TISBE: A Public Web Platform for the Consensus-Based Explainable Prediction of Developmental Toxicity.

Despite being extremely relevant for the protection of prenatal and neonatal health, the developmental toxicity (Dev Tox) is a highly complex endpoint whose molecular rationale is still largely unknown. The lack of availability of high-quality data as well as robust nontesting methods makes its understanding even more difficult. Thus, the application of new explainable alternative methods is of utmost importance, with Dev Tox being one of the most animal-intensive research themes of regulatory toxicology. Descending from TIRESIA (Toxicology Intelligence and Regulatory Evaluations for Scientific and Industry Applications), the present work describes TISBE (TIRESIA Improved on Structure-Based Explainability), a new public web platform implementing four fundamental advancements for in silico analyses: a three times larger dataset, a transparent XAI (explainable artificial intelligence) framework employing a fragment-based fingerprint coding, a novel consensus classifier based on five independent machine learning models, and a new applicability domain (AD) method based on a double top-down approach for better estimating the prediction reliability. The training set (TS) includes as many as 1008 chemicals annotated with experimental toxicity values. Based on a 5-fold cross-validation, a median value of 0.410 for the Matthews correlation coefficient was calculated; TISBE was very effective, with a median value of sensitivity and specificity equal to 0.984 and 0.274, respectively. TISBE was applied on two external pools made of 1484 bioactive compounds and 85 pediatric drugs taken from ChEMBL (Chemical European Molecular Biology Laboratory) and TEDDY (Task-Force in Europe for Drug Development in the Young) repositories, respectively. Notably, TISBE gives users the option to clearly spot the molecular fragments responsible for the toxicity or the safety of a given chemical query and is available for free at https://prometheus.farmacia.uniba.it/tisbe.

CIRCE: Web-Based Platform for the Prediction of Cannabinoid Receptor Ligands Using Explainable Machine Learning.

The endocannabinoid system, which includes cannabinoid receptor 1 and 2 subtypes (CB1R and CB2R, respectively), is responsible for the onset of various pathologies including neurodegeneration, cancer, neuropathic and inflammatory pain, obesity, and inflammatory bowel disease. Given the high similarity of CB1R and CB2R, generating subtype-selective ligands is still an open challenge. In this work, the Cannabinoid Iterative Revaluation for Classification and Explanation (CIRCE) compound prediction platform has been generated based on explainable machine learning to support the design of selective CB1R and CB2R ligands. Multilayer classifiers were combined with Shapley value analysis to facilitate explainable predictions. In test calculations, CIRCE predictions reached ∼80% accuracy and structural features determining ligand predictions were rationalized. CIRCE was designed as a web-based prediction platform that is made freely available as a part of our study.

TIRESIA: An eXplainable Artificial Intelligence Platform for Predicting Developmental Toxicity.

Herein, a robust and reproducible eXplainable Artificial Intelligence (XAI) approach is presented, which allows prediction of developmental toxicity, a challenging human-health endpoint in toxicology. The application of XAI as an alternative method is of the utmost importance with developmental toxicity being one of the most animal-intensive areas of regulatory toxicology. In this work, the established CAESAR (Computer Assisted Evaluation of industrial chemical Substances According to Regulations) training set made of 234 chemicals for model learning is employed. Two test sets, including as a whole 585 chemicals, were instead used for validation and generalization purposes. The proposed framework favorably compares with the state-of-the-art approaches in terms of accuracy, sensitivity, and specificity, thus resulting in a reliable support system for developmental toxicity ensuring informativeness, uncertainty estimation, generalization, and transparency. Based on the eXtreme Gradient Boosting (XGB) algorithm, our predictive model provides easy interpretative keys based on specific molecular descriptors and structural alerts enabling one to distinguish toxic and nontoxic chemicals. Inspired by the Organisation for Economic Co-operation and Development (OECD) principles for the validation of Quantitative Structure-Activity Relationships (QSARs) for regulatory purposes, the results are summarized in a standard report in portable document format, enclosing also details concerned with a density-based model applicability domain and SHAP (SHapley Additive exPlanations) explainability, the latter particularly useful to better understand the effective roles played by molecular features. Notably, our model has been implemented in TIRESIA (Toxicology Intelligence and Regulatory Evaluations for Scientific and Industry Applications), a free of charge web platform available at http://tiresia.uniba.it.

Polymer Encapsulated Liposomes for Oral Co-Delivery of Curcumin and Hydroxytyrosol.

Curcumin (Cur) is a hydrophobic polyphenol from the rhizome of Curcuma spp., while hydroxytyrosol (HT) is a water-soluble polyphenol from Olea europaea. Both show outstanding antioxidant properties but suffer from scarce bioavailability and low stability in biological fluids. In this work, the co-encapsulation of Cur and HT into liposomes was realized, and the liposomal formulation was improved using polymers to increase their survival in the gastrointestinal tract. Liposomes with different compositions were formulated: Type 1, composed of phospholipids and cholesterol; Type 2, also with a PEG coating; and Type 3 providing an additional shell of Eudragit® S100, a gastro-resistant polymer. Samples were characterized in terms of size, morphology, ζ-potential, encapsulation efficiency, and loading capacity. All samples were subjected to a simulated in vitro digestion and their stability was investigated. The Eudragit®S100 coating demonstrated prevention of early releases of HT in the mouth and gastric phases, while the PEG shell reduced bile salts and pancreatin effects during the intestinal digestion. In vitro antioxidant activity showed a cumulative effect for Cur and HT loaded in vesicles. Finally, liposomes with HT concentrations up to 40 µM and Cur up to 4.7 µM, alone or in combination, did not show cytotoxicity against Caco-2 cells.

Multistep Continuous Flow Synthesis of Isolable NH2 -Sulfinamidines via Nucleophilic Addition to Transient Sulfurdiimide.

The growing interest in novel sulfur pharmacophores led to recent advances in the synthesis of some S(IV) and S(VI) motifs. However, preparation and isolation of uncommon primary sulfinamidines, the aza-analogues of sulfinamides, is highly desirable. Here we report a multistep continuous flow synthesis of poorly explored NH2 -sulfinamidines by nucleophilic attack of organometallic reagents to in situ prepared N-(trimethylsilyl)-N-trityl-λ4 -sulfanediimine (Tr-N=S=N-TMS). The transformation can additionally be realized under mild conditions, at room temperature, via a highly chemoselective halogen-lithium exchange of aryl bromides and iodides with n-butyllithium. Moreover, the synthetic potential of the methodology was assessed by exploring further manipulations of the products and accessing novel S(IV) analogues of celecoxib, tasisulam, and relevant sulfinimidoylureas.

PLATO: A Predictive Drug Discovery Web Platform for Efficient Target Fishing and Bioactivity Profiling of Small Molecules.

PLATO (Polypharmacology pLATform predictiOn) is an easy-to-use drug discovery web platform, which has been designed with a two-fold objective: to fish putative protein drug targets and to compute bioactivity values of small molecules. Predictions are based on the similarity principle, through a reverse ligand-based screening, based on a collection of 632,119 compounds known to be experimentally active on 6004 protein targets. An efficient backend implementation allows to speed-up the process that returns results for query in less than 20 s. The graphical user interface is intuitive to give practitioners easy input and transparent output, which is available as a standard report in portable document format. PLATO has been validated on thousands of external data, with performances better than those of other parallel approaches. PLATO is available free of charge (http://plato.uniba.it/ accessed on 13 April 2022).

Synthesis and Investigation of Novel CHCA-Derived Matrices for Matrix-Assisted Laser Desorption/Ionization Mass Spectrometric Analysis of Lipids.

A significant area of study and upgrading for increasing sensitivity and general performances of matrix-assisted laser-desorption ionization (MALDI) mass spectrometry (MS) is related to matrix design. Several efforts have been made to address the challenge of low-mass-region interference-free for metabolomics analysis and specifically for lipidomics. To this aim, rationally designed matrices as 4-chloro-α-cyanocinnamic acid (ClCCA) were introduced and reported to provide enhanced analytical performances. We have taken this rational design one step further by developing and optimizing new MALDI matrices with a range of modifications on the CHCA core, involving different functionalities and substituents. Of particular interest was the understanding of the electron-withdrawing (e.g., nitro-) or donating (e.g., methoxy-) effects along with the extent of conjugation on the ionization efficiency. In the present work, ten matrices were designed on a reasonable basis, synthesized, and characterized by NMR and UV spectroscopies and laser desorption ionization. With the assistance of these putative MALDI matrices, samples containing phospholipids (PL), and neutral di-/tri-acylglycerols (DAG, TAG) were investigated using milk, fish, blood, and human plasma extracts. In comparison with CHCA and ClCCA, four of them, viz. [(2E,4E)-2-cyano-5-(4-methoxyphenyl)penta-2,4-dienoic acid] (1), [(2E,4E)-2-cyano-5-(4-nitrophenyl)penta-2,4-dienoic acid] (2), [(E)-2-cyano-3-(6-methoxynaphthalen-2-yl)acrylic acid] (6) and [(E)-2-cyano-3-(naphthalen-2-yl)acrylic acid] (7) displayed good to even excellent performances as MALDI matrices in terms of ionization capability, interference-free spectra, S/N ratio, and reproducibility. Especially compound 7 (cyano naphthyl acrylic acid, CNAA) was the election matrix for PL analysis and matrix 2 (cyano nitrophenyl dienoic acid, CNDA) for neutral lipids such as DAG and TAG in positive ion mode.

Structural Insights into the Vapochromic Behavior of Pt- and Pd-Based Compounds.

Anionic complexes having vapochromic behavior are investigated: [K(H2O)][M(ppy)(CN)2], [K(H2O)][M(bzq)(CN)2], and [Li(H2O)n][Pt(bzq)(CN)2], where ppy = 2-phenylpyridinate, bzq = 7,8-benzoquinolate, and M = Pt(II) or Pd(II). These hydrated potassium/lithium salts exhibit a change in color upon being heated to 380 K, and they transform back into the original color upon absorption of water molecules from the environment. The challenging characterization of their structure in the vapochromic transition has been carried out by combining several experimental techniques, despite the availability of partially ordered and/or impure crystalline material. Room-temperature single-crystal and powder X-ray diffraction investigation revealed that [K(H2O)][Pt(ppy)(CN)2] crystallizes in the Pbca space group and is isostructural to [K(H2O)][Pd(ppy)(CN)2]. Variable-temperature powder X-ray diffraction allowed the color transition to be related to changes in the diffraction pattern and the decrease in sample crystallinity. Water loss, monitored by thermogravimetric analysis, occurs in two stages, well separated for potassium Pt compounds and strongly overlapped for potassium Pd compounds. The local structure of potassium compounds was monitored by in situ pair distribution function (PDF) measurements, which highlighted changes in the intermolecular distances due to a rearrangement of the crystal packing upon vapochromic transition. A reaction coordinate describing the structural changes was extracted for each compound by multivariate analysis applied to PDF data. It contributed to the study of the kinetics of the structural changes related to the vapochromic transition, revealing its dependence on the transition metal ion. Instead, the ligand influences the critical temperature, higher for ppy than for bzq, and the inclination of the molecular planes with respect to the unit cell planes, higher for bzq than for ppy. The first stage of water loss triggers a unit cell contraction, determined by the increase in the b axis length and the decrease in the a (for ppy) or c (for bzq) axis lengths. Consequent interplane distance variations and in-plane roto-translations weaken the π-stacking of the room-temperature structure and modify the distances and angles of Pt(II)/Pd(II) chains. The curve describing the intermolecular Pt(II)/Pd(II) distances as a function of temperature, validated by X-ray absorption spectroscopy, was found to reproduce the coordinate reaction determined by the model-free analysis.

Quantitative Polypharmacology Profiling Based on a Multifingerprint Similarity Predictive Approach.

We present a new quantitative ligand-based bioactivity prediction approach employing a multifingerprint similarity search algorithm, enabling the polypharmacological profiling of small molecules. Quantitative bioactivity predictions are made on the basis of the statistical distributions of multiple Tanimoto similarity θ values, calculated through 13 different molecular fingerprints, and of the variation of the measured biological activity, reported as ΔpIC50, for all of the ligands sharing a given protein drug target. The application data set comprises as much as 4241 protein drug targets as well as 418 485 ligands selected from ChEMBL (release 25) by employing a set of well-defined filtering rules. Several large internal and external validation studies were carried out to demonstrate the robustness and the predictive potential of the herein proposed method. Additional comparative studies, carried out on two freely available and well-known ligand-target prediction platforms, demonstrated the reliability of our proposed approach for accurate ligand-target matching. Moreover, two applicative cases were also discussed to practically describe how to use our predictive algorithm, which is freely available as a user-friendly web platform. The user can screen single or multiple queries at a time and retrieve the output as a terse html table or as a json file including all of the information concerning the explored similarities to obtain a deeper understanding of the results. High-throughput virtual reverse screening campaigns, allowing for a given query compound the quick detection of the potential drug target from a large collection of them, can be carried out in batch on demand.

Accelerating Drug Discovery by Early Protein Drug Target Prediction Based on a Multi-Fingerprint Similarity Search.

In this continuing work, we have updated our recently proposed Multi-fingerprint Similarity Search algorithm (MuSSel) by enabling the generation of dominant ionized species at a physiological pH and the exploration of a larger data domain, which included more than half a million high-quality small molecules extracted from the latest release of ChEMBL (version 24.1, at the time of writing). Provided with a high biological assay confidence score, these selected compounds explored up to 2822 protein drug targets. To improve the data accuracy, samples marked as prodrugs or with equivocal biological annotations were not considered. Notably, MuSSel performances were overall improved by using an object-relational database management system based on PostgreSQL. In order to challenge the real effectiveness of MuSSel in predicting relevant therapeutic drug targets, we analyzed a pool of 36 external bioactive compounds published in the Journal of Medicinal Chemistry from October to December 2018. This study demonstrates that the use of highly curated chemical and biological experimental data on one side, and a powerful multi-fingerprint search algorithm on the other, can be of the utmost importance in addressing the fate of newly conceived small molecules, by strongly reducing the attrition of early phases of drug discovery programs.

Azetidine-Borane Complexes: Synthesis, Reactivity, and Stereoselective Functionalization.

The present study reports, for the first time, the synthesis and structural features of azetidine-borane complexes, as well as their reactivity in lithiation reactions. A temperature-dependent stereoselectivity has been disclosed in the reaction of borane with N-alkyl-2-arylazetidines, allowing for a stereoselective preparation of azetidine-borane complexes 2 and 3. A regioselective hydrogen/lithium permutation, at the benzylic position, was observed in lithiation reactions of complexes possessing a syn relationship, between the ring proton and the BH3 group. In contrast, scarce or no reactivity was noticed in complexes lacking such a stereochemical requirement. The configurational stability of the lithiated intermediates has also been investigated, in order to shed some light on the stereoselectivity of the lithiation/electrophile trapping sequence. Calculations helped in supporting experimental observations, concerning structure and reactivity of these azetidine-borane complexes. Data suggest that the BH3 group could promote the lithiation reaction likely by an electrostatic complex induced proximity effect. Interestingly, a new synthetic strategy for the synthesis of N-alkyl-2,2-disubstituted azetidines has been developed.

Semiquinone oscillations as a tool for investigating the ubiquinone binding to photosynthetic reaction centers.

Semiquinone oscillations induced by light pulses in the presence of exogenous electron donors are a valuable source of information on the kinetics and thermodynamics of ubiquinone chemistry relevant to the QB site of the photosynthetic reaction center (RC). In previous attempts to achieve the quantitative interpretation of data, the ubiquinone concentration was considered constant during the experiment since it was much bigger than that of RC. In this work, we extended existing models to low ubiquinone concentrations revealing several hidden processes taking place during the ubiquinone photoreduction and enabling the evaluation of the ubiquinone binding constant K Q at the QB site. The proposed approach provides for the first time the evaluation of K Q without any preliminary treatment of ubiquinone extraction from the binding site, thereby better preserving its native structure.

The binding of quinone to the photosynthetic reaction centers: kinetics and thermodynamics of reactions occurring at the QB-site in zwitterionic and anionic liposomes.

Liposomes represent a versatile biomimetic environment for studying the interaction between integral membrane proteins and hydrophobic ligands. In this paper, the quinone binding to the QB-site of the photosynthetic reaction centers (RC) from Rhodobacter sphaeroides has been investigated in liposomes prepared with either the zwitterionic phosphatidylcholine (PC) or the negatively charged phosphatidylglycerol (PG) to highlight the role of the different phospholipid polar heads. Quinone binding (K Q) and interquinone electron transfer (L AB) equilibrium constants in the two type of liposomes were obtained by charge recombination reaction of QB-depleted RC in the presence of increasing amounts of ubiquinone-10 over the temperature interval 6-35 °C. The kinetic of the charge recombination reactions has been fitted by numerically solving the ordinary differential equations set associated with a detailed kinetic scheme involving electron transfer reactions coupled with quinone release and uptake. The entire set of traces at each temperature was accurately fitted using the sole quinone release constants (both in a neutral and a charge separated state) as adjustable parameters. The temperature dependence of the quinone exchange rate at the QB-site was, hence, obtained. It was found that the quinone exchange regime was always fast for PC while it switched from slow to fast in PG as the temperature rose above 20 °C. A new method was introduced in this paper for the evaluation of constant K Q using the area underneath the charge recombination traces as the indicator of the amount of quinone bound to the QB-site.

Virtual Reverse Screening Approach to Target Type 2 Cannabinoid Receptor.

A screening pool consisting of 617710 drug-like query molecules properly filtered from the ChEMBL database was employed for a ligand-based reverse screening toward the type 2 cannabinoid receptor (CB2) target. By using our recently developed PLATO polypharmacological web platform, 233 out of 617710 drug-like molecules were prioritized on the basis of the predicted bioactivity values, better than 0.2 µM with a probability of about 98%, toward the CB2 target. Building on these results, the occurrence of putative CB2-related targets was also investigated for prospective repurposing studies.

Recent Advances on Type-2 Cannabinoid (CB2) Receptor Agonists and their Therapeutic Potential.

In the last decade, selective modulators of type-2 cannabinoid receptor (CB2) have become a major focus to target endocannabinoid signaling in humans. Indeed, heterogeneously expressed within our body, CB2 actively regulates several physio-pathological processes, thus representing a promising target for developing specific and safe therapeutic drugs. If CB2 modulation has been extensively studied since the very beginning for the treatment of pain and inflammation, the more recent involvement of this receptor in other pathological conditions has further strengthened the pursuit of novel CB2 agonists in the last five years. Against this background, here we discuss the most recent evidence of the protective effects of CB2 against pathological conditions, emphasizing central nervous system disorders, bone and synovial diseases, and cancer. We also summarize the most recent advances in the development of CB2 agonists, focusing on the correlation between different chemical classes and diverse therapeutic applications. Data mining includes a review of the CB2 ligands disclosed in patents also released in the last five years. Finally, we discuss how the recent elucidation of CB2 tertiary structure has provided new details for the rational design of novel and more selective CB2 agonists, thus supporting innovative strategies to develop effective therapeutics. Our overview of the current knowledge on CB2 agonists provides pivotal information on the structure and function of different classes of molecules and opens possible avenues for future research.

Making sense of chemical space network shows signs of criticality.

Chemical space modelling has great importance in unveiling and visualising latent information, which is critical in predictive toxicology related to drug discovery process. While the use of traditional molecular descriptors and fingerprints may suffer from the so-called curse of dimensionality, complex networks are devoid of the typical drawbacks of coordinate-based representations. Herein, we use chemical space networks (CSNs) to analyse the case of the developmental toxicity (Dev Tox), which remains a challenging endpoint for the difficulty of gathering enough reliable data despite very important for the protection of the maternal and child health. Our study proved that the Dev Tox CSN has a complex non-random organisation and can thus provide a wealth of meaningful information also for predictive purposes. At a phase transition, chemical similarities highlight well-established toxicophores, such as aryl derivatives, mostly neurotoxic hydantoins, barbiturates and amino alcohols, steroids, and volatile organic compounds ether-like chemicals, which are strongly suspected of the Dev Tox onset and can thus be employed as effective alerts for prioritising chemicals before testing.

Where developmental toxicity meets explainable artificial intelligence: state-of-the-art and perspectives.

INTRODUCTION: The application of Artificial Intelligence (AI) to predictive toxicology is rapidly increasing, particularly aiming to develop non-testing methods that effectively address ethical concerns and reduce economic costs. In this context, Developmental Toxicity (Dev Tox) stands as a key human health endpoint, especially significant for safeguarding maternal and child well-being. AREAS COVERED: This review outlines the existing methods employed in Dev Tox predictions and underscores the benefits of utilizing New Approach Methodologies (NAMs), specifically focusing on eXplainable Artificial Intelligence (XAI), which proves highly efficient in constructing reliable and transparent models aligned with recommendations from international regulatory bodies. EXPERT OPINION: The limited availability of high-quality data and the absence of dependable Dev Tox methodologies render XAI an appealing avenue for systematically developing interpretable and transparent models, which hold immense potential for both scientific evaluations and regulatory decision-making.

Exploring the 1,3-benzoxazine chemotype for cannabinoid receptor 2 as a promising anti-cancer therapeutic.

The discovery of selective agonists of cannabinoid receptor 2 (CB2) is strongly pursued to successfully tuning endocannabinoid signaling for therapeutic purposes. However, the design of selective CB2 agonists is still challenging because of the high homology with the cannabinoid receptor 1 (CB1) and for the yet unclear molecular basis of the agonist/antagonist switch. Here, the 1,3-benzoxazine scaffold is presented as a versatile chemotype for the design of CB2 agonists from which 25 derivatives were synthesized. Among these, compound 7b5 (CB2 EC50 = 110 nM, CB1 EC50 > 10 µM) demonstrated to impair proliferation of triple negative breast cancer BT549 cells and to attenuate the release of pro-inflammatory cytokines in a CB2-dependent manner. Furthermore, 7b5 abrogated the activation of extracellular signal-regulated kinase (ERK) 1/2, a key pro-inflammatory and oncogenic enzyme. Finally, molecular dynamics studies suggested a new rationale for the in vitro measured selectivity and for the observed agonist behavior.

Rational Discovery of Antiviral Whey Protein-Derived Small Peptides Targeting the SARS-CoV-2 Main Protease.

In the present work, and for the first time, three whey protein-derived peptides (IAEK, IPAVF, MHI), endowed with ACE inhibitory activity, were examined for their antiviral activity against the SARS-CoV-2 3C-like protease (3CLpro) and Human Rhinovirus 3C protease (3Cpro) by employing molecular docking. Computational studies showed reliable binding poses within 3CLpro for the three investigated small peptides, considering docking scores as well as the binding free energy values. Validation by in vitro experiments confirmed these results. In particular, IPAVF exhibited the highest inhibitory activity by returning an IC50 equal to 1.21 µM; it was followed by IAEK, which registered an IC50 of 154.40 µM, whereas MHI was less active with an IC50 equal to 2700.62 µM. On the other hand, none of the assayed peptides registered inhibitory activity against 3Cpro. Based on these results, the herein presented small peptides are introduced as promising molecules to be exploited in the development of "target-specific antiviral" agents against SARS-CoV-2.

Stochastic simulations of mixed-lipid compartments: from self-assembling vesicles to self-producing protocells.

The computational platform ENVIRONMENT, developed to simulate stochastically reaction systems in varying compartmentalized conditions [Mavelli and Ruiz-Mirazo: Philos Trans R Soc Lond B Biol Sci 362:1789-1802, 2007; Physical Biology 7(3): 036002, 2010], is here applied to study the dynamic properties and stability of model protocells that start producing their own lipid molecules (e.g., phospholipids), which get inserted in previously self-assembled vesicles, made of precursor amphiphiles (e.g., fatty acids). Attention is mainly focused on the changes that this may provoke in the permeability of the compartment, as well as in its eventual osmotic robustness.