Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Dendritic cells (DC) are the most potent antigen-presenting cells (APCs) of the immune system and are specialized in the generation of primary immune responses. DC exist in two main stages of maturation. As immature cells, DC are very effective in capturing and processing antigens. Mature DC lose their capacity to capture antigens and become high immunostimulatory cells that can trigger naive T cells. The ability of mature DC to activate naive CD4⁺ and CD8⁺ T cells is related to their high major histocompatibility complex (MHC) expression and to the presence of adhesion and costimulatory molecules, as well as to their production of cytokines such as interleukin (IL)-12 (1).

DC are widely distributed in the blood, the skin, and in most tissues, especially in lymphoid organs, where they display an important phenotypic diversity, the significance of which remains unclear (3). Murine and human lung DC are localized within airway epithelium, within alveolar septa, and in the connective tissue surrounding pulmonary vessels (4). Lung DC are believed to form a sentinel network of antigen-capturing cells, particularly around the airways.

The difficulty in obtaining large numbers of DC from human tissues has been palliated by the recent development of methods to generate DC in vitro. Different DC development pathways have been described. The methods used differ in two aspects: (1) the use of cord blood precursors or adult blood monocytes (Mo); and (2) the use of different cytokines. Caux and colleagues (5, 6) have shown that human CD34⁺ cord blood progenitors can differentiate in granulocyte macrophage colony-stimulating factor (GM-CSF) and tumor necrosis factor (TNF)- along two unrelated pathways: the Langerhans cells (LC) and the CD14-derived DC with discrete biologic functions. Siena and associates (7) reported that the FLT3 ligand (FL) significantly enhanced (GM-CSF + TNF-)-dependent generation of either DC from bone marrow or CD34⁺ cord blood cells. Interestingly, Maraskovsky and coworkers (8) have shown that injection of FL into mice resulted in a significant accumulation of functionally active DC within spleen, lymph nodes, peripheral blood, bone marrow, Peyer's patches, liver, and skin. Sallusto and colleagues (2, 9, 10) and Romani and associates (11) have derived DC from human Mo. Mo were cultured for 6 to 7 d in GM-CSF and IL-4 to obtain "immature" DC. Immature DC were characterized by a high endocytic activity that can be measured by fluorescein isothiocyanate (FITC)-dextran incorporation (10) but a low capacity to stimulate T cells. When challenged by inflammatory stimuli such as TNF-, lipopolysaccharide (LPS), or monocyte-conditioned medium (11, 12), these cells become mature DC. They lose their capacity to incorporate FITC-dextran but become strong T-cell stimulators and express high levels of CD80, CD40, CD86, and CD83. This latter (CD83) is a specific cell-surface marker of mature DC whose function is unknown (13, 14).

The aim of the current study was to characterize DC purified from human lung (15) using the in vitro system developed by Sallusto and coworkers (2, 9, 10, 16). Thus, phenotype and function of freshly isolated lung DC were compared with blood-derived DC at various stages of maturation to better understand the function and the state of differentiation of DC in human lung.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents, Media, and Antibodies

Ficoll-Paque and Percoll were from Pharmacia (Uppsala, Sweden). Cells were washed with Hanks' balanced salt solution without Ca²⁺ and Mg²⁺ (HBSS), and cultured using RPMI 1640 supplemented with 2 mM glutamine, penicillin (100 U/ml), streptomycin (100 U/ml) (all three from GIBCO BRL, Grand Island, NY), and 10% fetal calf serum. Human recombinant IL-4 was from R&D Systems (Minneapolis, MN), GM-CSF from Immugenex Corp. (Los Angeles, CA), and LPS (Escherichia coli 055:B5) from Difco (Detroit, MI). The following primary antibodies were used for cytometry: CD83 (HB15A) and CD40 (MAB89) (Immunotech, Marseille, France); CD80 (L307.4) (Becton Dickinson, San Jose, CA); CD86 (IT2.2), CCR5 (2D7), and CD64 (10.1) (PharMingen, San Diego, CA); CD14 (MY4) (Coulter Clone, Hialeah, FL); CD1a (OKT6) (Ortho-Diagnostic System, Raintain, NJ); and CCR1 (53504.11) (R&D Systems). Secondary antibody was a FITC isotype-specific goat antimouse immunoglobulin (Ig)G (Immunotech). CTLA-4 Ig was a kind gift from Dr. C. Gimmi, University of Basel (Basel, Switzerland).

DC Culture

The method described by Sallusto and colleagues (9) was employed in the present study. Briefly, human peripheral blood mononuclear cells (PBMC) from healthy donors were isolated from a Ficoll-Paque density gradient centrifugation. Monocytes were isolated on Percoll gradient. The monocytes were recovered and plated in six-well tissue culture plates (Costar, Cambridge, MA) at 5 × 10⁶ cells/well in 5 ml of RPMI 1640 complete medium where GM-CSF and IL-4 were added at final concentrations of 10 ng/ml. Fresh media containing GM-CSF and IL-4 were added on Day 3. On Day 6, maturation of immature DC was induced by addition of LPS (1 µg/ml) for 3 additional days.

Isolation of Human Lung DC

Lung tissue was obtained from surgical specimens. Tissue distant from a limited primary lung carcinoma was collected. Lung fragments were rinsed with HBSS to remove residual blood. Tissue was minced and digested with type IA collagenase (Sigma, St. Louis, MO) for 1 h at 37°C. Because this enzyme treatment is an obligatory step to prepare cell suspension, we have this collagenase treatment on the in vitro-generated DC. No differences in the expression of the cell-surface molecules were detected between treated and untreated DC (data not shown). The enzyme-digested fragments were taped through a stainless-steel screen and separated on a Ficoll-Paque density gradient to obtain pulmonary mononuclear cells (PMC). PMC (4 × 10⁶ cells/ml) were cultured in P10 petri dishes in complete medium. The nonadherent cells were removed after 1 h, using three rinses of HBSS. The adherent cells were incubated for an additional 16-h period at 37°C in complete medium. The cells released after three rinses of HBSS are referred to as loosely adherent mononuclear cells (LAM).

Unstained LAM were separated into DC and autofluorescent macrophages (AM) with a Coulter EPICS V according to the presence or absence of autofluorescent inclusions with a coherent INNOVA 90 light source, using a 488-nm wavelength for excitation and a 588-nm filter for emission. The gates were set to remove cell debris and to select mature AM and nonphagocytic DC. These latter cells are in contrast to phagocytic macrophages, potent stimulators of allogeneic T lymphocytes (15).

Fluorescence-Activated Cell Sorter Analysis

Cell staining was performed using specific and control isotype monoclonal antibodies (mAbs), followed by FITC-conjugated goat antimouse antibodies. The samples were analyzed on a FACScan (Coulter, EPICS XL-MCL, or Becton Dickinson). Dead cells were gated out on the basis of their light-scattering properties. The results are reported as mean fluorescence intensity (MFI) index: MFI_(Cell+mAb)/MFI_{(Cells+IgG control1)}.

Isolation of Peripheral Blood T Cells

PBMC were obtained from the buffy coat of healthy volunteers from the blood transfusion center. After a Ficoll-Paque density gradient, the interface mononuclear cells were washed three times in HBSS. After adhesion on plastic culture dishes, the lymphocytes were removed by washing and purified by being passed through a nylon wool column twice.

Mixed Leukocyte Reactions

The APCs were irradiated with a cesium source (3000 Rad). Purified allogeneic T-lymphocyte cells (1.5 × 10⁵) were cocultured in the presence or absence of APC (2.5 × 10⁴) in 96-well tissue culture plates (Costar). Mixed leukocyte reaction (MLR) was performed in 200 µl of complete medium in triplicate for 5 to 6 d at 37°C. Supernatants were collected (100 µl) and cells pulsed with 0.5 µCi [³H]thymidine (NEN, Boston, MA) for 18 h before being harvested. For antibody inhibition experiments, DC were preincubated for 1 h in the presence of mAbs against specific accessory molecules CD80 (BB1; Ancell, Bayport, MN), CD86, and CD40 (mAbs had already been used for phenotypic studies), all at 10 µg/ml. T cells were subsequently added for 5 additional days. Antibody effects are reported as percent of incorporation (% incorporation: [cpm_{(T cells+DC+mab)}/cpm_{(T cells+DC)}] × 100).

Uptake of FITC-Dextran

Cells (1 × 10⁶/ml) were resuspended in RPMI 1640 complete medium. FITC-dextran (Molecular Probes, Eugene, OR) was added at the final concentration of 1 mg/ml. The cells were incubated for 1 h at either 4°C (nonspecific binding) or 37°C. Cells were washed three times with cold complete medium containing 0.01% NaN₃, and analyzed on a FACScan (Coulter, EPICS XL-MCL, or Becton Dickinson).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Lung DC Expression of CD40, CD80, and CD86

The expression of the costimulatory molecules has been compared on Mo immature DC, mature DC, and lung DC (Figure 1). CD40 and CD86 were weakly expressed on Mo, whereas CD80 expression was undetectable. GM-CSF and IL-4 stimulation induced CD80 expression on immature DC and increased CD40 and CD86 expression with an MFI index increasing by 15- and 2-fold, respectively. All these markers were further enhanced when immature DC were challenged by LPS to become mature DC. Effectively, more than 90% of mature DC express CD80, CD86, and CD40, with MFI indexes that are 3, 15, and 2.5 times higher than immature DC, respectively.

View larger version (31K): [in this window] [in a new window]   Figure 1.   Phenotypic analysis of in vitro-generated DC and lung DC. Purified Mo were cultured for 6 d with GM-CSF (10 ng/ml) and IL-4 (10 ng/ml) to obtain immature DC, followed by a maturation period from Day 6 to Day 9 in the presence of LPS (1 µg/ml) to obtain mature DC. Human pulmonary DC were isolated as described in MATERIALS AND METHODS. The expression of the costimulatory molecules CD80, CD86, and CD40 were analyzed by flow cytometry. The control is an irrelevant isotypic IgG (dashed lines). Data are representative of at least three experiments performed.

Surprisingly, lung DC express moderate levels of CD40, CD80, and CD86 on their cell surfaces. CD86 was similarly expressed on lung DC and Mo. Lung DC, however, expressed higher levels of CD40 and CD80 than did Mo, but much lower levels than immature and mature in vitro- derived DC (Figure 1).

Lung DC Expression of CD83, CD14, and CD64

Other cell-surface molecules are useful markers for following the maturation of Mo-derived DC. Mature dendritic cells acquire the CD83 expression (14) and decrease or lose CD14 and CD64 expression (9). CD83, which is weakly expressed on 4.8 ± 1% (mean ± standard error of the mean [SEM]; n = 6) of monocytes and on 18.5 ± 2% of immature DC, is strongly upregulated by LPS. Indeed, 78 ± 5% of mature DC are CD83⁺. CD83, however, is weakly expressed on lung DC (Figure 2). In contrast, CD14, which is present on 84 ± 5% of monocytes and 85 ± 5% of immature DC, is strongly downregulated by LPS. Thus, only 38 ± 9% of mature DC are still CD14⁺-positive with an MFI index about 10 times lower than that of Mo. Similar to mature DC, lung DC express low levels of CD14. In the in vitro system, the expression of the CD1a, a marker present on LC, varies according to the blood donor. Lung DC had no CD1a on their surfaces. CD64, a phagocytic surface molecule, is present on Mo. This molecule is rapidly downregulated and weakly expressed on DC, with no differences between immature and mature DC. Lung DC express lower levels of CD64 than do Mo, but higher levels than in vitro-derived mature or immature DC (Figure 2).

View larger version (38K): [in this window] [in a new window]   Figure 2.   Phenotypic analysis of Mo, immature DC, mature DC, and human lung DC (prepared as described in Figure 1). The expression of CD83, CD14, CD1a, and CD64 was analyzed by flow cytometry. The control is an irrelevant IgG (dashed lines). Data are representative of at least three experiments performed.

CCR1 and CCR5 Expression on the Surface of Lung DC

Surface chemokine receptors were investigated on lung DC and compared with those on Mo and immature and mature DC. As already shown by Sallusto and colleagues (16), CCR1 and CCR5 are present on monocytes and immature DC but disappear on mature DC (Figure 3). A high proportion of monocytes express CCR5 compared with CCR1, but both were clearly present on most immature DC. CCR1 and CCR5 were absent from mature DC. Lung DC showed a definitive shift of the whole population compared with the control population of both CCR1 and CCR5 markers, despite some autofluorescent background. However, the intensity of these markers on lung DC appeared to be intermediate between the levels expressed by immature and mature DC.

View larger version (21K): [in this window] [in a new window]   Figure 3.   Phenotypic analysis of Mo, immature DC, mature DC, and human lung DC (prepared as described in Figure 1). Expression of CCR1 and CCR5 was analyzed by flow cytometry. The control is an irrelevant IgG (dashed lines). Data are representative of at least three experiments performed.

T-Cell Stimulatory Capacity of Lung DC

Because the ability to induce a high MLR is a characteristic of mature dendritic cells, the capacity of peripheral Mo and immature and mature DC to induce allogeneic T cell proliferation was compared with that of AM or lung DC. As shown in Figure 4, mature DC and lung DC were both strong stimulators of allogeneic T cells by inducing similar levels of T-cell proliferation. If no difference was present in the T-cell proliferation induced by mature DC or lung DC, mature DC were 3.8 ± 0.7-fold (mean ± SEM; n = 6) more efficient than immature DC. Lung DC were a 4.2 ± 0.8- and 7.6 ± 0.9-fold stronger inducer of allogeneic T-cell proliferation than were Mo or AM, respectively.

View larger version (35K): [in this window] [in a new window]   Figure 4.   T-cell stimulatory capacity of DC. Irradiated Mo, immature DC, mature DC, and human AM or lung DC (prepared as described in Figure 1) (2.5 × 104) were cocultured with T cells (1.5 × 105). On the fifth day, cells were pulsed with [3H]thymidine and harvested 20 h later. Results are expressed as mean cpm of [3H]thymidine incorporation ± SEM (n = 6). #P < 0.05 by unpaired Student's t tests.

Inhibitory Effects of mAbs on T-Cell Proliferation in the MLR

Because lung DC have a high capacity to stimulate the proliferation of T cells, despite rather low CD40, CD80, and CD86 expression, the functional significance of these accessory molecules was tested using blocking mAb (10 µg/ml) against CD80, CD86, and CD40 (Figure 5) in allogeneic MLR assays. T-cell proliferation was found to be strongly inhibited by the presence of mAbs against either CD80 or CD86. This inhibition of T-cell proliferation with antibodies against CD80 or CD86 was 48 ± 6% (mean ± SEM, n = 3) and 38 ± 4% (mean ± SEM, n = 5), respectively. When we used the recombinant molecule CTLA4 (at 10 µg/ml) to block CD80 and CD86 simultaneously, the MLR was decreased by 47.8 ± 16% (n = 4; data not shown). In contrast, anti-CD40 (n = 5) had no significant effect on T-cell proliferation. These antibodies to CD40, CD80, CD86, and CTLA4 had similar inhibitory properties on the MLR induced by monocytes and immature and mature DC derived from Mo (data not shown).

View larger version (41K): [in this window] [in a new window]   Figure 5.   Functional significance of accessory molecules expression. T cells were cultured with lung DC in the presence of 10 µg/ml mAbs against specific accessory molecules CD80 (n = 3), CD86 (n = 5), CD40 (n = 5). mAb effects were reported as percent of incorporation (% incorporation = [(cpm(T cell+DC+mAb)/cpm(T cell+DC)) × 100] ± SEM). #P < 0.05 by unpaired Student's t tests.

Capacity of Lung DC to Uptake FITC-Dextran

Mannose receptors are crucial for the phagocytosis of soluble antigens as well as of many pathogens (10, 17). Although immature DC have been shown to have the capacity to endocytose large amounts of FITC-dextran particles, mature DC lose this function (10). The capacity of Mo and lung DC to uptake FITC-dextran particles was compared with in vitro-matured DC. During 1 h assay, 84 ± 2% of immature DC were able to endocytose FITC-dextran particles, compared with 7 ± 3% of Mo. Mature DC had an MFI index that was 3.5-fold lower than immature DC, with only 40 ± 5% of cells positively labeled. In contrast, lung DC had between 40 and 68% positive cells with an MFI index between immature and mature DC, and showed a capacity to use their mannose receptors much more efficiently than did Mo or even mature DC (Figure 6).

View larger version (13K): [in this window] [in a new window]   Figure 6.   Capacity of lung DC to uptake FITC-dextran. Mo, immature DC, mature DC, and human lung DC (prepared as described in Figure 1) were incubated with 1 mg/ ml FITC-dextran for 1 h at 37°C, washed three times, and analyzed by flow cytometry as cells were incubated 1 h on ice to assess nonspecific binding (dashed lines). Data are representative of at least three experiments performed.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In this study we used the in vitro model of blood-derived DC developed by Sallusto and associates (9) to better characterize human lung DC. We showed that lung DC possess an immature phenotype with likewise a rather good capacity for the endocytosis of FITC-dextran. They express low levels of classic costimulatory molecules and few or no CD83, which are predominantly present on mature blood-derived DC (13, 14). In contrast, they are very powerful stimulators of allogeneic T cells.

DC isolated from human lung have a phenotype similar to freshly isolated blood DC (18, 19). They express high levels of LFA3 and MHC Class II molecules (20, 21), and only few cells weakly express CD14 molecules. They also have a weak CD83 expression, a glycoprotein with unknown function which is specifically present on mature DC (13, 14). CD1a, a characteristic cell-surface molecule present on epidermal LC (22), was shown to be present on blood-derived DC as reported by others (18). However, CD1a was consistently absent on lung DC, confirming earlier former observations (4).

Lung DC, like blood DC (18), show only limited expression of CD40, CD80, and CD86, whereas human epidermal DC (22, 23) and murine spleen or skin DC (24) do not express CD80 but instead constitutively express a significant level of CD86. Similarly, murine lung DC are said to express CD86 but a subset of this population seems to express CD80. This change in phenotype is perhaps induced by the use of latex beads to isolate DC from macrophages in this study (25). Cryostat sections of mice tissues (24) confirm the low expression of CD80 and CD86; their staining was weak or negative in most nonlymphoid organs but strong in selected sites in lymphoid organs, especially in the T-cell regions of the spleen. Although it appears that the expression of both B7 molecules can be upregulated on mice DC isolated from spleen, or from skin after 1 to 3 d in vitro culture without intentional stimulation (24), our results and others show that DC in nonlymphoid tissue are in a so-called immature phenotype.

Although lung DC express low levels of costimulatory molecules, lung DC shared a strong capacity, like mature DC, to stimulate T-cell proliferation. The role of CD80 and CD86 on lung DC was demonstrated by using blocking antibodies. Indeed anti-CD80 or anti-CD86 mAb both substantially decreased T-cell proliferation in MLR between lung DC and allogeneic T cells. These results are in agreement with Scheinecker and associates (18), who showed that either anti-CD80 or anti-CD86 mAb efficiently inhibits T-cell proliferation in the allo-MLR; these findings, however, contrast with those reported by others. For instance, Masten and coworkers (25), using lung DC from mice, stated that T-cell proliferation was not affected with anti-CD86 mAb but was blocked with anti-CD80 mAb. Inaba and associates (24), using splenic or epidermal DC from mice, and Rattis and coworkers (23), using human epidermal DC, reported that anti-CD86 mAb but not CD80 mAb inhibited the allogeneic T-cell proliferation. The results of Rattis and associates could be explained by a lack of blocking activity of the antibody used. Indeed, when we tested the anti-CD80 mAb used by Rattis and colleagues, this latter did not block the lung DC-induced MLR (data not shown). Thus, the nature of the mAb appears to be an important factor in the interpretation of the various studies. However, several studies demonstrated that both anti-CD80 and anti-CD86 mAbs have an effect, underlining the importance of B7 molecules (24, 25).

In vitro, the ligation to CD40-CD40L can increase the expression of CD80 and CD86 and also the production of IL-12. These effects lead to a maturation of immature DC and to a stronger T-cell stimulation (26). A similar function of CD40 has been shown on human epidermal cell (16) and on mouse lung DC (25). Surprisingly, the same CD40 mAb clone inhibiting the MLR induced by human epidermal DC (16) had no effect on the MLR induced by human lung DC. This could imply that enough costimulatory signals were already present when the experiments were performed. Lung DC have few CD80 and CD86 molecules on their surfaces, implying that they are rather like immature DC in lung tissue and may potentially change their phenotype-like blood-derived DC as soon as they are in the vicinity of inflammatory signals. The in vitro culture conditions have been shown to enhance MHC Class II expression and the APC activity of rat lung DC (27). Similar activation may have occured in our isolation procedure of human lung DC. However, lung DC maintained overall an immature phenotype, like freshly isolated rat DC seems to have done (28). This strong capacity of human lung DC to stimulate an MLR despite the low expression of CD80, CD86, or CD40 may suggest either a rapid upregulation of these molecules during the MLR or the presence of other costimulatory signals that are yet unknown, perhaps obtained during their isolation process. This latter hypothesis may be supported by the only partial inhibition obtained when using the potent inhibitor CTLA4 Ig, which blocked both CD80 and CD86 in our experiments.

The function of DC is intimately connected to their capacity to migrate. Their maturation results in a downregulation of receptors for inflammatory chemokines and upregulation of receptors for chemokines produced in secondary lymphoid organs (16). Immature blood-derived DC express both CCR1 and CCR5 (16). These cells exhibit a migratory response to regulated on activation, normal T cells expressed and secreted (RANTES), which is suppressed significantly with mAb to CCR1 but slightly with mAb to CCR5 (27). It must be noted that inflammatory chemokines such as IL-8, monocyte chemotactic protein-1, and in particular RANTES, are constitutively expressed by epithelial cells, where immature DC accumulate (16). Lung epithelia may thus attract lung DC as they express both CCR1 and CCR5, whereas mature DC migrating toward lymphoid structure no longer express these two receptors. Sato and colleagues (29), however, described how DC treatment with anti-CCR1 mAb but not anti-CCR5 mAb markedly inhibits allogeneic T-cell response, including proliferation and interferon- secretion, via suppression of chemotactic migration between T cells and DC. Thus, the expression of chemokine receptors is tightly regulated to facilitate their migration and their interaction with T cells. The low presence of CCR1 and CCR5 on the lung DC may be important to keep them near the epithelia while still allowing them to migrate toward lymphoid structure under the proper stimuli.

The capacity of DC to uptake and process antigens is highly dependent on the stage of differentiation of DC. Immature DC have the capacity to phagocytose particles and microbes (30). To do so they express receptors that mediate adsorptive endocytosis including the mannose receptor as well as Fc and Fc receptors (10, 17, 19). The mannose receptor is the most often found, if not the exclusive, receptor involved in endocytosis of FITC-dextran by DC. Inflammatory stimuli result in loss of capturing machinery and in an increase of T-cell stimulatory function. Lung DC have an ability to endocytose dextran, which is in-between immature and mature blood-derived DC. Lung DC uptake dextran and the number of molecules internalized, as judged by the MFI, is rather high. In addition, lung DC, like monocytes and blood DC, express CD64, a receptor important in triggering the internalization of antigen-IgG complexes (19). Because the level of phagocytic ability is correlated with the surface expression of CD64 (19), this function is probably of some importance for lung DC. The CD64 expression is downregulated by IL-4, and this may explain why it cannot be detected on the immature and mature blood-derived DC. Further, immunohistochemical studies in normal skin have revealed that CD64 stains only few DC and individual cells in the interstitial areas of lung section (31).

We have demonstrated that human lung DC present an immature phenotype with a rather good endocytic capacity, through their mannose receptor, while maintaining a low CD64 expression. They are, however, already potent stimulators of allogeneic T cells. The role of infections or inflammatory signals on the evolution and function of human lung DC remains to be determined.