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

The IGG curve dropped by the same pattern for non responders and responders. So IGG drop doesn’t seem to be the signal. As per the Dara trial.

However, could Lupus be a disease of the SLPC/plasmablast and CFS being of an LLPC?

That would explain why:

1. Lupus has memory, the process is generating new SLPCs every time and patients need to be on continuous treatment.
2. The one shot Dara treatment (not permanent) can induce remission in CFS by eliminating the faulty LLPC.

And of course, as for why there was “some” response in Rituximab P1 and P2, that is due to NK cell variation and also whether the faulty LLPC expressed CD20.

And for cyclo P1, it only kills LLPC in some patients. Since cyclo attacks dividing cells and LLPCs do not divide, it could be killing the supporting cast that provide nutrients to the LLPC itself?

So Lupus faulty is more upstream than CFS (which occurs at the last stage of the pipeline).

Both are lifelong but for diff reasons:

• Lupus has a continually renewing SLPC production
• CFS has a one time LLPC batch that lives for decades.

But the difference is in Lupus you have cells that continuously regenerate after being killed off, and in CFS you have a one time batch of cells that are immortal.

all this is just pure speculation of course. Just piecing things together given data. I’m not claiming anything here.


Lastly on subgroups I’m not keen on the idea personally, I believe any subgroup should be due to variation in the kind of AAB produced by the LLPC. Which of course any virus or stress can trigger.
 
Last edited:
ryanc97 said:
For example a commenter says for Lupus "I have been on Rituximab for 5 years now".

Which seems to imply that the faulty antibodies production cycle starts again if you stop?

Which then implies the fault is very upstream/far back in the pipeline like at the stem cell/B cell level? Is this is a correct understanding?
Click to expand...

Most physicians using these drugs do not think through the sort of question you are asking. They just maximise the number of chargeable treatments or follow some dumbed down guideline.

Both for rheumatoid and lupus the intelligent thing to do is to treat with rituximab once and monitor events closely from 5-12 months. Some patients relapse repeatedly at 6 monyhs and need roughly 6-monthly infusions. Immunosuppression is then more or less continuous. Other patients may rturn to a rather stable state for years and not need repeated treatments. Some have a relapse cycle of around a year.

But, yes, in many but not all cases the faulty antibody production cycle starts up again. That does not mean, however, that the fault is 'upstream' because, for a cycle, there is no upstream - upstream is downstream is upstream is downstream.

One possibility is that the fault is the continued survival of clones producing autoantibody, which we know have the machinery to educate further clones to join them in a sort of life long relay race. If that is the case the key is to clear away all antibody against the relevant antigens by removing all plasma cells and memory B cells with that commitment.

The other possibility is in fact 'upstream' in the sense that it belongs to another cycle, that can get the first one going again, that involves a shift in T cell clonality. Nobody has ever shown this but it may be that once you have lupus you grow up some T cells that will help new B cells start making autoantibody in a way they don't in normal people.

Over the last 25 years since we started this the key question has been whether if you clear out all antibody production as in explanation 1, you get long term remission. (You still have the genetic risk so you might get a new dose of lupus twenty yeas later, although this may not be that likely.) The second question about whether there are rogue T cells, I don't think we know how to handle. We know that if you drastically deplete helper CD4 T cells in rheumatoid it doesn't even get the disease to go away, so if there are rogue T cells it seems that they may not be needed all the time or only a tiny number are needed to keep the disease going.

The disappointment has been that up until the arrival of CAR-T therapy upping the potency of B cell depletion has not seemed to make much difference to length of remission. But nocking down plasma cells has not really been tried much, except perhaps recently in lupus. Unofrtunately, the physicians doing the trials tend not to realise that you need to clear out all the antibodies to hope to get a cycle break.
 
ryanc97 said:
It would seem the leak for Lupus is very much far upstream/closer to source (aka the stem cells) than CFS?
Click to expand...

This isn't going to have anything to do with stem cells because they are not committed to any particular antibody. However, the defect may relate in part to failure of weeding out of bad B cells at the pre-B stage in bone marrow. That would fit with the complement gene risk (complement does the weeding) and also in RA with the PTPN22 gene risk, since that molecule is involved as well.

But the cycle is almost certainly going to occur later in mature B cells that make use of existing antibody to educate each other as whether or not to proliferate.
 
Arfmeister said:
There is actually Cyclo data on IgG levels.
Click to expand...

Useful to see that. As I would have expecteed, the drops are modest and probably not even below normal levels. It is hard to see how you explain symptom improvement on that basis unless there is a selective killing of short lived plasma cells producing all the bad stuff, but then rituximab should have worked.
 
Jonathan Edwards said:
Useful to see that. As I would have expecteed, the drops are modest and probably not even below normal levels. It is hard to see how you explain symptom improvement on that basis unless there is a selective killing of short lived plasma cells producing all the bad stuff, but then rituximab should have worked.
Click to expand...

Since the drop pattern is almost identical in responders and non responders it would seem IGG drop is not particularly predictive unlike NK cell, this is seen by the scatterplot. Of course why nobody knows why


On this topic is there a chance cyclo can kill LLPCs? Not directly but indirectly. Because under an LLPC theory for cyclo to work in some which it did the LLPC must have been hit.

Is there any medical knowledge that supports this? I know cyclo targets dividing cells which LLPCs are not but still. Maybe it kills the supporting cast
 
Last edited:
Arfmeister said:
Why can’t we state that Ritux still might have worked for a subgroup?
Click to expand...

Yes, but bananas might work, or sitting on a clod floor for six hours. The data provide us with no good reason to think the drug did more than bananas would.

Arfmeister said:
I’ve notice drug trials are the highest standards, rightly so, we need for scientific rigor.
But it doesn’t leave room for the middle ground: Either it is placebo, or the drug works.
Basically saying it’s placebo means patients were in remission because of psychological effect.
(Or they are the relapse remitting type.)
Click to expand...

No, drug trials are not the highest standard. They are the minimum usable standard. And there are a dozen reasons why people get better during trials. Some involve psychology, some do not.
 
Arfmeister said:
PS: Isn’t it common in many autoimmune diseases for e.g. Rheuma. A. or M.S. ?
- certain drugs work for a small % of patients, others drugs work for another group and some patients don’t respond at all ?
Click to expand...

For the new biological class of drug it is usual to see most patients show a convincing improvement in biological measures of disease if the drug works. Some have too much damage to get much symptom benefit. Some are drug resistant because of immune responses or have adverse reactions. When I first chose 5 patients I thought would respond to rituximab if our theorising was right 5 out of 5 had what is regarded as a major response of ACR50 and 3 out of 5 had the highest rated response of ACR70. For a drug to have major benefit in such a small proportion that nothing shows up on the statistical group analysis is unheard of.
 
@Jonathan Edwards you mentioned that antibody producing cells can “teach” their neighbours to produce the same antibody, so as long as some remain, they will come back.

Educate further clones to join then. How does this mechanism work?

Could the same logic apply to long lived plasma cells too?

I understand these guys are fairly entrenched and quite specialized, can they teach new cells to become LLPCs and produce the same antibody they do?
 
Last edited:
EndME said:
Of course subgroups responding to Rituximab theoretically might exist but you cannot argue that the trial provides evidence of this because it doesn't, it only provides evidence of no efficacy of the drug in treating ME/CFS
Click to expand...

I think many points you are making are very valid. I was actually editing my post to be more clear.

But for me, the heterogeneity of ME-patients and the lack of clear bio markers makes it very difficult to make bold statements.
- in almost all the ME-drug trials I’ve read upon there’s always a responder group - be it big or small
- Placebo & poor quality trial design aside: for me it’s clear that ME patients can react very different to different drugs or supplements
- Just a day of reading on this forum and you get the picture

Adding complexity to that is the relapse remitting type vs progressive worsening type - and severe vs mild
- I think that makes conclusions from drug trials even more difficult to make

E.g. Sometimes a few detailed case studies from (a) long-term ME (b) severe bedbound patients can be more telling than a poor quality design drug trial.

So I don’t discard that Ritux could’ve been effective for a few patients in the trial.
- It definitely is not an effective treatment for ME CFS - as long as we don’t know more about possible phenotypes/sub groups


PS: Personally, I think it would be great if F & Mella were able to start classifying their trial patients into the following groups.

IMG_0242.jpeg
 
Arfmeister said:
I think many points you are making are very valid. I was actually editing my post to be more clear.

But for me, the heterogeneity of ME-patients and the lack of clear bio markers makes it very difficult to make bold statements.
- in almost all the ME-drug trials I’ve read upon there’s always a responder group - be it big or small
- Placebo & poor quality trial design aside: for me it’s clear that ME patients can react very different to different drugs or supplements
- Just a day of reading on this forum and you get the picture

Adding complexity to that is the relapse remitting type vs progressive worsening type - and severe vs mild
- I think that makes conclusions from drug trials even more difficult to make

E.g. Sometimes a few detailed case studies from (a) long-term ME (b) severe bedbound patients can be more telling than a poor quality design drug trial.

So I don’t discard that Ritux could’ve been effective for a few patients in the trial.
- It definitely is not an effective treatment for ME CFS - as long as we don’t know more about possible phenotypes/sub groups


PS: Personally, I think it would be great if F & Mella were able to start classifying their trial patients into the following groups.

View attachment 28213
Click to expand...
Under this framework I would attribute any patient on Ritux having remission to the combo of having enough NK cells for ADCC + having faulty LLPCs that have CD20 so Ritux can get at them. Probably the intersection of these two conditions is so low you have like a 10% remission rate and this gets drowned out when you do pooling
 
Arfmeister said:
- in almost all the ME-drug trials I’ve read upon there’s always a responder group - be it big or small
Click to expand...
But the whole point is that this has nothing do with ME/CFS. One thinks to see "responders" in any intervention for any condition because of things like regression to the mean, natural recovery, Hawthorne effect, placebo-like effects, subjective outcomes etc, there's countless papers and books on these subjects. That's the whole point of trial designs. And of course one should expect Hawthorne like effects to be particularly big in ME/CFS trials and that seemingly might have also been the case. One also has to be honest: Most trials in ME/CFS are of an abysmal quality and claims about (super)responders in many trials have been very laughable and ridiculous, which isn't different to other conditions with low quality research either. There's also reasons why many people have been advocating for long-term and objective outcomes measures in ME/CFS trials. ME/CFS is not particularly special in that sense.
Arfmeister said:
- Placebo & poor quality trial design aside: for me it’s clear that ME patients can react very different to different drugs or supplements
Click to expand...
Maybe, but that doesn't matter for drug efficacy for the illness. On a whole ME/CFS patients reacted the same to Rituximab as they did for the placebo, so actually the picture is rather consistent and not the opposite. Maybe one can have a closer look at only the placebo group to gauge the reaction to a "drug" vs the reaction to a trial (in which Hawthorne and the like always play a role).

I do appreciate that there are some things that make trials in ME/CFS tricky compared to some very simple situations, for example remissions in ME/CFS, but to be honest the same applies to many other conditions as well, where people have shown what succesful trials look like with the difference between intervention and non-intervention being night and day. I think the situation will be fairly straightforward: One day the situation will become very clear when someone actually shows success of an intervention in ME/CFS, but because people have never seen success they still wander what kind of dirt they can interpret as success. Right now many people are mistaking dirt for gold, simply because they've never seen gold before. For some that dirt may be Rituximab for others it is CBT/GET.

Arfmeister said:
PS: Personally, I think it would be great if F & Mella were able to start classifying their trial patients into the following groups.
Click to expand...

That's actually exactly what F&M have done. I think it's been discussed on the thread for the new Daratumab trial, but they have done some long-term activity monitoring to understand the situation better and are doing some pre-trial tracking of patients and have adjusted recruitment criteria of their trial to hope to filter out additional noise.
 
EndME said:
That's actually exactly what F&M have done. …and have adjusted recruitment criteria of their trial to hope to filter out additional noise.
Click to expand...

Well, on that aspect, they are gold - in accordance to your analogy ;)

Wrt Ritux: I agree, it is not a convincing drug at all.
- I just happen to know 2 severe cases that benefited a lot, which is coloring my opinion on ‘it’s all placebo’
 
ryanc97 said:
Educate further clones to join then. How does this mechanism work?

Could the same logic apply to long lived plasma cells too?
Click to expand...

The strange thing about B cell clones that are to expand (into plasma cells, either short or long lived) to make useful amounts of antibody is that they do so in response to picking up antigen already bound to an antibody. So to make lots of antibody you have to have made some antibody already to get started. The advantage of the system is that it provides an explosive chain reaction of B cell clone expansion when antigen arrives (as long as you have some primer antibody). Newborn infants use mother's antibodies to prime but throughout life our immune systems work on the basis that we are always producing a little bit of antibody to everything, weeding out antibodies to self through a mixture of processes in bone marrow and lymph node as best we can.

How autoantibody production gets out of hand in autoimmune disease remains mysterious. Bone marrow and lymph node cytotoxic T cells should cut it out but for certain self antigens we know there are glitches in the control mechanisms that may be exploited.

The biggest glitch of course is the risk that if you have already made quite a bit of autoantibody that can bind to loads of self antigen and any new B cells that can make more antibody to that antigen will pick up the antigen bound to antibody and get a signal to proliferate. So old B clones can educate new B clones to make antibody to the same antigen. This is where plasma cells come in, though, because the antibody bound to antigen will have been made by plasma cells, not B cells as such. So LLPC are an intrinsic part of the education process.

There is an interesting complication to this, which relates to the normal switch off mechanism. If you are making antibody to a foreign antigen and make enough antibody to bind all over and cover up the antigen it becomes invisible to B cells and so the chain reaction stops. More generally, the process relies on antigens having at least two 'sides' or epitopes. If you already have plasma cells from clone C making antibody to side R but not side L of the antigen then side L is free to bind to surface antibody on B cells of clone D, which will then make plasma cells that make antibody to side L. So B cell clones 9via their plasma cells) always educate other B cells to produce new antibodies to a different side or epitope of the antigen. Many antigens may be big enough for four or more antibodies to bind at once to different epitopes although I suspect when most immune complex aggregates form each antigen probably only binds 2 or 3 antibodies.

Our original 1999 Immunology paper (Edwrds, Cambridge, Abrahams) on self-perpetuating autoreactive B cells had diagrams to show how this all works. I am not sure if it is open access now.
 
ryanc97 said:
LLPCs that have CD20 so Ritux can get at them.
Click to expand...

From all the historic data we have I think we can be pretty sure that ritux doesn't kill a significant number of LLPC. A very tiny number of people get rapid hypogammaglobulinaemia but not moe than 1 in 100 in my experience. Otherwise LLPC as judged by antibody levels to things like tetanus really don't budge.
 
Jonathan Edwards said:
From all the historic data we have I think we can be pretty sure that ritux doesn't kill a significant number of LLPC. A very tiny number of people get rapid hypogammaglobulinaemia but not moe than 1 in 100 in my experience. Otherwise LLPC as judged by antibody levels to things like tetanus really don't budge.
Click to expand...
I see. I am just quoting Mella because in his video he said that up to 20% of LLPC express CD20 and he thought maybe that is the action of Ritux in responders.

I guess for me I just look at ALL the data and try to fit a theory. So I try to fit something that can explain all the trials, Ritux P1, P2 and P3, Cyclo P1 and Dara P1.

So the theory must account for any remission under any of those drugs and I don’t just discount it as placebo.

The main unknowns are

• If it really is LLPC only then there needs to be some mechanism by which Ritux and Cyclo both can cause remission by hitting the LLPC
• If SLPCs are involved then Ritux and Cyclo success could explain remissions.

But because Dara has the highest success rate so far (and a mechanism to explain non responders quite well) and Ritux has “failed”, then it being LLPC is the most likely theory given the evidence.

Also because the disease is unknown there is definitely a gap in medical knowledge so far that could mean conventional medical wisdom could be lacking something.

But anyway, it’s all just speculation…

It comes from my background where I looked at lots of data in commodity futures and tried to code up systematic trading strategies in Python for my firm (when i was healthy and brain worked flawlessly…)

The signal to noise ratio was very low. You needed to look at data and form a hypothesis that explained why your strategy or signal worked or didn’t work. So to me, discarding any strange data point and saying it’s invalid (aka placebo) would be madness to me.

You would be able to make money if your signal had an R2 of 0.10, so when I saw the Dara scatterplot with Spearman’s rho 0.77, it basically jumped out at me.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom