The Open Source Malaria (OSM) project has a paper out (open-access, fittingly) in ACS Central Science, and it’s an interesting read. This is the effort from Mat Todd at Sydney and many, many others around the world to build on the malaria phenotypic screening results released in 2010 (and prioritized in 2011) by GSK. They’ve been following a true open-source model – every part of the project is out there for people to see, and for anyone to join in with suggestions. That’s always going to be trickier to do than it is in the software world, where the amount of equipment needed is far lower and there are correspondingly few barriers to entry, but this work shows that (under favorable conditions) it can be done. There’s a lot of rot talked about open-source drug discovery, but this is one of the closest examples to the open-source-coding world that I’m aware of.
The paper’s actually a good read, and I think it’s the open-source nature of the work that does it. More than almost any paper I’ve ever read, it goes into the ins and outs of the work on the various lead series. It’s actually a very accurate look into how these things work – synthetic difficulties, re-routing, structure-activity surprises, arguments about which structures are worth moving on with, tradeoffs at every turn. This is exactly how medicinal chemistry works, and in the future I’ll be recommending this paper to show people a blow-by-blow account of it.
In the end, the arylpyrrole series being investigated here looks like it will be put to one side in favor of pursuing other leads. The OSM-S-5 and OSM-S-6 compounds shown (5 has an H at R1, 6 has a heterocycle) were hits from the GSK work. The problem with them is that ester functionality buried in the middle of the structure, which was expected to be labile (experiments to determine blood levels in mice and followup mouse plasma stability assays were consistent with this worry). Unfortunately, changing that to an amide, any sort of amide, completely killed the biological activity. While the team was working on these, they were also looking at the structurally related TCMDC hits, also from the GSK data set. These had more robust SAR (albeit with potency coming along with greater lipophilicity), but the team seems to have had quite a debate about whether these structures had a future or not. They are indeed getting into PAINs territory, and were also filtered out by the ALARM NMR assay, which is food for thought. In the end, since the compounds were heading into even more unfavorable property space along with these potential liabilities, they were deprioritized.
A range of other arylpyrrole structures were then investigated (see the paper for details). The team had a big meeting online, with a prioritized “hit list” of desired structures emerging from it, and in the end about half of these got made (or purchased, in a few cases, when something was available). As the paper shows, though, none of the modifications were tolerated. One series was unstable on storage, and both it and the other modifications were much less active (or completely dead) in the assays. There are, though, some promising ideas that haven’t been investigated yet, such as replacement of that ester with a suitable oxadiazole. That’s an isosteric substitution (along with some other heterocycles) that has worked in many previous med-chem efforts, and it’s probably the outstanding thing on the list. And since this is an open-source effort, anyone who wants to try to make it (or any other analog in this area) is welcome to do so.
One of the unique features of this project, the open source research method, ensures that the unexplored lines of inquiry remain open alongside the attendant data posted online that makes it straightforward for others to resume any portion of the research project as fully fledged participants, with access to both positive and negative data, details of all procedures as they were carried out (to aid reproducibility), and anecdotal insight into project loose ends that are easy to explore. The machine-readability of the present project (for example the use of cheminformatic strings in the online electronic lab notebook) permits an unusually straightforward link between a high throughput screening result in a public database and a “live” research project that has investigated that compound.
The original GSK assay was phenotypic, so it’s still an open question how any of these compounds might work. The team has done a number of assays against different stages of the malaria parasite to try to narrow things down, but the target (or targets) of these compounds remains unknown. That brings up the traditional phenotypic assay project decision – press on with what you have, in the knowledge that an in vivo assay will probably wipe out most of your compounds (but what’s left may be gold), and you may well stumble around a lot, or take the time to identify the target and run screens against it. In that case, you may find yourself working on things that will be harder to develop in vivo (although not necessarily), but you’re also working with a lot more tools to try to get a compound series going. Neither of these is a slam dunk, and if you have the resources, there’s no reason not to try going on with both, if possible, since they don’t rule each other out. The team has done some work into mechanisms of action, but there’s a lot more to do if something more interesting comes up.
This, then, ends up being a pretty typical medicinal chemistry project – if you didn’t know that it was being done by an assortment of different people all over the world, you wouldn’t be able to guess it. I look forward to seeing where things go from here, and what tools the project puts together to keep things moving.