Blood-Brain Barrier

What the blood-brain barrier is

Inject a blue dye into the bloodstream of an animal and almost every organ turns blue — except the brain and spinal cord, which stay pale. That experiment, run by Paul Ehrlich and his student Edwin Goldmann between roughly 1885 and 1913, is the original demonstration of the blood-brain barrier (BBB). Goldmann's reciprocal version sealed it: injecting the same dye directly into the cerebrospinal fluid stained the brain but left the rest of the body clear. Something between blood and brain was blocking the dye in both directions.

That "something" is not a single membrane but a property of the brain's smallest blood vessels. Brain capillaries are lined by endothelial cells just like capillaries everywhere, but here the cells are welded together by continuous bands of tight junctions and carry almost none of the leaky transport vesicles found elsewhere. The result is a living tube whose wall is essentially uncrossable — unless a molecule passes through the endothelial cells via a specific molecular gate. The brain gets to pick exactly what comes in, keeping its delicate electrical environment stable while excluding toxins, pathogens, neurotransmitters circulating in the blood, and most drugs.

How the barrier is built — the neurovascular unit

The BBB is best understood as a layered assembly called the neurovascular unit, built around each brain capillary:

• Endothelial cells — the wall itself. A single flattened sheet of brain microvascular endothelial cells lines the capillary lumen. Unlike peripheral endothelium, these cells are sealed by tight junctions, have few fenestrations, almost no pinocytotic (transcytotic) vesicles, and a high density of mitochondria to power their many active transporters.
• Tight junctions — the seal. The junctions are protein zippers spanning the gap between adjacent endothelial cells. The load-bearing transmembrane proteins are claudin-5 (the dominant size-selective component) and occludin, linked to the actin cytoskeleton through cytoplasmic scaffolds ZO-1, ZO-2, and ZO-3. Adherens junctions (VE-cadherin) sit just below and provide mechanical adhesion. Knocking out claudin-5 in mice opens the barrier to molecules up to ~800 Da and is lethal within hours of birth.
• Pericytes — the regulators. Contractile mural cells embedded in the basement membrane, covering roughly 30% of the abluminal capillary surface. The brain has the highest pericyte coverage of any organ. They tune barrier tightness, suppress transcytosis, and control capillary diameter. Loss of pericytes (as in PDGFRβ-mutant mice) produces a leaky barrier and is linked to neurodegeneration.
• Basement membrane. A shared sheet of extracellular matrix (collagen IV, laminin, fibronectin) sandwiching the pericytes and anchoring everything.
• Astrocyte end-feet — the inducers. Star-shaped glia send broad "foot" processes that blanket more than 95% of the capillary surface. They don't form the seal, but they secrete the developmental signals (Sonic hedgehog, Wnt/β-catenin, angiopoietin-1, retinoic acid) that induce and maintain the endothelial barrier phenotype. Their end-feet are studded with aquaporin-4 water channels and Kir4.1 potassium channels for buffering the brain's fluid and ion balance.

Once the wall is built, traffic is sorted into a few routes. (1) Paracellular — between cells — is closed by tight junctions, so water-soluble molecules cannot squeeze through. (2) Passive transcellular diffusion — through the lipid membranes — works only for small (under ~400–500 Da), lipophilic, weakly hydrogen-bonding molecules: oxygen, carbon dioxide, ethanol, caffeine, nicotine, and many anesthetics cross this way. (3) Carrier-mediated transport uses specific solute carriers: GLUT1 (SLC2A1) for glucose, LAT1 (SLC7A5) for large neutral amino acids, MCT1 for ketone bodies and lactate. (4) Receptor-mediated transcytosis ferries large molecules — the transferrin receptor carries iron-loaded transferrin, the insulin receptor carries insulin, LRP1 carries lipoproteins. (5) Efflux pumps — P-glycoprotein (ABCB1) and BCRP (ABCG2) on the blood-facing membrane — actively pump lipophilic xenobiotics back out using ATP, which is why even drugs that diffuse in often don't accumulate.

The numbers that make it a wall

The barrier's selectivity is quantifiable, and the figures are extreme compared to ordinary capillaries:

• Transendothelial electrical resistance (TEER): ~1500–2000 Ω·cm² in brain capillaries versus ~3–30 Ω·cm² in peripheral capillaries — roughly 100× tighter. High resistance means the paracellular gaps barely leak ions.
• Total surface area: The human brain holds an estimated 400+ miles (~640 km) of capillaries presenting roughly 12–20 m² of exchange surface — about the floor area of a small room, packed into ~1.4 kg of tissue. Every neuron sits within ~25 µm of a capillary.
• Passive size cutoff: ~400–500 Da, and even then only if lipophilic with fewer than ~8–10 hydrogen bonds. Add a charge or polarity and the molecule is excluded.
• Glucose throughput: ~120 g/day across GLUT1 — roughly 60% of the body's resting glucose consumption, feeding a brain that is ~2% of body mass.
• Vesicle density: Brain endothelium has ~5–10× fewer pinocytotic vesicles than peripheral endothelium, slamming the transcytotic route shut by default.
• Drug exclusion: ~98% of small-molecule drugs and ~100% of large-molecule biologics fail to reach therapeutic brain levels.

Brain capillary vs ordinary capillary

Property	Brain capillary (BBB)	Peripheral capillary
Tight junctions	Continuous, claudin-5 / occludin	Sparse or absent
Electrical resistance (TEER)	~1500–2000 Ω·cm²	~3–30 Ω·cm²
Fenestrations (pores)	None (continuous)	Often fenestrated (kidney, gut, endocrine)
Pinocytotic vesicles	Very few — transcytosis suppressed	Abundant
Pericyte coverage	~30% (highest in body)	~10–25%
Glial wrapping	Astrocyte end-feet >95% coverage	None
Efflux pumps	High P-gp / BCRP density on luminal side	Low
Default permeability	Selective — gated entry only	Broadly permeable to small solutes

How molecules actually cross — by route

Route	Mechanism	Example cargo	Energy
Paracellular	Between cells (blocked by tight junctions)	Water-soluble ions — mostly excluded	None (closed)
Passive transcellular	Diffusion through lipid membranes	O₂, CO₂, ethanol, caffeine, nicotine, anesthetics	None (down gradient)
Carrier-mediated (facilitated)	Solute carrier flips conformation	Glucose (GLUT1), amino acids (LAT1), lactate/ketones (MCT1)	None (down gradient)
Receptor-mediated transcytosis	Receptor binds, vesicle shuttles across	Transferrin (iron), insulin, LRP1 ligands	ATP (vesicle traffic)
Adsorptive transcytosis	Charge-based vesicle uptake	Cationized albumin, cell-penetrating peptides	ATP
Active efflux	ATP pump ejects cargo back to blood	Many drugs, toxins (P-gp / ABCB1, BCRP / ABCG2)	ATP

Where it shows up — disease, drugs, and exceptions

• The drug-delivery problem. The BBB is the single biggest reason CNS drug development fails. Antibiotics that clear an infection in the body can fail in meningitis; chemotherapy that shrinks tumors elsewhere can't reach a glioblastoma. Strategies to get past it include designing small lipophilic molecules, hijacking carriers ("molecular Trojan horses" fused to transferrin- or insulin-receptor antibodies), inhibiting P-glycoprotein, intranasal delivery along olfactory nerves, and focused ultrasound with microbubbles to transiently pop the junctions open.
• L-DOPA, the classic Trojan horse. Dopamine cannot cross the BBB, so Parkinson's disease can't be treated with dopamine directly. L-DOPA, its precursor, can cross — it rides LAT1, the large-neutral-amino-acid carrier — and is then converted to dopamine inside the brain. This is textbook exploitation of a native transporter.
• Stroke and edema. Ischemia degrades tight junctions within hours; the leaking barrier drives vasogenic edema, and the resulting swelling inside the rigid skull can be lethal. Reperfusion can worsen the breakdown.
• Multiple sclerosis. Focal barrier opening lets autoreactive T-cells invade; the gadolinium contrast that "lights up" active MS lesions on MRI is literally marking where the BBB has failed. Natalizumab works by blocking the α4-integrin lymphocytes use to cross.
• Brain tumors. Tumor neovasculature is disorganized and leaky (the "blood-tumor barrier"), which is why contrast agents highlight a glioma on MRI — but the leak is patchy, so drug delivery to the tumor edge stays poor.
• Pathogens that breach it. Neisseria meningitidis, Streptococcus pneumoniae, Listeria monocytogenes, and HIV-infected immune cells (the "Trojan horse" route) actively cross or are carried across to cause CNS infection.
• GLUT1 deficiency syndrome. Loss of one SLC2A1 copy starves the brain of glucose despite normal blood sugar, causing infantile seizures and developmental delay; a ketogenic diet rescues it by supplying ketone bodies that cross on MCT1 instead.
• Neurodegeneration. Chronic, subtle barrier leak and pericyte loss are increasingly implicated in Alzheimer's disease and small-vessel dementia, partly by impairing clearance of amyloid-β through LRP1.

Common misconceptions

• "The blood-brain barrier is a membrane around the brain." No — it is a property of the capillary walls inside the brain, distributed across hundreds of miles of vessels. There is no single enclosing sheet; every microvessel is its own barrier.
• "Astrocytes form the seal." Astrocyte end-feet induce and maintain the barrier, but the actual seal is the endothelial tight junctions. Astrocytes are the signal, not the wall.
• "If a molecule is small enough, it gets in." Size is necessary but not sufficient. The molecule must also be lipophilic and form few hydrogen bonds — and even then P-glycoprotein may pump it right back out. Many small drugs are excellent P-gp substrates and never accumulate.
• "The whole brain is sealed." The circumventricular organs (area postrema, median eminence, subfornical organ, neurohypophysis, pineal) have leaky fenestrated capillaries on purpose, to sense or secrete into the blood. The choroid plexus forms a separate blood-CSF barrier at its epithelium instead.
• "The barrier just passively keeps things out." It is metabolically expensive and active: ATP-driven efflux pumps eject toxins, and dedicated carriers actively import nutrients. Brain endothelium is mitochondria-rich precisely because the barrier is a busy, selective gate, not a passive wall.
• "Once it's broken, it's broken." The barrier is dynamic and can repair. And opening it can be done deliberately and reversibly — focused ultrasound plus microbubbles transiently separates tight junctions for hours, a route now in clinical trials for delivering chemotherapy and antibodies into tumors and Alzheimer's plaques.