:500	10004301	1:10,000	HSMSTD	301740
EP2	5	101750	1:500	10004301	1:10,000	10004722	301750
EP3	10	101760	1:500	10004301	1:10,000	HSMSTD	301760
EP4	10	101775	1:500	10004301	1:10,000	HSMSTD	301775

All primary antibodies were purchased from Cayman Chemical (Ann Arbor, MI). All commercially purchased positive and negative controls were obtained from Cayman, except mPGES-2, which was obtained from Santa Cruz Biotechnology (Santa Cruz, CA). PGE₂, prostaglandin E₂; COX, cyclooxygenase; HSMSTD, human skeletal muscle standard (vastus lateralis, obtained from our laboratory), which was used for quantification normalization when a positive control was not available. All positive controls are human except for mPGES-2 (mouse) and EP2 (ram).

Each gel was loaded with soleus and vastus lateralis samples from the same subject or young and old samples in alternating fashion, a molecular mass ladder (MagicMark XP Western Protein Standard; Invitrogen/Life Technologies, Grand Island, NY), and a positive control protein or human skeletal muscle standard (Table 1). Intensities of the protein bands were quantified and normalized to the positive control or standard and expressed as arbitrary units (AU). Verification of equal protein loading was accomplished by Ponceau S staining.

Western blot identification considerations.

Identification of immunodetected proteins was confirmed by considering the reported molecular mass of the protein and any known posttranslational modifications, along with migration relative to the molecular mass ladder and positive control (Fig. 2). Negative control (blocking peptide) experiments were also completed for COX-2, EP1, EP2, EP3, and EP4. For COX-1, the identified band migrated equivalent to the positive control protein and appropriately relative to the molecular mass ladder as we have previously reported (5, 47). COX-2 was not detected in any of the muscle samples, with a banding pattern similar to the blots we have previously reported and discussed in detail (5, 47). The cPGES-identified band migrated slightly beyond the positive control protein and just above the 20-kDa molecular mass band, which is consistent with previous reports in other human tissues (15, 27, 29). Identification of a mPGES-1 band was not completely conclusive, although a band migrating somewhat beyond the 20-kDa molecular mass band was identified. However, this band did not migrate with or beyond the hexahistidine-tagged positive control (reported as 19 kDa), as might be expected considering the molecular mass of mPGES-1 is ∼16 kDa (13, 15, 27, 29). No other bands at or below 20 kDa were detectable. However, the identified band was quantified and was not different between soleus and vastus lateralis or young and old muscle (data not shown). The mPGES-2 identified band migrated slightly beyond the positive control and just above the 30-kDa molecular mass band, which is consistent with the commonly reported 33-kDa version of the protein (15, 27, 29). No data are presented for the ∼42-kDa version, which did appear as a very faint band near the expected migration distance on the blots.

Fig. 2.

Representative blot images of the PGE₂ synthases and receptors. Representative images of the COX enzymes have been presented and described in detail previously (5, 47). Images for a given protein include molecular mass band(s) (MW), a positive control or human skeletal muscle standard (Table 1), and representative samples from the muscle/fiber type specific or aging comparison. See Western blot identification considerations for a complete description of protein identification. Images presented are from noncontiguous lanes of the same blot (as indicated by vertical lines) at the same exposure and reassembled to reproduce exactly the relative migration patterns.

For EP1, two bands migrating near the 40-kDa molecular mass band were identified. One band migrated near (likely representing the expected molecular mass protein, ∼42 kDa) and one band migrated beyond (∼35 kDa), as has been reported previously in myocytes and other tissues (31, 37). These two bands were reduced in intensity with the blocking peptide. The EP2 identified band migrated just above the 50-kDa molecular mass band but beyond the 60-kDa band and (nonhuman) positive control protein. The identified band and positive control band were eliminated with the blocking peptide. This migration is consistent with previous reports in other human tissues (23). EP3 was identified as two distinct bands migrating near the 50 and 60-kDa molecular mass bands, and these two bands were eliminated with the blocking peptide. This banding pattern is consistent with previously reported glycosylations of the EP3 receptor in nonhuman tissues and bands reported in other human tissues (17, 26). The EP4-identified band migrated near the 50-kDa molecular mass band and was eliminated with the blocking peptide. This migration is consistent with what would be expected from human and other tissues (4, 26).

Statistical analysis.

Data were analyzed with paired (soleus vs. vastus lateralis) or unpaired (young vs. old) t-tests. Data are presented as means ± SE.

RESULTS

PGE₂/COX pathway components in human skeletal muscle.

COX-1, cPGES, mPGES-2, and all four PGE₂ receptors (EP1, EP2, EP3, and EP4) were detected in all young and old muscle samples examined. mPGES-1 was not reliably identified and COX-2 was not detected in any of the muscle samples (see Western blot identification considerations).

Soleus vs. vastus lateralis skeletal muscle.

On average, COX-1, cPGES, and mPGES-2 were ∼20% higher in soleus compared with vastus lateralis (Fig. 3). EP1 (35 kDa: 1.34 ± 0.36 vs. 1.51 ± 0.46 AU; 42 kDa: 0.55 ± 0.11 vs. 0.64 ± 0.13 AU) and EP2 (0.56 ± 0.04 vs. 0.50 ± 0.05 AU) were not different between soleus and vastus lateralis, while EP3 was 99% higher and EP4 was 57% lower in soleus (P < 0.05) (Fig. 3).

Fig. 3.

Protein levels for COX-1, cPGES, mPGES-2, EP3, and EP4 in soleus (a surrogate for type I muscle fibers) and vastus lateralis (a surrogate for type II muscle fibers, relative to soleus). *P < 0.05 vs. vastus lateralis. **P < 0.1 vs. vastus lateralis. The y-axis is protein abundance normalized to the positive control or human skeletal muscle standard (Table 1) in arbitrary units. The EP3 data represent the 50-kDa band (no difference was found between muscles for the 60-kDa band, data not shown).

Aging skeletal muscle.

COX-1 (1.26 ± 0.29 vs. 1.00 ± 0.12 AU) and mPGES-2 (0.45 ± 0.03 vs. 0.42 ± 0.03 AU) were not different between young and old, while cPGES was 45% higher in old (P < 0.05) (Fig. 4). On average, the PGE₂ receptors were ∼33% lower in old compared with young skeletal muscle (Fig. 4).

Fig. 4.

Protein levels for cPGES and all four PGE₂ receptors in young and old skeletal muscle. *P < 0.05 vs. young. **P < 0.1 vs. young. The y-axis is protein abundance normalized to the positive control or human skeletal muscle standard (Table 1) in arbitrary units. The EP1 and EP3 data represent the 35 and 60-kDa bands, respectively (no difference was found between young and old for the other quantified bands for those receptors, data not shown).

DISCUSSION

The primary aim of this investigation was to define the PGE₂/COX pathway enzymes and receptors in human skeletal muscle, as well as to understand the potential influence of muscle fiber type and aging on the pathway. Understanding this pathway is important because it is involved in skeletal muscle adaptations to aging and exercise training, and it is regulated by commonly consumed COX inhibiting drugs (49). The findings show that human skeletal muscles with markedly different type I and II muscle fiber composition, as well as skeletal muscle from aging individuals, have altered PGE₂ production capacity and PGE₂ receptor abundance.

The vastus lateralis and soleus muscle samples studied in this investigation were from 28 healthy men and women ranging from 20 to 93 yr old. Thus the findings provide a broad representation of muscle fiber type, age, and gender. The results define for the first time that COX-1, the PGE₂ synthases cPGES and mPGES-2, and the PGE₂ receptors EP1, EP2, EP3, and EP4 are constitutively expressed at the protein level in young and old human skeletal muscle and in skeletal muscles with vastly different fiber types. This latter finding provides strong support that all of these proteins are expressed in both type I and type II muscle fibers of men and women. These protein data are supported by the findings that COX-1, cPGES, mPGES-2, and EP4 are constitutively expressed at the mRNA level at relatively high abundance in skeletal muscle (5, 33, 47, 51, 53). However, EP1, EP2, and EP3 are inconsistently expressed at low transcript levels (33). It should be noted that our homogenate measurements could contain some nonmuscle fiber contributions (e.g., intramuscular vasculature, inflammatory cells). This contribution is likely very small in our subjects, since they did not have overt muscle or other inflammatory disease (18, 25, 30). In addition, some of these proteins (and related mRNA) (e.g., COX-1, cPGES, mPGES-2, EP4) have been detected inside muscle fibers using immunolocalization and isolated muscle fiber measurements (18, 25, 30, 33).

Whether or not mPGES-1 was present in the muscle samples studied in this investigation is not completely clear (see Western blot identification considerations). However, it does appear that if present in muscle homogenates, mPGES-1 is expressed at very low levels at both the protein and mRNA levels (18, 25, 33, 51). Furthermore, any contributions to the homogenate measurements likely come from nonmuscle fiber components, even in muscle inflammatory disease states (18, 25). These findings are consistent with the current and previously reported findings of no or very low level expression of COX-2 protein and mRNA in human skeletal muscle (5, 18, 25, 30, 33, 47). This is also consistent with the idea that COX-2 and mPGES-1 appear to be linked in the production of PGE₂ in inflammatory diseases of skeletal muscle and pathophysiological conditions in other tissues (1, 18, 25, 27, 29). Collectively, these findings provide further evidence that COX-1 is likely the major source of general PG production in young and old human skeletal muscle (5, 18, 25, 30, 47, 49) and suggest PGE₂ specific synthesis is directed by cPGES and mPGES-2. Thus these three constitutively expressed proteins likely regulate the homeostatic processes controlled, at least in part, by PGE₂. In skeletal muscle, these processes would include the regulation of hypertrophy and atrophy in young and old individuals.

The influence of muscle fiber type on the PGE₂/COX pathway profile was one focus of this investigation because the contractile and metabolic characteristics between fiber types are fundamentally different (7, 35, 44) and respond differently to aging and exercise training (11, 32, 35, 45). The soleus and vastus lateralis were chosen as surrogates for type I and type II fibers, respectively, because numerous studies have shown that homogenate measurements from these two muscles compared with isolated fiber type specific measurements consistently yield similar results (6–8, 10, 14, 21, 22). These comparisons include measurements at the protein, enzymatic activity, protein metabolism, and energy substrate levels. Thus the current results likely reflect the PGE₂/COX pathway in type I and II muscle fibers, although the differences found in the current study may be even greater in a “pure” comparison. The higher COX-1 and mPGES-2 levels in the soleus compared with the vastus lateralis suggest an increased capacity to generate PGE₂ via enhanced PGH₂ generation from arachidonic acid and PGE₂ synthesis from PGH₂. As PGE₂ is an important mediator of protein turnover (49), this increased PGE₂ generation capacity in type I compared with type II fibers may be involved in the regulation of fiber type-specific adaptations to exercise training. Additionally, these data further suggest the possibility of a fiber type-specific impact of COX inhibiting drugs on protein metabolism-related adaptations to exercise training. Indeed, it has recently been shown that chronic COX inhibition during resistance exercise training in older adults, likely working through a PGE₂-related mechanism, enhances the typical muscle mass gains in a fiber type-specific fashion (49–51). That is, the enhanced myocellular hypertrophy induced by COX inhibition was more pronounced in type I muscle fibers. These findings support further investigation into the effects of common COX inhibitors and future PGES inhibitors (19, 36) on skeletal muscle.

The fiber type-specific differences in PGE₂ receptor levels (higher EP3 and lower EP4 in the soleus compared with vastus lateralis) are somewhat more difficult to interpret, primarily due to the limited number of EP receptor-related studies in skeletal muscle and the complexity of EP receptor biology (4, 28, 41). From other tissue types it is clear that different EP receptor subtypes can have similar and opposing intracellular effects, which can be modified by changing cofactors such as the cytokine environment (4, 23, 28, 41, 42, 52). The limited available skeletal muscle information shows that homogenate vastus lateralis mRNA levels of EP4 increase with resistance exercise training (51), suggesting this subtype and its resultant downstream signaling effects are involved in the normal adaptive response to resistance exercise. In addition, EP4 mRNA is higher in isolated type IIa compared with type I human muscle fibers (33), which corresponds with the notion that type II fibers are typically more responsive to resistance exercise. These isolated muscle fiber findings also correspond to the current findings of a higher EP4 protein level in vastus lateralis compared with the primarily type I soleus muscle.

The initial PGE₂ receptor data presented in the current investigation provide a basis to develop ex vivo studies of human skeletal muscle EP signaling and extend those studies to in vivo investigations in humans. These studies are possible because there is a well-developed line of EP1-4 receptor agonists and antagonists and specific downstream inhibitors of the known EP-activated pathways (4, 28, 41, 52, 55). Some of these same pathways have been shown to regulate skeletal muscle signaling and adaptations (2, 24, 40). These tools can also be incorporated with the commonly available line of COX-inhibiting drugs for use in humans.

The influence of aging on the PGE₂/COX pathway profile in skeletal muscle was another focus of this investigation because of the significance of sarcopenia in the aging population (9, 12) and the apparent role of this pathway in the etiology and exercise-related treatment of sarcopenia (49). The higher level of cPGES in old muscle suggests an increased capacity for PGE₂ production. Although there is no information regarding skeletal muscle PGE₂ levels with aging in humans, data from aging monkeys show higher PG levels in skeletal muscle (54). Furthermore, data from aging humans show higher rates of skeletal muscle proteolysis (46), to which PGE₂ may be linked. Although speculative, the age-related PGE₂ receptor downregulation in the current study (Fig. 4) may be the result of chronic elevations in PGE₂. This potential adaptation is important considering the likelihood of elevated skeletal muscle PGE₂ levels creating an environment that is not ideal for growth or may even contribute to atrophy (40, 47, 51). Interestingly, cPGES mRNA levels in skeletal muscle are reduced ∼20% with resistance exercise training that induces hypertrophy in older individuals (51). Finally, the similar levels of the COX enzymes in young and old skeletal muscle suggest that a given COX inhibitor dose, notwithstanding gastrointestinal absorption or metabolic clearance differences, should be equally efficacious at blocking PGH₂ production (see Fig. 1) in young and old individuals. However, the higher cPGES level in old muscle suggests more PGE₂ production for a given amount of PGH₂; thus a higher dose of COX inhibitor may be needed in an older individual to control intramuscular PGE₂ levels.

In summary, the current investigation provides the first comprehensive profile of the PGE₂/COX pathway in human skeletal muscle. PGE₂ production capacity and receptor levels are different in skeletal muscles with markedly different type I and II muscle fiber composition and are altered in aging skeletal muscle. The current findings coupled with the existing literature provide a platform for investigations into the PGE₂/COX pathway as it relates to 1) the understanding and treatment of the age-related loss of skeletal muscle mass and function, and 2) the regulation of fiber type specific adaptations to exercise in young and old individuals. This information will also add to our understanding of the role of COX inhibitors, as well as potential PGE₂ synthase and EP receptor modulators in the adaptations of skeletal muscle to exercise training and aging.

GRANTS

This study was funded by National Institute of Aging Grants AG-020532 and AG-038576.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: S.Z.L., B.J., and T.A.T. conception and design of research; S.Z.L., B.J., K.M.L., B.E.L., and T.A.T. performed experiments; S.Z.L., B.J., K.M.L., and T.A.T. analyzed data; S.Z.L., B.J., K.M.L., B.E.L., S.W.T., and T.A.T. interpreted results of experiments; S.Z.L., K.M.L., and T.A.T. prepared figures; S.Z.L. and T.A.T. drafted manuscript; S.Z.L., B.J., K.M.L., B.E.L., S.W.T., and T.A.T. edited and revised manuscript; S.Z.L., B.J., K.M.L., B.E.L., S.W.T., and T.A.T. approved final version of manuscript.