MS is a B cell Disease and Memory B cells are the Target for therapy

Baker D, Marta M, Pryce G, Giovannoni G, Schmierer KMemory B cells are major targets for effective immunotherapy in relapsing multiple sclerosis. EBioMedicine 2017

Although multiple sclerosis (MS) is considered to be a CD4,Th17-mediated autoimmune disease, supportive evidence is perhaps circumstantial, often based on animal studies, and is questioned by the perceived failure of CD4-depleting antibodies to control relapsing MS. Therefore it was of interest to find that current MS-treatments, believed to act via T cell inhibition including: beta-interferons, glatiramer acetate, cytostatic agents, dimethyl fumarate, fingolimod, cladribine, daclizumab, rituximab/ ocrelizumab physically, or functionally in the case of natalizumab, also depleted CD19+,CD27+ memory B cells. This depletion was substantial and long-term following CD52 and CD20-depletion, and both also induced long-term inhibition of MS with few treatment cycles, indicating induction-therapy activity. Importantly, memory B cells were augmented by B cell activating factor (atacicept) and tumor necrosis factor (infliximab) blockade that are known to worsen MS. This creates a unifying concept centered on memory B cells that is consistent with therapeutic, histopathological and etiological aspects of MS.

Following on from our recent post (major target for treatment), it is clear that some of you don’t want to read the full story (if you do click here) and would like it summarised


  • Memory B cells express CD19+, CD27+  and CD21, CD20, CD25, CD49d, CD52, S1P1, and have higher endogenous proliferative background than T cells.
  • Memory B cells express CD21, which is the EBV receptor
  • Memory B cells are elevated in MS and accumulate in the CNS in MS
  • Memory B cells are depleted or functionally silenced by all treatments of MS.
  • Memory B cell depletion potential correlates with drug efficacy
  • Memory B cell augmentation is associated with MS worsening
  • Memory B cell-tailored therapy may allow personalised and safer induction treatment
ProfG was set to go to Lisbon to present at an antibody engineering meeting to talk on “How do antibodies work in MS”. Rather than try present his 65 slides in 30min, we aimed to see if we could answer the question. 

So do alemtuzumab, rituximab/ocrelizumab, daclizumab and natalizumab have anything in common? 

Although alemtuzumab depletes T cells, it was evident that memory B cell depletion is marked and it set us thinking as an alternative to the T cell hypothesis.

Surprisingly all agents that inhibit MS could cause depletion of memory B cells. Interestingly those that worsen MS cause elevated memory B cell responses in some people.
Surprise surprise, we found papers where some people had similar thoughts (there is no such things as an original idea) but in most cases this was a throw-away idea and was dumped in the midst of many other concepts. However, we had more meat on the story by that time, we saw others work.

One of the odd one being (F below) natalizumab that augmented numbers in the blood, but as the idea is that it stops trafficking into the brain, this should be expected.

Daclizumab is another one of the odd treatments. However once you realise that memory B cells express CD25 and you can see that dacliziumab depletes CD19 B cells, you get in a groove and start looking for it, and find supportive evidence pop-up all over the place.  There is also a poster on teriflunomide which supports this concept too.

Some publishers wanted too much money for us to include every example for every treatment in the publication. But we give some references if you want to look and the paper is “open source”.

So this was revelation part one. The next one has additional implications.

Importantly the B memory cell idea gives more insight on ocrelizumab to suggest that CD20 treatment may be over-used to achieve benefit.

The data can be interpreted to suggest that ocrelizumab and alemtuzumab could work in the same way (bad news for both companies, sorry). 

They are probably “both” induction therapies of PIRT (pulsed immune reconsitution therapies)

One wonders why such important data has not been properly published? It was dumped in an ECTRIMS abstract in 2012 and an AAN abstract in 2013 (where we found them. Do neurologists have enough time in a couple of hours to see and digest a 1000 posters?) and so there is no excuse from the companies or authors that they have not had enough time to publish the work.  There are enough journals desperate for papers. 

Now look at the figure above and you will see there is a clear inference that memory B cell levels may be a biomarker for treatment responsiveness. 

Looking at figure E it shows only a few people on beta interferon have less than 20% of memory CD19 B cells. Are these the responders? Fingolimod (E) has more, but some failures. In (B) there is more depletion with mitoxantrone and (C) rituximab shows a lot of depletion, which is consistent with its high efficacy in MS.

There are studies in other conditions that use rituximab showing that CD19, CD27 numbers rise in the 4 months before a relapse and can be used as a biomarker to inform on retreatment such that one may be able to use 2/3 less rituximab, which is both safer and cheaper and takes us some way forward to personalised medicine.

In some treatments there are some indivudals that make more memory B cells and treatments that do this have in some cases been associated with MS worsening

Importantly memory B cells are the population of cells that accumulates in the CNS in MS and they express the EBV receptor, so depletion of the memory B cell subset depletes the antigen presenting cell that can activate T cells, or perhaps it eliminates the viral reservoir and removes the trigger of MS. 

In many cases both unswitched (Have not change the antibody subtype they can produce) and switched (IgM and IgD are down regulated and IgG is upregulated from the unswtichted to switched) memory B cells are affected. Which subset is the critical one, the one that has already been reported to be increased in MS.

The T cell immunologists, will have a hard time with this concept, and their stock answer it is because it blocks antigen presentation, but the simple answer is what data argues against this concept? 

Is the data that suggests a T cell action solid or is it simply circumstantial. Have a read of the paper for your self it is open access (CLICK to DOWNLOAD).

We suggests ways forward to perhaps make treatments safer and better.

Maybe the data miners out there can do some trawling and see what’s against the idea.  Send them (paper titles) in and lets see if the idea flies. 

We have thrown down the gauntlet, and if read, people will either agree or disagree.  

About the author



  • I believe in the B cell hypothesis you are offering here, but I think that there is something else at play that is still elusive. If you are selective in what you look for in terms of treatment impact, I wonder if you would find other changes in the body occurring with medical intervention for MS? Perhaps you would show that there are other depletions… and that would send you down a totally different rabbit hole.

    Clearly B cells are an issue. But are they the whole issue?

    And if you do identify rogue B cells as the culprit – then what does that mean, not just for prevention, but for the root cause of the disease? What are the correlations between B cells and Vitamin D? B cells and EBV?

    Once upon a time T cells were seen, again and again, as the target cell. Research went out to support that dogma, and did so again and again. Beware the mistakes of the past.

    • Following dogma is a problem as it does become self sustaining

      So must get next paper written to question that EAE experiments are the way to go. Another lead balloon coming your way….Yes we should save the science lemmings and save a load of beasties in the process.

  • Please read the final paragraph of the paper. The relationship between B cell memory and EBV is there. These cells are the B cells infected by EBV. I suspect vitamin D is important in repertoire and immune development. There is a clear link with diabetes and vitamin D and remember that diabetes starts in childhood so the inference is that the impact will be early in life hence the month of birth effects.

    It seems to be a week for rabbit holes and so you are right about putting all eggs in one basket.Quite dangerous but for a long long time most of those eggs have been T cell eggs.

    There is a problem that this phenotyping is done using flow cytometry and this outputs in proportions so as one goes down something else goes up. Is it B reg?

    But there are some testable questions we are waiting for pharma to trawl their trial data

    • hahaha… while (i thought) your efforts to link b cells to ebv were a tiny bit less convincing (in the context of causation) then your table setting out the control/destruction of memory b cells by current dmts thought to control ms on basis of t cell action, i was so very impressed by your dedication to tie it all together with a convincing theory 🙂

      you made such a compelling argument: and i agree – disagree with it by using evidence against the info in table 4: if you can do that, we will be all ears to hear the arguments because they could be informative.

      but if you can't do that.. are you just arguing a religion?

    • Causation and EBV is a profG thang and there have been numerous reviews on this. I need to do more reading on this.

      OK wait till Monday and I,ll use the same logic to tie memory T cells into the table 4. How,s that?

    • As to evidence against I can't remember seeing much using the searches done, maybe we can pick populations and do it all again.

    • CD21 is the receptor EBV uses to infect cells. This is a B cell receptor that binds inactivated C3 complement C3d

    • I'm not going to give personal or anecdotal reports, but months before my first relapse in my life I had a relapse of a mononucleosis that I thought would kill me.
      If EBV is linked even to cancers, why would not it be associated with other diseases?
      Until the scientific community and society itself, it's thought that the outbreak of Zika Virus was directly linked to the numerous cases of Microcephaly in infants in Brazil, there was a last flurry of disbelief and denial.

    • If EBV is linked even to cancers, why would not it be associated with other diseases?

      I suspect (but could be wrong) that there could be a number of diseases thought to be linked to EBV (but not listed in the 'official' list of EBV disease associations). EBV Zika et al….. why on earth would our world pollies care about them while there is no direct association? zika turned out to be visible and so some money was spent on it (on research and on other measures)… but if you can't see it, why care about it? most ebv related diseases don't come with pictures of babies with abnormally shaped heads…

    • You may even be right that many of the EBV/related diseases "don't cause deformations", but what if EBV really is the driving force behind the development of MS in genetically susceptible individuals?
      If you have MS you may not have an "apparent" strain on your head because of a virus, but you are having your CNS eaten "inside."
      So your statement is somewhat incoherent.

    • Every hypothesis has its "Achilles heel." For example the gender gap the development of MS. Women further develop the disease, it's not an inherited disease but has common genetic factors involved. But someone asks, "How do you explain the cases of MS in men? Is it really related because GA plus oestradiol did not stop the disease?" And a lot of whys who could put hypotheses to the ground. But the important question is: why Ocrelizumab can act on Primary Progressive MS, and the other DMTs have not been able to do this? It is already clear that there is something with B lymphocytes.
      So checking if EBV can be behind this is perfectly plausible, since it infects B cells.
      If it were so nobody would but try to find out the cause and involved factors of anything, illness some.

    • why ocrelizumab works in progressive MS. I suspet that because it was loaded with young pwMs with active disease that would respond to peripherally acting immunosuppressive agents

  • The funny thing about this journal as it asks for the outstanding questions so our ideas for free again…Happy experimenting

    • I know what you mean about ideas for free MD. I was doing my postgrad and my supervisor said they hadn't heard before of one idea I suggested. Fast forward a year and I look on their university website page and it said that they are currently working with that idea!
      I know try and keep my research ideas quiet which is hard to do when I'm enthusiastic.

  • This looks to me like a good explanation for the natalizumab withdrawal rebound effect. Natalizumab increases b-cells in the blood but blocks them from entering the CNS.
    When one goes off natalizumab the increased number of b-cells in the blood are now free to enter CNS and lead to the dreaded rebound effect.

  • How long do you think it will be before Pharma accepts this and focus primarily on B-cells? What does this means for drugs in the pipeline such as ublituximab? Or do you think we need to start from scratch again in developing an effective, targeted treatment?

    • Pharma are already focusing on B cells, but they have not always got it right like atacicept and tabalumab, but there are other ways to target these cells.

      Ublituximab is another CD20 antibody, like ocrelizumab and ofatumumab they will need to do their trials just like Roche and Novartis have done.

      However you can start from scratch to make more focused treatments

  • How does your B cell hypothesis Relate to this :
    B cells
    For a long time, B cells were thought to be important in MS in their capacity
    to differentiate into plasma cells and produce antibodies.36 The advent of B
    cell depleting therapies has challenged this view. It was noted that clinical
    improvement in patients treated with rituximab often preceded reduction in
    autoantibody levels.52 Further, in a phase II trial of atacicept, a fusion protein
    that blocks plasma cell function and the late stages of B cell development,
    treatment was found to aggravate MS.53 This is in contrast to rituximab,
    which in clinical trials has been proven to be effective in MS.54, 55 The above
    suggests that B cells are important in some other capacity than antibody
    production. Rituximab treatment results in a noticeable decline of T cell
    numbers in CNS of treated patients, suggesting that B cells sustain
    pathogenic T cell responses. Recently, it was demonstrated that IL-6
    production is the major mechanism of B cell contribution to the pathogenesis
    of EAE (experimental autoimmune encephalomyelitis, an experimental
    model of MS), and also that this inflammatory pathway was increased in
    RRMS patients.52 Another possibility is that the B cells act as antigen
    presenting cells.56
    Joachim Burman

  • Where are the anti-CD79a antibodies? There's robust evidence that plasma cells play a role in MS pathogenesis. Intrathecal rituximab + a cd79a antibody could be a true game changer, maybe even a perfect therapy.

    • Intrathecal is probably not the answer as it is at the bottom of the outflow mechanism..this was shown by the lack of B cell depletion following intrathecal served to wiper out B cells in the blood. However, your point about targeting the plasma cells is well taken

    • Thanks for replying
      "There was a
      concomitant depletion of total CSF lymphocytes
      and B lymphocytes, albeit with low baseline counts.
      Although the rationale for administering rituximab
      IT was to achieve an effective therapeutic antibody
      concentration within the BBB, we observed a rapid
      and potent effect on lymphocytes in the peripheral

      Only 2 patients had a high enough baseline
      CSF total cell count to be able to reliably discern a
      change posttreatment. Both showed an initial depletion
      of B lymphocytes with a return to baseline by 6
      months. Total CSF lymphocyte counts also decreased
      to almost zero in these 2 patients, which cannot be
      explained by B-lymphocyte depletion alone. This is in
      agreement with previous data on CSF T cells after
      rituximab treatment, but the exact mechanisms are
      unknown.12 The recently described effect of rituximab
      on a subset of T cells expressing CD20 is an
      interesting mechanism that needs to be investigated

      A weakness in our study is that we cannot determine
      whether changes occurring among lymphocyte
      subsets in the CSF are an effect of rituximab within
      the CNS compartment or are a result of a peripheral
      depletion and subsequently less recruitment into the
      CSF from PB. It is presently not known whether
      appropriate effector mechanisms exist in the CSF to
      mediate lysis by injected monoclonal antibodies.
      However, there is now believed to be an inflammatory
      milieu along the meninges in many cases of
      PMS14 that could possibly facilitate both
      complement-mediated lysis and antibody-dependent
      cytotoxicity. There are also data indicating increased
      complement activation in MS, with the highest
      occurrence among progressive patients.15 Furthermore,
      it was shown that IT administration of anti-
      CD20 monoclonal antibody could reduce the
      amount of B lymphocytes in the meninges in EAE
      with a concomitant modest amelioration of the clinical

      There were significant reductions in CD20+ and CD19+ cells in blood from baseline (p<0.0001) but not in CSF.
      While CSF B cells were killed robustly (median -79.71%, P = 0.0176), B cells in CNS tissue were depleted inadequately (~-10-20%, P < 0.0001). Consequently, the T-cell-specific CSF biomarker sCD27 decreased slightly (-10.97%, P = 0.0005), while axonal damage marker, neurofilament light chain did not change. Insufficient saturation of CD20, lack of lytic complement, and paucity of cytotoxic CD56(dim) NK cells contribute to decreased efficacy of rituximab in the CNS.

      The Peripheral Blood B-lymphocyte count dropped rapidly after the first 2 injections (total dose of 3.5 mg IT rituximab) to undetectable levels. Three 25-mg doses given once per week depleted peripheral B lymphocytes entirely for the following 3-6 month period. Depletion occurred
      Peripheral CD20+ B cells were reduced, while oligoclonal bands were unaffected

      By contrast, the initially completely suppressed B and plasma cells in both the blood and CSF reappeared after 5 and 10 months, CSF cells.

      There was reduction of T cells after rituximab, but if you look at the data this was months after the treatment. So rather than antibody killing T cells to control disease, one could argue the control of disease lead to drop in cells being retained in the CNS.

      Do you think The CD20 expression by T cells is the answer or is it the cop-out that allows t cell immunolologists to keep going 🙁

    • Interesting MD. Very interesting. It seems the intrathecal route is not the answer. What would happen if the antibodies were delivered into the CSF of the cranium? I think at that point HSCT using small molecules becomes a lot safer and easier but it's interesting to think about at least.

    • Thanks
      For your contribution .Only have acess to the second paper
      The others papers that csf b cells depletion is modest and OCB continue
      Do you think that OCB positive are a good marker for poor disease outcome?

      Thanks Luis

    • I dont think t cell expression cd20 is the answer (they make up only 10% of t cell population)
      I think one as to look only the b cell and t cells ,maybe even in inate immunity

    • I heard there was talk of intraventricular delivery, personally I think we should be looking for a small molecule inhibitor that gets in the CNS.

    • Cyclophosphamide has robust CNS penetration and kills every immune cell there is. Sort of reinforces that HSCT really is the holy grail.

    • Kills every immune cell, hair cell, sperm cell, intestine cell..OK I'm being dramatic however, it only kills proliferating cells and so may miss stuuf and unless HSCT uses cyclophosphamide it doesn't target the CNS

    • In this study Allogenic Hsct the manage to eradicate OCB

      A handful of cases of MS patients treated with allo-HCT for malignant hematological disease have been published [1]. In two recent reports continued disease activity was observed after transplantation [9, 10]. In our patient allo-HCT was associated with the disappearance of IgG oligoclonal bands. This has also been observed in auto-HCT-treated MS patients [2]. However, serial CSF studies showed that even though IgG oligoclonal bands disappeared, CXCL13 and MMP-9 concentrations were increased. CXCL13 is expressed in perivascular infiltrates and within the brain parenchyma in MS, and CSF concentrations of CXCL13 are increased in secondary progressive MS [5, 11]. Secondary progressive MS patients with a more severe disease course harbor lymphoid follicle-like structures in the meninges, and CXCL13 may be involved in the generation of these structures [12]. MMP-9 in CSF has also been associated with disease activity in MS [13]. In an autopsy study it was observed that although lymphocytes and plasma cells were almost absent after auto-HCT, there was ongoing microglia activation, demyelination, and axonal pathology [14]. We hypothesize that CXCL13 and MMP-9 detected in the CSF of our patient may reflect ongoing, pathogenic, immune activation associated with continuing tissue damage even after the eradication of intrathecal IgG synthesis. This is consistent with the notion that progressive MS may depend more on innate than on adaptive immune activation

By MouseDoctor



Recent Posts

Recent Comments