BAFF Suppresses IL-15 Expression in B Cells

B cell–activating factor is a crucial factor that regulates B cell maturation, survival, and function. An excess of BAFF has been shown to lead to the development of autoimmune disorders in animal models, and BAFF concentrations are known to be higher in patients with various autoimmune conditions compared with normal subjects (1). The pathogenic role of BAFF in systemic lupus erythematosus (SLE) was revealed in a study that showed BAFF-transgenic mice develop a lupus-like illness with the production of anti-DNA Abs and development of glomerulonephritis (2–4). Increased levels of BAFF reported in some studies correlate with disease activity (5). In multiple sclerosis (MS) patients, BAFF is expressed by astrocytes that are closely associated with BAFFR–expressing cells (6) within ectopic lymphoid follicles of the meninges (7). BAFF has been regarded as a new therapeutic target in many autoimmune diseases (8).

On March 9, 2011, the Food and Drug Administration approved belimumab, a fully human anti-BAFF mAb, as a new B cell–specific treatment for SLE. Belimumab selectively reduces the number of CD20⁺ naive B cells, activated B cells, and plasmablasts and results in a transient increase in the number of memory cells (9). Atacicept (TACI)-IgG shows similar clinical results to belimumab by binding a portion of the receptor TACI to block the effects of survival factors BAFF and a proliferating-inducing ligand (APRIL). It reduces levels of circulating mature B cells and plasma cells in the spleen and bone marrow as well as inhibiting T cell activation, but not memory B cells (10, 11). As long as the reactive memory is maintained, it is probably acceptable that in principle humoral immunological response can be restored from immunological memory (12).

IL-15 provided survival signals that maintain memory T and B cells in the absence of Ag (13–15). Therefore, we propose that BAFF suppresses memory B cells by downregulating IL-15 expression. In the current study, we found that BAFF suppressed IL-15 expression in B cells from lupus-like or experimental allergic encephalomyelitis (EAE) mice. When BAFF was blocked with TACI-IgG, IL-15 expression was upregulated. The study provides hints for the clinical application of BAFF- and IL-15–specific therapeutic agents.

The blood samples were taken with the approval of the local ethical committee (Clinical Trial Center, Beijing Institute of Basic Medical Science), and all participants gave written informed consent to participate in the current study.

Care, use, and treatment of mice in this study were in strict agreement with international guidelines for the care and use of laboratory animals. This study was approved by Animal Ethics Committee of Beijing Institute of Basic Medical Sciences.

Blood samples were obtained after the approval from the Beijing Institute of Basic Medical Sciences, consent from 15 normal human subjects and 11 SLE patients from Clinical Trial Center (Beijing 307 Hospital), and consent from 15 MS patients from Department of Neurology, Beijing Chaoyang Hospital.

PBMCs were isolated from the heparin-treated whole blood by density gradient centrifugation, and CD19⁺ B cells were isolated using human CD19 MicroBeads (AutoMACS; Miltenyi Biotec). PBMC or CD19⁺ B cells were stimulated with 1 μg/ml LPS or 10 μg/ml goat F(ab′)₂ anti-human IgM (Southern Biotechnology Associates, Birmingham, AL) to activate the cells in the presence of 50 ng/ml human rBAFF (PeproTech, Rocky Hill, NJ).

Nine-week-old C57BL/6, BALB/c mice, and (New Zealand Black/New Zealand White)F₁ mice (Chinese Academy of Medical Sciences, Beijing, China) were bred in our animal facilities under specific pathogen-free conditions.

EAE induction was performed, as previously described (16). Briefly, C57BL/6 mice at 9 wk of age received a s.c. injection of 125 μg myelin oligodendrocyte glycoprotein (MOG)35–55 peptide (Mimotopes) emulsified 1:1 (v/v) in CFA containing 4 mg/ml Mycobacterium tuberculosis H37Ra (Difco, Detroit, MI), to the base of the tail and both flanks. Pertussis toxin (300 ng pertussis toxin in PBS; List Biological) was injected i.p. at the time of induction, and a second dose was administered 3 d later. The same volume of CFA and pertussis toxin was injected into each mouse in the control group. Animals were weighed, monitored, and clinically assessed according to the following grading scale: 0 = no signs; 1 = distal tail weakness; 1.5 = tail weakness and some hind limb weakness; 2 = complete tail paralysis; 2.5 = complete tail paralysis and partial hind limb weakness; 3 = complete hind limb weakness; 3.5 = inability to right when placed on back or significant forelimb weakness; 4 = euthanized or spontaneous death (17, 18). Mice were euthanized if they lost 20% of their starting weight, displayed a clinical score of 3 for 72 h, or reached a clinical score of 3.5. Mice were examined for up to 21 d postimmunization. Number of mice per treatment group was 12.

EAE and lupus-like mice were divided into the following six groups: 1, control mice; 2, untreated; 3, human IgG treated; 4, TACI-IgG treated; 5, TACI-IgG plus rabbit IgG treated; 6, TACI-IgG plus anti–IL-15. Twelve EAE mice per group were injected i.v. with 2 mg/kg TACI-IgG and/or 0.5 mg/kg anti-mouse IL-15 neutralizing Ab (R&D Systems) on days 4, 8, 12, and 16 (one time per day) after EAE was induced with MOG33–55. Twelve (New Zealand Black/New Zealand White)F₁ mice (lupus-like mice) per group were injected i.p. with 5 mg/kg TACI-IgG and/or 1 mg/kg rabbit anti-mouse IL-15 neutralizing Ab (R&D Systems) at 1, 2, 3, and 4 wk (twice per week) after the mice reached 6 mo of age.

There is no marker for memory B cells in mice (12). Thus, we used a two-step procedure to isolate the memory B cells from six mice per group by memory B cell isolation kit (MACS Miltenyi Biotec; Order 130-095-838), according to the manufacturer’s protocol (Miltenyi Biotec). Briefly, we first depleted nonmemory B cells with nonmemory B cell biotin–Ab mixture and anti-biotin microbeads. Additional labeling was done with anti-IgG1 allophycocyanin and anti-IgG2ab allophycocyanin. The magnetically labeled cells were passed through a column placed in the magnetic field of a MACS separator. Secondly, anti-allophycocyanin microbeads were used for indirect magnetic labeling of memory B cells. Positive selection of memory B cells was sorted by the magnetic field of a MACS separator. Memory B cells were fluorescently stained with B220-VioBlue (130-094-287), CD38 PE (130-097-087), anti-IgG1 allophycocyanin, and anti-Ig2ab allophycocyanin, followed by FACS analysis. Flow cytometric analysis demonstrated >90% memory B cells (data not shown). Finally, we counted the absolute number of memory B cells per mouse.

CD19⁺ B cells were isolated using mouse CD19 MicroBeads (AutoMACS; Miltenyi Biotec) and cultured in RPMI 1640 medium containing 10% FBS, 2 mM glutamine, penicillin (100 IU/ml), streptomycin (100 μg/ml), and 50 mM 2-ME. Cells were stimulated with 1 μg/ml LPS, 10 μg/ml goat F(ab′)₂ anti-mouse IgM (Southern Biotechnology Associates), MOG35–55 peptide (20 μg/ml precoated in 96-well plates), or dsDNA (20 μg/ml precoated in 96-well plates) in the presence of 50 ng/ml rBAFF (PeproTech).

All cell experiments were strictly prepared on ice, unless otherwise stated in other specific procedures. Cells (1 × 10⁶ cells/sample) were washed with FACS staining buffer (PBS, 2% FBS or 1% BSA, 0.1% sodium azide). All samples were incubated with 2.4G2 anti-FcRs (BD Pharmingen), prior to incubation with other Abs diluted in FACS buffer supplemented with 2% anti-FcR Ab. For intracellular cytokine staining, 50 ng/ml PMA and 1 μg/ml ionomycin (all from Sigma-Aldrich) were added, and then 10 μg/ml brefeldin A and 2 μM monensin were added 3 h later. Another 3 h later, cells were collected and fixed for 20 min with 1 mL fixation buffer (Fix and Perm cell permeabilization kit; eBioscience). After washing, the fixed cells were stained. The samples were filtered immediately before analysis or cell sorting to remove remaining clumps. Dead cell exclusion/discrimination dyes were used to eliminate dead cells from analysis and sorted. For nonfixed cells, we used DAPI, SYTOX Blue, Green, or Red; for fixed cells, we use LIVE/DEAD Fixable Blue, Green, or Red. Choice of dead cell exclusion dye depended on the color combination of fluorochromes within the sample. The following Abs were purchased from eBioscience: anti-mouse CD27, CD3, B220, CD19, CD93, CD21, CD23, and CD4 Abs. Anti–IL-15 Ab was purchased from R&D Systems. Data collection and analyses were performed on a FACSCalibur flow cytometer using CellQuest software.

The concentration of cytokines (IL-15, BAFF, and dsDNA) was measured by ELISA kits. Anti-mouse IL-15, BAFF, and dsDNA ELISA kits were purchased from R&D Systems. Briefly, diluted supernatants were added in triplicate to the plate for 1 h at 37°C. Then after washing, biotin rat anti-mouse IL-15, BAFF, and dsDNA (4 μg/ml) Abs were added to the plate and were incubated for another hour at 37°C. Thereafter, unbinding Abs were washed off, followed by addition of avidin-HRP (1/1000 diluted) (all Abs were obtained from eBioscience). Plates were incubated for 1 h at 37°C. Finally, the color was developed by incubation with o-phenylenediamine. The OD was read at 492 nm with an ELISA reader (Bio-Rad). Standard curves were established to quantitate the amounts of the respective cytokines.

Lymphocytes were collected from lymph nodes (LN) in different groups of EAE. The cells (5 × 10⁵ cells/well) were stimulated for 72 h with 0, 1, 5, 15, and 30 μg/ml MOG35–55. Forty-eight hours later, the cultures were pulsed with ³H-thymidine (0.5 μCi), and data were expressed as mean cpm ± SE of responses of five replicate cultures. The results were expressed as the stimulation index (cpm with Ag divided by cpm with medium alone). The data shown were from only one of the Ag concentrations; however, results were obtained with various concentrations of the Ags.

All RNA samples were DNA free. cDNA synthesis and quantitative PCR (qPCR) analyses were performed, as previously described (19). Each gene-specific primer pair used for qPCR analysis spanned at least an intron. Primers and probes used for qPCR were purchased from Applied Biosystems, and mRNA expression was normalized to the levels of β-actin gene.

For analysis of transitional B cells, multicolor flow cytometry (three, four, or five colors) was performed by gating on CD19⁺CD93⁺ B cells that were either CD21⁻/CD23⁺ (transitional 1 B [T1B] cells) or CD21⁺CD23⁺ (transitional 2 B [T2B] cells). For analysis of mature B cells, multicolor flow cytometry (three, four, or five colors) was performed by gating on CD19⁺CD93⁻ B cells that were either CD21⁺/CD23⁺ (follicular B [FB] cells) or CD21⁺CD23⁻ (marginal zone B [MZB] cells). All flow cytometry data were acquired with FACSCanto, FACSCantoII, or FACSAria (BD Biosciences), gated on live lymphocyte-sized cells on the basis of forward and side scatter, and analyzed using FlowJo software (Tree Star, Ashland, OR). The following Abs were purchased from eBioscience: PerCP-conjugated anti-mouse CD19, FITC-conjugated anti-mouse CD93, PE-conjugated anti-mouse CD21, and allophycocyanin-conjugated anti-mouse CD23.

Foxo1-specific short hairpin RNA (shRNA) with GFP (Santa Cruz Biotechnology, Santa Cruz, CA) were transfected into CD19⁺ B cells. The B cells were stimulated for 72 h with 1 μg/ml LPS. The GFP⁺ (Foxo1 was reduced) and GFP⁻ T2B cells were sorted from cultured cells by flow cytometry. All RNA samples were DNA free. cDNA synthesis and qPCR analyses were done to detect Foxo1 and IL-15 expressions. Mouse recombinant Foxo1 construct pEGFP-N1/Foxo1 was generated by rPCR. Plasmid pEGFP-N1/Foxo1 was transduced into CD19⁺ B cells by Lipofectamine 2000 (Invitrogen, Carlsbad, CA). Cells were stimulated for 72 h with 1 μg/ml LPS and analyzed by FACS. Cells were also sorted by flow cytometry. cDNA synthesis and qPCR analyses were done to detect IL-15 expression.

To clone series of the human IL-15 gene promoter and to construct six luciferase reporter gene vectors pGL3 containing IL-15 promoter regions, six DNA fragments of the 5′-flanking region of human IL-15 gene were isolated from genomic DNA of HaCaT cells by PCR with IL-15 promoter primer sets (Table I). Empty vector or pEGFP-N1/Foxo1 and luciferase reporter vector pGL3/IL-15 promoter were cotransduced into 293T cells. Dual luciferase reporter gene expression was analyzed, and the results were shown as the ratio of firefly to Renilla luciferase activity.

CD19⁺ B cells were stimulated with 10 μg/ml goat F(ab′)₂ anti-mouse IgM (Southern Biotechnology Associates) in the presence of 50 ng/ml human rBAFF (PeproTech). Cells were stimulated for 72 h. For Foxo1 phosphorylation analysis, CD19⁺ B cells were stimulated for 0, 30, and 60 min with 10 μg/ml goat F(ab′)₂ anti-mouse IgM (Southern Biotechnology Associates) in the presence of 50 ng/ml human rBAFF (PeproTech). The whole-cell lysates were prepared. Blots were probed with anti–β-actin, Foxo1, and p-Foxo1 (Santa Cruz Biotechnology). Preimmune serum was used in parallel as controls and signals were detected with HRP-conjugated secondary F(ab′)₂ (Zymed Laboratories, San Francisco, CA) using ECL system (Amersham, Arlington Heights, IL).

Statistics were generated using t test in GraphPad Prism (version 5.0; GraphPad Software), and values are represented as mean ± SEM. Results were considered statistically significant at p < 0.05.

BAFF levels have been shown to be increased in many autoimmune diseases, such as SLE, MS, and their animal models (Supplemental Fig. 1). When BAFF was blocked with TACI-IgG, numbers of mature B cells (Supplemental Fig. 2A–C) and the titer of anti-dsDNA autoantibodies (Supplemental Fig. 2D) were decreased in lupus-like mice. Previous studies have shown that TACI-IgG does not reduce memory B cells (10, 11). However, there is no marker for memory B cells in mice (12). Thus, we used a two-step procedure to isolate the memory B cells from six mice per group with a memory B cell isolation kit. The absolute memory B cell number per mouse was counted, and the results suggested that TACI-IgG treatment significantly upregulated memory B cell number in lupus-like mice (Fig. 1A). These results are in accordance with the clinical data suggesting that the total number of circulating memory B cells is preserved (or sometimes even increased) in belimumab- or TACI-IgG–treated patients.

IL-15 is an important cytokine for the survival and maintenance of memory T and B cells (13–15). Thus, we propose that blocking BAFF with TACI-IgG upregulates memory B cells by inducing IL-15 in lupus-like mice. As expected, TACI-IgG treatment upregulated IL-15 expression in the serum (Fig. 1B), spleen, and LN (Fig. 1C) from lupus-like mice. Because BAFF, APRIL, and their receptors play important roles in the B cell arm of the immune system (20), we assessed whether TACI-IgG upregulated B cells to express IL-15 by qPCR and flow cytometry (FACS) analysis. We showed that IL-15 expression was upregulated mainly in B cells from TACI-IgG–treated lupus-like mice (Fig. 1D–F).

To further prove the effect of TACI-IgG on IL-15 expression in B cells, TACI-IgG was also used to treat EAE mice, resulting in the reduction of B cell numbers (Supplemental Fig. 3A–C), clinical scores (Supplemental Fig. 3D), and autoantigen-induced immune response (Supplemental Fig. 3E). However, TACI-IgG treatment also upregulated both the memory B cell number (Fig. 2A) and IL-15 expression in the serum (Fig. 2B) of EAE mice. Further experiments showed that TACI-IgG induced B cells to upregulate the expression of IL-15 (Fig. 2C–F).

The evidence for the role of IL-15 in autoimmune condition is compelling. ELISA has demonstrated that IL-15 levels are increased in many autoimmune diseases, such as SLE, MS, and their animal models (Supplemental Fig. 1). qPCR and FACS analysis have further shown that IL-15 was upregulated mainly in the spleen in autoimmune diseases (Fig. 3). IL-15/IL-15R has been used as a therapeutic target in inflammatory autoimmune diseases (21). In addition, our results presented above suggest that TACI-IgG treatment upregulates memory B cells by inducing IL-15 in B cells. Thus, the combination of TACI-IgG and anti–IL-15 Ab was used to treat lupus-like and EAE mice. Once IL-15 was blocked with neutralizing anti–IL-15 Ab, the memory B cell number was significantly reduced in TACI-IgG–treated lupus-like (Fig. 1A) and EAE (Fig. 2A) mice. This study provides hints for the clinical application of BAFF- and IL-15–specific therapeutic agents.

The above experiments demonstrated that IL-15 was upregulated mainly in the spleen but not in LN (Figs. 1C–F, 2C–F, 3). Thus, we focused on splenic B cells. To determine the effect of TACI ligand (BAFF or APRIL) on IL-15 expression in B cells, we used autoantigen dsDNA (precoated in 96-well plate) plus BAFF or APRIL to stimulate B cells from lupus-like mice. FACS assay showed that BAFF but not APRIL significantly suppressed IL-15 expression in dsDNA-stimulated B cells from lupus-like mice (Fig. 4A). Furthermore, qPCR assay also proved that BAFF but not APRIL significantly suppressed IL-15 mRNA expression in B cells (Fig. 4B).

To further prove the effect of BAFF on IL-15 expression in B cells, we used MOG35–55 (precoated in 96-well plates) plus BAFF to stimulate B cells from EAE mice. BAFF significantly suppressed IL-15 expression in MOG35–55-stimulated B cells from EAE mice (Fig. 4C, 4D). In comparison, BAFF also slightly suppressed IL-15 expression in B cells from wild-type (WT) mice (Fig. 4A, 4B). To further prove these results, we also used different conditions such as LPS or anti-IgM to stimulate B cells from WT mice. qPCR analysis showed that BAFF suppressed IL-15 expression in LPS (Fig. 5A)- or anti-IgM (Fig. 5B)–stimulated B cells from WT mice. Similarly, BAFF also suppressed IL-15 expression in anti-human IgM-stimulated B cells from healthy donors (Fig. 5C). These results suggest that BAFF suppresses IL-15 expression in B cells induced by different stimuli. Furthermore, BAFF suppresses the secretion of IL-15 by CD19⁺ B cells but not CD4⁺ T cells (Fig 4C).

Next, we determined which population of splenic B cells was crucial in the BAFF suppression of IL-15. We used autoantigen MOG35–55 (precoated in 96-well plates) plus BAFF to stimulate B cells from control and EAE mice. Different populations of B cells were analyzed by FACS (Fig. 6A). The results showed that T2B cells expressed a high level of IL-15, whereas T1B cells, MZB cells, and FB cells expressed low levels of IL-15. Among these cell types, BAFF suppressed the expression of IL-15 from T2B cells (Fig. 6A, 6B). Furthermore, we sorted T1B, T2B, FB, and MZB cells from BAFF-treated B cells, as described in Fig. 6A, by FACS (Supplemental Fig. 4) and then used qPCR to detect IL-15 expression. The results suggest that IL-15 is expressed mainly in T2B cells but not in T1B, MZB, or FB cells (Fig. 6C). In addition, BAFF suppressed IL-15 expression in T2B cells (Fig. 6C). Meanwhile, we found that BAFF promoted the transformation of T1B cells into T2B cells and T2B cells into FB cells (Fig. 6D, 6E). These results suggest that BAFF suppresses T2B cells from secreting IL-15 and promotes the transformation of T2B cells into mature B cells.

Next, we examined which transcription factor is responsible for IL-15 expression in T2B cells. Foxo1 was mainly expressed in T2B cells but not in T1B, MZB, or FB cells (Fig. 7A, upper panel), which is in accordance with IL-15 expression (Fig. 7A, lower panel). We used Foxo1-specific shRNA to reduce Foxo1 expression in B cells (Fig. 7B, upper panel) and then sorted the T2B cells. qPCR analysis showed a lower IL-15 expression in Foxo1-reduced T2B cells than Foxo1-unreduced T2B cells (Fig. 7B, lower panel). When overexpressed in B cells, Foxo1 upregulated IL-15 mRNA (Fig. 7C, lower panel) and protein (Fig. 7C, upper panel) expression. These results suggest that Foxo1 regulates IL-15 expression in T2B cells. To prove that Foxo1 directly regulates IL-15 expression, we constructed six luciferase reporter gene vectors pGL3 containing IL-15 promoter regions. By dual luciferase reporter gene expression analysis, we found that Foxo1 bound the region from 1500 to 1800 bp upstream of the IL-15 promoter (Fig. 7D, Table I).

Subsequently, we detected whether BAFF suppressed Foxo1 expression. We used autoantigen dsDNA or MOG35–55 (precoated in 96-well plates) plus BAFF to stimulate B cells from lupus-like or EAE mice, respectively. BAFF significantly suppressed Foxo1 expression in dsDNA-stimulated B cells from lupus-like mice (Fig. 8A) and MOG35–55-stimulated B cells from EAE mice (Fig. 8B). To further prove the results, we used different conditions such as LPS or anti-IgM to stimulate B cells from WT mice. qPCR analysis showed that BAFF suppressed Foxo1 expression in LPS (Fig. 9A)- or anti-IgM (Fig. 9B)–stimulated B cells from WT mice. Similarly, BAFF also suppressed Foxo1 expression in anti-human IgM-stimulated B cells from healthy donors (Fig. 9C). These results suggest that BAFF suppresses Foxo1 expression induced by different stimuli. Further Western blot analysis also showed that BAFF suppressed Foxo1 expression in CD19⁺ B cells (Fig. 8C). Additionally, we found that BAFF induced Foxo1 phosphorylation (Fig. 8D), which suggests that BAFF could also reduce Foxo1 level in the nuclei by inducing Foxo1 phosphorylation. Furthermore, blocking BAFF with TACI-IgG induced Foxo1 expression in B cells from lupus-like (Fig. 8E) and EAE (Fig. 8F) mice. These results altogether suggest that BAFF suppresses IL-15 in B cells by downregulating Foxo1 expression.

BAFF demonstrates sp. act. toward B cells and supports B cell proliferation, differentiation, and survival (22). BAFF transgenic mice harbor increased numbers of B220⁺ B cells and plasma cells in the spleen and LN; they also develop anti-dsDNA Abs, proteinuria, and glomerulonephritis consistent with a systemic lupus-like autoimmunity as they age (2–4). By contrast, BAFF knockout mice have a markedly reduced mature B cell population and decreased serum Ig levels (22, 23). This study demonstrates that BAFF is increased in autoimmune diseases, such as SLE, MS, and lupus-like and EAE animal models (Supplemental Fig. 1). BAFF is regarded as a potential new therapeutic target in autoimmune diseases (8). These studies certainly suggest that BAFF is a positive immune regulator in autoimmune diseases; however, our current study shows that BAFF could suppress IL-15 expression in B cells stimulated with autoantigens (Fig. 4) as well as nonspecific stimulation reagents (such as LPS and anti-IgM) (Fig. 5), which implies that BAFF has a different role in autoimmune response.

Many trials have been done to study the effects of BAFF-specific inhibition, such as belimumab and TACI-IgG on SLE and other autoimmune diseases. Our data demonstrate that TACI-IgG treatment effectively controls the disease in lupus-like (Supplemental Fig. 2) and EAE (Supplemental Fig. 3) mice by reducing the number of mature B cells. However, memory B cells were upregulated in TACI-IgG–treated lupus-like (Fig. 1A) and EAE (Fig. 2A) mice. These data showed a strong correlation with previous studies suggesting that belimumab and TACI-IgG effectively reduce the number of circulating mature B cells but not memory B cells (10, 11).

Previous studies have shown that the survival signals maintaining memory T and B cells in the absence of Ag are provided by IL-15 (13–15). Our studies demonstrate that blockade of BAFF with TACI-IgG upregulates IL-15 levels in B cells but not in other cells, such as T cells (Figs. 1–3). These results suggest that BAFF inhibitors upregulate IL-15 expression in B cells so that they could not control memory B cells. Once the combination of TACI-IgG and anti–IL-15 Ab was used to treat lupus-like and EAE mice, anti–IL-15 Ab could effectively reduce memory B cells in lupus-like (Fig. 1A) and EAE (Fig. 2A) mice. IL-15 was upregulated mainly in the spleen (Figs. 1C–F, 2C–F) but not in LN, which is in accordance with previous publications suggesting that memory cells reside in the spleen (24, 25).

IL-15 is an inflammatory cytokine, and disordered IL-15 expression has been reported in patients with an array of inflammatory autoimmune diseases (26–28). Our study showed that IL-15 was increased in autoimmune diseases and EAE or lupus-like mice (Supplemental Fig. 1). A series of therapeutic agents that inhibit IL-15 action have been introduced, including anti–IL-15 neutralizing Abs (29). Further studies show that IL-15 was expressed mainly in T2B cells and BAFF suppressed IL-15 expression in T2B cells (Fig. 5). These results suggest that BAFF suppresses late immune response (memory B cells) by regulating IL-15 expression when BAFF positively activates earlier stage B cells. The basic mechanism in the immune response is a self-limited process, and, based on this, the possible harmful effects are self-limited (30). CD4⁺ T cells can also mediate pathologic immune responses, such that Th1 cells self-limit their inflammatory activity by IFN-γ (31). When it expands B cell responses, BAFF limits the late response (memory B cells) by suppressing IL-15 expression in B cells.

TACI-IgG neutralizes both BAFF and APRIL. To determine the effect of BAFF and APRIL on IL-15 expression in B cells, we used autoantigen dsDNA (precoated in 96-well plates) plus BAFF or APRIL to stimulate B cells from lupus-like mice. The results suggest that BAFF but not APRIL significantly suppresses IL-15 expression in dsDNA-stimulated B cells from lupus-like mice (Fig. 4A). Moreover, we found that IL-15 was expressed mainly in T2B cells and BAFF suppressed IL-15 expression in these cells (Fig. 5). Thus, our data suggest that BAFF but not APRIL could suppress IL-15 expression in earlier B cells, such as T2B cells. This is in accordance with previous studies suggesting that the interaction of BAFF and BAFFR is essential for viability of B cells, whereas APRIL-B cell maturation Ag and APRIL-TACI interactions are implicated more in later processes that culminate in Ab production (32–36).

A recent study has shown that Foxo3a expression does not change in B cell differentiation (37). Foxo1 can directly regulate the transcription of RAGs during B cell development (38). Our study showed that Foxo1 directly regulates the IL-15 gene transcription in T2B cells (Fig. 7). BAFF suppresses memory B cells by regulating the Foxo1–IL-15 signal pathway (Fig. 7) when it promoted B cell maturation (Fig. 6D, 6E). These results are consistent with a recent study showing that Foxo1 actively represses effectors or terminal differentiation processes to promote the memory CD8⁺ T cell development (39).

In conclusion, BAFF suppressed IL-15 expression in T2B cells by downregulating the Foxo1 signaling pathway. Once BAFF was blocked with belimumab or TACI-IgG, Foxo1 and IL-15 levels increased, leading to an increase in the number of memory B cells in autoimmune diseases. As long as the immune memory B cells are maintained, humoral response can in principle be restored from it (12). If essential components of the reactive memory are eliminated through combination therapy targeting memory B cells (40), patients will become immunologically naive. As a result, it is necessary to use BAFF inhibitors in combination with anti–IL-15 Abs to treat autoimmune diseases. This provides hints for the clinical application of combination of BAFF and IL-15 inhibitor in human autoimmune diseases.

The authors have no financial conflicts of interest.