E3 ubiquitin ligase Fbw7 regulates the survival of mature B cells

Abstract

Mature naïve B cells expressing B cell receptors (BCRs) of the IgM and IgD isotypes respond to antigen in secondary lymphoid organs. However, the vast majority of B cells do not undergo productive antigen encounter, and have finite life spans dependent on survival signals propagated by the BCR and the Baff receptor. Here we show that the E3 ubiquitin ligase Fbw7, is required for the maintenance of mature B cell populations in mice. BCR stimulation of B cells induced substantial apoptosis along with proliferative and growth defects upon loss of Fbw7. Analysis of B cell proteomes revealed aberrant signaling patterns including lower Bcl2 and diminished NF-κB signaling. Further, excessive accumulation of Fbw7 substrate c-Myc, increased Bim expression and loss of PI3K signaling mediated apoptosis downstream of BCR signaling. In accordance, strong pro-survival signals delivered through ectopic expression of BCL2 in B cells could largely rescue apoptotic cells in the absence of Fbw7. Overall, this study reveals an unexpected role for Fbw7 in the survival and fitness of mature B cells.

Introduction

During B cell maturation, several checkpoints ensure the integrity of the uniquely generated B cell receptor (BCR) on an individual B cell. Upon completion, mature B cells populate secondary lymphoid organs on standby with the potential to receive activatory stimuli. Signaling via both BCR isotypes, IgM or IgD, is crucial during the initiation of an immune response, but also in the absence of antigenic stimulation the BCR delivers survival signals, known as tonic BCR signaling (1). The phosphatidyl inositol-3 kinase (PI3K) signaling pathway is a key pathway supporting tonic BCR signals required for the survival of peripheral mature B cells (2). Tonic BCR signaling appears to be less dependent on canonical NF-κB signaling, however this pathway is critical for downstream signaling after antigenic BCR activation via the signaling platform consisting of Card11, Bcl10 and Malt1 (CBM complex) (3). PI3K signaling is also crucial during BCR activation, thereby controlling tonic signals and the dynamics of antigenic activation. We have recently shown that glycogen synthase kinase 3 (Gsk3), which is inhibited by PI3K signaling, regulates the metabolic needs of B cells (4). The phosphorylation of substrates by Gsk3 often primes these for degradation via the proteasome through functional recognition by the E3 ubiquitin ligase F-box and WD repeat domain-containing 7 (Fbw7) (5). Controlled proteolysis is crucial to maintain cellular homeostasis, but also to effectively activate or terminate signaling pathways. The Fbxw7 gene encodes three Fbw7 protein isoforms (α, β and γ) with different tissue distribution and subcellular localization, of which Fbw7α is the most ubiquitously expressed and predominantly found in the nucleus (5, 6). Among the substrates of Fbw7 are the transcription factors c-Myc (7, 8), Notch (9), NF-κB2 (10-12) and the pro-survival factor Mcl1 (13), which have fundamental roles in B cells. Expression of c-Myc is crucial in developing B cells and during immune responses for germinal center B cell selection and expansion (14-16). In B cells, Notch signaling appears to primarily depend on Notch2 whereas Notch1 is dispensable, however enforced expression of active Notch1 can greatly shift B cell fate to marginal zone B differentiation, which is normally attributed to Notch2 (17, 18). Deficiency of NF-κB2 (p100/p52) leads to reduced B cell numbers in the spleen and impaired germinal center formation after immunization (19, 20). Deletion of NF-κB2 in ongoing germinal centers does not affect their progression, but impairs plasma cell differentiation (21). Notably, aberrant accumulation of p100 in the absence of Fbw7 induces cell death in multiple myeloma cell lines (12). Early ablation of Mcl1 leads to a drastic block during early B cell development, but Mcl1 is also required for the survival of mature B cells (22, 23).

In immune cells, loss of Fbw7 in hematopoietic stem cells reduces self-renewal capacity and affects T and B lymphopoiesis (24). T cell-specific deletion of Fbw7 leads to increased proliferation of DP thymocytes, and ultimately malignant transformation (25). However, the role of Fbw7 in B cells has not been elucidated yet, which prompted us to investigate its function during B cell development. Here we show, that Fbw7 plays an important role in mature B cell survival by regulating NF-κB and PI3K signaling pathways, and restricting c-Myc and Bim protein levels to prevent BCR-mediated apoptosis. Regulators of BCR signaling are becoming increasingly important to modulate B cell responses in autoimmune diseases, treat B cell cancers and inhibit pro-tumorigenic B cells.

Material and Methods

Mice

Fbxw7^L/L (Jax 017563), Mb1^Cre (Jax 020505), Eμ-BCL-2-22 (Jax 002319), ^LSLYFP (Jax 006148), ^LSLMYC (Jax 020458), Cd19^Cre (Jax 006785), SW_HEL and hCD20-Cre-ER^T2 were all kept on a C57BL/6 background. A minimum of 3 (female and/or male) animals were used per experiment. Age-matched Fbxw7^+/+, Mb1^Cre/+ or co-housed Fbxw7^L/L littermates were used as controls, unless otherwise noted. All mice were kept under specific-pathogen-free conditions at the animal facility of SBP. Experimental procedures were in accordance with IACUC regulations.

Cell culture

B cells from the spleen were enriched by MACS, collecting the negative fraction of CD43 (Ly48) Microbeads (Miltenyi Biotec). Cells were cultured in RPMI 1640 (Corning) supplemented with 10% FBS (MilliporeSigma), 1x Penicillin-Streptomycin, 1x MEM Nonessential Amino Acids (Corning), 1mM Sodium Pyruvate, 2mM GlutaMax, and 55μM 2-Mercaptoethanol (Thermo Fisher Scientific). The following concentrations were used for cell stimulation: 10μg/ml anti-mouse IgM F(ab’)2 (Jackson ImmunoResearch), 10ng/ml mouse rBaff (R&D Systems), 5μg/ml anti-mouse CD40 (1C10) and 10ng/ml mouse rIL4 (Thermo Fisher Scientific). Total BM cells were cultured in IMDM with GlutaMax (Thermo Fisher Scientific) supplemented with 10% FBS (MilliporeSigma), 1x Penicillin-Streptomycin (Corning), 55μM 2-Mercaptoethanol (Thermo Fisher Scientific) and in the presence of 10ng/ml mouse rIL7 (PrepoTech).

Flow cytometry

Single cell suspensions from tissues were ACK lysed and stained in 1% FBS in PBS containing 0.05% sodium azide. Cells were gated according to size and granularity based on FSC-A and SSC-A. Doublets were excluded by FSC-A vs FSC-H gating. The following antibodies were used: B220 (RA3-6B2) APC-eFluor780, CD19 (1D3) PerCP-Cy5.5 or APC, IgM (II/41) APC or PerCP-eFluor710, IgD (11-26c) APC or FITC, CD23 (B3B4) PE, CD21 (8D9) PE-Cy7, CD5 (53-7.3) FITC, CD25 (PC61.5) PE and Streptavidin PE (Thermo Fisher Scientific). Lysozyme Biotin (GTX82960) (GeneTex). CD43 (S7) PE or FITC, FAS (Jo2) PE-Cy7 and GL7 FITC (BD Biosciences). CD5 (53-7.3) PE (Biolegend). Apoptosis assays were performed by using the PE Annexin V Apoptosis Detection Kit I (BD Biosciences) according to the manufacturer’s protocol. Cell proliferation was analyzed using the eBioscience Cell Proliferation Dye eFluor670 (Thermo Fisher Scientific). For Ca²⁺ measurements, B splenocytes were loaded with Indo-1 AM and treated with or without probenecid before acquisition (Thermo Fisher Scientific). Samples were acquired on a FACSCanto or LSRFortessa X-20 (BD Biosciences) and analyzed with Flowjo (Becton, Dickinson and Company).

Immunoblotting

Cells were lysed either in CelLytic M (Millipore Sigma) supplemented with Protease/Phosphatase Inhibitor Cocktail (Cell Signaling Technology) or in 1% SDS buffer. Immunoblotting was performed following standard procedures. PVDF membranes were blocked with 5% Milk in TBST and probed with primary antibodies overnight on a rotating platform at 4C. The following antibodies were used: phospho-Erk1/2 T202/Y204 (D13.14.4E), phospho-Akt S473 (D9E), phospho-S6 S235/236 (D57.2.2E), phospho-Gsk3α/β S21/9 (#9331), phospho-c-Myc T58/S62 (#9401), c-Myc (D84C12), Mcl1 (D35A5), Erk1/2 (L34F12), β-actin (13E5), Vinculin (E1E9V), p65 (D14E12), Notch2 (D76A6), cleaved Notch1 (D3B8), NF-κB2 p100/p52 (#4882), Bim (C34C5) and IκBα (#9242) (Cell Signaling Technology). Ship1 (Clone32) and Bcl2 (Clone7) (BD Biosciences). The following horseradish-peroxidase-coupled antibodies were used as secondaries: donkey anti-rabbit IgG and goat anti-mouse IgG (Jackson ImmunoResearch). Protein was detected on a ChemiDoc Imaging System (Bio-Rad). For protein detection on the Odyssey system (LI-COR Biosciences) the following products and workflow was used. After transferring onto PVDF, the Revert total protein stain was used to normalize the loading of protein. The membranes were blocked using Odyssey blocking buffer, stained with the primary antibody, followed by staining with an IRDye 800CW conjugated donkey anti-rabbit IgG.

Immunofluorescence & Histology

Spleens embedded in Tissue TEK O.C.T. (Sakura Finetek) were sectioned on a Microtome Cryostat HM 505 E (Microm). Sections were fixed with acetone and blocked with 5% FBS in PBS. Florescent images were acquired on an Axio Imager.M1 (Zeiss) microscope equipped with an Orca-ER (Hamamatsu) camera. SlideBook (3i) was used as the imaging software. Gimp (GNU Image Manipulation Program) was used for image editing. The following antibodies were used: B220 (RA3-6B2) APC, IgM (II/41) APC, IgD (11-26c) PE (Thermo Fisher Scientific), Moma1 Biotin (Abcam), and Streptavidin-Cy3 (Jackson ImmunoResearch). For immunohistochemical analysis, spleens were formalin-fixed and paraffin-embedded. B220 (RA3-6B2) Biotin (BD Biosciences) and Myc (Y69) (Abcam) and were used as primary antibodies and horseradish-peroxidase-coupled Streptavidin or anti-Rabbit IgG were used as secondary antibodies. TUNEL staining was performed by using the ApopTag Peroxidase In Situ Apoptosis Detection Kit (MilliporeSigma) according to the manufacturer’s protocol.

Proteome analysis

B splenocytes were purified and pellets were flash frozen in liquid nitrogen. Cells were lysed in 8M urea and 50mM ammonium bicarbonate supplemented with benzonase. After cell lysis, proteins were digested overnight in Trypsin/Lys-C Mix (Promega Corporation). Samples were then acidified with formic acid and desalted on an AssayMap Bravo liquid handling platform (Agilent Technologies). Dried samples were reconstituted in 2% acetonitrile, 0.1% formic acid and analyzed by LC-MS/MS on an EASY nLC system coupled to an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific). Mass spectra were analyzed with MaxQuant (maxquant.org) using Mus musculus protein sequence database (uniprot.org) as reference. Calculated peptide intensities were log₂-transformed and normalized across samples. Protein-level quantification and testing for differential abundance were performed using MSstats (bioconductor.org/packages/MSstats). Morpheus (software.broadinstitute.org/morpheus) was utilized to calculate z-scores and visualize expression values as a heatmap.

Immunization

Acute deletion of Fbw7 was induced with 2 to 4 doses of 1mg tamoxifen in 100μl olive oil by oral gavage. Mice were immunized i.p. with 5 or 50μg of either NP₅₃-AECM-Ficoll or NP_0.3-LPS (Biosearch Technologies) in 100μl PBS. Serum was collected on indicated days and antibody titers were determined by ELISA on plates coated with NP₂₃-BSA. For germinal center responses, mice were immunized i.p. with 50μg NP₂₀-CGG (Biosearch Technologies) in 100μl PBS. Sera was collected on indicated days and antibody titers were determined on plates coated with NP₄- or NP₂₃-BSA.

qRT-PCR

RNA was extracted with RNeasy Mini Kit (Qiagen), cDNA was synthesized with iScript cDNA Synthesis Kit and reactions were run with iTaq Universal SYBR green (Bio-Rad) on a LightCycler 96 (Roche). The following program was used: preincubation at 95C for 5 mins, 2-step amplification of 45 cycles at 95C for 10s and 60C for 40s. The following primers were used: Actin (Fwd 5′-GGCTGTATTCCCCTCCATCG-3′; Rev 5′-CCAGTTGGTAACAATGCCATGT-3′), Hprt (5′-GGGGGCTATAAGTTCTTTGC-3′ ; 5′-TCCAACACTTCGAGAGGTCC-3′), Fbxw7 (Fwd 5′-TTCATTCCTGGAACCCAAAG-3′; Rev 5′-TCCATGGGCTGTGTATGAAA-3′), Bcl2a1 (Fwd 5′-TACAGGTACCCGCCTTTGAG-3′; Rev 5′-TCCACGTGAAAGTCATCCAA-3′), Bax (Fwd 5′-TGCAGAGGATGATTGCTGAC-3′; Rev 5′-GATCAGCTCGGGCACTTTAG-3′), Bcl2l11 (Fwd 5′-GGAGATACGGATTGCACAGG-3′; Rev 5′-TCAATGCCTTCTCCATACCA-3′). Primers for mouse specific Fbxw7α, Fbxw7β and Fbxw7γ isoforms were described by Matsumoto et al. before (6).

Statistical analysis

Prism (GraphPad Software) was used for statistical analysis and p values from unpaired two-tailed tests are indicated in the figures: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Results

Peripheral B cells are decreased in the absence of Fbw7

During B lymphopoiesis in the bone marrow, lineage differentiation depends on timed expression of key transcription factors and phases of cytokine versus pre-BCR driven differentiation and expansion. To investigate the functional effects of Fbw7 in the B cell compartment, ablation of Fbw7 (Fbxw7^L/L) (24) was induced in mice during early B cell development (Mb1^Cre) (26). Two major checkpoints in the bone marrow facilitate proper BCR construction in a stepwise manner (27). First, pre-B cells express the successfully rearranged heavy chain locus of the BCR assembled with surrogate light chains (in form of a pre-BCR) to undergo clonal expansion, secondly, the completed BCR is expressed on immature B cells to monitor autoreactivity. Defects in the BCR signaling cascade are therefore frequently evident by developmental blocks at the pre-B cell stage. Fbw7-deficient B cells progressed through early developmental stages and showed no obvious signs of maturation defects in the bone marrow (Fig. 1A). Albeit pre-B cells were increased and proliferated more in response to IL7 signaling (Supplemental Fig. 1A), differentiation to immature B cells appeared normal compared to control. After immature B cells egress the bone marrow to finalize maturation and populate peripheral lymphoid organs, some return to the bone marrow as mature recirculating B cells, which were nearly absent in Fbw7-deficient mice. Correspondingly, peripheral B cells in the spleen were overall reduced in number and showed specific decreases in follicular B cells, T2 and marginal zone B cells. However, T1 cell numbers were comparable to control (Fig. 1B). Notably, Fbw7-deficient mature B cells downregulate CD23 expression, commonly used to identify B cell subsets in the spleen, but could be identified by IgM, CD21 and IgD surface expression, albeit the latter two showed increased expression levels and marginal downregulation of IgM was detectable (Supplemental Fig. 1B). The expression of BAFF receptor was normal on splenic B cells. To determine if B cells were localized normally in the periphery, we examined spleens from Fbw7-deficient mice by immunofluorescence for any abnormalities in the splenic architecture. Intact follicles with the presence of marginal zone B cells were detectable (Fig. 1C). However, follicles appeared to be smaller in size by gross analysis of the whole spleen (Fig. 1D). Following the reductions in mature B splenocytes, we determined that the third major mature B cell subset of B1 cells, an innate-like B cell compartment, was affected. Interestingly, B1 cells were drastically reduced in the peritoneal cavity, which affected both B1a and B1b cell subsets (Fig. 1E). In accordance with splenic B cell reductions, B2 cells in the peritoneal cavity were also greatly reduced.

Figure 1. Peripheral B cells are decreased in the absence of Fbw7.

(A) Absolute numbers of B cells (B220⁺) or B cell subsets in the bone marrow (BM) of Fbw7-deficient or control mice. The following B cells subsets per femur are shown: pro-B (B220⁺CD43⁺), pre-B (B220⁺CD25⁺), immature B (B220^loIgM⁺) and mature B (B220^hiIgM⁺) cells. Representative flow plot shows are shown for immature and mature B cell gates. (B) Absolute numbers of B cells (B220⁺) or B cell subsets in Fbw7-deficient or control mice in the spleen, including T1, T2, marginal zone (MZ) and follicular (FO) B cells. Representative flow plots show gate frequencies of T1 (CD21^loIgM^hi), T2/MZ (CD21^hiIgM^hi) and FO (CD21^midIgM⁺) within B220⁺ gate. (C) Immunofluorescence of the spleen shows follicles with B cells (B220⁺ or IgM⁺IgD⁺) and depicts the marginal zone (MOMA⁺ or IgM⁺IgD^lo). Scale bar, 100 μm. (D) Immunohistochemical analysis of the whole spleen for B220 expression (B cells). Scale bar, 3 mm. (E) Absolute numbers of peritoneal B2 (B220⁺CD19⁺), B1 (B220^lo/−CD19⁺), B1a (B220^lo/−CD19⁺CD5⁺) and B1b (B220^lo/−CD19⁺CD5⁻) cells and corresponding representative flow plots.

Acute deletion of Fbw7 in mature B1 and B2 cells and functional outcomes of TI-antigen immunization

The early absence of Fbw7 drastically reduced mature B cell population and may have constituted unfavorable conditions preceding the mature stage to engage in functional immune responses. Thus, inducible deletion of Fbw7 in mature B1 and B2 cells (post-B lymphopoiesis) was preferred. It has not been reported if hCD20-Cre-ER^T2 mice (28) have specific Cre expression in mature B1 cells (in addition to B2 cells), but this seemed likely due to CD20 expression on B1 cells (29). To test the efficacy of Cre deletion in mature B cells, we crossed hCD20-Cre-ER^T2 to Fbxw7^L/L animals and subsequently to ^LSLYFP mice (30) as a reporter. After administration of tamoxifen by oral gavage, we analyzed B cells in the bone marrow, spleen and peritoneal cavity after 3 weeks. Notably, YFP⁺ cells were detectable in all mature B cell compartments including B1 and B2 cells in the peritoneal cavity (Fig. 2A). Compared to control cells, acute deletion of Fbw7 recapitulated loss of mature B cells in the bone marrow and spleen. However, B1 cells were not as drastically lost as compared to early B cell deletion by Mb1^Cre, but trended lower. Using this system, we induced Fbw7 deletion and subsequently immunized mice with T cell-independent type I and II antigens, NP-LPS and NP-Ficoll, respectively to test the role of Fbw7 in B1 cell activation in vivo. Fbw7-deficient mice showed comparable early antibody responses on Day 3 for both antigens, however after 7 days serum antibodies were lower after NP-LPS immunization (Fig. 2C), but not with NP-Ficoll (Fig. 2B).

Figure 2. Acute deletion of Fbw7 regulates TI-type I antigen responses.

(A) Acute deletion of Fbw7 was induced by tamoxifen and YFP⁺ cells were analyzed in B cell compartments of the bone marrow (BM), Spleen (Sp) and peritoneal cavity (PerC) after 3 weeks. (B and C) Mice were immunized with NP-Ficoll or NP-LPS, respectively 4 days after the last dose of tamoxifen. Sera was collected on indicated days and antibody titers were determined by ELISA.

Analysis of B cell proteomes reveals defective NF-κB signaling in the absence of Fbw7

The reduction of B cells in Fbw7-deficient mice was surprising as Fbw7 has been largely conceptualized as a negative regulator of growth, proliferation and survival (5). To gain a better understanding of the specific role of Fbw7 in B cells, we analyzed the global proteome of purified B cell population from the spleen of Fbw7-deficient and control animals, which yielded more than 3500 detectable proteins in our sample set. Among the significant perturbations, several notable proteins involved in survival signaling (Bcl2 and Caspase 3) and BCR signaling (Cd79a, Card11, p65, Stim1, Ship1) were differentially regulated upon Fbw7 deficiency (Fig. 3A). Some proteins were solely detectable in the control samples (potentially due to the limit of detection), but not in the Fbw7-deficient cells, which included Cd23 and Bcl2. Notably, CD23 was already validated by flow cytometry (Supplemental Fig. 1B) and lower protein levels of Bcl2 was confirmed in Fbw7-deficient cells by immunoblot (Fig. 3B). There was also evidence of lower canonical NF-κB signaling due to lower levels of Card11 (also known as Carma1) and p65 (also known as Rela). Card11 acts upstream of the negative regulator IκBα, which sequesters p65 in the cytoplasm (31), so we tested whether p65 was lower and IκBα was more stabilized. Fbw7-deficient cells showed less degraded IκBα protein and lower p65 levels, highlighting a potential insufficient utilization of this pathway. Further, the lipid phosphatase Ship1 was drastically downregulated in the Fbw7-deficient cells (Fig. 3B). Consistent with the role of Ship1 in antagonizing Akt activity, Akt phosphorylation was heightened in Fbw7-deficient cells (Fig. 3C). Active Akt mediates inhibitory phosphorylation of Gsk3, which then restricts Gsk3-mediated phosphorylation of c-Myc leading to its stabilization. We found increased phosphorylation of Gsk3α and Gsk3β isoforms and heightened c-Myc protein levels. However, the phosphodegron motif in c-Myc protein was present highlighting the stabilization of c-Myc predominantly through insufficient protein degradation in the absence of Fbw7.

Figure 3. Loss of Fbw7 lowers NF-κB signaling and stabilizes c-Myc.

(A) Freshly isolated primary B cells from the spleen of Fbw7-deficient or control mice were analyzed by mass spectrometry. Heat map analysis shows relative abundance of indicated proteins among triplicates. Black circles represent undetectable protein in that sample. (B) Validation of differentially regulated proteins and their respective pathways by immunoblot. (C) Analysis of c-Myc protein levels downstream of the Akt-Gsk3 axis.

Fbw7 regulates BCR-mediated survival, growth and proliferation

Two major signaling pathways have been implicated thus far for B cell homeostasis in the spleen, namely Baff receptor signaling and tonic BCR signaling (32). Thus, we treated B cells ex vivo in the presence of Baff, anti-IgM or left without stimulation in media. Notably, the viability of Fbw7-deficient cells was greatly reduced without any stimuli within one day (Fig. 4A). Addition of Baff increased the viability slightly compared to media alone. The addition of BCR stimulation via anti-IgM had drastic consequences on Fbw7-deficient cells, where the majority of cells underwent apoptosis and were unable to grow (Fig. 4A, Supplemental Fig. 3B). Importantly, this cell death preceded the BCR-induced proliferation, which was also diminished in the remaining viable Fbw7-deficient cells (Supplemental Fig. 3B). The majority of these cells were unable to undergo cell division with only a small fraction of cells dividing in contrast to control cells.

Figure 4. Fbw7 regulates BCR-mediated survival signaling.

(A) Purified B cells from the spleen of Fbw7-deficient or control mice were stimulated in the presence of anti-IgM (10μg/ml), Baff (10ng/ml), anti-CD40 (5μg/ml), IL4 (10ng/ml), indicated combinations or media alone. Regulation of apoptosis was assessed with anti-Annexin V and 7AAD dye before stimulation (Day 0) or after 1 day in culture. (B) Calcium mobilization is presented by the ratio of Ca²⁺-bound Indo to Ca²⁺-unbound Indo. After acquisition of baseline for 20 seconds anti-IgM was added to purified B cells from Fbw7-deficient or control mice. (C) qRT-PCR analysis of Fbxw7 and Bcl2a1 (A1) mRNA abundance in Fbw7-deficient versus control cells after 1 hour of stimulation with anti-IgM or expression of Bax and Bcl2l11 (Bim) after 4 hours. (D) Purified B cells were stimulated with anti-IgM for the indicated times and probed for phospho-Akt and phospho-S6. Measurements show relative fluorescence units (RFU) normalized for total protein per lane loaded. (E) Purified B cells were stimulated with anti-IgM for 1 hour and assessed for IκBα degradation. (F) Purified B cells were stimulated with anti-IgM for 16 hours and probed for phospho-Gsk3α/β and c-Myc. (G) Purified B cells were stimulated with anti-IgM or Baff for 16 hours. Lysates were probed for p100/p52 (NF-κB2) and Bim.

Notably, Fbw7-deficient cells did not proliferate or grow in the presence of Baff or media alone, which is not mitogenic under normal conditions, but may have supported proliferative and growth signals due to higher signaling of Akt in the absence of Fbw7. To rule out that the difference in survival was due to aberrant signaling thresholds in the absence of Fbw7, we examined the induction of apoptosis in the presence of varying concentrations of anti-IgM for 24 hours. Deficiency of Fbw7 resulted in high apoptosis despite the concentration of IgM in the cell culture media (Supplemental Fig. 3A). Addition of Baff, anti-CD40 or IL4 was able to partially rescue apoptosis in Fbw7-deficient IgM-stimulated cells (Fig. 4A). Early adverse effects were evident in the calcium mobilization, which was diminished in Fbw7-deficient cells (Fig. 4B), although activation with anti-IgM induced comparable phosphorylation of p-Akt as early as 5 minutes (Fig. 4D, Supplemental Fig. 3E). However, activated Akt signaling and downstream effector S6 was diminished after 1 hour post-stimulation and persisted lower. Also, canonical NF-κB signaling was defective after BCR-stimulation (Fig. 4E), however non-canonical NF-κB signaling appeared to be normal when stimulated with Baff (Fig. 4G). The strong induction of c-Myc protein downstream of BCR signaling was further substantiated in Fbw7-deficient cells, although Gsk3 protein was less inhibited by phosphorylation (Fig. 4F). Notably, pro-apoptotic members of the Bcl2 family Bax and Bim (Bcl2l11) were transcriptionally increased (Fig. 4C) and Bim protein was enhanced in Fbw7-deficient cells (Fig. 4G).

Enforced c-MYC expression induces apoptosis downstream of BCR signaling

To test if the excessive accumulation of c-Myc can induce apoptosis in B cells, we utilized mice with B cell-restricted human c-MYC (^LSLMYC) (16) under the control of Cd19^Cre/+ (33). Purified B cells from the spleen of single copy (^LSLMYC;Cd19^Cre/+), double copy (^LSLMYC/MYC;Cd19^Cre/+) or control (^LSLMYC) mice were cultured in the presence of anti-IgM, anti-CD40, a combination of the two or media alone. After 1 day in culture the viability was comparable in all conditions except double copy c-MYC cells had increased apoptosis when left untreated (Fig. 5A). Apoptosis was further enhanced after 3 days without stimulation. Notably, BCR stimulation induced apoptosis in a dose dependent-manner of c-MYC expression, whereas CD40 signaling was comparable and able to rescue some of the survival loss seen with IgM stimulation. Although c-Myc levels are higher in mice carrying double copies of ^LSLMYC compared to single copy, Fbw7-deficient cells had comparable c-Myc protein induction (Fig. 5B). This elevation of c-Myc protein appears to occur in most Fbw7-deficient cells within the splenic follicles (Fig. 5C).

Figure 5. Excessive c-MYC protein induces BCR-mediated apoptosis.

(A) Purified B cells from the spleen of mice with single (^LSLMYC;Cd19^Cre/+), double (^LSLMYC/MYC;Cd19^Cre/+) or unexpressed (^LSLMYC) c-MYC transgene were stimulated in the presence of anti-IgM, anti-CD40, anti-IgM plus anti-CD40 or media alone. Regulation of apoptosis was assessed with anti-Annexin V and 7AAD dye before stimulation after 1 day or 3 days in culture. (B) Expression of c-Myc levels by immunoblot in freshly isolated B cells from Fbw7^−/− (KO) mice compared to single (M/+) and double (M/M) c-MYC transgene mice determined by immunoblot. (C) Immunohistochemical analysis of spleens from Fbw7-deficient and control mice for c-Myc expression. Scale bar, 300 μm.

Enforced BCL2 expression overcomes the apoptotic fate of Fbw7-deficient B cells

We hypothesized that the decrease of mature B cells in Fbw7-deficient mice may occur due to apoptosis in vivo. Notably, TUNEL staining revealed increased numbers of apoptotic cells in splenic follicles of Fbw7-deficient mice compared to controls (Fig. 6A). Thus, we posited the question if strong pro-survival signals by Bcl2 can overcome the apoptotic fate and deficits of Fbw7-deficient B cells. For this, we introduced a B cell-restricted transgenic human BCL2 (Eμ-BCL-2-22) (34) to Fbw7-deficient animals and analyzed the effects. In the spleen, ectopic BCL2 expression in Fbw7-deficient mice surpassed B cell numbers of wildtype controls, albeit still lower than BCL2 transgene controls, nonetheless BCL2 efficiently rescued the loss of B splenocytes in the absence of Fbw7 (Fig. 6B). In the peritoneum, the BCL2 transgene increased B1 cells (Fig. 6B) compared to Fbw7 deficiency alone, but could not reinstate the typical levels of B1 cells in the peritoneum of control mice. TUNEL staining was very low in the spleen of Fbw7-deficient mice with BCL2 and comparable to control BCL2 mice (Fig. 6C). Interestingly, ex vivo Fbw7-deficient B cell cultures were also more viable in the presence of BCL2 after 1 day post-IgM stimulation (Supplemental Fig. 4B) than Fbw7-deficiency alone (Fig. 4A). Overall, apoptosis appeared to be a major outcome in Fbw7-deficient cells, which could be partially rescued by BCL2.

Figure 6. Enforced BCL2 expression overcomes the apoptotic fate of Fbw7-deficient B cells.

(A) TUNEL staining of spleen sections from Fbw7-deficient or control mice. Scale bar, 200 μm (B) Left: Absolute number of B cells in the spleen (Sp) of Fbw7-deficient or control mice with or without ectopic expression of BCL2. Right: Absolute number of B1 cells in the peritoneal cavity (PerC) of these mice. (C) TUNEL staining of spleen sections from mice with Fbw7-deficiency carrying BCL2 or BCL2 controls. Scale bar, 200 μm

Discussion

In this study, we show for the first time, to our knowledge, that Fbw7 regulates the survival of mature B cells. Early ablation of Fbw7 during B lymphopoiesis manifests in reduced mature B cells in the bone marrow, spleen and peritoneal cavity. Follicles in the spleen show increased apoptosis in the absence of Fbw7. Importantly, Fbw7-deficient B cells undergo heightened apoptosis ex vivo without any stimuli highlighting their lower fitness compared to wildtype cells (Fig. 4A). Our global proteome analysis revealed deficiencies in the NF-κB pathway and lower Bcl2 levels in the absence of Fbw7 (Fig. 3A, B). Importantly, these cells have a significant elevation of c-Myc protein, which is insufficiently degraded in the absence of Fbw7 (Fig. 3C). BCR-stimulation further substantiates c-Myc levels in Fbw7-deficient B cells, but also increases a strong albeit lower induction of c-Myc in the wildtype cells (Fig. 4F). The relative level of c-Myc protein appears to be mirrored in Bim levels (by mRNA and protein), which has been shown to mediate c-Myc-induced apoptosis (35). Importantly, mature B cells that are lost in the spleen and bone marrow of Fbw7-deficient mice are significantly accumulated in Bim-deficient mice (Mb1^Cre), suggesting the mechanistic contribution of Bim to apoptosis in these cell stages (36). Interestingly, increased levels of ectopic c-MYC expression (single or double copy) in B cells induced specifically BCR-dependent apoptosis, but not via CD40 signaling, which is integral during the germinal center response (Fig. 5A). Although, elevated c-MYC protein induces apoptosis downstream of BCR signaling, this effect occurred later than in Fbw7-deficient cells. The specific makeup of Fbw7-deficient cells based on aberrant pathway signaling, including PI3K, Erk and NF-κB signaling suggests that these additional defects are contributing to the apoptotic fate. Fbw7-deficient cells present initially with higher PI3K signaling, probably due to secondary effects and downregulation of the negative regulator Ship1, however the PI3K pathway is diminished post-BCR stimulation (Fig. 4D). This defect may play a crucial role in the apoptotic fate as excessive c-Myc levels require supplementation by PI3K signaling to enable proliferative and growth potential, and to potentially prevent c-Myc-induced apoptosis (37). Overall, the combined accumulation of c-Myc and Bim, and the lower levels of Bcl2, PI3K, Erk and NF-κB signaling induce apoptosis in Fbw7-deficient mature B cells. Further, the partial rescue of apoptosis by BCL2 transgenic expression in the absence of Fbw7 reveals the missing counterbalance to pro-apoptotic induction (in part by Bim) in Fbw7-deficient B cells.

B1 cells were drastically reduced with early deletion of Fbw7, but less affected after acute deletion. B1 cells are innate-like cells that form mostly during fetal development, before conventional B2 cells, and bear a restricted BCR repertoire geared to self-antigens and mostly mucosal pathogens (38). Unlike follicular or marginal zone B cells, B1 cells depend on self-renewal and are not continuously replenished, which may explain their drastic disappearance in Fbw7-deficient animals. BCR signaling strength is an important contributor to fate decisions. B1 cells depend on strong BCR signaling during development contrary to follicular and marginal zone B cells (39). The IgM-induced apoptosis in Fbw7-deficient B cells may be indicative for a potential role of Fbw7 regulating BCR signaling in early B1 cells. The strong positive selection during development may lead to the apparent loss in this compartment. Also, the acute deletion of Fbw7 shows less dependencies in the homeostatic role and T cell-independent antigen activation of B1 cells, highlighting potential developmental defects. Notably, BCR transgenic (HEL⁺) B cells are not lost the in absence of Fbw7 (Supplemental Fig. 2A). Compared to wildtype cells, HEL⁺ cells have lower tonic BCR signaling (40) and are nearly absent within the B1 cell pool (41). It is not clear, what source of BCR signaling is mediating B cell apoptosis in vivo in the absence of Fbw7, but it is possible that B cells with heightened (antigen-independent) tonic BCR signaling succumb.

In B cell malignancies, missense mutations of FBW7 have been identified in ~3.4% of ALL (42), ~1.6% of DLBCL (43) and ~2.5% of CLL (44) patients. Notably, we found increased pre-B cell proliferation in the absence of Fbw7, which may reflect some of its roles in their malignant counterparts in ALL. At this stage, pre-B cell expansion is controlled in part by c-Myc, which must be suppressed to halt proliferation and enable further differentiation (45). Fbw7-deficient pre-B cells potentially are unable to effectively remove c-Myc protein, and thereby exit the cell cycle delayed and accumulate in number. DLBCL arise from malignant germinal center B cells (46). Low FBW7 expression in DLBCL patients has shown to correlate with poor survival outcome (47). Surprisingly, the germinal center response was intact upon Fbw7 deficiency and appeared normal, aside from lower IgG1 antibody secretion (Supplemental Fig. 2B). Although, c-Myc is involved in germinal center expansion, the dynamics of germinal center B cells are much more complex and involve a multitude of pathways and selection mechanisms. It appears that BCR stimulation with real antigen can effectively stimulate Fbw7-deficient B cells in vivo, which then are able expand and form functional germinal center B cells. Notably, CD40 and IL4 signaling (typically provided by T follicular helper cells in the germinal center) can partially rescue the survival defect of Fbw7-defcient cells after IgM-stimulation (Fig. 4A). Further, Fbw7-deficient B cells proliferate normally when stimulated with CD40 and IL4 (data not shown). Thus, the difference in early, mature and germinal center B cell proliferation upon Fbw7 loss might be due to the distinct activated stimulatory pathways among these cell population. The cellular origins for CLL remain unclear, but functionally CLL cells have been ascribed to B1 cells (48). Notably, recent work addressed the mechanistic role of FBW7 in CLL, showing that functional loss of FBW7 elevates NOTCH1 protein levels and its associated signaling pathway (49). Although, we found Notch proteins to be induced in Fbw7-deficient B cells (Supplemental Fig. 3D), enhanced marginal zone B cell differentiation or proliferative responses appeared not to be induced. Thus, it would be important to study the molecular roles of Fbw7 in specific B cell subsets with context-dependent constituents (i.e. cooperating mutations and microenvironmental cues) to directly relate these findings to B cell pathogenesis, specifically in patients with FBXW7 mutations.

Key Points.

• Loss of Fbw7 during B cell development reduces mature B1 and B2 cells.

• Fbw7 controls survival pathway signaling to prevent apoptosis in mature B cells.

• Ectopic expression of BCL2 partially overcomes apoptosis upon Fbw7 deficiency.