Germinal Center B Cells Replace Their Antigen Receptors in Dark Zones and Fail Light Zone Entry when Immunoglobulin Gene Mutations are Damaging

Abstract

Adaptive immunity involves the development of bespoke antibodies in germinal centers (GCs) through immunoglobulin somatic hypermutation (SHM) in GC dark zones (DZs) and clonal selection in light zones (LZs). Accurate selection requires that cells fully replace surface B cell receptors (BCRs) following SHM, but whether this happens before LZ entry is not clear. We found that most GC B cells degrade pre-SHM receptors before leaving the DZ, and that B cells acquiring crippling mutations during SHM rarely reached the LZ. Instead, apoptosis was triggered preferentially in late G1, a stage wherein cells with functional BCRs re-entered cell cycle or reduced surface expression of the chemokine receptor CXCR4 to enable LZ migration. Ectopic expression of the anti-apoptotic gene Bcl2 was not sufficient for cells with damaging mutations to reach the LZ, suggesting that BCR-dependent cues may actively facilitate the transition. Thus, BCR replacement and pre-screening in DZs prevents the accumulation of clones with non-functional receptors and facilitates selection in the LZ.

Graphical abstract

Figure 1

Reduced BCR Levels on DZ GC B Cells (A and B) Mice were immunized with SRBCs and splenic IgD^low CD95⁺ GL7⁺ GC B cells were analyzed on day 7. (A) BCR levels on DZ populations (CXCR4^high) were analyzed and compared to B220^neg non-B cells. The gate indicates a BCR^low DZ population. (B) DZ and LZ cells are gated as shown and BCR levels overlaid with IgD⁺ CD95^low follicular B cells. (C–G) HEL-specific SW_HEL or MD4 B cells were adoptively co-transferred with OT-II CD4⁺ T cells and mice were immunized with HEL-OVA/adjuvant adjuvant. Splenic IgD^low CD95⁺ GL7⁺ CD45.1⁺ GC B cells were analyzed on day 7. (C) Surface and intracellular BCR (IgMa) levels were co-stained in fixed/permeabilized MD4 GC B cells revealing a subset of cells that is negative for both. (D and E) DZ cells with low BCR levels in MD4 transgenic (D) and SW_HEL gene targeted GC B cell populations (E). (F and G) Single CXCR4^high BCR^high and CXCR4^high BCR^low DZ SW_HEL GC B cells (gated IgD^low, CD38^low, GL7⁺, CD45.1⁺, or CD45.2⁺) were index sorted and cultured for 8 days on NB21 feeder cells. (F) The BCR and CXCR4 staining profiles (at the time of sort) of single GC B cells that grew and secreted detectable antibody after culture are shown. The plots for cells that did not grow, or that grew did not secrete detected antibody, are also shown and overlaid with the total GC B cell population for reference. Plots for additional mice are shown in Figure S1. (G) Summary of results. Numbers of cells in each group per experiment are indicated. (H) Surface BCR levels on Aicda^+/+ and Aicda^−/− IgD^low CD95⁺ GL7⁺ GC cells from mixed bone marrow chimeric mice 8 days after SRBC immunisation. Numbers in (C–E) and (H) represent mean ± SEM from 7 (C and D), 10 (E) and 11 (H) mice from 2–4 experiments. See also Figure S1.

Figure 2

DZ Cells Acquiring Detrimental Mutations in Their Immunoglobulin V Region Genes Accumulate in BCR^low Gates SW_HELxFucci2 B cells were co-transferred with OT-II CD4⁺ T cells into WT hosts that were subsequently immunized with HEL-OVA/adjuvant. Single IgD^low CD95⁺ GL7⁺ (mice 1-3), or IgD^low CD95⁺ (mice 4 or 5), CD45.1/2⁺ SW_HEL GC B cells were index sorted on day 8 and their Ighv region genes were PCR amplified, Sanger sequenced, and analyzed for the presence of premature stop codons. (A) Frequency (mean ± S.E.M., n = 5) of premature stop codons in BCR^high and BCR^low CXCR4^high CD83^low DZ populations is shown. (B) Summary of results from multiple mice and experiments. Numbers of cells sequenced are indicated above bars and the proportion of total DZ stop codons that are in the BCR^low subset is marked below. Results are normalized for rare population enrichments performed during FACS sorting. (C) BCR and CXCR4 levels on individual indexed sequenced cells. Colors indicate presence or absence of Ighv region premature stop codons in cells of the indicated phenotype (pos, positive, neg, negative). Data are overlaid with cells that were not sequenced (gray), for reference. Less frequent BCR^low cells were often enriched during FACS sorting. See also Figure S2.

Figure 3

Correlating Cell Cycle and the Acquisition of Stop Codons (A) Verification of utility of Fucci2 reporter mice for tracking cell cycle in IgD^low CD95⁺ GL7⁺ GC B cells on day 10 following SRBC immunisation. mVenus accumulates during S phase as is evidenced by its correlation with DNA content; mCherry is degraded during S phase but accumulates during G1. (B) S/G2/M (mVenus⁺) and early G1 (mCherry- mVenus^low) CXCR4^high CD83^low DZ cells were gated (i) and assessed for BrdU incorporation 30 minutes and 2 hours after a single injection (ii). S phase cells label at 30 mins while labeling of early G1 cells require completion of mitosis. (C) mCherry levels in CXCR4^low CD83^high LZ and CXCR4^high CD83^low DZ cells. (D and E) SW_HELxFucci2 CXCR4^high CD83^low DZ cells 8 days after immunization with HEL-OVA/adjuvant (as in Figure 2). (D) Gates indicate the cell cycle status of DZ SW_HELxFucci2 GC B cells (annotated in left plot). Cells with high and low BCR levels that have acquired premature Ighv region stop codons, as well as sequenced cells in which stop codons were not detected, are shown (right). Data are overlaid with cells that were not sequenced (gray), for reference. (E and F) Frequency of stop codons among DZ BCR^high (E) and DZ BCR^low (F) cells at the different cell-cycle stages. (E) includes results from 3 additional mice from which rare BCR^low S and early G1 phase cells were enriched to increased cell numbers. Plots in (A) are representative of 8 mice from 2 experiments, (B) of 4 mice from 2 experiments, (C) of 10 mice from 3 experiments. See also Figure S3.

Figure 4

Cells Carrying Premature Stop Codons Fail to Accumulate in the LZ (A) The frequency of cells carrying Ighv region premature stop codons in SW_HELxFucci2 LZ and DZ populations was determined. Plot and gate are from a representative mouse, numbers indicating means (+/− SEM) from five independent experiments. (B) The phenotype of LZ cells with premature stop codons is shown for all three mice. Data are overlaid with total GC B cells that were not sequenced (gray), for reference. (C) Summary from five independent experiments. Numbers above bars indicate number of each cell type sequenced. Numbers below indicate frequency of detected LZ stop codons that were BCR^high and have been normalized for where rare populations were enrichments performed during FACS sorting.

Figure 5

BCR^low GC B Cells Undergo Apoptosis during Late G1 in the DZ Fucci2 mice were immunized with SRBCs and splenic IgD^low CD95⁺ GL7⁺ GC B cells were analyzed on day 10. (A and B) GC B cells were stained with LZ/DZ markers and with anti- act-Casp3. (C) The timing of cell cycle re-entry from the DZ or surface CXCR4 decreases by G1 GC B cells was determined by examining mCherry levels. Yellow area indicates a likely fate junction where cells must be committed one of these processes. (D and E) DZ, act-Casp3⁺ DZ, and BCR^low act-Casp3⁺ DZ cells were gated and levels of the cell cycle regulated Fucci2 florescent reporters were examined. Cell-cycle stages are gated in (D) and mean frequencies ± SD are shown, with summaries in (E). (F) BCR (Igβ) levels on total GC B cells and act-Casp3⁺ GC B cells. (G) Frequency of BCR^high and BCR^low DZ cells that were act-Casp3⁺. (H) mCherry levels in LZ, DZ, and Act-Casp3⁺ cells are shown. Plots and graphs in (A–C) are representative of 13 mice from 4 experiments; plots and means (+/− SD) in (D–H) are compiled from, or representative of (F and H), of 9 mice from 3 experiments. Each point in (B, E, G) represents a single mouse. Analysis was performed using an unpaired two-tailed Student’s t test. ^∗∗∗p < 0.001. See also Figure S4.

Figure 6

Promotion of Survival Is Not Sufficient to Permit Cells with Non-Functional BCRs to Enter the LZ (A and B) Fucci2 mice with or without (WT) a BCL2 transgene (BCL2-tg) were immunized with SRBCs and analyzed day 10 (A) Frequencies of CXCR4^high CD83^low DZ cells and (B) CXCR4^high BCR(Igβ)^low cells in GCs (CD38^low IgD^low GL7⁺). (C–E) SW_HELxBCL2-tg B cells were co-transferred with OT-II CD4⁺ T cells into WT hosts which were subsequently immunized with HEL-OVA/adjuvant and their GCs (IgD^low CD95⁺ GL7⁺) were analyzed on day 8. (C) Frequencies of Bcl2-tg SW_HEL DZ cells with high and low BCR levels that carry Ighv region premature stop codons are shown (left). The phenotype of sequenced cells with stop codons is shown to right. (D) Summary of results from multiple mice and experiments. Numbers of cells sequenced are indicated. Mouse 1 DZ is gated using just CXCR4. (E) Frequencies of DZ and LZ cells with premature stop codons. Numbers are means from two independent experiments which are shown separately in Figure S5F. Data in (A) and (B) are pooled from 3 experiments with each point representing a single mouse. Analysis was performed using an unpaired two-tailed Student’s t test, ^∗∗∗p < 0.001. See also Figure S5.