Polarization of naive T cells into Th1 or Th2 by ... - Semantic Scholar

Uncorrected Version. Published on June 16, 2005 as DOI:10.1189/jlb.1104631

Polarization of naive T cells into Th1 or Th2 by distinct cytokine-driven murine dendritic cell populations: implications for immunotherapy Maryam Feili-Hariri,*,†,1 Dewayne H. Falkner,† and Penelope A. Morel†,‡,2 Departments of *Surgery, †Immunology, and ‡Medicine, University of Pittsburgh School of Medicine, Pennsylvania

Abstract: Dendritic cells (DCs) activate T cells and regulate their differentiation into T helper cell type 1 (Th1) and/or Th2 cells. To identify DCs with differing abilities to direct Th1/Th2 cell differentiation, we cultured mouse bone marrow progenitors in granulocyte macrophage-colony stimulating factor (GM), GM ⴙ interleukin (IL)-4, or GM ⴙ IL-15 and generated three distinct DC populations. The GM ⴙ IL-4 DCs expressed high levels of CD80/ CD86 and major histocompatibility complex (MHC) class II and produced low levels of IL12p70. GM and GM ⴙ IL-15 DCs expressed low levels of CD80/CD86 and MHC class II. The GM ⴙ IL-15 DCs produced high levels of IL-12p70 and interferon (IFN)-␥, whereas GM DCs produced only high levels of IL-12p70. Naive T cells stimulated with GM ⴙ IL-4 DCs secreted high levels of IL-4 and IL-5 in addition to IFN-␥. In contrast, the GM ⴙ IL-15 DCs induced higher IFN-␥ production by T cells with little or no Th2 cytokines. GM DCs did not induce T cell polarization, despite producing large amounts of IL-12p70 following activation. A similar pattern of T cell activation was observed after in vivo administration of DCs. These data suggest that IL-12p70 production alone, although necessary for Th1 differentiation, is not sufficient to induce Th1 responses. These studies have implications for the use of DC-based vaccines in immunotherapy of cancer and other clinical conditions. J. Leukoc. Biol. 78: 000 – 000; 2005. Key Words: T cell activation 䡠 costimulation 䡠 IL-12p70 䡠 IFN-␥

INTRODUCTION Dendritic cells (DCs) are unique bone marrow (BM)-derived antigen-presenting cells (APC) that play a key role in the initiation and modulation of immune responses. DC-based vaccines are being used clinically for the induction of antitumor immune responses in cancer [1] and are being considered for the induction of tolerance in transplantation or autoimmune disease [2]. The factors that determine whether DCs induce immunity or tolerance are of intense interest [3]. DCs, in an immature state, reside in the peripheral tissues, are 0741-5400/05/0078-0001 © Society for Leukocyte Biology

highly endocytic, and express low levels of major histocompatibility complex (MHC) class II and costimulatory molecules. It has been suggested that these “steady state” DCs are capable of maintaining tolerance by inducing regulatory T cells or the deletion of autoreactive T cells [4]. Following maturation, DCs lose their endocytic ability, up-regulate MHC class II and costimulatory molecules, and present antigen to T cells for the initiation of immune responses [5]. DCs influence T cell proliferation and effector function by providing costimulation and establishing the cytokine environment at the time of T cell priming [6 – 8]. Following activation, naive CD4⫹ T cells differentiate into effector cells and most commonly develop into T helper cell type 1 (Th1) or Th2 cells. In general, type 1 responses are dominated by T cells producing primarily interferon (IFN)-␥. In contrast, type 2 responses are dominated by T cells producing interleukin (IL)-4, IL-5, and IL-10 [9]. Many factors influence the differentiation of Th1/Th2 cells, including the dose of antigen, the nature and degree of costimulation, and the cytokine milieu surrounding the differentiating cells. The level and/or type of costimulation influence the differentiation of Th1 or Th2 cells, and the development of Th2 responses, in particular, is strictly dependent on costimulation mediated via CD28/B7 interactions [10]. In addition, cytokines, secreted by DCs at the time of initial T cell stimulation, play an important role in the subsequent differentiation of effector T cells. The dominant cytokine produced by activated DCs following CD40 ligation or Toll-like receptor (TLR) ligation is IL-12p70, which induces IFN-␥ production and is necessary for the generation of Th1 cells [11]. DCs that fail to produce IL-12p70 are assumed to promote the generation of a Th2 response or to down-regulate Th1 responses through the action of costimulatory molecules and/or other DC-produced cytokines such as IL-12p40, IL-6, or IL-10 [12–16]. We have identified a DC population, derived from BM cells propagated in granulocyte macrophage-colony stimulating factor (GM-CSF; GM) ⫹ IL-4. This DC population prevents the development of diabetes in nonobese diabetic (NOD) mice, a

1 Correspondence: Department of Surgery, Division of Plastic Surgery, University of Pittsburgh, Scaife Hall, Suite 666, 3550 Terrace St., Pittsburgh, PA 15261. E-mail: [email protected] 2 Correspondence: Department of Immunology, University of Pittsburgh, Biomedical Science Tower E1048, 200 Lothrop St., Pittsburgh, PA 15261. E-mail: [email protected] Received November 2, 2004; revised April 20, 2005; accepted May 17, 2005; doi: 10.1189/jlb.1104631

Journal of Leukocyte Biology Volume 78, September 2005

Copyright 2005 by The Society for Leukocyte Biology.

1

murine model for type 1 diabetes [17, 18]. The therapeutic DC population expressed high levels of costimulatory molecules and produced low levels of IL-12p70 following activation [18]. Furthermore, we demonstrated that treatment of NOD mice with this DC population induced a Th2 response in vivo [19, 20]. When DCs were generated in the presence of GM alone, they produced high levels of IL-12p70 following CD40 ligation and had no effect on the development of the disease [17, 19]. In particular, we did not observe any enhancement of the Th1 response or exacerbation of disease in mice given GM DCs, suggesting that high levels of IL-12p70 production were not indicative of strong Th1 polarization by GM DCs. Recently, it was shown that the culture of human monocytes in GM and IL-15 led to the generation of DCs with features of Langerhans cells [21]. In addition, culture of murine splenic DCs in IL-15 induced IFN-␥ production by these DCs, and IL-15-treated, splenic DCs were more effective stimulators of CD8⫹ T cell proliferation [22]. Furthermore, a more recent report demonstrated that culture of BM cells in GM and IL-15 generated DCs with a heightened ability to stimulate CD8⫹ T cells and induced higher levels of IFN-␥ production by CD8⫹ T cells [23]. These data suggest that DCs that produce IL-12p70 and IFN-␥ might be more effective in driving Th1 responses than DCs that only produce IL-12p70. In this study, we have compared the ability of three distinct BM-derived DC populations to induce Th1 and Th2 responses in vitro and in vivo. We have cultured BM progenitors from NOD mice in GM alone, GM ⫹ IL-4, or GM ⫹ IL-15. These DCs differed in their ability to produce IL-12p70 and/or IFN-␥ and in their expression of costimulatory molecules. Using naive T cells from syngeneic, allogeneic, or T cell receptor (TCR) transgenic mice, we show that GM ⫹ IL-4 DCs are strong inducers of Th2 cells in vitro and in vivo. In contrast, GM ⫹ IL-15 DCs produced high levels of IFN-␥ as well as IL-12p70 and induced Th1 differentiation. Despite the fact that GM DCs produced high levels of IL-12p70, they were unable to induce strong Th1 polarization. These results suggest that Th1 differentiation is more robust when DCs produce IL-12p70 and IFN-␥ and that reliance on IL-12p70 levels alone to predict DC function might not be appropriate in all cases.

MATERIALS AND METHODS Mice Female NOD/Lt and SWR mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and were maintained in the pathogen-free animal facility of the University of Pittsburgh (PA) under Institutional Animal Care and Use Committee guidelines and approved protocols. BDC2.5-TCR transgenic mice were a gift from Dr. Christophe Benoist (Joslin Diabetes Center, Boston, MA) and were bred onto female NOD mice in our animal facility.

DC generation, phenotype, and activation Myeloid progenitor cells from NOD BM cells were prepared as described by depleting MHC class II⫹ cells, T cells, and B cells using monoclonal antibody (mAb) and complement [17]. The progenitor cells (0.5⫻106/ml) were cultured in 4 ml/well in six-well plates in the presence of GM alone (1 ng/ml, R&D Systems, Minneapolis, MN), GM ⫹ IL-4 (1 ng/ml, PeproTech, Rocky Hill, NJ), or GM ⫹ IL-15 (20 – 40 ng/ml, PeproTech) in RPMI 1640 (Gibco, Grand Island, NY) [17]. Cells were fed on day 2, and DCs were purified on day 4. The

2


GM ⫹ IL-4 DCs were purified using metrizamide gradients (purity ⬎87%) [18, 19], and the GM and GM ⫹ IL-15 DCs were purified using CD11c magnetic beads (Miltenyi Biotec, Auburn, CA; purity ⬎90%). Purified DC populations were stained with fluorescein isothiocyanate (FITC)- or phycoerythrin (PE)-conjugated mAb specific for CD80, CD86, CD40, class II, or the DC markers CD11c and CD205. DC purity was determined by the expression of DC markers and lack of staining with mAb against CD3, B220, Gr.1, and F4/80 molecules [17]. The ability of DCs to respond to inflammatory stimuli and produce cytokines was determined in vitro. Purified DCs (106/ml) were stimulated via TLR-3 by poly I:C (20 ␮g/ml, Sigma Chemical Co., St. Louis, MO), TLR-4 by lipopolysaccharide (LPS; 10 ␮g/ml, Sigma Chemical Co.), or TLR-9 by CpG (4 ␮g/ml, DNA Synthesis Facility, University of Pittsburgh) in the presence or absence of IFN-␥ (50 ng/ml, BD PharMingen, San Diego, CA) [18, 24]. Cells were cultured for 24 h, and culture supernatants were collected and kept frozen at –20°C for subsequent analyses of IL-12p70, IL-12p40, IL-6, IFN-␥, IL-10, and nitric oxide (NO), as described previously [18, 20].

Syngeneic and allogeneic mixed leukocyte reaction (MLR) Naive CD4⫹ splenic T cells from NOD or SWR mice were purified using CD4 magnetic beads (Miltenyi Biotec; purity 90 –95%). T cells (2⫻105/100 ␮l) were cultured in 96-well flat-bottom plates with varying numbers of irradiated (20 Gy) DCs per 100 ␮l medium for 3 days. The plates were pulsed with 0.5 ␮Ci/well 3H-thymidine for last 18 h, and cells were harvested as described [17].

In vitro DC:T cell cocultures Purified, splenic CD4⫹ T cells (2⫻106/ml) from NOD or SWR were cocultured with (2⫻105/ml) NOD DC populations in 12-well plates. Supernatants from these primary cultures were collected after 3– 4 days and kept frozen (–20°C). Freshly isolated CD4⫹ T cells or purified T cells from primary DC:T cell cocultures were stained with Cy-Chrome-conjugated anti-CD4 (BD PharMingen) in combination with PE-conjugated anti-CD69, CD25, and CD62L (BD PharMingen), and cells were analyzed by fluorescein-activated cell sorter (FACS). Freshly isolated CD4⫹ T cells and purified T cells from primary DC:T cell cocultures were stimulated (0.5⫻106/well) in a total volume of 1 ml in 24-well plates precoated with 10 ␮g/ml anti-TCR␣␤ mAb [19]. Plates were incubated at 37°C for 2 days, after which the culture supernatants were collected and kept frozen (–20°C). Th1 (IFN-␥) and Th2 (IL-4, IL-5, IL-10) cytokines were quantified in supernatants from primary and secondary cultures by enzyme-linked immunosorbent assay (ELISA). Naive CD62L⫹ T cells from the spleens of TCR transgenic BDC2.5-NOD mice [25] were purified using magnetic beads (Miltenyi Biotec; purity ⬎85%). BDC2.5 mice express TCRV␤4 and proliferate in response to an 11-amino acid synthetic peptide, which is structurally similar to the islet antigen GAD65 (524 –541) [26]. Purified CD62L⫹ T cells (106/well) were cocultured with DCs (2⫻105/well) and 50 ␮g/ml antigenic peptide in a total volume of 2 ml in 12-well plates for 2 days. Following incubation, T cells from these primary cultures were purified on Ficoll gradients and stained with anti-Cy-ChromeCD4 and anti-FITC-TCRV␤4 mAb in combination with PE-conjugated antiactivation/regulatory molecules (CD69, CD25, CD62L). Purified T cells from these cultures were rested for 4 –5 h at 37°C. The T cells were restimulated (0.5⫻106/well) in a total volume of 1 ml in 24-well plates precoated with 10 ␮g/ml anti-TCR␣␤ mAb. Culture supernatants were collected and kept frozen (–20°C) for detection of Th1 and Th2 cytokines by ELISA as described above.

In vivo effect of DCs on T cell responses Seven-week-old NOD mice (n⫽7/group) were injected intravenously (i.v.) with 106 DC/mouse in 100 ␮L phosphate-buffered saline (PBS) or left untreated as controls (n⫽4). Prior to injection, DC populations were washed three times in ice-cold PBS. Five days following DC injections, spleens were harvested, and CD4⫹ T cells were purified using magnetic beads. For the determination of intracellular cytokine production, splenic T cells (1.5–2⫻106/ml) were stimulated with phorbol 12-myristate 13-acetate (PMA; 10 ng/ml), ionomycin (500 ng/ml, Sigma Chemical Co.), and Brefeldin A (10 ␮g/ml, Epicentre, Madison, WI) for 5 h in 12-well plates. Following activation, cells were stained with anti-Cy-Chrome-CD4 in combination with FITC-IFN-␥/PE-IL-4 mAb (BD

http://www.jleukbio.org

PharMingen) and then analyzed by FACS for the production of intracellular cytokines by CD4⫹ T cells [27]. To detect cytokine secretion by T cells, 106/ml was stimulated with antiTCR␣␤ (10 ␮g/ml) mAb, precoated on 24-well plates in the presence or in some experiments, in the absence of soluble anti-CD28 mAb (2 ␮g/ml, BD PharMingen) for 48 h in a total of 2 ml. Culture supernatants were collected and kept frozen for ELISA as described above.

Statistics The statistical analyses were determined using one-way ANOVA. Turkey’s multiple comparison test was used amongst three groups for the level of significance. The two-tailed Student’s t-test was applied between two groups. P value ⱕ0.05 denotes significance. Data are presented as x ⫾ SD.

RESULTS Influence of cytokines on DC yield, phenotype, and cytokine production The recovery of the BM-derived DCs using three different culture conditions was similar (P⬎0.05). The DC yield (1⫻106 per 1⫻107 BM cells) from GM cultures was 0.7 ⫾ 0.22, n ⫽ 8; the GM ⫹ IL-4 was 0.76 ⫾ 0.26, n ⫽ 9; and the GM ⫹ IL-15 was 0.85 ⫾ 0.41, n ⫽ 9. DCs were analyzed by FACS for the expression of DC-specific marker CD11c and costimulatory molecules. The analysis of these results revealed that the GM ⫹ IL-4 DCs expressed high levels of CD80, CD86, CD40, and MHC class II, whereas the GM DCs and GM ⫹ IL-15 DCs expressed low levels of these molecules (Fig. 1). We then examined the DC populations for the pattern of cytokine production following activation via TLR-3 and TLR-4, as described in Materials and Methods. Although the GM ⫹ IL-4 DCs produced low levels of IL-12p70, the GM and GM ⫹

IL-15 DCs produced large amounts of IL-12p70 within 24 h of activation [LPS: GM⫹IL-4 vs. GM DCs (*, P⬍0.001) or GM⫹IL-15 DCs (*, P⬍0.01); poly I:C: GM⫹IL-4 vs. GM DCs (**, P⬍0.001) or GM⫹IL-15 DCs (**, P⬍0.01; Fig. 2, A and C)]. Similar results were obtained when DCs were stimulated via TLR-9 ligation [CpG, n⫽2, GM DCs (404⫾38.2), GM⫹IL-4 DCs (0), and GM⫹IL-15 DCs (921.5⫾518)]. In addition, we observed that the pattern of IL-12p70 production by three DC populations was similar when DCs were stimulated with LPS in the presence of IFN-␥ [LPS⫹IFN-␥: GM⫹IL-4 DCs (147⫾90, n⫽10) vs. GM DCs (4524⫾2068, n⫽6, *, P⬍0.001) or GM⫹IL-15 DCs (3785⫾2513, n⫽7, *P⬍0.01)]. A similar pattern was seen with DCs stimulated with poly I:C ⫹ IFN-␥ (data not shown). Furthermore, the GM ⫹ IL-15 DCs produced IFN-␥, whereas GM or GM ⫹ IL-4 DCs did not (IFN-␥: LPS or poly I:C, ⫹IL-15 DCs vs. GM or ⫹IL-4 DCs; †, P⬍0.05, and ††, P⬍0.05, respectively; Fig. 2D). We also measured the levels of IFN-␥ secretion by DCs following TLR-9 ligation using CpG. The GM ⫹ IL-15 DC was the only DC population to produce IFN-␥ (107.7⫾125, n⫽2), although this was lower than that seen with LPS or poly I:C. None of the DC populations produced any IL-12p70 or IFN-␥ when cultured in the absence of stimulation (data not shown). In addition, GM and GM ⫹ IL-15 DCs produced higher levels of IL12p40 compared with GM ⫹ IL-4 DCs (**, P⬍0.05); however, none of the three DC populations produced any IL-10 following activation with poly I:C (Table 1). The GM ⫹ IL-15 DCs produced higher levels of NO compared with GM ⫹ IL-4 DCs (††, P⬍0.05); however, no differences in NO production were found between GM and GM ⫹ IL-15 DCs (Table 1).

Fig. 1. Influence of cytokines on DC phenotype. DCs were generated from BM progenitors in the presence of GM-CSF alone (GM DC), GM ⫹ IL-4 (⫹IL-4 DC), or GM ⫹ IL-15 (⫹IL-15 DC). The ⫹IL-4 DCs were purified on metrizamide gradients, whereas GM and ⫹IL-15 DCs were purified using CD11c beads. The purified DCs were stained with indicated mAb (solid lines) or immunoglobulin isotype controls (dashed lines) and analyzed on an open gate for their phenotype by FACS. The numbers represent the percentage of cells staining positive for the indicated marker. Histograms represent one of the three FACS experiments.

Feili-Hariri et al. Polarization of naive T cells by cytokine-driven DCs

3

Fig. 2. Cytokine production by DC populations following activation. DCs were stimulated with the indicated TLR ligands. Supernatants were collected after 24 h and were tested for the production of (A and C) IL12p70 and (B and D) IFN-␥ by ELISA. Each symbol represents an independent experiment. IL-12p70 [LPS: ⫹IL-4 vs. GM (*, P⬍0.001) or ⫹IL-15 (*, P⬍0.01); poly I:C: ⫹IL-4 vs. GM (**, P⬍0.001) or ⫹IL-15 (**, P⬍0.01)]. IFN-␥ [LPS or poly I:C: ⫹IL-15 vs. GM (†, P⬍0.05) or ⫹IL-4 (††, P⬍0.05)].

Stimulatory function of cytokine-driven DCs in MLR Next, we examined the capacity of DC populations to stimulate naive T cells from NOD or SWR. Analysis of these data revealed that the GM ⫹ IL-4 DCs were potent stimulators of naı¨ve, allogeneic CD4⫹ T cells compared with GM or GM ⫹ IL-15 DCs (Fig. 3, right panel). DC populations stimulated the naı¨ve, syngeneic CD4⫹ T cells similarly (Fig. 3, left panel). The higher T cell stimulatory capacity of GM ⫹ IL-4 DCs in the primary allogeneic MLR was reflected in higher levels of IFN-␥ secretion by T cells during the primary stimulation (data not shown) as we previously published [17]. In addition, the higher T cell stimulatory capacity of the GM ⫹ IL-4 DCs in allogeneic MLR was correlated with the higher expression of costimulatory molecules on this DC population (Fig. 1).

In vitro effect of cytokine-driven DCs on T cell activation The nature of syngeneic (NOD) or allogeneic (SWR) T cell responses induced by DC populations was determined following secondary stimulation using anti-TCR mAb. Purified T cells were TABLE 1.

Cytokines and NO Production by BM-Derived DCa Populations following Activation with poly I:Cb GM DC

IL-12p70 (pg/ml) IL-12p40 (ng/ml) IFN-␥ (pg/ml) IL-6 (ng/ml) IL-10 (pg/ml) NO (␮M)

cultured with the three DC populations for 3– 4 days. The T cells were restimulated on anti-TCR-coated plates as described in Materials and Methods. Analysis of these results revealed that the naı¨ve, syngeneic (Fig. 4, upper panels) or allogeneic (Fig. 4, lower panels) T cells stimulated with GM ⫹ IL-4 DCs produced large amounts of IL-4 (P⬍0.007 vs. GM DC or ⬍0.02 vs. GM⫹IL-15-stimulated, syngeneic T cells) and IL-5 (P⬍0.002 vs. GM DC or P⬍0.02 vs. GM⫹IL-15-stimulated, syngeneic T cells) in addition to type 1 (IFN-␥) cytokines, indicating induction of a Th2 response. In contrast, the GM ⫹ IL-15 DCs stimulated higher IFN-␥ production by allogeneic (P⬍0.02 vs. naı¨ve T cells) T cells, however, little or no IL-4 (P⬍0.05 vs. naı¨ve, allogeneic T cells and vs. GM⫹IL-4 stimulated, allogeneic T cells; Fig. 4, lower panels); thus, this DC subset induced a predominantly Th1 response. To our surprise, the pattern of cytokine production by T cells harvested from the primary GM DC:T cultures and restimulated via TCR demonstrated no T cell polarization by the syngeneic or allogeneic T cells (Fig. 4). As an alternative means to examine the differential ability of DC populations to induce T cell polarization, we used BDC2.5TCR transgenic mice. It has been shown that naive BDC2.5 T

398.6 ⫾ 194.8 (5) 67.2 ⫾ 42.2 (5) 110.7 ⫾ 57.9 (6) 7.85 ⫾ 10.5 (4) 0 (2) 50.6 ⫾ 21.5 (2)

c

GM ⫹ IL-4 DC

GM ⫹ IL-15 DC

32 ⫾ 55.8* (7) 9.9 ⫾ 5.9** (6) 56.8 ⫾ 54.0 (8) 0.34 ⫾ 0.47 (3) 0 (2) 27.3 ⫾ 5.6 (3)

302.7 ⫾ 149.6 (8) 58.5 ⫾ 34.3 (7) 827.4 ⫾ 133† (8) 7.77 ⫾ 10.2 (4) 0 (2) 85.0 ⫾ 27.3‡ (4)

DC populations were derived from BM progenitors in the presence of indicated cytokines. b Indicated DC populations (1⫻106/ml) were stimulated with poly I:C (20 ␮g/ml) for 24 h. c Number of side-by-side experiments where two or three DC populations were compared. *,** GM ⫹ IL-4 DC versus GM or GM ⫹ IL-15, P ⬍ 0.05. † GM ⫹ IL-15 DC versus GM or GM ⫹ IL-4, P ⬍ 0.05. ‡ GM ⫹ IL-15 DC versus GM ⫹ IL-4, P ⬍ 0.05. a

4



Fig. 3. Stimulatory function of cytokine-driven DCs in primary syngeneic (left) and allogeneic (right) MLR. Splenic CD4⫹ T cells from NOD or SWR mice (2⫻105/100 ␮l) were cultured in 96-well flat-bottom plates with increasing numbers of irradiated (20 Gy) DCs/100 ␮l medium for 3 days. The cells were pulsed with 0.5 ␮Ci/well [3H]-thymidine (TdR) for last 18 h, harvested, and counted as described in Materials and Methods. Data shown are from one representative syngeneic and allogeneic MLR of three experiments performed side-by-side using the indicated DC populations. CPM, Counts per minute.

cells can be induced to differentiate into Th1 or Th2 cells [28]. In this study, the naive CD62L⫹ BDC2.5 T cells were stimulated with DC populations or splenic APC in the presence of peptide, as described in Materials and Methods. Purified T cells from DC:T cell cultures were stained with anti-CD4 and anti-TCRV␤4 mAb, in combination with anti-CD69, CD25, or CD62L, as described in Materials and Methods. FACS analysis of these markers on the gated CD4⫹/V␤4⫹ T cells revealed that a higher percentage of T cells had an activated phenotype following culture with GM ⫹ IL-4 DCs compared with T cells cultured with GM DCs or GM ⫹ IL-15 DCs (Fig. 5A), correlating with higher levels of costimulatory molecules on the GM ⫹ IL-4 DC population (Fig. 1). No differences in T cell phenotype were observed between T cells stimulated with GM or GM ⫹ IL-15 DCs (Fig. 5A). The purified T cells were then restimulated on anti-TCR-coated plates as described in Materials and Methods (Fig. 5B). Analysis of cytokine production using culture supernatants collected from three independent experiments revealed that the purified T cells from GM ⫹ IL-4 DC cultures produced high levels of IL-4 and IL-5 in addition to IFN-␥, whereas the GM ⫹ IL-15 DCs mainly stimulated the T cells for the production of IFN-␥ (Fig. 5B). The T cells from the GM DC cultures produced minimal levels of cytokines following second stimulation (Fig. 5B).

Fig. 4. In vitro effect of cytokine-driven DCs on syngeneic (upper panels) and allogeneic (lower panels) CD4⫹ T cells following secondary stimulation via TCR. Splenic CD4⫹ T cells (2⫻106/ml) from NOD (syngeneic) or SWR (allogeneic) mice were cocultured with (2⫻105/ml) DC populations in 12-well plates for 3– 4 days. T cells were harvested and 5 ⫻ 105/well cultured in 1 ml in 24-well plates, precoated with 10 ␮g/ml anti-TCR␣␤ mAb. Culture supernatants were collected after 2 days for ELISA. Each symbol represents an independent experiment. Syngeneic: IL-4 [⫹IL-4 DCs (*, P⬍0.007, vs. GM DCs or *, P⬍0.02, vs. ⫹IL-15 DCs)] and IL-5 [⫹IL-4 DCs (*, P⬍0.002, vs. GM DCs or *, P⬍0.02, vs. ⫹IL-15 DCs)]. Allogeneic: IFN-␥ [⫹IL-15 DCs (†, P⬍0.02, vs. naı¨ve T cells)], IL-4 [⫹IL-15 DCs (†, P⬍0.05, vs. naı¨ve T cells and vs. ⫹IL-4 DCs)], and IL-5 [⫹IL-4 DCs (†, P⬍0.01, vs. naı¨ve T cells or †, P⬍0.03, GM DCs)]. Pur T, Purified T cells.


5

Figure 5. In vitro effect of cytokine-driven DCs on activation molecules and cytokine production by BDC2.5 T cells. Naive CD62L⫹ BDC2.5 T cells were stimulated with DC populations or splenic APC and peptide for 2 days. T cells were purified on Ficoll and then stained and analyzed by FACS (A) for the expression of the indicated activation markers on gated CD4⫹/V␤4⫹ T cells, or T cells were rested for 4 –5 h and restimulated for cytokine production (B) following a second stimulation via TCR, as described in Figure 4. Results shown in A are x ⫾ SD of the two independent experiments and in B, are three independent experiments.

In vivo effect of cytokine-driven DCs on T cell activation To determine the function of the DC populations in vivo, we examined the pattern of cytokine production by T cells following i.v. injection of DCs. Five days following DC injection, splenic CD4⫹ T cells from DC-treated or untreated NOD mice were harvested as described in Materials and Methods. We first examined the intracellular cytokine production by splenic CD4⫹ T cells following in vitro re-

stimulation with PMA and ionomycin using FACS. Analysis of IFN-␥ and IL-4 production by CD4⫹ T cells revealed that the percentage of T cells producing IL-4 from GM ⫹ IL-4 DC-injected mice (n⫽5) was higher compared with the T cells from untreated GM (n⫽5) or GM ⫹ IL-15 (n⫽5) DC-treated mice (*, P⬍0.05; Fig. 6A). There was no difference in the number of IFN-␥-producing T cells between GM or GM ⫹ IL-15 DC-treated mice (n⫽5, P⬎0.05; Fig. 6A).

Fig. 6. In vivo effect of cytokine-driven DCs on T cell activation. Seven-week-old NOD mice were injected i.v. with 1 ⫻ 106-indicated DC populations or left untreated. Five days following injection, the splenic CD4⫹ T cells were purified as described in Materials and Methods. (A) Analysis of intracellular cytokine production by purified, splenic CD4⫹ T cells (mice, n⫽5; 1.5–2⫻106/ml) following in vitro stimulation with PMA (10 ng/ml), ionomycin (500 ng/ml), and Brefeldin A (10 ␮g/ml) for 5 h. Following activation, T cells were stained with CyChrome-CD4 for cell-surface expression and FITCIFN-␥/PE-IL-4 mAb combination for the detection of an intracellular cytokine production by FACS. Cytokine production by (B) splenic CD4⫹ T cells measured by ELISA. Cells (106/ml) were stimulated via TCR in the absence (mice, n⫽3) or presence (mice, n⫽5) of soluble, anti-CD28 mAb for 48 h. Results shown are x ⫾ SD. (A) IL-4: ⫹IL-4 versus GM or ⫹IL-15 (*, P⬍0.05). (B) The P values for ␣TCR [IFN-␥: ⫹IL-15 vs. GM or ⫹IL-4 (*, P⬍0.01); IL-4: ⫹IL-4 vs. GM (*, P⬍0.01) and ⫹IL-15 (*, P⬍0.05); IL-5: ⫹IL-4 vs. GM (*, P⬍0.01) and ⫹IL-15 (*, P⬍0.05)]. The The P values for ⫹␣CD28 [IL-5: ⫹IL-4 vs. GM (†, P⬍0.01) or ⫹IL-15 (†, P⬍0.05)].

6



Next, we examined cytokine production by T cells from DC-injected mice using ELISA. The splenic CD4⫹ T cells were restimulated in vitro on anti-TCR-coated plates in the absence (n⫽3) or the presence (n⫽5) of soluble anti-CD28 for 2 days (Fig. 6B). When T cells were restimulated via TCR, the GM ⫹ IL-4 DC-treated mice produced predominantly Th2 cytokines [IL-4: GM⫹IL-4 vs. GM DCs (*, P⬍0.01) or GM⫹IL-15 DCs (*, P⬍0.05); IL-5: GM⫹IL-4 vs. GM DCs (*, P⬍0.01) or GM⫹IL-15 DCs (*, P⬍0.05); Fig. 6B]. This was in contrast to the T cells from GM ⫹ IL-15 DC-treated mice, which produced a predominant Th1 cytokine [IFN-␥: GM⫹IL-15 vs. GM DCs or GM⫹IL-4 DCs (*, P⬍0.01); Fig. 6B)], similar to the data obtained with the in vitro experiments. Following T cell restimulation in the presence of anti-CD28 mAb, the GM ⫹ IL-4 DC-treated mice produced only higher levels of IL-5 compared with GM (†, P⬍0.01) or GM ⫹ IL-15 (†, P⬍0.05; Fig. 6B). T cells from GM DC-treated or GM ⫹ IL-15 DC-treated mice produced low levels of IL-4, and much higher levels were produced by T cells from GM ⫹ IL-4 DC-treated mice, but this did not reach any statistical significance. This suggests that the GM ⫹ IL-4 DCs had an intrinsic ability to stimulate T cells for production of Th2 cytokines. In addition, these results suggest a critical role for the production of IFN-␥ by GM ⫹ IL-15 DCs on the observed enhancement of Th1.

DISCUSSION Several studies have demonstrated that subsets of myeloid DCs play important roles in the stimulation and differentiation of Th1 or Th2 cells during immune responses [11, 29, 30]. These studies have suggested that the level of costimulatory molecules on DCs and the local cytokine environment at the time of T cell priming are important, e.g., IL-12p70 directs Th1 differentiation, and IL-4 induces Th2 differentiation [6, 11, 29, 31, 32]. In our previous studies with NOD mice, we generated a DC population from BM that expressed high levels of costimulatory molecules (CD80, CD86, CD40) and produced low levels of IL-12p70 following activation. This GM ⫹ IL-4 DC population induced regulatory Th2 cells, which altered the Th1/Th2 balance in vivo and protected young NOD mice from diabetes development [18, 19]. In contrast, a DC (GM DC) population that expressed low levels of costimulatory molecules and produced high levels of IL-12p70 did not affect the development of diabetes in this model [17–19]. Despite the ability of GM to produce high levels of IL-12p70, we did not observe an increase in Th1 differentiation in NOD mice treated with GM DCs, suggesting that IL-12p70 production is not the only factor required for Th1 differentiation. In the present study, we have compared the differential role of three phenotypically and functionally distinct BM-derived DC populations in the activation and polarization of three distinct T cell populations. Our data demonstrate that GM⫹ IL-4 DCs can directly induce the differentiation of Th2 cells in allogeneic, syngeneic, and TCR transgenic models, suggesting that this DC population has an intrinsic ability to direct Th2 differentiation. In contrast, GM ⫹ IL-15 DCs are highly effective in the induction of Th1 differentiation, whereas GM DCs fail to induce any T cell polarization. The differential ability of

GM and GM ⫹ IL-15 DCs to induce Th1 differentiation cannot be attributed to IL-12p70 production, as both populations produced high levels of this cytokine. Conversely, we found that GM ⫹ IL-15 DCs were the only DCs capable of producing IFN-␥, an important factor for the up-regulation of IL-12R␤2 in T cells [33]. Cytokines are the major factors driving the differentiation of Th1 and Th2 cells, and the ability of DCs to induce Th1 differentiation has been related to their ability to produce high levels of IL-12p70. This is becoming relevant in clinical settings, where DCs are being used as immunotherapeutic vaccines in the context of cancer [34]. To our surprise, we found that two DC populations that produced equivalent levels of IL-12p70 differed markedly in their ability to drive Th1 differentiation. No T cell polarization was observed with GM DCs, although this population secreted large amounts of IL-12p70 following activation with TLR ligands or CD40 ligation [18]. These results are in contrast to previous reports stating that IL-12p70 production by DCs was sufficient to induce abundant IFN-␥ production in T cells and Th1 responses [29, 35]. Conversely, GM ⫹ IL-15 DCs, which produced IFN-␥ in addition to IL-12p70, were effective in Th1 polarization. These DCs could have primed the naive T cells to respond to IL12p70 through the induction of the IL-12R␤2 chain by DCderived IFN-␥. This is in agreement with a recent report demonstrating that the early presence of IFN-␥ can prime CD4⫹ T cells to be IFN-␥-producing effector cells [36]. IL-15 influences DC function and phenotype in human [21] and murine [22, 23] systems. Studies in the mouse have suggested that DCs generated in the presence of IL-15 are better able to stimulate CD8⫹ T cells. In addition, it was shown that splenic DCs produce large amounts of IFN-␥ when stimulated by IL-12 and IL-18, suggesting that an additional factor besides IL-12 was needed to contribute to Th1 differentiation [37]. Further, it has been shown that infection of mice with Listeria monocytogenes stimulated IFN-␥ production by splenic CD8⫹ DC and monocytes [38], demonstrating the potential importance of this phenomenon in vivo. The culture of BM progenitors in GM ⫹ IL-15 gives rise to a DC population with strong Th1-skewing capability, which could be of clinical significance. Thus, these results suggest that reliance on IL12p70 production as an indicator of whether DCs can induce Th1 differentiation may be unwarranted. The ability of GM ⫹ IL-4 DCs to induce Th2 differentiation could be related to the low level of IL-12p70 produced by these cells, but this DC population also expressed much higher levels of costimulatory molecules (CD80, CD86, and CD40) than the other two populations. Th2 differentiation is more dependent on the presence of strong CD28-induced signals than Th1 differentiation [39]. It has been shown that T cells from CD28 knockout mice fail to produce type 2 cytokines, whereas IFN-␥ is produced normally by these cells [37, 40]. In this study, GM ⫹ IL-4 DCs induced a greater degree of T cell activation, as determined by increased proliferation and expression of activation markers. This is likely to be related to the increased expression levels of MHC class II and costimulatory molecules on the DC. There have been several reports describing DCs with Th2-driving capabilities, and recently, it has been shown that human DC2 are characterized by in-


7

creased levels of OX40L expression and that this expression is required for induction of Th2 differentiation [35]. Another molecule, CD200, is homologous to CD80 and CD86 and has costimulatory function [41]. In addition, blocking the interaction between CD200 and its ligand (CD200R) enhances skin graft survival and alters the pattern of cytokine production by T cells in MLR [42]. We have recently completed microarray analyses comparing the GM ⫹ IL-4 and GM DCs and have found that the GM ⫹ IL-4 DCs expressed higher levels of OX40L and CD200 at the RNA level and by FACS analysis (manuscript in preparation). In addition, we found that the GM ⫹ IL-4 DCs express higher levels of RNA for chemokines implicated in Th2 responses, such as CC chemokine ligand (CCL)17 and CCL22. These data, coupled with the present study, strongly suggest that the GM ⫹ IL-4 DC population has intrinsic ability to induce Th2 differentiation under a variety of conditions. The next step is to determine which of these molecules plays critical roles in Th2 differentiation induced by GM ⫹ IL-4 DCs and to use this information in the design of clinical interventions aimed at inducing regulatory Th2 responses, such as in autoimmune disease. Overall, these studies demonstrate that manipulation of the conditions under which DCs are generated from BM results in DCs with distinct abilities to stimulate naive T cells [11]. We have identified and characterized culture conditions that allow for the generation of DCs with strong Th1 or Th2 differentiating capabilities. Differentiation of Th1 cells appears to require IFN-␥ and IL-12p70 production by DCs. The ability of DCs to stimulate Th2 differentiation may be related to the reduced level of IL-12p70 production but also to the high levels of CD80 and CD86 as well as the expression of other molecules, such as OX40L and specific chemokines. As DC therapy is being considered in many areas of clinical medicine, including autoimmunity, transplantation, and cancer, there is a need for a clear understanding of the molecular interactions necessary to drive the desired Th cell response. In the case of cancer, a Th1 response is necessary for the induction of antitumor cytotoxic responses, and thus, GM ⫹ IL-15 DCs would be suitable. In contrast, in autoimmune disease, characterized by a pathogenic Th1 response, the GM ⫹ IL-4 DCs would be more appropriate.

ACKNOWLEDGMENTS This work was supported by National Institute of Health Grant #CA73743 (P. A. M.). We kindly thank Dr. Angus W. Thomson for critical review of this manuscript. We thank Dr. Christopher Benoist for the kind gift of BDC2.5-TCR transgenic mice, and we thank Dr. Rosemary Hoffman for the kind gift of reagents for NO measurements.

REFERENCES 1. Cerundolo, V., Hermans, I. F., Salio, M. (2004) Dendritic cells: a journey from laboratory to clinic. Nat. Immunol. 5, 7–10. 2. Morel, P. A., Feili-Hariri, M., Coates, P. T., Thomson, A. W. (2003) Dendritic cells, T cell tolerance and therapy of adverse immune reactions. Clin. Exp. Immunol. 133, 1–10.

8


3. Steinman, R. M., Nussenzweig, M. C. (2002) Avoiding horror autotoxicus: the importance of dendritic cells in peripheral T cell tolerance. Proc. Natl. Acad. Sci. USA 99, 351–358. 4. Hawiger, D., Inaba, K., Dorsett, Y., Guo, M., Mahnke, K., Rivera, M., Ravetch, J. V., Steinman, R. M., Nussenzweig, M. C. (2001) Dendritic cells induce peripheral T cell unresponsiveness under steady state conditions in vivo. J. Exp. Med. 194, 769 –779. 5. Banchereau, J., Steinman, R. M. (1998) Dendritic cells and the control of immunity. Nature 392, 245–252. 6. O’Garra, A. (1998) Cytokines induce the development of functionally heterogeneous T helper cell subsets. Immunity 8, 275–283. 7. Weaver, D., Poligone, B., Bui, T., Abdel-Motal, U., Baldwin, A., Tisch, R. (2001) Dendritic cells from nonobese diabetic mice exhibit a defect in NF-␬B regulation due to a hyperactive I␬B kinase. J. Immunol. 167, 1461–1468. 8. Shortman, K., Liu, Y. J. (2002) Mouse and human dendritic cell subtypes. Nat. Rev. Immunol. 2, 151–161. 9. Morel, P. A., Oriss, T. B. (1998) Crossregulation between Th1 and Th2 cells. Crit. Rev. Immunol. 18, 275–303. 10. Salomon, B., Bluestone, J. A. (2001) Complexities of CD28/B7: CTLA-4 costimulatory pathways in autoimmunity and transplantation. Annu. Rev. Immunol. 19, 225–252. 11. Boonstra, A., Asselin-Paturel, C., Gilliet, M., Crain, C., Trinchieri, G., Liu, Y. J., O’Garra, A. (2003) Flexibility of mouse classical and plasmacytoid-derived dendritic cells in directing T helper type 1 and 2 cell development: dependency on antigen dose and differential Toll-like receptor ligation. J. Exp. Med. 197, 101–109. 12. Rothe, H., O’Hara, R. M., Martin, S., Kolb, H. (1997) Suppression of cyclophosphamide-induced diabetes development and pancreatic Th1 reactivity in NOD mice treated with the interleukin (IL)-12 antagonist IL-12(p40)2. Diabetologia 40, 641– 646. 13. Trembleau, S., Penna, G., Gregori, S., Gately, M. K., Adorini, L. (1997) Deviation of pancreas-infiltrating cells to Th2 by interleukin-12 antagonist administration inhibits autoimmune diabetes. Eur. J. Immunol. 27, 2330 –2339. 14. Rincon, M., Anguita, J., Nakamura, T., Fikrig, E., Flavell, R. A. (1997) Interleukin (IL)-6 directs the differentiation of IL-4-producing CD4⫹ T cells. J. Exp. Med. 185, 461– 469. 15. Rissoan, M-C., Grouard, G., Briere, F., Malefyt, R. W., Liu, Y-J. (1999) Reciprocal control of T helper cell and dendritic cell differentiation. Science 283, 1183–1186. 16. Diehl, S., Rincon, M. (2002) The two faces of IL-6 on Th1/Th2 differentiation. Mol. Immunol. 39, 531–536. 17. Feili-Hariri, M., Dong, X., Alber, S. M., Watkins, S. C., Salter, R. D., Morel, P. A. (1999) Immunotherapy of NOD mice with bone marrowderived dendritic cells. Diabetes 48, 2300 –2308. 18. Feili-Hariri, M., Morel, P. A. (2001) Phenotypic and functional characteristics of BM-derived DC from NOD and non-diabetes-prone strains. Clin. Immunol. 98, 133–142. 19. Feili-Hariri, M., Falkner, D. H., Morel, P. A. (2002) Regulatory Th2 response induced following adoptive transfer of DC in prediabetic NOD mice. Eur. J. Immunol. 32, 2021–2030. 20. Feili-Hariri, M., Falkner, D. H., Gambotto, A., Papworth, G. D., Watkins, S. C., Robbins, P. D., Morel, P. A. (2003) Dendritic cells transduced to express IL-4 prevent diabetes in nonobese diabetic mice with advanced insulitis. Hum. Gene. Ther. 14, 13–23. 21. Mohamadzadeh, M., Berard, F., Essert, G., Chalouni, C., Pulendran, B., Davoust, J., Bridges, G., Palucka, A. K., Banchereau, J. (2001) Interleukin 15 skews monocyte differentiation into dendritic cells with features of Langerhans cells. J. Exp. Med. 194, 1013–1020. 22. Mattei, F., Schiavoni, G., Belardelli, F., Tough, D. F. (2001) IL-15 is expressed by dendritic cells in response to type I IFN, double-stranded RNA, or lipopolysaccharide and promotes dendritic cell activation. J. Immunol. 167, 1179 –1183. 23. Pulendran, B., Dillon, S., Joseph, C., Curiel, T., Banchereau, J., Mohamadzadeh, M. (2004) Dendritic cells generated in the presence of GM-CSF plus IL-15 prime potent CD8⫹ Tc1 responses in vivo. Eur. J. Immunol. 34, 66 –73. 24. Vasquez, A. C., Feili-Hariri, M., Tan, R. J., Morel, P. A. (2004) Qualitative and quantitative abnormalities in splenic dendritic cell populations in NOD mice. Clin. Exp. Immunol. 135, 209 –218. 25. Katz, J. D., Wang, B., Haskins, K., Benoist, C., Mathis, D. (1993) Following a diabetogenic T cell from genesis through pathogenesis. Cell 74, 1089 –1100. 26. Judkowski, V., Pinilla, C., Schroder, K., Tucker, L., Sarvetnick, N., Wilson, D. B. (2001) Identification of MHC class II-restricted peptide ligands, including a glutamic acid decarboxylase 65 sequence, that stim-


27.

28.

29.

30.

31.

32.

33.

ulate diabetogenic T cells from transgenic BDC2.5 nonobese diabetic mice. J. Immunol. 166, 908 –917. Morel, P. A., Falkner, D., Plowey, J., Larregina, A. T., Falo, L. D. (2004) DNA immunization: altering the cellular localization of expressed protein and the immunization route allows manipulation of the immune response. Vaccine 22, 447– 456. Hoglund, P., Mintern, J., Waltzinger, C., Heath, W., Benoist, C., Mathis, D. (1999) Initiation of autoimmune diabetes by developmentally regulated presentation of islet cell antigens in the pancreatic lymph nodes. J. Exp. Med. 189, 331–339. Pulendran, B., Smith, J. L., Caspary, G., Brasel, K., Pettit, D., Maraskovsky, E., Maliszewski, C. R. (1999) Distinct dendritic cell subsets differentially regulate the class of immune response in vivo. Proc. Natl. Acad. Sci. USA 96, 1036 –1041. Langenkamp, A., Messi, M., Lanzavecchia, A., Sallusto, F. (2000) Kinetics of dendritic cell activation: impact on priming of Th1, Th2 and nonpolarized T cells. Nat. Immunol. 1, 311–316. Nakamura, T., Lee, R. K., Nam, S. Y., Podack, E. R., Bottomly, K., Flavell, R. A. (1997) Roles of IL-4 and IFN-␥ in stabilizing the T helper cell type 1 and 2 phenotype. J. Immunol. 158, 2648 –2653. Morelli, A. E., Zahorchak, A. F., Larregina, A. T., Colvin, B. L., Logar, A. J., Takayama, T., Falo, L. D., Thomson, A. W. (2001) Cytokine production by mouse myeloid dendritic cells in relation to differentiation and terminal maturation induced by lipopolysaccharide or CD40 ligation. Blood 98, 1512–1523. Szabo, S. J., Dighe, A. S., Gubler, U., Murphy, K. M. (1997) Regulation of the interleukin (IL)-12R ␤ 2 subunit expression in developing T helper 1 (Th1) and Th2 cells. J. Exp. Med. 185, 817– 824.

34. Schuler, G., Schuler-Thurner, B., Steinman, R. M. (2003) The use of dendritic cells in cancer immunotherapy. Curr. Opin. Immunol. 15, 138 –147. 35. de Jong, E. C., Vieira, P. L., Kalinski, P., Schuitemaker, J. H., Tanaka, Y., Wierenga, E. A., Yazdanbakhsh, M., Kapsenberg, M. L. (2002) Microbial compounds selectively induce Th1 cell-promoting or Th2 cell-promoting dendritic cells in vitro with diverse Th cell-polarizing signals. J. Immunol. 168, 1704 –1709. 36. Das, G., Sheridan, S., Janeway, C. A. (2001) The source of early IFN-␥ that plays a role in Th1 priming. J. Immunol. 167, 2004 –2010. 37. Hochrein, H., Shortman, K., Vremec, D., Scott, B., Hertzog, P., O’Keeffe, M. (2001) Differential production of IL-12, IFN-␣, and IFN-␥ by mouse dendritic cell subsets. J. Immunol. 166, 5448 –5455. 38. Ohteki, T., Fukao, T., Suzue, K., Maki, C., Ito, M., Nakamura, M., Koyasu, S. (1999) Interleukin 12-dependent interferon ␥ production by CD8␣⫹ lymphoid dendritic cells. J. Exp. Med. 189, 1981–1986. 39. Lenschow, D. J., Herold, K. C., Rhee, L., Patel, B., Koons, A., Qin, H. Y., Fuchs, E., Singgh, B., Thompson, C. B. (1996) CD28/B7 regulation of Th1 and Th2 subsets in the development of autoimmune diabetes. Immunity 5, 285–293. 40. Schweitzer, A. N., Sharpe, A. H. (1998) Studies using antigen-presenting cells lacking expression of both B7–1 (CD80) and B7–2 (CD86) show distinct requirements for B7 molecules during priminig versus restimulation of Th2 but not Th1 cytokine production. J. Immunol. 161, 2762–2771. 41. Borriello, F., Lederer, J., Scott, S., Sharpe, A. H. (1997) MRC OX-2 defines a novel T cell costimulatory pathway. J. Immunol. 158, 4548 – 4554. 42. Gorczynski, R. M. (2001) Transplant tolerance modifying antibody to CD200 receptor, but not CD200, alters cytokine production profile from stimulated macrophages. Eur. J. Immunol. 31, 2331–2337.


9