5 ms·

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 u

by jcims 6y ago

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.

umanwizard 6y ago
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.
- jcims 6y ago
  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.
- wgolsen 6y ago
  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
- 8note 6y ago
  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
DanielBMarkham 6y ago
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)
- perlgeek 6y ago
  Makes you wonder what it would take to get such a process (that produces a new medicine for every individual) approved...
  - IdoRA 6y ago
    The FDA has recently issued draft guidance for such: https://endpts.com/for-tailored-single-person-antisense-oligonucleotides-fda-offers-guidance-on-nonclinical-testing/ https://endpts.com/for-tailored-single-person-antisense-olig...
- inglor_cz 6y ago
  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 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.
  - dhosek 6y ago
    Although given how we've squandered (and continue to squander) antibiotics, future generations may view bacterial diseases the way past generations did.
hirenj 6y ago
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).
- elcritch 6y ago
  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.
  - 8note 6y ago
    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
    - elcritch 6y ago
      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.

by jcims 6y ago

umanwizard 6y ago

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.

jcims 6y ago
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.
wgolsen 6y ago
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
8note 6y ago
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

DanielBMarkham 6y ago

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)

perlgeek 6y ago
Makes you wonder what it would take to get such a process (that produces a new medicine for every individual) approved...
- IdoRA 6y ago
  The FDA has recently issued draft guidance for such: https://endpts.com/for-tailored-single-person-antisense-oligonucleotides-fda-offers-guidance-on-nonclinical-testing/ https://endpts.com/for-tailored-single-person-antisense-olig...
inglor_cz 6y ago
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 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.
- dhosek 6y ago
  Although given how we've squandered (and continue to squander) antibiotics, future generations may view bacterial diseases the way past generations did.

hirenj 6y ago

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).

elcritch 6y ago
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.
- 8note 6y ago
  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
  - elcritch 6y ago
    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.