Biological Results for First Set of Compounds

Published by MatTodd on 11 January 2012 - 12:12am

The screening data from three separate labs have been obtained for the first set of compounds on the project. Data were obtained from the Ralph Lab at the University of Melbourne, and a second data set was provided just before Christmas by the Avery Lab at Griffith University. Yesterday the third set was provided by GSK Tres Cantos in Spain, who originally discovered the hits we're starting with. The current list of available compounds in this open project is here, with those that have been evaluated by at least one lab indicated in the relevant column.

Having data on the same compounds from three labs using different screening methods is useful as it provides contrasting ways of assaying effectiveness. In any given screening experiment on this project it's going to be important to include known actives, so that we have benchmarks, and this was done in these cases. It's also very important to be 100% sure about the effectiveness of a compound before we become too attached to it...

The data (below, but all available through the relevant lab book) show that the original TCAMS compounds are certainly active, though perhaps not quite as active as suggested by the original screen. Paul Willis at MMV had suggested we also check out some "near neighbors" of these compounds that were in the original data set. We made a couple and one (a novel compound with the code PMY 14-1, shown below and synthesized here) has shown promising activity in all three screens, with Avery/GSK IC50s coming back as low nanomolar. (Note that this project will never involve patents or closed data, giving us the freedom to discuss the compounds freely.)

What's next? In the short term: We're waiting for confirmation of the Melbourne data via a re-run of some of the experiments. But what we need is an expert qualitative assessment of these bioactivity data by someone familiar with such screening assays. Either in comments below this post, or on G+, not by email. First item of business in the lab is to generate a few variants of PMY 14-1. We already have some new relevant compounds and are now planning others. What should we make - i.e. how ought we to change PMY 14-1? Sanjay Batra has students who are about to make steric variations in the aryl pyrrole, and these could then be employed in the synthesis of PMY 14-1 variants, for example, but shouldn't we also be interested in changes in the "upper half" of the molecule?

In the long term: It would be good to find other labs which already have analogous compounds to the actives. Paul and Zoe found a paper from the Roberts lab at Scripps describing a number of such compounds, and I will write to them to ask whether they are interested in having the compounds be screened for their antimalarial activity. If anyone knows of any other possible sources, that would be great, since using existing compounds saves a lot of time in the lab.



Great to see that that several analogues have been prepared and screened congratulations As one would expect, there is good agreement between the Avery and GSK data it would be helpful to detail the endpoints in each assay (apologies if I missed this data) It is encouraging that both PMY14-1 and PMY10-2 have low cytotoxicities (HEK 293) and similar potencies on the 3D7 and K1 (Chloroquine resistant) strains It would now be useful to further characterize these leads: solubility, in vitro metabolic stability (including GSH trapping?), lipophilicity, and plasma stability would be great. MMV can help generate some of this data is anyone else interested in contributing?

MatTodd's picture

(Further suggestions from Paul Willis at MMV, since our image upload is a little tortuous...)

·       Although PMY10-2 is less potent than PMY14-1, it can be considered a better lead due to its superior ligand-lipophilicity efficiency (4.7 vs 2.5) and ligand efficiency (0.41 vs 0.3)
·       A key liability with PYM10-2 is likely the ester linkage.  I would explore some simple ester replacements / isosteres, for example (main drive increase potency and stability)

MatTodd's picture

Paul - interesting. Those ligand efficiency values calculated by you or from Chembl? What's the sweet spot you're looking for?
Paul Ylioja is already on the case of the amide and ether-linked varieties - spreadsheet of currently-done compounds here. The amide bond is a pain to make for some reason, and one of the attractions of the near-neighbours is the requirement for a double bond there instead! Paul's going to summarize where he's at in the next day or so. The heterocycle isosteres should be simple from the carboxylic acid I'm guessing. Sanjay Batra's student is about to make variations in the fluorine that could be contributed to that series.
A question about the biological data: these arise from three different assay types. Is that a problem or a strength? In other words, is there a need for consistency?
I will contact Sue Charman at the outset to ask about ADME, unless a reader would like some of these compounds to play with.

Hi Matthew,

It can be useful to deconstruct your hit molecule before optimising.

Why is PMY 14-1 more active than 10-2 - increased lipophilicity (N-Ph); extra HBA (acetyl); more rigid; determine E or Z (is this functionalality a michael acceptor?); more optimal position of HBA (the one closest to the pyrrole)?
Likewsie for 10-2 - amine or amide; NH2, NHR or NRR; ester, ether, amide or carbon linker; is ester HBA in optimal positoin for interactions - determine before making isosteres (extend and reverse); what activity value is the pyrrole N-aryl group and it's substituent?
If you have acces to molecular modeling, an overlay of 14-1 and 10-2 might be useful. If not, ball and stick models can be good enough to generate ideas.


 Hi Martine, Thanks for the input. These are exactly the kind of questions we're after. We're currently running some DFT of PMY 14-1 and PMY 14-3 to get an idea about how they look and behave. We've synthesised a good range of N-aryl analogues for the 14-1 set and some variety on the N-Acyl position. These are going to be sent for testing very soon, at the end of next week hopefully. I discuss some of these ideas here. If you have access to modelling software then it would be fantastic if you would like to contribute by investigating the pharmacophore? The next set of data should be available in the not too distant future but we could already start using the small set of biological data we have and the much larger published GSK data set. 

MatTodd's picture

Given that the first round of compounds have been shown to have no significant cytotoxicity against HEK293 and have equivalent activity for 3D7 and K1 parasite strains do we only need 3D7 data for the second round of testing, given that the compounds are analogous to the first set?

MatTodd's picture

In response to the above question, some replies came in via email overnight:
Stuart Ralph, Uni Melbourne: we would go for a drug resistant strain like w2mef first rather than 3d7. However finding a differential sensitivity between a drug sensitive and resistant line might be illuminating later on for compounds of particular interest.
Javier Gamo-Benito from GSK: I´d suggest to maintain 3D7 as standard to construct and guide SAR. To test a different genetic background harboring resistances could be a good approach to be sure we are not getting intro troubles......for that W2 could be a good option.
Paul Willis from MMV: I agree with Javier: use 3D7 as our standard screen for SAR and once in the next few screening check a few representative compounds on a resistant strain: Dd2 or W2. Note in the initial HTS there was no difference in the % inhibition on 3D7 and Dd2, so there is flag of an issue.
So the consensus would seem to be to test this second batch of compounds on 3D7 only. Depending on the data, one or two might subsequently be tested on a resistant strain (with, presumably, a re-check for cytotoxicity).

It's great to see the interest in this area. One thing about PMY 14-1 druggability comments - what hasn't been raised is the potentially problematic nature of the groups therein. One never wants to become too jaded, especially in a wonderful endeavour such as this, but I'd just like to caution people on the potentially problematic nature of these groups. In fact, we have removed compounds that have similarities to both the top and or the bottom half, from recent HTS libraries because they light up our screens so often. Of course, not only may there be exceptions but also our focus is target-based HTS (this is changing though), and in phenotypic HTS (like here, just cidal readout) one obviously can view what may be biochemical probes differently. Without wanting to push my own publication, see Fut. Med. Chem. 2010 (2) 1529-1546. I should add that pulling electrons out of the pyrrole as in PMY 14-1 probably helps towards any tendency for retrosynthetic cyclocondensation and the imino thiazolidinone is not the hottest type of PAIN we see...i.e there are quite a few benign ones so in this context may be ok.

Just be cautious with these types of chemistries. Maybe getting access to the (blinded) HTS history for similar compounds to PMY 14-1 at GSK would be useful?

On the ester, it might be worth looking into being more it actually prone to esterases? It is in the 3 position of pyrrole, an e-rich aryl, and with a 2-methyl so may be more stable than anticipated?

MatTodd's picture

Jonathan - Thanks. This is an exceptionally important point. You know how much of a fan I am of your original paper. The importance of the work is to alert the community to promiscuous compounds and be more sceptical of molecules containing certain groups. Two questions:
1) I had thought that the problem was with certain HTS assays - we had three different low-throughput assays applied to these compounds, so the warning light hadn't gone off in my mind. Perhaps I misremembered that part of your talk/paper? Would the same caveats apply to the screens done to date on these compounds (as opposed to the original GSK screen)?
2) To evaluate directly whether this, or other compounds, are PAINS, what's the most direct test? Is there one?
We're about to send off round 2 for screening, which may throw some more light on this. I'll alert the screening teams to this post, though.

1) It is true that our assay conditions are particularly suited to pulling out inherently reactive compounds that may not light up other assays because we test at fairly high 25uM. But generally, we would say if our PAINS filters recognize a compound, deprioritize if a target-based screening hit and view cautiously even if phenotypic readout. But I don't want to become too dogmatic until we understand more on MOA of these.
2) Abbott's ALARM NMR is the industry standard to pick up protein reactives. I can put in a request to Andrew Petros there if you like to look at this cpd. A lot of places will now be setting up assays for cpds that may turn out to be protein-reactive intracellularly (but not necessarily in a test-tube).
With respect to medchem optimization of the the TCAMS pyrrole, I have emailed something that maybe you can upload on this site [Done below - Mat], which is how we would go about it. The message: optimize within your core first before extending. I'd be a little worried with your PMY cpd you may have moved into testing very different hypotheses.

Looking at the near neighbour compounds (PMY 14-1 analogues) in ChEMBL, they have IFIs of 0%, 4.7%, 3.65% and 6.67%, so while they have all clearly passed the thresholds they do have some propensity to hit different screens. Between the fact that GSK did take appropriate measures (the IFIs plus testing for assay interference, as well as doing cell based toxicity screens), and the fact that these are cell based assays, I'm not sure how we would see this sort of profile if the activity was completely non-specific - but I have had difficulty finding the question of false positives in cell based HTS assays vs binding type assays addressed directly in the literature, so I'm not sure how big a factor this is.
Regardless, if it's possible for us to get some of our compounds tested by ALARM-NMR, I think that would be fantastic as it will give us direct evidence of their reactivity towards proteins, and as I understand it can also provide an indication if activity is the result of a reactive contaminant.

Hi Jonathan,
Looking at the original GSK paper, the GSK hits were assigned an "Inhibition Frequency Index" (IFI) to exclude promiscuous compounds which was basically a percentage of the HTS screens each compound had been used in where they showed 50% or more inhibition. They used a variable threshold, from 5 to 25% of screens, depending on the number of HTS screens a given compound had been through.
Is this something you would see as an adequate measure? Our library doesn't give me access to your article in future med chem, but I had a look at your 2010 paper in J. med chem, and as I understood it you were taking something like activity in 4 out of 6 HTSs with one or two higher values as indicators of PAINS, so at face value the GSK thresholds would seem fairly reasonable?
Thanks very much,

Well answered Zoe. Yes that seems reasonable. One thing I would recommend with a pyrrole is to always retest on freshly HPLC-purified material, because an HTS history may vary with its age. That may have been done already for the original TCAMS you know if it has?

I checked ChEMBL and both TCMDC-123974 and TCMDC-123812 have a reported inhibition frequency index of 0 so there are no frequent hitter flags from the original screen - we could ask if GSK will reveal how many screens these compounds have now been tested in. This is an importnat issue: perhaps some groups will offer to screen the hits in their favourite assays so we can use the community to confirm these are not promiscious hitters. In addition, for information, the compounds have been resynthesised and the potencies confirmed in multiple assays using fresh samples.