To determine phenotypic and functional abnormalities of blood B cell subsets in patients with systemic sclerosis (SSc).
To determine phenotypic and functional abnormalities of blood B cell subsets in patients with systemic sclerosis (SSc).
Cell surface marker expression was determined by flow cytometry. Spontaneous apoptosis was evaluated by annexin V expression with flow cytometric analysis. IgG production by isolated IgD− memory B cells was examined by enzyme-linked immunosorbent assay.
The numbers of blood CD27− naive B cells from SSc patients were increased compared with normal control cells, while memory B cells expressing medium levels of CD27 and plasmablasts expressing high levels of CD27 were reduced. In contrast, plasmablasts were the predominant population in patients with systemic lupus erythematosus (SLE). Memory B cells in SSc showed increased expression of activation markers, including CD80, CD86, and CD95, relative to normal controls. Consistent with CD95 up-regulation, SSc memory B cells exhibited augmented spontaneous apoptosis after 24-hour incubation; augmented apoptosis may explain the reduced memory B cell number. Nonetheless, isolated IgD− SSc memory B cells treated with stimuli had an enhanced ability to produce IgG. Furthermore, expression of CD19, a critical signal transduction molecule of B cells that regulates autoantibody production, was significantly increased in memory B cells as well as in naive B cells in SSc. In contrast, CD19 expression was decreased in SLE B cells.
SSc patients have distinct abnormalities of blood homeostasis and B cell compartments, characterized by expanded naive B cells and activated but diminished memory B cells. Our results suggest that CD19 overexpression in SSc memory B cells is related to their hyperreactivity.
Systemic sclerosis (SSc) is a connective tissue disease characterized by excessive extracellular matrix deposition in the skin and other visceral organs. Because clinical manifestations of SSc are heterogeneous, various subtypes of SSc have been proposed, such as CREST syndrome (calcinosis, Raynaud's phenomenon, esophageal dysmotility, sclerodactyly, telangiectasias). The most widely accepted classification system is that proposed by LeRoy et al (1), which includes limited cutaneous SSc (lcSSc), a type similar to CREST syndrome, and diffuse cutaneous SSc (dcSSc), a more rapidly advancing type with more frequent visceral involvement. The presence of autoantibodies is a central feature of SSc; antinuclear antibodies (ANAs) have been detected in >90% of patients (2). SSc patients have autoantibodies that react to various intracellular components, such as DNA topoisomerase I, centromere, RNA polymerases, U1RNP, and U3RNP (2). In addition, hyper-gammaglobulinemia and polyclonal B cell hyperactivity are detected in SSc patients (3, 4). Furthermore, a recent study demonstrated that down-regulation of B cell function leads to improved skin fibrosis in the tight-skin mouse, a genetic model of SSc (5). Although the pathogenesis of SSc remains unknown, the B cell abnormalities characterized by autoantibody production and polyclonal B cell activation play an important role. In particular, long-lived memory B cells are considered to play a crucial role as autoreactive B cells.
Although memory T cells can be distinguished from naive T cells by expression of different CD45 isoforms, individual cell surface markers that can directly identify all memory B cells were, until recently, not known. The most definitive marker of memory B cells identified to date is the presence of somatically mutated, high-affinity antigen receptors (6), but accumulating evidence has shown that cell surface CD27 is a useful marker of human memory B cells (6–12). CD27 is a type I glycoprotein expressed on some B cells and the majority of T cells, and is a member of the tumor necrosis factor receptor family (10). The interaction of CD27 with its ligand on T cells, CD70, can induce quick activation that leads to differentiation into plasma cells (9, 10). Importantly, recent studies on circulating B cells at the single-cell level confirmed that essentially all circulating CD27+ B cells display hypermutated, rearranged VH genes, while no mutations are identified in CD27− B cells (6, 7, 10, 13, 14). B cells with high levels of CD27 (CD27high) express low levels of CD19 and surface Ig, high amounts of CD38 and CD138, and no CD20, a pattern found on plasma cells (13, 15). The majority of these cells do not, however, look like mature plasma cells, but like plasma cell precursors (plasmablasts or early plasma cells); analysis by sorting and Giemsa staining revealed that they have larger, less peripheral nuclei and less abundant cytoplasm (15). Thus, CD27 is a reliable and useful marker for characterizing peripheral blood B cell subpopulations and homeostasis.
Recent studies using CD27 as a marker of memory B cells have revealed a disturbance of peripheral B cell compartments and homeostasis in systemic autoimmune disorders. Patients with systemic lupus erythematosus (SLE) exhibit an expanded population of CD27high,CD38+ plasmablasts that correlates with disease activity, while the number of CD27− naive B cells and CD27+ memory B cells is reduced due to marked B lymphocytopenia (13, 15, 16). However, patients with Sjögren's syndrome show a normal number of naive B cells, but a significantly reduced number of CD27+ memory B cells without expansion of plasmablasts (14, 17). Thus, these findings indicate distinct types of abnormal B cell homeostasis in some systemic autoimmune disorders.
Despite the potential contribution of B cell abnormalities to the pathogenesis of SSc, no studies have examined phenotypic and functional abnormalities in peripheral blood B cell compartments from SSc patients. The current study showed distinct altered B lymphocyte homeostasis characterized by an expanded naive B cell population, a diminished but highly activated memory B cell population, and a reduced plasmablast population in SSc.
Blood samples for most experiments were obtained from 39 Japanese SSc patients (37 women and 2 men). All patients fulfilled the criteria for SSc classification proposed by the American College of Rheumatology (ACR; formerly, the American Rheumatism Association) (18). The mean ± SD age of the SSc patients was 47 ± 13 years and the duration of their disease was 4.6 ± 6.7 years. These patients were grouped according to the classification system proposed by LeRoy et al (1): 17 patients (all women) had lcSSc and 22 patients (20 women and 2 men) had dcSSc. Fifteen dcSSc patients were receiving low-dose corticosteroid (mean prednisolone dosage 9.1 mg/day; range 5–25 mg/day). None of the SSc patients was treated with D-penicillamine (D-Pen) or immunosuppressive therapy. ANAs were determined by indirect immunofluorescence using HEp-2 cells as the substrate, and autoantibody specificities were further measured by enzyme-linked immunosorbent assay (ELISA) and immunoprecipitation. Anticentromere antibody was positive in 17 patients, anti–topoisomerase I antibody in 12, anti-U1RNP antibody in 2, anti-U3RNP antibody in 3, anti-RNA polymerase I and III antibodies in 3, and Th/To antibody in 1. The remaining 1 patient was negative for autoantibodies.
Eighteen patients who fulfilled the ACR criteria for SLE classification (19) were also examined at their initial visit or at the time of a flare during followup in our clinic. These patients had active SLE as determined by the SLE Disease Activity Index (20) (mean 15.8; range 13–18). Fifteen SLE patients received low-dose steroid treatment (mean prednisolone dosage 8.4 mg/day; range 5–15 mg/day). However, none of the SLE patients received immunosuppressive therapy. Eighteen age- and sex-matched healthy Japanese individuals (17 women and 1 man; age 45 ± 10 years) were used as controls. The protocol was approved by the Kanazawa University Graduate School of Medical Science and Kanazawa University Hospital, and informed consent was obtained from all patients.
Two-color analysis was performed using phycoerythrin (PE)–conjugated anti-CD19 (B4; Beckman Coulter, Miami, FL) and fluorescein isothiocyanate (FITC)–conjugated CD27 (M-T271; BD PharMingen, San Diego, CA) monoclonal antibodies (mAb). For study of the expression of activation markers, 3-color analysis was conducted using a combination of peridin chlorophyll protein–conjugated anti-CD19 (Leu-12; BD PharMingen), FITC-conjugated CD27 (M-T271; BD PharMingen), and PE-conjugated mAb, either anti-CD80 (MAB104; Beckman Coulter), anti-CD86 (Ancell, Bayport, MN), or anti-CD95 (UB2; Beckman Coulter). For immunofluorescence staining, fresh heparinized whole blood samples were placed on ice immediately after collection. Blood samples (50 μl) were stained at 4°C using predetermined saturating concentrations of the test mAb for 20 minutes, as previously described (21). Blood erythrocytes were lysed after staining using the Whole Blood Immuno-Lyse kit as detailed by the manufacturer (Beckman Coulter). Cells were analyzed on a FACScan flow cytometer (BD PharMingen). Ten thousand cells with the forward and side light scatter properties of mononuclear cells were analyzed for each sample, with fluorescence intensity shown on a 4-decade log scale. Positive and negative populations of cells were determined using nonreactive isotype-matched mAb (Beckman Coulter) as controls for background staining. Background levels of staining were delineated using gates positioned to include 98% of the control cells. Absolute numbers of cells were calculated from the relative frequency of total B cells and B cell subpopulations and the absolute lymphocyte counts.
For analysis of CD80, CD86, and CD95 expression on B cells, the patients examined were different from those described above. None of these patients was treated with steroids, D-Pen, or immunosuppressive therapy. For analysis of CD95 expression, 15 SSc patients (14 women and 1 man; age 50 ± 9 years; 11 with lcSSc and 4 with dcSSc) and 14 healthy controls (13 women and 1 man; age 45 ± 12 years) were examined. For analysis of CD80 and CD86 expression, a subset of these individuals was examined: 11 SSc patients (10 women and 1 man; age 56 ± 9 years; 8 with lcSSc and 3 with dcSSc) and 9 healthy controls (8 women and 1 man; age 52 ± 10 years). The frequency and absolute number of B cells in these 15 SSc patients examined for CD80, CD86, or CD95 expression were added to the data from the 39 SSc patients examined for B cell subsets.
Early apoptotic cells were detected with annexin V, a specific marker for the early phase of apoptosis, using an Annexin V-PE apoptosis detection kit (BD PharMingen). Peripheral blood mononuclear cells (PBMCs) were separated using Ficoll-Paque (Pharmacia Biotech, Uppsala, Sweden) after centrifugation. PBMCs were suspended in RPMI 1640 (Sigma, St. Louis, MO) containing 10% fetal calf serum (FCS) and incubated in a 5% CO2 incubator at 37°C for 5 or 24 hours. Cells were harvested and stained for 3-color flow cytometric analysis using PE-conjugated annexin V in combination with peridin chlorophyll protein–conjugated anti-CD19 (Leu12) and FITC-conjugated anti-CD27 (M-T271) mAb. Six SSc patients (all women; age 48 ± 6 years; 3 with lcSSc and 3 with dcSSc) and 6 healthy controls (all women; age 50 ± 9 years) were examined. None of the SSc patients was treated with steroids, D-Pen, or immunosuppressive therapy.
IgD− B cells were isolated using a B cell negative isolation kit (Dynal, Lake Success, NY) in combination with anti-IgD mAb (IA6-2, mouse IgG2a; BD PharMingen). IgD− B cells consisted of ∼70% CD27+ memory B cells and included IgM+,IgD−, IgG+,IgD−, and IgA+,IgD− memory B cells, but not IgM+,IgD+ memory B cells (7). Briefly, PBMCs (1 × 107) were incubated for 20 minutes on ice with mAb mixtures containing mAb to CD2, CD3, CD7, CD14, CD16, and CD56 plus anti-IgD mAb. Then, IgD− B cells were isolated by incubating magnetic beads coated with an Fc-specific human IgG4 antibody against mouse IgG. The isolation step was repeated 3 times and, after the isolation, >90% of cells were IgD−,CD19+ by flow cytometric analysis.
Purified IgD− B cells (1.0 × 105) were cultured in 0.2 ml of RPMI 1640 containing 10% FCS in 96-well flat-bottomed plates with stimuli, at 37°C for 8 days in a 5% CO2 humidified atmosphere. The cells were stimulated with either 0.01% Staphylococcus aureus Cowan strain (SAC; Sigma) plus 50 units/ml of interleukin-2 (IL-2; Sigma) or 1 μg/ml of anti-CD40 mAb (mAb89; Beckman Coulter) plus 50 ng/ml of IL-10 (R&D Systems, Minneapolis, MN), as previously described (11). The culture supernatants were harvested, and IgG concentrations were measured by ELISA (Bethyl Laboratories, Montgomery, TX). Six SSc patients (all women; age 45 ± 7 years; 3 with lcSSc and 3 with dcSSc) and 6 healthy controls (all women; age 48 ± 10 years) were examined. None of these SSc patients was treated with steroids, D-Pen, or immunosuppressive therapy.
The Mann-Whitney U test was used to determine the level of significance of differences between sample means, and the Bonferroni adjustment was used for multiple comparisons. Spearman's rank correlation was used to examine the relationship between 2 continuous variables. All values are shown as the mean ± SD. P values less than 0.05 were considered significant.
The frequency and absolute number of CD19+ B cells in blood from patients and healthy controls were examined first. To assess the effect of steroid treatment on B cell homeostasis, SSc patients were placed into 3 groups according to disease classification and steroid treatment: 11 untreated dcSSc patients, 28 untreated lcSSc patients, and 15 dcSSc patients treated with low-dose oral steroids. The frequency of B cells in untreated dcSSc and lcSSc patients was increased by ∼30% compared with healthy controls (P < 0.01) (Figure 1A). Similarly, the absolute number of B cells in untreated dcSSc and lcSSc patients was significantly higher than that found in healthy controls (P < 0.05) (Figure 1B). The frequency and absolute number of B cells were similar in untreated dcSSc and untreated lcSSc patients. Steroid treatment in dcSSc patients tended to decrease the B cell frequency and significantly reduced the absolute B cell number relative to untreated dcSSc patients (P < 0.005). In contrast to the SSc patients, SLE patients exhibited the normal frequency, but a significantly diminished absolute number of B cells (P < 0.0005).
To assess which subsets of peripheral B cells explain the increased absolute number of B cells in SSc, blood B cells were stained with anti-CD19 and anti-CD27 mAb in flow cytometric analysis. CD19+ blood B cells were grouped into CD27− naive B cells, memory B cells expressing medium levels of CD27 (CD27medium), and CD27high plasmablasts (13, 14). The frequency of CD27− naive B cells in SLE patients was normal, while in untreated lcSSc patients it was ∼20% higher than that found in healthy controls (P < 0.0001) (Figure 3A). A similar increase was observed for untreated dcSSc patients (P < 0.0001) (Figures 2 and 3A). Although only 7 untreated dcSSc patients were used in the analysis of B cell subsets, this patient group was included for the comparison with steroid-treated dcSSc patients. Similarly, the absolute number of naive B cells in untreated dcSSc and lcSSc patients was increased by ∼60% relative to healthy controls (P < 0.05 and P < 0.01, respectively) (Figure 3B). In contrast, SLE patients, due to B lymphocytopenia, exhibited a significantly diminished absolute number of naive B cells compared with healthy controls (P < 0.001). Steroid treatment of dcSSc patients did not eliminate the increased frequency of naive B cells (P < 0.0001 versus healthy controls), but it significantly decreased the absolute number of naive B cells relative to untreated dcSSc patients (P < 0.05).
In contrast to the increased number of naive B cells in SSc, both the frequency and the absolute number of CD27medium memory B cells in untreated lcSSc patients were 35–50% lower than those found in healthy controls (P < 0.0001 and P < 0.01, respectively) (Figures 2 and 3C and D). Similarly, untreated dcSSc patients exhibited significantly reduced frequency and absolute number of memory B cells compared with healthy controls (P < 0.0001 and P < 0.05, respectively). There was no significant difference in the frequency of memory B cells between SLE patients and healthy controls, while the absolute number in SLE patients was significantly decreased (P < 0.0001). The frequency of memory B cells was similar between steroid-treated and untreated dcSSc patients, while the absolute number in steroid-treated dcSSc patients was further diminished relative to untreated dcSSc patients (P < 0.05) and relative to healthy controls (P < 0.0001).
Like CD27medium memory B cells, CD27high plasmablasts were significantly lower (by 50–75%) in frequency and absolute number in untreated lcSSc patients compared with healthy controls (P < 0.005) (Figures 2 and 3E and F). Similarly, untreated dcSSc patients showed significantly reduced frequency and absolute number of plasmablasts compared with healthy controls (P < 0.05). In contrast, the frequency of plasmablasts in SLE patients was elevated by 2.3-fold relative to healthy controls (P < 0.01); however, the absolute number was normal due to B lymphocytopenia in the SLE patients. Steroid treatment of dcSSc patients increased the frequency of plasmablasts by 2.8-fold compared with untreated dcSSc patients (P < 0.05); however, the absolute number remained reduced (P < 0.05 versus healthy controls). There was no significant difference in the frequency and number of each B cell subset between untreated dcSSc and untreated lcSSc patients. The frequency or number of each B cell subpopulation did not correlate with serum levels of IgM and IgG, titers of ANA, or levels of anticentromere and anti–topoisomerase I antibodies as measured by ELISA (data not shown). In summary, SSc patients exhibited altered peripheral B cell homeostasis characterized by an increased number of naive B cells and a reduced number of memory B cells and plasmablasts.
It was shown in a previous study that total blood B cells from SSc patients exhibited significantly increased expression of CD19 (21), which is a general “rheostat” that defines signaling thresholds critical for humoral immune responses and autoimmunity in B cells (22, 23). Therefore, we investigated the cell surface density of CD19 on blood B cell subsets from patients and healthy controls by flow cytometry. On CD27− naive B cells, CD19 expression levels were significantly higher in untreated dcSSc patients (mean fluorescence intensity 491 ± 44; P < 0.0001), untreated lcSSc patients (479 ± 50; P < 0.0001), and dcSSc patients treated with steroids (453 ± 52; P < 0.01) than in healthy controls (411 ± 28) (Figures 2 and 4A). In contrast, SLE patients showed significantly reduced CD19 expression on naive B cells compared with healthy controls (341 ± 61; P < 0.0001). To exclude the potential biases that may have been introduced by the differences in the naive B cell numbers, the naive B cell number in isolated PBMCs was measured and then similar numbers of naive B cells were compared for CD19 expression levels, but the results were similar (data not shown).
Like naive B cells, CD27medium memory B cells showed significantly elevated CD19 expression levels (P < 0.0001) in untreated dcSSc patients (550 ± 51), untreated lcSSc patients (529 ± 40), and steroid-treated dcSSc patients (520 ± 42) compared with those in healthy controls (449 ± 37) (Figures 2 and 4B). In contrast, SLE patients had significantly decreased CD19 expression levels on CD27medium memory B cells (396 ± 47) relative to healthy controls (P < 0.0005). Similar results were obtained when similar numbers of memory B cells were compared (data not shown). CD27high plasmablasts exhibited markedly diminished CD19 expression levels that were similar in untreated lcSSc patients, untreated dcSSc patients, dcSSc patients treated with steroids, SLE patients, and healthy controls (Figure 2 and data not shown). Furthermore, in healthy controls, CD19 expression levels on memory B cells were slightly up-regulated (by 9%) compared with naive B cells (P < 0.005). Similarly, memory B cells showed significantly increased CD19 expression levels (by 10–16%) relative to naive B cells in untreated dcSSc patients (P < 0.05), untreated lcSSc patients (P < 0.005), dcSSc patients treated with steroids (P < 0.001), and SLE patients (P < 0.005). Thus, CD19 expression was increased on both naive and memory B cells from SSc patients, with higher expression levels on memory B cells.
To assess the activation status of B cell subsets, we investigated expression of various activation markers, including CD23, CD62 ligand (CD62L), CD69, CD80, CD86, CD95 (Fas), and HLA–DR, by immunofluorescence staining with flow cytometry analysis, in SSc patients who had not been treated with steroids. The SSc patients for this assay were different from those examined for blood B cell subsets and CD19 expression. On naive B cells, expression of all the activation markers examined was similar in SSc patients and healthy controls (Figure 5A and data not shown). In contrast, on CD27+ B cells (∼95% of CD27medium memory B cells), CD80 and CD86 expression was up-regulated in SSc patients relative to healthy controls (Figure 5A); as a result, SSc patients exhibited a 20–30% higher frequency of CD80+ and CD86+ cells in memory B cells compared with healthy controls (P < 0.0005 and P < 0.005, respectively) (Figure 5B). Similarly, the frequency of CD95+ cells in memory B cells was significantly increased by 66% in SSc patients relative to healthy controls (P < 0.0001). Expression of CD23, CD62L, CD69, and HLA–DR was normal on CD27+ B cells from SSc patients (data not shown). There was no significant difference between lcSSc and dcSSc patients (data not shown). Thus, memory B cells from SSc patients showed an activated phenotype with increased expression of CD80, CD86, and CD95.
CD95 is up-regulated following B cell activation, and its interaction with CD95 ligand on T cells induces CD95-mediated apoptosis (activation-induced cell death), which is essential in the down-regulation of B cell activation (24). In SSc patients, the combination of increased CD95 expression on memory B cells along with the reduced number of memory cells (Figures 2, 3, and 5) suggests that activation-induced cell death could explain this reduction. Therefore, spontaneous apoptosis was evaluated in CD27+ memory B cells from SSc patients who had not been treated with steroids, by staining with annexin V, a specific marker for the early phase of apoptosis. Five hours after incubation in medium containing 10% FCS, memory B cells from SSc patients showed an increased frequency of apoptotic cells that was similar to that found in healthy controls (Figure 6). In contrast, after incubation for 24 hours, memory B cells from SSc patients exhibited a further increased frequency of apoptotic cells that was 21% higher than that in healthy controls (P < 0.05). There was no significant difference between lcSSc and dcSSc patients (data not shown). Thus, spontaneous apoptosis occurred more frequently in memory B cells from SSc patients.
Hypergammaglobulinemia is very common in patients with SSc (3). To assess whether memory B cells from SSc patients increase IgG production, we examined IgG secretion by isolated memory B cells treated with stimuli. Purified blood IgD− memory B cells, which consisted of ∼70% CD27+ memory B cells (7), were stimulated with either SAC plus IL-2 or anti-CD40 mAb plus IL-10 for 8 days, and IgG concentrations in the cultured supernatants were measured by ELISA. None of the SSc patients was treated with steroids. SSc memory B cells stimulated with SAC and IL-2 produced 31% more IgG than those from healthy controls (P < 0.05) (Figure 7). Similarly, stimulation with anti-CD40 mAb and IL-10 resulted in 2.5-fold higher IgG production by SSc memory B cells relative to that by memory B cells from healthy controls (P < 0.05). However, autoantibodies specific for centromere and DNA topoisomerase I were not detected in the cultured supernatants (data not shown). Serum IgG levels were positively correlated with levels of IgG produced by memory B cells stimulated with SAC plus IL-2 (r = 0.857, P < 0.05) or anti-CD40 mAb plus IL-10 (r = 0.955, P < 0.01) in SSc. However, levels of IgG produced by stimulated SSc memory B cells did not correlate with autoantibody levels determined by ELISA (data not shown). There was no significant difference between lcSSc and dcSSc patients (data not shown). Thus, although the number of SSc memory B cells was decreased, their capacity to produce IgG was significantly augmented.
The present study revealed disturbed homeostasis of peripheral B cell subsets in SSc. Both dcSSc and lcSSc patients exhibited a pattern of expanded numbers of naive B cells and reduced numbers of plasmablasts and memory B cells, which had increased expression of CD80, CD86, and CD95. CD80 and CD86 are critical costimulatory molecules of B cells, and B cell activation is required to up-regulate expression of both molecules; thus, increased CD80 and CD86 expression on memory B cells indicates that they are constantly activated in SSc. Furthermore, CD95 expression is up-regulated after B cell activation, and increased CD95 expression coincides with the acquisition of sensitivity to CD95-mediated apoptosis (24); thus, up-regulated CD95 expression on SSc memory B cells may result in their enhanced spontaneous apoptosis and diminished number. Moreover, it is possible that the continuous loss of memory B cells and plasmablasts increases the production of naive B cells in an attempt to maintain B cell homeostasis in SSc. Remarkably, although memory B cells were decreased in SSc patients, they had a significantly enhanced ability to produce IgG, which may result in hypergammaglobulinemia.
Studies investigating phenotypic abnormalities of blood B cells are limited in SSc. Furthermore, previous studies examined only total B cells, not the naive and memory B cell compartments. Previous studies suggested that SSc B cells were activated since the frequency of B cells expressing activation markers, including HLA–DR and CD25, was increased in SSc (4, 25, 26). However, the present study showed a normal frequency of HLA–DR–positive cells and normal expression levels of HLA–DR in both naive and memory B cells (data not shown). This discrepancy may be due to the different antibodies used for assay: all naive and memory B cells had detectable HLA–DR expression in this study, while HLA–DR expression was detected in only ∼10% of total B cells in previous studies (25, 26). Due to the marked reduction in memory B cells and normal expression of activation markers on expanded naive B cells in SSc, it would be difficult to detect phenotypic abnormalities of B cells by examining total B cells as a whole. Thus, this is the first study to reveal the distinct activated phenotypic abnormalities in memory B cells in SSc.
We previously reported that B cells from SSc patients overexpress CD19 by ∼20% (21). CD19 is a critical cell surface signal transduction molecule on B cells that regulates signaling thresholds important for humoral immune responses and autoimmunity (22, 23). Remarkably, transgenic mice with a similar increase in CD19 expression exhibit hypergammaglobulinemia and produce characteristic autoantibodies with specificities similar to autoantibodies in human SSc (21). Furthermore, B cells from tight-skin mice display enhanced CD19 signaling with a constantly activated phenotype, hypergammaglobulinemia, and spontaneous autoantibody production; these phenomena are completely eliminated by the loss of CD19 expression from B cells (5). The present study showed that CD19 is overexpressed in naive B cells and, to a greater extent, in memory B cells from SSc patients. This CD19 overexpression was specific to SSc, since SLE patients exhibited significantly down-regulated CD19 expression, which is associated with decreased expression of CD21, a C3d receptor (21, 27). In SLE, this down-regulation of CD19 and CD21 expression was suggested to be secondary to interaction with immune complexes bearing fragments of C3 (27). Furthermore, up-regulation of CD19 expression on memory B cells compared with naive B cells may render memory B cells more responsive to transmembrane signals, since CD19 expression levels correlate closely with B cell function in mice (28). These results suggest that the CD19 overexpression in SSc memory B cells induces both the activated phenotype of memory B cells and augmented IgG production.
Although autoantibody production is a common feature of systemic autoimmune disorders, B cell abnormalities have been shown to differ among systemic autoimmune disorders (13–15, 17). The present study confirms that plasmablasts are the predominant blood B cell population in SLE (13, 15). However, a previous study showed a reduced frequency of naive B cells in SLE (13), while this study showed a normal frequency, although the absolute number was similarly diminished in both studies due to B lymphocytopenia. This discrepancy may be due to differences in race and/or level of disease activity in the patient groups studied.
In contrast to SLE, SSc patients had significant reductions in both memory B cells and plasmablasts. A similar reduction in memory B cells was observed in patients with Sjögren's syndrome (14, 17). In Sjögren's syndrome, preferential accumulation of memory B cells in the inflamed parotid gland may lead to their reduction in the blood (14). However, this is unlikely in SSc because few, if any, B cells infiltrated into the affected tissues (29). Instead, the continuous loss of memory B cells by activation-induced apoptosis in SSc could explain the reduction in blood memory B cells, which may result in the subsequent reduction in plasmablasts. However, the decrease in plasmablasts in SSc may also result from their preferential localization in the lungs, since an increased plasma cell number is seen within the lungs (30). Unlike SLE and Sjögren's syndrome, SSc exhibits an expansion of naive B cells, which could be explained by increased B cell production in bone marrow to compensate for the reduction in memory B cells and plasmablasts. Recent studies demonstrated an increased presence of blood germinal center precursors (CD19+,CD27−,CD38+) in juvenile SLE or Sjögren's syndrome (15, 17); however, such a B cell subset was not increased in SSc (data not shown). Thus, SSc had a pattern of disturbed B cell homeostasis that was distinct from other systemic autoimmune disorders.
Although steroid treatment did not affect the increased frequency of naive B cells in dcSSc, it reduced the number of expanded naive B cells almost to the normal level. Similarly, there was no significant difference in the memory B cell frequency between steroid-treated and untreated dcSSc patients; however, steroid treatment further decreased the absolute number of memory B cells in dcSSc to a level similar to that in SLE patients, most of whom were treated with steroids. Since steroids are potent apoptosis inducers (31), the reduction in memory B cells in SSc may be due to a higher responsiveness of memory cells to steroids. However, this possibility is unlikely, since steroid treatment did not further decrease the frequency of memory B cells in dcSSc patients. These results suggest that both naive and memory B cells are highly sensitive to steroids. Alternatively, the reduction in naive and memory B cells may result from decreased early B cell precursors, which are also highly sensitive to steroids (32). In contrast, the frequency of plasmablasts increased and their absolute number did not further decrease after steroid treatment in dcSSc, suggesting that plasmablasts have a markedly lower responsiveness to steroids compared with naive and memory B cells. Although this analysis involved only a small number of untreated dcSSc patients, collectively the results indicate that steroids may have a differential effect on various B cell subsets.
A recent study showed that memory B cells expressing CD80 are able to secrete particularly large amounts of class-switched Ig, and can efficiently and rapidly present antigen to T cells and activate them (33). Therefore, the increase in CD80+ memory B cells in SSc may be responsible for the augmented IgG production by the memory B cells. Furthermore, since autoreactive B cells may function as effective antigen-presenting cells to naive T cells, and this function depends in part on CD80 expression (34), dysregulation of the CD80+ memory B cell subset in SSc may contribute to autoimmune pathogenesis. However, autoantibodies could not be produced by stimulated SSc memory B cells (data not shown). Consistent with this, it has been demonstrated that autoantibodies are not produced by B cells cultured alone, but are produced by B cells and T cells cultured together, suggesting that collaboration between autoreactive T and B cells is essential for in vitro autoantibody production in SSc (35). Collectively, these findings suggest that the hyperreactivity of SSc memory B cells can induce more efficient and stronger activation of helper CD4+ T cells, resulting in further activation of B cells by helper CD4+ T cells. This positive feedback loop of T and B cell collaboration may finally lead to hypergammaglobulinemia, abnormal cytokine secretion, and spontaneous autoantibody production in SSc.
Critical roles of B cells in the development of autoimmunity and disease expression in animal models of systemic autoimmune disorders have been reported. Elimination of B cells in lupus-prone mice results in a complete abrogation of glomerulonephritis, vasculitis, and skin disease (36). Furthermore, lupus-prone mice with B cells that cannot secrete antibodies still develop nephritis and vasculitis (36). This finding suggests that B cells, independent of autoantibodies, are essential for lupus pathogenesis either by serving as antigen-presenting cells or by contributing directly to local inflammation through cytokine secretion. Skin fibrosis in tight-skin mice is consistently improved by the diminished B cell function induced by loss of CD19, which causes a parallel decrease in IL-6 production by B cells (5). Pathogenic autoantibodies from B cells are also important, since K/BxN mice, a model for human rheumatoid arthritis, have hyperactive B cells that cause hypergammaglobulinemia and produce arthritogenic autoantibodies (37). Recent studies have consistently shown that B cell depletion by anti-CD20 mAb is effective in treating patients with rheumatoid arthritis or SLE (38–40), suggesting that B cells are essential for disease expression of human systemic autoimmune disorders. Taken together, the current findings showing disturbed B cell homeostasis and memory B cell hyperactivity in SSc suggest that B cells are a potential target in the treatment of SSc.
We thank Ms M. Matsubara for technical assistance.