Skip to main content

Chemical News

Screen Carefully

This recent paper asks the question “Can Invalid Bioactives Undermine Natural-Product Based Drug Discovery?” And as any of you who’ve done any natural product screening can attest, the answer is “You betcha”. There are any number of natural products (here’s one!) that, although active in a given screening assay, turn out to be active in a lot of screening assays. They are, in other words, pan-assay interference compounds (PAINs), and their origin shouldn’t be allowed to confuse that issue. Quercetin, gossypol, berberinegenistein, curcumin. . .the list goes on, and this paper does everyone a favor by providing just such a list, derived from the Napralert database.

You will notice many a quinone and polyphenol on it, and that is no coincidence at all. Natural products have a reputation (often earned) for exquisite selectivity, but they have also, in other instances, earned a reputation as fragmentation grenades. A compound like celastrol can show up as everything from an antioxidant to a protease inhibitor to an insecticide, and you know what? It’s all of those and more, thanks to that quinone methide functional group. Such compounds are almost certainly going to react promiscuously with sulfur nucleophiles inside a cell, show redox activity, and wreak all forms of havoc. There are covalent functional groups that are weakly reactive (epoxides), and there are covalent functional groups that are ready to party, and a quinone methide shows up already lurching to one side, with sloshing red party cups in both its hands. Everyone’s pal resveratrol is there, too, which probably won’t surprise people who’ve worked with it. Its polyphenol nature has raised eyebrows from the moment it was proposed as a sirtuin inhibitor, and its winding, frustrating, unresolved story since then shouldn’t come as too much of a surprise. What you want to do, though, is to avoid running the next project that’s heading in the same direction.

This would be a timely place to mention this open-access book chapter, just out from Jayme Dahlin, Jonathan Baell, and Michael Walters, on assay interference by reactive compounds. It’s a comprehensive guide to the subject, with recommendations for counterscreens at every step, and it’s well worth a look. It’s especially strongly recommended for groups that are just getting into a big screening campaign for the first time – the experience contained there can save you from a lot of wasted time and effort

As this paper shows, you’ll need to pay attention both to functional group problems and to things like aggregation. A good number of the compounds on the watch list are notorious aggregators under common assay conditions, and this behavior can throw false positives all over the place. There are also a lot of fluorescent interference compounds (emitters and quenchers) on the list, so if that’s your readout, caution is indicated. Luciferase-based readouts are all over the place in assay development, and plenty of natural products find ways to interfere with that system, too.

The paper also raises some bigger questions. How are these activities related to these various compounds’ actions in vivo? How do cells handle them, and why are they so common and biologically important? In many cases, their true functions in cells aren’t well understood at all. And for screening, what (and where) is the line between “very active compound that should be ignored” and “very active compound that will save your project”? Choose wisely!

7 comments on “Screen Carefully”

  1. MoBio says:

    Yes this is something which bedevils me and our screening center all the time. Particularly when the only ‘active’ is a known promiscuous compound (read “quercetin” here).

    What we and others have found helpful is to go to PubChem (or internal database) and note that these compounds are ‘active’ in many cases in 100’s of assays.

    It begs the question, which you raise of what to think about in those instances in which you see apparently therapeutic actions in animal models (see this for a Alzeimer’s Disease model …normally I assume this is related to the known anti-oxidant activities of these compounds but who knows for sure?

  2. Paul Brookes says:

    Typo in end of 2nd para… Resveratrol is (allgedly) a sirtuin activator, not a sirtuin inhibitor. But, maybe the typo is correct and you should leave it in, as an indication of the general FUBAR nature of this field.

  3. anon the II says:

    This stuff isn’t rocket science. Any old chemist who watched animal models go away and get replaced by in vitro screens knows there are a whole bunch of these things that started showing up in the 80’s. I first saw the a lot of them when our mouse paw edema models were replaced with PLA2 enzyme screens. All you need to do is hang on to a few old natural products chemists and a few old guys who’ve been associated with lead generation. Everybody knows all about this stuff.

    Oh wait! They did what? They fired, excuse me, laid off all those people. Well no wonder we’re seeing this stuff over and over again.

  4. Erebus says:

    Thanks for the very interesting link and summary.

    It’s interesting to note that there are 39 compounds on that study’s list. Many are polyphenols like quercetin and resveratrol, some are fatty acids, and the rest are mostly terpenes — monoterpenes, diterpenes, and triterpenes are all well represented. Tetrandrine and capsaicin are the only alkaoids. The medicinal compounds taxol and glycyrrhizin are also on the list.

    With respect to drawing the line between “very active compound that should be ignored” and “very active compound that will save your project”, I would suggest the following:

    1. Compounds with very high potency at one specific and measurable site, for e.g. capsaicin at TRPV1. Resiniferatoxin would be an even more extreme example. When you’ve got nanomolar/picomolar affinity for a specific target, a small amount of interference and some minor off-target effects shouldn’t be too much of a concern.

    2. Compounds which don’t appear to hit any target effectively at low concentrations, and which don’t have a well-defined medicinal function. These are typically ubiquitous dietary phytochemicals, e.g. quercetin, oleic acid, resveratrol, most of the monoterpenes, and so forth.

    3. Compounds which consistently hit several receptors with high potency, in a measurable and verifiable way. There could be no more appropriate example than celastrol. It is one hell of a dirty compound. In addition to everything you’ve already mentioned, it is an HSP-90 inhibitor, thus it down-regulates just about every nuclear steroid receptor. It’s something of an anti-progestin, and an anti-androgen, and an anti-glucocorticoid, and so forth…
    -“Celastrol regulates multiple nuclear transcription factors belonging to HSP90’s clients in a dose- and cell type-dependent way”, PMC3024068
    -“Celastrol Inhibits Hsp90 Chaperoning of Steroid Receptors by Inducing Fibrillization of the Co-chaperone p23”, PMC2823561

    (Speaking of “sloshing red party cups”, celastrol itself is exactly that same bright red color. Unusual for a triterpene.)

    …I’d use class #1 extensively, throw away #2, and use #3 very sparingly and only in exceedingly well-defined circumstances, where there might be no other options at-hand.

  5. Peter Kenny says:

    It is not always easy to make a clear distinction between fact and opinion when evaluating screening hits or deciding what to screen. Put more bluntly, we need to be brutally honest about what we know and what we believe. I’ve linked a blog post on this theme as the URL for this comment.

  6. Derek Freyberg says:

    “There are covalent functional groups that are weakly reactive (epoxides), and there are covalent functional groups that are ready to party, and a quinone methide shows up already lurching to one side, with sloshing red party cups in both its hands.”
    Dr. Lowe, for a non-drinker (even of coffee, as I recall), you have a flair for alcoholic analogy.

  7. Ash (Curious Wavefunction) says:

    I would spread the drugs under your feet:
    But I, being poor, have only my screens;
    I have spread my screens under your feet;
    Tread softly because you tread on my screens.
    – W. B. Yeats’s “He wishes for the cloths of heaven” rephrased

Comments are closed.