Circulating plasmablasts/plasma cells: a potential biomarker for IgG4-related disease

Forty-two untreated patients with IgG4-RD fulfilling the 2011 comprehensive IgG4-RD diagnostic criteria were enrolled in this study. The diagnosis of IgG4-RD was based on the following three manifestations: (1) clinical examination showing characteristic diffuse/localized swelling or masses in single or multiple organs; (2) hematological examination showing elevated serum IgG4 concentration (>135 mg/dl); and (3) histopathologic examination showing (a) marked lymphocyte and plasma cell infiltration and fibrosis or (b) infiltration of IgG4⁺ plasma cells (ratio of IgG4⁺/IgG⁺ cells >40% and >10% IgG4⁺ plasma cells per high-power field). Patients with cancer or lymphoma and other autoimmune diseases were excluded.

Clinical data, including age, gender, disease duration, and manifestations, were obtained for all patients. Laboratory findings were recorded, including erythrocyte sedimentation rate (ESR); C-reactive protein (CRP); and serum immunoglobulin IgG, IgA, IgM, and IgG subsets. IgG4-related Disease Responder Index (IgG4-RD RI) was calculated for each patient [18].

Peripheral blood mononuclear cells (PBMCs) from patients with IgG4-RD were separated by Ficoll gradient centrifugation. Different B-cell populations were stained with phycoerythrin (PE)-cyanine 7 (Cy7)-anti-CD19, fluorescein isothiocyanate-anti-CD24, allophycocyanin-anti-CD38, peridinin chlorophyll protein-CY5.5-anti-CD27, PE-anti-CD40, PE-anti-CD80, PE-anti-CD86, PE-anti-B-cell activating factor receptor (BAFF-R), PE-anti-transmembrane activator CAML (calcium modulator and cyclophilin ligand) interactor (TACI), PE-anti-CD95, PE-anti-CD138, PE-anti-interleukin-6 receptor (IL-6R), PE-anti-IgD, and PE-anti-CD59 monoclonal antibodies (mAbs) (BD Biosciences, San Jose, CA, USA), as well as PE-anti-B-cell maturation antigen (BCMA) mAb (BioLegend, San Diego, CA, USA), or isotype-matched controls. After incubation for 30 minutes at 4 °C, the cells were washed and resuspended in fluorescence-activated cell sorting staining buffer (BD Biosciences). All experiments were analyzed by gating on lymphocytes according to forward scatter/side scatter; dead or dying cells or granulocytes were excluded. B-cell subsets were gated on CD19 and then gated on CD24 and CD38. Flow cytometric analysis was performed immediately after sample preparation. All samples were analyzed using a BD FACSAria II system (BD Biosciences), and data were analyzed by FlowJo version 7.6.4 software (FlowJo, Ashland, OR, USA).

PBMCs were stained with CD19, CD24, and CD38 antibodies and then sorted with a MoFlo high-performance cell sorter (Beckman Coulter Life Sciences, Indianapolis, IN, USA). Different B-cell subsets were separated, including CD19⁺CD24⁺CD38⁻ memory B cells, CD19⁺CD24^intCD38^int naïve B cells, CD19⁺CD24^hiCD38^hi regulatory B cells (Bregs), and CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells. Sorted B-cell populations with purity greater than 95% were used for in vitro culture and RNA extraction.

Purified B-cell populations from patients with IgG4-RD were resuspended in RPMI 1640 medium supplemented with 10% fetal calf serum and antibiotics (penicillin 100 IU/ml, streptomycin 100 μg/ml; Life Technologies, Carlsbad, CA, USA) in 96-well U-bottomed plates (Nunc, Langenselbold, Germany) in a humidified atmosphere of 5% CO₂ at 37 °C with 1 × 10⁵ cells in each well. For each group, 100 ng/ml recombinant human CD40L (Abcam, Cambridge, MA, USA) and 0.1 μg/ml cytosine phosphate guanosine oligodeoxynucleotide 2006 (CpG ODN 2006; InvivoGen, San Diego, CA, USA) was added at the beginning. After 7 days of culture, the levels of Ig secretion in collected supernatants were tested.

Culture supernatant samples from different B-cell subsets were collected and stored at −80 °C until used. Cytometric bead array (CBA) analysis for IgG, IgA, IgM, IgG4, and IgE in supernatants was performed according to the manufacturer’s instructions (BD Biosciences). Data were analyzed using CBA analysis software obtained from BD Biosciences. The concentrations of IgG, IgA, IgM, IgG4, and IgE in supernatants were determined by reference to a standard curve.

Sorted B-cell subpopulations were placed in TRIzol reagent (Life Technologies) for RNA extraction following the manufacturer’s instructions. Isolated RNA was further purified with the RNeasy Mini Kit (Qiagen, Valencia, CA, USA) and processed for microarray analysis using the standard Affymetrix protocols (www.​affymetrix.​com; Affymetrix, Santa Clara, CA, USA). Briefly, 1–10 μg of RNA was reverse-transcribed into complementary DNA (cDNA) (Life Technologies). The template cDNA was purified for amplification and in vitro transcription to cRNA using the BioArray™ HighYield™ RNA Transcript Labeling Kit (T7) (Enzo Life Sciences, Inc., Farmingdale, NY, USA). cRNA was biotin-labeled, purified, and hybridized to Human Genome U133A GeneChips® (Affymetrix). GeneChips® were scanned on a high-resolution scanner using GCOS version 1.2 software (Affymetrix). Data analysis was conducted after standard Affymetrix algorithm analysis (MAS5).

Statistical analyses were performed using IBM SPSS Statistics version 19.0 software (IBM, Armonk, NY, USA) and Prism software version 5.0 (GraphPad Software, La Jolla, CA, USA). A P value <0.05 was considered significantly different. Data are reported as mean ± SD. Normal distribution data between two groups were analyzed using independent-samples t tests or paired-samples t tests, and one-way analysis of variance (ANOVA) was used to compare groups. The relationships between CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells and clinical features were analyzed by Pearson’s rank correlation test, and a P value <0.05 was considered significant. The level of gene expression was standardized by robust multiarray average and detection above background. One-way ANOVA was used to test different levels of gene expression. The Benjamini-Hochberg method was employed to determine the false discovery rate after multiple hypothesis testing. A false discovery rate <0.3 was used.

All 42 patients were newly diagnosed, untreated patients with IgG4-RD. Their demographic features as well as clinical and laboratory manifestations are listed in Table 1. Their average age was 55 (41.5–60) years old, and the male-to-female ratio was 2.23:1. Thirty-two (76.2%) of the patients were characterized as definite IgG4-RD, 1 (2.4%) was classified as probable IgG4-RD, and 9 (21.4%) were categorized as possible IgG4-RD. The majority of patients had multiple organ involvement. Forty-one (97.6%) patients had elevated serum IgG4.

Our previous studies revealed that CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells were significantly increased in the peripheral blood of patients with IgG4-RD (6.99 ± 6.24%), higher than that in patients with Sjögren’s syndrome (2.39 ± 2.64%, P < 0.001) and healthy control subjects (2.16 ± 1.65%, P < 0.001) [17]. The plot with the gating strategy for identifying the different B-cell subpopulations is shown in Fig. 1a. In this study, we further analyzed the association of these B-lineage populations with clinical and laboratory parameters of IgG4-RD. Figure 1b–h illustrates the correlations between CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells and clinical/laboratory parameters and shows that the ratio of CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells/CD19⁺ B cells correlated positively with serum IgG4 (R = 0.4852, P = 0.0001), the numbers of involved organs (R = 0.5039, P < 0.0001), and ESR (R = 0.3557, P = 0.0097), but that it had no correlation with serum IgG (P = 0.0874), serum IgE (P = 0.6020), or CRP (P = 0.2580). In addition, the absolute number of CD19⁺CD24⁻CD38^hi cells in patients with IgG4-RD was markedly elevated at 7100/ml (2711–16,297/ml), compared with 1475/ml (1021–2792/ml) in healthy subjects (P < 0.001). Importantly, the absolute number of CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells significantly correlated with serum IgG4 as well (R = 0.3088, P = 0.0466) (Fig. 1c).

To further analyze the characteristics of CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells, we collected PBMCs from 15 newly diagnosed patients with IgG4-RD, and the following B cell subsets were sorted by flow cytometry: Bregs (CD19⁺CD24^hiCD38^hi), memory B cells (CD19⁺CD24⁺CD38⁻), naïve B cells (CD19⁺CD24^intCD38^int), and plasmablasts/plasma cells (CD19⁺CD24⁻CD38^hi). Because the numbers of each B-cell subset were small, the subsets of B cells from individual patients were pooled. Total RNA was extracted for microarray analysis. Hierarchical clustering visualized by heat mapping indicated significant differences in gene expression by the four B-cell subsets (Fig. 2). Importantly, CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells were the most different from the other populations. Among genes that were differentially upregulated in CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells compared with naïve B cells and memory B cells, we found those that are known to alter the B-cell program and commit a B cell to plasma cell differentiation, such as interferon regulatory factor 4 (IRF4), PR domain containing 1 (PRDM1), and X-box binding protein 1 (XBP1). Ig heavy and light chains as well as J chain were also significantly upregulated, in addition to those previously reported to be upregulated in plasma cell genes, such as VDR, CAV1, SDC1, SSR4, ERP70, PPIB, ITGA6, CD38, and others listed in Tables 2 and 3 [19‐26]. Both IGK and IGL genes were upregulated, implying the polyclonal nature of the plasma cell expansion. Genes involved in plasma cell homing to tissue niches such as SDC1 were also upregulated, although CXCR4, which is also involved in homing of plasma cells to niches, was downregulated. Finally, tumor necrosis factor receptor superfamily member 17 (TNFRSF-17), which encodes BCMA, a BAFF-R involved in plasma cell survival, was greatly upregulated in CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells. As expected, BCR complex, CD19, CD20, CIITA, human leukocyte antigen (HLA)-DRA, and HLA-DRB1 gene expression was downregulated in CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells. Likewise, the genes characteristic of earlier stages of B-cell development or differentiation, such as SPIB, BCL6, BLK, FYN, BCL11A, CD37, CD1C, VAV3, CCR6, and CD22, were downregulated significantly.

The transcription factors regulating B-cell differentiation into plasmablasts/plasma cells were analyzed in the various B-cell subsets in greater detail (Fig. 3). In patients with IgG4-RD, B-cell lymphoma 6 protein (BCL6) and PRDM1, a pair of mutually restrictive transcription factors involved in B-cell differentiation, showed opposite expression in CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells compared with the other three B-cell subsets, with BCL6 expression being lower and PRDM1 expression higher in CD19⁺CD24⁺CD38^hi plasmablasts/plasma cells. In addition, compared with the other three B-cell subsets, the expression of paired box gene 5 (PAX5) in CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells was much lower. Genes related to PAX5, including IRF8, SPIB, Bach2, EBF, ID3, and CIITA, were also downregulated. These results confirm that CD19⁺CD24⁻CD38^hi cells are plasmablasts/plasma cells.

Next, we analyzed other important genes related to the differentiation of B cells. Compared with the other three B-cell subsets, CD22 was markedly downregulated in CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells, whereas HLA class II genes (HLA-DRA, HLA-DRB1), CIITA, and CD19 were significantly downregulated, compared with memory and naïve B cells (Fig. 4, Tables 2 and 3). High expression of IGHD, IGJ, IGK, and IGHM are important indicators for B-cell differentiation into plasma cells. In CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells from patients with IgG4-RD, IGH gene expression was also markedly increased, as was that of IGK and IGL genes. In addition, the expression of CD27 and Fas/CD95 genes was significantly higher than in the other three B-cell subsets, whereas CD40 gene expression was decreased. CD23 was also markedly decreased in CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells of patients with IgG4-RD.

Marker of proliferation Ki-67 (MKI67), encoding the Ki-67 protein and denoting recent cellular proliferation, was markedly increased in CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells compared with the other three B-cell subsets (Fig. 5). We also found that a number of genes involved in cell homing, including SELL, which encodes L-selectin (CD62L), was markedly increased, whereas expression of CCR7, CXCR5, and CXCR4 was greatly decreased, in CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells compared with the other three B-cell subsets. CCR10 expression was also markedly increased compared with memory B cells and naïve B cells (Fig. 5).

By flow cytometric analysis, we compared the expression of B-cell-activating factors and homing factors in different B-cell subsets in patients with IgG4-RD (Fig. 6). This analysis revealed that CD86, CD62L, HLA-DR, and CD95 were highly expressed in CD19⁺C24⁻CD38^hi plasmablasts/plasma cells compared with other B-cell subsets. In particular, the majority of CD19⁺CD24⁻CD38^hi B cells expressed CD27 (76.46 ± 7.16%), which demonstrated that the CD19⁺CD24⁻CD38^hi population was a larger population that included CD19⁺CD20⁻CD27⁺ plasmablasts/plasma cells. Whereas CD138, BAFF-R, and BCMA were expressed by a subset of CD19⁺CD24⁻CD38^hi plasma/plasma cells, there was almost no expression of TACI, CD23, CD40, IgE, and IgD by this population.

The CD19⁺C24⁻CD38^hi plasmablasts/plasma cells, Bregs, memory B cells, and naïve B cells from patients with IgG4-RD were isolated by cell sorting and cultured in vitro. At day 7, the supernatants were collected, and the secretion of immunoglobulin was tested.

The results showed that the IgG secreted by CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells was 618.4 ± 163.3 ng/ml, much higher than that of Bregs and naïve B cells (0.72 ± 0.59 ng/ml and 231.2 ± 44.83 ng/ml, respectively; (P = 0.003 and P = 0.02, respectively), but that it was not different from that of memory B cells (541.0 ± 252.1 ng/ml, P = 0.678). In addition, the IgG4 secreted by CD19⁺C24⁻CD38^hi plasmablasts/plasma cells was 176.1 ± 28.5 ng/ml, significantly higher than that secreted by Bregs, memory B cells, and naïve B cells (0.87 ± 0.59 ng/ml, P = 0.0005; 26.24 ± 7.08 ng/ml, P = 0.0009; and 5.34 ± 1.50 ng/ml, P = 0.0005, respectively). CD19⁺CD24⁻CD38^hi plasmablasts/plasma cells also secreted more IgE than other B-cell subsets (3.14 ± 0.99 ng/ml, 2.01 ± 0.01 ng/ml, 2.01 ± 0.01 ng/ml, and 2.00 ± 0.01 ng/ml, respectively; P = 0.054). Regarding other immunoglobulins, there was no significant difference in secreted IgA (20,573 ± 22,215 ng/ml, 12.39 ± 1.96 ng/ml, 5209 ± 1066 ng/ml, and 12105 ± 16970 ng/ml, respectively; P = 0.37) and secreted IgM (2090 ± 2401 ng/ml, 8.34 ± 1.75 ng/ml, 2385 ± 2402 ng/ml, and 2810 ± 2387 ng/ml, respectively; P = 0.41) compared with CD19⁺C24⁻CD38^hi plasmablasts/plasma cells and the other three subsets (Fig. 7).

Forty-two patients with IgG4-RD were assessed before treatment with glucocorticoids, and 15 patients with IgG4-RD were assessed both before and after treatment with glucocorticoids. Before treatment, both the ratio of CD19⁺CD24⁻CD38^hi cells/CD19⁺ B cells and the absolute number of CD19⁺CD24⁻CD38^hi cells correlated positively with disease activity in terms of IgG4-RD RI (R = 0.575, P < 0.0001; R = 0.373, P = 0.015, respectively) (Fig. 8a, b). After treatment, along with the decrease of IgG4-RD RI (from 13.48 ± 5.60 to 5.06 ± 2.38, P < 0.009) (Fig. 8f), serum IgG (from 25.74 g/L ± 12.55 g/L to 14.06 g/L ± 6.23 g/L, P < 0.001), and IgG4 (from 1740 mg/dl [7160–58,700] to 938 mg/dl [2900–17,300], P < 0.0001) (Fig. 8g, h), both the absolute number (from 5914 [1888–19,487]/ml to 2157 [822–3582]/ml, P = 0.005) and the ratio (from 7.25% ± 6.35% to 2.55% ± 1.50%, P < 0.001) of CD19⁺C24⁻CD38^hi plasmablasts/plasma cells decreased significantly (Fig. 8c–e). By contrast, circulating memory B cells, which were also increased in untreated patients [17], did not show significant changes after steroid treatment. Notably, however, there was no significant correlation between the magnitude of decrease of the ratio of CD19⁺CD24⁻CD38^hi plasma cells and the decrease of the IgG4-RD RI (R = 0.505, P = 0.055) (data not shown).