20 Comments
User's avatar
Metacelsus's avatar

The most expensive part of manufacturing biologic drugs is usually not the actual biomolecule production, but rather the separation and purification. Thus, while growing drugs in farm plants may seem attractive, yeast-based fermentation is much more economical. And cell-free systems may be even better than yeast, so I'm excited to see how this area develops over the next few years.

Another thought: for certain kinds of drugs produced in farm plants, it might be possible to just have the patient eat the plant, skipping the purification process entirely. However, this would only be possible for drugs that are active when taken orally (which excludes most proteins, as these are degraded by the digestive system).

Expand full comment
Metacelsus's avatar

For example a recent article (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8542935/) stated, "Downstream processing of biopharmaceuticals is therefore usually a resource-intensive section of overall processing, being cited as high as 80% of production costs (and contributions of input mass) for monoclonal antibody (mAb) therapeutics produced using mammalian cell cultures".

(This paper is also quite cool, it discusses manufacturing antibodies on space missions with resource-constrained environments.)

Expand full comment
Asimov Press's avatar

Great comments, as always.

Expand full comment
Sam Harsimony's avatar

Yeah the purification process still happens with pretty old and expensive techniques. Another idea I've been pushing is to design the proteins to self-separate:

https://splittinginfinity.substack.com/p/unlocking-precision-fermentation

Expand full comment
Jordan Wolfe's avatar

I just wanted to say that I really enjoy both of your writing and pieces. As someone without a scientific or engineering background, you really find a nice balance when explaining concepts around biomanufacturing and engineering biology.

Expand full comment
Katia Tarasava's avatar

This a fascinating topic. I'd love to hear more discussions about alternative production platforms.

Expand full comment
Mohd Anisul karim's avatar

Enjoyed reading this. It made me (naively) think, if one day we had a little bit of GLP-1s engineered in our rice, potatoes, wheat etc, would we solve the obesity crisis?

Expand full comment
Jamie Coldwell's avatar

Cool idea, but would only work if the obesity crisis was caused by consumption of too many whole foods, which seems to be becoming more and more unlikely. Ultraprocessing of foods is common for economical reasons, which separates macronutrients for transport and reconstitution down the line, resulting in loss of minerals and many other molecules which were present in the whole food. e.g. you don't get a lot of vitamin C in hash browns. Pharmaceuticals would likely be lost during processing too. There is increasing evidence that ultraprocessing of foods contributes heavily to obesity - so the very nature of the possible cause of the crisis would unfortunately render this idea moot!

Expand full comment
Thomas Murphy's avatar

First and foremost, I want to commend the author for noting the tremendous ethical shortcomings of egg production. He understood pointing this out wouldn't help his argument, and doing so is all the more ennobling.

Second, a huge drawback to the egg-strategy is the paper cited a yield determined before any purification!! I imagine if you looked at pre-purification yields from crude S cerevisiae lysate you'd get incredibly high numbers as well, and the protein density of egg whites vastly exceeds that of yeast cells. This could make attaining a pure product even more difficult. I don't think the comparison is totally fair.

Expand full comment
Klein's avatar

Well not all of them - large biologics can pass the digestive system when protected by an heavy cellulose - some plants are magic - look at PLANTIBODIES.tech maybe

Expand full comment
Tom X's avatar

Unfortunately it seems as though one promising and very exciting company mentioned, Tiamat Sciences founded five years ago, is "saying goodbye". They had just begun to announce their first exciting partnerships, successfully expressing human prolactin for human milk with BIOMILQ. :(

https://www.linkedin.com/posts/tiamat-sciences_after-five-years-of-dedicated-effort-and-activity-7246890304492957697-F_ZX?utm_source=share&utm_medium=member_desktop

They mention "financial constraints". It seems like some of the other companies mentioned in the article, like Cirsium producing viral vectors in plants and Future Fields recombinant proteins in fruit flies, are continuing forth. Perhaps the wrong distribution of target markets/at the wrong time brought Tiamat to kneel? I'm not at all knowledgeable on these matters.

Expand full comment
Sofia Sanchez's avatar

For edible drugs, I also wonder how the dose could be controlled. Does that mean controlling gene expression, fruit growth? Idk

Expand full comment
Sam Harsimony's avatar

One piece I'm missing: why can't semaglutide be made with peptide synthesis and organic synthesis? Can't the amino acid sections of the drug can be made with solid phase peptide synthesis?

https://en.wikipedia.org/wiki/Peptide_synthesis

Expand full comment
Thomas Murphy's avatar

peptide synthesis is *much* less efficient than fermentation-based recombinant expression, speaking from experience

Expand full comment
Sam Harsimony's avatar

Oh interesting. Even for short peptides (~20-50 AA)?

It looks like companies will synthesize these short peptides for pretty cheap, so I'm trying to figure out what I'm missing.

Expand full comment
Thomas Murphy's avatar

If my lab needed a small quantity (~50ug) of peptides (say 20AA) quickly, we'd probably order synthesized peptide to cut out the labor of cloning, verifying, expressing and purifying all for a tiny amount. But when I need 500+ ug it would be crazy to do synthesis (though I mostly work with whole proteins so I would happily defer to a peptide chemist on these details).

Each AA added in the synthetic reaction entails a loss of yield (there's ways to move around this.. but that entails time and resources), such that if each step gives even 97% yield, 0.97^(50AA) = 21.8% total. So the more you need x the longer the peptide, the less practical synthesis becomes over recombinant.

Expand full comment
Sam Harsimony's avatar

Right, good points. I wonder why peptide synthesis doesn't scale to large masses?

One thing I'm considering is whether we can live with the fact that peptide synthesis is impractical past ~50-100 AA's and instead assemble larger structures in a modular fashion:

https://splittinginfinity.substack.com/p/modular-peptide-nanotechnology

It seems odd to me that people aren't trying this much but some groups are working with modular proteins/peptides:

https://www.nature.com/articles/s41586-024-07188-4

Expand full comment
Friede's avatar

This is so interesting. I did my MSc on cell-free expression and it would be fascinating to see it implemented on a large scale.

Expand full comment
Jim the Mighty's avatar

I would be interested to hear if Nico has any thoughts on a company called Plant Health Care that use bioreactors to manufacture peptides to use as plant stimulants.

https://www.planthealthcare.com/new-technology

Expand full comment
Asimov Press's avatar

Haven't heard of it, but I'll check it out! Thanks for sharing.

Expand full comment