Hi,
This looks like a promising start for discussing TPPs.
If I could perhaps make a suggestion it would be to split these into TPPs (Target Product Profiles) and
TCPs (Target Candidate Profiles).
A TPP is clinician / patient focused - dosing regimen, patient population etc
A TCP is focused on drug discovery scientists - Target IC50s in assays, off target tox goals etc… all the good stuff we care about right now.
The TPP(s) come first (and the above doc covers this really well), and the TCPs are then created, informed by the TPP. There can be more TCPs than TPPs
It may seem like semantics but it’s really helpful to think in these different stages - to put it bluntly the clinicians don’t care what the IC50 of a drug is, so the TPP is better off without it - keep it as simple as possible.
There is a good publication which covers this TPP / TCP concept from an infectious disease perspective from GHIT, Gates Foundation, MMV, TB alliance, DNDI:
Katsuno et al, Nature Reviews Drug Discovery 2015, 14, 751
Also a good example of how MMV use this TCP vs TPP differentiation:
https://www.mmv.org/research-development/information-scientists/target-product-profiles-target-candidate-profiles
(Just my suggestion of course - the doc above is already a great starting point)
Great comments Ben, thank you for your suggestions. My thoughts in creating these were to just get us off the blocks with some simple 1 page per indication documents that connect the clinical position to the actions for the discovery team. The critical early discussions were around iv/oral/inhaled and then using these to inform a first test cascade (which is also about to go on the Forum, with even more discussion).
Hi Ed,
I agree that IV infusion would be the best option for critical care in hospital since this allows you to control plasma concentrations (good from both efficacy and toxicity perspectives). It may also be better to attack plaque from the ‘inside’ rather than the ‘outside’ (as would be the case with an inhaled drug) and it would be useful to get views from any experimental medicine folk to whom you have access.
I think reversibility needs to be addressed explicitly in the target product profiles (I see mention of “reversible and monitorable toxicities” in the safety section and I take this to mean that the toxicities need to be reversible rather than the enzyme inhibition). I have prepared some notes on reversibility in the context of design that may be helpful:
I do not believe that there will be any therapeutic advantages to using an irreversible inhibitor when dosing by IV infusion. The situation is less clear for oral dosing although the enzyme inhibition assay for irreversible inhibitors is more complex than for reversible inhibitors and assay throughput is likely to be lower.
Why is the solubility criterion (>10mg/ml at pH >5) for IV infusion not set at plasma pH? It may be useful to also set criteria for permeability and efflux ratio.
You may wish to specifically consider human cysteine proteases when setting safety criteria and it would be useful to get the views of people in industry working on these targets to see what should be avoided. Lysosomal accumulation can be an issue and if objectives can be achieved without basic centres (it also hERGs less) then there may be advantages to avoiding them.
Hope these comments are helpful.
Hi Pete, thanks for all the input, a couple of straight answers:
-
“reversible and monitorable toxicities” is the classic toxicologists phrase to mean exactly that - if the toxicities are irreversible or unmonitorable then it’s very hard to get into Ph I.
-
cysteine protease selectivity - agreed, we should add that to the testing cascade, but I wouldn’t look at it till there was sub 100nM potency in an enzyme assay
-
the solubility criteria is set from the reasonable limit of what is possible for an IV infusion, there are a few drugs that are infused at pH below 4.5, but you get site of injection issues. I have back up data if you want to see it.
-
I can see the mode of inhibition (reversible / irreversible) being in a TCP rather than a TPP as discussed by Ben Parry above - I’m just giving it some time for people to digest the TPPs (and I do have to work on the day job too…).
Agree with above comments, and with points Peter makes in his “covalent inhibitor” document. Adding to comments on covalent inhibitors:
-
Regarding “protease assay: IC50< 10 nM”: with irreversible covalent binders, the longer you pre-incubate, the lower the IC50 (so <10 nM will be easy
). With slowly-reversible covalent inhibitors, IC50 will asymptotically improve to a minimum once equilibrium is achieved; sometimes this is fast and sometimes it is not (warhead dependent) so you have to check. So for irreversible covalents, IC50s from an assay with x pre-incubation is not great for driving lead opt unless “all others things are equal” meaning within a series with similar exposure in vivo. You still need the assay, but caveat emptor with irreversibles. With slowly reversibles, once you know your pre-incubation is enough to get in the asymptotic regime for the IC50, you have a good number for driving lead opt although if it is really slow-on you may still have issues with exposure of course. -
I belong to the camp which is reluctant about irreversibles and in favor of reversible covalent inhibitors for reasons similar to those in Peter Kenny’s “Reversability…” white paper above. Additionally, for TPP-2 and TPP-3, a pan-specific coronavirus inhibitor (which we want and is higher probability with Mpro inhibitors) implies chronic (or at least episodic) dosing for health care workers, in which case the much higher risk of immunogenicity in the case of irreversibles is a correspondingly much higher concern. For critical care dosing (TPP-1) not so much. Finally, for irreversibles the accumulation of bound inhibitor in vivo may be required to be measured in a radioactivity assay.
-
Because of the high similarity of SARS-CoV-2 Mpro with SARS Mpro, for “mutant cover” the resistance mutant prior art for SARS Mpro, especially mutants proximal to the active site, should be plug-and-play applicable to SARS-CoV-2 to jump start the process. That said, I am not up to speed on that prior art or even if there is any. Maybe for GC376.
-
Still on the subject of “mutant cover”, I suppose a virus passaging assay could be used early on to bring resistance mutants to light. Didn’t see it in the TPP.
Hi Ed,
I’ll defer detailed TPP/TCP comments until separate TPP and TCP drafts are available. Even before that, I think the reversibility issue is one that the team needs to address. Chris Bayly and I have presented similar arguments for focusing on reversible inhibitors and it would be useful to get the views of the team on this point. If the team would prefer to focus on irreversible inhibitors then let’s at least debate the point.
Regarding the target product profile and bearing in mind the importance of resource-poor settings, perhaps it is worth adding something about cost and storage stability?
A rational for both covalent and noncovalent inhibitors.
The different mechanisms give diversity in strengths and weaknesses and it is not possible to know in advance which is the best approach. Irreversible inhibition may be necessary for efficacy or noncovalent inhibition may be required for safety.
Irreversible compounds have potential advantages in the clinic, including increased efficacy, extended dose interval and decreased off target toxicity (Swinney, D, 2009, Curr Op Drug Disc Dev 12, 31-39). These potential benefits are not always observed in practice and there may be increased on-target toxicity or adverse immune reactions. In a 2005 survey of FDA-approved medicines, around 35% of enzyme targets were irreversibly inhibited (Robertson, JG, 2005, Biochemistry 44, 5561). Progression from noncovalent to covalent inhibition has proved important for efficacy of kinase inhibitors to treat cancer.
Reversible inhibitors usually are easier to optimise, but in terms of characterisation against isolated enzyme, there may not be a great deal of additional work required for covalent inhibitors. Both reversible and irreversible compounds require dose-response studies, respectively to determine IC50 or apparent kinact and Ki values. For irreversible compounds a time-course is required, but this also is strongly recommended for reversible compounds to ensure the assay is at steady state. Some potent reversible compounds are slow binders and require time-courses. In any case, most proteases are characterised in continuous assays, which generate time-course data. An additional assay required for covalent inhibitors targetting Cys residues is stability in glutathione (GSH), but this requires only a single concentration of test compound. It is desirable to measure rate of regain of activity for selected irreversible compounds and reversible slow binders. When targetting reversible inhibition, it is prudent to test selected examples for regain of activity.
An excellent point. This is one of the elements being addressed by COVID-19 Clinical Research Coalition set up by DNDi, FIND, IDDO and a host of other global partners. For more detailed info see this recent comment in the Lancet.
Whether one incudes such aspects in the TPP or have separate TPP for resource-poor settings is something to be discussed. Personally I’d include it in a global TPP but have it as a “Ideal” property but not necessarily a “minimum” property.
Thanks @Wal-Ward, really interesting comments, and good point about the need to consider resource-poor settings in advance. We have been trying to move so fast on the science that is a very good idea to step back and consider the broader issues for a bit.
I am also going to just note that @pwkenny has previously had some interesting thoughts of the covalent / non-covalent issue in this thread A brief exploration of past SARS small-molecule inhibitors . I think your insights are a nice complement to that discussion.
I would certainly agree with @Wal-Ward that irreversible inhibitors can have advantages over reversible inhibitors and that one needs to check reversibility even for inhibitors that are believed to be reversible. I would argue that reversible inhibition would be preferred for IV infusion since dosing is continuous (no dosing interval) and slow binding to target may have implications for efficacy while complicating the setting of loading and maintenance doses. The vinyl sulfone K777 (an irreversible inhibitor) showed efficacy in animal models for Chagas disease and schistosomiasis. I believe that they were lining it up for clinical trials and I’m not sure exactly what happened (I’ve heard both tox and that they ran out of funding).