Myeloid dendritic cells (mDC) activated with a B7-DC-specific cross-linking IgM Ab (B7-DC XAb) take up and retain Ag and interact with T cell compartments to affect a number of biologic changes that together cause strong antitumor responses and blockade of inflammatory airway disease in animal models. The molecular events mediating the initial responses in mDC remain unclear. In this study we show that B7-DC XAb caused rapid phosphorylation of the adaptor protein DAP12 and intracellular kinases Syk and phospholipase C-γ1. Pretreatment of mDC with the Syk inhibitor piceatannol blocked B7-DC XAb-induced Ag uptake with a concomitant loss of tumor protection in mice. Vaccination with tumor lysate-pulsed wild-type B7-DC XAb-activated mDC, but not TREM-2 knockout XAb-activated mDC, protected mice from lethal melanoma challenge. Multimolecular caps appeared within minutes of B7-DC XAb binding to either human or mouse mDC, and FRET analysis showed that class II, CD80, CD86, and TREM-2 are recruited in tight association on the cell surface. When TREM-2 expression was reduced in wild-type mDC using short hairpin RNA or by using mDC from TREM-2 knockout mice, in vitro DC failed to take up Ag after B7-DC XAb stimulation. These results directly link TREM-2 signaling with one change in the mDC phenotype that occurs in response to this unique Ab. The parallel signaling events observed in both human and mouse mDC support the hypothesis that B7-DC cross-linking may be useful as a therapeutic immune modulator in human patients.
Dendritic cells (DC)3 are key targets in schemes to regulate immunity for the treatment of cancer, allergy, infectious diseases, transplant rejection, and autoimmunity (1). Activation of DC through the TLR family initiates DC maturation resulting in down-regulation of Ag uptake and migration to regional lymph nodes, where they encounter and activate naive T cells (2, 3). As DC mature, the cell surface expression of a series of costimulatory molecules is up-regulated (4, 5). These costimulatory molecules are critical for the activation of naive T cells (6). The activated DC also produce immunomodulating cytokines that influence the polarity of the ensuing immune response, determining the array of effector mechanisms brought to bear at the site of infection (7, 8).
We have recently described a new approach for modulating the activity of DC that is distinct from previously defined mechanisms, yet results in potent immunomodulatory signals. B7-DC (or, PD-L2) is a B7 family member normally expressed on the cell surface of DC (9). The receptor for this protein, PD-1, is expressed on activated T cells (10). Cross-linking B7-DC in vitro or in vivo with the human IgM Ab (B7-DC XAb) alters a wide variety of important functions of myeloid DC (mDC), inducing 1) enhanced survival, 2) increased ability to process and present soluble Ag in the class I Ag-presenting pathway, 3) enhanced ability to activate naive T cells, 4) increased efficiency of seeding draining lymph nodes, and 5) up-regulation of the key immunomodulating cytokine, IL-12 (11, 12). However, mDC treated with B7-DC XAb do not display traditionally defined maturation phenotypes (12) such as an up-regulation of the costimulatory markers CD80 or CD86 or a concomitant increase in cell surface expression of class II Ag-presenting molecules. Instead, treatment of immature mDC with B7-DC XAb results in increased Ag uptake and even restores the ability of TLR ligand-matured mDC to take up and retain Ag (13). The combination of TLR-9 ligand and B7-DC cross-linking results in a synergistic CTL response against peptide Ag (13). These differences in maturation lead to important biological distinctions between cells activated by traditional approaches compared with cells activated by cross-linking B7-DC. For example, mDC activated with B7-DC XAb are highly efficient modulators of the polarity of Th2 memory cells (14), effectively redirect T regulatory cells into effector cell phenotypes (15), and rapidly activate cytolytic T cell responses (16), whereas DC matured using TLR agonists are relatively inefficient inducers of these changes in T cell behavior.
B7-DC interaction with PD-1 has been shown to result in either a positive response (9) or a negative response (10) by T cells. This discrepancy could be due either to the different model systems used or to the ability of B7-DC to interact with more than one of the receptors that differentially govern T cell responsiveness (17). The positive immune response observed when we treated animals with B7-DC XAb might also be attributable to a blockade of interactions between B7-DC and PD-1. However, adoptive transfer experiments using mDC activated in vitro under conditions in which B7-DC XAb did not physically block expression of accessible B7-DC on the cell surface demonstrated full immunomodulatory capabilities (18).
How cross-linking B7-DC stimulates the immunomodulatory properties of mDC is not known. The plethora of immunomodulatory effects due to fundamental changes in mDC signaling is reflected in how mDC interact with T cell lineages. To understand the mechanisms involved, we previously showed that engagement of B7-DC by this IgM Ab protected mDC from cell death caused by cytokine withdrawal. This protection was mediated by intracellular signals involving PI3K, Akt, and NF-κB (11, 19). Thus, cross-linking of molecules on the surface of mDC by B7-DC XAb appears to elicit back-signals similar to events described following the ligation of B7 family members in studies using a CTLA-4-Ig fusion protein (20, 21). However, B7-DC XAb invoked responses remain poorly defined.
In this study we demonstrate that cross-linking of B7-DC molecules on mouse mDC derived from bone marrow precursors results in phosphorylation of the upstream adaptor protein DAP12 and the protein kinases Syk and phospholipase C (PLC)-γ. Syk→PLC-γ activation leads to the restoration of Ag uptake and retention in matured DC. Tumor lysate pulsed B7-DC XAb mDC pretreated with Syk inhibitor fail to take up and retain Ag or to protect mice against tumors upon adoptive transfer. Importantly, multimolecular complexes involving several cell surface molecules, including TREM-2 (Triggering receptor expressed by myeloid cells-2), form on B7-DC XAb-activated mDC. TREM cells is an activating receptor of the Ig superfamily and regulates the development and function of DC, microglia, and osteoclasts (22). Our studies with TREM-2-deficient mDC show that TREM-2 initiates signaling through DAP12 and Syk and mediates Ag uptake and retention that occurs in matured DC upon cross-linking with B7-DC XAb. These signals and changes in mDC function are necessary for B7-DC XAb to have its unique immunomodulating properties.
Materials and Methods
C57BL/6J and B6.129s4-CD80−/−CD86−/− mice at 6- to 8-wk-old were obtained from The Jackson Laboratory and used for generation of bone marrow-derived DC. OT-II mice for T cell activation assay were also from The Jackson Laboratory. Class II knockout mice (23) were a gift from Dr. C. David (Mayo Clinic, Rochester, MN). TREM-2 knockout mice (24) were bred in the mouse colony at Washington University School of Medicine (St. Louis, MO). All animals were maintained at the Mayo Clinic animal facility for at least 1 wk before use according to protocols approved by the Mayo Clinic Institutional Animal Care and Use Committee.
Abs and other reagents
Appropriate fluorophore labeled Abs against murine I-Ab (25-9-17), murine class II specific IgM (25-9-3), allophycocyanin-labeled anti-mouse CD11c (HL3) FITC-labeled anti-human class II (TU39), FITC-labeled anti-human CD28 (CD28.2), allophycocyanin-labeled anti-human CD28 (CD28.2), PE labeled anti-human CD4 (RPA-T4) and PE-labeled anti-human HLA-A, HLA-B, HLA-C (G46-2.6) were purchased from BD Pharmingen. Appropriate fluorophore labeled Abs against mouse class II (M5/114.15.2), CD80 (16.10A1), CD86 (GL-1), CD11c (N418), allophycocyanin-labeled Ab against human DR (LN3), PE labeled anti-human CD80 (2D10.4), CD86 (IT2.2), murine B7-DC specific IgG Ab (TY25), and human B7-DC IgG Ab (MIH18) were all purchased from eBioscience. All secondary appropriately fluorophore-labeled F(ab)2 fragment Abs used in this study were obtained from Jackson ImmunoResearch Laboratories. An IgM Ab (28-13-3) specific for mouse class I H-2 Kb was obtained from a hybridoma cell line from American Type Culture Collection (HB-41). Ab against the protein kinase Syk (4D10) was obtained from Santa Cruz Biotechnology. Anti-phosphotyrosine (4G10) and goat anti-mouse Abs were obtained from Upstate Cell Signaling Solutions. Anti-mouse TREM-2 Abs 237920 (for flow cytometry) and 237916 (for Western Blot) were purchased from R&D Systems. Rabbit Abs against PLC-γ1 (MC490) and DAP12 (MC457) were developed by Dr. P. Leibson (Mayo Clinic, Rochester, MN). OVA labeled with FITC or allophycocyanin was purchased from Molecular Probes. Protein A-Sepharose was purchased from Pierce. DAPI (4′,6-diamidino-2-phenylindole), DNase, and LPS were obtained from Sigma-Aldrich. All inhibitors used in this study were obtained from Calbiochem unless otherwise indicated. Piceatannol was obtained from Sigma-Aldrich. Rac-1 inhibitor, NSC23766 was a gift from Dr. D. Billadeau (Mayo Clinic, Rochester, MN). CpG oligonucleotides as described (25) were synthesized in Mayo Core Facility. The polynucleotide polyinosinic-polycytidylic acid poly(I:C) was purchased from Calbiochem. All human IgM Abs were purified as described (26) and used at 10 μg/ml.
Generation of mDC
The mDC from mouse bone marrow were generated as described (27). Bone marrow was isolated from the long bones of the hind legs. Erythrocytes were lysed by treatment with ammonium chloride/potassium bicarbonate/EDTA at 37°C. The remaining cells were plated 1 × 106 cells/ml in 6-well plates (BD Biosciences) in RPMI 1640 containing 10 ng/ml murine GM-CSF and 1 ng/ml murine IL-4 (PeproTech). The cells were incubated at 37°C with 5% CO2. After 48 h, the cells were washed and replated with RPMI 1640 containing the same concentration of GM-CSF and IL-4 for another 5 days. Human DC were derived from CD14+ mononuclear cells isolated from peripheral blood using magnetic bead sorting (Miltenyi Biotec). Buffy coat was obtained from a unit of blood donated by a normal human donor. The PBMC were isolated by centrifugation over Ficoll-Paque PLUS (Amersham Biosciences), and CD14+ cells were separated by positive magnetic cell sorting. The isolated cells were incubated in RPMI 1640 supplemented with 5% human AB serum (HP 10220; Valley Biomedical), 1% sodium pyruvate (Mediatech), 1% nonessential amino acids (Mediatech), 1 ng/ml IL-4 (R&D Systems), and 50 ng/ml GM-CSF (Berlex Laboratories) at 1 × 106 cells/ml, 3 ml/well in 6-well plates for 6–8 days at 37°C with 5% CO2. Maturation of DC was achieved by addition of TLR agonist oligonucleotides 1826 (11) for a period of 24 h before being used for Ag uptake assays. The conditions used to culture both bone marrow-derived mouse DC and peripheral blood-derived human DC are favorable toward myeloid-type DC, thus we refer to the DC in this study as mDC.
Production of short hairpin TREM-2 and short hairpin control virus
Oligonucleotides containing the short hairpin TREM-2 sequence 5′-TGATGCTGGAGATCTCTGGGTTCAAGAGACCCAGAGATCTCCAGCATCTTTTTTC-3′ and short hair Control sequence 5′-TGACTGCTGAAGGTCGCTTGTTTCAAGAGACCAAGCGACCTCCAGCATCTTTTTTC-3′ (28) were synthesized and cloned into the pSUPER RNAi System, as provided by Dr. D. Billadeau (Mayo Clinic, Rochester, MN) using the key restriction sites BglII and HindIII. All sequences were confirmed by automated sequencing of the vectors by the Mayo Clinic Molecular Biology Core Facility. The resulting vectors were cotransfected with VSV-G and gagpol plasmids, provided by Dr. R. Vile (Mayo Clinic, Rochester, MN) into 293T cells. Supernatant was collected at 48 and 72 h, pooled, filtered through a 0.45-micron filter and frozen until used for transduction. Viral titers, determined by counting the number of resistant HT1080 cell colonies after selection in puromycin 4 days, were 2.8 × 107 and 3.6 × 107 viral particles/ml for the short hairpin RNA (shRNA) and scrambled control encoding viruses, respectively.
Transduction of mDC
For transducing mDC with the virus, the mDC culture medium was replaced with fresh medium containing 1 ml of supernatant containing the scrambled virus or virus encoding shRNA against TREM-2 and 2 ml of RPMI 1640 (multiplicity of infection of ∼30, see previous). Cytokines were added to a final concentration of 10 ng/ml murine GM-CSF and 1 ng/ml murine IL-4 at day 2 of mDC culture. Cells were maintained for another 3 days before using the mDC for Ag uptake assay as mentioned or for analysis of phosphorylation status of DAP12 and Syk proteins. Using this transient transduction procedure, intracellular staining showed that TREM-2 protein levels were reduced by ∼60%.
Immunoprecipitations and Western blots
Whole cell lysates were prepared from mouse or human DC stimulated with control Ab or B7-DC XAb. In experiments involving inhibition of Syk kinase, cells were preincubated with 10 μM piceatannol for 30 min. For suppression of TREM-2, mDC were transduced as described and were stimulated with control Ab or B7-DC XAb on day 6. Cells were lysed on ice for 10 min in 1 ml buffer containing 20 mM Tris-HCl, 40 mM NaCl, 5 mM EDTA, 50 mM NaF, 30 mM Na4P2O7, 0.1% BSA, 1 mM Na3VO4, 1 mM PMSF, 5 μg/ml aprotinin, 10 μg/ml leupeptin, and 1% Triton X-100. Cellular debris was removed by centrifugation at 20,800 × g for 5 min at 4°C and the supernatant used for further analysis. For immunoprecipitation, Ab (10 μg) against mouse Syk (4D10) or PLC-γ1 (MC490) or DAP12 (MC457) was bound to protein A-Sepharose beads at 4°C for 2 h under constant rotation. Supernatant from cell lysate were added to the Ab-coupled beads and incubated for 2 h at 4°C with constant rotation. Protein complexes were then eluted in 40 μl of SDS sample buffer, resolved by SDS-PAGE, and transferred to Immobilon-P membranes (Millipore). Tyrosine-phosphorylated proteins were detected using the anti-phosphotyrosine specific Ab, 4G10, followed by goat anti-mouse IgG coupled to HRP (Santa Cruz Biotechnology) and the SuperSignal detection system (Pierce). Thereafter, total protein was visualized by staining the membrane with Ponceau staining solution (Pierce) for 30 s in case of analysis of whole cell lysate or in the case of immunoprecipitation assays, the membrane was stripped with 7M guanidine, blocked with BSA, probed with the Ab against the whole protein followed by protein A coupled to HRP (Amersham Biosciences) and the SuperSignal detection system. For analysis of coprecipitating signaling molecules, affinity purified Ab against mouse class II (I-Ab) (KH74) was used for immunoprecipitation. TREM-2 was detected by blot using mouse Ab (237920) and goat anti-mouse coupled to HRP.
Live cell imaging for visualization of multimolecular complex
Mouse mDC were stained with anti-class II-FITC (MF/114.15.2), and either anti-CD80-PE (16.10A1)/CD86-PE (GL-1) or anti-CD11c-PE (N418). Human mDC were stained with anti-class II-FITC (LN3) and anti-CD80-PE (2D10.4)/CD86-PE (IT2.2). DAPI was used to stain the nuclei. All incubations were conducted for 15 min at 37°C. The cells were subsequently stimulated with 10 μg/ml control Ab (sHIgM39) or B7-DC XAb and were observed every 5 min using time lapse confocal imaging at ×40 magnification with a LSM510 Laser scanning confocal microscope with a 37°C stage (Carl Zeiss).
Flow cytometry and fluorescence resonance energy transfer (FRET)
FRET occurs when certain fluorophores are in close enough proximity (<80 Å) such that when one has been excited (the donor), energy can be directly transferred to the other (the acceptor), causing it to fluoresce. A flow cytometry approach using fluorochrome-coupled Abs specific for cell surface molecules was used to study changes in cell surface interactions in response to cross-linking Ab treatment as previously described (29). Briefly, mouse mDC were stained with anti-class II allophycocyanin (M5/114.15.2) and anti-CD80-PE (16.10A1)/CD86-PE (GL-1) or anti-TREM-2-PE (237920). Human mDC were stained with allophycocyanin anti-class II (LN3) and anti-CD80-PE (2D10.4)/CD86-PE (IT2.2). All staining was for 15 min. In experiments involving blocking of B7-DC, both fluorophore-labeled Abs and purified anti-mouse B7-DC (TY-25) or purified anti-human B7-DC (MIH18) IgG mAb, was added at 10 μg/ml for 15 min. Cells were stimulated with control Ab or B7-DC XAb or purified anti-mouse class II IgM (25-9-3) and aliquots from different groups were taken at different time points. After 15 min of incubation, the cells were washed and fixed in 2% paraformaldehyde before analysis by FACS performed by the Mayo Flow Cytometry Core Facility using a FACSCalibur (BD Biosciences). Ab-induced generation of a FRET signal (upon excitation of PE at 488 nm and emission of allophycocyanin at 660 nm) was visualized in FL3 channel (650–670 nm LP). Data collected as log10 fluorescence were analyzed using CellQuest (BD Biosciences).
Ag uptake and retention assay
Experiments to assess the accumulation of Ag (reflecting both Ag uptake and retention) were conducted as previously described (13). Day 5 mDC were matured with TLR ligand CpG oligonucleotides (10 μg/ml) for mouse mDC (11) or poly(I:C) (10 μg/ml) for human mDC. The matured cells were incubated with OVA (1 mg/ml) labeled with FITC or allophycocyanin and control Ab or B7-DC XAb for 2 h, washed, and analyzed by FACSCalibur (BD Biosciences). In studies involving inhibitors, cells were pretreated 30 min as indicated in the experiments or at 10 μM concentration before the addition of control Ab or B7-DC XAb.
T cell stimulation by mDC
To ensure that inhibition of Ag uptake by various signaling inhibitors did not compromise all DC functions, mDC were tested for their ability to stimulate T cell proliferation. Naive mouse splenocytes were harvested from the OT-II mice, plated in triplicate (3 × 105), and stimulated in vitro for 3 days with titrated doses of mDC that were pretreated with different inhibitors (10 μg/ml), pulsed with 10 μM OT-II peptide (ISQAVHAAHAEINEAGR), and treated with control Ab or B7-DC XAb overnight. The cells were pulsed with [3H]thymidine for 18 h before harvest. Tritium incorporation was measured by liquid scintillation.
Tumor vaccine experiments
All in vivo tumor experiments were conducted as previously described (30). Briefly, all the groups of mice were injected with 0.5 × 106 B16 melanoma cells in the right flank. In addition, some mice were vaccinated with B16 melanoma lysate-pulsed wild-type or TREM-2 knockout mDC (2 × 106, i.p.) treated with control Ab or with B7-DC XAb. In some experiments, mice received mDC that were pretreated for 15 min with 10 μg/ml the Syk inhibitor piceatannol before the addition of lysate and B7-DC XAb. After 7 days, draining lymph node cells (from two mice in each group) were harvested, pooled, and used as effectors against the 51Cr-labeled B16 melanoma target cells. The remaining mice were monitored for the tumor growth and mice bearing tumors of size 17 x 17 were euthanized as per the Institutional Animal Care and Use Committee recommendations.
DAP12-Syk-PLC-γ mediates B7-DC XAb-induced Ag uptake and retention in mature human mDC
We have demonstrated earlier that cross-linking B7-DC on mature mDC from mice leads to restoration of the ability of the mDC to take up and retain protein Ag (13). Matured human mDC also responded to B7-DC XAb treatment by regaining the ability to take up and retain Ag (Fig. 1,A). Initial studies into the mechanisms regulating this process showed that global protein phosphorylation was induced as early as one minute after treatment of mDC with B7-DC XAb (data not shown). To determine what kinases were activated and what signaling intermediates mediated the Ag uptake response, we tested whether Src, Syk, and PLC-γ inhibitors blocked steps in the activation pathway. Blockade of Src kinases with PP2 resulted in the inhibition of Ag accumulation by matured mDC upon B7-DC cross-linking (Fig. 1,B). Similarly, pharmacologic agents that block calcium-dependent protein kinase C activity, PI3K, and the Rho family GTPase RAC1 were also inhibitory (Fig. 1,C). B7-DC XAb treatment of mDC resulted in the activation of ERK, but not p38 (data not shown). However, inhibitors of Rho-A, MEK, and p38 pathways did not influence B7-DC XAb-induced Ag accumulation (Fig. 1 C). Assessment of cell viability using Alamar Blue or Annexin V binding showed that 94–98% of the mDC treated with B7-DC XAb were viable and remained so in the presence of the various inhibitors (data not shown).
As no signaling elements in B7-DC have been identified, it seemed likely that an association with adapter molecules containing signaling domains would be required to mediate an mDC response. DAP12 is an adaptor molecule that can couple receptor molecules like TREM-2 that lack innate signaling capability to downstream activation pathways (31, 32, 33). DAP12 was rapidly phosphorylated on tyrosine upon cross-linking B7-DC (Fig. 2,A). Further analysis showed that B7-DC XAb stimulation of human mDC also caused the tyrosine phosphorylation of p72 Syk (Fig. 2,B), shown in other systems to be downstream of a Src kinase (34) and associated with Fc receptor-mediated Ag uptake by B cells, macrophages, and DC (35). PLC-γ1, a downstream substrate for Syk (36) was also phosphorylated (Fig. 2 C). Vav1, a RhoGEF activated by Syk and required for cytoskeletal rearrangements during Ag uptake in B cells (37) also was phosphorylated in DC following activation with B7-DC XAb (data not shown).
Pretreatment of mDC with piceatannol inhibited phosphorylation of Syk and PLC-γ1 (Fig. 2, D and E) and inhibited accumulation of tagged proteins by matured mDC in a dose responsive manner (Fig. 2,F). Inhibition of PLC-γ using U73122, but not with the inactive analog U73343 (Fig. 2 G) also blocked the accumulation of Ag. Thus, the ability of mDC to regain the ability to take up and retain Ag in response to B7-DC XAb is mediated by the activation of multiple kinases, especially Syk and PLC-γ, which are tied to B7-DC by DAP12.
Intact Syk pathway in mDC is necessary for the protection of mice against tumor
We next tested whether the mDC signaling intermediates we identified in vitro were important for the immunomodulatory affects observed using B7-DC XAb. We showed previously that administration of B7-DC XAb into mice leads to the generation of a potent CTL response and tumor clearance (30). Mice immunized with B16 tumor lysate and receiving B7-DC XAb treatment are protected against B16 melanoma tumor (38). Because Syk was the upstream kinase involved in inducing Ag uptake in mDC, we asked whether inhibition of Syk affected B7-DC XAb-induced tumor immunity in a B16 melanoma model. Myeloid DC were treated with piceatannol before being pulsed with the tumor cell lysate and treated with B7-DC XAb or control Ab. Adoptive transfer of the mDC pulsed with the B16 melanoma tumor cell lysate and activated with the B7-DC XAb induced a potent CTL response (Fig. 3,A) and protected the mice against a lethal challenge of B16 tumor (Fig. 3,C). However, mice that received piceatannol-treated mDC did not mount a cytotoxic response and failed to clear the tumor (Fig. 3, B and C). Peptide-pulsed mDC treated with piceatannol were still able to stimulate T cell proliferation in vitro (Fig. 3 D). Taken together, these findings indicate that the B7-DC XAb activation of the Syk pathway in mDC is required for the processing of Ags from tumor lysates and the subsequent induction of tumor immunity.
B7-DC XAb leads to multimolecular cap on the surface of the mDC
Because B7-DC has a short cytoplasmic tail and is thereby lacking inherent signaling capability, we hypothesize that B7-DC XAb may be able to cause topological changes or clustering of membrane molecules that could recruit and activate DAP12/Syk pathways in mDC. To test this ability, we looked for changes in membrane localization of key molecules involved in the mDC stimulation of naive T cells. We found that class II, CD80, and CD86 molecules were reorganized into a distinct cap-like cluster on the cell membranes of both mouse and human mDC within 15 min after Ab treatment (Fig. 4, A and B). The relationship between co-capped molecules and the kinetics of cap formation were investigated further using FRET. Within 10 min of B7-DC XAb treatment, class II, CD80, and CD86 molecules moved into close juxtaposition (<80 Å) as judged by the induction of a strong FRET signal (Fig. 5, A and B). However, no FRET signal could be detected when mDC were incubated with a different IgM Ab that binds to mouse class II molecules (Fig. 5,C). Moreover, if the acceptor fluorophore was quenched as a result of this photobleaching, emission of donor fluorophore increased, further confirming FRET and the implied proximity of the rearranged class II and CD80/CD86 molecules on the cell surface (data not shown). The ability of B7-DC XAb to form capped structures on the surface of human and mouse mDC was abolished by blockade of B7-DC XAb with the B7-DC-specific IgG Ab MIH18 on human mDC (Fig. 5,D, top row) and by the IgG Ab TY-25 on mouse mDC (Fig. 5,D, bottom row). These observations are consistent with our previous findings that the ability of B7-DC XAb to induce functional changes in mDC is dependent on direct binding to B7-DC (11, 12). To test whether the complex was internalized, cells were stripped of the bound fluorophore-labeled Abs using Hanks buffer (pH 2.5) and subsequently analyzed for FRET signal. Signal was lost when cells were stripped at 15 min, though a weak signal was detected if the cells were stripped at 30 min, suggesting internalization of some of the complex (data not shown). The independent recruitment of CD80, CD86, and class II molecules into the cap was investigated using mDC derived from CD80/CD86 double knockout or MHC class II knockout mice. Following treatment with B7-DC XAb, co-capping (FRET) of CD11c with class II still occurred in CD80−/−CD86−/− mDC and co-capping of CD80 with CD86 still occurred in class II−/− mDC (Fig. 5 E), indicating that B7-DC XAb can induce clustering of these particular molecules independently of one another.
TREM-2 mediates activation of a DAP12-Syk signaling pathway and is required for Ag accumulation by mature mDC treated with B7-DC XAb
TREM-2, a recently described pattern recognition receptor expressed on monocytes and cultured DC, is known to signal through DAP12 (39, 40). As visualized by FRET, an association of TREM-2 with class II molecules was observed within 5 min and was increased in 15 min after treatment of mDC with B7-DC XAb (Fig. 6,A). In contrast, TREM-2 was not associated with class II molecules on mDC following treatment with isotype control Ab. The association of TREM-2 with class II molecules was confirmed by coimmunoprecipitation of TREM-2 with the class II molecule I-Ab in lysates isolated from mDC 5 min after B7-DC XAb treatment (Fig. 6 B).
To evaluate the functional importance of TREM-2 in mDC activated with B7-DC XAb, we assessed the activation of Ag uptake and retention by matured TREM-2 deficient mDC using a knockdown strategy. Transduction of a retrovirus containing shDNA for TREM-2 into mouse mDC substantially reduced the expression of TREM-2 on the cell surface (Fig. 6,C). This reduction was associated with the absence of the phosphorylation of DAP12 and Syk (Fig. 6, D and E). Furthermore, using matured mDC, the shDNA-transduced cells accumulated very little OVA when treated with B7-DC XAb (Fig. 6,F, right). When mDC were transduced with virus containing a scrambled shDNA sequence, expression of TREM-2 (Fig. 6,C), phosphorylation of DAP12, and Syk (Fig. 6, D and E), and accumulation of OVA (Fig. 6 F, left) were similar to that observed for nontransduced mDC (13).
The important contribution of TREM-2 in the transduction of B7-DC XAb-induced signals was further confirmed using mDC derived from TREM-2 knockout mice. Although phosphorylation of DAP12 and Syk was readily induced in wild-type mDC activated with the MTAb B7-DC XAb, phosphorylation of these same signaling intermediates was not observed in TREM-2 knockout mDC (Fig. 7, A and B), and the matured knockout mDC did not regain the ability to accumulate OVA after Ab treatment (Fig. 7 C). Furthermore, whereas wild-type mDC pulsed with B16 tumor lysate and B7-DC XAb functioned efficiently as an antitumor vaccine (fully protecting 5/5 animals from B16 melanoma challenge), animals receiving a TREM-2 knockout mDC vaccine succumbed to the melanoma (5/5 animals; p < 0.05).
The ability of B7-DC XAb to induce a spectrum of biologic changes in mDC is consistent with the hypothesis that B7-DC may function as a signaling receptor on DC. Others have provided evidence for reverse signaling through a number of B7 family members, including CD80 and CD86 in response to CTLA-4-Ig or CD28-Ig ligands and B7-H1 in response to autoantibodies (20, 21, 41). Ligation of CD80 and CD86 by CTLA-4-Ig induces activation of NF-κB, p38 MAPKs, and STAT1, leading to the secretion of IFN-γ and the suppression of T cell proliferation (20). In contrast, CD28-Ig has been shown to activate NF-κB and MAPK pathways and the production of IL-6 and IFN-γ by DC, supporting potent T cell responses (21). In this study, we have used both human and mouse mDC to show that cross-linking B7-DC with the IgM Ab B7-DC XAb results in a reorganization of cell surface molecules, the activation of intracellular kinases, and the induction of Ag uptake and retention by mature mDC. These signaling events require TREM-2 and are mediated by the activation of Syk. Furthermore, both TREM-2 and Syk are required for B7-DC XAb-induced tumor immunity in mice. Although our data show that CD80 and CD86 are present in the molecular caps formed by B7-DC cross-linking, their role in Syk activation is unknown. Our previous studies showed that CD80 and CD86 are not required for Th1-polarization of recall responses induced by B7-DC XAb (12). We have also documented the expression of an array of immune mediators resulting from cross-linking B7-DC on mDC (26, 30) with a role in B7-DC XAb action are not yet clear. A key conclusion from all these experiments is that engagement of B7 family members with different ligands can induce a spectrum of immunomodulatory signals that can either up-regulate or down-regulate the immune response.
Engagement of B7-DC with B7-DC XAb results in the formation of multimolecular clusters including cell surface molecules involved in the generation of a productive T cell response (CD80, CD86, class II, and TREM-2). The ability of B7-DC XAb to induce these structures on the surface of cells is in contrast to the inability of an anti-class II IgM to induce comparable molecular clusters. The fact that bivalent Abs specific for B7-DC block membrane rearrangements induced by B7-DC XAb on both human and mouse mDC demonstrates that the induction of molecular complexes is mediated through B7-DC in both species.
The structures of human and mouse B7-DC differ with respect to the size of their cytoplasmic tails (9, 10). Although human B7-DC has a tail of 30 aa that could be a potential site for interactions with intracellular signaling molecules, mouse B7-DC is missing most of this structure, having only a few amino acids on the cytoplasmic side of the cell membrane. The B7-DC transmembrane domain is unremarkable in both species and contains no charged residues. Similar to all members of the B7 family, the extracellular domains of B7-DC are predicted to fold into two Ig-like domains. Molecules with these folds have a tendency to form heterodimers: CD8, Igs, TCR, and more recently B7 family members being notable examples (42). Whether the formation of multimolecular structures featuring the close juxtaposition of molecules such as class II, CD80, CD86, CD11c with B7-DC is critical for the generation of intracellular signals or is the consequence of signals generated by other molecular interactions remains to be tested. Nonetheless, the formation of molecular clusters containing TREM-2 in response to B7-DC XAb treatment demonstrates the principle that molecules other than B7-DC with intrinsic signaling capability may be recruited in this process.
One unanswered question is whether activation of mDC with B7-DC XAb represents a normal physiological process. It seems unlikely that cross-linking B7-DC in isolation is a means for intercellular communication. However, it is possible that reorganization of the DC membrane during interactions with the T cell surface induces intracellular signals in the DC also involving B7-DC. What is interesting is that Ab induced cross-linking of B7-DC is a stimulus that has distinct physiological consequences that might be exploited in the treatment or prevention of disease. The potential importance of this novel DC activation pathway is illustrated by our findings that B7-DC XAb has potent immunomodulatory effects in vitro and protects mice from allergic airway inflammation or lethal melanoma challenge (26, 30, 38).
Thus, with this report, we have new understanding of the molecular mechanisms linking B7-DC XAb binding on mDC, the recruitment of TREM-2 into membrane clusters, the activation of Syk to an enhanced uptake and retention of Ag in mature mDC, and the induction of tumor immunity in mice. Our interpretation is that DC residing in tumors or draining lymph nodes are activated by cross-linking B7-DC and that the induced changes enhance their ability to capture tumor Ags and mobilize a protective cellular immune response against developing melanoma nodules (see also11, 14, 30, 38). Moreover, changes in gene expression induced by B7-DC cross-linking of resident DC alters the polarity of the immune response against allergenic Ag introduced into the airways in a mouse model of allergic asthma (26). The observations documented in this study demonstrate that this human IgM activates signaling pathways in human DC that closely mirror activated pathways in the mouse, providing a basis for our hypothesis that B7-DC XAb will be a potent modulator of human immune responses and the rationale for developing this reagent to modulate DC function to treat human disease.
We thank Bogoljub Ciric, Virginia P. Van Keulen, Dianne Khurana, and Kristina Bruns for technical support and help from Timothy Kottke and Richard Vile in establishing viral titers.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Grants R01 CA104996-4 (to L.R.P.), R01 HL077296-3 (to L.R.P.), and R01 CA96859 (to L.R.P.) from the National Institutes of Health and by a grant from the Ralph Wilson Medical Research Foundation.
Abbreviations used in this paper: DC, dendritic cell; mDC, myeloid DC; poly(I:C), polyinosinic-polycytidylic acid; XAb, cross-linking Ab; PLC, phospholipase C; shRNA, short hairpin RNA.