Chapter Edited and Updated by: Vishal Thovarai & Kevin Smith
Rational Drug DesignEdit
Serendipity has, for a very long time, been a driving force in the Drug Design and Development process. A shift towards rationalization of the process has begun, accompanying the increasing understanding of corresponding areas such as human as well as pathogen biochemistry and pathophysiology. In a broad and general way the rational drug design approach involves a comprehensive study of the affected biochemical pathways, identification of key elements (generally proteins) and designing small molecules to modify/manipulate the functions of these proteins. Most of the approaches are target oriented and hence structure based i.e. drug targets are identified which is typically a component (a protein or some other biomolecule) of a relevant pathway
Structure based Drug DesignEdit
In the early 1980s, researchers were not able to take advantage of structure-based methods in the drug discovery process. This was due to a number of factors, the most important of which were a lack of computing power and docking programs that could test potential models as well as a lack of interest in the established community do in part to the aforementioned lack of tools. However, in the 1990s, computing power and available programs increased exponentially as well as the ability to obtain cheaper, more reliable x-ray crystallographic structures necessary for any type of computational study. This marked the start of a new era and the first attempts and successes were published, with the two most famous being the HIV-1 protease inhibitors and renin inhibitors (to combat hypertension) (Lunney). In contemporary drug discovery, structure-based methods are an integral part of the drug development process. This change can be attributed to the rapid advances in genomics and structural biology, as well as developments in information technology. The advancements of technology in several fields that are vital to the drug discovery process increased the pace of drug development. However, many years of research are still required until a drug which is both effective and tolerated by the human body is marketable (Anderson). Despite this new technology and increased funding by pharmeaceutical companies, the issues with finding safe, effective compounds have resulted in there being no significant increase in the rate of therapeutic agent release to the general public. (Lindsay)
Approaches to Structure Based Drug DesignEdit
The de novo approachEdit
The de novo approach involves constructing novel drug molecules from the scratch based on the receptor/target structure. This is a particularly challenging task, considering that the search space of potentially feasible structures is to the order of 10100 (Bohacek). The fundamentals of de novo drug design include assembling drug like compounds, evaluating their potential and also searching the sample space for such compounds.
The design of the drug molecule is generally driven by the receptor structure or more specifically it’s hypothetical interaction with the structure of the receptor. Therefore, modeling the binding site to the greatest possible degree of accuracy is one of the critical steps in the process. Once the binding site is defined, the next step is to place atoms or small molecule within it and evaluate the fit/ interaction. This is done using a scoring function. Depending on the software package used, there now exists a wide variety of scoring functions ranging from simple steric constraints to functions that approximate and evaluate binding free energy. The scoring functions are also responsible for guiding the growth and optimization of structures by assigning fitness values to the evaluated structures.
Different ways of constructing complete ligand molecules (in silico) include linking, growing (Figure 1), random structure change, and molecular dynamics based methods.
Linking: Atoms, functional groups or fragments are placed at pre-determined interaction sites and are joined together to yield a complete molecule. The linking groups may be either pre-defined or generated on the fly in compliance with the constraints.
Growing: A single building block is chosen as a starting point or seed and placed in the interaction site of the receptor. Fragments or other groups are added to provide feasible interactions with the residues in the site. This process is continued until all fragments have been integrated into a single molecule.
Molecular Dynamics: The initial fragments are scattered randomly in the interaction site and then allowed to rearrange using Molecular dynamics simulations. Scoring functions are used to evaluate the resultant structures. One of the major drawbacks of de novo algorithms is that they generally ignore synthetic feasibility of the designed structures.
The Lead optimization approachEdit
The Lead optimization approach involves the modification and optimization of existing analogues (drug candidates) that demonstrate therapeutic potential. This technique is based on the study of the interaction between ligands and receptors. The data thus obtained is used to guide the modification of the structures of selected molecules or compounds. A large number of analogues are created and then screened through thereby improving efficacy and other desirable properties
Drug Target Selection and Identification of the Target SiteEdit
The selection of the drug target is mainly based on biological and biochemical considerations. Proteomics as a tool in this area is still relatively limited due to the sheer complexity of protein expression in any given cell. Despite this, proteomics’ usefulness has grown in other areas of the drug discovery process including biomarker identification and tracking. The ideal drug target for structure-based drug design should bind a small molecule and should be closely associated with the disease. The small molecule then either changes the function of the drug target, or in case of a pathogenic organism, inhibits the function of the target. This will ideally lead to the cell death of the pathogen. In the latter case, the drug target should only be present in the diseased cells or pathogen and should have a unique function that allows for and encourages this selectivity. Furthermore, the uniqueness of the drug target guarantees that another pathway cannot restore the function of an inhibited target. The structure-based search for anti-cancer and autoimmune drugs is much more challenging, since the drug targets regulate essential cell functions. Hence, these targets are not unique and isolated; inhibition of their function not only affects mutant or over/under-activated cells, but also normal cells. For example, the phosphoinositide 3-kinase (PI3K) pathway is both involved in cellular growth and has been directly linked with pancreatic cancer activation (Reddy). Trying to separate convoluted pathways such as this make the drug discovery process much more complex than with other diseases. Another option for targeting in cancer is DNA. Cisplatin and Bleomycin cross-link and cleave DNA respectively and are used to slow down cell division, particularly in cancerous tissues. (Singh)
The ligand binding site of a drug target should be a pocket containing both hydrogen donors and acceptors, as well as hydrophobic residues. In many cases, the selected target location is the active site of an enzyme, as with sildenafil citrate (Viagra), which targets the catalytic subunit of NADPH (Jeremy); however, it can also be an assembly or regulatory site, as is the case with the phosphotransferase regulatory domain the Bacillus subtilis Spo0f protein, a histidine kinase (Dai-Fu). Even protein-protein interaction sites, which are often large and planar, have been selected as target sites (2-oxoglutarate, a naturally occurring molecule, affects the monomer-monomer interactions of GlnK, an ammonia transporter [AcrB] inhibitor (Anderson, Stroud).
Proteomics as a Tool to Discover BiomarkersEdit
‘Biomarker’ is short for ‘biological marker.’ A biomarker is a molecule, indicator, or test that can be used to measure such processes as disease progression, infection stage, and drug efficacy, as well as other various biological functions. They can also be used as part of safety studies for therapeutic agents. Although the most common biomarkers in use today are small molecules and proteins, the growing fields of pharmacogenetics and pharmacogenomics are attempting to utilize genotypes, haplotypes, and single nucleotide polymorphisms as biomarkers (Frank).
With the growing emphasis on biomarkers as indicators, the National Institutes of Health (NIH) began a ‘Biomarkers and Surrogate Endpoint Working Group.’ This organization has set up a classification system for biomarkers.
Type 0 biomarkers are more symptomatic in nature and track disease progression over its complete history. They are used in phase 0 clinical studies. These clinical studies use well developed assay techniques in highly regulated populations for specific durations. The goal of these studies is to achieve simple positive or negative results for the drug or system being studied.
Type 1 biomarkers are used to track any type of compound injected into the biological system. The most familiar aspects are drug trials, where the researchers are looking for a specific effect. The effect observed may be positive or negative.
Type II biomarkers are used to determine a ‘surrogate endpoint.’ A surrogate endpoint, as defined by the FDA, “is a marker – a laboratory measurement or physical sign – that is used in the therapeutic trials as a substitute for a clinically meaningful endpoint that is a direct measure of how a patient feels, functions or survives and is expected to predict the effect of the therapy.” In other words, a surrogate endpoint moves beyond the concept of a single biomarker and into the realm where many or no biomarkers may be sufficient. Other symptoms including overall health and mortality may be studied to determine the efficacy of a treatment regimen. Although surrogate endpoints are still in the early stages, two examples that have been accepted are blood pressure and cholesterol, which have a firm connection to cardiovascular health and mortality (Frank).
In addition to the above effects, a biomarker should correlate well with the disease condition and minimize the number of false positives, as well as false negatives. Thus, a biomarker should be able to accurately discriminate between a normal and an infected condition with high reproducibility. The challenge in proteomics research is to identify unique biomarkers from complex biological mixtures that unequivocally correlate with the disease condition. Biomarkers can be utilized for many purposes. Also, established biomarkers are useful as risk factor indicators, capable of providing information to show that a person is susceptible to a disease. QT prolongation, a measure of change in the ventricular electrical cycle, is used as a assessment of a patient’s chances for survival after heart attack, as is troponin T, a cardiac enzyme whose levels rise following a heart attack. 5-hydroxytryptophan is a metabolic precursor to seratonin, and has been found to localize around neoplasms in neuroendocrine tissue. Labelling of these molecules allows for visualization of these events by fluorodeoxyglucose positron emission tomography (FDG-PET), which is used in visualizing many types of malignancies. Another PET application uses SPA-RQ, a neurokinin-1 binder, whose injection is used to track binding of the drug Aprepitant™, a drug used in chemotherapy patients to control vomiting (Frank).
There are currently many more biomarkers being studied and in development, and this number will continue to grow as our understanding of the body evolves. The combination of tracer molecules and imaging techniques, as in 5-hydroxytryptophan, creates a powerful method for visualizing areas that should not be touched surgically except as a last resort. The growing use of proteomics to find biomarkers is important as well. Recent advances in analytical techniques such as mass spectrometry and column chromatography have allowed for more comprehensive studies of protein expression. In addition, 2-dimensional electrophoresis provides an overall view of expression, in similar fashion to gene studies. Proteomics studies using these techniques have identified ovarian cancer, macular degeneration, and lipoprotein composition. Development of these techniques will undoubtedly increase in the future (Frank).
Asano, T.; Yao, Y.; Shin, S.; McCubrey, J.; Abbruzzese, J. L.; Reddy, S. A. Cancer Res. 2005, 65, 9164-9168.
Bleicher, K. H.; Bohm, H. J.; Muller, K.; Alanine, A. I. Nat. Rev. Drug Discov. 2003, 2, 369-378.
Bohacek, R. S.; McMartin, C.; Guida, W. C. “The art and practice of structure-based drug design: a molecular modelling perspective.” Med. Res. Rev. 1996, 16, 3–50
Cai, X. H.; Zhang, Q.; Shi, S. Y.; Ding, D. F. Acta Biochim. Biophys. Sin. (Shanghai) 2005, 37, 293-302.
Frank, R.; Hargreaves, R. Nat. Rev. Drug Discov. 2003, 2, 566-580.
Gruswitz, F.; O'Connell, J.,3rd; Stroud, R. M. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 42-47.
Lindsay, M. A. Nat. Rev. Drug Discov. 2003, 2, 831-838.
Lunney, E. Structure-Based Design and Two Aspartic Proteases. http://www.netsci.org/Science/Cheminform/feature01.html.
Muzaffar, S.; Shukla, N.; Srivastava, A.; Angelini, G. D.; Jeremy, J. Y. Br. J. Pharmacol. 2005, 146, 109-117.
Singh, S.; Malik, B. K.; Sharma, D. K. Bioinformation 2006.