Hacker News Reader

Comment by EMM_386 | Hacker News Reader

> MTDH is important for most major human cancers, not important for normal cells, and it can be eliminated with no obvious side effects

I'm not fully buying that.

If you look at "Function" on Wikipedia, it seems to have some important roles, including angiogenesis (which is still important outside of cancer). It's not like evolution would select for something that is useless outside of helping cancer.

I'd be curious what the side effects are in humans. How can it be claimed that it can be eliminated with "no obvious side effects" when they are still years away from human trials?

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

View on HN

Disclaimer: I’m not in these fields but I am curious enough about how the mechanisms work at a high level that I feel confident about the rest of this post.

Two things to keep in mind:

1. Biology has never selected genes to maximize longevity, only to maximize healthy breeding. There are very likely tons of genes that make sense as a growing child but are harmful at a later age, e.g. a post menopause/vasectomy adult does not need the genes to create gametes but those same genes may be causing cancerous issues for some.

2. Gene therapy (at least CRISPR-Cas9 like things) are in theory reversible[0]. Diagnose the cancer. Delete the notable gene or just make a point mutation to disable it. Treat/wait the cancer into remission. Re-add the notable gene or re-enable it with a point mutation. Hopefully such a strategy would minimize any long term effects and eliminate the cancer.

[0] There may be cases where a bio-engineer could design an effective yet irreversible edit. However, that seems like it would require serious effort and I doubt it would have enough value to be approved by the FDA for use.

> Biology has never selected genes to maximize longevity, only to maximize healthy breeding.

That is assumption which is hard to validate. There is a case for longevity as children with longer living dedicated parents in the context of abundant resources will out survive those that are ditched by their parents.

> Delete the notable gene or just make a point mutation to disable it.

The article say that effect is the same with deletion and drug in mice. Why would you delete the gene when you can temporally prevent its function ?

> Delete the notable gene or just make a point mutation to disable it. Treat/wait the cancer into remission.

Yet nobody did nothing remotely close in humans, even for stuff that looks like highly beneficial in all contexts such as GULO gene.

I agree with your sentiment, but this condition...

> in the context of abundant resources

...has only been true for a tiny number of generations and so likely has not had a large evolutionary influence.

Abundant is probably too strong word.

Obviously the system can't work to the extreme due to the limited resources, but there is lot of possible contexts that can last for a long time.

Also, evolution isn't "complete". It's not done optimising. And sometimes, some things evolve and there just isn't enough evolutionary pressure to remove stuff that does't hurt. Men have nipples.

> Men have nipples.

Just sayin', that has little to do with "remove stuff that does't hurt", but the fact that there is very little difference between men and women besides sex steroid levels. In particular, women do not have any genes men don't. The Y-chromosome adds the tiniest modification to gonadal development in embryos and that's it; androgens do the rest. With the right hormone levels, men can develop fully functional breasts; without estrogens, women don't.

Cancer definitely hurts.

>>> 1. Biology has never selected genes to maximize longevity, only to maximize healthy breeding. There are very likely tons of genes that make sense as a growing child but are harmful at a later age, e.g. a post menopause/vasectomy adult does not need the genes to create gametes but those same genes may be causing cancerous issues for some.

This is being debated, in the sense that there is a relationship between longevity and breeding. How ? For a human to be able to have kids every 2-3 years - instead of 6-7 in apes - older individuals have to be healthy enough to bring a calorie surplus to the mothers.

Adding to point 2, gene therapy‘s effects seem to wear off after a couple of years.

The MTDH gene clearly has some utility in the human body. We know that because natural human knockouts are under-represented in the population. The Gnomad project has sequenced very many people's DNA, and they have been able to work out a formula for how many loss-of-function mutations they would expect to see in a gene (which mainly depends on the size of the gene). For the MTDH gene [0] they would expect to see on average 30.5 LoF mutations, but they only actually see 7. This is reasonably strong evidence that knockouts of this gene are selected against - that is, having two working copies of this gene conveys some advantage of some kind.

Knocking out the gene and gaining protection from cancer is likely a bigger advantage than whatever you lose, especially if you currently have cancer. Also, the proposed drug doesn't knock out the gene - it just prevents the protein from performing one of its interactions. Side effects are likely to take a while to work out, but if they're milder than the side effects of chemotherapy (which isn't hard) then that's a win.

The other thing I would like to point out is that humans and mice differ. It looks like most of the work they have done so far has been in mice. Knockout of this gene in mice doesn't seem to cause any ill effects. Working out whether the same is true in humans is a long and very careful process.

[0] https://gnomad.broadinstitute.org/gene/ENSG00000147649?datas...

Angiogenesis is going to be important in wound healing, I suppose, but compared to the current side effects of chemo this seems relatively minor.

I agree that "no obvious side effects" seems excessive if you interpret that as the expected result in humans. Still, on paper this looks much more specific that anything we have.

Thats a pretty important consideration here. If I was looking at chemotherapy & its side effects vs alternatives... the side effect list from these angiogenesis inhibitors looks a lot more focused, the sort of thing you can develop a focused regime around the monitoring and management of.

Considering that people still get prescribed Thalidomide to limit angiogenesis as a cancer treatment, and we know about its side effects (lets ignore birth defects for now), I see no particular difference with this drug.

You're challenging someone with 15 years experience studying this particular gene... by quoting Wikipedia?

That someone has skin in the game and a reason to manipulate other peoples understanding for his own gain though. Not saying he is, but I dont think the original comment was unjustified

You acting like posters on HN don't have skin in the game as they try to appear smart in front of their peers.

“Do your own research”

Ah the classic appeal-to-authority-fallacy ...

https://yourlogicalfallacyis.com/appeal-to-authority

From the link you posted:

> It's important to note that this fallacy should not be used to dismiss the claims of experts, or scientific consensus. Appeals to authority are not valid arguments, but nor is it reasonable to disregard the claims of experts who have a demonstrated depth of knowledge unless one has a similar level of understanding and/or access to empirical evidence.

At no point have I dismissed any expert.

Thanks for the quote tho, whats your point?

The point is quite clear. It's not an appeal to authority. Experts are generally more qualified to express opinions on their topic of expertise, but this doesn't make them immune from criticism. It does, however, make it a bit harder to call their view into question, and calls for a more sophisticated approach, usually by referring to problems with their evidence or absence of consensus.

Suppose an expert in a field has offered an opinion on their topic of expertise. You criticise that view based on unreliable evidence. I point out to you that your evidence is unreliable, and that the expert's view should be preferred due to its higher reliability. It's no answer for you to invoke the appeal to authority fallacy. You need to show that your evidence really is reliable, or why it's wrong to characterise your evidence as unreliable.

Suppose on the other hand that an expert in a particular field offers a view outside of mainstream consensus. I say that his view should be accepted. You say his view is not mainstream consensus. I say that he's an expert so he must be right. This is an appeal to authority.

I think it's important to distinguish the authority of expertise from the authority of power.

This isn't an appeal to authority, though. Also note there's a fallacy on there about using fallacies to discredit someone.

> It's not like evolution would select for something that is useless outside of helping cancer.

Cancers often have genetic basis so in a sense evolution has already selected for it in some people or at least not selected against it.

IANAdoctor/biologist but i assume diseases that tend to strike after child bearing age have less negative selection pressure than other diseases.

> IANAdoctor/biologist but i assume diseases that tend to strike after child bearing age have less negative selection pressure than other diseases.

It gets even better: there is selective pressure for adults that have borne children to die so that their children can procreate. First noted in pea plants that send a chemical signal to kill themselves up from the pod, can be extended to animals in things like cancer rates.

Except that humans are not pea plants and having grandparents play an active role in child rearing significantly improves the survival odds for their grandchildren (who carry a quarter of their genes).

Humans aren’t peas. Pea grandparents don’t help out in child rearing but human grandparents do. There is no indication that humans face selective pressure to die early.

>There is no indication that humans face selective pressure to die early.

While I do agree that pea plants are different, human grandparents may help in child rearing, but they may also consume time and energy for care.

We're social animals that would not kill our own grandparents for having more food/time available. But nature might find it's useful for procreation if they don't live too long.

Elderly care is a huge time/energy sink. And longevity makes the input/output unfavorable in a lot of cases.

If you're in an environment like our ancestors where food was most of the time scarce, then that adds weight to this equation.

There is evidence that earlier death can help animals, so it is appropriate to extend it to all animals barring evidence to the contrary.

Long lived, already procreated animals limit the amount of evolution that can take place by drawing out the time between procreation events. So if you have 2 species, one which dies earlier but has more generations, you expect them to outcompete a longer lived species.

I didn’t read the article fully but the parts of those two papers that are public and available to me (see Extended data Fig.4 comment at the bottom of the first paper), those do not talk about turning off a gene that encodes MTDH.

The paper is talking about a compound (two compound candidates) that inhibit binding of two proteins: AEG-1 (encoded by MTDH) and SND1, those compounds do that by binding to either AEG-1’s or SND1’s site where otherwise SND or AEG-1 would have bind, and they likely do this with stronger bond than SND1 and AEG1 would have had.

So in short, this compound does not leave you without some vital protein but disrupt formation of some bonds with SND1 protein.

No obvious side effects in mice with knockout gene and using the drug.

The drug doesn't disable the gene but make the protein unable to do its function by preventing the binding with another protein. If it can make cancer chronic disease, any side effects it may have will probably be less important then death.

A lot of proteins are involved in angiogenesis. Not to say it isn't important, but the body has many redundancies to make sure small errors don't break the whole system. That alone doesn't necessarily mean much, hence the MTDH knock out mice(model mice with the MTDH gene turned off) shows no obvious side effects. I suppose we'll see with time whether that holds up when applied in practice.

There’s also a lot of redundancy in biology. A protein could be essential for cancers and non-essential for healthy cells. That’s the holy grail.

I know nothing about biology, medicine or evolution, but I don't think that's how it works. Lots of people inherit and pass on all kinds of bad genes that do terrible things to them, even kill them. And for such people, switching them off would be better pretty much irrespective of any side effects. (to the degree there even are any side effects at all)

It’s one thing for a “bad” gene to stick around for a few generations or even spread through a species. It’s wildly different for a “bad” gene to be shared by species who’s common ancestor was 60+ million years ago. On that timescale anything conserved is doing something useful.

Useful, yes, but for whom? Not necessarily the individual. The ultimate goal of the genes is to propagate themselves, possibly in another body.

In the contemporary research of aging, the hypothesis that "aging is actually at least partly a programmed process and some genes useful in youth may turn against us later" is taken seriously. There seems to be an intriguing negative correlation between age of sexual maturity and lifespan in mammalian species.

> some genes useful in youth may turn against us later

That’s very much still useful for the individual, from an evolutionary standpoint. Of course finding out what something does isn’t the same as finding it needed for you. People get vasectomies which are hardly something evolution is going to do.

The extreme prevalence of testicular and breast cancer are likely the kind of trade offs that modern individuals don’t want, but they are a likely a consequence of some useful feature for our ancestors evolutionary environment.

> I know nothing about biology, medicine or evolution, but

well, by all means, don't let that stop you

this site is FULL of people who make wildly unqualified, unsourced, incorrect claims and passing them off as fact

but you have to have a dig at the one guy who clearly and openly admits they don't know anything before continuing with "but I think"

wow

Entirely possible that while it’s involved in angiogenesis, it’s not a critical component of that process, and angiogenesis can continue just fine without it.

The statement “not important for normal cells” doesn’t mean it serves zero purpose outside of cancer, all it means is that it’s not critical for body functions.

Might be important for sustained metabolism, but it might be eliminated for long enough to starve the disease.

I don’t know, but I do know the reality provides more combinations than I can imagine, especially when presented with binary and mutually exclusive scenarios.

Reply on news.ycombinator.com