An open study of B lymphocyte depletion in systemic lupus erythematosus




To gain preliminary evidence for the safety and efficacy of B lymphocyte depletion therapy in refractory systemic lupus erythematosus (SLE).


Six female patients with active SLE, resistant to standard immunosuppressive therapy, were treated on an open-label basis. During a 2-week period, each patient received two 500-mg infusions of rituximab, two 750-mg infusions of cyclophosphamide, and high-dose oral corticosteroids.


No significant adverse events were observed during followup. Patient 1 had not improved at 3 months but was then lost to followup. At 6 months, all 5 remaining patients had improved, as evidenced by improvement in British Isles Lupus Assessment Group global scores, from a median of 14 (range 9–27) at baseline to a median of 6 (range 3–8) at 6 months. Manifestations of SLE such as fatigue, arthralgia/arthritis, and serositis responded particularly well to this protocol. Hemoglobulin levels increased in patients 2, 3, 5, and 6. The erythrocyte sedimentation rate decreased in patients 2, 3, 4, and 5 and was stable in patient 1. In patients 4 and 5, the urinary protein–to-creatinine ratio decreased significantly. C3 serum levels increased in all 5 patients who had low levels at baseline; in two of these patients, patients 2 and 5, C3 values were normal at 6 months. The variation in the level of anti–double-stranded DNA antibody was different in individual patients.


This study provides sufficient evidence for the safety and possible efficacy of B lymphocyte depletion therapy in SLE to justify a formal controlled trial.

Systemic lupus erythematosus (SLE) is an autoimmune rheumatic disorder that is thought to involve disturbances in both innate and adaptive immune mechanisms, including complex interactions between T lymphocytes, B lymphocytes, and other antigen-presenting cells. Tissue damage and dysfunction are thought to be mediated chiefly by autoantibodies, via both immune complex formation and direct effects on cells (1). Although disease mortality has improved, optimal therapy has yet to be determined. Rituximab is a chimeric anti-CD20 monoclonal antibody that has proved to be very effective in depleting B lymphocytes in vivo (2). CD20 is a B lymphocyte–restricted antigen that is expressed on B cells from the pre-B stage to the mature B stage but not on hematopoietic stem cells and normal plasma cells (3).

B lymphocyte depletion using rituximab has been introduced for the treatment of several autoimmune disorders, including rheumatoid arthritis (RA), IgM-associated neuropathies, and immune thrombocytopenia, with encouraging early results (4–6). The rationale for B lymphocyte depletion has been the lowering of autoantibody levels through reducing the generation of autoantibody-producing daughter plasma cells, although the precise mechanisms involved remain debatable (7). There is evidence for selective reduction in levels of certain autoantibodies correlating with clinical improvement (4, 5, 8). Given the evidence for a probable direct pathogenic role for at least some autoantibodies in SLE, notably anti–double-stranded DNA (anti-dsDNA) (9), it was considered likely that B lymphocyte depletion might be of similar benefit in this condition.

At the time this study was initiated, available evidence suggested that cyclophosphamide contributed significantly to the therapeutic benefit of B lymphocyte depletion, possibly helping to achieve a higher level of B lymphocyte depletion. Based on our experience with dose–response in RA, the first subjects with SLE were treated with two 500-mg doses of rituximab in combination with two 750-mg doses of cyclophosphamide and high-dose oral prednisolone. We now report 6-month followup data for the first 6 subjects treated in the current study.



Six patients entered the study. All patients fulfilled at least 4 of the revised American College of Rheumatology criteria for the diagnosis of SLE (10). All patients had active disease, having failed conventional immunosuppressive treatment, including intravenous cyclophosphamide (7–13 pulses) in patients 1, 4, 5, and 6. Characteristics of the 6 patients are summarized in Table 1. The study was approved by the hospital ethics committee, and all subjects gave written informed consent.

Table 1. Characteristics at study entry of 6 female patients with systemic lupus erythematosus*
PatientAge, yearsDisease duration, yearsDisease manifestations at baselinePrevious therapiesTherapy before study entry
  • *

    Anti-dsDNA = anti–double-stranded DNA; ESR = erythrocyte sedimentation rate; nephritis (IV) = World Health Organization class IV glomerulonephritis.

1387Fever, fatigue, mouth ulcers, sicca syndrome, organic brain syndrome, arthritis, Raynaud's, lymphopenia, ↓ C3, ↑ anti-dsDNAPrednisolone, azathioprine, cyclophosphamidePrednisolone, 10 mg/day; hydroxychloroquine, 400 mg/day
23511Fatigue, anorexia, mild alopecia, mouth ulcers, migraine headaches, arthritis, serositis, esophageal dysmotility, Raynaud's, skin vasculitis, anemia, lymphopenia, ↑ ESR, ↓ C3, ↑ IgG, ↑ anti-dsDNAPrednisolone, cyclosporinePrednisolone, 10 mg/day; hydroxychloroquine, 400 mg/day; methotrexate, 5 mg/day
33610Fatigue, lymphadenopathy, sclerodactyly, arthritis, serositis, Raynaud's, anemia, lymphopenia, ↑ anti-dsDNAPrednisolone, azathioprine, cyclosporinePrednisolone, 10 mg/day; hydroxychloroquine, 400 mg/day; methotrexate, 15 mg/week
44011Fatigue, skin rash, arthritis, serositis, shrinking lung syndrome, Raynaud's, skin vasculitis, nephritis (IV), anemia, lymphopenia, ↑ ESR, ↓ C3, ↑ anti-dsDNAPrednisolone, sulfasalazine, azathioprine, cyclosporine, cyclophosphamidePrednisolone, 9 mg/day; hydroxychloroquine, 400 mg/day; mycophenolate, 1.5 mg/day
5174Fatigue, anorexia, skin rash, migraine headaches, arthritis, serositis, Raynaud's, skin vasculitis, nephritis (IV), anemia, lymphopenia, ↑ ESR, ↓ C3, ↑ IgG, ↑ anti-dsDNAPrednisolone, azathioprine, cyclophosphamide, mycophenolatePrednisolone, 15 mg/day; hydroxychloroquine, 400 mg/day; azathioprine, 175 mg/day
62012Fatigue, skin rash, arthritis, Raynaud's, nephritis (IV), anemia, lymphopenia, ↓ C3, ↑ anti-dsDNAPrednisolone, azathioprine, cyclophosphamidePrednisolone, 8 mg/day; hydroxychloroquine, 400 mg/day; cyclophosphamide, 750 mg every 3 months

Treatment protocol.

Patients were treated with two 500-mg infusions of rituximab (600–700 mg/m2 of body surface area in total) and two 750-mg infusions of intravenous cyclophosphamide under oral prednisolone cover (30 mg or 60 mg for 5 days, starting the day before rituximab infusion) given 2 weeks apart. To further minimize infusion reactions, subjects were premedicated with chlorpheniramine (10 mg intravenously) and paracetamol (1 gm orally). Other immunosuppressive agents were stopped before day 0 (see Table 1). Patients were allowed to continue taking hydroxychloroquine and oral prednisolone.


Clinical and laboratory assessments were performed at the time of recruitment, before treatment, and monthly for the first 6 months following treatment. At each visit, patients were evaluated for clinical manifestations of SLE and for any side effects of therapy. Laboratory measurements included full blood count, erythrocyte sedimentation rate (ESR), renal and liver serum function tests, ratio of urinary protein to creatinine excretion as measured in a random specimen of urine (normal <13), serum C3 by laser nephelometry (normal 0.90–1.80 gm/liter) serum immunoglobulin levels by immunoturbidometry (IgA normal 0.7–4.0 gm/liter, IgG normal 7.0–16.0 gm/liter, and IgM normal 0.4–2.3 gm/liter), antinuclear antibodies by immunofluorescence, and IgG anti-dsDNA antibodies by enzyme-linked immunosorbent assay (Shield Diagnostics, Dundee, UK) (normal <50 IU/ml). Circulating B lymphocytes (CD19+ count by flow cytometry; normal 0.03–0.40 × 109/liter) and total lymphocytes were measured before treatment, monthly after treatment for the first 6 months, and then every 2 months until normal levels were reached. The British Isles Lupus Assessment Group (BILAG) index (11) was used for both individual organ system assessment and also as a global index, whereby an A score in any system = 9 points, B = 3, C = 1, and D/E = 0. Treatment efficacy was evaluated on the basis of improvement in both clinical and laboratory indices of active disease.


Clinical outcome.

Patient 1 did not experience any significant clinical change. She was lost to followup at 3 months. All other patients improved significantly. Baseline BILAG global scores were 13, 15, 9, 13, 27, and 18, respectively (median 14, range 9–27). BILAG global scores at 3 months were 13, 12, 5, 6, 10, and 4, respectively (median 8, range 4–13). BILAG global scores at 6 months for patients 2, 3, 4, 5, and 6 were 5, 6, 8, 6, and 3, respectively (median 6, range 3–8) (Figure 1). Manifestations such as fatigue, arthralgia/arthritis, serositis, and skin vasculitis responded particularly well to this protocol. In all patients with nephritis, creatinine serum levels remained stable and within the normal range. In patient 5, serum albumin levels normalized at 3 months. In patients 4 and 5, renal disease responded well, with significant decreases in the ratio of urinary protein to creatinine, from 250 at baseline to 78 and 16, respectively, at 6 months. In patient 6, the ratio of urinary protein to creatinine was stable at 23 at 6 months. In patient 4, pulmonary function tests improved. Oral daily prednisolone doses at baseline for patients 1, 2, 3, 4, 5, and 6 were 10 mg, 10 mg, 10 mg, 9 mg, 15 mg, and 8 mg, respectively (mean 10.3 mg, range 8–15 mg). At 6 months, prednisolone doses in patients 2, 3, 4, 5, and 6 had been slightly decreased to 9 mg, 9 mg, 8 mg, 10 mg, and 7 mg, respectively (mean 8.6 mg, range 7–10 mg).

Figure 1.

British Isles Lupus Assessment Group (BILAG) global scores at baseline and during followup.

Other laboratory indices.

All patients were lymphopenic at baseline and remained so throughout the study. Total lymphocyte counts fluctuated with no particular pattern. The period of B lymphocyte depletion, defined as a CD19+ count <0.005 × 109/liter in peripheral blood, for patients 1, 2, 4, 5, and 6 was 3, 5, 4, 4, and 6 months, respectively (mean 4.4 months, range 3–6 months). Patient 3 is still B lymphocyte–depleted 16 months after being treated. Hemoglobulin levels increased in patients 2, 3, 5, and 6. The ESR decreased in patients 2, 3, 4, and 5 and was stable in patient 1. In patients 1, 2, 4, 5, and 6, C3 serum levels increased, normalizing in patients 1, 2, and 5 (Figure 2). Patient 3 had normal C3 at baseline, having received methotrexate until the week before entry. At 2 weeks, her C3 level had lowered to 0.76 but increased afterward, reaching a normal value at 6 months (Figure 2). The level of serum anti-dsDNA varied in different patients (Figure 2). In patient 1, anti-dsDNA increased at 1 month and then decreased. In patients 2 and 3, anti-dsDNA increased to a maximum at 3 months and then decreased. In patients 4 and 5, anti-dsDNA levels decreased steadily until 5 months but were slightly higher at 6 months. Both patients relapsed soon after (patient 4 at 7 months, patient 5 at 8 months).

Figure 2.

Serum levels of C3 (normal 0.90–1.80 gm/liter) and anti–double-stranded DNA (anti-dsDNA) (normal <50 IU/ml) at baseline and during followup.

Serum total IgG and IgM decreased moderately. IgA dropped by a mean of 0.2 gm/liter (range 0–0.6), IgG dropped by a mean of 3.9 gm/liter (range 0–6.3), and IgM dropped by a mean of 0.5 gm/liter (range 0–1.6). Patient 6 had low IgM at baseline, which did not decrease further with treatment. In all other patients, immunoglobulin serum levels remained within the normal range.

Further followup.

Followup for patients 2, 3, 4, 5, and 6 is currently 18, 16, 14, 9, and 6 months, respectively (mean 12.6 months, range 6–18 months). Patients 2 and 3, despite fluctuating evidence of disease activity, are still well enough that treatment with other immunosuppressive drugs has not been restarted. Their BILAG global scores at the last followup visit were 5 and 3, respectively. For patient 2, the dose of oral prednisolone was very slowly reduced, from a baseline dose of 10 mg to 5 mg. At 16 months, patient 3 was still taking 9 mg of prednisolone. Patients 4 and 5 relapsed at 7 and 8 months, respectively, and have been retreated, although with different protocols.

Adverse events.

Treatment was safe and well tolerated. No significant infusion-related adverse effects of rituximab were observed. Subjects 1, 2, 3, and 4 had nausea and/or vomiting after receiving cyclophosphamide. Other adverse events during the study were as follows: acute respiratory infection (3 episodes), acute gastroenteritis (3 episodes), shingles (1 episode), folliculitis (1 episode), oral candidiasis (1 episode), and tinea pedis (1 episode). The infectious episodes reported were either self-limited or responded well to conventional antibiotic treatment.


Use of rituximab in combination with cyclophosphamide and corticosteroids in this cohort of subjects with severe SLE provides evidence for a good level of safety and sufficient evidence for possible efficacy to merit proceeding to a controlled trial. All patients improved except patient 1, whose CNS manifestations, in retrospect, were probably due to damage rather than active disease. The safety profile is consistent with a preliminary report of 12 patients treated with lower-dose rituximab alone (12) and isolated case reports of patients with complex, life-threatening lupus manifestations (13). As in patients with lymphoma and RA, B lymphocyte depletion does not appear to be associated with an increased risk of major, and in particular, opportunistic, infection. This probably relates to the generally well-preserved total immunoglobulin levels.

In this cohort, as seen in RA and certain other conditions, clinical improvement was not limited to the period of B lymphocyte depletion. Subjects 2 and 3 remain well enough that immunosuppressive therapy was not reintroduced at 18 and 16 months of followup, respectively. It seems that a proportion of subjects with autoimmune disease relapse immediately on repopulation of circulating B lymphocytes, but a proportion remain well for many months (up to 17 in RA). These extended periods of improvement lend support to the idea that disease perpetuation depends on a restricted population of B lymphocyte clones with a particular ability to engage a vicious cycle of autoantibody production, and that relapse may require the reemergence of such clones through random immunoglobulin gene mutation (7).

The protocol used was based on our experience with dose–response in RA at the time (14). In other series, subjects with autoimmune diseases were usually treated with full-dose rituximab (375 mg/m2 of body surface area × 4) monotherapy. The optimal total dose, treatment schedule, and use of adjuvants for rituximab in different autoimmune diseases remain to be determined. However, in parallel with experience in lymphoma, 2 options appear to be emerging.

Rituximab monotherapy appears to be safe, with the rare exception of infusion reactions. In patients with indolent oncologic disorders and in cases of autoimmunity, periodic rituximab monotherapy might, therefore, be considered to have an important advantage. However, it often may have to be used every 6–9 months for an indefinite period of time, and there is uncertainty about long-term responsiveness as well as potential immunodeficiency. Low-dose rituximab monotherapy in SLE has been associated with frequent human antichimeric antibody responses. This may not occur with full dosage, but it raises the possibility that resistance may eventually occur.

Combination therapy is the usual approach in B cell neoplasia (15), for which long-term cure is the key objective. In RA, there is preliminary evidence for extended benefit with combination therapy, and long-term cure remains a theoretical goal in autoimmune disease. However, even with full-dose rituximab and cyclophosphamide, most patients relapse, and the impression is that, as in lymphoma, if the aim is to achieve prolonged remission, the regimens in use are still not potent enough. Because increasing the dose of rituximab is costly, efforts are being directed chiefly at finding safer, more potent adjuvant agents.

For the time being, this line of argument has persuaded us to return to use of full-dose rituximab for further treatments on the grounds of urgent clinical need. Nothing clearly superior to cyclophosphamide is available; therefore, we have continued to use it, despite the potential risks. The priority now is to set up formal controlled trials to assess the true benefit of B lymphocyte depletion in SLE. The most convenient options are probably to compare rituximab monotherapy with placebo in patients with nonrenal disease, and to compare a combination of rituximab and the National Institutes of Health (NIH) cyclophosphamide regimen against the NIH regimen alone in patients with renal disease.