This is the third article in our run through our regulatory safety testing suite, and after skin irritation and skin corrosion, the eye is where the approach behind both ran into difficulty.
Skin irritation and skin corrosion both use reconstructed human tissue, MTT viability and a threshold. Apply the substance, measure what survives, classify against a cut-off. The eye uses the same building blocks and it wasn’t enough, because eye damage is about depth rather than surface, and the classification depends on whether that damage is reversible.
The first in vitro guideline for the eye, TG 492, got part of the way there. TG 492B is the one that finished the job.
Why the eye was harder than the skin
The Draize eye test, OECD Test Guideline 405, has been the standard for decades and became the image most people associate with animal testing in cosmetics: a substance instilled into a rabbit’s eye and the resulting damage scored over days. Replacing it was an obvious priority, and it took considerably longer than replacing the equivalent skin test.
Skin irritation is largely a surface event, which is why reconstructed human epidermis works so well for TG 439. Eye damage is about depth. A substance that stays in the outermost corneal layers causes reversible irritation. One that penetrates further can cause permanent damage to vision. The regulatory categories reflect that distinction: Category 1 for serious, irreversible eye damage, Category 2 for irritation that resolves within 21 days, and No Category for everything below that threshold.
An in vitro method therefore has to do more than detect damage. It has to distinguish between kinds of damage, and predict reversibility from a tissue that only models the corneal epithelium.
TG 492 and the partial answer
TG 492 uses reconstructed human cornea-like epithelium and is based on the depth of injury model of Maurer and Jester: the degree of irritation correlates with how far a substance penetrates the cell layers and the resulting loss of viability. Cell viability is measured by MTT conversion, the same principle used in the skin irritation guideline.
The threshold is 60%. Above it, the substance does not require classification, and that result is definitive with no further testing needed. At or below it, the guideline is explicit that no prediction can be made.
That’s the limitation. A positive result under TG 492 isn’t a classification. It’s an unresolved question, and the substance goes on to further testing, either another in vitro method or, depending on the regulatory framework, back to the rabbit. The guideline was designed to sit inside a sequential testing strategy rather than to answer the question alone.
For a great many products this is perfectly adequate. If you’re reasonably confident a formulation is benign and you need that confirmed for a registration, TG 492 confirms it. The difficulty arises when the result comes back the other way, because you’ve spent time and sample and you still don’t have a classification.
TG 492B and the time-to-toxicity approach
TG 492B takes a different route. Rather than measuring how much viability is lost at a fixed exposure, it measures how long the tissue takes to lose it.
The rationale is that severity correlates with time to toxicity. A substance causing serious eye damage will reduce viability quickly. An irritant takes longer. Something benign may not reach the threshold at all. By running multiple exposure durations, the method produces a profile rather than a single reading, and that profile maps onto the full GHS classification.
The result is that TG 492B discriminates between all three categories in one test: Category 1, Category 2, and No Category. It is recommended as a full replacement for the in vivo Draize eye test, a status TG 492 was never able to claim. There are separate protocols for liquids and solids, both using reconstructed human cornea-like epithelium with MTT as the endpoint.
Which one to use
Both are valid regulatory methods and both remain in the OECD catalogue. The choice comes down to what you need the data to do.
TG 492 makes sense when you have strong grounds to expect a No Category outcome and the study is confirmatory. It is the older, more widely recognised method and some submissions or clients will specifically request it.
TG 492B makes sense in most other situations, and particularly when you don’t already know the answer. One study, one sample, one classification, whatever the result turns out to be. For a novel ingredient, or a formulation where the outcome is genuinely uncertain, that resolves in a single step what TG 492 might leave hanging.
If you’re not sure which fits, that’s a conversation worth having before the study is placed rather than after.
Where this sits now
Both guidelines are accepted for hazard identification under the UN GHS and are appropriate for compliance with REACH and the EU Cosmetics Regulation 1223/2009. Data generated to GLP is recognised across all territories participating in the OECD’s Mutual Acceptance of Data arrangement.
There’s a deadline attached now, too. The UK government’s strategy for replacing animals in science commits to ending animal testing for skin and eye irritation by the end of 2026. With TG 492B available as a full replacement for the Draize eye test, the scientific obstacle to that commitment has already been cleared. What remains is adoption.
We run both methods to GLP under our MHRA accreditation, in fully animal-product-free conditions. If you need eye irritation data for a registration or submission, get in touch and we’ll work out which route suits your product.
Email info@x-cellr8.com or call +44 (0)1925 607 134.
———
Contact us if you’d like to learn how you can support our mission, become a corporate sponsor or donate to our crowdfunder.
Email info@x-cellr8.com
Call +44 (0)1925 607 134.
Follow Us!