Independent Roles for BAFF in B Cell Factor

The Journal of Immunology

TNF Family Member B Cell-Activating Factor (BAFF) Receptor-Dependent and -Independent Roles for BAFF in B Cell Physiology1

Yoshiteru Sasaki,2* Stefano Casola,* Jeffery L. Kutok,† Klaus Rajewsky,* and Marc Schmidt-Supprian2*

The cytokine TNF family member B cell-activating factor (BAFF; also termed BLyS) is essential for B cell generation and maintenance. Three receptors have been identified that bind to BAFF: transmembrane activator, calcium modulator, and cyclophilin ligand interactor (TACI); B cell maturation Ag (BCMA); and BAFF-R. Recently, it was shown that A/WySnJ mice, which contain a dramatically reduced peripheral B cell compartment due to decreased B cell life span, express a mutant BAFF-R. This finding, together with normal or enhanced B cell generation in mice deficient for BCMA or TACI, respectively, suggested that the interaction of BAFF with BAFF-R triggers signals essential for the generation and maintenance of mature B cells. However, B cells in mice deficient for BAFF differ phenotypically and functionally from A/WySnJ B cells. Residual signaling through the mutant BAFF-R could account for these differences. Alternatively, dominant-negative interference by the mutant receptor could lead to an overestimation of the importance of BAFF-R. To resolve this issue, we generated BAFF-R-null mice. Baff-r؊/؊ mice display strongly reduced late transitional and follicular B cell numbers and are essentially devoid of marginal zone B cells. Overexpression of Bcl-2 rescues mature B cell development in Baff-r؊/؊ mice, suggesting that BAFF-R mediates a survival signal. CD21 and CD23 .surface expression are reduced on mature Baff-r؊/؊ B cells, but not to the same extent as on mature B cells in BAFF-deficient mice -In addition, we found that Baff-r؊/؊ mice mount significant, but reduced, Ag-specific Ab responses and are able to form spon taneous germinal centers in mesenteric lymph nodes. The reduction in Ab titers correlates with the reduced B cell numbers in the mutant mice. The Journal of Immunology, 2004, 173: 2245–2252.

he TNF family member B cell-activating factor (BAFF,3 and MZ B cells, whereas B1 B cells are not affected. This was also known as BLyS, TALL-1, zTNF4, or THANK) is initially attributed to a complete block of B cell development at the T required for generation and maintenance of the mature B T1 stage (11, 12). However, later analyses showed that some ma- cell pool. BAFF is expressed on a variety of cells of myeloid and ture B cells can be generated in the absence of BAFF (13, 14). by guest on October 1, 2021. Copyright 2004 Pageant Media Ltd. other origin, and its expression pattern may control the compart- These B cells lack expression of the CD21 and CD23 surface mentalization of B cells in the peripheral immune system, such as markers, suggesting that expression of these receptors depends on the localization of B cells to B cell follicles (1–5). Newly formed BAFF (14). B cells are recruited into germinal centers (GC) upon B cells in the bone marrow (BM) reach the spleen as transitional Ag stimulation in the absence of BAFF, albeit in a transient fash- type 1 (T1) B cells. These cells subsequently develop into transi- ion (13, 15). tional type 2 (T2), transitional type 3 (T3), and into the mature In contrast, transgenic mice engineered to overexpress BAFF follicular and marginal zone (MZ) B cells (6, 7). BAFF promotes show increased numbers of peripheral B cells and elevated serum the survival of T2, T3, and mature B cells (8–10). Genetic dis- Ig levels, including anti-DNA Abs, rheumatoid factor, and circu- ruption of the Baff gene leads to a profound defect in mature B2

http://classic.jimmunol.org lating immune complexes, and develop autoimmune symptoms re- sembling human systemic lupus erythematosus and Sjo¨gren’s syndrome as they age. Elevated serum BAFF levels have been *CBR Institute for Biomedical Research, Harvard Medical School, Boston, MA detected in patients with rheumatoid arthritis, systemic lupus ery- 02115; and †Department of Pathology, Brigham and Women’s Hospital, Boston, MA 02115 thematosus, and Sjo¨gren’s syndrome (1–5). Received for publication March 24, 2004. Accepted for publication June 1, 2004. BAFF can bind to three receptors of the TNFR family: transmembrane activator, calcium modulator, and cyclophilin ligand Downloaded from The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance interactor (TACI); B cell maturation Ag (BCMA); and BAFF-R with 18 U.S.C. Section 1734 solely to indicate this fact. (also known as BLyS receptor 3 or BR3). TACI and BMCA also 1 This work was supported by the National Institutes of Health Grants AI057947 and bind to another TNF family member, a proliferation-inducing li- 1R37AI054636 and the Uehara Memorial Foundation. gand (1–5). A proliferation-inducing ligand, BCMA, and TACI are 2 Address correspondence and reprint requests to Dr. Yoshiteru Sasaki or Dr. Marc Schmidt-Supprian, CBR Institute for Biomedical Research, 200 Longwood Avenue, all dispensable for B cell development (16–20), indicating that Boston, MA 02115. E-mail address: [email protected] or supprian@cbr. triggering of BAFF-R by BAFF is essential for the generation of med.harvard.edu mature B cells. A/WySnJ mice contain 10% of the peripheral B 3 Abbreviations used in this paper: BAFF, TNF family member B cell-activating cells found in control A/J mice, due to a reduced life span of B2, factor; BCMA, B cell maturation Ag; BM, bone marrow; CG, chicken ␥-globulin; ES, embryonic stem; GC, germinal center; HPRT, hypoxanthine phosphoribosyltrans- but not of B1 cells (21–24). These mice harbor a transposon in- ferase; LN, lymph node; mLN, mesenteric LN; MZ, marginal zone; NP, 4-hydroxy- sertion that replaces the last 8 aa of the C terminus of BAFF-R 3-nitrophenylacetyl; PNA, peanut agglutinin; T1/2/3, transitional type 1/2/3; TACI, transmembrane activator, calcium modulator, and cyclophilin ligand interactor; TD, T with 21 aa encoded by the insertion (25, 26). The mutant receptor dependent; TI, T independent. is expressed on the surface of A/WySnJ B cells and is capable of

binding BAFF (27), suggesting that its signaling capacities are Technologies). Baff-r and Hprt (hypoxanthine phosphoribosyltransferase) impaired. This notion is supported by the fact that BAFF admin- cDNA was amplified by PCR with the following intron-spanning primers: Ј Ј Ј istration increases B cell numbers in A/J, but not in A/WySnJ mice BAFF-R-1, 5 -CCCAGACTCGGAACTGTCCC-3 ; BAFF-R-2, 5 -GTA GAGATCCCTGGGTTCCAG-3Ј; Hprt-1, 5Ј-ATTAGCGATGATGAAC (24). However, significant differences exist between A/WySnJ and CAGGTTA-3Ј; Hprt-2, 5Ј-CCAGTTAAAGTTGAGATCATCTCCAC-3Ј. Ϫ/Ϫ Ϫ/Ϫ Baff mice; A/WySnJ mice, in constrast to Baff mice, con- PCR conditions were as follows: 94¡C for 30 s, 58¡C for 30 s, and 72¡C tain CD21- and CD23-expressing follicular and CD21highCD23low for 30 s for 40 cycles with1UofKODHiFi polymerase (Novagen, MZ B cells (12, 26). They exhibit IgM serum levels comparable to Madison, WI). A/J mice (28), whereas in BaffϪ/Ϫ mice IgM levels in the blood are strongly reduced (12). A/WySnJ mice mount efficient IgM re- Flow cytometry sponses against T-dependent (TD) and T-independent (TI) Ags, Single-cell suspensions prepared from various lymphoid organs and the but are profoundly deficient in producing primary or secondary peritoneum were stained with the following mAbs conjugated to FITC, PE, IgG responses (28). The Ag-specific Ab production by BaffϪ/Ϫ CyChrome, allophycocyanin, or biotin: anti-CD5 (53-7.3), anti-CD19 (ID3), anti-CD21 (7G6), anti-CD23 (B3B4), anti-CD38 (90), anti-CD1d mice in response to TD and TI Ags, in contrast, is dramatically (1B1), anti-Fas (Jo2), anti-IgD (11-26c.2a) (all from BD Pharmingen, San reduced compared with wild-type mice, suggesting that production Diego, CA), and anti-AA4.1 (eBioscience, San Diego, CA). Additional, of all Ig subclasses is affected (12). “homemade” mAbs to IgM (R33-2412) and B220 (RA3-6B2) were used. The differences observed in B cell phenotypes and function be- Rabbit polyclonal Abs to mouse BAFF-R were generated by immunization tween A/WySnJ and BaffϪ/Ϫ mice can be explained by the nature with a synthesized mouse BAFF-R peptide (aa 1Ð17). PNA conjugates were PNA-FITC and PNA-biotin (Vector Laboratories, Burlingame, CA). of the A/WySnJ mutation, which leaves most of the BAFF-R in- Biotin conjugates were visualized with CyChrome or allophycocyanin (BD tact. The mutant receptor could therefore retain some signaling Pharmingen). Anti-BAFF-R Ab was visualized with goat anti-rabbit IgG functions. Alternatively, it could heterotrimerize with other TNF- FITC (Zymed Laboratories, San Francisco, CA). All samples were ac- R-like molecules, such as BCMA or TACI, and act as a dominant- quired on a FACSCalibur (BD Pharmingen), and results were analyzed with CellQuest software. negative molecule. Therefore, B cell physiology could be affected more by expression of a mutant BAFF-R, than by lack of the Immunohistochemistry receptor, leading to an overinterpretation of the role of BAFF-R. However, it is also possible that BAFF exerts some of its functions For staining of B cells and macrophages, frozen 6-␮m sections were through its other receptors, BCMA and TACI, or in a non-B cell- thawed, air dried, fixed in acetone, and stained for1hatroom temperature in a humidified chamber with biotinylated rat anti-CD19 (BD Pharmingen) autonomous fashion through as yet unidentified receptors. and rat anti-MOMA-1 (Cedarlane Laboratories, Hornby, Canada), fol- To more precisely define the role of BAFF-R in B cell devel- lowed by HRP-conjugated goat anti-rat IgG and alkaline phosphatase-con- opment, we generated mice that are deficient for this receptor. jugated streptavidin. For staining of B and T cells, sections were stained for Baff-rϪ/Ϫ mice demonstrate a block in B cell development at the 1 h at room temperature in a humidified chamber with biotinylated anti- CD19 and armenian hamster anti-CD3⑀ (eBioscience). Bcl-6 expression transitional stage. The mutant mice show a 40-fold reduction in was detected with a rabbit anti-Bcl-6 (Santa Cruz Biotechnology, Santa numbers of mature B2 cells and contain very few, if any, MZ B Cruz, CA), followed by incubation with HRP-conjugated goat anti-rabbit cells. Baff-rϪ/Ϫ B cells produce reduced, but significant Ag-spe- IgG (DakoCytomation, Carpinteria, CA). cific Ab titers in response to immunizations with TI and TD Ags,

by guest on October 1, 2021. Copyright 2004 Pageant Media Ltd. and they develop into GC B cells in mesenteric lymph nodes Immunization and serum analyses (mLNs) in response to gut-derived Ags. The main function of Mice were immunized with 100 ␮g of TD Ag 4-hydroxy-3-nitrophenyla- BAFF-R appears to provide a survival signal to mature B cells, as cetyl chicken ␥-globulin (NP-CG) in alum i.p. Mice were boosted with 10 overexpression of Bcl-2 rescues mature B cell development in ␮g of NP-CG without adjuvant on day 48. TI immune responses were Ϫ Ϫ ␮ Baff-r / mice. elicited by i.p. injections of 50 g of NP-Ficoll. Mice were bled before and after immunization from tail veins. Ig serum concentrations were deter- Materials and Methods mined by ELISA, as described previously (33). Generation of Baff-rϪ/Ϫ mice and maintenance of mice Results Construction of a conditional Baff-rFL targeting vector. An EcoRI fragment http://classic.jimmunol.org Generation of BAFF-R-deficient mice of the mouse Baff-r genomic locus in pBluescript II was used. The targeting vector was designed to flank exons 3 and 4 of the Baff-r gene with two loxP To study the physiological role of BAFF-R in vivo, we generated r sites. An frt site-flanked selection cassette, containing a neo gene and the BAFF-R-deficient mice by gene targeting. The targeting vector Flpe6 cDNA cloned under control of the ACE promoter (29), was placed into the fourth intron of the Baff-r gene. A 1.9-kb AvrII-MfeI fragment was was designed to place exons 3 and 4, which encode the entire used as 5Ј homology region, a 1.3-kb MfeI-AvrII fragment was placed extracellular and transmembrane domains and part of the cytoplas- between the two loxP sites, and a 5.0-kb AvrII-BamHI fragment was used mic domain, between loxP sites (Fig. 1A). Two C57BL/6-derived Ј Downloaded from as 3 homology region. A thymidine kinase gene was used for negative ES clones (30) with the appropriately targeted allele were obtained selection of clones with random integration of the targeting vector. (Fig. 1B), injected into BALB/c blastocysts to generate chimeric Bruce-4 embryonic stem (ES) cells (30) derived from C57BL/6 mice were transfected, cultured, and selected, as previously described (31). Of mice, which transmitted the targeted loxP-flanked Baff-r allele 475 G418 and gancyclovir-resistant colonies, 2 were identified as homol- (Baff-rFL) to their progeny (Fig. 1C). All mice were maintained on ogous recombinants with cointegration of the second loxP site by Southern a pure C57BL/6 genetic background. The frt-flanked neor casette Ј blot analysis of BglII-digested DNA, using a 465-bp fragment as 5 exter- can be removed through intercrossing with Flpe-deleter mice (34). nal probe (Fig. 1). Mice were bred and maintained in specific pathogen-free Ϫ FL/ϩ conditions; all mouse protocols were approved by the Harvard University To obtain a Baff-r null allele (Baff-r ), Baff-r mice were Institutional Animal Care and Use Committee and by the CBR Institute for crossed with the deleter strain (35) (Fig. 1C). The expression of Ϫ Ϫ Biomedical Research. The E␮-bcl-2-22 (32) transgenic mouse strain was BAFF-R in BAFF-R-deficient mice (Baff-r / ) was examined obtained from The Jackson Laboratory (Bar Harbor, ME). both at the RNA and the protein level. No Baff-r mRNA could be Ϫ/Ϫ RT-PCR analysis detected by RT-PCR analysis in Baff-r mice (Fig. 1D). FACS analysis using a polyclonal anti-BAFF-R Ab revealed absence of Total RNA was extracted from splenocytes with TRIzol (Invitrogen Life ϩ Ϫ/Ϫ Technologies, Carlsbad, CA), according to the manufacturer’s protocol. BAFF-R expression on the surface of B220 Baff-r splenic One microgram of total RNA was used for first-strand cDNA synthesis cells (Fig. 1E). Despite the reduction of BAFF-R surface levels on ϩ Ϫ with random hexamer primers and Superscript RT III (Invitrogen Life Baff-r / B cells (Fig. 1E), we did not detect any statistically The Journal of Immunology 2247

FIGURE 1. Generation of BAFF-R-deﬁcient mice. A, Schematic rep- resentation of the targeting strategy. a, Structure of Baff-r cDNA. b, Sketch of the Baff-r genomic locus. c, The targeting construct was designed to ﬂank exons 3 and 4 by two loxP sites (triangles) and a Neor casette by two frt sites (ovals). d, The targeted allele. The probe used for Southern blot analysis is shown as a bar. e, The structure of the Baff-rϪ allele. B, Southern blot analysis of targeted ES clones. Genomic DNA from wild-type (lane 1)

by guest on October 1, 2021. Copyright 2004 Pageant Media Ltd. and two targeted ES clones (lanes 2 and 3) was digested with BglII and probed with the external probe shown in Ad. C, Southern blot analysis of Ϫ Ϫ tail DNA of Baff-rϪ/Ϫ mice. DNA was digested with BglII (left panel)or FIGURE 2. Flow cytometric analysis of B cell population in Baff-r / ϩ Ϫ SacI(right panel) and hybridized with the probe shown in Ad. D, RT-PCR mice. A, Comparison of BM B cells from wild-type, Baff-r / , and Baff- Ϫ Ϫ analysis to detect expression of Baff-r in spleen. HPRT expression is shown r / mice. B, Analysis of splenocytes for the surface expression of IgM as control. E, FACS analysis of BAFF-R expression on the surface of and IgD. C, Analysis of LN lymphocytes for the surface expression of IgM B220ϩ splenocytes using rabbit anti-BAFF-R serum. ϩ/ϩ, ϩ/Ϫ, and Ϫ/Ϫ, and IgD. D, Analysis of B cell subpopulations in the peritoneal cavity (PC). Wild-type, heterozygous, and homozygous BAFF-R knockout mice, re- B2 cells are B220highCD19low, and B1 cells are B220lowCD19high. B1a ϩ spectively. The shaded histogram (control) shows staining with preimmune cells are CD5 , whereas B1b cells are CD5low. The numbers refer to the rabbit serum. percentages of cells in individual gates of total live lymphocytes, or of cells http://classic.jimmunol.org in the gate indicated in brackets above the dot plots.

signiﬁcant differences in B cell development in our analyses between Baff-rϩ/Ϫ and wild-type mice, which are therefore collec- tively referred to as control mice below. detected in LNs from Baff-rϪ/Ϫ in comparison with control mice (Fig. 3C). In the peritoneal cavity of the mutants, the Ϫ/Ϫ Decreased numbers of mature B cells in Baff-r mice high intemediate Downloaded from B220 CD19 B2 cell population was reduced Ͼ10-fold, B cell development in Baff-rϪ/Ϫ mice was examined by ﬂow cy- but the CD19highB220low B1 cell population was not affected (Fig. ϩ Ϫ tometric analyses. In the BM of Baff-rϪ/Ϫ mice, early B cell de- 2D, upper panels). The ratio of B1a (CD5 ) to B1b (CD5 ) cells Ϫ Ϫ velopment was essentially normal. The only difference between was also normal in Baff-r / mice (Fig. 2D, lower panels). Anal- Ϫ Ϫ BM B cells of control and Baff-rϪ/Ϫ mice was a strong reduction ysis of T cell development in Baff-r / mice yielded normal num- of mature recirculating B cells in the latter (Fig. 2A). The fraction bers of conventional peripheral CD4 and CD8 T cells. The gen- ϩ ϩ of IgD-positive B cells was Ͼ10-fold reduced in Baff-rϪ/Ϫ BM eration of memory-type T cells, regulatory CD4 CD25 T cells, compared with control mice (data not shown). In the spleen, B cell and NK-like T cells was also not affected by the absence of numbers were strongly reduced (Table I). The reduction occurred BAFF-R (data not shown). mainly in the IgMlowIgDϩ mature B cell subset (Fig. 2B). Immu- ϩ B cell maturation is blocked at the transitional stages in nohistological staining of spleen sections revealed small CD19 B Ϫ/Ϫ cell clusters, but essentially normal-sized CD3ϩ T cell areas in Baff-r mice Baff-rϪ/Ϫ mice (Fig. 3A). Lymph nodes (LNs) of Baff-rϪ/Ϫ mice Newly formed or immature B cells emigrate from the BM, reach also had a Ͼ20-fold reduced IgMlowIgDϩ mature B cell fraction the spleen as so-called transitional B cells, and subsequently ma- (Fig. 2C). Accordingly, few and very small B cell clusters were ture into follicular and MZ B cells. Transitional B cells can be 2248 BAFF-R SIGNALS FOR B CELL SURVIVAL

Table I. Number of splenic B cell populations in Baff-rϩ/ϩ, Baff-rϩ/Ϫ, and Baff-rϪ/Ϫ micea

Mature B Cell MZ B Cell Immature B Cell T1 T2 T3 (B220ϩ, AA4.1Ϫ) (CD19ϩ, CD1dhigh) (B220ϩ, AA4.1ϩ) (IgMϩ, CD23Ϫ) (IgMϩ, CD23ϩ) (IgMlow, CD23ϩ)

Baff-rϩ/ϩ 63.3 Ϯ 7.5 3.9 Ϯ 0.6 16.2 Ϯ 4.0 4.4 Ϯ 2.4 4.3 Ϯ 1.7 4.4 Ϯ 1.6 Baff-rϩ/Ϫ 49 Ϯ 7.9 2.4 Ϯ 1.2 14 Ϯ 6 3.6 Ϯ 1.8 4.2 Ϯ 1.8 3.6 Ϯ 2 Baff-rϪ/Ϫ 1.4 Ϯ 0.3 0.05 Ϯ 0.04 5.5 Ϯ 4.6 2.8 Ϯ 2.9 0.9 Ϯ 0.8 0.5 Ϯ 0.3

a Values shown indicate the mean cell number (ϫ106) and the SD for three mice (Baff-rϩ/ϩ), six mice (Baff-rϩ/Ϫ), and ﬁve mice (Baff-rϪ/Ϫ).

subdivided into three distinct subsets (T1Ð3) according to the sur- specific IgG1 response was significantly reduced in Baff-rϪ/Ϫ com- face expression of AA4.1 (CD93, C1qRp), IgM, and CD23 (36). pared with control mice (Fig. 6). Forty-eight days after the primary This staining procedure offers the advantage that B1 cells can be immunization, we immunized all mice with a second dose of separated from the transitional subsets, which is not possible by NP-CG without adjuvant to evaluate the secondary Ab response. using the surface markers IgM and IgD or CD21 and CD23 for B Baff-rϪ/Ϫ mice mounted a secondary response of NP-specific IgM cell subpopulation analysis. Only transitional stage B cells express similar to that of control mice. However, we could not detect a AA4.1, whereas mature B cells do not. In spleens from Baff-rϪ/Ϫ secondary NP-specific IgG1 response in Baff-rϪ/Ϫ mice (Fig. 6). mice, the majority of B220ϩ cells are AA4.1ϩ transitional B cells Ϫ/Ϫ (Fig. 4A and Table I). Further analysis of the transitional stages GCs are formed in mLNs of Baff-r mice showed that in Baff-rϪ/Ϫ mice, the sizes of the T2 (IgMϩCD23ϩ) In mLNs, spontaneous GCs occur in response to gut-derived bac- and T3 (IgMlowCD23ϩ) populations were reduced, whereas the terial Ags (40, 41). Because GC formation can be initiated, but not proportion of T1 B cells (IgMϩCD23Ϫ) was increased (Fig. 4B maintained, in BaffϪ/Ϫ and A/WySnJ mice (13, 15), we analyzed and Table I). Absolute T1 B cell numbers were comparable to GC B formation in mLNs of Baff-rϪ/Ϫ and control mice by flow those of control mice (Table I). Recently, it was reported that the cytometry. This analysis revealed that B cells in the mLNs of surface expression of CD23 is regulated by BAFF (14). Therefore, Baff-rϪ/Ϫ mice can continuously differentiate into FashighPNAhigh the use of CD23 as a B cell maturation marker is problematic in GC B cells, presumably in response to microbial Ags derived from Baff-rϪ/Ϫ mice. However, analysis of IgM expression levels on the gut (Fig. 7A). Histological analysis of mLN confirmed the pres- AA4.1ϩ Baff-rϪ/Ϫ and control B cells confirmed a developmental ence of GC B cells expressing high levels of the transcription block in B cell development from IgMϩ cells (T1 ϩ T2) to IgMlow factor Bcl-6 in Baff-rϪ/Ϫ mice. cells (T3) (Fig. 4C). Analysis of CD23 expression on AA4.1Ϫ mature B cells revealed two populations, one with low CD23 ex- Overexpression of the antiapoptotic protein Bcl-2 rescues the Ϫ/Ϫ pression and one with CD23 expression comparable to control B development of mature B cells in Baff-r mice cells (Fig. 4D). This indicates that some B cells can express CD23 BAFF is considered to be a B cell survival factor (2). The fact that in the absence of BAFF-R signaling. MZ B cells are usually de- BaffϪ/Ϫ and Baff-rϪ/Ϫ mice display a similar reduction in B cell ϩ high low by guest on October 1, 2021. Copyright 2004 Pageant Media Ltd. fined as CD19 CD21 CD23 populations (37). However, be- numbers (Ref. 12 and this study) suggests that BAFF-R is indeed cause the expression of both CD21 and CD23 is regulated by the receptor through which BAFF exerts its antiapoptotic effects. Ϫ Ϫ BAFF (14), we assessed MZ B cell development in Baff-r / mice BAFF treatment was shown to increase the levels of the antiapop-

by using the cell surface markers CD1d and CD38, known to be totic proteins Bcl-xL, Bcl-2, and A1 in B cells (9, 42Ð45). We expressed at higher levels on MZ than on follicular B cells (37, addressed this issue by attempting to rescue B cell development in 38). This analysis demonstrated that MZ B cells are essentially BAFF-R-deficient mice by overexpression of the antiapoptotic Ϫ Ϫ absent in Baff-r / mice (Fig. 4E and Table I). We confirmed this protein Bcl-2. This was achieved through intercrossing Baff-rϪ/Ϫ finding by immunohistological staining of spleen sections with the mice with E␮-bcl-2 transgenic mice, which express Bcl-2 in B B cell marker CD19 and the metallophilic macrophage marker cells (32). FACS analysis revealed that overexpression of Bcl-2 ϩ ϩ Ϫ http://classic.jimmunol.org MOMA-1. This staining showed little, if any, B cells in the MZ restored the mature IgD IgMlow and B220 AA4.1 B cell com- Ϫ Ϫ area of spleen sections of Baff-r / in contrast to control mice partments in the compound mutants to a similar size as in control (Fig. 3B). Baff-rϩ/Ϫ/E␮-bcl-2 mice (Fig. 8, A and B). However, mature Baff- Ϫ/Ϫ ␮ Ϫ/Ϫ r /E -bcl-2 B cells expressed lower levels of CD21 than control Baff-r mice mount reduced Ag-specific Ab responses B cells (Fig. 8C), indicating that the regulation of CD21 by Measurement of Ig isotype levels in the blood of unimmunized BAFF-R is independent of its prosurvival function.

Downloaded from mice revealed decreased titers of IgM, IgG1, IgG2a, IgG2b, and IgG3 in Baff-rϪ/Ϫ compared with control mice (Fig. 5A). In con- Discussion trast, the serum concentration of IgA was similar in Baff-rϪ/Ϫ and The TNF family member BAFF, which plays a critical role in the control mice, most likely reflecting normal numbers of B1 cells in development of B2 and MZ B cells, binds to three receptors: Baff-rϪ/Ϫ mice (39). BCMA, TACI, and BAFF-R. Ablation of BAFF-R results in a very To determine whether Baff-rϪ/Ϫ mice can respond to TI and TD strong reduction of mature B2 cell numbers, leaving the B1 com- Ags, we immunized mice i.p. with NP-Ficoll, a TI type II Ag partment unaffected, essentially recapitulating the phenotype of (TI-II), or with NP-CG, a TD Ag. Monitoring of the TI-II humoral BaffϪ/Ϫ and A/WySnJ mice (11, 12, 23). This deficiency in mature immune responses at different time points after immunization with B cells is due to a survival defect caused by absence of BAFF-R, NP-Ficoll revealed that Baff-rϪ/Ϫ mice produced lower titers of because overexpression of Bcl-2 allows the development and NP-specific IgM and IgG3 than control mice (Fig. 5, B and C). maintenance of mature Baff-rϪ/Ϫ IgMlowIgDhigh B2 cells. We can- Determination of TD responses by measuring NP-specific Ab titers not exclude a mild effect of heterozygous loss of BAFF-R on B cell at different time points after immunization with saturating doses of development in Baff-rϩ/Ϫ mice (see Table I). However, this effect NP-CG showed that Baff-rϪ/Ϫ mice produced only slightly lower is clearly much less significant than the reduction of B cell num- levels of NP-specific IgM than control mice. However, the NP- bers by ϳ50% in offspring of A/WySnJ mice crossed to either A/J The Journal of Immunology 2249

FIGURE 3. Histological analysis of spleen and LN from wild-type, Baff-rϩ/Ϫ, and Baff-rϪ/Ϫ mice. A, Anti-CD19 (blue) and anti-CD3 (red)

by guest on October 1, 2021. Copyright 2004 Pageant Media Ltd. staining of spleen sections. The upper panels are shown in 100-fold magnification, and the lower panels in 200-fold magnification. B, Anti-CD19 (blue) and anti-MOMA-1 (red) staining of spleen sections. Magnification of the upper panels is ϫ100, and that of the lower panels is ϫ200. C, Anti-CD19 (blue) and anti-CD3 (red) staining of LN sections. Magnifica- tion is ϫ200.

(28) or BALB/c mice (24). Therefore, it seems that in a heterozygous situation, the A/WySnJ BAFF-R can act as a dominant- http://classic.jimmunol.org negative molecule. Our analysis of B cell development in Baff-rϪ/Ϫ mice revealed FIGURE 4. Flow cytometric analysis of B cell subpopulations in the a developmental block at the transitional stages, most likely at the spleen. A, Analysis of splenocytes for surface expression of AA4.1 and T1 to T2 transition. This corresponds well with the up-regulation B220. Transitional B cells are AA4.1ϩ, and mature B cells are AA4.1Ϫ. B, of BAFF-R expression from the T1 to the T2 transitional stage (9, Analysis of B220ϩAA4.1ϩ transitional B cells for surface expression of 46). It seems possible that BAFF-R engagement at the T1 transi- IgM and CD23; T1, IgMϩCD23Ϫ; T2, IgMϩCD23ϩ; T3, IgMlowCD23ϩ. ϩ ϩ Downloaded from tional stage of B cell development promotes further differentiation C, Histogram of IgM expression on B220 AA4.1 transitional B cells: T1, to the T2 stage (10). Additionally, binding of BAFF to BAFF-R T2, IgMϩ; T3, IgMlow. D, Histogram of CD23 expression on ϩ Ϫ enhances survival of T2, T3, follicular, and MZ B cell subsets. We B220 AA4.1 mature B cells. E, Analysis of surface expression of CD1d find Ͻ2 million mature B cells in the spleens of Baff-rϪ/Ϫ mice. (upper panel) and CD38 (lower panel) on splenic B cells. The numbers These cells express reduced levels of CD21 and, to a lesser extent, refer to the percentages of cells in individual gates of total live lymphocytes, or of cells in the gate indicated in brackets above the dot plots. CD23. Staining with Abs against CD1d and CD38 and immunohistochemistry revealed a near absence of MZ B cells, suggesting that the MZ B cell subset is exquisitely dependent on BAFF-R signals. The requirements for GC formation in mLNs and Peyer’s Monitoring B cell function in immune responses showed that patches in response to microbial Ags differ from those for GC Baff-rϪ/Ϫ B cells are able to produce class-switched Abs in re- formation in the spleen upon antigenic stimulation (47). Therefore, sponse to TI-II and TD Ags. The reduction of Ab titers in immu- we investigated the role of BAFF-R in this context and show that nized Baff-rϪ/Ϫ mice can be explained by the overall reduction of this receptor is dispensable for chronic GC responses triggered by B cell numbers in the case of TD and the deficiency in MZ B cells Ags derived from the gut microflora. The fact that GALT GC in case of TI-II Ab responses. reactions occur independently of BAFF-R and the intact B1 B cell 2250 BAFF-R SIGNALS FOR B CELL SURVIVAL

FIGURE 6. The humoral immune response to a TD Ag in wild-type (green circles, n ϭ 4), Baff-rϩ/Ϫ (blue circles, n ϭ 3), and Baff-rϪ/Ϫ (red by guest on October 1, 2021. Copyright 2004 Pageant Media Ltd. circles, n ϭ 3) mice. The serum titers of NP-speciﬁc IgM (A) and IgG1s (B) in primary and secondary responses after initial immunization with 100 ␮g and subsequent boost immunization with 10 ␮g of NP-CG are shown.

BAFF-R had no effect on the generation of B1 cells. These results suggest that the generation and maintenance of mature B2 and MZ B cells depend on the presence of at least two receptors, the BCR http://classic.jimmunol.org (48) and BAFF-R. Previous analyses of the role of BAFF signaling in B cell physiology in BaffϪ/Ϫ and A/WySnJ mice expressing a mutant BAFF-R yielded some differences. However, due to the complex nature of the A/WySnJ mutation, it is impossible to distinguish FIGURE 5. Serum Ig titers and the humoral immune response to a TI-II whether these differences are due to BAFF-R-independent signals Downloaded from Ag in wild-type (green circles), Baff-rϩ/Ϫ (blue circles), and Baff-rϪ/Ϫ (red elicited by BAFF or residual signaling through the mutant receptor circles) mice. The bars indicates geometric means. A, Titers of Ig isotypes in A/WySnJ B cells. By generating BAFF-R-deficient mice, we in the serum of unimmunized mice (n ϭ 5Ð8). B and C, Immune responses attempted to clarify these isssues. to NP-Ficoll (wild type, n ϭ 3; Baff-rϩ/Ϫ, n ϭ 4; Baff-rϪ/Ϫ, n ϭ 3). NP- Recently, it was suggested that expression of CD21 and CD23 specific IgM (B) and IgG3 (C) in sera of mice after immunization with 50 depends on BAFF (14). Residual B cells from A/WySnJ mice, in ␮g of NP-Ficoll are shown. contrast, have normal surface levels of CD23 and seem to express only slightly decreased levels of CD21 (26). Our results indicate that BAFF regulates CD21 and CD23 expression through BAFF- population in Baff-rϪ/Ϫ mice most likely both contribute to the R-dependent and -independent ways. normal IgA serum levels observed in these mice. In both BaffϪ/Ϫ (14) and A/WySnJ (26) mice, normal propor- Taken together, we show in this work that BAFF-R is essential tions of cells with a MZ phenotype were observed among the re- for the survival of peripheral B2 cells from the T2 transitional sidual splenic B cells. However, inhibition of BAFF function stage on. In addition, BAFF-R seems to be indispensable for the through transgenic expression of a TACI-Ig fusion protein pre- development of MZ B cells and plays a role in the regulation of vented MZ B cell development altogether (49). In accordance with CD21 and, to a lesser extent, CD23 surface expression. Ablation of this finding, our results suggest a very strong reduction of MZ B The Journal of Immunology 2251

why the latter mice mount normal Ag-specific IgM TI-II responses upon immunizations with TI-II Ags (28), whereas Baff-rϪ/Ϫ mice produce lower levels of Ag-specific IgM than control mice. In response to a TD Ag, Baff-rϪ/Ϫ mice mount a robust, only slightly reduced Ag-specific IgM response, whereas the TD IgG1 primary and secondary immune responses are significantly reduced. Sim- ilar results were previously obtained in A/WySnJ mice and corre- lated to the general B cell deficiency (28). However, despite the prominent role of BAFF-R in B cell survival, BAFF-R could have additional functions in the control of GC responses (13, 15) and class-switch recombination (53) and could contribute in this way to the deficiencies of TD Ab responses in Baff-rϪ/Ϫ mice. In contrast, BaffϪ/Ϫ mice, which have a similar defect in peripheral B cell numbers as Baff-rϪ/Ϫ and A/WySnJ mice, display a much more pronounced defect in humoral immunity (12). Be- cause TaciϪ/Ϫ mice also show defects in TI-II responses, despite FIGURE 7. Analysis of GC B cells in mLNs. A, Flow cytometric analysis of lymphocytes from mLNs for the surface expression of Fas and significantly elevated peripheral B cell numbers (20, 54), it seems PNA. The numbers refer to the percentages of CD19ϩ cells of total live plausible that BAFF promotes TI-II Ab responses also through lymphocytes. B, Histological analysis of GC in mLNs from wild-type, TACI. It has been reported that BAFF has costimulatory activity Baff-rϩ/Ϫ, and Baff-rϪ/Ϫ mice; anti-CD19 (blue) and anti-Bcl-6 (red) stain- for T cells (55, 56), which could account for a role of BAFF in TD ings of mLN sections are shown. Magnification is ϫ400. immune responses independent of any of the described receptors. Although the exact mechanisms still need to be uncovered, we show in this study that in the control of humoral immune responses Ϫ/Ϫ high cells in Baff-r mice. It is therefore possible that the IgM Ig- BAFF must operate at least partially independently of BAFF-R. Dlow fraction found in BaffϪ/Ϫ mice (14) is composed entirely of T1 B cells. We propose that BAFF-R signaling is essential for MZ B cell development. A/WySnJ mice, in contrast, appear to contain Acknowledgments normal proportions of CD21highCD23low MZ B cells. This sug- We are grateful to S. Willms for Ab determinations, and to A. Egert, V. Dreier, gests that the A/WySnJ mutation impairs BAFF-R-mediated sur- D. Ghitza, and S. Linehan for help with ES cell injections and mouse work. vival signals, but does not completely interfere with MZ B cell We thank K. Otibopy for critical reading of the manuscript, and M. L. Scott, S. Shulga-Morskaya, and M. Dobles for sharing unpublished results with us. differentiation. MZ B cells are thought to play an important role in TI-II immune responses (50Ð52). The stronger reduction in MZ B cell References numbers in Baff-rϪ/Ϫ compared with A/WySnJ mice could explain 1. Do, R. K., and S. Chen-Kiang. 2002. Mechanism of BLyS action in B cell immunity. Cytokine Growth Factor Rev. 13:19.

by guest on October 1, 2021. Copyright 2004 Pageant Media Ltd. 2. Mackay, F., P. Schneider, P. Rennert, and J. Browning. 2003. BAFF AND APRIL: a tutorial on B cell survival. Annu. Rev. Immunol. 21:231. 3. Melchers, F. 2003. Actions of BAFF in B cell maturation and its effects on the development of autoimmune disease. Ann. Rheum. Dis. 62 (Suppl. 2):ii25. 4. Rolink, A. G., and F. Melchers. 2002. BAFFled B cells survive and thrive: roles of BAFF in B-cell development. Curr. Opin. Immunol. 14:266. 5. Schneider, P., and J. Tschopp. 2003. BAFF and the regulation of B cell survival. Immunol. Lett. 88:57. 6. Allman, D., B. Srivastava, and R. C. Lindsley. 2004. Alternative routes to ma- turity: branch points and pathways for generating follicular and marginal zone B cells. Immunol. Rev. 197:147. 7. Cariappa, A., and S. Pillai. 2002. Antigen-dependent B-cell development. Curr. http://classic.jimmunol.org Opin. Immunol. 14:241. 8. Batten, M., J. Groom, T. G. Cachero, F. Qian, P. Schneider, J. Tschopp, J. L. Browning, and F. Mackay. 2000. BAFF mediates survival of peripheral immature B lymphocytes. J. Exp. Med. 192:1453. 9. Hsu, B. L., S. M. Harless, R. C. Lindsley, D. M. Hilbert, and M. P. Cancro. 2002. Cutting edge: BLyS enables survival of transitional and mature B cells through distinct mediators. J. Immunol. 168:5993. 10. Rolink, A. G., J. Tschopp, P. Schneider, and F. Melchers. 2002. BAFF is a

Downloaded from survival and maturation factor for mouse B cells. Eur. J. Immunol. 32:2004. 11. Gross, J. A., S. R. Dillon, S. Mudri, J. Johnston, A. Littau, R. Roque, M. Rixon, O. Schou, K. P. Foley, H. Haugen, et al. 2001. TACI-Ig neutralizes molecules critical for B cell development and autoimmune disease: impaired B cell maturation in mice lacking BLyS. Immunity 15:289. 12. Schiemann, B., J. L. Gommerman, K. Vora, T. G. Cachero, S. Shulga-Morskaya, M. Dobles, E. Frew, and M. L. Scott. 2001. An essential role for BAFF in the normal development of B cells through a BCMA-independent pathway. Science 293:2111. 13. Rahman, Z. S., S. P. Rao, S. L. Kalled, and T. Manser. 2003. Normal induction but attenuated progression of germinal center responses in BAFF and BAFF-R signaling-deﬁcient mice. J. Exp. Med. 198:1157. 14. Gorelik, L., A. H. Cutler, G. Thill, S. D. Miklasz, D. E. Shea, C. Ambrose, FIGURE 8. Flow cytometric analysis of B cell population in spleen of S. A. Bixler, L. Su, M. L. Scott, and S. L. Kalled. 2004. Cutting edge: BAFF Baff-rϪ/Ϫ/E␮-bcl-2 and control mice. A and B, Surface expression of IgM regulates CD21/35 and CD23 expression independent of its B cell survival func- and IgD (A), and of AA4.1 and B220 (B). C, Surface expression of CD21 tion. J. Immunol. 172:762. and CD23 on CD19ϩ splenocytes. The numbers refer to the percentages of 15. Vora, K. A., L. C. Wang, S. P. Rao, Z. Y. Liu, G. R. Majeau, A. H. Cutler, P. S. Hochman, M. L. Scott, and S. L. Kalled. 2003. Cutting edge: germinal cells in individual gates of total live lymphocytes, or of cells in the gate centers formed in the absence of B cell-activating factor belonging to the TNF indicated in brackets above the dot plots. family exhibit impaired maturation and function. J. Immunol. 171:547. 2252 BAFF-R SIGNALS FOR B CELL SURVIVAL

16. Castigli, E., S. Scott, F. Dedeoglu, P. Bryce, H. Jabara, A. K. Bhan, 37. Martin, F., and J. F. Kearney. 2002. Marginal-zone B cells. Nat. Rev. Immunol. E. Mizoguchi, and R. S. Geha. 2004. Impaired IgA class switching in APRIL- 2:323. deficient mice. Proc. Natl. Acad. Sci. USA 101:3903. 38. Oliver, A. M., F. Martin, G. L. Gartland, R. H. Carter, and J. F. Kearney. 1997. 17. Varfolomeev, E., F. Kischkel, F. Martin, D. Seshasayee, H. Wang, D. Lawrence, Marginal zone B cells exhibit unique activation, proliferative and immunoglob- C. Olsson, L. Tom, S. Erickson, D. French, et al. 2004. APRIL-deficient mice ulin secretory responses. Eur. J. Immunol. 27:2366. have normal immune system development. Mol. Cell. Biol. 24:997. 39. Berland, R., and H. H. Wortis. 2002. Origins and functions of B-1 cells with notes 18. Von Bulow, G. U., J. M. van Deursen, and R. J. Bram. 2001. Regulation of the on the role of CD5. Annu. Rev. Immunol. 20:253. T-independent humoral response by TACI. Immunity 14:573. 40. Cebra, J. J., A. George, and C. E. Schrader. 1991. A microculture containing TH2 19. Xu, S., and K. P. Lam. 2001. B-cell maturation protein, which binds the tumor and dendritic cells supports the production of IgA by clones from both primary necrosis factor family members BAFF and APRIL, is dispensable for humoral and IgA memory B cells and by single germinal center B cells from Peyer’s immune responses. Mol. Cell. Biol. 21:4067. patches. Immunol. Res. 10:389. 20. Yan, M., H. Wang, B. Chan, M. Roose-Girma, S. Erickson, T. Baker, D. Tumas, 41. Koni, P. A., and R. A. Flavell. 1999. Lymph node germinal centers form in the I. S. Grewal, and V. M. Dixit. 2001. Activation and accumulation of B cells in absence of follicular dendritic cell networks. J. Exp. Med. 189:855. TACI-deficient mice. Nat. Immunol. 2:638. 42. Claudio, E., K. Brown, S. Park, H. Wang, and U. Siebenlist. 2002. BAFF-induced 21. Lentz, V. M., M. P. Cancro, F. E. Nashold, and C. E. Hayes. 1996. Bcmd governs NEMO-independent processing of NF-␬B2 in maturing B cells. Nat. Immunol. recruitment of new B cells into the stable peripheral B cell pool in the A/WySnJ 3:958. mouse. J. Immunol. 157:598. 22. Lentz, V. M., C. E. Hayes, and M. P. Cancro. 1998. Bcmd decreases the life span 43. Do, R. K., E. Hatada, H. Lee, M. R. Tourigny, D. Hilbert, and S. Chen-Kiang. of B-2 but not B-1 cells in A/WySnJ mice. J. Immunol. 160:3743. 2000. Attenuation of apoptosis underlies B lymphocyte stimulator enhancement 23. Miller, D. J., and C. E. Hayes. 1991. Phenotypic and genetic characterization of of humoral immune response. J. Exp. Med. 192:953. a unique B lymphocyte deficiency in strain A/WySnJ mice. Eur. J. Immunol. 44. Hatada, E. N., R. K. Do, A. Orlofsky, H. C. Liou, M. Prystowsky, ␬ 21:1123. I. C. MacLennan, J. Caamano, and S. Chen-Kiang. 2003. NF- B1 p50 is required ␬ 24. Harless, S. M., V. M. Lentz, A. P. Sah, B. L. Hsu, K. Clise-Dwyer, D. M. Hilbert, for BLyS attenuation of apoptosis but dispensable for processing of NF- B2 p100 C. E. Hayes, and M. P. Cancro. 2001. Competition for BLyS-mediated signaling to p52 in quiescent mature B cells. J. Immunol. 171:761. through Bcmd/BR3 regulates peripheral B lymphocyte numbers. Curr. Biol. 45. He, B., A. Chadburn, E. Jou, E. J. Schattner, D. M. Knowles, and A. Cerutti. 11:1986. 2004. Lymphoma B cells evade apoptosis through the TNF family members 25. Yan, M., J. R. Brady, B. Chan, W. P. Lee, B. Hsu, S. Harless, M. Cancro, BAFF/BLyS and APRIL. J. Immunol. 172:3268. I. S. Grewal, and V. M. Dixit. 2001. Identification of a novel receptor for B 46. Gorelik, L., K. Gilbride, M. Dobles, S. L. Kalled, D. Zandman, and M. L. Scott. lymphocyte stimulator that is mutated in a mouse strain with severe B cell de- 2003. Normal B cell homeostasis requires B cell activation factor production by ficiency. Curr. Biol. 11:1547. radiation-resistant cells. J. Exp. Med. 198:937. 26. Amanna, I. J., J. P. Dingwall, and C. E. Hayes. 2003. Enforced Bcl-xL gene 47. Casola, S., K. L. Otipoby, M. Alimzhanov, S. Humme, N. Uyttersprot, expression restored splenic B lymphocyte development in BAFF-R mutant mice. J. L. Kutok, M. C. Carroll, and K. Rajewsky. 2004. B cell receptor signal strength J. Immunol. 170:4593. determines B cell fate. Nat. Immunol. 5:317. 27. Thompson, J. S., S. A. Bixler, F. Qian, K. Vora, M. L. Scott, T. G. Cachero, 48. Lam, K. P., R. Kuhn, and K. Rajewsky. 1997. In vivo ablation of surface im- C. Hession, P. Schneider, I. D. Sizing, C. Mullen, et al. 2001. BAFF-R, a newly munoglobulin on mature B cells by inducible gene targeting results in rapid cell identified TNF receptor that specifically interacts with BAFF. Science 293:2108. death. Cell 90:1073. 28. Miller, D. J., K. D. Hanson, J. A. Carman, and C. E. Hayes. 1992. A single 49. Tardivel, A., A. Tinel, S. Lens, Q. G. Steiner, E. Sauberli, A. Wilson, F. Mackay, autosomal gene defect severely limits IgG but not IgM responses in B lympho- A. G. Rolink, F. Beermann, J. Tschopp, and P. Schneider. 2004. The anti- cyte-deficient A/WySnJ mice. Eur. J. Immunol. 22:373. apoptotic factor Bcl-2 can functionally substitute for the B cell survival but not 29. Bunting, M., K. E. Bernstein, J. M. Greer, M. R. Capecchi, and K. R. Thomas. for the marginal zone B cell differentiation activity of BAFF. Eur. J. Immunol. 1999. Targeting genes for self-excision in the germ line. Genes Dev. 13:1524. 34:509. 30. Kontgen, F., G. Suss, C. Stewart, M. Steinmetz, and H. Bluethmann. 1993. Tar- 50. Balazs, M., F. Martin, T. Zhou, and J. Kearney. 2002. Blood dendritic cells geted disruption of the MHC class II Aa gene in C57BL/6 mice. Int. Immunol. interact with splenic marginal zone B cells to initiate T-independent immune 5:957. responses. Immunity 17:341. 31. Schmidt-Supprian, M., W. Bloch, G. Courtois, K. Addicks, A. Israel, 51. Guinamard, R., M. Okigaki, J. Schlessinger, and J. V. Ravetch. 2000. Absence of K. Rajewsky, and M. Pasparakis. 2000. NEMO/IKK ␥-deficient mice model in- marginal zone B cells in Pyk-2-deficient mice defines their role in the humoral continentia pigmenti. Mol. Cell 5:981. response. Nat. Immunol. 1:31. by guest on October 1, 2021. Copyright 2004 Pageant Media Ltd. 32. Strasser, A., A. W. Harris, and S. Cory. 1991. bcl-2 transgene inhibits T cell death and perturbs thymic self-censorship. Cell 67:889. 52. Martin, F., A. M. Oliver, and J. F. Kearney. 2001. Marginal zone and B1 B cells 33. Roes, J., and K. Rajewsky. 1993. Immunoglobulin D (IgD)-deficient mice reveal unite in the early response against T-independent blood-borne particulate anti- an auxiliary receptor function for IgD in antigen-mediated recruitment of B cells. gens. Immunity 14:617. J. Exp. Med. 177:45. 53. Litinskiy, M. B., B. Nardelli, D. M. Hilbert, B. He, A. Schaffer, P. Casali, and 34. Rodriguez, C. I., F. Buchholz, J. Galloway, R. Sequerra, J. Kasper, R. Ayala, A. Cerutti. 2002. DCs induce CD40-independent immunoglobulin class switch- A. F. Stewart, and S. M. Dymecki. 2000. High-efficiency deleter mice show that ing through BLyS and APRIL. Nat. Immunol. 3:822. FLPe is an alternative to Cre-loxP. Nat. Genet. 25:139. 54. Seshasayee, D., P. Valdez, M. Yan, V. M. Dixit, D. Tumas, and I. S. Grewal. 35. Schwenk, F., U. Baron, and K. Rajewsky. 1995. A cre-transgenic mouse strain 2003. Loss of TACI causes fatal lymphoproliferation and autoimmunity, estab- for the ubiquitous deletion of loxP-flanked gene segments including deletion in lishing TACI as an inhibitory BLyS receptor. Immunity 18:279. germ cells. Nucleic Acids Res. 23:5080. 55. Huard, B., L. Arlettaz, C. Ambrose, V. Kindler, D. Mauri, E. Roosnek, 36. Allman, D., R. C. Lindsley, W. DeMuth, K. Rudd, S. A. Shinton, and J. Tschopp, P. Schneider, and L. E. French. 2004. BAFF production by antigen-

http://classic.jimmunol.org R. R. Hardy. 2001. Resolution of three nonproliferative immature splenic B cell presenting cells provides T cell co-stimulation. Int. Immunol. 16:467. subsets reveals multiple selection points during peripheral B cell maturation. 56. Huard, B., P. Schneider, D. Mauri, J. Tschopp, and L. E. French. 2001. T cell J. Immunol. 167:6834. costimulation by the TNF ligand BAFF. J. Immunol. 167:6225. Downloaded from