Long-lived plasma cell (LLPC) theory - Similarities between CFS and Lupus?

[ryanc97]

Sure, maybe, but that’s exactly my point. You can pretty much spin the wheel for any part of the body and claim that it’s somehow driving symptoms by a completely unknown mechanism that deviates from everything else we’d expect. We could just as well say it’s the thyroid. Or a specific subtype of tissue resident myeloids that hasn’t been well characterized yet.

As I already stated, the main and really only explanatory benefit of autoantibodies is that we know they can sustain long term disease. But they’re far from the only things that do that, and other explanatory options don’t require you to basically hold out hope for an undiscovered unicorn that magically behaves in exactly the way we want it to, contrary to every other example of that mechanism at play.

I think it’s more like because to me the evidence (just working backwards from the trial data without assuming any medical knowledge) points to the AAB / LLPC which then has a magical hole in knowledge.

Whereas coming from a more research background with medical knowledge (you weigh prior/current knowledge more) you end up at the other conclusion.

Obviously us patients have no prior knowledge so we don’t know the current science, so we draw conclusions or at least me purely from trial results whether the conclusion is medically feasible or not.

 

[forestglip]

The more practical problem will be that if the efficacy is so low that you'd need a study of thousands of patients or even more in an ideal environment, then you can not justify the risks in the first place and you will likely never be able to show efficacy in the first place due to practical problems (the NIH struggled recruiting just 20 patients rigorously, you would probably struggle to recruit that many people in the current ME/CFS world without ending up with a problematic cohort).

Oh yeah, I don't think there's any reason to care much about proving rituximab works with giant sample sizes. It clearly isn't very effective.

I'm just saying that I think there is reason to believe that either cyclo or rituximab (or both) actually affects some mechanism that improves ME/CFS, even if the effect is tiny and not clinically meaningful, because it seems exceedingly unlikely to me that just by chance these initial results were false positives but that they would go on and happen to find a drug with an incredibly similar mechanism of action that does work.

 

[EndME]

I think it’s more like because to me the evidence (just working backwards from the trial data without assuming any medical knowledge) points to the AAB / LLPC which then has a magical hole in knowledge.

Whereas coming from a more research background with medical knowledge (you weigh prior/current knowledge more) you end up at the other conclusion.

Obviously us patients have no prior knowledge so we don’t know the current science, so we draw conclusions or at least me purely from trial results whether the conclusion is medically feasible or not.

I think everybody would be working backwards from trial data if we had any meaningful data. If we'd have positive data from a sufficiently controlled study I think everybody will work backwards. But the problem is that we currently don't know that but already know that it's easy to not have any meaningful data and think the opposite.

Garner also argues that he can work back from meaningless lightning process study data, but anybody will be quick to point out the problem in that approach.

 

[EndME]

I'm just saying that I think there is reason to believe that either cyclo or rituximab (or both) actually affects some mechanism and improves ME/CFS, even if the effect is tiny and not clinically meaningful, because it seems exceedingly unlikely to me that just by chance these initial results were false positives but that they would go on and happen to find a drug with an incredibly similar mechanism of action that does work.

Sure, but this just keeps on ignoring the fact that people in the placebo group improved just as much. There's no need to invoke false positives or chance. Sure, some effects can be related to efficacy, but we already know the opposite as well, you can have "positives" without any efficacy, just ask people in the placebo group.

 

[ryanc97]

Well in a few years once the Dara P2 is complete we will know and if the results are positive people will investigate further into the cause

 

[forestglip]

But this just keeps on ignoring the fact that people in the placebo group improved just as much. There's no need to invoke false positives.

Yes, I'm only referring to the phase 2 trial, and assuming the phase 3 trial was underpowered to replicate the effect. But it does make more sense that cyclo would be the drug with the common mechanism, since it was the one that started this whole thing.

 

[EndME]

Yes, I'm only referring to the phase 2 trial, and assuming the phase 3 trial was underpowered to replicate the effect. But it does make more sense that cyclo would be the drug with the common mechanism, since it was the one that started this whole thing.

But both studies were negative and the phase 3 trial had a much larger sample size. It couldn't have been underpowered unless we are invoking recruitment criteria or something else, but I don't think there's much reason to do that. And of course the effect replicated rather than the opposite: Both studies failed (I guess you can argue about adapting the primary outcome in the phase 2 to something else and that this didn't replicate in the phase 3 but such arguments don't seem sensible to me).

 

[ryanc97]

I think everybody would be working backwards from trial data if we had any meaningful data. If we'd have positive data from a sufficiently controlled study I think everybody will work backwards. But the problem is that we currently don't know that but already know that it's easy to not have any meaningful data and think the opposite.

Garner also argues that he can work back from meaningless lightning process study data, but anybody will be quick to point out the problem in that approach.

Yes but the difference is that I’m willing to accept that data quality and have a plan of action from there (boost NK cells then take Dara)n rather than sit and wait

It’s just personal preference

 

[Jonathan Edwards]

But if dara wiped out all LLPCs, would tetanus antibodies quickly be automatically re-made?

No, because there is no tetanus antigen around.But for autoantigens they are around all the time so you expect new plasma cells and antibody to be made straight away - you are re-vaccinating yourself against self every day.

If some wonky LLPCs got made when a person got a COVID infection and which cause ME/CFS, but without a persistent antigen, I wouldn't expect them to quickly be replaced if they got killed.

This is the sort of story that Jackie Cliff and Jo C and I raised in the Qeios piece. Jackie would like it to be herpes related but she is open to other options. The question is in what way are these antibodies wonky? The idea that you can make autoantibodies as a result of meeting an infection has been popular over the years but as far as I know it never happens and the explanation for it happening doesn't make any real sense. Which is why we suggested that these aren't typical autoantibodies.

But, as we explore in the Qeios article, there is another serious problem with this theory in that ME/CFS seems to follow lots of different infections and so there is no good reason why you should make wonky antibodies that all produce the same disease after all these different infections.

 

[ryanc97]

But, as we explore in the Qeios article, there is another serious problem with this theory in that ME/CFS seems to follow lots of different infections and so there is no good reason why you should make wonky antibodies that all produce the same disease after all these different infections.

I mean the conclusion is that the gap in medical knowledge is right there, clearly it is happening but science has not discovered how it works yet

So do you say “it can’t be possible given our current knowledge and it must be something else”

Or

“It can be possible but we just don’t know how yet”

 

[jnmaciuch]

I think it’s more like because to me the evidence (just working backwards from the trial data without assuming any medical knowledge) points to the AAB / LLPC which then has a magical hole in knowledge.

Whereas coming from a more research background with medical knowledge (you weigh prior/current knowledge more) you end up at the other conclusion.

Obviously us patients have no prior knowledge so we don’t know the current science, so we draw conclusions or at least me purely from trial results whether the conclusion is medically feasible or not.

unfortunately I think that ends up being a perfect demonstration for why introductions and conclusions can’t be taken at face value—reading a discussion of Dara that only mentions LLPCs will give the false impression that any efficacy of Dara will be attributable to LLPCs when there was not remotely enough evidence to zero in on that explanation.

In reality, it’s one small tangent on a long list of things CD38 does. It’s basically like saying that if you didn’t eat for 48 hours and feel much more energetic after eating a fistful of almonds, it must be because you needed the Manganese from the almonds. Theoretically that could be true, but wisdom indicates that it’s probably low on the list for reasons why eating almonds made you feel better.

 

[forestglip]

But both studies were negative and the phase 3 trial had a much larger sample size. It couldn't have been underpowered unless we are invoking recruitment criteria or something else, but I don't think there's much reason to do that. And of course the effect replicated rather than the opposite: Both studies failed.

Ok, yeah, rituximab might have been nothing, a wrong way to try to replicate cyclo. But I think it is likely that whatever explanation dara offers, it will have to explain why cyclo causes a response as well.

Yes, it was possibly an unrelated placebo effect, as we've seen with lots of interventions, and maybe my intuition is wrong. But to be so convinced by a placebo effect that you devote your lab to immune cell killing drugs for years, and then actually find one that works, seems implausible.

 

[Jonathan Edwards]

I mean the conclusion is that the gap in medical knowledge is right there, clearly it is happening but science has not discovered how it works yet

Ah, but if you invoke a T cell mechanism you don't have this problem of getting the same disease after lots of infections - that is the norm for T cells.

That is because with antibody the effector mechanism is a chemically specific binding of antibody to a specific protein and the illness is what happens when that protein is screwed up. There are some more generalised mechanisms involving immune complexes but the self proteins that can do that are still very few. For T cells, however, the effector mechanism is generic - either killing of a cell that is misbehaving or secreting warning cytokines or whatever. The T cell never actually sees the whole protein is doesn't like.

So this is only a problem for invoking antibodies. For T cells we know exactly how it would work. And for the T cell diseases we get the same thing, such as Reiter's, with any number of different infections. It all fits very well. (Which is another piece of circumstantial evidence for jnmaciuch.)

 

[ryanc97]

unfortunately I think that ends up being a perfect demonstration for why introductions and conclusions can’t be taken at face value—reading a discussion of Dara that only mentions LLPCs will give the false impression that any efficacy of Dara will be attributable to LLPCs when there was not remotely enough evidence to zero in on that explanation.

In reality, it’s one small tangent on a long list of things CD38 does. It’s basically like saying that if you didn’t eat for 48 hours and feel much more energetic after eating a fistful of almonds, it must be because you needed the Manganese from the almonds. Theoretically that could be true, but wisdom indicates that it’s probably low on the list for reasons why eating almonds made you feel better.

Well my main argument is that we see the same NK cell correlation in MM where we know Dara is killing LLPC therefore it is the same thing in ME.

but of course you are saying it is invalid to draw such a conclusion but that’s how I see it.

 

[forestglip]

so there is no good reason why you should make wonky antibodies that all produce the same disease after all these different infections.

I agree. The specifics of your paper were somewhat over my head. But maybe nothing to do with the specific type of antibody, but instead something about location? Is it possible maybe the normal covid or EBV antibody LLPCs somehow find their way into the brain and set up camp after an infection? I like brain but maybe another location. Would dara deplete B cells in the brain?

 

[EndME]

Yes, it was possibly an unrelated placebo effect, as we've seen with lots of interventions, and maybe my intuition is wrong. But to be so convinced by a placebo effect that you devote your lab to immune cell killing drugs for years, and then actually find one that works, seems implausible.

But why should this be the case? Maybe they've simply gone up the chain to explore what other ideas are left? If you can't solve the problem with a certain approach, you have to do something else, but sometimes you will first explore something that is similar in approach and often that will work. That's how countlessly many problems are solved without it meaning that the first approach actually works.

If I remember correctly there were discussion about going up the chain pretty much when the Rituximab ideas started in ME/CFS.

 

[ryanc97]

Ah, but if you invoke a T cell mechanism you don't have this problem of getting the same disease after lots of infections - that is the norm for T cells.

That is because with antibody the effector mechanism is a chemically specific binding of antibody to a specific protein and the illness is what happens when that protein is screwed up. There are some more generalised mechanisms involving immune complexes but the self proteins that can do that are still very few. For T cells, however, the effector mechanism is generic - either killing of a cell that is misbehaving or secreting warning cytokines or whatever. The T cell never actually sees the whole protein is doesn't like.

So this is only a problem for invoking antibodies. For T cells we know exactly how it would work. And for the T cell diseases we get the same thing, such as Reiter's, with any number of different infections. It all fits very well. (Which is another piece of circumstantial evidence for jnmaciuch.)

Well that’s fair, I have no idea about biology so how I would do it is under this theory look at possible drugs that target these T cells and then scan forums for anecdotes of people improving on said drugs to validate this theory.

No biology needed, just only anecdotes. But of course to people here that is not the proper way to do science

 

[jnmaciuch]

Well my main argument is that we see the same NK cell correlation in MM where we know Dara is killing LLPC therefore it is the same thing in ME.

but of course you are saying it is invalid to draw such a conclusion but that’s how I see it.

Again, that’s like saying that another study found that manganese levels correlated with how good people feel therefore it must be why someone felt good after eating almonds

 

[ryanc97]

Again, that’s like saying that another study found that manganese levels correlated with how good people feel therefore it must be why someone felt good after eating almonds

Yeah so we differ in how we draw conclusions from data. That’s pretty much the crux of it. I’m not a biologist I was a (quant) so that’s how I do it.

My job was to draw conclusions with very crap data so that’s how I am doing it here lol

Put it this way i would look at scatterplots with R2 of 0.05 and make a decision off it

 

[forestglip]

But why should this be the case? Maybe they've simply gone up the chain to explore what other ideas are left. If you can't solve the problem with a certain approach, you have to do something else, but sometimes you will first explore something that is similar in approach and often that will work. That's how countlessly many problems are solved without it meaning that the first approach actually works.

It's possible. I can't construct a good argument that satisfyingly refutes that, just intuition that may very well be wrong.