[iandanforth]

The most interesting point of the article for me is that this vaccine appears to be a series of shots which act as a curriculum for the immune system. Each one slightly different and targeting a different stage of B-cell development. I've never thought of vaccine series working like that, so it was a new and impressive idea.

[dist-epoch]

The way the HIV evades effective antibodies is by tricking the immune system to generate antibodies for fake decoy targets that the virus will immediately mutate.

So each turn, B-cells are presented with the latest version of the virus, they generate various antibodies to the various parts, and are graded at the end by how well the generated antibodies bind to the virus. The problem is that you have 1 million cells which bind strongly to the fake decoy targets and 1 cell which will bind not as strong to the real effective target. So this 1 cell never gets "promoted".

What this "germline targeting" multi-shot vaccine tries to do is to introduce a series of targets that stimulate that 1 in a million cell which will attack the right part of the HIV virus, so it gets "promoted", so if the real virus appears, the body will still go for the decoy targets, but will also generate a lot of these really effective 1 in a million cells, which got "promoted" previously by the vaccine.

To be more precise, you need to "guide" a lets call it B1 cell that all of us have in our repertoire to mutate into B2 then B3 than B4, because this B4 version will be capable of creating the right antibodies for HIV, the issue being that the intermediary states, B2, B3 are not naturally promoted so you very rarely get to the B4 state without this intervention.

[mrguyorama]

HIV isn't the only virus that generates decoys to trick antibody development, so having techniques that bypass that decoy strategy is very useful.

[Melatonic]

What else does this?

[greazy]

I think all viruses do to some extent. We refer to them as DIPs https://en.wikipedia.org/wiki/Defective_interfering_particle

List includes RSV, flu, sc2 and of course hiv

[M95D]

All single-stranded RNA viruses do that. But it's being done to them is the more correct term. A single strand of RNA is unstable and prone to mutations due to enviromental factors.

[wwweston]

A fascinating everyday arms race I had no idea was playing out around me.

[pazimzadeh]

Here’s a fun read on that arms race

The acquired immune system: a vantage from beneath

https://pubmed.ncbi.nlm.nih.gov/15539148/

Basically saying our immune systems ensure that microbes get ever more devious

However there’s ways that we counter that, like mucus that basically slowly feeds and tames your bacteria. Doesn’t work on viruses though

[Fire-Dragon-DoL]

Everything I hear is about bacterias though. I wish we found more about viruses.

[pazimzadeh]

viruses and transposons are a major source of creativity in life. the adaptive immune system of jawed vertebrates like us, which defends against viruses, only exists because the RAG transposon inserted itself into the gonads of our fish ancestors 500 million years ago

Origin of immunoglobulins and T cell receptors: A candidate gene for invasion by the RAG transposon

https://pubmed.ncbi.nlm.nih.gov/40614193/

[bongoman37]

Bacteria is plural, bacterium is singular.

[calf]

What's the risk of collateral damage or autoimmune issues from this kind of process, it sounds more difficult to evaluate for safety because it is more complicated.

[dist-epoch]

It's still the body doing the work, the antibodies will still be tested for self-attack in the regular ways, so there shouldn't be any impact. But immunity is super-complicated, so who knows...

[mjevans]

This sounds a lot like the firewall technique of port knocking. Poke this thing to enable a ruleset, then another thing in that state for a different ruleset, etc.

[christkv]

This is the way a lot of modern cancer vaccines are being developed. The idea is to create custom vaccines in the future targeting the patients specific cancer. So they are therapeutic vaccines. Obviously some cancers cannot be reach by a vaccine or are so aggressive that they overwhelm the patient before they can help.

[dekhn]

I work in pharma and a coworker asked if they could use my code (a model inference server). I asked what they were doing and it's an ML model for determining clinical dosing for a cancer trial. They sequence a person's cancer, then develop custom RNAs that train the immune system to eliminate that cancer (similar to https://www.nature.com/articles/s41392-026-02769-3) which they inject into the patient.

I would like to thank all the scientists who have dedicated all their hard work and ingenuity to eliminating the impact of cancer. Also, fuck cancer.

[epistasis]

I'm not aware of cancer vaccines that use this sort of b-cell priming, though I've heard lots about mRNA vaccines, as well as neo-epitope based cancer therapeutic vaccines. Are you analogizing the neoepitope approaches to the B cell progression here? That is somewhat similar, but the chain of new vaccines does seem quite different and novel to me.

[christkv]

I cannot answer those questions as much as I wish I could. My dads papers around the stuff they are working on is here if it’s of any use https://www.sciencedirect.com/author/7005329466/gunnar-kvalh...

[mohamedkoubaa]

Well, they kind of all work that way, it's just that most of the curriculum occurs naturally through interactions with wild pathogens

[mullingitover]

From what I understand, random interactions with wild pathogens can prime the immune system in the wrong way, leading it down the wrong path. This is why some people can get lucky with immunity and be super resistant to flu, because they happened to luck into a response that targets a common factor in most flu strains. Other people have bad luck and target some area that mutates frequently, and they proceed to get walloped by every flu strain, every year.

This is why it's hard to make a 'super vaccine' for flu, becuase it takes many rounds of shots to retrain some people's immune systems once they've latched on to the wrong recipe for antibodies.

[pazimzadeh]

And every time you generate an immune response you are honing your natural receptors so although they are more specific to the specific virus at hand, they lose general response, so this can backfire for example if you over-vaccinate with the wrong strain of flu and this one year the flu changes direction, mutationally.

this is called “original antigenic sin”

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

[Fire-Dragon-DoL]

Wait, so what are we supposed to do. If you over vaccinate is bad, if you under vaccinate is bad, but if you vaccinate a little bit you can end up on the wrong strain. Great! /s

[pazimzadeh]

actually, the 'little bit' one is probably correct, unless you are in a susceptible group like the elderly then probably vaccinate a lot

we can also get better at predicting what each year's mutations are going to be, though that takes funding

but the idea that more vaccination is always better is technically not true, especially for a virus that is extremely good at mutating/shuffling its genome like the flu does.

but the reasons given by anti-vaxxers are generally BS.

[Fire-Dragon-DoL]

Yeah I'm not in the anti-vaxxers thing, I was thinking more about the COVID vaccine we are given basically every year. After... I think 6 iterations (or 7?!) me and my family decided to go for the flu shot only, mainly because the flu shot can be given as a nasal shot, which is way less problematic.

Also the COVID shot is damn painful for some reason

[mrkstu]

As someone whose had rheumatic fever, your immune system can definitely prime itself for the wrong target.

[mohamedkoubaa]

Right I'm saying it's always path dependent not that it necessarily improves outcomes.

An analogy of this is having the wrong math instruction in middle school could make your math classes harder at the university level.

[shevy-java]

I do not see what is impressive here at all. The immune system constantly yields new antibodies targeting different epitopes - all the time. THAT is what is impressive, not multi-vaccines trying to induce shifts towards certain epitopes over others.

[hydrogen7800]

Are you saying that the result of millions of years of evolution is more impressive than decades of medical research?

[api]

All vaccines are training data for the immune system, which is a learning machine whose operation is a little bit like a genetic algorithm. It generates and sieves variation using a vocabulary of pre-encoded building blocks to fit the targets it's presented with.

It's an incredible system, and a product of probably a billion years of evolution since the first multicellular organisms had to protect themselves from infection.