Does Skin Actually Convert an Ester Into Its Active Form?
Research Team
Partly — and much less completely than the premise assumes. Human skin does carry carboxylesterases, but they are unevenly distributed. In dermatomed human skin, 0.97 ± 0.25 % of an applied methyl salicylate dose crossed as free salicylic acid against 3.59 ± 0.35 % as the intact ester; in heat-separated epidermis the hydrolysed fraction fell to 0.03 ± 0.01 % (Telaprolu et al., 2023).
Published 26 October 2026 · Last reviewed 29 October 2026
What an ester is doing in a topical formulation
Many cosmetic actives are supplied not as the free molecule but as an ester of it. The reason is physical rather than biological: attaching a fatty chain to a polar molecule raises its lipophilicity, and lipophilicity is one of the two properties — together with molecular weight — that govern how readily a molecule partitions into the lipid matrix of the stratum corneum. Ascorbic acid and its lipophilic ester tetrahexyldecyl ascorbate are the textbook pair; retinyl esters and tocopheryl acetate follow the same logic.
The arrangement only works if a second step occurs. The ester has to be cleaved so that the parent molecule is released in the tissue. That cleavage is enzymatic, and the enzymes in question are a definable class.
Definitions used in this article
- Carboxylesterase. An enzyme that hydrolyses a carboxylic ester bond, splitting an ester into its parent acid or alcohol and the corresponding fatty component. In mammals these enzymes are abundant in liver and are also present, at lower activity, in skin.
- Stratum corneum. The outermost, non-living layer of the epidermis, composed of corneocytes embedded in a lipid matrix. It is the principal barrier to percutaneous absorption and contains no appreciable living-cell metabolism.
- Dermatomed skin. A skin membrane cut to a defined thickness with a dermatome, so that it retains stratum corneum, viable epidermis and part of the dermis.
- Heat-separated epidermis. A membrane obtained by immersing skin in warm water and peeling the epidermis from the dermis. The procedure is convenient but exposes the tissue to heat.
- Intrinsic clearance (Vmax/Km). A measure of how fast an enzyme system converts a substrate when the substrate is scarce. It allows two tissues to be compared independently of substrate concentration.
What was measured in human skin
Telaprolu and colleagues ran in-vitro permeation tests on four human skin membrane preparations with methyl salicylate, an ester whose hydrolysis product, salicylic acid, can be quantified separately. After 24 hours they reported the percentage of the applied dose recovered as each species.
| Membrane preparation | Intact ester (% dose) | Free acid (% dose) | Total actives (% dose) | Acid/ester ratio |
|---|---|---|---|---|
| Dermis | 4.44 ± 0.68 | 1.16 ± 0.31 | 5.61 ± 0.74 | 0.26 |
| Dermatomed skin | 3.59 ± 0.35 | 0.97 ± 0.25 | 4.56 ± 0.56 | 0.27 |
| Dispase-separated epidermis | 3.88 ± 0.46 | 0.05 ± 0.004 | 3.93 ± 0.47 | 0.013 |
| Heat-separated epidermis | 2.88 ± 0.79 | 0.03 ± 0.01 | 3.86 ± 0.80 | 0.010 |
Two things are worth separating here. The total quantity of material crossing the membrane varied little between preparations — from 3.86 ± 0.80 % to 5.61 ± 0.74 % of the applied dose, a spread of roughly 1.45-fold, and the authors report no significant difference in total actives between membranes. The proportion that arrived hydrolysed varied by about 39-fold across the same four preparations, from 0.03 ± 0.01 % to 1.16 ± 0.31 %. Transport and activation are therefore not the same variable, and they do not respond to the same things.
The acid/ester ratio in heat-separated epidermis (0.010) was roughly 27 times lower than in dermatomed skin (0.27). Esterase distribution in that study was assessed by α-naphthyl acetate staining — a qualitative localisation method; no activity values in units were reported. One limitation should be stated plainly: the paper's abstract describes the membranes as coming from the same donor while the methods section refers to three different donors, so donor-to-donor variability cannot be resolved from the published text. The percentages above are the paper's own; the acid-as-share-of-permeated-actives figures (approximately 21 % for dermatomed skin and dermis, approximately 1 % for the two epidermal preparations) are ratios computed from Table 1 rather than values the authors state.
What was measured about the enzyme itself
A mechanistic counterpart comes from Imai and colleagues, who applied a caproyl ester of propranolol to excised rat skin and measured both species in the receptor fluid. Over 12 hours across full-thickness skin they recovered 140 ± 35 nmol of the parent drug against 1.7 ± 1.4 nmol of the intact prodrug; steady-state flux was 20 ± 6.2 nmol·h⁻¹ for the parent and 0.40 ± 0.28 nmol·h⁻¹ for the ester. Non-enzymatic hydrolysis accounted for under 0.1 % in 12 h at pH 4.0, so the conversion was enzymatic rather than chemical.
Pre-treating the membrane with 2 mM bis-p-nitrophenyl phosphate, a carboxylesterase inhibitor, cut parent-drug recovery from 140 ± 35 to 21 ± 18 nmol and its flux from 20 ± 6.2 to 3.5 ± 2.6 nmol·h⁻¹, while intact prodrug in the receptor fluid rose about two-fold. Applied to the free drug, the same inhibitor changed nothing (2,600 ± 990 versus 2,600 ± 970 nmol), which is the control that makes the first result interpretable.
| Parameter | Skin S9, PNPA | Skin S9, PNPB | Liver microsomes |
|---|---|---|---|
| Km (μM) | 55.5 ± 8.18 | 97.1 ± 30.2 | not reported here |
| Vmax (nmol·min⁻¹·mg⁻¹ protein) | 188 ± 30.3 | 460 ± 127 | 3,570 ± 810 (PNPA); 4,220 ± 521 (PNPB) |
| Vmax/Km (mL·min⁻¹·mg⁻¹ protein) | 3.39 ± 0.332 | 4.77 ± 0.268 | 21-fold (PNPA) and 24-fold (PNPB) higher |
For the propranolol ester specifically, rat skin S9 gave Km 19.7 ± 5.36 μM and Vmax 9.24 ± 2.95 nmol·min⁻¹·mg⁻¹. Hydrolysis of the model substrates in skin ran around 20-fold slower than in liver microsomes. The same authors estimated from their partitioning data that about 75 % of the applied dose was retained in the stratum corneum, with only about 25 % transferring into viable skin. These are rat-skin figures and should not be read as human values; rodent skin is generally more permeable than human skin, and esterase expression differs between species.
What follows for formulation, and what does not
Three conclusions are supported by the data above, and a fourth is not.
Supported. First, esterification is a transport strategy whose second step is a separate problem. A molecule can be made to partition well and still arrive largely unhydrolysed. Second, the enzymatic capacity lies predominantly in the viable tissue rather than in the stratum corneum, which is consistent with the observation that preparations retaining dermis hydrolysed an order of magnitude more than epidermal sheets. Third, the measurement method is part of the result: heat separation of the epidermis reduced the hydrolysed fraction some 27-fold in the human study, which means that comparisons between published penetration figures obtained with different membranes are not interchangeable.
Not supported. The data do not license the inference that a given ester in a given finished formulation will convert to a particular extent in living human skin. Methyl salicylate and caproyl-propranolol are not ascorbyl or retinyl esters; chain length, steric hindrance around the ester bond and the vehicle all bear on hydrolysis rate, and none of that was varied in these studies. A brand that states a conversion percentage for its own ester, without having measured that ester in that vehicle, is extrapolating.
This is the practical content of formulating according to compounding-pharmacy criteria: the ester is selected for a defined physicochemical reason, the vehicle is specified, and the claim stops where the measurement stops. The precision is citable; the slogan is not.
Scope note
This article describes findings published in the peer-reviewed literature on percutaneous permeation and cutaneous esterase activity. It is written for researchers and healthcare professionals, is not medical advice, and does not constitute a claim about any product. Cosmetic products are not intended to diagnose, treat, cure or prevent disease.
References
- TTS Clinical Scientific Team (2026). Topical Vitamin C: the pH Constraint on L-Ascorbic Acid and the Oxidative Stability of Tetrahexyldecyl Ascorbate, version 1.0. Zenodo. CC BY 4.0. DOI 10.5281/zenodo.22491021 (concept DOI, all versions: 10.5281/zenodo.22491019).
- Telaprolu KC, Grice JE, Mohammed YH, Roberts MS. Human Skin Drug Metabolism: Relationships between Methyl Salicylate Metabolism and Esterase Activities in IVPT Skin Membranes. Metabolites. 2023;13(8):934. DOI 10.3390/metabo13080934. PMID 37623877.
- Imai T, Takase Y, Iwase H, Hashimoto M. Involvement of Carboxylesterase in Hydrolysis of Propranolol Prodrug during Permeation across Rat Skin. Pharmaceutics. 2013;5(3):371–384. DOI 10.3390/pharmaceutics5030371. PMID 24300511.