Frontiers in Computational Chemistry

Drug discovery and development is a time-consuming, complex, and expensive process. Usually, it takes about 15 years in the best scenario since drug candidates have a high attrition rate. Therefore, drug development projects rarely take place in low and middle-income countries (LMICs). Traditionally, this process consists of four sequential stages: (1) target identification and early drug discovery, (2) preclinical studies, (3) clinical development, and (4) review, approval and monitoring by regulatory agencies.

During the last decades, computational tools have offered interesting opportunities for Research and Development (R & D) in LMICs, since these techniques are affordable, reduce wet lab experiments in the first steps of the drug discovery process, reduce animal testing by aiding experiment design, and also provide key knowledge involving clinical data management as well as statistical analysis.

This book chapter aims to highlight different computational tools to enable early drug discovery and preclinical studies in LMICs for different pathologies, including cancer. Several strategies for drug target selection are discussed: identification, prioritization and validation of therapeutic targets; particularly focusing on high-throughput analysis of different “omics” approaches using publicly available data sets. Next, strategies to identify and optimize novel drug candidates as well as computational tools for costeffective drug repurposing are presented. In this stage, chemoinformatics is a key emerging technology. It is important to note that additional computational methods can be used to predict possible uses of identified human-aimed drugs for veterinary purposes. Application of computational tools is also possible for predicting pharmacokinetics and pharmacodynamics as well as drug-drug interactions. Drug safety is a key issue and it has a profound impact on drug discovery success. Finally, artificial intelligence (AI) has also served as a potential tool for drug design and discovery, expected to be a revolution for drug development in several diseases.

It is important to note that the development of drug discovery projects is feasible in LMICs and in silico tools are expected to potentiate novel therapeutic strategies in different diseases.

This book chapter aims to highlight different computational tools to enable early drug discovery and preclinical studies in LMICs for different pathologies, including cancer. Several strategies for drug target selection are discussed: identification, prioritization and validation of therapeutic targets; particularly focusing on high-throughput analysis of different “omics” approaches using publicly available data sets. Next, strategies to identify and optimize novel drug candidates as well as computational tools for costeffective drug repurposing are presented. In this stage, chemoinformatics is a key emerging technology. It is important to note that additional computational methods can be used to predict possible uses of identified human-aimed drugs for veterinary purposes.

Application of computational tools is also possible for predicting pharmacokinetics and pharmacodynamics as well as drug-drug interactions. Drug safety is a key issue and it has a profound impact on drug discovery success.

Finally, artificial intelligence (AI) has also served as a potential tool for drug design and discovery, expected to be a revolution for drug development in several diseases.Application of computational tools is also possible for predicting pharmacokinetics and pharmacodynamics as well as drug-drug interactions. Drug safety is a key issue and it has a profound impact on drug discovery success. Finally, artificial intelligence (AI) has also served as a potential tool for drug design and discovery, expected to be a revolution for drug development in several diseases.

 

The study of new nanomaterials with potential applications as drug carriers and biosensors is based on the interactions between adsorbate (drug/biomolecule) and adsorbent (nanomaterial). Experimentally, the study of these cases has several economic efforts because of the high cost of carrying out all experiments. In this sense, computational chemistry is beginning to become a useful tool for designing and developing new nanostructures with the possible application as drug carriers and biosensors, with less economic resources. In literature, several works evidence the usefulness of computational chemistry in this area, promoting the proposal of new nanomaterials with peculiar characteristics.

In this regard, the present chapter shows an overview of the study of drug carriers and biosensors from an adsorption process point of view. Also, some adsorbent materials are exemplified, as well as the main interactions present at the adsorbate-adsorbent complex formation. Later, a bridge between computational chemistry and the adsorption phenomena is highlighted, as well as some electronic parameters in the framework of the density functional theory useful in these studies. Finally, two cases are represented: the application of molecular modeling for the study of drug-carrier nanostructures, and the design and modeling of biosensors based on nanostructures.

Most computational studies of biologically relevant systems have used Molecular Mechanics (MM). While MM is generally reliable for many applications, chemical reactions and bond formations/breakage are not describable in MM. In contrast, Quantum Mechanics (QM) is an approach that utilizes wave functions and/or electron density functions for property and structural analyses and hence does not suffer from such limitations. QM methods can be classified into two main frameworks, ab initio and semi-empirical. Semi-empirical methods utilize experimental or ab initio results to make additional approximations, thereby using a combination of some ab initio calculations and fitted experimental data. Despite the accuracy and general applicability of QM, the major disadvantages are limitations due to the system size. Not surprisingly, hybrid methods that partition the problem at hand into subsystems have been developed. Some of these methods mix QM with MM, and others are strictly QM, but limit the range of interactions. As a result, there exists a plethora of methods, some with fanatical followers, with the result that researchers are often faced with bewildering choices.

This review, perhaps more accurately described as a mini-review or perspective, examines recent calculations on biologically relevant (including biomimetic molecules) in which QM is necessary, to a greater or lesser degree, to obtain results that are consistent with the experiment. The review is not an exposition on the theoretical foundations of different methods, but rather a practical guide for the researcher with an interest in using computational methods to produce biologically, or at least biochemically, useful results. Because of our own specific interests, the Arg-Gly-Asp sequence, or so-called RGD, figures prominently in the work, in terms of size, including oligomers of RGD, and strengths of interactions. A key feature of RGD is its role in the binding of cells to the Extra Cellular Matrix (ECM) depending on the cell type and receptor protein on the cell itself. The ECM is comprised of spectra of biological compounds such as proteoglycans and fibrous proteins; RGD is located and found as a motif on these fibrous proteins. The cell bindings to the ECM are done via integrin-RGD binding. Because metal interactions and hydrogen bonding significantly affect integrin-RGD binding, theoretical methodology beyond MM is needed. IntegrinRGD binding affects the adhesion and movement of cells along the ECM. Hence, these interactions are highly relevant to understanding the spread of cancer in an organism.

The G-protein coupled receptor GPCR family is the most numerous and diversified set of membrane receptors linked with various neurological disorders like Epilepsy, Alzheimer's disease, Fronto-temporal dementia, Vascular dementia, Parkinson's disease, and Huntington's disease. They provide messages to the cell by interacting with various ligands, which include hormones, neurotransmitters, and photons. They are the focus of roughly one-third of the medications on the market today. Similarly, the subtype of the serotonin receptor, 5-hydroxytryptamine 2B (5-HT2B), belongs to the G-protein receptor (GPCR) class-A family and is a sensitive class prone to deactivation and activation. There has been an increasing interest in the structural geometry of the receptor upon ligand binding to the allosteric site. The cavities at the receptor-lipid interface are an unusual allosteric binding region that presents numerous issues concerning ligand interactions and stability, binding site conformation, and how the lipid molecules alter all these molecular modeling mechanisms provide an insight into the docking and binding of drug and structural variations. For instance, ligand recognition in the neuronal adenosine receptor type 2A (hA2AR), a GPCR related to various neurodegenerative disorders, was investigated for its affinity against an inhibitor in a solvated neuronal-like membrane in metadynamics. The study provided a factual description of atomic interactions between the ligand and the receptor. It was supported by in vitro binding affinity studies for highlighting the importance of membrane lipids and protein extracellular loop regions, thus, providing valuable input for ligand design and targeting GPCR. Since 5HT is essential as a target for various pharmaceutical and recreational drugs, studies are gaining pace regarding its seven subtypes. In research, general molecular design is carried out, including homology modeling, docking, dynamics, and a hallucinogen-specific chemogenomics database for pharmacological analysis of small molecules and their potential targets. The analogs of piperidine and piperazine moieties were investigated against the 5HT2A receptor via pharmacophore modeling, 3D-Quantitative Structure-Activity Relationship (3D-QSAR), Molecular docking, and Absorption Distribution Metabolism Excretion (ADME) studies. With the onset of multiscale molecular modeling, it is now possible to apply multiple levels of theory to a system of interest, such as assigning chemically relevant regions to high quantum mechanics (QM) theory while treating the rest of the system with a classical force field (molecular mechanics (MM) potential). Several groups have explored the atomic level of interaction between the ligand and the allosteric site via molecular docking and dynamics simulations, followed by quantum chemical calculations to achieve specific results and strengthen the analysis. Quantum Mechanics/Molecular Mechanics (QM/MM) is employed by considering conformational plasticity to identify the critical binding site residues responsible for modifying GPCR function. By this path, the geometry of the receptor is analyzed either by fixing its position w.r.t. to the ligand or by choosing a bound ligand. Finally, structure-based drug design (SBDD) methodologies will be more efficient. Density Functional Theory (DFT) calculations reveal the stabilization of the molecular structure to depict the interactions. Various study groups also practice Fragment-based lead discovery methods for GPCR-based drug discovery. Creating leads from fragments is complicated, accurate, and dependable computational methods are employed to explore G protein-coupled receptor as a target via molecular dynamics simulations and the free energy perturbation approaches (MD/FEP). The overall knowledge of GPCR-mediated signaling can be expanded using such computational approaches.

Inflammation is a natural response to external stimuli related to the protection of the organism. However, their exaggerated reaction can cause severe damage to the body and is related to several diseases, including allergies, rheumatoid arthritis, diabetes, cancer, and various infections. Furthermore, inflammation is mainly characterized by increased temperature, pain, flushing, and edema due to the production of pro-inflammatory cytokines, such as prostaglandins, and can be controlled using anti-inflammatory drugs. In this sense, selective prostaglandin E2 (PGE2 ) inhibition has been targeted and explored for designing new compounds for anti-inflammatory drugs because it can show fewer side effects than non-steroidal antiinflammatory drugs (NSAIDs) and corticosteroids. It is a bioactive lipid overproduced during an inflammatory process, produced mainly by COX-1, COX-2, and microsomal prostaglandin E2 synthase-1 (mPGES-1). Recently, studies have demonstrated that mPGES-1 inhibition is an excellent strategy for designing anti-inflammatory drugs, which could protect against pain, arthritis, acute inflammation, autoimmune diseases, and different types of cancers. Also, in recent years, Computer-Aided Drug Design (CADD) approaches have been increasingly used to design new inhibitors, decreasing costs and increasing the probability of discovering active substances and constantly applying them to discover mPGES-1 inhibitors. Thus, here, this chapter will approach the latest advances in computational methods to discover new mPGES-1 inhibitors that can be promising against several inflammatory conditions. The focus is on techniques such as molecular docking and dynamics, virtual screenings, pharmacophore modeling,fragment-based drug design, quantitative structure-activity relationship (QSAR), and others explored by researchers worldwide that can lead to the design of a promising drug against this target.