Helminth-Conditioned Dendritic Cells Prime CD4⁺ T Cells to IL-4 Production In Vivo

Allergic and Th2 immune responses are induced by exposure to allergens, parasites, and adjuvants such as aluminum hydroxide. They involve the production of the cytokines IL-4, IL-5, and IL-13, elevated eosinophil numbers, and IgE production (1, 2). The precise conditions leading to the initiation of these responses are, however, controversial and intensely researched.

Allergens and parasites mostly lack typical microbial- and danger-associated molecular patterns, thus failing to induce the strong dendritic cell (DC) activation usually observed in bacterial or viral infections (3). For this reason, the initiation of Th2 immune responses in vivo was proposed to involve APC populations that are alternative or additional to DC and that would provide the specific signals required for Th2 activation (4–8). Ultimately, studies using in vivo depletion of CD11c⁺ cells and infection with the helminths Schistosoma mansoni, Heligmosomoides polygyrus, and Nippostrongylus brasiliensis (9–11), or following exposure to house dust mite (HDM) or OVA and papain (7, 12), confirmed that DC are required for the development of innate and adaptive Th2 responses. Therefore, these studies established that DC are necessary, but not necessarily sufficient, for the initiation of Th2 responses in vivo.

The main candidate accessory signal for the induction of Th2 responses by DC is the costimulatory molecule OX40 ligand (OX40L) (13–15). OX40L expression on DC is induced by innate cytokines, including thymic stromal lymphopoietin (TSLP), IL-25, and IL-33. Blockade of OX40L results in markedly diminished allergic inflammation in models of allergic airway, skin, and gastrointestinal disease (13, 16, 17); however, OX40/OX40L deficiency does not block IL-4 production during parasite infections (18, 19). In addition, it is not yet firmly established whether OX40–OX40L is a true Th2-polarizing signal or whether its function is mainly costimulatory (20, 21). Other studies have reported that DC may require signals from B cells (22), basophils (7), or innate lymphoid cells (23) to prime Th2 immune responses to intestinal parasites or allergens, respectively. Together, these studies suggest that, with the appropriate conditioning, DC may acquire the accessory signals that direct naive CD4⁺ T cells to a Th2 phenotype. Again, these studies do not establish whether DC are sufficient for the priming of Th2 cells in vivo.

In addition to the differential expression of accessory signals as possible drivers of Th2 induction, distinct DC populations may exist that specialize in inducing Th2 immune responses. Three recent studies have described a novel Th2-promoting DC population in skin-draining lymph nodes (dLN), which express the cell-surface markers CD301b (MGL2) and programmed death ligand 2 (PDL2) and are distinct from CD103⁺ dermal DC (d-DC) and Langerhans cells (24–26). Interestingly, although CD301b⁺ DC were required for induction of Th2 immune responses in vivo, they were unable to promote Th2 differentiation by naive T cells in vitro and in vivo, suggesting a requirement for additional cell populations. These results differ from studies in which CD11b⁺ DC from the lungs of HDM-treated mice were sufficient to induce Th2 priming and sensitize to airway challenge with HDM, whereas transfer of PBS-treated DC was ineffective (27, 28). Several reasons could explain the discrepancy between these results, including the use of different Th2 immune response models, the different DC subsets involved, and the different degree to which Ag carried over with the DC could have contributed to the induction of the T cell responses being measured.

Infection with the mouse-adapted parasite nonviable N. brasiliensis represents an ideal model for studying Th2 priming in vivo. N. brasiliensis elicits strong Th2 immune responses in lung and gut (29), and administration of nonviable N. brasiliensis L3 larvae (Nb) into ear skin triggers a strong and highly polarized Th2 response in the ear-dLN (30). In this paper, we combine the powerful Nb model of Th2 immune response with a highly sensitive readout of IL-4 production in vivo to investigate the induction of Th2 responses by DC. We show that Nb-conditioned DC were able to prime CD4⁺ T cells of unrelated Ag specificity to IL-4 production in vivo and that this property was restricted to skin-derived DC, as LN-resident DC from Nb-treated mice, and DC from untreated mice, could support CD4⁺ T cell expansion but were unable to prime for IL-4 production in vivo. Thus, the ability of DC to prime IL-4 responses is independent of Ag, but requires prior exposure to the appropriate conditioning environment in vivo.

Specific pathogen-free mice were bred at the Malaghan Institute of Medical Research. C57BL/6 mice were originally from The Jackson Laboratory (Bar Harbor, ME). CD45-congenic B6.SJL-Ptprc^a mice were from the Animal Resources Centre (Perth, Australia). OTII mice carrying a transgenic TCR specific for IA^b and OVA_323–339 were obtained from Melbourne University. Heterozygous G4/IL-4 mice that are IL-4 sufficient and express GFP as a reporter for IL-4 (31), and TSLP receptor^−/− (32) G4/G4 mice (TSLPR^−/− G4/G4) were kindly donated by Dr. William Paul (National Institute of Allergy and Infectious Diseases, National Institutes of Health). OTII mice were crossed to G4 mice to obtain OTII G4/G4 mice. All experimental protocols were approved by the Victoria University Animal Ethics Committee and performed according to institutional guidelines.

N. brasiliensis infective L3 larvae were prepared, washed in sterile PBS, killed by three freeze-thaw cycles, and injected into the ear pinna of anesthetized mice as described (30). For fluorescent labeling, nonviable L3 were incubated in 0.05 M NaHCO₃ buffer and 0.1 mg Alexa Fluor 488 (AF488) Microscale Protein labeling dye (Molecular Probes, Invitrogen) and then washed with 0.1 M Tris buffer. When indicated, 0.5 μg pertussis toxin (Ptx; Sigma-Aldrich) was added to the L3 preparation prior to injection. Heat-killed Mycobacterium smegmatis (MC2155) was injected intradermally (i.d.) at 2 × 10⁶ CFU/mouse.

To block OX40L in vivo, recipient mice were injected i.p. with 300 μg LEAF-purified anti-mouse CD252 (RM134L/rat IgG2b/κ) (BioLegend, San Diego, CA) 1 d before and 1 d following Nb challenge.

The concentration of IL-33 in ear lysates was determined 6 and 24 h after intradermal treatments. Mice were euthanized, and ears were removed and placed immediately on dry ice. Ear tissue was homogenized in Halt Protease Inhibitor Cocktail (Thermo Scientific Pierce) using a Polytron Rotor Stator Homogenizer (Fisher Scientific), and the aqueous layer was separated by centrifugation. Cytokine was measured using a sandwich ELISA (eBioscience) according to the manufacturer’s instructions and read at 450 nm on a tunable VERSAmax microplate reader (Molecular Devices, Victoria, Australia).

Single-cell suspensions were prepared from both auricular LN from 30–40 donor mice per group and digested using DNAse I and Liberase TL (Roche) as described (33). CD11c⁺ cells were enriched by positive magnetic selection (Miltenyi Biotec). Enriched cells were flow sorted into CD11c^hi MHC class II (MHCII)^int and CD11c^int MHCII^hi populations using anti-CD11c (BD Pharmingen) and anti–MHCII, clone 3JP (made in-house) mAb, and a FACSVantage Diva (BD Biosciences, San Jose, CA) with BD FACSDiva software, version 5.0.3 (BD Biosciences).

DC populations were incubated with 10 μM OVA_323–339 peptide (Mimotopes, Victoria, Australia) for 1 h at 37°C, washed to remove excess Ag, and 1 × 10⁵ cells was injected i.d. into the ear pinna in 30 μl.

OTII G4 cells were prepared from naive spleens using a Dynabeads CD4 Positive Isolation kit (Invitrogen). A total of 2 × 10⁶ cells was injected i.v. into B6.SJL-Ptprc^a recipients 1 d before DC transfer.

For staining of cell-surface molecules, cells were incubated with anti-mouse CD16/32 (clone 2.4G2, affinity purified from hybridoma culture supernatant) prior to labeling with cocktails of fluorescent Abs specific for: CD11c (HL3), CD11b (M1/70), CD86 (GL1), CXCR5 (2G8), CD4 (RM4-5), Ly6C (AL-21) (all from BD Pharmingen, San Jose, CA); CD103 (2E7), Ly6C (HK1.4), CD64 (X54-5/7.1) CD326 (G8.8), CD301b (MGL2, clone URA-1), CD252 (OX40L, clone RM134L) (all from BioLegend); CD24 (M1/69), CD172α (SIRPα; clone P84), CD273 (PDL2, clone TY25), CD45.1 (A20), CD45.2 (104), CD339 (Jagged-1; clone HMJ1-29), CD275 (ICOS ligand, clone HK5.3) (all from eBioscience, San Diego, CA); IL-33Rα (ST2, clone DJ8) (MD Bioproducts, Zurich, Switzerland); and MHCII (3JP, prepared in-house).

For intracellular cytokine staining, cells were surface stained, fixed, permeabilized using the BD Cytofix/Cytoperm kit, and labeled with anti–IFN-γ (XMG1.2) or isotype control (all from BD Pharmingen).

For IFN regulatory factor 4 (IRF4) intranuclear staining, cells were surface labeled before being fixed and permeabilized with Foxp3 Fixation/Permeabilization Concentrate and Diluent (eBioscience) and then labeled with anti-IRF4 (3E4) or control rat IgG1 Abs (Santa Cruz Biotechnology). Dead cells were identified and excluded from analysis using DAPI labeling or the LIVE/DEAD Fixable Blue dead cell stain kit (both from Molecular Probes, Invitrogen). Compensation was set in each experiment using BD CompBeads (BD Pharmingen). All samples were collected on a LSRII SORP flow cytometer (BD Biosciences) and analyzed using FlowJo version 9.6.2 (Tree Star).

Seven days following i.d. challenge, mice were sacrificed, and auricular LN were collected for analysis. To assess IL-4 production using IL-4–GFP reporter expression, live CD4⁺CD45.2⁺ T cells were identified by flow cytometry, and the percentage of IL-4–GFP⁺ cells was determined by comparison with an IL-4–GFP^neg C57BL/6 control. To assess other cytokines, LN cells were cultured in complete IMDM in the presence of 5 μg/ml anti-CD3 (clone 2C11), 2 μg/ml anti-CD28 (clone 37.51), and GolgiStop (BD Pharmingen) for 4 h and then fixed and permeabilized as described (33).

Statistical analyses were performed using Prism 5.0 (GraphPad). Mean ± SEM is shown in all graphs. Data were analyzed using one-way or two-way ANOVA with Bonferroni posttest. The p values <0.05 were considered significant.

To define the role of DC in the initiation of Th2 immune responses in the skin, we used injection of Nb given i.d. into the ear (30). Nb were labeled with the fluorescent dye AF488 (Nb-AF488), injected into the ear pinna, and total cells were isolated from the dLN at different times following injection. As shown in Fig. 1A, cells labeled with fluorescent parasite material could be detected in the dLN as early as day 1, peaked at day 2, and rapidly decreased by day 4 after injection. Total LN cellularity did not increase until day 4 after Nb injection, possibly suggesting that it was initiated by the entry of parasite material into the LN (Fig. 1A, right panel).

To establish whether entry of parasite material in the dLN involved active transport via a migratory cell population, Nb-AF488 was coadministered with Ptx, a G1 protein inhibitor that blocks chemokine receptor signaling and prevents leukocyte trafficking. As shown in Fig. 1B, treatment with Ptx significantly decreased, but did not eliminate, AF488⁺ cells in dLN. In addition, treatment with Ptx strongly decreased priming of IL-4–producing CD4⁺ T cells, measured as frequency and number of CD4⁺IL-4–GFP⁺ T cells in the dLN at day 7 after challenge (Fig. 1C, 1D). Thus, the priming of IL-4–producing CD4⁺ T cells requires the active transport of parasite material from the skin to the dLN by a migratory cell population.

We used flow cytometry to identify the AF488⁺ LN cell populations in Nb-AF488–treated mice. As shown in Fig. 2A, essentially all AF488⁺ cells were also CD11c⁺. Further characterization revealed that ∼45% of the AF488⁺ population in the dLN expressed high levels of MHCII and was mostly CD11b⁺Ly6C⁻, a phenotype characteristic of a subset of d-DC. The second population of AF488⁺ cells, also including ∼45% of the total AF488⁺ population, instead expressed intermediate levels of MHCII and was CD11b^hiLy6C⁺ (Fig. 2A). Together with expression of CD64 and FcεRI (not shown), these markers identify these DC as inflammatory monocyte-derived DC (mo-DC) (12).

The increase in numbers of AF488⁺Ly6C⁻ d-DC and Ly6C⁺ mo-DC in the dLN of Nb-AF488–treated mice showed similar kinetics and magnitude (Fig. 2B). Thus, either of these DC subsets might be involved in the priming of IL-4–producing T cells. To begin to separate the role of d-DC and mo-DC in the priming of IL-4–producing T cells, we examined the effect of Ptx treatment. We found that, compared with Nb-AF488 only, coadministration of Nb-AF488 and Ptx resulted in significantly decreased percentages and numbers of AF488⁺ d-DC in the dLN. In contrast, the AF488⁺ mo-DC subset was only partially affected (Fig. 2C). This finding aligns with the data in Fig. 2A and suggests that Nb-AF488 material in the dLN is associated with two distinct DC subsets, a migratory d-DC subset that presumably enters the LN from the skin and a mo-DC subset presumably originating from blood monocytes. Together with the results in Fig. 1D, this finding also suggests that the mo-DC subset is not sufficient for the priming of Th2 cells, whereas the d-DC subset is required for a response.

Further experiments were carried out to compare the phenotypes of MHCII^hi AF488⁺ and AF488⁻ DCs in Nb-AF488–treated and untreated mice (Fig. 3A). We found that the majority of AF488⁺ cells were CD11b⁺, with some CD11b^low cells also observed in some experiments. All of the AF488⁺ cells were reproducibly CD103⁻, CD326⁻, CD24^low, and SIRPα⁺ (Fig. 3B) and did not express CD4 or CD207 (not shown), confirming their putative identification as d-DC.

We also examined AF488⁺ DC for expression of costimulatory molecules and markers that have been associated with Th2 immune responses. As shown in Fig. 3B, AF488⁺ DC expressed higher levels of CD86 compared with the AF488⁻ DC population. Expression of OX40L was moderate but again higher on AF488⁺ DC compared with the AF488⁻. As OX40L is normally expressed at low levels (28), this low expression may nonetheless be functionally relevant. Additionally, AF488⁺ DC were CD301b⁺ and expressed high levels of PDL2, which are both markers recently associated with a population of Th2-promoting skin DC (24–26). IRF4 and CXCR5 were expressed regardless of Nb treatment. Expression of IL-33Rα and Jagged-1 was low in all subsets but marginally higher in AF488⁺ DC, and ICOS ligand was negative.

Lastly, as the uptake of Nb-AF488 by DC was variable, we gated the AF488⁺ DC into bright and dim populations (Fig. 3C). The AF488^bright DC were homogeneously positive for expression of CD11b, PDL2, and CD301b, whereas the AF488^dim and AF488⁻ DC expressed variable levels of these markers (Fig. 3D). This result suggests that Nb material may be preferentially taken up by a specific DC subset.

The innate cytokine TSLP is reported to exert Th2-promoting activities by upregulating OX40L expression on DC (14). To evaluate the contribution of TSLP in the response to Nb treatment, we examined Th2 induction in TSLPR knockout (KO) mice. The numbers and percentages of CD4⁺IL-4–GFP⁺ cells were similar in TSLPR wild-type and KO mice (Fig. 4A). To indirectly address the involvement of other innate cytokines that also mediate their effect via OX40L upregulation (13, 34), we carried out in vivo blocking experiments using a neutralizing anti-OX40L mAb. As shown in Fig. 4B, i.d. challenge with Nb resulted in increased frequencies and numbers of CD4⁺ IL-4–GFP⁺ cells in dLN. Treatment with anti-OX40L did not affect this response and had no significant effect on LN cellularity, suggesting that the observed low expression of OX40L on d-DC was not necessary for the response to Nb. Therefore, OX40L upregulation by innate cytokines such as TSLP is not required for the priming of IL-4–producing CD4⁺ cells after i.d. challenge with Nb.

To determine whether DC from Nb-challenged mice were solely capable of directing naive CD4⁺ T cells to Th2 differentiation in vivo, we assessed their ability to prime IL-4–producing T cells upon in vivo transfer. Total CD11c⁺ cells were prepared from the dLN of mice treated with Nb or PBS 2 d earlier; these preparations contained no detectable basophils (not shown). CD11c⁺ cells were transferred into naive, IL-4–sufficient heterozygous IL-4–GFP reporter mice, and the priming of T cell responses was examined in the dLN 7 d after DC transfer by assessing the frequency of CD4⁺ CD44^hi cells as a measure of the T cell response and the frequency of CD4⁺ IL-4–GFP⁺ T cells as a measure of Th2 priming. Representative flow plots for this analysis are shown in Fig. 5A. Compared to PBS-treated mice or mice receiving CD11c⁺ cells from PBS-treated donors, mice injected with CD11c⁺ cells from Nb-challenged donors showed increased frequency of both CD4⁺CD44^hi and CD4⁺IL-4–GFP⁺ cells in the dLN (Fig. 5B, 5C, respectively). In contrast, CD11c⁺ cells from mice treated with an agent that does not induce Th2 activation, M. smegmatis, failed to increase the frequency of CD4⁺ IL-4–GFP⁺ cells compared with DC from Nb-treated mice (Fig. 5D).

To determine whether transfer of Nb-conditioned DC into naive mice induced priming of IFN-γ–producing CD4⁺ T cells, we used intracellular cytokine staining on d 7 after DC transfer. As shown in Fig. 5E, the frequency of IFN-γ–producing cells in mice receiving Nb-conditioned DC or no DC was similarly low, indicating that DC transfer did not result in the activation of IFN-γ–producing CD4⁺ T cells.

The experiments in Fig. 5 compared DC loaded with parasite material to DC from untreated mice. To assess whether the differential ability of these DC populations to prime IL-4–producing cells in vivo was simply due to Ag carryover, we performed experiments in which purified DC were loaded with equal amounts of OVA_323–339 before in vivo transfer, as outlined in Fig. 6A. This experimental design enabled us to assess the expansion of OTII cells and their ability to express IL-4–GFP at the same time.

DC populations were isolated from the dLN of mice 2 d following i.d. Nb or PBS challenge, flow sorted into CD11c⁺MHCII^hi and CD11c⁺MHCII^int DC populations (Fig. 6B), and loaded with OVA_323–339 before in vivo transfer. As shown in Fig. 6C and 6D, all DC populations were able to induce expansion of OTII donor cells in the dLN. MHCII^hi DC appeared to induce a better expansion compared with the MHCII^int DC, but this difference was not statistically significant. No OTII cell expansion was observed in controls that did not receive DC.

We used the sensitivity of the IL-4–GFP reporter system to follow the frequency and numbers of IL-4–GFP⁺ OTII cells after DC transfer (Fig. 6C, 6E). IL-4–GFP⁺ OTII T cells were clearly detected in mice receiving MHCII^hi DC from Nb-treated mice, whereas MHCII^int DC, or MHCII^hi DC from PBS-treated mice, induced a negligible IL-4–GFP response. Together, these results suggest that Nb can instruct d-DC to prime CD4⁺ T cells to IL-4 production and that this property of DC is independent of the specificity of the responding CD4⁺ T cells.

In this manuscript, we show that DC from Nb-treated mice transferred into naive recipients were able to prime CD4⁺ T cells to IL-4 production, whereas similarly isolated DC from naive, PBS-treated, or M. smegmatis–treated mice were unable to do so. The Nb-conditioned DC’s ability to prime CD4⁺ T cells to IL-4 production was not due to a general, nonspecific improvement of the T cell stimulatory capacity of DC from Nb-treated mice compared with PBS-treated, as Nb-treated and PBS-treated DC were similar in their ability to induce expansion of OTII cells in vivo and were both unable to prime IFN-γ production. Thus, DC from Nb-treated mice had acquired the selective ability to prime CD4⁺ T cells to IL-4 production in vivo.

Our experiments used the expression of an IL-4–GFP reporter to detect production of the hallmark cytokine IL-4, and Th2 priming, in vivo. The extended t_1/2 of the GFP protein compared with IL-4, and the ability to directly detect cytokine production ex vivo without in vitro restimulation, leads to a much greater sensitivity of detection of IL-4–producing T cells in this assay as compared with intracellular cytokine staining. In addition, recent studies from our laboratory showed that expression of IL-4–GFP in CD4⁺ T cells correlated with increased mast cell and eosinophil numbers in tissues upon secondary Ag challenge (30), confirming that increased IL-4–GFP is a good indicator of a Th2 response.

The ability of in vitro–generated DC to prime Th2 responses in vivo or in vitro has been examined in a number of previous studies (3, 35–38); however, few have used transfer of ex vivo DC as used in this study. DC generated in vitro from bone marrow precursors do not reflect the phenotypes and heterogeneity of DC populations in vivo; therefore, ex vivo studies are necessary to obtain information that is relevant to physiological situations. By using OTII cells as readouts of the T cell stimulatory ability of DC, we were also able to control for several of the limitations that may affect other DC transfer experiments (26–28). As discussed above, we were able to compare IL-4 production by T cells in conditions of similar T cell expansion, which enabled us to separate quantitative and qualitative aspects of T cell priming. By using OVA_323–337 as an Ag, we were also able to control for the possibility that the priming of IL-4–producing T cells was due to intrinsic properties of the Ag itself. For example, it has been proposed that the affinity with which Ag engage the TCR has a critical role in determining cytokine production by the responding T cells (39). In our study, all DC were presenting the same peptide epitope and thus should engage the TCR of the responding T cells with the same affinity; nonetheless, DC still differed in their ability to prime IL-4–producing T cells. The affinity of the OTII TCR for I-A^b plus OVA_323–337 is also unlikely to affect the results of our experiments, as any intrinsic propensity of these cells to differentiate into Th2 should be reflected in the induction of Th2 responses in all conditions, and this was not observed. Lastly, other bystander effects of carried over Nb Ag on the T cell response are also insufficient to explain our results given that, as shown in Figs. 1 and 6, MHCII^int DC loaded with Nb were unable to prime IL-4–GFP⁺ cells in Ptx-treated hosts or when transferred into naive mice. Together, these results are consistent with the notion that selected DC subsets can provide CD4⁺ T cells with specific signals that direct them to IL-4 production.

The innate cytokines TSLP, IL-33, and IL-25 have been shown to be required for optimal Th2 priming through their ability to upregulate OX40L expression on DC (13, 17, 34). Experiments in TSLPR KO mice and using OX40L-blocking Abs showed that the TSLP–OX40L axis was not critical for the priming of IL-4–producing T cells, a result that is consistent with other studies using helminth parasites including live N. brasiliensis (19, 40). Similarly, our recent experiments in IL-25 KO mice (41) indicated that this cytokine is also not necessary for the priming of IL-4–producing cells after injection of live or nonviable N. brasiliensis. We did not directly assess the role of IL-33 in our model: IL-33 is constitutively expressed in skin (42) and was not upregulated after Nb or M. smegmatis treatment (Supplemental Fig. 1). IL-33 was recently proposed to promote DC migration and the priming of Th2 responses via its ability to induce IL-13 production by group 2 innate lymphoid cells in the lung (23). However, IL-4 responses are intact in STAT6 KO and IL-33 KO mice infected with N. brasiliensis (29, 43), implying a nonessential role of IL-13 in Th2 priming in our model of immune response.

Our experiments did not clarify the nature of the DC signals that direct the priming of IL-4–producing T cells. Previous work from our laboratory used TLR4-KO mice to show that TLR4 expression is not necessary for the priming of IL-4–producing T cells after Nb injection (30). In the present paper, DC phenotyping experiments examined the expression of a number of markers that have been associated with induction of Th2 responses by DC. We show that Nb-loaded DC expressed increased levels of CD86 compared with nonloaded DC and modestly increased OX40L. CD86 is clearly important for the optimal priming of all T cell responses, including Th2 (44, 45). In contrast, as discussed above, OX40L was not required. Expression of CXCR5, the marker of a DC subset found in B cell follicles and that has been implicated with induction of Th2 responses (22, 46, 47), was low on both Nb-loaded and other DC. Our in vivo studies using B cell–deficient mice showed that B cells are not required for the priming of IL-4–producing T cells and for Th2 responses to N. brasiliensis (48), again suggesting that B cell follicles and the associated cell populations are unlikely to be critical for IL-4 priming. We also examined expression of PDL2 (CD273) and CD301b, both of which have recently been reported to identify a subset of d-DC that are required, but not sufficient, to induce Th2 responses (24–26) in an IRF4-dependent fashion (25). Nb-loaded DC were CD301b⁺ and expressed higher levels of PDL2 compared with other DC or DC from PBS-treated mice, suggesting a preferential involvement of such a subset in the Nb-induced response. The transcription factor IRF4, which is required for the maturation, migration, and function of skin DC in LN (25, 49, 50), was also expressed, but its levels were similar in all DC regardless of Nb loading or treatment. This latter finding aligns with recent reports showing that IRF4 is necessary for the generation of Th2 responses (25, 51), but also suggests that IRF4 is unlikely to represent a polarizing signal that determines the ability of DC from Nb-treated mice to prime for IL-4 production. Thus, our data suggest that PDL2⁺CD301b⁺IRF4⁺ DC require exposure to the appropriate stimuli to prime CD4⁺ T cells to IL-4 production.

The d-DC taking up Nb Ag expressed lineage markers that are consistent with a subset of steady-state DC normally found in the skin and skin-dLN. Our data imply that these DC must be appropriately conditioned by as yet unidentified signals, elicited by Nb exposure, before they can prime T cells to IL-4 production. We do not know whether these signals operate in the skin or in the skin-dLN or whether they act directly on DC or via other cellular intermediaries. Basophils have been suggested to be required, together with DC, for Th2 induction (7); however, it is not clear whether their function is upstream of DC or at another point in the chain of events leading to Th2 activation. Other immune cell populations such as mast cells or innate lymphocytes in the skin might also be involved (52). Importantly, together with recent data indicating that papain- or dibutylphthalate-conditioned DC are unable to prime IL-4–producing cells in vivo or in vitro, respectively (24, 26), our data also suggest that the events that lead to the conditioning of DC to Th2 induction need not be common to all Th2 responses and that different Th2-inducing Ag, such as parasites, allergens, cysteine proteases, etc., may condition DC by different mechanisms and also to different extents. In this respect, N. brasiliensis may represent a highly effective model in which to examine DC conditioning, as it is known as a powerful inducer of IL-4 responses, thereby making it possible to carry out sensitive adoptive transfer experiments using rare DC populations such as those described in this study. It should also be noted that, unlike previous studies using highly purified DC subsets (24, 26), our experiments used total MHCII^hi DC populations. Therefore, it remains possible that the induction of IL-4–producing T cells may require the cooperation of multiple DC subsets.

The dLN of mice treated with Nb also contained a population of mo-DC, identified on the basis of their expression of Ly6C and CD64, which presumably originated from blood monocytes. These mo-DC were also loaded with Nb material, suggesting that they had migrated to the LN via the skin (53, 54), or they were preferentially involved in the uptake of lymph-borne Nb material after reaching the LN (55). Unexpectedly, on the basis of two lines of evidence, mo-DC did not appear to play a direct role in the priming of IL-4–producing CD4⁺ T cells. First, local Ptx treatment at the site of Nb injection had only a partial effect on the numbers of mo-DC in LN, but drastically inhibited the priming of IL-4–producing T cells. Secondly, transfer of MHCII^int DC, which include the mo-DC population, did not result in priming of OVA-specific CD4⁺ T cells to IL-4 production. The reason for the lack of activity of mo-DC is unclear. Experiments using OVA-specific OTII–IL-4–GFP cells indicated that insufficient uptake of Nb material was unlikely to explain the lack of response, as Nb uptake was not directly required for T cell priming. Therefore, insufficient Th2-polarizing signals appear to be the most likely explanation for our finding. This result also implies that, in this model of immune response, the uptake of parasite material is not sufficient to confer DC with the ability to prime IL-4–producing T cells. Although mo-DC were not sufficient for the priming of IL-4–producing cells, they may nonetheless contribute to the establishment of the appropriate environment for conditioning d-DC to Th2 induction. Mo-DC depletion studies will be necessary to test this possibility.

In conclusion, we report that Nb injection can impart d-DC with the ability to prime IL-4–producing CD4⁺ T cells in the dLN, independently of the Ag specificity of the responding T cells. The implication of these results is that DC exposed to the appropriate parasite-conditioned environment can express all of the signals that are required to instruct CD4⁺ T cells to Th2 differentiation, without requiring the cooperation of additional cell populations. The molecular identification of the upstream signals that instruct DC to prime Th2 cells, and the downstream signals that direct CD4⁺ T cells to Th2 differentiation, will greatly facilitate the dissection of the pathways for establishing protective immunity to parasites and that lead to allergic disease in humans.

We thank Drs. William Paul and Warren Leonard for the generous gift of mouse strains, staff at the Biomedical Research Unit and Cell Technology Suite of the Malaghan Institute of Medical Research for expert support, and all colleagues at the Malaghan Institute for discussion and useful suggestions.

The authors have no financial conflicts of interest.