24th International Colloquium Tribology - January 2024 235 Using Molecular Modelling to Anticipate Future Toxicity Classifications of Anti-oxidants and Identify Safer Structures Siegfried Lucazeau 1 , Grégoire Hervé 2 , Florence Séverac 3 1 NYCO, Paris, France 2 NYCO, Paris, France 3 NYCO, Paris, France 1. Regulatory background 1.1 Toxicity of classical chemistries As more toxicity studies are carried out, new information on the toxicity of classicial chemistries used in lubrication, once considered as safe, have caused them to now be classified as toxic and possibly make the formulated product itself hazardous. Such changes have required reformulation work, and changes in additive purchasing strategies. 1.2 Lack of confidence in the future Outside of the immediate impact of additive classification changes, a number of questions arise as more chemistries might be found toxic in the future. In order to regain confidence about the long term safety and availability of commonly used additives, we need to be able to identify truly safe compounds and anticipate the toxicity profile of existing or new molecules. This work covers the area of anti-oxidant additives. 2. Computational chemistry to assess toxicity 2.1 Similar molecules have similar activities Computational chemistry uses computer simulation to solve chemical problems; it is widely used in drug design. In particular, QSAR (Quantitiave Structure Activity Relationship) modelling, also called 2D modelling, may be used as a predicting tool. QSAR models are complex mathematical equations that associate the biological activity of a known molecule with its structural features (descriptors or predictors), in order to anticipate the activity of non evaluated compounds. 2.2 Robustness of QSAR models QSAR models have to be educated and evaluated using sets of compounds with known, measured toxicity features. A part of a set is used to build the model, while the other part is used to test the model. The quality of the model may be described by metrics such as sensitivity, specifity, accuracy, or AUC (Area Under the Curve). 3. Additional in-vivo testing Freshwater planarians are flatworms that possess a cholinergic system. They may be used as living models to assess neurodevelopmental toxicology and deliver additional information on potential toxicity of tested compounds. Cholinesterase activity on planarians 4. Results on antioxidants: aminics, phenolics and polyaminic oligomers 4.1 Scope Various classes of anti-oxidants were evaluated using 4 dedicated QSAR models: • Carcinogenicity • Mutagenicity • Reprotoxicity • Neurotoxicity Assessed compounds include alkylated diphenylamines with various alkyl chains, various phenolic anti-oxidants, as well as oligomeric anti-oxidants. 4.2 Results 4.2.1 General comments A few examples of well-known chemical structures used as anti-oxidants in lubricants do confirm the robustness of the QSAR models that have been designed for the purpose. Results on planarians are in good agreement with those of QSAR models, thus giving us more confidence in the overall evaluation system. 4.2.2 Takeaway • Toxicity of phenylalphanaphthylamine is known and confirmed by QSAR modelling on the 4 categories; alkylation of this molecule seems to reduce mutagenicity and neurotoxicty, however carcinogenicity and reprotoxicity may still be an issue. Using Molecular Modelling to Anticipate Future Toxicity Classifications of Anti-oxidants and Identify Safer Structures 236 24th International Colloquium Tribology - January 2024 Alkylated phenyalphanaphthylamine • Diphenylamines carrying iso C4 alkyl groups do show indications of high carcinogenicity and reprotoxicity potential; unfortunately, it appears like none of the evaluated alkyl groups on diphenylamines are able to exclude any risk of toxicity. Alkylated diphenylamines • The evaluation of several phenolic anti-oxidant structures leads us to believe that phenolics do not represent a satisfactory, long term safer alternative. Methylene bis(diterbutyl paraphenol) One specific polyaminic oligomeric antioxidant, however, shows an excellent toxicity profile according to our models, on all 4 toxicity categories; this compound has no effect on planarians either, confirming the credibility of this concept. More structures will be investigated. Oligomer of aminic anti-oxidants 5. Anti-oxidancy performance of oligomers Oligomeric anti-oxidants have long been used in the aviation industry to maximize resistance to thermo-oxidation and cleanliness of jet engine oils. This concept has also been introduced in high temperature industrial applications such as high temperature chain oils, greases, and compressor oils, with proven performance. The question remains as to how the oligomeric anti-oxidant that was identified as safe will perform in the variety of standard, less stringent conditions usually met with classical anti-oxidants; this is an ongoing project. 6. Conclusion Formulators, more than ever, are in need of future-proof, safe technologies to design their lubricants. Carefully selected polyaminic oligomeric anti-oxidants may represent such a beneficial technology, as an alternative to current anti-oxidants. The level of confidence in the toxicity profile of this technology is high, thanks to the metrics of the QSAR models that were used, as well as supporting data from additional in-vivo testing.