Assessing the safety of thermal mineral water for cosmetic applications: an integrated approach using physicochemical, cheminformatics, and bioinformatics techniques

Abstract

Thermal and mineral waters represent a complex multifunctional natural resource that has been used for various purposes throughout human history. The physico-chemical characterization of thermal and mineral waters is a comprehensive process that integrates knowledge and practice from different scientific fields. When used in direct contact with human skin, whether for bathing or for use in topical products, a toxicological analysis of thermal and mineral waters must also be performed. This work is an example of a multidisciplinary approach to investigate the safety of concrete thermal and mineral water from the Pannonian Basin for use in cosmetics. A detailed physicochemical characterization was performed together with the subsequent safety assessment of the final cosmetic product, coupled with cheminformatics and bioinformatics tools used to predict physicochemical properties, pharmacokinetics, determination of descriptors to assess bioactive potential and evaluation of possible biological pathways and interactions. The results show that the tested thermal and mineral water is a promising resource for use in cosmetic products that can help maintain skin integrity and improve its condition. The toxicological evaluation showed that the tested water is acceptable as an ingredient in a face cream for adults, excluding pregnant and breastfeeding women. The results are discussed in detail and guidance and comments on outstanding issues are provided.

Fig. 1. Graphic presentation of NOM classification.

Fig. 2. Results of fractionation of NOM from thermal mineral water Novi Bečej (FAF-fraction of fulvic acid, HAF-fraction of humic acids, HPI-A-hydrophilic acid fraction, HPI-NA-hydrophilic non-acidic fraction).

Fig. 3. Humic acid affected pathways fold enrichment: significance levels (−log 10 of FDR – false discover rate) are indicated by different colours as presented. Number of genes associated with each pathway are indicated by different sizes of points.

Fig. 4. Protein–protein interaction network for HA gathered using the STRING database. Nodes labelled with gene names presents the proteins. Lines and edges present the interaction between the proteins. 35 proteins and 547 interactions are detected in the network.

Fig. 5. FA affected pathways fold enrichment. Significance levels (−log 10 of FDR – false discover rate) are indicated by different colours as presented. Number of genes associated with each pathway are indicated by different sizes of points.

Fig. 6. Protein–protein interaction network for FA gathered using the STRING database. Nodes labelled with gene names presents the proteins. Lines and edges present the interaction between the proteins. 26 proteins and 186 interactions are detected in the network.