395 – Brain lipidology: understanding APOE, cholesterol homeostasis, Alzheimer’s disease, & more

KEYPOINTS

• 01

  The brain runs its own cholesterol system, almost completely separate from the blood.

  Lipidologist Tom Dayspring's central point: apoB particles (LDL, VLDL) are far too big to cross the blood-brain barrier, so how much cholesterol is in plasma has 'zero to do with' brain cholesterol. A 2-year-old with LDL around 30 mg/dL is building brain tissue faster than at any other point in life, which shows that low circulating cholesterol doesn't starve the brain. That undercuts the common fear that drug-lowered LDL harms cognition.

• 02

  LDL's main job is bringing cholesterol back to the liver, not delivering it to cells.

  Every cell can make all the cholesterol it needs. HDLs (about 90% of all lipoprotein particles) pick up excess cholesterol, then hand much of it to long-lived LDLs (3–5 day lifespan), which the liver clears through LDL receptors. This is indirect reverse cholesterol transport. When LDL falls, cells are simply in balance and less cholesterol needs returning.

• 03

  Plasma holds only a sliver of the body's cholesterol; the brain holds the most.

  Of roughly 140 g of cholesterol in the body, the brain holds about 20–25 g versus 3–5 g in the liver, which is a transfer station rather than a vault. Brain cholesterol has a half-life of about five years, versus days in the periphery. Red blood cell membranes carry more blood cholesterol than lipoproteins do. Halving plasma cholesterol from 200 to 100 cuts total body cholesterol by only a couple of percent.

• 04

  Atherosclerosis comes from apoB particle number, mainly through poor clearance.

  Above a threshold concentration, apoB particles diffuse into even healthy artery walls, get trapped, oxidize and are engulfed by macrophages that become foam cells. Every trial and Mendelian randomization study shows that lowering apoB lowers events. Insulin resistance, inflammation, blood pressure, smoking and autoimmune disease make entry easier. The rare 90-year-old smoker with high LDL and no plaque is genetically lucky, and Dayspring warns nobody can know in advance whether they're one of them.

• 05

  After about age 10, neurons stop making cholesterol and outsource it to astrocytes.

  Building one cholesterol molecule takes 37 enzymatic steps and over 30 ATP, and neurons need their energy for firing. Oligodendrocytes make about 70% of brain cholesterol, mostly for myelin; astrocytes package cholesterol onto apoE into HDL-sized 'brain HDLs' that travel through the brain's tissue matrix. Neurons take them up mainly through LRP1 (which recognizes only apoE), plus LDL receptors and SR-B1. Dayspring suggests calling the LDL receptor the 'apoB/apoE receptor'.

• 06

  Plasma desmosterol tracks brain cholesterol synthesis; 24S-hydroxycholesterol signals neurons getting rid of excess.

  The periphery mostly uses the lathosterol pathway, while astrocytes use the desmosterol (Bloch) pathway, so blood desmosterol correlates closely with levels in spinal fluid and brain. Neurons convert surplus cholesterol into water-soluble 24S-hydroxycholesterol, which crosses the barrier and goes to the liver for excretion. Rising plasma levels suggest a neuron under stress; only research labs measure it.

• 07

  APOE4 makes dysfunctional brain HDLs, disturbing neuron membranes and tilting APP processing toward toxic amyloid-β42.

  Peter's numbers: about 55% of people are E3/E3, 20–25% E3/E4, 1–2% E4/E4, and E2/E2 is under 0.5%. One E4 copy roughly doubles Alzheimer's risk (some series say triples); two copies are now usually put at 8–12x, down from older estimates of 20–25x. E4 particles bind neuron receptors poorly, so cholesterol lodges in membranes but doesn't reach the cell interior. Membrane cholesterol excess favors beta/gamma-secretase cleavage of APP into Aβ42 over the alpha-secretase route; brains with Alzheimer's at autopsy are cholesterol-overloaded.

• 08

  Statins are the only lipid drugs that reach the brain, and randomized trials show neutral or beneficial cognitive effects.

  Lipophilic statins enter faster, but at steady state hydrophilic ones like rosuvastatin get in too, so Dayspring doesn't pick a statin by solubility. Peter's earlier review of meta-analyses found neutrality or improvement for MCI and dementia, though dementia was always a secondary outcome. The hypothesis: modestly slowing brain synthesis may help, while oversuppression may explain the occasional statin 'brain fog' that clears when the drug stops.

• 09

  HDL's real value is its protein cargo and immune function, not its cholesterol count.

  More than 200 proteins have been found on HDL particles, each particle carrying one or two. The tiniest, protein-laden apoA-I species can cross the blood-brain barrier and join apoE brain HDLs, possibly delivering anti-inflammatory and antioxidant proteins. No commercial test measures whether HDL is functional or dysfunctional.