Send Orders for Reprints to
[email protected] Current Topics in Medicinal Chemistry, 2017, 17, 1-12
1
REVIEW ARTICLE
Recent Updates on Computer-aided Drug Discovery: Time for a Paradigm Shift Talambedu Usha1, Dhivya Shanmugarajan2, Arvind Kumar Goyal3, Chinaga Suresh Kumar4 and Sushil Kumar Middha2,* 1
Department of Biochemistry, Bangalore University, Bengaluru, Karnataka, India; 2DBT-BIF facility, Department of Biotechnology, Science Research Centre, Maharani Lakshmi Ammanni, College for Women, Bengaluru, Karnataka, India; 3Centre for Bamboo Studies, Department of Biotechnology, Bodoland University, Kokrajhar, Assam; 4Barrix Agro Sciences Pvt. Ltd., Department of Chemistry, Bengaluru, Karnataka, India
ARTICLE HISTORY Received: October 11, 2017 Revised: November 09, 2017 Accepted: November 22, 2017 DOI: 10.2174/1568026618666180101163651
Abstract: Computer-aided drug designing (CADD) has gained a wide popularity among biologists and chemists as a part of interdisciplinary drug discovery approach. It plays a vital role in the discovery, design and analysis of drugs in pharmaceutical industry. It is extensively used to reduce cost, time and speed up the early stage development of biologically new active molecules. In the current review we presented a brief review of CADD, merits and demerits, DNA, protein and enzyme as targets, types of CADD: Structure based drug designing (SBDD), ligand based drug designing (LBDD), Pharmacophore based drug designing (PBDD) and fragment based drug designing (FBDD), theory behind the types of CADD and their applications. The review also focuses on the in-silico pharmokinetic, pharmacodynamic and toxicity filters or predictions that play a major role in identifying the drug like molecules. Currently in pharmaceutical sciences computational tools and software are exhibiting imperative role in the different stages of drug discovery hence the review throws light on various commercial and freeware available for each step of CADD.
Keywords: Drug discovery, Pharmacophore, receptor-ligand, Molecular dynamics, in-silico. 1. INTRODUCTION Computational chemistry is a combinatorial subject/ research area comprising Chemistry and Information technology. This branch of Bioinformatics is gaining popularity amongst chemists and biologists due to its wide application in the discovery and design of the pharmaceutical products. Computational methods are more sophisticated and less time consuming compared to the experimental procedures which are laborious and time consuming. World’s leading pharmaceutical industries are widely using computational tools for the design and discovery of therapeutic products for various life threatening diseases [1]. In today’s time, the drug discovery process is largely driven by receptors, the biological targets, transgenic systems, genetic studies, next generation sequencing studies and protein technology. Drug discovery process, in general, is a daunting task for organic chemist due to the complexity of the pharmacophore feature that increases the property and efficiency of a drug. Life style and bad food habits are leading to a wide variety of dreadful diseases and new outbreaks of viral, bacterial diseases often leads to an international health emergency [2, 3]. Thus, there is always an urgent need for the discovery of new drugs with target specificity in a short time *Address correspondence to this author at the DBT-BIF Facility, Department of Biotechnology, Maharani Lakshmi Ammanni College for Women, Malleswaram, Bengaluru, Karnataka-560012, India; Tel: +91 80 23346781; Email:
[email protected] 1568-0266/17 $58.00+.00
frame and Computer-aided drug discovery is one such approach of drug discovery and design. Drug development is a complex, time consuming process involving preclinical testing followed by clinical trials in human subjects and costs approximately $ 1.2 billion with high chances of failure of a drug at any stage of drug development. On an average, in order to introduce one drug into the market it takes 10-16 years which includes target identification, validation, lead identification, synthesis, candidate optimization, preclinical studies, clinical trials (phase I, II and III), evaluation of the FDA approval and Phase IV studies [3]. Preclinical studies have been an integral part of pharmacology and medicine which require the sacrifice of a huge number of animals. In contrast, computational strategies like high through put screening of molecules help prioritize and identify small molecules for screening and subsequent toxicity testing while saving cost and time, minimize experimental testing and optimize overall uses of resources. They are also widely used to predict the target mutations leading to drug resistance, one of the biggest challenges in drug discovery [4, 5]. The tools and software suite are designed using various methods like machine learning, mathematical modeling algorithm and experimental data with significance. The statistic integrated tools are always user-friendly which avoid false positive and true negative on a final outcome. The tools and softwares for Computer-aided drug discovery are developed © 2017 Bentham Science Publishers
2 Current Topics in Medicinal Chemistry, 2017, Vol. 17, No. 30
and designed using significant experimental data, statistical approaches, machine learning and mathematical modelling algorithms. In silico drug designing or Computer-aided drug designing (CADD) can be carried out using numerous approaches that are briefly discussed in this review. The different forms of drug designing are Structure based drug designing (SBDD), ligand based drug designing (LBDD), Pharmacophore based drug designing (PBDD) and fragment based drug designing (FBDD). SBDD, LBDD, PBDD and FBDD are widely used in drug research for high throughput screening of the molecules that hastens the process of drug discovery. Recently, natural product research is gaining more importance among researchers. It is mainly due to reduced side effects of natural compounds compared to synthetic compounds [6-10]. In synthetic drugs, the initial precursors are preferably the parent natural compounds with a pharmacological activity. [8, 11]. The major advantage of CADD is to generate an ample number of compounds and screen them for target specific binding. Also, molecular dynamics simulation based compound validation is popular among computational chemistry. This kind of study will reveal atom- atom interaction of the compound with its active amino acid stability with respect to time. This review will briefly discuss recent trends in Computer-aided drug designing and its application with their work flow. 2. HISTORY OF CADD The concept of CADD was accepted by scientists in the beginning of the 20th century. In 1900, the concept of lock and key was put forward by P. Eh- rich (1909) and E. Fisher (1894) [12-14]. After few decades, the statistical approaches integrated with biological activities gained further importance and are known as Quantitative Structure-Activity Relationships (QSAR) [13]. In the beginning of 1980, CADD was combined with molecular biology concepts, structure prediction through multi-dimensional NMR molecular modeling, X-ray crystallography along with computer graphics gained more attention [15]. In 1990, the modern techniques in the chemistry and biology started to emerge and researchers showed a keen interest in various domains like the human genome, bioinformatics, combinatorial chemistry and high-throughput screening. Later in the 21st century, many models and tools were developed and still play a crucial role in drug discovery and its allied fields. Table 1 depicts all strategies involved in CADD. Table 1.
Strategies in CADD.
Protein
Ligand Known
Unknown
Known
Structure based drug designing
Denovo Drug designing
Unknown
Ligand based drug designing /Pharmacophores Alignment
Literature study and Rational screening (Library Design/ Analysis Diversity)
Usha et al.
3. THEORETICAL DRUG DISCOVERY PROCESS In silico drug designing begins with data collection of probable compounds that could be small molecules or peptides. Sketching of molecules and saving them in required formats for further studies can be done easily using appropriate tools and softwares (Table 2). Further, small molecule databases can be used to build similar or diverse compound library. Few extensively used small molecule databases are Pubchem (https://pubchem. ncbi.nlm.nih.gov/), drugbank (https://www.drugbank.ca/), Zinc (http://zinc.docking.org/),KEGGdrug (http://www. genome.jp/kegg/drug/), chembank (http://chembank. broadinstitute. org/), chemspider(http://www.chemspider.com/), WOMBAT (http://dud.docking.org/ wombat/), ChEBI (http://www.ebi.ac.uk/chebi/), Super Natural Database, and CancerPP( http://crdd.osdd.net/raghava/cancerppd/), THPdb (http://crdd.osdd.net/raghava/thpdb/), CAMPR3 (http:// www.camp.bicnirrh.res.in/) are few peptide databases from various sources that can be retrieved from omics tool (https://omictools.com/peptide-prediction-category). The next step is to screen large chemical libraries for structures that are most likely to bind to the target, which is also known as virtual screening. It involves use of numerous chemoinformatics tools to identify a hit molecule before being synthesized and experimented. In virtual screening, large chemical libraries are subjected to a series of filters like ADMET, TOPKAT, drug likeliness, and are considered mandatory for all kinds of in-silico drug designing (Fig. 1). Further, it is necessary to choose an appropriate strategy of drug designing based on the ligand or target information (Table 1). 4. ADME AND TOPKAT (TOXICITY PREDICTION BY KOMPUTER ASSISTED TECHNOLOGY) Pharmacokinetics is the study of the movement of the drug in and out of the body inclusive of absorption, distribution, metabolism, excretion (ADME). The efficiency of a drug is directly proportional to the concentration of the drug at the site of action. The route of administration, dose, an onset of action and duration of action of a drug can be efficiently determined by investigating the pharmacokinetic properties of the lead molecules. Computer-aided ADME predictions are done using open source and commercial tools (Table 3). These tools make predictions exclusively based on the structure of the molecule. The importance of these tools is associated with an screening of molecules at the reduced cost, failure and increased time efficacy to identify a subset of molecules for the subsequent steps [16]. Safety and efficacy are the two important facets of drug discovery and many drugs fail in phase IV of clinical trials or postmarketing surveillance due to side effects and adverse side effects. The side effects can be broadly classified into two types viz- predictable and non-predictable. The predictable side effects like vomiting nausea, depression, anxiety, restlessness, and adverse events like death are caused due to drug- drug interactions, drug-food interactions, mutagenic, carcinogenic and teratogenic nature of the drugs. Therefore, it is required to predict the toxicity of the lead molecules at the first drug research stage to ensure that the drug discovery process is foolproof, cost-effective and efficient. The computational tools indicate the toxicity of the molecules based on
Recent Updates on Computer-aided Drug Discovery
Table 2.
Current Topics in Medicinal Chemistry, 2017, Vol. 17, No. 30
3
List of tools and software for drawing and exporting chemical structures
S. no.
Software
Web Link
1
PubChem Sketcher
https://pubchem.ncbi.nlm.nih.gov/edit2/index.html
2
Marvin Sketch
https://www.chemaxon.com/download/marvinsuite/#marvin
3
ACD/ChemSketch
http://www.acdlabs.com/resources/freeware/chemsketch/
4
Chemdoodle
https://www.chemdoodle.com/
5
ChemWriter
https://chemwriter.com/
6
BIOVIA Draw 4.2
http://en.bio-soft.net/chemdraw/ISISDRAW.html
7
ChemDraw
http://en.bio-soft.net/chemdraw/ChemDraw.html
8
Zem
http://en.bio-soft.net/chemdraw/zem.html
9
ChemicPen 2.6
http://www.cetramax.com/
Fig. (1). Virtual screening process.
the physiochemical properties of the lead or ligand. Table 3 lists the available open source and commercial tools to identify the toxicity of the molecules [17]. 5. DRUG LIKENESS Drug likeliness describes a set of standard rules designed based on the molecular and physicochemical properties of the molecules used extensively for fast screening of drug like molecules. These rules are effective and efficient,however they not mandatory. Lipinski’s rule of five is thumb rule to evaluate the drug likeliness of a compound. He stated that a molecule should have Alogp
