Exercise has been shown to acutely elevate several metabolic processes in tendon tissue, including collagen turnover and blood flow, and chronically induce changes in tendon properties. Many of these acute metabolic responses to exercise are regulated by the cyclooxygenase (COX) enzymes. We measured the expression levels of COX-1 [variants 1 and 2 (COX-1v1 and COX-1v2)], COX-2, and the recently discovered intron 1-retaining COX-1 variants (COX-1b1, COX-1b2, and COX-1b3) at rest and after resistance exercise (RE). Patellar tendon biopsy samples were taken from six individuals (3 men and 3 women) before and 4 h after a bout of RE (3 sets of 10 repetitions at ∼70% of 1 repetition maximum) and from a separate group of six individuals (3 men and 3 women) before and 24 h after RE and analyzed by real-time RT-PCR. The COX-1 variants were the most abundant COX mRNAs before exercise and remained unchanged (P > 0.05) after exercise. COX-2 was also expressed in tendon tissue at rest and was unchanged (P > 0.05) after exercise. The intron 1-retaining COX-1 variants were not detectable in tendon tissue before or after exercise. COX-1 and COX-2 were expressed at much higher levels by the patellar tendon than by quadriceps skeletal muscle, although the overall COX mRNA expression patterns were similar in skeletal muscle and tendon (COX-1v2 > COX-1v1, P < 0.05; ratio of COX-1 to COX-2 ≅ 4:1). These results suggest that COX-1 and COX-2 are constitutively expressed at relatively high levels in human patellar tendon and are likely targets of COX-inhibiting drugs at rest and after physical activity.
- real-time reverse transcription-polymerase chain reaction
human patellar and Achilles tendons have been shown to be highly responsive to resistance and aerobic exercise. Acutely, exercise increases tendon collagen synthesis (29, 30), collagen breakdown (24, 25), blood flow (6, 23), and glucose uptake (2, 16). Chronically, exercise alters the tendon mechanical properties in humans (20–22, 39) and biochemical properties in animals (51). In tendon tissue, as in skeletal muscle (5, 27, 31, 35, 44, 45, 47), prostaglandins (PGs) produced by the cyclooxygenase (COX) enzymes regulate several metabolic responses to exercise (19, 23), loading (49, 54, 55, 58), and injury (37, 40, 48, 50).
There are two commonly known isoforms of COX, COX-1 and COX-2 (11, 53). Additionally, there are two variants of COX-1, COX-1 variant 1 (COX-1v1) and COX-1 variant 2 (COX-1v2) (10, 43), although almost all previous investigations examining COX-1 do not distinguish between COX-1v1 and COX-1v2 (56). Recent evidence suggests that a third isoform exists in human, rat, and mouse tissues (7, 18, 33, 36). This isoform was provisionally named COX-3 (7) and, more recently, referred to as COX-1b (18, 36), because it is derived from the COX-1 gene and is distinctive because of the retention of intron 1 in the mRNA. Furthermore, three splice variants of COX-1b, COX-1b1, COX-1b2, and COX-1b3, have recently been reported (36). The specific role of each of these COX isoforms and variants in the regulation of the acute metabolic responses to exercise, as well as the normal and pathological responses to chronic physical activity, is not clear. It is generally understood that COX-1 is constitutively expressed and helps maintain the normal homeostatic functions in human tissues, whereas COX-2 is the inducible isoform and is more involved with inflammatory responses in tissues (11, 53). In tendon tissue, more recent evidence suggests that, in addition to healing responses to injury, COX-2 may be involved in important nonpathological responses in the tendon, such as blood flow and protein turnover responses to exercise and loading (23, 48, 49, 54, 55, 58), as has been shown in skeletal muscle (3–5, 45, 56).
The basis for the present investigation was severalfold. 1) Chronic exercise has been shown to cause profound adaptations in the patellar tendon and quadriceps muscle (21, 28, 39, 41), and the acute protein metabolism response to exercise in these two tissues appears to be coordinated (30). 2) In the quadriceps muscle, this acute metabolic response appears to be regulated by one or more of the COX isoforms (46, 47), possibly COX-2, inasmuch as this isoform is induced by exercise (56). 3) In vitro studies have shown that human patellar tendon fibroblasts express COX-1 and COX-2 (54) and that uniaxial mechanical stretching of cultured human patellar tendon fibroblasts induces COX-2 in a load-dependent fashion (58). 4) Millions of men and women consume nonspecific and specific COX-inhibiting drugs daily, which may interfere with the normal metabolic processes regulated by the COX enzymes.
Given this information and the relative dearth of knowledge about the factors that regulate tendon adaptations in humans, we sought to examine the expression of the known COX isoforms and variants in the human patellar tendon in response to a bout of resistance exercise. Our primary focus was the two main isoforms of COX, and we hypothesized that COX-1 would be constitutively expressed and COX-2 would be induced within the first 24 h after a bout of resistance exercise. Discrimination of the COX-1 variants (COX-1v1 and COX-1v2) and measurement of the intron 1-retaining COX-1b variants were also completed, inasmuch as these potentially important COX variants have not been described in human tendon tissue.
MATERIALS AND METHODS
Subjects and Experimental Design
Twelve recreationally active individuals (6 men and 6 women) were divided into two groups of six subjects each with equal numbers of men and women in each group, a 4-h group (4h: 27 ± 1 yr old, 173 ± 1 cm height, and 74.8 ± 4.1 kg body wt) and a 24-h group (24h: 23 ± 1 yr old, 167 ± 1 cm height, and 62.6 ± 4.8 kg body wt). All subjects were nonobese (body mass index ≤27 kg/m2), were nonsmokers, and did not consume any prescription or nonprescription analgesic or anti-inflammatory drug(s), chronically or for the duration of the study. After approval by the Institutional Review Board at Ball State University, all the procedures, risks, and benefits associated with the study were explained to the subjects before they signed a consent form adhering to the guidelines of the Institutional Review Board. The experimental design is presented in Fig. 1 and described in detail below.
Dietary and Activity Control
The evening meals were supplied in liquid form (Ensure Plus, Ross, Columbus, OH; 53% carbohydrate, 15% protein, and 32% fat) and provided 50% of the subject's estimated caloric need [1.5 times the subject's predicted resting metabolic rate (12)] before the resting and 24-h postexercise tendon biopsies. This level of dietary control standardized the composition, amount, and timing (i.e., duration of fast) of the final meal consumed before the morning tendon biopsies and has been used in similar studies of COX (56). In addition, subjects were asked to refrain from physical activity or exercise training 3 days before the start of the study and for the duration of the study. Subjects in the 4h group rested quietly in the laboratory during the 4-h postexercise period, whereas those in the 24h group were discharged and returned on the following morning for the postexercise biopsy.
Resistance Exercise Protocol
Each subject completed a bout of isotonic unilateral knee extension exercise with the right leg. The workload was set to ∼70% of the subject's concentric one repetition maximum, which was determined ≥1 wk before the resting tendon biopsy procedure. After 5 min of light (∼50-W) cycling and 2 sets of 5 repetitions with light loads, each subject completed 3 sets of 10 repetitions with 2 min of rest between sets (38).
After 30 min of supine rest, a patellar tendon biopsy was taken before exercise (left leg) and 4 h (4h group) or 24 h (24h group) after the resistance exercise bout (right leg). All biopsies were taken from the central portion of the patellar tendon by the same technician in a consistent fashion for all subjects. Tendon tissue was obtained after administration of local anesthetic (1.0 ml of 1% lidocaine HCl) using a Magnum biopsy instrument and 14-gauge needle (models MG1522 and MN1410, C. R. Bard, Covington, GA) (17, 30, 32). The tissue was immediately inspected and cleaned of any nontendon tissue under ×10 to ×20 magnification and placed in a preweighed RNase- and DNase-free tube containing 400 μl of RNAlater (Ambion, Austin, TX). Each tube was reweighed to obtain the tissue weight (5.04 ± 0.26 mg), incubated at 4°C for 24 h, and then stored at −20°C until mRNA analysis (see below).
Tendon RNA Analysis
The goal of the RNA analysis was to determine the mRNA levels of the known COX isoforms and variants, COX-1 [COX-1v1 and COX-1v2 (10, 43)], COX-2, and the intron-retaining COX-1 (also referred to as COX-3) (7) variants (COX-1b1, COX-1b2, and COX-1b3), as we previously described in detail from measurements in human skeletal muscle (56).
Disruption and homogenization of tissue.
Each tendon sample was removed from RNAlater and powdered/ground into small pieces in a liquid nitrogen-filled mortar that was kept on ice. Before they were used, the mortar and pestle were treated with RNaseZap (Ambion), autoclaved, and stored at −20°C. The liquid nitrogen-tissue suspension was transferred to a liquid nitrogen-cooled six-well culture plate that was kept on ice. The liquid nitrogen was allowed to evaporate from the well, and then, with care taken to prevent thawing of the sample, 600 μl of lysis buffer (RNeasy Mini Kit, Qiagen, Valencia, CA) were added to the well. The solution was pipetted up and down several times to collect the ground tissue into the tip, and the solution was loaded onto a QIAshredder spin column (Qiagen). The solution was homogenized by passage through the QIAshredder spin column homogenizer during a 3-min spin at 15,000 g. The supernatant was carefully transferred to a new tube, leaving a well-formed pellet on the bottom of the spin column collection tube.
Total RNA extraction and RNA quality check.
RNA was extracted from the supernatant using the RNeasy Mini Kit (Qiagen) according to the manufacturer's protocol. Elimination of genomic DNA contamination was done on-column by DNase digestion following the Qiagen RNeasy Mini Handbook protocol.
One microliter of each total RNA extract (1:1 vol/vol with water) was analyzed using the RNA 6000 Pico LabChip kit on a bioanalzyer (model 2100, Agilent Technologies, Palo Alto, CA). This system reported detailed information about quantity and integrity of the RNA samples. Each RNA sample was electrophoretically separated into two peaks of 18S and 28S rRNA. Data were displayed as a gel-like image and an electropherogram (15, 38). Sample analyses were performed as described by the manufacturer.
Oligo(dT)-primed first-strand cDNA was synthesized using SuperScript II RT (Invitrogen, Carlsbad, CA). This system was optimized for sensitive RT-PCR on small amounts of RNA. A first reaction mix of 12 μl for each sample, consisting of 8 μl of RNA extract (2.4 ng), 1 μl of 10 mM dNTP, 1 μl of oligo(dT)12–18 (0.5 μg/μl), and 2.0 μl of DNase- and RNase-free water, was incubated at 65°C for 5 min and then placed on ice for 1 min. A second reaction mix of 7 μl, consisting of 4 μl of 5× first-strand buffer, 2 μl of 0.1 M DTT, and 1 μl of RNaseOUT recombinant RNase inhibitor, was added to the first reaction mix and incubated at 42°C for 2 min. Finally, 1 μl (50 U) of SuperScript II RT was added to each tube (giving a total volume of 20 μl), incubated at 42°C for 50 min and then at 70°C for 15 min to terminate the reaction, and chilled to −4°C. These cDNA samples were diluted with water to a final volume of 40 μl. All thermal incubations and chilling were done in a Peltier thermal cycler with dual-block DNA engine (MJ Research, Waltham, MA) to provide temperature homogeneity and identical temperature ramping for all samples.
mRNA transcription was quantified in duplicate in a 72-well Rotor-Gene 3000 centrifugal real-time cycler (Corbett Research, Mortlake, NSW, Australia). The PCR mix consisted of 12.5 μl of SYBR Green JumpStart Taq Ready Mix (Sigma-Aldrich, St. Louis, MO), 0.5 μl of forward and reverse primers at 10 μM each, 2.5 μl of cDNA, and RNase-free water to a final volume of 25 μl. All primers were mRNA specific and designed for gene expression real-time PCR analysis (Vector NTI Advance 9 software, Invitrogen) using SYBR Green chemistry. Details about primer characteristics and sequences, as well as the real-time PCR parameters and amplicon melting curve analysis, have been reported previously (56).
Three housekeeping genes (HKGs), GAPDH (NM_002046), 18S rRNA (NR_003286), and RPLP0 (ribosomal protein, large, P0; NM_053275; Table 1), were evaluated for expression stability after the exercise bout (2−ΔCT, where ΔCT is difference in threshold cycle). Total RNA (1.2 ng) from before and 4 and 24 h after exercise was amplified with specific primers for each HKG. Also, the RPLP0 and 18S rRNA were normalized to GAPDH and to each other (2−ΔΔCT, i.e., fold change) by comparison of postexercise (4h or 24h) with preexercise levels to observe the relationship between the various HKGs in the same tissue. With use of this approach, if the HKG expression levels are not affected by time or exercise, the mean fold change at each time point should be ∼1, since 2° = 1 (26). The geNorm program (http://medgen.ugent.be/∼jvdesomp/genorm/download.php) was also applied to determine the most stable reference gene for the tested HKGs in a given cDNA sample (52). This program ranked the tested HKGs, indicating the average expression stability value (M) for each gene, from the least stable to the most stable (M < 1.5).
With use of gene-specific primers, a serial dilution (1, 0.5, 0.250, 0.125, 0.062, and 0.031) curve for each gene was included in each real-time PCR assay to evaluate reaction efficiencies. To make the dilution curve, cDNA from 50 ng of human skeletal muscle RNA (Ambion) was used. The real-time PCR amplification calculated by the Rotor-Gene software was specific and highly efficient (average efficiency for all PCRs = 1.08 ± 0.03, R2 = 0.99 ± 0.002, slope = 3.19 ± 0.03) for all groups, with an intra-assay coefficient of variation of 3.66.
The gene of interest (GOI) expression was evaluated by the 2−ΔCT relative quantification method (26, 42, 56, 59). This method is based on the fact that ΔCT between the GOI and the HKG is proportional to the relative expression level of the GOI. The previously described validation of the HKGs was performed to ensure that the chosen HKG expression was unaffected by time or exercise and, therefore, can be used as a normalization gene.
Patellar tendon and quadriceps muscle comparison.
Our initial results from the patellar tendon suggested that the relative amounts of the COX isoforms were substantially different from our previous data from the quadriceps skeletal muscle (56). That is, between tendon and skeletal muscle, similar CT values were obtained for each of the exon-only COX isoforms and variants, with a substantially lower amount of total RNA used for the tendon (2.4 ng) than for the muscle (50 ng) analyses. Therefore, separate analyses were performed on COX-1v1, COX-1v2, COX-2, and an HKG (GAPDH) to allow for direct comparison between the tendon and skeletal muscle tissue. Similar amounts of total RNA (1.2 and 1.6 ng of tendon and muscle, respectively) were amplified with the specific COX primers under the same reaction conditions and in the same RT and PCR total reaction dilution.
The main objective of the statistical analysis was to determine the influence of acute resistance exercise on the mRNA expression of each of the COX isoforms and variants. For each group, a paired t-test was used to compare the pre- and postexercise mRNA levels for each isoform and variant. An additional objective was to determine the appropriateness of each of the HKGs. These comparisons were made with a t-test. A t-test was used for sex-specific comparisons of the average mRNA levels (independent of time) of each COX isoform and variant. Significance was set at an α-level of 0.05 for all comparisons. Values are means ± SE.
RNA Quality and Concentration
The quality of RNA was confirmed by the presence of ribosomal peaks with no additional signals (such as partially sheared genomic DNA contamination or RNA degradation) around the ribosomal bands and no shifts to lower fragments (Fig. 2). On average, the yield of total RNA from the tendon tissue was 8.07 ± 1.16 ng/mg tendon wet wt.
Validation of HKGs
Analyses for the validation of the three HKGs (using 1.2 ng of RNA) revealed average CT values for GAPDH, RPLP0, and 18S rRNA of 23.29 ± 0.17, 23.54 ± 0.11, and 23.20 ± 0.15, respectively. When these HKGs were cross-normalized and calibrated to preexercise levels, all exhibited stability in expression throughout the time points (i.e., fold change ∼1; Fig. 3). In addition, all three HKGs showed a good stability value (M < 1.5), with GAPDH and RPLP0 being the most stable (0.448–0.495) and 18S rRNA the least stable (0.569–1.016). GAPDH was chosen as the housekeeping normalization gene simply for consistency with our previous work on expression of COX genes in skeletal muscle (56).
COX mRNA Levels
The mRNA levels for COX-1v1, COX-1v2, and COX-2 measured before and 4 and 24 h after exercise are presented in Figs. 4 and 5. Every subject expressed each of these isoforms and variants at every time point. COX-1v1 and COX-1v2 were the most abundant COX mRNA before exercise (∼4 times greater than COX-2) and remained unchanged (P > 0.05) after exercise. Independent of time, COX-1v2 levels were significantly higher (P < 0.05) than COX-1v1 levels (4.61 ± 0.44 vs. 4.15 ± 0.40 arbitrary units). Although COX-2 mRNA expression tended to be higher before exercise in the 24h group, it was unchanged (P > 0.05) from rest at 4 and 24 h after exercise. Since there were no differences over time for each of these isoforms and variants, comparisons between men and women (independent of time) showed no differences (P > 0.05) in any of the mRNA levels (Table 2).
As described previously in muscle (56), two primer sets were used to detect the COX-1b variants [COX-1b1 and COX-1b2 (COX-1b1,2) and COX-1b1 and COX-1b3 (COX-1b1,3)]. However, there were no detectable amounts of any of the three intron 1-retaining COX-1b variants in any subject at any time.
Tendon vs. Muscle Comparison
When similar amounts of total RNA were analyzed together from the patellar tendon and quadriceps muscle (56), our initial impression that the tendon contained substantially higher amounts of COX mRNA (see materials and methods) was confirmed. In the tendon, the CT values for COX-1v1, COX-1v2, and COX-2 were 31.06 ± 0.08, 30.55 ± 0.10, and 31.72 ± 0.09, respectively. In the muscle, there was no measurable amplification up to 45 cycles for these same three mRNAs. CT values for GAPDH were 23.29 ± 0.03 in the tendon and 18.73 ± 0.03 in the muscle, supporting the notion that GAPDH is found in much lower amounts in tendon than in skeletal muscle.
The present investigation is the first to examine the in vivo expression of the known COX isoforms and variants in human tendon tissue. The primary findings were that COX-1 (COX-1v1 and COX-1v2) and COX-2 were constitutively expressed in healthy (i.e., noninjured) human patellar tendon tissue subjected to an exercise stimulus that has been shown to induce significant changes in tendon material properties when performed chronically. The patellar tendon expressed much higher levels of COX-1 and COX-2 than quadriceps skeletal muscle (56), although the overall expression patterns were similar in skeletal muscle and tendon.
COX-1 and COX-1b
Investigations of the COX enzymes in tendon tissue have focused primarily on chronic-overuse injuries and recovery from surgical repair and usually center around the positive and negative influence of COX-inhibiting drugs (37, 40, 48, 54, 58). However, more recently, it has become evident that the COX enzymes likely play an important role in many of the normal (i.e., nonpathological) metabolic processes in tendon tissue (19, 23, 49, 54, 55, 58). Because COX-1 is constitutively expressed under most conditions in many different tissues (11, 53) and, as reported here, in human patellar tendon, it could be involved in the regulation of these normal metabolic processes in tendon tissue. Too few studies have been completed in this area to allow a full understanding of the role of each of the COX enzymes in producing the various PGs, even though a large amount of information is available about the function of specific PGs. Of interest is the relative expression of the two variants of COX-1, COX-1v1 and COX-1v2. The fact that there are two variants of COX-1 (10, 43) and that the distinction between these two variants continues to be relatively ignored is surprising. The alternative splicing of the COX-1 mRNA, which eliminates a portion of exon 9 to produce COX-1v2, has been shown in human lung fibroblasts to be regulated by several cytokines (10), including IL-1β, which is a potent regulator of COX expression and PG production in human tendon fibroblasts (50). The results of the present study in human patellar tendon show that COX-1v2 is the most abundant COX mRNA, although the abundance of both COX-1 mRNAs was relatively high compared with that of the other COX isoforms and variants. Whether the truncated variant 2 produces a functioning COX protein in human tendon or simply serves to regulate the COX-1 enzyme activity (10) is not clear. Of course, constitutive expression of these mRNAs does not necessarily mean that the COX-1 isoform is not producing PGs at rest or altering PG production in response to resistance exercise loading. Both variants of COX-1 likely play a role in PG-mediated metabolism in the tendon, and more investigation into the specific roles of the two variants of COX-1 is needed.
The intron 1-retaining COX-1b variants (COX-1b1, COX-1b2, and COX-1b3) were not detectable in the patellar tendon tissue of men or women at rest or after resistance exercise in the present study. The recent discovery of COX-3 (7) and subsequent identification of three variants (36) suggested that functioning COX proteins are produced by one or more of these messages, which can produce PGs and can be inhibited by various nonsteroidal anti-inflammatory drugs and acetaminophen. From these previous studies, which did not report any measurements in tendon tissue, and the results from the present investigation, it appears that the intron 1-retaining COX variants likely do not play a role in the regulation of tendon metabolism in humans.
Tendon Loading and Sex-Specific Responses
Several in vitro models have been developed to examine the influence of loading on mechanotransduction mechanisms that regulate gene expression in human tendon fibroblasts (49, 54, 58). Fibroblasts are the cells in connective tissues, such as tendon, that sense mechanical loading and ultimately regulate the metabolism and structure of connective tissue by expressing the genes responsible for the synthesis and degradation of the extracellular matrix (e.g., collagens, proteoglycans, and matrix metalloproteinases) (19, 55). As such, fibroblasts are at the center of normal and pathological connective tissue physiology (1, 8, 9, 14). Several studies using the in vitro systems that cyclically load (i.e., stretch) fibroblasts show a large induction of the genes that are involved in the maintenance of the extracellular matrix (49, 54, 58). Specifically, COX-2 is consistently shown to be induced, along with a concomitant increase of PGs, after cyclic stretching of human tendon fibroblasts (49, 54, 58). Yang et al. (58) reported a load-dependent increase in COX-2 expression in human patellar tendon fibroblasts. In fact, at higher levels of stretching of human patellar tendon fibroblasts, COX-1 was also induced (54). In these studies, stretching was maintained for 2–24 h and the amount of stretching was varied relative to the size of the plates on which the fibroblasts were cultured. Thus it is difficult to directly compare or specifically contrast these data with those obtained in the present study. One could speculate that the in vivo levels of the COX-1 and COX-2 enzymes in the patellar tendon before exercise are sufficient to produce the PGs that are necessary to stimulate the molecular and metabolic responses to the relatively high-load and short-duration (<10 min) resistance exercise stimulus used in the present study. This speculation is generally supported by the notion of Yang et al. that small-magnitude stretching is anti-inflammatory and large-magnitude stretching is proinflammatory. The comparisons with the in vitro stretching data should be kept in the context that loading of the patellar tendon with resistance exercise similar to the bout used in the present study a few days per week for several weeks induces significant changes in the mechanical properties of the patellar tendon (21, 39). In addition, the regulation of tendon metabolism and adaptations to loading are complex (19, 55) and likely regulated by numerous factors independent of COX.
In vivo and in vitro studies suggest that some, but not all, patellar tendon mechanical properties, as well as the ability of the tendon to adapt to aerobic exercise training, may be different between men and women (13, 34, 57). In addition, patellar tendon collagen synthesis at rest and 72 h after an aerobic exercise bout is sex specific but is not influenced by the phase of the menstrual cycle (29, 30). The present results show no difference in the expression levels of COX-1 or COX-2 at rest or after resistance exercise (Table 2) in the patellar tendon of men and women. Thus any sex-specific difference in tendon metabolism and adaptations may not be related to the mRNA expression levels of the various COX isoforms and variants found in the human patellar tendon.
Tendon vs. Skeletal Muscle
The mRNA levels of the same COX isoforms and variants have recently been examined in the quadriceps femoris (vastus lateralis) muscle with use of a similar experimental design (i.e., before exercise and 4 and 24 h after knee extension resistance exercise) (56). We compared the present results from the patellar tendon with those from the quadriceps muscle, inasmuch as the metabolic response to exercise in these two tissues appears to be coordinated (30). The substantially higher expression levels of COX-1v1, COX-1v2, and COX-2 in the tendon than in the skeletal muscle suggest that the tendon tissue may be a more likely target of COX-inhibiting drugs. In addition, the general expression patterns of the COX isoforms and variants were very similar between the tendon and muscle (Fig. 6). The COX-1 variants were the most abundant, with COX-1v2 being higher than COX-1v1, and the ratio of the COX-1 variants to COX-2 was ∼4:1. The significance of the similar relative abundances of the two main COX isoforms is not clear but suggests that the COX enzymes regulate metabolism in tendon and muscle in a similar fashion. In addition, the intron 1-retaining COX variants were not expressed (tendon), or very few individuals expressed any of these variants and at very low and sporadic levels (muscle). The main difference between these two tissues was that skeletal muscle did not express COX-2 mRNA, although COX-2 protein was present, at rest (56). These general comparisons have implications for tissue-specific targets of nonspecific and specific COX-inhibiting drugs, as it relates to normal, pathological, and repair responses in tendon and skeletal muscle.
Although simultaneous measurements of COX protein levels would have added to the interpretation of the study findings, the tissue yield from the tendon biopsy technique precluded these measurements. Future studies examining the protein levels of the COX-1 variants and COX-2 are clearly warranted. Also, although COX-2 was constitutively expressed in both groups in response to the exercise bout, the levels of the COX-2 mRNA between the groups (Fig. 5) suggest a relatively high level of variability in this message in human tendon. In fact, when the data from both groups are combined, the variability of the COX-2 mRNA expression levels is less in tendon than in skeletal muscle (56). Thus this impression is simply a function of the two-group study design used to ensure that no COX measurements in the tendon were influenced by a previous biopsy.
Perspectives and Significance
The constitutive expression of COX-1 (COX-1v1 and COX-1v2) and COX-2 in healthy (i.e., noninjured) patellar tendon at rest and in response to quadriceps muscle resistance exercise in men and women suggests that the COX enzymes are likely involved in normal tendon metabolism and adaptations to loading. The high COX expression levels in human tendon compared with other tissues such as skeletal muscle also suggest that tendon tissue may be a significant target of nonspecific and isoform-specific COX-inhibiting drugs. Further examination of the PGs produced from the specific COX enzymes and the role of these PGs in the normal metabolic responses to physical activity to induce changes in the material properties of the tendon are needed. Larger-scale human clinical studies of the role of chronically consumed COX-inhibiting drugs on tendon metabolism and adaptations are needed for a better understanding of the findings of the present investigation and other acute in vivo human studies, as well as studies of isolated tendons and tendon cells. The present results and future studies will also provide insight into tendon pathology and treatment, including exercise and pharmacological approaches.
This work was supported by National Institute on Aging Grant R01 AG-020532 (T. A. Trappe).
The authors thank all the subjects for their participation.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society