|Home | About | Journals | Submit | Contact Us | Français|
Antibody responses to T cell-independent type 2 (TI-2) antigens (Ag), such as bacterial capsular polysaccharides, are critical for host defense. In mice, B-1b cells expressing a CD11b+FSChiCD21lo/-CD19hi phenotype play a key role in producing antibodies against T cell-independent type 2 (TI-2) antigens (Ag). In primates, a distinct IgM+CD27+ “memory” B cell population is thought to generate TI-2 antibody responses and evidence for a B-1b-like cell population participating in these responses is lacking. Herein, we demonstrate that nonhuman primates (NHP; African green monkeys and cynomolgus macaques) harbor serosal B cells expressing a CD11b+FSChiCD21lo/-CD80+/-CD19hi phenotype, constitutively active Stat3, and increased reactivity with phosphorylcholine, similar to murine peritoneal B-1a and B-1b cell populations. Similar to what is observed for murine B-1b cells, NHP CD11b+FSChiCD21lo/-CD19hi B cells dominate the Ag-specific B cell response and antibody production against the TI-2 Ag, TNP-Ficoll. Although Ag-specific IgM+ B cells expressing CD27 were not detected prior to immunization, Ag-specific CD11b+CD19hi B cells expressed and maintained an IgM+IgDloCD27+CD80+ phenotype following immunization. Thus, the murine and NHP B cell populations responding to TNP-Ficoll are highly similar, with the main exception being that Ag-specific NHP B-1-like cells express CD27 following TI-2 Ag encounter. Therefore, murine B-1b and primate IgM+CD27+ “memory” B cell subsets proposed to produce TI-2 antibody responses may be highly related if not identical. Overall, these data not only support that B-1-like cells are present in NHP, but provide evidence that these cells perform the same functions attributed to murine B-1b cells.
The murine B-1 cell compartment is comprised of phenotypically and functionally distinct B cell subsets important for host defense and immune regulation (1, 2). B-1a (CD5+) and B-1b (CD5-) cells display a unique phenotype (CD11b+CD21loCD23loCD19hiIgMhi), a preferential localization to serosal cavities and omentum, and derive from a progenitor that is distinct from that which gives rise to conventional (B-2) cells (3). Rothstein and colleagues have recently presented evidence for a B-1a-like population in human peripheral blood exhibiting a CD20+CD27+CD43+CD70- phenotype with the capacity for spontaneous IgM secretion, T cell stimulation, heightened tonic intracellular signaling, and typical murine B-1a specificities (4, 5). Despite these findings, the existence of B-1 cells in humans has remained a matter of substantial controversy (6-9). Moreover, evidence for human B cells with the functional and phenotypic characteristics of B-1 cells present in tissues typically enriched in B-1 cells in mice (ie., serosal cavities and omentum) is lacking.
Murine B-1a and B-1b cells are distinct, as they have different developmental requirements (10), differential responsiveness to Ag receptor signaling (11), and perform unique functions in the immune system (1). B-1a cells play a major role in producing natural Abs important for homeostasis and immune defense (1, 12), but may also participate in Ag-specific Ab responses (13, 14). Murine B-1b cells appear to serve a more critical role in producing Abs in response to classical TI-2 Ags such as pneumococcal polysaccharides (PPS), α1,3 dextran, and haptenated Ficoll (10, 15-17) as well as other TI Ags (18-20). It is clear that human B cells can produce Abs against the same Ags and pathogens that elicit murine B-1 cell responses (10, 18, 21, 22). However, a TI-2 Ab-producing B-1b-like subset is generally not thought to exist in primates (23). Instead, IgM+CD27+ “memory” B cells have been proposed to generate TI-2 Ab responses in primates (24-27). Although IgM+CD27+ B cells express mutated Ag receptors, it has been argued that they may not be true memory cells but that they have undergone a process of Ag-independent somatic hypermutation during developmental repertoire diversification (26). Despite the controversy surrounding the origin, functions, and memory status of IgM+CD27+ “memory” B cells (27), recent studies nonetheless support a role for CD27+ B cells in either producing IgM and IgG against PPS (22, 25) or increasing in frequency following PPS immunization in humans (28). Human IgM+CD27+ memory cells have therefore been proposed to perform the functions of murine B-1 cells (25); however, the relationship between these cells and murine B-1a and B-1b cells is not clear.
Evidence for primate B cells exhibiting preferential localization within serosal cavities with additional features characteristic of murine B-1 cells is currently lacking. Moreover, the extent to which a primate B-1-like subset participates in antibody responses to TI-2 Ags is unexplored. To definitively assess whether primates have a B cell subset that is both phenotypically and functionally similar to murine B-1b cells, it is necessary to perform side-by-side analyses of murine and primate Ag-specific B cell subset (cellular) responses to a well-defined TI-2 Ag that neither species has previously encountered. This type of study has not yet been performed. These types of studies are challenging in humans since 1) identifying human subjects that are naïve to bacterial polysaccharides justified for use in human immunizations is difficult, and 2) B cell subset analysis is often limited to peripheral blood cells. To bypass these limitations, we assessed multiple tissues in NHP for the presence of B cells with phenotypic and functional characteristics shared with murine B-1a and B-1b cells. As part of this analysis, we assessed the functional role of NHP B-1 cells in the response to the synthetic TI-2 Ag, TNP-Ficoll, which we have previously demonstrated to yield a robust Ag-specific B-1b cell response in normal mice (15). Collectively, our results reveal striking similarities between primate and murine B-1 cells, including a conserved functional role for primate B-1-like cells in immunity to TI-2 Ags.
African green monkeys (AGM) and cynomolgus macaques were from Wake Forest Primate Center breeding colonies. Male cynomolgus macaques and AGM ~3 yr of age were immunized i.v. with 0.25 mg/kg TNP65-Ficoll (Biosearch Technologies) in 400-500 μl saline.. Studies were approved by the Wake Forest University Animal Care and Use Committee.
Blood was collected in heparinized tubes and centrifuged at 400 × g for 10 min. Plasma was removed and remaining blood was lysed using 10-20 volumes of RBC lysis buffer (0.15 M NH4Cl/0.01M KHCO3), followed by centrifugation and washing in PBS containing 2% bovine calf serum (PBS-2% BCS). Omental tissue, spleen, and axillary lymph nodes were processed using glass-ground homogenizers. Peritoneal cells were isolated by lavaging the peritoneal cavity with 40-150 ml PBS. Spleen homogenate pellets were lysed in RBC lysis buffer followed by washing. Cells were filtered through 80 μm mesh and in some cases, further purified using Histopaque 1077 (Sigma) according to manufacturer's instructions. Single cell blood, spleen, lymph node and peritoneal cavity leukocyte suspensions (2 × 107/ml) were incubated in PBS-2% BCS with 20 μg/ml TNP30-Ficoll-Fluorescein8 (Biosearch Technologies) for 30 min. at room temperature, followed by subsequent staining with fluorochrome-labeled mAbs on ice for 25 min as previously described (15). Biotinylated- or fluorochrome-conjugated Abs used included: goat anti-monkey IgM and IgG (Fitzgerald Industries); goat anti-human IgD (Southern Biotechnology Associates, Inc.); goat anti-monkey Ig (H+L) (AbD Serotec); CD5 (MA1-82405; ThermoScientific), monkey/human CD19 (J3.119; Beckman Coulter and NHP Reagent Resource); CD11b (M1/70), CD20 (2H7), CD21 (LT21), and PD1 (EH12.2H7) all from Biolegend; CD27 (LG.7F9) and CD20 (2H7) from eBioscience; CD86 (FUN-1) and CD80 (L307.4) from BD Pharmingen; and CD38 (OKT10, NHP Reagent Resource). PC-reactive B cells were identified as previously described (4), which involved incubating cells with 20 μg/ml PC-Fluorescein-BSA (Biosearch Technologies) along with CD19- and CD11b-fluorochrome labeled mAbs at room temperature for 25 min. followed by washing and analysis. pY705-Stat3 (4/P-STAT3, BD Pharmingen) and Ki-67 (SolA15; eBioscience) intracellular staining were performed according to manufacturer's instructions. The following Abs did not react with cynomolgus macaque or AGM cells (CD43; clones 1G10, 84-3C1, and DTF1), CD138 (M115 and DL-101), CD70 (Ki-74), CD1c (AD5-8E7). Cells were analyzed using FACSCalibur and FACSCantoII flow cytometers (Becton Dickinson, San Jose, CA). Doublets/triplets were excluded using FSC-A/FSC-H gating on the CantoII instrument as previously described (15). Positive and negative cell populations were determined using unreactive isotype-matched Abs (Biolegend, eBioscience, and BD Pharmingen). Additional staining controls for Ag-specific B cells used TNP30-Ficoll-Fluorescein8 and CD20 APC-eFluor780 mAb (eBioscience) along with the appropriate PE-, APC-, PercpCy5.5-, and PE-Cy7-labeled isotype controls. Data was analyzed using FlowJo analysis software (Treestar).
ELISAs were as described (15). Plasma was harvested as described above and stored at -80 °C. Prior to ELISA, samples were thawed and diluted in TBS containing 1% BSA (Sigma). TNP-specific Ab levels were measured by adding diluted plasma samples to plates that had been coated with 5 μg/ml TNP-BSA (Biosearch Technologies) in 0.1 M borate buffered saline overnight at 4°C. Alkaline phosphatase (AP)-conjugated polyclonal goat anti-monkey IgG (Fitzgerald Industries) or biotinylated goat anti-monkey IgM (Fitzgerald Industries) or biotinylated goat anti-monkey IgA Abs (AbD Serotec) followed by streptavidin-AP (Southern Biotechnology Associates) and pNPP (Sigma) were used to detect TNP-specific Ab.
ELISPOTs were performed on single cell suspensions from splenocytes, axillary lymph nodes, bone marrow cells or CD11b- and CD11b+ PBMCs sorted using Miltenyi beads according to manufacturer's instructions (Miltenyi Biotec). Purity and B cell (CD20+) frequencies of sorted populations were assessed by flow cytometry as described above with CD11b expression by B (CD20+) cells within purified CD11b+ and CD11b- populations detected by anti-CD11b (M1/70)-PE combined with anti-rat Ig-PE (to detect bound unlabeled CD11b mAb associated with beads). ELISPOT 96-well plates (Immobilon P, Millipore) were precoated with TNP-BSA (5 μg/ml) in PBS overnight at 4°C, washed 2 times with PBS, and blocked 1 hr at 37°C with cRPMI containing 10% FCS (Gibco BRL). Cells were plated at concentrations ranging from 105-107 cells/ml in cRPMI containing 10% FCS and cultured 18-20 hrs. Polyclonal goat anti-monkey IgM and IgG Abs (Fitzgerald) described above were used in conjunction with NBT/BCIP substrate (Promega, Madison, WI) to develop plates according to manufacturer instructions. Membranes were dried and spots were enumerated.
Differences between sample means were assessed using Student's t-test. Results are shown as mean values (±SEM) unless otherwise indicated.
CD11b is expressed on murine B-1a (CD5+) and B-1b (CD5-) cell subsets in serosal cavities and omentum, but is diminished upon exit from these tissues (1, 29). CD11b expression is therefore used to distinguish B-1 cell populations from conventional B cells in mice, although identification of CD11b- B-1 cells requires the use of additional markers (30). We assessed CD11b expression on NHP B cells from multiple tissues to examine the possibility of B-1-like cells in NHP that share the same phenotype and distribution pattern as murine B-1 cells. CD19 was used in combination with CD20 or Ig (H+L) detection to identify NHP B cells. An example of the gating strategy used to identify these cells is shown in Fig. 1A. CD11b was expressed by a substantial fraction of B cells isolated from peritoneal lavage and omentum of AGM (Fig. 1A). B cells expressing lower levels of CD11b were also present in blood, but few spleen B cells (<10%) expressed CD11b. Similar results were found for cynomolgus macaques (not shown). Cellular yields obtained from NHP peritoneal lavages and processed omentum were typically ~1-2 ×106 cells/animal. B cell frequencies in peritoneal cavities were typically <10%, and ~25-40% of these B cells expressed CD11b (Fig. 1B). Interestingly, a significant increase in the frequency of peritoneal B cells expressing CD11b occurred between 1-90d and 1.5-3 yrs of age, despite comparable total peritoneal B cell frequencies among these age groups (Fig. 1B).
Consistent with what has been described for mouse B-1 cells, primate CD11b+ B cells expressed higher FSC, IgM, and CD19, and lower levels of CD21 relative to CD11b- B cells (Fig. 1C). Notably, CD27 was only expressed on a fraction of CD11b+ B cells (<50%) in peritoneal cavity, spleen and blood (Fig. 1D), suggesting that this marker alone cannot be used to define primate B-1 cells. The percentage of AGM and cynomolgus spleen and blood B cells found to express CD27 is consistent with a previous report for cynomolgus macaques of a similar age (31), although older rhesus macaques were reported to have increased frequencies (32) likely owing to increased memory B cell accumulation with age. CD5+ B cells were present in spleen, blood, peritoneal cavity, and omentum (Fig. 1E), but did not correlate with CD11b expression, even in serosal tissues. Thus, CD5 alone is not a reliable marker for defining the B-1a subset in NHP as is true in other species (33) and CD27 is not expressed on all primate peritoneal B-1 cells that are defined by CD11b expression.
Consistent with previous findings for murine B-1 cells (34), primate and murine peritoneal CD11b+ B cells expressed constitutively active Stat3, whereas CD11b- peritoneal B cells and spleen B cells did not (Fig. 1F). Moreover, the frequency of PC-binding B cells, a specificity enriched in murine B-1a cells (33), was increased in the peritoneal cavities of NHP relative to spleen and these cells largely belonged to the CD11b+ B cell subset (Fig. 1G). Finally, because some murine peritoneal B-1 cells constitutively express CD80, we examined CD80 expression by peritoneal AGM B cells. Similar to what is observed in mice, a fraction of CD11b+, but not CD11b-, peritoneal B cells expressed CD80 (Fig. 1H). Thus, primate CD11b+ B cells exhibit a distribution pattern and surface phenotype that is similar to murine B-1 cells (FSChiSSChiCD11b+CD19hiIgMhiCD21lo/-CD80+/-), with variable CD27 expression and delayed development relative to CD11b- B cells. This, along with the finding that NHP peritoneal CD11b+ B cells constitutively express active Stat3 and exhibit enhanced reactivity with PC strongly supports the existence of a functional peritoneal B-1a-like population in higher primates. Moreover, these results demonstrate that, like mouse peritoneal CD11b+ B-1 cells which can be divided into B220loCD21lo/-CD5+ B-1a cells, and B220hiCD21int and B220loCD21lo CD5- B-1b cell subsets (11), heterogeneity exists within the primate B-1 cell pool with respect to the expression of markers such as CD21, CD80, and CD5.
To determine whether NHP B-1-like cells are functionally responsive to TI-2 Ags, and thereby perform functions of murine B-1b cells, we examined NHP Ag-specific B cell responses to TNP-Ficoll, a classical synthetic TI-2 Ag which induces murine Ag-specific B-1b cells to expand, differentiate, and produce TNP-specific Ab, regardless of whether Ag is delivered i.p. or i.v. (15) or whether MZ B cells are present (11). We have validated the use of this model system to track Ag-specific B cell responses to TNP-Ficoll in mice (11, 15). TNP-binding B cells are found within all murine B cell subsets; however, B-1b cells predominantly respond to TNP-Ficoll immunization (15), similar to what has been reported for NP-Ficoll (16).
As we have found in mice, TNP-Ficoll-binding B cells are present in cynomolgus macaques and AGM prior to immunization (Fig. 2A). TNP-Ficoll immunization significantly increases the number of Ag-specific blood B cells (d5; Fig. 2A and Supplemental Figure 1). Relative to non-Ag-specific B cells and d0 Ag-specific B cells, Ag-specific B cells from immune animals (d5) had significantly increased FSC, SSC, PD-1, and CD86 expression and decreased CD21 (Fig. 2B-C). IgM expression was also decreased, possibly due to Ag-induced receptor internalization. Importantly, TNP-Ficoll immunization resulted in significant increases in the frequencies of Ag-specific B cells exhibiting an FSChiCD11b+ phenotype (Fig. 2C-D). These cells co-expressed the PD-1+CD86+CD21lo phenotype. These phenotypic changes also occurred in Ag-specific AGM blood B cells 5 days following TNP-Ficoll immunization and are similar to that which is observed with the Ag-specific B-1b cell response to TNP-Ficoll in mice (Supplemental Figure 2A-B and data not shown and ref.(15)).
Ag-specific CD20lo B cells expressing a CD11b+/CD21lo phenotype significantly increased in the blood between d5 and 21 post immunization but returned to naive levels by d31 (Fig. 3A-B and Supplemental Figures 1 and 2C). Notably, CD21loCD11b+ Ag-specific B cell frequencies were selectively increased in cynomolgus macaques, AGM, and mice following immunization (Fig. 3C and ref.(15). Non-Ag-specific CD20loCD21lo B cells only represented 2-3% of the total CD20+ blood B cell population in naïve AGM, and B cells bearing the CD20loCD19hi phenotype were not detected in the peritoneal cavity (data not shown). CD11b expression was only present on ~1/3 of these blood B cells, and hence, CD11b+CD20loCD21loCD19+ B cells represented only 1.1 ± 0.6% of the total CD20+ blood B cell pool in naïve AGM. These results show that, similar to the Ag-specific B-1b cell response to TNP-Ficoll in mice (Supplemental Figure 2A-B and ref.(15), TNP-Ficoll immunization in NHP transiently induces Ag-specific blood B cells with an FSChiSSChiCD20+/loCD86+PD-1+IgMloCD19hiCD21loCD11b+ phenotype. This finding is consistent with the participation of NHP B-1 cells in the Ag-specific B cell response to TNP-Ficoll.
Given the role that CD27+ “memory” B cells are thought to have in human immune responses to polysaccharide Ags, we examined CD27 expression by Ag-specific blood B cells before and following immunization. In contrast to non-Ag-specific blood B cells (~25-50% CD27+), Ag-specific B cells lacked CD27 expression on d0 but clearly expressed CD27 following immunization (Fig. 3D). Ag-specific CD11b+ cells, regardless of CD20 expression levels, expressed higher levels of CD27 than CD11b- Ag-specific B cells (Fig. 3E and Supplemental Figure 2D). CD20loCD11b+ cells were also CD38lo, while CD20hiCD11b+ B cells were CD38hi (Fig. 3E). Consistent with their B-1b cell phenotype, both CD11b+ B cell populations expressed higher levels of CD19 relative to CD11b-Ag-specific B cells and non-Ag-specific B cells.
To assess whether Ag-specific B-1 cells were actively secreting Ag-specific Ab, we sorted CD11b+ and CD11b- PBMC populations and measured Ab production by ELISPOT. As shown in Fig. 3F, TNP-specific ASC frequencies were significantly higher (30-fold) in the CD11b+ B cell fraction compared to the CD11b- fraction on d5. This is consistent with the CD27+CD38lo phenotype of Ag-specific CD20loCD11b+ cells, which suggests plasmablast transition (35). Thus, NHP B cell responses to haptenated-Ficoll are similar to responses in mice, with Ag-specific activated CD11b+CD19hiCD21lo blood (B-1 cells) B cells significantly increasing in the blood of NHP following immunization and declining by 1 month post immunization, and with Ag-specific Ab production originating from the B-1 cell population. However, NHP Ag-specific blood B-1 cells additionally express CD27 following, but not before, immunization.
We assessed TNP-specific ASC in multiple tissues 40 days post immunization. In cynomolgus macaques as well as in AGM, IgM-secreting cell frequencies were significantly higher in spleen than in bone marrow (Fig. 4A-B). A similar finding was made for mouse (Fig. 4C). IgM-secreting Ag-specific spleen cells outnumbered IgG-secreting spleen cells in all species (Fig. 4 A-C), consistent with the limited isotype switching that occurs with TI-2 Ags. IgG-secreting Ag-specific cell frequencies were more evenly distributed between spleen and bone marrow. Notably, very few ASCs were detected in LNs following TNP-Ficoll immunization (Fig. 4A), consistent with what has been reported for TNP-Ficoll and other TI-2 Ags in mouse. Finally, for both primate species, Ag-specific serum IgM and IgG, but not IgA, levels increased following immunization and were maintained up to 4 weeks post immunization (Fig. 4A-C). Thus, Ab responses to TNP-Ficoll are very similar between mouse and NHP, with IgM predominantly being produced and a large fraction of Ag-specific ASC being located in the spleen.
Ag-specific B-1b cells significantly increase in spleens of mice following TNP-Ficoll immunization and can be identified at least 5 wks post i.p. immunization, with numbers of Ag-specific splenic B-1b cells numbers remaining increased over naïve animals by ~3-fold (15). These cells may be plasmablasts and/or memory cells based on previous studies (16-18, 36, 37). Therefore, we assessed whether primate Ag-specific B-1 cells participate in the splenic B cell response to TNP-Ficoll. Total Ag-specific B cell frequencies in immune cynomolgus macaque spleens were significantly increased (1.6-fold) over naïve frequencies at d40 (Fig. 5A-B). This was due to a selective increase (>10-fold) in the frequency of CD11b+ Ag-specific B cells (Fig. 5C-D). An increase in CD11b+ Ag-specific B cells was also observed in immune AGM 7 weeks post immunization (Supplemental Figure 2E). Moreover, the frequency of Ag-specific B cells expressing an IgMhiIgDlo phenotype was significantly increased in immune spleens (Fig. 5E-F) and these Ag-specific IgMhiIgDlo B cells coexpressed CD11b in contrast to their IgMlo counterparts (Fig. 5G). CD11b was clearly expressed at much higher levels on Ag-specific B cells relative to the non-Ag-specific CD21hi (MZ B cell-enriched) B cell population (Fig. 5H) and thereby phenotypically distinguishes the Ag-specific CD11b+IgMhiIgDlo population from MZ B cells. Additional analysis of Ag-specific CD11b+IgMhiIgDlo B cells revealed that they were CD19hi with increased FSC and CD20 expression relative to CD11b- Ag-specific B cells and non-Ag-specific B cells (Fig. 5I). Thus, similar to results with mice, Ag-specific B-1 cell frequencies in NHP remain significantly increased at least one month following TNP-Ficoll immunization (15).
In naïve NHP, Ag-specific splenic B cells did not exhibit CD27 expression over MFI values obtained for isotype control staining (Fig. 5J and Supplemental Figure 2F), similar to results with blood (Fig. 3D). In contrast, a significant fraction of Ag-specific splenic B cells expressed CD27 following immunization (Fig. 5J). Importantly, expression of CD27, along with CD80, was restricted to CD11b+ Ag-specific B cells from immune animals (Fig. 5I). We have similarly found that CD80 is induced and maintained on TNP-specific B cells in mice following TNP-Ficoll immunization (data not shown). Consistent with previous reports showing increased proliferation by Ag-specific murine splenic B-1b cells at least 1 mo beyond immunization (16, 36), the frequency of proliferating (Ki-67+) Ag-specific B cells at d40 was significantly increased over that detected for non-Ag-specific B cells (Fig. 5K) and was restricted to CD11b+ cells (Fig. 5L). Ki67+ cells represented 5.1 ± 0.8 % of the Ag-specific CD11b+CD20+ population. Ki67+ non-Ag-specific B cells were largely CD11b-. In summary, TNP-Ficoll immunization in NHP induced a significant and selective increase in splenic CD11b+CD19hiCD20hiIgMhiIgDloCD27+CD80+/-FSChiSSChi Ag-specific B cells with an increased proliferative capacity 6 wks post immunization. With the exception of CD27 expression, similar markers are expressed by Ag-specific B-1b cells in spleens of immune mice (15).
Collectively, our novel findings demonstrate that: 1) primate B cells in serosal tissues exhibit striking phenotypic and functional similarities with murine B-1 cells, 2) there is remarkable similarity between mouse and primate Ag-specific B cell responses to TNP-Ficoll, with primate Ag-specific B-1 cells playing a dominant role in the humoral immune response to this model TI-2 Ag, and 3) Ag-specific primate B-1 cells enter the IgM+IgDloCD27+ compartment upon TI-2 Ag exposure, thereby linking the mouse B-1b cell population with the human/primate IgM+IgD+CD27+ “memory” population. Our findings may aid in resolving the controversies surrounding the existence of B-1 cells in higher primates and the origin of IgM+CD27+ “memory” B cells.
Previous studies have not directly performed side-by side phenotypic and kinetic analyses of murine and primate Ag-specific B cell responses to a TI-2 Ag. This, along with the use of different markers and subset definitions in mice and primates has made it impossible to conclude whether the B cell subsets responding to TI-2 Ags truly differ between rodents and primates. Immunization of mice with the synthetic TI-2 Ag, TNP-Ficoll, results in clearly identifiable Ag-specific CD11b+ B cells in blood, spleen and lymph nodes (15), despite the very low overall CD11b+ B cell frequencies found in these tissues. In the current study, we utilized CD11b expression along with other B-1 cell-associated markers to assess the potential for NHP B-1 participation in the response to TNP-Ficoll. By using the identical Ag and phenotyping scheme that we used to discover the role of murine B-1b cells in the response to TNP-Ficoll, we have been able to definitively show that the primate B cell response to TNP-Ficoll is highly similar to that of the B-1b cell response in mice with respect to: 1) the phenotype (FSChiSSChiCD20+/loCD86+PD-1+IgMloCD19hiCD21loCD11b+), expansion and activation kinetics of Ag-specific blood B-1 cells, 2) the differentiation of Ag-specific B-1 cells into Ab-producing cells, 3) the maintenance of increased numbers of Ag-specific B-1 cells in the spleen out to 6 weeks post immunization along with their increased proliferative capacity, and 4) the localization of ASCs to primarily spleen and bone marrow. Taken together, these results strongly support that the functional roles of murine B-1b cells in TI-2 Ag responses have been conserved in NHP.
The main difference found between the primate and murine Ag-specific B cell response to TNP-Ficoll was that primate Ag-specific B-1 cells expressed (and maintained) CD27 at the cell surface following immunization. This is not surprising, as CD27 expression in mouse and primates is differentially regulated, with murine CD27 being expressed at much lower frequencies and transiently on Ag-activated murine B cells (38). Thus, our findings do not dispute the conclusions that had previously been drawn regarding the importance of primate CD27+IgM+ B cells in generating TI-2 Ab responses since TI-2-Ag-activated IgM+ B (B-1) cells are likely induced to express CD27 and actively secrete Ab. However, because CD27+ TNP-specific B cells were not found prior to TNP-Ficoll immunization, our data do not support the conclusion that naïve TNP-Ficoll-specific IgM+CD27+ B cells exist in NHP prior to Ag exposure, as has been suggested to be the case with human IgM+CD27+ B cells that are reactive against bacterial polysaccharide (26). Although the somatic hypermutation which accumulates in the polysaccharide-responsive IgM+CD27+ human B cell population has been proposed to be driven by an Ag- and germinal-center-independent mechanism (26), the possibility that TI-2 Ag encounter leads to CD27 expression and somatic hypermutation in IgM+CD27+ “memory” B cells (39), even in cases where productive Ab responses are not generated (ie., in infants), cannot be excluded. Thus, our results suggest that TI-2 Ag immunization induces B-1 cells to co-express the IgM+CD27+ “memory” phenotype, and given the maintenance of this population out to 6 weeks post immunization and sustained expression of the CD80 memory marker (40), raises the possibility that persisting TI-2 Ag-experienced IgM+CD27+ B-1 cells may represent a memory cell population. As TI-2 Ag-specific memory B cells are unable to optimally respond to secondary Ag encounter (37), future studies assessing the mechanistic basis for the functional unresponsiveness of this subset may be essential in developing strategies for enhancing carbohydrate vaccine responses.
In the current study we also show that primate B-1-like cells bearing a CD11b+CD19hiCD21lo/-CD27+/-CD80+/-IgMhiFSChi phenotype are selectively enriched in NHP peritoneal cavities and omental tissue as in mice. Moreover, a fraction of these primate peritoneal CD11b+ B-1 cells expressed constitutively active Stat3 and exhibited increased specificity for phosphorylcholine, both of which are characteristics of murine B-1a cells (33, 34). Notably, our data from young developing primates shows that the frequency of CD11b expression among peritoneal B cells significantly increases between the first 3 months of age to 1.5-3 years of age, raising the possibility that delayed development of the B-1b-like cell compartment could contribute to the impaired TI-2 Ab responses observed in very young primates and mice. That primate B-1-like cells were heterogeneous in their expression levels of CD21 and CD80 is consistent with the heterogeneity observed in the expression levels of these markers among mouse B-1a and B-1b cell populations. For example, whereas B-1a cells express very low levels of CD21/35, CD19+CD11b+CD5- B-1b cells can be subdivided into CD21int and CD21lo/- populations (11), CD23+ and CD23- populations (41), as well as B220hiCD80- and B220loCD80+ populations (unpublished observations). Thus, B-1a and B-1b subsets may be conserved in primates. Unfortunately, CD5 does not reliably distinguish B-1a cells from other B cell populations in primates, and is even problematic as a sole marker of B-1a cells in mice, as BCR-activated B-1b cells and B-2 cells (15), anergic conventional B cells (42), and B10 regulatory B cells all express CD5 (43). For this reason, we are presently unsure as to whether CD5 phenotypically distinguishes NHP B-1a and B-1b subsets. Nonetheless, as in mice, CD5+ and CD5- CD11b+ B cells were identified in NHP.
While the primate B-1 cells identified in our study have phenotypic and functional similarities with the highly-studied murine B-1a and B-1b cell populations, it is difficult to draw robust comparisons between primate B-1 cells and the recently-characterized B-1 cell population in human peripheral blood (4). First, while our data provide supporting evidence for the localization of primate B-1 cells in the peritoneal cavity and omental tissue, there is no evidence that the human B-1-like CD20+CD27+CD43+CD70- population is similarly enriched in serosal tissues, as would be expected for a B-1-like population. In addition to the fact that analysis was limited to blood B cells, the relative expression levels of markers characteristic of murine B-1 cells used in our study (eg., IgMhi, CD21lo/-, CD19hi) were not assessed in human B-1 study. Of note, CD27 was only expressed on a fraction of NHP B-1 (CD11b+) B cells in blood, spleen, and peritoneal cavity. Therefore, CD27 may not be an all-inclusive B-1 marker in NHP as has been proposed for human B-1 cells. Moreover, its induction on TI-2 Ag-specific B-1b-like cells supports that it may functionally mark Ag-experienced B-1 cells. Finally, CD43, a marker expressed on 60-80% of murine peritoneal B-1 cells (1, 44), was used to identify human B-1 B cells in blood. Unfortunately, the 3 human CD43 mAb clones we tested were not reactive with AGM or cynomolgus macaque cells. For this reason, we relied on CD11b expression as a marker for primate B-1 cells. While CD11b is an ideal marker for detecting serosal B-1 cells and activated circulating and splenic Ag-specific B-1 cells in mice and primates following TI-2 Ag immunization (15), its diminished expression on “resting” B-1 cells outside of the serosal tissues in both mice and primates may hinder the identification of B-1 cells. Indeed, this may explain why only a small percentage of CD11b+ cells (~10%) is found among the human B-1-like CD20+CD27+CD43+CD70- blood population (5). How this human B cell population relates to the CD11b+ blood and spleen B cells found in NHP is not yet clear. Interestingly, the human CD20+CD11b+CD27+CD43+CD70- population lacked CD80 expression whereas the Ag-experienced CD27+CD11b+ splenic and blood B cell populations in NHP coexpressed CD80. The functional roles for the CD11b+CD27- and CD11b+CD27+ CD19+ human B cell subsets also remain to be elucidated, although based on our results, these heterogeneous cell populations may harbor naïve and Ag-activated/experienced B-1 cells, respectively. Future studies are necessary to determine whether this is indeed the case.
In summary, our findings strongly support the existence of a primate B-1 cell population which participates in TI-2 Ab responses in a manner similar to murine B-1b cells. Similar to our findings, a recent study showed that human CD20+CD27+CD5- (CD43+) B-1-like cells produced PPS-specific Ab following Pneumovax immunization (54). Given the similarities between B cell populations of humans and NHP (31, 45, 46), it is clear that our identification of functional B-1 cells in two species of NHP is a significant step in our understanding of human B cell responses to TI Ags. Murine B-1a and B-1b cells have key roles in host defense and immune homeostasis (1, 2, 10, 17-19), immune regulation via IL-10 production (47-50), autoimmunity (50, 51), and transplant rejection (20). Therefore, it will be critical to investigate how human B-1 B cells function in multiple aspects of health and disease. That CD11b+ B cells comprise a fraction of the normal human B cell compartment (52) as well as B-CLL malignancies (53) with as-yet unknown functions that may significantly impact human disease (5) highlights the need for future studies of both murine and primate B-1 cells. Ultimately, studies of this unique population may lead to advanced vaccine strategies for improving TI Ab responses as well as novel therapies for the treatment of human diseases.
The authors thank the Wake Forest Primate Center (colonies funded by NIH P40RR019963-08 and P40RR02138005) for access to animals and Drs. Matthew Jorgenson and Glicerio Ignacio for assistance with sample collection.