What is target validation?

[mbhebhe]

We are currently trying to validate our series 3 target. @danaklug has done some reading on target validation and we would love some feedback on the definition.

Our definition can be found here or just read the text below

What is target validation?

Given the continuing emergence of P. falciparum strains that are resistant to currently-used malaria therapeutics, drugs that act through novel mechanisms of action are highly desirable. A 2018 assessment of the current antimalarial pipeline suggested that phenotypic screens have failed in this regard, instead producing multiple chemotypes that act through the same mechanisms (Chaparro 2018). Target-based drug discovery programs focused on novel mechanisms have the potential to address this need. The attrition seen in target-based programs is often caused by poor target validation (Chaparro 2018, Jones 2016, Smith 2003).

What does “Target validation” mean? The definition is a challenge because the word “validation” appears to mean different things to different organizations, but in general, the validation of a potential target requires evidence that manipulation of the target can be both safe and effective. In other words, the target should be experimentally linked to disease phenotype and potential on- or off-target toxicity should be ruled out as much as possible. For an antimalarial, this includes evidence of the following:

• Modulation of target activity results in the desired phenotypic outcome: the target is essential for malarial survival; there are no compensatory mechanisms; in vivo efficacy is demonstrated in an appropriate animal model.
• Modulation of target activity does not necessarily result in on- or off-target toxicity: selectivity over host homologues can be demonstrated; in vivo safety is demonstrated.
• Target activity can be modulated by an optimizable chemical series: dose-response curves can be generated for a series of structurally similar small molecules.

There are many experimental tools which provide pieces of required evidence, including genetic manipulation, the use of chemical probes, proteomics experiments, and animal models (Smith 2003, Bunnage 2013, Smurnyy 2014, Batista 2020).

References

Batista FA et al. “New directions in antimalarial target validation.” Expert Opin. Drug Disc. 2020 (15), 189-202. DOI: 10.1080/17460441.2020.1691996
Bunnage, ME; Piatnitski Chekler, EL; and Jones, LH. “Target validation using chemical probes.” Nature Chem Biol. 2013 (9), 195-199.
Chaparro, MJ et al. “Efforts aimed to reduce attrition in antimalarial drug discovery: a systematic evaluation of the current antimalarial targets portfolio.” ACS ID 2018 (4), 568-576. DOI: 10.1021/acsinfecdis.7b00211
Jones, LH. “An industry perspective on drug target validation.” Expert Opin. Drug Disc. 2016(11), 623-625. DOI: 10.1080/1746044.2016.1182484
Smith, C. “Drug target validation: Hitting the target.” Nature 2003 (422), 341-347.
Smurny, Y et al. “DNA sequencing and CRISPR-Cas9 gene editing for target validation in mammalian cells.” Nature Chem. Biol. 2014 (10), 623-627.

[mattodd]

I asked the opinions of Julian Rayner and Pedro Beltrao. They emphasised that they are involved very early in target identification, so far from experts in drug development, but kindly shared the following thoughts (my paraphrasing from emails):

1. The definition of validation is not clear in any field, let alone malaria, and different people probably use different definitions depending on their purpose.

2. Our current definition (above) goes beyond they would have thought of as the typical one, because it includes toxicity and tractability. Other definitions would usually stop at the first point - modulation of the target (protein/gene) results in the desired phenotypic outcome - that would make something a validated target to pursue for further development (which would certainly involve toxicity and tractability).

3. The definition we link to includes in vivo validation, but In the case of malaria the role of in vivo validation is complex and somewhat disputed. Malaria animal models are of widely differing relevance to human P. falciparum infections because P. falciparum naturally only infects very few primates, so while other animal models are available (humanised mice, different Plasmodium species such as P. berghei), and are hugely useful for some purposes in particular (eg non-blood stages), they are not always thought to be on the critical path, at least at the early stages of target identification.

4. The second two points (toxicity and compound optimisation) are in our opinion further down the pipeline, after validation. These steps are about whether you can actually develop a drug for the target, or how easy it is to develop that drug, not whether the target itself is a real one.

5. The open targets portal avoids "validation" and talks about "target identification" and "target prioritization": "target identification defines targets with significant associations to the disease biology. Target prioritisation refines the list of targets on the strength of the association evidence and additional target and disease parameters such as target tractability by drug modality, and likely target-based safety risks."

Thanks Julian and Pedro, greatly appreciated.

Other people to maybe ping here: Andrew Leach at EBI and folks at the Wellcome Centre for Anti-Infectives Research.

[mattodd]

I also reached out to Darren Creek at Monash who kindly offered his thoughts on this: "I think validation is a difficult thing to distill into a single definition. Overall your definition is very good. The only thing I wouldn't agree with is regarding in vivo safety and off-target toxicity. I think you can have a validated parasite target that has a good chemical inhibitor that causes off-target tox, because if the tox is genuinely off-target you could theoretically remove that liability while maintaining target affinity. It is reasonable to retain on-target tox in the definition (and this could include cases where the binding site of a human homologue is near identical to the parasite target).

You also elude to the range of techniques available to validate a target... I often classify these into genetic validation (i.e. gene shown to be essential to parasite growth) and chemical validation (i.e. inhibitor compound shown to inhibit growth), and sometimes biochemical validation (i.e. compound inhibits the activity of pure protein and also kills cells with similar potency and SAR). A truly convincing target validation would involve all three of these, and ideally that genetic mutation of the target protein alters the potency of compounds in the whole-cell and pure-protein assays."

Many thanks to Darren for sharing thoughts on this with me by email.

[drc007]

I think the definition is pretty good, but I'd differentiate between the mechanism of toxicity. On target toxicity, or any toxicity that occurs due to modulation of the target (e.g. compensatory mechanisms) would impact target validation. Off-target toxicity that is completely unrelated to the intended mechanism would not compromise validation.

A target is still valid even if at the moment you don't have a small molecule to modulate the target, there may also be other modalities that could be used.

[bendndi]

Agree with @drc007 . Validation of a target should be completely decoupled from the tool used to do the validation - one can even validate a target without any chemical matter, e.g. via knock-out work / CRISPR-CAS9 etc. It basically answers the question: If I mess with this target will it impact the disease in a positive way?

However there is also clinical validation of a target, which comes much later when you can point to a clear example in patients that there is the desired impact on human disease via modulation of the target.