BAFF antagonism via the BAFF receptor 3 binding site attenuates BAFF 60-mer-induced classical NF-κB signaling and metabolic reprogramming of B cells

Abstract

Human recombinant B cell activating factor (BAFF) is secreted as 3-mers, which can associate to form 60-mers in culture supernatants. However, the presence of BAFF multimers in humans is still debated and it is incompletely understood how BAFF multimers activate the B cells. Here, we demonstrate that BAFF can exist as 60-mers or higher order multimers in human plasma. In vitro, BAFF 60-mer strongly induced the transcriptome of B cells which was partly attenuated by antagonism using a soluble fragment of BAFF receptor 3. Furthermore, compared to BAFF 3-mer, BAFF 60-mer strongly induced a transient classical and prolonged alternate NF-κB signaling, glucose oxidation by both aerobic glycolysis and oxidative phosphorylation, and succinate utilization by mitochondria. BAFF antagonism selectively attenuated classical NF-κB signaling and glucose oxidation. Altogether, our results suggest critical roles of BAFF 60-mer and its BAFF receptor 3 binding site in hyperactivation of B cells.

1. Introduction

B cell-activating factor (BAFF, BLyS, or TALL-1) is a member of the tumor necrosis factor family (TNF) of cytokines that inhibits apoptosis of B cells and promotes differentiation, proliferation, antibody class switching, and antibody secretion by B cells [1, 2]. BAFF is synthesized by macrophages, monocytes, dendritic cells, neutrophils, T-cells, and some nonhematopoietic cells such as adipocytes. After synthesis, BAFF is expressed as a membrane-bound protein and then cleaved to a soluble 3-mer [3]. The secreted 3-mer can form a 60-mer by the association of 20 units of 3-mer [4]. The 60-mer is known to be the highly active form of BAFF that induces proliferation of B cells that are activated by ligation of B cell receptors [4].

BAFF can bind to three receptors: BAFF receptor 3 (BAFFR, BR3), Transmembrane activator and CAML interactor (TACI), and B-cell maturation antigen (BCMA). The interaction of BAFF multimers to BAFF receptors is well characterized in humans. The BAFF 3-mer preferentially signals via BR3, whereas the 60-mer, apart from BR3, can signal via TACI [5–8]. A proliferation inducing ligand (APRIL), a BAFF-related factor, can activate TACI and BCMA, but not the BR3 receptor. Both BAFF and BR3 are critical for the maturation of the transitional 1 (T1) B cells to T2 B cells and further maturation into B cell subsets. Therefore, BAFF or BR3 deficiency leads to the arrest of B cell maturation at the T1 stage [9]. BR3 and TACI are required for the differentiation of B2 cells to antibody-producing plasma B cells, whereas BCMA is required for the survival of plasma B cells. Memory B cells primarily express BR3 and TACI [2]. Mice genetically deficient in either BAFF or BR3 exhibit a similar immune phenotype, such as lower numbers of mature B cells, a lower number of B2 cells in the spleen, lower expression levels of CD21 and CD23 on B cells (B2 cell maturation markers), and reduced antigen-specific antibody response [10, 11]. BAFF can regulate CD21 and CD23 expression independent of its B cell survival activity supporting differential roles of BAFF multimers and BAFF receptors [12]. Peripheral B cell populations and antibody responses are attenuated in the BAFF knockout mice which are reversed by injection of recombinant human BAFF 3-mer [3]. However, injection of the human BAFF 60-mer increases the number of B2 cells compared to WT mice.

NF-κB signaling pathway is crucial for B cell development and function [13]. Activation of NF-κB2 in the alternative/noncanonical NF-κB signaling in B cells by BAFF is well documented [14]. In this pathway, BAFF exposure leads to the degradation of TNF receptor-associated factor 3 (TRAF3), a negative regulator of alternate signaling. This stabilizes NF-κB inducing kinase (NIK), which phosphorylates IκB kinase α (IKKα), which in turn phosphorylates p100 resulting in the degradation of p100 to p52 [15]. p52 dimerizes with RelB and translocates to the nucleus to activate genes involved in B cell survival [14, 16]. Some reports also indicate that BAFF can activate both the classical and alternate NF-κB signaling pathways [17, 18]. In the classical/canonical NF-κB signaling, IκB kinase γ activates the IκB kinase complex by phosphorylating IKKβ, which in turn phosphorylates inhibitor of NF-κB resulting in phosphorylation of p65, dimerization of phospho-p65 with p50, nuclear translocation of the activated dimer, and activation of genes involved in inflammation. So far, it is not clearly understood if BAFF multimers can differentially regulate classical and alternate NK-κB signaling.

The metabolic profile of lymphocytes plays an essential role in the fate of the cell. In activated T cells, dendritic cells, and macrophages, glycolysis predominates over oxidative metabolism resulting in a reliance on aerobic glycolysis, that is oxidation of glucose to lactate [19–21]. Antigenic stimulation of B cells leads to an increase in both aerobic glycolysis and oxidative phosphorylation (OXPHOS) [22]. Choice of one or both pathways significantly affect B cell function [23, 24]. B cells from BAFF transgenic mice favor aerobic glycolysis, which is essential for proliferation and antibody production [23]. It is unknown if BAFF multimers affect the metabolic reprogramming of B cells and if such metabolic reprogramming is regulated by BAFF-induced NF-κB signaling.

A critical role of BAFF in abnormal activation of B cells is identified in non-Hodgkin lymphoma, autoimmune and cardiovascular diseases, and diabetes. Multiple BAFF and BR3 antagonists such as Belimumab and Ianalumab (VAY736), respectively, which suppress abnormal activation of B cells in autoimmune diseases, are approved for treatment or in clinical trials [25, 26]. However, it is incompletely understood whether BAFF 3-mer and 60-mer differentially impact B cell activation and if BAFF 60-mer is merely a highly active form of BAFF. Furthermore, it is incompletely understood how BAFF abnormally activates the B cells.

Presence of large multimers of BAFF, such as BAFF 60-mer, in humans is still questionable [6, 27]. Using size exclusion chromatography (SEC), we identified 3-mer, 60-mer, and higher order multimers of BAFF in human plasma. Since mouse models are essential in the process of drug discovery and multimer formation is well-charactered for human BAFF, we examined the roles of human BAFF multimers on the activation of murine B cells. We first determined the avidity of interaction of human BAFF multimers with murine BAFF receptors BR3, TACI, and BCMA. Next, we performed global transcriptomics of B cells which revealed BAFF 60-mer hyperactivates B cells whereas BAFF 3-mer also activates B cells, but at a lower intensity than 60-mer. We further identified critical roles of BAFF multimers in NF-κB signaling, mitochondrial function, metabolic reprogramming, and surface expression of CD23, a marker for differentiation of transitional B cell to mature B cells, and surface expression of MHC II, a marker for B cell activation. Since BAFF 3-mer and 60-mer shares a common receptor BR3, we examined if BAFF-mediated B cell activation can be attenuated by blocking the BR3 binding site on BAFF with a soluble fragment of BR3 (mBaffR-Fc). Our studies revealed that the BAFF 60-mer uniquely reprograms B cells for both cellular and metabolic activation. Furthermore, we identified a crucial role of canonical NF-κB signaling relayed via BR3 binding site on BAFF 60-mer in B cell hyperactivation.

2. Materials and methods

2.1. Mice

Seven-week-old male C57BL/6J (stock # 000664) and BAFF receptor 3 knockout mice (stock# 007212) were obtained from Jackson Laboratory (US). All mice were given water and normal mouse diet Prolab IsoPro RMH 3000 5P76 ad libitum. All protocols were approved by East Carolina University and University of Virginia Animal Care and Use Committee.

2.2. Fractionation of culture supernatant and plasma sample using size exclusion chromatography

pReceiver-M68 containing BM40 secretion signal and soluble human BAFF (NM_006573, amino acid 134-285) was transfected to HEK293T cells using Invitrogen Lipofectamine reagent (94756, MA, USA) and cultured with Opti-MEM I Reduced-Serum Medium. After 3 days of transfection, culture supernatant was collected, filtered with 0.2 μm filter, and 500 μl of the supernatant was fractionated using the SEC column Superose 6 Increase 10/300 GL column connected to ACTA pure 25 chromatography system. Typically, the column was equilibrated using phosphate buffered saline (PBS), pH 7.4, and loaded proteins were eluted as 1 ml fractions in the same buffer. The 1 ml fractions were acid-precipitated, analyzed on an 8-16% Tris-Glycine SDS-PAGE under reducing conditions (with 2-Mercaptoethanol) and Western blotting using an anti-BAFF antibody (PA1-30556, ThermoFisher Scientific, MA, USA).

Plasma samples from healthy humans were purchased from Innovative Research (MI, USA). The selection of plasma samples was based on age (40-50 years), body mass index (19-26), and cholesterol (160-250 mg/dl) of female donors. Typically, 50 μl of plasma was diluted to 800 μl in PBS, 500 μl of which was fractionated similarly as described for culture supernatant. The presence of BAFF in the fractions collected from SEC was determined by a BAFF ELISA (AG-45B-0001-KI01, Adipogen, CA, USA). For the ELISA, the pH of the fractions was adjusted to 5.5 with 0.5 M 2-(N-Morpholino) ethanesulfonic acid, 4-Morpholineethanesulfonic acid monohydrate for maximum detection of BAFF as reported before [27]. To detect BAFF using Western blot, albumin and immunoglobulins were depleted from 400 ul of plasma using Thermo Scientific^™ High Select^™ HSA/Immunoglobulin Depletion Midi Spin Columns, fractionated using the Superose 6 Increase 10/300 GL column, proteins from 1 ml fraction were acid precipitated, and 1/4^th of the samples was separated using SDS-PAGE and the presence of BAFF in the fractions were determined by Western blotting using an anti-BAFF antibody. Untagged soluble human recombinant BAFF (Cat# 310-13, PeproTech, NJ, USA) was used to identify the BAFF band. Molecular weights of proteins were determined by analyzing the elution profile of a mixture of markers (Gel Filtration LMW and HMW Calibration Kits, Cytiva, USA) in 500 μl of PBS containing Thyroglobulin (Mr 669 000), Ferritin (Mr 440 000), Aldolase (Mr 158 000), Conalbumin (Mr 75 000), Carbonic anhydrase (Mr 29 000), Ribonuclease A (Mr 13 700) and Apoprotein (Mr 6 500).

2.3. Protein-protein interaction assays by surface plasmon resonance (SPR)

Recombinant BAFF 3-mer (Flag-tagged BAFF amino acid 134-285, AG-40B-0016) and 60-mer (histidine-tagged BAFF amino acid 134-285, AG-40B-0112) were obtained from Adipogen, USA. mBAFFR-Fc was obtained from Biogen Idec, US. TACI-Fc (1041-TC-050) and BCMA-Fc (593-BC-050) were obtained from Novus Biologicals (CO, USA), SPR experiments were performed on a Biacore T200 instrument at 25° C using a flowrate of 30 μL min-1 in a running buffer of HBS-T (10 mM HEPES pH 7.3, 140 mM NaCl, 0.005% Tween 20). Proteins were immobilized on a CMD200 biosensor chip (Xantec Bioanalytics) using ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC)/N-hydroxysuccinimide/(NHS) coupling chemistry. Immobilization densities, reported in resonance units (RU), were as follows: mBaffR-Fc (814 RU), TACI-Fc (584 RU), BMCA-Fc (378 RU), and anti-BAFF antibody (2,760 RU; 750 RU; 610 RU). A reference surface was created by EDC/NHS activation followed by injection of 1M ethanolamine (pH 8.5). A serial dilution of BAFF 60-mer (0.2, 0.64, 2, 6.4, 20 nM, each 60-mer was considered as 20x 3-mer BAFF molecule) or BAFF 3-mer (1, 3.2, 10, 32, 100 nM, each 3-mer was considered as a single BAFF molecule) was injected over each surface in a single cycle. Each sample was injected for three minutes, and dissociation was monitored for 30 minutes following the final injection. Surfaces were regenerated to baseline by three two-minute injections of 10 mM glycine (pH 2.2), and 2.5 M NaCl. Each injection series was performed in triplicate. The resulting reference-corrected sensorgrams were blank subtracted using an injection of HBS-T and fit to a steady-state model of binding using T200 Evaluation software. All figures were prepared using Graphpad Prism 8.0.

2.4. Murine B cell isolation

Murin B cells were isolated using the Pan B Cell Isolation Kit II (Miltenyi Biotec, 130-104-443) or EasySep^™ Mouse Pan-B Cell Isolation Kit (StemCell, 19844) that labels the non-B cells with a cocktail of biotinylated CD4, CD11c, CD49b, CD90.2, Gr-1, and Ter-119 antibodies and deplete from the splenocytes to provide highly pure B cells. Following isolation, the B cells were suspended in a complete RPMI medium, that is RPMI 1640 Medium containing 10% heat-inactivated fetal bovine serum, 1x antibiotic-antimycotic, 20 mM HEPES Buffer solution, 1x GlutaMAX-I, 1mM Sodium Pyruvate, 1x MEM Non-Essential Amino Acids Solution, and 55 μM 2-Mercaptoeathanol from Gibco. B cell density was determined using Countess II FL and counting chamber slides from ThermoFisher. The purity of the B cells was found to be >96% as determined by cell surface expression of CD19 (PerCP-Cy5.5 anti-CD19), and B220 (Alexa Fluor 488 anti-B220) using flow cytometry.

2.5. Flow Cytometry Quantification

The cells were blocked with purified anti-mouse CD16/32 and stained with the following fluorescent conjugated antibodies for 30 minutes on ice: PerCP-Cy5.5 anti-CD19, APC anti-CD23, and Brilliant Violet 785 anti-MHC II from BioLegend, US. Invitrogen^™ DAPI (4’,6-Diamidino-2-Phenylindole, Dihydrochloride) (#D1306) was used for staining the dead cells. After the staining, the samples were run on the Flow cytometer machine Becton Dickinson LSRFortessa, USA or Cytek Biosciences Cytek Aurora, USA. Data analysis and quantification were performed using FlowJo v10.6.2.

2.6. Quantification of sCD23

B cells (1 million/sample) were left untreated (UT) or treated with 5 nM BAFF 3-mer or 60-mer for 20 hours, and the level of soluble CD23 (sCD23) in the culture supernatant was determined using Mouse CD23 ELISA Kit (ab206984, Abcam, USA).

2.7. RNA sequencing, Real-time RT-PCR, and data analysis

Mouse splenic total B cells (6 million cells/sample) were pre-treated with the 30 nM of control Ab (IgG1κ, BioXCell, #BE0083, USA) or mBaffR-Fc (obtained from Biogen Idec. USA) for 1 hour, and then treated with 5 nM of BAFF 3-mer (AG-40B-0016, AdipoGen, US), 60-mer (AG-40B-0112, AdipoGen) or left untreated for the next 4 hours. Endotoxin content of the recombinant BAFF proteins was <0.01EU/μg purified protein (LAL test; Lonza). Each treatment was performed in triplicate. Total RNA was extracted from the samples using QIAGEN RNeasy Mini Kit and was sent to Novogene Corporation (Wilmington, DE, USA) to determine the gene expression levels. The RNA integrity was determined using Agilent 2100. Sequence libraries were prepared from the RNA samples and sequenced using an Illumina Hiseq 4000. The original raw data were transformed to Sequenced Reads by base calling and the data were recorded in a FASTQ file. The percentage of bases whose correct base recognition rates are greater than 99% in total bases (Q20) was >97%. TopHat2 was used for the algorithm for mapping sequences. Total number of reads that can be mapped to the reference genome was >91%. The Pearson coefficient, R², between two samples of a treatment was >0.92 suggesting very low variability. FPKM (expected number of Fragments Per Kilobase of transcript sequence per Million base pairs sequenced) was used for estimating gene expression levels. DESeq was used for differential gene expression analysis from read counts obtained from gene expression level analysis, and cluster heat maps were generated. Using the FPKM values, log2 fold change (BAFF 3-mer vs untreated and 60-mer vs untreated) values were calculated and analyzed using QIAGEN Ingenuity Pathway Analysis software (IPA). Alternatively, following RNA extraction, B cells were removed from the culture by washing 3 times with PBS and treated with Qiagen Buffer RLT for RNA extraction. cDNA was synthesized and real-time PCR was performed on a CFX Connect^™ Real-Time PCR Detection System as described previously [28, 29]. The abundance of Gapdh mRNA was used for relative quantification of Apoe, S1pr1, and Nfkbia mRNAs in unstimulated and stimulated cells. Sequences of the primers used for the PCR were: Apoe forward: CAGACCCTGGAGGCTAAGGA and reverse: CTCCATCAGTGCCGTCAGTT; S1pr1 forward: CTCAGGGAACTTTGCGAGTGA and reverse: GGGTGGTATTTCTCCAGGCAA; Nfkbia forward: TAGCAGTCTTGACGCAGACC and reverse: AGGTAAGCTGGTAGGGGGAG; Gapdh forward: TGTGTCCGTCGTGGATCTGA and reverse: TTGCTGTTGAAGTCGCAGGAG.

2.8. RNA sequencing data availability

The RNA-sequencing data is available from the Gene Expression Omnibus database (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE161072.

2.9. Western Blotting for NF-κB signaling

B cells were isolated from mouse spleen and 6 million B cells/sample were treated with 5 nM BAFF 3-mer, 60-mer, or left untreated for 3 hours or 24 hours. Alternatively, B cells were treated with 30 nM of the control antibody or the mBaffR-Fc and BAFF proteins. After the treatments, the cells were washed, and total cellular protein was extracted using 1x RIPA buffer containing protease inhibitor cocktail and phosphatase inhibitors (SigmaAldrich). The extracted protein was analyzed using 4-12% Tris-Glycine SDS-PAGE. The primary antibodies used are Phospho-NF-κB p65 (Ser536) Antibody (#3031), NF-κB p65 (D14E12) XP^® Rabbit mAb (#8242), NF-κB2 p100/p52 Antibody (#4882), and β-Actin Antibody (#4967) from Cell Signaling Technology, OxPhos Rodent WB Antibody Cocktail (#45-8099) from ThermoFisher Scientific, the IRDye-conjugated secondary antibodies were from LI-COR, and the blots were detected, and band intensities were quantified using LI-COR’s Image Studio software version 5.2.5.

2.10. An ELISA method to identify mBaffR, TACI, and BCMA binding to BAFF 60mer

Murine TACI-Fc and BCMA-Fc were biotinylated using Thermo Scientific^™ EZ-Link^™ Sulfo-NHS-Biotin. Excess reactive biotin was removed using Thermo Scientific^™ Pierce^™ Protein Concentrators PES 10K MWCO. A 96-well Nunc MaxiSorp^™ flat-bottom (ThermoFisher Scientific, Cat# 44-2404-21) plate was coated with 100 μl of 30 nM mBaffR, blocked with 2% BSA in PBS, added 100 μl of 5 nM BAFF 60mer for 1 h, washed, and incubated with 100 μl of 30 nM biotinylated TACI or BCMA (30 nM) for 1 h. As a negative control, biotinylated TACI or BCMA was added to mBaffR. Unbound proteins were washed and bound biotinylated TACI or BCMA was identified by avidin/biotin-based peroxidase (VECTASTAIN^® Elite^® ABC-HRP Kit, Peroxidase, Cat# PK-6100, Vector Laboratories)) and Pierce^™ TMB Substrate Kit (ThermoFisher Scientific, Cat# 34021) after addition of 1N sulfuric acid at OD 450 nm in an ELISA reader.

2.11. Seahorse assays

The Seahorse assays were performed using a Seahorse XFe24 Extracellular Flux Analyzer as described previously [30]. A day before the assay, XFe24 well microplates were coated with Poly-L-lysine solution (P4832, Sigma-Aldrich, USA) and seeded with B cells (1 million/sample) isolated from murine spleen and treated with BAFF and other reagents as listed in the figure legends. One hour before the assay, the culture medium was replaced with Dulbecco’s Modified Eagle’s Medium high glucose (1280017, ThermoFisher) for Mitochondrial Stress Test or with Dulbecco’s Modified Eagle’s Medium with no glucose (D5030, ThermoFisher) for the Glycolysis Stress Test. The tests were performed with minor modifications from methods provided by Agilent.com. In the Mitochondrial Stress Test, three drugs were serially injected: 1^st injection was Oligomycin A (10 μM, Sigma-Aldrich, #75351), 2^nd injection was BAM15 (20 μM, Cayman Chemicals, #17811), and the 3^rd injection was Antimycin A (100 μM, Sigma-Aldrich, #A8674) + Rotenone (10 μM, Sigma-Aldrich, #557368). Oxygen consumption rate (OCR) was measured three times before the injection of Oligomycin A and after each injection. In the Glycolysis Stress Test, 1^st injection was D-(+)-Glucose (20 mM, Sigma-Aldrich, G7021), 2^nd injection was Oligomycin A (10 μM), and the 3^rd was 2-Deoxy-D-glucose (80 mM, Sigma-Aldrich, #D6134). Extracellular acidification rate (ECAR) was measured three times before the injection of Glucose and after each injection. The fuel flexibility test was a Mitochondrial Stress Test after injection with oxidation pathway inhibitors UK5099 (2 μM, inhibits glucose oxidation, Cayman Chemicals, #16980), BPTES (3 μM, inhibits glutamine oxidation, Cayman Chemicals, #19284), or Etomoxir (4 μM, inhibits oxidation of long chain fatty acids, Cayman Chemicals, #11969).

For experiments involving mBaffR-Fc, mouse splenic total B cells were treated with the 30 nM of control Ab (IgG1κ, BioXCell, #BE0083, USA) or mBaffR-Fc (obtained from Biogen Idec. USA) and 5 nM of BAFF 3-mer, BAFF 60-mer or left untreated for the next 20 hours, and a mitochondrial or glycolysis stress test was performed. For experiments involving mBaffR-Fc and NF-κB inhibitor BMS-345541, dimethyl sulfoxide (DMSO) was added to mBaffR-Fc or control wells (0.1% final concentration)

2.12. Measurement of mitochondrial density by microscopy

After treating B cells with BAFF proteins in a complete culture medium for 24 hours on Chambered Cover glasses, the culture medium was removed, and the cells were treated with 50 nM Invitrogen^™ MitoTracker^™ Green FM (Fisher Scientific, M7514) in RPMI 1640 media for 20 minutes at 37 °C. The cells were washed with the RPMI media and covered with ProLong Live mounting media (ThermoFisher Scientific) containing nuclear counterstain Hoechst 33342. Confocal images to a depth of 4.5 μm (at an interval of 0.45 μm) were acquired on a ZEISS LSM 700 microscope with laser lines 405 nm, 488 nm, and 555 nm. 3-D images followed by single plane top view images were generated using the Zen black software. MitoTracker stained area/cell was calculated using the ImageJ software. Authors quantifying the images were blinded to the treatment strategy.

2.13. Measurement of mitochondrial density and membrane potential, and cellular reactive oxygen species by flow cytometry

Purified B cells (0.5 million/sample) were treated with 5 nM of BAFF 3-mer or 60-mer proteins in a complete culture medium in 96 well plates for 20 hours, the cells were stained with 200 nM MitoTracker^™ Green FM (Fisher Scientific, M7514), 100 nM Tetramethylrhodamine, Ethyl Ester, Perchlorate (TMRE) (Fisher Scientific, T669), and 2.5 μM Invitrogen^™ CellROX^™ Green Reagent (Fisher Scientific, C10444) in RPMI 1640 media for 20 minutes at 37 °C. LPS (1 μg/ml) was used as a positive control for B cell activation and 20 μM BAM15 was used as a negative control for dissipation of mitochondrial membrane potential detected by TMRE. The cells were pre-treated for 1 hour with a vehicle (complete culture medium with 0.1 % Dimethyl sulfoxide or various inhibitors of NF-κB signaling in 0.1% Dimethyl sulfoxide: BMS-345541 (5 μM, #16667, Cayman Chemicals, US), BI 605906 (5 uM, #5300, Tocris, US), and IMD 0354 (1 uM, #2611, Tocris) before treating with BAFF proteins. After the staining, the samples were run on the Flow cytometer machine Becton Dickinson LSRFortessa (equipped with laser lines 355nm, 405nm, 488nm, and 640nm). Dead cells were discriminated by using DAPI. Data analysis and quantification were performed using FlowJo v10.6.2.

2.14. Mitochondrial substrate metabolism assays

Mitochondrial substrate metabolism was assessed by using MitoPlate S-1 (Biolog, USA). In brief, 0.5 million B cells were treated with the vehicle (DMSO 1 μl/ml), LPS (1 μg/ml with DMSO 1 μl/ml), BMS-345541 (5 μM, final DMSO concentration 0.1%), and mBaffR-Fc (30 nM with DMSO 0.1%) for 1 hour at 37°C in three different experimental setups. Then, the cells were treated with 5 nM of BAFF 3-mer, BAFF 60-mer, or left untreated for 20 hours. One hour before the assay, 30 μl of assay mix supplemented with 2X mitochondrial assay solution (MAS), 6X redox dye, and 24X saponin was dispensed into the 96 wells of MitoPlate S1 and incubated at 37°C. The treated cells were washed with 1X MAS medium to remove the residual media. Then, 30 μl cell suspension in 1X MAS was transferred to MitoPlate S-1 and incubated for 18 hours at 37°C in a humidified chamber. Color formation in MitoPlate was measured on Varioskan Lux multimode microplate reader (Thermo Scientific) using OD 590 nm [31]. Before the assay, saponin concentration was optimized for better permeabilization of the B cell plasma membrane without the loss of mitochondrial function.

2.15. Glucose uptake assay

Glucose uptake by the B cells was examined by using the Glucose Uptake-Glo Assay kit (Promega, USA) according to the manufacturer’s instructions. Briefly, in a 96 well plate, 0.5 × 10⁶ B cells were treated with the vehicle (DMSO 1μl/ml), LPS (1 μg/ml with 0.1% DMSO), BMS-345541 (5 μM, final DMSO concentration 0.1%), and mBaffR-Fc (30 nM with 0.1% DMSO) for 1 hour at 37°C. Then, the cells were treated with 5 nM of BAFF 3-mer, BAFF 60-mer, or left untreated for 20 hours. Next, the cells were washed with PBS and incubated with 50 μl of 1 mM 2-deoxyglucose (2-DG) for 30 minutes followed by the addition of the stop buffer and neutralization buffer. Subsequently 100 μl 2-deoxyglucose-6-phosphate detection reagent was added and mixed properly by shaking. Cells were incubated for 1 hour at room temperature. Luminescence was recorded with one-second integration on a Varioskan Lux multimode microplate reader (Thermo Scientific, USA).

2.16. Statistical Analysis

The data were analyzed using GraphPad Prism 8 and Excel and presented as means + SEM or means + SD. In vitro assays were performed in quadruplicates and repeated at least two times. The difference between mean values of the two groups was determined using a t-test. D’Agostino-Pearson normality test was performed on each group. If the P value is not significant (>0.05), a parametric test was used, and if the P value is significant (<0.05), a non-parametric t-test (Mann-Whitney test) to determine significant differences between the groups. Means of multiple groups were compared using one-way or two-way ANOVA followed by Dunnett’s or Tukey’s multiple comparisons test. Differences between the groups were considered significant when the p-value is <0.05. p-values >0.05 were indicated in the graphs.

3. Results

3.1. Human plasma contains higher order BAFF multimers

A rigor method to determine if BAFF 60-mer is present in human plasma would be to fractionate the plasma proteins based on molecular size by SEC and find out which fractions contain BAFF. Before using the fractionation method, we first fractionated culture supernatant from HEK293T cells which were transfected with a plasmid containing human soluble BAFF (amino acid 134-285) with BM40 secretion signal. Western blot of the fractions against anti-BAFF antibody revealed two major multimers of BAFF: the 60-mer that was eluted at 13 ml (669 kDa) and the 3-mer eluted at 17 ml (75 kDa) (Fig. 1A and B). This confirms the previously reported results that BAFF exists as 3-mers and 60-mers in culture supernatants [6]. Apart from these two major multimers, there are other multimers suggesting maintenance of homeostasis in BAFF multimerization.

Figure 1: Presence of higher order BAFF multimers and BAFF 3-mers in human plasma.

A, SEC elution profile of molecular weight markers in Superose 6 Increase 10/300 GL column. B, Western blot against anti-BAFF antibody of SEC fractions of culture supernatant from HEK293T overexpressing human soluble BAFF with BM40 secretion signal. C, ELISA to detect BAFF in SEC fractions of plasma collected from two doners. D, Western blot against anti-BAFF antibody of SEC fractions of human plasma. ‘*’ indicates a non-specific band.

In human plasma, low abundance of BAFF (2-6 ng/ml [32]) and high abundance of albumin and immunoglobulin possess major challenges in the detection of BAFF using Western blot. Therefore, first, we used a sensitive ELISA assay to detect BAFF in SEC fractions of plasma. Similar to the culture supernatants, BAFF 3-mers were detected at 17 ml fractions. Multiple very high-molecular multimers of BAFF were detected at 8-11 ml fractions (Fig. 1C) suggesting the presence of higher order multimers of BAFF in humans. However, the level of higher order multimers was disproportionate among the four plasma samples tested in this study. Next, we depleted albumin and immunoglobulin from plasma before fractionating on the size exclusion chromatography. Following fractionation, the proteins in the samples were precipitated, and separated on an SDS-PAGE, and presence of BAFF in the fractions was determined by Western blot. Although this procedure did not completely remove albumin and immunoglobulin, BAFF was identified at fractions of 8-12 ml suggesting the presence of higher order multimers of BAFF.

3.2. Human BAFF multimers bind to murine BAFF receptors and differentially induce the expression of CD23 and MHC II on B cells

Various methods have been used to study the BAFF-BAFF receptor interaction (SPR, enzyme-linked immunosorbent assay, cell-based and flow cytometry assays) and different binding avidities are documented because of the presence of multivalent forms and the methods of immobilization of the ligands and the receptors (avidity effects) [7, 33–35]. There is a general consensus that human BAFF 60-mer is the highly active form of BAFF, and activates BR3 and TACI receptors, whereas BAFF 3-mer, preferentially activates the BR3 receptor [6]. Murine splenic B cells are known to express three BAFF receptors: BR3, TACI, and BCMA [36]. We determined the binding affinities of human BAFF multimers with Fc-conjugated murine soluble dimeric BAFF receptors mBaffR-Fc (BR3), mTACI-Fc, and mBCMA-Fc or with an anti-BAFF antibody (clone Sandy-2) which were immobilized on SPR biosensor chips (Suppl. Fig. 1A). The avidity of BAFF 3-mer was determined as a single molecule and BAFF 60-mer as 20 molecules (representing 20 copies of BAFF 3-mer). As expected, BAFF 60-mer bound to murine BR3, TACI, and BCMA [7]. However, the BAFF 3-mer bound only to BR3, not to TACI or BCMA. The anti-BAFF antibody bound to both the BAFF 60-mer and 3-mer.

To determine the optimal concentration of BAFF to activate B cells, we isolated total splenic B cells from C57BL/6 (wild-type, or WT) mice and treated the cells with various concentrations of BAFF 3-mer or 60-mer and then assessed the surface expression of CD23 and MHC II (Suppl. Fig. 1B). Concentration of BAFF multimers was expressed as nM of BAFF monomer. At 0.05 nM, BAFF 60-mer significantly increased CD23 and MHC II expression, which reached a plateau at 0.5 to 5 nM. At the tested concentrations (0.05 nM to 50 nM), compared to the BAFF 3-mer, BAFF 60-mer strongly induced CD23 and MHC II expression. Furthermore, compared to the BAFF 3-mer, BAFF 60-mer at 5 nM also significantly increased the production of soluble CD23 after treatment for 20 hours (Suppl. Fig. 1C). Therefore, 5 nM of the BAFF 3-mer and the 60-mer were selected for further studies. At this concentration, BAFF 3-mer marginally increased CD23 (also shown in [3]) and MHC II expression (Suppl. Fig. 1B and E). Therefore, the B cells were treated with IL-4 and IL-6, along with BAFF multimers as described before [6]. Compared to the control and the 3-mer-treated cells, BAFF 60-mer significantly increased CD23 expression, whereas both the 3-mer and 60-mer marginally increased MHC II expression (Suppl. Fig. 1D). Altogether, these results suggest differential roles of BAFF 3-mer and 60-mer in the expression of CD23 and MHC II.

BR3 is a critical receptor for BAFF as both BAFF 3-mer and 60-mer bind to and also activate the receptor [4]. To determine if BAFF antagonism via the BR3 binding site on BAFF affects CD23 and MHC II expression, murine splenic B cells were treated with either the BAFF 3-mer or the 60-mer at 5 nM, in the presence of the control Ab or the mBaffR-Fc at 6-fold molar excess, that is 30 nM. As a control, the B cells were treated with BAFF in presence of an anti-BAFF Ab (clone Sandy-2). After 20 hours, the surface expression of CD23 and MHC II was examined via flow cytometry. At 30 nM, mBaffR-Fc did not affect the 60-mer-induced CD23 expression but attenuated the expression of MHC II (Suppl. Fig. 1E). At the same concentration, the anti-BAFF Ab attenuated the 60-mer-induced expression of CD23 and MHC II. This suggests a critical role of BR3 binding site on BAFF on MHC II, but not CD23 expression of B cells.

3.3. mBaffR-Fc partly attenuates BAFF 60-mer-induced B cell transcriptional hyperactivation

To further determine the critical role of the BR3 binding site on BAFF in the activation of B cells, we performed a global transcriptomics study by RNA sequencing (RNA-Seq). In this study, B cells purified from mouse spleen were treated with 5 nM of BAFF 3-mer, 60-mer for 4 hours or left untreated either in the presence of 30 nM of a control antibody (IgG1) or mBaffR-Fc. Total RNA was isolated and RNA sequencing was performed. Using the Fragments Per Kilobase of transcript per Million (FPKM) values, relative gene expression (3-mer or 60-mer-treated relative to untreated) was calculated as log2 fold change and analyzed for enrichment of genes in ‘Diseases and Biological Functions’ using QIAGEN Ingenuity Pathway Analysis (IPA) software. B cell activation is tightly linked to NF-κB signaling [37]. Therefore, we analyzed the relative expression of individual genes involved in the activation of B cells (Fig. 2A), NF-κB signaling (Fig. 2B), and death of B cells (Fig. 2C). Increased or decreased expression of some of the genes identified in the RNA-seq data was verified using the FPKM values (Fig. 2D) and real-time RT-PCR (Fig. 2E). Compared to the untreated and the 3-mer-treated cells, cells treated with the 60-mer upregulated synthesis of genes involved in B cell activation (Cd21, Cd23, Il-9r, Cd44, Cd40) and NF-κB signaling (NFkb2, Nfkbie, Nfkbia, and Traf3), and downregulated genes involved in the death of B cells (Casp8 and Casp9). Furthermore, expression levels of CD22 and PirB, the negative co-receptors for B cell receptor activation [38], were lower, suggesting the 60-mer-treated B cells were in an activated state. Sphingosine-1-phosphate receptor 1 (S1pr1) expression is essential for the egress of B cells from secondary lymphoid organs [39], and BAFF 60-mer significantly decreased S1pr1 gene expression (Fig. 2C, D and E). Analysis of individual genes reveals that BAFF 3-mer promoted expression of genes involved in activation of B cells, NF-κB signaling, and survival of B cells similar to the BAFF 60-mer, however, at a much lower intensity (Fig. 2A–C). IPA compares user data sets with available literature and computes a statistical quantity called the activation z-score that provides information on activation states of biological functions. The z-score analysis revealed significant positive z-scores for activation, proliferation, and expansion of B cells, and negative z-scores for apoptosis and cell death in response to BAFF 60-mer (Suppl. Fig. 2). However, the modest transcriptional activation of B cells by 3-mer (Fig. 2A–E) was scored ‘0’ by IPA. These results identify BAFF 60-mer as the critical player of B cell activation. Interestingly, the expression of most of the genes involved in B cell activation and NF-κB signaling was partly reduced or not affected in 60-mer-treated B cells in the presence of mBaffR-Fc. The z-score analysis of these cells revealed a negative score for activation of B cells (Suppl. Fig. 2). Altogether, these results suggest that BAFF 60-mer polarizes B cells to a highly activated phenotype, which is partly mediated by the BR3 binding site on BAFF.

Figure 2: BAFF 60-mer induces expression of genes required for B cell activation, which are partly suppressed by mBaffR-Fc.

Transcriptomics studies on murine splenic B cells were performed after 4 hours of treatment with 5 nM of BAFF 3-mer, 60-mer or left untreated (UT) in presence of 30 nM of the control Ab or mBaffR-Fc. A, B and C, Genes involved in the ‘Activation of B cells’, ‘NF-κB signaling’ and ‘Cell death of B cells’ respectively, were sorted and genes with FPKM ≥10 are displayed as log2 (fold change). D, FPKM values from RNA sequencing of selected genes are shown. Values are expressed as means + SEM, n=3/group. “**” and “***” indicate p<0.01 and 0.001, respectively determined by a Two-way ANOVA with Tukey’s multiple comparisons test. E, Messenger RNA expression of Apoe, S1pr1, and Nfkbia genes in B cells after 5 hours of treatment with 5 nM of BAFF 3-mer, 60-mer or left untreated. Values are expressed as means + SEM, n=4/group. “*” indicates p<0.05 determined by Two-way ANOVA with Bonferroni’s multiple comparisons test.

3.4. BAFF 60-mer activates both classical and alternate NF-κB signaling, and mBaffR-Fc blocks only the classical, but not the alternate signaling

The transcriptomics study revealed induction of NF-κB signaling by BAFF 3-mer and 60-mer; however, it was not clear if the BAFF multimers differentially activate the classical and alternate signaling pathways. Therefore, we analyzed p65 phosphorylation at Serine 536 (a part of classical) and degradation of p100 to p52 (a part of alternate) using Western blot. Total splenic B cells from C57BL/6 mice were treated with 5 nM of the 3-mer and 60-mer for 3 hours or 24 hours or left untreated and analyzed for NF-κB signaling (Fig. 3A). After 3 hours of treatment, the 60-mer significantly induced the phosphorylation p65, which was diminished by 24 hours, whereas the 3-mer did not affect p65 phosphorylation. BAFF 3-mer marginally induced p100 degradation at the 3-hour and 24-hour time points; however, the 60-mer response was stronger than that of the 3-mer and was statistically significant.

Figure 3: BAFF 60-mer induces both the classical and alternate NF-κB signaling and mBaffR-Fc attenuates only the classical signaling.

Murine splenic B cells were treated with 5 nM of BAFF 3-mer, 60-mer or left untreated (UT) for 3 h or 24 hours in A, and for 3 hours in B. In B, the cells were treated with BAFF in the presence of 30 nM of a control antibody or mBaffR-Fc. Following the treatments, the total cell extracts were resolved on a 4-15% SDS-PAGE, and NF-κB1 and NF-κB2 signaling were determined by Western blotting. p65 phosphorylation and p100 degradation were quantified using LI-COR software. Values are expressed as means + SEM and n=3. “*”, “**” and “***” indicate p<0.05, 0.01 and 0.001, respectively determined by a Two-way ANOVA with Tukey’s multiple comparisons test. C, ELISA plate was coated with 100 ul of mBaffR (30 nM), blocked, added BAFF 60-mer (5 nM), washed, and biotinylated TACI or BCMA (30 nM) was added. As a negative control, biotinylated TACI or BCMA was added to mBaffR. Bound biotinylated TACI or BCMA was identified by avidin/biotin-based peroxidase and TMB substrate.

Most of the BAFF signaling studies have centered on alternate NF-κB signaling [40–42] and it is unknown whether the BAFF 3-mer and 60-mer differentially signal through the BR3 receptor to induce the NF-κB signaling pathways. Therefore, we blocked the BR3 binding site on BAFF using the mBaffR-Fc and examined NF-κB signaling via Western blot. Interestingly, mBaffR-Fc did not affect p100 degradation induced by the 60-mer in 3 hours of treatment (Fig. 3B). However, mBaffR-Fc attenuated the BAFF 60-mer-induced phosphorylation of p65. To determine if mBaffR-Fc blocked the binding of TACI or BCMA to BAFF 60-mer, we performed an ELISA assay. First, mBaffR-Fc was coated on an ELISA plate, blocked, added BAFF 60-mer (5 nM), washed, and added biotinylated murine TACI-Fc or BCMA-Fc. As a negative control, biotinylated TACI-Fc or BCMA-Fc was added to mBaffR captured wells. The bound biotinylated receptors were identified by avidin/biotin-based peroxidase and TMB (3,3’,5,5’-Tetramethylbenzidine) substrate. Signals were detected only from the wells that received BAFF 60-mer (Fig. 3C). This suggests that murine TACI or BCMA can bind to BAFF 60-mer:mBaffR-Fc complexes. Altogether, the BR3 binding site on BAFF 60-mer is critical for BAFF-medicated canonical signaling in B cells.

3.5. BAFF 60-mer induces both aerobic glycolysis and mitochondrial OXPHOS, which are attenuated by mBaffR-Fc

Energy metabolism is critical in B cell activation. Therefore, we examined how BAFF 3-mer and 60-mer affect energy metabolism in B cells using a Seahorse XFe24 Extracellular Flux Analyzer. Total B cells isolated from spleens of WT mice were left untreated or treated with various concentrations of the BAFF 3-mer, the BAFF 60-mer, anti-CD40 monoclonal antibody, or lipopolysaccharide (LPS) for 20 hours. Anti-CD40 antibody and LPS are positive controls for B cell activation [43]. Following the incubation, a mitochondrial stress test that quantifies oxidative phosphorylation (OXPHOS)/mitochondrial respiration under stressed conditions or a glycolysis stress test that quantifies aerobic glycolysis/lactate production under stressed conditions was performed. The results demonstrate increasing concentrations of the BAFF 60-mer increased mitochondrial maximal respiration and reserve respiratory capacity (Fig. 4A). Similarly, increasing concentrations of the BAFF 60-mer significantly increased aerobic glycolysis (Suppl. Fig. 3 A), glycolytic capacity, and reserve (Fig. 4B). At 5 nM concentration, compared to the BAFF 3-mer, BAFF 60-mer significantly increased OXPHOS and aerobic glycolysis (Fig. 5A and B). As expected, anti-CD40 Ab and LPS increased both OXPHOS and aerobic glycolysis. It was possible that activation of the B cell receptor would attenuate the 60-mer induced energy metabolism. Therefore, we pre-treated the B cells with an anti-IgM antibody before the addition of the BAFF 3-mer or 60-mer. Similar to the results of BAFF addition to inactivated or resting B cells, the addition of BAFF 60-mer to activated B cells significantly increased OXPHOS (Suppl. Fig. 3 B). Since OXPHOS was significantly increased following BAFF 60-mer treatment, we determined if BAFF treatment increases mitochondrial density in B cells. Splenic B cells were treated with the BAFF 3-mer or 60-mer for 20 hours, stained with MitoTracker Green, and observed under a confocal microscope. MitoTracker Green accumulates in the mitochondrial matrix independent of membrane potential [44]. BAFF 60-mer significantly increased the mitochondrial density of the B cells (Suppl. Fig. 4 A). These results are in agreement with an increase in size and granularity of B cells detected after BAFF 60-mer treatment (Suppl. Fig. 4 B). Western blot analysis for OXPHOS proteins (complex I, II, III, IV, and V compared to beta-actin) of B cell lysate from BAFF 3-mer and 60-mer treated samples revealed a partial decrease in complex III and V intensity, but no changes in complex II intensity (Suppl. Fig. 4 C). These changes were found at 3 hours after the treatment with the BAFF proteins and were persistent until 24 hours. These results suggest the BAFF 60-mer treatment increases mitochondrial efficiency for enhanced energy metabolism.

Figure 4: BAFF 60-mer induces both the aerobic glycolysis and glucose-dependent OXPHOS which are attenuated by mBaffR-Fc.

A and B, Cellular respiration of B cells was determined after treatment with various concentrations of BAFF 3-mer or 60-mer for 20 hours, 1 μg/ml LPS, 10 μg/ml anti-CD40 Ab or left untreated (UT) by a mitochondrial stress test (OXPHOS) and a glycolytic stress test (aerobic glycolysis) on a Seahorse XF24 Analyzer. OCR, Oxygen consumption rate; ECAR, extracellular acidification rate. C, Flexibility of B cells to oxidize glucose, glutamine and long-chain fatty acids in response to BAFF 60-mer was determined by addition of oxidation pathway inhibitors UK5099 (2 μM), BPTES (3 μM), or Etomoxir (4 μM) and a mitochondrial stress test. D and E, OXPHOS and aerobic glycolysis, respectively, of WT B cells were determined from splenic B cells treated with 5 nM of BAFF 3-mer or 60-mer, or left untreated in the presence of 30 nM of the control Ab or mBaffR-Fc for next 20 hours. F, OXPHOS of splenic B cells isolated from the WT and BR3 knockout mice and treated with BAFF 60-mer for 20 hours. Values are expressed as means + SEM, n=4-5/group and “*”, “**” and “***” indicate p<0.05, 0.01 and 0.001, respectively determined by Two-way ANOVA with Dunnett’s multiple comparisons test.

Figure 5: Blocking NF-κB signaling or mBaffR-Fc treatment attenuates BAFF 60-mer-induced B cell activation.

A and B, OXPHOS and aerobic glycolysis, respectively, were determined on murine splenic B cells that were pre-treated with the vehicle or BMS-345541 (5 μM) for 1 hour, and then treated with 5 nM of BAFF 60-mer or left untreated for the next 20 hours. OCR, Oxygen consumption rate and ECAR, extracellular acidification rate (n=5). C-E, Splenic B cells were treated left untreated (UT) or with 5 nM of BAFF 3-mer or 60-mer 20 hours, in presence of BMS-345541 (5 μM) or mBaffR-Fc (30 nM), stained with 200 nM MitoTracker Green (C), 100 nM TMRE (D), or 2.5 μM CellROX Green (E), and mean fluorescence intensity (M.F.I.) of the dyes in live B cells was determined using flow cytometry (n=4). LPS (1 μg/ml) was used as a positive control for B cell activation and 20 μM BAM15 was used as a negative control for dissipation of mitochondrial membrane potential detected by TMRE. Representative histograms of the flow cytometry data are shown in the left panels. F and G, Mitochondrial substrate metabolism assay performed on permeabilized B cells and utilization of glycerol 3-PO4 and succinate is shown. Before permeabilization, the cells were left untreated (UT) or treated with 5 nM of BAFF 3-mer or 60-mer for 20 hours in the presence of BMS-345541 (5 μM) or mBaffR-Fc (30 nM), n=3. H, B cells were left untreated (UT) or treated with 5 nM of BAFF 3-mer or 60-mer 20 h in presence of BMS-345541 (5 μM) or mBaffR-Fc (30 nM), and a glucose uptake assay was performed, n=4. Values are expressed as means + SEM, n=4-5/group and “*”, “**” and “***” indicate p<0.05, 0.1 and 0.001, respectively determined by One-way ANOVA with Tukey’s multiple comparisons test.

Next, to determine the fuel used by mitochondria for the increased OXPHOS in the 60-mer-treated B cells, we performed a Fuel Flex Test. In this test, the BAFF 60-mer-activated B cells were treated with inhibitors of glucose oxidation (UK5099, inhibits pyruvate transport from cytosol to mitochondria), glutamine oxidation (BPTES, inhibits the catalysis of glutamine to glutamate in mitochondria), or long-chain fatty acid oxidation (Etomoxir, inhibits transport of fatty acyl-CoA into mitochondria) pathway, and a mitochondrial stress test was performed. The results demonstrate that OXPHOS was dampened only by the inhibition of glucose oxidation (Fig. 4C), suggesting glucose is the primary fuel for OXPHOS in the BAFF 60-mer-treated B cells.

To determine if binding of BAFF to BR3 is critical for metabolic activation, mitochondrial and glycolysis stress tests were performed on B cells isolated from the spleen of WT mice were treated with BAFF 3-mer, 60-mer, or left untreated in presence of a control antibody or mBaffR-Fc. mBaffR-Fc attenuated the maximal and reserve OXPHOS, and the glycolytic capacity and reserve (Fig. 4D and E). To determine if BAFF signaling via BR3 is critical for the 60-mer-induced metabolic activation, B cells isolated from the spleen of WT mice and BR3 knockout (KO) mice were treated with BAFF 60-mer or left untreated, and a mitochondrial stress test was performed. Although mature B cells are depleted, spleen from BR3 KO mice retains T1 B cells which are known to express the BAFF receptor TACI [45]. Since the cell suspension of B cells was prepared from the spleen of different mice (WT and BR3 KO), the oxygen consumption rate was normalized to cellular protein concentration. As expected, BAFF 60-mer significantly increased basal, maximal, and reserve OXPHOS in the WT B cells; however, these effects were significantly attenuated in the BR3 KO B cells (Fig. 4F). Altogether, these results demonstrate a critical role of BAFF 60-mer binding to BR3 in the metabolic activation of B cells.

3.6. Blocking NF-κB signaling, or mBaffR-Fc treatment attenuates BAFF 60-mer-induced cellular and metabolic activation of B cells

To determine if BAFF 60-mer-mediated NF-κB signaling is critical for metabolic activation, B cells isolated from spleens of WT mice were treated with BMS-345541 (BMS) before treatment with BAFF. BMS inhibits both IKKα and IKKβ and thereby inhibits both NF-κB1 and NF-κB2 signaling and the downstream transcription of target genes. A mitochondrial and a glycolysis stress test revealed BMS significantly attenuated BAFF 60-mer induced OXPHOS and aerobic glycolysis (Fig. 5A and B).

Mitochondrial dysfunction leads to increased production of cellular reactive oxygen species (ROS) production, primarily because of the dissipation of mitochondrial membrane potential (MMP). Using tetramethylrhodamine (TMRE), a positively charged dye that accumulates in active mitochondria (with high membrane potential) of live cells, and CellROX Green, which becomes fluorescent after reacting with cellular ROS, we determined if BAFF multimers differentially affect mitochondrial health and cellular ROS production, respectively, using flow cytometry. As a control, the cells were stained with MitoTracker Green to determine mitochondrial density. The results revealed a modest increase in mitochondrial density and MMP, but a modest decrease in cellular ROS (Fig. 5C, D, and E) in response to BAFF 60-mer indicating the generation of healthy mitochondria and cells. Although BMS attenuated BAFF 60-mer-induced increase in mitochondrial density and MMP, and decrease in cellular ROS content, the basal level of mitochondrial density, and ROS level were increased, whereas, MMP was decreased, suggesting generation of unhealthy mitochondria. BAFF treatment did not improve mitochondrial health in BMS-treated B cells. mBaffR-Fc attenuated BAFF 60-mer-induced changes in mitochondrial density, MMP, and ROS suggesting a critical role of the BR3 binding site on BAFF for the generation of highly active and efficient mitochondria (Fig. 5C, D, and E). As expected, LPS-activated B cells demonstrated increased mitochondrial density and MMP, and decreased ROS suggesting the generation of healthy mitochondria similar to the BAFF 60-mer activated B cells.

To further interrogate changes in mitochondrial activity in B cells in response to BAFF multimers, B cells were treated with BAFF for 20 hours, and plasma membranes were permeabilized by saponin and assayed on MitoPlate S-1 plates, which contain 24 substrates (one substrate/well) used by mitochondria. In the MitoPlate, utilization of the substrates which generate NADH or FDH₂, promotes electron flow through the electron transport chain and finally delivers to a redox dye as a terminal electron acceptor that becomes purple upon reduction. These plates also contain 7 substrates used in the cytoplasm as a control for permeabilization of the plasma membrane. Saponin permeabilizes the plasma membrane by selective removal of cholesterol and does not affect mitochondrial integrity and function. However, a high concentration of saponin can affect mitochondria function. Therefore, we first optimized the concentration of saponin and found 75 μg/ml of saponin (Suppl. Fig. 5), which is within the range (≤100 μg/ml) of optimal permeabilization of cardioblasts to investigate mitochondrial function [46]. In B cells, only two substrates, glycerol 3-PO4 (D,L-α-Glycerol-PO4) and succinate showed very strong signals (Suppl. Fig. 6). Although glycerol 3-PO4 is primarily oxidized in the cytoplasm by the activity of glycerol-3-phosphate dehydrogenase (cGPDH), it can also be oxidized by cytosol-facing mitochondrial mGPDH which is localized on mitochondrial membranes. BAFF 60-mer, BMS, or mBaffR-Fc did not affect glycerol 3-PO4 oxidation (Fig. 5F). In contrast, the 60-mer significantly increased succinate oxidation, even more than BAFF 3-mer treatment, suggesting enhanced activity of succinate dehydrogenase (Fig. 5G). BMS attenuated the BAFF 60-mer-induced increase in succinate utilization and suppressed the basal level of succinate oxidation. Interestingly, mBaffR-Fc did not affect BAFF 60-mer induced oxidation of succinate. A glucose uptake assay demonstrated that glucose uptake was significantly higher in the BAFF 60-mer-treated B cells compared to the untreated and 3-mer-treated cells (Fig. 5H). BMS completely attenuated the BAFF 60-mer-induced glucose uptake (Fig. 5H). However, mBaffR-Fc had a partial attenuation (Fig. 5H). These results suggest a direct role of NF-κB signaling in cellular glucose absorption and the metabolic activity of mitochondria. Blocking BR3 binging site on BAFF attenuates glucose absorption, however, this does not affect succinate utilization suggesting a critical role of BAFF 60-mer mediated cytoplasmic processes in the modulation of mitochondria function.

3.7. Inhibition of classical NF-κB signaling attenuates BAFF 60-mer-induced metabolic activation of B cells.

Since mBaffR-Fc primarily attenuated classical NF-κB signaling, we examined if classical NF-κB signaling is critical for BAFF 60-mer-induced glucose oxidation in B cells. To inhibit classical NF-κB signaling, we used two inhibitors, BI 605906 and IMD 0354 which block phosphorylation and activation of IKKβ, a critical component of the IκB kinase complex that phosphorylates p65. Since the effectiveness of these inhibitors is unknown in B cells, we pretreated B cells with these inhibitors or with BMS for 1 hour and then treated the cells with BAFF 60-mer. BMS attenuated BAFF 60-mer induced p65 phosphorylation and p100 degradation (Fig. 6A). While BI 605906 and IMD 0354 both did not affect p100 degradation, only BI 605906 attenuated p65 phosphorylation in response to BAFF 60-mer. In line with this, BI 605906 attenuated BAFF 60-mer induced both OXPHOS and aerobic glycolysis, whereas IMD 0354 attenuated only the OXPHOS (Fig. 6B and C). These results suggest a critical role of classical NF-κB signaling in BAFF 60-mer-induced metabolic activation of B cells.

Figure 6: Inhibition of classical NF-κB signaling attenuates BAFF 60-mer-induced metabolic activation of B cells.

Murine splenic B cells were pretreated with BMS-345541 (5 μM), BI 605906 (5 μM), IMD 0354 (1 μM), or with vehicle, and then with 5 nM of BAFF 3-mer, 60-mer or left untreated (UT) for 3 hours. Following the treatments, the total cell extracts were resolved on a 4-15% SDS-PAGE, and classical and alternate NF-κB signaling pathways were determined by Western blotting. p65 phosphorylation and p100 degradation were quantified using LI-COR software. Values are expressed as means + SEM and n=3. “*”, “**” and “***” indicate p<0.05, 0.01 and 0.001, respectively determined by a One-way ANOVA with Dunnett’s multiple comparisons test. C and D, OXPHOS and aerobic glycolysis, respectively, were determined on murine splenic B cells that were pre-treated with the vehicle, BI 605906 (5 μM), or IMD 0354 (1 μM) for 1 hour, and then treated with 5 nM of BAFF 60-mer or left untreated for the next 20 hours. OCR, Oxygen consumption rate and ECAR, extracellular acidification rate (n=5). Values are expressed as means + SEM, n=4-5/group and “*”, “**” and “***” indicate p<0.05, 0.01 and 0.001, respectively determined by Two-way ANOVA with Dunnett’s multiple comparisons test.

3.8. Inhibition of mitochondrial respiration, glucose oxidation or NF-κB signaling impairs BAFF 60-mer-induced expression of CD23 and MHC II

Next, we determined if BAFF-induced metabolic activation or NF-κB signaling affects CD23 and MHC II expression on the B cell surface. We blocked OXPHOS using rotenone and antimycin A (Rot + AA), which inhibits the transfer of electrons from mitochondrial complex I and III or blocked glycolysis using 2-Deoxy-D-glucose (2-DG). Cells were then treated with BAFF 3-mer and 60-mer for 20 hours, and the expression of CD23 and MHC II was determined by flow cytometry on live B cells. Rot + AA and 2-DG attenuated the basal level of CD23 expression (Fig. 7A). Whereas Rot + AA completely attenuated, 2-DG partly retained BAFF 60-mer-induced CD23 and MHC II expression (Fig. 7A and B). Both inhibitions led to significantly lower levels of CD23 and MHC II expression in response to BAFF 60-mer.

Figure 7: Inhibition of mitochondrial respiration, glucose oxidation or NF-κB signaling impairs BAFF 60-mer-induced expression of CD23 and MHC II.

A-B, Splenic B cells were pre-treated with antimycin A (5 μM) and rotenone (50 μM), 2-dexoyglucose (12.5 mM), or left untreated for 1 hour, and then treated with 5 nM of BAFF 3-mer, 60-mer, or left untreated for next 20 hrs. Surface expression of CD23 (A) and MHC II (B) on live cells was determined by flow cytometry (M.F.I., mean fluorescence intensity), n=4. C-D, Murine splenic B cells were pre-treated with the vehicle or BMS-345541 (5 μM) (C) and BI 605906 (5 μM), or IMD 0354 (1 μM) (D) for 1 hour, and then treated with 30 nM of 60-mer or left untreated for next 20 hours. Surface expression of MHC II and CD23 was determined by flow cytometry. Representative histograms of the flow cytometry data are shown. Values are expressed as means + SEM, n=4/group and “***” indicates p< 0.001 determined by One-way ANOVA with Dunnett’s multiple comparisons test.

BMS attenuated BAFF 60-mer-induced increase in CD23 and MHC II expression (Fig. 7 C), although, the basal level of MHC II was increased. BI 605906 and IMD 0354 completely attenuated BAFF 60-mer induced CD23 expression (Fig. 7D). BI 605906 attenuated BAFF 60-mer induced MHC II expression, but IMD 0354 did not. Altogether, these results suggest a complex interplay of metabolic and NF-κB signaling on the expression of CD23 and MHC II. Nonetheless, inhibiting glucose oxidation or classical NF-κB signaling attenuated BAFF 60-mer-induced CD23 and MHC II expression.

4. Discussion

Differentiation and proliferation of B2 cells or the mature B cell subsets require BAFF and the BAFF receptor BR3. Proteolytic cleavage of membrane-bound BAFF to soluble BAFF 3-mer is critical for the survival of B2 cells and mice deficient in the production of soluble BAFF shares a similar phenotype with BAFF knockout mice [3]. Interestingly, soluble histidine-tagged human BAFF (amino acid 134-285) exists as a 3-mer at pH ≤6 buffer and as a 60-mer at pH ≥7.4 buffer [5]. Recombinant myc-tagged and Flag-tagged BAFF (amino acid 134-285) exist as 3-mers, even at pH≥7.4, and are used as BAFF 3-mer in cell culture experiments [4]. Both the recombinant human and mouse BAFF associate to form highly active 60-mers [6]. Since 60-mer formation is well-characterized for human BAFF, and BAFF receptors BR3, TACI, and BCMA are expressed on murine B cell surface, we first determined the relative binding properties of human BAFF 3-mer and 60-mer with extracellular soluble fragments of murine BAFF receptors using SPR. As reported for humans [6], human BAFF multimers bound to murine BR3, TACI, and BCMA, and compared to the BAFF 3-mer, BAFF 60-mer has a stronger apparent avidity for BR3 (Supplemental Fig. 1A). Previously, an enzyme-linked immunosorbent assay (ELISA) method identified binding of BAFF 3-mer with human TACI-Fc and BCMA-Fc [6]. Our SPR did not show binding of BAFF 3-mer with murine TACI-Fc or BCMA-Fc. This would be because of the shorter interaction time of ligands with receptors in SPR compared to ELISA. Nonetheless, the SPR results agree with the biological role of BAFF 3-mer as the 3-mer does not signal through TACI and BCMA [6].

Injection of 40 μg/mouse of human BAFF 3-mer into BAFF knockout mice normalizes B cell populations and B cell responses, whereas injection of 40 μg/mouse of human 60-mer (histidine-tagged soluble BAFF) significantly increases the number of B2 cells compared to WT mice [3]. In this line, treatment of splenic B cells (isolated from mice deficient in the production of soluble BAFF) with human BAFF 60-mer strongly induces CD23 expression [3]. These findings prompted us to hypothesize that BAFF 60-mer abnormally activates the B cells. Here, we used CD23 and MHC II to track the maturation and activation of B cells in response to BAFF multimers. BAFF 60-mer treatment not only induced surface expression of CD23 but also the production of sCD23, a prognostic factor for chronic lymphocytic leukemia patients [47]. Although MHC II is highly expressed following antigenic stimulation of B cells and presentation of antigens to CD4+ T cells, MHC II expression on B cells also promotes maturation and proliferation of B cells [48]. Interestingly, BAFF 60-mer-induced CD23 expression was not affected, whereas MHC II expression was attenuated by mBaffR-Fc. These results were further corroborated by our RNA-seq data (Fig. 2) as Fcer2a (CD23a is specifically expressed by B cells) levels were unaffected by mBaffR-Fc treatment and a report by Bossen et al. showed crucial role of BAFF 60-mer-activated TACI in the expression of MHC II [6]. In this line, in mice, 7 days after the 2 mg/kg dose of mBaffR-Fc partly attenuated CD23, whereas strongly attenuated MHC II expression on splenic B cells [49]. Interestingly, mBaffR-Fc treatment attenuated classical NF-κB signaling and selective inhibition of classical NF-κB signaling or glucose oxidation attenuated CD23 expression. This suggests that CD23 expression is primarily supported by BAFF receptors other than BR3. Nonetheless, our RNA-seq data provide conclusive evidence that the 3-mer and the 60-mer differentially regulate gene synthesis in B cells. Moreover, the RNA-seq data analysis using IPA and individual gene and protein expression analysis confirm that BAFF 3-mer mildly induces the genes involved in B cell activation and NF-κB signaling, whereas BAFF 60-mer hyperactivates these genes.

Glucose and glutamine are considered essential nutrients for activated B cells [50, 51]. BAFF 60-mer increased both the mitochondrial respiration and aerobic glycolysis in B cells, similar to effector T cells [52]. Such dependence on dual-energy pathways helps the cells to generate metabolic intermediates which increase the cell’s mass and sufficient adenosine triphosphate essential cellular activity. Lam et al. demonstrated long-lived plasma cells undergo significantly higher OXPHOS compared to short-lived plasma cells and utilize mitochondrial pyruvate import for long-term survival and antibody production [24]. Our metabolic studies demonstrate that, in response to BAFF 60-mer, glucose not only participates in aerobic glycolysis but is also the primary substrate for energy production via OXPHOS in mitochondria. Furthermore, BAFF 60-mer-mediated OXPHOS and glycolysis were critical for the surface expression of CD23 and MHC II on B cells. mBaffR-Fc attenuated both the aerobic glycolysis and OXPHOS. Since mBaffR-Fc may have off-target effects, we used BR3 deficient B cells and found similar results. The BAFF 60-mer mediated increase in OXPHOS was linked to an increase in mitochondrial density and membrane potential, which were attenuated by mBaffR-Fc treatment. Furthermore, BAFF 60-mer-induced decrease in ROS was attenuated by mBaffR-Fc. While these results suggest BAFF 60-mer binding to BR3 is required for increased mitochondrial health, the mitochondrial function assay revealed no significant differences in increased succinate utilization after mBaffR-Fc treatment (Fig. 5G). In the mitochondrial function assay, the plasma membrane of B cells was permeabilized, which inhibits further signaling process in the cytoplasm [53], substrate partitioning/succinate transport into mitochondria, or activity of succinate dehydrogenase. Interestingly, BMS, which inhibits both the NF-κB1 and NF-κB2 signaling pathways impaired mitochondrial health, glucose absorption, and succinate utilization. Therefore, it can be hypothesized that BAFF 60-mer induces critical signaling pathways in the cytoplasm that affects mitochondrial function. In support of this hypothesis, NF-κB is known to be a regulator of mitochondrial respiration [54] and our studies on the Seahorse (Fig. 6) demonstrate that BI 605906, which blocks classical, but not alternate NF-κB signaling, attenuated both the BAFF 60-mer induced OXPHOS and glycolysis and the surface expression of CD23 and MHC II. All these data fit a model in which BAFF 60-mer binding to BR3 activates classical NF-κB signaling which promotes glucose utilization via aerobic glycolysis and OXPHOS and hyperactivates B cells. While BAFF 60-mer induced MHC II expression is regulated partly by signaling via BR3, CD23 expression is regulated by binding of the 60-mer to other BAFF receptors (Suppl. Fig. 7). This model is supported by the finding that BAFF overexpressing mice abnormally express TACI in T1 B cells that secrete class-switched autoantibodies and cause humoral autoimmunity [45].

A recent report identified that, in the absence of a second signal (such as a signal from T helper cells or a pathogen product), activation of B cells via B cell receptor ligation leads to mitochondrial dysfunction in 24 hours [22]. The BAFF 60-mer increased mitochondrial content in B cells, and importantly, the mitochondria were healthy as revealed from multiple studies. Interestingly, the BAFF 60-mer-induced classical NF-κB signaling (not alternate) as well as the increased OXPHOS and glycolysis were dependent on signaling via the BR3 binding site. Furthermore, blocking both classical and alternate NF-κB signaling attenuated the BAFF 60-mer-induced CD23 and MHC II expression, mitochondrial respiration, and glycolysis. Both the RNA-seq and flow cytometry data demonstrate a partial effect of mBaffR-Fc treatment on BAFF 60-mer-induced B cell activation (no effect on CD23 and lower MHC II expression). Therefore, more studies are required to understand how BAFF 60-mer drives CD23 expression on the B cell surface.

Our study has several limitations. According to a previous report, serum from normal humans has BAFF 60-mer dissociating factor, and majority of the BAFF is found as 3-mers in serum collected from BAFF receptor 3 deficient humans [27]. In our method of detection of BAFF multimers, we used plasma from normal humans and found the presence of BAFF 60-mer and higher order multimers. The discrepancy in the results may be different materials used (plasma vs serum) and different methods used for the identification of BAFF multimers. Furthermore, it is unknown what drives the disproportionate level of BAFF multimers in human plasma samples. In our in vitro studies, we used human recombinant BAFF on murine B cells. Murine BAFF may have different effects than human BAFF on murine B cells, however, similar functions of human and murine BAFF are corroborated by the fact that BAFF deficient mice mimicked the phenotype of WT mice after administration of human BAFF 3-mer [3]. Furthermore, replacing mouse BAFF with human BAFF in a humanized mouse model did not affect B cell maturation [55]. Splenic B cells primarily express BR3 and TACI, whereas, plasma cells, which represent a minor B cell subtype, express BCMA [56]. Kim et al. proposed a model for the BAFF-TACI interaction in which two cysteine-rich domains of one TACI molecule contain one conserved Asp-x-Leu motif per domain, each of which binds to one BAFF molecule via the BR3 binding sites [57]. This suggests that the mBaffR-Fc used in this study would have blocked the interaction of BAFF both with BR3 and TACI on the surface of B cells. A novel recombinant BAFF antagonist, BAFF-Trap, blocks BR3 and TACI binding sites on BAFF and inhibits both the classical and alternate NF-κB signaling [58]. Based on our result that mBaffR-Fc does not affect alternate signaling, it can be speculated that mBaffR-Fc: BAFF 60-mer complex can still bind to TACI and activate the alternate NF-κB signaling. This was supported by our ELISA assay in which murine TACI-Fc and BCMA-Fc bound to mBaffR-Fc: BAFF 60-mer complex (Fig. 3C). More studies are needed to determine if murine TACI-Fc and BCMA-Fc remove some of the mBaffR-Fc to bind with BAFF 60-mer in the BAFF 60-mer complex or TACI-Fc and BCMA-Fc bind to other sites than BR3 binding site on the complex. Lastly, our study involves metabolic studies on total splenic B cells, which are made up of multiple B cell subsets, and each subset may be metabolically different. For example, germinal center B cells rely primarily on fatty acid oxidation [59].

BAFF is also produced as the site of inflammation in tissues [49], however, because of technical limitations, so far, BAFF 60-mer has not been detected in tissues. Our study is the first to show the presence of BAFF 60-mer and higher order BAFF multimers in humans. In vitro, BAFF 60-mer induces both the classical and alternate NF-κB signaling pathways. Importantly, BAFF 60-mer via the BR3 binding site induces classical NF-κB signaling and a profound metabolic reprogramming that hyperactivates the B cells. BAFF 3-mer induced transcriptome of B cells that involves activation of B cells, NF-κB signaling, and survival of B cells similar to the BAFF 60-mer, however, at a much lower intensity. However, BAFF 60-mer is not simply a highly active form of BAFF 3-mer, the activities of the 60-mer are different than those of 3-mer. Belimumab completely blocks 3-mer activity [60]. Furthermore, belimumab binds to the BR3 binding site on BAFF, which is crucial for the survival of B2-cells, and belimumab treatment depletes B2-cells [34, 61]. Based on the results of BAFF 60-mer hyperactivating B cells, it can be hypothesized that reagents that block BAFF 60-mer formation without affecting BAFF 3-mer activity would reduce B cells hyperactivation but retain the survival of B cells. Altogether, our studies contribute to the understanding of the role of BAFF multimers the in cellular and metabolic activation of B cells and reveal a direct link between cellular signaling pathways with metabolism, which will be helpful for the development and validation of biologics targeting B cells in murine models.

Supplementary Material

Supplementary Material

Supplemental Figure 1: Differential binding of BAFF 3-mer and 60-mer to BAFF receptors and optimization of BAFF treatment on B cells. A, Protein-protein interaction analysis by SPR: 60-mer BAFF (black lines) or 3-mer BAFF (blue lines) were injected over mBaffR-Fc, TACI-Fc, BMCA-Fc, or anti-BAFF antibody biosensors. A series of increasing analyte concentrations of BAFF 60-mer (0.2, 0.64, 2, 6.4, 20 nM) or BAFF 3-mer (1, 3.2, 10, 32, 100 nM) were injected over each surface using a single cycle injection strategy. Injection phases are denoted by an ‘*’. The resulting sensorgrams were fit to a model of single cycle kinetics and the resulting fits are shown overlaid as red lines. All experiments were performed in triplicate and representative sensorgrams are shown. Immobilization densities are reported in resonance units (RU). ‘K_D’ indicates apparent mean equilibrium dissociation constant, and ‘n.d.’ indicates ‘not determined’. B, Optimization of BAFF concentration: Splenic B cells were left untreated (UT) or treated with various concentrations of BAFF 3-mer or 60-mer for 20 hours, and surface expression of CD23 and MHCII on live cells were analyzed using flow cytometry. Significant differences between 3-mer and 60-mer-treated cells are shown. C, B cells were left untreated (UT) or treated with 5 nM BAFF 3-mer or 60-mer for 20 hours, and the level of soluble CD23 (sCD23) was determined using ELISA. D, B cells were treated with 200 ng/ml of murine IL-4 and IL-6, and left untreated (UT) or treated with 5 nM BAFF 3-mer or 60-mer for 20 hours, surface expression of CD23 and MHC II on live cells was determined by flow cytometry. M.F.I. is mean fluorescence intensity. E, Differential activity of BAFF multimers on CD23 and MHC II expression: B cells were left untreated (UT) or treated with 5 nM BAFF 3-mer or 60-mer in presence of 30 nM of the control Ab, mBaffR-Fc, or anti-BAFF Ab for 20 hours, and surface expression of CD23 and MHC II on live cells was determined by flow cytometry. Representative histograms of the flow cytometry data are shown. Values are expressed as means + SD and n = 4. “***” indicates p<0.001 determined by a One-way ANOVA with Dunnett’s multiple comparisons test.

Supplemental Figure 2: The z-score analysis of B cell transcriptome in response to BAFF 3-mer and 60-mer. QIAGEN Ingenuity Pathway Analysis was used to analyze the ‘Diseases & Biofunctions’ on the relative expression of BAFF 3-mer and 60-mer-treated B cells compared to the untreated B cells (supplement to Figure 2). Biological functions with positive (≥1) and negative (≤−1) z-score are displayed with p-values of the functions. The BAFF 60-mer/control Ab samples were sorted for high z-score and genes with ≥0.2 log2 values were used for the analysis.

Supplemental Figure 3: BAFF 60-mer metabolically reprograms B cells. A, Basal respiration, and glycolysis of B cells were determined after treatment with various concentrations of BAFF 3-mer or 60-mer, 1 μg/ml LPS, 10 μg/ml anti-CD40 Ab or left untreated (UT) for 20 hours by a mitochondrial stress test and a glycolytic stress test, respectively, on a Seahorse XF24 Analyzer. OCR, Oxygen consumption rate; ECAR, extracellular acidification rate. B, OXPHOS was determined in splenic B cells that were treated with an anti-IgM antibody (αIgM, 20 μg/ml) for 1 hr and then with various concentrations of BAFF 3-mer or 60-mer for 20 hours or left untreated (UT) and a mitochondrial stress test was performed on Seahorse. Values are expressed as means + SEM and n = 4. “*” and “***” (compared to the UT) indicate p<0.05 and 0.001, respectively determined by Two-way ANOVA with Dunnett’s multiple comparisons test.

Supplemental Figure 4: BAFF 60-mer increases mitochondrial density and granularity of B cells. A, Splenic B cells were left untreated (UT) or treated with 5 nM of BAFF 3-mer or 60-mer 20 hours, stained with 50 mM MitoTracker Green, visualized under a confocal microscope, and images were collected to a depth of 4.5 μm. Single plane top view images were generated (scale, 10 μm) and MitoTracker stained area was calculated using Image J. Values are expressed as means + SEM and n = 6 images, each image has 20 to 40 B cells. B, Murine splenic B cells were treated with 5 nM of BAFF 3-mer, 60-mer or left untreated (UT) for 3 h and 24 h, the total cell extracts were resolved on a 4-15% SDS-PAGE, and the level of OXPHOS proteins were determined by Western blotting. Values are expressed as means + SEM and n=3. *, p<0.05; and **, p<0.01 determined by a One-way ANOVA with Dunnett’s multiple comparisons test.

Supplemental Figure 5: Optimization of saponin concentration for the permeabilization of plasma membrane of unfixed B cells for the mitochondrial substrate metabolism assay on MitoPlate S-1. Splenic B cells were permeabilized with various concentrations of saponin and incubated on MitoPlate S1 at 37°C in a humidified chamber. Six hours after the incubation, color formation in MitoPlate was measured on at OD 590 nm. For background correction, a MitoPlate was incubated in similar conditions without B cells and OD 590 nm was recorded. The final MitoPlate readings of B cells were subtracted from the background readings.

Supplemental Figure 6: Mitochondrial substrate metabolism assay by using MitoPlate S-1: B cells were treated with the vehicle, LPS (1 μg/ml), BMS-345541 (5 μM), and mBaffR-Fc (30 nM) for 1 hour at 37°C in three different experimental setups. Then, the cells were treated with 5 nM of BAFF 3-mer, BAFF 60-mer, or left untreated for 20 hours. The cells were then permeabilized with saponin, incubated on MitoPlate S1 for 18 hours at 37°C in a humidified chamber, color formation was measured on at OD 590 nm (n=3).

Supplemental Figure 7: A model for BAFF 60-mer-mediated cellular and metabolic activation of B cells. BAFF 60-mer activates both the classical and alternate NF-κB signaling pathways. Particularly, BAFF 60-mer binding to BR3 promotes glucose uptake and activates classical NF-κB signaling which promotes glucose utilization via aerobic glycolysis and OXPHOS, and hyperactivates B cells. While BAFF 60-mer induced MHC II expression is regulated by signaling via BR3, CD23 expression is regulated by binding of the 60-mer to unknown BAFF receptors and activation of alternate NF-κB signaling. mBaffR-Fc blocks BR3 receptor binding site on BAFF, BMS-345541 inhibits both classical and alternate signaling pathways by inhibiting the kinase activity of IKKα and IKKβ, whereas, BI 605906 inhibits only the classical pathway by inhibiting the kinase activity of IKKβ.

Non-standard abbreviations and acronyms:

BAFF

B cell activating Factor

BR3

BAFF receptor 3, BAFF-R

TACI

Transmembrane activator and CAML interactor

BCMA

B-cell maturation antigen

mBaffR-Fc

Murine BR3 ecto-domain fused to murine IgG1 Fc fragment, blocks BR3 binding site on BAFF

NF-κB

Nuclear factor kappa-light-chain-enhancer of activated B cells

IKKα

Inhibitor of NF-κB kinase α

IKKβ

Inhibitor of NF-κB kinase β

SEC

Size exclusion chromatography

FPKM

Fragments Per Kilobase of transcript sequence per Millions base pairs sequenced

IPA

Ingenuity Pathway Analysis

LPS

Lipopolysaccharide

OXPHOS

Oxidative phosphorylation

OCR

Oxygen consumption rate

ECAR

Extracellular acidification rate