Mice rendered deficient in p52, a subunit of NF-κB, or in Bcl-3, an IκB-related regulator that associates with p52 homodimers, share defects in the microarchitecture of secondary lymphoid organs. The mutant mice are impaired in formation of B cell follicles and are unable to form proper follicular dendritic cell (FDC) networks upon antigenic challenge. The defects in formation of B cell follicles may be attributed, at least in part, to impaired production of the B lymphocyte chemoattractant (BLC) chemokine, possibly a result of defective FDCs. The p52- and Bcl-3-deficient mice exhibit additional defects within the splenic marginal zone, including reduced numbers of metallophilic macrophages, reduced deposition of the laminin-β2 chain and impaired expression of a mucosal addressin marker on sinus-lining cells. Whereas p52-deficient mice are severely defective in all of these aspects, Bcl-3-deficient mice are only partially defective. We determined that FDCs or other non-hemopoietic cells that underlie FDCs are intrinsically impaired in p52-deficient mice. Adoptive transfers of wild-type bone marrow into p52-deficient mice failed to restore FDC networks or follicles. The transfers did restore metallophilic macrophages to the marginal zone, however. Together, the results suggest that p52 carries out functions essential for a proper splenic microarchitecture in both hemopoietic and non-hemopoietic cells and that Bcl-3 is important in enhancing these essential activities of p52.

The NF-κB transcription factor family is intimately involved in numerous cellular responses, especially those associated with stress, injury, and pathogens (1, 2, 3, 4). Upon pathogenic challenge, NF-κB factors are activated and can induce the expression of many genes that encode proteins critical to the defense of the organism. Mice rendered deficient for various NF-κB subunits exhibit diverse immunologic defects (5, 6, 7, 8, 9, 10, 11, 12). The intimate relationship between this transcription factor and immune responses is quite ancient in evolutionary terms, since the Drosophila host defense responses to pathogens have been shown to depend on NF-κB protein homologues (13). NF-κB factors are present in many cell types, where they may be activated by signals derived from pathogenic challenge, either directly or indirectly.

NF-κB is a family of dimeric complexes composed of closely related polypeptides that share a Rel homology domain (1). This domain of about 300 amino acids encodes DNA-binding, dimerization, and nuclear localization functions. In mammalian systems, the family of complexes is composed of the subunits p50, p52, RelA (p65), c-Rel, and RelB. The prototypical complex is a p50-RelA heterodimer, but most of the other possible dimer combinations have been observed as well, depending on cell type and activation status. One exception is RelB, which is known to dimerize with p50 or p52 only. RelA, c-Rel, and RelB have definite trans-activation domains, while p50 and p52 do not. Therefore, all dimers that include the former subunits can directly trans-activate κB site- dependent reporters. NF-κB activation is regulated primarily through inhibitory IκB proteins, which normally complex with and retain the transcription factor in the cytoplasm. These inhibitors can be rapidly degraded in response to a wide variety of signals, thus freeing the complexes to translocate into the nucleus and carry out their functions (14, 15, 16, 17).

Although the various NF-κB factors have redundant functions, they also harbor unique functions, as evidenced by defects in mice lacking individual factors. We as well as others have previously generated mice deficient in the p52 subunit (9, 10). These mice have impaired Ab responses to T-dependent Ags. The reason for this may relate to a partially impaired microarchitecture in spleen and lymph nodes that no longer supports proper germinal center formation. Germinal centers are the principal sites where Ag-activated B cells undergo affinity maturation and isotype switching, and where both plasma and memory B cells are generated (18). Other subunits therefore are unable to substitute for some activities conducted by p52, this despite the fact that the p50 subunit is structurally and functionally very similar to p52. In addition, p50 is abundantly and apparently ubiquitously expressed (1, 19). Although p50 knockout mice have certain defects associated with B cells (5), these mice do not appear to have the problems with lymphoid microarchitecture noted in the p52 knockout mice (9, 10). One important way in which these two subunits may differ relates to their interaction with the Bcl-3 protein. Based on its structure, Bcl-3 is a member of the IκB family, but functionally it appears to be quite distinct. Rather than inhibit, Bcl-3 may be able act as a transcriptional coactivator, as shown in transient transfection experiments (20, 21). Although Bcl-3 can complex with both p50 and p52, its interaction with p52 homodimers in particular appears to lead to potent trans-activation of a κB-dependent reporter construct in some cell types. However, no specific physiologic target of this unusual association of p52 and Bcl-3 has been defined to date. Nevertheless, consistent with a shared function, the defects observed in mutant mouse strains lacking p52 or Bcl-3 are similar (9, 22, 23). In contrast, mutant mice lacking p50 did not share any obvious defects with those lacking Bcl-3. Thus, p52 and Bcl-3 may together have a unique function, one not shared with p50.

In the present study, we explore the phenotypes of mice lacking p52 or Bcl-3 to identify the cells directly affected by the lack of these proteins. Ultimately this will set the stage to elucidate molecular mechanism(s). Adoptive transfer of wild-type bone marrow into p52 knockout animals revealed that at least two cell types are responsible for the observed defects. A non-hemopoietic cell, most likely the follicular dendritic cell (FDC)2 itself, is critically dependent on p52 in formation of B cell follicles and of FDC networks. In addition, certain marginal zone (MZ) defects appear to be the result of an impaired non-hemopoietic cell. On the other hand, a hemopoietic cell is critically dependent on p52 for the appearance of metallophilic macrophages in the MZs, a subset of macrophages that is lacking in p52 knockout mice (9, 10, 24). We also demonstrate that Bcl-3 significantly enhances the critical activities of p52 in these cells, but that it is not absolutely required.

Mice deficient in p52 or Bcl-3 protein were generated by targeted disruption in embryonic stem (ES) cells as described previously (9, 22).

2,4,6-Trinitrophenyl-keyhole limpet hemocyanin (TNP-KLH), SRBC, and HRP (Sigma, St. Louis, MO) were used to elicit T-dependent immune responses in mice. SRBC were washed three times in PBS, and then a 10% solution in PBS was made. Mice were injected with 250 μl i.p. of this solution and sacrificed 7 days later. TNP-KLH was prepared as described previously (9). Briefly, 20 mg of lyophilized KLH (Pierce, Rockford, IL) was dissolved in 4 ml of potassium borate buffer (0.25 M at pH 9.2), and 3 mg of TNBS (2,4,6-trinitro-benzyl sulfonic acid; Sigma) was added along with 16 μl of sodium carbonate (1 M). The reaction was allowed to take place overnight, after which the protein derivatives were dialyzed against PBS (pH 7.4). TNP-KLH conjugates were frozen at −20°C until use.

The following Abs/markers were used at a final concentration of 10 μg/ml: anti-B220 (clone RA3-6B2), anti-CD35 (clone 8C12), anti-mucosal addressin cell adhesion molecule-1 (MAdCAM-1) (clone MECA-367) (all from PharMingen, San Diego CA), anti-mouse laminin β-2 chain (Upstate Biotechnology, Lake Placid, NY), anti-p52, and anti-Bcl-3 (Santa Cruz, Santa Cruz, CA). MOMA-1 (BACHEM Bioscience, Torrence, CA) and HRP-conjugated peanut agglutinin (PNA) were used at a 1:20 dilution (PNA; Dako, Carpinteria, CA). Anti-FDC Abs FDC-M1 and FDC-M2 were diluted 1:150 and 1:200, respectively (produced in house). Biotinylated goat anti-rat and goat anti-rabbit Abs (Vector Laboratories, Burlingame, CA) were used as secondary Abs at a 1:1000 dilution.

Bone marrow cells isolated from femora of 6-wk-old wild-type mice were injected (3 × 106 cells per mouse) into either p52 knockout recipients or control littermates (6–18 wk old); recipient mice had been irradiated with 900 rad or 600 rad 24 h previously. Seven to eight weeks after the adoptive transfers, recipients were injected i.p. with TNP-KLH (100 μg) adsorbed to alum. For cryosections, spleen and lymph nodes were collected 9 days after immunization.

Peripheral lymph nodes were collected from mice immunized with SRBC and then digested with an enzyme mixture. Briefly, the lymphatic tissue was incubated for 45 min at 37°C with RPMI 1640 (Life Technologies) containing collagenase class IV (450 U/ml, Worthington, Freehold, NJ) and DNase I while being stirred constantly. At 15-min intervals, the released cells were collected, and a new stock of enzyme solution was added to the remaining nondissolved tissue fragments. The final cell suspension was washed with RPMI 1640 to remove any nondigested tissue debris, overlayed onto a continuous Percoll gradient, and centrifuged at 600 × g for 30 min. Low density cells were collected, twice washed in RPMI 1640, and incubated with FDC-M2 Ab for 15 min at 4°C. Subsequently, the cell preparation was washed in RPMI 1640 and incubated with goat anti-rat IgG Microbeads (Miltenyi Biotec, Bergisch-Gladbach, Germany) for 15 min at 4°C and applied to a MiniMACS separation column. Unlabeled cells were eliminated with washes of RPMI 1640. Cells attached to the column by conjugation with the Microbeads were eluted after removal of the magnetic field by rinsing with RPMI 1640.

Single color IHC.

Spleens were extracted, placed in OCT freezing medium (Miles Laboratories, Elkhart, IN) and flash frozen. Ten-micrometer thick acetone-fixed sections were processed as described previously. Briefly, tissue sections were rehydrated in PBS and then blocked for 30 min with Dako protein block. After blocking, sections were incubated for 30 min in the presence of relevant Ab (60 min in the case of FDC-M2), diluted in Dako Ab diluent, washed in PBS, and incubated an additional 30 min with biotinylated goat anti-rat secondary Ab in Dako Ab diluent. After quenching endogenous peroxide activity in 0.3% H2O2 for 15 min, tissue sections were incubated for a further 30 min with HRP-conjugated streptavidin (Vector Laboratories). Slides were then washed in PBS, and the avidin/biotin complexes were revealed with diaminobenzidine (DAB) tetrachloride chromogen (Vector Laboratories) according to the manufacturer’s instructions. Finally, slides were rinsed, counterstained with Methyl green (Vector Laboratories), and mounted with Permount (Fisher Scientific, Pittsburgh, PA).

Two color IHC.

Double IHC was performed on acetone-fixed cryosections as described previously (9). Briefly, tissue sections were dehydrated in PBS for 10 min, blocked for 20 min in Dako protein block, and subsequently incubated for 30 min in the presence of the first biotinylated primary Ab, followed by a 30 min incubation with alkaline phosphatase (AP)-conjugated streptavidin diluted 1:50 (Vector Laboratories). The sections were then incubated for a further 30 min in the presence of the next relevant Ab, conjugated with HRP (except for PNA, which was for 1 h). As a final step, HRP-conjugated streptavidin was applied for 30 min. Endogenous peroxidase activity was quenched with 0.3% H2O2 in PBS for 15 min. All washings were done as described above, and all incubations were at room temperature. Alkaline phosphatase and HRP enzymatic activities were finally revealed with the Fast Red (Dako Corporation) and DAB chromogens, respectively, and specimens were mounted in aqueous mounting medium (Dako Corporation).

Unfixed, frozen tissue sections were washed in PBS two times for 20 min and blocked in 8% BSA in PBS for 1 h at room temperature. Sections were then washed with PBS four times, 15 min each at room temperature. Anti-mouse laminin β-2 chain Ab (10 μg/ml) diluted in 1% albumin in PBS was applied overnight at 4°C. Subsequently, sections were washed twice for 30 min, quenched with 0.3% H2O2, and further incubated in the presence of goat anti-rat secondary Ab diluted at room temperature for the next 1 h. After three washings, 15 min each, the HRP reaction was revealed with DAB.

Acetone-fixed tissue sections were washed in PBS two times for 10 min and then blocked in the same way as described for laminin. After extensive washings in PBS, sections were permeabilized with 0.5% Triton X-100 for 5 min at room temperature, followed by two washes in PBS, 15 min each. Rabbit anti-mouse p52 or rabbit anti-mouse Bcl-3 Abs diluted in PBS/1% BSA/0.01% Triton X-100 were then applied for 2 h at room temperature. Sections were quenched with 0.3% H2O2 in PBS for 15 min. Biotinylated goat anti-rabbit Ab diluted in PBS/1% BSA/0.01% Triton X-100 was applied for 30 min followed by HRP-conjugated streptavidin. All incubations were done at room temperature. After the final washes, the HRP reaction was revealed with DAB.

RNA from splenocytes was isolated using Trizol reagent (Life Technologies, Grand Island, NY). cDNAs were synthesized using the Clontech (Palo Alto, CA) PCR kit starting with the same quantity of RNA (1 μg). The following primers were used: for B lymphocyte chemoattractant (BLC), 5′ primer, TCA CCT AGG ATG AGG CTC AGC ACA GCA, and 3′ primer, CAC TCA TTC TCT TCT CGA CGG GAA TTC CAC (the amplified PCR fragment size is 364 bp); for B220, 5′ primer, GTG TAC AGC TGA TCT GGG ACG TGA AC, and 3′ primer, GAA GAT AAT AGT TGA AAG TTT ATT ATG G (the amplified PCR fragment size is 520 bp); and for GAPDH, 5′ primer, GGT GAA GGT CGG TGT GAA CGG A, and 3′ primer, TGT TAG TGG GGT CTC GCT CCT G. PCRs were optimized for each set of primers and were performed using different numbers of cycles to ensure that amplification occurred in a linear range. After amplification, the products were electrophoresed in a 1% agarose gel and detected by ethidium bromide.

Mice deficient in the p52 subunit of NF-κB were found to be impaired in their Ab response to T-dependent Ags (9, 10). In particular, the mutant mice lacked Abs with switched isotypes, whereas the IgM response was higher than controls. Similarly, mice lacking Bcl-3 were defective in the generation of Abs with a switched isotype (IgG2a), in response to infection with flu (22). The ability to switch, however, was not seriously impaired, since challenge with high dose Ags in the presence of adjuvants allowed for the generation of switched isotypes in both types of mutant mice.

Further analysis of these mice by immunohistochemistry revealed defects that underlie the impaired Ab responses. Spleens were isolated 7 days after challenge with 100 μg of TNP-KLH Ag adsorbed to alum and cryosections prepared. p52 knockout mice were unable to generate FDC networks, as judged by the lack of cells expressing high levels of CD35, confirming our previous data (9) (Fig. 1,A, wild-type control; Fig. 1,B, p52 KO; CD35 red). Typical germinal centers were not observed either in the p52 knockouts as judged by PNA labeling (wild-type, Fig. 1,A; p52 knockout, Fig. 1 B; PNA blue). Instead, PNA+ clusters were detected in spleens of challenged p52 knockouts but were located exclusively in the T cell zone and were profoundly reduced in numbers.

FIGURE 1.

Splenic microarchitecture of wild-type (A, D, and G), p52-deficient (B, E, and H), and Bcl-3-deficient (C, F, and I) mice. Mice were immunized i.p. with 100 μg/ml TNP-KLH adsorbed to alum and sacrificed 7 days later. Splenic cryosections were labeled with both anti-CD35 (red) and PNA (blue in A and B; brown in C) or with FDC-M2 (brown; D, E, and F) or with anti-MAdCAM-1 (brown; G, H, and I). WP, white pulp; ac, central arteriole; MZ, marginal zone; GC, germinal center; RP, red pulp.

FIGURE 1.

Splenic microarchitecture of wild-type (A, D, and G), p52-deficient (B, E, and H), and Bcl-3-deficient (C, F, and I) mice. Mice were immunized i.p. with 100 μg/ml TNP-KLH adsorbed to alum and sacrificed 7 days later. Splenic cryosections were labeled with both anti-CD35 (red) and PNA (blue in A and B; brown in C) or with FDC-M2 (brown; D, E, and F) or with anti-MAdCAM-1 (brown; G, H, and I). WP, white pulp; ac, central arteriole; MZ, marginal zone; GC, germinal center; RP, red pulp.

Close modal

The absence of FDC networks in p52 knockout animals was also revealed via the FDC-M2 Ab (wild-type, Fig. 1,D; p52 knockouts, Fig. 1 E). FDC-M2 labels FDC networks in challenged wild-type animals and, in addition, weakly stains some reticular structures throughout the white pulp (25). We have previously shown lack of labeling with FDC-M1 in these mice (9), an Ab that recognizes a distinct Ag from FDC-M2 (M. H. Kosco-Vilbois, unpublished observation). Combined, these data using the two anti-FDC mAbs indicate an apparent complete lack of mature FDC networks in p52-deficient mice.

We further processed these sections for MAdCAM-1, an adhesion receptor present on sinus-lining cells in MZs and present on some FDCs (26, 27). MAdCAM-1 expression could not be detected in p52 knockout mice (Fig. 1,H), whereas it was readily observed in the MZ and in follicular areas of wild-type littermate controls (see Fig. 1,G). Therefore, together with the data presented above, no mature FDCs could be detected in p52 knockout mice. Beyond this, p52 knockout mice appear to have a defect associated with sinus-lining cells. Possibly related to the absence of MAdCAM-1 expression on the sinus-lining cells, a lack of deposition of the laminin β-2 chain was observed (Fig. 2,A, wild-type control; Fig. 2,C, p52 KO). This basement membrane component is usually detected near sinus-lining cells in wild-type mice (Fig. 2,D, MAdCAM-1; Fig. 2 E, MAdCAM-1 and laminin β-2 chain). In addition to possible defects in or of sinus-lining cells, there were other defects associated with splenic MZs. As demonstrated previously (9), p52 knockouts lack metallophilic macrophages that reside in the MZs of wild-type animals.

FIGURE 2.

Impaired expression of the basement membrane component laminin β-2 chain in the splenic MZ of p52-deficient and Bcl-3-deficient mice. Splenic sections of control littermates (A), Bcl-3 knockout (B), and p52 knockout mice (C) were labeled with anti-laminin β-2 Abs (red/pink). Sections of wild-type mice were also labeled for MAdCAM-1 (brown; D) or double-labeled for MAdCAM-1 and laminin β-2 (brown and red, respectively; E). See Fig. 1 for abbreviations used.

FIGURE 2.

Impaired expression of the basement membrane component laminin β-2 chain in the splenic MZ of p52-deficient and Bcl-3-deficient mice. Splenic sections of control littermates (A), Bcl-3 knockout (B), and p52 knockout mice (C) were labeled with anti-laminin β-2 Abs (red/pink). Sections of wild-type mice were also labeled for MAdCAM-1 (brown; D) or double-labeled for MAdCAM-1 and laminin β-2 (brown and red, respectively; E). See Fig. 1 for abbreviations used.

Close modal

Bcl-3-deficient mice exhibit defects similar to those of p52-deficient mice, except that these deficiencies are generally less severe. Challenged Bcl-3 knockout mice did form PNA+ centers in response to high-dose Ags plus adjuvants, even though the numbers of these centers appeared to be well below those seen in wild-type mice (see Ref. 22 and Fig. 1,C; PNA brown). Furthermore, the PNA+ B cell clusters of Bcl-3-deficient animals were located in B cell-enriched zones, although these zones did not have a normal follicular appearance (22). However, Bcl-3-deficient mice did not develop detectable FDC networks in response to a single antigenic stimulus, as judged by the lack of expression of FDC-M1 (22) or of high levels of CD35 (Fig. 1,C, CD35 red; compare with control in Fig. 1,A). Spleens from Bcl-3 knockouts did contain some FDC-M2-expressing cells, but there was no evidence for a network of such cells in the follicles (Fig. 1 F). Instead, FDC-M2 expression was dispersed throughout the white pulp and was highest around the periphery.

Bcl-3 knockouts also expressed some MAdCAM-1 in their MZs (Fig. 1,I), but only at very low levels compared with wild-type (Fig. 1,G). Similarly, laminin β-2 chain deposits near the sinus-lining cells of the MZ were significantly reduced in these mutant mice as well, although the reduction was not as severe as that seen in p52-deficient mice (Fig. 2 B).

These data, together with previously presented results, demonstrate that both p52 and Bcl-3 knockout animals do not form proper FDC networks under the conditions used here for stimulation. But whereas p52 knockout mice lack known markers for FDCs (CD35, FDC-M1, MAdCAM-1, and FDC-M2), Bcl-3 knockouts appear to contain some FDC-M2+ and MAdCAM-1+ cells, especially in the MZ areas. Therefore, it is conceivable that the Bcl-3 knockouts contain FDC precursor-like cells and that full differentiation of these cells into networks does not occur in response to a single dose of Ag. To investigate this more rigorously, the ability to retain Ag, a functional feature of FDCs, was tested. Mice were injected s.c. with 100 μg of the Ag HRP, on three separate occasions, and then sacrificed 24 h after the last injection. At this time, Ag is known to be bound to FDCs in wild-type mice, largely in the form of Ag-Ab complexes both on cell processes and iccosomes (28, 29). The presence of the HRP Ag was then detected with peroxidase anti-peroxidase (PAP), and sections were also labeled for CD35. Fig. 3 shows the presence of CD35 (A, red; C, brown) and HRP Ag (A and B, brown) in wild-type mice, whereas p52 knockout mice not only lack CD35 clusters (G and I) but also do not contain any deposits of Ag (G and H). Thus, p52 knockout mice lack cells capable of capturing Ag, and they also lack markers classically associated with FDCs. In contrast, Bcl-3 knockout mice have some deposits of Ag (D and E; brown) and, surprisingly, also contain some CD35+ cell clusters (D, red; F, brown), although much less so than the wild-type. The presence of CD35 clusters was unexpected in light of the total absence of such clusters in the previous experiments. However, in contrast to these earlier experiments, Ag has been administered three times, conditions that apparently allowed a weak but detectable development of an FDC network. Therefore, Bcl-3 knockouts are not absolutely blocked for the generation of an FDC network, whereas p52 knockouts appear to be totally impaired.

FIGURE 3.

Impaired trapping of immune complexes in p52-deficient and Bcl-3-deficient mice. Mice were immunized three times s.c. with HRP, and splenic cryosections were examined for the presence of HRP-Ab complexes. Sections from control littermates (A, B, and C), Bcl-3-deficient (D, E, and F), and p52-deficient mice (G, H, and I) were double labeled with anti-CD35 (red) and PAP (brown) (A, D, and G) or single stained with either PAP (B, E, and H) or anti-CD35 (C, F, and I). Sections B, C, E, F, H, and I were counterstained with hematoxylin and eosin (HE). T, T cell zone; see Fig. 1 for other abbreviations used.

FIGURE 3.

Impaired trapping of immune complexes in p52-deficient and Bcl-3-deficient mice. Mice were immunized three times s.c. with HRP, and splenic cryosections were examined for the presence of HRP-Ab complexes. Sections from control littermates (A, B, and C), Bcl-3-deficient (D, E, and F), and p52-deficient mice (G, H, and I) were double labeled with anti-CD35 (red) and PAP (brown) (A, D, and G) or single stained with either PAP (B, E, and H) or anti-CD35 (C, F, and I). Sections B, C, E, F, H, and I were counterstained with hematoxylin and eosin (HE). T, T cell zone; see Fig. 1 for other abbreviations used.

Close modal

To better define the identity of the cell responsible for the observed defects, p52 knockout mice were reconstituted with wild-type bone marrow. Six to eight weeks after adoptive transfer into either lethally or sublethally irradiated p52 knockout mice, the resulting chimeric mice were challenged with 100 μg TNP-KLH adsorbed to alum and were sacrificed 9 days later. IHC analyses of the recipient knockout mice still presented a lack of proper B cell follicles (Fig. 4,A; B cells red), whereas such follicles readily formed in recipient wild-type mice (Fig. 4,B). PNA+ B cell clusters still localized in the T cell zone in these recipient mutant mice, rather than in the B cell follicles like wild-type recipients (Fig. 4, A and B, respectively; PNA brown). The p52 knockout recipients also still lacked cell clusters highly positive for CD35 (Fig. 4,C; control in D), and cells positive for MAdCAM-1 (Fig. 4,E; control in F), and FDC-M1 (Fig. 4,G; control in H). Although FDC-M2+ cell clusters were not present either, some weak and widely scattered staining for FDC-M2 was seen throughout the white pulp (Fig. 4,I; control in J). The level of CD35 labeling in the MZs of p52 knockout recipients was quite strong (Fig. 4,C) (although still well below that of mature FDC networks), which may reflect an increased number of B cells that now accumulate in this location (Fig. 4 A). Together, these data clearly indicate that p52 knockout mice receiving wild-type bone marrow were still unable to form B cell follicles, FDC networks, or proper germinal centers.

FIGURE 4.

Splenic microarchitecture of p52-deficient mice adoptively transferred with wild-type bone marrow (BM). Six to seven weeks after reconstitution of irradiated p52-deficient (A, C, and E; and G, I, and K) or wild-type littermate control mice (B, D, and F; and H, J, and L) with wild-type BM, recipients were immunized i.p. with 100 μg/ml of TNP-KLH adsorbed to alum and sacrificed 9 days later. Splenic cryosections were double labeled with B220 (red) and PNA (brown) (A and B) or singly with anti-CD35 (brown; C and D) or with anti-MAdCAM-1 (brown; E and F) or FDC-M1 (brown; G and H), or FDC-M2 (brown; I and J) or MOMA-1 (brown; K and L). See Figs. 1 and 3 for abbreviations used.

FIGURE 4.

Splenic microarchitecture of p52-deficient mice adoptively transferred with wild-type bone marrow (BM). Six to seven weeks after reconstitution of irradiated p52-deficient (A, C, and E; and G, I, and K) or wild-type littermate control mice (B, D, and F; and H, J, and L) with wild-type BM, recipients were immunized i.p. with 100 μg/ml of TNP-KLH adsorbed to alum and sacrificed 9 days later. Splenic cryosections were double labeled with B220 (red) and PNA (brown) (A and B) or singly with anti-CD35 (brown; C and D) or with anti-MAdCAM-1 (brown; E and F) or FDC-M1 (brown; G and H), or FDC-M2 (brown; I and J) or MOMA-1 (brown; K and L). See Figs. 1 and 3 for abbreviations used.

Close modal

In contrast to FDC networks, MOMA-1+ metallophilic macrophages were restored to the MZs of p52 knockouts after transfer of wild-type bone marrow (Fig. 4,K; control in L). Therefore, the wild-type bone marrow cells corrected a specific defect of the original knockout mice. This unexpected result suggests that the absence of metallophilic macrophages in p52 KO mice is independent of the lack of FDCs or B cell follicles. In contrast to the restored metallophilic macrophages in the MZs, MAdCAM-1 expression was not restored to sinus-lining cells (neither was it restored to FDCs; see above) (Fig. 4 E; control in F), indicating that not all defects in the MZ can be corrected.

The inability of naive B cells to form follicles into which Ag-activated PNA+ B cells can migrate could be due to a defect intrinsic to B cells in receiving appropriate signals, or to a defect in the generation of such signals by other cells. Since mature p52-deficient B cells migrate properly into B cell follicles when placed into recombination-activating gene (RAG)-deficient animals, it is more likely that p52 knockout mice lack the ability to send a proper signal to B cells. BLC is a recently described chemokine that may be important in directing B cells to migrate into the follicles of the spleen (30, 31). PCR analysis of reverse transcribed RNA extracted from spleens 48 h after administration of SRBC demonstrated an apparent lack of BLC transcripts in p52 knockout mice, whereas such transcripts were readily detected in the wild-type controls (Fig. 5,A). However, when analyzed by PCR on day 7 after Ag stimulation, some BLC transcripts could be detected in the p52 knockout animals (Fig. 5,B). Therefore, the expression of BLC in p52 knockout mice appears to be partially disrupted, which may contribute to the observed lack of B cell follicles and FDC networks. Impaired expression of BLC in p52 knockout mice is also consistent with an impaired generation of FDC networks, given that FDCs/stromal cells have been suggested to express BLC (30, 31, 32). In support of this hypothesis, we detected BLC by PCR in cell preparations from wild-type mice enriched for FDCs (Fig. 5,C; left). Although the FDC preparations contain significant numbers of B cells (see B220 transcripts in Fig. 5,C; left), B cells were not the source of BLC, since purified B cells lacked detectable levels of transcripts for this chemokine (Fig. 5 C; right). Therefore, FDCs are the most likely source of BLC transcripts in wild-type mice, but this remains to be demonstrated directly.

FIGURE 5.

Impaired expression of BLC transcripts in p52-deficient mice. A, RT-PCR analyses of splenocytes obtained from wild-type control, p52-deficient, and Bcl-3-deficient mice 48 h after Ag stimulation. B, RT-PCR analyses of splenocytes from wild-type control and p52-deficient mice 7 days after Ag stimulation. Successive dilutions of input cDNAs are shown, along with control PCRs for GAPDH. C, Expression of B220 and of BLC in an FDC preparation (left panel) and a B cell preparation (right panel) from a wild-type mouse. Successive dilutions of input cDNAs are shown.

FIGURE 5.

Impaired expression of BLC transcripts in p52-deficient mice. A, RT-PCR analyses of splenocytes obtained from wild-type control, p52-deficient, and Bcl-3-deficient mice 48 h after Ag stimulation. B, RT-PCR analyses of splenocytes from wild-type control and p52-deficient mice 7 days after Ag stimulation. Successive dilutions of input cDNAs are shown, along with control PCRs for GAPDH. C, Expression of B220 and of BLC in an FDC preparation (left panel) and a B cell preparation (right panel) from a wild-type mouse. Successive dilutions of input cDNAs are shown.

Close modal

Given the critical importance of p52 and, to a lesser degree, of Bcl-3 for formation of FDC networks, we investigated the expression of these proteins in splenic sections (Fig. 6). Bcl-3 (Fig. 6,A) and p52 (Fig. 6,C) have a high level of expression in the follicles/FDC networks subsequent to antigenic challenge. To control for the specificity of the Ab, the anti-p52 peptide Ab was competed for with the reactive peptide, and in this case no specific staining was observed (Fig. 6,D). In the case of the polyclonal anti-Bcl-3 Ab, the specificity was demonstrated by its lack of labeling in Bcl-3 knockout animals (Fig. 6 B). To confirm the localization, we have stained sections for both CD35 and p52 as well (data not shown).

FIGURE 6.

Immunohistochemical detection of Bcl-3 and p52 proteins in splenic cryosections of wild-type mice. Detection with anti-Bcl-3 (brown; A) and anti-p52 (brown; C) Abs reveals the highest levels of expression in follicular areas of spleens after SRBC stimulation. No labeling is seen with anti-Bcl-3 Abs in sections from SRBC-stimulated Bcl-3 knockout mice (B) nor with anti-p52 Abs in sections from SRBC-stimulated wild-type mice in the presence of competing amounts of the p52 peptide to which the Ab was raised (D).

FIGURE 6.

Immunohistochemical detection of Bcl-3 and p52 proteins in splenic cryosections of wild-type mice. Detection with anti-Bcl-3 (brown; A) and anti-p52 (brown; C) Abs reveals the highest levels of expression in follicular areas of spleens after SRBC stimulation. No labeling is seen with anti-Bcl-3 Abs in sections from SRBC-stimulated Bcl-3 knockout mice (B) nor with anti-p52 Abs in sections from SRBC-stimulated wild-type mice in the presence of competing amounts of the p52 peptide to which the Ab was raised (D).

Close modal

The present manuscript delineates causes underlying impaired production of Abs to T-dependent Ags in p52- and Bcl-3-deficient mice. In both mutant mouse strains, FDC networks do not develop in response to a single antigenic stimulation, even if given in the presence of adjuvants. The lack of such networks appears to be related to the inability to form proper B cell follicles and thus proper germinal center reactions. The data further suggest that p52 and Bcl-3 function within a common path that is required to form the FDC network and that FDCs themselves may be intrinsically impaired if these proteins are missing. In addition, sinus-lining cells in the MZ may be defective, although nothing is known about whether this could be related to impaired FDCs or not. On the other hand, metallophilic macrophages are defective in a way unrelated to the lack of FDCs, and these cells may be intrinsically impaired. Thus, at least two, and possibly three, cell types may require p52 and Bcl-3 for proper differentiation. The contributions of these two proteins can be distinguished, however, because the defects appear to be absolute in p52-deficient animals, whereas the related defects in Bcl-3-deficient mice are only partial. For example, repeated antigenic stimulation eventually leads to the development of some FDC networks in Bcl-3 knockouts, but never in p52 knockouts. Bcl-3 may play a modulating function, whereas p52 appears absolutely necessary.

Despite the inability to form any FDC networks or B cell follicles, p52 knockout mice nevertheless generate switched Abs to T dependent Ags, but only in the presence of adjuvants, as noted previously (9). Most likely, this is due to small PNA+ B cell clusters that develop in the T cell zones under these conditions. The activated B cells within these clusters may receive sufficient local stimulation to partially overcome the lack of FDC networks; as a result they appear able to undergo some affinity maturation and switching. High dose Ag plus adjuvant may provide sufficient levels of Ag and locally produced cytokines to bypass the need for a proper germinal center. Less potent and presumably more physiologic antigenic stimulation, however, may require the formation of proper germinal centers in association with an FDC network, a condition met only in wild-type mice, but not in p52 knockouts. Therefore, one function of FDCs may be to lower the threshold of stimulation required for effective Ab responses.

The PNA+ B cells that form clusters in the T cell zones of p52 knockout mice in response to Ag plus adjuvant may not be able to migrate properly for the same reason that B cells do not form follicles in these mice. The BLC chemokine is thought to provide an important signal to localize B cells into follicles, and transcripts for this chemokine cannot be detected in spleens of p52 knockout mice early after antigenic stimulation, although they are detectable at later times, although at apparently lower levels. In contrast, wild-type animals contain easily detectable levels at early and later times after antigenic stimulation, whereas Bcl-3 knockout mice exhibit partially reduced levels at early times (our unpublished observations). It has been suggested that BLC is produced by FDCs or underlying stromal cells within the B cell areas (32), and our enriched populations of FDCs do express this chemokine. Therefore, the impaired expression of BLC in p52 knockouts is consistent with the absence of FDC networks in these cells.

An important question concerns the identity of the cells responsible for the lack of an FDC network. Previously we had ruled out mutant mature B or T cells as sources for this defect. Adoptive transfers of bone marrow from mice lacking p52 or Bcl-3 resulted in near normal Ab responses in recombination-activating gene-deficient recipients, as well as normal germinal centers (9) and fully developed FDC networks (L. Poljak and U. Siebenlist, unpublished observations). It was important to rule out intrinsic defects in B cells, since FDC networks have previously been shown to depend on the production of lymphotoxin-α by B cells (33). We demonstrate now that transfer of wild-type bone marrow into p52-deficient animals does not correct the FDC defects. The recipient mice are still unable to form FDC networks. Therefore, the source of the defects leading to lack of FDCs cannot be readily transferable hemopoietic cells but instead appear to be non-hemopoietic cells. It remains to be demonstrated whether FDCs are intrinsically impaired or whether the defects reside in stromal cells distinct from FDCs that are necessary for formation of FDC networks. Consistent with intrinsic defects in FDCs/stromal cells, IHC analyses of sections (Fig. 6) and of partially purified FDC preparations (L. Poljak and U. Siebenlist, unpublished results) revealed high levels of expression of both p52 and Bcl-3 proteins in FDC-like cells. Indeed, p52 levels had previously been reported to be high in FDCs (34).

In contrast to the continued inability to form FDC networks in transfers of wild-type bone marrow into p52-deficient mice, this procedure did result in the appearance of metallophilic macrophages in the MZ. Therefore, lack of metallophilic macrophages in the original mutant strain must be due to a defect in a transferable, presumably hemopoietic cell type, possibly the metallophilic macrophages themselves. Interestingly, Bcl-3 knockouts had a significant, but only partial reduction of metallophilic macrophages, suggesting that Bcl-3 may again play a modulating role in this instance as well (22). The data imply that p52 has critical functions in at least two distinct cell types, possibly in FDCs and metallophilic macrophages, and that these functions may be enhanced by Bcl-3.

Unlike metallophilic markers, the expression of the MECA-367 epitope of the addressin MAdCAM-1 on sinus-lining cells in the MZ was not restored by transfer of wild-type bone marrow into p52 knockouts. MAdCAM-1 is reportedly expressed on FDCs and on sinus-lining cells in wild-type mice (26, 27), and its expression in Peyer’s patches has been reported to be necessary for extravasation of blood lymphocytes into these lymphoid organs (35). Adoptive transfers do not restore expression of the MECA-367 epitope to either FDCs or sinus-lining cells. The defect associated with sinus-lining cells may or may not be intrinsic to these cells. In this regard, it could be relevant that laminin β-2 chain deposits near sinus-lining cells are significantly reduced in the p52 knockouts. Therefore, the basement membrane may be disrupted, and this in turn may affect other cells, including FDCs, that receive critical signals from this matrix via their integrin receptors (36, 37, 38). In fact, signaling via α21 integrin receptors has been reported to induce protein translation of preformed Bcl-3 mRNAs in platelets, although it is not known whether such a control mechanism also operates in other cells (39). It remains to be investigated whether lack of expression of MAdCAM-1 on sinus-lining cells is related to the defect in laminin β-2 chain deposition and/or the defect in FDC network formation.

Disruptions of the microarchitecture of secondary lymphoid organs have been observed in several other mutant mice. In particular, mice deficient in TNF-α, lymphotoxin (LT)α, LTβ, LTβR, or TNFRI all lack FDC networks (40, 41, 42, 43). In addition, some of these mice share other phenotypes with the p52 knockout mice, although all of the mutants present with unique disruptions as well. Nevertheless, an interdependent network of these TNF/LT ligands, their receptors, and NF-κB is suggested (p50/p52 double knockout mice have very severe disruptions in lymphoid architecture). Although it is possible that loss of p52 could impair the expression of these particular mediators and thus cause the observed phenotypes, PCR analyses have not revealed major changes in the mRNAs for these proteins in total splenocytes from our knockouts (L. Poljak and U. Siebenlist, unpublished observations). Furthermore, adoptive transfers of wild-type hemopoietic precursors into the knockout mice did not cure the majority of the defects, arguing against intrinsic defects in hematopoietic cells. Thus activation of NF-κB complexes downstream of the activities of the TNF and lymphotoxin ligands may be critical to lymphoid architecture, with the p52 protein in particular playing an essential role for B cell follicle formation, FDC network development, and a proper MZ.

We thank R. Dreyfuss and S. Everett for photography, and A. S. Fauci for continued support and encouragement.

2

Abbreviations used in this paper: FDC, follicular dendritic cell; BLC, B lymphocyte chemoattractant; TNP-KLH, 2,4,6-trinitrophenyl-keyhole limpet hemocyanin; PNA, peanut agglutinin; MZ, marginal zone; IHC, immunohistochemistry; PAP, peroxidase anti-peroxidase; LT, lymphotoxin.

1
Siebenlist, U., G. Franzoso, K. Brown.
1994
. Structure, regulation and function of NF-κB.
Annu. Rev. Cell. Biol.
10
:
405
2
Baldwin, A. S..
1996
. The NF-κB and IκB proteins: new discoveries and insights.
Annu. Rev. Immunol.
14
:
649
3
Ghosh, S., M. J. May, E. B. Kopp.
1998
. NF-κB and Rel proteins: evolutionarily conserved mediators of immune responses.
Annu. Rev. Immunol.
16
:
225
4
Gerondakis, S., R. Grumont, I. Rourke, M. Grossmann.
1998
. The regulation and roles of Rel/NF-κB transcription factors during lymphocyte activation.
Curr. Opin. Immunol.
10
:
353
5
Sha, W. C., H. C. Liou, E. I. Tuomanen, D. Baltimore.
1995
. Targeted disruption of the p50 subunit of NF-κB leads to multifocal defects in immune responses.
Cell
80
:
321
6
Kontgen, F., R. J. Grumont, A. Strasser, D. Metcalf, R. Li, D. Tarlinton, S. Gerondakis.
1995
. Mice lacking the c-rel proto-oncogene exhibit defects in lymphocyte proliferation, humoral immunity, and interleukin-2 expression.
Genes Dev.
9
:
1965
7
Weih, F., D. Carrasco, S. K. Durham, D. S. Barton, C. A. Rizzo, R. P. Ryseck, S. A. Lira, R. Bravo.
1995
. Multiorgan inflammation and hematopoietic abnormalities in mice with targeted disruption of RelB, a member of the NF-κB/Rel family.
Cell
80
:
331
8
Doi, T. S., T. Takahashi, O. Taguchi, T. Azuma, Y. Obata.
1997
. NF-κB RelA-deficient lymphocytes: normal development of T cells and B cells, impaired production of IgA and IgG1 and reduced proliferative responses.
J. Exp. Med.
185
:
953
9
Fransozo, G., L. Carlson, L. Poljak, E. W. Shores, S. Epstein, A. Leonardi, A. Grinberg, T. Tran, T. Sharton-Kersten, M. Anver, P. Love, K. Brown, U. Siebenlist.
1998
. Mice deficient in nuclear factor (NF)-κB/p52 present with defects in humoral responses, germinal center reactions and splenic microarchitecture.
J. Exp. Med.
187
:
147
10
Caamano, J. H., C. A. Rizzo, S. K. Durham, D. S. Barton, C. Raventos-Suarez, C. M. Snaper, R. Bravo.
1998
. Nuclear factor (NF)-κB2 (p100/p52) is required for normal splenic microarchitecture and B cell-mediated immune responses.
J. Exp. Med.
187
:
185
11
Beg, A. A., W. C. Sha, R. T. Bronson, S. Ghosh, D. Baltimore.
1995
. Embryonic lethality and liver degeneration in mice lacking the RelA component of NF-κB.
Nature
376
:
167
12
Burkly, L., C. Hession, L. Ogata, C. Reilly, L. A. Marconi, D. Olsen, R. Tizard, R. Cate, D. Lo.
1995
. Expression of relB is required for development of thymic medulla and dendritic cells.
Nature
373
:
531
13
Belwin, M. P., K. V. Anderson.
1996
. A conserved signaling pathway: the Drosophila Toll-dorsal pathway.
Annu. Rev. Cell. Dev. Biol.
12
:
393
14
Brown, K., S. Park, T. Kanno, G. Fransozo, U. Siebenlist.
1993
. Mutual regulation of the transcriptional coactivator NF-κB and its inhibitor IκBα.
Proc. Natl. Acad. Sci. USA
90
:
2532
15
Brown, K., S. Gerstberger, L. Carlson, G. Franzoso, U. Siebenlist.
1995
. Control of IκB proteolysis by site-specific signal-induced phosphorylation.
Science
267
:
1485
16
May, M. J., S. Ghosh.
1997
. Rel/NF-κB and IκB proteins: an overview.
Semin. Cancer Biol.
8
:
63
17
Whiteside, S. T., A. Israel.
1997
. IκB proteins: structure, function and regulation.
Semin. Cancer Biol.
2
:
75
18
Tarlinton, D..
1998
. Germinal centers: form and function.
Curr. Opin. Immunol.
10
:
245
19
Huguet, C., P. Crepieux, V. Laudet.
1997
. Rel/NF-κB transcription factors and IκB inhibitors: evolution from unique common ancestor.
Oncogene
15
:
2965
20
Bours, V., G. Franzoso, V. Azarenko, S. Park, T. Kanno, K. Brown, U. Siebenlist.
1993
. The oncoprotein Bcl-3 directly transactivates through κB motifs via association with DNA-binding p50B homodimers.
Cell
72
:
729
21
Fujita, T., G. P. Nolan, H. C. Liou, M. L. Scott, D. Baltimore.
1993
. The candidate proto-oncogene bcl-3 encodes a transcriptional coactivator that activates through NF-κB p50 homodimers.
Genes Dev.
7
:
1354
22
Fransozo, G., L. Carlson, T. Sharton-Kersten, E. W. Shores, S. Epstein, A. Grinber, T. Tran, E. Shacter, A. Leonardi, P. Love, A. Sher, U. Siebenlist.
1997
. Critical roles for the bcl-3 oncoprotein in T-cell mediated immunity, splenic microarchitecture and germinal center formation.
Immunity
6
:
479
23
Schwartz, E. M., P. Krimpenfort, A. Berns, I. M. Verma.
1997
. Immunological defects in mice with targeted disruption in Bcl-3.
Genes Dev.
11
:
187
24
Kraal, G..
1992
. Cells in the marginal zone of the spleen.
Int. Rev. Cytol.
132
:
31
25
Kosco-Vilbois, M. H., H. Zentgraf, J. Gerdes, J. Y. Bonnefoy.
1997
. To “B” or not to “B” a germinal center?.
Immunol. Today
5
:
225
26
Kraal, G., K. Schornagel, P. R. Streeter, B. Holzmann, E. C. Butcher.
1995
. Expression of mucosal vascular addressin, MAdCAM-1 on sinus-lining cells in the spleen.
Am. J. Pathol.
147
:
763
27
Szabo, M. C., E. Butcher, L. M. McEvoy.
1997
. Specialization of mucosal follicular dendritic cells revealed by mucosal addressin cell adhesion molecule-1 display.
J. Immunol.
158
:
5584
28
Chen, L. L., A. M. Frank, J. C. Adams, R. M. Steinman.
1978
. Distribution of horseradish peroxidase (HRP)-anti-HRP immune complexes in mouse spleen with special reference to follicular dendritic cells.
J. Exp. Med.
79
:
184
29
Szakal, A. K., M. H. Kosco, J. G. Tew.
1988
. A novel in vivo follicular dendritic cell-dependent icosome-mediated mechanism for delivery of antigen to antigen-processing cells.
J. Immunol.
140
:
341
30
Gunn, M. D., V. N. Ngo, K. M. Ansel, E. H. Ekland, J. G. Cyster, L. T. Williams.
1998
. A B-cell homing chemokine made in lymphoid follicles activates Burkitt’s lymphoma receptor-1.
Nature
391
:
799
31
Legler, D. F., M. Loetscher, R. S. Roos, I. Clark-Lewis, M. Baggiolini, B. Moser.
1998
. B cell-attracting chemokine 1, a human CXC chemokine expressed in lymphoid tissues, selectively attracts B lymphocytes via BLR1/CXCR5.
J. Exp .Med.
187
:
655
32
Ngo, V. N., H. Korner, M. D. Gunn, K. N. Schmidt, D. S. Riminton, M. D. Cooper, J. L. Browning, J. D. Sedgwick, J. G. Cyster.
1999
. Lymphotoxin α and tumor necrosis factor are required for stromal cell expression of homing chemokines in B and T cell areas of the spleen.
J. Exp. Med.
189
:
403
33
Fu, Y. X., G. Huang, Y. Wang, D. D. Chaplin.
1998
. B lymphocytes induce formation of follicular dendritic cell clusters in lymphotoxin α dependent fashion.
J. Exp. Med.
187
:
1009
34
Feuillard, J., M. Korner, A. Israel, J. Vassy, M. Raphael.
1996
. Differential nuclear localization of p50, p52 and RelB proteins in human accessory cells of the immune response in situ.
Eur. J. Immunol.
26
:
2547
35
Berg, E. L., L. M. McEvoy, C. Berlin, R. F. Bargatze, E. C. Butcher.
1993
. L-selectin mediated lymphocyte rolling on MAdCAM-1.
Nature
366
:
695
36
Gullberg, D., P. Ekblom.
1995
. Extracellular matrix and its receptors during development.
Int. J. Dev. Biol.
39
:
845
37
Ogata, T., M. Yamakawa, Y. Imai, T. Takahashi.
1996
. Follicular dendritic cells adhere to fibronectin and laminin fibres via their respective receptors.
Blood
88
:
2995
38
Shyy, J. Y., S. Chien.
1997
. Role of integrins in cellular response to mechanical stress and adhesion.
Curr. Opin. Cell. Biol.
9
:
707
39
Pabla, R., A. S. Weyrich, D. A. Dixon, P. F. Bray, T. M. McIntyre, S. M. Prescott, G. A. Zimmerman.
1999
. Integrin-dependent control of translation engagement of integrin αIIβ3 regulates synthesis of proteins in activated human platelets.
J. Cell. Biol.
144
:
175
40
Futterer, A., K. Mink, A. Luz, M. H. Kosco-Vilbois, K. Pfeffer.
1998
. The lymphotoxin β receptor controls organogenesis and affinity maturation in peripheral lymphoid tissues.
Immunity
9
:
59
41
Matsumoto, M., Y.-X. Fu, H. Molina, G. Huang, J. Kim, D. A. Thomas, M. H. Nahm, D. D. Chaplin.
1997
. Distinct roles of lymphotoxin α and the type I tumor necrosis factor (TNF) receptor in establishment of follicular dendritic cells from non-bone marrow-derived cells.
J. Exp. Med.
186
:
1997
42
Matsumoto, M., Y.-X. Fu, H. Molina, D. Chaplin.
1997
. Lymphotoxin-α-deficient and TNF receptor-I-deficient mice define developmental and functional characteristics of germinal centers.
Immunol. Rev.
156
:
137
43
Koni, P. A., R. Sacca, P. Lawton, J. L. Browning, N. H. Ruddle, R. A. Flavell.
1997
. Distinct roles in lymphoid organogenesis for lymphotoxin α and β revealed in lymphotoxin β-deficient mice.
Immunity
6
:
491