Probing skin photoallergens in reconstructed human epidermis: An EPR spin trapping investigation

Abstract

Photoallergic contact dermatitis is a skin disease caused by combined exposure to photoreactive chemicals and sunlight. Exposure to allergens and the risk of skin sensitization is an essential regulatory issue within the industry. Yet, only few non‐validated assays for photoallergy assessment exist as the pathogenesis is not fully deciphered. Improving such assays and/or developing new ones require an understanding of the chemical mechanisms involved. The first key event in the photosensitization process, namely chemical binding of the photoallergen to endogenous proteins, is thought to proceed via photo‐mediated radicals arising from the photoallergen. Moreover, the mechanism of action of these radicals if formed in the epidermis is not known and far from being unraveled. We present here an original proof‐of‐concept methodology to probe radical generation from allergens in contact with photoexposed skin, using electron paramagnetic resonance and spin trapping in a reconstructed human epidermis model mimicking real‐life exposure scenarios.

We present an original proof‐of‐concept methodology to probe radical generation from allergens in contact with photoexposed skin, using electron paramagnetic resonance and spin trapping in a reconstructed human epidermis model mimicking real‐life exposure scenarios. The aim was to provide a starting point untangling the chemical mechanisms involved in photoallergic contact dermatitis. We focused on benzophenone derivatives being among the main indexed photoallergy culprits, such as the non‐steroidal anti‐inflammatory topical drug ketoprofen, and UV filters octocrylene and oxybenzone contained in cosmetic products.

Abbreviations

AM 1.5 G

air mass 1.5 global

BP

benzophenone

DEPMPO

5‐diethoxyphosphoryl‐5‐methyl‐1‐pyrroline N‐oxide

EC

European Commission

EPR

electron paramagnetic resonance

EPR‐ST

electron paramagnetic resonance spin trapping technique

HEPES

4‐(2‐hydroxyethyl)‐1‐piperazine ethane sulfonic acid

hfccs

hyperfine coupling constants

KP

ketoprofen

OBZ

oxybenzone

OCT

octocrylene

PA

photoallergen

PACD

photoallergic contact dermatitis

REACH

registration, evaluation and authorization of chemicals

RHE

reconstructed human epidermis

ROS

reactive oxygen species

S/N

signal to noise ratio

ST

spin trap

UVA

(long wave) ultraviolet, λ ∈ [315,400] nm

UVB

(medium wave) ultraviolet, λ ∈ [280,315] nm

INTRODUCTION

Photoallergic contact dermatitis (PACD) is a critical skin disease sharing the spotlight with photoirritation. Both are adverse reactions caused by combined exposure to photoreactive chemicals and light (UVA, UVB, and/or visible). ¹ While photoirritation is mainly caused by the immediate generation of reactive oxygen species (ROS) provoking cellular damage, PACD is a well‐structured immunological reaction occurring through (i) an initial sensitization phase acquired after first contact with a photoallergen (PA), followed by (ii) an elicitation clinical phase after new exposure of the sensitized individual to the PA. PACD presents as a highly limiting photodistributed eczematous dermatitis, with sharp demarcation between photoexposed and photoprotected sites (e.g., border of textiles and naturally shaded areas).

Exposure to chemicals and the risk of skin sensitization is a regulatory issue within the industry. Nowadays, it is mandatory to predict the skin sensitization potential of chemicals before their large‐scale use in consumer products to succeed constructive risk assessment. Indeed, information on skin sensitization potential is required by several European regulations, and notably Registration, Evaluation and Authorization of Chemicals (REACH) Regulation, aiming to ensure a high level of protection for human health and the environment. ² Many in vitro and in vivo assays have been developed for photosafety assessment. ³ While photoirritation testing tools have been validated, there are only few non‐validated assays for PACD assessment as the pathogenesis is not yet fully understood. Research needs to focus on understanding the precise photoallergenicity chemical mechanisms to improve the applicability and robustness of current assays. ³ The first key event in the PACD process is chemical. Upon uptake of the PA into the epidermis, excitation by light of the appropriate wavelength gives rise to reactive species which bind to skin proteins through the formation of stable covalent bonds. The PA‐protein conjugate acts then as an antigen triggering the adaptive immune response. ³ The mechanism of PA‐protein binding is believed to proceed via the formation of radicals upon PA irradiation, able then to react with skin amino acids via radical mechanisms. ⁴ , ⁵ , ⁶ Though, the link between such radicals in the epidermis and their mechanism of action is far from being fully unraveled.

We report here a proof‐of‐concept method to probe radical production from PACD culprits using electron paramagnetic resonance (EPR) and spin trapping (ST) in a reconstructed human epidermis (RHE) model. Radicals in biological media are often too short‐lived or of too low concentration to be directly detected by EPR. ST relies on the use of an EPR silent diamagnetic molecule (spin trap), which reacts with transient radicals to form more persistent ones EPR active (spin adduct) whose fingerprints depend on the trapped radicals allowing their detection and identification. ⁷ Our approach derives from our previously developed EPR‐ST methodology identifying radicals derived from skin sensitizers. ⁸ RHE models are histologically similar to human skin regarding morphology, reconstructed from a monolayer culture of normal human keratinocytes. Here, we set the path for a methodology upon daylight illumination to provide a starting point untangling the chemical mechanisms involved in PACD. This work focused on benzophenone (BP) derivatives being among main indexed PACD culprits, ⁹ such non‐steroidal anti‐inflammatory topical drug ketoprofen (KP) and UV filters octocrylene (OCT) and oxybenzone (OBZ) contained in cosmetic products (Figure 1).

FIGURE 1.

Target compounds benzophenone (BP), ketoprofen (KP), oxybenzone (OBZ), and octocrylene (OCT), and schematic description of the methodology for photoexposed reconstructed human epidermis (RHE).

MATERIALS AND METHODS

Materials

BP (CAS 119‐61‐9), KP (CAS 22071‐15‐4), OCT (CAS 6197‐30‐4), OBZ (CAS 131‐57‐7), Dispase II (0.5 U/mg), and reagents for buffer solutions were purchased from Sigma‐Aldrich. Acetone (99.8%) was acquired from Carlo Erba Reagents. 4‐(2‐Hydroxyethyl)‐1‐piperazine ethane sulfonic acid (HEPES) buffer (10 mM, pH 6.8) was prepared with 1.19 g HEPES in 400 mL deionized water, 4 g sodium chloride, and 0.1 g potassium chloride. To attain pH 6.8, sodium hydroxide pellets were added. If the pH went too high, it was lowered back by adding hydrogen chloride until it remained stable to 6.8. Deionized water was added to give a final volume of 500 mL. Spin trap 5‐diethoxyphosphoryl‐5‐methyl‐1‐pyrroline N‐oxide (DEPMPO) was synthetized as reported in the literature. ¹⁰

Equipment

EPR acquisitions were performed at room temperature (295 K) with a continuous‐wave EPR X‐band spectrometer (EMXplus, Bruker Biospin GmbH) equipped with a high‐sensitivity resonator (Bruker Biospin GmbH). The g calibration standard was Bruker “strong pitch” of acknowledged g factor 2.0028. Principal experimental parameters: 10 scan accumulation, microwave power of ca. 5 mW in solution and ca. 10 mW in RHE, modulation amplitude 2 G and sweep time 40 s. EPR spectra were simulated with labmade scripts built on the Easyspin toolbox under the Matlab (Mathworks) environment. ¹¹ An Oriel Sol AAA solar simulator using a Xenon lamp to reach AM 1.5G emission spectrum was used. ¹² All hyperfine coupling constants (hfccs) characteristic of the spin adducts observed in this study are summarized in Table S1.

Methodology

Prior to RHE investigations, control studies were performed in solution, that is, the simplest environment (procedure in Appendix S1). AlternaSkin™ RHE samples (0.6 cm², Cell Alternativ®, Trosly Breuil, France), kept in agarose gel upon receipt, were placed in culture medium (Cell Alternativ®, 1.5 mL/well for a 12‐well plate) and stored 24 h in an incubator (37°C, 5% CO₂, saturated humidity atmosphere). The culture medium was replaced (1.5 mL/well) every 24 h for RHE waiting to be used. For the experimentations, RHE was first put in Dispase II solution (1.5 mL/well, 5 min, basal application) to allow further removal from the polycarbonate backing. Dispase II was prepared in HEPES diluted in water (75:25) to have an enzymatic activity of 2.4 U/mL. RHE was further pre‐treated with DEPMPO (30 μL, 400 mM, dimethylsulfoxide/HEPES 1:1) and incubated during 15 min (37°C, 5% CO₂). After this time, pre‐treated RHE was taken out from the insert and washed with HEPES solution before being placed in an EPR tissue cell equipped with a silica window (Willmad, #ER162TC‐Q). The target compound was then topically applied (30 μL in acetone) in a concentration coherent with the dose per area used in photopatch tests diagnosing PACD ¹³ : 44 mM for BP (2.2 μmol/cm²), 32 mM for KP (1.6 μmol/cm²), 350 mM for OBZ (17.5 μmol/cm²), and 220 mM for OCT (11 μmol/cm²). The EPR tissue cell was closed and sealed, and EPR spectra first recorded prior photo‐exposition (t₀, blank). Furthermore, the tissue cell was alternatively photo‐exposed ex situ for given periods of time (min) and EPR spectra were recorded (schematic procedure shown in Figure 1).

RESULTS AND DISCUSSION

We report a new EPR‐ST approach to study the ability of PAs to generate radical intermediates in the skin. RHE models are a suitable alternative to human or animal tissues, amending ethical issues. To be in conditions close to sun exposure, we use a solar simulator reproducing natural exposure experienced on the Earth's surface (AM 1.5G spectrum). ¹² For a better screening of the radicals formed, studies were first managed in solution to set up the optimal experimental conditions affording a good signal‐to‐noise ratio (S/N) as potential reducing pathways are here downplayed.

KP and OBZ have BP as a common core, hidden in the case of OCT. BP is one of the most used photoactive compounds in organic‐bioorganic chemistry because of its well‐known photochemical properties. ¹⁴ By absorption of a photon, BP is excited from the fundamental singlet state to a triplet state forming the biradical form of the molecule (${\mathrm{BP}}^{••}$) (see Figure 3 top reaction). ¹⁴ , ¹⁵ In our experimental conditions, BP/DEPMPO studies in solution with acquisitions at different light exposure times (from 30 s to accumulated time of 30 min) did not show any EPR signal. In contrast, irradiation of RHE incubated with BP/DEPMPO lead to EPR evidence of photo‐mediated oxygen (O−) centered radicals despite modest S/N (Figure 2A). Nevertheless, the absence of signal in solution for BP is actually biased. Indeed, photo‐generated ROS have been undoubtedly emphasized in solution through a spin‐scavenging complementary study, where an oxidative stress has been probed by the reduction in a stable paramagnetic molecule (Figure S1). Thus, in the ST studies, paramagnetic spin adducts are probably quickly reduced to EPR silent by‐products in the same way by the too‐strong photo‐mediated oxidative stress. The absence of EPR‐ST signal was therefore not surprising.

FIGURE 3.

Formation of BP biradical and proposed photodegradation of KP leading to the formation of reactive radical intermediates.

FIGURE 2.

(A) Benzophenone (BP) (2.2 μmol/cm²)/5‐diethoxyphosphoryl‐5‐methyl‐1‐pyrroline N‐oxide (DEPMPO) (20 μmol/cm²) in the reconstructed human epidermis (RHE) irradiated 5 min: Experimental spectrum (Exp) and overlayed simulation (Sim) agreeing with DEPMPO‐OH spin adduct (a _H = 13.1 G, a _N = 14.0 G, a _P = 47.1 G); (B) Ketoprofen (KP) (1.6 μmol/cm²)/DEPMPO (20 μmol/cm²) in RHE irradiated 5 min: (a) experimental spectrum (Exp) plus simulation (Sim), and deconvolution affording (b) spectra of spin adduct DEPMPO‐C‐radical (a _H = 21.1 G, a _N = 14.6 G, a _P = 47.3 G), and (c) spectra of spin adduct DEPMPO‐O‐radical (a _H = 12.0 G, a _N = 13.8 G, a _P = 46.3 G). All g factors were found to be of ca. 2.0060 ± 0.0005. Control experiments with only BP, KP or DEPMPO and also sole DEPMPO irradiated 5 min did not give any signal.

While ${\mathrm{BP}}^{••}$ is involved in a wide range of applications involving C‐radical intermediates, ¹⁴ we were surprised that no C‐radicals were detected in RHE in our experimental conditions, but ROS. It is well known that skin exposure to sunlight triggers ROS formation, ¹⁶ though in our studies these were not noticeable in control experiments in irradiated RHE without BP. Accordingly, irradiation of RHE in the presence of BP would appear to promote ROS production.

Since the 1980s, KP is known because of the growing number of patients diagnosed photosensitive with severe and long‐lasting clinical symptoms. ¹⁷ KP degrades under the effect of light producing various radical intermediates (Figure 3), potential candidates to induce PACD. ⁴ , ¹⁸ , ¹⁹ Besides, it was shown that products from KP photo‐excitation react with ethanol when used as a solvent, suggesting a radical delocalization within the aromatic moiety. ¹⁸

EPR studies in solution after irradiating KP/DEPMPO gave a spectral signature agreeing with the mixture of at least two spin adducts stressing trapping of carbon (C‐) and O‐centered radicals (Figure S2). Studies in irradiated RHE showed trapping of similar radicals (Figure 2B). RHE pre‐incubated with DEPMPO were treated topically with KP at a dose per unit area (1.6 μmol/cm²) alike to the concentration applied in clinical skin photopatch tests (1% in petrolatum). ¹³ After 5 min irradiation, the fingerprint of a DEPMPO‐C‐radical spin adduct was identified together with the outline of a O‐centered radical spin adduct. Thus, KP generated ROS but also C‐radicals in RHE at relatively short light exposure times. The non‐detection of C‐radicals for BP but for KP suggested that trapped C‐radicals were not carried by the BP moiety of KP. Supporting this, the detection of ROS was compatible with photoionization as a possible photodegradation pathway, leading to the generation of a benzyl‐type C‐radical (Figure 3). Yet, to unravel this hypothesis, synthesis of KP containing a ¹³C‐substitution at the benzyl position might help thanks to the specific EPR fingerprint of the DEPMPO‐¹³C‐radical spin adduct. ²⁰

OBZ and OCT, listed in Annex VI to Regulation (EC) 1223/2009 freshly amended as regards safe concentrations in consumer products, ²¹ are well‐known PAs. ²² OBZ was the most frequent giving positive photopatch tests from sunscreens and cosmetics in the 1970s–1990s. While replaced or its concentration reduced, it is still associated with a high number of positive photoallergic patients most at present related to KP photoallergy. ⁹ , ²³ OCT, introduced in the late 1990s, also triggers PACD in patients formerly sensitized to KP in topical medications. ²⁴ Assays in solution with OBZ/DEPMPO or OCT/DEPMPO showed dissimilar results. As with BP and despite numerous attempts at different light exposure times and concentrations, no spectral signature was obtained for OBZ. Likewise, no radical production in irradiated RHE was detected after the application of OBZ at a dose per unit area (17.5 μmol/cm²) alike to skin photopatch tests concentration (10% in petrolatum). ¹³ These results were in agreement with EPR spin‐scavenging investigations, where no photo‐generated ROS were highlighted in contrast to BP (Figure S1). In the case of OCT in solution, an EPR signature was detected but after a longer light exposure compared with the other molecules studied. A signal appeared after 15 min and it was after 20 min of illumination that it was clearly possible to identify at least two spin adducts (Figures S3 and S4). Overall, EPR fingerprints found in RHE following topical application of OCT at a dose per unit area (11 μmol/cm²), similar to the concentration applied in photopatch tests (10% in petrolatum), ¹³ were compatible with those obtained in solution, pointing to similar O‐ and C‐radicals (Figure S5). Although poor S/N, analysis of the signal was attempted relying on the signature found beforehand in the solution (Figures S4 and S5).

As KP photosensitivity leads to OBZ and OCT photocontact allergy, we expected insights for these cross‐reactions given in terms of the radical species formed in irradiated RHE. Yet, the obtained results proved to be very eclectic. OBZ is a BP with a methoxy substituent and a phenol‐type hydroxyl function, electron‐donating groups that would favor ${\mathrm{BP}}^{••}$. Unexpectedly, no radical generation was observed. This was surprising when compared to the production of O‐radicals by BP in irradiated RHE. On the contrary, OCT does not have a BP moiety per se but can lead to BP by oxidative cleavage of the double bond or by hydrolysis and retro‐aldol condensation. ²⁵ In this case, similarly to KP, the generation of O‐ and C‐radicals within irradiated RHE exposed to OCT was demonstrated but after longer exposure to light (>15 min) compared with KP. Although we believe radicals detected in RHE were issued from the target compounds as control experiments were EPR silent, the results do not allow to elucidate their involvement in the cross‐reactions observed at photoptach testing. If the major KP photodegradation product is considered to be 3‐ethylbenzophenone, ²⁶ proof of benzyl radical formation (Figure 3), ⁴ the observed C‐radical from OCT should be different as hidden BP released after hydrolysis is non‐substituted. Moreover, no radicals were observed from OBZ. If at first sight, KP, OBZ, and OCT appear to have structural similarities based on the BP core, more careful studies must be conducted to precisely identify the origin and the nature of the C‐radicals observed.

An EPR‐ST methodology is here nevertheless presented as a proof‐of‐concept to study the formation of radicals derived from xenobiotics in photoexposed skin. Even if photoreactivity of the PAs considered here seems to be much more complex than what has been described so far and requires further in‐depth studies, the methodology is promising to evaluate the behavior of sensitizing molecules whose haptenization mechanisms are not well understood by exposure to the “Sun.” Photo‐exposure studies conducted in RHE open thus the path to understand photo‐induced xenobiotic‐derived unresolved dermatological issues.

Supporting information

Appendix S1

Port‐Lougarre Y, Voegeli G, Vileno B, Giménez‐Arnau E. Probing skin photoallergens in reconstructed human epidermis: An EPR spin trapping investigation. Photochem Photobiol. 2025;101:275‐281. doi: 10.1111/php.14010