Bone marrow stromal cell Ag-1 (BST-1; CD157)-deficient mice were generated to examine the immunologic roles of the molecule in vivo. In BST-1−/− mice, the development of peritoneal B-1 cells was delayed, and CD38low/− B-lineage cells were increased in the bone marrow and spleen. Partial impairment of thymus-independent (TI-2) and thymus-dependent (TD) Ag-specific immune responses was noted in the systemic and mucosal compartments of BST-1−/− mice, respectively. Although serum Ig levels as well as TD and TI-1 Ag-specific systemic immune responses were normal, the TI-2 Ag-induced IgG3 response was selectively impaired. Oral immunization of BST-1−/− mice with cholera toxin, a potent TD Ag for the induction of IgA response, resulted in the poor production of Ag-specific Abs at the intestinal mucosa accompanied by the reduced number of Ag-specific IgA-producing cells in the lamina propria. These results indicate that BST-1 has roles in B cell development and Ab production in vivo.
We identified human bone marrow stromal cell Ag-1 (BST-1)3 as a surface molecule that was expressed highly on the bone marrow stromal cell lines and synovial cell lines, from patients with rheumatoid arthritis, and as a molecule exhibiting the ability to support the growth of a pre-B cell line (1). In the sixth Human Leukocyte Differentiation Antigens workshop, a myeloid cell Ag Mo5 (2) was found to be identical with BST-1, and both were designated CD157 (3). BST-1 is a glycosyl-phosphatidylinositol-anchored membrane protein, having about 30% homology with Aplysia ADP-ribosyl cyclase or CD38 at the level of amino acid. BST-1 and CD38 actually have both ADP-ribosyl cyclase and cyclic ADP-ribose hydrolase activities in vitro (1, 4, 5, 6, 7). Furthermore, human genes for BST-1 and CD38 are closely located in chromosome 4p15 (8, 9), and murine genes are located in chromosome 5 (8, 10). Genomic structure analysis reveals the striking structural similarity between BST-1 and CD38, indicating that BST-1 and CD38 are evolved by gene duplication from an ancestral gene (8, 11, 12). These molecules are in a novel family of ectoenzyme, but the functions in vivo are largely unknown.
Our group as well as Cooper’s colleagues independently cloned the cDNA for murine BST-1 and BP-3, respectively, and found that both are identical molecules (4, 5). In the hemopoietic system and lymphoid tissues, the surface expression of murine BST-1/BP-3 is detected on T or B cell progenitors, immature B cells, mature myeloid cells, and reticular cells in the splenic white pulps, lymph nodes, and Peyer’s patches (PP) (13, 14, 15, 16). The expression of murine BST-1 in fetus, is maximally detected on thymocytes and liver B progenitors on days 15 to 16 and 17 of gestation, respectively, and gradually decreased until birth. In adult mice, the surface expression of BST-1 on both B and T cell progenitors begins around the stage when gene rearrangement of the Ag receptors is initiated (15), suggesting its role in lymphopoiesis. Actually, an anti-BST-1/BP-3 mAb, IF-7, has synergistic effects on anti-CD3-induced proliferation of pre-T cells and promotes the generation of TCRαβ+ cells in fetal thymic organ culture (16). This may indicate that BST-1 functions as a receptor. Along this line, we previously demonstrated that cross-linking of BST-1 induces tyrosine phosphorylation (17). These studies suggested that BST-1 may affect B and T cell development.
Although the expression of BST-1 is down-regulated on mature B cells in the secondary lymphoid tissues, it should be noted that these B cells are surrounded by the reticular cells expressing BST-1. This suggests that BST-1 on the reticular cells may play an important role in supporting the growth, differentiation, and survival of mature B cells.
It is also interesting to note that BST-1 is highly expressed in gut (5, 14). However, the role of BST-1 in the intestinal immune system is unknown. The mucosal immune system has been shown to possess a unique immunoregulatory mechanism that distinctively differs from the systemic immune compartment for the induction of IgA Ab responses (18, 19). For example, immunity to Ags administered orally is achieved by the stimulation of Th cells and IgA committed sIgA+ B cells in the gut-associated lymphoreticular tissues. It has been shown that oral immunization of mice with tetanus toxoid and cholera toxin (CT) resulted in the generation of Ag-specific Th2-type CD4+ T cells for the induction of IgA-producing cells (20, 21). These Ag-triggered Th2 cells and sIgA+ B cells in gut-associated lymphoreticular tissues then migrate to distant mucosal effector sites, such as the lamina propria (LP) region of the intestine, where terminal differentiation of the sIgA+ B cell into IgA-producing plasma cells takes place under the influence of Th2 cytokines, including IL-5 and IL-6. Since BST-1 is expressed by the brush border of the intestinal epithelial cells and the reticular cells in PP (14), it is interesting to postulate that BST-1 may be involved in the Ag-specific mucosal immune response.
To evaluate the roles of BST-1 in vivo, our initial effort was aimed at the establishment of BST-1 knockout mice using homologous recombination. The lymphocyte development of BST-1−/− mice was almost normal, except for the delayed appearance of peritoneal B-1 B cells and the increase in CD38low/− B-lineage cells. Among humoral immune responses in the systemic compartment, the thymus-independent (TI-2) Ag-induced IgG3, but not the IgM, response was selectively impaired. Furthermore, oral immunization with CT resulted in low production of Ag-specific IgA and IgG Abs in fecal extracts due to the reduced number of Ag-specific Ab-producing cells in the intestinal LP. These findings demonstrated that BST-1 is an important molecule for the selected population of B cell responses associated with TI-2 Ag-induced IgG3 and thymus-dependent (TD) Ag-induced IgA in the systemic and mucosal compartments, respectively.
Materials and Methods
Generation of BST-1 knockout mice
Mouse genomic DNA clones were isolated by screening a genomic DNA library derived from the 129/SV mouse strain (Stratagene, San Diego, CA). The targeting vector was prepared by replacing the KpnI-SalI fragment containing exons with pMC1neo poly(A) in the opposite transcriptional orientation. The ClaI-linearized targeting vector was electroporated into the E14.1 embryonic stem (ES) cell line. G418- and ganciclovir-resistant clones were selected and screened by PCR with a primer corresponding to the 5′ external sequence of the targeting vector (5′-TCCCTTTGACAAGGAAGCCCACTGAGTAGC-3′) and a primer complementary to the neo gene sequence (5′-GAGGGGATCGGCAATAAAAAGACAGAATAAAAC-3′). Homologous recombination was also confirmed by Southern blotting. Genomic DNA from individually selected clones were digested with HindIII and hybridized with the 5′-flanking probe. A DNA restriction fragment of 6.4 kb corresponded to the wild allele, while a 4.8-kb fragment resulted from the targeted allele. Five clones containing both alleles were identified from 453 colonies resistant to G418 and ganciclovir. Targeted ES clones were then injected into C57BL/6 (B6) blastocysts. Chimeric offspring generated from one clone transmitted the mutation to their progeny and were mated to B6 females. Mice carrying the mutation in the heterozygous state (BST-1+/−) were intercrossed to produce homozygous mutants (BST-1−/−).
Flow cytometric analysis
Details of the method for flow cytometric analysis and most of the mAbs used in this study were described previously (15). Briefly, single-cell suspensions from spleen, thymus, bone marrow, and peritoneal cavity were prepared and washed with ice-cold PBS without Ca2+ and Mg2+ containing 2% FCS and 0.1% sodium azide (washing buffer). Cells were blocked with a mixture of culture supernatant of 2.4G2 (anti-Fcγ receptor II/III) and 6% rat serum for 10 min, and then stained with respective mAbs directly labeled with FITC, phycoerythrin (PE), or biotin for 15 min on ice. Red 670-conjugated streptavidin (Life Technologies, Grand Island, NY) was used to reveal biotin-coupled Abs. After the staining, samples were analyzed with FACSort using the CellQuest program (Becton Dickinson, Mountain View, CA). Abs used for the staining are as follows: PE-conjugated RA3-6B2 (anti-CD45R); FITC-, PE- and biotin-conjugated AM/3 (anti-μ); PE-11-26c-1 (anti-δ); FITC-S7 (anti-CD43); PE-53-7.3 (anti-CD5); FITC-90 (anti-CD38, PharMingen, San Diego, CA); PE-RM4-4 (anti-CD4, PharMingen); FITC-53-5.8 (anti-CD8β, PharMingen); and FITC- and biotin-M1/70 (anti-Mac-1, CD11b).
For the TD immune response, mice (2–3 mo old) were immunized i.p. with 100 μg of DNP-KLH (LSL, Tokyo, Japan) in CFA on day 0 and were boosted with 100 μg of DNP-KLH in IFA on day 21. For TI immune responses, mice were immunized i.p. with 50 μg of 2,4,6-trinitrophenyl-conjugated LPS (TNP-LPS) or 20 μg of TNP-Ficoll in saline. Mice were bled before and after immunization. TNP-Ficoll was donated by Dr. S. Ono (Osaka University, Osaka, Japan). For oral immunization, mice were deprived of food for 2 h and given 0.5 ml of a solution containing eight parts HBSS and two parts 7.5% sodium bicarbonate by gastric intubation to neutralize stomach acidity for 30 min. Then, the mice were immunized orally with 0.25 ml of PBS containing CT (25 μg/mouse; Sigma, St. Louis, MO) by gastric intubation. CT was administered to mice once per week for 3 consecutive wk.
Fecal extracts or sera were collected for the analysis of Ag-specific and total Abs. The ELISA plates were coated with DNP-BSA, TNP-BSA, or CT for the Ag-specific Ab. To measure the isotypes and levels of Abs, plates were coated with isotype-specific rabbit or goat anti-mouse Abs. Diluted serum and secretion samples were incubated in the plates and revealed by alkaline phosphatase-labeled secondary Ab. The Abs were purchased from Zymed Laboratories (South San Francisco, CA), except for the unlabeled or alkaline phosphatase-labeled anti-IgG3 Abs, which were purchased from Southern Biotechnology Associates (Birmingham, AL).
Mononuclear cells of spleen, PP, MLN, and LP of the small intestine were isolated. Single cells were suspended in RPMI 1640 containing 10% FCS. An ELISPOT assay was used to detect cells producing CT-specific IgM, IgG, and IgA Ab (22). Ninety-six-well nitrocellulose-based plate (Millititer HA, Millipore, Bedfold, MA) were coated with 2 μg/ml of CT (100 μl/well). Mononuclear cells (1 × 103 to 1 × 105/well) were added to each well and then incubated at 37°C in air with 10% CO2 and 90% humidity for 4 h. After washing the plate, 1 μg/ml of horseradish peroxidase-conjugated goat anti-mouse μ-, γ-, and α-specific Abs (Southern Biotechnology Associates) were added to individual wells. Following overnight incubation at 4°C, the plate was reacted with 3-amino-9-ethylcarbazole for the development of isotype-specific spots.
B cell proliferation
Splenic B cells were enriched by eliminating T cells, macrophages, and NK cells using a Minimacs column (Miltenyi Biotec, Auburn, CA) following the manufacturer’s instruction. At first, T cells, macrophages, and NK cells were coated with the following rat mAbs: culture supernatant of RL172.4 (anti-CD4), GK1.5 (anti-CD4), 3.155 (anti-CD8), 53.6 (anti-CD8), and M1/70 (anti-CD11b) and purified 29B (anti-CD3, Life Technologies) and TMβ1 (anti-IL-2R β-chain, provided by Dr. T. Tanaka, Osaka University). The cells coated with rat mAbs were then incubated with colloidal superparamagnetic beads conjugated with goat anti-rat IgG and depleted by passing through a Minimacs column. The eluted cells were enriched with B cells (90% purity) and used in the proliferation assay. B cells were cultured in 96-well flat-bottom microtiter plates at 2 × 105/0.2 ml of RPMI 1640 medium supplemented with 10% FCS, 2-ME (50 μM), penicillin (50 μg/ml), and streptomycin (50 μg/ml). These cells were stimulated with LPS (Escherichia coli 0111:B4, Sigma), CS/2 (anti-CD38), 3/23 (anti-CD40, PharMingen), F(ab′)2 goat anti-mouse IgM polyclonal Ab (Jackson ImmunoResearch Laboratory, West Grove, PA), and Con A (Pharmacia, Uppsala, Sweden). Cultured cells were pulsed for the last 6 h of a 48-h culture period with 0.5 μCi/well of 3H-labeled thymidine, followed by scintillation counting.
To make polyclonal anti-mouse BST-1 Ab, the chimeric molecule of mouse BST-1 and the human IgG1Fc portion (mBST-1Fc) was prepared as described previously (15). Rabbit antiserum was then obtained by immunization with mBST-1Fc. Abs to human IgG1Fc were depleted by passing through the column of human IgG Sepharose 4B (Pharmacia), and cross-reactive Abs to mouse CD38 (mCD38) were removed by the absorption with mCD38 transfectant BAFmCD38. The specificity of the antiserum was confirmed by Western blotting using human IgG and cell lysate of BAFmBST-1 and BAFmCD38 (23).
Frozen sections (6 μm) of spleens were air-dried and subsequently fixed in cold acetone. The sections were blocked with 20% goat serum, then incubated with the anti-mouse BST-1 serum diluted 1/200. After washing, the sections were treated with biotinylated goat anti-rabbit Ig, washed with 0.05% Triton X-100 in PBS, then visualized with FITC-avidin.
Generation of BST-1-deficient mice
To obtain genomic BST-1 DNA, we screened a genomic 129/SV library using mouse BST-1 cDNA as a probe. Positive clones were mapped, and a targeting vector was constructed as shown in Figure 1,A. In brief, the sequences from the KpnI site in exon 1 up to the SalI site in exon 3 were deleted and replaced by neomycin resistance gene. The herpes simplex virus-thymidine kinase gene was placed at the end of the targeting vector for negative selection against nonhomologous integration. The targeting vector was linearized and introduced into E14 ES cells by electroporation, and transfectants were selected with G418 and ganciclovir. Homologous recombinants were identified by PCR and Southern blot analysis. Five clones were selected and injected into C57BL/6 blastocysts to generate chimeric mice. Chimeras from one clone transmitted the mutated allele through germline. Homozygous mice were generated by intercrosses of heterozygous mice (Fig. 1 B).
To verify that the targeted BST-1 gene was not expressed, RT-PCR analysis with primers of exon 1 and exon 4 were conducted using reverse transcribed cDNA from bone marrow, spleen, and thymus. The expected product was only detected using RNA from BST-1+/+ and BST-1+/− mice (Fig. 1,C). Moreover, production of BST-1 protein was examined by immunohistochemical analyses of the spleen and PP using a polyclonal antiserum reactive with BST-1 (Fig. 1 D). In the wild-type mice, BST-1 was expressed by the reticular cells of the spleen and PP, as reported in the immunohistochemical analysis using BP-3-specific Ab (14). However, in the BST-1−/− mice, BST-1 was not detected, confirming that these mice are deficient in BST-1. BST-1-deficient mice thrived and reproduced as well as their wild-type littermates.
Developmental defect in peritoneal B cells and abnormal expression of CD38 in B lymphocytes
Since the BST-1 expression is detected in the early ontogeny of lymphoid cells and is restricted at the critical stages of both B and T cell development, our experiment was aimed to elucidate whether BST-1 deficiency could affect lymphocyte development.
Interestingly, in 2- to 3-wk-old BST-1−/− mice, the number of peritoneal μbright δ− B cells, i.e., B-1 cells expressing CD5 or Mac1, was reduced by 30 to 70% compared with that in control mice (Fig. 2). However, the percentage of B-1 cells was recovered to the normal level in adult BST-1−/− mice older than 6 wk. Furthermore, the surface expression of CD38 was compared using BST-1−/− mice and control mice, since the molecule has homology of amino acid sequences with BST-1 (1, 4, 5). In mice, it is known that almost all B-lineage cells (CD45R+ cells) in the bone marrow and spleen express CD38 (24, 25, 26). An increase in a minor B-lineage population expressing low to no CD38 (CD38low/−) was noted in the adult bone marrow and spleen of BST-1−/− mice. In the wild-type littermates, the CD38low/− populations in the bone marrow and spleen were 1.5 and 0.6%, respectively, whereas in BST-1−/− mice, the populations increased to 8.9 and 3.5%, respectively (Fig. 3 a).
Inasmuch as our previous study reported that the expression of CD38 was up-regulated at the transition from pro-B to pre-B cells in the bone marrow (15), we next examined the expression of CD38 in the developmental stages of bone marrow B-lineage cells. In wild-type littermates, CD38-positive cells increased in the R2 fraction (CD45intermediate CD43−) containing pre-B and immature B cells; however, such an increase was not observed in BST-1−/− mice. This result suggested that the up-regulation of CD38 expression was delayed during B lymphopoiesis in the bone marrow of BST-1−/− mice (Fig. 3 b). However, the expression levels of CD38 by the recirculating B cells (in the R3 fraction) of BST-1−/− mice were normal, suggesting that the delayed up-regulation of CD38 caught up during peripheral B cell maturation.
Except for those defects, no difference was observed in the cell number of bone marrow, thymus, and spleen between adult BST-1−/− mice and wild-type littermates (data not shown). We observed no differences in the populations of B lymphoid progenitors defined by CD45R and IgM in the bone marrow and of T lymphoid progenitors by CD4 and CD8 in the thymus (Fig. 4, a and b). The mature B and T cell populations defined by IgM, IgD, CD4, and CD8 in the spleen were also normal (Fig. 4 c). Our analysis was extended to other surface markers, including CD24, BP-1, c-kit, CD19, Ig-β, CD23, Igκ, Igλ, Thy1, IL-2Rα, TCRαβ, Mac-1, and Gr-1 (data not shown). The differences in these surface expressions were also indistinguishable between normal and BST-1−/− mice, indicating that the major pathway of lymphoid development was unaffected by depletion of the BST-1 gene.
Defects in TI-2 Ag-induced IgG3 production in BST-1 knockout mice
The expression of murine BST-1/BP-3 on the reticular cells of splenic white pulps, lymph nodes, and PP suggested that BST-1 can support the functions of mature B cells. To evaluate the role of BST-1 in the induction of humoral immune responses, we first analyzed the concentrations of serum Igs by IgM-, IgG subclass-, and IgA-specific ELISA. The serum Ig levels of all isotypes examined in BST-1−/− mice were the same as those in wild-type littermates (Fig. 5,a). In the next series of experiments, the influence of BST-1 deficiency on the induction of Ag-specific humoral immune responses was examined using a group of TD, TI-1, and TI-2 Ags. When BST-1−/− mice were systemically immunized with DNP-KLH as a TD Ag and with TNP-LPS as a TI-1 Ag, BST-1−/− mice mounted normal levels of Ag-specific primary and secondary responses of IgM as well as IgG classes (Fig. 5, b and c). However, when TNP-Ficoll, as a TI-2 Ag, was used, a reduction of TNP-specific IgG3, but not IgM production, was observed (Fig. 5 c).
Normal proliferative B cell responses in BST-1 deficient mice
Anti-CD38 Ab induces potent proliferative and antiapoptotic responses in splenic B cells, except for some B cell populations such as xid and neonatal B cells in which these B cells were unresponsive to anti-CD38 Ab and TI-2 Ag (24, 25, 27, 28). An increase in CD38low/− splenic B cells and a defective TI-2 response in BST-1−/− mice prompted us to examine whether B cells of BST-1−/− mice have defects in responses to anti-CD38 Ab. However, BST-1−/− B cells have no apparent defects in response to anti-CD38 Ab (Fig. 6). Also no differences were observed between the B cells of wild-type and BST-1−/− mice in their responses to anti-μ Ab, anti-CD40 Ab, or LPS (Fig. 6). Thus, the splenic BST-1−/− B cells do not have inherent defects in BCR, CD40, LPS, or CD38 signaling. Furthermore, immunohistochemical analysis revealed that the structure of T and B cell areas in splenic white pulps was intact in BST-1−/− mice (data not shown). These data suggested that the defective IgG3 response to TI-2 Ags in BST-1−/− mice was ascribed to the lack of BST-1 normally expressed by splenic reticular cells.
Impaired mucosal immune responses in BST-1−/− mice
BST-1 is also expressed by the reticular cells in the PP and the brush border of the intestinal epithelial cells (14). Such a localization of BST-1 suggests the involvement of BST-1 in the mucosal immune responses. To this end, we addressed the possible role of BST-1 in the induction and regulation of mucosal immunity. When the concentration of total IgA in the sera and fecal extracts was examined and compared between wild-type and BST-1−/− mice, unaltered levels of total IgA were seen in both systemic and mucosal compartments (Fig. 5 a and data not shown). These results indicated that normal numbers of IgA-producing plasma cells were developed in BST-1−/− mice.
To examine possible influences of BST-1 deletion on the induction of the Ag-specific mucosal immune response, BST-1−/− and control mice were orally immunized with CT, a strong immunogen as well as adjuvant for the induction of mucosal IgA production. When the levels of CT in sera and fecal extracts were examined, it was interesting to note that the levels of anti-CT IgA and IgG Abs in the fecal extracts from BST-1−/− mice were at least twofold lower than those in wild-type littermates (Fig. 7,a). However, in the sera no difference was observed between BST-1−/− and wild-type mice. To further examine the influence of BST-1 deficiency on the CT-specific mucosal immune response, the frequency of CT specific Ab-forming cells (AFCs) in different mucosa-associated tissues was measured by ELISPOT assay. In BST-1−/− mice, the number of CT-specific IgA AFC was low in the intestinal LP, which is considered to be a major mucosal effector site for IgA production. On the other hand, in the spleen and in MLN and PP, representing the systemic compartments and the mucosal inductive tissues, respectively, the frequencies of CT-specific IgA and IgG AFCs in BST-1−/− mice were comparable to those seen in wild-type mice (Fig. 7 b). These results indicated that BST-1 could be an important molecule for the induction of the Ag-specific IgA B cell response in the mucosa-associated effector tissue.
In this study, BST-1 knockout mice were generated by targeted disruption of the gene to determine whether there was an obligatory involvement of BST-1 in the induction of the immune response in vivo.
Role of BST-1 in lymphocyte development and its relationship to CD38
As described in the introduction, the accumulated evidence has indicated that BST-1 may play an important role in lymphoid cell development. The present study produced the interesting results that BST-1 gene deletion led to developmental defects in peritoneal B-1 cells and to an increase in the CD45R+CD38low/− population in bone marrow and spleen. Since B-1 cells are considered to be of fetal liver origin, and BST-1 is expressed on the B progenitor of fetal liver maximally on day 17 and is decreased on day 18 (15, 29), BST-1 may enhance a narrow window of the maturation or migration of B-1-lineage cells. It would be interesting to examine the expression of BST-1 in IL-5−/− or IL-5R−/− mice, since the deletion of IL-5 or IL-5R genes has been shown to influence B-1 cell development (30, 31).
BST-1 and CD38 molecules are members of a novel family of ectoenzymes (26, 32), but the functional relation or difference between BST-1 and CD38 is still largely unknown. To date, a different expression pattern on B cell development was observed between these molecules. The expression of CD38 is low on pro-B cells, up-regulated at the transition from pro-B to pre-B cells, and increased further during B cell maturation, while that of BST-1 is restricted on pro-B, pre-B, and immature B cells and then down-regulated at the mature B cell stage (15). However, in secondary lymphoid organs such as the spleen, PP, and lymph node, BST-1 is expressed on the reticular cells that may provide positive signals for the Ab production. Since both BST-1 and CD38 are expressed on B progenitor cells from pro-B to immature B cell stages, the functions of both molecules may complement each other at these stages. This may explain why BST-1−/− mice showed almost normal B cell development, except for CD38 expression. Unexpectedly, this study revealed that the loss of BST-1 was not compensated by the up-regulation of the expression of CD38. Instead, CD38 up-regulation at the pre-B stage was delayed in the bone marrow. Moreover, in the spleen of BST-1 knockout mice, the CD45R+CD38low/− population was increased. In normal immunized mice, the CD45R+CD38low/− population in peripheral lymphoid organs was considered to be germinal center B cells and mature plasma cells (33). These germinal center B cells have been shown to have down-regulated surface expression of IgM and IgD. However, the CD45R+CD38low/− population observed in BST-1 knockout mice expressed high levels of IgM and IgD (data not shown), suggesting that this population may not be germinal center B or mature plasma cells. Therefore, it is possible that some populations of splenic B cells could not express high levels of CD38 due to the lack of BST-1.
Involvement of BST-1 in systemic TI-2 Ag-induced IgG3 response
Analyses of the systemic humoral responses in BST-1 knockout mice revealed the decreased IgG3 response to a TI-2 Ag, TNP-Ficoll. The selective defects in the TI-2, but not the TD or TI-1, response, suggested that Ag-specific Ab responses that were dependent on BST-1 seemed to be limited. To interpret the relationship between the loss of BST-1 and the impaired TI-2 response, expression of BST-1 on the cells at the site of Ab production, the architecture of lymphoid tissues, and the responsibility of B cells to the stimuli important in differentiation to Ab-producing cells should be considered. BST-1 is not expressed on mature B cells, but is expressed on reticular cells in secondary lymphoid organs in normal mice. B cells of BST-1 knockout mice were essentially normal because the splenic B cell responses of BST-1 knockout mice to various proliferative stimuli were comparable to those of wild-type mice. Furthermore, the architectures of T and B cell areas in the secondary lymphoid organs were not altered in BST-1 knockout mice, except for the lack of BST-1 on reticular cells (data not shown). Taken together, these findings suggested that defective responses to TI-2 Ag in BST-1 knockout mice were probably due to the absence of BST-1 expression on the reticular cells of splenic white pulps, although this issue remains to be resolved.
Defective responses to TI-2 Ags were also observed in xid mice, knockout mice of Btk, Ig-α, IL-5R, protein kinase C, and CD22 (31, 34, 35, 36, 37, 38, 39). These molecules are expressed by mature B cells, and the absence of these molecules affected signaling in B cells in response to various proliferative stimulations. Among cytokines, TNF-lymphotoxin-α double-knockout mice lacked a TI-2 response (40). The cell lineages affected by the deficient TNF-lymphotoxin-α are not clear, but the poorly organized architecture of lymphoid tissues seems to be responsible for this defect. Compared with the molecules affecting the TI-2 response in vivo, BST-1 is unique, in that BST-1 is not a cytokine but, rather, a membrane protein expressed by the reticular cells in microenvironment of Ab production and is required only for the TI-2 response of the IgG3 class.
TI Ags are classified as the Ags that stimulate Ab production in the absence of MHC class II-restricted T cell help. However, TI-2 Ag by itself is not sufficient to stimulate B cells for subsequent Ig synthesis, and it must therefore be accompanied by other lymphoreticular cells that can interact with B cells either directly or indirectly (41, 42). There is evidence implicating NK cells, T cells, and follicular dendritic cells as playing important roles in response to TI-2 Ag (41). Our results suggest that reticular cells expressing BST-1 could be included in the auxiliary cells for the TI-2 response of the IgG3 class. Although TI-2-specific IgG3 production was impaired in BST-1 knockout mice, BST-1 is not an essential factor that induces the class switch to IgG3 such as IFN-γ, because IgG3 production was normal when TI-1 or TD Ags were immunized. IFN-γ is a switch factor for IgG3 when B cells are activated with anti-Ig-dextran, a model Ag for the TI-2 response in vitro (43, 44). Furthermore, immnohistochemical analyses revealed that immunization with TNP-Ficoll induced the appearance of IFN-γ-producing T cells and NK cells in the outer parts of periarteriolar lymphoid sheath, where TNP-AFC localize (45). At present, we do not have any evidence that IFN-γ production was impaired in BST-1−/− mice, because they mounted the normal Th1 response in which IFN-γ plays a major role (H. Tsutsui and K. Nakanishi, unpublished observation). Moreover, IFN-γ has been shown to induce BST-1 expression on peritoneal macrophages in vitro (M. Itoh and K. Ishihara, unpublished observation). These findings suggested that BST-1 may be up-regulated in situ by IFN-γ and may function as the effector molecule.
Although the exact mechanisms by which BST-1 supports the TI-2 response are currently unknown, it is possible that BST-1 transduces certain signals to activate the reticular cells to support B cells or that putative BST-1 ligands transduce signals to B cells. Actually, we showed that BST-1 facilitated the growth of a pre-B cell line (1) and recently found the expression of a BST-1 binding molecule on a B cell line (K. Ishihara, unpublished data). There is another possibility that the lowered expression of CD38, which is capable of augmenting BCR signaling (46), impaired the TI-2 response, since it was reported that marginal zone (MZ) B cells that respond to TI-2 Ag are CD38high (47). However, our findings do not support this possibility, since CD45R+CD38low/− cells were not typical MZ B cells, and the expression of CD38 by MZ B cells was not significantly altered in BST-1 knockout mice (data not shown). Since enzyme activity of CD38 has been suggested to affect the growth of B cells (7), it is likely that the enzyme activities of BST-1 on reticular cells are also involved in supporting the survival, growth of B cells, or class switching to IgG3. This issue remains to be resolved.
BST-1 is required for the full response of TD Ag-induced mucosal IgA responses
It was interesting to note that Ag (CT)-specific mucosal immune responses after oral immunization were impaired in BST-1 knockout mice. BST-1 is expressed on the reticular cells of PP, which is considered to be IgA-inductive tissue containing high numbers of IgA precursors cells (sIgA+ B cells) (14, 18). Following Ag stimulation, these sIgA+ B cells leave the tissue and migrate to IgA effector sites such as intestinal LP. When the frequency of CT-specific IgA AFC was compared in different tissues of BST-1−/− and control mice following oral immunization, the reduction of Ag-specific IgA-producing cells was most obvious in the intestinal LP of BST-1−/− mice. On the other hand, the numbers of CT-specific IgA AFC were comparable in the spleen, PP, and MLN in BST-1−/− and wild-type mice. In addition, it was important to indicate that the levels of CT-specific IgG serum responses induced by oral immunization were similar between these two mouse groups. These results indicated that Ag-specific mucosal, but not systemic, responses were impaired in BST-1−/− mice orally immunized with TD Ag such as CT.
Since PP-derived dendritic cells have the ability to promote commitment to IgA (48), it is possible that BST-1 gene deletion may influence the function of PP dendritic cells for their ability to induce IgA-committed sIgA+ B cells. However, the committed sIgA+ cells are comparable between BST-1−/− and wild-type mice in PP, and serum levels of IgA and systemic anti-CT Ab production in BST-1 knockout mice were normal (Fig. 7), suggesting that the commitment of IgA B cells was not impaired. Thus, BST-1 expressed in PP may support the maturation or homing of B cell blasts that migrate to the LP.
Since the majority of B cells in the LP are B-1 cells, and there is traffic between the peritoneal cavity and the LP (49, 50), it is possible that the impaired development of peritoneal B-1 cells of BST-1−/− mice affects the function of peritoneal B-1 cells. This issue has yet to be examined.
Previous studies have demonstrated that cytokines such as IL-4, IL-5, IL-6, and TGF-β regulate IgA B cell differentiation in vitro (51, 52, 53, 54, 55). For example, both mucosal and systemic Ab responses to orally administered TD (e.g., tetanus toxoid and KLH) Ags were impaired in IL-4 knockout mice (56, 57). Although conflicting results were generated in IL-6 gene-deleted mice (58, 59), a study demonstrated the reduction of IgA responses in both mucosal and systemic sites. In contrast to these studies, mucosal immunodeficiency in BST-1 knockout mice was unique, because only local immune responses to TD Ag were impaired.
Although our current findings do not offer any explanation for the mechanisms of the partial immunodeficiency observed in BST-1 knockout mice, gene targeting clearly demonstrated that BST-1 is a unique surface molecule expressed by reticular cells and has a function to support the systemic B cell differentiation induced by TI-2 Ag. Furthermore, BST-1 knockout mice could be useful animal models that could be extended to demonstrate the importance of mucosal immunity, especially in the role of secretory IgA as a first line of defense against different infectious agents, since BST-1 knockout mice could be considered mucosal immunodeficient mice with intact systemic immune responses to orally administered Ags.
We thank Drs. H. Kikutani, T. Yasui, and T. Abe for useful advice and help in making knockout mice; Dr. J. Miyagawa for useful advice in preparing immunohistochemical analysis; K. Nishikawa for technical assistance with the ELISA assay; Drs. K. Nakanishi, H. Tsutsui, and S. Kashiwamura for assays of Th1 and Th2 responses in vivo; Dr. M. Cooper for providing BP-3 Ab; and Drs. Y. Dohi and S. Ono for helpful advice and suggestions. We also thank R. Masuda and T. Kimura for their secretarial assistance.
This work was supported in part by a grant-in-aid for COE Research from the Ministry of Education, Science, Sports, and Culture in Japan; by the Osaka Foundation for Promotion of Clinical Immunology; by the Mochida Memorial Foundation for Medical and Pharmaceutical Research in Japan; and by the Japan Health Science Foundation. M.I. is supported by Research Fellowships from the Japanese Society for the Promotion of Science for Young Scientists.
Abbreviations used in this paper: BST-1, bone marrow stromal cell Ag-1; PP, Peyer’s patch; sIgA, surface IgA; CT, cholera toxin; LP, lamina propria; TD, thymus dependent; TI, thymus independent; ES, embryonic stem; PE, phycoerythrin; KLH, keyhole limpet hemocyanin; TNP, 2,4,6-trinitrophenyl; ELISPOT, enzyme-linked immunospot; MLN, mesenteric lymph node; mCD38, mouse CD38; AFC, Ab-forming cell; MZ, marginal zone.