Peripheral B Cell Tolerance and Function in Transgenic Mice Expressing an IgD Superantigen

This information is current as Bao Hoa Duong, Takayuki Ota, Djemel Aït-Azzouzene, of October 2, 2021. Miyo Aoki-Ota, José Luis Vela, Christoph Huber, Kevin Walsh, Amanda L. Gavin and David Nemazee J Immunol 2010; 184:4143-4158; Prepublished online 15 March 2010; doi: 10.4049/jimmunol.0903564 http://www.jimmunol.org/content/184/8/4143 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2010/03/15/jimmunol.090356

Material 4.DC1 http://www.jimmunol.org/ References This article cites 120 articles, 67 of which you can access for free at: http://www.jimmunol.org/content/184/8/4143.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on October 2, 2021

• No Triage! Every submission reviewed by practicing scientists

• 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 © 2010 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Peripheral B Cell Tolerance and Function in Transgenic Mice Expressing an IgD Superantigen

Bao Hoa Duong,*,† Takayuki Ota,* Djemel Aı¨t-Azzouzene,*,1 Miyo Aoki-Ota,* Jose´ Luis Vela,*,†,2 Christoph Huber,* Kevin Walsh,* Amanda L. Gavin,* and David Nemazee*

Transitional B cells turn over rapidly in vivo and are sensitive to apoptosis upon BCR ligation in vitro. However, little direct evidence addresses their tolerance sensitivity in vivo. A key marker used to distinguish these cells is IgD, which, through alternative RNA splicing of H chain transcripts, begins to be coexpressed with IgM at this stage. IgD is also expressed at high levels on naive follicular (B-2) and at lower levels on marginal zone and B-1 B cells. In this study, mice were generated to ubiquitously express a membrane- bound IgD-superantigen. These mice supported virtually no B-2 development, a greatly reduced marginal zone B cell population,

but a relatively normal B-1 compartment. B cell development in the spleen abruptly halted at the transitional B cell population 1 to 2 Downloaded from stage, a block that could not be rescued by either Bcl-2 or BAFF overexpression. The developmentally arrested B cells appeared less mature and turned over more rapidly than nontransgenic T2 cells, exhibiting neither conventional features of anergy nor appreciable receptor editing. Paradoxically, type-2 T-independent responses were more robust in the transgenic mice, although T-dependent responses were reduced and had skewed IgL and IgH isotype usages. Nevertheless, an augmented memory response to secondary challenge was evident. The transgenic mice also had increased serum IgM, but diminished IgG, levels mirrored + by the increased numbers of IgM plasma cells. This model should facilitate studies of peripheral B cell tolerance, with the http://www.jimmunol.org/ advantages of allowing analysis of polyclonal populations, and of B cells naturally lacking IgD. The Journal of Immunology, 2010, 184: 4143–4158.

ince shortly after the discovery of IgD, its main role in the of IgD expression to be regulated by RNA splicing is evolutionarily immune system has been believed to be as a BCR (1–6) (for ancient and probably predates the process of class switching (14). S review see Ref. 7). IgD is not appreciably expressed by Surface expression of IgD varies with B cell subset. It is highest early developing B lymphocytes of the bone marrow (BM) or end in long-lived recirculating B cells that are abundant in follicles of stage Ab-secreting plasma cells. Instead, through alternative RNA the spleen (SP) and lymph nodes (LNs), on which mIgD levels can by guest on October 2, 2021 splicing of primary IgH locus transcripts (8, 9) IgD is coexpressed exceed those of mIgM by 10-fold (24). IgD is expressed at lower with IgM as a membrane Ig (mIg) by most mature, quiescent B levels on B-1 cells, which are abundant in the peritoneal cavity cells in mammals (1, 2, 10) and most other vertebrate species (11– (PC), lamina propria, and pleural tissues (25), and on distinct B 14) [although not in rabbits and birds (15–17)]. Alternative RNA cells of the splenic marginal zone (MZ) (26). IgD is also expressed splicing ensures that on any given cell, the Ag specificity of IgM on many recently generated semimature transitional B cells found and IgD molecules is identical (18–21) and allows coexpression of in the SP and BM, most of which are short lived and express CD93 IgD and IgM at different relative abundances (22, 23). This ability (27–31). B cells initially transitioning from the BM to the SP are CD212CD232IgDlo/2 and are referred to as transitional B cell population 1 (T1) cells. Upon maturation to the T2 stage, all three *Department of Immunology and Microbial Science and †Doctoral Program in surface molecules are upregulated along with CD62L. T3 and Chemical and Biological Sciences, Kellogg School of Science and Technology, naive follicular (B-2) cells continue to retain high surface IgD The Scripps Research Institute, La Jolla, CA 92037 (sIgD) expression while downregulating their sIgM (31). IgD 1 Current address: Neostasis/Biogen Idec, San Diego, CA. expression is therefore an important marker for B cell de- 2 Current address: La Jolla Institute for Allergy and Immunology, La Jolla, CA. velopmental stage and maturation. However, IgM and IgD can Received for publication November 3, 2009. Accepted for publication February 9, also be functionally redundant. For example, in IgM2/2 mice and 2010. in IgD transgenic (Tg) mice, IgD can substitute for IgM in pro- This work was supported by National Institutes of Health Grant RO1AI059714 moting early B cell development (32, 33). and by Training Grant T32AI007606. K.W. was supported by Grant AI74564 to M. Oldstone. Immature B cells and B cells from fetal and neonatal mice express Address correspondence and reprint requests to Prof. David Nemazee, The Scripps exclusively IgM (1, 2, 34) and tend to be more susceptible to im- Research Institute, 10550 North Torrey Pines Road, La Jolla, CA 92037. E-mail mune tolerance than are more mature cells that coexpress IgD (35, address: [email protected] 36), correlations that prompted the hypothesis that IgD expression The online version of this article contains supplemental material. could protect B cells from immune tolerance (37, 38). However, Abbreviations used in this paper: ASC, Ab-secreting cell; BM, bone marrow; cIgM, mature IgD+ B cells encountering foreign or self-ligands can be cytoplasmic IgM; DTg, double transgenic; HEL, hen egg lysozyme; LN, lymph node; MHC II, MHC class II; MZ, marginal zone; mIg, membrane Ig; PB, Pacific Blue; PC, rendered tolerant by clonal elimination or functional inactivation peritoneal cavity; p-Tyr, phosphotyrosine; pUb, part of the ubiquitin promoter; sIg, (33, 39–42). Moreover, functionally hyporesponsive anergic B cells surface Ig; SP, spleen; T, transitional B cell population; TD, T cell-dependent; Tg, are characterized by high levels of mIgD combined with low levels transgenic; TI-2, T cell-independent type 2; TNP-Hy, trinitrophenol-hemocyanin. of mIgM and generally turn over rapidly when competing non- Copyright Ó 2010 by The American Association of Immunologists, Inc. 0022-1767/10/$16.00 tolerant B cells are present (43–45). On the other hand, IgD www.jimmunol.org/cgi/doi/10.4049/jimmunol.0903564 4144 TOLERANCE IN IgD SUPERANTIGEN MICE

expression is correlated with developmental maturation and the loss cag gac ctg gc-39); and AMS9 VH R(59-atc acc tcc cgg gtt tct ggg ggc tgt of the ability of B cells to undergo self-Ag–induced receptor ed- tgt ttc agc tga gga gac ggt gac cgc ggt ccc tgt gcc-39). iting (46, 47). However, transitional cells might retain the capacity IgDa-macroself mice to edit given the appropriate microenvironment (48, 49). Animal studies have been reviewed and approved by the The Scripps Like mIgM, mIgD can transmit signals through the associated Research Institute institutional review committee. All IgDa-macroself Tg signal transducers Ig-a/b (CD79a/b) (50–52), triggering a tyrosine mice described in this work had been bred and maintained on C57BL/6 kinase cascade and the activation of additional signaling path- background congenic for IgHa/a alleles (B6.Cg-Igha Thy1a Gpi1a/J, stock number 001317, JAX) for at least eight generations. In some experiments, ways. Early biochemical changes triggered by IgM and IgD in a/a mature B cells are biochemically similar, although signaling they were also bred to overexpress BAFF-R but retain IgH alleles. BAFF Tg mice have been described (66). To genotype for the presence of transgene through IgD may be more prolonged (53). However, upon acti- in each breeding, 35 cycles of PCR were performed using AMS9.1 VL CDR3 vation by Ag, IgD expression on B cells is lost (54–56). With rare S(59-cca aca atc tta taa tct att ca-39) and AMS9.1 VH CDR3 AS (59-cag aca exceptions, IgD is not expressed by plasma cells, although the tcg aag tac cag taa t-39) primers and the following conditions: melting (94˚C), secretory domain of IgD is intact and conserved in most species annealing (58˚C), and elongation (72˚C) for 40 s each. studied. IgD is secreted at relatively high levels in rare in- BM and SP cell transfers flammatory diseases, such as mevalonate kinase deficiency (57). In all BM transfer experiments, each mouse host was lethally irradiated at Rare cells or tumors with dedicated IgD secretion can be found 1000 Rad 1 d prior to receiving i.v. 107 BM cells, which were isolated from that have undergone a recombination event similar to classical both upper and lower leg bones of donors using standard protocols. Donor switch recombination (58). In humans, a subset of IgD class- mice were either IgHa/a C57BL/6 congenic mice (described above) or Bcl2 a/a switched cells with unique properties exists (59), and it has been Tg (67) bred onto IgH C57BL/6 congenic background. In splenic transfer experiments, total SP cells were isolated from either non-Tg or Downloaded from suggested that basophils carry an Fc receptor for IgD (60). IgDa-macroself mice by standard procedures. Subsequently, 3 3 107 total However, IgD-deficient mice have been generated and shown to SP cells were injected i.v. into nonirradiated IgHa/a congenic mice. Be- have only subtle defects, mainly in the kinetics of T cell- cause all donor and recipient mice had been crossed onto C57BL/6 dependent (TD) Ab responses (61, 62). background for at least eight generations, no depletion of donor T cells was We recently developed a custom superantigen technology based performed. I.v. injections were performed retro-orbitally. on the expression of membrane tethered single-chain Abs reactive Flow cytometry analyses to Ab constant regions (63–65). The goal of the current study was http://www.jimmunol.org/ Flow cytometric analyses for surface markers were performed using to assess the ability of an IgD-reactive superantigen expressed as standard protocols. Intracellular proteins were stained for by using Cytofix/ a ubiquitously expressed transgene in mice to promote B cell Cytoperm Kit (BD Biosciences, San Jose, CA) according to manufacturer’s tolerance and to alter B cell development and function. These recommendations. The following Abs were used at 1:200 in all experi- studies were carried out with a view to establishing more precisely ments described: B220 (RA3-6B2; allophycocyanin/Alexa 750; eBio- science, San Diego, CA); CD19 (1D3; PE/Cy7; eBioscience); CD21 (7E9; the cell lineage and precursor/product relationships among B cell FITC, allophycocyanin, or Pacific Blue [PB]; Biolegend, San Diego, CA); subsets found in the secondary lymphoid tissues and their modes CD22 (OX-97; Alexa 647; AbD Serotec, Raleigh, NC); CD43 (S7; PE; BD of tolerance induction. Biosciences); CD62L (MEL-14; PE; eBioscience); CD86 (GL1; PE; BD Biosciences); CD93 (AA4.1; PE/Cy7 or allophycocyanin; eBioscience); CD138 (281-2; PE; BD Biosciences); BAFF-R (7H22-E16; Alexa 647; by guest on October 2, 2021 Materials and Methods eBioscience); I-Ab (25-9-17; FITC; BD Biosciences); IgD (11-26; PE; a b l IgD -macroself construct eBioscience); IgM (AF6-78; FITC; Biolegend); and Ig 1,2,3 (R26-46; FITC; BD Bioscience). Abs against mouse IgM (M41; PB), IgMa (DS-1; The transgene encoding IgDa-macroself Ag was generated based on Alexa 647), and Igk (187.1; Alexa 647) were labeled in-house and used at a plasmid template described for Igk-macroself (63). Total RNA from either 0.5 mg (M41) or 0.1 mg (DS-1 and 187.1) per 2 3 106 cells (in 200 ml) AMS9.1 hybridoma was isolated using Trizol (Invitrogen, Carlsbad, CA) for both surface and intracellular stains. To stain for intracellular IgG, FITC- according to manufacturer’s instruction. VL and VH cDNA sequences were labeled mAbs against each isotype purchased from BD Biosciences (IgG1, obtained by 59 RACE (Ambion, Austin, TX) using Ck and IgG2b antisense A85-1; IgG2a, R19-15; IgG2b, R12-3; and IgG3, R40-82) were mixed and primers and cloned into Zero Blunt TOPO vector using methods suggested used at 0.15 mg each per 2 3 106 cells in 200 ml. To gate out most T and by the manufacturer (Invitrogen). Using the Igk-macroself construct as myeloid cells, most flow analyses included a dump Ab mixture consisting template, a cDNA fragment containing part of the ubiquitin promoter of PerCP/Cy5.5-labeled Abs against CD4 (RM4-5), CD8 (53-6.7), Gr-1 (pUb) and the 59 end of the VL was generated by PCR using pUb F and (RB6-8C5), and F4/80 (BM8) (all purchased from Biolegend), also used at iLAMS9 R primers. From the TOPO vector containing the VL cDNA, the 1:200 dilution each. All samples were read on an LSR-II instrument (BD VL fragment was reamplified using extended primers (AMS9VL F and Biosciences) and analyzed using the FlowJo program (Tree Star, Ashland, AMS9VL R) that contained the antisense sequence of the linker at its 39 OR). Mean fluorescence intensity values given were calculated by geo- end. Similarly, the VH fragment was reamplified using primers (AMS9VH metric means of the indicated fluorophore staining of the gated cells. F and AMS9VH R) that contained sequence for the linker at its 59 end and part of rat IgG1 Fc at its 39. PCRs for all three fragments were performed BrdU labeling analyses for 25 cycles using Accuprime Pfx (Invitrogen). Melting (94˚C), annealing (55˚C), and elongation (68˚C) times were 45 s each. Subsequently, a long All mice were initially injected i.p. with 1 mg/mouse of BrdU (Sigma- Aldrich, St. Louis, MO) in PBS and immediately placed on drinking water insert containing all three fragments was generated by nested PCR using containing 1 mg/ml BrdU for 7 d. Groups of three mice per time point were the pUb F and AMS9VH R primers. A total of 25 cycles were performed using the following conditions: melting (94˚C, 1 min), annealing (50˚C, 1 sacrificed on days 0, 1, 2, 4, and 7. Total cells harvested from their BM and min), and elongation (68˚C, 1 min). The insert was subsequently gel- SPs were subjected to intracellular BrdU staining (BD Biosciences) per purified, digested with SpeI and XmaI, and cloned into the previously manufacturer’s instruction. Surface staining was performed using the fol- described Igk-macroself construct (after removing the anti-Igk cDNA in- lowing Abs: CD93-PE (AA4.1, eBioscience); CD23-PE/Cy7 (B3B4, sert by SpeI and XmaI digestion). After linearization by NotI digestion, eBioscience); CD21-allophycocyanin (7E9, Biolegend); B220-allophyco- plasmid was sent to The Scripps Research Institute Transgenic Core (La cyanin/Alexa 750 (RA3-6B2, eBioscience); and PerCP/Cy5.5-labeled dump Jolla, CA) for pronuclei injection. Primer sequences used are as follows: Ab mixture (see above). Intracellular staining of IgM using PB-labeled clone Ck antisense (59-ctg ctc act gga tgg tgg gaa gat gg-39); IgG2b antisense M41 mAb was performed along with BrdU-FITC (BD Biosciences). (59-gga cag gga tcc aga gtt cca agt-39); pUb F (59-ttt tct ccg tcg cag gac gca In vitro stimulation for phospho-protein Western analyses ggg ttc ggg-39); iLAMS9 R (59-atc aca atg tcc tct gca cag gca cct gta ata att aat agg c-39); AMS9 VL F(59-tac agg tgc cTgc aga gga cat tgt gat gtc aca Total B cells were purified from pooled mouse SPs by CD4/CD43-negative gtc tcc ctc ctc cct ggc tgt gtc-39); AMS9 VL R(59-cct ccc gag cca ccg cct selection using the B Cell Isolation Kit (Miltenyi Biotec, Auburn, CA) per ccg ctg cct ccg cct ccc cgt tct att tcc aac ttt gtc ccc g-39); AMS9 VH F(59- the manufacturer’s instruction. For Western analyses of phospho-protein gca gcg gag gcg gtg gct cgg gag gcg gag gct cgc agg tgc agc tga agg agt levels, purified B cells were resuspended at 107 cells/ml in HBSS buffer The Journal of Immunology 4145 without supplement, prewarmed to 37˚C in water bath for ∼10 min, and (BD Biosciences) diluted at 1023 in ELISA buffer was used to report then stimulated for 10 min with 20 mg/ml of goat F(ab9)2 anti-mouse IgM signals using 1-Step Ultra-TMB Substrate (Pierce) per the manufacturer’s (Jackson ImmunoResearch Laboratories, West Grove, PA). Pelleted cells instructions. All wells were washed 43 with TBST in between steps. were lysed in (100 ml per 107 cells) TBS buffer containing 1% Nonidet Signals were recorded at 450 nm using VersaMax microplate reader P-40, Complete EDTA-Free Protease Inhibitor Cocktail (Roche Applied (Molecular Devices, Sunnyvale, CA). Science, Indianapolis, IN), 5 mM EDTA, 10 mM NaF, and 1 mM Na3VO4. After removal of insoluble cellular debris by centrifugation, 10 ml each TD and T cell-independent anti-TNP responses m lysate was run in 4–12% NuPAGE gels (Invitrogen). A total of 0.2 m IgDa-macroself and non-Tg littermates were bled on day 0 prior to receiv- nitrocellulose membranes (Bio-Rad, Hercules, CA) were used for all ing i.p. 20 mg TNP-Ficoll in PBS or 100 mg trinitrophenol-hemocyanin Western blotting analyses. The following primary Abs were used at the (TNP-Hy) in Imject Alum adjuvant (Pierce) for subsequent measurements indicated concentrations, diluted in TBST (TBS plus 0.1% Tween-20) of their T cell-independent and TD responses, respectively. Mice were bled containing either 5% milk or 1% BSA: anti-pTyr (4G10, 1:1000 in milk; at 7-d intervals. To assess memory responses, the latter group of mice re- Millipore, Bedford, MA), anti-GAPDH (6C5, 1:10,000 in milk; Millipore), ceived a secondary challenge of 100 mg TNP-Hy in PBS i.p. on day 44. Sera anti-pTyr531 CD19 (1:1000 in milk; Cell Signaling Technology, Beverly, were collected prior to secondary challenge and 7 d later on day 51. To MA), rabbit serum anti-total CD19 (1:1000 in milk; generous gift from Dr. measure anti-TNP responses by ELISA, Nunc Maxisorp plates (Nunc) were J. Cambier at National Jewish Health, Denver, CO), rabbit anti-pERK1/ coated with 10 mg/ml of TNP -BSA in PBS overnight, blocked with ELISA 2 mAb (197G2, 1:1000 in BSA; Cell Signaling Technology), and rabbit 20 buffer for 1 h, and incubated with sera serially diluted 23 in ELISA buffer anti-total ERK1/2 pAb (1:1000 in BSA; Cell Signaling Technology). All for 1 to 2 h. All subsequent detection and reporting steps were performed as HRP-conjugated secondary polyclonal Abs were purchased from Jackson described above for isotype ELISAs. Response titer at each time point per ImmunoResearch Laboratories and used at 1:30,000 diluted in TBST plus mouse was defined as the theoretical factor of serum dilution required to 5% milk. Signals were developed using SuperSignal West Pico Chemilu- achieve half-maximum OD of the assay (normally at values of 1.6–1.9), minescent Substrate (Pierce, Rockford, IL). To strip blots for reprobing, extrapolated from four-parameter plots. membranes were incubated at 65˚C with gentle agitation for 15–20 min in TBS buffer (pH 2.3). Downloaded from In vitro proliferation assays Results Generation of IgDa-macroself Tg mice Two-fold serial dilutions of goat F(ab9)2 anti-mouse IgM (Jackson Im- munoResearch Laboratories) or LPS (055:B5, Sigma-Aldrich) in media To assess the in vivo effects of the expression of an IgD-reactive (DMEM containing 10% FBS, 55 mM 2-ME, L-glutamine, and penicillin- superantigen, a membrane tethered single-chain Ab gene was streptomycin) were prepared in 96-well plates (100 ml/well). B cells, pu- engineered from the variable H and L regions of anti-IgDa hy- rified from pooled mouse SPs using a B Cell Isolation Kit (Miltenyi bridoma AMS9.1 (68) using previously described methods (63) http://www.jimmunol.org/ Biotec), were resuspended at 2 3 106 cells/ml in media and subsequently added into each well (100 ml/well). After ∼24 h of stimulation, 1 mCi of (Supplemental Fig. 1). Tg mice that ubiquitously expressed this a [3H]thymidine (GE Healthcare, Piscataway, NJ) was added per well. Cells construct were generated (and are called IgD -macroself mice were allowed to label for ∼15 h prior to harvest for scintillation counts. All hereafter). Three lines with high-, intermediate-, or low-level ex- treatments were done in triplicate. pression were selected for further study. As the phenotypes of Ca2+ flux assays these mice were virtually identical, most data shown in this study

a involve studies carried out with mice carrying an intermediate Total SP cells isolated from IgD -macroself mice and their non-Tg litter- level of macroself expression (line #16). All IgDa-macroself mice mates were pooled and resuspended at 1.6 3 107 cells/ml in HBSS 2+ 2+ 6 (containing Ca and Mg ). A total of 8 3 10 SP cells were then pre- subsequently analyzed had been bred onto the C57BL/6 back- by guest on October 2, 2021 a incubated with 500 ml Fluo-4 NW reagent (Invitrogen), 5 mg Fc Block ground congenic for the IgH allele for at least eight generations. m (2.4G2; BD Biosciences), and 1 g each PerCP-Cy5.5–labeled CD4 and a CD8 (RM4-5 and 53-6.7; Biolegend) for 30 min at 37˚C, followed by Altered mature B cell subsets in IgD -macroself mice 6 another 30 min at room temperature. Aliquots of 4 3 10 cells (0.5 ml) Flow cytometry analysis was used to assess the effect of super- were then stimulated at room temperature with 10 or 40 mg/ml goat F(ab9)2 2+ antigen expression on the numbers of B cells in different tissues. anti-mouse IgM (Jackson ImmunoResearch Laboratories). Ca signals a were recorded for 200 s, gating on CD4/CD8 double-negative cells. Lethally irradiated IgD -macroself mice were reconstituted with IgHa BM, and their B cell development was compared with that of Quantitative RT-PCR the similarly treated non-Tg littermates. Initial analysis was car- B cells were purified from pooled SPs or BM using B Cell Isolation Kit ried out with CD19/B220 costaining, which identifies virtually all (Miltenyi Biotec), and their total RNA was isolated using Trizol Reagent B cells and can reveal distinct subsets (Fig. 1A, Supplemental (Invitrogen) per kit instructions. After DNase-treatment (Ambion), 4 mg Table I). As expected, no significant changes were seen in im- total RNA was used to synthesize first-strand cDNA from random hexamers using Transcriptor First Strand cDNA Synthesis Kit (Roche Applied Sci- mature BM B cells. By contrast, recirculating mature B-2 cells in ence) according to the manufacturer’s recommendations. PCRs were per- the BM were virtually absent. This was similarly reflected by the formed using SYBR GreenER qPCR Supermix (Invitrogen) and analyzed ∼96% reduction in total B cell numbers in the LNs, where the on the 7900HT Fast Real-Time PCR System (Applied Biosystems, Foster most developmentally mature recirculating B-2 cells normally City, CA). Annealing temperature per cycle was set to 60˚C. Quantifica- reside. In the SP, B cell percentages were reduced by .90%. A tions of RAG1 and RAG2 were normalized with concomitant amplifica- tions of b-actin cDNA in each triplicate sample. The following primers more detailed analysis for the two major mature B cell subsets in + were used: RAG1 F (59-gga ggc ctg tgg agc aag gta g-39); RAG1 R (59-gtt this compartment revealed that B-2 cells bearing the CD23 ctg acc acc agg ctt ctc t-39); RAG2 F (59-ccc agt gca tgg att tgg aag aac-39); CD932B220+ surface phenotype were severely reduced in fre- RAG2 R (59-gga gtt tgc aat gct ctt gct atc tg-39); b-actin F (59-caa ccgt gaa quency and number (Fig. 1B,1E). However, CD21hiCD1dhiB220+ aag atg acc cag a-39); and b-actin R (59-cac agc ctg gat ggc tac gta c-39). 2 MZ B cells of CD93 mature phenotype were clearly present, Serum Ig ELISAs albeit at only ,15% of normal numbers (Fig. 1C,1E, Supple- m mental Table I). In the peritoneum, total B cell numbers were also Nunc Maxisorp plates (Nunc, Roskilde, Denmark) were coated with 2 g/ hi lo ml rat IgG-adsorbed donkey anti-mouse IgG (H+L) (Jackson Immuno- reduced although a significant number of CD19 B220 B cells Research Laboratories) overnight at room temperature, blocked with 1% still remained. These cells were confirmed to be B-1 by CD43 BSA in TBST (ELISA buffer) for 1 h, and incubated for 1 h with mouse marker expression (Fig. 1D), suggesting that B-1 cells could still sera diluted in ELISA buffer. The following biotinylated anti-mouse Abs be reconstituted in IgDa-macroself mice by adult BM. Surface were used at 0.25 mg/ml in ELISA buffer for detection: IgM (M41), IgG1 + (A85-1; BD Biosciences), IgG2a (RMG2a-62; Biolegend), IgG2b (clone IgM B cells were clearly present in all lymphoid organs tested R12-3; BD Biosciences), IgG3 (R40-82; BD Biosciences), IgA (C10-1; (Supplemental Fig. 2A). However, surface IgD expression (mea- BD Biosciences), and IgE (R35-72; BD Biosciences). Streptavidin-HRP sured using an Ab that is not affected by AMS9.1 binding) was 4146 TOLERANCE IN IgD SUPERANTIGEN MICE

FIGURE 1. Impairment of peripheral B cell de- velopment in IgDa-macroself mice. Non-Tg and IgDa- macroself mice were lethally irradiated, reconstituted with 107 IgHa BM cells, and analyzed 10 wk later for B cell development in the BM, SP, LNs, and PC. A, Representative CD19 and B220 expression profiles of lymphocytes within each compartment are shown. B-1 and B-2 cells in PC are represented by the CD19hi B220lo and CD19+B220hi populations, respectively. Reconstituted recipients were evaluated for presence of splenic CD23+CD932 B-2 cells (B), CD21hCD1dhi MZ B cells (C), and peritoneal CD43+CD19+ B-1 cells (D). Downloaded from Plots shown were gated on non-T and nonmyeloid cells. Maturity of MZ B cells was also assessed by CD93 expression, shown in lower panels of C. E, Total cell numbers for each immune tissue and the three mature B cell compartments were calculated. Shown are mean + SD analyzed from six non-Tg and six IgDa-

macroself age-matched hosts. Values of p by two-tailed http://www.jimmunol.org/ Student t tests are also shown where there is statistical significance. by guest on October 2, 2021

undetectable on these cells (Supplemental Fig. 2B). These alter- CD62L, IgD, and CD19 (Fig. 2E). Importantly, most CD23+B220+ ations in B cell development were reproducible in nonirradiated B cells in the SP of IgDa-macroself mice were CD93+, suggesting IgDa-macroself mice bred onto a C57BL/6 background homozy- that B cell development did not progress beyond this point (Fig. gously congenic for the IgHa allele, except that normal numbers of 2C). Thus, immature B cells appeared to develop normally in the B-1 cells were present in the peritoneum (Supplemental Fig. 3 and BM of IgDa-macroself mice but were abruptly halted in the SP data not shown). Thus, presence of the IgDa-macroself Ag im- shortly after the T1 stage. Except for CD23, the few B cells that paired development of all three mature B cell subsets, although developed beyond the T1 stage failed to upregulate any other B-1 cell numbers could reach the normal range in intact IgHa mice markers normally associated with conventional T2 cells defined by carrying the macroself Ag. Allman et al. (31). Developmental block at transitional B cell stages Increased turnover of CD23+CD93+ transitional cells To determine where developmental block occurred in IgDa- To test the hypothesis that the transitional B cells in IgDa-macroself macroself mice, B cells were evaluated for expression of matu- mice were eliminated coincident with their acquisition of CD23 ration markers. Development of mature B cells in the periphery is expression, we measured their turnover. IgDa-macroself and non- thought to occur via two cellular pathways, one in the BM and Tg mice carrying IgHa alleles were continuously labeled with a second in the SP (69, 70). Recently formed CD23+CD93+ T2- BrdU, and uptake in B cells at the different developmental stages like B cells in the BM of IgDa-macroself mice did not appear was assessed. As predicted, the splenic CD23+CD93+ transitional significantly different from those of non-Tg littermates, either B cells of Tg mice exhibited a more rapid rate of turnover than quantitatively or phenotypically (Fig. 2A,2D). Similarly, the those of non-Tg littermates (Fig. 3A, third panel from left). number of CD212CD232CD93+ T1 cells in the SP appeared However, no significant differences were seen at the less mature normal (Fig. 2B,2D). However, in contrast to the T2-like cells in splenic CD232CD93+ T1 or BM CD23+CD93+ T2-like stages the BM, the numbers of CD23+CD93+ transitional B cells in the (Fig. 3A, first two panels on left). Interestingly, MZ B cells in IgDa- SP were severely reduced (Fig. 2C,2D). Unlike their T2 and T3 macroself mice turned over more slowly than normal, suggesting counterparts in non-Tg mice (defined as CD23+CD93+IgMhi and either a slower rate of homeostatic proliferation or reduced re- CD23+CD93+IgMlo, respectively), these CD23+CD93+ cells in plenishment of this compartment by recently formed B cells. IgDa-macroself mice failed to upregulate markers normally as- Levels of sIgM appeared to be actively downregulated on some sociated with maturation from the T1 stage, such as CD21, CD22, of the CD232CD93+ T1 cells and remained downregulated upon The Journal of Immunology 4147 Downloaded from http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 2. Developmental block of B cell maturation at CD23+CD93+ transitional stage in IgDa-macroself mice. Lethally irradiated IgDa-macroself and non-Tg littermates were reconstituted with syngeneic IgHa BM cells. Two months later, numbers and phenotypes were analyzed for CD23+CD93+ T2-like cells in BM (A), splenic CD212CD232CD93+ T1 cells (B), and splenic CD23+CD93+ transitional cells (C). Also shown are their surface IgM expression profiles (lower panels). D, Cell counts for each transitional B cell subset were compared between the two groups of mice. Shown are mean + SD analyzed from six mice per group; p values by two-tailed Student t tests. E, Surface marker phenotypes of the splenic transitional B cells in IgDa-macroself mice were compared against their counterparts in the non-Tg littermates. All analyses were performed after gating out CD4+CD8+F4/80+Gr1+ cells in the dump channel. However, in the lower panels of A–C, the dump channel cells were used to establish sIgM background (gray histograms).

CD23 expression (Fig. 2B,2C, lower panels). These IgMlo cells on the cell surface, whereas MZ cells remained IgD-negative expressed highly elevated levels of I-A (MHC class II [MHC II]), (Supplemental Fig. 4A,4B). Thus, IgD cross-linking caused a prominent feature of activated B cells (Fig. 3B,3C). When transitional B cells to downregulate IgM, but also to upregulate isolated from SPs of BM-reconstituted IgDa-macroself mice and I-A, starting at the CD212CD232 T1 stage. Subsequently, most B cultured for ∼17 h in medium alone, virtually all B cells expressed cells were eliminated prior to transitioning to the conventionally elevated levels of I-A and CD86 (Supplemental Fig. 4C), although defined T2 stage, as reflected by their increased BrdU in- still retaining a CD212CD232 immature phenotype (Supple- corporation, except for the few MZ B cells that did not express mental Fig. 4B). Interestingly, if relieved from IgD ligation by IgD. Notably, the levels of I-A on the CD23+CD93+ transitional B removal of IgDa-macroself Ag, they also began to accumulate IgD cells from IgDa-macroself mice were significantly higher than 4148 TOLERANCE IN IgD SUPERANTIGEN MICE

FIGURE 3. Increased turnover rate and activation phenotype of transitional B cells in IgDa-macroself mice. A, IgHa congenic C57BL/6 mice carrying or lacking the IgDa-macroself transgene were labeled with BrdU over 7 d. Cells from BM and SPs were harvested 0, 1, 2, 4, or 7 d later for analysis of BrdU incorporation by B cells. Gates used to define each transitional B cell subset were the same as those de- picted in Fig. 2A–C. pp , 0.05; pppp , 0.005. B, Levels of surface MHC II expression were compared between transitional B cells that expressed high and low sIgM illustrated in Fig. 2B,2C. Note the sIgMlo CD23+CD93+ T3 population in non-Tg mice (middle right panel, shaded histogram), unlike the sIgMlo CD23+CD93+ cells in IgDa-macroself mice (solid line), did not express significantly increased levels of Downloaded from MHC II when compared with non-Tg CD932CD23+ B-2 cells (dashed line). C, Mean fluorescence in- tensity (MFI) of I-Ab staining for each B cell subset is shown, error bars representing SD of six IgDa- macroself and four non-Tg littermates. Values of p by two-tailed Student t tests are shown for each set

of comparisons where there is statistical significance. http://www.jimmunol.org/

those on the IgMloCD23+CD93+ anergic T3 population of non-Tg their proliferative responses to all concentrations of anti-IgM

mice as defined by Merrell et al. (71) (Fig. 3B, middle right tested were severely reduced (Fig. 4C, left panel). B cells in some by guest on October 2, 2021 panel). By contrast, I-A expression was not significantly elevated models of anergy have also been reported to exhibit reduced re- on MZ B cells of Tg mice and only slightly elevated on B-1 cells, sponsiveness to LPS (75). However, this was not the case with B suggesting that the majority of these cells were not activated. cells from IgDa-macroself mice either, as their responses to LPS Thus, the IgD+ transitional population appeared to mainly die were comparable to those of wild-type B cells (Fig. 4C, right rapidly subsequent to abortive activation, whereas remaining MZ panel). Although already expressing higher levels of I-A and and B-1 populations appeared stable and more quiescent. CD86 than non-Tg B cells when cultured in medium alone, B cells isolated from IgDa-macroself mice failed to upregulate CD86 to Absence of conventional anergy phenotype the extent of non-Tg B cells upon IgM cross-linking, but did fully a To determine if B cells from IgD -macroself mice exhibited fea- upregulate I-A (Supplemental Fig. 4C). Altogether, these ob- tures of anergy (reviewed in Ref. 72), they were stimulated in vitro servations suggest that B cells from IgDa-macroself mice dis- and responses assessed. Upon BCR ligation, phosphotyrosine played a phenotype resembling immature non-Tg B cells rather a (p-Tyr) was similar in non-Tg and IgD -macroself B cells, except than tolerized anergic cells, with the exception that they expressed for some changes in species around ∼95 to ∼100 kDa (Fig. 4A). higher levels of the activation markers CD86 and I-A. One of these bands proved to be CD19, which was poorly phos- phorylated at Tyr531 in IgDa-macroself B cells; however, total Altered L chain repertoire, but lack of peripheral receptor CD19 levels of these cells were also reduced. Because most B editing cells in IgDa-macroself mice appeared to be CD212CD232 T1 To determine if receptor editing actively occurred in peripheral B cells, we wished to assess if the phosphorylation differences were cells of IgDa-macroself mice, we assessed L chain isotype usage due to the immaturity of the cells or to changes induced by the and RAG expression. Increased l-chain usage often correlates macroself Ag. Because BAFF2/2 mice also have mostly B cells with receptor editing (76). In BM, l versus k usages were not that resemble CD212CD232 T1 cells (73, 74), their BCR-induced significantly different between the IgM+ immature B cell pop- p-Tyr profile was compared with that of wild-type B cells. In- ulations of non-Tg and IgDa-macroself mice (Fig. 5A,5E). l L terestingly, stimulated B cells from BAFF2/2 mice exhibited chain representation, however, was significantly increased in total a profile that was remarkably similar to those from IgDa-macroself splenic B cells of IgDa-macroself mice (Fig. 5B,5E). Mature mice, including reduced CD19 Tyr531 phosphorylation and total peritoneal B-1 and splenic MZ B cells also had comparably in- CD19 expression, suggesting that they are common characteristics creased usage of l-chain (Fig. 5C–E). If receptor editing was of immature B cells (Fig. 4A). responsible for increased l usage, we predicted that increases Another hallmark of B cell anergy is a reduced Ca2+ response to in RAG1 and RAG2 expression would be apparent in splenic BCR cross-linking, but IgDa-macroself B cells appeared to induce B cells. Quantitative RT-PCR analyses were performed on splenic a normal Ca2+ flux upon BCR cross-linking (Fig. 4B). By contrast, B cells purified by CD4/CD43-negative selection, revealing that The Journal of Immunology 4149

FIGURE 4. Absence of conventional B cell anergy in IgDa-macroself mice. A, B cells were purified from SPs of IgDa-macroself and non-Tg littermates by CD4/CD43-depletion, left unstimulated or stimulated with goat F(ab9)2 anti-mouse IgM for 10 min, and analyzed by Western blotting for levels of p-Tyr, p- CD19 (on Tyr531), and p-ERK as described in Ma- terials and Methods. To compare with B cells of Downloaded from mostly CD212CD232CD93+ T1 phenotype, B cells from wild-type and BAFF2/2 mice were similarly purified, stimulated, and analyzed. B, Splenic cells from IgDa-macroself and non-Tg littermates were subjected to stimulation by 10 or 40 mg/ml goat 2+ F(ab9)2 anti-mouse IgM and monitored for Ca flux. http://www.jimmunol.org/ C, In vitro B-cell proliferative responses to increasing concentrations of LPS and goat F(ab9)2 anti-mouse IgM polyclonal Abs were measured by tritiated thymidine incorporation. Cells were pulsed for ∼16 h after ∼24 h of stimulation in culture. by guest on October 2, 2021

IgDa-macroself B cells had levels of RAG message ,0.4% of consistent with a recent report indicating that CD93 is functionally wild-type BM B cells (Fig. 5F). Although this level was slightly important for plasma cell maintenance (77). The frequency of IgM+ higher than that found in non-Tg B cells, it likely reflected ASCs increased with age, whereas numbers of IgG+ ASCs were low a higher proportion of immature cells in the population because but not reliably different between non-Tg and IgDa-macroself mice this value was less than that found in non-Tg BAFF2/2 splenic B at all ages tested (not shown). The IgM bias may reflect the effector cells. Thus, induction of RAG-mediated L chain editing in SPs of properties of MZ and B-1 populations (see Discussion). IgDa-macroself mice was unlikely to be responsible for the altered Altered serum Ab levels and humoral immunity l/k ratios observed. a + To assess the in vivo functionality of B cells in IgD -macroself Increased numbers of splenic IgM Ab-secreting cells mice, we first measured their serum Ig levels at 2 mo of age. IgM Considering that IgDa-macroself mice had a severely reduced MZ B levels were significantly elevated in IgDa-macroself mice (Fig. cell compartment and virtually nonexistent B-2 cells, we next de- 6A), mirroring the observed increases of splenic cIgM+ ASCs in termined whether Ab-secreting cells (ASCs, defined as possessing these mice. However, the levels of IgG1 and IgG2a were reduced, high levels of intracellular Ig) existed at steady state. Surprisingly, and IgG2b showed a slight (although statistically nonsignificant) more intracellular IgM+ (cytoplasmic IgM [cIgM]+) ASCs were trend toward reduction. The levels of IgG3 and IgA were similar found in SPs of IgDa-macroself mice (Supplemental Fig. 5A,5C). between the two groups of mice, but IgE was not reliably de- These cells expressed high levels of surface IgM (in addition to tectable in either group. their high cIgM levels), high levels of CD138, intermediate levels Next, we compared IgDa-macroself to non-Tg littermates for of CD19, low levels of B220, and, interestingly, extremely high Ab responses to the T cell-independent type 2 (TI-2) Ag TNP- levels of CD93 (Supplemental Fig. 5B). This last observation is Ficoll. Measurement of anti-TNP titers of the two isotypes mainly 4150 TOLERANCE IN IgD SUPERANTIGEN MICE

FIGURE 5. Altered ratios of l/k L chain representation by B cells in IgDa- macroself mice, but absence of signifi- cant RAG1/2 upregulation. Intracellular stains of k and l Ig L chains are shown for BM intracellular IgM+ immature B cells (A); total splenic (SP) lymphocytes negative for CD4, CD8, Gr1, and F4/80 markers (B); peritoneal CD19+CD43+ B-1 cells (C); and MZ B cells (D). E, l/k Ratios were calculated for each B cell subset shown in A–C (mean + SD; n = 4 per group; p values by two-tailed Student t tests). F, Quantitative RT-PCR analyses of RAG1 and 2 expression in splenic B cells isolated from non-Tg, a 2/2

IgD -macroself, and BAFF mice Downloaded from (triplicate measurements from B cell pools of SPs from the following number of mice/group; n = 6, 6, 4, respectively). Signals were compared with those ob- tained from BM B cells isolated from non-Tg mice, which were set arbitrarily

to 1000. http://www.jimmunol.org/

responsible for TI-2 responses, IgM and IgG3, revealed, surpris- macroself Ag. Isolated SP cells from IgHb Ly5a IgDa-macroself ingly, that IgDa-macroself mice mounted not only robust IgM mice were introduced into nonirradiated IgHa Ly5b mice pos- a

responses that stayed stronger than those in non-Tg mice for least sessing normal B cell compartments of IgH allotype (Fig. 7A). by guest on October 2, 2021 3 wk, but also comparable IgG3 responses (Fig. 6B). Phenotypes of recipient B cells were assessed 7 d later. The To assess TD responses, mice were primed with TNP-Hy in alum lymphoid organs of these mice recapitulated the phenotype of and boosted 44 d later with TNP-Hy in buffered saline. Re- IgHa IgDa-macroself mice, including severe reduction of follicular markably, the primary responses of IgDa-macroself mice were also type B cells in all tissues tested (Fig. 7B, Supplemental Table II), higher for IgM but subnormal for IgG1, IgG2a, and IgG2b ∼80% reduction of MZ B cells in the SP (Fig. 7D), but only (Fig. 6C), resembling their resting serum Ig levels. After sec- a slight and nonstatistically significant reduction of peritoneal B-1 ondary challenge, the IgM response remained higher in IgDa- cells (Fig. 7B). Surface IgD was also downregulated in all com- macroself mice, whereas IgG1, IgG2b, and IgA responses were all partments tested (Fig. 7C), and most B cells in the SP were CD93+ significantly lower than those of non-Tg mice. Secondary IgG2a (Fig. 7E). These results thus argue that tolerance mediated by the and IgE responses were comparable in the two groups. Regardless IgD superantigen can occur in fully mature follicular B cells. of isotype, IgDa-macroself mice were clearly capable of mounting Reduced competitiveness of IgHa+ B cells among innocuous a memory response to the secondary challenge that was signifi- IgHb+cells cantly more robust than their primary response. These findings are consistent with our preliminary studies in which IgDa-macroself As anergic B cells have poor survival in the presence of non-anergic mice showed no ill effects to a primary challenge with a sublethal B cells (78, 79), we assessed the consequences of B cell compe- a a/b dose of influenza or a secondary challenge with a lethal dose of tition in the IgD -macroself model by characterizing IgH mice, a the same virus (K. Walsh, unpublished observations). Strikingly, in which only half of the B cells are predicted to express IgD .To primary and secondary anti-TNP responses in IgDa-macroself monitor B cells independently expressing each IgH allele, we mice were dominated by Abs carrying l L chains, whereas k+ Abs measured the frequencies of IgM cells carrying the a and b allo- a a+ predominated in all responses of control mice (Fig. 6C, lower types. In IgD -macroself mice, IgM B cells were greatly un- panels). These experiments illustrate that despite skewed usages derrepresented in all compartments except BM (Fig. 8). LNs of a a+ of L chains and H chain isotypes and reduced primary response IgD -macroself mice were virtually devoid of IgM cells, but a+ among certain isotypes to TD Ag, IgDa-macroself mice appeared significant subpopulations of IgM B cells were still present in a+ capable of mounting normal (if not exaggerated) TI-2 responses the SP and PC. Although the cell surface markers of the IgM a/b a and relatively unimpaired memory responses to TD Ags. peritoneal B cells in IgH IgD -macroself mice were consistent with a mature B-1 phenotype, splenic IgMa+ cells were mostly Induction of tolerance in mature B cells by IgD superantigen CD212CD232 T1 cells. IgMa+ MZ and plasma B cells were Because B cell tolerance to IgDa-macroself Ag could simply in- virtually nonexistent compared with homozygous IgHa/a IgD- volve developmental arrest of immature B cells, we next de- macroself mice (Fig. 8B,8C). Quantitation of IgMa/IgMb ratios termined whether fully mature cells could also be tolerized by the revealed, as expected from the BrdU-labeling analyses, that the The Journal of Immunology 4151

FIGURE 6. Altered serum Ig levels and Ab responses to T cell-independent and TD Ags in IgDa-macroself mice. A, Serum Ig levels in 2-mo-old naive mice were measured by ELISA. B,IgDa -macroself and non-Tg littermates were immunized i.p. with 20 mg/mouse of TNP-Ficoll. Anti-TNP IgM and IgG3 responses were measured by ELISA every7dfor3wk.C, Mice were primed on day 0 i.p. with 100 mg/ mouse TNP-Hy in alum and boosted on day44with100mg/mouse TNP-Hy in PBS also administered i.p. Left and center panels show anti-TNP titers of the indicated isotypes determined by ELISA. Right panels show only IgA and IgE titers measured on day 7 post secondary immunization because pri-

mary responses of these two Ig classes Downloaded from were not reliably detectable. In all ca- ses, wild-type values are shown with black fill and IgDa-macroself values by unfilled symbols. All values represent mean + SD obtained from six to eight mice per group. Asterisks indicate sta- tistical significance by two-tailed Stu- http://www.jimmunol.org/ dent t tests: pp , 0.05; ppp , 0.01. Absence of asterisks indicates no sta- tistical significance.

most dramatic loss of IgMa+ B cells occurred between the CD232 terestingly, B cells in IgDa-macroself mice retained a slightly to CD23+ transition, although there was already an ∼50% loss of higher l/k ratio than those in non-Tg mice, but this trend started in a+ 2 2

IgM B cells at the CD21 CD23 T1 stage due to the presence the immature stage of the BM. by guest on October 2, 2021 of competing cells (Fig. 8D). Accumulation of MZ B cells and plasma cells induced by BAFF Rescue of putative preplasma but not mature B cells by Bcl2 Tg overexpression To investigate whether deletion in the IgDa-macroself model in- We generated double Tg mice expressing both IgDa-macroself and volved a Bcl-2–inhibitable pathway, we transferred BM from IgHa BAFF transgenes to assess the effects of BAFF overexpression on B mice overexpressing Bcl-2 in the B lineage (Bcl2 Tg) into lethally cell tolerance in the IgDa-macroself model. BAFF is an important irradiated wild-type or IgDa-macroself mice and investigated their homeostatic cytokine for follicular, MZ, and plasma B cells that is B cell phenotype. B cells were abundant in IgDa-macroself re- known to regulate the B cell competition-dependent survival dis- cipients, but their numbers were still reduced by at least ∼70% in advantage of anergic self-reactive B cells (78, 79). However, its the SP and ∼85% in the LNs (Fig. 9A, Supplemental Table I). overexpression in the IgDa-macroself mice failed to rescue B-2 Bcl-2 overexpression only caused an accumulation of CD212 cells (Fig. 10A). The LNs of IgDa-macroself/BAFF double Tgs CD232CD93+ B cells in the SP, but did not significantly rescue were nearly devoid of CD23+CD932 B-2 cells. However, with B-2, MZ, or even CD23+CD93+ B cells (Fig. 9B). Interestingly, BAFF overexpression, the numbers of MZ B cells in IgDa-macro- the accumulated CD212CD232 cells expressed not only very high self mice were increased by nearly 100% (Fig. 10B), which rep- levels of CD93 but also CD138 (Fig. 9C). That these cells did not resented ∼50% of what was found in wild-type mice and ,10% of express high levels of intracellular IgM suggested that they were BAFF single Tgs. In both the SP and LNs, excess BAFF promoted preplasma cells. However, in contrast to IgDa-macroself mice not increased numbers of IgM+ plasma cells (Fig. 10B,10C), but not overexpressing Bcl-2, which had higher numbers of IgM+ ASCs in IgG+ ASCs or memory cells (Supplemental Fig. 6A). Surprisingly, the SP, enforced Bcl-2 expression did not further increase accu- the numbers of CD232 and CD23+ transitional B cells were not mulation of these cells. In fact, the numbers of IgM+ ASCs were significantly different among IgDa-macroself mice, regardless of now comparable between mice that lacked or carried the macro- BAFF expression, possibly indicating that they were eliminated self Ag (Fig. 9D). Overexpression of Bcl-2 also failed to rescue prior to acquisition of responsiveness to BAFF. Analysis of BAFF- splenic and LN IgG+CD19+ B cells—by definition, memory B R(TNFRSF13C) expression in IgDa-macroself mice revealed that cells (Fig. 9D). In models of central B cell tolerance, Bcl-2 MZ and both CD232 and CD23+ populations of transitional B cells overexpression improves the efficiency of receptor editing by expressed lower levels than their non-Tg counterparts (Supple- extending the window of time for secondary gene rearrangements mental Fig. 6B). These results indicate that, unlike models of (80). As expected from our failure to detect induction of RAG anergy in which excess BAFF could promote survival of autor- expression in SPs of IgDa-macroself mice, overexpression of Bcl- eactive B cells, BAFF overexpression in the context of IgDa- 2 did not alter the l/k ratio from the newly formed B cell stage in macroself Ag merely supports elevated steady-state levels of MZ BM to the more advanced stages in the periphery (Fig. 9E). In- and IgM+ plasma B cells. 4152 TOLERANCE IN IgD SUPERANTIGEN MICE

FIGURE 7. Ability of SP cells carrying IgDa- macroself Ag to induce B cell elimination upon transfer to adult IgHa congenic mice. A, Unirradiated Ly5b IgHa recipient mice were given 3 3 107 SP cells from IgDa-macroself mice (Ly5a IgHb) 7 d preanalysis. B, Shown are representative plots depicting total B cell a loss in SP, LN, and PC of IgD -macroself mice by Downloaded from CD19 and B220 flow cytometry analysis. Effects on numbers of host IgD+ B cells (C), MZ B cells (D), and levels of CD93 expression (E) were evaluated by flow cytometry. Plots shown in B, C, and E were gated on lymphocytes, excluding IgMb-positive donor cells. Additionally, T and myeloid cells of both donor and recipient origins were excluded in the analyses for MZ http://www.jimmunol.org/ B cells shown in D. by guest on October 2, 2021

Discussion cytoplasmic m (90). Similar studies were carried out using anti- IgDa-macroself mice were generated to determine if and how IgD treatment in experimental animals (39, 91–104). These polyclonal B cells can undergo tolerance at a late, post-BM de- studies differed in many details, but suggested that anti-IgD velopmental stage. Because IgD itself is a very useful marker for treatment could only partly modulate Ab responses, depending B cell maturity, the effect of IgDa-macroself Ag should by defi- upon the context of treatment, the availability of competing, Ab nition be restricted to a B cell that is developmentally advanced. nonreactive B cells, and the type of immunogenic challenge. IgD is absent from pre-B cells and is poorly expressed by fraction Moreover, these studies were performed before good markers of E immature B cells in the BM and splenic T1 cells (30, 31, 81). B cell subsets were available and before the process of receptor This delayed expression of IgD relative to IgM in turn allowed us editing was well defined. to use IgD superantigen to further delineate B cell developmental Data obtained in the current study support previous tolerance stages and to test the prediction that self-Ags encountered at studies using conventional Ig Tg mice, suggesting that IgD can a post-BM developmental stage promote deletion or anergy, but transmit tolerogenic signals (33) and that IgD+ cells are largely not editing. unable to undergo receptor editing (46, 47, 105). Infusion of Studies involving treatment of animals with Abs to Ig constant superantigen-carrying cells into intact mice led to deletion of regions have been instrumental in elucidating precursor/product follicular B cells and most MZ B cells, whereas B-1 cell num- relationships among naive, immature, and memory B lymphocyte bers were only slightly affected, suggesting that cells carrying subsets. In the developing chicken, anti-IgM H chain (m-chain) a high level of IgD were eliminated by apoptosis. Remaining suppression revealed that expression of mIgM preceded in on- splenic B cells had an immature phenotype, including CD93 togeny the expression of other isotypes because both IgM and IgG expression, or were MZ B cells lacking IgD. These phenotypes isotypes were suppressed (82, 83). Similar studies in mice sup- were similar to those of radiation BM chimeras in which IgDa ported these results (84–89) and facilitated the discovery of pre-B -superantigen-carrying mice were reconstituted by wild-type cells as a population resistant to anti-m treatment that expressed IgHa BM,exceptthatB-1cellswerelessnumerousinBM The Journal of Immunology 4153

FIGURE 8. Development of IgHa B cells among competing IgHb cells in IgDa-macroself mice. A, IgHa/b mice lacking or carrying IgDa- macroself transgene were assessed for contri- bution of each IgH allele in B cell development by allotypic Abs. Non-Tg and IgDa-macroself mice possessing either IgHa/a or IgHa/b alleles were evaluated by cell surface staining for the presence of IgMa+ MZ B cells (B) and ASCs

(C). Boxes in B and C indicate MZ and plasma Downloaded from cell populations, respectively. Gated populations were indicated above plots. D, Ratios of IgMa- to IgMb-positive B cells were assessed in splenic transitional and MZ B cell subsets as well as peritoneal B-1 cells. Shown are mean + SD from four mice per group; statistical analyses by two- tailed Student t tests. pppp , 0.005. http://www.jimmunol.org/ by guest on October 2, 2021

chimeras. Thus, IgD superantigen can delete preformed follic- more mature population that includes a BAFF-responsive subset, ular and a major subset of MZ B cells, but likely also eliminates which is missing from IgDa-macroself mice. their precursors. We also observed B cell populations in IgDa-macroself mice that Several lines of evidence from the current study indicate that B were reminiscent of T3 and T39 transitional stages (31, 71, 108), cells destined for the follicular B cell lineage undergo a de- based on the following phenotypes: T3-like cells were CD93+ velopmental block and enhanced turnover at the T1-to-T2 transi- CD23+sIgMlo (but cIgM+); T39-like were CD93+CD232sIgMlo tional stage, where IgD is highly upregulated (30, 31). In IgDa- (but cIgM+). As originally described, T3 cells (IgMloIgD+CD23+ macroself mice, T1 cells are largely spared from deletion, probably CD93+) were considered a late transitional stage derived from T2 because of a lack of sufficient IgD, whereas CD23+CD93+ transi- cells (31), but have recently been viewed as anergic B cells (71), tional cells are reduced in number, increased in turnover rate, and possibly similar to those generated in soluble hen egg lysozyme lacked markers associated with phenotype of T2 cells. T2 cells in the (HEL)/anti-HEL Tg mice (45). T3 cells differ from T2 cells in SP have been defined as IgMhiCD93+CD23+ (31). In the BM of their distinctly lower sIgM levels, which presumably result from wild-type mice, CD93+CD23+ T2-like cells are also present, al- interaction with autoantigens. In IgDa-macroself mice, we found though they appear to be at a slightly earlier transitional stage than that IgM2/lo T3-like cells had intracellular IgM levels similar to splenic T2 cells in that they express significantly lower levels of splenic CD93+CD23+ cells in these mice, indicating that they markers upregulated upon T1-T2 transition, including CD62L, IgD, internalized this receptor (not shown), but their levels of surface and CD21 (30, 69). In IgDa-macroself mice, the remaining CD23+ MHC II and turnover rates were higher than T3 cells from non-Tg CD93+ cells present in the SP have abnormally low levels of CD21 controls. One caveat in distinguishing T39-like cells from T1 cells and CD62L, thus exhibiting a phenotype similar to wild-type BM in IgDa-macroself mice is that in these mice, we failed to see sIgD T2-like cells. Splenic T2 cells have been suggested to include on any cells, and we were unable to reliably detect intracellular a fraction that is proliferative and responsive in particular to BAFF IgD. Moreover, all splenic CD93+ transitional B cells in IgDa- (30, 69, 106, 107). But in IgDa-macroself mice carrying a second macroself mice, regardless of CD23 expression, had identical transgene that overexpresses BAFF, CD93+CD23+ B cells were not BrdU-labeling kinetics. We suggest that the cells are T3-like and elevated in number. Thus, our results suggest that the T2 population T39-like because of their IgM downregulation. In any case, these in SP consists of at least two differentiative states: a less mature results suggest that in IgDa-macroself mice, CD93+CD23+ (pre- population reminiscent of BM T2-like cells and a developmentally sumably T2- and T3-like) cells undergo BCR activation and as 4154 TOLERANCE IN IgD SUPERANTIGEN MICE

FIGURE 9. Rescue of putative preplasma cells, but not total B cell loss, in IgDa-macroself mice by en- forced Bcl-2 expression. IgHa Bcl2 Tg BM was used to reconstitute lethally irradiated non-Tg or IgDa-macroself recipients, which were analyzed ∼8 wk later. A, Shown are representative flow cytometry plots for total CD19+ B220+ cells in non-Tg and IgDa- macroself mice. B, Numbers of tran- Downloaded from sitional and mature B cell subsets in non-Tg and IgDa-macroself mice were compared. C, Shown are representa- tive flow plots illustrating increased representation of intracellular IgM+ CD138+ preplasma cells in IgDa- macroself recipients. Also evaluated http://www.jimmunol.org/ were the numbers of IgM+ ASCs (D), IgG+CD19+ putative memory B cells (D), and l versus k L chain usages (E) in both sets of recipients. All graphs shown represent mean + SD; p values given where there is statistical signif- icance by two-tailed Student t tests; n = 4 mice per group. by guest on October 2, 2021

a population turn over about twice as rapidly as normal (t1/2 of 2 d mostly T1 type (73, 74, 110). Likewise, transitional B cells from versus 4 d). To our knowledge, this is the clearest evidence that wild-type mice have been shown to exhibit reduced proliferative transitional B cells can undergo accelerated deletion upon auto- responsiveness and CD86 upregulation to IgM cross-linking (106), antigen encounter in vivo. explaining the reduced responsiveness of B cells from IgDa- Prolonged exposure to a weakly self-reactive Ag has been shown macroself mice to such stimuli. to result in B cell anergy (109). Although details differ among the Splenic B cells in IgDa-macroself mice expressed extremely low models, a common hallmark of anergic B cells is reduced Ca2+ levels of RAG1 and 2 mRNA. If editing occurs in this population, signaling in response to IgM cross-linking (72). The T3 B cell it is likely to be at a low level and restricted to transitional B cells. population present in wild-type mice has been suggested to be However, we cannot exclude that T2-like cells in the BM can a collection of anergic B cells also based upon this fact (71). undergo receptor editing. Thus, our data are compatible with the However, in B cells isolated from IgDa-macroself mice, IgM conclusions of Sandel et al. (48, 111), who suggested that tran- stimulation triggered robust Ca2+ signals, consistent with our in- sitional B cells provided with a BM microenvironment are com- terpretation that the transitional B cells in IgDa-macroself mice are petent to undergo receptor editing. Although this could in theory distinct from wild-type (anergic) T3 cells. Although upon IgM explain the higher frequency of l+ B cells in IgDa-macroself mice, ligation, B cells from Tg mice do exhibit reduced tyrosine phos- as explained below, we rather think that this is a secondary result phorylation in some proteins, including CD19, this appears to be a of the skewing in B cell subset makeup in these mice. result of their immaturity, rather than their state of un- T39 cells are of interest because they have been suggested to be responsiveness, as their p-Tyr profile is remarkably similar to that splenic B cells undergoing receptor editing. T39 cells have been of B cells from BAFF-deficient mice, which are thought to be of defined based on experiments with 3H9H-56R anti-DNA H chain The Journal of Immunology 4155

+

FIGURE 10. Rescue of MZ and IgM Ab secreting Downloaded from B cells in IgDa-macroself mice by overexpression of BAFF. A, IgDa-macroself mice on IgHa congenic C57BL/6 background were bred to overexpress BAFF and analyzed for presence of CD23+CD932 B-2 cells. B, Shown are plots summarizing the effects of excess BAFF on average numbers of B cells at different

stages of development (+ SD) analyzed from four http://www.jimmunol.org/ mice per group; p values by two-tailed Student t tests. pp , 0.05; ppp , 0.01; pppp , 0.005. C, Evidence for increase of cIgM+CD138+ plasma cells in the SP and LNs by excess BAFF is shown by representative flow cytometry plots. DTg, double transgenic. by guest on October 2, 2021

Tg mice, which have expanded numbers of a small splenic IgMlo known to be resistant to peripheral tolerance and had an enlarged IgDhiCD93+CD232 subpopulation expressing RAG1 and 2 and CD19+CD232CD93+ compartment, failed to show a l/k skewing carrying dsDNA breaks at the Jk locus (108) (this study also on the IgDa-macroself background when compared with the newly showed some RAG expression and double-stranded breaks in T3 formed B cells in the BM. cells). Anergic and late immature B cells in the HEL/anti-HEL If editing in semimature transitional B cells is not occurring to model have also been suggested to undergo editing in vivo, but a substantial extent in the SPs of IgDa-macroself mice, then why do only at the IgM+IgD2 (presumably T1) stage (112). Although they have elevated proportions of l-L chain-bearing cells and splenic T39 cells, if they indeed exist in IgDa-macroself mice, Abs? As explained by Lindsley et al. (69), a relatively high pro- should fully engage their IgD receptors, there is no evidence that portion of newly formed and transitional B cells in normal mice their numbers are expanded. Nor is there appreciable receptor express l (∼12–15%), and they tend to be counterselected by an editing activity associated with the splenic B cells in the IgDa- unknown process during developmental progression to the follic- macroself model, because RAG levels were at background levels ular compartment, which has only ∼5% l. Similar observations despite the presence in these samples of a relatively high pro- have been made by Dingjan et al. (113). Consistent with this, portion of transitional cells. Moreover, Bcl2 Tg B cells, which are splenic B cells in BAFF2/2 mice have elevated frequencies of l 4156 TOLERANCE IN IgD SUPERANTIGEN MICE cells (data not shown). Moreover, B-1 and MZ B cells also have B-2 cells. The IgDa-macroself model may be useful in the study of higher frequencies of l than do follicular B cells (114, 115). Ab responses generated by nonfollicular B cells. Without the contribution of follicular cells, which are counter- A most unexpected feature of the IgDa-macroself model was the selected in IgDa-macroself mice, the remaining cells are thus presence of elevated numbers of IgM+ plasma cells. The simplest enriched in l by selection rather than editing. explanation for their abundance is that in the absence of follicular The MZ B cell population that develops in IgDa-macroself mice B cells, the MZ compartment undergoes more TD stimulation and is of interest for several reasons. Normal MZ B cells in rodents attendant Ab formation, with the aforementioned skewing to the may be heterogeneous, including a higher frequency of cells with IgM class. Similarly, B-1 responses, which contribute greatly to a fetal-type repertoire (26, 115–117). They are believed to derive serum IgM levels, may be enhanced in these mice and preferen- from a subset of CD23+CD21hi cells now termed MZ precursors tially yield IgM responses. The increased numbers of IgM+ plasma (118). The fact that MZ B cell numbers are subnormal in IgDa cells may also be stabilized by a presumed increase in BAFF -macroself mice might indicate that the remaining cells represent levels due to a total reduction in B cell numbers in these mice. a distinct subpopulation or lineage. Alternatively, the MZ B cells Consistent with this idea, overexpression of BAFF gives rise to developing in IgDa-macroself mice may be a transient population a further increase in their numbers. Serum IgM, but not IgG, from or rare variants that emerge and self-renew upon stringent selec- Tg mice also exhibited increased reactivity to dsDNA (data not tion. Unlike immature B cells present in IgDa-macroself mice, shown). It remains to be determined whether this reflects abnor- upon in vitro culture in the absence of superantigen, MZ B cells mal selection events, as this could be a very useful model to study from these mice do not upregulate IgD, indicating that the Tg systemic autoimmunity. environment selects for MZ B cells permanently lacking IgD. Consistent with this finding, MZ B cells in IgDa-macroself mice Acknowledgments Downloaded from turn over slowly and therefore are a stable population. Although We thank Martin L. Scott, Biogen/Idec, for providing BAFF2/2 mice and these cells may represent a special subset, preliminary studies John Cambier for anti-CD19 reagent. indicate that their H chain N-region length diversity is not ap- preciably different from that of wild-type MZ B cells (not shown). Disclosures Positive selection by autoantigens has been suggested to play The authors have no financial conflicts of interest. a role in the development of MZ B cells (26). However, in this http://www.jimmunol.org/ a study, the IgD -macroself autoantigen appears to solely play a role References in negative selection, possibly because of the relatively high af- 1. Rowe, D. S., K. Hug, L. Forni, and B. Pernis. 1973. Immunoglobulin D as finity of this ligand for the BCR. In any case, the fact that MZ a lymphocyte receptor. J. Exp. Med. 138: 965–972. B cells are formed at all suggests that those cells may have been 2. Abney, E. R., and R. M. Parkhouse. 1974. Candidate for immunoglobulin D present on murine B lymphocytes. Nature 252: 600–602. derived from a precursor lacking IgD, possibly a T1 population or 3. Melcher, U., E. S. Vitetta, M. McWilliams, M. E. Lamm, J. M. Phillips- + + the early CD23 CD93 T2-like stage. Quagliata, and J. W. Uhr. 1974. Cell surface immunoglobulin. X. Identification By contrast, B-1 cells, which also have low but variable levels of of an IgD-like molecule on the surface of murine splenocytes. J. Exp. Med. 140: 1427–1431. IgD, are not reduced in number in nonirradiated IgDa-macroself

4. Vitetta, E. S., and J. W. Uhr. 1975. Immunoglobulin-receptors revisited. Science by guest on October 2, 2021 mice. Because B-1 cells are relatively resistant to apoptosis with 189: 964–969. the ability to self-renew, homeostatic expansion and accumulation 5. Martin, L. N., and G. A. Leslie. 1977. Lymphocyte surface IgD and IgM in Macaca monkeys: ontogeny, tissue distribution and occurrence on individual of the rare IgD-negative B-1 cells are likely to be responsible for lymphocytes. Immunology 33: 865–872. their normal numbers. Thus, their numbers remain low in irradi- 6. Ruddick, J. H., and G. A. Leslie. 1977. Structure and biologic functions of a human IgD. XI. Identification and ontogeny of a rat lymphocyte immuno- ated IgD -macroself hosts reconstituted with adult BM cells, globulin having antigenic cross-reactivity with human IgD. J. Immunol. 118: which are known to have a decreased propensity to generate B-1 1025–1031. cells. Rather than positive selection, our results suggest that IgD 7. Geisberger, R., M. Lamers, and G. Achatz. 2006. The riddle of the dual ex- pression of IgM and IgD. Immunology 118: 429–437. cross-linking eliminates B-1 cells that had the tendency to express 8. Maki, R., W. Roeder, A. Traunecker, C. Sidman, M. Wabl, W. Raschke, and high levels of IgD while sparing those that expressed lower or S. Tonegawa. 1981. The role of DNA rearrangement and alternative RNA absent levels. Interestingly, although BAFF overexpression can processing in the expression of immunoglobulin delta genes. Cell 24: 353–365. a 9. Cheng, H. L., F. R. Blattner, L. Fitzmaurice, J. F. Mushinski, and P. W. Tucker. expand B-1 cell number (66), this is not observed in IgD 1982. Structure of genes for membrane and secreted murine IgD heavy chains. -macroself mice, which may also have elevated levels of BAFF Nature 296: 410–415. 10. Vitetta, E. S., and J. W. Uhr. 1977. IgD and B cell differentiation. Immunol. Rev. due to reduced total number of B cells. 37: 50–88. a Consistent with the dearth of follicular B cells in IgD -macroself 11. Wilson, M., E. Bengte´n, N. W. Miller, L. W. Clem, L. Du Pasquier, and Ag Tg mice, their Ab responses to TD Ag given in alum were G. W. Warr. 1997. A novel chimeric Ig heavy chain from a teleost fish shares similarities to IgD. Proc. Natl. Acad. Sci. U.S.A. 94: 4593–4597. significantly reduced in both primary and secondary responses of 12. Zhao, Y., I. Kacskovics, Q. Pan, D. A. Liberles, J. Geli, S. K. Davis, most Ab classes, with the major exception of IgM, levels of which H. Rabbani, and L. Hammarstrom. 2002. Artiodactyl IgD: the missing link. J. were elevated. It is not clear if these responses come from MZ, Immunol. 169: 4408–4416. 13. Zhao, Y., Q. Pan-Hammarstro¨m, I. Kacskovics, and L. Hammarstro¨m. 2003. B-1, or transitional B cell subsets. Song and Cerny (119) assessed The porcine Ig delta gene: unique chimeric splicing of the first constant region the ability of different B cell subsets to undergo TD Ab responses domain in its heavy chain transcripts. J. Immunol. 171: 1312–1318. 14. Ohta, Y., and M. Flajnik. 2006. IgD, like IgM, is a primordial immunoglobulin upon transfer into SCID mice along with carrier primed T cells. class perpetuated in most jawed vertebrates. Proc. Natl. Acad. Sci. U.S.A. 103: They found that MZ B cells are able to participate in TD Ab re- 10723–10728. sponses and to undergo a germinal center reaction, class switching, 15. Zhao, Y., H. Rabbani, A. Shimizu, and L. Hammarstro¨m. 2000. Mapping of the chicken immunoglobulin heavy-chain constant region gene locus reveals an inverted and somatic mutation (119). Under these conditions, immature alpha gene upstream of a condensed upsilon gene. Immunology 101: 348–353. cells could also contribute to germinal center formation. Their 16. Lundqvist, M. L., D. L. Middleton, S. Hazard, and G. W. Warr. 2001. The study also showed that MZ B cells were more effective than fol- immunoglobulin heavy chain locus of the duck. Genomic organization and expression of D, J, and C region genes. J. Biol. Chem. 276: 46729–46736. licular B cells at IgM responses in the primary and secondary re- 17. Lanning, D. K., S. K. Zhai, and K. L. Knight. 2003. Analysis of the 39 Cmu sponse. Phan et al. (120) concluded from experiments with anti- region of the rabbit Ig heavy chain locus. Gene 309: 135–144. 18. Fu, S. M., R. J. Winchester, and H. G. Kunkel. 1975. Similar idiotypic speci- HEL Tg B cells that MZ cells were able to generate memory and ficity for the membrane IgD and IgM of human B lymphocytes. J. Immunol. class switched responses, but at a reduced efficiency compared with 114: 250–252. The Journal of Immunology 4157

19. Salsano, F., S. S. Froland, J. B. Natvig, and T. E. Michaelsen. 1974. Same idi- 49. Nemazee, D., and M. Weigert. 2000. Revising B cell receptors. J. Exp. Med. otype of B-lymphocyte membrane IgD and IgM. Formal evidence for mono- 191: 1813–1817. clonality of chronic lymphocytic leukemia cells. Scand. J. Immunol. 3: 841–846. 50. Venkitaraman, A. R., G. T. Williams, P. Dariavach, and M. S. Neuberger. 1991. 20. Pernis, B., J. C. Brouet, and M. Seligmann. 1974. IgD and IgM on the mem- The B-cell antigen receptor of the five immunoglobulin classes. Nature 352: brane of lymphoid cells in macroglobulinemia. Evidence for identity of 777–781. membrane IgD and IgM antibody activity in a case with anti-IgG receptors. Eur. 51. Campbell, K. S., E. J. Hager, and J. C. Cambier. 1991. Alpha-chains of IgM and J. Immunol. 4: 776–778. IgD antigen receptor complexes are differentially N-glycosylated MB-1-related 21. Goding, J. W., and J. E. Layton. 1976. Antigen-induced co-capping of IgM and molecules. J. Immunol. 147: 1575–1580. IgD-like receptors on murine B cells. J. Exp. Med. 144: 852–857. 52. Wienands, J., and M. Reth. 1991. The B cell antigen receptor of class IgD can 22. Yuan, D., P. L. Witte, J. Tan, J. Hawley, and T. Dang. 1996. Regulation of IgM be expressed on the cell surface in two different forms. Eur. J. Immunol. 21: and IgD heavy chain gene expression: effect of abrogation of intergenic tran- 2373–2378. scriptional termination. J. Immunol. 157: 2073–2081. 53. Kim, K. M., and M. Reth. 1995. The B cell antigen receptor of class IgD in- 23. Reed, D. J., J. Hawley, T. Dang, and D. Yuan. 1994. Role of differential mRNA duces a stronger and more prolonged protein tyrosine phosphorylation than that stability in the regulated expression of IgM and IgD. J. Immunol. 152: 5330–5336. of class IgM. J. Exp. Med. 181: 1005–1014. 24. Havran, W. L., D. L. DiGiusto, and J. C. Cambier. 1984. mIgM:mIgD ratios on 54. Monroe, J. G., W. L. Havran, and J. C. Cambier. 1983. B lymphocyte activa- B cells: mean mIgD expression exceeds mIgM by 10-fold on most splenic B tion: entry into cell cycle is accompanied by decreased expression of IgD but cells. J. Immunol. 132: 1712–1716. not IgM. Eur. J. Immunol. 13: 208–213. 25. Hayakawa, K., R. R. Hardy, M. Honda, L. A. Herzenberg, A. D. Steinberg, and 55. Bourgois, A., K. Kitajima, I. R. Hunter, and B. A. Askonas. 1977. Surface L. A. Herzenberg. 1984. Ly-1 B cells: functionally distinct lymphocytes that immunoglobulins of lipopolysaccharide-stimulated spleen cells. The behavior secrete IgM autoantibodies. Proc. Natl. Acad. Sci. U.S.A. 81: 2494–2498. of IgM, IgD and IgG. Eur. J. Immunol. 7: 151–153. 26. Martin, F., and J. F. Kearney. 2002. Marginal-zone B cells. Nat. Rev. Immunol. 56. Hodgkin, P. D., J.-H. Lee, and A. B. Lyons. 1996. B cell differentiation and isotype 2: 323–335. switching is related to division cycle number. J. Exp. Med. 184: 277–281. 27. Allman, D. M., S. E. Ferguson, and M. P. Cancro. 1992. Peripheral B cell 57. Haas, D., and G. F. Hoffmann. 2006. Mevalonate kinase deficiencies: from maturation. I. Immature peripheral B cells in adults are heat-stable antigenhi mevalonic aciduria to hyperimmunoglobulinemia D syndrome. Orphanet J. and exhibit unique signaling characteristics. J. Immunol. 149: 2533–2540. Rare Dis. 1: 13. 28. Allman, D. M., S. E. Ferguson, V. M. Lentz, and M. P. Cancro. 1993. Peripheral 58. Kluin, P. M., H. Kayano, V. J. Zani, H. C. Kluin-Nelemans, P. W. Tucker, Downloaded from B cell maturation. II. Heat-stable antigen(hi) splenic B cells are an immature E. Satterwhite, and M. J. Dyer. 1995. IgD class switching: identification of a novel developmental intermediate in the production of long-lived marrow-derived B recombination site in neoplastic and normal B cells. Eur.J. Immunol.25: 3504–3508. cells. J. Immunol. 151: 4431–4444. 59. Koelsch, K., N. Y. Zheng, Q. Zhang, A. Duty, C. Helms, M. D. Mathias, 29. Rolink, A. G., J. Andersson, and F. Melchers. 1998. Characterization of im- M. Jared, K. Smith, J. D. Capra, and P. C. Wilson. 2007. Mature B cells class mature B cells by a novel monoclonal antibody, by turnover and by mitogen switched to IgD are autoreactive in healthy individuals. J. Clin. Invest. 117: reactivity. Eur. J. Immunol. 28: 3738–3748. 1558–1565. 30. Loder, F., B. Mutschler, R. J. Ray, C. J. Paige, P. Sideras, R. Torres, 60. Chen, K., W. Xu, M. Wilson, B. He, N. W. Miller, E. Bengte´n, E. S. Edholm, M. C. Lamers, and R. Carsetti. 1999. B cell development in the spleen takes P. A. Santini, P. Rath, A. Chiu, et al. 2009. Immunoglobulin D enhances im- place in discrete steps and is determined by the quality of B cell receptor- mune surveillance by activating antimicrobial, proinflammatory and B cell- http://www.jimmunol.org/ derived signals. J. Exp. Med. 190: 75–89. stimulating programs in basophils. Nat. Immunol. 10: 889–898. 31. Allman, D., R. C. Lindsley, W. DeMuth, K. Rudd, S. A. Shinton, and 61. Roes, J., and K. Rajewsky. 1993. Immunoglobulin D (IgD)-deficient mice re- R. R. Hardy. 2001. Resolution of three nonproliferative immature splenic B cell veal an auxiliary receptor function for IgD in antigen-mediated recruitment of subsets reveals multiple selection points during peripheral B cell maturation. J. B cells. J. Exp. Med. 177: 45–55. Immunol. 167: 6834–6840. 62. Nitschke, L., M. H. Kosco, G. Ko¨hler, and M. C. Lamers. 1993. Immunoglobulin 32. Lutz, C., B. Ledermann, M. H. Kosco-Vilbois, A. F. Ochsenbein, D-deficient mice can mount normal immune responses to thymus-independent R. M. Zinkernagel, G. Ko¨hler, and F. Brombacher. 1998. IgD can largely and -dependent antigens. Proc. Natl. Acad. Sci. U.S.A. 90: 1887–1891. substitute for loss of IgM function in B cells. Nature 393: 797–801. 63. Aı¨t-Azzouzene, D., L. Verkoczy, J. Peters, A. Gavin, P. Skog, J. L. Vela, and 33. Brink, R., C. C. Goodnow, J. Crosbie, E. Adams, J. Eris, D. Y. Mason, D. Nemazee. 2005. An immunoglobulin C kappa-reactive single chain antibody S. B. Hartley, and A. Basten. 1992. Immunoglobulin M and D antigen receptors fusion protein induces tolerance through receptor editing in a normal polyclonal are both capable of mediating B lymphocyte activation, deletion, or anergy immune system. J. Exp. Med. 201: 817–828. after interaction with specific antigen. J. Exp. Med. 176: 991–1005. 64. Ait-Azzouzene, D., A. L. Gavin, P. Skog, B. Duong, and D. Nemazee. 2006. by guest on October 2, 2021 34. Vitetta, E. S., U. Melcher, M. McWilliams, M. E. Lamm, J. M. Phillips- Effect of cell:cell competition and BAFF expression on peripheral B cell tol- Quagliata, and J. W. Uhr. 1975. Cell surface immunoglobulin. XI. The ap- erance and B-1 cell survival in transgenic mice expressing a low level of Ig- pearance of an IgD-like molecule on murine lymphoid cells during ontogeny. J. kappa-reactive macroself antigen. Eur. J. Immunol. 36: 985–996. Exp. Med. 141: 206–215. 65. Aı¨t-Azzouzene, D., L. Verkoczy, B. Duong, P. Skog, A. L. Gavin, and 35. Klinman, N. R. 1996. The “clonal selection hypothesis” and current concepts of D. Nemazee. 2006. Split tolerance in peripheral B cell subsets in mice ex- B cell tolerance. Immunity 5: 189–195. pressing a low level of Igkappa-reactive ligand. J. Immunol. 176: 939–948. 36. Nossal, G. J. 1983. Cellular mechanisms of immunologic tolerance. Annu. Rev. 66. Gavin, A. L., B. Duong, P. Skog, D. Aı¨t-Azzouzene, D. R. Greaves, M. L. Scott, Immunol. 1: 33–62. and D. Nemazee. 2005. deltaBAFF, a splice isoform of BAFF, opposes full- 37. Cambier, J. C., J. R. Kettman, E. S. Vitetta, and J. W. Uhr. 1976. Differential length BAFF activity in vivo in transgenic mouse models. J. Immunol. 175: susceptibility of neonatal and adult murine spleen cells to in vitro induction of 319–328. B-cell tolerance. J. Exp. Med. 144: 293–297. 67. Strasser, A., S. Whittingham, D. L. Vaux, M. L. Bath, J. M. Adams, S. Cory, 38. Carsetti, R., G. Ko¨hler, and M. C. Lamers. 1993. A role for immunoglobulin D: and A. W. Harris. 1991. Enforced BCL2 expression in B-lymphoid cells pro- interference with tolerance induction. Eur. J. Immunol. 23: 168–178. longs antibody responses and elicits autoimmune disease. Proc. Natl. Acad. Sci. 39. Finkelman, F. D., J. M. Holmes, O. I. Dukhanina, and S. C. Morris. 1995. U.S.A. 88: 8661–8665. Cross-linking of membrane immunoglobulin D, in the absence of T cell help, 68. Stall, A. M., and M. R. Loken. 1984. Allotypic specificities of murine IgD and kills mature B cells in vivo. J. Exp. Med. 181: 515–525. IgM recognized by monoclonal antibodies. J. Immunol. 132: 787–795. 40. Goodnow, C. C., J. Crosbie, H. Jorgensen, R. A. Brink, and A. Basten. 1989. 69. Lindsley, R. C., M. Thomas, B. Srivastava, and D. Allman. 2007. Generation of Induction of self-tolerance in mature peripheral B lymphocytes. [see com- peripheral B cells occurs via two spatially and temporally distinct pathways. ments] Nature 342: 385–391. Blood 109: 2521–2528. 41. Parry, S. L., J. Hasbold, M. Holman, and G. G. Klaus. 1994. Hypercross-linking 70. Cariappa, A., C. Chase, H. Liu, P. Russell, and S. Pillai. 2007. Naive re- surface IgM or IgD receptors on mature B cells induces apoptosis that is re- circulating B cells mature simultaneously in the spleen and bone marrow. Blood versed by costimulation with IL-4 and anti-CD40. J. Immunol. 152: 2821–2829. 109: 2339–2345. 42. Russell, D. M., Z. Dembic´, G. Morahan, J. F. Miller, K. Bu¨rki, and D. Nemazee. 71. Merrell, K. T., R. J. Benschop, S. B. Gauld, K. Aviszus, D. Decote-Ricardo, 1991. Peripheral deletion of self-reactive B cells. Nature 354: 308–311. L. J. Wysocki, and J. C. Cambier. 2006. Identification of anergic B cells within 43. Cyster, J. G., S. B. Hartley, and C. C. Goodnow. 1994. Competition for fol- a wild-type repertoire. Immunity 25: 953–962. licular niches excludes self-reactive cells from the recirculating B-cell reper- 72. Seo, S. J., L. Mandik-Nayak, and J. Erikson. 2003. B cell anergy and systemic toire. [see comments] Nature 371: 389–395. lupus erythematosus. Curr. Dir. Autoimmun. 6: 1–20. 44. Fulcher, D. A., and A. Basten. 1994. Reduced life span of anergic self-reactive 73. Schiemann, B., J. L. Gommerman, K. Vora, T. G. Cachero, S. Shulga- B cells in a double-transgenic model. J. Exp. Med. 179: 125–134. Morskaya, M. Dobles, E. Frew, and M. L. Scott. 2001. An essential role for 45. Goodnow, C. C., J. Crosbie, S. Adelstein, T. B. Lavoie, S. J. Smith-Gill, BAFF in the normal development of B cells through a BCMA-independent R. A. Brink, H. Pritchard-Briscoe, J. S. Wotherspoon, R. H. Loblay, K. Raphael, pathway. Science 293: 2111–2114. et al. 1988. Altered immunoglobulin expression and functional silencing of 74. Schneider, P., H. Takatsuka, A. Wilson, F. Mackay, A. Tardivel, S. Lens, self-reactive B lymphocytes in transgenic mice. Nature 334: 676–682. T. G. Cachero, D. Finke, F. Beermann, and J. Tschopp. 2001. Maturation of 46. Hertz, M., and D. Nemazee. 1997. BCR ligation induces receptor editing in marginal zone and follicular B cells requires B cell activating factor of the IgM+IgD- bone marrow B cells in vitro. Immunity 6: 429–436. tumor necrosis factor family and is independent of B cell maturation antigen. J. 47. Tiegs, S. L., D. M. Russell, and D. Nemazee. 1993. Receptor editing in self- Exp. Med. 194: 1691–1697. reactive bone marrow B cells. J. Exp. Med. 177: 1009–1020. 75. Noorchashm, H., A. Bui, H. L. Li, A. Eaton, L. Mandik-Nayak, C. Sokol, 48. Sandel, P. C., and J. G. Monroe. 1999. Negative selection of immature B cells K. M. Potts, E. Pure´, and J. Erikson. 1999. Characterization of anergic anti- by receptor editing or deletion is determined by site of antigen encounter. DNA B cells: B cell anergy is a T cell-independent and potentially reversible Immunity 10: 289–299. process. Int. Immunol. 11: 765–776. 4158 TOLERANCE IN IgD SUPERANTIGEN MICE

76. Nemazee, D. 2006. Receptor editing in lymphocyte development and central 99. Skelly, R. R., Y. Baine, A. Ahmed, B. Xue, and G. J. Thorbecke. 1983. Cell tolerance. Nat. Rev. Immunol. 6: 728–740. surface phenotype of lymphoid cells from normal mice and mice treated with 77. Chevrier, S., C. Genton, A. Kallies, A. Karnowski, L. A. Otten, B. Malissen, monoclonal anti-IgD from birth. J. Immunol. 130: 15–18. M. Malissen, M. Botto, L. M. Corcoran, S. L. Nutt, and H. Acha-Orbea. 2009. 100. Dresser, D. W., R. M. Parkhouse, and A. M. Popham. 1983. Pretreatment of CD93 is required for maintenance of antibody secretion and persistence of mice with anti-IgD serum does not result in suppression of immune re- plasma cells in the bone marrow niche. Proc. Natl. Acad. Sci. U.S.A. 106: sponsiveness induced by TNP-AECM-Ficoll or lipopolysaccharide. Immunol- 3895–3900. ogy 48: 497–501. 78. Lesley, R., Y. Xu, S. L. Kalled, D. M. Hess, S. R. Schwab, H. B. Shu, and 101. Xue, B., T. Hirano, B. Pernis, Z. Ovary, and G. J. Thorbecke. 1984. Physiology J. G. Cyster. 2004. Reduced competitiveness of autoantigen-engaged B cells of IgD. III. Effect of treatment with anti-IgD from birth on the magnitude and due to increased dependence on BAFF. Immunity 20: 441–453. isotype distribution of the immune response in the spleen. Eur. J. Immunol. 14: 79. Thien, M., T. G. Phan, S. Gardam, M. Amesbury, A. Basten, F. Mackay, and 81–86. R. Brink. 2004. Excess BAFF rescues self-reactive B cells from peripheral 102. Denis, O., F. Macedo-Soares, D. Latinne, F. Nisol, and H. Bazin. 1994. In vivo deletion and allows them to enter forbidden follicular and marginal zone niches. study of mIgM and mIgD cross-linking on murine B cells. Scand. J. Immunol. Immunity 20: 785–798. 39: 625–632. 80. Lang, J., B. Arnold, G. Hammerling, A. W. Harris, S. Korsmeyer, D. Russell, 103. Finkelman, F. D., J. J. Mond, and E. S. Metcalf. 1983. Effects of neonatal anti- A. Strasser, and D. Nemazee. 1997. Enforced Bcl-2 expression inhibits antigen- delta antibody treatment on the murine immune system. I. Suppression of de- mediated clonal elimination of peripheral B cells in an antigen dose-dependent velopment of surface IgD+ B cells and expansion of a surface IgM+ IgD- manner and promotes receptor editing in autoreactive, immature B cells. J. Exp. B lymphocyte population. J. Immunol. 131: 593–600. Med. 186: 1513–1522. 104. Metcalf, E. S., J. J. Mond, and F. D. Finkelman. 1983. Effects of neonatal anti- 81. Hardy, R. R., C. E. Carmack, S. A. Shinton, J. D. Kemp, and K. Hayakawa. delta antibody treatment on the murine immune system. II. Functional capacity 1991. Resolution and characterization of pro-B and pre-pro-B cell stages in of a stable sIgM+sIa+sIgD- B cell population. J. Immunol. 131: 601–605. normal mouse bone marrow. J. Exp. Med. 173: 1213–1225. 105. Melamed, D., R. J. Benschop, J. C. Cambier, and D. Nemazee. 1998. De- 82. Kincade, P. W., A. R. Lawton, D. E. Bockman, and M. D. Cooper. 1970. velopmental regulation of B lymphocyte immune tolerance compartmentalizes Suppression of immunoglobulin G synthesis as a result of antibody-mediated clonal selection from receptor selection. Cell 92: 173–182. suppression of immunoglobulin M synthesis in chickens. Proc. Natl. Acad. Sci. 106. Meyer-Bahlburg, A., S. F. Andrews, K. O. Yu, S. A. Porcelli, and D. J. Rawlings. U.S.A. 67: 1918–1925. 2008. Characterization of a late transitional B cell population highly sensitive to 83. Leslie, G. A., and L. N. Martin. 1973. Suppression of chicken immunoglobulin BAFF-mediated homeostatic proliferation. J. Exp. Med. 205: 155–168. Downloaded from ontogeny by F(ab9)2 fragments of anti-mu chain and by anti-L chain. Int. Arch. 107. Hsu, B. L., S. M. Harless, R. C. Lindsley, D. M. Hilbert, and M. P. Cancro. Allergy Appl. Immunol. 45: 429–438. 2002. Cutting edge: BLyS enables survival of transitional and mature B cells 84. Lawton, A. R., 3rd, R. Asofsky, M. B. Hylton, and M. D. Cooper. 1972. through distinct mediators. J. Immunol. 168: 5993–5996. Suppression of immunoglobulin class synthesis in mice. I. Effects of treatment 108. Kiefer, K., P. B. Nakajima, J. Oshinsky, S. H. Seeholzer, M. Radic, with antibody to -chain. J. Exp. Med. 135: 277–297. G. C. Bosma, and M. J. Bosma. 2008. Antigen receptor editing in anti-DNA 85. Manning, D. D., and J. W. Jutila. 1972. Immunosuppression in mice injected transitional B cells deficient for surface IgM. J. Immunol. 180: 6094–6106. with heterologous anti-immunoglobulin antisera. J. Immunol. 108: 282–285. 109. Goodnow, C. C., J. Crosbie, S. Adelstein, T. B. Lavoie, S. J. Smith-Gill,

86. Manning, D. D., and J. W. Jutila. 1972. Immunosuppression of mice injected D. Y. Mason, H. Jorgensen, R. A. Brink, H. Pritchard-Briscoe, M. Loughnan, http://www.jimmunol.org/ with heterologous anti-immunoglobulin heavy chain antisera. J. Exp. Med. 135: et al. 1989. Clonal silencing of self-reactive B lymphocytes in a transgenic 1316–1333. mouse model. Cold Spring Harb. Symp. Quant. Biol. 54: 907–920. 87. Murgita, R. A., C. A. Mattioli, and T. B. Tomasi, Jr. 1973. Production of 110. Thompson, J. S., S. A. Bixler, F. Qian, K. Vora, M. L. Scott, T. G. Cachero, a runting syndrome and selective A deficiency in mice by the administration of C. Hession, P. Schneider, I. D. Sizing, C. Mullen, et al. 2001. BAFF-R, a newly anti-heavy chain antisera. J. Exp. Med. 138: 209–228. identified TNF receptor that specifically interacts with BAFF. Science 293: 88. Cooper, M. D., J. F. Kearney, W. E. Gathings, and A. R. Lawton. 1980. Effects 2108–2111. of anti-Ig antibodies on the development and differentiation of B cells. Im- 111. Sandel, P. C., M. Gendelman, G. Kelsoe, and J. G. Monroe. 2001. Definition of munol. Rev. 52: 29–53. a novel cellular constituent of the bone marrow that regulates the response of 89. Gordon, J., R. A. Murgita, and T. B. Tomasi, Jr. 1975. The immune response of immature B cells to B cell antigen receptor engagement. J. Immunol. 166: mice treated with anti-mu antibodies: the effect on antibody-forming cells, their 5935–5944. precursors and helper cells assayed in vitro. J. Immunol. 114: 1808–1812. 112. Tze, L. E., K. L. Hippen, and T. W. Behrens. 2003. Late immature B cells

90. Raff, M. C., M. Megson, J. J. Owen, and M. D. Cooper. 1976. Early production (IgMhighIgDneg) undergo a light chain receptor editing response to soluble by guest on October 2, 2021 of intracellular IgM by B-lymphocyte precursors in mouse. Nature 259: 224– self-antigen. J. Immunol. 171: 678–682. 226. 113. Dingjan, G. M., S. Middendorp, K. Dahlenborg, A. Maas, F. Grosveld, and 91. Layton, J. E., G. R. Johnson, D. W. Scott, and G. J. Nossal. 1978. The anti-delta R. W. Hendriks. 2001. Bruton’s tyrosine kinase regulates the activation of gene suppressed mouse. Eur. J. Immunol. 8: 325–330. rearrangements at the lambda light chain locus in precursor B cells in the 92. Parkhouse, R. M., and D. W. Dresser. 1978. Effect of anti-IgD serum on im- mouse. J. Exp. Med. 193: 1169–1178. mune responses. Adv. Exp. Med. Biol. 107: 43–51. 114. Hayakawa, K., R. R. Hardy, and L. A. Herzenberg. 1986. Peritoneal Ly-1 93. Jacobson, E. B., Y. Baine, Y. W. Chen, T. Flotte, M. J. O’Neil, B. Pernis, B cells: genetic control, autoantibody production, increased lambda light chain G. W. Siskind, G. J. Thorbecke, and P. Tonda. 1981. Physiology of IgD. I. expression. Eur. J. Immunol. 16: 450–456. Compensatory phenomena in B lymphocyte activation in mice treated with 115. Carey, J. B., C. S. Moffatt-Blue, L. C. Watson, A. L. Gavin, and A. J. Feeney. anti-IgD antibodies. J. Exp. Med. 154: 318–332. 2008. Repertoire-based selection into the marginal zone compartment during 94. Tokuhisa, T., F. T. Gadus, L. A. Herzenberg, and L. A. Herzenberg. 1981. B cell development. J. Exp. Med. 205: 2043–2052. Monoclonal antibody to an IgD allotype induces a new type of allotype sup- 116. Dammers, P. M., A. Visser, E. R. Popa, P. Nieuwenhuis, and F. G. Kroese. 2000. pression in the mouse. J. Exp. Med. 154: 921–934. Most marginal zone B cells in rat express germline encoded Ig VH genes and 95. Thorbecke, G. J., Y. Baine, B. Xue, Y. W. Chen, B. Pernis, G. W. Siskind, and are ligand selected. J. Immunol. 165: 6156–6169. E. B. Jacobson. 1982. Effects of allotype specific anti-IgD on the immune re- 117. Schelonka, R. L., J. Tanner, Y. Zhuang, G. L. Gartland, M. Zemlin, and sponses of homozygous mice. Ann. N. Y. Acad. Sci. 399: 304–309. H. W. Schroeder, Jr. 2007. Categorical selection of the antibody repertoire in 96. Metcalf, E. S., J. J. Mond, I. Scher, M. A. LaVeck, and F. D. Finkelman. 1982. splenic B cells. Eur. J. Immunol. 37: 1010–1021. B-cell function in mice treated with anti-IgD from birth. Ann. N. Y. Acad. Sci. 118. Allman, D., and S. Pillai. 2008. Peripheral B cell subsets. Curr. Opin. Immunol. 399: 351–359. 20: 149–157. 97. Baine, Y., Y. W. Chen, E. B. Jacobson, B. Pernis, G. W. Siskind, and 119. Song, H., and J. Cerny. 2003. Functional heterogeneity of marginal zone B cells G. J. Thorbecke. 1982. Chronic suppression in mice with anti-IgD: role of B- revealed by their ability to generate both early antibody-forming cells and cell numbers. Ann. N. Y. Acad. Sci. 399: 360–367. germinal centers with hypermutation and memory in response to a T-dependent 98. Baine, Y., Y. W. Chen, E. B. Jacobson, B. Pernis, G. W. Siskind, and antigen. J. Exp. Med. 198: 1923–1935. G. J. Thorbecke. 1982. Physiology of Igd. II. Lack of humoral immune re- 120. Phan, T. G., S. Gardam, A. Basten, and R. Brink. 2005. Altered migration, sponsiveness in lymph nodes of mice treated with anti-IgD from birth. Eur. J. recruitment, and somatic hypermutation in the early response of marginal zone Immunol. 12: 882–886. B cells to T cell-dependent antigen. J. Immunol. 174: 4567–4578.