A simple method for estimating in vitro air-tissue and in vivo blood-tissue partition coefficients

In Silico Acute Aquatic Hazard Assessment and Prioritization Using a Grouped Target Site Model: A Case Study of Organic Substances Reported in Permian Basin Hydraulic Fracturing Operations

Hydraulic fracturing (HF) is commonly used to enhance onshore recovery of oil and gas during production. This process involves the use of a variety of chemicals to support the physical extraction of oil and gas, maintain appropriate conditions downhole (e.g., redox conditions, pH), and limit microbial growth. The diversity of chemicals used in HF presents a significant challenge for risk assessment. The objective of the present study is to establish a transparent, reproducible procedure for estimating 5th percentile acute aquatic hazard concentrations (e.g., acute hazard concentration 5th percentiles [HC5s]) for these substances and validating against existing toxicity data. A simplified, grouped target site model (gTSM) was developed using a database (n = 1696) of diverse compounds with known mode of action (MoA) information. Statistical significance testing was employed to reduce model complexity by combining 11 discrete MoAs into three general hazard groups. The new model was trained and validated using an 80:20 allocation of the experimental database. The gTSM predicts toxicity using a combination of target site water partition coefficients and hazard group-based critical target site concentrations. Model performance was comparable to the original TSM using 40% fewer parameters. Model predictions were judged to be sufficiently reliable and the gTSM was further used to prioritize a subset of reported Permian Basin HF substances for risk evaluation. The gTSM was applied to predict hazard groups, species acute toxicity, and acute HC5s for 186 organic compounds (neutral and ionic). Toxicity predictions and acute HC5 estimates were validated against measured acute toxicity data compiled for HF substances. This case study supports the gTSM as an efficient, cost-effective computational tool for rapid aquatic hazard assessment of diverse organic chemicals. Environ Toxicol Chem 2024;00:1-12. © 2024 ExxonMobil Petroleum and Chemical BV. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC.

Multimodal Model to Predict Tissue-to-Blood Partition Coefficients of Chemicals in Mammals and Fish

Tissue-to-blood partition coefficients (Ptb) are key parameters for assessing toxicokinetics of xenobiotics in organisms, yet their experimental data were lacking. Experimental methods for measuring Ptb values are inefficient, underscoring the urgent need for prediction models. However, most existing models failed to fully exploit Ptb data from diverse sources, and their applicability domain (AD) was limited. The current study developed a multimodal model capable of processing and integrating textual (categorical features) and numerical information (molecular descriptors/fingerprints) to simultaneously predict Ptb values across various species, tissues, blood matrices, and measurement methods. Artificial neural network algorithms with embedding layers were used for the multimodal modeling. The corresponding unimodal models were developed for comparison. Results showed that the multimodal model outperformed unimodal models. To enhance the reliability of the model, a method considering categorical features, weighted molecular similarity density, and weighted inconsistency in molecular activities of structure-activity landscapes was used to characterize the AD. The model constrained by the AD exhibited better prediction accuracy for the validation set, with the determination coefficient, root mean-square error, and mean absolute error being 0.843, 0.276, and 0.213 log units, respectively. The multimodal model coupled with the AD characterization can serve as an efficient tool for internal exposure assessment of chemicals.

Multi-task machine learning models for simultaneous prediction of tissue-to-blood partition coefficients of chemicals in mammals

Tissue-to-blood partition coefficients (Ptb) are crucial for assessing the distribution of chemicals in organisms. Given the lack of experimental data and laborious nature of experimental methods, there is an urgent need to develop efficient predictive models. With the help of machine learning algorithms, i,e., random forest (RF), and artificial neural network (ANN), this study developed multi-task (MT) models that can simultaneously predict Ptb values for various mammalian tissues, including liver, muscle, brain, lung, and adipose. Single-task (ST) models using partial least squares regression, RF, and ANN algorithms for each endpoint were established for comparison. Overall, the performances of MT models were superior to those of ST models. The MT model using ANN algorithms showed the highest prediction accuracy with determination coefficients ranging from 0.704 to 0.886, root mean square errors between 0.223 and 0.410, and mean absolute errors ranging from 0.178 to 0.285 log units. Results showed that lipophilicity and polarizability of molecules significantly influence their partition behavior in organisms. Applicability domains (ADs) of the models were characterized by weighted molecular similarity density, and weighted inconsistency in molecular activities of structure-activity landscapes. When constrained by ADs, the models displayed enhanced predictive accuracy, making them valuable tools for the risk assessment and management of chemicals.

Recent advances for estimating environmental properties for small molecules from chromatographic measurements and the solvation parameter model

The transfer of neutral compounds between immiscible phases in chromatographic or environmental systems can be described by six solute properties (solute descriptors) using the solvation parameter model. The solute descriptors are size (McGowan's characteristic volume), V, excess molar refraction, E, dipolarity/polarizability, S, hydrogen-bond acidity and basicity, A and B, and the gas-liquid partition constant on n-hexadecane at 298.15 K, L. V and E for liquids are accessible by calculation but the other descriptors and E for solids are determined experimentally by chromatographic, liquid-liquid partition, and solubility measurements. These solute descriptors are available for several thousand compounds in the Abraham solute descriptor databases and for several hundred compounds in the WSU experimental solute descriptor database. In the first part of this review, we highlight features important in defining each descriptor, their experimental determination, compare descriptor quality for the two organized descriptor databases, and methods for estimating Abraham solute descriptors. In the second part we focus on recent applications of the solvation parameter model to characterize environmental systems and its use for the identification of surrogate chromatographic models for estimating environmental properties.

Descriptors for some Compounds with Pharmacological Activity; Calculation of Properties

Abraham model solute descriptors have been determined for nisoldipine, nizatidine, loratadine, zonisamide, oxaprozin, rebamipide, domperidone, temozolomide, 'florfenicol', florfenicol A, dapsone, chrysin, benorilate, β-lapachone, and Ipriflavone based on published partition coefficients, molar solubilities and gas chromatographic retention indices. The calculated solute descriptors, combined with our previously published Abraham model correlations, are used to predict several important physicochemical and biological properties, such as air-water, air-blood, air-lung, air-fat, air-skin, water-lipid, water-membrane and water-skin partition coefficients, as well as permeation from water through skin.

Abraham Model Descriptors for Melatonin; Prediction of Solution, Biological and Thermodynamic Properties

Literature solubilities have been used to obtain properties or descriptors of melatonin. These indicate the chemical nature of melatonin: it is dipolar and has moderate hydrogen bond acidity and hydrogen bond basicity. The descriptors can be combined with equations that we have previously constructed to estimate water–solvent partition coefficients and solubilities in a huge number of organic solvents. In the same way, a range of biological properties can be estimated. These include blood–tissue partitions, water–skin partition and permeability through skin.

Development and intercomparison of single and multicompartment physiologically-based toxicokinetic models: Implications for model selection and tiered modeling frameworks

This study describes the development and intercomparison of generic physiologically-based toxicokinetic (PBTK) models for humans comprised of internally consistent one-compartment (1Co-) and multi-compartment (MCo-) implementations (G-PBTK). The G-PBTK models were parameterized for an adult male (70 kg) using common physiological parameters and in vitro biotransformation rate estimates and subsequently evaluated using independent concentration versus time data (n = 6) and total elimination half-lives (n = 15) for diverse organic chemicals. The model performance is acceptable considering the inherent uncertainty in the biotransformation rate data and the absence of model calibration. The G-PBTK model was then applied using hypothetical neutral organics, acidic ionizable organics and basic ionizable organics (IOCs) to identify combinations of partitioning properties and biotransformation rates leading to substantial discrepancies between 1Co- and MCo-PBTK calculations for whole body concentrations and half-lives. The 1Co- and MCo-PBTK model calculations for key toxicokinetic parameters are broadly consistent unless biotransformation is rapid (e.g., half-life less than five days). When half-lives are relatively short, discrepancies are greatest for the neutral organics and least for the acidic IOCs which follows from the estimated volumes of distribution (e.g., VD_SS = 9.6-15.4 L/kg vs 0.3-1.6 L/kg for the neutral and acidic compounds respectively) and the related approach to internal chemical equilibrium. The model intercomparisons demonstrate that 1Co-PBTK models can be applied with confidence to many exposure scenarios, particularly those focused on chronic or repeat exposures and for prioritization and screening-level decision contexts. However, MCo-PBTK models may be necessary in certain contexts, particularly for intermittent, short-term and highly variable exposures. A key recommendation to guide model selection and the development of tiered PBTK modeling frameworks that emerges from this study is the need to harmonize models with respect to parameterization and process descriptions to the greatest extent possible when proceeding from the application of simpler to more complex modeling tools as part of chemical assessment activities.

A dynamic model of the body gas stores for carbon dioxide, oxygen, and inert gases that incorporates circulatory transport delays to and from the lung

Many models of the body’s gas stores have been generated for specific purposes. Here, we seek to produce a more general purpose model that: 1) is relevant for both respiratory (CO₂ and O₂) and inert gases; 2) is based firmly on anatomy and not arbitrary compartments; 3) can be scaled to individuals; and 4) incorporates arterial and venous circulatory delays as well as tissue volumes so that it can reflect rapid transients with greater precision. First, a “standard man” of 11 compartments was produced, based on data compiled by the International Radiation Protection Commission. Each compartment was supplied via its own parallel circulation, the arterial and venous volumes of which were based on reported tissue blood volumes together with data from a detailed anatomical model for the large arteries and veins. A previously published model was used for the blood gas chemistry of CO₂ and O₂. It was not permissible ethically to insert pulmonary artery catheters into healthy volunteers for model validation. Therefore, validation was undertaken by comparing model predictions with previously published data and by comparing model predictions with experimental data for transients in gas exchange at the mouth following changes in alveolar gas composition. Overall, model transients were fastest for O₂, intermediate for CO₂, and slowest for N₂. There was good agreement between model estimates and experimentally measured data. Potential applications of the model include estimation of closed-loop gain for the ventilatory chemoreflexes and improving the precision associated with multibreath washout testing and respiratory measurement of cardiac output.

NEW & NOTEWORTHY A model for the body gas stores has been generated that is applicable to both respiratory gases (CO₂ and O₂) and inert gases. It is based on anatomical details for organ volumes and blood contents together with anatomical details of the large arteries. It can be scaled to the body size and composition of different individuals. The model enables mixed venous gas compositions to be predicted from the systemic arterial compositions.

Predicting plant cuticle-water partition coefficients for organic pollutants using pp-LFER model

Predicting plant cuticle-water partition coefficients (K_cw) and understanding the partition mechanisms are crucial to assess environmental fate and risk of organic pollutants. Up to now, experimental K_cw values are determined for only hundreds of compounds because of high experimental cost. For this reason, computational models, which can predict K_cw values based on chemical structures, are promising approaches to evaluate new compounds. In this study, a large dataset consisting of 279 logK_cw values for 125 unique compounds were collected and curated. A poly-parameter linear free energy relationship (pp-LFER) model was developed with stepwise multiple linear regression based on this dataset. The resulted pp-LFER model has good predictability and robustness as indicated by determination coefficient (R²_adj,tra) of 0.93, bootstrapping coefficient (Q²_BOOT) of 0.92, external validation coefficient (Q²_ext) of 0.94 and root mean square error of 0.52 log units. Contribution analysis of different interactions indicated that dispersion and hydrophobic interactions have the highest positive contribution (56%) to increase the partition of pollutants onto plant cuticles. In addition, for organic pollutions containing benzene ring (13-31%), double bond (9-17%) or nitrogen-containing heterocycles (9-17%), π/n-electron pairs interactions exhibit obvious positive contributions to logK_cw. In conclusion, the proposed pp-LFER model is beneficial for predicting logK_cw of potential organic pollutants directly from their molecular structures.

Target site model: Predicting mode of action and aquatic organism acute toxicity using Abraham parameters and feature-weighted k-nearest neighbors classification

A database of 1480 chemicals with 47 associated modes of action compiled from the literature encompasses a wide range of chemical classes (alkanes, polycyclic aromatic hydrocarbons, pesticides, and polar compounds) and includes toxicity data for 79 different aquatic genera. The data were split into a calibration group and a validation group (80/20) to apply k-nearest neighbors (k-NN) methodology to predict the toxic mode of action for the compound. Other approaches were tested (support vector machines and linear discriminant analysis) as well as variations in the k-NN technique (distance weighting, feature weighting). Best-prediction results were found with k = 3, in a voting platform with optimized feature weighting. Using the predicted mode of action, the appropriate polyparameter target site model for that mode of action is applied to calculate the 50% lethal concentration (LC50). Predicted LC50s for the validation database resulted in a root-mean squared error (RMSE) of 0.752. This can be compared to an RMSE of 0.655 for the same validation set using the reference mode of action labels. The complete database resulted in an RMSE of 0.793 for reference mode of action labels. This confirms that the classification model has sufficient accuracy for predicting the mode of action and for determining toxicity using the target site model. Environ Toxicol Chem 2019;38:375-386. © 2018 SETAC.

Can poly-parameter linear-free energy relationships (pp-LFERs) improve modelling bioaccumulation in fish?

A wide range of studies have characterized different types of biosorbent, with regard to their interactions with chemicals. This has resulted in the development of poly-parameter linear free energy relationships (pp-LFERs) for the estimation of partitioning of neutral organic compounds to biological phases (e.g., storage lipids, phospholipids and serum albumins). The aims of this study were to explore and evaluate the influence of implementing pp-LFERs both into a one-compartment fish model and a multi-compartment physiologically based toxicokinetic (PBTK) fish model and the associated implications for chemical risk assessment. For this purpose, fish was used as reference biota, due to their important role in aquatic food chains and dietary exposure to humans. The bioconcentration factor (BCF) was utilized as the evaluation metric. Overall, our results indicated that models incorporating pp-LFERs (R² = 0.75) slightly outperformed the single parameter (sp) LFERs approach in the one-compartmental fish model (R² = 0.72). A pronounced enhancement was achieved for compounds with log K_OW between 4 and 5 with increased R² from 0.52 to 0.71. The minimal improvement was caused by the overestimation of lipid contribution and underestimation of protein contribution by the sp-approach, which cancelled each other out. Meanwhile, a greater improvement was observed for multi-compartmental PBTK models with consideration of metabolism, making all predictions fall within a factor of 10 compared with measured data. For screening purposes, the K_OW-based (sp-LFERs) approach should be sufficient to quantify the main partitioning characteristics. Further developments are required for the consideration of ionization and more accurate quantification of biotransformation in biota.

Descriptors for Pentane-2,4-dione and Its Derivatives

We have used equations for partition coefficients of compounds from water and the gas phase to various solvents to obtain descriptors for pentane-2,4-dione and 21 of its derivatives. These descriptors can then be used to estimate further partition coefficients into a wide variety of solvents. The descriptors also yield information about the properties of pentane-2,4-dione and its derivatives. Pentane-2,4-dione and its alkyl derivatives are quite polar, with substantial hydrogen bond basicity but with no hydrogen bond acidity. In contrast 1,1,1-trifluoropentane-2,4-dione and hexafluoropentan-2,4-dione have significant hydrogen bond acidities.

A simple and facile NMR method for the determination of hydrogen bonding by amide N–H protons in protein models and other compounds

It is shown that measurements of the ¹H chemical shifts of amide N–H protons in chloroform and in DMSO solvents are sufficient to determine the extent of hydrogen bonding of the N–H protons in a variety of compounds containing the amide group.

Equations for water-triolein partition coefficients for neutral species; comparison with other water-solvent partitions, and environmental and toxicological processes

Linear free energy relationships, LFERs, have been constructed for water-triolein partition coefficients for neutral species. It is shown that separate equations are required for wet and dry triolein. From a comparison of the equation coefficients for water-wet triolein with those for 52 other water-solvent systems it is shown that there is little correspondence between triolein and any of the 52 other solvents - only the water-isopropyl myristate system is close to the water-wet triolen system. A comparison of equation coefficients for the water-wet triolein system with LFER coefficients of 16 environmentally important processes shows that wet triolein is not a suitable model for any of the processes, although a number of other water-solvent systems are possible models for some of the environmental processes. A comparison of LFER coefficients with those of 17 aqueous toxicological processes reveals that most of the water-solvent systems, including water-wet triolein, will be poor models for any of the toxicological systems, but the water-lower alcohol systems show promise as models for a number of the toxicological systems. Our method of comparison of coefficients for LFERs that have exactly the same independent variables can be extended to various other types of system.

Analysis of the distribution of organic compounds and drugs between biological tissues in the framework of solute partitioning in aqueous two-phase systems

Distribution of organic compounds between different biological tissues may be considered in the framework of solute partitioning in aqueous two-phase systems.

Solvation descriptors for porphyrins (porphines)

Descriptors for porphyrin show that it is dipolar, a weak hydrogen-bond acid, a moderately strong hydrogen-bond base, and very hydrophobic.

Using water-solvent systems to estimate in vivo blood-tissue partition coefficients

BACKGROUND

Blood-tissue partition coefficients indicate how a chemical will distribute throughout the body and are an important part of any pharmacokinetic study. They can be used to assess potential toxicological effects from exposure to chemicals and the efficacy of potential novel drugs designed to target certain organs or the central nervous system. In vivo measurement of blood-tissue partition coefficients is often complicated, time-consuming, and relatively expensive, so developing in vitro systems that approximate in vivo ones is desirable. We have determined such systems for tissues such as brain, muscle, liver, lung, kidney, heart, skin, and fat.

RESULTS

Several good (p < 0.05) blood-tissue partition coefficient models were developed using a single water-solvent system. These include blood-brain, blood-lung, blood-heart, blood-fat, blood-skin, water-skin, and skin permeation. Many of these partition coefficients have multiple water-solvent systems that can be used as models. Several solvents-methylcyclohexane, 1,9-decadiene, and 2,2,2-trifluoroethanol-were common to multiple models and thus a single measurement can be used to estimate multiple blood-tissue partition coefficients. A few blood-tissue systems require a combination of two water-solvent partition coefficient measurements to model well (p < 0.01), namely: blood-muscle: chloroform and dibutyl ether, blood-liver: N-methyl-2-piperidone and ethanol/water (60:40) volume, and blood-kidney: DMSO and ethanol/water (20:80) volume.

CONCLUSION

In vivo blood-tissue partition coefficients can be easily estimated through water-solvent partition coefficient measurements.Graphical abstract:Predicted blood-brain barrier partition coefficients coloured by measured log BB value.

Response to "comment on 'structural determinants of drug partitioning in surrogates of phosphatidylcholine bilayer strata'"

We used the solvatochromic correlation to explain the influence of characteristics of studied compounds on the partition coefficients (P) measured using n-hexadecane (C16) and the novel headgroup surrogate (diacetyl phosphatidylcholine, DAcPC), and compared them with those in other systems, including the C16/water (W) system. The comment analyzes why our correlation for the C16/W system has the standard deviation (SD) higher than that published previously. The main reason is that in our, much smaller, data set the measured P values are complemented by the P values predicted by a reliable, unrelated method. We believe that this approach is acceptable for the aforementioned comparison. We did not use just experimental values, as suggested in the comment, because the solvatochromic correlation, although exhibiting 35% reduction in the SD, was accompanied by a sign change of one of the regression coefficients. The recommended use of special solvatochromic solute characteristics for a few compounds and replacement of a predicted PC16/W value by the experimental value resulted in improved correlations. The observed differences between our correlation and those published in the comment and in a previous article do not affect our main conclusions regarding the solvation of solutes in the surrogates (DAcPC and C16) of intrabilayer strata.