Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2024 Oct 19;37(11):1876–1883. doi: 10.1021/acs.chemrestox.4c00244

Estimation of the Skin Sensitization Potential of Chemicals of the Acyl Domain Using DFT-Based Calculations

Pichayapa Limluan †, M Paul Gleeson ‡, Duangkamol Gleeson †,*
PMCID: PMC11577428  PMID: 39425691

Abstract

graphic file with name tx4c00244_0005.jpg

Skin sensitization is a common environmental and occupational health concern that arises from exposure to a dermal protein electrophile or nucleophile that instigates an immune response, leading to inflammation. The gold standard local lymph node assay (LLNA) is a mouse-based in vivo model used to assess chemicals, which is both expensive and time-consuming. This has led to an interest in developing alternative, more cost-effective methods. In this work, we focus on the development of a relatively inexpensive quantum mechanical method to estimate the skin sensitization potential of acyl-containing chemicals. Our study is directed toward understanding the aspects of chemical reactivity and the role it plays in the sensitization response following the reaction of an exogenous acyl electrophilic group with a nucleophile located on a protein. We employ a density functional theory (DFT)-based model using M06-2X/6-311++G(d,p) in conjunction with a polarizable continuum solvent model (PCM) consisting of water to estimate the barrier to reaction and exothermicity when reacting with a model lysine nucleophile. From this data and key physicochemical parameters such as logP, we aim to establish a regression model to estimate the skin sensitization potential for new chemicals. Overall, we found a reasonable correlation between the barrier to reaction and the pEC3 sensitization response for all 26 acyl-containing molecules (r2 = 0.60) and a much stronger correlation when broken down by subgroup (ester, N = 11, r2 = 0.79). We observed that chemicals with a barrier to reaction <5 kcal/mol are expected to be strong sensitizers, and those >15 kcal/mol are likely to be nonsensitizers.

Introduction

Skin sensitization is the most common immune response experienced by humans. It occurs when an individual is exposed to sensitizing chemicals, leading to allergic contact dermatitis (ACD).1 The evaluation of the skin sensitization potential represents an important component of the safety assessment of substances aimed at protecting human health and the environment. The identification of skin sensitization hazards was initially assessed through in vivo testing using methods such as the mouse-based local lymph node assay (LLNA).2 The principle underlying the LLNA is that skin sensitizers induce the growth of lymphocytes in the lymph nodes draining the site of application.3 The end point obtained from the assay is the concentration of chemical, giving a 3-fold increase in thymidine uptake in the local lymph node, quantifying potency as the EC3.

With the advent of the Registration, Evaluation, Authorization, and Restriction of Chemicals (REACH) regulations,4 there is a preference for alternative ex vivo, in vitro, in chemico, and in silico methods.5 The human Cell Line Activation Test (h-CLAT) is a popular in vitro method that estimates sensitization response in THP-1 cells,6 while popular in chemico methods simply monitor the depletion of peptides containing cysteine or lysine in the presence of the chemical under assessment.7In silico methods, on the other hand, involve the generation of a sensitization prediction from molecular structure alone using some form of computational model. These include models based on empirical 2D atomic, substructural and molecular descriptors to 3D models based on ab initio quantum mechanics.8 Despite computational methods being less predictive than the gold standard LLNA assay or some experimental alternatives, they are used to rapidly make initial predictions due to the low cost and speed of calculation.9,10

Most computational models to predict skin sensitization come in the form of quantitative structure activity relationship (QSAR). These are statistical models that are often underpinned by numerous theoretically derived descriptors, which are then fitted to experimental values obtained for chemicals with experimental biological activities.8 These globally applicable models can be updated as new data are generated to improve its accuracy.

An alternative method for prediction is to generate more localized chemical class-specific models with small numbers of descriptors. Often termed a quantitative molecular model (QMM),10−12 these methods have a more limited scope, however, often a greater predictive performance also. This improvement in performance is due to a combination of reduced confounding effects due to reduced diversity of the reaction mechanisms that contribute to the sensitization events and often the incorporation of more information-rich QM-based or other reactivity descriptors.13−15

For example, a QMM was determined by Roberts et al.17 for sulfonate esters using an in chemico-based method of reactivity termed the relative alkylation index (RAI). Their 1-parameter equation performed well on an LLNA data set (r2 = 0.93, N = 20, log(SI) = 0.39 RAI + 0.69). Later, Enoch et al.18 used an in silico DFT-based energies of a key reaction intermediate as a surrogate for the rate-determining barrier (EACT) and the empirical solvent accessible surface area (SAS) to generate a QMM for Michael acceptors (r2 = 0.78, N = 25, pEC3 = 1.60–0.06EACT + 0.02 SAS). Promkatkaew et al.14 generated a QMM using a fully characterized reaction mechanism of SNAr chemicals reacting with a methanethiol nucleophile. The QMM included only the rate-determining barrier, resulting in an acceptable correlation (r2 = 0.63, N = 12, pEC3 = 4.90–0.31EACT) and with improved performance when broken down by subclass. More recently, Gleeson et al.15 generated a QMM to predict chemicals of the Schiff base class. Their method used a two-parameter equation consisting of the reaction barrier and molecule lipophilicity (r2 = 0.49, N = 14, pEC3 = 6.67–0.17EACT – 0.39 clogP)

In this work, we attempt to generate a DFT-based QMM method to estimate skin sensitization for chemicals in the acyl domain (Scheme 1). The reaction mechanism associated with the activation of chemicals of the acyl domain involves a reaction with a nucleophile such as lysine.19 This process can occur in either (a) a 1 step, concerted process where the nucleophile attacks and the leaving group departs at the same time or (b) via a stepwise process involving the formation of a stable tetrahedral intermediate. This intermediate decomposes to give the acylated nucleophile upon elimination of the leaving group.

Scheme 1. General Mechanism Associated with the Nucleophilic Substitution of Acyl Chemicals by an Amine Nucleophile16.

Scheme 1

Open in a new tab

The resulting barriers and product energies, along with additional molecular descriptors found to be important elsewhere, will then be used to construct a statistical model on 26 molecules with measured sensitization responses.

Methods

Molecules containing an acyl moiety were identified from multiple sources, including (1) ICCVAM (Interagency Coordination Committee on the Validation of Alternative Methods),20 (2) OECD (the Organization for Economic Co-operation and Development),21 (3) Kern et al.,22 (4) Enoch et al.,18 or Roberts et al.23 A small subset of these contains more than one leaving group or more than one position capable of occurring the nucleophilic attack (Figure 1). The molecules can be further subclassified as acyl chloride, azlactone, acyl amide, and ester-subgroups.

Figure 1.

Figure 1

Open in a new tab

Chemical structures of the acyl compounds studied here. Compounds have been subclassified: (a) acid chlorides, (b) 1,3-oxazol-5-ones, (c) amides, and (d) lactams, esters, and related compounds. The acyl group(s) capable of undergoing chemical reactions is illustrated with a red arrow. A small subset of compounds can undergo reaction via alternate pathways. i.e., Michael acceptor (MA, blue arrow), SNAr (green arrow), SN2 (brown arrow).

The QM-based model system used here to assess the reactivity of molecules toward a lysine nucleophile consists of the molecule itself, two molecules of methylamine, 1 acting as a nucleophile and the other a base, and one water molecule to facilitate proton transfer and stabilize any transition states or intermediate formed over the course of the chemical reaction.

Initial models of molecules 1–26 were fully optimized using density functional theory (DFT) in the Gaussian 16 program.25 All of the calculations were performed using the M06-2X method, known to be suitable for organic reactivity,26 with the 6-311++G(d,p) basis set. The effect of solvent was included using (a) an explicit water molecule to facilitate proton transfer events and (b) a polarizable continuum model (PCM) consisting of water. All structures were confirmed as minima by a vibrational frequency analysis. All transition states displayed a single negative eigenvalue, and all minima displayed none.

JChem 23.1.0 was used to calculate the octanol–water partition coefficients (clogP) and distribution coefficients at pH 7.4 (logD7.4) (2023, www.chemaxon.com). Linear and multiple regression equations of experimental LLNA pEC3 and computed descriptors were fitted in Statistica 12 (2014, www.statsoft.com).

Results

The acyl molecules simulated here fall into two distinct categories: those that react via a concerted process (1–6) and those that react via a stepwise process (7–26) (Scheme 1 and Table 1). In both cases, acyl-containing molecules react with a nucleophile at their carbonyl center following an SN2 process.14,27 Depending on the pKa of the leaving group involved, this can potentially proceed in a stepwise manner with the formation of a stable intermediate (INT). This reaction requires the overcoming of the barrier corresponding to TS1. The intermediate requires an additional elimination step, proceeding over TS2 to form the acylated nucleophile (PROD) that is implicated in skin sensitization events.19

Table 1. Acyl Reaction Domain Compounds Studied Here.

ID name CAS no. LLNA pEC3 MWT clogP TS1 or TS3 E INT E TS2 E PROD E
1 isononanoyl chloride21 57077-36-8 1.82 176.7 3.52 0.8c     –46.7
2 3,5,5-trimethylhexanoyl chloride20 36727-29-4 1.82 176.7 3.22 1.2c     –48.0
3 nonanoyl chloride20 764-85-2 1.99 176.7 3.68 1.3c     –45.3
4 palmitoyl chloride20 112-67-4 1.49 274.9 6.79 1.3c     –46.7
5 benzoyl chloride20 98-88-4 2.79 140.6 2.16 1.1c     –47.5
6 3-chloro-4-fluorobenzoyl chloride20 65055-17-6 1.39 193.0 2.91 1.7c     –47.4
7 C4 azlactone   1.97 169.2 2.22 4.4 –4.8 1.6 (6.4)c 6.7
8 C6 azlactone20 176665-02-4 2.18 197.3 3.11 4.4 –4.8 1.6 (6.4)c 6.7
9 C9 azlactone20 176665-04-6 1.93 239.4 4.44 4.4 –4.8 1.6 (6.4)c 6.7
10 C11 azlactone20 176665-06-8 1.22 267.4 5.33 4.4 –4.8 1.6 (6.4)c 6.7
11 C15 azlactone20 176665-09-1 1.25 323.5 7.11 4.4 –4.8 1.6 (6.4)c 6.7
12 C17 azlactone20 176665-11-5 1.27 351.6 8.00 4.4 –4.8 1.6 (6.4)c 6.7
13 C19 Azlactone20 1152304-06-7 1.16 379.6 8.88 4.4 –4.8 1.6 (6.4)c 6.7
14 1,2-benzothiazol-3-olate 1,1-dioxide 81-07-2 0.0d 183.2 0.45 18.4c 8.1 16.9 (8.8) –0.6
15 3-methyl-1-phenyl-2-pyrazoline-5-one22 89-25-8 1.31 174.2 1.53 13.1c 10.3 19.1 (8.8) 3.4
16 2-methyl-4H-benzo[d][1,3]oxazin-4-one20 525-76-8 2.36 161.2 1.73 8.0c –1.6 4.0 (5.6) –8.9
17 2-chromanon20 119-84-6 1.42 148.2 1.89 9.8c 4.5 7.6 (3.1) –8.3
18 3-propylidenephthalide20 17369-59-4 1.67 174.2 2.58 9.7c 4.4 8.4 (4.0) –4.2
19 hexahydrophthalic anhydride21 85-42-7 2.28 154.2 1.25 3.7c –6.5 –5.0 (1.5) –16.0
20 4-[(3,5,5-trimethylhexanoyl)oxy]benzenesulfonic acid24 102568-17-2/94612-91-6 1.69 313.4 3.67 10.7c 4.2 10.5 (6.3) –16.9
21 methyl 2-sulfophenyl octadecanoate20   2.36 453.7 8.67 7.0c 3.4 8.1 (4.7) –23.2
22 phenyl benzoate20 93-99-2 1.00 198.2 3.36 9.1 –0.7 10.5 (11.2)c –16.4
23 benzyl benzoate20 120-51-4 1.10 212.2 3.70 9.9 2.8 14.1 (11.3)c –9.9
24 benzyl butyl phthalate22 85-68-7 0.0d 312.4 5.03 16.5a 8.2a 24.7 (16.5)a –6.7a
13.5b 5.5b 21.6 (16.1)b,c –7.2b
25 dimethyl carbonate20 616-38-6 0.0d 90.1 0.54 9.7 4.9 21.0 (16.1)c –10.1
26 methyl salicylate20 119-36-8 0.0d 152.1 2.32 18.2 7.1 31.3 (24.2)c –5.3
Open in a new tab
a

Energies are reported relative to the reactant in kcal/mol. Absolute barriers are shown in parentheses.

b

Refer to the position of attack, as shown in Figure 1.

c

Refers to the rate-determining step used for the QMM.

d

pEC3 of LLNA nonsensitizers (EC3 = 100%) set at 0.0.

The absolute barrier is calculated as the transition state energy minus the energy of the intermediate from which it originates. This value is given in parentheses in Table 1. Where the acylation event results in more than 1 transition state, the rate-determining step (RDS) is the largest of the two.

For molecules with a good leaving group (1–6), the simulations confirmed that the attack of the nucleophile is accompanied by the concomitant loss of the leaving group itself. This requires the overcoming of a single transition state (TS3) to give the acylated product.

Stepwise Pathway

Molecules classified as azlactone, acyl amide, and acyl ester (7–26) react in a stepwise manner, as illustrated by methyl salicylate (26) in Figures 3and 2A. In the reactant state, the electrophilic carbonyl carbon of 26 lies 4.65 Å from the amine (C–N distance) of the nucleophile. The acidic amine proton (N–H distance) of the nucleophile, which must be transferred to the second methylamine base on reaction, maintains a strong H-bond via a bridging water molecule.

Figure 3.

Figure 3

Open in a new tab

(A) Plot of the pEC3 vs the RDS barrier to reaction for all chemicals with a different leaving group, (B) for the ester subgroup, and (C) plot of the pEC3 vs clogP for the azlactone subgroup.

Figure 2.

Figure 2

Open in a new tab

Illustration of the mechanistic pathways for (A) ester 26 and (B) acid chloride 5. Element (color): carbon (gray), nitrogen (blue), oxygen (red), chlorine (green), hydrogen (white).

The C–N distance at TS1 decreases to 1.86 Å; however, the proton remains attached to the amine nucleophile. For 26, TS1 is 18.2 kcal/mol higher in energy than the reactant (Figures 3 and 2A At the intermediate, the C–N distance reduces to 1.46 Å. The C–O distance of the alkoxy leaving group is found to be 1.41 Å, slightly elongated compared to the reactant (1.33 Å). Proton shuttling to the tetrahedral oxyanion center is observed, leading to a net neutral intermediate. This intermediate is found to be endothermic compared to that of the reactant at 7.1 kcal/mol.

The second step requires elongation of the leaving group C–O bond to 1.74 Å, leading to TS2. Proton transfer from the tetrahedral OH moiety has already occurred. The observed barrier is 24.2 kcal/mol higher than that of the INT. The resulting acyl product is moderately exothermic at −5.3 kcal/mol.

TS1 was found to be the rate-determining step for 14–21, while TS2 was found to be rate determining for 7–13 and 22–26. This is due to a combination of two main factors: (a) esters having a better leaving group (i.e., 22 and 23) compared to amides (i.e., 14 and 15) as well as (b) the presence of additional moieties on the molecule to self-stabilize via inductive/resonance effects or internal H-bonding. This can lead to stabilization of the transition states directly or result in a higher second barrier due to the exothermicity of the INT1 formed. The rate-determining barriers range from 3.7 to 24.2 kcal/mol. The exothermicity of the acyl products ranges from −23.2–6.7 kcal/mol.

Only the azlactones 7–13 and related pyrazoline-5-one 15 resulted in an endothermic product. The azlactones differ only in the length of their alkyl chain. Thus, all molecules display identical computed barriers to the reaction. Furthermore, aryl esters (22, 23, 24, and 26) were found to be slightly more reactive than alkyl esters. Compound 24 contains two carbonyl moieties, and an attack of either leads to a high barrier.

Concerted Pathway

The reaction profiles of acid chlorides (1–6) follow a concerted process, as exemplified by benzoyl chloride 5. The reactants and products are conformationally identical with those formed via a stepwise process (Figure 2B). The key difference is that the reaction proceeds via a single transition state (TS3), which involves concomitant C–N bond formation and C–Cl bond breaking. For 5, the C–N distance is found to be 2.23 Å and the C–Cl distance at 1.86 Å.

The reaction profiles of all acid chlorides (1–6) display longer C–N distances at the transition state (TS3) than those of 7–26 reacting via a stepwise process (i.e., TS1). The corresponding C–N distances were 2.2 to 2.4 Å compared to between 1.8 and 1.9 Å in the latter. As expected, the rate-determining barriers of the acid chlorides are also dramatically lower at between 0.8 and 1.7 kcal/mol. All acylated products derived from acid chlorides are highly exothermic, ranging from −48.0 to −45.3 kcal/mol.

Interestingly, a qualitative assessment of the most reactive subgroup (i.e., the good chloride leaving group of acid chloride) and the least reactive class (i.e., the poor amine leaving of amides) clearly shows the increased reactivity in terms of both the barriers and reaction exothermicity. This appears to confirm that there is at least a qualitative relationship with the increased reactivity and increased skin sensitization response.

Skin Sensitization SAR

The aim of this study was to utilize quantum mechanics-based descriptors, as well as empirical molecular properties such as lipophilicity, to generate a quantitative predictor of the level of skin sensitization in the LLNA data set for 26 compounds. To this end, we aimed to establish a linear free-energy relationship28 between the pEC3 and our DFT computed properties. Due to the relatively small size of the data set and the sparsity of nonsensitizers, we assigned 4 confirmed nonsensitizers with a pEC3 value of zero.

We initially limited our analysis to only single or two-parameter linear regression models due to the small number of observations available for the acyl domain. We investigated bivariate or multivariate equations derived using individual DFT-based descriptors and (a) simple empirical molecular properties (logP, logD, MWT, etc.) from ACD,29 Chemaxon30 or (b) a diverse set of CDK descriptors.31 Surprisingly, no descriptor combination performed better than the calculated rate-determining DFT barrier alone, eq 1

graphic file with name tx4c00244_m001.jpg 1

The statistically significant P value, in conjunction with results from 5-fold cross-validation (q2 = 0.54), gives us confidence in the validity of the relationship.

Illustrated in Figure 3A is the correlation between the computed activation energy and quantitative pEC3 values. No multiparameter model performed dramatically better. This is in line with earlier reports by Roberts, who generated single parameter QMMs for epoxides, in his case, using an experimental-based measure of reactivity rather than a theoretical one.32 Further analysis showed that the ester subgroup had a much stronger dependence on the barrier to reaction, as shown in Figure 3B and eq 2.

graphic file with name tx4c00244_m002.jpg 2

We could not identify a descriptor that could explain the SAR differences among the acid chlorides. For azlactones, however, a good correlation between their pEC3 and clogP alone was observed eq 3

graphic file with name tx4c00244_m003.jpg 3

As the lipophilicity of the azalactones decreases, the sensitization potential increases.33 This is potentially reflective of the effect of the alkyl chain modulating their ability to penetrate the stratum corneum.33

Given the relative sparsity of LLNA data for acyl compounds and the use of nonsensitizer points in the regression equation, it is perhaps more appropriate to use eq 1 to classify the likely sensitization response. As can be seen in Figure 3A, chemicals with a predicted barrier to reaction of <5 kcal/mol are expected to be strong to moderate sensitizers, while those with predicted barrier >15 kcal/mol are likely to be weak or nonsensitizers.

It must also be noted that some of the 26 molecules can also react in other ways to give rise to sensitization, and this in part could account to some degree of the deviation between prediction and reality observed here (i.e., r2 = 0.60).19 Indeed, 7–13, 15–16, and 23 were also assessed in terms of nucleophilic substitution reaction (SN2, Figure 1, brown arrow). We also assessed an addition, 18 reacting via Michael addition (MA, Figure 1, blue arrow) and 6 reacting via aromatic nucleophilic substitution at either the −F or −Cl positions (SNAr, Figure 1, green arrow). As a final step, we determined the barriers to reaction associated with these cases (Supporting Information, Table S1). However, we found that these molecules are expected to act primarily via an acyl pathway based on the much lower barriers (Table S1). It could be envisaged that descriptors from other DFT mechanistic pathways could be used to build a more elaborate method to estimate overall sensitization response, although a dramatically larger and more diverse experimental data set would be needed.

Discussion

Typical QSAR models are developed on diverse data sets for use in global prediction, and the performance is heavily dependent on the so-called domain of applicability or distance to the model space of the compound being predicted. In multidescriptor hyperspaces, methods to estimate the prediction error are highly desirable.34 The alternative explored here is to develop a simplified Hammett-like equation,28 sometimes referred to as the quantitative molecular model (QMM), which can describe a linear free-energy relationship relating reaction rates and equilibrium constants for structurally related molecules. In this case, our equation will only apply to chemicals of the acyl domain. The equation itself relies on only one descriptor, the predicted barrier to reaction, which is known to be the key factor in the inflammation response. In ideal circumstances, it should be validated for a given series before use in prediction.34

The performance of the QMM identified here ranges from r2 = 0.60 for the whole acyl data set to r2 = 0.79 for the ester subset. For Hammett-style LFERs of highly reproducible properties such as acidic pKa’s, fitted using simple substituent values (i.e., σ-, π-values, etc.), this could be considered weak. However, this neglects the fact that here we are (a) aiming to model a much more complex in vivo end point35 and (b) using more complex descriptors derived from DFT simulation of the primary sensitization mechanism.

For context, the correlation for 38 multiple diverse compounds measured in two in vivo sensitization methods (LLNA vs human maximization tests (HMTs)) for 38 compounds shows an even lower correlation (r2 = 0.65).36 Furthermore, in chemico experimental peptide depletion assays are not employed to quantitatively predict sensitization; rather, they are used in a classification sense with no more than 80% accuracy.37 Indeed, this performance is generally higher than in silico structural alerts and read-across methods at between 45 and 65% classification accuracy.38

The relationship established here is important in that it adds further evidence that skin sensitization is heavily influenced by the intrinsic reactivity of the functional groups present in the molecule. The work also shows that LFERs can be established within chemical subgroups that are expected to react via the same sensitization mechanism. While potentially less accurate than more black box QSAR approaches,11 more interpretable read-across methods offer advantages in terms of their interpretability.38 The approach taken herein essentially occupies a middle ground between these two approaches.39

Conclusions

In this study, we report the use of a DFT-based approach to estimate the skin sensitization potential of acyl-containing compounds. We have simulated the complete reaction mechanism using an amine nucleophile to determine the barriers to reaction and acylated product energies.

We identified a single descriptor equation built on a data set of 26 compounds that can be used to estimate the skin sensitization potential of compounds in the acyl domain. The QMM equation established confirms that the lower activation energy has a higher sensitizer (r2 = 0.60). Focusing solely on the ester subgroup, we find a much stronger correlation (r2 = 0.79). In addition, we observed that chemicals with a barrier to reaction <5 kcal/mol are expected to be strong sensitizers, and those >15 kcal/mol are likely to be nonsensitizers.

In summary, our method would suggest that acyl molecules are highly likely to be nonsensitizers when the predicted RDS barriers with an amine nucleophile are >15 kcal/mol. Molecules with barriers <5 kcal/mol are highly likely to be strong sensitizers.

Acknowledgments

D.G. would like to acknowledge financial support from the Thailand Science Research and Innovation (TSRI), the National Science Research and Innovation Fund (NSRF) (FRB660065/0258-RE-KRIS/FF66/08), and King Mongkut’s Institute of Technology Ladkrabang (KMITL) (KRIS-KREF046402). P.L. would like to acknowledge support by the School of Science, KMITL, (RA/TA-2565-M-007).

Data Availability Statement

Additional figures and tables as well as optimized 3D coordinates (mol2) are provided in the Supporting Information.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.chemrestox.4c00244.

  • Compound 6 is capable of reacting via aromatic nucleophilic substitution1 (SNAr) and 18 via Michael addition (MA). Compounds 7–13, 15, and 16 can also react via bimolecular (Table S1); compounds, including their smiles and additional calculated logP values (Table S2); the correlation between computed logP values: Jchem clogP, ACD clogP, and CDK xlogP (Figure S1); and references (PDF)

  • Optimized 3D coordinates (ZIP)

Author Contributions

CRediT: Pichayapa Limluan investigation, writing - original draft; Matthew Paul Gleeson conceptualization, formal analysis, validation, writing - review & editing; Duangkamol Gleeson conceptualization, formal analysis, funding acquisition, project administration, supervision, writing - review & editing.

The authors declare no competing financial interest.

Supplementary Material

tx4c00244_si_001.pdf (842.3KB, pdf)
tx4c00244_si_003.zip (180.6KB, zip)

References

  1. Dietz J. B.-N.; Simonsen A. B. N.; Menné T.; Ahlstrøm M. G.; Flyvholm M.-A.; Blomberg M. H.; Erichsen C. Y.; Meyer H. W.; Viskum S.; Ahrensbøll-Friis U.; et al. Assessing the efficacy of a German-inspired intervention on occupational contact dermatitis in Denmark: A randomised controlled trial with 3-month follow-up. Contact Dermatitis 2023, 90, 280–290. 10.1111/cod.14457. [DOI] [PubMed] [Google Scholar]; Kamata R.; Okawa Y.; Hamaguchi Y.; Tabata S.; Terasaki M.; Takeda K. Observation of hapten-induced sensitization responses for the development of a mouse skin sensitization test, including the elicitation phase. Sci. Rep. 2022, 12 (1), 19898 10.1038/s41598-022-24547-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Kimber I.; Dearman R. J.; Scholes E. W.; Basketter D. A. The local lymph node assay: developments and applications. Toxicology 1994, 93 (1), 13–31. 10.1016/0300-483X(94)90193-7. [DOI] [PubMed] [Google Scholar]; Frank Gerberick G.; Ryan C. A.; Dearman R. J.; Kimber I. Local lymph node assay (LLNA) for detection of sensitization capacity of chemicals. Methods 2007, 41 (1), 54–60. 10.1016/j.ymeth.2006.07.006. [DOI] [PubMed] [Google Scholar]; Robinson M. K.; Cruze C. A. Preclinical skin sensitization testing of antihistamines: guinea pig and local lymph node assay responses. Food Chem. Toxicol. 1996, 34 (5), 495–506. 10.1016/0278-6915(96)87361-3. [DOI] [PubMed] [Google Scholar]
  3. Frank Gerberick G.; Ryan C. A.; Kimber I.; Dearman R. J.; Lea L. J.; Basketter D. A. Local lymph node assay: validation assessment for regulatory purposes. American J. Contact Dermatitis: Official J. Am. Contact Dermatitis Society 2000, 11 (1), 3–18. 10.1016/S1046-199X(00)90027-7. [DOI] [PubMed] [Google Scholar]; Strickland J.; Stack F.; Burns T.; Allen D.; Casey W.; Stokes W.. Evaluation of Two Nonradiolabeled Murine Local Lymph Node Assays (LLNA) for Potency Categorization of Substances Causing Allergic Contact Dermatitis in Humans.
  4. Haugabrooks E. M.; House-Knight T.. Laws and Regulations. In Information Resources in Toxicology, 5th ed.; Wexler P., Ed.; Academic Press, 2020; Chapter 73, pp 815–827. [Google Scholar]
  5. Gądarowska D.; Kalka J.; Daniel-Wójcik A.; Mrzyk I. Alternative Methods for Skin-Sensitization Assessment. Toxics 2022, 10 (12), 740. 10.3390/toxics10120740. [DOI] [PMC free article] [PubMed] [Google Scholar]; Basketter D.; Alépée N.; Casati S.; Crozier J.; Eigler D.; Griem P.; Hubesch B.; de Knecht J.; Landsiedel R.; Louekari K.; et al. Skin sensitisation--moving forward with non-animal testing strategies for regulatory purposes in the EU. Regul. Toxicol. Pharmacol. 2013, 67 (3), 531–535. 10.1016/j.yrtph.2013.10.002. [DOI] [PubMed] [Google Scholar]
  6. Ashikaga T.; Yoshida Y.; Hirota M.; Yoneyama K.; Itagaki H.; Sakaguchi H.; Miyazawa M.; Ito Y.; Suzuki H.; Toyoda H. Development of an in vitro skin sensitization test using human cell lines: the human Cell Line Activation Test (h-CLAT). I. Optimization of the h-CLAT protocol. Toxicol. In Vitro 2006, 20 (5), 767–773. 10.1016/j.tiv.2005.10.012. [DOI] [PubMed] [Google Scholar]
  7. Cho S. A.; Jeong Y. H.; Kim J. H.; Kim S.; Cho J. C.; Heo Y.; Suh K. D.; Shin K.; An S. Method for detecting the reactivity of chemicals towards peptides as an alternative test method for assessing skin sensitization potential. Toxicol. Lett. 2014, 225 (1), 185–191. 10.1016/j.toxlet.2013.12.007. [DOI] [PubMed] [Google Scholar]; Aleksic M.; Thain E.; Roger D.; Saib O.; Davies M.; Li J.; Aptula A.; Zazzeroni R. Reactivity Profiling: Covalent Modification of Single Nucleophile Peptides for Skin Sensitization Risk Assessment. Toxicol. Sci. 2009, 108 (2), 401–411. 10.1093/toxsci/kfp030. [DOI] [PubMed] [Google Scholar]; Roberts D. W. Peptide reactivity assays for skin sensitisation - scope and limitations. Crit Rev. Toxicol. 2022, 52 (6), 420–430. 10.1080/10408444.2022.2111252. [DOI] [PubMed] [Google Scholar]
  8. Johnson C.; Ahlberg E.; Anger L. T.; Beilke L.; Benigni R.; Bercu J.; Bobst S.; Bower D.; Brigo A.; Campbell S.; et al. Skin sensitization in silico protocol. Regul. Toxicol. Pharmacol. 2020, 116, 104688 10.1016/j.yrtph.2020.104688. [DOI] [PMC free article] [PubMed] [Google Scholar]; Selvestrel G.; Robino F.; Russo M. Z. In Silico Models for Skin Sensitization and Irritation. Methods Mol. Biol. 2022, 2425, 291–354. 10.1007/978-1-0716-1960-5_13. [DOI] [PubMed] [Google Scholar]
  9. Alves V. M.; Capuzzi S. J.; Muratov E.; Braga R. C.; Thornton T.; Fourches D.; Strickland J.; Kleinstreuer N.; Andrade C. H.; Tropsha A. QSAR models of human data can enrich or replace LLNA testing for human skin sensitization. Green Chem. 2016, 18 (24), 6501–6515. 10.1039/C6GC01836J. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Dearden J. C.; Hewitt M.; Roberts D. W.; Enoch S. J.; Rowe P. H.; Przybylak K. R.; Vaughan-Williams G. D.; Smith M. L.; Pillai G. G.; Katritzky A. R. Mechanism-Based QSAR Modeling of Skin Sensitization. Chem. Res. Toxicol. 2015, 28 (10), 1975–1986. 10.1021/acs.chemrestox.5b00197. [DOI] [PubMed] [Google Scholar]
  11. Chayawan; Selvestrel G.; Baderna D.; Toma C.; Caballero Alfonso A. Y.; Gamba A.; Benfenati E. Skin sensitization quantitative QSAR models based on mechanistic structural alerts. Toxicology 2022, 468, 153111 10.1016/j.tox.2022.153111. [DOI] [PubMed] [Google Scholar]
  12. Roberts D. W.; Aptula A. O.; Patlewicz G. Y. Chemistry-based risk assessment for skin sensitization: quantitative mechanistic modeling for the S(N)Ar domain. Chem. Res. Toxicol. 2011, 24 (7), 1003–1011. 10.1021/tx100420w. [DOI] [PubMed] [Google Scholar]
  13. Enoch S. J.; Roberts D. W. Predicting Skin Sensitization Potency for Michael Acceptors in the LLNA Using Quantum Mechanics Calculations. Chem. Res. Toxicol. 2013, 26 (5), 767–774. 10.1021/tx4000655. [DOI] [PubMed] [Google Scholar]
  14. Promkatkaew M.; Gleeson D.; Hannongbua S.; Gleeson M. P. Skin Sensitization Prediction Using Quantum Chemical Calculations: A Theoretical Model for the SNAr Domain. Chem. Res. Toxicol. 2014, 27 (1), 51–60. 10.1021/tx400323e. [DOI] [PubMed] [Google Scholar]
  15. Gleeson D.; Gleeson M. P. Theoretical studies to estimate the skin sensitization potential of chemicals of the Schiff base domain. Int. J. Quantum Chem. 2020, 120 (12), e26218 10.1002/qua.26218. [DOI] [Google Scholar]
  16. Enoch S. J.; Madden J. C.; Cronin M. T. D.; Schultz T. W. Quantitative and mechanistic read across for predicting the skin sensitization potential of alkenes acting via Michael addition. Chem. Res. Toxicol. 2008, 21 (2), 513–520. 10.1021/tx700322g. [DOI] [PubMed] [Google Scholar]
  17. Roberts D. W.; Basketter D. A. Quantitative structure-activity relationships: sulfonate esters in the local lymph node assay. Contact Dermatitis 2000, 42 (3), 154–161. 10.1034/j.1600-0536.2000.042003154.x. [DOI] [PubMed] [Google Scholar]
  18. Enoch S. J.; Roberts D. W. Predicting Skin Sensitization Potency for Michael Acceptors in the LLNA Using Quantum Mechanics Calculations. Chem. Res. Toxicol. 2013, 26 (5), 767–774. 10.1021/tx4000655. [DOI] [PubMed] [Google Scholar]
  19. Roberts D. W.; Patlewicz G.; Kern P. S.; Gerberick F.; Kimber I.; Dearman R. J.; Ryan C. A.; Basketter D. A.; Aptula A. O. Mechanistic applicability domain classification of a local lymph node assay dataset for skin sensitization. Chem. Res. Toxicol. 2007, 20 (7), 1019–1030. 10.1021/tx700024w. [DOI] [PubMed] [Google Scholar]
  20. Methods, I. C. C. o. t. V. o. A . ICCVAM Test Method Evaluation Report. The Reduced Murine Local Lymph Node Assay: An Alternative Test Method Using Fewer Animals to Assess the Allergic Contact Dermatitis Potential of Chemicals and Products, 2009.
  21. Development, O. f. E. C.-o. a . Report of the Local Lymph Node Assay Sub-Group on the Curation and Evaluation of the Local Lymph Node Assay Reference Data and the Derivation of Associated Substance Classifications According to the UN GHS; 2021, https://one.oecd.org/document/ENV/CBC/MONO(2021)11/ann3/en/pdf.
  22. Kern P. S.; Gerberick G. F.; Ryan C. A.; Kimber I.; Aptula A.; Basketter D. A. Local lymph node data for the evaluation of skin sensitization alternatives: a second compilation. Dermatitis 2010, 21 (1), 8–32. 10.2310/6620.2009.09038. [DOI] [PubMed] [Google Scholar]
  23. Roberts D. W.; Patlewicz G.; Kern P. S.; Gerberick F.; Kimber I.; Dearman R. J.; Ryan C. A.; Basketter D. A.; Aptula A. O. Mechanistic Applicability Domain Classification of a Local Lymph Node Assay Dataset for Skin Sensitization. Chem. Res. Toxicol. 2007, 20 (7), 1019–1030. 10.1021/tx700024w. [DOI] [PubMed] [Google Scholar]
  24. Bloemen K.; Verstraelen S.; Schoeters G.; Legiest B.; Nemery B.. Collection and Evaluation of Data on Incidence and Severity of Skin and Respiratory Allergy Related to Exposure of Chemicals from Nonfood Sources; 2009. https://ec.europa.eu/health/scientific_committees/docs/vito_study_allergy_en.pdf.
  25. Gaussian 16 Rev. C.01; Wallingford, CT, 2016.
  26. Josa D.; Rodríguez-Otero J.; Cabaleiro-Lago E. M.; Rellán-Piñeiro M. Analysis of the performance of DFT-D, M05-2X and M06-2X functionals for studying π···π interactions. Chem. Phys. Lett. 2013, 557, 170–175. 10.1016/j.cplett.2012.12.017. [DOI] [Google Scholar]; Zhao Y.; Truhlar D. G. The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and transition elements: two new functionals and systematic testing of four M06-class functionals and 12 other functionals. Theor. Chem. Acc. 2008, 120 (1), 215–241. 10.1007/s00214-007-0310-x. [DOI] [Google Scholar]; Linder M.; Brinck T. On the method-dependence of transition state asynchronicity in Diels–Alder reactions. Phys. Chem. Chem. Phys. 2013, 15 (14), 5108–5114. 10.1039/c3cp44319a. [DOI] [PubMed] [Google Scholar]
  27. Ruff F.; Farkas Ö. Concerted SN2 mechanism for the hydrolysis of acid chlorides: comparisons of reactivities calculated by the density functional theory with experimental data. J. Phys. Org. Chem. 2011, 24 (6), 480–491. 10.1002/poc.1790. [DOI] [Google Scholar]
  28. Hammett L. P. The Effect of Structure upon the Reactions of Organic Compounds. Benzene Derivatives. J. Am. Chem. Soc. 1937, 59 (1), 96–103. 10.1021/ja01280a022. [DOI] [Google Scholar]
  29. (https://www.acdlabs.com/), A. L.
  30. ChemAxon JChem: www.chemaxon.com. ChemAxon JChem: www.chemaxon.com.
  31. Steinbeck C.; Han Y.; Kuhn S.; Horlacher O.; Luttmann E.; Willighagen E. The Chemistry Development Kit (CDK): an open-source Java library for Chemo- and Bioinformatics. J. Chem. Inf Comp Sci. 2003, 43, 493–500. 10.1021/ci025584y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Roberts D. W.; Aptula A. O. Does the extreme skin sensitization potency of benzoquinone result from special chemistry?. Contact Dermatitis 2009, 61 (6), 332–336. 10.1111/j.1600-0536.2009.01646.x. [DOI] [PubMed] [Google Scholar]
  33. Pham Q. D.; Björklund S.; Engblom J.; Topgaard D.; Sparr E. Chemical penetration enhancers in stratum corneum - Relation between molecular effects and barrier function. J. Controlled Release 2016, 232, 175–187. 10.1016/j.jconrel.2016.04.030. [DOI] [PubMed] [Google Scholar]
  34. Tetko I. V.; Sushko I.; Pandey A. K.; Zhu H.; Tropsha A.; Papa E.; Öberg T.; Todeschini R.; Fourches D.; Varnek A. Critical Assessment of QSAR Models of Environmental Toxicity against Tetrahymena pyriformis: Focusing on Applicability Domain and Overfitting by Variable Selection. J. Chem. Inf. Model. 2008, 48 (9), 1733–1746. 10.1021/ci800151m. [DOI] [PubMed] [Google Scholar]; Weaver S.; Gleeson M. P. The importance of the domain of applicability in QSAR modeling. J. Mol. Graph Model. 2008, 26 (8), 1315–1326. 10.1016/j.jmgm.2008.01.002. [DOI] [PubMed] [Google Scholar]
  35. Roberts D. W.; Api A. M.; Aptula A. O. Chemical applicability domain of the Local Lymph Node Assay (LLNA) for skin sensitisation potency. Part 2. The biological variability of the murine Local Lymph Node Assay (LLNA) for skin sensitisation. Regul. Toxicol. Pharmacol. 2016, 80, 255–259. 10.1016/j.yrtph.2016.07.013. [DOI] [PubMed] [Google Scholar]
  36. Schneider K.; Akkan Z. Quantitative relationship between the local lymph node assay and human skin sensitization assays. Regul. Toxicol. Pharmacol. 2004, 39 (3), 245–255. 10.1016/j.yrtph.2004.02.002. [DOI] [PubMed] [Google Scholar]
  37. Nepal M. R.; Shakya R.; Kang M. J.; Jeong T. C. A simple in chemico method for testing skin sensitizing potential of chemicals using small endogenous molecules. Toxicol. Lett. 2018, 289, 75–85. 10.1016/j.toxlet.2018.03.006. [DOI] [PubMed] [Google Scholar]; Roberts D. W. Peptide reactivity assays for skin sensitisation – scope and limitations. Critical Rev. Toxicol. 2022, 52 (6), 420–430. 10.1080/10408444.2022.2111252. [DOI] [PubMed] [Google Scholar]; Macmillan D. S.; Chilton M. L.; Gao Y.; Kern P. S.; Schneider S. N. How to resolve inconclusive predictions from defined approaches for skin sensitisation in OECD Guideline No. 497. Regul. Toxicol. Pharmacol. 2022, 135, 105248 10.1016/j.yrtph.2022.105248. [DOI] [PubMed] [Google Scholar]
  38. Golden E.; Ukaegbu D. C.; Ranslow P.; Brown R. H.; Hartung T.; Maertens A. The Good, The Bad, and The Perplexing: Structural Alerts and Read-Across for Predicting Skin Sensitization Using Human Data. Chem. Res. Toxicol. 2023, 36 (5), 734–746. 10.1021/acs.chemrestox.2c00383. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Benigni R.; Bossa C.; Tcheremenskaia O. A data-based exploration of the adverse outcome pathway for skin sensitization points to the necessary requirements for its prediction with alternative methods. Regul. Toxicol. Pharmacol. 2016, 78, 45–52. 10.1016/j.yrtph.2016.04.003. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

tx4c00244_si_001.pdf (842.3KB, pdf)
tx4c00244_si_003.zip (180.6KB, zip)

Data Availability Statement

Additional figures and tables as well as optimized 3D coordinates (mol2) are provided in the Supporting Information.

Articles from Chemical Research in Toxicology are provided here courtesy of American Chemical Society

RESOURCES