The Plasma Interleukin (IL)-35 Level and Frequency of Circulating IL-35+ Regulatory B Cells are Decreased in a Cohort of Chinese Patients with New-onset Systemic Lupus Erythematosus

Systemic lupus erythematosus (SLE) is a multisystemic autoimmune disease that is associated with the destruction of immune tolerance and activation of B cells. Interleukin (IL)-35 and IL-35-producing (IL-35+) regulatory B cells (Bregs) have been demonstrated to possess immunosuppressive functions, but their roles in the initiation and early development of SLE have not been explored. Here, we measured and compared the frequencies of blood regulatory B cell subsets and the concentrations of plasma IL-35, IL-10, IL-17A, tumor necrosis factor (TNF)-α, and interferon (IFN)-γ in 47 Chinese patients with newly diagnosed SLE and 20 matched healthy controls (HCs). The SLE patients had decreased percentages of IL-35+ B cells and IL-10+ B cells among the total blood B cells as well as decreased concentrations of plasma IL-35. In addition, higher levels of plasma IL-10, IFN-γ, TNF-α, and IL-17 along with higher frequencies of circulating plasma and memory B cells were observed in the SLE patients. The percentage of IL-35+ Bregs and the serum IL-35 level were inversely correlated with the SLE disease activity index and the erythrocyte sedimentation rate (ESR) levels. Our results indicate that IL-35+ Bregs and IL-35 may play protective roles in SLE initiation and progression.


Comparisons of demographic and clinical characteristics of SLE patients and HCs. As shown
in Table 1, comparisons of the clinical information between 47 new-onset SLE patients and 20 matched HCs indicated no significant difference in terms of age, gender, or white blood cell count. Compared with the HCs, the patients with newly diagnosed SLE demonstrated significantly elevated levels of plasma immunoglobulin (Ig) G, IgA, IgM, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP), while they had decreased levels of complement factors C3 and C4. In addition, varied SLE Disease Activity Index (SLEDAI) scores (mean value, 18; range,  were identified among the SLE patients. Furthermore, over half of these SLE patients had positive sera for anti-double stranded (ds) DNA antibody (59.57%) and anti-Smith (Sm) antibody (53.19%) ( Table 1). SLE patients exhibited decreased frequencies of circulating IL-35 + Bregs, IL-10 + Bregs, and IL-5 + Bregs among blood B cells. It has been demonstrated previously that Bregs can secrete IL-10 and IL-35 to suppress the overactivated immunity in autoimmune diseases 7 . To identify the potential role of IL-35 + B cells and other Breg subsets, we analyzed the percentages of circulating IL-35 + B cells, IL-10 + B cells, and CD5 + Bregs between newly diagnosed SLE patients and HCs by flow cytometry (Fig. 1A). The results indicated a significantly decreased frequency of IL-35 + B cells among the total CD3 − CD19 + blood B cells in the SLE patients, compared with that in the HCs (Fig. 1B). Similarly, the SLE patients also demonstrated a significant reduction in the frequencies of IL-10 + Bregs and CD5 + Bregs in peripheral blood B lymphocytes (Fig. 1B). Age (years, range) 27  29  Gender ratio (female/male) 41 Fig. 3I) were also identified. However, we did not observe any correlations between the level of plasma IL-10 or IFN-γ and the frequencies of Breg subsets or plasma IL-35 levels in the SLE patients (data not shown). The plasma IL-35 level was negatively correlated with the frequency of circulating CD27 + CD38 − memory B cells in new-onset SLE patients. Based on CD27 and CD38 expression, we further characterized the frequencies of CD27 + CD38 − memory B cells, CD27 + CD38 + plasma B cells, CD27 − CD38 + transitional B cells, and CD27 − CD38 − naive B cells among CD3 − CD19 + B lymphocytes by flow cytometry (Fig. 4A). Subsequent comparisons on the abundance of these distinct B cell subsets between the HCs and SLE patients revealed that the SLE patients had significantly higher frequencies of CD27 + CD38 − memory B cells, CD27 + CD38 + plasma B cells, and CD27 − CD38 + transitional B cells (Fig. 4B). Meanwhile, a significant decrease in the frequency of CD27 − CD38 − naive B cells was observed in the SLE patients (Fig. 4B). Furthermore, we noted that the plasma IL-35 level in the SLE patients was negatively correlated with the frequency of the CD27 + CD38 − memory B cell population (P = 0.0020, r = −0.4390;  Fig. 6F). In addition, we also identified a positive correlation between the frequency of CD5 + Bregs and IgA levels (P = 0.0226, r = 0.3511; Fig. 6G), a negative correlation between the levels of

Discussion
B cells are critical players in the initiation and perpetuation of autoimmunity in SLE 1,2 . Activated B cells not only produce large amounts of autoantibodies, secrete cytokines, and present antigens 14,15 , but they also act as negative regulators of immunity 4,5,7 . In this study, we found that the new-onset SLE patients had higher frequencies of circulating CD27 + CD38 − memory B cells, CD27 + CD38 + plasma B cells, and CD27 − CD38 + transitional B cells than the HCs. The percentage of the plasma B cell subset showed positive correlations with the SLEDAI score and other clinical indicators such as CRP and IgA levels in the SLE patients. Together with previous findings, these observations suggest that the enriched B cell subset may not only play a pathogenic role in the development  www.nature.com/scientificreports www.nature.com/scientificreports/ of SLE but also can be considered as a potential biomarker to assess disease severity. Moreover, we revealed a decreased frequency of CD27 − CD38 − naive B cells in the peripheral blood of SLE patients, which is consistent with previous studies 16 . Furthermore, a similar pattern of changes in the frequencies of different B cell subpopulations was found in the SLE patients. Collectively, these results indicate that imbalanced distributions of different B cell subsets occur in the peripheral blood of patients with newly diagnosed SLE.
In a mouse model of EAE, IL-35 + Bregs have been identified to be novel key players in immunosuppression by Shen et al. 13 . What's more, we observed a decreased frequency of blood IL-35 + Bregs and a reduced plasma IL-35 concentration in new-onset SLE patients. Notably, the abundance of this subset and its functional cytokines were negatively correlated with the SLEDAI score and ESR value. Our data proposed that IL-35 + Bregs and IL-35 might be involved in the pathogenesis of SLE and may act as regulators in autoimmunity. We noted that the percentage of circulating IL-35 + Bregs and the plasma IL-35 level were inversely correlated with the levels of the inflammatory cytokines IL-17 and TNF-α, which can be secreted by Th17 cells. Th17 cells are pathogenic and accelerate organ impairment in various animal models and patients with SLE [17][18][19][20] . Considering the findings of recent studies indicating the direct suppression of effector T cell proliferation and function by IL-35 in vitro 11,21 , we speculated that IL-35 + Bregs and their functional cytokine IL-35 might suppress the Th17/IL-17 axis, thus playing a protective role in the pathogenesis of SLE.
Comprising two subunits, IL-12A (p35) and Epstein-Barr virus-induced 3 (EBI3), IL-35 has recently been considered as a new immunosuppressive/anti-inflammatory cytokine 11,22 . IL-35 can mediate signaling either through the heterodimer of receptor chains IL-12R b2/gp130 or the homodimer of each chain 23 . IL-35 is expressed by both Bregs and regulatory T cells (Tregs) 10,24 . In our study, we observed that the IL-35 level was positively correlated with the percentages of circulating IL-35 + Bregs and IL-10 + Bregs in patients with SLE. These observations indicated that IL-35 was not only produced by the IL-35 + Breg subset, but it also might be expressed by IL-10 + Bregs, which is consistent with a previous study suggesting that IL-35 + Bregs and IL-10 + Bregs are overlapping cell subsets 25 . Moreover, we identified a negative correlation between IL-35 expression and the abundance of CD27 + CD38 + plasma cells. Previous studies also have established that IL-35 can expand the numbers of IL-35 + Bregs and IL-10 + Bregs as well as induce IL-10 production 12,25 , which, combined with our findings, led us to hypothesize that IL-35 may function as a pivotal regulatory cytokine to promote the generation of Bregs by inducing naive B cells to develop into IL-35 + Bregs and IL-10 + Bregs as well as inhibiting the conversion of pathogenic plasma B cells in the pathogenesis of SLE. We found that the percentage of blood IL-10 + Bregs was decreased in www.nature.com/scientificreports www.nature.com/scientificreports/ the SLE patients. This subset showed a negative correlation with SLE disease severity indicators, including the SLEDAI score and the ESR. Consistent with published reports 6,8,9 , our data also indicated that IL-10 + Bregs might possess regulatory and protective properties in SLE progression and that their dynamic changes might be associated with the progression of human SLE.
IL-10 production is the most-studied mechanism of IL-10 + Bregs 7,26 , which confers the ability of these cells to maintain tolerance to self-antigens and to suppress inflammation and autoimmune responses 25 . Interestingly, we observed elevated circulating IL-10 levels in the SLE patients. The increased IL-10 level showed no correlations with IL-10 + Bregs or other Breg subsets, but it did correlate with SLE disease severity parameters, including C4 (negatively correlated) and CRP (positively correlated), suggesting that the total amount of circulating IL-10 plays a proinflammatory role in SLE. IL-10 is derived from different cell types and exerts dual functions under different contexts in the pathogenesis of autoimmune diseases 27 . Apart from its anti-inflammatory and immunosuppressive effects, IL-10 is also a potent cofactor for B cell survival, proliferation, differentiation, and Ig secretion 28 . www.nature.com/scientificreports www.nature.com/scientificreports/ Besides Breg-derived IL-10, IL-10 is produced by monocytes and T cells in SLE patients [29][30][31] . Together with previous studies 6,8,9,32 , we speculate that IL-10 + B cell-derived IL-10 comprises only a small fraction of the pool of IL-10 in patients with SLE. In a mouse model of lupus, Kalampokis et al. identified that IL-10 + Bregs regulate autoimmune responses and have protective and potentially therapeutic effects 31 . However, besides the fact that IL-35 can be produced by IL-10 + Bregs as discussed above, other mechanisms of action of IL-10-independent Breg cell subsets, including the production of IgM, the generation of adenosine, and the expression of programmed death-ligand 1 or Fas ligand, still need to be clarified in future studies.
CD5 has been shown to be expressed on a subpopulation of B cells with unique properties 3 . CD5 + Bregs can produce pathogenic autoantibodies and present antigens to T cells. Alternatively, they can also secrete IL-10 7 . Thus, the CD5 + Breg subset is identified as an important regulator of autoimmunity and an inducer of immune tolerance 3 . Our study demonstrated a reduced frequency of circulating CD5 + Bregs in patients with newly diagnosed SLE, which was positively correlated with the IgA level. However, no correlation was found between the frequency of blood CD5 + Bregs and the plasma IL-10 level. CD5 + Bregs are more likely to be involved in SLE by the dynamic changes of their number and the pathogenic antibodies they produce. Since a reduced number of CD5 + Bregs may correspondingly result in the reduced production of IL-10 by CD5 + Bregs, IL-10 from other sources might play a predominant role in the pathogenesis of SLE.
In conclusion, our study suggests that IL-35 expression is reduced in patients with newly diagnosed SLE, as both the frequency of circulating IL-35 + Bregs and the level of plasma IL-35 are significantly decreased in new-onset SLE patients. Our work only focused on Breg subsets and their roles in SLE pathogenesis, and the involvement of other regulatory cell subsets like Tregs, plasmacytoid dendritic cells (pDCs) or myeloid-derived suppressor cells (MDSCs) cannot be ignored. They also play very important roles in the pathogenesis of SLE. Tregs can suppress the activation, expansion, and differentiation of multiple types of cells including CD4+ T helper cells, CD8+ T cells, and B cells [33][34][35] . And some studies have already reported reduced numbers or impaired function of circulating Tregs in SLE patients, though some others found no apparent abnormalities and even increased levels of Tregs in SLE as compared with healthy controls 36 . As to pDCs, an aberrant regulatory feedback was found between pDCs and Bregs in SLE 37 . There were studies showed that MDSCs significantly suppressed the proliferation of CD4+ T cells, and the production of IFN-γ was significantly decreased in CD4+ T cells co-cultured with MDSCs 38 . Although a further study with a larger sample size and multiple time points is required to decipher the molecular and cellular mechanisms of IL-35 + Bregs on SLE progression, our current study implies that IL-35 + Bregs are important negative autoimmune regulators in SLE initiation and progression, our current study implies that IL-35 + Bregs are important negative autoimmune regulators in SLE initiation and progression.

Materials and Methods
Study subjects. This study involved a total of 47 new-onset SLE patients who were recruited between November 2016 and January 2018 from the Department of Rheumatology, The First Hospital of Jilin University, Changchun, China. Twenty healthy volunteers, as controls, with matched ethnicity, age, and gender, were recruited from the Physical Examination Center of the Outpatient Department of The First Hospital of Jilin University. These SLE patients were diagnosed according to the revised SLE classification criteria established by the American College of Rheumatology 39 , and they demonstrated clinical symptoms for <3 months without receiving any treatments. The disease activity of these new-onset SLE patients was assessed by the SLEDAI 40 , and active disease was indicated by an SLEDAI score >6. A total of 20 HCs, with matched age, gender, and ethnicity, were enrolled from the Physical Examination Center of Jilin University, Changchun, China. The participants with any kind of other autoimmune disorder, malignancy, or microbial infection, or those who had been treated with immunosuppressive reagents within the past 6 months were excluded from this study. All the experimental procedures involving human samples were conducted with strict adherence to the guidelines of the Declaration Quantitation of plasma cytokine levels by cytometric bead array (CBA). The levels of plasma IL-10, IL-17A, TNF-α, and IFN-γ were measured using the CBA method, as described previously 42 . A FACSAria II flow cytometer was used for data acquisition. The kits were purchased from BD Biosciences, and the experimental procedures as well as data analysis were performed according to the manufacturer's instructions.
Statistical analysis. Data were analyzed using the Statistical Package for the Social Sciences software (version 21.0; IBM, Armonk, NY, USA). Quantitative data were represented as the mean ± standard error of the mean, with each dot in the plots denoting an individual value. The Kruskal-Wallis analysis of variance test followed by the Dunn-Bonferroni post hoc method or the Mann-Whitney U test was used to evaluate differences between the indicated groups. The Spearman's rank correlation test was employed to analyze correlations between the indicated groups. P values < 0.05 were considered statistically significant.