Anti-CD20 therapy has been shown to be highly effective at depleting the existing population of immune B cells in patients with multiple sclerosis (MS).1 The goal of therapy is to remove impaired B cells from the immune system, setting the stage for repopulation with a healthier generation of new B cells. Little is known about the maturation and activation of the repopulated cells, or any other immune cells, once a course of anti-CD20 therapy has been completed in this setting.

This Journal Club discussion will focus on a study by Nissimov and colleagues that evaluated the character of the new population of B cells after anti-CD20 therapy with rituximab, as well as the effects of repopulation on the frequency, differentiation, and activity of T cells and myeloid cells.1 The results, which showed differences in the B-cell populations before vs after anti-CD20 therapy, led our Journal Club reviewers to pose an intriguing question: Can anti-CD20 therapy be used for immune-system remodeling in the treatment of MS?


The study cohort included a heterogeneous population of 15 patients with relapsing-remitting MS who underwent B-cell depletion with rituximab 1000 mg.1 The frequency and function of B cells, T cells, and myeloid cells were analyzed before administration of the anti-CD20 agent and at different time intervals over a period of 24 months following treatment.

The majority of B-cell types identified in the 15 study patients were either naive B cells (45.5% ± 3.1%; mean ± SEM) or memory B cells (36.8% ± 3.1%). The patients were then stratified into phenotypes by the ratio of naive B cells to memory B cells. A ratio of 1 or less was designated as memory/balanced type and a ratio greater than 1 was designated as naive type. For the balanced type, the difference between the frequencies of naive vs memory B cells could not exceed 5%.


Anti-CD20 treatment was found to be effective for all study participants. When compared with the phenotypes recorded before the initiation of B-cell-depletion therapy, the repopulating B cells appeared both less mature and more activated.1 The reappearing population had a much higher percentage of transitional B cells (58.8% ± 5.2%) compared with that before B-cell depletion (10.1% ± 1.9%), and the percentage of mature naive phenotypes was significantly altered (before treatment, 45.5% ± 3.1%;  after treatment, 25.1% ± 3.5%). The frequency of memory B cells decreased following treatment (before, 36.7% ± 3.1%; after, 8.9% ± 1.7%).

Enhanced expression of activation markers CD25 and CD69 was noted in the new population of B cells. Expression of CD25 rose from 2.1% (± 0.4%) prior to treatment to 9.3% (± 2.1%) after treatment; CD69 expression rose from 5.9% (± 1.0%) prior to treatment to 21.4% (± 3.0%) after treatment. The investigators also observed that the reappearing B cells expressed significantly higher levels of costimulatory CD40 and CD86.

Changes in T cells were also evident, including increases in naive CD4+ T cells (before, 11.8% ± 1.3%; after, 18.4% ± 3.4%) and CD8+ T cells (before, 12.5% ± 1.4%; after, 16.5% ± 2.3%), as well as decreases in terminally differentiated subsets of both CD4+ T cells (before, 47.3% ± 3.2%; after, 34.4% ± 3.7%) and CD8+ T cells (before, 53.7% ± 2.1%; after, 49.1% ± 2.7%).


Benjamin Greenberg, MD, MHS

Natalia Gonzalez Caldito, MD

Victor Salinas, MD, PhD

Benjamin Greenberg, MD, MHS, professor and the Cain Denius Scholar in Mobility Disorders; Natalia Gonzales Caldito, MD, fourth-year resident; and Victor Salinas, MD, PhD, fourth-year resident, Department of Neurology, University of Texas Southwestern Medical Center, Dallas participated in this journal club discussion.

Benjamin Greenberg, MD, MHS: We selected this article to review for a variety of reasons. One goal was to evaluate the findings in the context of the therapies that we currently use for MS. As you are both aware, we separate our therapies into broad categories. We have immunomodulatory therapies, immunosuppressive therapies, and then what some people call the “immune-remodeling” therapies, which involve administering a drug that depletes the immune system in order to regrow a new immune system. Classic examples of drugs used for immune remodeling are alemtuzumab and cladribine.

For the last decade at least, we have thought of anti-CD20 therapies — the ones currently approved for MS by the US Food and Drug Administration are ocrelizumab

and ofatumumab — as immune-suppressive therapies. You are used to using these drugs for the treatment of a variety of forms of MS. The idea is to start patients on the drug and keep them on it because as long as the immune system is suppressed, patients remain in remission. There is a growing question, however, about whether you can consider using an anti-CD20 drug as an immune-remodeling therapy.

For example, after suppressing the immune system for a period of time by inducing apoptosis of CD20-positive B cells, and as the drug wears off, is the patient’s immune system different from how it was prior to treatment?

I thought this paper by Nissimov and colleagues from Göttingen, Germany, would be interesting to explore because it looks from an immunologic perspective at what happens to the immune system over time after exposure to anti-CD20 therapy.

So Victor, can you provide an overview of the study design?

Victor Salinas, MD, PhD: I would say that the investigators profiled the B-cell populations in patients at various times before and after receiving medication by using cell surface markers to characterize different B cells.

BG: Yes, they did these samplings, as you mentioned, of the B cells both before and after treatment, and then they profiled several different compartments within the immune system, such as B cells, T cells, and even monocytes. Natalia, looking at this, how would you explain their categorization?

Natalia Gonzales Caldito, MD: The investigators categorized them into 5 different groups: transitional B cells, naive B cells, antigen-experienced B cells, memory B cells, and plasmablasts. Then they also divided the 15 patients into 2 groups according to immunophenotype.

BG: Along those lines, what are your thoughts about the phenotyping, using a paradigm of naive-type vs balanced-type?

NGC:  I thought that was an interesting way to do phenotyping. The investigators did it by the ratio of memory B cells to naive B cells. In the balanced type, the ratio was 1 or less, whereas in the naive group it favored the naive cells (ratio >1). In this paper, graph B in Figure 1 shows that while they make the distinction, the ratio of memory B cells to naive B cells is actually very similar in some patients.

BG: As they described it, there was a “heterogeneity” between patients. The authors assigned a cutoff point where, depending on whether the ratio is more than or less than 1, patients were split into 2 groups, but the patients in the middle are quite similar. What they are saying is that even though all of these patients have MS, they come to their B-cell-depleting therapy with different ratios of these B-cell types.

Also, before the investigators even go on to evaluate B-cell depletion, they looked at the association between B-cell populations and patients’ expanded disability status scale (EDSS) scores, as depicted by the interesting graphs in Figure 1. They definitely documented statistical significance in terms of positively correlating the frequency of naive B cells with the EDSS score. You look at that and see an r value of 0.9, which is a very significant r value. Victor, as someone who has spent a lot of time in the world of immunology and MS, what do you take away from that?

VS: Given that the EDSS data used to generate this graph were obviously from the present, the key question is: If we were to follow these patients, how would this relationship fare if we were to start with patients who are completely treatment naive? The study authors tell us that most of these patients had been treated with some agent prior to enrollment, so maybe there is some confounder — and clearly they have been treated with different agents, so I do not think that is necessarily a fair comparison — but it is pretty remarkable. This [naive B-cell status] has possible prognostic value that is worth looking into, and there’s also the question of whether it would be a useful biomarker for determining the efficacy of these therapies.

Although this is a very limited study with just 15 patients, it is the kind of relationship that you could build a research program around.

BG:  Yes, I agree with you. We think of MS as being a disease in which the immune system matures over time with antigen specificity against the cerebrospinal fluid (CSF) antigen, so you would expect more memory B cells than you have — meaning they have been primed and activated, they are antigen-specific, and they are on the hunt for CSF antigens — and that this would associate with higher disability scores. In fact, though, Figure 1F of this study shows that the reverse has happened: the more naive a population of B cells the patients had, the worse they fared.

I want to draw your attention to one of the things that the authors do, which is to remind us of the cytokine profiles associated with naive B cells vs memory B cells. They tell us that the naive B cells are the major producers of interleukin (IL)-10, whereas the memory B cells are much more likely to produce tumor necrosis factor-alpha (TNF-α), lymphotoxin, and other cytokines. So although we think of B cells as being disease-propagating in MS, you could consider a paradigm in which an overabundance of naive B cells is the thing driving disability and maybe there are protective memory B cells. It is just a different way of looking at the disease in terms of intrinsic B-cell populations, and I agree that it was counterintuitive to what you might expect.

What struck each of you the most about the data after B-cell depletion?

VS: One thing that the authors emphasize is the degree of heterogeneity in their repopulation dynamics, but clearly there is a robust result after B-cell treatment showing a change in the population subset that looks not as heterogeneous afterward since you end up with a larger population of naive B cells as opposed to memory B cells. I think that is a very robust result.

Obviously the next question is: what is the phenotype of naive B cells? The study authors do characterize it with a subset of cytokine and activation markers. We gleaned that they have a more activated phenotype, but this also gets us thinking about what else could be different, what the overall pattern might be, and in what other directions we can we go with these data.

BG: You’re right, I think it was a robust result. I do not think it was a surprise to see an augmentation of transitional B cells during repopulation after treatment with an anti-CD20 drug because you are going to repopulate from the bone marrow and the first cells we are going to see are transitional; so that did not shock me. Their IL-6 data surprised me.

You mentioned that they saw a more activated cell population. What do we do with those data?  We have treated a patient with a monoclonal antibody, we have depleted this population of naive and memory B cells, they start to grow back with transitional B cells, but then they are actually secreting more IL-6 than they would in their baseline state. Do you think we have enough data to say whether that would be a good, bad, or indifferent outcome in a patient with MS or are we not there yet?

VS: The intuition is that because the cell population is more activated, you get concerned about the phenomenon of repletion and making sure that the frequency of scheduling of this medication is such that we prevent B-cell population. The other issue is that this study is just looking at IL-6. These cells could be more complicated. There could be other cell types that have this, even the regulatory B cells. For all we know, these repopulated B cells may represent some sort of spectrum between regulatory B cells and naive B cells. Here, in this study, we have only looked at a subset of B-cell cytokines, so we have to explore this further. The main concern is that B cells are perhaps more activated, and we have to make sure that we prevent a rebound of clinical disease.

NGC: I was also surprised by why the repopulation led to a significant increase in IL-6. There were no clinical relapses in these 15 patients. so I do not know how to interpret this IL-6 increase. It is thought to be a proinflammatory cytokine related to MS disease, with some indication that IL-6 can be a marker in the CSF,2 but we do not have enough information to see the significance in these patients. This needs to be studied further to ascertain whether IL-6 in the context of the repopulation of B cells is not as dangerous as it was before.

BG: Beyond the results related to activated B cells, both in terms of cell surface markers and cytokine production, there was a fascinating additional arm to this study, which was to phenotype the T-cell populations. I do not think this area has gotten enough attention over the years in terms of integrating the lymphocyte compartments between B cells and T cells.

Natalia, could you walk us through the first portion of T-cell data, outlining what they saw in terms of CD4+ and CD8+ T cells before and after B-cell depletion?

NGC: The main observation regarding the B-cell population was an increase in the CD4+ to CD8+ T-cell ratio, although the number of T cells did not change overall. For CD4+ T cells, the authors showed a proportional increase in naive, central memory, and effector CD4+ T cells, while levels of the terminally differentiated CD4+ T cells were decreased. Regarding CD8+ T cells, they saw similar changes. but the levels of memory CD8+ T cells were unchanged. This interesting observation made me think about how, although anti-CD20 therapies are supposed to target B cells, some studies show that T cells can also express CD20 on their surfaces.3,4 The question is whether that is due to a lack of assimilation with the B cells or due to anti-CD20 treatments also affecting the B cells.

BG: Yes, I agree that the investigators were not able to differentiate between the T-cell population changes being an indirect consequence of depleted B cells vs a direct effect of the monoclonal antibody on those rare CD20-positive T cells. Sorting that out would be a fascinating immunologic exercise that would have repercussions for patients.

Within the discussion of the T-cell data, there was one paragraph, Victor, that — as someone who likes to talk about finding biomarkers for prognosis — I found fascinating. It was the paragraph on drawing correlations between the baseline B-cell phenotyping and what happened in terms of T-cell differentiation after therapy. I call it the “crystal ball paragraph,” where they looked at whether you had a memory-balance type vs naive-type B-cell population and what happened to T cells after depletion. Did that stand out to you and do you want to highlight the results of that analysis?

VS: I will admit that it does not [stand out to me]. I harbored other questions. One of my questions in general relates to the body of data itself. It struck me that, despite having known data for most of the patients (at least 10 of 15), the data were analyzed at the level of the population sample rather than the individual. I don’t understand why. When I see the error bars showing a statistical difference, it may just be due to the fact that the investigators were looking at medians or averages. What I would like to see is intra-individual trajectories because I would argue that there’s a significance to the trajectories between individuals. I actually did not have much confidence in the data as reported. 

I think you are talking about Figure 5, where they compare changes in the frequencies of the different T-cell subsets within the 2 populations that Natalia noted. I wanted to bring up a small point: I would have liked to see the structure of the data population. I guess there is a rationale for separating the population into naive vs balanced types, but is there actual structure in the data that supports that artificial separation? The authors did not provide this information. Those 2 things influenced the way that I looked at the study.

That said, the data do show, at least for naive cells (Figure 5A) before anti-CD20 therapy, that when the B cells re-emerged, there was a slight increase in the frequency of naive CD4+ T cells. It looks as if this is the case for CD8+ T cells too, albeit with some variability. The data show a similar increase in the frequency of central memory CD8+ and CD4+ T cells, and a decrease in the terminally differentiated cells. I think the authors emphasized that as well.

In the lower panels of Figure 5, CD62 is an activation marker and it is upregulated in cells that home to secondary lymphoid organs. This is an interesting result. It looks as if there is increased activation in the CD62 expression of CD4+, again suggesting that the T cells that emerge after repletion have this quasi-activated phenotype. This bears looking into, but I would like to see the data presented differently so I can see just how robust their conclusions are about the findings.

BG: Based on these data and what you have both discussed, particularly this repopulation of T cells after depletion, is it reasonable to explore anti-CD20 therapy as an immune remodeler? Is there a scientific basis to think about this as a strategy in the management of MS? After answering, can each of you briefly summarize what you consider to be a key takeaway from the paper?

NGC: I definitely think that, after reading this article and seeing the change in the immunophenotype of the B cells and T cells, anti-CD20 therapy could be considered for remodeling therapy. For me, this study brings home how anti-CD20 therapies affect not only B cells but also T cells, and it highlights the complexities behind these therapies.

VS: I also think the perspective that we may be resetting the immune system in some way is very interesting. For me, one of the big takeaways is that we can look at this study as an experiment in how we understand the immune system and its abnormal function in MS.  I think it is worth trying similar approaches to get at this question by using data from a larger number of patients.

BG: My conclusion is a combination of what both of you have said. I think it is worth looking at anti-CD20 therapy as a possible immune remodeler in some patients using their baseline immunologic phenotype. One thing I would take away from this study is that rather than treating all patients with MS in a clinical trial with the same therapy based on their initial diagnosis, a patient’s immunologic phenotype at baseline may determine the individual response or lack of response to a certain therapeutic paradigm. We may find that certain patients can use anti-CD20 monoclonal antibodies as an immune remodeler while others may just use this therapy as an immunosuppressant. I think this is definitely worthy of further study. This would be a nice way to tie together a patient’s specific biomarkers and immunologic phenotype to predict that person’s response to a given therapeutic intervention. 

NGC: Besides my concern about the small size of this study, I was also worried about the scheduling of the rituximab. Not every patient received rituximab every 6 months, which is standard practice in the clinic. Ideally, this study would be repeated in a more protocolized manner in the future.

This Q&A was edited for clarity and length.


1. Nissimov N, Hajiyeva Z, Torke S, et al. B cells reappear less mature and more activated after their anti-CD20-mediated depletion in multiple sclerosis.  Proc Natl Acad Sci USA. 2020;13;117(41):25690-25699. doi:10.1073/pnas.2012249117

2. Stampanoni Bassi M, Iezzi E, Drulovic J, et al. IL-6 in the cerebrospinal fluid signals disease activity in multiple sclerosis. Front Cell Neurosci. 2020;14:120. doi:10.3389/fncel.2020.00120

3. Vlaming M, Bilemjian V, Freile JÁ, et al. CD20 positive CD8 T cells are a unique and transcriptionally-distinct subset of T cells with distinct transmigration propertiesSci Rep. 2021;11(1):20499. doi:10.1038/s41598-021-00007-0

4. Palanichamy A, Jahn S, Nickles D, et al. Rituximab efficiently depletes increased CD20-expressing T cells in multiple sclerosis patients. J Immunol. 2014;193(2):580-586. doi:10.4049/jimmunol.1400118

Posted by Haymarket’s Clinical Content Hub. The editorial staff of Neurology Advisor had no role in this content’s preparation.

Reviewed February 2022