Enhanced trafficking to the pancreatic lymph nodes and auto-antigen presentation capacity distinguishes peritoneal B lymphocytes in non-obese diabetic mice
NOD.Igµnull mice lacking mature B cells are highly resistant to diabetes and display poor CD4 T cell responses to autoantigens. Nevertheless, the degree to which different B cell subsets contribute to diabetes in NOD mice remains unresolved. Due to their role in the recognition of microbial and autoantigens, peritoneal B cell characteristics were examined in NOD mice to see if they differ developmentally, phenotypically or functionally in aspects relevant to diabetogenesis.
The population dynamics, activation state, migratory behaviour and antigen presentation function were investigated in NOD peritoneal B cells.
NOD peritoneal B cells were found to express abnormally high levels of co-stimulatory molecules (CD40, CD86 and CD69). In contrast, the expression of l-selectin and integrin α4β1 was markedly reduced in NOD mice compared with BALB/c and C57BL/6 mice. The number of B cells in the peritoneum was lower in NOD than in control mice throughout development; migration of B cells from the peritoneum to the pancreatic lymph nodes in NOD mice was enhanced tenfold. NOD B cells showed no chemotactic response to sphingosine-1-phosphate, which normally acts to retain B cells in the peritoneum. Peritoneal B cells of NOD mice also presented insulin autoantigen to CD4 T cells, inducing T cell proliferation.
NOD peritoneal B cells are hyperactivated, migrate to the pancreatic lymph nodes and are capable of driving insulin-specific CD4 T cell activation. These characteristics could make them important for inducing or amplifying T cell responses against islet-antigens.
KeywordsAntigen presentation Costimulatory molecules Migration Non-obese diabetic mouse Pancreatic lymph nodes Peritoneal B cells Sphingosine-1-phosphate
5,6-Carboxyfluorescein diacetate succinimidyl ester
Endothelial differentiation sphingolipid G-protein-coupled receptor 1
Pancreatic lymph nodes
The importance of B cells for the development of type 1 diabetes in NOD mice has become increasingly evident over the last 10 years [1, 2, 3, 4, 5, 6]. Several reports have demonstrated that eliminating B cells protects against type 1 diabetes in mice [1, 3, 6]. B1 cells moreover produce natural antibodies and recognise common bacterial antigens as well as self-antigens . The specific role of NOD marginal zone B cells has been evaluated in a type 1 diabetes context [4, 5]. These B cells expand with the onset of diabetes and are able to present insulin peptides to diabetogenic T cells .
Peritoneal B cells, however, have not as yet been researched to the same extent as other B cell subsets. Since peritoneal B cells comprise an exceptionally high concentration of B1 type B cells, which are of particular interest in autoimmunity, a thorough examination of the role of peritoneal B cells in type 1 diabetes is paramount. Nonetheless, some evidence exists for the importance of peritoneal B cells, and more specifically of the peritoneal B1 cell subset, in type 1 diabetes. Many islet-infiltrating B cells in NOD mice express the CD5 antigen, which is also expressed by most B1 cells . These B1-like B cells have been reported to decrease in number in infiltrates following hypotonic lysis of peritoneal cells. The depletion of peritoneal B cells also protects NOD mice from type 1 diabetes . Insulin autoantibodies measured by competitive radiobinding assays are typically of high affinity in humans and in NOD mice [8, 9]. However, low-affinity insulin-binding autoantibodies that interact with multiple antigens have been reported in young NOD mice . Thus some of the early repertoire of insulin autoantibodies in NOD mice appear to share characteristics with natural antibodies. B1 cells could hence be implicated as producers of autoreactive antibodies such as insulin autoantibodies and in the presentation of insulin- and/or other islet-antigens to diabetogenic T cells. This would, however, require that they endocytose islet-antigens and provide co-stimulatory signals for T cells.
Sphingosine-1-phosphate (S1P) has received considerable attention in recent years as a regulator of lymphocyte trafficking. In the peritoneum, S1P helps retain B lymphocytes within the peritoneal cavity . B cell unresponsiveness to chemotactic signals from S1P thus facilitates the emigration of peritoneal lymphocytes. An analogous modulator of peritoneal cell migration is integrin α4β1. High expression of this integrin on B1 lymphocytes keeps these lymphocytes in the peritoneal cavity . Activating signals such as lipopolysaccharide have the effect of impeding S1P signalling  as well downregulating expression of integrin α4β1, consequently permitting B cell exit.
An interesting aspect of the peritoneal B cell subset is its location adjacent to, and interrelation with the intestine. Peritoneal B cells in mice respond to microbial and inflammatory signals derived from the gastrointestinal tract . Moreover, Turley and collegues have suggested the existence of an inherent trafficking route between the peritoneum and the pancreatic lymph nodes (PaLN) . This provides a unique opportunity for peritoneal B cells to directly interlink environmental and microbial factors with autoimmune manifestations in the PaLN.
Non-obese diabetic mice were used at 5 to 10 weeks, unless otherwise indicated; age-matched BALB/c and C57/Bl6 mice were used as controls. The mice were raised and bred in the Central Animal Laboratory of Turku University. All animal experiments were approved by the National Laboratory Animal Care and Use Committee in Finland and conformed to the legal acts, regulations and requirements of the European Union concerning the protection of animals used for research.
Peritoneal cells were obtained by washing out the peritoneum with cold RPMI-1640 medium (Gibco, Grand Island, NY, USA). Single-cell suspensions of spleen cells and lymph node cells were acquired by gently pressing the tissues through a metal mesh. Erythrocytes were extracted by hypotonic lysis using 0.2% (wt/vol.) NaCl.
Immunofluorescence and flow cytometry
To assess B cell types and numbers, peritoneal cells were incubated either with biotinylated anti-CD45R (R&D Systems, Minneapolis, MN, USA) followed by incubation with streptavidin–allophycocyanin (Invitrogen, Carlsbad, CA, USA), or with allophycocyanin-conjugated anti-CD45R (Caltag Laboratories, Burlingame, CA, USA). To distinguish between B1 and B2 cells, either CD11b FITC (which stains B1a and B1b type B1 cells) (Immunotools, Friesoythe, Germany) or CD5 PE (for staining of B1a cells) (R&D systems) were used. After surface staining, cells were analysed using FACS Calibur. To examine the surface expression of various activation markers and integrins, the following PE-conjugated antibodies were used in addition to anti-CD45R and CD11b: CD40 (BD Pharmingen, Franklin Lakes, NJ, USA), CD86, Integrin α4 (BD Biosciences, San Jose, CA, USA, and AbD Serotec, Oxford, UK), Integrin β1 (AbD Serotec), Integrin α4 β7 heterodimer (DATK32; BD Pharmingen) and CD69 (Immunotools). The L-selectin antibody that was used was FITC-conjugated (Immunotools). S1P receptor 1 surface expression analysis was performed using rabbit anti-endothelial differentiation sphingolipid G-protein-coupled receptor 1 (EDG1) antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and Alexa-fluor 488-labelled anti-rabbit IgG (Invitrogen) as the second-step antibody. Flow cytometry was performed using FACS Calibur (BD Biosciences) and Cell Quest (BD) software.
Adoptive transfer experiments
Peritoneal cells from BALB/c and NOD mice were labelled with Vybrant DiI dye (Molecular Probes, Eugene, OR, USA) according to the manufacturer’s recommendations. We injected 2 × 106 cells intraperitoneally into each recipient mouse. The mice were killed 20 h later and peritoneal cells, mesenteric lymph nodes, spleens and PaLN were isolated. The cells obtained were stained with fluorochrome-conjugated CD45R and CD11b antibodies. Frequencies of adoptively transferred B cells in collected samples were analysed by flow cytometry and the results expressed as ‰ of total (DiI-negative) B cells collected for each sample of mesenteric lymph nodes, spleen and PaLN. The decrease in DiI-labelled B cells in the peritoneum at 20 h after transfer was determined by measuring DiI + B cells as percentage of total DiI + cells at the time of transfer vs 20 h post-transfer. Results are expressed as per cent DiI-labelled B cells left in peritoneum relative to input.
In vitro cell migration experiments
Migration experiments were performed using 96-well Transwell Permeable Supports (Corning, Corning, NY, USA) with 5 μm pore sizes. Peritoneal cells from NOD and BALB/c mice were isolated. Macrophages were removed by adherence to plastic for 1 to 2 h in RPMI medium containing 10% FCS. This step typically yielded a cell suspension with approximately 90% purity of B cells. Serum-free medium supplemented with 0.1% fatty acid free BSA was added to the lower wells with different concentrations (1, 5, 10, 30, 100, 500 or 1,000 nmol/l) of S1P (Biomol International, Plymouth, UK) or without S1P (control). The enriched B cells were washed in RPMI, resuspended in serum-free medium supplemented as described above and added to the upper wells (75,000 cells/well). In some experiments, B cells were pre-incubated for 1 h with 1 μmol/l VPC23019 (Avanti Polar Lipids, Alabaster, AL, USA), an antagonist of S1P receptors 1 and 3. VPC23019 was additionally added at 1 μmol/l to the upper and lower wells of the migration assay in these experiments. Cell migration was allowed to proceed for 1 h at 37ºC. Cells that had migrated to the lower wells were then collected and quantified by flow cytometry. Quantification was based on cell counts obtained by flow cytometry during a set time constant (100 s) with identical cell-suspension volumes. The margin of error using this method of analysis was approximately 10%, which was roughly the same as when using TruCount fluorescent beads for quantification with flow cytometry.
Peritoneal cells from NOD and BALB/c mice were isolated. Macrophages were removed by adherence to plastic as described above. Whole-cell lysates from B cell-enriched peritoneal cells were obtained by incubating the cells on ice for 1 h, under agitation in a lysis buffer containing 20 mmol/l Tris base, 150 mmol/l NaCl, 0.5% (vol./vol.) Nonidet P-40, 2 mmol/l EDTA, 1 mmol/l phenylmethylsulphonyl fluoride and aprotinin 0.01 trypsin inhibitor units per ml (pH 8.0). The cell suspension was then centrifuged at 15,000×g for 15 min at 4ºC, after which the supernatant fraction was collected. The protein concentration in the sample was determined with a BSA protein assay reagent kit (Pierce, Rockford, IL, USA) according to the manufacturer’s recommendations.
To determine peritoneal cell S1P receptor 1 and 3 levels, protein lysates were mixed 3:1 in 3× SDS buffer (30% [vol./vol.] glycerol, 3% [wt/vol.] SDS, 187.5 mmol/l Tris/HCl, pH 6.8, 0.015% (wt/vol.) Bromophenol Blue and 3% [wt/vol.] 2-mercaptoethanol). The mixture was then heated for 4 min at 95°C. Western blotting was performed in accordance with Balthasar et al. . Rabbit anti-EDG1 and rabbit anti-EDG3 (Santa Cruz Biotechnology) were used as primary antibodies and HRP-conjugated goat anti-rabbit IgG (BioRad Laboratories, Hercules, CA, USA) as the secondary antibody. Finally, membranes were stripped as described  and secondary probing with the loading control β-actin (Cell Signalling Technology, Danvers, MA, USA) was performed. Densitometric analysis of the blots was performed using ImageJ 1.41 software (US Department of Health and Human Services, USA; http://rsbweb.nih.gov/ij/) and S1P receptor blot intensities were correlated to the β-actin loading control blot intensities.
In vitro S1P-induced CD69 downregulation
Peritoneal cells were incubated overnight in serum-free medium supplemented with 0.1% fatty acid free BSA, 2 mmol/l l-glutamine and 100 units/ml penicillin and streptomycin, either in the presence or in the absence of S1P (200 nmol/l). The cells were collected and incubated with antibodies against B220/CD45R, CD11b and CD69 and analysed using flow cytometry.
In vitro spleen and peritoneal cell proliferation assay
Peritoneal and spleen cells from BALB/c and NOD mice were labelled with 5,6-carboxyfluorescein diacetate succinimidyl ester (CFSE) (5 μmol/l; Molecular Probes), diluted to a concentration of 1 × 106 cells/ml and cultured for 72 h on 96-well culture plates in culture medium alone or with the addition of anti-mouse CD40 antibody (3 μg/ml; BD Biosciences), CpG oligodeoxynucleotide (CpG) (1 μg/ml; CpG 1826; Coley Pharmaceutical, Langenfeld, Germany) or lipopolysaccharide (10 μg/ml; Sigma-Aldrich, St Louis, MO, USA). Finally, cells were collected from the wells, stained with anti CD45RB/B220-APC and CD5-PE conjugated antibodies, and subjected to flow cytometric analysis.
Antigen presentation assay
NOD and BALB/c mice were immunised subcutaneously on the hind flank with 50 μg insulin peptide (Insulin B [9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23]; Anaspec, San Jose, CA, USA). After 10 days, spleens were collected from these animals and splenic CD4+ T cells purified using CD4 MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany). These purified T cells were co-cultured (200,000 cells/well) with peritoneal or splenic B cells (150,000 cells/well) purified with B220 MicroBeads (Miltenyi Biotech). Additionally, either insulin peptide (4 μmol/l or 40 μmol/l; Insulin B 9-23; Anaspec) or intact insulin (20 μg/ml or 200 μg/ml; Sigma-Aldrich) was added to some of the wells. The cells were incubated for 72 h at 37ºC with the addition of [3H]thymidine (148 kBq/ml) during the last 16 h of incubation. Finally, cells were collected using an automatic cell harvester (Harvest 96; Tomtec, Frankfurt, Germany) and radioactivity was counted using a beta counter (Wallac, San Diego, CA, USA). Each experiment was performed in triplicate.
Statistical analysis of the data
The results are presented as averages of at least three independent experiments ± SEM. All individual in vitro cell migration experiments were performed in duplicates. The Student’s t test was used for comparison between two groups. One-way ANOVA and relevant post-hoc tests were used, if more than two groups were compared.
The results for the adoptive transfer experiments are presented as the fraction of DiI-labelled B cells found in the PaLN divided by the fraction of DiI-labelled cells in the peritoneum. This formula was employed to minimise any errors arising from potential variations in cell numbers injected into recipients.
The peritoneal B cell pool develops poorly in NOD mice
In vivo migration of NOD peritoneal cells
Defective integrin expression and S1P responsiveness in NOD peritoneal B cells
Moreover, NOD B1 cells underexpressed integrin β1 (Fig. 3c). Given that peritoneal B cells did not express integrin α4β7 in any of the mouse strains (not shown), we conclude that the heterodimer integrin α4β1 is dysregulated on peritoneal B1 cells of NOD mice. Peritoneal B2 cells were also investigated, but did not express integrin α4 in any of the three mouse strains.
S1P has received considerable attention in recent years as a regulator of lymphocyte trafficking . For peritoneal B lymphocytes, S1P signalling is required for retention of B-lymphocytes within the peritoneum . In agreement with a previous study , a moderate but significant migratory response of BALB/c peritoneal B cells towards S1P was observed (Fig. 3d). The migration rate peaked at 10 nmol/l S1P and returned to control levels when the S1P concentration reached 100 nmol/l, demonstrating a bell-shaped dose–response curve that is typical of chemokine receptor-mediated cell migration. Moreover, pre-incubation with the S1P receptor 1 and S1P receptor 3 inhibitor, VPC23019, completely abolished the migratory response of BALB/c peritoneal B cells towards 10 nmol/l S1P (Fig. 3e). However, NOD peritoneal B cells did not significantly respond to S1P in in vitro migration assays at any of the S1P concentrations (1, 10, 300, 500 or 1,000 nmol/l) (Fig. 3d).
S1P receptors and CD69 expression in NOD peritoneal B cells
Expression of co-stimulatory molecules on NOD peritoneal B cells
Peritoneal B cells from 5- to 10-week old NOD mice expressed higher levels of several activation markers compared with BALB/c and C57BL/6 mice of the same age (Fig. 5a–h). Levels of the co-stimulatory receptor CD40 was significantly higher on B1 as well as B2 cells from NOD than on those from BALB/c and C57BL/6 mice. CD86, which could only be detected on B1 cells, was also overexpressed in the NOD mouse.
The majority (56%) of NOD B2 cells expressed CD69, as compared with 24% of C57BL/6 B2 cells and only 8% of B2 cells from BALB/c mice. A smaller, yet significant increase in CD69 expression was also observed on NOD B1 cells. Less than 30% of NOD B1 cells expressed l-selectin, as compared with over 75% of BALB/c and C57BL/6 B1 cells. There was no significant difference in l-selectin expression between B2 cells from NOD, BALB/c and C57BL/6 mice.
Expression of CD40, CD86, CD69 and l-selectin was additionally analysed in 3-week old mice. CD40 and CD86 were not yet upregulated in NOD mice at this time point (result not shown), but CD69 was upregulated and l-selectin downregulated. Moreover, the increased CD69 expression and decreased l-selectin expression was still evident in NOD mice at >17 weeks of age (Fig. 5i-l).
NOD peritoneal B cells respond strongly to CD40 ligation
Antigen-presenting capacity of peritoneal B cells
The reduced number of B cells in the peritoneum of NOD mice appears to correlate with a strongly accelerated migration rate of these cells specifically to the PaLN. The results presented here show a tenfold increase in the migration rate of peritoneal B cells to PaLN in NOD mice, compared with BALB/c mice.
The retention of B1 cells in the peritoneal cavity is in part dependent on the expression of integrin α4β1. This is demonstrated by the rapid clearance of peritoneal B1 cells in reponse to lipopolysaccharide-stimulated downregulation of integrin α4β1 . Integrin α4β1 was found to be markedly downregulated on NOD peritoneal B1 cells from a very early age. Low expression of integrin α4β1 on NOD B1 cells, in addition to being indicative of cell activation, undoubtedly contributes to the enhanced egress of B1 cells from the peritoneum in NOD mice.
The function of S1P in the peritoneal cavity is to retain B cells in the peritoneum. . Insensitivity to S1P would thus allow increased migration of these cells out of the cavity. In contrast to BALB/c peritoneal B cells, NOD peritoneal B cells showed no in vitro migratory response to S1P. The reason for this was presumed to be either a lack of S1P receptors 1 and/ or 3, which are known to communicate the migratory response [13, 19], or alternatively a functional defect in signalling through these receptors. NOD peritoneal B cells were not in fact deficient in S1P receptor 1 or S1P receptor 3 proteins. However, when the surface expression of S1P receptor 1 was examined, it was found to be lower on NOD than on BALB/c peritoneal B cells. This could be explained by a cross-communication between CD69 and S1P receptor 1 . High levels of CD69 downregulate S1P receptor 1 surface expression specifically, without substantially affecting S1P receptor mRNA expression. Moreover, high levels of CD69 functionally inhibit S1P receptor 1, causing cells to lose their chemotactic response to S1P . The cross-communication between S1P receptor 1 and CD69 appears to be reciprocal, since stimulation with S1P downregulates CD69 on NOD T cells . The same effect was observed on NOD peritoneal B cells. It is therefore suggested that NOD peritoneal B cell migration towards S1P is probably impeded by cross-communication between CD69 and S1P receptor 1 on the B cell surface, rather than by a lack of S1P receptor 1 proteins.
High surface expression of the co-stimulatory molecules CD40, CD86 and CD69, low l-selectin expression and a strong response to CD40 ligation demonstrate that NOD peritoneal B cells are highly activated. Moreover, co-stimulatory molecules are important factors in the pathology of diabetes. NOD T cells expressing high levels of CD40 are capable of inducing clinical diabetes , while CD86 upregulation on islet infiltrating NOD B cells is associated with development of insulitis . CD69 is an early activation marker, but its sustained upregulation has been associated with a number of inflammatory and autoimmune diseases . The anomalous levels of CD69 and l-selectin were observed on NOD peritoneal B cells from 3 weeks of age. This indicates that the increase in NOD peritoneal B cell activation status is initiated prior to insulitis onset.
The discovery that NOD peritoneal B2 and B1 cells express higher levels of various activation markers was unexpected, since the importance of peritoneal B cells in type 1 diabetes was thought to depend mostly on the B1 cell subset . However, peritoneal B2 cells appear to constitute a unique B cell subset, which exhibits characteristics that are intermediate between peritoneal B1 cells and splenic B2 cells. Peritoneal B2 cells additionally acquire certain characteristics of B1 B cells upon residing in the peritoneal cavity . CD11b, moreover, has been suggested to distinguish different developmental stages of peritoneal B cells  and its expression depends upon the location within the peritoneal cavity .
Previous studies have highlighted the importance of B cells as antigen-presenting cells in the pathology of diabetes [5, 24, 28, 29]. Our work demonstrates that peritoneal B cells are indeed capable of triggering a T cell response. In fact, peritoneal B cells seem to activate T cells more efficiently than splenic B cells. No distinction in antigen-presenting capacity between peritoneal and splenic B cells was observed in BALB/c B cells, underlining the exceptional characteristics of the NOD peritoneal B cell subset in this respect.
Peritoneal B cells reportedly interrelate with the immune system of the intestine and react to events taking place there . In the study by Ha and co-workers , indomethacin was used to disrupt the gut epithelial barrier, resulting in downregulation of integrin α4 and an efflux of peritoneal B cells. Ha and co-workers  further reported that TLR stimulation causes upregulation of CD69 and decreased peritoneal B cell sensitivity to S1P chemotactic signals. All of these activation-related events coincide with the findings presented here regarding NOD peritoneal B cells. It is thus interesting to speculate whether the abnormalities observed in NOD peritoneal B cells are a consequence of disrupted gut epithelial barrier properties and/or TLR-stimulated events.
This study demonstrates that NOD peritoneal B cells are activated and traffic vigorously to the PaLN. Peritoneal B cells, moreover, efficiently present insulin autoantigen to self-reactive T cells, triggering T cell proliferation in vitro. These characteristics underline the role of peritoneal B cells in inducing or accentuating islet-specific autoimmune responses in NOD mice.
This work was supported by The Academy of Finland, The Päivikki and Sakari Sohlberg Foundation, Finland, The Finnish Diabetes Research Foundation, The Danisco Research Foundation and Turku University Foundation, Finland. We thank M. Vainio (Turku University of Applied Sciences) and S. Lindqvist (Central Animal Laboratory of Turku University) for skilful technical assistance.
Duality of interest
The authors declare that there is no duality of interest associated with this manuscript.