In vitro treatment of human transforming growth factor-β₁-treated monocyte-derived dendritic cells with haptens can induce the phenotypic and functional changes similar to epidermal Langerhans cells in the initiation phase of allergic contact sensitivity reaction

Abstract

Human monocyte-derived dendritic cells (MoDCs) obtained from peripheral blood monocytes (PBMC) cultured with granulocyte–macrophage colony-stimulating factor (GM-CSF) and interleukin-4 (IL-4) can be activated in vitro by a variety of simple chemicals such as haptens and several metals. Recently, it has been demonstrated that transforming growth factor-β₁ (TGF-β₁) can induce further differentiation of MoDCs to the cells that share some characteristics with epidermal Langerhans cells, i.e. they contain Birbeck granules and express E-cadherin. In this study, using such TGF-β₁-treated dendritic cells (TGF-β₁⁺ DCs), we examined the in vitro effects of representative haptens, i.e. NiCl₂ and dinitrochlorobenzene (DNCB), on their phenotypic and functional characteristics, comparing with those reported in vivo in epidermal Langerhans cells during the sensitization phase of a contact sensitivity reaction. Treatment of TGF-β₁⁺ DCs with NiCl₂ increased their expression of the molecules related to antigen presentation such as CD86, major histocompatibility complex class I and class II, and CD83, although weakly, in addition to that of those essential for their migration to the regional lymph nodes, such as CD49e, CD44 and its variant 6, while it down-regulated the expression of the molecules required for homing to the skin and staying in the epidermis, such as cutaneous leucocyte antigen (CLA) and E-cadherin. It also increased the production of tumour necrosis factor-α, but not that of IL-1β or IL-12. DNCB also increased their CD86 expression and down-regulated E-cadherin and CLA, but did not affect other phenotypic changes that were observed in TGF-β₁⁺ DCs treated with NiCl₂. TGF-β₁⁺ DCs treated with either NiCl₂ or DNCB increased their allogeneic T-cell stimulatory function. In addition, reverse transcribed polymerase chain reaction revealed augmented expression of chemokine receptor 7 mRNA by TGF-β₁⁺ DCs when treated with either NiCl₂ or DNCB. Moreover, consistent with this data, TGF-β₁⁺ DCs treated with these chemicals chemotactically responded to macrophage inflammatory protein-3β. These data suggest the possibility that TGF-β₁⁺ DCs present a good in vitro model to study the biology of epidermal Langerhans cells.

Introduction

We found that murine Langerhans cells (LCs) up-regulate their expression of class II major histocompatibility complex (MHC) antigen and antigen-presenting function after hapten painting on the skin, whereas the chemicals that simply irritate the skin rather than sensitize animals cannot induce this phenomenon.¹ Later, we demonstrated that the application of haptens to murine skin was accompanied by the up-regulation of several co-stimulatory molecules on LCs, i.e. CD40, CD54, CD80 and CD86.² Thus, haptens can induce the activation of LCs in vivo. In addition, Enk & Katz³ indicated a crucial role played by interleukin-1β (IL-1β) secreted by LCs themselves in their increased expression of class II MHC antigen after hapten application.

Recently, using monocyte-derived dendritic cells (MoDCs), which were obtained from the culture of purified monocytes in the presence of granulocyte–macrophage colony-stimulating factor (GM-CSF) and IL-4, we examined similar effects of simple chemicals on their activation.⁴ The obtained results clearly demonstrated that purified DCs responded to such chemicals as NiCl₂ and dinitrochlorobenzene (DNCB) in vitro by significantly augmenting their expression of CD54, CD86 and human leucocyte antigen (HLA)-DR and by increasing their production of pro-inflammatory cytokines. Furthermore, the augmented expression of CD86 on DCs treated with DNCB was suppressed by either anti-IL-1β or anti-tumour necrosis factor-α (TNF-α) antibody, while that induced with NiCl₂ was relatively insensitive to these antibody treatments, suggesting that different chemicals use different signal transduction pathway to stimulate DCs.

Geissmann et al.⁵ reported that transforming growth factor-β₁ (TGF-β₁), could induce LCs from peripheral blood monocytes (PBMC) in vitro in the presence of GM-CSF and IL-4. Recently, however, Jaksits et al.⁶ have demonstrated several phenotypic changes between epidermal LCs and DCs derived from the culture of monocytes in the presence of GM-CSF, IL-4 and TGF-β₁. Namely, although both share CD1a, E-cadherin and Birbeck granules, MoDCs with or without TGF-β₁ express CD11b and factor XIIIa.

The first purpose of this study was to examine whether these TGF-β₁-treated DCs (TGF-β₁⁺ DCs) also respond to the simple chemicals like MoDCs do, since there were several papers demonstrating that TGF-β₁ affects the intracellular signal transduction pathway in various cells, such as mitogen-activated protein kinase (MAP) kinases.^7,8 If TGF-β₁⁺ DCs can respond to simple chemicals, the next purpose was to reveal whether haptens can induce phenotypic and functional changes in these TGF-β₁⁺ DCs that have been reported to occur in epidermal LCs during the initiation phase of the contact hypersensitivity reaction in vivo, such as the down-regulation of E-cadherin⁹ and cutaneous leucocyte antigen (CLA),⁸ the induction of matrix metalloproteinase-9 (MMP-9) expression,¹⁰ and the augmentation of some of β₁-integrins,^11,12 CD44 and some of its variants,¹³ as well as the expression of chemokine receptor 7 (CCR7) mRNA that is involved in their chemotactic response to macrophage inflammatory protein-3β (MIP-3β).^14–16 In addition, since Enk & Katz³ have suggested that the maturation of epidermal LCs stimulated by cutaneous application of haptens is triggered by IL-1β secreted by LCs themselves, we examined whether the maturation of TGF-β₁⁺ DCs induced by haptens is also mediated by IL-1β.

Materials and methods

Media and reagents

The medium used in the study was RPMI-1640 including 25 mm HEPES buffer (Sigma Chemical Co., St. Louis, MO) supplemented with 2 mm l-glutamine, 1 mm sodium pyruvate, 1% penicillin, streptomycin and fungizone antibiotic solution (Sigma) and 10% fetal calf serum (FCS; Bioserum, Canterbury, Victoria, Australia) (complete medium). The buffer used for purification of CD14⁺ monocytes from (PBMC) was phosphate-buffered saline (PBS) supplemented with 1% bovine serum albumin (less than 1 ng/mg of detectable endotoxin) (Sigma) and 5 mm ethylenediaminetetraacetic acid (EDTA; MACS buffer). NiCl₂ (Sigma), and 2,4-dinitrochlorobenzene (DNCB) (Wako Pure Chemicals, Osaka, Japan) were used for the stimulation of DCs. The endotoxin content of the final dilution used was < 30 pg/ml, as determined by the Limulus amoebocyte lysate assay (Seikagaku Co Inc. Tokyo, Japan). We used the following monoclonal antibodies (mAbs) for immunostaining: fluorescein isothiocyanate (FITC)-anti-CLA, anti-CD40, anti-CD80, anti-CD86 antibodies, phycoerythrin (PE)-conjugated-anti-CD29, FITC or PE-conjugated isotype-matched mouse control antibodies [immunoglobulin G2a (IgG2a) and G2b; PharMingen, San Diego, CA], FITC- or PE-conjugated anti-HLA-DR antibody (Becton-Dickinson, San Jose, CA), FITC-conjugated anti-CD49e and -CD49f (Serotec Ltd, Oxford, UK), FITC-conjugated anti-CD49d, PE-conjugated anti-CD83 antibody (Immunotech, Marseilles, France), FITC-conjugated anti-CD54 antibody (Ancell, Bayport, MN), FITC-conjugated anti-CD44 antibody (Caltag Laboratories, Burlingame, CA), FITC-conjugated anti-HLA I-ABC antibody (Biosource, Camarillo, CA), PE-conjugated anti-CD1a antibody (Coulter, Hialeah, FL), monoclonal anti-human E-cadherin antibody (Takara Biomedicals, Tokyo, Japan), Lag (gift of Dr F. Furukawa, Hamamatsu University, Sizuoka, Japan) and isotype control antibody (IgG1) (Sigma). For examining the effects of cytokines on surface molecule expression by TGF-β₁⁺ DCs, we used mouse anti-IL-1β mAb, anti-TNF-α antibody (Genzyme Corporation, Cambridge, MA), and isotype matched control antibodies (PharMingen). Magnetic activated cell sorter (MACS) colloidal supermagnetic microbeads conjugated with anti-human CD14 mAb (CD14 microbeads) were purchased from Miltenyi Biotec Inc., Sunnyvale, CA. Recombinant human (rh) GM-CSF was a gift from Kirin Brewery Co., Tokyo, Japan, and rhIL-4 and rhTGF-β₁ were purchased from Genzyme Corporation, Cambridge, MA and R & D Systems, Minneapolis, MN. NiCl₂ was solubilized in distilled water, while DNCB was solubilized in dimethyl sulfphoxide (DMSO) at a concentration of 1 m. The final concentration of DMSO was always less than 0·1% and cultures of DCs with 0·1% DMSO were also examined as a control.

Culture of DCs from PBMC

PBMC were isolated from heparinized fresh leucocyte-enriched buffy coats from different donors using Lymphoprep (Nycomed Pharma As, Oslo, Norway). After several washes with PBS, 1 × 10⁸ PBMC were treated with 150 µl of CD14 microbeads in 600 µl of MACS buffer at 4° for 30 min. After washing with MACS buffer, the cells coated with CD14 microbeads were separated by a magnetic cell separator, MACS (Miltenyi Biotech), according to the manufacturer's protocol. Before culturing we examined the percentage of CD14⁺ cells in these preparations by flow cytometry and used cell specimens containing more than 98% of CD14⁺ cells in the experiments.

CD14⁺ monocytes (2 × 10⁶/ml) were cultured in complete medium containing 100 ng/ml rhGM-CSF and 100 ng/ml rhIL-4 with or without 100 ng/ml TGF-β₁ for 6 days. On the 5th day the cells were subcultured to expand the volume twice with the same combination of cytokines.

Chemical treatment of cultured TGF-β₁⁺ DCs

Six days after the culture with GM-CSF and IL-4 with or without TGF-β₁, the cultured cells were treated with various concentrations of NiCl₂, or DNCB for 48 hr. In some experiments, in addition to the chemicals, neutralizing antibody for IL-1β or TNF-α at each optimal concentration or control isotype-matched antibody was added to the culture.

Flow cytometry and immunocytological analysis

Forty-eight hours after treatment with the chemicals, the phenotypic changes of TGF-β₁⁺ DCs were analysed by flow cytometry. Cell staining was performed using a combination of FITC-conjugated antibody (25 µg/ml) and PE-conjugated antibody (25 µg/ml). For E-cadherin expression on TGF-β₁⁺ DCs, the TGF-β₁⁺ DCs were first incubated with anti-E-cadherin antibody and revealed by PE-conjugated anti-mouse immunoglobulins (Tago Immunologicals, Camarillo, CA). After washing with PBS supplemented with 1% BSA and 0·02% NaN₃ (FACS buffer), the cells were analysed by FACScan using cellquest software (Becton Dickinson). Dead cells were gated out after staining with 0·5 µg/ml propidium iodide solution.

Immunohistochemical analysis of LAG expression was performed on cytospin slides for MoDCs and TGF-β₁⁺ DCs. The slides were at first fixed and permeabilized by Fix & Perm Cell Permeabilization Kit (Caltag Laboratories, An Der Grub, Austria). After blocking with goat serum, the slides were incubated with Langerhans cell-associated granule (LAG) or non-reactive mouse IgG1 antibody, MOPC 21 and then stained with FITC-conjugated anti-mouse immunoglobulins (Tago Immunologicals, Camarillo, CA). The slides were mounted in permafluor (Lipshaw Immunon, Pittsburgh, PA) containing 0·5 µg/ml propidium iodide, and were observed under a confocal microscope (Leica microscope and system GmbH, Wetzlar, Germany).

Enzyme-linked immunosorbent assay (ELISA) for cytokine production

The culture supernatants of the TGF-β₁⁺ DCs treated with various chemicals were recovered 24 and 48 hr after culture. The production of IL-1β, TNF-α, IL-10, IL-12p70 and MMP-9 was measured by ELISA kit obtained from R & D Systems for IL-1β, Genzyme for TNF-α and IL-10, Endogen Inc., Soburn, MA for IL-12p70 and Amersham Life Science, Buckinghamshire, UK for MMP-9, respectively, using 96-well microtitre plates, according to each manufacturer's instructions. The levels of IL-1β, TNF-α, IL-10, IL-12p70, or MMP-9 were calculated by using a standard curve obtained with recombinant IL-1β (from 0 to 125 pg/ml), recombinant TNF-α (from 0·5 to 32·0 pg/ml), recombinant IL-10 (from 0 to 256 pg/ml), recombinant IL-12p70 (from 0 to 600 pg/ml), and recombinant MMP-9 (from 0 to 64 ng/ml).

Primary allogeneic T-cell stimulation by TGF-β₁⁺ DCs

PBMC were obtained by Ficoll–Paque (Pharmacia) gradient centrifugation of heparinized blood. These PBMC were treated with a pan-T-cell isolation kit (Miltenyi Biotec), according to the manufacturer's protocol. These enriched T cells consisted of more than 95% CD3⁺ cells. T cells (2 × 10⁵ cells/well) were co-cultured in 96-well flat-bottom microtitre plates with various numbers of LCs, which were precultured with studied chemicals for 2 days and washed. After 4 days of culture at 37° in a 5% CO₂ humidified atmosphere, the cells were pulsed with 10 µm bromodeoxyuridine (BrdU) (Boehringer Mannheim GmbH, Germany) during the last 16 hr of culture. The immunoenzymatic measurement of BrdU uptake by T cells was performed by 5-bromo-2′-deoxy-uridine labelling and detection kit III (Boehringer Mannheim). The intensity of the colour reaction was read in an automatic ELISA plate reader (Titertek Multiskan MC) using a 492-nm filter. The mean values ± SE from three replicates were taken from each sample culture. To decrease the background proliferation of LCs, LCs were irradiated with 3000 rads with a Softex (Softex Co. Ltd, Tokyo, Japan).

Reverse transcription–polymerase chain reaction (RT-PCR)

Total RNA was extracted from TGF-β₁⁺ DCs using the guanidinium thiocyanate method mentioned by the manufacturer (ISOGEN; Nippon gene Inc., Toyama, Japan), and RNA was quantified by spectrophotometry. First-strand cDNA was synthesized from total RNA extracted in RNAse-free conditions. The reaction was performed with 1 µg of total RNA using TaKaRa RNA PCR kit (AMV) (Takara Biochemicals, Osaka, Japan), as described by the manufacturer. PCR was performed in a Perkin Elmer 2400 thermal cycler, in a final volume of 100 µl reaction mixture containing 1 µm of each forward and reverse primer, using TaKaRa RNA PCR kit, according to the manufacturer's protocol. CCR7 primers, which were described by Dieu et al.,¹⁴ were designed. The + 154/CCR7 5′-GATTACATCGGAGACAACACC-3′ forward primer and −1202/CCR7 5′-TAGTCCAGGCAGAAGAGTCG-3′ reverse primer, and the primer set for human GAPDH (Maxim Biotech, Inc., San Francisco, CA) were used for RT-PCR. The reaction mixture was subjected to 30 cycles of PCR with the following conditions: 94° for 1 min, 55° for 1 min and 72° for 1 min. PCR products were visualized on 1·2% agarose gels containing 0·1 µg/ml ethidium bromide.

Chemotactic assay

Cell migration was evaluated using a chemotaxis microchamber technique (48-well Boyden microchamber; Neuroprobe, Pleasanton, CA). In brief, MIP-3β was diluted to concentrations ranging from 1 to 1000 ng/ml in RPMI-1640 medium, and was added to the lower wells of the chemotaxis chamber. Then, 5 × 10⁴ cells/well in 50 µl of RPMI-1640 medium were applied to the upper wells of the chamber, with a standard 5-µm pore polyvinylpyrrolidone-free polycarbonate filter (Neuroprobe) separating the lower wells. The chamber was incubated at 37° in humidified air with 5% CO₂ for 90 min. Then, cells that had migrated to the underside of the filter were stained with May–Gruenwald's solution and Giemsa's solution and were counted microscopically at × 200 magnification in five randomly selected fields per well. Each assay was performed in duplicate and the results were expressed as the mean ± SD of migrating cells per field.

Statistical analysis

The statistical significance of the effects of simple chemicals on allogeneic T-cell stimulatory function was analysed by using the Mann–Whitney test between non-treated TGF-β₁⁺ DCs and TGF-β₁⁺ DCs treated with simple chemicals.

Results

TGF-β₁⁺ DCs expressing E-cadherin and CLA can be induced from peripheral blood CD14⁺ cells in the presence of GM-CSF, IL-4 and TGF-β₁

According to the procedure reported by Geissmann et al.,⁵ we induced peripheral blood CD14⁺ monocytes to differentiate into TGF-β₁⁺ DCs, which were characterized by the expression of CD1a, HLA-DR antigen, CD86, E-cadherin and CLA (Fig. 1). The expression of Lag antigen, a Birbeck granule-related molecule, was observed in 10–20% of TGF-β₁⁺ DCs, while no MoDCs expressed Lag antigen (Fig. 2). These TGF-β₁⁺ DCs significantly differ phenotypically from MoDCs by the expression of E-cadherin and CLA. They did not express the maturation marker of DCs, CD83.

Figure 1.

Monocyte-derived DCs treated with TGF-β₁ induce E-cadherin and CLA. Peripheral blood CD14⁺ monocytes were cultured with GM-CSF and IL-4 (MoDCs) or with GM-CSF, IL-4 and TGF-β₁ (TGF-β₁⁺ DCs) for 6 days. The surface expression of several phenotypic markers for DCs or LCs were examined by flow cytometry. These are representative data from five different experiments that reproduced a similar staining pattern.

Figure 2.

Monocyte-derived DCs treated with TGF-β₁ induce LAG expression. The cytospin slides were at first fixed and permeabilized. After blocking with goat serum, the slides were incubated with LAG or non-reactive mouse IgG1 antibody, MOPC 21, and then stained with FITC-conjugated anti-mouse immunoglobulins. The slides were mounted in permafluor containing 0·5 µg/ml propidium iodide, and observed under a confocal microscope. Original magnification: × 200. Inset: original magnification: × 400.

Haptens can induce phenotypic changes in TGF-β₁⁺ DCs in vitro

We reported that NiCl₂, the representative hapten, increased the expression of CD54, CD86 and HLA-DR antigen on MoDCs, while DNCB augmented only the expression of CD86.⁴ Therefore, at first, we examined whether these chemicals could induce similar phenotypic changes on TGF-β₁⁺ DCs. In the following experiments, we adjusted concentrations of these chemicals at a three-fold dilution from their respective lethal doses (more than 1000 µm of NiCl₂ and more than 100 µm of DNCB were lethal for TGF-β₁⁺ DCs), and demonstrated only the data obtained with the concentrations that induced a maximum effect. Figure 3, which is one of the representative data from five different experiments, clearly demonstrates that 300 µm of NiCl₂ and 30 µm of DNCB induced phenotypic changes in TGF-β₁⁺ DCs similar to those noted in MoDCs. In addition, they increased CD83 expression on TGF-β₁⁺ DCs weakly. Various in vivo studies on epidermal LCs have demonstrated that LCs showed several phenotypic changes, such as the down-regulation of E-cadherin,⁹ and CLA,⁸ and the augmentation of some β₁-integrins,^11,12 CD44 and some of its variants,¹³ in addition to the augmentation of CD54, CD86 and HLA-DR antigen in the initiation phase of the allergic contact hypersensitivity reaction. Thus, we next examined whether these phenotypic changes would also occur in TGF-β₁⁺ DCs treated with the haptens. As for E-cadherin and CLA expression, TGF-β₁⁺ DCs stimulated with NiCl₂ or DNCB down-regulated their expression of these two molecules. When we examined the expression of CD49e, -f and -g and CD29, NiCl₂ increased the expression of CD49f and CD29 integrins, while DNCB rather down-regulated CD29 expression. Furthermore, we found that again only TGF-β₁⁺ DCs treated with NiCl₂ increased the expression of pan CD44 epitope and its variant exon V6.

Figure 3.

The treatment with NiCl₂ and DNCB induced several phenotypic changes on TGF-β₁⁺ DCs. TGF-β₁⁺ DCs differentiated from peripheral blood CD14⁺ monocytes by GM-CSF, IL-4 and TGF-β₁ were treated with 300 µm of NiCl₂ or 30 µm of DNCB for 48 hr. These hapten-treated TGF-β₁⁺ DCs were analysed by flow cytometry for their expression of the molecules related to the function of LCs. These are representative data from five different experiments that showed similar results.

TGF-β₁⁺ DCs treated with NiCl₂ increases the secretion of TNF-α, while MMP-9 is constitutively secreted by TGF-β₁⁺ DCs irrespective of stimulation

We examined the production of several cytokines and MMP-9. TGF-β₁⁺ DCs could not secrete IL-1β or IL-12 p70 with any kinds of stimulation (Table 1). Only TNF-α and IL-10 were detected in the culture supernatants of TGF-β₁⁺ DCs stimulated with NiCl₂. In contrast, TGF-β₁⁺ DCs constitutively secreted MMP-9, but this was not augmented by any stimulation.

Table 1.

Production of cytokines and MMP-9 by MoDCs (ng/ml)

Treatment	IL-1β	TNF-α	IL-10	IL-12	MMP-9
Exp. 1
Non-treated	0·0	0·0	0·0	0·0	22·4
300 μm Ni	0·0	18·1	0·0	0·0	21·1
30 μm DNCB	0·0	0·0	0·0	0·0	20·2
100 μm DNCB	0·0	0·0	0·0	0·0	21·4
10 μm DNFB	0·0	0·0	0·0	0·0	25·8
30 μm DNFB	0·0	0·0	0·0	0·0	24·9
exp. 2
Non-treated	0·0	36·8	23·0	0·0	61·2
300 μm Ni	0·0	57·9	36·3	0·0	57·4
Exp. 3
Non-treated	0·0	0·0	0·0	0·0	10·9
30 μm DNCB	0·0	49·0	0·0	0·0	10·4
30 μm DNCB	0·0	0·0	0·0	0·0	14·9
30 μm DNCB	0·0	0·0	0·0	0·0	14·5
Exp. 4
Non-treated	ND	0·0	0·0	ND	ND
100 μm Ni	ND	10·0	0·0	ND	ND
300 μm Ni	ND	10·0	0·0	ND	ND
1000 μm Ni	ND	80·0	0·0	ND	ND
30 μm DNCB	nd	0·0	0·0	ND	ND
100 μm DNCB	ND	5·0	0·0	ND	ND

nd, not determined.

TGF-β₁⁺ DCs treated with haptens increase their allogeneic T-cell stimulatory activity

Concomitantly with the augmented CD86 expression, TGF-β₁⁺ DCs pretreated with NiCl₂ or DNCB stimulated allogeneic T cells more vigorously than non-treated TGF-β₁⁺ DCs (Fig. 4).

Figure 4.

TGF-β₁⁺ DCs treated with haptens increase their allogeneic T-cell stimulatory function. Allogeneic T cells (2 × 10⁵ cells/well) were co-cultured in 96-well flat-bottom microtitre plates with various numbers of TGF-β₁⁺ DCs, which were precultured with 300 µm of NiCl₂ or 30 µm of DNCB for 2 days and washed. After 4 days of culture at 37° in a 5% CO₂ humidified atmosphere, the cells were pulsed with 10 µm bromodeoxyuridine (BrdU) during the last 16 hr of culture. The allogeneic T-cell stimulation by TGF-β₁⁺ DCs was evaluated by the immunoenzymatic measurement of BrdU uptake by T cells.

TGF-β₁⁺ DCs treated with haptens augment CCR7 mRNA expression and respond chemotactically to MIP-3β

Mature DCs stimulated by lipopolysaccharide (LPS), TNF-α, or CD40L have been reported to respond to MIP-3β, probably with CCR7.^14,17 Furthermore, Saeki et al.¹⁵ presented evidence suggesting that MIP-3β and CCR7 participate in the emigration of LCs from epidermis to the regional lymph nodes via lymphatics. Therefore, in this study, we examined CCR7 mRNA expression on TGF-β₁⁺ DCs under the treatment with the haptens and their chemotactic response to MIP-3β. As shown in Fig. 5, TGF-β₁⁺ DCs treated with either NiCl₂ or DNCB in addition to those treated with LPS and TNF-α, up-regulated CCR7 mRNA. Parallel to this data, TGF-β₁⁺ DCs treated with the haptens demonstrated a significant chemotactic response to MIP-3β, which was comparable to that of the response of TGF-β₁⁺ DCs treated with 100 ng/ml of TNF-α, but much weaker than that of TGF-β₁⁺ DCs treated with 10 ng/ml of LPS (Fig. 6). Unexpectedly, we recognized the migration of significant numbers of non-treated TGF-β₁⁺ DCs which did not express CCR7 mRNA. As this migration was totally independent on the concentration of MIP-3β, we considered that their migration was a random migration rather than a chemotactic response.

Figure 5.

TGF-β₁⁺ DCs treated with haptens increase CCR7 mRNA expression. Total RNA was isolated from TGF-β₁⁺ DCs, which were treated with 10 ng/ml of LPS, 100 ng/ml of TNF-α, 300 µm of NiCl₂ or 30 µm of DNCB, for 2 days, and used for RT-PCR with CCR7- and glyceraldehyde-3-phosphate dehydrogenase (G3PDH)-(loading control) specific primers. (a) Non-treated TGF-β₁⁺ DCs; (b) TGF-β₁⁺ DCs treated with 10 ng/ml of LPS; (c) TGF-β₁⁺ DCs treated with 100 ng/ml of TNF-α; (d)TGF-β₁⁺ DCs treated with 300 µm of NiCl₂; (e) TGF-β₁⁺ DCs treated with 30 µm of DNCB.

Figure 6.

TGF-β₁⁺ DCs treated with haptens respond to MIP-3β. TGF-β₁⁺ DCs, which were precultured with 10 ng/ml of LPS, 100 ng/ml of TNF-α, 300 µm of NiCl₂, or 30 µm of DNCB, for 2 days and washed, were assessed by a chemotaxis microchamber technique for their chemotactic responses to MIP-3β; 5 × 10⁴ cells/well in 50 µl of RPMI 1640 medium were applied to the upper wells of the chamber, with a standard 5-µm pore polyvinylpyrrolidone-free polycarbonate filter separating the lower wells. The chamber was incubated at 37° in humidified air with 5% CO₂ for 90 min. Then, cells that had migrated to the underside of the filter were counted microscopically at × 200 magnification in five randomly selected fields per well. Each assay was performed in duplicate and the results were expressed as the mean ± SD of migrating cells per field. Shown are two of three experiments with similar results. In these experiments, significant numbers of non-treated TGF-β₁⁺ DCs which did not express CCR7 mRNA and migrated independent of the concentration of MIP-3β, which suggested that their migration was a random migration rather than a chemotactic response.

The activation of TGF-β₁⁺ DCs mediated by NiCl₂ is mediated by mechanisms different from those in that by DNCB

Since Enk& Katz³ suggested that the cutaneous application of haptens triggered the maturation of epidermal LCs via IL-1β secreted by LCs themselves, and our data indicated that NiCl₂ and DNCB induced phenotypic changes and cytokine production in TGF-β₁⁺ DCs in a different fashion, we finally examined the mechanism involved in the hapten-induced activation of TGF-β₁⁺ DCs (Table 2). When we added anti-TNF-α or anti-IL-1β antibody to the culture of TGF-β₁⁺ DCs stimulated with the haptens, they suppressed significantly (P < 0·05), although partially, the augmentation of CD86 induced by DNCB, whereas they could not suppress that induced by NiCl₂ at all. In a control experiment where we added the relevant neutralizing antibody to the culture of TGF-β₁⁺ DCs stimulated with recombinant TNF-α or IL-1β, each antibody completely abrogated the activation of DCs induced by the corresponding cytokine.

Table 2.

The suppression of hapten-induced CD86 augmentation on MoDCs by anti-cytokine antibodies (%suppression)

	Antibodies
Chemicals	Control	Anti-TNF-α	Anti-IL-1β
Exp.1
DNCB	0	43	80
NiCl2	0	13	7·3
Exp. 2
DNCB	0	96	26
NiCl2	0	23	28
Exp. 3
DNCB	ND	57	57
NiCl2	ND	0	0
Exp. 4
DNCB	ND	2	60
NiCl2	ND	0	0
Exp. 5
DNCB	0	19	22
NiCl2	0	19	0

Per cent suppression of mean fluorescence (MFI) was defined as [(MFI of hapten-stimulated DCs) − (MFI of hapten-stimulated DCs treated with cytokine antibody)] / [(MFI of hapten-stimulated DCs) − (MFI of non-treated controls)] × 100.

Discussion

The initial purpose of this study was to answer the question whether TGF-β₁⁺ DCs are also stimulated directly by haptens like MoDCs. Our present study demonstrated that haptens such as NiCl₂ and DNCB directly stimulated TGF-β₁⁺ DCs to induce increased CD86 expression and augmented antigen presenting function, as noted in MoDCs.¹

The second purpose was to characterize the phenotypic and functional changes occurring in TGF-β₁⁺ DCs treated with the haptens. Several lines of evidence have demonstrated that DCs, including LCs, show two stages of maturation.¹⁸ Namely, the first is the differentiation from bone marrow precursor cells to immature DCs and the second is the maturation from immature DCs to mature DCs that have potent antigen-presenting function. Furthermore, recent progress on DC research has shown that the latter maturation process of DCs is characterized by their augmented expression of various co-stimulatory molecules, class II MHC antigens,^2,19–22 and CD83,^23,24 down-regulation of c-fms,²⁵ production of several pro-inflammatory cytokines and IL-12^26–28 and chemotactic response to MIP-3β, probably via CCR7.^14,17 In addition to these changes associated with DC maturation, epidermal LCs have been reported to acquire the following phenotypic and functional changes in the maturation process during the induction phase of allergic contact hypersensitivity, i.e. the down-regulation of E-cadherin⁹ and CLA,⁸ expression of MMP-9,¹⁰ and augmentation of β₁-integrin and some α-integrins,^11,12 and CD44 and some of its variants.¹³ In our experiments, we found that TGF-β₁⁺ DCs stimulated by haptens lacked several characteristics of fully matured DCs or LCs, because their expression of CD83 was weak in addition to the lack of IL-12 production. In contrast, however, they down-regulated E-cadherin and CLA, and chemotactically responded to MIP-3β, which paralleled their CCR7 mRNA expression. Furthermore, TGF-β₁⁺ DCs stimulated with NiCl₂ induced augmentation of β₁-integrin and α₅-integrin, CD44 and some of its variants, which are expected to play a role in the migration of DCs through the dermis to draining lymph nodes. They also produced MMP-9, although its production was not augmented by the hapten treatment.

These unique characteristics of hapten-stimulated TGF-β₁⁺ DCs indicate that haptens trigger maturation of TGF-β₁⁺ DCs, but not to a full extent, while they can supply the signals necessary for them to emigrate from the epidermis to regional lymph nodes via the lymphatics. In future, these observations made in TGF-β₁⁺ DCs should also be confirmed in epidermal LCs.

In this study, we demonstrated that inducibility of phenotypic changes of TGF-β₁⁺ DCs differs between different haptens. Namely, only the TGF-β₁⁺ DCs stimulated with NiCl₂, but none of those treated with DNCB, showed augmentation of CD54, HLA-DR, α₅β₁-integrin, CD44 and CD44v6. Consistent with these differences, the experiments using anti-TNF-α and anti-IL-1β antibodies demonstrated such a difference between NiCl₂ and DNCB in their stimulatory effects on TGF-β₁⁺ DCs. Namely, DNCB seemed to at first stimulate TGF-β₁⁺ DCs to secrete IL-1β or TNF-α, which then induced their CD86 expression. On the other hand, NiCl₂ directly induced the augmentation of CD86 expression on TGF-β₁⁺ DCs. These data indicate that haptens trigger maturation of TGF-β₁⁺ DCs at least via two different mechanisms. It has been demonstrated that the two transition metals, NiCl₂ and CoCl₂, directly induce the formation of reactive oxygen intermediates that activate NF-κB.^29–32 On the other hand, Arner et al.³³ reported that DNCB was an irreversible inhibitor of human thioredoxin reductase, which was accompanied by a large increase in NADPH oxidase activity. This increased NADPH oxidase activity produces reactive oxygen intermediates, which may activate NF-κB. Indeed, recently, Rescigno et al.³⁴ demonstrated that LPS induced DC maturation through NF-kB activation. However, our present study demonstrating the difference between DNCB and NiCl₂ in their stimulatory effects on TGF-β₁⁺ DCs, indicates the importance of the signals other than NF-kB in their activation of TGF-β₁⁺ DCs. We are currently investigating the signal transduction pathways involved in activation by these chemicals.

Finally, it is well known that the reaction of epidermal LCs to simple chemicals such as haptens or metals is the initial step in allergic contact sensitivity. Therefore, it is crucial to clarify the precise underlying mechanisms. For the observation of the direct effects of these chemicals on LCs, however, it is essential to culture large numbers of purified immature LCs. Although we want to examine human LCs in such a way, it is impossible to fulfil the required conditions. Thus our present data suggest that TGF-β₁⁺ DCs can be substituted for the immature epidermal LCs and provide a good in vitro model to study the biological role of epidermal LCs in allergic contact sensitivity.

Glossary

Abbreviations

DC

dendritic cell

MoDC

monocyte-derived dendritic cell

MoLC

monocyte-derived Langerhans cell