Hacker News Reader

Comment by jcims | Hacker News Reader

Many cancers exhibit immunosuppressive qualities that allow them to co-exist in a healthy immune system. It seems possible that mRNA-based therapies could be used to interfere with those mechanisms (eg binding to PD-L1, etc).

The delivery mechanism for the mRNA appears to be the primary innovation and I wonder if it would be possible to target it at specific cells. In the cancer context if you could differentially target the mRNA delivery to cancerous tissue, the 'payload' could be more generally cytotoxic but only affect tissue proximate to cancer cells.

This seems like it's on the order of CRISPR for potential to change the landscape of medicine in the next 50 years.

View on HN

I agree. It's easy to overstate things, but over the next several decades I could easily see a computational personalized approach to fighting cancer: read the unique immunosuppressive markers, sort out what's tractable and what's not, then deliver a mRNA payload for that particular cancer for that particular person.

For the incredibly small amount that I know, it looks like a game changer. (Like everything else, though, there's a long road from theory to practice. Implementation is going be very tough)

Makes you wonder what it would take to get such a process (that produces a new medicine for every individual) approved...

The FDA has recently issued draft guidance for such: https://endpts.com/for-tailored-single-person-antisense-olig...

I think this is precisely what BioNTech has been doing with melanoma. Not in a few decades, they already had living patients with personalized vaccination against their own melanomas in 2019:

https://www.nature.com/articles/d41586-019-03072-8

If this immunotherapy gets developed further, the next generations of humanity may look at cancer in the same way that we look upon bacterial diseases: unpleasant, the threat of treatment resistance is there, but not the fearsome serial killer that they used to be.

Although given how we've squandered (and continue to squander) antibiotics, future generations may view bacterial diseases the way past generations did.

i think necesity here is for to block third world from acquisition of new antibiotic drug. there is where new superbug are arise. maybe it kill many there, but it also save many American here in longer term.

in such 3world country, misuse/overuse of antibiotic are make for development of resistant. then poor sanatation are make for it easy for to spread to other persons. https://wwwnc.cdc.gov/eid/article/5/1/99-0103_article

Please stop using "Third World" as a term — it is no longer accurate, the AP has eliminated it from its stylebook, and is widely considered offensive: https://www.npr.org/sections/goatsandsoda/2021/01/08/9548203...

"Developing Nations" or "Developing World" is better.

"Third world" contributes to exactly the kinds of stereotypes that this comment reinforces — the backwardness of poor countries, where doctors don't know better, science is ignored, etc. It's just not accurate; medical knowledge and competence is on the rise all over the world, among both professionals and volunteers.

no. i am not stopping. i disagree. i see not any inaccuracy. i do not work for AP and do not care what it say in style book, AP change style book many time for to please the pc persons. example, AP change style book for to say singular "they" are now ok: https://www.apstylebook.com/blog_posts/7

i have not in previous heard it "offensive" and am not changing because some person sayes it. i have been to many nation, those i call "third world" are backward. improvement you describe make them moving away from third world, not make term wrong, this is move to second then to first. i will not call different word because somebody say so.

Dude, it's not developing country misuse of antibiotics that's the problem. It's pouring it into animal feed in factory farms. It's doctors in wealthy countries giving their patients an antibiotic prescription for a viral infection because their patients expect them to "do something." You're not getting downvoted for using "third world" you're getting downvoted for speaking from a position of ignorance.

> in the same way that we look upon bacterial diseases: unpleasant, the threat of treatment resistance is there

Those treatment resistances are quite different. Resistant bacteria spread to other people, while cancer is almost always limited to a single patient. So it the treatment works for a certain fraction of cancers, it'll stay at that level, unlike bacteria which become increasingly resistant over time.

(I guess that in the very long term that might not be true, since natural cancer resistance will offer less of an evolutionary advantage. But that assumes that humanity will remain in a similar state as currently, which seems unlikely.)

I know, it is an imperfect analogy. But thank you for noticing and explaining my shortcut to other readers.

Cancers are very patient-specific, and the main threat with immunotherapy is that the targeted cancer adapts quickly enough to escape the immune system again.

>Cancers are very patient-specific, and the main threat with immunotherapy is that the targeted cancer adapts quickly enough to escape the immune system again.

(I'm just an interested party and don't have any formal training on the subject.) I'm sure there are technical terms for it, but from what I've read the lack of regulatory features in the replication process of cancer cells tends to accumulate more mutations and genetic damage through each generation. PARP-inhibitors, for example, help fight cancer by suppressing DNA repair enzymes and letting the cancer cells get into non-viable states more often than healthy tissue.

In general though, this genetic entropy/volatility creates a scenario where adaptations can happen quite quickly.

I was thinking about this very subject this morning, and I realised the mRNA stuff actually isn’t the exciting part.

It’s the packaging in lipid particles that is much more interesting. We can get away with this approach for vaccines because we (largely) don’t care where we deliver the payload to, just as long as we get mRNA to a cell where it can make the protein. Not sure about current formulation, but I read most LNPs end up in the liver from circulation.

The next level of tech is targeting the particles, and then it gets as tricky as other contemporary techs, because you want your targeting mechanism on the prticles to be something resembling a receptor ligand (protein/carbohydrate).

Manufacturing of those (and putting them on a lipid particle) is still a slog. If we figure out nice ways to do that (without reasonable purity) then it doesn’t matter what is in the LNP (e.g put gold particles inside cancer targeting particles and zap your cancer cells dead).

An alternative approach would be to learn enough to make mRNA “programs” that once inside a cell could determine cancerous vs healthy cells. There’s been research on computations using DNA, and many natural pathways do similar “calculations”, that it seems feasible to create a “if specific xyz protein/carbohydrate/etc in cell exceeds threshold trigger cell death pathways”. Then injection in a tumor mass should be sufficient. The mRNA program might only need to be partially accurate to be more effective than many chemotherapy’s.

You could maybe skip the cell death part, and instead trigger some existing telemetry systems that the immune system already watches for?

At least then if something goes wrong, you could still be put on immunosupressants

From what I’ve read both are options. There’s benefits to getting the immune system to target cancerous cells Such as ongoing protection. Directly killing the cells is a good way to reduce tumor mass by direct injection however. The first t-cell immune treatments almost killed the patients due to liver/kidney overload from cellular debris as the immune system reacted too strongly, so being able to in cull tumor mass in a controlled fashion likely has merit as well. Either way the “detection method” would need to be very well tested and robust I imagine. Though the first treatments could be for “last chance” treatments that’d enable testing the safety of the particular detection method.

Possibly naive question: if you can differentially target any particular thing to cancer cells specifically, why not just deliver poison to kill the cells directly? I.e. chemotherapy but without any of the nasty side effects.

Not naive at all. I didn't word my question well as my interest is broadly if this new technology gives us any more ability to differentiate the delivery, either by selecting for cancer cells or by only 'activating' within cells that are cancerous.

In general there are a number of 'targeted' therapies being developed that don't specifically target the cancer cell, but try to make life more difficult for the cancer by broadly impacting a cell's ability to hide, replicate and accumulate...features that the rest of a healthy human tissue is less dependent upon.

One version of what you are describing is called an antibody-drug conjugate. There are about a dozen ADCs approved and many more in various stages of development

I'd prefer to have the poison not have to be in the same process as the detection? The detector likely needs to enter or interact with the cell regardless of whether it's cancerous

Reply on news.ycombinator.com