Eye drop journey across tear film, epithelium, stroma and endothelium into the anterior chamber
The overwhelming majority of an administered eye drop never reaches the cornea at all. Standard droppers dispense 30–56 µL per drop, but the human conjunctival sac can only hold roughly 7–10 µL without overflowing — and blinking pushes the excess out almost instantly. What survives that first flush still has to outrun continuous tear turnover and drainage through the nasolacrimal duct into the nose and throat.
The eye is not a passive reservoir — it is a self-cleaning surface. The moment a drop lands, three defensive reflexes activate almost simultaneously:
• Volume overflow: any liquid beyond the ~7–10 µL sac capacity spills over the lid margin immediately, often before a single blink occurs • Reflex blinking: mechanical and thermal stimulation of trigeminal nerve endings in the cornea and conjunctiva triggers rapid, forceful blinking, which acts like a squeegee spreading and expelling fluid • Reflex tearing: the same stimulation activates the lacrimal gland, diluting the drug and accelerating its clearance through the puncta
Together, these reflexes can remove more than 80% of the instilled dose within the first 15–30 seconds — long before meaningful corneal absorption has begun.
Because the ocular surface aggressively rejects excess volume, larger drops do not proportionally increase drug delivery — most ophthalmic formulations are volume-limited, not dose-limited, in the first seconds after instillation.
Surviving tear fluid is continuously pumped away from the ocular surface through the puncta (small openings at the inner corner of each lid) into the canaliculi, then the lacrimal sac, and finally down the nasolacrimal duct into the nasal cavity.
Normal (unstimulated) tear turnover already replaces roughly 16% of the tear film volume every minute. Reflex tearing after drop instillation accelerates this turnover several-fold, so the precorneal residence time of an unformulated aqueous eye drop is typically only 2–5 minutes before the bulk of it has drained away or been absorbed systemically via the nasal mucosa — a route that also contributes to unwanted systemic side effects.
Because so little of any single drop is retained, ophthalmic formulation science focuses heavily on extending precorneal residence time rather than simply increasing concentration:
• Viscosity enhancers (hydroxypropyl methylcellulose, carbomers) slow drainage by increasing tear film viscosity • Mucoadhesive polymers bind transiently to the mucin layer of the tear film, resisting the blink-driven sweep • In-situ gelling systems are liquid on instillation but gel on contact with the tear film, reducing runoff • Reduced-volume "micro-drop" dispensers cut wasted volume by dispensing closer to the eye's natural capacity, reducing overflow without lowering the effective dose reaching the surface
Once a drug survives precorneal loss, it meets the corneal epithelium — five to six layers of tightly-joined, lipid-rich cells covering the corneal surface. This layer is the single largest barrier to ocular drug absorption for hydrophilic molecules, while posing very little resistance to drugs that are sufficiently lipid-soluble.
Drug molecules can cross the epithelium by two distinct routes:
• Transcellular route: passing directly through the lipid bilayer of epithelial cell membranes. This route dominates for lipophilic (fat-soluble) drugs, which partition readily into the membrane and diffuse cell-to-cell. • Paracellular route: squeezing between adjacent cells, through the narrow intercellular spaces. This route is the main path for small hydrophilic molecules — but it is severely restricted by tight junctions.
The superficial epithelial cells are sealed together by zonula occludens (tight junctions) that form a near-continuous barrier, forcing most polar, water-soluble drugs into a slow, low-capacity paracellular trickle rather than a free flow.
Tight junctions in the outermost epithelial cell layer alone account for the majority of the epithelium's resistance to hydrophilic drug penetration — deeper epithelial layers are comparatively leaky.
The epithelial cell membrane is a phospholipid bilayer, so its "chemical environment" strongly favors molecules with moderate-to-high lipid solubility. Lipophilicity is usually quantified as log P (octanol-water partition coefficient): the higher the log P, the more a molecule prefers a lipid environment over water.
Drugs with log P in the range of roughly 2–3 partition efficiently into the epithelial membrane while still retaining enough aqueous solubility to dissolve in the tear film in the first place — this dual requirement is the seed of the "biphasic solubility" problem explored in Stage 4.
Highly hydrophilic drugs (very low log P), by contrast, essentially bounce off the epithelial surface, unable to partition into the membrane at all, and are largely reliant on the narrow, saturable paracellular route.
Because the epithelium is such a dominant barrier for hydrophilic drugs, several strategies are used to work around it:
• Permeation enhancers (benzalkonium chloride, EDTA) transiently loosen tight junctions to allow more paracellular flux — but at the cost of some epithelial toxicity with chronic use • Prodrug design: attaching a lipophilic ester group to a hydrophilic parent drug (the ester is cleaved by corneal esterases once inside the tissue), letting the prodrug cross the epithelium efficiently before reverting to the active hydrophilic form • Nanocarrier delivery (liposomes, nanoparticles, cyclodextrin complexes) can shuttle hydrophilic payloads across the lipid barrier by presenting a more lipid-compatible outer surface
Having crossed the epithelium, a drug encounters the stroma — a thick, water-rich, collagen-based layer that makes up roughly 90% of total corneal thickness. Here the challenge inverts: the stroma poses little resistance to hydrophilic molecules but actively restricts highly lipophilic ones, which struggle to partition out of the epithelium's lipid environment into this aqueous matrix.
The stroma is composed of regularly-spaced collagen fibrils (mostly type I) organized into roughly 200–250 stacked lamellae, embedded in a hydrated proteoglycan ground substance that is about 78% water by weight. This precise, orderly spacing of collagen fibrils — smaller than the wavelength of visible light — is what gives the healthy cornea its optical transparency, and incidentally also creates an essentially aqueous diffusion pathway for drug molecules.
For a hydrophilic drug that struggled to cross the epithelium's lipid membranes, the stroma is comparatively welcoming: once past the epithelial bottleneck, hydrophilic molecules diffuse relatively freely through the water-filled spaces between collagen fibrils.
Highly lipophilic drugs face the opposite problem here. Having partitioned efficiently into the lipid-rich epithelium, they now resist leaving that favorable lipid environment to enter the aqueous stroma — the same property that made them good at crossing the epithelium works against them at this interface.
This creates a well-documented "reservoir effect": very lipophilic drugs can accumulate and linger at the epithelium-stroma interface, sometimes even providing a slow-release depot effect, but at the cost of markedly slower net movement toward the anterior chamber. In extreme cases, a drug that is excellent at crossing the epithelium but too lipophilic for the stroma will show poor overall corneal permeability despite easily entering the tissue.
Corneal permeability studies classically show a parabolic (inverted-U) relationship between log P and net permeability — permeability rises with lipophilicity up to a point, then falls as the stroma becomes the rate-limiting barrier instead of the epithelium.
A few additional factors shape stromal drug transport:
• Molecular size: the collagen fibril spacing imposes a rough size-exclusion limit; larger macromolecules (biologics, antibody fragments) diffuse far more slowly regardless of solubility • Charge interactions: the negatively charged proteoglycan matrix (keratan sulfate, chondroitin sulfate) can transiently bind cationic drugs, further slowing diffusion • Stromal hydration state: pathological corneal edema (excess stromal water) can actually increase hydrophilic drug diffusion while further disadvantaging lipophilic drugs, since it dilutes and expands the aqueous compartment
The epithelium and stroma impose opposite solubility demands on a drug molecule: one favors fat-soluble molecules, the other favors water-soluble ones. Satisfying both simultaneously — biphasic (dual) solubility — is one of the central and most difficult challenges in designing an effective topical ophthalmic drug.
From the drug's perspective, the cornea behaves like two barriers in series with opposite chemical preferences:
Epithelium (lipid layers) → favors lipophilic (high log P) molecules Stroma (aqueous matrix) → favors hydrophilic (low log P) molecules
A drug that is optimized purely for one barrier tends to fail at the other. Extremely hydrophilic drugs are blocked at the epithelium before they even reach the stroma. Extremely lipophilic drugs cross the epithelium easily but then stall at the epithelium-stroma interface, unable to efficiently enter the water-rich stroma.
Only drugs with a "just right" balance of both properties — biphasic solubility — can traverse the full corneal thickness at a clinically useful rate.
This dual requirement is why topical ophthalmic drug discovery cannot simply optimize for maximum lipophilicity or maximum aqueous solubility — the target is a narrow log P window, roughly 2–3, that compromises between both barriers.
When a candidate drug's natural solubility profile does not fall in the ideal window, several strategies are used to engineer biphasic behavior:
• Prodrugs: a hydrophilic parent drug is temporarily masked with a lipophilic promoiety (commonly an ester) to raise its effective log P for epithelial crossing; corneal esterases then cleave the promoiety inside the tissue, regenerating the hydrophilic active drug for stromal diffusion and pharmacological action • Ion-pairing and complexation: pairing an ionized hydrophilic drug with a lipophilic counter-ion, or complexing with cyclodextrins, can transiently modify the effective partition behavior at each barrier • Permeation enhancers: chemically loosen epithelial tight junctions to let more hydrophilic drug through the paracellular route without needing extreme lipophilicity • Nanocarrier delivery: liposomes, polymeric nanoparticles, and nanomicelles physically ferry drug payloads (of nearly any intrinsic solubility) across the epithelium, sidestepping the biphasic solubility requirement altogether
Consider three hypothetical drug candidates delivered as identical eye drops:
• A highly hydrophilic candidate (very low log P) is stopped almost entirely at the epithelium — even though it would diffuse beautifully through the stroma, it never gets that far • A highly lipophilic candidate (very high log P) sails through the epithelium but accumulates at the epithelium-stroma junction, showing a slow, reservoir-limited trickle into the stroma and beyond • A balanced candidate (log P near 2–3) crosses the epithelium efficiently and continues diffusing through the stroma at a comparable rate — achieving the highest net corneal penetration efficiency of the three, despite not being the "best" molecule at either individual barrier
| Product | Indication | Trial Design | Key Result |
|---|---|---|---|
| Highly hydrophilic | Very low log P | Blocked at epithelial tight junctions; minimal transcellular uptake | Low penetration efficiency |
| Balanced / biphasic | log P ≈ 2–3 | Crosses epithelium transcellularly, then diffuses through stromal water phase | High penetration efficiency |
| Highly lipophilic | Very high log P | Crosses epithelium easily, but reservoir-trapped at epithelium–stroma interface | Moderate, slow penetration |
The innermost corneal layer, the endothelium, is a single sheet of hexagonal, non-regenerating cells. It is markedly more permeable than the epithelium — but it is still the final checkpoint a drug must cross before reaching the aqueous humor of the anterior chamber and producing an intraocular effect.
Unlike the multilayered, tight-junction-sealed epithelium, the endothelium is a single layer of hexagonal cells joined by looser junctional complexes (macula occludens rather than true zonula occludens). This makes the endothelium considerably more permeable to both hydrophilic and lipophilic molecules — it rarely becomes the rate-limiting barrier for a drug that has already survived the epithelium and stroma.
However, "more permeable" does not mean "no barrier." The endothelium still moderates the final passage of drug into the anterior chamber, and it must also maintain its primary physiological role — an active pump-and-leak system that keeps the stroma properly dehydrated for corneal transparency — even while permitting drug passage.
Once past the endothelium, the drug enters the aqueous humor filling the anterior chamber, where it can reach intraocular targets: the iris, ciliary body, trabecular meshwork, and lens. This is the therapeutic destination for topical drugs treating glaucoma, uveitis, and related intraocular conditions.
From here, the drug is subject to aqueous humor turnover (the aqueous humor itself is fully replaced roughly every 100 minutes) and outflow through the trabecular meshwork and uveoscleral pathways — meaning intraocular drug concentration is a transient peak rather than a stable plateau, echoing the same rise-and-decay dynamic seen back at the ocular surface in Stage 1.
Because each of the five layers described in this pathway multiplies additional loss, the fraction of an originally instilled eye drop dose that ultimately reaches the anterior chamber is typically well under 5% — for some poorly-formulated hydrophilic drugs, under 1%.
The corneal penetration pathway — tear film loss, epithelial lipid barrier, stromal aqueous barrier, and endothelial final gate — explains why ophthalmic drug development is so distinct from oral or systemic drug design:
• No single barrier can be optimized in isolation; the biphasic solubility requirement (Stage 4) is a direct consequence of having two opposite-personality barriers in series • Precorneal loss (Stage 1) means that even a perfectly-designed molecule benefits enormously from residence-time-extending formulation • The endothelium's non-regenerative nature means permeation enhancers or preservatives that are safe for the epithelium can still be too toxic for long-term endothelial exposure — endothelial cell density loss is irreversible
Understanding this full multi-layer pathway is why so few oral drug candidates can simply be reformulated as eye drops — most require dedicated ophthalmic-specific optimization from the earliest design stages.