Inorganic Nitrate Mimics Exercise-Stimulated Muscular Fiber-Type Switching and Myokine and γ-Aminobutyric Acid Release

Inorganic nitrate was considered a biologically inert metabolic end product of nitric oxide (NO) metabolism (1). However, the discovery of the nitrate-nitrite-NO pathway in mammals in which nitrate is reduced to the ubiquitous signaling molecule NO has led to renewed interest in the physiological role of nitrate (2,3). The discovery of the antiobesity and antidiabetic effects of nitrate in rodents has stimulated interest in nitrate as a therapeutic agent for the metabolic syndrome (4,5).

NO has multiple effects on skeletal muscle phenotypes, including mitochondrial biogenesis and fatty acid β-oxidation, glucose homeostasis, and contraction and fatigue (6,7). Augmenting the nitrate-nitrite-NO pathway may be a means of increasing bioavailable NO to improve muscle function. Systemic concentrations of nitrate can be enhanced through dietary consumption of nitrate-rich foods, such as green leafy vegetables (8). Studies in rodents and humans identifying that nitrate can improve exercise efficiency (9), tolerance (10), and performance (11) have led to nitrate gaining popularity as a nutritional exercise supplement.

We have demonstrated that dietary nitrate increases peroxisome proliferator–activated receptor (PPAR) γ coactivator 1α (PGC1α) expression in white adipose tissue (WAT) (12). In skeletal muscle, PGC1α is a master transcriptional regulator of the adaptive response to exercise (13). Transgenic PGC1α expression in skeletal muscle increases exercise performance and transcription of genes for oxidative slow-twitch type I and intermediate type IIa fibers (13). Endurance exercise training, partially through PGC1α transcriptional activity, also increases slow-twitch type I and intermediate type IIa myosin heavy chains (MYHs) within skeletal muscle (13,14). These type I and IIa fibers are characterized by high mitochondrial content and facilitate resistance to fatigue (15). Skeletal muscle also mediates many of the signals that contribute to the beneficial effects of exercise on health (16). PGC1α has been identified as a mediator of several nonmuscle metabolic effects of exercise through the regulation of exercise-dependent myokine signals (17,18).

Exercise is an effective intervention for both the prevention and the treatment of type 2 diabetes mellitus (T2DM) (19). Diabetic myopathy, characterized by reduced muscle functional capacity, atrophy, and a glycolytic fiber type, is a major diabetes complication (20). Furthermore, the proportion of type I fibers in muscle is positively correlated with systemic insulin sensitivity (21), and decreased levels of type I fibers are associated with insulin-resistant states, including T2DM (22). Therefore, we use transcriptional, protein, and metabolomic analyses to determine the effects of inorganic nitrate on markers of the adaptive response to exercise in skeletal muscle, including fiber-type phenotype, and PGC1α and exercise-dependent myokine production. We demonstrate that nitrate induces physiological adaptations in skeletal muscle and systemically that mimic exercise training and may underlie the beneficial effects of the small molecule on exercise tolerance, efficiency, and performance and on cardiometabolic risk.

Male Wistar rats (6 weeks old, 269 ± 2 g) (Charles River) were weight matched and received either distilled water containing 0.7 mmol/L NaCl (n = 9) or water containing 0.7 mmol/L sodium nitrate (NaNO₃) (n = 9) (Ultra Pure; Sigma-Aldrich) ad libitum for 18 days. Animals were housed in conventional cages at room temperature with a 12-h light/dark photoperiod. Morphological parameters are listed in Supplementary Table 1.

Muscle-specific PGC1α transgenic mice were generated and maintained as previously described (17). For the exercise experiments, 12-week-old B6 mice (The Jackson Laboratory) were conditioned with 3 weeks of free wheel running (n = 6) (14,17,18). Controls were age-matched sedentary littermates (n = 6). All procedures were carried out in accordance with U.K. Home Office protocols by a personal license holder.

Rats and mice were euthanized (sodium pentobarbital 200 mg/mL; Vétoquinol, Buckingham, U.K.). Blood was collected by cardiac puncture and centrifuged to obtain plasma. Soleus, gastrocnemius, or quadriceps muscle was removed and flash frozen in liquid nitrogen.

C2C12 myoblasts were grown to confluence in DMEM supplemented with 10% FBS. Differentiation was induced with DMEM containing 10% horse serum (Sigma-Aldrich) and 850 nmol/L insulin. During the 6-day differentiation, cells were cultured either with saline (control), 25 μmol/L NaNO₃, or 500 μmol/L NaNO₃ or, during the sodium nitrite (NaNO₂) studies, with saline (control), 500 nmol/L NaNO₂, 25 μmol/L NaNO₂, or 500 μmol/L NaNO₂ (Ultra Pure). Differentiation of C2C12 cells is unaffected by NaNO₃ (23). Pharmacological agent studies have used 2,2-bipyridyl (100 μmol/L), 2-phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl 3-oxide (PTIO) (50 μmol/L), l-N^G-nitro-l-arginine methyl ester (l-NAME) (1 mmol/L), ¹H-​[1,​2,​4]oxadiazolo[4,​3-​a]quinoxalin-​1-​one (ODQ) (1 μmol/L), GW742 (100 nmol/L), and sildenafil (1 μmol/L) (Sigma-Aldrich). Cells were treated with 2,2-bipyridyl, PTIO, l-NAME, or ODQ with and without 500 μmol/L NaNO₃. NaNO₃ and pharmacological agents were added at day 1 of differentiation and maintained for 6 days with the exception of 2,2-bipyridyl, which was added 6 h before the end of the experiment.

Primary adipose stromal vascular cells were isolated, cultured, and differentiated into adipocytes according to published methods (12,17). During differentiation, cells were cultured with either saline (control), 500 μmol/L NaNO₃ (Ultra Pure), 125 ng/mL irisin (BioVision), or 125 ng/mL irisin and 500 μmol/L NaNO₃.

FlexiTube small interfering RNA (siRNA) against PGC1α, AllStars Negative Control siRNA, and HiPerFect Transfection reagent were purchased from QIAGEN. C2C12 myotubes were transfected per the manufacturer’s instructions (75 ng siRNA, 2 μL transfection reagent per well, 10 nmol/L final siRNA concentration) on days 2 and 4 of differentiation.

Metabolites were extracted from soleus muscle of nitrate-treated rats; gastrocnemius and quadriceps muscle of exercise-trained mice; plasma of nitrate-treated rats, muscle creatine kinase (MCK)-PGC1α transgenic mice, exercise-trained mice, and nitrate-supplemented human volunteers; nitrate-treated and PPARδ agonist–treated C2C12 myotubes; and culture media as previously described (12).

Total RNA extraction from soleus and gastrocnemius muscle of nitrate-treated rats, nitrate- and irisin-treated adipocytes, and myocytes treated with nitrate, PPARδ agonist GW742, 2,2-bipyridyl, PTIO, l-NAME, ODQ, and sildenafil was performed as per published protocols (17). cDNA conversion and quantitative RT-PCR were also per published protocols (17). All data were normalized to 18S rRNA (mouse primary adipocytes, C2C12 myoblasts) or RLPL1 (rat muscle), and quantitative measures were obtained by using the ΔΔC_T method.

PGC1α, MYH7, and MYH2 were analyzed in nitrate-treated rat soleus. Irisin was analyzed in nitrate-supplemented human and rat plasma and nitrate-, GW742-, and sildenafil-treated myocyte media. Rat and human growth hormone was measured in nitrate-supplemented rat and human plasma, respectively. For all measurements, ELISA was performed per the manufacturer’s instructions (PGC1α and MYH7 kits: SEH337Ra and SED418Ra, respectively [Cloud-Clone, Houston, TX]; MYH2 kit: ABIN2093055 [antibodies-online.com, Aachen, Germany]; irisin kit: K4761-100 [BioVision, Milpitas, CA]; rat growth hormone kit: E-EL-R0029 [Elabscience, Beijing, China]; human growth hormone kit: EK0578 [Boster Biological Technology, Pleasanton, CA]).

Soleus and gastrocnemius cross-sections from nitrate-treated rats were prepared with a cryostat at a thickness of 10 μm. Immunohistochemical staining and fiber-type assessment were performed according to published methods (24). Antibodies were type I fiber, monoclonal antimyosin (skeletal, slow, clone NOQ7.5.4D; Sigma-Aldrich), and type II fiber, monoclonal antimyosin (skeletal, fast, alkaline phosphatase conjugate, clone MY-32; Sigma-Aldrich).

Gas chromatography–mass spectrometry (GC-MS) and data analysis were performed on metabolites from nitrate-treated rat soleus by using methods described previously (12).

Plasma nitrate from rats and humans was measured as described previously (12).

Liquid chromatography–mass spectrometry (LC-MS) analysis was performed on metabolites from the plasma of nitrate-supplemented humans and rats and on MCK-PGC1α mice, soleus of nitrate-treated rats, nitrate- and GW742-treated myotubes and culture media, and gastrocnemius and quadriceps muscles of exercise-trained mice by using a 4000 QTRAP mass spectrometer (Applied Biosystems/Sciex) coupled to an Acquity UPLC system (Waters Corporation, Manchester, U.K.) according to published methods (12,17).

A crossover, randomized, placebo-controlled, double-blind clinical trial of healthy, nonobese volunteers given either concentrated beetroot juice (2 × 70 mL/day, 12.0 mmol nitrate) or nitrate-depleted beetroot juice (placebo, 2 × 70 mL/day, 0.003 mmol nitrate) for 7 days was conducted. Nineteen participants (mean age 64.7 ± 3.0 years) completed the study. Participant characteristics are listed in Supplementary Table 2. Exclusion criteria were high resting blood pressure (systolic >180 mmHg and/or diastolic >110 mmHg), high physical activity level (>15,000 steps/day), a weight change of >3.0 kg in the past 2 months, and a diagnosis of metabolic, cardiovascular, and inflammatory conditions or diabetes. Participants were randomized to 1) nitrate supplementation (beetroot juice 2 × 70 mL/day containing ∼12.0 mmol nitrate) or 2) placebo (nitrate-depleted beetroot juice 2 × 70 mL/day, containing ∼0.003 mmol nitrate) for 7 days. Nitrate-depleted beetroot juice was prepared according to published methods (25). After a 7-day washout period, the second 7-day intervention was conducted. Blood serum was collected at baseline and termination of each intervention and aliquoted. Participants were fasted for 12 h and avoided exercise for 3 days before each visit. Serum nitrogen oxide (NOx) concentrations were measured by GC-MS as previously described (26). The trial was approved by the North of Scotland Research Ethics Committee 2, Aberdeen, U.K. (14/NS/0061).

Error bars represent SEM. P values were calculated by either t test or one-way or two-way ANOVA as stated with a Tukey post hoc test. Human study data were analyzed by using paired t test.

PGC1α activation may underlie the effects of nitrate on exercise; therefore, we analyzed the effect of nitrate on PGC1α expression in skeletal muscle in vivo. Wistar rats were treated with 0.7 mmol/L NaCl or 0.7 mmol/L NaNO₃ in drinking water for 18 days. Plasma nitrate concentrations increased from 7.1 ± 1.0 μmol/L in NaCl-treated control rats to 12.8 ± 1.1 μmol/L in nitrate-treated rats (P ≤ 0.001). Water and food intake was not significantly different between groups (Supplementary Table 1). Nitrate increased PGC1α expression in the soleus muscle of rats (Fig. 1A).

Hypoxia-inducible factor 2α (HIF2α) is a skeletal muscle–selective PGC1α transcriptional target that is positively regulated by exercise and plays a key role in muscle function (14). HIF2α expression was increased in the soleus of nitrate-treated rats (Fig. 1B). PGC1α and HIF2α coordinate to induce fiber-type switching from glycolytic type IIb fibers to intermediate type IIa and oxidative type I fibers as an adaptation to exercise training (13,14). Nitrate also increased the expression of MYH7, a fundamental constituent of type I muscle fiber, and MYH2, an important component of type IIa fibers (Fig. 1C). Concomitantly, the expression of MYH4, specific to fast-twitch type IIb fibers, was decreased. Expression of additional genes characteristic of slow-twitch muscle fibers, such as myoglobin (Mb) and calmodulin 2 (CALM2), was increased in the soleus by nitrate treatment (Fig. 1C).

To determine whether transcriptional changes in fiber-type–specific genes translated into corresponding changes in protein levels, the concentrations of PGC1α, MYH2, and MYH7 protein in the soleus of nitrate-treated rats were analyzed with ELISA. Nitrate increased the protein concentration of the three (Fig. 1D–F).

An increase in type I fibers drives a switch in energetic fuel usage away from glycolytic metabolism. Therefore, we used GC-MS–based metabolomics to analyze glycolytic intermediates from soleus muscle of nitrate-treated rats. Nitrate decreased the concentrations of the key glycolytic intermediates glucose-6-phosphate (Fig. 1G) and 3-phosphoglycerate (Fig. 1H). Consistent with these data, immunohistochemical analysis of soleus muscle from control and nitrate-treated rats confirmed that nitrate induces a decrease in constituent type II fibers and an increase in type I fibers (Fig. 1I and J).

To establish whether the effects in slow-twitch soleus were muscle type specific, the expression of genes determining fiber type were examined in the mixed fiber-type gastrocnemius muscle of nitrate-treated rats. Nitrate increased PGC1α and HIF2α expression in the gastrocnemius (Fig. 2A). The expression of the type I and type IIa fiber markers MYH7, Mb, and MYH2 was increased to a greater extent in gastrocnemius than in soleus after nitrate treatment, possibly representing the greater capacity for gastrocnemius to undergo a switch toward oxidative fiber types compared with the already highly oxidative soleus. Nitrate also decreased the expression of the type IIb fiber component MYH4 (Fig. 2A). Histological examination of the gastrocnemius muscle of nitrate-treated rats confirms an increase in type I muscle fibers and a concomitant decrease in type II muscle fibers (Fig. 2B and C). Taken together, these data suggest that nitrate is an inducer of muscle fiber-type switching, increasing the proportion of oxidative slow-twitch fibers to fast-twitch glycolytic fibers within muscle.

We next focused on establishing the mechanism underlying nitrate-induced fiber-type switching. C2C12 mouse myoblasts were differentiated into myotubes in the presence of nitrate to establish whether nitrate functions directly at the muscle to increase type I fiber–associated gene expression. Under normoxic conditions, HIF2α protein is hydroxylated by prolyl hydroxylase domain–containing enzymes (PHDs), ubiquitinated, and degraded by the proteasome (27). HIF2α protein must be stabilized after PGC1α-induced expression to activate fiber-type switching (14); therefore, the myotubes were also treated for 6 h with 2,2-bipyridyl to inactivate PHDs and stabilize HIF2α. Expression of PGC1α was significantly and similarly increased in myotubes treated with either nitrate alone or nitrate in combination with 2,2-bipyridyl (Fig. 3A). After HIF2α stabilization, treatment of myotubes with nitrate increased expression of type I and type IIa fiber-type genes and suppressed expression of type IIb fiber markers (Fig. 3B). Of note, nitrate in the absence of additional HIF2α stabilization also increased expression of type I and type IIa fiber-type markers, albeit to a lesser extent than nitrate in combination with 2,2-bipyridyl (Fig. 3B).

NO can be generated directly from nitrate through nonenzymatic processes and the enzymatic nitrate-nitrite-NO pathway in vivo (2,12). NO increases PGC1α expression in skeletal muscle and inhibits HIF complex prolyl hydroxylases, stabilizing the HIF complex (6,28). Therefore, C2C12 mouse myotubes were treated with a range of nitrate doses during differentiation (25 and 500 μmol/L) to determine whether nitrate-induced fiber-type switching would occur in a dose-dependent manner in the absence of additional HIF stabilization. The 500 μmol/L nitrate dose corresponded to plasma concentrations in mice exhibiting improved metabolic phenotypes when chronically dosed with 0.1 mmol/kg NaNO₃ (4). Nitrate treatment alone significantly increased PGC1α and HIF2α expression (Fig. 3C). Type I fiber–associated genes also increased in expression in nitrate-treated myotubes. The expression of type IIb fiber–associated genes was decreased by nitrate. These data suggest that nitrate-mediated induction of fiber-type gene expression occurs in skeletal muscle and does not require additional HIF stabilization.

We next examined whether the nitrate-stimulated expression of type I fiber–associated genes was NO dependent. Myotubes were differentiated in the presence of nitrate and the NO scavenger PTIO. NO sequestration by PTIO abrogated nitrate-induced expression of type I and type IIa fiber–associated genes (Fig. 3C). Nitrite is the initial reduction product in the nitrate-nitrite-NO pathway and was found to increase expression of type I fiber–associated genes in myotubes (Fig. 3D). The effect of nitrite was also found to be abrogated by NO sequestration (Supplementary Fig. 1).

To eliminate the canonical l-arginine–nitric oxide synthase (NOS)–NO pathway as the source of NO stimulating the fiber-type switching effect of nitrate, myotubes were cotreated with nitrate and the NOS inhibitor l-NAME. NOS inhibition did not prevent nitrate-stimulated expression of type I fiber–associated genes (Supplementary Fig. 2). Therefore, the nitrate/NO-mediated fiber-type switching effect is independent of NOS.

Nitrate may augment the fiber-type switching effect observed in muscle after exercise training (29); therefore, myotubes were treated with an exercise mimetic, PPARδ agonist GW742 (30), both alone and in combination with nitrate. Cotreatment of myotubes with nitrate and GW742 had an additive effect on type I fiber–associated gene expression (Fig. 3E). Thus, nitrate may augment muscle fiber-type switching in an exercise training background.

The increases in PGC1α expression and protein concentration in skeletal muscle and myotubes likely mediate the nitrate-induced fiber-type switch. Therefore, the expression of PGC1α in C2C12 myotubes was knocked down by ∼60% by using siRNA (Fig. 4A). Reduced PGC1α expression abrogated the nitrate-induced expression of type I and type IIa fiber genes in myotubes (Fig. 4B).

We next sought to identify the downstream signaling cascade that mediates nitrate/NO-induced muscle fiber-type switching. The principal downstream effector of NO is cyclic guanosine monophosphate (cGMP). NO activates cGMP signaling through stimulation of soluble guanylyl cyclase (sGC). An inhibitor of sGC, ODQ, abrogated the nitrate-induced expression of type I fiber–associated genes in myotubes (Fig. 4C).

The secondary messenger cGMP is degraded by cyclic nucleotide phosphodiesterase 5 (PDE5). The expression of PDE5 is increased in nitrate-treated myotubes and may represent a counterregulatory response (Fig. 4D). The PDE5 inhibitor sildenafil increases PGC1α expression in myocytes, reduces muscle fatigue, and improves exercise efficiency (31,32). Therefore, we hypothesized that PDE5 inhibition enhances nitrate-mediated fiber-type switching. The expression of type I muscle fiber–associated genes in myotubes was increased by sildenafil, an effect further enhanced by cotreatment with nitrate (Fig. 4E).

Exercise-conditioned skeletal muscle is the source of hormone-like signals, known as myokines, which participate in organ crosstalk (17,18). The myokine FNDC5 (fibronectin type III domain-containing 5)/irisin is expressed and secreted in a PGC1α-dependent manner (18). Because nitrate increases PGC1α expression in skeletal muscle, we asked whether the nitrate-induced exercise training–like effect in myotubes increased irisin production and secretion. Nitrate increased FNDC5 expression in myotubes in an NO-dependent manner (Fig. 5A). Of note, FNDC5 expression in myotubes was increased by PDE5 inhibition using sildenafil and further increased by cotreatment with both nitrate and sildenafil (Fig. 5B). The secretion of irisin from myotubes into serum-free culture media was increased after nitrate treatment, an effect also observed after treatment of myotubes with the exercise mimetic GW742 (30) (Fig. 5C).

Having identified nitrate-stimulated expression and secretion of irisin from myotubes in vitro, we investigated whether this effect was translated in vivo. FNDC5 expression in the soleus (Fig. 5D) and gastrocnemius (Fig. 5E) of nitrate-treated rats was increased compared with control animals. Increased irisin expression in the soleus and gastrocnemius muscles was reflected by a significant increase in the plasma irisin concentration in nitrate-treated rats (Fig. 5F).

We then asked whether the effect of nitrate on plasma irisin concentrations in rodents translated to humans. A crossover, randomized, placebo-controlled, double-blind clinical trial of participants given either concentrated beetroot juice (2 × 70 mL/day, 12.0 mmol nitrate) or nitrate-depleted beetroot juice (placebo, 2 × 70 mL/day, 0.003 mmol nitrate) for 7 days was implemented (Supplementary Table 2). After a 7-day washout period, participants were crossed over to the reciprocal intervention. Nitrate treatment significantly increased both the fasted serum NOx (Fig. 5G) and the serum irisin (Fig. 5H) concentration.

Irisin induces a brown adipose–like phenotype in WAT depots in a PPARα-dependent manner (18). Nitrate activates WAT browning through a PPARα-independent nitrate-NO-cGMP signaling axis (12). Given their distinct signaling mechanisms, we hypothesized that irisin and nitrate have an additive effect on brown adipocyte–associated gene expression in primary adipocytes. Murine primary adipocytes treated with a combination of irisin and nitrate exhibited a greater expression of brown adipocyte–associated genes than adipocytes treated with irisin or nitrate alone (Supplementary Fig. 3). Therefore, nitrate may enhance the exercise-induced irisin-mediated browning of WAT.

Exercise stimulates the release of β-aminoisobutyric acid (BAIBA) from skeletal muscle (17). BAIBA is an exercise training– and PGC1α-dependent myokine-like small molecule signal with effects on both WAT and hepatic metabolism. Therefore, we used LC-MS to profile metabolites in nitrate-treated myotubes and in serum-free conditioned cell media. Nitrate significantly increased the intracellular (Fig. 6A) and extracellular (Fig. 6B) BAIBA concentrations. Nitrate also increased the expression of BAIBA biosynthetic enzymes acyl-CoA dehydrogenase short chain (ACADS) and hydroxyacyl-CoA dehydrogenase (HADHA) in myotubes in an NO-dependent manner (Fig. 6C).

We next used LC-MS to profile the metabolites in soleus muscle and plasma from nitrate-treated rats. Nitrate significantly increased the BAIBA concentration (Fig. 6D) and the expression of BAIBA biosynthetic genes ACADS and HADHA in the soleus muscle (Fig. 6E). Plasma BAIBA concentration was also increased by nitrate treatment (Fig. 6F).

LC-MS profiling identified an increase in the plasma GABA concentration of nitrate-treated rats (Fig. 7A), with a concomitant decrease in the GABA metabolic precursor glutamine (Fig. 7B). This observation was reproduced in humans, with 7-day nitrate supplementation increasing the serum GABA (Fig. 7C) and decreasing the serum glutamine (Fig. 7D) concentration. GABA is also enriched in the media of transgenic PGC1α-expressing myocytes (17) and in the muscle of mice with muscle-specific transgenic expression of PGC1α (MCK-PGC1α) (33). We used LC-MS to analyze the plasma GABA concentrations of MCK-PGC1α mice. The plasma GABA concentration was significantly increased by PGC1α forced expression in muscle in vivo (Fig. 7E).

Because nitrate increased PGC1α expression in myotubes, we applied LC-MS profiling to serum-free media conditioned for 24 h on myotubes treated with nitrate or the exercise mimetic GW742 (30). GABA was significantly enriched in the media of both nitrate- and GW742-treated myotubes (Fig. 7F). The intracellular GABA concentration was also increased in nitrate-treated myotubes (Fig. 7G).

GABA was increased in myocyte media by nitrate treatment and by forced PGC1α expression and in the plasma of nitrate-treated rats and humans; therefore, we examined whether concentrations of GABA were increased after exercise training in the plasma (Fig. 7H), gastrocnemius (Fig. 7I), and quadriceps (Fig. 7J) of mice conditioned with 3 weeks of free wheel running (17,18). Exercise training increased the GABA concentration in both muscle and plasma. These data suggest that PGC1α stimulates the biosynthesis and release of GABA from skeletal muscle during exercise training and nitrate treatment.

Long-term exercise training, NO, and plasma GABA increase the plasma concentration of growth hormone (34–36). Nitrate increased circulating growth hormone concentrations in nitrate-treated rats (Fig. 7K) and nitrate-supplemented humans (Fig. 7L). Together, these studies demonstrate that nitrate treatment mimics several physiological effects of exercise training both on skeletal muscle and systemically and suggest that GABA is a novel, PGC1α-mediated, exercise- and nitrate-stimulated myokine-like small molecule signal.

Exercise training demonstrates significant antidiabetic effects (19). The ubiquitous dietary anion nitrate improves submaximal and high-intensity exercise efficiency (9), tolerance (10), and performance (11). As a result, dietary nitrate supplementation is growing in popularity as an augmentation to exercise training. We show that nitrate increases PGC1α expression and a switch from fast-twitch type IIb toward oxidative slow-twitch type I and intermediate type IIa fibers in both the soleus and the gastrocnemius muscle, a phenomenon typical of exercise training and consistent with the effects observed on exercise, mitochondrial fatty acid oxidation, and efficiency in muscle predominately consisting of type I and type IIa fibers (9,23).

T2DM is accompanied by perturbations to muscle physiology, resulting in decreased mitochondrial content and abnormal lipid deposition, impairing the insulin-mediated switch between fat and carbohydrate metabolism (37). Dietary nitrate supplementation exhibits antiobesity and antidiabetic properties in rodents, improving HOMA of insulin resistance and quantitative insulin sensitivity check index assessments in high-fructose diet–induced insulin-resistant rats (5) and improving glucose tolerance in NOS-deficient mice (4) while reducing adiposity in both models. We have shown that nitrate increases muscle β-oxidation and mitochondrial biogenesis (23) and, therefore, may have utility in treating the intramuscular lipid deposition–induced lipotoxicity and insulin resistance associated with diabetes. However, diabetic myopathy is a major complication of diabetes and is characterized by muscle atrophy and reduced physical performance and muscle capacity (20). Skeletal muscle from patients with diabetes exhibits a higher concentration of glycolytic fibers (38) and evidence of a switch toward a glycolytic phenotype (39). Indeed, the proportion of muscle type I fibers is positively correlated with insulin sensitivity (21), and decreased levels of type I fibers are associated with insulin resistance (22). Relative to glycolytic fibers, oxidative fibers are more resistant to atrophy after denervation or aging. Because the concentrations of nitrate used in the current study are achievable in humans after dietary intervention (4), nitrate supplementation may provide a means of stimulating exercise-like fiber-type switching effects in the muscle of patients with diabetes to treat perturbed muscle phenotypes. However, a small number of human trials assessing the antidiabetic effects of short-term nitrate supplementation have proven disappointing thus far (40); whether larger, longer-term studies in populations with diabetes or prediabetes (in which disease is less advanced) will demonstrate the beneficial effects observed in rodent models remains to be seen.

The current data suggest that nitrate may be additive to the effects of exercise training on muscle phenotypes, complementing studies by De Smet et al. (29) that showed that sprint interval training in combination with nitrate supplementation increases the proportion of type IIa fibers in human muscle. Indeed, type I and type IIa muscle fibers are more efficient at performing and have a greater capacity for oxidative metabolism than type IIb muscle fibers (41,42). A higher percentage of type I muscle correlates positively with exercise efficiency in exercise tests (43,44), meaning that a higher percentage of type I muscle fibers improves endurance performance by significantly increasing the power output generated for a given rate of oxygen consumption and energy expenditure (43,44). Therefore, the switching of muscle fibers from type IIb to types IIa and I (as we observe with nitrate supplementation) decreases the oxygen cost of endurance exercise, complementing previous observations that dietary nitrate lowers the oxygen cost of exercise. Subsequently, nitrate-induced fiber-type switching likely contributes to the observed effects of nitrate on exercise.

Furthermore, we identify that the nitrate-mediated induction of fiber-type switching occurs, at least in part, through an NO-PGC1α–dependent mechanism. Rasbach et al. (14) identified that PGC1α-induced fiber-type switching requires functional, stabilized HIF2α. Others have demonstrated that NO stabilizes HIF (6,28). We found that nitrate-induced fiber-type switching occurs without extrinsic HIF stabilization, suggesting that NO generated as a result of nitrate reduction may stabilize HIF2α during the nitrate-PGC1α–stimulated expression of type I and type IIa fiber-type markers.

The current in vitro mechanistic studies implicating NO in nitrate-induced expression of fiber-type markers were conducted in noncontracting myotubes; therefore, future studies should examine these mechanisms in models of exercising muscle. Moreover, although a physiological, enzymatic mechanism for the reduction of nitrate to NO in mammalian tissue has been suggested (2,45), numerous proteins have been implicated in this reductive activity (2,12,46,47). In addition, nitrate reduction has been observed in both hypoxia and normoxia (2,12,45). Therefore, the mechanisms underlying the reduction of nitrate to NO in skeletal muscle and the involvement of NO in the downstream effects of nitrate remain to be fully determined.

Exercise-induced PGC1α expression in skeletal muscle contributes to the beneficial cardiometabolic effects of exercise through myokine signaling (17,18). We found that nitrate regulates the exercise- and PGC1α-dependent myokine FNDC5/irisin and myokine-like small molecule BAIBA. BAIBA alters energy metabolism in adipose tissue and liver (17), whereas irisin stimulates the browning of adipose tissue (18). In the current study, nitrate and irisin functioned additively to induce brown adipocyte–specific gene expression in adipocytes; therefore, nitrate supplementation on an exercise-trained background may augment the exercise-dependent activation of WAT browning. Speculation that the nitrate-induced secondary release of myokine factors contributes to the antidiabetic effects of nitrate is also reasonable.

This study unexpectedly highlighted a PGC1α and exercise-induced mechanism mediating muscle production and release of GABA. Nitrate activated this pathway, increasing circulating GABA concentrations in rodents and humans. Peripheral GABA increases plasma growth hormone concentration in humans (35). Independently, NO signaling has been implicated as a pathway mediating the exercise-stimulated increase in circulating growth hormone (48) and directly increases growth hormone release (36). Exercise is one of the most effective stimuli for growth hormone secretion; however, the exact mechanisms remain to be defined. We observed that nitrate treatment increases the circulating growth hormone concentration in rats and humans. Therefore, GABA may function as a PGC1α-dependent, myokine-like small molecule signal, with biosynthesis and secretion triggered by both exercise and nitrate. The current investigations highlight an interaction linking NO and PGC1α to muscle GABA secretion and peripheral GABA concentration in a pathway that may contribute to exercise-stimulated growth hormone release in the adaptive response to exercise.

This investigation highlights that nitrate induces exercise training–like physiological adaptations both in skeletal muscle and systemically. GABA may function as an exercise and nitrate-stimulated, PGC1α-mediated, myokine-like small molecule, and the exercise mimetic effects of nitrate may contribute to the beneficial effects of nitrate on both exercise and cardiometabolic health.

Funding. L.D.R. is supported by the Diabetes UK RD Lawrence Fellowship (16/0005382) and the Medical Research Council-Human Nutrition Research Elsie Widdowson Fellowship. This work was supported by the Biotechnology and Biological Sciences Research Council (Bb/H013539/2, bb/I000933/I), the British Heart Foundation, and the Medical Research Council (UD99999906).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. L.D.R. conceived, designed, and performed the experiments and wrote the manuscript with input from all coauthors. T.A. performed animal and cell studies. B.D.M. performed cell experiments. S.A.M. performed GC-MS. B.O.F. and M.F. performed nitrate measurements. R.L. assisted with animal studies. M.S. and E.A.W. conceived, designed, and performed the human experiments. A.J.M. conceived and designed the animal studies. J.L.G. contributed to the manuscript. L.D.R. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.