Materials and methods

After informed consent was obtained, peripheral blood was drawn from 25 (24 female, 1 male) lupus patients hospitalised at our unit (Charité University Hospitals, Berlin, Germany). Patient characteristics are shown as supplementary material (available online only). In addition all methods are described in the supplement (supplementary text file, available online only).

Results

In the current lupus patient cohort, 12.2% (SD 10.6) of all circulating B cells were determined to be CD27⁺⁺CD20⁻CD19^dim. This subset has previously been found to be expanded in lupus patients comprising plasmablasts and plasma cells.4 5 As illustrated in fig 1, this expansion was mainly related to an increase of HLA-DR^high plasmablasts (median 86%; range 59–97%) compared with HLA-DR^low plasma cells. Interestingly and similar to peripheral blood, HLA-DR^high plasmablasts were also found to be expanded in the bone marrow of two lupus patients (fig 1). In contrast, only a low percentage of HLA-DR^high plasmablasts was found in bone marrow from healthy controls (1–1.5% of CD19⁺ cells or <0.1% of bone marrow mononuclear cells, respectively, fig 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

(A) Gating strategy. Peripheral blood mononuclear cells were labelled with fluochrome-tagged antibodies to CD19, CD20, CD27 and HLA-DR. After gating on CD19⁺ cells, CD27⁺⁺CD20⁻ cells were gated (plot) and HLA-DR^high and ^–low subsets (histogram) were analysed. The blue line shows the expression of HLA-DR, gated on the CD27⁺⁺CD20⁻ subset isolated from the peripheral blood, for comparison the corresponding bone marrow (BM) subset is shown in grey. (B) Expression of HLA-DR on peripheral blood (PB) and bone marrow B cells and plasmablasts as well as plasma cells in a representative patient with systemic lupus erythematosis (SLE). For comparison, the result of peripheral blood from a normal donor (ND) and the bone marrow from another normal subject is also shown.

Further analyses sought to investigate if any of these circulating CD27⁺⁺CD20⁻CD19^dim subsets correlate with disease activity in lupus patients. As a result, we identified a significant correlation between the absolute number or frequency of HLA-DR^high plasmablasts and the disease activity as assessed by the systemic lupus erythematosus disease activity index (SLEDAI) score (p<0.04 and p<0.004, respectively) or IgG anti-double-stranded DNA antibody levels (p<0.03, p<0.01, respectively) in 25 patients with SLE (table 1). Interestingly, analysis of the entire CD27⁺⁺CD20⁻CD19^dim subset revealed a correlation between the absolute cell number and lupus activity (p<0.005) and between its frequency and anti-dsDNA titres (p<0.03) but no other significant correlations. In contrast, HLA-DR^low plasma cells, total CD27⁺ memory B cells as well as CD27⁻ B cells did not correlate with disease activity or IgG anti-dsDNA antibody levels. These data clearly support that HLA-DR^high plasmablasts are the immediate product of immune activation in lupus and that this subset is associated with disease activity more closely than the entire CD27⁺⁺CD20⁻CD19^dim subset.

In line with these results, we found the percentage of HLA-DR^high plasmablasts within the CD27⁺⁺CD20⁻CD19^dim subset to be significantly correlated with the SLEDAI score (r_S  =  0.59, p<0.002, fig 2A) and the level of IgG anti-dsDNA antibodies (r_S  =  0.46, p<0.02, fig 2B). To test whether HLA-DR^high plasmablasts might be directly related to the IgG anti-dsDNA antibody serum levels, we analysed the frequency of IgG anti-dsDNA antibody-secreting cells among both the HLA-DR^high plasmablast subset and HLA-DR^low plasma cell subset in a lupus patient. Of note, the frequency of IgG anti-dsDNA antibody-secreting cells was more than twice as high in the HLA-DR^high plasmablast subset compared with the HLA-DR^low plasma cell subset (3.3% (5/153) vs 1.5% (3/198) of all IgG-secreting cells).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Significant correlation between the HLA-DR^high subset of CD27⁺⁺CD20⁻CD19^dim cells (%) and disease activity assessed using the systemic lupus erythematosus disease activity index (SLEDAI) (A) and IgG anti-dsDNA antibody levels (B). *Spearman rank correlation.

Discussion

This study identified that among circulating CD27⁺⁺CD20⁻CD19^dim cells known to be enhanced in active patients with SLE,4 5 recently generated HLA-DR^high plasmablasts predominate over HLA-DR^low plasma cells. Their absolute number and frequency correlate with disease activity more significantly than the entire CD27⁺⁺CD20⁻CD19^dim subset, which appears to be a mixed population that also contains HLA-DR^low plasma cells not related to disease activity. Although anti-dsDNA IgG antibody production could be identified in both subsets, there was an enhanced production by HLA-DR^high plasmablasts derived from peripheral blood further supporting their pathogenic role and relation with disease activity. Freshly generated HLA-DR^high tetanus-specific plasmablasts have been found to be expanded in the peripheral blood of healthy individuals 6–8 days after secondary immunisation.6 In comparison, most circulating CD27⁺⁺CD20⁻CD19^dim cells (86%; range 59–97%) express high levels of HLA-DR in patients with SLE. However, the expansion of HLA-DR^high plasmablasts does not appear to be restricted to the peripheral blood, because we could also readily identify them in bone marrow aspirates in two lupus patients. However, the source and fate of these HLA-DR^high plasmablasts as well as the mechanisms involved in their trafficking remain unknown.

In this context, a previous study identified an enlarged population of plasma cells in kidneys from patients with lupus nephritis adjacent to germinal centre-like structures and local Ig production in the affected organ,3 as has been shown in lupus nephritis8 and the spleen9 in animal models. These models offer the advantage of addressing various questions more thoroughly. Feeding NZB/W mice with bromodeoxyuridine, Hoyer et al.9 documented substantial populations of long-lived plasma cells and short-lived plasmablasts populating the spleen, with only the latter being sensitive to cyclophosphamide treatment. In line with these experimental data, the abundance of HLA-DR^high plasmablasts in different organs including peripheral blood and bone marrow is notable in patients. Whether these cells further differentiate into HLA-DR^low plasma cells and undergo specific selection processes or just reflect the characteristic overflow observed in the tissues of lupus patients remains to be elucidated.

Although a decline in HLA-DR^high plasmablasts noted almost immediately after methylprednisolone treatment as well as 2 months after B-cell depletion were noted in individual patients (supplementary fig 1, available online only), longitudinal studies of large cohorts of patients are required to assess the effect of particular drugs on HLA-DR^high plasmablasts. Several factors contributing to B-cell activation and differentiation into plasmablasts or plasma cells have been identified in patients with SLE and other autoimmune diseases. In this context, T-cell and antigen-dependent and independent pathways, certain cytokines and B-cell activating factor have been implicated to drive plasma cell differentiation from memory, naive and even transitional B cells in humans.10 It is thus also possible that the expansion of HLA-DR^high plasmablasts in lupus results from premature B-cell activation and differentiation. In this regard, CD27⁻IgD⁻CD95⁺ memory B cells showing an activated phenotype were found to be expanded in SLE patients with active disease.11

An important mechanism to control B-cell activation and differentiation at the post-germinal centre stage is mediated through FcγRIIb. Usually, FcγRII expression is upregulated on activated B cells and germinal centre cells.12 In contrast, memory B cells show a decreased surface expression of FcγRIIb associated with an impaired regulation of B-cell receptor signalling in a subset of patients with SLE,13 suggesting that the expansion of plasmablasts and autoantibody secretion in lupus patients might be related to an impairment of this post-germinal centre control mechanism, as also observed in autoimmune mouse models.14 15

In conclusion, expanded HLA-DR^high plasmablasts represent a more precise marker of B-cell activation in SLE closely related to disease activity, including the production of IgG anti-dsDNA antibodies. The expansion of these cells reflects a lack of immune regulation and selection, respectively, which requires further studies.