Baff−/− mice were as described12,20. C57BL/6OlaHsd mice were purchased from Envigo (Horst, Netherlands, stock# 057) and housed in a specific pathogen-free animal facility. BAFF E247K knock-in mice were generated according to standard procedures. Briefly, embryonic stem cells (129 SvEv) were electroporated with a construct containing the E247K mutation (GAG CTG → AAG CTT) in exon 6 of Baff and a floxed NeoR cassette inserted in introns 6–7 and screened by PCR and Southern blot for homologous recombination events (Supplementary Fig. 2). Recombined embryonic stem cells were injected into C57BL/6 blastocysts followed by transfer into the uterus of pseudo-pregnant recipients. Progeny of chimeric mice that transmitted the mutation was backcrossed for 4 generations onto the C57BL/6 background, then bred to B6.C-Tg(CMV-cre)1Cgn/J mice (sn#006054, The Jackson Laboratory) to remove the NeoR cassette. Heterozygous mice were bred to obtain cohorts of Baffwt/wt, Baffwt/E247K, and BaffE247K/E247K littermates, or bred to Baff−/− to obtain cohorts of Baffwt/wt, Baffwt/–, Baffwt/E247K, and BaffE247K/− littermates. Experiments were performed according to guidelines and under the authorization of the Swiss Federal Food Safety and Veterinary Office (authorization 1370.6 to PS).
Ear biopsies were digested for 2–16 h at 55 °C in 50 µl of the DirectPCR Lysis Reagent (Peqlab, 31-102-T) plus 2 µl of Proteinase K (Roche), followed by 45 min at 85 °C. Supernatants of BAFF-E247K ki lysates were used for PCR using primers fwd 5ʹ-ACCCTGTTCCGATGTATTCA-3ʹ and rev 5ʹ-TAAGAGAGTGCCAGGTCCC-3ʹ, with 30 cycles of 94 °C (7 s)/58 °C (20 s)/72 °C (40 s). Supernatants of BAFF-ko lysates were used for PCR using primers wt shared 5ʹ-GCAGATTGAGCAATCCATGGAAGGCCA-3ʹ, wt specific 5ʹ-CAAGTTGATGTCCTGACCCAAGGCACC-3ʹ, and neo 5ʹ-TGGCAGGGTCTTTGCAGACTCATCCAT-3ʹ, with 30 cycles of 94 °C (7 s)/60 °C (20 s)/72 °C (20 s). For the screening of recombined ES cells, genomic DNA was amplified with primers NeoR fwd 5ʹ-CCTTCTATCGCCTTCTTGAC-3ʹ and BAFF rev 5ʹ-GTGGAACAGATAAGGTGCCT-3ʹ, with 3 cycles of 95 °C (30 s)/64 °C [ramp-0.5 °C/cycle] (30 s)/68 °C (2 min), followed by 30 cycles of 95 °C (30 s)/54 °C (30 s)/68 °C (2 min), using TaKaRa LA polymerase (TaKaRa) at 5 U/ml.
Reagents and cell lines
The following antibodies were used: mouse IgG1 anti-FLAG M2 and biotinylated M2 (Sigma Aldrich F3165 and F9291), goat IgG 852 anti-mouse BAFF (R&D AF2106), biotinylated mouse IgG1 anti-human BAFFR huBR9.1 (Adipogen, Liestal, Switzerland, AG-20B-0016B), rat IgG1 anti-mouse BAFF 5A840, mouse IgG1 anti-mouse BAFF Sandy-2 and Sandy-541, rat IgM anti-human BAFF Buffy-29, and mouse IgG1 anti-EDA EctoD142. Human IgG1 anti-human BAFF antibody belimumab (registered trade name Benlysta) was purchased from the Pharmacy of Lausanne University Hospital (CHUV). hBAFFR-Fc was from Adipogen (AG-40B-0027) and hTACI (aa 31–110)-hIgG1 Fc (aa 245–470, L258E, A353S, P354S) (atacicept) was provided by Merck, KGaA. HEK 293, HEK 293 T, Jurkat, and BJAB cells were obtained from late Jürg Tschopp (University of Lausanne). CHO-S cells were from Thermoscientific (A1155701). Reporter cells Jurkat-JOM2-hBAFFR:Fas clone 2131,43,44 were as described and grown in RPMI supplemented with 10% fetal calf serum (FCS). HEK 293 T cells were grown in DMEM 10% FCS. BJAB and BJAB–TACI cells45 were grown in RPMI supplemented with 10% FCS. Cell lines were tested for mycoplasma using MycoAlert mycoplasma detection kit (Lonza, LT07-318) and found to be negative. As the identity of CHO-S, HEK 293, HEK 293 T, and Jurkat cells does not impact on result interpretation, these cell lines were not authenticated. The identity of BJAB cells was confirmed by microsatellite sequencing (cell line typing service, Microsynth, Balgach, Switzerland).
Immunoglobulin-depleted fetal calf serum
A volume of 500 ml of FCS was depleted from immunoglobulins by repeated passages (usually eight times) at 4 ml/min on a 5 ml Protein A-Sepharose column (GE Healthcare), until the fraction of immunoglobulins bound to the column became negligible. Immonoglobulin-depleted serum was sterilized by filtration at 0.2 µm.
Purification of Fc-containing recombinant protein
Fc-containing proteins (from plasmids listed in Supplementary Table 5) were produced from stable clones of HEK 293 or CHO-S cells. Cells were grown in serum-free OptiMEM medium or in DMEM/F12 (1:1, v/v) medium supplemented with 2% of immunoglobulin-depleted FCS. Proteins in conditioned supernatants were affinity-purified on 5 ml (or 1 ml) Protein A- or Protein G-Sepharose column (GE Healthcare), eluted with 50 mM citrate-NaOH pH 2.7 and neutralized with appropriate amounts of 1 M Tris–HCl pH 9. Proteins were concentrated, and buffer exchanged for PBS using 30 kDa cutoff centrifugal devices, then sterilized by filtration at 0.2 µm, quantified by absorbance at 280 nm using theoretical molar extinction coefficients and stored at −70 °C until use44,46.
Purification of His-tagged BAFF
Escherichiacoli M15 pRep4 bacteria expressing His-tagged hBAFF H218A (3-mer) and His-tagged hBAFF (60-mer)21 in early logarithmic phase of growth were grown overnight at 18 °C in L-Broth medium supplemented with 0.1 mM of isopropyl-thiogalactoside. Bacteria were harvested, lysed by sonication in 30 ml of 10 mM Tris–HCl pH 8, 0.5 M NaCl per liter of culture and insoluble material was removed by centrifugation (15 min, 17,000 × g). Supernatants were loaded (up to three liter equivalent of culture) on a 5 ml chelating-Sepharose column (GE Healthcare) pre-loaded with 0.5 M ZnSO4 and equilibrated in lysis buffer. The column was washed with two volumes of 10 mM Tris–HCl pH 7.4, 150 mM NaCl (TBS), five volumes of TBS 50 mM imidazole, two volumes of TBS, and eluted with three volumes of TBS, 50 mM ethylenediaminetetraacetic acid (EDTA). Proteins were concentrated (30 kDa cutoff) and fractionated by size-exclusion chromatography on a Superdex 200 column as described under ‘size-exclusion chromatography’. Fractions corresponding to BAFF 3-mer or BAFF 60-mer were filter-sterilized, quantified by absorbance at 280 nm and stored at −70 °C.
Generation of the Fab fragment of belimumab
Quantity of 120 mg of belimumab was mixed with 500 μl of immobilized papain beads (ThermoScientific) in 3 ml of 20 mM cysteine, 20 mM Na-phosphate pH 7, 10 mM EDTA, and digested for 4 days at 37 °C. At the end of the incubation, beads were removed and the Fab was recovered in the flowthrough of a Protein A affinity column. Fab were concentrated on 10 kDa cutoff centrifugal devices, then size-fractionated on a Superdex 200 Increase column in 20 mM Hepes pH 7.5, 130 mM NaCl.
Relative affinity of belimumab Fab for Fc-BAFF WT and H218A
Quantity of 300 µg of belimumab Fab in 300 µl of 0.1 M Na-borate pH 8.8 was biotinylated for 2 h at room temperature with 3 µl of sulfo-N-hydroxysuccinimide-LC-biotin (Pierce, 21335) at 30 µg/ml in DMSO. Reaction was terminated by addition of 10 µl of 1 M NH4Cl, and buffer was exchanged to PBS using a 30 kDa cutoff centrifugal device. Fc-hBAFF (WT or H218A, 100 µl) was coated at 1 µg/ml in PBS in an ELISA plate. After blocking, 50 µl of soluble Fc-BAFF (WT or H218A) were added at 2-fold the desired final concentration, followed by the addition of 50 µl of 100 ng/ml biotinylated Fab of belimumab and immediate mixing. After washing, bound Fab was detected with horseradish peroxidase-coupled streptavidin.
Complex between BAFF 3-mer and the Fab of belimumab
Quantity of 8.8 mg of His-hBAFF H218A was mixed with 50 mg of the Fab fragment of belimumab in 20 mM Hepes pH 7.5, 130 mM NaCl, incubated for 16 h at 4 °C, and size-fractionated by size-exclusion chromatography in the same buffer. Fractions of complex were concentrated to 28 mg/ml.
Generation of a complex between BAFF 60-mer and atacicept
Quantity of 0.9 mg of His-hBAFF 60-mer at 4.6 mg/ml was mixed with 9.7 mg atacicept at 75 mg/ml in 20 mM Hepes pH 8.2, 130 mM NaCl and incubated for 1 h on ice. The precipitate was recovered by centrifugation at 4 °C, then re-dissolved at room temperature in 20 mM Hepes pH 8.2, 130 mM NaCl and size-fractionated at room temperature by size-exclusion chromatography in the same buffer. Fractions of the complex were concentrated to 0.8 mg/ml.
HEK 293 T cells were transiently transfected with the polyethyleneimide method47 and grown for 7 days in serum-free OptiMEM, after which time supernatants were harvested and concentrated 20 times in 30 kDa cutoff centrifugal concentration devices.
Generation of BAFFR-deficient cell lines
BJAB and BJAB-TACI cells deficient for BAFFR were generated by lentiviral transduction of a CRISPR/Cas9-expression vector carrying a hBAFFR gRNA (Supplementary Table 5). Annealed oligonucleotides 5ʹ-CACCGGGCCGAGTGCTTCGACCTGC-3ʹ and 5ʹ-AAACGCAGGTCGAAGCACTCGGCCC-3ʹ were cloned in the BsmBI restriction site of lentrcrispr v2 plasmid (Supplementary Table 5)48. This plasmid was co-transfected with co-vectors pCMV-VSV-g and psPAX2 (Supplementary Table 5) into 293 T cells with polyethyleneimide. The next day, cells were washed with PBS and cultured for an additional day in RPMI 10% FCS. Cell supernatants filtered at 0.45 µm were supplemented with 8 µg/ml polybrene (Sigma, H9268) and 3 ml were added to 2 × 106 pelleted target cells that were subsequently cultured overnight in a 12-well plate, then expanded for 2 days in a 6-well plate. Cells were then submitted to two rounds of selection in RPMI 10% FCS, 1 µg/ml puromycin (EnzoLifeSciences). Surviving cells were cloned in RPMI 10% FCS. BAFFR-deficient clones were identified by flow cytometry after staining with biotinylated huBR9.1 at 2 µg/ml, followed by PE-coupled streptavidin at 1/500.
Size-exclusion chromatography was performed at a flow rate of 0.6 ml/min on a Superdex 200 Increase HR 10/30 column equilibrated in either PBS, or 20 mM Hepes pH 7.5, 130 mM NaCl, or 20 mM Hepes pH 8.2, 130 mM NaCl, as indicated. The column was calibrated with 100 µl of a mixture of ferritin (440 kDa) at 140 µg/ml and thyroglobulin (669 kDa), aldolase (158 kDa), ribonuclease A (13.7 kDa) (all from GE Healthcare), bovine serum albumin (67 kDa), ovalbumin (43 kDa), carbonic anhydrase (29 kDa), and aprotinin (6.5 kDa) (all from Sigma-Aldrich) at 1.4 mg/ml in PBS.
Western blot and quantification of BAFF
SDS-PAGE and western blot (using Buffy-2 at 0.5 µg/ml for human BAFF, 852 at 0.5 µg/ml for mouse BAFF and a peroxydase-coupled donkey anti-human IgG for belimumab) were performed according to standard procedures, using WesternBright ECL spray for detection (Advansta) (Supplementary Fig. 9). For the detection of BAFF in fractions of size-exclusion chromatography, proteins were precipitated for 10 min on ice with 5% trichloroacetic acid and 300 µl equivalent of fraction was used for western blot. Coomassie blue staining was performed with a semidry iD Stain System (Eurogentech). WT forms of purified FLAG-mBAFF and FLAG-hBAFF were quantified by SDS-PAGE and Coomassie blue staining against a standard curve of His-BAFF 60-mer41. They were used as standards to quantify FLAG-tagged BAFF 3-mers in Superdex-200 fractions (fraction 13 + 14 for mBAFF and 15 + 16 for hBAFF) and naturally cleaved BAFF in concentrated supernatants by western blot using biotinylated anti-FLAG M2 antibody (0.5 µg/ml) followed by IRDye 800CW-coupled streptavidin (100 ng/ml in PBS, 0.5% Tween-20, 1% powdered milk), or Buffy-2 (0.5 µg/ml) followed by IRDye anti-rat (100 ng/ml) or 852 (0.5 µg/ml) followed by IRDye anti-goat (100 ng/ml), and detection with a LI-COR Odyssey infrared fluorescence detector (LI-COR Biosciences).
Cytotoxic assays with BAFFR:Fas reporter cells
Untagged BAFF in fractions of the size elution column were tested at the indicated dilutions. 3-mer fractions of FLAG-tagged BAFF were tested at the indicated concentrations, in the presence of a fixed concentration of 1 µg/ml of cross-linking (anti-FLAG, 5A8, Sandy-5) or control (EctoD1) antibodies. Fc-BAFF and BAFF 60-mer were tested at the indicated concentrations. When inhibitors were present, they were added at the indicated concentrations and, unless stated otherwise, pre-incubated for several minutes with ligands (in a final volume of 50 µl medium) before addition of 50 µl of reporter cells (20,000–50,000 cells/well). Reporter cells were incubated with BAFF for 16 h, then assayed for viability by the addition of 20 µl of a solution of phenazine methosulfate at 0.9 mg/ml in PBS and 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium at 2 mg/ml in PBS (1:20, v/v) (PMS/MTS), followed by further incubation at 37 °C until adequate color development (2–8 h). Absorbance was measured at 490 nm44.
Receptor–ligand interaction ELISA
For the direct-binding ELISA, BAFFR-Fc, BCMA-Fc, or TACI-Fc (human or mouse) were adsorbed at 1 µg/ml in PBS overnight at room temperature into 96-well immunoplates. The following steps were then performed, with incubations at 37 °C: blocking (in PBS, 4% powdered skimmed milk, 0.5% Tween-20), washing (in PBS, 0.05% Tween-20), incubation with FLAG-tagged BAFF 3-mers of the corresponding species at the indicated concentrations for 1 h (in incubation buffer: PBS, 0.4% milk, 0.05% Tween-20), washing, incubation with biotinylated anti-FLAG at 0.5 µg/ml in incubation buffer for 1 h, washing, incubation with horseradish-coupled streptavidin at 1/4000 dilution in incubation buffer, washing, incubation with 100 µg/ml of 3,3ʹ,5,5ʹ-tetramethylbenzidine in 50 mM citrate-phosphate pH 5, 10% DMSO, 0.03% perborate until color development, addition of 1/3 volume of 2M H2SO4, and absorbance reading at 405 nm44. For the competitive ELISA, titrated amounts of naturally cleaved (untagged) wild-type BAFF were added to adsorbed BAFFR-Fc for 30 min, followed without intermediate washing steps by a fixed and non-saturating concentration of FLAG-tagged BAFF 3-mers (10 ng/ml for human and 50 ng/ml for mouse). Detection with biotinylated anti-FLAG was performed as described for the direct-binding ELISA.
BAFF in mouse sera was detected by AlphaLISA in white 384-well plates41. 5 µl of serum diluted 1:10 in assay buffer (PerkinElmer Life Sciences) was added to 20 µl of a mix of biotinylated Sandy-2 antibody at 75 ng/ml and 0.5 µg of 5A8-coupled acceptor beads in assay buffer and incubated for 1 h at room temperature. Streptavidin-coupled donor beads (1 µg) in 25 µl assay buffer was added. 15 min later, emission at 615 nm after excitation at 680 nm was measured with an Enspire plate reader (PerkinElmer Life Sciences).
hTACI-Fc (100 µg) was mixed to 1 mg of AlphaLISA acceptor beads in 400 µl of 27 mM Na-phosphate pH 8, 5 mM NaBH3CN, 0.03% Tween 20 and incubated for 24 h at 37 °C with agitation, after which time coupling was stopped by addition of 100 µl of carboxymethoxylamine at 65 mg/ml in 0.8 M NaOH for 1 h at 37 °C. Beads were washed twice in 0.1 M Tris–HCl, pH 8 and stored at 4 °C at a concentration of 5 mg/ml in PBS 0.05% Proclin-300. Measures of receptor binding-competent BAFF were performed by mixing 5 µl of mouse serum with 1.5 µg of hTACI-Fc acceptor beads in assay buffer for 3 h at room temperature. Beads were harvested for 10 min at 8000 rpm (6200 × g), washed with assay buffer, suspended in 25 µl of 0.45 µg/ml biotinylated Sandy-2, transferred in white 384-well assay plates, incubated for 1 h at room temperature, after which time 1.5 µg of streptavidin-coupled donor beads in 25 µl of assay buffer was added. Emission at 615 nm after excitation at 680 nm was recorded with an Enspire plate reader (Perkin Elmer).
For B cell stimulation experiments, Sandy-5, 5A8, or EctoD1 (control) antibodies were administered i.p. to BaffE247K/E247K mice at 2 mg/kg on days 0, 7, 14, 21, 28, and 35. Mice were killed by CO2 exposure at day 42 for the analysis of spleens and lymph nodes (inguinal, axillary, and brachial).
Secondary lymphoid organs were homogenized. Red blood cells were lysed for 5 min on ice in 150 mM NH4Cl, 10 mM NaHCO3, 100 µM Na2-EDTA. Lymphocytes were washed in PBS 2% FCS and filtered on a nylon mesh. Cells were incubated with antiCD16/32 (Fc-block, clone 93, 1/100, eBiosciences) and stained with a mix of anti-CD19-PE.Cy7 (clone eBio1D3, 1/200, eBiosciences), anti-CD93-biot (mAb493, 1/100, prepared in-house by Antonius Rolink), anti-CD1d-FITC (clone 1B1, 1/100, eBiosciences), and anti-CD3-APC (clone 17A2, 1/100, eBiosciences) for 20 min at 4 °C followed by PE-Cy5.5-coupled streptavidin (1/200) and analysis with a FACS Canto (BD Biosciences). Data were analyzed with the FlowJo software.
Mouse B splenocyte survival assay
Mouse splenocytes were purified with an EasySepTM mouse isolation kit (StemCell Technologies, #19854). Purified B splenocytes were labeled for 8 min at 37 °C in PBS, 1% FCS, 2 µM carboxyfluorescein succinimidylester (CFSE) and cultured in round-bottomed 96-well culture plates, at a density of 1.25 × 106 cells/ml in 200 µl of medium (RPMI, 10% FCS, 50 µM of 2-mercaptoethanol, 100 U/ml of penicillin, and 100 µg/ml of streptomycin medium) with titrated amounts of size-fractionated FLAG-mBAFF trimer in the presence or absence of 1 µg/ml of cross-linking (5A8, Sandy-2) or control (EctoD1) antibodies. After 72 h, cells were stained with 100 ng of propidium iodide per sample and analyzed by flow cytometry using an AccuriC6 flow cytometer (BD Bioscience). The percentage of viable cells was calculated as the number of live cells (CFSE+, PI-) divided by the number of cells (CFSE+) × 10041.
Crystals of His6-hBAFF H218A in complex with three Fab fragments of belimumab were obtained by mixing 0.1 µl protein solution (28 mg/ml in 20 mM Hepes pH 7.5, 150 mM NaCl) with 0.1 µl reservoir solution (9% (w/v) PEG4000, 0.1 M MgCl2, 0.1 M HEPES pH 7.5) using the sitting drop vapor diffusion method at 293 K. Before flash cooling in liquid nitrogen, crystals were cryo-protected by reservoir solution supplemented with 20% (v/v) ethylene glycol.
Diffraction data were collected at 100 K at the Swiss Light Source beamlight X06SA (SLS, Villigen, Switzerland) using an Eiger X 16 M detector (Dectris, Baden-Daettwil, Switzerland). Data were processed to 2.9 Å resolution using the programs XDS and XSCALE49. The crystals belong to space group P 21 with a solvent content of 64% (Matthews coefficient 3.45). They contain 2 hexameric complexes per asymmetric unit. Each complex, consisting of a BAFF 3-mer bound by 3 Fab fragments of belimumab, has a 3-fold non-crystallographic symmetry. The phase information necessary to determine and analyze the structure of the BAFF – belimumab Fab complex was obtained by molecular replacement using the program PHASER50 and the published structures of BAFF 3-mer (PBD-ID 1KD7) and of an Fab fragment (PDB-ID 7FAB). Subsequent model building and refinement was performed with the software packages CCP4 and COOT51,52. For the measure of the free R-factor, a measure to cross-validate correctness of the final model, about 0.6% of measured reflections were excluded from the refinement procedure (Supplementary Table 4). Several rounds of manual model building in COOT and bulk solvent correction, positional, B-factor and TLS refinement using REFMAC yielded the final model51,52. The Ramachandran Plot of the final model shows 89.4% of all residues in the most favored region, 10.0% in the additionally allowed region, and 0.5% in the generously allowed region. The residues Asp153(E), Asp153(I), and Asp153(N) of the Fab fragments are found in the disallowed region of the Ramachandran plot. Data collection and model statistics are shown in Supplementary Table 4. The atomic coordinates and structure factors have been deposited in the Protein Data Bank, www.pdb.org (PDB-ID code 6ERX). Buried surface areas were calculated with AreaIMol53. Images were generated with the PyMOL Molecular Graphics System, Schrödinger, LLC.
Electron microscopy data acquisition and image processing
Negatively stained specimens were prepared following an established protocol with minor modifications54. Specifically, 2.5 µl of sample was applied to glow-discharged copper electron microscopy (EM) grids covered with a thin layer of continuous carbon film, and the grids were stained with 2% (w/v) uranyl formate. The sample concentrations of BAFF 60-mer, BAFF 60-mer bound with atacicept, and BAFF 3-mer bound with Fab were 23, 20, and 8.8 μg/ml, respectively. These grids were imaged on a Tecnai T12 electron microscope (FEI) operated at 120 kV at a nominal magnification of 67,000× using a 4k × 4k CCD camera (UltraScan 4000, Gatan), corresponding to a calibrated pixel size of 1.68 Å on the specimen level.
The EM data were processed using SAMUEL and SamViewer55. Negative-stain EM images were binned over 2 × 2 pixels for further processing, yielding a pixel size of 3.36 Å. Particle picking was performed using a semi-automated procedure56. 2D classification of selected particle images was carried out with ‘samclasscas.py’, which uses SPIDER operations to run 10 cycles of correspondence analysis, K-means classification, and multi-reference alignment57.
Group sizes for animal experiments were chosen to detect 40% differences between conditions, with variation coefficients of 15–20% (4 animals/group). Animals were assigned randomly to groups, except for males and females that were distributed equally between groups. Analyses were not blinded. Normal distribution of data was confirmed by normality tests using Prism (D’Agostino and Pearson test for n ≥ 8, Kolmogorov–Smirnov test for n = 6 or 7), or was assumed for smaller size groups. One-way analysis of variance with Bonferroni’s multiple comparison tests was used to compare selected groups using Prism. One-way analysis of variance assumes that the standard deviation of the different groups is equal, an assumption that was not always met according to Bartlett’s test for equality of variance. But as this assumption is of little importance when group sizes are (almost) equal, this was ignored.
The authors declare that the data supporting the findings of this study are available within the article, in its supplementary information files, as a dataset58, or are available upon reasonable requests to the authors. The structural datasets are available in the Protein Data Bank repository, www.pdb.org, under the PDB-ID code 6ERX.