Apoptosis and antigen affinity limit effector cell differentiation of a single naïve B cell

See allHide authors and affiliations

Science  13 Feb 2015:
Vol. 347, Issue 6223, pp. 784-787
DOI: 10.1126/science.aaa1342

For a single B cell, many roads to take

To successfully fight a pathogen, immunological B cells must wage a multipronged attack: They can differentiate into antibody-secreting plasma cells or become T cell–helping germinal center cells, or even long-lived memory cells. But can a single B cell acquire all of these different fates? To find out, Taylor et al. tracked the responses of single B cells in mice. Although some B cells acquired only one fate, others differentiated into all three. The authors linked the ability of B cells to differentiate into multiple subsets to their ability to proliferate and resist cell death, and the affinity of their antigen receptor.

Science, this issue p. 784


When exposed to antigens, naïve B cells differentiate into different types of effector cells: antibody-producing plasma cells, germinal center cells, or memory cells. Whether an individual naïve B cell can produce all of these different cell fates remains unclear. Using a limiting dilution approach, we found that many individual naïve B cells produced only one type of effector cell subset, whereas others produced all subsets. The capacity to differentiate into multiple subsets was a characteristic of clonal populations that divided many times and resisted apoptosis, but was independent of isotype switching. Antigen receptor affinity also influenced effector cell differentiation. These findings suggest that diverse effector cell types arise in the primary immune response as a result of heterogeneity in responses by individual naïve B cells.

Antibody production results from a differentiation process that begins when the surface form of immunoglobulin (Ig) known as the B cell receptor (BCR) on a naïve B cell binds an antigen (1, 2). BCR signaling causes the B cell to migrate to the border of the T cell zone, where it receives signals from T cells (3, 4). These signals cause the B cell to proliferate and differentiate into several types of effector cells, including short-lived plasma cells, germinal center (GC) cells, and GC-independent memory cells (1, 2). GC cells then undergo somatic hypermutation in their Ig genes, and cells with mutations that improve BCR affinity for antigen are selected to become GC-dependent memory or plasma cells (1, 2).

Despite the importance of this process to immunity and vaccination, it is unclear how individual naïve B cells simultaneously produce all of the early effector cell types. Some studies suggest that different naïve B cell clones only produce a single effector subset, depending on BCR affinity for antigen (58) or intrinsic stochastic biases of the responding clonal population (9). Alternatively, each naïve B cell may produce all effector cell types, as suggested by recent work on naïve T cells (1013).

We addressed these possibilities by tracking the fates of antigen-specific naïve B cells during the primary immune response to the protein antigen allophycocyanin (APC). Using a sensitive antigen-based cell enrichment method (14), we found that the spleen and lymph nodes of a C57BL/6 (B6) mouse contained about 4000 polyclonal APC-specific naïve B cells, which produced ~100,000 effector cells 7 days after immunization with APC in complete Freund’s adjuvant (CFA) (Fig. 1, A and B). As expected, the effector cell population consisted of B220low Ighigh antibody-secreting plasma cells, CD38 GL7+ GC cells, CD38+ GL7 memory cells, and a few remaining undifferentiated CD38+ GL7+ activated precursors (APs) (15) (Fig. 1, C and D, and fig. S1).

Fig. 1 Assessing the polyclonal APC-specific B cell response.

(A) Detection and (B) quantitation of APC-specific B cells from pooled spleen and lymph node samples enriched using anti-APC microbeads from naïve mice (n = 18) or mice immunized with APC in CFA (n = 45) 7 days earlier. (C) Detection and (D) quantitation of APC-specific plasma cells (Ighigh), GC (Ig+ B220+ CD38- GL7+), naïve/memory (M) (Ig+ B220+ CD38+ GL7), and AP B cells (Ig+ B220+ CD38+ GL7+). Memory B cells were quantitated by the increase in CD38+ GL7 cells over uninjected controls. (E) Detection and (F) quantitation of CFSEhigh CD45.1+ APC-specific B cells from CD45.2+ mice that received 2 × 107 donor CD45.1+ B cells before immunization with APC in CFA, or CFA alone (n = 10). Numbers on the flow cytometry plots in (A), (C), and (E) reflect the percent of cells within the gated population. These percentages and knowledge of the total number of B cells in the enriched fraction were used to calculate the number of cells shown in (B), (D), and (F). The bars represent the mean and P values determined using an unpaired two-tailed Student’s t test. Data points were combined from two to six experiments.

In vivo limiting dilution was used to assess the multipotentiality of a single APC-specific naïve B cell. Before limiting dilution could be achieved, it was necessary to determine the fraction of APC-specific naïve B cells that responded to immunization. Twenty million B cells from CD45.1+ mice that were never exposed to APC were labeled with the cell division–tracking dye carboxyfluorescein succinimidyl ester (CFSE) (16) and transferred into CD45.2+ recipients. Donor-derived APC-specific B cells were CFSEhigh 7 days after immunization with CFA alone, which is indicative of cells that had not divided (Fig. 1E). After the injection of APC in CFA, most donor APC-specific B cells were CFSElow, and the CFSEhigh population was 33% smaller than in mice injected with CFA alone (Fig. 1, E and F). These results indicated that one in three APC-specific naïve B cells, or 1 in 60,000 total B cells, proliferated in mice immunized with APC. The 33% response frequency of APC-specific naïve B cells was not a limitation of the CFSE dilution assay, because 97 to 100% of naïve MD4 B cells proliferated (fig. S2) after the injection of hen egg lysozyme (HEL) or duck egg lysozyme (DEL), for which the MD4 BCR has a high or medium affinity, respectively (17). Thus, the 33% responder frequency was a feature of the polyclonal APC-specific B cell population under these immunization conditions.

Limiting dilution experiments were then performed, based on the above knowledge and the fact that 7.7 ± 2.8% (n = 116 recipients) of donor naïve B cells survive after transfer. 2 × 106 or 0.2 × 106 CD45.1+ B cells were transferred into CD45.2+ mice, with the expectation that an average of 3.3 or 0.33 APC-responsive CD45.1+ naïve B cells would survive per recipient. Seven days after APC immunization, mice that did not receive transferred B cells contained two or fewer CD45.1+ background events (Fig. 2A). All mice that received 2 × 106 B cells contained a defined population of CD45.1+ donor-derived APC-specific B cells that had proliferated in response to APC (Fig. 2, A and B). In contrast, 19% (74 out of 384) of mice that received the limiting number of 0.2 × 106 B cells contained donor-derived APC-responsive B cells (Fig. 2, B and C). Based on the Poisson distribution (18), over 91% of the donor-derived populations in this group were the progeny of a single naïve B cell.

Fig. 2 Assessing the response of an individual naïve APC-specific B cell.

(A) Detection of APC-specific donor cells from CD45.2+ recipients that received 0.2 or 2 × 106 CFSE or Celltrace violet (CTV)–labeled CD45.1+ B cells 1 to 3 days before immunization with APC in CFA. Samples were analyzed 7 days after immunization, after simultaneous CD45.1 and APC-based cell enrichment. The fourth and fifth plots show CFSE profiles for gated populations a and b from the second and third plots. Numbers on the plots reflect the percent of cells within the gated population. (B) Frequency of immunized recipient mice (n = 6 for mice that received 2 × 106 cells or n = 384 for those that received 0.2 × 106) containing an APC-specific CFSE/CTVlow donor population above the limit of detection (LOD) of two cells. (C) Total number of cells in APC-specific clonal populations from 74 mice that received 0.2 × 106 cells and contained a population above the LOD. (D) Frequency of each subset within polyclonal or clonal APC-specific populations. Subsets are gated as shown in Fig. 1C, and each row depicts an individual clone (n = 74) or the entire APC-specific population from a mouse (n = 45). (E) Frequency of APC-specific clones generating one, two, three, or four subsets. (F) Total number of cells produced by each clone, separated into groups based on the number of subsets produced. A Mann-Whitney test was used to generate the P values. Bars in (C) and (F) represent medians. Data points were combined from 12 experiments.

Extensive effector cell heterogeneity was observed in the progeny of individual naïve B cells. Single naïve B cells produced between 4 and 957 progeny, with a median of 16 (Fig. 2C). The polyclonal naïve cell populations of recipient origin produced all effector cell subsets, but 35 of the 74 clonal populations (44%) contained only plasma cells, only GC cells, only GC-independent memory cells, or only AP cells (Fig. 2, D and E). Four clonal populations contained all four subsets, and many contained two or three.

Large clonal populations were more likely to contain multiple subsets. The clonal populations that contained all four subsets had a median of 352 cells, whereas the populations that contained only one had a median of 10 cells (Fig. 2F). No relationship was observed between the size of a clonal population and the frequency of cells in it that divided seven or more times (Fig. 3, A and B). This finding pointed toward cell death as a basis for the differences in clonal population size. Consistent with the idea that cell death limits population size, most clonal populations contained less than 20% of the minimum number of daughter cells expected based on their CFSE profile (Fig. 3C). This effect was not uniform across all populations. Large clonal populations that contained all four subsets exceeded the expected minimum number of daughter cells, whereas the small populations that contained only one subset had a median of only 12% (Fig. 3D).

Fig. 3 Assessing proliferation and apoptosis of APC-specific clones.

(A) Frequency of cells in each CFSE/CTV division bin (Div) in wild-type (n = 74) or Bim-deficient (n = 34) APC-specific clonal populations. Clones are displayed in the same order as in Fig. 2D. (B) Total number of cells detected for each wild-type (open circle) or Bim-deficient (black triangle) clone compared to the frequency of cells completing at least seven divisions. (C and D) Number of cells detected for each clone displayed as a percentage of the minimum number predicted based on CFSE/CTV dilution analysis, with the clones grouped in (D) based on the number of subsets produced. (E) Total number of cells produced by APC-specific Bim-deficient clones. (F and G) Frequency of clones that produced (F) the indicated number or types of subsets or (G) any of the indicated subset. The bars in (C) to (E) represent medians. P values were determined in (C) and (D) using a Mann-Whitney test and in (F) and (G) using Fisher’s exact test. Data points were combined from 3 to 12 experiments.

These results suggested that the multipotentiality of a single naïve B cell was related to the production of progeny resistant to cell death. This idea was tested using B cells lacking the proapoptotic-mediator Bim (19). Bim-deficient and wild-type B cells fluxed calcium and proliferated equally in response to BCR signaling in vitro (fig. S3). Single Bim-deficient, APC-specific, naïve B cells, however, produced 3.4-fold more progeny than wild-type clones (P < 0.0001, compare Fig. 2C to Fig. 3E) in response to APC immunization. Bim-deficient clones were also more likely to produce multiple effector cell subsets (Fig. 3F), especially those containing plasma cells, GC cells, and AP cells (Fig. 3G). Unlike wild-type clones, Bim-deficient clonal populations showed a significant correlation between the number of cells and cell division (Fig. 3B), suggesting that they experienced less apoptosis. Consistent with apoptosis limiting population size, most Bim-deficient clones approached or exceeded the minimum number of daughter cells expected based on their CFSE profile (Fig. 3C). Thus, the capacity of a single naïve B cell to produce many effector cells and multiple subsets appears to be limited by Bim-mediated apoptosis, although suppression of proliferation by Bim could also contribute.

The multipotentiality of a single naïve B cell could also be influenced by BCR affinity for antigen. This was tested by comparing the response of single BCR transgenic MD4 B cells to high-affinity (HEL) or medium-affinity (DEL) antigens to that of activation-induced cytidine deaminase (AID)–deficient APC-specific B cells (Fig. 4, A and B), which like MD4 cells are unable to undergo class switching (20). Single AID-deficient APC-specific B cells produced a similar number (compare Fig. 2C to Fig. 4B) and diversity (Fig. 4C) of effector cells as their wild-type counterparts, indicating that class switching does not play a major role in differentiation at the early time point analyzed in these experiments. In contrast, single naïve MD4 cells stimulated with HEL or DEL produced more progeny than single APC-specific cells stimulated with APC (Fig. 4, A and B). 58% of HEL-stimulated single MD4 cells produced only plasma cells and GC cells, a combination that occurred in few clonal APC-specific B cell populations (Fig. 4C). This idiosyncrasy was related to BCR affinity, because only 12.5% of single MD4 cells produced this pattern when stimulated with the lower-affinity antigen DEL (Fig. 4C). In addition, HEL-stimulated clones generated more plasma cells and fewer memory cells than DEL-stimulated clones (Fig. 4D), which is consistent with earlier work in the MD4 system (5, 8). The tendency of clones with higher-affinity BCRs to produce plasma cells was also observed in the polyclonal repertoire. BCR affinity for antigen was indirectly measured as the amount of APC bound to memory cells, which express BCR at levels similar to naïve cells (fig. S1A). Among the APC-specific clones that produced memory cells, those that also produced plasma cells bound more APC than those that did not (Fig. 4E). Together, these data indicate that BCR affinity for antigen influences the precise effector cell pattern produced by a naïve B cell.

Fig. 4 Assessing the response of an individual BCR transgenic B cell.

(A) Representative detection of CD45.2+ IgMa CFSElow donor cells from recipients of a limited number (5 to 15 cells) of MD4 Rag1−/− B cells 1 to 3 days before immunization with HEL–OVA or DEL-OVA in CFA. Samples were analyzed 7 days after immunization after CD45.2-based cell enrichment. The second and third plots show gated populations a and b from the first and second plots. Numbers on the plots reflect the percent of cells within the gated population. (B) Total number of cells in HEL-stimulated MD4 (black circles, n = 31), MD4 DEL-stimulated (gray circles, n = 40), or AID-deficient APC-specific (white circles, n = 27) clonal populations generated by immunization. (C and D) Frequency of clones that produced (C) the indicated subset combinations or (D) or any of the indicated subset. (E) Amount of APC staining of memory cells in clonal APC-specific populations that produced memory cells and plasma cells or memory cells but not plasma cells. The bars in (B) and (E) represent medians. P values were determined in (B) and (E) using a Mann-Whitney test and in (C) and (D) using Fisher’s exact test. Data points were combined from 17 MD4 experiments, 3 AID-deficient experiments, and 12 wild-type experiments.

Overall, our results demonstrate that individual naïve B cells vary greatly with respect to the number and types of effector cells generated early during the primary response, as reported for naïve T cells (1013). Analogous to CD4+ T cells (12), the precise effector cell subset pattern produced by a single naïve B cell was influenced by BCR affinity for antigen. Unlike T cells, however, many individual naïve B cells only produced a single type of effector cell, which was associated with Bim-mediated apoptosis. This situation could come about because naïve B cells are biased toward the production of only one subset (9). The progeny of different clones may then experience different levels of trophic signals from T cells or cytokine receptors. Some clonal populations may prematurely stop receiving these trophic signals, resulting in apoptosis of some of their members and cessation of further differentiation. Other cells that continued to receive trophic signals may be protected by these signals from apoptosis, allowing further proliferation and the generation of additional effector cell subsets. Together, the combination of extrinsic heterogeneity in trophic signals and intrinsic heterogeneity in BCRs expressed by the population of naïve B cells specific for an antigen ensures that a diverse set of effector cells types is produced during the primary response.

Supplementary Materials

Materials and Methods

Figs. S1 to S3

References (21, 22)

References and Notes

  1. Acknowledgments: We thank J. Walter, A. Quade, K. Anderson, S. Voght, B. Debuysscher, the Jenkins Lab, and the University of Minnesota Flow Cytometry Resource for technical assistance and helpful discussions; M. Jankovic, M. Nussenzweig, D. Liao, and G. Kelsoe for providing AID-deficient cells; and L. Manlove, M. Farrar, S. Roepke, and M. Prlic for providing Bim-deficient cells. The data presented in this manuscript are tabulated in the main paper and in the supplementary materials. This work was supported by the Irvington Fellowship Program of the Cancer Research Institute (J.J.T), and the National Institutes of Health [P01AI035296, R01AI036914, and R37AI027998 (M.K.J.)]. The authors have no conflicting financial interests.
View Abstract

Stay Connected to Science

Navigate This Article