Skip to Content

Silicon Stays in the Shadows

I like this review, but I’ve seen it before. Well, not this exact manuscript, but every few years it seems there’s another one with a similar title, something about “Incorporating Silicon Into Drug Structures”. I am guilty of the exact same thing, though: here’s a blog post from 2004 on the topic, and here’s one from 2012.

This latest paper incorporates some recent references, but looking through it, I think it’s fair to say that we don’t know that much more about organosilanes as drugs than we did in 2012. At right is the paper’s comprehensive list of all the organosilanes that have entered human clinical trials, and unless I’m missing something, none of them are particularly recent. There have been papers with in vitro studies since then, or animal work, but I believe that the most recent clinical work reported has been the photodynamic phthalocyanine PC-4, which was published in 2010. If anyone knows of any Si-containing drug candidates that are currently in the clinic, I’ll be glad to correct that impression.

So why is this? The obvious answers, which I think in this case are also the correct ones, is that no one has yet identified a particular property that silicon substitution brings to a drug structure that is in much demand, and that any such properties (which are likely case-by-case things) have so far been deemed not worth the trouble or risk involved. Silicon is most certainly not carbon, and philosophically, you’d imagine that having a not-quite-carbon atom to mess around with might be the source of a lot of interesting structures and biological activity. But so far, that hasn’t quite been the case. It’s probably true that a larger or more concerted effort might well uncover some more interesting things, but who’s going to make such an effort? “This could be interesting, we’re not sure” is not quite enough of a rationale for the kinds of money and time that go into drug research. And some would look at the record so far and say “This could be interesting, but probably not”, even thought that might be unfair.

That’s not to say that silicon can’t someday bring something to the party. You’ll note from that graphic that most of the substitutions have been pretty straightforward trimethylsilane-for-tertiary-butyl sorts of things (the only organosilane I’ve ever synthesized on a drug project was exactly that, and boy, did I get the fish-eye for making it). There are surely a lot of silicon-containing heterocycles, for example, that we either don’t know how to synthesize yet, or have not really been evaluated in drug structures. One of these might be just what some project is looking for.

But as long as it’s in the “might be” category, the chances that someone will go to the trouble of finding that out are relatively small, compared to the option of sticking with the more common elements and just making more analogs. If there were more commercially available intermediates and building blocks of this type, chances would go up – but in order for someone to figure out how to make them and then offer them for sale, it would help if there were more interest among the customers. Chickens and eggs, once again. . .

34 comments on “Silicon Stays in the Shadows”

  1. Mark says:

    So when you say “no one has yet identified a particular property that silicon substitution brings to a drug structure that is in much demand” do you mean that the studies haven’t turned up anything interesting or that there aren’t enough studies that have been done yet?

    Is this a call to action?

    1. Derek Lowe says:

      I’m not sure myself! I think that the first is true, which means that the second is going to be a matter of opinion. . .

  2. Wavefunction says:

    From what I have seen, silicon, unlike boron, cannot usually make a decisive difference. At best it can lead to slightly better PK which might provide an advantage in patent busting and incremental me-too advances. The “killer app” for Si would presumably be very rare, although I wonder if someone has taken advantage of a displaceable group on silicon to form strong Si-O bonds with an enzyme to shut it down.

  3. Anon says:

    Can all these be considered as pro-drug in that one can expect proto-desilyalation in stomach?

  4. John Wayne says:

    During my career I’ve made a dozenish carbon-for-silicon analogues, and the pairs usually have similar in vitro potencies. The silicon analogues have had much higher LogD’s (about a log) with corresponding and predictable PK properties. If I am ever on a project that has a lead that is way too polar, this is in my toolbox.

  5. one_man_CRO says:

    no shortage of silicon containing compounds not moving toward the clinic……cyclopentasiloxane and the like

    1. Anonymous says:

      cyclopentasiloxane, D-5, D5: already used in cosmetics and other health / hair / beauty products and in personal lubricants. 🙂 It is also being used in dry cleaning, seeking its place as an “environmentally friendly” safe solvent.

      But there is some controversy. Environmental and biological persistence. When broken down, do the methyl groups generate CO2 (greenhouse gas) or stop at CH2O (carcinogen)? Some rat studies found an increased incidence of uterine cancers. It is ultimately prepared from chlorosilanes which means you really aren’t getting rid of chlorine (cf., perchloroethylene used in dry cleaning) … except in your advertising.

      When it comes to silicone based dry cleaning and washing, I guess that would be an example of greenwashing 🙂 https://en.wikipedia.org/wiki/Greenwashing .

  6. Red Agent says:

    I wonder about the metabolic fate of silicon in anything that would bypass, or get past, the GI tract. Unless it was bound up tightly, away from catabolic intervention, I envision silicon dioxide trails wherever it went.

  7. Derek — check your gmail — I sent you an invite to participate in CHI’s ‘Drug Discovery Chemistry’ event — San Diego April 3-5, 2018.

  8. Barry says:

    The carbon-silicon (single) bond length is 20% longer than the carbon-carbon. In a ring, that’s going to change all the bond angles, and therefore the vectors on which all the substituents are displayed. It’s very similar to the carbon-sulfur bond length, but without the lone pairs/oxidative liability.
    So even if TMS-for-(t)butyl is futile, we shouldn’t stop looking at these transmutations in the scaffold.

  9. Petros says:

    There were 2 or 3 startups floating around back at the beginning of the Millennium which had this as their primary/sole strategy.

    Amedis Pharmaceuticals was bought out and the strategy ditched. I don;t know what happened to the others

  10. Druid says:

    I like the structure of silatranes (chemical curiosity), and they certainly have some uses (rodenticides, lab reagents) but no medicines yet after more than 40 years.

  11. Matthew Statham says:

    The silanediol group as a bioisostere of the hydrated carbonyl group or hydrated amide transition state is the most interesting of the silicon based functional groups and offers the largest potential difference in properties. eg. Prof Scott McN. Sieburths work

    1. DrOcto says:

      The main issue with silanediols (as far as I’m aware) is that they polymerise as pure compounds, though they are quite stable in dilute aqueous conditions. A second issue may be that the hydroxy group is too easily substituted, which would give any number of off target effects.

      I think that pharmaceutically that there might be some hope for some sort of silanediol pro-drug, which you might develop after demonstratign efficacy of the diol itself at the target.

      1. Matthew Statham says:

        With enough steric hindrance they are pretty stable in my experience eg. diphenylsilanediol
        By switching a carbonyl/amide for a silanediol you create an isostere of the hydrated transition state which along with strong hydrogen bonding should be able to irreversibly bind to an enzyme and shut it down

  12. tlp says:

    Hey, how about all those silicone implants?
    Also carbon-silicon analogy calls for a bunch of lucrative mee-too strategies. That’s indeed a surprise no one tried to travel down this path. Maybe ‘entrepreneurs’ are still busy with hydrogen/deuterium low-hanging fruits.

  13. Christophe Verlinde says:

    From the review: “In fact, silabolin, a prodrug of the steroid nandrolone, was available in Russia and used by athletes”. Those Russians, they love their silicon drugs.

  14. Rhenium says:

    Oh, perfect timing for the 1982 Dewar and Healy “Why life exists paper” regarding carbon and silicon comparisons. It’s an oldie but a goody…

    Michael J. S. Dewar, Eamonn Healy
    http://pubs.acs.org/doi/abs/10.1021/om00072a029
    Organometallics, 1982, 1 (12), pp 1705–1708

    1. Me says:

      Yeah Star Trek did it way before then!

    2. Bagger Vance says:

      I like to think a paper that really answered the question “why life exists” would show up in a journal with a slightly better impact factor than Organometallics, thankyouverymuch.

      1. Rhenium says:

        It’s provocatively named that way for a reason!
        As for impact factor, Dewar could have published anywhere he wished, but this was (I think) in memoriam for Rowland Petit. We just covered this paper in my graduate inorganic course, hence why it was on my mind when Derek had his post.

      2. tlp says:

        like in a Bible or something?

  15. SillyCone says:

    Silicon has some rather unique properties it can impart on molecules; increased lipophilicity, reduced toxicity, and increased metabolite resistance. Unfortunately, the gains (when they do happen, which isn’t that often) are not significant enough in most cases that the cost of making those changes is desirable. Some of the most extensive work has been done by Prof. Reinhold Tacke. Essentially decades of silicon-carbon switching comes from his group. The problem is when you read his papers, it becomes apparent that it is very much hit or miss. Some swaps can give huge effects (good or bad) and others don’t do a damn thing.
    As was stated earlier by Statham, the silanediol really is the most interesting due to the lack of reliably generating and maintaining the gem-diol on carbon.

    I think the root of the problem is that we’re still learning a lot about silicon in small molecules. A lot of why we go after certain functional groups is because we try to mimic the functional groups that show activity in natural products, but there are no silicon containing natural products to show us what structures are bioactive. There is a single report of a silicon complex formed in a living system by algae (http://pubs.rsc.org/en/content/articlelanding/2002/dt/b105379p#!divAbstract).

    Other than that (which isn’t really a natural product), we’ve got nothing to go on. We’re trying to catch up to what evolution has done over billions of years with carbon; it’s gonna take more than the time we’ve put in so far (less than a century of silicon bioactivity studies) to come up with what silicon groups and structures affect biological systems (if there are any).

    Unfortunately, Derek is right, no one is going to put in a huge effort and truck loads of money into that sort of thing. My guess would be that, the slow, but steady, pace of academic silicon work out there may one day find the biological utility of silicon, but it will take quite a bit of time.

  16. Young Padawan says:

    I remember having seen silicon used in fragrance research, though I don’t know if any of this has ever made it to the market.
    10.1002/chem.201000549
    10.1002/cbdv.200890105
    I sure like the idea that you can imitate the scent of a complex polycyclic terpenoid like patchoulol with an achiral silcon derivative.

  17. Pennpenn says:

    So what you’re saying is that no one has time to deal with all this silaness?

    1. Me says:

      Very good pun, PunPun.

    2. a. nonymaus says:

      Would it be germane to move down the periodic table a second time?

      1. Derek Lowe says:

        Tip your waiters and waitresses generously, folks. . .

        1. db says:

          Puns. Many can’t stannum because they can’t tell where they lead.

  18. The annual North American Silicon Symposium will be held in May 30 & 31 in Edmonton, Alberta. International attendance and all things organosilicon including medicinal. See: https://sisymposium.com

  19. Patrick Lam says:

    Here is a novel drug design concept. There is an interesting property of silicone that is staring at us. We have all learnt that silicon likes to interact with oxygen nucleophile, besides fluoride, in our advanced organic chemistry course in graduate school. Can one use the silicone of trimethylsilylphenyl group to form noncovalent/semicovalent bonding with the oxygen of aspartate or glutamate (i.e. forming hypervalent silicone complex). This is much like using boron except there is less potential for toxicity. Trimethylsilyl group is so much bigger than t-butyl group that one needs a superwide pocket to test this hypothesis. I believe I had presented this concept in several of my “molecular recognition in drug design” talks.

  20. Barry says:

    That would have to be an Aspartate or Glutamate in a lipophilic pocket. I don’t know that such a thing is possible. The silane would have to displace the Asp/Glu hydration (potentially a strong H-bond) to form a weak coordination to silicon.

  21. no proof says:

    “even thought” => “even though”

  22. AVS-600 says:

    From a scientific perspective (which, understandably, is not the same as a cost/benefit risk-management perspective), the question probably shouldn’t be “is there strong evidence that silicon is OK in drug-like molecules?” but rather “is there any evidence that silicon is problematic in drug-like molecules”.

    Looking at that chart, there are only five compounds that appear to have gone into the clinic that have stable silicon centers (as opposed to pro-druggish silyl ethers and siloxanes). If drugs in the clinic have a 90% failure rate, there is a roughly 60% chance that all five would fail even if the null hypothesis of “silanes are, broadly speaking, about as metabolically stable as alkanes” holds true. Is there any evidence from any of those studies that the silicon atom caused a metabolic liability (additionally, did any of them fail due to tox, and can that be potentially blamed on the silicon atom)?

    Unlike fluorine (or deuterium!), it seems like silicon is really likely to be most useful as a structural tool rather than a functional one. One thing that comes to mind is that it might sometimes be the answer to something like “I wish I could make a 6.5-membered ring”.

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.