Mitochondrial fatty acid flux from adipocyte lipolysis to ATP, and how CPT1-targeting drugs redirect it
Every round of mitochondrial β-oxidation begins outside the mitochondrion — often outside the cell entirely. Adipose triglyceride stores are mobilized hormonally, released into plasma as non-esterified fatty acids (NEFA), and must be activated to a thioester before any oxidative machinery can touch them. This activation step is where the cell commits a fatty acid to either oxidation or re-esterification.
Adipocyte triglyceride is hydrolyzed by a coordinated three-enzyme cascade: adipose triglyceride lipase (ATGL) removes the first fatty acid to yield diacylglycerol, hormone-sensitive lipase (HSL) removes the second, and monoacylglycerol lipase (MGL) completes hydrolysis to free glycerol. HSL is the rate-controlling node — it is activated by PKA-mediated phosphorylation downstream of catecholamine β-adrenergic signaling (epinephrine, norepinephrine) and glucagon, and suppressed by insulin via PDE3B-mediated cAMP degradation.
During an overnight fast, plasma NEFA rise from a fed baseline of ~150–300 µM to 400–800 µM. Prolonged fasting (>24 h) or vigorous exercise (>65% VO2max) can drive NEFA above 1000–1500 µM as catecholamine tone rises and insulin falls. Each NEFA circulates non-covalently bound to serum albumin (each albumin molecule carries up to 6–7 fatty acids at high affinity sites), which keeps the highly amphipathic lipid soluble and buffers its delivery to tissues.
Long-chain fatty acids cross the plasma membrane through a combination of passive flip-flop diffusion and protein-facilitated transport, with the protein-mediated route dominating at physiological NEFA concentrations:
• CD36 (FAT/CD36): a scavenger-receptor family transporter that is the principal facilitator of long-chain FA uptake in heart, skeletal muscle and adipose tissue; insulin and muscle contraction both translocate CD36 from intracellular vesicles to the sarcolemma, acutely raising uptake capacity • FATP1–6 (fatty acid transport proteins): membrane-associated proteins with intrinsic acyl-CoA synthetase activity, effectively coupling import with activation • FABPpm (plasma-membrane fatty acid binding protein): a peripheral membrane protein that increases local FA concentration at the bilayer surface
Once inside the cytosol, cytosolic FABPs (FABP1 in liver, FABP3 in heart/muscle) chaperone the hydrophobic fatty acid to the endoplasmic reticulum and outer mitochondrial membrane, where acyl-CoA synthetases reside.
Acyl-CoA synthetase long-chain (ACSL1, dominant in liver, heart, adipose, and skeletal muscle) catalyzes a two-step reaction: the fatty acid first attacks ATP to form a high-energy fatty acyl-adenylate intermediate (releasing pyrophosphate, immediately hydrolyzed by pyrophosphatase to make the reaction thermodynamically irreversible), then CoA-SH displaces AMP to yield fatty acyl-CoA. Net cost: one ATP is converted to AMP + 2 Pi — equivalent to 2 ATP-hydrolysis events.
This single reaction sits at a critical metabolic fork: the resulting fatty acyl-CoA can be (a) esterified into triglyceride or phospholipid for storage/membrane synthesis, or (b) shuttled to the mitochondrion for oxidation. ACSL1 knockout studies in mouse heart and adipose show that loss of this enzyme reduces fatty acid oxidation rates by >50%, confirming that activation — not just transport — gates the entire downstream pathway. Km for palmitate is typically 5–20 µM, well within the physiological free (unbound) fatty acid range, so flux through this step scales roughly linearly with local FA delivery.
Long-chain fatty acyl-CoA cannot cross the inner mitochondrial membrane directly. Evolution solved this with the carnitine shuttle — a three-enzyme relay that converts acyl-CoA to acylcarnitine for transport, then reconstitutes acyl-CoA in the matrix. Carnitine palmitoyltransferase 1 (CPT1), the first and controlling enzyme, is the single most important pharmacological target for modulating whole-body fat oxidation.
The shuttle operates in three discrete reactions:
• CPT1 (outer mitochondrial membrane): transfers the acyl group from CoA to carnitine, forming acylcarnitine + free CoA-SH. Three tissue-specific isoforms exist — CPT1a (liver, kidney, most tissues), CPT1b (heart and skeletal muscle), and CPT1c (brain, regulatory rather than catalytic in the classical sense) • CACT (carnitine-acylcarnitine translocase, SLC25A20): an antiporter embedded in the inner mitochondrial membrane that exchanges cytosolic acylcarnitine for matrix free carnitine on a 1:1 basis • CPT2 (inner mitochondrial membrane, matrix-facing): reverses the CPT1 reaction, transferring the acyl group back onto matrix CoA-SH to regenerate acyl-CoA and release free carnitine for another shuttle cycle
CPT1 is overwhelmingly the flux-controlling enzyme of the triad: its Vmax under physiological conditions is far below that of CACT or CPT2, and it is the only member of the shuttle subject to potent allosteric regulation.
Malonyl-CoA, the first committed intermediate of de novo lipogenesis (produced by acetyl-CoA carboxylase, ACC), is also CPT1's endogenous allosteric inhibitor — an elegant piece of metabolic logic that prevents newly synthesized fatty acids from being immediately re-oxidized (a futile cycle).
CPT1b (muscle/heart) is roughly 100-fold more sensitive to malonyl-CoA than CPT1a (liver), with IC50 in the tens of nanomolar range. AMP-activated protein kinase (AMPK) phosphorylates and inhibits ACC in response to energy stress (exercise, hypoxia, metformin, biguanides), lowering malonyl-CoA and disinhibiting CPT1 — this is the principal molecular switch by which exercise and AMPK activators increase fat oxidation. Conversely, in the fed state, high insulin/glucose promotes ACC activity, malonyl-CoA rises, and CPT1 is suppressed — directing incoming fatty acids toward esterification and storage rather than oxidation.
Malonyl-CoA sensitivity differs ~100-fold between CPT1a and CPT1b, meaning the same circulating malonyl-CoA concentration can leave hepatic fat oxidation nearly untouched while almost completely suppressing cardiac and skeletal muscle fat oxidation — a key reason tissue-specific drug and genetic effects on FAO diverge so sharply.
Because CPT1 gates essentially all long-chain fatty acid entry into the mitochondrial oxidative machinery, it has been a long-standing pharmacological target:
• Etomoxir: an irreversible CPT1 inhibitor (covalently modifies the enzyme after CoA-ester formation) developed for type 2 diabetes and heart failure; halted in Phase III trials (ERGO/ERGO2, 2000s) due to hepatotoxicity, but remains the standard research tool for FAO inhibition in vitro and in vivo • Perhexiline: a reversible, weaker CPT1/CPT2 inhibitor used clinically (outside the US) for refractory angina; shifts cardiac substrate use toward glucose, improving efficiency of oxygen use per ATP produced • Oxfenicine: a CPT1a-preferring inhibitor used mainly as a research tool • Ranolazine: does not block CPT1 directly, but inhibits late INa current and downstream indirectly reduces fatty acid oxidation dominance in ischemic myocardium
These agents are of major interest because glucose oxidation yields ~12% more ATP per mole of O2 consumed than fatty acid oxidation — a meaningful efficiency gain in oxygen-limited states like heart failure and ischemia (the rationale for "metabolic modulator" cardiac therapy).
Once acyl-CoA is regenerated in the mitochondrial matrix, it enters a four-step enzymatic cycle that repeats, spiral-fashion, shortening the acyl chain by two carbons on every turn. For a typical C16 fatty acid (palmitate), this spiral turns seven times before terminating in a final round of thiolytic cleavage.
Each turn of the spiral consists of:
1. Acyl-CoA dehydrogenase (chain-length specific: VLCAD for C14–C20, MCAD for C6–C12, SCAD for C4–C6) removes two hydrogens across the Cα–Cβ bond, forming a trans-Δ² enoyl-CoA and reducing enzyme-bound FAD to FADH2, which passes electrons to electron-transfer flavoprotein (ETF) and then ETF-ubiquinone oxidoreductase (ETF-QO) into the CoQ pool 2. Enoyl-CoA hydratase adds water across the double bond, producing L-3-hydroxyacyl-CoA 3. 3-hydroxyacyl-CoA dehydrogenase oxidizes the 3-hydroxyl to a 3-keto group, reducing NAD+ to NADH 4. β-ketothiolase (3-ketoacyl-CoA thiolase) cleaves the bond between C2 and C3 using a second CoA-SH, releasing acetyl-CoA and a new acyl-CoA shortened by two carbons
The shortened acyl-CoA re-enters the cycle at step 1. For very-long and long-chain fatty acids, the entire spiral runs on the inner mitochondrial membrane surface via membrane-bound VLCAD and the trifunctional protein (MTP, which carries hydratase + dehydrogenase + thiolase activities in one complex); medium- and short-chain intermediates are processed by soluble matrix enzymes.
Defects in specific acyl-CoA dehydrogenase isoforms cause distinct, chain-length-specific clinical syndromes, all part of newborn screening panels via tandem mass spectrometry acylcarnitine profiling:
• MCAD deficiency (ACADM gene): the most common FAO disorder (~1:10,000–20,000 births in populations of Northern European descent); presents with hypoketotic hypoglycemia during fasting or intercurrent illness, risk of sudden death; diagnosed by elevated C8 (octanoylcarnitine) on newborn screening • VLCAD deficiency (ACADVL gene): more severe, presents with cardiomyopathy, hypoglycemia, rhabdomyolysis; elevated C14:1 (tetradecenoylcarnitine) • LCHAD/MTP deficiency (HADHA/HADHB genes): impairs the trifunctional protein; associated with maternal HELLP syndrome/acute fatty liver of pregnancy when the fetus is affected
Management centers on avoiding prolonged fasting, providing exogenous glucose during illness, and — for long-chain defects — medium-chain triglyceride (MCT) supplementation, since medium-chain fatty acids bypass the CPT1 gate (they diffuse into mitochondria independent of carnitine) and are processed by MCAD rather than the defective long-chain enzymes.
Complete β-oxidation of palmitoyl-CoA (16 carbons) requires 7 turns of the spiral (not 8, because the final turn directly yields two acetyl-CoA without a further cycle):
• 7 × FADH2 (via ETF → ETF-QO → CoQ → Complex III) • 7 × NADH (via Complex I) • 8 × acetyl-CoA (enters the TCA cycle via citrate synthase)
Unsaturated and odd-chain fatty acids require accessory enzymes: Δ³,Δ²-enoyl-CoA isomerase and 2,4-dienoyl-CoA reductase handle double bonds encountered mid-spiral (as in oleate or linoleate oxidation), while propionyl-CoA — the 3-carbon product of odd-chain fatty acid spirals — is carboxylated by propionyl-CoA carboxylase (biotin-dependent) and isomerized to succinyl-CoA, feeding directly into the TCA cycle rather than via acetyl-CoA.
The acetyl-CoA and reduced electron carriers generated by the spiral now feed into the two systems that actually make ATP: the TCA cycle, which fully oxidizes acetyl-CoA to CO2 while generating more NADH and FADH2, and the electron transport chain, which uses all of that reducing power to build the proton-motive force that drives ATP synthase.
Each acetyl-CoA condenses with oxaloacetate (OAA) via citrate synthase to form citrate, entering the eight-reaction TCA cycle. One full turn of the cycle regenerates OAA while releasing 2 CO2 and generating 3 NADH, 1 FADH2, and 1 GTP/ATP (substrate-level phosphorylation at succinyl-CoA synthetase).
For palmitate, the 8 acetyl-CoA produced by the spiral drive 8 TCA turns, contributing: 24 NADH, 8 FADH2, and 8 GTP from the TCA cycle itself, on top of the 7 NADH and 7 FADH2 generated directly by the spiral. Totaled and passed through oxidative phosphorylation (using ~2.5 ATP per NADH and ~1.5 ATP per FADH2 at the P/O ratios established for mammalian mitochondria), plus the 8 GTP and correcting for the initial 2 ATP-equivalent activation cost, the accepted net yield is ≈106 ATP per palmitate — roughly 3.3–3.5× the ATP yield of a single glucose molecule per mole of substrate, though ~12% less efficient per mole of O2 consumed.
NADH generated by β-oxidation and the TCA cycle donates electrons to Complex I (NADH:ubiquinone oxidoreductase), pumping 4 H+ per NADH. FADH2 generated within the spiral does not feed Complex II directly — it is bound to acyl-CoA dehydrogenase and passes electrons to electron-transfer flavoprotein (ETF), then to ETF-ubiquinone oxidoreductase (ETF-QO), which reduces coenzyme Q — bypassing Complex I and contributing fewer pumped protons per electron pair than NADH-derived electrons, which is the biochemical basis for the ~1.5 (FADH2) vs ~2.5 (NADH) ATP/carrier convention.
Both electron streams converge at the CoQ pool, pass through Complex III (cytochrome bc1) and cytochrome c to Complex IV (cytochrome c oxidase), which reduces O2 to H2O. The proton gradient built across the inner membrane drives ATP synthase (Complex V) via chemiosmotic coupling — the mechanism for which Peter Mitchell won the 1978 Nobel Prize in Chemistry.
When hepatic β-oxidation flux is high but OAA is depleted (OAA is diverted to gluconeogenesis during prolonged fasting, since the liver must maintain blood glucose), acetyl-CoA cannot fully enter the TCA cycle. The liver — uniquely among tissues, because it lacks significant ketolytic capacity itself — diverts excess acetyl-CoA to ketogenesis:
2 Acetyl-CoA → acetoacetyl-CoA (thiolase) → HMG-CoA (HMG-CoA synthase, the committed step) → acetoacetate (HMG-CoA lyase) → β-hydroxybutyrate (β-OHB, via β-hydroxybutyrate dehydrogenase) or acetone (spontaneous decarboxylation)
Circulating ketone bodies rise from a fed baseline of <0.1 mM to 1–2 mM after a 24–48 h fast, and can exceed 6–8 mM in prolonged starvation or diabetic ketoacidosis. Peripheral tissues (heart, skeletal muscle, and critically the brain after adaptation) take up β-OHB and reconvert it to acetyl-CoA for TCA oxidation, sparing glucose — the central adaptive rationale for ketogenesis during fasting.
Because CPT1 sits at the single controlling node of long-chain fatty acid oxidation, it is the primary lever pharmacologists pull to redirect cardiac and systemic substrate metabolism — with applications spanning heart failure, angina, and metabolic disease. Quantifying the resulting flux shift in real patients relies on isotope tracer methods and whole-body gas exchange.
The Randle cycle (glucose-fatty acid cycle, described by Philip Randle in 1963) describes the reciprocal biochemical relationship between fat and glucose oxidation. High fatty acid oxidation raises mitochondrial acetyl-CoA/CoA and NADH/NAD+ ratios, which inhibit pyruvate dehydrogenase — suppressing glucose oxidation even when glucose and insulin are abundant (a mechanism implicated in skeletal muscle insulin resistance). Blocking CPT1 collapses this brake: acetyl-CoA/CoA falls, pyruvate dehydrogenase is disinhibited, and glucose oxidation rises to compensate for cellular energy demand — the mechanistic basis for CPT1 inhibitors and partial FAO inhibitors improving glucose handling and cardiac efficiency.
• Etomoxir — irreversible CPT1 inhibitor; potent FAO blockade validated the Randle-cycle hypothesis in humans, but Phase III heart failure trials (ERGO, ERGO2) were halted after elevated liver transaminases; remains the reference tool compound in bench research • Perhexiline — reversible CPT1/CPT2 inhibitor, approved outside the US (Australia, New Zealand) for refractory angina and heart failure; requires therapeutic drug monitoring due to CYP2D6-dependent hepatotoxicity/neuropathy risk in poor metabolizers • Trimetazidine — partially inhibits long-chain 3-ketoacyl-CoA thiolase, cutting FAO flux by roughly 30–40% without fully blocking CPT1; widely used in Europe/Asia for stable angina; improves cardiac efficiency without hemodynamic side effects • Ranolazine — FDA-approved anti-anginal; primary mechanism is late INa current blockade, with indirect, smaller effects on fatty acid oxidation dominance • Malonyl-CoA decarboxylase (MCD) inhibitors — raise endogenous malonyl-CoA to naturally suppress CPT1; investigated preclinically for heart failure and ischemia
Because glucose oxidation yields ~12% more ATP per mole of O2 than fatty acid oxidation, even a partial (30–40%) pharmacologic shift of cardiac substrate flux from fat to glucose — as achieved by trimetazidine or moderate-dose CPT1 inhibition — can measurably improve mechanical efficiency in the failing or ischemic heart without any change in coronary blood flow.
Because β-oxidation cannot be observed directly in vivo, clinical and research flux is inferred through indirect methods:
• Indirect calorimetry / RER (respiratory exchange ratio, VCO2/VO2): RER ≈0.70 indicates near-pure fat oxidation, RER ≈1.00 indicates near-pure carbohydrate oxidation; whole-body RER shifts of 0.05–0.15 are readily detected after CPT1 inhibitor dosing or exercise-intensity changes • ¹³C-labeled palmitate or octanoate breath tests: labeled substrate is infused or ingested, and the appearance of ¹³CO2 in exhaled breath over 2–6 hours quantifies whole-body FAO flux non-invasively — the current clinical/research gold standard for suspected FAO disorders and drug-effect studies • Plasma acylcarnitine profiling (tandem mass spectrometry): the same assay used in newborn screening reveals which chain-length step is rate-limiting or blocked — a rise in long-chain acylcarnitines (C16, C18) with low free carnitine suggests a CPT1/CACT/CPT2 bottleneck, mirroring the pharmacologic signature of etomoxir or perhexiline • Arteriovenous balance (forearm or cardiac catheterization): direct sampling of NEFA and lactate/glucose across an organ bed quantifies substrate extraction in research settings, most classically applied to the human heart