Novel pyruvate tracer reveals dichloroacetate’s distinct effects on muscle metabolism

For decades, peering into the energy factories of a living muscle meant settling for indirect readings—measurements that hinted at what mitochondria were doing without ever quite capturing the chemistry itself. Now, a team at UT Southwestern Medical Center has demonstrated an imaging technique that tracks carbon atoms as they stream through the citric acid cycle of skeletal muscle in real time, and its first results are already overturning a core assumption of metabolism research. By pairing a specially engineered isotope of pyruvate with a subtle chemical trick involving heavy water, the researchers captured a hidden buffering system in action: a molecular holding tank that soaks up surplus fuel when the mitochondrial furnace is not burning hot enough to consume it. The observation may help explain why metabolism becomes rigid in obesity and type 2 diabetes, and why therapies meant to reignite mitochondrial energy burning often disappoint.

The study, published in the journal iScience, centers on the tricarboxylic acid (TCA) cycle, the revolving sequence of reactions that sits at the heart of mitochondrial energy production. Skeletal muscle is an energy-hungry tissue, and during sustained activity it depends overwhelmingly on oxidative phosphorylation and the TCA cycle to convert carbohydrates and fatty acids into adenosine triphosphate, the universal currency of cellular work. When these pathways falter, as they do in insulin resistance and type 2 diabetes, muscle loses its metabolic flexibility and struggles to oxidize substrates efficiently. Muscle is also remarkably adaptable, rewiring its energetic pathways within moments when exercise load surges, which makes its metabolic control points attractive therapeutic targets. Yet measuring the cycle’s throughput inside an intact organism has long been a stubborn technical problem. Established clinical tools such as phosphorus magnetic resonance spectroscopy and oxygen respirometry deliver bulk information about phosphate metabolites or whole-body energetics, but they cannot resolve the substrate-specific fluxes flowing through individual mitochondrial enzymes.

The new method belongs to a family of techniques known as hyperpolarized carbon-13 magnetic resonance spectroscopy. In hyperpolarization, a sample is cooled to near absolute zero inside a strong magnetic field, where a radical-mediated process called dynamic nuclear polarization aligns the spins of carbon-13 nuclei to a degree millions of times beyond their ordinary thermal equilibrium. When the polarized agent is rapidly dissolved and injected into the bloodstream, an MRI scanner can watch it transform into downstream metabolites, second by second. The field’s workhorse probe, [1-13C]pyruvate, has already reached human clinics and can gauge pyruvate dehydrogenase, or PDH—the gatekeeping enzyme that ushers pyruvate into the mitochondrion—by watching the labeled carbon wash into bicarbonate. But that readout carries an inherent blind spot: once the label is released as bicarbonate, it vanishes from view, leaving acetyl-coenzyme A and everything downstream of PDH invisible. Researchers have long equated PDH flux with TCA cycle flux, but the assumption had never been validated, because the standard tracer simply cannot see past the gate.

In principle, [2-13C]pyruvate solves the problem. Its labeled carbon survives PDH catalysis, rides into acetyl-CoA and propagates through the TCA cycle, ultimately tagging glutamate, a reliable signature of cycle activity. The obstacle is physics. The carbon-2 nucleus of pyruvate has a spin-lattice relaxation time, or T1, of roughly 39 seconds at 3 tesla, compared with about 67 seconds for [1-13C]pyruvate. Hyperpolarized signal decays exponentially, so a shorter T1 means the enhanced nuclear alignment drains away before the tracer ever reaches the mitochondrion, drowning the glutamate signal in noise—especially in resting muscle, where PDH flux is low. The team, led by Jae Mo Park, overcame this with two maneuvers. They replaced the three hydrogen atoms on pyruvate’s methyl group with deuterium, suppressing the dipolar and scalar interactions with neighboring protons that accelerate relaxation, and they dissolved the polarized substrate in deuterium oxide—heavy water—rather than ordinary water. Together these steps extend the T1 of the carbonyl carbon enough that [5-13C]glutamate becomes reliably detectable in living skeletal muscle even at rest, something the undeuterated tracer could only achieve in drug-stimulated tissue.

To put the probe through its paces, the researchers injected healthy Sprague-Dawley rats with hyperpolarized [2-13C,3-2H3]pyruvate twice: once at baseline and again 45 minutes after administering dichloroacetate, a drug that jolts PDH into action by blocking pyruvate dehydrogenase kinase, the enzyme that normally brakes PDH. Each tracer dose was polarized for three to four hours at 0.8 kelvin inside a dynamic nuclear polarization device, dissolved in superheated heavy water and delivered through a tail-vein catheter. Using a clinical 3 tesla MRI scanner fitted with a custom carbon-13 surface coil placed over the quadriceps, the team recorded spectra every three seconds across a 90-second window, normalizing every metabolite to the total hyperpolarized carbon-13 signal. The results were decisive. Glutamate labeled at its fifth carbon rose from 0.0116 to 0.0196 after dichloroacetate—an increase of roughly 78 percent—confirming that TCA cycle activity had accelerated. Lactate production, by contrast, barely moved, indicating the drug was acting on mitochondrial oxidation rather than on the cytosolic handling of pyruvate.

The surprise came from a second metabolite. Acetyl-L-carnitine, the ester through which the enzyme carnitine acetyltransferase parks excess acetyl groups in a cellular reservoir, surged far more dramatically: from 0.0330 to 0.1277, a leap of more than 400 percent. In other words, throwing open the mitochondrial gate did not proportionally speed up the Krebs wheel. The ratio of glutamate to the combined glutamate-plus-acetylcarnitine signal—a proxy for the balance between TCA cycle flux and PDH flux—fell by roughly 45 percent after treatment. The mitochondria were receiving acetyl-CoA far faster than they were oxidizing it, and they responded by stashing the surplus in the acetylcarnitine pool. Time-to-peak analysis confirmed the shift was not an artifact of altered perfusion: the labeled acetylcarnitine appeared no later than before, only in far greater quantity, while lactate kinetics stayed flat. Surplus acetyl units, it seems, are not wasted but banked—reversible storage that smooths the fuel supply for future bursts of demand.

This observation speaks directly to what biochemists call the acetyl-L-carnitine overflow pool hypothesis. Acetyl-CoA stands at a metabolic crossroads: when oxidative demand is high, it is fed to citrate synthase and into the TCA cycle; when demand lags, carnitine acetyltransferase converts it into acetylcarnitine for storage, ready to be remobilized when energy needs spike. Earlier work in perfused hearts established acetylcarnitine as such a reservoir, and hyperpolarized studies there showed acetylcarnitine production rising when PDH was chemically activated but staying near baseline when cardiac workload was raised with dobutamine—hinting that the buffer tracks fuel surplus rather than sheer demand. But watching that dynamic unfold in intact, resting skeletal muscle in vivo is new. The team’s data suggest that pharmacological PDH activation preferentially routes acetyl groups into the buffer rather than into accelerated oxidation, which means that measuring PDH flux alone, as the conventional bicarbonate-based approach does, could substantially overestimate the muscle’s true oxidative throughput. Human studies lend the idea physiological weight: obese individuals show reduced muscle acetylcarnitine and slower phosphocreatine recovery after exercise, while reduced carnitine acetyltransferase activity has been linked to metabolic inflexibility and insulin resistance.

Because hyperpolarized signals are fleeting and detection is demanding, the researchers validated their imaging findings with an independent method. A separate cohort of rats received a bolus of uniformly carbon-13-labeled pyruvate, and 90 seconds later their hamstring muscles were flash-frozen for gas chromatography–mass spectrometry. This ex vivo isotopomer analysis told the same story: doubly labeled glutamate, an unambiguous fingerprint of pyruvate-derived acetyl-CoA entering the cycle, roughly quadrupled in abundance in drug-treated muscle, and its concentration climbed from about 14 to 91 nanomoles per gram of tissue. The team also traced how the two-carbon label propagated across successive TCA intermediates—citrate, alpha-ketoglutarate, succinate, fumarate and malate. Fitting an exponential decay model to the pattern yielded propagation constants of 1.01 in controls versus 1.51 in treated animals, quantitative evidence of accelerated cycle flux. The labeling patterns simultaneously showed that pyruvate carboxylase, an auxiliary enzyme that could otherwise confound the analysis, contributed relatively little under these conditions.

The two techniques also diverged in instructive ways, and reconciling them clarifies what each actually measures. Hyperpolarized spectroscopy captures real-time kinetics over the first minutes after injection, whereas mass spectrometry provides a single endpoint snapshot; the acetylcarnitine pool equilibrates so quickly that a 90-second measurement understates its turnover, while glutamate labeling remains far from isotopic steady state at that moment. The study has caveats, too: the experiments involved only adult male rats, the TCA cycle activity being measured reflects pyruvate-derived carbon entry rather than total cycle turnover—fatty acids, ketone bodies and amino acids also feed the wheel—and even with the prolonged relaxation time, tracking the glutamate signal dynamically remained at the edge of detectability. The authors also note that alanine concentrations fell after dichloroacetate treatment, consistent with earlier findings that the drug suppresses cytosolic transamination of pyruvate, which reinforces confidence in the broader metabolic picture.

The implications reach well beyond basic physiology. Dichloroacetate has been investigated for decades as a treatment for mitochondrial disorders, lactic acidosis and certain cancers, because activating PDH pushes glucose toward oxidation and suppresses lactate accumulation. But the metabolic consequences of pushing more carbon through PDH likely depend on where the respiratory chain or the cycle itself is impaired, and clinicians currently lack tools that distinguish successful reactivation from futile overflow into storage. A probe that can simultaneously interrogate PDH flux, acetyl-CoA buffering and TCA turnover in the same living tissue could pinpoint bottlenecks distal to PDH and reveal whether an intervention genuinely restores oxidative capacity. The method also integrates readily with proton and phosphorus spectroscopy, opening the door to richer, multi-nuclear portraits of muscle energetics. With hyperpolarized carbon-13 pyruvate already carrying an established safety profile across multi-center human trials, the researchers argue that their deuterated carbon-2 version holds translational potential for studying exercise physiology, obesity and diabetes—conditions in which impaired metabolic flexibility and mishandled acetyl-CoA are defining features. For the first time, scientists can watch a mitochondrion deciding whether to burn its fuel—or bank it.