One of the small-molecule drugs that’s getting attention as a possible coronavirus treatment is Ivermectin, which is an interesting story from a couple of different directions. I’ve been getting some inquiries about it, so I thought it was time to have a look.
Background on Ivermectin
It’s part of a family of natural products called the avermectins, isolated (as have so many other interesting natural products) from soil actinomycetes. I go back quite a ways with this compound, back to my graduate school days when several folks in my old research group were trying to complete a total synthesis of the compound. It was (not so affectionately) known in the group as “Nevermectin”, which should give you some idea.
If you look down in the “southeast” of the molecule, you can see one reason it was no fun to make – the organic chemists in the audience will note that deconjugated enone (the double-bonded carbons that aren’t next to the carbonyl, but rather skip a carbon). As you might imagine, getting that thing in that position and (even worse) keeping it there was very annoying. The rest of the structure falls into the “doable with time and effort” category – the rest of the ring structure down there, the spiroketal in the northeast, the macrolide polyene ring, and the glycosyls coming off of it. The carbohydrate part is two oleandrose units in a row; I think that this is the most famous occurrence of that otherwise obscure sugar. Ivermectin (and parts of it) have been synthesized several times since the 1980s by different routes, but like the huge majority of natural products, no one makes avermectin from scratch – it’s made by fermentation of the soil organism, which is far better at it than we’ll be any time soon. A related group of compounds, without the carbohydrates on the side, are the milbemycins.
The drug’s two main discoverers and developers won the Nobel in 2015 for it, deservedly so, because it has been a tremendous force in the campaigns to eliminate untold amounts of human suffering in river blindness and filariasis. (Merck, I should note, has donated the drug for those purposes ever since the late 1980s). The drug is effective against a wide number of parasites and arthropods in general- pinworms, mites, lice, heartworms and fleas in dogs, parasitic worms in pasture animals, as an ingredient in ant baits, and more. It acts most strongly on glutamate-gated chloride channels, which vertebrates don’t even have, although it can also bind to other ligand-gated chloride channels at higher concentrations. It’s really just an accident of pharmacology that it doesn’t cause more problems – it’s very toxic to a number of turtle species, for example, and to some particular breeds of dogs (particularly collies and other herding types). In both those cases it seems to be differences in the P-glycoprotein transporter proteins in the blood-brain barrier; that’s one of the key “toss this compound back out” proteins in many cell membranes. Turtles apparently don’t pump the drug back out at all, and those dog breeds have picked up mutations in PGP (all that close breeding) that allow the drug to accumulate in the brain and set off severe neurotoxicity via those other ion channels.
Ivermectin and COVID-19
So what is this compound doing in the news with the coronavirus? Its ion-channel mechanism of action against parasites has no application to viruses. But it has been shown several times to have activity in cell assays against Dengue infection and to reduce viral loads in the mosquito vectors (see this paper and references therein), although the reasons for this are still unknown. There have been some human studies for efficacy, but all I can find out is that a preliminary read showed no reduction of viral load in the infected patients, and there seems to have been no update. But since denguevirus is also a single-strand positive-sense RNA virus, Ivermectin has come up as a possible coronavirus drug, mechanism or not.
Here’s where the puzzling part comes in. A group at Monash Univ. showed a month ago that Ivermectin could inhibit the coronavirus in vitro. That paper is also a good review of the attempts at repurposing the drug in the antiviral space. The thing is, if you crank up the concentration, probably a lot of stuff will inhibit any given virus in vitro (relevant XKCD). The IC50 in their assay was about 2µM, and that number is where many medicinal chemists started getting off this particular boat. Micromolar cell-assay activity can certainly be real, but you’d generally want more potency than that as you move in vivo. The thing about Ivermectin is that it is effective against things like the river blindless parasite at really low doses: one 12mg or 18mg dose every few months. That’s not because it hangs around in the blood in high concentrations for that long – the compound’s half-life is a few hours, and it’s extensively metabolized as it goes through the liver. It does distribute strongly out into tissue, though, and that combined with its intrinsically high activity against the roundworms does the job.
But let’s imagine that you want to hit that 2 micromolar IC50 concentration in the bloodstream: how much Ivermectin would you need to take? Well, this calculation will involve some hand-waving, because who knows what the pharmacokinetics do as you head upwards like that. Here’s why I don’t want to even guess: that last link will show you that an 18mg tablet gives a peak blood concentration of 31 nanograms/mL. Ivermectin has a molecular weight of 875 (rather a lot when you think about it, to quote an old Monty Python routine), so that Cmax of 31 ng/mL is about 35 picomolar in the blood. Told you it was potent! So we’re around five orders of magnitude off here, which just isn’t feasible. One would have to hope that the compound concentrates significantly in the site(s) of action – really, really significantly. And we have no evidence that that’s happening.
It doesn’t take much to do these calculations – people were throwing their hands up in the air as soon as the Monash paper came out. As those letters to Antiviral Research note, ivermectin will start to hit other targets as you jack up the doses. The reason it’s such a great antiparasitic drug is that you don’t have to dose up at all. It’s hard to imagine that it’s a safe compound if you really have to push the concentrations to even within binocular distance of those in vitro levels. For that reason, the FDA has been warning people not to take the drug for the coronavirus.
Nonetheless, there is said to be a clinical trial in France getting underway investigating the drug, and just the other day one went up at Clinicaltrials.gov. That one looks odd to me in general – it’s an open label comparison in coronavirus patients of Ivermectin and the antiandrogen drug bicalutamide (which is given for prostate cancer). I simply have no idea what that’s about; if bicalutamide has been publicly suggested as a repurposed drug for the pandemic, I’ve certainly missed seeing it. Update: there are actually nine trials showing at Clinicaltrials.gov ranging from Ivermectin along to things like Ivermectin plus nitazoxinide, which also seems futile to me.
A correspondent has sent me anecdotal reports from South America of people seeing efficacy with Ivermectin, but (as is usual with anecdotal reports) I don’t know what to make of those, either. At the normal doses of the drug (which is what those reports were doing, or perhaps 2x to 3x), it’s hard to see how any activity it might show would relate to the Monash study at all. I’m willing to wait to see some more controlled clinical data, but this really does look like the proverbial shot in the proverbial dark to me. I still have some hopes for repurposed drugs against the coronavirus, but not really in this case.
Update: see here for more on Ivermectin PK as regards coronavirus therapy.
Update: as per the comments, there’s work showing that Ivermectin apparently inhibits the actual of a different class of viruses based on inhibition of protein import into the nucleus. Could that be what’s going on here? Such a mechanism has been considered for antiviral therapy, but it’s in early stages of work.
Update: there is a report, using data from a company that has a real-time hospitalization database, of a retrospective matched-control set of coronavirus patients receiving Ivermectin versus those that did not. This analysis showed a pronounced effect in favor of Ivermectin, which is quite interesting. If this is real, it should definitely show in the other trials that are underway. More on this as data come in. . .Update: this company’s data-handling techniques have come under fire, and the report mentioned has been taken down, reasons unexplained.