Hacker News new | past | comments | ask | show | jobs | submit login
The curious DNA circles that make treating cancer so hard (acs.org)
119 points by panabee on Oct 2, 2021 | hide | past | favorite | 61 comments



Key Points:

* Scientists discovered a relationship between cancer and extrachromosomal DNA (ecDNA) in 1965, but considered ecDNAs rare and not worth investigating.

* Researchers subsequently learned that ecDNAs are common and central to the behavior of some of the most aggressive forms of cancer, enabling remarkably elevated levels of oncogene transcription, creating new gene regulatory interactions, and providing a powerful mechanism for rapid change that can drive very high oncogene copy numbers or allow cancer cells to resist treatment.

* Like Ptolemy’s flawed map that placed Earth at the center of the solar system, cancer researchers today may be analyzing genes and diseases with flawed maps.

Related Articles:

* https://www.the-scientist.com/features/cancer-may-be-driven-...

* https://chemh.stanford.edu/news/shining-light-extrachromosom...


What it doesn't say is whether the "DNA circles" or "double minutes" have any role in normal cell activity

Except, they seem to have that in yeast, which are distinguished by a need to reproduce fast like tumor cells. If they have been ignored all this time, maybe they are just a normal part of eukaryote biology?


Yes the recent paper [1] that confirmed they were circular mentioned the similarities to plasmids in bacteria too!

In the nematode world, we create extrachromosomal arrays of transgenes by injecting plasmids into the worm’s gonad that then get passed down to progeny; the expression we see is almost always variable and mosaic and not every cell ends up showing it.

I would guess this technique used for decades in the field is actually an induced version of these!

[1] https://www.nature.com/articles/s41586-019-1763-5


One of the articles posted suggests that weeds use DNA circles to quickly adapt to herbicides.


Would be interesting to see if the mechanisms underlying cancer's genetic agility could ultimately be put to use to search for solutions to other problems.


exactly. cancer is a scourge that must be eradicated, but on the way, we may discover the solution to other health issues.


Are cancer cells immortal? It would be very poetic that the key to cancer would also unlock amortality.


They're immortal in the sense that they no longer listen to tumor suppression genes that trigger apoptosis (programmed cell death). Based on our current understanding of senescence, that doesn't really help with the main causes of aging.


Why?

The presence of telomerase prevents telomere shortening in cancer cells and gives the ability to divide an unlimited number of times in a laboratory cell culture plate.

But the aging process is not reversed, so a multicellular complex organism would eventually die.



I, like so many others, had a sudden motivation to develop my understanding of cancer.

Once you’ve read your hundredth paper and followed as many clinical trials, it takes on another form. Near sentience, its the ultimate survivalist. It’s too bad it has such a devastating outcome on loved ones, we’d all admire it otherwise.


Sadly, I agree 100%. Time and again I see labs/papers/companies trying to "target" cancer cells for specific killing. What we are all awaking up to now in the field is the only way to reliably achieve durable responses you need to a) have a functional immune system, b) somehow convince it to kill cancer cells via surface antigen recognition/MHC all without getting checkpointed. It's a super fucking hard problem and I makes me so angry; but I agree there is a bit of awe in how clever/sneaky/insidious it is.


I’m not a biologist, but how I imagine it is that cancer is just evolution + entropy. We can obviously get a lot better at controlling it, but it will always be there because there’s always a way for one cell to go rouge and gobble up resources and make tons of copies to the detriment of the whole. Biology already puts tons of roadblocks in the way to this, and we can surely add more, but at the end of the day it’s such a huge incentive so things will go find the easier faster way to chemical equilibrium


We probably won't kill cancer through one miracle cure, but probably through death by a thousand cuts.

That and we probably need to repair the immune system. The incident of cancer increases with age, after all.


I believe, as you hint towards, that the ultimate way to reduce cancer risk is to cure ageing.

Most cancers are rare in your 20s and 30s, then get dramatically more common every year after from your mid 40s roughly.

To me, this suggests that the mechanisms that stop you getting cancer (and your body/DNA repairing in general) are themselves failing.

Of course, if everyone was 25 forever we'd still all eventually die of cancer, but it would take a lot longer than the current state of affairs.


Death by a thousand cuts is vastly saner than "Nuke it from orbit. It's the only way to be sure." when your body is the battlefield.

But most people don't really want to hear that. We hear "big, scary death label disease" and we want someone to promise they can nuke it from orbit. No one wants to hear that we need to care for the body and foster its health while battling this "demon."


How do you improve the immune system to convey the improvements we’ve made with mRNA to it?


We haven't really made any improvements to the immune system with mRNA, we're just giving it some training film to be ready when the fight comes.

An immune system that has a hair trigger to new antigens would likely cause more problems than it solves. Some kind of ability to have contagious antibodies might be an actual improvement but the fact that we don't see that in nature tells me it has downsides that are difficult to see.


> Some kind of ability to have contagious antibodies might be an actual improvement but the fact that we don't see that in nature tells me it has downsides that are difficult to see.

My biology education is a bit weak, but off the top of my head, autoimmunity being contagious sounds like a likely side-effect of contagious antibodies.


We could sample the actual cancer in the patient and custom print an mRNA cure to that specific person for maximum accuracy.

The Pfizer/BioNTech vaccine was created in hours after Covid was sequenced. So this might not be a too far fetched idea.


Easier said than done, but yes most mRNA companies are trying the non-personalized and personalized cancer vaccine idea. The issue is the immune system is held back, like a big bouncer holding you back in a fight. Checkpoint inhibitors remove that bouncer, but now you're drunk and throwing fists all over the place, hitting both your friends, the bouncer, and the hooligans at the same time...


The incidence of errors occurring and the incidence of error correction failing increases over time

I think that has much less to do with the immune system, although I think that can be a factor too

It’s just probability from a flawed replication process


The interesting thing going against your fear is that there are organisms like blue whales that do not die of cancer, or at least for them it is an exceedingly rare disease. The reasons for this are not yet understood, and they may not be entirely applicable to humans (e.g. the sheer size of a blue whale may play an important role, since cancer would take much, much longer to grow to any noticeable proportion of the whale than in a human-sized organism), but at least it proves it's not completely impossible to have life without cancer.


I can't actually find a good source for blue whales not dying of cancer. I do see lots of sources claiming the rates are way lower than in humans, and that is indeed interesting. It's evidence that we could do better at suppressing cancer, but I'm very convinced that cells that can evolve eventually deciding to go rouge isn't something that be precisely stopped. Just made less likely


Elephants have 20 copies of the TP53 tumor suppressor gene.

https://www.nature.com/articles/nature.2015.18534


Every time I read about it I'm baffled we are not doing the same for humans with gene editing.


I'm sure there are many side effects that could only be found and fixed on a mountain of dead babies. For example, if I understand correctly, this makes elephants very susceptible to radiation sickness, as their cells are on a hair trigger to destroy themselves at any sign of genetic damage.


I can envision a moment where non invasive intervention and monitoring can be precise enough to stop being a threat. Other than that yeah, I think it's a naturally emergent biological phenomenon that cannot be removed.


There is no sentience. It’s a trillion monkeys banging at a trillion typewriters.

If anything treating cancer is more akin to antibiotic resistance. One therapy usually just selects for the cancer that is resistant, so you relapse.

And cancer cells have notoriously fragile genomes. I remember reading a paper that sequence 1,000 breast cancer samples. There were about 800 different mutations that led to the cancer. Very little similarity between patients.


> There is no sentience. It’s a trillion monkeys banging at a trillion typewriters.

I get it but the human genome has ~4^3,200,000,000 possible combinations. Even single genes have 4^1000 possible combinations...that's a number with 600 zeroes in it. That's a trillion^50. It's just odd that the mutations always seem to zero in on the right spot pretty damn quickly.


It's a wonder we don't all have cancer. Our large functioning multi cellular bodies are a marvel.


Evidence suggest humans “get” cancer way more frequently than we’d expect. Cells are programmed to “self-destruct” if things go wrong and our immune system is primed to destroy cells that express certain proteins associated with uncontrolled growth.

It’s really the edge cases that get through.


I am curious if you have read, and of what you think of Jason Fung's book Cancer Code. Is it a fair overview of the papers you have read? What level of understanding does he achieve?


Survivalist? You do realize once the loved one dies,the cells have to die too. A bit counter-inituitive..


It is a tragedy of the commons scenario. The individual actors have no conscience or plan, of course, but it all amounts to the same: some cells can morph into freeloaders and those that morph into the most aggressive replicators "win" in the short term.


Unless perhaps if the cancer is transmissible.


Yes, quite. Sometimes it's too successful and outstrips its resource. Sounds familiar.

If you have a word you prefer I invite you to use it in the future.


I have been using "cancer", but "carcinoma" and "tumor" are common.


Those terms are not interchangeable.


Some cells survive the death of their organism. Like the HeLa line of cells. There's an interesting and tragic story behinds it too...


That’s because cancer only happens when it manages to survive or bypass every process in the human body that would normally prevent it.

It’s the pinnacle of (natural) selection bias.


I also had that moment where you're somehow in awe about all the reinforcing intricacies that allow cancer to persist. It's an odd form of anti-life.


Really fascinating! I wonder if these cells all have defects in innate immune sensing of cytosolic DNA. Like, do cGAS/MAPK/etc. require linear for recognition like RIG-I does for dsRNA or are they all mutated/shutdown in these cells. Would be interesting to see the innate immune transcriptome with and without ecDNA.


one of the best accelerants for science would be empowering non-researchers like you to hire researchers to conduct experiments like this -- or ideally conduct experiments yourself.

equipment, material, and people costs are prohibitively high for now. hopefully one day.


Ideas are cheap. Every biologist has way more ideas for what to investigate than they could ever get grants to look into. Grant committees, likewise, see applications for way more research programs than they have money to fund. (This is like startup ideas, founders, and VCs.)

But grant committees are controlled by the older members of each sub-field. They are biased toward what seemed important to them when they were young, but also against what experience has taught them tend to be dead ends. They want questions explored that have gone wanting for decades. Young researchers will eventually graduate to grant committees, and fund what they wish had been looked into decades before.

We can see systematic failure in, e.g., Alzheimer syndrome research, still funding suppression of amyloid plaques and tau tangles decades after every such avenue was found to fizzle. But we don't know a better way to organize evaluating grant applications. We would not know how to evaluate any change proposed. And, we don't know how to change committees over if we found a better one.

Each sub-field has its own dynamic, independent of all the others. Probably some have evolved a method that works well for their sub-field, for now, but would certainly be wrong for the next sub-field over, and will be wrong in its own in ten years. They are all playing catch-up all the time. The only thing that keeps it working at all is that they are run by smart people who have had to solve many different hard problems, and those people retire on a regular schedule.


agreed. however, lowering barriers so more people can help also means potentially clearing the idea backlog faster. the startup analogy is right: sometimes a promising idea requires another implementation (experiment in this case) for validation/invalidation. similar to ecDNAs, novobiocin, and other cases.

science funding is indeed broken.

only 7% of NIH grants go to researchers under 40.

tesla, apple, and many impactful companies were started by founders under 40. watson, crick, and currie were all under 40 when making seminal contributions to science. imagine how many breakthrough ideas are gated in science right now.

the NIH is amazing but not allowed to take risk: how can we fund more unestablished scientists?


industry. i'll just say it plainly. that's where you can fund more unestablished scientists. the vc world, the smart ones, are now funding moonshot ideas from younger scientists in ways academic science funding methods will never match. if you're a young scientist with crazy ideas, try to hitch up with vc's that run an incubator. the first step, though, is to move to boston.


i appreciate the vote of confidence! to assuage your concerns, i'm part of the research planning team at a major biopharma. so funding, executing, and analyzing these kinds of experiments is my day job : )


in the same way FAANG engineers build impactful projects on the side, it would only help science if we could empower talented people like you to conduct your own "side" experiments.

let's hope the day comes soon.


it boils down to costs. spinning up some code on your laptop is effectively free. of course your time, electricity, laptop depreciating, etc. cost money, but it's negligible if you're just taking a break from earning your living in that space. to execute a biological experiment, you need to move liquids around which adds a ton of costs for handling said liquids. also, you need to indirectly analyze the results, which requires equipment to amplify/change the output to a human readable format.

i've returned to issac asimov's response to innovation: find the smartest people, pay them a stipend contingent on completing some quick menial task, and then leave them be for a few weeks. out the other end you'll have innovations.


this is right. driving down costs is crucial in science (and other fields actually). as software grows more sophisticated, this will happen.

startups like https://mbcbiolabs.com are starting to lower lab costs, and this trend will only accelerate over time.

what are your favorite resources/labs for tracking research?


Is there ongoing research looking at adapting cancer fighting mechanisms of large mamals like elephants and whales? I read somewhere they had the lowest occurrence of cancers of all mamals.


Can't we develop some protein that locks onto these circles and then triggers cell death?


The issue is that a single protein might not be specific enough ie. you will kill lots of wrong things if you are too aggressive.

Some people have been working on combining two (or more) factors in order to only activate when in the right spot: https://www.science.org/content/blog-post/two-steps-activati...

My guess is that this is how the body actually works.

At the chemical level, the human body has many highly individually specific things (like enzymes and receptors).

However, at the system level, it's probably easier to take two slightly specific things and combine them to make something that is specific enough but not too brittle.


I expect the tricky part, or at least a tricky part, is distinguishing an ecDNA circle from healthy chromosomal DNA. More speculatively, bacterial DNA looks kind of similar, and you don't want to indiscriminately kill your gut bacteria.


The mitochondria also have circular DNA, exclusively. I wonder how similar these forms are.


Wouldn't that protein need to penetrate the cell wall? You'd need something more like a virus.


I guess the delivery mechanism is the next step in the development of such an agent.


or more likely find an organism that already has.


Which type of cancer?


They particularly mention certain brain and colon cancers. But that is just an artifact of specialization in the researchers. It is likely that some of the cancers that affect any body part exhibit the same phenomenon.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: