Integrating computer-based de novo drug design and multidimensional filtering for desirable drugs

In the pharmaceutical industry today, many of the potent compounds discovered using expensive technologies are eventually rejected because of poor physicochemical or absorption, distribution, metabolism, excretion and toxicology (ADME/Tox) properties. This problem can be addressed by placing fast and accurate computational technologies at the heart of drug discovery. Chemically diverse and potent compounds generated by de novo design algorithms are scored for ADME/Tox properties using rigorously validated statistical models. Every molecule passing through this in silico pipeline is thus associated with a wealth of predicted properties, thereby allowing for rapid assessment to determine which molecule should be further developed. Critical to this idea is a platform that allows for the efficient exchange of in silico and experimental data between all scientists regardless of specialization. By bridging the gap between the in silico and experimental cultures in this fashion, an information-driven, cost-effective drug discovery program can be realized.     

A parsimonious software sensor for estimating the individual dynamic pattern of methane emissions from cattle

Large efforts have been deployed in developing methods to estimate methane emissions from cattle. For large scale applications, accurate and inexpensive methane predictors are required. Within a livestock precision farming context, the objective of this work was to integrate real-time data on animal feeding behaviour with an in silico model for predicting the individual dynamic pattern of methane emission in cattle. The integration of real-time data with a mathematical model to predict variables that are not directly measured constitutes a software sensor. We developed a dynamic parsimonious grey-box model that uses as predictor variables either dry matter intake (DMI) or the intake time (IT). The model is described by ordinary differential equations.Model building was supported by experimental data of methane emissions from respiration chambers. The data set comes from a study with finishing beef steers (cross-bred Charolais and purebred Luing finishing). Dry matter intake and IT were recorded using feed bins. For research purposes, in this work, our software sensor operated off-line. That is, the predictor variables (DMI, IT) were extracted from the recorded data (rather than from an on-line sensor). A total of 37 individual dynamic patterns of methane production were analyzed. Model performance was assessed by concordance analysis between the predicted methane output and the methane measured in respiration chambers. The model predictors DMI and IT performed similarly with a Lin’s concordance correlation coefficient (CCC) of 0.78 on average. When predicting the daily methane production, the CCC was 0.99 for both DMI and IT predictors. Consequently, on the basis of concordance analysis, our model performs very well compared with reported literature results for methane proxies and predictive models. As IT measurements are easier to obtain than DMI measurements, this study suggests that a software sensor that integrates our in silico model with a real-time sensor providing accurate IT measurements is a viable solution for predicting methane output in a large scale context.     

Animal models of amyotrophic lateral sclerosis and Huntington’s disease

Amyotrophic lateral sclerosis (ALS) and Huntington’s disease (HD) are debilitating neurodegenerative conditions for which there is no effective cure. Genetic determinants of both diseases have been identified, providing insight into neuropathological mechanisms and opportunities for therapeutic intervention. Aggregation of mutant proteins is the most prominent phenotype of these neurodegenerative diseases as is the case in Alzheimer’s disease and Parkinson’s disease. Here we review transgenic animal models of ALS and HD in mouse, zebrafish, C. elegans, and Drosophila that have been developed to study different aspects of disease progression. We also cover some large mammal transgenic models that have been recently developed. To effectively tackle these conditions will likely require effective use of several of these animal models, as each offers distinct advantages and insights into disease pathology.     

Integrated stoichiometric, thermodynamic and kinetic modelling of steady state metabolism

The quantitative analysis of biochemical reactions and metabolites is at frontier of biological sciences. The recent availability of high-throughput technology data sets in biology has paved the way for new modelling approaches at various levels of complexity including the metabolome of a cell or an organism. Understanding the metabolism of a single cell and multi-cell organism will provide the knowledge for the rational design of growth conditions to produce commercially valuable reagents in biotechnology. Here, we demonstrate how equations representing steady state mass conservation, energy conservation, the second law of thermodynamics, and reversible enzyme kinetics can be formulated as a single system of linear equalities and inequalities, in addition to linear equalities on exponential variables. Even though the feasible set is non-convex, the reformulation is exact and amenable to large-scale numerical analysis, a prerequisite for computationally feasible genome scale modelling. Integrating flux, concentration and kinetic variables in a unified constraint-based formulation is aimed at increasing the quantitative predictive capacity of flux balance analysis. Incorporation of experimental and theoretical bounds on thermodynamic and kinetic variables ensures that the predicted steady state fluxes are both thermodynamically and biochemically feasible. The resulting in silico predictions are tested against fluxomic data for central metabolism in Escherichia coli and compare favourably with in silico prediction by flux balance analysis.     

Advanced models of human skeletal muscle differentiation,development and disease: Three-dimensional cultures,organoids and beyond

Advanced in vitro models of human skeletal muscle tissue are increasingly needed to model complex developmental dynamics and disease mechanisms not recapitulated in animal models or in conventional monolayer cell cultures. There has been impressive progress towards creating such models by using tissue engineering approaches to recapitulate a range of physical and biochemical components of native human skeletal muscle tissue. In this review, we discuss recent studies focussed on developing complex in vitro models of human skeletal muscle beyond monolayer cell cultures, involving skeletal myogenic differentiation from human primary myoblasts or pluripotent stem cells, often in the presence of structural scaffolding support. We conclude with our outlook on the future of advanced skeletal muscle three-dimensional cultures (e.g. organoids and biofabrication) to produce physiologically and clinically relevant platforms for disease modelling and therapy development in musculoskeletal and neuromuscular disorders.     

Network design meets in silico evolutionary biology

Cell fate is programmed through gene regulatory networks that perform several calculations to take the appropriate decision. In silico evolutionary optimization mimics the way Nature has designed such gene regulatory networks. In this review we discuss the basic principles of these evolutionary approaches and how they can be applied to engineer synthetic networks. We summarize the basic guidelines to implement an in silico evolutionary design method, the operators for mutation and selection that iteratively drive the network architecture towards a specified dynamical behavior. Interestingly, as it happens in natural evolution, we show the existence of patterns of punctuated evolution. In addition, we highlight several examples of models that have been designed using automated procedures, together with different objective functions to select for the proper behavior. Finally, we briefly discuss the modular designability of gene regulatory networks and its potential application in biotechnology.     

Cunninghamella--a microbial model for drug metabolism studies--a review

Drug metabolism studies constitute an important and necessary step in the evaluation of drug efficacy and safety. In vivo drug metabolism studies suffer from many disadvantages. Hence there is a rise in validation of in vitro microbial models. This review describes the transformation studies of drugs by the fungus, Cunninghamella and correlating them with the metabolism/biotransformation in animal systems and providing technical methods to develop microbial models. Emphasis is laid on the potential of Cunninghamella fungus to mimic mammalian drug biotransformations and to use as in vitro model for drug metabolism studies and for further toxicological and pharmacological studies of metabolites.     

Modeling of human M1 aminopeptidases for in silico screening of potential Plasmodium falciparum alanine aminopeptidase (PfA-M1) specific inhibitors

Plasmodium falciparum alanine M1-aminopeptidase (PfA-M1) is a validated target for anti-malarial drug development. Presence of significant similarity between PfA-M1 and human M1-aminopeptidases, particularly within regions of enzyme active site leads to problem of non-specificity and off-target binding for known aminopeptidase inhibitors. Molecular docking based in silico screening approach for off-target binding has high potential but requires 3D-structure of all human M1-aminopeptidaes. Therefore, in the present study 3D structural models of seven human M1-aminopeptidases were developed. The robustness of docking parameters and quality of predicted human M1-aminopeptidases structural models was evaluated by stereochemical analysis and docking of their respective known inhibitors. The docking scores were in agreement with the inhibitory concentrations elucidated in enzyme assays of respective inhibitor enzyme combinations (r2≈0.70). Further docking analysis of fifteen potential PfA-M1 inhibitors (virtual screening identified) showed that three compounds had less docking affinity for human M1-aminopeptidases as compared to PfA-M1. These three identified potential lead compounds can be validated with enzyme assays and used as a scaffold for designing of new compounds with increased specificity towards PfA-M1.     

Improving Compound Quality through in vitro and in silico Physicochemical Profiling

Many compounds entering clinical studies do not survive the numerous hurdles for a good pharmacological lead to a drug on the market. The reasons for attrition have been widely studied which resulted in more early attention to compound quality related to physical chemistry, drug metabolism and pharmacokinetics (DMPK), and toxicology/safety. This paper will briefly review current physicochemical in vitro assays and in silico predictions to support compound and library design through to lead optimization. The most important physicochemical properties include lipophilicity (log P/D), pK_a, solubility, and permeability. These drive key ADMET properties such as absorption, cell penetration, access to the brain, volume of distribution, plasma protein binding, metabolism, and toxicity, as well as biopharmaceutical behavior. Much data are now available from medium‐ to high‐throughput physchem and ADMET in vitro assays, either in the public domain (see, e.g., PubChem, PubMed) or in drug companies' in‐house databases. Such data are increasingly being computer‐modelled and used in predictive chemistry. New pipelining technology makes it easier to build and update QSAR models so that such models can use the latest available data to produce robust local and global predictive in silico ADMET models.     

The use of transgenic systems in pharmaceutical research.

Those pharmaceutical companies whose goal is to generate novel innovative drugs are faced with the challenge that only a fraction of the compounds tested in clinical trials eventually become a registered drug. This problem of attrition is compounded by the fact that the clinical trial or development stage is by far the most costly phase of bringing a new drug to market, consuming around 80 per cent of the total spend. Transgenic technology represents an attractive approach to reducing the attrition rate of compounds entering clinical trials by increasing the quality of the target and compound combinations making the transition from discovery into development. Transgenic technology can impact at many points in the discovery process, including target identification and target validation, and provides models designed to alert researchers early to potential problems with drug metabolism and toxicity, as well as providing better models for human diseases. In target identification, transgenic animals harbouring large DNA fragments can be used to narrow down genetic regions. Genetic studies often result in the identification of large genomic regions and one way to decrease the region size is to do complementation studies in transgenic animals using, for example, inserts from bacterial artificial chromosome (BAC) clones. In target validation, transgenic animals can be used for in vivo validation of a specific target. Considerable efforts are being made to establish new, rapid and robust tools with general utility for in vivo validation, but, so far, only transgenic animals work reliably on a wide range of targets. Transgenic animals can also be used to generate better disease models. Predictive animal models to test new compounds and targets will significantly speed up the drug discovery process and, more importantly, increase the quality of the compounds taken further in the research and development process. Humanised transgenic animals harbouring the human target molecule can be used to understand the effect of a compound acting on the human target in vivo. Also, models mimicking human drug metabolism will provide a means of assessing the effect of human-specific metabolites and of understanding the pharmacokinetic properties of potential drugs. In toxicology studies, transgenic animals are providing more predictive models. A good example of this are those models routinely used to look for carcinogenicity associated with new compounds.     

Stoichiometric network reconstruction and analysis of yeast sphingolipid metabolism incorporating different states of hydroxylation

The first elaborate metabolic model of Saccharomyces cerevisiae sphingolipid metabolism was reconstructed in silico. The model considers five different states of sphingolipid hydroxylation, rendering it unique among other models. It is aimed to clarify the significance of hydroxylation on sphingolipids and hence to interpret the preferences of the cell between different metabolic pathway branches under different stress conditions. The newly constructed model was validated by single, double and triple gene deletions with experimentally verified phenotypes. Calcium sensitivity and deletion mutations that may suppress calcium sensitivity were examined by CSG1 and CSG2 related deletions. The model enabled the analysis of complex sphingolipid content of the plasma membrane coupled with diacylglycerol and phosphatidic acid biosynthesis and ATP consumption in in silico cell. The flux data belonging to these critically important key metabolites are integrated with the fact of phytoceramide induced cell death to propose novel potential drug targets for cancer therapeutics. In conclusion, we propose that IPT1, GDA1, CSG and AUR1 gene deletions may be novel candidates of drug targets for cancer therapy according to the results of flux balance and variability analyses coupled with robustness analysis.     

Metabolite-centric approaches for the discovery of antibacterials using genome-scale metabolic networks

Development of genome-scale metabolic models and various constraints-based flux analyses have enabled more sophisticated examination of metabolism. Recently reported metabolite essentiality studies are also based on the constraints-based modeling, but approaches metabolism from a metabolite-centric perspective, providing synthetic lethal combination of reactions and clues for the rational discovery of antibacterials. In this study, metabolite essentiality analysis was applied to the genome-scale metabolic models of four microorganisms: Escherichia coli, Helicobacter pylori, Mycobacterium tuberculosis and Staphylococcus aureus. Furthermore, chokepoints, metabolites surrounded by enzymes that uniquely consume and/or produce them, were also calculated based on the network properties of the above organisms. A systematic drug targeting strategy was developed by combining information from these two methods. Final drug target metabolites are presented and examined with knowledge from the literature.     

Machine Learning Prediction of Cancer Cell Sensitivity to Drugs Based on Genomic and Chemical Properties

Predicting the response of a specific cancer to a therapy is a major goal in modern oncology that should ultimately lead to a personalised treatment. High-throughput screenings of potentially active compounds against a panel of genomically heterogeneous cancer cell lines have unveiled multiple relationships between genomic alterations and drug responses. Various computational approaches have been proposed to predict sensitivity based on genomic features, while others have used the chemical properties of the drugs to ascertain their effect. In an effort to integrate these complementary approaches, we developed machine learning models to predict the response of cancer cell lines to drug treatment, quantified through IC₅₀ values, based on both the genomic features of the cell lines and the chemical properties of the considered drugs. Models predicted IC₅₀ values in a 8-fold cross-validation and an independent blind test with coefficient of determination R² of 0.72 and 0.64 respectively. Furthermore, models were able to predict with comparable accuracy (R² of 0.61) IC50s of cell lines from a tissue not used in the training stage. Our in silico models can be used to optimise the experimental design of drug-cell screenings by estimating a large proportion of missing IC₅₀ values rather than experimentally measuring them. The implications of our results go beyond virtual drug screening design: potentially thousands of drugs could be probed in silico to systematically test their potential efficacy as anti-tumour agents based on their structure, thus providing a computational framework to identify new drug repositioning opportunities as well as ultimately be useful for personalized medicine by linking the genomic traits of patients to drug sensitivity.     

Small Molecule Proprotein Convertase Inhibitors for Inhibition of Embryo Implantation

Uterine proprotein convertase (PC) 6 plays a critical role in embryo implantation and is pivotal for pregnancy establishment. Inhibition of PC6 may provide a novel approach for the development of non-hormonal and female-controlled contraceptives. We investigated a class of five synthetic non-peptidic small molecule compounds that were previously reported as potent inhibitors of furin, another PC member. We examined (i) the potency of these compounds in inhibiting PC6 activity in vitro; (ii) their binding modes in the PC6 active site in silico; (iii) their efficacy in inhibiting PC6-dependent cellular processes essential for embryo implantation using human cell-based models. All five compounds showed potent inhibition of PC6 activity in vitro, and in silico docking demonstrated that these inhibitors could adopt a similar binding mode in the PC6 active site. However, when these compounds were tested for their inhibition of decidualization of primary human endometrial stromal cells, a PC6-dependent cellular process critical for embryo implantation, only one (compound 1o) showed potent inhibition. The lack of activity in the cell-based assay may reflect the inability of the compounds to penetrate the cell membrane. Because compound''s lipophilicity is linked to cell penetration, a measurement of lipophilicity (logP) was calculated for each compound. Compound 1o is unique as it appears the most lipophilic among the five compounds. Compound 1o also inhibited another crucial PC6-dependent process, the attachment of human trophoblast spheroids to endometrial epithelial cells (a model for human embryo attachment). We thus identified compound 1o as a potent small molecule PC6 inhibitor with pharmaceutical potential to inhibit embryo implantation. Our findings also highlight that human cell-based functional models are vital to complement the biochemical and in silico analyses in the selection of promising drug candidates. Further investigations for compound 1o are warranted in animal models to test its utility as an implantation-inhibiting contraceptive drug.     

A Novel Mathematical Model Describing Adaptive Cellular Drug Metabolism and Toxicity in the Chemoimmune System

Cells cope with the threat of xenobiotic stress by activating a complex molecular network that recognizes and eliminates chemically diverse toxic compounds. This “chemoimmune system” consists of cellular Phase I and Phase II metabolic enzymes, Phase 0 and Phase III ATP Binding Cassette (ABC) membrane transporters, and nuclear receptors regulating these components. In order to provide a systems biology characterization of the chemoimmune network, we designed a reaction kinetic model based on differential equations describing Phase 0–III participants and regulatory elements, and characterized cellular fitness to evaluate toxicity. In spite of the simplifications, the model recapitulates changes associated with acquired drug resistance and allows toxicity predictions under variable protein expression and xenobiotic exposure conditions. Our simulations suggest that multidrug ABC transporters at Phase 0 significantly facilitate the defense function of successive network members by lowering intracellular drug concentrations. The model was extended with a novel toxicity framework which opened the possibility of performing in silico cytotoxicity assays. The alterations of the in silico cytotoxicity curves show good agreement with in vitro cell killing experiments. The behavior of the simplified kinetic model suggests that it can serve as a basis for more complex models to efficiently predict xenobiotic and drug metabolism for human medical applications.     

Determining Disease Intervention Strategies Using Spatially Resolved Simulations

Predicting efficacy and optimal drug delivery strategies for small molecule and biological therapeutics is challenging due to the complex interactions between diverse cell types in different tissues that determine disease outcome. Here we present a new methodology to simulate inflammatory disease manifestation and test potential intervention strategies in silico using agent-based computational models. Simulations created using this methodology have explicit spatial and temporal representations, and capture the heterogeneous and stochastic cellular behaviours that lead to emergence of pathology or disease resolution. To demonstrate this methodology we have simulated the prototypic murine T cell-mediated autoimmune disease experimental autoimmune encephalomyelitis, a mouse model of multiple sclerosis. In the simulation immune cell dynamics, neuronal damage and tissue specific pathology emerge, closely resembling behaviour found in the murine model. Using the calibrated simulation we have analysed how changes in the timing and efficacy of T cell receptor signalling inhibition leads to either disease exacerbation or resolution. The technology described is a powerful new method to understand cellular behaviours in complex inflammatory disease, permits rational design of drug interventional strategies and has provided new insights into the role of TCR signalling in autoimmune disease progression.     

How to make predictions about future infectious disease risks

Formal, quantitative approaches are now widely used to make predictions about the likelihood of an infectious disease outbreak, how the disease will spread, and how to control it. Several well-established methodologies are available, including risk factor analysis, risk modelling and dynamic modelling. Even so, predictive modelling is very much the 'art of the possible', which tends to drive research effort towards some areas and away from others which may be at least as important. Building on the undoubted success of quantitative modelling of the epidemiology and control of human and animal diseases such as AIDS, influenza, foot-and-mouth disease and BSE, attention needs to be paid to developing a more holistic framework that captures the role of the underlying drivers of disease risks, from demography and behaviour to land use and climate change. At the same time, there is still considerable room for improvement in how quantitative analyses and their outputs are communicated to policy makers and other stakeholders. A starting point would be generally accepted guidelines for 'good practice' for the development and the use of predictive models.