Humoral Immune Responses in Cr2−/− Mice: Enhanced Affinity Maturation but Impaired Persistence

This information is current as Zhibin Chen, Sergei B. Koralov, Mariya Gendelman, of September 25, 2021. Michael C. Carroll and Garnett Kelsoe J Immunol 2000; 164:4522-4532; ; doi: 10.4049/jimmunol.164.9.4522 http://www.jimmunol.org/content/164/9/4522 Downloaded from

References This article cites 74 articles, 38 of which you can access for free at: http://www.jimmunol.org/content/164/9/4522.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

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

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

by guest on September 25, 2021 *average

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

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2000 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Humoral Immune Responses in Cr2؊/؊ Mice: Enhanced Affinity Maturation but Impaired Antibody Persistence1

Zhibin Chen,* Sergei B. Koralov,* Mariya Gendelman,* Michael C. Carroll,† and Garnett Kelsoe2*

Deficiency in CD21/CD35 by disruption of the Cr2 loci leads to impaired humoral immune responses. In this study, we detail the role of CD21/CD35 on Ab responses to the (4-hydroxy-3-nitrophenyl)acetyl conjugated to chicken gamma-globulin. Sur- ,prisingly, Cr2؊/؊ mice generate significant Ab responses and (GC) reactions to low doses of this Ag in alum although the magnitude of their responses is much reduced in comparison with those of Cr2؉/؊ and C57BL/6 controls. Increasing Ag dose partially corrected this deficit. In situ study of the somatic genetics of GC B cells demonstrated that VDJ hypermutation does not require CD21/CD35, and Cr2؊/؊ mice exhibited enhanced affinity maturation of serum Ab in the post-GC phase of the ؊ ؊ primary response. On the other hand, Cr2 / mice displayed accelerated loss of serum Ab and long-lived Ab-forming cells. These Downloaded from observations suggest that activation/survival signals mediated by CD21 and/or the retention of Ag by CD21/CD35 play important roles in the generation, quality, and maintenance of serum Ab. The Journal of Immunology, 2000, 164: 4522–4532.

he discovery of complement receptors on B cells by Nus- mains of human CD21 suppressed primary Ab responses to T- senzweig and colleagues (1–3) suggested that the com- dependent Ags in mice (17). Two mechanisms have been proposed T plement system might affect acquired . Pepys (4, to explain enhancement of by CD21 (18, 19): http://www.jimmunol.org/ 5) tested this possibility by transient depletion of C3 in mice with facilitation of Ag retention by FDC in germinal centers (GCs) cobra venom factor and observed suppression of Ab responses to (20–23) and enhanced recruitment of the CD21/CD19/CD81 co- T-dependent Ags. Subsequent studies in guinea pigs (6–9), dogs receptor into the B cell Ag receptor (BCR) complex (24–27). Ag (10), humans (11), and mice (12) demonstrated that genetic defi- bound to FDCs drives the GC reaction that generates the B cell ciencies in C2, C3, or C4 resulted in similar immunological de- memory compartment, whereas the CD19/CD21 coreceptor lowers fects. Recently, Dempsey et al. (13) generated T-dependent Ags B cell activation thresholds. comprised of hen egg lysozyme and two or three tandemly aligned Several groups have used gene disruption to clarify the role(s) of copies of C3d. These novel Ags were 1,000- to 10,000-fold more CD21/35 in immunity (28–30). In mice, CD21 and CD35 are al- by guest on September 25, 2021 immunogenic in mice than the unmodified hen egg lysozyme, dra- ternatively spliced gene products encoded by the Cr2 locus (31– matically illustrating the importance of complement in enhancing 33). Disruption of the Cr2 loci abolishes expression of both CD21 Ab responses. and CD35 and results in impaired humoral immune responses (28, The effect of complement on B cell responses is thought to be 29). However, the degree of impairment is controversial. One line mediated by two distinct complement receptors, CD21 (CR2) and of Cr2Ϫ/Ϫ mice failed to generate serum Ab responses to immu- CD35 (CR1), that are expressed on B cells and follicular dendritic 7 nization with 3 ϫ 10 PFU of bacteriophage ⌽X174 in the absence cells (FDCs).3 Administration of a rat mAb specific for both CD21 of adjuvant. Nonetheless, a 10-fold increase in Ag dose elicited Ab and CD35 suppressed primary Ab responses of mice to T-depen- production but only to levels significantly below that of wild-type dent Ags, whereas an Ab specific for CD35 alone had a more Ϫ Ϫ controls (28). Reconstitution of lethally irradiated Cr2 / mice modest immunosuppressive effect (14–16). This suggested that ϩ ϩ with bone marrow from MHC-matched Cr2 / donors repaired CD21 may play a more prominent role in mediating complement’s this defect, indicating that impaired Ab production was a defect of immune-enhancing effects than does CD35. Supporting this argu- B cells (28). Supporting this notion, mice lacking CD21/35 only on ment, a fusion protein containing the complement-binding do- B cells failed to generate Ab responses to 10 ␮g (4-hydroxy-3- nitrophenyl)acetyl (NP)-keyhole limpet hemocyanin in alum adju- *Department of Immunology, Duke University Medical Center, Durham, NC 27710; vant (30). An independently generated line of Cr2Ϫ/Ϫ mice also † and Center for Blood Research, Harvard Medical School, Boston, MA 02115 displayed markedly reduced primary humoral responses but re- Received for publication September 14, 1999. Accepted for publication February sponded to both high and low doses of SRBCs (29). In addition, 18, 2000. defective Ab responses were observed even when this line of The costs of publication of this article were defrayed in part by the payment of page Ϫ/Ϫ ϩ/ϩ charges. This article must therefore be hereby marked advertisement in accordance Cr2 mice was reconstituted with Cr2 bone marrow, sug- with 18 U.S.C. Section 1734 solely to indicate this fact. gesting a significant role for CD21/CD35 on FDCs (34). 1 This work was supported in part by U.S. Public Health Service Grants AI-24335 and CD21 has also been shown to promote the survival of GC B AG-13789 (to G.K.), AI-39246, AI-36389 (to M.C.C.), and AG-10207 (to G.K. and M.C.C.). cells. Stimulation of human tonsillar GC B cells with anti-CD21 mAbs in vitro induced the expression of Bcl-2 and reduced levels 2 Address correspondence and reprint requests to Dr. Garnett Kelsoe, Department of Ϫ/Ϫ Immunology, Box 3010, Duke University Medical Center, Durham, NC 27710. E- of apoptosis (35, 36). Adoptively transferred, Ag-specific Cr2 mail address: [email protected] B cells did not persist in the GCs of wild-type mice, despite ex- 3 Abbreviations used in this paper: FDC, follicular dendritic cell; GC, germinal cen- pression of BCRs with high affinity for Ag (37). Nonetheless, in ter; BCR, B cell Ag receptor; NP, (4-hydroxy-3-nitrophenyl)acetyl; AFC, Ab-forming the absence of competition from CD21/CD35ϩ cells, B lympho- cell; CG, chicken gamma-globulin; ELISPOT, enzyme-linked immunospot; PNA, Ϫ Ϫ peanut agglutinin. cytes in Cr2 / mice can form GCs (28, 29, 34), although the

Copyright © 2000 by The American Association of Immunologists 0022-1767/00/$02.00 The Journal of Immunology 4523

magnitude of the reaction is significantly reduced (28). The quality Mice were immunized with a single i.p. injection of NP12-CG precipitated of the GC reaction in the absence of CD21/CD35, e.g., hypermu- in alum. tation, and affinity maturation, and maintenance of B cell memory compartments, remains unknown. Quantification of serum NP-specific Abs To investigate further the function of CD21/CD35 in B cell responses, we bred the Cr2Ϫ/Ϫ mutation onto the Ighb genetic Serum IgM, IgG1, or ␭1 Ab specific for NP was quantified by ELISA on NP6-BSA or NP20-BSA (48). In brief, 96-well plates (Falcon 3912; Becton background of C57BL/6 mice to study immune responses to NP. In ␮ b ␭ Dickinson, Oxnard, CA) were coated with 20 g/ml NP-BSA in 0.1 M all inbred mouse strains carrying Igh and Ig 1, primary Ab re- carbonate buffer (pH 8.8) at 4°C overnight and then were blocked and sponses to immunogenic conjugates of NP are clonally restricted washed with PBS (pH7.4) containing 0.1% Tween 20 (Sigma, St. Louis, (38). Virtually all NP-specific Abs in Ighb mice bear the ␭1 light MO) and 1% BSA. Serially diluted sera were added to duplicate wells and chain (39, 40) and use V 186.2 and DFL16.1 gene segments to incubated for1hatroom temperature. Each plate included standard con- H ␭ ␥ ␭ encode the Ig heavy chain (41–43). This response is compartmen- trols of serially diluted monoclonal IgG1/ 1 Abs, H33L 1/ 1, and pE- ␥ Х ϫ 7 VHC 1 (50). These Abs bind NP with different affinities (Ka 2.0 10 talized into pauciclonal foci of Ab-forming cells (AFCs) along the Ϫ1 ϫ 6 Ϫ1 ␭ M and 1 10 M , respectively). The IgM/ 1 anti-NP Ab, B1-8 (Ka periarteriolar lymphoid sheath (PALS) and follicular GCs (42– Х 1.0 ϫ 106 MϪ1) (50), was used as a control to quantititate NP-specific 45). A common precursor(s) that is activated in the outer PALS serum IgM Ab. After washing, bound serum Ab was revealed by HRP- establishes both populations (42). NP initially activates B cells conjugated goat anti-mouse IgM or IgG1 (Southern Biotechnology Asso- ␭ expressing ␭1 light chain and heavy chains encoded by V gene ciates, Birmingham, AL) or biotinylated anti- 1 (Ls136) and then by HRP- H conjugated streptavidin (Southern Biotechnology Associates). HRP segments in the V186.2 and V3 subfamilies of the J558 VH ho- activity was visualized using TMB peroxidase substrate kit (Bio-Rad Lab- mology group (44); as the primary response progresses, GC B cells oratory, Hercules, CA). After the reaction was stopped with 1N sulfuric Downloaded from carrying V186.2-to-DFL16.1 VDJ rearrangements (so-called ca- acid, ODs were read at 450 nm on an Emax ELISA reader (Molecular nonical rearrangements) achieve dominance. By day 8 of the pri- Devices, Sunnyvale, CA) and analyzed with SOFTmax PRO software mary response, Ն80% of ␭1ϩ GC B cells express rearrangements (Molecular Devices). Concentrations of serum Abs were determined by the comparison of titrated samples to standard curves. of V186.2, whereas the VH gene segments that are common in the early primary response, V23, C1H4, CH10, and 24.8 (analogues),

become rare. By day 7–8, primary GC B cells acquire point mu- Enzyme-linked immunospot (ELISPOT) http://www.jimmunol.org/ tations and are subject to a affinity-driven selection. Crippling mu- tations in VDJ rearrangements are common in day 8 GCs but are NP-specific AFCs from bone marrow were estimated by ELISPOT on NP -BSA and NP -BSA substrates as described (48). rare at days 14 and 16 after immunization. On the other hand, 20 6 mutations that are known to increase the affinity to NP are more Affinity measurements of anti-NP serum Abs frequently observed in late GCs (43). The affinity of anti-NP serum Ab was estimated by calculating the ratios of

The GC reaction is necessary for long-lasting primary serum Ab NP6-binding Ab to NP20-binding Ab (51). Fluorescence quenching (52, 53) responses. A single exposure to Ag elicits persistent AFCs in bone was also performed to measure the Ka of pooled samples of serum Ab. marrow (46–49) and significant levels of Ab for as long as 120 Briefly, serum IgG was purified from the sera of four to eight mice using

a protein G-Sepharose column (Amersham Pharmacia Biotech, Picataway, by guest on September 25, 2021 days (48). Even after the GC reaction wanes, bone marrow AFCs NJ); recovered IgG was adjusted to a concentration of 50 ␮g/ml in PBS undergo affinity-driven clonal selection, and the affinity of the se- containing 0.02% Tween 20. Fluorescence quench was titrated over a rum Ab increases (48). The forces that drive these processes are three-log range (10Ϫ8–10Ϫ5 M) of monovalent hapten (NP-caproate; Ge- unknown. nosys). Nonspecific quenching by an irrelevant, dextran-specific Ab In this study we have investigated the primary Ab response and (IgG1), MOPC21 (ICN Pharmaceuticals, Costa Mesa, CA), was deter- mined in each assay for the calculation of NP-specific K . GC reaction to NP in Cr2Ϫ/Ϫ mice, to characterize the effects of a Ag dose, the persistence and affinity of serum Ab, and the somatic Immunohistochemistry genetics of the GC reaction. Six-micrometer-thick sections of frozen spleen were prepared (45) and stored at Ϫ80°C until use. Before staining, sections were rehydrated in Materials and Methods PBS and blocked with PBS containing 10% FCS and 0.1% Tween 20. Mice Hydrated sections were then stained in tandem with HRP-conjugated pea- nut agglutinin (PNA) (EY Laboratories, San Mateo, CA) and biotinylated Cr2Ϫ/Ϫ (28) mice were originally established on 129/Sv genetic back- Ls136 and then with streptavidin-alkaline phosphatase (Southern Biotech- ground. These mice were crossed with C57BL/6 mice (The Jackson Lab- nology Associates) as described (45). Bound HRP and alkaline phospha-

oratory, Bar Harbor, ME) and F2 offspring were typed at Cr2 loci and for tase activities were revealed with 3-amino-ethyl-carbazole (Sigma) and Igh haplotype. The status of the Cr2 loci was determined by Southern blot naphthol AS-MX phosphate/fast blue BB base (Sigma), respectively. analysis of DNA from tail tissue (28) and was verified by flow cytometric analysis of CD21/CD35 expression on peripheral blood B cells (see be- Flow cytometric analyses low). The Igh haplotype was determined by flow cytometry of peripheral Expression of CD21/CD35, IgMa, and IgMb on peripheral blood B cells blood B cells (see below) and confirmed by ELISA of serum IgM. More was determined by flow cytometry. In brief, Х100 ␮l of blood was taken than 10 male and female F2 animals carrying at least one disrupted Cr2 b/b from the tail vein of individual mice; PBMCs were isolated from these allele on an Igh background were selected for breeding to generate the F3 samples over Lympholyte-M (Accurate Chemical and Scientific, West- homozygous and heterozygous knockout mice used in this study. Hetero- ϩ/Ϫ bury, NY) density gradients. After washes with PBS containing 2% FCS zygote F3 (Cr2 ) animals served as normal phenotype controls to ensure and 0.08% sodium azide, cells were incubated with anti-Fc␥RI/RII Ab that any background genetic effects were distributed without regard to the (PharMingen, San Diego, CA) to block FcR-mediated binding. PBMCs disrupted gene loci. C57BL/6 mice were also used as normal controls. All were then stained with FITC-labeled anti-CD21/35 (PharMingen), PE-con- mice were maintained under identical specific-pathogen-free conditions at jugated anti-IgMa (PharMingen), and biotinylated anti-IgMb (clone AF6- the Duke University Medical Center Vivarium (Durham, NC) and were 78) plus Tri-color- or Red 613-labeled streptavidin (Caltag, Burlingame, used at the age of 2–3 mo. CA, and Life Technologies, Gaithersburg, MD, respectively). Ags and immunization To quantify splenic GC reactions by flow cytometry (54), splenocyte suspensions were depleted of RBCs cells by incubation in 0.83% NH4Cl The succinic anhydride esters of NP (Genosys, Woodland, TX) were re- and washed as above. After blocking FcR-mediated binding, cells were acted with chicken gamma-globulin (CG) (Pel-Freeze Biological, Rogers, stained with FITC-labeled GL-7 (PharMingen), PE-conjugated anti-B220 AR) or BSA (United States Biochemical, Cleveland, OH) as described (PharMingen), and 7-aminoactinomycin D (7-AAD; Molecular Probes, Eu- (45). Hapten substitution ratios were determined spectrophotometrically. gene, OR) for 30 min. 4524 HUMORAL RESPONSES IN Cr2Ϫ/Ϫ MICE

FIGURE 1. Serum Ab responses (a–c) and GC reactions (d) to different doses of Ag in Cr2Ϫ/Ϫ (filled bars), Cr2ϩ/Ϫ (stip- pled bars), and C57BL/6 (open bars) mice. Mice were immunized with 5–50 ␮gof

NP12-CG precipitated in alum and sacri- ficed at day 10 postimmunization. NP-spe- cific serum IgM (a), IgG1 (b), and ␭1(c) Downloaded from Ab responses were quantified by ELISA. The GC reaction was evaluated (d)inin- dividual mice by immunohistochemical staining with PNA. GCs were enumerated by microscopy in two distant sections for each spleen. Each bar represents the mean (ϮSEM) numbers of GCs per section from http://www.jimmunol.org/ three to nine mice. by guest on September 25, 2021

Amplification and sequencing of VDJ rearrangements from cloning, as was done for microdissected GC B cells. Transformed bacterial single GCs clones were subject to colony hybridization using a VH186.2-specific probe and a framework 1-binding probe that hybridizes to most members of the ϩ Cellular material (Х20 cells) was microdissected from individual ␭1 V 186.2 and V3 V subgroups (45, 55). The hybridization was repeated to ϩ H H PNA GCs identified by immunohistochemistry (see above) and digested screen 170–255 bacterial clones from each group. with proteinase K as described (42–44). After heat inactivation, the reac- tion mixture was subjected to two rounds of PCR amplification (42–44) with Pfu polymerase (Strategene, La Jolla, CA). Briefly, the first round of Results 40 amplification cycles used primers homologous to the genomic DNA 5Ј Dose dependence of serum Ab responses in Cr2Ϫ/Ϫ mice of the transcription start site of the V186.2 VH gene segment and to a Ϫ/Ϫ region in the JH2–JH3 intron. Two-microliter aliquots of this reaction mix- Cr2 and C57BL/6 mice were immunized i.p. with 5, 10, 20, or ture were reamplified for another 40 cycles using a second set of nested 50 ␮g of NP-CG in alum. NP-specific serum IgM and IgG1 re- primers complementary to the first 20 nucleotides of the V186.2 and to the sponses were determined by ELISA 10 days later (Fig. 1, a and b). terminal portion of J 2 segment. Both 5Ј primers are complementary to V H H With alum adjuvant, Cr2Ϫ/Ϫ mice mounted significant IgM and gene segments in the V186.2 and V3 subgroups of the J558 VH family. Amplified VDJ DNA was purified with the QIAquick PCR Purification kit IgG1 responses at Ag doses as low as 5 ␮g. However, the re- (Qiagen, Valencia, CA), digested with the BamHI and PstI restriction en- sponses of Cr2Ϫ/Ϫ mice were substantially less than those of zymes (New England Biolabs, Beverly, MA), and cloned into pBluescript ϩ/Ϫ ␮ ϩ Cr2 and C57BL/6 controls; 5 g NP-CG elicited 3- and 4-fold SK (Strategene). Plasmid DNA from randomly picked clones of bacterial Ϫ/Ϫ transformants was sequenced using an ABI 377 PRISM DNA sequencer lower IgM and IgG1 responses, respectively, in Cr2 mice than with the Perkin-Elmer Dye Terminator Sequencing system (PE Biosys- in either control group (Fig. 1, a and b). Increasing Ag dose to 50 Ϫ Ϫ tems, Foster City, CA). ␮g reduced the deficit in Cr2 / mice so that IgM Ab levels be- Ϫ/Ϫ Estimating frequencies of V 186.2 rearrangements in naive B cells came equivalent, and IgG1 rose in Cr2 to half of that observed H in C57BL/6 and Cr2ϩ/Ϫ mice. In Cr2Ϫ/Ϫ mice, a 10-fold increase ϩ/Ϫ Ϫ/Ϫ Splenic cells from three C57BL/6, Cr2 ,orCr2 mice (8 wk old) in Ag (from 5 to 50 ␮g) resulted in a 3-fold increase in IgM were pooled and stained with monoclonal anti-IgD-FITC (PharMingen), ϩ/Ϫ anti-IgMb-PE (PharMingen), and Ls136-biotin plus Red 613-labeled responses and a 7-fold increase in IgG1 responses. In Cr2 and streptavidin (Life Technologies). IgMϩIgDϩ␭1ϩ cells were isolated by C57BL/6 controls, Ag-specific IgM titers did not increase and FACS and subjected to proteinase K digestion, PCR amplification, and IgG1 levels increased about 4-fold over the same dose range (Fig. The Journal of Immunology 4525 Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

Ϫ/Ϫ ␮ FIGURE 2. Accelerated decay of Ab responses in Cr2 mice. Mice were immunized with 50 gofNP12-CG precipitated in alum. The serum levels of NP-specific IgM (a; days 3–30 postimmunization) and IgG1 (b; days 3–112 postimmunization) in individual mice were followed. Each line (F, Cr2Ϫ/Ϫ; E, C57BL/6) represents the kinetics of an individual mouse. Linear regression analysis of mean logarithmic values (ϮSEM) of IgG1 levels (c)in Cr2Ϫ/Ϫ(f), Cr2ϩ/Ϫ (ࡗ), and C57BL/6(Ⅺ) mice (four to eight in each group) from day 16 to day 112 after immunization reveals distinct rates of Ab decay.

1, a and b). Similar Ag-dose responses were observed for total ␭1ϩ 6. After day 6 of the response, serum IgM responses in the ma- anti-NP serum Ab in Cr2Ϫ/Ϫ and control mice (Fig. 1c). jority of Cr2Ϫ/Ϫ mice decreased more rapidly than controls did (day 10, 9.9 Ϯ 2.1 ␮g/ml; day 30, 0.3 Ϯ 0.1 ␮g/ml) (Fig. 2a). In Ϫ/Ϫ Serum Ab levels fall more rapidly in Cr2 mice Cr2Ϫ/Ϫ, Cr2ϩ/Ϫ, and C57BL/6 mice, serum anti-NP IgG1 in- Single immunizations with NP-CG elicit long-lasting serum Ab creased about 15-fold from day 6 to a peak at day 10. By day 16, that is produced by AFCs in the bone marrow (48). To follow the serum anti-NP IgG1 levels were about 60% of day 10 values in kinetics of Ab responses in the absence of CD21/CD35, Cr2Ϫ/Ϫ, both knockout and control animals. At later times, the rate of loss Ϫ Ϫ Cr2ϩ/Ϫ, and C57BL/6 mice were immunized with 50 ␮g NP-CG of specific IgG1 from the serum diverged between Cr2 / and in alum and bled at days 3, 6, 10, 16, 30, 50, 70, 90, and 112 C57BL/6 or Cr2ϩ/Ϫ controls (Fig. 2, b and c). Linear regression (Cr2ϩ/Ϫ at days 6, 10, 30, 70, and 112) after immunization. NP- analysis on the mean logarithmic values of NP-specific serum specific serum IgM and IgG1 Ab was below 0.13 ␮g/ml at day 3 IgG1 concentration (days 16–112) indicated that Ab responses de- postimmunization in both Cr2Ϫ/Ϫ and C57BL/6 mice. In Cr2ϩ/Ϫ cayed twice as fast in Cr2Ϫ/Ϫ mice (y ϭϪ0.012x ϩ 2.726, r ϭ and C57BL/6 controls, IgM responses were first observed at day 6 0.993) as in C57BL/6 controls (y ϭϪ0.006x ϩ 3.157, r ϭ 0.964) (5.5 Ϯ 0.9 and 5.4 Ϯ 0.7 ␮g/ml (x¯ Ϯ SEM), respectively), peaked and Cr2ϩ/Ϫ mice (y ϭϪ0.006x ϩ 3.089, r ϭ 0.956). The differ- at day 10 (14.6 Ϯ 1.3 and 18 Ϯ 1.4 ␮g/ml, respectively), and then ence of slopes between C57BL/6 and Cr2Ϫ/Ϫ mice is statistically decayed about 10-fold by day 30 (1.8 Ϯ 0.1 and 1.4 Ϯ 0.10 ␮g/ml, significant ( p Ͻ 0.001, Student’s t test for the homogeneity of respectively). Cr2Ϫ/Ϫ mice had NP-specific IgM levels (8.3 Ϯ 2.1 regression; Fig. 2c). In addition, regression analysis for Ab decay ␮g/ml) not less than those of Cr2ϩ/Ϫ and C57BL/6 controls at day in individual mice demonstrated that the decay rates in C57BL/6 4526 HUMORAL RESPONSES IN Cr2Ϫ/Ϫ MICE Downloaded from

FIGURE 3. Frequencies of NP-specific, IgG1 AFCs in the bone marrow of C57BL/6 vs Cr2Ϫ/Ϫ mice (a) and of C57BL/6 vs Cr2ϩ/Ϫ mice (b) early and Ϯ ϩ late after immunization. Complete bars represent the mean frequencies ( SEM) of AFCs detected by ELISPOT on NP20-BSA substrates (low- Ͻ ء high-affinity cells), and hatched regions represent AFCs detected on NP6-BSA (high-affinity cells); data are from groups of four mice. , p 0.05, Student’s t test. http://www.jimmunol.org/ and Cr2ϩ/Ϫ (slopes 0.006 Ϯ 0.001, both) differ significantly from Affinity maturation in Cr2ϩ/Ϫ and C57BL/6 mice was virtually those in Cr2Ϫ/Ϫ animals (0.012 Ϯ 0.002; p Ͻ 0.05, Student’s t identical (Fig. 4a). test) but not from each other. These decay rates are far in excess In an independent experiment, groups of Cr2Ϫ/Ϫ and C57BL/6 of the half-life of passively injected IgG1 (56). Thus, the more mice (n ϭ 4–8) were killed at different times after immunization, rapid loss of serum IgG1 in Cr2Ϫ/Ϫ animals in comparison to and IgG Ab was purified from the pooled sera. Fluorescence ϩ/Ϫ Cr2 and C57BL/6 mice likely reflects declining Ab production quenching assays were performed to measure the average Ka for rather than altered Ig catabolism. NP-caproate (50, 52, 53). Consistent with the relative affinity es-

timates by ELISA (Fig. 4a), the Ka of serum IgG for NP at day 10 by guest on September 25, 2021 Fewer NP-specific Ab-forming cells in the bone marrow of postimmunization were virtually identical in Cr2Ϫ/Ϫ and C57BL/6 Ϫ/Ϫ Cr2 mice late after immunization Ϫ/Ϫ mice. By 28 days after immunization, the Ka of Ab from Cr2 Persistent IgG1 serum Ab is maintained by long-lived AFCs that mice was slightly higher than that of controls, but by day 59, the Ϫ/Ϫ Ն reside in the bone marrow (46–49). These bone marrow AFCs are Ka of Ab from Cr2 mice was 4-fold greater than that of Ϫ/Ϫ ϩ/Ϫ present in Cr2 , Cr2 , and C57BL/6 mice in comparable C57BL/6 mice (Fig. 4b). These measured Ka confirm the ELISA numbers early in the primary response (day 14), but 62 days after affinity estimates of individual sera and demonstrate enhanced af- immunization, the bone marrow of Cr2Ϫ/Ϫ mice contains only finity maturation in the absence of CD21/35. about half as many NP-specific, IgG1 AFCs as do controls (Fig. 3). Ϫ/Ϫ The difference late in the response is due less to declining numbers Cr2 mice exhibit reduced GC reactions at low doses of Ag Ϫ Ϫ of AFCs in Cr2 / mice than in the failure of this compartment to CD21/CD35 affect B cell persistence in GCs (37). Thus, we eval- ϩ Ϫ expand. In Cr2 / and C57BL/6 mice, frequencies of total (Fig. 3) uated the ability of Cr2Ϫ/Ϫ mice to form GCs after immunization and high-affinity (hatched bars) NP-specific bone marrow AFCs with different doses of Ag. Cr2Ϫ/Ϫ mice injected with 5 ␮gof increase ϳ2-fold to day 64. In contrast, the frequencies of all NP- NP-CG have ϳ5-fold fewer GCs than do C57BL/6 or Cr2ϩ/Ϫ Ϫ Ϫ specific AFCs in Cr2 / mice (Fig. 3a) slightly decline over the controls (Fig. 1d); the size of GCs in Cr2Ϫ/Ϫ animals was also same interval, falling from 57 Ϯ 7to39Ϯ 14 AFCs per 106 bone reduced in comparison to controls (data not shown). Larger doses marrow cells. Virtually all of this decline was confined to the low- of Ag increased both the size and number of GCs in Cr2Ϫ/Ϫ mice er-affinity compartment of AFCs; frequencies of high-affinity so that at 50 ␮g of NP-CG in alum, GC reactions were comparable Ϫ Ϫ AFCs in the bone marrow of Cr2 / mice (Fig. 3a) remained in Cr2Ϫ/Ϫ and C57BL/6 mice (Fig. 1d). Surprisingly, and in con- unchanged from day 14 to day 62 of the response. In contrast, bone trast to the Ag-dose-dependent increases of serum Ab (Fig. 1, a–c), marrow AFC numbers and affinities were virtually identical in the GC reaction of C57BL/6 mice was reduced at a dose of 50 ␮g ϩ Ϫ C57BL/6 and Cr2 / mice (Fig. 3b). (Fig. 1d). This reduction was significant in the minimization of Ϫ/Ϫ Ϫ/Ϫ differences between the GC responses of Cr2 and control mice Enhanced affinity maturation in serum Ab of Cr2 mice at high Ag doses. The affinity of anti-NP serum Ab was estimated by calculating the We studied the kinetics of the GC reaction in Cr2Ϫ/Ϫ mice by ϩ ϩ ratios of Ab bound to NP6-BSA vs NP20-BSA (51). From day 6 to flow cytometric quantification of the B220 GL7 population of Ϫ/Ϫ ␮ day 10 postimmunization, NP6/NP20 binding ratios for Cr2 , GC B cells (54). At an Ag dose of 50 g, the kinetics of GC Cr2ϩ/Ϫ, and C57BL/6 mice increased 4- to 5-fold to reach similar reactions in Cr2Ϫ/Ϫ mice and C57BL/6 controls were very similar, averages by day 16. However, from day 30 to day 112, average with peak responses at day 10 and then monotonic declines to Ϫ/Ϫ NP6/NP20 binding ratios for Cr2 mice were always greater than frequencies slightly above naive levels by day 28 (Fig. 5). This those of Cr2ϩ/Ϫ and C57BL/6 mice, indicating increased affinity. pattern of response is typical of normal responses (48). The Journal of Immunology 4527

FIGURE 4. Enhanced affinity maturation of se- rum Ab in the absence of CD21/CD35. Serum samples (as in Fig. 2) from Cr2Ϫ/Ϫ (f), Cr2ϩ/Ϫ (ࡗ), and C57BL/6 (Ⅺ) mice were titrated in ELISA (a) for the amount of high-affinity (binds

NP6-BSA) vs total (binds NP20-BSA) NP-reactive IgG1. NP6/NP20 ratios were calculated and aver- aged for each group at each time point. Data rep- resent the mean ratios (ϮSEM) exhibited by four to eight mice. In an independent experiment (b), cohorts of four to eight mice were immunized ␮ with 50 gofNP12-CG in alum and sacrificed at various times after immunization; sera from mice in each group were pooled and IgG was purified from each pool. Fluorescence quenching assays

were performed to measure the equilibrium Ka of

the purified IgG for NP-caproate (48). Downloaded from http://www.jimmunol.org/ VDJ hypermutation in the absence of CD21/CD35 VH gene segment, and 70% (38/54) contained DFL16.1 (Table I). Ig heavy chain gene rearrangements present in GC B cells from The YYGS or YYGN motifs optimal for NP-binding were present Cr2Ϫ/Ϫ and C57BL/6 mice were amplified by a specific PCR and at the V-to-D junction in 31 (57%) of these VDJ fragments, and sequenced (43); 79 and 61 VDJ sequences, respectively, were ob- each V186.2 gene segment contained an average of 2.6 mutations (Fig. 5 and Table I). To our surprise, fewer than half (26; 45%) of tained from each group. Ratios of productive:nonproductive rear- Ϫ/Ϫ rangements were significantly lower in Cr2Ϫ/Ϫ mice (58:21) than the 58 productive rearrangements sequenced from Cr2 mice in C57BL/6 controls (54:7). Consistent with previous reports (41– contained V186.2, and exactly half (29/58) used DFL16.1. None- 43, 48), in the 54 productive VDJ rearrangements sequenced from theless, YYG(S/N) junctional motifs were slightly more common than in controls (69% vs 57%; Table I), suggesting active clonal C57BL/6 controls, a large majority (46/54; 85%) used the V186.2 by guest on September 25, 2021 selection in the absence of CD21/CD35. Somatic mutation was not reduced in Cr2Ϫ/Ϫ mice; average mutation numbers in knockout mice were somewhat higher than in C57BL/6 mice (3.8 vs 2.6; p Ͻ 0.05, Student’s t test) (Fig. 5 and Table I).

Table I. Somatic genetics of ␭1ϩ GC B cells from Cr2Ϫ/Ϫ and C57BL/6 micea

Cr2Ϫ/Ϫ C57BL/6

VDJ Rearrangements Containing:b V186.2 (% of total) 26 (45) 46 (85)

Other VH (% of total) 32 (55) 8 (15) DFL16.1 (%) 50 70 YYG(S/N) CDR3 motif (%) 69 57

Mutations Average no. in V186.2c 3.8 2.6 R/S ratiosd CDR1 (14.0/1)e Ͼ2.0/1f Ͼ1.0/1f CDR2 (4.2/1) Ͼ7.0/1f 4.0 /1 Framework (3.1/1) 4.0 /1 1.4 /1 a ␮ Mice were immunized with 50 gofNP12-CG in alum and sacrificed at day 10 postimmunization. Data are a summary of 10 GCs from three Cr2Ϫ/Ϫ mice and eight GCs from two C57BL/6 mice. The average number of unique CDR3 sequences re- covered from individual GCs was similar for both Cr2Ϫ/Ϫ and C57BL/6 mice at Ϫ/Ϫ f Ⅺ FIGURE 5. Kinetics of GC reactions in Cr2 ( ) and C57BL/6 ( ) 1.3 Ϯ 0.6 and 1.1 Ϯ 0.3, respectively. ␮ b mice are similar. Mice were immunized with 50 gofNP12-CG in alum Includes only productive rearrangements. and sacrificed at different time points after immunization. Spleen cells were c Includes both productive and nonproductive rearrangements. d analyzed with flow cytometry for the numbers of B220ϩGL-7ϩ cells. Data Replacement/silent (R/S) ratios for unique mutations in productive VDJ rear- Ϯ rangements containing V186.2. represent the mean frequencies ( SEM) of GC B cells present in three to e R/S ratio of V186.2 given random mutagenesis. eight mice. f No silent mutation was observed. 4528 HUMORAL RESPONSES IN Cr2Ϫ/Ϫ MICE Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 6. Representative heavy-chain VDJ DNA sequences recovered from day 10 ␭1ϩ GCs of Cr2Ϫ/Ϫ (upper group) and C57BL/6 (lower group) mice. A total of 79 and 61 VDJ sequences were recovered from three Cr2Ϫ/Ϫ and two C57BL/6 mice, respectively. All unique VDJ sequences from both groups are presented. Each sequence is identified by a number (01–10) and lowercase letter (a–d) that follow the capitals “AB.” The number identifies the GC origin, and the letter indicates the specific VDJ sequence. Sequences that share VDJ junctional regions presumably share clonal origins and are presented to show shared and unique mutations. VDJ rearangements were compared with germline VH186.2 and JH2 sequences; dashes indicate identity with these reference sequences. Sequences that overlap the 5Ј and 3Ј PCR primers are not shown. Underlined nucleotides are not found in homologous germline gene sequences available from GenBank, and may represent somatic mutation. “I” in 61AB03c between codons 52A and 53 indicates an insertion of 27 nucleotides (CTT GAG TGG ATT GGA AAT ATT AAT CCT), which is a repeat of the sequence from codons 45–52A. It is unknown whether this insertion resulted from an in vivo or in vitro process. Sequence data are available from GenBank under accession nos. AF146302-17 (C57BL/6) and AF146318-38 (Cr2Ϫ/Ϫ).

ϩ/Ϫ Lower numbers of VDJ rearrangements containing VH186.2 in Cr2 , and C57BL/6 mice by PCR amplification and cloning. GCs of Cr2 knockouts could reflect alterations of the GC reaction Virtually all cloned rearrangements hybridized to a framework re- or the naive repertoire. Thus, we recovered VDJ rearrangements gion probe that identifies the V3 and V186.2 VH subfamilies (98– from naive (IgMϩIgDϩ␭1ϩ) splenic B cells from Cr2Ϫ/Ϫ, 100%, C57BL/6; 91–99%, Cr2ϩ/Ϫ; 98–100%, Cr2Ϫ/Ϫ). However, The Journal of Immunology 4529 Downloaded from http://www.jimmunol.org/ by guest on September 25, 2021

FIGURE 6. (continued)

within this hybridizing set, VH186.2 frequencies in C57BL/6 and of C3b, which is rapidly converted to C3d on the pathogens’ sur- Cr2ϩ/Ϫ mice (29–32% and 27–31%, respectively) were 2- to face (19, 58). C3d-decorated Ag is thought to facilitate binding/ Ϫ Ϫ 3-fold higher than that of Cr2 / mice (11–14%). Thus, the retention by FDCs and the recruitment of the CD21 and coreceptor Ϫ Ϫ Cr2 / genotype modifies specific VDJ frequencies within the complex into close association with Ag-bound BCRs (18, 19, 57). pool of naive, splenic B cells. We studied the primary B cell responses to the NP hapten to dissect the role(s) of CD21/CD35 on serum Ab responses and the Discussion GC reaction. As shown in Fig. 1, reduced but significant serum Ϫ/Ϫ B cell responses to T-dependent Ags are mediated by the Ag- IgM and IgG1 responses were elicited in Cr2 mice even by low specific BCR and a nonspecific coreceptor comprising CD21, doses of NP-CG in alum. This impairment was dose-dependent; CD19, and CD81 (18, 19, 57). The complement system, a major increasing amounts of Ag gradually improved serum IgG1 Ab ti- Ϫ Ϫ ϩ Ϫ component of innate immunity, responds to pathogens by activa- ters in Cr2 / from ϳ25% to 50% of that in Cr2 / and C57BL/6 tion of the classic or alternative pathways; both activation path- controls (Fig. 1b). In a previous study, Cr2Ϫ/Ϫ mice failed to gen- ways converge at C3 with the consequence of covalent deposition erate Ab responses to low numbers (3 ϫ 107) of bacteriophage 4530 HUMORAL RESPONSES IN Cr2Ϫ/Ϫ MICE

⌽X174 in the absence of adjuvant but mounted modest responses could be more stringent in CD21/CD35-deficient mice. In fact, to a 10-fold increase in phage numbers (28). Similarly, high doses after day 30 of the response, the average affinity of serum Ab in of Ag mitigated the suppression of humoral immune responses by Cr2Ϫ/Ϫ mice was significantly higher than that present in Cr2ϩ/Ϫ anti-CD21/CD35 Ab treatment (14, 15). These observations con- and C57BL/6 controls (Fig. 4). At what site(s) is affinity matura- trast with a report (29) in which a 200-fold increase of Ag (SRBC) tion augmented in CD21/CD35-deficient mice? The activation of Ϫ Ϫ did not improve deficient primary IgG Ab responses in Cr2 / extra-follicular B cells in the spleens of Cr2Ϫ/Ϫ mice appears nor- Ϫ Ϫ mice. Differences in responsiveness by Cr2 / mice correlate with mal; initial Ab affinities are identical in Cr2Ϫ/Ϫ and control mice the use of inflammatory adjuvants, agents known to mitigate other (Fig. 4), and early GC B cell populations in knockout animals were humoral deficiencies. Generally, our findings indicate that com- even more diverse than in controls (Fig. 5 and Table I). These plement and its receptors enhance humoral immune responses to findings suggest that responding B cells in Cr2Ϫ/Ϫ mice expressed physiological (low) amounts of noninflammatory Ags, which is a broad range of BCR affinities and do not support predictions that consistent with the concept that the CD21/CD19/CD81 coreceptor limiting activation thresholds in the absence of CD21 restrict lowers the threshold for B cell activation (18, 19, 57). clonal diversity in immune responses. Ag dose also affects the GC reaction. Consistent with earlier Unusually intense selection may have taken place in the GCs of Ϫ/Ϫ findings (28), the GC reaction of Cr2 mice was only about 35% Cr2Ϫ/Ϫ mice. The average number of unique CDR3 sequences ϩ/Ϫ of that in Cr2 and C57BL/6 controls at lower Ag doses (Fig. recovered from individual GCs was Ϸ1 for both knockout and ␮ 1d). However, after immunization with 50 g NP-CG in alum, BL/6 mice. Ratios of replacement to silent mutations in the CDR2 Ϫ/Ϫ Cr2 and C57BL/6 mice generated comparable numbers of of VDJ rearrangements were higher in Cr2Ϫ/Ϫ mice than in those splenic GCs and GC B cells (Figs. 1d and 5). The GC reaction in recovered from C57BL/6 controls (Table I). However, given the Downloaded from Cr2Ϫ/Ϫ mice given this amount of Ag is kinetically similar to decreased representation of canonical VH186.2/JH2 rearrange- wild-type responses (Fig. 5), but equalization of responses de- ments present in Cr2Ϫ/Ϫ GCs (Table I), it is difficult to gauge pended as much on a diminished GC response in C57BL/6 mice as Ϫ/Ϫ selection intensity. Reduction in VH186.2 use in the GC B cells of it did on increased GC reactions in Cr2 animals (Fig. 1d). Cr2Ϫ/Ϫ mice may simply reflect its relative rarity in naive reper- Decreased GC but not Ab responses in control animals are unex- toire. VH186.2/JH2 B cells are about 2- to 3-fold less frequent in pected outcomes to higher Ag doses that merit further investiga- ␭1ϩ, naive splenic B cells from Cr2Ϫ/Ϫ mice than those from http://www.jimmunol.org/ tion. Reduced GC responses to 50 ␮g NP-CG in alum were not ϩ Ϫ C57BL/6 and Cr2 / mice. How CD21/CD35 alters the naive observed in mice deficient for CD21/CD35. Perhaps large quanti- repertoire is unknown, but changes could reflect disturbed selec- ties of Ag on FDCs synergize with coreceptor signals to limit the tion during B cell maturation (68). On the other hand, the remark- GC reaction (59–61). ably high frequency of nonproductive VDJ rearrangements present VDJ hypermutation is intact in Cr2Ϫ/Ϫ mice. By day 10 postim- in GCs of CD21/35-deficient mice might result from receptor re- munization, significantly more VDJ mutations were present in the Ϫ Ϫ vision in Ag-reactive cells (69). Nonetheless, B-cell clones ex- GC B cells of Cr2 / mice than were those from C57BL/6 mice pressing VH186.2/JH2 rearrangements are enriched in GCs of (Fig. 6 and Table I). This increase possibly reflects more stringent Ϫ/Ϫ

Cr2 mice to the same extent as for C57BL/6 mice. We do not by guest on September 25, 2021 selection. Prior work has shown that B cells lacking CD21/CD35 know the affinity of noncanonical BCRs in Cr2Ϫ/Ϫ GC B cells. do not persist in wild-type GCs, even if they bear high-affinity Despite their presence, the affinity of early serum Ab (day 16) is BCRs (37). However, in the absence of normal B cell competitors, Ϫ/Ϫ Ϫ/Ϫ comparable in both C57BL/6 and Cr2 mice (Fig. 4), suggesting Cr2 B cells respond to abundant Ag/adjuvant with GCs that are Ϫ/Ϫ quantitatively and qualitatively comparable to those of control that not all noncanonical clones in Cr2 mice have low affinities mice (Figs. 1d and 5, Table I, and Refs. 28 and 29). Thus, the to NP, even though noncanonical VDJ rearrangements from C57BL/6 GCs showed generally lower affinities (50). However, advantage that is provided by CD21/CD35 to GC B cells must be Ϫ/Ϫ relative rather than absolute. Amelioration of the deficient GC re- even if the starting affinities of noncanonical clones in Cr2 action of Cr2Ϫ/Ϫmice by adjuvant and abundant Ag supports the mice are low, they could achieve higher affinities by somatic mu- role of CD21/CD35 in Ag retention by FDC/stromal cells (34). tation and selection in GCs (50). Selection is evident in the early Ϫ/Ϫ One function of the alum adjuvant is to serve as an Ag depot (62). (day 10) GCs of Cr2 mice (Table I and Fig. 6), but the phe- Our observations raise the question of how Ag is presented in notype of higher Ab affinity is not present until 30 days after im- GCs without CD21/CD35. In GCs, Ags are trapped primarily on munization (Fig. 4). At this late stage of the B cell response, the the FDC surface in the form of Ag-Ab-complement complexes GC reaction is ending/ended, but affinity-driven selection still (20–23). This Ag resource initiates and sustains the GC reaction. shapes the long-lived AFC population of the bone marrow (48). Ϫ/Ϫ As Ag levels fall in GCs, increasingly intense selection for higher- The persistence of serum Ab is reduced in Cr2 mice (Fig. 2), affinity clones is believed to occur (38, 63, 64). Ag in GCs was even when elicited by a dose of Ag that produces GC responses ϩ/Ϫ thought to be retained on FDCs by Fc␥R and CD21/CD35 (20– similar to those in controls (Fig. 5). In comparison to Cr2 and Ϫ/Ϫ 23). However, targeted disruption of the common FcR ␥-chain C57BL/6 controls, Cr2 mice initially generate similar levels of gene increased rather than reduced Ag levels on FDCs (65). Sim- NP-specific serum IgM and 2-fold less IgG1 in response to 50 ␮g ilarly, mice deficient in Fc␥RIIb have elevated Ab responses (66). of NP-CG (Fig. 1, a and b). However, Ab titers decay twice as Ϫ Ϫ In contrast, disruption of the Cr2 loci dramatically decreased Ag rapidly in Cr2 / mice (Fig. 2). Associated with the accelerated retention by FDCs (19, 34), indicating that CD21/CD35 play the loss of IgG1 Ab was enhanced affinity maturation (Fig. 4). In- primary role in holding Ag in GCs. In the absence of CD21/CD35, creased rates of Ab decay and affinity maturation mirrored pref- FDCs likely use the Fc␥RIIb molecule as an alternative receptor erential loss of bone marrow AFCs secreting lower-affinity IgG1 for Ag retention. Fc␥RIIb is expressed at lower levels than CD21/ Ab (Fig. 3). CD35 on FDCs and is thought to not bind IgM (67); these limi- If CD21/CD35 play important roles in the GC reaction, why tations in potential efficiency may be responsible for the strong does affinity maturation of serum Ab appear to increase in their dose dependence of the GC reaction in Cr2Ϫ/Ϫ animals (Fig. 1d). absence? As the kinetics of the GC reaction in Cr2Ϫ/Ϫ and In a milieu of reduced Ag on FDCs/stromal cells and in the C57BL/6 mice are similar at high Ag doses (Fig. 5), we propose absence of complete coreceptors, affinity-driven clonal selection that enhanced affinity maturation in Cr2Ϫ/Ϫ mice reflects selection The Journal of Immunology 4531 of the AFCs in bone marrow or of their precursors. Although adop- 5. Pepys, M. B. 1974. Role of complement in induction of antibody production in tive transfer experiments indicate that bone marrow AFCs do not vivo: effect of cobra factor and other C3-reactive agents on thymus-dependent and thymus-independent antibody responses. J. Exp. Med. 140:126. depend on Ag for survival (49), we previously demonstrated that 6. Ellman, L., I. Green, F. Judge, and M. M. Frank. 1971. In vivo studies in C4- this population undergoes affinity maturation (Fig. 3) long after the deficient guinea pigs. J. Exp. Med. 134:162. GC reaction wanes (48). Thus, Ag plays a critical role in the fate 7. Ochs, H. D., R. J. Wedgwood, M. M. Frank, S. R. Heller, and S. W. Hosea. 1983. The role of complement in the induction of antibody responses. Clin. Exp. Im- of bone marrow AFCs, if only in their selection and not in their munol. 53:208. survival. In the absence of CD21/CD35, the bone marrow AFC 8. Bottger, E. C., S. Metzger, D. Bitter-Suermann, G. Stevenson, S. Kleindienst, and R. Burger. 1986. Impaired humoral immune response in complement C3-defi- population not only fails to expand but contracts by the preferential cient guinea pigs: absence of secondary antibody response. Eur. J. Immunol. loss of lower-affinity cells (Fig. 3a). This selective contraction co- 16:1231. incides with the conundrum of accelerated loss of serum IgG1 Ab 9. Bitter-Suermann, D., and R. Burger. 1986. Guinea pigs deficient in C2, C4, C3 or the C3a receptor. Prog. Allergy. 39:134. levels and increased Ab affinity (Fig. 4). 10. O’Neil, K. M., H. D. Ochs, S. R. Heller, L. C. Cork, J. M. Morris, and Where and how Ags act on the bone marrow AFCs are un- J. A. Winkelstein. 1988. Role of C3 in humoral immunity. Defective antibody known. Nonetheless, the accelerated loss of circulating Ab in production in C3-deficient dogs. J. Immunol. 140:1939. Ϫ/Ϫ 11. Jackson, C. G., H. D. Ochs, and R. J. Wedgwood. 1979. Immune response of a Cr2 mice (Fig. 2c) suggests that Ag depots influence bone patient with deficiency of the fourth component of complement and systemic marrow AFCs and implies that this Ag may be detected/retained lupus erythematosus. N. Engl. J. Med. 300:1124. through a mechanism(s) involving CD21/CD35. For example, in 12. Fischer, M. B., M. Ma, S. Goerg, X. Zhou, J. Xia, O. Finco, S. Han, G. Kelsoe, R. G. Howard, T. L. Rothstein, et al. 1996. Regulation of the B cell response to the absence of CD21/CD35, Ag deposition and coreceptor signal- T-dependent by classical pathway complement. J. Immunol. 157:549. ing could result in more stringent clonal selection for the bone 13. Dempsey, P. W., M. E. Allison, S. Akkaraju, C. C. Goodnow, and D. T. Fearon. marrow AFC compartment. This might lead to the selective loss of 1996. C3d of complement as a molecular adjuvant: bridging innate and acquired Downloaded from immunity. Science 271:348. lower-affinity clones and result in the enhanced affinity maturation 14. Heyman, B., E. J. Wiersma, and T. Kinoshita. 1990. In vivo inhibition of the of late IgG1 Ab. Indeed, a role for Ag dose on the rate of affinity antibody response by a complement receptor-specific . maturation has long been recognized (70). In the first two weeks of J. Exp. Med. 172:665. 15. Thyphronitis, G., T. Kinoshita, K. Inoue, J. E. Schweinle, G. C. Tsokos, primary Ab responses, Ab affinity is affected little by Ag dose, but E. S. Metcalf, F. D. Finkelman, and J. E. Balow. 1991. Modulation of mouse at later times, guinea pigs and rabbits given low doses of Ag pro- complement receptors 1 and 2 suppresses antibody responses in vivo. J. Immunol. 147:224. duce serum Abs with affinities manyfold higher than those of an- http://www.jimmunol.org/ 16. Gustavsson, S., T. Kinoshita, and B. Heyman. 1995. to murine com- imals given higher Ag doses (71, 72). Thus, the phenotype of plement receptor 1 and 2 can inhibit the antibody response in vivo without in- Cr2Ϫ/Ϫ mice may reflect reductions in available Ag. However, hibiting induction. J. Immunol. 154:6524. accelerated loss of serum Ab was not noted in earlier studies on Ag 17. Hebell, T., J. M. Ahearn, and D. T. Fearon. 1991. Suppression of the immune response by a soluble complement receptor of B lymphocytes. Science 254:102. dose and affinity maturation (70–72), and our analogy may well be 18. Fearon, D. T., and R. H. Carter. 1995. The CD19/CR2/TAPA-1 complex of B incomplete. An alternative explanation is that survival of bone lymphocytes: linking natural to acquired immunity. Annu. Rev. Immunol. 13:127. marrow AFCs generally declines in the absence of persistent Ag 19. Carroll, M. C. 1998. The role of complement and complement receptors in in- duction and regulation of immunity. Annu. Rev. Immunol. 16:545. Ϫ/Ϫ stores. Increased Ab affinity in Cr2 mice would then reflect the 20. Klaus, G. G., J. H. Humphrey, A. Kunkl, and D. W. Dongworth. 1980. The survival of higher-affinity clones generated by unusually intense follicular dendritic cell: its role in presentation in the generation of im- GC selection. munological memory. Immunol. Rev. 53:3. by guest on September 25, 2021 21. Mandel, T. E., R. P. Phipps, A. Abbot, and J. G. Tew. 1980. The follicular The effect of CD21/CD35 on the long-term Ab responses pre- dendritic cell: long term antigen retention during immunity. Immunol. Rev. 53:29. sents an interesting dilemma for humoral immunity to microbial 22. Schriever, F., and L. M. Nadler. 1992. The central role of follicular dendritic cells infection. Does the complement system recruit and maintain a in lymphoid tissues. Adv. Immunol. 51:243. 23. Tew, J. G., J. Wu, D. Qin, S. Helm, G. F. Burton, and A. K. Szakal. 1997. broad spectrum of specific AFC and memory B cells at the expense Follicular dendritic cells and presentation of antigen and costimulatory signals to of optimal affinity maturation? Does increased Ab concentration B cells. Immunol. Rev. 156:39. 24. Tuveson, D. A., J. M. Ahearn, A. K. Matsumoto, and D. T. Fearon. 1991. Mo- balance more modest affinity? In vitro, the affinity/avidity of an Ab lecular interactions of complement receptors on B lymphocytes: a CR1/CR2 com- correlates with its ability to fix complement (73) and neutralize plex distinct from the CR2/CD19 complex. J. Exp. Med. 173:1083. virus (74). However, in vivo tests of humoral protection from acute 25. Matsumoto, A. K., J. Kopicky-Burd, R. H. Carter, D. A. Tuveson, T. F. Tedder, and D. T. Fearon. 1991. Intersection of the complement and immune systems: a viral infection suggest that immunity by serum Ab depend solely signal transduction complex of the B lymphocyte-containing complement recep- on Ab concentration, given a minimal avidity threshold (74). It tor type 2 and CD19. J. Exp. Med. 173:55. would be interesting to test the effect of CD21/CD35 deficiency on 26. Bradbury, L. E., G. S. Kansas, S. Levy, R. L. Evans, and T. F. Tedder. 1992. The CD19/CD21 signal transducing complex of human B lymphocytes includes the the longevity of humoral protection from viral pathogens. Presum- target of antiproliferative antibody-1 and Leu-13 molecules. J. Immunol. 149: ably, the minimum serum concentration required for immunity 2841. would, over time, be compromised in Cr2Ϫ/Ϫ mice, but the resid- 27. Krop, I., A. L. Shaffer, D. T. Fearon, and M. S. Schlissel. 1996. The signaling activity of murine CD19 is regulated during cell development. J. Immunol. 157: ual Ab would exhibit enhanced (and compensating) affinity. 48. 28. Ahearn, J. M., M. B. Fischer, D. Croix, S. Goerg, M. Ma, J. Xia, X. Zhou, R. G. Howard, T. L. Rothstein, and M. C. Carroll. 1996. Disruption of the Cr2 Acknowledgments locus results in a reduction in B-1a cells and in an impaired B cell response to T-dependent antigen. Immunity 4:251. We thank Dr. T. F. Tedder for helpful discussions and review of the manu- 29. Molina, H., V. M. Holers, B. Li, Y. Fung, S. Mariathasan, J. Goellner, script and Lindsay Grey Cowell (North Carolina State University) for help J. Strauss-Schoenberger, R. W. Karr, and D. D. Chaplin. 1996. Markedly im- with statistical analysis. paired humoral immune response in mice deficient in complement receptors 1 and 2. Proc. Natl. Acad. Sci. USA 93:3357. 30. Croix, D. A., J. M. Ahearn, A. M. Rosengard, S. Han, G. Kelsoe, M. Ma, and References M. C. Carroll. 1996. Antibody response to a T-dependent antigen requires B cell expression of complement receptors. J. Exp. Med. 183:1857. 1. Lay, W. H., and V. Nussenzweig. 1968. Receptors for complement on leukocytes. 31. Fingeroth, J. D., M. A. Benedict, D. N. Levy, and J. L. Strominger. 1989. Iden- J. Exp. Med. 128:991. tification of murine complement receptor type 2. Proc. Natl. Acad. Sci. USA 2. Bianco, C., R. Patrick, and V. Nussenzweig. 1970. A population of lymphocytes 86:242. bearing a membrane receptor for antigen-antibody-complement complexes. I. 32. Kurtz, C. B., E. O’Toole, S. M. Christensen, and J. H. Weis. 1990. The murine Separation and characterization. J. Exp. Med. 132:702. complement receptor gene family. IV. Alternative splicing of Cr2 gene transcripts 3. Nussenzweig, V. 1974. Receptors for immune complexes on lymphocytes. Adv. predicts two distinct gene products that share homologous domains with both Immunol. 19:217. human CR2 and CR1. J. Immunol. 144:3581. 4. Pepys, M. B. 1972. Role of complement in induction of the allergic response. 33. Molina, H., T. Kinoshita, K. Inoue, J. C. Carel, and V. M. Holers. 1990. A Nature New Biol. 237:157. molecular and immunochemical characterization of mouse CR2: evidence for a 4532 HUMORAL RESPONSES IN Cr2Ϫ/Ϫ MICE

single gene model of mouse complement receptors 1 and 2. J. Immunol. 145: 53. Jones, P. T., P. H. Dear, J. Foote, M. S. Neuberger, and G. Winter. 1986. Re- 2974. placing the complementarity-determining regions in a human antibody with those 34. Fang, Y., C. Xu, Y. X. Fu, V. M. Holers, and H. Molina. 1998. Expression of from a mouse. Nature 321:522. complement receptors 1 and 2 on follicular dendritic cells is necessary for the 54. Han, S., B. Zheng, D. G. Schatz, E. Spanopoulou, and G. Kelsoe. 1996. Neoteny generation of a strong antigen-specific IgG response. J. Immunol. 160:5273. in lymphocytes: Rag1 and Rag2 expression in germinal center B cells. Science 35. Bonnefoy, J. Y., S. Henchoz, D. Hardie, M. J. Holder, and J. Gordon. 1993. A 274:2094. subset of anti-CD21 antibodies promote the rescue of germinal center B cells 55. Gu, H., D. Tarlinton, W. Muller, K. Rajewsky, and I. Forster. 1991. Most pe- from apoptosis. Eur. J. Immunol. 23:969. ripheral B cells in mice are ligand selected. J. Exp. Med. 173:1357. 36. Roberts, T., and E. C. Snow. 1999. Cutting edge: recruitment of the CD19/CD21 56. Vieira, P., and K. Rajewsky. 1988. The half-lives of serum immunoglobulins in coreceptor to B cell antigen receptor is required for antigen-mediated expression adult mice. Eur. J. Immunol. 18:313. of Bcl-2 by resting and cycling hen egg lysozyme transgenic B cells. J. Immunol. 57. Tedder, T. F., M. Inaoki, and S. Sato. 1997. The CD19-CD21 complex regulates 162:4377. signal transduction thresholds governing humoral immunity and . 37. Fischer, M. B., S. Goerg, L. Shen, A. P. Prodeus, C. C. Goodnow, G. Kelsoe, and Immunity 6:107. M. C. Carroll. 1998. Dependence of germinal center B cells on expression of 58. Fearon, D. T. 1998. The complement system and adaptive immunity. Semin. CD21/CD35 for survival. Science 280:582. Immunol. 10:355. 38. Kelsoe, G. 1995. In situ studies of the germinal center reaction. Adv. Immunol. 59. Pulendran, B., G. Kannourakis, S. Nouri, K. G. Smith, and G. J. Nossal. 1995. 60:267. Soluble antigen can cause enhanced apoptosis of germinal-centre B cells. Nature 39. Makela, O., and K. Karjalainen. 1977. Inherited immunoglobulin of the 375:331. mouse. Immunol. Rev. 34:119. 60. Shokat, K. M., and C. C. Goodnow. 1995. Antigen-induced B-cell death and 40. Reth, M., G. J. Hammerling, and K. Rajewsky. 1978. Analysis of the repertoire elimination during germinal-centre immune responses. Nature 375:334. of anti-NP antibodies in C57BL/6 mice by cell fusion. I. Characterization of 61. Han, S., B. Zheng, J. Dal Porto, and G. Kelsoe. 1995. In situ studies of the primary antibody families in the primary and hyperimmune response. Eur. J. Immunol. immune response to (4-hydroxy-3-nitrophenyl)acetyl. IV. Affinity-dependent, anti- 8:393. gen-driven B cell apoptosis in germinal centers as a mechanism for maintaining 41. Bothwell, A. L., M. Paskind, M. Reth, T. Imanishi-Kari, K. Rajewsky, and self-tolerance. J. Exp. Med. 182:1635. D. Baltimore. 1981. Heavy chain variable region contribution to the NPb family

␥ 62. Gupta, R. K., B. E. Rost, E. Relyveld, and G. R. Siber. 1995. Adjuvant properties Downloaded from of antibodies: somatic mutation evident in a 2a variable region. Cell 24:625. of aluminum and calcium compounds. Pharm. Biotechnol. 6:229. 42. Jacob, J., and G. Kelsoe. 1992. In situ studies of the primary immune response to 63. Kelsoe, G. 1996. Life and death in germinal centers (redux). Immunity 4:107. (4-hydroxy-3-nitrophenyl)acetyl. II. A common clonal origin for periarteriolar 64. MacLennan, I. C. 1994. Germinal centers. Annu. Rev. Immunol. 12:117. lymphoid sheath-associated foci and germinal centers. J. Exp. Med. 176:679. 65. Vora, K. A., J. V. Ravetch, and T. Manser. 1997. Amplified follicular immune 43. Jacob, J., J. Przylepa, C. Miller, and G. Kelsoe. 1993. In situ studies of the ␥ primary immune response to (4-hydroxy-3-nitrophenyl)acetyl. III. The kinetics of complex deposition in mice lacking the Fc receptor -chain does not alter mat- V region mutation and selection in germinal center B cells. J. Exp. Med. 178: uration of the B cell response. J. Immunol. 159:2116. 1293. 66. Takai, T., M. Ono, M. Hikida, H. Ohmori, and J. V. Ravetch. 1996. Augmented humoral and anaphylactic responses in Fc␥RII-deficient mice. Nature 379:346.

44. Jacob, J., G. Kelsoe, K. Rajewsky, and U. Weiss. 1991. Intraclonal generation of http://www.jimmunol.org/ antibody mutants in germinal centres. Nature 354:389. 67. Unkeless, J. C., E. Scigliano, and V. H. Freedman. 1988. Structure and function 45. Jacob, J., R. Kassir, and G. Kelsoe. 1991. In situ studies of the primary immune of human and murine receptors for IgG. Annu. Rev. Immunol. 6:251. response to (4-hydroxy-3-nitrophenyl)acetyl. I. The architecture and dynamics of 68. Prodeus, A. P., S. Goerg, L. M. Shen, O. O. Pozdnyakova, L. Chu, E. M. Alicot, responding cell populations. J. Exp. Med. 173:1165. C. C. Goodnow, and M. C. Carroll. 1998. A critical role for complement in 46. Slifka, M. K., M. Matloubian, and R. Ahmed. 1995. Bone marrow is a major site maintenance of self-tolerance. Immunity 9:721. of long-term antibody production after acute viral infection. J. Virol. 69:1895. 69. Hertz, M., V. Kouskoff, T. Nakamura, and D. Nemazee. 1998. V(D)J recombi- 47. Manz, R. A., A. Thiel, and A. Radbruch. 1997. Lifetime of plasma cells in the nase induction in splenic B lymphocytes is inhibited by antigen-receptor signal- bone marrow. Nature 388:133. ing. Nature 394:292. 48. Takahashi, Y., P. R. Dutta, D. M. Cerasoli, and G. Kelsoe. 1998. In situ studies 70. Goidl, E. A., W. E. Paul, G. W. Siskind, and B. Benacerraf. 1968. The effect of of the primary immune response to (4-hydroxy-3-nitrophenyl)acetyl. V. Affinity antigen dose and time after immunization on the amount and affinity of anti- maturation develops in two stages of clonal selection. J. Exp. Med. 187:885. hapten antibody. J. Immunol. 100:371.

49. Slifka, M. K., R. Antia, J. K. Whitmire, and R. Ahmed. 1998. Humoral immunity 71. Siskind, G. W., P. Dunn, and J. G. Walker. 1968. Studies on the control of by guest on September 25, 2021 due to long-lived plasma cells. Immunity 8:363. antibody synthesis. II. Effect of antigen dose and of suppression by passive an- 50. Dal Porto, J. M., A. M. Haberman, M. J. Shlomchik, and G. Kelsoe. 1998. An- tibody on the affinity of antibody synthesized. J. Exp. Med. 127:55. tigen drives very low affinity B cells to become plasmacytes and enter germinal 72. Steiner, L. A., and H. N. Eisen. 1967. Sequential changes in the relative affinity centers. J. Immunol. 161:5373. of antibodies synthesized during the immune response. J. Exp. Med. 126:1161. 51. Herzenberg, L. A., S. J. Black, and T. Tokuhisa. 1980. Memory B cells at suc- 73. Neuberger, M. S., and K. Rajewsky. 1981. Activation of mouse complement by cessive stages of differentiation: affinity maturation and the role of IgD receptors. monoclonal mouse antibodies. Eur. J. Immunol. 11:1012. J. Exp. Med. 151:1071. 74. Bachmann, M. F., U. Kalinke, A. Althage, G. Freer, C. Burkhart, H. Roost, 52. Eisen, H. N., and G. W. Siskind. 1964. Variations in affinities of antibodies M. Aguet, H. Hengartner, and R. M. Zinkernagel. 1997. The role of antibody during the immune response. Biochemistry 3:996. concentration and avidity in antiviral protection. Science 276:2024.