The first step towards developing a medicine for any disease is identifying your targets — which proteins are involved with the pathology. Only after a target is identified can scientists design a drug to combat and resolve the disease. However, the process of identifying and subsequently validating targets can be challenging. Simply knowing what makes a good target is up for debate, let alone which are the best methods for going out and finding them.
Dr Janet Brownlees is a Senior Director at MSD’s Discovery Centre whose main area of research is focused on developing treatments for neurodegenerative diseases. Dr Brownlees moderated Oxford Global’s Target Identification, Phenotypic, and Cell Based Screening panel discussion at Discovery UK 2021.
Joining Dr Brownlees was Dr Wei, a Senior Scientist, also at MSD, who specialises in bioinformatics and neurological disease. Dr Davide Gianni, a Senior Director of the Functional Genomics Centre at AstraZeneca, was our final panellist. Dr Gianni’s team uses genome editing technologies in conjunction with bioinformatics and AI to analyse drug screenings.
The Balance Between Novel and Validated Targets
The first question that the panel discussed was ‘what makes a good target?’ Dr Brownlees introduced the question by commenting that “most people say that a target is not validated until it gets in the clinic and works.” The question was then posed to the other panellists to discuss their opinions.
Dr Gianni said that selection of the right target was the most important decision to make in drug discovery
Dr Gianni said that selection of the right target was the most important decision to make in drug discovery, and how to make that decision had no single answer, but rather relied on a balance of multiple conditions. The conditions that Dr Gianni is interested in are the genetic link between target and disease, tractability, modalities used to interact with targets, and their safety profile.
These are all considerations that Dr Gianni explained are relevant when it comes to generating lists of targets to “hand over” to biologists: “how one goes about the prioritisation of those lists, I think is important.”
Dr Brownlees then handed the microphone over to the audience to talk about their experience in selecting the targets that bioinformaticians like Dr Gianni produce for them. Specifically, Dr Brownlees wanted to know biologists’ experiences with picking targets that have had very little research conducted on them.
One audience member said that “if we don’t know much about the target, then there’s going to be a cost to phenotypic validation.” They wondered what the biopharmaceutical industry’s demand for targets that had very little published about them. In Dr Brownlees’s experience, they were engaged by those novel targets in the interest of having a balanced portfolio, but that “slightly more validated” targets were most desired.
Dr Gianni agreed with the suggestion of striking the balance between novelty and well-validated targets, asserting that when dealing with an extremely novel target, “you have to ask yourself – do I understand enough about the biology of this target and its molecular mechanism to trigger the very expensive investment of hit generation, identification, and campaigns?”
Intelligent vs Random Occurrence Hits
Another audience member asked our panel whether it was worth using chance to identify targets rather than bioinformatic research: “if we look at the whole transcriptome, what is the probability that the set of molecules associated with this disease have occurred by random chance?
Dr Gianni handed this topic over to Dr Wei, asking what tools were in place that would allow researchers to identify what targets were intelligent hits rather than something that could have been discovered randomly. Dr Wei said that the data training perspective is similar to the perspectives of biologists for this question. That is, a list of potential targets can be ranked by the genetic evidence that describes their potential validity.
The same is true in transcriptomics, offered Dr Wei: if these targets are selectively expressed in any diseased T cell, they are given a higher rank. Furthermore, databases of targets can be ranked on their druggability and the small molecules that can be used to target them.
On the Genetic Link
It was posed by one audience member that although it is well understood that a genetic link to the disease is critical for target validation, can that really be enough? “Do we fool ourselves by saying we need to tick this box—we need a SNP that links the disease to the target, and then we’re good to go?”
Dr Brownlees agreed that there was some concern with wholly relying on genetic data for target validation: “we’ve got genetics in Alzheimer’s and in Parkinson’s, but it’s still difficult. We do not really understand all the biology around it, even after working on it for years.” Dr Brownlees added that it was valuable to know the genetic link, but that equally important was “understanding the pathways linked to the protein that has been mutated.”
“It’s also different depending on the indication,” added Dr Gianni, stating that the genetic link is needed very early in the project for cancer indications. This is because it provides essential information to understand which patient population to treat. Dr Gianni went on to say that he did not think it was a box ticking exercise, especially in the context of cancer treatment.
“In the oncology space,” said Dr Gianni, “trying to find that link genetically to a disease, may refer to a SNP but is more often specific mutations: gain of function, loss of function, or gene amplification.”
The question of how to reliably identify targets for drug discovery is the jumping-off point for the development of new therapeutics. Therefore, it is vital that researchers continue the conversation around navigating the intricate environment of human biology. Developments in AI and bioinformatics have been game-changing in making sense of this complex ecosystem. We look forward to seeing the world of drug discovery develop with these new technologies.
Exploring the latest advancements in phenotypic and target-based discovery, chemical biology as well as drug design at our two-day summit. Join us at Drug Discovery Europe and network with over 400 industry leaders at the renowned summit.