CD22 expression mediates the regulatory functions of peritoneal B-1a cells during the remission phase of contact hypersensitivity reactions

Abstract

While contact hypersensitivity (CHS) has been considered a prototype of T cell-mediated immune reactions, recently a significant contribution of regulatory B cell subsets in the suppression of CHS has been demonstrated. CD22, one of the Siglecs, is a B cell-specific molecule that negatively regulates B cell receptor signaling. To clarify the roles of B cells in CHS, CHS in CD22^-/- mice was investigated. CD22^-/- mice showed delayed recovery from CHS reactions compared with wild type mice. Transfer of wild type peritoneal B-1a cells reversed the prolonged CHS reaction seen in CD22^-/- mice, and this was blocked by the simultaneous injection with IL-10 receptor Ab. While CD22^-/- peritoneal B-1a cells were capable of producing IL-10 at wild type levels, intraperitoneal injection of differentially labeled wild type/CD22^-/- B cells demonstrated that a smaller number of CD22^-/- B cells resided in lymphoid organs 5 days after CHS elicitation, suggesting a defect in survival or retention in activated CD22^-/- peritoneal B-1 cells. Thus, our current study reveals a regulatory role for peritoneal B-1a cells in CHS. Two distinct regulatory B cell subsets cooperatively inhibit CHS responses. While splenic CD1d^hiCD5⁺ B cells have a crucial role in suppressing the acute exacerbating phase of CHS, peritoneal B-1a cells are likely to suppress the late remission phase as “regulatory B cells”. CD22 deficiency results in disturbed CHS remission by impaired retention or survival of peritoneal B-1a cells that migrate into lymphoid organs.

Introduction

While the prototypic function of B cells is secreting Abs, or mounting humoral immune responses, recent studies have revealed that a regulatory subset of B cells plays a significant suppressive role in various immune reactions and diseases (1-4), such as experimental autoimmune encephalomyelitis (EAE) (5, 6), inflammatory bowel diseases (7, 8), and collagen-induced arthritis (9-11). In addition to these complex disorders, we have recently demonstrated that regulatory B cells have an important role in suppression of contact hypersensitivity (CHS) (12, 13), a representative model of delayed type hypersensitivity that is mediated mainly by antigen-specific effector T cells.

CHS is a cutaneous immune reaction which develops in two distinct phases: sensitization and elicitation (14, 15). In mice, primary skin painting with reactive hapten induces the CHS sensitization phase, in which effector T cells are sensitized by antigen-presenting cells. Subsequently, the elicitation phase is induced by re-exposure to the same hapten. Small numbers of sensitized antigen-specific T cells migrate from the circulation into the extravascular space at the skin challenge site and then interact again with antigen/peptide-major histocompatibility complexes on antigen-presenting cells. Activated T cells release proinflammatory cytokines, which then activate local tissue cells, leading to the characteristic late effector responses at 24-48 hours (16-18). In the elicitation phase, the main effector cells have been demonstrated to be IFN-γ-producing CD8⁺ Tc1 cells (19-21). Thus, CHS is a prototypic T cell-mediated response.

The existence of “regulatory” B cells was originally suggested in delayed hypersensitivity reactions (22). Recently, we reported that CD19-deficient (CD19^-/-) mice mount augmented CHS responses, and that marginal zone (MZ) B cells, which are lacking in CD19^-/- mice, have a regulatory role in CHS (12). Subsequently, splenic IL-10-producing CD1d^hiCD5⁺ B cells were proven to normalize this augmented CHS reaction in CD19^-/- mice or in wild type mice depleted of CD20⁺ B cells (13). On the other hand, in a collagen-induced arthritis model, transitional 2-MZ precursor cells were suggested to serve as regulatory B cells (9). While it remains unclear whether these two populations are the same, IL-10 production appears a hallmark of these regulatory B cells. In addition to splenic regulatory B cells, studies have demonstrated that peritoneal B-1 cells are also an abundant source of IL-10 (23). Nonetheless, whether peritoneal B-1 cells play a regulatory role in inflammatory diseases remains to be determined.

B-cell development, activation, and survival are elaborately regulated by the BCR and functionally interrelated cell-surface receptors (24). CD22 is a B cell-specific transmembrane molecule which is a member of “sialic acid-binding immunoglobulin-like lectin (Siglec)” family of adhesion molecules (25). CD22 has ITIMs in its cytoplasmic domain, and becomes phosphorylated in response to BCR ligation and other stimuli (26, 27). CD22 serves as an inhibitory coreceptor, and modulates the BCR signal in response to cues from the local microenvironment (28). B cells from CD22-deficient (CD22^-/-) mice exhibit exaggerated Ca²⁺ mobilization in response to BCR stimulation in vitro (29-32), and display hyperimmune responses. CD22^-/- B cells express less surface IgM than those from wild type mice (33), suggesting that they have undergone chronic activation in vivo. In response to BCR stimulation, CD22^-/- B cells predominantly undergo apoptosis (29, 34), which can be rescued by CD40 coligation. Thus, CD22 has significant and complex functions that regulate B cell activation and survival.

In the current study, we examined the role of CD22 expression in CHS. CD22 deficiency resulted in prolonged CHS reactions, suggesting an inhibitory role of CD22 expression especially during the late phase in elicitation of CHS. The current study also demonstrates that peritoneal B-1a cells serve as regulatory B cells via IL-10 production.

Methods

Mice

C57BL/6 and BALB/c wild type mice were purchased from Clea Japan Inc. (Shizuoka, Japan). CD22^-/- (C57BL/6 × 129) mice were generated as described (32) and backcrossed onto a C57BL/6 strain 12 times. IL-10-deficient (IL-10^-/-) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). All mice used were 8 to 12 weeks of age and were housed in a specific pathogen-free barrier facility. All studies and procedures were approved by the Animal Committees of International Medical Center of Japan and Kanazawa University Graduate School of Medical Science.

Sensitization and elicitation of CHS

Mice were sensitized with 25 μl of 0.5% 2,4-dinitrofluorobenzene (DNFB; Sigma-Aldrich, St. Louis, MO), in acetone and olive oil (4:1), on shaved abdominal skin for 2 consecutive days. Five days later, CHS was elicited by applying 20 μl of 0.25% DNFB on the right ear. As a negative control, the left ear was treated with acetone/olive oil alone. In some experiments, mice were treated with anti-IL-10 receptor mAb (250 μg, 1B1.3a; BD PharMingen, San Diego, CA) or control IgG 2 days after DNFB challenge. Ear thickness was measured in a blinded manner with a micrometer (Mitutoyo, Kawasaki, Japan) before and after the challenge until day 10. The degree of CHS was expressed as a swelling of the hapten-challenged ear compared with that of the vehicle-treated ear in units of mm ×10^-2 (mean ± SEM). As for sensitization with FITC, 400 μl of 0.5% FITC (Sigma-Aldrich) in acetone and dibutylphthalate (1:1) was applied on shaved abdominal skin of each mouse, followed by loading 20 μl of 0.5% FITC on the right ear and acetone/ dibutylphthalate alone on the left ear (control) 5 days later. Ear thickness was evaluated as for DNFB-elicited mice.

Histology

Ear samples were taken before and 24 and 120 hours after DNFB challenge and were fixed in 4% formalin for routine histology with hematoxylin and eosin staining.

Immunofluorescence analysis

To examine splenic CD1d^hiCD5⁺ B cells and nodal regulatory T cells in CD22^-/- and wild type mice, spleens and inguinal lymph nodes (LNs) were harvested from these mice before and 5 days after the sensitization with DNFB. Spleens and LNs from non-treated mice were used as controls. Splenic cells were stained for three-color immunofluorescence analysis using FITC-conjugated anti-CD1d (1B1, BD PharMingen), PE-conjugated anti-CD5 (53-7.3, BD PharMingen), and PE-Cy5-conjugated anti-B220 Abs (RA3-6B2; BD PharMingen) for 20 minutes at 4°C. Nodal cells were stained with FITC-conjugated anti-CD4 (GK1.5, BD PharMingen) and PE-Cy5-conjugated anti-Thy-1.2 (30-H12; BD PharMingen), and PE-conjugated anti-Foxp3 Ab (MF-14, Biolegend, San Diego, CA) using a Cytofix/Cytoperm kit (BD PharMingen). Peritoneal B cells were stained using PE-conjugated anti-CXCR4 (2B11, eBioscience; San Diego, CA), PE-conjugated anti-CXCR5 (2G8, BD PharMingen), PE-conjugated anti-CCR6 (140706, R&D Systems; Minneapolis, MN), PE-conjugated anti-CCR7 (4B12, eBioscience), and PE-conjugated anti-CCR9 (eBioCW-1.2 eBioscience) after stimulation with LPS (10 μg/ml; Sigma-Aldrich) for 8 hr. Labeled cells were analyzed on an Epics XL flow cytometer (Beckman Coulter, Miami, FL) with fluorescence intensity shown on a 4-decade log scale. Positive and negative populations of cells were determined using unreactive isotype-matched Abs (Southern Biotechnology, Birmingham, AL) as controls for background staining.

To examine peritoneal B-1a/B-1b cell populations, 5 ml of PBS was injected into the abdominal cavity of sensitized mice 5 days after sensitization and then recollected. Single cell suspensions were stained using biotin-conjugated anti-CD11b. After washing, cells were stained by streptavidin-PE-Cy5 (BD Pharmingen), FITC-conjugated anti-CD5 Ab, and PE-conjugated anti-B220 Ab, followed by flow cytometric analysis. To evaluate IL-10-producing B cells, cells were resuspended with LPS (10 μg/ml; Sigma-Aldrich), PMA (50 ng/ml, Sigma-Aldrich), ionomycin (500 ng/ml; Sigma-Aldrich), and monensin (2 μM; eBioscience) for 5 hr (13). For IL-10 detection, Fc receptors were blocked with mouse Fc receptor-specific mAb (2.4G2; BD PharMingen) before cell-surface staining, and then cells were fixed, permeabilized, and stained with PE-conjugated IL-10 mAb (JES5-16E3; BD PharMingen).

Measurement of IL-10 concentrations

Secreted IL-10 was quantified using mouse IL-10 ELISA kit (Invitrogen, Carlsbad, CA). Splenic B cells were purified with B220 mAb-coupled microbeads (Miltenyi Biotech, Auburn, CA) from wild type mice and CD22^-/- mice. CD1d^hiCD5⁺ B cells were collected using a BD FACS Aria cell-sorting system (Becton Dickinson). Isolated CD1d^highCD5⁺ B cells as well as CD5^- B cells (3×10⁵) were cultured in 200 μl of RPMI 1640 culture medium containing 5% bovine serum albumin and 10 mM HEPES in a 96-well flat-bottom tissue culture plates at 37°C with 5% CO₂ in the presence of 10 μg/ml of LPS (Escherichia coli serotype 0111: B4, Sigma-Aldrich, St. Louis, MO). After incubating for 72 hr, IL-10 concentrations of cultured supernatant were quantified. All assays were carried out in triplicate.

Adoptive cell transfer

In adoptive cell transfers from sensitized donor mice to naive recipient mice, recipient mice were treated with DNFB as described above. Inguinal LNs, spleen or peritoneal lavage were harvested 5 days later. T cells (2 ×10⁶ cells) and B cells (2 ×10⁶ cells) were isolated using anti-Thy1.2- or anti-B220-coupled microbeads (Miltenyi Biotec), respectively, for positive selection via autoMACS (Miltenyi Biotec). These cells were adoptively transferred intravenously.

In adoptive cell transfers from naive donors to sensitized recipient mice, donor mice were sensitized by DNFB. T cells and B cells were isolated as above. To collect B-1a and B-1b B cells, peritoneal cells were stained with FITC-conjugated anti-CD5, PE-conjugated anti- B220, and PE-Cy5-conjugated anti-CD11b Abs, and then isolated using a BD FACS Aria cell-sorting system. These cells were also transferred intravenously to recipient mice. These cells were also transferred intravenously to recipient mice. In adoptive cell transfers from sensitized donors to sensitized recipient mice, donor mice and recipient mice were sensitized as above and 5 days later cells were isolated and transferred as described above.

Ear thickness of the recipients, at least five mice per group, was measured with a micrometer before the transfers. Twenty-four hours later, the recipients were challenged on the right ear with 20 μl of 0.25% DNFB as above and on the left ear with acetone/olive oil alone. The subsequent increases in ear thickness were determined at 24, 120, and 240 hours after challenge. The thickness of the control ears was subtracted from experimental responses to yield net ear swelling.

Quantitative reverse transcription polymerase chain reaction (Q-RT-PCR)

Peritoneal CD5⁺B220^dull/CD5^-B220⁺ cells taken before sensitization and two days after DNFB challenge as mentioned-above were homogenized in Isogen S (Wako, Toyo, Japan) and total RNA was isolated following the manufacturer's instructions. RNA concentration was determined using nanodrop (NanoDrop Technologies; Wilmington, DE) by A260 value of the samples. Total RNA was reverse transcribed to cDNA using the Reverse Transcription System with random hexamers (Promega, Madison, WI). Q-RT-PCR was performed using the TaqMan® system (Applied Biosystems, Foster City, CA) on an ABI Prism 7000 Sequence Detector (Applied Biosystems) according to the manufacturer's instructions. TaqMan® probes and primers for IL-10, TGF-β, and GAPDH were purchased from Applied Biosystems.

B cell migration assays

After wild type recipient mice were sensitized with DNFB, peritoneal cells were obtained from wild type and CD22^-/- donor mice, and peritoneal CD5⁺ B cells were isolated as above. B cells derived from wild type mice were labeled by calcein-AM (Invitrogen), and equivalent amounts of B cells from CD22^-/- mice were labeled by PKH-26 (Sigma-Aldrich). The ratios of calcein-AM to PKH-26 before intraperitoneal injection were examined by flow cytometry. Subsequently, both labeled cell populations were injected intraperitoneally into wild type recipient mice. One day after injection, recipient mice were elicited as above. Spleen and cervical LNs were obtained from recipient mice at day 1 and day 5 post elicitation and the ratios of calcein/PKH-26 labeled cells in these were examined by flow cytometry.

Statistic analysis

All data are shown as mean values ± SEM. Student's t-test was used for determining the level of significance of differences between two groups, and Steel-Dwass multiple comparison test was used among three or more groups. A P value less than 0.05 was considered statistically significant.

Results

Recovery from CHS is delayed in CD22^-/- mice

To assess whether CD22 expression plays a role in CHS, wild type and CD22^-/- mice were challenged with DNFB after sensitization, and ear swelling was measured before and after challenge. In both mice, ear thickness was increased equivalently on of elicitation day 1 post elicitation (Fig. 1A). In wild type mice, ear swelling responses reached a peak on day 3 and returned to normal by approximately on day 7. In By contrast, ear swelling was prolonged in CD22^-/- mice, resulting in a significant increase of ear swelling from day 5 to 10 compared with wild type mice. Thus, CD22^-/- mice exhibited prolonged CHS reactions to DNFB. This delayed recovery was not limited to CHS against DNFB, because CD22^-/- mice exhibited similarly prolonged responses to FITC application (Fig. 1B). Thus, CD22^-/- mice experience a are more persistent to CHS reaction.

Figure 1. CHS in CD22^-/- mice.

Mice were sensitized with 0.5 % DNFB (A) or 0.5 % FITC (B) for 2 consecutive days. Five days later, CHS was elicited by 0.25% DNFB (A) or 0.5 % FITC (B) on the right ear. Ear thickness was measured from one day prior to the day of challenge (day 0) and for before and after the challenge until day 10 days post challenge. [Note to authors: the way in which I am understanding what's written is that the day before challenge should be referred to as day -1, day of challenge as day 0, and first day post challenge as day 1. Would this be correct, or is there another way to clarify this on the graph?]. The degree of CHS was expressed determined as the degree of swelling of the hapten-challenged ear when compared with that of the vehicle-treated ear and was expressed in units of mm ×10^-2 (mean±SEM). Significant differences between the mean CHS responses between CD22^-/- mice and wild type mice are indicated. *, P<0.05; **, P<0.01. (C) Histopathology of CHS-elicited ear pinnae in wild type and CD22^-/- mice. Ear samples were taken before and at day 2 and day 10 after DNFB challenge and fixed in 4% formalin for routine histology with hematoxylin and eosin staining.

The degree of CHS reaction was also assessed histopathologically (Fig. 1C). Cellular infiltrates and edema were observed in the DNFB-painted ears of both CD22^-/-and wild type mice 2 days after challenge. In both groups of mice, Tthe majority of the infiltrating cells were polymorphonuclear cells and lymphoid cells in both mice. As for the ear specimens obtained By 10 days after challenge, cellular infiltration and edema had almost disappeared in the ears of wild type mice, whereas the ears of CD22^-/- mice still showed marked remained cellular infiltration and edema. Therefore, CD22^-/- mice exhibit a delayed recovery in CHS, both clinically and pathologically, compared to that ofwith wild type mice.

Splenic CD1d^hiCD5⁺ B cells are normally present in CD22^-/- mice have CD5⁺CD1d^hi regulatory B cells normally

Since CD22 expression is restricted to B cells, a delayed recovery of CHS in CD22^-/-mice suggests that regulatory mechanisms by B cells are impaired in CD22^-/- mice. IL-10-producing CD1d^hiCD5⁺ B cells in the spleen, termed B10 cells, have been demonstrated to play a suppressive role in CHS (13). IL-10-producing CD1d^hi[Note to authors: include a reference here]. CD5⁺ B cells have a CD21^hiCD23^lo phenotype, which resembles MZ B cells. Since CD22^-/- mice lack MZ B cells (29-32), we next examined splenic CD1d^hiCD5⁺ B cells in CD22^-/- mice to assess the possibility that the delayed recovery of CHS in CD22^-/- mice resulted from the loss of CD1d^hiCD5⁺ B cells. We examined the rates percentages of these cells in CD22^-/- and wild type mice before and after sensitization (Fig. 2A). The percentages of splenic CD1d^hiCD5⁺ B cells in total splenic B cells from CD22^-/- and wild type mice were 1.9±0.3% and 2.0±0.3% before sensitization, and 3.0±0.3% and 3.1±0.4% after sensitization [Note to authors: These numbers do not match those in Fig. 3: 1.8, 2.1, 3.0, 3.3. I assume these are the averages based on the fact that the figure is a representative, but you may want to state that here as well to avoid confusion; ie “the percentages of splenic … from 6 CD22-/- and 6 wild-type mice were…”]. Thus, although the phenotype of splenic CD1d^hiCD5⁺ B cells resembles that of MZ B cells, the CD1d^hiCD5⁺ B population is distinct from MZ B cells, and normally exists isand is present in untreated CD22^-/- mice.

Figure 2. Splenic CD1d^hiCD5⁺cells in wild type and CD22^/mice before and after sensitization.

(A) Spleens were harvested before and 5 days after the sensitization with DNFB. Splenicen B cells were stained for two-color immunofluorescence analysis using FITC-conjugated anti-CD1d and PE-conjugated anti-CD5. Labeled cells were analyzed on a flow cytometer with fluorescence intensity shown on a 4-decade log scale. Positive and negative populations of cells were determined using unreactive isotype-matched Abs as controls for background staining. RAll results shown are representative of mice those from from six or seven more different 2-month-old mice. Numbers represent the percentage of CD1d^hi CD5⁺ cells. (B) IL-10 secretion from splenic B cells. B220⁺CD1d^hiCD5⁺ B cells and B cells other than CD1d^hiCD5⁺ population (3×10⁵ for each) were obtained from wild type mice and CD22^-/- mice, and cultured with LPS for 72 hr. Supernatant IL-10 concentrations were measured by ELISA. Each group contains three mice. **, P<0.01. (C) IL-10 producing population in B220⁺CD1d^hiCD5⁺ B cells from CD22^-/- mice and wild type mice. Representative data of histograms demonstrate cytoplasmic IL-10 expression before and after LPS, PMA, and ionomycin stimulation. Percentages indicate IL-10⁺ cell freqiencies. Dashed lines represent negative controls stained with isotype-matched control Ab.

Figure 3. Adoptive transfer of LN cells in CHS reaction.

Whole cells (A), T cells (B) and B cells (C) were prepared from inguinal LNs of sensitized wild type or CD22^/ mice and transferred to unsensitized mice as indicated. CHS was elicited 1 day after transfer using DNFB. Ear swelling values indicate the difference between the ear thickness before and after each time point. Each experiment included three mice and was repeated at least three times. *, P<0.05, **, P<0.01.

Next, IL-10 secretion from splenic CD1d^hiCD5⁺ B cells was investigated. IL-10 secretions from CD1d^hiCD5⁺ B cells were significantly higher than those from non-CD1d^hiCD5⁺ B cells in both wild type mice and CD22^-/- mice (P<0.01 for each; Fig. 2B). However, the difference between CD1d^hiCD5⁺ B cells from wild type mice and those from CD22^-/- mice was not significant, although IL-10 secretion from CD22^-/- B cells was slightly increased (Fig. 2B). Therefore, CD1d^hiCD5⁺ B cells from CD22^-/- mice have a capability of secreting IL-10 similarly to those from wild type mice. Furthermore, the numbers of IL-10-producing CD1d^hiCD5⁺ B cells were significantly higher in CD22^-/- mice than wild type mice (P<0.05; Fig. 2C).

The rates percentages of CD4⁺Foxp3⁺ regulatory T cells in inguinal LNs were also examined. They represented were 2.8±0.4% of live cells present before sensitization, and 3.2±0.4% after sensitization in CD22^-/- mice. There was no significant difference from theose percentages present in wild type mice (2.7±0.5% before sensitization and 3.5±0.5% after sensitization; data not shown).

CD22 expression in recipient mice influenced the duration of CHS reactions

To examine the cause of delayed recovery from CHS in CD22^-/- mice, bulk inguinal LN cells from wild type mice were adoptively transferred intravenously to unsensitized CD22^-/- mice or unsensitized wild type mice 5 days after sensitization. Then one day after cellcell transfer, 20 μl of 0.25 % DNFB was loaded onto the ears of recipient CD22^-/- mice orf wild type mice. As shown in Fig. 3A, when they were injected with whole cells, ear swelling on day 1 and day 2 was comparable between recipient CD22^-/-mice and wild type mice. [Note to authors: There are several points to address in this section. First, Figure 4 does not show data for day 1, only days 2, 7 and 10. Please change your text accordingly. Also, although this is time-consuming, we recommend changing the graphs in Figure 4 to show in descending order: wild type donor into wild type recipient; wild type donor into CD22^-/- recipient; CD22^-/- donor into wild type recipient; and CD22^-/- donor into CD22^-/- recipient. This would make the graphical presentation of the data follow the order in which it is described in the text, and would thus make it easier for reviewers to follow as they read your manuscript. Next, please specify in the text whether you are referring to the whole cell transfers in this portion of the text, as the whole cell transfer data and T cell transfer data look much the same. Lastly, for the whole cell transfer data, it appears as if there is slightly greater ear swelling in the CD22^-/- vs the wildtype recipients at day 2. ]. By contrast, ear swelling remained significantly augmented in CD22^-/- mice transferred with whole cells on days 7 (p<0.05) and day day 10 (p<0.001) compared with wild type recipients., Likewise, when inguinal LN cells of from sensitized CD22^-/- mice were adoptively transferred to naive CD22^-/- mice or naive wild type mice, there was no significant difference in ear swelling between recipient CD22^-/- mice or naive wild type mice on day 1 and day 2, whereas recipient CD22^-/- mice showed a more significantly increased response than recipient wild type mice on day 7 (p<0.05) and day 10 (p<0.001). When the donor cells were from CD22^-/- mice, the difference of the magnitude of the responses was not significantobserved compared with wild type donor cells. [Note to authors: Use of the word “significant” to describe data implies that the differences are statistically significant. If this is the case, include a p value in parentheses after your sentences. If the differences are not statistically significant, you must use another word such as “recipient CD22^-/- mice showed substantially increased”] Therefore, although CHS is fully transferable to naive mice with sensitized LN cells from the donor regardless of CD22 deficiency in the donor mice, CD22 deficiency in the recipient mice prolongs the duration of the responses once CHS is induced. Next, T cells or B cells were purified from inguinal LNs from sensitized mice and were transferred to wild type or CD22^-/- mice. T cells were able to induce initial CHS reactions at the a magnitude similar with to those observed in whole cell transfers (Fig. 3, A and B). Between wild type recipients and CD22^-/- recipients transferred with wild type T cells (Fig. 3B, the middle box), there is more significant difference on day 7 (p<0.001) than whole cell transfers. CHS did not occur when B cells alone were transferred to naive mice (Fig. 3C). Collectively, these results suggest that CD22 expression in recipient mice is required for the optimal late-phase suppression of the CHS reactions.[Note to authors: the X axis on your T cell graphs (B) go to day 14 due to the error bars on the graph. It would be conventional to make the A and C graphs match by drawing them out to day 14 as well to allow for easier comparison].

Figure 4. Adoptive transfer of sera and B-cell subsets in CHS reaction.

(A) Sera, splenic B or/ T cells, peritoneal B or /T cells, and LN B or /T cells from unsensitized wild type mice were transferred to CD22^-/- mice. The bottom bars indicate ear swelling of non-transferred wild type mice. (B) Peritoneal CD5⁺/CD5^-B cells and peritoneal CD11b⁺/CD11b^- B cells from unsensitized wild type mice were transferred to CD22^-/- mice. CHS was elicited by DNFB 1 day after transfer. *, P<0.05, **, P<0.01.

Peritoneal B-1a cells of wild type mice suppress prolonged CHS reactions in recipient CD22^-/- mice

To further investigate the delayed recovery from CHS reaction in CD22^-/- mice, B cells or T cells from the spleen, inguinal LN, and peritoneal cavity were isolated from unsensitized wild type mice and transferred to CD22^-/- mice that were sensitized 5 days before (Fig. 4A). Serum transfer was also performed in order to examine the participation of Abs. Recipient CD22^-/- mice were elicited for CHS one day after transfer. In all cases, CHS was occurred equivalently to that in mice that were injected only with PBS. Ear swelling in CD22^-/- mice was not altered by the transfer of nodal or splenic B cells, all fractions of T cells, or sera from wild type mice. Notably, however, the CHS reaction in CD22^-/- mice was normally ceasedresolved in a manner comparable to that in wild type mice by peritoneal B-cell transfer from unsensitized wild type mice (Fig. 4A). Furthermore, this result was equally also observed when B cells and T cells were isolated negatively or positively, and when a sensitized wild type donor was used (data not shown).

The transfer of splenic, nodal, and peritoneal B and T cells as well as sera from unsensitized CD22^-/- mice to wild type mice that were sensitized 5 days before was also examined (Supplemental Fig. 1). Recipient wild type mice were elicited for CHS one day after transfer. In all cases, CHS was occurred and recovered resolved equivalently to mice that were injected only with PBS. This result was equally observed equally when sensitized wild type donor cells was were used. Consequently, these data suggest that peritoneal B cells from wild type mice, whether sensitized or not, were suggested to be capable of amending the prolonged CHS reactions in CD22^-/- mice. Also, peritoneal B cells from CD22^-/- mice did not have the ability to exacerbate a capability of exacerbating CHS reactions in wild type mice.

To investigate further the inhibitory functions feature of peritoneal B cells, peritoneal B cells of unsensitized wild wild type mice were separated into B220⁺CD5⁺CD11b⁺ cells (B-1a cells), B220⁺CD5^-CD11b⁺ cells (B-1b cells), and B220⁺CD5^-CD11b^- cells (B-2 cells). Each fraction was injected into CD22^-/- mice 5 days after sensitization (Fig. 4B). While peritoneal B-1b or B-2 cells from wild type mice did not affect ear swelling in CD22^-/- mice, the CHS reaction normally ceased when peritoneal B-1a B cells from wild type mice were transferred to CD22^-/- mice. This result was also equally observed using wild type donors that were sensitized 5 days before. Thus, while the possibility that Ab engagement during the B220⁺CD5⁺CD11b⁺ cell isolation may influence the results cannot be completely excluded, these data suggest that peritoneal B220⁺CD5⁺CD11b⁺ B cells viz. peritoneal B-1a cells were suggested to have the ability to capability of suppressing prolonged CHS reactions in CD22^-/- mice.

CD22^-/- mice have increased peritoneal B-1a cells and B-1b cells

It has been already demonstrated that CD22^-/- mice have increased numbers of peritoneal B-1 cells (29-32); however, while detailed characterization in a C57BL/6 background has not been reported. As is shown in Fig. 5A, the percentages ratios of B220⁺CD11b⁺ B cells (whole B-1 cells) in total peritoneal lymphocytes were 10.0±3.8% (1.6±0.6 ×10⁶ cells) in wild type mice and 28.8±5.1% (4.9±0.9 ×10⁶ cells) in CD22^-/- mice[Note to authors: these numbers are not the same as in Fig. 7: 10.2%, 26.8%-again, I assume they are averages based on the fact that the figure is a representative, but you may want to state that here as well to avoid confusion]. Thus, peritoneal B-1 cells were significantly increased in CD22^-/- mice (p<0.05). B220⁺CD11b⁺CD5⁺ B-1a cells were increased in CD22^-/- mice (2.9±0.5 ×10⁶; 17.2±3.8% of peritoneal B cells) compared with wild type mice (0.7±0.2 ×10⁶; 4.5±1.6%; p<0.05). were Peritoneal B-1b cells were also increased in CD22^-/- mice, as the percentages ratios of B220⁺CD11b⁺CD5^- B-1b cells were 5.6±1.9% (0.9±0.4 ×10⁶) in wild type mice and 11.5±2.3% (1.9±0.4 ×10⁶) in CD22^-/- mice (p<0.05). Thus, peritoneal B-1a cells and B-1b cells are both increased in CD22^-/- mice compared with wild type mice, although this did does not explain the mechanisms of how inby which CD22^-/- mice exhibited augmented CHS response.

Figure 5. IL-10 expression in peritoneal B-1a cells from wild type and CD22^/mice.

(A) Peritoneal B-1 cell profile of CD22^-/- mice. Single cell suspension of the peritoneal lavage from wild type and CD22^/ mice were stained using biotin-conjugated anti-CD11b. After washinged, the cells were stained by streptavidin-PE-Cy5, FITC-conjugated anti-CD5, and PE-conjugated anti-B220 Abs, followed by flow cytometricy analysis. These results are representative of those obtained with six 2-month-old mice.

(B) Peritoneal CD5⁺B220⁺ cells from wild type mice and CD22^-/-mice were stained for intracellular IL-10 before sensitization and 2 days after DNFB challenge. (C) Peritoneal CD5⁺B220⁺ cells before sensitization and 2 days after DNFB challenge were collected from wild type mice and CD22^-/-mice (five per group), and RNA was extracted. The mRNA levels of IL-10 were analyzed by quantitative RT-PCR and normalized with internal control GAPDH. Data are shown as mean ± SEM from five mice. (D) IL-10 secretion was determined by ELISA. Naive peritoneal CD5⁺B220⁺ cells from wild type mice and CD22^-/-mice were cultured in media alone or containing LPS for 24 hr. *, P<0.05, **, P<0.01.

Figure 7. Peritoneal B-1 cell migration.

Wild type recipient mice were sensitized with DNFB as above. Peritoneal cells were obtained from wild type and CD22^-/- donor mice, and peritoneal CD5⁺ B cells were isolated by autoMACS using anti-B220-coupled microbeads. The cells derived from wild type mice were labeled by calcein-AM, and an equivalent amount number of the cells from CD22^-/- mice were labeled by PKH-26. The ratios of calcein- to PKH-26 before intraperitoneal injection were examined by flow cytometry. Subsequently, both labeled cell populationss were injected intraperitoneally to wild wild type recipient mice. One day after injection, recipient mice were elicited as above. Spleen and cervical LNs were obtained from recipient mice one and five day after elicitation and the ratios of calcein to /PKH-26 labeled cells in these were examined by flow cytometry. We assessed B cell distribution by comparing the ratio of calcein- to PKH-26- labeled cells collected from cervical LNs and spleen at day 1 and day 5 after elicitation (Ro) with the ratio of calcein- to PKH-26-labeled cells before intraperitoneal injection (Ri). Cells with equivalent migratory properties distribute evenly and generate Ro/Ri ratios approaching 1.0. (A) Values represent the mean + SEM from three experiments. Representative flow cytometry data are shown in (B). *, P<0.05, **, P<0.01.

CD22 deficiency does not affect IL-10 production from peritoneal B cells after elicitation

Regulatory B cells have been demonstrated to suppress T cell-mediated inflammatory reactions through IL-10 secretion. While CD22^-/- mice had an increased number of B-1 cells, the their cytokine-producing abilities of these cells were unknown function may be disturbed. Therefore, we examined IL-10 production from by peritoneal B cells produce before sensitization and after elicitation. Peritoneal B cells were harvested from wild wild type and CD22^-/- mice before sensitization and 2 days after elicitation and analyzed by flow cytometry and Q-RT-PCR. Both in wild type mice and CD22^-/- mice, <0.5% of peritoneal B-1a cells were positive for intracellular IL-10 before sensitization and 2 days after elicitation (Fig. 5B). Fig. 5C shows IL-10 mRNA expression in peritoneal B cells. After elicitation, IL-10 mRNA expression was significantly increased in both types of mice compared with to that seen before sensitization (p<0.01). Between wild type mice and CD22^-/- mice, expression levels of IL-10 were equivalent before sensitization and after elicitation. Also, there was no significant difference in TGF-β mRNA expression between wild wild type and CD22^-/- peritoneal B cells before and after elicitation (data not shown). When IL-10 secretion was evaluated by ELISA, there was also no significant difference between wild type and CD22^-/- peritoneal B-1a cells (Fig. 5D). Collectively, IL-10 secretion from peritoneal B cells of CD22^-/- mice was increased after elicitation, similarly to that of wild type mice.

IL-10 contributes to CHS remission

While IL-10 secretion from splenic regulatory B cells has been demonstrated to play an essential role in their suppressive functions, this it may not be the case with peritoneal B cells. Therefore, whether IL-10 contributions contributes to the suppression of CHS by peritoneal B-1a cells were assessed with a function–blocking mAb against the IL-10 receptor. Co-iInjection of isotypic control mAb and wild type peritoneal B-1a cells into sensitized CD22^-/- recipient mice 2 days after elicitation resulted in suppression of the suppressing late phase CHS response to wild type levels, as was observed previously above (Fig. 6A). In By contrast, IL-10 receptor mAb injection during with transfer of wild type peritoneal B-1a cells into sensitized CD22^-/- mice significantly delayed CHS recovery (p<0.01 at day 10), which was similar to the course observed in CD22^-/- mice. IL-10 receptor mAb injection alone further augmented CHS response when compared with IL-10 receptor mAb injection with with wild type peritoneal B-1a cell transfer on days 2 and 7, although the difference was not significant (Fig. 6A).

Figure 6. The suppression of CHS by peritoneal B-1a cells depends upon IL-10.

(A) Peritoneal CD5⁺B cells (B-1a cells) from unsensitized wild type mice and/or control or IL-10 receptor-specific mAb were transferred to CD22^-/- mice 3 days after DNFB challenge. The bottom bars indicate ear swelling of non-transferred wild type mice. (B) Peritoneal B-1a cells from unsensitized wild type mice or IL-10^-/- mice were transferred to CD22^-/- mice 3 days after DNFB challenge. The bottom bars indicate ear swelling of non-transferred wild type mice. **, P<0.01.

That IL-10 secretion from peritoneal B cells is critical for the suppression of the late phase CHS response was confirmed using IL-10^-/- mice. When peritoneal CD5⁺ B cells from wild type or IL-10^-/- mice were injected into sensitized CD22^-/- recipient mice, IL-10^-/- B-1a cells failed to resolve the ear swelling on days 7 and 10 (Fig. 6B). Thus,it is demonstrated that disrupted the ability of regulatory function of peritoneal B-1a cells to bring about via IL-10 participates in prolonged CHS in CD22^-/- mice depends upon IL-10.

Fewer CD22^-/- peritoneal B-1a cells resided in lymphoid organs less than wild-type cells after elicitation

It has been suggested that CD22 binding to endogenous ligands is required for the migration of mature recirculating B cells to the bone marrow. As CD22^-/- mice had considerablye reduced numbers of recirculating B cells in the bone marrow, we assessed the migration of CD22^-/- B-1a cells to lymphoid organs using a two-color staining in vivo migration assay (35). Equivalent numbers of purified peritoneal CD5⁺ B cells from CD22^-/- mice labeled with the intravital fluorochrome PKH-26 and purified peritoneal CD5⁺ B cells those from wild type mice labeled with the intravital fluorochrome calcein, were subjected to intraperitoneal adoptive transfer into wild type recipient mice. One day after cell transfer, the right ears of recipient mice were elicited by DNFB. B cell distribution was assessed by comparing the ratio of calcein- to PKH-26-labeled cells collected from cervical LNs and spleens at day 1 and day 5 after elicitation (Ro) with the ratio of calcein- to PKH-26-labeled cells before intraperitoneal injection (Ri). Cells with equivalent migratory properties distribute evenly and generate Ro/Ri ratios approaching 1 (Fig. 7). On day 1, wild type and CD22^-/- B cells were distributed in spleen and cervical LNs at similar frequencies. However, Ro/Ri ratios were significantly increased to 1.58±0.48 in spleen (p<0.05) and 1.76±0.54 in LNs (p<0.01) on day 5. These data reflect that fewer less CD22^-/- peritoneal B cells, relative to wild type peritoneal B cells, existed in lymphoid organs 5 days after elicitation than wild-type peritoneal B cells. This was also similarly observed when these cells were injected into CD22^-/- mice (1.52x±0.55 in spleen, p<0.05; 1.83x±0.77 in LNs; p<0.05 on day 5). Therefore, a decreased number of CD22^-/- peritoneal B cells remained in lymphoid organs compared with wild type peritoneal B cells after elicitation. To examine the migratory functions of CD22^-/- peritoneal B cells, cell-surface densities of B cell-expressing chemokine receptors were examined in peritoneal CD5⁺ B cells from CD22^-/- and wild type mice. Constitutive expression levels of CXCR4, CXCR5, CCR6, CCR7, and CCR9 were almost identical between CD22^-/- and wild type B-1a cells (data not shown). After LPS stimulation, CXCR4 expression was slightly higher in CD22^-/-B-1a cells (Supplemental Fig. 2). By contrast, CCR6 expression was slightly lower in CD22^-/- B-1a cells. Nonetheless, these modest differences in expression levels of these chemokine receptors may not be likely to cause the difference in the migration.

Peritoneal B-1a cells from wild type mice, but not from CD22^-/- mice, improve recovery from CHS reactions in CD19^-/- mice

Recently, we have reported that CHS reactions are is augmented by a deficiency of CD19, a positive B cell response regulator, due to the absence of regulatory B cells in the spleen (12). When CHS reactions induced by DNFB in CD22^-/- mice were was compared with those that in CD19^-/- mice, CHS reactions in CD19^-/- mice were augmented during both early (day 1-day 3) and late (day 5-day 10) phases, while, in CD22^-/- mice, the acute phase of CHS reactions was similar to that in wild type mice and only recovery in the late phase (recovery phase) was delayed (Fig. 8A). Since we have demonstrated above that peritoneal B-1a cells have the potential to play a regulatory role in the recovery phase of CHS in CD22^-/- mice, we examined further the role of peritoneal B-1a cells in augmented CHS reactions of CD19^-/- mice by adoptive cell transfer. Unsensitized wild wild type or CD22^-/- peritoneal B-1a cells were purified and transferred to CD19^-/- mice that were sensitized 5 days before. Recipient CD19^-/-mice were elicited for CHS one day after transfer. As shown in Fig. 8B, in both cases, CHS in CD19^-/- recipient mice was more prominent than in control, unmanipulated non-transferred wild type mice, and showed significant delays in recovery. CD19^-/-recipient mice that were transferred withreceived wild type cells recovered from CHS equivalently to non-transferred wild type mice, while ear swelling remained on day 10 in CD19^-/- recipient mice that received were transferred CD22^-/- cells. Collectively, these results suggest that wild type peritoneal B-1a cells have an inhibitory role during the late phase of in CHS reactions in of CD19^-/- mice, but not during the early phase, and that CD22^-/- peritoneal B-1a cells were impaired in this inhibitory role.

Figure 8. Peritoneal B-1a cell transfer from wild type and CD22^-/-to CD19^-/-mice in CHS.

(A) CHS response of CD19^-/- mice, CD22^-/- mice, and wild type mice. They Mice were sensitized and elicited as above by DNFB. (B) Peritoneal B-1a cells from unsensitized wild type and CD22^-/- mice were transferred to CD19^-/- mice. The bottom bars indicate ear swelling of non-transferred wild type mice. CHS was elicited by DNFB one day after transfer. Ear swelling was assessed as described in Fig. 3. *, P<0.05, **, P<0.01.

Discussion

CD22 is a B cell-specific transmembrane molecule that negatively regulates B-cell responses to a range of extracellular signals (32). The current study has demonstrated that CHS was more prolonged in CD22^-/- mice (Fig. 1). When various fractions of splenic, nodal, and peritoneal lymphocytes derived from wild type donors were transferred to sensitized CD22^-/- recipients, only transfer of peritoneal CD5⁺CD11b⁺ B-1a cells extinguished prolonged ear swelling of CD22^-/- mice (Fig. 4) Furthermore, peritoneal B-1a cells from IL-10^-/- mice failed to suppress CHS response (Fig. 6B). [Does the suggested change in the previous sentence retain your original meaning?]. Recent studies have extensively clarified that inhibitory subsets of lymphocytes play crucial roles during inflammatory and immune responses. In addition to regulatory T cells (36), we have recently reported that splenic CD1d^hiCD5⁺ B cells play a suppressive role in CHS (13). The current study has demonstrated that peritoneal B-1 cells also serves as a regulatory B cells in CHS via IL-10 production.

B-1 cells are long-lived and self-renewing B cells that mainly reside in peritoneal and pleural cavities, although while they are also found in the spleen in smaller numbers. B-1 cells are defined by a pattern of surface marker expression, B220^lo, IgM^hi, IgD⁺, CD9⁺, CD43⁺, and CD23^lo, as opposed to conventional circulating B-2 cells that are B220^hi, IgM^hi/lo, IgD⁺, CD9^-, CD43^-, and CD23^hi(37-39). B-1a cells express CD5, while the other B-1 sister populations of B-1b cells share all the surface markers phenotypes with B-1a cells except CD5 expression. B-1 cells are the major source of natural Abs, which are poly-reactive and weakly autoreactive. They recognize antigens from many common pathogens, and thus are very important for the early response to bacterial and viral infections (40). Previous studies have demonstrated the possible possibility of contribution of B-1 cells to immune disorders. In CHS, a promoting, rather than inhibitory, role of peritoneal B-1 cells has been reported. Using Btk-defective xid mice on a CBA background, Tsuji et al. have demonstrated that antigen-specific IgM Abs from peritoneal B-1 cells are required for the recruitment of effector T cells in the early elicitation phase (41). By contrast, Btk^-/- mice on a C57BL/6 background have been reported to exhibit augmented CHS in response to DNFB (42). Similarly, we reported that CD19^-/- mice, which have decreased numbers of B-1a cells, but normal B-1b cell numberss, exhibited augmented CHS responses. The discrepancy of these results may be due to the differences in mouse backgrounds and/or haptens. Nonetheless, in the current study, the early elicitation phase of CHS in CD22^-/- mice was were similar to that in wild type mice, suggesting that the initiation of CHS responses is not impaired in CD22^-/- mice. Furthermore, serum transfer from CD22^-/- mice, whether they were sensitized or not, did not exacerbate CHS in wild type mice (Fig. 4). Collectively, it is unlikely that IgM Abs cause the prolonged CHS in CD22^-/- mice. RatherIn fact, cell transfer experiments have elucidated that peritoneal B-1 cells have a regulatory function in CHS. Intriguingly, Lyn-deficient mice, which display hyperactive B-1 cells and IgM hyperglobulinemia, display show the augmented severity of EAE (43). Lyn is a src-kinase family member abundantly expressed in B cells and is responsible for CD22 phosphorylation. For augmented EAE in Lyn-deficient mice, IgM Abs from B-1 cells are mainly suspected to be responsible. However, depletion of CD5⁺ B-1 cells during the induction phase of the disease resulted in an increase in the incidence of EAE and in the clinical score, while depletion during the effector phase of EAE also decreased the severity and the incidence (44). Recently, Shimomura et al. have also reported a regulatory role for of B-1 cells in chronic colitis (45). Thus, during inflammatory responses, B-1 cells may have complex roles that vary depending on the phase of the response.

Paradoxically, although CD22^-/- mice have an increased number of B-1a cells that can produce IL-10 at wild type levels, peritoneal B-1a cells from wild type mice, but not CD22^-/- mice, can condense converge the late phase of CHS. Itakura et al. have reported that peritoneal B-1 cells are activated to migrate to lymphoid organs immediately after sensitization (46). The migration assay of adoptively transferred peritoneal B cells in the current study showed that a significantly smaller number of peritoneal B cells from CD22^-/- mice, relative to those from wild type mice, were observed than that from wild-type mice in recipient mice's lymphoid organs in recipient mice 5 days after elicitation, while there was no significant difference on day 1. It has been demonstrated that CD22^-/- B cells predominantly undergo apoptosis in response to BCR stimulation and thus show reduced proliferation in vitro (29, 34). Thus, our data may reminiscent ofimply an impairment ofed survival of CD22^-/- peritoneal B cells. Alternatively, CD22 deficiency may result in defective cell-cell interactions in activated B cells. CD22 constitutively binds with other B cell-surface glycoproteins, such as CD45 and IgM, in cis (27, 47, 48). Thus, CD22 on most B-cell surfaces is ‘masked’ and unable to bind exogenous ligands;, it has been reported that CD22 becomes ‘unmasked’ upon cellular activation (49). Therefore, it is possible to hypothesizeThis suggests that CD22 deficiency may impaired the retention of peritoneal B cells by binding exogenous ligands in lymphoid organs. [We are unsure of your intended meaning in the preceding sentence. The sentence prior to this one does not seem connected to this statement. Do you mean to say something to the affect that ‘CD22 may become unmasked upon cellular activation leading to its role in cell migration which our data would support'?]. Also, CD22 ligands are expressed on sinusoidal endothelial cells of bone marrow, and interaction of CD22 has been implicated in the homing of recirculating B cells to the bone marrow (50). Taken together, either retention or survival of peritoneal B cells may be disrupted in CD22^-/- mice.

While both CD22^-/- mice and CD19^-/- mice showed augmented CHS responses, the reaction pattern appeared different. The most definitive difference is that the remission phase, and not the acute phase, is disturbed in CD22^-/- mice, whereas both in contrast to disturbance of both acute and remission phases are altered in CD19^-/- mice (Fig. 1, Fig. 8A). Like CD22, CD19 is a B cell specific cell-surface molecule. We have demonstrated that CD19^-/- mice lack splenic IL-10-producing CD1d^hiCD5⁺ B cells that suppress CHS (13)[insert reference]. Although the cell-surface phenotype of CD1d^hiCD5⁺ B cells resembles that of MZ B cells, these two populations appeared distinct since CD22^-/- mice lack MZ B cells but possess a normal number of CD1d^hiCD5⁺ B cells. These results imply that CHS is suppressed in the remission phase by a different pathway from that in the acute phase and that peritoneal B-1a cells have an inhibitory role in the remission phase of CHS. This is also supported by the adoptive cell transfer of wild wild type peritoneal B-1a cells into sensitized CD19^-/- recipients, which extinguished prolonged ear swelling of CD19^-/- mice in the late phase, since CD19^-/- mice have markedly decreased numbers of peritoneal B-1a cells (51). Alternatively, it might be possible that peritoneal B-1a cells that migrate to lymphoid tissue may develop into CD1d^hiCD5⁺ regulatory B cells. Collectively, in CHS, the elicitation phase can be divided into an acute phase and a remission phase, and splenic CD1d^hiCD5⁺ B cells and peritoneal B-1a cells have distinct inhibitory roles in CHS that are mediated via IL-10.

Abbreviations used in this paper

CD19^-/-

CD19-deficient

CD22^-/-

CD22-deficient

CHS

contact hypersensitivity

DNFB

2,4-dinitrofluorobenzene

EAE

experimental autoimmune encephalomyelitis

IL-10^-/-

IL-10-deficient

LN

lymph node

MZ

marginal zone