Th2 Differentiation Is Unaffected by Jagged2 Expression on Dendritic Cells¹

Polarization of Th cell response is greatly influenced by signals delivered by dendritic cells (DCs)⁵ and other APCs (1, 2). These cells use pattern recognition receptors to identify and respond to specific pathogens, and in so doing can produce a series of positive and negative signals that facilitate the differentiation of naive precursor Th cells into distinct lineages such as Th1, Th2, Th17, and regulatory T cells (3, 4). Although the strength of the activation signal delivered to Th cells and the cytokines produced by DCs as they activate Th cells have been seen as major factors influencing Th cell response polarization (1), recent work has additionally implicated Notch signaling in this process (5, 6, 7, 8, 9, 10). Thus MyD88-dependent signals associated with DC maturation and Th1 response induction promote expression of the Notch ligands Delta and Jagged1, whereas cholera toxin (CT) and PGE, which signal via G protein-coupled receptors and condition DCs to drive Th2 responses, stimulate expression of Jagged2 (6). Functional roles for these differentially expressed Notch ligands were implied by in vitro studies in which Delta or Jagged1 overexpression in MHC class II expressing fibroblasts were shown to promote Th1 or Th2 differentiation, respectively (6). Recent reports have reiterated roles for Notch in Th1 (7, 11) and Th2 (5, 6, 12) cell differentiation. The possibility that differential Notch ligand engagement is controlling disparate outcomes downstream of a common Notch receptor remains viable. However, although published reports implicate preferential expression of Jagged2 by DCs as being important for the induction of Th2 responses, data illustrating this relationship are lacking and expectations of the roles of Jagged2, if any, in such a process are unclear.

To examine whether Jagged2 expression is linked to the ability of DCs to drive Th2 responses, we have developed a system in which Jagged2 expression in DCs can be modulated through transduction with retrovirus-encoded Jagged2 cDNA or hairpin RNA targeting Jagged2. We show that resting bone marrow-derived DCs, which in our hands can induce Th2 differentiation in vitro, express Jagged2. We also show that expression of Jagged2 is increased by exposure to soluble Schistosoma mansoni egg Ag (SEA), an extract of the eggs of S. mansoni, which has inherent adjuvanticity and conditions DCs to promote Th2 responses (13). Moreover, we have found that Jagged2 expression is markedly suppressed by TLR signaling in DCs, a process that promotes transcriptional responses that favor the induction of Th1 responses (14). Nevertheless, suppression of the expression of Jagged2 has no discernible effect on the ability of resting or SEA-pulsed DCs to drive Th2 responses, and forced expression of Jagged2 in TLR-stimulated DCs does not allow them to induce Th2 responses. Our data indicate that Jagged2 expression by DCs is dispensable for Th2 cell differentiation.

C57BL/6 (B6), BALB/c, and B6 IL-12p40-deficient mice were purchased from The Jackson Laboratory. BALB/c 4get mice were a gift from M. Mohrs (Trudeau Institute, Saranac Lake, NY) and were crossed to DO11.10 mice. Bone marrow from MyD88-deficient mice was obtained from L. Turka (University of Pennsylvania, Philadelphia, PA). All mice were housed according to animal care guidelines at the University of Pennsylvania. Conjugated Abs used to detect CD4, Thy1.1, CD11c, CD40, B7-2, IFN-γ, IL-2, or IL-4 in FACS analyses were purchased from BD Pharmingen. Goat anti-Jagged2 Ab was purchased from Santa Cruz Biotechnology (catalog no. SC-8158) and detected using a FITC-labeled anti-goat Ab (Jackson ImmunoResearch Laboratories). Listeriolysin O (LLO)_190–201 (NEKYAQAYPNVS) was synthesized by Invitrogen.

Bone marrow-derived DCs were generated as described (15, 16). Briefly, bone marrow DC precursors were differentiated in the presence of GM-CSF (20 ng/ml) in complete DC medium (RPMI 1640 containing 10% FCS, 100 U/ml penicillin/streptomycin, 0.05 mM 2-ME, and 2 mM l-glutamine). On days 9–10 of culture, DCs were washed in complete DC medium containing 5 ng/ml GM-CSF and pulsed for 18 h with heat-killed Propionibacterium acnes (10 μg/ml; Van Kampen Group), SEA (50 μg/ml; prepared aseptically as described in Ref. 17), CpG oligodeoxynucleotide 1826 (1 μg/ml; Coley Pharmaceutical Group). Polyinosinic-polycytidylic acid (1 μg/ml; Sigma Aldrich), peptidoglycan (1 μg/ml; Sigma Aldrich), LPS (100 ng/ml, Escherichia coli serotype O111:B4; Sigma-Aldrich), cholera toxin (1 μg/ml; Calbiochem), or PGE₂ (10⁻⁶ M; Sigma-Aldrich). For T cell assays, DCs were copulsed with OVA (200 μg/ml; prepared as described in Ref. 18), OVA peptide over a range of concentrations, or LLO_190–201 (10 μg/ml).

Jagged2 cDNA, a gift from Y.-Y. Kong (Pohang University of Science and Technology, Kyungbuk, South Korea), was cloned upstream of the internal ribosomal entry site sequence (IRES) in the MSCV-ires-Thy1.1 (MIT) retroviral vector, provided by W. Sha (University of California, San Francisco, CA). A PCR-based strategy was used to generate a short hairpin RNA to target Jagged2. Using pEF6-hU6 as a template, primers containing the hairpin sequence were used to amplify the hairpin downstream of the U6 promoter, as described (19). The resulting PCR product was TOPO-TA cloned. The inserts in the correct orientation were excised with BamH1 and EcoRV and ligated to the BglII and HincIII sites of MIT. The target sequence of the Jagged2 hairpin sequence is 5′-ggcaactccttctacctgcc-3′ and the sense sequence was modified to 5′-ggcaactccttctacctgtt-3′. MSCV-CMV-Thy1 (MCT) was generated by replacing the IRES of MIT with a CMV promoter. The IRES was excised with HindII and NcoI, blunt-ended, and replaced with the CMV promoter from pCI. The U6/hairpin cassette was cloned into the Pme site of MCT. Virus was produced in 293T cells by cotransfection of the retroviral construct with a helper plasmid containing the gag, pol, and env genes. Bone marrow cells were removed for spin-infection (90 min, 2500 rpm, 30°C) 2 days after the cultures were initiated, at which time cell proliferation was maximal (data not shown). Viral supernatants were supplemented with 0.05 mM 2-ME, 2 mM l-glutamine, and polybrene (8 μg/ml). Following spin-infection, retrovirus-containing supernatants were replaced with complete DC medium containing GM-CSF (20 ng/ml). DCs were sorted on Thy1.1 either by FACS or by using Thy1.1-FITC staining followed by anti-FITC beads (Miltenyi Biotec).

RNA was isolated using either TRIzol or RNeasy kit (Qiagen) and treated with DNase (TURBO DNA-free; Ambion). For semiquantitative RT-PCR, the one-step Superscript III kit (Invitrogen) was used. Primer sequences are as follows: Jagged2 (forward) 5′-GGA TGG CCA CCG GGA TTG TA-3′, (reverse) 5′-GCA GGA GGA GGC GGT CGT GT-3′; and HPRT (forward) 5′-GTT GGA TAC AGG CCA GAC TTT GTT G-3′, (reverse) 5′-GAG GGT AGG CTG GCC TAT AGG CT-3′.

For real-time analysis of gene expression, cDNAs were synthesized using random hexamers and Superscript II polymerase. Real-time PCR analysis was performed using SYBR Green and run on an ABI7500 Fast Real-time PCR System. Relative expression of Jagged2 to HPRT was calculated using the threshold cycle expression fold value (2^−ΔΔCt) method. Dissociation curves were generated to verify the presence of a single amplicon. Real-time primers for Jagged2 were purchased from Qiagen. Primer sequences were HPRT (forward) 5′-CTC CGC CGG CTT CCT CCT CA-3′, (reverse) 5′-ACC TGG TTC ATC ATC GCT AAT C-3′.

CD4⁺ T cells were purified from OTII or 4get.DO11.10 mice by negative selection using magnetic sorting (Miltenyi Biotec). An anti-CD69-biotin Ab was included in the mixture to remove activated CD4 T cells. For proliferation studies, cells were labeled with 1 μM CFSE (Molecular Probes). A total of 2 × 10⁵ T cells were plated with 2 × 10⁴ DCs, which had been pulsed overnight with Ags or TLR ligands, as described for each experiment. Cells were analyzed on day 2 for surface or GFP expression by FACS analysis. For the analysis of cytokine production, cells were restimulated with PMA (100 ng/ml; Sigma-Aldrich) and ionomycin (500 ng/ml; Sigma-Aldrich) for 4 h with GolgiStop (BD Pharmingen) present during the last 2 h. Cells were fixed and stained for cytokines using the BD Cytofix/Cytoperm kit according to the manufacturer’s protocol. Data were collected using a BD Canto and LSRII and analyzed using FlowJo.

Following incubation with Ag, DCs were washed in PBS and 5 × 10⁵ DCs were i.p. injected per mouse. Seven days following immunization, splenocytes were harvested, resuspended at 15 × 10⁶/ml, and restimulated for 6 h with or without LLO₁₉₀ (1 μg/ml) in the presence of human IL-2 (50 U/ml; PeproTech) and GolgiStop in Iscove’s medium supplemented with 3% normal mouse serum, 100 U/ml penicillin/streptomycin, 0.05 mM 2-ME, and 2 mM l-glutamine. Cytokine production was determined using the BD Cytofix/Cytoperm kit. Data were collected using a BD Canto and LSRII and analyzed using FlowJo.

Notch signaling in T cells has been shown to be important for both the activation of naive T cells and their differentiation into the Th1 and Th2 subsets (5, 6, 7, 8, 9, 10, 11, 20, 21). DCs, which are potent activators of naive T cells and can direct their polarization, differentially express Notch ligands in response to pathogen exposure (2, 6, 22). There is evidence that this differential expression of Delta and Jagged plays a key role in directing Th cell differentiation (6). Helminth parasites, which induce strong Th2 responses, have been proposed to do so by up-regulating Jagged expression on DCs (23); however, data supporting this relationship are lacking. To determine whether Jagged1 and Jagged2 expression in DCs is controlled by pathogen-derived signals, DCs were pulsed for 18 h with SEA, heat-killed P. acnes, or defined TLR ligands (Fig. 1).

Gene expression analysis revealed that Jagged2 but not Jagged1 expression correlated with the ability of DCs to induce Th2 responses. Jagged1 and Jagged2 are both expressed in resting DCs. Jagged1 expression is moderately, but not consistently, enhanced by SEA stimulation, whereas Jagged1 expression is markedly induced by TLR ligation (Fig. 1, A and B). Jagged2 expression, however, was consistently induced by SEA stimulation and was notably down-regulated by TLR ligation by either defined TLR ligands or P. acnes (Fig. 1, A and B).

To determine whether Jagged2 was up-regulated specifically by SEA or by other agents that condition DCs to induce Th2 responses, we stimulated DCs with CT and PGE₂. Jagged2 expression was induced by both CT and PGE₂ (ranging from 4- to 10-fold), suggesting that expression of Jagged2 may be characteristic of DCs that induce Th2 responses. To examine Jagged2 protein expression on DCs in response to stimulation, DCs were stimulated with SEA, P. acnes, or LPS and surface-stained for Jagged2. FACS analysis revealed that Jagged2 expression is up-regulated by SEA and down-regulated by P. acnes and LPS, confirming the gene expression analyses (Fig. 1 D). Taken together, these data indicated that Jagged2 but not Jagged1 expression correlates with the ability of DCs to induce Th2 responses.

We examined the ability of Jagged2 to be regulated by multiple TLR ligands. We found that Jagged2 expression is down-regulated in response to both polyinosinic-polycytidylic acid, which is a ligand for the TLR3 MyD88-independent pathways, and peptidoglycan, which activates the TLR2 MyD88-dependent pathway, strongly indicating that either pathway can regulate Jagged2 expression (Fig. 1,E). This was confirmed by an analysis of TLR ligand-regulated Jagged2 expression in DCs from myd88^−/− mice. Jagged2 expression was down-regulated within 5 h by P. acnes and LPS comparably in myd88^+/+ and myd88^−/− cells, confirming that Jagged2 expression can be regulated by MyD88-independent pathways (Fig. 1 F). In contrast, the ability of CpG to induce down-regulation of expression of Jagged2 was apparent in myd88^+/+ but not myd88^−/− mice, consistent with the fact that TLR9 signaling is entirely MyD88-dependent (14). Additionally, basal levels of Jagged2 expression were higher in MyD88-deficient DCs, suggesting that tonic MyD88-dependent signaling may down-regulate Jagged2 expression.

Jagged2 expression correlates with the ability of DCs to induce Th2 responses. However, it remains to be determined whether Jagged2 is responsible for the induction of the Notch signaling, which leads a T cell to adopt a Th2 phenotype. To determine whether Jagged2 up-regulation in response to Th2-inducing Ag (Fig. 1) confers the ability of these DCs to induce Th2 responses, the expression of Jagged2 was manipulated by targeted knockdown or forced expression in bone marrow-derived DCs. An MSCV-based retroviral vector expressing Thy1.1 as a reporter was constructed to express a hairpin sequence targeting Jagged2 mRNA (J2hp) or the Jagged2 cDNA. Because DCs are generally activated by viruses, bone marrow cultures were transduced 2 days after seeding, at which point the DC precursors were in the process of proliferating. When harvested 6 days later, transduced cells were easily sorted on the basis of Thy1.1 expression (Fig. 2, A–C). As monitored by costimulatory molecule up-regulation, transduction with retrovirus encoding either Jagged2 or J2hp did not result in activation of the DCs (Fig. 2, B and C). Transduced cells remained responsive to stimulation, up-regulating both CD40 and CD86 in response to P. acnes (Fig. 2, B and C). As anticipated from previous work, SEA did not induce costimulatory molecule up-regulation in DCs (Fig. 2,B). Cells expressing the J2hp showed reduced levels of Jagged2 mRNA (Fig. 2,D and data not shown), indicating that Jagged2 expression was successfully knocked down using this methodology. Importantly, Jagged1 mRNA levels were unaffected by J2hp (Fig. 2,D). Use of a separate retrovirus MCT expressing J2hp gave similar results (data not shown). Overexpression and knockdown of Jagged2 was confirmed by real-time RT-PCR analysis (Fig. 2,E and data not shown) and by FACS analysis (Fig. 2 F).

Many studies have revealed that Notch ligands can direct the differentiation of DCs, and hematopoietic cells in general (24, 25). We found that retroviral transduction of hematopoietic stem cell cultures, in the presence of GM-CSF, did not alter the differentiation of the DCs. Transduced DC cultures remained comparable to untransduced cultures and were CD11c^high, PDCA1⁻, CD11b⁺, Gr1⁻, CD8⁻, B220⁻, and expressed normal baseline levels of CD40, CD86, CD80, and MHC class II, demonstrating that the lineage of these cells had not been altered by manipulating Jagged expression (data not shown). Furthermore, the percentage of Gr1⁺ or CD11c⁻ cells in the cultures was comparable.

We first used an in vitro culture system to examine the effect of Jagged2 on Th cell differentiation. DCs copulsed with SEA and OVA were used to direct the differentiation of OTII CD4⁺ T cells to the Th2 lineage. Five days following stimulation, the cultures were restimulated and examined for cytokine production. Under these conditions, when DCs are transduced with control retrovirus MIT, >35% of responding CD4⁺ T cells commit to the Th2 lineage and become capable of making IL-4 (Fig. 3,A). Overexpression of Jagged2 in DCs had no consistent effect on this outcome (Fig. 3,A). Moreover, suppression of Jagged2 expression in DCs did not inhibit their ability to induce Th2 responses (Fig. 3,A). In these experiments, we also examined the ability of Th cells to make IL-2 because IL-2 plays a major role in Th2 commitment (26). We observed no effects of the retroviral expression or knockdown of Jagged2 in DCs on the ability of the Th cells they had activated to make IL-2 (Fig. 2,A). In all of these experiments, Th cells were labeled with CFSE before activation allowing us to monitor their proliferative responses. We noted no effects of Jagged on proliferation in these assays (Fig. 2 A).

Early IL-4 production by Th cells is indicative of their potential to commit to the Th2 lineage (27). To examine whether expression of Jagged2 by DCs has any effect on the production of IL-4 by Th cells early following activation, we used 4get.DO11.10 T cells, which express GFP under the control of the IL-4 promoter (28), to examine early IL-4 production. In these experiments, IL-12-deficient DCs were used, as Th2 responses are more readily induced in the absence of this cytokine, and early IL-4 production is more easily measured. Jagged2 expressing IL-12p40-deficient DCs were pulsed with OVA and cocultured with sorted 4get.DO11.10 CD4⁺ T cells. Two days following stimulation, CD4⁺ cells were examined for activation and GFP expression as a sensitive indicator of IL-4 production. We noted no effect of induced Jagged2 expression, or of Jagged2 knockdown in DCs, on the ability of Th cells to become activated (measured by CD25 expression), to make early IL-4 by day 2, or to commit to Th2 cell differentiation over the course of 5 days (Fig. 4). Furthermore, Jagged2 expression did not influence Th2 differentiation in response to varying doses of Ag (data not shown).

Finally, we asked whether Jagged2 expression influences Th2 cell development in vivo. For these experiments, mice were immunized with untransduced or transduced DCs copulsed with SEA and the IA^b-restricted peptide LLO₁₉₀. Seven days following immunization, spleens were harvested, and CD4⁺ cells were examined by intracellular staining (ICS) and flow cytometry for their ability to produce IL-4 in response to restimulation with LLO₁₉₀ (Fig. 3,C). A clear population of LLO₁₉₀-specific Th2 cells, absent in mice immunized with unpulsed DCs, was induced by this immunization protocol. In these assays, commitment of the LLO₁₉₀ response to the Th2 lineage was indistinguishable in mice immunized with control DCs, DCs transduced with control retroviruses, DCs expressing Jagged2, and DCs expressing J2hp (Fig. 3,C). CT, which is known to promote both Th1 and Th2 responses, induces a greater increase in Jagged2 expression than SEA. To examine whether Jagged2 plays a role in CT–induced Th cell responses, mice were immunized with DCs copulsed with CT and LLO₁₉₀. Seven days following immunization, there were an equal number of LLO₁₉₀-specific Th1 and Th2 cells (Fig. 3 D) in spleens from mice immunized with DCs transduced with control retrovirus and DCs expressing J2hp. Cumulatively, our data indicate that the expression of Jagged2 by DCs during Th cell activation is not essential for Th2 response development.

In an alternative approach to testing the role of Jagged2 in Th2 response induction, we asked whether forcing expression of Jagged2 in DCs that have down-regulated endogenous Jagged2 expression in response to TLR signaling could allow them to instruct Th cells to make IL-4. Under normal conditions, TLR-stimulated DCs powerfully suppress Th2 development (18).

DCs pulsed with TLR ligands make IL-12, which is a potent inducer of Th1 responses, and can suppress Th2 development, apparently even development resulting from overexpression in T cells of the Notch1 intracellular domain (6). This response suggests that IL-12 is dominant over Notch-related Th2-promoting signals. We used IL-12p40-deficient DCs for these experiments to exclude the possibility that IL-12 was simply overriding the ability of Jagged2 to induce IL-4 production. Transduced DCs were pulsed with OVA and P. acnes and cocultured with 4get.DO11.10 CD4⁺ cells. We found that forced Jagged2 expression failed to promote IL-4 production (Fig. 4). We also noted that in this setting, forced Jagged2 expression did not inhibit the production of IFN-γ by Th cells responding to TLR-stimulated DCs (Fig. 4).

We have found that the expression of Jagged2 correlates with the ability of DCs to induce Th2 responses. Specifically, Jagged2 expression is markedly down-regulated by exposure of DCs to TLR agonists and is increased when DCs are stimulated with SEA, CT, and PGE₂. Moreover, MyD88-deficient DCs have been reported to be capable of inducing Th2 responses (6). Interestingly, expression of Jagged2 is increased in these cells. Nevertheless, our data do not support the conjecture (23) that expression of Jagged proteins underlies the ability of worm Ag-conditioned DCs to promote Th2 responses because neither increasing nor decreasing Jagged2 expression had any measurable effect on DCs ability to induce Th2 responses.

Jagged1 expression has been shown to increase following TLR activation (6). However, combinations of TLR ligands, which act in synergy to modulate gene expression and promote Th1 responses, down-regulate Jagged1 expression (22) leaving open the possibility that Jagged1 is involved in Th2 response induction. Although induced expression of Jagged1 failed to significantly enhance Th2 cell differentiation in our system (data not shown), it remains possible that in its absence Th2 differentiation may be reduced. Further studies using Jagged1 or Jagged2 double-deficient DCs would be needed to address this possibility.

Notch signaling in T cells is clearly important for the development of both Th1 and Th2 responses (5, 6, 7, 8, 9, 10, 11, 12). The ability of Notch to differentially direct Th responses has been proposed to be due to differential expression of Notch ligands on Th1- and Th2-promoting DCs. Delta overexpression has been reported to increase IFN-γ production and promote Th1 differentiation (6, 11). Additionally, gamma-secretase inhibitors prevent the development of Th1 responses, indicating that Notch signaling is important for the development of Th1 cells (7). However, Notch signaling through RBPJk (recombination signal binding protein for Ig κ J region) and MAML (mastermind-like 1) pathways is not essential for Th1 responses (5), indicating that Delta-regulated pathways independent of RBPJk and MAML can contribute to the differentiation of Th1 cells. In T cells, RBPJk and MAML have been shown to be important for the development of Th2 responses. However, Notch ligand involvement in this process is not clear. Specifically, Notch intracellular domain functions in part to activate RBPJk, which activates GATA-3 and IL-4 transcription leading to the development of Th2 cells (6, 29). In addition, mice expressing a dominant negative form of MAML (an activator downstream of Notch) cannot generate functional Th2 responses, but can mount Th1 responses, further supporting the notion that Notch signaling regulates IL-4 production (5). Whether Notch ligands engage separate Notch receptors or differentially induce signaling pathway downstream of Notch is of central importance to understanding how Notch ligands influence Th cell differentiation. Further work elucidating the relationship between Notch ligand expression on DCs and Notch activation in T cells will provide insight into the mechanisms by which DCs direct Th cell differentiation.

Our failure to detect a role for induced Jagged2 expression on DCs in the promotion of Th2 responses by SEA suggests that Jagged2 does not provide a Th2-instructive signal to naive Th cells. As put forward by Amsen and colleagues (6) and expanded on by Lehar and Bevan (23), the idea that differential-induced Notch ligand expression underlies the abilities of different types of pathogens to preferentially induce Th1 or Th2 responses is very appealing, not least because it provides a candidate for a positive signal, deliverable by DCs, that can promote Th2 responses. This model allowed a move away from the inherently unsatisfying “default hypothesis,” which proposes that the absence of induced IL-12 production underlies the ability of DCs to induce Th2 responses (30). Although it remains feasible that Jagged2 is important for Th2 response induction in specific settings, our inability to define a role for it in a variety of in vitro and in vivo systems suggests that it is neither required or sufficient for the development of Th2 responses, and is therefore not a key signal delivered by DCs to naive Th cells to begin them on the road to becoming Th2 cells.

We thank R. Cinalli, R. G. Jones, C. B. Thompson, Y.-Y. Kong, W. Sha, L. Turka, P. Bates, and H. Shen for reagents. We thank E. L. Pearce, T. Freitas, J. Northrop, J. DiSpirito, F. Marshall, and J. Taylor for critical comments and E. Jung for excellent technical assistance.

The authors have no financial conflict of interest.