Replacing Animal Testing for Skin: The TG 439 and TG 431 Route End to End

The problem arrives in one of three forms. A regulator or a customer will no longer accept the rabbit data. An ethics committee has closed the route. Or a market, cosmetics in the European Union being the obvious one, simply does not permit it and never did for your product class.

In every case the question is the same: what do I buy instead, how many of them, from whom, and how long does it take. This page answers that, and it starts with the single fact that determines the whole project.

You cannot build your own reconstructed human epidermis and call the result TG 439 data. The guideline names specific validated test methods. Using one of them is what buys you Mutual Acceptance of Data. This is the difference between a dossier that is accepted and one that is not. Everything else on this page follows from it.

Step 1, Work out which guideline answers your question

Three different regulatory questions are routinely collapsed into “skin testing”, and they need different guidelines, different protocols and different tissue counts.

Your question	Guideline	Endpoint	Exposure	Post-treatment	What it cannot tell you
Is it an irritant?	TG 439	Reversible damage; UN GHS Category 2	One exposure, 15-60 min depending on method	42 hours at 37 °C	Whether it is Category 1 or 2 when viability is at or below 50 percent. It also cannot classify optional Category 3 (mild irritants)
Is it corrosive?	TG 431	Irreversible damage; UN GHS Category 1, with optional sub-categorisation	Two or three timepoints, 3 min and 60 min for EpiDerm SCT, SkinEthic RHE, epiCS and LabCyte	No 42-hour equivalent; viability read after the exposure sequence	It cannot separate Sub-category 1B from 1C. It reports a combination
Is it a skin sensitiser?	TG 442D and the wider defined-approach framework	Keratinocyte activation, one key event of several	Method-specific	Method-specific	Sensitisation is not answered by a single assay. This page does not cover it

Run them in the right order. A TG 439 result at or below 50 percent viability tells you the chemical classifies but not into which category. Corrosion data from TG 431, TG 430 or TG 435 is what resolves that. If you have any reason to suspect corrosivity, run TG 431 first, a corrosive call there means you never needed the irritation run at all, and you have saved the tissues.

The detailed guideline requirements, barrier function, MTT, controls, run acceptance criteria and the prediction models, are set out at length in what a compliant RhE model must demonstrate. This page is the sourcing side of the same problem.

Step 2, Choose a model, and understand what “compliant” means

TG 439 lists seven validated test methods, six of which remain commercially available. EpiSkin was removed in the 2024/25 update because it is no longer commercially available from December 2024, although data generated with it remains covered by Mutual Acceptance of Data.

Model	Developer	In TG 439	In TG 431	Barrier benchmark and acceptance window	Positive control criterion	Publishes a price
EpiDerm SIT (EPI-200)	MatTek	Validated Reference Method	Yes (EpiDerm SCT)	1% Triton X-100, ET-50 4.0-8.7 hr	< 20% viability	No, request a quote
SkinEthic RHE	EPISKIN	Validated	Yes	1% Triton X-100, ET-50 4.0-10.0 hr	< 40% viability	No, specifications published, no pricing
epiCS	,	Validated (GD 220 route, peer review 2018)	Yes	1% Triton X-100, ET-50 2.0-7.0 hr	< 20% viability	Not established in our survey
Skin+	,	Validated (GD 220 route, peer review 2018)	Not in the TG 431 method set we reviewed	1% Triton X-100, ET-50 4.0-9.0 hr	< 40% viability	Not established in our survey
LabCyte EPI-MODEL24 SIT	,	Validated	Yes (SCT version)	SDS 18 hr, IC50 1.4-4.0 mg/mL	< 40% viability	Not established in our survey
KeraSkin SIT	,	Validated (peer review 2020)	Not in the TG 431 method set we reviewed	SDS 18 hr, IC50 1.5-4.8 mg/mL	≤ 40% viability	Not established in our survey
EpiSkin (SM)	,	Removed, not commercially available from Dec 2024	Yes	SDS 18 hr, IC50 1.0-3.0 mg/mL	≤ 20% viability	Not applicable

Two things to take from that table.

The tighter positive control criterion is a throughput cost, not a quality bonus. EpiDerm and epiCS require the 5% SDS positive control to fall below 20 percent viability, against 40 percent for the others. A marginal positive control invalidates the whole run. Tighter is better data and more repeat runs, and repeat runs are tissues.

“Compliant” and “validated” are different words. A supplier describing its own model as “TG 439 compliant” is making a claim about its model’s performance. A model named in the guideline as a validated reference method has had that claim assessed externally. Mutual Acceptance of Data is guaranteed only for methods validated against the Performance Standards and reviewed and adopted by OECD. A model that meets the performance standards on paper without OECD adoption does not carry MAD, and that gap is exactly where a procurement decision goes quietly wrong.

If your endpoint is a REACH dossier, a cosmetics safety assessment or a medical device submission, that distinction is the entire reason to pay a premium for an established model. If your work is exploratory method development, it may not matter at all.

EPiTRI was validated against the TG 439 performance standards in a published four-laboratory study reporting 96 percent sensitivity, 70 percent specificity and 83 percent accuracy overall. Confirm its listing in the guideline version your regulator accepts, and confirm supply in your region, before specifying it. The head-to-head detail is in EpiDerm vs SkinEthic RHE vs EPiTRI.

Step 3, Count the tissues, because tissues are the cost

This is the arithmetic almost nobody does before requesting a quote, and it is the reason first-time budgets come in low by a factor of two or more.

MODELLED tissue counts. The replicate and control requirements below are the guideline’s. The totals are our arithmetic from those requirements, for a study design we have assumed. Your study plan may require more. This is not a quotation.

A TG 439 irritation run, per chemical

Component	Requirement	Tissues
Test chemical	At least 3 replicate tissues per chemical per run	3
Negative control (water or PBS)	Concurrent, at least 3 replicates	3
Positive control (5% aqueous SDS)	Concurrent, at least 3 replicates	3
Subtotal, one chemical, one run, no interference		9
Second chemical in the same run	Controls are shared across the run	+3
Third chemical in the same run		+3
Three chemicals, one run		15

The controls are the thing to notice. A single chemical costs nine tissues; three chemicals cost fifteen, because the six control tissues are shared. Batching chemicals into one run is the largest single cost lever available on this assay, and it is free. It also constrains your schedule, because everything in the batch has to be ready at the same time.

The three multipliers that inflate that number

Multiplier	When it applies	Extra tissues
Direct MTT reducer	The chemical turns fresh MTT solution blue or purple. Requires the NSMTT control on killed tissues	+3 per chemical
Coloured chemical	Coloured, or becomes coloured during treatment. Requires the NSCliving control on living tissues	+3 per chemical
Both, and it binds tissue	Requires NSCkilled as well, and a three-term correction	+3 more per chemical
Borderline result	Non-concordant replicates, or mean viability at 50 ± 5 percent. A second run should be considered, a third if the first two disagree	+3 to +6 per chemical, plus shared controls if run separately

A coloured, MTT-reducing, tissue-binding chemical that lands borderline can consume more than twice the tissues of the clean case. If you know your chemical is coloured or reducing, say so in the quote request and budget the controls up front. The guideline’s escape route for strongly coloured chemicals is HPLC/UPLC-spectrophotometry, which removes the need for NSCliving and NSCkilled entirely, but the system has to be qualified first, which is a method development project, not a line item.

A TG 431 corrosion run, per chemical

TG 431 uses at least two replicates per exposure time rather than three per chemical, but it runs two or three timepoints.

Component	Requirement	Tissues
Test chemical, 3-minute exposure	At least 2 replicates	2
Test chemical, 60-minute exposure	At least 2 replicates	2
Negative control, both timepoints	Concurrent	4
Positive control (8N KOH), both timepoints	Concurrent	4
Subtotal, one chemical, one run		12
Each additional chemical in the same run	Controls shared	+4

Again the controls dominate at low chemical counts, and again batching is the lever.

Proficiency, before any of this counts

Before routine use of a method under the guideline, a laboratory must demonstrate technical proficiency on ten specified proficiency substances, five UN GHS No Category and five Category 2. At three replicates plus shared controls per run, that is a substantial one-off tissue commitment before a single one of your own chemicals is tested.

Two of the ten are flagged in the guideline itself: 1-methyl-3-phenyl-1-piperazine and 1-bromohexane can give variable results between laboratories depending on the chemical supplier. If your proficiency run wobbles on those two specifically, that is a known effect rather than a technique failure.

Budget the proficiency exercise as a distinct project with its own tissue order and its own delivery week. It is the most commonly omitted line in a first-year plan.

Step 4, The prices that exist, and the ones that do not

We checked every route we could find on 2026-09-01. Here is the honest position.

Supplier	Product	Format	Published price	What it is
MatTek (Sartorius)	EpiDerm EPI-200	24 tissues per kit; 9 mm and 22 mm inserts; 96-well HTS	None, “Request a Quote”	The validated reference method. No price at the vendor or at the distributors we checked
EPISKIN	SkinEthic RHE	Multiple maturities and surfaces	None, specifications published, no pricing	Validated. No price
Phenion (Henkel)	FT SM-1 full-thickness skin model, 1.4 cm	Single tissue	EUR 140	Full-thickness, not an RhE. Not a TG 439 substitute and Phenion does not present it as one
Phenion (Henkel)	FT LONG-LIFE-1, 1.4 cm	Single tissue	EUR 190	Culturable at least 50 days after arrival. A different experimental window entirely
Phenion (Henkel)	FT LARGE-1, 3.1 cm	Single tissue	EUR 235	Larger format
Phenion (Henkel)	ALI CM-250 air-liquid interface medium	250 mL	EUR 75	Ordered separately. Phenion states 250 mL cultures 6 standard FT models in separate dishes for up to 10 days
Phenion (Henkel)	Phenol-red-free, hydrocortisone-free, antibiotic-free medium variants	250 mL	EUR 110 each	The variant is a price difference, not a preference
Phenion (Henkel)	FCA-15 Franz cell adapter	Each	EUR 150	For permeation work
NeoBioPharma	NeoEpidermis NEP3D1-6	Catalogue implies 6-well	$1,900.00	Supplier claims TG 439/431 compliance. We could not independently verify validation status
NeoBioPharma	NeoEpidermis NEP3D1-12	Catalogue implies 12-well	from $3,500.00	Page describes size as custom and the figure as a starting price per unit, so we do not assert a per-tissue number

All figures captured 2026-09-01. Phenion’s published volume discounts are 5 percent above EUR 3,000, 7 percent above EUR 6,000 and 9 percent above EUR 10,000, which is a real and unusually transparent tiering structure worth designing an order around.

We cannot tell you what EpiDerm costs, and no honest source can. Three structural reasons, all usable in a negotiation:

The product is a slot on a production line, not a unit. Tissue lots are produced weekly with lot numbers assigned per week. Publishing a price would expose capacity allocation.
Academic and commercial pricing genuinely differ. One distributor states explicitly on its product page to contact them for academic pricing. Publishing one number breaks the other.
Format variety. 9 mm inserts, 22 mm inserts and 96-well HTS are different products at different price points, plus four media variants.

The practical implication: you are expected to ask, and the first number you are given is not necessarily the only number available. State your institution type, your annual tissue volume and your format in the first email. The full price survey is in the RhE price comparison.

Step 5, Plan against the production calendar, not against a lead time

This is the scheduling fact that catches every first-time buyer, and it cannot be bought past.

Living tissue models ship on a fixed weekly cadence. One published example produces and ships every Monday, delivering Tuesday morning by priority courier in the US and Tuesday to Thursday internationally, with a four-day shelf life that explicitly includes transit.

Read that again in operational terms:

You cannot expedite. There is no premium that produces a tissue on Wednesday.
The usable window is days, not weeks. Tissue that arrives Tuesday and is not used by the end of the week is gone.
Your assay date is set by their production date, working backwards through transit.
A missed delivery is a missed week, not a missed day.

Consequences for how you plan:

Week	Activity	The failure mode
−4 to −6	Request quotes. State model, format, kit size, media variant, delivery week and destination	Asking for “an RhE model” and receiving three non-comparable quotes
−3	Place the order against a named production week. Confirm the media variant lead time, variants have agents removed at least 3 days before shipment, which affects the schedule as well as the price	Ordering a phenol-red-free variant at the last minute and missing the production week
−1	Confirm the ship date and the delivery day in writing. Confirm someone is in the building	A Tuesday delivery into a bank holiday
0, delivery day	Receive. Inspect. Verify barrier function on receipt as the producer specifies. Record the batch number and the batch ET-50 or IC50	Skipping receipt verification, which the guideline recommends specifically because transit degrades barrier function
0 to +4 days	Run the assay inside the shelf life	Discovering on day five that the tissues are out of window
+2 days from exposure	The 42-hour post-treatment incubation completes; read MTT	Reading early, which gives you a different assay

The shelf life is the schedule. Everything else is negotiable; four days including transit is not. See production calendars and shelf life for how this behaves across the tissue-model category, and shipping live cells and tissue for the format and dangerous-goods treatment.

Step 6, What your supplier owes you with every delivery

The guideline is explicit that the model may only be used if the developer demonstrates that each batch meets defined production release criteria for viability, barrier function and morphology, and that these data should be provided to users so they can be included in the test report.

So the batch data is a required input to your own report, not a courtesy. Ask for all of it, every time:

The batch number of the tissues shipped.
The measured ET-50 or IC50 for that batch, with the benchmark chemical and exposure stated, and the acceptance window it sits in.
The batch viability release data.
The morphology documentation for the model, showing the multi-layered stratum corneum.
Confirmation of freedom from bacteria, viruses, mycoplasma and fungi.
The model SOP for the version you are running.
The producer’s instructions for verifying barrier function on receipt.

Two lots at opposite ends of a 4.0-to-8.7-hour ET-50 window are not equivalent tissues for a borderline chemical. A statement that a batch passed is not the same as the measured value, and you are entitled to the measured value.

Where the route does not work

Worth knowing before you commit a budget, because these are stated in the guideline itself.

Gases and aerosols. Not assessed in any validation study; the guideline does not permit testing them.
UN GHS Category 3, mild irritants. Cannot be classified by TG 439.
Category 1 versus Category 2. Cannot be resolved by TG 439. You need corrosion data.
Agrochemical formulations. The guideline cites a study of 65 agrochemical formulations giving overall accuracy of 54 percent, sensitivity 44 percent and specificity 60 percent, and states that this indicates a lack of applicability of the RhE-based test for agrochemical formulations. That is an unusually direct statement for a test guideline and it should be heeded rather than argued with.
Mixtures generally. Applicable, but the guideline advises considering upfront whether the result will be scientifically meaningful.
Sub-category 1B versus 1C. TG 431 reports a combination and cannot separate them. It is also candid that roughly 29 to 33 percent of Sub-category 1A calls across the four main methods may actually be 1B or 1C, that is, over-classifications.

What it does cover: solids, liquids, semi-solids and waxes; aqueous and non-aqueous liquids; water-soluble and insoluble solids. Solids should be ground to a fine powder where possible, and the epidermis surface moistened before application.

Adjacent tissue routes, when skin is not the endpoint

If the animal test you are replacing is not the skin test, the same commercial structure applies in the adjacent categories, with the same absence of published pricing.

Eye irritation uses reconstructed corneal-like epithelium. We checked all four TG 492 model developers and their distributors on 2026-09-01: every route is quote only and we do not estimate one. See /catalog/corneal-ocular-model.
Oral, gingival and mucosal models are the same picture, no supplier in the category publishes a price. See oral, gingival and mucosal models.
Full-thickness skin, dermis plus epidermis, is where the published prices actually are, as the table above shows, and it answers permeation and genotoxicity questions rather than TG 439 irritation. See /catalog/full-thickness-skin-model.

The specification to send

Send the identical document to every supplier. Quotes for RhE tissue are non-comparable by default because the format, kit size and media variant are all separately priced.

Model and regulatory intent

Model, by name and developer: [e.g. EpiDerm SIT EPI-200 / SkinEthic RHE]. Do not ask for “an RhE model”
Guideline: TG 439 / TG 431 / both
Regulatory purpose: REACH dossier / cosmetics safety assessment / medical device ISO 10993-23 / internal screening
Whether you require the model to be a validated reference method named in the guideline, or whether a compliant model is acceptable

Format and quantity

Format: 9 mm insert / 22 mm insert / 96-well HTS. These are different prices
Kit size and number of kits. State what a partial or larger order does to the unit price
Tissue count derived from your run design, stated explicitly: chemicals, replicates, controls, interference controls, and an allowance for one repeat run
Whether the proficiency exercise on the ten guideline substances is included in this order or ordered separately

Media and consumables

Media variant: standard / phenol red-free / antibiotic-free / anti-fungal-free / hydrocortisone-free, and the lead time effect of the variant
Maintenance medium volume for the protocol duration, quoted as a separate line
Any plasticware, filters or spacers included or excluded

Schedule

Required delivery week, against the supplier’s named production cadence
Delivery day, and confirmation that a Friday delivery will not be used
Destination, because international transit consumes the shelf life
The stated shelf life, and whether it includes transit

Documentation, as a contractual deliverable

Batch number, batch ET-50 or IC50 with benchmark and exposure, batch viability, morphology documentation, contamination clearance, model SOP, receipt-verification instructions

Commercial

Institution class (academic, non-profit, commercial), state it, because tiering is real
Annual volume estimate, state it, because tiering is real
Remedy for a batch outside its release window on arrival, and the claim window in days
Repeat-order price at your stated annual volume

State your institution type and annual volume in the first email. In a market where nobody publishes a price, those two facts are what determines the number you are given.

Where we fit

We put that specification to named suppliers, normalise the responses onto a per-tissue and per-run basis, and check the production week against your assay date before you commit, because on this category the schedule risk is not lead time, it is a four-day shelf life against a fixed weekly cadence.

Sometimes the correct answer is a full-thickness model with a published price rather than an RhE. The corrosion test usually belongs first, to save the irritation tissues. And a supplier’s claim of compliance is a different thing from the validation your regulator will ask for.

Replacing animal skin testing: the OECD TG 439 and TG 431 route