Nfkbie-deficiency leads to increased susceptibility to develop B-cell lymphoproliferative disorders in aged mice

Aberrant NF-κB activation is a hallmark of most B-cell malignancies. Recurrent inactivating somatic mutations in the NFKBIE gene, which encodes IκBε, an inhibitor of NF-κB-inducible activity, are reported in several B-cell malignancies with highest frequencies in chronic lymphocytic leukemia and primary mediastinal B-cell lymphoma, and account for a fraction of NF-κB pathway activation. The impact of NFKBIE deficiency on B-cell development and function remains, however, largely unknown. Here, we show that Nfkbie-deficient mice exhibit an amplification of marginal zone B cells and an expansion of B1 B-cell subsets. In germinal center (GC)-dependent immune response, Nfkbie deficiency triggers expansion of GC B-cells through increasing cell proliferation in a B-cell autonomous manner. We also show that Nfkbie deficiency results in hyperproliferation of a B1 B-cell subset and leads to increased NF-κB activation in these cells upon Toll-like receptor stimulation. Nfkbie deficiency cooperates with mutant MYD88 signaling and enhances B-cell proliferation in vitro. In aged mice, Nfkbie absence drives the development of an oligoclonal indolent B-cell lymphoproliferative disorders, resembling monoclonal B-cell lymphocytosis. Collectively, these findings shed light on an essential role of IκBε in finely tuning B-cell development and function.

The NF-κB transcription factor family members are maintained as inactive homo-or heterodimers in the cytoplasm by IκB, including the classical IκB protein family: IκBα, IκBβ and IκBε (ref. 1 ). IκBε is an inhibitor of inducible NF-κB activity, which traps the Rel proteins in the cytoplasm at the resting state 1, 18,19 . Its expression is higher in primary murine splenic B cells than in T cells and has a specific role in the late retro-control of B-cell response to external stimuli, including signaling through the B-cell receptor (BCR) and Toll-like receptors (TLRs) [20][21][22] . IκBε is more expressed in IgM+ B cells than in IgG+ B cells and presents a slower turnover than IκBα and IκBβ, following in vitro activation of primary splenic B cells with LPS or a combination of anti-IgM and CD40L (ref. 21 ). IκBε shows preferential specificity for NF-κB Rel homodimers, with respect to other NF-κB transcription factor subunits 21 . IκBε-deficient B cells present higher expression and activity of Rel at early time points of activation with IgM and IL4 compared to WT cells [20][21][22] .
To get more insights into the role of NFKBIE in normal and malignant B-cell differentiation, we studied Nfkbiedeficient mice and showed that the loss of Nfkbie results in marginal zone B (MZB) and B1 cells expansion, and a higher sensitivity to T-cell-dependent and -independent stimulation. We also show that Nfkbie deficiency cooperates with mutant MYD88 signaling and causes enhanced B-cell proliferation. In aged mice, Nfkbie absence drives development of an oligoclonal indolent Bcell lymphoproliferative disorders, resembling monoclonal B-cell lymphocytosis (MBL).

Materials and methods
Additional information can be found in the Supplemental Methods.

Mice
Inactivated Nfkbie allele on a mixed Sv129xDBA-2xC57BL/6 J background has been described previously; 23 >20 back-crosses were performed on the C57BL/6 J background to give rise to a pure congenic Nfkbie −/− C57BL/6 J strain. Mutant mice in this pure C57BL/6 J background were housed in the animal facility of the Gustave Roussy Institute (SCEA, Gustave Roussy, Villejuif, France), crossed with WT C57BL/6 J mice and agematched animals of selected genotypes were sacrificed at the indicated times. Bone marrow transplantation were performed as described 24,25 . Animal experiments were conducted in compliance with the Gustave Roussy Institutional guidelines and authorized by the Direction Départementale des Services Vétérinaires du Val de Marne.

Mice immunization
For the analysis of germinal center (GC) formation in Tcell-dependent immune response, Nfkbie-deficient and WT age-matched 2-3-month-old mice were immunized with sheep red blood cells (SRBCs; 1 × 10 8 cells per mouse) via intraperitoneal injection. Mice were sacrificed 10 days after immunization and spleens were analyzed for GC B cells and plasma cells.
For in vivo TLR9 activation, CpG ODN (Invivogen) was administered intraperitoneally at a dose of 333 μg per mouse in 200 μL phosphate-buffered saline. Mice were sacrificed 10 days after immunization; inguinal lymph nodes and spleens were analyzed for GC B cells.

In vitro generation of induced GC B cells
Splenic untouched resting murine B cells were first isolated from 2 to 3-month-old animals using the mouse CD43 (Ly-48) B-cell isolation kit (MicroBeads, Miltenyi Biotec) according to the manufacturer's instructions (yield > 90% CD19+B220+). MZB and FOB cells were flow-sorted as described in the "Cell isolation and culture" section. 40LB-cells expressing CD40L and producing BAFF were a kind gift of Pr. D. Kitamura 26 and cultured in DMEM media (Invitrogen) with 10% FBS and penicillin/ streptomycin.

MBL diagnosis
The appearance of a B220lowCD19+CD5+ population and its percentage among peripheral blood mononuclear cells (PBMCs) were used for initial MBL-like disease diagnosis. The diagnostic criterion for MBL in mice was arbitrary defined as the appearance of oligoclonal/monoclonal B220lowCD19+CD5+ cells, constituting over 5% of PBMCs without a detectable increase in total white blood cell count.

Retroviral infection and in vivo cell transfer
MYD88WT and that containing L265P mutation (MYD88L265P) cDNA were synthesized by GenScript and subcloned into MSCV-eGFP backbone. Viral particles and transduction procedures were described previously 3 .

Statistical analysis
Statistical significance of differences between the results was assessed using a two-tailed unpaired Student's t-test with Welch's correction, performed using Prism (GraphPad software, version 5.03). Statistically significant p values: *p values < 0.05; **p values < 0.01 and ***p values < 0.005. Error bars displayed throughout the paper represent s.e.m. or s.d. as indicated in figure legends. No statistical method was used to predetermine sample size. No blinding and no randomization of samples were applied. No data was excluded.

Results
Nfkbie −/− mice show amplification of MZB and B1 B cells Initial analyses of immune cells in Nfkbie −/− mice on a mixed genetic background showed only very subtle phenotypic consequences on thymic lymphoid T-cell subsets 23 . Our analysis showed that, compared to 5-8month-old WT, aged-matched Nfkbie −/− mice on a pure C57BL/6 J background presented an increase in cell number and frequency of a population expressing low levels of the B220 marker (B220low) in spleen, thymus, lymph nodes, peritoneal cavity, and peripheral blood (Fig. 1a). Detailed immunophenotyping in spleen and peritoneal cavity was compatible with a B1 B-cell population: CD19+, B220low/−, CD43+, CD23−, CD21−, CD93−, IgDlow, and IgM+ ( Supplementary Fig. 1a).
Together, these data indicate that the absence of Nfkbie affects mature B-cell subsets differentiation and leads to expansion of MZB and B1a B cells. These B-cell subsets are known to mediate the innate functions of the B lineage. Both populations are particularly sensitive to variations in NF-κB activity and strongly influenced by BCR specificity and strength of signaling [28][29][30] . We then evaluated the frequency and numbers of B1 Bcell progenitors (Lin−CD93+CD19+B220-/low) in mutant bone marrow 31 . Significantly higher frequency of Lin−CD93+CD19+B220−/low cells in 2-month-old Nfkbie −/− mice were observed compared to those of WT or Nfkbie +/− mice, with a trend for Nfkbie −/− B1 progenitor cell number increases (Fig. 2a).
To check the balance between B1 and B2 cells, we analyzed their proportions in the spleen and peritoneal cavity. Percentages of total CD19+ B cells increased significantly in peritoneal cavity but remained stable in the spleen of Nfkbie −/− mice, as compared to WT animals (Fig. 2b). Percentages of splenic and peritoneal B2 cells (CD19+B220+) were also similar between both genotypes (Fig. 2b); however, percentages of B1 cells were significantly higher in Nfkbie −/− mice. These data indicate that there is no shift toward the B1 cell lineage in mutant mice and that the increase in B1 cell numbers might result from the increase of the B1 B-cell progenitor cells.
we investigated ex vivo proliferation and apoptosis. No significant differences in the proportion of Ki67+ or viable MZB, FoB, B1 B cells between Nfkbie −/− and WT mice were detected (Supplementary Fig. 3a, b). As the frequency and cell numbers of splenic transitional B cells were equivalent between Nfkbie −/− and WT mice, these findings suggest that Nfkbie deficiency biases the differentiation of transitional B cell into MZB cell fate.
Overall, these data indicate that Nfkbie is important for follicular versus MZB cell fate decision and that its loss may affect the size of the B1 B-cell progenitor compartment.

Biased differentiation toward MZB cell and expansion of B1 B-cell subsets in absence of Nfkbie is cell-autonomous
To investigate whether Nfkbie deficiency-associated changes were cell-autonomous, we performed competitive bone marrow reconstitutions ( Fig. 3a for scheme). FACS analysis in peripheral blood showed that recipients of WT CD45.2+ cell had a stable reconstitution with 30% donor cells, whereas there was a steady increase in the percentage of donor cells in recipients of Nfkbie +/− and Nfkbie −/− CD45.2+ cells that reached statistical significance at 22 weeks with over 40% of donor cells (Fig. 3b).
Analysis of the distribution of cell lineages showed that Nfkbie +/− and Nfkbie −/− LSK (Lin−Sca1+Kit+) cell transfer resulted in a greater proportion of CD19+ B cells than that of WT cells in recipient mice (Fig. 3c). Consistent with our initial observations, at 22 weeks Nfkbie −/− cells contributed to a significantly higher proportion of splenic MZB cells and B220lowCD19+ B1 cells than WT ones (Fig. 3d). These data establish the cell-autonomous nature of the Nfkbie-deficient-associated phenotypes.
To assess the long-term impact of Nfkbie deficiency, we monitored monthly a cohort of ten Nfkbie −/− , ten Nfkbie +/− mice, and ten control WT mice for the presence of the B220lowCD19+ population in blood. After 12 months, five out of ten (50%) Nfkbie −/− mice started to develop MBL (defined as over 5% of B220lowCD19+ within PBMCs; Fig. 4a). Cytological analysis of blood smears from MBL developing Nfkbie −/− mice showed larger lymphocytes with more abundant cytoplasm around the nucleus (Supplementary Fig. 4a). Nfkbie −/− mice that developed MBL-like disease showed a clear expansion of CD5+ cells among the B220lowCD19+ population (Fig. 4b). Analysis of IGH gene rearrangements revealed an oligoclonal pattern of the

Nfkbie-deficient B1a cells hyper-proliferate in response to TLR stimulation and exhibit enhanced NF-κB signaling
We then investigated the impact of Nfkbie deficiency on the proliferative response of splenic and peritoneal B-cell subsets to T-cell independent stimuli, such as TLR agonists. These stimuli are known to induce NF-κB activity in B cells 1,8,[20][21][22]32 . FACS-sorted splenic B-cell subsets, FoB (CD19+B220+CD23+CD21+), MZB (CD19+B220 +CD23lowCD21hi), and B1 (CD19+B220low) cells were stimulated with anti-IgM, LPS, or CpG oligodeoxynucleotides, and cell division was measured by CFSE dilution and cell count after 72 h of culture. We found that splenic B1 (Fig. 5a) and MZB ( Supplementary Fig. 5a) cells lacking Nfkbie displayed increased proliferation rate in response to LPS and CpG compared with WT B cells. Furthermore, Nfkbie −/− MZB cells showed significantly higher proliferation rate in response to IgM stimulation compared to WT MZB cells ( Supplementary Fig. 5a). There was no significant difference in FoB cells proliferation in response to LPS, CpG, and anti-IgM between mutant and WT cells ( Supplementary Fig. 5b). As WT B1 cells, Nfkbie −/− B1 cells did not respond to anti-IgM stimulation (data not shown). This confirms that MZB and B1 and to a lesser extent FoB cells require IκBε to limit their proliferative response in a B-cell-intrinsic manner.
We next focused our analysis on B-cell subsets from the peritoneal cavity. Peritoneal B2 B cells were sorted as CD19+B220+. Similarly to splenic FoB cells, increased proliferation of Nfkbie −/− peritoneal B2 B-cells in response to stimuli did not reached statistical significance when compared with WT cells (Supplementary Fig. 5c).
Since IκBε deficiency leads to enhanced NF-κB activation, we hypothesized that Nfkbie-deficient B1a B cells would be differentially sensitive to NF-κB inhibition. We thus treated Nfkbie −/− and WT B1a B cells with increasing concentration of the NF-κB inhibitor IT-901 (ref. 34 ), in presence or absence of LPS stimulation. Cells were analyzed for the growth inhibition and apoptosis induction. We found that IT-901 inhibited cell growth of both Nfkbie −/− and WT B1a B cells in a dose-dependent manner; however, the IC50 value was higher for Nfkbie −/− B1a B cells (3 µM) compared to that of WT B1a B cells (2.4 μM; Fig. 5f). At 4 µM, no significant difference in cell growth inhibition was observed (Fig. 5f). Analysis of apoptosis showed a significant difference in survival between Nfkbie −/− and WT B1a B cells when cells where vehicle treated or IT-901 treated at 0.5 µM, 1 µM, and 2 µM, whereas such difference was lost at 4 µM and 8 µM (Fig. 5g). These data show that Nfkbie −/− B1a B cells are generally less sensitive to NF-κB activity inhibition than WT B1a B cells.
Overall these data demonstrate that IκBε is critical to limit B-cell proliferation in response to stimuli through the control of p65 phosphorylation and cRel nuclear translocation. They also establish a link between IgM expression levels and proliferation of Nfkbie-deficient B cells.
Nfkbie deficiency enhances GC B-cell proliferation GC reaction is necessary for maturation of the humoral immune response, including production of high-affinity plasma cells and memory B-cells. We explored the impact of Nfkbie absence in GC by immunizing mice with SRBCs. FACS analysis revealed an increase in both percentages and absolute cell numbers of GC B cells in Nfkbie −/− mice compared with those of WT mice (Fig. 6a). Frequency of follicular T-helper cells remained similar in both genotypes (data not shown). In Nfkbie −/− mice, IgG1+ GC B cells were significantly increased and IgM+ GC B cells were significantly decreased (Fig. 6b). Percentages but not absolute cell numbers of plasmablasts/plasma cells were reduced in Nfkbie −/− mice compared to those of WT mice (Fig. 6c). Nfkbie −/− mice had significantly higher percentages and cell numbers of GL7-IgG1+ memory B cells (Fig. 6d). These data suggest that IκBε plays a role in the generation/proliferation of GC B cells, isotype switching, and terminal differentiation of GC B cells.
We further investigated iGCB cell number changes in this system by analyzing cell cycle and apoptosis. iGCB cells derived from Nfkbie −/− FoB cells displayed a (see figure on previous page) Fig. 5 In B cells, IκBε sets the proliferation threshold in response to TLR or T-dependent antigen stimulations. a Sorted splenic B1 cells were labeled with CFSE and cultured in the presence of LPS or CpG. Cell division was measured by CFSE dilution at 72 h. Percentages of cells that underwent at least one division and cell numbers at 72 h are shown. Each symbol represents one mouse (n = 3). b Sorted peritoneal Nfkbie +/+ B1a, Nfkbie −/− B1a IgMhi, and Nfkbie −/− B1a IgMlow cells were labeled with CFSE, cultured in the presence of LPS or CpG, and analyzed as in a. Each symbol represents one mouse. Nfkbie +/+ (n = 4), Nfkbie +/− mice (n = 5), and Nfkbie −/− mice (n = 5). c Sorted peritoneal Nfkbie +/+ B1a, Nfkbie −/− B1a IgMhi, and Nfkbie −/− B1a IgMlow cells were cultured in complete medium without any stimulant for 24 h. Spontaneous cell death was assessed by sytox staining. Each symbol represents one mouse. Nfkbie +/+ (n = 4), Nfkbie +/− mice (n = 5), and Nfkbie −/− mice (n = 5). d significant increase of S and G2/M phases as compared to iGCB cells derived from Nfkbie +/+ FoB cells (Fig. 7d). No such difference was observed for MZB cells (Fig. 7d).
Analysis of the apoptosis rate showed no difference between Nfkbie −/− and Nfkbie +/+ in both FoB and MZB cells subsets (Fig. 7e).  To demonstrate the direct impact of Nfkbie deficiency on proliferation via cell cycle, we used a CRISPR/Cas9 approach and successfully generated three knockout Nfkbie clones of the pro-B-cell line Ba/F3 ( Supplementary  Fig. 6a). We confirmed the knockout of Nfkbie gene at the protein level in the three clones ( Supplementary Fig. 6a). Proliferation survival and cell cycle analysis were performed in control and edited cells in the presence of IL-3 at 24 h, 48 h, and 72 h. Nfkbie knockout clones exhibited increased proliferation when compared to WT cells ( Supplementary Fig. 6b). Nfkbie knockout clones showed a decrease in the percentages of cells in G0/G1 phase, and an increase in S and G2/M phases as compared to controls ( Supplementary Fig. 6c). Analysis of apoptosis failed to detect any changes between knockout and WT cells ( Supplementary Fig. 6d). Altogether, these data indicate that the effect of Nfkbie deficiency on proliferation of Ba/ F3 cells is due to an increased cell cycle entry and is apoptosis independent.
Nfkbie loss cooperates with MYD88 signaling MYD88 codes for an proximal signaling adaptor downstream of IL-1R and mammalian TLRs. Mutations in MYD88, essentially the missense p.L265P mutation, are observed in a wide range of B-cell malignancies 8,10,13,[35][36][37][38][39] . They have been shown to activate the JAK and NF-κB pathways 36 . Since both NFKBIE and MYD88 mutations may control TLR signaling and co-occur in human B-cell malignancies (Table 1), we used mutant MYD88 to document the interactions between MYD88 and Nfkbie mutations in primary B cells. Purified WT or Nfkbie −/− follicular (CD19+B220+CD23+CD21+) B cells were activated with anti-IgM and anti-CD40, transduced with MYD88 (WT) or MYD88L265P (LP) retrovirus, and cultured with anti-CD40 for 36 h and the proportion of fluorescent cells was monitored in the culture, as described 40 (Fig. 8a for scheme). As already reported 40 , expression of mutant MYD88 conferred a growth advantage over B cells expressing WT constructs. However, ectopic expression of MYD88LP conferred a significant growth advantage to the Nfkbie-deficienttransduced cells in comparison to other genotypes (Fig. 8b). Nfkbie-deficient-MYD88LP-transduced cells showed an increased cell number compared to other conditions (Fig. 8c). This indicates that targeting the TLR-NFKB axe endows the double-mutant cell with a growth advantage over the other cells.
MYD88 is the canonical adaptor for signaling pathways downstream of the members of the TLR family, including TLR9. Genetic or pharmacological inactivation of TLR9 was shown to inhibits MYD88L265P-induced proliferation in murine and human B-cell lines 40,41 . TLR9 is an essential component of the oncogenic BCR signaling supercomplex (MyD88-TLR9-BCR) identified in ABC-DLBCL and leads to NF-κB activation 41 . Therefore, to examine the cooperation between Nfkbie loss and MYD88 signaling in vivo, we immunized mice intraperitoneally with the TLR9 ligand, CpG oligodeoxynucleotides, and analyzed inguinal lymph nodes 5 days later for GC B-cell formation. We observed a significant higher GC B-cell frequency and absolute cell numbers in Nfkbie −/− compared to Nfkbie +/+ mice (Fig. 8d). In the spleen, there was no induction of GC B-cell generation as the percentage of GC B cells was similar to unimmunized mice (data not shown). These data indicate that signaling through TLR9/MYD88 may cooperate with Nfkbie deficiency to promote GC B-cell formation and proliferation.

Discussion
The prevalent inactivation of NFKBIE in human BCL and the link between polymorphisms in the NFKBIE gene, and human autoimmune and infectious diseases 42,43 suggest a role for IκBε in B-cell development, function, and transformation. To address this question, we analyzed in detail young and aging Nfkbie-deficient mice. Our data indicate that IκBε activity is required for correct B-cell development on the one  development of mature B1 B cells 28,30 . Mice deficient in components of the alternative NF-κB pathway (e.g., NF-κB2) or of the classical NF-κB pathway (e.g., NF-κB1 and Rel), present a drastic reduction in the number of FoB and MZB cells, and B1 B cells, respectively 28 . In keeping with its negative regulator function 20,21 , absence of Nfkbie elicits elevated NF-κB activity downstream BCR and cytokine receptors 2,20-22 that likely accounts for the bias in mature B-cell development observed here.
We also uncovered that Nfkbie activity is required for correct development and function of B1 B cells. B1 cells might differentiate from fetal liver and adult bone marrow B1 progenitors 31 . The bone marrow of Nfkbie −/− mice contains more B1 progenitors than WT mice. However, we did not observe increased proliferation or survival of the B1 B cells directly isolated from Nfkbie −/− and WT mice. These findings suggest a role for IκBε in the generation or expansion of B1 progenitor pool. BCR signaling plays an instructive role in B1a cell lineage determination and maintenance 28,30,44,45 . In addition, recent studies showed that BCR signaling cooperates with TLR signaling in controlling expansion and activation of B1a B cells [46][47][48][49] . NF-κB is one of the key transcription factors activated by BCR and TLR signaling. Consequently, increase of NF-κB activation upon BCR and TLR signaling likely drives the expansion of B1a cell compartment observed in the Nfkbie-deficient mice.
Consistent with IκBε being required to limit B-cell activation 20,22 , our findings revealed that Nfkbie deficiency enhances proliferation of MZB and B1a IgMhi B cells in response to the TLR4 and TLR9 ligand, LPS, and CpG oligodeoxynucleotides respectively. B1a cells of Nfkbiedeficient mice show a significant downregulation of IgM expression in peritoneal B1a cells, with respect to WT and presented two B220lowCD19+CD5+B1a populations with different levels of IgM expression (B1a IgMhi and B1a IgMlow). Nfkbie −/− B1a IgMlow population showed impaired proliferation response to TLRs stimuli and reduced ex vivo survival compared to B1a IgMhi cells, in agreement with the key role of BCR in B1 cells expansion and activation 28,30,[45][46][47][48][49] . Downregulation of IgM on B cells contributes to a physiological B-cell immune tolerance mechanism called anergy, and serves to restrain their response to endogenous/self-antigens and to chronic BCR stimulation 50 . Our results strongly suggest that the B1a IgMlow population is phenotypically and functionally anergic as defined by in vivo downregulation of surface IgM expression, CD5 expression, in vitro hyporesponsiveness to TLR-dependent stimuli and ex vivo-reduced survival [50][51][52][53][54] . Importantly, these properties are hallmark features of human CLL 44,54 , in which NFKBIE is mutated in~10% of cases 2,3,15,16 . Recent work has uncovered cooperation between BCR and TLR in response to innate mitogenic stimuli. BCR-deficient B cells show impaired proliferation in response to LPS or CpG, indicating that correct BCR surface expression and signaling are critical for B-cell proliferation in response to such innate immune stimuli 48,49 . However, whether regaining IgM surface expression reverses anergic functions in these cells, as reported for autoreactive anergic cell model studies 55 , needs to be tested.
Nfkbie-deficient mice develop several features of human lymphoproliferative B-cell disorders, in particular, MBL, a low-penetrance premalignant stage 13,56 . Firstly, Nfkbie deficiency in mice leads to the expansion of an oligoclonal CD19 +CD5+B-cell subset, and in humans, mature CD19+CD5 +B cells, show high transcriptional and functional similarity to CLL B cells and are candidate for being a normal cellular counterpart of CLL B cells 13,45,54 . Secondly, Nfkbie-deficient B1a cells exhibited increased survival, proliferation, and enhanced NF-κB activity in response to stimulation consistently with the observations made for NFKBIE-mutated human CLL B-cells 2 . Thirdly, B220lowCD19+CD5+B cells expanded dramatically in the peripheral blood of old Nfkbiedeficient mice and infiltrated several organs, including spleen and lymph nodes. Finally, the lymphoproliferative disease developed by Nfkbie-deficient mice is indolent and has low penetrance, resembling the mostly indolent course of MBL and CLL, diseases of elderly population. Aging is associated with activation of innate immune cells, including a subset of B1a B cells, due to the increased auto-and pathogen-derived antigen availability, age-associated immune dysfunction, and age-induced gut permeability 45,57,58 . These age-associated changes could participate in the generation of MBL-like and/or CLL-like malignancies 45,54,59 . In this perspective, the indolent MBL development in Nfkbie −/− mice might largely be attributed to the increased survival and hyperproliferation of B1a IgMhi B cells, in response to immune stimulation that would be more available with age and the anergic nature of B1a IgMlow B cells and their hyporesponsiveness to stimulation. Indeed, in mice, MBL and CLL develops from B1a CD5+B cells and several study demonstrate a crucial role for BCR expression, reactivity, and signaling in the malignant transformation of these cells 45,[60][61][62] .
NF-κB activity plays an important role in the regulation of GC reaction 32,63 . Conditional deletion of Rel and Rela in GC B cells revealed that Rel is required to maintain GC populations, whereas Rela promoted Blimp-1-mediated plasma cell development 63 . Our data show that under Tcell-dependent immunization condition, Nfkbie deficiency promotes the excessive generation of GC B cells but not TFH cells in vivo. The amplification of GC B cells seems to be caused by an increased proliferation of Nfkbiedeficient GC B cells, suggesting a role for IκBε in controlling proliferation of GC B cells. In line with these findings, it was recently shown that activating mutations in CARD11 (observed in~5% of DLBCL (ref. 5 )), a NF-κB activator protein downstream BCR signaling, increases GC formation and GC B-cell proliferation through enhanced cycling 64 . GC B cells represent the origin of several BCLs including DLBCL (ref. 32 ). Deregulated NF-κB signaling is a hallmark of GC-derived B-cell malignancies 3,5,6,8 . Mutations in NFKBIE gene are detected iñ 5% of DLBCL and 23% of PMBCL (refs. 12,17 ). Our results provide a functional link between Nfkbie mutations and GC-derived BCL. As pointed above, the absence of Nfkbie induces a bias of B2 B-cell development toward MZB cells. MZB B cells are the normal cellular counterpart of splenic marginal zone lymphoma, a BCL in which NFKBIE is also mutated in~3% of cases 12 . Our data strongly suggest that IκBε plays an essential role in regulating proliferation and differentiation of the B cells from which these lymphomas derive.
In CLL and several BCL, aberrant NF-κB activation is the result of genetic alterations in pathways leading to NF-κB activation and aberrant stimulation from the microenvironment (through BCR, CD40, and TLR) 2,3,5,6,8,[10][11][12]32 . In vivo injection of the TLR9 ligand, CpG oligodeoxynucleotide, increased generation, and proliferation of Nfkbie −/− GC B cells compared to that of WT ones. Accordingly, MYD88L265P mutation, a mutation observed across many subtypes of human lymphoid malignancies, increased proliferation of Nfkbie −/− cells compared to that of WT B cells in vitro. In line with these findings, in Waldenström macroglobulinemia and DLBCL human cell lines, inactivation of TNFAIP3 (encoding A20, a negative regulator of NF-κB signaling) also cooperates with MYD88LP mutation to enhance NF-κB activation and resistance to BCR signaling inhibitor, ibrutinib 65,66 . Together, these observations demonstrate that inactivation of Nfkbie in mice predisposes to MBL and CLL, and upon collaboration with oncogenic events, it might promote BCL development.