Venous blood samples were collected from pSS patients (n = 13 females ages 32 to 64 years (average age = 50.5)) and normal individuals (n = 12 females ages 26 to 60 years (average age = 43.5)) after receiving their informed consent. Patients fulfilled the American-European Consensus Group criteria for pSS 33 . At the time of the collection of blood samples, two patients (15.4%) were receiving prednisolone at a daily dose < 5 mg. The remaining patients were free of medication. This study was approved by the ethics committees at Keio University School of Medicine and Saitama Medical University.

Peripheral monocytes were isolated as follows: Whole blood was mixed with RosetteSep Human Monocyte Enrichment Cocktail (StemCell Technologies, Vancouver, BC, Canada) and centrifuged over Ficoll-Hypaque (Beckman Coulter, Fullerton, CA, USA). A monocyte-enriched fraction was collected and cultured overnight in RPMI 1640 (American Tissue Culture Collection, Manassas, VA, USA) supplemented with 10% FCS in a humidified incubator (7% CO₂) at 37°C so that the expression of various stress-induced genes subsided. The cells were then washed once with the medium to remove debris. Fluorescence-activated cell sorting (FACS) analysis of the cells demonstrated that > 96% of the living cells were CD14-positive.

The monocytes were cultured in the absence or presence of various concentrations of IFN-γ or sBAFF, and the cumulative production of sBAFF and/or IL-6 was examined by ELISA. The production was dependent on the incubation period. The optimal incubation period was found to be 96 hours. The production of the cytokines increased almost in proportion to the concentration of stimuli up to 200 ng/ml IFN-γ or 2 μg/ml sBAFF.

An anti-BAFF mAb for ELISA was prepared in our laboratory 6 . A rabbit polyclonal anti-BAFF antibody and recombinant human sBAFF were purchased from Chemicon International (Temecula, CA, USA). Recombinant human IFN-γ, a control mouse IgG1, and mAbs for measurement of the amount of IL-6 by ELISA (MQ2-13A5 and MQ2-39C3 for capture and detection, respectively) and for FACS analysis (CD4-APC (RPA-T4) for T cells, CD14-PE-Cy7 (M5E2) for monocytes, CD20-APC-Cy7 (L27) for B cells and CD268-FITC (11C1) for BAFF-R) were purchased from BD Biosciences/Pharmingen (San Diego, CA, USA). An anti-TACI antibody for FACS analysis (CD267-PE (FAB1741P)) was purchased from R&D Systems (Minneapolis, MN, USA).

Monocytes were cultured at 2.5 × 10⁵/ml for 96 hours in a 24-well plate (2 ml/well) in the presence of stimuli (that is, recombinant human IFN-γ or recombinant human sBAFF). The amounts of sBAFF (in response to IFN-γ as a stimulus) and IL-6 (in response to IFN-γ or sBAFF as stimuli) in the culture supernatants were measured by sandwich ELISA according to previously described methods 6 , except for the concentrations of capture and detection antibodies for IL-6, which were prepared at 0.5 μg/ml.

For quantitation of sBAFF, we used our own anti-BAFF mAb, which specifically detects sBAFF and does not react with a proliferation-inducing ligand (APRIL) 6 . In our hands, the sensitivity of our ELISA system was better than that of commercially available ELISA kits (R&D Systems) in the range of 0.4 to 100 ng/ml sBAFF (data not shown).

The expression levels of BAFF, BAFF-R, TACI, NF-IL6, NF-IL6β, NF-κB1, NF-κB2 and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were quantitated by using a method described previously 6 . The following oligonucleotides were used as primers for PCR: 5'-ggaatctctgatgccacagctc and 5'-accttcaagggctgtcaaagatg (BAFF-R); 5'-agcatcctgagtaatgagtggcc and 5'-gagcttgttctcacagaagtatgc (TACI); 5'-aaaactttggcactggggcacttg and 5'-catctttaagcgattactcagggc (NF-IL6); 5'-agatgcagcagaagttggtggag and 5'-tagcttctctcgcagtttagtgg (NF-IL6β); 5'-atgggatctgcactgtaactgc and 5'-tcatagatggcgtctgataccacg (NF-κB1); 5'-cctgactttgagggactgtatcca and 5'-gcagcatttagcagcaaggtcttc (NF-κB2). Primer sets for BAFF and GAPDH were designed as described previously 6 . The expression level of each gene underwent dual normalization against GAPDH expression and expression of the same gene in unstimulated normal monocytes.

FACS and data analyses were carried out on a MACSQuant Analyzer (Miltenyi Biotec, Bergisch Gladbach, Germany). FACS analysis of cells in whole blood was carried out according to methods recommended by the manufacturer of the antibodies (BD Biosciences/Pharmingen).

Differences between groups were examined for statistical significance by using the two-tailed Student's t-test for single comparisons. Two-way analysis of variance (ANOVA) was also employed when appropriate. A P value less than 0.05 denoted the presence of a statistically significant difference.

Peripheral monocytes were prepared from primary pSS patients and normal individuals. The clinical characteristics of the pSS patients involved in this study are listed in Table 1. The cells were cultured for 96 hours in the absence or presence of IFN-γ (200 ng/ml), which is known to activate monocytes 34 and upregulate the expression of BAFF 2 . Stimulation of the cells was confirmed by the induction of interferon-gamma inducible protein 10 (data not shown). pSS monocytes released a significantly higher amount of sBAFF (5.4 ± 0.8 ng/ml) into the culture media than normal monocytes did (1.6 ± 0.3 ng/ml), even in the absence of stimulation, suggesting dysregulated production of sBAFF in pSS monocytes (Figure 1A, "Normal -" and "pSS -"). IFN-γ stimulation (Figure 1A, "Normal +" and "pSS +") resulted in an increase in sBAFF in both normal (6.6 ± 1.6 ng/ml) and pSS monocytes (21.1 ± 2.1 ng/ml).

RT-PCR analysis indicated that the expression of the BAFF gene in pSS monocytes was distinctly elevated upon stimulation with IFN-γ (Figure 1B). Quantitative RT-PCR analysis indicated that the relative expression level of the BAFF gene was about three times higher in pSS monocytes than in normal monocytes under unstimulated conditions. The expression levels increased about sixfold in both normal and pSS monocytes upon stimulation with IFN-γ. These data are basically consistent with the results derived by ELISA. Therefore, we postulated that the elevated production of sBAFF was the consequence of the enhanced expression of the BAFF gene.

We also investigated whether the production of IL-6 by pSS monocytes was abnormal. As indicated in Figure 2, pSS monocytes produced significantly higher amounts of IL-6 than normal monocytes without stimulation (Figure 2, open column vs checkerboard column; P < 0.01). Stimulation of pSS monocytes with 200 ng/ml IFN-γ induced a striking increase (8.6-fold; P < 0.001) in IL-6 production (Figure 2, checkerboard column vs closed column). Since IFN-γ induced the expression of BAFF (Figure 1) and BAFF is able to activate monocytes 35 36 , these results suggest that BAFF produced by monocytes may act in an autocrine fashion to augment the expression of IL-6. To test this hypothesis, we stimulated pSS monocytes with IFN-γ in the presence of an anti-BAFF mAb 6 . Interestingly, the mAb suppressed IL-6 production in part, but significantly so (P < 0.05) (Figure 2, closed column vs hatched column, whereas a control antibody had no effect (Figure 2, closed column vs gray column). These results suggest that the signal transduction pathway mediated by BAFF is implicated in the regulation of IL-6 production by IFN-γ-primed monocytes.

If this is really the case, then exogenously supplemented sBAFF should affect the production of IL-6 by monocytes. As expected, recombinant human sBAFF induced the production of IL-6 by both normal (Figure 3, closed circles) and pSS (Figure 3, open circles) monocytes in a dose-dependent manner. pSS monocytes produced approximately six times more abundant IL-6 than normal monocytes in the presence of 2 μg/ml sBAFF (Figure 3). It should be noted that two-way ANOVA revealed that disease status (normal or pSS) had significantly stronger effects than stimulation with sBAFF (P < 0.001 for cell type × stimulation interaction) on IL-6 production by monocytes.

A 2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide (XTT) assay in a separate experiment indicated that 2 μg/ml sBAFF supported the survival of both normal and pSS monocytes during the culture period. However, there was no significant difference in survival rates between normal and pSS monocytes (data not shown), suggesting that higher production of IL-6 was not simply a consequence of enhanced survival of monocytes. These data suggest that the regulatory mechanism for IL-6 production is aberrant in pSS monocytes.

Although it has been reported that BAFF receptors are mainly expressed in lymphocytes 37 , our results suggest that BAFF receptors are also expressed in monocytes. RT-PCR detected mRNA for BAFF-R in monocytes. Notably, the expression level of BAFF-R was significantly elevated in pSS monocytes (2.1-fold; P < 0.001) compared to normal monocytes (Table 2). In accordance with these data, FACS analysis indicated that approximately 60% of pSS monocytes were BAFF-R-positive (mean fluorescence intensity (MFI) = 50), whereas only about 25% of normal monocytes were positive to the same antibody (MFI = 20) (Figure 4A). The proportion of TACI-positive monocytes was relatively low compared to BAFF-R-positive cells (Figure 4B). Although the population of TACI-positive monocytes seemed to increase slightly in pSS compared to control monocytes (Figure 4B), the expression level of the TACI gene was not significantly increased (Table 2).

FACS analysis of lymphocytes in whole blood indicated that there were no significant differences between pSS patients and normal individuals in the population of BAFF-R-positive B and T cells (Figures 4C and 4D). All these data suggest that the expression of BAFF-R is dysregulated in pSS monocytes and that this dysregulation seems to be specific to monocytes among the cells examined thus far.

In an attempt to elucidate a possible mechanism of sBAFF-mediated overproduction of IL-6 by pSS monocytes, we investigated the expression levels of transcription factors involved in the expression of the IL-6 gene (that is, NF-IL6 (CCAAT/enhancer-binding protein β), NF-IL6β (CCAAT/enhancer-binding protein δ), NF-κB1 and NF-κB2). The relative expression levels of all the transcription factors were significantly elevated in pSS monocytes compared with the control (Table 3). Remarkably, the relative expression level of NF-IL6 was more than six times higher in pSS monocytes than in normal monocytes. These data indicate that the expression of IL-6-regulating transcription factors was abnormally upregulated in pSS monocytes.