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

« IBM And The Limits of Transferable Tech Expertise | Main | A New Book on Longevity Research »

April 24, 2013

Watching PARP1 Inhibitors Fail To Work, Cell By Cell

Email This Entry

Posted by Derek

Here's something that's been sort of a dream of medicinal chemists and pharmacologists, and now can begin to be realized: single-cell pharmacokinetics. For those outside the field, you should know that we spend a lot of time on our drug candidates, evaluating whether they're actually getting to where we want them to. And there's a lot to unpack in that statement: the compound (if it's an oral dose) has to get out of the gut and into the bloodstream, survive the versatile shredding machine of the liver (which is where all the blood from from the gut goes first), and get out into the general circulation.

But all destinations are not equal. Tissues with greater blood flow are always going to see more of any compound, for starters. Compounds can (and often do) stick to various blood components preferentially (albumin, red blood cells themselves, etc.), and ride around that way, which can be beneficial, problematic, or a complete non-issue, depending on how the med-chem gods feel about you that week. The brain is famously protected from the riff-raff in the blood supply, so if you want to get into the CNS, you have more to think about. If your compound is rather greasy, it may find other things it likes to stick to rather than hang around in solution anywhere.

And we haven't even talked about the cellular level yet. Is your target on the outside of the cells, or do you have to get in? If you do, you might find your compounds being pumped right back out. There are ongoing nasty arguments about compounds being pumped in in the first place, too, as opposed to just soaking through the membranes. The inside of a cell is a strange place, too, once you're there. The various organelles and structures all have their own affinities for different sorts of compounds, and if you need to get into the mitochondria or the nucleus, you've got another membrane barrier to cross.
At this point, things really start to get fuzzy. It's only been in recent years that it's been possible to follow the traffic of individual species inside a cell, and it's still not trivial, by any means. Some of the techniques used to do it (fluorescent tags of various kinds) also can disturb the very systems you're trying to study. This latest paper uses such a fluorescent label, so you have to keep that in mind, but it's still quite impressive. The authors took a poly(ADP) ribose polymerase 1 (PARP1) inhibitor (part of a class that has had all kinds of trouble in the clinic, despite a lot of biological rationale), attached a fluorescent tag, and watched in real time as it coursed through the vasculature of a tumor (on a time scale of seconds), soaked out into the intracellular space (minutes), and was taken up into the cells themselves (within an hour). Looking more deeply, they could see the compound accumulating in the nucleus (where PARP1 is located), so all indications are that it really does reach its target, and in sufficient amounts to have an effect.

But since it doesn't, there must be something about PARP1 and tumor biology that we're not quite grasping. Inhibiting DNA repair by this mechanism doesn't seem to be the death blow that we'd hoped for, but we now know that that's the place to figure out the failure of these inhibitors. Blaming some problems of delivery and distribution won't cut it.

Comments (24) + TrackBacks (0) | Category: Cancer | Pharmacokinetics


1. HTSguy on April 24, 2013 10:31 AM writes...

I am not a chemist and am curious of the reactions of the med chemists out there. The probe molecule is much larger than the original compound and is over 40% (by heavy atom count) "label". Is anybody else concerned about the generalizability of this (even back to the original compound)?

Permalink to Comment

2. Chrispy on April 24, 2013 11:05 AM writes...

I would just like to take a moment to complain about Nature Publishing. Despite the many thousands of dollars that my large University ponies up for their journals, they want to charge me $32 to read this article.

I will be glad when their dinosaur publishing model is dead. It is an impediment to science.

Permalink to Comment

3. barry on April 24, 2013 11:09 AM writes...

HTSguy is right on. A fluorescent tag can often be added to a protein without meaningfully changing its PK properties. But for a small molecule, that seems a terrible wager. Swap proton for tritium or 12C for 14C and I agree that you're unlikely to change transport and/or binding meaningfully (although of course primary isotope effects can still change metabolism).

But to change the probe molecules MW and logP and then assume that it behaves like the small PPAR modulator would take a leap of faith.

Permalink to Comment

4. A Non Mousse on April 24, 2013 11:10 AM writes...

Chrispy: I share your chagrin at the high cost of journal articles, but if your university is shelling out thousands of dollars for them shouldn't you be able to access the article for free?

Permalink to Comment

5. anon2 on April 24, 2013 11:16 AM writes...

Regarding comments by #3.....sounds analogous to the Heizenberg uncertainty principle in that if you can see it, you don't really know what it is and what it may do biologically, and if you can't see if, well then you don't know where it is.

Permalink to Comment

6. RB Woodweird on April 24, 2013 11:21 AM writes...

Can anyone tell me why in the HPLC analysis of the dye-labeled drug they used an ELS detector instead of UV/Vis?

Permalink to Comment

7. newnickname on April 24, 2013 12:11 PM writes...

@6, RBW: First guess: Convenience? ELS was already in-line, so why switch?

@3, barry: I agree -- and like -- the 14C and 3H suggestions. The paper says that radiography doesn't have the resolution of fluorescence. Any new methods out there for better localization of emitters? Fix the cells, infuse scintillant and do single photon counting?

Supp Info has a movie! I queued it up but by the time I got back with my popcorn, it was over.

Permalink to Comment

8. paperclip on April 24, 2013 1:18 PM writes...

I totally agree with the above commenters raising the point about the fluorescent tag affecting cell uptake. (I haven't read this particular paper yet, so just speaking generally here.) Sometimes a compound tagged with a particular dye will get in, but the same compound tagged with another dye will be shut out. The happiest scenario is when the compound is naturally fluorescent, so no tag is needed. Barring that, it should be checked that the tagged inhibitor (or whatever it is) still acts as an inhibitor (or whatever it does). Too many papers skip this part.


Permalink to Comment

9. isodope on April 24, 2013 1:20 PM writes...

@3 and @7: 3H tritium is the only isotope that is useful. the technique is called microscopic autoradiography. 3H is useful because of the short path to extinction of the low energy beta particle. this allows very high (subcellular) resolution with the proper x-ray film.

Permalink to Comment

10. paperclip on April 24, 2013 1:25 PM writes...

Oops that should be @7 Not an expert but I think you can separate the nucleus from the cytoplasm and put both in a scintillation counter?

Permalink to Comment

11. Sam on April 24, 2013 2:16 PM writes...

Yeah... so if you see the same inhibition of the target with or without the tag, then you can be pretty sure the tag isn't interfering *too* much. I don't know if they did that comparison using an assay that could really distinguish.

For now, fluorescent tags are the way to go. Radioactive stuff will be great when we have single-cell MRI types of imaging, but it's not there yet. Time resolution is key, and we can't do that with tritium. Seconds and minutes are speed of interesting reactions in cells. It's gone by the time you could get x-ray film on there.

FWIW, re: whether inhibiting DNA repair should work or not, IMO it has to do with checkpoints. Inhibiting repair alone doesn't matter if the cell doesn't signal death in response to a block.

And DNA repair is notorious for having backup mechanisms.

And the role of PARP in DNA repair is hardly well understood, actually. NHEJ is not as simple as it sounds.

Also, this question about the disconnect between in vitro (cells) and in vivo (animals/patients) is an important one... Rapidly dividing cells in a dish have a lot fewer support systems, so they die pretty easily if you inhibit DNA repair alone, just from the troubles they encounter with normal DNA replication & cell division (& oxidation, etc). But inside an animal, the surrounding cells provide a lot of support to the tumor, so it might take more than one mechanism to convince the tumor cells to die. Like say inhibiting PARP and some kind of checkpoint signaling kinase (I don't know, maybe something obvious like ATM? Hasn't anyone tried that?).

Permalink to Comment

12. B on April 24, 2013 2:26 PM writes...

@6: Potential photobleaching or UV mediated destruction of the fluorescent tag seems like a possibility, no?

Permalink to Comment

13. RKN on April 24, 2013 3:09 PM writes...


If I'm not mistaken the intention with PARP1 inhibitors is to take advantage of synthetic lethality. When a patient has certain BRCA1 mutations that prevent the oncoprotein from functioning in the homologous recombination (HR) pathway, DNA repair is shunted to the base excision repair (BER) pathway, the activity of which requires PARP1. So giving a PARP1 inhibitor to patients with these BRCA1 mutations should result in both HR and BER being "turned off," and hence signal cell death.

Permalink to Comment

14. Puff the Mutant Dragon on April 24, 2013 3:11 PM writes...

I totally agree with #1 and #3. I feel like it's unreasonable to assume your results reflect what's going on with the original compound when you've added this big ol' fluorescent label that changes the logP and could affect transport in ways you know nothing about. I guess the question is how else you do this if you don't use fluorescent labels though.

Permalink to Comment

15. dearieme on April 24, 2013 4:20 PM writes...

At the risk of sounding terribly pious, may I just thank Derek for the clarity of his exposition, and you commenters for the calm rationality of what you have to say?

Permalink to Comment

16. barry on April 24, 2013 4:22 PM writes...

re: Sam (#11)
John Tukey famously said:
"Far better an approximate answer to the right question, which is often vague, than an exact answer to the wrong question, which can always be made precise."
To be sure, fluorescence has time resolution and spatial resolution and quantification and those are cool. But--unless you've tested this fluorescent-tagged inhibitor in vivo--you're not measuring the species of interest.

Permalink to Comment

17. newnickname on April 24, 2013 4:41 PM writes...

Their AZD2281-BODIPY chimera migrates and binds but the piperazine-NH is now a piperazine-NAmide. Could that NH be more crucial to function than to binding?

Which could lead to many possible modifications: link the dye via another position (add a side chain to a piperazine C); link using a cleavable linker; link using an "NH" bearing component rather than that C=O amide O; etc..

Permalink to Comment

18. Hap on April 24, 2013 4:53 PM writes...

I haven't read the paper, but wouldn't the function of the fluorescently labeled inhibitor be a reasonable sanity check on whether the label is perturbing the compound's mechanism of action? If the derivatized compound behaves similarly to the underivatized compound (similar binding in vitro, similar biogical effect in vivo), then wouldn't that place an upper bound on how much perturbation the label creates? It doesn't guarantee that the mechanism hasn't changed (a problem for a tool compound), but it should give an idea of whther there are any gross changes.

Would mass spec be possible here (or could you freeze cells at specific timepoints and sample by MS)?

Permalink to Comment

19. Andy on April 24, 2013 9:30 PM writes...

#13: When a patient has certain BRCA1 mutations that prevent the oncoprotein from functioning in the homologous recombination (HR) pathway, DNA repair is shunted to the base excision repair (BER) pathway, the activity of which requires PARP1

You have this backwards. The BER pathway is actively engaged in all cells dealing with the 10,000 or so DNA single strand breaks that occur due to oxidative damage. In the presence of functioning PARP, these are repaired by BER, in the absence of PARP, the residual single strand breaks are converted to double strand breaks by DNA replication, at which point homologous recombination is the rescue mechanism. If HR is defective (due to lack of BRCA1/2 activity), the double strand breaks don't get repaired which is a cell-lethal event.

Permalink to Comment

20. Crimso on April 24, 2013 9:42 PM writes...

So the drug is apparently reaching its target. If (and I know this is a big if as some of the above comments point out) the fluorescent tag isn't having any negative effect on the supposed function of the drug, how about this: PARP1 undoubtedly interacts with other molecules (large and small), perhaps there is something there blocking the binding of the drug. Or perhaps the drug gets to the vicinity of its target, but finds soemthing else in the area that it likes much better. I'm guessing there are proteomics data that go against the latter (though maybe not); but as to the former, what are the chances that every interaction PARP1 has in vivo are known? I'm thinking close to 0%.

These are just a few things that occur right off the top of my head (I'm not a drug discovery guy, btw). Thinking about this for a long while would probably result in a number of other potential complications that might be interfering with the drug. You have to admit that the story would be much more boring if they found the drug never entered the target cells (or was cleared by the liver, etc.). The fact that it SHOULD be working but isn't will probably lead to some important findings that will be applicable to other drugs as well. And thusly do we lurch forward.

Permalink to Comment

21. overthetop on April 25, 2013 9:02 AM writes...

I can't access the journal article through the paywall, but have a question for those that can.

Did the authors use the fluorescent tag by itself as a control and observe that it did not accumulate in the nucleus without the inhibitor present?

Permalink to Comment

22. RKN on April 25, 2013 11:54 AM writes...


Correction noted. Thank you.

Permalink to Comment

23. Imaging guy on April 25, 2013 1:15 PM writes...

You can get the full-text article from Pubmed Central (PMC) for free. Supplementary information can be downloaded as a single pdf file from nature website for free too.

In order to prove that fluorescent drug is similar to its parent compound (olaparib), they did an enzyme inhibitory assay which showed that the IC50 of fluorescent drug is about 7 times higher than the parent compound (12.2 nM vs. 1.7 nM) (Supplementary fig S3C). Since the concentration of fluorescent drug inside the cell was about 1.2 ┬ÁM (ten times IC50), I assume that it was still able to inhibit 90% of PARP1. In order to prove that the fluorescent drug was binding to PARP, they did a challenge test and found that excess parent compound completely blocked the binding of fluorescent drug in cell culture (Supplementary fig S3D). Moreover they showed that fluorescent drug colocalized with antibody against PARP1 in tissue sections.

Permalink to Comment

24. Kilo on April 29, 2013 4:08 AM writes...

The article is available free from PMC.

Perplexing to see a detailed study performed in a cell background in which Olaparib is known to be inactive.

Permalink to Comment


Remember Me?


Email this entry to:

Your email address:

Message (optional):

Gitcher SF5 Groups Right Here
Changing A Broken Science System
One and Done
The Latest Protein-Protein Compounds
Professor Fukuyama's Solvent Peaks
Novartis Gets Out of RNAi
Total Synthesis in Flow
Sweet Reason Lands On Its Face