Dr. Mulleman has received consulting fees and speaking fees from Pfizer and MSD (less than $10,000 each), has participated on behalf of his institution in clinical trials sponsored by Abbott, Roche, Bristol-Myers Squibb, Pfizer, UCB, and MSD, and has been invited to attend international congresses by MSD, Roche, Bristol-Myers Squibb, and Abbott.
Université François-Rabelais de Tours, CNRS, UMR 7292, and Centre Hospitalier Régional Universitaire de Tours, Tours, France
Dr. Goupille has received consulting fees, speaking fees, and/or honoraria from Abbott, Bristol-Myers Squibb, Eli Lilly, MSD, Novartis, Pfizer, Roche, and UCB (less than $10,000 each), has participated on behalf of his institution in clinical trials sponsored by Abbott, Roche, Bristol-Myers Squibb, Eli Lilly, Novartis, Pfizer, UCB, and MSD, and has been invited to attend international congresses by MSD, Roche, Bristol-Myers Squibb, and Abbott.
Centre Hospitalier Régional Universitaire de Tours, Tours, France
Rituximab, a monoclonal antibody specifically targeting CD20, induces B cell depletion and is effective in the treatment of rheumatoid arthritis (RA). This study was undertaken to evaluate whether routine monitoring of lymphocyte subpopulations, especially T cells, may be useful in patients receiving rituximab for RA.
We examined data on all RA patients receiving rituximab between July 2007 and November 2012 in our center. Peripheral blood CD3+, CD4+, CD8+, CD3−CD56+, and CD19+ lymphocyte counts before and during the first course of rituximab were measured by flow cytometry. The Mann-Whitney nonparametric test was used to compare lymphocyte subpopulation counts before and during treatment.
Data on 52 patients were examined. Rituximab induced unexpected and substantial depletion of T cells, mainly CD4+ cells, in most patients. The CD4+ cell count decreased by a mean ± SD of 37 ± 33% as compared to baseline at week 12, reaching <200 cells/μl in 3 patients. Importantly, lack of CD4+ cell depletion was associated with no clinical response. Therefore, the mechanism of action of rituximab may depend at least in part on T cells.
Rituximab induces substantial T cell depletion, mainly of CD4+ cells, which is associated with the clinical response in RA. Routine monitoring of T cells may be useful in the clinical setting of RA.
T cells, especially CD4+ T cells, and B cells are both considered to be involved in the pathogenesis of rheumatoid arthritis (RA). Decreasing lymphocyte activity with conventional or monoclonal antibody–based treatments can reduce disease activity. The first evidence that CD4+ T lymphocytes were involved in the pathogenesis of RA was the association of the disease with polymorphic HLA class II alleles, such as HLA–DRB1, which present antigenic peptides to CD4+ T cells (). Consistent with this finding, targeting of CD4+ T cells with anti-CD4 monoclonal antibodies has resulted in clinical improvement, although modest, in preliminary clinical trials ([2, 3]). Moreover, abatacept, a CTLA-4-Fc recombinant fusion protein that inhibits CD4+ T cell activation by antigen-presenting cells (APCs), has been approved for the treatment of RA ().
However, B cells are responsible for the production of autoantibodies such as rheumatoid factor (RF) and anti–citrullinated protein antibodies (ACPAs), which are hallmarks of RA. Rituximab, a chimeric anti-CD20 monoclonal antibody that induces the depletion of mature B cells and pre–B cells, is effective and currently used in the treatment of RA ([5, 6]). The roles of B and T cells in the pathogenesis of RA and autoimmune diseases may be tightly linked because of the APC functions or cytokine secretion capacities of B cells. Activated B cells are thought to be fundamental in coordinating T cell functions. Indeed, transgenic mice prone to a spectrum of autoimmune diseases (glomerulonephritis, vasculitis, and skin disease) were protected after B cell depletion, which is itself associated with a marked decrease in CD4+ and CD8+ cell populations (). Sfikakis et al highlighted that after B cell depletion, decreased T helper cell activation was associated with clinical remission of lupus nephritis and suggested that B cells play a role in promoting the disease, independent of autoantibody production (). Saadoun et al showed that T cell abnormalities found in mixed cryoglobulinemia, such as CD8+ T cell activation, could be reversed after treatment with rituximab, and that this reversal was associated with a complete clinical response ().
Although rituximab seems to be modestly more effective in RF-positive than in RF-negative patients, its efficacy is not restricted to patients who are autoantibody positive, so T cell functions may also be hampered by rituximab treatment ([10, 11]). Several cases of opportunistic infections, usually observed in CD4+-deficient patients, have been reported in rituximab-treated patients with RA ([12-17]), which supports the concept of an action of rituximab on the T cell arm of the pathogenesis. In addition, some of these patients, who had low pretreatment CD4+ levels, showed a further decrease in CD4+ counts at the time of the opportunistic infection ([12, 14]).
Although lymphocyte phenotyping can be recommended before each treatment course of rituximab to identify patients at high risk of infection (), changes in lymphocyte subpopulation counts during rituximab treatment have been essentially documented for the B cell compartment only. In a pilot study, Vital et al, using highly sensitive flow cytometry, showed that the degree of B cell depletion rather than dose of rituximab determined the clinical response (). However, an increase in the IgD+CD27+ memory B cell subset at the time of B cell population recovery was associated with nonresponse (). In this study, we investigated lymphocyte count changes, with particular attention to T cells, during rituximab treatment and examined the association between changes in lymphocyte population counts and clinical response in RA patients.
PATIENTS AND METHODS
Patients and study protocol
The study included patients with RA who received rituximab between July 2007 and November 2012 in the rheumatology department of the University Hospital Centre of Tours, France. The treatment protocol was designed in accordance with the guidelines of the French Society of Rheumatology (). Patients received a 1,000-mg infusion of rituximab preceded by a 100-mg intravenous pulse of methylprednisolone, unless contraindicated, at baseline and week 2. Patients were followed up and evaluated for the recurrence of symptoms at week 12, week 24, and once between weeks 36 and 48. Disease activity was assessed using the Disease Activity Score in 28 joints (DAS28) () before treatment and at each followup visit. In our center, lymphocyte phenotyping by flow cytometry is routinely performed to check the level of B cell depletion induced by rituximab before the first and second infusion of rituximab and at each followup visit. Therefore, ethics approval and written consent were not sought for this analysis. Data on demographic, clinical, and biologic variables were collected at the time of initiation of treatment. Patients were classified at week 24, according to the European League Against Rheumatism (EULAR) criteria (), as good responders, moderate responders, or nonresponders.
Lymphocyte phenotyping by flow cytometry
Phenotype analysis was performed at the Immunology Laboratory, University Hospital Center of Tours, according to a standard no-wash whole-blood procedure using a PrepPlus workstation (Beckman Coulter). Blood samples (100 μl) were incubated with fluorescein isothiocyanate (FITC)–conjugated anti-CD3, phycoerythrin (PE)–conjugated anti-CD56, and PE-Cy5–conjugated anti-CD19 antibodies (20 μl), or with FITC-conjugated anti-CD8 antibody, or with PE-conjugated anti-CD4, FITC-conjugated anti-CD45RA, PE-conjugated anti-CD45RO, and PE-Cy5–conjugated anti-CD4 antibodies (all from Beckman Coulter) for 15 minutes at 18–20°C. Red blood cell lysis and cell fixation were performed using a TQ-Prep workstation and ImmunoPrep reagent system (Beckman Coulter). Cells were analyzed with an Epics XL-MCL flow cytometer (Beckman Coulter). Gates were set on lymphocytes with forward and side scatters, and data were collected for a minimum of 5 × 103 events for each determination.
The Mann-Whitney test was used to compare lymphocyte counts before treatment, at week 2, week 12, week 24, and between weeks 36 and 48. Absolute lymphocyte counts, percentage depletion, and absolute values of depletion were compared before and after treatment. Percentages and absolute values of depletion at week 24 were compared between good responders, moderate responders, and nonresponders, and the association of patient characteristics with categories and numbers of CD4+ T cells before treatment and with CD4+ T cell depletion was assessed by Mann-Whitney test. For patient characteristics represented by continuous variables, Pearson correlation testing was used to analyze correlations with the number of CD4+ T cells before treatment and with CD4+ T cell depletion.
Characteristics of the patients
Among the 64 RA patients who received rituximab during the study period, 52 patients for whom flow cytometry data were available before treatment and from at least one time during the first course of rituximab were included in this study. The characteristics of these 52 patients are shown in Table 1.
Table 1. Baseline characteristics of the 52 patients with rheumatoid arthritis treated with rituximab*
Except where indicated otherwise, values are the median (range).
Sex, no. (%) female
Disease duration, years
Previous treatment, no (%)
Anti–tumor necrosis factor α
Disease Activity Score in 28 joints
Swollen joint count
Tender joint count
Erythrocyte sedimentation rate, mm/hour
C-reactive protein, mg/liter
Rheumatoid factor positive, no. (%)
Anti–citrullinated protein antibody positive, no. (%)
Radiologic evidence of erosion, no. (%)
CD4+ CD8+ cells/μl
Depletion of T lymphocytes during rituximab treatment
As expected, total B cell depletion occurred at week 2 and persisted for 6 months, followed by a partial B cell recovery observed between 9 and 12 months (Figure 1). In accordance with this, the percentages of T cells and natural killer cells were slightly increased from week 2 to week 24, whereas those of CD4+ and CD8+ cells remained stable (data not shown). Importantly, although the CD3+, CD4+, and CD8+ absolute cell counts were unchanged at week 2 as compared with baseline, they were substantially decreased later on. At week 12, the average percentages of depletion were 35%, 37%, and 24% for CD3+, CD4+, and CD8+ cells, respectively. One-half of the patients showed a decrease in the percentage of CD4+ T cells ranging from 21% to 62%, and for the lowest quartile, the depletion ranged from 62% to 77% (Figure 1). In contrast, some patients showed a substantial increase (up to 57%) in CD4+ T cells. The results observed at week 24 were similar, and the decrease was less marked although still significant between weeks 36 and 48, with a trend toward a recovery (Figure 1C). Thus, the T cell depletion, which mainly affected CD4+ cells, was delayed as compared to the very early B cell depletion, but the patterns of change in T and B cell counts were similar in that both lasted up to 6 months, with partial recovery later on.
When the results were expressed as absolute counts, several patients showed depletion of >2,000 CD4+ cells/μl (data not shown). Of note, the CD4+ T cell count was <200 cells/μl in 3 of our patients (5.8%), including one who presented with an extensive oropharyngeal candidiasis. In some patients who received successive courses of rituximab, the depletion and reconstitution of the T cell population occurred after each course with the same time-related pattern, indicating that T cell changes, although delayed as compared to B cell changes, were unambiguously related to rituximab treatment. An example is presented in Figure 2.
Finally, we analyzed the change in the number of naive and memory CD4+ T cells by examining CD45RA and CD45RO markers in 10 RA patients with CD4+ T cell depletion. The proportion of CD4+CD45RA+ and CD4+CD45RO+ cells was unchanged during treatment (data not shown), indicating that rituximab-induced CD4+ T cell depletion affected naive and memory cells similarly.
Association between lack of T lymphocyte depletion and lack of clinical response to rituximab
To examine the association of change in T cell count with the response to rituximab, we used data on 41 patients for whom flow cytometry data and information on clinical response at week 24 were available. We compared changes in CD3+, CD4+, and CD8+ cell counts in patients divided into 3 groups according to EULAR response (i.e., nonresponders [n = 11], moderate responders [n = 15], and good responders [n = 15]) (Figure 3). The decreases in CD4+ and CD3+ cell counts were significantly greater for moderate responders (on average, 43% and 37%, respectively) and good responders (on average, 47% and 38%, respectively) than for nonresponders (on average, 7% and 7%, respectively). The groups did not differ significantly with regard to CD8+ T cell changes although there was a trend toward a decrease in CD8+ cell counts in good responders compared to nonresponders. Thus, a lack of CD4+ T cell depletion was associated with a lack of response to rituximab at week 24.
Moreover, we analyzed the association of CD4+ T cell depletion at week 24 with time between the first and second course of rituximab in 24 patients with available data (Figure 4). There was a trend toward greater CD4+ T cell depletion relative to baseline among patients who received a second course of rituximab after 12 months compared to those who received a second course between 6 and 12 months (46% versus 15%; P = 0.06).
To our knowledge, this is the first study to show that rituximab induces substantial T cell depletion, mainly affecting CD4+ cells, in most RA patients, for a final count of <200 CD4+ cells/μl in some patients. Moreover, a lack of CD4+ T cell depletion was associated with no clinical response.
We found that rituximab has a long-lasting and reversible depletion effect on CD4+ cell counts (both naive and memory cells), and, to a lesser extent, on CD8+ cell counts. To our knowledge, only 2 previous studies have investigated the peripheral T cell compartment in rituximab-treated RA patients ([23, 24]). The aim of the first one was to investigate the impact of rituximab treatment on Treg cells. The absolute numbers of CD3+, CD4+, and CD8+ cells remained unchanged during B cell depletion and the regeneration phase compared to baseline (). However, that study included only 17 patients, and 6 of them received a lower rituximab dose than that received by the patients in our study. Moreover, the numbers of CD3+ cells observed in patients before treatment in the previous study were unusually low (mean 867/μl), a finding that has not been reported previously. As stated by the authors, this may be due to extensive treatment preceding rituximab in their patients. In addition, T cell counts were obtained using a dual-platform method after Ficoll-Paque separation, which may result in different T cell counts than the single-platform method used in our study.
In the second study the main objective was to investigate synovial tissue in 24 patients with RA treated with rituximab in order to identify predictors of clinical response. The total numbers of T cells and T cell subsets in peripheral blood were not significantly decreased at week 16, in contrast to the number of T cells in synovial tissue (). It is of note that patients included in that study were selected on the basis of very active disease (mean DAS28 6.5, versus 5.4 in the present study) with a joint amenable for arthroscopy. In addition, all patients had erosions, were receiving methotrexate, and were RF and/or ACPA positive. Thus, these differences in patient characteristics and the limited number of patients included in the previous study may explain the discrepancy with the results of the present study.
Besides the major quantitative effect of rituximab on T cells that we observed in RA patients, a few studies have demonstrated qualitative modifications in patients with autoimmune diseases. For instance, the expression of T cell activation markers, such as CD40L, CD69, or HLA–DR, is decreased along with B cell depletion in lupus patients treated with rituximab ([8, 25, 26]). Different mechanisms have been postulated to explain this phenomenon. First, rituximab targets some CD20-expressing T cells. However, Leandro et al highlighted that only 3.2% of peripheral CD3+ T cells expressed low levels of CD20 (), so a direct effect of rituximab on T cells is unlikely. An indirect effect of rituximab on T cells seems more credible.
Rituximab is thought to act by reducing autoantibody production by decreasing autoreactive B cell counts. However, its efficacy is not restricted to RF- or ACPA-positive patients, which implies that there are other mechanisms of action. Rituximab may inhibit an activation pathway of CD4+ cells initiated by the APC function of B cells or by their ability to stimulate CD4+ cell proliferation after priming by dendritic cells (). In addition, B cells are known to produce large amounts of cytokines and chemokines. Rituximab-induced B cell depletion may therefore affect the equilibrium of the cytokines and chemokines involved in cell migration and retention (). Although this remains to be demonstrated, it could explain why both naive and memory CD4+ cells (as well as CD8+ cells) were affected by rituximab treatment in our patients. Whatever the mechanism(s), this scenario may explain the decrease in T cells that occurs after B cell depletion in some patients and the return to baseline values with the diminishing effect of treatment.
An interesting finding is the relationship between the CD4+ T cell depletion and clinical outcomes, whereas B cell depletion occurred in all patients whatever the response. Indeed, the absence of CD4+ T cell depletion was associated with nonresponse to treatment at week 24. In accordance with this, the extent of CD4+ depletion at week 24 seems to be related to the time between the first and second rituximab course (i.e., disease flare). This finding could help guide clinicians' decisions regarding whether to discontinue rituximab in patients with a lack of CD4+ T cell depletion and poor response after a first course, or when to administer another course in others.
To date, lymphocyte phenotyping in RA patients receiving rituximab has focused on identifying patients at high risk of infection. We found an 80% decrease in CD4+ cells, leading to a count of <200 cells/μl, in some patients, a threshold below which opportunistic infections may occur (). Several cases of opportunistic infections, usually observed in patients deficient in CD4+ T cells, have been reported in RA patients receiving rituximab. Given our results, clinicians should pay particular attention to CD4+ cell counts before and after a course of anti-CD20 monoclonal antibodies because rituximab may favor the occurrence of opportunistic infection, particularly in patients with low CD4+ cell counts ([12, 14]). Consideration of the primary benefit of antibiotic prophylaxis for these patients seems essential.
In conclusion, we found that rituximab may significantly decrease the numbers of peripheral blood T cells, particularly CD4+ cells, in most patients with RA. Interestingly, lack of CD4+ cell depletion was associated with no response to treatment at week 24. Our results support the usefulness of T cell monitoring in RA patients receiving rituximab. CD4+ cell counts may help clinicians in decision making. It is less likely that a clinical response will be achieved in patients who do not experience a decrease in CD4+ cell count than in those who do show a decreased CD4+ cell count. Patients with a low CD4+ cell count at baseline should be monitored to prevent opportunistic infections.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Mulleman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Mulleman, Goupille, Thibault.
Acquisition of data. Mélet, Mulleman.
Analysis and interpretation of data. Mélet, Mulleman, Goupille, Ribourtout, Watier, Thibault.
We thank Drs. Saloua Mammou, Isabelle Griffoul, Emilie Ducourau, and Virginie Martaillé for helping with clinical assessment. We are indebted to Nelly Jaccaz-Vallée, Sergine Gosset, Valérie Angebeau, Laetitia Cornec, Adeline Coutellier, Corinne Depont, Vanessa Fougeray, Valérie Fuseau, Pascale Guibout, Sophie Joncheray, Céline Letot, Isabelle Romier, and Elodie Vigneron for blood sampling and their commitment to taking care of patients and to Claude Gautier and Elisabeth Billant who performed blood sample staining and flow cytometry analysis. We thank Dr. Michael Hahne for critical reading of the manuscript and helpful discussions.