Target Identification and Validation in Drug Development

The discovery of a new drug begins by understanding thebiological origin of a disease and identifying potential drug targets for intervention. The process of identifying and validating drug targets may look simple in theory, but it is easier said than done.

Many drug candidatesfailto reach the market, with numerous promising treatment options failing at the last minute in late-stage trials.Identifying suitable drug targets and carrying outeffective validation studies at an early stage can help avoid costly clinical failures.

This article will highlight the properties of an attractive drug target, outline the approaches used to identify targets and discuss the key steps involved in target validation.

Selection of the target

The process of drug discovery begins with the identification of a possible biological target and elucidating its role in the disease. A target is a biochemical entity (a protein, RNA, or gene) to which a drug can bind and elicit a physiological change. A target must possess an active site or binding pocket to which a potential drug can bind. Anoptimal targetshould be druggable, safe, efficient and able to fulfill commercial requirements.

Dr. Barbara Zdrazil, ChEMBL Coordinator,European Bioinformatics Institute说,“传统上,G protein-coupled受体(GPCRs) and kinases have served as very good drug targets for many years. These days, we see the emergence of new modalities for the treatment of disease which include previously less tractable targets. One example is the recent advances in cancer therapy by the approval of the first selective KRAS inhibitor, Sotorasenib, in 2021 for the treatment of tumors with theKRASG12C突变。此外,bromodomain抑制剂emerged as a promising class of anticancer agents. The concept of protein degradation by heterobifunctional small-molecule proteolysis-targeting chimeras (PROTACs) has also gained a lot of interest in recent years, as well as gene therapy (e.g., the use of CRISPR technology) which can offer personalized treatments for patients.”

Identification of drug–target interactions

A drug’spolypharmacologicalprofile can lead to desirable as well as adverse effects. Hence, profiling drug–target interactions is essential in drug discovery to maximize the therapeutic action and minimize undesirable effects.

Conventionalapproachesto identifying drug–target interactions (using biological experiments) were expensive and laborious, and have since been replaced with computational approaches. Due to their high efficiency and low costs, computational approaches have proved to be valuable methods for the prediction of drug–target interactions. Additionally, with advances in network pharmacology and systems biology, drug discovery approaches have drifted from the linear mode (one drug, one target and one disease approach) to the network mode (multi-drug, multi-target and multi-disease approach). This means that rather than selectively binding to one target, a single drug may work on multiple targets.

Pharmacophore-basedmethods to predict drug–target interactions include structure-based and ligand-based pharmacophore mapping.Structure-based methods rely on thethree-dimensional (3D)structures of targets.Dr.Behnoush Hajian, research scientist II at theBroad Institute of MIT and Harvardexplains, “Structure-based drug design (SBDD) continues to be a key tool in guiding the design of new therapeutic agents. Recent advances in structural biology have significantly accelerated the process of obtaining the 3D structures of drug targets bound to ligands or lead compounds.SBDD allows us to explore new chemical probes more confidentlyand improve the potency and selectivity of lead compounds in order to see the fruits of long and expensive drug discovery campaigns.”

Molecular dockingis a widely used structure-basedin silicomethod. Based on the 3D structures of targets, this method uses scoring functions to predict drug–target interactions and provides quantitative docking scores correlated with binding affinities. Ligand-based pharmacophore modelinguses structural information about the active ligands that bind to the target if the target structure is not available.

Network-based methodspredict novel drug target genes or drugs for repositioning through multiple algorithms. Compared to molecular docking-based methods, network-based methods are simple, fast and independent of the 3D structures of drug targets.

Recently, a variety of novel methodologies for target identification have also emerged. Thermal proteome profiling (TPP) is a recently developed tool that enables the monitoring of changes in protein thermal stability across the proteome using quantitative mass spectrometry.Dr.André Mateus, assistant professor at the Department of Chemistry at Umeå University, explains, “Thermal proteome profiling (TPP) is based on the principle that when proteins are heated, they denature and become insoluble. TPP is an extension of the cellular thermal shift assay (CETSA), where this principle was first applied directly in cells. By using quantitative mass spectrometry-based proteomics, TPP enables determining the thermal stability of each protein in a biological system (cells, tissues, biological fluids).”

“Any physical interaction of a protein with a ligand has the potential to alter its thermal stability. By monitoring proteome-wide thermal stability, it is possible to see which proteins change in thermal stability when cells or tissues are treated with a drug – those are likely to be the targets. Since the method is performed in cells, proteins maintain their natural environment. TPP is one of the only chemical biology methods that thus enables measuring drug–target engagement directlyin situ,” says Mateus.

Aside from this,machine learningand deep learning techniques are rapidly developing tools that are being used to predict drug–target interactions.

Target validationis a critical step in drug discovery

Like target identification, target validation is a crucial step in drug discovery. Target validation ensures thata molecular target is directly involved in a disease mechanism and that modulation of the target is likely to have a therapeutic effect.

Target validation may involve determining the structure–activity relationship, genetic manipulation of target genes (knockdown or overexpression), generating a drug-resistant mutant of the presumed target,usingdegradation-based tools to anticipate the effects of the target and monitoring signaling pathways downstream of the presumed target.

What are some recent trends in drug target identification?

Elaborating on some recent trends in drug target identification, Zdrazil continues, “One recent trend in drug target identification is to include genetic evidence for the target–disease associations, since targets with a genetically proven disease association appear to be moresuccessfulin delivering effective drugs.Another expanding area is the inclusion of pathway and network information which has helped researchers infer new disease associations for genes with no direct genetic support. In addition, many computational scientists are leveraging drug target information from bioinformatics databases such asChEMBL– which hosts mainly small molecule bioactivity data along with assay, target and lots of other types of information – as well as theOpen Targets Platform, which integrates genetic, genomics and other data for systematic drug target prioritisation.”