Endothelial Plasmalemma Vesicle–Associated Protein Regulates the Homeostasis of Splenic Immature B Cells and B-1 B Cells

The innate immune response is the host’s first and most rapid response to infection with a pathogen, whereas the adaptive immune response involves a complex process including activation, expansion, and differentiation of pathogen-specific B and T cells. The development of adaptive immunity requires several days to weeks to generate a long-standing effector and memory immune response (1, 2). A key transition from innate to adaptive immunity is mediated by the marginal zone (MZ) B and B-1 cells because they produce the first set of low-affinity Abs against the pathogen (3). MZ B and B-1 cells are localized in the marginal sinus and peritoneal cavity, respectively, where they are favored as the first cells to sample Ags in the blood and gut. Moreover, MZ and B-1 B cells are well characterized as having a low activation threshold and their BCRs recognize a wide range of microbial Ags (4). Both B cell subsets significantly contribute to levels of serum IgM and the production of natural Abs. Natural Abs in many cases can be specific to pathogen-encoded molecules and be critical in the rapid neutralization of both viruses and bacteria (5).

MZ B cells arise from bone marrow precursors through transitional B cells, which colonize the periarteriolar lymphoid sheath (5). The differentiation of transitional B cells to MZ B cells is driven by a weak BCR activity through a dependent pathway Bruton’s tyrosine kinase (6–8). This and the interaction of NOTCH expressed on transitional B cells with the ligand, Δ-like 1, on endothelial cells induce the differentiation to MZ B cells (9). The homing of MZ B cells is dependent on circulating sphingosine-1-phosphate (S1P) binding to S1P₁ and S1P₃ receptors expressed in the endothelial cells of blood vessels of MZ (10, 11). After migration, MZ B cells are retained by the interaction of αLβ2 and α4β1 with ICAM1 and VCAM1, respectively (12).

In contrast, B-1 cells are competently produced before birth and throughout the first couple weeks after birth. The precursors for B-1 cells have been discovered in the splanchnopleural region, yolk sac and intraembryonic hemogenic endothelium, and fetal liver, but they are absent from adult bone marrow (13–16). B-1 cells constantly circulate to and from the peritoneal space across the omentum in a process that involves CXCL13, which is likely produced by macrophages (17). Collectively, these findings show that B cell progenitor migration is highly regulated by molecules expressed on endothelial cells. However, it is not known whether molecules expressed on endothelial cells are involved in B cell differentiation and trafficking.

Plasmalemma vesicle-associated protein (Plvap) is a vertebrate gene (18, 19) whose product, Plvap, is a heparin-binding (20), homodimeric, single-span type II membrane glycoprotein (21–23) critical for the formation of the stomatal diaphragms of caveolae, transendothelial channels, and vesiculo-vacuolar organelles, as well as the diaphragms of fenestrae in both mice (24–27) and humans (28). Microscopic (19, 22) and genetic (26, 28, 29) lines of investigation led to the conclusion that Plvap is specifically expressed in the endothelial cells of blood vessel capillaries and venules in select vascular beds and in the heart endocardium, and is absent from lymphatic endothelial cells. This pattern of expression was fully supported by a large body of literature obtained with two endothelial-specific mAbs that bind Plvap, such as MECA-32 (30, 31) in the mouse and PAL-E (32–34) in humans (reviewed in Refs. 35, 36). However, recently, Plvap expression was also demonstrated in the sinus lymphatic endothelial cells of peripheral lymph nodes (PLNs) while confirming its absence from peripheral lymphatics elsewhere (27) (http://immgen.org). At the whole organism level, endothelial diaphragms formed by Plvap (24, 37, 38) are critical for maintaining basal permeability of fenestrated blood vessels, with their absence resulting in disrupted blood homeostasis and reduced survival (26, 28). In PLNs, Plvap⁺ diaphragms in the sinus lymphatic endothelial cells control the entry of soluble Ags and lymphocytes in PLN parenchyma (27).

The expression of Plvap and diaphragm formation is increased in activated endothelial states associated with inflammation (39, 40) and physiological and pathological angiogenesis (35, 41, 42) where it has active roles. Plvap is required for cancer progression (43) and diapedesis of leukocytes into inflammation sites in vivo (31). In vitro Plvap knockdown and Ab-mediated blockade experiments suggest that endothelial Plvap is important for the transcellular transmigration, but not for adhesion and rolling of lymphoblasts, with no effect on neutrophils transmigration (31). Plvap is thought to control the transcellular migration of lymph-borne lymphocytes into PLN parenchyma (27). Deletion of Plvap results in defective PLN morphogenesis with mild decreases in the T cell compartment (both CD4 and CD8 T cells), hyperplastic B cell follicles, and increases in both PLN B and T cell activation. Intriguingly, Plvap deletion increases the entry of adoptively transferred lymph-borne splenocytes (both B and T cells), whereas its ligation with MECA-32 Ab inhibits the recruitment of these subsets (27). The mechanism of how Plvap mediates transendothelial migration of immune cells is currently unclear.

In this study, we have examined whether Plvap plays a role in the development and homeostasis of hematopoietic lineages, taking advantage of recently created genetic models of Plvap gain and loss of function and endothelial-specific reconstitution (26). Our studies show that deletion of Plvap results in an intense reduction of IgM⁺ B cells in both spleen and peritoneal cavity. Tissue-specific deletion of Plvap demonstrates that the defect is B cell extrinsic, because B cell and pan-hematopoietic Plvap deletion has no effect on IgM⁺ B cell numbers. Endothelial-specific deletion of Plvap recapitulates full Plvap knockout, whereas endothelial-specific reconstitution of Plvap rescues the IgM⁺ B cell phenotype. Taken together, these results demonstrate that Plvap expression on endothelial cells is key in the maintenance of IgM⁺ B cells into spleen and peritoneal cavity.

Homozygous Plvap^loxP (Plvap^L/L) mice were generated by knock-in using homologous recombination in mice, as already described (26). Plvap^L/L mice express Plvap at normal levels and have no overt phenotype (26). The Plvap^L/L mice were bred to mice expressing the cre recombinase under the control of different promoters to generate compound mice where Plvap was deleted in: 1) the germline (label Plvap^−/−, genotype Plvap^−/−;CMV-cre^tg/+) using CMV-cre^tg/+ transgenic mice (JAX strain BALB/c-Tg^{(CMV-cre)1Cgn/J}); 2) endothelial and hematopoietic cells in the embryo using Ins-VEC-cre transgenic mice (44) (label Plvap^ECKO-VEC, genotype Plvap^L/L;Ins-VEC-cre^tg/+); 3) endothelial and hematopoietic cells in the embryo using Tie2/Tek-cre^tg/+ mice [JAX strain B6.Cg-Tg(Tek-cre)12Flv/J] (label Plvap^ECKO-Tie2, genotype Plvap^L/L;Tek-cre^tg/+); 4) B cells in the embryo using CD19-cre^KI/+ knock-in mice (JAX strain B6.Cg-Cd19^tm1(cre)Cgn/J) (45) (label CD19^CrexPlvap^L/L, genotype Plvap^L/L;CD19-cre^ki/+); and 5) hematopoietic cells in the embryo using Vav1-cre^+/− transgenic mice (46), for deletion of Plvap in all the hematopoietic cell lineages, but not in endothelium (44, 46). For inducible deletion of Plvap in the endothelial cells of the adult mice, we used end-SCL-Cre-ERT^tg/+ transgenic mice to generate mice with the genotype Plvap^L/L; end-SCL-CreER^{T tg/+} (label Plvap^iECKO). Deletion of Plvap was achieved by dosing 4-wk-old mice (both males and females) by gavage with seven doses of 4 mg of tamoxifen spaced at 48 h. Experiments were carried out 2 wk after the last tamoxifen dose administration. Control animals for germline deletion were sex- and age-matched wild type (WT) or CMV-cre^tg/+ (labeled WT) and Plvap^+/− or Plvap^+/−;CMV-cre^tg/+ (labeled Plvap^+/−) littermates. Control animals for tissue-specific deletions were sex- and age-matched WT or Plvap^L/L littermates.

VEC-Plvap-HA^tg/+ transgenic mice (26) that express Plvap-HA fusion protein specifically in the endothelial cells under the control of cadherin 5 (VE Cadherin) promoter were used to reconstitute Plvap in endothelial cells in the context of the Plvap^−/− (label PV1^ECRC) as described (26). The same strategy was used to reconstitute Plvap in the context of Plvap^iECKO (label Plvap^iECRC, genotype Plvap^L/L; end-SCL-CreER^{T tg/+}; VEC-Plvap-HA^tg/+).

All animals were maintained in a pathogen-free facility at Dartmouth College. All procedures were approved by the local Institutional Animal Care and Use Committee.

The following Abs and staining reagents were used: IgG1 (clone A85-1), IgG2a/b (clone R2-40), CD138 (clone 281-2), IgM (clone 11–41), CD24 (clone M1/69), CD4 (clone RM4-5), CD21/35 (clone 7E9), CD8 (clone 53-6.7), CD25 (clone 3C7), CD62L (clone MEL-14), CD69 (clone H1.2F3), CD44 (clone lM7), CD11c (clone HL3), CD80 (clone 16-10A1), CD86 (clone GL1), CD11b (clone M1/70), c-Kit (CD117, clone 2B8), and streptavidin-PerCP were from BD Pharmingen; Plvap (clone MECA-32) was from Abcam; CD38 (clone 90), B220 (clone 6B2), CD23 (clone B3B4), IgD (clone 11-26c), and FceRI (clone MAR-1) were from eBioscience; and peanut agglutinin was from Vector Laboratories. MECA-32 mAb (anti-mouse Plvap rat IgG2a) secreting hybridoma was obtained from the Developmental Studies Hybridoma Bank (University of Iowa) and was produced in serum-free conditions by BioXCell (Lebanon, NH). The Ab was labeled with Alexa Fluor (AF) fluorochromes using the protein labeling kits for AF488 and AF647 (Invitrogen), as per manufacturer’s instructions. Flow cytometry was performed on either a refurbished FACSCAN or a FACSCalibur running CellQuest software (BD), or a FACSCanto running FACSDiva software within the Norris Cotton Cancer Center DartLab Immune Monitoring Facility. The data analysis was performed using FlowJo (Tree Star).

To analyze B cells, T cells, and monocytes, we prepared single-cell suspensions of lymphocytes from spleens by mechanical disruption in HBSS followed by passing the cells through a 70-μm cell strainer. In experiments where dendritic cells were profiled, an enzymatic digestion step (45 min, 37°C) was included using DNAse (10 mg/ml; Roche) and Liberase (12.5 mg/ml; Roche) before mechanical disruption. Cells were collected by centrifugation (5 min, 500 × g, 4°C), and RBCs were lysed (2 min, 37°C) using RBC Lysis buffer (BioLegend). Total number of cells and cell viability were determined using either a hemacytometer and trypan blue or a Guava system (Millipore). Cells (10⁵–10⁶) were stained with Ab cocktails, as noted. Analysis was performed after costaining with the mixture of Abs by flow cytometry. For isolation of peritoneal cells, the peritoneal cavity was flushed with 5 ml of warm (37°C) PBS, 2% BSA, 2 mM EDTA, 0.02% sodium azide, and 10 U/ml heparin.

Female WT C57Bl6/J mice were treated with either 50 μg of LPS (E. coli 055:B5; Sigma) in PBS or an equal volume of PBS alone (control mice) i.p. 12 h before spleens were harvested, enzymatically dissociated, as noted earlier, followed by Ab staining of splenocytes, and analyzed by flow cytometry.

Tissues were snap-frozen in optimal cutting temperature medium and sectioned to 8 μm. Sections were collected on charged slides (Surgipath), fixed (−20°C, 10 min) with cold methanol, rinsed (3 × 2 min, room temperature [RT]) in PBS, encircled with hydrophobic barriers (PapPen), blocked (30 min, RT) with 10% rat serum in PBS containing 10 μg/ml mouse Fc block, incubated (1 h, RT, in dark) with various fluorescently labeled primary Ab cocktails in blocking buffer, rinsed (3 × 5 min, RT, in the dark) again in PBS, stained (10 min, RT, in the dark) with 300 nM DAPI (D1306; Life Technologies), and washed (3× 5 min, RT) in PBS. Labeled sections were mounted under #1.5 coverslips using a polymerizing mounting medium (Fluoromount G; Southern Biotech). The Abs used were rat anti-mouse Plvap-AF568 (clone MECA-32), rat anti-mouse CD169-FITC (clone MOMA-1; AbD Serotec), and rat anti-mouse/human B220-AF647 (clone 6B2). Labeled sections were analyzed using a Zeiss LSM510 Meta confocal microscope equipped with appropriate lasers (405, 488, 532, 633 nm) and filters, all within the Norris Cotton Cancer Center microscopy facility. The acquired images were processed for brightness and contrast and analyzed using ImageJ (http://imagej.nih.gov/ij/), and the figures were mounted using Adobe Photoshop and Adobe Illustrator CS6.

Results are expressed as mean ± SEM. A two-tailed Student t test with unequal variance was used to evaluate the statistical significance of the data.

Plvap is an endothelial protein that is involved in the diapedesis of leukocytes at sites of inflammatory challenge and PLN sinuses. Diapedesis is a process that is central to the development and homeostasis of hematopoietic lineages. To understand the role of Plvap in these processes, we used recently generated Plvap^−/− mice (26) to characterize the function of this molecule on the numbers and subset composition of leukocytes in the blood, spleen, PLNs, and Peyer’s patches. Plvap^−/−, Plvap^+/−, and WT littermate control mice were profiled by flow cytometry to determine whether there are modifications in terms of leukocyte subset numbers, frequency, or function. No differences were found in cellular composition of peripheral blood, LN, and Peyer’s patches with respect to percentages, number and viability of granulocytes, T and B lymphocytes, NK cells, and monocytes (Fig. 1A, Supplemental Fig. 1, and data not shown). In the spleen, there was a drastic reduction in the percentage and absolute number of IgM⁺ IgD⁻ B cells in Plvap^−/− mice (Fig. 1B, 1C), whereas IgD⁺ B cells were not affected (Fig. 1B).

In the spleen, several prominent B cell subsets are represented (47). To evaluate whether the reduction of this population is due to a reduction of a specific subpopulation of splenic B cells, we stained with a panel of Abs that identify those subpopulations (47). Neither the percentage of follicular (CD21/35^int IgM^int) or MZ B cells (CD21/35⁺ IgM⁺) is affected in Plvap^−/− mice (Fig. 1D). Interestingly, CD21/35^lo IgM⁺ B cells are reduced in the Plvap^−/− mice compared with the WT or Plvap^+/− mice (Fig. 1D, 1E). Lastly, CD21/35^lo IgM⁺ B cells can be subdivided in transitional 1 (T1) or transitional 2 (T1) B lymphocytes using the expression of CD23 and HSA (Fig. 1F). Our results show that the proportion of T1 or T2 B cells is not affected in the Plvap^−/− mice (Fig. 1F, 1G), indicating that both T1 and T2 B cells are reduced. Taken together, these results show that among the splenic B cell populations, in Plvap-deficient mice there is a selective reduction in transitional IgM⁺ B cells.

Similarly, no difference in the proportion of transitional B cells in Plvap^−/− mice (data not shown) was found when expression of CD93 was used for the analysis of transitional B cells (48). Thus, irrespective of the markers used to analyze transitional B cells, we obtain the same results.

Based on the earlier observations showing a decrease of IgM⁺ B cells in the spleen of Plvap^−/− mice, we hypothesized that the same reduction would also be found in peritoneal B-1 B cells. Profiling of the resident leukocytes in the peritoneum obtained by peritoneal lavage demonstrated a drastic reduction in the total number of cells in the peritoneum of Plvap^−/− mice compared with WT or Plvap^+/− littermates (Fig. 2A). The reduction in total viable leukocyte numbers in the peritoneal cavity was accompanied by a low percentage and absolute number of IgD⁺ B cells and the absence of IgM⁺ IgD^low B cells (Fig. 2B–D). Taken together, these results suggest that Plvap plays a role in the recruitment or retention of B cells into the peritoneal cavity and/or their survival.

Because of the drastic reduction of cells in the peritoneum cavity of Plvap^−/− mice, we also analyzed the percentage of mast cells and monocytes in this compartment. Our results show that mast cell (cKit⁺ FceRI⁺ cells; Supplemental Fig. 2A) and monocyte (CD11b⁺ Gr-1⁺ cells, data not shown) percentages were also reduced in Plvap^−/− mice compared with control mice, indicating that Plvap is an important molecule in migration to or retention into the peritoneum.

To understand the underlying mechanisms responsible for altered B cell frequencies in the spleen and peritoneum, we sought to determine whether these effects are B cell intrinsic or extrinsic. First, we inquired whether Plvap was expressed on hematopoietic cells, and second, we established the impact on B cell frequencies of genetic B cell–specific and pan-hematopoietic deletion of Plvap.

LPS-activated or resting B cells or dendritic cells did not express Plvap, as detected with Ab staining and using Plvap^−/− splenocytes as controls. Furthermore, CD4⁺ and CD8⁺ T cells were also negative for Plvap expression (Fig. 3A–C). Examination of Plvap expression by confocal microscopy revealed that expression was limited to the splenic blood vessels in the MZ area (determined by MOMA-1⁺ macrophage localization) (Fig. 3D). Taken together, these data suggest that Plvap is not expressed within the hematopoietic subsets studied, in agreement with recent data published by the Immunological Genetic Consortium (http://immgen.org) (49).

Even though expression analysis conclusively established the lack of hematopoietic expression of Plvap, genetic deletion of Plvap in hematopoietic cells and subsets was used to confirm these findings. Plvap^L/L mice were interbred with CD19-Cre (45) or Vav1-Cre mice (46) to obtain compound mice lacking Plvap in the B cells (CD19^Cre × Plvap^L/L) or all hematopoietic cells (Vav^Cre × Plvap^L/L) (26), respectively. Vav^Cre × Plvap^L/L mice have >97% Plvap deletion in the hematopoietic compartment (26). B cells isolated from CD19^Cre × Plvap^L/L mice using magnetic separation showed >95% deletion of Plvap (data not shown). Immune profiling of the peritoneal lavage showed that there was no effect on B frequencies or phenotype in the spleen and peritoneum in either CD19^Cre × Plvap^L/L (Fig. 4A–C) or Vav1^Cre × Plvap^L/L (Fig. 4D–F) mice. The combination of microscopy, flow cytometry, and genetic data clearly demonstrate that Plvap is not expressed in the hematopoietic compartment and that the maintenance of IgM⁺ B cells in both spleen and peritoneal cavity is due to Plvap expression outside the hematopoietic compartment.

Plvap is a molecule specifically expressed in endothelial cells of blood vessels (26, 29) and PLN sinus (27), suggesting that endothelial cells might impact on the presence of IgM⁺ B cells in spleen and peritoneum. For this purpose, we made use of Tie2^Cre × Plvap^L/L mice (26) where Plvap is efficiently (>95%) deleted in the endothelial cells and the hematopoietic compartment at the embryonic stage. The results obtained in Tie2^Cre × Plvap^L/L mice phenocopied those obtained in Plvap^−/− mice. There was a reduction in IgM⁺ IgD⁻ B cell percentage in the spleen from Tie2^Cre × Plvap^L/L compared with control littermates (Fig. 5A–C). In addition, there also was a striking reduction in the percentage of IgD⁺ and IgM^pos IgD⁻ B cells in the peritoneal cavity of Tie2^Cre × Plvap^L/L compared with control mice (Fig. 5D–F). These data suggest that the expression of Plvap on endothelial cells regulates the numbers of IgM⁺ B cells in the spleen and peritoneal cavity.

To demonstrate that Plvap^−/− phenotype is not due to a distortion of other genetic loci close to the Plvap locus, we transgenically complemented the Plvap deficiency by the expression of Plvap-HA in endothelial cells of Plvap^−/− mice (26). Previously, we have generated mouse lines (VEC-Plvap-HA^+/tg) (26) expressing Plvap under the control of the VE Cadherin promoter and 5′ intronic enhancer (50). We used VEC-Plvap-HA^+/tg and Plvap^+/− mice to generate compound Plvap^−/−; VEC-Plvap-HA^+/tg (Plvap^ECRC) mice, which display between 30 and 50% reconstitution of native endothelial Plvap levels (26).

As shown in Fig. 6, IgM⁺ IgD^low B cells are recovered to normal levels in both spleen (Fig. 6A–C) and peritoneum (Fig. 6D–F) of Plvap^ECRC mice compared with Plvap^−/− mice. Moreover, IgD⁺ B cells were also recovered in the peritoneum of Plvap^ECRC mice (Fig. 6D–F). Together, all these results suggest that the expression of Plvap in endothelial cells is necessary to maintain normal levels of B cells in the spleen and peritoneum compartment.

To determine whether Plvap deletion leads to a defect in B cell development in bone marrow, we profiled B cells progenitors in the marrow of Plvap^L/L; end-SCL-CreER^{T tg/+} compound mice (labeled Plvap^iECKO) and mice with endothelial reconstitution of Plvap in the context of the Plvap^iECKO (labeled Plvap^iECRC mice) (see 2Materials and Methods). The promoter driving the expression of CreER^T fusion protein consists of SV40 virus minimal promoter and the endothelial enhancer of the stem cell leukemia gene (SCL) (51). This compound promoter confers activity in endothelial cells from select vascular beds, bone marrow included, as previously shown (51, 52). To minimize the effects on the peritoneal cavity, we administered tamoxifen by gavage.

As shown in Fig. 7A and 7B, IgM⁺ IgD^low B cells are reduced in both spleen (Fig. 7A, 7B) and peritoneum (Fig. 7C, 7D) of Plvap^iECKO mice compared with control (tamoxifen-treated Plvap^L/L) mice, whereas endothelial reconstitution in Plvap^iECRC mice reverses the effect. In addition, IgD⁺ B cells were also reduced in the peritoneum of Plvap^iECKO mice (Fig. 7C, 7D), indicating that endothelial deletion in adult mice has a similar phenotype as that obtained in Plvap^−/− mice with respect to B-1 and MZ B cells (Fig. 1).

Using these models, we analyzed the proportion of immature B cells (B220⁺IgM^hiIgD⁺), mature B cells (B220⁺IgM^intIgD⁺), and pre-B cells (B220⁺IgM⁺IgD⁺) from bone marrow (53, 54). However, no changes were detected in B cell subsets in the bone marrow of Plvap^iECKO mice as compared with controls or Plvap^iECRC mice (Fig. 7E–G). These results together suggest that the defect observed in the splenic immature B cells and peritoneum B cells are not due to a defect in the B cell development in the bone marrow.

The development of effective immune responses is dependent on endothelial cell–mediated leukocyte migration into sites of inflammation. To our knowledge, the results presented in this article are the first to determine the critical role of Plvap in the maintenance of IgM⁺ B cells in spleen and peritoneum. We show that the lack of Plvap results in a reduction of transitional splenic IgM⁺ B cells as well as B-1 B cells in the peritoneum. The tissue-specific deletion of Plvap clearly demonstrates that the role of Plvap is extrinsic to the B cell compartment, because B cell and pan-hematopoietic Plvap deletion has no effect on the IgM⁺ B cells frequencies in either the spleen or the peritoneal cavity. In contrast, endothelial-specific deletion of Plvap recapitulates full Plvap knockout phenotype, and endothelial-specific reconstitution of Plvap in the context of germline Plvap knockout rescues the IgM⁺ B cell phenotype. Taken together, these results conclusively demonstrate that Plvap expression in endothelial cells is key in maintenance of IgM⁺ B cells in spleen and peritoneal cavity. The reduced number of immature B cells in spleen and B-1 cells in the peritoneum of Plvap-deficient mice could also explain the previously reported marked reduction of IgM and IgA titers in Plvap-deficient mice (26), further suggesting that Plvap may regulate the abundance of B cells that produce natural Abs.

Previous reports have shown Plvap expression in spleen (19, 55) (http://immgen.org), especially on the endothelial cells of the marginal sinus of the spleen (55). Our observations confirm these findings by showing Plvap expression on vessels in the MZ of spleen (identified by colocalization of MOMA-1⁺ macrophages) as well as capillaries in the red pulp. Thus, Plvap is expressed at the site where it could regulate IgM⁺ B cell trafficking or retention in this histological site. Although human Plvap expression has been reported in circulating human lymphocytes and monocytes (i.e., PBMCs) by intracellular staining (31), we could not find Plvap expression by Ab staining in a variety of circulating or parenchymal murine immune subsets. Our data are in accordance with RNA sequencing data on Plvap expression, published by http://immgen.org (49, 56). In addition, we found that neither B cells nor dendritic cells express Plvap in basal or under inflammatory (LPS stimulation) conditions. Furthermore, when we abrogate Plvap expression on B cells or in the hematopoietic compartment, there is no effect on the abundance of IgM⁺ B cells in the spleen or peritoneum. These findings suggest that if Plvap has a B cell trafficking or retention function in the spleen or peritoneum, this role is not B cell or hematopoietic cell intrinsic. In contrast, when we used a mouse model where the lack of Plvap was specific to endothelial cells, the defect in B cell migration or retention was indistinguishable from that observed in the global Plvap-deficient mice. Furthermore, the transgenic overexpression of Plvap expression on endothelial cells in Plvap-deficient mice rescued the phenotype. Together, these results demonstrate that Plvap expression in endothelial cells is key in the B cell trafficking or retention in the spleen and peritoneum.

Considering the data in the literature, there are several mechanisms by which Plvap may reduce the number of B-1 and MZ B cells. First, Plvap might hamper the generation of the B-1 B cell hematopoietic progenitors in the vascular plexus of the amniotic sac and hemogenic endothelium (16, 57, 58) where Plvap is expressed very early (18, 59). Preliminary data show that the number of B-1 B cell progenitors in the amniotic sac (M. Yoshimoto, M.C. Yoder, and R.V. Stan, unpublished observations) at embryonic day 9 in the WT, Plvap^L/L, and Plvap^−/− mice is similar. In addition, B cell precursors in the bone marrow are also similar in the Plvap-deficient mice (Fig. 7), which makes Plvap involvement in B cell development unlikely.

Second, Plvap controls B-1 and MZ B cell recruitment and/or retention in the peritoneum and spleen, respectively. Integrins, chemokines, and other adhesion molecules have been shown to be involved in B cell migration (12, 60–62). Different types of adhesion molecules regulate distinct events in lymphocyte extravasation. We know from these studies that integrins play a role in leukocyte tethering, whereas chemokines play a role in the rolling process (63). Immature B cells express αLβ2 and α4β1, which retain B cells in the splenic marginal sinus, where they interact with their ligand, ICAM1 and VCAM1 expressed in endothelial cells (12). In addition, an increased gradient of CXCL13 in the periarterial lymphocyte sheath induces the migration of B cells from the MZ to the follicles (64). In contrast, S1P₁ expressed on B cells is able to induce the migration of B cells to the sinus of the MZ and peritoneum (10, 11, 65). Conceivably, S1P₁ expressed on B cells could interact with Plvap, through vim, to permit the extravasation of B cell progenitors to the peritoneum and the splenic marginal sinus. However, additional studies are necessary to test this hypothesis.

Plvap may also be an integral mediator of inflammation-induced migration. It has been previously observed that the blocking of Plvap in vivo reduced the number of monocytes found in the peritoneum after the induction of peritonitis (31). Our results show that Plvap has a novel role in the progression of leukocyte migration to peritoneum, not just in inflammation but also in the steady-state. The global deletion or specific obliteration of Plvap in endothelial cells reduced the migration not just of B cells but also others leukocytes by ∼80% in steady-state. In addition, when we restore the expression of Plvap exclusively in endothelial cells, we observed the rescue of B cells number in steady-state, suggesting that Plvap expression in endothelial cells is the master regulator of immune cell migration to peritoneum.

Third, Plvap^−/− vessels leak plasma components leading to hypoproteinemia and formation of ascites, a condition known to reduce viability of resident peritoneal macrophages and other immune cell subsets. Deletion of Plvap may induce conditions that deplete the peritoneal progenitors of B-1 and MZ B cells. Experiments involving Plvap deletion at adult stages should shed light on which mechanism(s) are involved.

In conclusion, the expression of the glycoprotein Plvap on endothelial cells is a key regulator of B cells in the peritoneum, as well as immature IgM⁺ B cells in the spleen. It also has a physiological role in facilitating the production of natural IgM and IgA Abs by peritoneal and splenic IgM⁺ B cells. Collectively, our findings point to a novel immunological significance for Plvap in innate humoral immunity. Future work will determine the precise mechanism by which Plvap regulates B cell subsets in the peritoneum and spleen.

We thank Dr. T. Graf (Centre for Genomic Regulation, Barcelona), Dr. N. Speck (University of Pennsylvania), Dr. M. Chen (Harvard University), Dr. Patricia A. Ernst (University of Colorado Denver), and Dr. M. Yoshimoto and Dr. M. Yoder (University of Indiana) for reagents and suggestions.

The authors have no financial conflicts of interest.