Son of a nonAddicting Painkiller

I’ve decided to republish yesterday’s post with some added material to show exactly how to find one of the world’s most useful, not to say lucrative drugs. The response of a friend led me do this. If you’ve read yesterday’s post — start reading after the ****

Synthetic organic chemists, molecular modelers and X-ray crystallographers fire up your engines. A great target is available, discovered by the humble centipede of all things.

Several times during my career as a neurologist, a new drug was ballyhoo’d as a non-addicting narcotic — Talwin springs to mind, among others. The hope was that a molecule with both agonist and antagonist properties wouldn’t be addicting. All basically inhibited neurotransmitter release (like the opiates), to which the synapse responded by making more receptors for the neurotransmitter, explaining tolerance, addiction to the new drug and why junkies can take doses of morphine that would kill you and I.

Have a look at Proc. Natl. Acad. Sci. vol. 110 pp. 17534 – 17539 ’13. Here’s some background. To conduct a nerve impulse, neurons have to let sodium ions flow through their membranes. This is done through proteins called sodium channels. We have 9 distinct genes for them, and most neurons express more than one.

Consider the humble pufferfish. It sometimes makes a beautiful organic molecule called tetrodotoxin ( which blocks all but 2 of our sodium channels stopping them from conducing sodium ions. It can kill you and is 100x more potent than cyanide. Even so, it’s a very pretty molecule (to an organic chemist) with some resemblance to adamantane.

The centipede makes a 42 amino acid protein which essentially blocks only one sodium channel (NaV1.7). It’s quite potent, doing this at a concentration of 25 nanoMolar. Clinically, it is even more potent than morphine (well, in animal models anyway)

So why get excited? Because as far as we can tell, its action is on peripheral nerve fibers, not the brain. For some reason nerve fibers carrying painful impulses (a philosophic conundrum — impulses themselves are no more painful than a fire feels hot to itself) from the periphery to the spinal cord and brain use lots NaV1.7. So addiction shouldn’t be a problem.

There were some clues already — there is a disorder called congenital insensitivity to pain [Nature vol. 444 pp. 831 – 832 ’06], due to loss of function mutations in NaV1.7. Other mutations here cause several painful syndromes — erythromelalgia [J. Med. Genet. vol. 41 pp. 171 – 174 ’04 ], chronic rectal pain. These mutations cause a hyper functioning sodium channel which stays open too much.

Total absence of pain isn’t good, as we need it to warn us that we’re stressing our joints excessively. In the bad old days when there was a lot of syphilis around, it sometimes caused peripheral nerve degeneration, resulting in something called Charcot joints —

So get cracking guys, if tetrodotoxin, a small compact molecule can nonspecifically block most sodium channels, surely you should be able to find something smaller then a 42 amino acid protein to block NaV1.7 (selectively of course). I’m not sure that we have a structure of NaV1.7, but others are known, along with all their amino acid sequences, so it should be possible to model binding sites by analogy.


I received the following from a computational chemist friend. ” Although good luck finding a peptidomimetic for a 42 amino acid peptide…”

I don’t think we’ll need a peptide at all. Likely, it will be a modified tetrodotoxin-like molecule. So we know both where to look and how to look.

First some more background, which I should have put in the original post. People with mutations in NaV1.7 and congenital insensitivity to pain, are cognitively normal (except for pain sensation). This implies that just blocking the NaV1.7 with even a drug getting into the brain won’t have cognitive and/or addictive side effects (and any medicinal chemist will tell you how hard it is to get drugs into the brain).

Consider the Grasshopper mouse of the Arizona desert [ Science vol. 342 pp. 428 – 429, 441 – 446 ’13 ]. It eats the fearsome Bark scorpion, without suffering pain. The venom slows the inactivation of the NaV1.7, effectively activating it and causing pain in every other animal. Mutations in another sodium channel (NaV1.8) cause binding of the toxin, so it doesn’t get near NaV1.7. The variants are very near the pore of the channel (where the ions go through). Just a switch in the sequence order of two amino acids (Glutamic acid and glutamine) in NaV1.8 is enough to make NaV1.8 inhibited by the venom.

What does all this have to do with finding a drug to inhibit NaV1.7?

We have the amino acid sequences of all 9 of our sodium channels? NaV1.7 is sensitive to tetrodotoxin (e.g blocked by it), NaV1.8 is Insensitive. The Xray crystallographers need to give us structures of both channels, with and without tetrodotoxin bound. The computational chemist has to dock a tetrodotoxin analog into NaV1.7 similar to the way it fits into NaV1.8 blocking it. Not easy, but we know the target, and we have an excellent candidate to start with. There is no need to mess with 42 amino acid proteins, even though nature and the Grasshopper mouse did. The synthetic organic chemist will be kept busy making tetrodotoxin variants, perhaps those suggested by the computational chemists.

The really hard part in all this will be finding a drug that isn’t a super tetrodotoxin, blocking NaV1.7 and everything else. It should be doable.

The rewards for the successful company doing this are enormous. I’ve got enough money and am too lazy to suit up and go back into the lab (assuming anyone would have me). The reward of pointing the way will be enough. See

Addendum 8 Nov ’13 — Tetrodotoxin should be particularly easy to model, as its stereochemistry should be relatively simple. The adamantine-like structure doesn’t move at all, being locked in place. There is a CH2OH side chain which can flop about, and a cyclohexane ring fused to the rigid part. Even there, the guanido group embedded there is likely to be planar. It’s easy to see why it’s water soluble with all those hydroxyls, and the fact that it is a zwitterion. Modeling solvent interactions is never easy, but modeling hydrogen bonds to amino acid side chains is probably easier, which is what’s important here.

Post a comment or leave a trackback: Trackback URL.


  • Curious Wavefunction  On November 7, 2013 at 12:01 pm

    Yes, that’s why I said small molecule peptidomimetic. It certainly sounds like a lucrative target, but look at how much time and resources were sunk into finding a small molecule analog for the natural peptide that inhibits renin (as a blood-pressure lowering target). It took 20+ years of efforts before they discovered Aliskiren (interestingly, with a healthy contribution from computational chemistry). Not saying that the effort would not be worth it, but it might be hard to convince an already ailing pharmaceutical industry to fund that kind of long-term, ambitious effort.

  • Jack  On November 11, 2013 at 5:08 am

    Not to burst your bubble, but you know pharma already has multiple programs targeting Nav1.7 right? Convergence Pharmaceuticals has CNV 1014802. Maybe Afferent Pharmaceuticals has something? Pfizer has something via Neusentis I think. For all I know they have already all failed or been shuttered, but at least some of these made it to phase 1. I will also bet none of them are TTX analogs, but who knows.

    • luysii  On November 14, 2013 at 6:00 pm

      No I didn’t know that. I did think of it myself (so I’m happy to know people are working on it. As I said at the end of the post “I’ve got enough money and am too lazy to suit up and go back into the lab (assuming anyone would have me). The reward of pointing the way will be enough.”

      Also it’s worth looking at Proc. Natl. Acad. Sci. vol. 110 pp. 17534 – 17539 ’13 and [ Science vol. 342 pp. 428 – 429, 441 – 446 ’13 cited in the post, to see just how potent such a medication would be.

      I’m not sure why you think they wouldn’t be TTX analogues — it’s already the perfect drug, if it just didn’t clobber the other 6 channels. Designing such an analog would be the midterm, final exam and PhD thesis for protein chemists and molecular modelers. If they really deeply understood what is going on at the drug protein interface, the way synthetic chemists stalk their target, it should be a snap — but they don’t as the honest ones readily admit — see Curious Wavefunction on this point. Here’s a link —

      Thanks for commenting. The delay in responding to your note had to do with the birth of our first grandchild.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: