Evaluating the Impact of In Silico Modelling: Drug Development and Clinical Research

Some of the biggest pharmaceutical leaders are prioritising the integration of computational modelling within their drug development process. In silico models are becoming increasingly popular not only in drug discovery, but across clinical research. The implementation of in silico in R&D and clinical research has demonstrated a number of advantages over conventional methods, and with technology evolving at an exponential rate, in silico methods could become a key part of drug development. 

In silico refers to experiments performed via computer simulation in the form of equations or rules. This form of computational modelling provides researchers with the tools to qualitatively and quantitatively evaluate different types of treatments for a specific disease, as well as testing a larger set of different conditions (e.g. dosing). 

In comparison to in-vivo techniques, which are performed in whole, living organisms, in silico modelling offers more practical, economical experiments. Furthermore, computational methods limit the use of animal models in research which is becoming an increasingly desirable approach, both in terms of ethical considerations and from a time/cost perspective. 

Pharmaceutical applications

Drug development and repurposing 

The process of drug development continues to experience issues with rising costs, increasing demand and shorter timelines. Innovative approaches are required to better identify drug targets and efficacy predictions in a shorter period of time for growing patient populations. 

Computer-aided drug design (CADD) is a group of computational methods which offer a cost-effective way of identifying drug candidates. There are two types of CADD approach: structure-based design (SB-CADD) and ligand-based design (LB-CADD). 

SB- CADD methods analyse the 3D structural information of macromolecules, typically proteins or RNA, in order to identify key sites and interactions that are important for their respective biological functions. LB-CADD, on the other hand, focuses on known ligands for a target in order to establish a relationship between their physicochemical properties and activities, referred to as a structure-activity relationship (SAR). 

While SB-CADD relies on the knowledge of the target protein structure, LB-CADD uses reference structures collected from the compounds which interact with the target, and analyses their 2D/3D structure. The objective aim is to predict the mechanism and strength of binding for a specific molecule to the target. 

CADD methods offer a higher probability of discovering drug candidates with the desired properties by better understanding the target’s structure and ligand interactions. This is particularly beneficial for pharma companies, for which this will increase the likelihood of a compound overcoming the barriers of preclinical testing. 

In addition to optimising drug design, in silico models have been used in drug repurposing. Drug repurposing is an important strategy for identifying “new uses for approved or investigational drugs that are outside the scope of the original medical indication”. 

Network-based drug-repurposing (NB-DRP) is a prime example of the integration of in silico modelling into drug development. In NB-DRP, the relationships between biological compounds are organised into networks in order to identify emerging properties at a network level. The network allows users to examine how cellular systems undergo different biological phenotypes, under various conditions. 

The network is created as a connected graph, in which each node represents a drug or biological target within a target pathway. The benefit this network brings is a perspective to complex diseases which arise from the interaction of many biological networks. Understanding the implication of other networks within disease pathology can highlight other potential diseases which could be treated with the same drug candidate. 

Molecular docking

Molecular docking is a useful tool used to quantify the interaction of a protein with a small-molecule ligand in a complex. Studies using molecular docking have been used to determine whether a drug has the potential to bind to other targets. This method has been exploited to repurpose drugs for different targets including SARS-Cov-2. 

A popular in silico method, molecular docking, is a convenient method of rapidly screening extensive libraries of ligands and targets – this is critical for accelerating drug discovery, as well as identifying the most suitable drug candidate. 

One of the main drawbacks with molecular docking is ensuring appropriate scoring functions and algorithms are implemented, which could otherwise compromise molecular screening. However, post-processing docking results have been developed to overcome this problem with more accurate scoring functions. 

Clinical applications

In silico imaging in clinical trials

While conventional clinical trials can inform whether a product/technology is unsafe or ineffective, they often fail to explain why or how to improve it. In silico clinical trials utilise computer simulation in the development or evaluation of a medical device, intervention or product. This overcomes the challenges of conventional clinical trials, by creating algorithms that identify an error or simulating potential improvements.

In silico imaging is described as the computational simulation for an entire imaging system. The simulation creates the source, the objection, detection and interpretation that are typically used to evaluate new technologies. The primary endpoint of these clinical trials is to determine the scientific value of imaging technology, in comparison to the standard of care.  

As a computational model, in silico clinical trials require ‘virtual patients’ (VPs). The first step in creating a virtual patient population is using the Virtual Physiological Human (VPH). The VPH is a collective framework that shares resources by many organisations, to integrate computer models of the mechanical, physical, and biochemical functions of a living human body. 

VPs offer significant benefits in comparison with human volunteers. In the case of COVID-19, VPs could have potentially predicted whether specific vaccines were likely to work, as well as the potential side effects without the need to test on living candidates, which saves time and cost. However, the reporting of adverse events, the placebo effect, and treatment preference are essential for clinical research, and whether VPs could replicate this is yet to be observed. 

Computational-based approaches show promise in accelerating drug development and revolutionising clinical research. In the future, In silico modelling has shown great potential to support drug discovery for precision medicine, developing effective and safe therapeutic options for a host of complex diseases.

Charlotte Di Salvo, Editor & Lead Medical Writer
PharmaFeatures