The Recirculating Pool Contains Two Functionally Distinct, Long-Lived, Posttransitional, Follicular B Cell Populations

This information is current as Annaiah Cariappa, Cristian Boboila, Stewart T. Moran, of September 29, 2021. Haoyuan Liu, Hai Ning Shi and Shiv Pillai J Immunol 2007; 179:2270-2281; ; doi: 10.4049/jimmunol.179.4.2270 http://www.jimmunol.org/content/179/4/2270 Downloaded from

References This article cites 56 articles, 28 of which you can access for free at: http://www.jimmunol.org/content/179/4/2270.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

by guest on September 29, 2021 *average

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 © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

The Recirculating B Cell Pool Contains Two Functionally Distinct, Long-Lived, Posttransitional, Follicular B Cell Populations1

Annaiah Cariappa, Cristian Boboila, Stewart T. Moran, Haoyuan Liu, Hai Ning Shi, and Shiv Pillai2

Disparate models for the development of peripheral B cells may reflect significant heterogeneity in recirculating long-lived B cells that have not been previously accounted for. We show in this study that the murine recirculating B cell pool contains two distinct, long-lived, posttransitional, follicular B cell populations. Follicular Type I IgMlow B cells require Ag-derived and Btk-dependent signals for their development and make up the majority of cells in the recirculating follicular B cell pool. Follicular type II B cells do not require Btk- or Notch-2-derived signals, make up about a third of the long-lived recirculating B cell pool, and can develop Downloaded from in the absence of Ag. These two follicular populations exhibit differences in basal tyrosine phosphorylation and in BCR-induced proliferation, suggesting that they may represent functionally distinct populations of long-lived recirculating B cells. The Journal of Immunology, 2007, 179: 2270–2281.

uring binary cell fate decisions, distinct signaling recep- levels of CD21/CR2 and CD23 (16) represent newly formed or 3

tors ensure that daughter cells representing each of two transitional type 1 (NF/T1) B cells. They may reside in an extra- http://www.jimmunol.org/ D distinct fates are generated, and that development is not follicular location in the before their entry into B lymphoid completely skewed in favor of one type of daughter population. follicles. These NF/T1 B cells have a short half-life, and they ma- During the generation of B and T cells, for instance, the Flk2 ture after a brief delay into transitional type 2 (T2) cells. T2 cells receptor tyrosine kinase, in conjunction with IL-7-derived signals, still express high levels of CD24 and CD93, reflecting their rela- may drive the commitment of progenitors toward the B lineage, tively recent generation in the bone marrow, express high levels of while signals from Notch1 ensure the simultaneous commitment of IgM, and are CD23ϩ (13, 17, 18). A population of CD93ϩ B cells progenitors to a fate (1–3). Broadly similar mechanisms transitional type 3 (T3) cellsthat are IgMlow has also been de- involving reciprocal pairs or sets of pairs of signaling receptors scribed (18). T2 and T3 B cells are believed to represent transient may operate during other commitment events that appear to in- intermediates on their way to becoming long-lived recirculating by guest on September 29, 2021 volve binary cell fate decisions. One such binary cell fate decision follicular B cells that remain CD23ϩ, but no longer express CD93 involves the commitment of peripheral B cells to a follicular or a (Fig. 1A). In a recent study, it has been suggested that T3 B cells marginal zone B cell fate. represent anergic B cells (19). In a different classification, nonre- The development of naive peripheral B cells depends on the circulating splenic IgMhighIgDhigh CD23ϩCD21high cells were also signals delivered by the BCR, as well as on survival and differ- designated as T2 cells (16). We and others have suggested that entiation signals derived from other receptors and transcriptional these latter cells represent specialized splenic precursors of mar- regulators, including B cell activating factor receptor (BAFF-R), ginal zone (MZ) B cells called MZ precursor, or MZP, B cells (7, Notch2, and NF␬B1 (4–10). Developmental signals that originate 9, 10, 20–23). MZP B cells reside only in splenic follicles in ro- from the BCR are postulated to include both constitutive ligand- dents and are nonrecirculating B cells. independent signals, as well as signals that depend on interactions A number of mutations that result in striking defects in MZ B between distinct self-Ags and their cognate BCRs, apparently be- cell development are accompanied by normal or enhanced follic- low the avidity threshold required to trigger receptor editing and ular B cell numbers. MZ B cell development is defective in the clonal deletion (11, 12). absence of negative regulators of BCR signaling, such as Aiolos Immature B cells in the bone marrow represent the first defined and CD22, but mature follicular B cell numbers are enhanced in stage of B cell development in which the BCR is expressed on the these mice (20, 24, 25). The loss of MZ and MZP B cells in cell surface. These cells also express high levels of CD24 and AiolosϪ/Ϫ mice was reversed in mice harboring defects in both CD93 (13–15). IgMhighIgDlow/ϪCD93ϩ B cells that express low Aiolos and Btk, suggesting that enhanced Btk signaling compro- mises MZ B cell development (20), but contributes to normal or increased numbers of IgDhighIgMlow follicular B cells. Although Cancer Center, Massachusetts General Hospital, Harvard Medical School, Boston, MA 02129 this signal strength model (Fig. 1B) is consistent with the normal Received for publication November 9, 2006. Accepted for publication June 1, 2007. development of MZ B cells observed in the absence of Btk in 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 with 18 U.S.C. Section 1734 solely to indicate this fact. 3 1 Abbreviations used in this paper: NF/T1, newly formed or transitional type 1; T2, This work was supported by Grants AI06493 and CA102793 from the National transitional type 2; T3, transitional type 3; MZ, marginal zone; MZP, MZ precursor; Institutes of Health. APC, allophycocyanin; HEL, hen egg lysozyme; SA, streptavidin; BrdU, bromode- 2 Address correspondence and reprint requests to Dr. Shiv Pillai, MGH Cancer oxyuridine; FO II, follicular type II; FO I, follicular type I. Center, Building 149, 13th Street, Charlestown, MA 02129. E-mail address: [email protected] Copyright © 2007 by The American Association of Immunologists, Inc. 0022-1767/07/$2.00 www.jimmunol.org The Journal of Immunology 2271

FIGURE 1. Models for transitional B cells and the role of Ag in splenic B cell differentiation. A, Sequential progression through transitional T1, T2, and T3 stages leads to the generation of mature B cells. T1 cells are AA4.1ϩCD23ϪCD45ϩ B cells, T2 B cells are AA4.1ϩCD23ϩIgMhigh B cells, and T3 B cells are AA4.1ϩCD23ϩIgMlow B cells. Ma- ture B cells are AA4.1ϪCD23ϩ IgMϩ B cells (after Allman et al., Ref. 18). B, A signal strength model for peripheral B cell development. NF/T1 B cells are IgMhighIgDlowCD21low cells. T2/FP cells are IgMhighIgDhighCD21int B cells. Relatively strong

BCR signals and Btk are required for the generation Downloaded from of IgDhighIgMlow FO B cells, while relatively weak signals collaborate with Notch2 and NF-␬B1 to yield MZ B cells (after Pillai et al.; Ref. 10). C,Ag mediated MZ B cell development. Ag mediates MZ B cell development while CD24lowCD23ϩ follicular B cells develop in the absence of Ag (after Wen et al.; Ref. 27). http://www.jimmunol.org/ by guest on September 29, 2021 replete mice, other non-BCR-linked genes, such as Notch2, RBP- those locations. Indeed, mutations in Pyk2, a tyrosine kinase linked J␬, and NF␬B1 are also required for MZ B cell development (6– to integrin signaling (32), the Rho guanine nucleotide exchange 8). We have therefore suggested that developing B cells that factor Lsc, (33), the adaptor DOCK-2 (34), the Rac2 Rho family receive weaker BCR signals may be receptive to inductive signals GTPase (35), and integrins (36) exhibit defective MZ B cell de- delivered by intracellular Notch2 and by NF␬B1 to differentiate velopment because these genes presumably contribute to the mi- first into MZP B cells and to then migrate into and be retained in gration or retention of MZ B cells in the marginal zone. the marginal zone area as MZ B cells (10). NF␬B activation occurs Most of the evidence for presumed stages of peripheral B cell ␬ downstream of the BAFF receptor, and Notch2 and NF B1 func- development is inferential and has been obtained from the analysis tion synergistically during MZ B cell development (26). Another of mutants or the study of the sequential emergence of B cell study using BCR transgenic mice has suggested that weaker BCR populations following nonablative irradiation. Direct evidence for signals contribute to follicular B cell development rather than MZ precursor-product type relationships of specific populations during B cell development (27; Fig. 1C), and this view will be discussed B development is generally lacking. A major con- below in the context of our current findings. founding factor in the study of peripheral B cell development is the In contrast to defects in the Btk/PLC␥2 pathway that compro- phenomenon of homeostatic proliferation. Transfer of purified B mise follicular but not MZ B cell development, a number of other cell populations into nonirradiated replete mice results in very poor mutations in genes encoding signaling proteins result in a marked subsequent recovery of the transferred cells, making precursor- reduction of MZ B cells. Mutations in CD19 (28), the p50␣-p55␣- product type studies on cellular populations by cell transfer virtu- p85␣ regulatory subunits of PI3K (29), and the p110 ␦ subunit of PI3K (30, 31) all present defects in BCR/Btk/Plc␥2 activation, but ally impossible. Transfer of B cell populations into mice that lack also have defects in the development of both B-1 and MZ B cells. as a consequence of immunodeficiency or irradiation It is recognized that PI3 kinase not only contributes to the activa- results in homeostatic proliferation, the expression of high levels tion of Btk, but also is required for the activation of Akt and the of CD21, and the generation preferentially of MZ B cells (37–40). subsequent inhibition of Forkhead family proteins. It has been sug- In these circumstances, following transfer into lymphopenic hosts, gested that the peripheral B cell developmental defects seen in the transitional and follicular B cells primarily give rise to MZ B cells. absence of PI3K signaling (but not in the absence of Btk and Lymphocyte development appears to proceed in a distinct manner PLC␥2) may reflect the failure to activate the Akt pathway in PI3K in replete mice. Some genes that are required for homeostatic pro- mutant mice (29). CD19 functions as an adaptor for PI3K activa- liferation in lymphopenic mice are not required for normal MZ B tion, and it is possible that, during B cell development, these mol- cell development in replete mice. As an example, homeostatic B ecules contribute to the migration of B cells to the marginal zone cell proliferation in lymphopenic recipients is impaired in the ab- and/or the peritoneal cavity, as well as to the retention of cells in sence of Btk, but MZ B cell development proceeds normally in 2272 TWO DISTINCT POPULATIONS OF MATURE FOLLICULAR B CELLS

replete mice that lack Btk (39). Similarly, although NF␬B1 con- Lymphocyte proliferation assay tributes to MZ B cell development (8), it is not required for the The method described previously (24) was followed with some modifica- homeostatic proliferation of B cells (39). tions. In brief, sorted cell fractions were plated at 1 ϫ 105 cells/well in Our studies, described below, indicate that two functionally dis- triplicate in 96-well flat-bottom microtiter plates with medium alone or tinct, long-lived, recirculating, posttransitional, follicular B cell with medium and one of the following: goat anti-mouse IgM (Fab’)2 (␮- ␮ populations can be distinguished, one of which represents a pool of chain specific), 10 g/ml (Jackson ImmunoResearch Laboratories), poly- clonal goat anti-mouse IgD antiserum (activating), 10 ␮g/ml (eBioscience), B cells that develops in the apparent absence of self-Ag-mediated purified hamster anti-mouse CD40, 1 ␮g/ml (BD Pharmingen), LPS from BCR ligation and does not require Btk, and is also preserved in the Escherichia coli serotype O55:B5, 10 ␮g/ml (Sigma Aldrich). The cells absence of Notch2-derived signals. Therefore, this long-lived pop- were cultured for 48 h, then pulsed with 2 ␮Ci/well of [3H]thymidine ulation could represent the cell type that undergoes a cell fate (PerkinElmer) for an additional 18 h, and harvested and read in a betaplate reader. decision, giving rise to either MZ B cells or IgMlow follicular B cells. The latter IgMlow B cell population requires Ag and Btk Affymetrix DNA GeneChip probe array assay signals for development, presumably resulting from the relatively IgMhighIgDhighCD21int follicular type II (FO II) B cells and IgMlowIgDhigh avid interaction of developing B cells in vivo with cognate self- CD21int follicular type I (FO I) B cells were sorted directly into RNAlater structures. Possible reasons for the existence of these two distinct (Qiagen). Total cellular RNA was prepared using RNeasy columns (Qia- follicular B cell populations will be discussed below. gen). RNA was amplified twice using RiboAmp RNA Amplification Kit (Arcturus) and single-stranded biotinylated cRNA probes were generated using the cDNA as template, and the BioArray RNA Transcript Labeling Materials and Methods Kit (Enzo Biochem), which uses T7 RNA polymerase, and biotin-labeled Mice nucleotides. The probes were purified using Qiagen columns, fragmented, Downloaded from and hybridized to the murine genome MOE430A set GeneChip (Af- The HyHEL10 IgH and IgL knock-in mice have been described previously fymetrix) at the Harvard Center for Genomic Research (Cambridge, MA). (41, 42) and were kindly made available by Dr. J. Cyster (University of Hybridization signals were detected by a GeneChip Scanner 2500 (Af- California, San Francisco, CA) and Dr. M. Nussenzweig (Rockefeller Uni- fymetrix) and checked for uniformity. Probe sets were all scaled to the versity, New York). Notch2ϩ/Ϫ mice (7) were made available by Dr. Y. same target intensity of 500 using global scaling, and the data were ana- Hamada (University of Tokyo, Tokyo, Japan). C57BL/6 mice were pur- lyzed using the Affymetrix Microarray Analysis Suite (MAS 5.0). The samples were validated using two sets of internal controls: Affymetrix in- chased from The Jackson Laboratories and housed in a pathogen-free fa- http://www.jimmunol.org/ cility. All mice used for experiments were between 8 and 12 wk of age. ternal control genes, and a known set of genes that should be up- or down- Animal procedures were cleared by the Subcommittee on Research Animal regulated in the two follicular B cell samples. Care at Massachusetts General Hospital. In vitro differentiation Flow cytometric analysis and fluorescent-activated cell sorting Flow-purified splenocytes were plated in 48-well tissue culture plates at a density of 0.5–1 ϫ 105 cells/well in triplicate in 300 ␮l of complete Flow cytometry and flow sorting were performed as previously described ␣-DMEM with or without HEL, and cultured in a CO incubator at 37°C (43, 44). The following murine mAb conjugates were used: allophycocya- 2 and 5% CO2. Two days later, each of three wells were pooled, filtered, and nin (APC)- and r-PE-Cy7-1B4B1 (anti-IgM, rat IgG), and R-PE-, biotin- stained for flow cytometry using the same fluorochrome combination as ␬ ylated-, and Cy5-11-26 (anti-IgD, rat IgG2a, , all from Southern Bio- used for the sort. technology Associates, Birmingham, AL), and purified 2.4G2 (anti-CD16/ by guest on September 29, 2021 CD32 {Fc␥ III/II receptor}, rat IgG2b, ␬, culture supernatant), APC- and Statistical analysis ␬ Pacific Blue-RA3-6B2 (anti-CD45R/B220, rat IgG2a, ), PE-Cy7-RA3- 6B2, PE-Cy7-R6-60.2 (anti-IgM, rat {LOU}, rat IgG2a, ␬), FITC- and The p values for differences between groups were calculated by the Mann- PE-7G6, (anti-CD21/CD35, rat IgG2b), PE-M1/69 (anti-CD24, Heat Sta- Whitney U test using StatView (version 5.0.1). BrdU decay curves were ble Ag, rat {DA} IgG2b, ␬), PE-S7 (anti-CD43, Leukosialin, rat IgG2a, ␬), generated by the least squares nonlinear regression analysis method using PE-B3B4 (anti-CD23, rat IgG2b, ␬), and PE-AA4.1 (Early B Lineage, rat MacCurveFit 1.5 (Kevin Raner Software). {Harlan Sprague Dawley} IgG2b, ␬) all from BD Pharmingen, and APC- and PE-AA4.1 (eBioscience). Hen egg lysozyme (HEL; Sigma-Aldrich) Results was biotinylated, and was revealed by streptavidin (SA)-PE-Texas Red or In these studies, we sought to test the hypothesis that there are two PerCP-Cy5.5, and biotinylated IgD was revealed by SA-PE-Cy5.5 or SA- distinct, long-lived, recirculating, posttransitional follicular B cell PerCP (all from BD Pharmingen). For intracellular cytokine staining, the populations, one of which owes its origins to relatively strong Btk- following R-PE-conjugated Abs were used: XMG1.2 (anti-IFN-␥, rat IgG1), MP6-XT22 (anti-TNF, rat IgG1), JES6-5H4 (anti-IL-2, rat IgG2b), dependent BCR signals (and may be consequently prohibited from 11B11 (anti-IL-4, rat IgG1), MP5-20F3 (anti-IL-6, rat IgG1), and C15.6 differentiation into the MZ B cell lineage), whereas the other re- (anti-IL-12, rat IgG1), all from BD Pharmingen. Intracellular cytokine quires weaker BCR signals and might represent a distinct type of staining was performed after surface staining, following treatment with mature follicular B cell that develops in the absence of cognate Ag ␮ 0.1% saponin (Sigma-Aldrich), and GolgiPlug 1 l/ml (brefeldin A; BD in a “default” manner, dependent perhaps on constitutive BCR Pharmingen). Intracellular phosphotyrosine residues were stained with FITC-4G10 (anti-phosphotyrosine mAb; Upstate Biotechnology) follow- signaling and exposure to BAFF. Such a population of follicular B ing surface staining and treatment with 0.1% saponin. Staining at 4°C or cells would theoretically retain the ability to be drawn into the MZ after fixation with paraformaldehyde gave similar results. B cell pool when required. Flow cytometric analysis was performed on a dual laser FC500 (Beckman Coulter), and sorting was performed on a MoFlo sorter IgMhighIgDhigh B cells in the spleen can be separated into two (DakoCytomation), a FACSAria (BD Biosciences), and an Epics Altra distinguishable B cell populations hypersort system (Beckman Coulter). The purity of sorted samples al- ways exceeded 96%. Processed sample data was analyzed using FloJo Hardy et al. (45) described two categories of IgDhigh B cells dis- v8.1.1 (Tree Star) software. IgM/IgD gates in the spleen were set ac- tinguishable by their different levels of surface IgM, and by the cording to Hardy et al. (45). fact that one category was markedly diminished in Xid mice. Xid mice carry a mutant version of the TEC kinase, Btk. IgDhighIgMlow Bromodeoxyuridine (BrdU) labeling (Fraction I) B cells are lost in the absence of Btk-derived signals, Continuous labeling with BrdU (Sigma Aldrich) was performed as de- whereas IgDhighIgMhigh (Fraction II) B cells were maintained in scribed earlier (43). In brief, 0.25 mg/ml BrdU and 2 mg/ml glucose were Xid mice. It has sometimes been assumed that IgDhighIgMhigh administered in drinking water for 22 days, and the decay in labeling was followed for 4 wk after cessation of BrdU administration. The BrdU decay Fraction II B cells are also transitional T2 B cells (16), but this has data curves were generated by the least squares nonlinear regression anal- never been rigorously established or disproved, and this is an issue ysis method. that will be addressed below. The Journal of Immunology 2273

tions of transitional B cells in the FO II B cell pool in adult mice. A small proportion of FO II B cells (ϳ10%) expressed AA4.1 or were CD24high (Fig. 3B). These data indicate that the vast majority of FO II B cells are AA4.1neg posttransitional B cells, and that these cells are therefore potentially long-lived B lymphocytes. FO II B lymphocytes are long-lived cells To determine whether FO II B cells are long-lived or short-lived B cells, we performed continuous BrdU labeling studies. Labeling studies have been reported on unfractionated Fraction II B cells (48), but we wished to separate these IgMhighIgDhigh B cells into MZP and FO II B cell pools. Mice were fed BrdU continuously for 3 wk, labeling was stopped abruptly, and BrdU-positive cells in different B cell populations in the spleen were enumerated for the following 4 wk as the label decayed and were plotted using re- gression analysis. As can be seen from Fig. 4, FO I, MZP, and FO II B cells are long-lived splenic follicular B cell populations, whereas NF/T1 cells, as is well established, represent a short-lived

B cell population. MZ B cells accumulate higher levels of BrdU Downloaded from initially, but are also long-lived B cells. Because FO II cells are FIGURE 2. Splenic IgMhighIgDhigh B cells can be separated into two long-lived B cells, can be identified in lymph nodes and the distinct populations. IgMhighIgDhighCD21high B cells are MZP cells. IgDhigh bone marrow, and show clear evidence of recirculation when IgMhighCD21int cells represent FO II B cells. FO II and MZP B cells were assessed using a parabiosis approach (44), they represent a dis- separated by flow sorting and recharacterized as shown. tinct pool of long-lived recirculating B cells, easily distinguish-

able phenotypically from FO I B cells. Because these B cells http://www.jimmunol.org/ have a long half-life similar to those of FO I B cells, the vast As seen in Fig. 2, IgDhighIgMhigh Fraction II B cells can be majority of the FO II cells are clearly posttransitional B cells, ϩ separated into two distinct populations. Both are CD23 . One pop- in keeping with the data presented in Fig. 3. BrdU labeling ulation expresses high levels of CD21. These cells also express studies were also performed on bone marrow B cells (44) and high levels of CD1d and CD9 (46, 47; AC and SP, unpublished FO II B cells identified in the bone marrow were as long lived observations) and correspond to the cells that we and others have as FO I B cells (data not shown). previously called MZP B cells (7, 9, 10, 20–23). A distinct sub- ␮ population of Fraction II B cells expresses intermediate levels of Alteration of Ig mRNA levels in FO I vs FO II B cells CD21 (20), and also lower levels of CD1d and CD9 than MZP The specific reduction in cell surface IgM levels in FO I vs FO II by guest on September 29, 2021 cells (47). We refer to these Btk-independent IgMhighIgDhigh cells may reflect an alteration of gene expression at a transcrip- CD21int B cells as FO II or FO II B cells to differentiate these cells tional or posttranscriptional level, or might reflect the posttransla- from the Btk-dependent Fraction I IgDhighIgMlow population, tional down-regulation specifically of IgM BCRs. We compared which we categorize as FO I or FO I B cells. the expression, at the level of mRNA accumulation, of selected There were a number of issues regarding FO II B cells that we specific B cell genes in purified FO I and FO II B cells using a sought answers to. Are FO II B cells long-lived B cells or are they microarray approach. As can be seen in Fig. 5, although mRNA part of the transitional pool? Although Fraction II cells survive in accumulation of CD21, CD1d, and CD24 are comparable be- the absence of Btk, it is unclear whether the surviving cells are FO tween FO I and FO II B cells, there is a striking reduction of II B cells or MZP B cells. Do FO II B cells develop in a Btk- ␮-chain mRNA accumulation noted in FO I B cells. Although independent manner, and are they preserved in mice in which B there might or might not be a posttranslational component to cells develop in the apparent absence of a cognate Ag? Are cells of surface IgM regulation in FO I B cells, a major differentiation- the Btk-dependent FO I B cell population lost in the absence of a related alteration between FO I and FO II B cells occurs at the cognate Ag? level of the regulation of ␮-mRNA levels by transcriptional or posttranscriptional means. The majority of FO II B cells are posttransitional B cells The elegant studies of Allman et al. (17, 18) have revealed the FO I and FO II B cells exhibit differences in basal tyrosine presence of a population of AA4.1/CD93ϩ B cells that express phosphorylation, basal cytokine synthesis, and proliferation CD23 and are referred to as T2 cells. We have attempted to de- following BCR ligation lineate the relationship between all peripheral B cell populations, Basal tyrosine phosphorylation of FO II B cells is set at a slightly as categorized by different groups, using six-color flow cytometry. higher level than that of FO I B cells, suggesting intrinsic differ- The vast majority of AA4.1ϩ B cells reside in the newly ences between these two populations. Basal tyrosine phosphory- formed/T1 B cell pool, and in the T2 pool. As shown in Fig. 3A, lation levels of FO II B cells resemble the levels observed in a large proportion (ϳ80%) of the AA4.1ϩ CD23ϩ T2 population NF/T1 B cells, and the MZ and MZP B cell populations (Fig. 6A). is made up of IgMhighIgDlow B cells that do not overlap with the However, mobilization of calcium following BCR crosslinking FO II pool. A proportion of the IgMhighIgDhigh CD23ϩ FO II pool was similar in FO I and FO II B cells (data not shown). Sorted FO may be made up of relatively recent emigrants from the bone mar- I cells proliferated less than sorted FO II cells in response to BCR row that have just entered the follicular B cell compartment, and ligation with either anti-IgD or anti-IgM (Fig. 6B). FO I cells also we wished to determine what proportion of FO II B cells express proliferated less well in response to CD40 ligation, but these cells markers of the transitional/T2 population. We used AA4.1, as well do not have a global impairment in proliferation because the re- as the widely used CD24 marker, to identify the relative propor- sponse to LPS was robust (Fig. 6B). A relatively small proportion 2274 TWO DISTINCT POPULATIONS OF MATURE FOLLICULAR B CELLS Downloaded from

FIGURE 3. A small proportion of FO II B cells express transitional markers. A, Most AA4.1ϩCD23ϩ IgMhigh T2 cells are IgDlow B cells. Six-color flow cytometric analysis

was performed using Abs against http://www.jimmunol.org/ CD45R, CD23, CD21, IgM, IgD, and CD93. B, A small proportion of IgMhighIgDhighCD21int FO II B cells express transitional markers such as AA4.1 (i) or CD24 (ii); n ϭ 3 mice. by guest on September 29, 2021

of activated B cells synthesizes cytokines (49), and we have noted FO II B cells are preserved in mice heterozygous for Notch2 basal synthesis of cytokines in a small number of ex vivo follicular and in the absence of Btk-derived signals B cells. However, the proportion of cytokine-synthesizing cells is consistently greater in the FO II population than in FO I B cells MZ B cells are lost in mice in which the recombination signal ␬ (Fig. 7). The proportions of FO II B cells that synthesize cytokines binding protein RBPJ or Notch2 has been conditionally deleted in under basal conditions more closely resemble the proportions seen the B lineage, and are significantly diminished in mice that are in newly formed B cells, and in MZP and MZ B cells (data not heterozygous for Notch2 (6, 7, 50). We wished to establish shown), broadly mirroring the results of studies on basal tyrosine whether the long-lived FO II B cell population is generated inde- phosphorylation. Taken together, the differences in basal tyrosine pendently of both Btk-derived and Notch2-dependent differentia- phosphorylation, the in vitro proliferation data, and the comparison tion signals. We examined FO II B cells and other splenic B cell of cytokine synthesis in unstimulated cells suggest that FO I and populations in wild-type mice and in Notch2 heterozygotes. As can FO II B cells may represent functionally distinct populations. be observed in Fig. 8, A and B, although there is a significant The Journal of Immunology 2275

In replete as well as lymphopenic BCR knock-in mice lacking cognate Ag, the FO II B cell population is preserved, whereas FO I B cells are markedly diminished A number of BCR transgenic mice have been identified that ex- hibit a marginal zone B cell bias with a relative decrease in fol- licular B cells. One possible explanation for this phenomenon is that these mice are relatively lymphopenic and that homeostatic proliferation drives the generation of MZ B cells (40). We decided to examine mice with an Ag-specific BCR in a nonlymphopenic or a relatively lymphopenic setting. Anti-HEL transgenic BCR ex- pressing B cells have been extensively studied (41, 42, 51). We examined the distribution of splenic peripheral B cell populations in an anti-HEL H and L chain BCR knock-in mouse in the absence of HEL. Heterozygous BCR knock-in mice essentially express “physiological” levels of the BCR, and knock-in Ig genes are sub- ject to somatic alterations that may occur during an immune re- sponse (52). Our expectations were that, in the absence of any FIGURE 4. FO II B cells are long-lived cells. Following continuous known cognate Ag, the FO I population would be markedly di- BrdU labeling for 22 days, labeling was abruptly stopped and allowed to minished (as has been noted previously in anti-HEL transgenic Downloaded from decay duringa4wkperiod. Regression analysis of the data is shown. Error mice), but that FO II, MZP, and MZ B cells would be preserved. bars, SEM. NF, Newly formed B cells. n ϭ 3 mice per time point. As seen in Fig. 9A, in nonlymphopenic mice heterozygous for the anti-HEL BCR knock-in H and L chain genes, only a propor- tion of all B cells were HEL specific. One explanation for this decrease in MZ B cells in Notch2 heterozygotes, and a significant phenomenon is the presence of ongoing receptor editing and V increase in FO I B cells, the relative absence of Notch2 does not gene replacement (data not shown). HEL-binding B cells were http://www.jimmunol.org/ significantly influence the FO II B cell population, which is largely excluded from the FO I compartment, suggesting that en- slightly more abundant in Notch2 heterozygotes. As seen in Fig. dogenous Ags that cross-react sufficiently with HEL and poten- 8C, it is clear that at steady-state, FO II B cells are present in tially drive FO I B cell development may be of very low abun- normal numbers in Xid mice, although, as has long been estab- dance. The IgD/IgM profile of HEL-binding B cells resembles that lished, Btk-derived signals are required for the preservation of FO described for Xid mice, in keeping with the view that Ag and I B cells (45). These data suggest that although the generation or Btk-derived BCR signals may be required for the generation of the survival of FO I B cells depends on Btk, and MZ B cells represent FO I phenotype. In mice homozygous for anti-HEL H and L chain a Notch2-dependent population, long-lived FO II B cells do not genes, the IgD/IgM profiles of HEL-binding cells were virtually by guest on September 29, 2021 require Btk or Notch2 for their generation or maintenance. identical to those of heterozygous knock-in mice (Fig. 9A). These mice exhibit a marked reduction of all IgDhighIgMlow FO I B cells, presumably because there are fewer non-HEL binding B cells in the homozygous mice and HEL-specific B cells are largely ex- cluded from the FO I fraction. We tested the assumption that the few, but clearly identifiable, non-HEL-binding B cells in homozygous knock-in mice may have arisen because of “spontaneous” receptor editing and V gene re- placement events in the absence of Ag. All ␬L chain positive non- HEL-binding B cells in the FO I compartment in homozygous

anti-HEL knock-in mice had undergone VH gene replacement (data not shown; Ref. 53). This phenomenon may explain, in part, why only a fraction of the B cells in heterozygous BCR knock-in mice express anti-HEL BCRs. Although FO I HEL-binding B cells were diminished in homozygous knock-in mice, immunohisto- chemical studies on homozygous knock-in mice revealed the presence of HEL-binding B cells in splenic follicles, and these presumably represent primarily FO II and MZP B cells (data not shown). Although the absence of specific Ag leads to a marked reduction in HEL-specific FO I B cells, FO II B cell numbers were preserved in both relatively normopenic heterozygous knock-in mice and in lymphopenic homozygous mice, suggest- ing that this long-lived follicular B cell population, in contrast to FO I B cells, develops and is maintained in the absence of any known self Ag.

FIGURE 5. FO I and FO II B cells can be distinguished at the level of Spontaneous development of FO II B cells in the absence of Ag IgM mRNA accumulation. mRNA levels of CD1d, CD21, CD22, CD23, and IgM were compared in purified FO I and FO II cells, on an Affymetrix The expression of cell-surface CD93 decreases soon after periph- microarray. GAPDH expression was similar in both samples in each of two eral B cells mature, whereas CD23 levels and IgD levels increase separate analyses. during the process of maturation. T1 cells presumably develop into 2276 TWO DISTINCT POPULATIONS OF MATURE FOLLICULAR B CELLS Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021 FIGURE 6. FO I cells proliferate less than FO II cells following BCR stimulation and CD40 ligation. A, Basal tyrosine phosphorylation was compared by flow cytometry in FO I and FO II B cells. B, Proliferation was examined in the context of anti-IgD-mediated activation (left), anti-CD40 triggering (left), anti-IgM-mediated activation (right), and LPS-mediated activation (right). n ϭ 3 mice per group. Error bars, SEM.

T2 cells that are initially IgDlow, which go on to mature into quired in vitro, in the absence of any cognate Ag, by newly IgDhigh T2 cells, which might then in turn mature into the cells formed/T1 cells. We also wished to examine whether FO II B cells with the phenotype of AA4.1Ϫ long-lived FO II B cells. We sought have the potential to assume a FO I-like phenotype in vitro to determine whether the IgMhighIgDhigh phenotype may be ac- following exposure to cognate Ag. There were many reasons to

FIGURE 7. Under basal conditions, cytokine synthesis is more frequently observed in the FO II population of follicular B cells. The synthesis of TNF, IFN-␥, IL-2, IL-4, IL-6, and IL-12 was compared in ex vivo unstimulated FO I and FO II B cells using intracellular staining and flow cytometry. Representative results are shown, but similar results were noted in six mice. The Journal of Immunology 2277 Downloaded from

FIGURE 8. FO II cells develop independently of Notch2 and Btk. A, Peripheral splenic B cells were an- alyzed by flow cytometry in wild-type mice and Notch2 heterozygotes. The FO II B cell population was pre-

served. n ϭ three mice per group. B, Quantitation of FO http://www.jimmunol.org/ I, FO II, MZP, and MZ B cell populations in wild-type and Notch2 heterozygotes, based on mice analyzed in A. Error bars, SEM. C, FO II B cells are preserved in Xid mice. by guest on September 29, 2021

choose an in vitro approach. The differentiation potential of avoid any potential activation of B cells by Abs against IgM and specific B cell populations cannot be ascertained in vivo by cell IgD. When incubated for 48 h in the absence of Ag, these newly transfer because of the confounding effect of homeostatic pro- formed B cells spontaneously assumed an IgMhighIgDhigh pheno- liferation, and cell transfer studies in replete mice are not readily type resembling that of Fraction II B cells. However, in the pres- interpretable. Clearly, exposure of certain BCR transgene-expressing ence of low concentrations of HEL (10 ng/ml), these cells assumed B cells to self-Ags leads to the induction of tolerance in vivo. An in a FO I B cell-like IgDhighIgMlow phenotype after 48 h (Fig. 9B). vitro approach permits the easy titration of cognate Ag and the use of This negative selection approach for cell sorting was initially un- specific B cell populations, but focuses specifically on differentiation dertaken because there is no method available to purify FO II B potential, without addressing issues that relate to tolerance. cells without using Abs against surface Igs. In separate experi- We purified newly formed B cells by an immunodepletion ap- ments, we also purified FO II B cells from homozygous anti-HEL proach, selecting cells that were CD43ϪCD23ϪCD21Ϫ by flow knock-in mice using what are described as nonactivating Abs to cytometry from a homozygous anti-HEL BCR knock-in mouse IgM, IgD, and CD21, and exposed these cells to the same concen- (Fig. 9B). This negative selection approach was initially used to trations of Ag used in the experiments with newly formed B cells. 2278 TWO DISTINCT POPULATIONS OF MATURE FOLLICULAR B CELLS

FIGURE 9. Differentiation of newly formed B cells into IgMhighIgDhigh B cells occurs in the absence of Ag. A, Analysis of IgM- and IgD-ex- pressing HEL-binding B cells in of anti-HEL BCR knock-in heterozygous and homozygous mice. FO I, FO II, and MZ B cell popula- tions are quantitatively represented on the right. B, Newly formed B cells differentiate into IgMhighIgDhigh B cells in the absence of Ag. CD43neg Downloaded from CD21negCD23neg splenic B cells from homozygous knock-in mice were pu- rified using flow cytometry, reana- lyzed after sorting, and incubated for 48 h in the absence or presence of HEL. Cells were analyzed for the ex- http://www.jimmunol.org/ pression of IgM and IgD using fluo- rochromes not used during the initial sort. C, IgMhighIgDhighCD21int FO II B cells were flow purified from ho- mozygous anti-HEL BCR knock-in mice and incubated in the absence or presence of HEL. Cells were ana- lyzed after 48 h using the same colors for anti-IgM and anti-IgD as used ini- high tially for purification. D, IgM by guest on September 29, 2021 IgDlowCD21low newly formed B cells were purified using flow sorting from homozygous anti-HEL BCR knock-in mice and incubated in the absence or presence of HEL. Cells were ana- lyzed after 48 h using the same colors for anti-IgM and anti-IgD as used ini- tially for purification. E, IgMhigh IgDlowCD21low newly formed B cells and IgMhighIgDhighCD21int FO II B cells were flow purified from wild- type mice and individually incubated with or without HEL for 48 h. Cells were analyzed after 48 h using the same colors for anti-IgM and anti-IgD as used initially for purification.

As seen in Fig. 9C, incubation of FO II B cells in the absence of linked to the specific Ag and not any potential contaminant (such Ag for 48 h caused no discernible change in the IgMhighIgDhigh as LPS, for instance), because the same HEL preparation did not phenotype of FO II B cells. However, incubation of these cells induce the differentiation of Ag-nonspecific B cells (Fig. 9E), nor with low concentrations of HEL in vitro resulted in most of these did it induce an alteration in the size of HEL-specific cells as cells assuming an FO I-like phenotype after 48 h. Differentiation is assessed by forward scatter (data not shown). Purification of The Journal of Immunology 2279 Downloaded from http://www.jimmunol.org/

FIGURE 10. A schematic model for peripheral B cell development in the absence of Ag. See text for details.

NF/T1 cells from homozygous anti-HEL knock-in mice (Fig. 9D) Notch2 for development. It appears therefore that FO II B cells or from wild-type mice (Fig. 9E, left) using “nonactivating” Abs represents a “default” follicular B cell population that requires by guest on September 29, 2021 against IgM, IgD, and CD21 also led to the in vitro assumption of BAFF, but does not require Btk activation or Notch2 derived sig- an FO II-like phenotype by NF/T1 B cells similar to the results nals for development. It is likely that a “default” follicular B cell observed in Fig. 9B. These studies suggest that both the depletion population noted in a previous study, but not further characterized, and positive selection approaches to purification of T1 and FO II is actually the FO II B cell pool (27). A model for B cell differ- B cells are effective and equally valid, and that newly formed cells entiation, building in part on the models for are already programmed to differentiate spontaneously and development generated by Allman et al. (17, 18), is provided in quantitatively into cells that express high levels of IgM and Fig. 10. It is emphasized that this model, although supported by the high high int IgD. Long-lived IgM IgD CD21 B cells might retain analysis of loss of function mutant mice, remains speculative. T1 the FO II-like phenotype in vivo because they are not tickled by cells progress to the T2 stage and then express higher levels of cognate self-Ags. These data are consistent with the view that surface IgD to enter the FO II B cell pool. We speculate that if a one subset of long-lived follicular B cells develops in a “de- specific B cell clone encounters a self-Ag of the appropriate avid- fault”, Ag-independent manner, while the other requires cog- ity it may be selected at the T2-IgDhigh stage and differentiate nate Ag-derived and Btk-dependent signals. The absence of a further into an FO I cell possibly passing through an intermediate known cognate Ag in a transgenic animal does not rule out the T3 stage; alternatively, such a clone may be selected by self-Ag at presence of some unidentified cross-reacting self-Ag in the host the FO II stage and mature into an FO I B cell. IgDhigh B cells of animal that is recognized by the transgenic BCR. However, in the HEL knock-in model, if such a postulated self-Ag exists, it the FO II type can develop in an Ag-independent manner. Ag- is incapable of allowing differentiation of FO II B cells into FO independence is always inferred, because it is impossible to rule I B cells in vivo. out the presence of some cross-reactive, unidentified, self Ag for any BCR. Studies showing the retention of FO II cells in the ab- Discussion sence of MZP and MZ B cells in certain mutant mice strains sug- Our studies demonstrate the existence of two distinguishable long- gest, but do not prove, that FO II B cells may represent long-lived lived recirculating follicular B cell populations. FO II B cells are follicular precursors of MZ B cells. posttransitional cells that develop in a Btk-independent manner. Our in vitro differentiation studies are limited in scope and have FO I B cells require Btk, which in turn is presumably activated by focused on the crucial alterations in surface IgM and IgD that self-Ags, for their development. Relatively strong BCR signals are accompany differentiation. They were initiated primarily because already known to facilitate FO I B cell development, but are not of the lack of suitable in vivo approaches. Although it is unlikely required for FO II and MZ B cell development. Although the latter that all the phenotypic changes that occur in the specialized fol- populations can both develop in the absence of Btk-derived sig- licular environment in vivo can be recreated in vitro, the quanti- nals, MZ B cells require Notch2, while FO II B cells do not require tative shifts in the phenotypes of purified B cell populations that 2280 TWO DISTINCT POPULATIONS OF MATURE FOLLICULAR B CELLS were observed speak to the validity of using these surface markers 7. Saito, T., S. Chiba, M. Ichikawa, A. Kunisato, T. Asai, K. Shimizu, T. to obtain information regarding the differentiation process. Yamaguchi, G. Yamamoto, S. Seo, K. Kumano, et al. 2003. Notch2 is preferen- tially expressed in mature B cells and indispensable for marginal zone B lineage FO I and FO II B cells are present in all follicular compartments development. Immunity 18: 675–685. as well as in the bone marrow perisinusoidal niche (44). It is likely 8. Cariappa, A., H. C. Liou, B. H. Horwitz, and S. Pillai. 2000. Nuclear factor ␬B is required for the development of marginal zone B lymphocytes. J. Exp. Med. that these two distinguishable types of cells respond to immuno- 192: 1175–1182. gens in a broadly similar fashion, given the ability of mice, such as 9. Pillai, S., A. Cariappa, and S. T. Moran. 2004. Positive selection and lineage the MD4 anti-HEL transgenic mice, that have very few FO I B commitment during peripheral B-lymphocyte development. Immunol. Rev. 197: 206–218. cells, to readily make Abs in response to immunization with HEL. 10. Pillai, S., A. Cariappa, and S. T. Moran. 2005. Marginal zone B cells. Annu. Rev. Purified follicular B cells (deficient in FO I cells) from anti-HEL Immunol. 23: 161–196. transgenic mice have been shown to be functional (54), and it is 11. Pillai, S. 1999. The chosen few? Positive selection and the generation of naive B lymphocytes. Immunity 10: 493–502. therefore likely that FO II B cells are functional, competent B 12. Tze, L. E., B. R. Schram, K. P. Lam, K. A. Hogquist, K. L. Hippen, J. Liu, lymphocytes. A likely explanation for why two distinct follicular S. A. Shinton, K. L. Otipoby, P. R. Rodine, A. L. Vegoe, et al. 2005. Basal B cell populations have evolved is that Btk-dependent BCR signals immunoglobulin signaling actively maintains developmental stage in immature B cells. PLoS Biol. 3: E82. and Notch2 represent opposing inducers that respectively drive FO 13. Allman, D. M., S. E. Ferguson, and M. P. Cancro. 1992. Peripheral B cell mat- I B cell and MZ B cell generation, ensuring the reciprocal gener- uration, I: immature peripheral B cells in adults are heat-stable antigenhi and ation of these two distinct terminally differentiated sublineages of exhibit unique signaling characteristics. J. Immunol. 149: 2533–2540. 14. Li, Y. S., K. Hayakawa, and R. R. Hardy. 1993. The regulated expression of B B-2 B cells. FO I B cells are rarely depleted in a dramatic fashion lineage associated genes during B cell differentiation in bone marrow and fetal during the course of an infection, but -borne pathogens can liver. J. Exp. Med. 178: 951–960. cause a drastic activation and depletion of the MZ B cell pool (55, 15. Rolink, A. G., J. Andersson, and F. Melchers. 1998. Characterization of immature Downloaded from B cells by a novel monoclonal antibody, by turnover and by mitogen reactivity. 56). FO II B cells may constitute a long-lived cellular reservoir that Eur. J. Immunol. 28: 3738–3748. can potentially be recruited as it transits the spleen to replenish the 16. Loder, F., B. Mutschler, R. J. Ray, C. J. Paige, P. Sideras, R. Torres, MZ B cell pool following a severe and acute challenge by a blood- M. C. Lamers, and R. Carsetti. 1999. B cell development in the spleen takes place in discrete steps and is determined by the quality of B cell receptor-derived borne pathogen. signals. J. Exp. Med. 190: 75–89. The Ag/BCR/Btk dependent generation of FO I B cells in turn 17. Allman, D. M., S. E. Ferguson, V. M. Lentz, and M. P. Cancro. 1993. Peripheral B cell maturation, II: heat-stable antigenhi splenic B cells are an immature de- presumably ensures that not all developing B cells are driven by http://www.jimmunol.org/ velopmental intermediate in the production of long-lived marrow-derived B cells. Notch2 ligand-dependent signals into the MZ B cell pool. Strong J. Immunol. 151: 4431–4444. BCR signals may ensure that a subset of B cells is driven irrevers- 18. Allman, D., R. C. Lindsley, W. DeMuth, K. Rudd, S. A. Shinton, and R. R. Hardy. 2001. Resolution of three nonproliferative immature splenic B cell ibly to a follicular (FO I) fate and thus may be actively inhibited subsets reveals multiple selection points during peripheral B cell maturation. from committing to entry into the MZ B cell pool. It is this re- J. Immunol. 167: 6834–6840. quirement for reciprocal inhibitory events during development that 19. Merrell, K. T., R. J. Benschop, S. B. Gauld, K. Aviszus, D. Decote-Ricardo, L. J. Wysocki, and J. C. Cambier. 2006. Identification of anergic B cells within might have driven the evolution of two distinguishable and devel- a wild-type repertoire. Immunity 25: 953–962. opmentally distinct long-lived follicular B cell populations. Fur- 20. Cariappa, A., M. Tang, C. Parng, E. Nebelitskiy, M. Carroll, K. Georgopoulos, ther studies are called for to examine how strong BCR signals may and S. Pillai. 2001. The follicular versus marginal zone B lymphocyte cell fate

decision is regulated by Aiolos, Btk, and CD21. Immunity 14: 603–615. by guest on September 29, 2021 antagonize inductive signals for MZ B cell development, and to 21. Amano, H., E. Amano, T. Moll, D. Marinkovic, N. Ibnou-Zekri, E. Martinez-Soria, determine how apparently weaker BCR signals may also be re- I. Semac, T. Wirth, L. Nitschke, and S. Izui. 2003. The Yaa mutation promoting quired, as clearly suggested in the work of Wen et al. (27), during murine lupus causes defective development of marginal zone B cells. J. Immunol. 170: 2293–2301. the program of MZ B cell development. 22. Srivastava, B., W. J. Quinn, III, K. Hazard, J. Erikson, and D. Allman. 2005. Characterization of marginal zone B cell precursors. J. Exp. Med. 202: Acknowledgments 1225–1234. 23. Cariappa, A., and S. Pillai. 2002. Antigen-dependent B-cell development. Curr. We thank Michael Carroll, Michel Nussenzweig, Jason Cyster, Yoshio Opin. Immunol. 14: 241–249. Hamada, and Christopher Klug for making mice available. We thank 24. Wang, J. H., N. Avitahl, A. Cariappa, C. Friedrich, T. Ikeda, A. Renold, Joanne Yetz-Aldape, John Daley, Suzan Lazo-Kallanian, and Michelle K. Andrikopoulos, L. Liang, S. Pillai, B. A. Morgan, and K. Georgopoulos. 1998. Connole for help with flow cytometry. Sharon Germana is thanked for help Aiolos regulates B cell activation and maturation to effector state. Immunity 9: 543–554. with the proliferation assays. Jennifer Couget is thanked for help with 25. Samardzic, T., D. Marinkovic, C. P. Danzer, J. Gerlach, L. Nitschke, and microarrays. T. Wirth. 2002. Reduction of marginal zone B cells in CD22-deficient mice. Eur. J. Immunol. 32: 561–567. Disclosures 26. Moran, S. T., A. Cariappa, H. Liu, M. Beth, D. Sgroi, C. Boboila, and S. Pillai. 2007. Synergism between NFkB1/p50 and Notch2 during the development of The authors have no financial conflict of interest. marginal zone B lymphocytes. J. Immunol. 179: 195–200. 27. Wen, L., J. Brill-Dashoff, S. A. Shinton, M. Asano, R. R. Hardy, and References K. Hayakawa. 2005. Evidence of marginal-zone B cell-positive selection in 1. Singh, H., K. L. Medina, and J. M. Pongubala. 2005. Contingent gene regulatory spleen. Immunity 23: 297–308. networks and B cell fate specification. Proc. Natl. Acad. Sci. USA 102: 28. Martin, F., and J. F. Kearney. 2000. Positive selection from newly formed to 4949–4953. marginal zone B cells depends on the rate of clonal production, CD19, and btk. 2. Radtke, F., A. Wilson, G. Stark, M. Bauer, J. van Meerwijk, H. R. MacDonald, Immunity 12: 39–49. and M. Aguet. 1999. Deficient T cell fate specification in mice with an induced 29. Dai, X., Y. Chen, J. Schuman, Z. Hua, J. W. Adamson, R. Wen, and D. Wang. ␥ inactivation of Notch1. Immunity 10: 547–558. 2006. Distinct roles of phosphoinositide-3 kinase and phospholipase C 2inB- 3. Pui, J. C., D. Allman, L. Xu, S. DeRocco, F. G. Karnell, S. Bakkour, J. Y. Lee, cell receptor-mediated signal transduction. Mol. Cell. Biol. 26: 88–99. T. Kadesch, R. R. Hardy, J. C. Aster, and W. S. Pear. 1999. Notch1 expression 30. Clayton, E., G. Bardi, S. E. Bell, D. Chantry, C. P. Downes, A. Gray, in early influences B versus T lineage determination. Immunity 11: L. A. Humphries, D. Rawlings, H. Reynolds, E. Vigorito, and M. Turner. 2002. 299–308. A crucial role for the p110␦ subunit of phosphatidylinositol 3-kinase in B cell 4. Lam, K.-P., R. Kuhn, and K. Rajewsky. 1997. In vivo ablation of surface im- development and activation. J. Exp. Med. 196: 753–763. munoglobulin on mature B cells by inducible gene targeting results in rapid 31. Okkenhaug, K., A. Bilancio, G. Farjot, H. Priddle, S. Sancho, E. Peskett, death. Cell 90: 1073–1083. W. Pearce, S. E. Meek, A. Salpekar, M. D. Waterfield, et al. 2002. Impaired B 5. Schiemann, B., J. L. Gommerman, K. Vora, T. G. Cachero, S. Shulga-Morskaya, and T cell antigen receptor signaling in p110␦ PI 3-kinase mutant mice. Science M. Dobles, E. Frew, and M. L. Scott. 2001. An essential role for BAFF in the 297: 1031–1034. normal development of B cells through a BCMA-independent pathway. Science 32. Guinamard, R., M. Okigaki, J. Schlessinger, and J. V. Ravetch. 2000. Absence of 293: 2111–2114. marginal zone B cells in Pyk-2-deficient mice defines their role in the humoral 6. Tanigaki, K., H. Han, N. Yamamoto, K. Tashiro, M. Ikegawa, K. Kuroda, response. Nat. Immunol. 1: 31–36. A. Suzuki, T. Nakano, and T. Honjo. 2002. Notch-RBP-J signaling is involved in 33. Girkontaite, I., K. Missy, V. Sakk, A. Harenberg, K. Tedford, T. Potzel, cell fate determination of marginal zone B cells. Nat. Immunol. 3: 443–450. K. Pfeffer, and K. D. Fischer. 2001. Lsc is required for marginal zone B cells, The Journal of Immunology 2281

regulation of lymphocyte motility and immune responses. Nat. Immunol. 2: 45. Hardy, R. R., K. Hayakawa, D. R. Parks, and L. A. Herzenberg. 1983. Demon- 855–862. stration of B cell maturation in X-linked immunodeficient mice by simultaneous 34. Fukui, Y., O. Hashimoto, T. Sanui, T. Oono, H. Koga, M. Abe, A. Inayoshi, three colour immunofluorescence. Nature 306: 270–272. M. Noda, M. Oike, T. Shirai, and T. Sasazuki. 2001. Haematopoietic cell-specific 46. Amano, M., N. Baumgarth, M. D. Dick, L. Brossay, M. Kronenberg, CDM family protein DOCK2 is essential for lymphocyte migration. Nature 412: L. A. Herzenberg, and S. Strober. 1998. CD1 expression defines subsets of fol- 826–831. licular and marginal zone B cells in the spleen: ␤ 2-microglobulin-dependent and 35. Croker, B. A., D. M. Tarlinton, L. A. Cluse, A. J. Tuxen, A. Light, F. C. Yang, independent forms. J. Immunol. 161: 1710–1717. D. A. Williams, and A. W. Roberts. 2002. The Rac2 guanosine triphosphatase 47. Cariappa, A., T. Shoham, H. Liu, S. Levy, C. Boucheix, and S. Pillai. 2005b. The regulates B lymphocyte antigen receptor responses and chemotaxis and is re- CD9 tetraspanin is not required for the development of peripheral B cells or for quired for establishment of B-1a and marginal zone B lymphocytes. J. Immunol. humoral immunity. J. Immunol. 175: 2925–2930. 168: 3376–3386. 48. Fulcher, D. A., and A. Basten. 1997. Influences on the lifespan of B cell sub- 36. Lu, T. T., and J. G. Cyster. 2002. Integrin-mediated long-term B cell retention in populations defined by different phenotypes. Eur. J. Immunol. 27: 1188–1199. the splenic marginal zone. Science 297: 409–412. 49. Harris, D. P., L. Haynes, P. C. Sayles, D. K. Duso, S. M. Eaton, N. M. Lepak, 37. Sprent, J., M. Schaefer, M. Hurd, C. D. Surh, and Y. Ron. 1991. Mature murine L. L. Johnson, S. L. Swain, and F. E. Lund. 2000. Reciprocal regulation of B and T cells transferred to SCID mice can survive indefinitely and many main- polarized cytokine production by effector B and T cells. Nat. Immunol. 1: tain a virgin phenotype. J. Exp. Med. 174: 717–728. 475–482. 38. Agenes, F., and A. A. Freitas. 1999. Transfer of small resting B cells into im- 50. Witt, C. M., W. J. Won, V. Hurez, and C. A. Klug. 2003. Notch2 haploinsuffi- munodeficient hosts results in the selection of a self-renewing activated B cell ciency results in diminished B1 B cells and a severe reduction in marginal zone population. J. Exp. Med. 189: 319–330. B cells. J. Immunol. 171: 2783–2788. 39. Cabatingan, M. S., M. R. Schmidt, R. Sen, and R. T. Woodland. 2002. Naive B 51. Goodnow, C. C., J. Crosbie, S. Adelstein, T. B. Lavoie, S. J. Smith-Gill, lymphocytes undergo homeostatic proliferation in response to B cell deficit. R. A. Brink, H. Pritchard-Briscoe, J. S. Wotherspoon, R. H. Loblay, K. Raphael, J. Immunol. 169: 6795–6805. et al. 1988. Altered immunoglobulin expression and functional silencing of self- 40. Martin, F., and J. F. Kearney. 2002. Marginal-zone B cells. Nat. Rev. Immunol. reactive B lymphocytes in transgenic mice. Nature 334: 676–682. 2: 323–335. 52. Hippen, K. L., B. R. Schram, L. E. Tze, K. A. Pape, M. K. Jenkins, and 41. Casellas, R., T. A. Shih, M. Kleinewietfeld, J. Rakonjac, D. Nemazee, T. W. Behrens. 2005. In vivo assessment of the relative contributions of deletion, K. Rajewsky, and M. C. Nussenzweig. 2001. Contribution of receptor editing to anergy, and editing to B cell self-tolerance. J. Immunol. 175: 909–916. Downloaded from the antibody repertoire. Science 291: 1541–1544. 53. Zhang, Z., P. D. Burrows, and M. D. Cooper. 2004. The molecular basis and 42. Pape, K. A., V. Kouskoff, D. Nemazee, H. L. Tang, J. G. Cyster, L. E. Tze, biological significance of VH replacement. Immunol. Rev. 197: 231–242. K. L. Hippen, T. W. Behrens, and M. K. Jenkins. 2003. Visualization of the 54. Attanavanich, K., and J. F. Kearney. 2004. Marginal zone, but not follicular B genesis and fate of isotype-switched B cells during a primary immune response. cells, are potent activators of naive CD4 T cells. J. Immunol. 172: 803–811. J. Exp. Med. 197: 1677–1687. 55. Martin, F., A. M. Oliver, and J. F. Kearney. 2001. Marginal zone and B1 B cells 43. Cariappa, A., T. J. Kim, and S. Pillai. 1999. Accelerated emigration of B lym- unite in the early response against T-independent blood-borne particulate anti- phocytes in the Xid mouse. J. Immunol. 162: 4417–4423. gens. Immunity 14: 617–629.

44. Cariappa, A., I. B. Mazo, C. Chase, H. N. Shi, H. Liu, Q. Li, H. Rose, H. Leung, 56. Achtman, A. H., M. Khan, I. C. MacLennan, and J. Langhorne. 2003. Plasmo- http://www.jimmunol.org/ B. J. Cherayil, P. Russell, et al. 2005. Perisinusoidal B cells in the bone marrow dium chabaudi chabaudi infection in mice induces strong B cell responses and participate in T-independent responses to blood-borne microbes. Immunity 23: striking but temporary changes in splenic cell distribution. J. Immunol. 171: 397–407. 317–324. by guest on September 29, 2021