Hacker News Reader

Hacker News

Comment by ethbr0 | Hacker News Reader

Full thread

ethbr0

This pulls together so many advancements that it would make a 90s oncologist cry.

1) Being able to rapidly sequence the genomes of cancer cells to detect common mutations

2) Computationally simulating those mutations to look for viable T-cell targets

3) Custom building T-cell receptor proteins [0] capable of recognizing those targets

4) Inserting those custom receptor proteins into the patients' own T-cells with CRISPR

Truly, we're living in the days of future medicine...

[0] https://en.m.wikipedia.org/wiki/T-cell_receptor

View on HN

I might not understand it correctly, but the t4 helper cells don't directly work on antigens/pathogens.

They do stimulate macrophages, and activate b-cells.

This from some cartoons, and the associated book by Detmer, Immune.

Only tangentially my field, but both cytotoxic and helper T-cells bind on antigens, and indeed this work CRISPR edited anything with a TCR to their (3) preferred, cancer-targeting receptors (for each patient).

The press release has a few more details (in addition to the Nature paper): https://pactpharma.com/news/pact-pharma-reports-data-from-fi...

Your CD4 T cells are basically the directors of your immune system. They license CD8 T cells (T killer cells), B cells and recruit NK and immune cells to sites. The type thought to be most anti-cancer are the Th1, which recruit CD8 T cells and NKs, and are optimised for intracellular pathogens. Cancer can be thought of as an intracellular pathogen because it harbors mutations.

> 1) Being able to rapidly sequence the genomes of cancer cells to detect common mutations

Only of the cancer cells? How are mutations detected in the 0.4% of the DNA that differs in different human beings? Can't there be non-cancerous changes in the other 99.6% that are present in all cells? Is it just a matter of cost, sequencing both too expensive?

My initial guess would be that they would compare cancerous vs non-cancerous cells within the same human being

The article mentions: "...sequencing DNA from blood samples and tumour biopsies, to look for mutations that are found in the tumour but not in the blood."

This is pretty common for most cancer genotyping tests. (Sometimes they compare tumor with saliva or some other "benign" source, but the principle is the same.)

> Custom building T-cell receptor proteins [0] capable of recognizing those targets

How does this part work? Is this some sort of selection of cells in the lab, or done in silico?

There are processes such as this one. Sequencing specific T-cell DNA gives you the fragments you need to insert into the target genome.

I have a theory that every 10-15 years or so we look back at the previous 10-15 years in medicine and wonder how we lived through it at all.

It’s good progress but I want to say it’s still not a magical cure - the fundamental idea is flawed - the tumor can and likely will in many patients just mutate the immunogenic epitope.

How fast can a tumor mutate to resistance? If we're iterating our therapy faster than this cadence do we win the battle?

An iterative approach could work, yes. My guess would be that it takes months to grow resistance to a new therapy.

I haven't seen evidence of cancers re-mutating mutations. They tend to develop new mutations (as in the clonal evolution hypothesis), and they can completely change their cell surface expression. But I suspect that they are unlikely to mutate an already mutated gene.

I guess for the treatment to be most effective you need to target some of the earlier mutations, rather than a small clone.

From my perspective, and not having read the paper, I thought the technology to predict what TCR would bind a particular peptide on a particular MHC-1 was not there yet.

> But I suspect that they are unlikely to mutate an already mutated gene

Could you elaborate? I'd like to understand what you mean, as I don't work on onc. Aren't recurring mutations in response to treatment in f.e EGFR is the reason we keep developing multiple generations of small molecule therapies for it?

Yes, I didn't mean impossible, I just mean compared to accumulating new mutations elsewhere, and compared to downregulating surface proteins (which can be the issue in CAR-T).

Different treatment, but tumors receptive to EGFR-treatment almost always develop secondary resistance to treatment by additional mutations.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4367712/

Yes good point.

Cancers have higher mutation rates and also constantly rearrange their genome including deletions. Given that they’re looking for Novel epitopes it goes to reason these are not high copy number hence deletions can also remove them fully. It’ll be hard for me to believe that a mutagenic cancer (which is a given here since you’re looking to treat cancers with actual mutations) will not eventually gain resistance to this therapy.

Yes you are right, I should have said that re-mutation of mutations happen far less often than accumulation of additional mutations.

Novel epitopes should be high copy number - and driver mutations will be present in 80-100% of the cancer cells. It depends how many cancer cells you get in your biopsy that you sequence I guess.

It is easier for a cancer cell to mutate or remove a cell surface protein than to mutate the same mutation targeted, but you are right - that can happen and I'm sure will be a form of treatment resistance for these types of treatment in the future.

KRAS does this. Doesn’t necessarily invalidate your point, but it is interesting.

https://www.sciencedirect.com/science/article/pii/S155608642...

Thanks for breaking out all these crucial steps!

Do you know of any journal articles that cover step 3? I'm very interested in how this whole process works and can't seem to find much (paywalled at Nature for this one.)

Reply on news.ycombinator.com