|Home | About | Journals | Submit | Contact Us | Français|
Dendritic cells (DCs) play an essential role in regulation of immune responses. In the periphery antigen presentation by DCs is critical for adaptive responses and because of this reason DCs are often targets of adjuvants that enhance vaccine responses. Activated mature DCs enhance B cell activation and differentiation by providing cytokines like BAFF and APRIL. However, the role of immature DCs in B cell tolerance is not well studied. Recently mouse immature bone marrow-derived DCs (iBMDCs) have been shown to suppress anti-IgM induced B cell activation. Here we tested the ability of mouse DCs to modulate B cell functions during Toll-like receptors (TLR) activation. We found that iBMDCs potently suppressed proliferation and differentiation of various B cell subsets upon TLR stimulation. However, iBMDCs did not affect CD40 mediated B cell activation. Optimal suppression of B cell activation by iBMDCs required cell contact via the CD22 receptor on B cells. The B cell suppression was a property of iBMDCs or DCs resident in the bone marrow but not mature BMDCs (mBMDCs) or DCs resident in the spleen. Presence of iBMDCs also enhanced the antigen induced apoptotic response of bone marrow B cells suggesting that the suppressive effects of iBMDCs may have a role in B cell tolerance.
Dendritic cells (DCs) play a significant role in initiation and regulation of the antigen-specific T cell immune response. Immature DCs are very efficient in antigen (Ag) capture and processing. Ag uptake initiates the closely-linked process of maturation and migration (1–3). Maturation of DCs continues during migration to the draining lymphoid organs and is completed during the DC-T cell interaction. Several molecules, such as CD40, IL-1 receptor, Toll-like receptor (TLR) and the TNF receptor family members have been shown to regulate multiple aspects of DC maturation (1, 4, 5). Mature DCs activate naïve T cells by presenting antigenic peptides bound to major histocompatibility complex (MHC) class II, and by providing co-stimulatory signals via CD80/86 (4).
DCs, besides their well-studied role in naïve T cell activation, have also been shown to directly affect B cell function. DCs enhance proliferation and differentiation of B cells that have been stimulated through CD40 ligands on activated T cells (6, 7). DCs also induce surface IgA expression in CD40-activated B cells through transforming growth factor beta (TGF-β) secretion. However, direct interaction of DCs with B cells in the presence of TGF- β and IL-10 is critical for immunoglobulin class-switching to IgA1 and IgA2 (8). DCs also secrete B cell activating factors belonging to the TNF family such as BAFF/BLyS and a proliferation-inducing ligand (APRIL), which have been shown to enhance B cell survival, proliferation, differentiation and class-switching (9). Follicular DCs, a very specialized type of DCs, take part in the organization of primary B cell follicles and the germinal center reaction (1).
Injection of DCs pulsed with various bacterial, viral or protein Ags induces Ag-specific antibody production in different in vivo studies, suggesting a role for DC-mediated antigen presentation in B cell responses (10–12). In fact, a study by Qi et al. showed that newly arriving naïve B cells examine lymph node DCs for Ags before they enter into lymph node follicles (13). This study also demonstrated that interaction between Ag-specific B cells and Ag-carrying DCs leads to B cell receptor (BCR) signaling and extrafollicular activation of B cells. In other studies DCs have been shown to interact with B cells in lymph nodes and spleen (14) via the integrin, LFA-1(CD11a/18) (15). Blood-derived DCs have been shown to capture particulate Ags and present them to marginal zone B cells in the spleen. This Ag presentation by DCs to marginal zone B cells leads to their differentiation into IgM-secreting plasmablasts (16). These observations altogether suggest a possible role for DCs in the direct activation of Ag-specific B cells during the in vivo immune responses.
TLRs are pathogen-recognition receptors that recognize pathogens via specific components conserved among microorganisms known as pathogen-specific molecular patterns (PAMPs) (17). TLRs are differentially expressed among leukocytes and also within the different subsets of DCs. DCs are mainly divided into two major subsets, myeloid DC and plasmacytoid DC, both of which have unique phenotypes and functions (1, 18). Myeloid DCs express CD11b and are involved in Th1-type immune responses through IL-12 production, while plasmacytoid DCs express B220 and play an important role during viral infection by producing type 1 interferon (IFN) or IFN-α (19, 20). Many different TLR ligands such as peptidoglycan (PGN), lipopolysaccharide (LPS), poly (I:C) and CpG induce DC maturation (21, 22). Stimulation of DCs with TLR ligands induces major morphological and functional changes, such as elevated surface expression of MHC class II and co-stimulatory molecules (CD80, CD86) that are essential for DC-mediated activation of the adaptive immune responses (4). In addition to the positive effects of DCs in enhancing immune responses, DCs have been shown to have a role in negative selection of T cells in thymus, peripheral T cell tolerance and induction of regulatory T cells (23–25). However their role in inducing B cell tolerance has not been investigated extensively.
Recently, it was shown that bone marrow-derived DCs (BMDCs) inhibit BCR-mediated proliferation of B cells (26). However, the role of DCs during the TLR-induced B cell immune responses has not been examined carefully. This is important considering the trend of increasing use of TLR ligands as vaccine adjuvants to activate DCs and B cells to enhance the vaccine-induced antibody response in numerous studies and clinical trials and in the context of several endogenous ligands that appear to act via TLR receptors (27, 28). In view of the importance of DC-B cell cross-talk during TLR-induced responses, we examined the role of DCs in B cell activation in response to various TLR ligands. Our results show that immature bone marrow-derived DCs (iBMDCs) and bone marrow resident DCs (BM-RDCs) profoundly inhibited TLR-induced, but not CD40-induced, B cell proliferation and differentiation. The inhibition of B cells by iBMDCs was dependent on CD22 expression on B cells. Maturation of iBMDCs with TLR ligands before their interaction with B cells abrogated their inhibitory effect on TLR-induced B cell proliferation. Unlike BM-RDCs, splenic resident DCs (Spl-RDCs) had no effect on TLR-mediated proliferation of B cells. Finally, we present the novel observation that iBMDCs and BM-RDCs induced apoptosis of BM B cells in an Ag- and BCR-dependent manner suggesting that BMDCs may influence B cell tolerance mechanisms by providing stronger negative signals when self-antigens are presented by DCs.
C57BL/6 mice were obtained from Harlan (Indianapolis, IN, USA). Hen egg white lysozyme (HEL) specific BCR expressing transgenic mice were obtained from the Jackson Laboratory (BarHarbor, ME, USA). CD22 knock out (KO) mice were a generous gift from Dr. Thomas Tedder. ST6Gal-I sialyl transferase knock out (ST6-Gal I KO) mice generated by Hennet et al. were utilized (29). Mice were housed under specific pathogen-free conditions in micro-isolator cages under the Institutional Animal Care and Use Committee (IACUC) approved protocol. The University of Kentucky IACUC protocol number for this study is 00680M2004. The described studies are approved under this protocol.
TLR agonists: LPS and Pam3CSK4 were obtained from Sigma Chemical Co. (St. Louis, MO, USA) and Calbiochem (San Diego, CA, USA), respectively. Poly (I:C) was obtained from Invivogen (San Diego, CA, USA). The polyclonal goat anti-IgM F(ab′)2 was obtained from ICN/MP Biomedicals, (Irvine, CA, USA). Anti-CD40 (1C10 clone) was a gift from Dr. Maureen Howard and was used as ascites fluid. Anti-B220/CD45RB~FITC/APC, anti- CD86~PE, anti-CD138~PE, anti-CD19~APC-Cy7, anti-IgM~PerCP-Cy5.5, anti-B220~AF700, anti-AA4.1~APC, anti-CD21/35~PE, anti-CD23~PE-Cy7 and anti-I-Ab~PE, were obtained from BD Pharmingen (San Jose, CA, USA). PMA and ionomycin were obtained from Sigma-Aldrich (St. Louis, MO).
BM cells were collected from femur and tibia bones of C57BL/6 or ST6Gal-I KO mice. A single-cell suspension of BM cells was obtained by flushing the bones with RPMI 1640 medium (HyClone, Thermo scientific) supplemented with 10% fetal bovine serum (Atlanta Biologicals), L-glutamine, HEPES and Penicillin/Streptomycin (Invitrogen). Cells were plated in 100×15 mm petri dishes (BD Falcon) at a concentration of 4×106 cells in 10 ml dendritic cell media (cell culture media supplemented with 20% cell supernatant of F10.9 cells that were transfected with murine GM-CSF gene or with recombinant GM-CSF (20ng/ml). Cells were supplemented with fresh dendritic cell media on day 3 and day 6. On day 9, both adherent and non-adherent fraction of cells were harvested and washed free of DC media. For immature and mature DC preparations, day 9 cells were stimulated with media, 1 μg/ml LPS or 50 μg/ml poly (I:C) for 24 hours in IF-12 + 10% fetal bovine serum (FBS) media, respectively.
Splenic B cells were purified as previously described (30) and by positive selection with CD19-magnetic beads and the autoMACS Pro Separator (Miltenyi Biotec.). BM B cells were purified by positive selection with B220-magnetic beads and the autoMACS Pro Separator. T cells were purified by negative selection using magnetic beads from the CD4+ T cell isolation kit (Miltenyi Biotec.), and DCs from BM and spleen were purified by positive selection with CD11c-magnetic beads and the autoMACS Pro Separator (Miltenyi Biotec. Auburn, CA).
Splenic or BM B cells (1 × 105) were stimulated in the presence or absence of different numbers/subsets of DCs with 5 μg/ml LPS, 5 μg/ml Pam3CSK4, 50 μg/ml poly (I:C), PMA (30 ng/ml), ionomycin (100 ng/ml) or anti-CD40 (1C10) (20 μg/ml) ascites in 200 μlIF-12 media (1:1 mixture of Iscove’s DMEM and Ham’s F12 with 10% fetal calf serum). CD4+ T cells (1 × 105) were stimulated with 10 μg/ml anti-CD3 in RPMI media. Costar 96 well plates were used for the cell cultures. Cultures were pulsed with 3[H] thymidine on day 2 or day 3 and were harvested after 4 hours (Packard, Meriden, CT); the incorporated radioactivity was then measured using a Matrix 96 β-counter (Packard, Downers Grove, IL). Results are represented as mean ± SD of counts per minute (CPM) from triplicate cultures.
Splenic B cells (resuspended at 107/ml in PBS/0.1% BSA, 10 μM CFSE) were incubated at 37°C for 20 min, and then washed with IF-12 medium + 10% FBS as described earlier (31). CFSE-labeled B cells (2.5×105) were cultured with LPS (5μg/ml) in the presence or absence of DCs for 2 days at 37°C with 5% CO2 and analyzed by flow cytometry. Results are represented as mean ± SD of triplicate cultures.
Splenic B cells (2.5 × 105) were cultured with 5 μg/ml LPS for 2 days in the presence or absence of iBMDCs. For cell cycle analysis B cells were fixed in 70% (v/v) ethanol for at least 1 h at 4°C, after which the cells were incubated in a mixture of 1 μg/ml propidium iodide (PI) (Sigma-Aldrich)and 25 μg/ml RNase A (Sigma-Aldrich) at 37°C for atleast 30 min. The level of PI fluorescence was measured by flow cytometry in the B220+ B cell population. Cell populations at subG1, G1, S, G2/M phase were calculated using the ModFit analytical program. Results are represented as mean ± SD of triplicate cultures.
Splenic B cells were cultured with 5 μg/ml LPS for 5 days in the presence or absence of iBMDCs. At the end of culture periods, plates were centrifuged, culture supernatants were collected and total IgM levels in the supernatants were estimated using ELISA. O.D. was measured on an ELISA plate reader (Multiskan MCC/340, Thermo Scientific, USA) at dual wavelengths of 405nm and 630nm. Results are represented as mean ± SD of triplicate cultures.
BM recirculating mature B cells and pre + immature B cells were sort-purified as live singlet CD19+B220hiAA4.1− and B220lowAA4.1+CD43− cells, respectively. BM-RDCs were sort-purified as live singlet CD11c+ cells. All the sorted B cells were labeled with CFSE. Sorted CFSE labeled Fo B cells, recirculating mature B cells or pre + immature B cells (1 × 105) were stimulated with 5 μg/ml LPS in the presence or absence of BM-RDCs (5× 104) for 48 hours. After 48 hours of culture, recirculating mature B cells were separated from cultured BM-RDCs as AA4.1-CD19+ CFSE+ cells. Immature B cells were separated as IgM+CD19+ CFSE+ from cultured BM-RDCs and pre-B cells in FACS analysis. B cell death was analyzed by inclusion of live/dead marker-DAPI among above-mentioned gated B cells.
BM B cells (1 × 105) were stimulated with 25 μg/ml anti-IgM F(ab′)2 in the presence or absence of iBMDCs. Alternatively, BM B cells (1 × 105) from HEL transgenic mice were incubated with HEL pulsed (10 μg/ml, 2 hours) iBMDCs for 36 hours. At the end of the culture period cells were stained with CD19/B220~APC and Annexin V~FITC at room temperature for 15 min in the dark. Then 2 μl of PI solution (0.5 μg/ml) was added and samples were analyzed by flow cytometry within one hour.
Paired student’s t-test was used to determine statistical significance of differences between various groups.
iBMDCs have been recently shown to inhibit BCR-induced proliferation of B cells (26). We questioned if this was unique to BCR-mediated growth responses and therefore tested the effect of iBMDCs on TLR-dependent B cell activation. We found that GM-CSF cultured iBMDCs (85–95% CD11c+) strongly inhibited the LPS (TLR4)-induced proliferative response of splenic B cells in a dose dependent manner (Fig. 1A). The inhibition of the TLR response was much stronger than that of the BCR response (data not shown), as one iBMDC per 64 B cells inhibited the LPS-induced proliferation significantly compared to one iBMDC per 16 B cells required for appreciable inhibition of BCR-induced proliferation, as reported by Santos et al.(26) (Fig. 1A). DCs and B cells have been shown to form tight clusters (6) and DCs can directly enhance growth, differentiation and antibody production by CD40 mediated activation of B cells (6, 32). This predicted that iBMDCs should not inhibit the B cell responses to CD40 ligation. Accordingly, when splenic B cells were stimulated with a low dose of anti-CD40 that does not induce DC maturation (as shown by lack of CD86 and Class II up-regulation, data not shown), there was no inhibition of CD40 stimulated B cell proliferation (Fig. 1B).
PMA and ionomycin are known to induce T and B cell activation and proliferation, bypassing the early signaling steps of receptor activation and directly elevating Ca2+ fluxes and PKC activity. We stimulated splenic B cells with PMA and ionomycin in the presence of iBMDCs to determine if the inhibition of B cell proliferation by iBMDCs is a pre-or post-PKC activation event. There was no inhibition of B cell proliferation by iBMDCs over a wide range of DC/B cell ratios when B cells were stimulated with PMA and ionomycin, even as high as one iBMDC per four B cells (Fig. 1C). Immature DCs are deficient at processing and presenting antigens to T cells and at activation of naïve T cells (2). We used CD4+ T cells as control cells to determine if iBMDCs also inhibit T cell proliferation. CD4+ T cells were stimulated by anti-CD3 in the presence or absence of a graded number of iBMDCs. There was no inhibition of the anti-CD3 induced T cell proliferation (Fig. 1D) suggesting that inhibitory effects of iBMDCs are specific to TLR4 and BCR induced B cell responses. Moreover, the lack of inhibition of PMA and ionomycin induced B cell proliferation, suggests that iBMDCs affect receptor induced (TLR, BCR) signaling pathways at the proximal receptor level. B cells undergo blast transformation before they enter into S phase of cell cycle. We tested if this earlier step in B cell response to TLR4 stimulation was affected by iBMDCs. Indeed, TLR4 induced blast transformation response, as measured by light scatter, was suppressed by iBMDCs in a dose dependent manner (Fig. 1E). In addition we also measured changes in expression of MHCII, CD80, CD69 and CD86. Interestingly, expression of MHCII but not CD80, CD69 and CD86 on B cells was suppressed by iBMDCS (data not shown).
Putative differences in the sensitivities of different splenic B cells subsets was tested by phenotyping CFSE labeled B cells cultured with different ratios of iBMDCs in the presence of TLR4 ligand. Top panel in Figure 1F shows CFSE dilution of each B cell subset when cultured with or without iBMDCs at a DC:B-cell ratio of 1:4. The complete dose response of DC: B-cell ratios for each B cell subset is shown in the lower panel of Figure 1F. Clearly TLR4 induced proliferation responses of all B cell subsets (follicular, marginal zone and transitional) examined were suppressed by the iBMDCs, although there are some subtle differences. In a separate experiment, these B cell subsets were first separated by FACS sorting and then were cultured with iBMDCs but the results were similar in that TLR4 responses of all three B cells subsets (Fo, MZ and Tr) were suppressed by iBMDCs (data not shown).
Since mature DCs have been shown to help T and B cell responses (1, 2, 6, 10, 33), we questioned if this inhibition of the B cell growth response by iBMDCs is abrogated upon maturation. iBMDCs were stimulated with LPS for 24 hours in vitro for maturation. There was an up-regulation of CD86 and MHC class II expression on BMDCs after LPS stimulation compared to unstimulated cells (Fig. 2A), confirming that TLR signaling induced iBMDC maturation. Magnetic bead purified CFSE-labeled splenic B cells were stimulated with LPS along with either iBMDCs or mature BMDCs (mBMDCs) (LPS matured, as shown in Fig. 2A) at different ratios and B cell division was determined by CFSE dilution. There was a strong inhibition of cell division when B cells were cultured with iBMDCs (60% inhibition at DC/B cell ratio of 1/8) (Fig. 2B, left panel). Interestingly, LPS-induced cell division of B cells was not significantly reduced by mBMDCs except at a high DC/B cell ratio of 1/8 (Fig. 2B, right panel). Similar results were found in a thymidine incorporation assay when we used TLR3 ligand [poly (I:C)] matured BMDCs. We confirmed that poly (I:C) significantly increased expression of CD86 and MHC class II on iBMDCs compared to media treatment (data not shown). LPS induced proliferation of splenic B cells was inhibited with iBMDCs, but not with poly I:C matured BMDCs over a wide range of DC/B cell ratios (Fig. 2C).
TLR4 is unique among TLRs, as it can activate both MyD88-dependent and TRIF-dependent signaling and is the only known TLR which utilizes all four TIR domain-containing adaptors (34). These unique characteristics of TLR4 led us to examine if iBMDCs can inhibit proliferation of B cells mediated by other TLRs and whether maturation of iBMDCs can overcome the inhibitory effects. We found that iBMDCs inhibited the proliferation response of splenic B cells when stimulated with the TLR2 ligand, Pam3CSK4, (Fig. 2D) or the TLR3 ligand, poly (I:C) (Fig. 2E). Similar to LPS, maturation of iBMDCs with poly (I:C) overcame the inhibitory effects of iBMDCs on B cells stimulated with Pam3CSK4 (Fig. 2D) or poly (I:C) (Fig. 2E).
Proliferation is usually arrested when B cells begin to differentiate into plasma cells or memory cells. DCs are known to enhance the differentiation of naïve B cells into plasma cells upon CD40 mediated activation (6). Therefore, it is possible that iBMDCs help B cells become plasma cells during TLR-mediated activation of B cells, which results in decreased proliferation of B cells. To examine this possibility, we evaluated plasma cell formation upon LPS stimulation of B cells in the presence or absence of iBMDCs. B cells express the plasma cell marker CD138 upon differentiation into plasmablasts and plasma cells (35). B cells also start losing B220 expression on the surface during differentiation into plasma cells. As such, we measured the number of CD138+B220lo plasma cells upon LPS stimulation in the presence or absence of iBMDCs. Large numbers of plasma cells were generated by day 3 upon LPS stimulation of B cells, but the addition of iBMDCs to the B cell cultures inhibited plasma cell formation up to 76% (Fig. 3A). There was also a significant decrease in total IgM production after day 5 by LPS stimulated B cells upon addition of iBMDCs (Fig. 3B). These results suggest that iBMDCs do not help in B cell differentiation but inhibit B cell activation.
To evaluate the inhibitory effect of iBMDCs on B cell activation, we examined the effect of iBMDCs on cell cycle progression. Splenic B cells were cultured with LPS in the presence or absence of different doses of iBMDCs for 48 hours, stained with propidium iodide (PI) and analyzed by flow cytometry. Representative results are shown in Figure 3C, and a summary of the analysis is shown in Figure 3D. The presence of iBMDCs induced predominantly G1 arrest in B cells (cells in G1 phase increased from 41% to 75%, Fig. 3C & 3D). Accordingly, there was a significant decrease in the percentage of B cells in S phase of the cell cycle (from 41% to 13%, Fig. 3C & 3D). In the absence of iBMDCs the G1/S ratio of LPS stimulated B cells was one, and increased to 5.8 in the presence of iBMDCs when cultured at a DC/B cell ratio of 1/4. These results suggest that iBMDCs inhibit progression of LPS activated B cells from G1 to S phase. To determine if G1-S arrest leads to apoptosis, we isolated B220+ cells by magnetic beads and quantified apoptotic cells (annexin ad PI staining) upon TLR stimulation in the presence of iBMDCs. Figure 3E shows that iBMDCs induce apoptosis in these highly purified splenic B cells activated via TLR4.
DCs have been shown to enhance growth and differentiation of CD40-activated B cells (6). Further, it has been shown that DCs can activate T cells that subsequently induce activation and differentiation of B cells through CD40 signaling (1). In this study iBMDCs did not inhibit CD40 stimulated B cell proliferation (Fig. 1B), but rather increased CD40 stimulated B cell proliferation when cultured with mBMDCs (data not shown), which parallels results reported by Dubois et al. (6). Hence, we evaluated whether CD40 co-stimulation, a surrogate for T-dependent B-cell activation, can overcome the iBMDC-mediated inhibition of TLR-induced B cell activation. B cells were cultured with LPS or LPS+anti-CD40 in the presence of iBMDCs. There was an 80% inhibition of B cell proliferation when stimulated with LPS alone in the presence of iBMDCs (1/4 ratio), which was abrogated when anti-CD40 was included in these cultures (Fig. 3F). This result suggests that cognate signaling from T cells can overcome the inhibitory effect of iBMDCs on B cell activation.
BM is a primary site for B cell development and also provides a niche for mature B cells and plasma cells (36, 37). Sapoznikov and colleagues have shown that BM-RDCs play an important role in the survival of mature B cells in BM by producing macrophage migration inhibitory factor (38). Since the effects of BM-RDCs on BCR or TLR induced B cell responses have not been well studied, we examined if BM-RDCs also have an inhibitory effect on BM B cells. We used CD11c and B220 magnetic beads to purify BM-RDCs and BM B cells, respectively. BM B cells were stimulated with LPS in the presence or absence of BM-RDCs. Interestingly, BM-RDCs behaved like the iBMDCs in inhibiting the LPS-mediated BM B cell proliferation response in a dose dependent manner (Fig. 4A, left panel). Similar to cultured iBMDCs, BM-RDCs also inhibited splenic B cell response to TLR4 stimulation (Fig. 4A, right panel). However, BM-RDCs helped the proliferative response of B cells from either BM or spleen when stimulated via CD40 (Fig. 4B), suggesting that the inhibitory effects of BM-RDCs are TLR4 specific.
Dendritic cells in humans and mice are mainly divided into two subsets, plasmacytoid DCs and myeloid DCs. Myeloid DCs are potent antigen presenting cells and help in initiation of the adaptive immune response by activating T cells, while plasmacytoid DCs have a modest capacity to activate T cells (39), but plasmacytoid DCs produce very high amounts of type I interferon upon viral stimulation (40, 41). Both plasmacytoid and myeloid DCs are present in mouse bone marrow (BM). To identify the inhibitory subset of iBMDCs we first characterized the DCs in GM-CSF treated BM cell cultures and found that more than 70% of the DCs in the culture were myeloid DCs (CD11c+CD11b+, data not shown). We purified the CD11c+CD11b+ myeloid DCs from these cultures by flow cytometric sorting and tested their ability to inhibit B cell proliferation. Highly purified myeloid DCs inhibited the LPS mediated, but not anti-CD40 mediated, splenic B cell proliferation (Fig. 4C &D), suggesting that myeloid DCs play a major role in regulation of TLR4 induced B cell responses.
Spl-RDC function has been shown to be regulated by microbial products such as LPS, which induce maturation of Spl-RDCs and migration to T cell areas where they regulate T cell function (42). Spl-RDCs also interact with B cells in the red pulp (14) and T cell-B cell borders (16) in the spleen. This interaction supports plasma blast formation, survival and their differentiation into plasma cells (14, 16). One report suggests that Spl-RDCs can also tolerize B cells and inhibit activation of B cells with low doses of antigen (43). To understand whether Spl-RDC-B cell interaction can also inhibit B cell responses, as did iBMDCs, we evaluated sort-purified Spl-RDCs (> 95% CD11c+). Upon maturation with LPS, Spl-RDCs showed upregulation of CD86 and MHC class II expression when compared to just media-treated cells (Fig. 5A). Unlike iBMDCs or BM-RDCs, Spl-RDCs did not inhibit the splenic B cell response to LPS (Fig. 5B). LPS matured Spl-RDCs behaved like media treated Spl-RDCs and had neither a stimulatory nor an inhibitory effect on splenic B cell proliferation (Fig. 5B). Similarly, the splenic B cell response to anti-CD40 and receptor independent (PMA + ionomycin) stimulation remained unchanged when cultured with Spl-RDCs (Fig. 5C). Moreover, Spl-RDCs neither inhibited nor enhanced CD4+ T cell proliferation when cultured at different DC/T cell ratios (Fig. 5D).
Immature DCs have been shown to induce tolerance in T cells (25, 43) and to inhibit BCR-mediated proliferation of B cells (26) through contact dependent events. When B cells were separated from iBMDCs by a membrane barrier in a trans-well culture there was a significant decrease in the inhibition of LPS induced B cell proliferation (Fig. 6A). Thus, when iBMDC and B cells were co-cultured at a DC/B cell ratio of 1/32 the LPS response of B cells was reduced by 60%, and this inhibition was almost completely abrogated when B cells were separated from iBMDCs by a membrane. However, at the high DC/B cell ratio (1/10), there was still appreciable inhibition of B cell proliferation even when iBMDCs and B cells were not in contact suggesting a role for a soluble mediator (Fig. 6A). Because of our recent data about the potent inhibitory effects of IL-10 on TLR induced B cell responses (31), we tested if antibodies to the IL-10 receptor could overcome iBMDC mediated inhibitory effects. However, inhibition of IL-10 signaling did not overcome the iBMDC induced growth inhibition (data not shown). Thus, at low DC/B cell ratios, iBMDC mediated inhibition of B cells required cell-to-cell contact between iBMDCs and B cells, possibly through inhibitory surface receptor(s).
A number of inhibitory receptors have been described on B cells (44), and among these CD22 has been shown to inhibit the BCR mediated B cell response when engaged by iBMDCs (26). CD22, a member of the Siglec (sialic acid binding Ig-like lectin) family, is known to recruit the Src homology region 2 domain-containing phosphatase-1 (SHP-1) upon activation and to negatively regulate BCR mediated signaling in B cells (45, 46). CD22 can participate in cell-cell interactions by binding to its ligand, α2-6-linked sialic acid (47). Several cell types, including DCs, express ST6Gal I which is required for the production of α2-6-linked sialic acid. We examined if CD22 signaling plays an inhibitory role in iBMDC mediated regulation of B cells upon TLR signaling. B cells from wild type and CD22 KO mice were stimulated with different TLR ligands in the presence or absence of iBMDCs. Unlike wild type mice, CD22 KO B cell proliferation was not inhibited by iBMDCs when stimulated with Pam3CSK4 (Fig. 6B) or LPS (Fig. 6D) at different DC/B cell ratios. These results demonstrate that iBMDCs inhibit various TLR-induced B cell proliferation responses via CD22. We also examined if, in the absence of CD22 mediated negative regulatory signaling, mBMDCs aid in the B cell proliferation response. We did not see any significant increase in proliferation of CD22 KO B cells or wild type B cells when co-cultured with mBMDCs and stimulated with Pam3CSK4 (Fig. 6C) or LPS (Fig. 6E). Since the major CD22 ligands are generated by the sialyl transferase ST6Gal I (29), we tested if iBMDCs from ST6-Gal I KO mice are deficient in their ability to regulate B cell responses to TLR4. Surprisingly, but in agreement with the previous studies on BCR responses (26), iBMDCs from ST6-Gal I KO mice were as efficient as wild type iBMDC in suppressing TLR4 responses (Fig. 6F).
The role of DCs during T cell negative selection is well studied (24, 25). However, it is not known what role DCs play during B cell negative selection in BM. We asked if interactions with DCs can lead to apoptosis of BM B cells during the TLR or BCR mediated responses. We observed increased death of BM immature B cells in the presence of BM-RDCs upon TLR4 stimulation. Similarly, BM-RDCs also induce death of BM recirculating mature B cells (Fig. 7A). We also examined the survival of highly purified BM B cells upon BCR stimulation in the presence or absence of iBMDCs. Immature B cells from BM have been shown to undergo apoptosis in vitro when cultured with anti-IgM antibodies (48, 49). As expected, we found a modest increase in apoptosis of BM B cells in the presence of low dose anti-IgM compared to untreated BM B cells (Fig. 7B). Notably, addition of iBMDCs induced a dose dependent increase in BM B cell apoptosis during BCR stimulation. iBMDCs alone even at the highest DC/B cell ratio did not increase BM B cell apoptosis (Fig. 7B). These results suggest that iBMDCs increase apoptosis of BM B cells when they are triggered via BCR signaling. To examine the Ag-specific apoptosis of BM B cells by iBMDCs, we used transgenic mice that have high frequencies of B cells expressing a high affinity BCR against hen-egg white lysozyme (HEL). Good now and colleagues demonstrated elegantly that soluble antigen induces BM B cell anergy in HEL transgenic mice, whereas membrane bound self-antigen has been shown to induce clonal deletion (50, 51). Therefore, we determined if the presentation of HEL to transgenic B cells via iBMDCs can induce apoptosis of these Ag-specific B cells. Purified BM B cells from HEL transgenic mice were cultured with HEL pulsed iBMDCs. iBMDCs alone reduced spontaneous apoptosis of BM B cells but strongly enhanced the apoptosis of BM B cells when they were pulsed with HEL (Fig 7C). Pulsing of iBMDCs with HEL alone did not alter expression of MHC II, CD80 and CD86 on DCs (data not shown). These results suggest that DCs may play a crucial role during the B cell negative selection in the BM.
In recent years, naturalor synthetic TLR ligands have been used to boost DC driven Ag-specific immune responses in mice and humans (27, 28). A better understanding of the role of DCs during TLR-mediated B cell activation would provide important information about the use of TLR ligands as adjuvants during DC-based vaccine development. In this regard, we performed comprehensive in vitro studies, including proliferation, cell cycle progression, differentiation, antibody production, and apoptosis of B cells in the presence or absence of iBMDCs, mBMDCs, BM-RDCs and Spl-RDCs. iBMDCs or BM-RDCs, but not mBMDCs or Spl-RDCs, strongly inhibited TLR (2, 3 & 4)-induced B cell proliferation.
Among various mitogens used in our study iBMDCs specifically inhibited TLR2, TLR3 and TLR4 as well as BCR-induced proliferation, but did not inhibit anti-CD40 or PMA-ionomycin induced B cell proliferation. These results suggest that iBMDC-mediated inhibition of B cells is signaling pathway specific and does not have global growth inhibitory effects. The cell cycle analysis data revealed that iBMDCs block B cell proliferation by inducing G1-S growth arrest. Further our study showed that the cell cycle arrest of B cells in the presence of iBMDCs is not due to their differentiation, as plasma cell formation and antibody production were also decreased in the presence of iBMDCs. Immature DCs have been shown to inhibit and tolerize T cells in vivo due to the absence of a second (co-stimulatory) signal, whereas maturation of DCs by TLR or CD40 overcomes this inhibitory effect on T cells (25, 52, 53). Similar to the DC-mediated regulation of T cells, maturation of iBMDCs with TLR ligands overcame the inhibitory effect of DCs on B cells. Many different TLR ligands can mature DCs in either a MyD88 dependent or MyD88 independent manner (21). Even though different TLR ligands are able to induce DC maturation, there were noticeable differences observed in terms of cytokine and chemokine secretion patterns depending on the TLR ligand (54). We found that the maturation of iBMDCs via the MyD88 dependent pathway [LPS, Peptidoglycan (data not shown)] or the MyD88 independent pathway [Poly (I:C)] had similar effects in overcoming the inhibitory effect of iBMDCs on B cells. Co-stimulating B cells with anti-CD40 also overcame inhibitory effects of iBMDCs on B cells.
The iBMDC-mediated inhibition of B cells was dependent significantly on cell contact. This is in agreement with Santos et al. who showed that iBMDCs inhibited BCR mediated B cell responses in a contact dependent manner through CD22 mediated signaling on B cells (26). We observed that CD22 expression is also critical for the iBMDC-mediated inhibition of TLR-induced B cell proliferation. The absence of CD22 on B cell surfaces completely abrogated iBMDC-mediated inhibitory effects during TLR3 or TLR4 induced B cell proliferation. Interestingly, maturation of BMDCs has been reported to decrease the expression of sialic acid ligands (55), which is consistent with the role of CD22 in DC-B cell interaction. In parallel with finding of Santos et al. (26), we found that expression of ST6Gal I, one of the enzymes that induces sialylation, was not required for DC-B cell interaction. Presently we cannot rule out the possibility that CD22 on B cells binds some other ligand on iBMDCs and inhibits their response to BCR and TLR mediated signaling. Despite the importance of CD22, present studies also cannot rule out a role for other sialic acid binding proteins like Siglec-G, which have been shown to be important for B-1 cell responses and for antigen induced tolerance in B cells (56). Moreover, the inability of mature DCs to inhibit B cells could also be due to generation of rescue signals such as the gene expression pattern induced by CD40 ligation.
Our study with resident DCs from spleen and BM showed that inhibitory effect of DCs on B cells is a characteristic of BM-RDCs but not Spl-RDCs. To our knowledge, such a difference between BM-RDCs and Spl-RDCs in affecting B cell activation was never reported before. Spl-RDCs without stimulation were immature, characterized by low CD86 and Class II expression on the surface, compared to LPS matured Spl-RDCs. However, Spl-RDCs did not inhibit TLR induced B cell responses. In contrast, DCs isolated from the BM showed strong inhibition of TLR-induced proliferation of B cells from both BM and spleen. BM-RDCs resembled iBMDCs in enhancing CD40 mediated proliferation of B cells from both BM and spleen. This difference between BM-RDCs and Spl-RDCs is interesting in the context of the requirement for CD22, as there is no data about differences in the ubiquitously expressed ligands containing sialic acids on these two DC populations. Although iBMDCs normally are a heterogeneous population of DCs, flow cytometric sorting established that myeloid DCs in this population have potent inhibitory effects on B cells. It is of interest that Kilmon et al found that myeloid DCs and macrophages, but not plasmacytoid DCs inhibited auto-antibody production by self-reactive B cells but not antibody production by normal B cells (57).
Finally, the inhibitory effects of iBMDCs in BM B cell growth responses led us to examine the role of iBMDCs in Ag-dependent selection of B cells. Negative selection of self-reactive B cells may occur via receptor editing, anergy or deletion. B cell tolerance involves anergy when soluble antigen is present but utilizes clonal deletion mechanisms when self-antigens are membrane bound. This was well established by the studies of Good now et al using soluble and membrane bound HEL and by Nemazee et al using membrane bound MHC class I molecules (51, 58). Therefore DCs may provide an opportunity to present self-antigens in a membrane bound form leading to deletion of self-reactive B cells. We found that HEL specific BM B cells showed increased apoptosis when HEL was presented by iBMDCs. Currently we don’t know the underlying mechanism. In the literature there is evidence for presentation of native antigen by DCs (59). Hence one scenario for increased apoptosis with HEL pulsed BM B cells could be simultaneous binding of antigen to BCR and CD22 ligands on DCs to CD22 on B cells, although other explanations cannot be ruled out at this time. In thymus, DCs present self-antigen in the context of their MHC and play a decisive role in the negative selection of T cells (60). Our study is first to suggest that DCs in the BM might play a similar role in B cell negative selection by presenting self-antigens to B cells in the context of CD22/SIGLEC mediated inhibitory signals. In this context, recently it has been shown that decoration of T-independent antigens with sialic acid epitopes made them tolerogenic through their ability to cross-link SIGLEC family proteins (56). It was proposed that self-antigens that behave like T-independent antigens may utilize this pathway for self-tolerance. In support of such an idea, mice doubly deficient for CD22 and Siglec G developed autoantibodies and a moderate form of immune complex mediated glomerular nephritis (61). Although the significance of inhibition of TLR responses of BM B cells is at present unclear, it must be noted that several endogenous TLR ligands have been identified and have been implicated in the breakdown of self-tolerance in several autoimmune models (62–65). Some of the endogenous TLR ligands such as high-mobility group box 1 and heat shock proteins have their origin in cell death (66), which is known to occur extensively during B cell development. Defects in clearance of dead cells have a critical role in development of autoimmune diseases like lupus (67). In the periphery DCs in mucosal areas and epidermal Langerhans cells (sites of extensive cell turnover) have anti-inflammatory properties. Further in vivo studies are warranted to address the potential role of DCs in the negative selection of B cells in the BM.
Overall, our study gives valuable insight on the role that immature DCs play in B cell function. As there is increasing use of BM derived DC vaccines and TLR ligands as adjuvants, a careful selection of DC subset and maturation state is warranted to generate B cell specific immune responses.
We thank Dr. Lakshman Chelvarajan and Dr. Alan M. Kaplan for helpful discussions on these studies and on manuscript. We thank Ms. Jennifer Strange and Dr. Greg Baumann for help with flow cytometry.
This work was supported by National Institutes of Health Grants AI076956 to S.B and AI56363 and AQI057157 to T.T.