About this Author
DBL%20Hendrix%20small.png College chemistry, 1983

Derek Lowe The 2002 Model

Dbl%20new%20portrait%20B%26W.png After 10 years of blogging. . .

Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases. To contact Derek email him directly: Twitter: Dereklowe

Chemistry and Drug Data: Drugbank
Chempedia Lab
Synthetic Pages
Organic Chemistry Portal
Not Voodoo

Chemistry and Pharma Blogs:
Org Prep Daily
The Haystack
A New Merck, Reviewed
Liberal Arts Chemistry
Electron Pusher
All Things Metathesis
C&E News Blogs
Chemiotics II
Chemical Space
Noel O'Blog
In Vivo Blog
Terra Sigilatta
BBSRC/Douglas Kell
Realizations in Biostatistics
ChemSpider Blog
Organic Chem - Education & Industry
Pharma Strategy Blog
No Name No Slogan
Practical Fragments
The Curious Wavefunction
Natural Product Man
Fragment Literature
Chemistry World Blog
Synthetic Nature
Chemistry Blog
Synthesizing Ideas
Eye on FDA
Chemical Forums
Symyx Blog
Sceptical Chymist
Lamentations on Chemistry
Computational Organic Chemistry
Mining Drugs
Henry Rzepa

Science Blogs and News:
Bad Science
The Loom
Uncertain Principles
Fierce Biotech
Blogs for Industry
Omics! Omics!
Young Female Scientist
Notional Slurry
Nobel Intent
SciTech Daily
Science Blog
Gene Expression (I)
Gene Expression (II)
Adventures in Ethics and Science
Transterrestrial Musings
Slashdot Science
Cosmic Variance
Biology News Net

Medical Blogs
DB's Medical Rants
Science-Based Medicine
Respectful Insolence
Diabetes Mine

Economics and Business
Marginal Revolution
The Volokh Conspiracy
Knowledge Problem

Politics / Current Events
Virginia Postrel
Belmont Club
Mickey Kaus

Belles Lettres
Uncouth Reflections
Arts and Letters Daily
In the Pipeline: Don't miss Derek Lowe's excellent commentary on drug discovery and the pharma industry in general at In the Pipeline

In the Pipeline

« Layoffs at Human Genome Sciences | Main | The Absolute Bottom of the Publishing Barrel »

October 19, 2012

Another Controversial Scaffold?

Email This Entry

Posted by Derek

My post the other day on a very unattractive screening hit/tool compound prompted a reader to mention this paper. It's one from industry this time (AstraZeneca), and at first it looks like similarly foul chemical matter. But I think it's worth a closer look, to see how they dealt with what they'd been given by screening.
This team was looking for hits against PIM kinases, and the compound shown was a 160nM hit from high-throughput screening. That's hard to ignore, but on the other hand, it's another one of those structures that tell you that you have work to do. It's actually quite similar to the hit from the previous post - similar heterocycle, alkylidene branching to a polyphenol.

So why am I happier reading this paper than the previous one? For one, this structure does have a small leg up, because this thiazolidinedione heterocycle doesn't have a thioamide in it, and it's actually been in drugs that have been used in humans. TZDs are certainly not my first choice, but they're not at the bottom of the list, either. On the other hand, I can't think of a situation where a thioamide shouldn't set off the warning bells, and not just for a compound's chances of becoming a drug. The chances of becoming a useful tool compound are lower, too, for the same reasons (potential reactivity / lack of selectivity). Note that these compounds are fragment-sized, unlike the diepoxide we were talking about the other day, which means that they're likely to be able to fit into more binding sites.

But there's still that aromatic ring. In this case, though, the very first thing this paper says after stating that they decided to pursue this scaffold is: "We were interested to determine whether or not we could remove the phenol from the series, as phenols often give poor pharmacokinetic and drug-like properties.". And that's what they set about doing, making a whole series of substituted aryls with less troublesome groups on them. Basic amines branching off from the ortho position led to very good potency, as it turned out, and they were able to ditch the phenol/catechol functionality completely while getting well into (or below) single-digit nanomolar potency. With these compounds, they also did something else important: they tested the lead structures against a panel of over four hundred other kinases to get an idea of their selectivity. These is just the sort of treatment that I think the Tdp-1 inhibitor from the Minnesota/NIH group needs.

To be fair, that other paper did show a number of attempts to get rid of the thioamide head group (all unsuccessful), and they did try a wide range of aryl substituents (the polyphenols were by far the most potent). And it's not like the Minnesota/NIH group was trying to produce a clinical candidate; they're not a drug company. A good tool compound to figure out what selective Tdp-1 inhibition does is what they were after, and it's a worthy goal (there's a lot of unknown biology there). If that had been a drug company effort, those two SAR trends taken together would have been enough to kill the chemical series (for any use) in most departments. But even the brave groups who might want to take it further would have immediately profiled their best chemical matter in as many assays as possible. Nasty functional groups and lack of selectivity would surely have doomed the series anywhere.

And it would doom it as a tool compound as well. Tool compounds don't have to have good whole-animal PK, and they don't have to be scalable to pilot plant equipment, and they don't have to be checked for hERG and all the other in vivo tox screens. But they do have to be selective - otherwise, how do you interpret their results in an assay? The whole-cell extract work that the group reported is an important first step to address that issue, but it's just barely the beginning. And I think that sums up my thoughts when I saw the paper: if it had been titled "A Problematic Possible Tool Compound for Tdp-1", I would have applauded it for its accuracy.

The authors say that they're working on some of these exact questions, and I look forward to seeing what comes out of that work. I'd have probably liked it better if that had been part of the original manuscript, but we'll see how it goes.

Comments (19) + TrackBacks (0) | Category: Academia (vs. Industry) | Chemical News


1. Practical Fragments on October 19, 2012 9:23 AM writes...

Last year Pfizer researchers described a similar TZD inhibitor of PI3 kinases, obtained a co-crystal structure, and also showed that it is reasonably selective and cell-active. It's a good reminder that you can't necessarily discount compounds based on chemical structure alone, but if a structure is dubious one does need to take extra precautions. A bad tool compound can be more misleading than no tool at all.

Permalink to Comment

2. James M. on October 19, 2012 9:26 AM writes...

I had and enjoyed O Chem and P Chem in undergrad but I was a physics major and am now a physics grad student. I apologize for my ignorance but I was hoping someone could explain what a 160nM hit on high throughput screening is. And if single digit nanomolar potency is a good thing.

Permalink to Comment

3. Tom Womack on October 19, 2012 10:04 AM writes...

I have a very poor handle on how much the various experiments that you suggest be conducted for a proper analysis of a screening hit actually cost.

For example, how available in the academic-services world is screening against a large bank of kinases - is that a service which a good biochemistry department's support techs offer, or which you could ask a CRO to do in three months for $100,000, or which requires the resources of a major pharma company who would only do it in exchange for full rights to any arising compound?

Permalink to Comment

4. Anon Academic on October 19, 2012 10:11 AM writes...

Are there really any useful Tdp-1 inhibitors out there? [I am not familiar with the field, however, based on this review:, I would say zero.] While far from an ideal scaffold, having chemical matter for a new target is important and publishable.

Meanwhile, how many PIMK inhibitors do we have in the literature?

This blog is often decidedly anti-academic. And I totally understand why industry laughs at academics who think they are going to discover lots of drugs in academia. I'm in academia and I laugh about that too. However, there is a lot more to medicinal chemistry / chemical biology than just discovering a drug with the perfect looking structure that will satisfy all the industry experts. If you step away from that narrow view, you will see compounds like the Tdp-1 inhibitor do serve a useful purpose. Even if it's not ready for cell work, having a small molecule inhibitor can be a useful starting point for improved probes.

Permalink to Comment

5. Anon Academic on October 19, 2012 10:21 AM writes...

Tom: kinase profiling (450+ kinases) can be done for

Permalink to Comment

6. Anon Academic on October 19, 2012 10:23 AM writes...

My comment got cutoff..

anyway, kinome profiling is less than $5,000 from a number of companies (KinomeScan, Invitrogen, Reaction Biology). And they get it done in less than 2 weeks.

So it's affordable for an academic for sure.

Now whether it makes sense to profile a Tdp-1 inhibitor against a kinase panel is another question.

Permalink to Comment

7. Crimso on October 19, 2012 10:26 AM writes...

"But they do have to be selective - otherwise, how do you interpret their results in an assay?"

If you define "tool compound" as something never used to study purified enzyme, I agree. But sometimes you can learn things about your favorite enzyme by studying its interactions with other molecules, with no concern whatsoever (nor any needed) for the PK, toxicity, etc. I understand such "navel-gazing" is abhorrent to industry, but academia has different goals, attitudes, and approaches. Reading the comments to the previous related post, I started wondering whether the commenters wanted to put Pommier on a giant set of scales to see if he weighs the same as a duck.

Permalink to Comment

8. yourmama on October 19, 2012 10:38 AM writes...

that thing will have off-target PDE4 inhibitory activity.

Permalink to Comment

9. a on October 19, 2012 11:10 AM writes...

James M

This is a pharma blog; there will be serious technical discussions on it, it's not our job to educate you.

Some reading, in future, google is your friend, o graduate student!

Permalink to Comment

10. Howie on October 19, 2012 11:48 AM writes...

I agree with Crimso. I feel many posts by Derek talk about the pharmaceutical attractiveness of early "leads" in a very narrow way. It is easy to imagine missing possible discoveries by such unofounded bias of what could be become a drug, or what looks attractive to a chemist. It is better to look for pharmacological potential (how to overcome shortcomings) and let the data do the talking.

Permalink to Comment

11. Hap on October 19, 2012 12:23 PM writes...

The entire point of a tool compound is selective inhibition - inhibition of one enzyme over lots of others. That way, inhibition of your target can tell you something more than "this triepoxide that inhibits everything binds to our enzyme too", and so that you can separate the consequences of inhibiting your target from the consequences of inhibiting other enzymes.

If you have functionality that is likely to be reactive to many conditions (the triphenol lead is likely susceptible to acid, base, oxidizing conditions, and the presence of nucleophiles), then it is correspondingly less likely that it is specific for your enzyme (because it can react with lots of things), and that it is a noncovalent inhibitor (not required, but it might make it easier to control inhibition in vivo). In addition, it also makes the compound less likely to be stable on storage so that unless you're careful, you may accidentally changes in activity over time.

Leads that are unpredictably reactive are likely not to have much pharmacological potential. In addition, as one person before said, an unreliable tool compound is worse than none at all.

Permalink to Comment

12. partial agonist on October 19, 2012 12:29 PM writes...

Since if this structure came out in an academic paper, 5 people here would say "typical academic crap"

shouldn't somebody say "typical industrial crap"??

... just sayin'

I'm in academia with, as much as I can replicate on a tight budget, the same near-term mindset, attitudes, and approach as early lead pharma drug discovery, where I used to work.

I want tool compounds. I want them to be potent, target-selective, tractable, patentable, have good animal PK, be non-toxic, be scaleable. I will pay for a kinase selectivity panel if I can. I will get hERG profile and dozens of other targets. I will do SAR not just for potency but to fix a Cmax problem, or an AUC problem, or T1/2 problem, or whatever, because yes, I want them to work in animals.

We just call them tool compounds because they hit a target that no decent published molecules hit.

Long-term, yeah, I know I will need a pharma partner to get it anywhere, and I don't have such a huge team to make 20 back-ups or to get PK on everything.

It's way more likely my tool compound will just get published and lead to somebody else in industry discovering something somewhat similar that they can patent. And that's fine for science.

Permalink to Comment

13. HTSguy on October 19, 2012 1:26 PM writes...

Sorry to repeat myself (from the other string), but I still don't see the evidence that this "Tdp-1 inhibitor" works by anything other than an non-specific mechanism (which would make for a very poor tool compound). The Biacore data indicate the following:

1. Extremely slow kinetics for a 50 uM (or greater) Kd binder.
2. Multiphasic off kinetics
3. Rmax of >=200RU when 45RU is the theoretical maximum (usual real-world results are lower).

These are the characteristics of compounds that inhibit by "coating" the target protein:


Permalink to Comment

14. Pete on October 19, 2012 3:19 PM writes...

The phenol in the hit molecule can donate a hydrogen bond to the oxygen of the ethoxy group. As such it can be argued that it is less likely to be contributing to binding by donating a hydrogen bond to an acceptor in the protein. This in turn increases confidence that the phenol can be replaced.

Permalink to Comment

15. Howie on October 19, 2012 6:36 PM writes...

@Hap: I'm sure there are many definitions of what a tool compound is. If you take, as an objective, a selective tool compound, it would be good to have a compound that has activity against the target, even if it is not selective. You could take that compound and build in selectivity. From this approach, you could take even lowly staurosporine and come up with LY333531. If a molecule like this could be co-crystallized, it would also be useful as from the standpoint of FBDD.

Permalink to Comment

16. metaphysician on October 20, 2012 9:12 AM writes...


That was unnecessarily arrogant. A big part of why people come to read In The Pipeline is that Derek is very good at making complicated chemistry and biology concepts understandable to laymen. This would be why we come to read his blog, and not *yours*.

Permalink to Comment

17. Hap on October 22, 2012 8:19 AM writes...

16: Ethylene oxide could inhibit the enzyme, too - but that wouldn't tell you anything useful about how the enzyme works or what fragments might be useful to inhibit it. Compounds that inhibit an enzyme of interest nonselectively only tell you what inhibits proteins in general or just lots of other related enzymes.

One of the main things people interested in the enzyme want to know is what inhibits it, and not other enzymes - thus they need to know what interactions between an enzyme and an inhibitor are specific to the enzyme. A cocrystal of a nonselective compound with the enzyme won't tell you that - it will tell you more of how to inhibit that class of enzymes or proteins in general, which people probably knew already.

The entire nature of tool compounds is selectivity, not inhibition (well, selectivity > inhibition - if they inhibit the enzyme really poorly, they are also not useful) - they tell you 1) what kinds of structure inhibit an enzyme and not related ones, so that you can make better inhibitors, and 2) what happens to an organism or system if you inhibit an enzyme (and not any others) so that you can test whether a protein is likely to be useful as a target for drug intervention, and to determine what the enzyme does. A nonselective tool compound is an oxymoron because in both of these cases its use will mislead the people using it, and waste time, money, and lives chasing illusions.

Permalink to Comment

18. a on October 23, 2012 11:52 AM writes...

#16: OK, metaphysician, get educating and explaining!

Permalink to Comment

19. James M on October 24, 2012 9:12 AM writes...

Thank you for providing the links. I had tried using google but my google-fu that morning was very poor because I had just gotten off a double shift of running the accelerator and was quite tired. I apologize if I offended you in some fashion.

Permalink to Comment


Remember Me?


Email this entry to:

Your email address:

Message (optional):

The Last Post
The GSK Layoffs Continue, By Proxy
The Move is Nigh
Another Alzheimer's IPO
Cutbacks at C&E News
Sanofi Pays to Get Back Into Oncology
An Irresponsible Statement About Curing Cancer
Oliver Sacks on Turning Back to Chemistry