Bypassing the blood-brain barrier: intrathecal CSF injection vs. intranasal olfactory-nerve delivery
The blood-brain barrier (BBB) is a highly selective interface formed by brain capillary endothelial cells sealed together with tight junctions, wrapped in pericytes and astrocyte end-feet. It protects the CNS from toxins and pathogens — but the same seal excludes the overwhelming majority of therapeutic molecules, especially biologics, gene therapies, and many antibiotics or chemotherapeutics.
Brain capillaries differ fundamentally from peripheral capillaries. Endothelial cells are fused by tight junctions (claudins, occludins, ZO proteins) that eliminate the paracellular gaps found elsewhere in the body. Pericytes wrap the abluminal surface, and astrocyte end-feet cover roughly 99% of the capillary wall, releasing signals that maintain barrier integrity. Efflux transporters (P-glycoprotein, BCRP, MRPs) actively pump many molecules that do cross the membrane back out into the blood, further reducing net penetration.
The result is a barrier that is not merely a passive wall but an active, selective gatekeeper — permitting only small, lipophilic, non-effluxed molecules (or those with a dedicated carrier, like glucose via GLUT1) to pass at meaningful rates.
Lipinski-type rules of thumb for BBB permeability (molecular weight <400–500 Da, high lipophilicity, few hydrogen-bond donors) exclude most peptides, antibodies, oligonucleotides, and many small-molecule chemotherapeutics from systemic dosing reaching the brain.
One tempting fix is simply giving more drug intravenously or orally. In practice this rarely works: increasing systemic exposure to reach therapeutic brain concentrations often means crossing toxic thresholds in the liver, kidneys, bone marrow, or gut long before enough drug accumulates in the CNS. Efflux transporters at the BBB also saturate poorly, meaning the fraction reaching brain tissue stays a nearly fixed, small percentage of plasma concentration regardless of dose escalation.
This dose-limiting toxicity problem is especially severe for CNS malignancies, CNS infections, and neurodegenerative disease, where the target tissue is precisely the compartment the barrier is built to exclude drugs from.
Pharmaceutical science has pursued two broad strategies: (1) engineer drugs or carriers to cross the barrier — receptor-mediated transcytosis shuttles, nanoparticle carriers, focused-ultrasound-mediated transient BBB opening; or (2) bypass the barrier geographically by delivering the drug to a CNS compartment that never requires crossing it. This simulation focuses on the second strategy: intrathecal injection into the cerebrospinal fluid, and intranasal delivery via olfactory/trigeminal nerve pathways. Both routes place the drug inside — or adjacent to — CNS tissue without asking it to cross the endothelial tight-junction wall at all.
Intrathecal (IT) administration injects drug directly into the subarachnoid space via lumbar puncture, placing the molecule immediately into the cerebrospinal fluid (CSF) that bathes the spinal cord and brain. Because the CSF compartment sits inside the barrier, IT delivery achieves direct CNS exposure that completely sidesteps the BBB.
A spinal needle is introduced between lumbar vertebrae (commonly L3–L4 or L4–L5, below the level where the spinal cord itself ends, to avoid cord injury) and advanced into the subarachnoid space. Free flow of CSF confirms correct placement, and drug is injected directly into that fluid. Because the needle bypasses skin, epidural fat, and dura only — never blood vessels feeding the brain — the drug enters a CNS-contiguous compartment in a single step.
IT delivery is used clinically for intrathecal chemotherapy (e.g., methotrexate, cytarabine for leptomeningeal disease), intrathecal baclofen for severe spasticity, intrathecal opioids for refractory pain, and antisense oligonucleotide therapies (e.g., nusinersen for spinal muscular atrophy) that are far too large and polar to cross the BBB from blood.
Because CSF turns over roughly 3–4 times per day, an intrathecally delivered drug is also cleared relatively quickly from the CSF unless it partitions into CNS tissue or a sustained-release/repeat-dosing strategy is used.
After injection, drug movement within the CSF is governed largely by bulk CSF flow (driven by choroid plexus production, cardiac-cycle pulsations, and postural/respiratory pressure changes) rather than by simple diffusion, since diffusion alone is far too slow across the scale of the neuraxis. Rostral (upward, toward the brain) spread depends on injection volume, patient positioning, CSF density relative to the drug formulation (baricity), and time.
Because the injection originates at the lumbar cistern, concentration is typically highest near the spinal cord and lower CNS structures immediately after dosing, with a concentration gradient declining toward the cerebral ventricles and cortex unless specific measures (large volume, prolonged positioning, intraventricular catheters) are used to promote cranial distribution.
IT delivery is invasive: it requires a trained clinician, sterile technique, and carries risks of post-lumbar-puncture headache, infection (meningitis), bleeding, and nerve irritation. Repeat dosing may require an implanted intrathecal catheter and pump (common for chronic baclofen or opioid therapy) to avoid repeated punctures. Despite the invasiveness, IT delivery remains the most direct and reliable way to guarantee CNS exposure for large or polar molecules that have no other path into the CSF compartment.
Intranasal (IN) administration deposits drug on the nasal epithelium, where a fraction is taken up by olfactory sensory neurons and trigeminal nerve fibers whose axons pass through the cribriform plate directly into the olfactory bulb and brainstem — a non-invasive route that, like intrathecal injection, bypasses the BBB, though via a completely different anatomical path.
The olfactory region sits at the roof of the nasal cavity. Olfactory sensory neurons there are unusual: they are the only neurons in the body with cell bodies exposed to the external environment (via the nasal mucosa) while their axons project directly into the CNS, passing through tiny perforations in the cribriform plate of the skull to synapse in the olfactory bulb. Drug applied to this epithelium can travel along the extracellular or intracellular (axonal transport) space surrounding these nerve fibers, reaching the olfactory bulb and, from there, adjacent forebrain structures — without ever entering the bloodstream or crossing the BBB.
The trigeminal nerve, which also innervates the nasal mucosa and enters the brainstem, offers a second, parallel nerve-mediated route, relevant for reaching caudal (brainstem-adjacent) structures.
Because part of any intranasal dose is also absorbed into the nasal mucosal bloodstream and distributed systemically, "intranasal" delivery is a mix of direct nose-to-brain transport and conventional systemic absorption — the direct-CNS fraction varies widely by drug and formulation.
Intranasal delivery requires no needle, no sterile procedure, and no trained proceduralist — a nasal spray or drop can be self-administered or given by a caregiver in minutes. It avoids first-pass hepatic metabolism (unlike oral dosing) and avoids the infection and bleeding risks of lumbar puncture entirely. This makes it attractive for chronic dosing, emergency/rescue situations (e.g., intranasal naloxone, intranasal benzodiazepines for seizure rescue), and pediatric or needle-averse populations.
Esketamine nasal spray (approved for treatment-resistant depression) and intranasal insulin (studied for Alzheimer's disease and cognition) are prominent examples of the route's clinical and research interest, alongside decades of nose-to-brain research for peptides, growth factors, and small CNS-active molecules.
The olfactory epithelium is a small target (a few square centimeters, high up in the nasal cavity), difficult to reach reliably with standard nasal sprays that mostly deposit drug on the respiratory epithelium instead. Only a small fraction of an administered dose typically reaches the olfactory region and CNS via the direct nerve pathway. Formulation science (droplet size, viscosity enhancers, targeted delivery devices, mucoadhesives, permeation enhancers) is an active area aimed at increasing the direct-transport fraction and reducing dose lost to swallowing, mucociliary clearance, or systemic absorption.
Intrathecal and intranasal delivery do not distribute drug evenly across the CNS — each route has a characteristic regional bias rooted in its anatomical entry point. Understanding these patterns is essential to matching a route to a target lesion rather than assuming any BBB-bypass route reaches the whole brain equally.
Because IT injection deposits drug in the lumbar CSF cistern, concentration is typically highest near the spinal cord and cauda equina immediately after dosing, then declines as a gradient moving rostrally toward the cerebral ventricles and cortex, since cranial spread depends on relatively slow bulk CSF circulation rather than fast diffusion. Higher injected volumes, hyperbaric or hypobaric formulations relative to CSF, and patient positioning (e.g., Trendelenburg) can promote more cranial spread, but even with these measures the lower spinal cord and nerve roots generally receive the highest exposure.
This makes IT delivery especially well suited to conditions localized to the spinal cord and lower CNS: chronic spasticity, lumbar/sacral nerve root pain, and leptomeningeal disease with spinal involvement.
Because IN delivery enters the CNS through the cribriform plate at the roof of the nasal cavity, concentration is typically highest in and around the olfactory bulb and adjacent frontal/forebrain structures, with a steep fall-off in reaching deeper or more caudal regions such as the brainstem or spinal cord (the parallel trigeminal pathway offers some access to brainstem-adjacent structures, but at lower efficiency than the olfactory route reaches the forebrain).
This makes IN delivery attractive for conditions localized to forebrain and olfactory-adjacent regions — for example, research targeting frontal cortex, hippocampal, or limbic pathology — while making it a poor fit for a lesion confined to the lumbar spinal cord.
A common misconception is that any BBB-bypass route delivers drug "to the brain" uniformly. In practice, both IT and IN delivery are regionally biased by their anatomical entry point, and neither reliably achieves uniform whole-neuraxis exposure from a single dose. Achieving broader coverage typically requires strategy-specific adaptations: intraventricular catheters or repeated/continuous IT infusion for broader CSF coverage, or combination dosing strategies and enhanced nasal formulations for IN delivery. Matching the expected distribution gradient to the anatomical location of the target lesion is therefore a central part of route selection.
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Intrathecal (IT) | Spinal cord, cauda equina, lower CNS, leptomeninges | CSF bulk flow from lumbar cistern, gradient declines rostrally | Reliable, high local concentration near injection level |
| Intranasal (IN) | Olfactory bulb, frontal/forebrain, limbic structures | Olfactory & trigeminal nerve transport through cribriform plate | Non-invasive, rapid onset, repeatable self-administration |
Choosing between intrathecal and intranasal delivery — or deciding neither is appropriate — requires weighing three factors together: where in the CNS the target lesion sits, how much invasiveness the clinical situation can tolerate, and whether the specific drug molecule is physically and chemically compatible with each route's transport mechanism.
The single strongest predictor of route suitability is where the pathology sits along the neuraxis. A lesion confined to the lumbar spinal cord or cauda equina is poorly served by intranasal delivery, which barely reaches past the forebrain and olfactory bulb. Conversely, a target concentrated in frontal cortex or olfactory-adjacent limbic structures is well matched to intranasal delivery, while intrathecal injection would need to rely on slow, incomplete rostral CSF spread to reach the same structures from a lumbar injection site.
When a target spans both regions — for example, diffuse leptomeningeal disease — clinicians may combine strategies (e.g., intrathecal dosing plus an intraventricular reservoir) rather than relying on either route alone.
Intrathecal injection requires a proceduralist, sterile conditions, and carries procedural risk (headache, infection, bleeding) — real costs that are easiest to justify when the disease is severe, chronic dosing infrastructure (an implanted pump) is available, or the drug simply has no other path into the CNS. Intranasal delivery, being non-invasive and self- or caregiver-administrable, is preferred whenever it can adequately reach the target: for rescue medications needing rapid, needle-free administration, for chronic outpatient dosing, and for pediatric or needle-averse populations.
The tradeoff is not simply "safer route wins" — if the non-invasive route cannot achieve adequate concentration at the target location, choosing it anyway sacrifices efficacy for convenience.
A well-matched decision aligns anatomical reach with clinical tolerance for invasiveness: use intrathecal delivery when the target is caudal/spinal or when no alternative route can reach a large or polar molecule into the CNS at all; use intranasal delivery when the target is rostral/forebrain and the drug is compatible with mucosal-nerve transport.
Beyond location and invasiveness, the molecule itself constrains route choice. Large biologics, antisense oligonucleotides, and viral vectors are frequently only practical intrathecally, since CSF injection places them directly where they need to act without requiring nerve-mediated axonal transport, which favors smaller, more diffusible or actively transported molecules. Formulation stability also matters: a drug must tolerate direct CSF exposure (pH, osmolality, preservative-free formulation) for IT use, or must be compatible with nasal mucosal absorption enhancers and resist mucociliary clearance for IN use.
In practice, route selection is an integrated decision — location sets the anatomical requirement, invasiveness sets the clinical tolerance, and molecular properties set the feasibility ceiling for each option.