LUBAC Suppresses IL-21-Induced Apoptosis in CD40-Activated Murine B Cells and Promotes Germinal Center B Cell Survival and the T-Dependent Antibody Response

B cell activation by Tfh cells, i.e., through CD154 engagement of CD40 and IL-21, and survival within GCs are crucial for the T-dependent Ab response. LUBAC, composed of HOIP, SHARPIN, and HOIL-1, catalyzes linear ubiquitination (Linear M1-Ub) to mediate NF-κB activation and cell survival induced by TNF receptor superfamily members, which include CD40. As shown in this study, B cells expressing the Sharpin null mutation cpdm (Sharpincpdm) could undergo proliferation, CSR, and SHM in response to immunization by a T-dependent Ag, but were defective in survival within GCs, enrichment of a mutation enhancing the BCR affinity, and production of specific Abs. Sharpincpdm B cells stimulated in vitro with CD154 displayed normal proliferation and differentiation, marginally impaired NF-κB activation and survival, but markedly exacerbated death triggered by IL-21. While activating the mitochondria-dependent apoptosis pathway in both Sharpin+/+ and Sharpincpdm B cells, IL-21 induced Sharpincpdm B cells to undergo sustained activation of caspase 9 and caspase 8 of the mitochondria-dependent and independent pathway, respectively, and ultimately caspase 3 in effecting apoptosis. These were associated with loss of the caspase 8 inhibitor cFLIP and reduction in cFLIP Linear M1-Ub, which interferes with cFLIP poly-ubiquitination at Lys48 and degradation. Finally, the viability of Sharpincpdm B cells was rescued by caspase inhibitors but virtually abrogated – together with Linear M1-Ub and cFLIP levels – by a small molecule HOIP inhibitor. Thus, LUBAC controls the cFLIP expression and inhibits the effects of caspase 8 and IL-21-activated caspase 9, thereby suppressing apoptosis of CD40 and IL-21-activated B cells and promoting GC B cell survival.


INTRODUCTION
B lymphocytes are central to humoral immunity, as they differentiate into "effector" cells that secrete class-switched (IgG, IgA, and IgE) and high-affinity mature antibodies (Abs) or "helper" cells that produce regulatory cytokines (1)(2)(3). During the Ab response, B cells expressing antigen (Ag)-specific B cell receptors (BCRs) are engaged by activated T follicular helper cells (Tfh) in secondary lymphoid organs, leading to their robust proliferation and differentiation in germinal centers (GCs), from which plasma cells emerge as long-lived Ab secreting cells (ASCs) and memory B cells are also generated to establish recallable immunity (3)(4)(5). Among the Tfh cell-expressed factors, CD154 (CD40 ligand, CD40L) engages CD40, a member of the TNF receptor superfamily, on target B cells and functions as a "primary" stimulus by inducing epigenetic modulation, activating transcription factor (e.g., NF-kB) and expanding the transcriptome to drive B cell proliferation and differentiation (2,6). The sustained GC reaction requires IL-21, the hallmark cytokine of Tfh cells (7,8), to boost B cell proliferation and induce plasma cell differentiation by upregulating BLIMP-1, the master transcription factor of plasma cells (9)(10)(11)(12). IL-21 also synergizes with CD154 to induce AID, an activated B cell-restricted cytidine deaminase that introduces DNA lesions in the immunoglobulin (Ig) genes to initiate class switch DNA recombination (CSR) and somatic hypermutation (SHM), which underpin Ab class-switching (e.g., from IgM to different IgG isotypes) and affinity maturation, respectively (2,13). Other CD4 + T helper cells, such as Th1 and Th2 cells, also participate in the GC reaction, e.g., through their hallmark cytokines. In particular, Th2 cytokine IL-4 synergizes with CD154 to activate NF-AT to enhance B cell proliferation as well as induce AID expression and IgH Ig1-Cg1 and Ie-Ce germline transcription to initiate CSR to IgG1 and IgE (2,14).
CSR, most notably IL-4-dependent CSR to IgG1, is likely completed before the full development of IL-21 + Tfh cells and GCs (15,16). Within the GC dark zone, class-switched B cells (and some IgM + cells) acquire a high load of mutations in the V (D)J region DNA to alter the affinity of BCRs to the Ag. B cells bearing the BCRs of higher affinities are positively selected in the GC light zone and then shuttle back into the dark zone for new rounds of proliferation and SHM, an iterative process that eventually leads to affinity maturation (17). This can be manifested by the appearanceat the peak of the GC reactionof several dominant BCR families whose members carry the signature mutation obtained by the founders. The positive selection of founder B cells and a subset of progenies in later rounds could be based on their superior survival in GCs, a notion consistent with that a high proportion of GC B cells are apoptotic and likely subclones outcompeted and failing to cross the survival threshold (18)(19)(20). Increasing evidence suggests that higher-affinity BCRs result in more efficient Ag uptake and MHC presentation to Tfh cells, which in turn endow GC B cells with a better survival capacity (21,22). However, the stimuli that trigger the death of GC B cells and survival mechanisms employed by B cells are still poorly understood.
The linear ubiquitin chain assembly complex (LUBAC), which is composed of the HOIP catalytic subunit, SHARPIN structural subunit, and HOIL-1, is at the nexus of regulation of NF-kB activation and cell survival induced by TNF receptor superfamily members (23,24). Prompted by this, we hypothesize that LUBAC plays an important role in CD40-mediated B cell proliferation, survival, and/or differentiation in response to immunization to T-dependent Ags. To test our hypothesis, we used a well-defined in vitro culture system to recapitulate the opposing impact of Tfh cell stimuli, i.e., induction of B-cell death by IL-21 and maintenance of survival by CD154, on B cells expressing the cpdm null mutation of Sharpin (Sharpin cpdm ) and/ or treated with HOIPIN-8, a recently discovered small molecule inhibitor of HOIP (25,26). We also explored the molecular mechanisms underpinning IL-21-induced B cell apoptosis and the role of SHARPIN and LUBAC in regulating this process. We further addressed the B cell-intrinsic role of SHARPIN in specifically mediating the GC B cell survival and Ab responses to a T-dependent Ag in vivo by generating mice with B cellspecific expression of Sharpin cpdm (B-Sharpin cpdm ) and mice in which Sharpin cpdm B cells directly competed against wildtype B cells within the same GC environment. Finally, we performed SHM analysis of over 20,000 BCR-encoding sequences to provide molecular evidence that B-cell SHARPIN promotes positive selection for high-affinity Abs.
All mice were maintained in a pathogen-free vivarium at the University of Texas Health Science Center at San Antonio (UTHSCSA). For immunization, mice were injected intraperitoneally (i.p.) with 100 mg of NP-CGG (in average 16 molecules of NP, 4-hydroxy-3-nitrophenyl acetyl, conjugated to one molecule of CGG, chicken g-globulin; Biosearch Technologies) in the presence of 100 ml of alum (Imject ® Alum adjuvant, ThermoFisher) in the central-left abdomen area. Both male and female mice were used. All protocols were in accordance with the rules and regulations of the Institutional Animal Care and Use Committee (IACUC) of UTHSCSA.

B Cell Isolation, Culture, and Stimulation
Mouse immune cells were isolated from single cell suspensions prepared from the spleen. Spleen cells were resuspended in ACK Lysis Buffer (Lonza) to lyse red blood cells. After quenching with RPMI 1640 medium supplemented with 10% FBS, 50 mM bmercaptoethanol and 1x antibiotic-antimycotic mixture (Invitrogen) (RPMI-FBS), cells were resuspended in PBS for flow cytometry analysis or further preparation. To isolate B cells, splenocytes were subjected to negative selection (against CD43, CD4, CD8, CD11b, CD49b, CD90.2, Gr-1 or Ter-119) using EasySep ™ Mouse B cell Isolation Kit (StemCell ™ Technologies) following the manufacturer's instructions, resulting in the preparation of more than 99% IgM + IgD hi B cells. After pelleting, B cells were directly used for genomic DNA extraction for genotyping, RNA extraction, or resuspended in RPMI-FBS for stimulation. In genotyping experiments, non-B cells, i.e., those that bound the Ab cocktail and magnetic beads, were also subjected to genomic DNA extraction.

Flow Cytometry and B Cell Proliferation Analysis
To analyze B cells and other immune cells ex vivo, spleen or blood cells (2x10 6 ) were first stained in Hank's Buffered Salt Solution plus 0.1% BSA (HBSS-BSA) for 20 m with fluorophorelabeled mAbs to surface markers (Supplementary Table 1) in the presence of mAb Clone 2.4G2, which blocks FcgIII and FcgII receptors, and fixable viability dye (FVD) or 7-AAD without permeabilization. After washing, cells were resuspended in HBSS for FACS analysis. FACS data were analyzed by the FlowJo ® software (BD).
To analyze B cell proliferation in vivo, mice were injected twice i.p. with 2 mg of bromodeoxyuridine (BrdU) in 200 ml PBS, with the first and second injection at 24 h and 1 h prior to sacrificing, respectively. Splenocyte (2x10 6 ) were washed with BSA-HBSS and stained with Abs specific for surface markers in the presence of FVD. After washing, cells were resuspended in the BD Cytofix/Cytoperm ™ buffer (250 ml) and incubated at 4°C for 20 m. After washing twice with the BD Perm/Wash ™ buffer, cells were counted again and 10 6 cells were resuspended in 100 ml of BD Cytofix/Cytoperm ™ buffer for staining with fluorochrome-conjugated anti-BrdU mAb and/or 7-AAD for 30 m. After washing with BD Perm/Wash ™ buffer, cells were analyzed by FACS. All data were analyzed by FlowJo ® (BD).
To analyze B cells stimulated in vitro, cells were harvested, stained with FVD or 7-AAD without permeabilization in HBSS-BSA for 20 m and analyzed by flow cytometry for expression of Igg3, Igg1 and Igg2b expression (CSR to IgG3, IgG1, and IgG2b), CD138 (plasma cell marker) and other B cell surface molecules. To analyze B cell proliferation in vitro, CellTrace ™ Yellow Cell Proliferation Kit (ThermoFisher) was used following the manufacturer's instructions with minor modifications. Briefly, 10x10 6 purified naïve B cell were stained with CellTrace ™ Yellow (10 mM, diluted from the 5 mM stock solution in 1 ml DPBS) for 5 m at 37°C with protection from light. After incubation, cells were pelleted by centrifugation for 5 m at 800g, washed with 10 ml RPMI-FBS before cultured for stimulation. Up to 10 cell divisions could be traced after 96 h of stimulation by CD154. B cells were harvested and analyzed by flow cytometry (Ex532nm, Em555/580nm).

Cell Viability, Apoptosis, and Caspase Activity Analysis
For viability analysis, B cells were stained with 7-AAD, which enters the cell that had a compromised plasma membrane integrity, together with Abs specific surface marker in HBSS-BSA for 20 m, washed once, and resuspended in HBSS-BSA for FACS analysis. To analyze apoptosis and necrosis using 7-AAD and Annexin V, which binds to the phosphatidylserine that was located at the intracellular leaflet of the plasma membrane and exposed to Annexin V once "flipped" outside, B cells were stained with Abs specific for surface markers in HBSS-BSA on ice for 20 m. After washing, cells were resuspended in 100 ml Annexin V binding buffer (BioLegend) containing 7-AAD and 1:50 diluted FITC-conjugated Annexin V (BioLegend) at RT for 15 m. After staining, 400 ml of Annexin V binding buffer was added and cells were immediately analyzed by flow cytometry to analyze apoptotic (Annexin V + 7-AAD lo ) cells and necrotic (Annexin V + 7-AAD hi ) cells. The caspase activity in freshly harvested spleen B cells from immunized mice was detected using the CaspACE ™ FITC-VAD-FMK (Promega), a FITCconjugated analog of the pan caspase inhibitor Z-VAD-FMK that could diffuse into cells and irreversibly bind to all activated caspases. Splenocyte (10 6 ) were cultured with CaspACE ™ FITC-VAD-FMK (10 mM) in RPMI-FBS at 37°C for 15 m and then harvested for staining with Abs specific for surface markers in HBSS-BSA on ice for 20 m. After washing once and resuspended in HBSS-BSA, cells were analyzed by FACS. To analyze cleaved caspase 3 in B cells by intracellular staining, splenocyte (10 6 ) were washed with BSA-HBSS and stained with Abs specific for surface markers in the presence of FVD. After washing, cells were fixed with the BD Cytofix/Cytoperm ™ (200 ml) at 4°C for 20 m. After washing twice with BD Perm/Wash ™ buffer, cells were stained in the same buffer with PE-conjugated Ab for cleaved caspase 3.

Mitochondrial Membrane Potential Assay
Mitochondrial membrane potential (Dym) was measured by the JC-1 Mitochondrial Membrane Potential Assay Kit (Abcam) following the manufacturer's instructions. When the m i t o c h o n d r i a l m e m b r a n e p o t e n t i a l i s l o w , J C -1 (tetraethylbenzimidazolylcarbocyanine iodide, a cationic dye) becomes a monomer that yields green fluorescence (JC-1 green) detectable in the FITC channel (530 nm Em) in flow cytometry. When the mitochondrial membrane potential is high, the dye accumulates into high concentrations in mitochondria, thereby aggregating and yielding red fluorescence (JC-1 red) detectable in the PE channel (590 nm Em). Stimulated B cells were washed with 1x dilution buffer and then stained with 10 mM JC-1 in 1x dilution buffer for 15 m at 37°C. Cells were washed once with 1x dilution buffer and stained for surface markers before flow cytometry analysis.
For ELISPOT analysis of NP 7 -binding and total IgM + and IgG1 + ASCs, Multi-Screen ® filter plates (Millipore) were activated with 35% ethanol, washed with PBS, and coated with anti-IgM, anti-IgG or NP 7 -BSA (all 5 mg/ml) in PBS. Single spleen or bone marrow cell suspensions were cultured at 250,000 cells/ml (in plates coated with NP 7 -BSA) or 125,000 cells/ml (in plates coated with anti-IgM or anti-IgG) in RPMI-FBS at 37°C for 16 h. After supernatants were removed, plates were incubated with biotinylated goat anti-mouse IgM or -IgG1 Ab, as indicated, for 2 h and, after washing, incubated with HRPconjugated streptavidin. Plates were developed using the Vectastain AEC peroxidase substrate kit (Vector Laboratories). The stained area in each well was quantified using the CTL Immunospot software (Cellular Technology) and depicted as the number of spots for quantification.

RNA Isolation, qRT-PCR and RNA-Seq
Total RNA was extracted from 5 x 10 6 B cells using the RNeasy Mini Kit (Qiagen). First-strand complementary DNA (cDNA) was synthesized from equal amounts of total RNA (4 mg) using the SuperScript III System (Invitrogen) and an oligo-dT primer. cDNA was analyzed by qPCR using SYBR Green (Bio-Rad) and appropriate primers (Supplementary Table 2). PCR was performed in a CFX96 ™ Real-Time PCR System (Bio-Rad Laboratories) according to the following protocol: 95°C for 30 s, 40 cycles of 95°C for 10 s, 55°C for 30 s, 72°C for 30 s. Melting curve analysis was performed at 72°C-95°C. The DDCt method was used to analyze transcript levels and data were normalized to the expression of Cd79b, which encodes BCR Igb, as constitutively expressed in B cells.
For RNA-Seq, after RNA integrity was verified using an Agilent Bioanalyzer 2100 ™ (Agilent), RNA was processed using an Illumina TruSeq RNA sample prep kit v2 (Illumina). Clusters were generated using TruSeq Single-Read Cluster Gen. Kit v3-cBot-HS on an Illumina cBot Cluster Generation Station. After quality control procedures, individual RNA-Seq libraries were pooled based on their respective 6-bp index portion of the TruSeq adapters and sequenced at 50 bp/sequence using an Illumina HiSeq 3000 sequencer. The resulting reads, typically 16 million reads per sample, were checked by assurance (QA) pipeline and initial genome alignment (Alignment). Demultiplexing with CASAVA was employed to generate a Fastq file for each sample. After removing the adaptor and poorquality reads using Trim Galore, all sequencing reads were aligned with the (GRCm38/mm10) reference genome by using HISAT2 with default settings, yielding Bam files, which were further processed using HTSeq-count to obtain counts for each gene. RNA expression levels were determined using GENCODE annotation. Differential expression analysis was performed using the Deseq2 package in R post-normalization based on a Benjamini-Hochberg false discovery rate (FDR)-corrected threshold for statistical significance of p value <0.01 and log 2 FC >1 The count of differentially expressed genes was used to generate heatmaps using Clustvis software. For gene set enrichment analysis, the 112 RelA target genes identified by Ngo et al (27) were used for GSEA analysis of the enrichment of the Deseq2 package normalized RNA-Seq data.

High-Throughput Repertoire and SHM Analysis
To analyze the repertoire usage and SHM in the V 186.2 region DNA, RNA was extracted from splenic B cells for cDNA synthesis, as described above. Rearranged V 186.2 DJ H -Cm and V 186.2 DJ H -Cg1 cDNA was amplified using Phusion ™ highfidelity DNA polymerase (New England BioLabs) and a V 186.2 leader-specific forward primer and a reverse primer specific for the Cm or Cg1 exon (Supplementary Table 2), followed by the second round of PCR using the same forward primer and a nest reverse primer tagged with an Illumina clustering adapter (Supplementary Table 2), using the following protocol: 98°C for 10 s, 60°C for 45 s and 72°C for 1 m for 30 cycles. The amplified library was tagged with barcodes for sample multiplexing and analyzed by 300-bp pair ended sequencing (the Illumina Mi-Seq system).
The BCR repertoire usage and mutations in V 186.2 (V1-72) segments were analyzed using the web-interfaced International ImmunoGeneTics Information System ® IMGT/HighV-QUEST (http://www.imgt.org). The mutation collection process is described in our previous study (28). Briefly, IgBLAST v1.15.0 (http://www.ncbi.nlm.nih.gov/igblast/) was applied for the alignment of the datasets of SHM sequencing (29). Change-O v1.0.0 (https://changeo.readthedocs.io/en/stable/) python package was applied for processing the output of V(D)J sequencing data (30). For mutation counting, Fasta.fmt7 files were generated by MakeDb.py with reference of imgt_mouse_ ig_v.fasta, imgt_mouse_ig_d.fasta, imgt_mouse_ig_j.fasta and ParseDb.py was applied with -if SEQUENCE_ID -sf SEQUENCE_IMGT -mf V_CALL DUPCOUNT. Mutations in the V 186.2 (IMGT V1-72) segment were filtered and aligned based on a single nucleotide with uncertain or missing bases replaced by "N" or "NA". The final step was applied by metric summary. Only unique sequences were further analyzed. The R pipeline was used to count point-mutations.

Mitochondria-Free Cytosolic Fractionation, Immunoblotting, and Immunoprecipitation (IP)
To prepare whole-cell lysates, B cells (10 7 ) were harvested by centrifugation at 500 g for 5 m, resuspended in 0.5 ml of lysis buffer (20 mM Tris-Cl, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 1% (v/v) NP-40) supplemented with Halt ™ Protease & Phosphatase Inhibitors Cocktail (ThermoScientific). To prepare mitochondria-free cytosolic fractions, B cells (10 7 ) were centrifuged at 500 g for 5 m. Cell pellets were processed to obtain the mitochondria-free cytosolic fraction using the Mitochondrial Isolation Kit (ThermoFisher, Cat# 89874) following the manufacturer's instructions. The whole cell lysates and the mitochondria-free cytosolic fraction were subjected to SDS-PAGE, and immunoblotting involving specific Abs (Supplementary Table 1). Membranes were then stripped with Restore ™ PLUS Western Blot Stripping Buffer (ThermoScientific) for re-immunoblotting. Signals were quantified by ImageJ ® (NIH).
For IP, spleen B cells (10 7 ) were resuspended in lysis buffer (20 mM Tris-Cl, pH 7.5, 100 mM NaCl, 1 mM EDTA, 0.5% (v/v) NP-40) supplemented with Halt ™ Protease and Phosphatase Inhibitors Cocktail. After sonication and centrifugation, protein lysates were precleared with equilibrated Protein A/Gconjugated Sepharose ™ 4B beads (ThermoFisher, 50 ml) and incubated with anti-cFLIP mouse Ab in 500 ml of lysis buffer at 4°C for 4 h in the presence of Protein A/G Sepharose ™ 4B beads. After washing with lysis buffer 3 times, immunoprecipitated proteins were eluted in SDS sample buffer for immunoblotting.
In the second step, Compound 2 (483 mg, 2.17 mmol) and Compound 5 (500 mg, 2.17 mmol) were dissolved in EtOH (12 mL) in a 100-ml glass round-shaped flask. The suspension was cooled to 0°C and, with NaOH (2.9 ml, 6M, 17.34 mmol in total) added, allowed to warm to RT. The mixture was stirred for 16 h until the reaction was quenched with HCl (12 ml, 2M), transferred to a separator funnel, and extracted with chloroform:isopropanol (3:1) three times. The organic layers were collected and dried over Na 2 SO 4. After filtration and concentration, the residue was purified by C18 reversed phase flash chromatography using a 0-55% acetonitrile:water gradient to give the HOIPIN-8 base. Finally, the (HOIPIN-8) freebase (80 mg, 0.184 mmol) was dissolved in EtOH (1.5 ml) in a 100-ml glass round-shaped flask, added NaOH (0.2 mL, 1M, 0.184 mmol) at 0°C and stirred for 1 h, warmed to RT and stirred for 2 h. The suspension was then lyophilized to obtain HOIPIN-8 a yellow powder (25% yield) with the following profile: 1

Immunofluorescence Imaging
Spleens were embedded in OCT (Tissue-Tek) and snap-frozen on dry ice. Cryostat sections (5 mm) were fixed in pre-chilled acetone for 10 m, air dried at 25°C, washed with PBS, and blocked with 5% FBS in DPBS for 1 h. Sections were stained with FITC-conjugated anti-B220 mAb (1:500) and PerCP-Cy5.5conjugated anti-GL-7 mAb (1:100) in a humidified chamber overnight at 4°C. Slides were mounted using ProLong ® Gold with DAPI for analysis under a Zeiss LM710 confocal microscope. All images are pseudocolored.

Statistical Analysis
Statistical analysis was performed by either GraphPad (Prism ® ) or Excel (Microsoft) software to determine p values by Student ttest. p values less than 0.05 were considered significant. Correlation analyses were performed using Prism ® .  Figure 1B), culminating in decreased formation of ASCs that produced NP-binding IgM and IgG1 Abs as well as much reduced titers of NP-specific IgM and high-affinity (NP 7 -binding) IgG1, IgG2b, IgG3 and IgA ( Figures 1E, F) total NP-specific (NP 34 -binding) IgG1 titers were also reduced to similar extents ( Figure 1G). ASCs secreting non-specific IgM and IgG1 were reduced too, resulting in decreased titers of circulating IgM and IgG1also decreased were IgG2a and IgA, as elicited by both T-dependent and Tindependent antigens, as well as IgG3 and IgG2b, as elicited mainly by T-independent antigens ( Figures 1H, I). Residual B-Sharpin cpdm GC B cells, however, expressed IgG1 at high levels, indicating their normal CSR ( Figure 1J).

B-Cell SHARPIN Promotes T-Dependent Ab Responses and a BCR Affinity-Enhancing Mutation
Upon NP-CGG immunization, B cells with a recombined IgH V 186.2 region, which encodes NP-binding BCRs, enter GCs and accumulate V 186.2 DNA mutations to generate BCR mutants for selection (32)(33)(34). As shown by analysis of IgM-encoding V 186.2 DJ H -Cm and IgG1-encoding V 186.2 DJ H -Cg1 transcript, NP-CGG-immunized B-Sharpin cpdm mice had similar BCR repertoires in IgG1 + B cells as B-Sharpin +/+ mice, but more diverse ones in IgM + B cells (Figure 2A and Supplementary Figure 1C). V 186.2 DNA mutations occurred at high levels in V 186.2 DJ H -Cg1 of both B-Sharpin +/+ and B-Sharpin cpdm mice, averaging 1.0x10 -2 and 0.6x10 -2 change per base, respectively ( Figure 2B and Supplementary Figure 1D). They also frequently occurred in V 186.2 DJ H -Cm and featured G!A and C!T transitions as major substitutions ( Figure 2B), likely due to the "replication over" of uracils generated after AID deamination of cytidines (pairing "A" with "U" instead of "G" with the original C). Such transitions were also dominant, albeit not as much, in V 186.2 DJ H -Cg1, likely reflecting mutation spreading by error-prone DNA synthesis. In both B-Sharpin +/+ and B-Sharpin cpdm mice, base changes in CDR1 and CDR2 of V 186.2 DJ H -Cg1 DNA were dominated by replacement mutations (Figure 2C), consistent with the notion that amino acid residues in CDRs are altered to serve as the substrates for selection. In particular, a G!T transversion substitution (G 98 !T 98 ) changes the CDR1 Trp 33 residue (TGG) to Leu (TTG) to raise the BCR/Ab affinity to NP by 10-fold (34,35). It was the most frequent mutation within the CDR regions of V 186.2 DJ H -Cg1 in B-Sharpin +/+ mice, occurring in 19.7% of B cells in these mice, much higher than the 4.18% frequency in B-Sharpin cpdm B cellsall other CDR mutations showed smaller differences ( Figure 2D and Table 1). Furthermore, among the ten most frequent mutations in the entire V 186.2 DJ H -Cg1 DNA, all but G 98 !T 98 were in the FR regions (five in FR1, one in FR2, and three in FR3) and showed less difference between B-Sharpin +/+ and B-Sharpin cpdm mice ( Figure 2D and Table 1). In addition, G 98 !T 98 was unique to V 186.2 DJ H -Cg1 DNA, in contrast to the frequent occurrence of all other dominant mutations also in V 186.2 DJ H -Cm DNA ( Figure 2D). Finally, while the CDR1 in V 186.2 DJ H -Cm did not show overall enrichment of replacement mutations ( Figure 2C), the CDR1 G 92 !A 92 and CDR2 G 197 !A 197 or T 197 replacement mutations occurred much less frequently in B-Sharpin cpdm mice ( Figure 2D and Table 1).
Thus, SHARPIN plays a B cell-autonomous role in promoting the GC reaction, accumulation of the affinity-augmenting G 98 !T 98 substitution and maturation of the T-dependent Ab response.
B Cell-Intrinsic Role of SHARPIN in Promoting GC B Cell Survival B cell clones with higher-affinity BCRs need to survive in GCs towards maturation of the Ab responses. In association with muchdecreased high-affinity NP-specific IgG Abs and enrichment of the affinity-enhancing G 98 !T 98 mutation in B-Sharpin cpdm mice, FAS hi GL-7 hi GC B cells in these mice had increased death, including more apoptosis and necrosis ( Figure 3A). Among live B-Sharpin cpdm GC B cells, more than half showed caspase activation (Z-VAD-FMK + ), a hallmark of pre-apoptotic and apoptotic cells ( Figure 3B). The proportions of apoptotic, necrotic, and Z-VAD-FMK + cells in non-GC (FAS -GL-7 -) B cells were all much lower than those in GC B cells in both B-Sharpin +/+ and B-Sharpin cpdm mice. They were comparable in these mice, showing that SHARPIN deficiency selectively affects GC B cell survival.
Overall, SHARPIN plays a B cell-intrinsic role in inhibiting apoptosis of GC B cells and mediating their survival during the T-dependent Ab response.

SHARPIN Suppresses IL-21-Induced Death in CD154-Stimulated B Cells
To understand the mechanisms underlying the increased death of Sharpin cpdm GC B cells, we subjected purified Sharpin cpdm B cells to stimulation by membrane-bound CD154, which mimics CD154 expressed on the plasma membrane of Tfh cells that potently engages CD40 to initiate and sustain the GC reaction (6,38,39). CD154-stimulated Sharpin cpdm B cells only mildly impaired in canonical NF-kB activation, as indicated by phosphorylation of the p65 subunit at Ser333, and were largely normal in expressing p65 target genes ( Figures 4A, B). Consequently, they showed normal proliferation and IL-4directed CSR to IgG1 ( Figure 4C). While neither Sharpin +/+ nor Sharpin cpdm B cells differentiated into plasma cells at a high level after 96 h of stimulation with CD154 and IL-4, they both could do so after 48 h of stimulation, and being washed to remove IL-4 and re-stimulation for 48 h with CD154 and IL-21 ( Figure 4D). Accordingly, CD154-stimulated Sharpin cpdm B cells were fully capable of inducing Prdm1 (encoding BLIMP-1) and upregulating AID in the presence of IL-21 as well as undergoing IL-4-and IFNg-directed Ig1-Cg1 and Ig2a-Cg2a germline transcription, respectively ( Figure 4E). They also shared a similar transcriptome with their Sharpin +/+ B cell counterparts, The AG92C93 (Ser) codon has a high mutability and all mutations are replacement mutations except for AGT93 (36). ‡ The TG98G (Trp) to TT98G (Leu) replacement mutation changes the affinity of Abs to NP by 10-fold (34,35). The frequency of G98 ®T98 mutation were 33.6x10-2, 11.0 x10-2, and 9.8x10-2 change/base in three B-Sharpin+/+ mice, and 7.8 x10-2, 2.9x10-2, and 0.2x10-2 change/base in three B-Sharpincpdm mice. Since only unique sequences were analyzed, the frequencies of each mutation were equivalents of the proportions of B cell clones carrying such mutations. with only 76 (23 downregulated and 53 upregulated) and 83 (36 downregulated and 47 upregulated) differentially expressed genes upon stimulation by CD154 and CD154 plus IL-4, respectively ( Figure 4F and Supplementary Figure 2A). Finally, Sharpin cpdm B cells underwent robust proliferation, CSR, and plasma cell differentiation upon stimulation by LPS (Supplementary Figure 2B), showing that SHARPIN did not mediate B cell responses to T-dependent or T-independent stimuli. IL-21 enhanced proliferation and induced plasma cell differentiation of CD154-stimulated B cells, but triggered their death, resulting in a net loss of such cells (Figures 5A-C). The killing effect of IL-21 was exacerbated by the Sharpin deficiency, leading to as few as 10% of Sharpin cpdm B cells able to survive, virtually all of which had completed high numbers of division ( Figures 5B, C). By contrast, Sharpin cpdm B cells stimulated with CD154 alone or with IL-4 showed high viability, which was only marginally lower than that of Sharpin +/+ B cells ( Figure 4C). Sharpin cpdm B cells were also normal in survival upon stimulation by aIgM plus IL-4, but defective upon stimulation IL-21 alone, LPS or LPS plus IL-4 -aIgM (alone or with IFNg) or LPS plus IL-21 could not maintain the viability of even wildtype B cells (Supplementary Figure 2B-D). High doses of CD154 countered the death-inducing effect of IL-21 in Sharpin +/+ B cells and more effectively in Sharpin cpdm B cells ( Figures 5A, D, E), showing a SHARPIN-independent pro-survival activity of strong CD40 signals.
CD40 could be engaged in vitro by aCD40, which activated NF-kB in wildtype B cells, albeit not as quickly or potently as CD154 ( Figures 4A, 5F), and require IL-21 to induce these B cells to proliferate (Figures 5D, E). More than half of dividing Sharpin +/+ cells, however, were unable to survive, even in the presence of high doses of aCD40. aCD40-stimulated Sharpin cpdm B cells showed much weaker NF-kB activation at early time points when B cells were particularly sensitive to the killing by IL-21 ( Figure 5F) and failed to robustly proliferate even in the presence of IL-21, which instead triggered death in both dividing and non-dividing B cells (Figures 5D, E). Upon stimulation with aCD40 plus IL-4, Sharpin cpdm , by contrast, showed comparable survival as Sharpin +/+ B cells ( Figure 5G).
To summarize, SHARPIN does not play a major role in CD154-induced B-cell NF-kB activation, gene expression, proliferation, CSR, or plasma cell differentiation. Rather, it specifically suppresses IL-21-induced B cell death through a mechanism not entirely overlapping with the one underlying the effect of CD154 in countering the killing by IL-21.

SHARPIN Inhibits Caspase 8 and Caspase 9 Activation and B Cell Apoptosis
IL-21 induced more apoptosis (AnnexinV + 7-AAD lo ) in CD154 or aCD40-stimulated Sharpin cpdm B cells than Sharpin +/+ B cells, in addition to extensive but similar degrees of necrosis (AnnexinV + 7-AAD hi ) in both -aCD40 alone triggered more necrosis than CD154 ( Figure 6A). The heightened apoptosis in Sharpin cpdm B cells could be alleviated by pre-treatment with CD154, BCR-crosslinking aIgD/dex or TLR9 ligand CpG, but not other TLR ligands, which instead reduced aCD40-triggered necrosis ( Figures 6A-C and Supplementary Figures 3A, B). The viability of CD154 and IL-21-stimulated Sharpin cpdm B cells could be rescued by the pan-caspase inhibitor Z-VAD-FMK in a dose-dependent manner to a level comparable to that of Sharpin +/+ B cells ( Figure 6D and Supplementary Figure 3C). It could also be partially restored by Z-IETD-FMK and Z-LEHD-FMK, which specifically inhibits initiator caspase 8 and caspase 9, respectively ( Figure 6D and Supplementary Figure 3C), showing that Sharpin cpdm B cell apoptosis was mediated by both the mitochondria-independent and mitochondriadependent (intrinsic) apoptosis pathways. The integrity of mitochondria membranes was severely compromised by IL-21 in CD154-stimulated Sharpin cpdm B cells, but much less so in Sharpin +/+ B cells ( Figure 6E). This, together with the failure of high doses of CD154 to readily improve the mitochondria membrane integrity in either cell types, indicated that IL-21 overrode CD154 to activate the intrinsic apoptosis pathway that, if not controlled by SHARPIN, led to cell death.
Like Sharpin +/+ B cells, Sharpin cpdm B cells responded to IL-21 to activate STAT3 and STAT5, downregulate the expression of BCL2 and BCL-XL to supersede the effect of CD154 in inducing these anti-apoptosis members of the BCL2 family, and upregulate BIM, a pro-apoptotic member ( Figures 7A, B). As compared to Sharpin +/+ B cells, they also expressed comparable transcript levels of anti-apoptotic (Bcl2, Bcl2l1/Bcl2-Xl, Mcl2, which were all downregulated by IL-21) and pro-apoptotic (Bcl2l11/Bim, Bad and Bax) genes, all of which are involved in the regulation of mitochondrial membrane permeability, but showed more cytochrome C accumulation in the cytoplasm and activation of the effector caspase 3 ( Figures 7B-D), indicating that SHARPIN deficiency amplified IL-21-triggered intrinsic apoptosis without altering the expression of the BCL2 family factors, likely by boosting downstream caspase activation. Indeed, upon IL-21 induction, CD154-stimulated Sharpin +/+, and Sharpin cpdm B cells underwent similar changes in the levels of BCL2, BCL-XL, BIM, BAD, and (anti-apoptotic) BAD phosphorylation, but Sharpin cpdm B cells showed more caspase 9 activation, which depends on cytochrome C release from the mitochondria and in turn activates caspase 3 ( Figure 7E and Supplementary Figure 4). The caspase 9 activation started 2 h after stimulation and continued to increase until 48 h. At this time point, caspase 8 of the mitochondria-independent apoptosis pathway was also activated in Sharpin cpdm B cells, but not in Sharpin +/+ B cells despite its activation at early time points, when caspase 9 was marginally activated ( Figure 7E and Supplementary Figure 4). These, together with the rescue of the viability of Sharpin cpdm B cells by the inhibitor of caspase 8 or caspase 9 ( Figure 6D), showed that the SHAPRIN deficiency resulted in the synchronized activation of these two principal initiator caspases, which together triggered the irreversible apoptosis process in Sharpin cpdm B cells.
The activation and function of caspase 8 are tightly regulated by a catalytically inactive homologous factor cFLIP, which includes the longer (cFLIP L ) and shorter (cFLIP R ) forms generated through alternative splicing of the Cflar transcript in mice, with cFLIP L acting as either an inhibitor or a potentiator depending on its protein expression level while cFLIP R acting exclusively as an inhibitorboth isoforms are unstable proteins whose concentrations determine the sensitivity of cells to mitochondria-independent apoptosis (40). In Sharpin +/+ B cells, the kinetics of the caspase 8 activation, i.e., first induced as early as 2 h after stimulation, peaking at 24 h and then decreased to the pre-stimulation level at 48 h, was mirrored by the kinetics of the cleavage of the full-length cFLIP L at Asp377 (equivalent to Asp376 of human cFLIP L ) to generate p43-cFLIP L and also by the kinetics of the expression of full-length or total cFLIP L , likely reflecting a feedback control (Figures 7E, F). By contrast, full-length cFLIP L was continuously cleaved to yield p43-cFLIP L in Sharpin cpdm B cells, with its protein level briefly induced after 2 h, starting to decrease at 6 h and diminished at 48 h, showing an inverse correlation with caspase 8 activation (Figures 7E, F). The reduced cFLIP L protein expression in Sharpin cpdm B cells occurred despite normal Cflar gene transcription and was instead associated with decreased linear ubiquitination at the Met1 (linear M1-Ub) and increased polyubiquitination at Lys48 (K48 poly-Ub) of cFLIP L ( Figures 7G, H), consistent with the notion that the cFLIP L protein level is controlled by proteasome degradation in a manner dependent on its K48 poly-Ub, which is catalyzed by the ITCH/AIP4 E3 ubiquitin ligase and was suggested to be hampered by linear M1-Ub of cFLIP L (41,42). The overall linear M1-Ub was modestly reduced in Sharpin cpdm B cells, likely reflecting the residual catalytic activities of the HOIP-HOIL sub-complex formed in the absence of SHARPIN. Finally, expression of XIAP, a potent inhibitor of the initiator caspase 7 as well as caspase 9 and caspase 3, was largely comparable in Sharpin +/ + and Sharpin c p d m B cells ( Figure 7E and Supplementary Figure 4), highlighting the specificity of SHARPIN regulation of cFLIP L . Thus, SHARPIN deficiency modulates post-translational modifications of cFLIP L , abolishes the expression of this caspase 8 regulator and boosts caspase 8 activation, and synchronizes the activation of caspase 8 with that of caspase 9 in the intrinsic apoptosis pathway without affecting BCL2 family factors.

SHARPIN Deficiency and HOIPIN-8 Synergize to Promote IL-21-Induced cFLIP L Loss and Apoptosis
Like the genetic ablation of Sharpin, LUBAC inhibitor HOIPIN-8 hampered activation of NF-kB by aCD40 in B cells, starting at 50 mM (Supplementary Figure 5A). At 20 mM, HOIPIN-8 did not affect NF-kB activation and had no impact on proliferation, IL-4-directed CSR to IgG1 or IL-21-triggered plasma cell differentiation in CD154-stimulated wildtype B cells (Supplementary Figure 5B). Neither did it adversely affect the normal proliferation or CSR in Sharpin cpdm B cells ( Figure 8A). In Sharpin +/+ B cells, total linear M1-Ub was induced upon stimulation by CD154, but returned to the basal level when IL-21 was added -IL-4 had no such dampening effect ( Figure 8D). Treatment with HOIPIN-8 (at 20 mM) and SHARPIN deficiency each reduced linear M1-Ub, but together abrogated it in CD154 and IL-21-stimulated B cells ( Figure 8D and Supplementary Figure 5D), likely due to the partial and full sensitivity of the HOIP-HOIL-SHARPIN complex and HOIP-HOIL subcomplex, respectively, to HOIPIN-8 at sub-optimal concentrations. Combining the HOIPIN-8 treatment and SHARPIN deficiency also decreased, by about 75%, linear M1-Ub induced by CD154 or CD154 plus IL-4, while each alone led to a 30-40% reduction ( Figure 8D and Supplementary Figure  5D). Finally, the level of linear M1-Ub was positively correlated with the viability of B cells stimulated by CD154 or CD154 plus IL-21 ( Figure 8E and Supplementary Figure 5E).
Unlike linear M1-Ub, the total K48 poly-Ub level was barely inducible; nor was it impaired by SHARPIN deficiency or the HOIPIN-8 treatment, which did not change the expression of anti-apoptotic factor BCL-XL and pro-apoptotic factor BAD -CD154 and IL-21-stimulated Sharpin cpdm B cells did show 35% reduction of K48 poly-Ub after HOIPIN-8 treatment ( Figure 8D and Supplementary Figures 5D, E). Accordingly, the levels of K48 poly-Ub level, BCL-XL, and BAD showed weak or no correlation with B cell viability ( Figure 8E). By contrast, the expression of pro-apoptotic BIM and the level of the PARP cleavage, which is mediated by effector caspases during apoptosis, were synergistically elevated by the SHARPIN

DISCUSSION
This study has described the B cell-intrinsic role of LUBAC in promoting B cell survival from IL-21-triggered apoptosis, as relevant to the GC reaction, positive selection, and production of class-switched high-affinity Abs. Our data, as stemming from experiments involving B-Sharpin cpdm and Cd45.1 + -Sharpin +/+ / Cd45.2 + -Sharpin cpdm chimera mice, have significantly extended findings from a previous report, which showed that mb1 +/cre Hoip fl/fl mice displayed much reduced Ab responses, including those to Tindependent Ags, but did not identify the in vivo defect of mb1 +/cre Hoip fl/fl B cells underlying such impairment (31,43). By contrast, our extensive analyses of B cell proliferation, survival, differentiation (CSR and plasma cell differentiation) as well as SHM and  Figure 5D). **p < 0.01; ***p < 0.001; ****p < 0.0001; t-test. enrichment of a mutation that augments the affinity of BCRs have identified a key defect of SHARPIN-deficient B cells, i.e., reduced survival within GCs, that would provide a parsimonious explanation for the impaired T-dependent Ab response in mice with B cell-specific ablation of LUBAC. Confirming a causative role of the defective Sharpin cpdm GC B cell survival in this process would require generating mice with simultaneous Sharpin (or Hoip) deletion and enforced expression of an apoptosis inhibitor, such as a dominant negative mutant of effector caspase 3 (44), specifically in GC B cells (e.g., through Cg1-cre or Aicda-cre)genetic alterations of a single factor in the mitochondriadependent intrinsic pathway or the mitochondria-independent pathway would be less likely to maintain the full GC B cell viability. The viability of GC B cells in B-Sharpin cpdm mice may also be restored by caspase 3 inhibitor Z-VAD-FMK (or possibly Z-IETD-FMK combined with Z-LEHD-FMK), but further compromised by HOIPIN-8. Finally, Sharpin cpdm B cells were also defective in survival upon TLR ligand stimulation, likely explaining the reduced production of IgG3 and IgG2b Abs. Shortly after the discovery of IL-21 and its crucial role in plasma cell differentiation, the potent B cell killing activity of this cytokine was recognized (10,45,46), although how this killing activity is circumvented in GCs and its relevance to the Tdependent Ab response remain unaddressed. This was in part due to the difficulty of untangling its effect from the important role of IL-21 in promoting the proliferation of B cells activated by suboptimal CD40 signals (e.g., as initiated by aCD40 or low doses of CD154). As shown here, the SHARPIN deficiency decoupled the survival from the proliferation/differentiation of GC B cells in vivo and B cells activated by CD154 and IL-21 in vitro, thereby providing an opportunity to reveal the role of IL-21-triggered B cell death in the GC reaction, including the positive selection of high-affinity BCR mutants, and underlying mechanisms. CD40activated B cells would need at least two mechanisms to survive the assault by IL-21, i.e., potent CD40 signals that induce antiapoptotic BCL2 and BCL-XL to counter the effect of IL-21 in activating the mitochondria-dependent intrinsic apoptosis pathway ( Figure 9B) as well as LUBAC that inhibits caspase 8, thereby preventing synchronized caspase 8 and -9 activation, full mobilization of caspase 3 and irreversible apoptosis ( Figure 9A). The requirement for LUBAC would be important to preserve all GC B cells, irrespective of the affinity of their BCRs, particularly when such B cells are at the early phase of the iterative proliferation and selection process, i.e., completing fewer divisions. The requirement for strong CD40 signals would be consistent with the notion that GC B cell clones are differentially selected based on the quality and quantity of their contacts with Tfh cells due to their differences in BCR affinity and, therefore, the Ag uptake and presentation to Tfh cells. However, more contacts would also lead to more exposure to IL-21 produced by the same Tfh cells and, therefore, a higher threshold of B cell survival. Thus, the entwined and counteracting effects of CD154 and IL-21 would make them the prime candidate for enforcing a continuous positive selection process underpinned by the co-evolution (i.e., co-upregulation) of the death-inducing signals and the survival signal after each round of the selection. Such co-evolution would explain the stepwise improvement of the Ab affinity, a crucial aspect of the positive selection that had, however, been largely unexplored. Notably, Tfh cells were largely normal in B-Sharpin cpdm mice despite the reduced GC reaction, a phenotype similar to that of IL-21-and IL-21R-deficient mice (11,47), perhaps reflecting a compensatory role of CD154 plus IL-4activated B cells in the engagement and maintenance of Tfh cellssuch B cells would not be affected by the Sharpin deficiency in their proliferation, survival, CSR or plasma cell differentiation, as shown here. Complementing the role of Tfh cells, elevated signaling from high-affinity BCRs may also be involved in maintaining GC B cells despite being downregulated in these cells (48,49), but unlikely to induce B cell death or involved in the stepwise affinity maturation process. Finally, the Sharpin cpdm B cells activated in vitro showed normal expression of Myc and mTORC1 (our RNA-Seq data, not shown), which promote the expansion of selected B cell clones (50)(51)(52)(53), suggesting that the defective positive selection of B-Sharpin cpdm mice was not due to reduced proliferation of high-affinity clones.
The high-throughput V 186.2 -DJ H sequences reported here could be used to construct "lineage trees". Each tree would be rooted in a distinct progenitor, acquiring a dominant branch stemming from a high-affinity founder sub-clone (e.g., that carrying the G 98 !T 98 mutation) and further branched with new rounds of SHM, thereby opening a window for observation into the positive selection process and its impairment in Sharpin cpdm B cells, as lack of robust positive selection would result in overall less complex lineage trees. In addition, the exclusive occurrence of G 98 !T 98 in V 186.2 -DJ H -Cg1, but not V 186.2 -DJ H -Cm, provides additional, albeit indirect, evidence for the notion that IL-4-directed CSR to IgG1 preceded IL-21-mediated development of GCs, within which peak SHM and positive selection unfold (16). Also along this line, preactivation of Sharpin +/+ or Sharpin cpdm B cells with CD154 plus IL-4 did not make them resistant to death induced by IL-21, the production of which lags the IL-4 production by NKT cells and Th2 cells before the GC reaction (15,54), indicating that sensitivity to IL-21-triggered death is an important feature intrinsic to GC B cells. By contrast, pre-activation by TLR ligands promoted the survival of B cells activated by suboptimal CD40 signaling (e.g., as initiated by aCD40) and IL-21, likely mediating their differentiation into IL-27-producing B cells (55). Finally, the ability of strong BCR crosslinking by aIgD/ dex to endow Sharpin cpdm B cells resistance to IL-21 suggests that BCR signaling partially contributed to the GC B cell survival in B-Sharpin cpdm mice, possibly by upregulating MYC (56).
As previously shown (45,46) and further extended here, IL-21 induced apoptosis through the intrinsic pathway by modulating the expression of the BCL2 family of pro-apoptotic and antiapoptotic members that control the mitochondria membrane permeability and cytochrome C release for caspase 9 activation. Without a death domain (DD), the IL-21 receptor is unable to recruit DD-containing adaptors that initiate the death effector domain (DED)-dependent caspase 8 activation. CD40 has no DD either, in contrast to FAS (and other TNF receptor superfamily members, such as TNFR and TRAIL), raising the possibility that, in CD154 and IL-21-stimulated B cells, caspase 8 was activated by a receptor-independent apoptosis process, e.g., through a macromolecular complex nucleated by the auto-aggregation of DD-and DED-containing adaptor FADD followed by procaspase 8 recruitment and self-activation, as occurring in cancer cells (57). Such caspase 8 activation would be subsequently suppressed in Sharpin +/+ B cells due to upregulation of cFLIP, but not in Sharpin cpdm B cells, in which cFLIP level was significantly reduced, as shown here. Nevertheless, the full and sustained activation of caspase 8 might require caspase 9 activation by IL-21 and ultimately caspase 3, perhaps within a recently suggested caspase network (58) ( Figure 9C). Within such a network, positive feedback loops (e.g., those among the initiator caspases and between the initiator and effector caspases) would synchronize the activation of multiple caspases, as likely underpinned by caspase cleavage-mediated inactivation of antiapoptosis factors and activation of pro-apoptotic factors, including caspase 8 activation of BID to trigger caspase 9 (59). The rescue of the viability of IL-21-stimulated B cells by enforced BCL2 overexpression or deficiency in BIM (45,46) emphasizes the role of the intrinsic apoptosis pathway but does not rule out its collaboration with activated caspase 8 to fully activate caspase 3. Finally, IL-21 readily induced apoptosis in B cells deficient in FAS, TNFRI, or TNFRII (46), showing that IL-21 did not cause cell death indirectly by inducing these death receptors.
Although SHARPIN was originally identified as a key LUBAC component for NF-kB activation by aCD40 in Ramos B lymphoma cells (23,24), the LUBAC pathway did not play a FIGURE 9 | Illustration of LUBAC-mediated suppression of IL-21-induced apoptosis in CD154-stimulated B cells. (A) CD40 activation in B cells upregulates LUBAC-catalyzed linear M1-Ub, including that of cFLIP, which leads to cFLIP stabilization and inhibition of caspase 8 activation. LUBAC plays a minor role to the major role of RAB7 in CD40-triggered NF-kB activation in B cells. (B) CD40 activation also induces the expression of anti-apoptotic factor BCL2 and BCL-XL. IL-21, however, dampens such induction and also upregulates expression of pro-apoptotic factor BIM. The combined effects of BCL2/BCL-XL downregulation and BIM upregulation would activate pro-apoptotic BAX/BAK complex to increase the mitochondria permeability and cytochrome C release into the cytoplasm to activate caspase 9. (C) In the absence of LUBAC, caspase 8 activation would amplify a putative caspase network, within which multiple caspases, including caspase 9 and caspase 3, would be activated in a synchronized manner due to positive-feedback loops, leading to the irreversible apoptosis process. major role in NF-kB activation in CD154-stimulated primary B cells, which rather use CD40 endocytosis and the RAB7dependent endosomal pathway to activate NF-kB (39). Instead, CD154 upregulated, in addition to BCL2 and BCL-XL, and Linear M1-Ub ( Figures 9A, B), which is exclusively catalyzed by different forms of complexes containing HOIP, HOIL-1, and/ or SHARPIN and, as shown here, had a strong a positive correlation with the viability of B cells stimulated with CD154 plus different cytokine combinations. This correlation was unveiled by the partial defect of Sharpin cpdm B cells in catalyzing Linear M1-Ub and the use of a sub-optimal dose of HOIPIN-8. Furthermore, despite the strong correlation among the B-cell viability, Linear M1-Ub, and c-FLIP L levels, a definitive role of linear ubiquitinated c-FLIP L in maintaining the B cell viability remains to be proven, possibly by the specific disruption of interaction of c-FLIP L with LUBAC and, conversely, the rescue of the viability of HOIPIN-8-treated Sharpin cpdm B cells with enforced expression of c-FLIP L , preferably a form that carries a K48 mutation and, therefore, is resistant to K48 polyubiquitination and proteasome degradation. Although it has long been recognized that cFLIP plays a role in the activation and survival of immune cells (60)(61)(62)(63)(64)(65), including T cells and B cells, exactly how its level (a limiting factor in the anti-apoptosis activity) in immune cells is regulated, including by LUBAC, remains to be defined. As cFLIP also controls the pivoting between apoptosis and necroptosis, either through receptor-induced RIP kinase or receptor-independent assembly of riptosomes, whether and how cFLIP regulates necroptosis to influence the outcome of B cell viability and GC reaction needs to be explored, particularly in light of recent findings showing the unexpected role of caspase 9 in inhibiting necroptosis to promote GC B cell maintenance (66). Finally, cFLIP downregulation sensitized B lymphoma cells to TRAIL-induced apoptosis and breast cancer cells to ligand-independent but FADD-, caspase 8and caspase 9-dependent apoptosis (65,67), suggesting that HOIPIN-8, which downregulated cFLIP expression, could potentially be developed into therapeutics for B cell lymphoma as well as systemic lupus, particularly the disease caused by heightened apoptosis threshold due to lack of FAS.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: NCBI SRA data PRJNA704065, BioSample accessions: SAMN18029793, SAMN18029794.

ETHICS STATEMENT
The animal study was reviewed and approved by The Institutional Animal Care and Use Committee (IACUC) of UTHSCSA.

AUTHOR CONTRIBUTIONS
Conceptualization (JW, HY, and ZX), investigation (JW, CR, HY, and ZX), visualization (JW, TL, HZ, and HY), funding acquisition (ZX), and supervision (HY and ZX). All authors contributed to the article and approved the submitted version.

FUNDING
This work was supported by NIH AI 124172, AI 131034, AI 135599, AI 153506, and DOD BC170448 grants. The UTHSCSA Flow Cytometry Core facility is supported by NIH P30 CA054174 and UL1 TR001120, and Genome Sequencing Facility supported by NIH P30 CA054174, S10 OD021805, and CPRIT Core RP160732 grants.