QSAR classification models for the screening of the endocrine-disrupting activity of perfluorinated compounds

Prediction of critical micelle concentration for per- and polyfluoroalkyl substances

In this study, we focus on the development of Quantitative Structure-Property Relationship (QSPR) models to predict the critical micelle concentration (CMC) for per- and polyfluoroalkyl substances (PFASs). Experimental CMC values for both fluorinated and non-fluorinated compounds were meticulously compiled from existing literature sources. Our approach involved constructing two distinct types of models based on Support Vector Machine (SVM) algorithms applied to the dataset. Type (I) models were trained exclusively on CMC values for fluorinated compounds, while Type (II) models were developed utilizing the entire dataset, incorporating both fluorinated and non-fluorinated compounds. Comparative analyses were conducted against reference data, as well as between the two model types. Encouragingly, both types of models exhibited robust predictive capabilities and demonstrated high reliability. Subsequently, the model having the broadest applicability domain was selected to complement the existing experimental data, thereby enhancing our understanding of PFAS behaviour.

Per- and Polyfluoroalkyl Substances (PFAS) in Breast Milk: Concerning Trends for Current-Use PFAS

This is the first study in the last 15 years to analyze per- and polyfluoroalkyl substances (PFAS) in breast milk collected from mothers (n = 50) in the United States, and our findings indicate that both legacy and current-use PFAS now contaminate breast milk, exposing nursing infants. Breast milk was analyzed for 39 PFAS, including 9 short-chain and 30 long-chain compounds, and 16 of these PFAS were detected in 4-100% of the samples. The ∑PFAS concentration in breast milk ranged from 52.0 to 1850 pg/mL with a median concentration of 121 pg/mL. Perfluorooctanesulfonic acid (PFOS) and perfluorooctanoic acid (PFOA) were the most abundant PFAS in these samples (medians 30.4 and 13.9 pg/mL, respectively). Two short-chain PFAS, including perfluoro-n-hexanoic acid (PFHxA, C6) and perfluoro-n-heptanoic acid (PFHpA, C7), were detected in most of the samples with median concentrations of 9.69 and 6.10 pg/mL, respectively. Analysis of the available breast milk PFAS data from around the world over the period of 1996-2019 showed that while the levels of the phased-out PFOS and PFOA have been declining with halving times of 8.1 and 17 years, respectively, the detection frequencies of current-use short-chain PFAS have been increasing with a doubling time of 4.1 years.

Endocrine disruption by PFAS: A major concern associated with legacy and replacement substances

Perand poly-fluorinated alkyl substances (PFAS) have been used for decades in a great variety of processes and products by virtue of their exceptional properties, versatility and chemical stability. Nevertheless, it is increasingly recognized that these substances can represent a serious hazard to human health and living organisms due to their persistence, long-range transport potential and tendency to accumulate in biota. For this reason, some efforts have been made across the EU to identify alternative molecules, with a shorter carbon chain and theoretically safer profile, that might replace the previous generation of legacy PFAS. Unfortunately, this strategy has not been entirely successful and serious concerns are still posed by PFAS in different human populations. Among others, an emerging aspect is represented by the adverse effects that both legacy and alternative PFAS can exert on the human endocrine system, with respect to vulnerable target subpopulations. In this review we will briefly summarize PFAS properties, uses and environmental fate, focusing on their effects on human reproductive capacity and fertility, body weight control and obesity as well as thyroid function.

Exploring QSAR models for assessment of acute fish toxicity of environmental transformation products of pesticides (ETPPs)

Environmental transformation products of pesticides (ETPPs) have a great deal of ecological impact owing to their ability to cause toxicity to the aquatic organisms, which can then be translated to the humans. The limited experimental data on biochemical and toxic effects of ETPPs, the high test costs together with regulatory limitations and the international push to reduce animal testing encourage greater dependence on predictive in silico techniques like quantitative structure-activity relationship (QSAR) models. The aim of the present work was to explore the key structural features, which regulate the toxicity towards fishes, for 85 ETPPs using a partial least squares (PLS) regression based chemometric model developed according to Organisation for Economic Co-operation and Development (OECD) guidelines. The model was extensively validated using both internal and external validation metrics, and the results so obtained justify the reliability and usefulness of the developed model (Q² = 0.648, R²_pred or Q²_F1 = 0.734 and Q²_F2 = 0.733). From the developed model, we can conclude that lipophilicity, polarity, presence of branching and the functional form of O-atom in the transformed structures of pesticides are the important features that are to be considered during ecotoxicity assessment of ETPPs. The information obtained from the descriptors of the developed model could be utilized in the future for assessing ETPPs with the benefit of providing an early warning of their potentially detrimental effect on fishes for regulatory purposes.

Principles of QSAR Modeling

At the end of her academic career, the author summarizes the main aspects of QSAR modeling, giving comments and suggestions according to her 23 years' experience in QSAR research on environmental topics. The focus is mainly on Multiple Linear Regression, particularly Ordinary Least Squares, using a Genetic Algorithm for variable selection from various theoretical molecular descriptors, but the comments can be useful also for other QSAR methods. The need for rigorous validation, also external, and for applicability domain check to guarantee predictivity and reliability of QSAR models is particularly highlighted. The commented approach is the “predictive” one, based on chemometrics, and is usefully applied to the prioritization of environmental pollutants. All the discussed points and the author's ideas are implemented in the software QSARINS, as a legacy to the QSAR community.

Using Machine Learning to Classify Bioactivity for 3486 Per- and Polyfluoroalkyl Substances (PFASs) from the OECD List

A recent OECD report estimated that more than 4000 per- and polyfluorinated alkyl substances (PFASs) have been produced and used in a broad range of industrial and consumer applications. However, little is known about the potential hazards (e.g., bioactivity, bioaccumulation, and toxicity) of most PFASs. Here, we built machine-learning-based quantitative structure-activity relationship (QSAR) models to predict the bioactivity of those PFASs. By examining a number of available molecular data sets, we constructed the first PFAS-specific database that contains the bioactivity information on 1012 PFASs for 26 bioassays. On the basis of the collected PFAS data set, we trained 5 different machine learning models that cover a variety of conventional models (e.g., random forest and multitask neural network (MNN)) and advanced graph-based models (e.g., graph convolutional network). Those models were evaluated based on the validation data set. Both MNN and graph-based models demonstrated the best performance. The average of the best area-under-the-curve score for each bioassay is 0.916. For predictions on the OECD list, most of the biologically active PFASs have perfluoroalkyl chain lengths less than 12 and are categorized into fluorotelomer-related compounds and perfluoroalkyl acids and their precursors.

In vitro and in silico modeling of perfluoroalkyl substances mixture toxicity in an amphibian fibroblast cell line

Poly and perfluoroalkyl substances (PFAS) are a large group of emerging organic pollutants that can persist in the environment and bioaccumulate in biota. They are found in complex mixtures, and although the exact number of PFAS is unknown, it has been estimated to be in the thousands. The objective of this study was two-fold. First, we examined the cytotoxicity of PFAS singly and in binary mixtures using an amphibian fibroblast cell line. Second, we used this experimental data to develop quantitative structure-activity relationship (QSAR) models for single and binary mixtures. We tested the cytotoxicity of four common PFAS: perfluorooctane sulfonate (PFOS); perfluorooctanoic acid (PFOA); perfluorohexane sulfonate (PFHxS); and perfluorohexanoic acid (PFHxA). PFOS was the most toxic and PFHxA the least cytotoxic. Binary mixtures allowed for the construction of isobolograms to test for additivity, synergism, or antagonism. Using this data, QSAR modeling was used for predicting the toxicity of 24 single and 1380 binary mixtures (theoretically generated). Overall, our experimental and modeling results showed that mixtures were approximately additive, with the exception of PFOS and PFOA, which were found to be weakly synergistic. This data shows that certain mixtures of PFAS may have increased toxicity potential above what the simple sum of PFAS concentrations would suggest. More studies are needed that test the toxicity of PFAS mixtures.

Single or mixture halogenated chemicals? Risk assessment and developmental toxicity prediction on zebrafish embryos based on weighted descriptors approach

Halogenated chemicals including perfluoroalkyl substances (PFASs) represent an emerging class of endocrine-disrupting pollutants for human populations across the globe. Distress related to their environmental fate and toxicity has initiated several research projects, but the amount of experimental data available for these pollutants is limited. The objective of this study is to assess the toxicity of potentially "safer" alternatives, in relation to their existing counterparts. Developmental toxicity data on zebrafish (Danio rerio) embryos of single and tertiary halogenated mixtures were modeled employing quantitative structure-toxicity relationship (QSTR) tool. The computed models are then employed for toxicity prediction of theoretically generated binary and tertiary mixtures (which have no experimental data) to check their possible threshold and mode of toxicity for future risk assessment. Further, for toxicity screening, we have prepared a huge external dataset consists of single (24), binary (276) and tertiary (2024) mixtures of PFASs. It was accomplished by combination method and predicted through developed models for interpretation of toxicity threats for individuals and mixtures along with identification of diverse range and combination of toxicity thresholds. We found that chemicals in mixtures displayed concentration addition of individual chemical suggesting a similar mode of toxic action and non-interaction of chemicals. Not only that, mixtures of halogenated compounds including PFASs showed less toxicity than their single counterparts and the obtained toxicity trend is: Single chemical > Binary mixture > Tertiary mixture.

Endocrine-disrupting activity of per- and polyfluoroalkyl substances: Exploring combined approaches of ligand and structure based modeling

Exposure to perfluorinated and polyfluoroalkyl substances (PFCs/PFASs), endocrine disrupting halogenated pollutants, has been linked to various diseases including thyroid toxicity in human populations across the globe. PFASs can compete with thyroxine (T4) for binding to the human thyroid hormone transport protein transthyretin (TTR) which may lead to reduce thyroid hormone levels leading to endocrine disrupting adverse effects. Environmental fate and endocrine-disrupting activity of PFASs has initiated several research projects, but the amount of experimental data available for these pollutants is limited. In this study, experimental data for T4-TTR competing potency of 24 PFASs obtained in a radioligand-binding assay were modeled using classification- and regression-based quantitative structure-activity relationship (QSAR) tools with simple molecular descriptors obtained from chemical structure of these compounds in order to identify the responsible structural features and fragments of the studied PFASs for endocrine disruption activity. Additionally, docking studies were performed employing the crystal structure complex of TTR with bound 2', 6'-difluorobiphenyl-4-carboxylic acid (PDB: 2F7I) in order to constitute the receptor model for human TTR. The results corroborate evidence for these binding interactions and indicate multiple high-affinity modes of binding. The developed in silico models therefore advance our understanding of important structural attributes of these chemicals and may provide important information for the design of future synthesis of PFASs as well as may serve as an efficient query tool for virtual screening of large PFAS databases to check their endocrine toxicity profile.

Different binding mechanisms of neutral and anionic poly-/perfluorinated chemicals to human transthyretin revealed by In silico models

Chemical forms-dependent binding interactions between phenolic compounds and human transthyretin (hTTR) have been elaborated previously. However, it is not known whether the binding interactions between ionizable halogenated alphatic compounds and hTTR also have the same manner. In this study, poly-/perfluorinated chemicals (PFCs) were selected as model compounds and molecular dynamic simulation was performed to investigate the binding mechanisms between PFCs and hTTR. Results show the binding interactions between the halogenated aliphatic compounds and hTTR are related to the chemical forms. The ionized groups of PFCs can form electrostatic interactions with the -NH⁺₃ groups of Lys 15 residues in hTTR and form hydrogen bonds with the residues of hTTR. By analyzing the molecular orbital energies of PFCs, we also found that the anionic groups (nucleophile) in PFCs could form electron donor - acceptor interactions with the -NH⁺₃ groups (electrophile) in Lys 15. The aforementioned orientational interactions make the ionized groups of the PFCs point toward the entry port of the binding site. The roles of fluorine atoms in the binding interactions were also explored. The fluorine atoms can influence the binding interactions via inductive effects. Appropriate molecular descriptors were selected to characterize these interactions, and two quantitative structure-activity relationship models were developed.

A new adaptive L1-norm for optimal descriptor selection of high-dimensional QSAR classification model for anti-hepatitis C virus activity of thiourea derivatives

A high-dimensional quantitative structure-activity relationship (QSAR) classification model typically contains a large number of irrelevant and redundant descriptors. In this paper, a new design of descriptor selection for the QSAR classification model estimation method is proposed by adding a new weight inside L1-norm. The experimental results of classifying the anti-hepatitis C virus activity of thiourea derivatives demonstrate that the proposed descriptor selection method in the QSAR classification model performs effectively and competitively compared with other existing penalized methods in terms of classification performance on both the training and the testing datasets. Moreover, it is noteworthy that the results obtained in terms of stability test and applicability domain provide a robust QSAR classification model. It is evident from the results that the developed QSAR classification model could conceivably be employed for further high-dimensional QSAR classification studies.

Binding interactions of perfluoroalkyl substances with thyroid hormone transport proteins and potential toxicological implications

Perfluoroalkyl substances (PFASs) have been shown to cause abnormal levels of thyroid hormones (THs) in experimental animals, but the molecular mechanism is poorly understood. Here, a fluorescence displacement assay was used to determine the binding affinities of 16 PFASs with two major TH transport proteins, transthyretin (TTR) and thyroxine-binding globulin (TBG). Most of the tested PFASs bound TTR with relative potency (RP) values of 3×10(-4) to 0.24 when compared with that of the natural ligand thyroxine, whereas fluorotelomer alcohols did not bind. Only perfluorotridecanoic acid and perfluorotetradecanoic acid bound TBG, with RP values of 2×10(-4) when compared with that of thyroxine. Based on these results, it was estimated that displacement of T4 from TTR by perfluorooctane sulfonate and perfluorooctanoic acids would be significant for the occupationally exposed workers but not the general population. Structure-binding analysis revealed that PFASs with a medium chain length and a sulfonate acid group are optimal for TTR binding, and PFASs with lengths longer than 12 carbons are optimal for TBG binding. Three mutant proteins were prepared to examine crucial residues involved in the binding of PFASs to TH transport proteins. TTR with a K15G mutation and TBG with either a R378G or R381G mutation showed decreased binding affinity to PFASs, indicating that these residues play key roles in the interaction with the compounds. Molecular docking showed that the PFASs bind to TTR with their acid group forming a hydrogen bond with K15 and the hydrophobic chain towards the interior. PFASs were modeled to bind TBG with their acid group forming a hydrogen bond with R381 and the hydrophobic chain extending towards R378. The findings aid our understanding of the behavior and toxicity of PFASs on the thyroid hormone system.

Investigation of the influence of protein corona composition on gold nanoparticle bioactivity using machine learning approaches

The understanding of the mechanisms and interactions that occur when nanomaterials enter biological systems is important to improve their future use. The adsorption of proteins from biological fluids in a physiological environment to form a corona on the surface of nanoparticles represents a key step that influences nanoparticle behaviour. In this study, the quantitative description of the composition of the protein corona was used to study the effect on cell association induced by 84 surface-modified gold nanoparticles of different sizes. Quantitative relationships between the protein corona and the activity of the gold nanoparticles were modelled by using several machine learning-based linear and non-linear approaches. Models based on a selection of only six serum proteins had robust and predictive results. The Projection Pursuit Regression method had the best performances (r(2) = 0.91; Q(2)loo = 0.81; r(2)ext = 0.79). The present study confirmed the utility of protein corona composition to predict the bioactivity of gold nanoparticles and identified the main proteins that act as promoters or inhibitors of cell association. In addition, the comparison of several techniques showed which strategies offer the best results in prediction and could be used to support new toxicological studies on gold-based nanomaterials.

Use of computational modeling approaches in studying the binding interactions of compounds with human estrogen receptors

Estrogens have a whole host of physiological functions in many human organs and systems, including the reproductive, cardiovascular, and central nervous systems. Many naturally-occurring compounds with estrogenic or antiestrogenic activity are present in our environment and food sources. Synthetic estrogens and antiestrogens are also important therapeutic agents. At the molecular level, estrogen receptors (ERs) mediate most of the well-known actions of estrogens. Given recent advances in computational modeling tools, it is now highly practical to use these tools to study the interaction of human ERs with various types of ligands. There are two common categories of modeling techniques: one is the quantitative structure activity relationship (QSAR) analysis, which uses the structural information of the interacting ligands to predict the binding site properties of a macromolecule, and the other one is molecular docking-based computational analysis, which uses the 3-dimensional structural information of both the ligands and the receptor to predict the binding interaction. In this review, we discuss recent results that employed these and other related computational modeling approaches to characterize the binding interaction of various estrogens and antiestrogens with the human ERs. These examples clearly demonstrate that the computational modeling approaches, when used in combination with other experimental methods, are powerful tools that can precisely predict the binding interaction of various estrogenic ligands and their derivatives with the human ERs.

Structure-activity relationship study of trifluoromethylketone inhibitors of insect juvenile hormone esterase: comparison of several classification methods

Juvenile hormone esterase (JHE) plays a key role in the development and metamorphosis of holometabolous insects. Its inhibitors could possibly be targeted for insect control. Conversely, JHE may also be involved in endocrine disruption by xenobiotics, resulting in detrimental effects in beneficial insects. There is therefore a need to know the structural characteristics of the molecules able to monitor JHE activity, and to develop SAR and QSAR studies to estimate their effectiveness. For a large diverse population of 181 trifluoromethylketones (TFKs) - the most potent JHE inhibitors known to date - we recently proposed a binary classification (active/inactive) using a support vector machine and Codessa structural descriptors. We have now examined, using the same data set and with the same descriptors, the applicability and performance of five other machine learning approaches. These have been shown able to handle high dimensional data (with descriptors possibly irrelevant or redundant) and to cope with complex mechanisms, but without delivering explicit directly exploitable models. Splitting the data into five batches (training set 80%, test set 20%) and carrying out leave-one-out cross-validation, led to good results of comparable performance, consistent with our previous support vector classifier (SVC) results. Accuracy was greater than 0.80 for all approaches. A reduced set of 15 descriptors common to all the investigated approaches showed good predictive ability (confirmed using a three-layer perceptron) and gives some clues regarding a mechanistic interpretation.

Structure-activity relationship study of trifluoromethylketones: inhibitors of insect juvenile hormone esterase

The juvenile hormone esterase (JHE) regulates juvenile hormone titre in insect hemolymph during its larval development. It has been suggested that JHE could be targeted for use in insect control. This enzyme can also be considered as involved in the phenomenon of endocrine disruption by xenobiotics in beneficial insects. Consequently, there is a need to know the characteristics of the molecules able to act on the JHE. Trifluoromethylketones (TFKs) are the most potent JHE inhibitors found to date and different quantitative structure-activity relationships (QSARs) have been derived for this group of chemicals. In this context, a set of 181 TFKs (118 active and 63 inactive compounds), tested on Trichoplusia ni for their JHE inhibition activity and described by physico-chemical descriptors, was split into different training and test sets to derive structure-activity relationship (SAR) models from support vector classification (SVC). A SVC model including 88 descriptors and derived from a Gaussian kernel was selected for its predictive performances. Another model computed only with 13 descriptors was also selected due to its mechanistic interpretability. This study clearly illustrates the difficulty in capturing the essential structural characteristics of the TFKs explaining their JHE inhibitory activity.

QSAR prediction of the competitive interaction of emerging halogenated pollutants with human transthyretin

The determination of the potential endocrine disruption (ED) activity of chemicals such as poly/perfluorinated compounds (PFCs) and brominated flame retardants (BFRs) is still hindered by a limited availability of experimental data. Quantitative structure-activity relationship (QSAR) strategies can be applied to fill this data gap, help in the characterization of the ED potential, and screen PFCs and BFRs with a hazardous toxicological profile. This paper proposes the modelling of T4-TTR (thyroxin-transthyretin) competing potency and relative binding potency toward T4 (logT4-REP) of PFCs and BFRs by regression and classification QSAR models. This study is a follow up of a former work, which analysed separately the interaction of BFRs and PFCs with the carrier TTR. The new results demonstrate the possibility of developing robust and predictive QSARs, which include both BFRs and PFCs in the training set, obtaining larger applicability domains than the existing models developed separately for BFRs and PFCs. The selection of modelling molecular descriptors confirms the importance of structural features, such as the aromatic OH or the molecular length, to increase the binding of the studied chemicals to TTR. Additionally, the need of experimental tests for some chemicals, and in particular for some of the BFRs, is highlighted.