Direct Effects of Polymyxin B on Human Dendritic Cells Maturation

Polymyxin B is a lipopolysaccharide binding antibiotic used to inactivate potential lipopolysaccharide contaminations when evaluating the activity of different agents on innate immune cells. We report that polymyxin B is able to induce directly in monocyte-derived human dendritic cells (DCs) several functional and molecular modifications characteristic of DCs undergoing a maturation process. DCs incubated with polymyxin B up-regulate the expression of HLA class I and II, the co-stimulatory CD86 molecule, and show an increase in the fraction of adherent cells at short time, which persist at 48 h of incubation. Adhesion to the plate was required for the polymyxin B-induced DCs maturation. A transient activation of IκB-α/NF-κB and ERK1/2 pathways at short time and a further ERK1/2 activation at long term were also detected. Neither up-regulation of the maturation marker CD83 nor activation of p38 nor induction of cytokines secretion was observed in DCs treated with polymyxin B. We demonstrated that inhibition of IκB-α/NF-κB pathway abolishes polymyxin B effects. ERK1/2 inhibition instead allowed DCs treated with polymyxin B to progress in their maturation process as revealed by the increased up-regulation of the CD83 co-stimulatory molecules, the activation of p38, and the reduced adhesion to culture plates at 48 h of incubation. Our results indicate that polymyxin B induces a partial maturation of human DCs through increased adhesion to a substrate and activation of the IκB-α/NF-κB pathway. The increased ERK1/2 activation observed, even though correlating with the initial phases of the maturation process, actually inhibits the occurrence of full maturation.


Polymyxin B is a lipopolysaccharide binding antibiotic used to inactivate potential lipopolysaccharide contaminations when evaluating the activity of different agents on innate immune cells. We report that polymyxin B is able to induce directly in monocyte-derived human dendritic cells (DCs) several functional and molecular modifications characteristic of DCs undergoing a maturation process. DCs incubated with polymyxin B up-regulate the expression of HLA class I and II, the co-stimulatory CD86 molecule, and show an increase in the fraction of adherent cells at short time, which persist at 48 h of incubation. Adhesion to the plate was required for the polymyxin B-induced DCs maturation. A transient activation of IB-␣/NF-B and ERK1/2 pathways at short time and a further ERK1/2 activation at long term were also detected. Neither up-regulation of the maturation marker CD83 nor activation of p38 nor induction of cytokines secretion was observed in DCs treated with polymyxin B. We demonstrated that inhibition of IB-␣/NF-B pathway abolishes polymyxin B effects. ERK1/2 inhibition instead allowed DCs treated
with polymyxin B to progress in their maturation process as revealed by the increased up-regulation of the CD83 co-stimulatory molecules, the activation of p38, and the reduced adhesion to culture plates at 48 h of incubation. Our results indicate that polymyxin B induces a partial maturation of human DCs through increased adhesion to a substrate and activation of the IB-␣/NF-B pathway. The increased ERK1/2 activation observed, even though correlating with the initial phases of the maturation process, actually inhibits the occurrence of full maturation.
Dendritic cells (DC(s)), 1 the most efficient antigen-presenting cells of the immune system, are present in almost all tissues in an immature state and are characterized by a high ability for antigen uptake and processing (1,2). The exposure to inflammatory stimuli (e.g. TNF-␣, interleukin-1␤) or microbial agents (e.g. lipopolysaccharide (LPS)) is able to induce in DCs a complex maturation process that enhances their antigen presentation ability to T cells (2)(3)(4). Hallmarks of this maturation process are the up-regulation of cell surface major histocompatibility complex class I and II and co-stimulatory molecules (e.g. CD40, CD80, CD86, CD83). Moreover, several cytokines as IL-12, TNF-␣, and IL-10 are released by DCs during maturation.
Migration of mature DCs into lymphoid organs, where antigen presentation to T cells occurs, is achieved by changes in expression of chemokine surface receptors (2,5) and requires modulation of DCs adhesion and motility properties (5,6). Human DCs induced to mature by TNF-␣ or LPS up-regulate CD49d integrin expression that parallels that of the CD83 marker (7). Some studies suggest the involvement of extracellular matrix proteins in DCs maturation. In fact, DCs of murine or human origin up-regulate co-stimulatory molecules and secrete cytokines when grown on collagen type 1 in the absence of any other stimulus (8,9).
The signal transduction events underlying the DCs maturation process have been only recently investigated. Two mitogen-activated protein kinases, the extracellular signal-regulated protein kinases (ERK1/2) and the p38 stress-activated protein kinase (p38), are activated by TNF-␣ or LPS in both murine and human immature DCs (10 -13). The nuclear factor (NF)-B also play a central role in the maturation process induced by LPS or TNF-␣ of both murine (14) and human DCs (10,11). The transcription factor NF-B is present in the cytosol in an inactive form bound to the inhibitory IB-␣ protein.
Phosphorylation and subsequent degradation of IB-␣ results in the release and nuclear translocation of active NF-B transcription factor (15,16). The availability of specific inhibitors for ERK1/2, p38, and IB-␣/NF-B pathways allowed for evaluation of their respective role in the maturation process. Several studies have concordantly shown that p38 inhibition by SB203580 prevents the occurrence of DCs maturation, inhibiting the up-regulation of several maturation markers and cytokines release (10 -12). IB-␣/NF-B pathway inhibition also blocks the maturation process in DCs stimulated with LPS (10,14). More contradictory are the results obtained with the ERK1/2 inhibitor PD98059. Some reports have shown that ERK1/2 inhibition does not affect any of the LPS-induced maturation parameters analyzed (10,12), although other studies have demonstrated that ERK1/2 inhibition enhances TNF-␣ and LPS-induced phenotypic and functional maturation (11,13). In support of a negative role of ERK1/2, it has been recently reported that IL-12 production by DCs stimulated with LPS or CD40L is increased by inhibition of ERK1/2 (17,18).
In vitro maturation of DCs represents a valid system to evaluate the potential immunomodulating properties of chemical compounds and purified proteins (12,19). In this contest, the LPS-binding polypeptide polymyxin B (20) has been widely used as LPS-neutralizing reagent to avoid the possible inter-ference of LPS contaminations (19,21,22). It is generally assumed that polymyxin B does not have any direct activity on innate immune cells. Nevertheless, polymyxin B has been also used as an inhibitor of protein kinase C in eukaryotic cells (23,24). Other reports suggest that polymyxin B is a Ca 2ϩ -activated K ϩ channel inhibitor (25) and a stimulator of mitochondrial sn-glycerol-3-phosphate acyltransferase (26).
In this article, we studied the possible direct effects of polymyxin B on the maturation process of human monocyte-derived DCs. In our experimental system, polymyxin B was able to induce a partial DC maturation characterized by up-regulation of early co-stimulatory molecules and increased adhesion properties. This partial maturation was adhesion-and IB-␣/NF-B pathway-mediated and correlated with ERK1/2 activation, which actually inhibited the progress of DCs into full maturation.

MATERIALS AND METHODS
Reagents-Bacterial LPS from Salmonella abortus equi (Sigma-Aldrich, St. Louis, MO) was used at a final concentration of 100 ng/ml. Polymyxin B was purchased from Sigma-Aldrich, dissolved in distilled water, and used at a final concentration of 70 units/ml. PD98059 (Cell Signaling Technology, Beverly, MA) was dissolved in dimethysulfoxide (Me 2 SO) and used at a final concentration of 50 M. TPCK (Sigma-Aldrich) was dissolved in ethanol and used at a final concentration of 20 M. Granulocyte macrophage colony-stimulating factor and IL-4 used at a final concentration of 800 units/ml and 100 units/ml, respectively, were purchased from AL-ImmunoTools (Friesoythe, Germany). Fetal bovine serum was purchased by HyClone (South Logan, UT) and heat inactivated at 56°C for 60 min.
Generation of Monocyte-derived DCs-Human DCs were generated from the adherent fraction of peripheral blood mononuclear cells of healthy donors in the presence of granulocyte macrophage colony-stimulating factor and IL-4, as described previously (27). At day 5 of culture, only detached, floating cells were collected and used in experiments as immature DCs. These cells are in fact fully differentiated DCs as shown by their phenotype (CD14 negative and CD1a positive).
PolyHEMA-covered Plates-Culture plates (Corning Incorporated, Corning, NY) were coated with a film of polyHEMA (Sigma-Aldrich), as described previously (28,29). Culture plates were covered with 0.95 l per mm 2 of polyHEMA solution, left to dry at room temperature, and washed before use in RPMI 1640.
Quantification of Cell Adhesion and Flow Cytometric Analysis-Immature DCs were seeded at 10 5 cells/cm 2 in 0.5 ml of RPMI 1640 supplemented with 10% fetal bovine serum, 800 units/ml granulocyte macrophage colony-stimulating factor, 100 units/ml IL-4. To quantify cell adhesion, after 4 or 48 h of incubation cells in suspension were collected with two washes of phosphate-buffered saline, and adherent cells were detached by incubation in phosphate-buffered saline with 2 mM EDTA. Both suspended and adherent viable cells were counted at the microscope by trypan blue exclusion. The analysis of cell surface antigens was performed at 48 h on the whole population of suspended and adherent cells collected as describe above. Cells were incubated with the anti-human CD86-FITC, CD83-PE, CD40-FITC, CD80-PE, or HLA-DR-FITC (anti-HLA Class II)-conjugated antibodies (Pharmingen, BD Biosciences) for 30 min on ice. HLA Class I was detected by indirect immunofluorescence using the specific monoclonal antibody W6/32 for 30 min on ice. Appropriate isotype controls (Pharmingen) were used. Samples were analyzed using the Beckman-Coulter flow cytometer.
Immunoblots-At the indicated time points for each experimental condition, cells in suspension were collected, washed in ice-cold phosphate-buffered saline, centrifuged, and lysed along with the corresponding phosphate-buffered saline-washed adherent cells without detaching them from the plate. Cells were lysed, and equal amounts of total proteins were subjected to western immunoblotting as described previously (30). ERK1/2 and p38 activities or IB-␣ status were evaluated by immunoblotting using antibodies directed against specific phosphorylated amino acids present only in the active form of the proteins, according to the manufacturer's instruction (New England Biolabs, Beverly, MA). We used an anti-phospho-ERK1/2 (Thr 202 /Tyr 204 ) monoclonal antibody, an anti-phospho-p38 mitogen-activated protein kinase (Thr 180 / Tyr 182 ) polyclonal antibody, and an anti-phospho-IB-␣ (Ser 32 ) poly- Detection of Cytokines-Supernatants of cells incubated in the different experimental conditions were collected at 48 h and subjected to cytokine analysis. IL-10 and IL-12 were evaluated by enzyme-linked immunosorbent assay according to the manufacturer's instruction using couples of antibodies from Pharmingen and Endogen, respectively. TNF-␣ released by DCs in the supernatant was determined by testing its cytotoxic effect on WEHI-164.13 (31) in a 3-(4,5-dimethylthiazol-2yl)-2,5-diphenyltetrazolium bromide colorimetric assay.
Analysis of the Allostimulatory Potential of DCs-DCs untreated or treated as indicated for 48 h were used as stimulator in an allogeneic mixed leukocyte reaction. To evaluate lymphocytes proliferation, DCs were irradiated (5000 rads) and incubated in 96 flat-bottomed microplates at different densities (ranging from 5 ϫ 10 4 to 2 ϫ 10 2 cell per well) with a constant number of allogeneic adult peripheral blood mononuclear cells (1 ϫ 10 5 cells per well). On day 5, after a 18 h pulse with 3 H-thymidine (1 Ci/well, specific activity of 5 Ci/mmol, Amersham Biosciences), thymidine incorporation was measured. The level of 3 Hthymidine uptake by DCs alone was always the same and ranged between 300 -200 cpm. For the evaluation of interferon-␥ produced by activated lymphocytes, irradiated DCs were seeded in dilution (ranging from 5 ϫ 10 3 to 3 ϫ 10 2 ) with a constant number of allogeneic adult lymphocytes (2 ϫ 10 4 per well). After 72 h of incubation supernatants were collected and subjected to interferon-␥ detection by enzyme-linked immunosorbent assay (Genzyme Corp., Cambridge, MA) according to the manufacturer's recommendations.

Polymyxin B-induced Up-regulation of DC Maturation
Markers-To investigate whether polymyxin B, beside its LPS neutralizing activity, was able to induce any direct biological effect on human monocyte-derived DCs, we first evaluated the ability of polymyxin B to modify the expression of molecules that are up-regulated during DC maturation. After 48 h of incubation with polymyxin B (Fig. 1, A and B), DCs showed an increase in the expression of CD86 and HLA-class I and II molecules evaluated as the percentage of positive cells or as mean fluorescence of the whole population. On the contrary, CD40, CD80, and CD83 remained unchanged. DCs stimulated with LPS were fully able to up-regulate both CD86 and CD83 (Fig.  1A) and all the other maturation markers analyzed (Fig. 1B).

Polymyxin B Modulated the Adhesion Property of DCs-Then
we analyzed the ability of polymyxin B to modulate the adhesion properties of DCs. Indeed, modulation of integrin expression and motility properties have been reported to occur in mature DCs (5,7). In the first few hours of incubation, we observed a progressive increase in the attachment and spreading to the culture plate of DCs treated with polymyxin B or LPS. To quantify this event, after 4 h of incubation cells in suspension or adherent to the culture plate were separately collected and counted, and the percentage of adherent cells was calculated. In several independent experiments with DCs derived from different donors (a representative experiment is shown in Fig. 2A), we consistently observed increased adherence of cells incubated with polymyxin B or LPS with respect to untreated cells. The ability of untreated DCs to spontaneously adhere to the culture plate was variable ranging from a 8.6 -44.6% (median 29.8%, n ϭ 6). Nevertheless, treatment with polymyxin B or LPS always led to a significantly (p Ͻ 0.01) higher percentage of adherent DCs, ranging between 27.3 and 67.1% for polymyxin B (median 47.8%) and between 44.1 and 79.8% for LPS (median 74.2%). The median percent increase in adhesion induced by polymyxin B or LPS compared with untreated DCs was 59.2 and 160.6%, respectively. Treatment with control protein bovine serum albumin at the same molar concentration used for polymyxin B did not alter the adhesion profile ( Fig. 2A).
After 48 h of incubation, a significantly (p Ͻ 0.01) higher percentage of polymyxin B-treated DCs were still adherent to the plate when compared with the untreated control (Fig. 2B). On the contrary, DCs treated with LPS were primarily detached as expected for fully mature DCs that have modified their adhesion properties to migrate into the lymphoid tissues (2,5). Also at this time point, we observed variations in the percentage of adherent untreated DCs derived from different donors (median 19.3%, range 2.9 -58.1%, n ϭ 11). However, the modulation of adherence induced by polymyxin B or LPS was always consistent and statistically significant (p Ͻ 0.01). In fact, the median percentage of adherence for polymyxin B treated DCs was 37.9% (range 9.9 -86.3%), which represents a median increase of 48.5% in comparison to untreated DCs. On the contrary, the median percentage of adherence for LPS treated DCs was 4.7% (range 0.2-38.9%), which represents a median decrease in comparison to untreated DCs of 86.2%. In some experiments (one of which is shown in Fig. 2B) we observed that DCs treated with LPS or polymyxin B were in higher number at 48 h with respect to untreated cells, in agreement with reported observations on prolonged survival of DCs maturing in vitro (10,32,33).
These experiments suggest that polymyxin B is able to induce early adhesion of DC to culture plates, a biological effect common to the process of DC maturation induced by LPS. However, although LPS-treated DCs continue their maturation program and modify their adhesion properties by detaching from the plate, polymyxin B-treated DCs seem to be blocked in the adherent condition.
Polymyxin B-induced Activation of ERK1/2 and IB-␣/ NF-B Pathways-To characterized the signal transduction events involved in the functional and phenotypic modifications induced by polymyxin B, we analyzed its ability to induce activation of ERK1/2, p38, and IB-␣/NF-B pathways, which are implicated in the maturation process of human DCs (10 -13). Polymyxin B is able to induce a transient activation of ERK1/2 (Fig. 3A) with a kinetic similar to that of LPS but at lower level. On the contrary, p38 activation was induced only by LPS (Fig. 3A). Polymyxin B incubation resulted also in the phosphorylation of IB-␣, (a marker of NF-B activation (15,16)) even if at very a low level when compared with LPSinduced phosphorylation of IB-␣ (Fig. 3A). Then we analyzed the activation state of ERK1/2, p38 kinases, and IB-␣ protein at 48 h. We considered this long term analysis feasible because monocyte-derived human DCs do not proliferate in culture (33). As shown in Fig. 3B, a second phase of ERK1/2 activation was observed in DCs treated with either polymyxin B or LPS, whereas p38 and IB-␣/NF-B pathways were activated only by LPS treatment. We found a higher level of IB-␣ in LPStreated DCs with comparison to untreated DCs (Fig. 3B). This feature has been described in mature DCs (11,34), and it correlates with the increase in NF-B activity evaluated as DNA-binding ability (11). We tried to better characterize the kinetics of appearance of the second wave of ERK1/2 activation. However, the time of appearance of phosphorylation was variable in DCs from different donors (data not shown). This observation correlates with the heterogeneity observed in the level of adhesion increase and in the level of CD86 expression induced by polymyxin B stimulation. The ERK kinase predominantly expressed in human DCs was ERK2 (Fig. 3C) as shown by comparison with fibroblast lysates. However, low levels of ERK1 expression were detected in some experiments. Treatment with control protein bovine serum albumin at the same molar concentration used for polymyxin B did not activate ERK1/2 (data not shown). In conclusion, the phenotypic, functional, and molecular results obtained suggest that in human DCs polymyxin B is able to start a maturation process similar to that induced by LPS (characterized by up-regulation of HLA-I, -II, and CD86 molecules, increased cell adhesion, and ERK1/2 and IB-␣/NF-B pathways activation) without being able to sustain and complete it (lack of CD83 up-regulation, of reduced adhesion to the plate, and of p38 and IB-␣/NF-B pathways activation at long term).

Role of Adhesion in Polymyxin B-induced DCs
Maturation-To elucidate the role of polymyxin B-induced adhesion to culture plates in the induction of CD86 up-regulation and signaling pathways activation, we performed DC stimulation in anchorage-independent culture conditions using polyHEMAcovered plates. A direct role for cell adherence to a substrate in ERK1/2 activation has been described to occur in different cellular types (35)(36)(37), but no data are reported for human DCs. In DCs cultured in polyHEMA plates, ERK1/2 was activated at short term but with lower extent with respect to regular culture plates (Fig. 4A), thus suggesting that polymyxin B activates ERK1/2 mainly through the induction of cell adherence to the plate. The IB-␣ phosphorylation induced by polymyxin B was also strongly inhibited in polyHEMA culture condition (Fig. 4B), supporting the hypothesis that adhesion is mediating the activation of the signaling pathways that induce maturation in polymyxin B-treated DCs.
In fact, DCs treated with polymyxin B in the absence of adhesion to the plate showed a strong decrease in the percentage of CD86 positive cells with respect to polymyxin B-treated DCs cultured in regular plate. In a representative experiment shown in Fig. 4C, the percentage of CD86 positive cells de- creased from 65% in regular plates to 21% in polyHEMA plates. Similar results were observed in untreated DCs, where the spontaneous CD86 expression (33% of positive cells) was inhibited in anchorage-independent culture conditions (10% of positive cells). On the contrary, in LPS-treated DCs the absence of adhesion to the plate led to a small decrease in CD86 and CD83 expression (Fig. 4C) with respect to DCs stimulated in regular culture plate. In agreement with these data, when LPS-treated DCs were denied attachment to the plate, activation at short term of ERK1/2 (Fig. 4A) and IB-␣/NF-B pathways (Fig. 4B) was not affected. These results demonstrate that CD86 expression induced by polymyxin B is dependent on DCs adhesion to a substrate and correlate with ERK1/2 and IB-␣/NF-B pathways activation.
Effect of PD98059 on Polymyxin B-induced DCs Maturation-To further investigate the role of ERK1/2 in the maturation process induced by polymyxin B, we analyzed maturation of DCs in the presence of the ERK1/2 inhibitor PD98059. As shown in Fig. 5D, no ERK1/2 activation was detectable at 48 h of incubation with polymyxin B and PD98059, thus demonstrating the specificity of the PD98059 inhibitory activity. Similar results were obtained at short time (data not shown).
Inhibition of ERK1/2 did not alter the adhesive ability of cells incubated with polymyxin B at short term (Fig. 5A). At long term, instead, as shown in a representative experiment in Fig.  5B, DCs incubated with PD98059 and polymyxin B showed a very low percentage of adherence (6%), similar to the adherence of LPS-treated DCs (5%) and much lower than the adherence of DCs treated with polymyxin B alone (39%). Also in DCs treated with polymyxin B and PD98059, as previously observed in LPS-treated DCs, the reduced adherence at 48 h correlated with an increase in the expression of the co-stimulatory molecules CD86 and CD83 (a representative experiment is shown in Fig. 5C) and in the activation of p38 (Fig. 5D). On the contrary, in this last condition no increase of phospho-IB-␣ or total IB-␣ level was observed (Fig. 5D). In cells induced to mature with LPS, ERK1/2 inhibition by PD98059 did not increase the expression of CD86 and CD83 (data not shown), probably because the inhibitory effect of ERK1/2 activation had been already counteracted by the p38 activation. In fact, CD86 and CD83 molecules were already expressed at high percentage in LPS-treated DCs. Treatment with PD98059 or Me 2 SO alone did not alter the CD86 expression and did not activate p38 or IB-␣/NF-B pathways. In conclusion, these experiments demonstrate that inhibition of ERK1/2 in polymyxin B-treated cells promotes the DCs maturation process as revealed by CD86 and CD83 up-regulation, decreased adhesion to the plate, and p38 activation. Effect of TPCK on Polymyxin B-induced DCs Maturation-To further elucidate the role of IB-␣/NF-B pathway in polymyxin B-induced maturation, DCs were incubated with the IB-␣/ NF-B inhibitor TPCK (38,39). The polymyxin B-or LPS-induced IB-␣ phosphorylation at 60 and 90 min was prevented by the preincubation with TPCK (Fig. 6A). Moreover, the increased expression of the maturation marker CD86 induced by polymyxin B was completely blocked by the IB-␣/NF-B inhibitor (Fig. 6B). TPCK had no effects on untreated DCs and only partially blocked the LPS-induced CD86 expression (Fig. 6B). These data confirmed the hypothesis that polymyxin B mediates DC partial maturation by the IB-␣/NF-B pathway.
Analysis of Polymyxin B-treated DC Functionality-To investigate whether the partial maturation induced by polymyxin B alone or in combination with PD98059 modulated the DCs stimulating properties, DCs treated with polymyxin B with or without PD98059 were compared with untreated and LPStreated DCs for their ability to release cytokines involved in lymphocyte activation and to stimulate alloreactive T cells. Supernatants from DCs incubated for 48 h in the different experimental conditions were tested for the presence of TNF-␣, IL-10, and IL-12 (Fig. 7, A-C). Only DCs treated with LPS released into the medium relevant quantities of TNF-␣, IL-12, and IL-10. There was no release of all the cytokines tested from cells treated by polymyxin B alone or in combination with PD98059. Concerning the allostimulatory functions, LPStreated DCs were effective in stimulating T cells to release interferon-␥ and to proliferate (Fig. 7, D and E), whereas DCs treated with polymyxin B or polymyxin B plus PD98059 did not. These results suggest that the partial maturation induced by polymyxin B alone or in combination with PD98059 does not correspond to a functional maturation of DCs.

DISCUSSION
It has been shown that the antibiotic polymyxin B binds to LPS and neutralizes the LPS effects (40 -42). For this property, polymyxin B has been utilized to exclude the presence of potential LPS contaminations in different experimental systems (43)(44)(45) and in particular in experiments intended to evaluate the activity of potential modulators of innate immune cells (19,21,22). In the present study we analyzed the possible molecular, phenotypic, and functional modifications induced directly by polymyxin B in human monocyte-derived DCs.
We demonstrate that polymyxin B is not immunologically inert. In fact, upon incubation with polymyxin B, DCs undergo partial maturation, as revealed by up-regulation of activation markers (HLA class I, II, and CD86 molecules), activation of ERK1/2 at short and long term, activation of the IB-␣/NF-B pathway at short term, and by an increased adhesion to the culture plate that is required for maturation. This partial maturation, however, did not correspond to a functional state as indicated by the absence of cytokines release and the lack of allostimulatory properties of polymyxin B-treated DCs. Polymyxin B-treated DCs differ from DCs treated with LPS, a well characterized DCs maturation stimulus, for the absence of p38 activation and CD83 expression and for the reduced ability to detach from the plate at long term. All these features can be partially restored by inhibition of ERK1/2, thus suggesting that ERK1/2 activation induced during maturation actually inhibits the progression of the maturation process. On the contrary, inhibition of the IB-␣/NF-B pathway completely blocks the maturation induced by polymyxin B.
A correlation between modulation of adhesion molecules and the process of DC maturation has been reported (7). Moreover, adhesion to collagen type 1 has been shown to induce DC maturation (8,9). In this context, we demonstrated that human DCs induced to mature by polymyxin B or LPS are character-ized by a rapid increase of adhesiveness at short term (i.e. 4 h). At late time points (i.e. 48 h) during the maturation process, we observed a decrease adherence only for LPS-treated DCs that correlates with a full maturation, whereas DCs treated with polymyxin B, which undergo a partial maturation, are blocked in the adherent state.
In support of a role of adhesion to a substrate in the maturation process of DCs, we showed that DCs incubated with polymyxin B in anchorage-independent culture conditions were not able to up-regulate CD86 and fully activate ERK1/2 and IB-␣/NF-B pathways, demonstrating that adhesion is required for the activation of early events of the maturation process. However, a strong maturation stimulus like LPS was able to bypass almost completely the adhesion requirement for maturation, probably through the stronger activation of IB-␣/ NF-B pathway and the activation of p38 kinase. Our data together with the evidence reported by Mahnke et al. (8) and Brand et al. (9) suggest that weak stimuli may induce different degree of DCs maturation depending on the features of the surrounding extracellular matrix.
DCs maturation is a multistep process that requires the sequential switching on of several genes (46) and can be assessed by the analysis of different phenotypic and functional features that appear in a temporal order (6). The partial maturation of DCs induced by polymyxin B is an interesting system to study the role of different signal transduction pathways in the induction of early maturation markers without the interference of other pathways activated by stronger stimuli (e.g. LPS) promoting a full maturation. In agreement with reported data obtained with LPS or TNF-␣ treated DCs (11,13), we found that ERK1/2 and IB-␣/NF-B pathways are activated transiently immediately after the stimulation in those DCs that start a maturation process by up-regulating CD86 expression (i.e. DCs stimulated by polymyxin B or LPS). Activation of p38 at short term instead occurs only in LPS-treated DCs that undergo a full maturation process ending with CD83 up-regulation and cytokine secretion. Human monocyte-derived DCs are differentiated, nondividing primary cells (33); the absence of proliferation allowed us to study kinase activation at long term (i.e. 24 -48 h), which reveals secondary, direct, or indirect effects of the maturation stimulus on the signal transduction elements.
We reported for the first time that a second wave of activation occurs at long term and ERK1/2, p38, and IB-␣/NF-B pathways can be found activated even at 48 h of stimulation in DCs undergoing LPS-induced full maturation. In DCs undergoing a partial polymyxin B-induced maturation, only ERK1/2 is activated at long term.
The second wave of activation of the three signaling pathways analyzed is probably induced directly by cytokines secreted by DCs during the first step of maturation and acting in an autocrine fashion. TNF-␣, for example, induces ERK1/2, p38, and IB-␣/NF-B pathways activation in human and murine DCs (11,13). However, we showed that only LPS-stimulated DCs release TNF-␣ in the medium, suggesting a potential role for this cytokine in the late signaling pathways activation only for LPS-treated DCs. Concerning polymyxin B-treated DCs, we cannot exclude that the secretion of other cytokines may be involved in the late activation of ERK1/2. Because polymyxin B treatment significantly increased the adhesive-ness of DCs to a culture plate, we speculated that this event could also mediate the long term activation of ERK1/2. In fact, there is increasing evidence for a role of adhesion and integrin cross-linking in mediating biological effects through ERK1/2 activation in several cell types (35-37, 47, 48). Polymyxin Btreated DCs cultured in polyHEMA plates still showed ERK1/2 phosphorylation at long term (data not shown), suggesting that an adhesion-independent activation of ERK1/2 did occur. However, in anchorage-independent culture conditions, we observed that DCs form intercellular aggregates (data not shown) that in fibroblasts cultured in suspension are reported to form and to induce activation of focal adhesion kinase (49), an upstream activator of the ERK1/2 pathway (50,51). Nevertheless, this cell-to-cell contact is not able to replace cell-to-plate adhesion in mediating the polymyxin B induced maturation.
To assess the role of ERK1/2 and IB-␣/NF-B pathways in DC maturation, we utilized the PD98059 and TPCK inhibitors. We observed that inhibition of ERK1/2 activity enhances the maturation process induced by polymyxin B by leading to reduction in adhesiveness, up-regulation of CD86 and CD83, and activation of p38. These results are in agreement with those obtained by Puig-Kroger et al. (11) and Yanagawa et al. (13), which show that DCs stimulated with TNF-␣ or LPS presented a partial up-regulation of the maturation markers that was increased by PD98059 treatment. On the contrary, inhibition of IB-␣/NF-B pathway completely inhibits polymyxin B-in- duced maturation in agreement with reported data obtained with LPS-treated DCs (10,14). Altogether these results strongly suggest that ERK1/2 activation induced by LPS or polymyxin B is an inhibitory signal arising during the first step of maturation, which has to be inhibited (by PD98059) or counteracted by p38 activation to allow a full maturation process.
Last, we analyzed whether the partial maturation induced by polymyxin B alone or in combination with PD98059 confers a functional state to DCs. The absence of cytokines release and the inability to stimulate an allogeneic immune response in a classical mixed lymphocyte culture both demonstrate that DCs treated by polymyxin B even in the presence of PD98059 are in a state of partial maturation that is not functional at least when evaluated in vitro.
In conclusion, we demonstrated that polymyxin B is able to activate some early steps of the complex process of DCs maturation through the transient IB-␣/NF-B pathways activation. Because of the inability of polymyxin B to sustain the activation of IB-␣/NF-B pathway and to activate p38, ERK1/2 activation blocks further steps of maturation. It will be of interest to analyze whether the maturation mechanism described for polymyxin B is common to other weak stimuli of maturation.