Tom opens with the foundational principle: cholesterol is essential for cell membrane integrity in every cell of the body, so evolution gave every cell the ability to synthesize it. But excess cholesterol crystallizes and kills cells, so evolution also gave cells a way to export it. Since plasma is aqueous and cholesterol is hydrophobic, lipoproteins evolved as carriers — proteins that bind lipids and wrap them into water-soluble particles. Tom walks through the ApoB family (LDL, VLDL, IDL, chylomicrons) and the ApoA family (HDL), explaining how the liver and intestine each produce ApoB particles with ApoB-100 and ApoB-48 respectively. He highlights the long plasma residence time of LDL (3–5 days) and introduces a concept that challenges common understanding: LDL's primary function is to return cholesterol to the liver via LDL receptors, not to deliver it to peripheral cells. HDLs, he explains, both transport cholesterol directly and transfer it to LDL for indirect return — making 'reverse cholesterol transport' a dual direct-plus-indirect system. He closes by teasing the brain discussion: the fear that lowering LDL will deprive the brain of cholesterol is physiologically unfounded.

Peter synthesizes Tom's opening framework by drawing an analogy between plasma and the body's highway system, reinforcing why things that can't dissolve in water need carriers. He then poses the crucial question: if LDL is so important for reverse cholesterol transport, does lowering LDL deprive the body? Tom explains that lower LDL simply reflects a system in balance — cells are not effluxing as much cholesterol, so there's less need for it to return to the liver. Peter then asks where most of the body's cholesterol actually lives. Tom reveals a layered surprise: most cholesterol is in cells, not plasma; and within the bloodstream, the biggest carrier is not lipoproteins but red blood cells. The brain — not the liver as most people assume — holds the most cholesterol of any organ, roughly 20–25 grams versus the liver's 3–5 grams. Peter closes with the essential reassurance: if your LDL falls from 200 to 100 mg/dL, total body cholesterol has declined by only a couple of percent ^{[1] — Peter Attia "Total body cholesterol in plasma vs. cells: The vast majority of the body's cholesterol is stored within cells; plasma cholesterol represen…" 19:48 Play} , making fears about 'cholesterol deprivation' physiologically baseless.

Tom opens with an unambiguous declaration: without cholesterol accumulation in the artery wall, atherosclerosis cannot exist. The question becomes how cholesterol gets there. Once ApoB-containing particles — primarily LDL, given its long plasma residence time — exceed a threshold concentration, simple diffusion carries them across the arterial endothelium. Inside, they are trapped, oxidized, engulfed by macrophages, and become foam cells that coalesce into plaque. Since every ApoB-containing particle carries exactly one ApoB molecule, measuring ApoB gives a perfect particle count — more reliable than LDL cholesterol, which is merely a proxy. Tom explains that the liver's failure to express enough LDL receptors is the dominant cause of elevated ApoB; every particle the liver fails to clear is one that can enter an artery wall ^{[1] — Tom Dayspring "Once ApoB-containing particles exceed a concentration threshold, they diffuse into the arterial wall, get oxidized, trigger macrophage infi…" 20:40 Play} . Peter then raises the puzzle of heterogeneous risk: two people with identical ApoB levels can have radically different outcomes. Tom attributes this to metabolic health (insulin resistance, type 2 diabetes), chronic inflammation, blood pressure, oxidative stress, and collagen diseases — plus unknown genetic protective factors. He cautions against assuming personal immunity from cardiovascular disease based on a family history of longevity.

Tom opens with the definitive statement: what is happening with brain cholesterol has zero connection to what is circulating in plasma. ApoB particles — the main cholesterol carriers in blood — cannot cross the blood-brain barrier. Instead, the fetal brain begins synthesizing its own cholesterol in the second and third trimesters, ramping up production because neurons and glial cells need massive amounts for their membranes and for myelin. Oligodendrocytes produce roughly 70% of brain cholesterol, primarily used to wrap axons and dendrites in the myelin sheaths that enable fast neural signaling. Microgliocytes serve as the brain's immune cells, and astrocytes handle much of the remaining synthesis. To move cholesterol between cells through the brain's interstitial space — the matrosome — the astrocytes package it into ApoE-containing lipoproteins that resemble plasma HDL in density but are fundamentally different in composition, carrying copies of ApoE rather than ApoA-1 ^{[1] — Tom Dayspring "The brain and body run completely separate cholesterol systems. ApoB-containing particles that carry most of your plasma cholesterol are fa…" 29:05 Play} . These particles mature in the matrosome, and neurons take them up via LDL receptors and especially LRP1, both of which recognize ApoE. Tom argues the receptor should be renamed the ApoB/ApoE receptor to reflect its true ligand range across both body compartments.

Peter asks Tom to explain precisely how an LDL particle 'docks' with an LDL receptor. Tom describes the receptor-binding domain on ApoB — specific amino acids that create an electrostatic surface charge complementary to the recognition domain on the LDL receptor. Crucially, this conformational alignment is size-sensitive: large or small LDL particles have distorted ApoB conformations that poorly expose the binding domain, explaining why individuals with predominantly small or large LDL particles can have elevated particle counts despite seemingly normal LDL cholesterol. Tom then reveals a recent NIH finding: LDL receptors function as dimers — two receptors pair up like lobster claws and grab two LDL particles simultaneously. He also explains that in the periphery, ApoE on rare LDL particles can also bind the receptor, and that in the brain, the same receptor exclusively handles ApoE-containing particles, reinforcing his call to rename it the ApoB/ApoE receptor.

Peter asks about the two cholesterol synthesis pathways that share a common trunk before bifurcating at their penultimate sterol. Tom explains that in peripheral tissues, the lathosterol pathway dominates; in the brain, it is the desmosterol pathway — used by astrocytes — that is predominant. When neurons are young and actively synthesizing cholesterol, they use the lathosterol pathway; after age 10, when neurons stop making cholesterol, lathosterol production in the brain essentially ceases. Desmosterol, by contrast, continues to be produced by astrocytes throughout life and reflects ongoing brain cholesterol synthesis ^{[1] — Tom Dayspring "Desmosterol as brain cholesterol marker: Plasma desmosterol correlates highly with cerebrospinal fluid desmosterol and brain tissue cholest…" 1:18:20 Play} . The remarkable clinical implication: plasma desmosterol correlates highly with cerebrospinal fluid desmosterol, meaning a routine blood test can serve as a window into the brain's cholesterol factory. A study using mass spectrometry found that people with low desmosterol have higher rates of cognitive impairment and Alzheimer's disease. Tom also notes that desmosterol is commercially measurable, unlike 24S-hydroxycholesterol, which remains confined to research settings — though both are valuable biomarkers of brain cholesterol health.

Peter explicitly distinguishes APOE the gene (written in all caps) from apoE the protein (written with a lower-case 'apo'), a nomenclature distinction that often causes confusion. Tom explains that the ApoE protein comes in three isoforms — E2, E3, and E4 — differing by just one amino acid, but that single substitution changes the protein's ability to bend and bind, altering its function as a ligand on lipoproteins. Since individuals inherit one allele from each parent, six genotype combinations are possible. Peter provides the population frequencies: roughly 55% are E3/E3 (wild type), 20–25% are E3/E4, and 1–2% are E4/E4. E2/E2 is exceedingly rare. The risk associated with Alzheimer's disease scales nonlinearly: E3/E4 individuals face approximately 2–3 times the risk of E3/E3 carriers ^{[1] — Peter Attia "A single amino acid difference between ApoE isoforms changes the protein's shape and function. Carrying two E4 copies raises Alzheimer's ri…" 49:35 Play} , while E4/E4 homozygotes face 8–12-fold higher risk. Earlier estimates of 20–25-fold for E4/E4 have been revised downward as studies have matured. The underlying reason — which the next chapter will elaborate — is that the ApoE4 protein is structurally inferior at its job of transporting cholesterol to neurons.

Tom uses the observation that 90% of all lipoproteins are HDL particles to make a powerful point: if HDL were primarily about cholesterol transport, the body is massively overbuilt for that task. HDL cholesterol — the number everyone knows — tells us almost nothing about what HDL particles are actually doing. The real action is in the protein cargo: over 200 distinct proteins have been identified on HDL particles in peripheral circulation, each conferring different functions — anti-inflammatory, anti-oxidative, immune-modulating, coagulation-related. Because individual HDL particles are tiny, each carries only one or two of these proteins, creating a diverse army of functionally specialized HDL subpopulations. The smallest, most protein-dense HDL particles — essentially free ApoA-1 or ApoA-1 bound to one or two accessory proteins — can cross the blood-brain barrier ^{[1] — Tom Dayspring "The brain and body run completely separate cholesterol systems. ApoB-containing particles that carry most of your plasma cholesterol are fa…" 29:05 Play} , likely via the scavenger receptor B1, and fuse with brain ApoE HDL particles. This is how the brain receives ApoA-1 it cannot synthesize itself. If these peripheral HDL particles carry the right protective proteins, they may confer anti-inflammatory or anti-oxidative benefits inside the brain. If they are dysfunctional — carrying harmful proteins — they may worsen neurological disease.

This is the mechanistic heart of the episode. Tom traces the full pathological chain: APOE4-carrying astrocytes produce an inferior ApoE4 protein that, when assembled into brain HDL particles, binds poorly to neuronal receptors (LDLR and LRP1). The result is decreased internalization of the cholesterol-laden particle. Instead of being fully taken into the lysosome where cholesterol can be liberated into the cytosol, the particle only deposits cholesterol onto the outer cell membrane — creating a paradox of membrane overload and cytosolic deficiency simultaneously ^{[1] — Tom Dayspring "APOE4 produces inferior ApoE proteins that bind poorly to neuronal receptors, starving neurons of cholesterol. Without proper membrane chol…" 59:40 Play} . With too much cholesterol in the neuronal cell membrane, beta and gamma secretases preferentially cleave amyloid precursor protein into amyloid-beta-42, the more toxic and aggregation-prone form ^{[2] — Tom Dayspring "When neurons accumulate too much cholesterol, they convert it to 24S-hydroxycholesterol — a water-soluble form that can escape through the …" 1:04:10 Play} . The proper ratio of cholesterol favors alpha-secretase cleavage into the less-toxic amyloid-beta-40. Tom also explains the neuron's cholesterol overflow mechanism: 24S-hydroxycholesterol, an oxysterol that is more water-soluble and can escape through the blood-brain barrier into plasma. Elevated plasma 24S-hydroxycholesterol signals active neuronal cholesterol overload — a biomarker researchers use to track Alzheimer's drug efficacy. Tom notes that statins have been shown to reduce these plasma levels, suggesting they may be correcting neuronal cholesterol excess.

Peter pivots to pharmacology, and Tom draws an immediate boundary: only statins among all ApoB-lowering drug classes can penetrate the blood-brain barrier. PCSK9 inhibitors, bile acid sequestrants, ezetimibe, and CETP inhibitors all work outside the brain. In steady state, even hydrophilic statins (like rosuvastatin) ultimately reach the brain, so the lipophilicity distinction matters less than commonly assumed for long-term therapy. Since Alzheimer's pathology involves excessive neuronal cholesterol, the hypothesis that statins might benefit the brain by moderately reducing brain cholesterol synthesis is logically coherent and is supported by the data: meta-analyses of statin trials show statins are either neutral or beneficial with respect to dementia outcomes — no study has shown harm ^{[1] — Tom Dayspring "All statins — hydrophilic or lipophilic — reach the brain in steady state and can reduce cholesterol synthesis there. Meta-analyses show st…" 1:10:00 Play} . But the flip side is real: cholesterol is also essential for brain function, and some patients develop brain fog on statins that promptly resolves on discontinuation. Tom proposes this is the over-suppression subgroup — individuals in whom the statin too aggressively reduces brain cholesterol synthesis. Monitoring plasma desmosterol could identify these patients before symptoms emerge, allowing dose adjustments. Similarly, statins have been shown to reduce plasma 24S-hydroxycholesterol, which Tom interprets as evidence of normalized neuronal cholesterol levels.

Tom synthesizes the desmosterol and 24S-hydroxycholesterol threads that have run through the episode. Plasma desmosterol is commercially measurable and correlates strongly with CSF desmosterol — a published study using mass spectrometry confirmed high correlation and found that low plasma desmosterol is associated with increased rates of cognitive impairment and Alzheimer's disease ^{[1] — Tom Dayspring "24S-hydroxycholesterol as neuronal distress marker: Elevated plasma 24S-hydroxycholesterol signals that neurons are excreting excess choles…" 1:06:40 Play} . If a patient is on a statin and their plasma desmosterol falls too low, that could indicate over-suppression of brain cholesterol synthesis, warranting a dose reduction or switch to a non-statin ApoB-lowering strategy. 24S-hydroxycholesterol, produced when neurons need to shed excess cholesterol, is elevated in early Alzheimer's disease and falls with statin use — another marker of improved neuronal cholesterol balance. Unfortunately, 24S-hydroxycholesterol is not yet available through commercial laboratories. Tom expresses the hope that it will become a routine clinical tool, since it would allow clinicians to simultaneously confirm adequate statin effect on brain cholesterol (falling 24S-hydroxycholesterol) and absence of over-suppression (stable desmosterol).

Peter asks about ezetimibe — a drug that works in the gut by blocking intestinal cholesterol absorption and might seem irrelevant to brain health. Tom acknowledges it shouldn't have a brain effect based on first principles, but then reveals that ezetimibe's main metabolite, ezetimibe glucuronide, can cross the blood-brain barrier in small amounts ^{[1] — Tom Dayspring "Ezetimibe works in the gut and can't cross the blood-brain barrier — but its metabolite, ezetimibe glucuronide, can. Animal studies show it…" 1:19:40 Play} . In animal studies, it interferes with hexokinase and reduces glycosylation of brain proteins, producing anti-inflammatory effects. Neurologists who specialize in Alzheimer's prevention — Tom names Richard Isaacson and Kelly Endiotes — have anecdotally observed cognitive benefits in patients taking ezetimibe, beyond what would be expected from ApoB lowering alone. Tom frames a plausible clinical strategy: for APOE4 carriers with elevated ApoB, a combination of low-dose statin plus ezetimibe may target both cardiovascular and brain health simultaneously. He notes that existing trial biobanks from ezetimibe studies could in principle be mined for pre- and post-treatment P-tau and amyloid-beta-40/42 data without needing a new trial.

Tom explains the complete journey of dietary omega-3s from supplement to neuron. After intestinal absorption as free fatty acids or lysophospholipids, EPA and DHA are repackaged into chylomicrons, hydrolyzed in plasma, and then bound by phospholipid transfer protein — a specialized shuttle that carries omega-3-containing phospholipids to the blood-brain barrier, where a specific receptor mediates their entry into the brain ^{[1] — Tom Dayspring "DHA and EPA reach brain cells via a phospholipid transfer protein shuttle that docks at a specific blood-brain barrier receptor. Both fatty…" 1:23:54 Play} . Once inside, they can be incorporated into brain HDL particles or diffuse directly into nearby cells. Tom notes a shift in the field: EPA is no longer considered secondary to DHA; both fatty acids serve important roles in brain cell membrane composition. The evidence hierarchy is mixed: observational data is broadly supportive, and Bill Harris's work suggests an omega-3 index of 8–9% reflects adequate systemic saturation, but no randomized controlled trial has established a brain-specific outcome benefit at any particular index level. Tom is candid that this places omega-3s in 'plausible but unproven' territory for brain health specifically, while the cardiovascular evidence from the REDUCE-IT trial (EPA in high-risk patients with controlled ApoB) is more robust.

Peter closes the pharmacology discussion with the most forward-looking topic: obicetrapib, a CETP inhibitor that inhibits the transfer of cholesterol from HDL to ApoB particles, causing HDL to enlarge and ApoA-1 to rise. The biological rationale for neuroprotection comes from genetics: people with natural CETP loss-of-function variants have lower rates of Alzheimer's disease and cognitive impairment. Pharmacologically mimicking this state could theoretically confer the same benefit ^{[1] — Tom Dayspring "Obicetrapib, a CETP inhibitor, raises ApoA-1 and generates tiny protein-rich HDL particles that can cross the blood-brain barrier. The BROA…" 1:31:50 Play} . Tom explains the cascade: obicetrapib enlarges HDL particles, the liver interprets the reduced free ApoA-1 as a deficiency and ramps up ApoA-1 production, and circulating ApoA-1 rises. These free ApoA-1 molecules and the tiny protein-rich HDL particles they form can cross the blood-brain barrier, fuse with brain ApoE HDL particles, and potentially convert dysfunctional APOE4 brain HDL into functional units capable of properly delivering cholesterol to neurons. The BROADWAY trial, conducted primarily for cardiovascular endpoints, measured Alzheimer's biomarkers — phosphorylated tau, amyloid-40/42 ratios, and fibrillary markers — and found movement in the right direction. New Amsterdam Pharma is now planning further trials with cognitive endpoints and possibly PET imaging. Tom and Peter are cautiously optimistic, noting that getting the timing and patient selection right — catching APOE4 carriers before irreversible pathology accumulates — will be the key challenge, citing the PREDIMED trial's early-halt success as a model for well-targeted primary prevention.