Inside your cells: mitochondria, ATP, and insulin

TL;DR. Every cell needs ATP. Mitochondria make it from food. Insulin tells cells where to store extra fuel. When this machinery is overwhelmed by modern food, you get "metabolic dysfunction" — the upstream cause behind type 2 diabetes, nonalcoholic fatty liver disease, PCOS, atherosclerosis, and Alzheimer's. Casey Means cites a UNC analysis showing 93.2 percent of U.S. adults are metabolically unhealthy. Knowing the machinery lets you see why a handful of dietary changes — less sugar, less industrial seed oil, more fiber, more real food — keep showing up as the answer to wildly different-looking diseases. They take pressure off the same engine.

What you'll learn

• What ATP actually is and why every cellular function depends on it.
• How mitochondria turn food into ATP, and what "Bad Energy" means at the level of an individual cell.
• What insulin does, why it matters more than glucose, and how it switches the body between storage mode and burn mode.
• Why insulin resistance and metabolic syndrome show up years before any standard lab test catches them.
• The eight nutrient-sensing pathologies Lustig uses to explain why "not druggable but foodable" is a real category.
• Why fructose is metabolically not "just sugar."
• Three practical levers that take pressure off the machinery.

The energy hierarchy: food → ATP → everything

Every cell in your body — about 37 trillion of them — runs on a single rechargeable molecule called adenosine triphosphate, or ATP. ATP is not stored in any meaningful quantity; the average adult makes and uses roughly 88 pounds of it per day. ATP powers protein synthesis, DNA repair, ion pumping across membranes, muscle contraction, neurotransmitter release, immune defense, autophagy (the cell's clean-up crew), and the basic homeostasis that keeps you alive between heartbeats. When ATP runs short, cellular work backs up, the cell sends out distress signals, and over time tissues stop doing what they're supposed to do.

ATP is made almost entirely inside mitochondria — tiny organelles inside almost every cell, descended from a free-living bacterium that was engulfed by a larger cell about 1.5 billion years ago. Each cell carries somewhere between several hundred and several thousand mitochondria, with the highest counts in the tissues that burn the most energy: heart muscle (about 5,000 per cell), skeletal muscle, kidney, liver, and brain. Eggs at ovulation carry hundreds of thousands. When you see "metabolically active tissue," read "mitochondria-rich tissue."

The biochemistry is famously intricate but the essence is short. Carbohydrates, fats, and protein are broken down to a common molecule called acetyl-CoA, which feeds into the Krebs cycle (also called the TCA cycle) inside the mitochondrion. The Krebs cycle strips electrons off acetyl-CoA and loads them onto two carrier molecules (NADH and FADH2), which drop those electrons down the electron transport chain — a series of protein complexes in the inner mitochondrial membrane. Energy released in that drop pumps protons across the membrane, creating a gradient. Protons flow back through an enzyme called ATP synthase, rotating a molecular turbine that welds a phosphate onto ADP to make ATP. Oxygen sits at the end of the chain and accepts the spent electrons, which is the entire reason you breathe.

When mitochondria are damaged, overwhelmed, or simply outnumbered relative to the fuel coming in, this whole assembly stalls. Pyruvic acid backs up, gets shunted out of the mitochondrion, and is converted into fat by the liver (de novo lipogenesis). Electrons leak off the chain and produce reactive oxygen species — the source of "oxidative stress." Cells that can't make enough ATP cannot do their jobs and start to scream chemically for help. Casey Means calls the healthy state "Good Energy" and the dysfunctional state "Bad Energy"; the rest of the module is about how a modern diet reliably drives cells toward the second one.

Insulin: the storage signal

Insulin is a small protein hormone produced by the beta cells of the pancreas, and it is the body's master fuel-storage signal. Every time you eat carbohydrate, your blood glucose rises, your pancreas releases insulin, and insulin tells cells across the body what to do with the incoming fuel: muscle cells, take up glucose and burn it or store it as glycogen; liver cells, store glycogen and, if there's still excess, convert it to triglyceride; fat cells, take up fatty acids and lock them away. Insulin is simultaneously a "store this" signal and a "stop burning fat" signal — when insulin is high, fat cells essentially cannot release fat for fuel.

Until 1960, no one could measure insulin in human blood quantitatively. Solomon Berson and Rosalyn Yalow developed the radioimmunoassay that year at the Bronx VA, and the technique earned Yalow a 1977 Nobel Prize (Berson had died). Suddenly endocrinology became quantitative, and the first thing the new measurement revealed was that obese people and type 2 diabetics were not just hyperglycemic but hyperinsulinemic — their pancreases pumping out far more insulin than normal. That single finding rewired the question of obesity: eating raises insulin, chronically elevated insulin keeps fat locked in fat cells, and the older European hormonal-regulatory view of obesity suddenly had a measurable hormone behind it.

The body is supposed to swing between two states. Fed state: you ate, glucose is up, insulin is up, you're in storage and growth mode. Fasted state: blood glucose drops, insulin falls, glucagon rises, the liver releases stored glycogen, then fat cells release fatty acids, the liver converts some of those to ketones, and your body burns through its own fuel reserves. Both states are normal. What is not normal is the modern pattern: snacks, sweetened drinks, processed carbs from waking to bedtime, with insulin elevated essentially around the clock. Cells stop responding to the constant signal — exactly the way a doorbell becomes background noise if it never stops ringing — and the pancreas compensates by ringing the bell harder.

That compensation is what we call insulin resistance, and it is the single most important upstream change in modern chronic disease. Fasting glucose can stay "normal" for years while fasting insulin climbs. Eventually, when the pancreas can no longer keep up, glucose drifts upward and gets diagnosed as prediabetes, then type 2 diabetes. By that point the underlying problem has been brewing for decades.

Insulin resistance and metabolic syndrome

In his 1988 Banting Lecture to the American Diabetes Association, the Stanford endocrinologist Gerald Reaven proposed something heretical: that obesity, type 2 diabetes, hypertension, atherosclerosis, dyslipidemia, and glucose intolerance were not separate problems that happened to cluster, but downstream branches of a single upstream defect — insulin resistance and the compensatory hyperinsulinemia it produced. He initially called it Syndrome X. The field eventually settled on the name "metabolic syndrome."

The current diagnostic criteria (NCEP ATP III, with later updates) require any three of five: waist circumference above 40 inches in men or 35 in women, fasting triglycerides at or above 150 mg/dL, HDL cholesterol below 40 in men or 50 in women, blood pressure at or above 130/85, and fasting glucose at or above 100 mg/dL. The cluster is mechanistically coherent. Hyperinsulinemia drives the kidneys to retain sodium, which raises blood pressure. It drives the liver to produce more triglyceride-rich VLDL, which lowers HDL and shifts LDL toward the small-dense, atherogenic species. It drives storage of fat in places fat isn't supposed to live — liver, pancreas, muscle — which is the visible signature of metabolic syndrome.

In 2022, Joana Araújo and colleagues at the University of North Carolina published an analysis using NHANES data showing that only 6.8 percent of U.S. adults met criteria for optimal cardiometabolic health on all five measures. By the inverse, 93.2 percent of American adults are metabolically unhealthy. Means treats that figure as the central public-health fact of the era, and it is not really contested — it is just usually buried. Most of those 93 percent are not yet diagnosed with anything. They have rising fasting insulin, rising HOMA-IR (insulin times glucose, divided by 405), and rising triglyceride-to-HDL ratios — none of which appear on a standard annual physical, because most physicians do not order fasting insulin and the lab's "normal range" for fasting glucose runs up to 99 mg/dL.

This is why Lustig and Means both push fasting insulin as the single most valuable lab test. By the time HbA1c moves, the machinery has been broken for ten or twenty years. Insulin resistance precedes hyperglycemia by decades. Bad Energy is silent.

The eight cellular pathologies (Lustig's framework)

Robert Lustig, the UCSF pediatric endocrinologist who launched the modern anti-sugar argument with his 2009 lecture "Sugar: The Bitter Truth," argues in Metabolical that what we call chronic disease is actually eight subcellular processes going wrong. None of them has a clean drug target, which is why pharma has spent billions on drugs for nonalcoholic fatty liver disease and Alzheimer's with essentially nothing to show for it. All of them respond to food.

Glycation is the Maillard reaction happening inside you: sugars binding nonenzymatically to proteins, forming advanced glycation end products (AGEs) that stiffen tissues, scar arteries, and trigger receptors that drive inflammation. Fructose glycates about seven times faster than glucose; its breakdown product methylglyoxal is 250 times faster. Oxidative stress is the leakage of reactive oxygen species off the electron transport chain, overwhelming the cell's antioxidant defenses (glutathione, superoxide dismutase, catalase, the Nrf2-driven response). Mitochondrial dysfunction is the structural and functional decline of the mitochondria themselves — fewer of them, fatter and lazier cristae, less throughput.

Insulin resistance is the section above. Membrane integrity is the loss of healthy cell-membrane structure when industrial seed oils displace fish-derived omega-3s; the U.S. dietary omega-6:omega-3 ratio has shifted from roughly 1:1 in our evolutionary past to about 20:1 today, and the ratio in adipose tissue can shift in a matter of days with dietary change. Inflammation is chronic low-grade signaling driven by leaky gut, dietary AGEs, palmitate from de novo lipogenesis, and visceral fat behaving like an endocrine organ.

Epigenetics is the layer of methylation and acetylation marks that turn genes on and off without changing the underlying DNA; maternal diet, stress, and toxin exposure leave epigenetic marks that can persist multiple generations. Autophagy is the cell's nightly clean-up — damaged proteins and worn-out mitochondria recycled, the brain's glymphatic system clearing metabolic waste. Autophagy is triggered by fasting, exercise, and certain food compounds (urolithin A from pomegranate; sulforaphane from broccoli) and suppressed by constant eating and constant insulin.

Lustig's punchline: these pathways are nutrient-sensing. They respond to what you eat, when you eat it, and what your gut microbiome does with it. They do not respond well to drugs, because drugs are designed to hit single targets and these pathways are integrated. This is the structural reason a single intervention — eat real food, not industrial formulations — keeps showing up as effective across diseases that look completely different from the outside. Type 2 diabetes, fatty liver, PCOS, Alzheimer's, depression, and atherosclerosis all sit on the same upstream axis.

Fructose: the special case

Fructose deserves its own paragraph because it is metabolically not interchangeable with glucose. Glucose can be used by every cell in the body — your brain alone burns about 120 grams per day. Fructose, on the other hand, is almost entirely metabolized in the liver. When a 12-ounce can of soda hits the gut, the glucose half distributes across the body and gets used. The fructose half loads into the liver in a single bolus.

Inside the liver, fructose bypasses the regulatory step (phosphofructokinase) that normally limits glycolysis when energy is sufficient. Without that brake, fructose carbons get converted to fat through de novo lipogenesis — the same pathway by which ethanol makes a fatty liver. Lustig's argument is biochemically literal: a steady high-dose fructose load is to the liver what a steady alcohol load is. Fructose drives the Maillard reaction seven times faster than glucose. Its byproduct methylglyoxal is 250 times faster. It raises uric acid, which independently impairs mitochondrial function. It crosses the placenta and shows up in breast milk in proportion to maternal soda consumption.

This is why an apple is fine and a 12-ounce soda is not, even at similar fructose totals. The apple delivers its fructose slowly, packaged in a fiber matrix that the gut absorbs over an hour or more, while the colon's bacteria ferment what's left into short-chain fatty acids. The soda delivers its fructose as a free, fast load with no fiber to slow it down — a pure hepatic hit. Children now get adult diseases of alcohol (fatty liver, type 2 diabetes) without ever taking a drink. Up to 20 percent of U.S. children, and 42 percent of Hispanic men 25 to 30, have nonalcoholic fatty liver disease.

What this means for action

If the upstream problem is the same engine being overrun by the same handful of inputs, the interventions don't have to be exotic. Three categories take pressure off the machinery, in roughly descending order of leverage.

First, reduce continuous insulin signaling. The single largest gain for most people is cutting sugar-sweetened drinks: a can of soda is a clean hepatic fructose hit with nothing slowing it down. The second largest is shifting the carbohydrate portion of meals toward intact, fiber-bound forms — beans, whole grains, fruit, vegetables — and the third is letting the body experience a low-insulin state every day. A 12- to 14-hour overnight eating gap (last meal three hours before bed, breakfast the next morning) gives the liver time to clear fat and gives autophagy room to do its job. None of this requires hunger; it requires not eating something every two hours.

Second, reduce oxidative and inflammatory load. The omega-6:omega-3 ratio in your adipose tissue can shift in a few days with dietary change — fewer refined seed oils (soybean, corn, cottonseed, "vegetable oil"), more cold-water fish or omega-3 eggs, more olive oil, more whole-plant antioxidants from real food. The Nrf2 pathway — your cells' master antioxidant response — is activated by isothiocyanates in cruciferous vegetables (broccoli, brussels sprouts, arugula) and by polyphenols in berries, tea, and dark chocolate. This is not "supplementation"; it's eating plants on a regular schedule.

Third, build mitochondrial capacity. Mitochondria are not a fixed inheritance — they expand and contract in response to demand. Zone-2 cardio (brisk walking, easy cycling, anything you can sustain a conversation through) is the single most studied stimulus for mitochondrial biogenesis. Resistance training builds metabolically active tissue that takes glucose out of the bloodstream without much insulin and adds capacity that buffers everything else. Adequate sleep (seven to nine hours; six days of four-hour sleep induces measurable prediabetes in healthy adults) supports the autophagy and glymphatic clean-up that mitochondria need. Cold exposure activates brown fat, which is mitochondria-dense by design.

The framing matters. You are not "doing six different things"; you are taking pressure off one piece of machinery in three different ways.

Frequently Asked Questions

Does this mean low-carb is the answer?

Low-carb works for many people because it lowers chronic insulin signaling. So does Mediterranean. So does whole-food plant-based. So does time-restricted eating. The unifying mechanism is "stop hammering the machinery," not "eliminate this macronutrient." Lustig's six-word version is "protect the liver, feed the gut" — and several diets do both.

Are insulin sensitivity and insulin resistance really opposite ends of a spectrum?

Roughly, yes. Sensitivity means your cells respond strongly to a small amount of insulin; resistance means they need more and more to do the same job. HOMA-IR is the most common single-number summary, calculated from fasting insulin and fasting glucose. Below about 1.0 is excellent; above 2.5 to 2.8 indicates meaningful resistance.

Are GLP-1 drugs (Ozempic, Wegovy) treating insulin resistance?

Indirectly. GLP-1 agonists slow gastric emptying, blunt postprandial glucose spikes, and suppress appetite. The weight loss they produce does improve insulin sensitivity. But they are not fixing the upstream pathway — they are imposing a chemical brake on intake. Stop the drug, and unless the underlying diet has changed, the resistance comes back. That is consistent with the rest of the chronic-disease-drug pattern.

Is fasting really good for mitochondria?

Yes, in the modest sense most people mean: a 12- to 14-hour overnight gap, occasional 16-hour windows, or occasional 24-hour fasts trigger autophagy and mitophagy (selective clearing of damaged mitochondria) and let AMP-kinase activate. Extreme protocols (multi-day fasts, prolonged ketosis) have separate trade-offs and are not necessary for the core benefit.

How is metabolic syndrome different from prediabetes?

Prediabetes is defined by glucose alone (fasting glucose 100–125 mg/dL, or HbA1c 5.7–6.4 percent). Metabolic syndrome is the broader cluster — waist, blood pressure, triglycerides, HDL, glucose — that captures the same underlying problem earlier and across more tissues. You can have metabolic syndrome with perfectly normal fasting glucose.

Why do thin people get metabolic syndrome?

"TOFI" — thin on the outside, fat on the inside — describes people whose visible body fat is normal but whose visceral and liver fat are pathological. Roughly 40 percent of normal-weight U.S. adults are metabolically unhealthy. BMI misses them entirely; waist circumference, fasting insulin, triglyceride-to-HDL ratio, and ALT catch them.

Why does this matter if I'm not diabetic?

Because metabolic dysfunction sits upstream of most of what eventually kills Americans. Alzheimer's risk roughly doubles in type 2 diabetics; the literature now calls late-onset Alzheimer's "type 3 diabetes." Cardiovascular disease, the leading cause of death in the U.S., tracks insulin resistance better than it tracks LDL cholesterol. PCOS, infertility, NAFLD, gout, and many cancers share the same upstream mechanism. The diagnosis comes late; the machinery breaks early.

Is this the same as the "carnivore diet is the answer" position online?

No. The case in this module is that the upstream problem is overrun cellular machinery, and that real food in any of several patterns — Mediterranean, whole-food plant-based, traditional low-carb, time-restricted whole-food — gives that machinery room to recover. The carnivore claim that all plants are the problem is not what the biochemistry says.

Sources

1. Means, C. Good Energy: The Surprising Connection Between Metabolism and Limitless Health (2024). The 93.2 percent figure comes from Araújo, J., Cai, J., Stevens, J. "Prevalence of Optimal Metabolic Health in American Adults." Metabolic Syndrome and Related Disorders, 2019. DOI: 10.1089/met.2018.0105.
2. Lustig, R. H. Metabolical: The Lure and the Lies of Processed Food, Nutrition, and Modern Medicine (2021). Eight subcellular pathologies; fructose biochemistry.
3. Taubes, G. The Case Against Sugar (2016). Yalow and Berson 1960 radioimmunoassay; insulin as the dominant lipogenic hormone.
4. Gropper, S. S., Smith, J. L., Carr, T. P. Advanced Nutrition and Human Metabolism, 8th edition (2022). Mitochondrial ATP production; AMPK and mTOR; carnitine shuttle; de novo lipogenesis.
5. Reaven, G. M. "Banting Lecture 1988: Role of insulin resistance in human disease." Diabetes, 1988;37(12):1595–1607. DOI: 10.2337/diab.37.12.1595.
6. Yalow, R. S., Berson, S. A. "Immunoassay of endogenous plasma insulin in man." Journal of Clinical Investigation, 1960;39(7):1157–1175. DOI: 10.1172/JCI104130.
7. Hall, K. D., et al. "Ultra-processed diets cause excess calorie intake and weight gain." Cell Metabolism, 2019;30(1):67–77.e3. DOI: 10.1016/j.cmet.2019.05.008.

Related modules

• C1: What is food, really? (the modern industrial food question)
• C3: Sugar — the clearest case
• C4: You're not average — bio-individuality and the limits of dietary guidelines

Related glossary terms

• Mitochondria
• ATP
• Insulin
• Insulin resistance
• AMPK
• Metabolic syndrome
• Oxidative stress
• Nrf2