And Memory Response Post-Germinal Center B Cell

Role of MHC Class II on Memory B Cells in Post-Germinal Center B Cell Homeostasis and Memory Response

This information is current as Michiko Shimoda, Tao Li, Jeanene P. S. Pihkala and of September 27, 2021. Pandelakis A. Koni J Immunol 2006; 176:2122-2133; ; doi: 10.4049/jimmunol.176.4.2122 http://www.jimmunol.org/content/176/4/2122 Downloaded from

References This article cites 56 articles, 32 of which you can access for free at: http://www.jimmunol.org/content/176/4/2122.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 27, 2021 *average

Subscription Information about subscribing to The Journal of Immunology 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 © 2006 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Role of MHC Class II on Memory B Cells in Post-Germinal Center B Cell Homeostasis and Memory Response1

Michiko Shimoda,2*† Tao Li,* Jeanene P. S. Pihkala,* and Pandelakis A. Koni2*‡

We investigated the role of B cell Ag presentation in homeostasis of the memory B cell compartment in a mouse model where a conditional allele for the ␤-chain of MHC class II (MHC-II) is deleted in the vast majority of all B cells by cd19 promoter-mediated expression of Cre recombinase (IA-B mice). Upon T cell-dependent immunization, a small number of MHC-II؉ B cells in IA-B mice dramatically expanded and restored normal albeit delayed levels of germinal center (GC) B cells with an affinity-enhancing somatic mutation to Ag. IA-B mice also established normal levels of MHC-II؉ memory B cells, which, however, subsequently lost MHC-II expression by ongoing deletion of the conditional iab allele without significant loss in their number. Furthermore, in vivo Ag restimulation of MHC-II؊ memory B cells of IA-B mice failed to cause differentiation into plasma cells (PCs), even in the ؉ presence of Ag-specific CD4 T cells. In addition, both numbers and Ag-specific affinity of long-lived PCs during the late post-GC Downloaded from phase, as well as post-GC serum affinity maturation, were significantly reduced in IA-B mice. These results support a notion that MHC-II-dependent T cell help during post-GC phase is not absolutely required for the maintenance of memory B cell frequency but is important for their differentiation into PCs and for the establishment of the long-lived PC compartment. The Journal of Immunology, 2006, 176: 2122–2133.

ajor histocompatibility complex class II (MHC-II) - lived PCs are long-lived (8, 9) and can be maintained without http://www.jimmunol.org/ restricted B cell Ag presentation is essential for the immunizing Ag persistence (10–12). Other studies also suggested M induction of Ag-specific humoral immune response that memory B cells were maintained in mice deprived of CD4ϩ T and generation of immunological memory. Upon infection or im- cells (13) or in mice deficient in complement receptor 2 (Cr2) (14). munization with T cell-dependent (TD) Ag, Ag-specific B cells Furthermore, recent studies have shown that polyclonal stimula- ϩ present Ag to Ag-specific CD4 T cells primed by dendritic cells tion such as CpG DNA or bystander T cell help can provide pro- (DCs), and this cognate T cell-B cell interaction promotes B cell liferation and even differentiation of memory B cells (11, 12). In proliferation and differentiation either into extrafollicular plasma addition, virus-specific memory B cells can be activated and dif- cells (PCs) or germinal center (GC) B cells (1, 2). During ensuing ferentiated into PCs in the absence of CD4ϩ T cells (15). Thus, massive proliferation, GC B cells undergo somatic mutations in the these studies support a notion that established memory B cells are by guest on September 27, 2021 V region of their Ig BCR, which can change the affinity of the BCR mainly maintained independent of Ag itself or Ag-specific T cell to Ag. High-affinity B cells survive by preferential interaction with help, and their differentiation into PCs is CD4ϩ T cell independent. follicular DCs and T cells, while low-affinity B cells die by apoptosis In contrast, there are also findings suggesting that even long-lived (3). This GC selection process is responsible for the generation of PCs are replenished by differentiation of high-affinity precursors “non-Ab-secreting” memory B cells and “Ab-secreting” long-lived during the post-GC phase, reflected in post-GC serum Ig affinity PCs in bone marrow (BM) (4), both of which are necessary for long- maturation, which itself is dependent on Ag-deposition (5, 16–20). term immunological memory and protection (4–6). Irradiated Cr2Ϫ/Ϫ chimeric mice reconstituted with Ag-primed B In contrast to the extensive studies on GC B cell differentiation cells and T cells showed reduced frequency of memory B cells and and selection, the mechanisms delineating memory B cells and long-lived PCs from GC B cells and the maintenance of these two Ab-secreting cells (ASCs) (19). In addition, post-GC B cells failed B cell populations long after their generation are largely unknown. to induce Blimp-1, XBP-1, and Bcl-2, which resulted in failure to Previous studies revealed that both memory B cells (7) and long- generate the precursor population of long-lived PCs (14). These results strongly suggest that there is an Ag-dependent mechanism for the maintenance of long-lived PC compartment by differenti- *Program in Molecular Immunology, Immunotherapy Center, †Department of Pathol- ation of precursors, which is required for the maximum retention ‡ ogy, and Department of Medicine, Medical College of Georgia, Augusta, GA 30912 of long-term protective humoral immunity. Received for publication August 17, 2005. Accepted for publication November Classically, Ag-specific B220ϩ memory B cells are defined as 25, 2005. class-switched IgDϪIgMϪ B cells (21, 22) and highly express 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 CD38 (23, 24). However, several other studies now also show that with 18 U.S.C. Section 1734 solely to indicate this fact. Ag restimulation promotes B220ϩCD138Ϫ memory B cells to Ϫ Ϫ 1 This work was supported in part by Medical College of Georgia Combined Intra- generate B220 CD138 preplasma memory B cells that are also mural Grants Program grants-in-aid (to M.S. and P.A.K.). maintained in the post-GC phase and differentiate into PCs upon 2 Address correspondence and reprint requests to Dr. Michiko Shimoda or Dr. Pan- Ag recall (6, 25, 26). In addition, by using a transgenic BCR delakis A. Koni, Program in Molecular Immunology, Immunotherapy Center, De- partment of Medicine, Medical College of Georgia, 1120, 15th Street, Augusta, GA model, precursors for long-lived PCs were identified in BM, which 30912. E-mail address: [email protected] or [email protected] proliferate and differentiate into PCs in the absence of Ag (18). It 3 Abbreviations used in this paper: MHC-II, MHC class II; PC, plasma cell; GC, is not clear at present whether memory B cells and long-lived PCs germinal center; CGG, chicken gammaglobulin; NP, (4-hydroxy-3-nitrophenyl)acetyl; ASC, Ab-secreting cell; TD, T cell dependent; TI, T cell independent; BM, bone are largely maintained independently as a separate compartment marrow. once derived from GC B cells (4) or whether, in fact, these B cell

subsets in the post-GC B cell compartment sequentially or inde- 6.7), CD90.1 (HIS51), TCR␤ (H57-597), CD19 (1D3), CD138 (281-2), pendently act as precursors of long-lived PCs (6). To support the 120G8 (Ref. 32; provided by Dr. G. Trinchieri, Schering-Plough Research ␣b latter, it has been suggested that virus-speciﬁc long-term Ab mem- Institute, Kenilworth, NJ), IA (Y3P; provided by our colleague Dr. L. ϩ Ignatowicz, Medical College of Georgia, Augusta, GA), and PNA-FITC ory is provided by Ag-driven and CD4 TD continuous differen- (Vector Laboratories). Abs were from BD Biosciences unless otherwise

tiation of memory B cells into short-lived PCs (20). Therefore, indicated. NP-haptenated PE (NP20-PE) was prepared as described else- these post-GC B cells are considered to be points of selection of where (31). Cells were analyzed on a FACSCalibur (BD Biosciences) with high-affinity clones responsible for post-GC serum affinity matu- CellQuest software (BD Biosciences). Ag-specific GC B cells and B220ϩ memory B cells were identified by ration, and their terminal differentiation into PCs is controlled by five-color analysis as previously reported (33, 34). Briefly, cells were certain mechanisms to achieve maximal affinity in the long-lived stained on ice for 45 min with a mixture of biotinylated specific Abs to PC compartment. IgM, IgD, CD43 (S7), CD5, Gr-1 (RB6-8C5), CD11b, CD49b (DX5), and In this article, we study the role of MHC-II-restricted B cell Ag CD90.2 (30-H12) to exclude naive B cells, PCs, B-1 cells, macrophages, presentation for memory B cell maintenance and their terminal NK T cells, and T cells (lineage depletion Ab mixture). For analysis of secondary response, cell-bound Igs were stripped by incubating spleen differentiation into PCs during memory response with a new cells at 37°C for 30 min followed by washing twice with PBS/1% FCS mouse model of B cell-restricted MHC-II deficiency by crossing before staining with specific Abs (35) to exclude Ab-capturing non-B cells the cd19cre allele (27) onto an iabneo/neo mouse background (28) that masquerade as memory B cells (36). neo/neo cre/ϩ neo ␤b (iab cd19 : IA-B mice). The conditional iab allele was After washing, cells were stained with anti-IA -FITC, NP20-PE, streptavidin-PE-Cy5, anti-CD38-allophycocyanin (clone 90; eBioscience) generated such that loxP sites flank the promoter and the first exon and B220-allophycocyanin-Cy7. Cells were washed and finally suspended of the iab gene encoding IA␤ chain of MHC-II (28). Cre recom- in 5 ␮g/ml 7-aminoactinomycin-D (Sigma-Aldrich) for analysis with a binase-mediated deletion of this iabneo allele is known to cause MoFlo (DakoCytomation). At least 2,000,000 events were collected, and Downloaded from ϩ loss of cell surface MHC-II expression (28), as is observed in the the frequency of NP-binding GC B cells, B220 memory B cells in the case of conventional IA␤ chain knockout mice (29). Despite effi- viable lymphocyte gate was determined with Summit software version 4 (DakoCytomation). cient MHC-II deletion from the vast majority (ϳ95%) of B cells, these IA-B mice generated normal numbers of GC B cells and Cell sorting memory B cells upon immunization with TD Ag by dramatic ex- ϩ To enrich memory B cells, spleen cells were pretreated with Fc-block and pansion of a small number of MHC-II B cells. However, these incubated for 45 min with a lineage depletion biotinylated Ab mixture (as http://www.jimmunol.org/ memory B cells subsequently lost MHC-II due to ongoing cd19cre- described in Flow cytometry) including anti-CD138 Ab in RPMI 1640/2% Ϫ b mediated deletion of the conditional iabneo allele without loss in FCS. To enrich only MHC-II memory B cells, biotinylated anti-IA␤ Ab was also included in the mixture. After washing, cells were incubated with number. With this unique mouse model, we demonstrate in this ϩ anti-biotin microbeads (Miltenyi Biotec). Biotin cells were depleted with study the impact of MHC-II on B cells during GC differentiation an AutoMACS (Miltenyi Biotec). For isolation of NP-specific memory B and further address the requirements for MHC-II-restricted B cell cells, the enriched cells were incubated with a mixture of anti-IA␤b-FITC,

NP20-PE, streptavidin-PE-Cy5, anti-CD38-allophycocyanin, and B220- Ag presentation in homeostasis of memory B cell and long-lived Ϫ ϩ PCs and memory response. allophycocyanin-Cy7. Among the NP-binding/lineage /B220 cells, CD38ϩ cells, and CD38Ϫ cells were purified by MoFlo as B220ϩ memory B cells and GC B cells, respectively. Materials and Methods by guest on September 27, 2021 Animals and immunization Adoptive transfer ϩ Ϫ C57BL6/J mice were purchased from The Jackson Laboratory and were MHC-II or MHC-II B cells were enriched from control mice or IA-B bred in our facility. Mice with a conditional loxP-targeted IA␤ chain mice after depleting non-B cells from spleen cells by magnetic cell sorting (iabneo) allele were described previously (28). The deleted (null allele) or complement lysis. For magnetic cell sorting, B cells were incubated with version (iab⌬) was created by crossing a male iabneo/neo mouse with a a lineage depletion biotinylated Ab mixture (as described in Flow cytom- Ϫ ␤b female TIE2Cre mouse (30) and will be described elsewhere. Mice that etry). To enrich only MHC-II memory B cells, biotinylated anti-IA Ab was also included in the mixture. For complement lysis, spleen cells were lack MHC-II specifically from B cells were generated by interbreeding ␮ iabneo/neo mice and iab⌬/⌬ mice with cd19cre/ϩ mice. Mice were genotyped incubated with 10 g/ml anti-CD90.2 Ab and anti-NK.1 Ab for 30 min on for cre by PCR with primers described elsewhere (30). B cell-specific ice. After washing, the cells were incubated with rabbit complement for 30 min at 37°C then washed three times with RPMI 1640/10% FCS. CD4ϩ T MHC-II deletion was confirmed by analyzing B cells in peripheral blood ␮ by flow cytometry at 6 wk of age. All experiments used iabneo/neocd19cre/ϩ cells were enriched from C57BL6/J mice that had been primed with 50 g (IA-B) mice and their cd19cre-negative littermates or iabneo/⌬cd19cre/ϩ of CGG plus alum 30 days before by using magnetic cell sorting with ϩ ϫ 7 ϫ mice and their cd19cre-negative littermates. (4-Hydroxy-3-nitrophenyl)a- CD4 magnetic beads (Miltenyi Biotec). B cells (2 10 ) and T cells (2 106) were injected i.v. into RAG-1Ϫ/Ϫ mice and, 24 h later, were chal- cetyl (NP13)-chicken gammaglobulin (CGG) was prepared as described ␮ elsewhere (31). Mice 7–10 wk of age were given an i.p. challenge with 100 lenged i.p. with 50 g of soluble NP-CGG. Anti-NP response was esti- ␮ ␮ ␮ mated by ELISA and FACS at day 10 days after the challenge. l of PBS containing 5 or 50 gofNP13-CGG absorbed in alum or 50 g of NP -Ficoll (Bioresearch Technologies). For secondary immunization, 24 Histology mice were given an i.p. challenge with 100 ␮l of PBS containing 50 ␮gof ␮ soluble NP13-CGG or 20 gofNP24-Ficoll. Blood was collected from Spleens were snap frozen in OCT embedding compound in a dry ice/ immunized mice by tail vein bleeding for serum Ab determination at var- methylbutane bath. Frozen spleen tissue sections of 7-␮m thickness were ious times. All mice were maintained under specific pathogen-free condi- prepared with a cryostat microtome (Leica Microsystems), fixed in cold tions and all the studies have been reviewed and approved by an appro- acetone for 20 min, air-dried, and stored at Ϫ80°C until staining. Thawed priate institutional committee. sections were rehydrated in PBS and then preblocked before being incu- ␤b ␣ bated with either anti-I-A -PE, NP20-PE, anti-CD8 -allophycocyanin, or Flow cytometry anti-CD19-PE, and biotin-anti-IgD followed by streptavidin Alexa 488 (Molecular Probes). Images were acquired using an LS51confocal micro- Single-cell suspensions were prepared from spleens by mechanical disrup- scope system (Zeiss). tion of small fragments of organ between frosted glass slides followed by depletion of RBC with ACK lysing buffer (BioWhittaker). Spleen cells ELISA and ELISPOT assays were then washed twice with PBS and filtered through nylon mesh in RPMI 1640/1% FCS. For general analysis, single cell suspensions of ELISA and ELISPOT assays were performed as previously described (34). spleen were pretreated with Fc-block (2.4G2; anti-CD16/CD32; BD Bio- Total and high-affinity NP-specific Abs in sera were measured by using a sciences) on ice for 30 min. Cells were then incubated on ice for 30 min Clonotyping System-AP with pNPP substrate (Southern Biotechnology ␤b with specific Abs to IA (AF6-120.1), CD4 (RM4-5), CD11c (HL3), Associates) using plates coated with NP15-BSA and NP2-BSA, respec- CD11b (M1/70), CD5 (53-7.3), B220 (RA3-6B2), IgM (R6-0.2), CD21/ tively. The reciprocal endpoint titer for total (NP15-BSA) and high-affinity ␣ CD35 (7G6), CD23 (B3B4), CD24 (M1/69), IgD (217-170), CD8 (53- (NP2-BSA) NP-specific Abs in sera was defined as the dilution of serum 2124 MHC-II-DEPENDENT PC DIFFERENTIATION OF MEMORY B CELLS giving an OD at 405 nm of 0.05. Pooled sera from NP-CGG-immunized Results control mice between 12 and 20 wk postimmunization were used as a Generation of mice lacking MHC-II on B cells standard control. Similarly, total and high-affinity NP-specific ASCs in spleen and BM were determined by 3-h culture on plates coated with NP15- We generated mice that lack MHC-II specifically from B cells by BSA and NP - cre neo/neo 2-BSA, respectively, followed by detection with a Clonotyp crossing the cd19 allele (27) onto an iab mouse back- ing System-AP as above but with a Liquid BCIP/NBT Substrate Kit neo/neo cre/ϩ (Zymed Laboratories). ground (28) (iab cd19 : IA-B mice). The conditional iabneo allele was generated such that loxP sites flank the promoter Sequence analysis of VDJ DNA segments and the first exon of the iab gene encoding IA␤ chain of MHC-II neo VDJ sequences were determined as previously described (34). Total RNA (28). Cre recombinase-mediated deletion of this iab allele is was prepared from sorted TRIzol-solubilized GC B cells and memory B known to cause loss of cell surface MHC-II expression (28), as is cells from pooled spleen cells of at least three mice at each group. First- observed in the case of conventional IA␤ chain knockout mice strand cDNA was synthesized with an oligo(dT) primer by use of a Su- ϩ perscript Kit (Invitrogen Life Technologies). Two microliters of cDNA (29). Selective loss of cell surface MHC-II on B220 B cells was solution was used as a template in a reaction volume of 50 ␮l for two confirmed by staining MHC-II ␣-chain or ␤-chain, with no appar- rounds of nested PCR for amplifying the V186.2 gene rearranged to the ent loss of MHC-II among CD11cϩ DCs (Fig. 1A). Both CD8␣ϩ JH2 region by use of Pfu DNA polymerase (Stratagene). PCR products and CD8␣Ϫ subpopulations among B220ϪCD11chigh DCs (data were size-fractionated by agarose gel electrophoresis and purified with a low ϩ PCR Purification Kit (Qiagen). The purified fragments were cloned into the not shown) as well as CD11c 120G8 plasmacytoid DCs (32) PCR-Script Amp cloning vector (Stratagene) according to the manufactur- (Fig. 1B) in IA-B mice showed normal MHC-II expression. Loss er’s instructions, and the sequence of the VDJ segment was determined. of MHC-II was also observed among peritoneal B-2 ϩ Ϫ Ϫ low ϩ Ϫ Assignment of gene usage and somatic mutations was performed with the (B220 CD11b CD5 ), B-1a (B220 CD11b CD5 ), and B-1b Downloaded from BLAST and CLUSTALW programs provided by the DNA Data Bank of ϩ ϩ Ϫ Japan (͗www.ddbj.nig.ac.jp/͘). (B220 CD11b CD5 ) B cells (Fig. 1C). The fraction of B220ϩ B cells in 6- to 8-wk-old naive IA-B mice Statistical analysis that lost MHC-II was 95.9 Ϯ 2.3% (n ϭ 10) with no apparent The Student t test (two-tailed) was used. A value of p Ͻ 0.05 was con- variation in mice even up to 6 mo of age (data not shown). This sidered to indicate a significant difference. loss was further increased to 97.2 Ϯ 1.0% (n ϭ 18) in mice bearing http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 1. Selective loss of MHC-II expression on B cells in IA-B mice. A, Cell surface MHC-II on B220ϩCD11cϪ B cells and B220ϪCD11cϩ DCs in spleen was stained with Abs speciﬁc to MHC-II ␣-chain (IA␣)or␤-chain (IA␤) (bold open histograms), with anti-CD90.1 as an isotype control (gray histograms). C57BL/6J mice and iabneo/neo mice express MHC-II on B cells and DCs, whereas iab⌬/⌬ mice lack MHC-II from both B cells and DCs. IA-B (iabneo/neocd19cre/ϩ) mice lack MHC-II selectively from B cells. B, MHC-II expression on spleen plasmacytoid DCs (CD11clow120G8ϩ) was compared between wild-type mice (thin line) and IA-B mice (bold line) by staining with anti-IA␤-PE, using anti-TCR␤-PE as a negative control (gray histogram). C, MHC-II expression on peritoneal B-2 (B220ϩCD11bϪCD5Ϫ), B-1a (B220lowCD11bϩCD5ϩ), and B-1b (B220ϩCD11bϩCD5Ϫ) cells (56) in wild-type mice and IA-B mice is shown by staining with anti-IA␤b-FITC (ﬁlled histograms) using anti-CD90.1-FITC as an isotype control (open histograms). D, CD19 expression was compared between MHC-IIϩ (thin line) and MHC-IIϪ (bold line) B cells and B220Ϫ CD4ϩ T cells (gray histogram) in the spleen of IA-B mice. Data in A–D are representative of one of at least three mice per group. The Journal of Immunology 2125

one MHC-II null allele (i.e., iabneo/⌬cd19cre/ϩ mice). The remain- anti-NP IgM, IgG3, and IgG1 response during the first 2 wk postim- ing MHC-IIϩ B cells in IA-B mice had a level of CD19 expression munization (Fig. 2A). To investigate TD response, IA-B mice and similar to that of the MHC-IIϪ B cells (Fig. 1D), suggesting that control mice were immunized with 50 ␮g of NP-CGG with alum as Cre recombinase (i.e., cd19cre), which is under the control of the an adjuvant. Age-matched littermate (iabneo/neocd19ϩ/ϩ) mice were de novo CD19 genomic locus (27), was at least being expressed. used as control mice because heterozygous expression of CD19 due to Among the MHC-IIϩ B cells, 70% were IgMϩIgDϩCD38ϩB220ϩ the presence of cd19cre in IA-B mice did not affect Ag-specific Ab cells, indicative of mature virgin B cells. It is possible that CD19ϩ production and serum Ig affinity maturation (data not shown). naive B cells might have deleted the iabneo allele but still ex- In IA-B mice, anti-NP serum titer of IgG1 and all other Ig iso- press MHC-II protein on their surface. The rest were types were Ͼ10-fold defective during the first 2 wk postimmuni- IgMϩIgDϪB220ϩCD38dull/Ϫ and highly expressed MHC-II, all of zation (Fig. 2B). Even the anti-NP IgM titer, which is the primary which is consistent with an activated B cell phenotype. Fur- Ig produced independent of class-switching, was defective. The thermore, GC B cells chronically developing in Peyer’s patches numbers of IgG1 and IgM ASCs in spleen were also greatly re- of IA-B mice contained CD38dull/Ϫ B cells highly expressing duced in IA-B mice (data not shown). Thus, Ag-specific Ab pro- MHC-II (data not shown). In IA-B mice, a small subset of MHC-IIϪ duction in the early primary TD response was severely impaired in B cells also had an activated phenotype, but at ϳ10-fold lower fre- IA-B mice. However, the IgG1 titer dramatically increased by 100- quency than among MHC-IIϩ B cells (data not shown). Therefore, it fold by 4 wk postimmunization and reached a level similar to that ϩ is possible that MHC-II B cells in IA-B mice were preferentially of control mice by 8 wk postimmunization (Fig. 2B). The strength activated in a TD manner by environmental Ag. and kinetics of Ag-specific IgG1 response also seemed to be pro- ϩ The numbers of B cells, DCs, marginal zone B cells, T-1 and portional to the frequency of MHC-II B cells. That is, IA-B mice Downloaded from ϩ T-2 immature B cells, as well as CD4 T cells in the spleen of and iabneo/⌬cd19cre/ϩ mice carrying Ͼ2% MHC-IIϩ B cells before IA-B mice were normal, and there was no apparent histological immunization (as judged by flow cytometry of peripheral blood at abnormality in follicular structure and no difference in serum Ig 6 wk of age) achieved normal levels of anti-NP IgG1 response by titer for all the isotypes between wild-type and IA-B mice at 8–12 8 wk postimmunization despite the severe defect in anti-NP titer wk of age (data not shown). during the first 2 wk of early primary response (Figs. 2B and 3). In neo/⌬ cre/ϩ ϩ contrast, iab cd19 mice that had Ͻ2% MHC-II B cells http://www.jimmunol.org/ IA-B mice develop delayed but otherwise normal humoral had undetectable levels of Ag-specific IgG1 during the first 2 wk response in association with GC formation postimmunization and were never able to mount a robust response When immunized with a T cell-independent (TI) Ag, NP-hapte- even by 8 wk postimmunization (Fig. 3). In fact, the deletion levels ϩ ϩ nated Ficoll (NP24-Ficoll), IA-B mice developed relatively normal of MHC-II among naive (IgM IgD ) NP-binding cells of IA-B by guest on September 27, 2021

FIGURE 2. Early TD but not TI immune response is impaired in IA-B mice. A, IA-B mice and control mice ␮ were immunized with 50 gofNP24-Ficoll and sera were collected at each time point. Endpoint titers of serum anti-NP Ig of each isotype for control mice (E) and IA-B mice (F) were deﬁned as the dilution of serum giving an OD at 405 nm of 0.05 by ELISA. Data shown are representative of one of three experiments with at least three mice per time point. B, Similar to A, except IA-B mice and control mice were immunized with 50 ␮g of TD Ag, NP-CGG with alum as an adjuvant. Data shown are representative of one of three experiments with at least three mice per time point. 2126 MHC-II-DEPENDENT PC DIFFERENTIATION OF MEMORY B CELLS

Thus, only the GC B cells expressed MHC-II, whereas the vast majority of IgDϩ follicular naive B cells remained MHC-II deﬁ- cient in IA-B mice.

Ongoing MHC-II deletion does not affect the frequency of established memory B cells During GC reaction, high-affinity GC B cells are selected and differentiate into memory B cells and long-lived PCs (4–6). To fur- FIGURE 3. Mice having Ͼ2% MHC-IIϩ B cells develop normal levels ther study whether or not IA-B mice can restore and maintain of anti-NP humoral immune response upon immunization with TD Ag. normal numbers of GC and memory B cells, the kinetics of dif- ϩ Various iabneo/⌬cd19cre/ϩ mice having slightly different frequencies of ferentiation of GC and B220 memory B cells in IA-B mice and MHC-IIϩ B cells (as determined by FACS of peripheral blood cells at 6 wk the expression levels of MHC-II in these populations were fol- of age) and wild-type control mice were immunized with NP-CGG in alum, lowed by FACS. After gating out non-B cells, B-1, and plasma B and endpoint anti-NP IgG1 titers in sera were defined as the dilution of cells, as well as naive (IgMϩ/IgDϩ) B cells with lineage markers E F serum giving an OD at 405 nm of 0.05 by ELISA at 4 ( )and8( )wk (CD90.2, Gr-1, CD49b, CD11b, CD5, CD43, IgD, and IgM), the postimmunization. The endpoint titer for each individual mouse was then ϩ Ϫ ϩ ϩ ϩ B220 CD38 GC and B220 CD38 memory B cells (23, 24, 33, plotted against the frequency of MHC-II B cells in each mouse, determined at 6 wk of age before immunization. 34) were identified among NP-binding B cells (Fig. 5A). IA-B

mice proved to have delayed kinetics of NP-binding GC B cell Downloaded from generation (Fig. 5B). This did not appear to be a result of an in- mice determined by flow cytometry were similar to or even greater trinsic difference in NP-specific precursor frequency between con- than that in the total B cell population at 98.0 Ϯ 0.39% (n ϭ 5). trol mice and IA-B mice but a difference in the frequency of MHC- Therefore, the analysis leads to the estimation that as little as 2% IIϩ cells among the anti-NP repertoire because IA-B mice and of the normal NP Ag-specific B cell population can restore a nor- control mice had a similar frequency of NP-binding naive ϩ ϩ ϩ ϳ mal level of humoral response. IgM IgD B220 cells ( 0.03% of total spleen cells) before im- http://www.jimmunol.org/ In addition, the ratio of high-affinity spleen ASCs dramatically munization. However, IA-B mice restored a maximum frequency increased at 4 wk postimmunization (data not shown), suggesting of GC B cells similar to or even higher than that in control mice GC development in IA-B mice. In fact, PNAhighIgDϪ GCs were between 2 and 10 wk postimmunization (Fig. 5B). Also, genetic clearly stained in spleen by 2 wk after immunization both in con- analysis of V186.2DJ rearrangements, which dominate in anti-NP trol and IA-B mice. All of the IgDϪ GCs but not IgDϩ follicular response in C57BL/6 background (37), showed that clones with B cells in IA-B mice were positively stained with Ab against somatic mutations as well as high-affinity clones carrying an af- MHC-II (IA-␤) (Fig. 4). GCs in IA-B mice were grossly normal finity-enhancing Try to Leu mutation at amino acid position 33 and clearly stained with CD19 as in control mice (data not shown). (38, 39) were enriched among GC B cells of IA-B mice at 4 wk by guest on September 27, 2021

FIGURE 4. MHC-IIϩ GC development in IA-B mice. Frozen sections of spleen from control mice (A) and IA-B mice (B) at 2 wk postimmunization with NP-CGG in alum were stained with anti-IgD-biotin and streptavidin Alexa 488 (green), anti-IA␤-PE (red), and anti-CD8␣-allophycocyanin (blue). T cell areas are stained as CD8␣ϩ. In IA-B mice, a typical IgDϪ GC (indicated as GC) is clearly stained with anti-IA␤-PE, whereas IgDϩ follicular B cells are not. In a merged image, the GC is stained in red in both control mice and IA-B mice, whereas follicular B cells stained in red in IA-B mice and yellow in control mice. Data represent results from one of at least three mice per group. The Journal of Immunology 2127 Downloaded from http://www.jimmunol.org/

FIGURE 5. Memory B cells in IA-B mice lose MHC-II expression without significant loss of frequency. A, Representative flow cytometry profile of spleen cells from a control mouse 2 wk postimmunization with 50 ␮g of NP-CGG in alum. Among lineage-negative (CD90.2, Gr-1, CD49b, CD11b, CD5, ϩ dull/Ϫ CD43, IgD, and IgM) B220 B cells (R1), NP-binding cells identified with NP20-PE were separated by differential expression of CD38 as CD38 GC cells (R2) and CD38ϩ memory B cells (R3). B, The frequency of NP-binding GC B cells of control mice (E) and IA-B mice (F) was analyzed as in A at the indicated weeks postimmunization. Data represent the average of at least four mice per time point. NP-binding cells were not detectable (Ͻ5 ϫ 10Ϫ5% ϩ ϩ of total spleen cells) in the lineage-negative B220 cell population in naive mice (i.e., wk 0). C, The percentage of MHC-II B cells among NP-binding by guest on September 27, 2021 GC B cells was determined at the indicated weeks postimmunization. Data represent the average of at least four mice per time point. D, Similar to B, the frequency of NP-binding memory B cells was analyzed. E, Similar to C, the percentage of MHC-IIϩ B cells among NP-binding memory B cells was analyzed. F, Representative flow cytometry profiles of spleen cells from wild-type mice (left) and IA-B mice (right) at 4, 12, 20, and 30 wk postimmunization as in A. Histograms show MHC-II IA␤b expression on memory B cells (gray plot) as well as anti-CD90.1 isotype control staining (empty plot). postimmunization (Table I). Thus, despite the delayed kinetics, memory B cells. However, at 4 wk postimmunization, only ϳ75% accumulation of somatic mutations and GC selection in IA-B mice of memory B cells in IA-B mice expressed MHC-II (Fig. 5F), seemed to be normal. Also, Ͼ93% of the GC cells in IA-B mice which was somewhat lower than that among GC B cells. This led expressed MHC-II (Fig. 5C), emphasizing the strict requirement us to speculate that the cd19cre was still deleting the iabneo allele for MHC-II expression on B cells for GC differentiation. in memory B cells, and that cell surface MHC-II was slowly being The frequency of NP-binding B220ϩCD38ϩ memory B cells in lost. Indeed, the fraction of memory B cells that were MHC-IIϩ immunized IA-B mice and control mice was ϳ0.01% of total continuously decreased during the analysis and MHC-II expres- spleen cells and did not differ significantly throughout the analysis sion was almost completely lost from memory B cells of IA-B (Fig. 5D). Thus, IA-B mice also generated normal numbers of mice by 20 wk postimmunization (Fig. 5, E and F). The loss of

Table I. Summary of NP-binding B cells with GC (CD38Ϫ) or memory (CD38ϩ) phenotype

Week Phenotype of Puriﬁed Number of Number of Number of Mutation Postimmunization NP-Binding B Cells Mice Clonesa L33Y/99b W33/G99b Averagec

Control 5/8 4 0 2.5 2GC LA-B 6/9 0 0 5.3 Control 6/8 0 4 7.9 4GC IA-B 10/14 10 0 11.3 MHC-IIϩ memory Control 8/10 0 6 4.8 30 Ϫ MHC-II memory IA-B 10/12 0 6 5.0

aNumber of mutated clones per number of recovered V186.2 clones. GC cells (1 ϫ 105 cells) and memory B cells (2 ϫ 104 cells) sorted from spleen cells pooled from three to six mice were used for preparation of RNA. At least 15 clones were analyzed for each time point, and only the clones with unique sequence were subjected to the analysis. bNumber of clones with a replacement mutation from tryptophan to leucine (L) (the L33/Y99 group) and those carrying a glycine residue at position 99 in CDR3 with a number of somatic mutations except for the L33 replacement (the W33/G99 group) generally have high affinity to NP (33, 38, 39). The L33/Y99 group appears to be enriched among NP-specific B cells in the early primary response while the W33/G99 group becomes dominant in the NP-specific memory B cell compartment in the late primary response (33, 39). cAverage mutation frequency for all V186.2 clones. 2128 MHC-II-DEPENDENT PC DIFFERENTIATION OF MEMORY B CELLS

MHC-II remained a consistent phenotype among B220ϩCD38ϩ memory B cells in IA-B mice recovered at 30 wk postimmunization, without significant loss in the memory B cell numbers (Fig. 5, E and F). MHC-IIϪ memory B cells of IA-B mice expressed a similar level of CD19 compared with MHC-IIϩ memory B cells of control mice at 30 wk postimmunization (data not shown). How- ever, it should be noted that 4.8 Ϯ 1.3% (n ϭ 4) of the memory B cells of IA-B mice remained MHC-IIϩ even at 30 wk postimmunization (Fig. 5, E and F), although the level of MHC-II expression among these MHC-IIϩ cells estimated by mean fluorescence in- tensity was reduced compared with that in control mice (165 Ϯ 59 (n ϭ 4) and 337 Ϯ 45 (n ϭ 4), respectively ( p ϭ 0.004)). Genetic analysis showed that comparable frequencies of high- affinity clones, which are generally recovered from the late memory B cell compartment (33, 39), were recovered at 30 wk postimmunization from both purified MHC-IIϪ and MHC-IIϩB220ϩ memory B cells of IA-B mice and control mice, respectively (Ta- ble I). Thus, these results suggest that once selected to differentiate into memory B cells, MHC-II expression (i.e., cognate T cell-B Downloaded from cell interaction) is no longer required for the maintenance of B220ϩ memory B cell frequency. Impaired establishment of long-lived PC compartment in IA-B mice

To study whether or not IA-B mice established and maintained a http://www.jimmunol.org/ normal long-lived PC compartment, affinity maturation of serum anti-NP IgG1 and BM anti-NP IgG1 ASCs in IA-B mice was followed. When mice were immunized with a low dose (5 ␮g) of NP-CGG rather than a high dose (50 ␮g) as above, IA-B mice failed to show normal levels of Ag-specific serum Ig titer (Fig. 6A), and post-GC serum affinity maturation was not evident between 8 and 20 wk postimmunization even in control mice (Fig. 6B). Thus, consistent with other studies (17, 19), Ag supply was by guest on September 27, 2021 critical for post-GC affinity maturation. FIGURE 6. Post GC affinity maturation of Ag-specific serum Ab and In association with GC development, NP-specific serum Ig titer generation of high-affinity ASCs in BM is impaired in IA-B mice. A, IA-B in IA-B mice challenged with 50 ␮g of NP-CGG was normal albeit mice (F) and control mice (E) were immunized with 5 ␮g of NP-CGG with slightly delayed kinetics (Figs. 2 and 6C). Serum Ig affinity with alum as an adjuvant and sera were collected at each time point shown. maturation in IA-B mice was also normal until 8 wk postimmu- Endpoint titers of anti-NP IgG1 Ab in sera defined as the dilution of serum nization (Fig. 6D). However, although the ratio of high-affinity giving an OD at 405 nm of 0.05 by ELISA were determined with NP15- IgG1 in control mice kept increasing between 8 and 30 wk postim- BSA-coated plates (detecting high-affinity Abs) and NP2-BSA-coated munization, that in IA-B mice did not (Fig. 6D). The total number plates (detecting total Abs). Data represent one of four similar experiments with the average of at least four mice per time point. B, Serum Ig affinity of ASCs in BM (i.e., long-lived PCs) was significantly lower in maturation in the same mice shown in A was estimated as the ratio of IA-B mice at around 4 wk postimmunization, in association with high-affinity anti-NP IgG1 Ab captured with NP -BSA-coated plates vs delayed GC differentiation, but recovered to a level similar to that 2 total anti-NP IgG1 Ab captured with NP15-BSA-coated plates. C, Similar of control mice as the frequency of GC B cells in IA-B mice to A, except mice were immunized with 50 ␮g of NP-CGG with alum. D, recovered to a level even higher than that of control mice during Similar to B, serum Ig affinity maturation in the same mice shown in C was 8–12 wk postimmunization. However, total numbers of long-lived estimated. E, The numbers of anti-NP ASCs in BM of mice immunized PCs as well as total anti-NP IgG1 serum titers in IA-B mice re- with 50 ␮g of NP-CGG plus alum were measured by ELISPOT at various duced to ϳ20% of the wild-type level between 20 and 30 wk time points. Data represent the average of three to eight mice per time postimmunization (Fig. 6, C and E). In addition, continuous accu- point. F, The ratio of high-affinity anti-NP IgG1 BM ASC captured with NP -BSA-coated plates vs total anti-NP IgG1 BM ASC captured with mulation of high-affinity PCs into the long-lived PC compartment 2 NP -BSA-coated plates in the same mouse shown in E was measured by in BM, which was obvious in control mice, seemed to be absent in 15 .(Statistically significant (p Ͻ 0.05 ,ء .ELISPOT at various time points IA-B mice during the post-GC phase (Fig. 6F). Thus, the post-GC B cell differentiation pathway responsible for accumulation of high-affinity long-lived PCs seems to be impaired in IA-B mice. memory B cells still expressed MHC-II (Fig. 5E), IA-B mice generated a relatively normal memory response with increased high- Ag stimulation alone cannot induce differentiation of memory B affinity anti NP-IgG1 titers (data not shown). Furthermore, IA-B cells mice generated robust memory response at 20ϳ30 wk postimmu- The absolute requirement of cognate T cell help in PC differenti- nization even after the majority (ϳ96%, Fig. 5E) of memory B ation of memory B cells is still questionable (15). If cognate T cell cells lost MHC-II expression. The anti-NP IgG1 titer in both con- help is required, Ag-specific memory response in IA-B mice, trol mice and IA-B mice dramatically increased by day 7 after which lost MHC-II from memory B cells, would be defective. secondary challenge, and was even more remarkable in IA-B mice When IA-B mice received a secondary challenge with 50 ␮gof (Fig. 7A). The ratio of high-affinity anti-NP Ab (0.75 Ϯ 0.15, n ϭ soluble NP-CGG at 8 wk postimmunization, when the majority of 6) was similar to that of control mice (0.88 Ϯ 0.18, n ϭ 6) (data The Journal of Immunology 2129 Downloaded from http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 7. Robust Ag-specific memory response in IA-B mice that lack MHC-II from the majority of memory B cells. A, Wild-type mice and/or IA-B mice were given an i.v. secondary challenge with 50 ␮g of soluble NP-CGG or 20 ␮g of soluble NP-Ficoll at 30 wk postprimary immunization. Anti-NP IgG1 serum titer defined as the dilution of serum giving an OD at 405 nm of 0.05 by ELISA was measured immediately before (Ⅺ) and 7 days after (u) the secondary challenge. Representative data of four similar experiments with at least four mice per group are shown. B, Spleen cells were recovered from the same mice shown in A 7 days after the secondary challenge with NP-CGG or NP-Ficoll. CD138ϩ NP-PEϩ Ag-specific PCs (boxed) are shown after pregating on B220ϪCD38Ϫ cells. C, GC B cells (R2 gate) and memory B cells (R3 gate) were gated among lineage-negative (CD90.2, Gr-1, CD49b, CD11b, CD5, CD43, IgD, and IgM) B220ϩ spleen cells (R1 gate, FACS profiles not shown) from the same wild-type mice and IA-B mice shown in A. Histograms show MHC-II IA␤b expression on GC B cells (R2 gate) and memory B cells (R3 gate). Anti-CD90.1 isotype control staining is shown as the empty plots. Representative flow cytometry profiles of four similar experiments with at least three mice per group were shown. D, Ag capturing B220ϩ cells with memory phenotype (R4 gate) in the same wild-type mice shown in C were analyzed with or without pretreatment to strip cell surface Ig by incubation at 37°C for 30 min.

not shown). Thus, despite the loss of MHC-II expression from the should be noted that this population remarkably reduced without majority (ϳ96%) of the established memory B cells, IA-B mice significant change in the frequency of NP-binding GC B cells and mounted normal memory response. However, when mice received memory B cells when spleen cells were preincubated at 37°C be- ␮ a secondary challenge with 20 g of TI Ag NP-Ficoll, neither fore FACS staining to strip Ig off from the cell surface (35) (Fig. control mice (Fig. 7A) nor IA-B mice (data not shown) generated 7D). Therefore, these B220ϩ cells with memory phenotype seem detectable memory response. to have bound serum anti-NP Ig on their cell surface at some point When memory response was analyzed by FACS at day 7 after ϩ and then captured Ag (NP-PE) during FACS staining, like Ag- secondary immunization, NP-binding CD138 PCs in spleen (Fig. capturing cells that masquerade as memory B cells (36). The fre- 7B) and BM (data not shown) were dramatically increased in control mice and IA-B mice challenged with soluble NP-CGG but not quency of memory B cells in control mice and IA-B mice after ϳ with NP-Ficoll. Differentiation of NP-binding GC B cells in IA-B secondary challenge was about 0.01% of total spleen cells, mice was even more significant in IA-B mice compared with con- which was similar to that of mice before secondary challenge with trol mice (Fig. 7C). In addition, we found that a discrete population NP-CG and that of mice challenged with NP-Ficoll. of MHC-IIϪ NP-binding cells with memory B cell phenotype ap- Because NP-Ficoll challenge failed to generate memory re- peared particularly in IA-B mice (Fig. 7C). This NP-binding cell sponse in both control and IA-B mice, these results suggest that Ag population also appeared in some control mice in association with stimulation itself is not sufficient and cognate T cell help is re- high-titer of serum anti-NP Ab (data not shown). However, it quired for PC differentiation of memory B cells. Thus, the robust 2130 MHC-II-DEPENDENT PC DIFFERENTIATION OF MEMORY B CELLS memory response in IA-B mice was mounted by the ϳ4% of mem- cipient 4), and those that received only MHC-IIϩ memory B cells ory B cells that were MHC-IIϩ. without T cells and challenged with NP-CGG (recipient 5) (Fig.

Ϫ ϩ 8A). Thus, consistent with the results of Fig. 7, these results sug- MHC-II B220 memory B cells do not differentiate into PCs gest that Ag stimulation is not sufficient for PC differentiation of even in the presence of T cells memory B cells and cognate T cell help through MHC-II-restricted To further test whether cognate T cell help is required for PC Ag presentation by memory B cells is required. differentiation of memory B cells, MHC-IIϪ or MHC-IIϩ B cells To further examine the mechanisms of PC differentiation of Ϫ Ϫ from IA-B mice or control mice, respectively, at 30 wk postim- memory B cells, Ag-specific cellular response in RAG-1 / re- munization (B220ϩ cells, Ͼ86%) were transferred into RAG-1Ϫ/Ϫ cipient mice was analyzed by FACS at day 10 postchallenge. The ϩ ϩ mice with CD4ϩ T cells purified from CGG-primed B6 mice. ratios of B220 B cells and CD4 T cells in spleen were similar Ϫ Ϫ Ϫ ϩ These RAG-1Ϫ/Ϫ recipients were then challenged with soluble between RAG-1 / mice receiving MHC-II and MHC-II B NP-CGG. MHC-IIϩ cells (including MHC-IIϩ memory B cells) in cells after transfer and Ag restimulation (data not shown). As IA-B mice were either depleted by complement lysis or magnetic shown in Fig. 8B, B220 and MHC-II expression levels of cell sorting before transfer. The remaining MHC-IIϩB220ϩ B NP-binding cells (CD4Ϫ NP-binding cells (R1 gate) in the first cells estimated by FACS were Ͻ0.8%. The frequency of NP-bind- column) among total spleen cells of individual RAG-1Ϫ/Ϫ recip- ing cells among B220ϩ B cells was similar between control and ient mice were analyzed. It should be noted that the majority of IA-B mice before transfer (data not shown). NP-binding cells (R1) that appear in no. 1 and no. 2 recipient mice The RAG-1Ϫ/Ϫ mice that received MHC-IIϩ memory B cells are B220Ϫ cells, which are most likely Ab capturing non-B cells with T cells and challenged with NP-CGG (recipient 1) generated that bound serum anti-NP Ig and are detected as NP-binding cells Downloaded from robust anti-NP IgG1 response in association with high-affinity anti- (36). NP Ab production (Fig. 8A) at day 10 postchallenge, which was Consistent with anti-NP serum IgG1 response, a large number of also confirmed by ELISPOT (data not shown). In contrast, anti-NP both B220ϩ memory B cells (R2) and B220low PC precursors (R3) IgG1 response was greatly reduced (Ͼ1000-fold) in the RAG- were observed in no. 1 recipient mice that received MHC-IIϩ 1Ϫ/Ϫ mice that received MHC-IIϪ memory B cells with T cells memory B cells with T cells and challenged with NP-CGG. This ϩ/low and challenged with NP-CGG (recipient 2) (Fig. 8A). Anti-NP B220 population was not observed in no. 4 recipient mice that http://www.jimmunol.org/ IgG1 response was undetectable in recipient mice challenged with received MHC-IIϩ naive B cells. As shown in the MHC-II histo- PBS (recipient 3), those that received naive MHC-IIϩ B cells (re- grams of no. 1 recipient mice, B220ϩ memory B cells expressed by guest on September 27, 2021

FIGURE 8. MHC-II-deficient memory B cells transferred into RAG-1Ϫ/Ϫ mice fail to differentiate into PCs upon Ag restimulation even in the presence of Ag-specific CD4ϩ T cells. A, MHC-IIϩ B cells and MHC-IIϪ B cells were purified from wild-type mice and IA-B mice, respectively, at 30 wk postprimary immunization and transferred i.v. into RAG-1Ϫ/Ϫ mice with (no. 1–4) or without (no. 5) CD4ϩ T cells from CGG-primed mice. As a control, some recipients received MHC-IIϩ B cells from naive control mice (no. 4). Twenty-four hours after transfer, recipients were given an i.p. challenge of 50 ␮g of soluble NP-CGG (no. 1, 2, 4, and 5) or PBS (no. 3). Anti-NP IgG1 serum titers defined as the dilution of serum giving an OD at 405 nm of 0.05 Ⅺ u were measured by ELISA using NP15-BSA-coated plates (total Abs: ) and NP2-BSA-coated plates (high-affinity Abs: ) 10 days after the challenge. Data shown are representative of four similar experiments with at least four mice per group. B, Spleen cells were recovered from the same RAG-1Ϫ/Ϫ recipient mice shown in A 10 days after the challenge. Among the viable lymphocyte gate, CD4Ϫ NP-PEϩ Ag-specific B cells (R1) were identified and their MHC-II and B220 expression levels were analyzed in the second column. Among the NP-binding cells (R1), B220ϩ memory B cells (R2), and B220low PC precursors (R3) are identified according to their B220 expression levels. For recipient mice nos. 1–3, histograms of MHC-II IA␤b (bold empty histograms) expression on NP-binding B220ϩ memory B cells (R2) and NP-binding B220low PC precursors (R3) are also shown with anti-CD90.1 isotype control staining (gray histograms). Representative flow cytometry profiles and histograms of six similar experiments with at least three mice per group are shown. The Journal of Immunology 2131 higher levels of MHC-II, whereas B220low PC precursors down- MHC-IIϩ memory B cells in IA-B mice, which mounted even regulated MHC-II expression. Also, significant numbers of B220ϩ greater anti-NP IgG1 response compared with that in control mice memory B cells (R2) and B220low PC precursors (R3) were iden- within just 7 days after Ag restimulation (Fig. 7). It is not clear tified in no. 2 recipient mice that received MHC-IIϪ memory B why IA-B mice could generate such a robust memory response cells with T cells and challenged with NP-CGG. The numbers of even though the frequency of MHC-IIϩ memory B cells is small B cell populations in R2 and R3 gate of no. 2 recipient mice were compared with that in control mice. It has been reported that de- less than those in no. 1 recipient mice but much more than those liberate removal of T cell help promotes specific Ab production in no. 3 recipient mice that received MHC-IIϪ memory B cells with virus-neutralizing ability (42). In addition, NP-binding mem- with T cells and challenged with PBS. Thus, NP-CGG challenge ory B cells at 20ϳ30 wk postimmunization are more enriched in increased the number of MHC-IIϪ B220ϩ memory B cells. In the total memory B cells of IA-B mice compared with that of addition, as evident from the MHC-II histograms of no. 2 recipient control mice, as a consequence of the fact that IA-B mice have mice, NP-CGG challenge also induced expansion of MHC-IIϩ fewer naturally activated IgDϪ B cells (data not shown). There- cells among B220ϩ memory B cells (R2) and B220low PC precur- fore, it is possible that less competition between other MHC- sors (R3). In contrast, MHC-IIϩ memory cells hardly responded by IIϩB220ϩ memory B cells may help anti-NP MHC-IIϩB220ϩ NP-CGG challenge in the absence of memory T cells (no. 5). memory B cells of IA-B mice to gain efficient access to memory These results suggest that even MHC-IIϪ memory B cells can T cells upon secondary Ag stimulation. proliferate to a certain extent if Ag and bystander T cell help are Memory B cells recovered from mice 6 wk after CD4ϩ T cell provided (note that even though there is no cognate T-B interaction depletion could still generate memory response when transferred ϩ in no. 2 recipient mice due to lack of MHC-II on memory B cells, into mice with CD4 T cells and Ag (13), suggesting that memory Downloaded from memory T cells can provide bystander help such as cytokines). B cells can be maintained with little CD4ϩ T cell help. In this However, MHC-IIϪ memory B cells cannot further differentiate context, the current study demonstrates that B220ϩ memory B into PCs because they cannot receive optimum help from Ag-spe- cells do not absolutely require cognate interaction with Ag-specific cific T cells through MHC-II-restricted Ag presentation. CD4ϩ T cells through MHC-II restricted Ag presentation for the Collectively, these results led to the conclusion that memory maintenance of their number because neither a loss of overall ϩ response in IA-B mice at 30 wk postimmunization was by selec- B220 memory B cell frequency nor predominance of MHC- http://www.jimmunol.org/ tive reactivation of MHC-IIϩ memory B cells and that only 4% of IIϩB220ϩ B cells in the memory compartment was seen in IA-B the wild-type level of MHC-IIϩ memory B cells can restore nor- mice. However, it should be noted that a small fraction (ϳ4%) of mal memory response within 7 days after secondary challenge. B220ϩ memory B cells still expressed detectable levels of MHC-II Furthermore, the current results demonstrate that Ag stimulation is even at 30 wk postimmunization. Thus, it remains possible that not sufficient for the differentiation of memory B cells and cognate MHC-II expressing B220ϩ memory B cells in IA-B mice acquired T cell help through MHC-II-restricted Ag presentation is required. some advantageous survival signals provided by Ag-specific CD4ϩ T cells. Alternatively, it is conceivable that the remaining Discussion MHC-IIϩ memory B cells in IA-B mice represent a distinct subset MHC-II is a critical player for TD humoral immune response. of memory B cells, but we have no evidence at present to support by guest on September 27, 2021 However, the role of MHC-II-restricted B cell Ag presentation in such a possibility. Nonetheless, the survival of B220ϩ memory B humoral immune response and B cell differentiation has never cells seems to rely mainly on factors other than cognate T cell help. been directly studied in an intact whole animal model such as the Certainly, memory B cells regain expression of anti-apoptotic one we now present. In this context, the current study provides genes such as bcl-2 (33, 43) and TOSO (43) after GC selection. several findings. First, the current study demonstrated the great Cytokine receptors such as IL-2R␤ (43) and IL-4R as well as impact of clonal expansion of MHC-IIϩ B cells during the TD CD130, the signal transducer for IL-6 (44), are significantly up- primary response as well as memory response. In the primary re- regulated in human memory B cells compared with naive B cells. sponse, as little as 2% of the wild-type level of NP Ag-specific Because memory B cells seem to be preferentially localized in MHC-IIϩ B cells was enough to establish normal levels of GC B particular sites such as marginal zone areas of the spleen (1, 45) cells and memory B cells albeit only with a large Ag dose and with and mucosal epithelia (46), the particular localization in addition to delayed kinetics. In contrast, differentiation of extrafollicular PCs the intrinsic survival ability of memory B cells may ensure that in the primary response was greatly impaired. In fact, Ag-specific memory B cells have access to soluble survival factors and/or to ASCs in the spleen of IA-B mice were barely detectable during the cells providing survival signals. primary response until GC-derived PCs developed (data not It has been generally considered that differentiation of memory shown). Presumably, this is because naive MHC-IIϩ B cells acti- B cells into PCs upon secondary Ag stimulation requires T cell vated by Ag and T cells show only very limited proliferation be- help (6, 47). Memory B cells differentiate into short-lived PCs in fore they differentiate into either extrafollicular PCs or GC B cells. vivo upon Ag stimulation only in the presence of Ag-specific However, once they differentiated into GC B cells, they prolifer- CD4ϩ T cells (20). However, activation of virus-specific memory ated dramatically and clearly were able to reach wild-type maxi- B cells was intact even in the absence of CD4ϩ T cells (15). In mal levels and restore the humoral response. In this regard, it is addition, polyclonal stimulation such as CpG ODN or bystander T known at least that IgG-BCR-mediated signaling on IgGϩ B cells cells can induce proliferation and differentiation of human memory ensures efficient proliferation and IgG production upon Ag re- B cells in vitro (11, 12). Thus, the absolute requirement of Ag- stimulation compared with those expressing IgM and/or IgD specific CD4ϩ T cell help during memory B cell differentiation is signaling (40, 41). still questionable. In this context, the current study demonstrated Even more significant was the dramatic memory response by that MHC-IIϪ B cells fail to differentiate into PCs upon Ag re- MHC-IIϩ memory B cells in IA-B mice after Ag restimulation. stimulation even in the presence of Ag-specific CD4ϩ T cells. In Given the fact that MHC-IIϪ memory B cells failed to differentiate addition, RAG-1Ϫ/Ϫ mice receiving only either MHC-IIϩ or into PCs in vivo even in the presence of Ag-specific CD4ϩ T cells MHC-IIϪ B cells without T cells did not generate memory re- (Fig. 8), the secondary response in IA-B mice appears to be gen- sponse upon Ag restimulation. Therefore, our results support a erated by as little as ϳ4% of the wild-type level of NP Ag-specific notion that PC differentiation of memory B cells requires both 2132 MHC-II-DEPENDENT PC DIFFERENTIATION OF MEMORY B CELLS

MHC-II expression on memory B cells and Ag-specific CD4ϩ T this process is impaired in IA-B mice because of loss of cells. The discrepancies between the current study and a previous MHC-II from memory B cells. study (15) in regard to the requirement of CD4ϩ T cell help in Previous studies showed that polyclonal activation such as CpG memory B cell activation may depend on the Ag used for exper- DNA and bystander T cell help are able to drive PC differentiation iments. It is possible that virus Ag may be able to activate memory of memory B cells in vitro (11, 12). This type of stochastic mech- B cells through innate signaling pathways such as TLRs even in anism may play a role in maintaining a basal level of humoral the absence of cognate T-B interaction. In relation to this, early immunity when Ag is diminished. However, an “instructed” mech- CD4ϩ TD antiviral IgA response can be generated in the absence anism by Ag-specific T cells is anticipated to have a great advan- of MHC-II or CD40 on B cells (48). tage in controlling the “quality” of the long-lived PC population Certainly, the role of Ag-specific CD4ϩ T cells is to intimately when pathogen (Ag) is still present. First of all, high-affinity B provide memory B cells with cytokines such as IL-4 and IL-5 that cells preferentially capture Ag and present it to Th cells (52). promote PC differentiation through MHC-II-restricted interaction Through MHC-II-restricted Ag presentation to T cells, high-affin- (5). In addition, several studies have reported that MHC-II trans- ity memory B cells, and/or PC precursors differentiate into PCs duces signals to B cells upon Ag presentation (49, 50). Therefore, and provide higher affinity Abs that can clear residual pathogen it is possible that MHC-II on memory B cells may be required for more efficiently. Second, cognate T cell-B cell interaction would delivering specific signals for their efficient proliferation and dif- allow differentiation of useful Ag-specific PCs but presumably not ferentiation into PCs. In this context, MHC-IIϪ memory B cells help the differentiation of self-reactive PCs. Third, it remains pos- proliferated to a certain extent in vivo when Ag and bystander T sible that cognate T cell-B cell interaction provides Ag-specific cell help were provided (Fig. 8B). In addition, in vitro culture of memory B cells with signals that maintain their capability to dif- Downloaded from purified MHC-IIϪ memory B cells of IA-B mice at 30 wk postim- ferentiate into PCs. Finally, in fact, cognate T cell-B cell interac- munization in the presence of LPS and cytokines induced a similar tion may be important for the maintenance of memory T cell func- level of anti-NP IgG1 production compared with MHC-IIϩ mem- tion. In support of this, memory T cells that had been maintained ory B cells of control mice (data not shown). However, it remains in the absence of MHC-II showed reduced functional activity upon possible that MHC-II expression on memory B cells (i.e., cognate Ag re-encounter (53). CD4ϩ T cell help) may be required to maintain their full compe- Our study showed that cognate interaction between memory T http://www.jimmunol.org/ cells is important for PC differentiation by memory B cells. Sev- tence for subsequent PC differentiation and function upon Ag stim- eral studies have shown that both memory B cells and T cells home ulation. Furthermore, we have to concede the possibility that on- to BM (20, 54, 55). In this context, we are interested in the pos- going deletion of MHC-II in IA-B mice might select a certain sibility that memory B cells make cognate interaction with mem- subset of memory B cells that can survive in the absence of ory T cells and differentiate into long-lived PCs in BM during MHC-II but has lost PC differentiation capability. These possibil- immunological surveillance. In this regard, IA-B mice should pro- ities are under investigation with V Tg memory B cells in an H vide an excellent tool to further investigate the mechanisms of TD IA-B mouse background. and TI memory B cell maintenance and differentiation.

Even though the selection of GC B cells and the generation and by guest on September 27, 2021 maintenance of B220ϩ memory B cells seemed to be fairly normal, the generation of high-affinity long-lived PCs in IA-B mice Acknowledgments was reduced in association with impaired post-GC serum affinity We thank T. Takemori, T. Tsubata, Y. Takahashi, M. Iwashima, and A. Mellor for critical discussions; Y. Takahashi for technical suggestions maturation (Fig. 6). That is, GC B cell frequency in IA-B mice on the memory B cell sorting and culture; L. Ignatowicz for Y3P Ab; recovered to a similar or even higher level than that of control mice G. Trinchieri for 120G8 Ab; and K. Miyake for confocal image analysis. by 2 wk postimmunization (Fig. 5B). The selection of high-affinity variants in GC of IA-B mice was also normal despite the delayed Disclosures kinetics (Table I). In addition, the frequency of memory B cells in The authors have no financial conflict of interest. IA-B mice was normal throughout the analysis (Fig. 5D). In contrast, numbers of high-affinity as well as total numbers of ASCs in References BM were significantly reduced in IA-B mice during the post-GC 1. Liu, Y. J., S. Oldfield, and I. C. MacLennan. 1988. Memory B cells in T cell- phase (Fig. 6, E and F) in association with impaired post-GC se- dependent antibody responses colonize the splenic marginal zones. Eur. J. Im- rum affinity maturation in IA-B mice (Fig. 6D). Long-lived PCs munol. 18: 355–362. 2. Jacob, J., R. Kassir, and G. Kelsoe. 1991. In situ studies of the primary immune are generated by preferential differentiation of high-affinity vari- response to (4-hydroxy-3-nitrophenyl)acetyl. I. The architecture and dynamics of ants from GC (51). It is possible that generation of high-affinity responding cell populations. J. Exp. Med. 173: 1165–1175. PCs from GC is impaired in IA-B mice because of delayed GC 3. Kelsoe, G. 1996. Life and death in germinal centers (redux). Immunity 4: ϩ 107–111. development and/or a limited MHC-II Ag-specific repertoire in 4. Blink, E. J., A. Light, A. Kallies, S. L. Nutt, P. D. Hodgkin, and D. M. Tarlinton. IA-B mice. However, this alone cannot explain the reduced anti- 2005. Early appearance of germinal center-derived memory B cells and plasma cells in blood after primary immunization. J. Exp. Med. 201: 545–554. NP IgG1 long-lived PC numbers and serum titers particularly in 5. Calame, K. L., K. I. Lin, and C. Tunyaplin. 2003. Regulatory mechanisms that the late post-GC phase (Fig. 6, C and D). Alternatively, these re- determine the development and function of plasma cells. Annu. Rev. Immunol. 21: 205–230. sults may suggest that accumulation of long-lived PCs during 6. McHeyzer-Williams, L. J., and M. G. McHeyzer-Williams. 2005. Antigen-spe- the post-GC phase rather than their generation from GC itself is cific memory B cell development. Annu. Rev. Immunol. 23: 487–513. impaired in IA-B mice. Whether the long-lived PC compart- 7. Schittek, B., and K. Rajewsky. 1990. Maintenance of B-cell memory by long- lived cells generated from proliferating precursors. Nature 346: 749–751. ment is established only from the early GC emigrants (4) or by 8. Slifka, M. K., R. Antia, J. K. Whitmire, and R. Ahmed. 1998. Humoral immunity continuous differentiation of precursor cells during the post-GC due to long-lived plasma cells. Immunity 8: 363–372. phase (18) is not yet clear. Considering the fact that MHC-IIϪ 9. Manz, R. A., A. Thiel, and A. Radbruch. 1997. Lifetime of plasma cells in the bone marrow. Nature 388: 133–134. memory B cells were not able to differentiate into PCs in vivo 10. Maruyama, M., K. P. Lam, and K. Rajewsky. 2000. Memory B-cell persistence even in the presence of CD4ϩ T cells, it is interesting to spec- is independent of persisting immunizing antigen. Nature 407: 636–642. 11. Bernasconi, N. L., E. Traggiai, and A. Lanzavecchia. 2002. Maintenance of se- ulate that the number and affinity of long-lived PCs are main- rological memory by polyclonal activation of human memory B cells. Science tained by continuous differentiation of memory B cells and that 298: 2199–2202. The Journal of Immunology 2133

12. Poeck, H., M. Wagner, J. Battiany, S. Rothenfusser, D. Wellisch, V. Hornung, selection of high affinity memory B cells in nasal-associated lymphoid tissue. B. Jahrsdorfer, T. Giese, S. Endres, and G. Hartmann. 2004. Plasmacytoid den- J. Exp. Med. 194: 1597–1607. dritic cells, antigen, and CpG-C license human B cells for plasma cell differen- 35. Kumagai, K., T. Abo, T. Sekizawa, and M. Sasaki. 1975. Studies of surface tiation and immunoglobulin production in the absence of T-cell help. Blood 103: immunoglobulins on human B lymphocytes. I. Dissociation of cell-bound im- 3058–3064. munoglobulins with acid pH or at 37 degrees C. J. Immunol. 115: 982–987. 13. Vieira, P., and K. Rajewsky. 1990. Persistence of memory B cells in mice de- 36. Bell, J., and D. Gray. 2003. Antigen-capturing cells can masquerade as memory prived of T cell help. Int. Immunol. 2: 487–494. B cells. J. Exp. Med. 197: 1233–1244. 14. Gatto, D., T. Pfister, A. Jegerlehner, S. W. Martin, M. Kopf, and 37. Cumano, A., and K. Rajewsky. 1985. Structure of primary anti-(4-hydroxy-3- M. F. Bachmann. 2005. Complement receptors regulate differentiation of bone nitrophenyl)acetyl (NP) antibodies in normal and idiotypically suppressed marrow plasma cell precursors expressing transcription factors Blimp-1 and C57BL/6 mice. Eur. J. Immunol. 15: 512–520. XBP-1. J. Exp. Med. 201: 993–1005. 38. Allen, D., T. Simon, F. Sablitzky, K. Rajewsky, and A. Cumano. 1988. Antibody 15. Hebeis, B. J., K. Klenovsek, P. Rohwer, U. Ritter, A. Schneider, M. Mach, and engineering for the analysis of affinity maturation of an anti-hapten response. T. H. Winkler. 2004. Activation of virus-specific memory B cells in the absence EMBO J. 7: 1995–2001. of T cell help. J. Exp. Med. 199: 593–602. 39. Furukawa, K., A. Akasako-Furukawa, H. Shirai, H. Nakamura, and T. Azuma. 16. Takahashi, Y., P. R. Dutta, D. M. Cerasoli, and G. Kelsoe. 1998. In situ studies 1999. Junctional amino acids determine the maturation pathway of an antibody. of the primary immune response to (4-hydroxy-3-nitrophenyl)acetyl. V. Affinity Immunity 11: 329–338. maturation develops in two stages of clonal selection. J. Exp. Med. 187: 885–895. 40. Pogue, S. L., and C. C. Goodnow. 2000. Gene dose-dependent maturation and 17. Wang, Y., G. Huang, J. Wang, H. Molina, D. D. Chaplin, and Y. X. Fu. 2000. receptor editing of B cells expressing immunoglobulin (Ig)G1 or IgM/IgG1 tail Antigen persistence is required for somatic mutation and affinity maturation of antigen receptors. J. Exp. Med. 191: 1031–1044. immunoglobulin. Eur. J. Immunol. 30: 2226–2234. 41. Wakabayashi, C., T. Adachi, J. Wienands, and T. Tsubata. 2002. A distinct sig- 18. O’Connor, B. P., M. Cascalho, and R. J. Noelle. 2002. Short-lived and long-lived naling pathway used by the IgG-containing B cell antigen receptor. Science 298: bone marrow plasma cells are derived from a novel precursor population. J. Exp. 2392–2395. Med. 195: 737–745. 42. Recher, M., K. S. Lang, L. Hunziker, S. Freigang, B. Eschli, N. L. Harris, 19. Barrington, R. A., O. Pozdnyakova, M. R. Zafari, C. D. Benjamin, and A. Navarini, B. M. Senn, K. Fink, M. Lotscher, et al. 2004. Deliberate removal M. C. Carroll. 2002. B lymphocyte memory: role of stromal cell complement and of T cell help improves virus-neutralizing antibody production. Nat. Immunol. 5: Fc␥RIIB receptors. J. Exp. Med. 196: 1189–1199. 934–942. Downloaded from 20. Ochsenbein, A. F., D. D. Pinschewer, S. Sierro, E. Horvath, H. Hengartner, and 43. Klein, U., Y. Tu, G. A. Stolovitzky, J. L. Keller, J. Haddad, Jr., V. Miljkovic, R. M. Zinkernagel. 2000. Protective long-term antibody memory by antigen- G. Cattoretti, A. Califano, and R. Dalla-Favera. 2003. Transcriptional analysis of driven and T help-dependent differentiation of long-lived memory B cells to the B cell germinal center reaction. Proc. Natl. Acad. Sci. USA 100: 2639–2644. short-lived plasma cells independent of secondary lymphoid organs. Proc. Natl. 44. Feldhahn, N., I. Schwering, S. Lee, M. Wartenberg, F. Klein, H. Wang, G. Zhou, Acad. Sci. USA 97: 13263–13268. S. M. Wang, J. D. Rowley, J. Hescheler, et al. 2002. Silencing of B cell receptor 21. Hayakawa, K., R. Ishii, K. Yamasaki, T. Kishimoto, and R. R. Hardy. 1987. signals in human naive B cells. J. Exp. Med. 196: 1291–1305. Isolation of high-affinity memory B cells: phycoerythrin as a probe for antigen- 45. Dunn-Walters, D. K., P. G. Isaacson, and J. Spencer. 1995. Analysis of mutations

binding cells. Proc. Natl. Acad. Sci. USA 84: 1379–1383. in immunoglobulin heavy chain variable region genes of microdissected marginal http://www.jimmunol.org/ 22. Gray, D., and H. Skarvall. 1988. B-cell memory is short-lived in the absence of zone (MGZ) B cells suggests that the MGZ of human spleen is a reservoir of antigen. Nature 336: 70–73. memory B cells. J. Exp. Med. 182: 559–566. 23. Oliver, A. M., F. Martin, and J. F. Kearney. 1997. Mouse CD38 is down-regu- 46. Liu, Y. J., C. Barthelemy, O. de Bouteiller, C. Arpin, I. Durand, and lated on germinal center B cells and mature plasma cells. J. Immunol. 158: J. Banchereau. 1995. Memory B cells from human tonsils colonize mucosal ep- 1108–1115. ithelium and directly present antigen to T cells by rapid up-regulation of B7-1 and 24. Ridderstad, A., and D. M. Tarlinton. 1998. Kinetics of establishing the memory B7-2. Immunity 2: 239–248. B cell population as revealed by CD38 expression. J. Immunol. 160: 4688–4695. 47. McHeyzer-Williams, M. G. 2003. B cells as effectors. Curr. Opin. Immunol. 15: 25. McHeyzer-Williams, L. J., M. Cool, and M. G. McHeyzer-Williams. 2000. An- 354–361. tigen-specific B cell memory: expression and replenishment of a novel B220Ϫ 48. Sangster, M. Y., J. M. Riberdy, M. Gonzalez, D. J. Topham, N. Baumgarth, and memory B cell compartment. J. Exp. Med. 191: 1149–1166. P. C. Doherty. 2003. An early CD4ϩ T cell-dependent immunoglobulin A re- 26. Driver, D. J., L. J. McHeyzer-Williams, M. Cool, D. B. Stetson, and sponse to influenza infection in the absence of key cognate T-B interactions. Ϫ M. G. McHeyzer-Williams. 2001. Development and maintenance of a B220 J. Exp. Med. 198: 1011–1021. by guest on September 27, 2021 memory B cell compartment. J. Immunol. 167: 1393–1405. 49. St. Pierre, Y., and T. H. Watts. 1991. Characterization of the signaling function 27. Rickert, R. C., K. Rajewsky, and J. Roes. 1995. Impairment of T-cell-dependent of MHC class II molecules during antigen presentation by B cells. J. Immunol. B-cell responses and B-1 cell development in CD19-deficient mice. Nature 376: 147: 2875–2882. 352–355. 50. Lang, P., J. C. Stolpa, B. A. Freiberg, F. Crawford, J. Kappler, A. Kupfer, and 28. Hashimoto, K., S. K. Joshi, and P. A. Koni. 2002. A conditional null allele of the J. C. Cambier. 2001. TCR-induced transmembrane signaling by peptide/MHC major histocompatibility IA-␤ chain gene. Genesis 32: 152–153. class II via associated Ig-␣/␤ dimers. Science 291: 1537–1540. 29. Cosgrove, D., D. Gray, A. Dierich, J. Kaufman, M. Lemeur, C. Benoist, and 51. Smith, K. G., A. Light, G. J. Nossal, and D. M. Tarlinton. 1997. The extent of D. Mathis. 1991. Mice lacking MHC class II molecules. Cell 66: 1051–1066. affinity maturation differs between the memory and antibody-forming cell com- 30. Koni, P. A., S. K. Joshi, U. A. Temann, D. Olson, L. Burkly, and R. A. Flavell. partments in the primary immune response. EMBO J. 16: 2996–3006. 2001. Conditional vascular cell adhesion molecule 1 deletion in mice: impaired 52. Batista, F. D., and M. S. Neuberger. 2000. B cells extract and present immobi- lymphocyte migration to bone marrow. J. Exp. Med. 193: 741–754. lized antigen: implications for affinity discrimination. EMBO J. 19: 513–520. 31. Koni, P. A., and R. A. Flavell. 1999. Lymph node germinal centers form in the 53. Kassiotis, G., S. Garcia, E. Simpson, and B. Stockinger. 2002. Impairment of absence of follicular dendritic cell networks. J. Exp. Med. 189: 855–864. immunological memory in the absence of MHC despite survival of memory T 32. Asselin-Paturel, C., G. Brizard, J. J. Pin, F. Briere, and G. Trinchieri. 2003. cells. Nat. Immunol. 3: 244–250. Mouse strain differences in plasmacytoid dendritic cell frequency and function 54. Paramithiotis, E., and M. D. Cooper. 1997. Memory B lymphocytes migrate to revealed by a novel monoclonal antibody. J. Immunol. 171: 6466–6477. bone marrow in humans. Proc. Natl. Acad. Sci. USA 94: 208–212. 33. Takahashi, Y., H. Ohta, and T. Takemori. 2001. Fas is required for clonal selec- 55. Di Rosa, F., and R. Pabst. 2005. The bone marrow: a nest for migratory memory tion in germinal centers and the subsequent establishment of the memory B cell T cells. Trends Immunol. 26: 360–366. repertoire. Immunity 14: 181–192. 56. Haas, K. M., J. C. Poe, D. A. Steeber, and T. F. Tedder. 2005. B-1a and B-1b cells 34. Shimoda, M., T. Nakamura, Y. Takahashi, H. Asanuma, S. Tamura, T. Kurata, exhibit distinct developmental requirements and have unique functional roles in T. Mizuochi, N. Azuma, C. Kanno, and T. Takemori. 2001. Isotype-specific innate and adaptive immunity to S. pneumoniae. Immunity 23: 7–18.