IKKβ Is Required for Peripheral B Cell Survival and Proliferation¹

Nuclear factor-κB designates a family of transcription factors activated by a diverse array of proinflammatory cytokines, pathogen-associated molecular patterns (PAMPs),⁴ cell-bound ligands, Ags, and physical stresses (1). Expression of NF-κB target genes is essential for mounting innate immune responses to infectious microorganisms (2), but is also important for the proper development and cellular compartmentalization of secondary lymphoid organs necessary to orchestrate an adaptive immune response (3, 4, 5). In mammals, the NF-κB family consists of five members, RelA(p65), RelB, c-Rel, p50(NF-κB1), and p52(NF-κB2), that form hetero- and homodimeric sequence-specific transcriptional regulators (2). Regulation of NF-κB activity occurs at multiple levels, primarily being dependent on inducible degradation of IκB inhibitory proteins, which hold NF-κB dimers in the cytoplasm, and the proteolytic processing of the p105 and p100 precursor proteins to the mature p50 and p52 subunits, respectively (1). The processing of p105 is thought to occur constitutively (1), whereas p100 cleavage is an inducible process (6, 7). IκB degradation depends on site-specific serine phosphorylation, which triggers their polyubiquitination and proteasome-mediated destruction (1). The release of NF-κB from IκBs allows nuclear translocation and binding to specific DNA sequences (κB sites) at the regulatory regions of specific target genes (1, 2). While other mechanisms, some of which are based on post-translational modifications of Rel/NF-κB proteins, do exist, the key regulatory step in the canonical NF-κB activation pathway is phosphorylation of IκB proteins by the IκB kinase (IKK) complex (2).

Three key components of the IKK complex were identified: IKKα, IKKβ, and IKKγ (2). IKKα and IKKβ are catalytic subunits showing significant sequence similarity, whereas IKKγ (also called NEMO) is unrelated in sequence and performs a regulatory role (8). Contrary to initial predictions, however, examination of IKKα-, IKKβ-, and IKKγ-deficient mice has revealed that activation of NF-κB by most proinflammatory stimuli and PAMPs does not require IKKα, but is dependent on IKKβ and IKKγ function (9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19). Ikkβ^−/− mice die in midgestation due to extensive apoptosis of fetal hepatocytes culminating in liver failure (12, 14, 19). This phenotype is shared by Rela^−/− mice and is attributed to TNF-α-induced programmed cell death, as compound mutants defective in either Ikkβ or Rela and either Tnfr1 or Tnfα genes are rescued from embryonic lethality (14, 20, 21). Further examination of Ikkβ^−/−Tnfr1^−/− double-mutant mice revealed reduced thymocyte cellularity that was attributed to impaired thymocyte proliferation (22). Notably, Rela^−/−Tnfr1^−/− mice can survive up to 6 mo before they succumb to opportunistic infections (20), whereas most Ikkβ^−/−Tnfr1^−/− mice die ∼1 wk after birth from rampant infections (Z.-W. Li, unpublished observations). Thus, IKKβ activity is necessary for activation of innate immune responses, which depend in part on RelA (p65)-containing dimers. These findings also suggest that IKKβ is responsible for activation of several different forms of NF-κB besides those that contain RelA. Therefore, its elimination results in a larger decrease in total NF-κB activity than the mere elimination of RelA, thereby causing a more dramatic increase in sensitivity to infections. Although the general role of IKKβ and RelA in activation of innate immune response is fairly well established, there are reasons to believe that these proteins are also involved in adaptive immune responses, probably through effects on lymphocyte development and activation (23, 24).

The large decrease in total NF-κB activity caused by the loss of IKKβ provides an opportunity to learn more about the functions of NF-κB in lymphoid cells and a context for comparing previous findings in mice deficient for specific members of the Rel/NF-κB family. Previous results have shown that the loss of RelA blocks organogenesis of Peyer’s patches and lymph nodes, and splenic microarchitecture is perturbed as well (4). This effect appears to be primarily associated with stromal cell function and may involve signaling through lymphotoxin β receptor (LTβR) and TNF receptor 1 (TNFR1), as RelA/TNFR1 double-deficient hemopoietic cells can efficiently reconstitute lethally irradiated wild-type (wt) recipients (20). Targeted deletion of Nfkb1 results in selective impairment in marginal zone (MZ) B cell formation (25), while Nfkb2^−/− mice have a dramatic, but B cell extrinsic, defect in germinal center formation (3, 5). Similarly, RelB is required for homing as well as proper formation of the splenic microarchitecture, germinal center formation, and MZ organization (26). B cell differentiation in Nfkb1^−/−Nfkb2^−/− mice is blocked at the transitional 1 (T1) stage (IgM^highIgD^low→IgM^highIgD^high) (27), similar to that in Rela^−/−c-Rel^−/− double-mutant mice, which is blocked at the transitional 2 (T2) stage (IgM^highIgD^high) (24), whereas Nfkb1^−/−c-Rel^−/− mice exhibit blocks in B cell activation and function (28). These findings support a role for Rel/NF-κB factors in B cell differentiation and survival, but outstanding questions of functional redundancy of NF-κB members remain. We found that lethally irradiated mice reconstituted with Ikkβ^−/− fetal liver progenitor cells exhibit a complete absence of B and T cells when analyzed 6–8 wk post-transplantation (22). However, further analysis indicated that the absence of T cells is due to the increased sensitivity of IKKβ-deficient thymocytes to TNF-α-induced apoptosis, as stem cells derived from Ikkβ^−/− Tnfr1^−/− double-mutant embryos were capable of reconstituting normal T lymphopoiesis in lethally irradiated mice (22). Nonetheless, IKKβ-deficient T cells display a proliferation defect similar to that exhibited by T cells lacking the protein kinase C θ isozyme, which is required for NF-κB activation in response to ligation of the TCR (29). As the function of NF-κB subunits is perhaps best characterized in B cells (30, 31), we sought to investigate the role of IKKβ in B cell development and function.

In general agreement with the results of deletion of individual NF-κB subunits, we found that B cell-specific depletion of IKKβ results in an increased propensity for spontaneous apoptosis in the absence of antigenic stimulation and decreased proliferative responses to a variety of B cell mitogens. Thus, IKKβ, through activation of the canonical NF-κB pathway, plays a key role in the maintenance and expansion of the B cell compartment. This function is distinct from that of the IKKα catalytic subunit, which is specifically required for late B cell maturation and formation of secondary lymphoid organs (6, 32).

Ikkβ^F/+ mice were generated according to the standard two-step procedure (33). A targeting vector was designed to insert loxP sites and the selection marker genes, Neo^r and TK, into the Ikkβ genomic locus. The floxed 1.65-kb HindIII/XbaI fragment of Ikkβ contains exon 3, which codes for the ATP binding site of the IKKβ kinase domain. A 3.4-kb HindIII fragment and a 5.2-kb XbaI fragment were used as 5′ and 3′ homologous arms, respectively. After electroporation, the transfected embryonic stem (ES) cells (clone GS; Incyte Genomics, St. Louis, MO) were selected with G418 (0.2 mg/ml), and the homologous recombinants were identified by Southern blot analysis of KpnI-digested ES cell genomic DNA using a 1.7-kb XbaI/KpnI fragment downstream of the 3′ homologous arm as the probe. One of seven homologous recombinants was further cultured and transfected with a Cre recombinase (Cre) expression vector to delete the selection marker genes and to generate Ikkβ^F/+ ES cells. Two Ikkβ^F/+ ES clones identified by PCR were injected into E3.5 C57BL/6 blastocysts, and the progeny with the highest degree of chimerism were crossed with C57BL/6 mice to derive lines. PCR genotyping was performed using the primers 5′-GTC ATT TCC ACA GCC CTG TGA-3′ and 5′-CCT TGT CCT ATA GAA GCA CAA C-3′, which amplify both the Ikkβ⁺ (220-bp) and Ikkβ^F (310-bp) alleles. B cell-specific, IKKβ-deficient (CD19CreIkkβ^F/F) mice were generated by breeding Ikkβ^F/F mice with CD19Cre^+/− knockin mice expressing Cre under control of the endogenous CD19 promoter (34).

Six- to 12-wk-old mice were sacrificed, and bone marrow, spleen, lymph nodes, peripheral blood, and peritoneal fluid were collected. Resting splenic B cells were isolated by depletion of CD43⁺ cells with magnetic beads (MACS; Miltenyi Biotec, Auburn, CA) and were used for all experiments except Southern blot analysis. For Southern blot analysis, bone marrow and spleen single-cell suspensions were purified with anti-B220 beads. Alternatively, B cells were purified by complement lysis of T cells using anti-Thy1.2 Abs (FD75 and HO11.34) and rabbit complement (Cedarlane, Hornby, Canada). Resting B cells isolated by all methods were at least 90% B220⁺ as verified by flow cytometry.

Single-cell suspensions were prepared from bone marrow, spleen, lymph node, peripheral blood, and peritoneal cavity of wt (+/+) and CD19CreIkkβ^F/F (F/F) mice. Bone marrow and spleen were depleted of RBC by hypotonic lysis with ACK buffer, and PBMC were isolated by Ficoll (Pharmacia Biotech, Piscataway, NJ) gradient centrifugation before staining. The Abs used for staining include anti-B220-PE, anti-B220-CyChrome, anti-Thy1.2-CyChrome, anti-CD3-FITC, anti-CD5-biotin, anti-CD23-PE, anti-CD21-FITC, anti-CD43-biotin, anti-IgM^a-PE, anti-IgM^b-PE, anti-IgD-FITC, anti-CD24/heat-stable Ag (HSA)-biotin (M1/69), and streptavidin-PerCP (all purchased from BD PharMingen, San Diego, CA) and anti-B220-allophycocyanin and streptavidin-Tricolor (Caltag Laboratories, Burlingame, CA). Flow cytometric analysis was performed using a FACScan or a FACSCalibur and CellQuest software (BD Biosciences, Mountain View, CA).

Cryostat sections (8–10 μm) prepared from OCT (Tissue-Tek; Sakura, Torrance, CA)-embedded spleen were dried briefly and then stored in humidified chambers overnight at 4°C. Sections were fixed in ice-cold acetone for 10 min, dried, then blocked for ≥1 h at room temperature with PBS supplemented with 5% FBS. Sections were incubated with anti-IgM^a-PE and anti-IgD-FITC (BD PharMingen) in blocking buffer for 2 h at room temperature, followed by three washes with PBS supplemented with 0.05% Tween 20. Slides were mounted in Gel/Mount (Biomeda, Hayward, CA) and viewed under a fluorescent microscope (E800; Nikon, Melville, NY).

Resting B cells were cultured at 2–5 × 10⁶/ml in six-well plates for 2 h at 37°C before being stimulated with 20 μg/ml LPS (Sigma-Aldrich, St. Louis, MO) for 1 h. Total cell lysates were used for Western blotting and kinase assays as previously described (12). The Abs used are anti-IKKγ (BD PharMingen; clone 73-764) for immune precipitation; anti-IKKα (Imgenex, San Diego, CA; clone 14A231) and anti-IKKβ (Upstate Biotechnology, Lake Placid, NY; clone 10AG2) for Western blot analysis.

Resting B cells (10⁶/ml) at 10⁵/well in triplicate in 96-well plates were cultured for 48 h in the presence of complete RPMI medium alone, IL-4 (5 ng/ml; BD PharMingen), anti-IgM (10 μg/ml; Jackson ImmunoResearch Laboratories, West Grove, PA), LPS (20 μg/ml; Sigma-Aldrich), or anti-CD40 (5 μg/ml, clone 3/23; BD PharMingen), respectively. [³H]thymidine (5 μCi/ml; Amersham Pharmacia Biotech, Arlington Heights, IL) was added for the final 16 h of the culture, and incorporation was measured by scintillation counting. For CFSE-labeled in vitro proliferation assay, B cells (10⁷/ml) were incubated in the dark with 5 μM CFSE (Molecular Probes, Eugene, OR) for 10 min in PBS at 37°C, washed with medium, and cultured with the additives as described above for 3 days before flow cytometric analysis. To measure apoptosis, cells stimulated for 2 days under the aforementioned conditions were incubated for 10 min on ice with hypoPI buffer (0.1% sodium citrate, 0.1% Triton X-100, 50 μg/ml RNase A, and 100 μg/ml propidium iodide (PI)), and the fraction of subdiploid cells was measured by flow cytometry.

Genomic DNA was isolated (DNeasy Tissue Kit, Qiagen) from resting B cells cultured for 0, 1, and 2 days. One microliter of serially diluted genomic DNA (12.5, 25, 50, and 100 ng) was added to 24 μl of a mixture of Sybr Green PCR Master Mix (12.5 μl), primers (1 μl at 25 μM each), and H₂O (10.5 μl). Real-time PCR was performed for 40 cycles at 95°C for 15 s and at 60°C for 1 min using an ABI PRISM 7700 Sequence Detector (PE Applied Biosystems, Foster City, CA). Primers 5′-TAG TCC AAC TGG CAG CGA ATA C-3′ and 5′-CGC CTA GGT AAG ATG GCT GTC T-3′ were used to amplify the Ikkβ^Δ allele, primers 5′-AAG ATG GGC AAA CTG TGA TGT G-3′ and 5′-CAT ACA GGC ATC CTG CAG AAC A-3′ were used to amplify the Ikkβ^F allele, and primers 5′-ATT CGC CAA TGA CAA GAC GCT GG-3′ and 5′-GGC TGC AGT CCA CGC ACT GG-3′ were used to amplify the Tnfr1 gene as a control. The ratio of Ikkβ^Δ and Ikkβ^F was calculated after normalization to the Tnfr1 signal.

Mice were administered BrdU (Sigma-Aldrich; 1 mg/ml plus 2% glucose) in drinking water for 3 or 6 days before sacrifice. B220⁺ B cells were isolated from bone marrow and spleen by MACS, permeabilized, and stained with anti-BrdU-FITC or an FITC-conjugated isotype-matched control Ab (BD PharMingen) in accordance with the company protocol.

Splenocytes of CD19CreIkkβ^F/F and Ikkβ^F/F mice were labeled with CFSE (1 μM) and washed with PBS. Two hundred microliters of labeled cells at 1–2 × 10⁸/ml were injected i.v. into 8 wk old 129 × C57BL/6 F₁ male mice (The Jackson Laboratory, Bar Harbor, ME). Labeled cells were also stained with anti-Thy1.2-PE and anti-B220-CyChrome and analyzed by flow cytometry before transfer. Seven days later splenocytes from the recipients were stained with anti-Thy1.2-PE and anti-B220-CyChrome and analyzed by flow cytometry. The ratio of B220⁺ to Thy1.2⁺ CFSE-positive cells was calculated and normalized to the ratio before transfer.

Mice were immunized with T-independent (nonencapsulated, type 2 Streptococcus pneumoniae (R36A), 10⁸ CFU/mouse) or T-dependent (2,4,6-trinitrophenyl (TNP)-OVA, 100 μg/mouse) Ags by i.p. injection. Serial dilutions of preimmune and immune sera from immunized animals were incubated on 96-well microtiter plates coated with goat anti-mouse IgM, goat anti-mouse IgG (both from Southern Biotechnology Associates, Birmingham, AL), TNP-BSA (Biosearch Technologies, Novato, CA), or phosphocholine-BSA (gift from Dr. G. Silverman, University of California-San Diego). Plates were developed using alkaline phosphatase-conjugated anti-mouse Igκ, anti-mouse IgM, and anti-mouse IgG and the substrate p-nitrophenylphosphate (all from Southern Biotechnology Associates) and were read at 405 nm.

To study the role of IKKβ in B cell development, function, and survival, mice were generated bearing a conditional Ikkβ allele (Ikkβ^F) subject to inactivation by Cre recombinase-mediated deletion (33). This was accomplished by construction of a gene targeting vector that introduces loxP sites into regions flanking exon 3 of the Ikkβ gene following homologous recombination in ES cells (Fig. 1 A). Homologous recombinants were identified by PCR and confirmed by Southern blot analysis. Selected recombinants were subjected to transient Cre expression and selected on the basis of acquired G418 resistance. ES cells that underwent a partial deletion to generate a loxP-flanked (floxed, F) exon 3 were chosen for blastocyst injection and eventual germline transmission of the conditional Ikkβ^F allele. Mice homozygous for the Ikkβ^F allele were viable and phenotypically normal, confirming that insertion of loxP sequences did not alter Ikkβ gene function.

To obtain B cell-specific inactivation of Ikkβ, Ikkβ^F/F mice were bred with CD19Cre (CD19^+/−, Cre⁺) mice that express Cre in the B lineage by virtue of targeted insertion of Cre into the Cd19 locus (34). As shown in Fig. 1,B, recombination of the Ikkβ^F allele is made evident by the appearance of a deletion product (Ikkβ^Δ) first detected in developing B cells in the bone marrow and more highly represented in the mature population of splenic B cells. B cells from homozygous (Ikkβ^F/F) mice are rendered IKKβ-deficient in the presence of Cre, which is evident from reduced IKKβ protein levels and total IKK kinase activity in splenic B cells following LPS stimulation (Fig. 1,C). Nonetheless, a significant proportion of B cells isolated from CD19CreIkkβ^F/F mice retain a functional Ikkβ^F allele. This finding does not appear to be due to poor deletion efficiency or specificity of Cre expression, since splenic B cells from mice heterozygous for the conditional allele (CD19CreIkkβ^+/F) show almost complete conversion of the Ikkβ^F allele to the Ikkβ^Δ form (Fig. 1,B). Rather, it suggests that the loss of IKKβ poses a severe disadvantage to B cell survival, thus selecting for variants that have managed to avoid deletion. Support for this argument is provided by the assessment of deletion efficiency in splenic B cells from young, adult, and aged mice (Fig. 1 B), showing a greater accumulation of cells bearing an intact Ikkβ^F allele(s) and decreased representation of the Ikkβ^Δ allele in older mice.

Flow cytometric analysis of bone marrow cells indicated that B cell generation was not affected in CD19CreIkkβ^F/F mice (data not shown). We therefore focused on the effect of IKKβ deletion in peripheral B cells. CD19CreIkkβ^F/F mice were compared with CD19CreIkkβ^+/+ cohorts to assess the composition of the B cell compartment in secondary lymphoid organs and peripheral blood. As shown in Fig. 2,A, the relative frequency of B cells was significantly reduced in all examined peripheral compartments of CD19CreIkkβ^F/F mice. The most dramatic reduction (5-fold) in B cell number was seen in the lymph nodes, which primarily contain long-lived recirculating B cells (IgM^low, IgD^high, CD23^high, CD21^int). The B cell deficiency was also evident in the peritoneal B-1 cell population (Fig. 2 A), suggesting that IKKβ is necessary for the persistence of B-1 as well as conventional (B-2) cells.

In addition to recirculating B cells present in the follicles, the splenic B cell compartment includes MZ (IgM^high, IgD^low, CD23^neg, CD21^high) B cells, which are long-lived cells that do not recirculate but occupy the region of the spleen peripheral to the marginal sinus. Formation of MZ B cells is notably impaired in NF-κB1^−/− mice and, to a lesser extent, c-Rel^−/− and Rela^−/− mice (24, 25). B cells from CD19CreIkkβ^F/F mice express lower levels of CD23, and it appears that the B220^high, CD23^neg, CD21^high MZ B cell population is reduced (Fig. 2,B). To substantiate this finding, we assessed the representation of MZ B cells in situ using immunohistologic staining to detect follicular (IgD^high) and MZ (IgM^high) B cells (Fig. 2 C). In agreement with the flow cytometric analysis, B cells from CD19CreIkkβ^F/F mice are capable of populating the MZ, but are under-represented relative to those in control CD19CreIkkβ^+/+ mice. Insofar as MZ B cells are thought to be an activated subset primed by internal Ags, the observed reduction may be due to impaired activation or survival in the absence of IKKβ.

B cells undergo clonal expansion in response to microbial products and T cell-derived cytokines or upon efficient cross-linking of membrane Ig molecules. To evaluate the response of IKKβ-deficient B cells to key B cell mitogens and growth factors, purified splenic B cells from CD19CreIkkβ^F/F and CD19CreIkkβ^+/+ mice were incubated with optimal concentrations of anti-IgM F(ab′)₂, LPS, anti-CD40, IL-4, or medium alone. As measured by [³H]thymidine incorporation, B cells of CD19CreIkkβ^F/F mice were hypoproliferative to all stimuli (Fig. 3,A). To confirm and extend these findings, B cells were labeled with the fluorescent membrane dye CFSE and subjected to the same set of stimuli. We observed that most B cells from CD19CreIkkβ^F/F mice failed to undergo cell division (Fig. 3,B). The minor population that does respond probably represents B cells that have retained a functional Ikkβ^F/F allele. Moreover, when analyzed for DNA content by PI staining, a reduced percentage of CD19CreIkkβ^F/F B cells were in cycle, but an increased fraction contained subdiploid DNA indicative of apoptosis (Fig. 3,B). IL-4 is nonmitogenic, but appears to confer increased cell survival in an IKKβ-independent manner when other costimuli are not present. These findings were confirmed by staining for annexin V as an indicator of apoptosis (data not shown). Real-time PCR analysis revealed a decreased presence of the Ikkβ^Δ allele with time in culture relative to the Ikkβ^F allele (Fig. 3 C), suggesting that homozygous deletion of Ikkβ results in accelerated cell death. Thus, reduced [³H]thymidine incorporation in IKKβ-deficient B cells is attributed to both reduced proliferation and increased cell death.

Most recirculating B cells in the adult animal are relatively long-lived, possessing a half-life of several weeks. In the spleen, less mature B cells representing recent emigrants from the bone marrow and mature MZ B cells are phenotypically identified as expressing high levels of IgM and HSA (CD24), whereas recirculating mature follicular B cells are IgM^low/HSA^low. In CD19CreIkkβ^F/F mice, a lower proportion of splenic B cells are IgM^low/HSA^low, while the absolute number of IgM^high/HSA^high B cells is modestly decreased (Fig. 4,A and data not shown), most likely representing the reduction in MZ B cells in CD19CreIkkβ^F/F mice as described above. This finding suggests that B cells from CD19CreIkkβ^F/F mice enter the peripheral B cell pool fairly efficiently, but are short-lived. To address this issue, mice were administered BrdU in the drinking water for 3 or 6 days, at which time B220⁺ cells were isolated from bone marrow and spleen and analyzed by flow cytometry for BrdU incorporation during the labeling period. The percentage of BrdU-labeled bone marrow B cells was similar in CD19CreIkkβ^F/F and CD19CreIkkβ^+/+ mice (data not shown). However, we observed a significant increase in the percentage of BrdU-labeled spleen B cells in CD19CreIkkβ^F/F mice (Fig. 4,B), indicating that B cell turnover is increased in the absence of IKKβ. To confirm this result, total spleen cell suspensions from CD19CreIkkβ^F/F and CD19CreIkkβ^+/+ mice were labeled with CFSE, and the relative percentages of B cells and T cells were enumerated by flow cytometry. The CFSE-labeled cells were then injected i.v. into MHC-matched, nonirradiated recipients. At 7 days post-transfer, spleens were harvested from the recipients and analyzed by flow cytometry for the relative distribution of donor (CFSE-positive) B and T cells. The B to T cell ratio was calculated and normalized to that measured before transfer. As indicated by the unaltered B/T cell ratio in transferred and recovered CD19CreIkkβ^+/+ cells, wild-type T and B cells showed a similar degree of survival over the 7-day period. However, a precipitous and specific loss occurred with transferred CD19CreIkkβ^F/F B cells (Fig. 4 C), thus supporting the idea that IKKβ-deficient B cells are prone to apoptosis, resulting in their rapid disappearance.

Measurement of serum Ig levels by ELISA revealed decreases in the basal levels of IgM and IgG in CD19CreIkkβ^F/F mice (Fig. 5,A). The reduced IgM levels are probably indicative of the reduced B-1 and MZ B cell populations that contribute a significant proportion of the natural Ab titer. To examine the responsiveness of Ikkβ^F/F mice to intact bacteria, mice were immunized with nonencapsulated, type 2 S. pneumoniae (R36A) and monitored for phosphocholine-specific Ab production 7 days postimmunization. We found that CD19CreIkkβ^F/F mice mounted a poor response to R36A (Fig. 5,B), consistent with the reduction in MZ and B-1 cells and minimal responses to anti-IgM stimulation in vitro. The Ab response to TD Ags was also examined by immunization with TNP-OVA. We found that CD19CreIkkβ^F/F mice were greatly impaired in the production of TNP-specific Abs (Fig. 5 C), which is consistent with an impaired response to anti-CD40 and a general B cell survival defect. Of note, the reduction in TD Ab production was more dramatic than predicted from the fraction of resting splenic B cells that retained IKKβ expression. Thus, in agreement with the in vitro stimulation data, it appears that in addition to decreased survival of IKKβ-deficient B cells, their decreased mitogenic response contributes to the overall decrease in Ab production.

Characterization of the IKK complex led to the identification of two catalytic subunits, IKKα and IKKβ, which in vitro can phosphorylate IκB proteins at sites that are required for ubiquitin-dependent IκB degradation and translocation of NF-κB dimers into the nucleus (1). Surprisingly, however, targeted deletion of the Ikkα gene revealed minimal defects in NF-κB activity induced by proinflammatory stimuli, but an absolute requirement for IKKα in epidermal differentiation (11, 13, 18). By contrast, Ikkβ^−/− mice die in midgestation due to TNF-α-induced apoptosis of hepatocytes (12, 14, 19). This phenotype is shared by Rela^−/− mice (36) and underscores the important antiapoptotic function of NF-κB (37). When the Ikkβ null mutation is bred onto the Tnfr1^−/− background, double-mutant mice are rescued from embryonic lethality (14, 22). However, most of these mice die before weaning as a result of severe infections (Z.-W. Li, unpublished observations), thus precluding the analysis of Ikkβ function in the adult. We overcame this difficulty through the generation of a conditional Ikkβ allele and used this system to probe Ikkβ function in B lymphocytes. We found that Ikkβ is required for B cell survival and mitogenic responsiveness to diverse stimuli, including those that act via the B cell receptor.

To ablate IKKβ function in B cells, mice expressing a floxed Ikkβ^F allele were generated and bred with mice expressing Cre recombinase under transcriptional control of the CD19 promoter (34). In this system the onset of Cre expression occurs at the pro-B cell stage and continues throughout B cell development and differentiation. Correspondingly, CD19CreIkkβ^F/F mice show a strong reduction in peripheral B cell numbers that affects all B cell subsets, including MZ and B-1 cells. By contrast, early B cell development in the bone marrow was not affected, as determined by flow cytometry and BrdU labeling experiments. These findings are consistent with the phenotype of Nfkb1^−/−Nfkb2^−/− and Rela^−/−c-Rel^−/− double mutants, which show intact generation of IgM-positive immature B cells, but a strong impairment in the generation or maintenance of mature B cells (24, 27). Thus, while early B cell development in bone marrow is intact, B cell homeostasis in peripheral lymphoid tissues of CD19CreIkkβ^F/F mice is impaired, resulting in increased turnover of IKKβ-deficient B cells.

A significant proportion of the mature B cells in CD19CreIkkβ^F/F mice were found to retain at least one functional Ikkβ allele. Since B cells from CD19CreIkkβ^+/F mice showed complete deletion of the floxed allele, it became evident that Cre-mediated deletion of the Ikkβ^F allele is efficient and that strong selection is imposed to deplete CD19CreIkkβ^F/F B cells that had deleted both Ikkβ alleles. The interpretation that IKKβ is essential for B cell survival is supported by our finding that B cells from CD19CreIkkβ^F/F mice showed increased turnover in vivo, as measured by BrdU labeling kinetics and tracking of transferred CFSE-labeled cells. Moreover, cultured CD19CreIkkβ^F/F B cells showed a considerable increase in the rate of apoptosis, as measured by DNA staining and as inferred by the decreased presence of B cells bearing an Ikkβ^Δ allele. These in vitro studies indicate that IKKβ-deficient B cells possess an intrinsic defect in survival as opposed to increased susceptibility to an extrinsic factor. In the specific case of TNF-α-induced apoptosis, this conclusion is supported by our findings that IKKβ-deficient B cells from CD19CreIkkβ^F/F mice bred onto the Tnfr1^−/− background do not show any increased survival capacity relative to B cells from CD19CreIkkβ^F/F mice (Z.-W. Li, unpublished observations). Thus, IKKβ-dependent protection from TNF-α may be necessary for thymocyte survival, but does not appear to be applicable to B cells. These findings suggest that the basal level of nuclear NF-κB in resting B cells is biologically significant and is required for continued synthesis of antiapoptotic proteins (37).

NF-κB is activated downstream of a diverse group of surface proteins that include members of the Toll, TNF, and Ig receptor families (38, 39). We found that IKKβ is essential for signaling by representative members of these families, in that CD19CreIkkβ^F/F B cells showed severe defects in the proliferative response to anti-IgM, LPS, or anti-CD40 stimulation. These stimuli activate NF-κB by distinct upstream adaptors and kinases, not all of which are fully elucidated (38, 39). However, all these stimuli appear to promote B cell proliferation by using the IKKβ-dependent canonical NF-κB activation pathway that cannot be fully activated by IKKα (2). Reduced CD19CreIkkβ^F/F B cell proliferation as measured by [³H]thymidine incorporation does not appear to be due simply to increased cell death. We derive this conclusion from the observation that gating on live CFSE-labeled CD19CreIkkβ^F/F B cells reveals little evidence of cell division, and the finding that stimulated, viable CD19CreIkkβ^F/F B cells do not show increased cell size. These findings are in agreement with in vivo defects in Ab responses to both TI-2 and TD Ags. Together, these data indicate that IKKβ-dependent activation of NF-κB is necessary for proliferation induced by polyclonal mitogenic stimuli and for Ag-driven clonal expansion.

The observed defects in CD19CreIkkβ^F/F mice are most consistent with the combined loss of Nfkb1 and c-Rel as well as Rela and c-Rel. These mice also have reduced follicular, B-1, and MZ B cell populations; increased B cell turnover; and a failure to proliferate in response to anti-IgM or LPS (24, 28). Single mutants in c-Rel or Nfkb1 show less severe and selective defects (40, 41), and c-Rel^−/− B cells are not prone to apoptosis in the quiescent state. However, mitogenic stimulation with anti-IgM, LPS, or anti-CD40 results in increased apoptosis due to an apparent failure to up-regulate Bcl-x_L and A1 (42, 43, 44). Nfkb1^−/− B cells are also impaired in responding to mitogenic stimuli, but have an additional defect in cell survival in the quiescent state. By contrast, RelA is not necessary for B cell autonomous functions in vivo (4), and Rela^−/− B cells respond normally to LPS, CD40, and B cell receptor stimulation in vitro (23). Instead, RelA appears to be required in nonhemopoietic tissues for the formation and organization of secondary lymphoid tissues (4). In the absence of both Nfkb1 and Rela, early B cell development is halted; however, this defect is not B cell autonomous (45).

This report and other recent findings from mice deficient for IKKα or expressing mutated forms of IKKα or IKKβ complex (6, 32, 46, 47) have led to a new appreciation for the individual functions of these components in B cells. Loss of Ikkα also results in reduced cellularity of the B cell compartment in fetal liver-reconstituted mice (6, 32). However, the nature and severity of the defect are more closely allied with a partial arrest in B cell maturation rather than acute responses to inflammatory stimuli (6, 32). Indeed, IKKα is not required for activation of the canonical NF-κB signaling pathway in B cells and instead is required for the NIK-dependent induction of p100 processing (6). Thus, it is possible that hyporesponsiveness of Ikkα^−/− B cells to mitogens is attributed to their immature nature. Unlike IKKβ-deficient mice, Ikkα^−/− chimeric mice exhibit striking defects in lymph node organogenesis and splenic cellular architecture (6, 32). These defects may be B cell dependent, yet rely mainly upon the maintenance of a proper stromal environment. A major challenge for the future entails the identification of the target genes responsible for the distinct biological functions of the two NF-κB activation pathways.

While this paper was under review, Pasparakis et al. (48) reported on the role of IKKβ in B cells using a similar genetic approach. In agreement with our findings, these authors show that total peripheral B cell numbers are reduced in the absence of IKKβ due to specific losses in the follicular and MZ B cell compartments. Using in vivo anti-IL-7R mAb treatment, they show that the remaining peripheral B cells have either only recently deleted the second Ikkβ^F allele and still presumably retain some IKKβ protein or have escaped deletion and still retain one intact Ikkβ^F allele. Importantly, the data presented by Pasparakis et al. (48) and our present study clearly demonstrate that in addition to its established role in suppression of receptor-induced apoptosis (37), the IKKβ-dependent canonical NF-κB signaling pathway is also of importance for sustaining nonstimulated resting B cells. Additionally, we have shown that IKKβ may also play a role in B cell proliferation and activation in response to various mitogenic stimuli and both TI and TD Ags. Dissecting IKKβ-dependent survival vs activation mechanisms remains a challenging topic for future study.

We thank Dr. Dennis Otero and Kirsten Vroom for their assistance with some of the experiments, and Dr. Gregg Silverman for providing the S. pneumoniae (R36A) extract and PC.