2010

2010

December

  • The future of toxicity testing: a focus on in vitro methods using a quantitative high-throughput screening platform. Shukla SJ. et al. Drug Discov Today. (5256)
  • KEYWORDS: PROJECT – RISK ASSESSMENT

    The Tox21 collaboration is presented as a bridge between the experimental toxicology expertise of the NTP, high-throughput screening technology at the NCGC and computational toxicology expertise at the EPA. The role of the NCGC in the Tox21 collaboration, the NCGC chemical library collection and technology used for Tox21 assays (see Table 1 for a list of current Tox21 assays (Cell viability Apoptosis, Membrane integrity, Mitochondrial toxicity, DNA damage, Cytokine, Nuclear receptor and toxicity pathway)) are detailed together with the goals of this collaboration combining technology, biology and computational methods to advance in vitro testing for toxicology concerns.

  • Pharmacogenomic biomarkers: new tools in current and future drug therapy. Sim SC. et al. Trends Pharmacol Sci. (5255)
  • KEYWORDS: ADME – BIOMARKERS – CYP450 – DRUG DISCOVERY – METABOLISM – PHARMACOLOGY – TRANSPORTERS

    This review provides a list of advantages and recommendations for the identification of new pharmacogenomic biomarkers thanks to new genome analysis techniques (see Box 1) and discusses several examples of both pharmacokinetic and pharmacodynamic biomarkers in different therapeutic areas (Analgesia, Cancer, Cardiovascular, CNS, Hyperuricemia and Infection, see Table 1 with drug information) and provides (see Box 1).

  • Report from the EPAA workshop: In vitro ADME in safety testing used by EPAA industry sectors. Schroeder K. et al. Toxicol In Vitro. (5254)
    Involved Partner: BSP
  • KEYWORDS: ADME – PROJECT – DRUG DISCOVERY – QSAR MODELING – REGULATORY GUIDELINES

    The European Partnership for Alternative Approaches to animal Testing (EPAA) working group reviews the outcome of the discussions and recommendations (reduction of number of animals used to determine ADME properties and considerations regarding in vitro and in silico assays) emerged during a workshop celebrated in Germany (ending 2008) with participants of different industry sectors (pharmaceuticals, cosmetics, industrial- and agro-chemicals). Firstly, an overview of differences in agencies regulatory guidelines is presented for pharmaceuticals (EMA), chemicals (ECHA), pesticides (EFSA) and cosmetics (SCCS). Secondly, reader can find a comprehensive compilation of in vitro and in silico models used for identifying ADME and specific organ toxicity (see tables 1 and 2); and also a description of recent supporting activities from industry, European Commission and Academia (EPPA, ECVAM, AcuteTox, Predict-IV, OSIRIS, START-UP 3Rs and EU Directorate General for research. In addition, authors present a list of in vitro liver models and their characteristics and limitations (see Table 4). Interestingly, most of the recommendations listed match perfectly with eTOX project proposes.

  • Identifying and characterizing chemical skin sensitizers without animal testing: Colipa’s research and method development program. Aeby P. et al. Toxicol In Vitro. (5253)
  • KEYWORDS: RISK ASSESSMENT – SKIN SENSITIZATION

    Colipa, the European trade association representing the interests of the cosmetic, toiletry and perfumery industry formed by 25 national associations, 18 major international companies, four supporting associations and two corespondent members currently collaborating with several academic and industrial research groups, presents this review where a series of alternative methodologies or techniques to avoid animal testing in case of skin sensitizers identification are discussed.

  • Cardiotoxicity. Brana I. et al. Ann Oncol. (5245)
  • KEYWORDS: CARDIOTOXICITY – DRUG SAFETY

    An other recent review of the incidence and the underlying mechanisms of cardiotoxicity (left ventricular dysfunction (anthracyclines), rhythm disturbances (QT interval prolongation, protein kinase C inhibitors), ischaemia (fluoropyrimides, taxanes, bevacizumab, sorafenib) induced by antineoplastic drugs. Proposal of prevention starts for an identification of the high-risk population, need of monitoring cardiac function, and development of cardioprotective drugs and early assessment of cardiotoxicity.

  • Cardiovascular effects of systemic cancer treatment. Senkus E. et al. Cancer Treat Rev. (5244)
  • KEYWORDS: CARDIOTOXICITY – DRUG SAFETY

    A recent widespread review of the cardiovascular complication of current anticancer treatments. Authors provide a list 29 antineoplastic drugs (see tables 1 and 2) that have been detected to induce cardiovascular morbidity (chronic heart failure, left ventricular dysfunction, pericarditis/myocarditis, arrhythmias, hypertension, cardiac ischemia, endomyocardial fibrosis, cardiac tamponade, thrombo-embolic complications, hypotension, angioedema, edema, QT prolongation, and bradycardia).

  • Can mutagenicity information be useful in an Integrated Testing Strategy (ITS) for skin sensitization?. Patlewicz G. et al. SAR QSAR Environ Res. (5243)
  • KEYWORDS: ASSAY DATA – MUTAGENICITY – QSAR MODELING – SKIN SENSITIZATION – STRUCTURE-BASED PREDICTION

    This study analyzes the correlation between mutagenicity information and skin sensitization event by use of Integrated Testing Strategy, based on a dataset of 100 chemicals. Authors provide 4 groups of data: 33 substances that were positive in both mutagenicity and sensitization assays, 35 that were negative in both assays, 15 outliers substances identified as mutagenic chemicals that not cause skin sensitization, and 17 outliers substances identified as skin-sensitizing chemicals that not cause mutagenicity disorder (see tables 1, 2, 3 and 4 respectively). From these results authors can conclude that there is a high probability to have a skin sensitizer when there is evidence of mutagenic effect of a chemical, meanwhile the viceversa relation is less probable and predictive.

  • Non-steroidal anti-inflammatory drugs: What is the actual risk of liver damage?. Bessone F. World J Gastroenterol. (5242)
  • KEYWORDS: DRUG SAFETY – HEPATOTOXICITY

    Based on epidemiologic studies, this article reviews the current knowledge of induced liver damage concerning a series of non-steroidal anti-inflammatory drugs (aspirin, diclofenac, sulindac, ibuprofen, coxibs, oxicams, and nimesulide) prescription.

  • The Drug Discovery Today: Technologies compiles a selection of articles devoted, as its Editorial presents, to 3D pharmacophore elucidation and virtual screening Wolber G. (5250)
    Involved Partner: IL
    • Fragment library design: efficiently hunting drugs in chemical space. Boyd SM. Drug Discov Today. (5257)
    • KEYWORDS: DATABASES – DRUG DISCOVERY – PHARMACOLOGY – STRUCTURE-BASED PREDICTION

      An excellent review of commercial suppliers of available fragment libraries (see Table 1), physico-chemical properties and ligand efficiency, sintetic considerations during the library design process, overview of published library characteristics (see Table 2, some from eTOX involved partners (AZ, Novartis, and ROCHE)) and further discussions about known drugs/biologically active compounds, natural compounds, novel scaffolds and a short analysis of diversity and complexity regarding current fragment libraries that evidences the need of focussed set prioritization.

    • History of 3D pharmacophore searching: commercial, academic and open-source tools. Van Drie JH. Drug Discov Today. (5248)
    • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – SOFTWARE – STRUCTURE-BASED PREDICTION

      A reference article for those who look for an entry-point into the past and current commercial (CoMFA, ALADDIN, MACCS-3D, APOLLO, Catalyst, DISCO, WIZARD, CONCORD, CAVEAT, Unity-3D, BioCAD, Apex-3D, COMPASS, CORINA, OMEGA, MOE, PHASE, LeadIT), academic (GASP, MIMIC, FEPOPS) and open source (CDK in Bioclipse, RDKIT.ORG) tools in the 3D pharmacophore searching field, with a final discussion devoted to answer why the pharmacophore concept contribute to facilitate the drug discovery process.

    • Docking compared to 3D-pharmacophores: the scoring function challenge. Hein M. et al. Drug Discov Today. (5249)
    • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – STRUCTURE-BASED PREDICTION

      This article provides the comparison of the pharmacophore-based and docking-based workflow for application in the structure-based virtual screening (see scheme in Figure 1, and principal features in Table 2) and lists 22 target studies recently reported employing pharmacophore-based methods or docking or both combined approaches (see Table 1).

    • Fragments: past, present and future. Whittaker M. et al. Drug Discov Today. (5252)
    • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – STRUCTURE-BASED PREDICTION

      A brief overview of the past (from the idea that ‘the activity of a drug molecule results from the sum of its parts’ to the fragment expansion (or evolution) concept, through the idea that ‘is to find the minimum active fragment (pharmacophore) of interest in a complex active molecule’), present (see Table 1 for a comparison between X-ray crystallography, Nuclear magnetic resonance and Bioassay of fragment screening methods) and future (Fragment Molecular Orbital, see box 1 for a brief description) methodologies regarding fragment-based drug discovery. The current limitation of these methodologies is the need of high resolution X-ray crystal structures of the fragment compounds for more realistic interpretations of the ligand-protein interactions. Examples of use from pharmaceutical companies site are reported (AZ applies fragment approaches to GPCRs, Novartis applies its in-house ‘Virtual Fragment Linking’ method also for GPCRs).

    • 3D pharmacophores as tools for activity profiling. Schuster D. Drug Discov Today. (5241)
    • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – SOFTWARE – STRUCTURE-BASED PREDICTION

      Review of several pharmacophore-based methodologies application in different scenarios (drug repositioning, natural product profiling, multitarget activity profiling and safety profiling). The authors summarize the most common quality parameters used in models evaluation (true/false positive hit rates, true/false negative hit rates, enrichment factor, goodnes of hits, or the receiver operation characteristic curve-area under the curve (ROC-AUC)) and highlight the diversity in model for a certain pharmacological target depending on different biding sites (active site vs. allosteric site), different binding modes to the same binding site, restrictive models for cherry-picking and general models for scaffold hopping and antitarget profiling.

    • In silico docking and scoring of fragments. Chen Y. et al. Drug Discov Today. (5239)
      Involved Partner: GSK
    • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – SOFTWARE – STRUCTURE-BASED PREDICTION

      Comparison of different fragment-based docking methodologies based on docking functional groups, fragment database screening or drug-like compound screening (see Table 1) and different scoring schemes in fragment-based docking: physics based, empirical, knowledge based or by concensus (see Table 2).

    • Strategies for 3D pharmacophore-based virtual screening. Seidel T. et al. Drug Discov Today. (5238)
      Involved Partner: IL
    • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – SOFTWARE – STRUCTURE-BASED PREDICTION

      Summary of the recent software strategies (Catalyst, Phase, LigandScout and MOE) to identify 3D pharmacophores as approaches to perform fast and accurate virtual screening (see Table 1 where Conformer generation and flexible search, Aligment strategy, geometric accuracy and scoring, Exclusion volume sphere interpretation, and Supported feature types are detailed). New developments should envisage for pattern matching and full tolerance sub-sampling to provide a higher geometric accuracy and more consistent results to the computational chemists using these software packages.

    • Conformations and 3D pharmacophore searching. Schwab CH. Drug Discov Today. (5231)
    • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – SOFTWARE – STRUCTURE-BASED PREDICTION

      Currently methods (Confirm/Catalyst, CEASAR, MOE, CORINA/ROTATE, OMEGA, MacroModel, Con/Gen and CONFORT) in the area of 3D structure and automatic conformation generation are presented with comparison between their features as conformers generators (see Table 1) to discuss their limits. With the aim to achieve accurate prediction of the bioactive conformaiton of a ligand, there is a room not only for the improvement of the existing technologies but also for new developments.

  • Thesaurus for histopathological findings in publically available reports of repeated-dose oral toxicity studies in rats for 156 chemicals. Nishikawa S. et al. J Toxicol Sci. (5246)
  • KEYWORDS: DATABASES – TEXT MINING

    Thanks to the collaboration of 13 institutions in Japan(see Table 1), this article reports a hystopathological findings thesaurus (see Figure 1) based on data gathered from reports of 28-day repeated-dose toxicity studies (RTS) for 156 chemicals, where majors parts of the thesaurus are devoted to the findings in the liver, kidney, stomach, adrenal, thyroid and testis organs. Note: The RTS database with this thesaurus will be publically available in 2012.

  • The kidney as a target organ in pharmaceutical research. Prunotto M. et al. Drug Discov Today. (5237)
    Involved Partner: ROCHE
  • KEYWORDS: BIOMARKERS – DRUG DISCOVERY – NEPHROTOXICITY – PATHWAYS – PHARMACOLOGY

    A well documented review about the kidney and their related mechanisms. A bried discussion of the relevant pathways known (RAAS, inflamation, hypoxia, phenotypical modulation processes, direct inhibition of effector cells and ECM remodeling (see Figure 3)) in chronic kidney disease (CKD) is provided, as well as description of in vitro and in vivo models, a list of 25 relevant targets in CKD (see Table 1), and a list of animal models in CKD available for pharmaceutical development (see Table 2).

  • Mitochondrial Membrane Potential Measurement of H9c2 cells Grown in High-Glucose and Galactose Containing Media Does Not Provide Additional Predictivity Towards Mitochondrial Assessment. Rana P. et al. Toxicol In Vitro. (5235)
    Involved Partner: PFIZER
  • KEYWORDS: ASSAY DATA – MITOCHONDRIAL TOXICITY – PHARMACOLOGY – RISK ASSESSMENT

    Based on data of 28 drugs evaluated on their mitochondrial membrane potential and cel viability (in two types of media: glucose and galactose) after 4 and 24 hours (see tables 2 and 3, respectively), authors check the relation between both types of data and evidence that mitochondrial membrane potential does not provide help to predict mitochondiral toxicity.

  • What can be learnt from an ecotoxicity database in the framework of the REACh regulation?. Henegar A. et al. Sci Total Environ. (5234)
  • KEYWORDS: DATABASES – QSAR MODELING – REGULATORY GUIDELINES – RISK ASSESSMENT

    A comprehensive analysis of a database of 3848 new chemicals registered in France and Europe during the last 20 years with aquatic ecotoxicity data with respect to 3 trophic levels (Algae EC50 72h, Daphnia EC50 48h and Fish LC50 96h) is presented based on only the ecotoxicological and physicochemical properties. The synthetic flowchart applied for the data-pruning procedure is shown in Figure 1, and the OECD categories contents of the database is presented in Table 1 using the the OECD (Q)SAR Application Toolbox.

  • Automated Selection of Compounds with Physicochemical Properties To Maximize Bioavailability and Druglikeness. Oashi T. et al. J Chem Inf Model. (5232)
  • KEYWORDS: ADME – DRUG DISCOVERY – MOLECULAR DESCRIPTORS – PHARMACOKINETICS

    This article presents a novel method to select compounds with a combination of physicochemical properties that maximize bioavailability and druglikeness based on compounds in the Workd Drug Index database (approximately 50,000 compounds consisting of all marketed drugs and pharmacologically active compounds with medical data: usage, drug target, mechanism of action, activity keywords, and adverse drug effects). Additionally some public databases (Maybridge, ChemBrigde and ChemDiv) were analyzed in terms of four physicochemical properties: MW, log P, hydrogen bond donor and acceptor (see Table 1). Validation of 4D-BA as a novel descriptor of bioavailability of druglike compounds is presented following the Biopharmaceutics Classification System (BCS) which is the scheme used by the FDA to categorize the orally administrate drugs (Class 1: high solubility and permeability, that are >90% absorbed; Class 2: low solubility and high permeability; Class 3: high solubility and low permeability; and Class 4: low solubility and permeability). Two examples of application are reported, one with a protein tyrosine phosphatase (SHP2) and one with the heme oxygenase, and evidence the utility of this descriptor at lead optimization level to prioritize compounds for chemical synthesis and biological assays.

  • Risk assessment and mitigation strategies for reactive metabolites in drug discovery and development. Thompson RA. et al. Chem Biol Interact. (5230)
    Involved Partner: AZ
  • KEYWORDS: DRUG DISCOVERY – HEPATOTOXICITY – RISK ASSESSMENT

    Authors propose use of an in vitro Hepatic Liability Panel (THLE cell toxicity, HepG2 cytotoxicity in galactose vs glucose medium and inhibition of human BSEP) alongside in vitro methods for the detection of reactive metabolites (RM) distinguishing between drug-related and patient-related factors (see Figure 1). An interesting discussion is reported regarding strategies for RM avoidance and minimization and the key components of an integrated RM safety hazard assessment in case of hepatotoxicity drug safety. An application of this strategy is shown by means of an internal study about troglitazone, rosiglitazone and pioglitazone drugs prescribed for Type 2 diabetes.

November

  • ‘Fuzziness’ in pharmacophore-based virtual screening and de novo design. Klenner A. et al. Drug Discov Today. (5229)
  • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY

    A helpful review in ‘fuzzy’ pharmacophore-based virtual screening methods (see Table 1) to be updated on this field of the drug discovery with discussion of concepts (fuzzy descriptor sets based on fuzzy set theory, fuzzy descriptor binning, fuzzy feature pairs/triplets/quadruplets, conformational ensembles, spatial tolerance in pharmacophore point assignment, Gaussian pharmacophoric points and graph kernel functions) and their application.

  • The graphical representation of ADME-related molecule properties for medicinal chemists. Ritchie TJ. et al. Drug Discov Today. (5228)
  • KEYWORDS: ADME

    An overview on different graphical representation for those molecular properties related to ADME that currently are available for drug discovery since evidences of link between the size, lipophilicity and H-bonding characteristics of drug molecules and their oral bioavailability was demonstrated in the 90s. List of graphics presented: Simple cell highlighting, Traffic light colouring, Craig plots, Flower plots, Egg plots, Oral bioavailability graphs, Golden Triangle, Face diagrams, Time series plots, Bioavailability plots, and Molecular ‘healthiness’ using a traffic-light pie chart.

  • Cross-study and cross-omics comparisons of three nephrotoxic compounds reveal mechanistic insights and new candidate biomarkers. Matheis KA. et al. Toxicol Appl Pharmacol. (5227)
    Involved Partner: BI
  • KEYWORDS: BIOMARKERS – HEPATOTOXICITY – NEPHROTOXICITY

    The InnoMed-PreTox project presents a comprehensive example of data extracted from a study based on omics analysis regarding three nephrotoxic compounds (see all tables for an evaluation of different data collected).

  • Molecular Topology Analysis of the Differences between Drugs, Clinical Candidate Compounds, and Bioactive Molecules. Chen H. et al. J Chem Inf Model. (5222)
    Involved Partner: AZ
  • KEYWORDS: DRUG DISCOVERY – MOLECULAR DESCRIPTORS – STRUCTURE-BASED PREDICTION

    Comparison between the number of compounds, and the media of number of heavy atoms and ClogP derives authors to define different topology classes for drugs, clinical candidates and bioactive compounds (see Table 1 for data values) based in different topological layers defined (see Figure 1).

  • Compound Set Enrichment: A Novel Approach to Analysis of Primary HTS Data. Varin T. et al. J Chem Inf Model. (5221)
    Involved Partner: NOVARTIS
  • KEYWORDS: DRUG DISCOVERY

    A novel effort to present a method to allow identification of latent and hidden hits with the aim of improve their activity and provide alternative to lead optimization process. Based on seven sets from PubChem database authors evidence the potential of this method into the rational selection of active type of compounds instead of individual recognition.

  • Polypharmacology Directed Compound Data Mining: Identification of Promiscuous Chemotypes with Different Activity Profiles and Comparison to Approved Drugs. Hu Y. et al. J Chem Inf Model. (5220)
  • KEYWORDS: ASSAY DATA – DATA MINING – PHARMACOLOGY

    Over 35000 compounds active against human targets with at least 1 microM potency are analysed in terms of promiscuity to define a list of chemotype for 458 targets belonging to 19 target families (see Table 1). Authors discuss about promiscuous scaffolds and chemotypes, concretely for a set of 1247 drugs extracted from DrugBank where 39 scaffolds represented 215 drugs.

  • In silico research in the era of cloud computing. Dudley JT. et al. Nat Biotechnol. (5219)
  • KEYWORDS: SOFTWARE

    Recent insights on the potential of cloud computing methodologies to be applied in silico research as innovative solutions that allow researchers to build complex computational workflows throgh drag-and-drop visual interfaces and to share standardized representations of workflows (e.g.: GenePattern, Taverna). Comparison between traditional challenges and cloud-computing solutions in terms of Data sharing, Software and applications, System and technical, and Access and preservation (see Table 1).

  • Mechanisms of Drug Toxicity and Relevance to Pharmaceutical Development. Guengerich FP. DMPK. (5218)
  • KEYWORDS:

    .

  • A structure-based approach for mapping adverse drug reactions to the perturbation of underlying biological pathways. Wallach I. et al. PLoS One. (5217)
  • KEYWORDS: DRUG SAFETY – STRUCTURE-BASED PREDICTION – RISK ASSESSMENT

    A recent example of data integration (730 drugs, 830 protein targets, human pathways and 506 side-effects) that predicts 185 side-effects pathway associations, which 32 were found supported by literature.

  • Probing Small-Molecule Binding to Cytochrome P450 2D6 and 2C9: An In Silico Protocol for Generating Toxicity Alerts. Rossato G. et al. ChemMedChem. (5216)
  • KEYWORDS: ASSAY DATA – CYP450 – METABOLISM – PHARMACOLOGY – QSAR MODELING

    Based on sets of 56 and 85 compounds for binding CYP2D6 and CYP2C9, respectively, two models are presented and validated as QSAR models (see workflow of Figure 5, and experimental binding affinities for external datasets of compounds that bind CYP2D6 and CYP2C9, respectively in tables 2 and 3).

  • The enhanced value of combining conventional and “omics” analyses in early assessment of drug-induced hepatobiliary injury. Ellinger-Ziegelbauer H. et al. Toxicol Appl Pharmacol. (5213)
    Involved Partners: BSP, SAD, ROCHE, BI
  • KEYWORDS: ASSAY DATA – HEPATOTOXICITY – RISK ASSESSMENT – TOXICOGENOMICS

    The InnoMed PredTox consortium presents a comprehensive example of combination of different ‘omics’ analyses applied to early assess of a hepatotoxicity event, drug-induced hepatobiliary injury, based in data from short-term toxicological studies in rats. Authors provide a list of 34 representative examples of genes found to be characteristically deregulated in response to drugs associated with hepatobiliary toxicity (see Table 1 and support material).

  • Comparative Analysis of QSAR Models for Predicting pKa of Organic Oxygen Acids and Nitrogen Bases from Molecular Structure. Yu H. et al. J Chem Inf Model. (5207)
  • KEYWORDS: ADME – ASSAY DATA – QSAR MODELING

    Based on 1143 organic compounds (580 oxygen acids and 563 nitrogen bases, see support material for experimental data) pKa prediction is presented provided by performance of ACD, SPARC, and 2 calibration of a semiempirical quantum chemical AM1 approach. The results for 6 acid subsets (see Table 4) and 9 base subsets (see Table 5) demonstrate an overall better performance for acids, although there is a substantial variation across subsets.

  • Grid-derived structure-based 3D pharmacophores and their performance compared to docking. Cross S. et al. Drug Discov Today. (5206)
  • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – STRUCTURE-BASED PREDICTION

    Since docking methods are solely designed for pose prediction and estimating the binding energy, they are complementary approaches to those structure-based pharmacophores approaches. Authors present the application of both techniques into the drug discovery process from virtual screening to metabolism prediction with a comparison between DOCK and FLAP programs based on several datasets from DUD database (see Figure 4). Their proposed GRID-derived structure-based pharmacophores approach combined with classical docking clearly illustrates regions where ligands can be modified to guide design improvements.

  • The scaffold hopping potential of pharmacophores. Hessier G. et al. Drug Discov Today. (5205)
    Involved Partner: SAD
  • KEYWORDS: ADME – DRUG DISCOVERY – MOLECULAR DESCRIPTORS – PHARMACOLOGY – STRUCTURE-BASED PREDICTION

    An overview of scaffold hopping techniques: core replacement (CAVEAT focuses on reproducing the geometry of the scaffold; Recore uses pharmacophoric constraints in addition to the basic geometric requirements; BROOD covers shape as well as the match of chemical features in addition to the geometric requirements of the exit vectors; CAVEAT strategy of geometric matching of exit vectors with the MOE pharmacophore tools; ParaFrag uses quantum chemically derived properties on the molecular surface in addition to the distances of related features for database searching; SHOP defines exit vectors on a reference scaffold as anchor points and generates molecular interaction fields for scaffold replacement), virtual screening (entire molecule is used to search in databases of available or virtual compounds), 2D approaches (CATS descriptor translates each heavy atom of the molecule into its corresponding pharmacophoric atom-type, hydrogen-bond donor, hydrogen-bond acceptor, liphophilic, among others, identifies of up to ten bonds the shortest path in bond length between all pairs of pharmacophoric features, and eventually maps these pairs into a vector suitable for comparison on Euclidean distance; Feature Tree descriptor maps physicochemical properties of chemical fragments onto a tree-like representation of the molecular graph preserving the relative arrangement of the molecular fragments; FLAP uses pharmacophoric quadruplets derived from molecular interaction fields) and 3D approaches (ROCS aligns compounds based on optimal shape overlap, uses atom-centered Gaussians for shape description and additionally, hydrogen-bond donor or acceptor; Discovery Studio extracts LUDI-type interaction features within the binding site which can be converted into pharmacophoric points, LigandScout detects pharmacophores by analyzing protein-ligand interactions, with focus on proven interaction points). Authors remark the main limit of 2D approaches, these methods do not consider conformational flexibility, and highlight 3D pharmacophores as a gold standard for rescaffolding of chemical series in lead optimization accompanied by data available (i.e. ADMET models).

  • A hybrid approach to advancing quantitative prediction of tissue distribution of basic drugs in human. Polin P. et al. Toxicol Appl Pharmacol. (5204)
  • KEYWORDS: ADME – ASSAY DATA – PHARMACOKINETICS

    A compilation of data regarding tissue distribution of basic drugs in human and rats (Principal mode of action are potent blockers of Na, K or Ca channel (14, 7), Dual mode of action (including potent blockers of Na or Ca channel (3, 2) and Other pharmacological mode of action (28,21); see Table 1 and figures 1 and 2, and Table 2 and figures 3 and 4, respectively) is presented to validate a correlation method for predicting tissue Kpu and Vss of basic drugs in comparison to previous methodologies existing in the literature, as Oie-Tozer model. The authors propose this hybrid approach to facilitate the prediction of phospholipidosis of basic drugs although it is limited to a chemical space defined by 45 drugs.

  • Application of toxicogenomics in hepatic systems toxicology for risk assessment: Acetaminophen as a case study. Kienhus AS. et al. Toxicol Appl Pharmacol. (5203)
  • KEYWORDS: HEPATOTOXICITY – RISK ASSESSMENT – TOXICOGENOMICS

    Based on acetaminophen case study, authors discuss about the application of the integration of toxicogenomics techniques (transcriptomics, proteomics, and metabolomics) with traditional toxicology measures (clinical chemistry, histopathology, etc) in hepatotoxicity concerns to take advantatge of its potential role in the risk assessment process. They provide an schematic overview of the risk assessment paradigm for hepatic systems toxicology (see Figure 2) and the toxicogenomics-based parallelogram (see Figure 3), together with a compilation of several toxicogenomics in vivo and in vitro studies (see tables 1 and 2, respectively). Finally they also shortly discuss about quantitative dose-response modeling in risk assessment for this compound.

  • Biological Reactive Intermediates (BRIs) Formed from Botanical Dietary Supplements. Dietz BM. et al. Chem Biol Interact. (5202)
  • KEYWORDS: DRUG SAFETY – RISK ASSESSMENT

    This article provides a summary of the toxicity concerns and resulting FDA actions of several botanical biological reactive intermediates (BRIs, see Table 1) in order to emphasize how their reactivity and selectivity have major influence on their toxicity and/or chemoprevention properties, as well as dose and time of exposure. Description of different types of BRIs is presented together with a brief overview of this class of chemical (see Figure 1 for examples of the most common types).

  • An integrated reactive metabolite evaluation approach to assess and reduce safety risk during drug discovery and development.Reese M. et al. Chem Biol Interact. (5200)
  • KEYWORDS: CYP450 – DRUG DISCOVERY – ENZYMES – HEPATOTOXICITY – METABOLISM – RISK ASSESSMENT

    A set of 114 hepatotoxic and 111 non-hepatotoxic of marketed drugs with in vitro bioactivation data and clinical dose information (GSH adduct assay, CYP MDI experiments, Microsomal covalent binding analysis, Hepatic Nrf2 gene expression) is analyzed to check their concordance with clinical hepatotoxicity and to develop an early reactive metabolite strategy. The results highlight the significant association of clinical dose and GSH adduct formation with hepatotoxicity and also demonstrate that a high percentage of compounds with covalent binding >200 pmol/mg were hepatotoxic in man. Strategies towards attenuation of metabolic bioactivation risk should reduce compound attrition and provide safer drugs.

  • PubChem as a public resource for drug discovery. Li Q. et al. Drug Discov Today. (5198)
  • KEYWORDS: DATABASES – DRUG DISCOVERY – ENZYMES – NUCLEAR RECEPTORS – PATHWAYS – PHARMACOLOGY

    Analysis of PubChem contents by means of function, 3D structure and biological pathway for each protein target stored there to evaluate the potency, selectivity and promiscuity of their bioactive compounds. Sequence similarity searches show that 2206 protein targets fell into 671 unique protein superfamilies (15% kinases superfamily, 2-3% other superfamilies: nuclear receptors, trypsin-like serine protease, src homology protein and zinc-dependent metalloprotease, and the rest, 67%, contains only one or two bioassay targets). This analysis remarks that 78% of these targets annotated have corresponding 3D structures with 100% sequence identity in the PDB database and that 507 (23%) of the 2206 protein targets are annotated to be involve in 287 pathways defined in the KEGG database.

October

  • Nature special issue: The 1000 Genomes Project
    The 1000 Genomes Project aims to provide a deep characterization of human genome sequence variation as a foundation for investigating the relationship between genotype and phenotype. This resource will aid our understanding of the role of genetic variation in human history, evolution and disease. This Nature Special accompanies the publication of the result of the pilot phase of the project, designed to develop and compare different strategies for genome-wide sequencing with high-throughput platforms. The site will be updated with companion papers as they become available.
  • Molecular dynamics simulations of protein dynamics and their relevance to drug discovery. Salsbury FR. Curr Opin in Pharmacol. (5201)
  • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY

    Since molecular dynamics have been successful in studying the protein folding problem, a brief overview of their basics (structure (experimental or from models coordinates, forcefield (CHARM, AMBER, GROMACS and NAMD) and parameters and solvation aspects) introduces their remark on the need of the development of a general method for finding conformations in and determining the free energies form simulations.

  • Structure of a cation-bound multidrug and toxic compound extrusion transporter. He X. et al. Nature. (5197)
  • KEYWORDS: ADME – METABOLISM – PHARMACOLOGY – TRANSPORTERS

    X-ray structure of MATE transporter NorM from Vibrio cholerae publication. MATE (multidrug and toxic compounds extrusion) family are vital in metabolite transport in plants, and mediate multidrug resistance (MDR) in bacteria and mammals, modulating the efficacy of many pharmaceutical drugs used in the treatment of a variety of diseases. Extensive discussions about its outward-facing conformation and also its cation-binding site in close proximity to critical residues for transport.

  • A sample storage management system for biobanks. Voegele C. et al. Bioinformatics. (5196)
  • KEYWORDS: DATABASES – SOFTWARE

    A single sample management system (SMS), both an Oracle database and an Oracle Forms web application, is presented by the International Agency for Research on Cancer (IARC) to afford limited number of analytic workflows that current commercial and academic systems offer with high restrictions to conduct with biobanks data. For further details check the supplementary material provided in this article and visit their site.

  • Navigating the human metabolome for biomarker identification and design of pharmaceutical molecules. Kouskoumvekaki I. et al. J Biomed Biotechnol. (5195)
    Involved Partner: DTU
  • KEYWORDS: DATABASES – METABOLISM – PATHWAYS – PHARMACOLOGY

    A comprehensive review of the most up-to-date metabolite and metabolic pathways resources (see Table 2), together with a summary of the statistical and machine-learning tools (see Table 1) used for the analysis of clinical metabolomics data. Authors provide detailed introduction to Metabolomics field, discussion for three main resources (HMDB, SMPDB and T3DB), point out the identification of biomarkers in case of Cancer, Diabetes, and Neurological and Other diseases, and finally, review the role of metabolomics in drug discovery in several polypharmacology studies, with special mention of polypharmacology of natural compounds.

  • Development of Quantitative Structure-Activity Relationship (QSAR) Models to Predict the Carcinogenic Potency of Chemicals. II. Using Oral Slope Factor as a Measure of Carcinogenic Potency. Wang NCY. et al. Regul Toxicol Pharmacol. (5194)
  • KEYWORDS: ASSAY DATA – CARCINOGENICITY – DATABASES – QSAR MODELING

    This study aims to circumvent the lack of animals or human studies in the literature to determine oral slope factors (OSFs) as measure of carcinogenic potential by developing QSAR models based on male/female human, rat and mouse biossays data for 70 chemicals from IRIS database (see Table 1). After externally validation, authors remark that although their models show promising predictions, they are not ready for regulatory decision, but they could be considered for evaluating and/or prioritizing chemicals.

  • Characteristics of Known Drug Space. Natural Products, their Derivatives and Synthetic Drugs. Bade R. et al. Eur J Med Chem. (5169)
  • KEYWORDS: DATABASES – DRUG DISCOVERY – MOLECULAR DESCRIPTORS

    After compilation of 1000 marketed drugs and their categorization in different drug types (Natural (subcategories: Macrocycle, Nucleotide/nucleoside, Penicillin, Peptide, Polycyclic, Simple, Steroid and Sugar, see Figure 2), Natural derivative, Synthetic, and Assumed synthetic, see Figure 3), this study provides statistics (see Table 1) by means of several molecular descriptors (molecular weight, logP, Polar surface area, rotatable bonds, and chiral centres) to analyze the Known Drug Space (KDS), since KDS gives drug designers a larger volume to work in compared to drug-like chemical space (see Figure 1 and 12).

  • Predicting Phospholipidosis Using Machine Learning. Lowe R. et al. Mol Pharm. (5192)
  • KEYWORDS: ASSAY DATA – MOLECULAR DESCRIPTORS – PHOSPHOLIPIDOSIS

    Using a data set mined from the literature (185 compounds, 102 positive and 83 negative) for phospholipidosis (see Support material data), this study stands out that circular fingerprints perform better models respect to E-Dragon descriptors or a combination of both options. The authors remark that for more reliable and robust predictivity, there is a clear need for a much larger publicly available database of phospholipidosis inducing potential.

  • Development of a mechanistic SAR model for the detection of phototoxic chemicals and use in an integrated testing strategy. Ringeissen S. et al. Toxicol In Vitro. (5191)
  • KEYWORDS: ASSAY DATA – PHOTOTOXICITY – QSAR MODELING – REGULATORY GUIDELINES – STRUCTURE-BASED PREDICTION

    This article presents the first attempt to develop a non-local mechanistic SAR model to predict phototoxic effects mediated by the production of reactive oxygen species (ROS). Authors describe their phototoxicity testing strategy (see Figure 8 ) where the SAR model is used as a pre-screen to help in reducing further experimental testing. Their proposed model predicts chemicals to be ‘‘phototoxic or photodegradable”, or ‘‘non-phototoxic and non-photodegradable” (see Support material data).

  • A Maximum Common Subgraph Kernel Method for Predicting the Chromosome Aberration Test. Mohr J. et al. J Chem Inf Model. (5190)
    Involved Partner: BSP
  • KEYWORDS: ASSAY DATA – GENOTOXICITY – STRUCTURE-BASED PREDICTION

    Taken into account Chromosome Aberration (CA) data from information contained in the literature and in VITIC software, this article describes a machine learning approach for predicting the outcome of CA assay based on the structure of a certain investigated compound. The final CA data set provided contains 940 unique compounds as canonical SMILES together with the corresponding in vitro CA test results and data sources (see Support Material files) after removing compounds with contradictory results in different sources.

  • WENDI: A tool for finding non-obvious relationships between compounds and biological properties, genes, diseases and scholarly publications. Zhu Q. et al. J Cheminform. (5188)
  • KEYWORDS: DATABASES – DATA MINING

    Eli Lilly presents their public web-service based tools for integrative mining of chemical and biological information, called WENDI (Web Engine for Non-obvious Drug Information). The WENDI interface is organized into six major sections: 1) Predictive models results (including ToxTree analysis), 2) Activities of similar compounds (PubChem, Gene Ontology, DrugBankand MRTD), 3) Similar compounds from chemogenomics data (CTD, ChEMBL), 4) Similar compounds from Systems data (KEGG), 5) Similar compounds in the literature (Medline), and 6) Inactivities of similar compounds (PubChem). See Figure 1 for an overall architecture of storage, interface, aggregation and interaction layers employed in WENDI.

  • Prediction of Michael-Type Acceptor Reactivity toward Glutathione. Schwöbel JAH. et al. J Chem Inf Model. (5185)
    Involved Partner: LJMU
  • KEYWORDS: ASSAY DATA – ENZYMES – METABOLISM

    This article introduces a model to predict the reactivity, in terms of GSH adduct formation, of apha,beta-unsaturated compounds based on the experimental and predicted rate constants for Adduct Formation with Glutathione (GSH) and the RC50 values of Schultz’s Chemoassay for a dataset of 66 Michael-type acceptor compounds (see Table 1).

  • PZIM: A Method for Similarity Searching Using Atom Environments and 2D Alignment. Berglund AE. et al. J Chem Inf Model. (5184)
    Involved Partner: PFIZER
  • KEYWORDS: DATABASES – SOFTWARE

    Description of a new similarity searching algorithm, called PZIM, that incorporates a number of concepts, including atom environment, atom similarity, and molecular alignment, within a single adjustable approach. Its usefulness was compared to seven other similarity-searching methods (see tables 1 and 2) on nine data sets of different enzymes inhibitors and GPCRs antagonists (see Table 3).

  • Lead Optimization Using Matched Molecular Pairs: Inclusion of Contextual Information for Enhanced Prediction of hERG Inhibition, Solubility, and Lipophilicity. Papadatos G. et al. J Chem Inf Model. (5173)
    Involved Partner: GSK
  • KEYWORDS: ADME – DATA MINING – hERG – PHARMACOKINETICS – PHARMACOLOGY

    Since the use of Matched Molecular Pairs (MMP) has became recently of extended interest to assist the medicinal chemists in the lead optimization stage of drug discovery, this article provides the 30 most frequent transformations for (a) hERG Inhibition, (b) Solubility, and (c) Lipophilicity, along with the percentages of occurrences where each effect was Favorable, Unfavorable, or Neutral (see Table 3) based on data from the GlaxoSmithKline (GSK) corporate database relating to these three important ADME properties.

  • Modeling liver-related adverse effects of drugs using knearest neighbor quantitative structure-activity relationship method. Rodgers AD. et al. Chem Res Toxicol. (5171)
  • KEYWORDS: ASSAY DATA – ENZYMES – HEPATOTOXICITY – MOLECULAR DESCRIPTORS – QSAR MODELING

    Based on data (see support material files) from the Human Liver Adverse Effects Database with approximately 500 compounds with five serum enzyme markers of toxicity information (alkaline phosphatase (ALP), alanine aminotransferase (ALT), aspartate aminotransferase (AST), lactate dehydrogenase (LDH), and γ-glutamyl transpeptidase (GGT)), the development of QSAR models predictive of human liver adverse effects of drugs is presented (see Figure 1 for their workflow proposal), and results suggest that they should be considered as complementary information to prioritize compounds for preclinical screening.

September

  • In silico approaches to predicting cancer potency for risk assessment of genotoxic impurities in drug substances. Bercu JP. et al. Regul Toxicol Pharmacol. (5168)
  • KEYWORDS: ASSAY DATA – CARCINOGENICITY – GENOTOXICITY – RISK ASSESSMENT – STRUCTURE-BASED PREDICTION

    Based on data of 694 compounds extracted from the CPDB (see distribution of carcinogenicity data in Table 1 and in the supplementary material), this article presents that MultiCASE classification models allow the prediction of carcinogenic potency class and VISDOM regression models predict a numerical TD50, and in addition proposes a step-wise approach to calculated predicted numerical TD50 values for those not potent compounds. Validation statistics for both types of models is detailed in Table 2 and Figure 2 shows the proposed decision tree.

  • Use of Genotoxicity Information in the Development of Integrated Testing Strategies (ITS) for Skin Sensitization. Mekenyan O. et al. Chem Res Toxicol. (5164)
  • KEYWORDS: ASSAY DATA – GENOTOXICITY – MUTAGENICITY – REGULATORY GUIDELINES – SKIN SENSITIZATION

    Application of Integrated Testing Strategies on genotoxicity data to stand out skin sensitization potential of sets of compounds provides lists of 1) alerts causing DNA binding and protein binding by the same mechanism (Table 1), by practically the same mechanism (Table 2), by different mechanism (Table 4); 2) DNA alerts which partially overlap with alerts present for skin sensitization (Table 3); 3) protein binding alerts with DNA reactivity and with the same active form (Table 6) or with a different active form (Table 7), or that are likely to be DNA reactive but lack in any experimental evidence (Table 8 ) or those evaluated not to be DNA reactive (Table 9); and 4) sensitization alerts not causing Ames and Chromosome Aberration (Table 11). For inclusion into TIMES (Tissue Metabolism Simulator, a hybrid expert system) this study postulates a number or new alerts for mutagenicity (Table 5).

  • HNF4α — role in drug metabolism and potential drug target?. Hwang-Verslues WW. et al. Curr Opin Pharmacol. (5162)
  • KEYWORDS: CYP450 – METABOLISM – NUCLEAR RECEPTOR – PHARMACOLOGY

    Recent update on drug metabolism focused on HNF4alpha nuclear receptor as a master regulator of liver-specific gene expression, especially those genes involved in lipid transport and glucose metabolism (see Table 1 for a list of verified and predicted HNF4alpha Phase I and Phase II human target genes).

  • Ontologies in Quantitative Biology: A Basis for Comparison, Integration, and Discovery. Jensen LJ. et al. PLoS Biol. (5161)
  • KEYWORDS: DRUG DISCOVERY – TEXT MINING

    A brief overview of data integration in biomedical ontologies that discusses about typical structures of ontologies (Hierarchy Tree, Directed acyclic graph and Direct cyclic graph, see Figure 1), their use for discovery and the semantic annotation of scientific publications (see Box 1)..

  • Bioactivation of Isothiazoles: Minimizing the Risk of Potential Toxicity in Drug Discovery. Teffera Y. et al. Chem Res Toxicol. (5159)
  • KEYWORDS: ENZYMES – METABOLISM – PHARMACOKINETICS – RISK ASSESSMENT

    Focused in isothiazoles chemotype case, this article reports the identification of a novel glutathione conjugate for a methyl-substituted isothiazole containing compound based on several in vitro and in vivo analysis and provides lists of metabolites for two compounds resulting of the replacement of the isothiazole ring with an isoxazole or a pyrazole which reduce their bioactivation retaining the pharmacokinetics and pharmacodynamics desirable.

  • Toxicogenomics and cancer risk assessment: A framework for key event analysis and dose–response assessment for nongenotoxic carcinogens. Bercu JP. et al. Regul Toxicol Pharmacol. (5158)
  • KEYWORDS: ASSAY DATA – RISK ASSESSMENT – SYSTEMS BIOLOGY – TOXICOGENOMICS

    Several key toxicogenomics analysis are presented in order to determine a threshold for nongenotoxic carcinogens considering a non-linear dose-response. This article provides data from a KEGG and a GO analysis following oral exposure to fenofibrate during 2 days (see tables 2 and 3), and to methapylrilene during 7 days (see tables 4 and 5).

  • Trainable structure-activity relationship model for virtual screening of CYP3A4 inhibition. Didziapetris R. et al. J Comput Aided Mol Des. (5157)
  • KEYWORDS: ASSAY DATA – CYP450 – MOLECULAR DESCRIPTORS – PHARMACOLOGY – QSAR MODELING – STRUCTURE-BASED PREDICTION

    Based on two sets of compounds from literature and PubChem (see tables 2 and 3), a new SAR model is presented for predicting CYP3A4 inhibition using the novel GALAS modeling method (see Figure 1 for the workflow of the GALAS model trainability testing) which is a combination of baseline global QSAR model and local similarity based corrections and a useful tool for virtual screening performance in order to help in candidates prioritization.

  • Discovering drug–drug interactions: a text-mining and reasoning approach based on properties of drug metabolism. Tari L. et al. Bioinformatics. (5156)
    Involved Partner: ROCHE
  • KEYWORDS: DRUG-DRUG INTERACTIONS – METABOLISM – NLP – TEXT MINING

    A novel approach that integrates text mining and automated reasoning to derive drug-drug interactions (DIDs) is presented and a study, based on a selection of 265 drugs extracted from DrugBank that resulted in a gold standard of 494 DIDs, evidences its applicability to real-world scenarios (see Figure 2 for an overview of their approach in extracting drug interactions, extraction and reasoning phases).

  • Using Open Source computational tools for predictiong human metabolic stability and additional ADME/TOX properties. Gupta R. et al. Drug Metab Dispos. (5155)
    Involved Partner: PFIZER
  • KEYWORDS: ADME – ASSAY DATA – METABOLISM – P-gp – QSAR MODELING

    This article provides ADME/Tox models based on 4 different datasets: 1) human liver microsomal stability (200,000 compounds, measurement of the apparent intrinsic clearance (Clint) of a compound in human liver); 2) permeability (70,000 compounds, cellular passive apparent permeability (Papp)); 3) P-gp efflux activity (60,000 compounds, measured by cell line transfected with the human MDR-1 gene); and 4) solubility data (1,300 compounds, from literature).

  • Collaborative development of predictive toxicology applications. Hardy B. et al. J Cheminform. (5151)
  • KEYWORDS: PROJECT – DRUG SAFETY – QSAR MODELING – REGULATORY GUIDELINES – STRUCUTRE-BASED PREDICTION

    This research article presents a comprehensive review on the collaborative approaches regarding predictive toxicology within the OpenTox Framework, from their objectives and design principles to description of the toxicity data considered, vocabularies standardization in ontologies, and QSAR approaches. Moreover, user requirements and system architecture are detailed facing the REACH endpoints and OECD guidelines, and their developed applications (ToxPredict and ToxCreate) are also presented.

  • A review of the electrophilic reaction chemistry involved in covalent DNA binding. Enoch SJ. et al. Crit Rev Toxicol. (5141)
    Involved Partner: LJMU
  • KEYWORDS: CARCINOGENICITY – MUTAGENICITY – STRUCTURE-BASED PREDICTION

    A comprehensive outline of the current knowledge in the area of structural alerts for both mutagenicity and genotoxic carcinogenicity endpoints. A list of 62 new alerts is provided (see Table 3) together with their assignment to a set of different mechanism domains defined (Acylation, Michael addition, Schiff base formation, aromatic nucleophilic substitution, unimoleclar aliphatic nucleophilic substitution, bimolecular aliphatic nucleophilic substitution, and reaction involving free radicals).

  • Integrating (Q)SAR models, expert systems and read-across approaches for the prediction of developmental toxicity. Hewitt M. et al. Reprod Toxicol. (5154)
    Involved Partner: LJMU
  • KEYWORDS: ASSAY DATA – QSAR MODELING – REPRODUCTIVE AND DEVELOPMENTAL TOXICITY – SOFTWARE – STRUCTURE-BASED PREDICTION

    Discussion, focused in the case of reproductive and developmental toxicity prediction, of limitations of current modeling methods when they are used in isolation. This article provides an evaluation of combining knowledge from the different predictive approaches in an effort to increase predictive performance considering the information from all methods (CAESAR (Q)SAR model, Derek for Windows, OECD (Q)SAR Application Toolbox oestrogen receptor-binding profiler and Read-across (where available)) by application of a weight-of-evidence approach. Study based on 233 training set compounds and 57 test compounds with FDA classification and binary classification of teratogenicity outcome (see Table 1 for original data, and Table 2 for predictions obtained for the test set from each method individually and overall weight-of -evidence prediction).

  • Mapping drug physico-chemical features to pathway activity reveals molecular networks linked to toxicity outcome. Antczak P. et al. PLoS One. (5147)
  • KEYWORDS: ASSAY DATA – BIOMARKERS – MOLECULAR DESCRIPTORS – NEPHROTOXICITY – PATHWAYS

    A computational strategy is presented to address a systematic approach to model the interaction between chemical features, molecular networks and toxicity outcomes. Authors exemplify its application with experimental data from a 28-day repeated-dose study on rat kidney profiles after 5 days of drug exposure (88 chemicals, where 22 were known to induce renal tubular degeneration, see Support Material) and highlight a list of 21 KEGG pathways that were found to be significantly perturbed by nephrotoxic chemicals (see Table 1), and a list of 19 KEGG pathways whose activity could be predicted by combinations of Physical Chemical Features (see Table2).

  • Optimized hydrophobic interactions and hydrogen bonding at the target-ligand interface leads the pathways of drug-designing. Patil R. et al. PLoS One. (5146)
  • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – QSAR MODELING – STRUCTURE-BASED PREDICTION

    A docking-based study (Discovery Studio software modules LigandFit, CDOCKER,and ZDOCK) of a selection of 39 molecules of 4-amino- substituted 1H-pyrazolo [3,4-d]pyramide compounds (see Table 1 in Support Material for their biological activity ans structural alignment) demonstrates that hydrogen bonding and optimized hydrophobic interactions both stabilize the ligands at the target site, and have some role to improve both binding affinity and drug efficacy. 3D-QSAR and Molecular Field Analysis calculations are presented to quantify the role of those weak interactions.

  • VoteDock: Consensus docking method for prediction of protein-ligand interactions. Plewczynski D. et al. J Comput Chem. (5145)
  • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – SOFTWARE

    Based on a set of 1300 protein-ligand pairs from refined PDBbind database, this article presents novel methods (MetaPose, MetaScore and VoteDock) as consensus approach to predict both protein-ligand complex structure and its corresponding binding affinity, that use as input the results from 7 widely used docking programs (Surflex, LigandFit, Glide, GOLD, FlexX, eHiTS, and AutoDock, see Figure 1 for an scheme of the algorithm applied).

  • Oral LD50 toxicity modeling and prediction of per- and polyfluorinated chemicals on rat and mouse. Bhhatarai B. et al. Mol Divers. (5144)
  • KEYWORDS: ASSAY DATA – MOLECULAR DESCRIPTORS – QSAR MODELING – REGULATORY GUIDELINES – STRUCTURE-BASED PREDICTION

    The EU project CADASTER presents an example of QSAR modeling for a set of per- and polyfluorinated (PFCs) compounds (58 for mouse oral and 50 for rat oral data, see tables of Support Material), collected from ChemID plus after structural filtering curation. See Figure 5 for the structural depiction of some of the long chain PFCs prioritized by this study as highly toxic which are within the applicability domain of both mouse and rat LD50 inhalation and LD50 oral models beyond PC1 cut-off of 1.25. A reduced set of more than 600 molecular descriptors (DRAGON computation) was subjected to variable selection method using Genetic algorithm in each dataset. See their presentation untitled ‘Modeling of physicochemical properties for perfluorinated chemicals using a data integration approach’ presented at the 2nd international workshop on new developments of fluorinated surfactants, Idstein, Germany, June 17-19, 2010 by Wolfram Teetz.

  • Cause-effect relationships in medicine: a protein network perspective. Fliri AF. et al. Trends Pharmacol Sci. (5142)
    Involved Partner: BI
  • KEYWORDS: DRUG DISCOVERY – NETWORKS – PATHWAYS – PHARMACOLOGY

    A brief discussion on uses for disease network topology (targets identification, multiple network targets, biomarkers, drug combinations, drug repurposing) focused on information flow in different cause-effect relationships analyses for a greater understanding at the organism, organ, tissue, cellular, and molecular level.

August

  • Genetic polymorphism in metabolism and host defense enzymes: Implications for human health risk assessment. Ginsberg G. et al. Crit Rev Toxicol. (5140)
  • KEYWORDS: METABOLISM – PHARMACOKINETICS – RISK ASSESSMENT

    This extensive review analyzes widely those polymorphic enzymes with remarkable role in metabolism concerns by providing background on each enzyme function, genotype effect on phenotype, the frequency of influential alleles in the general population, and the pharmacokinetic (clinical drug trials) and disease susceptibility (molecular epidemiology studies) evidence that these polymorphisms can alter xenobiotic fate and susceptibility to human disease (see Table 9 for a summary of Implications of SNPs of 11 enzymes (6 SULT, 4 UGT, 1 EPHX and 1 NQO) for human health risk assessment).

  • Analysis of published data for top concentration considerations in mammalian cell genotoxicity testing. Parry JM. et al. Mutagenesis. (5139)
  • KEYWORDS: ASSAY DATA – GENOTOXICITY – REGULATORY GUIDELINES

    Since the OECD genotoxicity test guidelines require testing of chemicals using mammalian cells in vitro to concentrations as high as 10 mM (5000 mg/ml) and recently, a draft revision of the ICH of Technical Requirements for Registration of Pharmaceuticals for Human Use genotoxicity test guidelines recommend that testing concentrations should be reduced to 1 mM (500 mg/ml), the authors present their evaluation to assess the impact that this lowering would have on the outcome of in vitro genotoxicity. Based on a database of 384 chemicals classified as rodent carcinogens (see Supplementary Material), they report Ames test results and the test concentrations that produced positive results in the mouse lymphoma assay (MLA), in vitro chromosome aberration (CA) assay and in vitro micronucleus test, and their results suggest that current 10 mM top concentration can be reduced without any loss of sensitivity in detecting rodent carcinogens.

  • A safety assessment of branched chain saturated alcohols when used as fragrance ingredients. Belsito D. et al. Food Chem Toxicol. (5138)
  • KEYWORDS: ASSAY DATA – CARCINOGENICITY – DRUG SAFETY – METABOLISM – MUTAGENICITY– PHARMACOKINETICS – PHOTOTOXICITY – REPRODUCTIVE AND DEVELOPMENTAL TOXICITY – RESPIRATORY SENSITIZATION – SKIN SENSITIZATION

    Based on a group of 12 primary, 5 secondary and 3 tertiary alcohol compounds (see Table 1 for their CAS Number, synonyms, structure and some physicochemical data (Henry’s law, log Kow, MW, Vapor pressure and Water solubility)), this review presents a full detailed safety assessment of branched chain saturated alcohols. Appropriate discussion and experimental data gathered in several tables are provided for different toxicological studies (Acute toxicity, Repeated dose toxicity, Genotoxicity, Mutagenicity, Carcinogenicity, Reproductive and Developmental toxicity, Skin and Respiratory sensitization, Phototoxicity and Photoallergenicity).

  • Computational quantum chemistry and adaptive ligand modeling in mechanistic QSAR. De Benedetti PG. et al. Drug Discov Today. (5136)
  • KEYWORDS: MOLECULAR DESCRIPTORS – QSAR MODELING – STRUCTURE-BASED PREDICTION

    This article illustrates the concept of adaptive drugs and the role of Computational Quantum Chemistry-derived intermolecular interaction propensity descriptors of the ligands in interpretative and predictive QSAR models (local, pharmacophore-based and global) with a series of selected examples of ligand-based QSAR modeling studies (enzyme-inhibitor interactions and G protein-coupled receptor ligands, see Table 1 for QSAR models equations) to evidence the potential of such methods/approaches.

  • Ligand efficiency indices for an effective mapping of chemico-biological space: the concept of an atlas-like representation. Abad-Zapatero C. et al. Drug Discov Today. (5133)
    Involved Partner: LJMU
  • KEYWORDS: DATABASES – DRUG DISCOVERY – MOLECULAR DESCRIPTORS – PHARMACOLOGY

    Following the concept of chemical-biological space, this article presents the concept of an atlas-like representation of both domains. Authors propose mapping of compounds (empirical formula or SMILES string) with a measure of affinity (Ki) or activity (IC50) by extraction of certain molecular descriptors (NPOL (number of N, O atoms), NHEA (number of non-hydrogen atoms), MW and tPSA (or any other estimate of the PSA)) to calculate several Ligand Efficiency Indices (see Table 1 for LEIs definition). Approaches like this should help to point the way towards a more effective and efficient drug discovery process for an extensive public and proprietary target-ligand databases of inhibitors and commercial drugs at different stages.

  • Semi-automated ontology generation within OBO-Edit. Wächter T. et al. Bioinformatics. (5132)
  • KEYWORDS: DATA MINING – NLP – SOFTWARE – TEXT MINING

    A new system, the Dresden Ontology Generator for Directed Acyclic Graphs (DOG4DAG), which supports the creation and extension of OBO ontologies by semi-automatically generating terms, definitions and parent–child relations from text in PubMed, the web and PDF repositories is presented. DOG4DAG is seamlessly integrated into OBO-Edit. It generates terms by identifying statistically significant noun phrases in text. For definitions and parent–child relations it employs pattern-based web searches. Authors discuss evaluation of each generation step using manually validated benchmarks and evidence that there is not other validated system that achieves comparable results. See Table 1 for an Screenshot of the OBO-Edit ontology generation tool showing the three steps ‘Term Generation’, ‘Definition Generation’ and ‘Add To Ontology’ for a certain example.

  • Phospholipidosis as a function of basicity, lipophilicity, and volume of distribution of compounds. Hanumegowda UM. et al. Chem Res Toxicol. (5131)
  • KEYWORDS: ADME – ASSAY DATA – PHOSPHOLIPIDOSIS

    This article reports how the combination of three physicochemical properties (Vd, pKa of the most basic group and clog P) of compounds provides greater concordance in predicting drug-induced phospholipidosis (PLD) in vivo, concretely for a set of 103 compounds extracted from the literature (53 PLD inducing and 50 non-PLD inducing, see Table 1 for all data related).

  • A Semantic Web Ontology for Small Molecules and Their Biological Targets. Choi J. et al. J Chem Inf Model. (5126)
  • KEYWORDS: DATABASES – PHARMACOLOGY – TEXT MINING

    Lastly, Semantic Web technologies have been proposed to avoid semantic incompatibilities as a solution to integrate data from diverse sources to address complex biological problems. This article presents the generation of a Small Molecule Ontology (SMO) employing both the Semantic Web technologies Resource Description Framework (RDF) and the Web Ontology Language (OWL) to represent concepts and provide unique identifiers for biologically relevant properties of small molecules and their interactions with biomolecules, such as proteins. Their results, based on a combination of three public data sources (DrugBank, PubChem and UniProt), evidence the potential of these technologies in drug discovery (chemical, biological and pharmacological concerns).

  • QSAR-based solubility model for drug-like compounds. Gozalbes R. et al. Bioorg Med Chem. (5125)
  • KEYWORDS: ADME – ASSAY DATA – DATABASES – MOLECULAR DESCRIPTORS – QSAR MODELING – STRUCTURE-BASED PREDICTION

    Another example of QSAR modeling about solubility data (51 compounds, see Table 4 for drug names, solubility values and prediction performed (low, medium or high). This study compares chemical structures from the FDAMDD and PHYSPROP databases to allow the selection of properties that are more efficient in discriminating drug-like compounds from other chemicals (see Table 1A for a list of properties and counts used to characterize the structures from both databases, Table 1B for a list of selected parameters and cut-off values applied, and Table 2 for a list of Descriptors included in the Axis-Parallel Decision Tree model).

  • WizePairZ: A Novel Algorithm to Identify, Encode, and Exploit Matched Molecular Pairs with Unspecified Cores in Medicinal Chemistry. Warner DJ. et al. J Chem Inf Model. (5124)
    Involved Partner: AZ
  • KEYWORDS: ADME – ASSAY DATA – DATABASES – SOFTWARE

    A new automated expert system for medicinal chemistry, WizePairZ, based on matched molecular pairs analysis is presented in this article with an extensive description as a collection of algorithms. They exemplify its applicability in a set of 11 histone deacetylase (HDAC) inhibitors (see Table 1 for structures, and log D7.4 and pIC50 values) and evidence its utility for a given simple data set containing structures, biological activity, and lipophilicity information, to extract key SARs and use the relationships to propose further chemical targets for synthesis in an informed and intelligent fashion.

  • The Chemistry Central Journal presents a special issue with proceedings from the ‘CAESAR QSAR Models for REACH’ conference celebrated in Milan, Italy, 10-11 March 2009. Here we highlight some of them:
    • The CAESAR project for in silico models for the REACH legislation. Benfenati E. Chem Cent J. (5130)
    • KEYWORDS: PROJECT – QSAR MODELING – REGULATORY GUIDELINES

      This introductory article explains how CAESAR project develops in silico models specifically designed for the REACH legislation and dedicates special attention to facilitate user access to them by a graphical interface on their website.

    • An open source multistep model to predict mutagenicity from statistical analysis and relevant structural alerts. Ferrari T. Chem Cent J. (5122)
    • KEYWORDS: ASSAY DATA – MOLECULAR DESCRIPTORS – MUTAGENICITY – QSAR MODELING – REGULATORY GUIDELINES – STRUCTURE-BASED PREDICTION

      This article presents validation on a large public set of molecular structures (4204 compounds, see Support Material flie 1 for data information) of a cascade model developed to generate models for mutagenicity prediction from statistical analysis and relevant structural alerts (see Support Material file 4 for both SAs ruleset considered (1: acyl halides, propiolactones or propiosultones, quinones, hydrazines, aliphatic azo and azoxy, alkyl carbamate and thiocarbamate, polycyclic aromatic hydrocarbons, alkyl and aryl N-nitroso groups, azide and triazene groups, aromatic nitroso group, aromatic mono- and dialkylamine and aromatiz diazo; and 2: epoxides and aziridines, aliphatic halogens, heterocyclic polyclic aromatic hydrocarbons and nitro-aromatic).

    • New public QSAR model for carcinogenicity. Fjodorova N. et al. Chem Cent J. (5129)
    • KEYWORDS: ASSAY DATA – CARCINOGENICITY – MOLECULAR DESCRIPTORS – QSAR MODELING – REGULATORY GUIDELINES – STRUCTURE-BASED PREDICTION

      This article provides two alternative models for prediction carcinogenicity (model A: with 8 MDL descriptors (see Table 8), and model B: with 12 Dragon descriptors (see Table 9)) based on a dataset of 805 non-congeneric chemicals extracted from Carcinogenic Potency Database, CPDBAS, (see Support Material for Names, CAS number, and TD50 Rat_mg values). Models are generated according to the OECD principles for the validation of QSAR by application of Counter Propagation Artificial Neural Network (CP ANN) algorithm. Additionally, they report the structural diversity of this CAESAR dataset by presence of specific structural alerts (SAs) extracted from ToxTree program (see Table 6).

    • CAESAR models for developmental toxicity. Cassano A. et al. Chem Cent J. (5123)
    • KEYWORDS: ASSAY DATA – REPRODUCTIVE AND DEVELOPMENTAL TOXICITY – QSAR MODELING – REGULATORY GUIDELINES

      This article describes two QSAR models (Random forest (RF) model and Adaptive Fuzzy Partition (AFP) model) for developmental toxicity of 292 compounds (see Support Material for Names, CAS number, FDA and CAESAR classifications) and as well the platform developed to make these models accessible.

    • Global QSAR models of skin sensitisers for regulatory purposes. Chaudry Q. et al. Chem Cent J. (5128)
    • KEYWORDS: QSAR MODELING – REGULATORY GUIDELINES – SKIN SENSITIZATION

      This article reports two global QSAR models for assessing skin sensitization. Both developed and tested under stringent quality criteria to fulfil the principles laid down by the OECD, using two different computational approaches (Adaptive Fuzzy Partition (AFP) model and Multilayer Perceptron (MLP), neural network model) applied to a set of 209 compounds previously published (‘Compilation of historical local node data for evaluation of skin sensitization alternative methods’), which contribute to the hybrid model freely available on their CAESAR website.

  • Hashing Algorithms and Data Structures for Rapid Searches of Fingerprint Vectors. Nasr R. et al. J Chem Inf Model. (5121)
  • KEYWORDS: DATABASES – STRUCTURE MASKING

    In order to speed up databases searches, this article proposes addition of a a short signature integer vector of length M to each molecule fingerprint (see Figure 1 for an illustration of modulo hashing). Authors have corroborated this new approach by large-scale simulations using molecules from ChemDB and perform their simulations using fingerprints based on labelled paths of lengths up to eight, but this approach can be applied with any kind of fingerprint system.

  • Drug Discovery in a Multidimensional World: Systems, Patterns, and Networks. Dudley JT. et al. J Cardiovasc Transl Res. (5119)
  • KEYWORDS: DRUG DISCOVERY – NETWORKS – SYSTEMS BIOLOGY

    Emphasis of recent advances in molecular profiling technologies. This article stands out how the development of sophisticated computational approaches for analyzing existing data currently provides systems-oriented approaches (connectivity maps, chemical systems biology, chemical similarity networks, and causal network inference) towards drug discovery.

  • Mechanistic Category Formation for the Prediction of Respiratory Sensitization. Enoch, SJ. et al. Chem. Res. Toxicol. (5118)
    Involved Partner: LJMU
  • KEYWORDS: ASSAY DATA – RESPIRATORY SENSITIZATION

    An example of respiratory sensitization prediction evaluating the acylation, Michael addition, Schiff base formation, SN2, SNAr, and no eletrophilic mechanism domains with a set of 39 respiratory sensitizers and 40 nonrespiratory sensitizers (see Support Material for Names, CAS number, SMILES, respiratory sensitizer effect and eletrophilic index). This study demonstrates that the potential of low molecular weight chemicals to cause respiratory sensitization can be understood in terms of covalent protein binding.

  • Genomic indicators in the blood predict drug-induced liver injury. Huang J. et al. Pharmacogenomics J. (5116)
  • KEYWORDS: BIOMARKERS – DILI – HEPATOTOXICITY – PATHWAYS

    Based on previously published set of gene expression data, this article stands a list of 10 identified genes (see Table 2) and a list of 30 biological processes in the blood (see Table 4) that are predictive of liver necrosis.

  • Coexistence of passive and carrier-mediated processes in drug transport. Sugano K. et al. Nat Rev Drug Discov. (5114)
    Involved Partners: ROCHE, PFIZER, UNIVIE, NOVARTIS
  • KEYWORDS: ADME – ASSAY DATA – P-gp – PHARMACOKINETICS – TRANSPORTERS

    This perspective article provides opinion on the coexistence of passive and carrier-mediated processes in drug transport, based on a notable literature review. Authors evidence their conclusions firstly with an overview of basics of membrane permeation in both passive and carried-mediated (see Box 1 for a schematic representation to illustrate the terminology of transport routes and direction of transport, and Box 2 for a list of general features of both transport types), secondly with a discussion of current methods to evaluate permeation mechanisms (PAMPA, Caco-2, MDCK, HeLa cells, HEK293 cells and Xenopus oocytes) and finally with a compilation of references describing relationship between biological membrane transport and passive transport indicators (See Table 1). Moreover, they present data from knockout mice for bioavailability and brain to plasma ratio (see Table 2) to evidence that carried-mediated transport has a more important role in blood-brain barrier permeation, renal and hepatic secretions than it does in the intestine. A comprehensive glossary is presented to help reader with specific concepts mentioned along the article (Active transport, Bioavailability, Biopharmaceutics classification system (BCS), Black lipid membrane, Caco-2 cells, Carrier-mediated transport, CYP3A4, Efflux ratio, Fa, First pass metabolism, Lipophilicity, log Doct, log Poct, Oral bioavailability, Papp, Passive transport, Peff, P-glycoprotein, pH partition theory, pKa, Tissue distribution coefficient, Total permeation, and Ussing chamber).

July

  • Similarity−Potency Trees: A Method to Search for SAR Information in Compound Data Sets and Derive SAR Rules. Wawer M. et al. J Chem Inf Model. (5112)
  • KEYWORDS: SOFTWARE – STRUCTURE-BASED PREDICTION

    Based on 6 data sets annotated to 6 different targets (structures from BindingDB (349 and 874) and PubChem (1243, 3348, 1765 and 1991)), this article introduces a new intuitive and generally applicable analysis method, termed similarity-potency tree (SPT), as a tool to mine structure-activity relationship (SAR) information in compound data sets of any source (see Figure 1 for the SPT data structure, Figure 2 for a prototype of SPT, Figure 3 for a SPT comparison and Figure 7 for a SPT analysis of HTS data sets).

  • Generative Models for Chemical Structures. White D. et al. J Chem Inf Model. (5111)
  • KEYWORDS: DRUG DISCOVERY – MOLECULAR DESCRIPTORS – PHARMACOLOGY – SOFTWARE

    This article provides details about a novel method of generating new chemical structures produced according to a statistical model of structural variation which is learned from the structures present in an input set of molecules. Authors overview those De novo methods for ligand design developed during the last 15 years (CHEMICS, COCOA, GenStar, GroupBuild, LigBuilder, GROW, BUILDER, DBMARKER, BOOMS-LANG, MOLMARKER, DREAM++, TOPAS, SYNOPSIS, and SPROUT). As stated, in drug discovery the properties required of a drug candidate generation are three-fold: 1) the shape of the intended active site determines the overall structure of the molecule, 2) it must be possible to synthesis the molecule and 3) the molecule must be drug-like; and their method (see Figure 1 for a diagrammatic description) focuses only on the first requirement, that is structural requirement, and assumes that populating the input set with active molecules to a certain target will provide generation of new molecules with also high affinity to this target.

  • Using an In Vitro Cytotoxicity Assay to Aid in Compound Selection for In Vivo Safety Studies. Greene N. et al. Bioorg Med Chem Lett. (5109)
    Involved Partner: PFIZER
  • KEYWORDS: CYTOTOXICITY – DRUG DISCOVERY

    A set of 290 Pfizer proprietary compounds, representing a diverse spectrum of chemical structures, intended target pharmacologies and disease indications, is selected to evaluate use of in vitro cytotoxicity assay data (in this case, compound ability to cause cytotoxicity in transformed human liver epithelial (THLE) cells) to help prioritize compounds for in vivo studies.

  • The challenges of in silico contributions to drug metabolism in lead optimization. Vaz RJ. et al. Expert Opin Drug Metab Toxicol. (5107)
    Involved Partner: LMD
  • KEYWORDS: CYP450 – METABOLISM – PHARMACOLOGY – SOFTWARE

    A comprehensive review of existing tools and methods (briefly: TIMES, META, MetaDrug, MetabolExpert, KnowItAll, METEOR, SPORCalc, MetaPrint2D, COMPACT, StarDrop, SMART CYP, Mlite and extensively: MetaSite) to predict the site of metabolism (SMO) as examples of software utility in overcoming metabolic stability in drug optimization by blocking the metabolic softspots and rationally modifying the inastable molecule to avoid interaction with the CYP450 pocket.

  • Importance of structural information in predicting human acute toxicity from in vitro cytotoxicity data. Lee S. et al. Toxicol Appl Pharmacol. (5103)
  • KEYWORDS: ASSAY DATA – CYTOTOXICITY – MOLECULAR DESCRIPTORS – PHARMACOKINETICS – STRUCTURE-BASED PREDICTION

    In vitro basal cytotoxicity data of 67 drugs (see Table 1 and Support Material data) selected from the literature by the ACuteTox project was considered to perform an analysis of the importance of structural information to predict human acute toxicity. The authors test several multiple linear regression models (see Table 2) based on a prior evaluation of series of molecular descriptors categorized in: constitutional, physicochemical, electrostatic, topological and geometrical. Results obtained provide an example of intends to close the gap between in vivo and in vitro toxicity, and suggest the use of structural descriptors related to pharmacokinetic parameters to correct the difference between cytotoxicity and human acute toxicity in terms of predictability.

  • In Silico Binary Classification QSAR Models Based on 4D-Fingerprints and MOE Descriptors for Prediction of hERG Blockage. Su BH. et al. J Chem Inf Model. (5101)
  • KEYWORDS: ASSAY DATA – CARDIOTOXICITY – DATABASES – hERG – MOLECULAR DESCRIPTORS – QSAR MODELING -STRUCTURE-BASED PREDICTION

    Based on 250 structurally diverse compounds screened for hERG activity defined as a training set, two different QSAR models are presented (a continuous partial least-squares (PLS) model and an optimized binary classification model) and evaluated by two external test sets (1315 compounds (from a condensed PubChem bioassay set after filtering (cleaning and constraints applied (Lipinski’s rule of five and relative lipophilicity (log P > 4.1 active, < 2.8 inactive)) and 106 compounds found in the literature, see Support Material for structures data information). Details of molecular descriptors selection to perform the continuous QSAR model are gathered in tables 5, 6 and 8 together with a comprehensive discussion. A comparison of accuracy of both training and testing set prediction (see Table 7) highlights the quality of the binary QSAR model derived in this study as one of the best modeling methodology and in this case study, it permits a structural interpretation to guide hERG blockage and detects an increasing trend when a polar negative group is present at a distance of 6-8 Å from a hydrogen bond donor.

  • Use of clearance concepts and modeling techniques in the prediction of metabolic drug-drug interactions. Ito K. et al. Trends Pharmacol Sci. (5106)
  • KEYWORDS: ADME – CYP450 – DRUG-DRUG INTERACTION – METABOLISM – PHARMACOKINETICS – TRANSPORTERS

    This opinion article reviews the methods for quantitative prediction of in vivo of drug-drug interactions (DDIs) caused by inhibition of metabolic enzymes based on clearance concepts and pharmacokinetic modeling. Emphasis on quantitative prediction of DDIs is presented by discussion about prediction of the area under the plasma concentration-time curve (AUC) increase caused by competitive enzyme inhibitors (basically CYP450 subfamily), physiologically based pharmacokinetic modeling for DDI prediction, and prediction of the interactions caused by mechanism-based inhibition of a certain enzyme.

  • Prediction of intrinsic solubility of generic drugs using MLR, ANN and SVM analyses. Louis B. et al. Eur J Med Chem. (5099)
  • KEYWORDS: ADME – ASSAY DATA – MOLECULAR DESCRIPTORS – QSAR MODELING

    Comparison of two different machine learning methods (artificial neural network (ANN) and support vector machine (SVM)) applied to model intrinsic solubility of a set of 74 generic drugs (see Table 1 for experimental data values) is presented in this article and demonstrates that the SVM model predicts the intrinsic solubility of drugs more accurately than ANN, though the test set prediction was better for ANN model. These results, together with the applicability of some theoretical calculated topological descriptors (distance or path based), identify some insights into what structural features are responsible for drug solubility and help in its prediction.

  • Getting physical in drug discovery: a contemporary perspective on solubility and hydrophobicity. Hill AP. et al. Drug Discov Today. (5102)
    Involved Partner: GSK
  • KEYWORDS: ADME – DRUG DISCOVERY – MOLECULAR DESCRIPTORS

    Since the optimization of physical properties is fundamental to success in drug discovery, this review studies solubility and hydrophobicity relationship based on one approximately 100 k compounds set with measured chemi-luminiscent nitrogen detection (CLND) solubility and other approximately 20 k compounds set with measured octanol-pH 7.4 buffer hydrophobicity values. A modification (log DpH instead of log P) of the General Solubility Equation (GSE, log S = – log P – 0.01 * (Mpt – 25) + 0.5), where Mpt refers to the melting point of a certain compound) is considered to reflect the contribution of charge to solubility and to evaluate the relationship between both properties under analysis. Additionally, authors suggest a logical extension to the finding that binned clog DpH7.4 shows enhanced resolution of solubility classes (low, medium and high) defining the Solubility Forecast Index (SFI = clog DpH + number of aromatic rings, where SFI < 5 means that there is a reasonable chance of having good solubility) and evidence somewhat solubility and aromaticity relationship. They conclude that the limits of GSE or any other solubility predictor must be governed by the quality of hydrophobicity estimates.

  • Predicting skin sensitization potential and inter-laboratory reproducibility of a human Cell Line Activation Test (h-CLAT) in the European Cosmetics Association (COLIPA) ring trials. Sakaguchi H. et al. Toxicol In Vitro. (5100)
  • KEYWORDS: ASSAY DATA – SKIN SENSITIZATION

    An example of efforts to validate non-animal test methods for skin sensitization after the prohibition of regulatory polices in Europe to test cosmetic ingredients in animals for a number of toxicological endpoints. The European Cosmetics Association (COLIPA) coordinated five independent laboratories (A: The Procter & Gamble Company, B: L’Oreal, Advanced Research,C: Henkel AG & Co., KgaA, D: Shiseido Safety Quality Assurance Center and E: Kao Safety Science Research Laboratories) to evaluate the inter-laboratory reproducibility of a human Cell Line Activation Test (h-CLAT) for a small set of chemicals (see Table 1 for experimental data and CAS number of compounds tested).

  • SMILES-based optimal descriptors: QSAR modeling of carcinogenicity by balance of correlations with ideal slopes. Toropov AA. et al. Eur J Med Chem. (5105)
  • KEYWORDS: ASSAY DATA – CARCINOGENICITY – QSAR MODELING – STRUCTURE-BASED PREDICTION

    A three version of the Monte Carlo optimization (the classic scheme (training set, test set), the balance of correlations (subtraining set, calibration set, test set), and the balance of correlations with ideal slopes (subtraining set, calibration set, test set)) is presented in this study with carcinogenicity data (log TD50) from a set of 401 compounds (see Support Material for experimental and calculated values). Results obtained (see tables 2-4) demonstrate that the correlation balance with ideal slopes gives best prediction for the test set for all three splits considered.

  • Discovery of metabolomics biomarkers for early detection of nephrotoxicity. Boudonck KJ. et al. Toxicol Pathol. (5104)
  • KEYWORDS: BIOMARKERS – METABOLISM – NEPHROTOXICITY

    Apart from the seven new biomarkers (KIM-1, Albumin, Total Protein, b2-Microglobulin, Cystatin C, Clusterin, and Trefoil Factor-3) to evaluate kidney damage accepted recently by FDA, EMEA and PSTC; this study confirms, in the framework of metabolomics increasing interest in the field of toxicology, other novel biomarkers associated with early nephrotoxicity (polyamines, 1,5-AG, monoethanolamine, phosphate, glycylproline, glucosamine, sorbitol, 5MeTHF, and others, see tables 3 and 4) and validates its analysis confirming as well some of the earlier published markers of nephrotoxicity (amino acids, glucose, and osmolytes) which highlights the fact that many of these metabolites are very early indicators of kidney malfunction.

  • Consortia and commodities. [No authors listed] Nat Biotechnol. (5098)
  • KEYWORDS: PROJECT – DRUG DISCOVERY

    This editorial makes reflection on purposes and advantages of recent collaborations between pharmaceutical companies, and government agencies, non-profit and academic institutions, taking the form of public-private consortia. Innovative Medical Initiatives (IMIs) are highligthed as the biggest European Union partnerships currently in progress.

  • The computational prediction of genotoxicity. Naven RT. et al. Expert Opin Drug Metab Toxicol. (5086)
    Involved Partner: PFIZER
  • KEYWORDS: CARCINOGENICITY – GENOTOXICITY – MOLECULAR DESCRIPTORS – MUTAGENICITY – QSAR MODELING – STRUCTURE-BASED PREDICTION

    This review updates readers about recent scientific developments regarding prediction of Ames mutagenicity and in vitro chromosome damage, discusses about several strategies for general genotoxicity prediction in drug development, and provides an evaluation of computational approaches (online applications: LAZAR and CAESAR; commercial software: Derek for Windows, ACD/Tox, MCASE/MC4PC and SciQSAR) limitations distinguishing between global or local QSAR models generation, based on entire data sets of diverse compounds or congeneric class of compounds, respectively. Authors stand out the need to identify toxicological knowledge gaps and to carry out focused testing in order to understand the mechanistic SARs behind toxicity data analysis.

June

  • Trust, But Verify: On the Importance of Chemical Structure Curation in Cheminformatics and QSAR Modeling Research. Fourches D. et al. J Chem Inf Model. (5110)
  • KEYWORDS: DATABASES – MOLECULAR DESCRIPTORS – QSAR MODELING – SOFTWARE

    This article emphasizes the need for a standardized chemical data curation strategy in any molecular modeling investigation. Authors suggest a general data set curation workflow (removal of inorganics, organometallics, counterions, and mixtures; structural conversion and cleaning, ring aromatization, normalization of specific chemotypes, curation of tautomeric forms; deletion of duplicates; and manual checking of the structures and biological activities, see Figure 1), a summary of major procedures and corresponding relevant software for every step of data curation process (see Table 1) and 6 examples relying on data set curation: DILI, nitroaromatics (2 species), ToxRefDB, Ames mutagenicity and bioavailability sets, with 1061, 28, 95, 320, 7090, 805) compounds in the original set and 951, 28, 95, 292, 6542, 734.after curation, respectively.

  • Estimation of ADME Properties with Substructure Pattern Recognition. Shen J. et al. J Chem Inf Model. (5097)
  • KEYWORDS: ADME – ASSAY DATA – MOLECULAR DESCRIPTORS – STRUCTURE-BASED PREDICTION

    Application of a support vector machine (SVM) algorithm to build classification models for two different data sets (578 compounds with human intestinal absorption (HIA) data, and 1593 compounds with blood-brain barrier penetration (BBB) data, see Support Material) by description of molecules using substructure pattern fingerprints generated based on a predefined pattern dictionary.

  • Ion channels in toxicology. Restrepo-Angulo I. et al. J Appl Toxicol. (5096)
  • KEYWORDS: CARDIOTOXICITY – CARCINOGENICITY – TRANSPORTERS

    A recent review on the important role of ion channels in toxicology, since they are involved in cardiac contraction, neural transmission, temperature sensing, insulin release, regulation of apoptosis, cellular pH, oxidative stress, etc., and their non-specific binding to several drugs is responsible for undesired side-effects. Apart from a brief overview of sodium, chloride, calcium and potassium channels types and features, the authors focus their discussion on toxins targeting ion channels (see Table 1 for a list of toxins, targeted channel and known effect).

  • Gaussian Processes for Classification: QSAR Modeling of ADMET and Target Activity. Obrezanova O. et al. J Chem Inf Model. (5081)
  • KEYWORDS: ADME – ASSAY DATA – QSAR MODELING

    Based on seven different data sets (BBB 1591, hERG 168 (see Suport Material), HIA 225, COX-2 303, BZR 306, DHFR 393, and ER 446 compounds), the authors perform a comparison between the performance of Gaussian processes (GP) for classification and other known computational methods (decision trees, random forest, support vector machines, and probit partial least squares). This study highlights the GP classifier as the one that often produces more predictive models.

  • Pharmaceutical Perspectives of Nonlinear QSAR Strategies. Michielan L. et al. J Chem Inf Model. (5077)
  • KEYWORDS: ADME – MOLECULAR DESCRIPTORS – QSAR MODELING – STRUCTURE-BASED PREDICTION

    This perspective article discusses the state-of-the-art in the field of nonlinear QSAR strategies focused in predictive toxicology about modeling methods (molecular descriptors, machine learning methodologies), and their application in pharmacodynamics (regression and classification strategies). The authors present a workflow from the data collection to the in silico prediction of the properties for new compounds, based on the use of pharmacodynamic (Table 1), pharmacokinetic (Table 2) and toxicity databases models (Table 3).

  • Drug profiling: knowing where it hits. Merino A. et al. Drug Discov Today. (5094)
  • KEYWORDS: ADME – DRUG DISCOVERY – PHARMACOLOGY

    Discussion of several emerging computational approaches, in the field of drug discovery, concerning the in vitro and in silico drug promiscuity analysis as complementary techniques to prevent attrition and maximize the utility of new compounds, or even in the case of old drugs.

  • Developing Structure-Activity Relationships for the Prediction of Hepatotoxicity. Greene N. et al. Chem Res Toxicol. (5091)
    Involved Partners: PFIZER, LL
  • KEYWORDS: DATABASES – DILI – HEPATOTOXICITY – STRUCTURE-BASED PREDICTION

    Based on the reference book ‘The Adverse Effects of Drugs and Other Chemical on the Liver‘ and some reviews of hepatotoxicity data, the authors build a structure-searchable database with 1116 chemicals extended to 1266 (see Supplementary Material for names and CAS numbers) after first stage of structural alerts creation using Derek for Windows (DdW). Their approach was evaluated with a Pfizer-developed set of 626 compounds (see Supplementary Material for names, CAS numbers and classification associated to hepatotoxicity [NE (no evidence for hepatotoxicity in any species, 152 total), WE (weak evidence for human hepatotoxicity, 62 total), AH (animal hepatotoxicity observed, not tested in humans, 139 total), and HH (evidence for hepatotoxicity in humans, 273 total)] which has significant overlap and a similar classification schema to a set of 344 compounds described in the literature consulted.

  • Extending pKa prediction accuracy: High-throughput pKa measurements to understand pKa modulation of new chemical series. Miletti F. et al. Eur J Med Chem. (5095)
  • KEYWORDS: PHARMACOKINETICS – SOFTWARE

    Application of a recent developed piece of software, MoKa (benchmarked against a set of 5581 pKa values taken from the Roche in-house library), to extend pKa prediction accuracy by presenting new results on important pKa determinants (inductive, steric, and hydrogen bonding effects) focused on series particularly important for the medicinal chemist (amines, amides, and sulfonamides).

  • Integration of in silico and in vitro platforms for pharmacokinetic-pharmacodynamic modeling. Sung JH. et al. Expert Opin Drug Metab Toxicol. (5092)
  • KEYWORDS: ADME – DRUG DISCOVERY – PHARMACOKINETICS

    This article, firstly, updates the reader about the concept of physiologically-based pharmacokinetic (PBPK) and pharmacodynamic (PD) modeling (see Figure 1 and Figure 2C) together with its applications and limitations, and secondly, provides a comprehensive summary of the research efforts towards ‘artificial organs’ (Liver (biotransformation), Kidney (excretion), Lung (gas exchange), Gastrointestinal tract (absorption) and Vascular network (gas and nutrient transport)) to evidence how the combination of microscale technology and PK-PD modeling should contribute to the development of a novel in vitro/in silico platform for more physiologically-realistic drug screening.

  • Data structures and computational tools for the extraction of SAR information from large compound sets. Wawer M. et al. Drug Discov Today. (5090)
  • KEYWORDS: DATA MINING – DRUG DISCOVERY – MOLECULAR DESCRIPTORS – STRUCTURE-BASED PREDICTION

    An extended discussion of computational data mining and visualization techniques currently relevant in the structure-activity relationship (SAR) information field (see Table 1 for a list of SAR information extraction methods representing the four principal approaches to SAR-relevant compound data mining.).

  • The Drug Discovery Portal: a resource to enhance drug discovery from academia. Clark R. et al. Drug Discov Today. (5089)
  • KEYWORDS: DATABASES – DRUG DISCOVERY

    Review on several coordinated and multi-institutional drug discovery operations between academia, and national and international industry to present the recent initiative of the Drug Discovery Portal.

  • High-Throughput Screening System for Identifying Phototoxic Potential of Drug Candidates Based on Derivatives of Reactive Oxygen Metabolites. Onoue S. et al. Pharm Res. (5087)
  • KEYWORDS: ASSAY DATA – PHOTOTOXICITY

    A comparative study of a new in-house high-throughput screening strategy, a derivatives-of-reactive-oxygenmetabolites (D-ROM) assay, together with a reactive oxygen species (ROS) assay, a DNA-photocleavage assay, and a 3T3 neutral red uptake phototoxicity test (3T3 NRU PT) applied on a set of 25 model compounds (20 phototoxic drugs and 5 non-phototoxic chemicals) states unclear correlation between D-ROM assay and ROS assay to predict photochemistry and phototoxicity of this set, meanwhile D-ROM assay partly indicates photogenotoxic risk identified by the DNA photocleavage test, and phototoxic potencial of these drugs as 3T3 NRU PT evidences (see Table II for photochemical and photobiological data of compounds). This study results suggest D-ROM assay as an alternative to identify phototoxic potential and avoid undesired side effects in the early stages of drug discovery.

  • Screening for phospholipidosis induced by central nervous drugs: Comparing the predictivity of an in vitro assay to high throughput in silico assays. Mesens N. et al. Toxicol In Vitro. (5083)
    Involved Partner: J&J PRD
  • KEYWORDS: ASSAY DATA – PHOSPHOLIPIDOSIS – RISK ASSESSMENT

    Test of 33 CNS-compounds (24 in vivo negative and 9 in vivo positive phospholipidosis-inducers, see Table 1 for their generic name, IUPAC name/chemical name, CAS number and Pharmacological action/indication information, and Table 2 and 3 for their in vitro and in vivo data) in an in-house developed in vivo phospholipidosis screening assay in comparison to well stablished in silico prediction models (Ploemen model, modified Ploemen model and Tomizawa model). As Table 5 statistics show, the in silico assays made a poor distinction between the negative and positive phospholipidosis-inducers based on physicochemical properties and evidence preference for an in vivo low throughput assay to predict phospholipidosis induced by central nervous drugs.

  • Druggable pockets and binding site centric chemical space: a paradigm shift in drug discovery. Pérot S. et al. Drug Discov Today. (5079)
  • KEYWORDS: DRUG DISCOVERY – PHARMACOLOGY – SOFTWARE

    This review presents a thorough compilation of the main algorithms to search for binding pockets (see Table1, classified in Geometric and genomic (2), Geometric (20), Geometic and energy-based (1), Energy-based (3), Probe-mapping/energy-based (3) and Docking (1)) and a list of 19 methods for evaluating binding site similarities and related databases (see Table2), and discuss how this methods assist binding site identification, prediction of druggability and binding site comparison.

  • Prediction of phospholipidosis-inducing potential of drugs by in vitro biochemical and physicochemical assays followed by multivariate analysis. Kuroda Y. et al. Toxicol In Vitro. (5084)
  • KEYWORDS: ASSAY DATA – METABOLISM – PHOSPHOLIPIDOSIS

    A slight multivariate analysis on 8 cationic amphiphilic drugs (CADs) to predict their phospholipidosis-inducing potential by application of principal component analysis using from in vitro assays several biochemical parameters expressing binding to PL, inhibition of enzymatic degradation of PL, metabolic stability of CADs, as well as physicochemical parameters, pKa, and clog P.

  • An in Silico Method for Predicting Ames Activities of Primary Aromatic Amines by Calculating the Stabilities of Nitrenium Ions. Bentzien J. et al. J Chem Inf Model. (5076)
    Involved Partner: BI
  • KEYWORDS: ASSAY DATA – MUTAGENICITY – QSAR MODELING – STRUCTURE-BASED PREDICTION

    Based on a set of 257 primary aromatic amines compounds from Organon Ames Data (Kazius et al. (2005)), this study shows how use of quantum mechanical calculations (at different levels of theory, semiempirical and ab initio) can correctly differenciate between Ames active and inactive compounds. Authors perform predictions applying 3 commercial software packages (DEREK, MULTICASE, and the MOE-Toxicophore descriptor) and evidence comparable results between their method and MOE-Toxicophore model.

  • Cheminformatics Analysis of Assertions Mined from Literature That Describe Drug-Induced Liver Injury in Different Species. Fourches D. et al. Chem Res Toxicol. (5075)
  • KEYWORDS: ASSAY DATA – DILI – HEPATOTOXICITY – QSAR MODELING – STRUCTURE-BASED PREDICTION – TEXT MINING

    An extensive example of combining text mining and cheminformatics data analysis with Drug-induced Liver Injury (DILI) data is presented in this paper, based on a data set of 951 compounds reported to produce a wide range of effects in the liver in different species (human, rodents, non-rodents) in 650000 MEDLINE records compiled, using BioWisdom’s Sofia platform. The good predictive powder of the QSAR models generated with this data opens up new opportunities for generating and modeling chemical toxicology data.

May

  • Computational science in drug metabolism and toxicology. Valerio Jr LG. Expert Opin Drug Metab Toxicol. (5070)
  • KEYWORDS: CARCINOGENICITY – CYP450 – DATABASES – GENOTOXICITY – hERG – METABOLISM – MOLECULAR DESCRIPTORS – MUTAGENICITY – PHARMACOKINETICS – QSAR MODELING – STRUCTURE-BASED PREDICTION

    This editorial introduces to the special issue (early online) of Expert Opinion on Drug Metabolism and Toxicology focused on Computational Science in Drug Metabolism and Toxicology. A series of articles that discuss about regulatory perspectives, predictive performance by QSAR models, molecular descriptors, large databases, genotoxicity, mutagenicity, and carcinogenicity data, machine learning algorithms for screening the hERG protein and CYP isoforms, metabolism, pharmacokinetics properties, etc.

  • State-of-the-art genomics approaches in toxicology. Van Hummelen P. et al. Mutat Res. (5068)
  • KEYWORDS: TOXICOGENOMICS

    This review highlights needs of interdisciplinary working groups to bring biologists, bioinformaticians, statisticians and toxicologists together to change the current paradigm of toxicology by means of genomics analysis that can provide supportive evidence to understand mechanisms of toxicity and can offer alternative to reduce the number of in vivo experiments, animal usage and animal suffering. See Figure 1 for a representative example of an end-to-end approach for a toxicogenomics investigation, and tables 1, 2 and 3 with lists of microarray studies mentioned in this review, toxicogenomics studies published in ArrayExpress, and toxicogenomics data sets published in GEO, respectively.

  • Next-generation biomarkers for detecting kidney toxicity. Bonventre JV. et al. Nat Biotechnol. (5067)
    Involved Partner: NOVARTIS
  • KEYWORDS: BIOMARKERS – NEPHROTOXICITY

    One of the ten articles from The Predictive Safety Testing Consortium (PSTC) published into a special issue in Nature Biothecnology. This consortium faced several criteria as key characteristics of a renal safety biomarker. These criteria were as follows: Identifies kidney injury early (well before the renal reserve is dissipated and levels of serum creatinine increase); Reflects the degree of toxicity, in order to characterize dose dependencies; Displays similar reliability across multiple species, including humans; Localizes site of kidney injury; Tracks progression of injury and recovery from damage; Is well characterized with respect to limitations of its capacities; and Is accessible in readily available body fluids or tissues. See Table 1 for a list of 19 urinary biomarkers of kidney toxicity.

  • Characterizing and predicting carcinogenicity and mode of action using conventional and toxicogenomics methods. Waters MD. et al. Mutat Res. (5071)
  • KEYWORDS: CARCINOGENICITY – GENOTOXICITY – TOXICOGENOMICS

    This review of nine toxicogenomics investigations over the past 6 years, including study design and performance characteristics, highlights the potential of molecular expression analysis to more properly classification of genotoxic and nongenotoxic carcinogens and to perform prediction of carcinogenicity of untested chemicals.

  • In Vitro Screening of Environmental Chemicals for Targeted Testing Prioritization: The ToxCast Project. Judson RS. et al. Environ Health Perspect. (5066)
  • KEYWORDS: ASSAY DATA – PATHWAYS

    The ToxCast Project outlines in this article their evaluation of the use of high-throughput in vivo assays to prioritize chemicals for more detailed testing and which tests should be run in preference. They provide data (see supplementary material) for a set of 309 mostly pesticide active chemicals in 467 assays across nine technologies (Cell-free HTS, Cell-based HTS, High-content cell imaging, Quantitative Nuclease protection, Multiplex transcription reporter, Biologically multiplexed activity profiling (BioMAP), Phase I and II XME cytotoxicity, HTS genotoxicity, and Real-time cell electronic sensing).

  • Gene set-level network analysis using a toxicogenomics database. Kiyosawa N. et al. Genomics. (5065)
  • KEYWORDS: BIOMARKERS – DATABASES – DATA MINING – NETWORKS – SYSTEMS BIOLOGY – TOXICOGENOMICS

    Based on a set of 118 compounds (see Table 1 for chemical information (Name, Abbreviation and Dose), the authors establish a gene set-level (see Table 2 for a list of genes) and phenotype-level (see Table 3 for a summary of the phenotype information) relational network using TG-GATE, a large-scale TGx database. Their results evidence that accumulation of such robust gene sets with toxicity-associated subnetwork structures would lead to a better understanding of the molecular mechanisms of drug-elicited toxicities.

  • In silico approaches to predicting cancer potency for risk assessment of genotoxic impurities in drug substances. Bercu JP. et al. Regul Toxicol Pharmacol. (5064)
  • KEYWORDS: ASSAY DATA – CARCINOGENICITY – GENOTOXICITY – RISK ASSESSMENT – SOFTWARE – STRUCTURE-BASED PREDICTION

    Discussion on the application of two software tools (MultiCASE and VISDOM (in-house software Eli Lilly and Co.)) to derive cancer potency predictions of 694 compounds of a total of 1515 compounds extracted from the carcinogenicity potency database CPDB. This study proposes a step-wise approach to calculate predicted numerical TD50 values for compounds categorized as not potent which can be used to establish safe levels greater than the threshold toxicological concern (TTC) for genotoxic impurities (GTIs) in a drug substance.

  • Toxicogenomic profiling of chemically exposed humans in risk assessment. McHale CM. et al. Mutat Res. (5063)
  • KEYWORDS: BIOMARKERS – DATABASES – RISK ASSESSMENT – SYSTEMS BIOLOGY – TOXICOGENOMICS

    This article reviews current approaches in the prediction of human toxicological outcomes and overviews several human toxicogenomics studies (see Table 1 for a list of Toxicogenomics studies of exposed populations) showing the application of the OMIC technologies to human health risk assessment (see Figure 1 where the interrelationship between Adductomics, Transcriptomics, Proteomics, and Epigenomics is shown that facilitates the examination of gene-environment interactions).

  • Prediction of carcinogenicity for diverse chemicals based on substructure grouping and SVM modeling. Tanabe K. et al. Mol Divers. (5062)
  • KEYWORDS: CARCINOGENICITY – MOLECULAR DESCRIPTORS – QSAR MODELING – STRUCTURE-BASED PREDICTION

    Based on a collection of experimental carcinogenicity data on about 1500 chemicals, 911 organic chemicals are selected to perform QSAR modelling by calculation of 1504 kinds of molecular descriptors with Dragon software and analyzed with a support vector machine (SVM) method. See Table 1 for a list of types, numbers and examples of Dragon descriptors used, Table 2 for a list of atom and functional group count descriptors, numbers of chemicals containing those atoms or functional groups, and statistical significance for positive/negative ratio and Table 3 for the resulting 20 models according to contained substructures obtained with this study.

  • Integrating background knowledge from internet databases into predictive toxicology models. Edelstein M. et al. SAR QSAR Environ Res. (5061)
  • KEYWORDS: DATABASES – DATA MINING – STRUCTURE-BASED PREDICTION

    One of the last efforts in integration of toxicology data from public databases to generate predictive toxicology models. Background knowledge information for three data sets gathered from DSSTox network (EPA Fathead Minnow Acute Toxicity (617 compounds), FDA Maximum (Recommended) Daily Dose (1215 compounds), and National Center for Toxicological Research Estrogen Receptor Binding databases (228 compounds)) was quested to three public chemical databases (PubChem, ChemBank and Sigma-Aldrich) in order to test integration approaches to predictive SAR models. Results obtained clearly evidence that the integration of background knowledge from public databases can significantly improve prediction performance.

  • Structure-based Drug Metabolism Predictions for Drug Design. Sun H. et al. Chem Biol Drug Des. (5060)
    Involved Partner: PFIZER
  • KEYWORDS: ADME – CYP450 – DRUG DISCOVERY – METABOLISM – STRUCTURE-BASED PREDICTION

    An extended review about the important progress in drug metabolism and currently in silico techniques adopted to predict drug regioselectivity, stereoselectivity, reactive metabolites, induction, inhibition and mechanism-based inactivation, and their implementation in hit-to-lead drug discovery. Discussion from Lipinski ‘Rule-of-Five’, ADMET prediction to Docking, Molecular dynamics and quantum chemical calculation, with special attention devoted to the case of CYP450 (see Table 1 for a complete list of available structures of human CYP450s determined by X-ray crystallography). CYP450s are major enzymes to catalyze the metabolic activation of xenobiotics ⁄ drugs, sterols, fatty acids, eicosanoids and vitamins. (Expressed in human: liver CYP1A2, 2A6, 2B6, 2C8, 2C9, 2C18, 2C19, 2D6, 2E1, 3A4 and 3A5, lungs CYP1A1, 1B1, 2A6, 2B6, 2E1, 2F1, 2J2, 2S1, 3A5 and 4B1 or in minor sites like kidney, brain, small intestine, peripheral blood cells, platelets, neutrophils, and seminal vesicles. CYP1A2, 2C9, 2C19, 2D6, 2E1, and 3A4 ⁄ 5 are major ones generally used for in vitro drug metabolism profiling studies in current discovery settings).

April

  • Nuclear receptors as drug targets in cholestasis and drug-induced hepatotoxicity. Holzer P. Pharmacol Ther. (5050)
  • KEYWORDS: HEPATOTOXICITY – METABOLISM – NUCLEAR RECEPTORS

    Part of hepatic effects of drugs, metabolites and herbal or synthetic compounds can be explained by their binding to specific nuclear receptors (NR) in the liver, this article provides an updated overview for a set of this protein family members (FXR, SHP, PXR, CAR, VDR, HNF4alpha, LRH-1, PPAR alpha and gamma, LXRalpha and GR) involved as therapeutic targets in cholestasis since they regualte a large number of target genes mediating transport, synthesis and detoxification of biliary constituents including bile acids. They emphasize the fact that induction of drug-clearance pathways by NR ligands can have important clinical consequences for drug interactions, increased clearance of other drugs can result in a decreased therapeutic effect, while enhanced bio-activation of drugs may contribute to hepatotoxicity via increased formation of reactive intermediate metabolites.

  • A structural feature-based computational approach for toxicology predictions. Valerio LG. et al. Expert Opin Drug Metab Toxicol. (5035)
  • KEYWORDS: RISK ASSESSMENT – SOFTWARE – STRUCTURE-BASED PREDICTION

    Evaluation of a computational toxicology software, the Leadscope Model Applier (LMA), as a decision support tool in safety and risk assessment to predict preclinical and clinical endpoints, drug metabolism, pharmacokinetics and mechanisms responsible for toxicity. The reader can find a full discussion about the underlying principles of the QSAR methods applied in this technology and a wide description regarding: i) training data sets; ii) molecular descriptors (predictors and structural features); and iii) model building algorithms.

  • Hepatotoxicity and hepatic metabolism of available drugs: current problems and possible solutions in preclinical stages. Giri S. et al. Expert Opin Drug Metab Toxicol. (5052)
  • KEYWORDS: DRUG SAFETY – HEPATOTOXICITY – METABOLISM – MITOCHONDRIAL TOXICITY

    Based on a set of 900 drugs, this review highligths the unexpected evidence that more than 64 allopathic drugs can induce potentially life-threatening hepatotoxicity with diverse clinical features, and lists the associated pattern hepatotoxicity for 36 of these drugs (see Table 1). Since chemically reactive metabolites have been proposed to be responsible for most types of drug-induced injury, their discussion focuses on drug metabolites formed during biotransformation in a widely accepted organotypical in vitro cellular model with relevance to in vivo two-compartment approaches. They stand out that drug-induced hepatoxicity depends on many factors, such as the host’s metabolic condition, age, sex, nutritional factors, multiple medications, physiological changes (pregnancy and renal impairment), genotype, duration and dosage of drug-enzyme induction, among others; and explain the different patterns of drug-induced hepatoxicity by these six basic mechanisms: i) disruption of calcium homeostasis leading to cell surface blebbing and lysis; ii) canalicular injury; iii) metabolic bioactivation of chemicals via CYP to reactive species; iv) stimulation of autoimmunity; v) stimulation of apoptosis and vi) mitochondrial injury.

  • Troubleshooting computational methods in drug discovery. Kortagere S. et al. J Pharmacol Toxicol Methods. (5037)
  • KEYWORDS: ADME – DATA MINING – DRUG DISCOVERY – QSAR MODELLING

    Description of various in silico approaches, ligand-based and structured-based methods (virtual screening, QSAR, pharmacophore based modeling and screening, homology modeling, docking, molecular dynamics simulation protocols, data mining of genomes or small molecule databases, network or pathway analysis, machine learning techniques,…). The authors make stress on the limitations of such technologies and recommend to alert the user’s attention during their workflows similarly as they exemplify in three flowcharts (homology modeling protocol, hybrid structure-based in silico screening method and ADMET model generation).

  • Prospective Validation of a Comprehensive In silico hERG Model and its Applications to Commercial Compound and Drug Databases. Doddareddy MR. et al. ChemMedChem. (5056)
  • KEYWORDS: ASSAY DATA – DATABASES – hERG – STRUCTURE-BASED PREDICTION – QSAR MODELING

    This article offers the largest public dataset described to date to generate ligand-based in silico hERG models for 2644 compounds (including literature compounds and FDA-approved drugs, see Supporting information) using Linear Discriminant Analysis (LDA) and Support Vector Machines (SVM). Biological evaluation of 50 predicted hERG blockers and 10 predicted non-hERG blockers confirms high models prediction quality, even better in predicting non-blockers since they were indeed found to be inactive in the assay. Furthermore, this article reports the results of a virtual screening of commercial available databases with selected models. At least two of the models are in agreement and predict 23% of Chembridge, 38% of Chemdiv, 31% of Asinex and 18% of Maybridge as potential hERG blockers. Remarkably, their set of models fails to predict most of the hERG blockers deposited in the PubChem database, and it can be understandable because of different assay formats used and claims for a future comparison between methods applied.

  • Membrane transporters in drug development. Giacomini KM. et al. Nat Rev Drug Discov. (5022)
  • KEYWORDS: PROJECT – DRUG-DRUG INTERACTION – DRUG SAFETY – PHARMACOKINETICS – TRANSPORTERS

    More than 400 membrane transporters (basically belong to ‘ATP-binding cassette’ (ABC) and ‘solute carrier’ (SLC) superfamilies) are under interest in drug development, with particular attention to those expressed in epithelia of the intestine, liver and kidney, and in the endothelium of the blood–brain barrier. Studies suggested their role in in vivo drug disposition, therapeutic efficacy and adverse drug reactions. This review expresses the consensus on the current knowledgment of transporters in drug development regarding the conduct of transporter assays and data interpretation summed up by the International Transporter Consortium after the workshop they held in 2008. Three main discussions are gathered in this review: 1) an overview of transporters (see tables 1 and 2 with ABC and SLC transporters of emerging clinical importance in the absoption and disposition of drugs, respectively), and Table 3 with a selection of transporter-mediated clinical drug-drug interactions); 2) a list of methods for studying transporters (Membrane-based systems. Cell-based assay systems, Intact organ/in vivo models, Contribution of transporters in vivo, interplay of efflux transporters and enzymes, Transporters and enzymes in drug clearance, and Computational models; and 3) recommendations and examples of guidelines that should be considered in drug development.

  • Gender differences in drug toxicity. Nicolson TJ. et al. Trends Pharmacol Sci. (5033)
  • KEYWORDS: CYP450 – DRUG DISCOVERY – LQT – REGULATORY GUIDELINES – RISK ASSESSMENT

    Since the 1998 FDA regulation demands equal male and female clinical trials, this article reviews key gender differences to take into account in the drug discovery process: differences in cardiac channels densities have relation to direct hormonal effects (repolarisation of cardiac ion channels, expression levels of ion channels and the densities of the ion channels); higher rates of CYP3A substrate metabolism in women; different body composition, women tend to have higher percentatge of body fat, plasma volume and organ blood flow, in addition to an altered drug/plasma protein binding profile; slower gastrointestinal transit in women; higher sensitivitiy to adverse drug reactions associated to opiods receptor agonists, specially Kappa opioids are more effective in women. All pharmacodynamics factors are highlighted due to their relation to cardiotoxicity and drug-induced long QT syndrome.

  • Behavioral toxicology in the 21st century: Challenges and opportunities for 2 behavioral scientists. Bushnell PJ. et al. Neurotoxicol Teratol. (5030)
  • KEYWORDS: NUCLEAR RECEPTORS – RISK ASSESSMENT

    Summary of a symposium presented at the annual meeting of the Neurobehavioral Teratology Society in 2009, where discussion about the potential challenges (predicting behavior using models of complex neurobiological pathways, standardizing study designs and dependent variables to facilitate creation of databases, and managing the cost and efficiency of behavioral assessments) and opportunities (identifying and characterizing toxicity pathways, informing the conditions and limits of extrapolation, addressing issues of susceptibility and variability, providing reality-checks on selected positives and negatives from screens, and performing targeted testing and dose-response assessments of chemicals flagged during screening) for behavioral scientists in developing took place.

  • A framework for using structural, reactivity, metabolic and physicochemical similarity to evaluate the suitability of analogs for SAR-based toxicological assessments. Wu S. et al. Regul Toxicol Pharmacol. (5023)
  • KEYWORDS: ADME – METABOLISM – RISK ASSESSMENT – STRUCTURE-BASED PREDICTION

    An example of a systematic expert-driven process based upon chemical and biochemical principles with an emphasis on bioactivation mechanisms to identify and evaluate analogs for read across in Structure activity relationship toxicological assessments. With some illustrative case studies extracted from literature (suitable analogs, analogs suitable with interpretation, analogs suitable with pre-condition and unsuitable analogs), they exemplify the structure, reactivity, physicochemical properties and metabolism evaluation to ranking analogs and discuss of uncertainty related to using analogs for risk assessment.

  • An introduction to QT interval prolongation and non-clinical approaches to assessing and reducing risk. Pollard CE. et al. Br J Pharmacol. (5041)
  • KEYWORDS: DRUG SAFETY – hERG – LQT – RISK ASSESSMENT

    Both, genetic and pharmacological evidence highlighting the pivotal role the human ether-a-go-go-related gene (hERG) channel is a critical step in understanding how to prevent drug-induced QT. This review describes, based on a generic drug discovery, the development process in relation to a non-clinical QT strategy by adopting a hERG-focused strategy to develop effective but safe medicines reducing safety-related attrition and shorting timescales required in early drug discovery.

  • QSAR with quantum topological molecular similarity indices: toxicity of aromatic aldehydes to Tetrahymena pyriformis. Kar S. et al. SAR QSAR Environ Res. (5046)
  • KEYWORDS: DRUG SAFETY – hERG – LQT – RISK ASSESSMENT

    Both, genetic and pharmacological evidence highlighting the pivotal role the human ether-a-go-go-related gene (hERG) channel is a critical step in understanding how to prevent drug-induced QT. This review describes, based on a generic drug discovery, the development process in relation to a non-clinical QT strategy by adopting a hERG-focused strategy to develop effective but safe medicines reducing safety-related attrition and shorting timescales required in early drug discovery.

  • Estimation of reliability of predictions and model applicability domain evaluation in the analysis of acute toxicity (LD50). Sazonovas A. et al. SAR QSAR Environ Res. (5047)
  • KEYWORDS: QSAR MODELLING

    Report of results obtained during the development of predictive models for six considered acute toxicity systems (acute rat toxicity after oral andintraperitoneal administration; acute mouse toxicity after oral, intraperitoneal, intravenous, and subcutaneous administration) of nearly 75000 compounds cleaned automatically (by removing any non-covalent complexes, salts, compounds with incorrect structures and unusually high deviations in interspecies correlations (animal vs. animal, and administration vs. Administration)) from RTECS database using GALAS modelling methodology(*) demonstrates the superiority of this approach compared with ordinary PLS, as its application resulted in the improvement of the statistical model performance.

    (*)Note: GALAS model can be viewed as a combination of two systems: 1) A structure-based QSAR model for the prediction of the property of interest (LD50 in this particular case) – baseline model; and 2) A similarity-based routine which identifies the most similar compounds contained in the training set and, considering their experimental values, calculates systematic deviations produced by the baseline QSAR model for any submitted test compound.

  • The International Transporter Consortium: A Collaborative Group of Scientists From Academia, Industry, and the FDA. Huang SM. et al. Clin Pharmacol Ther. (5051)
  • KEYWORDS: PROJECT – DRUG SAFETY – TRANSPORTERS

    The International Transporter Consortium (ITC) was formed in 2007 as a group of individuals working together interested in the role of transporters in drug safety and efficacy. This perspective article summarizes the key findings emerged from a workshop that expanded discussion about key transporters that have a role in drug absorption and disposition, and provided examples of currently used and potential technologies in studies of drug-transporter interactions and some guidances for industrial scientists in determining the role of transporters in the pharmacokinetics and pharmacodynamics of new medical products.

  • New cytochrome P450 mechanisms: implications for understanding molecular basis for drug toxicity at the level of the cytochrome. Shakunthala N. Expert Opin Drug Metab Toxicol. (5024)
  • KEYWORDS: CYP450

    Since CYPs, a large group of hemoproteins enzymes, catalyze mainly hydroxylation reactions in the presence of specific electron transport system and molecular oxygen, this review brings together published information indicating that i) CYPs cannot be treated as single-site enzymes, and that the active site has two different interacting subsites geared for entirely different types of functionally relevant interactions; ii) how substrates as well as products interact with the two sites is important for determining the functional state of the enzyme during catalysis; iii) these interactions may have necessary features to contribute in a major way to coupling/uncoupling mechanisms; and iv) product may play significant role in these mechanisms.

  • Computational toxicology―a tool for early safety evaluation. Merlot C. Drug Discov Today. (5007)
  • KEYWORDS: DRUG SAFETY – PHARMACOLOGY

    Focus on recent developments in computational toxicology is addressed to remark the current trend of making simpler predictions, closer to the mechanism of action, and follow them up with in vitro or in vivo assays as appropriate. New computational approaches should move from predicting global toxicity endpoints (carcinogenesis, mutagenesis,…) to the complexity of the predicted endpoints, rather than to the poor performance of data analysis methods. The focus, therefore, is on modeling more simple endpoints, such as off-target activity, to increase accuracy and to combine the results with experiments from other fields (e.g. -omics) to try to make a link with potential modes of action.

  • Using In Silico Tools in a Weight of Evidence Approach to Aid Toxicological Assessment. Ellison CM. et al. Mol Inf. (5013)
  • KEYWORDS: REGULATORY GUIDELINES – SOFTWARE – STRUCTURE-BASED PREDICTION

    Discussion about the use of the OECD (Q)SAR Application Toolbox, Derek for Windows, the CAESAR global model and SMARTS rules for reactivity within a Weight of Evidence approach(*) to assess chemical toxicity effectively. This study shows how inconclusive results are better than incorrect results and alert about which compounds need further analysis to be taken into account when making predictions (in chemico/in vitro testing or in vivo testing). In order to reduce the number of inconclusive results, the strategy must consists in feeding newly acquired data back into the modelling process to reduce eventually the need for further testing in the future.

    (*)Note: a Weight of Evidence (WoE) approach is used where data from a single source are considered insufficient to draw a conclusion but concordant information from several sources may be sufficient if certain conditions are met.The predictions from several methods can therefore be combined and may result in a more reliable outcome.