Rituximab Therapy for Primary Sjögren's Syndrome: An Open-Label Clinical Trial and Mechanistic Analysis


  • ClinicalTrials.gov identifier: NCT0012101829.



To study the safety and clinical efficacy of rituximab therapy for primary Sjögren's syndrome, as well as to investigate its mechanisms.


Patients with primary Sjögren's syndrome were enrolled in an open-label trial, were given rituximab (1 gm) infusions on days 1 and 15, and were monitored through week 52. The primary end point was safety, with secondary end points evaluating clinical and biologic efficacy. Blood was obtained for enumeration of lymphocyte subsets, measurement of serum autoantibody and BAFF levels, and analysis of gene expression.


Twelve female patients with primary Sjögren's syndrome were administered rituximab. They had a median age of 51 years (range 34–69 years) and a median disease duration of 8.0 years (range 2–18 years). We observed no unexpected toxicities from the rituximab therapy. Modest improvements were observed at week 26 in patient-reported symptoms of fatigue and oral dryness, with no significant improvement in the objective measures of lacrimal and salivary gland function. The recovery of blood B cells following the nadir from rituximab therapy was characterized by a predominance of transitional B cells and a lack of memory B cells. While blood B cell depletion was associated with an increase in serum BAFF levels, no significant changes were observed in the levels of serum anti-Ro/SSA, anti-La/SSB, and anti–type 3 muscarinic acetylcholine receptor autoantibodies or in the blood interferon signature.


In patients with primary Sjögren's syndrome, a single treatment course of rituximab was not associated with any unexpected toxicities and led to only modest clinical benefits despite effective depletion of blood B cells.

Primary Sjögren's syndrome is among the most common of the connective tissue diseases. For women, its prevalence in the UK has been estimated to be 0.1–0.6% (1). The disease is characterized by the presence of keratoconjunctivitis sicca (dry eyes), xerostomia (dry mouth), serum antinuclear antibodies, and chronic salivary gland inflammation, as well as the occurrence of systemic features, such as profound fatigue, polyarthralgia/polyarthritis, interstitial lung disease, peripheral neuropathy, and leukocytoclastic vasculitis (2, 3). Patients with primary Sjögren's syndrome are also at increased risk of developing B cell lymphoma (4).

The treatment of primary Sjögren's syndrome is largely based on alleviation of symptoms and includes the use of topical cyclosporine (for the management of dry eyes), sialogogues (oral muscarinic agonists), hydroxychloroquine, and low doses of prednisone (5). Patients with more serious systemic manifestations may require more intensive therapy with glucocorticoids and other immunosuppressive agents. However, no drugs have been shown in well-designed clinical trials of patients with primary Sjögren's syndrome to reduce disease activity or prevent damage.

The potential clinical utility of rituximab therapy has recently been investigated in primary Sjögren's syndrome (6–10), owing to its proven efficacy in other chronic inflammatory diseases, such as rheumatoid arthritis (11, 12) and systemic vasculitis (13), and its effects on potential disease-inciting B cells. The importance of abnormal B cell responses in the mechanisms of primary Sjögren's syndrome is strongly suggested by the presence of serum autoantibodies, most notably, anti-Ro/SSA and anti-La/SSB antibodies (3). The benign and malignant B cell monoclonal proliferations in the blood and salivary gland tissues of patients with primary Sjögren's syndrome (14), as well as the abnormalities in B cell memory (15), provide further evidence that B cells play an important role in the pathophysiology of this condition. We therefore conducted an open-label study of rituximab, a potent B cell depleter, to evaluate the safety and possible clinical efficacy of this approach in primary Sjögren's syndrome, as well as to determine its effects on blood B cell subsets, autoantibodies, cytokines, and gene transcripts.


Study design and treatment.

The study was a prospective, open-label, single-arm, phase I study of rituximab therapy for patients with primary Sjögren's syndrome (ClinicalTrials.gov identifier NCT0012101829). Twelve patients received two 1,000-mg infusions of rituximab 2 weeks apart, using a standard protocol, with escalation of the infusion rate to a maximum of 400 mg/hour. All patients were pretreated with 50 mg of oral diphenhydramine, 650 mg of oral acetaminophen, and 100 mg of intravenous methylprednisolone ∼30 minutes before each of the infusions. The patients returned for followup visits at weeks 4, 8, 14, 26, 30, 36, and 52. Diphtheria and tetanus toxoid as well as a pneumococcal polyvalent-23 vaccine were administered at week 26 to 8 of the patients for assessment of immune competence.

The study was approved by the institutional review boards (IRBs) at Duke University Medical Center and the University of Pennsylvania. All patients provided informed consent.

Patient population.

Eligible patients were adults between the ages of 18 and 75 years who met the classification criteria for primary Sjögren's syndrome as defined by the American–European Consensus Group (16). In addition, patients had one or more of the following systemic manifestations: fatigue (>50 mm on a 100-mm visual analog scale [VAS]), joint pain (>50 mm on a 100-mm VAS), severe parotid gland swelling, peripheral neuropathy, interstitial lung disease, leukocytoclastic vasculitis, interstitial nephritis, or other extraglandular disease causing organ-system dysfunction. Patients of reproductive potential agreed to use an acceptable method of birth control during treatment and for 12 months following treatment.

Concomitant therapy with nonsteroidal antiinflammatory drugs, cevimeline, pilocarpine, ophthalmic cyclosporine, and hydroxychloroquine was allowed, provided the dosages were maintained at the baseline levels. Concurrent therapy with prednisone (≤10 mg/day) was permitted if the dosage had been stable for at least 2 weeks prior to study entry and was kept constant during the study.

Patients were excluded from the study if they had previously been treated with rituximab or if they had been recently treated with the following medications: cyclophosphamide within 24 weeks, methotrexate, azathioprine, cyclosporine, or mycophenolate mofetil within 4 weeks, etanercept within 4 weeks, adalimumab within 8 weeks, or infliximab within 12 weeks. Those taking potent anticholinergic agents, such as tricyclic antidepressants, antihistamines, phenothiazine, and antiparkinsonian drugs were not allowed to participate in the study. Patients were also excluded if they had active infection, chronic or persistent infection that might be worsened by immunosuppressive therapy (e.g., human immunodeficiency virus, hepatitis B or C, tuberculosis), known coronary artery disease or a history of significant cardiac arrhythmias or severe congestive heart failure, pregnancy, ongoing oral anticoagulant therapy, a history of alcohol or substance abuse, prior head and neck radiation therapy, history of sarcoidosis, history of a positive result on purified protein derivative test without documentation of treatment for active or latent tuberculosis, history of severe pulmonary disease (forced vital capacity <50% predicted, diffusing capacity for carbon monoxide <50% predicted, resting oxygen saturation <95%), history of malignancy, except for resected basal cell or squamous cell carcinoma of the skin, cervical dysplasia, or in situ cervical cancer grade 1 within the last 5 years, abnormal laboratory results (absolute neutrophil count <1,000/mm3, platelets <100,000/mm3, hemoglobin <9 gm/dl, serum creatinine ≥2.0 mg/dl, aspartate aminotransferase or alanine aminotransferase >2 times the upper limit of normal), or administration of a live vaccine within the previous 3 months.

Study end points.

The primary safety end point was the proportion of patients experiencing a grade 3, grade 4, or grade 5 adverse event (AE) according to the Common Terminology Criteria for Adverse Events of the National Cancer Institute that was judged by the investigator to be possibly, probably, or definitely related to rituximab therapy. The other goal of the study was to obtain preliminary evidence of clinical and biologic activity using measures of exocrine gland function and other disease features, as well as immune system function. For the clinical and biologic end points, we focused our analysis on the changes between weeks 0 and 26.

Clinical assessments.

Safety was evaluated at each visit by monitoring for AEs, including changes in the values of routine laboratory parameters. Assessments of clinical efficacy were performed at weeks 8, 14, 26, 36, and 52 and included a Sjögren's Syndrome Symptom Survey (17), physician's and patient's global assessments of disease activity (100-mm VAS), unanesthetized Schirmer's test, slit-lamp examination with lissamine green staining, and measurement of unstimulated and stimulated whole salivary flow rate. For the assessment of salivary flow, patients were instructed to withhold the morning dose of secretagogue and take nothing by mouth for at least 60 minutes prior to measurement of salivary flow. Unstimulated salivary flow was determined by having the patient expectorate into a preweighed 50-cm3 centrifuge tube for 15 minutes. Saliva samples were weighed on an analytical balance to quantify the volume over the 15-minute collection period (1 gm = 1 ml). Subsequently, patients were treated with 5 mg of oral pilocarpine to stimulate salivary flow, and 60 minutes later, another 15-minute saliva sample was collected. In addition, quality of life was examined at weeks 0 and 26 using the Short Form 36 (SF-36) health survey (18).

Flow cytometry.

Venous blood that was less than 24 hours old was processed for flow cytometry as described previously (19, 20). After washing and Fc blockade, cells were stained with the following fluorochrome-conjugated antibodies in multiple 4-color combinations (all antibodies were purchased from BD Biosciences): fluorescein isothiocyanate–conjugated CD27 (M-7271) and CD3 (SK7); phycoerythrin-conjugated CD38 (HIT2), CD8 (HIT8a), CD16 (3G8), and CD56 (B159); PerCP–Cy5.5–conjugated CD20 (L27), CD4 (SK3), and CD8 (SK1); and allophycocyanin-conjugated CD19 (HIB19), CD45 (H130), and CD14 (M5E2). After staining and lysis of red blood cells, the white blood cell pellets were washed and fixed in paraformaldehyde. Data from stained cells were acquired with a FACSCalibur instrument and analyzed using CellQuest Pro software version 5.2 (both from BD Biosciences). Approximately 10,000 CD19+ lymphocytes were analyzed per tube or, in samples with low numbers of B cells, the maximum number of events was acquired by draining the tube.

For comparison with the patients in the clinical trial, blood was obtained, with IRB approval, from 15 subjects (11 women and 4 men) at the University of Pennsylvania. These subjects served as controls. Their median age was 39.5 years (interquartile range 27–49 years). Control subjects were not part of the current trial and were younger than the patients enrolled in the study herein.

B cell subset counts.

Fresh whole blood (less than 24 hours old) was analyzed by flow cytometry as described above. To calculate the absolute counts of B cell subsets, the absolute lymphocyte count (in cells/μl of whole blood) was obtained by Coulter analysis or through the complete blood cell count at each of the local sites. The B cell fraction was obtained by multiplying the absolute lymphocyte count by the CD19+ fraction. B cell subset counts were computed by multiplying the B cell absolute count by the fraction of B cells in each of the subsets. CD38 and CD27 staining was used to define B cell subsets, which were then classified as follows: transitional (CD38++CD27−), mature naive (CD38+ CD27−), mature activated memory (CD38+CD27+), resting memory (CD38−CD27+), plasmablast (CD38++CD27++), and double negative (CD38−CD27−), as described previously (20, 21).

Measurement of autoantibodies.

Serum IgM rheumatoid factor was determined by nephelometry. Serum antibodies to Ro/SS-A and La/SS-B were measured by enzyme-linked immunosorbent assay (ELISA) in the Clinical Immunology Laboratory at the Hospital of the University of Pennsylvania. For assay of anti–type 3 muscarinic acetylcholine receptor (anti-M3R) antibodies, patient sera were incubated with Flp-In Chinese hamster ovary cells that had been transfected with a human M3R construct, as previously described (21). Transfected or nontransfected control cells were incubated with 5 μl of serum for 2 hours on ice, washed once, and stained with fluorescein isothiocyanate–conjugated goat anti-rabbit antibodies to human IgG or IgA. After washing, cells were analyzed with a FACScan flow cytometer (Becton Dickinson). The results are expressed as the mean fluorescence intensity (MFI).

Determination of serum BAFF levels.

ELISA was performed using anti-human BAFF monoclonal antibody clone B4H7.2 for coating and biotin-labeled anti-human BAFF clone A9C9.1 for detection. Standard curves were constructed with recombinant BAFF (all from Biogen Idec). To control for interassay variations in each freshly prepared standard curve, a 2-point interassay control standard was applied. Serum was diluted 1:10 (or higher if necessary for high BAFF levels) and tested in triplicate.

Gene expression profiling.

Peripheral blood was collected into 2 PAXgene RNA tubes (PreAnalytiX), and total RNA was extracted using PAXgene 96 Blood RNA kits (Qiagen). Excess globin transcripts were removed using Ambion GlobinClear (Life Technologies), following the manufacturer's protocols. RNA concentrations and quality were assessed using an Agilent 2100 Bioanalyzer (quality threshold 28S:18S ribosomal RNA ratios >1.0; RNA concentration >70 ng/μl and no more than 14 μl for optimum cleaning). Biotinylated, amplified RNA was produced from 300 ng of RNA using a modification of the Eberwine protocol (22) as described in the Illumina TotalPrep RNA Amplification kit (Ambion). The complementary RNA was hybridized overnight at 58°C to Human WG-6 Expression BeadChip microarrays (Illumina), washed under high-stringency conditions, labeled with streptavidin–Cy3, and scanned.

Raw intensity values were background subtracted using Illumina BeadStudio software. Probe level analysis was performed using the BeadConductor Lumi package for R, a package specifically written to process Illumina microarray data (www.bioconductor.org). The raw gene expression data were subjected to variance-stabilizing transformation; robust splines were then applied for normalization. The 8 samples with data for each time point (baseline and 8 weeks, 26 weeks, and 52 weeks following rituximab) were included in the analysis. The expression data were not normalized for the numbers of B cells in the blood. Paired t-tests and P values were calculated at each 8-week, 26-week, and 52-week time point compared to baseline. IPA 9.0 software (Ingenuity Systems) was used to map transcripts to canonical pathways.

Statistical analysis.

The sample size of 12 patients was chosen to obtain sufficient information about safety in this disease population, with a secondary objective of obtaining preliminary information about its possible clinical and biologic activity. A sample size of 12 patients is the minimum size required to generate a one-sided 90% confidence interval that excludes a prevalence of 30% for an event if no more than one treatment-related AE of unacceptable severity is observed. The frequencies of different B cell subsets were compared between patients with primary Sjögren's syndrome and healthy controls, using the Mann-Whitney 2-tailed exact test. P values for all other tests were based on Wilcoxon's signed rank test and examined the null hypothesis that the median difference score between the baseline visit and the week 26 visit is equal to 0. P values less than or equal to 0.05 were considered statistically significant. No adjustments were made for multiple comparisons. Calculations were performed using SAS version 9.1 software, or higher (SAS Institute).


Recruitment took place between April 2005 and July 2006, with 6 patients enrolled from Duke University Medical Center and 6 from the University of Pennsylvania. All 12 patients received the full dose of rituximab, completed the study followup through week 52, and were included in the safety and efficacy analyses. Their baseline characteristics are shown in Table 1. Most of the enrolled participants had relatively low baseline rates of unstimulated and stimulated whole salivary flow. Eight patients were taking hydroxychloroquine, while only 3 patients were receiving oral prednisone ≤10 mg/ day. Four of the patients were receiving an oral secretagogue (2 taking pilocarpine and 2 taking cevimeline).

Table 1. Characteristics of the 12 patients with primary Sjögren's syndrome*
  • *

    LSG = labial salivary gland; VAS = visual analog scale (0–100 mm).

Age, median (range) years51 (34–69)
% female100
Race, no. (%) 
 Caucasian11 (91.7)
 Black/African American1 (8.3)
Disease duration, median (range) years8.0 (2–18)
Ocular symptoms, no. (%)12 (100)
Oral symptoms, no. (%)12 (100)
Ocular signs, no. (%)12 (100)
 Schirmer I test score, median (range) mm 
  Left eye11 (0–35)
  Right eye5 (0–19)
 Lissamine green staining score, median (range) (0–18 scale) 
  Left eye9.0 (2–18)
  Right eye9.5 (4–18)
Whole salivary flow rate, median (range) ml/minute 
 Unstimulated0.025 (0.0–0.22)
 Stimulated0.05 (0.0–0.65)
Minor LSG focus score ≥1 per 4 mm2, no./no. biopsied3/4
Fatigue (VAS score >50 mm), no. (%)10 (83.3)
Joint pain (VAS score >50 mm), no. (%)9 (75.0)
Severe parotid gland swelling, no. (%)3 (25.0)
Other extraglandular disease, no. (%) 
 Peripheral neuropathy, no. (%)4 (33.3)
 Interstitial lung disease, no. (%)1 (8.3)
Antibodies, no. (%) 
 Anti-Ro/SSA10 (83.3)
 Anti-La/SSB5 (41.7)
 IgM rheumatoid factor10 (83.3)
Short Form 36 health survey, median (range) 
 Physical function60.0 (20–100)
 Mental function70.0 (60–90)


The rituximab infusions were generally well-tolerated in this study (Table 2). Two patients experienced a serious adverse event (SAE), including 1 patient who had a grade 2 reaction to a pneumococcal vaccine, consisting of local (pain, swelling, and numbness) and systemic (fever, myalgia) symptoms that resulted in an emergency department evaluation and treatment with parenteral and oral antibiotics. Of note, another patient in the study had a grade 2 vaccination reaction, and an additional patient had a grade 2 vaccination reaction associated with fever and chills; neither of these AEs was considered an SAE. None of the other 5 patients who received both a pneumococcal and a diphtheria and tetanus toxoid vaccine had AEs from those immunizations. In all 3 cases, these AEs resolved without apparent sequelae. No subsequent vaccinations were administered following these AEs, owing to the unanticipated severity of these reactions. The other SAE was a squamous cell carcinoma of the skin that occurred 301 days after administration of rituximab that was considered to be possibly related to the study drug. We did not observe any serum sickness–like reactions in this study.

Table 2. Adverse events in the 12 patients with primary Sjögren's syndrome*
  • *

    AEs = adverse events; SAEs = serious adverse events.

  • Patient was diagnosed as having squamous cell carcinoma of the skin 301 days after rituximab therapy (considered possibly related to the study drug).

No. of AEs reported162
No. (%) of patients with AEs12 (100)
No. of SAEs2
No. of SAEs related to rituximab1
No. (%) of patients with SAEs2 (16.7)
No. (%) of AEs, by severity 
 Mild (grade 1)138 (85.2)
 Moderate (grade 2)20 (12.3)
 Severe (grade 3)4 (2.5)
 Life-threatening (grade 4)0
 Fatal (grade 5)0
No. (%) of patients with AEs, by severity 
 Mild (grade 1)5 (41.7)
 Moderate (grade 2)5 (41.7)
 Severe (grade 3)2 (16.7)

Clinical efficacy.

The results showed significant but modest levels of improvement between week 0 and week 26 in the both the physician's (median decrease 26 mm; P = 0.012) and patient's (median decrease 8.5 mm; P = 0.009) global rating of disease activity. Although positive trends toward subjective improvement in dryness were observed in many items of the survey, only the changes in the ratings of tongue dryness (median decrease 11.1 mm; P = 0.007), level of thirst (median decrease 35.5 mm; P = 0.005), level of oral discomfort (median decrease 23.5 mm; P = 0.02), and level of overall fatigue (median decrease 18.4 mm; P = 0.042) reached statistical significance. There was no significant improvement in joint pain (median decrease 4.0 mm; P = 0.077). We also did not find any statistically significant improvement between week 0 and week 26 in the unstimulated (median change 0.01; P = 0.287) or stimulated (median change 0.005; P = 0.718) whole salivary flow. There were also no significant changes in tear production, as measured by the unanesthetized Schirmer's test, or ocular surface dryness, as determined by a modified von Bijsterveld scoring system (0–18 scale). Although no significant improvement was observed between week 0 and week 26 in the summary measures on the SF-36 for physical and mental functioning, a statistically significant increase was found during this period in the scores on the vitality scale (P = 0.006).

Peripheral blood B cell depletion and reconstitution.

The absolute peripheral blood CD19+ lymphocyte (B cell) counts at baseline ranged from 45 cells/μl to 341 cells/μl (median 178). Following rituximab treatment, all 12 patients showed >95% depletion of blood B cells by week 8 or week 14 (Figure 1). The blood B cell counts had a nadir at weeks 8 and 14 and began to rise by week 26; they returned to ≥62% of baseline values by week 52 in 8 of the 12 patients.

Figure 1.

Total numbers of blood CD19+ B cells before and after treatment with rituximab in patients with primary Sjögren's syndrome. Absolute B cell counts were obtained by multiplying the absolute lymphocyte count obtained from the complete blood cell count by the CD19+ fraction obtained by flow cytometry. Plotted are the whole blood CD19+ lymphocyte counts as a function of time. Time zero is the baseline assessment and corresponds to the day of the first rituximab infusion. Each symbol represents a different patient; lines connect the individual data points.

The subsets of circulating B cells were analyzed according to the surface expression of CD38 and CD27, a previously validated approach (20,23). Using these markers, we defined 6 different subsets: transitional B cells (CD38++CD27−), mature naive B cells (CD38+CD27−), mature activated memory B cells (CD38+CD27+), resting memory B cells (CD38−CD27+), plasmablasts (CD38++CD27++), and so-called double- negative B cells (CD38–CD27−). Compared to our healthy controls, the patients with primary Sjögren's syndrome at baseline had a greater number of transitional B cells and a lower number of memory B cells (data available upon request from the corresponding author). Therefore, the distribution of circulating B cell subsets in the 12 patients from our study was similar to that in patients with primary Sjögren's syndrome described previously (10, 15).

After rituximab therapy, the initial wave of repopulating blood B cells comprised mainly transitional B cells (Figure 2 and Table 3). At week 26, the median number of transitional B cells (CD38++CD27−) in the circulation was 10.3 cells/μl (65% of total CD19+ B cells). By comparison, the naive mature B cells at week 26 comprised, on average, only 8.3% (median 0.57 cells/μl) of the total CD19+ B cells. Mature activated and resting memory B cells (CD38+CD27+ and CD38−/CD27+) were relatively rare during this initial phase of reconstitution, comprising ∼4% (median) of the total circulating CD19+ B cells. By week 52, the median number of circulating transitional B cells had increased to 21.3 cells/μl (19.3% of total CD19+ B cells). While at the same time, the median number of mature naive B cells (CD38+CD27−) had increased to 75.9 cells/μl (73.1% of total CD19+ B cells). The median number of circulating mature activated (CD38+CD27+) and resting memory (CD38–CD27+) B cells remained diminished, both in absolute and in relative terms, at week 52, representing only 3.2% and 0.4%, respectively, of the total CD19+ B cells. Thus, 52 weeks following rituximab therapy, the more mature B cell subsets had not yet fully repopulated the circulating pool.

Figure 2.

Analysis of blood B cell subsets following rituximab therapy in patients with primary Sjögren's syndrome. The average percentages of B cells in each of the different B cell subsets (CD19+ lymphocytes) in all 12 patients are plotted as a function of time following rituximab treatment. Time zero is the baseline assessment and corresponds to the day of the first rituximab infusion. B cell subsets were defined according to the expression of CD27 and CD38, as described in Patients and Methods.

Table 3. Analysis of peripheral blood B cell subsets before and after rituximab therapy*
B cell subset, time pointNo. of cells/μl, median (range)
  • *

    Time points are the number of weeks following rituximab administration. B cell subsets were determined as described in Patients and Methods. Data on B cell subsets were not available for patient 08-096 at baseline or for patient 04-098 at week 26.

Transitional, CD38++CD27− 
 Baseline21.24 (1.5–47.7)
 8 weeks0.01 (0.0–0.1)
 14 weeks0.01 (0.0–0.2)
 26 weeks10.30 (0.1–58.3)
 36 weeks13.89 (0.5–46.8)
 52 weeks21.32 (7.2–49.3)
Mature naive, CD38+CD27− 
 Baseline134.9 (18.9–280.0)
 8 weeks0.04 (0.0–0.2)
 14 weeks0.02 (0.0–0.1)
 26 weeks0.57 (0.1–74.8)
 36 weeks37.69 (0.3–96.0)
 52 weeks75.94 (1.4–192.7)
Mature activated memory, CD38+CD27+ 
 Baseline10.71 (5.1–40.1)
 8 weeks0.18 (0.0–0.5)
 14 weeks0.11 (0.0–0.9)
 26 weeks0.60 (0.3–2.0)
 36 weeks1.46 (0.5–7.1)
 52 weeks2.16 (1.2–23.6)
Resting memory, CD38−CD27+ 
 Baseline3.69 (0.0–14.9)
 8 weeks0.02 (0.0–0.1)
 14 weeks0.0 (0.0–1.0)
 26 weeks0.10 (0.0–0.5)
 36 weeks0.25 (0.1–3.5)
 52 weeks0.36 (0.1–8.4)
Plasmablast, CD38++CD27++ 
 Baseline1.01 (0.3–3.9)
 8 weeks0.06 (0.0–2.3)
 14 weeks0.06 (0.0–0.5)
 26 weeks0.24 (0.0–1.1)
 36 weeks0.43 (0.0–5.0)
 52 weeks0.78 (0.2–15.9)
Double negative, CD38−CD27− 
 Baseline4.28 (1.3–7.4)
 8 weeks1.54 (0.0–3.6)
 14 weeks0.76 (0.0–3.5)
 26 weeks0.87 (0.0–6.5)
 36 weeks1.47 (0.6–5.0)
 52 weeks1.28 (0.4–8.7)

Blood T cell, natural killer (NK) cell, and monocyte levels.

Rituximab therapy was not associated with any substantial changes in the numbers or percentages of blood CD3+, CD4+, and CD8+ T cells, CD16+CD56+ NK cells, or CD14+ and CD11b monocytes (data not shown).

Autoantibody levels.

Rituximab therapy had little effect on the serum levels of anti-Ro/SSA and anti-La/SSB antibodies (data not shown). However, there was a trend toward a decrease in the levels of serum rheumatoid factor, but this difference failed to reach statistical significance (median 150.0 units/liter [25th, 75th percentiles 20, 1930] at week 0 versus median 87 units/liter [25th, 75th percentiles 20, 1560]; P = 0.109) at week 26.

Serum antibodies to M3R have been reported to occur in primary Sjögren's syndrome and may be associated with impaired cholinergic transmission (24, 25). Baseline data for serum anti-M3R antibodies were available for 11 of the 12 patients in the study. In general, we observed no significant changes in the serum levels of anti-MR3 antibodies between baseline and week 26, except for a lone patient whose values decreased from an MFI of 35 to an MFI of 17. Of note, this particular patient did not show an increase in either the unstimulated (0.12 ml/minute at baseline and 0.06 ml/minute at week 26) or stimulated (0.33 ml/minute at baseline and 0.36 ml/minute at week 26) whole salivary flow rate over this time period.

Serum BAFF levels.

BAFF is a key survival factor for B cells and is important for the maintenance of peripheral B cell homeostasis (26). Serum BAFF levels are elevated in primary Sjögren's syndrome compared with healthy controls (27) and have been shown to increase after B cell depletion therapy and then gradually return to baseline following reconstitution of the circulating pool (28, 29). We found in our study that median serum BAFF levels followed similar kinetics, rising substantially while circulating CD19+ B cells were maximally depleted and then returning toward baseline with reconstitution of the circulating B cell pool (Figures 3A and B).

Figure 3.

Relationship between blood B cell depletion and serum BAFF levels. A, Median serum BAFF levels (relative [Rel.] units) are significantly elevated above baseline at weeks 8 and 14, the time of maximal blood B cell depletion, and slowly decline to near baseline levels by week 52, consistent with reconstitution of the blood B cell pool. B, The relationship between the numbers of blood B cells and serum BAFF levels is plotted for individual patients at weeks 0, 8, 14, 26, 36, and 52.

Effect of rituximab therapy on the interferon (IFN) signature and expression of other gene transcripts.

In 8 patients, we analyzed the effects of rituximab therapy on IFN signature transcripts, which are up-regulated in primary Sjögren's syndrome (30), as well as the expression of other gene transcripts by comparing the transcript levels at baseline with those at weeks 8, 26, and 52 (31). Overall, we found a significant change in the expression of 94, 77, and 342 genes at these time points, respectively, compared to baseline (P < 0.001). Among 63 IFN-related transcripts (IFNs, interferon regulatory factors [IRFs], and interferon inducible [IFI] genes), only IRF-4, IRF-8, interferon-induced transmembrane protein 1 (IFITM-1), IFI-30, and IFITM-4P showed statistically significant changes (P < 0.01) between baseline and any of these 3 subsequent time points.

We observed a significant decrease after rituximab therapy in the expression of several B cell–related genes, including CD79A, LOC652493 (Ig κ-chain V-I region HK102-like), IGKV3D-20 (Igκ variable 3D-20), FCRLA, LOC647450 (similar to Ig κ-chain V–I region HK101 precursor), LOC652775 (similar to Ig κ-chain V–V region L7 precursor), VPREB3, and BLK (data available upon request from the corresponding author). Since the expression data were not normalized for the numbers of blood B cells, these changes probably reflect the depleting effects of rituximab. These changes were most pronounced at week 8 and returned toward baseline values by week 52.

We also performed a pathway analysis, with the top 3 canonical pathways mapping to primary immunodeficiency signaling (P < 0.01 for baseline versus week 8), altered T cell and B cell signaling (P < 0.01 for baseline versus week 8 and versus week 26), phosphatidylinositol 3-kinase signaling in B lymphocytes (P < 0.01 for baseline versus week 26), and B cell development (P < 0.01 for baseline versus week 8 and versus week 26).


Our results show that two 1-gm infusions of rituximab given 2 weeks apart produce effective depletion of circulating B cells in patients with primary Sjögren's syndrome, with kinetics and pattern of B cell subset reconstitution similar to those observed in another study of patients with the same disease (10). The corresponding increases in serum BAFF levels following the depletion of blood B cells confirm earlier observations (31). The exploratory analysis of gene transcripts and pathways provides further evidence that rituximab therapy substantially alters B cell responses in this patient population.

Importantly, we did not detect any unexpected safety signals, except for the possibility of exaggerated vaccine reactions. Three of the patients in our trial had an unusually severe reaction to the pneumococcal vaccine given at week 26. However, these vaccines were administered to 68 patients with rheumatoid arthritis who were treated with rituximab in a randomized open-label trial without apparent untoward effects beyond the usual occurrence of itching, rash, and soreness at the injection site, and malaise (32). Serum sickness has occurred in some patients with primary Sjögren's syndrome following a rituximab infusion (6, 7, 9), but such an event was not observed in the current trial. We attempted to minimize this risk by premedicating patients with 100 mg of methylprednisolone in addition to diphenhydramine and acetaminophen.

Despite the effective depletion of blood B cells, rituximab therapy was not associated with striking clinical benefits in the current trial. Any improvements observed in an open-label study must be interpreted with caution because of the inherent subjectivity of many of the disease measures and the possibility of observer bias. Previous studies investigated the potential clinical efficacy and safety of rituximab therapy for primary Sjögren's syndrome (6–9), but they too, were limited by their small sample size, open-label design in some cases, and the lack of standardized treatment outcomes.

The results of the present study are consistent with those of a 122-patient, randomized, placebo-controlled study of rituximab therapy in primary Sjögren's syndrome, which was reported in the form of an abstract (33). In that study, rituximab therapy was not significantly more effective than placebo in improving by ≥30 mm the scores on at least 2 of the 4 100-mm VAS scales evaluating dryness, pain, fatigue, and global disease activity. Since relatively few patients in our study had extraglandular manifestations beyond constitutional symptoms and joint pain, we were unable to explore the effects of rituximab therapy on severe systemic features. Rituximab therapy has been shown in some cases, however, to benefit such systemic features as refractory pulmonary disease, synovitis, and mixed cryoglobulinemia (34). To date, rituximab treatment has not been consistently associated with an increase in lacrimal and salivary gland function. It has been suggested that patients whose lacrimal and salivary glands have minimal secretory capacity due to long-standing disease may be particularly refractory to disease-modifying treatment (6). Indeed, 7 of the 12 patients in our trial had stimulated salivary flow rates <0.1 ml/minute at entry, and it may be argued that their glandular function had limited potential for improvement.

Rituximab therapy did not appear to affect the serum levels of anti-SSA/Ro and anti-SSB/La antibodies, although serum rheumatoid factors trended lower, as has been seen in earlier studies (8, 9). Rituximab treatment also had no impact on the serum levels of anti-M3R antibodies. This result is of interest because of the possible role of these autoantibodies in the mechanisms of impaired salivary flow. We also hypothesized that rituximab treatment would alter the blood IFN signature in primary Sjögren's syndrome, based on a previous study showing that it can induce type I IFN activity in patients with rheumatoid arthritis (35). However, we were unable to demonstrate such an effect in our study, which is not unexpected, based on the established relationship between type I IFN and BAFF. Elevated serum levels of BAFF, which result from B cell depletion, are not known to induce type I IFN activity. Rather, type I IFN has been shown to induce BAFF production (36).

In summary, rituximab treatment for primary Sjögren's syndrome in this small open-label trial was associated with no unexpected toxicities, except possibly for exaggerated vaccine reactions. It led to only modest improvements in symptoms and no beneficial changes in lacrimal or salivary gland function. In addition to detailed studies of blood B cell subsets, exploratory analyses of gene transcripts and pathways in the peripheral blood suggest rituximab therapy substantially alters B cells, while having little impact on the IFN signature. Larger randomized placebo-controlled clinical trials such as those reported by Devauchelle-Pensec and colleagues (33) are needed to further evaluate the clinical efficacy and safety of rituximab therapy for primary Sjögren's syndrome.


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. St.Clair 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. St.Clair, Levesque, Vivino, Wedgwood, Sivils, Cohen.

Acquisition of data. St.Clair, Levesque, Prak, Vivino, Alappatt, Wedgwood, Sivils, Fisher, Cohen.

Analysis and interpretation of data. St.Clair, Levesque, Prak, Spychala, McNamara, Sivils, Cohen.


The study was designed by the investigators and coordinated by the National Institutes of Allergy and Infectious Diseases (NIAID;the study sponsor) and Rho, Inc., which managed the collection and quality control of the data and performed the statistical analyses. Rituximab was provided at no cost by Genentech, which had no other role in the design, conduct, or analysis of the study. Publication of this article was not contingent upon approval by the NIAID, Rho, Inc., or Genentech.


The authors wish to acknowledge Dr. Thomas A. McGraw (Duke University Medical Center) for assistance with the conduct of the study, Dr. Yang-Zhu Du (University of Pennsylvania) for technical assistance with the flow cytometry experiments, Dr. Dennis Wallace (formerly of Rho, Inc.) for assistance with the study design, and Beverly Welch (National Institute of Allergy and Infectious Diseases) for project management.