Production of IL-12 by Human Monocyte-Derived Dendritic Cells Is Optimal When the Stimulus Is Given at the Onset of Maturation, and Is Further Enhanced by IL-4¹

Dendritic cells are APCs specialized to initiate primary immune responses (1, 2). Several well-developed functional properties enable them to successfully fulfill this task. The generation of immunogenic MHC/peptide complexes from protein Ags is efficiently done by immature dendritic cells. Signals delivered by inflammatory cytokines set off a maturation process whereby, in vivo, dendritic cells migrate to the T cell areas of draining lymphoid organs. There they present antigenic peptides to naive T cells that are passing by in large numbers and select and bind the Ag-specific T cells from this circulating pool. The interactions of MHC/peptide and TCR and of costimulatory molecules with their counterreceptors lead to the activation of T cells that, in turn, results in their proliferation and cytokine synthesis. An additional crucial factor in the moment of dendritic cell/T cell interaction in the lymphoid organ is the cytokine IL-12. This heterodimeric cytokine (consisting of one p35 and p40 chain, each) critically regulates the balance between Th1 and Th2 responses (3): IL-12 potently induces IFN-γ-secreting Th1 cells.

Dendritic cells have repeatedly been shown to produce IL-12 both in an unstimulated state (4) and, in much larger amounts, when stimulated by either bacteria or bacterial products (5, 6), virus (7), or by ligation of their CD40 and/or MHC class II molecules (5, 8, 9). In the human system, this may only be true for dendritic cells of the myeloid lineage, so-called dendritic cells type 1 (10). Dendritic cell-derived IL-12 was functional in that it skewed primary T cell responses toward a Th1 pattern (4, 11). Most studies hitherto, except one recent report (12), have not specifically defined the maturational status of dendritic cells analyzed. This aspect is important, though, because it is the mature, T cell-activating dendritic cell in which IL-12 would presumably be most relevant for the generation of specific Th1 immunity. This prompted us to systematically study IL-12 production of dendritic cells as a function of their maturational state as well as of different maturation stimuli. In the light of recently reported feedback loops on IL-12 by Th2 cytokines (10), we have specifically investigated the influence of IL-4 and IL-10 on dendritic cell-derived IL-12. We emphasized the study of monocyte-derived dendritic cells, a population that is preferably used for immunotherapeutic approaches.

Culture medium used throughout was RPMI 1640 supplemented with 10% FCS (endotoxin <0.06 ng/ml), gentamicin (all obtained from PAA, Linz, Austria), and 2-ME (Sigma, St. Louis, MO). Alternatively, dendritic cell cultures were also set up in 1% autologous plasma, as described (13, 14).

Dendritic cells were generated from adherent mononuclear cells in human blood according to established standard procedures (13, 14). Blood cells were from freshly drawn blood or from buffy coats that were obtained from the local blood center. Briefly, an initial 7-day priming culture in the presence of GM-CSF (800 U/ml) and IL-4 (1000 U/ml) was followed by a 3-day differentiation culture in the additional presence of monocyte-conditioned medium (MCM).³ GM-CSF and IL-4 were still present during this period. In the majority of experiments, populations of immature dendritic cells were split in half on day 7 of culture. They were cultured for 3 more days in the presence or absence of MCM (13, 14). On day 10, cells were collected, and immature (i.e., those without MCM) and mature (i.e., those with MCM) dendritic cells were analyzed for IL-12 production in parallel. Alternatively, in a few experiments, immature dendritic cells on day 7 of culture were compared with mature dendritic cells on day 10 of culture, after maturation in the presence of MCM. Both types of immature dendritic cells were identical with regard to phenotype and IL-12 production. GM-CSF was obtained from Novartis (Basel, Switzerland; Leukomax, sp. act., 1.1 × 10⁶ U/mg), and IL-4 was purchased from Genzyme (Cambridge, MA; sp. act., 5 × 10⁷ U/mg). Alternatively, we used culture supernatant (5% v/v) of a cell line transfected with human IL-4 (IL-4-62) that was provided by Dr. A. Lanzavecchia (Basel, Switzerland).

Fixed Staphylococcus aureus Cowan I strain (SACS, 10 μg/ml Ig-binding capacity, Pansorbin cells, catalogue number 507861) was obtained from Calbiochem (La Jolla, CA). Murine myeloma cells transfected with the human CD154/CD40 ligand molecule (P3 × TBA7 cells) were used to ligate the CD40 molecule on the surface of dendritic cells (15). Wild-type cells served as negative control (P3 × 63Ag8.653-WT). Alternatively, we cross-linked CD40 with anti-CD40 mAbs G28-5 (gift of Dr. E. Clark, Seattle, WA (16)) and MAB089 (Immunotech-Coulter, Marseille, France) as well as with total and ultracentrifuged culture supernatants of CD40 ligand-transfected cells containing CD40 ligand bound to membrane fragments and soluble CD40 ligand, respectively (17). IL-10 (sp. act., 1 × 10⁷ U/mg) was a gift of Dr. Ann O’Garra (DNAX Research Institute, Seattle, WA). IL-4 was from Genzyme (see above).

Immature or mature dendritic cells were washed out (3×) of cytokine-containing culture media. They were counted under the hemocytometer and analyzed for CD83 expression by flow cytometry, and 1 × 10⁶ dendritic cells/ml were plated into 24-well or 48-well multiwell tissue culture plates in total volumes of 1 and 0.5 ml of culture medium, respectively. (mAb HB-15a, anti-CD83 was a gift of Dr. Thomas F. Tedder, Durham, NC; FITC-conjugated anti-CD83 was from Coulter-Immunotech, Marseille, France.) Supernatants were taken at 48 and 72 h and stored at −80°C until analysis by ELISA. For most experiments, we used a sandwich ELISA, which was generously provided by Drs. D. H. Presky and M. K. Gately from Hoffmann-LaRoche (Nutley, NJ; capture mAb, 20C2; detection mAb, peroxidase-conjugated 4D6). The exact protocol has been described previously (18, 19). Few experiments were analyzed by means of a commercial IL-12 ELISA (Quantikine; R&D Systems, Minneapolis, MN). The capture Abs used in both tests specifically recognize the p70 heterodimer, but not the free p40 chains. Detection limits were 20 pg/ml of IL-12.

PE-conjugated mouse mAb C11.5, directed against the p40 subunit of human IL-12, was used to stain saponin-permeabilized cell populations that had been stimulated for 30 h with CD40 ligand-expressing cells in the presence (last 5 h) of brefeldin A to achieve some accumulation of cytokine within the cell (20). All reagents and the staining protocol were from BD PharMingen (San Diego, CA).

Human IL-12 p40 and p35 mRNA was detected by PCR and liquid hybridization, as described previously (21). IFN-γ was measured with a commercial ELISA (BioSource-Medgenix, Fleurus, Belgium). Binding of mAb dendritic cell-lysosome-associated membrane glycoprotein (DC-LAMP) (mouse IgG1) (22) on acetone-fixed cytospins was visualized by a biotinylated anti-mouse Ig (Amersham-Pharmacia, Amersham, U.K.), followed by Texas Red-conjugated streptavidin (Amersham); after blocking of residual binding sites with an excess of mouse γ-globulin (100 μg/ml), dendritic cells were counterstained with an FITC-conjugated anti-HLA-DR mAb (clone L243; BD PharMingen, San Jose, CA). DC-LAMP was a gift of Dr. Serge Lebecque, Laboratory for Immunological Research, Schering-Plough (Dardilly, France). Neutralizing mAbs against human IFN-γ (clone B27) and isotype-matched control Abs were purchased from BD Phar-Mingen and used at final concentrations of 20 μg/ml.

In preliminary experiments, we found no difference in IL-12 protein levels between 48- and 72-h incubation periods. Therefore, the 48-h time point was used for all further ELISA analyses. IL-12 p70 heterodimer secreted by unstimulated dendritic cells (4) was almost always below the threshold of detectability of the ELISA. As another preliminary, CD40 expression was comparatively assessed on immature and mature dendritic cells and found to be similarly high on both populations (Fig. 1). This contrasted with the expression of CD83 that clearly distinguished the two maturational states (Figs. 1 and 2 A). In addition, the up-regulated expression of CD86, the lack of CD115 expression, and a pronounced veiled morphology under phase contrast were used as markers for mature dendritic cells (data not shown). It should be emphasized in this work that we use “immature dendritic cell” as the term to describe a monocyte differentiated for 6–7 days in the presence of GM-CSF and IL-4 (23, 24). It is clearly more advanced than the early (25) immature dendritic cell such as a Langerhans cell in situ, in that it has already sizeable levels of surface MHC class II molecules and it expresses high levels of CD40. Yet, it still requires maturation stimuli to further differentiate and to acquire those features that characterize the terminally mature dendritic cell (13, 14): 1) no reversion back to a macrophage, 2) greatly augmented T cell-stimulatory capacity for MLR and CTL, and 3) de novo expression of markers such as CD83 or DC-LAMP (22). These features are similar to those of dendritic cells that matured spontaneously from blood after 2 days of culture (26) or of cultured epidermal Langerhans cells (27, 28).

In 11 independent experiments, IL-12 values from immature dendritic cells ranged between 2.1 and 0.02 ng/ml; values from corresponding populations of mature dendritic cells ranged from 0.5 to 0 (i.e., below detection threshold) ng/ml. However, in all experiments, immature dendritic cells produced higher levels than mature dendritic cells (Fig. 3). It is of note that in 7 of 11 populations of mature dendritic cells, IL-12 in the supernatants was below the level of detection, i.e., <20 pg/ml. Analysis of data by means of the two-sample t test showed that the differences between immature and mature dendritic cells were statistically significant (p < 0.05).

SACS has been described as a potent maturation stimulus for dendritic cells (14). FACS analyses revealed that also under the very conditions of the IL-12 assays (i.e., 1 × 10⁶ cells/ml, 48-h incubation), maturation occurred as detected by the induction of CD83 expression (Fig. 2 A).

Initially, we tested the conditions for cross-linking CD40 by means of CD40 ligand-expressing cells (TBA7 cells). CD40 ligand expression of TBA7 cells was high; it was clearly more than the levels reported for activated T cells (data not shown). Maximal IL-12 release by dendritic cells was achieved at a ratio of one TBA7 cell to two dendritic cells. This ratio was kept for all additional experiments. When proliferation of TBA7 cells was stopped by irradiation with 15–30 Gy from a Cs source, they elicited considerably lower amounts of IL-12 from dendritic cells (data not shown). Therefore, viable TBA7 cells were used in all assays.

In 11 independent experiments, IL-12 values from immature dendritic cells ranged from >100 to 1.26 ng/ml; values from corresponding populations of mature dendritic cells were between 29 and 0.11 ng/ml. Like with SACS stimulation, immature dendritic cells produced higher levels than mature dendritic cells in all experiments (Fig. 4). However, in contrast to SACS stimulation, populations of mature dendritic cells did elaborate substantial amounts of IL-12 in all experiments, the lowest concentration measured being 110 pg/ml. Analysis of data by means of the two-sample t test showed that the differences between immature and mature dendritic cells were statistically significant (p < 0.05). These data were confirmed with dendritic cells that had been cultured in the presence of 1% autologous human plasma (n = 22) rather than 10% FCS, as in the experiments reported above (Fig. 4). IL-12 production in plasma-supplemented cultures was lower than in FCS-containing media, though. The down-regulation of IL-12 production upon maturation of dendritic cells did not only occur when MCM was used as a maturation stimulus, but also when a defined cytokine cocktail consisting of IL-1β, IL-6, TNF-α, and PGE₂ (29) was employed: Experiment 1, 19 ng/ml in immature vs 1.3 ng/ml in mature dendritic cells; experiment 2, 1 vs 0.5 ng/ml. Stimulation of dendritic cells with control wild-type cells was consistently negative. Addition of a mAb against the CD40 ligand completely abrogated IL-12 induction (data not shown).

Ligation of CD40 can induce maturation of dendritic cells (5). FACS analyses revealed that also under the very conditions of the IL-12 assays (i.e., 1 × 10⁶ cells/ml, 48-h incubation), maturation occurred in the presence of CD40 ligand-expressing cells, as detected by the expression of CD83 (Fig. 2,A) as well as of intracellular DC-LAMP (22) (Fig. 2,B). It is also noteworthy that CD83 of already mature dendritic cells (on day 10) remained stably on the surface during the 2-day duration of the stimulation culture, even in controls in which no stimulus was added (Fig. 2 A).

Ligation of CD40 by anti-CD40 mAbs yielded inconsistent results. mAb G28-5 did not induce IL-12 in some experiments; in some it did. In those experiments, the same phenomenon was noted: immature dendritic cells made more IL-12 than mature dendritic cells in response to the Ab (immature vs mature dendritic cells, IL-12 p70 in pg/ml: Expt. 1, >500 vs 52; Expt. 2, 456 vs 46). MAB089 never induced IL-12. Likewise, total and ultracentrifuged culture supernatants of CD40 ligand-transfected cells, containing CD40 ligand bound to membrane fragments and soluble CD40 ligand, respectively (17), were inactive in our assays.

Finally, we wished to test the possibility that immature and mature dendritic cells made similar amounts of IL-12, but that populations of mature dendritic cells contained proteases that would efficiently degrade the cytokine. Therefore, rIL-12 was added to immature and mature dendritic cells and incubated for 48 h. Supernatants analyzed by ELISA showed no substantial degradation of IL-12 in either population (data not shown).

Dendritic cells emigrated from whole skin explants were also tested for their ability to produce bioactive IL-12. These cells are a mixture consisting of epidermal Langerhans cells and dermal dendritic cells. All dendritic cells within these populations were fully mature, as previously shown (30, 31), and as monitored by morphology under phase contrast and by CD83 and CD86 expression (data not shown). In three independent experiments, SACS did not induce any measurable IL-12 p70. In two different experiments using CD40 ligand-transfected cells as stimulus, IL-12 p70 production was also below the threshold of detection.

Because the stimuli that brought about IL-12 production (CD40 ligation and SACS) also induced maturation, we wondered whether MCM, the classical stimulus for maturation (13, 14, 32), would also do so. Immature dendritic cells on day 7 or 10 were assayed for IL-12 in the presence or absence of MCM. In six independent experiments, virtually no IL-12 was induced by MCM; however, dendritic cells from parallel cultures that were stimulated with CD40 ligand or SACS did elaborate the cytokine (Fig. 5). FACS analyses (Fig. 2 A) proved that MCM rendered dendritic cells stably mature also under the specific culture conditions used for collecting supernatants for ELISA (i.e., 1 × 10⁶ cells/ml; 48 h).

It should be noted that in supernatants from standard maturation cultures (i.e., from day 7–10 in six-well plates at a cell density of 0.7 × 10⁵ cells/ml in the presence of 33% MCM), IL-12 was found only in 4 of 47 cultures, albeit at a low concentration.

From Figs. 1 and 4 it can be read that CD40 ligation is the strongest of the three stimuli tested. This becomes more apparent when the same data are plotted as side-by-side comparisons within different individual experiments (Fig. 5). When CD40-induced IL-12 production of immature dendritic cells is set equal to 100%, SACS elicits on average about one-fifth of this amount (19.9 ± 37%; range 0.1–139%; n = 16). MCM induce only very little IL-12 (4.4 ± 6, 9%; range 0.1–17, 9%; n = 6).

Intracellular FACS staining of CD40 ligand-stimulated dendritic cell populations using a mAb against the p40 subunit of IL-12 showed unequivocally that IL-12 had accumulated in CD83⁺ cells (Fig. 6). This indicated that IL-12 synthesis and maturation had proceeded simultaneously. It also means that the high levels of IL-12 are not produced by immature, but rather by maturing dendritic cells. FACS analyses also confirmed that already mature dendritic cells made less IL-12 than maturing dendritic cells.

Next we examined whether the differential secretion of IL-12 heterodimer protein in immature and mature dendritic cell was due to de novo synthesis. To this end, the expression of mRNA for the p35 and p40 subunits of IL-12 was investigated. A semiquantitative PCR analysis revealed that, in response to ligation of CD40, immature dendritic cells expressed more mRNA for both p35 and p40 than mature dendritic cells (Fig. 7). This was not as pronounced when SACS was used as a stimulus for IL-12 production.

IFN-γ has been described as necessary costimulus for IL-12 production by dendritic cells (33). Therefore, this cytokine was measured in parallel with the IL-12 assays. In a series of 13 independent experiments (in medium containing autologous plasma), a significant (r = 0.81) positive correlation between the values of IL-12 and IFN-γ became apparent (data not shown): Most cultures of CD40-stimulated immature dendritic cells contained more IFN-γ (maximum, >60 ng/ml IFN-γ) than cultures of CD40-stimulated mature cells (maximum, 9.2 ng/ml IFN-γ). Few experiments with SACS-stimulated dendritic cells revealed the same correlation (data not shown). It was not further investigated whether this cytokine was produced by dendritic cells or by few contaminating T cells. NK cells were ruled out as producers of IFN-γ: In five independent FACS analyses, we detected virtually no CD56⁺ cells, i.e., NK cells.

When IL-4 was present during the stimulation of dendritic cells with CD40 ligand-expressing cells, a marked increase of IL-12 secretion was observed. Up to almost the 10-fold amount of IL-12 was induced by IL-4. This observation was made both in FCS-containing cultures (Fig. 8) and in cultures with 1% autologous plasma: 698, 108, 995, 961, 139, 546, and 356% for immature dendritic cells in seven experiments, and 125, 438, 183, and 243% for mature dendritic cells in four experiments; IL-12 production in the absence of IL-4 was set equal to 100%. Thus, IL-4 enhanced IL-12 production irrespective of the state of dendritic cell maturation. IL-4 did not alter the degree of maturation of dendritic cells, as determined by morphology under phase contrast, CD83 expression by FACS (data not shown), and DC-LAMP expression on cytospins (Fig. 2,B). In a series of five independent experiments, we found that IL-4 did not lead to increased amounts of IFN-γ in the cultures, but to clearly enhanced levels of IL-12 (698, 108, 995, 961, and 139%). Conversely, the neutralization of IFN-γ with a mAb did not prevent the IL-4-induced augmentation of IL-12 (Table I).

IL-10 was shown to inhibit IL-12 synthesis in murine dendritic cells (8). In this study, we investigated the effects of IL-10 on CD40-induced IL-12 production in human dendritic cells. In six independent experiments, a concentration of 10 U/ml (i.e., 1 ng/ml) IL-10 did not consistently inhibit IL-12 production. When 100 U/ml (i.e., 10 ng/ml) IL-10 was present during the 48-h stimulation with CD40 ligand, a clear-cut reduction of IL-12 secretion was observed (Fig. 8). The mean reduction with 100 U/ml IL-10 was 61% for populations of immature dendritic cells, and 66% for mature dendritic cells. Thus, dendritic cells at both states of maturation were inhibited to a similar degree by the high dose of IL-10.

In this study, we further dissect the intricate regulation of IL-12 as a function of dendritic cell maturation (immature vs maturing vs terminally mature; see below). We find that human dendritic cells produce the bioactive IL-12 p70 heterodimer most abundantly when they begin to respond to certain maturation stimuli. They reduce this powerful capacity as maturation proceeds, as described recently (12). We extend these data in several regards, the most important ones being that 1) surprisingly, the Th2 cytokine IL-4 markedly enhances IL-12 synthesis and secretion by both immature and mature dendritic cells; 2) this phenomenon occurs also in the clinically relevant culture system with autologous plasma (instead of FCS); 3) down-regulation of IL-12 production is also observed with skin-derived dendritic cells (complementing one recent report dealing specifically with Langerhans cells (34)); and 4) only select maturation stimuli (CD40, bacteria) can induce IL-12 production in dendritic cells.

We observed that CD40 ligation and (less though) bacteria induced massive IL-12 p70 production in immature dendritic cells. In contrast, MCM, the classical maturation stimulus (13, 14), did not induce substantial IL-12 production when applied in an identical experimental setting as the other stimuli (nor did the combination of TNF-α and PGE₂ in two experiments). This latter combination was reported to induce IL-12 p40 secretion, though (35). Thus, of the three stimuli tested in this study, two induced maturation and IL-12 secretion (CD40 ligation and bacteria), whereas one (MCM) led to maturation without concomitant IL-12 secretion. This is similar to the findings of Cella et al. (5), who noted IL-12 p70 induction in dendritic cells only with CD40 ligation and to some degree with viral infection (7), but not with other stimuli such as LPS or TNF-α. The discrepancy with regard to bacterially induced IL-12 (in this study, good IL-12 induction; Cella et al. (5), no IL-12 induction) may be due to different reagents (staphylococci, FCS). In vivo it would be advantageous if a dendritic cell made the potent cytokine IL-12 only if threatened by microbes or if in physical touch with T cells, rather than in response to any inflammatory cytokine milieu.

In vivo (e.g., in the epidermis), dendritic cells receive maturation and migration stimuli by inflammatory cytokines (36), often in the absence of microbes. In that case, IL-12 is needed only when dendritic cells have arrived in the lymph nodes and interact with T cells. Maturation in the presence of MCM may be regarded as an in vitro equivalent for this case. In a scenario in which microbes are present, high levels of bacterially induced dendritic cell-derived IL-12 could be beneficial in that they would contribute to the inflammation (activation of NK cells, maintenance and enhancement of Th1 state of infiltrating Th cells, and, as a consequence, macrophage activation) and thus help with the clearance of microorganisms. Inflammation (e.g., in the skin) might be further fueled by high levels of IL-12 derived from the interaction of CD40-expressing migrating dendritic cells that encounter effector or memory T cells expressing CD40 ligand. In vivo examples underscore our in vitro data: Dendritic cells in the spleen of mice exposed in vivo to Toxoplasma Ags respond with vigorous IL-12 production (6). The initial strong IL-12 immunostaining was predominantly found at the edge of the T cell area, but also in the periarteriolar region, implying that these dendritic cells were not fully mature in situ. Similarly, Leishmania-infected dendritic cells in the spleen produce IL-12 in situ (37).

Upon encounter with Ag-specific T cells in the T cell areas of lymphoid organs, IL-12 is crucial for the establishment of a Th1 response. When a dendritic cell that arrives in the lymph node via the afferent lymphatics finds and binds an Ag-specific T cell in the T areas (38), it signals to the resting naive T cell via its MHC/peptide complexes (signal 1) and costimulatory molecules (signal 2). T cell activation ensues and within a period of few hours, activated T cells up-regulate CD154 (CD40 ligand) expression. CD154, in turn, engages with CD40 on the surface of the dendritic cell and now signals flow in the inverse direction: The T cell induces terminal maturation (i.e., further up-regulation of costimulatory molecules) and IL-12 production in the dendritic cells (39, 40). This determines the default pathway of dendritic cells to induce Th1 responses. Probably in the microenvironment of the intracellular spaces in lymph nodes or spleen, small quantities of IL-12 might suffice to reach biologically active concentrations. It should be emphasized that IL-12 production was found to be markedly down-regulated in mature dendritic cells; it did not completely disappear, though. Substantial quantities (up to some hundred pg/ml) were still made by mature dendritic cells in most experiments. Higher concentrations of IL-12 may even be harmful.

The presence of IL-4 during the stimulation period strongly increased the levels of IL-12 in response to CD40 ligation. The reason for this somewhat unexpected finding may be a previous conditioning of dendritic cells during the 7-day culture in the continuous presence of IL-4, perhaps similar, but clearly not identical with what was described by D’Andrea et al. (41) as priming: These authors had observed an increased IL-12 production by IL-4-pretreated PBMCs in response to bacteria. We observed an increased IL-12 production by IL-4-pretreated dendritic cells in response to CD40 ligation plus IL-4. Thus, the otherwise IL-12-inhibiting cytokine IL-4 turned out to be IL-12 enhancing when the cells had been pretreated (conditioned) with IL-4. From very recent data by Hochrein et al. (42), who used dendritic cells that had been generated in the absence of IL-4, it appears that the IL-12-enhancing effect of IL-4 does not depend on a prior exposure to the same cytokine. A similar observation was made with CD40 ligand-stimulated murine dendritic cell-containing populations by Takenaka et al. (43). IFN-γ appears not to mediate the IL-4 effect, because its production was not induced by IL-4 in our hands, and moreover, IL-4 has typically been described to inhibit IFN-γ rather than enhancing it (44, 45). In addition, neutralization of IFN-γ in the stimulation assays did not prevent the IL-4-induced increase in IL-12 production. This is in line with a recent report by Kalinski et al. (46), who observed the same phenomenon induced by IL-4 derived from a Th2 clone that was deficient in IFN-γ production. IL-4 also does not act by influencing the maturation status of dendritic cells: maturation markers CD83 and DC-LAMP were up-regulated in response to CD40 ligation, irrespective of IL-4 treatment. Although IL-4 can augment IL-12 production in such a potent way, it is not a prerequisite for the large amounts of IL-12 made by dendritic cells. This was first concluded by Cella et al. (5), who showed high levels of CD40-induced IL-12 in freshly isolated dendritic cells that had never encountered IL-4 in vitro. It is underscored by the data of Koch et al. (8): murine spleen dendritic cells were induced to make large amounts of IL-12 p70 by cross-linking with anti-CD40 mAb. These dendritic cells also never had contact with IL-4 during their generation.

When exogenous IL-4 is added to cocultures of APCs and T cells, a Th2 response (i.e., IL-4-producing T cells) is the consequence (47). If a Th2 response, e.g., to fungal hyphae (48) or to helminthic parasites (49), occurs in the environment of a lymph node, one might expect that the resulting T cell-derived IL-4 would skew all other ongoing or beginning immune responses toward a Th2 pattern. Our data would indicate that there may be some balancing mechanism ensuring that Th1 responses are not necessarily suppressed in an IL-4-rich milieu. This finding seems important for immunotherapy (see below). While our work was in the final stages of review, Hochrein et al. (42) reported that IL-4, and even more so IL-4 plus IFN-γ, enhance the IL-12 p70 production of murine and human dendritic cells, and Kalínski et al. (46) found that IL-4 secreted by Th2 cells mediates high level IL-12 p70 production by immature human dendritic cells. We confirm these data and extend them in that we show the effect of IL-4 in an FCS-free culture system relevant for adoptive immunotherapy and for populations of mature monocyte-derived dendritic cells, i.e., those dendritic cells that are able to prime naive T cells and are preferentially used in immunotherapy (50, 51).

The inhibition of IL-12 synthesis by IL-10 in mature dendritic cells was also somewhat unexpected. Steinbrink et al. (52) and Thurner et al. (53) have demonstrated that mature dendritic cells are resistant to the effects of IL-10 in the MLR. Three explanations are conceivable. First, it is possible that the dose of 100 U/ml (i.e., 10 ng/ml) of IL-10 is unphysiologically high and already toxic. We have not further explored this possibility, except for simple trypan blue staining of cell populations at the end of the 48-h IL-12 assay. However, by this criterium, no IL-10-induced toxicity was detected. The low dose of IL-10 (1 ng/ml) did not consistently inhibit mature dendritic cells, the average inhibition being −22%. Second, one may assume that the amounts of IL-12 that were measured in populations of mature dendritic cells were derived from few, still immature or maturing dendritic cells that were still susceptible to inhibition by IL-10. Third, IL-10 may have differential effects on mature dendritic cells. We observed in this study that the same dose of 100 U/ml of IL-10 did not affect the phenotypical (CD83, CD86 expression) and morphological (nonadherence, veils) characteristics of mature dendritic cells, whereas in parallel cultures it inhibited IL-12 secretion, as described. This might shift the Th1/Th2 balance in ensuing T cell responses toward Th2. Neither IL-12 nor the IFN-γ/IL-4 balance of resulting T cell responders was measured in previous work pinpointing the stability of mature dendritic cells (52, 53). This hypothesis is underscored by our previous finding with a population of classically mature dendritic cells, namely mouse spleen dendritic cells (8): IL-10 totally blocked IL-12 p70 secretion. When mature spleen dendritic cells were used to repetitively stimulate allogeneic T cells, the presence of IL-10 led to the development of a Th2 pattern of T cell cytokines (F. Koch, personal communication).

Dendritic cells have been widely used and (successfully) tested in animal models of tumor therapy, and a number of clinical trials are currently running (e.g., 51, 54, 55). Monocyte-derived mature dendritic cells are often used as a convenient source of large numbers of human dendritic cells (53). Three sets of data from our experiments may be of relevance in a clinical setting: 1) Our finding that mature dendritic cells were less responsive to CD40 ligation (i.e., T cell interaction) in terms of IL-12 production seems counterproductive at first glance. Yet, it is likely that the small amounts of IL-12 still produced by mature dendritic cells will suffice for Th1 skewing within the microenvironment of the lymph nodes. 2) Our observation that IL-10 inhibits CD40-induced IL-12 production in dendritic cells should alert us that under circumstances of high IL-10 levels in the body, e.g., in tumor situations (56), dendritic cell therapies might be impaired (57) and might need adjuvant treatment such as cytokines. Induction of anergy in melanoma-specific CTL by IL-10-pretreated immature (i.e., during the maturation culture) dendritic cells was recently demonstrated (58). 3) Finally and most importantly, the fact that dendritic cells make much more IL-12 when IL-4 is present seems encouraging for strategies in which a predominant Th1 response is desired, for example, therapy of tumors or microbial infections. It seems conceivable that Ag-pulsed dendritic cells that arrive in a lymph node with an IL-4-rich milieu (e.g., atopic state) would still be able to skew a T cell response toward a Th1 pattern, perhaps even better. Additionally, IL-4 may allow for the development of Th2 mechanisms that also appear to be critically involved in tumor immunity (59, 60), without inhibiting therapeutically administered dendritic cells.

We thank Susanne Neyer for generously providing her practical expertise and skills in molecular biology, Hella Stössel for immunocytochemistry, Karin Salzmann for help with ELISAs, and Dr. Franz Koch for critical discussions.