Peripheral blood CD27⁺ IgG⁺ B cells rapidly proliferate and differentiate into immunoglobulin-secreting cells after exposure to low CD154 interaction

Abstract

In vitro CD40 stimulation of human B cells isolated from lymphoid organs is dominated by memory B cells undergoing faster proliferation and higher differentiation than naive B cells. In contrast, we previously reported that blood memory B cells mainly differentiate into immunoglobulin-secreting cells in response to CD40 stimulation. However, variations in CD40–CD154 interaction are now recognized to influence B-cell fate. In this study, we have compared the in vitro response of blood CD27⁻ and CD27⁻ IgG⁻ to CD27⁺ and CD27⁺ IgG⁺ B cells following low-density exposure to CD154 in the presence of a mixture of interleukin-2 (IL-2), IL-4 and IL-10. The evolution of these cell populations was monitored during initiation and following long-term stimulation. Over a 5-day period, CD27⁺ B cells underwent differentiation into immunoglobulin-secreting cells more readily than CD27⁻ cells, and CD27⁺ IgG⁺ B cells gave rise to a near homogeneous population of CD19⁺ CD27⁺⁺ CD38⁺ IgG^lo cells capable of high immunoglobulin G (IgG) secretion. During the same period, CD27⁻ IgG⁻ B cells partially became CD19⁺⁺ CD27⁻ CD38⁻ IgG⁺⁺ cells but showed no IgG secretion. Long-term stimulation revealed that CD27⁺ IgG⁺ B cells retained a high expansion capacity and could maintain their momentum towards differentiation over naive B cells. In addition, long-term stimulation was driving CD27⁻ IgG⁻ and total CD19⁺ B cells to evolve into similar CD27⁺ and CD27⁻ subsets, suggesting naive homeostatic proliferation. Overall, these results tend to reconcile memory B cells from blood and lymphoid organs regarding their preferential differentiation capacity compared to naive cells, and further suggest that circulating memory IgG⁺ cells may be intrinsically prone to rapid activation upon appropriate stimulation.

Introduction

In humans, naive and memory B cells constitute the pool of mature B cells found in blood and in secondary lymphoid organs, such as tonsils, lymph nodes, Peyer’s patches and spleen. From the bloodstream, B cells migrate to lymphoid organs and screen for antigens against which they will mount effective primary and secondary immune responses.

In blood, naive and memory B cells can be distinguished by their differential expression of CD27,^1–3 a membrane protein belonging to the tumour necrosis factor receptor family,⁴ CD27 expression is associated with the presence of somatic mutations in immunoglobulin variable genes generated in germinal centres.^1,2 CD27 is routinely used to distinguish CD27⁻ naive and CD27⁺ memory B cells, representing 60% and 40% of blood B cells, respectively.^2,5 However, small immunoglobulin A-positive (IgA⁺) and IgG⁺ memory B-cell subsets lacking CD27 expression are now believed to be part of the memory B-cell population.^6–9 In tonsils, germinal centre B cells become CD27⁺ without necessarily entering the memory compartment.^10–12

Several studies focusing on human splenic B cells have highlighted the distinct behaviour of CD27⁺ memory and CD27⁻ naive B cells following a T-cell-dependent stimulation through CD40–CD154 interaction in the presence of various combinations of interleukin-2 (IL-2), IL-4 and IL-10.^13–16 The competitive advantage of memory B cells over naive cells towards proliferation and immunoglobulin secretion is related to the capacity of IgM⁺ and IgG/IgA⁺ memory B cells to enter active cell cycling.^14,17 Furthermore, splenic memory B cells are committed to differentiate into CD27⁺ immunoglobulin-secreting cells more readily than naive cells,^13,15 whereas naive B cells require in vitro isotype switching before gaining the capacity for IgG secretion.¹⁶ All these observations underscore the involvement of CD40–CD154 interaction taking place in vivo between antigen-activated B and T cells, as it promotes proliferation, isotype switching, generation of memory B cells and immunoglobulin production.^18–20

In contrast with these findings, we previously observed that a high level of CD154 interaction drives blood memory B cells into differentiation, whereas only naive cells rapidly proliferate and differentiate in response to the same stimulus.²¹ However, we also reported that variations in CD40–CD154 signal intensity influence the proliferation and differentiation of human peripheral blood B cells²² and a recent study using a human B-cell line stimulated with variable levels of CD154 revealed differential capacities to engage alternative nuclear factor-κB pathways.²³ Overall, these studies suggest that the quantity and quality of CD154 lead to distinct functional B-cell responses (reviewed in refs ^{24, 25}) and could reflect differential effects on naive and memory B-cell proliferation and differentiation. We therefore investigated whether the in vitro response of blood memory B cells differed from that previously reported²¹ when using a lower level of CD154 signal intensity in the presence of IL-2, IL-4 and IL-10. As performed elsewhere for splenic B cells,^13–16 naive and memory B cells were isolated according to CD27 expression and submitted to conditions of low CD40 stimulation supplemented with a mix of IL-2, IL-4 and IL-10²² for short-term (5 days) and long-term (14 days) culture periods. Furthermore, sorted CD19⁺ CD27⁻ IgG⁻ naive B cells were directly compared with CD19⁺ CD27⁺ IgG⁺ memory B cells. Our results showed that CD40-activated blood CD27⁺ B cells more readily entered cell cycling during the first days of stimulation and showed higher differentiation into IgM- and IgG-secreting cells than CD27⁻ cells. After long-term activation, both CD27⁺ memory and CD27⁻ naive B cells expanded to comparable degrees but memory B cells showed a higher differentiation phenotype. In addition, CD27⁺ IgG⁺ cells showed the highest differentiation potential and were the fastest at entering the cell cycle and maintaining their momentum over naive B cells during long-term activation. These results showed that peripheral blood memory B cells proliferate and differentiate more readily than naive B cells following a low level of CD154 signal intensity.

Materials and methods

Peripheral blood B-cell isolation and cell sorting

This study has been reviewed and approved by the Héma-Québec Ethics Committee. Blood samples or leucoreduction filters from blood-collecting devices were obtained from healthy individuals after obtaining informed consent. B cells were isolated from peripheral blood mononuclear cells as previously described,^21,26 using the StemSep™ CD19 cocktail (Stem Cell Technologies, Vancouver, Canada). B-cell purity, as determined by flow cytometry, was higher than 95% in all experiments reported herein. Cell sorting of CD19⁺ B cells according to CD27 and IgG expression was performed using an Epics Coulter or an Epics Elite ESP (Beckman Coulter, Burlington, Canada), after staining with phycoerythrin (PE)-conjugated anti-CD27 and fluorescein isothiocyanate (FITC)-conjugated anti-IgG (the source of these conjugates is given below). All subsets were more than 93% pure and used immediately after sorting.

Human B-cell culture and exposure to defined culture conditions

Purified B cells were seeded at 0·75 × 10⁵ to 1·5 × 10⁵ cells/ml in Primaria plates (BD Biosciences, Mississauga, Canada) in the presence of γ-irradiated (75 Gy; 7500 rad) L4.5 cells expressing CD154.²⁷ A constant ratio of 25 B cells per L4.5 cell, which corresponds to about 1000 CD154 molecules per B cell as reported previously,²² was used in all assays. B cells were cultured in Iscove’s modified Dulbecco’s medium supplemented with 10% heat inactivated Ultra Low IgG fetal bovine serum (Gibco/BRL, Burlington, ON, Canada), 5·5 μg/ml transferrin, 6·7 mg/ml sodium selenite, antibiotics (all from Invitrogen, Burlington, Canada), 50 U/ml IL-2, 25 μg/ml IL-10 (PeproTech Inc., Rocky Hill, NJ) and 100 U/ml IL-4 (R&D Systems, Minneapolis, MN). If necessary, L4.5 cells were renewed every 4–5 days, and half of the culture medium was replaced every 2–3 days. Cell counts and viability were evaluated in triplicates using trypan blue dye exclusion. Generation time (t_gen) was calculated within the initiation phase of the growth curve according to the formula:

Flow cytometry and CFSE labelling

Peridinin chlorophyll protein-cyanin 5.5 (PerCP-Cy5.5)-anti-CD19, PE-anti-CD27, FITC-anti-IgG, allophycocyanin (APC)-anti-CD27, APC-anti-CD38 and PerCP-Cy5.5-, PE-, FITC- and APC-conjugated isotype controls were used in triple or quadruple staining procedures. All antibodies were mouse monoclonal IgG1 obtained from BD Pharmingen (Mississauga, Canada). Stainings and analyses were performed as described previously,^21,22 using a BD FACScalibur™ flow cytometer and the BD cellquest™ Pro software (BD Biosciences). Data were subsequently analysed using fcsexpress software (De Novo Software, Thornhill, Canada). Purified CD19⁺ cells were also stained with 5 μm carboxyfluorescein succinimidyl ester (CFSE; Molecular Probes, Eugene, OR) before being stimulated as described above. CFSE cell content was analysed on day 5 as a function of CD27 expression using APC-anti-CD27 and data were subsequently analysed using modfit lt software (version 3.0; Verity Software House, Topsham, ME) to determine the percentage of subsets within each daughter generation.

Cell proliferation assays

Cellular proliferation was monitored using 5-bromo-2′-deoxy-uridine (BrdU) incorporation from day 4 to 5 following the manufacturer’s instructions (Roche Molecular Biochemicals, Indianapolis, IN). Briefly, CD19⁺ blood B cells and sorted CD19⁺ CD27⁻, CD19⁺ CD27⁺, CD19⁺ CD27⁻ IgG⁻ and CD19⁺ CD27⁺ IgG⁺ were seeded in triplicate at 7·5 × 10⁴ cells/ml in flat-bottomed 96-well plates in the presence of 0·3 × 10⁴γ-irradiated L4.5 cells/ml and IL-2, IL-4 and IL-10 as described above. The γ-irradiated L4.5 cells did not incorporate BrdU.

Quantification of IgG and IgM secretion

Concentrations of IgG and IgM were determined in supernatants by a standard enzyme-linked immunosorbent assay.²¹ For the determination of IgG and IgM secretion rate in long-term assays, cells were harvested, extensively washed with phosphate-buffered saline and seeded at 1 × 10⁶ cells/ml in Iscove’s modified Dulbecco’s medium alone for 24 hr.

Reverse transcription–polymerase chain reaction (RT-PCR) amplification of human IgG transcript V_H regions, subcloning and sequencing

Total RNA from resting blood B cells, and from CD19⁺, CD19⁺ CD27⁻ and CD19⁺ CD27⁺ B cells cultured for 14 days in the presence of L4.5 cells and IL-2, IL-4 and IL-10, was isolated using TRIzol Reagent (Invitrogen).⁶ A strategy broadening the repertoire of amplified transcripts was adopted to overcome the previously observed bias for the large V_H3 family.²¹ Briefly, IgG messenger RNA was amplified with primers specific to V_H1, V_H2, V_H3, V_H4, V_H5 and V_H6 families (5′) (all described in ref. 28) paired with a Cγ-specific primer (3′) (5′-AAGTAGTCCTTGACC-3′). Using the Titan™ One Tube RT-PCR System (Roche Diagnostics Canada, Laval, Canada) and the following program (30 min reverse transcription at 50° followed by a 35-cycle PCR of 94° for 30 seconds, 45° for 40 seconds and 72° for 60 seconds, with a final elongation step at 72° for 10 min), 500–650 nucleotide amplicons were generated. All IgG V_H amplicons were resolved on 1% agarose gel electrophoresis, purified using the QIAquick Gel Extraction Kit (Qiagen, Mississauga, Canada), pooled and subcloned into the pDRIVE vector from the Qiagen PCR Cloning kit. Plasmid DNA was purified and inserts were sequenced by automated fluorescent DNA sequencing (ABI 373; Perkin-Elmer Applied Biosystems, Foster City, CA).⁶ The error rate of the Taq DNA polymerase mix was 0·34%, representing one mutation every 294 nucleotides.²¹

Mutational analysis of IgG V_H genes

Mutational analysis of V_H transcripts was performed using the international ImMunoGeneTics database (IMGT), which is available on the internet (http://imgt.cines.fr).²⁹ The distribution and number of mutations were determined after the alignment of each sequenced transcript with the germline gene presenting the highest homology from the FR1 to FR3 regions. The proportions of replacement (R) mutations in framework regions (FR) and complementarity determining regions (CDR), comprising FR1 (26 amino acids), FR2 (17 amino acids) and FR3 (39 amino acids) and CDR1 (12 amino acids) and CDR2 (10 amino acids), respectively, were calculated and used as an index of antigenic selection.

Results

Differential responses of CD27⁺ and CD27⁻ B cells to low levels of CD154 interaction

The in vitro evolution of peripheral blood naive and memory cells was compared using sorted CD27⁻ and CD27⁺ B cells (≥ 95% pure) stimulated for 14 days in conditions of low exposure to CD154 in the presence of IL-2, IL-4 and IL-10 (Fig. 1). The response of total CD19⁺ cells, used as control for the low CD154 stimulus, was consistent with our previous observations.²² During the first 5 days, the viability of all cell populations was higher than 90% (data not shown). Both CD27⁺ and CD27⁻ cells showed comparable expansions (approximately fivefold; Fig. 1a). This initial expansion phase was also characterized by IgG and IgM secretion, which was at least fourfold higher in CD27⁺ compared with CD27⁻ B cells (Fig. 1b). Similar levels of IgM and IgG secretion were observed within CD27⁻ cells; such a rapid IgG secretion within the CD27⁻ population suggested that the minor CD27⁻ IgG⁺ memory B-cell subset⁶ was rapidly induced to differentiate into IgG-secreting cells, as observed for CD27⁺ cells. Following long-term stimulation, the expansions of CD27⁻ and CD27⁺ B cells were comparable, albeit about twofold lower than that of CD19⁺ cells (Fig. 1a). On day 14, the viability of all populations had decreased but was still above 70% (data not shown). For all populations, IgM secretion was overwhelmed by that of IgG by up to eightfold (Fig. 1c). Noticeably, differentiation of CD27⁺ memory B cells was characterized by a high rate of IgG secretion, approximately threefold higher than CD27⁻ and CD19⁺ cells. Overall, these results showed that blood CD27⁺ B cells are more competent than CD27⁻ cells to differentiate into immunoglobulin-secreting cells following low CD40 interaction.

Figure 1.

CD27⁺ B cells are more efficient than CD27⁻ cells at immunoglobulin secretion. Total CD19⁺ and sorted CD19⁺ CD27⁻ and CD19⁺ CD27⁺ peripheral blood B cells were stimulated with CD154 in the presence of interleukin-2 (IL-2), IL-4 and IL-10. (a) Expansion was determined on days 5, 9 and 14 and (b) immunoglobulin G (IgG) and IgM concentrations were measured at the end of the 5-day culture period. (c) IgG and IgM secretion rates after 14 days of culture. (d) CFSE labelling and CD27 staining were used to track cell division of CD27⁺ (proliferation index 7·9) and CD27⁻ (proliferation index 9·2) populations within total CD19⁺ cells over a 5-day-period. These results are representative of three independent experiments, all performed with different blood samples. Each experiment was performed in triplicate. Error bars can be smaller than symbols and represent variations among triplicates in this experiment.

The proliferative advantage of splenic memory B cells over naive cells was reported in studies using incorporation of [H³]thymidine¹⁷ as well as CFSE labelling.¹⁴ The relative proliferation of blood memory and naive cells grown in conditions of low CD154 interaction was evaluated using CFSE labelling to track cell divisions as a function of CD27 expression (Fig. 1d). In agreement with the similar expansions of CD27⁺ and CD27⁻ cells observed on day 5, CFSE labelling indicated that CD27⁻ cells had a slight advantage over CD27⁺ cells, given that on day 5 about 95% of CD27⁻ cells had already undergone four or more generations compared to about 86% for CD27⁺ cells (Fig. 1d). In fact, a small fraction (approximately 15%) of CD27⁺ cells was lagging, while another (approximately 2%) was already at its sixth generation, suggesting heterogeneity within the CD27⁺ cell population itself. Differentiating cells are expected to be delayed in their cell cycle progression, so this small CD27⁺ population could represent differentiated cells sustaining a very high immunoglobulin secretion rate, whereas most of the cells could still be characterized by a high proliferation rate.

Heterogeneous evolution of blood CD27⁺ and CD27⁻ B cells

Based on the above observations, the evolution of CD27⁻ and CD27⁺ B cells was monitored using CD19, CD27 and IgG staining after exposure to low levels of CD154 stimulation for 5 and 14 days (Fig. 2). Total CD19⁺ cells were used as the control in parallel. After 5 days, a large proportion of CD27⁺ B cells maintained expression of this surface marker; within this population, the average CD27 expression was increased (Fig. 2a), and the proportion of surface IgG⁺ cells was comparable to that found on day 0 (Fig. 2b). At the same time, the phenotypic pattern of CD27⁺ cells became more heterogeneous, with populations that were CD19^lo IgG⁻ (approximately 10%), CD19⁺ IgG⁻ (approximately 50%) and CD19⁺ IgG⁺ (35%). These subsets could have distinct secretion and proliferation capacities responsible for the results described above (Fig. 1). After 5 days in similar culture conditions, fractions of cells that were initially CD27⁻ acquired expression of CD27 (24%) and surface IgG (14%). After 14 days, a majority of sorted CD27⁺ cells (62%) were differentiated into CD27⁺⁺ cells that were also mostly negative for surface IgG (Fig. 2c). When CD19⁺ cells were cultured in similar conditions, the phenotypic analyses revealed that both CD27⁺ and CD27⁻ cells could still be detected after 5 days, but the proportion of CD27⁺ cells was substantially reduced after 14 days. In fact, long-term stimulation of CD19⁺ and sorted CD27⁻ cells led to very similar phenotypic patterns regarding CD27 (27% for CD19⁺, and 25% for CD19⁺CD27⁻), IgG (63% for CD19⁺, and 69% for CD19⁺CD27⁻) and IgG⁺ CD27⁻ (43% for CD19⁺, and 49% for CD19⁺ CD27⁻) expression.

Figure 2.

Heterogeneous populations emerge from blood CD27⁺ and CD27⁻ B cells. CD19, CD27 and surface immunoglobulin G (IgG) expression was evaluated by flow cytometry (> 5000 events) of cultures started with total CD19⁺ and sorted CD19⁺ CD27⁻ (CD27⁻) and CD19⁺ CD27⁺ (CD27⁺) B cells. (a) CD19 and CD27 expression and (b) CD19 and surface IgG expression were measured on days 0, 5 and 14. (c) Phenotypic CD27 and IgG expression on day 14 are shown. These results are representative of two independent experiments and similar results were observed on day 5 in a supplemental short-term assay.

To verify whether the recently described CD27⁻ IgG⁺ memory cell population⁶ was still present within sorted CD27⁻ cells following long-term stimulation, or else had completely differentiated into IgG-secreting cells during the first 5 days of culture, we performed mutational analysis of IgG transcripts from resting blood CD19⁺ B cells and CD40-activated, sorted CD19⁺, CD27⁻ and CD27⁺ cells following 14 days of stimulation (Fig. 3). To discriminate mutated from unmutated immunoglobulin genes³⁰ and to exclude in vitro-generated mutations,^21,31 only transcripts with four or more mutated amino acids were considered as hypermutated. The analysis of IgG transcripts amplified from resting blood B cells, resulting from in vivo antigenic selection, indicated for each clone a higher frequency of R mutations in CDRs than in FRs (Fig. 3a). A majority (70%) of sequenced IgG transcripts amplified from sorted CD27⁺ cells carried 6–30 mutated amino acids, with a higher proportion of R mutations in CDRs than in FRs (Fig. 3d). In contrast, total CD19⁺ and CD27⁻ cells had 17% and 20%, respectively, of mutated IgG transcripts, but only 10% and 13%, respectively, presented mutations consistent with in vivo antigenic selection (Fig. 3b,c). These observations seemed consistent with the expansion of the minor IgG⁺ memory cell subset within the CD27⁻ population after long-term stimulation. The contribution of CD27⁺ IgG⁺ cells in long-term expanded CD19⁺ cells appeared rather modest because very few transcripts had mutations consistent with in vivo antigenic selection after 14 days of culture, even though CD27⁺ IgG⁺ cells outnumbered CD27⁻ IgG⁺ cells at the start of the culture. Such similarities between the resulting mutated IgG transcripts suggest that low-abundance CD27⁻ IgG⁺ cells were maintained during long-term expansion of CD40-activated CD19⁺ cells and CD27⁻ cells. Whether these long-term memory IgG⁺ cells emerged from the initial CD27⁺ IgG⁺ population or the minor CD27⁻ IgG⁺ memory cells remains to be investigated. Additionally, molecular as well as phenotypic similarities between total CD19⁺ and CD27⁻ populations suggest that these distinct populations tend to converge towards similar cell phenotypes following long-term stimulation. Overall, these results revealed heterogeneous subset evolution within sorted CD27⁻ and CD27⁺ populations.

Figure 3.

CD27⁻ IgG⁺ cells persist within the CD27⁻ cells following long-term stimulation. (a) Nine IgG transcripts from resting blood CD19⁺ B cells were amplified, sequenced and analysed for the presence of somatic mutations, in parallel with transcripts from (b) unsorted CD19⁺ B cells, (c) sorted CD27⁻ B cells and (d) sorted CD27⁺ B cells stimulated for 14 days with CD154, interleukin-2 (IL-2), IL-4 and IL-10. All characterized IgG transcripts from two donors (1A or 1B) are presented; the number of clones, their origin, and the number of mutated amino acids are indicated. The distribution of R (Replacement: ) and S (Silent; ) mutations is shown only when four or more mutated amino acids were observed. The proportions (%) of R mutations found in FR1, FR2 and FR3 (framework regions; cumulative length: 82 amino acids), and in CDR1 and CDR2 (complementarity determining regions; cumulative length: 22 amino acids) are shown for each sequence. *Clones considered to have arisen independently from antigen selection.

Blood CD27⁺ IgG⁺ memory B cells are highly efficient at expansion and immunoglobulin secretion

The above-described CD27⁺ B-cell population comprises CD27⁺ IgM⁺ IgD⁺ cells, which are associated with the splenic marginal zone, arising independently of germinal centres and considered to be responsible for T-cell-independent immune responses.^32–34 Indeed, the relationship of this B-cell subset with the memory pool is still under debate.³⁵ On the other hand, the above-described naive CD27⁻ cells include a subset of memory cells of phenotype CD27⁻ IgG⁺/IgA⁺.^7,9 Even though most studies regarding memory and naive B cells rely upon sorting according to CD27 expression,^13–16 we decided to refine our investigations to specifically target CD27⁺ IgG⁺ memory B cells. Therefore, blood B cells (CD19⁺) were sorted into CD19⁺ CD27⁻ IgG⁻ and CD19⁺ CD27⁺ IgG⁺ cells with purities ranging from 93 to 100% (Fig. 4; day 0). The two B-cell subsets were stimulated in conditions of low levels of CD154 interaction and their proliferation and differentiation were monitored after 5 and 13 days of culture (Table 1). During the first 5 days, the average t_gen of CD27⁺ IgG⁺ cells was 33 hr, compared to 72 hr for CD27⁻ cells, and BrdU incorporation revealed an increased DNA synthesis for CD27⁺ IgG⁺ cells. In agreement with these observations, expansion of CD27⁺ IgG⁺ cells on day 5 was approximately fourfold higher than that of CD27⁻ IgG⁻ and CD19⁺ cells (data not shown). During this same time-frame, IgG concentration from secreting CD27⁺ IgG⁺ cells reached very high levels (> 20 μg/ml). Meanwhile, CD27⁻ IgG⁻ naive cells were secreting negligible amounts of IgG, whereas IgM concentration (460 ng/ml, data not shown) was similar to that of sorted CD27⁻ cells (Fig. 1), suggesting that in vitro isotype switching to IgG occurred at a later time-point.

Figure 4.

CD27⁺ IgG⁺ cells give rise to CD38⁺ IgG⁺ and CD38⁻ IgG^lo cells. Total CD19⁺ and sorted CD19⁺ CD27⁻ IgG⁻ and CD19⁺ CD27⁺ IgG⁺ B cells, stimulated as described in Table 1, were analysed after 5 days in culture for CD19, CD27 and immunoglobulin G (IgG) expression by flow cytometry (> 500 events). Cell purity before stimulation is shown (d = 0). The proportion of IgG⁺ cells as a function of CD27 expression is presented for each subset. These results are representative of two independent experiments.

Table 1.

Peripheral blood CD27⁺ IgG⁺ B cells show the highest proliferation and immunoglobulin G secretion rate

	Short term2			Long term3
Phenotype1	tgen (hr)	BrdU (OD 450 nm)	IgG ng/ml	Expansion (fold)	IgG ng/106 cells/hr
CD19+	61 ± 3	1·35 ± 0·06	1114 ± 100	151 ± 9	1519 ± 65
CD27− IgG−	72 ± 5	1·34 ± 0·18	14 ± 1	46 ± 1	906 ± 68
CD27+ IgG+	33 ± 2	1·84 ± 0·07	21548 ± 1044	300 ± 18	2126 ± 3

t_gen, generation time; IgG, immunoglobulin G; BrdU, 5-bromo-2′-deoxy-uridine.

¹Total CD19⁺ and sorted CD19⁺ CD27⁻ IgG⁻ and CD19⁺ CD27⁺IgG⁺ B cells were stimulated with CD154 in the presence of interleukin-2 (IL-2), IL-4 and IL-10. These results are representative of two independent experiments. SD represented variations among triplicate values obtained in this particular experiment.

²t_gen and total IgG accumulations were determined over the first 5 days, and BrdU incorporation was measured from day 4 to 5. Viability of all the cells remained > 93% (data not shown).

³Expansion and IgG secretion were determined on day 13.

In contrast to the entire population of CD27⁺ cells described in Fig. 1, CD27⁺ IgG⁺ memory cells maintained a higher expansion than CD27⁻ IgG⁻ B cells during long-term stimulation (Table 1). This preferential expansion was, however, mainly related to the momentum acquired during the initial phase of the culture (data not shown). Furthermore, in agreement with previous observations²² and the results of Table 1, IgG secretion on day 13 overwhelmed IgM secretion by 10-fold and 20-fold in CD27⁻ IgG⁻ and CD19⁺ cells, respectively (data not shown). Overall, during short-term activation, the proliferation and differentiation of CD27⁺ IgG⁺ memory B cells were both higher than that of naive CD27⁻ IgG⁻ cells. Moreover, CD27⁺ IgG⁺ cells were able to maintain their superiority over naive cells following long-term stimulation, as shown by their twofold higher IgG secretion rate on day 13, which further illustrates their commitment towards differentiation into immunoglobulin-secreting cells.

CD27⁺ IgG⁺ cells evolve towards CD38⁺ IgG⁺ and CD38⁻ IgG^lo cells

To investigate the short-term evolution of CD40-activated CD27⁻ IgG⁻ and CD27⁺ IgG⁺ cells, the expression of CD19, CD27, CD38 and IgG was determined and compared to that of CD19⁺ peripheral blood B cells (Fig. 4). After 5 days of culture, about 90% of CD27⁺ IgG⁺ cells showed slightly increased CD27 expression, characterized by a mean fluorescence intensity (MFI) of 95 compared to 60 on day 0. In contrast, the proportion of IgG⁺ cells (Fig. 4; 55%), as well as the level of surface IgG expression had decreased, from an initial MFI of 190 to an average of 40. After short-term activation, CD27⁺ IgG⁺ cells could be subdivided into CD19⁺ CD27⁺⁺ CD38⁻ IgG⁺, and more differentiated CD19⁺ CD27⁺⁺ CD38⁺ IgG^lo cells associated with an elevated IgG secretion rate (Table 1). During the same 5-day period, only a small fraction of CD27⁻ IgG⁻ B cells expressed high surface IgG (6%; MFI approximately 100). Additionally, most cells (> 80%), including those IgG⁺ cells, maintained a high expression of CD19⁺ (MFI approximately 200) but remained negative for CD27 and CD38. Of note, even though IgG⁺ cells were readily detected within this cell population, IgG secretion was still negligible (Table 1). Similarly, CD27⁻ IgG⁺ cells (approximately 6%) were present among CD40-activated CD19⁺ cells and can therefore be associated with cells differentiated in vitro and characterized by the CD19⁺⁺ IgG⁺⁺ CD27⁻ CD38⁻ phenotype. As described above (Fig. 2a), based on the phenotypes of B cells according to IgG, CD27 and CD38 expression, populations of both CD27⁻ IgG⁻ and CD27⁺ IgG⁺ cells seemed to be concomitantly present within the total CD19⁺ cell population on day 5.

CD27⁺ IgG⁺ cells differentiate into CD19^lo CD27⁺⁺ CD38⁺⁺ IgG⁻ cells following long-term stimulation

To further investigate the in vitro differentiation of the above-described B-cell populations, the phenotypes of CD40-activated CD27⁻ IgG⁻, CD27⁺ IgG⁺ and total CD19⁺ B cells were determined as above after 14 days of culture (Fig. 5). Such long-term activation of CD27⁺ IgG⁺ cells resulted in a fairly homogeneous phenotype characterized by high levels of CD27 (MFI approximately 300) and CD38 (MFI approximately 200), low levels of CD19 (MFI approximately 30) and loss of surface IgG. Of note, these cells showed intracellular IgG and were negative for surface IgM expression (data not shown) and their homogeneous phenotype CD19^lo CD27⁺⁺ CD38⁺⁺ IgG⁻ is consistent with a high differentiation status towards plasma cells.^36–39

Figure 5.

Long-term stimulation drives CD27⁺ IgG⁺ cells into plasma cell differentiation. Unsorted blood B cells (CD19⁺) and sorted CD19⁺ CD27⁻ IgG⁻ and CD19⁺ CD27⁺ IgG⁺ populations were stimulated with CD154, interleukin-2 (IL-2), IL-4 and IL-10 for 14 days and monitored for expression of CD19, CD27, CD38 and surface immunoglobulin G (IgG) by flow cytometry (> 10 000 events). All sorted populations were ≥ 94% pure. On day 0, the frequencies of CD27⁺, CD27⁻, CD27⁺ IgG⁺, CD27⁻ IgG⁺ and CD27⁻ IgG⁻ cells within CD19⁺ cells were 21%, 79%, 6%, 4%, and 75%, respectively. The level of CD38 expression (mean fluorescence intensity) on resting B cells varied from 7 to 27. These results are representative of two independent experiments.

Consistent with the above results (Figs 2 and 3), CD27⁻IgG⁻ and total CD19⁺ cells evolved similarly, giving rise to two major populations, CD19⁺ CD27⁻ and CD19^lo CD27⁺ cells, representing about 40% and 60% of total cells, respectively (Fig. 5). In both cases, the CD19⁺ CD27⁻ subset showed low expression of CD38 (MFI approximately 20) and was divided between IgG⁺ and IgG⁻ cells (data not shown). A similar distribution between IgG⁺ and IgG⁻ cells was also observed for the CD19^lo CD27⁺ cell subset, which was characterized by a high level of CD38 expression, with MFIs of approximately 100 (IgG⁺) and approximately 170 (IgG⁻), respectively (data not shown). The unique phenotype acquired by long-term activated CD27⁺ IgG⁺ cells appeared almost absent from these two populations. Once more (day 14; Fig. 2), CD19⁺ and CD27⁻ IgG⁻ populations led to similar phenotypes following long-term stimulation, suggesting again a tendency to reach a common equilibrium within these B-cell populations. In addition, these results also suggested that the memory CD27⁺ cells initially present within CD19⁺ cells were unable to expand significantly following long-term CD40 stimulation, such that most CD27⁺ cells detected at the end of the culture arose from a CD27⁻ population.

Discussion

In this study, we showed that blood CD27⁺ IgG⁺ memory B cells have a preferential growth advantage over CD27⁻ IgG⁻ naive B cells, as demonstrated by their enhanced capacity to rapidly enter into the active cell cycle and proliferate. Furthermore, CD27⁺ IgG⁺ cells rapidly differentiated into immunoglobulin-secreting cells during the initial phase of the culture and maintained this characteristic following long-term CD40 stimulation, reaching a fairly homogeneous CD19^lo CD38⁺⁺ CD27⁺⁺ cell population resembling the plasma cell phenotype.^13,40 Accordingly, such higher proliferation and IgG secretion of blood CD27⁺ IgG⁺ cells are in agreement with observations made for CD27⁺ splenic B cells using a cell-bound¹⁷ or soluble membrane^13–16 source of CD154 to activate CD40 in the presence or absence of cytokines. In vivo, higher and sustained expansion of CD27⁺ IgG⁺ memory cells following limited CD40 stimulation could be further enhanced by activation through the IgG cytoplasmic tail, which is able to sustain a division burst and B-cell survival during T-cell-dependent responses.⁴¹

Our findings underscore that enhanced activation of CD27⁺ IgG⁺ memory cells was observed in conditions of low levels of CD40 stimulation and in the presence of IL-2, IL-4 and IL-10, and these results are consistent with our previous observations.²² These observations also suggest that the proliferative advantage of naïve over memory B cells as observed before²¹ was related to the higher level of CD154 interaction (approximately 10-fold) in the presence of IL-4 or IL-10 alone. In vivo, memory B cells are believed to preferentially interact with memory T cells,⁴² which can indeed secrete these three cytokines simultaneously.⁴³ Several studies in mice and humans indicate that T-cell activation induces surface expression of CD154, and that this expression is transient and highly regulated.^44–47 Both naive and memory T cells are induced to express high levels of surface CD154, but only within hours following cellular activation (reviewed in ref. 48). In humans, it has been reported that tonsil memory T cells, which have preformed CD154 molecules, can express low levels of CD154 on their cell membrane within 5–15 min after cellular activation.⁴⁶ Low levels of CD154 expression have also been reported in subsets of human memory T cells⁴⁷ as well as in naive CD4⁺ T cells in mouse models.⁴⁹ In mice, memory T cells have lower and transient expression of CD154⁴⁴ and can also have preformed CD154 molecules.⁴⁵ Consequently, our results suggest that IgG⁺ memory B cells might be able to respond to a second antigenic challenge even when CD154 expression is limited on activated memory T cells found in the outer zone of germinal centres and margins of the T zone.⁴⁶ This response to low levels of CD154 interaction might also be relevant in vivo to the highly motile germinal centre B cells, whose access to helper T cells in the light zone is very limited in time.⁵⁰

In addition, the fact that short-term stimulation of sorted CD27⁻ IgG⁻ cells led to an almost exclusive IgM secretion suggests that CD27⁻ IgG⁺ memory cells within CD40-activated CD27⁻ cells were probably responsible for most of the observed IgG secretion during the same period (Fig. 1b) and so have differentiation capacities similar to those of CD27⁺ IgG⁺ cells. In addition, the presence of about 13% IgG⁺ cells carrying mutated V_H genes within sorted CD27⁻ cells indicates that these CD27⁻ IgG⁺ cells were proliferating and stably maintained following long-term activation, suggesting a strong renewal capacity for this subset. These observations support the notion that this small subset belongs to the memory B-cell compartment. Whether the IgA⁺ cells, representing about 2% of the CD27⁻ populations^7,9 and therefore also present within the CD27⁻ IgG⁻ cells, could follow the same pathway as their IgG⁺ counterpart or influence the other populations remains to be investigated.

B-cell differentiation towards immunoglobulin-secreting cells is linked to an arrest in cell cycle progression.^51,52 Therefore, differentiation and proliferation are opposite manifestations of B-cell maturation underlying divergent responses among CD27⁺ IgG⁺cells. It has also been suggested that following a second antigenic challenge, memory B cells divide into undifferentiated blasts to maintain the memory pool as well as antibody-secreting plasma cells.⁵³ Indeed, results presented in this study showed that CD40-activated CD27⁺ IgG⁺ memory cells rapidly evolved into two main populations, namely CD19⁺ CD27⁺⁺ CD38⁻ IgG⁺ and CD19⁺ CD27⁺⁺ CD38⁺ IgG^lo cells, which can be respectively associated with cells in transition towards the memory pool¹³ and towards differentiation into plasma cells.⁴⁰ Therefore, the high and sustained proliferation observed within activated CD27⁺ IgG⁺ cells could originate from highly proliferative CD19⁺ CD27⁺⁺ CD38⁻ IgG⁺ cells, while immunoglobulin secretion could be derived from more differentiated CD19⁺ CD27⁺⁺ CD38⁺ IgG^lo cells. However, at the end of the long-term stimulation, a trend favouring plasma cell differentiation could be observed, suggesting that continuous CD40 activation might have exhausted their regeneration capacity towards the memory cell pool.

The results of this study further indicate that long-term CD40 stimulation of sorted CD27⁻ IgG⁻ and total CD19⁺ B cells generated similar phenotypic patterns, implying that CD27⁺ populations did not have a selective growth advantage over unsorted CD19⁺ cells. The negative impact of CD27⁻ cells on CD27⁺ cell expansion could be mediated by homotypic interactions involving costimulatory molecules such as CD27 and CD70.⁵⁴ In fact, long-term stimulation of CD19⁺ B cells leads to the emergence of both CD27⁺ as well as CD70⁺ cells (data not shown). The CD27–CD70 interaction induces reciprocal activation (reviewed in refs ^{54, 55}); the binding of CD27 promotes differentiation of CD27⁺ cells⁵⁶ while CD70⁺ cells are driven into the cell cycle.⁵⁷ Therefore, CD27⁺ cells could enhance their differentiation when interacting with CD70⁺ cells emerging in response to CD40 activation, and in turn CD70⁺ cells could increase their proliferation. Such interactions could also allow CD27⁻ IgG⁺ cells to reach higher expansion than CD27⁺ IgG⁺ cells within the CD19⁺ population, as long as these cells were not induced to express CD27. This hypothesis is currently under investigation.

Similar phenotypic patterns to sorted CD27⁻ IgG⁻ and total CD19⁺ B cells also suggest a possible connection with homeostatic B-cell proliferation.^58,59 Homeostasis is important in maintaining appropriate numbers of immune cells in the periphery and in lymphoid organs (reviewed in refs ^{60, 61}). Homeostatic proliferation of naive B cells has been previously observed in response to a B-cell deficit in a mouse model.^58,62 In humans, homeostatic proliferation of naive B cells (reviewed in refs ^{63, 64}) has been highlighted within peripheral blood B cells by examining their replication history.⁵⁹ In our study, enhanced proliferation in response to a B-cell deficit generated through cell sorting could be involved in long-term stimulation of CD27⁻ IgG⁻ cells. On the other hand, the involvement of balanced proliferation within the entire CD19⁺ cell population could be related to intrinsic characteristics of naive B cells or induced following the progressive decline of those memory B cells initially present in the culture. Whether CD27–CD70 interactions drive both CD27⁻ IgG⁻ and CD19⁺ cells to reach the observed final subsets deserves further investigation.

In light of these observations, it could be speculated that in the presence of memory T cells expressing low levels of CD154, blood CD27⁺ IgG⁺ memory B cells entering lymphoid organs would have a selective advantage compared to CD27⁻ IgG⁻ naive B cells for proliferation, differentiation, and immunoglobulin secretion. Taken together, our findings also suggest that memory B cells from blood and lymphoid organs have comparable superiorities over naive B cells in terms of proliferation and differentiation. Finally, our in vitro observations regarding the long-term equilibrium reached by naive and CD19⁺ B cells could provide supplemental clues regarding communications between naive and memory B cells and homeostatic proliferation.