Memory B Cells a Unique Population of Human Splenic Stimulating

IL-21 and BAFF/BLyS Synergize in Stimulating Plasma Cell Differentiation from a Unique Population of Human Splenic Memory B Cells This information is current as of September 25, 2021. Rachel Ettinger, Gary P. Sims, Rachel Robbins, David Withers, Randy T. Fischer, Amrie C. Grammer, Stefan Kuchen and Peter E. Lipsky J Immunol 2007; 178:2872-2882; ; doi: 10.4049/jimmunol.178.5.2872 Downloaded from http://www.jimmunol.org/content/178/5/2872

References This article cites 50 articles, 18 of which you can access for free at: http://www.jimmunol.org/ http://www.jimmunol.org/content/178/5/2872.full#ref-list-1

Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists by guest on September 25, 2021 • Fast Publication! 4 weeks from acceptance to publication

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

IL-21 and BAFF/BLyS Synergize in Stimulating Plasma Cell Differentiation from a Unique Population of Human Splenic Memory B Cells1

Rachel Ettinger,2 Gary P. Sims, Rachel Robbins, David Withers, Randy T. Fischer, Amrie C. Grammer, Stefan Kuchen, and Peter E. Lipsky

Both constitutive Ig secretion by long-lived plasma cells (PC) and the recurrent differentiation of memory (mem) B cells into PC contribute to the maintenance of serologic mem. However, the relative contribution of each is unknown. In this study, we describe a novel population of human postswitched mem B cells that rapidly differentiate into PC and thus contribute to serologic mem. These IgG؉ B cells reside in the region of human spleen analogous to the murine marginal zone and have not previously been ؉ examined. These cells are highly responsive to IL-21 in the context of CD40 stimulation. Uniquely, IgG marginal zone analog B Downloaded from cells are exquisitely sensitive to the combination of IL-21 and B cell-activating factor belonging to the TNF family (BAFF/BLyS) that synergize in the absence of further costimulation to induce up-regulation of B lymphocyte-induced maturation protein-1 and drive PC differentiation. Other cytokine combinations are not active in this regard. This is the first demonstration that this unique population of mem B cells can respond specifically and exclusively to IL-21 and BAFF/BLyS by differentiating into IgG-secreting PC, and thus contributing to serologic mem in an Ag-independent manner. The Journal of Immunology, 2007, 178: 2872–2882. http://www.jimmunol.org/ ollowing vaccination, protective Abs (serologic memory ever, the phenotypic and functional characteristics of IgDϪ MZA (mem)3) persist for decades, and are maintained by the B cells have not been analyzed, and their potential role in main- F combination of Ab secretion from long-lived PC as well as taining serologic mem has not been assessed. recurrent differentiation of mem B cells into plasma cells (PC) (1). IL-21 belongs to a family of cytokines that bind receptors that Although human blood mem B cells can differentiate into PC in signal through the common ␥ receptor. IL-21 is produced by ac- response to various nonspecific stimuli and T cell-derived factors tivated T cells or spontaneously by CXCR5ϩ follicular helper T (1, 2), it is unclear whether a unique population of mem B cells cells, and contributes to responses of T cells, B cells and NK cells, contributes to serologic mem in humans, such as preplasma mem each of which express the IL-21R (10–12). Previously, we and cells in mice (3). In humans, unlike mice, CD27ϩ mem B cells others have demonstrated that IL-21 has the ability to induce class by guest on September 25, 2021 with somatically mutated Ig genes are found in the region of the switch recombination (CSR) and PC differentiation from both spleen analogous to the murine marginal zone (MZ) (4–6). This mouse splenic B cells and human blood B cells by costimulating region contains both IgDϩCD27ϩ and IgDϪCD27ϩ mem B cells. up-regulation of activation-induced cytidine deaminase (AID) and In contrast, in mice, MZ cells are largely naive, long-lived, non- B lymphocyte-induced maturation protein-1 (BLIMP-1) (13–16). recirculating cells that respond rapidly to T-independent type 2 Although the combination of IL-2, IL-10, and anti-CD40 has the Ags (TI-2) (7). Little is known regarding the role of human MZ ability to stimulate PC differentiation from human mem B cells analog (MZA) B cells in protective immunity, but studies have (17, 18), IL-21 coactivation is unique in its ability to induce CSR suggested that IgDϩ MZA B cells may be involved in responses to and PC differentiation from both naive and mem peripheral blood T-independent Ags (8), and these cells have been suggested to be B cells (15). The observations that helper T cells within secondary similar to circulating blood IgDϩCD27ϩ mem B cells (9). How- lymphoid organs constitutively produce IL-21 (11) and overpro- duction of IL-21 results in hypergammaglobulinemia and autoimmunity (14, 19), make IL-21 a prime candidate cytokine to be Autoimmunity Branch, National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, MD 20892 involved in the regulation of serologic mem. Received for publication October 3, 2006. Accepted for publication December 8, 2006. Whereas IL-21 is clearly one of the major T cell-derived factors The costs of publication of this article were defrayed in part by the payment of page involved in humoral immune responses, dendritic cell (DC)-de- charges. This article must therefore be hereby marked advertisement in accordance rived factors are also known to play a role. B cell-activating factor with 18 U.S.C. Section 1734 solely to indicate this fact. belonging to the TNF family (BAFF/BLyS) belongs to the TNF 1 This work was supported by the intramural program of the National Institute of family and binds three receptors, BAFF/BLyS receptor (BAFFR), Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health. transmembrane activator and calcium modulator cyclophilin li- 2 Address correspondence and reprint requests to Dr. Rachel Ettinger, National Institute of gand interactor (TACI), and B cell maturation Ag (BCMA) (20). Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Building 10, Room 6D-47B, Bethesda, MD 20892. E-mail address: ettingerr@mail. BAFF/BLyS is known to be a B cell survival factor, and overex- nih.gov pression of BAFF/BLyS results in hypergammaglobulinemia as 3 Abbreviations used in this paper: mem, memory; PC, plasma cell; MZ, marginal zone; well as autoimmunity (20). It therefore is another prime candidate TI-2, T-independent type 2 Ag; MZA, MZ analog; CSR, class switch recombination; to contribute to serologic mem. In humans, DC can induce PC AID, activation induced cytidine deaminase; BLIMP-1, B lymphocyte-induced maturation protein-1; DC, dendritic cell; BAFFR, BAFF/BLyS receptor; TACI, transmembrane differentiation (21–23), and BAFF/BLyS and a proliferation-in- activator and calcium modulator cyclophilin ligand interactor; BCMA, B cell maturation ducing ligand (APRIL) expressed by activated DC may induce Ag; APRIL, a proliferation-inducing ligand; FO, follicular; ␤-2M, ␤-2 microglobulin; AP, alkaline phosphatase; IHC, immunohistochemistry; AF, Alexa Fluor; LYVE, lymphatic CSR (24). Importantly, the relative contribution of T cell- and endothelial hyaluronan receptor; FDC, follicular DC. DC-derived factors in protective immunity is not well understood. www.jimmunol.org The Journal of Immunology 2873

In this study, we report a novel population of human splenic mem B cells poised to differentiate into PC following encounter with T cell-derived and DC-derived soluble factors. These IgGϩ B cells reside in the region of human spleen analogous to the murine MZ and are exquisitely sensitive to IL-21. Notably, the combination of IL-21 and BAFF/BLyS synergizes in the absence of further costimulation to induce BLIMP-1 expression and PC differentiation from IgGϩ human MZA B cells. The ability of IgGϩ MZA B cells to respond to the combination of IL-21 and BAFF/BLyS demonstrates that these cells are uniquely receptive to Ag-independent signals from activated T cells and DC and are capable of rapid differentiation into PC, thereby replenishing serologic mem in a bystander fashion.

Materials and Methods Isolation of human B cells All human studies have been approved by the Warren G. Magnuson Clin- ical Center Institutional Review Board, and informed consent was obtained according to the Declaration of Helsinki. Spleens were obtained from organ donors. For total CD19ϩ splenic cultures, B cells were isolated by posi- Downloaded from tively selecting B cells with anti-CD19 magnetic beads (Miltenyi Biotec). These were typically Ͼ95% pure. Splenic B cell subpopulations were iden- tiﬁed by expression of CD19, CD21, CD23, and IgG or CD19, CD23, and CD27, and isolated with a MoFlo cell sorter (DakoCytomation). Brieﬂy, splenic cells were incubated at 107 cells/ml in staining buffer with mAbs for 30 min at 4°C, washed, and sorted into follicular (FO): CD19ϩ ϩ ϩ Ϫ ϩ ϩ Ϫ ϩ

CD21 CD23 IgG or CD19 CD23 CD27 and total MZA; CD19 http://www.jimmunol.org/ CD27ϩCD23Ϫ or CD19ϩCD21highCD23Ϫ B cells. To distinguish IgGϪ from the IgGϩ MZA subpopulation (and exclude CD27ϩCD21Ϫ/low CD23Ϫ conventional mem B cells and PC), IgGϩ or IgGϪ MZA were sorted as CD19ϩCD23ϪCD21highIgGϪ or CD19ϩCD23ϪCD21highIgGϩ subsets. Cells isolated using CD21 and CD23 or CD27 and CD23 responded similarly. Preparations were typically Ͼ98% pure. Cells sorted with CD21-FITC, CD23-PE, CD19-PerCP-Cy5.5, and IgG-allophycocyanin were analyzed after 5–12 days of culture by staining with IgD-FITC, ϩorϪ ϩ ϩ CD19-PerCP-Cy5.5, and CD38-allophycocyanin to distinguish IgDϪCD38high FIGURE 1. Human spleen MZA B cells are IgM CD19 CD27 ϩ ϩ Ϫ Ϫ PC. Carryover of small amounts of CD21-FITC were observed in the IgD- CD21 CD1c CD23 Ki-67 . Serial cryosections of human spleen were FITC channel of CD21highCD19ϩCD23Ϫ MZA B cells following culture. No doubled stained with anti-IgD-HRP (red) and either anti-CD19, CD27, by guest on September 25, 2021 carryover of mAb used for isolation was observed in the other channels. CD21, IgM, CD1c, CD3, CD23, and Ki-67 mAbs (blue). Magnification shown is ϫ200 and bar shows 60 ␮m. Staining is representative of 15 Cell culture spleens. Arrows show central arterioles. Purified B cells were cultured at 1 ϫ 106 cells/ml in either 1 ml in 24-well culture plates or 100 ␮l in 96-well round-bottom culture plates. The cells were incubated with a combination of human IL-21 (100 ng/ml; BioSource Determination of Ig levels International) with either 1 ␮g/ml anti-human CD40 (R&D Systems), 5 ␮g/ml anti-IgM (Jackson ImmunoResearch Laboratories), or both. In some Secreted Ig in the culture supernatant was quantitated by ELISA. Briefly, experiments, the cells were incubated with a combination of human IL-2 96-well flat-bottom Nunc-Brand Immuno Plates (Nalge Nunc Interna- (100 U/ml; Roche), human IL-4 (100 ng/ml; R&D Systems), human IL-10 tional) were coated with 5 ␮g/ml either affinity-purified goat anti-human (25 ng/ml; R&D Systems), and human IL-6 (400 ng/ml; R&D Systems). IgM or goat anti-human IgG-Fc (Bethyl Laboratories) overnight at 4°C. All cytokines and stimuli were added at the initiation of culture. Human Wells were then washed and blocked with a 0.2% BSA/PBS, and then BAFF/BLyS was added at 200 ng/ml, and APRIL was used at 500 ng/ml titered culture supernatant was added and incubated overnight at 4°C. (both obtained from R&D Systems). Bound Ig was detected with 0.5 ␮g/ml alkaline phosphatase (AP)-conjugated goat anti-human IgM or IgG (Bethyl Laboratories) and developed Flow cytometry with p-nitrophenyl phosphate tablets (Sigma-Aldrich). Specific absorbance was measured and OD quantified at 410 nm by a Powerwave X 96-well Four-color flow cytometry was performed using a FACSCalibur (BD Bio- plate reader (Bio-Tek Instruments). sciences). Briefly, all cells were harvested from 96-well cultures and stained for 30 min on ice with a combination of mAbs. The combination of B cell proliferation anti-IgD-FITC, anti-CD27-PE, anti-CD19-PerCP-Cy5.5, and anti-CD38- allophycocyanin (clone HB7) was routinely used (BD Biosciences). In ad- To assess proliferative responses of cultured cells, 105 purified B cells were dition anti-CD21-FITC (Immunotech) with anti-CD23-allophycocyanin or cultured as described above in 96-well round-bottom plates. After 3 days anti-CD23-PE with anti-CD2-allophycocyanin with anti-CD19-PerCP- of culture, [3H]thymidine (37 Kbq/well) was added to the cultures for an Cy5.5 were used with PE-conjugated Abs to IgD, IgG, CD27, CD95, B220, additional 16 h. Thymidine uptake was measured using a liquid scintilla- CXCR4 and allophycocyanin conjugated to CD38 and CD44 or FITC- tion counter. conjugated to CD45RA (all obtained from BD Biosciences), CXCR5-PE (R&D Systems), IgM-PE (Serotec), CD148-AF647 (BioSource Interna- Real-time quantitative PCR tional) and anti-BCMA-biotin, anti-TACI -biotin or goat anti-BAFFR and goat IgG-biotin control was used (R&D Systems) followed by streptavi- Purified B cell populations were stimulated as described above. After 3 din-PE or anti-goat-FITC. Viable cells were identified by gating on lym- days in culture, RNA was isolated using the RNeasy mini kit (Qiagen). phocytes and cells were analyzed immediately. All samples were collected Reverse transcription reactions were prepared using the SuperScript One- for 30 s, and as a result the density of the dot plots reveals relative cell Step PCR System with platinum Taq polymerase and ROX reference dye numbers. In some experiments, to obtain total B cell numbers more accu- (Invitrogen Life Technologies). Fifty nanograms of isolated RNA was added rately, AccuCount Particles (Spherotech) were added and total cell num- per reaction with 1.2 mM MgSO4. TaqMan Assays on Demand Gene expres- bers were determined per the manufacturer’s instructions. sion primer/probe sets (Applied Biosystems) were used for BLIMP-1 2874 IL-21 AND BAFF STIMULATE HUMAN MZA B CELLS

FIGURE 2. MZA B cells are composed of both pre- and postswitched mem B cells. A, Serial frozen tissue sections of human spleen imaged by confocal micros- copy are shown at ϫ200. Single staining with triple overlays of CD27 (green), IgM (red), and IgD (blue) or CD3 (green; n ϭ 12), IgM (red), and CD11c (blue; n ϭ 6), or CD27 (green), IgM (red), and LYVE-1 (blue; n ϭ 6) are shown. Arrows indicate the central arterioles. B, High-power (ϫ400) triple staining shows CD11cϩ (red) cells located within the MZA area, as well as IgMϩ (blue) cells. Note the many IgDϪIgMϪCD27ϩB cells in this region. Bar shows 50 ␮m. Downloaded from http://www.jimmunol.org/ (Hs00153357_m1), AID (Hs00221068_m1), Bim (Hs00197982_m1), and ␤-2 quantiﬁcation were completed as described by the manufacturer’s instructions microglobulin (␤-2M) (Hs99999907_m1). Final concentrations were 1.8 ␮M (Applied Biosystems). mRNA expression for each gene was calculated using for primers and 0.5 ␮M for probes. RT-PCR was performed using the ABI the comparative cycle threshold method with efﬁciency calculations and with Prism 7700 Sequence Detection System, and cycle conditions and relative all mRNA levels normalized to ␤-2M. by guest on September 25, 2021

FIGURE 3. MZA B cells express a unique set of markers. Freshly obtained splenic cells were stained for expression of CD19, CD21, and CD23. Expression of various Ags in FO (CD21ϩCD23ϩ), MZA cells (CD21highCD23Ϫ), and CD21Ϫ/lowCD23Ϫ cells was determined as indicated. Numbers represent mean fluorescence intensity of the total population, and numbers in parentheses represent mean fluorescence intensity of the positive population. Data are representative results for 1 of 6 similar experiments for analysis of the total Ab profile and 11 experiments for analysis of CD27, IgD, IgM, IgG, and CD38 by populations defined by CD21/CD23expression.ThemeanfrequencyofCD21ϩ CD23ϩ FO cells was 22%, range 1–38%, and the mean frequency of CD21highCD23ϪMZA cells was 34%, range 11–74% (n ϭ 11). The Journal of Immunology 2875

FIGURE 4. Pre- and postsort analysis of splenic B cell subpopulations. A, Splenic B cells were analyzed based on expression of CD19, CD38, IgD, CD21, CD23, and CD27. Cells were sorted based on expression of CD19, CD27, and CD23. FO B cells were iden- tiﬁed as CD19ϩCD23ϩCD27Ϫ and MZA B cells, and a small percentage of mem B cells were identiﬁed as CD19ϩCD23ϪCD27ϩ. Pre- and postsort analysis of expression of CD23 and CD27 is shown as well as cell surface phenotype of presorted CD19ϩ B cells: CD19 and CD38; IgD and CD38; and CD21 and CD23. Pre- sorted spleens contained 1% PC on day 0 (CD19ϩ CD38high and IgDϪCD38high “boxed” cells), and no PC

were observed in the postsort gate as demonstrated by Downloaded from the absence of CD23ϪCD27high B cells. B, Splenic B cells were analyzed for expression of IgD, CD38, CD21, CD23, and IgG. Cells were sorted based on expression of CD19, CD21, CD23, and IgG. FO B cells were identiﬁed as CD19ϩCD23ϩCD21ϩIgGϪ, and total MZA B cells were identiﬁed as CD19ϩCD23Ϫ high ϩ high Ϫ Ϫ CD21 . IgG CD21 CD23 MZA or IgG http://www.jimmunol.org/ CD21highCD23ϪMZA B cells are also shown. Pre- and postsort analysis is shown as well as IgD vs CD38 cell surface phenotype of presorted cells. Pre- sorted spleens contained 0.3% PC on day 0 (IgDϪ CD38high boxed cells). Postsort analysis shows a maximum of 0.7% PC in IgGϩ MZA B cells, as determined by the frequency of IgGϪCD21ϪCD23Ϫ B cells. by guest on September 25, 2021

Immunohistochemistry (IHC) anti-human IgD-HRP (Southern Biotech) was detected using Vector NovaRED substrate (Vector Laboratories). Tissue samples for IHC were embedded in Tissue-Tek OCT compound (Miles). Six-micrometer-thick sections were cut, fixed in acetone, and Statistics blocked with 1.5% horse serum diluted in PBS/1% BSA. Abs used (diluted in PBS/1% BSA) were anti-human CD3-Alexa Fluor (AF) 647, CD11c- Statistics were preformed using a one-tail, two-sample (assuming unequal (p Ͻ 0.005 ,ءءء ;p Ͻ 0.01 ,ءء ;p Ͻ 0.05 ,ء) AF546, CD27-AF647, IgD-AF546, and IgM-Pacific Blue (BD Pharmin- variances) t test with p value gen). All Abs were directly conjugated using the AF mAb Labeling Kits compared with cultures with no cytokine for each group. (Invitrogen Life Technologies) according to the manufacturer’s instructions. Rabbit anti-human lymphatic endothelial hyaluronan receptor Results (LYVE)-1 Abs (Abcam) were detected with goat anti-rabbit IgG-AF546 (Invitrogen Life Technologies). Cross-talk-free confocal images were ob- Human spleen MZ-like areas contain a novel mem B cell subset tained on a Zeiss LSM 510 equipped with either a ϫ20 or ϫ40 lens and Initially, we analyzed the organization of human splenic B cell com- 405-, 543-, and 633-nm lasers. Figures were prepared using Adobe Pho- toshop (Adobe Systems). Because of uneven brightness of the three fluo- partments. Unlike mouse spleen, in which B cell follicles form around rochrome-labeled Abs, visual insistency was increased to the same extent T cell areas, the white pulp of human spleen consisted largely of ϩ ϩ of all three fluorochromes to delineate the distribution of the individual CD19 B cell aggregates with small patches of CD3 T cells outside markers in the single-color panels. For colored substrate IHC, sections this region (Figs. 1 and 2) (4, 5, 9). The CD19ϩ B cell regions were were first treated with 20% acetic acid to inactivate endogenous AP activity composed of a central follicle containing IgDϩIgMϩ B cells that was and 0.03% H2O2 in PBS to inactivate endogenous peroxidase. Sections were then washed in PBS, and nonspecific binding was blocked with 1.5% often associated with central arterioles (arrows). In the spleens exam- horse serum (diluted in PBS/1% BSA) before incubation with mAbs ined (n ϭ 15), ϳ10% of FO areas included germinal centers that were appropriately diluted in PBS/1% BSA. Mouse mAbs directed against Bcl-6ϩ and contained Ki-67ϩ-proliferating lymphocytes, CD21high human CD3, CD27, IgM, Ki-67 (BD Pharmingen), CD19, CD24, CD38 CD23ϩ follicular DC (FDC), and a few CD3ϩ T cells (Fig. 1 and data (Caltag Laboratories), CD21, CD23 (Leinco), and CD1c (a gift from ϩ Ϫ not shown). Surrounding the follicle was a zone of CD21 CD23 Dr. M. Brenner, Harvard Medical School, Boston, MA) were detected Ϫ ϩ ϩ using the Vectastain ELITE ABC kit (mouse IgG-biotin then avidin- Ki-67 CD27 CD1c B cells (Fig. 1) that was located in a region AP) and the Vector Blue AP substrate (Vector Laboratories). Mouse analogous to the murine MZ. A portion of these MZA B cells were 2876 IL-21 AND BAFF STIMULATE HUMAN MZA B CELLS

FIGURE 5. MZA are hyperresponsive to IL-21 costimulation compared with FO B cells. A and B, Splenic FO and MZA/mem B cells (sorted based on expression of CD23 and CD27), or CD21high MZA B cells (sorted based on high expression of CD21 from a spleen with no CD21highCD23ϩ B cells) were assessed for CD19,

IgD, and CD38 expression after 7 days of stimulation as Downloaded from indicated (A). IgDϪCD38high PC are boxed. B, Prolif- eration (day 3), PC number (day 5), and secreted IgG and IgM (day 7–10) were also determined. MZA B cell responsiveness was assessed only in nil and anti-CD40- stimulated cultures. IgM production by all subpopulations was assessed only in nil and anti-CD40-activated cultures. Data are from one representative experiment http://www.jimmunol.org/ (FO, n ϭ 8; MZA/mem, n ϭ 3; MZA, n ϭ 1). by guest on September 25, 2021

IgMϩIgDϪ/dim and the remainder were IgMϪIgDϪ, presumably MZA B cells brightly express B220 and CD45RA that are nor- postswitched B cells (Fig. 2) (4, 9, 25, 26). No marginal sinus was mally expressed by naive blood B cells and are usually down- present in this region (8, 26, 27), nor is the region vascularized by regulated when CD27 is expressed (29). In addition, CD95 that LYVE-1-expressing endothelium, which is present in the red pulp is normally expressed by mem B cells is only modestly up- (Fig. 2) (28). Importantly, the MZA contained CD3ϩ T cells as well regulated by MZA B cells. However, CXCR4, the receptor for as CD11cϩ DC, which were concentrated at the edge of the CD27ϩ CXCL12, is expressed at a considerably lower density on MZA mem B cell zone (Fig. 2). This collection of mem B cells surrounding B cells compared with conventional splenic mem B cells (Fig. the FO area was not noted in tonsil or lymph nodes (data not shown). 3). CD148 expression, which has previously been shown to be Based on the immunohistology, splenic FO B cells were iden- associated with human MZA B cells (4), was not detected by tiﬁed as CD19ϩCD21ϩCD23ϩ cells and mem MZA B cells were immunohistology (data not shown) but by ﬂow cytometry, and CD19ϩCD21ϩCD23Ϫ. Flow cytometry was used to characterize was found to be expressed by IgDlowIgMhigh MZA, but not these B cell subpopulations more completely. FO B cells were IgGϩ MZA B cells (data not shown). Moreover, MZA B cells found to be small and CD21ϩCD23ϩCD27ϪIgDhighIgMlowIgGϪ were CD80Ϫ/lowCD86Ϫ/low, and routinely CD25ϪCD69ϪHLA- CD148ϪTACIϪBCMAϪBAFFRϩCD95Ϫ (Fig. 3). CD21high DRϩ (data not shown). IL-21R was expressed at low density by CD23Ϫ MZA B cells all expressed the human mem B cell all splenic populations (data not shown). The third population marker, CD27 at high density. This population of B cells con- of B cells was CD21Ϫ/lowCD23Ϫ and comprised a mixture of tained both pre- and postswitched B cells, ϳ70% were IgDlow CD27high PC, CD27ϩ mem B cells, and CD27Ϫ transitional B IgMhigh and 30% were IgGϩ (mean 32%, range 16–53%; n ϭ cells. Taken together, the data demonstrate that MZA B cells 11). Furthermore, these CD27high MZA B cells were large and have a distinct phenotype compared with other splenic B cell predominately CD148ϩTACIhighBAFFRhighCD44highCD95high subpopulations or other mem B cell populations. B220highCD45RAhighCXCR4low compared with FO B cells. This display of phenotypic markers documents that MZA B MZA B cells are hyperresponsive to IL-21 coactivation cells are not only distinct from FO B cells but also from con- To elucidate whether IL-21 differentially impacts the responsive- ventional CD23Ϫ/CD21low/Ϫ splenic mem B cells. CD27high ness of human B cell subpopulations, splenic B cells were initially The Journal of Immunology 2877

FIGURE 6. IL-21 costimulation induces AID and BLIMP-1 expression and PC differentiation from both pre- and postswitched MZA B cells. A and B, Splenic FO, IgGϪ MZA and IgGϩ MZA B cells (Fig. 4B) were cultured as indicated. A, The ability of IL-21 to induced PC differentiation of IgGϪ MZA and IgGϩ MZA B cells was assessed by ﬂow cytometry (CD19ϩIgDϪ CD38high boxed cells) after 5 days of culture. B, Mean Ϯ SEM number of PC (day 5–7), and secreted IgG and IgM (day 7–8) were determined from IgGϪ ϩ MZA cells (n ϭ 6) and IgG MZA B Downloaded from cells (n ϭ 4). AID and BLIMP-1 mRNA (day 3) was determined in nil or anti-CD40-stimulated cultures with or without IL-21 in FO (n ϭ 2), IgGϪ MZA cells (n ϭ 4), and IgGϩ MZA B cells (n ϭ 4). Data were normalized to ␤-2M mRNA and shown as fold http://www.jimmunol.org/ change compared with cultures without stimuli. Mean Ϯ SEM of one representative experiment is shown of three rep- licate RNA samples from the same ,p Ͻ 0.01 ,ءء ,p Ͻ 0.05 ,ء .spleen donor -p Ͻ 0.005 compared with cul ,ءءء and tures with no cytokine for each group. by guest on September 25, 2021

sorted into CD19ϩCD27ϪCD23ϩ naive FO as well as CD19ϩ fested extensive proliferation (Fig. 5B). Finally, MZA B cells not CD27ϩCD23Ϫ mem MZA, which includes a small population of only responded more robustly than FO B cells to IL-21 costimu- conventional mem B cells based on expression of CD19, CD23, lation but also more rapidly. By day 5, MZA B cells costimulated and CD27 (Fig. 4A). Fig. 4A indicates the phenotype of the splenic with IL-21, anti-IgM, and anti-CD40 generated significant num- B cells analyzed and the results of the postsort analysis. Compar- bers of PC (16,329 and 21,900 PC for day 5 and day 7, respec- ison with Fig. 3 indicates the heterogeneity between spleen sam- tively), whereas FO B cells stimulated under identical conditions ples. Although the same B cell populations are found in all spleens required a longer incubation for the differentiation of substantial examined, the relative proportions of the various populations differ numbers of PC (1,501 and 7,715 PC for day 5 and day 7, respec- considerably. The sorted B cell subpopulations were assessed for tively). At all time points examined, the number of PC generated their functional responsiveness. from FO B cells was substantially less then that from MZA/mem As previously described for murine splenic and human blood B or MZA B cells. cells (14, 15), the majority of naive FO B cells were killed by the Because the CD21high MZA B cell population contains both combination of BCR and IL-21 stimulation (Fig. 5A). In contrast, preswitch IgMhighIgDϩ/Ϫ and postswitched IgGϩ mem B cells, the combination of anti-CD40 and IL-21 induced FO B cells to CD19ϩCD21highCD23Ϫ splenic B cells were sorted into IgGϪ or differentiate into CD27highIgDϪCD38high PC. This response was IgGϩ cells (as shown in Fig. 4B) to assess their function. Both enhanced by engagement of the BCR (Fig. 5A). Compared with IgGϪ and IgGϩ MZA B cells were induced to differentiate into PC FO B cells, PC differentiation was significantly greater when either and produce IgG following stimulation with IL-21 and anti-CD40 CD21low/high MZA/mem B cells were stimulated with anti-CD40 or anti-CD40 and anti-IgM (Fig. 6, A and B), demonstrating that and IL-21, and BCR engagement did not enhance this response IL-21 coactivation induces both CSR and PC differentiation. IL-21 (Fig. 5A). Similar results were obtained when MZA splenic B cells costimulation of IgGϪ MZA B cells induced considerable levels of were sorted based on high expression of CD21 that excluded the IgM, especially when both CD40 and the BCR were engaged, as CD21Ϫ/low conventional mem B cell population (Fig. 5A). Fur- well as low levels of IgG (Fig. 6B). Importantly, postswitched thermore, MZA/mem and MZA B cells generated significantly IgGϩ MZA B cells generated significantly more PC and secreted more PC and produced greater amounts of Ig than FO B cells in much greater amounts of IgG in response to IL-21 costimulation response to IL-21 costimulation, although all populations mani- compared with IgGϪ MZA and FO B cells (Fig. 6B). Notably, in 2878 IL-21 AND BAFF STIMULATE HUMAN MZA B CELLS

FIGURE 7. BAFF/BLyS does not increase FO B cell responsiveness to IL-21. A, Total CD19ϩ splenic B cells were cultured as indicated, and mean Ϯ SEM number of total B cells was determined on day 7 (n ϭ 4) using Accu- p Ͻ 0.05 compared ,ء .Count Particles with nil for each group. B, CD21ϩ CD23ϩ FO B cells were sorted as shown in Fig. 4B, and mean Ϯ SEM number of total B cells was determined on day 7 of culture in three independent ,p Ͻ 0.01 ,ءء ,p Ͻ 0.05 ,ء .experiments

p Ͻ 0.005 compared with cul- Downloaded from ,ءءء and tures with no cytokine for each group. C, FO B cells were cultured as indicated for 7 days. Cell phenotype was determined with anti-IgD and anti- CD38 on CD19ϩ cells. The numbers represent the total number of PC (IgDϪ

CD38high boxed cells). D, mRNA was http://www.jimmunol.org/ puriﬁed from total CD19ϩ B cells following 3 days of culture as indicated, and Bim mRNA was normalized to ␤-2M mRNA and shown as fold change compared with cultures without stimuli. Data represent one of two independent experiments for splenic CD19ϩ B cells and one with sorted splenic FO B cells. by guest on September 25, 2021

all populations, IL-21- and anti-CD40 costimulation induced AID Finally, IL-21 alone induced IgGϩ MZA B cells to produce low and BLIMP-1 up-regulation (Fig. 6B), although human MZA B levels of IgG (mean 435 ng/ml; n ϭ 9) and signiﬁcantly up-reg- cells have previously been reported not to express AID in situ (30). ulates BLIMP-1, but not AID, expression (Fig. 6B). These results

FIGURE 8. IL-21 and BAFF/BLyS stimulate splenic B cells in the absence of further costimulation. Splenic CD19ϩ B cells were isolated and cultured as indicated. Mean Ϯ SEM number of PC was determined on day 5 (n ϭ 4) and IgG secretion after 10–12 days (n ϭ 4). After 3 days of culture, RNA was isolated and AID and BLIMP-1 mRNA were quantified, and data were normalized to ␤-2M mRNA and shown as fold change compared with cultures without stimuli. Data are representative re- p Ͻ ,ءء .sults of one of three similar experiments p Ͻ 0.005 compared with cultures ,ءءء and 0.01 with no cytokine for each group. The Journal of Immunology 2879 Downloaded from http://www.jimmunol.org/ FIGURE 10. BAFF/BLyS does not affect B cell responsiveness to IL-21 and anti-CD40 activation. Splenic B cells were sorted as described in Fig. 9 and day 0 shown in Fig. 4. The cells were cultured as indicated with cytokines and stimuli. After 5 days of culture, the cells were harvested and stained with anti-IgD, anti-CD38, and anti-CD19. The numbers represent the number of PC (CD19ϩIgDϪCD38high boxed cells) as determined with AccuCount Particles. Data are representative of results of one of three FIGURE 9. IL-21 and BAFF/BLyS synergize to induce PC differentiation similar experiments. from IgGϩ MZA B cells. A, Splenic CD19ϩ FO B cells or total MZA and by guest on September 25, 2021 mem B cells (Fig. 4A) were cultured as indicated. After 8 days, cell surface phenotype was determined from two independent experiments. Numbers in- vival of FO B cells in the absence of IL-21, it did not significantly dicate absolute numbers of PC (CD19ϩIgDϪCD38high boxed cells). B and C, ( p ϭ 0.20) decrease the death of FO B cells stimulated by anti- FO, IgGϪ MZA, and IgGϩ MZA B cells (Fig. 4B) were cultured as indicated. IgM and IL-21 (Fig. 7B). Because Bim has been implicated in Mean Ϯ SEM number of PC after 5–7 days (n ϭ 5) and IgG production after IL-21-mediated cell death in mice (32) and BAFF/BLyS has been 7 days (n ϭ 5) were determined. D, RNA was isolated after 3 days, and mRNA shown to down-regulate Bim (33), we examined the impact of for BLIMP-1 (n ϭ 3) and AID (n ϭ 2) was determined as described in Fig. 6. BAFF/BLyS on Bim expression by splenic B cells stimulated with Ͻ ءءء Ͻ ء , p 0.05 and , p 0.005 compared with cultures with no cytokine for IL-21. As expected from the functional results, neither IL-21 nor each group. BAFF/BLyS affected expression of Bim in total splenic CD19ϩ B cells or purified FO B cells alone or after stimulation with anti-IgM (Fig. 7D and data not shown). indicate that IgGϩ MZA B cells are responsive to IL-21 in the absence of further costimulation. IL-21 and BAFF/BLyS synergize to induce PC differentiation from IgGϩ MZA B cells BAFF/BLyS does not promote naive FO B cell responsiveness Although BAFF/BLyS did not increase the responsiveness of FO to IL-21 B cells to IL-21, strikingly, the combination of IL-21 and BAFF/ FO B cells respond poorly to the combination of anti-CD40 and BLyS in the absence of other stimulation or with anti-IgM induced IL-21 and died when stimulated with anti-IgM and IL-21. There- PC differentiation from CD19ϩ splenic B cells (Fig. 8). Moreover, fore, we sought to determine whether provision of a survival factor the combination of IL-21 and BAFF/BLyS also induced significant might enhance the responsiveness of FO B cells to IL-21R signal- IgG production from total splenic CD19ϩ B cells regardless of the ing. BAFF/BLyS is known to promote the survival of certain hu- addition of anti-IgM (Fig. 8). IL-21 and BAFF/BLyS-induced PC man B cell subsets (20, 31). We therefore addressed whether differentiation of CD19ϩ B cells was preceded by induction of BAFF/BLyS could promote responsiveness to IL-21. As shown in BLIMP-1 mRNA (Fig. 8). In contrast, AID mRNA was not in- Fig. 7A and B, BAFF/BLyS clearly increased survival of total duced by the combination of IL-21 and BAFF/BLyS, but rather splenic CD19ϩ B cells and FO B cells, that were maintained with- was noted following stimulation with IL-21 and anti-IgM with or out other stimuli or in the presence of anti-IgM. without BAFF/BLyS (Fig. 8). Finally, BAFF/BLyS did not in- However, BAFF/BLyS did not increase the responsiveness of crease the responsiveness of CD19ϩ B cells to IL-21 and anti- these cells following stimulation with anti-CD40 in the presence of CD40 with or without anti-IgM, nor did it induce PC differentia- IL-21 (Fig. 7, A and B), but did marginally increase the number tion from B cells costimulated with anti-CD40 (Fig. 8). of PC generated from FO B cells stimulated with IL-21, anti-IgM, Because BAFF/BLyS had minimal effect on the responsiveness and anti-CD40 (Fig. 7C). Although BAFF/BLyS increased sur- of naive FO B cells, we examined the effect of this cytokine on 2880 IL-21 AND BAFF STIMULATE HUMAN MZA B CELLS

FIGURE 11. IL-21 and BAFF/ BLyS uniquely induce PC differentiation compared with other cytokine combinations. A, Splenic CD19ϩ B cells were isolated and cultured with cytokines and stimuli as indicated. PC number and IgG was determined after 5 and 7 days, respectively. B, CD19ϩ CD21highCD23Ϫ MZA B cells were sorted and cultured as indicated. After 5 and 7 days, cell surface phenotype and secreted IgG, respectively, were determined. Numbers indicate absolute numbers of PC. Forty-eight, 291, 292, 39, and 142 PC were found in control cultures with no stimulation and nil, IL-2, IL-21, BAFF/BLyS, and IL-10, respectively. Downloaded from

CD23ϪCD27ϩ MZA B cells that contained small numbers of BAFF/BLyS had little effect on PC differentiation of any splenic B CD23ϪCD27ϩ conventional mem B cells. We found that culture cell populations stimulated with anti-CD40 and IL-21 with or

with BAFF/BLyS alone increased the survival of both FO and without BCR engagement (Fig. 10). Importantly, BAFF/BLyS also http://www.jimmunol.org/ MZA/mem B cells, but only induced minimal PC differentiation did not induce PC differentiation from splenic B cell subsets stim- (Fig. 9A). Culture with IL-21 alone had no significant effect on B ulated with anti-CD40 or anti-IgM and anti-CD40 in the absence cell survival, and only small numbers of PC were induced from of IL-21 (Figs. 7 and 8). MZA/mem B cells (Fig. 9A and data not shown). Strikingly, how- Notably, the BAFFRs TACI and BAFFR were expressed at high ever, the combination of IL-21 and BAFF/BLyS induced MZA/ density on MZA B cells (Fig. 3) (34). Because both BAFF/BLyS mem (but not FO) B cells to differentiate into PC in the absence of and APRIL bind TACI, we cultured IgGϩ MZA B cells with further costimulation (Fig. 9A). Similar results were obtained when APRIL and IL-21. In three independent experiments, APRIL and MZA splenic B cells were sorted based on high expression of IL-21 induced only a fraction of the PC differentiation and Ig se- CD21, which excluded the CD21Ϫ/low conventional mem B cell cretion (3, 5, and 33%) of cells stimulated by IL-21 and BAFF/ by guest on September 25, 2021 population. When MZA B cells were purified from conventional BLyS. These data suggest that BAFFR (which does not bind mem B cells and subdivided into pre- and postswitched B cells, APRIL) is likely to be the major receptor inducing PC differenti- culture with IL-21 alone, but not BAFF/BLyS was again noted to ation and IgG production. However, signaling through TACI alone induce low levels of PC differentiation and IgG production from or combined engagement of both TACI and BAFFR may provide IgGϩ MZA B cells (Fig. 9, B and C). Notably, however, IL-21 optimal signaling for PC differentiation. synergized with BAFF/BLyS resulting in substantial PC differentiation and IgG production from IgGϩ MZA B cells in the absence Other cytokine combinations do not induce extensive PC of further costimulation (Fig. 9, B and C). In contrast, only min- differentiation of MZA B cells compared with IL-21 imal numbers of PC and minimal IgM or IgG secretion were gen- and BAFF/BLyS erated from IgGϪ MZA B cells stimulated with the combination of The combination of IL-2 and IL-10 is a known inducer of PC from IL-21 and BAFF/BLyS (Fig. 9, B and C). Moreover, compared mem splenic B cells costimulated via CD40 (4). Moreover, as with IgGϩ MZA B cells, CD27ϩCD21ϩ/lowCD23ϪIgGϩ conven- shown in Fig. 11, A and B, the ability of IL-2 and IL-10 alone to tional splenic mem B cells responded less well to IL-21 and BAFF/ induce PC differentiation from CD21high MZA was significantly BLyS or IL-21 and anti-CD40 with regard to both PC differenti- less than that of IL-21 and BAFF/BLyS. Notably, the combination ation and IgG production (Fig. 10 and data not shown). of IL-21 and BAFF/BLyS induced comparable numbers of PC and The ability of IL-21 and BAFF/BLyS to induce PC differentiation levels of IgG secretion from MZA B cells as the combination of from IgGϩ MZA B cells was not related to sorting with anti-IgG, and IL-2, IL-10, and anti-CD40 (Fig. 11B). Furthermore, we found that thus signaling through the BCR, because IL-21 and BAFF/BLyS in- none of the other cytokine combinations we analyzed resulted in duced Ig production from total splenic B cells and total MZA B cells significant PC differentiation from total CD19ϩ splenic B cells not purified with anti-IgG (Figs. 8 and 9). Moreover, costimulation of (Fig. 11A). These data support the conclusion that the combination IgGϪ (IgMhigh) MZA B cells with IL-21, BAFF/BLyS, and anti-IgM of IL-21 and BAFF/BLyS is unique in its ability to induce PC did not induce large numbers of PC (Fig. 10). differentiation from MZA B cells. IL-21 and BAFF/BLyS-induced PC differentiation was preceded by up-regulation of expression of BLIMP-1 mRNA in both pop- Discussion ulations of MZA B cells, although the magnitude of BLIMP-1 The current results indicate that a unique population of increase was greater in the IgGϩ subpopulation (Fig. 9D). AID IgGϩCD21highCD23lowCD27high mem B cells resides in the “MZ- was not induced in IgGϩ MZA B cells by the combination of IL-21 like” structure of human spleen. Previously, these cells have not and BAFF/BLyS. been examined (9, 35), although they clearly are an important Despite the finding that the combination of IL-21 with BAFF/ component of the human MZA. Phenotypically, these post- BLyS induced marked PC differentiation from IgGϩ MZA B cells, switched IgGϩ cells are very similar to preswitched MZA B cells, The Journal of Immunology 2881 differing only in BCR isotype and expression of CD148. Notably, rine MZ. There is no marginal sinus and no metallophilic macro- however, this is the first demonstration that IgGϩ MZA B cells are phages, although resident stromal cells may contribute to migration, functionally distinct. A unique feature of these cells is their general retention, and responsiveness of lymphocytes (28, 43). The peri-FO hyperresponsiveness to IL-21 and their capacity to respond exclu- zone just outside of the MZA, however, is vascularized and contains sively to the combination of IL-21 and BAFF/BLyS with rapid and sialoadhesin-expressing macrophages (43). Importantly, human MZA substantial differentiation of PC. Neither of these functional activ- B cells are also considerably different than murine MZ B cells in that ities can be explained by the density of IL-21R expression, which they are CD27high mem B cells consisting of both pre- and post- did not vary between pre- and postswitched MZA B cells. switched cells (Refs. 4, 26, and this study). As opposed to the murine In most circumstances, human B cells require engagement of CD40 MZ, which are largely nonmigratory cells, human IgDϩ MZA are and/or BCR to respond with significant PC differentiation. However, thought to appear in the peripheral circulation (9). Despite the marked IgGϩ MZA mem B cells are unique in requiring only signals from T differences between murine MZ and human MZA there are certain cell-derived IL-21 and DC-derived BAFF/BLyS for this response. similarities, including the surface phenotype (CD21highCD23low/Ϫ) IL-21 is a well-known T cell-derived cytokine that can costimulate and the ability of murine MZ and human IgDϩCD27high MZA B cells differentiation of human B cells into PC (15, 16), but previously has to respond to TI-2 (7). Finally, both populations can respond to been shown to require activation signals induced through engagement BAFF/BLyS stimulation and appear to be hyperresponsive to IL-21 of CD40 or CD40 and the BCR for this response. In humans, DC can costimulation (Ref. 42 and this study). Thus, in our preliminary stud- induce PC differentiation (21–23), and BAFF/BLyS and APRIL ex- ies, we found that murine MZ B cells respond robustly to IL-21 co- pressed by activated DC also can contribute to CSR (24). Importantly, stimulation (data not shown). Interestingly, mice that acutely (IL-21

BAFF/BLyS shares certain signaling pathways with CD40 (36, 37). plasmid-injected) or constitutively (BXSB yaa mice) express high Downloaded from In the IgGϩ MZA B cell population, BAFF/BLyS may mimic certain levels of IL-21 do not have MZ B cells (14, 44), perhaps because of CD40 signaling capabilities, and therefore allow synergistic interac- IL-21-induced PC differentiation. tions between IL-21 and BAFF/BLyS. Support for this concept de- The postswitched human IgGϩ MZA B cells have not previ- rives from the observation that BAFF/BLyS appears to have little ously been examined. These cells appear to reside at the outer edge effect on B cell responsiveness or survival when cells are also stim- of the MZA structure in a region interspersed with CD3ϩ T cells ϩ ulated through CD40 or activated via anti-CD40 and IL-21, suggest- and CD11c DC. They are a polyclonal population with heavily http://www.jimmunol.org/ ing that overlapping signaling pathways with CD40 may explain its mutated Ig genes (5, 6, 39). Importantly, one might assume that the activity. BAFF/BLyS has also been shown to promote the survival, MZA is a survival niche for these cells that reflect the humoral but not proliferation, of human splenic plasmablasts by preventing mem of the subject. The finding that they can vigorously respond their apoptosis (31). Such a mechanism could contribute to the current to the combination of IL-21 and BAFF/BLyS suggests that any results, although a number of findings make this possibility unlikely. stimulus that can simultaneously induce production of these cyto- First, stimulation with BAFF/BLyS did not induce PC differentiation kines can enhance serologic mem polyclonally by causing rapid or IgG production above background even when PC were present in induction of PC differentiation from IgGϩ MZA. Our previous the initial cultured cell population. Secondly, BAFF/BLyS and IL-21 finding that IL-21 induces differentiation of long-lived nondividing increased the generation of PC in experiments in which IL-21 induced PC (15) suggests that the contribution of IL-21 and BAFF/BLyS to by guest on September 25, 2021 minimal PC differentiation alone. Thirdly, the addition of BAFF/ serologic mem may be persistent and not transient, as is seen with BLyS to IL-21-stimulated cultures was essential for the induction of stimulation of murine MZ B cells. BLIMP-1 expression that is required for PC differentiation. In our Importantly, we show that IgGϩ MZA B cells are in close associ- experiments with sorted B cells, no PC were present in the initial ation with T cells and DC that are potentially capable of producing culture. Thus, the combination of BAFF and IL-21 is unlikely to be IL-21 and BAFF/BLyS, respectively. This proximity suggests that effective merely by increasing survival of PC present in the initial locally produced factors can stimulate IgGϩ MZA B cells to differ- culture. Finally, the combination of IL-21 and BAFF/BLyS led to a entiate into PC. Alternatively, because both IL-21 and BAFF/BLyS significant increase in the absolute number of PC that could not be can be detected in the plasma in autoimmune disease (Ref. 45 and our explained by an influence on plasmablast apoptosis rather than pre- unpublished results), MZA B cells may be poised to react in an en- cursor MZA B cell expansion. These results strongly suggest that the docrine manner to products of distantly activated T cells and DC. In combination of BAFF/BLyS exerts a synergistic effect to promote PC humans, BAFF/BLyS has also been shown to be produced by FDC differentiation rather than plasmablast survival. (46, 47). These and our results suggest that FO helper T cells (which The white pulp area of both the murine and human spleen contains spontaneously produce IL-21 (11), and BAFF/BLyS-expressing FDC a population of B cells that surround the B cell FO area. In the mouse, could also play a role in driving the rapid differentiation of MZA mem these MZ B cells are largely nonmigratory, naive, unactivated B cells B cell responses in an Ag-independent manner. These considerations with the phenotype, IgDϪIgMhighCD21highCD23low (7, 38). Murine raise the intriguing possibility that the induction of Ag-specific ger- MZ B cells are especially responsive to TI-2 with the capacity to minal center responses may lead to increased production of IL-21 differentiate rapidly into short-lived IgM-secreting PC, but manifest from Ag-specific T cells and BAFF/BLyS from FDC and thereby minimal CSR, somatic hypermutation, and generation of mem to TI-2 drive polyclonal activation of MZA mem B cells in adjacent peri-FO (7, 38). However, these cells can undergo CSR, somatic hypermuta- areas. This possibility is consistent with two previous observations. tion, and PC differentiation in response to T-dependent Ags (39, 40). First, the MZA region appears to be more prominent around follicles Moreover, in the rat, mem B cells with mutated Ig genes have been containing germinal centers (26), and second, specific immunization reported to migrate to the MZ following immunization (41). Impor- is frequently associated with polyclonal hyperglobulinemia in both tantly, the murine MZ is more than a collection of B cells, but rather mice and humans (48–50). an organized anatomic structure immediately outside the marginal In summary, our results strongly suggest that human splenic sinus that contains a specific population of metallophilic macrophages IgGϩ MZA B cells are a reservoir of cells that reflect the natural (38). This architecture appears to contribute to the function of MZ B history of Ag exposure of the individual. Moreover, these IgGϩ cells because it fosters the rapid entry of bacteria-containing DC that MZA B cells can respond to soluble factors from T cells and DC (or can facilitate MZ B cell differentiation in a BAFF/BLyS-dependent FDC) with rapid differentiation of PC and, therefore, are prime can- manner (42). The human MZA is considerably different than the mu- didates to replenish serologic mem in an Ag-independent manner. 2882 IL-21 AND BAFF STIMULATE HUMAN MZA B CELLS

Acknowledgments 24. Litinskiy, M. B., B. Nardelli, D. M. Hilbert, B. He, A. Schaffer, P. Casali, and A. Cerutti. 2002. DCs induce CD40-independent immunoglobulin class switch- We gratefully acknowledge James Simone and Rafael Villasmil of the ing through BLyS and APRIL. Nat. Immunol. 3: 822–829. National Institute of Arthritis and Musculoskeletal and Skin Diseases Flow 25. Timens, W., A. Boes, and S. Poppema. 1989. Human marginal zone B cells are Cytometry Facility for sorting cell subsets. All imaging was done at the not an activated B cell subset: strong expression of CD21 as a putative mediator National Institute of Arthritis and Musculoskeletal and Skin Diseases Light for rapid B cell activation Eur. J. Immunol. 19: 2163–2166. 26. Steiniger, B., E. M. Timphus, R. Jacob, and P. J. Barth. 2005. CD27ϩ B cells in Imaging section, with the assistance of Evelyn Ralston and Kristien Zaal. human lymphatic organs: re-evaluating the splenic marginal zone. Immunology 116: 429–442. Disclosures 27. Du, M. Q., H. Z. Peng, A. Dogan, T. C. Diss, H. Liu, L. X. Pan, R. P. Moseley, The authors have no financial conflict of interest. M. J. Briskin, J. K. Chan, and P. G. Isaacson. 1997. Preferential dissemination of B-cell gastric mucosa-associated lymphoid tissue (MALT) lymphoma to the splenic marginal zone. Blood 90: 4071–4077. References 28. Martinez-Pomares, L., L. G. Hanitsch, R. Stillion, S. Keshav, and S. Gordon. 1. Lanzavecchia, A., N. Bernasconi, E. Traggiai, C. R. Ruprecht, D. Corti, and 2005. Expression of mannose receptor and ligands for its cysteine-rich domain in F. Sallusto. 2006. Understanding and making use of human memory B cells. venous sinuses of human spleen. Lab. Invest. 85: 1238–1249. Immunol. Rev. 211: 303–309. 29. Bleesing, J. J. H., and T. A. Fleisher. 2003. Human B cells express a CD45 2. Bernasconi, N. L., E. Traggiai, and A. Lanzavecchia. 2002. Maintenance of se- isoform that is similar to murine B220 and is downregulated with acquisition of rological memory by polyclonal activation of human memory B cells. Science the memory B-cell marker CD27. Cytometry B Clin. Cytom. 51: 1–8. 298: 2199–2202. 30. Willenbrock, K., B. Jungnickel, M. L. Hansmann, and R. Kuppers. 2005. Human 3. McHeyzer-Williams, L. J., M. Cool, and M. G. McHeyzer-Williams. 2000. An- splenic marginal zone B cells lack expression of activation-induced cytidine Ϫ tigen-specific B cell memory: expression and replenishment of a novel B220 deaminase. Eur. J. Immunol. 35: 3002–3007. memory B cell compartment. J. Exp. Med. 191: 1149–1166. 31. Avery, D. T., S. L. Kalled, J. I. Ellyard, C. Ambrose, S. A. Bixler, M. Thien, 4. Tangye, S. G., Y. J. Liu, G. Aversa, J. H. Phillips, and J. E. de Vries. 1998. R. Brink, F. Mackay, P. D. Hodgkin, and S. G. Tangye. 2003. BAFF selectively

Identiﬁcation of functional human splenic memory B cells by expression of enhances the survival of plasmablasts generated from human memory B cells. Downloaded from CD148 and CD27. J. Exp. Med. 188: 1691–1703. J. Clin. Invest. 112: 286–297. 5. Dunn-Walters, D. K., P. G. Isaacson, and J. Spencer. 1995. Analysis of mutations 32. Jin, H., R. Carrio, A. Yu, and T. R. Malek. 2004. Distinct activation signals in immunoglobulin heavy chain variable region genes of microdissected marginal determine whether IL-21 induces B cell costimulation, growth arrest, or Bim- zone (MGZ) B cells suggests that the MGZ of human spleen is a reservoir of dependent apoptosis. J. Immunol. 173: 657–665. memory B cells. J. Exp. Med. 182: 559–566. 33. Craxton, A., K. E. Draves, A. Gruppi, and E. A. Clark. 2005. BAFF regulates B 6. Weller, S., C. A. Reynaud, and J. C. Weill. 2005. Splenic marginal zone B cells cell survival by downregulating the BH3-only family member Bim via the ERK in humans: where do they mutate their Ig receptor. Eur. J. Immunol. 35: pathway. J. Exp. Med. 202: 1363–1374. 2789–2792.

34. Ng, L. G., A. P. Sutherland, R. Newton, F. Qian, T. G. Cachero, M. L. Scott, http://www.jimmunol.org/ 7. Pillai, S., A. Cariappa, and S. T. Moran. 2005. Marginal zone B cells. Annu. Rev. J. S. Thompson, J. Wheway, T. Chtanova, J. Groom, et al. 2004. B cell-activating Immunol. 23: 161–196. factor belonging to the TNF family (BAFF)-R is the principal BAFF receptor 8. Spencer, J., M. E. Perry, and D. K. Dunn-Walters. 1998. Human marginal-zone facilitating BAFF costimulation of circulating T and B cells. J. Immunol. 173: B cells. Immunol. Today 19: 421–426. 807–817. 9. Weller, S., M. C. Braun, B. K. Tan, A. Rosenwald, C. Cordier, M. E. Conley, A. Plebani, D. S. Kumararatne, D. Bonnet, O. Tournilhac, et al. 2004. Human 35. Shen, Y., J. Iqbal, L. Xiao, R. C. Lynch, A. Rosenwald, L. M. Staudt, S. Sherman, blood IgM “memory” B cells are circulating splenic marginal zone B cells har- K. Dybkaer, G. Zhou, J. D. Eudy, et al. 2004. Distinct gene expression profiles boring a prediversified immunoglobulin repertoire. Blood 104: 3647–3654. in different B-cell compartments in human peripheral lymphoid organs. BMC 10. Parrish-Novak, J., S. R. Dillon, A. Nelson, A. Hammond, C. Sprecher, Immunol. 5: 20–37. J. A. Gross, J. Johnston, K. Madden, W. Xu, J. West, et al. 2000. Interleukin 21 36. Qian, Y., J. Qin, G. Cui, M. Naramura, E. C. Snow, C. F. Ware, R. L. Fairchild, and its receptor are involved in NK cell expansion and regulation of lymphocyte S. A. Omori, R. C. Rickert, M. Scott, B. L. Kotzin, and X. Li. 2004. Act1, a negative function. Nature 408: 57–63. regulator in CD40- and BAFF-mediated B cell survival. Immunity 21: 575–587. 11. Chtanova, T., S. G. Tangye, R. Newton, N. Frank, M. R. Hodge, M. S. Rolph, and 37. Lin-Lee, Y. C., L. V. Pham, A. T. Tamayo, L. Fu, H. J. Zhou, L. C. Yoshimura, by guest on September 25, 2021 C. R. Mackay. 2004. T follicular helper cells express a distinctive transcriptional G. L. Decker, and R. J. Ford. 2006. Nuclear localization in the biology of the profile, reflecting their role as non-Th1/Th2 effector cells that provide help for B CD40 receptor in normal and neoplastic human B lymphocytes. J. Biochem. 281: cells. J. Immunol. 173: 68–78. 18878–18887. 12. Leonard, W. J., and R. Spolski. 2005. Interleukin-21: a modulator of lymphoid 38. Mebius, R. E., and G. Kraal. 2005. Structure and function of the spleen. Nat. Rev. proliferation, apoptosis and differentiation. Nat. Rev. Immunol. 5: 688–698. Immunol. 5: 606–616. 13. Ozaki, K., R. Spolski, C. G. Feng, C. F. Qi, J. Cheng, A. Sher, H. C. Morse, III, 39. Song, H., and J. Cerny. 2003. Functional heterogeneity of marginal zone B cells C. Liu, P. L. Schwartzberg, and W. J. Leonard. 2002. A critical role for IL-21 in revealed by their ability to generate both early antibody-forming cells and ger- regulating immunoglobulin production. Science 298: 1630–1634. minal centers with hypermutation and memory in response to a T-dependent 14. Ozaki, K., R. Spolski, R. Ettinger, H. P. Kim, G. Wang, C. F. Qi, P. Hwu, antigen. J. Exp. Med. 198: 1923–1935. D. J. Shaffer, S. Akilesh, D. C. Roopenian, et al. 2004. Regulation of B cell 40. Attanavanich, K., and J. F. Kearney. 2004. Marginal zone, but not follicular B differentiation and plasma cell generation by IL-21, a novel inducer of Blimp-1 cells, are potent activators of naive CD4 T cells. J. Immunol. 172: 803–811. and Bcl-6. J. Immunol. 173: 5361–5371. 41. Liu, Y. J., S. Oldfield, and I. C. MacLennan. 1988. Memory B cells in T cell- 15. Ettinger, R., G. P. Sims, A. M. Fairhurst, R. Robbins, Y. S. da Silva, R. Spolski, dependent antibody responses colonize the splenic marginal zones. Eur. J. Im- W. J. Leonard, and P. E. Lipsky. 2005. IL-21 induces differentiation of human munol. 18: 355–362. naive and memory B cells into antibody-secreting plasma cells. J. Immunol. 176: 42. Balazs, M., F. Martin, T. Zhou, and J. Kearney. 2002. Blood dendritic cells 7867–7879. interact with splenic marginal zone B cells to initiate T-independent immune 16. Pene, J., J. F. Gauchat, S. Lecart, E. Drouet, P. Guglielmi, V. Boulay, A. Delwail, responses. Immunity 17: 341–352. D. Foster, J. C. Lecron, and H. Yssel. 2004. Cutting edge: IL-21 is a switch factor 43. Steiniger, B., P. Barth, and A. Hellinger. 2001. The perifollicular and marginal for the production of IgG1 and IgG3 by human B cells. J. Immunol. 172: zones of the human splenic white pulp: do fibroblasts guide lymphocyte immi- 5154–5157. gration. Am. J. Pathol. 159: 501–512. 17. Arpin, C., J. Banchereau, and Y.-J. Liu. 1997. Memory B cells are biased towards 44. Amano, H., E. Amano, T. Moll, D. Marinkovic, N. Ibnou-Zekri, E. Martinez-Soria, terminal differentiation: a strategy that may prevent repertoire freezing. J. Exp. I. Semac, T. Wirth, L. Nitschke, and S. Izui. 2003. The Yaa mutation promoting Med. 186: 931–940. murine lupus causes defective development of marginal zone B cells. J. Immunol. 18. Arpin, C., J. Dechanet, C. Van Kooten, P. Merville, G. Grouard, F. Briere, 170: 2293–2301. J. Banchereau, and Y. J. Liu. 1995. Generation of memory B cells and plasma 45. Stohl, W. 2005. BlySfulness does not equal blissfulness in systemic lupus ery- cells in vitro. Science 268: 720–722. thematosus: a therapeutic role for BLyS antagonists. Curr. Dir. Autoimmun. 8: 19. Vinuesa, C. G., M. C. Cook, C. Angelucci, V. Athanasopoulos, L. Rui, 289–304. K. M. Hill, D. Yu, H. Domaschenz, B. Whittle, T. Lambe, et al. 2005. A RING- type ubiquitin ligase family member required to repress follicular helper T cells 46. Zhang, X., C. S. Park, S. O. Yoon, L. Li, Y. M. Hsu, C. Ambrose, and Y. S. Choi. and autoimmunity. Nature 435: 452–458. 2005. BAFF supports human B cell differentiation in the lymphoid follicles 20. Ng, L. G., C. R. Mackay, and F. Mackay. 2005. The BAFF/APRIL system: life through distinct receptors. Int. Immunol. 17: 779–788. beyond B lymphocytes. Mol. Immunol. 42: 763–772. 47. Hase, H., Y. Kanno, M. Kojima, K. Hasegawa, D. Sakurai, H. Kojima, N. Tsuchiya, 21. Fayette, J., I. Durand, J. M. Bridon, C. Arpin, B. Dubois, C. Caux, Y. J. Liu, K. Tokunaga, N. Masawa, M. Azuma, et al. 2004. BAFF/BLyS can potentiate B-cell J. Banchereau, and F. Briere. 1998. Dendritic cells enhance the differentiation of selection with the B-cell coreceptor complex. Blood 103: 2257–2265. naive B cells into plasma cells in vitro. Scand. J. Immunol. 48: 563–570. 48. Ratcliffe, M. J., and M. H. Julius. 1983. Two classes of bystander B cell response: 22. Bjorck, P., L. Flores-Romo, and Y. J. Liu. 1997. Human interdigitating dendritic activation requirements reflect those of B cells in general. J. Immunol. 131: 581–586. cells directly stimulate CD40-activated naive B cells. Eur. J. Immunol. 27: 49. Moticka, E. J., and J. W. Streilein. 1978. Hypothesis: nonspecific polyclonal 1266–1274. activation of memory B cells by antigen as a mechanism for the preservation of 23. Jego, G., A. K. Palucka, J. P. Blanck, C. Chalouni, V. Pascual, and J. Banchereau. long term immunologic anamnesis. Cell. Immunol. 41: 406–413. 2003. Plasmacytoid dendritic cells induce plasma cell differentiation through type 50. Hyypia, T., J. Eskola, M. Laine, A. Salmi, and O. Meurman. 1985. Polyclonal I interferon and interleukin 6. Immunity 19: 225–234. activation of B cells during rubella infection. Scand. J. Immunol. 21: 615–617.