Distinct cellular pathways select germline-encoded and somatically mutated antibodies into immunological memory

Abstract

One component of memory in the antibody system is long-lived memory B cells selected for the expression of somatically mutated, high-affinity antibodies in the T cell-dependent germinal center (GC) reaction. A puzzling observation has been that the memory B cell compartment also contains cells expressing unmutated, low-affinity antibodies. Using conditional Bcl6 ablation, we demonstrate that these cells are generated through proliferative expansion early after immunization in a T cell-dependent but GC-independent manner. They soon become resting and long-lived and display a novel distinct gene expression signature which distinguishes memory B cells from other classes of B cells. GC-independent memory B cells are later joined by somatically mutated GC descendants at roughly equal proportions and these two types of memory cells efficiently generate adoptive secondary antibody responses. Deletion of T follicular helper (Tfh) cells significantly reduces the generation of mutated, but not unmutated, memory cells early on in the response. Thus, B cell memory is generated along two fundamentally distinct cellular differentiation pathways. One pathway is dedicated to the generation of high-affinity somatic antibody mutants, whereas the other preserves germ line antibody specificities and may prepare the organism for rapid responses to antigenic variants of the invading pathogen.

Figure 1.

Targeted insertion of loxP sites into the Bcl6 gene. (A). A DNA fragment which contains exons 7–9 of the Bcl6 gene was flanked by loxP sites (solid triangles) with an FRT-flanked pGK-Neo cassette inserted next to it. The pGK-Neo cassette was removed from the mouse germline by breeding heterozygous mice to FLPe-expressing deleter mice. Xb, XbaI site; Hd, HindIII site. (B) Southern blot analysis of XbaI digested genomic DNA from a WT and a targeted ES clone using the probe depicted by the thick bar in A. (C) Removal of the pGK-Neo cassette in a heterozygous floxed mouse was confirmed by genomic PCR. Location of the primer set is indicated by the two arrowheads on the bottom of A. (D–G) Bcl6^flox mice were crossed to mb1-cre mice. Conditional Bcl6-deletion in B cells does not affect BM B cell development (D), B cell numbers in the spleen (E), and B cell characteristics before immunization (F and G). BM cells and splenocytes were prepared from Bcl6^+/+ (+/+), Bcl6^+/f (+/f), and Bcl6^f/f (f/f) mice heterozygous for mb1-cre (n = 3). The frequency of B cell subsets in BM (D) and T and B cells in the spleen (E) was analyzed by FACS. Each symbol represents the number of cells in an individual mouse. Bars represent the mean number for each group. (F) Representative FACS plots of splenic B cells with expression of CD44 and CD69 (activated B cell phenotype). Numbers in the plots indicate the percentage of cells in quadrants. (G) The levels of cytokines in culture supernatants produced by splenic B cells upon stimulation with anti-IgM and anti-CD40 mAb or LPS for 3 d (see Materials and methods). Bars represent means for each group. The data are representative of two independent experiments in D–G.

Figure 2.

IgG1⁺ memory B cells develop independently of the GC reaction. Splenocytes were recovered from Bcl6^+/+, Bcl6^+/f, and Bcl6^f/f mice heterozygous for Cγ1-cre (A) or mb1-cre (B) at day 7 or 40 after immunization with NP-CG in alum and NP-specific/IgG1⁺ memory and GC B cells were enumerated from 10⁶ cells within the lymphocyte gate (see Materials and methods). Circles represent the number of cells in individual mice (n = 3–5). Bars indicate the mean number in each group. (C) BM cells from Bcl6^f/f mice heterozygous for mb1-cre (CD45. 2⁺) and CD45.1⁺ WT or heterozygous mb1-cre mice were mixed 1:1 and transferred into CD45.2⁺Rag-1^−/− mice (n = 4). After 8 wk, the recipient mice were immunized with NP-CG. Splenocytes were stained with anti-CD45.1 mAb to distinguish the WT (45.1) and conditional Bcl6 KO (45.2) compartments at days 7 and 40 after immunization. The number of B cells and NP-specific memory and GC B cells in individual spleens was determined as in A. Data are representative of two (A) and three (B and C) independent experiments. See also Fig. S1.

Figure 3.

The cell cycle position and V_H186.2 gene mutations in NP-specific memory and GC B cells. (A) DNA content in memory (Me) and GC B cells in conditional Bcl6-deficient (Bcl6^f/f/mb-1-cre^+/−) and control (Bcl6^+/+/mb-1-cre^+/−) mice at days 7 and 40 after immunization with NP-CG. Each symbol represents the percentage of positive cells in an individual mouse (n = 3). Data are representative of three independent experiments. (B) The cell cycle position of memory B cells (Me) and GC B cells (GC) in FUCCI transgenic mice (n = 3) at indicated times. Each symbol represents the percentage of cells positive for Azami Green in an individual mouse. Data are representative of three (day 7) and two (day 40) independent experiments. (C and D) Single NP-specific/IgG1⁺ memory and GC B cells were purified from the pooled spleens of Bcl6^+/+ (+/+) and Bcl6^f/f (f/f) mice heterozygous for mb1-cre (n = 6–13) at day 7 (C) or day 40 (D) after immunization and subjected to RT-PCR to amplify rearranged V_H186.2-Cγ1 cDNA for sequencing (see Materials and methods). (E) CD45.1⁺ (45.1) WT memory and GC B cells and CD45.1⁻ (45.2) Bcl6-deficient memory B cells were purified as single cells from pooled spleens of BM chimeras (n = 9) at day 40 after immunization and their V_H genes were sequenced as in C. Number of mutated clones/number of V186.2 genes sequenced are also shown in C–E. Circles represent the number of mutations in individual clones. Closed circles represent W33L clones.

Figure 4.

GC-independent memory B cells are maintained for a long period and later joined by somatically mutated GC descendents. (A) Bcl6^+/+ (+/+), Bcl6^+/f (+/f), and Bcl6^f/f (f/f) mice heterozygous for mb1-cre (n = 5–25) were injected with BrdU twice at days 4–6 after immunization with NP-CG. Memory (Me) and GC B cells (GC) were then purified at indicated times. BrdU⁺ B cells were determined as in Materials and methods. Circles represent the percentage of BrdU⁺ cells in three independent experiments. (B) C57BL/6 mice (n = 5–25) were immunized and injected with BrdU as described in A. On days 6, 8, and 10 after immunization, mice were either not treated (−) or injected with either hamster anti–CD40-ligand mAb (MR1) or normal hamster IgG (Ig) and sacrificed at day 20 or 40 for purification of NP-specific/IgG1⁺ memory (Me) and GC B cells (GC). BrdU incorporation was determined as in A. Circles represent the percentage of BrdU⁺ cells in three independent experiments. (C) Accumulation of mutations in V_H186.2 genes that were PCR amplified from single NP-specific/IgG1⁺ memory B cells in immunized WT mice (n = 10–22) at the indicated times. Number of mutated clones/number of V186.2 genes sequenced are also shown. Symbols are as in Fig. 3. The results were evaluated statistically by Mann-Whitney nonparametric test, with P < 0.05 regarded as significant. (D) C57BL/6 mice (n = 5) were immunized with NP-CG in alum and injected with or without (−) anti-CD40 ligand mAb (MR1) or normal Ig (Ig) as described in B. At day 20 or 40, the number of NP-specific/IgG1⁺ memory and GC B cells was analyzed by FACS. Each symbol represents the number of cells in individual mice. The data are representative of two independent experiments. (E) Immunofluorescence analysis of splenic cryosections from NP-CG–immunized C57BL/6 mice treated with MR1 or normal Ig was performed. Sections were stained with anti-B220 (blue) and PNA (red; see Materials and methods). Representative images of three mice are shown. Bars, 300 µm. (F) Mutational analysis of V_H186.2 rearrangements from single day 40 memory B cells of immunized mice (n = 20–40). Mice were either not treated (−) or injected with anti-CD40 ligand mAb (MR1) or normal Ig (Ig) as described in B. Number of mutated clones/number of V186.2 genes sequenced are also shown. Symbols are as in Fig. 3.

Figure 5.

Memory B cells attain functional maturity as the immune response progresses. (A) Splenocytes were recovered from either naive or NP-CG–primed Bcl6^+/+ (+/+), Bcl6^+/f (+/f), and Bcl6^f/f (f/f) mice (n = 4–14) heterozygous for mb1-cre (mb1-cre^+/−) at days 7 and 40 after immunization. B cells depleted of plasma cells were enriched by MACS and memory B cell frequency was estimated by FACS. B cell populations containing 1.5 × 10³ NP-specific/IgG1⁺ memory B cells were transferred into Rag-1^−/− mice, together with CG-primed CD4⁺ T cells and naive B cells, followed by immunization with soluble NP-CG. The number of total and high-affinity anti-NP/IgG1⁺ ASCs in the spleen was measured by ELISPOT at day 10 after immunization. Shown are the total number of anti-NP/IgG1⁺ ASCs (left) and the ratio of high-affinity ASCs/total number of ASCs (right). (B) On days 6, 8, and 10 after immunization, mice were either not treated (−) or injected with either anti-CD40 ligand mAb (MR1) or normal Ig (Ig) and sacrificed at day 7 (d7, n = 5) or 40 (d40, n = 22–25) for purification of NP-specific/IgG1⁺ memory B cells. Memory B cells (1 × 10³), CG-primed CD4⁺ T cells (4 × 10⁶), and naive B cells (10⁶), as filler cells, were transferred into Rag-1^−/− mice (n = 4), followed by immunization as in A. Anti-NP/IgG1⁺ ASCs in the spleen were determined as in A. (C) B cells were prepared from mutant (Bcl6^f/f/mb-1-cre^+/−, n = 4) or control mice (Bcl6^+/+/mb-1-cre^+/−, n = 4) at day 70 (d70) after immunization with NP-CG. NP-primed B cells or naive B cells (Naive) were transferred into Rag-1^−/− mice (n = 5), together with CG-primed T cells, followed by immunization with NP-CG. The number of total and high-affinity anti-NP/IgM ASCs in the spleen was measured by ELISPOT at day 5 after immunization. Data are representative of two (C) and three (A and B) independent experiments. Circles represent cell numbers in individual mice.

Figure 6.

Gene expression profiles between days 7 and 40 memory B cells are closely related. (A) A hierarchical clustering of triplicate samples. Colors in the heat map depict the Pearson’s correlation coefficient between a pair of samples. A higher value is represented by dark blue as shown in the vertical bar. An AU (approximately unbiased) p-value (percentage) was calculated and placed on a branch of a cluster dendrogram. The values in red color indicate that the microarray pairs are significantly clustered. NP-specific/IgG1⁺ memory (Me) and GC B cells (GC) and plasma cells (Pc, day 7 only) were purified from WT mice at day 7 (n = 14–20) and day 40 (n = 30–39) after immunization for extraction of total RNA. FO and MZ B cells were purified from unimmunized mice (n = 3). FO B cells were stimulated with anti-IgM and anti-CD40 mAb in vitro for 6 h (FO + stimulation). (B) qRT-PCR analysis for candidate genes identified in the microarray data analysis. NP-specific/IgG1⁺ memory (Me) and GC B cells and plasma cells (Pc) were purified from WT mice (n = 10) at day 7 after immunization. Day 40 memory B cells were purified from the pooled spleens of immunized mice (n = 18–25), treated with anti-CD40L mAb (MR1), control Igs, or left untreated. cDNA synthesized from total RNA was used for qRT-PCR with the indicated gene-specific primers (see Table S1) and standardized to the relative expression of β-actin. Shown here are gene expression profiles of naive FO (a) and MZ B cells (b), and memory (c), GC (d), and plasma cells (e) at day 7 after immunization. Also shown are day 40 memory B cells that developed in untreated mice (f) or those treated with either control Igs (h) or MR1 (j), together with day 40 GC B cells in untreated mice (g) or those treated with Igs (i). FO B cells stimulated as in A for 6 h (k) and 24 h (l) are also shown. Data are representative of two independent experiments. Error bars represent ±SD. See also Fig. S3.

Figure 7.

Different subsets of T cells support GC-independent and -dependent memory B cell development. (A) CD40 KO (−/−) and control (+/+) mice were immunized with NP-CG and generation of NP-specific/IgG1⁺ B cells was analyzed by FACS (left). Circles represent the number of cells in individual mice (n = 3, right). (B–D) Frequencies of Tfh (B), NP-specific/IgG1⁺ GC (C), and memory B cells (D) in spleens of Bcl6^+/+ (+/+) and Bcl6^f/f (f/f) mice carrying a CD4-cre transgene at days 7 and day 40 after immunization with NP-CG. Unimmunized mice were used as a control in B. Circles represent the number of cells in individual mice (n = 4–6). (E) The frequency of day 7 NP-specific/IgG1⁺ memory and GC B cells in the spleens of mutant (Bcl6^f/f/mb1-cre^+/− and CD4-cre) and control (Bcl6^+/+/mb1-cre^+/− and CD4-cre) mice. Circles represent the number of cells in individual mice (n = 4). (F and G) Mutational analysis of V_H186.2 rearrangements from NP-specific/IgG1⁺ memory and GC B cells of Bcl6^+/+ (+/+) and Bcl6^f/f (f/f) mice carrying a CD4-cre transgene at days 7 (F) and 40 (G) after immunization (n = 4–6). Circles represent the number of mutations in individual clones. Number of mutated clones/number of V186.2 genes sequenced are also shown in F and G. Closed circles represent high-affinity clones (W33L). (H) NP-specific/IgG1⁺ memory B cells were purified from Bcl6^+/+ (+/+) and Bcl6^f/f (f/f) mice carrying a CD4-cre transgene (n = 10–19) at day 40 after immunization and their secondary response was assessed as in Fig. 5. Response by naive B cells was also assessed as a control. Circles represent the number of anti-NP/IgG1⁺ ASCs in the spleens of individual mice at day 10 after immunization (n = 4). The data are representative of two independent experiments in A–E and H.

Figure 8.

Expression of PD-L2 and CD80 on GC-independent and -dependent memory B cells. (A) Histograms represent the relative intensity of indicated surface markers on naive B cells (darkly shaded), day 7 and day 40 memory B cells (solid line), and GC B cells (dashed line) in mutant (Bcl6^f/f/mb-1-cre) and control (Bcl6^+/+/mb-1-cre) mice. Representative results of two independent experiments with three mice per group are shown. (B) Accumulation of mutations in V_H186.2 genes that were PCR amplified from single CD73⁺ and CD73⁻ NP-specific/IgG1⁺ memory B cells and GC B cells in immunized WT mice (n = 5) at day 40 after immunization. Number of mutated clones/number of V186.2 genes sequenced are also shown. Symbols are as in Fig. 3.

Figure 9.

Antigen-engaged IgG1⁺ B cells differentiate into memory cells prior to the appearance of Bcl6 expressing pre-GC B cells. (A) C57BL/6 (CD45.2) mice (n = 3–4) were injected with B220⁺ B cells from B1-8^hi mice (CD45.1) and immunized with NP-CG, followed by FACS analysis. NP-specific/B220⁺ B cells of donor origin (CD45.1⁺) in the immunized or unimmunized recipients were gated and their PNA binding and IgG1 expression were analyzed (left). (B) Numbers of NP-specific/IgG1⁺/PNA⁻ (−) and PNA⁺ (+) B cells from day 3–4 (left) and NP-specific/IgG1⁺ memory B cells (CD38⁺/PNA⁻, Me) and GC B cells (CD38^dull/PNA⁺, GC) at day 7 (right) after immunization. MR1 was administered from days 3 to 6 after immunization as in Fig. 4 B. (C) Expression of surface PD-L2 and CD80 and intracellular Bcl6 analyzed by FACS in NP-binding B cells from unimmunized recipients (control B cells) and NP-specific/IgG1⁺ B cells and GC and memory B cells in recipients at different time points after immunization. Memory B cells were purified from recipients treated or not treated with MR1. Representative results of three independent experiments are shown in A–C. (D) Gene expression profiles in NP-specific/IgG1⁺ donor B cells at day 3 (b) and day 4 (c) after immunization and day 7 NP-specific/IgG1⁺ memory B cells in MR1-treated (e) or -untreated (d) recipients, together with GC B cells in untreated recipients (f). NP-binding B cells from unimmunized recipients served as controls (a). Total RNA was purified from B cells and cDNAs were synthesized for qRT-PCR with the indicated gene-specific primers. Also shown are gene expression profiles of naive FO B cells (g), MZ B cells (h), day 7 and day 40 memory B cells (i and l), GC B cells (j and m), plasma cells (k), and activated FO B cells (compare Figure 6; FO + Stim). Error bars represent ±SD. Data are representative of two independent experiments and standardized to the expression of β-actin.