opensourceantibiotics / Series-2-Diarylimidazoles

Open Source Antibiotics Series 2
10 stars 1 forks source link

GSK Involvement #4

Open danaklug opened 4 years ago

danaklug commented 4 years ago

It would be great to get GSK involved in this project at some level, since they have published structures closely related to the hit compound SB-400868 (see Surrounding Chemical Space).

I'm not sure the best way to go about contacting them about this - any suggestions?

drc007 commented 4 years ago

Perhaps contact the authors of the publication.

mattodd commented 4 years ago

This is an important point - we need to be sure that we have access to all the available data and all the available samples - at least, whatever can be shared, and whatever people are willing to share. I wonder if the UNC people have links to the original team/samples (the ones Dana linked to here).

@TwistedSwisster can I please loop you in here? Do you have a sense of whether we can access this through you guys, or should we be reaching out to someone currently at GSK to start this conversation? It'd be great to involve GSK on an ongoing basis, but we should take this one step at a time.

mattodd commented 3 years ago

I received by email a few days ago a final bit of text from James Callahan at GSK who has received clearance for me to share the following. James kindly also reached out to their anti-bac group, who have our details and are aware of the project and may reach out to us.

Summary of a discussion held on 31 July 2020 with Professor Matthew H. Todd, Chair of Drug Discovery, University College London to be shared with an open source community designing inhibitors of MRSA.

I spoke with James Callahan, a GSK scientist based in the US who was involved in the project that developed these molecules for Alk-5 inhibition. He kindly gave me the story here and has agreed that I can share the following.

James lead the early medicinal chemistry effort on the project at SmithKline Beecham. The medicinal chemistry leadership was later passed on to John Harling at SmithKline Beecham-UK (John is currently at GSK Stevenage). The structures feature a bicyclic imidazole (Callahan, et al., J. Med. Chem. 2002, 45, 999-1001) that raised some concerns of a potential metabolism/toxicology risk at the time based on some previous experience with the core template. The team were trying to identify a selective Alk5 kinase inhibitor. They moved away from this core structure (the one we're interested in) quite early in the project, so there are not a lot of data on things like metabolic clearance, etc. Also we need to bear in mind that the team members were attempting to optimize the compounds for Alk5 kinase activity and selectivity (the 2-pyridine substitution provided selectivity over p38 and other kinases), not MRSA activity, so some of the SAR learnings are not relevant to us.

GSK (Glaxo back then) pursued a somewhat similar biaryl pyrazole core (Gellibert, et al. J. Med. Chem. 2004, 47, 4494-4506) (we need to add this to our "similarity landscape" wiki page). The dihydro-pyrrolo imidazole core compounds in the GSK collection probably came from early p38 chemistry and are analogous to an early lead in that program, SK&F-86002 (Gallagher, et al. Bioorg. Med. Chem. 1997, 5, 49-64) (https://www.sigmaaldrich.com/catalog/product/sigma/s0193?lang=en&region=GB).

The GSK team played with the position of the nitrogen in the pyridine ring, again optimizing kinase inhibition and more importantly selectivity for Alk5. Playing with these isomers (and things like a pyrimidine ring) strongly influenced kinase activity and selectivity. The methylene dioxo-phenyl (benzo[d][1,3]dioxole) is a known metabolic risk we should be avoiding. James asked a good question about whether we have data on these compounds in a broad kinase screen - we need to dig this up, since clearly, we want to optimize away from the activity that the GSK group were trying to build in.

Action here is to follow up on this kinase screen idea.

danaklug commented 3 years ago

Regarding the kinase activity of SB-400868: This compound has a CHEMBL Compound Report Card which I used to create an easily-readable Google sheet containing kinase inhibition data. The only kinase showing >30% inhibition at 1 uM is CK1a.

drc007 commented 3 years ago

@danaklug Link to report card is broken. https://www.ebi.ac.uk/chembl/compound_report_card/CHEMBL1909377/

Also looks like this is the same compound

https://www.ebi.ac.uk/chembl/compound_report_card/CHEMBL268889/

Inhibition of TGFR1 IC50 = 460 nM

mattodd commented 3 years ago

Just trying to tie up loose ends on the kinase activity. The two ChEMBL compounds above are just amine and HCl salt, and the only kinases that the compound hits appear to be CK1a and TGFR1 (wasn't sure if your results, @drc007, implied TGFR1 was the only hit on ChEMBL)? This seems like a low level of worry (for now) but one that we might want to keep an eye on should potency improve (i.e. no kinase screening follow-up needed for now). If this is the case I can close this issue after incorporating the page in brief into the wiki. We're always glad to hear from the GSK team further, but I don't think we're missing out on a trove of data to which we don't have access.