Two Models of Insulin Resistance
The conventional understanding of insulin resistance, associated with the work of Gerald Shulman and colleagues, frames it as a consequence of ectopic lipid accumulation: excess diacylglycerol and ceramides in skeletal muscle and liver activate serine kinases (IKKβ, JNK) that phosphorylate insulin receptor substrate-1 (IRS-1) at serine residues, impairing the downstream PI3K/Akt pathway and reducing GLUT4 translocation. In this model, insulin resistance is a pathological failure of insulin signaling driven by lipotoxicity and chronic hyperinsulinemia.
A more mechanistically precise framework, developed over the past two decades by veterinarian and physiologist Peter Dobromylskyj (writing as "Peter" on the Hyperlipid blog) and supported by the independent work of biochemist Dave Speijer at the University of Amsterdam, reframes insulin resistance as a normal, physiological satiety signal generated by the mitochondria. In this model, the critical variable is not the total quantity of substrate entering the cell but the ratio of FADH2 to NADH produced during its oxidation. This is the Protons hypothesis.
The FADH2:NADH Ratio and Reverse Electron Transport
During beta-oxidation of fatty acids, each cycle of the spiral produces one FADH2 via the electron-transferring flavoprotein (ETF) dehydrogenase, which feeds electrons directly into the CoQ pool, and one NADH, which feeds electrons into Complex I. The ratio of FADH2 to NADH (the F/N ratio) produced by a given substrate determines the redox state of the CoQ couple and, critically, whether electrons flow forward or backward through Complex I.
Mechanism Diagram
The Insulin/ROS Hypothesis
Step through the mechanism from dietary substrate to metabolic disease.
When the F/N ratio is high (as it is during oxidation of saturated fatty acids like palmitate at 0.484 or stearate at 0.486), the CoQ pool becomes highly reduced. A highly reduced CoQ pool, combined with a high mitochondrial membrane potential (ΔΨm), drives electrons backward through Complex I in a process called reverse electron transport (RET). RET at Complex I is the primary source of mitochondrial superoxide (O₂•⁻), which is rapidly converted to hydrogen peroxide (H₂O₂) by superoxide dismutase. This H₂O₂ is the physiological signal that the cell is energy-replete. It inhibits insulin signaling, reducing further nutrient ingress. This is cellular satiety.
"Insulin resistance is not a malfunction. It is the cell's physiological satiety signal, generated by reverse electron transport at Complex I when the FADH2:NADH ratio is high. The pathology is not too much insulin resistance but too little - a failure of cellular satiety driven by the displacement of saturated fat by linoleic acid.
The F/N ratio can be calculated precisely for any fatty acid. For saturated fatty acids of carbon chain length n, F/N = (n−1)/(2n−1). For unsaturated fatty acids, each double bond reduces the FADH2 yield by one, giving F/N = (n−1−db)/(2n−1), where db is the number of double bonds. Glucose, oxidized entirely through NADH-generating pathways, has an F/N ratio of approximately 0.2. The practical consequence is that the type of fat consumed, not merely the quantity of carbohydrate, is a major determinant of the cell's ability to generate its own satiety signal.
The Physiological States of the Electron Transport Chain
The Protons hypothesis identifies three physiologically distinct states of the ETC. In the fasting or ketogenic state, mitochondrial membrane potential is low, Complex I throughput is low, and FADH2 input from beta-oxidation of saturated fats is high relative to NADH. Despite the high F/N ratio, the low ΔΨm makes RET thermodynamically unfavorable, so superoxide production is minimal. Insulin signaling is rapidly aborted by the continuing action of tyrosine phosphatases. This is the physiological insulin resistance of starvation, a protective mechanism that preserves glucose for the brain.
In the normally fed state with active insulin signaling, ΔΨm is high, NADH input from glucose oxidation dominates, and insulin-mediated suppression of lipolysis minimizes FADH2 input. High Complex I throughput with minimal RET produces little superoxide. Insulin signaling proceeds normally.
Signaling Pathway
Insulin Signaling Cascade
From receptor binding to GLUT4 translocation. Select any node to read its function.
Select any node in the pathway diagram to read its molecular function and clinical relevance.
Key clinical point: ROS-mediated serine phosphorylation of IRS-1 is the primary molecular mechanism of insulin resistance. This is why mitochondrial substrate overflow, not dietary fat, is the upstream driver of T2DM.
Under hypercaloric conditions with both glucose and fatty acids elevated, the combination of high ΔΨm and high FADH2 input from beta-oxidation produces significant RET at Complex I, significant superoxide, and appropriate cellular insulin resistance. This is the normal physiological response to caloric overload: the cell generates its own satiety signal and reduces further nutrient ingress. The development of individual cell insulin resistance under these conditions is, in Dobromylskyj's words, "utterly normal."
Linoleic Acid: The Failure of Cellular Satiety
Linoleic acid (LA, C18:2ω-6) is an 18-carbon fatty acid with two double bonds. Its F/N ratio is 0.429, calculated as (18−1−2)/(2×18−1) = 15/35. This places it below the threshold for generating significant RET-driven superoxide, which Dobromylskyj estimates falls between approximately 0.457 (oleate, MUFA) and 0.486 (stearate, saturated). In practical terms, linoleic acid oxidizes like a glucose-like substrate: it generates insufficient FADH2-driven ROS to signal cellular satiety.
The consequence is that a cell oxidizing linoleic acid cannot generate the mitochondrial H₂O₂ signal that would normally terminate insulin signaling when the cell is energy-replete. Insulin continues to act. Nutrients continue to enter the cell. Excess calories are diverted to intracellular triglycerides. In adipocytes, this manifests as progressive lipid droplet expansion. As adipocytes distend, adipocyte triglyceride lipase (ATGL) activity increases in proportion to lipid droplet size, driving basal lipolysis outside the control of insulin. The resulting elevation in circulating free fatty acids eventually produces systemic insulin resistance through the conventional lipotoxicity pathways (ceramide, diacylglycerol, Randle cycle competition). Hyperinsulinemia follows.
This is the Protons hypothesis account of metabolic syndrome: not a simple story of carbohydrate excess, but a story in which the displacement of saturated and monounsaturated fats by linoleic acid from industrial seed oils undermines the cell's ability to generate its own satiety signal, producing pathological insulin sensitivity that eventually transitions, through adipocyte overflow and FFA release, into pathological insulin resistance.
The Conventional Model and the Protons Hypothesis: What Each Explains
The conventional Shulman/Unger model and the Protons hypothesis are not mutually exclusive. The conventional model accurately describes the downstream lipotoxicity pathways (ceramide, DAG, IKKβ/JNK serine phosphorylation of IRS-1) that produce insulin resistance once adipocyte overflow occurs. The Protons hypothesis provides the upstream explanation for why adipocyte overflow occurs in the first place: the failure of linoleic acid to generate the RET-driven ROS satiety signal allows inappropriate nutrient accumulation in adipocytes long before the lipotoxicity threshold is reached.
For the clinician, the practical implication extends beyond carbohydrate restriction. The composition of dietary fat matters mechanistically. Saturated and monounsaturated fats generate appropriate cellular satiety signals. Linoleic acid, the dominant fatty acid in industrial seed oils (corn, soybean, sunflower, safflower, cottonseed), does not. The dramatic increase in linoleic acid consumption in the United States since the 1960s, from approximately 3 percent of calories to 8–10 percent, parallels the rise of metabolic disease across the same period and is a mechanistically coherent candidate for a contributing cause.
Joseph Kraft's work on insulin patterns complements this framework. By measuring insulin responses to oral glucose tolerance tests in over 14,000 patients, Kraft demonstrated that the majority of patients with normal glucose tolerance already have significantly abnormal insulin patterns, with hyperinsulinemic responses that precede glucose abnormalities by years to decades. The Protons hypothesis provides a mechanistic explanation for why hyperinsulinemia precedes glucose dysregulation: the failure of cellular satiety signaling drives adipocyte expansion and FFA release, which drives compensatory hyperinsulinemia, long before beta cell exhaustion produces the glucose abnormalities that standard screening detects.