World's first plastic antibody works in mice | Hacker News Reader

World's first plastic antibody works in mice

newscientist.com

42 pointsbshep8 comments

Man, that is amazingly cool. Does this have implications for other problems that have to this point been limited by the difficulties of predicting the outcome of protein folding?

Nice! Hoping to see the "no known antidote exists" tag disappear by the next decade or so.

Well, not to be a downer, but there are a number of caveats to be attached to this discovery. First, melittin is over 7.5 kDa which, while small for a protein, is still much, much larger than your typical household toxin. It is also larger than the smallest molecules that antibodies can be trained against.

Second, the interactions here appear to be entirely steric (that means size/shape related). One of the advantages of natural antibodies is that they can bind their targets via H-bonding, ionic interactions, van der Waals interactions, and hydrophobic interactions in addition to steric interactions.

Third, clearance by the liver runs the risk of just having all of the toxin concentrated in the liver, and it's not clear that the toxins will remain bound to the bead until they are inactivated. Certainly, if we're talking about a snake-venom which is a neurotoxin, and the "down-side" to taking this and not dying is that you need a liver transplant, I think it would still be useful, but this is not any sort of panacea I would think.

Edit: Also, it seems like the "imprinting" would require a fairly large interaction surface with the target molecule. This means that its probably not even very useful for molecular biology work, since natural antibodies can distinguish between proteins with very small differences and only require half-a-dozen or so amino acids for a match. Really, I think the lesson of this article should be just how amazingly cool natural antibodies are, and that it's taken until now to maybe come up with something that could possibly replace one of these uses.

> clearance by the liver runs the risk of just having all of the toxin concentrated in the liver

Could you bind the toxins, then do a full body blood transfusion?

Well, my training is as a biochemist, not a doctor, but my understanding is that complete transfusions are extremely difficult or outright impossible (there are a lot of tiny little capillaries for things to get stuck in...). Not to mention that the liver is really good at what it does, and it would probably have soaked up most everything by the time you got to the hospital.

The real concern is not that it ends up in the liver, but that the plastic antibodies don't bind as tightly as a natural antibody so that, once in the liver, the toxins slowly "leak" out and poison the liver.

Honestly, we've gotten really, really good at manufacturing antibodies against almost anything. Antibodies are a mainstay of modern biomedical research. Sure, they're expensive, but other than cost I don't see anything that plastic antibodies could do that natural antibody can't (and even in terms of cost, if you need one molecule of target to serve as the "stamp" for each plastic antibody, these are probably going to be more expensive since, once you've made the initial antibody, scaling up natural antibodies is mostly a solved problem.

I think the term antibody here is mostly hype. The plastic nano-particles don't act as antibodies beyond being specific for binding the toxin protein. It is still a cool technology demonstration but "antibody" overstates the case. To quote the article:

"However, Holliger doubts whether they could perform other important functions of natural antibodies, such as priming the body's immune system to fight future infections. Unlike natural antibodies, they are not equipped to communicate with other cells and components of the immune system"

I believe there has been research taking these sorts of molecule imprinted polymers beyond strictly steric interactions. I'm fairly certain it was Nicholas Peppas that I saw talking about it at a conference; sorry I don't recall a more specific name closely connected to it. Iirc, they were first binding several other molecules to the target molecule, then enveloping the target plus extras in the polymer, and then getting rid of the target molecule but not the others. So not only is the shape of the target imprinted in the polymer, but also, around the target-shaped hole are several molecules that would bind to the target were it to get in the hole again.

Now that I think about it, I don't recall if it helped with the kind of selectivity wanted for this "antibody". They were more concerned, I think, with using the binding of the target to the extra molecules around the hole as a trigger for degrading the polymer.

Yes, it's nice and all that it works in mice.

But does it work in the human body?

Reply on news.ycombinator.com