Peritoneal B-2 cells comprise a distinct B-2 cell population with B-1b-like characteristics

Authors

  • William D. Hastings,

    1. Department of Microbiology, Boston University School of Medicine, Boston, MA, USA
    2. Immunobiology Unit, Evans Memorial Department of Clinical Research, Boston University Medical Center, Boston, MA, USA
    Search for more papers by this author
  • Joseph R. Tumang,

    1. Department of Medicine, Boston University School of Medicine, Boston, MA, USA
    2. Immunobiology Unit, Evans Memorial Department of Clinical Research, Boston University Medical Center, Boston, MA, USA
    Search for more papers by this author
  • Timothy W. Behrens,

    1. Department of Medicine and Center for Immunology, University of Minnesota Medical School, Minneapolis, MN, USA
    Search for more papers by this author
  • Thomas L. Rothstein Dr.

    Corresponding author
    1. Department of Microbiology, Boston University School of Medicine, Boston, MA, USA
    2. Department of Medicine, Boston University School of Medicine, Boston, MA, USA
    3. Immunobiology Unit, Evans Memorial Department of Clinical Research, Boston University Medical Center, Boston, MA, USA
    • Center for Oncology and cell Biology, The Feinstein Institute for Medical Research, 350 Community Drive, Manhasset, NY 11030 USA, Fax: +1-516-562-1022
    Search for more papers by this author

Abstract

B-1 and B-2 cells are lymphocyte populations that differ in development, surface marker expression, tissue localization, and function. Though mainly found in the spleen, lymph nodes, and circulation of mice, small numbers of B-2 cells are found in the peritoneal cavity, a site predominantly populated by B-1 cells. Here, we characterized peritoneal B-2 cells, and determined their relationship to B-1 cells. We found that peritoneal B-2 cells appear to be intermediate between splenic B-2 and peritoneal B-1 cells in terms of surface marker expression of B220, CD80, and CD43, expression of several marker genes, and in vitro viability and IgM secretion. Adoptive transfer of peritoneal B-2 cells into severe combined immunodeficiency mice resulted in the acquisition of a phenotype reminiscent of B-1b cells, as shown by up-regulation of Mac-1 and CD43, and down-regulation of CD23. Moreover, adoptively transferred peritoneal B-2 cells recapitulated B-1 cell function by producing natural IgM in recipient mice. These data suggest that peritoneal B-2 cells express some characteristics of B-1b cells and that this similarity increases with additional time in the peritoneal cavity.

Introduction

B lymphocytes can be subdivided into different subpopulations based on a variety of phenotypic and functional characteristics. B-2 cells, also known as conventional or follicular (FO) B cells, are abundant in the spleen, lymph nodes, and peripheral blood of mice, and arise continually from bone marrow precursors. They cooperate with T cells in the germinal center and undergo somatic hypermutation of their immunoglobulin genes to produce high-affinity antibody in the humoral immune response. B-1 cells arise early in life from fetal and neonatal progenitors 1, are enriched in the peritoneal and pleural cavities of mice, and can be distinguished by the expression of several surface markers that are not expressed by B-2 cells (CD5, Mac-1, and CD43), as well as by differential expression of traditional B cell markers (B-1 cells are B220lowIgMhigh/IgDlow/CD23neg/low, whereas B-2 cells are B220+IgMlow/IgDhigh/CD23high). Peritoneal B-1 cells themselves are comprised of two different populations, B-1a and B-1b, which are CD5+ and CD5, respectively, but are otherwise similar in their surface marker phenotypes.

The origin of B-1 cells has been a point of contention for the past several decades, with two main hypotheses predominating. Studies utilizing adoptive transfers of fetal and adult precursors originally led investigators to conclude that B-1 cells arise separately from B-2 cells before surface BCR expression and thus constitute a separate lineage 1, 2. This view is supported by the recent identification of a B-1 cell progenitor with preferential expression in fetal life 3. However, subsequent studies seemed to indicate that B-1 cells could derive from mature B-2 cells, or their precursors. It was shown that B-2 cells are induced to express CD5 by cross-linking the B cell receptor with anti-IgM, a mimic of T-independent type II stimulation 4. In addition, several BCR-transgenic mouse models have shown that the specificity (and perhaps signal strength) of the BCR determines B cell fate, i.e. whether a precursor becomes a B-2 or a B-1 cell 58. In these studies, transgenic mice expressing B-1 cell-derived BCR (such as the phosphatidylcholine-specific heavy/light chain combination VH12/Vk4) developed very large numbers of B-1 cells (in both the spleen and the peritoneal cavity), while mice expressing B-2 cell-derived BCR developed primarily B-2 cells. In general, it appears that strong BCR signaling results in the development of B-1 cells, while weaker BCR signals tend to favor B-2 cell development. In support of this, mice with mutations that weaken BCR signaling, such as X-linked immunodeficient (xid) mice or PI3-kinase knockouts, tend to have fewer B-1 cells than normal mice 911 , while mice with mutations that augment BCR signaling tend to have elevated numbers of B-1 cells 12, 13. A recent report using transgenic BCR-less mice expressing either high or low levels of a BCR surrogate also suggests that BCR signal strength determines whether a cell expresses the B-1 or B-2 cell phenotype 14.

Regardless of origin, it has become increasingly clear that B-1 and B-2 cells differ functionally, with both subsets contributing to host protection in a variety of disease models. B-1 cells have been shown in many studies to be the major source of natural antibody 1517, polyreactive, weakly autoreactive immunoglobulin that is present in the sera of pre-immune animals and is thought to protect the organism early in the immune response. B-2 cells, on the other hand, produce high affinity, somatically mutated immunoglobulin only after immunization. Antibodies from both cell types have been shown to be necessary for protection from influenza virus 18, while B-1 cell-derived natural antibodies are needed for protection from endotoxin shock 19 and ameliorate bacterial peritonitis induced by cecal ligation and puncture 20. B-1 cells have also been shown to cooperate with marginal zone B cells in the T-independent antibody response to phosphorylcholine (PC)-positive bacteria 21.

Studies have shown that B-1 cells in different physiological locations are not necessarily identical. Splenic B-1 cells differ from peritoneal B-1 cells in many characteristics, including lack of constitutively activated STAT3 and IL-10 mRNA expression, and failure to respond to PMA 22. Unlike peritoneal B-1 cells, splenic B-1 cells do not survive well in vitro, nor do they spontaneously secrete IgM 23. Further, in a transgenic mouse model, splenic B-1 cells were shown to differ from their peritoneal counterparts in antibody repertoire 24. Based on this evidence, it seems reasonable to hypothesize that B-2 cells from different locations are not identical. We set out to test this hypothesis by FACS-purifying splenic and peritoneal B-2 cells, and peritoneal B-1 cells, and testing them for several known characteristics, including surface marker expression, gene expression, and in vitro survival and IgM secretion. We found that peritoneal B-2 cells are not identical to their splenic B-2 cell counterparts, and in adoptive transfer experiments these differences are accentuated with increased time in the peritoneal cavity.

Results

Peritoneal B-2 cells exhibit unique surface phenotypic features

To determine whether peritoneal B-2 cells differ from conventional splenic B-2 cells, we analyzed the expression of several surface markers known to be differentially expressed by B-1 and B-2 cells. Peritoneal B-2 cells were defined as B220-positive cells that were both Mac-1 negative and CD5 negative, as described in the Materials and methods section. We found that peritoneal B-2 cells did not differ significantly from their splenic counterparts in size or complexity, as measured by forward and side light scatter (data not shown). We also found that they did not differ significantly from their splenic counterparts in relatively low expression of IgM [MFI: B-2P, 736 ± 16; B-2S, 430 ± 51; B-1P, 2250 ± 143; mean ± SEM, n = 3], relatively high expression of IgD (MFI: B-2P, 2643 ± 163; B-2S, 1902 ± 173; B-1P, 339 ± 25; n = 3), and moderate expression of CD23 (MFI: B-2P, 881 ± 108; B-2S, 612 ± 46; B-1P, 104 ± 17; n = 3), each of which differed from the levels expressed by peritoneal B-1 cells. However, we found that the level of B220 expression was reduced by approximately 40% on peritoneal B-2 cells as compared to splenic B-2 cells (MFI, 446 ± 28 and 713 ± 53, respectively, mean ± SEM, n = 6; see Fig. 1a). This difference is statistically significant (p = 0.003, paired Student's t-test). Still, the level of B220 expression by peritoneal B-1 cells was much lower than for either population of B-2 cells. In addition, we found that the expression levels of CD80 and CD43 were elevated on peritoneal B-2 cells in comparison to splenic B-2 cells. In each case, surface marker expression appeared as two relatively distinct positive peaks rather than as a general change in MFI, as was the case with B220 (Fig. 1b). Whereas the proportion of splenic B-2 cells expressing CD80 was 4.7 ± 0.25%, the proportion of peritoneal B-2 cells expressing this marker was 21.3 ± 0.33% (mean ± SEM, n = 3). This difference in expression between splenic and peritoneal B-2 cells is statistically significant (p = 0.0006). Whereas the proportion of splenic B-2 cells expressing CD43 was 2.8 ± 0.38%, the proportion of peritoneal B-2 cells expressing this marker was 9.5 ± 0.5% (mean ± SEM, n = 3). This difference in expression is also statistically significant (p = 0.009). By way of comparison, the proportion of B-1 cells expressing CD80 and CD43 was over 90%. Thus, compared to splenic B-2 cells, peritoneal B-2 cells express distinct levels of three surface markers that are each hallmarks of the B-1 cell phenotype, although for all three, peritoneal B-2 cell expression, while in the same direction as peritoneal B-1 cell expression, still falls short of peritoneal B-1 cell levels.

Figure 1.

Peritoneal B-2 cells exhibit unique surface phenotypic features. Peritoneal B-2 cells (B-2P, B220+/Mac-1/CD5), peritoneal B-1 cells (B-1P, B220lo/Mac-1+/CD5+) and splenic B-2 cells (B-2S, B220+/Mac-1/CD5) were isolated by cell sorting from peritoneal washouts and spleen suspensions of BALB/c mice, as described in the Materials and methods section. B cells were re-stained with anti-B220 FITC and stained with either anti-CD80 PE or anti-CD43 PE, and analyzed by flow cytometry. (A) B220 expression is shown for each population, as indicated. The shaded histogram represents staining with an isotype control antibody. Results are representative of six independent experiments for B-2P cells and B-2S cells and three independent experiments for B-1P cells. (B) CD80 and CD43 expression is shown for the three cell populations. Gates were set based on isotype control staining. Results are representative of three independent experiments for B-2S and B-2P cells.

To rule out the possibility that differential expression of these markers resulted from contamination of the peritoneal B-2 cell population with B-1 cells, we re-analyzed sorted B cells, and specifically compared peritoneal B-2 cells with both subsets (B-1a and B-1b) of B-1 cells (Fig. 2). The peritoneal B-2 cell population was less than 1% positive for Mac-1 expression, much like splenic B-2 cells, whereas peritoneal B-1a and B-1b cells were almost entirely (94%) Mac-1 positive. Thus, putative B-1 cell contamination of the peritoneal B-2 cell population cannot explain the differential expression levels of B220, CD80, and CD43.

Figure 2.

Post-sort analysis of purified B cells reveals highly pure populations. Peritoneal B-2 cells (B-2P, B220+/Mac-1/CD5), peritoneal B-1a cells (B-1aP, B220lo/Mac-1+/CD5+), peritoneal B-1b cells (B-1bP, B220lo/Mac-1+/CD5) and splenic B-2 cells (B-2S, B220+/Mac-1/CD5) were isolated by cell sorting from peritoneal washouts and spleen suspensions of BALB/c mice. B cells were re-analyzed for B220 and Mac-1 expression by flow cytometry. A representative experiment is shown.

Peritoneal B-2 cell gene expression does not match that of splenic B-2 cells nor peritonealB-1 cells

We further examined peritoneal B-2 cells in terms of transcriptomic phenotype. A preliminary genechip analysis of RNA isolated from different B cell subsets suggested that peritoneal B-2 cells are intermediate in the expression of a number of genes that are differentially expressed by peritoneal B-1 and splenic B-2 cells. To confirm this, we isolated RNA from FACS-purified splenic and peritoneal B-2 cells and peritoneal B-1 cells, generated cDNA by reverse transcription, and performed quantitative real-time PCR using the primers listed in the Materials and methods section. Relative expression was calculated by comparison with the housekeeping gene β-2 microglobulin. Results are shown in Fig. 3, and are representative of three independent experiments. We found the gene encoding elfin, an Enigma family cytoplasmic protein associated with the cytoskeleton and protein trafficking 25, was expressed at a relatively high level in peritoneal B-1 cells (0.058 ± 0.0043, mean ± SEM), but at a much lower level in splenic B-2 cells (0.00094 ± 0.00030). Elfin expression by peritoneal B-2 cells fell in between these values (0.0082 ± 0.0013). These expression levels were all significantly different from each other, with p-values <0.05 for each population compared to the others.

Figure 3.

Peritoneal B-2 cell gene expression does not match that of splenic B-2 cells nor peritoneal B-1 cells. Peritoneal B-2 cells (B-2P), peritoneal B-1 cells (B-1P), and splenic B-2 cells (B-2S) were isolated by cell sorting as described above. RNA was extracted, cDNA was prepared, and gene expression was determined by quantitative PCR and normalized to expression of β2-microglobulin, using primer sets described in the Materials and methods section. Mean normalized values for three independent experiments are shown along with lines indicating SEM.

In other situations, peritoneal B-2 cell gene expression approximated the level characteristic of peritoneal B-1 cells or splenic B-2 cells. An example of the former is the gene for integrin beta-1, a molecule associated with lymphocyte localization 26, which was expressed at a relatively high level in peritoneal B-1 cells (0.0066 ± 0.0016, mean ± SEM), but was undetectable in splenic B-2 cells. In this case, expression by peritoneal B-2 cells (0.0035 ± 0.00097) was similar to peritoneal B-1 cell expression. It should be noted that aside from the purity of sorted peritoneal B-2 cells (see Fig. 2 and the Materials and methods section) it is unlikely that expression of integrin beta-1 in peritoneal B cells is the result of macrophage contamination because splenic B-1 cells also express high levels of this gene, and would not be expected to have macrophage contamination (data not shown). An example of the latter is the gene for CD23, the low affinity Fc receptor for IgE, which is known to be expressed at high levels on the surface of B-2 cells, but not B-1 cells, and which we found was transcriptionally expressed at a relatively high level in splenic B-2 cells (0.014 ± 0.0019, mean ± SEM), but at a much lower level in peritoneal B-1 cells (0.0016 ± 0.00012). In this case, expression by peritoneal B-2 cells was similar to splenic B-2 cell expression (0.012 ± 0.00011). This selected transcriptomic profile of peritoneal B-2 cells demonstrates that in some cases peritoneal B-2 cells are similar to peritoneal B-1 cells (integrin beta-1), and in other cases peritoneal B-2 cells are similar to splenic B-2 cells (CD23), and in still other situations peritoneal B-2 cells may lie somewhere between the two (elfin).

To further define the relationship between peritoneal B-2 cells and peritoneal B-1 cells, we examined expression of VH11 and VH12, two variable region genes that are overexpressed in peritoneal B-1 cells and that, when paired with appropriate light chains, give rise to antigen receptors specific for phosphatidylcholine. We found that both VH11 and VH12 were expressed at relatively high levels by peritoneal B-1 cells (0.0058 ± 0.00029 and 0.0056 ± 0.00094, respectively, n = 3), but at much lower levels in splenic B-2 cells (0.00033 ± 0.000061 and 0.00036 ± 0.000016, respectively), as expected. Peritoneal B-2 cell expression of both VH11 and VH12 was very low (0.00036 ± 0.000014 and 0.00054 ± 0.000093, respectively) and approximated the low level of expression observed in splenic B-2 cells, as previously reported based on a different purification protocol 27. These results suggest that peritoneal B-2 cells are not positively selected for BCR expression in the same manner as peritoneal B-1 cells, and further indicate that sorted peritoneal B-2 cells are not contaminated by meaningful levels of B-1 cells.

Peritoneal B-2 cells survive and secreteIgM in vitro

Peritoneal B-1 cells survive in culture to a much greater extent than splenic B-2 cells 28. In light of the fact that peritoneal B-2 cells exist in a similar physiological niche, we hypothesized that they might exhibit similar survival characteristics to B-1 cells, and we tested this notion by culturing all three cell types (splenic and peritoneal B-2 cells and peritoneal B-1 cells) in vitro for 3 days. Viability was assessed by staining with propidium iodide and measuring the percentage of cells containing intact (2N) DNA. Results are shown in Fig. 4 and depict the means of four independent experiments. We found that the majority of splenic B-2 cells did not survive 3 days in culture, as expected 23, with 36 ± 2.8% (mean ± SEM, n = 4) viability at the end of the incubation. In contrast, peritoneal B-1 cells remained largely viable over the entire culture period (76 ± 6.2%), also as expected 23. Peritoneal B-2 cells approximated the enhanced viability in vitro of peritoneal B-1 cells, with 70 ± 6.1% expressing 2N DNA at 3 days. This level of viability was significantly increased over that of splenic B-2 cells (p = 0.006). Beyond the purity of input peritoneal B-2 cells, this could not represent expansion of contaminating B-1 cells because of insufficient time for completion of more than one cell cycle. Thus, peritoneal B-2 cells mimic peritoneal B-1 cells in terms of in vitro survival.

Figure 4.

Peritoneal B-2 cells survive in vitro. Peritoneal B-2 cells (B-2P), peritoneal B-1 cells (B-1P), and splenic B-2 cells (B-2S) were isolated by cell sorting and cultured for 3 days. Viable cells were assessed by staining with propidium iodide and measuring the proportion of cells with 2N DNA content. Mean normalized values for four independent experiments are shown along with lines indicating SEM.

In addition to enhanced survival, B-1 cells spontaneously secrete IgM when placed in culture for several days 29. This may reflect their function as natural antibody producing cells in vivo 15. In view of the results on in vitro survival discussed above, we speculated that peritoneal B-2 cells might secrete IgM like peritoneal B-1 cells. To address this, we collected culture supernatants after varying periods of time and assayed IgM content by ELISA. As expected, IgM was not detected in splenic B-2 cell supernatants at any time point examined, whereas peritoneal B-1 cells produced detectable IgM as early as 2 days, with increasing amounts on subsequent days (Fig. 5). Peritoneal B-2 cells also produced IgM, although to a lesser extent than B-1 cells. In general, peritoneal B-2 cells secreted about one third to one-half as much IgM as peritoneal B-1 cells. Since these populations were highly pure (Fig. 2 and the Materials and methods section), a small contaminating B-1 cell population could not account for peritoneal B-2 cell IgM secretion. Thus, peritoneal B-2 cells mimic peritoneal B-1 cells in the capacity to secrete IgM in vitro.

Figure 5.

Peritoneal B-2 cells secrete IgM in vitro. Peritoneal B-2 cells (B-2P), peritoneal B-1 cells (B-1P), and splenic B-2 cells (B-2S) were isolated by cell sorting and cultured for 5 days. SN were harvested each day and assayed for the presence of IgM by ELISA. Mean values for four independent experiments for days 1–4, and two independent experiments for day 5, are shown along with lines indicating SEM.

Peritoneal B-2 cells accentuate B-1b-like characteristics upon adoptive transfer

Based on their intermediate phenotype, we reasoned that peritoneal B-2 cells might be influenced by the peritoneal environment to progressively acquire B-1 cell-like characteristics. To assess this, we adoptively transferred, by intraperitoneal inoculation, purified (Fig. 2) BALB/c splenic and peritoneal B-2 cells into congenic CB.17 SCID mice, which are deficient for B and T lymphocytes 30. Recipient mice were sacrificed 11 days post-inoculation. Mice that received splenic or peritoneal B-2 cells yielded substantial numbers of peritoneal lymphocytes as compared to control mice that received PBS alone. Cell yields averaged 26.4 ± 8.4% of input cells for peritoneal B-2 cells and 7.8 ± 3.7% for splenic B-2 cells (mean ± SEM, n = 3, p = .03), indicating that substantial numbers of donor peritoneal B-2 cells remained after 11 days, and also showing an increased propensity for peritoneal B-2 cells, as compared to splenic B-2 cells, to remain in, and be recovered from, the peritoneal cavity. Staining for IgMb, which is expressed by CB.17 mice but not by BALB/c mice, demonstrated that there were no resident recipient-derived B cells present in peritoneal washouts (data not shown); thus, all of the peritoneal B cells were donor-derived (as confirmed by IgMa staining). These adoptively transferred B cells were subsequently analyzed by flow cytometry (Fig. 6a). During 11 days of peritoneal residence, a substantial proportion of peritoneal B-2 cells became Mac-1 positive (41.0 ± 3.6%; mean ± SEM, n = 3, Fig. 6b), whereas a much smaller proportion of splenic B-2 cells acquired Mac-1 expression (17.8 ± 7.6%, mean ± SEM). Further, a substantial proportion of peritoneal B-2 cells became CD43 positive (28.4 ± 6.5%; increased from 9.1 ± 0.83% pre-sort), whereas very few splenic B-2 cells acquired CD43 expression (7.9 ± 0.60%; increased from 3.3 ± 0.61% pre-sort). Finally, a substantial proportion of peritoneal B-2 cells became CD23 negative (25.3 ± 5.0%; increased from 8.9 ± 1.5% pre-sort), whereas splenic B-2 cells did not lose CD23 expression (9.0 ± 1.2%; unchanged from 8.2 ± 1.4% pre-sort). The differences in expression of Mac-1, CD43, and CD23 by the two types of adoptively transferred B cells were statistically significant for all three surface markers (p = 0.015, 0.037, and 0.049, respectively, n = 3). Changes in Mac-1, CD43, and CD23 among adoptively transferred peritoneal B-2 cells appear to represent to some degree a coordinate phenotypic transition in that Mac-1+, CD43+ and CD23 peritoneal B-2 cells each contained very low B220 expressing B cells (10.5% ± 2.1, 18.7% ± 4.5, and 16.1% ± 4.1, respectively; n = 3), along with a general reduction in B220 expression, whereas Mac-1, CD43 and CD23+ peritoneal B-2 cells contained few, if any, B220lo B cells (Fig. 6c). Notably, the observation that a higher proportion of CD43+ B cells and CD23 B cells than Mac-1+ B cells expressed low B220, combined with the fact that more peritoneal B-2 cells acquired Mac-1 expression than gained CD43 expression or lost CD23 expression, suggests that only Mac-1+CD43+CD23 B cells down-regulate B220. Thus, a substantial number of peritoneal B-2 cells accentuate expression of surface markers typical of B-1b cells, whereas only a much smaller number of splenic B-2 cells experience any change, and that primarily with respect to Mac-1, and not other markers.

Figure 6.

Peritoneal B-2 cells can acquire a B-1b-like phenotype upon adoptive transfer. Peritoneal B-2 cells (B-2P), and splenic B-2 cells (B-2S) were isolated by cell sorting from BALB/c mice (IgMa) and inoculated intraperitoneally into CB.17 SCID recipients (IgMb) at approximately 1 ×X 106 cells per mouse in each experiment. Control mice received sterile PBS. Eleven days later, mice were sacrificed and peritoneal cell populations were stained with immunofluorescent antibodies and analyzed by flow cytometry. (A) Peritoneal lymphocyte populations were analyzed for expression of IgMa, Mac-1, CD5, CD43, and CD23. One of three comparable experiments is shown. (B) Peritoneal lymphocyte populations were analyzed for expression of B220+Mac-1+, B220+CD43+, and B220+CD23 B cells. Mean values for three independent experiments are shown along with lines indicating SEM. (C) Peritoneal lymphocyte populations obtained from peritoneal B-2 cell-inoculated SCID recipient mice were analyzed for expression of IgMa, Mac-1, CD43 and CD23; and then Mac-1+vs. Mac-1, CD43+vs. CD43, and CD23+vs. CD23 populations of IgMa+ B cells were analyzed for expression levels of B220, as indicated. One of three comparable experiments is shown.

Peritoneal B-2 cells can function like B-1 cellsin vivo

To determine if peritoneal B-2-derived cells function like B-1 cells, we analyzed serum IgM in SCID recipients of splenic and peritoneal B-2 cells, as compared to peritoneal B-1 cell recipients. Peripheral blood was collected at 11 days post-transfer, and serum was analyzed by IgM ELISA. Mean values for three independent experiments are shown in Fig. 7. As expected, no serum IgM was detected in the PBS control recipients at any time point (data not shown). Further, no serum IgM was detected in the splenic B-2 recipients. However, serum samples from recipient mice of both peritoneal B cell populations (peritoneal B-2 cells and peritoneal B-1 cells) showed detectable serum IgM that averaged 24.5 ± 14.6 μg/mL (mean ± SEM) for peritoneal B-2 cells and 50.2 ± 14.1 μg/mL for peritoneal B-1 cells. Thus, peritoneal B-2 cells recapitulate to some degree the peritoneal B-1 cell capacity to partially reconstitute serum IgM levels in SCID recipient mice. Of note, over longer periods of time, some serum IgM was detected in SCID recipients of splenic B-2 cells, although this was variable from experiment to experiment and never approached the levels produced by peritoneal B-2 cells.

Figure 7.

Peritoneal B-2 cells can function like B-1 cells in vivo. Peritoneal B-2 cells (B-2P), peritoneal B-1 cells (B-1P), and splenic B-2 cells (B-2S) were isolated by cell sorting and inoculated intraperitoneally into SCID mouse recipients, as indicated. Control mice received only PBS. Mice were bled 11 days thereafter, and serum samples assayed for IgM by ELISA. Mean values for three independent experiments are shown along with lines indicating SEM.

Discussion

This work provides two new insights into peritoneal B-2 cells: (1) peritoneal B-2 cells are distinct from splenic B-2 cells in a number of static and dynamic characteristics, some of which parallel B-1 cell features; and, (2) peritoneal B-2 cells acquire some aspects of the B-1 cell phenotype during continued residence in the peritoneal cavity.

The data obtained in these experiments strongly suggest that peritoneal B-2 cells are distinct from splenic B-2 cells. In comparison to splenic B-2 cells, peritoneal B-2 cells express less B220 antigen, more integrin beta-1 and elfin mRNA, and enhanced viability and increased IgM secretion in vitro. Many of these features, such as surface marker and gene expression, survival, and IgM secretion, are similar to the characteristics of B-1 cells. Because peritoneal B-2 cells share some characteristics with peritoneal B-1 cells, we cannot completely rule out the remote possibility that they are, in fact, B-1 cells. However, based on the evidence presented here, they would be B-1 cells that do not express CD5 or Mac-1 just as B-2 cells do not, that express elevated CD23, low IgM and high IgD, just as B-2 cells do, and that do not show evidence of repertoire selection for PtC-binding VH genes, just as B-2 cells do not. In other words, they would be B-1 cells with a unique set of B-2 cell-like features that has not been described to date and thus would, of necessity, constitute a new and interesting category of peritoneal B cells even if not peritoneal B-2 cells. It is considered more likely that, in accordance with Occam's razor, and the usual definition of B-2 cells, the B220+/IgMlo/IgDhi/CD23+/Mac-1/CD5 B cells studied here are peritoneal B-2 cells that differ from splenic B-2 cells in the ways described above, which differences become more accentuated with time. However, these B cells constitute an important and up to now understudied peritoneal population whether they are considered B-2 cells with B-1 cell-like characteristics or B-1 cells with B-2 cell-like characteristics.

The possibility that the unique characteristics of peritoneal B-2 cells represent a single temporal point in a continuing developmental or differentiative pathway was explored through an adoptive transfer strategy in which inoculated B cells could be tracked. Over time, a substantial proportion of peritoneal B-2 cells acquired Mac-1 and CD43 expression, and lost CD23 expression. Although some peritoneal B-2 cells expressed CD43 (and some did not express CD23) prior to adoptive transfer (Fig. 1b), virtually none expressed Mac-1, and so acquisition of this surface antigen represents acquisition of a previously unexpressed marker, rather than up-regulation (or down-regulation) of a marker expressed to some extent by unmanipulated peritoneal B-2 cells. These results suggest that peritoneal B-2 cells are changing within the peritoneal environment, adopting at least some characteristics of B-1 cells, a view reinforced by the observation that adoptively transferred peritoneal B-2 cells that express Mac-1 or CD43, or fail to express CD23, preferentially down-regulated B220 expression. The generation of serum IgM in adoptive hosts in the absence of immunization further supports the idea that peritoneal B-2 cells express B-1-like characteristics during peritoneal residence.

The induction of B-1 cell phenotypic characteristics in (splenic) B-2 cells has been reported before 4, 31, 32. However, the current work differs from these previous reports in two key respects. In earlier work, induction of B-1 cell characteristics depended on B-2 cell stimulation, whereas in the present study B-2 cells acquired B-1b-like phenotypic features during residence in the peritoneal cavity without stimulation; in earlier work induction of B-1 cell characteristics prominently included acquisition of CD5 expression, whereas in the present study peritoneal B-2 cells did not up-regulate CD5. Thus, over time, peritoneal B-2 cells acquired characteristics of B-1b, not B-1a, B cells. Previous work has shown that B-1b cells can develop from bone marrow progenitors in the adult, whereas B-1a cells do not or do so at a vanishingly small rate 33. The apparent transition of peritoneal B-2 cells toward a B-1b-like phenotype is consistent with the derivation of B-1b cells from the bone marrow. The unexpected observation that small numbers of splenic B-2 cells also acquire some phenotypic characteristics of B-1b cells, primarily Mac-1 expression, during forced peritoneal residence may indicate that these cells are somehow fated to enter into a developmental pathway to the B-1b phenotype. Thus, some splenic B-2 cells may receive an activation signal that induces them to migrate to the peritoneum where as peritoneal B-2 cells they begin to acquire B-1b cell characteristics. In this respect it is notable that many B-1 cell-like phenotypic characteristics of B-2 cells, such as elevation of CD43 and CD80, and decreased expression of B220 and CD23, are also exhibited by activated B-2 cells 4, 3436, supporting the notion that peritoneal B-2 cells may have encountered activation signals. Alternatively, Mac-1 may reflect a unique situation, with induction produced by specific factors in the peritoneal environment, so that acquisition of Mac-1 expression by splenic B-2 cells does not imply any concordance with peritoneal B-2 cells, and these B-2 cell populations may be completely distinct, as represented by CD23 expression, which is lost by peritoneal B-2 cells in vivo but not by splenic B-2 cells. This latter scenario especially raises the possibility that some or all of the B-1b-like characteristics displayed and/or acquired by peritoneal B-2 cells, including enhanced viability and IgM secretion, may represent the influence of the peritoneal environment. A role for the peritoneal environment is supported by previous work. It was shown there that BCR signaling for Ca2+ mobilization in VH11Vκ9 transgenic splenic B-1 cells was modulated after forced residence in the peritoneal cavity 37, although this study did not address the fate of B-2 cells, and the extent to which B-2 cells might be influenced by the peritoneal environment is entirely unknown at the present time. Moreover, it is not known whether the dichotomy of peritoneal B-2 cells, in which some but not all express B-1 cell-like characteristics, reflects a precursor/product relationship between CD43/CD80 and CD43+/CD80+ B-2 cells, or reflects a scenario in which the latter population is in transition to a B-1b-like phenotype, but the former population is not, perhaps because it is unaffected by putative signals that exist in the peritoneal cavity. However, the fact that CD43 expression rose further when peritoneal B-2 cells were "parked" in SCID peritoneal cavities for longer periods of up to 44 days (data not shown) suggests that potentially all peritoneal B-2 cells are malleable and can enter a pathway leading to B-1b-like features.

Alternatively, it might be thought that small numbers of contaminating B-1 cells could preferentially survive or expand following adoptive transfer in recipient mice, leading to the appearance of Mac-1+/CD43+/CD23/B220lo cells. This is highly unlikely for several reasons. Peritoneal B-2 cells exhibited similar survival characteristics to peritoneal B-1 cells (Fig. 3), and thus there is no evidence to suggest that B-1 cells would survive preferentially in recipient mice. Post-sort, pre-adoptive transfer, analysis of peritoneal and splenic B-2 cells showed a very high degree of purity, with less than 1% Mac-1+ cells present (Fig. 2), indicating that any putative expansion would start with very few cells and would thus demand a very high rate of proliferation. Moreover, yields of peritoneal B-2 cells were relatively high in the adoptive recipients (average 26% of input cells), further suggesting that extensive proliferation would be necessary were B-1-like cells to have arisen from the small number of Mac-1+ potential contaminants. Finally, when input peritoneal B-2 cells were labeled with CFSE and analyzed 11 days after adoptive transfer, only a small proportion (10%) of recovered B cells had undergone cell division, a number that is clearly insufficient to account for the level of Mac-1 and CD43 acquisition observed (amounting to about 20% of recovered B cells) (data not shown). Thus, it is highly unlikely that preferential survival or expansion could account for the peritoneal B-2 cell changes on adoptive transfer.

Taken together with other studies 4, 31, 32, our results support the hypothesis that peritoneal B-2 cells can acquire a B-1b-like phenotype under the appropriate conditions. This may reflect the ability of peritoneal B-2 cells to directly differentiate into B-1 cells as has been postulated by several investigators 38, 39. Alternatively, it may simply reflect the acquisition of similar phenotypes by different cell types that reside in the same physiological niche. In any case, our work clearly indicates that peritoneal B-2 cells are a unique B cell population, distinct from splenic B-2 cells. The role of the peritoneal environment in dictating B-1b-like features and/or triggering a putative B-2 to B-1 cell transition remains to be clarified.

Materials and methods

Animals

Male BALB/cByJ and congenic CB.17(SCID) mice were obtained from The Jackson Laboratory (Bar Harbor, ME) and maintained in the animal facilities at Boston University Medical Center in accordance with NIH guidelines.

Cell purification and flow cytometry

Peritoneal washouts and splenocytes were obtained from 8–14-week-old mice, and stained with fluorescence labeled antibodies to B220, Mac-1, and CD5 (peritoneal washout cells) or B220 and CD5 (splenocytes). B cell populations (splenic B-2: B220hi/CD5; peritoneal B-2: B220+/Mac-1/CD5; and peritoneal B-1: B220lo/Mac-1+/CD5+/–) were then purified by FACS using a MoFlo sorter (Cytomation, Ft. Collins, CO). Post-sort peritoneal and splenic B-2 cells were less than 1% CD3+ and less than 1.2% CD14+ by re-analysis of surface marker expression. B-2 cells constituted 21.0 ± 0.88% (mean ± SEM, n = 12) of peritoneal B cells. Splenic and peritoneal B-2 cells were also stained with antibodies to CD23, CD5, Mac-1, CD43, and CD80, and evaluated with a FACSCAN (Becton Dickinson, Franklin Lakes, NJ). Data were analyzed using FloJo software (Treestar, Ashland, OR).

In vitro viability and IgM secretion

Sorted cell populations were cultured for 3 days in wells of 96-well round-bottom plates, at 1 × 105 cells in 0.2 mL RPMI-1640 medium (BioWhittaker, Walkersville, MD) supplemented with 5% FBS, 10 mM HEPES (pH 7.25), 50 μM 2-mercapto-ethanol, 2 mM L-glutamine, 100 U/mL penicillin and 100 μg/mL streptomycin. To measure viability, B cells were stained with propidium iodide and the percentage of cells with intact (2N) DNA was determined by flow cytometry. To measure IgM secretion, 100 µL of supernatant was collected at 1, 2, 3, 4, and 5 days of culture, and subsequently assayed by ELISA. Briefly, 96-well flexible plates (Becton Dickinson) were coated with goat anti-mouse Ig (H + L) (Southern Biotechnology Associates, Birmingham, AL). Serum was diluted in PBS/0.05% Tween/0.5% BSA, and incubated on the plates for 3 h, along with IgM standards. IgM was detected using anti-mouse IgM conjugated to horseradish peroxidase (Sigma), followed by development with 2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (Sigma) and determination of absorbance at 405 nm.

Gene expression

Messenger RNA was isolated from sorted cell populations using the µMACS mRNA Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany), following the manufacturer's instructions. cDNA was prepared using AMV reverse transcriptase (Roche Applied Sciences, Indianapolis IN). Gene expression was assessed by quantitative real-time PCR using an MX3000P (Stratagene, La Jolla, CA) and the following primers: Beta2-microglobulin [CCCGCCTCACATTGAAATCC / GCGTATGTATCAGTCTCAGTGG]; elfin [ATGTGTGCACCGACTGTGGC / GGTCCTCTCCATCACTCGCG]; Integrin Beta-1 [AACAAAACTGCACCAGCCCA/ CAAACCGCAACCTGCATGA]; VH11 [GCAATAAACTACGCACCATCCA / TGTCCTCCGATCGCACATT]; VH12: [CTTCTACAACCCATCCCTCCAG / TACATGGCTGTGTCCTCTGTGG]; CD23 [CGGGAGCCCTGTGGGTTATA / CGGGAGCCCTGTGGGTTATA].

Adoptive transfer of B-2 cells

Peritoneal and splenic B-2 cells were FACS purified from male BALB/cByJ mice, aged 8–14 weeks. Cells (approximately 1 × 106) were then injected i.p. into age- and sex-matched CB.17 SCID mice in a volume of 1 mL sterile PBS. Control mice were injected with PBS alone. Except where otherwise noted, mice were sacrificed 11 days post-injection, and peritoneal cell populations were analyzed by flow cytometry using the following antibodies: anti-IgMa FITC, anti-IgMb FITC, anti-Mac-1 PE, anti-CD5 PE, anti-CD43 PE, and anti-CD23 PE. Peripheral blood was collected from recipient mice and serum IgM was measured by ELISA.

Acknowledgements

The authors thank Allen Parmelee and Sean Gurdak for excellent technical assistance in FACS purification of cells and experimental implementation, respectively. This work was supported by United States Public Health Service Grants AI29690 and AI60896 awarded by the National Institutes of Health.

Footnotes

  1. 1

    WILEY-VCH

  2. 2

    WILEY-VCH

  3. 3

    WILEY-VCH

  4. 4

    WILEY-VCH

  5. 5

    WILEY-VCH

  6. 6

    WILEY-VCH

  7. 7

    WILEY-VCH

Ancillary