IgMhighCD21high Enriched in the Splenic Marginal Zone Generate Effector Cells More Rapidly Than the Bulk of Follicular B Cells This information is current as of September 24, 2021. Alyce M. Oliver, Flavius Martin and John F. Kearney J Immunol 1999; 162:7198-7207; ; http://www.jimmunol.org/content/162/12/7198 Downloaded from

References This article cites 70 articles, 23 of which you can access for free at: http://www.jimmunol.org/content/162/12/7198.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 24, 2021

Subscription Information about subscribing to The Journal of is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. IgMhighCD21high Lymphocytes Enriched in the Splenic Marginal Zone Generate Effector Cells More Rapidly Than the Bulk of Follicular B Cells1

Alyce M. Oliver, Flavius Martin, and John F. Kearney2

Ag encounter will recruit Ag-specific cells from the pool of mature B lymphocytes in the and activate them to perform effector functions: generation of Ab-forming cells (plasma cells) and presentation of Ag to T cells. We have compared the ability of mature follicular and marginal zone cells to develop into effector B cells. The generation of marginal zone B cells and their localization in the marginal sinus area are and CD40 ligand independent, suggesting that they do not represent a postger- minal center population. Compared with mature recirculating follicular B cells, they express several characteristics of previous antigenic experience, including higher levels of B7.1 (CD80) and B7.2 (CD86) when freshly isolated and following in vitro stim- Downloaded from ulation. After a brief 6- to 8-h in vitro stimulation with LPS or anti-CD40 Abs, marginal zone B cells become potent APCs. In addition, their ability to proliferate and differentiate into plasma cells in response to low doses of T-independent polyclonal stimuli (LPS) is far greater than that of follicular B cells. These findings indicate a functional heterogeneity within splenic mature B lymphocytes, with marginal zone B cells having the capacity to generate effector cells in early stages of the immune response against particulate Ags scavenged efficiently in this special anatomical site. The Journal of Immunology, 1999, 162: 7198–7207. http://www.jimmunol.org/ he stages of B in adult bone marrow have sets. However, it is clear that the marginal zone of the spleen is been defined by the use of a large panel of cell surface and enriched in IgMhighIgDlowCD21highCD23low B cells, while the fol- T genetic markers, and as a result, many of the develop- licles contain mostly IgMlowIgDhighCD21intCD23high B cells (12, mental events resulting in the production of newly formed B cells 14–16). are well characterized (1–3). In contrast, the cellular and molecular It has been shown that memory B cells involved in T-dependent events involved in the transition of a newly formed, immature B antigenic responses are long lived, have receptors with increased cell into a long-lived, mature B populating peripheral antigenic affinity resulting from somatic mutations, and sometimes lymphoid organs are less well understood. Mature B lymphocytes switch isotypes (17–20). At least some of the long-lived memory are able to respond to antigenic challenge, and there is an abun- B cells generated in both T-dependent and T-independent re- by guest on September 24, 2021 dance of experimental evidence related to the cellular and molec- sponses reside in the splenic marginal zone for long periods of ular interactions of the immune response to T-dependent Ags, time after antigenic challenge (21, 22). Memory B cells, as de- leading to formation and subsequent production of scribed in humans, have a longer life span, a lower threshold of long term memory cells (4–7). activation, sometimes a higher affinity for Ag, and an increased In rodents and humans two categories of mature, long-lived B ability to present Ag to T cells (4). cells, termed memory and naive, have been defined that either have Although the phenotype of different long-lived subpopu- or have not encountered the nominal Ag under study (8–11). Phe- lations is well described, their functional capabilities in immune notypic and immunohistological analyses subdivided mature splenic B cells into follicular (FO)3 comprising 80–90% and mar- responses are less well understood (13, 23, 24). It has been shown ginal zone (MZ) accounting for 5–10% of adult mouse splenic B previously that there are differences in the ability of mature B cell high low high low high cells (12, 13). Because of complex intrasplenic trafficking path- subsets, defined as IgM CD23 CD21 and IgM CD23 int ways that are not fully understood, there is not a perfect correlation CD21 ,to respond in vitro to a variety of stimuli (12, 25, 26). between the phenotype and the topography of splenic B cell sub- Despite the fact that mature splenic B cells are constantly traffick- ing, and that the MZ of the spleen contains other cell types, in- cluding memory B cells, for the purpose of simplicity we refer to the two phenotypes as FO (IgMlowIgDhighCD21intCD23high) and Division of Developmental and Clinical Immunology, Department of Microbiology, high low high low University of Alabama, Birmingham, AL 35294 MZ (IgM IgD CD21 CD23 ) B cells (12). Received for publication December 14, 1998. Accepted for publication March MZ B cells are present in nude and thymectomized rats, sug- 24, 1999. gesting that they do not represent postgerminal center T-dependent The costs of publication of this article were defrayed in part by the payment of page B cells, although this assumption needs further confirmation (27, charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 28). Human splenic MZ lymphomas overlap significantly in im- munophenotype with the extranodal disease of mucosa-associated 1 This work was supported by National Institutes of Health Grants AI14782 and CA13148. lymphoid tissue generically referred as the postfollicular MZ B 2 Address correspondence and reprint requests to Dr. John F. Kearney, University of cell (29). Taken together with recent results from lymphotoxin Alabama, 378 Wallace Tumor Institute, 1824 6th Avenue South, Birmingham, AL gene-targeted mice in which both mucosa-associated lymphoid tis- 35294-3300. E-mail address: [email protected] sue and the MZ of the spleen are lacking, these observations fur- 3 Abbreviations used in this paper: FO, follicular B cell; MZ, marginal zone B cell; TG, transgenic; CMFDA, 5-chloromethylfluorescein diacetate; LM, littermate; BCR, ther suggest a link between mature long-lived B lymphocytes in B cell receptor for Ag; MFI, mean fluorescence intensity. these distinct anatomical sites (5, 30).

Copyright © 1999 by The American Association of Immunologists 0022-1767/99/$02.00 The Journal of Immunology 7199

Here we report that compared with their nontransgenic litter- 5-Chloromethylfluorescein diacetate (CMFDA) labeling of mates, certain Ig heavy chain transgenic (TG) mice have propor- B cells tionally larger MZ compartments that may constitute up to 50% of Sorted FO and MZ B cells were loaded with CellTracker Green CMFDA splenic B cells, and that the development of this compartment is (Molecular Probes, Eugene, OR) according to the manufacturer’s recom- indeed independent of T cells. We have used these TG mice as a mendations. In brief, sorted B cells were washed and resuspended at 107 source of large quantities of purified MZ B cells to facilitate the cells/ml in serum-free RPMI 1640 with 1.25 ␮M CMFDA and were in- analysis of their functional potential in early stages of in vitro cubated for 30 min at 37°C. Cells were washed and incubated for another 30 min at 37°C in serum-free RPMI 1640, washed again, and plated in activation. We show that MZ B cells respond better than FO B 96-well plates at 5 ϫ 105/ml in 200 ␮l of medium containing either LPS cells with respect to early in vitro activation events, resulting in (20 ␮g/ml) or anti-CD40 Ab (10 ␮g/ml). up-regulation of T cell costimulatory molecules and a concomitant ability to function as APCs. We also show that MZ B cells pro- Lymphocyte proliferation assays liferate much faster in response to very low doses of polyclonal Sorted B cells were cultured in triplicate at 1 ϫ 105 cells/well in RPMI mitogens and terminally differentiate into plasma cells within 1640 containing 10% FCS (HyClone Laboratories, Logan, UT) in 96-half- hours of activation. Collectively, these findings suggest that MZ B well flat-bottom plates (Costar, Cambridge, MA) in a total volume of 125 ␮l. LPS (20 ␮g/ml; Sigma, St. Louis, MO), polyclonal goat anti-mouse cells have unique signaling and subsequent differentiative poten- Ј ␮ F(ab )2 anti-IgM (20 g/ml; Southern Biotechnology Associates), poly- Ј ␮ tials that might permit them to react faster than the majority of clonal goat anti-rabbit F(ab )2 anti-IgM (20 g/ml), anti-CD40 (FGK45, 10 splenic B cells (FO) in the earliest stages of an in vivo immune ␮g/ml, a gift from Dr. Jan Anderson, Basel Institute Immunology, Basel, response. These functional differences apply to both Ag presenta- Switzerland), and IL-4 (100 U/ml) were added in different combinations to the plated cells. IL-4 was obtained from the supernatant of P3 ϫ tion and the production of short-lived IgM Ab-forming cells that 63Ag8.653 cells transfected with the appropriate expression constructs. Downloaded from are involved in the early defense against -borne pathogens. Cells were incubated for 48 h at 37°C in the presence of 10% CO2. Four hours before harvesting, 1 ␮Ci of [3H]thymidine was added to each well. Cells were harvested on Filter MATs (Skatron Instruments, Sterling, VA), and [3H]thymidine incorporation was measured in a scintillation counter Materials and Methods (Wallac, Gaithersburg, MD). Animals For the T cell alloreaction, sorted 81X TG C57BL/6 FO and MZ B cells

were activated for 8 or 24 h with LPS (20 ␮g/ml), anti-CD40 Ab (10 http://www.jimmunol.org/ Eight- to twelve-week-old C57BL/6 and BALB/c mice were purchased ␮g/ml), or both in culture medium at 37°C, washed twice, irradiated at from Charles River Laboratories (Raleigh, NC) or were bred in our mouse 2500 rad, and plated in 2-fold dilutions starting at 2.5 ϫ 105/well. Sorted facility. All other mice were used at 8–16 wk of age; 81X-C57BL/6 (31), BALB/c T cells (2 ϫ 105) were added to each well. The plate was pulsed 81X-JHko (32), MD2 (gift from Dr. C. C. Goodnow, Australian National 5 days later with [3H]thymidine and read on day 6. University, Canberra, Australia), MD4 and TCR T (TCR double knockout, C57BL6/J Tcrbtm1MomTcrdtm1Mom, purchased from The Jackson Labora- Cell culture tory, Bar Harbor, ME), and 81X-TCR T-C57BL/6 (generated by crossing 81X-C57BL/6 with TCR T-C57BL/6 and screening for the TCR double Cells were sorted as described above and were plated in 0.2–1 ml at a total knockout by PCR) were bred and housed in a pathogen-free facility in density of 5 ϫ 105/ml in flat-well (24, 48, or 96 wells, depending on accordance with institutional policies for animal care and usage. volume) tissue culture plates (Costar). Cells were cultured separately or in

a combination with anti-CD40 (10 ␮g/ml), LPS (20 ␮g/ml), anti-IgM (20 by guest on September 24, 2021 ␮g/ml), and IL-4 (100 U/ml) for 1–4 days. At the end of the incubation, the Flow cytometric analysis and sorting cells were isolated, washed, and stained with the appropriate Abs to de- termine changes in the expression of cell surface Ags or were used for Fluoresceinated (FITC) anti-CD5, anti-CD23, and anti-CD69 as well as cytospin preparations. biotinylated anti-CD62L, anti-CD25, anti-MHC class I, anti-MHC class II, anti-CD95, anti-syndecan-1, anti-B7.1, and anti-B7.2 mAbs were pur- Immunofluorescence analysis of tissues sections and cytospins chased from PharMingen (San Diego, CA). The anti-CD21 (7G6) Ab- embedded in OCT compound (Lab-Tek Products, Naperville, IL) secreting hybridoma was a gift from Dr. Michael Holers (University of were flash-frozen in liquid nitrogen. Frozen sections were cut, air-dried, Colorado Health Science Center, Denver, CO) (33, 34). The purified Ab fixed in ice-cold acetone, blocked with normal horse serum, and stained was conjugated to PE by Southern Biotechnology Associates (Birming- with MOMA-1 (rat, IgG2a ␬, a gift from Dr. Georg Kraal, specific for ham, AL). Two-, three-, and four-color surface stainings were performed as mouse splenic metallophilic macrophages) developed with goat anti-rat previously described (31). Briefly, 5 ϫ 105 cells depleted of RBC by lysis IgG-biotin and followed by a third layer of strepavidin-AMCA (Vector with an ammonium chloride-containing buffer were incubated with a mix- Laboratories, Burlingame, CA). The sections were washed and then stained ture of fluorescein-, PE-, and biotin-conjugated Abs followed by SA- with a mixture of FITC rat anti-mouse CD5 (clone 53-7.3, 10 ␮g/ml; PECY5 (Southern Biotechnology Associates) or streptavidin-allophyco- PharMingen) and rhodamine isothiocyanate goat anti-mouse IgM (2 ␮g/ cyanin (PharMingen). Cells were incubated for 15 min at each step and ml; Southern Biotechnology Associates). Cytospins were fixed in alcohol were washed with 2% FCS/PBS between steps. Cells were resuspended in for 30 min at Ϫ20°C and stained with rhodamine isothiocyanate goat anti- either 1% paraformaldehyde or 2% FCS/PBS with 2 ␮g/ml of propidium mouse IgM (2 ␮g/ml; Southern Biotechnology Associates). Sections and iodide to exclude dead cells. Data from stained cell samples were acquired cytospins were washed and mounted in Fluormount G (Southern Biotech- using FACSCalibur and CellQuest software (Becton Dickinson, Mountain nology Associates) and viewed with a Leica/Leitz DMRB fluorescence View, CA) and were analyzed with WinList 2.01 (Verity Software House, microscope equipped with appropriate filter cubes (Chromatechnology, La Jolla, CA) and WinMDI 2.0 ([email protected]) software programs. Battleboro, VT). Images were acquired with a C5810 series digital color FO and MZ cells were separated by cell sorting using anti-CD23-FITC camera (Hamamatsu Photonic System, Bridgewater, NJ). Images were pro- and anti-CD21-PE Abs as previously described (12). Briefly, single-cell cessed with Adobe Photo Shop and IP LAB Spectrum software (Signal suspensions were made from five or six mouse spleens. T cells were re- Analytics Software, Vienna, VA). moved from RBC-depleted spleen cells by treatment with anti-Thy1.2 and anti-CD4 Abs and rabbit complement (Cedarlane Laboratories, Hornby, Ig secretion by mature B cell subsets Canada). Viable cells were recovered by centrifugation over a Lympholyte M gradient (Cedarlane Laboratories) at 900 ϫ g and were stained with MZ and FO B cells were sorted based on their differential expression of anti-CD23-FITC and anti-CD21-PE as described above. MZ and FO B CD21 and CD23 as described above and were plated in 96-well flat-bottom cells were sorted based on their differential expression of CD21 and CD23 plates (Costar) at 1 ϫ 105 cells in 200 ␮l with 20 ␮g/ml LPS. Supernatants using a FACStar Plus (Becton Dickinson). When sorting T cells for the Ag were harvested at each of three time points (3, 7, or 10 days). Igs secreted presentation assays, BALB/c lymph nodes were used for the single-cell into the supernatant were detected by ELISA. Briefly, 96-well plates were suspension and were stained with B220-FITC and CD43-PE. To control for coated with purified goat anti-mouse IgM (2 ␮g/ml; Southern Biotechnol- a possible stimulating effect of CD21 ligation during the sorting process, ogy Associates) overnight at 4°C. The plates were washed and then some experiments were repeated using anti-CD38 (12, 35) with identical blocked with 1% BSA in PBS for1hatroom temperature. Supernatants results. were diluted and added to the plates along with titrated purified mouse IgM 7200 MARGINAL ZONE B CELLS AS EARLY EFFECTOR CELLS

FIGURE 1. Heavy chain TG mice have a large MZ population that develops independently of T cells. Spleen sections and FACS profiles of C57BL/6, 81X-C57BL/6, MD2-C57BL/6, 81X TCR T-C57BL/6 (TCR targeted)-C57BL/6, and TCR T-C57BL/6 are shown. Sections were stained for B cells ϫ

(anti-IgM, red), T cells (anti-CD5, green), and marginal sinus metallophilic macrophages (anti-MOMA-1, blue; original magnification, 100). Note the lack Downloaded from of CD5-positive T cells in the TCR T mice. Flow cytometry was performed on spleen suspensions stained with anti-CD23 and anti-CD21 Abs. FACS profiles were obtained by gating on lymphoid cells, and the percentages of FO and MZ cells were calculated from the lymphoid gate. Profiles are representative of at least five adult mice of each genotype. isotype control (Southern Biotechnology Associates) to establish a stan- TG mice, in 81X mice crossed to the TCR␤, ␦ double knockout, dard concentration curve. The plates were incubated for4hat37°C and the percentage of MZ B cells was also about 3-fold higher than that http://www.jimmunol.org/ then washed, and goat anti-mouse IgM conjugated to alkaline phosphatase in the LM controls (Fig. 1). In a parallel experiment CD40L-de- (1/500; Southern Biotechnology Associates) was added and incubated for ficient mice bred with 81X TG mice had a MZ B-cell compartment 4 h. The ELISAs were developed by adding alkaline phosphatase substrate ϩ (1 mg/ml; Sigma). The ODs of wells were measured using a TiterTek similar in size to that in the 81X CD40L controls (data not Multiskan Plus (Labsystems, Helsinki, Finland) at 405 nm, and data were shown), again suggesting that the establishment and maintenance analyzed using ELISA LITE software (Meddata, New York, NY). of MZ B cells are independent of T cells and also germinal centers. However, T cells, when present, might still participate in the dif- Results ferentiation and overall function of MZ B cells. Generation of CD21highCD23low B cells (MZ) is T cell

independent by guest on September 24, 2021 We have used the differential expression of the CD23 and CD21 TG and LM MZ B cells have similar functional properties surface Ags by MZ and FO B cells to identify these mature subsets The increased percentage of MZ B cells in Ig heavy chain TG in the mouse spleen. MZ B cells express high levels of CD21 and mouse models compared with normal mice permits the isolation of low levels of CD23, while FO B cells express CD23 and moderate greater numbers of these cells and facilitates subsequent functional levels of CD21 (12, 13, 36). In adult C57BL/6 mice, MZ B cells analyses. However, before using these mice as a source of MZ B comprise approximately 5% of the lymphocytes (Fig. 1). However, cells, we first determined whether their ability to respond differ- in heavy chain VH81X TG (81X) mice (37), the MZ is almost 3 ently from FO B cells in vitro (12) was preserved. As shown in times higher proportionally than that in littermate (LM) control Fig. 2A, TG MZ B cells isolated from mice that either have (81X) mice (81X TG, 15.1 Ϯ 1.3%; LM, 6 Ϯ 0.9%; n ϭ 5). This in- or do not have T cells (81X TCR T) share functional characteris- creased percentage of MZ cells is relative and is due to an overall tics similar to those reported previously for MZ B cells from nor- decrease in the absolute numbers of FO B cells (3.4 Ϯ 0.25 vs mal BALB/c mice (12) responding more vigorously to CD40 10.1 Ϯ 1.5 ϫ 106/spleen, respectively, in TG and LM; n ϭ 5). In (40,471 Ϯ 4,325 cpm) and to LPS stimulation (228,959 Ϯ 2,540 contrast, the absolute number of MZ B cells is remarkably con- cpm) compared with FO B cells (7,225 Ϯ 640, 138,416 Ϯ 1,765 served between TG and LM (1.85 Ϯ 0.1 and 1.9 Ϯ 0.1 ϫ 106/ cpm, respectively). Similar to normal mice, TG FO B cells spleen; n ϭ 5). Topographic evidence of smaller follicles and (47,359 Ϯ 5,140 cpm) proliferate better in response to BCR cross- larger MZ is observed in spleen sections stained to display the MZ linking (Fig. 2A) than do MZ B cells (391 Ϯ 130 cpm), which are B cells separated from the FO B cell population by MOMA-1- very sensitive to BCR ligation and die by apoptosis (12). Due to positive macrophages (Fig. 1). In another heavy chain TG line, the the fact that a similar increase in the percentage of MZ B cells is anti-HEL heavy chain MD2 (38) MZ are also proportionally in- observed when 81X TG is crossed onto a JH knockout background creased (15.3 Ϯ 1.1% in TG vs 7.5 Ϯ 0.4% in LM; n ϭ 4), al- (39) (data not shown), cells derived from either 81X-C57BL/6 or though the overall B cell compartment is reduced due to a decrease 81X-JH T mice were used for functional experiments and gave in the absolute numbers of FO B cells (2.36 Ϯ 0.6 ϫ 106/spleen in similar results. TG and 6.95 Ϯ 0.5 ϫ 106/spleen in LM; n ϭ 4). Again, the ab- By following the decrease in fluorescence of CMFDA-labeled solute numbers of splenic MZ B cells were similar in TG and LM purified B cells with the number of cell cycles, we observed that MD2 adult mice (1.5 Ϯ 0.14 ϫ 106/spleen in TG and 1.33 Ϯ 0.1 ϫ MZ cells begin to divide more quickly when stimulated by CD40 106/spleen in LM; n ϭ 4; Fig. 1). The establishment and mainte- ligation or LPS (Fig. 2B). The relative decrease in CMFDA levels nance of this MZ B-cell population do not appear to depend on the also indicated that MZ divide more than FO B cells within the first presence of T cells, because mice lacking the ␤- and ␦-chains of 48 h of culture with anti-CD40 Abs (1.5-fold) or LPS (2- to the TCR have nearly equivalent numbers of MZ B cells compared 3-fold). By 4 days, the difference is even more distinct for anti- with control mice (TCR T in Fig. 1). Similar to the original 81X CD40 stimulation, while for LPS the MZ cultures are overgrown, The Journal of Immunology 7201

FIGURE 2. TG MZ B cells display functional characteristics similar to Downloaded from those of MZ B cells from normal mice. A, Proliferation assays of TG MZ and FO B cells from 81X TG and 81X TCR knockout mice with anti-CD40 (10 ␮g/ml), LPS (20 ␮g/ml), and anti-IgM (20 ␮g/ml). B, MZ cells rapidly enter into cell cycle after LPS and CD40 stimulation. MZ and FO B cells from 81X TG mice were loaded with CMFDA (see Materials and Meth- ods) and stimulated in culture with LPS or anti-CD40. The relative de- crease in CMFDA levels indicates that MZ divide more than FO B cells http://www.jimmunol.org/ within the first 48 h of culture with anti-CD40 Abs (1.5-fold) or LPS (2- to 3-fold). By 4 days, the difference is more distinct for anti-CD40 stimula- tion, while for LPS the MZ cultures are overgrown, but surviving MZ and FO cells have gone through similar numbers of cycles. A representative experiment of three is shown. FIGURE 3. MZ B cells express higher levels of T cell cosignaling mol- ecules that are up-regulated to higher levels after in vitro stimulation for but surviving MZ and FO cells have gone through similar numbers 24 h. A, Splenocytes from 81X TG mice were stained for CD21 and CD23, of cycles. and basal levels of B7.1 or B7.2 are shown for MZ and FO populations. B by guest on September 24, 2021 and C, Sorted FO and MZ B cells from 81X TG mice were cultured with MZ B cells are not activated but can be quickly stimulated to anti-IgM (20 ␮g/ml), IL-4 (100 U/ml), LPS (20 ␮g/ml), and anti-CD40 (10 up-regulate T cell costimulatory molecules ␮g/ml) for 24 h and stained for B7.1 and B7.2 (B) or for CD25 and CD69 (C) expression. The vertical lines indicate the MFI of each Ag expressed on Since MZ B cells divide vigorously in response to stimuli, it was FO cells. One representative experiment of 10 is shown. of interest to determine whether these cells constitutively express surface Ags that would indicate their readiness to collaborate with T cells and engage in an immune response. MZ cells are not ac- or anti-CD40 Abs. The expression of other activation Ags was also tivated per se, since they do not express increased levels of typical analyzed after stimulation with either anti-CD40 or LPS, and sim- activation markers, including CD69, CD25, MHC class I, MHC ilar up-regulation of CD25 and CD69 was observed on both MZ class II, and CD95, compared with FO B cells (data not shown) and FO B cells, although levels of CD25 on MZ B cells were (40, 41). However, these cells do express surface Ags character- slightly higher after LPS stimulation (Fig. 3C). istic of a memory phenotype indicative of a history of prior Ag We also determined the kinetics of B7.1 and B7.2 expression on encounter. In contrast to FO B cells, MZ B cells express decreased sort-purified FO and MZ B cells cultured with anti-CD40 (10 ␮g/ levels of CD62L and IgD, but have increased levels of CD44 (data ml), anti-IgM (20 ␮g/ml), or LPS (20 ␮g/ml) or alone. At inter- not shown) (41). Moreover, the basal level of the costimulatory vals, cells were isolated and stained to determine the relative mean molecules B7.1 and B7.2 on MZ B cells is higher than that on FO fluorescence intensity (MFI) of B7.1 or B7.2 (Fig. 4A). Anti-IgM, B cells (Fig. 3A). These data further support the hypothesis that anti-CD40, and LPS are all able to up-regulate B7.2 to comparable MZ B cells have been previously primed, and together with their levels, although anti-IgM is slightly more potent. As previously ability to rapidly enter the cell cycle, suggest that these B cells are reported, expression of B7.2 increases in the first 8 h after activa- potential candidates to provide T cell costimulation. tion and reaches peak levels at 12–24 h, declining by 48 h (42, 43). Higher constitutive levels of B7.1 and B7.2 on MZ B cells sug- Both MZ and FO populations exhibit similar kinetics of B7.2 ex- gested that MZ B cells might further increase the surface expres- pression, but at all times, MZ cells respond more vigorously to sion of these molecules after stimulation. Indeed, there was up- stimulation with anti-IgM Abs (MFI, 828 vs 329), anti-CD40 Abs regulated expression of B7.2, but not B7.1, in sorted purified cells (MFI, 503 vs 160), or LPS (MFI, 473 vs 169). In contrast to B7.2, activated with anti-␮ Abs for 24 h, an increase that was higher for in the first 48 h of stimulation, B7.1 is up-regulated quickly (peak- MZ than for FO B cells (Fig. 3B). Stimulation with IL-4, LPS, or ing at 12 h) on MZ B cells by anti-CD40 and LPS and only mar- anti-CD40 also increased the expression of B7.2 on both FO and ginally on FO cells. Thus, MZ B cells quickly express more B7.1 MZ B cells, with the MZ B cells always expressing higher levels. and B7.2 molecules on their cell surface after acute stimulation. In addition, B7.1 expression was amplified on MZ B cells to a When combinations of LPS and anti-CD40 or LPS and IL-4 greater extent than on FO B cells when stimulated with either LPS were added to FO and MZ B cells for 24 h, B7.2 expression was 7202 MARGINAL ZONE B CELLS AS EARLY EFFECTOR CELLS

FIGURE 4. A, Zero to 48-h kinetics of B7.1 and B7.2 up-regulation on MZ and FO B cells. Sorted MZ and FO cells were stimulated as follows and were stained for B7 expression at the indicated time points: ‚, control FO; Œ, control MZ; Ⅺ, anti-CD40 FO; f, anti-CD40 MZ; छ, LPS FO; ࡗ, LPS MZ; E, anti- IgM-FO; F, anti-IgM-MZ. Results represent one of three similar experiments. B, Combinations of mito- gens are even more potent in up-regulating B7 mol- ecules. Sorted FO and MZ B cells from 81X TG mice were stimulated with LPS (thin line), anti-CD40, or IL-4 (normal line) or with a combination of either LPS plus anti-CD40 or LPS plus IL-4 (dotted line), and their B7 expression was assayed at 24 h by flow Downloaded from cytometry. Results represent one of three similar ex- periments. C, Homogeneous populations of MZ and FO B cells with the same specificity (anti-HEL) have different functional properties. Sorted FO and MZ B cells from MD4 anti-HEL TG mice were cultured with LPS (20 ␮g/ml) and stained for B7 at 24 and 72h(left) or were assayed for proliferation with anti- http://www.jimmunol.org/ CD40 (10 ␮g/ml) and anti-IgM (20 ␮g/ml; middle)or LPS (20 ␮g/ml; right). Results represent one of three similar experiments. by guest on September 24, 2021

significantly increased on FO, but not MZ, cells over that observed when stimulated with anti-CD40 Abs, while FO (12,380 Ϯ 870 following addition of the individual stimuli (Fig. 4B). By contrast, cpm) out-proliferate MZ B cells (1,622 Ϯ 280 cpm) when treated B7.1 expression was further up-regulated by the combined stimuli with anti-IgM Abs. The ability to proliferate in response to LPS on MZ cells. The dramatic increase in B7.1 on FO cells after LPS still follows the same pattern, with MZ cells cycling better than FO plus anti-CD40 treatment is of interest, since both agents showed B cells, although the differences between these two subsets are a minimal effect when added alone. All these data predict a com- smaller than those seen with B cells from normal or heavy chain plex picture in which the B cell subsets’ ability to provide T cell TG mice. These data show that independently generated B cell costimuli will depend on the nature, timing, dose, and combination clones with identical receptors are heterogeneous with respect to of signals that the B cell received. different stimuli (anti-IgM, anti-CD40) depending on their FO or MZ phenotypes. This functional heterogeneity indicates that un- Monoclonal and polyclonal MZ and FO B cells have similar known stochastic selective events influence the fate of a B cell functional characteristics depending on whether it is present in a certain microenvironment All previous experiments used a polyclonal population of B cells and not another at a particular time point in its life span. from normal or Ig heavy chain TG mice. Since it was possible that the results reported in the above experiments were influenced by MZ B cells can function as potent APCs certain B cell clones in the mixture, we determined whether a mono- Since MZ B cells express high basal levels of B7.1 and B7.2, clonal population of B cells with either FO (CD21intCD23high) which are rapidly up-regulated within 8–24 h, we next analyzed or MZ (CD21highCD23low) phenotype from the anti-HEL TG mice the capacity of FO and MZ B cells to present Ag and stimulate MD4 had similar characteristics (44, 45). As shown in Fig. 4C, alloreactive T cells. Purified FO and MZ B cells, when stimulated B7.1 and B7.2 expression is higher on MZ than on FO B cells 24 h with either LPS or anti-CD40 Abs for 8 h, increased their levels of after incubation with LPS. However, the differences between MZ B7.1, B7.2, MHC class I, and MHC class II molecules (Fig. 4A and and FO B7 up-regulation were slightly smaller than those observed data not shown). Stimulated or unstimulated B cells from 81X with a polyclonal B cell population and faded by 72 h, indicating heavy chain TG C57BL/6 mice were irradiated and mixed with that individual clones might differ in their ability to provide B7 sort purified BALB/c T cells. As shown in Fig. 5A, purified un- cosignals (Figs. 3B and 4C). Also, similar to TG mice in which B stimulated B cells were unable to induce T cell proliferation. In cells contain only a common heavy chain, MD4 MZ (14,754 Ϯ contrast, LPS- or anti-CD40-stimulated MZ, but not FO, B cells 2,760 cpm) proliferate better than FO B cells (386 Ϯ 190 cpm) induced vigorous T cell proliferation. T cells stimulated with 2.2 ϫ The Journal of Immunology 7203

FIGURE 5. Activated MZ B cells function as APCs for alloreactive T cells. Sorted FO and MZ B cells from 81X C57Bl/6 mice were treated in vitro for 8 or 24 h with LPS (20 ␮g/ml), anti-CD40 (10 ␮g/ml; A and B), or a mixture of the two (B); washed; irradiated; and added in 2-fold dilutions to- gether with sorted BALB/c T cells. MZ present alloantigen better than FO cells after short stimulation with only one of the stimuli, while FO Downloaded from cells become effective APCs only after treatment with a combination of both agents. One of five sim- ilar experiments is shown. http://www.jimmunol.org/ by guest on September 24, 2021

105 LPS-activated MZ B cells proliferated much more (132,486 Ϯ proliferated about twice as much as FO B cells (297,799 Ϯ 7,240 9,915 cpm) than T cells incubated with LPS-activated FO B cells vs 154,180 Ϯ 4,650 cpm) when stimulated with 20 ␮g/ml LPS. (31,068 Ϯ 4,390 cpm). Similarly, T cells stimulated with 1.1 ϫ Remarkably, a similar degree of MZ B cell proliferation was seen 105 MZ B cells that had been treated with anti-CD40 Abs even with a 10-fold lower LPS concentration (297,799 Ϯ 7,240 (55,024 Ϯ 5,870 cpm) proliferated better than T cells stimulated cpm at 20 ␮g/ml vs 292,379 Ϯ 34,010 cpm at 2 ␮g/ml LPS), while with their FO counterparts (3,849 Ϯ 290 cpm). Thus, MZ, but not FO B cells showed very little response at this dose (Fig. 6A). Even FO, B cells when activated by LPS or anti-CD40 for a brief time at 0.2 ␮g/ml LPS stimulation, MZ B cells proliferated (28,053 Ϯ can up-regulate key costimulatory molecules on the cell surface 1,450 cpm), while FO B cells were totally unresponsive. A kinetic that help to induce allogenic T cell proliferation. To determine study of LPS (20 ␮g/ml)-induced cell division showed a faster whether combinations of stimuli and a longer activation time will peak response for MZ cells (48 h) that had decreased by 72 h enable FO to present Ags as well as MZ cells, a similar experiment compared with that for FO cells (Fig. 6B). was performed using a mixture of LPS and anti-CD40 to stimulate To address whether this sensitivity to LPS extended to their the sorted B lymphocyte populations for 8 or 24 h. Although at 8 h ability to differentiate into mature Ab-forming cells, sorted FO and the LPS-stimulated or LPS- and anti-CD40-stimulated MZ cells MZ B cells were stimulated with either LPS or anti-CD40 and presented alloantigen much better than FO cells, by 24 h LPS- and assessed for their expression of syndecan-1, a surface Ag that is anti-CD40-activated FO B cells induced T cell proliferation com- up-regulated on plasmablasts (46). In Fig. 6C it can be seen that parable to that in similarly treated MZ cells (Fig. 5B). This corre- some MZ B cells begin to express syndecan within 24 h after LPS lated with the equivalent level of B7 expression achieved on MZ treatment, while FO B cells are still negative. By day 3 both MZ and FO cells after stimulation with LPS and anti-CD40 in combi- and FO B cells have subpopulations expressing syndecan, but nation but not with each stimulus alone (Fig. 4B). compared with FO, MZ has a greater number of very bright syn- decan cells. Similar to human tonsillar memory B cells (47), sorted MZ B cells are highly sensitive to LPS stimulation MZ and FO B cells treated with anti-CD40 do not express synde- Although we have clearly shown that MZ B cells express activa- can-1 (data not shown). In addition to syndecan expression, MZ B tion molecules that would potentially enable them to participate in cells enlarge very rapidly to form blasts within 24 h, and by day 3 B-T cell interactions, we also studied their involvement in T-in- mature to plasma cells containing high levels of cytoplasmic Ig dependent responses, which is another function potentially asso- compared with FO B cells-derived blasts (Fig. 6D). Furthermore, ciated with these B cells. Although MZ B cells overall are better MZ B cells from both 81X and MD4 TG mice secrete significantly LPS responders than FO B cells, we next investigated this effect at greater amounts of IgM in a shorter period of time than FO B cells limiting mitogen concentrations. As already shown, MZ B cells after in vitro LPS stimulation (Table I). These data show that MZ 7204 MARGINAL ZONE B CELLS AS EARLY EFFECTOR CELLS

FIGURE 6. T-independent polyclonal activators are much more potent for MZ than for FO B cells. FO and MZ sorted B cells from 81X TG mice were assayed for proliferation in response to LPS (decreas- Downloaded from ing doses: 20, 2, and 0.2 ␮g/ml (A); in- creasing time periods: 24, 48, and 72 h (B)) and for up-regulation of syndecan-1 after LPS (20 ␮g/ml) stimulation (C). Note that cells expressing very high levels of syndecan-1 are much more frequent in

the MZ than in the FO population on day http://www.jimmunol.org/ 3. Increased cell size and accumulation of cytoplasmic IgM are observed after LPS stimulation (20 ␮g/ml; original magnifica- tion, ϫ250; D). One of five similar exper- iments is shown. by guest on September 24, 2021

B cells differentiate into mature plasma cells and secrete large (Figs. 1 and 2) (28). They have the ability to perform effector quantities of Abs in response to LPS stimulation. functions after brief periods of in vitro activation and either be- come potent APCs for T cells through up-regulation and expres- Discussion sion of B7 costimulatory molecules (Figs. 3–5) or differentiate into The findings reported in this paper establish a clear functional plasma cells in response to very low doses of polyclonal mitogens heterogeneity in the compartment of long-lived splenic B lympho- (Fig. 6). These findings raise interesting questions about the role of cytes. CD21hiIgMhi B cells enriched in the MZ are not postger- these B cells and their functionality in the course of in vivo im- minal center B cells, since they develop independently of T cells mune responses. It is well documented that besides long-lived B cells, the MZ of the spleen contains several other cell types potentially involved in Table I. LPS-stimulated marginal zone B cells secrete large amounts of the early stages of immune responses. MZ and metallophilic mac- IgM earlier than follicular transgenic B cells rophages are able to scavenge Ags from the blood, and dendritic

Mice Days in Culture FO (␮g/ml) MZ (␮g/ml) cells transport them to the T cell zones in the first few hours after antigenic exposure (48–51) (F. Martin and J. F. Kearney, unpub- 81X JH T 3 0.23 Ϯ 0.03 15.19 Ϯ 4.2 lished observations). MZ B cells also migrate from their normal 7 45.06 Ϯ 15.2 97.19 Ϯ 17.35 10 100.49 Ϯ 34.97 192.62 Ϯ 71.2 sites very early after in vivo immunization, with some of them traveling to the T cell zone and subsequently to the B cell follicles MD4 3 0.033 Ϯ 0.002 3.12 Ϯ 0.38 The Journal of Immunology 7205

extra in vitro activation, while mouse MZ B cells need brief pe- riods of CD40 or LPS activation (57). Although no true memory appears to be involved in T-independent responses, postimmune Ag-specific cells were characterized in rats (21, 58), and since these cells do not depend on germinal center formation, they ex- hibit even greater similarity to mouse MZ B cells. The ability of B cells to up-regulate both B7.1 and B7.2 co- stimulatory molecules has been associated with polyclonal mito- gens (LPS), various cytokines, anti-CD40, and, recently, anti- CD21 and anti-CD19 cross-linking signals (59–61). In contrast, anti-BCR signals, either through Ag or anti-IgM Abs, preferen- tially up-regulate B7.2 (42). The differences in these events be- FIGURE 7. Working model showing the proposed differential temporal tween MZ and FO B cells is not simply one of kinetics, since FO ability of FO and MZ B cells to become effector cells that can present Ags cells never reach the levels of B7 expressed by MZ B cells (Fig. 4). and differentiate into plasma cells. The black arrowhead represents the These data suggest that there are intrinsic differences within these encounter of B cells with the appropriate stimulus. The question marks populations regarding B7 expression following activation, and that represent possible developmental pathways that are not as yet clear. Al- the resulting activated MZ and FO effector B cells differ in their though MZ B cells seem to have the potential to collaborate early in the ability to use these pathways of costimulation. When these B cell response with T cells, it is not known whether they are able to participate subpopulations are stimulated with a mixture of two different re- Downloaded from in the germinal center reaction. agents (Fig. 4B, LPS plus CD40 or IL-4), for some combinations there is a dramatic increase in B7 molecules compared with the use of each stimulus alone. As a result, the levels of B7.2 expression (15). Our results predict that in the first hour of an immune re- on FO B cells become comparable to those on MZ B cells, which sponse, MZ B cells generate activated effector B cells that can require only one of the stimuli alone to be maximally stimulated.

either present Ags to T cells or differentiate into plasma cells for an The attainment of similar levels of expression of B7.2 by both http://www.jimmunol.org/ early boost of IgM Ab production. The rapid secretion of these Abs populations is reflected by their equivalent ability to present al- would function locally to facilitate trapping of Ag and formation of loantigens. These findings predict that depending on the type of in immune complexes that enhance Ag processing and presentation vivo Ag stimulation, MZ or both subsets will participate in Ag to T cells by APCs or to B cells by follicular dendritic cells (52– presentation. These results also suggest that there is a qualitative 54). Although similar effector functions might also be performed at difference in the regulation of B7.2 and subsequently of the co- a slower pace by the bulk of follicular B cells (Fig. 7), the majority stimulatory activity between FO and MZ B cell subsets. Recently, of the Ags, at least for those of a particulate nature, is probably constitutive expression of the B7.2 molecule has been shown to scavenged and processed in the marginal zone and reaches the B break the inability of tolerant FO B cells to become activated (62). cell follicles mostly on the surface of follicular dendritic cells. The higher levels of B7 constitutive expression and its rapid up- by guest on September 24, 2021 Functional distinctions have been made previously between ma- regulation on MZ B cells raise questions about mechanisms that ture splenic B cells based on the low and high expression of CD24 control these B cells to avoid autoimmunity and suggest that their (heat stable Ag). In these in vivo experiments, the CD24high B cells repertoire selection and activation are stringently controlled in a were shown to contain precursors for plasma cells, while the healthy organism. CD24low B cells were able to initiate potent germinal center reac- In humans, chronic activation is considered one of the potential tions, leading to the hypothesis that there are different B cell pre- steps involved in the induction of lymphomas; in particular for MZ cursors involved in primary and secondary antigenic encounters lymphomas, molecular studies support the concept that they are (55, 56). It is very likely that the CD24high cells contained a mix- derived from a postfollicular, postgerminal center or a memory B ture of short-lived newly formed, and mature long-lived MZ B cell (29). In NFS.Vϩ mice, used as a model of lymphomagenesis, cells (12, 41), while the CD24low cells are basically follicular B Ͼ40% of the spontaneous developing lymphomas derive from the lymphocytes. The events of in vitro B cell activation that we de- MZ and, surprisingly, express low levels of CD5 and B220, a scribe here are also mirrored in vivo (F. Martin, A. M. Oliver, and characteristic of the B1 subset of mouse B lymphocytes (63). In J. F. Kearney, manuscript in preparation). The early difference in normal mice, splenic B1 cells are infrequent (13) and have not generating effector cells associated with FO and MZ would not been assayed functionally, but the similarities between peritoneal have been observed in the previously described experiments using B1 cells and MZ B cells (lack of proliferation to anti-IgM, larger CD24 as a marker, since the immune response was assayed at a size, ability to rapidly secrete Abs) suggest some kind of link, if later time point. In addition, since the CD24high population con- not in their developmental origins at least in function. However, tains a mixture of at least two B cell subsets, the distinctions we the selection and developmental expansion of B1 cells, in addition observed between MZ and FO subpopulations would not have to being dependent on costimulation, are probably also linked to been as dramatic. Nevertheless, it implies that MZ B cells partic- the specificity of their receptors (64–67). Since B1 cells are de- ipate poorly in the germinal center reaction (Fig. 7), and it will be pleted in several Ig TG mouse models that have normal or in- interesting to test this hypothesis in vivo in an Ag-specific immune creased MZ populations (MD4, 81X), these B cells might not have response. receptors with permissive specificities that select them into the B1 MZ B cells share some similar functional properties with Ag- compartment (16, 31). specific memory B cells that have been previously described in Recently, several groups have reported that the level of expres- mice, rats, and humans (4, 21, 57). They are long lived, localize in sion of CD1 on MZ B cells is high and comparable to that found the same compartment of the spleen, and have lower activation on dendritic cells, but there are no data yet on the functional sig- thresholds for specific Ag in the case of memory B cells (4) or in nificance of this finding (68, 69). NK T cells, which respond to Ags response to costimulation (CD40 or LPS) for MZ B cells. Human presented by CD1 molecules, are thought to be early participants memory cells from tonsils are able to present Ags without any in an immune response. The evidence we have presented indicates 7206 MARGINAL ZONE B CELLS AS EARLY EFFECTOR CELLS that MZ B cells show phenotypic and functional characteristics 23. Hayakawa, K., Y. S. Li, R. Wasserman, S. Sauder, S. Shinton, and R. R. Hardy. that would indicate that they are also early responders to antigenic 1997. B lymphocyte developmental lineages. Ann. NY Acad. Sci. 815:15. 24. Allman, D. M., S. E. Ferguson, V. M. Lentz, and M. P. Cancro. 1993. Peripheral challenge. We propose that certain nonprotein components of bac- B cell maturation. II. Heat-stable antigen(hi) splenic B cells are an immature teria cell walls (Gram-negative derived mitogens such as LPS, developmental intermediate in the production of long-lived marrow-derived B newly described mitogens from Gram-positive bacteria (70), and cells. J. Immunol. 151:4431. 25. Snapper, C. M., H. Yamada, D. Smoot, R. Sneed, A. Lees, and J. J. Mond. 1993. glycolipids presented by CD1 (71)) would be targeted, in the first Comparative in vitro analysis of proliferation, Ig secretion, and Ig class switching stage of an immune response, to this special subset of MZ B cells by murine marginal zone and follicular B cells. J. Immunol. 150:2737. 26. Waldschmidt, T. J., F. G. Kroese, L. T. Tygrett, D. H. Conrad, and R. G. Lynch. and prepare them for the subsequent phases of gener- 1991. The expression of B cell surface receptors. III. The murine low-affinity IgE ation and B-T cell interaction. Fc receptor is not expressed on Ly 1 or ‘Ly 1-like’ B cells. Int. Immunol. 3:305. Based on the strategic localization of MZ B cells in the spleen 27. Kumararatne, D. S., and I. C. MacLennan. 1981. Cells of the marginal zone of the spleen are lymphocytes derived from recirculating precursors. Eur. J. Immunol. and the findings described in this paper, we propose that MZ B 11:865. cells constitute the first cohort of B cells that respond and partic- 28. Kumararatne, D. S., and I. C. MacLennan. 1982. The origin of marginal-zone ipate in an immune response, in particular in conditions of limited cells. Adv. Exp. Med. Biol. 149:83. 29. Jennings, C. D., and K. A. Foon. 1997. Recent advances in flow cytometry: T cell help, low concentration of TI mitogens, or low numbers of application to the diagnosis of hematologic malignancy. Blood 90:2863. blood-borne particulate Ags reaching the spleen. 30. Fu, Y. X., G. Huang, M. Matsumoto, H. Molina, and D. D. Chaplin. 1997. Independent signals regulate development of primary and secondary follicle structure in spleen and mesenteric lymph node. Proc. Natl. Acad. Sci. USA Acknowledgments 94:5739 . 31. Martin, F., X. Chen, and J. F. Kearney. 1997. Development of VH81X transgene- We thank Lisa Jia, Stephanie Schwichtenberg, Jeffrey Sides, and Ann-

bearing B cells in fetus and adult: sites for expansion and deletion in conventional Downloaded from Elizabeth Ratcliffe for invaluable technical help; Dr. Larry Gartland and and CD5/B1 cells. Int. Immunol. 9:493. Keith Hawk for cell sorting; and Ann Brookshire for editorial help. We are 32. Martin, F., W. J. Won, and J. F. Kearney. 1998. Generation of the germline grateful to Drs. C. C. Goodnow, M. Holers, D. Huszar, and G. Kraal for peripheral B cell repertoire: VH81X-␭ B cells are unable to complete all devel- mice and reagents, and to P. D. Burrows and T.-H. Tai for comments on the opmental programs. J. Immunol. 160:3748. 33. Kinoshita, T., J. Takeda, K. Hong, H. Kozono, H. Sakai, and K. Inoue. 1988. manuscript. Monoclonal antibodies to mouse complement receptor type 1 (CR1). Their use in a distribution study showing that mouse erythrocytes and platelets are CR1-neg- ative. J. Immunol. 140:3066.

References http://www.jimmunol.org/ 34. Kinoshita, T., G. Thyphronitis, G. C. Tsokos, F. D. Finkelman, K. Hong, 1. Rolink, A., and F. Melchers. 1996. B-cell development in the mouse. Immunol. H. Sakai, and K. Inoue. 1990. Characterization of murine complement receptor Lett. 54:157. type 2 and its immunological cross-reactivity with type 1 receptor. Int. Immunol. 2. Hardy, R. R., and K. Hayakawa. 1995. B-lineage differentiation stages resolved 2:651. by multiparameter flow cytometry. Ann. NY Acad. Sci. 764:19. 35. Oliver, A. M., F. Martin, and J. F. Kearney. 1997. Mouse CD38 is down-regu- 3. Burrows, P. D., and M. D. Cooper. 1997. B cell development and differentiation. lated on germinal center B cells and mature plasma cells. J. Immunol. 158:1108. Curr. Opin. Immunol. 9:239. 36. Takahashi, K., Y. Kozono, T. J. Waldschmidt, D. Berthiaume, R. J. Quigg, 4. Liu, Y. J., and C. Arpin. 1997. Germinal center development. Immunol. Rev. Baron, and V. M. Holers. 1997. Mouse complement receptors type 1 (CR1; 156:111. CD35) and type 2 (CR2;CD21): expression on normal B cell subpopulations and 5. Eugster, H. P., M. Muller, U. Karrer, B. D. Car, B. Schnyder, V. M. Eng, Woerly, decreased levels during the development of autoimmunity in MRL/lpr mice. H. M. Le, P. F. Di, M. Aguet, et al. 1996. Multiple immune abnormalities in J. Immunol. 159:1557. tumor necrosis factor and lymphotoxin-␣ double-deficient mice. Int. Immunol. 37. Chen, X., F. Martin, K. A. Forbush, R. M. Perlmutter, and J. F. Kearney. 1997. by guest on September 24, 2021 8:23. Evidence for selection of a population of multi-reactive B cells into the splenic 6. Kelsoe, G., and B. Zheng. 1993. Sites of B-cell activation in vivo. Curr. Opin. marginal zone. Int. Immunol. 9:27. Immunol. 5:418. 38. Hartley, S. B., and C. C. Goodnow. 1994. Censoring of self-reactive B cells with 7. McHeyzer-Williams, M. G., M. J. McLean, P. A. Lalor, and G. J. Nossal. 1993. a range of receptor affinities in transgenic mice expressing heavy chains for a Antigen-driven B cell differentiation in vivo. J. Exp. Med. 178:295. lysozyme-specific antibody. Int. Immunol. 6:1417. 8. Nossal, G. J. 1994. Differentiation of the secondary B-lymphocyte repertoire: the 39. Chen, J., M. Trounstine, F. W. Alt, F. Young, C. Kurahara, J. F. Loring, and germinal center reaction. Immunol. Rev. 137:173. D. Huszar. 1993. Immunoglobulin gene rearrangement in B cell deficient mice 9. Zinkernagel, R. M., M. F. Bachmann, T. M. Kundig, S. Oehen, H. Pirchet, and generated by targeted deletion of the JH locus. Int. Immunol. 5:647. H. Hengartner. 1996. On immunological memory. Annu. Rev. Immunol. 14:333. 40. Timens, W., A. Boes, and S. Poppema. 1989. Human marginal zone B cells are 10. Cumano, A., and K. Rajewsky. 1986. Clonal recruitment and somatic mutation in not an activated B cell subset: strong expression of CD21 as a putative mediator the generation of immunological memory to the hapten NP. EMBO J. 5:2459. for rapid B cell activation. Eur. J. Immunol. 19:2163. 11. Rajewsky, K. 1993. B-cell lifespans in the mouse–why to debate what? Immunol. 41. Waldschmidt, T., K. Snapp, T. Foy, L. Tygrett, and C. Carpenter. 1992. B-cell Today 14:40. subsets defined by the Fc epsilon R. Ann. NY Acad. Sci. 651:84. 12. Oliver, A. M., F. Martin, G. L. Gartland, R. H. Carter, and J. F. Kearney. 1997. 42. Lenschow, D. J., A. I. Sperling, M. P. Cooke, G. Freeman, L. Rhee, D. C. Decker, Marginal zone B cells exhibit unique activation, proliferative and immunoglob- G. Gray, L. M. Nadler, C. C. Goodnow, and J. A. Bluestone. 1994. Differential ulin secretory responses. Eur. J. Immunol. 27:2366. up-regulation of the B7–1 and B7–2 costimulatory molecules after Ig receptor 13. Stall, A. M., S. M. Wells, and K. P. Lam. 1996. B-1 cells: unique origins and engagement by antigen. J. Immunol. 153:1990. functions. Semin. Immunol. 8:45. 43. Thompson, C. B. 1995. Distinct roles for the costimulatory ligands B7-1 and 14. Kumararatne, D. S., H. Bazin, and I. C. MacLennan. 1981. Marginal zones: the B7-2 in differentiation? Cell 81:979. major B cell compartment of rat spleens. Eur. J. Immunol. 11:858. 15. Gray, D., D. S. Kumararatne, J. Lortan, M. Khan, and I. C. MacLennan. 1984. 44. Goodnow, C. C. 1992. Transgenic mice and analysis of B-cell tolerance. Annu. Relation of intra-splenic migration of marginal zone B cells to antigen localiza- Rev. Immunol. 10:489. tion on follicular dendritic cells. Immunology 52:659. 45. Goodnow, C. C., J. Crosbie, S. Adelstein, T. B. Lavoie, S. J. Smith-Gill, 16. Cyster, J. G., and C. C. Goodnow. 1995. Protein tyrosine phosphatase 1C neg- R. A. Brink, H. Pritchard-Briscoe, J. S. Wotherspoon, R. H. Loblay, and atively regulates antigen receptor signaling in B lymphocytes and determines K. Raphael. 1988. Altered immunoglobulin expression and functional silencing thresholds for negative selection. Immunity 2:13. of self-reactive B lymphocytes in transgenic mice. Nature 334:676. 17. Milstein, C., and M. S. Neuberger. 1996. Maturation of the immune response. 46. Sanderson, R. D., P. Lalor, and M. Bernfield. 1989. B lymphocytes express and Adv. Protein Chem. 49:451. lose syndecan at specific stages of differentiation. Cell Regul. 1:27. 18. Foote, J., and C. Milstein. 1991. Kinetic maturation of an immune response. 47. Arpin, C., J. Dechanet, K. C. Van, P. Merville, G. Grouard, F. Briere, Nature 352:530. J. Banchereau, and Y. J. Liu. 1995. Generation of memory B cells and plasma 19. Smith, K. G., A. Light, G. J. Nossal, and D. M. Tarlinton. 1997. The extent of cells in vitro. Science 268:720. affinity maturation differs between the memory and antibody-forming cell com- 48. Buiting, A. M., R. Z. De, G. Kraal, and R. N. Van. 1996. Humoral immune partments in the primary immune response. EMBO J. 16:2996. responses against particulate bacterial antigens are dependent on marginal met- 20. Rajewsky, K. 1996. Clonal selection and learning in the antibody system. Nature allophilic macrophages in the spleen. Scand. J. Immunol. 43:398. 381:751. 49. Seiler, P., P. Aichele, B. Odermatt, H. Hengartner, R. M. Zinkernagel, and 21. Liu, Y. J., J. Zhang, P. J. Lane, E. Y. Chan, and I. C. MacLennan. 1991. Sites of R. A. Schwendener. 1997. Crucial role of marginal zone macrophages and mar- specific B cell activation in primary and secondary responses to T cell-dependent ginal zone metallophils in the clearance of lymphocytic choriomeningitis virus and T cell-independent antigens. Eur. J. Immunol. 21:2951. infection. Eur. J. Immunol. 27:2626. 22. Dunn-Walters, D. K., P. G. Isaacson, and J. Spencer. 1995. Analysis of mutations 50. Kraal, G. 1992. Cells in the marginal zone of the spleen. Int. Rev. Cytol. 132:31. in immunoglobulin heavy chain variable region genes of microdissected marginal 51. Oldfield, S., J. E. Lortan, M. A. Hyatt, and I. C. MacLennan. 1988. Marginal zone zone (MGZ) B cells suggests that the MGZ of human spleen is a reservoir of B cells and the localisation of antigen on follicular dendritic cells. Adv. Exp. Med. memory B cells. J. Exp. Med. 182:559. Biol. 237:99. The Journal of Immunology 7207

52. Humphrey, J. H., D. Grennan, and V. Sundaram. 1984. The origin of follicular 62. Rathmell, J. C., S. Fournier, B. C. Weintraub, J. P. Allison, and C. C. Goodnow. dendritic cells in the mouse and the mechanism of trapping of immune complexes 1998. Repression of B7. 2 on self-reactive B cells is essential to prevent prolif- on them. Eur. J. Immunol. 14:859. eration and allow Fas-mediated deletion by CD4ϩ T cells. J. Exp. Med. 188:651. 53. Fang, Y., C. Xu, Y. X. Fu, V. M. Holers, and H. Molina. 1998. Expression of 63. Dunn-Walters, D. K., L. Boursier, J. Spencer, and P. G. Isaacson. 1998. Analysis complement receptors 1 and 2 on follicular dendritic cells is necessary for the of immunoglobulin genes in splenic marginal zone lymphoma suggests ongoing generation of a strong antigen-specific IgG response. J. Immunol. 160:5273. mutation. Hum. Pathol. 29:585. 64. Inaoki, M., S. Sato, B. C. Weintraub, C. C. Goodnow, and T. F. Tedder. 1997. 54. Kosco-Vilbois, M. H., D. Gray, D. Scheidegger, and M. Julius. 1993. Follicular CD19-regulated signaling thresholds control peripheral tolerance and autoanti- dendritic cells help resting B cells to become effective antigen-presenting cells: body production in B lymphocytes. J. Exp. Med. 186:1923. induction of B7/BB1 and upregulation of major histocompatibility complex class 65. Ahearn, J. M., M. B. Fischer, D. Croix, S. Goerg, M. Ma, J. Xia, X. Zhou, II molecules. J. Exp. Med. 178:2055. R. G. Howard, T. L. Rothstein, and M. C. Carroll. 1996. Disruption of the Cr2 55. Linton, P. J., D. Lo, L. Lai, G. J. Thorbecke, and N. R. Klinman. 1992. Among locus results in a reduction in B-1a cells and in an impaired B cell response to naive precursor cell subpopulations only progenitors of memory B cells originate T-dependent antigen. Immunity 4:251. germinal centers. Eur. J. Immunol. 22:1293. 66. Arnold, L. W., C. A. Pennell, S. K. McCray, and S. H. Clarke. 1994. Develop- 56. Linton, P. L., D. J. Decker, and N. R. Klinman. 1989. Primary antibody-forming ment of B-1 cells: segregation of phosphatidyl choline-specific B cells to the B-1 cells and secondary B cells are generated from separate precursor cell subpopu- population occurs after immunoglobulin gene expression. J. Exp. Med. 179:1585. lations. Cell 59:1049. 67. Murakami, M., K. Nakajima, K. Yamazaki, T. Muraguchi, T. Serikawa, and T. Honjo. 1997. Effects of breeding environments on generation and activation of 57. Liu, Y. J., C. Barthelemy, B. O. de, C. Arpin, I. Durand, and J. Banchereau. 1995. autoreactive B-1 cells in anti-red blood cell autoantibody transgenic mice. J. Exp. Memory B cells from human tonsils colonize mucosal epithelium and directly present Med. 185:791. antigen to T cells by rapid up-regulation of B7–1 and B7–2. Immunity 2:239. 68. Roark, J. H., S. H. Park, J. Jayawardena, U. Kavita, M. Shannon, and 58. MacLennan, I. C., and Y. J. Liu. 1991. Marginal zone B cells respond both to A. Bendelac. 1998. CD1. 1 expression by mouse antigen-presenting cells and polysaccharide antigens and protein antigens. Res. Immunol. 142:346. marginal zone B cells. J. Immunol. 160:3121. 59. Hathcock, K. S., G. Laszlo, C. Pucillo, P. Linsley, and R. J. Hodes. 1994. Com- 69. Brossay, L., D. Jullien, S. Cardell, B. C. Sydora, N. Burdin, R. L. Modlin, and parative analysis of B7–1 and B7–2 costimulatory ligands: expression and func- M. Kronenberg. 1997. Mouse CD1 is mainly expressed on hemopoietic-derived Downloaded from tion. J. Exp. Med. 180:631. cells. J. Immunol. 159:1216. 70. Snapper, C. M., F. R. Rosas, L. Jin, C. Wortham, M. R. Kehry, and J. J. Mond. 60. Hathcock, K. S., and R. J. Hodes. 1996. Role of the CD28–B7 costimulatory 1995. Bacterial lipoproteins may substitute for cytokines in the humoral immune pathways in T cell-dependent B cell responses. Adv. Immunol. 62:131. response to T cell-independent type II antigens. J. Immunol. 155:5582. 61. Kozono, Y., R. Abe, H. Kozono, R. G. Kelly, T. Azuma, and V. M. Holers. 1998. 71. Ernst, W. A., J. Maher, S. Cho, K. R. Niazi, D. Chatterjee, D. B. Moody, Cross-linking CD21/CD35 or CD19 increases both B7-1 and B7-2 expression on G. S. Besra, Y. Watanabe, P. E. Jensen, S. A. Porcelli, et al. 1998. Molecular murine splenic B cells. J. Immunol. 160:1565. interaction of CD1b with lipoglycan antigens. Immunity 8:331. http://www.jimmunol.org/ by guest on September 24, 2021