Altered Marginal Zone B Cell Selection in the Absence of IκBNS

Marginal zone (MZ) B cells represent a distinct B cell subset that plays a central role in Ab responses against blood-borne Ags. In rodents, this cell population is located in the splenic MZ, outside the marginal sinus surrounding the lymphoid follicles (1). Adjacent to the MZ B cells, the MZ contains specialized macrophages, including MZ macrophages and metallophilic marginal macrophages (MMMs), as well as dendritic cell subtypes and MZ-specific stromal cells. In contrast to the follicular (FO) B cells that represent the majority of conventional B-2 cells and occupy the FO area of the spleen and lymph nodes, MZ B cells are found in the MZ of the spleen (2, 3).

In addition to their distinct anatomical localization, MZ B cells can be distinguished from FO B cells by the expression of several surface markers. They exhibit IgM^hiIgD^loCD21/35^hiCD23^lo/− phenotype, whereas FO B cells are IgM⁺IgD^hiCD21/35^intCD23^hi (4, 5). MZ B cells also express higher levels of MHC class II (MHCII), CD1d, and CD9 (6) and exhibit higher basal levels of the activation markers CD80 (B7-1) and CD86 (B7-2) than FO B cells (5, 7). The special location and preactivated state of MZ B cells allow them to interact with circulating pathogens and induce rapid T cell–independent type 2 (TI-2) Ab responses. Thus, MZ B cells provide the first immune defense against blood-borne pathogens, particularly against encapsulated bacteria, such as Streptococcus pneumoniae, Neisseria meningitidis, and Haemophilus influenzae (8, 9). MZ B cells shuttle between the MZ and the follicle (10), transporting captured Ags from the blood into the follicles, thereby participating in T cell–independent (TI) responses against foreign Ags (11, 12). Additionally, MZ B cells express a BCR repertoire that has been linked to the polyreactivity of the natural Ab pool (13–15). Although MZ B cells have a distinct phenotype, it has been shown that it is a heterogenous population, consisting of a mixture of naive and memory B cells (16). In relation to MZ B development, it is known that immature surface IgM⁺ B cells generated in the bone marrow migrate to peripheral lymphoid tissues where they continue their differentiation into mature naive B cells. In the spleen, these newly formed B lymphocytes progress through distinct transitional B cell stages (transitional-1 [T1], transitional-2 [T2], and transitional-3 [T3]) (17–19). T2 B cells are direct precursors of mature FO B cells, but a subset of these cells likely represents the immediate precursors of MZ B cells (T2–MZ precursor [T2-MZP]) (20, 21). FO B cells may also function as a potential reservoir for replenishing the MZ B cell compartment (18, 22–24). Direct development of MZ B cells from the T1 B subset has also been proposed (25, 26).

Commitment toward FO B cells versus MZ B cells depends on signals transduced by multiple external stimuli, including the BCR (3, 19, 27–29), Notch 2 (21, 30), and the receptor for B cell–activating factor (18, 31), all of which require functional NF-κB signaling. Additionally, signals involved in the migration and retention of B cells in the MZ may contribute to shape the fate of this cell population (18, 32, 33).

In mammals, NF-κB transcription factors are composed of five members [NF-κB1 (p50), NF-κB2 (p52), RelA (p65), RelB, and c-Rel], which form various homo- or heterodimers to mediate appropriate signals (34). Under resting conditions, these NF-κB complexes are sequestered in the cytoplasm in an inactive form complexed to the IκB proteins. Upon upstream signals, IκB proteins are proteasomally degraded, allowing NF-κB dimers to translocate to the nucleus where they induce or repress the transcription of a variety of genes (35). In addition to the classical IκB proteins (IκBα, IκBβ, and IκBε; precursors p100/IκBδ and p105/IκBγ), which are primarily found in the cytoplasm and associate with p50/p65 heterodimers, the atypical IκB proteins (BCL-3, IκBζ, IκBη, and IκBNS) localize in the nucleus, where IκBNS can interact with p50 homodimers (36), p52, and the Rel proteins (37).

We previously described a mouse strain, bumble, which carries a non-sense mutation in the nfkbid gene encoding the regulatory IκBNS protein. bumble mice exhibit reduced frequencies of splenic MZ B cells, a near-complete loss of peritoneal B-1 cells, and a diminished humoral immune response to TI and T cell–dependent (TD) Ags (38, 39). Because these mice have normal B-2 cell frequencies, the bumble strain represents a useful model to study the development of MZ B and B-1 cells. Further understanding of the biology of MZ B cells has clinical relevance, because splenectomized individuals, as well as children younger than 2 y of age (who still have “immature” MZ B cell populations), exhibit increased susceptibility to infections caused by encapsulated bacteria.

In this article, we demonstrate that the reduced MZ B cell compartment observed in bumble mice aged 6–10 wk is not due to an altered MZ cellular microenvironment but rather is caused by an intrinsic defect in MZ B cell development. Interestingly, by 3–4 mo of age, the vacant MZ niche in bumble mice is gradually populated with a cell type that resembles MZ B cells based on the expression of CD21 and CD23 markers; however, by several other read-outs, they are distinct from wild-type (wt) MZ B cells.

IκBNS-deficient bumble mice, generated by ENU mutagenesis of C57BL/6J mice, and their wt C57BL/6J counterparts have been described previously (38). For bone marrow chimera experiments, CD45.1 and RAG^−/− mice on a C57BL/6J background were used. All animals were bred and housed at Karolinska Institutet animal facility. All animal procedures were approved by the Stockholm North Ethical Committee on Animal Experiments, Sweden (Stockholms Norra djurförsöksetiska nämnd) and performed according to the given guidelines. Nfkb1-deficient Finlay mice, generated by ENU mutagenesis, are described at https://mutagenetix.utsouthwestern.edu. They were housed at the UT Southwestern Medical Center vivarium, and all experimental procedures were performed in accordance with institutionally approved protocols.

Single-cell suspensions of splenocytes were prepared using a 70-μm cell strainer and a syringe plunger. Peritoneal cells were isolated by flushing with 10 ml of cold PBS (Sigma, St. Louis, MO) plus 1% FCS (HyClone, Logan, UT). Femurs and tibias were flushed with a 26-gauge needle. Cell suspensions were collected in RPMI 1640 medium (Sigma) supplemented with 10% FCS (HyClone), 2 mM l-glutamine, penicillin (100 IU)-streptomycin (100 μg/ml) (Sigma), and 5 × 10⁻⁵ M 2-ME (Life Technologies, Waltham, MA) (complete RPMI medium). Cell suspensions were washed once in Ca²⁺- and Mg²⁺-free PBS (Sigma) and treated with RBC lysis buffer before further processing. RBC lysis buffer was omitted for the isolation of peritoneal cells and when cells for bone marrow chimeras were prepared.

Single-cell suspensions from the indicated tissues were washed in PBS–1% FBS (wash buffer). Cells were incubated with anti-CD16/32 Ab (2.4G2) and then surface stained with the following biotin- or fluorochrome-conjugated anti-mouse Abs: B220 (RA3-6B2), CD93 (AA4.1), CD21/35 (7E9), CD23 (B3B4), CD24 (M1/69), CD1d (1B1), CD80 (16-10A1), CD86 (GL1), CD5 (53-7.3), CD9 (KMC8), CD19 (1D3), CD43 (S7), CD45.1 (A20), CD45.2 (104), CD138 (281-2), IgM (II/41), Ig, κ L chain (LC; 187.1), Ig, λ1, λ2, and λ3 LC (R26-46), and Ig, λ1 LC (R11-153) (all from BD Biosciences, Franklin Lakes, NJ). IgM (RMM-1), and MHCII (M5/114.15.2) were purchased from BioLegend (San Diego, CA). CD38 (90) was purchased from eBioscience (San Diego, CA). Biotinylated primary reagents were detected using streptavidin–Alexa Fluor 488 (Invitrogen, Eugene, OR) or streptavidin–PerCP–Cy5.5 (eBioscience). All staining steps were performed in PBS–1% FBS at 4°C. Cells were collected using a FACSCalibur (BD) or a BD LSR II (BD) flow cytometer, and data were analyzed by FlowJo software version 8.7.1 or 10.0.8 (TreeStar, Ashland, OR).

For the isolation of MZ B and FO B cells, splenocytes from three wt and from three bumble mice were pooled, stained with B220, CD21/35, and CD23 Abs, and sorted using a BD FACSAria II cell sorter. Purity of sorted cells was >95%.

Nonlethally irradiated RAG^−/− recipient mice (600 rad, ¹³⁷Cs source) were reconstituted by i.v. injection with a 1:1 or 1:4 mix of CD45.1 wt and CD45.2 wt or bumble bone marrow cells (10 × 10⁷ cells total). Recipient mice were given antibiotics in the drinking water for 21 d. Chimeras were analyzed 8–12 wk after reconstitution.

For TI-2 Ag responses, age- and sex-matched wt and bumble mice were injected i.p. or i.v. with 50 μg of NP40-Ficoll (Biosearch Technologies, Novato, CA) in 100 μl of PBS. Anti-NP IgM Ab titers were measured by ELISA at day 7.

Age-matched wt and bumble mice were injected i.v. in the tail vein with 100 μg of 2,4,6-trinitrophenyl TNP-Ficoll (TNP-AECM-Ficoll; Biosearch Technologies) in 200 μl of sterile PBS. Uptake of TNP-Ficoll by MZ B cells was assessed by flow cytometry 30 min after injection, using biotin-coupled anti-TNP Ab (clone 49.2; BD).

NP-specific ELISA was performed by coating 96-well ELISA plates (Nunc, Rochester, NY) with 5 μg/ml NP25-BSA (Biosearch Technologies) diluted in PBS and incubating overnight at 4°C. Following washing in PBS–0.05% Tween-20 (wash buffer) and blocking with PBS containing 2% (w/v) dry milk for 1 h, serum was added in 3-fold serial dilutions in blocking buffer and incubated for 1.5 h at room temperature (RT). After washing six times in wash buffer, 100 μl per well of the secondary Ab, HRP-coupled goat anti-mouse IgM (Southern Biotechnology Associates, Birmingham, AL), was added, and plates were incubated for 1.5 h at RT. Plates were washed with wash buffer and developed with 3,3′,5,5′-tetramethylbenzidine substrate (KPL, Gaithersburg, MD). The reaction was stopped by 1 M H₂SO₄, and absorbance was measured at 450 nm using an Asys Expert 96 ELISA reader (Biochrom, Cambridge, U.K.).

An ELISPOT assay was used for the detection of NP-specific IgM-producing cells. MultiScreen-IP Filter Plates (Millipore, Billerica, MA) were pretreated with 70% ethanol and washed in sterile PBS. Plates were coated with 5 μg/ml NP30-BSA (Biosearch Technologies), diluted in PBS, and incubated overnight at 4°C. The following day, plates were washed in sterile PBS, blocked in complete RPMI medium for 2 h at 37°C, and cells were added in triplicates to the wells in 2-fold serial dilutions, starting at 2 × 10⁵ cells per well. Plates were incubated for 17 h at 37°C in 5% CO₂. Cells were then removed by washing the plates six times in PBS, and 0.1 μg of biotinylated goat anti-mouse IgM (Mabtech) diluted in PBS was added to each well. After 2 h of incubation at RT, plates were washed, and 100 μl of alkaline phosphatase–conjugated streptavidin (Mabtech) diluted in PBS was added to the wells, and plates were incubated for 45 min at RT. After washing, plates were developed with 100 μl of 5-bromo-4-chloro-3-indolyl phosphate/NBT-plus substrate (Mabtech). The reaction was stopped when distinct spots were observed by rinsing the plates extensively with tap water. Spots were counted by an ELISPOT reader and analyzed using BioSpot suite (both from CTL).

Spleens from age-matched wt and bumble mice were embedded in Tissue-Tek OCT compound (Sakura Finetek, Zoeterwoude, the Netherlands), snap-frozen in liquid nitrogen, and stored at −80°C. Sections (6–8 μm) were mounted on SuperFrost Plus slides (VWR, Leuven, Belgium), air dried, and fixed in ice-cold acetone for 20 min. Prior to immunostaining, sections were rehydrated in PBS and blocked with PBS containing 5% normal horse serum for 30 min at RT. Endogenous biotin and avidin–binding activity was blocked using an Avidin/Biotin Blocking Kit (Vector Laboratories, Burlingame, CA), according to the manufacturer’s protocol. The sections were stained at RT for 30 min with the following Abs in the combinations indicated in the relevant figures: anti-IgM–FITC (polyclonal; Southern Biotech Associates), anti-CD1d–PE (BD), Alexa Fluor 647–anti-IgD (mAb 11-26c.2a; BioLegend), biotinylated anti–SIGN-R1 (clone ER-TR9; BMA) or unconjugated anti–SIGN-R1 (clone ER-TR9; Abcam, Cambridge, U.K.), biotinylated MOMA-1 (clone MOMA-1; Abcam), unconjugated anti-MARCO (clone ED31; Hycult Biotech), and unconjugated MECA-367 Ab (anti–MAdCAM-1; BD). Biotinylated primary Abs were detected with Alexa Fluor 555–conjugated streptavidin (Life Technologies) or Cy3-conjugated streptavidin (Jackson ImmunoResearch). Unconjugated rat Abs were detected using goat anti-rat Ab conjugated with Alexa Fluor 647 (Invitrogen). All staining steps were performed in PBS–5% normal horse serum at RT. After the immunoreactions, sections were coverslipped using 2.5% DABCO (Sigma) in glycerol. The sections were examined with a Zeiss LSM 510 Meta confocal system installed on a Zeiss Axioplan 2 microscope equipped with 5× (N.A. 0.16), 10× (N.A. 0.45), 20× (N.A. 0.75), 40× oil (N.A. 1.3), 63× oil (N.A. 1.4), and 100× oil (N.A. 1.45) objectives (Carl Zeiss, Jena, Germany). Digital images were modified to optimize brightness and contrast using ZEN 2011 (Carl Zeiss) and Adobe Photoshop 7.0 software (Adobe Systems, San Jose, CA). Image assembly, processing, and quantification were performed using ImageJ software (ImageJ1.47t, National Institutes of Health).

A total of 2 × 10⁵ FACS-sorted MZ B or FO B cells was cultured in 48-well plates (TPP, Trasadingen, Switzerland) in 1 ml of complete RPMI medium in the absence or presence of 10 μg/ml Escherichia coli LPS (0111B4; Sigma) for the indicated time at 37°C in a 5% CO₂ humidified incubator. Expression of plasma cell (PC) markers CD138 and B220 was assessed by flow cytometry. Cell culture supernatants were collected after 6 d of stimulation, and IgM titer was analyzed by ELISA.

Statistical differences between groups were analyzed by a nonparametric Mann–Whitney U test or by a two-tailed unpaired t test, as indicated. GraphPad Prism software (version 6.0; GraphPad Software, San Diego, CA) was used for analysis, and data were considered significant at *p ≤ 0.05, **p ≤ 0.01, and ***p ≤ 0.001.

IκBNS-deficient bumble mice exhibit normal early B cell development in the bone marrow but severely reduced frequencies of CD23⁻CD21⁺ MZ B cells and a complete loss of B-1a cells (38, 39). To further investigate the MZ B cell defect, we first examined the frequencies of MZ B cells in 6–10-wk-old bumble mice using two MZ B cell marker combinations, CD23/IgM and CD23/CD1d, in addition to the CD23/CD21 staining used previously (38). These complementary stainings confirmed that bumble mice exhibited severely reduced frequencies of MZ B cells (Fig. 1A). To further verify the reduced size of the MZ B cell compartment, we performed immunohistochemical staining of frozen spleen sections from age-matched 6–10-wk-old wt and bumble mice. We used Abs against IgM and MOMA to visualize MZ B cells and MMMs at the border between the FO and MZ areas, respectively (40). In agreement with the flow cytometry data, the thickness of the IgM⁺ MZ B cell population peripheral to MOMA-1⁺ cells at the border of the MZ was markedly reduced in bumble mice in comparison with age-matched wt control mice. In contrast, we found comparable MOMA-1⁺ staining in the two genotypes, indicating normal frequencies and distribution of MMMs in bumble mice (Fig. 1B).

We next examined potential reasons for the reduced frequency of MZ B cells in bumble mice. Based upon previous studies, we reasoned that this may be a consequence of defective late-stage MZ B cell development or altered B cell migration or retention (18). To determine whether IκBNS is required for B cell maturation in the periphery, we investigated the different transitional B cell subsets (T1, T2, T3, and T2-MZP) in the spleen by flow cytometry (Fig. 1C, 1D). Representative dot plots in Fig. 1C (left panel) demonstrated that all three immature transitional B cell subsets (staining with B220 CD93 [AA4.1] CD23 IgM), T1, T2, and T3, are present in bumble mice. An additional staining panel (B220 CD21 IgM and CD23) was used to distinguish the CD21^hiIgM^hiCD23⁺ T2-MZP and CD21^hiIgM^hiCD23⁻ MZ B populations in wt and bumble mice (Fig. 1D, left panel). Determination of frequencies and absolute numbers revealed no significant differences in the T1 population between wt and bumble mice, and although the percentages of T2 cells were significantly increased in bumble mice relative to wt mice, the absolute number of these cells was comparable in the two genotypes. In contrast, the T3 subpopulation was significantly reduced in terms of frequencies and numbers in IκBNS-deficient bumble mice (Fig. 1C, right panel). Furthermore, the frequencies and absolute numbers of T2-MZP cells were significantly higher in bumble mice compared with wt mice, whereas mature MZ B cells in bumble mice were severely reduced, as expected (Fig. 1D, right panel). Altogether, these data demonstrate that the splenic MZ B cell niche in bumble mice at 6–10 wk of age is largely vacant. Additionally, the elevated number of T2-MZP cells suggests that development of mature MZ B cells is blocked at the T2-MZP stage.

Earlier studies demonstrate that the MZ microenvironment is critical for the maintenance and retention of MZ B cells (41, 42). To investigate whether the reduced MZ B cell number in bumble mice is a result of an altered MZ structure, we stained spleen sections from age-matched (6–10 wk) wt and bumble mice. We did not detect any alterations in the location of MARCO⁺ and SIGN-R1⁺ MZ macrophages, MAdCAM-1⁺ sinus-lining cells, or ER-TR7⁺ stromal cells in sections from bumble mice compared with wt mice (Fig. 1E–G).

Taken together, these results suggest that the observed reduction in the MZ B cell compartment in bumble mice is more likely to be a consequence of impaired B cell development than an altered MZ structure.

To investigate whether the decreased MZ B cell compartment in IκBNS-deficient bumble mice is due to an intrinsic defect in the B cell lineage, we performed a competitive bone marrow–reconstitution assay. Bone marrow cells from wt (CD45.1 allotype) and bumble (CD45.2 allotype) mice were mixed at a 1:1 or 1:4 ratio and injected into nonlethally irradiated recipient RAG-1^−/− mice, which were analyzed 8–12 wk after transplantation. The fate of the transplanted cells was followed by FACS analysis (Fig. 2A, lower panels). The gating strategy for the different B cell subpopulations is shown in Fig. 2A (upper panels). As controls, RAG-1^−/− mice received cells from bumble mice (CD45.2) only or from wt (CD45.1)/wt (CD45.2) mice. To determine the contribution of bumble and wt cells, absolute cell numbers of CD45.1⁺ (wt) versus CD45.2⁺ (bumble) cells in the different splenic B cell subsets were compared (Fig. 2B). For FO B, T1, T2-MZP, and T2-FO populations, the cell numbers were similar in the mice that received bone marrow cells from bumble mice and those that received wt cells. In contrast, mice that received only bumble bone marrow cells exhibited reduced numbers of MZ B cells relative to those that were injected with wt cells. bumble bone marrow cells transferred alone gave rise to some MZ B cells, whereas in the presence of wt cells, bumble-derived MZ B cells could not be recovered. This result suggests a competitive disadvantage for bumble MZ B or MZ B precursor cells. Staining for T2-MZP cells also showed that this population was mainly wt derived in the mixed bone marrow chimeras, thus suggesting that, in a competitive setting, bumble MZ B cell development may be blocked prior to the T2-MZP stage. Overlaid FACS plots summarize the distribution of the different wt- and bumble-derived splenic B cell populations in wt/bumble chimera mice (Fig. 2C). Additionally, in the mixed chimera animals, T1 and T2-FO B cells derived predominantly from bumble bone marrow cells, possibly indicating a partial block in bumble B cell development at the T1 stage in a competitive setting.

Taken together, these data demonstrate that the reduction in MZ B cells in IκBNS-deficient mice is a consequence of a cell autonomous defect and that bumble MZ B cells have a competitive disadvantage compared with wt MZ B cells.

Previous studies of mice lacking NF-κB1 (p50), which exhibit reduced MZ B cell numbers, showed an increase in the MZ B pool as the mice aged (43). Using Nfkb1-deficient Finlay mice, we confirmed that CD23⁻CD21⁺ MZ B cells accumulate with age in the absence of functional NF-κB1 (Supplemental Fig. 2). To investigate whether this was also the case in bumble mice, we performed a kinetic study of bumble and age-matched wt controls of different ages (4–22 wk). We observed significantly lower B cell frequencies in 4-wk-old bumble mice compared with age-matched wt mice; however, this difference was not apparent in older mice (Supplemental Fig. 1A). The frequency and absolute numbers of CD23⁻CD21⁺ MZ B cells in bumble mice increased gradually with age and even reached wt levels by 14 wk of age. Similar observations were made in wt control mice, but this difference was much smaller and was not significant (Supplemental Fig. 1B). In further experiments, we compared 6–10-wk-old and 5–6-mo-old mice. Representative FACS staining showed an accumulation of CD23⁻CD21⁺ MZ B cells in 5–6-mo-old bumble mice (Fig. 3A, top panel).

Quantitative analysis of the MZ B cell compartment revealed a significant increase between 6–10-wk-old and 5–6-mo-old bumble mice that was not observed in age-matched wt mice. In contrast, we did not see alterations in frequencies or absolute numbers of FO B cells with aging of bumble mice (Fig. 3A, middle and bottom panels). To determine whether the accumulated bumble MZ B cells occupied the MZ niche, we stained spleen sections from age-matched 6–10-wk-old and 5–6-mo-old bumble mice and wt controls. In agreement with the FACS data, we observed an increased size of the IgM⁺IgD⁻ cell population in the MZ area, peripheral to the MOMA-1⁺ marginal sinus, in 5–6-mo-old bumble mice (Fig. 3B). Examination of the frequency and numbers of T2-MZP cells in 6–10-wk-old and 5–6-mo-old bumble mice revealed a similar increase with age that was not observed in age-matched wt control mice (Fig. 3C, Supplemental Fig. 1C). Finally, we found that the numbers of peritoneal B-1 cells increased with age in wt mice but remained completely absent in 5–6-mo-old bumble mice (Fig. 3D) (39).

Altogether, we found that the accumulation of MZ B cells in aging bumble mice was not restricted to MZ B cells; it also was observed for the T2-MZP cell pool.

MZ B cells are involved in the initiation of immune responses to TI-2 Ags (44) (i.e., capsular polysaccharides of encapsulated bacteria). To determine whether 5–6-mo-old bumble mice with an accumulated pool of MZ B cells were able to induce a humoral immune response to polysaccharides, we immunized 5–6-mo-old wt and age-matched bumble mice i.v. or i.p. with the TI-2 model Ag, NP-Ficoll. NP-specific IgM and IgG levels were analyzed by ELISA 7 d postinjection. In contrast to wt mice, 5–6-mo-old bumble mice failed to produce NP-specific Abs in response to NP-Ficoll injection (Fig. 4A). Moreover, when analyzing NP-specific IgM Ab-secreting cells (ASCs) in splenocytes from mice 6 d after i.v. NP-Ficoll injection, we observed significant responses in immunized wt mice but no detectable NP-specific IgM ASC responses in bumble mice (Fig. 4B). Interestingly, low, but detectable, NP-specific IgM responses were also observed in splenocytes from naive wt mice but not in splenocytes from naive bumble mice, suggesting a difference in the naive B cell repertoire between the strains, perhaps due to altered selection processes in bumble mice.

Localization of MZ B cells in the MZ is critical for the uptake of blood-borne Ag (10, 45). To assess whether MZ B cells from 5–6-mo-old bumble mice were able to capture polysaccharides in the blood, we injected TNP-Ficoll i.v., which is a TI-2 Ag. Thirty minutes after injection, CD23⁻CD21⁺ MZ B cells were analyzed by flow cytometry for the binding of TNP-Ficoll. We found that bumble MZ B cells exhibited reduced capture of TNP-Ficoll compared with wt control mice (Fig. 4C).

To further explore the identity of the B cells that accumulate in the MZ of aged bumble mice, we examined the expression of additional MZ B cell markers. First, we stained splenocytes from 6–10 wk and 5–6-mo-old bumble mice and age-matched wt controls for B220, CD93 (AA4.1), CD23, and CD1d expression to compare the sizes of the accumulated MZ B cell populations in the two strains. Five- to six-month-old bumble mice showed elevated frequencies of MZ B cells relative to mice aged 6–10 wk (Fig. 5A, upper panel), which was significant in terms of frequencies and numbers (Fig. 5A, lower panel). Immunohistochemical staining of spleen sections from age-matched 6–10-wk-old and 5–6-mo-old mice revealed that the accumulating CD1d⁺IgD⁻ B cells in older bumble mice occupied the MZ area (Fig. 5B); however, the thickness of the CD1d⁺ B cell rim outside of the MOMA⁺ MMM ring appeared thinner in bumble mice compared with wt mice.

These results led us to investigate the expression of other surface markers associated with the accumulating B220⁺CD23⁻CD21⁺ cells in the MZ of 5–6-mo-old bumble mice. In addition to the reduced CD1d expression, we found several abnormalities compared with wt MZ B cells, including reduced CD9, CD24 (HSA), CD86, and CD38 expression. In contrast, the expression of MHCII was significantly higher on bumble B220⁺CD23⁻CD21⁺ cells, whereas IgM levels were comparable (Fig. 5C). T2-MZP cells also express higher levels of CD1d and CD9 than FO B cells (46, 47). However, T2-MZPs derived from bumble mice showed lower CD1d and CD9 expression levels in comparison with age-matched wt mice (Supplemental Fig. 3). MZ B cells have been described to be larger in size than FO B cells (48, 49). Likewise, accumulating MZ B cells in the bumble spleen appeared larger in size than FO B cells and were comparable to the size of wt MZ B cells (Fig. 5D).

Taken together, these results indicate that B cells that occupy the vacant MZ niche in 5–6-mo-old bumble mice are different from wt MZ B cells with regard to the expression of several characteristic MZ B cell markers.

Previous findings showed increased Igλ LC usage in the BCR of B-1a and MZ B cell subsets (25, 50). To investigate the BCR repertoire, we compared Igλ expression in B220⁺CD21⁺CD23⁻ MZ B cell and B220⁺CD21⁺CD23⁺ FO B cell subsets in adult wt mice and found a moderate increase in Igλ expression in MZ B cells compared with FO B cells (Fig. 6A, 6B). Further analysis of Igλ⁺ cells in accumulated MZ B cells derived from 5–6-mo-old mice revealed significantly lower Igλ usage in IκBNS-deficient bumble mice compared with wt controls (Fig. 6C, 6D).

In mice, three subtypes of Ig λ LC are known: λ1, λ2, and λ3 (51, 52). Therefore, we further analyzed the λ LC repertoire in the aged (5–6 mo old) bumble MZ B cell population. We found a significant reduction in the frequency of λ2,3⁺ B cells in the bumble late-appearing MZ B cell subset (Fig. 6E, 6F). In addition, unlike T1 B cells, T2-MZP cells, the presumed precursors of MZ B cells, showed reduced λ LC usage in bumble mice compared with the corresponding population in wt mice. Altered λ LC usage was also observed in FO B cells, suggesting a more general defect in B cell selection in bumble mice (Supplemental Fig. 4A, 4B). Strikingly, similar to the late-appearing MZ B cells, all bumble B cell populations analyzed (FO B cells, T2-MZP cells, and T1 B cells) exhibited impaired λ2,3 LC expression compared with the corresponding populations in wt control mice (Supplemental Fig. 4C, 4D). Thus, bumble mice showed reduced frequencies of Igλ⁺ B cells in the periphery, indicating that IκBNS plays an important role in Ig LC expression.

MZ B cells are known to proliferate and generate effector cells rapidly after polyclonal in vitro stimulation (5). After 24 h of activation by the classical TI type 1 (polyclonal) agent LPS, MZ B cells express the PC marker syndecan-1 (CD138) (9). We examined the capacity of MZ B cells from aged bumble mice to generate PCs. Sort-purified MZ B cells from 5–6-mo-old wt and bumble mice were stimulated with LPS for the indicated times, and PC development was assessed by flow cytometry (Fig. 7A). These experiments demonstrated that a large proportion of wt MZ B cells express the PC marker CD138 (syndecan-1) in response to LPS at day 1. In contrast, in bumble mice, the proportion of CD138⁺ cells was substantially smaller on day 1 but was similar to day 0. At day 6, a massive CD138⁺B220⁻ PC maturation was observed in cultures from wt mice, whereas the proportion of PCs was reduced in cultures from bumble mice. At day 6, the CD138⁺B220⁺ plasmablast (PB) proportion was increased in bumble mice compared with day 1, indicating delayed PC maturation in bumble mice (Fig. 7B). Measurement of the levels of IgM secreted in 6-d culture supernatants confirmed the reduced capacity of IκBNS-deficient bumble MZ B cells to differentiate into ASCs in response to LPS (Fig. 7C). Collectively, these results indicate that IκBNS is required for intact PC differentiation and IgM production by MZ B cells in response to LPS stimulation.

In this study, we investigated MZ B cells in bumble mice deficient for the atypical IκB protein IκBNS (38, 53). We found impaired MZ B cell development due to a B cell–intrinsic block at the T1 and/or T2-MZP developmental stage. We also demonstrated that the vacant MZ niche was repopulated by B cells in aging bumble mice, but these late-appearing cells differed from wt B220⁺CD23⁻CD21⁺ MZ B cells in phenotype and function, as shown by reduced expression of CD1d, CD9, CD80, and CD38 and a lack of responsiveness to TI NP-Ficoll immunization. Furthermore, the late-appearing bumble MZ B-like cells had reduced Igλ expression (especially Igλ2 and Igλ3) and, upon sorting and in vitro culturing in the presence of LPS, they exhibited impaired PC maturation compared with similarly sorted MZ B cells from wt mice.

NF-κB activity is essential for the development and survival of MZ B cells in the spleen; impaired expression of canonical and noncanonical NF-κB signaling pathway members (NF-κB1/p50, c-Rel, Rel-A, NF-κB2, Rel-B, CARMA1, Malt1, Bcl-10, Btk) was shown to affect MZ B cells and B-1 cell development (reviewed in Ref. 54). Moreover, mice with single deficiency in NF-κB1/p50 (55), NF-κB2, or Rel-B completely lack MZ B cells (56–58), whereas a less severe reduction in MZ B cell numbers was observed in the absence of c-Rel or Rel-A (p65) (55). Functional NF-κB signaling requires the polyubiquitination and proteasomal degradation of IκB proteins. Recent studies have revealed important roles for the atypical IκB protein IκBNS in lymphopoiesis and immune responses (38, 53, 59). The previously reported bumble mouse strain lacks functional IκBNS due to an ENU-induced non-sense mutation in the Nfkbid gene (38). bumble mice exhibit reduced numbers of MZ B and B-1 cells and defective Ab responses to TI and TD Ags (38). By flow cytometry analysis, Touma et al. (53) showed that CD23⁻CD21⁺ MZ B cells accumulated in 10-mo-old IκBNS-knockout (KO) mice, but these late-appearing cells were not characterized further.

In the current study, analysis of bumble splenocytes revealed a marked expansion of the CD23⁻CD21⁺ MZ B cell population between 4 and 22 wk of age. MZ B cells are primarily defined by their anatomical localization (60). Immunohistochemical staining of splenic sections of aged bumble mice revealed numerous IgM⁺ B cells in the MZ, proximal to the MOMA⁺ marginal sinus, indicating that MZ B cells accumulate with age in IκBNS-deficient mice. The localization of B cells that filled the vacant MZ niche was confirmed by additional staining using CD1d and MOMA markers to delineate the MZ B cells and the marginal sinus, respectively. The results were reminiscent of observations in p50-KO mice, in which the number of CD23⁻CD21⁺ MZ B cells reached wt level by 6 mo of age, and these cells also occupied the MZ niche (43).

In addition to their characteristic anatomical localization, several surface markers are used to distinguish MZ B cells from FO B and B-1 B cells. MZ B cells typically express high levels of surface IgM, CD21 (CR2), CD1d, and CD9 and low levels of surface IgD, CD23, CD5, and CD11b. Moreover, MZ B cells exhibit elevated basal expression of MHCII, CD80, and CD86, indicating that these cells are in a preactivated state. Phenotypic characterization of the late-appearing MZ B cells from aged bumble mice revealed that, in contrast to the accumulated CD1d⁺ and CD9⁺ B cells in p50-KO mice, bumble MZ B-like cells expressed CD1d and CD9 at much reduced levels. In wt C57BL/6 mice, MZ B cells express the highest levels of CD1d among immune cells. This MHC class I–like β2-microglobulin–associated molecule presents phospho- and glycosphingolipid Ags to invariant NK T cells. In murine B cells, CD1d expression relies on the ETS family of transcription factor, Elf-1 (61), to positively regulate its transcription. Because an evolutionarily conserved physical interaction was observed between NFATc/NFκBp50 and Elf-1 (62), and reduced levels of NFATc1 were described in IκBNS-KO mice (63), it will be of interest to investigate whether IκBNS is involved in the regulation of CD1d expression in B cells via activation of NFATc.

The tetraspanin family member CD9 is expressed at high levels on MZ B cells, T2-MZP cells, B-1 cells, and PCs (6, 47), whereas the accumulated MZ B and T2-MZP cells in aged IκBNS-deficient mice were CD9⁻. Of relevance to this, a recent article shows that IL-9/IL-9R signaling is involved in the regulation of CD9 expression on MZ B cells (64). This is of interest because IL-9 expression is regulated via cooperation between NF-κB p65 (Rel-A) and NFATc2, where NFATc2 binding to the IL-9 promoter facilitates subsequent NF-κB(p65)Rel-A binding to induce transcription of IL-9 (65). Because IκBNS interacts with p50 and Rel-A, this adds to the question of whether IκBNS affects the expression of genes regulated via NFATc.

Derudder et al. (66) showed that NF-κB signaling is required for Igλ⁺ B cell development. MZ B cells were reported to express higher surface Igλ LC levels than FO B cells (25). In this article, we report that bumble mice exhibited significantly reduced Igλ LC usage not only in the accumulated MZ B-like cells, but also in FO B cells and T2-MZPs compared with the corresponding B cell populations in age-matched wt mice. The reduced frequency of Igλ⁺ cells in bumble mice indicates that intact NF-κB signaling via IκBNS is required for the development of Igλ-expressing cells. Further investigation of the reduced Igλ⁺ cell populations in bumble mice revealed that λ1⁺ B cells predominated over λ2⁺ and λ3⁺ B cells in MZ B-like cells, as well as in the FO B, T2-MZP, and T1 subsets. These results suggest a general role for IκBNS in Igλ LC expression.

MZ B cells and B-1 cells are the predominant responders to TI Ags, but it is not known to what extent these cells contribute to the overall IgM response to NP-Ficoll immunization. Some studies support a major role for MZ B cells (44, 67), whereas other studies emphasize the role of B-1 cells in TI responses (68, 69). We previously demonstrated that adoptive transfer of wt peritoneal B-1 cells to bumble mice, followed by immunization with NP-Ficoll, resulted in a partially restored TI-2 response, confirming a role for B-1 cells in the NP-specific IgM response. However, this experiment did not reveal whether the partial effect was due to incomplete repopulation of the B-1 cell compartment following transfer or whether MZ B cells also contribute to the NP-specific response observed after NP-Ficoll immunization in wt mice (39). To investigate a potential contribution of MZ B cells in the NP-Ficoll response, we analyzed splenocytes from mice injected i.v. with NP-Ficoll. We detected a robust NP-specific IgM ELISPOT response in splenocytes from aged wt mice, whereas no NP-specific IgM ELISPOT response was detected in splenocytes from aged bumble mice. These experiments demonstrate that NP-specific IgM Ab responses can originate in the spleen of wt mice and suggest that the reappearing MZ B-like cells in aged bumble mice are nonfunctional.

The TLR agonist LPS can directly induce MZ B cells to differentiate into IgM-secreting PCs (70). However, the accumulated MZ B cells in aged IκBNS-deficient bumble mice were severely defective in their ability to differentiate into PCs in response to LPS stimulation, suggesting that IκBNS is necessary for intact PC generation. The role of IκBNS in PC differentiation in bumble mice is currently under investigation in the Karlsson Hedestam laboratory (Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet) to determine whether this is a general defect in IκBNS-deficient B cells.

In summary, our data indicate that, in the absence of functional IκBNS, MZ B cell commitment is blocked during the T1 or T2-MZP B cell stage, leaving a vacant MZ B cell niche in bumble mice. With age, B cells accumulate in the bumble MZ; however, these cells differ from wt MZ B cells in phenotype and function. Our future studies aim to determine the origin of these late-appearing MZ B-like cells, which is still an open question.

We thank the personnel at the animal facility and the FACS facility at the Department of Microbiology, Tumor, and Cell Biology at Karolinska Institutet.

The authors have no financial conflicts of interest.