Statins exacerbate exercise-induced skeletal muscle injury. Muscle-specific microRNAs (myomiRs) increase in plasma after prolonged exercise, but the patterns of myomiRs release after statin-associated muscle injury have not been examined. We examined the relationships between statin exposure, in vitro and in vivo muscle contraction, and expression of candidate circulating myomiRs. We measured plasma levels of myomiRs, circulating microRNA-1 (c-miR-1), c-miR-133a, c-miR-206, and c-miR-499-5p from 28 statin-using and 28 nonstatin-using runners before (PRE), immediately after (FINISH), and 24 h after they ran a 42-km footrace (the 2011 Boston marathon) (POST-24). To examine these cellular-regulation myomiRs, we used contracting mouse C2C12 myotubes in culture with and without statin exposure to compare intracellular and extracellular expression of these molecules. In marathoners, c-miR-1, c-miR-133a, and c-miR-206 increased at FINISH, returned to baseline at POST-24, and were unaffected by statin use. In contrast, c-miR-499-5p was unchanged at FINISH but increased at POST-24 among statin users compared with PRE and runners who did not take statins. In cultured C2C12 myotubes, extracellular c-miR-1, c-miR-133a, and c-miR-206 were significantly increased by muscle contraction regardless of statin use. In contrast, extracellular miR-499-5p was unaffected by either isolated statin exposure or isolated carbachol exposure but it was increased when muscle contraction was combined with statin exposure. In summary, we found that statin-potentiated muscle injury during exercise is accompanied by augmented extracellular release of miR-499-5p. Thus c-miR-499-5p may serve as a biomarker of statin-potentiated muscle damage.
- hydroxymethylglutaryl-CoA reductase inhibitors
- muscle contraction
- skeletal muscle
hydroxymethylglutaryl co-a reductase inhibitors, or statins, reduce blood cholesterol and the risk of atherosclerotic cardiovascular disease (43). Although well-tolerated by most individuals, statin use may result in skeletal muscle complications that include myalgia, muscle weakness, and rhabdomyolysis (35). The molecular pathways underlying statin-induced muscle damage are incompletely defined, but statins can potentiate the efflux of muscle enzymes such as creatine kinase (CK) into the blood stream (35). Vigorous physical exercise can also induce muscle fiber damage and increase cellular membrane permeability, and statins exacerbate exercise-induced skeletal muscle injury (33, 45). Although blood CK levels are used to quantify skeletal muscle injury in the settings of both exercise and statin intolerance, CK levels alone do not reliably correlate with clinical symptoms (13, 36) or the degree of muscle damage (19, 27). Indeed, additional biomarkers of statin-associated muscle complaints and injury would be of great utility, but such alternative statin-relevant laboratory measurements are not currently available.
MicroRNAs (miRNAs) are small, nonprotein-coding RNA molecules that regulate cellular function at the posttranscriptional level (2). In response to tissue stress, miRNAs may be released into the bloodstream. Circulating miRNAs (c-miRNAs) may facilitate cell-to-cell communication and be useful biomarkers of source tissue activity both in health and disease (9, 38). Certain miRNAs are specifically expressed in cardiac and skeletal muscle (known as myomiRs) and participate in the regulation of various muscle functions, including myoblast proliferation, differentiation, contractility, and stress responsiveness (including skeletal muscle fiber type switch and protection from cardiomyocyte apoptosis) (48). These myomiRs can be detected in the circulation and have recently been examined as markers of acute muscle damage and chronic muscle disease (14, 16).
Dynamic and molecule-specific regulation of circulating myomiRs has been recently reported during exercise (4, 6, 12, 25, 28, 46). Specifically, we found that plasma levels of certain circulating myomiRs such as c-miR-133a rise and fall rapidly following prolonged aerobic exercise, whereas in contrast, others, including c-miR-499-5p, rise more gradually and remain elevated for longer periods of time (5). Although these results suggest a potential role of specific c-myomiRs as unique markers of exercise-induced tissue response, c-myomiR secretory patterns at the cellular level have yet to be described and their responses to statin-associated muscle injury remain unknown. We hypothesized that specific c-myomiRs are differentially regulated both by muscle contraction and statin exposure. To interrogate this hypothesis, we used a combination of in vivo human exercise physiology and in vitro cell culture modeling to examine myomiR expression in the context of repetitive muscle contraction with and without concomitant statin exposure. Specifically, we examined plasma profiles of c-miR-1, c-miR-133a, c-miR-206, and c-miR-499-5p in statin-using and nonstatin-using marathon runners before, immediately after, and 24 h after a marathon. To confirm that the observed in vivo expression patterns reflect the fundamental cellular characteristics of muscle undergoing sustained exercise, we measured intracellular and extracellular levels of representative myomiRs in fully differentiated C2C12 myotubes after experimentally induced cellular contraction with and without statin treatment.
MATERIALS AND METHODS
Participants were healthy marathon runners free of known cardiovascular or metabolic disease except for hypercholesterolemia who participated in the 42-km Boston Athletic Association Marathon held in April 2011 (5, 33). For the present study, the statin group consisted of runners who used statins continuously for >6 mo before study enrollment for whom we had adequate banked plasma for the completion of all c-miRNA analyses, whereas the control group consisted of age-, gender-, and race finish time-matched runners who were not taking statins. All participants were nonsmokers, aged ≥35 yr, and not taking oral contraceptives or hormonal therapy. Participants provided written informed consent at the time of study enrollment. All protocol procedures complied with the Declaration of Helsinki and were approved by the institutional review board at Hartford Hospital (Hartford, CT).
In vivo study design.
Participants were enrolled on the day prior to the marathon following a tapering period that consisted of 1-2 wk of reduced exercise training. Body mass, height, resting blood pressure, and heart rate were measured at the time of enrollment. Conventional biomarker and c-miRNA profiles were assessed using a prospective, longitudinal, and repeated-measures study design. Specifically, venous blood was obtained on the day before the marathon (PRE), immediately after completion of the marathon (FINISH), and at ∼24 h after FINISH (POST-24). Venous blood (10 ml) was collected in standard anticoagulant (EDTA)-treated vacutainer tubes at each of the three time points using a 20-gauge intravenous catheter placed into a hand or arm vein. All blood samples were centrifuged at 2,000 g for 10 min to pellet cellular elements immediately after each blood draw. The supernatant plasma was then aliquoted and immediately frozen at −80°C to minimize freeze-thaw degradation. CK was measured as previously reported (33).
RNA extraction from human plasma samples.
The thawed plasma samples were further centrifuged (21,100 g, 10 min) to remove any remaining cellular contents. The plasma supernatant was then aliquoted into 150-μl volumes, which were analyzed or stored at −80°C for future analysis. The number of freeze-thaw cycles of plasma samples was minimized. Additionally, all samples from a given individual were processed and analyzed in a single batch to reduce variability of results due to plasma handling. On the basis of previous observations, equivalent levels of exogenous miR-422b were used for quantitative normalization of c-miRNA plasma levels (4, 24). More specifically, 4 fmol of the chemically synthesized miRNA duplex mimic of miR-422b (Life Technologies, Carlsbad, CA) was added to 150 μl of each plasma sample. Total RNA extraction was performed using a MicroRNA Extraction Kit (Benevbio, Mission Viejo, CA). The reliability of the c-miRNA extraction data was supported by additional extraction of the known quantities of the exogenous miR-422b mimic, as previously performed (4).
C2C12 mouse myoblasts (American Type Culture Collection, Manassas, VA) were cultured in DMEM (Thermo Fisher Scientific, Waltham, MA) containing high glucose (4.5 g/liter) supplemented with 10% FBS (Thermo Fisher Scientific) and 1% penicillin-streptomycin (Thermo Fisher Scientific) in a humidified incubator with 5% CO2 at 37°C. Experiments were performed at passages 4–9. On reaching confluence, the medium was substituted with a medium consisting of DMEM supplemented with 2% horse serum (Mediatech, Manassas, VA) and 1% penicillin-streptomycin to induce differentiation into myotubes as described (18, 26). The differentiation medium was changed every 24 h.
In vitro study design.
After 5–7 days of differentiation, simvastatin (InSolution simvastatin; EMD Millipore, Billerica, MA) was added in the fully differentiated C2C12 myotubes at a concentration of 1.0 μM. The same amount of differentiation medium was used as vehicle control. After 22 h, C2C12 myotubes were washed with PBS, replenished with new differentiation medium, and then incubated with 100 μM carbamoylcholine chloride (carbachol; Sigma-Aldrich, St. Louis, MO) for 2 h to induce cellular contraction, as previously described (30, 31, 58). The cells and media were harvested for RT-quantitative PCR (RT-qPCR) 2 h later. For comparison, a baseline group was prepared in which statin or control incubation was maintained for 24 h without adding carbachol.
In vitro cytotoxicity assay.
Lactate dehydrogenase (LDH) released from the cytosol of damaged cells (44) was measured using the Cytotoxicity Detection Kit Plus (LDH; Roche Diagnostics, Indianapolis, IN) following the manufacturer's instructions. The percent cytotoxicity was calculated by dividing the activity of LDH in culture media by the total activity of LDH in media and lysed cells.
For use in standard polyacrylamide gel electrophoresis and immunoblotting, primary antibodies included a 5′-AMP-activated protein kinase (AMPK) α-antibody (Cell Signaling Technology, Beverly, MA) and a phospho-AMPK-α (Thr172) rabbit monoclonal antibody (Cell Signaling Technology). As a secondary antibody, a goat anti-rabbit IgG-HRP antibody (Santa Cruz Biotechnology, Dallas, TX) was used.
RNA extraction from cultured cells and conditioned media.
After specific treatment of C2C12 myotubes, conditioned media was removed and centrifuged (21,100 g, 10 min) to remove cells and cellular debris from the supernatant prior to RNA extraction. A chemically synthesized miRNA duplex mimic of miR-422b (4 fmol; Life Technologies) was added to 300 μl of filtered conditioned media for quantitative normalization of extracellular miRNA levels in culture media. miRNA extraction was then performed via column-based extraction (BenevBio) as previously described (32). In parallel, RNA was extracted from cultured C2C12 myotubes as previously described (7) by the RNeasy Mini kit (Qiagen, Valencia, CA). Total cellular RNA concentration was calculated by Nanodrop spectrophotometry (Thermo Fisher Scientific).
Quantification of miRNAs.
Four skeletal muscle-related myomiRs (miR-1, miR-133a, miR-206, and miR-499-5p) were selected in this study. As we previously described (5), the brain-enriched miR-134 was chosen as an unrelated control to facilitate comparison with these candidate miRNA species. In C2C12 myotube samples, miR-134 was not detected, and therefore miR-15a, a ubiquitous miRNA with specific functions in angiogenesis and cardiomyocyte apoptosis (17, 42), was selected as a comparator.
RT-qPCR was used to quantify levels of the candidate c-miRNAs. Reverse transcription was performed to generate cDNA (TaqMan MicroRNA Reverse Transcription Kit; Thermo Fisher Scientific). To amplify cDNA via fluorescently labeled TaqMan probe and primer sets, a Fast Real Time PCR device (7900HT; Applied Biosystems) was used. TaqMan primer/probes for quantification were used as follows (Thermo Fisher Scientific): hsa-miR-1 [AB 4427975 (002222)], hsa-miR-15a [AB 4427975 (000389)], hsa-miR-133a [AB 4427975 (002246)], hsa-miR-134 [AB 4427975 (001186)], hsa-miR-206 [AB 4427975 (000510)], hsa-miR-499-5p [AB 4427975 (001352)], hsa-miR-422b [AB 4427975 (000575)], and snoRNA202 [AB 4427975 (001232)].
The raw Ct values obtained by RT-qPCR were converted to absolute copy numbers of candidate miRNAs using the standard curves for each miRNA, as described in previous studies (24, 39). As we previously described (41), RT-qPCR allows the ability to detect at least three copies of a specific miRNA for a given reaction, thus allowing ample ability to quantify even relatively low levels of expression in plasma. To avoid potential inconsistency during RNA extraction, copy numbers were then normalized using a ratio calculated from levels of a miR-422b exogenously spiked into each given sample compared with an equivalent amount of miR-422b spiked in water (“gold standard”) (4). For intracellular miRNA quantification in C2C12 myotubes, copy numbers were normalized using a ratio calculated from levels of snoRNA202, an endogenous control in each sample compared with a mean value of snoRNA202 from baseline controls. These normalized copy numbers were then used to calculate fold change by normalizing to 1 the mean copy number measured prior to the marathon (PRE) or in the control group, to which each individual miRNA measurement was compared.
Normality of data was assessed using the Shapiro-Wilk normality test. Demographics are expressed as means ± SD for continuous data and numbers and percentages for categorical data. Differences in continuous variables were analyzed using the Student's t-test and categorical variables were analyzed using the χ2 test. Levels of c-miRNAs in plasma are presented as median (interquartile range). Plasma levels of c-miRNAs within groups were compared using the Friedman test followed by the Dunn's multiple comparison test. Across-group comparisons of plasma levels of c-miRNAs at each time point were compared using the Mann-Whitney test. For correlation analyses, Spearman's rank correlation coefficient was calculated for all nonnormally distributed data.
The results from the in vitro immunoblot assays for AMPK were expressed as means ± SE and were compared by Student's t-test. The results from the in vitro cytotoxic assays for LDH release were expressed as median ± interquartile range and were compared by Kruskal-Wallis one-way ANOVA, followed by a series of Mann-Whitney tests with a Bonferroni correction for post hoc analysis. The results from in vitro assays of miRNA levels derived from myotubes were also expressed as the median ± interquartile range. The levels of myotube-specific miRNAs in different conditions were compared using the Friedman test followed by the Dunn's multiple comparison test. In vitro data were derived from three or more independent experiments, each performed at least in triplicate. A two-tailed value of P < 0.05 was considered statistically significant. All statistical analyses were performed with SPSS version 19.0 (SPSS, Chicago, IL).
Influence of Statins on c-myomiRs During Prolonged Aerobic Exercise
Baseline characteristics of marathon participants.
There were no baseline differences between the statin-using and control runners with the exception of low-density lipoprotein (LDL) cholesterol and high-density lipoprotein cholesterol, both of which were lower in the group of runners that used statins (Table 1). The statin users used a variety of agents and doses (Table 2), and statin exposure was thus standardized as “atorvastatin equivalents” as follows: rosuvastatin 2.5 mg = atorvastatin 5 mg = simvastatin 10 mg = lovastatin 20 mg = pravastatin 20 mg = fluvastatin 40 mg (37). The average statin use potency was 14.9 ± 14.3 mg of atorvastatin, and statin potency was inversely related to total and LDL cholesterol levels in the statin group (r = −0.554, P = 0.002; r = −0.509, P = 0.006, respectively). There were no significant changes in hemoglobin levels at the conclusion of the marathon in statin users (PRE, 14.5 ± 1.3 g/dl vs. FINISH, 14.6 ± 1.4 g/dl, P = 0.405) and nonstatin users (PRE, 14.4 ± 0.9 g/dl vs. FINISH, 14.3 ± 1.0 g/dl, P = 0.094) but hemoglobin levels at POST-24 were significantly lower in both groups (13.4 ± 0.9 g/dl and 13.5 ± 1.2 g/dl, respectively; both P < 0.001 vs. PRE). Thus contraction of plasma volume after exercise was excluded as an explanation for any substantial change in c-miRNA expression levels at any time point of analysis.
Alterations in circulating miRNAs in response to prolonged exercise.
Before the marathon, the baseline plasma expression levels of all candidate c-miRNAs (c-miR-1, c-miR-133a, c-miR-206, c-miR-499-5p, and the unrelated control, c-miR-134) were extremely low and did not differ between the two groups (data are not shown). In both groups, c-miR-134 levels were increased at FINISH and displayed a trend to decline at POST-24 (Fig. 1A). Importantly, alterations in c-miR134 were modest compared with those of the c-myomiRs described below (Fig. 1, B–E) and thus changes in c-miR-134 provided a reference for the interpretation of changes in other candidate c-miRNAs.
Levels of c-miR-1, c-miR-133a, and c-miR-206 increased significantly at FINISH and returned to baseline at POST-24 (Fig. 1, B–D). Expression profiles of these three c-myomiRs did not differ between the statin and control groups at any time point. In contrast, c-miR-499-5p was unchanged at FINISH in both groups. However, at POST-24, runners who used statins demonstrated higher levels of c-miR-499-5p at the POST-24 time point both compared with their prerace baseline [2.9 (1.3, 8.6) fold change, P < 0.001] and with nonstatin-using runners [1.4 (0.9, 3.2) fold change, P = 0.027; Fig. 1E]. This profile was similar to that observed for CK, which correlated with c-miR-499-5p across all time points (R = 0.56, P < 0.001) and also demonstrated significantly higher levels among statin users at POST-24 (Fig. 1F). Therefore, the temporal expression profile of c-miR-499-5p differed from those of c-miR-1, c-miR-133a, and c-miR-206 and was impacted by statin use, similar to CK, a well-established marker of muscle damage. There were no significant correlations between age, body mass index, or finishing time with change in the levels of any candidate c-myomiR from PRE to FINISH and from PRE to POST-24 in either participant group (data not shown).
Influence of Muscle Contraction and Statin Exposure on c-myomiR Release from C2C12 Myotubes
Exposure of fully differentiated C2C12 myotubes to carbachol (100 μM) successfully induced cellular contraction, as evidenced by increased phosphorylation of AMPK (Fig. 2A), an established confirmatory marker of in vitro skeletal myocyte contraction (18, 26). To model the exposure of skeletal muscle to statins, 1.0 μM simvastatin was selected for in vitro study on the basis of previous work (11, 20) reporting statin toxicity without widespread cell death in cultured myotubes. Consistent with prior reports of statin toxicity in vitro (11, 20), extracellular media derived from C2C12 myotubes exposed to 1 μM simvastatin for 24 h demonstrated an increase in LDH activity compared with control (Fig. 2B). However, carbachol-induced myotube contraction did not increase LDH activity in either control or statin-treated myotubes (Fig. 2B).
Intracellular levels of miR-133a, miR-499-5p, and the nonrelated ubiquitous miR-15a remained relatively stable and did not vary from their baseline values with isolated statin exposure, isolated carbachol exposure, or simultaneous statin and carbachol exposure (Fig. 3, A, B, and E). Intracellular levels of miR-1 and miR-206 (Fig. 3, C and D) were unaffected by isolated statin exposure but declined modestly in the settings of isolated carbachol exposure and simultaneous statin and carbachol exposure.
Extracellular levels of miR-15a were not altered by statin or myotube contraction and remained at the border of detectability in all tested conditions (Fig. 4A). In contrast, extracellular levels of the miR-1 were increased by similar magnitude following isolated carbachol exposure or simultaneous statin and carbachol exposure but were unchanged after isolated statin exposure (Fig. 4C). Extracellular miR-133a and miR-206 were modestly increased by isolated statin exposure, but similar to miR-1, were not subject to any additive or synergistic effects of combination treatment with carbachol and statin (Figs. 4, B and D). In contrast, miR-499-5p was not significantly affected by either isolated statin exposure or isolated carbachol exposure, but it increased significantly in the setting of simultaneous statin and carbachol exposure (Fig. 4E).
We used a combination of in vivo and in vitro modeling of exercising muscle to evaluate the release of c-myomiRs during skeletal muscle contraction with and without statin exposure. We found a distinct statin-associated release program for extracellular miR-499-5p compared with other c-myomiRs in both marathon participants and cultured myotubes (Fig. 5). To our knowledge, this is the first study to evaluate the effects of skeletal muscle contraction and statin exposure on c-myomiR release and the first to report a specific c-myomiR as a potential marker of statin-potentiated muscle damage. These findings provide a framework for studies designed to examine the biology of c-miRNAs in statin-induced muscle injury.
Statins can adversely affect skeletal muscle (22, 35), but mechanisms underlying statin-induced myopathic symptoms are not understood. Although statin-induced expression of specific miRNAs in circulation of patients with unstable angina has been reported (15), the precise responses of c-myomiR to statin-associated muscle injury have remained undefined until now. Here, statin-related alterations with muscle contraction were specific to c-miR-499-5p, because other candidate c-myomiRs, including c-miR-1, c-miR-133a, and c-miR-206 were not influenced by statins with prolonged exercise in vivo (Fig. 1) and were not potentiated by statins during myotube contraction in vitro (Fig. 4). As an intracellular molecule, miR-499 regulates cardiomyocyte differentiation (55), protection from cardiomyocyte apoptosis (52, 53), and skeletal muscle fiber type switching (49). Notably, recent data have also demonstrated a role for miR-499 in the regulation of mitochondrial function (57), a process that has also been implicated as a cause of statin-induced muscular dysfunction (23). It remains unknown whether extracellular miR-499 may also function as a circulating endocrine messenger to induce such metabolic alterations systemically.
Beyond the implications of this study on our molecular understanding of statin biology, the statin-specific alterations to c-miR-499-5p observed here suggest that this circulating miRNA may serve as a unique marker of statin-induced muscle damage, augmenting the relevance of prior work proposing miR-499 as a biomarker of myocardial (1, 8, 51, 56) or skeletal muscle (16) damage in nonstatin contexts. To date, there are no widely accepted clinical biomarkers to diagnose statin-associated muscle complaints. CK has been used most frequently in this regard as a marker for statin-related muscle injury after brief (13, 45) and more prolonged exercise (33). Yet, myalgia can develop without CK elevation (13, 36, 40), CK levels alone also do not correlate with the degree of muscle damage (19, 27), and CK can increase with statin use in asymptomatic subjects (34) and certainly in other statin-independent contexts of myocardial or skeletal damage. The positive correlation of c-miR-499-5p with CK values across time points in this marathon runner cohort (Fig. 1F) suggests the relevance of this c-miRNA to the fundamental mechanisms of statin-induced muscle injury. Yet perhaps more importantly, identification of an alternative yet complementary marker of statin-associated muscle injury beyond CK could aid in surmounting the above issues of sensitivity and specificity when CK is used in isolation. Future work will be required to determine whether alteration of statin dosing, timing of administration, or exercise intensity or volume affect c-miR-499-5p kinetics. Clinical studies will be needed to examine the potential diagnostic and prognostic roles of this specific miRNA in specific patient populations.
Data derived from our in vitro myotube contraction model help clarify several key areas of uncertainty surrounding miRNA biology. Although previous work has documented that c-myomiRs increase during both eccentric (6, 54) and more prolonged aerobic exercise (5), mechanisms underlying this phenomenon are unclear. An emerging question includes whether exercise increases c-myomiRs by augmented intracellular miRNA production or by efflux of preexisting miRNAs (10, 21, 29). We observed no significant increase in intracellular myomiRs after 24 h of cellular contraction. This suggests that greater myomiR levels after skeletal muscle contraction both in vitro and in vivo result from cellular efflux rather than augmented production. Furthermore, our observation of differing extracellular expression profiles of miR-1, miR-133a, and miR-206 compared with miR-499-5p in both marathoners and cultured myotubes suggests that muscle contraction and/or muscle damage-induced c-myomiR release indeed occur with at least some degree of molecular specificity. A variety of miRNA efflux mechanisms have been described previously (3, 47, 50), but how each of these molecular packaging programs contributes to the rise and clearance in specific c-myomiRs is unknown.
The present study has several limitations. We examined only 28 statin-using and 28 control runners. However, the use of a longitudinal repeated-measures study design permitted each participant to serve as his or her own control thereby minimizing the effect of interindividual data variability. Statin use was determined by self-report, but LDL levels were lower in statin users and inversely correlated with the potency of the statin used, suggesting accurate self-reporting. Participants in the statin group used a variety of agents and doses as typical in a community-based cohort of individuals treated for dyslipidemia and primary prevention of cardiovascular disease. Given the size of our cohort, we were not able to effectively evaluate the relationships between statin agent, statin dose, and circulating microRNA profiles during marathon participation. We did not test the utility of miR-499-5p as a marker of statin-associate muscle problems in patients with verified statin-associate muscle complaints, an important area of future work, so the applicability of this marker in that population is unknown. Moreover, at this time, we cannot determine whether spontaneous statin-induced muscle injury produces similar changes in miR-499-5p. Despite these limitations, both our in vivo and in vitro models consistently indicate that c-miR-499-5p does not increase solely with statin use and requires muscle activation in statin users. Thus these pilot experiments suggest specific regulatory programs for myomiR release with the combination of statins and skeletal muscle contraction.
In conclusion, statin-potentiated muscle injury during exercise is accompanied by augmented extracellular release of miR-499-5p. These findings suggest a potential role of c-miR-499-5p as a biomarker of statin-potentiated skeletal muscle damage.
This work was supported by the McArthur-Radovsky, Lerner, Harris, and Watkins Funds (to S.Y. Chan) and the Hassenfeld Scholars Fund (to A.L. Baggish).
B.A. Taylor has served on a pharmacovigilance monitoring board for Amgen, P.D. Thompson has received an NIH Small Business Collaborative Grant via Genomas. P.D. Thompson has served as a consultant and received an Industry Sponsored Research Grant from Regeneron/Sanofi and Pfizer. P.D. Thompson has given lectures to Merck and AstraZeneca. No other conflicts of interest, financial or otherwise, are declared by the remaining authors.
P.-K.M., S.Y.C., and A.B. conception and design of research; P.-K.M., J.P., S.I., B.A.T., P.D.T., C.T., P.D., S.D., and A.B. performed experiments; P.-K.M., B.A.T., P.D.T., S.Y.C., and A.B. analyzed data; P.-K.M., B.A.T., P.D.T., S.Y.C., and A.B. interpreted results of experiments; P.-K.M., S.Y.C., and A.B. prepared figures; P.-K.M., S.Y.C., and A.B. drafted manuscript; P.-K.M., J.P., S.I., B.A.T., P.D.T., C.T., P.D., S.D., S.Y.C., and A.B. edited and revised manuscript; P.-K.M., J.P., S.I., B.A.T., P.D.T., C.T., P.D., S.D., S.Y.C., and A.B. approved final version of manuscript.
We thank S. Tribuna and D. Margaria for expert administrative assistance.
- Copyright © 2016 the American Physiological Society