We’re doing something a bit different over the next few weeks. We’ve been asked more than once recently to explain what actually sits behind the regulatory tests we run, so we’re working through our safety testing suite one guideline at a time: where each one came from, what it does, and what it can and can’t tell you. Skin corrosion and eye hazard identification are coming later this month.
Starting with skin irritation, and with the test it replaced.
The Draize skin irritation test was adopted as an OECD guideline in 1981, though the method itself dates back to the 1940s. It involved applying a substance to shaved rabbit skin and scoring erythema and oedema by eye over several days. For decades it was simply how skin irritation was assessed, and the regulatory infrastructure of most countries was built around it.
Test Guideline 439 replaced it. Adopted in July 2010, it was among the first OECD guidelines to fully replace an animal test rather than reduce or refine it, and the route it took to get there is worth understanding, not least because it set the template for how non-animal methods gain regulatory acceptance.
Why reconstructed epidermis
The technical obstacle was never enthusiasm. It was demonstrating that a laboratory-grown tissue could predict human outcomes at least as reliably as a rabbit (although validation to human data would have been the ideal).
Reconstructed human epidermis is built from living human keratinocytes cultured at the air-liquid interface, producing a multi-layered, differentiated tissue with organised basal cells that flatten progressively toward the apical surface, mirroring the structure of in vivo epidermis. It has a functional barrier with an in vivo-like lipid profile, the key differentiation markers match those seen in human skin, and the cells are metabolically and mitotically active. They also release the pro-inflammatory cytokines involved in irritation and inflammation.
That last point matters. The model isn’t a passive membrane standing in for skin. It responds.
The validation route
Several RhE models were developed through the late 1990s and early 2000s. Validation was coordinated by ECVAM, the European Centre for the Validation of Alternative Methods, and ran as a multi-laboratory exercise across several years, assessing reproducibility within and between labs alongside predictive accuracy against existing data.
ECVAM’s Scientific Advisory Committee issued its statement on the scientific validity of in vitro skin irritation testing in November 2008. The performance standards were revised in 2009 to account for the UN GHS classification system, and those updated standards were used to define TG 439 when it was adopted the following year.
The guideline recognises EpiSkin™ as the Validated Reference Method, since it was used to develop the performance standards themselves. EpiDerm™ SIT and SkinEthic™ RHE were validated against those standards and are equally acceptable for regulatory use. LabCyte EPI-MODEL24 was added following peer review in 2011, and the guideline has been updated several times since as further models have been assessed.
This is the mechanism that makes the guideline durable. Rather than naming a fixed list of approved products, TG 439 sets out performance standards that any new RhE model can be measured against. Testing capacity isn’t tied to a single supplier.
What the test actually does
Skin irritation is defined as reversible damage to the skin following exposure to a substance or mixture for up to four hours.
At XCellR8, tissue models are pre-warmed at 37°C and 5% CO₂, medium is replaced, and the tissues are returned to the incubator for a further 18 hours before dosing.
Test items are applied directly to the tissue surface, which is what makes this a reasonable model of real-life exposure. Triplicate models receive 30μl of liquid or 25mg of solid at the apical surface. Sterile water serves as the negative control and 5% SDS as the positive control. Exposure runs for 60 minutes, after which test items and controls are washed off and the models are returned to the incubator for a 42-hour recovery period.
Viability is then measured by MTT conversion. MTT solution is applied and the models incubated for three hours, during which viable cells convert the dye to a blue formazan salt. That metabolite is extracted into isopropanol, and the extract is quantified by absorbance spectrophotometry at 570nm. Readings from treated models are compared against negative controls to calculate percentage viability.
The threshold is 50%. At or below it, the test item is classified as a skin irritant, UN GHS Category 2 or EU classification R38. Above it, non-irritant, with no label required. A defined set of acceptance criteria must be met for the run to be valid.
The limitations worth knowing
TG 439 gives a binary answer: does the test item reduce tissue viability below the irritation threshold, yes or no?
That has two consequences worth understanding before you place a study.
The first is that TG 439 cannot distinguish between GHS Category 1 (corrosive) and Category 2 (irritant) on its own. If viability falls below 50%, you know the item is hazardous, but not which of the two it is. Determining that requires the follow-up test, OECD TG 431, which we’ll cover in a separate post later this month.
The second is that the guideline won’t tell you how mild something is, only that it sits above the irritant threshold. For products where mildness is the selling point rather than a compliance requirement, a different approach is needed, which is where the ET50 modification of the same tissue system comes in.
Where it sits now
The results are accepted for hazard identification under the UN GHS, and the method is appropriate for compliance across REACH, the EU Cosmetics Regulation 1223/2009 and the CLP Regulation 1272/2008. It applies to single substances and to mixtures, including finished products.
We conduct the test to GLP under our MHRA accreditation, which means data can be used at regulatory level and, through the OECD’s Mutual Acceptance of Data arrangement, is recognised in all participating territories rather than needing to be repeated market by market.
There’s a deadline attached to this now. The UK government’s strategy for replacing animals in science, published in November 2025, commits to ending animal testing for skin and eye irritation by the end of 2026. For this endpoint the validated replacement has been sitting on the shelf since 2010, so the strategy is less about developing something new than about closing the gap between what the science can do and what still gets requested.
Fifteen years on from adoption, TG 439 has become unremarkable. That’s the measure of how well it worked.
If you need skin irritation data for a registration or a submission, get in touch and we’ll talk through what your product needs.
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!