Is the key pathology of ME/CFS in bone marrow?

Jonathan Edwards

Jonathan Edwards

Senior Member (Voting Rights)
This is flying another kite to see if it gets off the ground.

The idea is that the reason we cannot see the pathology of ME/CFS is that at least a crucial part of it is in the red (haemopoietic) bone marrow.

Red marrow is a bit like a fish market. If you wander around you will find piles of fish (cells) of various sorts and you will often find piles of the same fish on lots of stalls but on some they are the main species there, on others there are just a few. The whole thing seems a crazy muddle but each stall holder knows exactly what price to ask and how many fish are likely to be needed that day.

The point being that if you sample bone marrow you can get a rough idea which cells are most frequent but all sorts of things are going on in microenvironments that are very hard to assess, even if you take a solid trephine biopsy that preserves these.

I have tended to focus on lymph node and spleen when thinking of immune mechanisms. (Blood is unlikely to tell us much because it is just a route for cells to traffic through. So it is not surprising that in most of the best known chronic immune diseases, where we know B or T cells have specific roles, blood shows little or nothing useful.) But bone marrow may be just as obvious a place to look.

Bone marrow contains very young cells, forming from stem cells, and also very mature cells - plasma cells. It doesn't handle the lymphocyte clonal selection of the adaptive immune response in between.

Bone marrow is also well innervated, at least in sensory terms. It is highly sensitive to pressure change in particular.


In simple terms, the idea I am suggesting is that some shift in bone marrow cell dynamics, like growing up and sending out a skewed population of myeloid or lymphoid cells, feeds unhelpful cell interactions in other tissues, including lymph nodes and spleen that bring back signals to bone marrow that perpetuate the original skewing.
 
Lupus may be an example of this sort of picture. Complement deficiency promotes lupus. At least in part this is probably because complement is needed in bone marrow to weed out autoreactive B cells. If complement is not working B cells can go out and proliferate to make plasma cells that make autoantibodies that can further interfere with complement - anti-C1q antibodies being the paradigm.

The idea is that ME/CFS depends on a different circuit but it may be facilitated by a factor associated with female sex that also facilitates lupus.

A shift in bone marrow cell output following an infection, perhaps for a period of weeks or months might make sense as part of innate immunity. I am attracted by the idea that a reason for female predominance in both lupus and ME/CFS may be the option of a double dose of TLR7 on two X chromosomes, implicating the Xist gene that is normally involved in the silencing of one X.

Maybe TLR7 expression feeds in to pathways, that may involve interferons, which TLR7 is good at inducing, that shift the cell production microenvironments in marrow to produce more of certain sorts of myeloid or lymphoid subsets. Something a bit like seasonal changes in the fish market.

TLR7 could explain the 3:1 female predominance in most autoimmune disease, I guess, through an effect on interferons that simply made B cell survival in germinal centres a bit easier, making an autoimmune clonal error easier. But lupus is 9:1 so maybe TLR7 can have another effect through another environment, which might be bone marrow. If this operates in lupus and ME/CFS, but not in most other autoimmune disease, that might explain why lupus and ME/CFS both have incidence peaks earlier on - if the effect was not directly related to accumulation of random Ig gene rearrangements and mutations over a lifespan.
 
There is one further element to the broad scenario I think relevant at this point.

Jnmaciuch and I have debated the relative importance of type 1 and 2 interferons. A bone marrow model could easily accommodate a shift in type 1 interferon dynamics.

However, I now realise that I have two reasons for being worried about relying entirely on innate mechanisms for a 'disease memory-loop'. One is that although it is relatively easy to see how specific clones of B or T cells could subvert normal feedback rules in a way that seems harder to see for myeloid or stromal cells that have simpler rules. The other is that it looks as if genetics and environment probably leave a substantial causal gap for ME/CFS much as they do for diseases like RA and that a stochastic element is likely needed. Lymphocyte development and dynamics is inherently tied to random events. For myeloid cells that is not so.

In autoimmune disease we see rogue B cell clones setting up. In seronegative spondarthropathies we see rogue T cell subsets setting up. It is not clear how clonal these are but their sudden appearance suggests preferential explansion of certain clonal groups at least. In ME/CFS we do not obviously see either but it would not be hard for rogue clones that just fed back signals on to bone marrow dynamics to go unnoticed. Plasma cells would be ideal of course, since they set up shop there. But T cells involved in regulating other cells (not necessarily Treg per se) would do.
 
Sasha said:
If that was so, wouldn't TLR7 have shown up on a bunch of GWAS by now?
Click to expand...

No. This is the key point. GWAS says nothing about these genes that make women more at risk than men because we are not talking of gene variants here, which is all GWAS can show. We are talking of genes that everybody has, either in one or in two doses.

Or to put it another way, a GWAS on twenty patients would show TLR7 was in a risk factor DNA segment (X) - assuming that you controlled against a group of normal controls not selected with a 3:1 sex ratio.People with ME/CFS are much more likely to have two X chromosomes than the general population unselected. So all the SNPs on the X chromosome in DecodeME would give a p value of 10-^100 or something if compared with an unselected population.
 
Last edited:
Jonathan Edwards said:
Lupus may be an example of this sort of picture. Complement deficiency promotes lupus. At least in part this is probably because complement is needed in bone marrow to weed out autoreactive B cells. If complement is not working B cells can go out and proliferate to make plasma cells that make autoantibodies that can further interfere with complement - anti-C1q antibodies being the paradigm.

The idea is that ME/CFS depends on a different circuit but it may be facilitated by a factor associated with female sex that also facilitates lupus.
Click to expand...
Are there any tests that aren't commonly done but would be able to test your idea?

I was just looking at blood tests of mine, and noticing that both times it was measured, C4 was right on the lower border (0.14), and C3 was on the lower border once.

I may not be typical though, as I have a consistently positive speckled ANA with most recent titre 1:320, though I do not have lupus or any other diagnosed autoimmune disease, or any antibodies other than TPO. Family history of all things autoimmune.

In the NIH study, 24% of patients had positive ANA vs 5% of controls.

Regarding Decode ME, there were numerous mentions of lupus in the candidate gene document:
Evergreen said:
There are also numerous mentions of lupus in the candidate gene document:
RC3H1 "Lupus-like autoimmunity"
TNFSF4 "Region upstream of TNFSF4 is associated with systemic lupus erythematosus risk"
TRIM38 "A minority of lupus and Sjogren’s syndrome patients have autoantibodies reactive to TRIM38"
Click to expand...
Would any of those support your idea?

Edited to correct % of patients in NIH study with positive ANA to 24% (not 20% as originally typoed).
 
V.R.T. said:
So if we were convinced the pathology was partly in the bone marrow in ME/CFS, what would the next steps be?
Click to expand...

Try and figure out which cells are most likely involved.
The hard part of this is building a plausible overall model. When I worked on RA I made hardly any new observations on tissues (one or two crucial ones maybe) - almost all the necessary evidence, much of it negative evidence, was out there.
 
Evergreen said:
In the NIH study, 24% of patients had positive ANA vs 5% of controls.
Click to expand...

I wouldn't take the NIH cohort as representative of anything though. I think most studies have not found much difference if any.

There is a persistent suggestion that thyroid autoantibodies are more common in people with ME/CFS, which maybe should be taken seriously. But again, there are all sorts of ways one might explain that which don;t necessarily imply the ME/CFS itself requires autoantibodies.
 
Jonathan Edwards said:
I wouldn't take the NIH cohort as representative of anything though. I think most studies have not found much difference if any.

There is a persistent suggestion that thyroid autoantibodies are more common in people with ME/CFS, which maybe should be taken seriously. But again, there are all sorts of ways one might explain that which don;t necessarily imply the ME/CFS itself requires autoantibodies.
Click to expand...
I must look further into other studies. I find ANA is rarely mentioned, but I've probably missed some good studies.

Indeed. I remember reading a bunch of recovery stories where people with a diagnosis of CFS recovered with thyroid hormone treatment - surely that's just misdiagnosis.
 
Evergreen said:
Indeed. I remember reading a bunch of recovery stories where people with a diagnosis of CFS recovered with thyroid hormone treatment - surely that's just misdiagnosis.
Click to expand...

Or whatever, yes.
But thyroid autoantibodies are increased in a range of other autoimmune diseases and may be an indirect clue to autoimmune goings on. At least one US study claimed a higher rate in ME/CFS.
 
Jonathan Edwards said:
No. This is the key point. GWAS says nothing about these genes that make women more at risk than men because we are not talking of gene variants here, which is all GWAS can show. We are talking of genes that everybody has, either in one or in two doses.

Or to put it another way, a GWAS on twenty patients would show TLR7 was in a risk factor DNA segment (X) - assuming that you controlled against a group of normal controls not selected with a 3:1 sex ratio.People with ME/CFS are much more likely to have two X chromosomes than the general population unselected. So all the SNPs on the X chromosome in DecodeME would give a p value of 10-^100 or something if compared with an unselected population.
Click to expand...
Aren't all GWAS control samples selected with the same sex ratio as the disease group?

Would a WGS study also help look at this, because it can detect repeat segments (i.e. extra dosing of the gene)?
 
Could you collect bone marrow post mortum and still get a useful sample? Because ME/CFS patients die every week, many of which would donate their bodies to science if there is someone wanting to take the body and do more detailed studies on it. When it comes to invasive tests like this having registered sufferers willing to donate their body might well be the way to go, assuming what you need to look for is likely still viewable.
 
You must log in or register to reply here.
Back
Top Bottom