Novel eccentric (lengthening contraction) exercise typically results in muscle damage, which manifests as prolonged muscle dysfunction, delayed onset muscle soreness, and leakage of muscle proteins into circulation. There is a large degree of variability in the damage response of individuals to eccentric exercise, with higher responders at risk for potentially fatal rhabdomyolysis. We hypothesized that single nucleotide polymorphisms (SNPs) in chemokine ligand 2 (CCL2) and its receptor chemokine receptor 2 (CCR2) associate with the high degrees of variability in the muscle damage response. We based this hypothesis on CCL2's roles in macrophage and satellite cell signaling in injured muscle. DNA was obtained from 157 untrained men and women following maximal eccentric exercise. Strength loss, soreness, serum creatine kinase (CK), and myoglobin levels before and during recovery from a single exercise bout were tested for association with 16 SNPs in CCL2 and CCR2. The rare alleles for rs768539 and rs3918358 (CCR2) were significantly (P < 0.05) associated with lower preexercise strength in men, whereas CCL2 SNPs (rs13900, rs1024611, and rs1860189) and CCR2 (rs1799865) were associated with altered preexercise CK levels in women. During recovery, the rs3917878 genotype (CCL2) was associated with attenuated strength recovery in men and an elevated CK response in women. CCR2 variants were associated with slower strength recovery in women (rs3918358) and elevated soreness (rs1799865) across all subjects. In summary, we found that SNPs in CCL2 and CCR2 are associated with exercise-induced muscle damage and that the presence of certain variants may result in an exaggerated damage response to strenuous exercise.
- monocyte chemoattractant protein-1
- single-nucleotide polymorphisms
exertional rhabdomyolysis is a condition in which unaccustomed, intense exercise causes a breakdown of muscle tissue, resulting in myofibrillar content leakage [e.g., creatine kinase (CK) and myoglobin (Mb)] into circulation. This condition can become clinically relevant when Mb is increased to such high levels in the blood as to be toxic to the kidneys, potentially leading to kidney failure and possible death (6, 7, 8, 10, 19). CK is commonly used as a surrogate for Mb, since blood CK activity is easier and less expensive to measure than Mb. Other typical symptoms of clinically relevant rhabdomyolysis include dark urine (myoglobinuria), muscle tenderness, stiffness, and weakness, and possible complications include compartment syndrome, hyperkalemia, and lactic acidosis (2).
Clinically relevant rhabdomyolysis incidence rates are low for nondiseased populations (i.e., no existing kidney disease and no use of drugs such as statins) (29) but can be elicited in certain individuals under conditions such as dehydration and heat stress (6, 11, 17, 30). Why some individuals in these conditions suffer significant complications of exertional rhabdomyolysis and others do not is currently unknown.
We have reported that in response to controlled laboratory exercise (designed to induce moderate muscle damage that does not result in significant complications) (8), the extent of the damage response across individuals is notoriously variable (10, 22, 24). In fact, there seems to be a population of “high responders” to lengthening exercise, which describes subjects that experience a large degree of muscle dysfunction (>70% strength loss from baseline) after exercise with prolonged strength recovery over time (22) and highly elevated CK and Mb values in the blood (8). These high responders would likely suffer complications of rhabdomyolysis in situations such as those where there is heat stress and/or dehydration. Factors known to be associated with muscle damage variability include age, sex, race, training status, hydration status, and body mass (2, 23), as well as genetic variability (9, 13, 15, 33, 34).
We recently described several genetic polymorphisms associated with exertional muscle damage (i.e., functional losses or muscle protein efflux) variability, including those in the myosin light chain kinase (MYLK), α-actinin 3 (ACTN3), and insulin-like growth factor II (IGF2) genes (9, 13). Clarkson et al. (9) found that the C49T and C37885A rare alleles of MYLK were associated with increased CK, Mb, and strength losses postexercise, whereas the ACTN3 577X allele was associated with lower preexercise CK compared with the R577 allele. Devaney et al. (13) reported that the rare alleles for IGF2 polymorphisms, rs3213220 and rs680, were found to be associated with increased strength loss, soreness, and CK activity following eccentric exercise. Others also have reported associations between CK response and polymorphisms in the muscle-specific CK (CK-MM) (15), angiotensin-converting enzyme (ACE) (15, 33), interleukin-6 (IL6), and tumor necrosis factor-α (TNFA) genes (34). These studies indicated that genes involved in muscle structure (ACTN3) or that contribute to growth (IGF2), inflammation (TNFA and IL6), or force production (MYLK) can harbor polymorphisms that affect baseline CK and exacerbate the muscle damage response to eccentric exercise, generally predisposing those with the rare alleles to greater indexes of damage.
In a recent study (16), we found increased chemokine (C-C motif) ligand 2 (CCL2) mRNA following repeated bouts of eccentric exercise. CCL2 (NM_002982; 17q11.2) is a chemotactic cytokine known to be produced by both macrophages and satellite cells and was recently found to increase satellite cell proliferation in culture (32). CCR2 (NM_000647; 3p21.31) acts as a receptor for CCL2, CCL7, and CCL13 and mediates action via calcium mobilization. Warren et al. (31) showed that CCR2 knockout mice demonstrated impaired regeneration following freeze injury, and we (16) previously demonstrated a 9.0-fold increase in CCL2 mRNA expression following a novel bout of eccentric exercise and a subsequent additional 2.6-fold increase in CCL2 mRNA expression following a repeated exercise bout in humans. Together, these data suggest that CCL2 and CCR2 play significant roles in the muscle damage, repair, and adaptation responses to eccentric exercise.
We hypothesized in the present study that CCL2 and CCR2 genetic variants are associated with variability in exercise-induced damage markers following eccentric exercise. Specifically, we expected variants that have been shown to (or are predicted to) decrease CCL2/CCR2 levels would be associated with higher indexes of damage and attenuated recovery.
MATERIALS AND METHODS
Data for this study were derived from a larger clinical trial assessing the efficacy of a topical analgesic on delayed onset muscle soreness (8, 24), which is the same cohort as the ACTN3, MYLK, and IGF2 SNP association reports previously published (9, 13). Markers of muscle damage included muscle function, muscle soreness, CK, and Mb levels. Associations in the current study were not significantly influenced by the treatment; therefore, we included all subjects in the cohort, with all statistical analyses covaried for treatment group.
Subjects provided informed written consent for the overall study during screening and a supplemental informed consent for the additional blood sample used for DNA sequencing. Both consent forms as well as the study protocols were approved by the University of Massachusetts Human Subjects Review committee.
Subjects agreed to refrain from analgesic use, muscle treatments, strenuous or new physical activity, and alcohol during the study. Potential subjects having occupations that required heavy weight-lifting were excluded from participation, as well as those that had participated in a resistance training program in the previous 6 mo and those with baseline blood values outside of normal range, known muscle disorders, existing myopathy, diabetes mellitus, or hyperthyroidism. A physician screened all potential subjects and determined whether subjects were healthy and eligible for the study.
During visit 1, subjects were screened for inclusion/exclusion criteria and signed informed consent documents. Within 1 wk following visit 1, subjects were seen by the study physician, height and weight were recorded, and subjects had blood drawn for baseline analyses (visit 2). Visit 3 occurred 2 days following visit 2, during which preexercise soreness and strength were assessed, followed by the eccentric exercise bout, followed immediately by reassessment of strength. Twelve hours after visit 3, subjects returned for treatment (visit 4). Strength recovery was assessed at 3 (visit 5), 4 (visit 6), 7 (visit 7), and 10 days postexercise (visit 8). Blood was drawn at 4, 7, and 10 days postexercise.
Maximal voluntary isometric strength of the elbow flexors was assessed before and immediately after the exercise. Strength was assessed on a modified preacher bench, with the elbow flexed at 90°. A strain gauge (model 32628CTL; Lafayette Instrument, Lafayette, IN) was used to measure the force produced. Three trials were averaged, with 1 min of rest between sets.
All blood samples were collected via venipuncture from the antecubital vein into the appropriate vacutainers. Samples were shipped to a certified clinical laboratory (Holyoke Hospital, Holyoke, MA) for analysis of serum CK activity and Mb concentration. Blood samples were taken at preexercise and 4, 7, and 10 days postexercise.
The single exercise bout consisted of 50 maximal eccentric (muscle lengthening) contractions of the elbow flexor muscles of the nondominant arm (8, 9, 20, 23, 24). Two sets of 25 contractions were separated by a 5-min rest period. Each contraction was 3 s long, followed by 12 s of rest. Study personnel verbally encouraged subjects to maintain maximal effort to maximally contract their elbow flexor muscles during the exercise, resisting the downward motion of the lever as an instructor moved the lever until the arm was fully extended. Hydration was maintained by requiring subjects to drink water before and after exercise. Subjects were also instructed to maintain hydration during the 10-day recovery period.
SNP Selection and Genotyping
The blood samples for DNA sequencing were drawn in EDTA vacutainer tubes and shipped deidentified to Children's National Medical Center (Washington, DC) for DNA extraction and subsequent genotyping. Genomic DNA was extracted from whole blood samples using the PUREGENE DNA purification system (Gentra Systems, Minneapolis, MN) according to the manufacturer's instructions.
SNPs from CCL2 known to affect CCL2 protein levels (rs1024610 and rs1024611; Ref. 18) were included for analysis, as well as potentially functional SNPs selected using two different programs, FASTSNP (35) and PupaSuite (12, 21). Both of these programs offer selection of SNPs with potential phenotypic effects such as silencers and microRNAs including their targets, as well as additional methods for predicting SNPs in transcription factor binding sites and splice sites. SNP details, including location, type, and minor alleles, are listed in Table 1.
Genotypes for all SNPs in this study were obtained with the use of a TaqMan allelic discrimination assay that employs the 5′ nuclease activity of Taq polymerase to detect a fluorescent reporter signal generated during PCR reactions. The PCR reactions for the each SNP contained 20 ng of DNA, 900 nM primers, 200 nM probes, and TaqMan Universal PCR Master Mix, No AmpErase UNG (Applied Biosystems, Foster City, CA) in a final volume of 15 μl. PCR was performed on a MJ Research Tetrad thermal cycler (Waltham, MA). For each SNP, the primers and probes are listed in Table 2. The PCR profile was 10 min at 95°C (denaturation), 44 cycles of 15 s at 92°C, and 1 min at an annealing temperature of 60°C.
For this study, we used several methods of quality control for genotyping: negative controls, sequence-confirmed positive control (known genotypes for each plate from the HapMap samples), and genotyping of replicate samples (at least 5% of total). In addition, SNPs were also checked for Hardy-Weinberg equilibrium, and SNPs out of Hardy-Weinberg equilibrium (P < 0.01) were excluded. Genotype analysis was performed using an ABI 7900 (Foster City, CA) and the latest version of the 7900HT Sequence Detection Software (SDS version 2.3) with genotype calls confirmed by manual checking by the technician.
Hardy-Weinberg equilibrium was determined for each SNP using a χ2 test to compare the observed genotype frequencies with those expected under Hardy-Weinberg equilibrium. Analyses of demographic/physical characteristics (sex, age, and body mass) consisted of χ2 tests and one-way analysis of variance (ANOVA) with Sidak post hoc tests, where appropriate. Linkage disequilibrium between SNPs (within CCL2 and within CCR2) was determined by calculating r2 values between each SNP pair.
Normality of each quantitative trait was tested using the Shipiro-Wilk normality test. Phenotype measurements were compared in relation to SNP genotypes using analysis of covariance (ANCOVA) methods. The ANCOVAs again used Sidak post hoc tests to control for multiple tests. To account for any possible effects of sex, we used sex-specific models and covaried analyses for age, body mass, and treatment. The percentage of variation attributable to each SNP was determined with a likelihood-ratio test comparing the full model containing genotype and all covariates to the constrained model containing only covariates. Phenotypes taken at multiple time points were analyzed in a sex-specific model using a repeated-measures ANCOVA that included a genotype × time interaction term. Covariates were age, body mass, and treatment.
Given that other SNP associations have been previously published from this cohort (9, 13), we ran further statistical models incorporating ACTN3 and IGF2 loci from those publications as additional dependent variables (when associations overlapped on specific phenotypes).
Data are means ± SE or adjusted means ± SE where appropriate. All analyses were performed using Stata version 8.2 (StataCorp, College Station, TX). We set the significance level at P < 0.05, but given that we performed multiple statistical tests, those tests with 0.05 > P < 0.01 should be viewed with caution and require validation in future studies.
Demographics for the subject cohort are as follows: n = 157 (75 men, 82 women); mean age, 25.3 ± 5.4 yr; mean height, 170.8 ± 9.9 cm; and mean body mass, 73.5 ± 17.0 kg. Mean body mass index was 24.8. Average preexercise CK was 115 IU/l.
Average Exercise Response and Variability
Average strength loss immediately postexercise for all subjects was −53.1%. At 3 days postexercise, average strength loss was −42.8%, whereas 4-, 7-, and 10-day strength losses were −38.3, −25.7, and −18.5%, respectively. CK activity increased to 7,262 IU at 4 days, 3,057 IU at 7 days, and 491 IU at 10 days after the lengthening exercise bout. Average preexercise Mb was 38 μg/l, with increases to 392 μg/l at 4 days, 96.5 μg/l at 7 days, and 40.5 μg/l at 10 days after the lengthening exercise bout. Together, these data indicate the presence of moderate exercise-induced muscle damage, with recovery toward preexercise levels over the 10 days postexercise.
There was a large degree of variability in the response of individuals to exercise [described in detail in Sewright et al. (24)]. For example, CK at 4 days postexercise ranged from 55 to 80,550 IU/l. Changes in maximal voluntary contraction (MVC) strength immediately postexercise ranged from +29 to −90.8%, whereas MVC loss at 7 days postexercise ranged from +36 to −85.5%. Soreness at 12 h postexercise ranged from 0 to 94 mm on the 100-mm scale. There tended to be sex differences for several damage markers in the total cohort (i.e., ∼8% greater immediate strength loss in women and higher CK response in men), and these differences are also reported elsewhere (24). To account for this, all genotype-phenotype association analyses reported in the present study were sex specific.
Genotype-Phenotype Significant Associations
Genotype and allele frequencies for all SNPs tested are listed in Table 3. We found significant phenotype associations with four CCL2 SNPs (rs1024611, rs1860189, rs3917878, and rs13900) and three CCR2 SNPs (rs768539, rs3918358, and rs1799865). All of the SNPs tested were in Hardy-Weinberg equilibrium. Some of the polymorphisms tested were in linkage disequilibrium with one another (Tables 4 and 5), and similar effects were found in SNPs in high linkage disequilibrium.
Phenotypes found to be associated with genotypes in CCL2 were preexercise CK in women with rs13900, rs1024611, and rs1860189 (Table 6); CK change over time (genotype × time interaction) in women with rs3917878 (Fig. 1); and strength recovery (genotype × time interaction) in men with rs3917878 (Fig. 2). Percentages of variability in phenotypes explained by genotype are listed in Table 6. In all but one case, the presence of the minor allele was associated with greater indexes of damage (i.e., higher postexercise CK and attenuated strength recovery).
Phenotypes found to be associated with genotypes in CCR2 were preexercise CK in women with rs1799865 (Table 6); preexercise strength in men with rs768539 and rs3918358 (Table 6); soreness over time in the total cohort (not shown; P = 0.03); and strength recovery in women with rs3198358 (Fig. 3). Again, the presence of the minor allele in these SNPs was associated with greater indexes of damage (i.e., higher CK, attenuated strength recovery, and elevated soreness) in all but one case. Soreness was affected at 7 and 10 days postexercise (P = 0.03) with the following means: day 4, genotype TT = 49 ± 2 mm (SE), genotype CC/CT = 48 ± 2 mm; day 7, TT = 48 ± 3 mm, CC/CT = 53 ± 3 mm; and day 10, TT = 41 ± 2 mm, CC/CT = 48 ± 2 mm.
Interactions Between CCL2/CCR2 Loci and ACTN3 and IGF2 Loci
Five significant genotype-phenotype associations from the current study overlapped in phenotype with previously reported ACTN3 and IGF2 associations (no overlap with MYLK). In all but one case, additional loci variables did not significantly affect P value or percent variability explained by the CCL2/CCR2 loci (all P values <0.05). The one exception was the association of CCL2 rs1860189 with baseline CK in women, for which the P value of the CCL2 association increased from 0.049 to 0.061 with the addition of ACTN3 and IGF2 genotypes, becoming nonsignificant using the 0.05 a priori cutoff.
There is a large degree of variability in the damage response of individuals to eccentric exercise, with higher responders at risk for potentially fatal rhabdomyolysis. We and others (8, 13, 15, 33, 34) have reported that some of this variability is associated with genetic polymorphisms across a subset of genes. In this study, we have demonstrated that variations in the CCL2 and CCR2 genes are associated with severity of exercise-induced muscle damage markers. In all but one case, the presence of the rare alleles exacerbated strength loss, prolonged strength recovery, and elevated soreness and circulating CK activity.
CCL2 [also known as monocyte chemoattractant protein-1 (MCP-1)] is a small chemokine known to play key roles in inflammation and immunoregulation. CCR2 is a membrane-spanning G protein-coupled receptor that binds CCL2 as well as CCL7 and CCL13. We and others have previously demonstrated that CCL2 mRNA expression is dramatically upregulated after injury, in both animal and human models. No data suggest differential regulation of either CCL7 or CCL13 after exercise.
Using microarray screening, Summan et al. (25) first demonstrated a large increase (∼8-fold by microarray; 32-fold by quantitative RT-PCR) in CCL2 mRNA expression following freeze injury in the mouse model. In a recent investigation (16) targeting possible molecular mechanisms mediating the repeated bout effect (i.e., the adaptation that takes place in skeletal muscle after a single bout of eccentric exercise, rendering it less susceptible to damage from subsequent exercise bouts), we were surprised to find further elevations in CCL2 mRNA expression following a second eccentric exercise bout. In that study (16), CCL2 mRNA expression increased 9-fold after the first bout of exercise and another 2.6-fold from bout 1 to bout 2. CCR2 mRNA expression was unchanged between bouts. We colocalized CCL2 protein expression to both satellite cells and macrophages present in the muscle tissue (16), theorizing that upregulations in CCL2 expression were driving communications between these two cell populations, potentially enhancing recovery.
Preexercise Genotype-Phenotype Associations
In the current study, we found moderate associations between CCL2 and CCR2 genotypes and phenotypes at baseline, suggesting that CCL2 activity (as a product of CCL2 expression and the availability of its receptor) could be important to long-term muscle health. In men, CCR2 genotype (2 different SNPs predicted to alter transcription factor binding sites) affected preexercise strength, with the minor alleles (see Table 2) causing ∼10% lower maximal elbow flexion strength, explaining ∼4% of variability in this phenotype. Further work is needed to identify mechanisms underlying these associations, as well as their sexual dimorphism. Our previous studies with this cohort did not find any associations between genotype and baseline strength, and this association was not affected by previously reported loci.
Four CCL2/CCR2 variants were associated with preexercise CK levels in women (but not men). Women generally demonstrate higher baseline CK levels (29), which could enrich variability for SNP association. Two CCL2 SNPs in high linkage disequilibrium were associated with 35% elevation in CK levels [accounting for a large amount (11%) of CK variability] and two SNPs (1 each in CCL2 and CCR2) that are associated with lower CK levels ∼20% (accounting for ∼5% variability). Although only one of these SNPs (rs1024611) has been previously demonstrated to affect CCL2 protein levels (18) [with the minor allele (G) driving higher CCL2 levels], each of these polymorphisms is predicted to alter CCL2 levels via enhanced or repressed transcription factor binding sites (27, 28). How alterations of CCL2 levels affect levels of CK activity in blood at baseline and why these effects would be sexually dimorphic are currently unknown. The association between rs1860189 and baseline CK was slightly modified (P value from 0.049 to 0.06) by the addition of ACTN3 and IGF2 genotype information, indicating that multiple factors impact baseline CK.
Damage Response and Recovery
Exercise-induced muscle damage is modeled as product of strain-induced structural defects incurred during the exercise itself and secondary damage caused by inflammation and dysfunction in excitation-contraction coupling (10). These data suggest that CCL2 and CCR2 variants do not have significant effects during initial stages of exercise-induced muscle damage, as indicated by the lack of associations between the SNPs tested and strength losses immediately after exercise. This is consistent with the proposed mechanism of CCL2's action, that it is a key molecule in satellite cell and macrophage signaling (4, 5, 16). Macrophage infiltration into damaged tissue typically peaks in the days following exercise and is integral to muscle repair (5, 26, 31). Recently, studies have demonstrated a shift in macrophage phenotype from proinflammatory at the time of recruitment to anti-inflammatory supporters of myogenesis (1). The actions of the muscle's satellite cells are also integral to muscle repair (3, 14). Both cell types have been demonstrated to express CCL2, and coculture with both cell types enhances chemotaxis (5).
We found that genetic variants in CCL2 and CCR2 were associated with higher soreness and CK efflux following eccentric exercise and delayed strength recovery. None of these associations was modified by the addition of previously reported loci (ACTN3 and IGF2) to the model. The effects of CCL2 and CCR2 genotype were specific to the time frame (i.e., 4–10 days postexercise) during which muscle repair is peaking. Each of the SNPs associated with soreness, CK, and strength recovery are predicted to either enhance splicing or change transcription factor binding sites. For example, the T allele in the rs3918358 SNP (CCL2) is predicted to eliminate several CCAAT binding sites. In a previous study, we found transcriptional increases in CCAAT-enhancing binding protein-δ (CEBPD) along with increases in CCL2 mRNA with a repeated bout of eccentric exercise (16). CEBPD binds to CCL2 using these CCAAT binding sites, so it is possible that this SNP eliminates this coactivator of CCL2 expression, impairing the ability of muscle to repair. Inefficient repair can also keep CK levels in circulation high (via impaired sealing of membrane) and increase soreness via residual chemotactic signals.
We acknowledge several potential limitations to the current study. Although our population is large for an exercise-induced muscle damage study, some of the genetic variants that we studied had smaller frequencies for the minor allele than others, which limited some genotype groups to a few individuals. In these cases, we have reported dominant or recessive association models rather than full models. Further validations using larger sample sizes would be valuable. In addition, although we took several factors (i.e., age and sex) into account in our analyses, other factors such as body composition or fitness levels could contribute some variability to the damage response and should be examined in future studies. Finally, although the addition of ACTN3 and IGF2 loci to the CCL2/CCR2 association models in the current study did not contribute significantly to variability in damage indexes, it is still possible that multiple loci interact to drive damage response variability.
In summary, these data add evidence to our previous work with ACTN3, MYLK, and IGF2 showing that genetic variability, at least in part, underlies susceptibility to exertional rhabdomyolysis. We found that variations in CCL2 and its receptor are associated with preexercise muscle function and CK values, as well as muscle damage markers during repair from eccentric exercise. Indexes of primary muscle damage were not associated with CCL2 genotypes in the current study.
This study was funded by Medinova, Incorporated.
No conflicts of interest, financial or otherwise, are declared by the author(s).