An integrated In-Silico-In-Chemico-In-Vitro (iSiCiV) Approach to identify biomarkers to predict the skin sensitisation potential of phytochemicals

Background

Skin sensitisation is one of the important regulatory endpoints used to determine whether a given test chemical can elicit an allergic response in susceptible individuals. The in-vitro approaches for predicting skin sensitisation potential for chemicals have been well-developed and widely accepted. Continuous efforts are being made to identify novel solutions to overcome the limitations of the existing approaches.

Purpose

The herbal preparations used in dermal therapeutics are frequently found to elicit skin sensitisation, requiring caution. The herbal preparations tend to contain more than one phytochemical. The expected active ingredient may not possess skin sensitisation. However, the other phytochemical traces could pose synergistic and potentiating effects towards enhanced/suppressed skin sensitisation. In this regard, isolation, purification, yield, and in-vitro / in-vivo evaluation of skin sensitisation for all the traces would be time, cost, and animal-consuming. Hence, the regulatory agencies appreciate using computational tools to substantiate in-vitro methods. The regulators have also endorsed using in-silico approaches like quantitative structure-activity relationships (QSAR) and read-across. These approaches are limited to structure-based predictions and in combination with system-level approaches, which comprehensively facilitate a better understanding of complex biological systems.

Study design

In the present study, 603 phytochemicals were screened in-silico using OECD QSAR Toolbox. The limitations of the QSAR Toolbox in predicting skin sensitisers were identified.

Method

The systems biology-based approach to complement the QSAR-based prediction was developed, and 41 biomarkers were identified as a determinant of skin sensitisation. The molecules attributed to these biomarkers were predicted as skin sensitisers.

Results

As a result, 31 phytochemicals were predicted as skin sensitisers. Acrolein, eugenol, formaldehyde, and glycerol were selected for experimental validation of the in-silico predictions using direct peptide reactivity assay (DPRA), KeratinoSens™, and h-CLAT assays representing each key event in skin sensitisation. Further, identified biomarkers were validated in THP-1 and KeratinoSens cell lines using qRT-PCR. Among all genes, IL-8 and CCL4 were upregulated by 6.4 and 5.6 folds, respectively, in THP-1, while AKR1C2 and AKR1C3 were upregulated by 7.8 and 11.1 folds, respectively, in KeratinoSens. IL-8 level was analysed by ELISA and found to be in line with qRT-PCR results.

Conclusion

The study proposes a possible strategy based on the “iSiCiV” approach to enhance the prediction of the dermal sensitisation potential of a test compound. The approach offers a robust platform and highlights an opportunity to shift the paradigm of skin sensitisation assay with a better understanding of toxicology mechanisms.