High-throughput prediction of blood-brain partitioning: a thermodynamic approach

Combined Micellar Liquid Chromatography Technique and QSARs Modeling in Predicting the Blood-Brain Barrier Permeation of Heterocyclic Drug-like Compounds

The quantitative structure-activity relationship (QSAR) methodology was used to predict the blood-brain permeability (log BB) for 65 synthetic heterocyclic compounds tested as promising drug candidates. The compounds were characterized by different descriptors: lipophilicity, parachor, polarizability, molecular weight, number of hydrogen bond acceptors, number of rotatable bonds, and polar surface area. Lipophilic properties of the compounds were evaluated experimentally by micellar liquid chromatography (MLC). In the experiments, sodium dodecyl sulfate (SDS) as the effluent component and the ODS-2 column were used. Using multiple linear regression and leave-one-out cross-validation, we derived the statistically significant and highly predictive quantitative structure-activity relationship models. Thus, this study provides valuable information on the expected properties of the substances that can be used as a support tool in the design of new therapeutic agents.

Development of QSAR models to predict blood-brain barrier permeability

Assessing drug permeability across the blood-brain barrier (BBB) is important when evaluating the abuse potential of new pharmaceuticals as well as developing novel therapeutics that target central nervous system disorders. One of the gold-standard in vivo methods for determining BBB permeability is rodent log BB; however, like most in vivo methods, it is time-consuming and expensive. In the present study, two statistical-based quantitative structure-activity relationship (QSAR) models were developed to predict BBB permeability of drugs based on their chemical structure. The in vivo BBB permeability data were harvested for 921 compounds from publicly available literature, non-proprietary drug approval packages, and University of Washington's Drug Interaction Database. The cross-validation performance statistics for the BBB models ranged from 82 to 85% in sensitivity and 80-83% in negative predictivity. Additionally, the performance of newly developed models was assessed using an external validation set comprised of 83 chemicals. Overall, performance of individual models ranged from 70 to 75% in sensitivity, 70-72% in negative predictivity, and 78-86% in coverage. The predictive performance was further improved to 93% in coverage by combining predictions across the two software programs. These new models can be rapidly deployed to predict blood brain barrier permeability of pharmaceutical candidates and reduce the use of experimental animals.

Effect of tricyclic 1,2-thiazine derivatives in neuroinflammation induced by preincubation with lipopolysaccharide or coculturing with microglia-like cells

Background

Alzheimer’s disease (AD) is considered the most common cause of dementia among the elderly. One of the modifiable causes of AD is neuroinflammation. The current study aimed to investigate the influence of new tricyclic 1,2-thiazine derivatives on in vitro model of neuroinflammation and their ability to cross the blood–brain barrier (BBB).

Methods

The potential anti-inflammatory effect of new tricyclic 1,2-thiazine derivatives (TP1, TP4, TP5, TP6, TP7, TP8, TP9, TP10) was assessed in SH-SY5Y cells differentiated to the neuron-like phenotype incubated with bacterial lipopolysaccharide (5 or 50 μg/ml) or THP-1 microglial cell culture supernatant using MTT, DCF-DA, Griess, and fast halo (FHA) assays. Additionally, for cultures preincubated with 50 µg/ml lipopolysaccharide (LPS), a cyclooxygenase (COX) activity assay was performed. Finally, the potential ability of tested compounds to cross the BBB was evaluated by computational studies. Molecular docking was performed with the TLR4/MD-2 complex to assess the possibility of binding the tested compounds in the LPS binding pocket. Prediction of ADMET parameters (absorption, distribution, metabolism, excretion and toxicity) was also conducted.

Results

The unfavorable effect of LPS and co-culture with THP-1 cells on neuronal cell viability was counteracted with TP1 and TP4 in all tested concentrations. Tested compounds reduced the oxidative and nitrosative stress induced by both LPS and microglia activation and also reduced DNA damage. Furthermore, new derivatives inhibited total COX activity. Additionally, new compounds would cross the BBB with high probability and reach concentrations in the brain not lower than in the serum. The binding affinity at the TLR4/MD-2 complex binding site of TP4 and TP8 compounds is similar to that of the drug donepezil used in Alzheimer's disease. The ADMET analysis showed that the tested compounds should not be toxic and should show high intestinal absorption.

Conclusions

New tricyclic 1,2-thiazine derivatives exert a neuroregenerative effect in the neuroinflammation model, presumably via their inhibitory influence on COX activity and reduction of oxidative and nitrosative stress.

Supplementary Information

The online version contains supplementary material available at 10.1007/s43440-022-00414-8.

Effect of pyrrolo[3,4-d]pyridazinone derivatives in neuroinflammation induced by preincubation with lipopolysaccharide or coculturing with microglia-like cells

Alzheimer's disease is one of the most serious disorders of the 21st century. There is still no effective therapy for this condition. The study investigated the potential regenerative effect of four pyrrolo[3,4-d]pyridazinone derivatives in cultures of SH-SY5Y neuron-like cells preincubated with lipopolysaccharide (LPS) or cocultured with microglia-like cells. In addition to the traditional investigation of the effect on viability, the level of free radicals and nitric oxide, the average length of neurites was also measured. Via in silico studies, the possibility of penetration of the blood-brain barrier (BBB) by the tested compounds was assessed. The administration of LPS to the culture of SH-SY5Y cells as well as coculturing with microglia-like cells had a significant negative effect on the results of all the assays performed. The treatment with the tested derivatives in most cases significantly reduced this negative effect. The obtained results suggest that the compound L2 may have a beneficial impact on neuronal damage caused by LPS or proinflammatory cytokines secreted by microglia-like cells. Importantly, tested compounds can pass through the BBB, which allows them to enter the brain.

Biomimetic Chromatographic Studies Combined with the Computational Approach to Investigate the Ability of Triterpenoid Saponins of Plant Origin to Cross the Blood-Brain Barrier

Biomimetic (non-cell based in vitro) and computational (in silico) studies are commonly used as screening tests in laboratory practice in the first stages of an experiment on biologically active compounds (potential drugs) and constitute an important step in the research on the drug design process. The main aim of this study was to evaluate the ability of triterpenoid saponins of plant origin to cross the blood-brain barrier (BBB) using both computational methods, including QSAR methodology, and biomimetic chromatographic methods, i.e., High Performance Liquid Chromatography (HPLC) with Immobilized Artificial Membrane (IAM) and cholesterol (CHOL) stationary phases, as well as Bio-partitioning Micellar Chromatography (BMC). The tested compounds were as follows: arjunic acid (Terminalia arjuna), akebia saponin D (Akebia quinata), bacoside A (Bacopa monnieri) and platycodin D (Platycodon grandiflorum). The pharmacokinetic BBB parameters calculated in silico show that three of the four substances, i.e., arjunic acid, akebia saponin D, and bacoside A exhibit similar values of brain/plasma equilibration rate expressed as logPS_Fubrain (the average logPS_Fubrain: -5.03), whereas the logPS_Fubrain value for platycodin D is -9.0. Platycodin D also shows the highest value of the unbound fraction in the brain obtained using the examined compounds (0.98). In these studies, it was found out for the first time that the logarithm of the analyte-micelle association constant (logK_MA) calculated based on Foley's equation can describe the passage of substances through the BBB. The most similar logBB values were obtained for hydrophilic platycodin D, applying both biomimetic and computational methods. All of the obtained logBB values and physicochemical parameters of the molecule indicate that platycodin D does not cross the BBB (the average logBB: -1.681), even though the in silico estimated value of the fraction unbound in plasma is relatively high (0.52). As far as it is known, this is the first paper that shows the applicability of biomimetic chromatographic methods in predicting the penetration of triterpenoid saponins through the BBB.

In Silico Studies on Triterpenoid Saponins Permeation through the Blood-Brain Barrier Combined with Postmortem Research on the Brain Tissues of Mice Affected by Astragaloside IV Administration

As the number of central nervous system (CNS) drug candidates is constantly growing, there is a strong need for precise a priori prediction of whether an administered compound is able to cross the blood–brain barrier (BBB). The aim of this study was to evaluate the ability to cross the BBB of triterpenoid saponins occurring in Astragalus mongholicus roots. The research was carried out using in silico methods combined with postmortem studies on the brain tissues of mice treated with isolated astragaloside IV (AIV). Firstly, to estimate the ability to cross the BBB by the tested saponins, new quantitative structure–activity relationship (QSAR) models were established. The reliability and predictability of the model based on the values of the blood–brain barrier penetration descriptor (logBB), the difference between the n-octanol/water and cyclohexane/water logP (ΔlogP), the logarithm of n-octanol/water partition coefficient (logP_ow), and the excess molar refraction (E) were both confirmed using the applicability domain (AD). The critical leverage value h* was found to be 0.128. The relationships between the standardized residuals and the leverages were investigated here. The application of an in vitro acetylcholinesterase-inhibition test showed that AIV can be recognized as the strongest inhibitor among the tested compounds. Therefore, it was isolated for the postmortem studies on brain tissues and blood using semi-preparative HPLC with the mobile phase composed of water, methanol, and ethyl acetate (1.7:2.1:16.2 v/v/v). The results of the postmortem studies on the brain tissues show a regular dependence of the final concentration of AIV in the analyzed brain samples of animals treated with 12.5 and 25 mg/kg b.w. of AIV (0.00012299 and 0.0002306 mg, respectively, per one brain). Moreover, the AIV logBB value was experimentally determined and found to be equal to 0.49 ± 0.03.

Combined computational-experimental approach to predict blood-brain barrier (BBB) permeation based on "green" salting-out thin layer chromatography supported by simple molecular descriptors

The objective of this paper is to build QSRR/QSAR model for predicting the blood-brain barrier (BBB) permeability. The obtained models are based on salting-out thin layer chromatography (SOTLC) constants and calculated molecular descriptors. Among chromatographic methods SOTLC was chosen, since the mobile phases are free of organic solvent. As consequences, there are less toxic, and have lower environmental impact compared to classical reserved phases liquid chromatography (RPLC). During the study three stationary phase silica gel, cellulose plates and neutral aluminum oxide were examined. The model set of solutes presents a wide range of log BB values, containing compounds which cross the BBB readily and molecules poorly distributed to the brain including drugs acting on the nervous system as well as peripheral acting drugs. Additionally, the comparison of three regression models: multiple linear regression (MLR), partial least-squares (PLS) and orthogonal partial least squares (OPLS) were performed. The designed QSRR/QSAR models could be useful to predict BBB of systematically synthesized newly compounds in the drug development pipeline and are attractive alternatives of time-consuming and demanding directed methods for log BB measurement. The study also shown that among several regression techniques, significant differences can be obtained in models performance, measured by R² and Q², hence it is strongly suggested to evaluate all available options as MLR, PLS and OPLS.

A method to predict different mechanisms for blood-brain barrier permeability of CNS activity compounds in Chinese herbs using support vector machine

The blood-brain barrier (BBB), a highly selective barrier between central nervous system (CNS) and the blood stream, restricts and regulates the penetration of compounds from the blood into the brain. Drugs that affect the CNS interact with the BBB prior to their target site, so the prediction research on BBB permeability is a fundamental and significant research direction in neuropharmacology. In this study, we combed through the available data and then with the help of support vector machine (SVM), we established an experiment process for discovering potential CNS compounds and investigating the mechanisms of BBB permeability of them to advance the research in this field four types of prediction models, referring to CNS activity, BBB permeability, passive diffusion and efflux transport, were obtained in the experiment process. The first two models were used to discover compounds which may have CNS activity and also cross the BBB at the same time; the latter two were used to elucidate the mechanism of BBB permeability of those compounds. Three optimization parameter methods, Grid Search, Genetic Algorithm (GA), and Particle Swarm Optimization (PSO), were used to optimize the SVM models. Then, four optimal models were selected with excellent evaluation indexes (the accuracy, sensitivity and specificity of each model were all above 85%). Furthermore, discrimination models were utilized to study the BBB properties of the known CNS activity compounds in Chinese herbs and this may guide the CNS drug development. With the relatively systematic and quick approach, the application rationality of traditional Chinese medicines for treating nervous system disease in the clinical practice will be improved.

A Generally Applicable Computer Algorithm Based on the Group Additivity Method for the Calculation of Seven Molecular Descriptors: Heat of Combustion, LogPO/W, LogS, Refractivity, Polarizability, Toxicity and LogBB of Organic Compounds; Scope and Limits of Applicability

A generally applicable computer algorithm for the calculation of the seven molecular descriptors heat of combustion, logP_{octanol/water}, logS (water solubility), molar refractivity, molecular polarizability, aqueous toxicity (protozoan growth inhibition) and logBB (log (c_blood/c_brain)) is presented. The method, an extendable form of the group-additivity method, is based on the complete break-down of the molecules into their constituting atoms and their immediate neighbourhood. The contribution of the resulting atom groups to the descriptor values is calculated using the Gauss-Seidel fitting method, based on experimental data gathered from literature. The plausibility of the method was tested for each descriptor by means of a k-fold cross-validation procedure demonstrating good to excellent predictive power for the former six descriptors and low reliability of logBB predictions. The goodness of fit (Q²) and the standard deviation of the 10-fold cross-validation calculation was >0.9999 and 25.2 kJ/mol, respectively, (based on N = 1965 test compounds) for the heat of combustion, 0.9451 and 0.51 (N = 2640) for logP, 0.8838 and 0.74 (N = 1419) for logS, 0.9987 and 0.74 (N = 4045) for the molar refractivity, 0.9897 and 0.77 (N = 308) for the molecular polarizability, 0.8404 and 0.42 (N = 810) for the toxicity and 0.4709 and 0.53 (N = 383) for logBB. The latter descriptor revealing a very low Q² for the test molecules (R² was 0.7068 and standard deviation 0.38 for N = 413 training molecules) is included as an example to show the limits of the group-additivity method. An eighth molecular descriptor, the heat of formation, was indirectly calculated from the heat of combustion data and correlated with published experimental heat of formation data with a correlation coefficient R² of 0.9974 (N = 2031).

Molecular determinants of blood-brain barrier permeation

The blood-brain barrier (BBB) is a microvascular unit which selectively regulates the permeability of drugs to the brain. With the rise in CNS drug targets and diseases, there is a need to be able to accurately predict a priori which compounds in a company database should be pursued for favorable properties. In this review, we will explore the different computational tools available today, as well as underpin these to the experimental methods used to determine BBB permeability. These include in vitro models and the in vivo models that yield the dataset we use to generate predictive models. Understanding of how these models were experimentally derived determines our accurate and predicted use for determining a balance between activity and BBB distribution.

Qualitative prediction of blood-brain barrier permeability on a large and refined dataset

The prediction of blood-brain barrier permeation is vitally important for the optimization of drugs targeting the central nervous system as well as for avoiding side effects of peripheral drugs. Following a previously proposed model on blood-brain barrier penetration, we calculated the cross-sectional area perpendicular to the amphiphilic axis. We obtained a high correlation between calculated and experimental cross-sectional area (r = 0.898, n = 32). Based on these results, we examined a correlation of the calculated cross-sectional area with blood-brain barrier penetration given by logBB values. We combined various literature data sets to form a large-scale logBB dataset with 362 experimental logBB values. Quantitative models were calculated using bootstrap validated multiple linear regression. Qualitative models were built by a bootstrapped random forest algorithm. Both methods found similar descriptors such as polar surface area, pKa, logP, charges and number of positive ionisable groups to be predictive for logBB. In contrast to our initial assumption, we were not able to obtain models with the cross-sectional area chosen as relevant parameter for both approaches. Comparing those two different techniques, qualitative random forest models are better suited for blood-brain barrier permeability prediction, especially when reducing the number of descriptors and using a large dataset. A random forest prediction system (n(trees) = 5) based on only four descriptors yields a validated accuracy of 88%.

4D-QSAR: perspectives in drug design

Drug design is a process driven by innovation and technological breakthroughs involving a combination of advanced experimental and computational methods. A broad variety of medicinal chemistry approaches can be used for the identification of hits, generation of leads, as well as to accelerate the optimization of leads into drug candidates. The quantitative structure–activity relationship (QSAR) formalisms are among the most important strategies that can be applied for the successful design new molecules. This review provides a comprehensive review on the evolution and current status of 4D-QSAR, highlighting present challenges and new opportunities in drug design.

Validation of topochemical models for the prediction of permeability through the blood-brain barrier

Recently published topochemical models for permeability through the blood-brain barrier were validated and cross-validated in the present study. Five models based on three topochemical indices, Wiener's topochemical index - a distance-based topochemical descriptor, molecular connectivity topochemical index - an adjacency-based topochemical descriptor and eccentric connectivity topochemical index - an adjacency-cum-distance based topochemical descriptor, for permeability of structurally and chemically diverse molecules through blood-brain barrier were used in the present investigation. A data set comprising 62 structurally and chemically diverse compounds was selected. This data set was divided into two sets of 31 compounds each - one to serve as the validation set and other as the cross-validation set. The values of all the three-topochemical indices in the original as well as in the normalized form for each of the 31 compounds of the validation set were computed using an in-house computer program. Resultant data was analyzed and each compound was assigned a permeability characteristic using topochemical models, which was then compared with the reported permeability through the blood-brain barrier. Accuracy of prediction of these models was calculated. The same procedure was similarly followed for the cross-validation set. Studies revealed accuracy of prediction of the order of 70-80% during validation. Surprisingly, very high predictability of the order of 77-91% was observed during cross-validation. High predictability observed during validation as well as cross-validation authenticates topochemical models for prediction of permeability through the blood-brain barrier.

Prediction of blood-brain partitioning: a model based on ab initio calculated quantum chemical descriptors

A new model for the prediction of log BB, a penetration measure through the blood-brain barrier, based on a molecular set of 82 diverse molecules is developed. The majority of the descriptors are derived from quantum chemical ab initio calculations, augmented with a number of classical descriptors. The quantum chemical information enables one to compute fundamental properties of the molecules. The best set of descriptors was selected by sequential selection and multiple linear regression was used to develop the QSAR model. The predictive capability of the model was tested using internal and external test procedures and the domain of applicability was determined to identify reliable predictions. The selected set of descriptors shows a significant correlation with the experimental log BB. The proposed model could reproduce the data with an error approaching the experimental uncertainty and satisfies the available validation procedures. The obtained results indicate that the use of quantum chemical information in describing molecules improves the behavior of the model.

Prediction of Blood-Βrain Partitioning and Human Serum Albumin Binding Based on COSMO-RS σ-Moments

Models for the prediction of blood-brain partitioning (logBB) and human serum albumin binding (logK(HSA)) of neutral molecules were developed using the set of 5 COSMO-RS sigma-moments as descriptors. These sigma-moments have already been introduced earlier as a general descriptor set for partition coefficients. They are obtained from quantum chemical calculations using the continuum solvation model COSMO and a subsequent statistical decomposition of the resulting polarization charge densities. The model for blood-brain partitioning was built on a data set of 103 compounds and yielded a correlation coefficient of r2 = 0.71 and an rms error of 0.40 log units. The human serum albumin binding model was built on a data set of 92 compounds and achieved an r2 of 0.67 and an rms error of 0.33 log units. Both models were validated by leave-one-out cross-validation tests, which resulted in q2 = 0.68 and a qms error of 0.42 for the logBB model and in q2 = 0.63 and a qms error of 0.35 for the logK(HSA) model. Together with the previously published models for intestinal absorption and for drug solubility the presented two models complete the COSMO-RS based set of ADME prediction models.

A data base for partition of volatile organic compounds and drugs from blood/plasma/serum to brain, and an LFER analysis of the data

Literature values of the in vivo distribution (BB) of drugs from blood, plasma, or serum to rat brain have been assembled for 207 compounds (233 data points). We find that data on in vivo distribution from blood, plasma, and serum to rat brain can all be combined. Application of our general linear free energy relationship (LFER) to the 207 compounds yields an equation in log BB, with R2=0.75 and a standard deviation, SD, of 0.33 log units. An equation for a training set predicts the test set of data with a standard deviation of 0.31 log units. We further find that the in vivo data cannot simply be combined with in vitro data on volatile organic and inorganic compounds, because there is a systematic difference between the two sets of data. Use of an indicator variable allows the two sets to be combined, leading to a LFER equation for 302 compounds (328 data points) with R2=0.75 and SD=0.30 log units. A training equation was then used to predict a test set with SD=0.25 log units.

In silico prediction of blood brain barrier permeability: an Artificial Neural Network model

This paper has two objectives: first to develop an in silico model for the prediction of blood brain barrier permeability of new chemical entities and second to find the role of active transport specific to the P-glycoprotein (P-gp) substrate probability in blood brain barrier permeability. An Artificial Neural Network (ANN) model has been developed to predict the ratios of the steady-state concentrations of drugs in the brain to those in the blood (logBB) from their molecular structural parameters. Seven descriptors including P-gp substrate probability have been used for model development. The developed model is able to capture a relationship between P-gp and logBB. The predictive ability of the ANN model has also been compared with earlier computational models.

A recursive-partitioning model for blood–brain barrier permeation

A series of bagged recursive partitioning models for log(BB) is presented. Using a LGO-CV, three sets of physical property descriptors are evaluated and found to have Q2 values of 0.51 (CPSA), 0.53 (Ro5x) and 0.53 (MOE). Extrapolating these models to Pfizer chemical space is difficult due to P-glycoprotein (P-gp) mediated efflux. Low correlation coefficients for this test set are improved (R2 = 0.39) when compounds known to be P-gp substrates or statistical extrapolations are removed. The use of simple linear models for specific chemical series is also found to improve the correlation over a limited chemical space.

Constructing optimum blood brain barrier QSAR models using a combination of 4D-molecular similarity measures and cluster analysis

A new method, using a combination of 4D-molecular similarity measures and cluster analysis to construct optimum QSAR models, is applied to a data set of 150 chemically diverse compounds to build optimum blood-brain barrier (BBB) penetration models. The complete data set is divided into subsets based on 4D-molecular similarity measures using cluster analysis. The compounds in each cluster subset are further divided into a training set and a test set. Predictive QASAR models are constructed for each cluster subset using the corresponding training sets. These QSAR models best predict test set compounds which are assigned to the same cluster subset, based on the 4D-molecular similarity measures, from which the models are derived. The results suggest that the specific properties governing blood-brain barrier permeability may vary across chemically diverse compounds. Partitioning compounds into chemically similar classes is essential to constructing predictive blood-brain barrier penetration models embedding the corresponding key physiochemical properties of a given chemical class.

In silico ADME modelling: prediction models for blood-brain barrier permeation using a systematic variable selection method

Quantitative Structure-Property Relationship models (QSPR) based on in vivo blood-brain permeation data (logBB) of 88 diverse compounds, 324 descriptors and a systematic variable selection method, namely 'Variable Selection and Modeling method based on the prediction (VSMP)', are reported. Of all the models developed using VSMP, the best three-descriptors model is based on Atomic type E-state index (SsssN), AlogP98 and Van der Waal's surface area (r=0.8425, q=0.8239, F=68.49 and SE=0.4165); the best four-descriptors model is based on Kappa shape index of order 1, Atomic type E-state index (SsssN), Atomic level based AI topological descriptor (AIssssC) and AlogP98 (r=0.8638, q=0.8472, F=60.982 and SE=0.3919). The performance of the models on three test sets taken from the literature is illustrated and compared with the results from other reported computational approaches. Test set III constitutes 91 compounds from the literature with known qualitative BBB indication and is used for virtual screening studies. The success rate of the reported models is 82% in the case of BBB+ compounds and a similar success rate is observed with BBB- compounds. Finally, as the models reported herein are based on computed properties, they appear as a valuable tool in virtual screening, where selection and prioritization of candidates is required.

How to measure drug transport across the blood-brain barrier

The extent to which a substance in the circulation gains access to the CNS needs to be determined for potential neuropharmaceuticals as well as for drug candidates with primary targets in the periphery. Characteristics of the in vivo methods, ranging from classical pharmacokinetic techniques (intravenous administration and tissue sampling) over brain perfusions to microdialysis and imaging techniques, are highlighted. In vivo measurements remain unmatched with respect to sensitivity and for the characterization of carrier-mediated uptake, receptor-mediated transport, and active efflux. Isolated microvessels are valuable tools for molecular characterization of transporters. Endothelial cell culture models of the blood-brain barrier (BBB) are pursued as in vitro systems suitable for screening procedures. Recent applications of conditionally immortalized cell lines indicate that a particular weakness of culture models because of downregulation of BBB-specific transporter systems can be overcome. In silico approaches are being developed with the goal of predicting brain uptake from molecular structure at early stages of drug development. Currently, the predictive capability is limited to passive, diffusional uptake and predominantly relies on few molecular descriptors related to lipophilicity, hydrogen bonding capacity, charge, and molecular weight. A caveat with most present strategies is their reliance on surrogates of BBB transport, like CNS activity/inactivity or brain-to-blood partitioning rather than actual BBB permeability data.

Effect of Selection of Molecular Descriptors on the Prediction of Blood−Brain Barrier Penetrating and Nonpenetrating Agents by Statistical Learning Methods

The ability or inability of a drug to penetrate into the brain is a key consideration in drug design. Drugs for treating central nervous system (CNS) disorders need to be able to penetrate the blood-brain barrier (BBB). BBB nonpenetration is desirable for non-CNS-targeting drugs to minimize potential CNS-related side effects. Computational methods have been employed for the prediction of BBB-penetrating (BBB+) and -nonpenetrating (BBB-) agents at impressive accuracies of 75-92% and 60-80%, respectively. However, the majority of these studies give a substantially lower BBB- accuracy, and thus overall accuracy, than the BBB+ accuracy. This work examined whether proper selection of molecular descriptors can improve both the BBB- and the overall accuracies of statistical learning methods. The methods tested include logistic regression, linear discriminate analysis, k nearest neighbor, C4.5 decision tree, probabilistic neural network, and support vector machine. Molecular descriptors were selected by using a feature selection method, recursive feature elimination (RFE). Results by using 415 BBB+ and BBB- agents show that RFE substantially improves both the BBB- and the overall accuracy for all of the methods studied. This suggests that statistical learning methods combined with proper feature selection is potentially useful for facilitating a more balanced and improved prediction of BBB+ and BBB- agents.

Predictive model of blood-brain barrier penetration of organic compounds

AIM

To build up a theoretical model of organic compounds for the prediction of the activity of small molecules through the blood-brain barrier (BBB) in drug design.

METHODS

A training set of 37 structurally diverse compounds was used to construct quantitative structure-activity relationship (QSAR) models. Intermolecular and intramolecular solute descriptors were calculated using molecular mechanics, molecular dynamics simulations, quantum chemistry and so on. The QSAR models were optimized using multidimensional linear regression fitting and stepwise method. A test set of 8 compounds was evaluated using the models as part of a validation process.

RESULTS

Significant QSAR models (R=0.955, s=0.232) of the BBB penetration of organic compounds were constructed. BBB penetration was found to depend upon the polar surface area, the octanol/water partition coefficient, Balaban Index, the strength of a small molecule to combine with the membrane-water complex, and the changeability of the structure of a solute-membrane-water complex.

CONCLUSION

The QSAR models indicate that the distribution of organic molecules through BBB is not only influenced by organic solutes themselves, but also relates to the properties of the solute-membrane-water complex, that is, interactions of the molecule with the phospholipid-rich regions of cellular membranes.

TOPS-MODE approach for the prediction of blood-brain barrier permeation

The blood-brain barrier permeation has been investigated by using a topological substructural molecular design approach (TOPS-MODE). A linear regression model was developed to predict the in vivo blood-brain partitioning coefficient on a data set of 119 compounds, treated as the logarithm of the blood-brain concentration ratio. The final model explained the 70% of the variance and it was validated through the use of an external validation set (33 compounds of the 119, MAE = 0.33), a leave-one-out crossvalidation (q(2) = 0.65, S(press) = 0.43), fivefold full crossvalidation (removing 28 compounds in each cycle, MAE = 33, RMSE = 0.43) and the prediction of +/- values for an external test set (85.7% of good prediction). This methodology evidenced that the hydrophobicity increase the blood-brain barrier permeation, while the polar surface and its interaction with the atomic mass of compounds decrease it; suggesting the capacity of the TOPS-MODE descriptors to estimate brain penetration potential of new drug candidates. Finally, by the present approach, positive and negative substructural contributions to the brain permeation were identified, and their possibilities in the lead generation and optimization processes were evaluated.

Quantitative Structure-Pharmacokinetic Relationships for Drug Distribution Properties by Using General Regression Neural Network

Quantitative Structure-Pharmacokinetic Relationships (QSPkR) have increasingly been used for developing models for the prediction of the pharmacokinetic properties of drug leads. QSPkR models are primarily developed by means of statistical methods such as multiple linear regression (MLR). These methods often explore a linear relationship between the pharmacokinetic property of interest and the structural and physicochemical properties of the studied compounds, which are not applicable to those agents with nonlinear relationships. Hence, statistical methods capable of modeling nonlinear relationships need to be developed. In this work, a relatively new kind of nonlinear method, general regression neural network (GRNN), was explored for modeling three drug distribution properties based on diverse sets of drugs. The three properties are blood-brain barrier penetration, binding to human serum albumin, and milk-plasma distribution. The prediction capability of GRNN-developed models was compared to those developed using MLR and a nonlinear multilayer feedforward neural network (MLFN) method. For blood-brain barrier penetration, the computed r(2) and MSE values of the GRNN-, MLR-, and MLFN-developed models are 0.701 and 0.130, 0.649 and 0.154, and 0.662 and 0.147, respectively, by using an independent validation set. The corresponding values for human serum albumin binding are 0.851 and 0.041, 0.770 and 0.079, and 0.749 and 0.089, respectively, and that for milk-plasma distribution are 0.677 and 0.206, 0.224 and 0.647, and 0.201 and 0.587, respectively. These suggest that GRNN is potentially useful for predicting QSPkR properties of chemical agents.

In silico prediction of central nervous system activity of compounds. Identification of potential pharmacophores by the TOPS–MODE approach

The central nervous system (CNS) activity has been investigated by using a topological substructural molecular approach (TOPS-MODE). A discriminant analysis to classify CNS and non-CNS drugs was developed on a data set (302 compounds) of great structural variability where more than 81% (247/302) were well classified. Randic's orthogonalization procedures was carried out to allow the interpretation of the model and to avoid the collinearity among descriptors. The discriminant model was assessed by a leave-n-out (when n varies from 2 to 20) cross-validation procedure (79.94% of correct classification), an external prediction set composed by 78 CNS/non-CNS drugs (80.77% of correct classification) and a 5-fold full cross-validation (removing 78 compounds in each cycle, 80.00% of good classification). With this methodology was demonstrated that the hydrophobicity increase the CNS activity, while the dipole moment and the polar surface area decrease it; evidencing the capacity of the TOPS-MODE descriptors to estimate CNS activity for new drug candidates. The structural contributions to the CNS activity for two compounds are presented on the basis of fragment contributions. The model has also been able to identify potential structural pharmacophore, showing its possibilities in the lead generation and optimization processes.

Relationship between exposure and nonspecific binding of thirty-three central nervous system drugs in mice

Unbound fractions in mouse brain and plasma were determined for 31 structurally diverse central nervous system (CNS) drugs and two active metabolites. Three comparisons were made between in vitro binding and in vivo exposure data, namely: 1) mouse brain-to-plasma exposure versus unbound plasma-to-unbound brain fraction ratio (fu(plasma)/fu(brain)), 2) cerebrospinal fluid-to-brain exposure versus unbound brain fraction (fu(brain)), and 3) cerebrospinal fluid-to-plasma exposure versus unbound plasma fraction (fu(plasma)). Unbound fraction data were within 3-fold of in vivo exposure ratios for the majority of the drugs examined (i.e., 22 of 33), indicating a predominately free equilibrium across the blood-brain and blood-CSF barriers. Some degree of distributional impairment at either the blood-CSF or the blood-brain barrier was indicated for 8 of the 11 remaining drugs (i.e., carbamazepine, midazolam, phenytoin, sulpiride, thiopental, risperidone, 9-hydroxyrisperidone, and zolpidem). In several cases, the indicated distributional impairment is consistent with other independent literature reports for these drugs. Through the use of this approach, it appears that most CNS-active agents freely equilibrate across the blood-brain and blood-CSF barriers such that unbound drug concentrations in brain approximate those in the plasma. However, these results also support the intuitive concept that distributional impairment does not necessarily preclude CNS activity.

Prediction of blood-brain barrier permeation using quantum chemically derived information

A model for the prediction of the blood-brain distribution (logBB) is obtained by multiple regression analysis of molecular descriptors for a training set of 90 compounds. The majority of the descriptors are derived from quantum chemical information using semi-empirical AM1 calculations to compute fundamental properties of the molecules investigated. The polar surface area of the compounds can be described appropriately by six descriptors derived from the molecular electrostatic potential. This set shows a strong correlation with the observed logBB. Additional quantum chemically computed properties that contribute to the final model comprise the ionization potential and the covalent hydrogen-bond basicity. Complementary descriptors account for the presence of certain chemical groups, the number of hydrogen-bond donors, and the number of rotatable bonds of the compounds. The quality of the fit is further improved by including variables derived from principal component analysis of the molecular geometry.

Blood-Brain Barrier Permeation Models: Discriminating between Potential CNS and Non-CNS Drugs Including P-Glycoprotein Substrates

The aim of this article is to present the design of a large heterogeneous CNS library (approximately 1700 compounds) from WDI and mapping CNS drugs using QSAR models of blood-brain barrier (BBB) permeation and P-gp substrates. The CNS library finally includes 1336 BBB-crossing drugs (BBB+), 259 molecules non-BBB-crossing (BBB-), and 91 P-gp substrates (either BBB+ or BBB-). Discriminant analysis and PLS-DA have been used to model the passive diffusion component of BBB permeation and potential physicochemical requirement of P-gp substrates. Three categories of explanatory variables (Cdiff, BBBpred, PGPpred) have been suggested to express the level of permeation within a continuous scale, starting from two classes data (BBB+/BBB-), allowing that the degree to which each compound belongs to an activity class is given using a membership score. Finally, statistical data analyses have shown that some very simple descriptors are sufficient to evaluate BBB permeation in most cases, with a high rate of well-classified drugs. Moreover, a "CNS drugs" map, including P-gp substrates and accurately reflecting the in vivo behavior of drugs, is proposed as a tool for CNS drug virtual screening.

Recent advances in the prediction of blood-brain partitioning from molecular structure

The widely distributed software tools Cerius2 and ACD/log D Suite have been used to develop a new method for the prediction of the ratio of concentrations of a drug in the brain and blood (BB,quantified as log BB) from structure. The performances of all known blood-brain partitioning prediction methods are compared to give an up-to-date account on their accuracy, limitations, and usefulness. It is demonstrated that the new log BB prediction method is superior to other methods with regard to low-to-medium throughput log BB prediction, whereas the C2-ADME log BB two-dimensional (2D) method seems to offer the best compromise between speed and accuracy for ultra-high throughput processing of large compound databases for log BB prediction.

Predicting blood-brain barrier partitioning of organic molecules using membrane-interaction QSAR analysis

PURPOSE

Membrane-interaction quantitative structure-activity relationship (OSAR) analysis (MI-QSAR) has been used to develop predictive models of blood-brain barrier partitioning of organic compounds by, in part, simulating the interaction of an organic compound with the phospholipid-rich regions of cellular membranes.

METHOD

A training set of 56 structurally diverse compounds whose blood-brain barrier partition coefficients were measured was used to construct MI-QSAR models. Molecular dynamics simulations were used to determine the explicit interaction of each test compound (solute) with a model DMPC monolayer membrane model. An additional set of intramolecular solute descriptors were computed and considered in the trial pool of descriptors for building MI-QSAR models. The QSAR models were optimized using multidimensional linear regression fitting and a genetic algorithm. A test set of seven compounds was evaluated using the MI-QSAR models as part of a validation process.

RESULTS

Significant MI-QSAR models (R2 = 0.845, Q2 = 0.795) of the blood-brain partitioning process were constructed. Blood-brain barrier partitioning is found to depend upon the polar surface area. the octanol/water partition coefficient, and the conformational flexibility of the compounds as well as the strength of their binding" to the model biologic membrane. The blood-brain barrier partitioning measures of the test set compounds were predicted with the same accuracy as the compounds of the training set.

CONCLUSION

The MI-QSAR models indicate that the blood-brain barrier partitioning process can be reliably described for structurally diverse molecules provided interactions of the molecule with the phospholipids-rich regions of cellular membranes are explicitly considered.

Modeling blood-brain barrier partitioning using the electrotopological state

The challenging problem of modeling blood-brain barrier partitioning is approached through topological representation of molecular structure. A QSAR model is developed for in vivo blood-brain partitioning data treated as the logarithm of the blood-brain concentration ratio. The model consists of three structure descriptors: the hydrogen E-State index for hydrogen bond donors, HS(T)(HBd); the hydrogen E-State index for aromatic CHs, HS(T)(arom); and the second order difference valence molecular connectivity index, d(2)chi(v) (q(2) = 0.62.) The model for the set of 106 compounds is validated through use of an external validation test set (20 compounds of the 106, MAE = 0.33, rms = 0.38), 5-fold cross-validation (MAE = 0.38, rms = 0.47), prediction of +/- values for an external test set (27/28 correct), and estimation of logBB values for a large data set of 20 039 drugs and drug-like compounds. Because no 3D structure information is used, computation of logBB by the model is very fast. The quality of the validation statistics supports the claim that the model may be used for estimation of logBB values for drug and drug-like molecules. Detailed structure interpretation is given for the structure indices in the model. The model indicates that molecules that penetrate the blood-brain barrier have large HS(T)(arom) values (presence of aromatic groups) but small values of HS(T)(HBd) (fewer or weaker H-Bond donors) and smaller d(2)chi(v) values (less branched molecules with fewer electronegative atoms). These three structure descriptors encode influence of molecular context of groups as well as counts of those groups.

Computational approaches to the prediction of the blood-brain distribution

This review attempts to summarise present knowledge related to the theoretical modelling of drug transport across the blood-brain barrier. Several computational protocols are described ranging from quantum mechanics-based approaches through molecular mechanics-related techniques to simple and fast procedures based on only the 2-D graph of the investigated structures. Amazingly, few descriptors have been shown to influence the derived relationships in a significant manner and a cornerstone in most of the described models are terms describing hydrogen bonding. A very quick quantitative assessment of the brain partitioning of a compound has also been devised using the following two rules: If N+O (the number of nitrogen and oxygen atoms) in a molecule is less than or equal to five, it has a high chance of entering the brain. The second rule predicts that if log P-(N+O) is positive then log BB is positive.

Solubility predictions for crystalline nonelectrolyte solutes dissolved in organic solvents based upon the Abraham general solvation model

The Abraham general solvation model is used to predict the saturation solubility of crystalline nonelectrolyte solutes in organic solvents. The derived equations take the form of log (CS/CW) = c + rR2 + sπ2H + aΣα2H + bΣβ2H + vVx and log (CS/CG) = c + rR2 + sπ2H + aΣα2H + bΣβ2H + l log L(16) where CS and CW refer to the solute solubility in the organic solvent and water, respectively, CG is a gas-phase concentration, R2 is the solute's excess molar refraction, Vx is McGowan volume of the solute, Σα2H and Σβ2H are measures of the solute's hydrogen-bond acidity and hydrogen-bond basicity, π2H denotes the solute's dipolarity and (or) polarizability descriptor, and log L(16) is the solute's gas-phase dimensionless Ostwald partition coefficient into hexadecane at 298 K. The remaining symbols in the above expressions are known equation coefficients, which have been determined previously for a large number of gas–solvent and water–solvent systems. Computations show that the Abraham general solvation model predicts the observed solubility behavior of anthracene, phenanthrene, and hexachlorobenzene to within an average absolute deviation of about ±35%.Key words: solubility predictions, organic solvents, nonelectrolyte solutes, partition coefficients.

Correlation and prediction of a large blood-brain distribution data set--an LFER study

We report linear free energy relation (LFER) models of the equilibrium distribution of molecules between blood and brain, as log BB values. This method relates log BB values to fundamental molecular properties, such as hydrogen bonding capability, polarity/polarisability and size. Our best model of this form covers 148 compounds, the largest set of log BB data yet used in such a model, resulting in R(2)=0.745 and e.s.d.=0.343 after inclusion of an indicator variable for carboxylic acids. This represents rather better accuracy than a number of previously reported models based on subsets of our data. The model also reveals the factors that affect log BB: molecular size and dispersion effects increase brain uptake, while polarity/polarisability and hydrogen-bond acidity and basicity decrease it. By splitting the full data set into several randomly selected training and test sets, we conclude that such a model can predict log BB values with an accuracy of less than 0.35 log units. The method is very rapid-log BB can be calculated from structure at a rate of 700 molecules per minute on a silicon graphics O(2).