Contact allergy to capryloyl salicylic acid: a mechanistic chemistry and structure–activity perspective

Dear Sir, de Groot et al. have reported contact allergy, in 2 patients, to capryloyl salicylic acid (5-CSA, 1, CAS no. 78418-01-6), also known as 2-hydroxy-5octanoylbenzoic acid, which is used as a skin conditioner in various cosmetic products (1). The structure is shown in Fig. 1. Here, we consider whether 5-CSA is likely to be a contact allergen in its own right, or whether the reported allergy is more likely to be to a by-product of its manufacture. The chemical principles of skin sensitization have been, and continue to be, quite extensively investigated. The fundamental mechanistic basis of structure–activity relationships for sensitization is that, for a chemical to sensitize it must be reactive, i.e. able, either as such or after in cutaneo activation, to covalently modify the structure of cutaneous proteins or peptides (2–5). 5-CSA has a very simple structure, and the only plausible way in which it could react covalently with a protein is by the Schiff base mechanism, which involves a protein nucleophile attacking the carbonyl group in the 5-position. Is this carbonyl group reactive enough? The Schiff base mechanism for skin sensitization has been investigated with quantitative mechanistic modelling (QMM) (3) and experimental chemistry (6) approaches. The overall position can be summarized as follows. Simple aldehydes are reactive enough to sensitize. Their potency, at least in the murine local lymph node assay (LLNA), depends on the combined electronegativity of the groups bonded to the carbonyl group, which can be modelled by the sum of their Taft substituent constants Σσ*, together with their hydrophobicity, which can be modeled by logP, where P is the octanol/water partition coefficient (3). Taft σ* values have been determined experimentally and compiled (7) for a large number of chemical groups. Where experimental σ* values are not available, σ* can often be estimated with the methods described by Perrin et al. (7). Simple ketones are not reactive enough to sensitize. This is reflected in the lower values of Σσ* for the ketones RCOR′ than for the aldehydes RCHO – a difference of 0.49 when R′ is methyl, and a larger difference when R′ is a higher alkyl (7). However, ketones activated by another group (e.g. a second carbonyl group in the α position or β position) can be reactive enough to sensitize. Many aromatic aldehydes are non-sensitizers and, in particular, a hydroxyl group in the para position has been found to be deactivating. Thus, Natsch et al. (6) found negligible reactivity towards model peptides for the non-sensitizer p-hydroxybenzaldehyde. Bearing in mind the deactivating effect of the para hydroxyl group, and the generally lower potency of ketones than of aldehydes, it seems unlikely that the contact allergy observed for 5-CSA is the result of 5-CSA sensitizing by the Schiff base mechanism. It is relevant at this point to note that the European Chemical Agency database (8) contains details of two guinea-pig maximization tests with substantially different results (Table 1). This is consistent with the interpretation that 5-CSA is not itself a sensitizer, but can contain an allergenic impurity, at levels that can vary from sample to sample, that is responsible for the sensitization observed. It is often the case that the level of a sensitizer required to elicit a response in an already sensitized individual is substantially lower than the level required to sensitize (9), and bearing this in mind we can envisage a scenario whereby an individual becomes sensitized by exposure to a product containing a relatively high level of a allergenic impurity, and can subsequently react to the same product even with much lower levels of the sensitizing impurity. We now consider how a sensitizing impurity might be present in 5-CSA as a by-product of the manufacturing process. The synthesis of 5-CSA is described in a patent published in 1989 (10). We assume that the manufacturing process is similar to the synthetic method (Fig. 1), which is based on Friedel–Crafts acylation of methyl salicylate by capryloyl chloride and aluminium trichloride to