Skip to main content

Diabetes and Obesity

Proteins as Oral Drugs: Possible, But Not Probable

I have been writing this blog for some time (!) That occurs to me on seeing this article in Nature Reviews Drug Discovery on oral dosing of peptide drugs – I say that because of this 2002 post on the very same subject, at the time directed towards the then-still-somewhat-hot topic of Judah Folkman’s endostatin angiogenesis inhibitor (which was itself a peptide). I pointed out that such compounds have a rough time in development as oral therapeutics, and of course, they still do.

That new article goes into useful detail on why that’s the case, but fundamentally it’s because our digestive systems are completely focused on breaking proteins down into tiny absorbable pieces, and not towards letting intact peptides into the bloodstream. Digestive enzymes, mucosal barriers, the permeability of the various membranes involved, the biases of active transport mechanisms: the gut really does not want to let foreign proteins into the vasculature, for a lot of excellent reasons that have been under pretty hard evolutionary pressure since our distant ancestors first had anything like a gut. You’re fighting against all that when you want your protein drug to be an oral therapeutic, which is why there are very, very few of them.

The review goes into the various strategies that have been tried over the years – trying to enhance permeability (as always, a double-edged sword of a strategy), ways to harden-up the protein structures themselves, delivery devices such as intestinal patches and hydrogels, etc. And it lists several proteins that have had extensive work done on them, with insulin probably at the head of the group. An orally available insulin could be quite useful, and many people have taken a crack at it with no one making it all the way through yet. That’s actually a double challenge – just getting insulin as an oral agent is already a massive challenge, but even if you manage that you face the well-known trickiness of insulin dosing in general. The insulin response is very strong, very fast, and very brittle. One of the main ideas behind modifying the injectable protein over the years has been to try to mitigate those effects so that the dosing is smoothed out a bit into a more controlled longer-lasting form. All attempts to modify its dosing (oral, inhaled, what have you) have to contend with that behavior.

There is, though, a recently approved oral form of semaglutide, a GLP-1 agonist for type II diabetes that was first approved as a once-a-week injectable. It’s an engineered analog of GLP1 itself, acylated and modified to be resistant to cleavage by the DPP-IV protease, among other things. The oral form is a pill with a rather large dose of SNAC (sodium salcaprozate), a salicylamide-derived permeation enhancer. Its mechanism of action is still up for debate, and it probably acts via more than one, but it does indeed seem to enhance uptake of peptides. For semaglutide, this happens in the stomach (rather than the small intestine) and there’s a U-shaped dose-response – that is, you have have too little SNAC dosed along with the drug and you can also have too much, for reasons that are also not really clear. It needs to be taken with water, first thing in the morning in a fasted state, with no food for 30 minutes. There are some gastrointestinal effects in some patients, as you might expect, and there is variability both from day to day in individual patients and between patients, which means that physicians need to work out the dosing in each case. But the drug is effective and safe – no worse than the injected formulation and certainly more convenient to take.

So don’t just immediately assume that it’s impossible to get an oral peptide formulation. But don’t assume that the problem has been solved, either – SNAC and similar compounds have been around for quite a while now, and they are nowhere near the universal answer. But this plan can work, once in a while, if you’re willing to put enough time into the development and the clinical trials.

36 comments on “Proteins as Oral Drugs: Possible, But Not Probable”

  1. ‾\_(ツ)_/‾ says:

    Just use norovirus capsid to deliver them

    1. no body says:

      another promising method is encapsulation with milk proteins or similar

  2. milkshake says:

    is there any way to do transdermal delivery of peptides? Semaglutide would be just the kind of medication that could benefit from being slowly released from a patch

    1. DT says:

      Microneedle patches was my first thought.

  3. FoodScientist says:

    Any idea on the permeability of the buccal mucosa to small peptides, or under the tongue? I believe pro-athletes use these routes to administer IGF. Lol it could be a pill you put in your mouth and just don’t swallow.

  4. DevicesRus says:

    All of these things have been tried (transdermal, iontophoresis, buccal,) and despite the semiglutide oral success (the once a week injection has slightly better clinical outcomes) the only thing other than injection that works well is inhaled insulin (exceptional kinetics)

  5. Barry says:

    I grant that aspartame is not a drug, but surely it must be mentioned in the discussion of orally available peptides. The gastrin project went nowhere, but a lucrative sweetener did come out of it.

    1. Magrinho says:

      Not the case – aspartame is quickly broken down in the small intestine and there is no appreciable blood plasma concentration. But, yes, it is a decent high-potency sweetener. Oddly enough, even though there are better ones now, Diet Coke drinkers will scream if you try to make them switch. They like that cardboard-y, bitter aftertaste, I guess.

      1. Barry says:

        aspartame reaches its target w/o any need for systemic exposure, of course.

  6. some guy says:

    One of the approaches for oral delivery of proteins was to derivatize them with PEGylated functional groups maintaining the bioactivity and, ostensibly, increasing oral absorption. From what I saw of this approach, they could go from ~0% bioavailability to 1-2%, or so, but who can afford to basically throw away 98% of your protein. They tend to be expensive. Also, inter patient variability is a big issue. The difference between 0.5% bioavailability in one patient and 1.5% in another patient is pretty devastating if you don’t have an incredibly wide therapeutic window between therapeutic efficacy and side effects/toxicity. There were several companies that crashed and burned on this approach. We’re still waiting on pegylated oral insulin.

    Also, how well tolerated is a large dose of SNAC, in general (without any drug in it).

  7. NotThatLipinsky says:

    FYI – I believe there is an approved insulin inhaler formulation…

    1. Barry says:

      yes, inhalation and insufflation are quite distinct from oral administration. They skip the proteases of the stomach

      1. milkshake says:

        sorry for asking the obvious – but did anyone try suppositories of therapeutic peptides?

    2. A Nonny Mouse says:

      We once had an orally active renin inhibitor (peptide based) that was fantastic until the person doing the dosing went on holiday and it was then found to be inactive; seems that the perdon doing the animal dosing had been putting it into the lungs. So, yes, a good way of getting peptides in there.

  8. myma says:

    A lot of money has been spent on alternate non-parenteral peptide and protein delivery. Shed loads of money. Books written. Theses written. More books. Sheesh, every Journal of Controlled Release has an article or three or the topic.

    To people’s comments here: every possible alternate route has been tried. Every possible magic foo foo dust. Every orifice. Every super fancy delivery vehicle gizmo ultrasound clever whiz bang instrument. OK, I don’t recall intra-ear canal delivery for systemic protein delivery, but aside from that everything.

    Back in my consulting days over a decade ago, I had an assignment to assemble all the oral approaches for protein delivery, for which I dutifully created a very very large spreadsheet with a ton of references. The assignment warped when the client asked us to give them bullet points for their pitch deck why “They will win, and everyone else has lost.” They got so insistent that They will Win that the partner walked away from the project forfeiting the rest of the fees. … fast forward, as expected, they weren’t any better and are now kaputsky.

  9. An Old Chemist says:

    An oral peptide drug for IBD (Irritable Bowel Disorder), Linaclotide, by Ironwood pharmaceuticals, last year it generated sales of $264 million:

    1. Barry says:

      Although Linaclotide reaches/binds its target without ever getting out of the gut (negligible absorption) it does survive proteolysis

  10. Cb says:

    0.2 mg desmopressin in tablets shows only 0.1 % bioavailability, but it works

    1. christmas chemist says:

      true. looking at the numbers I wonder, however, if we understand how it works?

  11. Simon Auclair says:

    Oral steak therapy saved my sanity!

  12. Lambchops says:

    Oral semaglutide sprung to my mind as soon as I saw the post title. An impressive effort for sure.

    The part about it “needing to be taken with water, first thing in the morning in a fasted state, with no food for 30 minutes” struck me when I first heard about this. I wonder if efficacy in clinical practice is a bit lower than in the trials due to people not complying with this?

    1. Hap says:

      I don’t know how other people work, but I could do that – take the med with water and then take my shower, shave, get the kids up, make lunches, and then eat. It isn’t a ridiculous schedule – for example, I thought the initial AIDS drugs had really annoying dosing schedules.

  13. Scott says:

    “It needs to be taken with water, first thing in the morning in a fasted state, with no food for 30 minutes. There are some gastrointestinal effects in some patients, as you might expect, and there is variability both from day to day in individual patients and between patients, which means that physicians need to work out the dosing in each case. But the drug is effective and safe – no worse than the injected formulation and certainly more convenient to take.”

    That’s really quite an accomplishment!

    Not that hard to take, either, as Hap mentions. Though I suppose the evil question would be “what happens when you take a bunch of other drugs at the same time?”

  14. Steve says:

    Semaglutide is a derivative of exendin-4, which was originally isolated from the spit of the Gila monster. Maybe it’s derivation in spit has something to do with oral bioavailability; always wondered what it was doing there in the first place.

    1. Barry says:

      Many snake venoms (presumable lizard venoms, too?) evolved from digestive enzymes which are necessarily proteins that resist (somewhat) proteolysis in the gut. But that says nothing about getting out into circulation

      1. milkshake says:

        snake venoms are not toxic when ingested, if you have good gums and no stomach ulcers. The old recommendation on treating snake bites – even though dubious – was to suck on the wound, to try to get some of the venom out

      2. loupgarous says:

        Since most venomous reptiles use venom to disable prey, it stands to reason their digestive tracts deal with their own venom while they’re reducing other peptides in prey animals they eat to amino acids. Unless I’m missing a subtlety here. Not saying they wouldn’t be hurt by their own venom when it enters their bloodstreams, but if it makes it as far as the gastric juices and proteases in the stomach without entering the circulation, the venom is rendered harmless, right?

        1. Barry says:

          Like the digestive enzymes they evolved from, snake venoms are probably pretty resistant to proteolysis. But since they don’t pass the gut barrier, they’re harmless p.o. not only to the venomous predator, but to other species.

        2. steve says:

          None of that answers why a GLP-1 antagonist would be part of a snake venom, though I as I alluded to in the beginning, if it evolved from a digestive enzyme it would at least explain its resistance to proteolysis. The systemic delivery isn’t intrinsic, it’s due to complexing it with SNAC, as Derrick indicated.

  15. Zemyla says:

    What about cyclic peptides? They’re stable against proteases, and the amatoxins show that they can potentially cross into the blood stream very easily. Can the ends of the insulin molecule be joined together without disrupting its bonding with the active site?

    1. mfernflower says:

      Death cap toxins ride on a transporter

      1. milkshake says:

        I would still give a try to cyclic peptides, especially if chemically modified to cover the NH bonds, and with some D-aminoacids into the mix. But proteins are pretty hopeless as oral drugs

        1. David Edwards says:

          While amatoxins are absorbed by the gut and survive the onslaught of gastric proteases, the related phallotoxins (also cyclic oligopeptides) are reported as not being absorbed by the gut. Possible research project?

          1. mfernflower says:

            Phalloidin if I recall isomerizes when exposed to HCL so that explains the lack of oral tox

  16. poor lab rat says:

    another year…..another set of rock bottom salaries for research scientists. glad the esteemed “leaders of science ” are doing such a wonderful job advocating for their profession.

Comments are closed.