- IL-15 DC:
DC cultured in the presence of GM-CSF plus IL-15
- IL-4 DC:
DC cultured in the presence of GM-CSF plus IL-4
Dendritic cells (DC) comprise a system of professional antigen-presenting cells, which induce the stimulation of very rare antigen-specific naive T cells. DC progenitors can be stimulated to differentiate into immature DC by various growth factors, including GM-CSF and IL-4. Here we show that IL-15, in combination with GM-CSF, is a growth factor for murine DC. Murine bone marrow cells, depleted of T cells, B cells, I-A+ cells and Gr-1+ granulocytes, and cultured in the presence of GM-CSF plus IL-15 (IL-15 DC), yielded DC expressing high levels of CD11c and MHC class II molecules, as well as CD11b. These cells expressed significant levels of CD40, CD80 and CD86, and could stimulate allogeneic CD4+ T cells efficiently. Interestingly, IL-15 DC were farsuperior to DC generated with GM-CSF plus IL-4 in stimulating allogeneic CD8+ T cells in vitro. Consistent with this, IL-15 DC induced much more potent antigen-specific CD8+ T cell responses with high levels of Th1 cytokines in vivo, compared to DC generated with GM-CSF plus IL-4, or with GM-CSF plus TGF-β, or with GM-CSF alone. Together, these data suggest that IL-15 promotes the development of DC, which induce potent Th1 and Tc1 responses in vivo. This suggests potential roles for these IL-15 DC cells in the immunotherapy of tumors and infectious diseases.
The DC system consists of several different subsets of cells, which differ in their phenotype, function and microenvironmental localization 1–6. In the secondary lymphoid organs of mice, at least three different subsets of mature DC have been defined: CD8α+ DC, CD8α– DC and Langerhans cell (LC)-derived DC 1–6. In addition, an immature DC subset with the capacity to secrete large amounts of IFN-α upon stimulation with viruses has been recently described 7–11. In peripheral tissues, such as the skin, at least two subsets of immature DC have been defined: these include LC in the epidermis, and dermal dendritic cells in the dermis 1–6. These distinct sub-populations of DC have been the focus of intense studies recently, particularly with respect to potential differences in their ability to modulate immune responses 12–14. For example, freshly isolated CD8α+ and CD8α– DC have been shown to induce Th1 and Th2 responses, respectively 12, 13, and in humans monocyte-derived DC and plasmacytoid DC have been shown to stimulate Th1 and Th2 responses, respectively 14. However, the relative paucity of these DC subsets has impeded their detailed study. One solution to this problem has been the identification of cytokines, which profoundly increase the numbers of such subsets in vivo. For example, in mice, Flt3 ligand (FL) induces a dramatic increase in the numbers of both CD8α+ and CD8α– DC, and GM-CSF preferentially expands CD8α– DC 6, 15–18. In addition, systemic administration of FL to mice appears to promote a significant increase in the numbers of dermal DC 19. Finally, our recent work suggests a critical role for IL-15 in the development of DC, which display many features of LC, and which are potent stimulators of CTL 20.
IL-15 is a 14–15-kDa cytokine expressed at the mRNA level in a wide variety of cell types, including placenta, activated monocytes, DC, osteoclasts, fibroblasts and keratinocytes, but not T cells 21–23. The receptor for IL-15 is a heterodimer, which consists of the IL-2Rβ and γ chain, together with a unique α chain, the IL-15Rα chain, which is alternatively spliced to yield three active forms, each capable of high-affinity binding to IL-15 23. IL-15 and IL-2 induce many similar effects on NK and T cells, which likely occur as a result of the common receptor. These effects include proliferation of mitogen-activated T cells and cytotoxicity in CD8+ T cells and survival of memory CD8+ T cells 24–27. Studies with IL-15Rα–/– and IL-15–/– mice show that IL-15 is important for the development and/or survival of NK cells, NKT cells, CD8α α+ intestinal intraepithelial lymphocytes and memory CD8+ T cells 27, 28. IL-15 receptors on T cells and NK (type 1 receptors) cells consist of IL-15Rα, IL-2Rβ and γc chain, whereas mast cells express a distinct type of IL-15 receptor (type 2 receptor) called IL-15RX 29. IL-15 type 1 receptors are also expressed in antigen-presenting cells 30–35, but little is known about the role of the IL-15-IL-15R system in their development and/or functional maturation. A recent report by Ohteki et al. 36 suggests normal numbers of spleen DC from IL-15–/– mice or γc –/– mice; however, such DC appear to be impaired in their ability to produce IL-12, IFN-γ and NO 36. Consistent with this, Mattei et al. 35 recently demonstrated a role for IL-15 in activating spleen DC. Furthermore, our recent study suggests a critical role for IL-15 in the differentiation of DC, which resemble LC-like DC derived from human monocytes 20. In the present study we wished to determine: (i) the role of IL-15 in inducing DC differentiation in mice; and (ii) potential functional differences between DC generated in the presence of IL-15 and those generated conventionally with GM-CSF + IL-4.
2 Results and discussion
2.1 Generation of CD11c+ DC from IL-15-supplemented bone marrow cultures
We have previously reported that IL-15 in combination with GM-CSF stimulates the differentiation of human monocytes into DC with LC-like properties 20. Here we were interested in determining whether IL-15, in combination with GM-CSF could stimulate DC generation from murine bone marrow cells in vitro. Mouse bone marrow cells were depleted of MHC class II+ cells, and cultured in the presence of GM-CSF and IL-15. As a control, such cells were also cultured with GM-CSF and IL-4, a cytokine combination that is known to generate DC in vitro1–6. After 7 days, both cultures contained small lymphoid-sized cells that grew in clusters. Giemsa staining of these cells revealed a typical dendritic morphology with extensive dendrites (Fig. 1A). Cells from the cultures containing GM-CSF + IL-15 (IL-15 DC; Fig. 1B), or GM-CSF + IL-4 (IL-4 DC; data not shown) expressed significant levels of CD11c, CD80, CD86 and CD40, and variable levels of MHC (Fig. 1B) class II, markers characteristic of mature DC. These cells did not express CD4, CD8, Ly6C, B220, CD19 or OX40-L (Fig. 1B). The maturation levels of IL-4 and IL-15 DC were similar, based on similar levels of expression of MHC class II (MFI 550), CD80 (MFI 180), CD86 (MFI 1000) and CD40 (MFI 70) (Fig. 1B, and data not shown). In addition, bone marrow cells were also cultured with GM-CSF alone, or GM-CSF + TGF-β (data not shown). Although DC cultured with GM-CSF alone showed a similar phenotype to IL-4 DC or IL-15 DC, DC generated with GM-CSF + TGF-β expressed much lower levels of class II, CD80, CD86 and CD40 (data not shown).
2.2 Stimulation of allogeneic CD4+ and CD8+ T cells in vitro by IL-4 and IL-15 DC
Cells from bone marrow cultures, supplemented with GM-CSF plus IL-4, or with GM-CSF plus IL-15, were harvested at day 7 and then assayed for their ability to stimulate CD4+ or CD8+ T cells, in an allogeneic MLR assay. Consistent with their mature phenotype, both populations of DC were potent at stimulating naive, allogeneic CD4+ T cells (Fig. 2A), However, IL-15 DC were far superior at stimulating CD8+ T cells (Fig. 2B). This latter result is consistent with our previous published data, which suggests that human IL-15 DC can stimulate CD8+ T cells more efficiently than IL-4 DC 20. To exclude the possibility that the superior ability of IL-15 DC to stimulate CD8+ T cells was caused by any direct effects of contaminant IL-15 in cultures, we used an antibody against IL-15 to block such effects. Even in the presence of an anti-IL-15 antibody, IL-15 DC were far superior to IL-4 DC in the stimulation of naive CD8+ T cells (data not shown).
2.3 Stimulation of antigen-specific CD4+ T cells in vivo
We investigated whether IL-15 and IL-4 DC could stimulate antigen-specific CD4+ T cells in vivo, using OVA-specific, MHC class II-restricted (I-Ab), α β T cell receptor (TCR) transgenic mice (OT-2 mice) 37, 38. In these mice, the CD4+ OVA-specific T cells express Vα2 and Vβ5. TCR transgenic T cells were adoptively transferred into Thy-1 congenic B6.PL.THY1a (B6.PL) mice, such that they constituted a small but detectable proportion of all T cells 39. In this system, the fate of OVA-specific transgenic T cells was followed using the Thy1.2 antibody, which stains the transferred cells, but not the host cells. Cells with the phenotype Thy1.2+ CD4+ Vα2+ Vβ5+ are considered OVA-specific CD4+ T cells. In some of the experiments we simply used Thy1.2 in combination with CD4, to detect the OVA-specific T cells.
DC in IL-4 or IL-15 day 7 bone marrow cultures were pulsed with OVA 323–339 peptide for 2 h at 37°C, or left unpulsed, then washed extensively and injected into the footpads of mice which had been previously reconstituted with OT-2 cells. Four days later, the clonal expansion of OVA-specific CD4+ T cells was assessed by flow cytometry. Both IL-4 DC and IL-15 DC pulsed with OVA induced potent clonal expansion of antigen-specific CD4+ T cells (Fig. 3A, B). Unpulsed IL-4 or IL-15 DC were much weaker (Fig. 3A, B).
To assess cytokine production by the OT-2 cells, we cultured single cell suspensions from the draining popliteal lymph nodes in the presence of varying concentrations of OVA peptide 323 – 339 for 72 h. Cytokine production by antigen-specific T cells was measured by assaying the culture supernatants for IL-2, IFN-γ, IL-4, IL-5 and IL-10. Assessment of cytokine production in these cultures revealed differences between mice injected with IL-4 versus IL-15 DC (Fig. 3C). While both IL-4 and IL-15 DC induced similar levels of and IFN-γ and IL-2 (data not shown), IL-15 DC induced lower levels of IL-4 and IL-10. Thus, IL-15 DC appear to induce a strong Th1 response, while IL-4 DC induced a mixed Th response, which contains both Th1 and Th2 cytokines. These data are consistent with our previous observations in the human 20, where IL-15 DC induced a more potent Th1 response than IL-4 DC.
In contrast to IL-15 DC, or IL-4 DC, those generated with GM-CSF alone, or with GM-CSF + TGF-β were much weaker at stimulating antigen-specific CD4+ T cells (data not shown).
2.4 Stimulation of antigen-specific CD8+ T cells in vivo
The different Th responses induced by IL-4 DC and IL-15 DC, suggested that there may be differences in antigen-specific CD8+ T cell responses. We investigated this using OT-1 mice (H-2Kb restricted, OVA-specific TCR-transgenic mice 38; 2.5×106 or 5×106 spleen cells from OT-1 mice (B6.PL, Thy1.2) were adoptively transferred into B6.PL (Thy 1.1) hosts 39. Cohorts of host mice were injected with either IL-4 or IL-15 DC alone, or pulsed with OVA SIINFEKL peptide. Clonal expansion of OVA-specific CD8+ T cells (CD8+ Thy1.2+) was assessed by flow cytometry. Both IL-15 and IL-4 DC enhanced the clonal expansion of OVA-specific CD8+ T cells, (Fig. 4A, B). Unpulsed IL-4 or IL-15 DC were weaker at stimulating OT-1 cells in vivo.
To assess cytokine production by the OT-1 cells, we cultured single-cell suspensions from the draining popliteal lymph nodes in the presence of varying concentrations of OVA SIINFEKL peptide for 72 h. Cytokine production by antigen-specific T cells was measured by assaying the culture supernatants for IL-2, IFN-γ, IL-4, IL-5 and IL-10. There were significant differences between mice injected with IL-4 vs. IL-15 DC. IL-15 DC induced much higher levels of IFN-γ than IL-4 DC (Fig. 4C); this is consistent with the secretion of Th1 cytokines by GM-CSF + IL-15-primed Th cells (Fig. 3C), and with our previous report 20 demonstrating the superior ability of human IL-15 DC to stimulate naive CD8+ T cells and activate cytotoxic T cells. In contrast, both IL-4 DC and IL-15 DC were able to induce low levels of IL-4, IL-5 and some IL-10. In contrast to IL-15 DC and IL-4 DC, those generated with GM-CSF alone, or with GM-CSF + TGF-β were much weaker at stimulating antigen-specific CD8+ T cells (data not shown).
Considering that IL-15 DC induce much more potent Tc1 responses than IL-4 DC, we wondered whether the DC could be induced to secrete different levels of cytokines that influenced polarization of Tc responses. Thus, DC generated with GM-CSF alone, GM-CSF plus IL-4, GM-CSF plus IL-15, or GM-CSF plus TGF-β were cultured in vitro with Staphylococcus aureus Cowan strain (SAC) plus IFN-γ, a combination which is known to stimulate IL-12p70 production in DC 1–6. All DC subsets could be induced to secrete similar levels of the pro-inflammatory cytokines IL-12p70, IL-6, TNF-α (data not shown). Since both IL-4 DC and IL-15 DC appear to secrete similar levels of IL-12p70, the superior Tc1 inducing ability of IL-15 DC is likely not attributable to IL-12p70 secretion. However, DC generated with GM-CSF alone, or with GM-CSF plus TGF-β secrete lower levels of IL-12p40 and IL-12p70. The failure of DC cultured with GM-CSF alone, or GM-CSF plus TGF-β to secrete high levels of IL-12 is consistent with their poor Th1 and Tc1 inducing abilities.
In summary, our data suggests that IL-15 plays an important role in the generation of mouse DC. Interestingly, IL-15 DC appear to be far superior to DC generated with GM-CSF and IL-4 in inducing the proliferation and differentiation of antigen-specific CD8+ Tc1 cells. Their potent roles in stimulating antigen-specific CD8+ T cells in vivo suggests important roles for these cells in the immunotherapy of tumors and infectious diseases.
3 Material and methods
OT-II TCR transgenic mice (strain 426–6), generated by Dr. W. Heath (Walter & Eliza Hall Institute, Melbourne, Australia) and Dr. F. Carbone (Monash University, Melbourne, Australia) 37 were obtained from Dr. J. Kapp (Emory University, Atlanta). OT-1 TCR transgenic 38, C57BL/6, and B6.PL.THY1a (B6.PL) mice were purchased from Jackson Laboratory (Bar Harbor, ME). All mice were kept in microisolator cages in a specific-pathogen free facility. For adoptive transfers, age-matched, male C57BL/6 or B6.PL.THY1a recipients were given 2.5×106–5.0×106 OT-II cells or OT-1 TCR transgenic T cells intravenously, as described previously 39.
3.2 DC generation
Mice femurs were removed and mechanically purified from surrounding tissues and bone marrow was flushed using cold PBS. Cells were treated with Tris-buffered ammonium chloride to lyse erythrocytes. Afterwards, B cells, T cells, I-A+ cells, and Gr-1+ granulocytes were removed positively by specific antibodies against CD19, CD3, MHC class II (I-Ab) and Gr-1 (PharMingen, San Diego, CA). The remaining cells were I-A–, and were cultured in RPMI complete medium plus 10% fetal bovine serum (FBS) with mouse GM-CSF alone (20 ng/ml; Sigma, St.Louis, MO), or GM-CSF + human IL-15 (20 ng/ml) or GM-CSF + mouse IL-4 (20 ng/ml; Sigma) or GM-CSF + TGF-β (10 ng/ml) in 6-well plates for 6 days. Every other day cultures were fed with fresh media containingcytokines. On day 6, cells were harvested and used for different experiments.
3.3 Flow cytometry of DC
DC (5×105) were stained with anti-CD11c, in combination with CD11b, MHC class II, CD40, CD80, CD86, OX40 ligand (PharMingen) for 20 min on ice, in the presence of 2.4G2, an Fc receptor blocking antibody (PharMingen). Cells were then washed and analyzed with a FACSCalibur.
3.4 Allogeneic MLR
Splenic CD4+ or CD8+ T cells were purified by positive magnetic selection, using the Miltenyi magnetic selection system. Triplicates of 5×104 CD4+ or CD8+ T cells were seeded into a 96-well round-bottom plate and titrated numbers of day 6 bone marrow-derived DC from C57BL/6 mice were added. Cells cultures were performed in complete RPMI and pulsed at day 4 with 1 μCi/well [3H]thymidine (ICN) for 16 h.
3.5 In vivo immune responses
C57BL/6 mice were reconstituted with either OT-2 cells, or OT-1 cells, as previously described 39. OT-2 CD4+ T cells recognize the OVA323–339 peptide, in the context of I-Ab38, 39. OT-1 CD8+ T cells recognize the OVA SIINFEKL peptide, in the context of H-2Kb37, 39. The mice were then injected in the footpads with either IL-15 DC alone, IL-15 DC pulsed with the relevant OVA peptide, IL-4 DC alone or IL-4 DC pulsed with the relevant OVA peptide. Footpad injections were given in a volume of 25 μl. At 4 days after immunization, flow cytometric analysis of OT-2 cells in the draining popliteal lymph nodes was performed. Briefly, cell suspensions were prepared from the draining popliteal lymph nodes, and incubated on ice with PE-labeled anti-Thy1.2 (PharMingen, San Diego, CA), FITC-labeled Vα2 (PharMingen), Cy-Chrome labeled CD4 (PharMingen) and Biotin-labeled Vβ5 (PharMingen), followed by streptavidin allophycocyanin (PharMingen). In some of the experiments we simply used Thy1.2 versus CD4. For analyses of OT-1 cells, cell suspensions of draining popliteal lymph nodes or spleens were stained with PE-labeled anti-Thy1.2 (PharMingen), FITC-labeled Vα2 (PharMingen), and Biotin labeled CD8 (PharMingen), followed by streptavidin allophycocyanin (PharMingen). In some of the experiments we simply used Thy1.2 versus CD8.
3.6 In vitro cultures
Four days after priming with DC, popliteal lymph node cells were plated in triplicate in 96-well round-bottom plates (Costar, Cambridge, MA) in 200 μl RPMI complete medium supplemented with 10% FBS, together with different concentrations of OVA323–339 (for OT-2 cells) or OVA SIINFEKL (for OT-1 cells). Proliferation were assessed after 72 h of culture in a humidified atmosphere of 5% CO2 in air. Cultures were pulsed with 1.0 μCi [3H]thymidine for 12 h and incorporation of the radionucleotide was measured by β-scintillation spectroscopy. For cytokine assays, aliquots of culture supernatants were removed after 72 h, pooled, and assayed for the presence of IFN-γ, IL-2, IL-4, IL-5 and IL-10 by ELISA.
3.7 Cytokine ELISA
Cytokines were quantified by ELISA kits from Pharmingen.
We thank Dr. Rafi Ahmed for advice and discussions. This work was supported by grants from NIH (RO1 DK57665–01, RO1 AI48638–01, and RO1 AI056499–01 ) to B.P, and DA016029 to MZ. J.B was supported by Baylor health Care System and the NIH grant, RO-1 CA78846.