QSAR Models of Human Data can Enrich or Replace LLNA Testing for Human Skin Sensitization
Vinicius M. Alves, Stephen J. Capuzzi, Eugene N. Muratov, Rodolpho C. Braga, Thomas E. Thornton, Denis Fourches, Judy Strickland, Nicole Kleinstreuer, Carolina H. Andrade and Alexander Tropsha.
Green Chemistry (2016) DOI: https://doi.org/10.1039/c6gc01836j PMID: 28630595
Skin sensitization is a major environmental and occupational health hazard. Although many chemicals have been evaluated in humans, there have been no efforts to model these data to date. We have compiled, curated, analyzed, and compared the available human and LLNA data. Using these data, we have developed reliable computational models and applied them for virtual screening of chemical libraries to identify putative skin sensitizers. The overall concordance between murine LLNA and human skin sensitization responses for a set of 135 unique chemicals was low (R = 28-43%), although several chemical classes had high concordance. We have succeeded to develop predictive QSAR models of all available human data with the external correct classification rate of 71%. A consensus model integrating concordant QSAR predictions and LLNA results afforded a higher CCR of 82% but at the expense of the reduced external dataset coverage (52%). We used the developed QSAR models for virtual screening of CosIng database and identified 1061 putative skin sensitizers; for seventeen of these compounds, we found published evidence of their skin sensitization effects. Models reported herein provide more accurate alternative to LLNA testing for human skin sensitization assessment across diverse chemical data. In addition, they can also be used to guide the structural optimization of toxic compounds to reduce their skin sensitization potential.
Figure 1. Human DSA05 vs. LLNA EC3 for all 53 sensitizers.
Human DSA05 vs. LLNA EC3 for all 53 sensitizers (y = 0.3111x + 0.8259; R² = 0.08) – Left Panel; and 50 sensitizers (y = 1.4383x + 2.961; R² = 0.05) remained after exclusion of three outliers (lilial, pyridine, and phenyl benzoate) – Right Panel.
- Figure 1 (139 KB)
Figure 2. Results of cluster analysis of 109 compounds with human skin sensitization data.
(A) Heatmap and (B) dendrogram of the distance matrix, both colored according to structural similarity (blue/violet = similar; yellow/red = dissimilar).
- Figure 2 (175 KB)
Figure 3. Example of structural transformation of human skin sensitizer phenyl benzoate.
Example of structural transformation of human skin sensitizer phenyl benzoate into various non-sensitizers using developed models.
- Figure 3 (262 KB)
Figure 4. Example of interpretation of QSAR models visualized as color-coding.
Example of interpretation of QSAR models visualized as color-coding according to atom contributions in change of sensitization potential: red – sensitization increase; blue – sensitization decrease; green – no significant contribution. Green arrows represent higher confidence of the prediction. This information helps to guide structural transformation of human skin sensitizer phenyl benzoate into various non-sensitizers (see Figure 3).
- Figure 4 (385 KB)
Table 1. Human vs. LLNA experimental outcomes for 109 compounds with defined chemical structure.
- Table 1 (22 KB)
Table 2. Human vs. LLNA multi-class annotation of experimental results for 109 compounds with defined chemical structure.
- Table 2 (22 KB)
Table 3. Statistical characteristics of LLNA results vs. external QSAR predictions.
Statistical characteristics of LLNA results vs. external QSAR predictions (5-fold external cross-validation) for predicting human skin sensitization.
- Table 3 (23 KB)
Table 4. Summary of cluster analysis showing the number of compounds correctly predicted by QSAR and LLNA.
Summary of cluster analysis showing the number of compounds correctly predicted by QSAR and LLNA when compared to the human data.
- Table 4 (29 KB)
Table 5. Human data vs. LLNA results vs. external QSAR predictions within selected chemical clusters.
- Table 5 (184 KB)
- Cluster Analysis of Human Skin Sensitization Dataset (2 MB)
- Predictions (251 KB)
- Table S1, S2, S3 (57 KB)